567 files changed, 103959 insertions, 22005 deletions

diff --git a/.gitignore b/.gitignore
index bac9ab97d..ff253c1da 100644
--- a/.gitignore
+++ b/.gitignore
@@ -22,22 +22,27 @@ Makefile
 stamp-h1
 
 # Generated files
-ufo/.tox
-ufo/test/unit/.coverage
+api/examples/__init__.py*
+api/examples/setup.py
+argp-standalone/libargp.a
 contrib/uuid/uuid_types.h
-extras/init.d/glusterd.plist
 extras/init.d/glusterd-Debian
 extras/init.d/glusterd-Redhat
 extras/init.d/glusterd-SuSE
+extras/init.d/glusterd.plist
 extras/ocf/glusterd
 extras/ocf/volume
-glusterfs.spec
+extras/who-wrote-glusterfs/gitdm
 glusterfs-api.pc
-libtool
-xlators/mount/fuse/utils/mount.glusterfs
-xlators/mount/fuse/utils/mount_glusterfs
-argp-standalone/libargp.a
+glusterfs.spec
 glusterfsd/src/glusterfsd
+libgfchangelog.pc
 libglusterfs/src/spec.lex.c
 libglusterfs/src/y.tab.c
 libglusterfs/src/y.tab.h
+libtool
+run-tests.sh
+ufo/.tox
+ufo/test/unit/.coverage
+xlators/mount/fuse/utils/mount.glusterfs
+xlators/mount/fuse/utils/mount_glusterfs
diff --git a/.mailmap b/.mailmap
new file mode 100644
index 000000000..6bcd95dea
--- /dev/null
+++ b/.mailmap
@@ -0,0 +1,31 @@
+# .mailmap, see 'git short-log --help' for details
+#
+# Listing of contributors that filed patches with different email addresses.
+# Format: <name> <main-email> <alias> [<alias> ...]
+#
+
+Amar Tumballi <amarts@redhat.com> <amar@gluster.com> <amar@del.gluster.com>
+Anand Avati <avati@redhat.com> <avati@gluster.com> <avati@dev.gluster.com> <avati@amp.gluster.com> <avati@blackhole.gluster.com>
+Anush Shetty <ashetty@redhat.com> <anush@gluster.com>
+Csaba Henk <csaba@redhat.com> <csaba@gluster.com> <csaba@lowlife.hu> <csaba@zresearch.com>
+Harshavardhana <fharshav@redhat.com> <harsha@gluster.com> <harsha@zresearch.com> <harsha@dev.gluster.com> <harsha@harshavardhana.net>
+Kaleb S. KEITHLEY <kkeithle@redhat.com> <kkeithle@f16node1.kkeithle.usersys.redhat.com>
+Kaushal M <kaushal@redhat.com> <kaushal@gluster.com>
+Kaushik BV <kbudiger@redhat.com> <kaushikbv@gluster.com>
+Krishna Srinivas <ksriniva@redhat.com> <krishna@gluster.com> <krishna@zresearch.com> <krishna@guest-laptop>
+Krishnan Parthasarathi <kparthas@redhat.com> <kp@gluster.com>
+Louis Zuckerman <louiszuckerman@gmail.com> <me@louiszuckerman.com>
+M S Vishwanath Bhat <vbhat@redhat.com> <msvbhat@gmail.com> <vishwanath@gluster.com>
+Pavan Sondur <pavan@gluster.com> <pavan@dev.gluster.com>
+Pete Zaitcev <zaitcev@kotori.zaitcev.us> <zaitcev@yahoo.com>
+Pranith Kumar K <pkarampu@redhat.com> <pranithk@gluster.com>
+Raghavendra Bhat <raghavendra@redhat.com> <raghavendrabhat@gluster.com>
+Raghavendra G <rgowdapp@redhat.com> <raghavendra@gluster.com> <raghavendra@zresearch.com>
+Rahul C S <rahulcs@redhat.com> <rahulcssjce@gmail.com>
+Rajesh Amaravathi <rajesh@redhat.com> <rajesh@gluster.com> <rajesh.amaravathi@gmail.com>
+Shehjar Tikoo <shehjart@gluster.com> <shehjart@zresearch.com>
+Venky Shankar <vshankar@redhat.com> <venky@gluster.com>
+Vijay Bellur <vbellur@redhat.com> <vijay@gluster.com> <vijay@dev.gluster.com>
+Vijaykumar Koppad <vkoppad@redhat.com> <vijaykumar.koppad@gmail.com>
+Vikas Gorur <vikas@gluster.com> <vikas@zresearch.com>
+shishir gowda <sgowda@redhat.com> <shishirng@gluster.com>
diff --git a/Makefile.am b/Makefile.am
index eba8ed12d..598ebb410 100644
--- a/Makefile.am
+++ b/Makefile.am
@@ -3,7 +3,8 @@ EXTRA_DIST = autogen.sh \
 	INSTALL README AUTHORS THANKS NEWS \
 	glusterfs.spec glusterfs-api.pc.in libgfchangelog.pc.in \
 	error-codes.json gf-error-codes.h.template \
-	gen-headers.py
+	gen-headers.py run-tests.sh \
+	$(shell find $(top_srcdir)/tests -type f -print)
 
 SUBDIRS = argp-standalone libglusterfs rpc api xlators glusterfsd \
 	$(FUSERMOUNT_SUBDIR) doc extras cli @SYNCDAEMON_SUBDIR@
diff --git a/api/Makefile.am b/api/Makefile.am
index af437a64d..f0ad1ee97 100644
--- a/api/Makefile.am
+++ b/api/Makefile.am
@@ -1 +1 @@
-SUBDIRS = src
+SUBDIRS = src examples
diff --git a/api/examples/Makefile.am b/api/examples/Makefile.am
index 6048bb1c8..05f40ff53 100644
--- a/api/examples/Makefile.am
+++ b/api/examples/Makefile.am
@@ -1,4 +1,6 @@
-noinst_PROGRAMS = glfsxmp
+EXTRA_PROGRAMS = glfsxmp
 glfsxmp_SOURCES = glfsxmp.c
 glfsxmp_CFLAGS = $(GLFS_CFLAGS) -Wall
-glfsxmp_LDADD = $(GLFS_LIBS)
\ No newline at end of file
+glfsxmp_LDADD = $(GLFS_LIBS) -lrt
+
+EXTRA_DIST = gfapi.py
diff --git a/api/examples/README b/api/examples/README
index 8f0bab697..4d2b521f7 100644
--- a/api/examples/README
+++ b/api/examples/README
@@ -17,3 +17,20 @@ install glusterfs-api RPM.
 Make sure your LDFLAGS includes -L/path/to/lib where libgfapi.so is
 installed and -I/path/to/include/glusterfs where the 'api' directory
 containing the headers are available.
+
+glfsxmp.c
+=========
+
+glfsxmp.c is an example application which uses libgfapi
+
+Compilation Steps For glfsxmp.c
+===============================
+
+1. $./autogen.sh
+2. $./configure
+
+Note: Before running ./configure , as mentioned above, you need to
+      take care of #1 or #2 i.e. pkg-config path or LDFLAGS and
+      -I/<path> with correct values.
+
+3. $make glfsxmp
diff --git a/api/examples/gfapi.py b/api/examples/gfapi.py
index 8d9ed7329..3ac67f4d5 100755
--- a/api/examples/gfapi.py
+++ b/api/examples/gfapi.py
@@ -1,6 +1,7 @@
 #!/usr/bin/python
 
 from ctypes import *
+from ctypes.util import find_library
 import os
 import sys
 import time
@@ -8,9 +9,9 @@ import types
 
 # Looks like ctypes is having trouble with dependencies, so just force them to
 # load with RTLD_GLOBAL until I figure that out.
-glfs = CDLL("libglusterfs.so",RTLD_GLOBAL)
-xdr = CDLL("libgfxdr.so",RTLD_GLOBAL)
-api = CDLL("libgfapi.so",RTLD_GLOBAL)
+glfs = CDLL(find_library("glusterfs"),RTLD_GLOBAL)
+xdr = CDLL(find_library("gfxdr"),RTLD_GLOBAL)
+api = CDLL(find_library("gfapi"),RTLD_GLOBAL)
 
 # Wow, the Linux kernel folks really play nasty games with this structure.  If
 # you look at the man page for stat(2) and then at this definition you'll note
@@ -98,10 +99,12 @@ class File(object):
         def write (self, data, flags=0):
                 return api.glfs_write(self.fd,data,len(data),flags)
 
-	def fallocate (self, mode, offset, len):
-		return api.glfs_fallocate(self.fd, mode, offset, len)
-	def discard (self, offset, len):
-		return api.glfs_discard(self.fd, offset, len)
+        def fallocate (self, mode, offset, len):
+            return api.glfs_fallocate(self.fd, mode, offset, len)
+
+        def discard (self, offset, len):
+            return api.glfs_discard(self.fd, offset, len)
+
 
 class Dir(object):
 
@@ -381,7 +384,7 @@ if __name__ == "__main__":
                 test_setxattr,
                 test_getxattr,
                 test_listxattr,
-		test_fallocate,
+                test_fallocate,
         )
 
         ok_to_fail = (
diff --git a/api/examples/glfsxmp.c b/api/examples/glfsxmp.c
index 644793a4b..600d72fb5 100644
--- a/api/examples/glfsxmp.c
+++ b/api/examples/glfsxmp.c
@@ -1,6 +1,8 @@
 #include <stdio.h>
+#include <stdlib.h>
 #include <errno.h>
 #include "api/glfs.h"
+#include "api/glfs-handles.h"
 #include <string.h>
 #include <time.h>
 
@@ -8,240 +10,1589 @@
 int
 test_dirops (glfs_t *fs)
 {
-	glfs_fd_t *fd = NULL;
-	char buf[512];
-	struct dirent *entry = NULL;
-
-	fd = glfs_opendir (fs, "/");
-	if (!fd) {
-		fprintf (stderr, "/: %s\n", strerror (errno));
-		return -1;
-	}
+        glfs_fd_t *fd = NULL;
+        char buf[512];
+        struct dirent *entry = NULL;
+
+        fd = glfs_opendir (fs, "/");
+        if (!fd) {
+                fprintf (stderr, "/: %s\n", strerror (errno));
+                return -1;
+        }
+
+        fprintf (stderr, "Entries:\n");
+        while (glfs_readdir_r (fd, (struct dirent *)buf, &entry), entry) {
+                fprintf (stderr, "%s: %lu\n", entry->d_name, glfs_telldir (fd));
+        }
+
+        glfs_closedir (fd);
+        return 0;
+}
 
-	fprintf (stderr, "Entries:\n");
-	while (glfs_readdir_r (fd, (struct dirent *)buf, &entry), entry) {
-		fprintf (stderr, "%s: %lu\n", entry->d_name, glfs_telldir (fd));
-	}
 
-	glfs_closedir (fd);
-	return 0;
+int
+test_xattr (glfs_t *fs)
+{
+        char *filename = "/filename2";
+        char buf[512];
+        char *ptr;
+        int ret;
+
+        ret = glfs_setxattr (fs, filename, "user.testkey", "testval", 8, 0);
+        fprintf (stderr, "setxattr(%s): %d (%s)\n", filename, ret,
+                 strerror (errno));
+
+        ret = glfs_setxattr (fs, filename, "user.testkey2", "testval", 8, 0);
+        fprintf (stderr, "setxattr(%s): %d (%s)\n", filename, ret,
+                 strerror (errno));
+
+        ret = glfs_listxattr (fs, filename, buf, 512);
+        fprintf (stderr, "listxattr(%s): %d (%s)\n", filename, ret,
+                 strerror (errno));
+        if (ret < 0)
+                return -1;
+
+        for (ptr = buf; ptr < buf + ret; ptr++) {
+                printf ("key=%s\n", ptr);
+                ptr += strlen (ptr);
+        }
+
+        return 0;
 }
 
 
 int
-test_xattr (glfs_t *fs)
+test_chdir (glfs_t *fs)
+{
+        int ret = -1;
+        char *topdir = "/topdir";
+        char *linkdir = "/linkdir";
+        char *subdir = "./subdir";
+        char *respath = NULL;
+        char pathbuf[4096];
+
+        ret = glfs_mkdir (fs, topdir, 0755);
+        if (ret) {
+                fprintf (stderr, "mkdir(%s): %s\n", topdir, strerror (errno));
+                return -1;
+        }
+
+        respath = glfs_getcwd (fs, pathbuf, 4096);
+        fprintf (stdout, "getcwd() = %s\n", respath);
+
+        ret = glfs_symlink (fs, topdir, linkdir);
+        if (ret) {
+                fprintf (stderr, "symlink(%s, %s): %s\n", topdir, linkdir, strerror (errno));
+                return -1;
+        }
+
+        ret = glfs_chdir (fs, linkdir);
+        if (ret) {
+                fprintf (stderr, "chdir(%s): %s\n", linkdir, strerror (errno));
+                return -1;
+        }
+
+        respath = glfs_getcwd (fs, pathbuf, 4096);
+        fprintf (stdout, "getcwd() = %s\n", respath);
+
+        respath = glfs_realpath (fs, subdir, pathbuf);
+        if (respath) {
+                fprintf (stderr, "realpath(%s) worked unexpectedly: %s\n", subdir, respath);
+                return -1;
+        }
+
+        ret = glfs_mkdir (fs, subdir, 0755);
+        if (ret) {
+                fprintf (stderr, "mkdir(%s): %s\n", subdir, strerror (errno));
+                return -1;
+        }
+
+        respath = glfs_realpath (fs, subdir, pathbuf);
+        if (!respath) {
+                fprintf (stderr, "realpath(%s): %s\n", subdir, strerror (errno));
+        } else {
+                fprintf (stdout, "realpath(%s) = %s\n", subdir, respath);
+        }
+
+        ret = glfs_chdir (fs, subdir);
+        if (ret) {
+                fprintf (stderr, "chdir(%s): %s\n", subdir, strerror (errno));
+                return -1;
+        }
+
+        respath = glfs_getcwd (fs, pathbuf, 4096);
+        fprintf (stdout, "getcwd() = %s\n", respath);
+
+        respath = glfs_realpath (fs, "/linkdir/subdir", pathbuf);
+        if (!respath) {
+                fprintf (stderr, "realpath(/linkdir/subdir): %s\n", strerror (errno));
+        } else {
+                fprintf (stdout, "realpath(/linkdir/subdir) = %s\n", respath);
+        }
+
+        return 0;
+}
+
+#ifdef DEBUG
+static void
+peek_stat (struct stat *sb)
+{
+        printf ("Dumping stat information:\n");
+        printf ("File type:                ");
+
+        switch (sb->st_mode & S_IFMT) {
+                case S_IFBLK:  printf ("block device\n");            break;
+                case S_IFCHR:  printf ("character device\n");        break;
+                case S_IFDIR:  printf ("directory\n");               break;
+                case S_IFIFO:  printf ("FIFO/pipe\n");               break;
+                case S_IFLNK:  printf ("symlink\n");                 break;
+                case S_IFREG:  printf ("regular file\n");            break;
+                case S_IFSOCK: printf ("socket\n");                  break;
+                default:       printf ("unknown?\n");                break;
+        }
+
+        printf ("I-node number:            %ld\n", (long) sb->st_ino);
+
+        printf ("Mode:                     %lo (octal)\n",
+                (unsigned long) sb->st_mode);
+
+        printf ("Link count:               %ld\n", (long) sb->st_nlink);
+        printf ("Ownership:                UID=%ld   GID=%ld\n",
+                (long) sb->st_uid, (long) sb->st_gid);
+
+        printf ("Preferred I/O block size: %ld bytes\n",
+                (long) sb->st_blksize);
+        printf ("File size:                %lld bytes\n",
+                (long long) sb->st_size);
+        printf ("Blocks allocated:         %lld\n",
+                (long long) sb->st_blocks);
+
+        printf ("Last status change:       %s", ctime(&sb->st_ctime));
+        printf ("Last file access:         %s", ctime(&sb->st_atime));
+        printf ("Last file modification:   %s", ctime(&sb->st_mtime));
+
+        return;
+}
+
+static void
+peek_handle (unsigned char *glid)
+{
+        int i;
+
+        for (i = 0; i < GFAPI_HANDLE_LENGTH; i++)
+        {
+                printf (":%02x:", glid[i]);
+        }
+        printf ("\n");
+}
+#else /* DEBUG */
+static void
+peek_stat (struct stat *sb)
+{
+        return;
+}
+
+static void
+peek_handle (unsigned char *id)
 {
-	char *filename = "/filename2";
-	char buf[512];
-	char *ptr;
-	int ret;
+        return;
+}
+#endif /* DEBUG */
 
-	ret = glfs_setxattr (fs, filename, "user.testkey", "testval", 8, 0);
-	fprintf (stderr, "setxattr(%s): %d (%s)\n", filename, ret,
-		 strerror (errno));
+glfs_t    *fs = NULL;
+char      *full_parent_name = "/testdir", *parent_name = "testdir";
 
-	ret = glfs_setxattr (fs, filename, "user.testkey2", "testval", 8, 0);
-	fprintf (stderr, "setxattr(%s): %d (%s)\n", filename, ret,
-		 strerror (errno));
+void
+test_h_unlink (void)
+{
+        char               *my_dir = "unlinkdir";
+        char               *my_file = "file.txt";
+        char               *my_subdir = "dir1";
+        struct glfs_object *parent = NULL, *leaf = NULL, *dir = NULL,
+                           *subdir = NULL, *subleaf = NULL;
+        struct stat         sb;
+        int                 ret;
+
+        printf ("glfs_h_unlink tests: In Progress\n");
+
+        /* Prepare tests */
+        parent = glfs_h_lookupat (fs, NULL, full_parent_name, &sb);
+        if (parent == NULL) {
+                fprintf (stderr, "glfs_h_lookupat: error on lookup of %s: from (%p),%s\n",
+                         full_parent_name, NULL, strerror (errno));
+                printf ("glfs_h_lookupat tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        dir = glfs_h_mkdir (fs, parent, my_dir, 0644, &sb);
+        if (dir == NULL) {
+                fprintf (stderr, "glfs_h_mkdir: error creating %s: from (%p),%s\n",
+                         my_dir, parent, strerror (errno));
+                printf ("glfs_h_unlink tests: FAILED\n");
+                goto out;
+        }
+
+        leaf = glfs_h_creat (fs, dir, my_file, O_CREAT, 0644, &sb);
+        if (leaf == NULL) {
+                fprintf (stderr, "glfs_h_creat: error creating %s: from (%p),%s\n",
+                         my_file, dir, strerror (errno));
+                printf ("glfs_h_unlink tests: FAILED\n");
+                goto out;
+        }
+
+        subdir = glfs_h_mkdir (fs, dir, my_subdir, 0644, &sb);
+        if (subdir == NULL) {
+                fprintf (stderr, "glfs_h_mkdir: error creating %s: from (%p),%s\n",
+                         my_subdir, dir, strerror (errno));
+                printf ("glfs_h_unlink tests: FAILED\n");
+                goto out;
+        }
+
+        subleaf = glfs_h_creat (fs, subdir, my_file, O_CREAT, 0644, &sb);
+        if (subleaf == NULL) {
+                fprintf (stderr, "glfs_h_creat: error creating %s: from (%p),%s\n",
+                         my_file, subdir, strerror (errno));
+                printf ("glfs_h_unlink tests: FAILED\n");
+                goto out;
+        }
+
+        /* unlink non empty directory */
+        ret = glfs_h_unlink (fs, dir, my_subdir);
+        if ((ret && errno != ENOTEMPTY) || (ret == 0)) {
+                fprintf (stderr, "glfs_h_unlink: error unlinking %s: it is non empty: %s\n",
+                         my_subdir, strerror (errno));
+                printf ("glfs_h_unlink tests: FAILED\n");
+                goto out;
+        }
+
+        /* unlink regular file */
+        ret = glfs_h_unlink (fs, subdir, my_file);
+        if (ret) {
+                fprintf (stderr, "glfs_h_unlink: error unlinking %s: from (%p),%s\n",
+                         my_file, subdir, strerror (errno));
+                printf ("glfs_h_unlink tests: FAILED\n");
+                goto out;
+        }
+
+        /* unlink directory */
+        ret = glfs_h_unlink (fs, dir, my_subdir);
+        if (ret) {
+                fprintf (stderr, "glfs_h_unlink: error unlinking %s: from (%p),%s\n",
+                         my_subdir, dir, strerror (errno));
+                printf ("glfs_h_unlink tests: FAILED\n");
+                goto out;
+        }
+
+        /* unlink regular file */
+        ret = glfs_h_unlink (fs, dir, my_file);
+        if (ret) {
+                fprintf (stderr, "glfs_h_unlink: error unlinking %s: from (%p),%s\n",
+                         my_file, dir, strerror (errno));
+                printf ("glfs_h_unlink tests: FAILED\n");
+                goto out;
+        }
+
+        /* unlink non-existant regular file */
+        ret = glfs_h_unlink (fs, dir, my_file);
+        if ((ret && errno != ENOENT) || (ret == 0)) {
+                fprintf (stderr, "glfs_h_unlink: error unlinking non-existant %s: invalid errno ,%d, %s\n",
+                         my_file, ret, strerror (errno));
+                printf ("glfs_h_unlink tests: FAILED\n");
+                goto out;
+        }
+
+        /* unlink non-existant directory */
+        ret = glfs_h_unlink (fs, dir, my_subdir);
+        if ((ret && errno != ENOENT) || (ret == 0)) {
+                fprintf (stderr, "glfs_h_unlink: error unlinking non-existant %s:  invalid errno ,%d, %s\n",
+                         my_subdir, ret, strerror (errno));
+                printf ("glfs_h_unlink tests: FAILED\n");
+                goto out;
+        }
+
+        /* unlink directory */
+        ret = glfs_h_unlink (fs, parent, my_dir);
+        if (ret) {
+                fprintf (stderr, "glfs_h_unlink: error unlinking %s: from (%p),%s\n",
+                         my_dir, dir, strerror (errno));
+                printf ("glfs_h_unlink tests: FAILED\n");
+                goto out;
+        }
+
+        printf ("glfs_h_unlink tests: PASSED\n");
+
+out:
+        if (dir)
+                glfs_h_close (dir);
+        if (leaf)
+                glfs_h_close (leaf);
+        if (subdir)
+                glfs_h_close (subdir);
+        if (subleaf)
+                glfs_h_close (subleaf);
+        if (parent)
+                glfs_h_close (parent);
+
+        return;
+}
 
-	ret = glfs_listxattr (fs, filename, buf, 512);
-	fprintf (stderr, "listxattr(%s): %d (%s)\n", filename, ret,
-		 strerror (errno));
-	if (ret < 0)
-		return -1;
+void
+test_h_getsetattrs (void)
+{
+        char               *my_dir = "attrdir";
+        char               *my_file = "attrfile.txt";
+        struct glfs_object *parent = NULL, *leaf = NULL, *dir = NULL;
+        struct stat         sb, retsb;
+        int                 ret, valid;
+        struct timespec     timestamp;
+
+        printf("glfs_h_getattrs and setattrs tests: In Progress\n");
+
+        /* Prepare tests */
+        parent = glfs_h_lookupat (fs, NULL, full_parent_name, &sb);
+        if (parent == NULL) {
+                fprintf (stderr, "glfs_h_lookupat: error on lookup of %s: from (%p),%s\n",
+                         full_parent_name, NULL, strerror (errno));
+                printf ("glfs_h_lookupat tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        dir = glfs_h_mkdir (fs, parent, my_dir, 0644, &sb);
+        if (dir == NULL) {
+                fprintf (stderr, "glfs_h_mkdir: error creating %s: from (%p),%s\n",
+                         my_dir, parent, strerror (errno));
+                printf ("glfs_h_unlink tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        leaf = glfs_h_creat (fs, dir, my_file, O_CREAT, 0644, &sb);
+        if (leaf == NULL) {
+                fprintf (stderr, "glfs_h_creat: error creating %s: from (%p),%s\n",
+                         my_file, dir, strerror (errno));
+                printf ("glfs_h_unlink tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        ret = glfs_h_getattrs (fs, dir, &retsb);
+        if (ret != 0) {
+                fprintf (stderr, "glfs_h_getattrs: error %s: from (%p),%s\n",
+                         my_dir, dir, strerror (errno));
+                printf ("glfs_h_getattrs and setattrs tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&retsb);
+        /* TODO: Compare stat information */
+
+        retsb.st_mode = 00666;
+        retsb.st_uid = 1000;
+        retsb.st_gid = 1001;
+        ret = clock_gettime (CLOCK_REALTIME, &timestamp);
+        if(ret != 0) {
+                fprintf (stderr, "clock_gettime: error %s\n", strerror (errno));
+                printf ("glfs_h_getattrs and setattrs tests: FAILED\n");
+                goto out;
+        }
+        retsb.st_atim = timestamp;
+        retsb.st_mtim = timestamp;
+        valid = GFAPI_SET_ATTR_MODE | GFAPI_SET_ATTR_UID | GFAPI_SET_ATTR_GID |
+        GFAPI_SET_ATTR_ATIME | GFAPI_SET_ATTR_MTIME;
+        peek_stat (&retsb);
+
+        ret = glfs_h_setattrs (fs, dir, &retsb, valid);
+        if (ret != 0) {
+                fprintf (stderr, "glfs_h_setattrs: error %s: from (%p),%s\n",
+                         my_dir, dir, strerror (errno));
+                printf ("glfs_h_getattrs and setattrs tests: FAILED\n");
+                goto out;
+        }
+
+        memset(&retsb, 0, sizeof (struct stat));
+        ret = glfs_h_stat (fs, dir, &retsb);
+        if (ret != 0) {
+                fprintf (stderr, "glfs_h_stat: error %s: from (%p),%s\n",
+                         my_dir, dir, strerror (errno));
+                printf ("glfs_h_getattrs and setattrs tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&retsb);
+
+        printf ("glfs_h_getattrs and setattrs tests: PASSED\n");
+out:
+        if (parent)
+                glfs_h_close (parent);
+        if (leaf)
+                glfs_h_close (leaf);
+        if (dir)
+                glfs_h_close (dir);
+
+        return;
+}
 
-	for (ptr = buf; ptr < buf + ret; ptr++) {
-		printf ("key=%s\n", ptr);
-		ptr += strlen (ptr);
-	}
+void
+test_h_truncate (void)
+{
+        char               *my_dir = "truncatedir";
+        char               *my_file = "file.txt";
+        struct glfs_object *root = NULL, *parent = NULL, *leaf = NULL;
+        struct stat         sb;
+        glfs_fd_t          *fd = NULL;
+        char                buf[32];
+        off_t               offset = 0;
+        int                 ret = 0;
+
+        printf("glfs_h_truncate tests: In Progress\n");
+
+        /* Prepare tests */
+        root = glfs_h_lookupat (fs, NULL, full_parent_name, &sb);
+        if (root == NULL) {
+                fprintf (stderr, "glfs_h_lookupat: error on lookup of %s: from (%p),%s\n",
+                         full_parent_name, NULL, strerror (errno));
+                printf ("glfs_h_truncate tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        parent = glfs_h_mkdir (fs, root, my_dir, 0644, &sb);
+        if (parent == NULL) {
+                fprintf (stderr, "glfs_h_mkdir: error creating %s: from (%p),%s\n",
+                         my_dir, root, strerror (errno));
+                printf ("glfs_h_truncate tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        leaf = glfs_h_creat (fs, parent, my_file, O_CREAT, 0644, &sb);
+        if (leaf == NULL) {
+                fprintf (stderr, "glfs_h_creat: error creating %s: from (%p),%s\n",
+                         my_file, parent, strerror (errno));
+                printf ("glfs_h_truncate tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        fd = glfs_h_open (fs, leaf, O_RDWR);
+        if (fd == NULL) {
+                fprintf (stderr, "glfs_h_open: error on open of %s: %s\n",
+                         my_file, strerror (errno));
+                printf ("glfs_h_truncate tests: FAILED\n");
+                goto out;
+        }
+
+        memcpy (buf, "abcdefghijklmnopqrstuvwxyz012345", 32);
+        ret = glfs_write (fd, buf, 32, 0);
+
+        /* run tests */
+        /* truncate lower */
+        offset = 30;
+        ret = glfs_h_truncate (fs, leaf, offset);
+        if (ret != 0) {
+                fprintf (stderr, "glfs_h_truncate: error creating %s: from (%p),%s\n",
+                         my_file, parent, strerror (errno));
+                printf ("glfs_h_truncate tests: FAILED\n");
+                goto out;
+        }
+        ret = glfs_h_getattrs (fs, leaf, &sb);
+        if (ret != 0) {
+                fprintf (stderr, "glfs_h_getattrs: error for %s (%p),%s\n",
+                         my_file, leaf, strerror (errno));
+                printf ("glfs_h_truncate tests: FAILED\n");
+                goto out;
+        }
+        if (sb.st_size != offset) {
+                fprintf (stderr, "glfs_h_truncate: post size mismatch\n");
+                printf ("glfs_h_truncate tests: FAILED\n");
+                goto out;
+        }
+
+        /* truncate higher */
+        offset = 32;
+        ret = glfs_h_truncate (fs, leaf, offset);
+        if (ret != 0) {
+                fprintf (stderr, "glfs_h_truncate: error creating %s: from (%p),%s\n",
+                         my_file, parent, strerror (errno));
+                printf ("glfs_h_truncate tests: FAILED\n");
+                goto out;
+        }
+        ret = glfs_h_getattrs (fs, leaf, &sb);
+        if (ret != 0) {
+                fprintf (stderr, "glfs_h_getattrs: error for %s (%p),%s\n",
+                         my_file, leaf, strerror (errno));
+                printf ("glfs_h_truncate tests: FAILED\n");
+                goto out;
+        }
+        if (sb.st_size != offset) {
+                fprintf (stderr, "glfs_h_truncate: post size mismatch\n");
+                printf ("glfs_h_truncate tests: FAILED\n");
+                goto out;
+        }
+
+        /* truncate equal */
+        offset = 30;
+        ret = glfs_h_truncate (fs, leaf, offset);
+        if (ret != 0) {
+                fprintf (stderr, "glfs_h_truncate: error creating %s: from (%p),%s\n",
+                         my_file, parent, strerror (errno));
+                printf ("glfs_h_truncate tests: FAILED\n");
+                goto out;
+        }
+        ret = glfs_h_getattrs (fs, leaf, &sb);
+        if (ret != 0) {
+                fprintf (stderr, "glfs_h_getattrs: error for %s (%p),%s\n",
+                         my_file, leaf, strerror (errno));
+                printf ("glfs_h_truncate tests: FAILED\n");
+                goto out;
+        }
+        if (sb.st_size != offset) {
+                fprintf (stderr, "glfs_h_truncate: post size mismatch\n");
+                printf ("glfs_h_truncate tests: FAILED\n");
+                goto out;
+        }
+
+        printf ("glfs_h_truncate tests: PASSED\n");
+out:
+        if (fd)
+                glfs_close (fd);
+        if (root)
+                glfs_h_close (root);
+        if (parent)
+                glfs_h_close (parent);
+        if (leaf)
+                glfs_h_close (leaf);
+
+        return;
+}
 
-	return 0;
+void
+test_h_links (void)
+{
+        char               *my_dir = "linkdir";
+        char               *my_file = "file.txt";
+        char               *my_symlnk = "slnk.txt";
+        char               *my_lnk = "lnk.txt";
+        char               *linksrc_dir = "dir1";
+        char               *linktgt_dir = "dir2";
+        struct glfs_object *root = NULL, *parent = NULL, *leaf = NULL,
+                           *dirsrc = NULL, *dirtgt = NULL, *dleaf = NULL;
+        struct glfs_object *ln1 = NULL;
+        struct stat         sb;
+        int                 ret;
+        char               *buf = NULL;
+
+        printf("glfs_h_link(s) tests: In Progress\n");
+
+        /* Prepare tests */
+        root = glfs_h_lookupat (fs, NULL, full_parent_name, &sb);
+        if (root == NULL) {
+                fprintf (stderr, "glfs_h_lookupat: error on lookup of %s: from (%p),%s\n",
+                         full_parent_name, NULL, strerror (errno));
+                printf ("glfs_h_link(s) tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        parent = glfs_h_mkdir (fs, root, my_dir, 0644, &sb);
+        if (parent == NULL) {
+                fprintf (stderr, "glfs_h_mkdir: error creating %s: from (%p),%s\n",
+                         my_dir, root, strerror (errno));
+                printf ("glfs_h_link(s) tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        leaf = glfs_h_creat (fs, parent, my_file, O_CREAT, 0644, &sb);
+        if (leaf == NULL) {
+                fprintf (stderr, "glfs_h_creat: error creating %s: from (%p),%s\n",
+                         my_file, parent, strerror (errno));
+                printf ("glfs_h_link(s) tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        dirsrc = glfs_h_mkdir (fs, parent, linksrc_dir, 0644, &sb);
+        if (dirsrc == NULL) {
+                fprintf (stderr, "glfs_h_mkdir: error creating %s: from (%p),%s\n",
+                         linksrc_dir, parent, strerror (errno));
+                printf ("glfs_h_link(s) tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        dirtgt = glfs_h_mkdir (fs, parent, linktgt_dir, 0644, &sb);
+        if (dirtgt == NULL) {
+                fprintf (stderr, "glfs_h_mkdir: error creating %s: from (%p),%s\n",
+                         linktgt_dir, parent, strerror (errno));
+                printf ("glfs_h_link(s) tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        dleaf = glfs_h_creat (fs, dirsrc, my_file, O_CREAT, 0644, &sb);
+        if (dleaf == NULL) {
+                fprintf (stderr, "glfs_h_creat: error creating %s: from (%p),%s\n",
+                         my_file, dirsrc, strerror (errno));
+                printf ("glfs_h_link(s) tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        /* run tests */
+        /* sym link: /testdir/linkdir/file.txt to ./slnk.txt */
+        ln1 = glfs_h_symlink (fs, parent, my_symlnk, "./file.txt", &sb);
+        if (ln1 == NULL) {
+                fprintf (stderr, "glfs_h_symlink: error creating %s: from (%p),%s\n",
+                         my_symlnk, parent, strerror (errno));
+                printf ("glfs_h_link(s) tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        buf = calloc (1024, sizeof(char));
+        if (buf == NULL) {
+                fprintf (stderr, "Error allocating memory\n");
+                printf ("glfs_h_link(s) tests: FAILED\n");
+                goto out;
+        }
+
+        ret = glfs_h_readlink (fs, ln1, buf, 1024);
+        if (ret <= 0) {
+                fprintf (stderr, "glfs_h_readlink: error reading %s: from (%p),%s\n",
+                         my_symlnk, ln1, strerror (errno));
+                printf ("glfs_h_link(s) tests: FAILED\n");
+                goto out;
+        }
+        if (!(strncmp (buf, my_symlnk, strlen (my_symlnk)))) {
+                fprintf (stderr, "glfs_h_readlink: error mismatch in link name: actual %s: retrieved %s\n",
+                         my_symlnk, buf);
+                printf ("glfs_h_link(s) tests: FAILED\n");
+                goto out;
+        }
+
+        /* link: /testdir/linkdir/file.txt to ./lnk.txt */
+        ret = glfs_h_link (fs, leaf, parent, my_lnk);
+        if (ret != 0) {
+                fprintf (stderr, "glfs_h_link: error creating %s: from (%p),%s\n",
+                         my_lnk, parent, strerror (errno));
+                printf ("glfs_h_link(s) tests: FAILED\n");
+                goto out;
+        }
+        /* TODO: Should write content to a file and read from the link */
+
+        /* link: /testdir/linkdir/dir1/file.txt to ../dir2/slnk.txt */
+        ret = glfs_h_link (fs, dleaf, dirtgt, my_lnk);
+        if (ret != 0) {
+                fprintf (stderr, "glfs_h_link: error creating %s: from (%p),%s\n",
+                         my_lnk, dirtgt, strerror (errno));
+                printf ("glfs_h_link(s) tests: FAILED\n");
+                goto out;
+        }
+        /* TODO: Should write content to a file and read from the link */
+
+        printf ("glfs_h_link(s) tests: PASSED\n");
+
+out:
+        if (root)
+                glfs_h_close (root);
+        if (parent)
+                glfs_h_close (parent);
+        if (leaf)
+                glfs_h_close (leaf);
+        if (dirsrc)
+                glfs_h_close (dirsrc);
+        if (dirtgt)
+                glfs_h_close (dirtgt);
+        if (dleaf)
+                glfs_h_close (dleaf);
+        if (ln1)
+                glfs_h_close (ln1);
+        if (buf)
+                free (buf);
+
+        return;
 }
 
+void
+test_h_rename (void)
+{
+        char               *my_dir = "renamedir";
+        char               *my_file = "file.txt";
+        char               *src_dir = "dir1";
+        char               *tgt_dir = "dir2";
+        struct glfs_object *root = NULL, *parent = NULL, *leaf = NULL,
+                           *dirsrc = NULL, *dirtgt = NULL, *dleaf = NULL;
+        struct stat         sb;
+        int                 ret;
+
+        printf("glfs_h_rename tests: In Progress\n");
+
+        /* Prepare tests */
+        root = glfs_h_lookupat (fs, NULL, full_parent_name, &sb);
+        if (root == NULL) {
+                fprintf (stderr, "glfs_h_lookupat: error on lookup of %s: from (%p),%s\n",
+                         full_parent_name, NULL, strerror (errno));
+                printf ("glfs_h_rename tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        parent = glfs_h_mkdir (fs, root, my_dir, 0644, &sb);
+        if (parent == NULL) {
+                fprintf (stderr, "glfs_h_mkdir: error creating %s: from (%p),%s\n",
+                         my_dir, root, strerror (errno));
+                printf ("glfs_h_rename tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        leaf = glfs_h_creat (fs, parent, my_file, O_CREAT, 0644, &sb);
+        if (leaf == NULL) {
+                fprintf (stderr, "glfs_h_creat: error creating %s: from (%p),%s\n",
+                         my_file, parent, strerror (errno));
+                printf ("glfs_h_rename tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        dirsrc = glfs_h_mkdir (fs, parent, src_dir, 0644, &sb);
+        if (dirsrc == NULL) {
+                fprintf (stderr, "glfs_h_mkdir: error creating %s: from (%p),%s\n",
+                         src_dir, parent, strerror (errno));
+                printf ("glfs_h_rename tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        dirtgt = glfs_h_mkdir (fs, parent, tgt_dir, 0644, &sb);
+        if (dirtgt == NULL) {
+                fprintf (stderr, "glfs_h_mkdir: error creating %s: from (%p),%s\n",
+                         tgt_dir, parent, strerror (errno));
+                printf ("glfs_h_rename tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        dleaf = glfs_h_creat (fs, dirsrc, my_file, O_CREAT, 0644, &sb);
+        if (dleaf == NULL) {
+                fprintf (stderr, "glfs_h_creat: error creating %s: from (%p),%s\n",
+                         my_file, dirsrc, strerror (errno));
+                printf ("glfs_h_rename tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        /* run tests */
+        /* Rename file.txt -> file1.txt */
+        ret = glfs_h_rename (fs, parent, "file.txt", parent, "file1.txt");
+        if (ret != 0) {
+                fprintf (stderr, "glfs_h_rename: error renaming %s to %s (%s)\n",
+                         "file.txt", "file1.txt", strerror (errno));
+                printf ("glfs_h_rename tests: FAILED\n");
+                goto out;
+        }
+
+        /* rename dir1/file.txt -> file.txt */
+        ret = glfs_h_rename (fs, dirsrc, "file.txt", parent, "file.txt");
+        if (ret != 0) {
+                fprintf (stderr, "glfs_h_rename: error renaming %s/%s to %s (%s)\n",
+                         src_dir, "file.txt", "file.txt", strerror (errno));
+                printf ("glfs_h_rename tests: FAILED\n");
+                goto out;
+        }
+
+        /* rename file1.txt -> file.txt (exists) */
+        ret = glfs_h_rename (fs, parent, "file1.txt", parent, "file.txt");
+        if (ret != 0) {
+                fprintf (stderr, "glfs_h_rename: error renaming %s to %s (%s)\n",
+                         "file.txt", "file.txt", strerror (errno));
+                printf ("glfs_h_rename tests: FAILED\n");
+                goto out;
+        }
+
+        /* rename dir1 -> dir3 */
+        ret = glfs_h_rename (fs, parent, "dir1", parent, "dir3");
+        if (ret != 0) {
+                fprintf (stderr, "glfs_h_rename: error renaming %s to %s (%s)\n",
+                         "dir1", "dir3", strerror (errno));
+                printf ("glfs_h_rename tests: FAILED\n");
+                goto out;
+        }
+
+        /* rename dir2 ->dir3 (exists) */
+        ret = glfs_h_rename (fs, parent, "dir2", parent, "dir3");
+        if (ret != 0) {
+                fprintf (stderr, "glfs_h_rename: error renaming %s to %s (%s)\n",
+                         "dir2", "dir3", strerror (errno));
+                printf ("glfs_h_rename tests: FAILED\n");
+                goto out;
+        }
+
+        /* rename file.txt -> dir3 (fail) */
+        ret = glfs_h_rename (fs, parent, "file.txt", parent, "dir3");
+        if (ret == 0) {
+                fprintf (stderr, "glfs_h_rename: NO error renaming %s to %s (%s)\n",
+                         "file.txt", "dir3", strerror (errno));
+                printf ("glfs_h_rename tests: FAILED\n");
+                goto out;
+        }
+
+        /* rename dir3 -> file.txt (fail) */
+        ret = glfs_h_rename (fs, parent, "dir3", parent, "file.txt");
+        if (ret == 0) {
+                fprintf (stderr, "glfs_h_rename: NO error renaming %s to %s (%s)\n",
+                         "dir3", "file.txt", strerror (errno));
+                printf ("glfs_h_rename tests: FAILED\n");
+                goto out;
+        }
+
+        printf ("glfs_h_rename tests: PASSED\n");
+
+out:
+        if (root)
+                glfs_h_close (root);
+        if (parent)
+                glfs_h_close (parent);
+        if (leaf)
+                glfs_h_close (leaf);
+        if (dirsrc)
+                glfs_h_close (dirsrc);
+        if (dirtgt)
+                glfs_h_close (dirtgt);
+        if (dleaf)
+                glfs_h_close (dleaf);
+
+        return;
+}
 
-int
-test_chdir (glfs_t *fs)
+void
+assimilatetime (struct timespec *ts, struct timespec ts_st,
+                struct timespec ts_ed)
 {
-	int ret = -1;
-	ino_t ino = 0;
-	struct stat st;
-	char *topdir = "/topdir";
-	char *linkdir = "/linkdir";
-	char *subdir = "./subdir";
-	char *respath = NULL;
-	char pathbuf[4096];
-
-	ret = glfs_mkdir (fs, topdir, 0755);
-	if (ret) {
-		fprintf (stderr, "mkdir(%s): %s\n", topdir, strerror (errno));
-		return -1;
-	}
-
-	respath = glfs_getcwd (fs, pathbuf, 4096);
-	fprintf (stdout, "getcwd() = %s\n", respath);
-
-	ret = glfs_symlink (fs, topdir, linkdir);
-	if (ret) {
-		fprintf (stderr, "symlink(%s, %s): %s\n", topdir, linkdir, strerror (errno));
-		return -1;
-	}
-
-	ret = glfs_chdir (fs, linkdir);
-	if (ret) {
-		fprintf (stderr, "chdir(%s): %s\n", linkdir, strerror (errno));
-		return -1;
-	}
-
-	respath = glfs_getcwd (fs, pathbuf, 4096);
-	fprintf (stdout, "getcwd() = %s\n", respath);
-
-	respath = glfs_realpath (fs, subdir, pathbuf);
-	if (respath) {
-		fprintf (stderr, "realpath(%s) worked unexpectedly: %s\n", subdir, respath);
-		return -1;
-	}
-
-	ret = glfs_mkdir (fs, subdir, 0755);
-	if (ret) {
-		fprintf (stderr, "mkdir(%s): %s\n", subdir, strerror (errno));
-		return -1;
-	}
-
-	respath = glfs_realpath (fs, subdir, pathbuf);
-	if (!respath) {
-		fprintf (stderr, "realpath(%s): %s\n", subdir, strerror (errno));
-	} else {
-		fprintf (stdout, "realpath(%s) = %s\n", subdir, respath);
-	}
-
-	ret = glfs_chdir (fs, subdir);
-	if (ret) {
-		fprintf (stderr, "chdir(%s): %s\n", subdir, strerror (errno));
-		return -1;
-	}
-
-	respath = glfs_getcwd (fs, pathbuf, 4096);
-	fprintf (stdout, "getcwd() = %s\n", respath);
-
-	respath = glfs_realpath (fs, "/linkdir/subdir", pathbuf);
-	if (!respath) {
-		fprintf (stderr, "realpath(/linkdir/subdir): %s\n", strerror (errno));
-	} else {
-		fprintf (stdout, "realpath(/linkdir/subdir) = %s\n", respath);
-	}
-
-	return 0;
+        if ((ts_ed.tv_nsec - ts_st.tv_nsec) < 0) {
+                ts->tv_sec += ts_ed.tv_sec - ts_st.tv_sec - 1;
+                ts->tv_nsec += 1000000000 + ts_ed.tv_nsec - ts_st.tv_nsec;
+        } else {
+                ts->tv_sec += ts_ed.tv_sec - ts_st.tv_sec;
+                ts->tv_nsec += ts_ed.tv_nsec - ts_st.tv_nsec;
+        }
+
+        if (ts->tv_nsec > 1000000000) {
+                ts->tv_nsec = ts->tv_nsec - 1000000000;
+                ts->tv_sec += 1;
+        }
+
+        return;
 }
 
+#define MAX_FILES_CREATE 10
+#define MAXPATHNAME      512
+void
+test_h_performance (void)
+{
+        char               *my_dir = "perftest",
+                           *full_dir_path="/testdir/perftest";
+        char               *my_file = "file_", my_file_name[MAXPATHNAME];
+        struct glfs_object *parent = NULL, *leaf = NULL, *dir = NULL;
+        struct stat         sb;
+        int                 ret, i;
+        struct glfs_fd     *fd;
+        struct timespec     c_ts = {0, 0}, c_ts_st, c_ts_ed;
+        struct timespec     o_ts = {0, 0}, o_ts_st, o_ts_ed;
+
+        printf("glfs_h_performance tests: In Progress\n");
+
+        /* Prepare tests */
+        parent = glfs_h_lookupat (fs, NULL, full_parent_name, &sb);
+        if (parent == NULL) {
+                fprintf (stderr, "glfs_h_lookupat: error on lookup of %s: from (%p),%s\n",
+                         full_parent_name, NULL, strerror (errno));
+                printf ("glfs_h_performance tests: FAILED\n");
+                goto out;
+        }
+
+        dir = glfs_h_mkdir (fs, parent, my_dir, 0644, &sb);
+        if (dir == NULL) {
+                fprintf (stderr, "glfs_h_mkdir: error creating %s: from (%p),%s\n",
+                         my_dir, parent, strerror (errno));
+                printf ("glfs_h_performance tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        /* create performance */
+        ret = clock_gettime (CLOCK_REALTIME, &o_ts_st);
+        if(ret != 0) {
+                fprintf (stderr, "clock_gettime: error %s\n", strerror (errno));
+                printf ("glfs_h_getattrs and setattrs tests: FAILED\n");
+                goto out;
+        }
+
+        for (i = 0; i < MAX_FILES_CREATE; i++) {
+                sprintf (my_file_name, "%s%d", my_file, i);
+
+                ret = clock_gettime (CLOCK_REALTIME, &c_ts_st);
+                if(ret != 0) {
+                        fprintf (stderr, "clock_gettime: error %s\n",
+                                 strerror (errno));
+                        printf ("glfs_h_getattrs and setattrs tests: FAILED\n");
+                        goto out;
+                }
+
+                leaf = glfs_h_lookupat (fs, dir, my_file_name, &sb);
+                if (leaf != NULL) {
+                        fprintf (stderr, "glfs_h_lookup: exists %s\n",
+                                 my_file_name);
+                        printf ("glfs_h_performance tests: FAILED\n");
+                        goto out;
+                }
+
+                leaf = glfs_h_creat (fs, dir, my_file_name, O_CREAT, 0644, &sb);
+                if (leaf == NULL) {
+                        fprintf (stderr, "glfs_h_creat: error creating %s: from (%p),%s\n",
+                                 my_file, dir, strerror (errno));
+                        printf ("glfs_h_performance tests: FAILED\n");
+                        goto out;
+                }
+
+                ret = clock_gettime (CLOCK_REALTIME, &c_ts_ed);
+                if(ret != 0) {
+                        fprintf (stderr, "clock_gettime: error %s\n",
+                                 strerror (errno));
+                        printf ("glfs_h_getattrs and setattrs tests: FAILED\n");
+                        goto out;
+                }
+
+                assimilatetime (&c_ts, c_ts_st, c_ts_ed);
+                glfs_h_close (leaf); leaf = NULL;
+        }
+
+        ret = clock_gettime (CLOCK_REALTIME, &o_ts_ed);
+        if(ret != 0) {
+                fprintf (stderr, "clock_gettime: error %s\n", strerror (errno));
+                printf ("glfs_h_getattrs and setattrs tests: FAILED\n");
+                goto out;
+        }
+
+        assimilatetime (&o_ts, o_ts_st, o_ts_ed);
+
+        printf ("Creation performance (handle based):\n\t# empty files:%d\n",
+                MAX_FILES_CREATE);
+        printf ("\tOverall time:\n\t\tSecs:%ld\n\t\tnSecs:%ld\n",
+                o_ts.tv_sec, o_ts.tv_nsec);
+        printf ("\tcreate call time time:\n\t\tSecs:%ld\n\t\tnSecs:%ld\n",
+                c_ts.tv_sec, c_ts.tv_nsec);
+
+        /* create using path */
+        c_ts.tv_sec = o_ts.tv_sec = 0;
+        c_ts.tv_nsec = o_ts.tv_nsec = 0;
+
+        sprintf (my_file_name, "%s1", full_dir_path);
+        ret = glfs_mkdir (fs, my_file_name, 0644);
+        if (ret != 0) {
+                fprintf (stderr, "glfs_mkdir: error creating %s: from (%p),%s\n",
+                         my_dir, parent, strerror (errno));
+                printf ("glfs_h_performance tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        ret = clock_gettime (CLOCK_REALTIME, &o_ts_st);
+        if(ret != 0) {
+                fprintf (stderr, "clock_gettime: error %s\n", strerror (errno));
+                printf ("glfs_h_getattrs and setattrs tests: FAILED\n");
+                goto out;
+        }
+
+        for (i = 0; i < MAX_FILES_CREATE; i++) {
+                sprintf (my_file_name, "%s1/%sn%d", full_dir_path, my_file, i);
+
+                ret = clock_gettime (CLOCK_REALTIME, &c_ts_st);
+                if(ret != 0) {
+                        fprintf (stderr, "clock_gettime: error %s\n",
+                                 strerror (errno));
+                        printf ("glfs_h_getattrs and setattrs tests: FAILED\n");
+                        goto out;
+                }
+
+                ret = glfs_stat (fs, my_file_name, &sb);
+                if (ret == 0) {
+                        fprintf (stderr, "glfs_stat: exists %s\n",
+                                 my_file_name);
+                        printf ("glfs_h_performance tests: FAILED\n");
+                        goto out;
+                }
+
+                fd = glfs_creat (fs, my_file_name, O_CREAT, 0644);
+                if (fd == NULL) {
+                        fprintf (stderr, "glfs_creat: error creating %s: from (%p),%s\n",
+                                 my_file, dir, strerror (errno));
+                        printf ("glfs_h_performance tests: FAILED\n");
+                        goto out;
+                }
+
+                ret = clock_gettime (CLOCK_REALTIME, &c_ts_ed);
+                if(ret != 0) {
+                        fprintf (stderr, "clock_gettime: error %s\n",
+                                 strerror (errno));
+                        printf ("glfs_h_getattrs and setattrs tests: FAILED\n");
+                        goto out;
+                }
+
+                assimilatetime (&c_ts, c_ts_st, c_ts_ed);
+                glfs_close (fd);
+        }
+
+        ret = clock_gettime (CLOCK_REALTIME, &o_ts_ed);
+        if(ret != 0) {
+                fprintf (stderr, "clock_gettime: error %s\n", strerror (errno));
+                printf ("glfs_h_getattrs and setattrs tests: FAILED\n");
+                goto out;
+        }
+
+        assimilatetime (&o_ts, o_ts_st, o_ts_ed);
+
+        printf ("Creation performance (path based):\n\t# empty files:%d\n",
+                MAX_FILES_CREATE);
+        printf ("\tOverall time:\n\t\tSecs:%ld\n\t\tnSecs:%ld\n",
+                o_ts.tv_sec, o_ts.tv_nsec);
+        printf ("\tcreate call time time:\n\t\tSecs:%ld\n\t\tnSecs:%ld\n",
+                c_ts.tv_sec, c_ts.tv_nsec);
+out:
+        return;
+}
+
+int
+test_handleops (int argc, char *argv[])
+{
+        int                 ret = 0;
+        glfs_fd_t          *fd = NULL;
+        struct stat         sb = {0, };
+        struct glfs_object *root = NULL, *parent = NULL, *leaf = NULL,
+                           *tmp = NULL;
+        char                readbuf[32], writebuf[32];
+        unsigned char       leaf_handle[GFAPI_HANDLE_LENGTH];
+
+        char *full_leaf_name = "/testdir/testfile.txt",
+             *leaf_name = "testfile.txt",
+             *relative_leaf_name = "testdir/testfile.txt";
+        char *leaf_name1 = "testfile1.txt";
+        char *full_newparent_name = "/testdir/dir1",
+             *newparent_name = "dir1";
+        char *full_newnod_name = "/testdir/nod1",
+             *newnod_name = "nod1";
+
+        /* Initialize test area */
+        ret = glfs_mkdir (fs, full_parent_name, 0644);
+        if (ret != 0 && errno != EEXIST) {
+                fprintf (stderr, "%s: (%p) %s\n", full_parent_name, fd,
+                        strerror (errno));
+                printf ("Test initialization failed on volume %s\n", argv[1]);
+                goto out;
+        }
+        else if (ret != 0) {
+                printf ("Found test directory %s to be existing\n",
+                        full_parent_name);
+                printf ("Cleanup test directory and restart tests\n");
+                goto out;
+        }
+
+        fd = glfs_creat (fs, full_leaf_name, O_CREAT, 0644);
+        if (fd == NULL) {
+                fprintf (stderr, "%s: (%p) %s\n", full_leaf_name, fd,
+                        strerror (errno));
+                printf ("Test initialization failed on volume %s\n", argv[1]);
+                goto out;
+        }
+        glfs_close (fd);
+
+        printf ("Initialized the test area, within volume %s\n", argv[1]);
+
+        /* Handle based APIs test area */
+
+        /* glfs_lookupat test */
+        printf ("glfs_h_lookupat tests: In Progress\n");
+        /* start at root of the volume */
+        root = glfs_h_lookupat (fs, NULL, "/", &sb);
+        if (root == NULL) {
+                fprintf (stderr, "glfs_h_lookupat: error on lookup of %s: from (%p),%s\n",
+                         "/", NULL, strerror (errno));
+                printf ("glfs_h_lookupat tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        /* lookup a parent within root */
+        parent = glfs_h_lookupat (fs, root, parent_name, &sb);
+        if (parent == NULL) {
+                fprintf (stderr, "glfs_h_lookupat: error on lookup of %s: from (%p),%s\n",
+                         parent_name, root, strerror (errno));
+                printf ("glfs_h_lookupat tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        /* lookup a leaf/child within the parent */
+        leaf = glfs_h_lookupat (fs, parent, leaf_name, &sb);
+        if (leaf == NULL) {
+                fprintf (stderr, "glfs_h_lookupat: error on lookup of %s: from (%p),%s\n",
+                         leaf_name, parent, strerror (errno));
+                printf ("glfs_h_lookupat tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        /* reset */
+        glfs_h_close (root); root = NULL;
+        glfs_h_close (leaf); leaf = NULL;
+        glfs_h_close (parent); parent = NULL;
+
+        /* check absolute paths */
+        root = glfs_h_lookupat (fs, NULL, "/", &sb);
+        if (root == NULL) {
+                fprintf (stderr, "glfs_h_lookupat: error on lookup of %s: from (%p),%s\n",
+                         "/", NULL, strerror (errno));
+                printf ("glfs_h_lookupat tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        parent = glfs_h_lookupat (fs, NULL, full_parent_name, &sb);
+        if (parent == NULL) {
+                fprintf (stderr, "glfs_h_lookupat: error on lookup of %s: from (%p),%s\n",
+                         full_parent_name, root, strerror (errno));
+                printf ("glfs_h_lookupat tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        leaf = glfs_h_lookupat (fs, NULL, full_leaf_name, &sb);
+        if (leaf == NULL) {
+                fprintf (stderr, "glfs_h_lookupat: error on lookup of %s: from (%p),%s\n",
+                         full_leaf_name, parent, strerror (errno));
+                printf ("glfs_h_lookupat tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        /* reset */
+        glfs_h_close (leaf); leaf = NULL;
+
+        /* check multiple component paths */
+        leaf = glfs_h_lookupat (fs, root, relative_leaf_name, &sb);
+        if (leaf == NULL) {
+                fprintf (stderr, "glfs_h_lookupat: error on lookup of %s: from (%p),%s\n",
+                         relative_leaf_name, parent, strerror (errno));
+                goto out;
+        }
+        peek_stat (&sb);
+
+        /* reset */
+        glfs_h_close (root); root = NULL;
+        glfs_h_close (parent); parent = NULL;
+
+        /* check symlinks in path */
+
+        /* TODO: -ve test cases */
+        /* parent invalid
+        * path invalid
+        * path does not exist after some components
+        * no parent, but relative path
+        * parent and full path? -ve?
+        */
+
+        printf ("glfs_h_lookupat tests: PASSED\n");
+
+        /* glfs_openat test */
+        printf ("glfs_h_open tests: In Progress\n");
+        fd = glfs_h_open (fs, leaf, O_RDWR);
+        if (fd == NULL) {
+                fprintf (stderr, "glfs_h_open: error on open of %s: %s\n",
+                         full_leaf_name, strerror (errno));
+                printf ("glfs_h_open tests: FAILED\n");
+                goto out;
+        }
+
+        /* test read/write based on fd */
+        memcpy (writebuf, "abcdefghijklmnopqrstuvwxyz012345", 32);
+        ret = glfs_write (fd, writebuf, 32, 0);
+
+        glfs_lseek (fd, 0, SEEK_SET);
+
+        ret = glfs_read (fd, readbuf, 32, 0);
+        if (memcmp (readbuf, writebuf, 32)) {
+                printf ("Failed to read what I wrote: %s %s\n", readbuf,
+                        writebuf);
+                glfs_close (fd);
+                printf ("glfs_h_open tests: FAILED\n");
+                goto out;
+        }
+
+        glfs_h_close (leaf); leaf = NULL;
+        glfs_close (fd);
+
+        printf ("glfs_h_open tests: PASSED\n");
+
+        /* Create tests */
+        printf ("glfs_h_creat tests: In Progress\n");
+        parent = glfs_h_lookupat (fs, NULL, full_parent_name, &sb);
+        if (parent == NULL) {
+                fprintf (stderr, "glfs_h_lookupat: error on lookup of %s: from (%p),%s\n",
+                         full_parent_name, root, strerror (errno));
+                printf ("glfs_h_creat tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        leaf = glfs_h_creat (fs, parent, leaf_name1, O_CREAT, 0644, &sb);
+        if (leaf == NULL) {
+                fprintf (stderr, "glfs_h_creat: error on create of %s: from (%p),%s\n",
+                         leaf_name1, parent, strerror (errno));
+                printf ("glfs_h_creat tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        glfs_h_close (leaf); leaf = NULL;
+
+        leaf = glfs_h_creat (fs, parent, leaf_name1, O_CREAT | O_EXCL, 0644,
+                            &sb);
+        if (leaf != NULL || errno != EEXIST) {
+                fprintf (stderr, "glfs_h_creat: existing file, leaf = (%p), errno = %s\n",
+                        leaf, strerror (errno));
+                printf ("glfs_h_creat tests: FAILED\n");
+                if (leaf != NULL) {
+                        glfs_h_close (leaf); leaf = NULL;
+                }
+        }
+
+        tmp = glfs_h_creat (fs, root, parent_name, O_CREAT, 0644, &sb);
+        if (tmp != NULL || !(errno == EISDIR || errno == EINVAL)) {
+                fprintf (stderr, "glfs_h_creat: dir create, tmp = (%p), errno = %s\n",
+                        leaf, strerror (errno));
+                printf ("glfs_h_creat tests: FAILED\n");
+                if (tmp != NULL) {
+                        glfs_h_close (tmp); tmp = NULL;
+                }
+        }
+
+        /* TODO: Other combinations and -ve cases as applicable */
+        printf ("glfs_h_creat tests: PASSED\n");
+
+        /* extract handle and create from handle test */
+        printf ("glfs_h_extract_handle and glfs_h_create_from_handle tests: In Progress\n");
+        /* TODO: Change the lookup to creat below for a GIFD recovery falure,
+         * that needs to be fixed */
+        leaf = glfs_h_lookupat (fs, parent, leaf_name1, &sb);
+        if (leaf == NULL) {
+                fprintf (stderr, "glfs_h_lookupat: error on lookup of %s: from (%p),%s\n",
+                         leaf_name1, parent, strerror (errno));
+                printf ("glfs_h_extract_handle tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        ret = glfs_h_extract_handle (leaf, leaf_handle,
+                                             GFAPI_HANDLE_LENGTH);
+        if (ret < 0) {
+                fprintf (stderr, "glfs_h_extract_handle: error extracting handle of %s: %s\n",
+                         full_leaf_name, strerror (errno));
+                printf ("glfs_h_extract_handle tests: FAILED\n");
+                goto out;
+        }
+        peek_handle (leaf_handle);
+
+        glfs_h_close (leaf); leaf = NULL;
+
+        leaf = glfs_h_create_from_handle (fs, leaf_handle, GFAPI_HANDLE_LENGTH,
+                                          &sb);
+        if (leaf == NULL) {
+                fprintf (stderr, "glfs_h_create_from_handle: error on create of %s: from (%p),%s\n",
+                         leaf_name1, leaf_handle, strerror (errno));
+                printf ("glfs_h_create_from_handle tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        fd = glfs_h_open (fs, leaf, O_RDWR);
+        if (fd == NULL) {
+                fprintf (stderr, "glfs_h_open: error on open of %s: %s\n",
+                         full_leaf_name, strerror (errno));
+                printf ("glfs_h_create_from_handle tests: FAILED\n");
+                goto out;
+        }
+
+        /* test read/write based on fd */
+        memcpy (writebuf, "abcdefghijklmnopqrstuvwxyz012345", 32);
+        ret = glfs_write (fd, writebuf, 32, 0);
+
+        glfs_lseek (fd, 0, SEEK_SET);
+
+        ret = glfs_read (fd, readbuf, 32, 0);
+        if (memcmp (readbuf, writebuf, 32)) {
+                printf ("Failed to read what I wrote: %s %s\n", writebuf,
+                        writebuf);
+                printf ("glfs_h_create_from_handle tests: FAILED\n");
+                glfs_close (fd);
+                goto out;
+        }
+
+        glfs_close (fd);
+        glfs_h_close (leaf); leaf = NULL;
+        glfs_h_close (parent); parent = NULL;
+
+        printf ("glfs_h_extract_handle and glfs_h_create_from_handle tests: PASSED\n");
+
+        /* Mkdir tests */
+        printf ("glfs_h_mkdir tests: In Progress\n");
+
+        ret = glfs_rmdir (fs, full_newparent_name);
+        if (ret && errno != ENOENT) {
+                fprintf (stderr, "glfs_rmdir: Failed for %s: %s\n",
+                         full_newparent_name, strerror (errno));
+                printf ("glfs_h_mkdir tests: FAILED\n");
+                goto out;
+        }
+
+        parent = glfs_h_lookupat (fs, NULL, full_parent_name, &sb);
+        if (parent == NULL) {
+                fprintf (stderr, "glfs_h_lookupat: error on lookup of %s: from (%p),%s\n",
+                         full_parent_name, root, strerror (errno));
+                printf ("glfs_h_mkdir tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        leaf = glfs_h_mkdir (fs, parent, newparent_name, 0644, &sb);
+        if (leaf == NULL) {
+                fprintf (stderr, "glfs_h_mkdir: error on mkdir of %s: from (%p),%s\n",
+                         newparent_name, parent, strerror (errno));
+                printf ("glfs_h_mkdir tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        glfs_h_close (leaf); leaf = NULL;
+
+        leaf = glfs_h_mkdir (fs, parent, newparent_name, 0644, &sb);
+        if (leaf != NULL || errno != EEXIST) {
+                fprintf (stderr, "glfs_h_mkdir: existing directory, leaf = (%p), errno = %s\n",
+                         leaf, strerror (errno));
+                printf ("glfs_h_mkdir tests: FAILED\n");
+                if (leaf != NULL) {
+                        glfs_h_close (leaf); leaf = NULL;
+                }
+        }
+
+        glfs_h_close (parent); parent = NULL;
+
+        printf ("glfs_h_mkdir tests: PASSED\n");
+
+        /* Mknod tests */
+        printf ("glfs_h_mknod tests: In Progress\n");
+        ret = glfs_unlink (fs, full_newnod_name);
+        if (ret && errno != ENOENT) {
+                fprintf (stderr, "glfs_unlink: Failed for %s: %s\n",
+                         full_newnod_name, strerror (errno));
+                printf ("glfs_h_mknod tests: FAILED\n");
+                goto out;
+        }
+
+        parent = glfs_h_lookupat (fs, NULL, full_parent_name, &sb);
+        if (parent == NULL) {
+                fprintf (stderr, "glfs_h_lookupat: error on lookup of %s: from (%p),%s\n",
+                         full_parent_name, root, strerror (errno));
+                printf ("glfs_h_mknod tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        leaf = glfs_h_mknod (fs, parent, newnod_name, S_IFIFO, 0, &sb);
+        if (leaf == NULL) {
+                fprintf (stderr, "glfs_h_mkdir: error on mkdir of %s: from (%p),%s\n",
+                         newnod_name, parent, strerror (errno));
+                printf ("glfs_h_mknod tests: FAILED\n");
+                goto out;
+        }
+        peek_stat (&sb);
+
+        /* TODO: creat op on a FIFO node hangs, need to check and fix
+        tmp = glfs_h_creat (fs, parent, newnod_name, O_CREAT, 0644, &sb);
+        if (tmp != NULL || errno != EINVAL) {
+                fprintf (stderr, "glfs_h_creat: node create, tmp = (%p), errno = %s\n",
+                        tmp, strerror (errno));
+                printf ("glfs_h_creat/mknod tests: FAILED\n");
+                if (tmp != NULL) {
+                        glfs_h_close(tmp); tmp = NULL;
+                }
+        } */
+
+        glfs_h_close (leaf); leaf = NULL;
+
+        leaf = glfs_h_mknod (fs, parent, newnod_name, 0644, 0, &sb);
+        if (leaf != NULL || errno != EEXIST) {
+                fprintf (stderr, "glfs_h_mknod: existing node, leaf = (%p), errno = %s\n",
+                         leaf, strerror (errno));
+                printf ("glfs_h_mknod tests: FAILED\n");
+                if (leaf != NULL) {
+                        glfs_h_close (leaf); leaf = NULL;
+                }
+        }
+
+        glfs_h_close (parent); parent = NULL;
+
+        printf ("glfs_h_mknod tests: PASSED\n");
+
+        /* unlink tests */
+        test_h_unlink ();
+
+        /* TODO: opendir tests */
+
+        /* getattr tests */
+        test_h_getsetattrs ();
+
+        /* TODO: setattr tests */
+
+        /* truncate tests */
+        test_h_truncate();
+
+        /* link tests */
+        test_h_links ();
+
+        /* rename tests */
+        test_h_rename ();
+
+        /* performance tests */
+        test_h_performance ();
+
+        /* END: New APIs test area */
+
+out:
+        /* Cleanup glfs handles */
+        if (root)
+                glfs_h_close (root);
+        if (parent)
+                glfs_h_close (parent);
+        if (leaf)
+                glfs_h_close (leaf);
+
+        return ret;
+}
 
 int
 main (int argc, char *argv[])
 {
-	glfs_t    *fs = NULL;
-	glfs_t    *fs2 = NULL;
-	int        ret = 0;
-	glfs_fd_t *fd = NULL;
-	glfs_fd_t *fd2 = NULL;
-	struct stat sb = {0, };
-	char       readbuf[32];
-	char       writebuf[32];
+        glfs_t    *fs2 = NULL;
+        int        ret = 0;
+        glfs_fd_t *fd = NULL;
+        glfs_fd_t *fd2 = NULL;
+        struct stat sb = {0, };
+        char       readbuf[32];
+        char       writebuf[32];
+
+        char      *filename = "/filename2";
 
-	char      *filename = "/filename2";
+        if (argc != 3) {
+                printf ("Expect following args\n\t%s <volname> <hostname>\n", argv[0]);
+                return -1;
+        }
 
-	fs = glfs_new ("fsync");
-	if (!fs) {
-		fprintf (stderr, "glfs_new: returned NULL\n");
-		return 1;
-	}
+        fs = glfs_new (argv[1]);
+        if (!fs) {
+                fprintf (stderr, "glfs_new: returned NULL\n");
+                return 1;
+        }
 
-//	ret = glfs_set_volfile (fs, "/tmp/posix.vol");
+//      ret = glfs_set_volfile (fs, "/tmp/posix.vol");
 
-	ret = glfs_set_volfile_server (fs, "tcp", "localhost", 24007);
+        ret = glfs_set_volfile_server (fs, "tcp", argv[2], 24007);
 
-//	ret = glfs_set_volfile_server (fs, "unix", "/tmp/gluster.sock", 0);
+//      ret = glfs_set_volfile_server (fs, "unix", "/tmp/gluster.sock", 0);
 
-	ret = glfs_set_logging (fs, "/dev/stderr", 7);
+        ret = glfs_set_logging (fs, "/dev/stderr", 7);
 
-	ret = glfs_init (fs);
+        ret = glfs_init (fs);
 
-	fprintf (stderr, "glfs_init: returned %d\n", ret);
+        fprintf (stderr, "glfs_init: returned %d\n", ret);
 
-	sleep (2);
+        sleep (2);
 
-	fs2 = glfs_new ("fsync");
-	if (!fs2) {
-		fprintf (stderr, "glfs_new: returned NULL\n");
-		return 1;
-	}
+        fs2 = glfs_new (argv[1]);
+        if (!fs2) {
+                fprintf (stderr, "glfs_new: returned NULL\n");
+                return 1;
+        }
 
 
-//	ret = glfs_set_volfile (fs2, "/tmp/posix.vol");
+//      ret = glfs_set_volfile (fs2, "/tmp/posix.vol");
 
-	ret = glfs_set_volfile_server (fs2, "tcp", "localhost", 24007);
+        ret = glfs_set_volfile_server (fs2, "tcp", argv[2], 24007);
 
-	ret = glfs_set_logging (fs2, "/dev/stderr", 7);
+        ret = glfs_set_logging (fs2, "/dev/stderr", 7);
 
-	ret = glfs_init (fs2);
+        ret = glfs_init (fs2);
 
-	fprintf (stderr, "glfs_init: returned %d\n", ret);
+        fprintf (stderr, "glfs_init: returned %d\n", ret);
 
-	ret = glfs_lstat (fs, filename, &sb);
-	fprintf (stderr, "%s: (%d) %s\n", filename, ret, strerror (errno));
+        ret = glfs_lstat (fs, filename, &sb);
+        fprintf (stderr, "%s: (%d) %s\n", filename, ret, strerror (errno));
 
-	fd = glfs_creat (fs, filename, O_RDWR, 0644);
-	fprintf (stderr, "%s: (%p) %s\n", filename, fd, strerror (errno));
+        fd = glfs_creat (fs, filename, O_RDWR, 0644);
+        fprintf (stderr, "%s: (%p) %s\n", filename, fd, strerror (errno));
 
-	fd2 = glfs_open (fs2, filename, O_RDWR);
-	fprintf (stderr, "%s: (%p) %s\n", filename, fd, strerror (errno));
+        fd2 = glfs_open (fs2, filename, O_RDWR);
+        fprintf (stderr, "%s: (%p) %s\n", filename, fd, strerror (errno));
 
-	sprintf (writebuf, "hi there\n");
-	ret = glfs_write (fd, writebuf, 32, 0);
+        sprintf (writebuf, "hi there\n");
+        ret = glfs_write (fd, writebuf, 32, 0);
 
-	glfs_lseek (fd2, 0, SEEK_SET);
+        glfs_lseek (fd2, 0, SEEK_SET);
 
-	ret = glfs_read (fd2, readbuf, 32, 0);
+        ret = glfs_read (fd2, readbuf, 32, 0);
 
-	printf ("read %d, %s", ret, readbuf);
+        printf ("read %d, %s", ret, readbuf);
 
-	glfs_close (fd);
-	glfs_close (fd2);
+        glfs_close (fd);
+        glfs_close (fd2);
 
-	filename = "/filename3";
-	ret = glfs_mknod (fs, filename, S_IFIFO, 0);
-	fprintf (stderr, "%s: (%d) %s\n", filename, ret, strerror (errno));
+        filename = "/filename3";
+        ret = glfs_mknod (fs, filename, S_IFIFO, 0);
+        fprintf (stderr, "%s: (%d) %s\n", filename, ret, strerror (errno));
 
-	ret = glfs_lstat (fs, filename, &sb);
-	fprintf (stderr, "%s: (%d) %s\n", filename, ret, strerror (errno));
+        ret = glfs_lstat (fs, filename, &sb);
+        fprintf (stderr, "%s: (%d) %s\n", filename, ret, strerror (errno));
 
 
-	ret = glfs_rename (fs, filename, "/filename4");
-	fprintf (stderr, "rename(%s): (%d) %s\n", filename, ret,
-		 strerror (errno));
+        ret = glfs_rename (fs, filename, "/filename4");
+        fprintf (stderr, "rename(%s): (%d) %s\n", filename, ret,
+                 strerror (errno));
 
-	ret = glfs_unlink (fs, "/filename4");
-	fprintf (stderr, "unlink(%s): (%d) %s\n", "/filename4", ret,
-		 strerror (errno));
+        ret = glfs_unlink (fs, "/filename4");
+        fprintf (stderr, "unlink(%s): (%d) %s\n", "/filename4", ret,
+                 strerror (errno));
 
-	filename = "/dirname2";
-	ret = glfs_mkdir (fs, filename, 0);
-	fprintf (stderr, "%s: (%d) %s\n", filename, ret, strerror (errno));
+        filename = "/dirname2";
+        ret = glfs_mkdir (fs, filename, 0);
+        fprintf (stderr, "%s: (%d) %s\n", filename, ret, strerror (errno));
 
-	ret = glfs_lstat (fs, filename, &sb);
-	fprintf (stderr, "lstat(%s): (%d) %s\n", filename, ret, strerror (errno));
+        ret = glfs_lstat (fs, filename, &sb);
+        fprintf (stderr, "lstat(%s): (%d) %s\n", filename, ret, strerror (errno));
 
-	ret = glfs_rmdir (fs, filename);
-	fprintf (stderr, "rmdir(%s): (%d) %s\n", filename, ret, strerror (errno));
+        ret = glfs_rmdir (fs, filename);
+        fprintf (stderr, "rmdir(%s): (%d) %s\n", filename, ret, strerror (errno));
 
-	test_dirops (fs);
+        test_dirops (fs);
 
-	test_xattr (fs);
+        test_xattr (fs);
 
-	test_chdir (fs);
+        test_chdir (fs);
 
-	// done
+        test_handleops (argc, argv);
+        // done
 
-	glfs_fini (fs);
-	glfs_fini (fs2);
+        glfs_fini (fs);
+        glfs_fini (fs2);
 
-	return ret;
+        return ret;
 }
diff --git a/api/examples/setup.py.in b/api/examples/setup.py.in
new file mode 100644
index 000000000..44b738094
--- /dev/null
+++ b/api/examples/setup.py.in
@@ -0,0 +1,29 @@
+from distutils.core import setup
+
+# generate a __init__.py for the package namespace
+fo = open('__init__.py', 'w')
+fo.write('__version__ = "@PACKAGE_VERSION@"\n')
+fo.close()
+
+DESC = """GlusterFS is a clustered file-system capable of scaling to
+several petabytes. It aggregates various storage bricks over Infiniband
+RDMA or TCP/IP interconnect into one large parallel network file system.
+GlusterFS is one of the most sophisticated file systems in terms of
+features and extensibility.  It borrows a powerful concept called
+Translators from GNU Hurd kernel. Much of the code in GlusterFS is in
+user space and easily manageable.
+
+This package contains the Python interface to the libgfapi library."""
+
+setup(
+    name='glusterfs-api',
+    version='@PACKAGE_VERSION@',
+    description='Python client library for the GlusterFS libgfapi',
+    long_description=DESC,
+    author='Gluster Community',
+    author_email='gluster-devel@nongnu.org',
+    license='LGPLv3',
+    url='http://gluster.org/',
+    package_dir={'gluster':''},
+    packages=['gluster']
+)
diff --git a/api/src/Makefile.am b/api/src/Makefile.am
index 8980c021c..7c5df3e20 100644
--- a/api/src/Makefile.am
+++ b/api/src/Makefile.am
@@ -1,9 +1,10 @@
 lib_LTLIBRARIES = libgfapi.la
 noinst_HEADERS = glfs-mem-types.h glfs-internal.h
-libgfapi_HEADERS = glfs.h
+libgfapi_HEADERS = glfs.h glfs-handles.h
 libgfapidir = $(includedir)/glusterfs/api
 
-libgfapi_la_SOURCES = glfs.c glfs-mgmt.c glfs-fops.c glfs-resolve.c
+libgfapi_la_SOURCES = glfs.c glfs-mgmt.c glfs-fops.c glfs-resolve.c \
+	glfs-handleops.c
 libgfapi_la_LIBADD = $(top_builddir)/libglusterfs/src/libglusterfs.la \
 	$(top_builddir)/rpc/rpc-lib/src/libgfrpc.la \
 	$(top_builddir)/rpc/xdr/src/libgfxdr.la \
@@ -17,6 +18,10 @@ libgfapi_la_CPPFLAGS = $(GF_CPPFLAGS) -D__USE_FILE_OFFSET64 \
 
 xlator_LTLIBRARIES = api.la
 xlatordir = $(libdir)/glusterfs/$(PACKAGE_VERSION)/xlator/mount
+# workaround for broken parallel install support in automake with LTLIBRARIES
+# http://debbugs.gnu.org/cgi/bugreport.cgi?bug=7328
+install_xlatorLTLIBRARIES = install-xlatorLTLIBRARIES
+$(install_xlatorLTLIBRARIES): install-libLTLIBRARIES
 
 api_la_SOURCES = glfs-master.c
 api_la_DEPENDENCIES = libgfapi.la
diff --git a/api/src/glfs-fops.c b/api/src/glfs-fops.c
index 61b07e987..10bb7d38b 100644
--- a/api/src/glfs-fops.c
+++ b/api/src/glfs-fops.c
@@ -13,21 +13,17 @@
 #include "glfs-mem-types.h"
 #include "syncop.h"
 #include "glfs.h"
+#include <limits.h>
 
-#define DEFAULT_REVAL_COUNT 1
-
-#define ESTALE_RETRY(ret,errno,reval,loc,label) do {	\
-	if (ret == -1 && errno == ESTALE) {	        \
-		if (reval < DEFAULT_REVAL_COUNT) {	\
-			reval++;			\
-			loc_wipe (loc);			\
-			goto label;			\
-		}					\
-	}						\
-	} while (0)
+#ifdef NAME_MAX
+#define GF_NAME_MAX NAME_MAX
+#else
+#define GF_NAME_MAX 255
+#endif
 
+#define READDIRBUF_SIZE (sizeof(struct dirent) + GF_NAME_MAX + 1)
 
-static int
+int
 glfs_loc_link (loc_t *loc, struct iatt *iatt)
 {
 	int ret = -1;
@@ -52,7 +48,7 @@ glfs_loc_link (loc_t *loc, struct iatt *iatt)
 }
 
 
-static void
+void
 glfs_iatt_to_stat (struct glfs *fs, struct iatt *iatt, struct stat *stat)
 {
 	iatt_to_stat (iatt, stat);
@@ -60,7 +56,7 @@ glfs_iatt_to_stat (struct glfs *fs, struct iatt *iatt, struct stat *stat)
 }
 
 
-static int
+int
 glfs_loc_unlink (loc_t *loc)
 {
 	inode_unlink (loc->inode, loc->parent, loc->name);
@@ -136,7 +132,8 @@ out:
 	if (ret && glfd) {
 		glfs_fd_destroy (glfd);
 		glfd = NULL;
-	} else {
+	} else if (glfd) {
+                glfd->fd->flags = flags;
 		fd_bind (glfd->fd);
 		glfs_fd_bind (glfd);
 	}
@@ -158,6 +155,11 @@ glfs_close (struct glfs_fd *glfd)
 	__glfs_entry_fd (glfd);
 
 	subvol = glfs_active_subvol (glfd->fs);
+        if (!subvol) {
+                ret = -1;
+                errno = EIO;
+                goto out;
+        }
 
 	fd = glfs_resolve_fd (glfd->fs, subvol, glfd);
 	if (!fd) {
@@ -388,8 +390,12 @@ retry:
 		goto out;
 	}
 
-	ret = syncop_create (subvol, &loc, flags, mode, glfd->fd,
-			     xattr_req, &iatt);
+	if (ret == 0) {
+		ret = syncop_open (subvol, &loc, flags, glfd->fd);
+	} else {
+		ret = syncop_create (subvol, &loc, flags, mode, glfd->fd,
+				     xattr_req, &iatt);
+	}
 
 	ESTALE_RETRY (ret, errno, reval, &loc, retry);
 
@@ -404,7 +410,8 @@ out:
 	if (ret && glfd) {
 		glfs_fd_destroy (glfd);
 		glfd = NULL;
-	} else {
+	} else if (glfd) {
+                glfd->fd->flags = flags;
 		fd_bind (glfd->fd);
 		glfs_fd_bind (glfd);
 	}
@@ -484,13 +491,13 @@ glfs_preadv (struct glfs_fd *glfd, const struct iovec *iovec, int iovcnt,
 
 	glfd->offset = (offset + size);
 
-	if (iov)
-		GF_FREE (iov);
-	if (iobref)
-		iobref_unref (iobref);
-
 	ret = size;
 out:
+        if (iov)
+                GF_FREE (iov);
+        if (iobref)
+                iobref_unref (iobref);
+
 	if (fd)
 		fd_unref (fd);
 
@@ -576,10 +583,6 @@ glfs_io_async_task (void *data)
 	ssize_t         ret = 0;
 
 	switch (gio->op) {
-	case GF_FOP_READ:
-		ret = glfs_preadv (gio->glfd, gio->iov, gio->count,
-				   gio->offset, gio->flags);
-		break;
 	case GF_FOP_WRITE:
 		ret = glfs_pwritev (gio->glfd, gio->iov, gio->count,
 				    gio->offset, gio->flags);
@@ -596,6 +599,9 @@ glfs_io_async_task (void *data)
 	case GF_FOP_DISCARD:
 		ret = glfs_discard (gio->glfd, gio->offset, gio->count);
 		break;
+        case GF_FOP_ZEROFILL:
+                ret = glfs_zerofill(gio->glfd, gio->offset, gio->count);
+                break;
 	}
 
 	return (int) ret;
@@ -603,23 +609,90 @@ glfs_io_async_task (void *data)
 
 
 int
+glfs_preadv_async_cbk (call_frame_t *frame, void *cookie, xlator_t *this,
+		       int op_ret, int op_errno, struct iovec *iovec,
+		       int count, struct iatt *stbuf, struct iobref *iobref,
+		       dict_t *xdata)
+{
+	struct glfs_io *gio = NULL;
+	xlator_t       *subvol = NULL;
+	struct glfs    *fs = NULL;
+	struct glfs_fd *glfd = NULL;
+
+
+	gio = frame->local;
+	frame->local = NULL;
+	subvol = cookie;
+	glfd = gio->glfd;
+	fs = glfd->fs;
+
+	if (op_ret <= 0)
+		goto out;
+
+	op_ret = iov_copy (gio->iov, gio->count, iovec, count);
+
+	glfd->offset = gio->offset + op_ret;
+out:
+	errno = op_errno;
+	gio->fn (gio->glfd, op_ret, gio->data);
+
+	GF_FREE (gio->iov);
+	GF_FREE (gio);
+	STACK_DESTROY (frame->root);
+	glfs_subvol_done (fs, subvol);
+
+	return 0;
+}
+
+
+int
 glfs_preadv_async (struct glfs_fd *glfd, const struct iovec *iovec, int count,
 		   off_t offset, int flags, glfs_io_cbk fn, void *data)
 {
 	struct glfs_io *gio = NULL;
 	int             ret = 0;
+	call_frame_t   *frame = NULL;
+	xlator_t       *subvol = NULL;
+	glfs_t         *fs = NULL;
+	fd_t           *fd = NULL;
+
+	__glfs_entry_fd (glfd);
+
+	subvol = glfs_active_subvol (glfd->fs);
+	if (!subvol) {
+		ret = -1;
+		errno = EIO;
+		goto out;
+	}
+
+	fd = glfs_resolve_fd (glfd->fs, subvol, glfd);
+	if (!fd) {
+		ret = -1;
+		errno = EBADFD;
+		goto out;
+	}
+
+	fs = glfd->fs;
+
+	frame = syncop_create_frame (THIS);
+	if (!frame) {
+		ret = -1;
+		errno = ENOMEM;
+		goto out;
+	}
 
 	gio = GF_CALLOC (1, sizeof (*gio), glfs_mt_glfs_io_t);
 	if (!gio) {
+		ret = -1;
 		errno = ENOMEM;
-		return -1;
+		goto out;
 	}
 
 	gio->iov = iov_dup (iovec, count);
 	if (!gio->iov) {
-		GF_FREE (gio);
+		ret = -1;
 		errno = ENOMEM;
-		return -1;
+		goto out;
 	}
 
 	gio->op     = GF_FOP_READ;
@@ -630,15 +703,23 @@ glfs_preadv_async (struct glfs_fd *glfd, const struct iovec *iovec, int count,
 	gio->fn     = fn;
 	gio->data   = data;
 
-	ret = synctask_new (glfs_from_glfd (glfd)->ctx->env,
-			    glfs_io_async_task, glfs_io_async_cbk,
-			    NULL, gio);
+	frame->local = gio;
 
+	STACK_WIND_COOKIE (frame, glfs_preadv_async_cbk, subvol, subvol,
+			   subvol->fops->readv, fd, iov_length (iovec, count),
+			   offset, flags, NULL);
+
+out:
 	if (ret) {
 		GF_FREE (gio->iov);
 		GF_FREE (gio);
+		STACK_DESTROY (frame->root);
+		glfs_subvol_done (fs, subvol);
 	}
 
+	if (fd)
+		fd_unref (fd);
+
 	return ret;
 }
 
@@ -1203,6 +1284,7 @@ glfs_readlink (struct glfs *fs, const char *path, char *buf, size_t bufsiz)
 	loc_t            loc = {0, };
 	struct iatt      iatt = {0, };
 	int              reval = 0;
+	char            *linkval = NULL;
 
 	__glfs_entry_fs (fs);
 
@@ -1226,7 +1308,11 @@ retry:
 		goto out;
 	}
 
-	ret = syncop_readlink (subvol, &loc, &buf, bufsiz);
+	ret = syncop_readlink (subvol, &loc, &linkval, bufsiz);
+	if (ret > 0) {
+		memcpy (buf, linkval, ret);
+		GF_FREE (linkval);
+	}
 
 	ESTALE_RETRY (ret, errno, reval, &loc, retry);
 out:
@@ -1608,6 +1694,15 @@ retrynew:
 		goto out;
 	}
 
+        /* Filling the inode of the hard link to be same as that of the
+           original file
+        */
+	if (newloc.inode) {
+		inode_unref (newloc.inode);
+		newloc.inode = NULL;
+	}
+        newloc.inode = inode_ref (oldloc.inode);
+
 	ret = syncop_link (subvol, &oldloc, &newloc);
 
 	if (ret == -1 && errno == ESTALE) {
@@ -1781,6 +1876,38 @@ glfs_discard_async (struct glfs_fd *glfd, off_t offset, size_t len,
 	return ret;
 }
 
+int
+glfs_zerofill_async (struct glfs_fd *glfd, off_t offset, size_t len,
+                      glfs_io_cbk fn, void *data)
+{
+        struct glfs_io *gio  = NULL;
+        int             ret  = 0;
+
+        gio = GF_CALLOC (1, sizeof (*gio), glfs_mt_glfs_io_t);
+        if (!gio) {
+                errno = ENOMEM;
+                return -1;
+        }
+
+        gio->op     = GF_FOP_ZEROFILL;
+        gio->glfd   = glfd;
+        gio->offset = offset;
+        gio->count  = len;
+        gio->fn     = fn;
+        gio->data   = data;
+
+        ret = synctask_new (glfs_from_glfd (glfd)->ctx->env,
+                            glfs_io_async_task, glfs_io_async_cbk,
+                            NULL, gio);
+
+        if (ret) {
+                GF_FREE (gio->iov);
+                GF_FREE (gio);
+        }
+
+        return ret;
+}
+
 
 void
 gf_dirent_to_dirent (gf_dirent_t *gf_dirent, struct dirent *dirent)
@@ -1799,7 +1926,7 @@ gf_dirent_to_dirent (gf_dirent_t *gf_dirent, struct dirent *dirent)
 	dirent->d_namlen = strlen (gf_dirent->d_name);
 #endif
 
-	strncpy (dirent->d_name, gf_dirent->d_name, 256);
+	strncpy (dirent->d_name, gf_dirent->d_name, GF_NAME_MAX + 1);
 }
 
 
@@ -1893,16 +2020,56 @@ glfd_entry_next (struct glfs_fd *glfd, int plus)
 }
 
 
+static struct dirent *
+glfs_readdirbuf_get (struct glfs_fd *glfd)
+{
+        struct dirent *buf = NULL;
+
+        LOCK (&glfd->fd->lock);
+        {
+                buf = glfd->readdirbuf;
+                if (buf) {
+                        memset (buf, 0, READDIRBUF_SIZE);
+                        goto unlock;
+                }
+
+                buf = GF_CALLOC (1, READDIRBUF_SIZE, glfs_mt_readdirbuf_t);
+                if (!buf) {
+                        errno = ENOMEM;
+                        goto unlock;
+                }
+
+                glfd->readdirbuf = buf;
+        }
+unlock:
+        UNLOCK (&glfd->fd->lock);
+
+        return buf;
+}
+
+
 int
-glfs_readdirplus_r (struct glfs_fd *glfd, struct stat *stat, struct dirent *buf,
+glfs_readdirplus_r (struct glfs_fd *glfd, struct stat *stat, struct dirent *ext,
 		    struct dirent **res)
 {
 	int              ret = 0;
 	gf_dirent_t     *entry = NULL;
+	struct dirent   *buf = NULL;
 
 	__glfs_entry_fd (glfd);
 
 	errno = 0;
+
+	if (ext)
+		buf = ext;
+	else
+		buf = glfs_readdirbuf_get (glfd);
+
+	if (!buf) {
+		errno = ENOMEM;
+		return -1;
+	}
+
 	entry = glfd_entry_next (glfd, !!stat);
 	if (errno)
 		ret = -1;
@@ -1931,6 +2098,28 @@ glfs_readdir_r (struct glfs_fd *glfd, struct dirent *buf, struct dirent **res)
 }
 
 
+struct dirent *
+glfs_readdirplus (struct glfs_fd *glfd, struct stat *stat)
+{
+        struct dirent *res = NULL;
+        int ret = -1;
+
+        ret = glfs_readdirplus_r (glfd, stat, NULL, &res);
+        if (ret)
+                return NULL;
+
+        return res;
+}
+
+
+
+struct dirent *
+glfs_readdir (struct glfs_fd *glfd)
+{
+        return glfs_readdirplus (glfd, NULL);
+}
+
+
 int
 glfs_statvfs (struct glfs *fs, const char *path, struct statvfs *buf)
 {
@@ -2737,6 +2926,36 @@ out:
 	return ret;
 }
 
+int
+glfs_zerofill (struct glfs_fd *glfd, off_t offset, size_t len)
+{
+        int               ret             = -1;
+        xlator_t         *subvol          = NULL;
+        fd_t             *fd              = NULL;
+
+        __glfs_entry_fd (glfd);
+
+        subvol = glfs_active_subvol (glfd->fs);
+        if (!subvol) {
+                errno = EIO;
+                goto out;
+        }
+
+        fd = glfs_resolve_fd (glfd->fs, subvol, glfd);
+        if (!fd) {
+                errno = EBADFD;
+                goto out;
+        }
+
+        ret = syncop_zerofill (subvol, fd, offset, len);
+out:
+        if (fd)
+                fd_unref(fd);
+
+        glfs_subvol_done (glfd->fs, subvol);
+
+        return ret;
+}
 
 int
 glfs_chdir (struct glfs *fs, const char *path)
diff --git a/api/src/glfs-handleops.c b/api/src/glfs-handleops.c
new file mode 100644
index 000000000..9c707a619
--- /dev/null
+++ b/api/src/glfs-handleops.c
@@ -0,0 +1,1278 @@
+/*
+ *  Copyright (c) 2013 Red Hat, Inc. <http://www.redhat.com>
+ *  This file is part of GlusterFS.
+ *
+ *  This file is licensed to you under your choice of the GNU Lesser
+ *  General Public License, version 3 or any later version (LGPLv3 or
+ *  later), or the GNU General Public License, version 2 (GPLv2), in all
+ *  cases as published by the Free Software Foundation.
+ */
+
+
+#include "glfs-internal.h"
+#include "glfs-mem-types.h"
+#include "syncop.h"
+#include "glfs.h"
+#include "glfs-handles.h"
+
+static void
+glfs_iatt_from_stat (struct stat *stat, int valid, struct iatt *iatt,
+		     int *glvalid)
+{
+	/* validate in args */
+	if ((stat == NULL) || (iatt == NULL) || (glvalid == NULL)) {
+		errno = EINVAL;
+		return;
+	}
+
+	*glvalid = 0;
+
+	if (valid & GFAPI_SET_ATTR_MODE) {
+		iatt->ia_prot = ia_prot_from_st_mode (stat->st_mode);
+		*glvalid |= GF_SET_ATTR_MODE;
+	}
+
+	if (valid & GFAPI_SET_ATTR_UID) {
+		iatt->ia_uid = stat->st_uid;
+		*glvalid |= GF_SET_ATTR_UID;
+	}
+
+	if (valid & GFAPI_SET_ATTR_GID) {
+		iatt->ia_gid = stat->st_gid;
+		*glvalid |= GF_SET_ATTR_GID;
+	}
+
+	if (valid & GFAPI_SET_ATTR_ATIME) {
+		iatt->ia_atime = stat->st_atime;
+		iatt->ia_atime_nsec = ST_ATIM_NSEC (stat);
+		*glvalid |= GF_SET_ATTR_ATIME;
+	}
+
+	if (valid & GFAPI_SET_ATTR_MTIME) {
+		iatt->ia_mtime = stat->st_mtime;
+		iatt->ia_mtime_nsec = ST_MTIM_NSEC (stat);
+		*glvalid |= GF_SET_ATTR_MTIME;
+	}
+
+	return;
+}
+
+struct glfs_object *
+glfs_h_lookupat (struct glfs *fs, struct glfs_object *parent,
+		 const char *path, struct stat *stat)
+{
+	int                      ret = 0;
+	xlator_t                *subvol = NULL;
+	inode_t                 *inode = NULL;
+	struct iatt              iatt = {0, };
+	struct glfs_object      *object = NULL;
+	loc_t                    loc = {0, };
+
+	/* validate in args */
+	if ((fs == NULL) || (path == NULL)) {
+		errno = EINVAL;
+		return NULL;
+	}
+
+	__glfs_entry_fs (fs);
+
+	/* get the active volume */
+	subvol = glfs_active_subvol (fs);
+	if (!subvol) {
+		errno = EIO;
+		goto out;
+	}
+
+	/* get/refresh the in arg objects inode in correlation to the xlator */
+	if (parent) {
+		inode = glfs_resolve_inode (fs, subvol, parent);
+		if (!inode) {
+			errno = ESTALE;
+			goto out;
+		}
+	}
+
+	/* fop/op */
+	ret = glfs_resolve_at (fs, subvol, inode, path, &loc, &iatt,
+			       0 /*TODO: links? */, 0);
+
+	/* populate out args */
+	if (!ret) {
+		if (stat)
+			glfs_iatt_to_stat (fs, &iatt, stat);
+
+		ret = glfs_create_object (&loc, &object);
+	}
+
+out:
+	loc_wipe (&loc);
+
+	if (inode)
+		inode_unref (inode);
+
+	glfs_subvol_done (fs, subvol);
+
+	return object;
+}
+
+int
+glfs_h_stat (struct glfs *fs, struct glfs_object *object, struct stat *stat)
+{
+	int              ret = -1;
+	xlator_t        *subvol = NULL;
+	inode_t         *inode = NULL;
+	loc_t            loc = {0, };
+	struct iatt      iatt = {0, };
+
+	/* validate in args */
+	if ((fs == NULL) || (object == NULL)) {
+		errno = EINVAL;
+		return -1;
+	}
+
+	__glfs_entry_fs (fs);
+
+	/* get the active volume */
+	subvol = glfs_active_subvol (fs);
+	if (!subvol) {
+		ret = -1;
+		errno = EIO;
+		goto out;
+	}
+
+	/* get/refresh the in arg objects inode in correlation to the xlator */
+	inode = glfs_resolve_inode (fs, subvol, object);
+	if (!inode) {
+		errno = ESTALE;
+		goto out;
+	}
+
+	/* populate loc */
+	GLFS_LOC_FILL_INODE (inode, loc, out);
+
+	/* fop/op */
+	ret = syncop_stat (subvol, &loc, &iatt);
+
+	/* populate out args */
+	if (!ret && stat) {
+		glfs_iatt_to_stat (fs, &iatt, stat);
+	}
+out:
+	loc_wipe (&loc);
+
+	if (inode)
+		inode_unref (inode);
+
+	glfs_subvol_done (fs, subvol);
+
+	return ret;
+}
+
+int
+glfs_h_getattrs (struct glfs *fs, struct glfs_object *object, struct stat *stat)
+{
+	int                      ret = 0;
+	xlator_t                *subvol = NULL;
+	inode_t                 *inode = NULL;
+	struct iatt              iatt = {0, };
+
+	/* validate in args */
+	if ((fs == NULL) || (object == NULL)) {
+		errno = EINVAL;
+		return -1;
+	}
+
+	__glfs_entry_fs (fs);
+
+	/* get the active volume */
+	subvol = glfs_active_subvol (fs);
+	if (!subvol) {
+		ret = -1;
+		errno = EIO;
+		goto out;
+	}
+
+	/* get/refresh the in arg objects inode in correlation to the xlator */
+	inode = glfs_resolve_inode (fs, subvol, object);
+	if (!inode) {
+		errno = ESTALE;
+		goto out;
+	}
+
+	/* fop/op */
+	ret = glfs_resolve_base (fs, subvol, inode, &iatt);
+
+	/* populate out args */
+	if (!ret && stat) {
+		glfs_iatt_to_stat (fs, &iatt, stat);
+	}
+
+out:
+	if (inode)
+		inode_unref (inode);
+
+	glfs_subvol_done (fs, subvol);
+
+	return ret;
+}
+
+int
+glfs_h_setattrs (struct glfs *fs, struct glfs_object *object, struct stat *stat,
+		 int valid)
+{
+	int              ret = -1;
+	xlator_t        *subvol = NULL;
+	inode_t         *inode = NULL;
+	loc_t            loc = {0, };
+	struct iatt      iatt = {0, };
+	int              glvalid = 0;
+
+	/* validate in args */
+	if ((fs == NULL) || (object == NULL) || (stat == NULL)) {
+		errno = EINVAL;
+		return -1;
+	}
+
+	__glfs_entry_fs (fs);
+
+	/* get the active volume */
+	subvol = glfs_active_subvol (fs);
+	if (!subvol) {
+		ret = -1;
+		errno = EIO;
+		goto out;
+	}
+
+	/* get/refresh the in arg objects inode in correlation to the xlator */
+	inode = glfs_resolve_inode (fs, subvol, object);
+	if (!inode) {
+		errno = ESTALE;
+		goto out;
+	}
+
+	/* map valid masks from in args */
+	glfs_iatt_from_stat (stat, valid, &iatt, &glvalid);
+
+	/* populate loc */
+	GLFS_LOC_FILL_INODE (inode, loc, out);
+
+	/* fop/op */
+	ret = syncop_setattr (subvol, &loc, &iatt, glvalid, 0, 0);
+out:
+	loc_wipe (&loc);
+
+	if (inode)
+		inode_unref (inode);
+
+	glfs_subvol_done (fs, subvol);
+
+	return ret;
+}
+
+struct glfs_fd *
+glfs_h_open (struct glfs *fs, struct glfs_object *object, int flags)
+{
+	int              ret = -1;
+	struct glfs_fd  *glfd = NULL;
+	xlator_t        *subvol = NULL;
+	inode_t         *inode = NULL;
+	loc_t            loc = {0, };
+
+	/* validate in args */
+	if ((fs == NULL) || (object == NULL)) {
+		errno = EINVAL;
+		return NULL;
+	}
+
+	__glfs_entry_fs (fs);
+
+	/* get the active volume */
+	subvol = glfs_active_subvol (fs);
+	if (!subvol) {
+		errno = EIO;
+		goto out;
+	}
+
+	/* get/refresh the in arg objects inode in correlation to the xlator */
+	inode = glfs_resolve_inode (fs, subvol, object);
+	if (!inode) {
+		errno = ESTALE;
+		goto out;
+	}
+
+	/* check types to open */
+	if (IA_ISDIR (inode->ia_type)) {
+		ret = -1;
+		errno = EISDIR;
+		goto out;
+	}
+
+	if (!IA_ISREG (inode->ia_type)) {
+		ret = -1;
+		errno = EINVAL;
+		goto out;
+	}
+
+	glfd = glfs_fd_new (fs);
+	if (!glfd) {
+		errno = ENOMEM;
+		goto out;
+	}
+
+	glfd->fd = fd_create (inode, getpid());
+	if (!glfd->fd) {
+		ret = -1;
+		errno = ENOMEM;
+		goto out;
+	}
+
+	/* populate loc */
+	GLFS_LOC_FILL_INODE (inode, loc, out);
+
+	/* fop/op */
+	ret = syncop_open (subvol, &loc, flags, glfd->fd);
+
+out:
+	loc_wipe (&loc);
+
+	if (inode)
+		inode_unref (inode);
+
+	if (ret && glfd) {
+		glfs_fd_destroy (glfd);
+		glfd = NULL;
+	} else {
+		glfd->fd->flags = flags;
+		fd_bind (glfd->fd);
+		glfs_fd_bind (glfd);
+	}
+
+	glfs_subvol_done (fs, subvol);
+
+	return glfd;
+}
+
+struct glfs_object *
+glfs_h_creat (struct glfs *fs, struct glfs_object *parent, const char *path,
+	      int flags, mode_t mode, struct stat *stat)
+{
+	int                 ret = -1;
+	struct glfs_fd     *glfd = NULL;
+	xlator_t           *subvol = NULL;
+	inode_t            *inode = NULL;
+	loc_t               loc = {0, };
+	struct iatt         iatt = {0, };
+	uuid_t              gfid;
+	dict_t             *xattr_req = NULL;
+	struct glfs_object *object = NULL;
+
+	/* validate in args */
+	if ((fs == NULL) || (parent == NULL) || (path == NULL)) {
+		errno = EINVAL;
+		return NULL;
+	}
+
+	__glfs_entry_fs (fs);
+
+	/* get the active volume */
+	subvol = glfs_active_subvol (fs);
+	if (!subvol) {
+		ret = -1;
+		errno = EIO;
+		goto out;
+	}
+
+	/* get/refresh the in arg objects inode in correlation to the xlator */
+	inode = glfs_resolve_inode (fs, subvol, parent);
+	if (!inode) {
+		errno = ESTALE;
+		goto out;
+	}
+
+	xattr_req = dict_new ();
+	if (!xattr_req) {
+		ret = -1;
+		errno = ENOMEM;
+		goto out;
+	}
+
+	uuid_generate (gfid);
+	ret = dict_set_static_bin (xattr_req, "gfid-req", gfid, 16);
+	if (ret) {
+		ret = -1;
+		errno = ENOMEM;
+		goto out;
+	}
+
+	GLFS_LOC_FILL_PINODE (inode, loc, ret, errno, out, path);
+
+	glfd = glfs_fd_new (fs);
+	if (!glfd)
+		goto out;
+
+	glfd->fd = fd_create (loc.inode, getpid());
+	if (!glfd->fd) {
+		ret = -1;
+		errno = ENOMEM;
+		goto out;
+	}
+
+	/* fop/op */
+	ret = syncop_create (subvol, &loc, flags, mode, glfd->fd,
+			     xattr_req, &iatt);
+
+	/* populate out args */
+	if (ret == 0) {
+		/* TODO: If the inode existed in the cache (say file already
+		   exists), then the glfs_loc_link will not update the
+		   loc.inode, as a result we will have a 0000 GFID that we
+		   would copy out to the object, this needs to be fixed.
+		*/
+		ret = glfs_loc_link (&loc, &iatt);
+		if (ret != 0) {
+			goto out;
+		}
+
+		if (stat)
+			glfs_iatt_to_stat (fs, &iatt, stat);
+
+		ret = glfs_create_object (&loc, &object);
+	}
+
+out:
+	if (ret && object != NULL) {
+		glfs_h_close (object);
+		object = NULL;
+	}
+
+	loc_wipe(&loc);
+
+	if (inode)
+		inode_unref (inode);
+
+	if (xattr_req)
+		dict_unref (xattr_req);
+
+	if (glfd) {
+		glfs_fd_destroy (glfd);
+		glfd = NULL;
+	}
+
+	glfs_subvol_done (fs, subvol);
+
+	return object;
+}
+
+struct glfs_object *
+glfs_h_mkdir (struct glfs *fs, struct glfs_object *parent, const char *path,
+	      mode_t mode, struct stat *stat)
+{
+	int                 ret = -1;
+	xlator_t           *subvol = NULL;
+	inode_t            *inode = NULL;
+	loc_t               loc = {0, };
+	struct iatt         iatt = {0, };
+	uuid_t              gfid;
+	dict_t             *xattr_req = NULL;
+	struct glfs_object *object = NULL;
+
+	/* validate in args */
+	if ((fs == NULL) || (parent == NULL) || (path == NULL)) {
+		errno = EINVAL;
+		return NULL;
+	}
+
+	__glfs_entry_fs (fs);
+
+	/* get the active volume */
+	subvol = glfs_active_subvol (fs);
+	if (!subvol) {
+		ret = -1;
+		errno = EIO;
+		goto out;
+	}
+
+	/* get/refresh the in arg objects inode in correlation to the xlator */
+	inode = glfs_resolve_inode (fs, subvol, parent);
+	if (!inode) {
+		errno = ESTALE;
+		goto out;
+	}
+
+	xattr_req = dict_new ();
+	if (!xattr_req) {
+		ret = -1;
+		errno = ENOMEM;
+		goto out;
+	}
+
+	uuid_generate (gfid);
+	ret = dict_set_static_bin (xattr_req, "gfid-req", gfid, 16);
+	if (ret) {
+		ret = -1;
+		errno = ENOMEM;
+		goto out;
+	}
+
+	GLFS_LOC_FILL_PINODE (inode, loc, ret, errno, out, path);
+
+	/* fop/op */
+	ret = syncop_mkdir (subvol, &loc, mode, xattr_req, &iatt);
+
+	/* populate out args */
+	if ( ret == 0 )  {
+		ret = glfs_loc_link (&loc, &iatt);
+		if (ret != 0) {
+			goto out;
+		}
+
+		if (stat)
+			glfs_iatt_to_stat (fs, &iatt, stat);
+
+		ret = glfs_create_object (&loc, &object);
+	}
+
+out:
+	if (ret && object != NULL) {
+		glfs_h_close (object);
+		object = NULL;
+	}
+
+	loc_wipe(&loc);
+
+	if (inode)
+		inode_unref (inode);
+
+	if (xattr_req)
+		dict_unref (xattr_req);
+
+	glfs_subvol_done (fs, subvol);
+
+	return object;
+}
+
+struct glfs_object *
+glfs_h_mknod (struct glfs *fs, struct glfs_object *parent, const char *path,
+	      mode_t mode, dev_t dev, struct stat *stat)
+{
+	int                 ret = -1;
+	xlator_t           *subvol = NULL;
+	inode_t            *inode = NULL;
+	loc_t               loc = {0, };
+	struct iatt         iatt = {0, };
+	uuid_t              gfid;
+	dict_t             *xattr_req = NULL;
+	struct glfs_object *object = NULL;
+
+	/* validate in args */
+	if ((fs == NULL) || (parent == NULL) || (path == NULL)) {
+		errno = EINVAL;
+		return NULL;
+	}
+
+	__glfs_entry_fs (fs);
+
+	/* get the active volume */
+	subvol = glfs_active_subvol (fs);
+	if (!subvol) {
+		ret = -1;
+		errno = EIO;
+		goto out;
+	}
+
+	/* get/refresh the in arg objects inode in correlation to the xlator */
+	inode = glfs_resolve_inode (fs, subvol, parent);
+	if (!inode) {
+		errno = ESTALE;
+		goto out;
+	}
+
+	xattr_req = dict_new ();
+	if (!xattr_req) {
+		ret = -1;
+		errno = ENOMEM;
+		goto out;
+	}
+
+	uuid_generate (gfid);
+	ret = dict_set_static_bin (xattr_req, "gfid-req", gfid, 16);
+	if (ret) {
+		ret = -1;
+		errno = ENOMEM;
+		goto out;
+	}
+
+	GLFS_LOC_FILL_PINODE (inode, loc, ret, errno, out, path);
+
+	/* fop/op */
+	ret = syncop_mknod (subvol, &loc, mode, dev, xattr_req, &iatt);
+
+	/* populate out args */
+	if (ret == 0) {
+		ret = glfs_loc_link (&loc, &iatt);
+		if (ret != 0) {
+			goto out;
+		}
+
+		if (stat)
+			glfs_iatt_to_stat (fs, &iatt, stat);
+
+		ret = glfs_create_object (&loc, &object);
+	}
+out:
+	if (ret && object != NULL) {
+		glfs_h_close (object);
+		object = NULL;
+	}
+
+	loc_wipe(&loc);
+
+	if (inode)
+		inode_unref (inode);
+
+	if (xattr_req)
+		dict_unref (xattr_req);
+
+	glfs_subvol_done (fs, subvol);
+
+	return object;
+}
+
+int
+glfs_h_unlink (struct glfs *fs, struct glfs_object *parent, const char *path)
+{
+	int                 ret = -1;
+	xlator_t           *subvol = NULL;
+	inode_t            *inode = NULL;
+	loc_t               loc = {0, };
+
+	/* validate in args */
+	if ((fs == NULL) || (parent == NULL) || (path == NULL)) {
+		errno = EINVAL;
+		return -1;
+	}
+
+	__glfs_entry_fs (fs);
+
+	/* get the active volume */
+	subvol = glfs_active_subvol (fs);
+	if ( !subvol ) {
+		ret = -1;
+		errno = EIO;
+		goto out;
+	}
+
+	/* get/refresh the in arg objects inode in correlation to the xlator */
+	inode = glfs_resolve_inode (fs, subvol, parent);
+	if (!inode) {
+		errno = ESTALE;
+		goto out;
+	}
+
+	ret = glfs_resolve_at (fs, subvol, inode, path, &loc, NULL, 0 , 0);
+	if (ret != 0) {
+		goto out;
+	}
+
+	if (!IA_ISDIR(loc.inode->ia_type)) {
+		ret = syncop_unlink (subvol, &loc);
+		if (ret != 0) {
+			goto out;
+		}
+	} else {
+		ret = syncop_rmdir (subvol, &loc);
+		if (ret != 0) {
+			goto out;
+		}
+	}
+
+	if (ret == 0)
+		ret = glfs_loc_unlink (&loc);
+
+out:
+	loc_wipe (&loc);
+
+	if (inode)
+		inode_unref (inode);
+
+	glfs_subvol_done (fs, subvol);
+
+	return ret;
+}
+
+struct glfs_fd *
+glfs_h_opendir (struct glfs *fs, struct glfs_object *object)
+{
+	int              ret = -1;
+	struct glfs_fd  *glfd = NULL;
+	xlator_t        *subvol = NULL;
+	inode_t         *inode = NULL;
+	loc_t            loc = {0, };
+
+	/* validate in args */
+	if ((fs == NULL) || (object == NULL)) {
+		errno = EINVAL;
+		return NULL;
+	}
+
+	__glfs_entry_fs (fs);
+
+	/* get the active volume */
+	subvol = glfs_active_subvol (fs);
+	if (!subvol) {
+		ret = -1;
+		errno = EIO;
+		goto out;
+	}
+
+	/* get/refresh the in arg objects inode in correlation to the xlator */
+	inode = glfs_resolve_inode (fs, subvol, object);
+	if (!inode) {
+		errno = ESTALE;
+		goto out;
+	}
+
+	if (!IA_ISDIR (inode->ia_type)) {
+		ret = -1;
+		errno = ENOTDIR;
+		goto out;
+	}
+
+	glfd = glfs_fd_new (fs);
+	if (!glfd)
+		goto out;
+
+	INIT_LIST_HEAD (&glfd->entries);
+
+	glfd->fd = fd_create (inode, getpid());
+	if (!glfd->fd) {
+		ret = -1;
+		errno = ENOMEM;
+		goto out;
+	}
+
+	GLFS_LOC_FILL_INODE (inode, loc, out);
+
+	/* fop/op */
+	ret = syncop_opendir (subvol, &loc, glfd->fd);
+
+out:
+	loc_wipe (&loc);
+
+	if (inode)
+		inode_unref (inode);
+
+	if (ret && glfd) {
+		glfs_fd_destroy (glfd);
+		glfd = NULL;
+	} else {
+		fd_bind (glfd->fd);
+		glfs_fd_bind (glfd);
+	}
+
+	glfs_subvol_done (fs, subvol);
+
+	return glfd;
+}
+
+ssize_t
+glfs_h_extract_handle (struct glfs_object *object, unsigned char *handle,
+		       int len)
+{
+	ssize_t ret = -1;
+
+	/* validate in args */
+	if (object == NULL) {
+		errno = EINVAL;
+		goto out;
+	}
+
+	if (!handle || !len) {
+		ret = GFAPI_HANDLE_LENGTH;
+		goto out;
+	}
+
+	if (len < GFAPI_HANDLE_LENGTH)
+	{
+		errno = ERANGE;
+		goto out;
+	}
+
+	memcpy (handle, object->gfid, GFAPI_HANDLE_LENGTH);
+
+	ret = GFAPI_HANDLE_LENGTH;
+
+out:
+	return ret;
+}
+
+struct glfs_object *
+glfs_h_create_from_handle (struct glfs *fs, unsigned char *handle, int len,
+			   struct stat *stat)
+{
+	loc_t               loc = {0, };
+	int                 ret = -1;
+	struct iatt         iatt = {0, };
+	inode_t            *newinode = NULL;
+	xlator_t           *subvol = NULL;
+	struct glfs_object *object = NULL;
+
+	/* validate in args */
+	if ((fs == NULL) || (handle == NULL) || (len != GFAPI_HANDLE_LENGTH)) {
+		errno = EINVAL;
+		return NULL;
+	}
+
+	__glfs_entry_fs (fs);
+
+	/* get the active volume */
+	subvol = glfs_active_subvol (fs);
+	if (!subvol) {
+		errno = EIO;
+		goto out;
+	}
+
+	memcpy (loc.gfid, handle, GFAPI_HANDLE_LENGTH);
+
+	newinode = inode_find (subvol->itable, loc.gfid);
+	if (newinode)
+		loc.inode = newinode;
+	else {
+		loc.inode = inode_new (subvol->itable);
+		if (!loc.inode) {
+			errno = ENOMEM;
+			goto out;
+		}
+	}
+
+	ret = syncop_lookup (subvol, &loc, 0, &iatt, 0, 0);
+	if (ret) {
+		gf_log (subvol->name, GF_LOG_WARNING,
+			"inode refresh of %s failed: %s",
+			uuid_utoa (loc.gfid), strerror (errno));
+		goto out;
+	}
+
+	newinode = inode_link (loc.inode, 0, 0, &iatt);
+	if (newinode)
+		inode_lookup (newinode);
+	else {
+		gf_log (subvol->name, GF_LOG_WARNING,
+			"inode linking of %s failed: %s",
+			uuid_utoa (loc.gfid), strerror (errno));
+		errno = EINVAL;
+		goto out;
+	}
+
+	/* populate stat */
+	if (stat)
+		glfs_iatt_to_stat (fs, &iatt, stat);
+
+	object = GF_CALLOC (1, sizeof(struct glfs_object),
+			    glfs_mt_glfs_object_t);
+	if (object == NULL) {
+		errno = ENOMEM;
+		ret = -1;
+		goto out;
+	}
+
+	/* populate the return object */
+	object->inode = newinode;
+	uuid_copy (object->gfid, object->inode->gfid);
+
+out:
+	/* TODO: Check where the inode ref is being held? */
+	loc_wipe (&loc);
+
+	glfs_subvol_done (fs, subvol);
+
+	return object;
+}
+
+int
+glfs_h_close (struct glfs_object *object)
+{
+	/* Release the held reference */
+	inode_unref (object->inode);
+	GF_FREE (object);
+
+	return 0;
+}
+
+int
+glfs_h_truncate (struct glfs *fs, struct glfs_object *object, off_t offset)
+{
+	loc_t               loc = {0, };
+	int                 ret = -1;
+	xlator_t           *subvol = NULL;
+	inode_t            *inode = NULL;
+
+	/* validate in args */
+	if ((fs == NULL) || (object == NULL)) {
+		errno = EINVAL;
+		return -1;
+	}
+
+	__glfs_entry_fs (fs);
+
+	/* get the active volume */
+	subvol = glfs_active_subvol (fs);
+	if (!subvol) {
+		ret = -1;
+		errno = EIO;
+		goto out;
+	}
+
+	/* get/refresh the in arg objects inode in correlation to the xlator */
+	inode = glfs_resolve_inode (fs, subvol, object);
+	if (!inode) {
+		errno = ESTALE;
+		goto out;
+	}
+
+	GLFS_LOC_FILL_INODE (inode, loc, out);
+
+	/* fop/op */
+	ret = syncop_truncate (subvol, &loc, (off_t)offset);
+
+	/* populate out args */
+	if (ret == 0)
+		ret = glfs_loc_unlink (&loc);
+
+out:
+	loc_wipe (&loc);
+
+	if (inode)
+		inode_unref (inode);
+
+	glfs_subvol_done (fs, subvol);
+
+	return ret;
+}
+
+struct glfs_object *
+glfs_h_symlink (struct glfs *fs, struct glfs_object *parent, const char *name,
+		const char *data, struct stat *stat)
+{
+	int                 ret = -1;
+	xlator_t           *subvol = NULL;
+	inode_t            *inode = NULL;
+	loc_t               loc = {0, };
+	struct iatt         iatt = {0, };
+	uuid_t              gfid;
+	dict_t             *xattr_req = NULL;
+	struct glfs_object *object = NULL;
+
+	/* validate in args */
+	if ((fs == NULL) || (parent == NULL) || (name == NULL) ||
+		(data == NULL)) {
+		errno = EINVAL;
+		return NULL;
+	}
+
+	__glfs_entry_fs (fs);
+
+	/* get the active volume */
+	subvol = glfs_active_subvol (fs);
+	if (!subvol) {
+		ret = -1;
+		errno = EIO;
+		goto out;
+	}
+
+	/* get/refresh the in arg objects inode in correlation to the xlator */
+	inode = glfs_resolve_inode (fs, subvol, parent);
+	if (!inode) {
+		errno = ESTALE;
+		goto out;
+	}
+
+	xattr_req = dict_new ();
+	if (!xattr_req) {
+		ret = -1;
+		errno = ENOMEM;
+		goto out;
+	}
+
+	uuid_generate (gfid);
+	ret = dict_set_static_bin (xattr_req, "gfid-req", gfid, 16);
+	if (ret) {
+		ret = -1;
+		errno = ENOMEM;
+		goto out;
+	}
+
+	GLFS_LOC_FILL_PINODE (inode, loc, ret, errno, out, name);
+
+	/* fop/op */
+	ret = syncop_symlink (subvol, &loc, data, xattr_req, &iatt);
+
+	/* populate out args */
+	if (ret == 0) {
+		/* TODO: If the inode existed in the cache (say file already
+		 * exists), then the glfs_loc_link will not update the
+		 * loc.inode, as a result we will have a 0000 GFID that we
+		 * would copy out to the object, this needs to be fixed.
+		 */
+		ret = glfs_loc_link (&loc, &iatt);
+		if (ret != 0) {
+			goto out;
+		}
+
+		if (stat)
+			glfs_iatt_to_stat (fs, &iatt, stat);
+
+		ret = glfs_create_object (&loc, &object);
+	}
+
+out:
+	if (ret && object != NULL) {
+		glfs_h_close (object);
+		object = NULL;
+	}
+
+	loc_wipe(&loc);
+
+	if (inode)
+		inode_unref (inode);
+
+	if (xattr_req)
+		dict_unref (xattr_req);
+
+	glfs_subvol_done (fs, subvol);
+
+	return object;
+}
+
+int
+glfs_h_readlink (struct glfs *fs, struct glfs_object *object, char *buf,
+		 size_t bufsiz)
+{
+	loc_t               loc = {0, };
+	int                 ret = -1;
+	xlator_t           *subvol = NULL;
+	inode_t            *inode = NULL;
+	char               *linkval = NULL;
+
+	/* validate in args */
+	if ((fs == NULL) || (object == NULL) || (buf == NULL)) {
+		errno = EINVAL;
+		return -1;
+	}
+
+	__glfs_entry_fs (fs);
+
+	/* get the active volume */
+	subvol = glfs_active_subvol (fs);
+	if (!subvol) {
+		ret = -1;
+		errno = EIO;
+		goto out;
+	}
+
+	/* get/refresh the in arg objects inode in correlation to the xlator */
+	inode = glfs_resolve_inode (fs, subvol, object);
+	if (!inode) {
+		errno = ESTALE;
+		goto out;
+	}
+
+	GLFS_LOC_FILL_INODE (inode, loc, out);
+
+	/* fop/op */
+	ret = syncop_readlink (subvol, &loc, &linkval, bufsiz);
+
+	/* populate out args */
+	if (ret > 0)
+		memcpy (buf, linkval, ret);
+
+out:
+	loc_wipe (&loc);
+
+	if (inode)
+		inode_unref (inode);
+
+	if (linkval)
+		GF_FREE (linkval);
+
+	glfs_subvol_done (fs, subvol);
+
+	return ret;
+}
+
+int
+glfs_h_link (struct glfs *fs, struct glfs_object *linksrc,
+	     struct glfs_object *parent, const char *name)
+{
+	int                 ret = -1;
+	xlator_t           *subvol = NULL;
+	inode_t            *inode = NULL;
+	inode_t            *pinode = NULL;
+	loc_t               oldloc = {0, };
+	loc_t               newloc = {0, };
+
+	/* validate in args */
+	if ((fs == NULL) || (linksrc == NULL) || (parent == NULL) ||
+		(name == NULL)) {
+		errno = EINVAL;
+		return -1;
+	}
+
+	__glfs_entry_fs (fs);
+
+	/* get the active volume */
+	subvol = glfs_active_subvol (fs);
+	if (!subvol) {
+		ret = -1;
+		errno = EIO;
+		goto out;
+	}
+
+	/* get/refresh the in arg objects inode in correlation to the xlator */
+	inode = glfs_resolve_inode (fs, subvol, linksrc);
+	if (!inode) {
+		errno = ESTALE;
+		goto out;
+	}
+
+	if (inode->ia_type == IA_IFDIR) {
+		ret = -1;
+		errno = EISDIR;
+		goto out;
+	}
+
+	GLFS_LOC_FILL_INODE (inode, oldloc, out);
+
+	/* get/refresh the in arg objects inode in correlation to the xlator */
+	pinode = glfs_resolve_inode (fs, subvol, parent);
+	if (!pinode) {
+		errno = ESTALE;
+		goto out;
+	}
+
+	/* setup newloc based on parent */
+	newloc.parent = inode_ref (pinode);
+	newloc.name = name;
+	ret = glfs_loc_touchup (&newloc);
+	if (ret != 0) {
+		errno = EINVAL;
+		goto out;
+	}
+
+	/* Filling the inode of the hard link to be same as that of the
+	 * original file
+	 */
+	newloc.inode = inode_ref (inode);
+
+	/* fop/op */
+	ret = syncop_link (subvol, &oldloc, &newloc);
+
+	if (ret == 0)
+		/* TODO: No iatt to pass as there has been no lookup */
+		ret = glfs_loc_link (&newloc, NULL);
+out:
+	loc_wipe (&oldloc);
+	loc_wipe (&newloc);
+
+	if (inode)
+		inode_unref (inode);
+
+	if (pinode)
+		inode_unref (pinode);
+
+	glfs_subvol_done (fs, subvol);
+
+	return ret;
+}
+
+int
+glfs_h_rename (struct glfs *fs, struct glfs_object *olddir, const char *oldname,
+	       struct glfs_object *newdir, const char *newname)
+{
+	int                 ret = -1;
+	xlator_t           *subvol = NULL;
+	inode_t            *oldpinode = NULL;
+	inode_t            *newpinode = NULL;
+	loc_t               oldloc = {0, };
+	loc_t               newloc = {0, };
+	struct iatt         oldiatt = {0, };
+	struct iatt         newiatt = {0, };
+
+	/* validate in args */
+	if ((fs == NULL) || (olddir == NULL) || (oldname == NULL) ||
+		(newdir == NULL) || (newname == NULL)) {
+		errno = EINVAL;
+		return -1;
+	}
+
+	__glfs_entry_fs (fs);
+
+	/* get the active volume */
+	subvol = glfs_active_subvol (fs);
+	if ( !subvol ) {
+		ret = -1;
+		errno = EIO;
+		goto out;
+	}
+
+	/* get/refresh the in arg objects inode in correlation to the xlator */
+	oldpinode = glfs_resolve_inode (fs, subvol, olddir);
+	if (!oldpinode) {
+		errno = ESTALE;
+		goto out;
+	}
+
+	ret = glfs_resolve_at (fs, subvol, oldpinode, oldname, &oldloc,
+			       &oldiatt, 0 , 0);
+	if (ret != 0) {
+		goto out;
+	}
+
+	/* get/refresh the in arg objects inode in correlation to the xlator */
+	newpinode = glfs_resolve_inode (fs, subvol, newdir);
+	if (!newpinode) {
+		errno = ESTALE;
+		goto out;
+	}
+
+	ret = glfs_resolve_at (fs, subvol, newpinode, newname, &newloc,
+			       &newiatt, 0, 0);
+
+	if (ret && errno != ENOENT && newloc.parent)
+		goto out;
+
+	if (newiatt.ia_type != IA_INVAL) {
+		if ((oldiatt.ia_type == IA_IFDIR) !=
+			(newiatt.ia_type == IA_IFDIR)) {
+			/* Either both old and new must be dirs,
+			 * or both must be non-dirs. Else, fail.
+			 */
+			ret = -1;
+			errno = EISDIR;
+			goto out;
+		}
+	}
+
+	/* TODO: check if new or old is a prefix of the other, and fail EINVAL */
+
+	ret = syncop_rename (subvol, &oldloc, &newloc);
+
+	if (ret == 0)
+		inode_rename (oldloc.parent->table, oldloc.parent, oldloc.name,
+			      newloc.parent, newloc.name, oldloc.inode,
+			      &oldiatt);
+
+out:
+	loc_wipe (&oldloc);
+	loc_wipe (&newloc);
+
+	if (oldpinode)
+		inode_unref (oldpinode);
+
+	if (newpinode)
+		inode_unref (newpinode);
+
+	glfs_subvol_done (fs, subvol);
+
+	return ret;
+}
diff --git a/api/src/glfs-handles.h b/api/src/glfs-handles.h
new file mode 100644
index 000000000..437f2cbc8
--- /dev/null
+++ b/api/src/glfs-handles.h
@@ -0,0 +1,143 @@
+/*
+  Copyright (c) 2013 Red Hat, Inc. <http://www.redhat.com>
+  This file is part of GlusterFS.
+
+  This file is licensed to you under your choice of the GNU Lesser
+  General Public License, version 3 or any later version (LGPLv3 or
+  later), or the GNU General Public License, version 2 (GPLv2), in all
+  cases as published by the Free Software Foundation.
+*/
+
+#ifndef _GLFS_HANDLES_H
+#define _GLFS_HANDLES_H
+
+#include "glfs.h"
+
+/* GLFS OBJECT BASED OPERATIONS
+ *
+ * The following APIs are introduced to provide an API framework that can work
+ * with gluster objects (files and directories), instead of absolute paths.
+ *
+ * The following API set can be related to the POSIX *at interfaces (like
+ * openat (2)). The intention of these APIs is to be able to operate based
+ * on parent object and looking up or creating child objects within, OR to be
+ * used on the actual object thus looked up or created, and retrieve information
+ * regarding the same.
+ *
+ * The APIs also provide for generating an opaque invariant handle to the
+ * object, that can later be used to lookup the object, instead of the regular
+ * glfs_h_* variants. The APIs that provide this behaviour are,
+ * glfs_h_extract_handle and glfs_h_create_from_handle.
+ *
+ * The object handles can be transitioned to fd based operations as supported
+ * by glfs.h calls, using the glfs_h_open call. This provides a way to move
+ * from objects to fd's akin to moving from path to fd for required operations.
+ *
+ * NOTE: The opaque invariant handle is the GFID of the object in reality, but
+ * maintained as an opaque data value, for potential internal changes to the
+ * same without impacting the caller.
+ *
+ * NOTE: Currently looking up an object can create multiple object handles to
+ * the same, i.e distinct glfs_object *. Hence each such looked up or received
+ * handle from other calls, would need to be closed. In the future, for a given
+ * object these pointers would be the same, and an ease of use API to forget all
+ * instances of this bject would be provided (instead of a per lookup close).
+ * This should not change the APIs in their current form.
+ *
+ */
+
+/* Values for valid falgs to be used when using XXXsetattr, to set multiple
+ attribute values passed via the related stat structure.
+ */
+#define GFAPI_SET_ATTR_MODE  0x1
+#define GFAPI_SET_ATTR_UID   0x2
+#define GFAPI_SET_ATTR_GID   0x4
+#define GFAPI_SET_ATTR_SIZE  0x8
+#define GFAPI_SET_ATTR_ATIME 0x10
+#define GFAPI_SET_ATTR_MTIME 0x20
+
+/* Handle length for object handles returned from glfs_h_extract_handle or
+ * glfs_h_create_from_handle */
+#define GFAPI_HANDLE_LENGTH 16
+
+__BEGIN_DECLS
+
+/*
+ * Notes:
+ *
+ * The file object handle. One per looked up, created file/directory
+ *
+ * This had been introduced to facilitate gfid/inode based gfapi
+ * - a requirement introduced by nfs-ganesha
+ */
+struct glfs_object;
+typedef struct glfs_object glfs_object_t;
+
+/* Handle based operations */
+/* Operations that generate handles */
+struct glfs_object *glfs_h_lookupat (struct glfs *fs,
+				     struct glfs_object *parent,
+				     const char *path, struct stat *stat);
+
+struct glfs_object *glfs_h_creat (struct glfs *fs, struct glfs_object *parent,
+				  const char *path, int flags, mode_t mode,
+				  struct stat *sb);
+
+struct glfs_object *glfs_h_mkdir (struct glfs *fs, struct glfs_object *parent,
+				  const char *path, mode_t flags,
+				  struct stat *sb);
+
+struct glfs_object *glfs_h_mknod (struct glfs *fs, struct glfs_object *parent,
+				  const char *path, mode_t mode, dev_t dev,
+				  struct stat *sb);
+
+struct glfs_object *glfs_h_symlink (struct glfs *fs, struct glfs_object *parent,
+				    const char *name, const char *data,
+				    struct stat *stat);
+
+/* Operations on the actual objects */
+int glfs_h_unlink (struct glfs *fs, struct glfs_object *parent,
+		   const char *path);
+
+int glfs_h_close (struct glfs_object *object);
+
+int glfs_caller_specific_init (void *uid_caller_key, void *gid_caller_key,
+			       void *future);
+
+int glfs_h_truncate (struct glfs *fs, struct glfs_object *object, off_t offset);
+
+int glfs_h_stat(struct glfs *fs, struct glfs_object *object, struct stat *stat);
+
+int glfs_h_getattrs (struct glfs *fs, struct glfs_object *object,
+		     struct stat *stat);
+
+int glfs_h_setattrs (struct glfs *fs, struct glfs_object *object,
+		     struct stat *sb, int valid);
+
+int glfs_h_readlink (struct glfs *fs, struct glfs_object *object, char *buf,
+		     size_t bufsiz);
+
+int glfs_h_link (struct glfs *fs, struct glfs_object *linktgt,
+		 struct glfs_object *parent, const char *name);
+
+int glfs_h_rename (struct glfs *fs, struct glfs_object *olddir,
+		   const char *oldname, struct glfs_object *newdir,
+		   const char *newname);
+
+/* Operations enabling opaque invariant handle to object transitions */
+ssize_t glfs_h_extract_handle (struct glfs_object *object,
+			       unsigned char *handle, int len);
+
+struct glfs_object *glfs_h_create_from_handle (struct glfs *fs,
+					       unsigned char *handle, int len,
+					       struct stat *stat);
+
+/* Operations enabling object handles to fd transitions */
+struct glfs_fd *glfs_h_opendir (struct glfs *fs, struct glfs_object *object);
+
+struct glfs_fd *glfs_h_open (struct glfs *fs, struct glfs_object *object,
+			     int flags);
+
+__END_DECLS
+
+#endif /* !_GLFS_HANDLES_H */
\ No newline at end of file
diff --git a/api/src/glfs-internal.h b/api/src/glfs-internal.h
index 30ff599f2..ec1d5579d 100644
--- a/api/src/glfs-internal.h
+++ b/api/src/glfs-internal.h
@@ -14,6 +14,46 @@
 
 #include "xlator.h"
 
+#define GLFS_SYMLINK_MAX_FOLLOW 2048
+
+#define DEFAULT_REVAL_COUNT 1
+
+#define ESTALE_RETRY(ret,errno,reval,loc,label) do {	\
+	if (ret == -1 && errno == ESTALE) {	        \
+		if (reval < DEFAULT_REVAL_COUNT) {	\
+			reval++;			\
+			loc_wipe (loc);			\
+			goto label;			\
+		}					\
+	}						\
+	} while (0)
+
+#define GLFS_LOC_FILL_INODE(oinode, loc, label) do {   \
+	loc.inode = inode_ref (oinode);                \
+	uuid_copy (loc.gfid, oinode->gfid);            \
+	ret = glfs_loc_touchup (&loc);                 \
+	if (ret != 0) {                                \
+		errno = EINVAL;                        \
+		goto label;                            \
+	}                                              \
+	} while (0)
+
+#define GLFS_LOC_FILL_PINODE(pinode, loc, ret, errno, label, path) do {   \
+	loc.inode = inode_new (pinode->table);                            \
+	if (!loc.inode) {                                                 \
+		ret = -1;                                                 \
+		errno = ENOMEM;                                           \
+		goto label;                                               \
+	}                                                                 \
+	loc.parent = inode_ref (pinode);                                  \
+	loc.name = path;                                                  \
+	ret = glfs_loc_touchup (&loc);                                    \
+	if (ret != 0) {                                                   \
+		errno = EINVAL;                                           \
+		goto label;                                               \
+	}                                                                 \
+	} while (0)
+
 struct glfs;
 
 typedef int (*glfs_init_cbk) (struct glfs *fs, int ret);
@@ -55,6 +95,15 @@ struct glfs_fd {
 	fd_t              *fd; /* Currently guared by @fs->mutex. TODO: per-glfd lock */
 	struct list_head   entries;
 	gf_dirent_t       *next;
+	struct dirent     *readdirbuf;
+};
+
+/* glfs object handle introduced for the alternate gfapi implementation based
+   on glfs handles/gfid/inode
+*/
+struct glfs_object {
+        inode_t         *inode;
+        uuid_t          gfid;
 };
 
 #define DEFAULT_EVENT_POOL_SIZE           16384
@@ -133,6 +182,19 @@ inode_t * glfs_refresh_inode (xlator_t *subvol, inode_t *inode);
 
 inode_t *glfs_cwd_get (struct glfs *fs);
 int glfs_cwd_set (struct glfs *fs, inode_t *inode);
+inode_t *glfs_resolve_inode (struct glfs *fs, xlator_t *subvol,
+			     struct glfs_object *object);
+int glfs_create_object (loc_t *loc, struct glfs_object **retobject);
 int __glfs_cwd_set (struct glfs *fs, inode_t *inode);
 
+int glfs_resolve_base (struct glfs *fs, xlator_t *subvol, inode_t *inode,
+		       struct iatt *iatt);
+int glfs_resolve_at (struct glfs *fs, xlator_t *subvol, inode_t *at,
+                     const char *origpath, loc_t *loc, struct iatt *iatt,
+                     int follow, int reval);
+int glfs_loc_touchup (loc_t *loc);
+void glfs_iatt_to_stat (struct glfs *fs, struct iatt *iatt, struct stat *stat);
+int glfs_loc_link (loc_t *loc, struct iatt *iatt);
+int glfs_loc_unlink (loc_t *loc);
+
 #endif /* !_GLFS_INTERNAL_H */
diff --git a/api/src/glfs-master.c b/api/src/glfs-master.c
index 09a92b7f0..c02534c18 100644
--- a/api/src/glfs-master.c
+++ b/api/src/glfs-master.c
@@ -24,6 +24,7 @@
 #include "glusterfs.h"
 
 #include "glfs-internal.h"
+#include "glfs-mem-types.h"
 
 
 int
@@ -114,6 +115,18 @@ notify (xlator_t *this, int event, void *data, ...)
 int
 mem_acct_init (xlator_t *this)
 {
+	int ret = -1;
+
+	if (!this)
+		return ret;
+
+	ret = xlator_mem_acct_init (this, glfs_mt_end + 1);
+	if (ret) {
+		gf_log (this->name, GF_LOG_ERROR, "Failed to initialise "
+                        "memory accounting");
+		return ret;
+	}
+
 	return 0;
 }
 
diff --git a/api/src/glfs-mem-types.h b/api/src/glfs-mem-types.h
index 590acd03f..3301b3da5 100644
--- a/api/src/glfs-mem-types.h
+++ b/api/src/glfs-mem-types.h
@@ -16,14 +16,16 @@
 #define GF_MEM_TYPE_START (gf_common_mt_end + 1)
 
 enum glfs_mem_types_ {
-        glfs_mt_glfs_t,
+        glfs_mt_glfs_t = GF_MEM_TYPE_START,
         glfs_mt_call_pool_t,
         glfs_mt_xlator_t,
 	glfs_mt_glfs_fd_t,
 	glfs_mt_glfs_io_t,
 	glfs_mt_volfile_t,
 	glfs_mt_xlator_cmdline_option_t,
-        glfs_mt_end
+	glfs_mt_glfs_object_t,
+	glfs_mt_readdirbuf_t,
+	glfs_mt_end
 
 };
 #endif
diff --git a/api/src/glfs-mgmt.c b/api/src/glfs-mgmt.c
index a76692bfd..6843e9cb3 100644
--- a/api/src/glfs-mgmt.c
+++ b/api/src/glfs-mgmt.c
@@ -114,7 +114,7 @@ mgmt_cbk_event (struct rpc_clnt *rpc, void *mydata, void *data)
 }
 
 
-rpcclnt_cb_actor_t gluster_cbk_actors[] = {
+rpcclnt_cb_actor_t mgmt_cbk_actors[] = {
 	[GF_CBK_FETCHSPEC] = {"FETCHSPEC", GF_CBK_FETCHSPEC, mgmt_cbk_spec },
 	[GF_CBK_EVENT_NOTIFY] = {"EVENTNOTIFY", GF_CBK_EVENT_NOTIFY,
 				 mgmt_cbk_event},
@@ -125,7 +125,7 @@ struct rpcclnt_cb_program mgmt_cbk_prog = {
 	.progname  = "GlusterFS Callback",
 	.prognum   = GLUSTER_CBK_PROGRAM,
 	.progver   = GLUSTER_CBK_VERSION,
-	.actors	   = gluster_cbk_actors,
+	.actors	   = mgmt_cbk_actors,
 	.numactors = GF_CBK_MAXVALUE,
 };
 
@@ -203,162 +203,6 @@ out:
 
 
 static int
-xlator_equal_rec (xlator_t *xl1, xlator_t *xl2)
-{
-	xlator_list_t *trav1 = NULL;
-	xlator_list_t *trav2 = NULL;
-	int	       ret   = 0;
-
-	if (xl1 == NULL || xl2 == NULL) {
-		gf_log ("xlator", GF_LOG_DEBUG, "invalid argument");
-		return -1;
-	}
-
-	trav1 = xl1->children;
-	trav2 = xl2->children;
-
-	while (trav1 && trav2) {
-		ret = xlator_equal_rec (trav1->xlator, trav2->xlator);
-		if (ret) {
-			gf_log ("glfs-mgmt", GF_LOG_DEBUG,
-				"xlators children not equal");
-			goto out;
-		}
-
-		trav1 = trav1->next;
-		trav2 = trav2->next;
-	}
-
-	if (trav1 || trav2) {
-		ret = -1;
-		goto out;
-	}
-
-	if (strcmp (xl1->name, xl2->name)) {
-		ret = -1;
-		goto out;
-	}
-out :
-	return ret;
-}
-
-
-static gf_boolean_t
-is_graph_topology_equal (glusterfs_graph_t *graph1,
-			 glusterfs_graph_t *graph2)
-{
-	xlator_t    *trav1    = NULL;
-	xlator_t    *trav2    = NULL;
-	gf_boolean_t ret      = _gf_true;
-
-	trav1 = graph1->first;
-	trav2 = graph2->first;
-
-	ret = xlator_equal_rec (trav1, trav2);
-
-	if (ret) {
-		gf_log ("glfs-mgmt", GF_LOG_DEBUG,
-			"graphs are not equal");
-		ret = _gf_false;
-		goto out;
-	}
-
-	ret = _gf_true;
-	gf_log ("glfs-mgmt", GF_LOG_DEBUG,
-		"graphs are equal");
-
-out:
-	return ret;
-}
-
-
-/* Function has 3types of return value 0, -ve , 1
- *   return 0	       =======> reconfiguration of options has succeeded
- *   return 1	       =======> the graph has to be reconstructed and all the xlators should be inited
- *   return -1(or -ve) =======> Some Internal Error occurred during the operation
- */
-static int
-glusterfs_volfile_reconfigure (struct glfs *fs, FILE *newvolfile_fp)
-{
-	glusterfs_graph_t *oldvolfile_graph = NULL;
-	glusterfs_graph_t *newvolfile_graph = NULL;
-	FILE		  *oldvolfile_fp    = NULL;
-	glusterfs_ctx_t	  *ctx		    = NULL;
-
-	int ret = -1;
-
-	oldvolfile_fp = tmpfile ();
-	if (!oldvolfile_fp)
-		goto out;
-
-	if (!fs->oldvollen) {
-		ret = 1; // Has to call INIT for the whole graph
-		goto out;
-	}
-	fwrite (fs->oldvolfile, fs->oldvollen, 1, oldvolfile_fp);
-	fflush (oldvolfile_fp);
-	if (ferror (oldvolfile_fp)) {
-		goto out;
-	}
-
-	oldvolfile_graph = glusterfs_graph_construct (oldvolfile_fp);
-	if (!oldvolfile_graph) {
-		goto out;
-	}
-
-	newvolfile_graph = glusterfs_graph_construct (newvolfile_fp);
-	if (!newvolfile_graph) {
-		goto out;
-	}
-
-	if (!is_graph_topology_equal (oldvolfile_graph,
-				      newvolfile_graph)) {
-
-		ret = 1;
-		gf_log ("glfs-mgmt", GF_LOG_DEBUG,
-			"Graph topology not equal(should call INIT)");
-		goto out;
-	}
-
-	gf_log ("glfs-mgmt", GF_LOG_DEBUG,
-		"Only options have changed in the new "
-		"graph");
-
-	ctx = fs->ctx;
-
-	if (!ctx) {
-		gf_log ("glfs-mgmt", GF_LOG_ERROR,
-			"glusterfs_ctx_get() returned NULL");
-		goto out;
-	}
-
-	oldvolfile_graph = ctx->active;
-
-	if (!oldvolfile_graph) {
-		gf_log ("glfs-mgmt", GF_LOG_ERROR,
-			"glusterfs_ctx->active is NULL");
-		goto out;
-	}
-
-	/* */
-	ret = glusterfs_graph_reconfigure (oldvolfile_graph,
-					   newvolfile_graph);
-	if (ret) {
-		gf_log ("glfs-mgmt", GF_LOG_DEBUG,
-			"Could not reconfigure new options in old graph");
-		goto out;
-	}
-
-	ret = 0;
-out:
-	if (oldvolfile_fp)
-		fclose (oldvolfile_fp);
-
-	return ret;
-}
-
-
-static int
 glusterfs_oldvolfile_update (struct glfs *fs, char *volfile, ssize_t size)
 {
 	int ret = -1;
@@ -416,6 +260,7 @@ mgmt_getspec_cbk (struct rpc_req *req, struct iovec *iov, int count,
 		gf_log (frame->this->name, GF_LOG_ERROR,
 			"failed to get the 'volume file' from server");
 		ret = -1;
+                errno = rsp.op_errno;
 		goto out;
 	}
 
@@ -450,7 +295,8 @@ mgmt_getspec_cbk (struct rpc_req *req, struct iovec *iov, int count,
 	*  return -1(or -ve) =======> Some Internal Error occurred during the operation
 	*/
 
-	ret = glusterfs_volfile_reconfigure (fs, tmpfp);
+	ret = glusterfs_volfile_reconfigure (fs->oldvollen, tmpfp, fs->ctx,
+                                             fs->oldvolfile);
 	if (ret == 0) {
 		gf_log ("glusterfsd-mgmt", GF_LOG_DEBUG,
 			"No need to re-load volfile, reconfigure done");
@@ -481,6 +327,7 @@ out:
         if (ENOTSUP == ret) {
                 gf_log ("mgmt", GF_LOG_ERROR, "Server is operating at an "
                         "op-version which is not supported");
+                errno = ENOTSUP;
                 glfs_init_done (fs, -1);
         }
 
@@ -491,8 +338,11 @@ out:
 		gf_log ("glfs-mgmt", GF_LOG_ERROR,
 			"failed to fetch volume file (key:%s)",
 			ctx->cmd_args.volfile_id);
-		if (!need_retry)
+		if (!need_retry) {
+                        if (!errno)
+                                errno = EINVAL;
 			glfs_init_done (fs, -1);
+                }
 	}
 
 	if (tmpfp)
@@ -583,9 +433,10 @@ mgmt_rpc_notify (struct rpc_clnt *rpc, void *mydata, rpc_clnt_event_t event,
 			gf_log ("glfs-mgmt", GF_LOG_INFO,
 				"%d connect attempts left",
 				cmd_args->max_connect_attempts);
-			if (0 >= cmd_args->max_connect_attempts)
+			if (0 >= cmd_args->max_connect_attempts) {
+                                errno = ENOTCONN;
 				glfs_init_done (fs, -1);
-				break;
+                        }
 		}
 		break;
 	case RPC_CLNT_CONNECT:
@@ -594,10 +445,11 @@ mgmt_rpc_notify (struct rpc_clnt *rpc, void *mydata, rpc_clnt_event_t event,
 		ret = glfs_volfile_fetch (fs);
 		if (ret && ctx && (ctx->active == NULL)) {
 			/* Do it only for the first time */
-			/* Exit the process.. there is some wrong options */
+			/* Exit the process.. there are some wrong options */
 			gf_log ("glfs-mgmt", GF_LOG_ERROR,
 				"failed to fetch volume file (key:%s)",
 				ctx->cmd_args.volfile_id);
+                        errno = EINVAL;
 			glfs_init_done (fs, -1);
 		}
 
diff --git a/api/src/glfs-resolve.c b/api/src/glfs-resolve.c
index 4a634df6e..4ca2eb6fc 100644
--- a/api/src/glfs-resolve.c
+++ b/api/src/glfs-resolve.c
@@ -67,6 +67,7 @@ __glfs_first_lookup (struct glfs *fs, xlator_t *subvol)
 	}
 	pthread_mutex_lock (&fs->mutex);
 	fs->migration_in_progress = 0;
+	pthread_cond_broadcast (&fs->cond);
 
 	return ret;
 }
@@ -128,6 +129,7 @@ __glfs_refresh_inode (struct glfs *fs, xlator_t *subvol, inode_t *inode)
 	}
 	pthread_mutex_lock (&fs->mutex);
 	fs->migration_in_progress = 0;
+	pthread_cond_broadcast (&fs->cond);
 
 	return newinode;
 }
@@ -191,7 +193,7 @@ out:
 }
 
 
-void
+int
 glfs_resolve_base (struct glfs *fs, xlator_t *subvol, inode_t *inode,
 		   struct iatt *iatt)
 {
@@ -210,6 +212,8 @@ glfs_resolve_base (struct glfs *fs, xlator_t *subvol, inode_t *inode,
 	ret = syncop_lookup (subvol, &loc, NULL, iatt, NULL, NULL);
 out:
 	loc_wipe (&loc);
+
+	return ret;
 }
 
 
@@ -337,7 +341,8 @@ glfs_resolve_at (struct glfs *fs, xlator_t *subvol, inode_t *at,
 	} else {
 		inode = inode_ref (subvol->itable->root);
 
-		glfs_resolve_base (fs, subvol, inode, &ciatt);
+		if (strcmp (path, "/") == 0)
+			glfs_resolve_base (fs, subvol, inode, &ciatt);
 	}
 
 	for (component = strtok_r (path, "/", &saveptr);
@@ -356,7 +361,8 @@ glfs_resolve_at (struct glfs *fs, xlator_t *subvol, inode_t *at,
 						   component, as the caller
 						   wants proper iatt filled
 						*/
-						(reval || !next_component));
+						(reval || (!next_component &&
+						iatt)));
 		if (!inode)
 			break;
 
@@ -368,6 +374,16 @@ glfs_resolve_at (struct glfs *fs, xlator_t *subvol, inode_t *at,
 			char *lpath = NULL;
 			loc_t sym_loc = {0,};
 
+			if (follow > GLFS_SYMLINK_MAX_FOLLOW) {
+				errno = ELOOP;
+				ret = -1;
+				if (inode) {
+					inode_unref (inode);
+					inode = NULL;
+				}
+				break;
+			}
+
 			ret = glfs_resolve_symlink (fs, subvol, inode, &lpath);
 			inode_unref (inode);
 			inode = NULL;
@@ -383,7 +399,7 @@ glfs_resolve_at (struct glfs *fs, xlator_t *subvol, inode_t *at,
 					       /* always recurisvely follow while
 						  following symlink
 					       */
-					       1, reval);
+					       follow + 1, reval);
 			if (ret == 0)
 				inode = inode_ref (sym_loc.inode);
 			loc_wipe (&sym_loc);
@@ -583,6 +599,15 @@ glfs_migrate_fd_safe (struct glfs *fs, xlator_t *newsubvol, fd_t *oldfd)
 
 	loc.inode = inode_ref (newinode);
 
+        ret = inode_path (oldfd->inode, NULL, (char **)&loc.path);
+        if (ret < 0) {
+                gf_log (fs->volname, GF_LOG_INFO, "inode_path failed");
+                goto out;
+        }
+
+        uuid_copy (loc.gfid, oldinode->gfid);
+
+
 	if (IA_ISDIR (oldinode->ia_type))
 		ret = syncop_opendir (newsubvol, &loc, newfd);
 	else
@@ -613,6 +638,7 @@ glfs_migrate_fd_safe (struct glfs *fs, xlator_t *newsubvol, fd_t *oldfd)
 		goto out;
 	}
 
+        newfd->flags = oldfd->flags;
 	fd_bind (newfd);
 out:
 	if (newinode)
@@ -642,6 +668,7 @@ __glfs_migrate_fd (struct glfs *fs, xlator_t *newsubvol, struct glfs_fd *glfd)
 	}
 	pthread_mutex_lock (&fs->mutex);
 	fs->migration_in_progress = 0;
+	pthread_cond_broadcast (&fs->cond);
 
 	return newfd;
 }
@@ -797,6 +824,9 @@ glfs_subvol_done (struct glfs *fs, xlator_t *subvol)
 	int ref = 0;
 	xlator_t *active_subvol = NULL;
 
+	if (!subvol)
+		return;
+
 	glfs_lock (fs);
 	{
 		ref = (--subvol->winds);
@@ -877,3 +907,63 @@ glfs_cwd_get (struct glfs *fs)
 
 	return cwd;
 }
+
+inode_t *
+__glfs_resolve_inode (struct glfs *fs, xlator_t *subvol,
+		    struct glfs_object *object)
+{
+	inode_t *inode = NULL;
+
+	if (object->inode->table->xl == subvol)
+		return inode_ref (object->inode);
+
+	inode = __glfs_refresh_inode (fs, fs->active_subvol,
+					object->inode);
+	if (!inode)
+		return NULL;
+
+	if (subvol == fs->active_subvol) {
+		inode_unref (object->inode);
+		object->inode = inode_ref (inode);
+	}
+
+	return inode;
+}
+
+inode_t *
+glfs_resolve_inode (struct glfs *fs, xlator_t *subvol,
+		    struct glfs_object *object)
+{
+	inode_t *inode = NULL;
+
+	glfs_lock (fs);
+	{
+		inode = __glfs_resolve_inode(fs, subvol, object);
+	}
+	glfs_unlock (fs);
+
+	return inode;
+}
+
+int
+glfs_create_object (loc_t *loc, struct glfs_object **retobject)
+{
+	struct glfs_object *object = NULL;
+
+	object = GF_CALLOC (1, sizeof(struct glfs_object),
+			    glfs_mt_glfs_object_t);
+	if (object == NULL) {
+		errno = ENOMEM;
+		return -1;
+	}
+
+	object->inode = loc->inode;
+	uuid_copy (object->gfid, object->inode->gfid);
+
+	/* we hold the reference */
+	loc->inode = NULL;
+
+	*retobject = object;
+
+	return 0;
+}
diff --git a/api/src/glfs.c b/api/src/glfs.c
index efda6b67e..29ed47c0c 100644
--- a/api/src/glfs.c
+++ b/api/src/glfs.c
@@ -20,7 +20,6 @@
   - handle SEEK_END failure in _lseek()
   - handle umask (per filesystem?)
   - make itables LRU based
-  - implement glfs_fini()
   - 0-copy for readv/writev
   - reconcile the open/creat mess
 */
@@ -51,6 +50,7 @@
 #include "glfs.h"
 #include "glfs-internal.h"
 #include "hashfn.h"
+#include "rpc-clnt.h"
 
 
 static gf_boolean_t
@@ -66,7 +66,7 @@ glusterfs_ctx_defaults_init (glusterfs_ctx_t *ctx)
 	call_pool_t   *pool = NULL;
 	int	       ret = -1;
 
-	xlator_mem_acct_init (THIS, glfs_mt_end);
+	xlator_mem_acct_init (THIS, glfs_mt_end + 1);
 
 	ctx->process_uuid = generate_glusterfs_ctx_id ();
 	if (!ctx->process_uuid) {
@@ -85,7 +85,7 @@ glusterfs_ctx_defaults_init (glusterfs_ctx_t *ctx)
 		goto err;
 	}
 
-	ctx->env = syncenv_new (0);
+	ctx->env = syncenv_new (0, 0, 0);
 	if (!ctx->env) {
 		goto err;
 	}
@@ -317,6 +317,20 @@ enomem:
 	return -1;
 }
 
+int glfs_setfsuid (uid_t fsuid)
+{
+	return syncopctx_setfsuid (&fsuid);
+}
+
+int glfs_setfsgid (gid_t fsgid)
+{
+	return syncopctx_setfsgid (&fsgid);
+}
+
+int glfs_setfsgroups (size_t size, const gid_t *list)
+{
+	return syncopctx_setfsgroups(size, list);
+}
 
 struct glfs *
 glfs_from_glfd (struct glfs_fd *glfd)
@@ -370,6 +384,9 @@ glfs_fd_destroy (struct glfs_fd *glfd)
 
 	if (glfd->fd)
 		fd_unref (glfd->fd);
+
+	GF_FREE (glfd->readdirbuf);
+
 	GF_FREE (glfd);
 }
 
@@ -475,16 +492,25 @@ int
 glfs_set_logging (struct glfs *fs, const char *logfile, int loglevel)
 {
 	int  ret = 0;
+        char *tmplog = NULL;
 
-	if (logfile) {
-		ret = gf_log_init (fs->ctx, logfile);
-		if (ret)
-			return ret;
-	}
+        if (!logfile) {
+                ret = gf_set_log_file_path (&fs->ctx->cmd_args);
+                if (ret)
+                        goto out;
+                tmplog = fs->ctx->cmd_args.log_file;
+        } else {
+                tmplog = (char *)logfile;
+        }
+
+        ret = gf_log_init (fs->ctx, tmplog, NULL);
+        if (ret)
+                goto out;
 
 	if (loglevel >= 0)
 		gf_log_set_loglevel (loglevel);
 
+out:
 	return ret;
 }
 
@@ -550,7 +576,7 @@ glfs_init_common (struct glfs *fs)
 	if (ret)
 		return ret;
 
-	ret = pthread_create (&fs->poller, NULL, glfs_poller, fs);
+	ret = gf_thread_create (&fs->poller, NULL, glfs_poller, fs);
 	if (ret)
 		return ret;
 
@@ -594,7 +620,54 @@ glfs_init (struct glfs *fs)
 int
 glfs_fini (struct glfs *fs)
 {
-	int  ret = -1;
-
-	return ret;
+        int             ret = -1;
+        int             countdown = 100;
+        xlator_t        *subvol = NULL;
+        glusterfs_ctx_t *ctx = NULL;
+        call_pool_t     *call_pool = NULL;
+
+        ctx = fs->ctx;
+
+        if (ctx->mgmt) {
+                rpc_clnt_disable (ctx->mgmt);
+                ctx->mgmt = NULL;
+        }
+
+        __glfs_entry_fs (fs);
+
+        call_pool = fs->ctx->pool;
+
+        while (countdown--) {
+                /* give some time for background frames to finish */
+                if (!call_pool->cnt)
+                        break;
+                usleep (100000);
+        }
+        /* leaked frames may exist, we ignore */
+
+        /*We deem glfs_fini as successful if there are no pending frames in the call
+         *pool*/
+        ret = (call_pool->cnt == 0)? 0: -1;
+
+        subvol = glfs_active_subvol (fs);
+        if (subvol) {
+                /* PARENT_DOWN within glfs_subvol_done() is issued only
+                   on graph switch (new graph should activiate and
+                   decrement the extra @winds count taken in glfs_graph_setup()
+
+                   Since we are explicitly destroying, PARENT_DOWN is necessary
+                */
+                xlator_notify (subvol, GF_EVENT_PARENT_DOWN, subvol, 0);
+                /* TBD: wait for CHILD_DOWN before exiting, in case of
+                   asynchronous cleanup like graceful socket disconnection
+                   in the future.
+                */
+        }
+
+        glfs_subvol_done (fs, subvol);
+
+        if (ctx->log.logfile)
+                fclose (ctx->log.logfile);
+
+        return ret;
 }
diff --git a/api/src/glfs.h b/api/src/glfs.h
index d9c1202cd..18fda496e 100644
--- a/api/src/glfs.h
+++ b/api/src/glfs.h
@@ -176,7 +176,9 @@ int glfs_set_volfile_server (glfs_t *fs, const char *transport,
   @fs: The 'virtual mount' object to be configured with the logging parameters.
 
   @logfile: The logfile to be used for logging. Will be created if it does not
-            already exist (provided system permissions allow.)
+            already exist (provided system permissions allow). If NULL, a new
+            logfile will be created in default log directory associated with
+            the glusterfs installation.
 
   @loglevel: Numerical value specifying the degree of verbosity. Higher the
              value, more verbose the logging.
@@ -218,6 +220,36 @@ int glfs_set_logging (glfs_t *fs, const char *logfile, int loglevel);
 int glfs_init (glfs_t *fs);
 
 
+/*
+  SYNOPSIS
+
+  glfs_fini: Cleanup and destroy the 'virtual mount'
+
+  DESCRIPTION
+
+  This function attempts to gracefully destroy glfs_t object. An attempt is
+  made to wait for all background processing to complete before returning.
+
+  glfs_fini() must be called after all operations on glfs_t is finished.
+
+  IMPORTANT
+
+  IT IS NECESSARY TO CALL glfs_fini() ON ALL THE INITIALIZED glfs_t
+  OBJECTS BEFORE TERMINATING THE PROGRAM. THERE MAY BE CACHED AND
+  UNWRITTEN / INCOMPLETE OPERATIONS STILL IN PROGRESS EVEN THOUGH THE
+  API CALLS HAVE RETURNED. glfs_fini() WILL WAIT FOR BACKGROUND OPERATIONS
+  TO COMPLETE BEFORE RETURNING, THEREBY MAKING IT SAFE FOR THE PROGRAM TO
+  EXIT.
+
+  PARAMETERS
+
+  @fs: The 'virtual mount' object to be destroyed.
+
+  RETURN VALUES
+
+   0 : Success.
+*/
+
 int glfs_fini (glfs_t *fs);
 
 /*
@@ -239,6 +271,32 @@ int glfs_fini (glfs_t *fs);
 struct glfs_fd;
 typedef struct glfs_fd glfs_fd_t;
 
+/*
+ * PER THREAD IDENTITY MODIFIERS
+ *
+ * The following operations enable to set a per thread identity context
+ * for the glfs APIs to perform operations as. The calls here are kept as close
+ * to POSIX equivalents as possible.
+ *
+ * NOTES:
+ *
+ *  - setgroups is a per thread setting, hence this is named as fsgroups to be
+ *    close in naming to the fs(u/g)id APIs
+ *  - Typical mode of operation is to set the IDs as required, with the
+ *    supplementary groups being optionally set, make the glfs call and post the
+ *    glfs operation set them back to eu/gid or uid/gid as appropriate to the
+ *    caller
+ *  - The groups once set, need to be unset by setting the size to 0 (in which
+ *    case the list argument is a do not care)
+ *  - Once a process for a thread of operation choses to set the IDs, all glfs
+ *    calls made from that thread would default to the IDs set for the thread.
+ *    As a result use these APIs with care and ensure that the set IDs are
+ *    reverted to global process defaults as required.
+ *
+ */
+int glfs_setfsuid (uid_t fsuid);
+int glfs_setfsgid (gid_t fsgid);
+int glfs_setfsgroups (size_t size, const gid_t *list);
 
 /*
   SYNOPSIS
@@ -304,6 +362,29 @@ glfs_t *glfs_from_glfd (glfs_fd_t *fd);
 int glfs_set_xlator_option (glfs_t *fs, const char *xlator, const char *key,
 			    const char *value);
 
+/*
+
+  glfs_io_cbk
+
+  The following is the function type definition of the callback
+  function pointer which has to be provided by the caller to the
+  *_async() versions of the IO calls.
+
+  The callback function is called on completion of the requested
+  IO, and the appropriate return value is returned in @ret.
+
+  In case of an error in completing the IO, @ret will be -1 and
+  @errno will be set with the appropriate error.
+
+  @ret will be same as the return value of the non _async() variant
+  of the particular call
+
+  @data is the same context pointer provided by the caller at the
+  time of issuing the async IO call. This can be used by the
+  caller to differentiate different instances of the async requests
+  in a common callback function.
+*/
+
 typedef void (*glfs_io_cbk) (glfs_fd_t *fd, ssize_t ret, void *data);
 
 // glfs_{read,write}[_async]
@@ -387,12 +468,31 @@ int glfs_link (glfs_t *fs, const char *oldpath, const char *newpath);
 
 glfs_fd_t *glfs_opendir (glfs_t *fs, const char *path);
 
+/*
+ * @glfs_readdir_r and @glfs_readdirplus_r ARE thread safe AND re-entrant,
+ * but the interface has ambiguity about the size of @dirent to be allocated
+ * before calling the APIs. 512 byte buffer (for @dirent) is sufficient for
+ * all known systems which are tested againt glusterfs/gfapi, but may be
+ * insufficient in the future.
+ */
+
 int glfs_readdir_r (glfs_fd_t *fd, struct dirent *dirent,
 		    struct dirent **result);
 
 int glfs_readdirplus_r (glfs_fd_t *fd, struct stat *stat, struct dirent *dirent,
 			struct dirent **result);
 
+/*
+ * @glfs_readdir and @glfs_readdirplus are NEITHER thread safe NOR re-entrant
+ * when called on the same directory handle. However they ARE thread safe
+ * AND re-entrant when called on different directory handles (which may be
+ * referring to the same directory too.)
+ */
+
+struct dirent *glfs_readdir (glfs_fd_t *fd);
+
+struct dirent *glfs_readdirplus (glfs_fd_t *fd, struct stat *stat);
+
 long glfs_telldir (glfs_fd_t *fd);
 
 void glfs_seekdir (glfs_fd_t *fd, long offset);
@@ -452,9 +552,15 @@ int glfs_fallocate(glfs_fd_t *fd, int keep_size, off_t offset, size_t len);
 
 int glfs_discard(glfs_fd_t *fd, off_t offset, size_t len);
 
+
 int glfs_discard_async (glfs_fd_t *fd, off_t length, size_t lent,
 			glfs_io_cbk fn, void *data);
 
+int glfs_zerofill(glfs_fd_t *fd, off_t offset, size_t len);
+
+int glfs_zerofill_async (glfs_fd_t *fd, off_t length, size_t len,
+                        glfs_io_cbk fn, void *data);
+
 char *glfs_getcwd (glfs_t *fs, char *buf, size_t size);
 
 int glfs_chdir (glfs_t *fs, const char *path);
diff --git a/cli/src/Makefile.am b/cli/src/Makefile.am
index ae5231026..216d1bb55 100644
--- a/cli/src/Makefile.am
+++ b/cli/src/Makefile.am
@@ -2,10 +2,7 @@ sbin_PROGRAMS = gluster
 
 gluster_SOURCES = cli.c registry.c input.c cli-cmd.c cli-rl.c \
 	 cli-cmd-volume.c cli-cmd-peer.c cli-rpc-ops.c cli-cmd-parser.c\
-	 cli-cmd-system.c cli-cmd-misc.c cli-xml-output.c
-if ENABLE_BD_XLATOR
-gluster_SOURCES += cli-cmd-volume-bdevice.c
-endif
+	 cli-cmd-system.c cli-cmd-misc.c cli-xml-output.c cli-cmd-snapshot.c
 
 gluster_LDADD = $(top_builddir)/libglusterfs/src/libglusterfs.la $(GF_LDADD)\
 		$(RLLIBS) $(top_builddir)/rpc/xdr/src/libgfxdr.la \
diff --git a/cli/src/cli-cmd-misc.c b/cli/src/cli-cmd-misc.c
index 5f3a77ca6..566d7c978 100644
--- a/cli/src/cli-cmd-misc.c
+++ b/cli/src/cli-cmd-misc.c
@@ -32,6 +32,7 @@ extern struct cli_cmd cli_probe_cmds[];
 extern struct cli_cmd cli_log_cmds[];
 extern struct cli_cmd cli_system_cmds[];
 extern struct cli_cmd cli_bd_cmds[];
+extern struct cli_cmd snapshot_cmds[];
 struct cli_cmd cli_misc_cmds[];
 
 int
@@ -46,10 +47,7 @@ cli_cmd_display_help (struct cli_state *state, struct cli_cmd_word *in_word,
                       const char **words, int wordcount)
 {
         struct cli_cmd        *cmd[] = {volume_cmds, cli_probe_cmds,
-                                       cli_misc_cmds,
-#ifdef HAVE_BD_XLATOR
-                                       cli_bd_cmds,
-#endif
+                                       cli_misc_cmds, snapshot_cmds,
                                        NULL};
         struct cli_cmd        *cmd_ind = NULL;
         int                   i = 0;
diff --git a/cli/src/cli-cmd-parser.c b/cli/src/cli-cmd-parser.c
index 484c4a34b..5ab208b8f 100644
--- a/cli/src/cli-cmd-parser.c
+++ b/cli/src/cli-cmd-parser.c
@@ -1,5 +1,5 @@
 /*
-   Copyright (c) 2010-2012 Red Hat, Inc. <http://www.redhat.com>
+   Copyright (c) 2010-2013 Red Hat, Inc. <http://www.redhat.com>
    This file is part of GlusterFS.
 
    This file is licensed to you under your choice of the GNU Lesser
@@ -26,6 +26,38 @@
 #include "protocol-common.h"
 #include "cli1-xdr.h"
 
+#define MAX_SNAP_DESCRIPTION_LEN 1024
+
+struct snap_config_opt_vals_ snap_confopt_vals[] = {
+        {.op_name        = "snap-max-hard-limit",
+         .question       = "Changing snapshot-max-hard-limit "
+                           "will lead to deletion of snapshots "
+                           "if they exceed the new limit.\n"
+                           "Do you want to continue?"
+        },
+        {.op_name        = "snap-max-soft-limit",
+         .question       = "Changing snapshot-max-soft-limit "
+                           "will lead to deletion of snapshots "
+                           "if they exceed the new limit.\n"
+                           "Do you want to continue?"
+        },
+        {.op_name        = "both",
+        .question        = "Changing snapshot-max-hard-limit & "
+                           "snapshot-max-soft-limit will lead to "
+                           "deletion of snapshots if they exceed "
+                           "the new limit.\nDo you want to continue?"
+        },
+        {.op_name        = NULL,
+        }
+};
+
+enum cli_snap_config_set_types {
+        GF_SNAP_CONFIG_SET_HARD = 0,
+        GF_SNAP_CONFIG_SET_SOFT = 1,
+        GF_SNAP_CONFIG_SET_BOTH = 2,
+};
+typedef enum cli_snap_config_set_types cli_snap_config_set_types;
+
 static const char *
 id_sel (void *wcon)
 {
@@ -159,22 +191,17 @@ cli_cmd_volume_create_parse (const char **words, int wordcount, dict_t **options
         int32_t index = 0;
         char    *bricks = NULL;
         int32_t brick_count = 0;
-        char    *opwords[] = { "replica", "stripe", "transport",
-#ifdef HAVE_BD_XLATOR
-                                "device",
-#endif
-                                NULL };
+        char    *opwords[] = { "replica", "stripe", "transport", NULL };
 
         char    *invalid_volnames[] = {"volume", "type", "subvolumes", "option",
                                        "end-volume", "all", "volume_not_in_ring",
-                                       NULL};
+                                       "description", "force",
+                                       "snap-max-hard-limit",
+                                       "snap-max-soft-limit", NULL};
         char    *w = NULL;
         int      op_count = 0;
         int32_t  replica_count = 1;
         int32_t  stripe_count = 1;
-#ifdef HAVE_BD_XLATOR
-        char     *dev_type = NULL;
-#endif
         gf_boolean_t is_force = _gf_false;
         int wc = wordcount;
 
@@ -311,26 +338,7 @@ cli_cmd_volume_create_parse (const char **words, int wordcount, dict_t **options
                                 goto out;
                         }
                         index += 2;
-                }
-#ifdef HAVE_BD_XLATOR
-                else if ((strcmp (w, "device")) == 0) {
-                        if (dev_type) {
-                                cli_err ("'device' option given more"
-                                         " than one time");
-                                goto out;
-                        }
-                        if ((strcasecmp (words[index+1], "vg") == 0)) {
-                                dev_type = gf_strdup ("vg");
-                        } else {
-                                gf_log ("", GF_LOG_ERROR, "incorrect"
-                                        " device type specified");
-                                ret = -1;
-                                goto out;
-                        }
-                        index += 2;
-                }
-#endif
-                else {
+                }              else {
                         GF_ASSERT (!"opword mismatch");
                         ret = -1;
                         goto out;
@@ -371,19 +379,6 @@ cli_cmd_volume_create_parse (const char **words, int wordcount, dict_t **options
                 goto out;
         }
 
-        /* BD xlator does not support multiple bricks */
-#ifdef HAVE_BD_XLATOR
-        if (brick_count > 1 && dev_type) {
-                cli_err ("Block Device backend volume does not support multiple"
-                         " bricks");
-                gf_log ("", GF_LOG_ERROR,
-                         "Block Device backend volume does not support multiple"
-                        " bricks");
-                ret = -1;
-                goto out;
-        }
-#endif
-
         if (brick_count % sub_count) {
                 if (type == GF_CLUSTER_TYPE_STRIPE)
                         cli_err ("number of bricks is not a multiple of "
@@ -417,14 +412,6 @@ cli_cmd_volume_create_parse (const char **words, int wordcount, dict_t **options
         if (ret)
                 goto out;
 
-#ifdef HAVE_BD_XLATOR
-        if (dev_type) {
-                ret = dict_set_dynstr (dict, "device", dev_type);
-                if (ret)
-                        goto out;
-        }
-#endif
-
         ret = dict_set_int32 (dict, "count", brick_count);
         if (ret)
                 goto out;
@@ -712,7 +699,7 @@ cli_is_key_spl (char *key)
 
 #define GLUSTERD_DEFAULT_WORKDIR "/var/lib/glusterd"
 static int
-cli_add_key_group (dict_t *dict, char *key, char *value)
+cli_add_key_group (dict_t *dict, char *key, char *value, char **op_errstr)
 {
         int             ret = -1;
         int             opt_count = 0;
@@ -726,6 +713,7 @@ cli_add_key_group (dict_t *dict, char *key, char *value)
         char            *tagpath = NULL;
         char            *buf = NULL;
         char            line[PATH_MAX + 256] = {0,};
+        char            errstr[2048] = "";
         FILE            *fp = NULL;
 
         ret = gf_asprintf (&tagpath, "%s/groups/%s",
@@ -738,6 +726,10 @@ cli_add_key_group (dict_t *dict, char *key, char *value)
         fp = fopen (tagpath, "r");
         if (!fp) {
                 ret = -1;
+                snprintf(errstr, sizeof(errstr), "Unable to open file '%s'."
+                         " Error: %s", tagpath, strerror (errno));
+                if (op_errstr)
+                        *op_errstr = gf_strdup(errstr);
                 goto out;
         }
 
@@ -750,6 +742,10 @@ cli_add_key_group (dict_t *dict, char *key, char *value)
                 tok_val = strtok_r (NULL, "=", &saveptr);
                 if (!tok_key || !tok_val) {
                         ret = -1;
+                        snprintf(errstr, sizeof(errstr), "'%s' file format "
+                                 "not valid.", tagpath);
+                        if (op_errstr)
+                                *op_errstr = gf_strdup(errstr);
                         goto out;
                 }
 
@@ -771,6 +767,10 @@ cli_add_key_group (dict_t *dict, char *key, char *value)
 
         if (!opt_count) {
                 ret = -1;
+                snprintf(errstr, sizeof(errstr), "'%s' file format "
+                         "not valid.", tagpath);
+                if (op_errstr)
+                        *op_errstr = gf_strdup(errstr);
                 goto out;
         }
         ret = dict_set_int32 (dict, "count", opt_count);
@@ -791,7 +791,8 @@ out:
 #undef GLUSTERD_DEFAULT_WORKDIR
 
 int32_t
-cli_cmd_volume_set_parse (const char **words, int wordcount, dict_t **options)
+cli_cmd_volume_set_parse (const char **words, int wordcount, dict_t **options,
+                          char **op_errstr)
 {
         dict_t                  *dict = NULL;
         char                    *volname = NULL;
@@ -849,7 +850,7 @@ cli_cmd_volume_set_parse (const char **words, int wordcount, dict_t **options)
                         goto out;
                 }
 
-                ret = cli_add_key_group (dict, key, value);
+                ret = cli_add_key_group (dict, key, value, op_errstr);
                 if (ret == 0)
                         *options = dict;
                 goto out;
@@ -1595,22 +1596,161 @@ gsyncd_glob_check (const char *w)
         return !!strpbrk (w, "*?[");
 }
 
+static int
+config_parse (const char **words, int wordcount, dict_t *dict,
+              unsigned cmdi, unsigned glob)
+{
+        int32_t            ret     = -1;
+        int32_t            i       = -1;
+        char               *append_str = NULL;
+        size_t             append_len = 0;
+        char               *subop = NULL;
+
+        switch ((wordcount - 1) - cmdi) {
+        case 0:
+                subop = gf_strdup ("get-all");
+                break;
+        case 1:
+                if (words[cmdi + 1][0] == '!') {
+                        (words[cmdi + 1])++;
+                        if (gf_asprintf (&subop, "del%s",
+                                         glob ? "-glob" : "") == -1)
+                                subop = NULL;
+                } else
+                        subop = gf_strdup ("get");
+
+                ret = dict_set_str (dict, "op_name", ((char *)words[cmdi + 1]));
+                if (ret < 0)
+                        goto out;
+                break;
+        default:
+                if (gf_asprintf (&subop, "set%s", glob ? "-glob" : "") == -1)
+                        subop = NULL;
+
+                ret = dict_set_str (dict, "op_name", ((char *)words[cmdi + 1]));
+                if (ret < 0)
+                        goto out;
+
+                /* join the varargs by spaces to get the op_value */
+
+                for (i = cmdi + 2; i < wordcount; i++)
+                        append_len += (strlen (words[i]) + 1);
+                /* trailing strcat will add two bytes, make space for that */
+                append_len++;
+
+                append_str = GF_CALLOC (1, append_len, cli_mt_append_str);
+                if (!append_str) {
+                        ret = -1;
+                        goto out;
+                }
+
+                for (i = cmdi + 2; i < wordcount; i++) {
+                        strcat (append_str, words[i]);
+                        strcat (append_str, " ");
+                }
+                append_str[append_len - 2] = '\0';
+                /* "checkpoint now" is special: we resolve that "now" */
+                if (strcmp (words[cmdi + 1], "checkpoint") == 0 &&
+                    strcmp (append_str, "now") == 0) {
+                        struct timeval tv = {0,};
+
+                        ret = gettimeofday (&tv, NULL);
+                        if (ret == -1)
+                                goto out; /* FIXME: free append_str? */
+
+                        GF_FREE (append_str);
+                        append_str = GF_CALLOC (1, 300, cli_mt_append_str);
+                        if (!append_str) {
+                                ret = -1;
+                                goto out;
+                        }
+                        strcpy (append_str, "as of ");
+                        gf_time_fmt (append_str + strlen ("as of "),
+                                     300 - strlen ("as of "),
+                                     tv.tv_sec, gf_timefmt_FT);
+                }
+
+                ret = dict_set_dynstr (dict, "op_value", append_str);
+        }
+
+        ret = -1;
+        if (subop) {
+                ret = dict_set_dynstr (dict, "subop", subop);
+                if (!ret)
+                      subop = NULL;
+        }
+
+out:
+        if (ret && append_str)
+                GF_FREE (append_str);
+
+        GF_FREE (subop);
+
+        gf_log ("cli", GF_LOG_DEBUG, "Returning %d", ret);
+        return ret;
+}
+
+static int32_t
+force_push_pem_parse (const char **words, int wordcount,
+                      dict_t *dict, unsigned *cmdi)
+{
+        int32_t            ret     = 0;
+
+        if (!strcmp ((char *)words[wordcount-1], "force")) {
+                if ((strcmp ((char *)words[wordcount-2], "start")) &&
+                    (strcmp ((char *)words[wordcount-2], "stop")) &&
+                    (strcmp ((char *)words[wordcount-2], "create")) &&
+                    (strcmp ((char *)words[wordcount-2], "push-pem"))) {
+                        ret = -1;
+                        goto out;
+                }
+                ret = dict_set_uint32 (dict, "force",
+                                       _gf_true);
+                if (ret)
+                        goto out;
+                (*cmdi)++;
+
+                if (!strcmp ((char *)words[wordcount-2], "push-pem")) {
+                        if (strcmp ((char *)words[wordcount-3], "create")) {
+                                ret = -1;
+                                goto out;
+                        }
+                        ret = dict_set_int32 (dict, "push_pem", 1);
+                        if (ret)
+                                goto out;
+                        (*cmdi)++;
+                }
+        } else if (!strcmp ((char *)words[wordcount-1], "push-pem")) {
+                if (strcmp ((char *)words[wordcount-2], "create")) {
+                        ret = -1;
+                        goto out;
+                }
+                ret = dict_set_int32 (dict, "push_pem", 1);
+                if (ret)
+                        goto out;
+                (*cmdi)++;
+        }
+
+out:
+        gf_log ("cli", GF_LOG_DEBUG, "Returning %d", ret);
+        return ret;
+}
+
+
 int32_t
 cli_cmd_gsync_set_parse (const char **words, int wordcount, dict_t **options)
 {
         int32_t            ret     = -1;
         dict_t             *dict   = NULL;
         gf1_cli_gsync_set  type    = GF_GSYNC_OPTION_TYPE_NONE;
-        char               *append_str = NULL;
-        size_t             append_len = 0;
-        char               *subop = NULL;
         int                i       = 0;
         unsigned           masteri = 0;
         unsigned           slavei  = 0;
         unsigned           glob    = 0;
         unsigned           cmdi    = 0;
-        char               *opwords[] = { "status", "start", "stop", "config",
-                                          "log-rotate", NULL };
+        char               *opwords[] = { "create", "status", "start", "stop",
+                                          "config", "force", "delete",
+                                          "push-pem", "detail", NULL };
         char               *w = NULL;
 
         GF_ASSERT (words);
@@ -1622,10 +1762,11 @@ cli_cmd_gsync_set_parse (const char **words, int wordcount, dict_t **options)
 
         /* new syntax:
          *
-         * volume geo-replication [$m [$s]] status
+         * volume geo-replication $m $s create [push-pem] [force]
+         * volume geo-replication [$m [$s]] status [detail]
          * volume geo-replication [$m] $s config [[!]$opt [$val]]
-         * volume geo-replication $m $s start|stop
-         * volume geo-replication $m [$s] log-rotate
+         * volume geo-replication $m $s start|stop [force]
+         * volume geo-replication $m $s delete
          */
 
         if (wordcount < 3)
@@ -1656,6 +1797,13 @@ cli_cmd_gsync_set_parse (const char **words, int wordcount, dict_t **options)
                 if (slavei == 3)
                         masteri = 2;
         } else if (i <= 3) {
+                if (!strcmp ((char *)words[wordcount-1], "detail")) {
+                        /* For status detail it is mandatory to provide
+                         * both master and slave */
+                        ret = -1;
+                        goto out;
+                }
+
                 /* no $s, can only be status cmd
                  * (with either a single $m before it or nothing)
                  * -- these conditions imply that i <= 3 after
@@ -1682,7 +1830,12 @@ cli_cmd_gsync_set_parse (const char **words, int wordcount, dict_t **options)
         if (!w)
                 goto out;
 
-        if (strcmp (w, "status") == 0) {
+        if (strcmp (w, "create") == 0) {
+                type = GF_GSYNC_OPTION_TYPE_CREATE;
+
+                if (!masteri || !slavei)
+                        goto out;
+        } else if (strcmp (w, "status") == 0) {
                 type = GF_GSYNC_OPTION_TYPE_STATUS;
 
                 if (slavei && !masteri)
@@ -1702,14 +1855,33 @@ cli_cmd_gsync_set_parse (const char **words, int wordcount, dict_t **options)
 
                 if (!masteri || !slavei)
                         goto out;
-        } else if (strcmp(w, "log-rotate") == 0) {
-                type = GF_GSYNC_OPTION_TYPE_ROTATE;
+        } else if (strcmp (w, "delete") == 0) {
+                type = GF_GSYNC_OPTION_TYPE_DELETE;
 
-                if (slavei && !masteri)
+                if (!masteri || !slavei)
                         goto out;
         } else
                 GF_ASSERT (!"opword mismatch");
 
+        ret = force_push_pem_parse (words, wordcount, dict, &cmdi);
+        if (ret)
+                goto out;
+
+        if (!strcmp ((char *)words[wordcount-1], "detail")) {
+                if (strcmp ((char *)words[wordcount-2], "status")) {
+                        ret = -1;
+                        goto out;
+                }
+                if (!slavei || !masteri) {
+                        ret = -1;
+                        goto out;
+                }
+                ret = dict_set_uint32 (dict, "status-detail", _gf_true);
+                if (ret)
+                        goto out;
+                cmdi++;
+        }
+
         if (type != GF_GSYNC_OPTION_TYPE_CONFIG &&
             (cmdi < wordcount - 1 || glob))
                 goto out;
@@ -1718,97 +1890,26 @@ cli_cmd_gsync_set_parse (const char **words, int wordcount, dict_t **options)
 
         ret = 0;
 
-        if (masteri)
+        if (masteri) {
                 ret = dict_set_str (dict, "master", (char *)words[masteri]);
+                if (!ret)
+                        ret = dict_set_str (dict, "volname",
+                                            (char *)words[masteri]);
+        }
         if (!ret && slavei)
                 ret = dict_set_str (dict, "slave", (char *)words[slavei]);
         if (!ret)
                 ret = dict_set_int32 (dict, "type", type);
-        if (!ret && type == GF_GSYNC_OPTION_TYPE_CONFIG) {
-                switch ((wordcount - 1) - cmdi) {
-                case 0:
-                        subop = gf_strdup ("get-all");
-                        break;
-                case 1:
-                        if (words[cmdi + 1][0] == '!') {
-                                (words[cmdi + 1])++;
-                                if (gf_asprintf (&subop, "del%s", glob ? "-glob" : "") == -1)
-                                        subop = NULL;
-                        } else
-                                subop = gf_strdup ("get");
-
-                        ret = dict_set_str (dict, "op_name", ((char *)words[cmdi + 1]));
-                        if (ret < 0)
-                                goto out;
-                        break;
-                default:
-                        if (gf_asprintf (&subop, "set%s", glob ? "-glob" : "") == -1)
-                                subop = NULL;
-
-                        ret = dict_set_str (dict, "op_name", ((char *)words[cmdi + 1]));
-                        if (ret < 0)
-                                goto out;
-
-                        /* join the varargs by spaces to get the op_value */
-
-                        for (i = cmdi + 2; i < wordcount; i++)
-                                append_len += (strlen (words[i]) + 1);
-                        /* trailing strcat will add two bytes, make space for that */
-                        append_len++;
-
-                        append_str = GF_CALLOC (1, append_len, cli_mt_append_str);
-                        if (!append_str) {
-                                ret = -1;
-                                goto out;
-                        }
-
-                        for (i = cmdi + 2; i < wordcount; i++) {
-                                strcat (append_str, words[i]);
-                                strcat (append_str, " ");
-                        }
-                        append_str[append_len - 2] = '\0';
-
-                        /* "checkpoint now" is special: we resolve that "now" */
-                        if (strcmp (words[cmdi + 1], "checkpoint") == 0 &&
-                            strcmp (append_str, "now") == 0) {
-                                struct timeval tv = {0,};
-
-                                ret = gettimeofday (&tv, NULL);
-                                if (ret == -1)
-                                         goto out; /* FIXME: free append_str? */
-
-                                GF_FREE (append_str);
-                                append_str = GF_CALLOC (1, 300, cli_mt_append_str);
-                                if (!append_str) {
-                                        ret = -1;
-                                        goto out;
-                                }
-                                strcpy (append_str, "as of ");
-                                gf_time_fmt (append_str + strlen ("as of "),
-                                             300 - strlen ("as of "),
-                                             tv.tv_sec, gf_timefmt_FT);
-                        }
-
-                        ret = dict_set_dynstr (dict, "op_value", append_str);
-                }
-
-                ret = -1;
-                if (subop) {
-                        ret = dict_set_dynstr (dict, "subop", subop);
-                        if (!ret)
-                                subop = NULL;
-                }
-        }
+        if (!ret && type == GF_GSYNC_OPTION_TYPE_CONFIG)
+                ret = config_parse (words, wordcount, dict, cmdi, glob);
 
 out:
         if (ret) {
                 if (dict)
                         dict_destroy (dict);
-                GF_FREE (append_str);
         } else
                 *options = dict;
 
-        GF_FREE (subop);
 
         return ret;
 }
@@ -2068,7 +2169,7 @@ cli_cmd_get_statusop (const char *arg)
         uint32_t   ret       = GF_CLI_STATUS_NONE;
         char      *w         = NULL;
         char      *opwords[] = {"detail", "mem", "clients", "fd",
-                                "inode", "callpool", NULL};
+                                "inode", "callpool", "tasks", NULL};
         struct {
                 char      *opname;
                 uint32_t   opcode;
@@ -2079,6 +2180,7 @@ cli_cmd_get_statusop (const char *arg)
                 { "fd",       GF_CLI_STATUS_FD       },
                 { "inode",    GF_CLI_STATUS_INODE    },
                 { "callpool", GF_CLI_STATUS_CALLPOOL },
+                { "tasks",    GF_CLI_STATUS_TASKS    },
                 { NULL }
         };
 
@@ -2193,8 +2295,9 @@ cli_cmd_volume_status_parse (const char **words, int wordcount,
 
                 if (!strcmp (words[3], "nfs")) {
                         if (cmd == GF_CLI_STATUS_FD ||
-                            cmd == GF_CLI_STATUS_DETAIL) {
-                                cli_err ("Detail/FD status not available"
+                            cmd == GF_CLI_STATUS_DETAIL ||
+                            cmd == GF_CLI_STATUS_TASKS) {
+                                cli_err ("Detail/FD/Tasks status not available"
                                          " for NFS Servers");
                                 ret = -1;
                                 goto out;
@@ -2203,14 +2306,21 @@ cli_cmd_volume_status_parse (const char **words, int wordcount,
                 } else if (!strcmp (words[3], "shd")){
                         if (cmd == GF_CLI_STATUS_FD ||
                             cmd == GF_CLI_STATUS_CLIENTS ||
-                            cmd == GF_CLI_STATUS_DETAIL) {
-                                cli_err ("Detail/FD/Clients status not "
+                            cmd == GF_CLI_STATUS_DETAIL ||
+                            cmd == GF_CLI_STATUS_TASKS) {
+                                cli_err ("Detail/FD/Clients/Tasks status not "
                                          "available for Self-heal Daemons");
                                 ret = -1;
                                 goto out;
                         }
                         cmd |= GF_CLI_STATUS_SHD;
                 } else {
+                        if (cmd == GF_CLI_STATUS_TASKS) {
+                                cli_err ("Tasks status not available for "
+                                         "bricks");
+                                ret = -1;
+                                goto out;
+                        }
                         cmd |= GF_CLI_STATUS_BRICK;
                         ret = dict_set_str (dict, "brick", (char *)words[3]);
                 }
@@ -2359,12 +2469,103 @@ out:
        return ret;
 }
 
+static int
+extract_hostname_path_from_token (const char *tmp_words, char **hostname,
+                                  char **path)
+{
+        int ret = 0;
+        char *delimiter = NULL;
+        char *tmp_host = NULL;
+        char *host_name = NULL;
+        char *words = NULL;
+
+        *hostname = NULL;
+        *path = NULL;
+
+        words = GF_CALLOC (1, strlen (tmp_words) + 1, gf_common_mt_char);
+        if (!words){
+                ret = -1;
+                goto out;
+        }
+
+        strncpy (words, tmp_words, strlen (tmp_words) + 1);
+
+        if (validate_brick_name (words)) {
+                cli_err ("Wrong brick type: %s, use <HOSTNAME>:"
+                        "<export-dir-abs-path>", words);
+                ret = -1;
+                goto out;
+        } else {
+                delimiter = strrchr (words, ':');
+                ret = gf_canonicalize_path (delimiter + 1);
+                if (ret) {
+                        goto out;
+                } else {
+                        *path = GF_CALLOC (1, strlen (delimiter+1) +1,
+                                           gf_common_mt_char);
+                        if (!*path) {
+                           ret = -1;
+                                goto out;
+
+                        }
+                        strncpy (*path, delimiter +1,
+                                 strlen(delimiter + 1) + 1);
+                }
+        }
+
+        tmp_host = gf_strdup (words);
+        if (!tmp_host) {
+                gf_log ("cli", GF_LOG_ERROR, "Out of memory");
+                ret = -1;
+                goto out;
+        }
+        get_host_name (tmp_host, &host_name);
+        if (!host_name) {
+                ret = -1;
+                gf_log("cli",GF_LOG_ERROR, "Unable to allocate "
+                        "memory");
+                goto out;
+        }
+        if (!(strcmp (host_name, "localhost") &&
+            strcmp (host_name, "127.0.0.1") &&
+            strncmp (host_name, "0.", 2))) {
+                cli_err ("Please provide a valid hostname/ip other "
+                         "than localhost, 127.0.0.1 or loopback "
+                         "address (0.0.0.0 to 0.255.255.255).");
+                ret = -1;
+                goto out;
+        }
+        if (!valid_internet_address (host_name, _gf_false)) {
+                cli_err ("internet address '%s' does not conform to "
+                          "standards", host_name);
+                ret = -1;
+                goto out;
+        }
+
+        *hostname = GF_CALLOC (1, strlen (host_name) + 1,
+                                       gf_common_mt_char);
+        if (!*hostname) {
+                ret = -1;
+                goto out;
+        }
+        strncpy (*hostname, host_name, strlen (host_name) + 1);
+        ret = 0;
+
+out:
+        GF_FREE (words);
+        GF_FREE (tmp_host);
+        return ret;
+}
+
+
 int
 cli_cmd_volume_heal_options_parse (const char **words, int wordcount,
                                    dict_t **options)
 {
         int     ret = 0;
         dict_t  *dict = NULL;
+        char    *hostname = NULL;
+        char    *path = NULL;
 
         dict = dict_new ();
         if (!dict)
@@ -2386,6 +2587,11 @@ cli_cmd_volume_heal_options_parse (const char **words, int wordcount,
                         ret = dict_set_int32 (dict, "heal-op",
                                               GF_AFR_OP_HEAL_FULL);
                         goto done;
+                } else if (!strcmp (words[3], "statistics")) {
+                        ret = dict_set_int32 (dict, "heal-op",
+                                              GF_AFR_OP_STATISTICS);
+                        goto done;
+
                 } else if (!strcmp (words[3], "info")) {
                         ret = dict_set_int32 (dict, "heal-op",
                                               GF_AFR_OP_INDEX_SUMMARY);
@@ -2396,28 +2602,66 @@ cli_cmd_volume_heal_options_parse (const char **words, int wordcount,
                 }
         }
         if (wordcount == 5) {
-                if (strcmp (words[3], "info")) {
+                if (strcmp (words[3], "info") &&
+                    strcmp (words[3], "statistics")) {
                         ret = -1;
                         goto out;
                 }
-                if (!strcmp (words[4], "healed")) {
-                        ret = dict_set_int32 (dict, "heal-op",
-                                              GF_AFR_OP_HEALED_FILES);
-                        goto done;
-                }
-                if (!strcmp (words[4], "heal-failed")) {
-                        ret = dict_set_int32 (dict, "heal-op",
-                                              GF_AFR_OP_HEAL_FAILED_FILES);
-                        goto done;
+
+                if (!strcmp (words[3], "info")) {
+                        if (!strcmp (words[4], "healed")) {
+                                ret = dict_set_int32 (dict, "heal-op",
+                                                      GF_AFR_OP_HEALED_FILES);
+                                goto done;
+                        }
+                        if (!strcmp (words[4], "heal-failed")) {
+                                ret = dict_set_int32 (dict, "heal-op",
+                                                   GF_AFR_OP_HEAL_FAILED_FILES);
+                                goto done;
+                        }
+                        if (!strcmp (words[4], "split-brain")) {
+                                ret = dict_set_int32 (dict, "heal-op",
+                                                   GF_AFR_OP_SPLIT_BRAIN_FILES);
+                                goto done;
+                        }
                 }
-                if (!strcmp (words[4], "split-brain")) {
-                        ret = dict_set_int32 (dict, "heal-op",
-                                              GF_AFR_OP_SPLIT_BRAIN_FILES);
-                        goto done;
+
+                if (!strcmp (words[3], "statistics")) {
+                        if (!strcmp (words[4], "heal-count")) {
+                                ret = dict_set_int32 (dict, "heal-op",
+                                               GF_AFR_OP_STATISTICS_HEAL_COUNT);
+                                goto done;
+                        }
                 }
                 ret = -1;
                 goto out;
         }
+        if (wordcount == 7) {
+                if (!strcmp (words[3], "statistics")
+                    && !strcmp (words[4], "heal-count")
+                    && !strcmp (words[5], "replica")) {
+
+                        ret = dict_set_int32 (dict, "heal-op",
+                                   GF_AFR_OP_STATISTICS_HEAL_COUNT_PER_REPLICA);
+                        if (ret)
+                                goto out;
+                        ret = extract_hostname_path_from_token (words[6],
+                                                              &hostname, &path);
+                        if (ret)
+                                goto out;
+                        ret = dict_set_dynstr (dict, "per-replica-cmd-hostname",
+                                               hostname);
+                        if (ret)
+                                goto out;
+                        ret = dict_set_dynstr (dict, "per-replica-cmd-path",
+                                               path);
+                        if (ret)
+                                goto out;
+                        else
+                                goto done;
+
+                }
+        }
         ret = -1;
         goto out;
 done:
@@ -2522,3 +2766,917 @@ out:
 
         return ret;
 }
+
+int32_t
+cli_snap_create_desc_parse (dict_t *dict, const char **words,
+                            size_t wordcount, int32_t desc_opt_loc)
+{
+        int32_t        ret      = -1;
+        char          *desc     = NULL;
+        int32_t        desc_len = 0;
+
+        desc = GF_CALLOC (MAX_SNAP_DESCRIPTION_LEN + 1, sizeof(char),
+                          gf_common_mt_char);
+        if (!desc) {
+                ret = -1;
+                goto out;
+        }
+
+
+        if (strlen (words[desc_opt_loc]) >= MAX_SNAP_DESCRIPTION_LEN) {
+                cli_out ("snapshot create: description truncated: "
+                         "Description provided is longer than 1024 characters");
+                desc_len = MAX_SNAP_DESCRIPTION_LEN;
+        } else {
+                desc_len = strlen (words[desc_opt_loc]);
+        }
+
+        strncpy (desc, words[desc_opt_loc], desc_len);
+        desc[desc_len] = '\0';
+        /* Calculating the size of the description as given by the user */
+
+        ret = dict_set_dynstr (dict, "description", desc);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Unable to save snap "
+                        "description");
+                goto out;
+        }
+
+        ret = 0;
+out:
+        if (ret && desc)
+                GF_FREE (desc);
+
+        return ret;
+}
+
+/* Function to check whether the Volume name is repeated */
+int
+cli_check_if_volname_repeated (const char **words, unsigned int start_index,
+                           uint64_t cur_index) {
+        uint64_t        i       =       -1;
+        int             ret     =        0;
+
+        GF_ASSERT (words);
+
+        for (i = start_index ; i < cur_index ; i++) {
+                if (strcmp (words[i], words[cur_index]) == 0) {
+                        ret = -1;
+                        goto out;
+                }
+        }
+out :
+        return ret;
+}
+
+/* snapshot create <snapname> <vol-name(s)> [description <description>]
+ *                                           [force]
+ * @arg-0, dict     : Request Dictionary to be sent to server side.
+ * @arg-1, words    : Contains individual words of CLI command.
+ * @arg-2, wordcount: Contains number of words present in the CLI command.
+ *
+ * return value : -1 on failure
+ *                 0 on success
+ */
+int
+cli_snap_create_parse (dict_t *dict, const char **words, int wordcount) {
+        uint64_t        i               =        0;
+        int             ret             =       -1;
+        uint64_t        volcount        =        0;
+        char            key[PATH_MAX]   =        "";
+        char            *snapname       =       NULL;
+        unsigned int    cmdi            =        2;
+        /* cmdi is command index, here cmdi is "2" (gluster snapshot create)*/
+
+        GF_ASSERT (words);
+        GF_ASSERT (dict);
+
+        if (wordcount <= cmdi + 1) {
+                cli_err ("Invalid Syntax.");
+                gf_log ("cli", GF_LOG_ERROR,
+                        "Too less words for snap create command");
+                goto out;
+        }
+
+        if (strlen(words[cmdi]) >= GLUSTERD_MAX_SNAP_NAME) {
+                cli_err ("snapshot create: failed: snapname cannot exceed "
+                         "255 characters.");
+                gf_log ("cli", GF_LOG_ERROR, "Snapname too long");
+
+                goto out;
+        }
+
+        snapname = (char *) words[cmdi];
+        for (i = 0 ; i < strlen (snapname); i++) {
+                /* Following volume name convention */
+                if (!isalnum (snapname[i]) && (snapname[i] != '_'
+                                           && (snapname[i] != '-'))) {
+                        /* TODO : Is this message enough?? */
+                        cli_err ("Snapname can contain only alphanumeric, "
+                                 "\"-\" and \"_\" characters");
+                        goto out;
+                }
+        }
+
+        ret = dict_set_str (dict, "snapname", (char *)words[cmdi]);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Could not save snap "
+                        "name");
+                goto out;
+        }
+
+        /* Filling volume name in the dictionary */
+        for (i = cmdi + 1 ; i < wordcount
+                            && (strcmp (words[i], "description")) != 0
+                            && (strcmp (words[i], "force") != 0); i++) {
+                volcount++;
+                /* volume index starts from 1 */
+                ret = snprintf (key, sizeof (key),"volname%ld", volcount);
+                if (ret < 0) {
+                        goto out;
+                }
+
+                ret = dict_set_str (dict, key, (char *)words[i]);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR, "Could not "
+                                "save volume name");
+                        goto out;
+                }
+
+                if (i >= cmdi + 2) {
+                        ret = -1;
+                        cli_err("Creating multiple volume snapshot is not "
+                                "supported as of now");
+                        goto out;
+                }
+                /* TODO : remove this above condition check once
+                 * multiple volume snapshot is supported */
+        }
+
+        if (volcount == 0) {
+                ret = -1;
+                cli_err ("Please provide the volume name");
+                gf_log ("cli", GF_LOG_ERROR, "Invalid Syntax");
+                goto out;
+        }
+
+        ret = dict_set_int32 (dict, "volcount", volcount);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Could not save volcount");
+                goto out;
+        }
+
+        /* Verify how we got out of "for" loop,
+         * if it is by reaching wordcount limit then goto "out",
+         * because we need not parse for "description" and "force"
+         * after this.
+         */
+        if (i == wordcount) {
+                goto out;
+        }
+
+        if ((strcmp (words[i], "description")) == 0) {
+                ++i;
+                if (i > (wordcount - 1)) {
+                        ret = -1;
+                        cli_err ("Please provide a description");
+                        gf_log ("cli", GF_LOG_ERROR,
+                                "Description not provided");
+                        goto out;
+                }
+
+                ret = cli_snap_create_desc_parse(dict, words, wordcount, i);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR, "Could not save snap "
+                                "description");
+                        goto out;
+                }
+
+                if ( i == (wordcount - 1))
+                        goto out;
+                i++;
+                /* point the index to next word.
+                 * As description might be follwed by force option.
+                 * Before that, check if wordcount limit is reached
+                 */
+        }
+
+        if ((strcmp (words[i], "force") != 0)) {
+                ret = -1;
+                cli_err ("Invalid Syntax.");
+                gf_log ("cli", GF_LOG_ERROR, "Invalid Syntax");
+                goto out;
+        }
+        ret = dict_set_int8 (dict, "snap-force", 1);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Could not save "
+                        "snap force option");
+                goto out;
+        }
+
+        /* Check if the command has anything after "force" keyword */
+        if (++i < wordcount) {
+                ret = -1;
+                gf_log ("cli", GF_LOG_ERROR, "Invalid Syntax");
+                goto out;
+        }
+
+        ret = 0;
+
+out :
+        return ret;
+}
+
+/* snapshot list [volname]
+ * @arg-0, dict     : Request Dictionary to be sent to server side.
+ * @arg-1, words    : Contains individual words of CLI command.
+ * @arg-2, wordcount: Contains number of words present in the CLI command.
+ *
+ * return value : -1 on failure
+ *                 0 on success
+ */
+int
+cli_snap_list_parse (dict_t *dict, const char **words, int wordcount) {
+        int             ret     =       -1;
+
+        GF_ASSERT (words);
+        GF_ASSERT (dict);
+
+        if (wordcount < 2 || wordcount > 3) {
+                gf_log ("cli", GF_LOG_ERROR, "Invalid Syntax");
+                goto out;
+        }
+
+        if (wordcount == 2) {
+                ret = 0;
+                goto out;
+        }
+
+        ret = dict_set_str (dict, "volname", (char *)words[2]);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR,
+                        "Failed to save volname in dictionary");
+                goto out;
+        }
+out :
+        return ret;
+}
+
+/* snapshot info [(snapname |  volume <volname>)]
+ * @arg-0, dict     : Request Dictionary to be sent to server side.
+ * @arg-1, words    : Contains individual words of CLI command.
+ * @arg-2, wordcount: Contains number of words present in the CLI command.
+ *
+ * return value : -1 on failure
+ *                 0 on success
+ */
+int
+cli_snap_info_parse (dict_t *dict, const char **words, int wordcount)
+{
+
+        int             ret             =       -1;
+        int32_t         cmd             =       GF_SNAP_INFO_TYPE_ALL;
+        unsigned int    cmdi            =        2;
+        /* cmdi is command index, here cmdi is "2" (gluster snapshot info)*/
+
+        GF_ASSERT (words);
+        GF_ASSERT (dict);
+
+        if (wordcount > 4 || wordcount < cmdi) {
+                gf_log ("", GF_LOG_ERROR, "Invalid syntax");
+                goto out;
+        }
+
+        if (wordcount == cmdi) {
+                ret = 0;
+                goto out;
+        }
+
+        /* If 3rd word is not "volume", then it must
+         * be snapname.
+         */
+        if (strcmp (words[cmdi], "volume") != 0) {
+                ret = dict_set_str (dict, "snapname",
+                                   (char *)words[cmdi]);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR, "Unable to save "
+                                "snapname %s", words[cmdi]);
+                        goto out;
+                }
+
+                /* Once snap name is parsed, if we encounter any other
+                 * word then fail it. Invalid Syntax.
+                 * example : snapshot info <snapname> word
+                 */
+                if ((cmdi + 1) != wordcount) {
+                        ret = -1;
+                        gf_log ("cli", GF_LOG_ERROR, "Invalid Syntax");
+                        goto out;
+                }
+
+                cmd = GF_SNAP_INFO_TYPE_SNAP;
+                ret = 0;
+                goto out;
+                /* No need to continue the parsing once we
+                 * get the snapname
+                 */
+        }
+
+        /* If 3rd word is "volume", then check if next word
+         * is present. As, "snapshot info volume" is an
+         * invalid command.
+         */
+        if ((cmdi + 1) == wordcount) {
+                ret = -1;
+                gf_log ("cli", GF_LOG_ERROR, "Invalid Syntax");
+                goto out;
+        }
+
+        ret = dict_set_str (dict, "volname", (char *)words[wordcount - 1]);
+        if (ret) {
+                gf_log ("", GF_LOG_ERROR, "Count not save "
+                        "volume name %s", words[wordcount - 1]);
+                goto out;
+        }
+        cmd = GF_SNAP_INFO_TYPE_VOL;
+out :
+        if (ret == 0) {
+                ret = dict_set_int32 (dict, "cmd", cmd);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR, "Could not save "
+                                "type of snapshot info");
+                }
+        }
+        return ret;
+}
+
+
+
+/* snapshot restore <snapname>
+ * @arg-0, dict     : Request Dictionary to be sent to server side.
+ * @arg-1, words    : Contains individual words of CLI command.
+ * @arg-2, wordcount: Contains number of words present in the CLI command.
+ *
+ * return value : -1 on failure
+ *                 0 on success
+ */
+int
+cli_snap_restore_parse (dict_t *dict, const char **words, int wordcount)
+{
+
+        int             ret             =       -1;
+
+        GF_ASSERT (words);
+        GF_ASSERT (dict);
+
+        if (wordcount != 3) {
+                gf_log ("cli", GF_LOG_ERROR, "Invalid Syntax");
+                goto out;
+        }
+
+        ret = dict_set_str (dict, "snapname", (char *)words[2]);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Unable to save snap-name %s",
+                        words[2]);
+                goto out;
+        }
+out :
+        return ret;
+}
+
+/* snapshot delete <snapname>
+ * @arg-0, dict     : Request Dictionary to be sent to server side.
+ * @arg-1, words    : Contains individual words of CLI command.
+ * @arg-2, wordcount: Contains number of words present in the CLI command.
+ *
+ * return value : -1 on failure
+ *                 0 on success
+ *                 1 if user cancel the operation
+ */
+int
+cli_snap_delete_parse (dict_t *dict, const char **words, int wordcount,
+                       struct cli_state *state) {
+
+        int             ret             =       -1;
+        const char      *question       =       NULL;
+        gf_answer_t     answer          =       GF_ANSWER_NO;
+
+        question = "Deleting snap will erase all the information about "
+                   "the snap. Do you still want to continue?";
+
+        GF_ASSERT (words);
+        GF_ASSERT (dict);
+
+        if (wordcount != 3) {
+                gf_log ("cli", GF_LOG_ERROR, "Invalid Syntax");
+                goto out;
+        }
+
+        ret = dict_set_str (dict, "snapname", (char *)words[2]);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Unable to save snapname %s",
+                        words[2]);
+                goto out;
+        }
+
+        answer = cli_cmd_get_confirmation (state, question);
+        if (GF_ANSWER_NO == answer) {
+                ret = 1;
+                gf_log ("cli", GF_LOG_DEBUG, "User cancelled "
+                        "snapshot delete operation");
+                goto out;
+        }
+out :
+        return ret;
+}
+
+/* snapshot status [(snapname | volume <volname>)]
+ * @arg-0, dict     : Request Dictionary to be sent to server side.
+ * @arg-1, words    : Contains individual words of CLI command.
+ * @arg-2, wordcount: Contains number of words present in the CLI command.
+ *
+ * return value : -1 on failure
+ *                 0 on success
+ */
+int
+cli_snap_status_parse (dict_t *dict, const char **words, int wordcount)
+{
+
+        int             ret  =        -1;
+        int32_t         cmd  =       GF_SNAP_STATUS_TYPE_ALL;
+        unsigned int    cmdi =        2;
+        /* cmdi is command index, here cmdi is "2" (gluster snapshot status)*/
+
+        GF_ASSERT (words);
+        GF_ASSERT (dict);
+
+        if (wordcount > 4 || wordcount < cmdi) {
+                gf_log ("cli", GF_LOG_ERROR, "Invalid Syntax");
+                goto out;
+        }
+
+        if (wordcount == cmdi) {
+                ret = 0;
+                goto out;
+        }
+
+        /* if 3rd word is not "volume", then it must be "snapname"
+        */
+        if (strcmp (words[cmdi], "volume") != 0) {
+                ret = dict_set_str (dict, "snapname",
+                                   (char *)words[cmdi]);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR, "Count not save "
+                                "snap name %s", words[cmdi]);
+                        goto out;
+                }
+
+                if ((cmdi + 1) != wordcount) {
+                        ret = -1;
+                        gf_log ("cli", GF_LOG_ERROR, "Invalid Syntax");
+                        goto out;
+                }
+
+                ret = 0;
+                cmd = GF_SNAP_STATUS_TYPE_SNAP;
+                goto out;
+        }
+
+        /* If 3rd word is "volume", then check if next word is present.
+         * As, "snapshot info volume" is an invalid command
+         */
+        if ((cmdi + 1) == wordcount) {
+                ret = -1;
+                gf_log ("cli", GF_LOG_ERROR, "Invalid Syntax");
+                goto out;
+        }
+
+        ret = dict_set_str (dict, "volname", (char *)words [wordcount - 1]);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Count not save "
+                        "volume name %s", words[wordcount - 1]);
+                goto out;
+        }
+        cmd = GF_SNAP_STATUS_TYPE_VOL;
+
+out :
+        if (ret == 0) {
+                ret = dict_set_int32 (dict, "cmd", cmd);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR, "Could not save cmd "
+                                "of snapshot status");
+                }
+        }
+        return ret;
+}
+
+
+int32_t
+cli_snap_config_limit_parse (const char **words, dict_t *dict,
+                             unsigned int wordcount, unsigned int index,
+                             char *key)
+{
+        int             ret             =       -1;
+        int             limit           =       0;
+
+        GF_ASSERT (words);
+        GF_ASSERT (dict);
+        GF_ASSERT (key);
+
+        if (index >= wordcount) {
+                ret = -1;
+                cli_err ("Please provide a value for %s.",key);
+                gf_log ("cli", GF_LOG_ERROR, "Value not provided for %s", key);
+                goto out;
+        }
+
+        limit = strtol (words[index], NULL, 0);
+        if (limit <= 0) {
+                ret = -1;
+                cli_err ("%s should be greater than 0.", key);
+                goto out;
+        }
+
+        ret = dict_set_int32 (dict, key, limit);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Could not set "
+                        "%s in dictionary", key);
+                goto out;
+        }
+
+out :
+        return ret;
+}
+
+/* function cli_snap_config_parse
+ * Config Syntax : gluster snapshot config [volname]
+ *                                         [snap-max-hard-limit <count>]
+ *                                         [snap-max-soft-limit <count>]
+ *
+   return value: <0  on failure
+                  1  if user cancels the operation
+                  0  on success
+
+  NOTE : snap-max-soft-limit can only be set for system.
+*/
+int32_t
+cli_snap_config_parse (const char **words, int wordcount, dict_t *dict,
+                       struct cli_state *state)
+{
+        int                            ret                 = -1;
+        gf_answer_t                    answer              = GF_ANSWER_NO;
+        gf_boolean_t                   vol_presence        = _gf_false;
+        struct snap_config_opt_vals_  *conf_vals           = NULL;
+        int8_t                         hard_limit          = 0;
+        int8_t                         soft_limit          = 0;
+        int8_t                         config_type         = -1;
+        const char                    *question            = NULL;
+        unsigned int                    cmdi               = 2;
+        /* cmdi is command index, here cmdi is "2" (gluster snapshot config)*/
+
+        GF_ASSERT (words);
+        GF_ASSERT (dict);
+        GF_ASSERT (state);
+
+        if ((wordcount < 2) || (wordcount > 7)) {
+                gf_log ("cli", GF_LOG_ERROR,
+                        "Invalid wordcount(%d)", wordcount);
+                goto out;
+        }
+
+        if (wordcount == 2) {
+                config_type = GF_SNAP_CONFIG_DISPLAY;
+                ret = 0;
+                goto set;
+        }
+
+        /* Check whether the 3rd word is volname */
+        if (strcmp (words[cmdi], "snap-max-hard-limit") != 0
+             && strcmp (words[cmdi], "snap-max-soft-limit") != 0) {
+                ret = dict_set_str (dict, "volname", (char *)words[cmdi]);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR, "Failed to set volname");
+                        goto out;
+                }
+                cmdi++;
+                vol_presence = _gf_true;
+
+                if (cmdi == wordcount) {
+                        config_type = GF_SNAP_CONFIG_DISPLAY;
+                        ret = 0;
+                        goto set;
+                }
+        }
+
+        config_type = GF_SNAP_CONFIG_TYPE_SET;
+
+        if (strcmp (words[cmdi], "snap-max-hard-limit") == 0) {
+                ret = cli_snap_config_limit_parse (words, dict, wordcount,
+                                                ++cmdi, "snap-max-hard-limit");
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR, "Failed to parse snap "
+                                "config hard limit");
+                        goto out;
+                }
+                hard_limit = 1;
+
+                if (++cmdi == wordcount) {
+                        ret = 0;
+                        goto set;
+                }
+        }
+
+        if (strcmp (words[cmdi], "snap-max-soft-limit") == 0) {
+                if (vol_presence == 1) {
+                        ret = -1;
+                        cli_err ("Soft limit cannot be set to individual "
+                                  "volumes.");
+                        gf_log ("cli", GF_LOG_ERROR, "Soft limit cannot be "
+                                "set to volumes");
+                        goto out;
+                }
+
+                ret = cli_snap_config_limit_parse (words, dict, wordcount,
+                                                ++cmdi, "snap-max-soft-limit");
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR, "Failed to parse snap "
+                                "config soft limit");
+                        goto out;
+                }
+
+                if (++cmdi != wordcount) {
+                        ret = -1;
+                        gf_log ("cli", GF_LOG_ERROR, "Invalid Syntax");
+                        goto out;
+                }
+                soft_limit = 1;
+        } else {
+                ret = -1;
+                gf_log ("cli", GF_LOG_ERROR, "Invalid Syntax");
+                goto out;
+        }
+        ret = 0; /* Success */
+
+set:
+        ret = dict_set_int32 (dict, "config-command", config_type);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Unable to set "
+                        "config-command");
+                goto out;
+        }
+
+        if (config_type == GF_SNAP_CONFIG_TYPE_SET) {
+                conf_vals = snap_confopt_vals;
+                if (hard_limit && soft_limit) {
+                        question = conf_vals[GF_SNAP_CONFIG_SET_BOTH].question;
+                } else if (soft_limit) {
+                        question = conf_vals[GF_SNAP_CONFIG_SET_SOFT].question;
+                } else if (hard_limit) {
+                        question = conf_vals[GF_SNAP_CONFIG_SET_HARD].question;
+                }
+
+                answer = cli_cmd_get_confirmation (state, question);
+                if (GF_ANSWER_NO == answer) {
+                        ret = 1;
+                        gf_log ("cli", GF_LOG_DEBUG, "User cancelled "
+                        "snapshot config operation");
+                }
+        }
+
+out:
+        return ret;
+}
+
+int
+validate_snapname (const char *snapname, char **opwords) {
+        int     ret     =       -1;
+        int     i       =       0;
+
+        GF_ASSERT (snapname);
+        GF_ASSERT (opwords);
+
+        for (i = 0 ; opwords[i] != NULL; i++) {
+                if (strcmp (opwords[i], snapname) == 0) {
+                        cli_out ("\"%s\" cannot be a snapname", snapname);
+                        goto out;
+                }
+        }
+        ret = 0;
+out :
+        return ret;
+}
+
+int32_t
+cli_cmd_snapshot_parse (const char **words, int wordcount, dict_t **options,
+                        struct cli_state *state)
+{
+        int32_t            ret        = -1;
+        dict_t             *dict      = NULL;
+        gf1_cli_snapshot   type       = GF_SNAP_OPTION_TYPE_NONE;
+        char               *w         = NULL;
+        char               *opwords[] = {"create", "delete", "restore", "start",
+                                         "stop", "list", "status", "config",
+                                         "info", NULL};
+        char               *invalid_snapnames[] = {"description", "force",
+                                                  "volume", NULL};
+
+        GF_ASSERT (words);
+        GF_ASSERT (options);
+        GF_ASSERT (state);
+
+        dict = dict_new ();
+        if (!dict)
+                goto out;
+
+        /* Lowest wordcount possible */
+        if (wordcount < 2) {
+                gf_log ("", GF_LOG_ERROR,
+                        "Invalid command: Not enough arguments");
+                goto out;
+        }
+
+        w = str_getunamb (words[1], opwords);
+        if (!w) {
+                /* Checks if the operation is a valid operation */
+                gf_log ("", GF_LOG_ERROR, "Opword Mismatch");
+                goto out;
+        }
+
+        if (!strcmp (w, "create")) {
+                type = GF_SNAP_OPTION_TYPE_CREATE;
+        } else if (!strcmp (w, "list")) {
+                type = GF_SNAP_OPTION_TYPE_LIST;
+        } else if (!strcmp (w, "info")) {
+                type = GF_SNAP_OPTION_TYPE_INFO;
+        } else if (!strcmp (w, "delete")) {
+                type = GF_SNAP_OPTION_TYPE_DELETE;
+        } else if (!strcmp (w, "config")) {
+                type = GF_SNAP_OPTION_TYPE_CONFIG;
+        } else if (!strcmp (w, "restore")) {
+                type = GF_SNAP_OPTION_TYPE_RESTORE;
+        } else if (!strcmp (w, "status")) {
+                type = GF_SNAP_OPTION_TYPE_STATUS;
+        }
+
+        if (type != GF_SNAP_OPTION_TYPE_CONFIG) {
+                ret = dict_set_int32 (dict, "hold_snap_locks", _gf_true);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR,
+                                "Unable to set hold-snap-locks value "
+                                "as _gf_true");
+                        goto out;
+                }
+        }
+
+        /* Check which op is intended */
+        switch (type) {
+               case GF_SNAP_OPTION_TYPE_CREATE:
+               {
+                        /* Syntax :
+                         * gluster snapshot create <snapname> <vol-name(s)>
+                         *                         [description <description>]
+                         *                         [force]
+                         */
+
+                        /* In cases where the snapname is not given then
+                         * parsing fails & snapname cannot be "description",
+                         * "force" and "volume", that check is made here
+                         */
+                       if (wordcount == 2){
+                                ret = -1;
+                                gf_log ("cli", GF_LOG_ERROR,
+                                        "Invalid Syntax");
+                                goto out;
+                       }
+
+                       ret = validate_snapname (words[2], invalid_snapnames);
+                       if (ret) {
+                               goto out;
+                       }
+
+                       ret = cli_snap_create_parse (dict, words, wordcount);
+                       if (ret) {
+                               gf_log ("cli", GF_LOG_ERROR,
+                                       "create command parsing failed.");
+                               goto out;
+                       }
+                       break;
+               }
+                case GF_SNAP_OPTION_TYPE_INFO:
+                {
+                        /* Syntax :
+                         * gluster snapshot info [(snapname] | [vol <volname>)]
+                         */
+                        ret = cli_snap_info_parse (dict, words, wordcount);
+                        if (ret) {
+                                gf_log ("cli", GF_LOG_ERROR, "Failed to parse "
+                                        "snapshot info command");
+                                goto out;
+                        }
+                        break;
+                }
+
+                case GF_SNAP_OPTION_TYPE_LIST:
+                {
+                        /* Syntax :
+                         * gluster snaphsot list [volname]
+                         */
+
+                        ret = cli_snap_list_parse (dict, words, wordcount);
+                        if (ret) {
+                                gf_log ("cli", GF_LOG_ERROR, "Failed to parse "
+                                        "snapshot list command");
+                                goto out;
+                        }
+                        break;
+                }
+
+                case GF_SNAP_OPTION_TYPE_DELETE:
+                {
+                        /* Syntax :
+                         * gluster snapshot delete <snapname>
+                         */
+                        ret = cli_snap_delete_parse (dict, words, wordcount,
+                                                     state);
+                        if (ret) {
+                                gf_log ("cli", GF_LOG_ERROR, "Failed to parse "
+                                        "snapshot delete command");
+                                goto out;
+                        }
+                        break;
+                }
+
+               case GF_SNAP_OPTION_TYPE_CONFIG:
+               {
+                /* snapshot config [volname]  [snap-max-hard-limit <count>]
+                 *                            [snap-max-soft-limit <percent>] */
+                       ret = cli_snap_config_parse (words, wordcount, dict,
+                                                    state);
+                       if (ret) {
+                                if (ret < 0)
+                                        gf_log ("cli", GF_LOG_ERROR,
+                                              "config command parsing failed.");
+                               goto out;
+                       }
+
+                       ret = dict_set_int32 (dict, "type",
+                                             GF_SNAP_OPTION_TYPE_CONFIG);
+                       if (ret) {
+                               gf_log ("cli", GF_LOG_ERROR, "Unable to set "
+                                       "config type");
+                               ret = -1;
+                               goto out;
+                       }
+                       break;
+               }
+                case GF_SNAP_OPTION_TYPE_STATUS:
+                {
+                        /* Syntax :
+                         * gluster snapshot status [(snapname |
+                         *                         volume <volname>)]
+                         */
+                        ret = cli_snap_status_parse (dict, words, wordcount);
+                        if (ret) {
+                                gf_log ("cli", GF_LOG_ERROR, "Failed to parse "
+                                        "snapshot status command");
+                                goto out;
+                        }
+                        break;
+                }
+
+               case GF_SNAP_OPTION_TYPE_RESTORE:
+               {
+                       /* Syntax:
+                        * snapshot restore <snapname>
+                        */
+                       ret = cli_snap_restore_parse (dict, words, wordcount);
+                       if (ret) {
+                               gf_log ("cli", GF_LOG_ERROR, "Failed to parse "
+                                       "restore command");
+                               goto out;
+                       }
+                       break;
+               }
+               default:
+                       gf_log ("", GF_LOG_ERROR, "Opword Mismatch");
+                       goto out;
+                       break;
+        }
+
+        ret = dict_set_int32 (dict, "type", type);
+        if (ret) {
+                gf_log ("", GF_LOG_ERROR,
+                        "Failed to set type.");
+                goto out;
+        }
+        /* If you got so far, input is valid */
+        ret = 0;
+out:
+        if (ret) {
+                if (dict)
+                        dict_destroy (dict);
+        } else
+                *options = dict;
+
+        return ret;
+}
diff --git a/cli/src/cli-cmd-snapshot.c b/cli/src/cli-cmd-snapshot.c
new file mode 100644
index 000000000..de492d683
--- /dev/null
+++ b/cli/src/cli-cmd-snapshot.c
@@ -0,0 +1,146 @@
+/*
+   Copyright (c) 2013-2014 Red Hat, Inc. <http://www.redhat.com>
+   This file is part of GlusterFS.
+
+   This file is licensed to you under your choice of the GNU Lesser
+   General Public License, version 3 or any later version (LGPLv3 or
+   later), or the GNU General Public License, version 2 (GPLv2), in all
+   cases as published by the Free Software Foundation.
+*/
+#include <stdio.h>
+#include <string.h>
+#include <stdlib.h>
+#include <stdint.h>
+#include <pthread.h>
+
+#ifndef _CONFIG_H
+#define _CONFIG_H
+#include "config.h"
+#endif
+
+#include "cli.h"
+#include "cli-cmd.h"
+
+extern rpc_clnt_prog_t *cli_rpc_prog;
+
+int
+cli_cmd_snapshot_help_cbk (struct cli_state *state, struct cli_cmd_word *in_word,
+                           const char **words, int wordcount);
+
+int
+cli_cmd_snapshot_cbk (struct cli_state *state, struct cli_cmd_word *word,
+                      const char **words, int wordcount)
+{
+        int                      ret       = 0;
+        int                      parse_err = 0;
+        dict_t                  *options   = NULL;
+        rpc_clnt_procedure_t    *proc      = NULL;
+        call_frame_t            *frame     = NULL;
+        cli_local_t             *local     = NULL;
+
+        proc = &cli_rpc_prog->proctable [GLUSTER_CLI_SNAP];
+        if (proc == NULL) {
+               ret = -1;
+                goto out;
+        }
+
+        frame = create_frame (THIS, THIS->ctx->pool);
+        if (frame == NULL) {
+                ret = -1;
+                goto out;
+        }
+
+        /* Parses the command entered by the user */
+        ret = cli_cmd_snapshot_parse (words, wordcount, &options, state);
+        if (ret) {
+                if (ret < 0) {
+                        cli_usage_out (word->pattern);
+                        parse_err = 1;
+                }
+                else {
+                        /* User might have cancelled the snapshot operation */
+                        ret = 0;
+                }
+                goto out;
+        }
+
+        CLI_LOCAL_INIT (local, words, frame, options);
+
+        if (proc->fn)
+                ret = proc->fn (frame, THIS, options);
+
+out:
+        if (ret && parse_err == 0)
+                cli_out ("Snapshot command failed");
+
+        CLI_STACK_DESTROY (frame);
+
+        return ret;
+}
+
+struct cli_cmd snapshot_cmds[] = {
+        { "snapshot help",
+          cli_cmd_snapshot_help_cbk,
+          "display help for snapshot commands"
+        },
+        { "snapshot create <snapname> <volname(s)> [description <description>] [force]",
+          cli_cmd_snapshot_cbk,
+          "Snapshot Create."
+        },
+        { "snapshot restore <snapname>",
+          cli_cmd_snapshot_cbk,
+          "Snapshot Restore."
+        },
+        { "snapshot status [(snapname | volume <volname>)]",
+          cli_cmd_snapshot_cbk,
+          "Snapshot Status."
+        },
+        { "snapshot info [(snapname | volume <volname>)]",
+          cli_cmd_snapshot_cbk,
+          "Snapshot Info."
+        },
+        { "snapshot list [volname]",
+          cli_cmd_snapshot_cbk,
+          "Snapshot List."
+        },
+        {"snapshot config [volname] [snap-max-hard-limit <count>] [snap-max-soft-limit <percent>]",
+          cli_cmd_snapshot_cbk,
+          "Snapshot Config."
+        },
+        {"snapshot delete <snapname>",
+          cli_cmd_snapshot_cbk,
+          "Snapshot Delete."
+        },
+        { NULL, NULL, NULL }
+};
+
+int
+cli_cmd_snapshot_help_cbk (struct cli_state *state,
+                           struct cli_cmd_word *in_word,
+                           const char **words,
+                           int wordcount)
+{
+        struct cli_cmd        *cmd = NULL;
+
+        for (cmd = snapshot_cmds; cmd->pattern; cmd++)
+                if (_gf_false == cmd->disable)
+                        cli_out ("%s - %s", cmd->pattern, cmd->desc);
+
+        return 0;
+}
+
+int
+cli_cmd_snapshot_register (struct cli_state *state)
+{
+        int  ret = 0;
+        struct cli_cmd *cmd = NULL;
+
+        for (cmd = snapshot_cmds; cmd->pattern; cmd++) {
+
+                ret = cli_cmd_register (&state->tree, cmd);
+                if (ret)
+                        goto out;
+        }
+out:
+        return ret;
+}
diff --git a/cli/src/cli-cmd-system.c b/cli/src/cli-cmd-system.c
index f73758ae3..8cfa5e70c 100644
--- a/cli/src/cli-cmd-system.c
+++ b/cli/src/cli-cmd-system.c
@@ -31,6 +31,12 @@ extern rpc_clnt_prog_t *cli_rpc_prog;
 int cli_cmd_system_help_cbk (struct cli_state *state, struct cli_cmd_word *in_word,
                              const char **words, int wordcount);
 
+int cli_cmd_copy_file_cbk (struct cli_state *state, struct cli_cmd_word *word,
+                           const char **words, int wordcount);
+
+int cli_cmd_sys_exec_cbk (struct cli_state *state, struct cli_cmd_word *word,
+                          const char **words, int wordcount);
+
 int
 cli_cmd_getspec_cbk (struct cli_state *state, struct cli_cmd_word *word,
                      const char **words, int wordcount)
@@ -423,10 +429,151 @@ struct cli_cmd cli_system_cmds[] = {
            cli_cmd_system_help_cbk,
            "display help for system commands"},
 
+        { "system:: copy file [<filename>]",
+           cli_cmd_copy_file_cbk,
+           "Copy file from current node's $working_dir to "
+           "$working_dir of all cluster nodes"},
+
+        { "system:: execute <command> <args>",
+           cli_cmd_sys_exec_cbk,
+           "Execute the command on all the nodes "
+           "in the cluster and display their output."},
+
         { NULL, NULL, NULL }
 };
 
 int
+cli_cmd_sys_exec_cbk (struct cli_state *state, struct cli_cmd_word *word,
+                      const char **words, int wordcount)
+{
+        char                   cmd_arg_name[PATH_MAX] = "";
+        char                  *command                = NULL;
+        char                  *saveptr                = NULL;
+        char                  *tmp                    = NULL;
+        int                    ret                    = -1;
+        int                    i                      = -1;
+        int                    cmd_args_count         = 0;
+        int                    in_cmd_args_count      = 0;
+        rpc_clnt_procedure_t  *proc                   = NULL;
+        call_frame_t          *frame                  = NULL;
+        dict_t                *dict                   = NULL;
+        cli_local_t           *local                  = NULL;
+
+        if (wordcount < 3) {
+                cli_usage_out (word->pattern);
+                goto out;
+        }
+
+        dict = dict_new ();
+        if (!dict)
+                goto out;
+
+        command = strtok_r ((char *)words[2], " ", &saveptr);
+        do {
+                tmp = strtok_r (NULL, " ", &saveptr);
+                if (tmp) {
+                        in_cmd_args_count++;
+                        memset (cmd_arg_name, '\0', sizeof(cmd_arg_name));
+                        snprintf (cmd_arg_name, sizeof(cmd_arg_name),
+                                  "cmd_arg_%d", in_cmd_args_count);
+                        ret = dict_set_str (dict, cmd_arg_name, tmp);
+                        if (ret) {
+                                gf_log ("", GF_LOG_ERROR, "Unable to set "
+                                        "%s in dict", cmd_arg_name);
+                                goto out;
+                        }
+                }
+        } while (tmp);
+
+        cmd_args_count = wordcount - 3;
+
+        ret = dict_set_str (dict, "command", command);
+        if (ret) {
+                gf_log ("", GF_LOG_ERROR, "Unable to set command in dict");
+                goto out;
+        }
+
+        for (i=1; i <= cmd_args_count; i++) {
+                in_cmd_args_count++;
+                memset (cmd_arg_name, '\0', sizeof(cmd_arg_name));
+                snprintf (cmd_arg_name, sizeof(cmd_arg_name),
+                          "cmd_arg_%d", in_cmd_args_count);
+                ret = dict_set_str (dict, cmd_arg_name,
+                                    (char *)words[2+i]);
+                if (ret) {
+                        gf_log ("", GF_LOG_ERROR, "Unable to set %s in dict",
+                               cmd_arg_name);
+                        goto out;
+                }
+        }
+
+        ret = dict_set_int32 (dict, "cmd_args_count", in_cmd_args_count);
+        if (ret) {
+                gf_log ("", GF_LOG_ERROR,
+                        "Unable to set cmd_args_count in dict");
+                goto out;
+        }
+
+        ret = dict_set_str (dict, "volname", "N/A");
+        if (ret) {
+                gf_log ("", GF_LOG_ERROR, "Unable to set volname in dict");
+                goto out;
+        }
+
+        proc = &cli_rpc_prog->proctable[GLUSTER_CLI_SYS_EXEC];
+        if (proc && proc->fn) {
+                frame = create_frame (THIS, THIS->ctx->pool);
+                if (!frame)
+                        goto out;
+                CLI_LOCAL_INIT (local, words, frame, dict);
+                ret = proc->fn (frame, THIS, (void*)dict);
+        }
+out:
+        return ret;
+}
+
+int
+cli_cmd_copy_file_cbk (struct cli_state *state, struct cli_cmd_word *word,
+                       const char **words, int wordcount)
+{
+        int                    ret      = -1;
+        rpc_clnt_procedure_t  *proc     = NULL;
+        call_frame_t          *frame    = NULL;
+        char                  *filename = "";
+        dict_t                *dict     = NULL;
+        cli_local_t           *local    = NULL;
+
+        if (wordcount != 4) {
+                cli_usage_out (word->pattern);
+                goto out;
+        }
+
+        dict = dict_new ();
+        if (!dict)
+                goto out;
+
+        filename = (char*)words[3];
+        ret = dict_set_str (dict, "source", filename);
+        if (ret)
+                 gf_log ("", GF_LOG_ERROR, "Unable to set filename in dict");
+
+        ret = dict_set_str (dict, "volname", "N/A");
+        if (ret)
+                 gf_log ("", GF_LOG_ERROR, "Unable to set volname in dict");
+
+        proc = &cli_rpc_prog->proctable[GLUSTER_CLI_COPY_FILE];
+        if (proc && proc->fn) {
+                frame = create_frame (THIS, THIS->ctx->pool);
+                if (!frame)
+                        goto out;
+                CLI_LOCAL_INIT (local, words, frame, dict);
+                ret = proc->fn (frame, THIS, (void*)dict);
+        }
+out:
+        return ret;
+}
+
+int
 cli_cmd_system_help_cbk (struct cli_state *state, struct cli_cmd_word *in_word,
                          const char **words, int wordcount)
 {
diff --git a/cli/src/cli-cmd-volume-bdevice.c b/cli/src/cli-cmd-volume-bdevice.c
deleted file mode 100644
index 19325754f..000000000
--- a/cli/src/cli-cmd-volume-bdevice.c
+++ /dev/null
@@ -1,270 +0,0 @@
-/*
-  CLI for BD translator
-
-  Copyright IBM, Corp. 2012
-
-  This file is part of GlusterFS.
-
-  Author:
-  M. Mohan Kumar <mohan@in.ibm.com>
-
-  This file is licensed to you under your choice of the GNU Lesser
-  General Public License, version 3 or any later version (LGPLv3 or
-  later), or the GNU General Public License, version 2 (GPLv2), in all
-  cases as published by the Free Software Foundation.
-*/
-
-#ifndef _CONFIG_H
-#define _CONFIG_H
-#include "config.h"
-#endif
-
-#include "cli.h"
-#include "cli-cmd.h"
-#include <string.h>
-
-extern rpc_clnt_prog_t *cli_rpc_prog;
-
-int
-cli_cmd_bd_help_cbk (struct cli_state *state, struct cli_cmd_word *in_word,
-                      const char **words, int wordcount);
-
-int32_t
-cli_cmd_bd_parse (dict_t *dict, const char **words)
-{
-        char          *volname  = NULL;
-        char          *buff     = NULL;
-        char          *buffp    = NULL;
-        int            ret      = -1;
-        char          *save     = NULL;
-        char          *path     = NULL;
-        char          *size     = NULL;
-        char          *eptr     = NULL;
-        gf_xl_bd_op_t  bd_op    = GF_BD_OP_INVALID;
-        char          *dest_lv  = NULL;
-
-
-        /* volname:/path */
-        if (!strchr (words[2], ':') || !strchr (words[2], '/')) {
-                cli_out ("invalid parameter %s, needs <volname:/path>",
-                                words[2]);
-                return -1;
-        }
-        buff = buffp = gf_strdup (words[2]);
-        volname = strtok_r (buff, ":", &save);
-        path = strtok_r (NULL, ":", &save);
-
-        ret = dict_set_dynstr (dict, "volname", gf_strdup (volname));
-        if (ret)
-                goto out;
-
-        ret = dict_set_dynstr (dict, "path", gf_strdup (path));
-        if (ret)
-                goto out;
-
-        if (!strcasecmp (words[1], "create"))
-                bd_op = GF_BD_OP_NEW_BD;
-        else if (!strcasecmp (words[1], "delete"))
-                bd_op = GF_BD_OP_DELETE_BD;
-        else if (!strcasecmp (words[1], "clone"))
-                bd_op = GF_BD_OP_CLONE_BD;
-        else if (!strcasecmp (words[1], "snapshot"))
-                bd_op = GF_BD_OP_SNAPSHOT_BD;
-        else
-                return -1;
-
-        ret = dict_set_int32 (dict, "bd-op", bd_op);
-        if (ret)
-                goto out;
-
-        if (bd_op == GF_BD_OP_NEW_BD) {
-                /* If no suffix given we will treat it as MB */
-                strtoull (words[3], &eptr, 0);
-                /* no suffix */
-                if (!eptr[0])
-                        gf_asprintf (&size, "%sMB", words[3]);
-                else
-                        size = gf_strdup (words[3]);
-
-                ret = dict_set_dynstr (dict, "size", size);
-                if (ret)
-                        goto out;
-        } else if (bd_op == GF_BD_OP_SNAPSHOT_BD ||
-                   bd_op == GF_BD_OP_CLONE_BD) {
-                /*
-                 * dest_lv should be just dest_lv, we don't support
-                 * cloning/snapshotting to a different volume or vg
-                 */
-                if (strchr (words[3], ':') || strchr (words[3], '/')) {
-                        cli_err ("invalid parameter %s, volname/vg not needed",
-                                 words[3]);
-                        ret = -1;
-                        goto out;
-                }
-                dest_lv = gf_strdup (words[3]);
-                ret = dict_set_dynstr (dict, "dest_lv", dest_lv);
-                if (ret)
-                        goto out;
-
-                /* clone needs size as parameter */
-                if (bd_op == GF_BD_OP_SNAPSHOT_BD) {
-                        ret = dict_set_dynstr (dict, "size",
-                                               gf_strdup (words[4]));
-                        if (ret)
-                                goto out;
-                }
-        }
-
-        ret = 0;
-out:
-        GF_FREE (buffp);
-        return ret;
-}
-
-/*
- * bd create <volname>:/path <size>
- * bd delete <volname>:/path
- * bd clone <volname>:/path <newbd>
- * bd snapshot <volname>:/<path> <newbd> <size>
- */
-int32_t
-cli_cmd_bd_validate (const char **words, int wordcount, dict_t **options)
-{
-        dict_t  *dict = NULL;
-        int     ret   = -1;
-        char    *op[] = { "create", "delete", "clone", "snapshot", NULL };
-        int     index = 0;
-
-        for (index = 0; op[index]; index++)
-                if (!strcasecmp (words[1], op[index]))
-                        break;
-
-        if (!op[index])
-                return -1;
-
-        dict = dict_new ();
-        if (!dict)
-                goto out;
-
-        if (!strcasecmp (words[1], "create")) {
-                if (wordcount != 4)
-                        goto out;
-        } else if (!strcasecmp (words[1], "delete")) {
-                if (wordcount != 3)
-                        goto out;
-        } else if (!strcasecmp (words[1], "clone")) {
-                if (wordcount != 4)
-                        goto out;
-        } else if (!strcasecmp (words[1], "snapshot")) {
-                if (wordcount != 5)
-                        goto out;
-        } else {
-                ret = -1;
-                goto out;
-        }
-
-        ret = cli_cmd_bd_parse (dict, words);
-        if (ret < 0)
-                goto out;
-
-        *options = dict;
-        ret = 0;
-out:
-        if (ret)
-                dict_unref (dict);
-        return ret;
-}
-
-int
-cli_cmd_bd_cbk (struct cli_state *state, struct cli_cmd_word *word,
-                         const char **words, int wordcount)
-{
-        int                     ret         = -1;
-        rpc_clnt_procedure_t    *proc       = NULL;
-        call_frame_t            *frame      = NULL;
-        int                     sent        = 0;
-        int                     parse_error = 0;
-        dict_t                  *options    = NULL;
-        cli_local_t             *local = NULL;
-
-        proc = &cli_rpc_prog->proctable[GLUSTER_CLI_BD_OP];
-
-        frame = create_frame (THIS, THIS->ctx->pool);
-        if (!frame)
-                goto out;
-
-        ret = cli_cmd_bd_validate (words, wordcount, &options);
-        if (ret) {
-                cli_usage_out (word->pattern);
-                parse_error = 1;
-                goto out;
-        }
-
-        CLI_LOCAL_INIT (local, words, frame, options);
-        if (proc->fn) {
-                ret = proc->fn (frame, THIS, options);
-        }
-
-out:
-        if (options)
-                dict_unref (options);
-
-        if (ret) {
-                cli_cmd_sent_status_get (&sent);
-                if ((sent == 0) && (parse_error == 0))
-                        cli_out ("BD op failed!");
-        }
-
-        CLI_STACK_DESTROY (frame);
-
-        return ret;
-}
-
-struct cli_cmd cli_bd_cmds[] = {
-        { "bd help",
-           cli_cmd_bd_help_cbk,
-           "display help for bd command"},
-        { "bd create <volname>:<bd> <size>",
-          cli_cmd_bd_cbk,
-          "\n\tcreate a block device where size can be "
-          "suffixed with KB, MB etc. Default size is in MB"},
-        { "bd delete <volname>:<bd>",
-          cli_cmd_bd_cbk,
-          "Delete a block device"},
-        { "bd clone <volname>:<bd> <newbd>",
-          cli_cmd_bd_cbk,
-          "clone device"},
-        { "bd snapshot <volname>:<bd> <newbd> <size>",
-          cli_cmd_bd_cbk,
-          "\n\tsnapshot device where size can be "
-          "suffixed with KB, MB etc. Default size is in MB"},
-        { NULL, NULL, NULL }
-};
-
-int
-cli_cmd_bd_help_cbk (struct cli_state *state, struct cli_cmd_word *in_word,
-                      const char **words, int wordcount)
-{
-        struct cli_cmd *cmd = NULL;
-
-        for (cmd = cli_bd_cmds; cmd->pattern; cmd++)
-                if (_gf_false == cmd->disable)
-                        cli_out ("%s - %s", cmd->pattern, cmd->desc);
-
-        return 0;
-}
-
-int
-cli_cmd_bd_register (struct cli_state *state)
-{
-        int  ret = 0;
-        struct cli_cmd *cmd = NULL;
-
-        for (cmd = cli_bd_cmds; cmd->pattern; cmd++) {
-                ret = cli_cmd_register (&state->tree, cmd);
-                if (ret)
-                        goto out;
-        }
-out:
-        return ret;
-}
diff --git a/cli/src/cli-cmd-volume.c b/cli/src/cli-cmd-volume.c
index aa1c77553..100be0b73 100644
--- a/cli/src/cli-cmd-volume.c
+++ b/cli/src/cli-cmd-volume.c
@@ -897,6 +897,7 @@ cli_cmd_volume_set_cbk (struct cli_state *state, struct cli_cmd_word *word,
         call_frame_t            *frame = NULL;
         dict_t                  *options = NULL;
         cli_local_t             *local = NULL;
+        char                    *op_errstr = NULL;
 
         proc = &cli_rpc_prog->proctable[GLUSTER_CLI_SET_VOLUME];
 
@@ -904,9 +905,14 @@ cli_cmd_volume_set_cbk (struct cli_state *state, struct cli_cmd_word *word,
         if (!frame)
                 goto out;
 
-        ret = cli_cmd_volume_set_parse (words, wordcount, &options);
+        ret = cli_cmd_volume_set_parse (words, wordcount, &options, &op_errstr);
         if (ret) {
-                cli_usage_out (word->pattern);
+                if (op_errstr) {
+                    cli_err ("%s", op_errstr);
+                    GF_FREE (op_errstr);
+                } else
+                    cli_usage_out (word->pattern);
+
                 parse_error = 1;
                 goto out;
         }
@@ -1823,10 +1829,12 @@ struct cli_cmd volume_cmds[] = {
           "list information of all volumes"},
 
         { "volume create <NEW-VOLNAME> [stripe <COUNT>] [replica <COUNT>] "
+          "[transport <tcp|rdma|tcp,rdma>] <NEW-BRICK>"
 #ifdef HAVE_BD_XLATOR
-          "[device vg] "
+          "?<vg_name>"
 #endif
-          "[transport <tcp|rdma|tcp,rdma>] <NEW-BRICK> ... [force]",
+          "... [force]",
+
           cli_cmd_volume_create_cbk,
           "create a new volume of specified type with mentioned bricks"},
 
@@ -1850,7 +1858,7 @@ struct cli_cmd volume_cmds[] = {
           cli_cmd_volume_add_brick_cbk,
           "add brick to volume <VOLNAME>"},
 
-        { "volume remove-brick <VOLNAME> [replica <COUNT>] <BRICK> ... {start|stop|status|commit|force}",
+        { "volume remove-brick <VOLNAME> [replica <COUNT>] <BRICK> ... [start|stop|status|commit|force]",
           cli_cmd_volume_remove_brick_cbk,
           "remove brick from volume <VOLNAME>"},
 
@@ -1887,7 +1895,8 @@ struct cli_cmd volume_cmds[] = {
          "reset all the reconfigured options"},
 
 #if (SYNCDAEMON_COMPILE)
-        {"volume "GEOREP" [<VOLNAME>] [<SLAVE-URL>] {start|stop|config|status|log-rotate} [options...]",
+        {"volume "GEOREP" [<VOLNAME>] [<SLAVE-URL>] {create [push-pem] [force]"
+         "|start [force]|stop [force]|config|status [detail]|delete} [options...]",
          cli_cmd_volume_gsync_set_cbk,
          "Geo-sync operations",
          cli_cmd_check_gsync_exists_cbk},
@@ -1907,11 +1916,11 @@ struct cli_cmd volume_cmds[] = {
            "volume top operations"},
 
         { "volume status [all | <VOLNAME> [nfs|shd|<BRICK>]]"
-          " [detail|clients|mem|inode|fd|callpool]",
+          " [detail|clients|mem|inode|fd|callpool|tasks]",
           cli_cmd_volume_status_cbk,
           "display status of all or specified volume(s)/brick"},
 
-        { "volume heal <VOLNAME> [{full | info {healed | heal-failed | split-brain}}]",
+        { "volume heal <VOLNAME> [{full | statistics {heal-count {replica <hostname:brickname>}} |info {healed | heal-failed | split-brain}}]",
           cli_cmd_volume_heal_cbk,
           "self-heal commands on volume specified by <VOLNAME>"},
 
diff --git a/cli/src/cli-cmd.c b/cli/src/cli-cmd.c
index aa7600350..b81f75b5b 100644
--- a/cli/src/cli-cmd.c
+++ b/cli/src/cli-cmd.c
@@ -231,11 +231,9 @@ cli_cmds_register (struct cli_state *state)
         if (ret)
                 goto out;
 
-#ifdef HAVE_BD_XLATOR
-        ret = cli_cmd_bd_register (state);
+        ret = cli_cmd_snapshot_register (state);
         if (ret)
                 goto out;
-#endif
 out:
         return ret;
 }
diff --git a/cli/src/cli-cmd.h b/cli/src/cli-cmd.h
index 0ad37d498..041729276 100644
--- a/cli/src/cli-cmd.h
+++ b/cli/src/cli-cmd.h
@@ -93,6 +93,8 @@ int cli_cmd_probe_register (struct cli_state *state);
 
 int cli_cmd_system_register (struct cli_state *state);
 
+int cli_cmd_snapshot_register (struct cli_state *state);
+
 int cli_cmd_misc_register (struct cli_state *state);
 
 struct cli_cmd_word *cli_cmd_nextword (struct cli_cmd_word *word,
@@ -119,7 +121,4 @@ gf_answer_t
 cli_cmd_get_confirmation (struct cli_state *state, const char *question);
 int cli_cmd_sent_status_get (int *status);
 
-#ifdef HAVE_BD_XLATOR
-int cli_cmd_bd_register (struct cli_state *state);
-#endif
 #endif /* __CLI_CMD_H__ */
diff --git a/cli/src/cli-rpc-ops.c b/cli/src/cli-rpc-ops.c
index 28b92ee2f..bfeb854ad 100644
--- a/cli/src/cli-rpc-ops.c
+++ b/cli/src/cli-rpc-ops.c
@@ -13,10 +13,6 @@
 #include "config.h"
 #endif
 
-#ifndef GSYNC_CONF
-#define GSYNC_CONF GEOREP"/gsyncd.conf"
-#endif
-
 /* Widths of various columns in top read/write-perf output
  * Total width of top read/write-perf should be 80 chars
  * including one space between column
@@ -26,6 +22,8 @@
 #define VOL_TOP_PERF_SPEED_WIDTH        4
 #define VOL_TOP_PERF_TIME_WIDTH         26
 
+#define INDENT_MAIN_HEAD "%-25s %s "
+
 #include "cli.h"
 #include "compat-errno.h"
 #include "cli-cmd.h"
@@ -62,8 +60,15 @@ char *cli_vol_status_str[] = {"Created",
                               "Stopped",
                              };
 
-char *cli_volume_backend[] = {"",
-                           "Volume Group",
+char *cli_vol_task_status_str[] = {"not started",
+                                   "in progress",
+                                   "stopped",
+                                   "completed",
+                                   "failed",
+                                   "fix-layout in progress",
+                                   "fix-layout stopped",
+                                   "fix-layout completed",
+                                   "fix-layout failed",
 };
 
 int32_t
@@ -278,7 +283,7 @@ gf_cli_output_pool_list (dict_t *dict, int count)
                 else
                         connected_str = "Disconnected";
 
-                cli_out ("%s\t%s\t%s ", uuid_buf, hostname_buf,
+                cli_out ("%s\t%-9s\t%s ", uuid_buf, hostname_buf,
                          connected_str);
                 i++;
         }
@@ -493,7 +498,11 @@ gf_cli_get_volume_cbk (struct rpc_req *req, struct iovec *iov,
         char                       key[1024]            = {0};
         char                       err_str[2048]        = {0};
         gf_cli_rsp                 rsp                  = {0};
-        int32_t                    backend              = 0;
+        char                      *caps                 = NULL;
+        int                        k __attribute__((unused)) = 0;
+        // snap_volume variable helps in showing whether a volume is a normal
+        //volume or a volume for the snapshot
+        int32_t                    snap_volume          = 0;
 
         if (-1 == req->rpc_status)
                 goto out;
@@ -615,6 +624,11 @@ xml_output:
                 if (ret)
                         goto out;
 
+                snprintf (key, sizeof (key), "volume%d.snap_volume", i);
+                ret = dict_get_int32 (dict, key, &snap_volume);
+                if (ret)
+                        goto out;
+
                 snprintf (key, 256, "volume%d.brick_count", i);
                 ret = dict_get_int32 (dict, key, &brick_count);
                 if (ret)
@@ -645,9 +659,6 @@ xml_output:
                 if (ret)
                         goto out;
 
-                snprintf (key, 256, "volume%d.backend", i);
-                ret = dict_get_int32 (dict, key, &backend);
-
                 vol_type = type;
 
                 // Distributed (stripe/replicate/stripe-replica) setups
@@ -658,9 +669,44 @@ xml_output:
                 cli_out ("Type: %s", cli_vol_type_str[vol_type]);
                 cli_out ("Volume ID: %s", volume_id_str);
                 cli_out ("Status: %s", cli_vol_status_str[status]);
+                if (snap_volume)
+                        cli_out ("Snap Volume: %s", "yes");
+                else
+                        cli_out ("Snap Volume: %s", "no");
 
-                if (backend)
+#ifdef HAVE_BD_XLATOR
+                k = 0;
+                memset (key, 0, sizeof (key));
+                snprintf (key, sizeof (key), "volume%d.xlator%d", i, k);
+                ret = dict_get_str (dict, key, &caps);
+                if (ret)
                         goto next;
+                do {
+                        j = 0;
+                        cli_out ("Xlator %d: %s", k + 1, caps);
+                        do {
+                                memset (key, 0, sizeof (key));
+                                snprintf (key, sizeof (key),
+                                          "volume%d.xlator%d.caps%d",
+                                          i, k, j++);
+                                ret = dict_get_str (dict, key, &caps);
+                                if (ret)
+                                        break;
+                                cli_out ("Capability %d: %s", j, caps);
+                        } while (1);
+
+                        memset (key, 0, sizeof (key));
+                        snprintf (key, sizeof (key),
+                                  "volume%d.xlator%d", i, ++k);
+                        ret = dict_get_str (dict, key, &caps);
+                        if (ret)
+                                break;
+                } while (1);
+
+next:
+#else
+                caps = 0; /* Avoid compiler warnings when BD not enabled */
+#endif
 
                 if (type == GF_CLUSTER_TYPE_STRIPE_REPLICATE) {
                         cli_out ("Number of Bricks: %d x %d x %d = %d",
@@ -682,12 +728,6 @@ xml_output:
                          ((transport == 0)?"tcp":
                           (transport == 1)?"rdma":
                           "tcp,rdma"));
-
-next:
-                if (backend) {
-                        cli_out ("Backend Type: Block, %s",
-                                 cli_volume_backend[backend]);
-                }
                 j = 1;
 
                 GF_FREE (local->get_vol.volname);
@@ -703,6 +743,12 @@ next:
                                 goto out;
 
                         cli_out ("Brick%d: %s", j, brick);
+#ifdef HAVE_BD_XLATOR
+                        snprintf (key, 256, "volume%d.vg%d", i, j);
+                        ret = dict_get_str (dict, key, &caps);
+                        if (!ret)
+                                cli_out ("Brick%d VG: %s", j, caps);
+#endif
                         j++;
                 }
 
@@ -1206,10 +1252,11 @@ gf_cli_defrag_volume_cbk (struct rpc_req *req, struct iovec *iov,
         char                     msg[1024] = {0,};
         gf_defrag_status_t       status_rcd = GF_DEFRAG_STATUS_NOT_STARTED;
         int32_t                  counter = 0;
-        char                    *node_uuid = NULL;
+        char                    *node_name = NULL;
         char                     key[256] = {0,};
         int32_t                  i = 1;
         uint64_t                 failures = 0;
+        uint64_t                 skipped = 0;
         double                   elapsed = 0;
         char                    *size_str = NULL;
         char                    *task_id_str = NULL;
@@ -1330,17 +1377,18 @@ gf_cli_defrag_volume_cbk (struct rpc_req *req, struct iovec *iov,
                 goto out;
         }
 
-        cli_out ("%40s %16s %13s %13s %13s %14s %s", "Node", "Rebalanced-files",
-                 "size", "scanned", "failures", "status", "run time in secs");
-        cli_out ("%40s %16s %13s %13s %13s %14s %16s", "---------",
+        cli_out ("%40s %16s %13s %13s %13s %13s %20s %18s", "Node",
+                 "Rebalanced-files", "size", "scanned", "failures", "skipped",
+                 "status", "run time in secs");
+        cli_out ("%40s %16s %13s %13s %13s %13s %20s %18s", "---------",
                  "-----------", "-----------", "-----------", "-----------",
-                 "------------", "--------------");
+                 "-----------", "------------", "--------------");
         do {
-                snprintf (key, 256, "node-uuid-%d", i);
-                ret = dict_get_str (dict, key, &node_uuid);
+                snprintf (key, 256, "node-name-%d", i);
+                ret = dict_get_str (dict, key, &node_name);
                 if (ret)
                         gf_log (frame->this->name, GF_LOG_TRACE,
-                                "failed to get node-uuid");
+                                "failed to get node-name");
 
                 memset (key, 0, 256);
                 snprintf (key, 256, "files-%d", i);
@@ -1378,34 +1426,23 @@ gf_cli_defrag_volume_cbk (struct rpc_req *req, struct iovec *iov,
                                 "failed to get failures count");
 
                 memset (key, 0, 256);
+                snprintf (key, 256, "skipped-%d", i);
+                ret = dict_get_uint64 (dict, key, &skipped);
+                if (ret)
+                        gf_log (frame->this->name, GF_LOG_TRACE,
+                                "failed to get skipped count");
+                memset (key, 0, 256);
                 snprintf (key, 256, "run-time-%d", i);
                 ret = dict_get_double (dict, key, &elapsed);
                 if (ret)
                         gf_log (frame->this->name, GF_LOG_TRACE,
                                 "failed to get run-time");
 
-                switch (status_rcd) {
-                case GF_DEFRAG_STATUS_NOT_STARTED:
-                        status = "not started";
-                        break;
-                case GF_DEFRAG_STATUS_STARTED:
-                        status = "in progress";
-                        break;
-                case GF_DEFRAG_STATUS_STOPPED:
-                        status = "stopped";
-                        break;
-                case GF_DEFRAG_STATUS_COMPLETE:
-                        status = "completed";
-                        break;
-                case GF_DEFRAG_STATUS_FAILED:
-                        status = "failed";
-                        break;
-                }
-
+                status = cli_vol_task_status_str[status_rcd];
                 size_str = gf_uint64_2human_readable(size);
                 cli_out ("%40s %16"PRIu64 " %13s" " %13"PRIu64 " %13"PRIu64
-                         " %14s %16.2f", node_uuid, files, size_str, lookup,
-                         failures, status, elapsed);
+                         " %13"PRIu64 " %20s %18.2f", node_name, files,
+                         size_str, lookup, failures, skipped, status, elapsed);
                 GF_FREE(size_str);
 
                 i++;
@@ -1413,10 +1450,18 @@ gf_cli_defrag_volume_cbk (struct rpc_req *req, struct iovec *iov,
 
 
 done:
-        if (rsp.op_ret)
-                cli_err ("volume rebalance: %s: failed: %s", volname, msg);
-        else
-                cli_out ("volume rebalance: %s: success: %s", volname, msg);
+        if (global_state->mode & GLUSTER_MODE_XML)
+                cli_xml_output_str ("volRebalance", msg,
+                                    rsp.op_ret, rsp.op_errno,
+                                    rsp.op_errstr);
+        else {
+                if (rsp.op_ret)
+                        cli_err ("volume rebalance: %s: failed: %s", volname,
+                                 msg);
+                else
+                        cli_out ("volume rebalance: %s: success: %s", volname,
+                                 msg);
+        }
         ret = rsp.op_ret;
 
 out:
@@ -1494,7 +1539,7 @@ gf_cli_reset_volume_cbk (struct rpc_req *req, struct iovec *iov,
 
         gf_log ("cli", GF_LOG_INFO, "Received resp to reset");
 
-        if (rsp.op_ret &&  strcmp (rsp.op_errstr, ""))
+        if (strcmp (rsp.op_errstr, ""))
                 snprintf (msg, sizeof (msg), "%s", rsp.op_errstr);
         else
                 snprintf (msg, sizeof (msg), "reset volume %s",
@@ -1512,7 +1557,7 @@ gf_cli_reset_volume_cbk (struct rpc_req *req, struct iovec *iov,
         if (rsp.op_ret)
                 cli_err ("volume reset: failed: %s", msg);
         else
-                cli_out ("volume reset: success");
+                cli_out ("volume reset: success: %s", msg);
 
         ret = rsp.op_ret;
 
@@ -1718,9 +1763,10 @@ gf_cli3_remove_brick_status_cbk (struct rpc_req *req, struct iovec *iov,
         char                     key[256] = {0,};
         int32_t                  i       = 1;
         int32_t                  counter = 0;
-        char                    *node_uuid = 0;
+        char                    *node_name = 0;
         gf_defrag_status_t       status_rcd = GF_DEFRAG_STATUS_NOT_STARTED;
         uint64_t                 failures = 0;
+        uint64_t                 skipped = 0;
         double                   elapsed = 0;
         char                    *size_str = NULL;
         int32_t                  command = 0;
@@ -1821,18 +1867,19 @@ xml_output:
         }
 
 
-        cli_out ("%40s %16s %13s %13s %13s %14s %s", "Node", "Rebalanced-files",
-                 "size", "scanned", "failures", "status", "run-time in secs");
-        cli_out ("%40s %16s %13s %13s %13s %14s %16s", "---------",
+        cli_out ("%40s %16s %13s %13s %13s %13s %14s %s", "Node",
+                 "Rebalanced-files", "size", "scanned", "failures", "skipped",
+                 "status", "run-time in secs");
+        cli_out ("%40s %16s %13s %13s %13s %13s %14s %16s", "---------",
                  "-----------", "-----------", "-----------", "-----------",
-                 "------------", "--------------");
+                  "-----------","------------", "--------------");
 
         do {
-                snprintf (key, 256, "node-uuid-%d", i);
-                ret = dict_get_str (dict, key, &node_uuid);
+                snprintf (key, 256, "node-name-%d", i);
+                ret = dict_get_str (dict, key, &node_name);
                 if (ret)
                         gf_log (frame->this->name, GF_LOG_TRACE,
-                                "failed to get node-uuid");
+                                "failed to get node-name");
 
                 memset (key, 0, 256);
                 snprintf (key, 256, "files-%d", i);
@@ -1868,6 +1915,11 @@ xml_output:
                         gf_log (frame->this->name, GF_LOG_TRACE,
                                 "Failed to get failure on files");
 
+                snprintf (key, 256, "failures-%d", i);
+                ret = dict_get_uint64 (dict, key, &skipped);
+                if (ret)
+                        gf_log (frame->this->name, GF_LOG_TRACE,
+                                "Failed to get skipped files");
                 memset (key, 0, 256);
                 snprintf (key, 256, "run-time-%d", i);
                 ret = dict_get_double (dict, key, &elapsed);
@@ -1891,14 +1943,17 @@ xml_output:
                 case GF_DEFRAG_STATUS_FAILED:
                         status = "failed";
                         break;
+                default:
+                        break;
                 }
 
                 size_str = gf_uint64_2human_readable(size);
                 
 		if (strcmp (status, "not started")) {
 			cli_out ("%40s %16"PRIu64 " %13s" " %13"PRIu64 " %13"
-				PRIu64 " %14s %16.2f", node_uuid, files, 
-				size_str, lookup, failures, status, elapsed);
+				PRIu64 " %13"PRIu64 " %14s %16.2f", node_name,
+                                files, size_str, lookup, failures, skipped,
+                                status, elapsed);
 		}
                 GF_FREE(size_str);
 
@@ -2916,142 +2971,6 @@ out:
         return ret;
 }
 
-#ifdef HAVE_BD_XLATOR
-int
-gf_cli_bd_op_cbk (struct rpc_req *req, struct iovec *iov,
-                             int count, void *myframe)
-{
-        gf_cli_rsp              rsp         = {0,};
-        int                     ret         = -1;
-        cli_local_t             *local      = NULL;
-        dict_t                  *dict       = NULL;
-        dict_t                  *input_dict = NULL;
-        gf_xl_bd_op_t           bd_op       = GF_BD_OP_INVALID;
-        char                    *operation  = NULL;
-        call_frame_t            *frame      = NULL;
-
-        if (-1 == req->rpc_status)
-                goto out;
-
-        frame = myframe;
-
-        ret = xdr_to_generic (*iov, &rsp, (xdrproc_t)xdr_gf_cli_rsp);
-        if (ret < 0) {
-                gf_log (frame->this->name, GF_LOG_ERROR,
-                        "Failed to decode xdr response");
-                goto out;
-        }
-
-        dict = dict_new ();
-        if (!dict) {
-                ret = -1;
-                goto out;
-        }
-
-        if (frame)
-                local = frame->local;
-
-        if (local) {
-                input_dict = local->dict;
-                ret = dict_get_int32 (input_dict, "bd-op",
-                                      (int32_t *)&bd_op);
-        }
-
-        switch (bd_op) {
-        case GF_BD_OP_NEW_BD:
-                operation = gf_strdup ("create");
-                break;
-        case GF_BD_OP_DELETE_BD:
-                operation = gf_strdup ("delete");
-                break;
-        case GF_BD_OP_CLONE_BD:
-                operation = gf_strdup ("clone");
-                break;
-        case GF_BD_OP_SNAPSHOT_BD:
-                operation = gf_strdup ("snapshot");
-                break;
-        default:
-                break;
-        }
-
-        ret = dict_unserialize (rsp.dict.dict_val, rsp.dict.dict_len, &dict);
-        if (ret)
-                goto out;
-
-        gf_log ("cli", GF_LOG_INFO, "Received resp to %s bd op", operation);
-
-        if (global_state->mode & GLUSTER_MODE_XML) {
-                ret = cli_xml_output_dict ("BdOp", dict, rsp.op_ret,
-                                           rsp.op_errno, rsp.op_errstr);
-                if (ret)
-                        gf_log ("cli", GF_LOG_ERROR,
-                                "Error outputting to xml");
-                goto out;
-        }
-
-        if (rsp.op_ret && strcmp (rsp.op_errstr, ""))
-                cli_err ("%s", rsp.op_errstr);
-        else
-                cli_out ("BD %s has been %s", operation,
-                                (rsp.op_ret) ? "unsuccessful":
-                                "successful.");
-        ret = rsp.op_ret;
-
-out:
-        cli_cmd_broadcast_response (ret);
-
-        if (dict)
-                dict_unref (dict);
-
-        if (operation)
-                GF_FREE (operation);
-
-        if (rsp.dict.dict_val)
-                free (rsp.dict.dict_val);
-        if (rsp.op_errstr)
-                free (rsp.op_errstr);
-        return ret;
-}
-
-int32_t
-gf_cli_bd_op (call_frame_t *frame, xlator_t *this,
-                      void *data)
-{
-        gf_cli_req      req    = { {0,} };
-        int             ret    = 0;
-        dict_t          *dict  = NULL;
-
-        if (!frame || !this || !data) {
-                ret = -1;
-                goto out;
-        }
-
-        dict = dict_ref ((dict_t *)data);
-        if (!dict)
-                goto out;
-
-        ret = dict_allocate_and_serialize (dict,
-                                           &req.dict.dict_val,
-                                           &req.dict.dict_len);
-
-
-        ret = cli_cmd_submit (&req, frame, cli_rpc_prog,
-                              GLUSTER_CLI_BD_OP, NULL,
-                              this, gf_cli_bd_op_cbk,
-                              (xdrproc_t) xdr_gf_cli_req);
-
-out:
-        if (dict)
-                dict_unref (dict);
-
-        if (req.dict.dict_val)
-                GF_FREE (req.dict.dict_val);
-
-        gf_log ("cli", GF_LOG_DEBUG, "Returning %d", ret);
-        return ret;
-}
-#endif
-
 int32_t
 gf_cli_create_volume (call_frame_t *frame, xlator_t *this,
                          void *data)
@@ -3752,8 +3671,11 @@ gf_cli_gsync_config_command (dict_t *dict)
         char *subop         = NULL;
         char *gwd           = NULL;
         char *slave         = NULL;
+        char *confpath      = NULL;
         char *master        = NULL;
         char *op_name       = NULL;
+        int   ret           = -1;
+        char  conf_path[PATH_MAX] = "";
 
         if (dict_get_str (dict, "subop", &subop) != 0)
                 return -1;
@@ -3772,9 +3694,17 @@ gf_cli_gsync_config_command (dict_t *dict)
         if (dict_get_str (dict, "op_name", &op_name) != 0)
                 op_name = NULL;
 
+        ret = dict_get_str (dict, "conf_path", &confpath);
+        if (!confpath) {
+                ret = snprintf (conf_path, sizeof(conf_path) - 1,
+                                "%s/"GEOREP"/gsyncd_template.conf", gwd);
+                conf_path[ret] = '\0';
+                confpath = conf_path;
+        }
+
         runinit (&runner);
         runner_add_args (&runner, GSYNCD_PREFIX"/gsyncd", "-c", NULL);
-        runner_argprintf (&runner, "%s/"GSYNC_CONF, gwd);
+        runner_argprintf (&runner, "%s", confpath);
         if (master)
                 runner_argprintf (&runner, ":%s", master);
         runner_add_arg (&runner, slave);
@@ -3786,67 +3716,639 @@ gf_cli_gsync_config_command (dict_t *dict)
 }
 
 int
-gf_cli_gsync_out_status (dict_t *dict)
+gf_cli_fetch_gsyncd_status_values (char *status,
+                                   gf_cli_gsync_status_t *sts_val)
 {
-        int   gsync_count              = 0;
-        int   i                        = 0;
-        int   ret                      = 0;
-        char             mst[PATH_MAX] = {0, };
-        char             slv[PATH_MAX] = {0, };
-        char             sts[PATH_MAX] = {0, };
-        char             nds[PATH_MAX] = {0, };
-        char             hyphens[100]  = {0, };
-        char *mst_val                  = NULL;
-        char *slv_val                  = NULL;
-        char *sts_val                  = NULL;
-        char *nds_val                  = NULL;
-
-        cli_out ("%-20s %-20s %-50s %-10s", "NODE", "MASTER", "SLAVE", "STATUS");
-
-        for (i=0; i<sizeof(hyphens)-1; i++)
-                hyphens[i] = '-';
+        int32_t   ret      = -1;
+        char     *tmp      = NULL;
+        char     *save_ptr = NULL;
+        char     *key      = NULL;
+        char     *value    = NULL;
 
-        cli_out ("%s", hyphens);
+        if (!status || !sts_val) {
+                gf_log ("", GF_LOG_ERROR, "status or sts_val is null");
+                goto out;
+        }
 
+        tmp = strtok_r (status, "\n", &save_ptr);
 
-        ret = dict_get_int32 (dict, "gsync-count", &gsync_count);
+        if (tmp)
+                sts_val->health = gf_strdup (tmp);
+
+        while (tmp) {
+                key = strtok_r (tmp, "=", &value);
+
+                if ((key) && (!strcmp(key, "Uptime")))
+                        sts_val->uptime = gf_strdup (value);
+
+                if ((key) && (!strcmp(key, "FilesSyncd")))
+                        sts_val->files_syncd = gf_strdup (value);
+
+                if ((key) && (!strcmp(key, "FilesPending")))
+                        sts_val->files_pending = gf_strdup (value);
+
+                if ((key) && (!strcmp(key, "BytesPending"))) {
+                        value = gf_uint64_2human_readable(atol(value));
+                        sts_val->bytes_pending = gf_strdup (value);
+                }
+
+                if ((key) && (!strcmp(key, "DeletesPending")))
+                        sts_val->deletes_pending = gf_strdup (value);
+
+                tmp = strtok_r (NULL, ";", &save_ptr);
+        }
+
+        if (sts_val->health)
+                ret = 0;
+
+        if (!sts_val->uptime)
+                sts_val->uptime = gf_strdup ("N/A");
+
+        if (!sts_val->files_syncd)
+                sts_val->files_syncd = gf_strdup ("N/A");
+
+        if (!sts_val->files_pending)
+                sts_val->files_pending = gf_strdup ("N/A");
+
+        if (!sts_val->bytes_pending)
+                sts_val->bytes_pending = gf_strdup ("N/A");
+
+        if (!sts_val->deletes_pending)
+                sts_val->deletes_pending = gf_strdup ("N/A");
+
+out:
+        gf_log ("", GF_LOG_DEBUG, "Returning %d.", ret);
+        return ret;
+}
+
+char*
+get_struct_variable (int mem_num, gf_cli_gsync_status_t *sts_val)
+{
+        switch (mem_num) {
+        case 0: return (sts_val->node);
+        case 1: return (sts_val->master);
+        case 2: return (sts_val->slave);
+        case 3: return (sts_val->health);
+        case 4: return (sts_val->uptime);
+        case 5: return (sts_val->files_syncd);
+        case 6: return (sts_val->files_pending);
+        case 7: return (sts_val->bytes_pending);
+        case 8: return (sts_val->deletes_pending);
+        default:
+                 goto out;
+        }
+
+out:
+        return NULL;
+}
+
+int
+gf_cli_print_status (char **title_values,
+                     gf_cli_gsync_status_t **sts_vals,
+                     int *spacing, int gsync_count,
+                     int number_of_fields, int is_detail)
+{
+        int     indents                  = 0;
+        int     i                        = 0;
+        int     j                        = 0;
+        int     ret                      = 0;
+        int     total_spacing            = 0;
+        char  **output_values            = NULL;
+        char   *tmp                      = NULL;
+        char   *hyphens                  = NULL;
+        char    heading[PATH_MAX]        = {0, };
+        char    indent_spaces[PATH_MAX]  = {0, };
+
+        /* calculating spacing for hyphens */
+        for (i = 0; i < number_of_fields; i++) {
+                /* Suppressing master and slave output for status detail */
+                if ((is_detail) && ((i == 1) || (i == 2))) {
+                        total_spacing++;
+                        continue;
+                } else if ((!is_detail) && (i > 4)) {
+                       /* Suppressing detailed output for
+                        * status */
+                        continue;
+                }
+                spacing[i] += 3; /* Adding extra space to
+                                    distinguish between fields */
+                total_spacing += spacing[i];
+        }
+        total_spacing += 4; /* For the spacing between the fields */
+
+        /* char pointers for each field */
+        output_values = GF_CALLOC (number_of_fields, sizeof (char *),
+                                   gf_common_mt_char);
+        if (!output_values) {
+                ret = -1;
+                goto out;
+        }
+        for (i = 0; i < number_of_fields; i++) {
+                output_values[i] = GF_CALLOC (spacing[i] + 1, sizeof (char),
+                                              gf_common_mt_char);
+                if (!output_values[i]) {
+                        ret = -1;
+                        goto out;
+                }
+        }
+
+        hyphens = GF_CALLOC (total_spacing + 1, sizeof (char),
+                             gf_common_mt_char);
+        if (!hyphens) {
+                ret = -1;
+                goto out;
+        }
+
+        ret = snprintf(heading, sizeof(heading), "MASTER: %s  SLAVE: %s",
+                       sts_vals[0]->master, sts_vals[0]->slave);
         if (ret) {
-                gf_log ("cli", GF_LOG_INFO, "No active geo-replication sessions"
-                        "present for the selected");
+                if (ret < sizeof(heading))
+                        heading[ret] = '\0';
+                else
+                        heading[sizeof(heading) - 1] = '\0';
                 ret = 0;
+        } else {
+                ret = -1;
                 goto out;
         }
 
-        for (i = 1; i <= gsync_count; i++) {
-                snprintf (nds, sizeof(nds), "node%d", i);
-                snprintf (mst, sizeof(mst), "master%d", i);
-                snprintf (slv, sizeof(slv), "slave%d", i);
-                snprintf (sts, sizeof(sts), "status%d", i);
+        if (is_detail) {
+                cli_out (" ");
+                if (strlen(heading) > total_spacing)
+                        cli_out ("%s", heading);
+                else {
+                        /* Printing the heading with centre justification */
+                        indents = (total_spacing - strlen(heading)) / 2;
+                        memset (indent_spaces, ' ', indents);
+                        indent_spaces[indents] = '\0';
+                        ret = snprintf (hyphens, total_spacing, "%s%s",
+                                        indent_spaces, heading);
+                        if (ret) {
+                                hyphens[ret] = '\0';
+                                cli_out ("%s", hyphens);
+                                ret = 0;
+                        } else {
+                                ret = -1;
+                                goto out;
+                        }
+                }
+                cli_out (" ");
+        }
 
-                ret = dict_get_str (dict, nds, &nds_val);
+        /* setting the title "NODE", "MASTER", etc. from title_values[]
+           and printing the same */
+        for (j = 0; j < number_of_fields; j++) {
+                /* Suppressing master and slave output for status detail */
+                if ((is_detail) && ((j == 1) || (j == 2))) {
+                        output_values[j][0] = '\0';
+                        continue;
+                } else if ((!is_detail) && (j > 4)) {
+                       /* Suppressing detailed output for
+                        * status */
+                       output_values[j][0] = '\0';
+                        continue;
+                }
+                memset (output_values[j], ' ', spacing[j]);
+                memcpy (output_values[j], title_values[j],
+                        strlen(title_values[j]));
+                output_values[j][spacing[j]] = '\0';
+        }
+        cli_out ("%s %s %s %s %s %s %s %s %s", output_values[0],
+                 output_values[1], output_values[2], output_values[3],
+                 output_values[4], output_values[5], output_values[6],
+                 output_values[7], output_values[8]);
+
+        /* setting and printing the hyphens */
+        memset (hyphens, '-', total_spacing);
+        hyphens[total_spacing] = '\0';
+        cli_out ("%s", hyphens);
+
+        for (i = 0; i < gsync_count; i++) {
+                for (j = 0; j < number_of_fields; j++) {
+                        /* Suppressing master and slave output for
+                         * status detail */
+                        if ((is_detail) && ((j == 1) || (j == 2))) {
+                                output_values[j][0] = '\0';
+                                continue;
+                        }  else if ((!is_detail) && (j > 4)) {
+                                /* Suppressing detailed output for
+                                 * status */
+                                output_values[j][0] = '\0';
+                                continue;
+                        }
+                        tmp = get_struct_variable(j, sts_vals[i]);
+                        if (!tmp) {
+                                gf_log ("", GF_LOG_ERROR,
+                                        "struct member empty.");
+                                ret = -1;
+                                goto out;
+                        }
+                        memset (output_values[j], ' ', spacing[j]);
+                        memcpy (output_values[j], tmp, strlen (tmp));
+                        output_values[j][spacing[j]] = '\0';
+                }
+
+                cli_out ("%s %s %s %s %s %s %s %s %s", output_values[0],
+                         output_values[1], output_values[2], output_values[3],
+                         output_values[4], output_values[5], output_values[6],
+                         output_values[7], output_values[8]);
+        }
+
+out:
+        if (output_values) {
+                for (i = 0; i < number_of_fields; i++) {
+                        if (output_values[i])
+                                GF_FREE (output_values[i]);
+                }
+                GF_FREE (output_values);
+        }
+
+        if (hyphens)
+                GF_FREE (hyphens);
+
+        return ret;
+}
+
+int
+gf_cli_read_status_data (dict_t *dict,
+                         gf_cli_gsync_status_t **sts_vals,
+                         int *spacing, int gsync_count,
+                         int number_of_fields)
+{
+        int     ret            = 0;
+        int     i              = 0;
+        int     j              = 0;
+        char    mst[PATH_MAX]  = {0, };
+        char    slv[PATH_MAX]  = {0, };
+        char    sts[PATH_MAX]  = {0, };
+        char    nds[PATH_MAX]  = {0, };
+        char   *status         = NULL;
+        char   *tmp            = NULL;
+
+        /* Storing per node status info in each object */
+        for (i = 0; i < gsync_count; i++) {
+                snprintf (nds, sizeof(nds), "node%d", i + 1);
+                snprintf (mst, sizeof(mst), "master%d", i + 1);
+                snprintf (slv, sizeof(slv), "slave%d", i + 1);
+                snprintf (sts, sizeof(sts), "status%d", i + 1);
+
+                /* Fetching the values from dict, and calculating
+                   the max length for each field */
+                ret = dict_get_str (dict, nds, &(sts_vals[i]->node));
                 if (ret)
                         goto out;
 
-                ret = dict_get_str (dict, mst, &mst_val);
+                ret = dict_get_str (dict, mst, &(sts_vals[i]->master));
                 if (ret)
                         goto out;
 
-                ret = dict_get_str (dict, slv, &slv_val);
+                ret = dict_get_str (dict, slv, &(sts_vals[i]->slave));
                 if (ret)
                         goto out;
 
-                ret = dict_get_str (dict, sts, &sts_val);
+                ret = dict_get_str (dict, sts, &status);
                 if (ret)
                         goto out;
 
-                cli_out ("%-20s %-20s %-50s %-10s", nds_val, mst_val,
-                         slv_val, sts_val);
+                /* Fetching health and uptime from sts_val */
+                ret = gf_cli_fetch_gsyncd_status_values (status, sts_vals[i]);
+                if (ret)
+                        goto out;
 
+                for (j = 0; j < number_of_fields; j++) {
+                        tmp = get_struct_variable(j, sts_vals[i]);
+                        if (!tmp) {
+                                gf_log ("", GF_LOG_ERROR,
+                                        "struct member empty.");
+                                ret = -1;
+                                goto out;
+                        }
+                        if (strlen (tmp) > spacing[j])
+                                spacing[j] = strlen (tmp);
+                }
         }
 
- out:
+out:
+        return ret;
+}
+
+int
+gf_cli_gsync_status_output (dict_t *dict, int status_detail)
+{
+        int                     gsync_count    = 0;
+        int                     i              = 0;
+        int                     j              = 0;
+        int                     ret            = 0;
+        int                     spacing[10]    = {0};
+        int                     num_of_fields  = 9;
+        char                    errmsg[1024]   = "";
+        char                   *master         = NULL;
+        char                   *slave          = NULL;
+        char                   *tmp            = NULL;
+        char                   *title_values[] = {"NODE", "MASTER", "SLAVE",
+                                                  "HEALTH", "UPTIME",
+                                                  "FILES SYNCD",
+                                                  "FILES PENDING",
+                                                  "BYTES PENDING",
+                                                  "DELETES PENDING"};
+        gf_cli_gsync_status_t **sts_vals = NULL;
+
+        /* Checks if any session is active or not */
+        ret = dict_get_int32 (dict, "gsync-count", &gsync_count);
+        if (ret) {
+                ret = dict_get_str (dict, "master", &master);
+
+                ret = dict_get_str (dict, "slave", &slave);
+
+                if (master) {
+                        if (slave)
+                                snprintf (errmsg, sizeof(errmsg), "No active "
+                                          "geo-replication sessions between %s"
+                                          " and %s", master, slave);
+                        else
+                                snprintf (errmsg, sizeof(errmsg), "No active "
+                                          "geo-replication sessions for %s",
+                                          master);
+                } else
+                        snprintf (errmsg, sizeof(errmsg), "No active "
+                                  "geo-replication sessions");
+
+                gf_log ("cli", GF_LOG_INFO, "%s", errmsg);
+                cli_out ("%s", errmsg);
+                ret = 0;
+                goto out;
+        }
+
+        for (i = 0; i < num_of_fields; i++)
+                spacing[i] = strlen(title_values[i]);
+
+        /* gsync_count = number of nodes reporting output.
+           each sts_val object will store output of each
+           node */
+        sts_vals = GF_CALLOC (gsync_count, sizeof (gf_cli_gsync_status_t *),
+                              gf_common_mt_char);
+        if (!sts_vals) {
+                ret = -1;
+                goto out;
+        }
+        for (i = 0; i < gsync_count; i++) {
+                sts_vals[i] = GF_CALLOC (1, sizeof (gf_cli_gsync_status_t),
+                                         gf_common_mt_char);
+                if (!sts_vals[i]) {
+                        ret = -1;
+                        goto out;
+                }
+        }
+
+        ret = gf_cli_read_status_data (dict, sts_vals, spacing,
+                                       gsync_count, num_of_fields);
+        if (ret) {
+                gf_log ("", GF_LOG_ERROR, "Unable to read status data");
+                goto out;
+        }
+
+        ret = gf_cli_print_status (title_values, sts_vals, spacing, gsync_count,
+                                   num_of_fields, status_detail);
+        if (ret) {
+                gf_log ("", GF_LOG_ERROR, "Unable to print status output");
+                goto out;
+        }
+
+out:
+        if (sts_vals) {
+                for (i = 0; i < gsync_count; i++) {
+                        for (j = 3; j < num_of_fields; j++) {
+                                tmp = get_struct_variable(j, sts_vals[i]);
+                                if (tmp)
+                                        GF_FREE (tmp);
+                        }
+                }
+                GF_FREE (sts_vals);
+        }
+
+        return ret;
+}
+
+static int32_t
+write_contents_to_common_pem_file (dict_t *dict, int output_count)
+{
+        char            *workdir                   = NULL;
+        char             common_pem_file[PATH_MAX] = "";
+        char            *output                    = NULL;
+        char             output_name[PATH_MAX]     = "";
+        int              bytes_writen              = 0;
+        int              fd                        = -1;
+        int              ret                       = -1;
+        int              i                         = -1;
+
+        ret = dict_get_str (dict, "glusterd_workdir", &workdir);
+        if (ret || !workdir) {
+                gf_log ("", GF_LOG_ERROR, "Unable to fetch workdir");
+                ret = -1;
+                goto out;
+        }
+
+        snprintf (common_pem_file, sizeof(common_pem_file),
+                  "%s/geo-replication/common_secret.pem.pub",
+                  workdir);
+
+        unlink (common_pem_file);
+
+        fd = open (common_pem_file, O_WRONLY | O_CREAT, 0600);
+        if (fd == -1) {
+                gf_log ("", GF_LOG_ERROR, "Failed to open %s"
+                        " Error : %s", common_pem_file,
+                        strerror (errno));
+                ret = -1;
+                goto out;
+        }
+
+        for (i = 1; i <= output_count; i++) {
+                memset (output_name, '\0', sizeof (output_name));
+                snprintf (output_name, sizeof (output_name),
+                          "output_%d", i);
+                ret = dict_get_str (dict, output_name, &output);
+                if (ret) {
+                        gf_log ("", GF_LOG_ERROR, "Failed to get %s.",
+                                output_name);
+                        cli_out ("Unable to fetch output.");
+                }
+                if (output) {
+                        bytes_writen = write (fd, output, strlen(output));
+                        if (bytes_writen != strlen(output)) {
+                                gf_log ("", GF_LOG_ERROR, "Failed to write "
+                                        "to %s", common_pem_file);
+                                ret = -1;
+                                goto out;
+                        }
+                        /* Adding the new line character */
+                        bytes_writen = write (fd, "\n", strlen("\n"));
+                        if (bytes_writen != strlen("\n")) {
+                                gf_log ("", GF_LOG_ERROR,
+                                        "Failed to add new line char");
+                                ret = -1;
+                                goto out;
+                        }
+                        output = NULL;
+                }
+        }
+
+        cli_out ("Common secret pub file present at %s", common_pem_file);
+        ret = 0;
+out:
+        if (fd)
+                close (fd);
+
+        gf_log ("", GF_LOG_DEBUG, "Returning %d", ret);
         return ret;
+}
+
+int
+gf_cli_sys_exec_cbk (struct rpc_req *req, struct iovec *iov,
+                     int count, void *myframe)
+{
+        int                     ret     = -1;
+        int                     output_count     = -1;
+        int                     i     = -1;
+        char                   *output  = NULL;
+        char                   *command = NULL;
+        char                    output_name[PATH_MAX] = "";
+        gf_cli_rsp              rsp     = {0, };
+        dict_t                  *dict   = NULL;
+        call_frame_t            *frame  = NULL;
+
+        if (req->rpc_status == -1) {
+                ret = -1;
+                goto out;
+        }
+
+        frame = myframe;
+
+        ret = xdr_to_generic (*iov, &rsp, (xdrproc_t)xdr_gf_cli_rsp);
+        if (ret < 0) {
+                gf_log (frame->this->name, GF_LOG_ERROR,
+                        "Failed to decode xdr response");
+                goto out;
+       }
+
+        dict = dict_new ();
+
+        if (!dict) {
+                ret = -1;
+                goto out;
+        }
+
+        ret = dict_unserialize (rsp.dict.dict_val, rsp.dict.dict_len, &dict);
+
+        if (ret)
+                goto out;
 
+        if (rsp.op_ret) {
+                cli_err ("%s", rsp.op_errstr ? rsp.op_errstr :
+                         "Command failed.");
+                ret = rsp.op_ret;
+                goto out;
+        }
+
+        ret = dict_get_int32 (dict, "output_count", &output_count);
+        if (ret) {
+                cli_out ("Command executed successfully.");
+                ret = 0;
+                goto out;
+        }
+
+        ret = dict_get_str (dict, "command", &command);
+        if (ret) {
+                gf_log ("", GF_LOG_ERROR,
+                        "Unable to get command from dict");
+                goto out;
+        }
+
+        if (!strcmp (command, "gsec_create")) {
+                ret = write_contents_to_common_pem_file (dict, output_count);
+                if (!ret)
+                        goto out;
+        }
+
+        for (i = 1; i <= output_count; i++) {
+                memset (output_name, '\0', sizeof (output_name));
+                snprintf (output_name, sizeof (output_name),
+                          "output_%d", i);
+                ret = dict_get_str (dict, output_name, &output);
+                if (ret) {
+                        gf_log ("", GF_LOG_ERROR, "Failed to get %s.",
+                                output_name);
+                        cli_out ("Unable to fetch output.");
+                }
+                if (output) {
+                        cli_out ("%s", output);
+                        output = NULL;
+                }
+        }
+
+        ret = 0;
+out:
+        if (dict)
+                dict_unref (dict);
+        cli_cmd_broadcast_response (ret);
+
+        free (rsp.dict.dict_val);
+
+        return ret;
+}
+
+int
+gf_cli_copy_file_cbk (struct rpc_req *req, struct iovec *iov,
+                      int count, void *myframe)
+{
+        int                     ret     = -1;
+        gf_cli_rsp              rsp     = {0, };
+        dict_t                  *dict   = NULL;
+        call_frame_t            *frame  = NULL;
+
+        if (req->rpc_status == -1) {
+                ret = -1;
+                goto out;
+        }
+
+        frame = myframe;
+
+        ret = xdr_to_generic (*iov, &rsp, (xdrproc_t)xdr_gf_cli_rsp);
+        if (ret < 0) {
+                gf_log (frame->this->name, GF_LOG_ERROR,
+                        "Failed to decode xdr response");
+                goto out;
+       }
+
+        dict = dict_new ();
+
+        if (!dict) {
+                ret = -1;
+                goto out;
+        }
+
+        ret = dict_unserialize (rsp.dict.dict_val, rsp.dict.dict_len, &dict);
+
+       if (ret)
+                goto out;
+
+        if (rsp.op_ret) {
+                cli_err ("%s", rsp.op_errstr ? rsp.op_errstr :
+                         "Copy unsuccessful");
+                ret = rsp.op_ret;
+                goto out;
+        }
+
+        cli_out ("Successfully copied file.");
+
+out:
+        if (dict)
+                dict_unref (dict);
+        cli_cmd_broadcast_response (ret);
+
+        free (rsp.dict.dict_val);
+
+        return ret;
 }
 
 int
@@ -3861,6 +4363,8 @@ gf_cli_gsync_set_cbk (struct rpc_req *req, struct iovec *iov,
         char                    *slave  = NULL;
         int32_t                  type   = 0;
         call_frame_t            *frame  = NULL;
+        gf_boolean_t             status_detail = _gf_false;
+
 
         if (req->rpc_status == -1) {
                 ret = -1;
@@ -3936,8 +4440,30 @@ gf_cli_gsync_set_cbk (struct rpc_req *req, struct iovec *iov,
                 break;
 
                 case GF_GSYNC_OPTION_TYPE_STATUS:
-                        ret = gf_cli_gsync_out_status (dict);
-                        goto out;
+                        status_detail = dict_get_str_boolean (dict,
+                                                              "status-detail",
+                                                              _gf_false);
+                        ret = gf_cli_gsync_status_output (dict, status_detail);
+                break;
+
+                case GF_GSYNC_OPTION_TYPE_DELETE:
+                        if (dict_get_str (dict, "master", &master) != 0)
+                                master = "???";
+                        if (dict_get_str (dict, "slave", &slave) != 0)
+                                slave = "???";
+                        cli_out ("Deleting " GEOREP " session between %s & %s"
+                                 " has been successful", master, slave);
+                break;
+
+                case GF_GSYNC_OPTION_TYPE_CREATE:
+                        if (dict_get_str (dict, "master", &master) != 0)
+                                master = "???";
+                        if (dict_get_str (dict, "slave", &slave) != 0)
+                                slave = "???";
+                        cli_out ("Creating " GEOREP " session between %s & %s"
+                                 " has been successful", master, slave);
+                break;
+
                 default:
                         cli_out (GEOREP" command executed successfully");
         }
@@ -3953,6 +4479,54 @@ out:
 }
 
 int32_t
+gf_cli_sys_exec (call_frame_t *frame, xlator_t *this, void *data)
+{
+        int                      ret    = 0;
+        dict_t                  *dict   = NULL;
+        gf_cli_req               req = {{0,}};
+
+        if (!frame || !this || !data) {
+                ret = -1;
+                gf_log ("cli", GF_LOG_ERROR, "Invalid data");
+                goto out;
+        }
+
+        dict = data;
+
+        ret = cli_to_glusterd (&req, frame, gf_cli_sys_exec_cbk,
+                               (xdrproc_t) xdr_gf_cli_req, dict,
+                               GLUSTER_CLI_SYS_EXEC, this, cli_rpc_prog,
+                               NULL);
+out:
+        GF_FREE (req.dict.dict_val);
+        return ret;
+}
+
+int32_t
+gf_cli_copy_file (call_frame_t *frame, xlator_t *this, void *data)
+{
+        int                      ret    = 0;
+        dict_t                  *dict   = NULL;
+        gf_cli_req               req = {{0,}};
+
+        if (!frame || !this || !data) {
+                ret = -1;
+                gf_log ("cli", GF_LOG_ERROR, "Invalid data");
+                goto out;
+        }
+
+        dict = data;
+
+        ret = cli_to_glusterd (&req, frame, gf_cli_copy_file_cbk,
+                               (xdrproc_t) xdr_gf_cli_req, dict,
+                               GLUSTER_CLI_COPY_FILE, this, cli_rpc_prog,
+                               NULL);
+out:
+        GF_FREE (req.dict.dict_val);
+        return ret;
+}
+
+int32_t
 gf_cli_gsync_set (call_frame_t *frame, xlator_t *this,
                      void *data)
 {
@@ -5636,43 +6210,57 @@ out:
         return;
 }
 
-
 static void
-cli_print_volume_tasks (dict_t *dict) {
-        int             ret = -1;
-        int             tasks = 0;
-        char            *op = 0;
-        char            *task_id_str = NULL;
-        int             status = 0;
-        char            key[1024] = {0,};
-        int             i = 0;
+cli_print_volume_status_tasks (dict_t *dict)
+{
+        int             ret         = -1;
+        int             i           = 0;
+        int             j           = 0;
+        int             count       = 0;
+        int             task_count  = 0;
+        int             status      = 0;
+        char           *op          = NULL;
+        char           *task_id_str = NULL;
+        char           *volname     = NULL;
+        char            key[1024]   = {0,};
+        char            task[1024]  = {0,};
+        char           *brick       = NULL;
+        char           *src_brick   = NULL;
+        char           *dest_brick  = NULL;
 
-        ret = dict_get_int32 (dict, "tasks", &tasks);
+        ret = dict_get_str (dict, "volname", &volname);
+        if (ret)
+                goto out;
+
+        ret = dict_get_int32 (dict, "tasks", &task_count);
         if (ret) {
-                gf_log ("cli", GF_LOG_ERROR,
-                        "Failed to get tasks count");
+                gf_log ("cli", GF_LOG_ERROR, "Failed to get tasks count");
                 return;
         }
 
-        if (tasks == 0) {
+        cli_out ("Task Status of Volume %s", volname);
+        cli_print_line (CLI_BRICK_STATUS_LINE_LEN);
+
+        if (task_count == 0) {
                 cli_out ("There are no active volume tasks");
+                cli_out (" ");
                 return;
         }
 
-        cli_out ("%15s%40s%15s", "Task", "ID", "Status");
-        cli_out ("%15s%40s%15s", "----", "--", "------");
-        for (i = 0; i < tasks; i++) {
+        for (i = 0; i < task_count; i++) {
                 memset (key, 0, sizeof (key));
                 snprintf (key, sizeof (key), "task%d.type", i);
                 ret = dict_get_str(dict, key, &op);
                 if (ret)
                         return;
+                cli_out ("%-20s : %-20s", "Task", op);
 
                 memset (key, 0, sizeof (key));
                 snprintf (key, sizeof (key), "task%d.id", i);
                 ret = dict_get_str (dict, key, &task_id_str);
                 if (ret)
                         return;
+                cli_out ("%-20s : %-20s", "ID", task_id_str);
 
                 memset (key, 0, sizeof (key));
                 snprintf (key, sizeof (key), "task%d.status", i);
@@ -5680,9 +6268,63 @@ cli_print_volume_tasks (dict_t *dict) {
                 if (ret)
                         return;
 
-                cli_out ("%15s%40s%15d", op, task_id_str, status);
+                snprintf (task, sizeof (task), "task%d", i);
+
+                /*
+                   Replace brick only has two states - In progress and Complete
+                   Ref: xlators/mgmt/glusterd/src/glusterd-replace-brick.c
+                */
+
+                if (!strcmp (op, "Replace brick")) {
+                        if (status)
+                                status = GF_DEFRAG_STATUS_COMPLETE;
+                        else
+                                status = GF_DEFRAG_STATUS_STARTED;
+
+                        memset (key, 0, sizeof (key));
+                        snprintf (key, sizeof (key), "%s.src-brick", task);
+                        ret = dict_get_str (dict, key, &src_brick);
+                        if (ret)
+                                goto out;
+
+                        cli_out ("%-20s : %-20s", "Source Brick", src_brick);
+
+                        memset (key, 0, sizeof (key));
+                        snprintf (key, sizeof (key), "%s.dst-brick", task);
+                        ret = dict_get_str (dict, key, &dest_brick);
+                        if (ret)
+                                goto out;
+
+                        cli_out ("%-20s : %-20s", "Destination Brick",
+                                 dest_brick);
+
+                } else if (!strcmp (op, "Remove brick")) {
+                        memset (key, 0, sizeof (key));
+                        snprintf (key, sizeof (key), "%s.count", task);
+                        ret = dict_get_int32 (dict, key, &count);
+                        if (ret)
+                                goto out;
+
+                        cli_out ("%-20s", "Removed bricks:");
+
+                        for (j = 1; j <= count; j++) {
+                                memset (key, 0, sizeof (key));
+                                snprintf (key, sizeof (key),"%s.brick%d",
+                                          task, j);
+                                ret = dict_get_str (dict, key, &brick);
+                                if (ret)
+                                        goto out;
+
+                                cli_out ("%-20s", brick);
+                        }
+                }
+                cli_out ("%-20s : %-20s", "Status",
+                         cli_vol_task_status_str[status]);
+                cli_out (" ");
         }
 
+out:
+        return;
 }
 
 static int
@@ -5739,10 +6381,12 @@ gf_cli_status_cbk (struct rpc_req *req, struct iovec *iov,
                                   "status information.");
 
                 if (global_state->mode & GLUSTER_MODE_XML) {
-                       cli_xml_output_str ("volStatus", msg, rsp.op_ret,
-                                           rsp.op_errno, rsp.op_errstr);
-                       ret = 0;
-                       goto out;
+                        if (!local->all)
+                                cli_xml_output_str ("volStatus", msg,
+                                                    rsp.op_ret, rsp.op_errno,
+                                                    rsp.op_errstr);
+                        ret = 0;
+                        goto out;
                 }
 
                 cli_err ("%s", msg);
@@ -5784,23 +6428,6 @@ gf_cli_status_cbk (struct rpc_req *req, struct iovec *iov,
         if ((cmd & GF_CLI_STATUS_NFS) || (cmd & GF_CLI_STATUS_SHD))
                 notbrick = _gf_true;
 
-        ret = dict_get_int32 (dict, "count", &count);
-        if (ret)
-                goto out;
-        if (count == 0) {
-                ret = -1;
-                goto out;
-        }
-
-        ret = dict_get_int32 (dict, "brick-index-max", &brick_index_max);
-        if (ret)
-                goto out;
-        ret = dict_get_int32 (dict, "other-count", &other_count);
-        if (ret)
-                goto out;
-
-        index_max = brick_index_max + other_count;
-
         if (global_state->mode & GLUSTER_MODE_XML) {
                 if (!local->all) {
                         ret = cli_xml_output_vol_status_begin (local,
@@ -5813,11 +6440,21 @@ gf_cli_status_cbk (struct rpc_req *req, struct iovec *iov,
                                 goto out;
                         }
                 }
-                ret = cli_xml_output_vol_status (local, dict);
-                if (ret) {
-                        gf_log ("cli", GF_LOG_ERROR,
-                                "Error outputting to xml");
-                        goto out;
+                if (cmd & GF_CLI_STATUS_TASKS) {
+                        ret = cli_xml_output_vol_status_tasks_detail (local,
+                                                                      dict);
+                        if (ret) {
+                                gf_log ("cli", GF_LOG_ERROR,"Error outputting "
+                                        "to xml");
+                                goto out;
+                        }
+                } else {
+                        ret = cli_xml_output_vol_status (local, dict);
+                        if (ret) {
+                                gf_log ("cli", GF_LOG_ERROR,
+                                        "Error outputting to xml");
+                                goto out;
+                        }
                 }
 
                 if (!local->all) {
@@ -5853,6 +6490,10 @@ gf_cli_status_cbk (struct rpc_req *req, struct iovec *iov,
                         cli_print_volume_status_callpool (dict, notbrick);
                         goto cont;
                         break;
+                case GF_CLI_STATUS_TASKS:
+                        cli_print_volume_status_tasks (dict);
+                        goto cont;
+                        break;
                 default:
                         break;
         }
@@ -5861,6 +6502,17 @@ gf_cli_status_cbk (struct rpc_req *req, struct iovec *iov,
         if (ret)
                 goto out;
 
+        ret = dict_get_int32 (dict, "brick-index-max", &brick_index_max);
+        if (ret)
+                goto out;
+
+        ret = dict_get_int32 (dict, "other-count", &other_count);
+        if (ret)
+                goto out;
+
+        index_max = brick_index_max + other_count;
+
+
         cli_out ("Status of volume: %s", volname);
 
         if ((cmd & GF_CLI_STATUS_DETAIL) == 0) {
@@ -5934,7 +6586,7 @@ gf_cli_status_cbk (struct rpc_req *req, struct iovec *iov,
         cli_out (" ");
 
         if ((cmd & GF_CLI_STATUS_MASK) == GF_CLI_STATUS_NONE)
-                cli_print_volume_tasks (dict);
+                cli_print_volume_status_tasks (dict);
 cont:
         ret = rsp.op_ret;
 
@@ -6010,12 +6662,6 @@ gf_cli_status_volume_all (call_frame_t *frame, xlator_t *this, void *data)
                 goto out;
         }
 
-        if (vol_count == 0) {
-                cli_err ("No volumes present");
-                ret = 0;
-                goto out;
-        }
-
         /* remove the "all" flag in cmd */
         cmd &= ~GF_CLI_STATUS_ALL;
         cmd |= GF_CLI_STATUS_VOL;
@@ -6030,6 +6676,12 @@ gf_cli_status_volume_all (call_frame_t *frame, xlator_t *this, void *data)
                 }
         }
 
+        if (vol_count == 0 && !(global_state->mode & GLUSTER_MODE_XML)) {
+                cli_err ("No volumes present");
+                ret = 0;
+                goto out;
+        }
+
         for (i = 0; i < vol_count; i++) {
 
                 dict = dict_new ();
@@ -6212,6 +6864,97 @@ gf_cli_umount (call_frame_t *frame, xlator_t *this, void *data)
 }
 
 void
+cmd_heal_volume_statistics_out (dict_t *dict, int  brick)
+{
+
+        uint64_t        num_entries = 0;
+        int             ret = 0;
+        char            key[256] = {0};
+        char            *hostname = NULL;
+        uint64_t        i = 0;
+        uint64_t        healed_count = 0;
+        uint64_t        split_brain_count = 0;
+        uint64_t        heal_failed_count = 0;
+        char            *start_time_str = NULL;
+        char            *end_time_str = NULL;
+        char            *crawl_type = NULL;
+        int             progress = -1;
+
+        snprintf (key, sizeof key, "%d-hostname", brick);
+        ret = dict_get_str (dict, key, &hostname);
+        if (ret)
+                goto out;
+        cli_out ("------------------------------------------------");
+        cli_out ("\nCrawl statistics for brick no %d", brick);
+        cli_out ("Hostname of brick %s", hostname);
+
+        snprintf (key, sizeof key, "statistics-%d-count", brick);
+        ret = dict_get_uint64 (dict, key, &num_entries);
+        if (ret)
+                goto out;
+
+        for (i = 0; i < num_entries; i++)
+        {
+                snprintf (key, sizeof key, "statistics_crawl_type-%d-%"PRIu64,
+                          brick, i);
+                ret = dict_get_str (dict, key, &crawl_type);
+                if (ret)
+                        goto out;
+
+                snprintf (key, sizeof key, "statistics_healed_cnt-%d-%"PRIu64,
+                          brick,i);
+                ret = dict_get_uint64 (dict, key, &healed_count);
+                if (ret)
+                        goto out;
+
+                snprintf (key, sizeof key, "statistics_sb_cnt-%d-%"PRIu64,
+                          brick, i);
+                ret = dict_get_uint64 (dict, key, &split_brain_count);
+                if (ret)
+                        goto out;
+                snprintf (key, sizeof key, "statistics_heal_failed_cnt-%d-%"PRIu64,
+                          brick, i);
+                ret = dict_get_uint64 (dict, key, &heal_failed_count);
+                if (ret)
+                        goto out;
+                snprintf (key, sizeof key, "statistics_strt_time-%d-%"PRIu64,
+                          brick, i);
+                ret = dict_get_str (dict, key,  &start_time_str);
+                if (ret)
+                        goto out;
+                snprintf (key, sizeof key, "statistics_end_time-%d-%"PRIu64,
+                          brick, i);
+                ret = dict_get_str (dict, key, &end_time_str);
+                if (ret)
+                        goto out;
+                snprintf (key, sizeof key, "statistics_inprogress-%d-%"PRIu64,
+                          brick, i);
+                ret = dict_get_int32 (dict, key, &progress);
+                if (ret)
+                        goto out;
+
+                cli_out ("\nStarting time of crawl: %s", start_time_str);
+                if (progress == 1)
+                        cli_out ("Crawl is in progress");
+                else
+                        cli_out ("Ending time of crawl: %s", end_time_str);
+
+                cli_out ("Type of crawl: %s", crawl_type);
+                cli_out ("No. of entries healed: %"PRIu64,
+                         healed_count);
+                cli_out ("No. of entries in split-brain: %"PRIu64,
+                        split_brain_count);
+                cli_out ("No. of heal failed entries: %"PRIu64,
+                         heal_failed_count);
+
+        }
+
+
+out:
+        return;
+}
+
+void
 cmd_heal_volume_brick_out (dict_t *dict, int brick)
 {
         uint64_t        num_entries = 0;
@@ -6277,6 +7020,53 @@ out:
         return;
 }
 
+
+void
+cmd_heal_volume_statistics_heal_count_out (dict_t *dict, int brick)
+{
+        uint64_t        num_entries = 0;
+        int             ret = 0;
+        char            key[256] = {0};
+        char           *hostname = NULL;
+        char           *path = NULL;
+        char           *status = NULL;
+        char            *shd_status = NULL;
+
+        snprintf (key, sizeof key, "%d-hostname", brick);
+        ret = dict_get_str (dict, key, &hostname);
+        if (ret)
+                goto out;
+        snprintf (key, sizeof key, "%d-path", brick);
+        ret = dict_get_str (dict, key, &path);
+        if (ret)
+                goto out;
+        cli_out ("\nBrick %s:%s", hostname, path);
+
+        snprintf (key, sizeof key, "%d-status", brick);
+        ret = dict_get_str (dict, key, &status);
+        if (status && strlen (status))
+                cli_out ("Status: %s", status);
+
+        snprintf (key, sizeof key, "%d-shd-status",brick);
+        ret = dict_get_str (dict, key, &shd_status);
+
+        if(!shd_status)
+        {
+                snprintf (key, sizeof key, "%d-hardlinks", brick);
+                ret = dict_get_uint64 (dict, key, &num_entries);
+                if (ret)
+                        cli_out ("No gathered input for this brick");
+                else
+                        cli_out ("Number of entries: %"PRIu64, num_entries);
+
+
+        }
+
+out:
+        return;
+}
+
+
 int
 gf_cli_heal_volume_cbk (struct rpc_req *req, struct iovec *iov,
                              int count, void *myframe)
@@ -6355,6 +7145,15 @@ gf_cli_heal_volume_cbk (struct rpc_req *req, struct iovec *iov,
                 case    GF_AFR_OP_SPLIT_BRAIN_FILES:
                         heal_op_str = "list of split brain entries";
                         break;
+                case    GF_AFR_OP_STATISTICS:
+                        heal_op_str =  "crawl statistics";
+                        break;
+                case    GF_AFR_OP_STATISTICS_HEAL_COUNT:
+                        heal_op_str = "count of entries to be healed";
+                        break;
+                case    GF_AFR_OP_STATISTICS_HEAL_COUNT_PER_REPLICA:
+                        heal_op_str = "count of entries to be healed per replica";
+                        break;
                 case    GF_AFR_OP_INVALID:
                         heal_op_str = "invalid heal op";
                         break;
@@ -6415,8 +7214,28 @@ gf_cli_heal_volume_cbk (struct rpc_req *req, struct iovec *iov,
                 goto out;
         }
 
-        for (i = 0; i < brick_count; i++)
-                cmd_heal_volume_brick_out (dict, i);
+        switch (heal_op) {
+                case GF_AFR_OP_STATISTICS:
+                        for (i = 0; i < brick_count; i++)
+                                cmd_heal_volume_statistics_out (dict, i);
+                        break;
+                case GF_AFR_OP_STATISTICS_HEAL_COUNT:
+                case GF_AFR_OP_STATISTICS_HEAL_COUNT_PER_REPLICA:
+                        for (i = 0; i < brick_count; i++)
+                                cmd_heal_volume_statistics_heal_count_out (dict,
+                                                                           i);
+                        break;
+                case GF_AFR_OP_INDEX_SUMMARY:
+                case GF_AFR_OP_HEALED_FILES:
+                case GF_AFR_OP_HEAL_FAILED_FILES:
+                case GF_AFR_OP_SPLIT_BRAIN_FILES:
+                        for (i = 0; i < brick_count; i++)
+                                cmd_heal_volume_brick_out (dict, i);
+                        break;
+                default:
+                        break;
+        }
+
         ret = rsp.op_ret;
 
 out:
@@ -6713,6 +7532,1024 @@ out:
         return ret;
 }
 
+int32_t
+cli_snapshot_remove_reply (gf_cli_rsp *rsp, dict_t *dict, call_frame_t *frame)
+{
+        int32_t       ret        = -1;
+        char         *snap_name  = NULL;
+
+        GF_ASSERT (rsp);
+        GF_ASSERT (dict);
+        GF_ASSERT (frame);
+
+        if (rsp->op_ret) {
+                cli_err("snapshot delete: failed: %s",
+                        rsp->op_errstr ? rsp->op_errstr :
+                        "Please check log file for details");
+                ret = rsp->op_ret;
+                goto out;
+        }
+
+        ret = dict_get_str (dict, "snapname", &snap_name);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Failed to get snapname");
+                goto out;
+        }
+
+        cli_out ("snapshot delete: %s: snap removed successfully",
+                 snap_name);
+        ret = 0;
+
+out:
+        return ret;
+}
+
+int
+cli_snapshot_config_display (dict_t *dict, gf_cli_rsp *rsp)
+{
+        char                buf[PATH_MAX]        = "";
+        char               *volname              = NULL;
+        int                 ret                  = -1;
+        int                 config_command       = 0;
+        uint64_t            value                = 0;
+        uint64_t            hard_limit           = 0;
+        uint64_t            soft_limit           = 0;
+        uint64_t            i                    = 0;
+        uint64_t            voldisplaycount      = 0;
+
+        GF_ASSERT (dict);
+        GF_ASSERT (rsp);
+
+        if (rsp->op_ret) {
+                cli_err ("Snapshot Config : failed: %s",
+                         rsp->op_errstr ? rsp->op_errstr :
+                         "Please check log file for details");
+                ret = rsp->op_ret;
+                goto out;
+        }
+
+        ret = dict_get_int32 (dict, "config-command", &config_command);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Could not fetch config type");
+                goto out;
+        }
+
+        ret = dict_get_str (dict, "volname", &volname);
+        /* Ignore the error, as volname is optional */
+
+        if (!volname) {
+                volname = "System";
+        }
+
+        ret = dict_get_uint64 (dict, "snap-max-hard-limit", &hard_limit);
+        /* Ignore the error, as the key specified is optional */
+        ret = dict_get_uint64 (dict, "snap-max-soft-limit", &soft_limit);
+
+        if (!hard_limit && !soft_limit
+                        && config_command != GF_SNAP_CONFIG_DISPLAY) {
+                ret = -1;
+                gf_log(THIS->name, GF_LOG_ERROR,
+                       "Could not fetch config-key");
+                goto out;
+        }
+
+        switch (config_command) {
+        case GF_SNAP_CONFIG_TYPE_SET:
+                if (hard_limit && soft_limit) {
+                        cli_out ("snapshot config: snap-max-hard-limit "
+                                "& snap-max-soft-limit for system set "
+                                 "successfully");
+                } else if (hard_limit){
+                        cli_out ("snapshot config: %s "
+                                 "for snap-max-hard-limit set successfully",
+                                 volname);
+                } else if (soft_limit) {
+                        cli_out ("snapshot config: %s "
+                                 "for snap-max-soft-limit set successfully",
+                                 volname);
+                }
+                break;
+
+        case GF_SNAP_CONFIG_DISPLAY :
+                cli_out ("\nSnapshot System Configuration:");
+                ret = dict_get_uint64 (dict, "snap-max-hard-limit",
+                                       &value);
+                if (ret) {
+                         gf_log ("cli", GF_LOG_ERROR, "Could not fetch "
+                                "snap_max_hard_limit for %s", volname);
+                         ret = -1;
+                         goto out;
+                }
+                cli_out ("snap-max-hard-limit : %"PRIu64, value);
+
+                ret = dict_get_uint64 (dict, "snap-max-soft-limit",
+                                       &soft_limit);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR, "Could not fetch "
+                               "snap-max-soft-limit for %s", volname);
+                        ret = -1;
+                        goto out;
+                }
+                cli_out ("snap-max-soft-limit : %"PRIu64"%%\n",
+                         soft_limit);
+
+                cli_out ("Snapshot Volume Configuration:");
+
+                ret = dict_get_uint64 (dict, "voldisplaycount",
+                                       &voldisplaycount);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR,
+                               "Could not fetch voldisplaycount");
+                        ret = -1;
+                        goto out;
+                }
+
+                for (i = 0; i < voldisplaycount; i++) {
+                        snprintf (buf, sizeof(buf), "volume%ld-volname", i);
+                        ret = dict_get_str (dict, buf, &volname);
+                        if (ret) {
+                                gf_log ("cli", GF_LOG_ERROR, "Could not fetch "
+                                       " %s", buf);
+                                ret = -1;
+                                goto out;
+                        }
+                        cli_out ("\nVolume : %s", volname);
+
+                        snprintf (buf, sizeof(buf),
+                                  "volume%ld-snap-max-hard-limit", i);
+                        ret = dict_get_uint64 (dict, buf, &value);
+                        if (ret) {
+                                gf_log ("cli", GF_LOG_ERROR, "Could not fetch "
+                                       " %s", buf);
+                                ret = -1;
+                                goto out;
+                        }
+                        cli_out ("snap-max-hard-limit : %"PRIu64, value);
+
+                        snprintf (buf, sizeof(buf),
+                                  "volume%ld-active-hard-limit", i);
+                        ret = dict_get_uint64 (dict, buf, &value);
+                        if (ret) {
+                                 gf_log ("cli", GF_LOG_ERROR, "Could not fetch"
+                                        " effective snap_max_hard_limit for "
+                                        "%s", volname);
+                                ret = -1;
+                                goto out;
+                        }
+                        cli_out ("Effective snap-max-hard-limit : %"PRIu64,
+                                 value);
+
+                        snprintf (buf, sizeof(buf),
+                                      "volume%ld-snap-max-soft-limit", i);
+                        ret = dict_get_uint64 (dict, buf, &value);
+                        if (ret) {
+                                gf_log ("cli", GF_LOG_ERROR, "Could not fetch "
+                                       " %s", buf);
+                                ret = -1;
+                                goto out;
+                        }
+                        cli_out ("Effective snap-max-soft-limit : %"PRIu64" "
+                                 "(%"PRIu64"%%)", value, soft_limit);
+                }
+                break;
+        default :
+                break;
+        }
+
+        ret = 0;
+out:
+        return ret;
+}
+
+/* This function is used to print the volume related information
+ * of a snap.
+ *
+ * arg - 0, dict       : Response Dictionary.
+ * arg - 1, prefix str : snaplist.snap{0..}.vol{0..}.*
+ */
+int
+cli_get_each_volinfo_in_snap (dict_t *dict, char *keyprefix,
+                              gf_boolean_t snap_driven) {
+        char             key[PATH_MAX]  =  "";
+        char            *get_buffer     =  NULL;
+        int              value          =  0;
+        int              ret            = -1;
+        char             indent[5]      = "\t";
+        char            *volname        =  NULL;
+
+        GF_ASSERT (dict);
+        GF_ASSERT (keyprefix);
+
+        if (snap_driven) {
+                ret = snprintf (key, sizeof (key), "%s.volname", keyprefix);
+                if (ret < 0) {
+                        goto out;
+                }
+
+                ret = dict_get_str (dict, key, &get_buffer);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR, "Failed to get %s", key);
+                        goto out;
+                }
+                cli_out ("%s" INDENT_MAIN_HEAD "%s", indent,
+                        "Snap Volume Name", ":", get_buffer);
+
+                ret = snprintf (key, sizeof (key),
+                                "%s.origin-volname", keyprefix);
+                if (ret < 0) {
+                        goto out;
+                }
+
+                ret = dict_get_str (dict, key, &volname);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_WARNING, "Failed to get %s", key);
+                        cli_out ("%-12s", "Origin:");
+                }
+                cli_out ("%s" INDENT_MAIN_HEAD "%s", indent,
+                        "Origin Volume name", ":", volname);
+
+
+                ret = snprintf (key, sizeof (key), "%s.snapcount",
+                                keyprefix);
+                if (ret < 0) {
+                        goto out;
+                }
+
+                ret = dict_get_int32 (dict, key, &value);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR, "Failed to get %s", key);
+                        goto out;
+                }
+                cli_out ("%s%s %s      %s %d", indent, "Snaps taken for",
+                        volname, ":", value);
+
+                ret = snprintf (key, sizeof (key), "%s.snaps-available",
+                                keyprefix);
+                if (ret < 0) {
+                        goto out;
+                }
+
+                ret = dict_get_int32 (dict, key, &value);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR, "Failed to get %s", key);
+                        goto out;
+                }
+                cli_out ("%s%s %s  %s %d", indent, "Snaps available for",
+                         volname, ":", value);
+        }
+
+
+        ret = snprintf (key, sizeof (key), "%s.vol-status", keyprefix);
+        if (ret < 0) {
+                goto out;
+        }
+
+        ret = dict_get_str (dict, key, &get_buffer);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Failed to get %s", key);
+                goto out;
+        }
+        cli_out ("%s" INDENT_MAIN_HEAD "%s", indent, "Status",
+                 ":", get_buffer);
+out :
+        return ret;
+}
+
+/* This function is used to print snap related information
+ * arg - 0, dict       : Response dictionary.
+ * arg - 1, prefix_str : snaplist.snap{0..}.*
+ */
+int
+cli_get_volinfo_in_snap (dict_t *dict, char *keyprefix) {
+
+        char            key[PATH_MAX]   = "";
+        int             i               = 0;
+        int             volcount        = 0;
+        int             ret             = -1;
+
+        GF_ASSERT (dict);
+        GF_ASSERT (keyprefix);
+
+        ret = snprintf (key, sizeof (key), "%s.vol-count", keyprefix);
+        if (ret < 0) {
+                goto out;
+        }
+
+        ret = dict_get_int32 (dict, key, &volcount);
+        for (i = 1 ; i <= volcount ; i++) {
+                ret = snprintf (key, sizeof (key),
+                                "%s.vol%d", keyprefix, i);
+                if (ret < 0) {
+                        goto out;
+                }
+                ret = cli_get_each_volinfo_in_snap (dict, key, _gf_true);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR, "Could not list "
+                                "details of volume in a snap");
+                        goto out;
+                }
+                cli_out (" ");
+        }
+
+out :
+        return ret;
+}
+
+int
+cli_get_each_snap_info (dict_t *dict, char *prefix_str,
+                        gf_boolean_t snap_driven) {
+        char            key_buffer[PATH_MAX] = "";
+        char           *get_buffer           = NULL;
+        int             ret                  = -1;
+        char            indent[5]            = "";
+
+        GF_ASSERT (dict);
+        GF_ASSERT (prefix_str);
+
+        if (!snap_driven)
+                strcat (indent, "\t");
+
+        ret = snprintf (key_buffer, sizeof (key_buffer), "%s.snapname",
+                        prefix_str);
+        if (ret < 0 ) {
+                goto out;
+        }
+
+        ret = dict_get_str (dict, key_buffer, &get_buffer);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Unable to fetch snapname %s ",
+                        key_buffer);
+                goto out;
+        }
+        cli_out ("%s" INDENT_MAIN_HEAD "%s", indent, "Snapshot",
+                ":", get_buffer);
+
+        ret = snprintf (key_buffer, sizeof (key_buffer), "%s.snap-id",
+                        prefix_str);
+        if (ret < 0 ) {
+                goto out;
+        }
+
+        ret = dict_get_str (dict, key_buffer, &get_buffer);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Unable to fetch snap-id %s ",
+                        key_buffer);
+                goto out;
+        }
+        cli_out ("%s" INDENT_MAIN_HEAD "%s", indent, "Snap UUID",
+                ":", get_buffer);
+
+        ret = snprintf (key_buffer, sizeof (key_buffer), "%s.snap-desc",
+                        prefix_str);
+        if (ret < 0 ) {
+                goto out;
+        }
+
+        ret = dict_get_str (dict, key_buffer, &get_buffer);
+        if (!ret) {
+                /* Ignore error for description */
+                cli_out ("%s" INDENT_MAIN_HEAD "%s", indent,
+                         "Description", ":", get_buffer);
+        }
+
+        ret = snprintf (key_buffer, sizeof (key_buffer), "%s.snap-time",
+                        prefix_str);
+        if (ret < 0 ) {
+                goto out;
+        }
+
+        ret = dict_get_str (dict, key_buffer, &get_buffer);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Unable to fetch snap-time %s ",
+                        prefix_str);
+                goto out;
+        }
+        cli_out ("%s" INDENT_MAIN_HEAD "%s", indent, "Created",
+                ":", get_buffer);
+
+        if (snap_driven) {
+                cli_out ("%-12s", "Snap Volumes:\n");
+                ret = cli_get_volinfo_in_snap (dict, prefix_str);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR, "Unable to list details "
+                                "of the snaps");
+                        goto out;
+                }
+        }
+out :
+        return ret;
+}
+
+/* This is a generic function to print snap related information.
+ * arg - 0, dict : Response Dictionary
+ */
+int
+cli_call_snapshot_info (dict_t *dict, gf_boolean_t bool_snap_driven) {
+        int             snap_count       =  0;
+        char            key[PATH_MAX]    =  "";
+        int             ret              =  -1;
+        int             i                =   0;
+
+        GF_ASSERT (dict);
+
+        ret = dict_get_int32 (dict, "snap-count", &snap_count);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Unable to get snap-count");
+                goto out;
+        }
+
+        if (snap_count == 0) {
+                cli_out ("No snapshots present");
+        }
+
+        for (i = 1 ; i <= snap_count ; i++) {
+                ret = snprintf (key, sizeof (key), "snap%d", i);
+                if (ret < 0) {
+                        goto out;
+                }
+                ret = cli_get_each_snap_info (dict, key, bool_snap_driven);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR,
+                                "Unable to print snap details");
+                        goto out;
+                }
+        }
+out :
+        return ret;
+}
+
+int
+cli_get_snaps_in_volume (dict_t *dict) {
+        int      ret             =  -1;
+        int      i               =  0;
+        int      count           =  0;
+        int      avail           =  0;
+        char     key[PATH_MAX]   =  "";
+        char    *get_buffer      =  NULL;
+
+        GF_ASSERT (dict);
+
+        ret = dict_get_str (dict, "origin-volname", &get_buffer);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Could not fetch origin-volname");
+                goto out;
+        }
+        cli_out (INDENT_MAIN_HEAD "%s", "Volume Name", ":", get_buffer);
+
+        ret = dict_get_int32 (dict, "snap-count", &avail);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Could not fetch snap-count");
+                goto out;
+        }
+        cli_out (INDENT_MAIN_HEAD "%d", "Snaps Taken", ":", avail);
+
+        ret = dict_get_int32 (dict, "snaps-available", &count);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Could not fetch snaps-available");
+                goto out;
+        }
+        cli_out (INDENT_MAIN_HEAD "%d", "Snaps Available", ":", count);
+
+        for (i = 1 ; i <= avail ; i++) {
+                snprintf (key, sizeof (key), "snap%d", i);
+                ret = cli_get_each_snap_info (dict, key, _gf_false);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR,
+                                "Unable to print snap details");
+                        goto out;
+                }
+
+                ret = snprintf (key, sizeof (key), "snap%d.vol1", i);
+                if (ret < 0) {
+                        goto out;
+                }
+                ret = cli_get_each_volinfo_in_snap (dict, key, _gf_false);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR, "Could not get volume "
+                                "related information");
+                        goto out;
+                }
+
+                cli_out (" ");
+        }
+out :
+        return ret;
+}
+
+int
+cli_snapshot_list (dict_t *dict) {
+        int     snapcount       =       0;
+        char    key[PATH_MAX]   =       "";
+        int     ret             =       -1;
+        int     i               =       0;
+        char    *get_buffer     =       NULL;
+
+        GF_ASSERT (dict);
+
+        ret = dict_get_int32 (dict, "snap-count", &snapcount);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Could not fetch snap count");
+                goto out;
+        }
+
+        if (snapcount == 0) {
+                cli_out ("No snapshots present");
+        }
+
+        for (i = 1 ; i <= snapcount ; i++) {
+                ret = snprintf (key, sizeof (key), "snapname%d",i);
+                if (ret < 0) {
+                        goto out;
+                }
+
+                ret = dict_get_str (dict, key, &get_buffer);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR, "Could not get %s ", key);
+                        goto out;
+                } else {
+                        cli_out ("%s", get_buffer);
+                }
+        }
+out :
+        return ret;
+}
+
+int
+cli_get_snap_volume_status (dict_t *dict, char *key_prefix)
+{
+        int             ret             =       -1;
+        char            key[PATH_MAX]   =       "";
+        char            *buffer         =       NULL;
+        int             brickcount      =       0;
+        int             i               =       0;
+        int             pid             =       0;
+
+        GF_ASSERT (dict);
+        GF_ASSERT (key_prefix);
+
+        ret = snprintf (key, sizeof (key), "%s.brickcount", key_prefix);
+        if (ret < 0) {
+                goto out;
+        }
+        ret = dict_get_int32 (dict, key, &brickcount);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Failed to fetch brickcount");
+                goto out;
+        }
+
+        for ( i = 0 ; i < brickcount ; i++ ) {
+                ret = snprintf (key, sizeof (key), "%s.brick%d.path",
+                                key_prefix, i);
+                if (ret < 0) {
+                        goto out;
+                }
+
+                ret = dict_get_str (dict, key, &buffer);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_INFO,
+                                "Unable to get Brick Path");
+                        continue;
+                }
+                cli_out ("\n\t%-17s %s   %s", "Brick Path", ":", buffer);
+
+                ret = snprintf (key, sizeof (key), "%s.brick%d.vgname",
+                                key_prefix, i);
+                if (ret < 0) {
+                        goto out;
+                }
+
+                ret = dict_get_str (dict, key, &buffer);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_INFO,
+                                "Unable to get Volume Group");
+                        cli_out ("\t%-17s %s   %s", "Volume Group", ":", "N/A");
+                } else
+                        cli_out ("\t%-17s %s   %s", "Volume Group", ":", buffer);
+
+                ret = snprintf (key, sizeof (key), "%s.brick%d.status",
+                                key_prefix, i);
+                if (ret < 0) {
+                        goto out;
+                }
+
+                ret = dict_get_str (dict, key, &buffer);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_INFO,
+                                "Unable to get Brick Running");
+                        cli_out ("\t%-17s %s   %s", "Brick Running", ":", "N/A");
+                } else
+                        cli_out ("\t%-17s %s   %s", "Brick Running", ":", buffer);
+
+                ret = snprintf (key, sizeof (key), "%s.brick%d.pid",
+                                key_prefix, i);
+                if (ret < 0) {
+                        goto out;
+                }
+
+                ret = dict_get_int32 (dict, key, &pid);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_INFO,
+                                "Unable to get pid");
+                        cli_out ("\t%-17s %s   %s", "Brick PID", ":", "N/A");
+                } else
+                        cli_out ("\t%-17s %s   %d", "Brick PID", ":", pid);
+
+                ret = snprintf (key, sizeof (key), "%s.brick%d.data",
+                                key_prefix, i);
+                if (ret < 0) {
+                        goto out;
+                }
+
+                ret = dict_get_str (dict, key, &buffer);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_INFO,
+                                "Unable to get Data Percent");
+                        cli_out ("\t%-17s %s   %s", "Data Percentage", ":", "N/A");
+                } else
+                        cli_out ("\t%-17s %s   %s", "Data Percentage", ":", buffer);
+
+                ret = snprintf (key, sizeof (key), "%s.brick%d.lvsize",
+                                key_prefix, i);
+                if (ret < 0) {
+                        goto out;
+                }
+                ret = dict_get_str (dict, key, &buffer);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_INFO, "Unable to get LV Size");
+                        cli_out ("\t%-17s %s   %s", "LV Size", ":", "N/A");
+                } else
+                        cli_out ("\t%-17s %s   %s", "LV Size", ":", buffer);
+
+        }
+out :
+        return ret;
+}
+
+
+
+int
+cli_get_single_snap_status (dict_t *dict, char *keyprefix)
+{
+        int             ret             =       -1;
+        char            key[PATH_MAX]   =       "";
+        int             i               =       0;
+        int             volcount        =       0;
+        char            *get_buffer     =       NULL;
+
+        GF_ASSERT (dict);
+        GF_ASSERT (keyprefix);
+
+        ret = snprintf (key, sizeof (key), "%s.snapname", keyprefix);
+        if (ret < 0) {
+                goto out;
+        }
+
+        ret = dict_get_str (dict, key, &get_buffer);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Unable to get snapname");
+                goto out;
+        }
+        cli_out ("\nSnap Name : %s", get_buffer);
+
+        ret = snprintf (key, sizeof (key), "%s.uuid", keyprefix);
+        if (ret < 0) {
+                goto out;
+        }
+
+        ret = dict_get_str (dict, key, &get_buffer);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Unable to get snap UUID");
+                goto out;
+        }
+        cli_out ("Snap UUID : %s", get_buffer);
+
+        ret = snprintf (key, sizeof (key), "%s.volcount", keyprefix);
+        if (ret < 0) {
+                goto out;
+        }
+
+        ret = dict_get_int32 (dict, key, &volcount);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Unable to get volume count");
+                goto out;
+        }
+
+        for (i = 0 ; i < volcount ; i++) {
+                ret = snprintf (key, sizeof (key), "%s.vol%d", keyprefix, i);
+                if (ret < 0) {
+                        goto out;
+                }
+
+                ret = cli_get_snap_volume_status (dict, key);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR,
+                                "Could not get snap volume status");
+                        goto out;
+                }
+        }
+out :
+        return ret;
+}
+
+int
+cli_snap_status_all (dict_t *dict) {
+        int             ret             =       -1;
+        char            key[PATH_MAX]   =       "";
+        int             snapcount       =       0;
+        int             i               =       0;
+
+        GF_ASSERT (dict);
+
+        ret = dict_get_int32 (dict, "status.snapcount", &snapcount);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Could not get snapcount");
+                goto out;
+        }
+
+        if (snapcount == 0) {
+                cli_out ("No snapshots present");
+        }
+
+        for (i = 0 ; i < snapcount; i++) {
+                ret = snprintf (key, sizeof (key), "status.snap%d",i);
+                if (ret < 0) {
+                        goto out;
+                }
+                ret = cli_get_single_snap_status (dict, key);
+        }
+out:
+        return ret;
+}
+
+
+int
+cli_snapshot_status_display (dict_t *dict, gf_cli_rsp *rsp)
+{
+        char            key[PATH_MAX]   =       "";
+        int             ret             =       -1;
+        int             status_cmd      =       -1;
+
+        GF_ASSERT (dict);
+        GF_ASSERT (rsp);
+
+        if (rsp->op_ret) {
+                cli_err ("Snapshot Status : failed: %s",
+                        rsp->op_errstr ? rsp->op_errstr :
+                        "Please check log file for details");
+                ret = rsp->op_ret;
+                goto out;
+        }
+
+        ret = dict_get_int32 (dict, "cmd", &status_cmd);
+        if (ret) {
+                gf_log ("cli", GF_LOG_ERROR, "Could not fetch status type");
+                goto out;
+        }
+        switch (status_cmd) {
+                case GF_SNAP_STATUS_TYPE_ALL :
+                {
+                        ret = cli_snap_status_all (dict);
+                        if (ret) {
+                                gf_log ("cli", GF_LOG_ERROR, "Could not fetch "
+                                        "status of all snap");
+                                goto out;
+                        }
+                break;
+                }
+
+                case GF_SNAP_STATUS_TYPE_SNAP :
+                {
+                        ret = snprintf (key, sizeof (key), "status.snap0");
+                        if (ret < 0) {
+                                goto out;
+                        }
+                        ret = cli_get_single_snap_status (dict, key);
+                        if (ret) {
+                                gf_log ("cli", GF_LOG_ERROR, "Could not fetch "
+                                        "status of snap");
+                                goto out;
+                        }
+                break;
+                }
+
+                case GF_SNAP_STATUS_TYPE_VOL :
+                {
+                        ret = cli_snap_status_all (dict);
+                        if (ret) {
+                                gf_log ("cli", GF_LOG_ERROR, "Could not fetch "
+                                        "status of snap in a volume");
+                                goto out;
+                        }
+                        break;
+                }
+                default :
+                        break;
+        }
+out :
+        return ret;
+}
+
+int
+gf_cli_snapshot_cbk (struct rpc_req *req, struct iovec *iov,
+                     int count, void *myframe)
+{
+        int                   ret                      = -1;
+        gf_cli_rsp            rsp                      = {0, };
+        dict_t               *dict                     = NULL;
+        char                 *snap_name                = NULL;
+        int32_t               type                     =  0;
+        call_frame_t         *frame                    = NULL;
+        gf_boolean_t         snap_driven               = _gf_false;
+
+        if (req->rpc_status == -1) {
+                ret = -1;
+                goto out;
+        }
+
+        frame = myframe;
+
+        ret = xdr_to_generic (*iov, &rsp, (xdrproc_t)xdr_gf_cli_rsp);
+        if (ret < 0) {
+                gf_log (frame->this->name, GF_LOG_ERROR,
+                        "Failed to decode xdr response");
+                goto out;
+        }
+
+        dict = dict_new ();
+
+        if (!dict) {
+                ret = -1;
+                goto out;
+        }
+
+        ret = dict_unserialize (rsp.dict.dict_val, rsp.dict.dict_len, &dict);
+
+        if (ret)
+                goto out;
+
+        ret = dict_get_int32 (dict, "type", &type);
+        if (ret) {
+                gf_log (frame->this->name, GF_LOG_ERROR, "failed to get type");
+                goto out;
+        }
+
+        switch (type) {
+        case GF_SNAP_OPTION_TYPE_CREATE:
+                if (rsp.op_ret) {
+                        cli_err("snapshot create: failed: %s",
+                                 rsp.op_errstr ? rsp.op_errstr :
+                                 "Please check log file for details");
+                                 ret = rsp.op_ret;
+                                 goto out;
+                }
+
+                ret = dict_get_str (dict, "snapname", &snap_name);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR,
+                                "Failed to get snap name");
+                        goto out;
+                }
+                cli_out ("snapshot create: %s: snap created successfully",
+                         snap_name);
+                break;
+
+        case GF_SNAP_OPTION_TYPE_RESTORE:
+                /* TODO: Check if rsp.op_ret needs to be checked here. Or is
+                 * it ok to check this in the start of the function where we
+                 * get rsp.*/
+                if (rsp.op_ret) {
+                        cli_err("snapshot restore: failed: %s",
+                                 rsp.op_errstr ? rsp.op_errstr :
+                                 "Please check log file for details");
+                                 ret = rsp.op_ret;
+                                 goto out;
+                }
+
+                ret = dict_get_str (dict, "snapname", &snap_name);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR,
+                                "Failed to get snap name");
+                        goto out;
+                }
+
+                cli_out ("Snapshot restore: %s: Snap restored "
+                                 "successfully", snap_name);
+
+                ret = 0;
+                break;
+
+        case GF_SNAP_OPTION_TYPE_INFO:
+                if (rsp.op_ret) {
+                        cli_err ("Snapshot info : failed: %s",
+                                  rsp.op_errstr ? rsp.op_errstr :
+                                  "Please check log file for details");
+                        ret = rsp.op_ret;
+                        goto out;
+                }
+
+                snap_driven = dict_get_str_boolean (dict, "snap-driven",
+                                                    _gf_false);
+                if (snap_driven == _gf_true) {
+                        ret = cli_call_snapshot_info  (dict, snap_driven);
+                        if (ret) {
+                                gf_log ("cli", GF_LOG_ERROR,
+                                        "Snapshot info failed");
+                                goto out;
+                        }
+                } else if (snap_driven == _gf_false) {
+                        ret = cli_get_snaps_in_volume (dict);
+                        if (ret) {
+                                gf_log ("cli", GF_LOG_ERROR,
+                                        "Snapshot info failed");
+                                goto out;
+                        }
+                }
+                break;
+
+        case GF_SNAP_OPTION_TYPE_CONFIG:
+                ret = cli_snapshot_config_display (dict, &rsp);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR, "Failed to display "
+                                "snapshot config output.");
+                        goto out;
+                }
+                break;
+
+        case GF_SNAP_OPTION_TYPE_LIST:
+                if (rsp.op_ret) {
+                        cli_err ("Snapshot list : failed: %s",
+                                 rsp.op_errstr ? rsp.op_errstr :
+                                "Please check log file for details");
+                        ret = rsp.op_ret;
+                        goto out;
+                }
+
+                ret = cli_snapshot_list (dict);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR, "Failed to display "
+                                "snapshot list");
+                        goto out;
+                }
+                break;
+
+        case GF_SNAP_OPTION_TYPE_DELETE:
+                ret = cli_snapshot_remove_reply (&rsp, dict, frame);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR,
+                                "Failed to delete snap");
+                        goto out;
+                }
+                break;
+
+        case GF_SNAP_OPTION_TYPE_STATUS:
+                ret = cli_snapshot_status_display (dict, &rsp);
+                if (ret) {
+                        gf_log ("cli", GF_LOG_ERROR, "Failed to display "
+                                "snapshot status output.");
+                        goto out;
+                }
+                break;
+
+        default:
+                cli_err ("Unknown command executed");
+                ret = -1;
+                goto out;
+        }
+out:
+        if (dict)
+                dict_unref (dict);
+        cli_cmd_broadcast_response (ret);
+
+        free (rsp.dict.dict_val);
+        free (rsp.op_errstr);
+
+        return ret;
+}
+
+int32_t
+gf_cli_snapshot (call_frame_t *frame, xlator_t *this,
+                 void *data)
+{
+        gf_cli_req        req     = {{0,}};
+        dict_t           *options = NULL;
+        int               ret     = -1;
+
+        if (!frame || !this || !data)
+                goto out;
+
+        options = data;
+
+        ret = cli_to_glusterd (&req, frame, gf_cli_snapshot_cbk,
+                               (xdrproc_t) xdr_gf_cli_req, options,
+                               GLUSTER_CLI_SNAP, this, cli_rpc_prog,
+                               NULL);
+out:
+        gf_log ("cli", GF_LOG_DEBUG, "Returning %d", ret);
+
+        GF_FREE (req.dict.dict_val);
+        return ret;
+}
+
 int
 cli_to_glusterd (gf_cli_req *req, call_frame_t *frame,
                  fop_cbk_fn_t cbkfn, xdrproc_t xdrproc, dict_t *dict,
@@ -6822,9 +8659,9 @@ struct rpc_clnt_procedure gluster_cli_actors[GLUSTER_CLI_MAXVALUE] = {
         [GLUSTER_CLI_STATEDUMP_VOLUME] = {"STATEDUMP_VOLUME", gf_cli_statedump_volume},
         [GLUSTER_CLI_LIST_VOLUME]      = {"LIST_VOLUME", gf_cli_list_volume},
         [GLUSTER_CLI_CLRLOCKS_VOLUME]  = {"CLEARLOCKS_VOLUME", gf_cli_clearlocks_volume},
-#ifdef HAVE_BD_XLATOR
-        [GLUSTER_CLI_BD_OP]            = {"BD_OP", gf_cli_bd_op},
-#endif
+        [GLUSTER_CLI_COPY_FILE]        = {"COPY_FILE", gf_cli_copy_file},
+        [GLUSTER_CLI_SYS_EXEC]         = {"SYS_EXEC", gf_cli_sys_exec},
+        [GLUSTER_CLI_SNAP]             = {"SNAP", gf_cli_snapshot},
 };
 
 struct rpc_clnt_program cli_prog = {
diff --git a/cli/src/cli-xml-output.c b/cli/src/cli-xml-output.c
index a2bd2b141..d8884d44b 100644
--- a/cli/src/cli-xml-output.c
+++ b/cli/src/cli-xml-output.c
@@ -14,6 +14,12 @@
 #include "compat.h"
 #include "syscall.h"
 
+
+enum gf_task_types {
+    GF_TASK_TYPE_REBALANCE,
+    GF_TASK_TYPE_REMOVE_BRICK
+};
+
 /*
  * IMPORTANT NOTE:
  * All exported functions in this file which use libxml need use a
@@ -1383,6 +1389,86 @@ out:
 
 #if (HAVE_LIB_XML)
 int
+cli_xml_output_remove_brick_task_params (xmlTextWriterPtr writer, dict_t *dict,
+                                         char *prefix)
+{
+        int             ret = -1;
+        char            key[1024] = {0,};
+        int             count = 0;
+        int             i = 0;
+        char            *brick = NULL;
+
+        /* <params> */
+        ret = xmlTextWriterStartElement (writer, (xmlChar *)"params");
+        XML_RET_CHECK_AND_GOTO (ret, out);
+
+        snprintf (key, sizeof (key), "%s.count", prefix);
+        ret = dict_get_int32 (dict, key, &count);
+        if (ret)
+                goto out;
+
+        for (i = 1; i <= count; i++) {
+                memset (key, 0, sizeof (key));
+                snprintf (key, sizeof (key), "%s.brick%d", prefix, i);
+                ret = dict_get_str (dict, key, &brick);
+                if (ret)
+                        goto out;
+                ret = xmlTextWriterWriteFormatElement (writer,
+                                                       (xmlChar *)"brick",
+                                                       "%s", brick);
+                XML_RET_CHECK_AND_GOTO (ret, out);
+                brick = NULL;
+        }
+
+        /* </param> */
+        ret = xmlTextWriterEndElement (writer);
+
+out:
+        gf_log ("cli", GF_LOG_DEBUG, "Returning %d", ret);
+        return ret;
+}
+
+int
+cli_xml_output_replace_brick_task_params (xmlTextWriterPtr writer, dict_t *dict,
+                                          char *prefix)
+{
+
+        int             ret = -1;
+        char            key[1024] = {0,};
+        char            *brick = NULL;
+
+        /* <params> */
+        ret = xmlTextWriterStartElement (writer, (xmlChar *)"params");
+        XML_RET_CHECK_AND_GOTO (ret, out);
+
+        snprintf (key, sizeof (key), "%s.src-brick", prefix);
+        ret = dict_get_str (dict, key, &brick);
+        if (ret)
+                goto out;
+        ret = xmlTextWriterWriteFormatElement (writer, (xmlChar *)"srcBrick",
+                                               "%s", brick);
+        XML_RET_CHECK_AND_GOTO (ret, out);
+
+
+        memset (key, 0, sizeof (key));
+        snprintf (key, sizeof (key), "%s.dst-brick", prefix);
+        ret = dict_get_str (dict, key, &brick);
+        if (ret)
+                goto out;
+        ret = xmlTextWriterWriteFormatElement (writer, (xmlChar *)"dstBrick",
+                                               "%s", brick);
+        XML_RET_CHECK_AND_GOTO (ret, out);
+
+
+        /* </param> */
+        ret = xmlTextWriterEndElement (writer);
+
+out:
+        gf_log ("cli", GF_LOG_DEBUG, "Returning %d", ret);
+        return ret;
+}
+
+int
 cli_xml_output_vol_status_tasks (cli_local_t *local, dict_t *dict) {
         int                     ret = -1;
         char                    *task_type = NULL;
@@ -1436,6 +1522,36 @@ cli_xml_output_vol_status_tasks (cli_local_t *local, dict_t *dict) {
                                                        "%d", status);
                 XML_RET_CHECK_AND_GOTO (ret, out);
 
+                if (!strcmp (task_type, "Replace brick")) {
+                    if (status) {
+                        status = GF_DEFRAG_STATUS_COMPLETE;
+                    } else {
+                        status = GF_DEFRAG_STATUS_STARTED;
+                    }
+                }
+
+                ret = xmlTextWriterWriteFormatElement (local->writer,
+                                                       (xmlChar *)"statusStr",
+                                                       "%s",
+                                             cli_vol_task_status_str[status]);
+
+                XML_RET_CHECK_AND_GOTO (ret, out);
+
+                memset (key, 0, sizeof (key));
+                snprintf (key, sizeof (key), "task%d", i);
+                if (!strcmp (task_type, "Replace brick")) {
+                        ret = cli_xml_output_replace_brick_task_params
+                                (local->writer, dict, key);
+                        if (ret)
+                                goto out;
+                } else if (!strcmp (task_type, "Remove brick")) {
+                        ret = cli_xml_output_remove_brick_task_params
+                                (local->writer, dict, key);
+                        if (ret)
+                                goto out;
+                }
+
+
                 /* </task> */
                 ret = xmlTextWriterEndElement (local->writer);
                 XML_RET_CHECK_AND_GOTO (ret, out);
@@ -1444,11 +1560,40 @@ cli_xml_output_vol_status_tasks (cli_local_t *local, dict_t *dict) {
         /* </tasks> */
         ret = xmlTextWriterEndElement (local->writer);
 
-
 out:
         gf_log ("cli", GF_LOG_DEBUG, "Returning %d", ret);
         return ret;
 }
+
+int
+cli_xml_output_vol_status_tasks_detail (cli_local_t *local, dict_t *dict)
+{
+        int    ret     = -1;
+        char  *volname = NULL;
+
+        /*<volume>*/
+        ret = xmlTextWriterStartElement (local->writer, (xmlChar *)"volume");
+        XML_RET_CHECK_AND_GOTO (ret, out);
+
+        ret = dict_get_str (dict, "volname", &volname);
+        if (ret)
+                goto out;
+        ret = xmlTextWriterWriteFormatElement (local->writer,
+                                               (xmlChar *)"volName", "%s",
+                                               volname);
+        XML_RET_CHECK_AND_GOTO (ret, out);
+
+        ret = cli_xml_output_vol_status_tasks (local, dict);
+        if (ret)
+                goto out;
+
+        /* </volume> */
+        ret = xmlTextWriterEndElement (local->writer);
+        XML_RET_CHECK_AND_GOTO (ret, out);
+
+out:
+        return ret;
+}
 #endif
 
 int
@@ -1568,7 +1713,6 @@ cli_xml_output_vol_status (cli_local_t *local, dict_t *dict)
                                         goto out;
                         }
                         break;
-
                 default:
                         break;
 
@@ -1578,7 +1722,11 @@ cli_xml_output_vol_status (cli_local_t *local, dict_t *dict)
                 XML_RET_CHECK_AND_GOTO (ret, out);
         }
 
-        if ((cmd & GF_CLI_STATUS_MASK) == GF_CLI_STATUS_NONE) {
+        /* Tasks are only present when a normal volume status call is done on a
+         * single volume or on all volumes
+         */
+        if (((cmd & GF_CLI_STATUS_MASK) == GF_CLI_STATUS_NONE) &&
+            (cmd & (GF_CLI_STATUS_VOL|GF_CLI_STATUS_ALL))) {
                 ret = cli_xml_output_vol_status_tasks (local, dict);
                 if (ret)
                         goto out;
@@ -2337,6 +2485,7 @@ cli_xml_output_vol_info (cli_local_t *local, dict_t *dict)
         int                     count = 0;
         char                    *volname = NULL;
         char                    *volume_id = NULL;
+        char                    *uuid = NULL;
         int                     type = 0;
         int                     status = 0;
         int                     brick_count = 0;
@@ -2348,7 +2497,9 @@ cli_xml_output_vol_info (cli_local_t *local, dict_t *dict)
         char                    key[1024] = {0,};
         int                     i = 0;
         int                     j = 1;
-
+        char                    *caps = NULL;
+        int                     k __attribute__((unused)) = 0;
+        char                    *snap_volume = NULL;
 
         ret = dict_get_int32 (dict, "count", &count);
         if (ret)
@@ -2390,6 +2541,18 @@ cli_xml_output_vol_info (cli_local_t *local, dict_t *dict)
                                                        "%d", status);
                 XML_RET_CHECK_AND_GOTO (ret, out);
 
+                memset (key, 0, sizeof (key));
+                snprintf (key, sizeof (key), "volume%d.snap_volume", i);
+                ret = dict_get_str (dict, key, &snap_volume);
+                if (ret)
+                        goto out;
+                if (snap_volume) {
+                        ret = xmlTextWriterWriteFormatElement (local->writer,
+                                                         (xmlChar *)"snapVol",
+                                                          "%s", snap_volume);
+                        XML_RET_CHECK_AND_GOTO (ret, out);
+                }
+
                 ret =xmlTextWriterWriteFormatElement
                         (local->writer, (xmlChar *)"statusStr", "%s",
                          cli_vol_status_str[status]);
@@ -2464,20 +2627,95 @@ cli_xml_output_vol_info (cli_local_t *local, dict_t *dict)
                                                        "%d", transport);
                 XML_RET_CHECK_AND_GOTO (ret, out);
 
+#ifdef HAVE_BD_XLATOR
+                /* <xlators> */
+                ret = xmlTextWriterStartElement (local->writer,
+                                                 (xmlChar *)"xlators");
+                XML_RET_CHECK_AND_GOTO (ret, out);
+
+                for (k = 0; ; k++) {
+                        memset (key, 0, sizeof (key));
+                        snprintf (key, sizeof (key),"volume%d.xlator%d", i, k);
+                        ret = dict_get_str (dict, key, &caps);
+                        if (ret)
+                                break;
+
+                        /* <xlator> */
+                        ret = xmlTextWriterStartElement (local->writer,
+                                                         (xmlChar *)"xlator");
+                        XML_RET_CHECK_AND_GOTO (ret, out);
+
+                        ret = xmlTextWriterWriteFormatElement
+                                (local->writer, (xmlChar *)"name", "%s", caps);
+                        XML_RET_CHECK_AND_GOTO (ret, out);
+
+                        /* <capabilities> */
+                        ret = xmlTextWriterStartElement (local->writer,
+                                                         (xmlChar *)
+                                                         "capabilities");
+                        XML_RET_CHECK_AND_GOTO (ret, out);
+
+                        j = 0;
+                        for (j = 0; ;j++) {
+                                memset (key, 0, sizeof (key));
+                                snprintf (key, sizeof (key),
+                                          "volume%d.xlator%d.caps%d", i, k, j);
+                                ret = dict_get_str (dict, key, &caps);
+                                if (ret)
+                                        break;
+                                ret = xmlTextWriterWriteFormatElement
+                                        (local->writer, (xmlChar *)"capability",
+                                         "%s", caps);
+                                XML_RET_CHECK_AND_GOTO (ret, out);
+                        }
+                        /* </capabilities> */
+                        ret = xmlTextWriterEndElement (local->writer);
+                        XML_RET_CHECK_AND_GOTO (ret, out);
+                        /* </xlator> */
+                        ret = xmlTextWriterEndElement (local->writer);
+                        XML_RET_CHECK_AND_GOTO (ret, out);
+                }
+                ret = xmlTextWriterFullEndElement (local->writer);
+                XML_RET_CHECK_AND_GOTO (ret, out);
+                /* </xlators> */
+#else
+                caps = 0; /* Avoid compiler warnings when BD not enabled */
+#endif
+                j = 1;
+
                 /* <bricks> */
                 ret = xmlTextWriterStartElement (local->writer,
                                                  (xmlChar *)"bricks");
                 XML_RET_CHECK_AND_GOTO (ret, out);
                 while (j <= brick_count) {
+                        ret = xmlTextWriterStartElement
+                                (local->writer, (xmlChar *)"brick");
+                        XML_RET_CHECK_AND_GOTO (ret, out);
+
+                        memset (key, 0, sizeof (key));
+                        snprintf (key, sizeof (key), "volume%d.brick%d.uuid",
+                                  i, j);
+                        ret = dict_get_str (dict, key, &uuid);
+                        if (ret)
+                                goto out;
+                        ret = xmlTextWriterWriteFormatAttribute
+                                (local->writer, (xmlChar *)"uuid", "%s",
+                                 uuid);
+                        XML_RET_CHECK_AND_GOTO (ret, out);
+
                         memset (key, 0, sizeof (key));
                         snprintf (key, sizeof (key), "volume%d.brick%d", i, j);
                         ret = dict_get_str (dict, key, &brick);
                         if (ret)
                                 goto out;
-                        ret = xmlTextWriterWriteFormatElement
-                                (local->writer, (xmlChar *)"brick", "%s",
-                                 brick);
+                        ret = xmlTextWriterWriteFormatString
+                                (local->writer, "%s", brick);
                         XML_RET_CHECK_AND_GOTO (ret, out);
+
+                        /* </brick> */
+                        ret = xmlTextWriterEndElement (local->writer);
+                        XML_RET_CHECK_AND_GOTO (ret, out);
+
                         j++;
                 }
                 /* </bricks> */
@@ -2844,22 +3082,29 @@ out:
 #if (HAVE_LIB_XML)
 /* Used for rebalance stop/status, remove-brick status */
 int
-cli_xml_output_vol_rebalance_status (xmlTextWriterPtr writer, dict_t *dict)
+cli_xml_output_vol_rebalance_status (xmlTextWriterPtr writer, dict_t *dict,
+                                     enum gf_task_types task_type)
 {
         int                     ret = -1;
         int                     count = 0;
         char                    *node_name = NULL;
+        char                    *node_uuid = NULL;
         uint64_t                files = 0;
         uint64_t                size = 0;
         uint64_t                lookups = 0;
         int                     status_rcd = 0;
         uint64_t                failures = 0;
+        uint64_t                skipped = 0;
         uint64_t                total_files = 0;
         uint64_t                total_size = 0;
         uint64_t                total_lookups = 0;
         uint64_t                total_failures = 0;
+        uint64_t                total_skipped = 0;
         char                    key[1024] = {0,};
         int                     i = 0;
+        int                     overall_status = -1;
+        double                  elapsed = 0;
+        double                  overall_elapsed = 0;
 
         if (!dict) {
                 ret = 0;
@@ -2881,13 +3126,22 @@ cli_xml_output_vol_rebalance_status (xmlTextWriterPtr writer, dict_t *dict)
                 XML_RET_CHECK_AND_GOTO (ret, out);
 
                 memset (key, 0, sizeof (key));
-                snprintf (key, sizeof (key), "node-uuid-%d", i);
+                snprintf (key, sizeof (key), "node-name-%d", i);
                 ret = dict_get_str (dict, key, &node_name);
                 if (ret)
                         goto out;
                 ret = xmlTextWriterWriteFormatElement (writer,
                                                        (xmlChar *)"nodeName",
                                                        "%s", node_name);
+
+                memset (key, 0, sizeof (key));
+                snprintf (key, sizeof (key), "node-uuid-%d", i);
+                ret = dict_get_str (dict, key, &node_uuid);
+                if (ret)
+                        goto out;
+                ret = xmlTextWriterWriteFormatElement (writer,
+                                                       (xmlChar *)"id",
+                                                       "%s", node_uuid);
                 XML_RET_CHECK_AND_GOTO (ret, out);
 
                 memset (key, 0, sizeof (key));
@@ -2934,6 +3188,27 @@ cli_xml_output_vol_rebalance_status (xmlTextWriterPtr writer, dict_t *dict)
                                                        "%"PRIu64, failures);
                 XML_RET_CHECK_AND_GOTO (ret, out);
 
+                /* skipped-%d is not available for remove brick in dict,
+                   so using failures as skipped count in case of remove-brick
+                   similar to logic used in CLI(non xml output) */
+                if (task_type == GF_TASK_TYPE_REBALANCE) {
+                    memset (key, 0, sizeof (key));
+                    snprintf (key, sizeof (key), "skipped-%d", i);
+                }
+                else {
+                    memset (key, 0, sizeof (key));
+                    snprintf (key, sizeof (key), "failures-%d", i);
+                }
+
+                ret = dict_get_uint64 (dict, key, &skipped);
+                if (ret)
+                        goto out;
+                total_skipped += skipped;
+                ret = xmlTextWriterWriteFormatElement (writer,
+                                                       (xmlChar *)"skipped",
+                                                       "%"PRIu64, skipped);
+                XML_RET_CHECK_AND_GOTO (ret, out);
+
                 memset (key, 0, sizeof (key));
                 snprintf (key, sizeof (key), "status-%d", i);
                 ret = dict_get_int32 (dict, key, &status_rcd);
@@ -2944,6 +3219,33 @@ cli_xml_output_vol_rebalance_status (xmlTextWriterPtr writer, dict_t *dict)
                                                        "%d", status_rcd);
                 XML_RET_CHECK_AND_GOTO (ret, out);
 
+                ret = xmlTextWriterWriteFormatElement (writer,
+                                                       (xmlChar *)"statusStr",
+                                                       "%s",
+                                         cli_vol_task_status_str[status_rcd]);
+
+                memset (key, 0, 256);
+                snprintf (key, 256, "run-time-%d", i);
+                ret = dict_get_double (dict, key, &elapsed);
+                if (ret)
+                        goto out;
+                ret = xmlTextWriterWriteFormatElement (writer,
+                                                       (xmlChar *)"runtime",
+                                                       "%.2f", elapsed);
+                XML_RET_CHECK_AND_GOTO (ret, out);
+
+                if (elapsed > overall_elapsed) {
+                    overall_elapsed = elapsed;
+                }
+
+                if (-1 == overall_status)
+                        overall_status = status_rcd;
+                else if ((GF_DEFRAG_STATUS_COMPLETE == overall_status ||
+                          status_rcd > overall_status) &&
+                         (status_rcd != GF_DEFRAG_STATUS_COMPLETE))
+                        overall_status = status_rcd;
+                XML_RET_CHECK_AND_GOTO (ret, out);
+
                 /* </node> */
                 ret = xmlTextWriterEndElement (writer);
                 XML_RET_CHECK_AND_GOTO (ret, out);
@@ -2970,7 +3272,22 @@ cli_xml_output_vol_rebalance_status (xmlTextWriterPtr writer, dict_t *dict)
                                                "%"PRIu64, total_failures);
         XML_RET_CHECK_AND_GOTO (ret, out);
 
-        // TODO : Aggregate status
+        ret = xmlTextWriterWriteFormatElement (writer,(xmlChar *)"skipped",
+                                               "%"PRIu64, total_skipped);
+        XML_RET_CHECK_AND_GOTO (ret, out);
+
+        ret = xmlTextWriterWriteFormatElement (writer,(xmlChar *)"status",
+                                               "%d", overall_status);
+        XML_RET_CHECK_AND_GOTO (ret, out);
+
+        ret = xmlTextWriterWriteFormatElement (writer,(xmlChar *)"statusStr",
+                                               "%s",
+                                      cli_vol_task_status_str[overall_status]);
+        XML_RET_CHECK_AND_GOTO (ret, out);
+
+        ret = xmlTextWriterWriteFormatElement (writer,(xmlChar *)"runtime",
+                                               "%.2f", overall_elapsed);
+        XML_RET_CHECK_AND_GOTO (ret, out);
 
         /* </aggregate> */
         ret = xmlTextWriterEndElement (writer);
@@ -3017,7 +3334,8 @@ cli_xml_output_vol_rebalance (gf_cli_defrag_type op, dict_t *dict, int op_ret,
         XML_RET_CHECK_AND_GOTO (ret, out);
 
         if ((GF_DEFRAG_CMD_STOP == op) || (GF_DEFRAG_CMD_STATUS == op)) {
-                ret = cli_xml_output_vol_rebalance_status (writer, dict);
+                ret = cli_xml_output_vol_rebalance_status (writer, dict,
+                                                      GF_TASK_TYPE_REBALANCE);
                 if (ret)
                         goto out;
         }
@@ -3068,7 +3386,8 @@ cli_xml_output_vol_remove_brick (gf_boolean_t status_op, dict_t *dict,
         }
 
         if (status_op) {
-                ret = cli_xml_output_vol_rebalance_status (writer, dict);
+                ret = cli_xml_output_vol_rebalance_status (writer, dict,
+                                                   GF_TASK_TYPE_REMOVE_BRICK);
                 if (ret)
                         goto out;
         }
diff --git a/cli/src/cli.c b/cli/src/cli.c
index 37015c33d..91b315ff1 100644
--- a/cli/src/cli.c
+++ b/cli/src/cli.c
@@ -168,7 +168,8 @@ logging_init (glusterfs_ctx_t *ctx, struct cli_state *state)
         char *log_file = state->log_file ? state->log_file :
                          DEFAULT_CLI_LOG_FILE_DIRECTORY "/cli.log";
 
-        if (gf_log_init (ctx, log_file) == -1) {
+        /* passing ident as NULL means to use default ident for syslog */
+        if (gf_log_init (ctx, log_file, NULL) == -1) {
                 fprintf (stderr, "ERROR: failed to open logfile %s\n",
                          log_file);
                 return -1;
@@ -352,6 +353,12 @@ cli_opt_parse (char *opt, struct cli_state *state)
                 return 0;
         }
 
+        oarg = strtail (opt, "glusterd-sock=");
+        if (oarg) {
+                state->glusterd_sock = oarg;
+                return 0;
+        }
+
         return -1;
 }
 
@@ -413,7 +420,6 @@ cli_state_init (struct cli_state *state)
         int                   ret = 0;
 
 
-        state->remote_host = "localhost";
         state->log_level = -1;
 
         tree = &state->tree;
@@ -500,23 +506,46 @@ cli_rpc_init (struct cli_state *state)
         if (!options)
                 goto out;
 
-        ret = dict_set_str (options, "remote-host", state->remote_host);
-        if (ret)
-                goto out;
+        /* Connect using to glusterd using the specified method, giving
+         * preference to unix socket connection. If nothing is specified connect
+         * to the default glusterd socket
+         */
+        if (state->glusterd_sock) {
+                gf_log ("cli", GF_LOG_INFO, "Connecting to glusterd using "
+                        "sockfile %s", state->glusterd_sock);
+                ret = rpc_transport_unix_options_build (&options,
+                                                        state->glusterd_sock,
+                                                        0);
+                if (ret)
+                        goto out;
+        } else if (state->remote_host) {
+                gf_log ("cli", GF_LOG_INFO, "Connecting to remote glusterd at "
+                        "%s", state->remote_host);
+                ret = dict_set_str (options, "remote-host", state->remote_host);
+                if (ret)
+                        goto out;
 
-        if (state->remote_port)
-                port = state->remote_port;
+                if (state->remote_port)
+                        port = state->remote_port;
 
-        ret = dict_set_int32 (options, "remote-port", port);
-        if (ret)
-                goto out;
+                ret = dict_set_int32 (options, "remote-port", port);
+                if (ret)
+                        goto out;
 
-        ret = dict_set_str (options, "transport.address-family", "inet");
-        if (ret)
-                goto out;
+                ret = dict_set_str (options, "transport.address-family",
+                                    "inet");
+                if (ret)
+                        goto out;
+        } else {
+                gf_log ("cli", GF_LOG_DEBUG, "Connecting to glusterd using "
+                        "default socket");
+                ret = rpc_transport_unix_options_build
+                        (&options, DEFAULT_GLUSTERD_SOCKFILE, 0);
+                if (ret)
+                        goto out;
+        }
 
         rpc = rpc_clnt_new (options, this->ctx, this->name, 16);
-
         if (!rpc)
                 goto out;
 
@@ -526,7 +555,7 @@ cli_rpc_init (struct cli_state *state)
                 goto out;
         }
 
-        rpc_clnt_start (rpc);
+        ret = rpc_clnt_start (rpc);
 out:
         if (ret) {
                 if (rpc)
diff --git a/cli/src/cli.h b/cli/src/cli.h
index da3d9c73c..8daa4b741 100644
--- a/cli/src/cli.h
+++ b/cli/src/cli.h
@@ -18,6 +18,7 @@
 #include "rpc-clnt.h"
 #include "glusterfs.h"
 #include "protocol-common.h"
+#include "logging.h"
 
 #include "cli1-xdr.h"
 
@@ -32,25 +33,10 @@
 #define CLI_DEFAULT_CMD_TIMEOUT              120
 #define CLI_TEN_MINUTES_TIMEOUT              600 //Longer timeout for volume top
 #define DEFAULT_CLI_LOG_FILE_DIRECTORY     DATADIR "/log/glusterfs"
-#define DEFAULT_LOG_FILE_DIRECTORY         DATADIR "/log/glusterfs"
-#define DEFAULT_VAR_RUN_DIRECTORY          DATADIR "/run/gluster"
 #define CLI_VOL_STATUS_BRICK_LEN              55
 #define CLI_TAB_LENGTH                         8
 #define CLI_BRICK_STATUS_LINE_LEN             78
 
-#define CLI_LOCAL_INIT(local, words, frame, dictionary) \
-        do {                                                 \
-                local = cli_local_get ();                    \
-                                                             \
-                if (local) {                                 \
-                        local->words = words;                \
-                        if (dictionary)                      \
-                                local->dict = dictionary;    \
-                        if (frame)                           \
-                                frame->local = local;        \
-                }                                            \
-        } while (0)
-
 enum argp_option_keys {
 	ARGP_DEBUG_KEY = 133,
 	ARGP_PORT_KEY = 'p',
@@ -66,6 +52,7 @@ struct cli_cmd;
 
 extern char *cli_vol_type_str[];
 extern char *cli_vol_status_str[];
+extern char *cli_vol_task_status_str[];
 
 typedef int (cli_cmd_cbk_t)(struct cli_state *state,
                             struct cli_cmd_word *word,
@@ -128,6 +115,8 @@ struct cli_state {
 
         char                 *log_file;
         gf_loglevel_t         log_level;
+
+        char                 *glusterd_sock;
 };
 
 struct cli_local {
@@ -147,6 +136,18 @@ struct cli_local {
 #endif
 };
 
+struct gf_cli_gsync_detailed_status_ {
+        char *node;
+        char *master;
+        char *slave;
+        char *health;
+        char *uptime;
+        char *files_syncd;
+        char *files_pending;
+        char *bytes_pending;
+        char *deletes_pending;
+};
+
 struct cli_volume_status {
         int            port;
         int            online;
@@ -165,6 +166,13 @@ struct cli_volume_status {
 #endif
 };
 
+struct snap_config_opt_vals_ {
+        char           *op_name;
+        char           *question;
+};
+
+typedef struct gf_cli_gsync_detailed_status_ gf_cli_gsync_status_t;
+
 typedef struct cli_volume_status cli_volume_status_t;
 
 typedef struct cli_local cli_local_t;
@@ -229,7 +237,7 @@ cli_cmd_quota_parse (const char **words, int wordcount, dict_t **opt);
 
 int32_t
 cli_cmd_volume_set_parse (const char **words, int wordcount,
-                          dict_t **options);
+                          dict_t **options, char **op_errstr);
 
 int32_t
 cli_cmd_volume_add_brick_parse (const char **words, int wordcount,
@@ -379,7 +387,14 @@ cli_xml_output_generic_volume (char *op, dict_t *dict, int op_ret, int op_errno,
 int
 cli_xml_output_vol_gsync (dict_t *dict, int op_ret, int op_errno,
                           char *op_errstr);
+int
+cli_xml_output_vol_status_tasks_detail (cli_local_t *local, dict_t *dict);
 
 char *
 is_server_debug_xlator (void *myframe);
+
+int32_t
+cli_cmd_snapshot_parse (const char **words, int wordcount, dict_t **options,
+                        struct cli_state *state);
+
 #endif /* __CLI_H__ */
diff --git a/configure.ac b/configure.ac
index cfccf3f9e..b3d1ed184 100644
--- a/configure.ac
+++ b/configure.ac
@@ -29,6 +29,8 @@ AC_CONFIG_HEADERS([config.h])
 AC_CONFIG_FILES([Makefile
                 libglusterfs/Makefile
                 libglusterfs/src/Makefile
+                geo-replication/src/peer_gsec_create
+                geo-replication/src/peer_add_secret_pub
                 glusterfsd/Makefile
                 glusterfsd/src/Makefile
                 rpc/Makefile
@@ -51,8 +53,8 @@ AC_CONFIG_FILES([Makefile
                 xlators/storage/Makefile
                 xlators/storage/posix/Makefile
                 xlators/storage/posix/src/Makefile
-                xlators/storage/bd_map/Makefile
-                xlators/storage/bd_map/src/Makefile
+                xlators/storage/bd/Makefile
+                xlators/storage/bd/src/Makefile
                 xlators/cluster/Makefile
                 xlators/cluster/afr/Makefile
                 xlators/cluster/afr/src/Makefile
@@ -65,6 +67,8 @@ AC_CONFIG_FILES([Makefile
                 xlators/performance/write-behind/src/Makefile
                 xlators/performance/read-ahead/Makefile
                 xlators/performance/read-ahead/src/Makefile
+                xlators/performance/readdir-ahead/Makefile
+                xlators/performance/readdir-ahead/src/Makefile
                 xlators/performance/io-threads/Makefile
                 xlators/performance/io-threads/src/Makefile
                 xlators/performance/io-cache/Makefile
@@ -109,6 +113,8 @@ AC_CONFIG_FILES([Makefile
                 xlators/features/marker/src/Makefile
                 xlators/features/read-only/Makefile
                 xlators/features/read-only/src/Makefile
+                xlators/features/compress/Makefile
+                xlators/features/compress/src/Makefile
                 xlators/features/mac-compat/Makefile
                 xlators/features/mac-compat/src/Makefile
                 xlators/features/quiesce/Makefile
@@ -117,12 +123,18 @@ AC_CONFIG_FILES([Makefile
                 xlators/features/index/src/Makefile
                 xlators/features/protect/Makefile
                 xlators/features/protect/src/Makefile
+                xlators/features/gfid-access/Makefile
+                xlators/features/gfid-access/src/Makefile
                 xlators/playground/Makefile
                 xlators/playground/template/Makefile
                 xlators/playground/template/src/Makefile
                 xlators/encryption/Makefile
                 xlators/encryption/rot-13/Makefile
                 xlators/encryption/rot-13/src/Makefile
+                xlators/encryption/crypt/Makefile
+                xlators/encryption/crypt/src/Makefile
+		xlators/features/qemu-block/Makefile
+		xlators/features/qemu-block/src/Makefile
                 xlators/system/Makefile
                 xlators/system/posix-acl/Makefile
                 xlators/system/posix-acl/src/Makefile
@@ -149,12 +161,15 @@ AC_CONFIG_FILES([Makefile
                 extras/ocf/glusterd
                 extras/ocf/volume
                 extras/LinuxRPM/Makefile
+                extras/geo-rep/Makefile
                 contrib/fuse-util/Makefile
                 contrib/uuid/uuid_types.h
                 glusterfs-api.pc
                 libgfchangelog.pc
                 api/Makefile
                 api/src/Makefile
+                api/examples/Makefile
+                api/examples/setup.py
 		geo-replication/Makefile
 		geo-replication/src/Makefile
 		geo-replication/syncdaemon/Makefile
@@ -291,16 +306,15 @@ if test "x$enable_fuse_client" != "xno"; then
 fi
 
 AC_ARG_ENABLE([bd-xlator],
-                AC_HELP_STRING([--enable-bd-xlator],
-                        [Build BD xlator]))
+              AC_HELP_STRING([--enable-bd-xlator], [Build BD xlator]))
 
 if test "x$enable_bd_xlator" != "xno"; then
   AC_CHECK_LIB([lvm2app],
-                  [lvm_init,lvm_lv_from_name],
-                  [HAVE_BD_LIB="yes"],
-                  [HAVE_BD_LIB="no"])
+               [lvm_init,lvm_lv_from_name],
+               [HAVE_BD_LIB="yes"],
+               [HAVE_BD_LIB="no"])
 
-  if test "x$HAVE_BD_LIB" = "xyes"; then
+if test "x$HAVE_BD_LIB" = "xyes"; then
     # lvm_lv_from_name() has been made public with lvm2-2.02.79
     AC_CHECK_DECLS(
                   [lvm_lv_from_name],
@@ -316,10 +330,10 @@ if test "x$enable_bd_xlator" = "xyes" -a "x$HAVE_BD_LIB" = "xno"; then
 fi
 
 BUILD_BD_XLATOR=no
-if test "x${enable_bd_xlator}" != "xno" -a "x${HAVE_BD_LIB}" = "xyes"; then
+if test "x${enable-bd-xlator}" != "xno" -a "x${HAVE_BD_LIB}" = "xyes"; then
   BUILD_BD_XLATOR=yes
-  AC_DEFINE(HAVE_BD_XLATOR, 1, [define if lvm2app library found and bd
-                  xlator enabled])
+  AC_DEFINE(HAVE_BD_XLATOR, 1, [define if lvm2app library found and bd xlator
+                            enabled])
   if test "x$NEED_LVM_LV_FROM_NAME_DECL" = "xyes"; then
     AC_DEFINE(NEED_LVM_LV_FROM_NAME_DECL, 1, [defined if lvm_lv_from_name()
                   was not found in the lvm2app.h header, but can be linked])
@@ -328,6 +342,26 @@ fi
 
 AM_CONDITIONAL([ENABLE_BD_XLATOR], [test x$BUILD_BD_XLATOR = xyes])
 
+# start encryption/crypt section
+
+AC_CHECK_HEADERS([openssl/cmac.h], [have_cmac_h=yes], [have_cmac_h=no])
+
+AC_ARG_ENABLE([crypt-xlator],
+	AC_HELP_STRING([--enable-crypt-xlator], [Build crypt encryption xlator]))
+
+if test "x$enable_crypt_xlator" = "xyes" -a "x$have_cmac_h" = "xno"; then
+   echo "Encryption xlator requires OpenSSL with cmac.h"
+   exit 1
+fi
+
+BUILD_CRYPT_XLATOR=no
+if test "x$enable_crypt_xlator" != "xno" -a "x$have_cmac_h" = "xyes"; then
+   BUILD_CRYPT_XLATOR=yes
+   AC_DEFINE(HAVE_CRYPT_XLATOR, 1, [enable building crypt encryption xlator])
+fi
+
+AM_CONDITIONAL([ENABLE_CRYPT_XLATOR], [test x$BUILD_CRYPT_XLATOR = xyes])
+
 AC_SUBST(FUSE_CLIENT_SUBDIR)
 # end FUSE section
 
@@ -348,6 +382,33 @@ fi
 AC_SUBST(FUSERMOUNT_SUBDIR)
 #end FUSERMOUNT section
 
+# QEMU_BLOCK section
+
+AC_ARG_ENABLE([qemu-block],
+                AC_HELP_STRING([--enable-qemu-block],
+                        [Build QEMU Block formats translator]))
+
+if test "x$enable_qemu_block" != "xno"; then
+  PKG_CHECK_MODULES([GLIB], [glib-2.0],
+                    [HAVE_GLIB_2="yes"],
+  		    [HAVE_GLIB_2="no"])
+fi
+
+if test "x$enable_qemu_block" = "xyes" -a "x$HAVE_GLIB_2" = "xno"; then
+   echo "QEMU Block formats translator requires libglib-2.0, but missing."
+   exit 1
+fi
+
+BUILD_QEMU_BLOCK=no
+if test "x${enable_qemu_block}" != "xno" -a "x${HAVE_GLIB_2}" = "xyes"; then
+  BUILD_QEMU_BLOCK=yes
+  AC_DEFINE(HAVE_QEMU_BLOCK, 1, [define if libglib-2.0 library found and QEMU
+                  	     Block translator enabled])
+fi
+
+AM_CONDITIONAL([ENABLE_QEMU_BLOCK], [test x$BUILD_QEMU_BLOCK = xyes])
+
+# end QEMU_BLOCK section
 
 # EPOLL section
 AC_ARG_ENABLE([epoll],
@@ -373,7 +434,7 @@ if test "x$enable_ibverbs" != "xno"; then
                [ibv_get_device_list],
                [HAVE_LIBIBVERBS="yes"],
                [HAVE_LIBIBVERBS="no"])
-  AC_CHECK_LIB([rdmacm], [rdma_create_id], [HAVE_RDMACM="yes"], [HAVE_RDMACM="no"]) 
+  AC_CHECK_LIB([rdmacm], [rdma_create_id], [HAVE_RDMACM="yes"], [HAVE_RDMACM="no"])
 fi
 
 if test "x$enable_ibverbs" = "xyes"; then
@@ -446,6 +507,17 @@ AC_SUBST(SYNCDAEMON_COMPILE)
 AC_SUBST(SYNCDAEMON_SUBDIR)
 # end SYNCDAEMON section
 
+# CDC xlator - check if libz is present if so enable HAVE_LIB_Z
+echo -n "checking if libz is present... "
+
+PKG_CHECK_MODULES([ZLIB], [zlib >= 1.2.0],
+                  [echo "yes (features requiring zlib enabled)"  AC_DEFINE(HAVE_LIB_Z, 1, [define if zlib is present])],
+                  [echo "no"] )
+
+AC_SUBST(LIBZ_CFLAGS)
+AC_SUBST(LIBZ_LIBS)
+# end CDC xlator secion
+
 # check for systemtap/dtrace
 BUILD_SYSTEMTAP=no
 AC_MSG_CHECKING([whether to include systemtap tracing support])
@@ -478,14 +550,24 @@ if test "x${DTRACE}" = "xyes" -a "x${SDT_H_FOUND}" = "xyes"; then
 fi
 # end of systemtap/dtrace
 
-#check if libxml is present if so enable HAVE_LIB_XML
-echo -n "checking if libxml2 is present... "
-
-AM_PATH_XML2([2.6.19])
-
-if test "x${no_xml}" = "x"; then
-    AC_DEFINE([HAVE_LIB_XML], [1], [Define to 1 if using libxml2.])
+# xml-output
+AC_ARG_ENABLE([xml-output],
+              AC_HELP_STRING([--disable-xml-output],
+                             [Disable the xml output]))
+BUILD_XML_OUTPUT="yes"
+if test "x$enable_xml_output" != "xno"; then
+    #check if libxml is present if so enable HAVE_LIB_XML
+    m4_ifdef([AM_PATH_XML2],[AM_PATH_XML2([2.6.19])], [no_xml=yes])
+    if test "x${no_xml}" = "x"; then
+        AC_DEFINE([HAVE_LIB_XML], [1], [Define to 1 if using libxml2.])
+    else
+        AC_MSG_WARN([libxml2 devel libraries not found disabling XML support])
+        BUILD_XML_OUTPUT="no"
+    fi
+else
+    BUILD_XML_OUTPUT="no"
 fi
+# end of xml-output
 
 dnl FreeBSD > 5 has execinfo as a Ported library for giving a workaround
 dnl solution to GCC backtrace functionality
@@ -521,11 +603,15 @@ if test "x${have_linkat}" = "xyes"; then
 fi
 AC_SUBST(HAVE_LINKAT)
 
+dnl check for Monotonic clock
+AC_CHECK_FUNC([clock_gettime], [has_monotonic_clock=yes], AC_CHECK_LIB([rt], [clock_gettime], , AC_MSG_WARN([System doesn't have monotonic clock using contrib])))
+
 dnl Check for argp
 AC_CHECK_HEADER([argp.h], AC_DEFINE(HAVE_ARGP, 1, [have argp]))
 AC_CONFIG_SUBDIRS(argp-standalone)
+
 BUILD_ARGP_STANDALONE=no
-if test "x${ac_cv_header_argp_h}" = "xno"; then 
+if test "x${ac_cv_header_argp_h}" = "xno"; then
    BUILD_ARGP_STANDALONE=yes
    ARGP_STANDALONE_CPPFLAGS='-I${top_srcdir}/argp-standalone'
    ARGP_STANDALONE_LDADD='${top_builddir}/argp-standalone/libargp.a'
@@ -557,7 +643,7 @@ if test "x${have_posix_fallocate}" = "xyes"; then
 fi
 
 
-# Check the distribution where you are compiling glusterfs on 
+# Check the distribution where you are compiling glusterfs on
 
 GF_DISTRIBUTION=
 AC_CHECK_FILE([/etc/debian_version])
@@ -591,9 +677,20 @@ else
     CFLAGS="$saved_CFLAGS $CFLAGS"
 fi
 
+# check for gcc -Werror=implicit-function-declaration
+saved_CFLAGS=$CFLAGS
+CFLAGS="-Werror=implicit-function-declaration"
+AC_MSG_CHECKING([whether $CC accepts -Werror=implicit-function-declaration])
+AC_COMPILE_IFELSE([AC_LANG_PROGRAM()], [cc_werror_implicit=yes], [cc_werror_implicit=no])
+echo $cc_werror_implicit
+if test "x$cc_werror_implicit" = "xno"; then
+    CFLAGS="$saved_CFLAGS"
+else
+    CFLAGS="$saved_CFLAGS $CFLAGS"
+fi
+
 case $host_os in
      linux*)
-     dnl        GF_LINUX_HOST_OS=1
         GF_HOST_OS="GF_LINUX_HOST_OS"
         GF_CFLAGS="${ARGP_STANDALONE_CPPFLAGS} -O0"
         GF_GLUSTERFS_CFLAGS="${GF_CFLAGS}"
@@ -785,7 +882,10 @@ echo "georeplication       : $BUILD_SYNCDAEMON"
 echo "Linux-AIO            : $BUILD_LIBAIO"
 echo "Enable Debug         : $BUILD_DEBUG"
 echo "systemtap            : $BUILD_SYSTEMTAP"
-echo "Block Device backend : $BUILD_BD_XLATOR"
+echo "Block Device xlator  : $BUILD_BD_XLATOR"
 echo "glupy                : $BUILD_GLUPY"
 echo "Use syslog           : $USE_SYSLOG"
+echo "XML output           : $BUILD_XML_OUTPUT"
+echo "QEMU Block formats   : $BUILD_QEMU_BLOCK"
+echo "Encryption xlator    : $BUILD_CRYPT_XLATOR"
 echo
diff --git a/contrib/fuse-include/fuse_kernel.h b/contrib/fuse-include/fuse_kernel.h
index 2f4e26ff0..60bb2f9f7 100644
--- a/contrib/fuse-include/fuse_kernel.h
+++ b/contrib/fuse-include/fuse_kernel.h
@@ -60,23 +60,75 @@
  * 7.13
  *  - make max number of background requests and congestion threshold
  *    tunables
+ *
+ * 7.14
+ *  - add splice support to fuse device
+ *
+ * 7.15
+ *  - add store notify
+ *  - add retrieve notify
+ *
+ * 7.16
+ *  - add BATCH_FORGET request
+ *  - FUSE_IOCTL_UNRESTRICTED shall now return with array of 'struct
+ *    fuse_ioctl_iovec' instead of ambiguous 'struct iovec'
+ *  - add FUSE_IOCTL_32BIT flag
+ *
+ * 7.17
+ *  - add FUSE_FLOCK_LOCKS and FUSE_RELEASE_FLOCK_UNLOCK
+ *
+ * 7.18
+ *  - add FUSE_IOCTL_DIR flag
+ *  - add FUSE_NOTIFY_DELETE
+ *
+ * 7.19
+ *  - add FUSE_FALLOCATE
+ *
+ * 7.20
+ *  - add FUSE_AUTO_INVAL_DATA
+ *
+ * 7.21
+ *  - add FUSE_READDIRPLUS
+ *  - send the requested events in POLL request
+ *
+ * 7.22
+ *  - add FUSE_ASYNC_DIO
  */
 
 #ifndef _LINUX_FUSE_H
 #define _LINUX_FUSE_H
 
-#include <sys/types.h>
-#define __u64 uint64_t
-#define __s64 int64_t
-#define __u32 uint32_t
-#define __s32 int32_t
-#define __u16 uint16_t
+#ifdef __KERNEL__
+#include <linux/types.h>
+#else
+#include <stdint.h>
+#endif
+
+/*
+ * Version negotiation:
+ *
+ * Both the kernel and userspace send the version they support in the
+ * INIT request and reply respectively.
+ *
+ * If the major versions match then both shall use the smallest
+ * of the two minor versions for communication.
+ *
+ * If the kernel supports a larger major version, then userspace shall
+ * reply with the major version it supports, ignore the rest of the
+ * INIT message and expect a new INIT message from the kernel with a
+ * matching major version.
+ *
+ * If the library supports a larger major version, then it shall fall
+ * back to the major protocol version sent by the kernel for
+ * communication and reply with that major version (and an arbitrary
+ * supported minor version).
+ */
 
 /** Version number of this interface */
 #define FUSE_KERNEL_VERSION 7
 
 /** Minor version number of this interface */
-#define FUSE_KERNEL_MINOR_VERSION 13
+#define FUSE_KERNEL_MINOR_VERSION 22
 
 /** The node ID of the root inode */
 #define FUSE_ROOT_ID 1
@@ -85,42 +137,42 @@
    userspace works under 64bit kernels */
 
 struct fuse_attr {
-	__u64	ino;
-	__u64	size;
-	__u64	blocks;
-	__u64	atime;
-	__u64	mtime;
-	__u64	ctime;
-	__u32	atimensec;
-	__u32	mtimensec;
-	__u32	ctimensec;
-	__u32	mode;
-	__u32	nlink;
-	__u32	uid;
-	__u32	gid;
-	__u32	rdev;
-	__u32	blksize;
-	__u32	padding;
+	uint64_t	ino;
+	uint64_t	size;
+	uint64_t	blocks;
+	uint64_t	atime;
+	uint64_t	mtime;
+	uint64_t	ctime;
+	uint32_t	atimensec;
+	uint32_t	mtimensec;
+	uint32_t	ctimensec;
+	uint32_t	mode;
+	uint32_t	nlink;
+	uint32_t	uid;
+	uint32_t	gid;
+	uint32_t	rdev;
+	uint32_t	blksize;
+	uint32_t	padding;
 };
 
 struct fuse_kstatfs {
-	__u64	blocks;
-	__u64	bfree;
-	__u64	bavail;
-	__u64	files;
-	__u64	ffree;
-	__u32	bsize;
-	__u32	namelen;
-	__u32	frsize;
-	__u32	padding;
-	__u32	spare[6];
+	uint64_t	blocks;
+	uint64_t	bfree;
+	uint64_t	bavail;
+	uint64_t	files;
+	uint64_t	ffree;
+	uint32_t	bsize;
+	uint32_t	namelen;
+	uint32_t	frsize;
+	uint32_t	padding;
+	uint32_t	spare[6];
 };
 
 struct fuse_file_lock {
-	__u64	start;
-	__u64	end;
-	__u32	type;
-	__u32	pid; /* tgid */
+	uint64_t	start;
+	uint64_t	end;
+	uint32_t	type;
+	uint32_t	pid; /* tgid */
 };
 
 /**
@@ -151,8 +203,22 @@ struct fuse_file_lock {
 /**
  * INIT request/reply flags
  *
+ * FUSE_ASYNC_READ: asynchronous read requests
+ * FUSE_POSIX_LOCKS: remote locking for POSIX file locks
+ * FUSE_FILE_OPS: kernel sends file handle for fstat, etc... (not yet supported)
+ * FUSE_ATOMIC_O_TRUNC: handles the O_TRUNC open flag in the filesystem
  * FUSE_EXPORT_SUPPORT: filesystem handles lookups of "." and ".."
+ * FUSE_BIG_WRITES: filesystem can handle write size larger than 4kB
  * FUSE_DONT_MASK: don't apply umask to file mode on create operations
+ * FUSE_SPLICE_WRITE: kernel supports splice write on the device
+ * FUSE_SPLICE_MOVE: kernel supports splice move on the device
+ * FUSE_SPLICE_READ: kernel supports splice read on the device
+ * FUSE_FLOCK_LOCKS: remote locking for BSD style file locks
+ * FUSE_HAS_IOCTL_DIR: kernel supports ioctl on directories
+ * FUSE_AUTO_INVAL_DATA: automatically invalidate cached pages
+ * FUSE_DO_READDIRPLUS: do READDIRPLUS (READDIR+LOOKUP in one)
+ * FUSE_READDIRPLUS_AUTO: adaptive readdirplus
+ * FUSE_ASYNC_DIO: asynchronous direct I/O submission
  */
 #define FUSE_ASYNC_READ		(1 << 0)
 #define FUSE_POSIX_LOCKS	(1 << 1)
@@ -161,7 +227,14 @@ struct fuse_file_lock {
 #define FUSE_EXPORT_SUPPORT	(1 << 4)
 #define FUSE_BIG_WRITES		(1 << 5)
 #define FUSE_DONT_MASK		(1 << 6)
+#define FUSE_SPLICE_WRITE	(1 << 7)
+#define FUSE_SPLICE_MOVE	(1 << 8)
+#define FUSE_SPLICE_READ	(1 << 9)
+#define FUSE_FLOCK_LOCKS	(1 << 10)
+#define FUSE_HAS_IOCTL_DIR	(1 << 11)
+#define FUSE_AUTO_INVAL_DATA	(1 << 12)
 #define FUSE_DO_READDIRPLUS	(1 << 13)
+#define FUSE_READDIRPLUS_AUTO	(1 << 14)
 #define FUSE_ASYNC_DIO		(1 << 15)
 
 /**
@@ -175,6 +248,7 @@ struct fuse_file_lock {
  * Release flags
  */
 #define FUSE_RELEASE_FLUSH	(1 << 0)
+#define FUSE_RELEASE_FLOCK_UNLOCK	(1 << 1)
 
 /**
  * Getattr flags
@@ -206,12 +280,16 @@ struct fuse_file_lock {
  * FUSE_IOCTL_COMPAT: 32bit compat ioctl on 64bit machine
  * FUSE_IOCTL_UNRESTRICTED: not restricted to well-formed ioctls, retry allowed
  * FUSE_IOCTL_RETRY: retry with new iovecs
+ * FUSE_IOCTL_32BIT: 32bit ioctl
+ * FUSE_IOCTL_DIR: is a directory
  *
  * FUSE_IOCTL_MAX_IOV: maximum of in_iovecs + out_iovecs
  */
 #define FUSE_IOCTL_COMPAT	(1 << 0)
 #define FUSE_IOCTL_UNRESTRICTED	(1 << 1)
 #define FUSE_IOCTL_RETRY	(1 << 2)
+#define FUSE_IOCTL_32BIT	(1 << 3)
+#define FUSE_IOCTL_DIR		(1 << 4)
 
 #define FUSE_IOCTL_MAX_IOV	256
 
@@ -261,9 +339,11 @@ enum fuse_opcode {
 	FUSE_DESTROY       = 38,
 	FUSE_IOCTL         = 39,
 	FUSE_POLL          = 40,
-
-	FUSE_FALLOCATE	   = 43,
+	FUSE_NOTIFY_REPLY  = 41,
+	FUSE_BATCH_FORGET  = 42,
+	FUSE_FALLOCATE     = 43,
 	FUSE_READDIRPLUS   = 44,
+
 	/* CUSE specific operations */
 	CUSE_INIT          = 4096,
 };
@@ -272,6 +352,9 @@ enum fuse_notify_code {
 	FUSE_NOTIFY_POLL   = 1,
 	FUSE_NOTIFY_INVAL_INODE = 2,
 	FUSE_NOTIFY_INVAL_ENTRY = 3,
+	FUSE_NOTIFY_STORE = 4,
+	FUSE_NOTIFY_RETRIEVE = 5,
+	FUSE_NOTIFY_DELETE = 6,
 	FUSE_NOTIFY_CODE_MAX,
 };
 
@@ -281,133 +364,143 @@ enum fuse_notify_code {
 #define FUSE_COMPAT_ENTRY_OUT_SIZE 120
 
 struct fuse_entry_out {
-	__u64	nodeid;		/* Inode ID */
-	__u64	generation;	/* Inode generation: nodeid:gen must
-				   be unique for the fs's lifetime */
-	__u64	entry_valid;	/* Cache timeout for the name */
-	__u64	attr_valid;	/* Cache timeout for the attributes */
-	__u32	entry_valid_nsec;
-	__u32	attr_valid_nsec;
+	uint64_t	nodeid;		/* Inode ID */
+	uint64_t	generation;	/* Inode generation: nodeid:gen must
+					   be unique for the fs's lifetime */
+	uint64_t	entry_valid;	/* Cache timeout for the name */
+	uint64_t	attr_valid;	/* Cache timeout for the attributes */
+	uint32_t	entry_valid_nsec;
+	uint32_t	attr_valid_nsec;
 	struct fuse_attr attr;
 };
 
 struct fuse_forget_in {
-	__u64	nlookup;
+	uint64_t	nlookup;
+};
+
+struct fuse_forget_one {
+	uint64_t	nodeid;
+	uint64_t	nlookup;
+};
+
+struct fuse_batch_forget_in {
+	uint32_t	count;
+	uint32_t	dummy;
 };
 
 struct fuse_getattr_in {
-	__u32	getattr_flags;
-	__u32	dummy;
-	__u64	fh;
+	uint32_t	getattr_flags;
+	uint32_t	dummy;
+	uint64_t	fh;
 };
 
 #define FUSE_COMPAT_ATTR_OUT_SIZE 96
 
 struct fuse_attr_out {
-	__u64	attr_valid;	/* Cache timeout for the attributes */
-	__u32	attr_valid_nsec;
-	__u32	dummy;
+	uint64_t	attr_valid;	/* Cache timeout for the attributes */
+	uint32_t	attr_valid_nsec;
+	uint32_t	dummy;
 	struct fuse_attr attr;
 };
 
 #define FUSE_COMPAT_MKNOD_IN_SIZE 8
 
 struct fuse_mknod_in {
-	__u32	mode;
-	__u32	rdev;
-	__u32	umask;
-	__u32	padding;
+	uint32_t	mode;
+	uint32_t	rdev;
+	uint32_t	umask;
+	uint32_t	padding;
 };
 
 struct fuse_mkdir_in {
-	__u32	mode;
-	__u32	umask;
+	uint32_t	mode;
+	uint32_t	umask;
 };
 
 struct fuse_rename_in {
-	__u64	newdir;
+	uint64_t	newdir;
 };
 
 struct fuse_link_in {
-	__u64	oldnodeid;
+	uint64_t	oldnodeid;
 };
 
 struct fuse_setattr_in {
-	__u32	valid;
-	__u32	padding;
-	__u64	fh;
-	__u64	size;
-	__u64	lock_owner;
-	__u64	atime;
-	__u64	mtime;
-	__u64	unused2;
-	__u32	atimensec;
-	__u32	mtimensec;
-	__u32	unused3;
-	__u32	mode;
-	__u32	unused4;
-	__u32	uid;
-	__u32	gid;
-	__u32	unused5;
+	uint32_t	valid;
+	uint32_t	padding;
+	uint64_t	fh;
+	uint64_t	size;
+	uint64_t	lock_owner;
+	uint64_t	atime;
+	uint64_t	mtime;
+	uint64_t	unused2;
+	uint32_t	atimensec;
+	uint32_t	mtimensec;
+	uint32_t	unused3;
+	uint32_t	mode;
+	uint32_t	unused4;
+	uint32_t	uid;
+	uint32_t	gid;
+	uint32_t	unused5;
 };
 
 struct fuse_open_in {
-	__u32	flags;
-	__u32	unused;
+	uint32_t	flags;
+	uint32_t	unused;
 };
 
 struct fuse_create_in {
-	__u32	flags;
-	__u32	mode;
-	__u32	umask;
-	__u32	padding;
+	uint32_t	flags;
+	uint32_t	mode;
+	uint32_t	umask;
+	uint32_t	padding;
 };
 
 struct fuse_open_out {
-	__u64	fh;
-	__u32	open_flags;
-	__u32	padding;
+	uint64_t	fh;
+	uint32_t	open_flags;
+	uint32_t	padding;
 };
 
 struct fuse_release_in {
-	__u64	fh;
-	__u32	flags;
-	__u32	release_flags;
-	__u64	lock_owner;
+	uint64_t	fh;
+	uint32_t	flags;
+	uint32_t	release_flags;
+	uint64_t	lock_owner;
 };
 
 struct fuse_flush_in {
-	__u64	fh;
-	__u32	unused;
-	__u32	padding;
-	__u64	lock_owner;
+	uint64_t	fh;
+	uint32_t	unused;
+	uint32_t	padding;
+	uint64_t	lock_owner;
 };
 
 struct fuse_read_in {
-	__u64	fh;
-	__u64	offset;
-	__u32	size;
-	__u32	read_flags;
-	__u64	lock_owner;
-	__u32	flags;
-	__u32	padding;
+	uint64_t	fh;
+	uint64_t	offset;
+	uint32_t	size;
+	uint32_t	read_flags;
+	uint64_t	lock_owner;
+	uint32_t	flags;
+	uint32_t	padding;
 };
 
 #define FUSE_COMPAT_WRITE_IN_SIZE 24
 
 struct fuse_write_in {
-	__u64	fh;
-	__u64	offset;
-	__u32	size;
-	__u32	write_flags;
-	__u64	lock_owner;
-	__u32	flags;
-	__u32	padding;
+	uint64_t	fh;
+	uint64_t	offset;
+	uint32_t	size;
+	uint32_t	write_flags;
+	uint64_t	lock_owner;
+	uint32_t	flags;
+	uint32_t	padding;
 };
 
 struct fuse_write_out {
-	__u32	size;
-	__u32	padding;
+	uint32_t	size;
+	uint32_t	padding;
 };
 
 #define FUSE_COMPAT_STATFS_SIZE 48
@@ -417,32 +510,32 @@ struct fuse_statfs_out {
 };
 
 struct fuse_fsync_in {
-	__u64	fh;
-	__u32	fsync_flags;
-	__u32	padding;
+	uint64_t	fh;
+	uint32_t	fsync_flags;
+	uint32_t	padding;
 };
 
 struct fuse_setxattr_in {
-	__u32	size;
-	__u32	flags;
+	uint32_t	size;
+	uint32_t	flags;
 };
 
 struct fuse_getxattr_in {
-	__u32	size;
-	__u32	padding;
+	uint32_t	size;
+	uint32_t	padding;
 };
 
 struct fuse_getxattr_out {
-	__u32	size;
-	__u32	padding;
+	uint32_t	size;
+	uint32_t	padding;
 };
 
 struct fuse_lk_in {
-	__u64	fh;
-	__u64	owner;
+	uint64_t	fh;
+	uint64_t	owner;
 	struct fuse_file_lock lk;
-	__u32	lk_flags;
-	__u32	padding;
+	uint32_t	lk_flags;
+	uint32_t	padding;
 };
 
 struct fuse_lk_out {
@@ -450,129 +543,135 @@ struct fuse_lk_out {
 };
 
 struct fuse_access_in {
-	__u32	mask;
-	__u32	padding;
+	uint32_t	mask;
+	uint32_t	padding;
 };
 
 struct fuse_init_in {
-	__u32	major;
-	__u32	minor;
-	__u32	max_readahead;
-	__u32	flags;
+	uint32_t	major;
+	uint32_t	minor;
+	uint32_t	max_readahead;
+	uint32_t	flags;
 };
 
 struct fuse_init_out {
-	__u32	major;
-	__u32	minor;
-	__u32	max_readahead;
-	__u32	flags;
-	__u16   max_background;
-	__u16   congestion_threshold;
-	__u32	max_write;
+	uint32_t	major;
+	uint32_t	minor;
+	uint32_t	max_readahead;
+	uint32_t	flags;
+	uint16_t	max_background;
+	uint16_t	congestion_threshold;
+	uint32_t	max_write;
 };
 
 #define CUSE_INIT_INFO_MAX 4096
 
 struct cuse_init_in {
-	__u32	major;
-	__u32	minor;
-	__u32	unused;
-	__u32	flags;
+	uint32_t	major;
+	uint32_t	minor;
+	uint32_t	unused;
+	uint32_t	flags;
 };
 
 struct cuse_init_out {
-	__u32	major;
-	__u32	minor;
-	__u32	unused;
-	__u32	flags;
-	__u32	max_read;
-	__u32	max_write;
-	__u32	dev_major;		/* chardev major */
-	__u32	dev_minor;		/* chardev minor */
-	__u32	spare[10];
+	uint32_t	major;
+	uint32_t	minor;
+	uint32_t	unused;
+	uint32_t	flags;
+	uint32_t	max_read;
+	uint32_t	max_write;
+	uint32_t	dev_major;		/* chardev major */
+	uint32_t	dev_minor;		/* chardev minor */
+	uint32_t	spare[10];
 };
 
 struct fuse_interrupt_in {
-	__u64	unique;
+	uint64_t	unique;
 };
 
 struct fuse_bmap_in {
-	__u64	block;
-	__u32	blocksize;
-	__u32	padding;
+	uint64_t	block;
+	uint32_t	blocksize;
+	uint32_t	padding;
 };
 
 struct fuse_bmap_out {
-	__u64	block;
+	uint64_t	block;
 };
 
 struct fuse_ioctl_in {
-	__u64	fh;
-	__u32	flags;
-	__u32	cmd;
-	__u64	arg;
-	__u32	in_size;
-	__u32	out_size;
+	uint64_t	fh;
+	uint32_t	flags;
+	uint32_t	cmd;
+	uint64_t	arg;
+	uint32_t	in_size;
+	uint32_t	out_size;
+};
+
+struct fuse_ioctl_iovec {
+	uint64_t	base;
+	uint64_t	len;
 };
 
 struct fuse_ioctl_out {
-	__s32	result;
-	__u32	flags;
-	__u32	in_iovs;
-	__u32	out_iovs;
+	int32_t		result;
+	uint32_t	flags;
+	uint32_t	in_iovs;
+	uint32_t	out_iovs;
 };
 
 struct fuse_poll_in {
-	__u64	fh;
-	__u64	kh;
-	__u32	flags;
-	__u32   padding;
+	uint64_t	fh;
+	uint64_t	kh;
+	uint32_t	flags;
+	uint32_t	events;
 };
 
 struct fuse_poll_out {
-	__u32	revents;
-	__u32	padding;
+	uint32_t	revents;
+	uint32_t	padding;
 };
 
 struct fuse_notify_poll_wakeup_out {
-	__u64	kh;
+	uint64_t	kh;
 };
 
 struct fuse_fallocate_in {
-	__u64	fh;
-	__u64	offset;
-	__u64	length;
-	__u32	mode;
-	__u32	padding;
+	uint64_t	fh;
+	uint64_t	offset;
+	uint64_t	length;
+	uint32_t	mode;
+	uint32_t	padding;
 };
 
 struct fuse_in_header {
-	__u32	len;
-	__u32	opcode;
-	__u64	unique;
-	__u64	nodeid;
-	__u32	uid;
-	__u32	gid;
-	__u32	pid;
-	__u32	padding;
+	uint32_t	len;
+	uint32_t	opcode;
+	uint64_t	unique;
+	uint64_t	nodeid;
+	uint32_t	uid;
+	uint32_t	gid;
+	uint32_t	pid;
+	uint32_t	padding;
 };
 
 struct fuse_out_header {
-	__u32	len;
-	__s32	error;
-	__u64	unique;
+	uint32_t	len;
+	int32_t		error;
+	uint64_t	unique;
 };
 
 struct fuse_dirent {
-	__u64	ino;
-	__u64	off;
-	__u32	namelen;
-	__u32	type;
+	uint64_t	ino;
+	uint64_t	off;
+	uint32_t	namelen;
+	uint32_t	type;
 	char name[];
 };
 
 #define FUSE_NAME_OFFSET offsetof(struct fuse_dirent, name)
-#define FUSE_DIRENT_ALIGN(x) (((x) + sizeof(__u64) - 1) & ~(sizeof(__u64) - 1))
+#define FUSE_DIRENT_ALIGN(x) \
+	(((x) + sizeof(uint64_t) - 1) & ~(sizeof(uint64_t) - 1))
 #define FUSE_DIRENT_SIZE(d) \
 	FUSE_DIRENT_ALIGN(FUSE_NAME_OFFSET + (d)->namelen)
 
@@ -587,15 +686,47 @@ struct fuse_direntplus {
 	FUSE_DIRENT_ALIGN(FUSE_NAME_OFFSET_DIRENTPLUS + (d)->dirent.namelen)
 
 struct fuse_notify_inval_inode_out {
-	__u64	ino;
-	__s64	off;
-	__s64	len;
+	uint64_t	ino;
+	int64_t		off;
+	int64_t		len;
 };
 
 struct fuse_notify_inval_entry_out {
-	__u64	parent;
-	__u32	namelen;
-	__u32	padding;
+	uint64_t	parent;
+	uint32_t	namelen;
+	uint32_t	padding;
+};
+
+struct fuse_notify_delete_out {
+	uint64_t	parent;
+	uint64_t	child;
+	uint32_t	namelen;
+	uint32_t	padding;
+};
+
+struct fuse_notify_store_out {
+	uint64_t	nodeid;
+	uint64_t	offset;
+	uint32_t	size;
+	uint32_t	padding;
+};
+
+struct fuse_notify_retrieve_out {
+	uint64_t	notify_unique;
+	uint64_t	nodeid;
+	uint64_t	offset;
+	uint32_t	size;
+	uint32_t	padding;
+};
+
+/* Matches the size of fuse_write_in */
+struct fuse_notify_retrieve_in {
+	uint64_t	dummy1;
+	uint64_t	offset;
+	uint32_t	size;
+	uint32_t	dummy2;
+	uint64_t	dummy3;
+	uint64_t	dummy4;
 };
 
 #endif /* _LINUX_FUSE_H */
diff --git a/contrib/qemu/block.c b/contrib/qemu/block.c
new file mode 100644
index 000000000..b56024113
--- /dev/null
+++ b/contrib/qemu/block.c
@@ -0,0 +1,4604 @@
+/*
+ * QEMU System Emulator block driver
+ *
+ * Copyright (c) 2003 Fabrice Bellard
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+#include "config-host.h"
+#include "qemu-common.h"
+#include "trace.h"
+#include "monitor/monitor.h"
+#include "block/block_int.h"
+#include "block/blockjob.h"
+#include "qemu/module.h"
+#include "qapi/qmp/qjson.h"
+#include "sysemu/sysemu.h"
+#include "qemu/notify.h"
+#include "block/coroutine.h"
+#include "qmp-commands.h"
+#include "qemu/timer.h"
+
+#ifdef CONFIG_BSD
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <sys/ioctl.h>
+#include <sys/queue.h>
+#ifndef __DragonFly__
+#include <sys/disk.h>
+#endif
+#endif
+
+#ifdef _WIN32
+#include <windows.h>
+#endif
+
+#define NOT_DONE 0x7fffffff /* used while emulated sync operation in progress */
+
+typedef enum {
+    BDRV_REQ_COPY_ON_READ = 0x1,
+    BDRV_REQ_ZERO_WRITE   = 0x2,
+} BdrvRequestFlags;
+
+static void bdrv_dev_change_media_cb(BlockDriverState *bs, bool load);
+static BlockDriverAIOCB *bdrv_aio_readv_em(BlockDriverState *bs,
+        int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
+        BlockDriverCompletionFunc *cb, void *opaque);
+static BlockDriverAIOCB *bdrv_aio_writev_em(BlockDriverState *bs,
+        int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
+        BlockDriverCompletionFunc *cb, void *opaque);
+static int coroutine_fn bdrv_co_readv_em(BlockDriverState *bs,
+                                         int64_t sector_num, int nb_sectors,
+                                         QEMUIOVector *iov);
+static int coroutine_fn bdrv_co_writev_em(BlockDriverState *bs,
+                                         int64_t sector_num, int nb_sectors,
+                                         QEMUIOVector *iov);
+static int coroutine_fn bdrv_co_do_readv(BlockDriverState *bs,
+    int64_t sector_num, int nb_sectors, QEMUIOVector *qiov,
+    BdrvRequestFlags flags);
+static int coroutine_fn bdrv_co_do_writev(BlockDriverState *bs,
+    int64_t sector_num, int nb_sectors, QEMUIOVector *qiov,
+    BdrvRequestFlags flags);
+static BlockDriverAIOCB *bdrv_co_aio_rw_vector(BlockDriverState *bs,
+                                               int64_t sector_num,
+                                               QEMUIOVector *qiov,
+                                               int nb_sectors,
+                                               BlockDriverCompletionFunc *cb,
+                                               void *opaque,
+                                               bool is_write);
+static void coroutine_fn bdrv_co_do_rw(void *opaque);
+static int coroutine_fn bdrv_co_do_write_zeroes(BlockDriverState *bs,
+    int64_t sector_num, int nb_sectors);
+
+static bool bdrv_exceed_bps_limits(BlockDriverState *bs, int nb_sectors,
+        bool is_write, double elapsed_time, uint64_t *wait);
+static bool bdrv_exceed_iops_limits(BlockDriverState *bs, bool is_write,
+        double elapsed_time, uint64_t *wait);
+static bool bdrv_exceed_io_limits(BlockDriverState *bs, int nb_sectors,
+        bool is_write, int64_t *wait);
+
+static QTAILQ_HEAD(, BlockDriverState) bdrv_states =
+    QTAILQ_HEAD_INITIALIZER(bdrv_states);
+
+static QLIST_HEAD(, BlockDriver) bdrv_drivers =
+    QLIST_HEAD_INITIALIZER(bdrv_drivers);
+
+/* If non-zero, use only whitelisted block drivers */
+static int use_bdrv_whitelist;
+
+#ifdef _WIN32
+static int is_windows_drive_prefix(const char *filename)
+{
+    return (((filename[0] >= 'a' && filename[0] <= 'z') ||
+             (filename[0] >= 'A' && filename[0] <= 'Z')) &&
+            filename[1] == ':');
+}
+
+int is_windows_drive(const char *filename)
+{
+    if (is_windows_drive_prefix(filename) &&
+        filename[2] == '\0')
+        return 1;
+    if (strstart(filename, "\\\\.\\", NULL) ||
+        strstart(filename, "//./", NULL))
+        return 1;
+    return 0;
+}
+#endif
+
+/* throttling disk I/O limits */
+void bdrv_io_limits_disable(BlockDriverState *bs)
+{
+    bs->io_limits_enabled = false;
+
+    while (qemu_co_queue_next(&bs->throttled_reqs));
+
+    if (bs->block_timer) {
+        qemu_del_timer(bs->block_timer);
+        qemu_free_timer(bs->block_timer);
+        bs->block_timer = NULL;
+    }
+
+    bs->slice_start = 0;
+    bs->slice_end   = 0;
+}
+
+static void bdrv_block_timer(void *opaque)
+{
+    BlockDriverState *bs = opaque;
+
+    qemu_co_queue_next(&bs->throttled_reqs);
+}
+
+void bdrv_io_limits_enable(BlockDriverState *bs)
+{
+    qemu_co_queue_init(&bs->throttled_reqs);
+    bs->block_timer = qemu_new_timer_ns(vm_clock, bdrv_block_timer, bs);
+    bs->io_limits_enabled = true;
+}
+
+bool bdrv_io_limits_enabled(BlockDriverState *bs)
+{
+    BlockIOLimit *io_limits = &bs->io_limits;
+    return io_limits->bps[BLOCK_IO_LIMIT_READ]
+         || io_limits->bps[BLOCK_IO_LIMIT_WRITE]
+         || io_limits->bps[BLOCK_IO_LIMIT_TOTAL]
+         || io_limits->iops[BLOCK_IO_LIMIT_READ]
+         || io_limits->iops[BLOCK_IO_LIMIT_WRITE]
+         || io_limits->iops[BLOCK_IO_LIMIT_TOTAL];
+}
+
+static void bdrv_io_limits_intercept(BlockDriverState *bs,
+                                     bool is_write, int nb_sectors)
+{
+    int64_t wait_time = -1;
+
+    if (!qemu_co_queue_empty(&bs->throttled_reqs)) {
+        qemu_co_queue_wait(&bs->throttled_reqs);
+    }
+
+    /* In fact, we hope to keep each request's timing, in FIFO mode. The next
+     * throttled requests will not be dequeued until the current request is
+     * allowed to be serviced. So if the current request still exceeds the
+     * limits, it will be inserted to the head. All requests followed it will
+     * be still in throttled_reqs queue.
+     */
+
+    while (bdrv_exceed_io_limits(bs, nb_sectors, is_write, &wait_time)) {
+        qemu_mod_timer(bs->block_timer,
+                       wait_time + qemu_get_clock_ns(vm_clock));
+        qemu_co_queue_wait_insert_head(&bs->throttled_reqs);
+    }
+
+    qemu_co_queue_next(&bs->throttled_reqs);
+}
+
+/* check if the path starts with "<protocol>:" */
+static int path_has_protocol(const char *path)
+{
+    const char *p;
+
+#ifdef _WIN32
+    if (is_windows_drive(path) ||
+        is_windows_drive_prefix(path)) {
+        return 0;
+    }
+    p = path + strcspn(path, ":/\\");
+#else
+    p = path + strcspn(path, ":/");
+#endif
+
+    return *p == ':';
+}
+
+int path_is_absolute(const char *path)
+{
+#ifdef _WIN32
+    /* specific case for names like: "\\.\d:" */
+    if (is_windows_drive(path) || is_windows_drive_prefix(path)) {
+        return 1;
+    }
+    return (*path == '/' || *path == '\\');
+#else
+    return (*path == '/');
+#endif
+}
+
+/* if filename is absolute, just copy it to dest. Otherwise, build a
+   path to it by considering it is relative to base_path. URL are
+   supported. */
+void path_combine(char *dest, int dest_size,
+                  const char *base_path,
+                  const char *filename)
+{
+    const char *p, *p1;
+    int len;
+
+    if (dest_size <= 0)
+        return;
+    if (path_is_absolute(filename)) {
+        pstrcpy(dest, dest_size, filename);
+    } else {
+        p = strchr(base_path, ':');
+        if (p)
+            p++;
+        else
+            p = base_path;
+        p1 = strrchr(base_path, '/');
+#ifdef _WIN32
+        {
+            const char *p2;
+            p2 = strrchr(base_path, '\\');
+            if (!p1 || p2 > p1)
+                p1 = p2;
+        }
+#endif
+        if (p1)
+            p1++;
+        else
+            p1 = base_path;
+        if (p1 > p)
+            p = p1;
+        len = p - base_path;
+        if (len > dest_size - 1)
+            len = dest_size - 1;
+        memcpy(dest, base_path, len);
+        dest[len] = '\0';
+        pstrcat(dest, dest_size, filename);
+    }
+}
+
+void bdrv_get_full_backing_filename(BlockDriverState *bs, char *dest, size_t sz)
+{
+    if (bs->backing_file[0] == '\0' || path_has_protocol(bs->backing_file)) {
+        pstrcpy(dest, sz, bs->backing_file);
+    } else {
+        path_combine(dest, sz, bs->filename, bs->backing_file);
+    }
+}
+
+void bdrv_register(BlockDriver *bdrv)
+{
+    /* Block drivers without coroutine functions need emulation */
+    if (!bdrv->bdrv_co_readv) {
+        bdrv->bdrv_co_readv = bdrv_co_readv_em;
+        bdrv->bdrv_co_writev = bdrv_co_writev_em;
+
+        /* bdrv_co_readv_em()/brdv_co_writev_em() work in terms of aio, so if
+         * the block driver lacks aio we need to emulate that too.
+         */
+        if (!bdrv->bdrv_aio_readv) {
+            /* add AIO emulation layer */
+            bdrv->bdrv_aio_readv = bdrv_aio_readv_em;
+            bdrv->bdrv_aio_writev = bdrv_aio_writev_em;
+        }
+    }
+
+    QLIST_INSERT_HEAD(&bdrv_drivers, bdrv, list);
+}
+
+/* create a new block device (by default it is empty) */
+BlockDriverState *bdrv_new(const char *device_name)
+{
+    BlockDriverState *bs;
+
+    bs = g_malloc0(sizeof(BlockDriverState));
+    pstrcpy(bs->device_name, sizeof(bs->device_name), device_name);
+    if (device_name[0] != '\0') {
+        QTAILQ_INSERT_TAIL(&bdrv_states, bs, list);
+    }
+    bdrv_iostatus_disable(bs);
+    notifier_list_init(&bs->close_notifiers);
+    notifier_with_return_list_init(&bs->before_write_notifiers);
+
+    return bs;
+}
+
+void bdrv_add_close_notifier(BlockDriverState *bs, Notifier *notify)
+{
+    notifier_list_add(&bs->close_notifiers, notify);
+}
+
+BlockDriver *bdrv_find_format(const char *format_name)
+{
+    BlockDriver *drv1;
+    QLIST_FOREACH(drv1, &bdrv_drivers, list) {
+        if (!strcmp(drv1->format_name, format_name)) {
+            return drv1;
+        }
+    }
+    return NULL;
+}
+
+static int bdrv_is_whitelisted(BlockDriver *drv, bool read_only)
+{
+    static const char *whitelist_rw[] = {
+        CONFIG_BDRV_RW_WHITELIST
+    };
+    static const char *whitelist_ro[] = {
+        CONFIG_BDRV_RO_WHITELIST
+    };
+    const char **p;
+
+    if (!whitelist_rw[0] && !whitelist_ro[0]) {
+        return 1;               /* no whitelist, anything goes */
+    }
+
+    for (p = whitelist_rw; *p; p++) {
+        if (!strcmp(drv->format_name, *p)) {
+            return 1;
+        }
+    }
+    if (read_only) {
+        for (p = whitelist_ro; *p; p++) {
+            if (!strcmp(drv->format_name, *p)) {
+                return 1;
+            }
+        }
+    }
+    return 0;
+}
+
+BlockDriver *bdrv_find_whitelisted_format(const char *format_name,
+                                          bool read_only)
+{
+    BlockDriver *drv = bdrv_find_format(format_name);
+    return drv && bdrv_is_whitelisted(drv, read_only) ? drv : NULL;
+}
+
+typedef struct CreateCo {
+    BlockDriver *drv;
+    char *filename;
+    QEMUOptionParameter *options;
+    int ret;
+} CreateCo;
+
+static void coroutine_fn bdrv_create_co_entry(void *opaque)
+{
+    CreateCo *cco = opaque;
+    assert(cco->drv);
+
+    cco->ret = cco->drv->bdrv_create(cco->filename, cco->options);
+}
+
+int bdrv_create(BlockDriver *drv, const char* filename,
+    QEMUOptionParameter *options)
+{
+    int ret;
+
+    Coroutine *co;
+    CreateCo cco = {
+        .drv = drv,
+        .filename = g_strdup(filename),
+        .options = options,
+        .ret = NOT_DONE,
+    };
+
+    if (!drv->bdrv_create) {
+        ret = -ENOTSUP;
+        goto out;
+    }
+
+    if (qemu_in_coroutine()) {
+        /* Fast-path if already in coroutine context */
+        bdrv_create_co_entry(&cco);
+    } else {
+        co = qemu_coroutine_create(bdrv_create_co_entry);
+        qemu_coroutine_enter(co, &cco);
+        while (cco.ret == NOT_DONE) {
+            qemu_aio_wait();
+        }
+    }
+
+    ret = cco.ret;
+
+out:
+    g_free(cco.filename);
+    return ret;
+}
+
+int bdrv_create_file(const char* filename, QEMUOptionParameter *options)
+{
+    BlockDriver *drv;
+
+    drv = bdrv_find_protocol(filename, true);
+    if (drv == NULL) {
+        return -ENOENT;
+    }
+
+    return bdrv_create(drv, filename, options);
+}
+
+/*
+ * Create a uniquely-named empty temporary file.
+ * Return 0 upon success, otherwise a negative errno value.
+ */
+int get_tmp_filename(char *filename, int size)
+{
+#ifdef _WIN32
+    char temp_dir[MAX_PATH];
+    /* GetTempFileName requires that its output buffer (4th param)
+       have length MAX_PATH or greater.  */
+    assert(size >= MAX_PATH);
+    return (GetTempPath(MAX_PATH, temp_dir)
+            && GetTempFileName(temp_dir, "qem", 0, filename)
+            ? 0 : -GetLastError());
+#else
+    int fd;
+    const char *tmpdir;
+    tmpdir = getenv("TMPDIR");
+    if (!tmpdir)
+        tmpdir = "/tmp";
+    if (snprintf(filename, size, "%s/vl.XXXXXX", tmpdir) >= size) {
+        return -EOVERFLOW;
+    }
+    fd = mkstemp(filename);
+    if (fd < 0) {
+        return -errno;
+    }
+    if (close(fd) != 0) {
+        unlink(filename);
+        return -errno;
+    }
+    return 0;
+#endif
+}
+
+/*
+ * Detect host devices. By convention, /dev/cdrom[N] is always
+ * recognized as a host CDROM.
+ */
+static BlockDriver *find_hdev_driver(const char *filename)
+{
+    int score_max = 0, score;
+    BlockDriver *drv = NULL, *d;
+
+    QLIST_FOREACH(d, &bdrv_drivers, list) {
+        if (d->bdrv_probe_device) {
+            score = d->bdrv_probe_device(filename);
+            if (score > score_max) {
+                score_max = score;
+                drv = d;
+            }
+        }
+    }
+
+    return drv;
+}
+
+BlockDriver *bdrv_find_protocol(const char *filename,
+                                bool allow_protocol_prefix)
+{
+    BlockDriver *drv1;
+    char protocol[128];
+    int len;
+    const char *p;
+
+    /* TODO Drivers without bdrv_file_open must be specified explicitly */
+
+    /*
+     * XXX(hch): we really should not let host device detection
+     * override an explicit protocol specification, but moving this
+     * later breaks access to device names with colons in them.
+     * Thanks to the brain-dead persistent naming schemes on udev-
+     * based Linux systems those actually are quite common.
+     */
+    drv1 = find_hdev_driver(filename);
+    if (drv1) {
+        return drv1;
+    }
+
+    if (!path_has_protocol(filename) || !allow_protocol_prefix) {
+        return bdrv_find_format("file");
+    }
+
+    p = strchr(filename, ':');
+    assert(p != NULL);
+    len = p - filename;
+    if (len > sizeof(protocol) - 1)
+        len = sizeof(protocol) - 1;
+    memcpy(protocol, filename, len);
+    protocol[len] = '\0';
+    QLIST_FOREACH(drv1, &bdrv_drivers, list) {
+        if (drv1->protocol_name &&
+            !strcmp(drv1->protocol_name, protocol)) {
+            return drv1;
+        }
+    }
+    return NULL;
+}
+
+static int find_image_format(BlockDriverState *bs, const char *filename,
+                             BlockDriver **pdrv)
+{
+    int score, score_max;
+    BlockDriver *drv1, *drv;
+    uint8_t buf[2048];
+    int ret = 0;
+
+    /* Return the raw BlockDriver * to scsi-generic devices or empty drives */
+    if (bs->sg || !bdrv_is_inserted(bs) || bdrv_getlength(bs) == 0) {
+        drv = bdrv_find_format("raw");
+        if (!drv) {
+            ret = -ENOENT;
+        }
+        *pdrv = drv;
+        return ret;
+    }
+
+    ret = bdrv_pread(bs, 0, buf, sizeof(buf));
+    if (ret < 0) {
+        *pdrv = NULL;
+        return ret;
+    }
+
+    score_max = 0;
+    drv = NULL;
+    QLIST_FOREACH(drv1, &bdrv_drivers, list) {
+        if (drv1->bdrv_probe) {
+            score = drv1->bdrv_probe(buf, ret, filename);
+            if (score > score_max) {
+                score_max = score;
+                drv = drv1;
+            }
+        }
+    }
+    if (!drv) {
+        ret = -ENOENT;
+    }
+    *pdrv = drv;
+    return ret;
+}
+
+/**
+ * Set the current 'total_sectors' value
+ */
+static int refresh_total_sectors(BlockDriverState *bs, int64_t hint)
+{
+    BlockDriver *drv = bs->drv;
+
+    /* Do not attempt drv->bdrv_getlength() on scsi-generic devices */
+    if (bs->sg)
+        return 0;
+
+    /* query actual device if possible, otherwise just trust the hint */
+    if (drv->bdrv_getlength) {
+        int64_t length = drv->bdrv_getlength(bs);
+        if (length < 0) {
+            return length;
+        }
+        hint = length >> BDRV_SECTOR_BITS;
+    }
+
+    bs->total_sectors = hint;
+    return 0;
+}
+
+/**
+ * Set open flags for a given discard mode
+ *
+ * Return 0 on success, -1 if the discard mode was invalid.
+ */
+int bdrv_parse_discard_flags(const char *mode, int *flags)
+{
+    *flags &= ~BDRV_O_UNMAP;
+
+    if (!strcmp(mode, "off") || !strcmp(mode, "ignore")) {
+        /* do nothing */
+    } else if (!strcmp(mode, "on") || !strcmp(mode, "unmap")) {
+        *flags |= BDRV_O_UNMAP;
+    } else {
+        return -1;
+    }
+
+    return 0;
+}
+
+/**
+ * Set open flags for a given cache mode
+ *
+ * Return 0 on success, -1 if the cache mode was invalid.
+ */
+int bdrv_parse_cache_flags(const char *mode, int *flags)
+{
+    *flags &= ~BDRV_O_CACHE_MASK;
+
+    if (!strcmp(mode, "off") || !strcmp(mode, "none")) {
+        *flags |= BDRV_O_NOCACHE | BDRV_O_CACHE_WB;
+    } else if (!strcmp(mode, "directsync")) {
+        *flags |= BDRV_O_NOCACHE;
+    } else if (!strcmp(mode, "writeback")) {
+        *flags |= BDRV_O_CACHE_WB;
+    } else if (!strcmp(mode, "unsafe")) {
+        *flags |= BDRV_O_CACHE_WB;
+        *flags |= BDRV_O_NO_FLUSH;
+    } else if (!strcmp(mode, "writethrough")) {
+        /* this is the default */
+    } else {
+        return -1;
+    }
+
+    return 0;
+}
+
+/**
+ * The copy-on-read flag is actually a reference count so multiple users may
+ * use the feature without worrying about clobbering its previous state.
+ * Copy-on-read stays enabled until all users have called to disable it.
+ */
+void bdrv_enable_copy_on_read(BlockDriverState *bs)
+{
+    bs->copy_on_read++;
+}
+
+void bdrv_disable_copy_on_read(BlockDriverState *bs)
+{
+    assert(bs->copy_on_read > 0);
+    bs->copy_on_read--;
+}
+
+static int bdrv_open_flags(BlockDriverState *bs, int flags)
+{
+    int open_flags = flags | BDRV_O_CACHE_WB;
+
+    /*
+     * Clear flags that are internal to the block layer before opening the
+     * image.
+     */
+    open_flags &= ~(BDRV_O_SNAPSHOT | BDRV_O_NO_BACKING);
+
+    /*
+     * Snapshots should be writable.
+     */
+    if (bs->is_temporary) {
+        open_flags |= BDRV_O_RDWR;
+    }
+
+    return open_flags;
+}
+
+/*
+ * Common part for opening disk images and files
+ *
+ * Removes all processed options from *options.
+ */
+static int bdrv_open_common(BlockDriverState *bs, BlockDriverState *file,
+    QDict *options, int flags, BlockDriver *drv)
+{
+    int ret, open_flags;
+    const char *filename;
+
+    assert(drv != NULL);
+    assert(bs->file == NULL);
+    assert(options != NULL && bs->options != options);
+
+    if (file != NULL) {
+        filename = file->filename;
+    } else {
+        filename = qdict_get_try_str(options, "filename");
+    }
+
+    trace_bdrv_open_common(bs, filename ?: "", flags, drv->format_name);
+
+    /* bdrv_open() with directly using a protocol as drv. This layer is already
+     * opened, so assign it to bs (while file becomes a closed BlockDriverState)
+     * and return immediately. */
+    if (file != NULL && drv->bdrv_file_open) {
+        bdrv_swap(file, bs);
+        return 0;
+    }
+
+    bs->open_flags = flags;
+    bs->buffer_alignment = 512;
+    open_flags = bdrv_open_flags(bs, flags);
+    bs->read_only = !(open_flags & BDRV_O_RDWR);
+
+    if (use_bdrv_whitelist && !bdrv_is_whitelisted(drv, bs->read_only)) {
+        return -ENOTSUP;
+    }
+
+    assert(bs->copy_on_read == 0); /* bdrv_new() and bdrv_close() make it so */
+    if (!bs->read_only && (flags & BDRV_O_COPY_ON_READ)) {
+        bdrv_enable_copy_on_read(bs);
+    }
+
+    if (filename != NULL) {
+        pstrcpy(bs->filename, sizeof(bs->filename), filename);
+    } else {
+        bs->filename[0] = '\0';
+    }
+
+    bs->drv = drv;
+    bs->opaque = g_malloc0(drv->instance_size);
+
+    bs->enable_write_cache = !!(flags & BDRV_O_CACHE_WB);
+
+    /* Open the image, either directly or using a protocol */
+    if (drv->bdrv_file_open) {
+        assert(file == NULL);
+        assert(drv->bdrv_parse_filename || filename != NULL);
+        ret = drv->bdrv_file_open(bs, options, open_flags);
+    } else {
+        if (file == NULL) {
+            qerror_report(ERROR_CLASS_GENERIC_ERROR, "Can't use '%s' as a "
+                          "block driver for the protocol level",
+                          drv->format_name);
+            ret = -EINVAL;
+            goto free_and_fail;
+        }
+        assert(file != NULL);
+        bs->file = file;
+        ret = drv->bdrv_open(bs, options, open_flags);
+    }
+
+    if (ret < 0) {
+        goto free_and_fail;
+    }
+
+    ret = refresh_total_sectors(bs, bs->total_sectors);
+    if (ret < 0) {
+        goto free_and_fail;
+    }
+
+#ifndef _WIN32
+    if (bs->is_temporary) {
+        assert(filename != NULL);
+        unlink(filename);
+    }
+#endif
+    return 0;
+
+free_and_fail:
+    bs->file = NULL;
+    g_free(bs->opaque);
+    bs->opaque = NULL;
+    bs->drv = NULL;
+    return ret;
+}
+
+/*
+ * Opens a file using a protocol (file, host_device, nbd, ...)
+ *
+ * options is a QDict of options to pass to the block drivers, or NULL for an
+ * empty set of options. The reference to the QDict belongs to the block layer
+ * after the call (even on failure), so if the caller intends to reuse the
+ * dictionary, it needs to use QINCREF() before calling bdrv_file_open.
+ */
+int bdrv_file_open(BlockDriverState **pbs, const char *filename,
+                   QDict *options, int flags)
+{
+    BlockDriverState *bs;
+    BlockDriver *drv;
+    const char *drvname;
+    bool allow_protocol_prefix = false;
+    int ret;
+
+    /* NULL means an empty set of options */
+    if (options == NULL) {
+        options = qdict_new();
+    }
+
+    bs = bdrv_new("");
+    bs->options = options;
+    options = qdict_clone_shallow(options);
+
+    /* Fetch the file name from the options QDict if necessary */
+    if (!filename) {
+        filename = qdict_get_try_str(options, "filename");
+    } else if (filename && !qdict_haskey(options, "filename")) {
+        qdict_put(options, "filename", qstring_from_str(filename));
+        allow_protocol_prefix = true;
+    } else {
+        qerror_report(ERROR_CLASS_GENERIC_ERROR, "Can't specify 'file' and "
+                      "'filename' options at the same time");
+        ret = -EINVAL;
+        goto fail;
+    }
+
+    /* Find the right block driver */
+    drvname = qdict_get_try_str(options, "driver");
+    if (drvname) {
+        drv = bdrv_find_whitelisted_format(drvname, !(flags & BDRV_O_RDWR));
+        qdict_del(options, "driver");
+    } else if (filename) {
+        drv = bdrv_find_protocol(filename, allow_protocol_prefix);
+        if (!drv) {
+            qerror_report(ERROR_CLASS_GENERIC_ERROR, "Unknown protocol");
+        }
+    } else {
+        qerror_report(ERROR_CLASS_GENERIC_ERROR,
+                      "Must specify either driver or file");
+        drv = NULL;
+    }
+
+    if (!drv) {
+        ret = -ENOENT;
+        goto fail;
+    }
+
+    /* Parse the filename and open it */
+    if (drv->bdrv_parse_filename && filename) {
+        Error *local_err = NULL;
+        drv->bdrv_parse_filename(filename, options, &local_err);
+        if (error_is_set(&local_err)) {
+            qerror_report_err(local_err);
+            error_free(local_err);
+            ret = -EINVAL;
+            goto fail;
+        }
+        qdict_del(options, "filename");
+    } else if (!drv->bdrv_parse_filename && !filename) {
+        qerror_report(ERROR_CLASS_GENERIC_ERROR,
+                      "The '%s' block driver requires a file name",
+                      drv->format_name);
+        ret = -EINVAL;
+        goto fail;
+    }
+
+    ret = bdrv_open_common(bs, NULL, options, flags, drv);
+    if (ret < 0) {
+        goto fail;
+    }
+
+    /* Check if any unknown options were used */
+    if (qdict_size(options) != 0) {
+        const QDictEntry *entry = qdict_first(options);
+        qerror_report(ERROR_CLASS_GENERIC_ERROR, "Block protocol '%s' doesn't "
+                      "support the option '%s'",
+                      drv->format_name, entry->key);
+        ret = -EINVAL;
+        goto fail;
+    }
+    QDECREF(options);
+
+    bs->growable = 1;
+    *pbs = bs;
+    return 0;
+
+fail:
+    QDECREF(options);
+    if (!bs->drv) {
+        QDECREF(bs->options);
+    }
+    bdrv_delete(bs);
+    return ret;
+}
+
+/*
+ * Opens the backing file for a BlockDriverState if not yet open
+ *
+ * options is a QDict of options to pass to the block drivers, or NULL for an
+ * empty set of options. The reference to the QDict is transferred to this
+ * function (even on failure), so if the caller intends to reuse the dictionary,
+ * it needs to use QINCREF() before calling bdrv_file_open.
+ */
+int bdrv_open_backing_file(BlockDriverState *bs, QDict *options)
+{
+    char backing_filename[PATH_MAX];
+    int back_flags, ret;
+    BlockDriver *back_drv = NULL;
+
+    if (bs->backing_hd != NULL) {
+        QDECREF(options);
+        return 0;
+    }
+
+    /* NULL means an empty set of options */
+    if (options == NULL) {
+        options = qdict_new();
+    }
+
+    bs->open_flags &= ~BDRV_O_NO_BACKING;
+    if (qdict_haskey(options, "file.filename")) {
+        backing_filename[0] = '\0';
+    } else if (bs->backing_file[0] == '\0' && qdict_size(options) == 0) {
+        QDECREF(options);
+        return 0;
+    }
+
+    bs->backing_hd = bdrv_new("");
+    bdrv_get_full_backing_filename(bs, backing_filename,
+                                   sizeof(backing_filename));
+
+    if (bs->backing_format[0] != '\0') {
+        back_drv = bdrv_find_format(bs->backing_format);
+    }
+
+    /* backing files always opened read-only */
+    back_flags = bs->open_flags & ~(BDRV_O_RDWR | BDRV_O_SNAPSHOT);
+
+    ret = bdrv_open(bs->backing_hd,
+                    *backing_filename ? backing_filename : NULL, options,
+                    back_flags, back_drv);
+    if (ret < 0) {
+        bdrv_delete(bs->backing_hd);
+        bs->backing_hd = NULL;
+        bs->open_flags |= BDRV_O_NO_BACKING;
+        return ret;
+    }
+    return 0;
+}
+
+static void extract_subqdict(QDict *src, QDict **dst, const char *start)
+{
+    const QDictEntry *entry, *next;
+    const char *p;
+
+    *dst = qdict_new();
+    entry = qdict_first(src);
+
+    while (entry != NULL) {
+        next = qdict_next(src, entry);
+        if (strstart(entry->key, start, &p)) {
+            qobject_incref(entry->value);
+            qdict_put_obj(*dst, p, entry->value);
+            qdict_del(src, entry->key);
+        }
+        entry = next;
+    }
+}
+
+/*
+ * Opens a disk image (raw, qcow2, vmdk, ...)
+ *
+ * options is a QDict of options to pass to the block drivers, or NULL for an
+ * empty set of options. The reference to the QDict belongs to the block layer
+ * after the call (even on failure), so if the caller intends to reuse the
+ * dictionary, it needs to use QINCREF() before calling bdrv_open.
+ */
+int bdrv_open(BlockDriverState *bs, const char *filename, QDict *options,
+              int flags, BlockDriver *drv)
+{
+    int ret;
+    /* TODO: extra byte is a hack to ensure MAX_PATH space on Windows. */
+    char tmp_filename[PATH_MAX + 1];
+    BlockDriverState *file = NULL;
+    QDict *file_options = NULL;
+
+    /* NULL means an empty set of options */
+    if (options == NULL) {
+        options = qdict_new();
+    }
+
+    bs->options = options;
+    options = qdict_clone_shallow(options);
+
+    /* For snapshot=on, create a temporary qcow2 overlay */
+    if (flags & BDRV_O_SNAPSHOT) {
+        BlockDriverState *bs1;
+        int64_t total_size;
+        BlockDriver *bdrv_qcow2;
+        QEMUOptionParameter *create_options;
+        char backing_filename[PATH_MAX];
+
+        if (qdict_size(options) != 0) {
+            error_report("Can't use snapshot=on with driver-specific options");
+            ret = -EINVAL;
+            goto fail;
+        }
+        assert(filename != NULL);
+
+        /* if snapshot, we create a temporary backing file and open it
+           instead of opening 'filename' directly */
+
+        /* if there is a backing file, use it */
+        bs1 = bdrv_new("");
+        ret = bdrv_open(bs1, filename, NULL, 0, drv);
+        if (ret < 0) {
+            bdrv_delete(bs1);
+            goto fail;
+        }
+        total_size = bdrv_getlength(bs1) & BDRV_SECTOR_MASK;
+
+        bdrv_delete(bs1);
+
+        ret = get_tmp_filename(tmp_filename, sizeof(tmp_filename));
+        if (ret < 0) {
+            goto fail;
+        }
+
+        /* Real path is meaningless for protocols */
+        if (path_has_protocol(filename)) {
+            snprintf(backing_filename, sizeof(backing_filename),
+                     "%s", filename);
+        } else if (!realpath(filename, backing_filename)) {
+            ret = -errno;
+            goto fail;
+        }
+
+        bdrv_qcow2 = bdrv_find_format("qcow2");
+        create_options = parse_option_parameters("", bdrv_qcow2->create_options,
+                                                 NULL);
+
+        set_option_parameter_int(create_options, BLOCK_OPT_SIZE, total_size);
+        set_option_parameter(create_options, BLOCK_OPT_BACKING_FILE,
+                             backing_filename);
+        if (drv) {
+            set_option_parameter(create_options, BLOCK_OPT_BACKING_FMT,
+                drv->format_name);
+        }
+
+        ret = bdrv_create(bdrv_qcow2, tmp_filename, create_options);
+        free_option_parameters(create_options);
+        if (ret < 0) {
+            goto fail;
+        }
+
+        filename = tmp_filename;
+        drv = bdrv_qcow2;
+        bs->is_temporary = 1;
+    }
+
+    /* Open image file without format layer */
+    if (flags & BDRV_O_RDWR) {
+        flags |= BDRV_O_ALLOW_RDWR;
+    }
+
+    extract_subqdict(options, &file_options, "file.");
+
+    ret = bdrv_file_open(&file, filename, file_options,
+                         bdrv_open_flags(bs, flags | BDRV_O_UNMAP));
+    if (ret < 0) {
+        goto fail;
+    }
+
+    /* Find the right image format driver */
+    if (!drv) {
+        ret = find_image_format(file, filename, &drv);
+    }
+
+    if (!drv) {
+        goto unlink_and_fail;
+    }
+
+    /* Open the image */
+    ret = bdrv_open_common(bs, file, options, flags, drv);
+    if (ret < 0) {
+        goto unlink_and_fail;
+    }
+
+    if (bs->file != file) {
+        bdrv_delete(file);
+        file = NULL;
+    }
+
+    /* If there is a backing file, use it */
+    if ((flags & BDRV_O_NO_BACKING) == 0) {
+        QDict *backing_options;
+
+        extract_subqdict(options, &backing_options, "backing.");
+        ret = bdrv_open_backing_file(bs, backing_options);
+        if (ret < 0) {
+            goto close_and_fail;
+        }
+    }
+
+    /* Check if any unknown options were used */
+    if (qdict_size(options) != 0) {
+        const QDictEntry *entry = qdict_first(options);
+        qerror_report(ERROR_CLASS_GENERIC_ERROR, "Block format '%s' used by "
+            "device '%s' doesn't support the option '%s'",
+            drv->format_name, bs->device_name, entry->key);
+
+        ret = -EINVAL;
+        goto close_and_fail;
+    }
+    QDECREF(options);
+
+    if (!bdrv_key_required(bs)) {
+        bdrv_dev_change_media_cb(bs, true);
+    }
+
+    /* throttling disk I/O limits */
+    if (bs->io_limits_enabled) {
+        bdrv_io_limits_enable(bs);
+    }
+
+    return 0;
+
+unlink_and_fail:
+    if (file != NULL) {
+        bdrv_delete(file);
+    }
+    if (bs->is_temporary) {
+        unlink(filename);
+    }
+fail:
+    QDECREF(bs->options);
+    QDECREF(options);
+    bs->options = NULL;
+    return ret;
+
+close_and_fail:
+    bdrv_close(bs);
+    QDECREF(options);
+    return ret;
+}
+
+typedef struct BlockReopenQueueEntry {
+     bool prepared;
+     BDRVReopenState state;
+     QSIMPLEQ_ENTRY(BlockReopenQueueEntry) entry;
+} BlockReopenQueueEntry;
+
+/*
+ * Adds a BlockDriverState to a simple queue for an atomic, transactional
+ * reopen of multiple devices.
+ *
+ * bs_queue can either be an existing BlockReopenQueue that has had QSIMPLE_INIT
+ * already performed, or alternatively may be NULL a new BlockReopenQueue will
+ * be created and initialized. This newly created BlockReopenQueue should be
+ * passed back in for subsequent calls that are intended to be of the same
+ * atomic 'set'.
+ *
+ * bs is the BlockDriverState to add to the reopen queue.
+ *
+ * flags contains the open flags for the associated bs
+ *
+ * returns a pointer to bs_queue, which is either the newly allocated
+ * bs_queue, or the existing bs_queue being used.
+ *
+ */
+BlockReopenQueue *bdrv_reopen_queue(BlockReopenQueue *bs_queue,
+                                    BlockDriverState *bs, int flags)
+{
+    assert(bs != NULL);
+
+    BlockReopenQueueEntry *bs_entry;
+    if (bs_queue == NULL) {
+        bs_queue = g_new0(BlockReopenQueue, 1);
+        QSIMPLEQ_INIT(bs_queue);
+    }
+
+    if (bs->file) {
+        bdrv_reopen_queue(bs_queue, bs->file, flags);
+    }
+
+    bs_entry = g_new0(BlockReopenQueueEntry, 1);
+    QSIMPLEQ_INSERT_TAIL(bs_queue, bs_entry, entry);
+
+    bs_entry->state.bs = bs;
+    bs_entry->state.flags = flags;
+
+    return bs_queue;
+}
+
+/*
+ * Reopen multiple BlockDriverStates atomically & transactionally.
+ *
+ * The queue passed in (bs_queue) must have been built up previous
+ * via bdrv_reopen_queue().
+ *
+ * Reopens all BDS specified in the queue, with the appropriate
+ * flags.  All devices are prepared for reopen, and failure of any
+ * device will cause all device changes to be abandonded, and intermediate
+ * data cleaned up.
+ *
+ * If all devices prepare successfully, then the changes are committed
+ * to all devices.
+ *
+ */
+int bdrv_reopen_multiple(BlockReopenQueue *bs_queue, Error **errp)
+{
+    int ret = -1;
+    BlockReopenQueueEntry *bs_entry, *next;
+    Error *local_err = NULL;
+
+    assert(bs_queue != NULL);
+
+    bdrv_drain_all();
+
+    QSIMPLEQ_FOREACH(bs_entry, bs_queue, entry) {
+        if (bdrv_reopen_prepare(&bs_entry->state, bs_queue, &local_err)) {
+            error_propagate(errp, local_err);
+            goto cleanup;
+        }
+        bs_entry->prepared = true;
+    }
+
+    /* If we reach this point, we have success and just need to apply the
+     * changes
+     */
+    QSIMPLEQ_FOREACH(bs_entry, bs_queue, entry) {
+        bdrv_reopen_commit(&bs_entry->state);
+    }
+
+    ret = 0;
+
+cleanup:
+    QSIMPLEQ_FOREACH_SAFE(bs_entry, bs_queue, entry, next) {
+        if (ret && bs_entry->prepared) {
+            bdrv_reopen_abort(&bs_entry->state);
+        }
+        g_free(bs_entry);
+    }
+    g_free(bs_queue);
+    return ret;
+}
+
+
+/* Reopen a single BlockDriverState with the specified flags. */
+int bdrv_reopen(BlockDriverState *bs, int bdrv_flags, Error **errp)
+{
+    int ret = -1;
+    Error *local_err = NULL;
+    BlockReopenQueue *queue = bdrv_reopen_queue(NULL, bs, bdrv_flags);
+
+    ret = bdrv_reopen_multiple(queue, &local_err);
+    if (local_err != NULL) {
+        error_propagate(errp, local_err);
+    }
+    return ret;
+}
+
+
+/*
+ * Prepares a BlockDriverState for reopen. All changes are staged in the
+ * 'opaque' field of the BDRVReopenState, which is used and allocated by
+ * the block driver layer .bdrv_reopen_prepare()
+ *
+ * bs is the BlockDriverState to reopen
+ * flags are the new open flags
+ * queue is the reopen queue
+ *
+ * Returns 0 on success, non-zero on error.  On error errp will be set
+ * as well.
+ *
+ * On failure, bdrv_reopen_abort() will be called to clean up any data.
+ * It is the responsibility of the caller to then call the abort() or
+ * commit() for any other BDS that have been left in a prepare() state
+ *
+ */
+int bdrv_reopen_prepare(BDRVReopenState *reopen_state, BlockReopenQueue *queue,
+                        Error **errp)
+{
+    int ret = -1;
+    Error *local_err = NULL;
+    BlockDriver *drv;
+
+    assert(reopen_state != NULL);
+    assert(reopen_state->bs->drv != NULL);
+    drv = reopen_state->bs->drv;
+
+    /* if we are to stay read-only, do not allow permission change
+     * to r/w */
+    if (!(reopen_state->bs->open_flags & BDRV_O_ALLOW_RDWR) &&
+        reopen_state->flags & BDRV_O_RDWR) {
+        error_set(errp, QERR_DEVICE_IS_READ_ONLY,
+                  reopen_state->bs->device_name);
+        goto error;
+    }
+
+
+    ret = bdrv_flush(reopen_state->bs);
+    if (ret) {
+        error_set(errp, ERROR_CLASS_GENERIC_ERROR, "Error (%s) flushing drive",
+                  strerror(-ret));
+        goto error;
+    }
+
+    if (drv->bdrv_reopen_prepare) {
+        ret = drv->bdrv_reopen_prepare(reopen_state, queue, &local_err);
+        if (ret) {
+            if (local_err != NULL) {
+                error_propagate(errp, local_err);
+            } else {
+                error_setg(errp, "failed while preparing to reopen image '%s'",
+                           reopen_state->bs->filename);
+            }
+            goto error;
+        }
+    } else {
+        /* It is currently mandatory to have a bdrv_reopen_prepare()
+         * handler for each supported drv. */
+        error_set(errp, QERR_BLOCK_FORMAT_FEATURE_NOT_SUPPORTED,
+                  drv->format_name, reopen_state->bs->device_name,
+                 "reopening of file");
+        ret = -1;
+        goto error;
+    }
+
+    ret = 0;
+
+error:
+    return ret;
+}
+
+/*
+ * Takes the staged changes for the reopen from bdrv_reopen_prepare(), and
+ * makes them final by swapping the staging BlockDriverState contents into
+ * the active BlockDriverState contents.
+ */
+void bdrv_reopen_commit(BDRVReopenState *reopen_state)
+{
+    BlockDriver *drv;
+
+    assert(reopen_state != NULL);
+    drv = reopen_state->bs->drv;
+    assert(drv != NULL);
+
+    /* If there are any driver level actions to take */
+    if (drv->bdrv_reopen_commit) {
+        drv->bdrv_reopen_commit(reopen_state);
+    }
+
+    /* set BDS specific flags now */
+    reopen_state->bs->open_flags         = reopen_state->flags;
+    reopen_state->bs->enable_write_cache = !!(reopen_state->flags &
+                                              BDRV_O_CACHE_WB);
+    reopen_state->bs->read_only = !(reopen_state->flags & BDRV_O_RDWR);
+}
+
+/*
+ * Abort the reopen, and delete and free the staged changes in
+ * reopen_state
+ */
+void bdrv_reopen_abort(BDRVReopenState *reopen_state)
+{
+    BlockDriver *drv;
+
+    assert(reopen_state != NULL);
+    drv = reopen_state->bs->drv;
+    assert(drv != NULL);
+
+    if (drv->bdrv_reopen_abort) {
+        drv->bdrv_reopen_abort(reopen_state);
+    }
+}
+
+
+void bdrv_close(BlockDriverState *bs)
+{
+    if (bs->job) {
+        block_job_cancel_sync(bs->job);
+    }
+    bdrv_drain_all(); /* complete I/O */
+    bdrv_flush(bs);
+    bdrv_drain_all(); /* in case flush left pending I/O */
+    notifier_list_notify(&bs->close_notifiers, bs);
+
+    if (bs->drv) {
+        if (bs->backing_hd) {
+            bdrv_delete(bs->backing_hd);
+            bs->backing_hd = NULL;
+        }
+        bs->drv->bdrv_close(bs);
+        g_free(bs->opaque);
+#ifdef _WIN32
+        if (bs->is_temporary) {
+            unlink(bs->filename);
+        }
+#endif
+        bs->opaque = NULL;
+        bs->drv = NULL;
+        bs->copy_on_read = 0;
+        bs->backing_file[0] = '\0';
+        bs->backing_format[0] = '\0';
+        bs->total_sectors = 0;
+        bs->encrypted = 0;
+        bs->valid_key = 0;
+        bs->sg = 0;
+        bs->growable = 0;
+        QDECREF(bs->options);
+        bs->options = NULL;
+
+        if (bs->file != NULL) {
+            bdrv_delete(bs->file);
+            bs->file = NULL;
+        }
+    }
+
+    bdrv_dev_change_media_cb(bs, false);
+
+    /*throttling disk I/O limits*/
+    if (bs->io_limits_enabled) {
+        bdrv_io_limits_disable(bs);
+    }
+}
+
+void bdrv_close_all(void)
+{
+    BlockDriverState *bs;
+
+    QTAILQ_FOREACH(bs, &bdrv_states, list) {
+        bdrv_close(bs);
+    }
+}
+
+/*
+ * Wait for pending requests to complete across all BlockDriverStates
+ *
+ * This function does not flush data to disk, use bdrv_flush_all() for that
+ * after calling this function.
+ *
+ * Note that completion of an asynchronous I/O operation can trigger any
+ * number of other I/O operations on other devices---for example a coroutine
+ * can be arbitrarily complex and a constant flow of I/O can come until the
+ * coroutine is complete.  Because of this, it is not possible to have a
+ * function to drain a single device's I/O queue.
+ */
+void bdrv_drain_all(void)
+{
+    BlockDriverState *bs;
+    bool busy;
+
+    do {
+        busy = qemu_aio_wait();
+
+        /* FIXME: We do not have timer support here, so this is effectively
+         * a busy wait.
+         */
+        QTAILQ_FOREACH(bs, &bdrv_states, list) {
+            if (!qemu_co_queue_empty(&bs->throttled_reqs)) {
+                qemu_co_queue_restart_all(&bs->throttled_reqs);
+                busy = true;
+            }
+        }
+    } while (busy);
+
+    /* If requests are still pending there is a bug somewhere */
+    QTAILQ_FOREACH(bs, &bdrv_states, list) {
+        assert(QLIST_EMPTY(&bs->tracked_requests));
+        assert(qemu_co_queue_empty(&bs->throttled_reqs));
+    }
+}
+
+/* make a BlockDriverState anonymous by removing from bdrv_state list.
+   Also, NULL terminate the device_name to prevent double remove */
+void bdrv_make_anon(BlockDriverState *bs)
+{
+    if (bs->device_name[0] != '\0') {
+        QTAILQ_REMOVE(&bdrv_states, bs, list);
+    }
+    bs->device_name[0] = '\0';
+}
+
+static void bdrv_rebind(BlockDriverState *bs)
+{
+    if (bs->drv && bs->drv->bdrv_rebind) {
+        bs->drv->bdrv_rebind(bs);
+    }
+}
+
+static void bdrv_move_feature_fields(BlockDriverState *bs_dest,
+                                     BlockDriverState *bs_src)
+{
+    /* move some fields that need to stay attached to the device */
+    bs_dest->open_flags         = bs_src->open_flags;
+
+    /* dev info */
+    bs_dest->dev_ops            = bs_src->dev_ops;
+    bs_dest->dev_opaque         = bs_src->dev_opaque;
+    bs_dest->dev                = bs_src->dev;
+    bs_dest->buffer_alignment   = bs_src->buffer_alignment;
+    bs_dest->copy_on_read       = bs_src->copy_on_read;
+
+    bs_dest->enable_write_cache = bs_src->enable_write_cache;
+
+    /* i/o timing parameters */
+    bs_dest->slice_start        = bs_src->slice_start;
+    bs_dest->slice_end          = bs_src->slice_end;
+    bs_dest->slice_submitted    = bs_src->slice_submitted;
+    bs_dest->io_limits          = bs_src->io_limits;
+    bs_dest->throttled_reqs     = bs_src->throttled_reqs;
+    bs_dest->block_timer        = bs_src->block_timer;
+    bs_dest->io_limits_enabled  = bs_src->io_limits_enabled;
+
+    /* r/w error */
+    bs_dest->on_read_error      = bs_src->on_read_error;
+    bs_dest->on_write_error     = bs_src->on_write_error;
+
+    /* i/o status */
+    bs_dest->iostatus_enabled   = bs_src->iostatus_enabled;
+    bs_dest->iostatus           = bs_src->iostatus;
+
+    /* dirty bitmap */
+    bs_dest->dirty_bitmap       = bs_src->dirty_bitmap;
+
+    /* job */
+    bs_dest->in_use             = bs_src->in_use;
+    bs_dest->job                = bs_src->job;
+
+    /* keep the same entry in bdrv_states */
+    pstrcpy(bs_dest->device_name, sizeof(bs_dest->device_name),
+            bs_src->device_name);
+    bs_dest->list = bs_src->list;
+}
+
+/*
+ * Swap bs contents for two image chains while they are live,
+ * while keeping required fields on the BlockDriverState that is
+ * actually attached to a device.
+ *
+ * This will modify the BlockDriverState fields, and swap contents
+ * between bs_new and bs_old. Both bs_new and bs_old are modified.
+ *
+ * bs_new is required to be anonymous.
+ *
+ * This function does not create any image files.
+ */
+void bdrv_swap(BlockDriverState *bs_new, BlockDriverState *bs_old)
+{
+    BlockDriverState tmp;
+
+    /* bs_new must be anonymous and shouldn't have anything fancy enabled */
+    assert(bs_new->device_name[0] == '\0');
+    assert(bs_new->dirty_bitmap == NULL);
+    assert(bs_new->job == NULL);
+    assert(bs_new->dev == NULL);
+    assert(bs_new->in_use == 0);
+    assert(bs_new->io_limits_enabled == false);
+    assert(bs_new->block_timer == NULL);
+
+    tmp = *bs_new;
+    *bs_new = *bs_old;
+    *bs_old = tmp;
+
+    /* there are some fields that should not be swapped, move them back */
+    bdrv_move_feature_fields(&tmp, bs_old);
+    bdrv_move_feature_fields(bs_old, bs_new);
+    bdrv_move_feature_fields(bs_new, &tmp);
+
+    /* bs_new shouldn't be in bdrv_states even after the swap!  */
+    assert(bs_new->device_name[0] == '\0');
+
+    /* Check a few fields that should remain attached to the device */
+    assert(bs_new->dev == NULL);
+    assert(bs_new->job == NULL);
+    assert(bs_new->in_use == 0);
+    assert(bs_new->io_limits_enabled == false);
+    assert(bs_new->block_timer == NULL);
+
+    bdrv_rebind(bs_new);
+    bdrv_rebind(bs_old);
+}
+
+/*
+ * Add new bs contents at the top of an image chain while the chain is
+ * live, while keeping required fields on the top layer.
+ *
+ * This will modify the BlockDriverState fields, and swap contents
+ * between bs_new and bs_top. Both bs_new and bs_top are modified.
+ *
+ * bs_new is required to be anonymous.
+ *
+ * This function does not create any image files.
+ */
+void bdrv_append(BlockDriverState *bs_new, BlockDriverState *bs_top)
+{
+    bdrv_swap(bs_new, bs_top);
+
+    /* The contents of 'tmp' will become bs_top, as we are
+     * swapping bs_new and bs_top contents. */
+    bs_top->backing_hd = bs_new;
+    bs_top->open_flags &= ~BDRV_O_NO_BACKING;
+    pstrcpy(bs_top->backing_file, sizeof(bs_top->backing_file),
+            bs_new->filename);
+    pstrcpy(bs_top->backing_format, sizeof(bs_top->backing_format),
+            bs_new->drv ? bs_new->drv->format_name : "");
+}
+
+void bdrv_delete(BlockDriverState *bs)
+{
+    assert(!bs->dev);
+    assert(!bs->job);
+    assert(!bs->in_use);
+
+    /* remove from list, if necessary */
+    bdrv_make_anon(bs);
+
+    bdrv_close(bs);
+
+    g_free(bs);
+}
+
+int bdrv_attach_dev(BlockDriverState *bs, void *dev)
+/* TODO change to DeviceState *dev when all users are qdevified */
+{
+    if (bs->dev) {
+        return -EBUSY;
+    }
+    bs->dev = dev;
+    bdrv_iostatus_reset(bs);
+    return 0;
+}
+
+/* TODO qdevified devices don't use this, remove when devices are qdevified */
+void bdrv_attach_dev_nofail(BlockDriverState *bs, void *dev)
+{
+    if (bdrv_attach_dev(bs, dev) < 0) {
+        abort();
+    }
+}
+
+void bdrv_detach_dev(BlockDriverState *bs, void *dev)
+/* TODO change to DeviceState *dev when all users are qdevified */
+{
+    assert(bs->dev == dev);
+    bs->dev = NULL;
+    bs->dev_ops = NULL;
+    bs->dev_opaque = NULL;
+    bs->buffer_alignment = 512;
+}
+
+/* TODO change to return DeviceState * when all users are qdevified */
+void *bdrv_get_attached_dev(BlockDriverState *bs)
+{
+    return bs->dev;
+}
+
+void bdrv_set_dev_ops(BlockDriverState *bs, const BlockDevOps *ops,
+                      void *opaque)
+{
+    bs->dev_ops = ops;
+    bs->dev_opaque = opaque;
+}
+
+void bdrv_emit_qmp_error_event(const BlockDriverState *bdrv,
+                               enum MonitorEvent ev,
+                               BlockErrorAction action, bool is_read)
+{
+    QObject *data;
+    const char *action_str;
+
+    switch (action) {
+    case BDRV_ACTION_REPORT:
+        action_str = "report";
+        break;
+    case BDRV_ACTION_IGNORE:
+        action_str = "ignore";
+        break;
+    case BDRV_ACTION_STOP:
+        action_str = "stop";
+        break;
+    default:
+        abort();
+    }
+
+    data = qobject_from_jsonf("{ 'device': %s, 'action': %s, 'operation': %s }",
+                              bdrv->device_name,
+                              action_str,
+                              is_read ? "read" : "write");
+    monitor_protocol_event(ev, data);
+
+    qobject_decref(data);
+}
+
+static void bdrv_emit_qmp_eject_event(BlockDriverState *bs, bool ejected)
+{
+    QObject *data;
+
+    data = qobject_from_jsonf("{ 'device': %s, 'tray-open': %i }",
+                              bdrv_get_device_name(bs), ejected);
+    monitor_protocol_event(QEVENT_DEVICE_TRAY_MOVED, data);
+
+    qobject_decref(data);
+}
+
+static void bdrv_dev_change_media_cb(BlockDriverState *bs, bool load)
+{
+    if (bs->dev_ops && bs->dev_ops->change_media_cb) {
+        bool tray_was_closed = !bdrv_dev_is_tray_open(bs);
+        bs->dev_ops->change_media_cb(bs->dev_opaque, load);
+        if (tray_was_closed) {
+            /* tray open */
+            bdrv_emit_qmp_eject_event(bs, true);
+        }
+        if (load) {
+            /* tray close */
+            bdrv_emit_qmp_eject_event(bs, false);
+        }
+    }
+}
+
+bool bdrv_dev_has_removable_media(BlockDriverState *bs)
+{
+    return !bs->dev || (bs->dev_ops && bs->dev_ops->change_media_cb);
+}
+
+void bdrv_dev_eject_request(BlockDriverState *bs, bool force)
+{
+    if (bs->dev_ops && bs->dev_ops->eject_request_cb) {
+        bs->dev_ops->eject_request_cb(bs->dev_opaque, force);
+    }
+}
+
+bool bdrv_dev_is_tray_open(BlockDriverState *bs)
+{
+    if (bs->dev_ops && bs->dev_ops->is_tray_open) {
+        return bs->dev_ops->is_tray_open(bs->dev_opaque);
+    }
+    return false;
+}
+
+static void bdrv_dev_resize_cb(BlockDriverState *bs)
+{
+    if (bs->dev_ops && bs->dev_ops->resize_cb) {
+        bs->dev_ops->resize_cb(bs->dev_opaque);
+    }
+}
+
+bool bdrv_dev_is_medium_locked(BlockDriverState *bs)
+{
+    if (bs->dev_ops && bs->dev_ops->is_medium_locked) {
+        return bs->dev_ops->is_medium_locked(bs->dev_opaque);
+    }
+    return false;
+}
+
+/*
+ * Run consistency checks on an image
+ *
+ * Returns 0 if the check could be completed (it doesn't mean that the image is
+ * free of errors) or -errno when an internal error occurred. The results of the
+ * check are stored in res.
+ */
+int bdrv_check(BlockDriverState *bs, BdrvCheckResult *res, BdrvCheckMode fix)
+{
+    if (bs->drv->bdrv_check == NULL) {
+        return -ENOTSUP;
+    }
+
+    memset(res, 0, sizeof(*res));
+    return bs->drv->bdrv_check(bs, res, fix);
+}
+
+#define COMMIT_BUF_SECTORS 2048
+
+/* commit COW file into the raw image */
+int bdrv_commit(BlockDriverState *bs)
+{
+    BlockDriver *drv = bs->drv;
+    int64_t sector, total_sectors;
+    int n, ro, open_flags;
+    int ret = 0;
+    uint8_t *buf;
+    char filename[PATH_MAX];
+
+    if (!drv)
+        return -ENOMEDIUM;
+    
+    if (!bs->backing_hd) {
+        return -ENOTSUP;
+    }
+
+    if (bdrv_in_use(bs) || bdrv_in_use(bs->backing_hd)) {
+        return -EBUSY;
+    }
+
+    ro = bs->backing_hd->read_only;
+    /* Use pstrcpy (not strncpy): filename must be NUL-terminated. */
+    pstrcpy(filename, sizeof(filename), bs->backing_hd->filename);
+    open_flags =  bs->backing_hd->open_flags;
+
+    if (ro) {
+        if (bdrv_reopen(bs->backing_hd, open_flags | BDRV_O_RDWR, NULL)) {
+            return -EACCES;
+        }
+    }
+
+    total_sectors = bdrv_getlength(bs) >> BDRV_SECTOR_BITS;
+    buf = g_malloc(COMMIT_BUF_SECTORS * BDRV_SECTOR_SIZE);
+
+    for (sector = 0; sector < total_sectors; sector += n) {
+        if (bdrv_is_allocated(bs, sector, COMMIT_BUF_SECTORS, &n)) {
+
+            if (bdrv_read(bs, sector, buf, n) != 0) {
+                ret = -EIO;
+                goto ro_cleanup;
+            }
+
+            if (bdrv_write(bs->backing_hd, sector, buf, n) != 0) {
+                ret = -EIO;
+                goto ro_cleanup;
+            }
+        }
+    }
+
+    if (drv->bdrv_make_empty) {
+        ret = drv->bdrv_make_empty(bs);
+        bdrv_flush(bs);
+    }
+
+    /*
+     * Make sure all data we wrote to the backing device is actually
+     * stable on disk.
+     */
+    if (bs->backing_hd)
+        bdrv_flush(bs->backing_hd);
+
+ro_cleanup:
+    g_free(buf);
+
+    if (ro) {
+        /* ignoring error return here */
+        bdrv_reopen(bs->backing_hd, open_flags & ~BDRV_O_RDWR, NULL);
+    }
+
+    return ret;
+}
+
+int bdrv_commit_all(void)
+{
+    BlockDriverState *bs;
+
+    QTAILQ_FOREACH(bs, &bdrv_states, list) {
+        if (bs->drv && bs->backing_hd) {
+            int ret = bdrv_commit(bs);
+            if (ret < 0) {
+                return ret;
+            }
+        }
+    }
+    return 0;
+}
+
+/**
+ * Remove an active request from the tracked requests list
+ *
+ * This function should be called when a tracked request is completing.
+ */
+static void tracked_request_end(BdrvTrackedRequest *req)
+{
+    QLIST_REMOVE(req, list);
+    qemu_co_queue_restart_all(&req->wait_queue);
+}
+
+/**
+ * Add an active request to the tracked requests list
+ */
+static void tracked_request_begin(BdrvTrackedRequest *req,
+                                  BlockDriverState *bs,
+                                  int64_t sector_num,
+                                  int nb_sectors, bool is_write)
+{
+    *req = (BdrvTrackedRequest){
+        .bs = bs,
+        .sector_num = sector_num,
+        .nb_sectors = nb_sectors,
+        .is_write = is_write,
+        .co = qemu_coroutine_self(),
+    };
+
+    qemu_co_queue_init(&req->wait_queue);
+
+    QLIST_INSERT_HEAD(&bs->tracked_requests, req, list);
+}
+
+/**
+ * Round a region to cluster boundaries
+ */
+void bdrv_round_to_clusters(BlockDriverState *bs,
+                            int64_t sector_num, int nb_sectors,
+                            int64_t *cluster_sector_num,
+                            int *cluster_nb_sectors)
+{
+    BlockDriverInfo bdi;
+
+    if (bdrv_get_info(bs, &bdi) < 0 || bdi.cluster_size == 0) {
+        *cluster_sector_num = sector_num;
+        *cluster_nb_sectors = nb_sectors;
+    } else {
+        int64_t c = bdi.cluster_size / BDRV_SECTOR_SIZE;
+        *cluster_sector_num = QEMU_ALIGN_DOWN(sector_num, c);
+        *cluster_nb_sectors = QEMU_ALIGN_UP(sector_num - *cluster_sector_num +
+                                            nb_sectors, c);
+    }
+}
+
+static bool tracked_request_overlaps(BdrvTrackedRequest *req,
+                                     int64_t sector_num, int nb_sectors) {
+    /*        aaaa   bbbb */
+    if (sector_num >= req->sector_num + req->nb_sectors) {
+        return false;
+    }
+    /* bbbb   aaaa        */
+    if (req->sector_num >= sector_num + nb_sectors) {
+        return false;
+    }
+    return true;
+}
+
+static void coroutine_fn wait_for_overlapping_requests(BlockDriverState *bs,
+        int64_t sector_num, int nb_sectors)
+{
+    BdrvTrackedRequest *req;
+    int64_t cluster_sector_num;
+    int cluster_nb_sectors;
+    bool retry;
+
+    /* If we touch the same cluster it counts as an overlap.  This guarantees
+     * that allocating writes will be serialized and not race with each other
+     * for the same cluster.  For example, in copy-on-read it ensures that the
+     * CoR read and write operations are atomic and guest writes cannot
+     * interleave between them.
+     */
+    bdrv_round_to_clusters(bs, sector_num, nb_sectors,
+                           &cluster_sector_num, &cluster_nb_sectors);
+
+    do {
+        retry = false;
+        QLIST_FOREACH(req, &bs->tracked_requests, list) {
+            if (tracked_request_overlaps(req, cluster_sector_num,
+                                         cluster_nb_sectors)) {
+                /* Hitting this means there was a reentrant request, for
+                 * example, a block driver issuing nested requests.  This must
+                 * never happen since it means deadlock.
+                 */
+                assert(qemu_coroutine_self() != req->co);
+
+                qemu_co_queue_wait(&req->wait_queue);
+                retry = true;
+                break;
+            }
+        }
+    } while (retry);
+}
+
+/*
+ * Return values:
+ * 0        - success
+ * -EINVAL  - backing format specified, but no file
+ * -ENOSPC  - can't update the backing file because no space is left in the
+ *            image file header
+ * -ENOTSUP - format driver doesn't support changing the backing file
+ */
+int bdrv_change_backing_file(BlockDriverState *bs,
+    const char *backing_file, const char *backing_fmt)
+{
+    BlockDriver *drv = bs->drv;
+    int ret;
+
+    /* Backing file format doesn't make sense without a backing file */
+    if (backing_fmt && !backing_file) {
+        return -EINVAL;
+    }
+
+    if (drv->bdrv_change_backing_file != NULL) {
+        ret = drv->bdrv_change_backing_file(bs, backing_file, backing_fmt);
+    } else {
+        ret = -ENOTSUP;
+    }
+
+    if (ret == 0) {
+        pstrcpy(bs->backing_file, sizeof(bs->backing_file), backing_file ?: "");
+        pstrcpy(bs->backing_format, sizeof(bs->backing_format), backing_fmt ?: "");
+    }
+    return ret;
+}
+
+/*
+ * Finds the image layer in the chain that has 'bs' as its backing file.
+ *
+ * active is the current topmost image.
+ *
+ * Returns NULL if bs is not found in active's image chain,
+ * or if active == bs.
+ */
+BlockDriverState *bdrv_find_overlay(BlockDriverState *active,
+                                    BlockDriverState *bs)
+{
+    BlockDriverState *overlay = NULL;
+    BlockDriverState *intermediate;
+
+    assert(active != NULL);
+    assert(bs != NULL);
+
+    /* if bs is the same as active, then by definition it has no overlay
+     */
+    if (active == bs) {
+        return NULL;
+    }
+
+    intermediate = active;
+    while (intermediate->backing_hd) {
+        if (intermediate->backing_hd == bs) {
+            overlay = intermediate;
+            break;
+        }
+        intermediate = intermediate->backing_hd;
+    }
+
+    return overlay;
+}
+
+typedef struct BlkIntermediateStates {
+    BlockDriverState *bs;
+    QSIMPLEQ_ENTRY(BlkIntermediateStates) entry;
+} BlkIntermediateStates;
+
+
+/*
+ * Drops images above 'base' up to and including 'top', and sets the image
+ * above 'top' to have base as its backing file.
+ *
+ * Requires that the overlay to 'top' is opened r/w, so that the backing file
+ * information in 'bs' can be properly updated.
+ *
+ * E.g., this will convert the following chain:
+ * bottom <- base <- intermediate <- top <- active
+ *
+ * to
+ *
+ * bottom <- base <- active
+ *
+ * It is allowed for bottom==base, in which case it converts:
+ *
+ * base <- intermediate <- top <- active
+ *
+ * to
+ *
+ * base <- active
+ *
+ * Error conditions:
+ *  if active == top, that is considered an error
+ *
+ */
+int bdrv_drop_intermediate(BlockDriverState *active, BlockDriverState *top,
+                           BlockDriverState *base)
+{
+    BlockDriverState *intermediate;
+    BlockDriverState *base_bs = NULL;
+    BlockDriverState *new_top_bs = NULL;
+    BlkIntermediateStates *intermediate_state, *next;
+    int ret = -EIO;
+
+    QSIMPLEQ_HEAD(states_to_delete, BlkIntermediateStates) states_to_delete;
+    QSIMPLEQ_INIT(&states_to_delete);
+
+    if (!top->drv || !base->drv) {
+        goto exit;
+    }
+
+    new_top_bs = bdrv_find_overlay(active, top);
+
+    if (new_top_bs == NULL) {
+        /* we could not find the image above 'top', this is an error */
+        goto exit;
+    }
+
+    /* special case of new_top_bs->backing_hd already pointing to base - nothing
+     * to do, no intermediate images */
+    if (new_top_bs->backing_hd == base) {
+        ret = 0;
+        goto exit;
+    }
+
+    intermediate = top;
+
+    /* now we will go down through the list, and add each BDS we find
+     * into our deletion queue, until we hit the 'base'
+     */
+    while (intermediate) {
+        intermediate_state = g_malloc0(sizeof(BlkIntermediateStates));
+        intermediate_state->bs = intermediate;
+        QSIMPLEQ_INSERT_TAIL(&states_to_delete, intermediate_state, entry);
+
+        if (intermediate->backing_hd == base) {
+            base_bs = intermediate->backing_hd;
+            break;
+        }
+        intermediate = intermediate->backing_hd;
+    }
+    if (base_bs == NULL) {
+        /* something went wrong, we did not end at the base. safely
+         * unravel everything, and exit with error */
+        goto exit;
+    }
+
+    /* success - we can delete the intermediate states, and link top->base */
+    ret = bdrv_change_backing_file(new_top_bs, base_bs->filename,
+                                   base_bs->drv ? base_bs->drv->format_name : "");
+    if (ret) {
+        goto exit;
+    }
+    new_top_bs->backing_hd = base_bs;
+
+
+    QSIMPLEQ_FOREACH_SAFE(intermediate_state, &states_to_delete, entry, next) {
+        /* so that bdrv_close() does not recursively close the chain */
+        intermediate_state->bs->backing_hd = NULL;
+        bdrv_delete(intermediate_state->bs);
+    }
+    ret = 0;
+
+exit:
+    QSIMPLEQ_FOREACH_SAFE(intermediate_state, &states_to_delete, entry, next) {
+        g_free(intermediate_state);
+    }
+    return ret;
+}
+
+
+static int bdrv_check_byte_request(BlockDriverState *bs, int64_t offset,
+                                   size_t size)
+{
+    int64_t len;
+
+    if (!bdrv_is_inserted(bs))
+        return -ENOMEDIUM;
+
+    if (bs->growable)
+        return 0;
+
+    len = bdrv_getlength(bs);
+
+    if (offset < 0)
+        return -EIO;
+
+    if ((offset > len) || (len - offset < size))
+        return -EIO;
+
+    return 0;
+}
+
+static int bdrv_check_request(BlockDriverState *bs, int64_t sector_num,
+                              int nb_sectors)
+{
+    return bdrv_check_byte_request(bs, sector_num * BDRV_SECTOR_SIZE,
+                                   nb_sectors * BDRV_SECTOR_SIZE);
+}
+
+typedef struct RwCo {
+    BlockDriverState *bs;
+    int64_t sector_num;
+    int nb_sectors;
+    QEMUIOVector *qiov;
+    bool is_write;
+    int ret;
+} RwCo;
+
+static void coroutine_fn bdrv_rw_co_entry(void *opaque)
+{
+    RwCo *rwco = opaque;
+
+    if (!rwco->is_write) {
+        rwco->ret = bdrv_co_do_readv(rwco->bs, rwco->sector_num,
+                                     rwco->nb_sectors, rwco->qiov, 0);
+    } else {
+        rwco->ret = bdrv_co_do_writev(rwco->bs, rwco->sector_num,
+                                      rwco->nb_sectors, rwco->qiov, 0);
+    }
+}
+
+/*
+ * Process a vectored synchronous request using coroutines
+ */
+static int bdrv_rwv_co(BlockDriverState *bs, int64_t sector_num,
+                       QEMUIOVector *qiov, bool is_write)
+{
+    Coroutine *co;
+    RwCo rwco = {
+        .bs = bs,
+        .sector_num = sector_num,
+        .nb_sectors = qiov->size >> BDRV_SECTOR_BITS,
+        .qiov = qiov,
+        .is_write = is_write,
+        .ret = NOT_DONE,
+    };
+    assert((qiov->size & (BDRV_SECTOR_SIZE - 1)) == 0);
+
+    /**
+     * In sync call context, when the vcpu is blocked, this throttling timer
+     * will not fire; so the I/O throttling function has to be disabled here
+     * if it has been enabled.
+     */
+    if (bs->io_limits_enabled) {
+        fprintf(stderr, "Disabling I/O throttling on '%s' due "
+                        "to synchronous I/O.\n", bdrv_get_device_name(bs));
+        bdrv_io_limits_disable(bs);
+    }
+
+    if (qemu_in_coroutine()) {
+        /* Fast-path if already in coroutine context */
+        bdrv_rw_co_entry(&rwco);
+    } else {
+        co = qemu_coroutine_create(bdrv_rw_co_entry);
+        qemu_coroutine_enter(co, &rwco);
+        while (rwco.ret == NOT_DONE) {
+            qemu_aio_wait();
+        }
+    }
+    return rwco.ret;
+}
+
+/*
+ * Process a synchronous request using coroutines
+ */
+static int bdrv_rw_co(BlockDriverState *bs, int64_t sector_num, uint8_t *buf,
+                      int nb_sectors, bool is_write)
+{
+    QEMUIOVector qiov;
+    struct iovec iov = {
+        .iov_base = (void *)buf,
+        .iov_len = nb_sectors * BDRV_SECTOR_SIZE,
+    };
+
+    qemu_iovec_init_external(&qiov, &iov, 1);
+    return bdrv_rwv_co(bs, sector_num, &qiov, is_write);
+}
+
+/* return < 0 if error. See bdrv_write() for the return codes */
+int bdrv_read(BlockDriverState *bs, int64_t sector_num,
+              uint8_t *buf, int nb_sectors)
+{
+    return bdrv_rw_co(bs, sector_num, buf, nb_sectors, false);
+}
+
+/* Just like bdrv_read(), but with I/O throttling temporarily disabled */
+int bdrv_read_unthrottled(BlockDriverState *bs, int64_t sector_num,
+                          uint8_t *buf, int nb_sectors)
+{
+    bool enabled;
+    int ret;
+
+    enabled = bs->io_limits_enabled;
+    bs->io_limits_enabled = false;
+    ret = bdrv_read(bs, 0, buf, 1);
+    bs->io_limits_enabled = enabled;
+    return ret;
+}
+
+/* Return < 0 if error. Important errors are:
+  -EIO         generic I/O error (may happen for all errors)
+  -ENOMEDIUM   No media inserted.
+  -EINVAL      Invalid sector number or nb_sectors
+  -EACCES      Trying to write a read-only device
+*/
+int bdrv_write(BlockDriverState *bs, int64_t sector_num,
+               const uint8_t *buf, int nb_sectors)
+{
+    return bdrv_rw_co(bs, sector_num, (uint8_t *)buf, nb_sectors, true);
+}
+
+int bdrv_writev(BlockDriverState *bs, int64_t sector_num, QEMUIOVector *qiov)
+{
+    return bdrv_rwv_co(bs, sector_num, qiov, true);
+}
+
+int bdrv_pread(BlockDriverState *bs, int64_t offset,
+               void *buf, int count1)
+{
+    uint8_t tmp_buf[BDRV_SECTOR_SIZE];
+    int len, nb_sectors, count;
+    int64_t sector_num;
+    int ret;
+
+    count = count1;
+    /* first read to align to sector start */
+    len = (BDRV_SECTOR_SIZE - offset) & (BDRV_SECTOR_SIZE - 1);
+    if (len > count)
+        len = count;
+    sector_num = offset >> BDRV_SECTOR_BITS;
+    if (len > 0) {
+        if ((ret = bdrv_read(bs, sector_num, tmp_buf, 1)) < 0)
+            return ret;
+        memcpy(buf, tmp_buf + (offset & (BDRV_SECTOR_SIZE - 1)), len);
+        count -= len;
+        if (count == 0)
+            return count1;
+        sector_num++;
+        buf += len;
+    }
+
+    /* read the sectors "in place" */
+    nb_sectors = count >> BDRV_SECTOR_BITS;
+    if (nb_sectors > 0) {
+        if ((ret = bdrv_read(bs, sector_num, buf, nb_sectors)) < 0)
+            return ret;
+        sector_num += nb_sectors;
+        len = nb_sectors << BDRV_SECTOR_BITS;
+        buf += len;
+        count -= len;
+    }
+
+    /* add data from the last sector */
+    if (count > 0) {
+        if ((ret = bdrv_read(bs, sector_num, tmp_buf, 1)) < 0)
+            return ret;
+        memcpy(buf, tmp_buf, count);
+    }
+    return count1;
+}
+
+int bdrv_pwritev(BlockDriverState *bs, int64_t offset, QEMUIOVector *qiov)
+{
+    uint8_t tmp_buf[BDRV_SECTOR_SIZE];
+    int len, nb_sectors, count;
+    int64_t sector_num;
+    int ret;
+
+    count = qiov->size;
+
+    /* first write to align to sector start */
+    len = (BDRV_SECTOR_SIZE - offset) & (BDRV_SECTOR_SIZE - 1);
+    if (len > count)
+        len = count;
+    sector_num = offset >> BDRV_SECTOR_BITS;
+    if (len > 0) {
+        if ((ret = bdrv_read(bs, sector_num, tmp_buf, 1)) < 0)
+            return ret;
+        qemu_iovec_to_buf(qiov, 0, tmp_buf + (offset & (BDRV_SECTOR_SIZE - 1)),
+                          len);
+        if ((ret = bdrv_write(bs, sector_num, tmp_buf, 1)) < 0)
+            return ret;
+        count -= len;
+        if (count == 0)
+            return qiov->size;
+        sector_num++;
+    }
+
+    /* write the sectors "in place" */
+    nb_sectors = count >> BDRV_SECTOR_BITS;
+    if (nb_sectors > 0) {
+        QEMUIOVector qiov_inplace;
+
+        qemu_iovec_init(&qiov_inplace, qiov->niov);
+        qemu_iovec_concat(&qiov_inplace, qiov, len,
+                          nb_sectors << BDRV_SECTOR_BITS);
+        ret = bdrv_writev(bs, sector_num, &qiov_inplace);
+        qemu_iovec_destroy(&qiov_inplace);
+        if (ret < 0) {
+            return ret;
+        }
+
+        sector_num += nb_sectors;
+        len = nb_sectors << BDRV_SECTOR_BITS;
+        count -= len;
+    }
+
+    /* add data from the last sector */
+    if (count > 0) {
+        if ((ret = bdrv_read(bs, sector_num, tmp_buf, 1)) < 0)
+            return ret;
+        qemu_iovec_to_buf(qiov, qiov->size - count, tmp_buf, count);
+        if ((ret = bdrv_write(bs, sector_num, tmp_buf, 1)) < 0)
+            return ret;
+    }
+    return qiov->size;
+}
+
+int bdrv_pwrite(BlockDriverState *bs, int64_t offset,
+                const void *buf, int count1)
+{
+    QEMUIOVector qiov;
+    struct iovec iov = {
+        .iov_base   = (void *) buf,
+        .iov_len    = count1,
+    };
+
+    qemu_iovec_init_external(&qiov, &iov, 1);
+    return bdrv_pwritev(bs, offset, &qiov);
+}
+
+/*
+ * Writes to the file and ensures that no writes are reordered across this
+ * request (acts as a barrier)
+ *
+ * Returns 0 on success, -errno in error cases.
+ */
+int bdrv_pwrite_sync(BlockDriverState *bs, int64_t offset,
+    const void *buf, int count)
+{
+    int ret;
+
+    ret = bdrv_pwrite(bs, offset, buf, count);
+    if (ret < 0) {
+        return ret;
+    }
+
+    /* No flush needed for cache modes that already do it */
+    if (bs->enable_write_cache) {
+        bdrv_flush(bs);
+    }
+
+    return 0;
+}
+
+static int coroutine_fn bdrv_co_do_copy_on_readv(BlockDriverState *bs,
+        int64_t sector_num, int nb_sectors, QEMUIOVector *qiov)
+{
+    /* Perform I/O through a temporary buffer so that users who scribble over
+     * their read buffer while the operation is in progress do not end up
+     * modifying the image file.  This is critical for zero-copy guest I/O
+     * where anything might happen inside guest memory.
+     */
+    void *bounce_buffer;
+
+    BlockDriver *drv = bs->drv;
+    struct iovec iov;
+    QEMUIOVector bounce_qiov;
+    int64_t cluster_sector_num;
+    int cluster_nb_sectors;
+    size_t skip_bytes;
+    int ret;
+
+    /* Cover entire cluster so no additional backing file I/O is required when
+     * allocating cluster in the image file.
+     */
+    bdrv_round_to_clusters(bs, sector_num, nb_sectors,
+                           &cluster_sector_num, &cluster_nb_sectors);
+
+    trace_bdrv_co_do_copy_on_readv(bs, sector_num, nb_sectors,
+                                   cluster_sector_num, cluster_nb_sectors);
+
+    iov.iov_len = cluster_nb_sectors * BDRV_SECTOR_SIZE;
+    iov.iov_base = bounce_buffer = qemu_blockalign(bs, iov.iov_len);
+    qemu_iovec_init_external(&bounce_qiov, &iov, 1);
+
+    ret = drv->bdrv_co_readv(bs, cluster_sector_num, cluster_nb_sectors,
+                             &bounce_qiov);
+    if (ret < 0) {
+        goto err;
+    }
+
+    if (drv->bdrv_co_write_zeroes &&
+        buffer_is_zero(bounce_buffer, iov.iov_len)) {
+        ret = bdrv_co_do_write_zeroes(bs, cluster_sector_num,
+                                      cluster_nb_sectors);
+    } else {
+        /* This does not change the data on the disk, it is not necessary
+         * to flush even in cache=writethrough mode.
+         */
+        ret = drv->bdrv_co_writev(bs, cluster_sector_num, cluster_nb_sectors,
+                                  &bounce_qiov);
+    }
+
+    if (ret < 0) {
+        /* It might be okay to ignore write errors for guest requests.  If this
+         * is a deliberate copy-on-read then we don't want to ignore the error.
+         * Simply report it in all cases.
+         */
+        goto err;
+    }
+
+    skip_bytes = (sector_num - cluster_sector_num) * BDRV_SECTOR_SIZE;
+    qemu_iovec_from_buf(qiov, 0, bounce_buffer + skip_bytes,
+                        nb_sectors * BDRV_SECTOR_SIZE);
+
+err:
+    qemu_vfree(bounce_buffer);
+    return ret;
+}
+
+/*
+ * Handle a read request in coroutine context
+ */
+static int coroutine_fn bdrv_co_do_readv(BlockDriverState *bs,
+    int64_t sector_num, int nb_sectors, QEMUIOVector *qiov,
+    BdrvRequestFlags flags)
+{
+    BlockDriver *drv = bs->drv;
+    BdrvTrackedRequest req;
+    int ret;
+
+    if (!drv) {
+        return -ENOMEDIUM;
+    }
+    if (bdrv_check_request(bs, sector_num, nb_sectors)) {
+        return -EIO;
+    }
+
+    /* throttling disk read I/O */
+    if (bs->io_limits_enabled) {
+        bdrv_io_limits_intercept(bs, false, nb_sectors);
+    }
+
+    if (bs->copy_on_read) {
+        flags |= BDRV_REQ_COPY_ON_READ;
+    }
+    if (flags & BDRV_REQ_COPY_ON_READ) {
+        bs->copy_on_read_in_flight++;
+    }
+
+    if (bs->copy_on_read_in_flight) {
+        wait_for_overlapping_requests(bs, sector_num, nb_sectors);
+    }
+
+    tracked_request_begin(&req, bs, sector_num, nb_sectors, false);
+
+    if (flags & BDRV_REQ_COPY_ON_READ) {
+        int pnum;
+
+        ret = bdrv_co_is_allocated(bs, sector_num, nb_sectors, &pnum);
+        if (ret < 0) {
+            goto out;
+        }
+
+        if (!ret || pnum != nb_sectors) {
+            ret = bdrv_co_do_copy_on_readv(bs, sector_num, nb_sectors, qiov);
+            goto out;
+        }
+    }
+
+    ret = drv->bdrv_co_readv(bs, sector_num, nb_sectors, qiov);
+
+out:
+    tracked_request_end(&req);
+
+    if (flags & BDRV_REQ_COPY_ON_READ) {
+        bs->copy_on_read_in_flight--;
+    }
+
+    return ret;
+}
+
+int coroutine_fn bdrv_co_readv(BlockDriverState *bs, int64_t sector_num,
+    int nb_sectors, QEMUIOVector *qiov)
+{
+    trace_bdrv_co_readv(bs, sector_num, nb_sectors);
+
+    return bdrv_co_do_readv(bs, sector_num, nb_sectors, qiov, 0);
+}
+
+int coroutine_fn bdrv_co_copy_on_readv(BlockDriverState *bs,
+    int64_t sector_num, int nb_sectors, QEMUIOVector *qiov)
+{
+    trace_bdrv_co_copy_on_readv(bs, sector_num, nb_sectors);
+
+    return bdrv_co_do_readv(bs, sector_num, nb_sectors, qiov,
+                            BDRV_REQ_COPY_ON_READ);
+}
+
+static int coroutine_fn bdrv_co_do_write_zeroes(BlockDriverState *bs,
+    int64_t sector_num, int nb_sectors)
+{
+    BlockDriver *drv = bs->drv;
+    QEMUIOVector qiov;
+    struct iovec iov;
+    int ret;
+
+    /* TODO Emulate only part of misaligned requests instead of letting block
+     * drivers return -ENOTSUP and emulate everything */
+
+    /* First try the efficient write zeroes operation */
+    if (drv->bdrv_co_write_zeroes) {
+        ret = drv->bdrv_co_write_zeroes(bs, sector_num, nb_sectors);
+        if (ret != -ENOTSUP) {
+            return ret;
+        }
+    }
+
+    /* Fall back to bounce buffer if write zeroes is unsupported */
+    iov.iov_len  = nb_sectors * BDRV_SECTOR_SIZE;
+    iov.iov_base = qemu_blockalign(bs, iov.iov_len);
+    memset(iov.iov_base, 0, iov.iov_len);
+    qemu_iovec_init_external(&qiov, &iov, 1);
+
+    ret = drv->bdrv_co_writev(bs, sector_num, nb_sectors, &qiov);
+
+    qemu_vfree(iov.iov_base);
+    return ret;
+}
+
+/*
+ * Handle a write request in coroutine context
+ */
+static int coroutine_fn bdrv_co_do_writev(BlockDriverState *bs,
+    int64_t sector_num, int nb_sectors, QEMUIOVector *qiov,
+    BdrvRequestFlags flags)
+{
+    BlockDriver *drv = bs->drv;
+    BdrvTrackedRequest req;
+    int ret;
+
+    if (!bs->drv) {
+        return -ENOMEDIUM;
+    }
+    if (bs->read_only) {
+        return -EACCES;
+    }
+    if (bdrv_check_request(bs, sector_num, nb_sectors)) {
+        return -EIO;
+    }
+
+    /* throttling disk write I/O */
+    if (bs->io_limits_enabled) {
+        bdrv_io_limits_intercept(bs, true, nb_sectors);
+    }
+
+    if (bs->copy_on_read_in_flight) {
+        wait_for_overlapping_requests(bs, sector_num, nb_sectors);
+    }
+
+    tracked_request_begin(&req, bs, sector_num, nb_sectors, true);
+
+    ret = notifier_with_return_list_notify(&bs->before_write_notifiers, &req);
+
+    if (ret < 0) {
+        /* Do nothing, write notifier decided to fail this request */
+    } else if (flags & BDRV_REQ_ZERO_WRITE) {
+        ret = bdrv_co_do_write_zeroes(bs, sector_num, nb_sectors);
+    } else {
+        ret = drv->bdrv_co_writev(bs, sector_num, nb_sectors, qiov);
+    }
+
+    if (ret == 0 && !bs->enable_write_cache) {
+        ret = bdrv_co_flush(bs);
+    }
+
+    if (bs->dirty_bitmap) {
+        bdrv_set_dirty(bs, sector_num, nb_sectors);
+    }
+
+    if (bs->wr_highest_sector < sector_num + nb_sectors - 1) {
+        bs->wr_highest_sector = sector_num + nb_sectors - 1;
+    }
+
+    tracked_request_end(&req);
+
+    return ret;
+}
+
+int coroutine_fn bdrv_co_writev(BlockDriverState *bs, int64_t sector_num,
+    int nb_sectors, QEMUIOVector *qiov)
+{
+    trace_bdrv_co_writev(bs, sector_num, nb_sectors);
+
+    return bdrv_co_do_writev(bs, sector_num, nb_sectors, qiov, 0);
+}
+
+int coroutine_fn bdrv_co_write_zeroes(BlockDriverState *bs,
+                                      int64_t sector_num, int nb_sectors)
+{
+    trace_bdrv_co_write_zeroes(bs, sector_num, nb_sectors);
+
+    return bdrv_co_do_writev(bs, sector_num, nb_sectors, NULL,
+                             BDRV_REQ_ZERO_WRITE);
+}
+
+/**
+ * Truncate file to 'offset' bytes (needed only for file protocols)
+ */
+int bdrv_truncate(BlockDriverState *bs, int64_t offset)
+{
+    BlockDriver *drv = bs->drv;
+    int ret;
+    if (!drv)
+        return -ENOMEDIUM;
+    if (!drv->bdrv_truncate)
+        return -ENOTSUP;
+    if (bs->read_only)
+        return -EACCES;
+    if (bdrv_in_use(bs))
+        return -EBUSY;
+    ret = drv->bdrv_truncate(bs, offset);
+    if (ret == 0) {
+        ret = refresh_total_sectors(bs, offset >> BDRV_SECTOR_BITS);
+        bdrv_dev_resize_cb(bs);
+    }
+    return ret;
+}
+
+/**
+ * Length of a allocated file in bytes. Sparse files are counted by actual
+ * allocated space. Return < 0 if error or unknown.
+ */
+int64_t bdrv_get_allocated_file_size(BlockDriverState *bs)
+{
+    BlockDriver *drv = bs->drv;
+    if (!drv) {
+        return -ENOMEDIUM;
+    }
+    if (drv->bdrv_get_allocated_file_size) {
+        return drv->bdrv_get_allocated_file_size(bs);
+    }
+    if (bs->file) {
+        return bdrv_get_allocated_file_size(bs->file);
+    }
+    return -ENOTSUP;
+}
+
+/**
+ * Length of a file in bytes. Return < 0 if error or unknown.
+ */
+int64_t bdrv_getlength(BlockDriverState *bs)
+{
+    BlockDriver *drv = bs->drv;
+    if (!drv)
+        return -ENOMEDIUM;
+
+    if (bs->growable || bdrv_dev_has_removable_media(bs)) {
+        if (drv->bdrv_getlength) {
+            return drv->bdrv_getlength(bs);
+        }
+    }
+    return bs->total_sectors * BDRV_SECTOR_SIZE;
+}
+
+/* return 0 as number of sectors if no device present or error */
+void bdrv_get_geometry(BlockDriverState *bs, uint64_t *nb_sectors_ptr)
+{
+    int64_t length;
+    length = bdrv_getlength(bs);
+    if (length < 0)
+        length = 0;
+    else
+        length = length >> BDRV_SECTOR_BITS;
+    *nb_sectors_ptr = length;
+}
+
+/* throttling disk io limits */
+void bdrv_set_io_limits(BlockDriverState *bs,
+                        BlockIOLimit *io_limits)
+{
+    bs->io_limits = *io_limits;
+    bs->io_limits_enabled = bdrv_io_limits_enabled(bs);
+}
+
+void bdrv_set_on_error(BlockDriverState *bs, BlockdevOnError on_read_error,
+                       BlockdevOnError on_write_error)
+{
+    bs->on_read_error = on_read_error;
+    bs->on_write_error = on_write_error;
+}
+
+BlockdevOnError bdrv_get_on_error(BlockDriverState *bs, bool is_read)
+{
+    return is_read ? bs->on_read_error : bs->on_write_error;
+}
+
+BlockErrorAction bdrv_get_error_action(BlockDriverState *bs, bool is_read, int error)
+{
+    BlockdevOnError on_err = is_read ? bs->on_read_error : bs->on_write_error;
+
+    switch (on_err) {
+    case BLOCKDEV_ON_ERROR_ENOSPC:
+        return (error == ENOSPC) ? BDRV_ACTION_STOP : BDRV_ACTION_REPORT;
+    case BLOCKDEV_ON_ERROR_STOP:
+        return BDRV_ACTION_STOP;
+    case BLOCKDEV_ON_ERROR_REPORT:
+        return BDRV_ACTION_REPORT;
+    case BLOCKDEV_ON_ERROR_IGNORE:
+        return BDRV_ACTION_IGNORE;
+    default:
+        abort();
+    }
+}
+
+/* This is done by device models because, while the block layer knows
+ * about the error, it does not know whether an operation comes from
+ * the device or the block layer (from a job, for example).
+ */
+void bdrv_error_action(BlockDriverState *bs, BlockErrorAction action,
+                       bool is_read, int error)
+{
+    assert(error >= 0);
+    bdrv_emit_qmp_error_event(bs, QEVENT_BLOCK_IO_ERROR, action, is_read);
+    if (action == BDRV_ACTION_STOP) {
+        vm_stop(RUN_STATE_IO_ERROR);
+        bdrv_iostatus_set_err(bs, error);
+    }
+}
+
+int bdrv_is_read_only(BlockDriverState *bs)
+{
+    return bs->read_only;
+}
+
+int bdrv_is_sg(BlockDriverState *bs)
+{
+    return bs->sg;
+}
+
+int bdrv_enable_write_cache(BlockDriverState *bs)
+{
+    return bs->enable_write_cache;
+}
+
+void bdrv_set_enable_write_cache(BlockDriverState *bs, bool wce)
+{
+    bs->enable_write_cache = wce;
+
+    /* so a reopen() will preserve wce */
+    if (wce) {
+        bs->open_flags |= BDRV_O_CACHE_WB;
+    } else {
+        bs->open_flags &= ~BDRV_O_CACHE_WB;
+    }
+}
+
+int bdrv_is_encrypted(BlockDriverState *bs)
+{
+    if (bs->backing_hd && bs->backing_hd->encrypted)
+        return 1;
+    return bs->encrypted;
+}
+
+int bdrv_key_required(BlockDriverState *bs)
+{
+    BlockDriverState *backing_hd = bs->backing_hd;
+
+    if (backing_hd && backing_hd->encrypted && !backing_hd->valid_key)
+        return 1;
+    return (bs->encrypted && !bs->valid_key);
+}
+
+int bdrv_set_key(BlockDriverState *bs, const char *key)
+{
+    int ret;
+    if (bs->backing_hd && bs->backing_hd->encrypted) {
+        ret = bdrv_set_key(bs->backing_hd, key);
+        if (ret < 0)
+            return ret;
+        if (!bs->encrypted)
+            return 0;
+    }
+    if (!bs->encrypted) {
+        return -EINVAL;
+    } else if (!bs->drv || !bs->drv->bdrv_set_key) {
+        return -ENOMEDIUM;
+    }
+    ret = bs->drv->bdrv_set_key(bs, key);
+    if (ret < 0) {
+        bs->valid_key = 0;
+    } else if (!bs->valid_key) {
+        bs->valid_key = 1;
+        /* call the change callback now, we skipped it on open */
+        bdrv_dev_change_media_cb(bs, true);
+    }
+    return ret;
+}
+
+const char *bdrv_get_format_name(BlockDriverState *bs)
+{
+    return bs->drv ? bs->drv->format_name : NULL;
+}
+
+void bdrv_iterate_format(void (*it)(void *opaque, const char *name),
+                         void *opaque)
+{
+    BlockDriver *drv;
+
+    QLIST_FOREACH(drv, &bdrv_drivers, list) {
+        it(opaque, drv->format_name);
+    }
+}
+
+BlockDriverState *bdrv_find(const char *name)
+{
+    BlockDriverState *bs;
+
+    QTAILQ_FOREACH(bs, &bdrv_states, list) {
+        if (!strcmp(name, bs->device_name)) {
+            return bs;
+        }
+    }
+    return NULL;
+}
+
+BlockDriverState *bdrv_next(BlockDriverState *bs)
+{
+    if (!bs) {
+        return QTAILQ_FIRST(&bdrv_states);
+    }
+    return QTAILQ_NEXT(bs, list);
+}
+
+void bdrv_iterate(void (*it)(void *opaque, BlockDriverState *bs), void *opaque)
+{
+    BlockDriverState *bs;
+
+    QTAILQ_FOREACH(bs, &bdrv_states, list) {
+        it(opaque, bs);
+    }
+}
+
+const char *bdrv_get_device_name(BlockDriverState *bs)
+{
+    return bs->device_name;
+}
+
+int bdrv_get_flags(BlockDriverState *bs)
+{
+    return bs->open_flags;
+}
+
+int bdrv_flush_all(void)
+{
+    BlockDriverState *bs;
+    int result = 0;
+
+    QTAILQ_FOREACH(bs, &bdrv_states, list) {
+        int ret = bdrv_flush(bs);
+        if (ret < 0 && !result) {
+            result = ret;
+        }
+    }
+
+    return result;
+}
+
+int bdrv_has_zero_init_1(BlockDriverState *bs)
+{
+    return 1;
+}
+
+int bdrv_has_zero_init(BlockDriverState *bs)
+{
+    assert(bs->drv);
+
+    if (bs->drv->bdrv_has_zero_init) {
+        return bs->drv->bdrv_has_zero_init(bs);
+    }
+
+    /* safe default */
+    return 0;
+}
+
+typedef struct BdrvCoIsAllocatedData {
+    BlockDriverState *bs;
+    BlockDriverState *base;
+    int64_t sector_num;
+    int nb_sectors;
+    int *pnum;
+    int ret;
+    bool done;
+} BdrvCoIsAllocatedData;
+
+/*
+ * Returns true iff the specified sector is present in the disk image. Drivers
+ * not implementing the functionality are assumed to not support backing files,
+ * hence all their sectors are reported as allocated.
+ *
+ * If 'sector_num' is beyond the end of the disk image the return value is 0
+ * and 'pnum' is set to 0.
+ *
+ * 'pnum' is set to the number of sectors (including and immediately following
+ * the specified sector) that are known to be in the same
+ * allocated/unallocated state.
+ *
+ * 'nb_sectors' is the max value 'pnum' should be set to.  If nb_sectors goes
+ * beyond the end of the disk image it will be clamped.
+ */
+int coroutine_fn bdrv_co_is_allocated(BlockDriverState *bs, int64_t sector_num,
+                                      int nb_sectors, int *pnum)
+{
+    int64_t n;
+
+    if (sector_num >= bs->total_sectors) {
+        *pnum = 0;
+        return 0;
+    }
+
+    n = bs->total_sectors - sector_num;
+    if (n < nb_sectors) {
+        nb_sectors = n;
+    }
+
+    if (!bs->drv->bdrv_co_is_allocated) {
+        *pnum = nb_sectors;
+        return 1;
+    }
+
+    return bs->drv->bdrv_co_is_allocated(bs, sector_num, nb_sectors, pnum);
+}
+
+/* Coroutine wrapper for bdrv_is_allocated() */
+static void coroutine_fn bdrv_is_allocated_co_entry(void *opaque)
+{
+    BdrvCoIsAllocatedData *data = opaque;
+    BlockDriverState *bs = data->bs;
+
+    data->ret = bdrv_co_is_allocated(bs, data->sector_num, data->nb_sectors,
+                                     data->pnum);
+    data->done = true;
+}
+
+/*
+ * Synchronous wrapper around bdrv_co_is_allocated().
+ *
+ * See bdrv_co_is_allocated() for details.
+ */
+int bdrv_is_allocated(BlockDriverState *bs, int64_t sector_num, int nb_sectors,
+                      int *pnum)
+{
+    Coroutine *co;
+    BdrvCoIsAllocatedData data = {
+        .bs = bs,
+        .sector_num = sector_num,
+        .nb_sectors = nb_sectors,
+        .pnum = pnum,
+        .done = false,
+    };
+
+    co = qemu_coroutine_create(bdrv_is_allocated_co_entry);
+    qemu_coroutine_enter(co, &data);
+    while (!data.done) {
+        qemu_aio_wait();
+    }
+    return data.ret;
+}
+
+/*
+ * Given an image chain: ... -> [BASE] -> [INTER1] -> [INTER2] -> [TOP]
+ *
+ * Return true if the given sector is allocated in any image between
+ * BASE and TOP (inclusive).  BASE can be NULL to check if the given
+ * sector is allocated in any image of the chain.  Return false otherwise.
+ *
+ * 'pnum' is set to the number of sectors (including and immediately following
+ *  the specified sector) that are known to be in the same
+ *  allocated/unallocated state.
+ *
+ */
+int coroutine_fn bdrv_co_is_allocated_above(BlockDriverState *top,
+                                            BlockDriverState *base,
+                                            int64_t sector_num,
+                                            int nb_sectors, int *pnum)
+{
+    BlockDriverState *intermediate;
+    int ret, n = nb_sectors;
+
+    intermediate = top;
+    while (intermediate && intermediate != base) {
+        int pnum_inter;
+        ret = bdrv_co_is_allocated(intermediate, sector_num, nb_sectors,
+                                   &pnum_inter);
+        if (ret < 0) {
+            return ret;
+        } else if (ret) {
+            *pnum = pnum_inter;
+            return 1;
+        }
+
+        /*
+         * [sector_num, nb_sectors] is unallocated on top but intermediate
+         * might have
+         *
+         * [sector_num+x, nr_sectors] allocated.
+         */
+        if (n > pnum_inter &&
+            (intermediate == top ||
+             sector_num + pnum_inter < intermediate->total_sectors)) {
+            n = pnum_inter;
+        }
+
+        intermediate = intermediate->backing_hd;
+    }
+
+    *pnum = n;
+    return 0;
+}
+
+/* Coroutine wrapper for bdrv_is_allocated_above() */
+static void coroutine_fn bdrv_is_allocated_above_co_entry(void *opaque)
+{
+    BdrvCoIsAllocatedData *data = opaque;
+    BlockDriverState *top = data->bs;
+    BlockDriverState *base = data->base;
+
+    data->ret = bdrv_co_is_allocated_above(top, base, data->sector_num,
+                                           data->nb_sectors, data->pnum);
+    data->done = true;
+}
+
+/*
+ * Synchronous wrapper around bdrv_co_is_allocated_above().
+ *
+ * See bdrv_co_is_allocated_above() for details.
+ */
+int bdrv_is_allocated_above(BlockDriverState *top, BlockDriverState *base,
+                            int64_t sector_num, int nb_sectors, int *pnum)
+{
+    Coroutine *co;
+    BdrvCoIsAllocatedData data = {
+        .bs = top,
+        .base = base,
+        .sector_num = sector_num,
+        .nb_sectors = nb_sectors,
+        .pnum = pnum,
+        .done = false,
+    };
+
+    co = qemu_coroutine_create(bdrv_is_allocated_above_co_entry);
+    qemu_coroutine_enter(co, &data);
+    while (!data.done) {
+        qemu_aio_wait();
+    }
+    return data.ret;
+}
+
+const char *bdrv_get_encrypted_filename(BlockDriverState *bs)
+{
+    if (bs->backing_hd && bs->backing_hd->encrypted)
+        return bs->backing_file;
+    else if (bs->encrypted)
+        return bs->filename;
+    else
+        return NULL;
+}
+
+void bdrv_get_backing_filename(BlockDriverState *bs,
+                               char *filename, int filename_size)
+{
+    pstrcpy(filename, filename_size, bs->backing_file);
+}
+
+int bdrv_write_compressed(BlockDriverState *bs, int64_t sector_num,
+                          const uint8_t *buf, int nb_sectors)
+{
+    BlockDriver *drv = bs->drv;
+    if (!drv)
+        return -ENOMEDIUM;
+    if (!drv->bdrv_write_compressed)
+        return -ENOTSUP;
+    if (bdrv_check_request(bs, sector_num, nb_sectors))
+        return -EIO;
+
+    assert(!bs->dirty_bitmap);
+
+    return drv->bdrv_write_compressed(bs, sector_num, buf, nb_sectors);
+}
+
+int bdrv_get_info(BlockDriverState *bs, BlockDriverInfo *bdi)
+{
+    BlockDriver *drv = bs->drv;
+    if (!drv)
+        return -ENOMEDIUM;
+    if (!drv->bdrv_get_info)
+        return -ENOTSUP;
+    memset(bdi, 0, sizeof(*bdi));
+    return drv->bdrv_get_info(bs, bdi);
+}
+
+int bdrv_save_vmstate(BlockDriverState *bs, const uint8_t *buf,
+                      int64_t pos, int size)
+{
+    QEMUIOVector qiov;
+    struct iovec iov = {
+        .iov_base   = (void *) buf,
+        .iov_len    = size,
+    };
+
+    qemu_iovec_init_external(&qiov, &iov, 1);
+    return bdrv_writev_vmstate(bs, &qiov, pos);
+}
+
+int bdrv_writev_vmstate(BlockDriverState *bs, QEMUIOVector *qiov, int64_t pos)
+{
+    BlockDriver *drv = bs->drv;
+
+    if (!drv) {
+        return -ENOMEDIUM;
+    } else if (drv->bdrv_save_vmstate) {
+        return drv->bdrv_save_vmstate(bs, qiov, pos);
+    } else if (bs->file) {
+        return bdrv_writev_vmstate(bs->file, qiov, pos);
+    }
+
+    return -ENOTSUP;
+}
+
+int bdrv_load_vmstate(BlockDriverState *bs, uint8_t *buf,
+                      int64_t pos, int size)
+{
+    BlockDriver *drv = bs->drv;
+    if (!drv)
+        return -ENOMEDIUM;
+    if (drv->bdrv_load_vmstate)
+        return drv->bdrv_load_vmstate(bs, buf, pos, size);
+    if (bs->file)
+        return bdrv_load_vmstate(bs->file, buf, pos, size);
+    return -ENOTSUP;
+}
+
+void bdrv_debug_event(BlockDriverState *bs, BlkDebugEvent event)
+{
+    if (!bs || !bs->drv || !bs->drv->bdrv_debug_event) {
+        return;
+    }
+
+    bs->drv->bdrv_debug_event(bs, event);
+}
+
+int bdrv_debug_breakpoint(BlockDriverState *bs, const char *event,
+                          const char *tag)
+{
+    while (bs && bs->drv && !bs->drv->bdrv_debug_breakpoint) {
+        bs = bs->file;
+    }
+
+    if (bs && bs->drv && bs->drv->bdrv_debug_breakpoint) {
+        return bs->drv->bdrv_debug_breakpoint(bs, event, tag);
+    }
+
+    return -ENOTSUP;
+}
+
+int bdrv_debug_resume(BlockDriverState *bs, const char *tag)
+{
+    while (bs && bs->drv && !bs->drv->bdrv_debug_resume) {
+        bs = bs->file;
+    }
+
+    if (bs && bs->drv && bs->drv->bdrv_debug_resume) {
+        return bs->drv->bdrv_debug_resume(bs, tag);
+    }
+
+    return -ENOTSUP;
+}
+
+bool bdrv_debug_is_suspended(BlockDriverState *bs, const char *tag)
+{
+    while (bs && bs->drv && !bs->drv->bdrv_debug_is_suspended) {
+        bs = bs->file;
+    }
+
+    if (bs && bs->drv && bs->drv->bdrv_debug_is_suspended) {
+        return bs->drv->bdrv_debug_is_suspended(bs, tag);
+    }
+
+    return false;
+}
+
+int bdrv_is_snapshot(BlockDriverState *bs)
+{
+    return !!(bs->open_flags & BDRV_O_SNAPSHOT);
+}
+
+/* backing_file can either be relative, or absolute, or a protocol.  If it is
+ * relative, it must be relative to the chain.  So, passing in bs->filename
+ * from a BDS as backing_file should not be done, as that may be relative to
+ * the CWD rather than the chain. */
+BlockDriverState *bdrv_find_backing_image(BlockDriverState *bs,
+        const char *backing_file)
+{
+    char *filename_full = NULL;
+    char *backing_file_full = NULL;
+    char *filename_tmp = NULL;
+    int is_protocol = 0;
+    BlockDriverState *curr_bs = NULL;
+    BlockDriverState *retval = NULL;
+
+    if (!bs || !bs->drv || !backing_file) {
+        return NULL;
+    }
+
+    filename_full     = g_malloc(PATH_MAX);
+    backing_file_full = g_malloc(PATH_MAX);
+    filename_tmp      = g_malloc(PATH_MAX);
+
+    is_protocol = path_has_protocol(backing_file);
+
+    for (curr_bs = bs; curr_bs->backing_hd; curr_bs = curr_bs->backing_hd) {
+
+        /* If either of the filename paths is actually a protocol, then
+         * compare unmodified paths; otherwise make paths relative */
+        if (is_protocol || path_has_protocol(curr_bs->backing_file)) {
+            if (strcmp(backing_file, curr_bs->backing_file) == 0) {
+                retval = curr_bs->backing_hd;
+                break;
+            }
+        } else {
+            /* If not an absolute filename path, make it relative to the current
+             * image's filename path */
+            path_combine(filename_tmp, PATH_MAX, curr_bs->filename,
+                         backing_file);
+
+            /* We are going to compare absolute pathnames */
+            if (!realpath(filename_tmp, filename_full)) {
+                continue;
+            }
+
+            /* We need to make sure the backing filename we are comparing against
+             * is relative to the current image filename (or absolute) */
+            path_combine(filename_tmp, PATH_MAX, curr_bs->filename,
+                         curr_bs->backing_file);
+
+            if (!realpath(filename_tmp, backing_file_full)) {
+                continue;
+            }
+
+            if (strcmp(backing_file_full, filename_full) == 0) {
+                retval = curr_bs->backing_hd;
+                break;
+            }
+        }
+    }
+
+    g_free(filename_full);
+    g_free(backing_file_full);
+    g_free(filename_tmp);
+    return retval;
+}
+
+int bdrv_get_backing_file_depth(BlockDriverState *bs)
+{
+    if (!bs->drv) {
+        return 0;
+    }
+
+    if (!bs->backing_hd) {
+        return 0;
+    }
+
+    return 1 + bdrv_get_backing_file_depth(bs->backing_hd);
+}
+
+BlockDriverState *bdrv_find_base(BlockDriverState *bs)
+{
+    BlockDriverState *curr_bs = NULL;
+
+    if (!bs) {
+        return NULL;
+    }
+
+    curr_bs = bs;
+
+    while (curr_bs->backing_hd) {
+        curr_bs = curr_bs->backing_hd;
+    }
+    return curr_bs;
+}
+
+/**************************************************************/
+/* async I/Os */
+
+BlockDriverAIOCB *bdrv_aio_readv(BlockDriverState *bs, int64_t sector_num,
+                                 QEMUIOVector *qiov, int nb_sectors,
+                                 BlockDriverCompletionFunc *cb, void *opaque)
+{
+    trace_bdrv_aio_readv(bs, sector_num, nb_sectors, opaque);
+
+    return bdrv_co_aio_rw_vector(bs, sector_num, qiov, nb_sectors,
+                                 cb, opaque, false);
+}
+
+BlockDriverAIOCB *bdrv_aio_writev(BlockDriverState *bs, int64_t sector_num,
+                                  QEMUIOVector *qiov, int nb_sectors,
+                                  BlockDriverCompletionFunc *cb, void *opaque)
+{
+    trace_bdrv_aio_writev(bs, sector_num, nb_sectors, opaque);
+
+    return bdrv_co_aio_rw_vector(bs, sector_num, qiov, nb_sectors,
+                                 cb, opaque, true);
+}
+
+
+typedef struct MultiwriteCB {
+    int error;
+    int num_requests;
+    int num_callbacks;
+    struct {
+        BlockDriverCompletionFunc *cb;
+        void *opaque;
+        QEMUIOVector *free_qiov;
+    } callbacks[];
+} MultiwriteCB;
+
+static void multiwrite_user_cb(MultiwriteCB *mcb)
+{
+    int i;
+
+    for (i = 0; i < mcb->num_callbacks; i++) {
+        mcb->callbacks[i].cb(mcb->callbacks[i].opaque, mcb->error);
+        if (mcb->callbacks[i].free_qiov) {
+            qemu_iovec_destroy(mcb->callbacks[i].free_qiov);
+        }
+        g_free(mcb->callbacks[i].free_qiov);
+    }
+}
+
+static void multiwrite_cb(void *opaque, int ret)
+{
+    MultiwriteCB *mcb = opaque;
+
+    trace_multiwrite_cb(mcb, ret);
+
+    if (ret < 0 && !mcb->error) {
+        mcb->error = ret;
+    }
+
+    mcb->num_requests--;
+    if (mcb->num_requests == 0) {
+        multiwrite_user_cb(mcb);
+        g_free(mcb);
+    }
+}
+
+static int multiwrite_req_compare(const void *a, const void *b)
+{
+    const BlockRequest *req1 = a, *req2 = b;
+
+    /*
+     * Note that we can't simply subtract req2->sector from req1->sector
+     * here as that could overflow the return value.
+     */
+    if (req1->sector > req2->sector) {
+        return 1;
+    } else if (req1->sector < req2->sector) {
+        return -1;
+    } else {
+        return 0;
+    }
+}
+
+/*
+ * Takes a bunch of requests and tries to merge them. Returns the number of
+ * requests that remain after merging.
+ */
+static int multiwrite_merge(BlockDriverState *bs, BlockRequest *reqs,
+    int num_reqs, MultiwriteCB *mcb)
+{
+    int i, outidx;
+
+    // Sort requests by start sector
+    qsort(reqs, num_reqs, sizeof(*reqs), &multiwrite_req_compare);
+
+    // Check if adjacent requests touch the same clusters. If so, combine them,
+    // filling up gaps with zero sectors.
+    outidx = 0;
+    for (i = 1; i < num_reqs; i++) {
+        int merge = 0;
+        int64_t oldreq_last = reqs[outidx].sector + reqs[outidx].nb_sectors;
+
+        // Handle exactly sequential writes and overlapping writes.
+        if (reqs[i].sector <= oldreq_last) {
+            merge = 1;
+        }
+
+        if (reqs[outidx].qiov->niov + reqs[i].qiov->niov + 1 > IOV_MAX) {
+            merge = 0;
+        }
+
+        if (merge) {
+            size_t size;
+            QEMUIOVector *qiov = g_malloc0(sizeof(*qiov));
+            qemu_iovec_init(qiov,
+                reqs[outidx].qiov->niov + reqs[i].qiov->niov + 1);
+
+            // Add the first request to the merged one. If the requests are
+            // overlapping, drop the last sectors of the first request.
+            size = (reqs[i].sector - reqs[outidx].sector) << 9;
+            qemu_iovec_concat(qiov, reqs[outidx].qiov, 0, size);
+
+            // We should need to add any zeros between the two requests
+            assert (reqs[i].sector <= oldreq_last);
+
+            // Add the second request
+            qemu_iovec_concat(qiov, reqs[i].qiov, 0, reqs[i].qiov->size);
+
+            reqs[outidx].nb_sectors = qiov->size >> 9;
+            reqs[outidx].qiov = qiov;
+
+            mcb->callbacks[i].free_qiov = reqs[outidx].qiov;
+        } else {
+            outidx++;
+            reqs[outidx].sector     = reqs[i].sector;
+            reqs[outidx].nb_sectors = reqs[i].nb_sectors;
+            reqs[outidx].qiov       = reqs[i].qiov;
+        }
+    }
+
+    return outidx + 1;
+}
+
+/*
+ * Submit multiple AIO write requests at once.
+ *
+ * On success, the function returns 0 and all requests in the reqs array have
+ * been submitted. In error case this function returns -1, and any of the
+ * requests may or may not be submitted yet. In particular, this means that the
+ * callback will be called for some of the requests, for others it won't. The
+ * caller must check the error field of the BlockRequest to wait for the right
+ * callbacks (if error != 0, no callback will be called).
+ *
+ * The implementation may modify the contents of the reqs array, e.g. to merge
+ * requests. However, the fields opaque and error are left unmodified as they
+ * are used to signal failure for a single request to the caller.
+ */
+int bdrv_aio_multiwrite(BlockDriverState *bs, BlockRequest *reqs, int num_reqs)
+{
+    MultiwriteCB *mcb;
+    int i;
+
+    /* don't submit writes if we don't have a medium */
+    if (bs->drv == NULL) {
+        for (i = 0; i < num_reqs; i++) {
+            reqs[i].error = -ENOMEDIUM;
+        }
+        return -1;
+    }
+
+    if (num_reqs == 0) {
+        return 0;
+    }
+
+    // Create MultiwriteCB structure
+    mcb = g_malloc0(sizeof(*mcb) + num_reqs * sizeof(*mcb->callbacks));
+    mcb->num_requests = 0;
+    mcb->num_callbacks = num_reqs;
+
+    for (i = 0; i < num_reqs; i++) {
+        mcb->callbacks[i].cb = reqs[i].cb;
+        mcb->callbacks[i].opaque = reqs[i].opaque;
+    }
+
+    // Check for mergable requests
+    num_reqs = multiwrite_merge(bs, reqs, num_reqs, mcb);
+
+    trace_bdrv_aio_multiwrite(mcb, mcb->num_callbacks, num_reqs);
+
+    /* Run the aio requests. */
+    mcb->num_requests = num_reqs;
+    for (i = 0; i < num_reqs; i++) {
+        bdrv_aio_writev(bs, reqs[i].sector, reqs[i].qiov,
+            reqs[i].nb_sectors, multiwrite_cb, mcb);
+    }
+
+    return 0;
+}
+
+void bdrv_aio_cancel(BlockDriverAIOCB *acb)
+{
+    acb->aiocb_info->cancel(acb);
+}
+
+/* block I/O throttling */
+static bool bdrv_exceed_bps_limits(BlockDriverState *bs, int nb_sectors,
+                 bool is_write, double elapsed_time, uint64_t *wait)
+{
+    uint64_t bps_limit = 0;
+    uint64_t extension;
+    double   bytes_limit, bytes_base, bytes_res;
+    double   slice_time, wait_time;
+
+    if (bs->io_limits.bps[BLOCK_IO_LIMIT_TOTAL]) {
+        bps_limit = bs->io_limits.bps[BLOCK_IO_LIMIT_TOTAL];
+    } else if (bs->io_limits.bps[is_write]) {
+        bps_limit = bs->io_limits.bps[is_write];
+    } else {
+        if (wait) {
+            *wait = 0;
+        }
+
+        return false;
+    }
+
+    slice_time = bs->slice_end - bs->slice_start;
+    slice_time /= (NANOSECONDS_PER_SECOND);
+    bytes_limit = bps_limit * slice_time;
+    bytes_base  = bs->slice_submitted.bytes[is_write];
+    if (bs->io_limits.bps[BLOCK_IO_LIMIT_TOTAL]) {
+        bytes_base += bs->slice_submitted.bytes[!is_write];
+    }
+
+    /* bytes_base: the bytes of data which have been read/written; and
+     *             it is obtained from the history statistic info.
+     * bytes_res: the remaining bytes of data which need to be read/written.
+     * (bytes_base + bytes_res) / bps_limit: used to calcuate
+     *             the total time for completing reading/writting all data.
+     */
+    bytes_res   = (unsigned) nb_sectors * BDRV_SECTOR_SIZE;
+
+    if (bytes_base + bytes_res <= bytes_limit) {
+        if (wait) {
+            *wait = 0;
+        }
+
+        return false;
+    }
+
+    /* Calc approx time to dispatch */
+    wait_time = (bytes_base + bytes_res) / bps_limit - elapsed_time;
+
+    /* When the I/O rate at runtime exceeds the limits,
+     * bs->slice_end need to be extended in order that the current statistic
+     * info can be kept until the timer fire, so it is increased and tuned
+     * based on the result of experiment.
+     */
+    extension = wait_time * NANOSECONDS_PER_SECOND;
+    extension = DIV_ROUND_UP(extension, BLOCK_IO_SLICE_TIME) *
+                BLOCK_IO_SLICE_TIME;
+    bs->slice_end += extension;
+    if (wait) {
+        *wait = wait_time * NANOSECONDS_PER_SECOND;
+    }
+
+    return true;
+}
+
+static bool bdrv_exceed_iops_limits(BlockDriverState *bs, bool is_write,
+                             double elapsed_time, uint64_t *wait)
+{
+    uint64_t iops_limit = 0;
+    double   ios_limit, ios_base;
+    double   slice_time, wait_time;
+
+    if (bs->io_limits.iops[BLOCK_IO_LIMIT_TOTAL]) {
+        iops_limit = bs->io_limits.iops[BLOCK_IO_LIMIT_TOTAL];
+    } else if (bs->io_limits.iops[is_write]) {
+        iops_limit = bs->io_limits.iops[is_write];
+    } else {
+        if (wait) {
+            *wait = 0;
+        }
+
+        return false;
+    }
+
+    slice_time = bs->slice_end - bs->slice_start;
+    slice_time /= (NANOSECONDS_PER_SECOND);
+    ios_limit  = iops_limit * slice_time;
+    ios_base   = bs->slice_submitted.ios[is_write];
+    if (bs->io_limits.iops[BLOCK_IO_LIMIT_TOTAL]) {
+        ios_base += bs->slice_submitted.ios[!is_write];
+    }
+
+    if (ios_base + 1 <= ios_limit) {
+        if (wait) {
+            *wait = 0;
+        }
+
+        return false;
+    }
+
+    /* Calc approx time to dispatch, in seconds */
+    wait_time = (ios_base + 1) / iops_limit;
+    if (wait_time > elapsed_time) {
+        wait_time = wait_time - elapsed_time;
+    } else {
+        wait_time = 0;
+    }
+
+    /* Exceeded current slice, extend it by another slice time */
+    bs->slice_end += BLOCK_IO_SLICE_TIME;
+    if (wait) {
+        *wait = wait_time * NANOSECONDS_PER_SECOND;
+    }
+
+    return true;
+}
+
+static bool bdrv_exceed_io_limits(BlockDriverState *bs, int nb_sectors,
+                           bool is_write, int64_t *wait)
+{
+    int64_t  now, max_wait;
+    uint64_t bps_wait = 0, iops_wait = 0;
+    double   elapsed_time;
+    int      bps_ret, iops_ret;
+
+    now = qemu_get_clock_ns(vm_clock);
+    if (now > bs->slice_end) {
+        bs->slice_start = now;
+        bs->slice_end   = now + BLOCK_IO_SLICE_TIME;
+        memset(&bs->slice_submitted, 0, sizeof(bs->slice_submitted));
+    }
+
+    elapsed_time  = now - bs->slice_start;
+    elapsed_time  /= (NANOSECONDS_PER_SECOND);
+
+    bps_ret  = bdrv_exceed_bps_limits(bs, nb_sectors,
+                                      is_write, elapsed_time, &bps_wait);
+    iops_ret = bdrv_exceed_iops_limits(bs, is_write,
+                                      elapsed_time, &iops_wait);
+    if (bps_ret || iops_ret) {
+        max_wait = bps_wait > iops_wait ? bps_wait : iops_wait;
+        if (wait) {
+            *wait = max_wait;
+        }
+
+        now = qemu_get_clock_ns(vm_clock);
+        if (bs->slice_end < now + max_wait) {
+            bs->slice_end = now + max_wait;
+        }
+
+        return true;
+    }
+
+    if (wait) {
+        *wait = 0;
+    }
+
+    bs->slice_submitted.bytes[is_write] += (int64_t)nb_sectors *
+                                           BDRV_SECTOR_SIZE;
+    bs->slice_submitted.ios[is_write]++;
+
+    return false;
+}
+
+/**************************************************************/
+/* async block device emulation */
+
+typedef struct BlockDriverAIOCBSync {
+    BlockDriverAIOCB common;
+    QEMUBH *bh;
+    int ret;
+    /* vector translation state */
+    QEMUIOVector *qiov;
+    uint8_t *bounce;
+    int is_write;
+} BlockDriverAIOCBSync;
+
+static void bdrv_aio_cancel_em(BlockDriverAIOCB *blockacb)
+{
+    BlockDriverAIOCBSync *acb =
+        container_of(blockacb, BlockDriverAIOCBSync, common);
+    qemu_bh_delete(acb->bh);
+    acb->bh = NULL;
+    qemu_aio_release(acb);
+}
+
+static const AIOCBInfo bdrv_em_aiocb_info = {
+    .aiocb_size         = sizeof(BlockDriverAIOCBSync),
+    .cancel             = bdrv_aio_cancel_em,
+};
+
+static void bdrv_aio_bh_cb(void *opaque)
+{
+    BlockDriverAIOCBSync *acb = opaque;
+
+    if (!acb->is_write)
+        qemu_iovec_from_buf(acb->qiov, 0, acb->bounce, acb->qiov->size);
+    qemu_vfree(acb->bounce);
+    acb->common.cb(acb->common.opaque, acb->ret);
+    qemu_bh_delete(acb->bh);
+    acb->bh = NULL;
+    qemu_aio_release(acb);
+}
+
+static BlockDriverAIOCB *bdrv_aio_rw_vector(BlockDriverState *bs,
+                                            int64_t sector_num,
+                                            QEMUIOVector *qiov,
+                                            int nb_sectors,
+                                            BlockDriverCompletionFunc *cb,
+                                            void *opaque,
+                                            int is_write)
+
+{
+    BlockDriverAIOCBSync *acb;
+
+    acb = qemu_aio_get(&bdrv_em_aiocb_info, bs, cb, opaque);
+    acb->is_write = is_write;
+    acb->qiov = qiov;
+    acb->bounce = qemu_blockalign(bs, qiov->size);
+    acb->bh = qemu_bh_new(bdrv_aio_bh_cb, acb);
+
+    if (is_write) {
+        qemu_iovec_to_buf(acb->qiov, 0, acb->bounce, qiov->size);
+        acb->ret = bs->drv->bdrv_write(bs, sector_num, acb->bounce, nb_sectors);
+    } else {
+        acb->ret = bs->drv->bdrv_read(bs, sector_num, acb->bounce, nb_sectors);
+    }
+
+    qemu_bh_schedule(acb->bh);
+
+    return &acb->common;
+}
+
+static BlockDriverAIOCB *bdrv_aio_readv_em(BlockDriverState *bs,
+        int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
+        BlockDriverCompletionFunc *cb, void *opaque)
+{
+    return bdrv_aio_rw_vector(bs, sector_num, qiov, nb_sectors, cb, opaque, 0);
+}
+
+static BlockDriverAIOCB *bdrv_aio_writev_em(BlockDriverState *bs,
+        int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
+        BlockDriverCompletionFunc *cb, void *opaque)
+{
+    return bdrv_aio_rw_vector(bs, sector_num, qiov, nb_sectors, cb, opaque, 1);
+}
+
+
+typedef struct BlockDriverAIOCBCoroutine {
+    BlockDriverAIOCB common;
+    BlockRequest req;
+    bool is_write;
+    bool *done;
+    QEMUBH* bh;
+} BlockDriverAIOCBCoroutine;
+
+static void bdrv_aio_co_cancel_em(BlockDriverAIOCB *blockacb)
+{
+    BlockDriverAIOCBCoroutine *acb =
+        container_of(blockacb, BlockDriverAIOCBCoroutine, common);
+    bool done = false;
+
+    acb->done = &done;
+    while (!done) {
+        qemu_aio_wait();
+    }
+}
+
+static const AIOCBInfo bdrv_em_co_aiocb_info = {
+    .aiocb_size         = sizeof(BlockDriverAIOCBCoroutine),
+    .cancel             = bdrv_aio_co_cancel_em,
+};
+
+static void bdrv_co_em_bh(void *opaque)
+{
+    BlockDriverAIOCBCoroutine *acb = opaque;
+
+    acb->common.cb(acb->common.opaque, acb->req.error);
+
+    if (acb->done) {
+        *acb->done = true;
+    }
+
+    qemu_bh_delete(acb->bh);
+    qemu_aio_release(acb);
+}
+
+/* Invoke bdrv_co_do_readv/bdrv_co_do_writev */
+static void coroutine_fn bdrv_co_do_rw(void *opaque)
+{
+    BlockDriverAIOCBCoroutine *acb = opaque;
+    BlockDriverState *bs = acb->common.bs;
+
+    if (!acb->is_write) {
+        acb->req.error = bdrv_co_do_readv(bs, acb->req.sector,
+            acb->req.nb_sectors, acb->req.qiov, 0);
+    } else {
+        acb->req.error = bdrv_co_do_writev(bs, acb->req.sector,
+            acb->req.nb_sectors, acb->req.qiov, 0);
+    }
+
+    acb->bh = qemu_bh_new(bdrv_co_em_bh, acb);
+    qemu_bh_schedule(acb->bh);
+}
+
+static BlockDriverAIOCB *bdrv_co_aio_rw_vector(BlockDriverState *bs,
+                                               int64_t sector_num,
+                                               QEMUIOVector *qiov,
+                                               int nb_sectors,
+                                               BlockDriverCompletionFunc *cb,
+                                               void *opaque,
+                                               bool is_write)
+{
+    Coroutine *co;
+    BlockDriverAIOCBCoroutine *acb;
+
+    acb = qemu_aio_get(&bdrv_em_co_aiocb_info, bs, cb, opaque);
+    acb->req.sector = sector_num;
+    acb->req.nb_sectors = nb_sectors;
+    acb->req.qiov = qiov;
+    acb->is_write = is_write;
+    acb->done = NULL;
+
+    co = qemu_coroutine_create(bdrv_co_do_rw);
+    qemu_coroutine_enter(co, acb);
+
+    return &acb->common;
+}
+
+static void coroutine_fn bdrv_aio_flush_co_entry(void *opaque)
+{
+    BlockDriverAIOCBCoroutine *acb = opaque;
+    BlockDriverState *bs = acb->common.bs;
+
+    acb->req.error = bdrv_co_flush(bs);
+    acb->bh = qemu_bh_new(bdrv_co_em_bh, acb);
+    qemu_bh_schedule(acb->bh);
+}
+
+BlockDriverAIOCB *bdrv_aio_flush(BlockDriverState *bs,
+        BlockDriverCompletionFunc *cb, void *opaque)
+{
+    trace_bdrv_aio_flush(bs, opaque);
+
+    Coroutine *co;
+    BlockDriverAIOCBCoroutine *acb;
+
+    acb = qemu_aio_get(&bdrv_em_co_aiocb_info, bs, cb, opaque);
+    acb->done = NULL;
+
+    co = qemu_coroutine_create(bdrv_aio_flush_co_entry);
+    qemu_coroutine_enter(co, acb);
+
+    return &acb->common;
+}
+
+static void coroutine_fn bdrv_aio_discard_co_entry(void *opaque)
+{
+    BlockDriverAIOCBCoroutine *acb = opaque;
+    BlockDriverState *bs = acb->common.bs;
+
+    acb->req.error = bdrv_co_discard(bs, acb->req.sector, acb->req.nb_sectors);
+    acb->bh = qemu_bh_new(bdrv_co_em_bh, acb);
+    qemu_bh_schedule(acb->bh);
+}
+
+BlockDriverAIOCB *bdrv_aio_discard(BlockDriverState *bs,
+        int64_t sector_num, int nb_sectors,
+        BlockDriverCompletionFunc *cb, void *opaque)
+{
+    Coroutine *co;
+    BlockDriverAIOCBCoroutine *acb;
+
+    trace_bdrv_aio_discard(bs, sector_num, nb_sectors, opaque);
+
+    acb = qemu_aio_get(&bdrv_em_co_aiocb_info, bs, cb, opaque);
+    acb->req.sector = sector_num;
+    acb->req.nb_sectors = nb_sectors;
+    acb->done = NULL;
+    co = qemu_coroutine_create(bdrv_aio_discard_co_entry);
+    qemu_coroutine_enter(co, acb);
+
+    return &acb->common;
+}
+
+void bdrv_init(void)
+{
+    module_call_init(MODULE_INIT_BLOCK);
+}
+
+void bdrv_init_with_whitelist(void)
+{
+    use_bdrv_whitelist = 1;
+    bdrv_init();
+}
+
+void *qemu_aio_get(const AIOCBInfo *aiocb_info, BlockDriverState *bs,
+                   BlockDriverCompletionFunc *cb, void *opaque)
+{
+    BlockDriverAIOCB *acb;
+
+    acb = g_slice_alloc(aiocb_info->aiocb_size);
+    acb->aiocb_info = aiocb_info;
+    acb->bs = bs;
+    acb->cb = cb;
+    acb->opaque = opaque;
+    return acb;
+}
+
+void qemu_aio_release(void *p)
+{
+    BlockDriverAIOCB *acb = p;
+    g_slice_free1(acb->aiocb_info->aiocb_size, acb);
+}
+
+/**************************************************************/
+/* Coroutine block device emulation */
+
+typedef struct CoroutineIOCompletion {
+    Coroutine *coroutine;
+    int ret;
+} CoroutineIOCompletion;
+
+static void bdrv_co_io_em_complete(void *opaque, int ret)
+{
+    CoroutineIOCompletion *co = opaque;
+
+    co->ret = ret;
+    qemu_coroutine_enter(co->coroutine, NULL);
+}
+
+static int coroutine_fn bdrv_co_io_em(BlockDriverState *bs, int64_t sector_num,
+                                      int nb_sectors, QEMUIOVector *iov,
+                                      bool is_write)
+{
+    CoroutineIOCompletion co = {
+        .coroutine = qemu_coroutine_self(),
+    };
+    BlockDriverAIOCB *acb;
+
+    if (is_write) {
+        acb = bs->drv->bdrv_aio_writev(bs, sector_num, iov, nb_sectors,
+                                       bdrv_co_io_em_complete, &co);
+    } else {
+        acb = bs->drv->bdrv_aio_readv(bs, sector_num, iov, nb_sectors,
+                                      bdrv_co_io_em_complete, &co);
+    }
+
+    trace_bdrv_co_io_em(bs, sector_num, nb_sectors, is_write, acb);
+    if (!acb) {
+        return -EIO;
+    }
+    qemu_coroutine_yield();
+
+    return co.ret;
+}
+
+static int coroutine_fn bdrv_co_readv_em(BlockDriverState *bs,
+                                         int64_t sector_num, int nb_sectors,
+                                         QEMUIOVector *iov)
+{
+    return bdrv_co_io_em(bs, sector_num, nb_sectors, iov, false);
+}
+
+static int coroutine_fn bdrv_co_writev_em(BlockDriverState *bs,
+                                         int64_t sector_num, int nb_sectors,
+                                         QEMUIOVector *iov)
+{
+    return bdrv_co_io_em(bs, sector_num, nb_sectors, iov, true);
+}
+
+static void coroutine_fn bdrv_flush_co_entry(void *opaque)
+{
+    RwCo *rwco = opaque;
+
+    rwco->ret = bdrv_co_flush(rwco->bs);
+}
+
+int coroutine_fn bdrv_co_flush(BlockDriverState *bs)
+{
+    int ret;
+
+    if (!bs || !bdrv_is_inserted(bs) || bdrv_is_read_only(bs)) {
+        return 0;
+    }
+
+    /* Write back cached data to the OS even with cache=unsafe */
+    BLKDBG_EVENT(bs->file, BLKDBG_FLUSH_TO_OS);
+    if (bs->drv->bdrv_co_flush_to_os) {
+        ret = bs->drv->bdrv_co_flush_to_os(bs);
+        if (ret < 0) {
+            return ret;
+        }
+    }
+
+    /* But don't actually force it to the disk with cache=unsafe */
+    if (bs->open_flags & BDRV_O_NO_FLUSH) {
+        goto flush_parent;
+    }
+
+    BLKDBG_EVENT(bs->file, BLKDBG_FLUSH_TO_DISK);
+    if (bs->drv->bdrv_co_flush_to_disk) {
+        ret = bs->drv->bdrv_co_flush_to_disk(bs);
+    } else if (bs->drv->bdrv_aio_flush) {
+        BlockDriverAIOCB *acb;
+        CoroutineIOCompletion co = {
+            .coroutine = qemu_coroutine_self(),
+        };
+
+        acb = bs->drv->bdrv_aio_flush(bs, bdrv_co_io_em_complete, &co);
+        if (acb == NULL) {
+            ret = -EIO;
+        } else {
+            qemu_coroutine_yield();
+            ret = co.ret;
+        }
+    } else {
+        /*
+         * Some block drivers always operate in either writethrough or unsafe
+         * mode and don't support bdrv_flush therefore. Usually qemu doesn't
+         * know how the server works (because the behaviour is hardcoded or
+         * depends on server-side configuration), so we can't ensure that
+         * everything is safe on disk. Returning an error doesn't work because
+         * that would break guests even if the server operates in writethrough
+         * mode.
+         *
+         * Let's hope the user knows what he's doing.
+         */
+        ret = 0;
+    }
+    if (ret < 0) {
+        return ret;
+    }
+
+    /* Now flush the underlying protocol.  It will also have BDRV_O_NO_FLUSH
+     * in the case of cache=unsafe, so there are no useless flushes.
+     */
+flush_parent:
+    return bdrv_co_flush(bs->file);
+}
+
+void bdrv_invalidate_cache(BlockDriverState *bs)
+{
+    if (bs->drv && bs->drv->bdrv_invalidate_cache) {
+        bs->drv->bdrv_invalidate_cache(bs);
+    }
+}
+
+void bdrv_invalidate_cache_all(void)
+{
+    BlockDriverState *bs;
+
+    QTAILQ_FOREACH(bs, &bdrv_states, list) {
+        bdrv_invalidate_cache(bs);
+    }
+}
+
+void bdrv_clear_incoming_migration_all(void)
+{
+    BlockDriverState *bs;
+
+    QTAILQ_FOREACH(bs, &bdrv_states, list) {
+        bs->open_flags = bs->open_flags & ~(BDRV_O_INCOMING);
+    }
+}
+
+int bdrv_flush(BlockDriverState *bs)
+{
+    Coroutine *co;
+    RwCo rwco = {
+        .bs = bs,
+        .ret = NOT_DONE,
+    };
+
+    if (qemu_in_coroutine()) {
+        /* Fast-path if already in coroutine context */
+        bdrv_flush_co_entry(&rwco);
+    } else {
+        co = qemu_coroutine_create(bdrv_flush_co_entry);
+        qemu_coroutine_enter(co, &rwco);
+        while (rwco.ret == NOT_DONE) {
+            qemu_aio_wait();
+        }
+    }
+
+    return rwco.ret;
+}
+
+static void coroutine_fn bdrv_discard_co_entry(void *opaque)
+{
+    RwCo *rwco = opaque;
+
+    rwco->ret = bdrv_co_discard(rwco->bs, rwco->sector_num, rwco->nb_sectors);
+}
+
+int coroutine_fn bdrv_co_discard(BlockDriverState *bs, int64_t sector_num,
+                                 int nb_sectors)
+{
+    if (!bs->drv) {
+        return -ENOMEDIUM;
+    } else if (bdrv_check_request(bs, sector_num, nb_sectors)) {
+        return -EIO;
+    } else if (bs->read_only) {
+        return -EROFS;
+    }
+
+    if (bs->dirty_bitmap) {
+        bdrv_reset_dirty(bs, sector_num, nb_sectors);
+    }
+
+    /* Do nothing if disabled.  */
+    if (!(bs->open_flags & BDRV_O_UNMAP)) {
+        return 0;
+    }
+
+    if (bs->drv->bdrv_co_discard) {
+        return bs->drv->bdrv_co_discard(bs, sector_num, nb_sectors);
+    } else if (bs->drv->bdrv_aio_discard) {
+        BlockDriverAIOCB *acb;
+        CoroutineIOCompletion co = {
+            .coroutine = qemu_coroutine_self(),
+        };
+
+        acb = bs->drv->bdrv_aio_discard(bs, sector_num, nb_sectors,
+                                        bdrv_co_io_em_complete, &co);
+        if (acb == NULL) {
+            return -EIO;
+        } else {
+            qemu_coroutine_yield();
+            return co.ret;
+        }
+    } else {
+        return 0;
+    }
+}
+
+int bdrv_discard(BlockDriverState *bs, int64_t sector_num, int nb_sectors)
+{
+    Coroutine *co;
+    RwCo rwco = {
+        .bs = bs,
+        .sector_num = sector_num,
+        .nb_sectors = nb_sectors,
+        .ret = NOT_DONE,
+    };
+
+    if (qemu_in_coroutine()) {
+        /* Fast-path if already in coroutine context */
+        bdrv_discard_co_entry(&rwco);
+    } else {
+        co = qemu_coroutine_create(bdrv_discard_co_entry);
+        qemu_coroutine_enter(co, &rwco);
+        while (rwco.ret == NOT_DONE) {
+            qemu_aio_wait();
+        }
+    }
+
+    return rwco.ret;
+}
+
+/**************************************************************/
+/* removable device support */
+
+/**
+ * Return TRUE if the media is present
+ */
+int bdrv_is_inserted(BlockDriverState *bs)
+{
+    BlockDriver *drv = bs->drv;
+
+    if (!drv)
+        return 0;
+    if (!drv->bdrv_is_inserted)
+        return 1;
+    return drv->bdrv_is_inserted(bs);
+}
+
+/**
+ * Return whether the media changed since the last call to this
+ * function, or -ENOTSUP if we don't know.  Most drivers don't know.
+ */
+int bdrv_media_changed(BlockDriverState *bs)
+{
+    BlockDriver *drv = bs->drv;
+
+    if (drv && drv->bdrv_media_changed) {
+        return drv->bdrv_media_changed(bs);
+    }
+    return -ENOTSUP;
+}
+
+/**
+ * If eject_flag is TRUE, eject the media. Otherwise, close the tray
+ */
+void bdrv_eject(BlockDriverState *bs, bool eject_flag)
+{
+    BlockDriver *drv = bs->drv;
+
+    if (drv && drv->bdrv_eject) {
+        drv->bdrv_eject(bs, eject_flag);
+    }
+
+    if (bs->device_name[0] != '\0') {
+        bdrv_emit_qmp_eject_event(bs, eject_flag);
+    }
+}
+
+/**
+ * Lock or unlock the media (if it is locked, the user won't be able
+ * to eject it manually).
+ */
+void bdrv_lock_medium(BlockDriverState *bs, bool locked)
+{
+    BlockDriver *drv = bs->drv;
+
+    trace_bdrv_lock_medium(bs, locked);
+
+    if (drv && drv->bdrv_lock_medium) {
+        drv->bdrv_lock_medium(bs, locked);
+    }
+}
+
+/* needed for generic scsi interface */
+
+int bdrv_ioctl(BlockDriverState *bs, unsigned long int req, void *buf)
+{
+    BlockDriver *drv = bs->drv;
+
+    if (drv && drv->bdrv_ioctl)
+        return drv->bdrv_ioctl(bs, req, buf);
+    return -ENOTSUP;
+}
+
+BlockDriverAIOCB *bdrv_aio_ioctl(BlockDriverState *bs,
+        unsigned long int req, void *buf,
+        BlockDriverCompletionFunc *cb, void *opaque)
+{
+    BlockDriver *drv = bs->drv;
+
+    if (drv && drv->bdrv_aio_ioctl)
+        return drv->bdrv_aio_ioctl(bs, req, buf, cb, opaque);
+    return NULL;
+}
+
+void bdrv_set_buffer_alignment(BlockDriverState *bs, int align)
+{
+    bs->buffer_alignment = align;
+}
+
+void *qemu_blockalign(BlockDriverState *bs, size_t size)
+{
+    return qemu_memalign((bs && bs->buffer_alignment) ? bs->buffer_alignment : 512, size);
+}
+
+/*
+ * Check if all memory in this vector is sector aligned.
+ */
+bool bdrv_qiov_is_aligned(BlockDriverState *bs, QEMUIOVector *qiov)
+{
+    int i;
+
+    for (i = 0; i < qiov->niov; i++) {
+        if ((uintptr_t) qiov->iov[i].iov_base % bs->buffer_alignment) {
+            return false;
+        }
+    }
+
+    return true;
+}
+
+void bdrv_set_dirty_tracking(BlockDriverState *bs, int granularity)
+{
+    int64_t bitmap_size;
+
+    assert((granularity & (granularity - 1)) == 0);
+
+    if (granularity) {
+        granularity >>= BDRV_SECTOR_BITS;
+        assert(!bs->dirty_bitmap);
+        bitmap_size = (bdrv_getlength(bs) >> BDRV_SECTOR_BITS);
+        bs->dirty_bitmap = hbitmap_alloc(bitmap_size, ffs(granularity) - 1);
+    } else {
+        if (bs->dirty_bitmap) {
+            hbitmap_free(bs->dirty_bitmap);
+            bs->dirty_bitmap = NULL;
+        }
+    }
+}
+
+int bdrv_get_dirty(BlockDriverState *bs, int64_t sector)
+{
+    if (bs->dirty_bitmap) {
+        return hbitmap_get(bs->dirty_bitmap, sector);
+    } else {
+        return 0;
+    }
+}
+
+void bdrv_dirty_iter_init(BlockDriverState *bs, HBitmapIter *hbi)
+{
+    hbitmap_iter_init(hbi, bs->dirty_bitmap, 0);
+}
+
+void bdrv_set_dirty(BlockDriverState *bs, int64_t cur_sector,
+                    int nr_sectors)
+{
+    hbitmap_set(bs->dirty_bitmap, cur_sector, nr_sectors);
+}
+
+void bdrv_reset_dirty(BlockDriverState *bs, int64_t cur_sector,
+                      int nr_sectors)
+{
+    hbitmap_reset(bs->dirty_bitmap, cur_sector, nr_sectors);
+}
+
+int64_t bdrv_get_dirty_count(BlockDriverState *bs)
+{
+    if (bs->dirty_bitmap) {
+        return hbitmap_count(bs->dirty_bitmap);
+    } else {
+        return 0;
+    }
+}
+
+void bdrv_set_in_use(BlockDriverState *bs, int in_use)
+{
+    assert(bs->in_use != in_use);
+    bs->in_use = in_use;
+}
+
+int bdrv_in_use(BlockDriverState *bs)
+{
+    return bs->in_use;
+}
+
+void bdrv_iostatus_enable(BlockDriverState *bs)
+{
+    bs->iostatus_enabled = true;
+    bs->iostatus = BLOCK_DEVICE_IO_STATUS_OK;
+}
+
+/* The I/O status is only enabled if the drive explicitly
+ * enables it _and_ the VM is configured to stop on errors */
+bool bdrv_iostatus_is_enabled(const BlockDriverState *bs)
+{
+    return (bs->iostatus_enabled &&
+           (bs->on_write_error == BLOCKDEV_ON_ERROR_ENOSPC ||
+            bs->on_write_error == BLOCKDEV_ON_ERROR_STOP   ||
+            bs->on_read_error == BLOCKDEV_ON_ERROR_STOP));
+}
+
+void bdrv_iostatus_disable(BlockDriverState *bs)
+{
+    bs->iostatus_enabled = false;
+}
+
+void bdrv_iostatus_reset(BlockDriverState *bs)
+{
+    if (bdrv_iostatus_is_enabled(bs)) {
+        bs->iostatus = BLOCK_DEVICE_IO_STATUS_OK;
+        if (bs->job) {
+            block_job_iostatus_reset(bs->job);
+        }
+    }
+}
+
+void bdrv_iostatus_set_err(BlockDriverState *bs, int error)
+{
+    assert(bdrv_iostatus_is_enabled(bs));
+    if (bs->iostatus == BLOCK_DEVICE_IO_STATUS_OK) {
+        bs->iostatus = error == ENOSPC ? BLOCK_DEVICE_IO_STATUS_NOSPACE :
+                                         BLOCK_DEVICE_IO_STATUS_FAILED;
+    }
+}
+
+void
+bdrv_acct_start(BlockDriverState *bs, BlockAcctCookie *cookie, int64_t bytes,
+        enum BlockAcctType type)
+{
+    assert(type < BDRV_MAX_IOTYPE);
+
+    cookie->bytes = bytes;
+    cookie->start_time_ns = get_clock();
+    cookie->type = type;
+}
+
+void
+bdrv_acct_done(BlockDriverState *bs, BlockAcctCookie *cookie)
+{
+    assert(cookie->type < BDRV_MAX_IOTYPE);
+
+    bs->nr_bytes[cookie->type] += cookie->bytes;
+    bs->nr_ops[cookie->type]++;
+    bs->total_time_ns[cookie->type] += get_clock() - cookie->start_time_ns;
+}
+
+void bdrv_img_create(const char *filename, const char *fmt,
+                     const char *base_filename, const char *base_fmt,
+                     char *options, uint64_t img_size, int flags,
+                     Error **errp, bool quiet)
+{
+    QEMUOptionParameter *param = NULL, *create_options = NULL;
+    QEMUOptionParameter *backing_fmt, *backing_file, *size;
+    BlockDriverState *bs = NULL;
+    BlockDriver *drv, *proto_drv;
+    BlockDriver *backing_drv = NULL;
+    int ret = 0;
+
+    /* Find driver and parse its options */
+    drv = bdrv_find_format(fmt);
+    if (!drv) {
+        error_setg(errp, "Unknown file format '%s'", fmt);
+        return;
+    }
+
+    proto_drv = bdrv_find_protocol(filename, true);
+    if (!proto_drv) {
+        error_setg(errp, "Unknown protocol '%s'", filename);
+        return;
+    }
+
+    create_options = append_option_parameters(create_options,
+                                              drv->create_options);
+    create_options = append_option_parameters(create_options,
+                                              proto_drv->create_options);
+
+    /* Create parameter list with default values */
+    param = parse_option_parameters("", create_options, param);
+
+    set_option_parameter_int(param, BLOCK_OPT_SIZE, img_size);
+
+    /* Parse -o options */
+    if (options) {
+        param = parse_option_parameters(options, create_options, param);
+        if (param == NULL) {
+            error_setg(errp, "Invalid options for file format '%s'.", fmt);
+            goto out;
+        }
+    }
+
+    if (base_filename) {
+        if (set_option_parameter(param, BLOCK_OPT_BACKING_FILE,
+                                 base_filename)) {
+            error_setg(errp, "Backing file not supported for file format '%s'",
+                       fmt);
+            goto out;
+        }
+    }
+
+    if (base_fmt) {
+        if (set_option_parameter(param, BLOCK_OPT_BACKING_FMT, base_fmt)) {
+            error_setg(errp, "Backing file format not supported for file "
+                             "format '%s'", fmt);
+            goto out;
+        }
+    }
+
+    backing_file = get_option_parameter(param, BLOCK_OPT_BACKING_FILE);
+    if (backing_file && backing_file->value.s) {
+        if (!strcmp(filename, backing_file->value.s)) {
+            error_setg(errp, "Error: Trying to create an image with the "
+                             "same filename as the backing file");
+            goto out;
+        }
+    }
+
+    backing_fmt = get_option_parameter(param, BLOCK_OPT_BACKING_FMT);
+    if (backing_fmt && backing_fmt->value.s) {
+        backing_drv = bdrv_find_format(backing_fmt->value.s);
+        if (!backing_drv) {
+            error_setg(errp, "Unknown backing file format '%s'",
+                       backing_fmt->value.s);
+            goto out;
+        }
+    }
+
+    // The size for the image must always be specified, with one exception:
+    // If we are using a backing file, we can obtain the size from there
+    size = get_option_parameter(param, BLOCK_OPT_SIZE);
+    if (size && size->value.n == -1) {
+        if (backing_file && backing_file->value.s) {
+            uint64_t size;
+            char buf[32];
+            int back_flags;
+
+            /* backing files always opened read-only */
+            back_flags =
+                flags & ~(BDRV_O_RDWR | BDRV_O_SNAPSHOT | BDRV_O_NO_BACKING);
+
+            bs = bdrv_new("");
+
+            ret = bdrv_open(bs, backing_file->value.s, NULL, back_flags,
+                            backing_drv);
+            if (ret < 0) {
+                error_setg_errno(errp, -ret, "Could not open '%s'",
+                                 backing_file->value.s);
+                goto out;
+            }
+            bdrv_get_geometry(bs, &size);
+            size *= 512;
+
+            snprintf(buf, sizeof(buf), "%" PRId64, size);
+            set_option_parameter(param, BLOCK_OPT_SIZE, buf);
+        } else {
+            error_setg(errp, "Image creation needs a size parameter");
+            goto out;
+        }
+    }
+
+    if (!quiet) {
+        printf("Formatting '%s', fmt=%s ", filename, fmt);
+        print_option_parameters(param);
+        puts("");
+    }
+    ret = bdrv_create(drv, filename, param);
+    if (ret < 0) {
+        if (ret == -ENOTSUP) {
+            error_setg(errp,"Formatting or formatting option not supported for "
+                            "file format '%s'", fmt);
+        } else if (ret == -EFBIG) {
+            const char *cluster_size_hint = "";
+            if (get_option_parameter(create_options, BLOCK_OPT_CLUSTER_SIZE)) {
+                cluster_size_hint = " (try using a larger cluster size)";
+            }
+            error_setg(errp, "The image size is too large for file format '%s'%s",
+                       fmt, cluster_size_hint);
+        } else {
+            error_setg(errp, "%s: error while creating %s: %s", filename, fmt,
+                       strerror(-ret));
+        }
+    }
+
+out:
+    free_option_parameters(create_options);
+    free_option_parameters(param);
+
+    if (bs) {
+        bdrv_delete(bs);
+    }
+}
+
+AioContext *bdrv_get_aio_context(BlockDriverState *bs)
+{
+    /* Currently BlockDriverState always uses the main loop AioContext */
+    return qemu_get_aio_context();
+}
+
+void bdrv_add_before_write_notifier(BlockDriverState *bs,
+                                    NotifierWithReturn *notifier)
+{
+    notifier_with_return_list_add(&bs->before_write_notifiers, notifier);
+}
diff --git a/contrib/qemu/block/qcow.c b/contrib/qemu/block/qcow.c
new file mode 100644
index 000000000..5239bd68f
--- /dev/null
+++ b/contrib/qemu/block/qcow.c
@@ -0,0 +1,914 @@
+/*
+ * Block driver for the QCOW format
+ *
+ * Copyright (c) 2004-2006 Fabrice Bellard
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+#include "qemu-common.h"
+#include "block/block_int.h"
+#include "qemu/module.h"
+#include <zlib.h>
+#include "qemu/aes.h"
+#include "migration/migration.h"
+
+/**************************************************************/
+/* QEMU COW block driver with compression and encryption support */
+
+#define QCOW_MAGIC (('Q' << 24) | ('F' << 16) | ('I' << 8) | 0xfb)
+#define QCOW_VERSION 1
+
+#define QCOW_CRYPT_NONE 0
+#define QCOW_CRYPT_AES  1
+
+#define QCOW_OFLAG_COMPRESSED (1LL << 63)
+
+typedef struct QCowHeader {
+    uint32_t magic;
+    uint32_t version;
+    uint64_t backing_file_offset;
+    uint32_t backing_file_size;
+    uint32_t mtime;
+    uint64_t size; /* in bytes */
+    uint8_t cluster_bits;
+    uint8_t l2_bits;
+    uint32_t crypt_method;
+    uint64_t l1_table_offset;
+} QCowHeader;
+
+#define L2_CACHE_SIZE 16
+
+typedef struct BDRVQcowState {
+    int cluster_bits;
+    int cluster_size;
+    int cluster_sectors;
+    int l2_bits;
+    int l2_size;
+    int l1_size;
+    uint64_t cluster_offset_mask;
+    uint64_t l1_table_offset;
+    uint64_t *l1_table;
+    uint64_t *l2_cache;
+    uint64_t l2_cache_offsets[L2_CACHE_SIZE];
+    uint32_t l2_cache_counts[L2_CACHE_SIZE];
+    uint8_t *cluster_cache;
+    uint8_t *cluster_data;
+    uint64_t cluster_cache_offset;
+    uint32_t crypt_method; /* current crypt method, 0 if no key yet */
+    uint32_t crypt_method_header;
+    AES_KEY aes_encrypt_key;
+    AES_KEY aes_decrypt_key;
+    CoMutex lock;
+    Error *migration_blocker;
+} BDRVQcowState;
+
+static int decompress_cluster(BlockDriverState *bs, uint64_t cluster_offset);
+
+static int qcow_probe(const uint8_t *buf, int buf_size, const char *filename)
+{
+    const QCowHeader *cow_header = (const void *)buf;
+
+    if (buf_size >= sizeof(QCowHeader) &&
+        be32_to_cpu(cow_header->magic) == QCOW_MAGIC &&
+        be32_to_cpu(cow_header->version) == QCOW_VERSION)
+        return 100;
+    else
+        return 0;
+}
+
+static int qcow_open(BlockDriverState *bs, QDict *options, int flags)
+{
+    BDRVQcowState *s = bs->opaque;
+    int len, i, shift, ret;
+    QCowHeader header;
+
+    ret = bdrv_pread(bs->file, 0, &header, sizeof(header));
+    if (ret < 0) {
+        goto fail;
+    }
+    be32_to_cpus(&header.magic);
+    be32_to_cpus(&header.version);
+    be64_to_cpus(&header.backing_file_offset);
+    be32_to_cpus(&header.backing_file_size);
+    be32_to_cpus(&header.mtime);
+    be64_to_cpus(&header.size);
+    be32_to_cpus(&header.crypt_method);
+    be64_to_cpus(&header.l1_table_offset);
+
+    if (header.magic != QCOW_MAGIC) {
+        ret = -EMEDIUMTYPE;
+        goto fail;
+    }
+    if (header.version != QCOW_VERSION) {
+        char version[64];
+        snprintf(version, sizeof(version), "QCOW version %d", header.version);
+        qerror_report(QERR_UNKNOWN_BLOCK_FORMAT_FEATURE,
+            bs->device_name, "qcow", version);
+        ret = -ENOTSUP;
+        goto fail;
+    }
+
+    if (header.size <= 1 || header.cluster_bits < 9) {
+        ret = -EINVAL;
+        goto fail;
+    }
+    if (header.crypt_method > QCOW_CRYPT_AES) {
+        ret = -EINVAL;
+        goto fail;
+    }
+    s->crypt_method_header = header.crypt_method;
+    if (s->crypt_method_header) {
+        bs->encrypted = 1;
+    }
+    s->cluster_bits = header.cluster_bits;
+    s->cluster_size = 1 << s->cluster_bits;
+    s->cluster_sectors = 1 << (s->cluster_bits - 9);
+    s->l2_bits = header.l2_bits;
+    s->l2_size = 1 << s->l2_bits;
+    bs->total_sectors = header.size / 512;
+    s->cluster_offset_mask = (1LL << (63 - s->cluster_bits)) - 1;
+
+    /* read the level 1 table */
+    shift = s->cluster_bits + s->l2_bits;
+    s->l1_size = (header.size + (1LL << shift) - 1) >> shift;
+
+    s->l1_table_offset = header.l1_table_offset;
+    s->l1_table = g_malloc(s->l1_size * sizeof(uint64_t));
+
+    ret = bdrv_pread(bs->file, s->l1_table_offset, s->l1_table,
+               s->l1_size * sizeof(uint64_t));
+    if (ret < 0) {
+        goto fail;
+    }
+
+    for(i = 0;i < s->l1_size; i++) {
+        be64_to_cpus(&s->l1_table[i]);
+    }
+    /* alloc L2 cache */
+    s->l2_cache = g_malloc(s->l2_size * L2_CACHE_SIZE * sizeof(uint64_t));
+    s->cluster_cache = g_malloc(s->cluster_size);
+    s->cluster_data = g_malloc(s->cluster_size);
+    s->cluster_cache_offset = -1;
+
+    /* read the backing file name */
+    if (header.backing_file_offset != 0) {
+        len = header.backing_file_size;
+        if (len > 1023) {
+            len = 1023;
+        }
+        ret = bdrv_pread(bs->file, header.backing_file_offset,
+                   bs->backing_file, len);
+        if (ret < 0) {
+            goto fail;
+        }
+        bs->backing_file[len] = '\0';
+    }
+
+    /* Disable migration when qcow images are used */
+    error_set(&s->migration_blocker,
+              QERR_BLOCK_FORMAT_FEATURE_NOT_SUPPORTED,
+              "qcow", bs->device_name, "live migration");
+    migrate_add_blocker(s->migration_blocker);
+
+    qemu_co_mutex_init(&s->lock);
+    return 0;
+
+ fail:
+    g_free(s->l1_table);
+    g_free(s->l2_cache);
+    g_free(s->cluster_cache);
+    g_free(s->cluster_data);
+    return ret;
+}
+
+
+/* We have nothing to do for QCOW reopen, stubs just return
+ * success */
+static int qcow_reopen_prepare(BDRVReopenState *state,
+                               BlockReopenQueue *queue, Error **errp)
+{
+    return 0;
+}
+
+static int qcow_set_key(BlockDriverState *bs, const char *key)
+{
+    BDRVQcowState *s = bs->opaque;
+    uint8_t keybuf[16];
+    int len, i;
+
+    memset(keybuf, 0, 16);
+    len = strlen(key);
+    if (len > 16)
+        len = 16;
+    /* XXX: we could compress the chars to 7 bits to increase
+       entropy */
+    for(i = 0;i < len;i++) {
+        keybuf[i] = key[i];
+    }
+    s->crypt_method = s->crypt_method_header;
+
+    if (AES_set_encrypt_key(keybuf, 128, &s->aes_encrypt_key) != 0)
+        return -1;
+    if (AES_set_decrypt_key(keybuf, 128, &s->aes_decrypt_key) != 0)
+        return -1;
+    return 0;
+}
+
+/* The crypt function is compatible with the linux cryptoloop
+   algorithm for < 4 GB images. NOTE: out_buf == in_buf is
+   supported */
+static void encrypt_sectors(BDRVQcowState *s, int64_t sector_num,
+                            uint8_t *out_buf, const uint8_t *in_buf,
+                            int nb_sectors, int enc,
+                            const AES_KEY *key)
+{
+    union {
+        uint64_t ll[2];
+        uint8_t b[16];
+    } ivec;
+    int i;
+
+    for(i = 0; i < nb_sectors; i++) {
+        ivec.ll[0] = cpu_to_le64(sector_num);
+        ivec.ll[1] = 0;
+        AES_cbc_encrypt(in_buf, out_buf, 512, key,
+                        ivec.b, enc);
+        sector_num++;
+        in_buf += 512;
+        out_buf += 512;
+    }
+}
+
+/* 'allocate' is:
+ *
+ * 0 to not allocate.
+ *
+ * 1 to allocate a normal cluster (for sector indexes 'n_start' to
+ * 'n_end')
+ *
+ * 2 to allocate a compressed cluster of size
+ * 'compressed_size'. 'compressed_size' must be > 0 and <
+ * cluster_size
+ *
+ * return 0 if not allocated.
+ */
+static uint64_t get_cluster_offset(BlockDriverState *bs,
+                                   uint64_t offset, int allocate,
+                                   int compressed_size,
+                                   int n_start, int n_end)
+{
+    BDRVQcowState *s = bs->opaque;
+    int min_index, i, j, l1_index, l2_index;
+    uint64_t l2_offset, *l2_table, cluster_offset, tmp;
+    uint32_t min_count;
+    int new_l2_table;
+
+    l1_index = offset >> (s->l2_bits + s->cluster_bits);
+    l2_offset = s->l1_table[l1_index];
+    new_l2_table = 0;
+    if (!l2_offset) {
+        if (!allocate)
+            return 0;
+        /* allocate a new l2 entry */
+        l2_offset = bdrv_getlength(bs->file);
+        /* round to cluster size */
+        l2_offset = (l2_offset + s->cluster_size - 1) & ~(s->cluster_size - 1);
+        /* update the L1 entry */
+        s->l1_table[l1_index] = l2_offset;
+        tmp = cpu_to_be64(l2_offset);
+        if (bdrv_pwrite_sync(bs->file,
+                s->l1_table_offset + l1_index * sizeof(tmp),
+                &tmp, sizeof(tmp)) < 0)
+            return 0;
+        new_l2_table = 1;
+    }
+    for(i = 0; i < L2_CACHE_SIZE; i++) {
+        if (l2_offset == s->l2_cache_offsets[i]) {
+            /* increment the hit count */
+            if (++s->l2_cache_counts[i] == 0xffffffff) {
+                for(j = 0; j < L2_CACHE_SIZE; j++) {
+                    s->l2_cache_counts[j] >>= 1;
+                }
+            }
+            l2_table = s->l2_cache + (i << s->l2_bits);
+            goto found;
+        }
+    }
+    /* not found: load a new entry in the least used one */
+    min_index = 0;
+    min_count = 0xffffffff;
+    for(i = 0; i < L2_CACHE_SIZE; i++) {
+        if (s->l2_cache_counts[i] < min_count) {
+            min_count = s->l2_cache_counts[i];
+            min_index = i;
+        }
+    }
+    l2_table = s->l2_cache + (min_index << s->l2_bits);
+    if (new_l2_table) {
+        memset(l2_table, 0, s->l2_size * sizeof(uint64_t));
+        if (bdrv_pwrite_sync(bs->file, l2_offset, l2_table,
+                s->l2_size * sizeof(uint64_t)) < 0)
+            return 0;
+    } else {
+        if (bdrv_pread(bs->file, l2_offset, l2_table, s->l2_size * sizeof(uint64_t)) !=
+            s->l2_size * sizeof(uint64_t))
+            return 0;
+    }
+    s->l2_cache_offsets[min_index] = l2_offset;
+    s->l2_cache_counts[min_index] = 1;
+ found:
+    l2_index = (offset >> s->cluster_bits) & (s->l2_size - 1);
+    cluster_offset = be64_to_cpu(l2_table[l2_index]);
+    if (!cluster_offset ||
+        ((cluster_offset & QCOW_OFLAG_COMPRESSED) && allocate == 1)) {
+        if (!allocate)
+            return 0;
+        /* allocate a new cluster */
+        if ((cluster_offset & QCOW_OFLAG_COMPRESSED) &&
+            (n_end - n_start) < s->cluster_sectors) {
+            /* if the cluster is already compressed, we must
+               decompress it in the case it is not completely
+               overwritten */
+            if (decompress_cluster(bs, cluster_offset) < 0)
+                return 0;
+            cluster_offset = bdrv_getlength(bs->file);
+            cluster_offset = (cluster_offset + s->cluster_size - 1) &
+                ~(s->cluster_size - 1);
+            /* write the cluster content */
+            if (bdrv_pwrite(bs->file, cluster_offset, s->cluster_cache, s->cluster_size) !=
+                s->cluster_size)
+                return -1;
+        } else {
+            cluster_offset = bdrv_getlength(bs->file);
+            if (allocate == 1) {
+                /* round to cluster size */
+                cluster_offset = (cluster_offset + s->cluster_size - 1) &
+                    ~(s->cluster_size - 1);
+                bdrv_truncate(bs->file, cluster_offset + s->cluster_size);
+                /* if encrypted, we must initialize the cluster
+                   content which won't be written */
+                if (s->crypt_method &&
+                    (n_end - n_start) < s->cluster_sectors) {
+                    uint64_t start_sect;
+                    start_sect = (offset & ~(s->cluster_size - 1)) >> 9;
+                    memset(s->cluster_data + 512, 0x00, 512);
+                    for(i = 0; i < s->cluster_sectors; i++) {
+                        if (i < n_start || i >= n_end) {
+                            encrypt_sectors(s, start_sect + i,
+                                            s->cluster_data,
+                                            s->cluster_data + 512, 1, 1,
+                                            &s->aes_encrypt_key);
+                            if (bdrv_pwrite(bs->file, cluster_offset + i * 512,
+                                            s->cluster_data, 512) != 512)
+                                return -1;
+                        }
+                    }
+                }
+            } else if (allocate == 2) {
+                cluster_offset |= QCOW_OFLAG_COMPRESSED |
+                    (uint64_t)compressed_size << (63 - s->cluster_bits);
+            }
+        }
+        /* update L2 table */
+        tmp = cpu_to_be64(cluster_offset);
+        l2_table[l2_index] = tmp;
+        if (bdrv_pwrite_sync(bs->file, l2_offset + l2_index * sizeof(tmp),
+                &tmp, sizeof(tmp)) < 0)
+            return 0;
+    }
+    return cluster_offset;
+}
+
+static int coroutine_fn qcow_co_is_allocated(BlockDriverState *bs,
+        int64_t sector_num, int nb_sectors, int *pnum)
+{
+    BDRVQcowState *s = bs->opaque;
+    int index_in_cluster, n;
+    uint64_t cluster_offset;
+
+    qemu_co_mutex_lock(&s->lock);
+    cluster_offset = get_cluster_offset(bs, sector_num << 9, 0, 0, 0, 0);
+    qemu_co_mutex_unlock(&s->lock);
+    index_in_cluster = sector_num & (s->cluster_sectors - 1);
+    n = s->cluster_sectors - index_in_cluster;
+    if (n > nb_sectors)
+        n = nb_sectors;
+    *pnum = n;
+    return (cluster_offset != 0);
+}
+
+static int decompress_buffer(uint8_t *out_buf, int out_buf_size,
+                             const uint8_t *buf, int buf_size)
+{
+    z_stream strm1, *strm = &strm1;
+    int ret, out_len;
+
+    memset(strm, 0, sizeof(*strm));
+
+    strm->next_in = (uint8_t *)buf;
+    strm->avail_in = buf_size;
+    strm->next_out = out_buf;
+    strm->avail_out = out_buf_size;
+
+    ret = inflateInit2(strm, -12);
+    if (ret != Z_OK)
+        return -1;
+    ret = inflate(strm, Z_FINISH);
+    out_len = strm->next_out - out_buf;
+    if ((ret != Z_STREAM_END && ret != Z_BUF_ERROR) ||
+        out_len != out_buf_size) {
+        inflateEnd(strm);
+        return -1;
+    }
+    inflateEnd(strm);
+    return 0;
+}
+
+static int decompress_cluster(BlockDriverState *bs, uint64_t cluster_offset)
+{
+    BDRVQcowState *s = bs->opaque;
+    int ret, csize;
+    uint64_t coffset;
+
+    coffset = cluster_offset & s->cluster_offset_mask;
+    if (s->cluster_cache_offset != coffset) {
+        csize = cluster_offset >> (63 - s->cluster_bits);
+        csize &= (s->cluster_size - 1);
+        ret = bdrv_pread(bs->file, coffset, s->cluster_data, csize);
+        if (ret != csize)
+            return -1;
+        if (decompress_buffer(s->cluster_cache, s->cluster_size,
+                              s->cluster_data, csize) < 0) {
+            return -1;
+        }
+        s->cluster_cache_offset = coffset;
+    }
+    return 0;
+}
+
+static coroutine_fn int qcow_co_readv(BlockDriverState *bs, int64_t sector_num,
+                         int nb_sectors, QEMUIOVector *qiov)
+{
+    BDRVQcowState *s = bs->opaque;
+    int index_in_cluster;
+    int ret = 0, n;
+    uint64_t cluster_offset;
+    struct iovec hd_iov;
+    QEMUIOVector hd_qiov;
+    uint8_t *buf;
+    void *orig_buf;
+
+    if (qiov->niov > 1) {
+        buf = orig_buf = qemu_blockalign(bs, qiov->size);
+    } else {
+        orig_buf = NULL;
+        buf = (uint8_t *)qiov->iov->iov_base;
+    }
+
+    qemu_co_mutex_lock(&s->lock);
+
+    while (nb_sectors != 0) {
+        /* prepare next request */
+        cluster_offset = get_cluster_offset(bs, sector_num << 9,
+                                                 0, 0, 0, 0);
+        index_in_cluster = sector_num & (s->cluster_sectors - 1);
+        n = s->cluster_sectors - index_in_cluster;
+        if (n > nb_sectors) {
+            n = nb_sectors;
+        }
+
+        if (!cluster_offset) {
+            if (bs->backing_hd) {
+                /* read from the base image */
+                hd_iov.iov_base = (void *)buf;
+                hd_iov.iov_len = n * 512;
+                qemu_iovec_init_external(&hd_qiov, &hd_iov, 1);
+                qemu_co_mutex_unlock(&s->lock);
+                ret = bdrv_co_readv(bs->backing_hd, sector_num,
+                                    n, &hd_qiov);
+                qemu_co_mutex_lock(&s->lock);
+                if (ret < 0) {
+                    goto fail;
+                }
+            } else {
+                /* Note: in this case, no need to wait */
+                memset(buf, 0, 512 * n);
+            }
+        } else if (cluster_offset & QCOW_OFLAG_COMPRESSED) {
+            /* add AIO support for compressed blocks ? */
+            if (decompress_cluster(bs, cluster_offset) < 0) {
+                goto fail;
+            }
+            memcpy(buf,
+                   s->cluster_cache + index_in_cluster * 512, 512 * n);
+        } else {
+            if ((cluster_offset & 511) != 0) {
+                goto fail;
+            }
+            hd_iov.iov_base = (void *)buf;
+            hd_iov.iov_len = n * 512;
+            qemu_iovec_init_external(&hd_qiov, &hd_iov, 1);
+            qemu_co_mutex_unlock(&s->lock);
+            ret = bdrv_co_readv(bs->file,
+                                (cluster_offset >> 9) + index_in_cluster,
+                                n, &hd_qiov);
+            qemu_co_mutex_lock(&s->lock);
+            if (ret < 0) {
+                break;
+            }
+            if (s->crypt_method) {
+                encrypt_sectors(s, sector_num, buf, buf,
+                                n, 0,
+                                &s->aes_decrypt_key);
+            }
+        }
+        ret = 0;
+
+        nb_sectors -= n;
+        sector_num += n;
+        buf += n * 512;
+    }
+
+done:
+    qemu_co_mutex_unlock(&s->lock);
+
+    if (qiov->niov > 1) {
+        qemu_iovec_from_buf(qiov, 0, orig_buf, qiov->size);
+        qemu_vfree(orig_buf);
+    }
+
+    return ret;
+
+fail:
+    ret = -EIO;
+    goto done;
+}
+
+static coroutine_fn int qcow_co_writev(BlockDriverState *bs, int64_t sector_num,
+                          int nb_sectors, QEMUIOVector *qiov)
+{
+    BDRVQcowState *s = bs->opaque;
+    int index_in_cluster;
+    uint64_t cluster_offset;
+    const uint8_t *src_buf;
+    int ret = 0, n;
+    uint8_t *cluster_data = NULL;
+    struct iovec hd_iov;
+    QEMUIOVector hd_qiov;
+    uint8_t *buf;
+    void *orig_buf;
+
+    s->cluster_cache_offset = -1; /* disable compressed cache */
+
+    if (qiov->niov > 1) {
+        buf = orig_buf = qemu_blockalign(bs, qiov->size);
+        qemu_iovec_to_buf(qiov, 0, buf, qiov->size);
+    } else {
+        orig_buf = NULL;
+        buf = (uint8_t *)qiov->iov->iov_base;
+    }
+
+    qemu_co_mutex_lock(&s->lock);
+
+    while (nb_sectors != 0) {
+
+        index_in_cluster = sector_num & (s->cluster_sectors - 1);
+        n = s->cluster_sectors - index_in_cluster;
+        if (n > nb_sectors) {
+            n = nb_sectors;
+        }
+        cluster_offset = get_cluster_offset(bs, sector_num << 9, 1, 0,
+                                            index_in_cluster,
+                                            index_in_cluster + n);
+        if (!cluster_offset || (cluster_offset & 511) != 0) {
+            ret = -EIO;
+            break;
+        }
+        if (s->crypt_method) {
+            if (!cluster_data) {
+                cluster_data = g_malloc0(s->cluster_size);
+            }
+            encrypt_sectors(s, sector_num, cluster_data, buf,
+                            n, 1, &s->aes_encrypt_key);
+            src_buf = cluster_data;
+        } else {
+            src_buf = buf;
+        }
+
+        hd_iov.iov_base = (void *)src_buf;
+        hd_iov.iov_len = n * 512;
+        qemu_iovec_init_external(&hd_qiov, &hd_iov, 1);
+        qemu_co_mutex_unlock(&s->lock);
+        ret = bdrv_co_writev(bs->file,
+                             (cluster_offset >> 9) + index_in_cluster,
+                             n, &hd_qiov);
+        qemu_co_mutex_lock(&s->lock);
+        if (ret < 0) {
+            break;
+        }
+        ret = 0;
+
+        nb_sectors -= n;
+        sector_num += n;
+        buf += n * 512;
+    }
+    qemu_co_mutex_unlock(&s->lock);
+
+    if (qiov->niov > 1) {
+        qemu_vfree(orig_buf);
+    }
+    g_free(cluster_data);
+
+    return ret;
+}
+
+static void qcow_close(BlockDriverState *bs)
+{
+    BDRVQcowState *s = bs->opaque;
+
+    g_free(s->l1_table);
+    g_free(s->l2_cache);
+    g_free(s->cluster_cache);
+    g_free(s->cluster_data);
+
+    migrate_del_blocker(s->migration_blocker);
+    error_free(s->migration_blocker);
+}
+
+static int qcow_create(const char *filename, QEMUOptionParameter *options)
+{
+    int header_size, backing_filename_len, l1_size, shift, i;
+    QCowHeader header;
+    uint8_t *tmp;
+    int64_t total_size = 0;
+    const char *backing_file = NULL;
+    int flags = 0;
+    int ret;
+    BlockDriverState *qcow_bs;
+
+    /* Read out options */
+    while (options && options->name) {
+        if (!strcmp(options->name, BLOCK_OPT_SIZE)) {
+            total_size = options->value.n / 512;
+        } else if (!strcmp(options->name, BLOCK_OPT_BACKING_FILE)) {
+            backing_file = options->value.s;
+        } else if (!strcmp(options->name, BLOCK_OPT_ENCRYPT)) {
+            flags |= options->value.n ? BLOCK_FLAG_ENCRYPT : 0;
+        }
+        options++;
+    }
+
+    ret = bdrv_create_file(filename, options);
+    if (ret < 0) {
+        return ret;
+    }
+
+    ret = bdrv_file_open(&qcow_bs, filename, NULL, BDRV_O_RDWR);
+    if (ret < 0) {
+        return ret;
+    }
+
+    ret = bdrv_truncate(qcow_bs, 0);
+    if (ret < 0) {
+        goto exit;
+    }
+
+    memset(&header, 0, sizeof(header));
+    header.magic = cpu_to_be32(QCOW_MAGIC);
+    header.version = cpu_to_be32(QCOW_VERSION);
+    header.size = cpu_to_be64(total_size * 512);
+    header_size = sizeof(header);
+    backing_filename_len = 0;
+    if (backing_file) {
+        if (strcmp(backing_file, "fat:")) {
+            header.backing_file_offset = cpu_to_be64(header_size);
+            backing_filename_len = strlen(backing_file);
+            header.backing_file_size = cpu_to_be32(backing_filename_len);
+            header_size += backing_filename_len;
+        } else {
+            /* special backing file for vvfat */
+            backing_file = NULL;
+        }
+        header.cluster_bits = 9; /* 512 byte cluster to avoid copying
+                                    unmodifyed sectors */
+        header.l2_bits = 12; /* 32 KB L2 tables */
+    } else {
+        header.cluster_bits = 12; /* 4 KB clusters */
+        header.l2_bits = 9; /* 4 KB L2 tables */
+    }
+    header_size = (header_size + 7) & ~7;
+    shift = header.cluster_bits + header.l2_bits;
+    l1_size = ((total_size * 512) + (1LL << shift) - 1) >> shift;
+
+    header.l1_table_offset = cpu_to_be64(header_size);
+    if (flags & BLOCK_FLAG_ENCRYPT) {
+        header.crypt_method = cpu_to_be32(QCOW_CRYPT_AES);
+    } else {
+        header.crypt_method = cpu_to_be32(QCOW_CRYPT_NONE);
+    }
+
+    /* write all the data */
+    ret = bdrv_pwrite(qcow_bs, 0, &header, sizeof(header));
+    if (ret != sizeof(header)) {
+        goto exit;
+    }
+
+    if (backing_file) {
+        ret = bdrv_pwrite(qcow_bs, sizeof(header),
+            backing_file, backing_filename_len);
+        if (ret != backing_filename_len) {
+            goto exit;
+        }
+    }
+
+    tmp = g_malloc0(BDRV_SECTOR_SIZE);
+    for (i = 0; i < ((sizeof(uint64_t)*l1_size + BDRV_SECTOR_SIZE - 1)/
+        BDRV_SECTOR_SIZE); i++) {
+        ret = bdrv_pwrite(qcow_bs, header_size +
+            BDRV_SECTOR_SIZE*i, tmp, BDRV_SECTOR_SIZE);
+        if (ret != BDRV_SECTOR_SIZE) {
+            g_free(tmp);
+            goto exit;
+        }
+    }
+
+    g_free(tmp);
+    ret = 0;
+exit:
+    bdrv_delete(qcow_bs);
+    return ret;
+}
+
+static int qcow_make_empty(BlockDriverState *bs)
+{
+    BDRVQcowState *s = bs->opaque;
+    uint32_t l1_length = s->l1_size * sizeof(uint64_t);
+    int ret;
+
+    memset(s->l1_table, 0, l1_length);
+    if (bdrv_pwrite_sync(bs->file, s->l1_table_offset, s->l1_table,
+            l1_length) < 0)
+        return -1;
+    ret = bdrv_truncate(bs->file, s->l1_table_offset + l1_length);
+    if (ret < 0)
+        return ret;
+
+    memset(s->l2_cache, 0, s->l2_size * L2_CACHE_SIZE * sizeof(uint64_t));
+    memset(s->l2_cache_offsets, 0, L2_CACHE_SIZE * sizeof(uint64_t));
+    memset(s->l2_cache_counts, 0, L2_CACHE_SIZE * sizeof(uint32_t));
+
+    return 0;
+}
+
+/* XXX: put compressed sectors first, then all the cluster aligned
+   tables to avoid losing bytes in alignment */
+static int qcow_write_compressed(BlockDriverState *bs, int64_t sector_num,
+                                 const uint8_t *buf, int nb_sectors)
+{
+    BDRVQcowState *s = bs->opaque;
+    z_stream strm;
+    int ret, out_len;
+    uint8_t *out_buf;
+    uint64_t cluster_offset;
+
+    if (nb_sectors != s->cluster_sectors) {
+        ret = -EINVAL;
+
+        /* Zero-pad last write if image size is not cluster aligned */
+        if (sector_num + nb_sectors == bs->total_sectors &&
+            nb_sectors < s->cluster_sectors) {
+            uint8_t *pad_buf = qemu_blockalign(bs, s->cluster_size);
+            memset(pad_buf, 0, s->cluster_size);
+            memcpy(pad_buf, buf, nb_sectors * BDRV_SECTOR_SIZE);
+            ret = qcow_write_compressed(bs, sector_num,
+                                        pad_buf, s->cluster_sectors);
+            qemu_vfree(pad_buf);
+        }
+        return ret;
+    }
+
+    out_buf = g_malloc(s->cluster_size + (s->cluster_size / 1000) + 128);
+
+    /* best compression, small window, no zlib header */
+    memset(&strm, 0, sizeof(strm));
+    ret = deflateInit2(&strm, Z_DEFAULT_COMPRESSION,
+                       Z_DEFLATED, -12,
+                       9, Z_DEFAULT_STRATEGY);
+    if (ret != 0) {
+        ret = -EINVAL;
+        goto fail;
+    }
+
+    strm.avail_in = s->cluster_size;
+    strm.next_in = (uint8_t *)buf;
+    strm.avail_out = s->cluster_size;
+    strm.next_out = out_buf;
+
+    ret = deflate(&strm, Z_FINISH);
+    if (ret != Z_STREAM_END && ret != Z_OK) {
+        deflateEnd(&strm);
+        ret = -EINVAL;
+        goto fail;
+    }
+    out_len = strm.next_out - out_buf;
+
+    deflateEnd(&strm);
+
+    if (ret != Z_STREAM_END || out_len >= s->cluster_size) {
+        /* could not compress: write normal cluster */
+        ret = bdrv_write(bs, sector_num, buf, s->cluster_sectors);
+        if (ret < 0) {
+            goto fail;
+        }
+    } else {
+        cluster_offset = get_cluster_offset(bs, sector_num << 9, 2,
+                                            out_len, 0, 0);
+        if (cluster_offset == 0) {
+            ret = -EIO;
+            goto fail;
+        }
+
+        cluster_offset &= s->cluster_offset_mask;
+        ret = bdrv_pwrite(bs->file, cluster_offset, out_buf, out_len);
+        if (ret < 0) {
+            goto fail;
+        }
+    }
+
+    ret = 0;
+fail:
+    g_free(out_buf);
+    return ret;
+}
+
+static int qcow_get_info(BlockDriverState *bs, BlockDriverInfo *bdi)
+{
+    BDRVQcowState *s = bs->opaque;
+    bdi->cluster_size = s->cluster_size;
+    return 0;
+}
+
+
+static QEMUOptionParameter qcow_create_options[] = {
+    {
+        .name = BLOCK_OPT_SIZE,
+        .type = OPT_SIZE,
+        .help = "Virtual disk size"
+    },
+    {
+        .name = BLOCK_OPT_BACKING_FILE,
+        .type = OPT_STRING,
+        .help = "File name of a base image"
+    },
+    {
+        .name = BLOCK_OPT_ENCRYPT,
+        .type = OPT_FLAG,
+        .help = "Encrypt the image"
+    },
+    { NULL }
+};
+
+static BlockDriver bdrv_qcow = {
+    .format_name	= "qcow",
+    .instance_size	= sizeof(BDRVQcowState),
+    .bdrv_probe		= qcow_probe,
+    .bdrv_open		= qcow_open,
+    .bdrv_close		= qcow_close,
+    .bdrv_reopen_prepare = qcow_reopen_prepare,
+    .bdrv_create	= qcow_create,
+    .bdrv_has_zero_init     = bdrv_has_zero_init_1,
+
+    .bdrv_co_readv          = qcow_co_readv,
+    .bdrv_co_writev         = qcow_co_writev,
+    .bdrv_co_is_allocated   = qcow_co_is_allocated,
+
+    .bdrv_set_key           = qcow_set_key,
+    .bdrv_make_empty        = qcow_make_empty,
+    .bdrv_write_compressed  = qcow_write_compressed,
+    .bdrv_get_info          = qcow_get_info,
+
+    .create_options = qcow_create_options,
+};
+
+static void bdrv_qcow_init(void)
+{
+    bdrv_register(&bdrv_qcow);
+}
+
+block_init(bdrv_qcow_init);
diff --git a/contrib/qemu/block/qcow2-cache.c b/contrib/qemu/block/qcow2-cache.c
new file mode 100644
index 000000000..2f3114ecc
--- /dev/null
+++ b/contrib/qemu/block/qcow2-cache.c
@@ -0,0 +1,323 @@
+/*
+ * L2/refcount table cache for the QCOW2 format
+ *
+ * Copyright (c) 2010 Kevin Wolf <kwolf@redhat.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#include "block/block_int.h"
+#include "qemu-common.h"
+#include "qcow2.h"
+#include "trace.h"
+
+typedef struct Qcow2CachedTable {
+    void*   table;
+    int64_t offset;
+    bool    dirty;
+    int     cache_hits;
+    int     ref;
+} Qcow2CachedTable;
+
+struct Qcow2Cache {
+    Qcow2CachedTable*       entries;
+    struct Qcow2Cache*      depends;
+    int                     size;
+    bool                    depends_on_flush;
+};
+
+Qcow2Cache *qcow2_cache_create(BlockDriverState *bs, int num_tables)
+{
+    BDRVQcowState *s = bs->opaque;
+    Qcow2Cache *c;
+    int i;
+
+    c = g_malloc0(sizeof(*c));
+    c->size = num_tables;
+    c->entries = g_malloc0(sizeof(*c->entries) * num_tables);
+
+    for (i = 0; i < c->size; i++) {
+        c->entries[i].table = qemu_blockalign(bs, s->cluster_size);
+    }
+
+    return c;
+}
+
+int qcow2_cache_destroy(BlockDriverState* bs, Qcow2Cache *c)
+{
+    int i;
+
+    for (i = 0; i < c->size; i++) {
+        assert(c->entries[i].ref == 0);
+        qemu_vfree(c->entries[i].table);
+    }
+
+    g_free(c->entries);
+    g_free(c);
+
+    return 0;
+}
+
+static int qcow2_cache_flush_dependency(BlockDriverState *bs, Qcow2Cache *c)
+{
+    int ret;
+
+    ret = qcow2_cache_flush(bs, c->depends);
+    if (ret < 0) {
+        return ret;
+    }
+
+    c->depends = NULL;
+    c->depends_on_flush = false;
+
+    return 0;
+}
+
+static int qcow2_cache_entry_flush(BlockDriverState *bs, Qcow2Cache *c, int i)
+{
+    BDRVQcowState *s = bs->opaque;
+    int ret = 0;
+
+    if (!c->entries[i].dirty || !c->entries[i].offset) {
+        return 0;
+    }
+
+    trace_qcow2_cache_entry_flush(qemu_coroutine_self(),
+                                  c == s->l2_table_cache, i);
+
+    if (c->depends) {
+        ret = qcow2_cache_flush_dependency(bs, c);
+    } else if (c->depends_on_flush) {
+        ret = bdrv_flush(bs->file);
+        if (ret >= 0) {
+            c->depends_on_flush = false;
+        }
+    }
+
+    if (ret < 0) {
+        return ret;
+    }
+
+    if (c == s->refcount_block_cache) {
+        BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_UPDATE_PART);
+    } else if (c == s->l2_table_cache) {
+        BLKDBG_EVENT(bs->file, BLKDBG_L2_UPDATE);
+    }
+
+    ret = bdrv_pwrite(bs->file, c->entries[i].offset, c->entries[i].table,
+        s->cluster_size);
+    if (ret < 0) {
+        return ret;
+    }
+
+    c->entries[i].dirty = false;
+
+    return 0;
+}
+
+int qcow2_cache_flush(BlockDriverState *bs, Qcow2Cache *c)
+{
+    BDRVQcowState *s = bs->opaque;
+    int result = 0;
+    int ret;
+    int i;
+
+    trace_qcow2_cache_flush(qemu_coroutine_self(), c == s->l2_table_cache);
+
+    for (i = 0; i < c->size; i++) {
+        ret = qcow2_cache_entry_flush(bs, c, i);
+        if (ret < 0 && result != -ENOSPC) {
+            result = ret;
+        }
+    }
+
+    if (result == 0) {
+        ret = bdrv_flush(bs->file);
+        if (ret < 0) {
+            result = ret;
+        }
+    }
+
+    return result;
+}
+
+int qcow2_cache_set_dependency(BlockDriverState *bs, Qcow2Cache *c,
+    Qcow2Cache *dependency)
+{
+    int ret;
+
+    if (dependency->depends) {
+        ret = qcow2_cache_flush_dependency(bs, dependency);
+        if (ret < 0) {
+            return ret;
+        }
+    }
+
+    if (c->depends && (c->depends != dependency)) {
+        ret = qcow2_cache_flush_dependency(bs, c);
+        if (ret < 0) {
+            return ret;
+        }
+    }
+
+    c->depends = dependency;
+    return 0;
+}
+
+void qcow2_cache_depends_on_flush(Qcow2Cache *c)
+{
+    c->depends_on_flush = true;
+}
+
+static int qcow2_cache_find_entry_to_replace(Qcow2Cache *c)
+{
+    int i;
+    int min_count = INT_MAX;
+    int min_index = -1;
+
+
+    for (i = 0; i < c->size; i++) {
+        if (c->entries[i].ref) {
+            continue;
+        }
+
+        if (c->entries[i].cache_hits < min_count) {
+            min_index = i;
+            min_count = c->entries[i].cache_hits;
+        }
+
+        /* Give newer hits priority */
+        /* TODO Check how to optimize the replacement strategy */
+        c->entries[i].cache_hits /= 2;
+    }
+
+    if (min_index == -1) {
+        /* This can't happen in current synchronous code, but leave the check
+         * here as a reminder for whoever starts using AIO with the cache */
+        abort();
+    }
+    return min_index;
+}
+
+static int qcow2_cache_do_get(BlockDriverState *bs, Qcow2Cache *c,
+    uint64_t offset, void **table, bool read_from_disk)
+{
+    BDRVQcowState *s = bs->opaque;
+    int i;
+    int ret;
+
+    trace_qcow2_cache_get(qemu_coroutine_self(), c == s->l2_table_cache,
+                          offset, read_from_disk);
+
+    /* Check if the table is already cached */
+    for (i = 0; i < c->size; i++) {
+        if (c->entries[i].offset == offset) {
+            goto found;
+        }
+    }
+
+    /* If not, write a table back and replace it */
+    i = qcow2_cache_find_entry_to_replace(c);
+    trace_qcow2_cache_get_replace_entry(qemu_coroutine_self(),
+                                        c == s->l2_table_cache, i);
+    if (i < 0) {
+        return i;
+    }
+
+    ret = qcow2_cache_entry_flush(bs, c, i);
+    if (ret < 0) {
+        return ret;
+    }
+
+    trace_qcow2_cache_get_read(qemu_coroutine_self(),
+                               c == s->l2_table_cache, i);
+    c->entries[i].offset = 0;
+    if (read_from_disk) {
+        if (c == s->l2_table_cache) {
+            BLKDBG_EVENT(bs->file, BLKDBG_L2_LOAD);
+        }
+
+        ret = bdrv_pread(bs->file, offset, c->entries[i].table, s->cluster_size);
+        if (ret < 0) {
+            return ret;
+        }
+    }
+
+    /* Give the table some hits for the start so that it won't be replaced
+     * immediately. The number 32 is completely arbitrary. */
+    c->entries[i].cache_hits = 32;
+    c->entries[i].offset = offset;
+
+    /* And return the right table */
+found:
+    c->entries[i].cache_hits++;
+    c->entries[i].ref++;
+    *table = c->entries[i].table;
+
+    trace_qcow2_cache_get_done(qemu_coroutine_self(),
+                               c == s->l2_table_cache, i);
+
+    return 0;
+}
+
+int qcow2_cache_get(BlockDriverState *bs, Qcow2Cache *c, uint64_t offset,
+    void **table)
+{
+    return qcow2_cache_do_get(bs, c, offset, table, true);
+}
+
+int qcow2_cache_get_empty(BlockDriverState *bs, Qcow2Cache *c, uint64_t offset,
+    void **table)
+{
+    return qcow2_cache_do_get(bs, c, offset, table, false);
+}
+
+int qcow2_cache_put(BlockDriverState *bs, Qcow2Cache *c, void **table)
+{
+    int i;
+
+    for (i = 0; i < c->size; i++) {
+        if (c->entries[i].table == *table) {
+            goto found;
+        }
+    }
+    return -ENOENT;
+
+found:
+    c->entries[i].ref--;
+    *table = NULL;
+
+    assert(c->entries[i].ref >= 0);
+    return 0;
+}
+
+void qcow2_cache_entry_mark_dirty(Qcow2Cache *c, void *table)
+{
+    int i;
+
+    for (i = 0; i < c->size; i++) {
+        if (c->entries[i].table == table) {
+            goto found;
+        }
+    }
+    abort();
+
+found:
+    c->entries[i].dirty = true;
+}
diff --git a/contrib/qemu/block/qcow2-cluster.c b/contrib/qemu/block/qcow2-cluster.c
new file mode 100644
index 000000000..cca76d4fc
--- /dev/null
+++ b/contrib/qemu/block/qcow2-cluster.c
@@ -0,0 +1,1478 @@
+/*
+ * Block driver for the QCOW version 2 format
+ *
+ * Copyright (c) 2004-2006 Fabrice Bellard
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#include <zlib.h>
+
+#include "qemu-common.h"
+#include "block/block_int.h"
+#include "block/qcow2.h"
+#include "trace.h"
+
+int qcow2_grow_l1_table(BlockDriverState *bs, uint64_t min_size,
+                        bool exact_size)
+{
+    BDRVQcowState *s = bs->opaque;
+    int new_l1_size2, ret, i;
+    uint64_t *new_l1_table;
+    int64_t new_l1_table_offset, new_l1_size;
+    uint8_t data[12];
+
+    if (min_size <= s->l1_size)
+        return 0;
+
+    if (exact_size) {
+        new_l1_size = min_size;
+    } else {
+        /* Bump size up to reduce the number of times we have to grow */
+        new_l1_size = s->l1_size;
+        if (new_l1_size == 0) {
+            new_l1_size = 1;
+        }
+        while (min_size > new_l1_size) {
+            new_l1_size = (new_l1_size * 3 + 1) / 2;
+        }
+    }
+
+    if (new_l1_size > INT_MAX) {
+        return -EFBIG;
+    }
+
+#ifdef DEBUG_ALLOC2
+    fprintf(stderr, "grow l1_table from %d to %" PRId64 "\n",
+            s->l1_size, new_l1_size);
+#endif
+
+    new_l1_size2 = sizeof(uint64_t) * new_l1_size;
+    new_l1_table = g_malloc0(align_offset(new_l1_size2, 512));
+    memcpy(new_l1_table, s->l1_table, s->l1_size * sizeof(uint64_t));
+
+    /* write new table (align to cluster) */
+    BLKDBG_EVENT(bs->file, BLKDBG_L1_GROW_ALLOC_TABLE);
+    new_l1_table_offset = qcow2_alloc_clusters(bs, new_l1_size2);
+    if (new_l1_table_offset < 0) {
+        g_free(new_l1_table);
+        return new_l1_table_offset;
+    }
+
+    ret = qcow2_cache_flush(bs, s->refcount_block_cache);
+    if (ret < 0) {
+        goto fail;
+    }
+
+    BLKDBG_EVENT(bs->file, BLKDBG_L1_GROW_WRITE_TABLE);
+    for(i = 0; i < s->l1_size; i++)
+        new_l1_table[i] = cpu_to_be64(new_l1_table[i]);
+    ret = bdrv_pwrite_sync(bs->file, new_l1_table_offset, new_l1_table, new_l1_size2);
+    if (ret < 0)
+        goto fail;
+    for(i = 0; i < s->l1_size; i++)
+        new_l1_table[i] = be64_to_cpu(new_l1_table[i]);
+
+    /* set new table */
+    BLKDBG_EVENT(bs->file, BLKDBG_L1_GROW_ACTIVATE_TABLE);
+    cpu_to_be32w((uint32_t*)data, new_l1_size);
+    cpu_to_be64wu((uint64_t*)(data + 4), new_l1_table_offset);
+    ret = bdrv_pwrite_sync(bs->file, offsetof(QCowHeader, l1_size), data,sizeof(data));
+    if (ret < 0) {
+        goto fail;
+    }
+    g_free(s->l1_table);
+    qcow2_free_clusters(bs, s->l1_table_offset, s->l1_size * sizeof(uint64_t),
+                        QCOW2_DISCARD_OTHER);
+    s->l1_table_offset = new_l1_table_offset;
+    s->l1_table = new_l1_table;
+    s->l1_size = new_l1_size;
+    return 0;
+ fail:
+    g_free(new_l1_table);
+    qcow2_free_clusters(bs, new_l1_table_offset, new_l1_size2,
+                        QCOW2_DISCARD_OTHER);
+    return ret;
+}
+
+/*
+ * l2_load
+ *
+ * Loads a L2 table into memory. If the table is in the cache, the cache
+ * is used; otherwise the L2 table is loaded from the image file.
+ *
+ * Returns a pointer to the L2 table on success, or NULL if the read from
+ * the image file failed.
+ */
+
+static int l2_load(BlockDriverState *bs, uint64_t l2_offset,
+    uint64_t **l2_table)
+{
+    BDRVQcowState *s = bs->opaque;
+    int ret;
+
+    ret = qcow2_cache_get(bs, s->l2_table_cache, l2_offset, (void**) l2_table);
+
+    return ret;
+}
+
+/*
+ * Writes one sector of the L1 table to the disk (can't update single entries
+ * and we really don't want bdrv_pread to perform a read-modify-write)
+ */
+#define L1_ENTRIES_PER_SECTOR (512 / 8)
+static int write_l1_entry(BlockDriverState *bs, int l1_index)
+{
+    BDRVQcowState *s = bs->opaque;
+    uint64_t buf[L1_ENTRIES_PER_SECTOR];
+    int l1_start_index;
+    int i, ret;
+
+    l1_start_index = l1_index & ~(L1_ENTRIES_PER_SECTOR - 1);
+    for (i = 0; i < L1_ENTRIES_PER_SECTOR; i++) {
+        buf[i] = cpu_to_be64(s->l1_table[l1_start_index + i]);
+    }
+
+    BLKDBG_EVENT(bs->file, BLKDBG_L1_UPDATE);
+    ret = bdrv_pwrite_sync(bs->file, s->l1_table_offset + 8 * l1_start_index,
+        buf, sizeof(buf));
+    if (ret < 0) {
+        return ret;
+    }
+
+    return 0;
+}
+
+/*
+ * l2_allocate
+ *
+ * Allocate a new l2 entry in the file. If l1_index points to an already
+ * used entry in the L2 table (i.e. we are doing a copy on write for the L2
+ * table) copy the contents of the old L2 table into the newly allocated one.
+ * Otherwise the new table is initialized with zeros.
+ *
+ */
+
+static int l2_allocate(BlockDriverState *bs, int l1_index, uint64_t **table)
+{
+    BDRVQcowState *s = bs->opaque;
+    uint64_t old_l2_offset;
+    uint64_t *l2_table;
+    int64_t l2_offset;
+    int ret;
+
+    old_l2_offset = s->l1_table[l1_index];
+
+    trace_qcow2_l2_allocate(bs, l1_index);
+
+    /* allocate a new l2 entry */
+
+    l2_offset = qcow2_alloc_clusters(bs, s->l2_size * sizeof(uint64_t));
+    if (l2_offset < 0) {
+        return l2_offset;
+    }
+
+    ret = qcow2_cache_flush(bs, s->refcount_block_cache);
+    if (ret < 0) {
+        goto fail;
+    }
+
+    /* allocate a new entry in the l2 cache */
+
+    trace_qcow2_l2_allocate_get_empty(bs, l1_index);
+    ret = qcow2_cache_get_empty(bs, s->l2_table_cache, l2_offset, (void**) table);
+    if (ret < 0) {
+        return ret;
+    }
+
+    l2_table = *table;
+
+    if ((old_l2_offset & L1E_OFFSET_MASK) == 0) {
+        /* if there was no old l2 table, clear the new table */
+        memset(l2_table, 0, s->l2_size * sizeof(uint64_t));
+    } else {
+        uint64_t* old_table;
+
+        /* if there was an old l2 table, read it from the disk */
+        BLKDBG_EVENT(bs->file, BLKDBG_L2_ALLOC_COW_READ);
+        ret = qcow2_cache_get(bs, s->l2_table_cache,
+            old_l2_offset & L1E_OFFSET_MASK,
+            (void**) &old_table);
+        if (ret < 0) {
+            goto fail;
+        }
+
+        memcpy(l2_table, old_table, s->cluster_size);
+
+        ret = qcow2_cache_put(bs, s->l2_table_cache, (void**) &old_table);
+        if (ret < 0) {
+            goto fail;
+        }
+    }
+
+    /* write the l2 table to the file */
+    BLKDBG_EVENT(bs->file, BLKDBG_L2_ALLOC_WRITE);
+
+    trace_qcow2_l2_allocate_write_l2(bs, l1_index);
+    qcow2_cache_entry_mark_dirty(s->l2_table_cache, l2_table);
+    ret = qcow2_cache_flush(bs, s->l2_table_cache);
+    if (ret < 0) {
+        goto fail;
+    }
+
+    /* update the L1 entry */
+    trace_qcow2_l2_allocate_write_l1(bs, l1_index);
+    s->l1_table[l1_index] = l2_offset | QCOW_OFLAG_COPIED;
+    ret = write_l1_entry(bs, l1_index);
+    if (ret < 0) {
+        goto fail;
+    }
+
+    *table = l2_table;
+    trace_qcow2_l2_allocate_done(bs, l1_index, 0);
+    return 0;
+
+fail:
+    trace_qcow2_l2_allocate_done(bs, l1_index, ret);
+    qcow2_cache_put(bs, s->l2_table_cache, (void**) table);
+    s->l1_table[l1_index] = old_l2_offset;
+    return ret;
+}
+
+/*
+ * Checks how many clusters in a given L2 table are contiguous in the image
+ * file. As soon as one of the flags in the bitmask stop_flags changes compared
+ * to the first cluster, the search is stopped and the cluster is not counted
+ * as contiguous. (This allows it, for example, to stop at the first compressed
+ * cluster which may require a different handling)
+ */
+static int count_contiguous_clusters(uint64_t nb_clusters, int cluster_size,
+        uint64_t *l2_table, uint64_t start, uint64_t stop_flags)
+{
+    int i;
+    uint64_t mask = stop_flags | L2E_OFFSET_MASK;
+    uint64_t offset = be64_to_cpu(l2_table[0]) & mask;
+
+    if (!offset)
+        return 0;
+
+    for (i = start; i < start + nb_clusters; i++) {
+        uint64_t l2_entry = be64_to_cpu(l2_table[i]) & mask;
+        if (offset + (uint64_t) i * cluster_size != l2_entry) {
+            break;
+        }
+    }
+
+	return (i - start);
+}
+
+static int count_contiguous_free_clusters(uint64_t nb_clusters, uint64_t *l2_table)
+{
+    int i;
+
+    for (i = 0; i < nb_clusters; i++) {
+        int type = qcow2_get_cluster_type(be64_to_cpu(l2_table[i]));
+
+        if (type != QCOW2_CLUSTER_UNALLOCATED) {
+            break;
+        }
+    }
+
+    return i;
+}
+
+/* The crypt function is compatible with the linux cryptoloop
+   algorithm for < 4 GB images. NOTE: out_buf == in_buf is
+   supported */
+void qcow2_encrypt_sectors(BDRVQcowState *s, int64_t sector_num,
+                           uint8_t *out_buf, const uint8_t *in_buf,
+                           int nb_sectors, int enc,
+                           const AES_KEY *key)
+{
+    union {
+        uint64_t ll[2];
+        uint8_t b[16];
+    } ivec;
+    int i;
+
+    for(i = 0; i < nb_sectors; i++) {
+        ivec.ll[0] = cpu_to_le64(sector_num);
+        ivec.ll[1] = 0;
+        AES_cbc_encrypt(in_buf, out_buf, 512, key,
+                        ivec.b, enc);
+        sector_num++;
+        in_buf += 512;
+        out_buf += 512;
+    }
+}
+
+static int coroutine_fn copy_sectors(BlockDriverState *bs,
+                                     uint64_t start_sect,
+                                     uint64_t cluster_offset,
+                                     int n_start, int n_end)
+{
+    BDRVQcowState *s = bs->opaque;
+    QEMUIOVector qiov;
+    struct iovec iov;
+    int n, ret;
+
+    /*
+     * If this is the last cluster and it is only partially used, we must only
+     * copy until the end of the image, or bdrv_check_request will fail for the
+     * bdrv_read/write calls below.
+     */
+    if (start_sect + n_end > bs->total_sectors) {
+        n_end = bs->total_sectors - start_sect;
+    }
+
+    n = n_end - n_start;
+    if (n <= 0) {
+        return 0;
+    }
+
+    iov.iov_len = n * BDRV_SECTOR_SIZE;
+    iov.iov_base = qemu_blockalign(bs, iov.iov_len);
+
+    qemu_iovec_init_external(&qiov, &iov, 1);
+
+    BLKDBG_EVENT(bs->file, BLKDBG_COW_READ);
+
+    /* Call .bdrv_co_readv() directly instead of using the public block-layer
+     * interface.  This avoids double I/O throttling and request tracking,
+     * which can lead to deadlock when block layer copy-on-read is enabled.
+     */
+    ret = bs->drv->bdrv_co_readv(bs, start_sect + n_start, n, &qiov);
+    if (ret < 0) {
+        goto out;
+    }
+
+    if (s->crypt_method) {
+        qcow2_encrypt_sectors(s, start_sect + n_start,
+                        iov.iov_base, iov.iov_base, n, 1,
+                        &s->aes_encrypt_key);
+    }
+
+    BLKDBG_EVENT(bs->file, BLKDBG_COW_WRITE);
+    ret = bdrv_co_writev(bs->file, (cluster_offset >> 9) + n_start, n, &qiov);
+    if (ret < 0) {
+        goto out;
+    }
+
+    ret = 0;
+out:
+    qemu_vfree(iov.iov_base);
+    return ret;
+}
+
+
+/*
+ * get_cluster_offset
+ *
+ * For a given offset of the disk image, find the cluster offset in
+ * qcow2 file. The offset is stored in *cluster_offset.
+ *
+ * on entry, *num is the number of contiguous sectors we'd like to
+ * access following offset.
+ *
+ * on exit, *num is the number of contiguous sectors we can read.
+ *
+ * Returns the cluster type (QCOW2_CLUSTER_*) on success, -errno in error
+ * cases.
+ */
+int qcow2_get_cluster_offset(BlockDriverState *bs, uint64_t offset,
+    int *num, uint64_t *cluster_offset)
+{
+    BDRVQcowState *s = bs->opaque;
+    unsigned int l2_index;
+    uint64_t l1_index, l2_offset, *l2_table;
+    int l1_bits, c;
+    unsigned int index_in_cluster, nb_clusters;
+    uint64_t nb_available, nb_needed;
+    int ret;
+
+    index_in_cluster = (offset >> 9) & (s->cluster_sectors - 1);
+    nb_needed = *num + index_in_cluster;
+
+    l1_bits = s->l2_bits + s->cluster_bits;
+
+    /* compute how many bytes there are between the offset and
+     * the end of the l1 entry
+     */
+
+    nb_available = (1ULL << l1_bits) - (offset & ((1ULL << l1_bits) - 1));
+
+    /* compute the number of available sectors */
+
+    nb_available = (nb_available >> 9) + index_in_cluster;
+
+    if (nb_needed > nb_available) {
+        nb_needed = nb_available;
+    }
+
+    *cluster_offset = 0;
+
+    /* seek the the l2 offset in the l1 table */
+
+    l1_index = offset >> l1_bits;
+    if (l1_index >= s->l1_size) {
+        ret = QCOW2_CLUSTER_UNALLOCATED;
+        goto out;
+    }
+
+    l2_offset = s->l1_table[l1_index] & L1E_OFFSET_MASK;
+    if (!l2_offset) {
+        ret = QCOW2_CLUSTER_UNALLOCATED;
+        goto out;
+    }
+
+    /* load the l2 table in memory */
+
+    ret = l2_load(bs, l2_offset, &l2_table);
+    if (ret < 0) {
+        return ret;
+    }
+
+    /* find the cluster offset for the given disk offset */
+
+    l2_index = (offset >> s->cluster_bits) & (s->l2_size - 1);
+    *cluster_offset = be64_to_cpu(l2_table[l2_index]);
+    nb_clusters = size_to_clusters(s, nb_needed << 9);
+
+    ret = qcow2_get_cluster_type(*cluster_offset);
+    switch (ret) {
+    case QCOW2_CLUSTER_COMPRESSED:
+        /* Compressed clusters can only be processed one by one */
+        c = 1;
+        *cluster_offset &= L2E_COMPRESSED_OFFSET_SIZE_MASK;
+        break;
+    case QCOW2_CLUSTER_ZERO:
+        if (s->qcow_version < 3) {
+            return -EIO;
+        }
+        c = count_contiguous_clusters(nb_clusters, s->cluster_size,
+                &l2_table[l2_index], 0,
+                QCOW_OFLAG_COMPRESSED | QCOW_OFLAG_ZERO);
+        *cluster_offset = 0;
+        break;
+    case QCOW2_CLUSTER_UNALLOCATED:
+        /* how many empty clusters ? */
+        c = count_contiguous_free_clusters(nb_clusters, &l2_table[l2_index]);
+        *cluster_offset = 0;
+        break;
+    case QCOW2_CLUSTER_NORMAL:
+        /* how many allocated clusters ? */
+        c = count_contiguous_clusters(nb_clusters, s->cluster_size,
+                &l2_table[l2_index], 0,
+                QCOW_OFLAG_COMPRESSED | QCOW_OFLAG_ZERO);
+        *cluster_offset &= L2E_OFFSET_MASK;
+        break;
+    default:
+        abort();
+    }
+
+    qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
+
+    nb_available = (c * s->cluster_sectors);
+
+out:
+    if (nb_available > nb_needed)
+        nb_available = nb_needed;
+
+    *num = nb_available - index_in_cluster;
+
+    return ret;
+}
+
+/*
+ * get_cluster_table
+ *
+ * for a given disk offset, load (and allocate if needed)
+ * the l2 table.
+ *
+ * the l2 table offset in the qcow2 file and the cluster index
+ * in the l2 table are given to the caller.
+ *
+ * Returns 0 on success, -errno in failure case
+ */
+static int get_cluster_table(BlockDriverState *bs, uint64_t offset,
+                             uint64_t **new_l2_table,
+                             int *new_l2_index)
+{
+    BDRVQcowState *s = bs->opaque;
+    unsigned int l2_index;
+    uint64_t l1_index, l2_offset;
+    uint64_t *l2_table = NULL;
+    int ret;
+
+    /* seek the the l2 offset in the l1 table */
+
+    l1_index = offset >> (s->l2_bits + s->cluster_bits);
+    if (l1_index >= s->l1_size) {
+        ret = qcow2_grow_l1_table(bs, l1_index + 1, false);
+        if (ret < 0) {
+            return ret;
+        }
+    }
+
+    assert(l1_index < s->l1_size);
+    l2_offset = s->l1_table[l1_index] & L1E_OFFSET_MASK;
+
+    /* seek the l2 table of the given l2 offset */
+
+    if (s->l1_table[l1_index] & QCOW_OFLAG_COPIED) {
+        /* load the l2 table in memory */
+        ret = l2_load(bs, l2_offset, &l2_table);
+        if (ret < 0) {
+            return ret;
+        }
+    } else {
+        /* First allocate a new L2 table (and do COW if needed) */
+        ret = l2_allocate(bs, l1_index, &l2_table);
+        if (ret < 0) {
+            return ret;
+        }
+
+        /* Then decrease the refcount of the old table */
+        if (l2_offset) {
+            qcow2_free_clusters(bs, l2_offset, s->l2_size * sizeof(uint64_t),
+                                QCOW2_DISCARD_OTHER);
+        }
+    }
+
+    /* find the cluster offset for the given disk offset */
+
+    l2_index = (offset >> s->cluster_bits) & (s->l2_size - 1);
+
+    *new_l2_table = l2_table;
+    *new_l2_index = l2_index;
+
+    return 0;
+}
+
+/*
+ * alloc_compressed_cluster_offset
+ *
+ * For a given offset of the disk image, return cluster offset in
+ * qcow2 file.
+ *
+ * If the offset is not found, allocate a new compressed cluster.
+ *
+ * Return the cluster offset if successful,
+ * Return 0, otherwise.
+ *
+ */
+
+uint64_t qcow2_alloc_compressed_cluster_offset(BlockDriverState *bs,
+                                               uint64_t offset,
+                                               int compressed_size)
+{
+    BDRVQcowState *s = bs->opaque;
+    int l2_index, ret;
+    uint64_t *l2_table;
+    int64_t cluster_offset;
+    int nb_csectors;
+
+    ret = get_cluster_table(bs, offset, &l2_table, &l2_index);
+    if (ret < 0) {
+        return 0;
+    }
+
+    /* Compression can't overwrite anything. Fail if the cluster was already
+     * allocated. */
+    cluster_offset = be64_to_cpu(l2_table[l2_index]);
+    if (cluster_offset & L2E_OFFSET_MASK) {
+        qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
+        return 0;
+    }
+
+    cluster_offset = qcow2_alloc_bytes(bs, compressed_size);
+    if (cluster_offset < 0) {
+        qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
+        return 0;
+    }
+
+    nb_csectors = ((cluster_offset + compressed_size - 1) >> 9) -
+                  (cluster_offset >> 9);
+
+    cluster_offset |= QCOW_OFLAG_COMPRESSED |
+                      ((uint64_t)nb_csectors << s->csize_shift);
+
+    /* update L2 table */
+
+    /* compressed clusters never have the copied flag */
+
+    BLKDBG_EVENT(bs->file, BLKDBG_L2_UPDATE_COMPRESSED);
+    qcow2_cache_entry_mark_dirty(s->l2_table_cache, l2_table);
+    l2_table[l2_index] = cpu_to_be64(cluster_offset);
+    ret = qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
+    if (ret < 0) {
+        return 0;
+    }
+
+    return cluster_offset;
+}
+
+static int perform_cow(BlockDriverState *bs, QCowL2Meta *m, Qcow2COWRegion *r)
+{
+    BDRVQcowState *s = bs->opaque;
+    int ret;
+
+    if (r->nb_sectors == 0) {
+        return 0;
+    }
+
+    qemu_co_mutex_unlock(&s->lock);
+    ret = copy_sectors(bs, m->offset / BDRV_SECTOR_SIZE, m->alloc_offset,
+                       r->offset / BDRV_SECTOR_SIZE,
+                       r->offset / BDRV_SECTOR_SIZE + r->nb_sectors);
+    qemu_co_mutex_lock(&s->lock);
+
+    if (ret < 0) {
+        return ret;
+    }
+
+    /*
+     * Before we update the L2 table to actually point to the new cluster, we
+     * need to be sure that the refcounts have been increased and COW was
+     * handled.
+     */
+    qcow2_cache_depends_on_flush(s->l2_table_cache);
+
+    return 0;
+}
+
+int qcow2_alloc_cluster_link_l2(BlockDriverState *bs, QCowL2Meta *m)
+{
+    BDRVQcowState *s = bs->opaque;
+    int i, j = 0, l2_index, ret;
+    uint64_t *old_cluster, *l2_table;
+    uint64_t cluster_offset = m->alloc_offset;
+
+    trace_qcow2_cluster_link_l2(qemu_coroutine_self(), m->nb_clusters);
+    assert(m->nb_clusters > 0);
+
+    old_cluster = g_malloc(m->nb_clusters * sizeof(uint64_t));
+
+    /* copy content of unmodified sectors */
+    ret = perform_cow(bs, m, &m->cow_start);
+    if (ret < 0) {
+        goto err;
+    }
+
+    ret = perform_cow(bs, m, &m->cow_end);
+    if (ret < 0) {
+        goto err;
+    }
+
+    /* Update L2 table. */
+    if (s->use_lazy_refcounts) {
+        qcow2_mark_dirty(bs);
+    }
+    if (qcow2_need_accurate_refcounts(s)) {
+        qcow2_cache_set_dependency(bs, s->l2_table_cache,
+                                   s->refcount_block_cache);
+    }
+
+    ret = get_cluster_table(bs, m->offset, &l2_table, &l2_index);
+    if (ret < 0) {
+        goto err;
+    }
+    qcow2_cache_entry_mark_dirty(s->l2_table_cache, l2_table);
+
+    for (i = 0; i < m->nb_clusters; i++) {
+        /* if two concurrent writes happen to the same unallocated cluster
+	 * each write allocates separate cluster and writes data concurrently.
+	 * The first one to complete updates l2 table with pointer to its
+	 * cluster the second one has to do RMW (which is done above by
+	 * copy_sectors()), update l2 table with its cluster pointer and free
+	 * old cluster. This is what this loop does */
+        if(l2_table[l2_index + i] != 0)
+            old_cluster[j++] = l2_table[l2_index + i];
+
+        l2_table[l2_index + i] = cpu_to_be64((cluster_offset +
+                    (i << s->cluster_bits)) | QCOW_OFLAG_COPIED);
+     }
+
+
+    ret = qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
+    if (ret < 0) {
+        goto err;
+    }
+
+    /*
+     * If this was a COW, we need to decrease the refcount of the old cluster.
+     * Also flush bs->file to get the right order for L2 and refcount update.
+     *
+     * Don't discard clusters that reach a refcount of 0 (e.g. compressed
+     * clusters), the next write will reuse them anyway.
+     */
+    if (j != 0) {
+        for (i = 0; i < j; i++) {
+            qcow2_free_any_clusters(bs, be64_to_cpu(old_cluster[i]), 1,
+                                    QCOW2_DISCARD_NEVER);
+        }
+    }
+
+    ret = 0;
+err:
+    g_free(old_cluster);
+    return ret;
+ }
+
+/*
+ * Returns the number of contiguous clusters that can be used for an allocating
+ * write, but require COW to be performed (this includes yet unallocated space,
+ * which must copy from the backing file)
+ */
+static int count_cow_clusters(BDRVQcowState *s, int nb_clusters,
+    uint64_t *l2_table, int l2_index)
+{
+    int i;
+
+    for (i = 0; i < nb_clusters; i++) {
+        uint64_t l2_entry = be64_to_cpu(l2_table[l2_index + i]);
+        int cluster_type = qcow2_get_cluster_type(l2_entry);
+
+        switch(cluster_type) {
+        case QCOW2_CLUSTER_NORMAL:
+            if (l2_entry & QCOW_OFLAG_COPIED) {
+                goto out;
+            }
+            break;
+        case QCOW2_CLUSTER_UNALLOCATED:
+        case QCOW2_CLUSTER_COMPRESSED:
+        case QCOW2_CLUSTER_ZERO:
+            break;
+        default:
+            abort();
+        }
+    }
+
+out:
+    assert(i <= nb_clusters);
+    return i;
+}
+
+/*
+ * Check if there already is an AIO write request in flight which allocates
+ * the same cluster. In this case we need to wait until the previous
+ * request has completed and updated the L2 table accordingly.
+ *
+ * Returns:
+ *   0       if there was no dependency. *cur_bytes indicates the number of
+ *           bytes from guest_offset that can be read before the next
+ *           dependency must be processed (or the request is complete)
+ *
+ *   -EAGAIN if we had to wait for another request, previously gathered
+ *           information on cluster allocation may be invalid now. The caller
+ *           must start over anyway, so consider *cur_bytes undefined.
+ */
+static int handle_dependencies(BlockDriverState *bs, uint64_t guest_offset,
+    uint64_t *cur_bytes, QCowL2Meta **m)
+{
+    BDRVQcowState *s = bs->opaque;
+    QCowL2Meta *old_alloc;
+    uint64_t bytes = *cur_bytes;
+
+    QLIST_FOREACH(old_alloc, &s->cluster_allocs, next_in_flight) {
+
+        uint64_t start = guest_offset;
+        uint64_t end = start + bytes;
+        uint64_t old_start = l2meta_cow_start(old_alloc);
+        uint64_t old_end = l2meta_cow_end(old_alloc);
+
+        if (end <= old_start || start >= old_end) {
+            /* No intersection */
+        } else {
+            if (start < old_start) {
+                /* Stop at the start of a running allocation */
+                bytes = old_start - start;
+            } else {
+                bytes = 0;
+            }
+
+            /* Stop if already an l2meta exists. After yielding, it wouldn't
+             * be valid any more, so we'd have to clean up the old L2Metas
+             * and deal with requests depending on them before starting to
+             * gather new ones. Not worth the trouble. */
+            if (bytes == 0 && *m) {
+                *cur_bytes = 0;
+                return 0;
+            }
+
+            if (bytes == 0) {
+                /* Wait for the dependency to complete. We need to recheck
+                 * the free/allocated clusters when we continue. */
+                qemu_co_mutex_unlock(&s->lock);
+                qemu_co_queue_wait(&old_alloc->dependent_requests);
+                qemu_co_mutex_lock(&s->lock);
+                return -EAGAIN;
+            }
+        }
+    }
+
+    /* Make sure that existing clusters and new allocations are only used up to
+     * the next dependency if we shortened the request above */
+    *cur_bytes = bytes;
+
+    return 0;
+}
+
+/*
+ * Checks how many already allocated clusters that don't require a copy on
+ * write there are at the given guest_offset (up to *bytes). If
+ * *host_offset is not zero, only physically contiguous clusters beginning at
+ * this host offset are counted.
+ *
+ * Note that guest_offset may not be cluster aligned. In this case, the
+ * returned *host_offset points to exact byte referenced by guest_offset and
+ * therefore isn't cluster aligned as well.
+ *
+ * Returns:
+ *   0:     if no allocated clusters are available at the given offset.
+ *          *bytes is normally unchanged. It is set to 0 if the cluster
+ *          is allocated and doesn't need COW, but doesn't have the right
+ *          physical offset.
+ *
+ *   1:     if allocated clusters that don't require a COW are available at
+ *          the requested offset. *bytes may have decreased and describes
+ *          the length of the area that can be written to.
+ *
+ *  -errno: in error cases
+ */
+static int handle_copied(BlockDriverState *bs, uint64_t guest_offset,
+    uint64_t *host_offset, uint64_t *bytes, QCowL2Meta **m)
+{
+    BDRVQcowState *s = bs->opaque;
+    int l2_index;
+    uint64_t cluster_offset;
+    uint64_t *l2_table;
+    unsigned int nb_clusters;
+    unsigned int keep_clusters;
+    int ret, pret;
+
+    trace_qcow2_handle_copied(qemu_coroutine_self(), guest_offset, *host_offset,
+                              *bytes);
+
+    assert(*host_offset == 0 ||    offset_into_cluster(s, guest_offset)
+                                == offset_into_cluster(s, *host_offset));
+
+    /*
+     * Calculate the number of clusters to look for. We stop at L2 table
+     * boundaries to keep things simple.
+     */
+    nb_clusters =
+        size_to_clusters(s, offset_into_cluster(s, guest_offset) + *bytes);
+
+    l2_index = offset_to_l2_index(s, guest_offset);
+    nb_clusters = MIN(nb_clusters, s->l2_size - l2_index);
+
+    /* Find L2 entry for the first involved cluster */
+    ret = get_cluster_table(bs, guest_offset, &l2_table, &l2_index);
+    if (ret < 0) {
+        return ret;
+    }
+
+    cluster_offset = be64_to_cpu(l2_table[l2_index]);
+
+    /* Check how many clusters are already allocated and don't need COW */
+    if (qcow2_get_cluster_type(cluster_offset) == QCOW2_CLUSTER_NORMAL
+        && (cluster_offset & QCOW_OFLAG_COPIED))
+    {
+        /* If a specific host_offset is required, check it */
+        bool offset_matches =
+            (cluster_offset & L2E_OFFSET_MASK) == *host_offset;
+
+        if (*host_offset != 0 && !offset_matches) {
+            *bytes = 0;
+            ret = 0;
+            goto out;
+        }
+
+        /* We keep all QCOW_OFLAG_COPIED clusters */
+        keep_clusters =
+            count_contiguous_clusters(nb_clusters, s->cluster_size,
+                                      &l2_table[l2_index], 0,
+                                      QCOW_OFLAG_COPIED | QCOW_OFLAG_ZERO);
+        assert(keep_clusters <= nb_clusters);
+
+        *bytes = MIN(*bytes,
+                 keep_clusters * s->cluster_size
+                 - offset_into_cluster(s, guest_offset));
+
+        ret = 1;
+    } else {
+        ret = 0;
+    }
+
+    /* Cleanup */
+out:
+    pret = qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
+    if (pret < 0) {
+        return pret;
+    }
+
+    /* Only return a host offset if we actually made progress. Otherwise we
+     * would make requirements for handle_alloc() that it can't fulfill */
+    if (ret) {
+        *host_offset = (cluster_offset & L2E_OFFSET_MASK)
+                     + offset_into_cluster(s, guest_offset);
+    }
+
+    return ret;
+}
+
+/*
+ * Allocates new clusters for the given guest_offset.
+ *
+ * At most *nb_clusters are allocated, and on return *nb_clusters is updated to
+ * contain the number of clusters that have been allocated and are contiguous
+ * in the image file.
+ *
+ * If *host_offset is non-zero, it specifies the offset in the image file at
+ * which the new clusters must start. *nb_clusters can be 0 on return in this
+ * case if the cluster at host_offset is already in use. If *host_offset is
+ * zero, the clusters can be allocated anywhere in the image file.
+ *
+ * *host_offset is updated to contain the offset into the image file at which
+ * the first allocated cluster starts.
+ *
+ * Return 0 on success and -errno in error cases. -EAGAIN means that the
+ * function has been waiting for another request and the allocation must be
+ * restarted, but the whole request should not be failed.
+ */
+static int do_alloc_cluster_offset(BlockDriverState *bs, uint64_t guest_offset,
+    uint64_t *host_offset, unsigned int *nb_clusters)
+{
+    BDRVQcowState *s = bs->opaque;
+
+    trace_qcow2_do_alloc_clusters_offset(qemu_coroutine_self(), guest_offset,
+                                         *host_offset, *nb_clusters);
+
+    /* Allocate new clusters */
+    trace_qcow2_cluster_alloc_phys(qemu_coroutine_self());
+    if (*host_offset == 0) {
+        int64_t cluster_offset =
+            qcow2_alloc_clusters(bs, *nb_clusters * s->cluster_size);
+        if (cluster_offset < 0) {
+            return cluster_offset;
+        }
+        *host_offset = cluster_offset;
+        return 0;
+    } else {
+        int ret = qcow2_alloc_clusters_at(bs, *host_offset, *nb_clusters);
+        if (ret < 0) {
+            return ret;
+        }
+        *nb_clusters = ret;
+        return 0;
+    }
+}
+
+/*
+ * Allocates new clusters for an area that either is yet unallocated or needs a
+ * copy on write. If *host_offset is non-zero, clusters are only allocated if
+ * the new allocation can match the specified host offset.
+ *
+ * Note that guest_offset may not be cluster aligned. In this case, the
+ * returned *host_offset points to exact byte referenced by guest_offset and
+ * therefore isn't cluster aligned as well.
+ *
+ * Returns:
+ *   0:     if no clusters could be allocated. *bytes is set to 0,
+ *          *host_offset is left unchanged.
+ *
+ *   1:     if new clusters were allocated. *bytes may be decreased if the
+ *          new allocation doesn't cover all of the requested area.
+ *          *host_offset is updated to contain the host offset of the first
+ *          newly allocated cluster.
+ *
+ *  -errno: in error cases
+ */
+static int handle_alloc(BlockDriverState *bs, uint64_t guest_offset,
+    uint64_t *host_offset, uint64_t *bytes, QCowL2Meta **m)
+{
+    BDRVQcowState *s = bs->opaque;
+    int l2_index;
+    uint64_t *l2_table;
+    uint64_t entry;
+    unsigned int nb_clusters;
+    int ret;
+
+    uint64_t alloc_cluster_offset;
+
+    trace_qcow2_handle_alloc(qemu_coroutine_self(), guest_offset, *host_offset,
+                             *bytes);
+    assert(*bytes > 0);
+
+    /*
+     * Calculate the number of clusters to look for. We stop at L2 table
+     * boundaries to keep things simple.
+     */
+    nb_clusters =
+        size_to_clusters(s, offset_into_cluster(s, guest_offset) + *bytes);
+
+    l2_index = offset_to_l2_index(s, guest_offset);
+    nb_clusters = MIN(nb_clusters, s->l2_size - l2_index);
+
+    /* Find L2 entry for the first involved cluster */
+    ret = get_cluster_table(bs, guest_offset, &l2_table, &l2_index);
+    if (ret < 0) {
+        return ret;
+    }
+
+    entry = be64_to_cpu(l2_table[l2_index]);
+
+    /* For the moment, overwrite compressed clusters one by one */
+    if (entry & QCOW_OFLAG_COMPRESSED) {
+        nb_clusters = 1;
+    } else {
+        nb_clusters = count_cow_clusters(s, nb_clusters, l2_table, l2_index);
+    }
+
+    /* This function is only called when there were no non-COW clusters, so if
+     * we can't find any unallocated or COW clusters either, something is
+     * wrong with our code. */
+    assert(nb_clusters > 0);
+
+    ret = qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
+    if (ret < 0) {
+        return ret;
+    }
+
+    /* Allocate, if necessary at a given offset in the image file */
+    alloc_cluster_offset = start_of_cluster(s, *host_offset);
+    ret = do_alloc_cluster_offset(bs, guest_offset, &alloc_cluster_offset,
+                                  &nb_clusters);
+    if (ret < 0) {
+        goto fail;
+    }
+
+    /* Can't extend contiguous allocation */
+    if (nb_clusters == 0) {
+        *bytes = 0;
+        return 0;
+    }
+
+    /*
+     * Save info needed for meta data update.
+     *
+     * requested_sectors: Number of sectors from the start of the first
+     * newly allocated cluster to the end of the (possibly shortened
+     * before) write request.
+     *
+     * avail_sectors: Number of sectors from the start of the first
+     * newly allocated to the end of the last newly allocated cluster.
+     *
+     * nb_sectors: The number of sectors from the start of the first
+     * newly allocated cluster to the end of the area that the write
+     * request actually writes to (excluding COW at the end)
+     */
+    int requested_sectors =
+        (*bytes + offset_into_cluster(s, guest_offset))
+        >> BDRV_SECTOR_BITS;
+    int avail_sectors = nb_clusters
+                        << (s->cluster_bits - BDRV_SECTOR_BITS);
+    int alloc_n_start = offset_into_cluster(s, guest_offset)
+                        >> BDRV_SECTOR_BITS;
+    int nb_sectors = MIN(requested_sectors, avail_sectors);
+    QCowL2Meta *old_m = *m;
+
+    *m = g_malloc0(sizeof(**m));
+
+    **m = (QCowL2Meta) {
+        .next           = old_m,
+
+        .alloc_offset   = alloc_cluster_offset,
+        .offset         = start_of_cluster(s, guest_offset),
+        .nb_clusters    = nb_clusters,
+        .nb_available   = nb_sectors,
+
+        .cow_start = {
+            .offset     = 0,
+            .nb_sectors = alloc_n_start,
+        },
+        .cow_end = {
+            .offset     = nb_sectors * BDRV_SECTOR_SIZE,
+            .nb_sectors = avail_sectors - nb_sectors,
+        },
+    };
+    qemu_co_queue_init(&(*m)->dependent_requests);
+    QLIST_INSERT_HEAD(&s->cluster_allocs, *m, next_in_flight);
+
+    *host_offset = alloc_cluster_offset + offset_into_cluster(s, guest_offset);
+    *bytes = MIN(*bytes, (nb_sectors * BDRV_SECTOR_SIZE)
+                         - offset_into_cluster(s, guest_offset));
+    assert(*bytes != 0);
+
+    return 1;
+
+fail:
+    if (*m && (*m)->nb_clusters > 0) {
+        QLIST_REMOVE(*m, next_in_flight);
+    }
+    return ret;
+}
+
+/*
+ * alloc_cluster_offset
+ *
+ * For a given offset on the virtual disk, find the cluster offset in qcow2
+ * file. If the offset is not found, allocate a new cluster.
+ *
+ * If the cluster was already allocated, m->nb_clusters is set to 0 and
+ * other fields in m are meaningless.
+ *
+ * If the cluster is newly allocated, m->nb_clusters is set to the number of
+ * contiguous clusters that have been allocated. In this case, the other
+ * fields of m are valid and contain information about the first allocated
+ * cluster.
+ *
+ * If the request conflicts with another write request in flight, the coroutine
+ * is queued and will be reentered when the dependency has completed.
+ *
+ * Return 0 on success and -errno in error cases
+ */
+int qcow2_alloc_cluster_offset(BlockDriverState *bs, uint64_t offset,
+    int n_start, int n_end, int *num, uint64_t *host_offset, QCowL2Meta **m)
+{
+    BDRVQcowState *s = bs->opaque;
+    uint64_t start, remaining;
+    uint64_t cluster_offset;
+    uint64_t cur_bytes;
+    int ret;
+
+    trace_qcow2_alloc_clusters_offset(qemu_coroutine_self(), offset,
+                                      n_start, n_end);
+
+    assert(n_start * BDRV_SECTOR_SIZE == offset_into_cluster(s, offset));
+    offset = start_of_cluster(s, offset);
+
+again:
+    start = offset + (n_start << BDRV_SECTOR_BITS);
+    remaining = (n_end - n_start) << BDRV_SECTOR_BITS;
+    cluster_offset = 0;
+    *host_offset = 0;
+    cur_bytes = 0;
+    *m = NULL;
+
+    while (true) {
+
+        if (!*host_offset) {
+            *host_offset = start_of_cluster(s, cluster_offset);
+        }
+
+        assert(remaining >= cur_bytes);
+
+        start           += cur_bytes;
+        remaining       -= cur_bytes;
+        cluster_offset  += cur_bytes;
+
+        if (remaining == 0) {
+            break;
+        }
+
+        cur_bytes = remaining;
+
+        /*
+         * Now start gathering as many contiguous clusters as possible:
+         *
+         * 1. Check for overlaps with in-flight allocations
+         *
+         *      a) Overlap not in the first cluster -> shorten this request and
+         *         let the caller handle the rest in its next loop iteration.
+         *
+         *      b) Real overlaps of two requests. Yield and restart the search
+         *         for contiguous clusters (the situation could have changed
+         *         while we were sleeping)
+         *
+         *      c) TODO: Request starts in the same cluster as the in-flight
+         *         allocation ends. Shorten the COW of the in-fight allocation,
+         *         set cluster_offset to write to the same cluster and set up
+         *         the right synchronisation between the in-flight request and
+         *         the new one.
+         */
+        ret = handle_dependencies(bs, start, &cur_bytes, m);
+        if (ret == -EAGAIN) {
+            /* Currently handle_dependencies() doesn't yield if we already had
+             * an allocation. If it did, we would have to clean up the L2Meta
+             * structs before starting over. */
+            assert(*m == NULL);
+            goto again;
+        } else if (ret < 0) {
+            return ret;
+        } else if (cur_bytes == 0) {
+            break;
+        } else {
+            /* handle_dependencies() may have decreased cur_bytes (shortened
+             * the allocations below) so that the next dependency is processed
+             * correctly during the next loop iteration. */
+        }
+
+        /*
+         * 2. Count contiguous COPIED clusters.
+         */
+        ret = handle_copied(bs, start, &cluster_offset, &cur_bytes, m);
+        if (ret < 0) {
+            return ret;
+        } else if (ret) {
+            continue;
+        } else if (cur_bytes == 0) {
+            break;
+        }
+
+        /*
+         * 3. If the request still hasn't completed, allocate new clusters,
+         *    considering any cluster_offset of steps 1c or 2.
+         */
+        ret = handle_alloc(bs, start, &cluster_offset, &cur_bytes, m);
+        if (ret < 0) {
+            return ret;
+        } else if (ret) {
+            continue;
+        } else {
+            assert(cur_bytes == 0);
+            break;
+        }
+    }
+
+    *num = (n_end - n_start) - (remaining >> BDRV_SECTOR_BITS);
+    assert(*num > 0);
+    assert(*host_offset != 0);
+
+    return 0;
+}
+
+static int decompress_buffer(uint8_t *out_buf, int out_buf_size,
+                             const uint8_t *buf, int buf_size)
+{
+    z_stream strm1, *strm = &strm1;
+    int ret, out_len;
+
+    memset(strm, 0, sizeof(*strm));
+
+    strm->next_in = (uint8_t *)buf;
+    strm->avail_in = buf_size;
+    strm->next_out = out_buf;
+    strm->avail_out = out_buf_size;
+
+    ret = inflateInit2(strm, -12);
+    if (ret != Z_OK)
+        return -1;
+    ret = inflate(strm, Z_FINISH);
+    out_len = strm->next_out - out_buf;
+    if ((ret != Z_STREAM_END && ret != Z_BUF_ERROR) ||
+        out_len != out_buf_size) {
+        inflateEnd(strm);
+        return -1;
+    }
+    inflateEnd(strm);
+    return 0;
+}
+
+int qcow2_decompress_cluster(BlockDriverState *bs, uint64_t cluster_offset)
+{
+    BDRVQcowState *s = bs->opaque;
+    int ret, csize, nb_csectors, sector_offset;
+    uint64_t coffset;
+
+    coffset = cluster_offset & s->cluster_offset_mask;
+    if (s->cluster_cache_offset != coffset) {
+        nb_csectors = ((cluster_offset >> s->csize_shift) & s->csize_mask) + 1;
+        sector_offset = coffset & 511;
+        csize = nb_csectors * 512 - sector_offset;
+        BLKDBG_EVENT(bs->file, BLKDBG_READ_COMPRESSED);
+        ret = bdrv_read(bs->file, coffset >> 9, s->cluster_data, nb_csectors);
+        if (ret < 0) {
+            return ret;
+        }
+        if (decompress_buffer(s->cluster_cache, s->cluster_size,
+                              s->cluster_data + sector_offset, csize) < 0) {
+            return -EIO;
+        }
+        s->cluster_cache_offset = coffset;
+    }
+    return 0;
+}
+
+/*
+ * This discards as many clusters of nb_clusters as possible at once (i.e.
+ * all clusters in the same L2 table) and returns the number of discarded
+ * clusters.
+ */
+static int discard_single_l2(BlockDriverState *bs, uint64_t offset,
+    unsigned int nb_clusters)
+{
+    BDRVQcowState *s = bs->opaque;
+    uint64_t *l2_table;
+    int l2_index;
+    int ret;
+    int i;
+
+    ret = get_cluster_table(bs, offset, &l2_table, &l2_index);
+    if (ret < 0) {
+        return ret;
+    }
+
+    /* Limit nb_clusters to one L2 table */
+    nb_clusters = MIN(nb_clusters, s->l2_size - l2_index);
+
+    for (i = 0; i < nb_clusters; i++) {
+        uint64_t old_offset;
+
+        old_offset = be64_to_cpu(l2_table[l2_index + i]);
+        if ((old_offset & L2E_OFFSET_MASK) == 0) {
+            continue;
+        }
+
+        /* First remove L2 entries */
+        qcow2_cache_entry_mark_dirty(s->l2_table_cache, l2_table);
+        l2_table[l2_index + i] = cpu_to_be64(0);
+
+        /* Then decrease the refcount */
+        qcow2_free_any_clusters(bs, old_offset, 1, QCOW2_DISCARD_REQUEST);
+    }
+
+    ret = qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
+    if (ret < 0) {
+        return ret;
+    }
+
+    return nb_clusters;
+}
+
+int qcow2_discard_clusters(BlockDriverState *bs, uint64_t offset,
+    int nb_sectors)
+{
+    BDRVQcowState *s = bs->opaque;
+    uint64_t end_offset;
+    unsigned int nb_clusters;
+    int ret;
+
+    end_offset = offset + (nb_sectors << BDRV_SECTOR_BITS);
+
+    /* Round start up and end down */
+    offset = align_offset(offset, s->cluster_size);
+    end_offset &= ~(s->cluster_size - 1);
+
+    if (offset > end_offset) {
+        return 0;
+    }
+
+    nb_clusters = size_to_clusters(s, end_offset - offset);
+
+    s->cache_discards = true;
+
+    /* Each L2 table is handled by its own loop iteration */
+    while (nb_clusters > 0) {
+        ret = discard_single_l2(bs, offset, nb_clusters);
+        if (ret < 0) {
+            goto fail;
+        }
+
+        nb_clusters -= ret;
+        offset += (ret * s->cluster_size);
+    }
+
+    ret = 0;
+fail:
+    s->cache_discards = false;
+    qcow2_process_discards(bs, ret);
+
+    return ret;
+}
+
+/*
+ * This zeroes as many clusters of nb_clusters as possible at once (i.e.
+ * all clusters in the same L2 table) and returns the number of zeroed
+ * clusters.
+ */
+static int zero_single_l2(BlockDriverState *bs, uint64_t offset,
+    unsigned int nb_clusters)
+{
+    BDRVQcowState *s = bs->opaque;
+    uint64_t *l2_table;
+    int l2_index;
+    int ret;
+    int i;
+
+    ret = get_cluster_table(bs, offset, &l2_table, &l2_index);
+    if (ret < 0) {
+        return ret;
+    }
+
+    /* Limit nb_clusters to one L2 table */
+    nb_clusters = MIN(nb_clusters, s->l2_size - l2_index);
+
+    for (i = 0; i < nb_clusters; i++) {
+        uint64_t old_offset;
+
+        old_offset = be64_to_cpu(l2_table[l2_index + i]);
+
+        /* Update L2 entries */
+        qcow2_cache_entry_mark_dirty(s->l2_table_cache, l2_table);
+        if (old_offset & QCOW_OFLAG_COMPRESSED) {
+            l2_table[l2_index + i] = cpu_to_be64(QCOW_OFLAG_ZERO);
+            qcow2_free_any_clusters(bs, old_offset, 1, QCOW2_DISCARD_REQUEST);
+        } else {
+            l2_table[l2_index + i] |= cpu_to_be64(QCOW_OFLAG_ZERO);
+        }
+    }
+
+    ret = qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
+    if (ret < 0) {
+        return ret;
+    }
+
+    return nb_clusters;
+}
+
+int qcow2_zero_clusters(BlockDriverState *bs, uint64_t offset, int nb_sectors)
+{
+    BDRVQcowState *s = bs->opaque;
+    unsigned int nb_clusters;
+    int ret;
+
+    /* The zero flag is only supported by version 3 and newer */
+    if (s->qcow_version < 3) {
+        return -ENOTSUP;
+    }
+
+    /* Each L2 table is handled by its own loop iteration */
+    nb_clusters = size_to_clusters(s, nb_sectors << BDRV_SECTOR_BITS);
+
+    s->cache_discards = true;
+
+    while (nb_clusters > 0) {
+        ret = zero_single_l2(bs, offset, nb_clusters);
+        if (ret < 0) {
+            goto fail;
+        }
+
+        nb_clusters -= ret;
+        offset += (ret * s->cluster_size);
+    }
+
+    ret = 0;
+fail:
+    s->cache_discards = false;
+    qcow2_process_discards(bs, ret);
+
+    return ret;
+}
diff --git a/contrib/qemu/block/qcow2-refcount.c b/contrib/qemu/block/qcow2-refcount.c
new file mode 100644
index 000000000..1244693f3
--- /dev/null
+++ b/contrib/qemu/block/qcow2-refcount.c
@@ -0,0 +1,1374 @@
+/*
+ * Block driver for the QCOW version 2 format
+ *
+ * Copyright (c) 2004-2006 Fabrice Bellard
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#include "qemu-common.h"
+#include "block/block_int.h"
+#include "block/qcow2.h"
+
+static int64_t alloc_clusters_noref(BlockDriverState *bs, int64_t size);
+static int QEMU_WARN_UNUSED_RESULT update_refcount(BlockDriverState *bs,
+                            int64_t offset, int64_t length,
+                            int addend, enum qcow2_discard_type type);
+
+
+/*********************************************************/
+/* refcount handling */
+
+int qcow2_refcount_init(BlockDriverState *bs)
+{
+    BDRVQcowState *s = bs->opaque;
+    int ret, refcount_table_size2, i;
+
+    refcount_table_size2 = s->refcount_table_size * sizeof(uint64_t);
+    s->refcount_table = g_malloc(refcount_table_size2);
+    if (s->refcount_table_size > 0) {
+        BLKDBG_EVENT(bs->file, BLKDBG_REFTABLE_LOAD);
+        ret = bdrv_pread(bs->file, s->refcount_table_offset,
+                         s->refcount_table, refcount_table_size2);
+        if (ret != refcount_table_size2)
+            goto fail;
+        for(i = 0; i < s->refcount_table_size; i++)
+            be64_to_cpus(&s->refcount_table[i]);
+    }
+    return 0;
+ fail:
+    return -ENOMEM;
+}
+
+void qcow2_refcount_close(BlockDriverState *bs)
+{
+    BDRVQcowState *s = bs->opaque;
+    g_free(s->refcount_table);
+}
+
+
+static int load_refcount_block(BlockDriverState *bs,
+                               int64_t refcount_block_offset,
+                               void **refcount_block)
+{
+    BDRVQcowState *s = bs->opaque;
+    int ret;
+
+    BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_LOAD);
+    ret = qcow2_cache_get(bs, s->refcount_block_cache, refcount_block_offset,
+        refcount_block);
+
+    return ret;
+}
+
+/*
+ * Returns the refcount of the cluster given by its index. Any non-negative
+ * return value is the refcount of the cluster, negative values are -errno
+ * and indicate an error.
+ */
+static int get_refcount(BlockDriverState *bs, int64_t cluster_index)
+{
+    BDRVQcowState *s = bs->opaque;
+    int refcount_table_index, block_index;
+    int64_t refcount_block_offset;
+    int ret;
+    uint16_t *refcount_block;
+    uint16_t refcount;
+
+    refcount_table_index = cluster_index >> (s->cluster_bits - REFCOUNT_SHIFT);
+    if (refcount_table_index >= s->refcount_table_size)
+        return 0;
+    refcount_block_offset = s->refcount_table[refcount_table_index];
+    if (!refcount_block_offset)
+        return 0;
+
+    ret = qcow2_cache_get(bs, s->refcount_block_cache, refcount_block_offset,
+        (void**) &refcount_block);
+    if (ret < 0) {
+        return ret;
+    }
+
+    block_index = cluster_index &
+        ((1 << (s->cluster_bits - REFCOUNT_SHIFT)) - 1);
+    refcount = be16_to_cpu(refcount_block[block_index]);
+
+    ret = qcow2_cache_put(bs, s->refcount_block_cache,
+        (void**) &refcount_block);
+    if (ret < 0) {
+        return ret;
+    }
+
+    return refcount;
+}
+
+/*
+ * Rounds the refcount table size up to avoid growing the table for each single
+ * refcount block that is allocated.
+ */
+static unsigned int next_refcount_table_size(BDRVQcowState *s,
+    unsigned int min_size)
+{
+    unsigned int min_clusters = (min_size >> (s->cluster_bits - 3)) + 1;
+    unsigned int refcount_table_clusters =
+        MAX(1, s->refcount_table_size >> (s->cluster_bits - 3));
+
+    while (min_clusters > refcount_table_clusters) {
+        refcount_table_clusters = (refcount_table_clusters * 3 + 1) / 2;
+    }
+
+    return refcount_table_clusters << (s->cluster_bits - 3);
+}
+
+
+/* Checks if two offsets are described by the same refcount block */
+static int in_same_refcount_block(BDRVQcowState *s, uint64_t offset_a,
+    uint64_t offset_b)
+{
+    uint64_t block_a = offset_a >> (2 * s->cluster_bits - REFCOUNT_SHIFT);
+    uint64_t block_b = offset_b >> (2 * s->cluster_bits - REFCOUNT_SHIFT);
+
+    return (block_a == block_b);
+}
+
+/*
+ * Loads a refcount block. If it doesn't exist yet, it is allocated first
+ * (including growing the refcount table if needed).
+ *
+ * Returns 0 on success or -errno in error case
+ */
+static int alloc_refcount_block(BlockDriverState *bs,
+    int64_t cluster_index, uint16_t **refcount_block)
+{
+    BDRVQcowState *s = bs->opaque;
+    unsigned int refcount_table_index;
+    int ret;
+
+    BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_ALLOC);
+
+    /* Find the refcount block for the given cluster */
+    refcount_table_index = cluster_index >> (s->cluster_bits - REFCOUNT_SHIFT);
+
+    if (refcount_table_index < s->refcount_table_size) {
+
+        uint64_t refcount_block_offset =
+            s->refcount_table[refcount_table_index] & REFT_OFFSET_MASK;
+
+        /* If it's already there, we're done */
+        if (refcount_block_offset) {
+             return load_refcount_block(bs, refcount_block_offset,
+                 (void**) refcount_block);
+        }
+    }
+
+    /*
+     * If we came here, we need to allocate something. Something is at least
+     * a cluster for the new refcount block. It may also include a new refcount
+     * table if the old refcount table is too small.
+     *
+     * Note that allocating clusters here needs some special care:
+     *
+     * - We can't use the normal qcow2_alloc_clusters(), it would try to
+     *   increase the refcount and very likely we would end up with an endless
+     *   recursion. Instead we must place the refcount blocks in a way that
+     *   they can describe them themselves.
+     *
+     * - We need to consider that at this point we are inside update_refcounts
+     *   and doing the initial refcount increase. This means that some clusters
+     *   have already been allocated by the caller, but their refcount isn't
+     *   accurate yet. free_cluster_index tells us where this allocation ends
+     *   as long as we don't overwrite it by freeing clusters.
+     *
+     * - alloc_clusters_noref and qcow2_free_clusters may load a different
+     *   refcount block into the cache
+     */
+
+    *refcount_block = NULL;
+
+    /* We write to the refcount table, so we might depend on L2 tables */
+    ret = qcow2_cache_flush(bs, s->l2_table_cache);
+    if (ret < 0) {
+        return ret;
+    }
+
+    /* Allocate the refcount block itself and mark it as used */
+    int64_t new_block = alloc_clusters_noref(bs, s->cluster_size);
+    if (new_block < 0) {
+        return new_block;
+    }
+
+#ifdef DEBUG_ALLOC2
+    fprintf(stderr, "qcow2: Allocate refcount block %d for %" PRIx64
+        " at %" PRIx64 "\n",
+        refcount_table_index, cluster_index << s->cluster_bits, new_block);
+#endif
+
+    if (in_same_refcount_block(s, new_block, cluster_index << s->cluster_bits)) {
+        /* Zero the new refcount block before updating it */
+        ret = qcow2_cache_get_empty(bs, s->refcount_block_cache, new_block,
+            (void**) refcount_block);
+        if (ret < 0) {
+            goto fail_block;
+        }
+
+        memset(*refcount_block, 0, s->cluster_size);
+
+        /* The block describes itself, need to update the cache */
+        int block_index = (new_block >> s->cluster_bits) &
+            ((1 << (s->cluster_bits - REFCOUNT_SHIFT)) - 1);
+        (*refcount_block)[block_index] = cpu_to_be16(1);
+    } else {
+        /* Described somewhere else. This can recurse at most twice before we
+         * arrive at a block that describes itself. */
+        ret = update_refcount(bs, new_block, s->cluster_size, 1,
+                              QCOW2_DISCARD_NEVER);
+        if (ret < 0) {
+            goto fail_block;
+        }
+
+        ret = qcow2_cache_flush(bs, s->refcount_block_cache);
+        if (ret < 0) {
+            goto fail_block;
+        }
+
+        /* Initialize the new refcount block only after updating its refcount,
+         * update_refcount uses the refcount cache itself */
+        ret = qcow2_cache_get_empty(bs, s->refcount_block_cache, new_block,
+            (void**) refcount_block);
+        if (ret < 0) {
+            goto fail_block;
+        }
+
+        memset(*refcount_block, 0, s->cluster_size);
+    }
+
+    /* Now the new refcount block needs to be written to disk */
+    BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_ALLOC_WRITE);
+    qcow2_cache_entry_mark_dirty(s->refcount_block_cache, *refcount_block);
+    ret = qcow2_cache_flush(bs, s->refcount_block_cache);
+    if (ret < 0) {
+        goto fail_block;
+    }
+
+    /* If the refcount table is big enough, just hook the block up there */
+    if (refcount_table_index < s->refcount_table_size) {
+        uint64_t data64 = cpu_to_be64(new_block);
+        BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_ALLOC_HOOKUP);
+        ret = bdrv_pwrite_sync(bs->file,
+            s->refcount_table_offset + refcount_table_index * sizeof(uint64_t),
+            &data64, sizeof(data64));
+        if (ret < 0) {
+            goto fail_block;
+        }
+
+        s->refcount_table[refcount_table_index] = new_block;
+        return 0;
+    }
+
+    ret = qcow2_cache_put(bs, s->refcount_block_cache, (void**) refcount_block);
+    if (ret < 0) {
+        goto fail_block;
+    }
+
+    /*
+     * If we come here, we need to grow the refcount table. Again, a new
+     * refcount table needs some space and we can't simply allocate to avoid
+     * endless recursion.
+     *
+     * Therefore let's grab new refcount blocks at the end of the image, which
+     * will describe themselves and the new refcount table. This way we can
+     * reference them only in the new table and do the switch to the new
+     * refcount table at once without producing an inconsistent state in
+     * between.
+     */
+    BLKDBG_EVENT(bs->file, BLKDBG_REFTABLE_GROW);
+
+    /* Calculate the number of refcount blocks needed so far */
+    uint64_t refcount_block_clusters = 1 << (s->cluster_bits - REFCOUNT_SHIFT);
+    uint64_t blocks_used = (s->free_cluster_index +
+        refcount_block_clusters - 1) / refcount_block_clusters;
+
+    /* And now we need at least one block more for the new metadata */
+    uint64_t table_size = next_refcount_table_size(s, blocks_used + 1);
+    uint64_t last_table_size;
+    uint64_t blocks_clusters;
+    do {
+        uint64_t table_clusters =
+            size_to_clusters(s, table_size * sizeof(uint64_t));
+        blocks_clusters = 1 +
+            ((table_clusters + refcount_block_clusters - 1)
+            / refcount_block_clusters);
+        uint64_t meta_clusters = table_clusters + blocks_clusters;
+
+        last_table_size = table_size;
+        table_size = next_refcount_table_size(s, blocks_used +
+            ((meta_clusters + refcount_block_clusters - 1)
+            / refcount_block_clusters));
+
+    } while (last_table_size != table_size);
+
+#ifdef DEBUG_ALLOC2
+    fprintf(stderr, "qcow2: Grow refcount table %" PRId32 " => %" PRId64 "\n",
+        s->refcount_table_size, table_size);
+#endif
+
+    /* Create the new refcount table and blocks */
+    uint64_t meta_offset = (blocks_used * refcount_block_clusters) *
+        s->cluster_size;
+    uint64_t table_offset = meta_offset + blocks_clusters * s->cluster_size;
+    uint16_t *new_blocks = g_malloc0(blocks_clusters * s->cluster_size);
+    uint64_t *new_table = g_malloc0(table_size * sizeof(uint64_t));
+
+    assert(meta_offset >= (s->free_cluster_index * s->cluster_size));
+
+    /* Fill the new refcount table */
+    memcpy(new_table, s->refcount_table,
+        s->refcount_table_size * sizeof(uint64_t));
+    new_table[refcount_table_index] = new_block;
+
+    int i;
+    for (i = 0; i < blocks_clusters; i++) {
+        new_table[blocks_used + i] = meta_offset + (i * s->cluster_size);
+    }
+
+    /* Fill the refcount blocks */
+    uint64_t table_clusters = size_to_clusters(s, table_size * sizeof(uint64_t));
+    int block = 0;
+    for (i = 0; i < table_clusters + blocks_clusters; i++) {
+        new_blocks[block++] = cpu_to_be16(1);
+    }
+
+    /* Write refcount blocks to disk */
+    BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_ALLOC_WRITE_BLOCKS);
+    ret = bdrv_pwrite_sync(bs->file, meta_offset, new_blocks,
+        blocks_clusters * s->cluster_size);
+    g_free(new_blocks);
+    if (ret < 0) {
+        goto fail_table;
+    }
+
+    /* Write refcount table to disk */
+    for(i = 0; i < table_size; i++) {
+        cpu_to_be64s(&new_table[i]);
+    }
+
+    BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_ALLOC_WRITE_TABLE);
+    ret = bdrv_pwrite_sync(bs->file, table_offset, new_table,
+        table_size * sizeof(uint64_t));
+    if (ret < 0) {
+        goto fail_table;
+    }
+
+    for(i = 0; i < table_size; i++) {
+        be64_to_cpus(&new_table[i]);
+    }
+
+    /* Hook up the new refcount table in the qcow2 header */
+    uint8_t data[12];
+    cpu_to_be64w((uint64_t*)data, table_offset);
+    cpu_to_be32w((uint32_t*)(data + 8), table_clusters);
+    BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_ALLOC_SWITCH_TABLE);
+    ret = bdrv_pwrite_sync(bs->file, offsetof(QCowHeader, refcount_table_offset),
+        data, sizeof(data));
+    if (ret < 0) {
+        goto fail_table;
+    }
+
+    /* And switch it in memory */
+    uint64_t old_table_offset = s->refcount_table_offset;
+    uint64_t old_table_size = s->refcount_table_size;
+
+    g_free(s->refcount_table);
+    s->refcount_table = new_table;
+    s->refcount_table_size = table_size;
+    s->refcount_table_offset = table_offset;
+
+    /* Free old table. Remember, we must not change free_cluster_index */
+    uint64_t old_free_cluster_index = s->free_cluster_index;
+    qcow2_free_clusters(bs, old_table_offset, old_table_size * sizeof(uint64_t),
+                        QCOW2_DISCARD_OTHER);
+    s->free_cluster_index = old_free_cluster_index;
+
+    ret = load_refcount_block(bs, new_block, (void**) refcount_block);
+    if (ret < 0) {
+        return ret;
+    }
+
+    return 0;
+
+fail_table:
+    g_free(new_table);
+fail_block:
+    if (*refcount_block != NULL) {
+        qcow2_cache_put(bs, s->refcount_block_cache, (void**) refcount_block);
+    }
+    return ret;
+}
+
+void qcow2_process_discards(BlockDriverState *bs, int ret)
+{
+    BDRVQcowState *s = bs->opaque;
+    Qcow2DiscardRegion *d, *next;
+
+    QTAILQ_FOREACH_SAFE(d, &s->discards, next, next) {
+        QTAILQ_REMOVE(&s->discards, d, next);
+
+        /* Discard is optional, ignore the return value */
+        if (ret >= 0) {
+            bdrv_discard(bs->file,
+                         d->offset >> BDRV_SECTOR_BITS,
+                         d->bytes >> BDRV_SECTOR_BITS);
+        }
+
+        g_free(d);
+    }
+}
+
+static void update_refcount_discard(BlockDriverState *bs,
+                                    uint64_t offset, uint64_t length)
+{
+    BDRVQcowState *s = bs->opaque;
+    Qcow2DiscardRegion *d, *p, *next;
+
+    QTAILQ_FOREACH(d, &s->discards, next) {
+        uint64_t new_start = MIN(offset, d->offset);
+        uint64_t new_end = MAX(offset + length, d->offset + d->bytes);
+
+        if (new_end - new_start <= length + d->bytes) {
+            /* There can't be any overlap, areas ending up here have no
+             * references any more and therefore shouldn't get freed another
+             * time. */
+            assert(d->bytes + length == new_end - new_start);
+            d->offset = new_start;
+            d->bytes = new_end - new_start;
+            goto found;
+        }
+    }
+
+    d = g_malloc(sizeof(*d));
+    *d = (Qcow2DiscardRegion) {
+        .bs     = bs,
+        .offset = offset,
+        .bytes  = length,
+    };
+    QTAILQ_INSERT_TAIL(&s->discards, d, next);
+
+found:
+    /* Merge discard requests if they are adjacent now */
+    QTAILQ_FOREACH_SAFE(p, &s->discards, next, next) {
+        if (p == d
+            || p->offset > d->offset + d->bytes
+            || d->offset > p->offset + p->bytes)
+        {
+            continue;
+        }
+
+        /* Still no overlap possible */
+        assert(p->offset == d->offset + d->bytes
+            || d->offset == p->offset + p->bytes);
+
+        QTAILQ_REMOVE(&s->discards, p, next);
+        d->offset = MIN(d->offset, p->offset);
+        d->bytes += p->bytes;
+    }
+}
+
+/* XXX: cache several refcount block clusters ? */
+static int QEMU_WARN_UNUSED_RESULT update_refcount(BlockDriverState *bs,
+    int64_t offset, int64_t length, int addend, enum qcow2_discard_type type)
+{
+    BDRVQcowState *s = bs->opaque;
+    int64_t start, last, cluster_offset;
+    uint16_t *refcount_block = NULL;
+    int64_t old_table_index = -1;
+    int ret;
+
+#ifdef DEBUG_ALLOC2
+    fprintf(stderr, "update_refcount: offset=%" PRId64 " size=%" PRId64 " addend=%d\n",
+           offset, length, addend);
+#endif
+    if (length < 0) {
+        return -EINVAL;
+    } else if (length == 0) {
+        return 0;
+    }
+
+    if (addend < 0) {
+        qcow2_cache_set_dependency(bs, s->refcount_block_cache,
+            s->l2_table_cache);
+    }
+
+    start = offset & ~(s->cluster_size - 1);
+    last = (offset + length - 1) & ~(s->cluster_size - 1);
+    for(cluster_offset = start; cluster_offset <= last;
+        cluster_offset += s->cluster_size)
+    {
+        int block_index, refcount;
+        int64_t cluster_index = cluster_offset >> s->cluster_bits;
+        int64_t table_index =
+            cluster_index >> (s->cluster_bits - REFCOUNT_SHIFT);
+
+        /* Load the refcount block and allocate it if needed */
+        if (table_index != old_table_index) {
+            if (refcount_block) {
+                ret = qcow2_cache_put(bs, s->refcount_block_cache,
+                    (void**) &refcount_block);
+                if (ret < 0) {
+                    goto fail;
+                }
+            }
+
+            ret = alloc_refcount_block(bs, cluster_index, &refcount_block);
+            if (ret < 0) {
+                goto fail;
+            }
+        }
+        old_table_index = table_index;
+
+        qcow2_cache_entry_mark_dirty(s->refcount_block_cache, refcount_block);
+
+        /* we can update the count and save it */
+        block_index = cluster_index &
+            ((1 << (s->cluster_bits - REFCOUNT_SHIFT)) - 1);
+
+        refcount = be16_to_cpu(refcount_block[block_index]);
+        refcount += addend;
+        if (refcount < 0 || refcount > 0xffff) {
+            ret = -EINVAL;
+            goto fail;
+        }
+        if (refcount == 0 && cluster_index < s->free_cluster_index) {
+            s->free_cluster_index = cluster_index;
+        }
+        refcount_block[block_index] = cpu_to_be16(refcount);
+
+        if (refcount == 0 && s->discard_passthrough[type]) {
+            update_refcount_discard(bs, cluster_offset, s->cluster_size);
+        }
+    }
+
+    ret = 0;
+fail:
+    if (!s->cache_discards) {
+        qcow2_process_discards(bs, ret);
+    }
+
+    /* Write last changed block to disk */
+    if (refcount_block) {
+        int wret;
+        wret = qcow2_cache_put(bs, s->refcount_block_cache,
+            (void**) &refcount_block);
+        if (wret < 0) {
+            return ret < 0 ? ret : wret;
+        }
+    }
+
+    /*
+     * Try do undo any updates if an error is returned (This may succeed in
+     * some cases like ENOSPC for allocating a new refcount block)
+     */
+    if (ret < 0) {
+        int dummy;
+        dummy = update_refcount(bs, offset, cluster_offset - offset, -addend,
+                                QCOW2_DISCARD_NEVER);
+        (void)dummy;
+    }
+
+    return ret;
+}
+
+/*
+ * Increases or decreases the refcount of a given cluster by one.
+ * addend must be 1 or -1.
+ *
+ * If the return value is non-negative, it is the new refcount of the cluster.
+ * If it is negative, it is -errno and indicates an error.
+ */
+static int update_cluster_refcount(BlockDriverState *bs,
+                                   int64_t cluster_index,
+                                   int addend,
+                                   enum qcow2_discard_type type)
+{
+    BDRVQcowState *s = bs->opaque;
+    int ret;
+
+    ret = update_refcount(bs, cluster_index << s->cluster_bits, 1, addend,
+                          type);
+    if (ret < 0) {
+        return ret;
+    }
+
+    return get_refcount(bs, cluster_index);
+}
+
+
+
+/*********************************************************/
+/* cluster allocation functions */
+
+
+
+/* return < 0 if error */
+static int64_t alloc_clusters_noref(BlockDriverState *bs, int64_t size)
+{
+    BDRVQcowState *s = bs->opaque;
+    int i, nb_clusters, refcount;
+
+    nb_clusters = size_to_clusters(s, size);
+retry:
+    for(i = 0; i < nb_clusters; i++) {
+        int64_t next_cluster_index = s->free_cluster_index++;
+        refcount = get_refcount(bs, next_cluster_index);
+
+        if (refcount < 0) {
+            return refcount;
+        } else if (refcount != 0) {
+            goto retry;
+        }
+    }
+#ifdef DEBUG_ALLOC2
+    fprintf(stderr, "alloc_clusters: size=%" PRId64 " -> %" PRId64 "\n",
+            size,
+            (s->free_cluster_index - nb_clusters) << s->cluster_bits);
+#endif
+    return (s->free_cluster_index - nb_clusters) << s->cluster_bits;
+}
+
+int64_t qcow2_alloc_clusters(BlockDriverState *bs, int64_t size)
+{
+    int64_t offset;
+    int ret;
+
+    BLKDBG_EVENT(bs->file, BLKDBG_CLUSTER_ALLOC);
+    offset = alloc_clusters_noref(bs, size);
+    if (offset < 0) {
+        return offset;
+    }
+
+    ret = update_refcount(bs, offset, size, 1, QCOW2_DISCARD_NEVER);
+    if (ret < 0) {
+        return ret;
+    }
+
+    return offset;
+}
+
+int qcow2_alloc_clusters_at(BlockDriverState *bs, uint64_t offset,
+    int nb_clusters)
+{
+    BDRVQcowState *s = bs->opaque;
+    uint64_t cluster_index;
+    uint64_t old_free_cluster_index;
+    int i, refcount, ret;
+
+    /* Check how many clusters there are free */
+    cluster_index = offset >> s->cluster_bits;
+    for(i = 0; i < nb_clusters; i++) {
+        refcount = get_refcount(bs, cluster_index++);
+
+        if (refcount < 0) {
+            return refcount;
+        } else if (refcount != 0) {
+            break;
+        }
+    }
+
+    /* And then allocate them */
+    old_free_cluster_index = s->free_cluster_index;
+    s->free_cluster_index = cluster_index + i;
+
+    ret = update_refcount(bs, offset, i << s->cluster_bits, 1,
+                          QCOW2_DISCARD_NEVER);
+    if (ret < 0) {
+        return ret;
+    }
+
+    s->free_cluster_index = old_free_cluster_index;
+
+    return i;
+}
+
+/* only used to allocate compressed sectors. We try to allocate
+   contiguous sectors. size must be <= cluster_size */
+int64_t qcow2_alloc_bytes(BlockDriverState *bs, int size)
+{
+    BDRVQcowState *s = bs->opaque;
+    int64_t offset, cluster_offset;
+    int free_in_cluster;
+
+    BLKDBG_EVENT(bs->file, BLKDBG_CLUSTER_ALLOC_BYTES);
+    assert(size > 0 && size <= s->cluster_size);
+    if (s->free_byte_offset == 0) {
+        offset = qcow2_alloc_clusters(bs, s->cluster_size);
+        if (offset < 0) {
+            return offset;
+        }
+        s->free_byte_offset = offset;
+    }
+ redo:
+    free_in_cluster = s->cluster_size -
+        (s->free_byte_offset & (s->cluster_size - 1));
+    if (size <= free_in_cluster) {
+        /* enough space in current cluster */
+        offset = s->free_byte_offset;
+        s->free_byte_offset += size;
+        free_in_cluster -= size;
+        if (free_in_cluster == 0)
+            s->free_byte_offset = 0;
+        if ((offset & (s->cluster_size - 1)) != 0)
+            update_cluster_refcount(bs, offset >> s->cluster_bits, 1,
+                                    QCOW2_DISCARD_NEVER);
+    } else {
+        offset = qcow2_alloc_clusters(bs, s->cluster_size);
+        if (offset < 0) {
+            return offset;
+        }
+        cluster_offset = s->free_byte_offset & ~(s->cluster_size - 1);
+        if ((cluster_offset + s->cluster_size) == offset) {
+            /* we are lucky: contiguous data */
+            offset = s->free_byte_offset;
+            update_cluster_refcount(bs, offset >> s->cluster_bits, 1,
+                                    QCOW2_DISCARD_NEVER);
+            s->free_byte_offset += size;
+        } else {
+            s->free_byte_offset = offset;
+            goto redo;
+        }
+    }
+
+    /* The cluster refcount was incremented, either by qcow2_alloc_clusters()
+     * or explicitly by update_cluster_refcount().  Refcount blocks must be
+     * flushed before the caller's L2 table updates.
+     */
+    qcow2_cache_set_dependency(bs, s->l2_table_cache, s->refcount_block_cache);
+    return offset;
+}
+
+void qcow2_free_clusters(BlockDriverState *bs,
+                          int64_t offset, int64_t size,
+                          enum qcow2_discard_type type)
+{
+    int ret;
+
+    BLKDBG_EVENT(bs->file, BLKDBG_CLUSTER_FREE);
+    ret = update_refcount(bs, offset, size, -1, type);
+    if (ret < 0) {
+        fprintf(stderr, "qcow2_free_clusters failed: %s\n", strerror(-ret));
+        /* TODO Remember the clusters to free them later and avoid leaking */
+    }
+}
+
+/*
+ * Free a cluster using its L2 entry (handles clusters of all types, e.g.
+ * normal cluster, compressed cluster, etc.)
+ */
+void qcow2_free_any_clusters(BlockDriverState *bs, uint64_t l2_entry,
+                             int nb_clusters, enum qcow2_discard_type type)
+{
+    BDRVQcowState *s = bs->opaque;
+
+    switch (qcow2_get_cluster_type(l2_entry)) {
+    case QCOW2_CLUSTER_COMPRESSED:
+        {
+            int nb_csectors;
+            nb_csectors = ((l2_entry >> s->csize_shift) &
+                           s->csize_mask) + 1;
+            qcow2_free_clusters(bs,
+                (l2_entry & s->cluster_offset_mask) & ~511,
+                nb_csectors * 512, type);
+        }
+        break;
+    case QCOW2_CLUSTER_NORMAL:
+        qcow2_free_clusters(bs, l2_entry & L2E_OFFSET_MASK,
+                            nb_clusters << s->cluster_bits, type);
+        break;
+    case QCOW2_CLUSTER_UNALLOCATED:
+    case QCOW2_CLUSTER_ZERO:
+        break;
+    default:
+        abort();
+    }
+}
+
+
+
+/*********************************************************/
+/* snapshots and image creation */
+
+
+
+/* update the refcounts of snapshots and the copied flag */
+int qcow2_update_snapshot_refcount(BlockDriverState *bs,
+    int64_t l1_table_offset, int l1_size, int addend)
+{
+    BDRVQcowState *s = bs->opaque;
+    uint64_t *l1_table, *l2_table, l2_offset, offset, l1_size2, l1_allocated;
+    int64_t old_offset, old_l2_offset;
+    int i, j, l1_modified = 0, nb_csectors, refcount;
+    int ret;
+
+    l2_table = NULL;
+    l1_table = NULL;
+    l1_size2 = l1_size * sizeof(uint64_t);
+
+    s->cache_discards = true;
+
+    /* WARNING: qcow2_snapshot_goto relies on this function not using the
+     * l1_table_offset when it is the current s->l1_table_offset! Be careful
+     * when changing this! */
+    if (l1_table_offset != s->l1_table_offset) {
+        l1_table = g_malloc0(align_offset(l1_size2, 512));
+        l1_allocated = 1;
+
+        ret = bdrv_pread(bs->file, l1_table_offset, l1_table, l1_size2);
+        if (ret < 0) {
+            goto fail;
+        }
+
+        for(i = 0;i < l1_size; i++)
+            be64_to_cpus(&l1_table[i]);
+    } else {
+        assert(l1_size == s->l1_size);
+        l1_table = s->l1_table;
+        l1_allocated = 0;
+    }
+
+    for(i = 0; i < l1_size; i++) {
+        l2_offset = l1_table[i];
+        if (l2_offset) {
+            old_l2_offset = l2_offset;
+            l2_offset &= L1E_OFFSET_MASK;
+
+            ret = qcow2_cache_get(bs, s->l2_table_cache, l2_offset,
+                (void**) &l2_table);
+            if (ret < 0) {
+                goto fail;
+            }
+
+            for(j = 0; j < s->l2_size; j++) {
+                offset = be64_to_cpu(l2_table[j]);
+                if (offset != 0) {
+                    old_offset = offset;
+                    offset &= ~QCOW_OFLAG_COPIED;
+                    if (offset & QCOW_OFLAG_COMPRESSED) {
+                        nb_csectors = ((offset >> s->csize_shift) &
+                                       s->csize_mask) + 1;
+                        if (addend != 0) {
+                            int ret;
+                            ret = update_refcount(bs,
+                                (offset & s->cluster_offset_mask) & ~511,
+                                nb_csectors * 512, addend,
+                                QCOW2_DISCARD_SNAPSHOT);
+                            if (ret < 0) {
+                                goto fail;
+                            }
+                        }
+                        /* compressed clusters are never modified */
+                        refcount = 2;
+                    } else {
+                        uint64_t cluster_index = (offset & L2E_OFFSET_MASK) >> s->cluster_bits;
+                        if (addend != 0) {
+                            refcount = update_cluster_refcount(bs, cluster_index, addend,
+                                                               QCOW2_DISCARD_SNAPSHOT);
+                        } else {
+                            refcount = get_refcount(bs, cluster_index);
+                        }
+
+                        if (refcount < 0) {
+                            ret = refcount;
+                            goto fail;
+                        }
+                    }
+
+                    if (refcount == 1) {
+                        offset |= QCOW_OFLAG_COPIED;
+                    }
+                    if (offset != old_offset) {
+                        if (addend > 0) {
+                            qcow2_cache_set_dependency(bs, s->l2_table_cache,
+                                s->refcount_block_cache);
+                        }
+                        l2_table[j] = cpu_to_be64(offset);
+                        qcow2_cache_entry_mark_dirty(s->l2_table_cache, l2_table);
+                    }
+                }
+            }
+
+            ret = qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
+            if (ret < 0) {
+                goto fail;
+            }
+
+
+            if (addend != 0) {
+                refcount = update_cluster_refcount(bs, l2_offset >> s->cluster_bits, addend,
+                                                   QCOW2_DISCARD_SNAPSHOT);
+            } else {
+                refcount = get_refcount(bs, l2_offset >> s->cluster_bits);
+            }
+            if (refcount < 0) {
+                ret = refcount;
+                goto fail;
+            } else if (refcount == 1) {
+                l2_offset |= QCOW_OFLAG_COPIED;
+            }
+            if (l2_offset != old_l2_offset) {
+                l1_table[i] = l2_offset;
+                l1_modified = 1;
+            }
+        }
+    }
+
+    ret = bdrv_flush(bs);
+fail:
+    if (l2_table) {
+        qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
+    }
+
+    s->cache_discards = false;
+    qcow2_process_discards(bs, ret);
+
+    /* Update L1 only if it isn't deleted anyway (addend = -1) */
+    if (ret == 0 && addend >= 0 && l1_modified) {
+        for (i = 0; i < l1_size; i++) {
+            cpu_to_be64s(&l1_table[i]);
+        }
+
+        ret = bdrv_pwrite_sync(bs->file, l1_table_offset, l1_table, l1_size2);
+
+        for (i = 0; i < l1_size; i++) {
+            be64_to_cpus(&l1_table[i]);
+        }
+    }
+    if (l1_allocated)
+        g_free(l1_table);
+    return ret;
+}
+
+
+
+
+/*********************************************************/
+/* refcount checking functions */
+
+
+
+/*
+ * Increases the refcount for a range of clusters in a given refcount table.
+ * This is used to construct a temporary refcount table out of L1 and L2 tables
+ * which can be compared the the refcount table saved in the image.
+ *
+ * Modifies the number of errors in res.
+ */
+static void inc_refcounts(BlockDriverState *bs,
+                          BdrvCheckResult *res,
+                          uint16_t *refcount_table,
+                          int refcount_table_size,
+                          int64_t offset, int64_t size)
+{
+    BDRVQcowState *s = bs->opaque;
+    int64_t start, last, cluster_offset;
+    int k;
+
+    if (size <= 0)
+        return;
+
+    start = offset & ~(s->cluster_size - 1);
+    last = (offset + size - 1) & ~(s->cluster_size - 1);
+    for(cluster_offset = start; cluster_offset <= last;
+        cluster_offset += s->cluster_size) {
+        k = cluster_offset >> s->cluster_bits;
+        if (k < 0) {
+            fprintf(stderr, "ERROR: invalid cluster offset=0x%" PRIx64 "\n",
+                cluster_offset);
+            res->corruptions++;
+        } else if (k >= refcount_table_size) {
+            fprintf(stderr, "Warning: cluster offset=0x%" PRIx64 " is after "
+                "the end of the image file, can't properly check refcounts.\n",
+                cluster_offset);
+            res->check_errors++;
+        } else {
+            if (++refcount_table[k] == 0) {
+                fprintf(stderr, "ERROR: overflow cluster offset=0x%" PRIx64
+                    "\n", cluster_offset);
+                res->corruptions++;
+            }
+        }
+    }
+}
+
+/* Flags for check_refcounts_l1() and check_refcounts_l2() */
+enum {
+    CHECK_OFLAG_COPIED = 0x1,   /* check QCOW_OFLAG_COPIED matches refcount */
+    CHECK_FRAG_INFO = 0x2,      /* update BlockFragInfo counters */
+};
+
+/*
+ * Increases the refcount in the given refcount table for the all clusters
+ * referenced in the L2 table. While doing so, performs some checks on L2
+ * entries.
+ *
+ * Returns the number of errors found by the checks or -errno if an internal
+ * error occurred.
+ */
+static int check_refcounts_l2(BlockDriverState *bs, BdrvCheckResult *res,
+    uint16_t *refcount_table, int refcount_table_size, int64_t l2_offset,
+    int flags)
+{
+    BDRVQcowState *s = bs->opaque;
+    uint64_t *l2_table, l2_entry;
+    uint64_t next_contiguous_offset = 0;
+    int i, l2_size, nb_csectors, refcount;
+
+    /* Read L2 table from disk */
+    l2_size = s->l2_size * sizeof(uint64_t);
+    l2_table = g_malloc(l2_size);
+
+    if (bdrv_pread(bs->file, l2_offset, l2_table, l2_size) != l2_size)
+        goto fail;
+
+    /* Do the actual checks */
+    for(i = 0; i < s->l2_size; i++) {
+        l2_entry = be64_to_cpu(l2_table[i]);
+
+        switch (qcow2_get_cluster_type(l2_entry)) {
+        case QCOW2_CLUSTER_COMPRESSED:
+            /* Compressed clusters don't have QCOW_OFLAG_COPIED */
+            if (l2_entry & QCOW_OFLAG_COPIED) {
+                fprintf(stderr, "ERROR: cluster %" PRId64 ": "
+                    "copied flag must never be set for compressed "
+                    "clusters\n", l2_entry >> s->cluster_bits);
+                l2_entry &= ~QCOW_OFLAG_COPIED;
+                res->corruptions++;
+            }
+
+            /* Mark cluster as used */
+            nb_csectors = ((l2_entry >> s->csize_shift) &
+                           s->csize_mask) + 1;
+            l2_entry &= s->cluster_offset_mask;
+            inc_refcounts(bs, res, refcount_table, refcount_table_size,
+                l2_entry & ~511, nb_csectors * 512);
+
+            if (flags & CHECK_FRAG_INFO) {
+                res->bfi.allocated_clusters++;
+                res->bfi.compressed_clusters++;
+
+                /* Compressed clusters are fragmented by nature.  Since they
+                 * take up sub-sector space but we only have sector granularity
+                 * I/O we need to re-read the same sectors even for adjacent
+                 * compressed clusters.
+                 */
+                res->bfi.fragmented_clusters++;
+            }
+            break;
+
+        case QCOW2_CLUSTER_ZERO:
+            if ((l2_entry & L2E_OFFSET_MASK) == 0) {
+                break;
+            }
+            /* fall through */
+
+        case QCOW2_CLUSTER_NORMAL:
+        {
+            /* QCOW_OFLAG_COPIED must be set iff refcount == 1 */
+            uint64_t offset = l2_entry & L2E_OFFSET_MASK;
+
+            if (flags & CHECK_OFLAG_COPIED) {
+                refcount = get_refcount(bs, offset >> s->cluster_bits);
+                if (refcount < 0) {
+                    fprintf(stderr, "Can't get refcount for offset %"
+                        PRIx64 ": %s\n", l2_entry, strerror(-refcount));
+                    goto fail;
+                }
+                if ((refcount == 1) != ((l2_entry & QCOW_OFLAG_COPIED) != 0)) {
+                    fprintf(stderr, "ERROR OFLAG_COPIED: offset=%"
+                        PRIx64 " refcount=%d\n", l2_entry, refcount);
+                    res->corruptions++;
+                }
+            }
+
+            if (flags & CHECK_FRAG_INFO) {
+                res->bfi.allocated_clusters++;
+                if (next_contiguous_offset &&
+                    offset != next_contiguous_offset) {
+                    res->bfi.fragmented_clusters++;
+                }
+                next_contiguous_offset = offset + s->cluster_size;
+            }
+
+            /* Mark cluster as used */
+            inc_refcounts(bs, res, refcount_table,refcount_table_size,
+                offset, s->cluster_size);
+
+            /* Correct offsets are cluster aligned */
+            if (offset & (s->cluster_size - 1)) {
+                fprintf(stderr, "ERROR offset=%" PRIx64 ": Cluster is not "
+                    "properly aligned; L2 entry corrupted.\n", offset);
+                res->corruptions++;
+            }
+            break;
+        }
+
+        case QCOW2_CLUSTER_UNALLOCATED:
+            break;
+
+        default:
+            abort();
+        }
+    }
+
+    g_free(l2_table);
+    return 0;
+
+fail:
+    fprintf(stderr, "ERROR: I/O error in check_refcounts_l2\n");
+    g_free(l2_table);
+    return -EIO;
+}
+
+/*
+ * Increases the refcount for the L1 table, its L2 tables and all referenced
+ * clusters in the given refcount table. While doing so, performs some checks
+ * on L1 and L2 entries.
+ *
+ * Returns the number of errors found by the checks or -errno if an internal
+ * error occurred.
+ */
+static int check_refcounts_l1(BlockDriverState *bs,
+                              BdrvCheckResult *res,
+                              uint16_t *refcount_table,
+                              int refcount_table_size,
+                              int64_t l1_table_offset, int l1_size,
+                              int flags)
+{
+    BDRVQcowState *s = bs->opaque;
+    uint64_t *l1_table, l2_offset, l1_size2;
+    int i, refcount, ret;
+
+    l1_size2 = l1_size * sizeof(uint64_t);
+
+    /* Mark L1 table as used */
+    inc_refcounts(bs, res, refcount_table, refcount_table_size,
+        l1_table_offset, l1_size2);
+
+    /* Read L1 table entries from disk */
+    if (l1_size2 == 0) {
+        l1_table = NULL;
+    } else {
+        l1_table = g_malloc(l1_size2);
+        if (bdrv_pread(bs->file, l1_table_offset,
+                       l1_table, l1_size2) != l1_size2)
+            goto fail;
+        for(i = 0;i < l1_size; i++)
+            be64_to_cpus(&l1_table[i]);
+    }
+
+    /* Do the actual checks */
+    for(i = 0; i < l1_size; i++) {
+        l2_offset = l1_table[i];
+        if (l2_offset) {
+            /* QCOW_OFLAG_COPIED must be set iff refcount == 1 */
+            if (flags & CHECK_OFLAG_COPIED) {
+                refcount = get_refcount(bs, (l2_offset & ~QCOW_OFLAG_COPIED)
+                    >> s->cluster_bits);
+                if (refcount < 0) {
+                    fprintf(stderr, "Can't get refcount for l2_offset %"
+                        PRIx64 ": %s\n", l2_offset, strerror(-refcount));
+                    goto fail;
+                }
+                if ((refcount == 1) != ((l2_offset & QCOW_OFLAG_COPIED) != 0)) {
+                    fprintf(stderr, "ERROR OFLAG_COPIED: l2_offset=%" PRIx64
+                        " refcount=%d\n", l2_offset, refcount);
+                    res->corruptions++;
+                }
+            }
+
+            /* Mark L2 table as used */
+            l2_offset &= L1E_OFFSET_MASK;
+            inc_refcounts(bs, res, refcount_table, refcount_table_size,
+                l2_offset, s->cluster_size);
+
+            /* L2 tables are cluster aligned */
+            if (l2_offset & (s->cluster_size - 1)) {
+                fprintf(stderr, "ERROR l2_offset=%" PRIx64 ": Table is not "
+                    "cluster aligned; L1 entry corrupted\n", l2_offset);
+                res->corruptions++;
+            }
+
+            /* Process and check L2 entries */
+            ret = check_refcounts_l2(bs, res, refcount_table,
+                                     refcount_table_size, l2_offset, flags);
+            if (ret < 0) {
+                goto fail;
+            }
+        }
+    }
+    g_free(l1_table);
+    return 0;
+
+fail:
+    fprintf(stderr, "ERROR: I/O error in check_refcounts_l1\n");
+    res->check_errors++;
+    g_free(l1_table);
+    return -EIO;
+}
+
+/*
+ * Checks an image for refcount consistency.
+ *
+ * Returns 0 if no errors are found, the number of errors in case the image is
+ * detected as corrupted, and -errno when an internal error occurred.
+ */
+int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
+                          BdrvCheckMode fix)
+{
+    BDRVQcowState *s = bs->opaque;
+    int64_t size, i, highest_cluster;
+    int nb_clusters, refcount1, refcount2;
+    QCowSnapshot *sn;
+    uint16_t *refcount_table;
+    int ret;
+
+    size = bdrv_getlength(bs->file);
+    nb_clusters = size_to_clusters(s, size);
+    refcount_table = g_malloc0(nb_clusters * sizeof(uint16_t));
+
+    res->bfi.total_clusters =
+        size_to_clusters(s, bs->total_sectors * BDRV_SECTOR_SIZE);
+
+    /* header */
+    inc_refcounts(bs, res, refcount_table, nb_clusters,
+        0, s->cluster_size);
+
+    /* current L1 table */
+    ret = check_refcounts_l1(bs, res, refcount_table, nb_clusters,
+                             s->l1_table_offset, s->l1_size,
+                             CHECK_OFLAG_COPIED | CHECK_FRAG_INFO);
+    if (ret < 0) {
+        goto fail;
+    }
+
+    /* snapshots */
+    for(i = 0; i < s->nb_snapshots; i++) {
+        sn = s->snapshots + i;
+        ret = check_refcounts_l1(bs, res, refcount_table, nb_clusters,
+            sn->l1_table_offset, sn->l1_size, 0);
+        if (ret < 0) {
+            goto fail;
+        }
+    }
+    inc_refcounts(bs, res, refcount_table, nb_clusters,
+        s->snapshots_offset, s->snapshots_size);
+
+    /* refcount data */
+    inc_refcounts(bs, res, refcount_table, nb_clusters,
+        s->refcount_table_offset,
+        s->refcount_table_size * sizeof(uint64_t));
+
+    for(i = 0; i < s->refcount_table_size; i++) {
+        uint64_t offset, cluster;
+        offset = s->refcount_table[i];
+        cluster = offset >> s->cluster_bits;
+
+        /* Refcount blocks are cluster aligned */
+        if (offset & (s->cluster_size - 1)) {
+            fprintf(stderr, "ERROR refcount block %" PRId64 " is not "
+                "cluster aligned; refcount table entry corrupted\n", i);
+            res->corruptions++;
+            continue;
+        }
+
+        if (cluster >= nb_clusters) {
+            fprintf(stderr, "ERROR refcount block %" PRId64
+                    " is outside image\n", i);
+            res->corruptions++;
+            continue;
+        }
+
+        if (offset != 0) {
+            inc_refcounts(bs, res, refcount_table, nb_clusters,
+                offset, s->cluster_size);
+            if (refcount_table[cluster] != 1) {
+                fprintf(stderr, "ERROR refcount block %" PRId64
+                    " refcount=%d\n",
+                    i, refcount_table[cluster]);
+                res->corruptions++;
+            }
+        }
+    }
+
+    /* compare ref counts */
+    for (i = 0, highest_cluster = 0; i < nb_clusters; i++) {
+        refcount1 = get_refcount(bs, i);
+        if (refcount1 < 0) {
+            fprintf(stderr, "Can't get refcount for cluster %" PRId64 ": %s\n",
+                i, strerror(-refcount1));
+            res->check_errors++;
+            continue;
+        }
+
+        refcount2 = refcount_table[i];
+
+        if (refcount1 > 0 || refcount2 > 0) {
+            highest_cluster = i;
+        }
+
+        if (refcount1 != refcount2) {
+
+            /* Check if we're allowed to fix the mismatch */
+            int *num_fixed = NULL;
+            if (refcount1 > refcount2 && (fix & BDRV_FIX_LEAKS)) {
+                num_fixed = &res->leaks_fixed;
+            } else if (refcount1 < refcount2 && (fix & BDRV_FIX_ERRORS)) {
+                num_fixed = &res->corruptions_fixed;
+            }
+
+            fprintf(stderr, "%s cluster %" PRId64 " refcount=%d reference=%d\n",
+                   num_fixed != NULL     ? "Repairing" :
+                   refcount1 < refcount2 ? "ERROR" :
+                                           "Leaked",
+                   i, refcount1, refcount2);
+
+            if (num_fixed) {
+                ret = update_refcount(bs, i << s->cluster_bits, 1,
+                                      refcount2 - refcount1,
+                                      QCOW2_DISCARD_ALWAYS);
+                if (ret >= 0) {
+                    (*num_fixed)++;
+                    continue;
+                }
+            }
+
+            /* And if we couldn't, print an error */
+            if (refcount1 < refcount2) {
+                res->corruptions++;
+            } else {
+                res->leaks++;
+            }
+        }
+    }
+
+    res->image_end_offset = (highest_cluster + 1) * s->cluster_size;
+    ret = 0;
+
+fail:
+    g_free(refcount_table);
+
+    return ret;
+}
+
diff --git a/contrib/qemu/block/qcow2-snapshot.c b/contrib/qemu/block/qcow2-snapshot.c
new file mode 100644
index 000000000..0caac9055
--- /dev/null
+++ b/contrib/qemu/block/qcow2-snapshot.c
@@ -0,0 +1,660 @@
+/*
+ * Block driver for the QCOW version 2 format
+ *
+ * Copyright (c) 2004-2006 Fabrice Bellard
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#include "qemu-common.h"
+#include "block/block_int.h"
+#include "block/qcow2.h"
+
+typedef struct QEMU_PACKED QCowSnapshotHeader {
+    /* header is 8 byte aligned */
+    uint64_t l1_table_offset;
+
+    uint32_t l1_size;
+    uint16_t id_str_size;
+    uint16_t name_size;
+
+    uint32_t date_sec;
+    uint32_t date_nsec;
+
+    uint64_t vm_clock_nsec;
+
+    uint32_t vm_state_size;
+    uint32_t extra_data_size; /* for extension */
+    /* extra data follows */
+    /* id_str follows */
+    /* name follows  */
+} QCowSnapshotHeader;
+
+typedef struct QEMU_PACKED QCowSnapshotExtraData {
+    uint64_t vm_state_size_large;
+    uint64_t disk_size;
+} QCowSnapshotExtraData;
+
+void qcow2_free_snapshots(BlockDriverState *bs)
+{
+    BDRVQcowState *s = bs->opaque;
+    int i;
+
+    for(i = 0; i < s->nb_snapshots; i++) {
+        g_free(s->snapshots[i].name);
+        g_free(s->snapshots[i].id_str);
+    }
+    g_free(s->snapshots);
+    s->snapshots = NULL;
+    s->nb_snapshots = 0;
+}
+
+int qcow2_read_snapshots(BlockDriverState *bs)
+{
+    BDRVQcowState *s = bs->opaque;
+    QCowSnapshotHeader h;
+    QCowSnapshotExtraData extra;
+    QCowSnapshot *sn;
+    int i, id_str_size, name_size;
+    int64_t offset;
+    uint32_t extra_data_size;
+    int ret;
+
+    if (!s->nb_snapshots) {
+        s->snapshots = NULL;
+        s->snapshots_size = 0;
+        return 0;
+    }
+
+    offset = s->snapshots_offset;
+    s->snapshots = g_malloc0(s->nb_snapshots * sizeof(QCowSnapshot));
+
+    for(i = 0; i < s->nb_snapshots; i++) {
+        /* Read statically sized part of the snapshot header */
+        offset = align_offset(offset, 8);
+        ret = bdrv_pread(bs->file, offset, &h, sizeof(h));
+        if (ret < 0) {
+            goto fail;
+        }
+
+        offset += sizeof(h);
+        sn = s->snapshots + i;
+        sn->l1_table_offset = be64_to_cpu(h.l1_table_offset);
+        sn->l1_size = be32_to_cpu(h.l1_size);
+        sn->vm_state_size = be32_to_cpu(h.vm_state_size);
+        sn->date_sec = be32_to_cpu(h.date_sec);
+        sn->date_nsec = be32_to_cpu(h.date_nsec);
+        sn->vm_clock_nsec = be64_to_cpu(h.vm_clock_nsec);
+        extra_data_size = be32_to_cpu(h.extra_data_size);
+
+        id_str_size = be16_to_cpu(h.id_str_size);
+        name_size = be16_to_cpu(h.name_size);
+
+        /* Read extra data */
+        ret = bdrv_pread(bs->file, offset, &extra,
+                         MIN(sizeof(extra), extra_data_size));
+        if (ret < 0) {
+            goto fail;
+        }
+        offset += extra_data_size;
+
+        if (extra_data_size >= 8) {
+            sn->vm_state_size = be64_to_cpu(extra.vm_state_size_large);
+        }
+
+        if (extra_data_size >= 16) {
+            sn->disk_size = be64_to_cpu(extra.disk_size);
+        } else {
+            sn->disk_size = bs->total_sectors * BDRV_SECTOR_SIZE;
+        }
+
+        /* Read snapshot ID */
+        sn->id_str = g_malloc(id_str_size + 1);
+        ret = bdrv_pread(bs->file, offset, sn->id_str, id_str_size);
+        if (ret < 0) {
+            goto fail;
+        }
+        offset += id_str_size;
+        sn->id_str[id_str_size] = '\0';
+
+        /* Read snapshot name */
+        sn->name = g_malloc(name_size + 1);
+        ret = bdrv_pread(bs->file, offset, sn->name, name_size);
+        if (ret < 0) {
+            goto fail;
+        }
+        offset += name_size;
+        sn->name[name_size] = '\0';
+    }
+
+    s->snapshots_size = offset - s->snapshots_offset;
+    return 0;
+
+fail:
+    qcow2_free_snapshots(bs);
+    return ret;
+}
+
+/* add at the end of the file a new list of snapshots */
+static int qcow2_write_snapshots(BlockDriverState *bs)
+{
+    BDRVQcowState *s = bs->opaque;
+    QCowSnapshot *sn;
+    QCowSnapshotHeader h;
+    QCowSnapshotExtraData extra;
+    int i, name_size, id_str_size, snapshots_size;
+    struct {
+        uint32_t nb_snapshots;
+        uint64_t snapshots_offset;
+    } QEMU_PACKED header_data;
+    int64_t offset, snapshots_offset;
+    int ret;
+
+    /* compute the size of the snapshots */
+    offset = 0;
+    for(i = 0; i < s->nb_snapshots; i++) {
+        sn = s->snapshots + i;
+        offset = align_offset(offset, 8);
+        offset += sizeof(h);
+        offset += sizeof(extra);
+        offset += strlen(sn->id_str);
+        offset += strlen(sn->name);
+    }
+    snapshots_size = offset;
+
+    /* Allocate space for the new snapshot list */
+    snapshots_offset = qcow2_alloc_clusters(bs, snapshots_size);
+    offset = snapshots_offset;
+    if (offset < 0) {
+        return offset;
+    }
+    ret = bdrv_flush(bs);
+    if (ret < 0) {
+        return ret;
+    }
+
+    /* Write all snapshots to the new list */
+    for(i = 0; i < s->nb_snapshots; i++) {
+        sn = s->snapshots + i;
+        memset(&h, 0, sizeof(h));
+        h.l1_table_offset = cpu_to_be64(sn->l1_table_offset);
+        h.l1_size = cpu_to_be32(sn->l1_size);
+        /* If it doesn't fit in 32 bit, older implementations should treat it
+         * as a disk-only snapshot rather than truncate the VM state */
+        if (sn->vm_state_size <= 0xffffffff) {
+            h.vm_state_size = cpu_to_be32(sn->vm_state_size);
+        }
+        h.date_sec = cpu_to_be32(sn->date_sec);
+        h.date_nsec = cpu_to_be32(sn->date_nsec);
+        h.vm_clock_nsec = cpu_to_be64(sn->vm_clock_nsec);
+        h.extra_data_size = cpu_to_be32(sizeof(extra));
+
+        memset(&extra, 0, sizeof(extra));
+        extra.vm_state_size_large = cpu_to_be64(sn->vm_state_size);
+        extra.disk_size = cpu_to_be64(sn->disk_size);
+
+        id_str_size = strlen(sn->id_str);
+        name_size = strlen(sn->name);
+        h.id_str_size = cpu_to_be16(id_str_size);
+        h.name_size = cpu_to_be16(name_size);
+        offset = align_offset(offset, 8);
+
+        ret = bdrv_pwrite(bs->file, offset, &h, sizeof(h));
+        if (ret < 0) {
+            goto fail;
+        }
+        offset += sizeof(h);
+
+        ret = bdrv_pwrite(bs->file, offset, &extra, sizeof(extra));
+        if (ret < 0) {
+            goto fail;
+        }
+        offset += sizeof(extra);
+
+        ret = bdrv_pwrite(bs->file, offset, sn->id_str, id_str_size);
+        if (ret < 0) {
+            goto fail;
+        }
+        offset += id_str_size;
+
+        ret = bdrv_pwrite(bs->file, offset, sn->name, name_size);
+        if (ret < 0) {
+            goto fail;
+        }
+        offset += name_size;
+    }
+
+    /*
+     * Update the header to point to the new snapshot table. This requires the
+     * new table and its refcounts to be stable on disk.
+     */
+    ret = bdrv_flush(bs);
+    if (ret < 0) {
+        goto fail;
+    }
+
+    QEMU_BUILD_BUG_ON(offsetof(QCowHeader, snapshots_offset) !=
+        offsetof(QCowHeader, nb_snapshots) + sizeof(header_data.nb_snapshots));
+
+    header_data.nb_snapshots        = cpu_to_be32(s->nb_snapshots);
+    header_data.snapshots_offset    = cpu_to_be64(snapshots_offset);
+
+    ret = bdrv_pwrite_sync(bs->file, offsetof(QCowHeader, nb_snapshots),
+                           &header_data, sizeof(header_data));
+    if (ret < 0) {
+        goto fail;
+    }
+
+    /* free the old snapshot table */
+    qcow2_free_clusters(bs, s->snapshots_offset, s->snapshots_size,
+                        QCOW2_DISCARD_SNAPSHOT);
+    s->snapshots_offset = snapshots_offset;
+    s->snapshots_size = snapshots_size;
+    return 0;
+
+fail:
+    return ret;
+}
+
+static void find_new_snapshot_id(BlockDriverState *bs,
+                                 char *id_str, int id_str_size)
+{
+    BDRVQcowState *s = bs->opaque;
+    QCowSnapshot *sn;
+    int i, id, id_max = 0;
+
+    for(i = 0; i < s->nb_snapshots; i++) {
+        sn = s->snapshots + i;
+        id = strtoul(sn->id_str, NULL, 10);
+        if (id > id_max)
+            id_max = id;
+    }
+    snprintf(id_str, id_str_size, "%d", id_max + 1);
+}
+
+static int find_snapshot_by_id(BlockDriverState *bs, const char *id_str)
+{
+    BDRVQcowState *s = bs->opaque;
+    int i;
+
+    for(i = 0; i < s->nb_snapshots; i++) {
+        if (!strcmp(s->snapshots[i].id_str, id_str))
+            return i;
+    }
+    return -1;
+}
+
+static int find_snapshot_by_id_or_name(BlockDriverState *bs, const char *name)
+{
+    BDRVQcowState *s = bs->opaque;
+    int i, ret;
+
+    ret = find_snapshot_by_id(bs, name);
+    if (ret >= 0)
+        return ret;
+    for(i = 0; i < s->nb_snapshots; i++) {
+        if (!strcmp(s->snapshots[i].name, name))
+            return i;
+    }
+    return -1;
+}
+
+/* if no id is provided, a new one is constructed */
+int qcow2_snapshot_create(BlockDriverState *bs, QEMUSnapshotInfo *sn_info)
+{
+    BDRVQcowState *s = bs->opaque;
+    QCowSnapshot *new_snapshot_list = NULL;
+    QCowSnapshot *old_snapshot_list = NULL;
+    QCowSnapshot sn1, *sn = &sn1;
+    int i, ret;
+    uint64_t *l1_table = NULL;
+    int64_t l1_table_offset;
+
+    memset(sn, 0, sizeof(*sn));
+
+    /* Generate an ID if it wasn't passed */
+    if (sn_info->id_str[0] == '\0') {
+        find_new_snapshot_id(bs, sn_info->id_str, sizeof(sn_info->id_str));
+    }
+
+    /* Check that the ID is unique */
+    if (find_snapshot_by_id(bs, sn_info->id_str) >= 0) {
+        return -EEXIST;
+    }
+
+    /* Populate sn with passed data */
+    sn->id_str = g_strdup(sn_info->id_str);
+    sn->name = g_strdup(sn_info->name);
+
+    sn->disk_size = bs->total_sectors * BDRV_SECTOR_SIZE;
+    sn->vm_state_size = sn_info->vm_state_size;
+    sn->date_sec = sn_info->date_sec;
+    sn->date_nsec = sn_info->date_nsec;
+    sn->vm_clock_nsec = sn_info->vm_clock_nsec;
+
+    /* Allocate the L1 table of the snapshot and copy the current one there. */
+    l1_table_offset = qcow2_alloc_clusters(bs, s->l1_size * sizeof(uint64_t));
+    if (l1_table_offset < 0) {
+        ret = l1_table_offset;
+        goto fail;
+    }
+
+    sn->l1_table_offset = l1_table_offset;
+    sn->l1_size = s->l1_size;
+
+    l1_table = g_malloc(s->l1_size * sizeof(uint64_t));
+    for(i = 0; i < s->l1_size; i++) {
+        l1_table[i] = cpu_to_be64(s->l1_table[i]);
+    }
+
+    ret = bdrv_pwrite(bs->file, sn->l1_table_offset, l1_table,
+                      s->l1_size * sizeof(uint64_t));
+    if (ret < 0) {
+        goto fail;
+    }
+
+    g_free(l1_table);
+    l1_table = NULL;
+
+    /*
+     * Increase the refcounts of all clusters and make sure everything is
+     * stable on disk before updating the snapshot table to contain a pointer
+     * to the new L1 table.
+     */
+    ret = qcow2_update_snapshot_refcount(bs, s->l1_table_offset, s->l1_size, 1);
+    if (ret < 0) {
+        goto fail;
+    }
+
+    /* Append the new snapshot to the snapshot list */
+    new_snapshot_list = g_malloc((s->nb_snapshots + 1) * sizeof(QCowSnapshot));
+    if (s->snapshots) {
+        memcpy(new_snapshot_list, s->snapshots,
+               s->nb_snapshots * sizeof(QCowSnapshot));
+        old_snapshot_list = s->snapshots;
+    }
+    s->snapshots = new_snapshot_list;
+    s->snapshots[s->nb_snapshots++] = *sn;
+
+    ret = qcow2_write_snapshots(bs);
+    if (ret < 0) {
+        g_free(s->snapshots);
+        s->snapshots = old_snapshot_list;
+        goto fail;
+    }
+
+    g_free(old_snapshot_list);
+
+#ifdef DEBUG_ALLOC
+    {
+      BdrvCheckResult result = {0};
+      qcow2_check_refcounts(bs, &result, 0);
+    }
+#endif
+    return 0;
+
+fail:
+    g_free(sn->id_str);
+    g_free(sn->name);
+    g_free(l1_table);
+
+    return ret;
+}
+
+/* copy the snapshot 'snapshot_name' into the current disk image */
+int qcow2_snapshot_goto(BlockDriverState *bs, const char *snapshot_id)
+{
+    BDRVQcowState *s = bs->opaque;
+    QCowSnapshot *sn;
+    int i, snapshot_index;
+    int cur_l1_bytes, sn_l1_bytes;
+    int ret;
+    uint64_t *sn_l1_table = NULL;
+
+    /* Search the snapshot */
+    snapshot_index = find_snapshot_by_id_or_name(bs, snapshot_id);
+    if (snapshot_index < 0) {
+        return -ENOENT;
+    }
+    sn = &s->snapshots[snapshot_index];
+
+    if (sn->disk_size != bs->total_sectors * BDRV_SECTOR_SIZE) {
+        error_report("qcow2: Loading snapshots with different disk "
+            "size is not implemented");
+        ret = -ENOTSUP;
+        goto fail;
+    }
+
+    /*
+     * Make sure that the current L1 table is big enough to contain the whole
+     * L1 table of the snapshot. If the snapshot L1 table is smaller, the
+     * current one must be padded with zeros.
+     */
+    ret = qcow2_grow_l1_table(bs, sn->l1_size, true);
+    if (ret < 0) {
+        goto fail;
+    }
+
+    cur_l1_bytes = s->l1_size * sizeof(uint64_t);
+    sn_l1_bytes = sn->l1_size * sizeof(uint64_t);
+
+    /*
+     * Copy the snapshot L1 table to the current L1 table.
+     *
+     * Before overwriting the old current L1 table on disk, make sure to
+     * increase all refcounts for the clusters referenced by the new one.
+     * Decrease the refcount referenced by the old one only when the L1
+     * table is overwritten.
+     */
+    sn_l1_table = g_malloc0(cur_l1_bytes);
+
+    ret = bdrv_pread(bs->file, sn->l1_table_offset, sn_l1_table, sn_l1_bytes);
+    if (ret < 0) {
+        goto fail;
+    }
+
+    ret = qcow2_update_snapshot_refcount(bs, sn->l1_table_offset,
+                                         sn->l1_size, 1);
+    if (ret < 0) {
+        goto fail;
+    }
+
+    ret = bdrv_pwrite_sync(bs->file, s->l1_table_offset, sn_l1_table,
+                           cur_l1_bytes);
+    if (ret < 0) {
+        goto fail;
+    }
+
+    /*
+     * Decrease refcount of clusters of current L1 table.
+     *
+     * At this point, the in-memory s->l1_table points to the old L1 table,
+     * whereas on disk we already have the new one.
+     *
+     * qcow2_update_snapshot_refcount special cases the current L1 table to use
+     * the in-memory data instead of really using the offset to load a new one,
+     * which is why this works.
+     */
+    ret = qcow2_update_snapshot_refcount(bs, s->l1_table_offset,
+                                         s->l1_size, -1);
+
+    /*
+     * Now update the in-memory L1 table to be in sync with the on-disk one. We
+     * need to do this even if updating refcounts failed.
+     */
+    for(i = 0;i < s->l1_size; i++) {
+        s->l1_table[i] = be64_to_cpu(sn_l1_table[i]);
+    }
+
+    if (ret < 0) {
+        goto fail;
+    }
+
+    g_free(sn_l1_table);
+    sn_l1_table = NULL;
+
+    /*
+     * Update QCOW_OFLAG_COPIED in the active L1 table (it may have changed
+     * when we decreased the refcount of the old snapshot.
+     */
+    ret = qcow2_update_snapshot_refcount(bs, s->l1_table_offset, s->l1_size, 0);
+    if (ret < 0) {
+        goto fail;
+    }
+
+#ifdef DEBUG_ALLOC
+    {
+        BdrvCheckResult result = {0};
+        qcow2_check_refcounts(bs, &result, 0);
+    }
+#endif
+    return 0;
+
+fail:
+    g_free(sn_l1_table);
+    return ret;
+}
+
+int qcow2_snapshot_delete(BlockDriverState *bs, const char *snapshot_id)
+{
+    BDRVQcowState *s = bs->opaque;
+    QCowSnapshot sn;
+    int snapshot_index, ret;
+
+    /* Search the snapshot */
+    snapshot_index = find_snapshot_by_id_or_name(bs, snapshot_id);
+    if (snapshot_index < 0) {
+        return -ENOENT;
+    }
+    sn = s->snapshots[snapshot_index];
+
+    /* Remove it from the snapshot list */
+    memmove(s->snapshots + snapshot_index,
+            s->snapshots + snapshot_index + 1,
+            (s->nb_snapshots - snapshot_index - 1) * sizeof(sn));
+    s->nb_snapshots--;
+    ret = qcow2_write_snapshots(bs);
+    if (ret < 0) {
+        return ret;
+    }
+
+    /*
+     * The snapshot is now unused, clean up. If we fail after this point, we
+     * won't recover but just leak clusters.
+     */
+    g_free(sn.id_str);
+    g_free(sn.name);
+
+    /*
+     * Now decrease the refcounts of clusters referenced by the snapshot and
+     * free the L1 table.
+     */
+    ret = qcow2_update_snapshot_refcount(bs, sn.l1_table_offset,
+                                         sn.l1_size, -1);
+    if (ret < 0) {
+        return ret;
+    }
+    qcow2_free_clusters(bs, sn.l1_table_offset, sn.l1_size * sizeof(uint64_t),
+                        QCOW2_DISCARD_SNAPSHOT);
+
+    /* must update the copied flag on the current cluster offsets */
+    ret = qcow2_update_snapshot_refcount(bs, s->l1_table_offset, s->l1_size, 0);
+    if (ret < 0) {
+        return ret;
+    }
+
+#ifdef DEBUG_ALLOC
+    {
+        BdrvCheckResult result = {0};
+        qcow2_check_refcounts(bs, &result, 0);
+    }
+#endif
+    return 0;
+}
+
+int qcow2_snapshot_list(BlockDriverState *bs, QEMUSnapshotInfo **psn_tab)
+{
+    BDRVQcowState *s = bs->opaque;
+    QEMUSnapshotInfo *sn_tab, *sn_info;
+    QCowSnapshot *sn;
+    int i;
+
+    if (!s->nb_snapshots) {
+        *psn_tab = NULL;
+        return s->nb_snapshots;
+    }
+
+    sn_tab = g_malloc0(s->nb_snapshots * sizeof(QEMUSnapshotInfo));
+    for(i = 0; i < s->nb_snapshots; i++) {
+        sn_info = sn_tab + i;
+        sn = s->snapshots + i;
+        pstrcpy(sn_info->id_str, sizeof(sn_info->id_str),
+                sn->id_str);
+        pstrcpy(sn_info->name, sizeof(sn_info->name),
+                sn->name);
+        sn_info->vm_state_size = sn->vm_state_size;
+        sn_info->date_sec = sn->date_sec;
+        sn_info->date_nsec = sn->date_nsec;
+        sn_info->vm_clock_nsec = sn->vm_clock_nsec;
+    }
+    *psn_tab = sn_tab;
+    return s->nb_snapshots;
+}
+
+int qcow2_snapshot_load_tmp(BlockDriverState *bs, const char *snapshot_name)
+{
+    int i, snapshot_index;
+    BDRVQcowState *s = bs->opaque;
+    QCowSnapshot *sn;
+    uint64_t *new_l1_table;
+    int new_l1_bytes;
+    int ret;
+
+    assert(bs->read_only);
+
+    /* Search the snapshot */
+    snapshot_index = find_snapshot_by_id_or_name(bs, snapshot_name);
+    if (snapshot_index < 0) {
+        return -ENOENT;
+    }
+    sn = &s->snapshots[snapshot_index];
+
+    /* Allocate and read in the snapshot's L1 table */
+    new_l1_bytes = s->l1_size * sizeof(uint64_t);
+    new_l1_table = g_malloc0(align_offset(new_l1_bytes, 512));
+
+    ret = bdrv_pread(bs->file, sn->l1_table_offset, new_l1_table, new_l1_bytes);
+    if (ret < 0) {
+        g_free(new_l1_table);
+        return ret;
+    }
+
+    /* Switch the L1 table */
+    g_free(s->l1_table);
+
+    s->l1_size = sn->l1_size;
+    s->l1_table_offset = sn->l1_table_offset;
+    s->l1_table = new_l1_table;
+
+    for(i = 0;i < s->l1_size; i++) {
+        be64_to_cpus(&s->l1_table[i]);
+    }
+
+    return 0;
+}
diff --git a/contrib/qemu/block/qcow2.c b/contrib/qemu/block/qcow2.c
new file mode 100644
index 000000000..0eceefe2c
--- /dev/null
+++ b/contrib/qemu/block/qcow2.c
@@ -0,0 +1,1825 @@
+/*
+ * Block driver for the QCOW version 2 format
+ *
+ * Copyright (c) 2004-2006 Fabrice Bellard
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+#include "qemu-common.h"
+#include "block/block_int.h"
+#include "qemu/module.h"
+#include <zlib.h>
+#include "qemu/aes.h"
+#include "block/qcow2.h"
+#include "qemu/error-report.h"
+#include "qapi/qmp/qerror.h"
+#include "qapi/qmp/qbool.h"
+#include "trace.h"
+
+/*
+  Differences with QCOW:
+
+  - Support for multiple incremental snapshots.
+  - Memory management by reference counts.
+  - Clusters which have a reference count of one have the bit
+    QCOW_OFLAG_COPIED to optimize write performance.
+  - Size of compressed clusters is stored in sectors to reduce bit usage
+    in the cluster offsets.
+  - Support for storing additional data (such as the VM state) in the
+    snapshots.
+  - If a backing store is used, the cluster size is not constrained
+    (could be backported to QCOW).
+  - L2 tables have always a size of one cluster.
+*/
+
+
+typedef struct {
+    uint32_t magic;
+    uint32_t len;
+} QCowExtension;
+
+#define  QCOW2_EXT_MAGIC_END 0
+#define  QCOW2_EXT_MAGIC_BACKING_FORMAT 0xE2792ACA
+#define  QCOW2_EXT_MAGIC_FEATURE_TABLE 0x6803f857
+
+static int qcow2_probe(const uint8_t *buf, int buf_size, const char *filename)
+{
+    const QCowHeader *cow_header = (const void *)buf;
+
+    if (buf_size >= sizeof(QCowHeader) &&
+        be32_to_cpu(cow_header->magic) == QCOW_MAGIC &&
+        be32_to_cpu(cow_header->version) >= 2)
+        return 100;
+    else
+        return 0;
+}
+
+
+/* 
+ * read qcow2 extension and fill bs
+ * start reading from start_offset
+ * finish reading upon magic of value 0 or when end_offset reached
+ * unknown magic is skipped (future extension this version knows nothing about)
+ * return 0 upon success, non-0 otherwise
+ */
+static int qcow2_read_extensions(BlockDriverState *bs, uint64_t start_offset,
+                                 uint64_t end_offset, void **p_feature_table)
+{
+    BDRVQcowState *s = bs->opaque;
+    QCowExtension ext;
+    uint64_t offset;
+    int ret;
+
+#ifdef DEBUG_EXT
+    printf("qcow2_read_extensions: start=%ld end=%ld\n", start_offset, end_offset);
+#endif
+    offset = start_offset;
+    while (offset < end_offset) {
+
+#ifdef DEBUG_EXT
+        /* Sanity check */
+        if (offset > s->cluster_size)
+            printf("qcow2_read_extension: suspicious offset %lu\n", offset);
+
+        printf("attempting to read extended header in offset %lu\n", offset);
+#endif
+
+        if (bdrv_pread(bs->file, offset, &ext, sizeof(ext)) != sizeof(ext)) {
+            fprintf(stderr, "qcow2_read_extension: ERROR: "
+                    "pread fail from offset %" PRIu64 "\n",
+                    offset);
+            return 1;
+        }
+        be32_to_cpus(&ext.magic);
+        be32_to_cpus(&ext.len);
+        offset += sizeof(ext);
+#ifdef DEBUG_EXT
+        printf("ext.magic = 0x%x\n", ext.magic);
+#endif
+        if (ext.len > end_offset - offset) {
+            error_report("Header extension too large");
+            return -EINVAL;
+        }
+
+        switch (ext.magic) {
+        case QCOW2_EXT_MAGIC_END:
+            return 0;
+
+        case QCOW2_EXT_MAGIC_BACKING_FORMAT:
+            if (ext.len >= sizeof(bs->backing_format)) {
+                fprintf(stderr, "ERROR: ext_backing_format: len=%u too large"
+                        " (>=%zu)\n",
+                        ext.len, sizeof(bs->backing_format));
+                return 2;
+            }
+            if (bdrv_pread(bs->file, offset , bs->backing_format,
+                           ext.len) != ext.len)
+                return 3;
+            bs->backing_format[ext.len] = '\0';
+#ifdef DEBUG_EXT
+            printf("Qcow2: Got format extension %s\n", bs->backing_format);
+#endif
+            break;
+
+        case QCOW2_EXT_MAGIC_FEATURE_TABLE:
+            if (p_feature_table != NULL) {
+                void* feature_table = g_malloc0(ext.len + 2 * sizeof(Qcow2Feature));
+                ret = bdrv_pread(bs->file, offset , feature_table, ext.len);
+                if (ret < 0) {
+                    return ret;
+                }
+
+                *p_feature_table = feature_table;
+            }
+            break;
+
+        default:
+            /* unknown magic - save it in case we need to rewrite the header */
+            {
+                Qcow2UnknownHeaderExtension *uext;
+
+                uext = g_malloc0(sizeof(*uext)  + ext.len);
+                uext->magic = ext.magic;
+                uext->len = ext.len;
+                QLIST_INSERT_HEAD(&s->unknown_header_ext, uext, next);
+
+                ret = bdrv_pread(bs->file, offset , uext->data, uext->len);
+                if (ret < 0) {
+                    return ret;
+                }
+            }
+            break;
+        }
+
+        offset += ((ext.len + 7) & ~7);
+    }
+
+    return 0;
+}
+
+static void cleanup_unknown_header_ext(BlockDriverState *bs)
+{
+    BDRVQcowState *s = bs->opaque;
+    Qcow2UnknownHeaderExtension *uext, *next;
+
+    QLIST_FOREACH_SAFE(uext, &s->unknown_header_ext, next, next) {
+        QLIST_REMOVE(uext, next);
+        g_free(uext);
+    }
+}
+
+static void GCC_FMT_ATTR(2, 3) report_unsupported(BlockDriverState *bs,
+    const char *fmt, ...)
+{
+    char msg[64];
+    va_list ap;
+
+    va_start(ap, fmt);
+    vsnprintf(msg, sizeof(msg), fmt, ap);
+    va_end(ap);
+
+    qerror_report(QERR_UNKNOWN_BLOCK_FORMAT_FEATURE,
+        bs->device_name, "qcow2", msg);
+}
+
+static void report_unsupported_feature(BlockDriverState *bs,
+    Qcow2Feature *table, uint64_t mask)
+{
+    while (table && table->name[0] != '\0') {
+        if (table->type == QCOW2_FEAT_TYPE_INCOMPATIBLE) {
+            if (mask & (1 << table->bit)) {
+                report_unsupported(bs, "%.46s",table->name);
+                mask &= ~(1 << table->bit);
+            }
+        }
+        table++;
+    }
+
+    if (mask) {
+        report_unsupported(bs, "Unknown incompatible feature: %" PRIx64, mask);
+    }
+}
+
+/*
+ * Sets the dirty bit and flushes afterwards if necessary.
+ *
+ * The incompatible_features bit is only set if the image file header was
+ * updated successfully.  Therefore it is not required to check the return
+ * value of this function.
+ */
+int qcow2_mark_dirty(BlockDriverState *bs)
+{
+    BDRVQcowState *s = bs->opaque;
+    uint64_t val;
+    int ret;
+
+    assert(s->qcow_version >= 3);
+
+    if (s->incompatible_features & QCOW2_INCOMPAT_DIRTY) {
+        return 0; /* already dirty */
+    }
+
+    val = cpu_to_be64(s->incompatible_features | QCOW2_INCOMPAT_DIRTY);
+    ret = bdrv_pwrite(bs->file, offsetof(QCowHeader, incompatible_features),
+                      &val, sizeof(val));
+    if (ret < 0) {
+        return ret;
+    }
+    ret = bdrv_flush(bs->file);
+    if (ret < 0) {
+        return ret;
+    }
+
+    /* Only treat image as dirty if the header was updated successfully */
+    s->incompatible_features |= QCOW2_INCOMPAT_DIRTY;
+    return 0;
+}
+
+/*
+ * Clears the dirty bit and flushes before if necessary.  Only call this
+ * function when there are no pending requests, it does not guard against
+ * concurrent requests dirtying the image.
+ */
+static int qcow2_mark_clean(BlockDriverState *bs)
+{
+    BDRVQcowState *s = bs->opaque;
+
+    if (s->incompatible_features & QCOW2_INCOMPAT_DIRTY) {
+        int ret = bdrv_flush(bs);
+        if (ret < 0) {
+            return ret;
+        }
+
+        s->incompatible_features &= ~QCOW2_INCOMPAT_DIRTY;
+        return qcow2_update_header(bs);
+    }
+    return 0;
+}
+
+static int qcow2_check(BlockDriverState *bs, BdrvCheckResult *result,
+                       BdrvCheckMode fix)
+{
+    int ret = qcow2_check_refcounts(bs, result, fix);
+    if (ret < 0) {
+        return ret;
+    }
+
+    if (fix && result->check_errors == 0 && result->corruptions == 0) {
+        return qcow2_mark_clean(bs);
+    }
+    return ret;
+}
+
+static QemuOptsList qcow2_runtime_opts = {
+    .name = "qcow2",
+    .head = QTAILQ_HEAD_INITIALIZER(qcow2_runtime_opts.head),
+    .desc = {
+        {
+            .name = "lazy_refcounts",
+            .type = QEMU_OPT_BOOL,
+            .help = "Postpone refcount updates",
+        },
+        {
+            .name = QCOW2_OPT_DISCARD_REQUEST,
+            .type = QEMU_OPT_BOOL,
+            .help = "Pass guest discard requests to the layer below",
+        },
+        {
+            .name = QCOW2_OPT_DISCARD_SNAPSHOT,
+            .type = QEMU_OPT_BOOL,
+            .help = "Generate discard requests when snapshot related space "
+                    "is freed",
+        },
+        {
+            .name = QCOW2_OPT_DISCARD_OTHER,
+            .type = QEMU_OPT_BOOL,
+            .help = "Generate discard requests when other clusters are freed",
+        },
+        { /* end of list */ }
+    },
+};
+
+static int qcow2_open(BlockDriverState *bs, QDict *options, int flags)
+{
+    BDRVQcowState *s = bs->opaque;
+    int len, i, ret = 0;
+    QCowHeader header;
+    QemuOpts *opts;
+    Error *local_err = NULL;
+    uint64_t ext_end;
+    uint64_t l1_vm_state_index;
+
+    ret = bdrv_pread(bs->file, 0, &header, sizeof(header));
+    if (ret < 0) {
+        goto fail;
+    }
+    be32_to_cpus(&header.magic);
+    be32_to_cpus(&header.version);
+    be64_to_cpus(&header.backing_file_offset);
+    be32_to_cpus(&header.backing_file_size);
+    be64_to_cpus(&header.size);
+    be32_to_cpus(&header.cluster_bits);
+    be32_to_cpus(&header.crypt_method);
+    be64_to_cpus(&header.l1_table_offset);
+    be32_to_cpus(&header.l1_size);
+    be64_to_cpus(&header.refcount_table_offset);
+    be32_to_cpus(&header.refcount_table_clusters);
+    be64_to_cpus(&header.snapshots_offset);
+    be32_to_cpus(&header.nb_snapshots);
+
+    if (header.magic != QCOW_MAGIC) {
+        ret = -EMEDIUMTYPE;
+        goto fail;
+    }
+    if (header.version < 2 || header.version > 3) {
+        report_unsupported(bs, "QCOW version %d", header.version);
+        ret = -ENOTSUP;
+        goto fail;
+    }
+
+    s->qcow_version = header.version;
+
+    /* Initialise version 3 header fields */
+    if (header.version == 2) {
+        header.incompatible_features    = 0;
+        header.compatible_features      = 0;
+        header.autoclear_features       = 0;
+        header.refcount_order           = 4;
+        header.header_length            = 72;
+    } else {
+        be64_to_cpus(&header.incompatible_features);
+        be64_to_cpus(&header.compatible_features);
+        be64_to_cpus(&header.autoclear_features);
+        be32_to_cpus(&header.refcount_order);
+        be32_to_cpus(&header.header_length);
+    }
+
+    if (header.header_length > sizeof(header)) {
+        s->unknown_header_fields_size = header.header_length - sizeof(header);
+        s->unknown_header_fields = g_malloc(s->unknown_header_fields_size);
+        ret = bdrv_pread(bs->file, sizeof(header), s->unknown_header_fields,
+                         s->unknown_header_fields_size);
+        if (ret < 0) {
+            goto fail;
+        }
+    }
+
+    if (header.backing_file_offset) {
+        ext_end = header.backing_file_offset;
+    } else {
+        ext_end = 1 << header.cluster_bits;
+    }
+
+    /* Handle feature bits */
+    s->incompatible_features    = header.incompatible_features;
+    s->compatible_features      = header.compatible_features;
+    s->autoclear_features       = header.autoclear_features;
+
+    if (s->incompatible_features & ~QCOW2_INCOMPAT_MASK) {
+        void *feature_table = NULL;
+        qcow2_read_extensions(bs, header.header_length, ext_end,
+                              &feature_table);
+        report_unsupported_feature(bs, feature_table,
+                                   s->incompatible_features &
+                                   ~QCOW2_INCOMPAT_MASK);
+        ret = -ENOTSUP;
+        goto fail;
+    }
+
+    /* Check support for various header values */
+    if (header.refcount_order != 4) {
+        report_unsupported(bs, "%d bit reference counts",
+                           1 << header.refcount_order);
+        ret = -ENOTSUP;
+        goto fail;
+    }
+
+    if (header.cluster_bits < MIN_CLUSTER_BITS ||
+        header.cluster_bits > MAX_CLUSTER_BITS) {
+        ret = -EINVAL;
+        goto fail;
+    }
+    if (header.crypt_method > QCOW_CRYPT_AES) {
+        ret = -EINVAL;
+        goto fail;
+    }
+    s->crypt_method_header = header.crypt_method;
+    if (s->crypt_method_header) {
+        bs->encrypted = 1;
+    }
+    s->cluster_bits = header.cluster_bits;
+    s->cluster_size = 1 << s->cluster_bits;
+    s->cluster_sectors = 1 << (s->cluster_bits - 9);
+    s->l2_bits = s->cluster_bits - 3; /* L2 is always one cluster */
+    s->l2_size = 1 << s->l2_bits;
+    bs->total_sectors = header.size / 512;
+    s->csize_shift = (62 - (s->cluster_bits - 8));
+    s->csize_mask = (1 << (s->cluster_bits - 8)) - 1;
+    s->cluster_offset_mask = (1LL << s->csize_shift) - 1;
+    s->refcount_table_offset = header.refcount_table_offset;
+    s->refcount_table_size =
+        header.refcount_table_clusters << (s->cluster_bits - 3);
+
+    s->snapshots_offset = header.snapshots_offset;
+    s->nb_snapshots = header.nb_snapshots;
+
+    /* read the level 1 table */
+    s->l1_size = header.l1_size;
+
+    l1_vm_state_index = size_to_l1(s, header.size);
+    if (l1_vm_state_index > INT_MAX) {
+        ret = -EFBIG;
+        goto fail;
+    }
+    s->l1_vm_state_index = l1_vm_state_index;
+
+    /* the L1 table must contain at least enough entries to put
+       header.size bytes */
+    if (s->l1_size < s->l1_vm_state_index) {
+        ret = -EINVAL;
+        goto fail;
+    }
+    s->l1_table_offset = header.l1_table_offset;
+    if (s->l1_size > 0) {
+        s->l1_table = g_malloc0(
+            align_offset(s->l1_size * sizeof(uint64_t), 512));
+        ret = bdrv_pread(bs->file, s->l1_table_offset, s->l1_table,
+                         s->l1_size * sizeof(uint64_t));
+        if (ret < 0) {
+            goto fail;
+        }
+        for(i = 0;i < s->l1_size; i++) {
+            be64_to_cpus(&s->l1_table[i]);
+        }
+    }
+
+    /* alloc L2 table/refcount block cache */
+    s->l2_table_cache = qcow2_cache_create(bs, L2_CACHE_SIZE);
+    s->refcount_block_cache = qcow2_cache_create(bs, REFCOUNT_CACHE_SIZE);
+
+    s->cluster_cache = g_malloc(s->cluster_size);
+    /* one more sector for decompressed data alignment */
+    s->cluster_data = qemu_blockalign(bs, QCOW_MAX_CRYPT_CLUSTERS * s->cluster_size
+                                  + 512);
+    s->cluster_cache_offset = -1;
+    s->flags = flags;
+
+    ret = qcow2_refcount_init(bs);
+    if (ret != 0) {
+        goto fail;
+    }
+
+    QLIST_INIT(&s->cluster_allocs);
+    QTAILQ_INIT(&s->discards);
+
+    /* read qcow2 extensions */
+    if (qcow2_read_extensions(bs, header.header_length, ext_end, NULL)) {
+        ret = -EINVAL;
+        goto fail;
+    }
+
+    /* read the backing file name */
+    if (header.backing_file_offset != 0) {
+        len = header.backing_file_size;
+        if (len > 1023) {
+            len = 1023;
+        }
+        ret = bdrv_pread(bs->file, header.backing_file_offset,
+                         bs->backing_file, len);
+        if (ret < 0) {
+            goto fail;
+        }
+        bs->backing_file[len] = '\0';
+    }
+
+    ret = qcow2_read_snapshots(bs);
+    if (ret < 0) {
+        goto fail;
+    }
+
+    /* Clear unknown autoclear feature bits */
+    if (!bs->read_only && s->autoclear_features != 0) {
+        s->autoclear_features = 0;
+        ret = qcow2_update_header(bs);
+        if (ret < 0) {
+            goto fail;
+        }
+    }
+
+    /* Initialise locks */
+    qemu_co_mutex_init(&s->lock);
+
+    /* Repair image if dirty */
+    if (!(flags & BDRV_O_CHECK) && !bs->read_only &&
+        (s->incompatible_features & QCOW2_INCOMPAT_DIRTY)) {
+        BdrvCheckResult result = {0};
+
+        ret = qcow2_check(bs, &result, BDRV_FIX_ERRORS);
+        if (ret < 0) {
+            goto fail;
+        }
+    }
+
+    /* Enable lazy_refcounts according to image and command line options */
+    opts = qemu_opts_create_nofail(&qcow2_runtime_opts);
+    qemu_opts_absorb_qdict(opts, options, &local_err);
+    if (error_is_set(&local_err)) {
+        qerror_report_err(local_err);
+        error_free(local_err);
+        ret = -EINVAL;
+        goto fail;
+    }
+
+    s->use_lazy_refcounts = qemu_opt_get_bool(opts, QCOW2_OPT_LAZY_REFCOUNTS,
+        (s->compatible_features & QCOW2_COMPAT_LAZY_REFCOUNTS));
+
+    s->discard_passthrough[QCOW2_DISCARD_NEVER] = false;
+    s->discard_passthrough[QCOW2_DISCARD_ALWAYS] = true;
+    s->discard_passthrough[QCOW2_DISCARD_REQUEST] =
+        qemu_opt_get_bool(opts, QCOW2_OPT_DISCARD_REQUEST,
+                          flags & BDRV_O_UNMAP);
+    s->discard_passthrough[QCOW2_DISCARD_SNAPSHOT] =
+        qemu_opt_get_bool(opts, QCOW2_OPT_DISCARD_SNAPSHOT, true);
+    s->discard_passthrough[QCOW2_DISCARD_OTHER] =
+        qemu_opt_get_bool(opts, QCOW2_OPT_DISCARD_OTHER, false);
+
+    qemu_opts_del(opts);
+
+    if (s->use_lazy_refcounts && s->qcow_version < 3) {
+        qerror_report(ERROR_CLASS_GENERIC_ERROR, "Lazy refcounts require "
+            "a qcow2 image with at least qemu 1.1 compatibility level");
+        ret = -EINVAL;
+        goto fail;
+    }
+
+#ifdef DEBUG_ALLOC
+    {
+        BdrvCheckResult result = {0};
+        qcow2_check_refcounts(bs, &result, 0);
+    }
+#endif
+    return ret;
+
+ fail:
+    g_free(s->unknown_header_fields);
+    cleanup_unknown_header_ext(bs);
+    qcow2_free_snapshots(bs);
+    qcow2_refcount_close(bs);
+    g_free(s->l1_table);
+    if (s->l2_table_cache) {
+        qcow2_cache_destroy(bs, s->l2_table_cache);
+    }
+    g_free(s->cluster_cache);
+    qemu_vfree(s->cluster_data);
+    return ret;
+}
+
+static int qcow2_set_key(BlockDriverState *bs, const char *key)
+{
+    BDRVQcowState *s = bs->opaque;
+    uint8_t keybuf[16];
+    int len, i;
+
+    memset(keybuf, 0, 16);
+    len = strlen(key);
+    if (len > 16)
+        len = 16;
+    /* XXX: we could compress the chars to 7 bits to increase
+       entropy */
+    for(i = 0;i < len;i++) {
+        keybuf[i] = key[i];
+    }
+    s->crypt_method = s->crypt_method_header;
+
+    if (AES_set_encrypt_key(keybuf, 128, &s->aes_encrypt_key) != 0)
+        return -1;
+    if (AES_set_decrypt_key(keybuf, 128, &s->aes_decrypt_key) != 0)
+        return -1;
+#if 0
+    /* test */
+    {
+        uint8_t in[16];
+        uint8_t out[16];
+        uint8_t tmp[16];
+        for(i=0;i<16;i++)
+            in[i] = i;
+        AES_encrypt(in, tmp, &s->aes_encrypt_key);
+        AES_decrypt(tmp, out, &s->aes_decrypt_key);
+        for(i = 0; i < 16; i++)
+            printf(" %02x", tmp[i]);
+        printf("\n");
+        for(i = 0; i < 16; i++)
+            printf(" %02x", out[i]);
+        printf("\n");
+    }
+#endif
+    return 0;
+}
+
+/* We have nothing to do for QCOW2 reopen, stubs just return
+ * success */
+static int qcow2_reopen_prepare(BDRVReopenState *state,
+                                BlockReopenQueue *queue, Error **errp)
+{
+    return 0;
+}
+
+static int coroutine_fn qcow2_co_is_allocated(BlockDriverState *bs,
+        int64_t sector_num, int nb_sectors, int *pnum)
+{
+    BDRVQcowState *s = bs->opaque;
+    uint64_t cluster_offset;
+    int ret;
+
+    *pnum = nb_sectors;
+    /* FIXME We can get errors here, but the bdrv_co_is_allocated interface
+     * can't pass them on today */
+    qemu_co_mutex_lock(&s->lock);
+    ret = qcow2_get_cluster_offset(bs, sector_num << 9, pnum, &cluster_offset);
+    qemu_co_mutex_unlock(&s->lock);
+    if (ret < 0) {
+        *pnum = 0;
+    }
+
+    return (cluster_offset != 0) || (ret == QCOW2_CLUSTER_ZERO);
+}
+
+/* handle reading after the end of the backing file */
+int qcow2_backing_read1(BlockDriverState *bs, QEMUIOVector *qiov,
+                  int64_t sector_num, int nb_sectors)
+{
+    int n1;
+    if ((sector_num + nb_sectors) <= bs->total_sectors)
+        return nb_sectors;
+    if (sector_num >= bs->total_sectors)
+        n1 = 0;
+    else
+        n1 = bs->total_sectors - sector_num;
+
+    qemu_iovec_memset(qiov, 512 * n1, 0, 512 * (nb_sectors - n1));
+
+    return n1;
+}
+
+static coroutine_fn int qcow2_co_readv(BlockDriverState *bs, int64_t sector_num,
+                          int remaining_sectors, QEMUIOVector *qiov)
+{
+    BDRVQcowState *s = bs->opaque;
+    int index_in_cluster, n1;
+    int ret;
+    int cur_nr_sectors; /* number of sectors in current iteration */
+    uint64_t cluster_offset = 0;
+    uint64_t bytes_done = 0;
+    QEMUIOVector hd_qiov;
+    uint8_t *cluster_data = NULL;
+
+    qemu_iovec_init(&hd_qiov, qiov->niov);
+
+    qemu_co_mutex_lock(&s->lock);
+
+    while (remaining_sectors != 0) {
+
+        /* prepare next request */
+        cur_nr_sectors = remaining_sectors;
+        if (s->crypt_method) {
+            cur_nr_sectors = MIN(cur_nr_sectors,
+                QCOW_MAX_CRYPT_CLUSTERS * s->cluster_sectors);
+        }
+
+        ret = qcow2_get_cluster_offset(bs, sector_num << 9,
+            &cur_nr_sectors, &cluster_offset);
+        if (ret < 0) {
+            goto fail;
+        }
+
+        index_in_cluster = sector_num & (s->cluster_sectors - 1);
+
+        qemu_iovec_reset(&hd_qiov);
+        qemu_iovec_concat(&hd_qiov, qiov, bytes_done,
+            cur_nr_sectors * 512);
+
+        switch (ret) {
+        case QCOW2_CLUSTER_UNALLOCATED:
+
+            if (bs->backing_hd) {
+                /* read from the base image */
+                n1 = qcow2_backing_read1(bs->backing_hd, &hd_qiov,
+                    sector_num, cur_nr_sectors);
+                if (n1 > 0) {
+                    BLKDBG_EVENT(bs->file, BLKDBG_READ_BACKING_AIO);
+                    qemu_co_mutex_unlock(&s->lock);
+                    ret = bdrv_co_readv(bs->backing_hd, sector_num,
+                                        n1, &hd_qiov);
+                    qemu_co_mutex_lock(&s->lock);
+                    if (ret < 0) {
+                        goto fail;
+                    }
+                }
+            } else {
+                /* Note: in this case, no need to wait */
+                qemu_iovec_memset(&hd_qiov, 0, 0, 512 * cur_nr_sectors);
+            }
+            break;
+
+        case QCOW2_CLUSTER_ZERO:
+            qemu_iovec_memset(&hd_qiov, 0, 0, 512 * cur_nr_sectors);
+            break;
+
+        case QCOW2_CLUSTER_COMPRESSED:
+            /* add AIO support for compressed blocks ? */
+            ret = qcow2_decompress_cluster(bs, cluster_offset);
+            if (ret < 0) {
+                goto fail;
+            }
+
+            qemu_iovec_from_buf(&hd_qiov, 0,
+                s->cluster_cache + index_in_cluster * 512,
+                512 * cur_nr_sectors);
+            break;
+
+        case QCOW2_CLUSTER_NORMAL:
+            if ((cluster_offset & 511) != 0) {
+                ret = -EIO;
+                goto fail;
+            }
+
+            if (s->crypt_method) {
+                /*
+                 * For encrypted images, read everything into a temporary
+                 * contiguous buffer on which the AES functions can work.
+                 */
+                if (!cluster_data) {
+                    cluster_data =
+                        qemu_blockalign(bs, QCOW_MAX_CRYPT_CLUSTERS * s->cluster_size);
+                }
+
+                assert(cur_nr_sectors <=
+                    QCOW_MAX_CRYPT_CLUSTERS * s->cluster_sectors);
+                qemu_iovec_reset(&hd_qiov);
+                qemu_iovec_add(&hd_qiov, cluster_data,
+                    512 * cur_nr_sectors);
+            }
+
+            BLKDBG_EVENT(bs->file, BLKDBG_READ_AIO);
+            qemu_co_mutex_unlock(&s->lock);
+            ret = bdrv_co_readv(bs->file,
+                                (cluster_offset >> 9) + index_in_cluster,
+                                cur_nr_sectors, &hd_qiov);
+            qemu_co_mutex_lock(&s->lock);
+            if (ret < 0) {
+                goto fail;
+            }
+            if (s->crypt_method) {
+                qcow2_encrypt_sectors(s, sector_num,  cluster_data,
+                    cluster_data, cur_nr_sectors, 0, &s->aes_decrypt_key);
+                qemu_iovec_from_buf(qiov, bytes_done,
+                    cluster_data, 512 * cur_nr_sectors);
+            }
+            break;
+
+        default:
+            g_assert_not_reached();
+            ret = -EIO;
+            goto fail;
+        }
+
+        remaining_sectors -= cur_nr_sectors;
+        sector_num += cur_nr_sectors;
+        bytes_done += cur_nr_sectors * 512;
+    }
+    ret = 0;
+
+fail:
+    qemu_co_mutex_unlock(&s->lock);
+
+    qemu_iovec_destroy(&hd_qiov);
+    qemu_vfree(cluster_data);
+
+    return ret;
+}
+
+static coroutine_fn int qcow2_co_writev(BlockDriverState *bs,
+                           int64_t sector_num,
+                           int remaining_sectors,
+                           QEMUIOVector *qiov)
+{
+    BDRVQcowState *s = bs->opaque;
+    int index_in_cluster;
+    int n_end;
+    int ret;
+    int cur_nr_sectors; /* number of sectors in current iteration */
+    uint64_t cluster_offset;
+    QEMUIOVector hd_qiov;
+    uint64_t bytes_done = 0;
+    uint8_t *cluster_data = NULL;
+    QCowL2Meta *l2meta = NULL;
+
+    trace_qcow2_writev_start_req(qemu_coroutine_self(), sector_num,
+                                 remaining_sectors);
+
+    qemu_iovec_init(&hd_qiov, qiov->niov);
+
+    s->cluster_cache_offset = -1; /* disable compressed cache */
+
+    qemu_co_mutex_lock(&s->lock);
+
+    while (remaining_sectors != 0) {
+
+        l2meta = NULL;
+
+        trace_qcow2_writev_start_part(qemu_coroutine_self());
+        index_in_cluster = sector_num & (s->cluster_sectors - 1);
+        n_end = index_in_cluster + remaining_sectors;
+        if (s->crypt_method &&
+            n_end > QCOW_MAX_CRYPT_CLUSTERS * s->cluster_sectors) {
+            n_end = QCOW_MAX_CRYPT_CLUSTERS * s->cluster_sectors;
+        }
+
+        ret = qcow2_alloc_cluster_offset(bs, sector_num << 9,
+            index_in_cluster, n_end, &cur_nr_sectors, &cluster_offset, &l2meta);
+        if (ret < 0) {
+            goto fail;
+        }
+
+        assert((cluster_offset & 511) == 0);
+
+        qemu_iovec_reset(&hd_qiov);
+        qemu_iovec_concat(&hd_qiov, qiov, bytes_done,
+            cur_nr_sectors * 512);
+
+        if (s->crypt_method) {
+            if (!cluster_data) {
+                cluster_data = qemu_blockalign(bs, QCOW_MAX_CRYPT_CLUSTERS *
+                                                 s->cluster_size);
+            }
+
+            assert(hd_qiov.size <=
+                   QCOW_MAX_CRYPT_CLUSTERS * s->cluster_size);
+            qemu_iovec_to_buf(&hd_qiov, 0, cluster_data, hd_qiov.size);
+
+            qcow2_encrypt_sectors(s, sector_num, cluster_data,
+                cluster_data, cur_nr_sectors, 1, &s->aes_encrypt_key);
+
+            qemu_iovec_reset(&hd_qiov);
+            qemu_iovec_add(&hd_qiov, cluster_data,
+                cur_nr_sectors * 512);
+        }
+
+        qemu_co_mutex_unlock(&s->lock);
+        BLKDBG_EVENT(bs->file, BLKDBG_WRITE_AIO);
+        trace_qcow2_writev_data(qemu_coroutine_self(),
+                                (cluster_offset >> 9) + index_in_cluster);
+        ret = bdrv_co_writev(bs->file,
+                             (cluster_offset >> 9) + index_in_cluster,
+                             cur_nr_sectors, &hd_qiov);
+        qemu_co_mutex_lock(&s->lock);
+        if (ret < 0) {
+            goto fail;
+        }
+
+        while (l2meta != NULL) {
+            QCowL2Meta *next;
+
+            ret = qcow2_alloc_cluster_link_l2(bs, l2meta);
+            if (ret < 0) {
+                goto fail;
+            }
+
+            /* Take the request off the list of running requests */
+            if (l2meta->nb_clusters != 0) {
+                QLIST_REMOVE(l2meta, next_in_flight);
+            }
+
+            qemu_co_queue_restart_all(&l2meta->dependent_requests);
+
+            next = l2meta->next;
+            g_free(l2meta);
+            l2meta = next;
+        }
+
+        remaining_sectors -= cur_nr_sectors;
+        sector_num += cur_nr_sectors;
+        bytes_done += cur_nr_sectors * 512;
+        trace_qcow2_writev_done_part(qemu_coroutine_self(), cur_nr_sectors);
+    }
+    ret = 0;
+
+fail:
+    qemu_co_mutex_unlock(&s->lock);
+
+    while (l2meta != NULL) {
+        QCowL2Meta *next;
+
+        if (l2meta->nb_clusters != 0) {
+            QLIST_REMOVE(l2meta, next_in_flight);
+        }
+        qemu_co_queue_restart_all(&l2meta->dependent_requests);
+
+        next = l2meta->next;
+        g_free(l2meta);
+        l2meta = next;
+    }
+
+    qemu_iovec_destroy(&hd_qiov);
+    qemu_vfree(cluster_data);
+    trace_qcow2_writev_done_req(qemu_coroutine_self(), ret);
+
+    return ret;
+}
+
+static void qcow2_close(BlockDriverState *bs)
+{
+    BDRVQcowState *s = bs->opaque;
+    g_free(s->l1_table);
+
+    qcow2_cache_flush(bs, s->l2_table_cache);
+    qcow2_cache_flush(bs, s->refcount_block_cache);
+
+    qcow2_mark_clean(bs);
+
+    qcow2_cache_destroy(bs, s->l2_table_cache);
+    qcow2_cache_destroy(bs, s->refcount_block_cache);
+
+    g_free(s->unknown_header_fields);
+    cleanup_unknown_header_ext(bs);
+
+    g_free(s->cluster_cache);
+    qemu_vfree(s->cluster_data);
+    qcow2_refcount_close(bs);
+    qcow2_free_snapshots(bs);
+}
+
+static void qcow2_invalidate_cache(BlockDriverState *bs)
+{
+    BDRVQcowState *s = bs->opaque;
+    int flags = s->flags;
+    AES_KEY aes_encrypt_key;
+    AES_KEY aes_decrypt_key;
+    uint32_t crypt_method = 0;
+    QDict *options;
+
+    /*
+     * Backing files are read-only which makes all of their metadata immutable,
+     * that means we don't have to worry about reopening them here.
+     */
+
+    if (s->crypt_method) {
+        crypt_method = s->crypt_method;
+        memcpy(&aes_encrypt_key, &s->aes_encrypt_key, sizeof(aes_encrypt_key));
+        memcpy(&aes_decrypt_key, &s->aes_decrypt_key, sizeof(aes_decrypt_key));
+    }
+
+    qcow2_close(bs);
+
+    options = qdict_new();
+    qdict_put(options, QCOW2_OPT_LAZY_REFCOUNTS,
+              qbool_from_int(s->use_lazy_refcounts));
+
+    memset(s, 0, sizeof(BDRVQcowState));
+    qcow2_open(bs, options, flags);
+
+    QDECREF(options);
+
+    if (crypt_method) {
+        s->crypt_method = crypt_method;
+        memcpy(&s->aes_encrypt_key, &aes_encrypt_key, sizeof(aes_encrypt_key));
+        memcpy(&s->aes_decrypt_key, &aes_decrypt_key, sizeof(aes_decrypt_key));
+    }
+}
+
+static size_t header_ext_add(char *buf, uint32_t magic, const void *s,
+    size_t len, size_t buflen)
+{
+    QCowExtension *ext_backing_fmt = (QCowExtension*) buf;
+    size_t ext_len = sizeof(QCowExtension) + ((len + 7) & ~7);
+
+    if (buflen < ext_len) {
+        return -ENOSPC;
+    }
+
+    *ext_backing_fmt = (QCowExtension) {
+        .magic  = cpu_to_be32(magic),
+        .len    = cpu_to_be32(len),
+    };
+    memcpy(buf + sizeof(QCowExtension), s, len);
+
+    return ext_len;
+}
+
+/*
+ * Updates the qcow2 header, including the variable length parts of it, i.e.
+ * the backing file name and all extensions. qcow2 was not designed to allow
+ * such changes, so if we run out of space (we can only use the first cluster)
+ * this function may fail.
+ *
+ * Returns 0 on success, -errno in error cases.
+ */
+int qcow2_update_header(BlockDriverState *bs)
+{
+    BDRVQcowState *s = bs->opaque;
+    QCowHeader *header;
+    char *buf;
+    size_t buflen = s->cluster_size;
+    int ret;
+    uint64_t total_size;
+    uint32_t refcount_table_clusters;
+    size_t header_length;
+    Qcow2UnknownHeaderExtension *uext;
+
+    buf = qemu_blockalign(bs, buflen);
+
+    /* Header structure */
+    header = (QCowHeader*) buf;
+
+    if (buflen < sizeof(*header)) {
+        ret = -ENOSPC;
+        goto fail;
+    }
+
+    header_length = sizeof(*header) + s->unknown_header_fields_size;
+    total_size = bs->total_sectors * BDRV_SECTOR_SIZE;
+    refcount_table_clusters = s->refcount_table_size >> (s->cluster_bits - 3);
+
+    *header = (QCowHeader) {
+        /* Version 2 fields */
+        .magic                  = cpu_to_be32(QCOW_MAGIC),
+        .version                = cpu_to_be32(s->qcow_version),
+        .backing_file_offset    = 0,
+        .backing_file_size      = 0,
+        .cluster_bits           = cpu_to_be32(s->cluster_bits),
+        .size                   = cpu_to_be64(total_size),
+        .crypt_method           = cpu_to_be32(s->crypt_method_header),
+        .l1_size                = cpu_to_be32(s->l1_size),
+        .l1_table_offset        = cpu_to_be64(s->l1_table_offset),
+        .refcount_table_offset  = cpu_to_be64(s->refcount_table_offset),
+        .refcount_table_clusters = cpu_to_be32(refcount_table_clusters),
+        .nb_snapshots           = cpu_to_be32(s->nb_snapshots),
+        .snapshots_offset       = cpu_to_be64(s->snapshots_offset),
+
+        /* Version 3 fields */
+        .incompatible_features  = cpu_to_be64(s->incompatible_features),
+        .compatible_features    = cpu_to_be64(s->compatible_features),
+        .autoclear_features     = cpu_to_be64(s->autoclear_features),
+        .refcount_order         = cpu_to_be32(3 + REFCOUNT_SHIFT),
+        .header_length          = cpu_to_be32(header_length),
+    };
+
+    /* For older versions, write a shorter header */
+    switch (s->qcow_version) {
+    case 2:
+        ret = offsetof(QCowHeader, incompatible_features);
+        break;
+    case 3:
+        ret = sizeof(*header);
+        break;
+    default:
+        ret = -EINVAL;
+        goto fail;
+    }
+
+    buf += ret;
+    buflen -= ret;
+    memset(buf, 0, buflen);
+
+    /* Preserve any unknown field in the header */
+    if (s->unknown_header_fields_size) {
+        if (buflen < s->unknown_header_fields_size) {
+            ret = -ENOSPC;
+            goto fail;
+        }
+
+        memcpy(buf, s->unknown_header_fields, s->unknown_header_fields_size);
+        buf += s->unknown_header_fields_size;
+        buflen -= s->unknown_header_fields_size;
+    }
+
+    /* Backing file format header extension */
+    if (*bs->backing_format) {
+        ret = header_ext_add(buf, QCOW2_EXT_MAGIC_BACKING_FORMAT,
+                             bs->backing_format, strlen(bs->backing_format),
+                             buflen);
+        if (ret < 0) {
+            goto fail;
+        }
+
+        buf += ret;
+        buflen -= ret;
+    }
+
+    /* Feature table */
+    Qcow2Feature features[] = {
+        {
+            .type = QCOW2_FEAT_TYPE_INCOMPATIBLE,
+            .bit  = QCOW2_INCOMPAT_DIRTY_BITNR,
+            .name = "dirty bit",
+        },
+        {
+            .type = QCOW2_FEAT_TYPE_COMPATIBLE,
+            .bit  = QCOW2_COMPAT_LAZY_REFCOUNTS_BITNR,
+            .name = "lazy refcounts",
+        },
+    };
+
+    ret = header_ext_add(buf, QCOW2_EXT_MAGIC_FEATURE_TABLE,
+                         features, sizeof(features), buflen);
+    if (ret < 0) {
+        goto fail;
+    }
+    buf += ret;
+    buflen -= ret;
+
+    /* Keep unknown header extensions */
+    QLIST_FOREACH(uext, &s->unknown_header_ext, next) {
+        ret = header_ext_add(buf, uext->magic, uext->data, uext->len, buflen);
+        if (ret < 0) {
+            goto fail;
+        }
+
+        buf += ret;
+        buflen -= ret;
+    }
+
+    /* End of header extensions */
+    ret = header_ext_add(buf, QCOW2_EXT_MAGIC_END, NULL, 0, buflen);
+    if (ret < 0) {
+        goto fail;
+    }
+
+    buf += ret;
+    buflen -= ret;
+
+    /* Backing file name */
+    if (*bs->backing_file) {
+        size_t backing_file_len = strlen(bs->backing_file);
+
+        if (buflen < backing_file_len) {
+            ret = -ENOSPC;
+            goto fail;
+        }
+
+        /* Using strncpy is ok here, since buf is not NUL-terminated. */
+        strncpy(buf, bs->backing_file, buflen);
+
+        header->backing_file_offset = cpu_to_be64(buf - ((char*) header));
+        header->backing_file_size   = cpu_to_be32(backing_file_len);
+    }
+
+    /* Write the new header */
+    ret = bdrv_pwrite(bs->file, 0, header, s->cluster_size);
+    if (ret < 0) {
+        goto fail;
+    }
+
+    ret = 0;
+fail:
+    qemu_vfree(header);
+    return ret;
+}
+
+static int qcow2_change_backing_file(BlockDriverState *bs,
+    const char *backing_file, const char *backing_fmt)
+{
+    pstrcpy(bs->backing_file, sizeof(bs->backing_file), backing_file ?: "");
+    pstrcpy(bs->backing_format, sizeof(bs->backing_format), backing_fmt ?: "");
+
+    return qcow2_update_header(bs);
+}
+
+static int preallocate(BlockDriverState *bs)
+{
+    uint64_t nb_sectors;
+    uint64_t offset;
+    uint64_t host_offset = 0;
+    int num;
+    int ret;
+    QCowL2Meta *meta;
+
+    nb_sectors = bdrv_getlength(bs) >> 9;
+    offset = 0;
+
+    while (nb_sectors) {
+        num = MIN(nb_sectors, INT_MAX >> 9);
+        ret = qcow2_alloc_cluster_offset(bs, offset, 0, num, &num,
+                                         &host_offset, &meta);
+        if (ret < 0) {
+            return ret;
+        }
+
+        ret = qcow2_alloc_cluster_link_l2(bs, meta);
+        if (ret < 0) {
+            qcow2_free_any_clusters(bs, meta->alloc_offset, meta->nb_clusters,
+                                    QCOW2_DISCARD_NEVER);
+            return ret;
+        }
+
+        /* There are no dependent requests, but we need to remove our request
+         * from the list of in-flight requests */
+        if (meta != NULL) {
+            QLIST_REMOVE(meta, next_in_flight);
+        }
+
+        /* TODO Preallocate data if requested */
+
+        nb_sectors -= num;
+        offset += num << 9;
+    }
+
+    /*
+     * It is expected that the image file is large enough to actually contain
+     * all of the allocated clusters (otherwise we get failing reads after
+     * EOF). Extend the image to the last allocated sector.
+     */
+    if (host_offset != 0) {
+        uint8_t buf[512];
+        memset(buf, 0, 512);
+        ret = bdrv_write(bs->file, (host_offset >> 9) + num - 1, buf, 1);
+        if (ret < 0) {
+            return ret;
+        }
+    }
+
+    return 0;
+}
+
+static int qcow2_create2(const char *filename, int64_t total_size,
+                         const char *backing_file, const char *backing_format,
+                         int flags, size_t cluster_size, int prealloc,
+                         QEMUOptionParameter *options, int version)
+{
+    /* Calculate cluster_bits */
+    int cluster_bits;
+    cluster_bits = ffs(cluster_size) - 1;
+    if (cluster_bits < MIN_CLUSTER_BITS || cluster_bits > MAX_CLUSTER_BITS ||
+        (1 << cluster_bits) != cluster_size)
+    {
+        error_report(
+            "Cluster size must be a power of two between %d and %dk",
+            1 << MIN_CLUSTER_BITS, 1 << (MAX_CLUSTER_BITS - 10));
+        return -EINVAL;
+    }
+
+    /*
+     * Open the image file and write a minimal qcow2 header.
+     *
+     * We keep things simple and start with a zero-sized image. We also
+     * do without refcount blocks or a L1 table for now. We'll fix the
+     * inconsistency later.
+     *
+     * We do need a refcount table because growing the refcount table means
+     * allocating two new refcount blocks - the seconds of which would be at
+     * 2 GB for 64k clusters, and we don't want to have a 2 GB initial file
+     * size for any qcow2 image.
+     */
+    BlockDriverState* bs;
+    QCowHeader header;
+    uint8_t* refcount_table;
+    int ret;
+
+    ret = bdrv_create_file(filename, options);
+    if (ret < 0) {
+        return ret;
+    }
+
+    ret = bdrv_file_open(&bs, filename, NULL, BDRV_O_RDWR);
+    if (ret < 0) {
+        return ret;
+    }
+
+    /* Write the header */
+    memset(&header, 0, sizeof(header));
+    header.magic = cpu_to_be32(QCOW_MAGIC);
+    header.version = cpu_to_be32(version);
+    header.cluster_bits = cpu_to_be32(cluster_bits);
+    header.size = cpu_to_be64(0);
+    header.l1_table_offset = cpu_to_be64(0);
+    header.l1_size = cpu_to_be32(0);
+    header.refcount_table_offset = cpu_to_be64(cluster_size);
+    header.refcount_table_clusters = cpu_to_be32(1);
+    header.refcount_order = cpu_to_be32(3 + REFCOUNT_SHIFT);
+    header.header_length = cpu_to_be32(sizeof(header));
+
+    if (flags & BLOCK_FLAG_ENCRYPT) {
+        header.crypt_method = cpu_to_be32(QCOW_CRYPT_AES);
+    } else {
+        header.crypt_method = cpu_to_be32(QCOW_CRYPT_NONE);
+    }
+
+    if (flags & BLOCK_FLAG_LAZY_REFCOUNTS) {
+        header.compatible_features |=
+            cpu_to_be64(QCOW2_COMPAT_LAZY_REFCOUNTS);
+    }
+
+    ret = bdrv_pwrite(bs, 0, &header, sizeof(header));
+    if (ret < 0) {
+        goto out;
+    }
+
+    /* Write an empty refcount table */
+    refcount_table = g_malloc0(cluster_size);
+    ret = bdrv_pwrite(bs, cluster_size, refcount_table, cluster_size);
+    g_free(refcount_table);
+
+    if (ret < 0) {
+        goto out;
+    }
+
+    bdrv_close(bs);
+
+    /*
+     * And now open the image and make it consistent first (i.e. increase the
+     * refcount of the cluster that is occupied by the header and the refcount
+     * table)
+     */
+    BlockDriver* drv = bdrv_find_format("qcow2");
+    assert(drv != NULL);
+    ret = bdrv_open(bs, filename, NULL,
+        BDRV_O_RDWR | BDRV_O_CACHE_WB | BDRV_O_NO_FLUSH, drv);
+    if (ret < 0) {
+        goto out;
+    }
+
+    ret = qcow2_alloc_clusters(bs, 2 * cluster_size);
+    if (ret < 0) {
+        goto out;
+
+    } else if (ret != 0) {
+        error_report("Huh, first cluster in empty image is already in use?");
+        abort();
+    }
+
+    /* Okay, now that we have a valid image, let's give it the right size */
+    ret = bdrv_truncate(bs, total_size * BDRV_SECTOR_SIZE);
+    if (ret < 0) {
+        goto out;
+    }
+
+    /* Want a backing file? There you go.*/
+    if (backing_file) {
+        ret = bdrv_change_backing_file(bs, backing_file, backing_format);
+        if (ret < 0) {
+            goto out;
+        }
+    }
+
+    /* And if we're supposed to preallocate metadata, do that now */
+    if (prealloc) {
+        BDRVQcowState *s = bs->opaque;
+        qemu_co_mutex_lock(&s->lock);
+        ret = preallocate(bs);
+        qemu_co_mutex_unlock(&s->lock);
+        if (ret < 0) {
+            goto out;
+        }
+    }
+
+    ret = 0;
+out:
+    bdrv_delete(bs);
+    return ret;
+}
+
+static int qcow2_create(const char *filename, QEMUOptionParameter *options)
+{
+    const char *backing_file = NULL;
+    const char *backing_fmt = NULL;
+    uint64_t sectors = 0;
+    int flags = 0;
+    size_t cluster_size = DEFAULT_CLUSTER_SIZE;
+    int prealloc = 0;
+    int version = 2;
+
+    /* Read out options */
+    while (options && options->name) {
+        if (!strcmp(options->name, BLOCK_OPT_SIZE)) {
+            sectors = options->value.n / 512;
+        } else if (!strcmp(options->name, BLOCK_OPT_BACKING_FILE)) {
+            backing_file = options->value.s;
+        } else if (!strcmp(options->name, BLOCK_OPT_BACKING_FMT)) {
+            backing_fmt = options->value.s;
+        } else if (!strcmp(options->name, BLOCK_OPT_ENCRYPT)) {
+            flags |= options->value.n ? BLOCK_FLAG_ENCRYPT : 0;
+        } else if (!strcmp(options->name, BLOCK_OPT_CLUSTER_SIZE)) {
+            if (options->value.n) {
+                cluster_size = options->value.n;
+            }
+        } else if (!strcmp(options->name, BLOCK_OPT_PREALLOC)) {
+            if (!options->value.s || !strcmp(options->value.s, "off")) {
+                prealloc = 0;
+            } else if (!strcmp(options->value.s, "metadata")) {
+                prealloc = 1;
+            } else {
+                fprintf(stderr, "Invalid preallocation mode: '%s'\n",
+                    options->value.s);
+                return -EINVAL;
+            }
+        } else if (!strcmp(options->name, BLOCK_OPT_COMPAT_LEVEL)) {
+            if (!options->value.s || !strcmp(options->value.s, "0.10")) {
+                version = 2;
+            } else if (!strcmp(options->value.s, "1.1")) {
+                version = 3;
+            } else {
+                fprintf(stderr, "Invalid compatibility level: '%s'\n",
+                    options->value.s);
+                return -EINVAL;
+            }
+        } else if (!strcmp(options->name, BLOCK_OPT_LAZY_REFCOUNTS)) {
+            flags |= options->value.n ? BLOCK_FLAG_LAZY_REFCOUNTS : 0;
+        }
+        options++;
+    }
+
+    if (backing_file && prealloc) {
+        fprintf(stderr, "Backing file and preallocation cannot be used at "
+            "the same time\n");
+        return -EINVAL;
+    }
+
+    if (version < 3 && (flags & BLOCK_FLAG_LAZY_REFCOUNTS)) {
+        fprintf(stderr, "Lazy refcounts only supported with compatibility "
+                "level 1.1 and above (use compat=1.1 or greater)\n");
+        return -EINVAL;
+    }
+
+    return qcow2_create2(filename, sectors, backing_file, backing_fmt, flags,
+                         cluster_size, prealloc, options, version);
+}
+
+static int qcow2_make_empty(BlockDriverState *bs)
+{
+#if 0
+    /* XXX: not correct */
+    BDRVQcowState *s = bs->opaque;
+    uint32_t l1_length = s->l1_size * sizeof(uint64_t);
+    int ret;
+
+    memset(s->l1_table, 0, l1_length);
+    if (bdrv_pwrite(bs->file, s->l1_table_offset, s->l1_table, l1_length) < 0)
+        return -1;
+    ret = bdrv_truncate(bs->file, s->l1_table_offset + l1_length);
+    if (ret < 0)
+        return ret;
+
+    l2_cache_reset(bs);
+#endif
+    return 0;
+}
+
+static coroutine_fn int qcow2_co_write_zeroes(BlockDriverState *bs,
+    int64_t sector_num, int nb_sectors)
+{
+    int ret;
+    BDRVQcowState *s = bs->opaque;
+
+    /* Emulate misaligned zero writes */
+    if (sector_num % s->cluster_sectors || nb_sectors % s->cluster_sectors) {
+        return -ENOTSUP;
+    }
+
+    /* Whatever is left can use real zero clusters */
+    qemu_co_mutex_lock(&s->lock);
+    ret = qcow2_zero_clusters(bs, sector_num << BDRV_SECTOR_BITS,
+        nb_sectors);
+    qemu_co_mutex_unlock(&s->lock);
+
+    return ret;
+}
+
+static coroutine_fn int qcow2_co_discard(BlockDriverState *bs,
+    int64_t sector_num, int nb_sectors)
+{
+    int ret;
+    BDRVQcowState *s = bs->opaque;
+
+    qemu_co_mutex_lock(&s->lock);
+    ret = qcow2_discard_clusters(bs, sector_num << BDRV_SECTOR_BITS,
+        nb_sectors);
+    qemu_co_mutex_unlock(&s->lock);
+    return ret;
+}
+
+static int qcow2_truncate(BlockDriverState *bs, int64_t offset)
+{
+    BDRVQcowState *s = bs->opaque;
+    int64_t new_l1_size;
+    int ret;
+
+    if (offset & 511) {
+        error_report("The new size must be a multiple of 512");
+        return -EINVAL;
+    }
+
+    /* cannot proceed if image has snapshots */
+    if (s->nb_snapshots) {
+        error_report("Can't resize an image which has snapshots");
+        return -ENOTSUP;
+    }
+
+    /* shrinking is currently not supported */
+    if (offset < bs->total_sectors * 512) {
+        error_report("qcow2 doesn't support shrinking images yet");
+        return -ENOTSUP;
+    }
+
+    new_l1_size = size_to_l1(s, offset);
+    ret = qcow2_grow_l1_table(bs, new_l1_size, true);
+    if (ret < 0) {
+        return ret;
+    }
+
+    /* write updated header.size */
+    offset = cpu_to_be64(offset);
+    ret = bdrv_pwrite_sync(bs->file, offsetof(QCowHeader, size),
+                           &offset, sizeof(uint64_t));
+    if (ret < 0) {
+        return ret;
+    }
+
+    s->l1_vm_state_index = new_l1_size;
+    return 0;
+}
+
+/* XXX: put compressed sectors first, then all the cluster aligned
+   tables to avoid losing bytes in alignment */
+static int qcow2_write_compressed(BlockDriverState *bs, int64_t sector_num,
+                                  const uint8_t *buf, int nb_sectors)
+{
+    BDRVQcowState *s = bs->opaque;
+    z_stream strm;
+    int ret, out_len;
+    uint8_t *out_buf;
+    uint64_t cluster_offset;
+
+    if (nb_sectors == 0) {
+        /* align end of file to a sector boundary to ease reading with
+           sector based I/Os */
+        cluster_offset = bdrv_getlength(bs->file);
+        cluster_offset = (cluster_offset + 511) & ~511;
+        bdrv_truncate(bs->file, cluster_offset);
+        return 0;
+    }
+
+    if (nb_sectors != s->cluster_sectors) {
+        ret = -EINVAL;
+
+        /* Zero-pad last write if image size is not cluster aligned */
+        if (sector_num + nb_sectors == bs->total_sectors &&
+            nb_sectors < s->cluster_sectors) {
+            uint8_t *pad_buf = qemu_blockalign(bs, s->cluster_size);
+            memset(pad_buf, 0, s->cluster_size);
+            memcpy(pad_buf, buf, nb_sectors * BDRV_SECTOR_SIZE);
+            ret = qcow2_write_compressed(bs, sector_num,
+                                         pad_buf, s->cluster_sectors);
+            qemu_vfree(pad_buf);
+        }
+        return ret;
+    }
+
+    out_buf = g_malloc(s->cluster_size + (s->cluster_size / 1000) + 128);
+
+    /* best compression, small window, no zlib header */
+    memset(&strm, 0, sizeof(strm));
+    ret = deflateInit2(&strm, Z_DEFAULT_COMPRESSION,
+                       Z_DEFLATED, -12,
+                       9, Z_DEFAULT_STRATEGY);
+    if (ret != 0) {
+        ret = -EINVAL;
+        goto fail;
+    }
+
+    strm.avail_in = s->cluster_size;
+    strm.next_in = (uint8_t *)buf;
+    strm.avail_out = s->cluster_size;
+    strm.next_out = out_buf;
+
+    ret = deflate(&strm, Z_FINISH);
+    if (ret != Z_STREAM_END && ret != Z_OK) {
+        deflateEnd(&strm);
+        ret = -EINVAL;
+        goto fail;
+    }
+    out_len = strm.next_out - out_buf;
+
+    deflateEnd(&strm);
+
+    if (ret != Z_STREAM_END || out_len >= s->cluster_size) {
+        /* could not compress: write normal cluster */
+        ret = bdrv_write(bs, sector_num, buf, s->cluster_sectors);
+        if (ret < 0) {
+            goto fail;
+        }
+    } else {
+        cluster_offset = qcow2_alloc_compressed_cluster_offset(bs,
+            sector_num << 9, out_len);
+        if (!cluster_offset) {
+            ret = -EIO;
+            goto fail;
+        }
+        cluster_offset &= s->cluster_offset_mask;
+        BLKDBG_EVENT(bs->file, BLKDBG_WRITE_COMPRESSED);
+        ret = bdrv_pwrite(bs->file, cluster_offset, out_buf, out_len);
+        if (ret < 0) {
+            goto fail;
+        }
+    }
+
+    ret = 0;
+fail:
+    g_free(out_buf);
+    return ret;
+}
+
+static coroutine_fn int qcow2_co_flush_to_os(BlockDriverState *bs)
+{
+    BDRVQcowState *s = bs->opaque;
+    int ret;
+
+    qemu_co_mutex_lock(&s->lock);
+    ret = qcow2_cache_flush(bs, s->l2_table_cache);
+    if (ret < 0) {
+        qemu_co_mutex_unlock(&s->lock);
+        return ret;
+    }
+
+    if (qcow2_need_accurate_refcounts(s)) {
+        ret = qcow2_cache_flush(bs, s->refcount_block_cache);
+        if (ret < 0) {
+            qemu_co_mutex_unlock(&s->lock);
+            return ret;
+        }
+    }
+    qemu_co_mutex_unlock(&s->lock);
+
+    return 0;
+}
+
+static int64_t qcow2_vm_state_offset(BDRVQcowState *s)
+{
+	return (int64_t)s->l1_vm_state_index << (s->cluster_bits + s->l2_bits);
+}
+
+static int qcow2_get_info(BlockDriverState *bs, BlockDriverInfo *bdi)
+{
+    BDRVQcowState *s = bs->opaque;
+    bdi->cluster_size = s->cluster_size;
+    bdi->vm_state_offset = qcow2_vm_state_offset(s);
+    return 0;
+}
+
+#if 0
+static void dump_refcounts(BlockDriverState *bs)
+{
+    BDRVQcowState *s = bs->opaque;
+    int64_t nb_clusters, k, k1, size;
+    int refcount;
+
+    size = bdrv_getlength(bs->file);
+    nb_clusters = size_to_clusters(s, size);
+    for(k = 0; k < nb_clusters;) {
+        k1 = k;
+        refcount = get_refcount(bs, k);
+        k++;
+        while (k < nb_clusters && get_refcount(bs, k) == refcount)
+            k++;
+        printf("%" PRId64 ": refcount=%d nb=%" PRId64 "\n", k, refcount,
+               k - k1);
+    }
+}
+#endif
+
+static int qcow2_save_vmstate(BlockDriverState *bs, QEMUIOVector *qiov,
+                              int64_t pos)
+{
+    BDRVQcowState *s = bs->opaque;
+    int growable = bs->growable;
+    int ret;
+
+    BLKDBG_EVENT(bs->file, BLKDBG_VMSTATE_SAVE);
+    bs->growable = 1;
+    ret = bdrv_pwritev(bs, qcow2_vm_state_offset(s) + pos, qiov);
+    bs->growable = growable;
+
+    return ret;
+}
+
+static int qcow2_load_vmstate(BlockDriverState *bs, uint8_t *buf,
+                              int64_t pos, int size)
+{
+    BDRVQcowState *s = bs->opaque;
+    int growable = bs->growable;
+    int ret;
+
+    BLKDBG_EVENT(bs->file, BLKDBG_VMSTATE_LOAD);
+    bs->growable = 1;
+    ret = bdrv_pread(bs, qcow2_vm_state_offset(s) + pos, buf, size);
+    bs->growable = growable;
+
+    return ret;
+}
+
+static QEMUOptionParameter qcow2_create_options[] = {
+    {
+        .name = BLOCK_OPT_SIZE,
+        .type = OPT_SIZE,
+        .help = "Virtual disk size"
+    },
+    {
+        .name = BLOCK_OPT_COMPAT_LEVEL,
+        .type = OPT_STRING,
+        .help = "Compatibility level (0.10 or 1.1)"
+    },
+    {
+        .name = BLOCK_OPT_BACKING_FILE,
+        .type = OPT_STRING,
+        .help = "File name of a base image"
+    },
+    {
+        .name = BLOCK_OPT_BACKING_FMT,
+        .type = OPT_STRING,
+        .help = "Image format of the base image"
+    },
+    {
+        .name = BLOCK_OPT_ENCRYPT,
+        .type = OPT_FLAG,
+        .help = "Encrypt the image"
+    },
+    {
+        .name = BLOCK_OPT_CLUSTER_SIZE,
+        .type = OPT_SIZE,
+        .help = "qcow2 cluster size",
+        .value = { .n = DEFAULT_CLUSTER_SIZE },
+    },
+    {
+        .name = BLOCK_OPT_PREALLOC,
+        .type = OPT_STRING,
+        .help = "Preallocation mode (allowed values: off, metadata)"
+    },
+    {
+        .name = BLOCK_OPT_LAZY_REFCOUNTS,
+        .type = OPT_FLAG,
+        .help = "Postpone refcount updates",
+    },
+    { NULL }
+};
+
+static BlockDriver bdrv_qcow2 = {
+    .format_name        = "qcow2",
+    .instance_size      = sizeof(BDRVQcowState),
+    .bdrv_probe         = qcow2_probe,
+    .bdrv_open          = qcow2_open,
+    .bdrv_close         = qcow2_close,
+    .bdrv_reopen_prepare  = qcow2_reopen_prepare,
+    .bdrv_create        = qcow2_create,
+    .bdrv_has_zero_init = bdrv_has_zero_init_1,
+    .bdrv_co_is_allocated = qcow2_co_is_allocated,
+    .bdrv_set_key       = qcow2_set_key,
+    .bdrv_make_empty    = qcow2_make_empty,
+
+    .bdrv_co_readv          = qcow2_co_readv,
+    .bdrv_co_writev         = qcow2_co_writev,
+    .bdrv_co_flush_to_os    = qcow2_co_flush_to_os,
+
+    .bdrv_co_write_zeroes   = qcow2_co_write_zeroes,
+    .bdrv_co_discard        = qcow2_co_discard,
+    .bdrv_truncate          = qcow2_truncate,
+    .bdrv_write_compressed  = qcow2_write_compressed,
+
+    .bdrv_snapshot_create   = qcow2_snapshot_create,
+    .bdrv_snapshot_goto     = qcow2_snapshot_goto,
+    .bdrv_snapshot_delete   = qcow2_snapshot_delete,
+    .bdrv_snapshot_list     = qcow2_snapshot_list,
+    .bdrv_snapshot_load_tmp     = qcow2_snapshot_load_tmp,
+    .bdrv_get_info      = qcow2_get_info,
+
+    .bdrv_save_vmstate    = qcow2_save_vmstate,
+    .bdrv_load_vmstate    = qcow2_load_vmstate,
+
+    .bdrv_change_backing_file   = qcow2_change_backing_file,
+
+    .bdrv_invalidate_cache      = qcow2_invalidate_cache,
+
+    .create_options = qcow2_create_options,
+    .bdrv_check = qcow2_check,
+};
+
+static void bdrv_qcow2_init(void)
+{
+    bdrv_register(&bdrv_qcow2);
+}
+
+block_init(bdrv_qcow2_init);
diff --git a/contrib/qemu/block/qcow2.h b/contrib/qemu/block/qcow2.h
new file mode 100644
index 000000000..3b2d5cda7
--- /dev/null
+++ b/contrib/qemu/block/qcow2.h
@@ -0,0 +1,437 @@
+/*
+ * Block driver for the QCOW version 2 format
+ *
+ * Copyright (c) 2004-2006 Fabrice Bellard
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#ifndef BLOCK_QCOW2_H
+#define BLOCK_QCOW2_H
+
+#include "qemu/aes.h"
+#include "block/coroutine.h"
+
+//#define DEBUG_ALLOC
+//#define DEBUG_ALLOC2
+//#define DEBUG_EXT
+
+#define QCOW_MAGIC (('Q' << 24) | ('F' << 16) | ('I' << 8) | 0xfb)
+
+#define QCOW_CRYPT_NONE 0
+#define QCOW_CRYPT_AES  1
+
+#define QCOW_MAX_CRYPT_CLUSTERS 32
+
+/* indicate that the refcount of the referenced cluster is exactly one. */
+#define QCOW_OFLAG_COPIED     (1LL << 63)
+/* indicate that the cluster is compressed (they never have the copied flag) */
+#define QCOW_OFLAG_COMPRESSED (1LL << 62)
+/* The cluster reads as all zeros */
+#define QCOW_OFLAG_ZERO (1LL << 0)
+
+#define REFCOUNT_SHIFT 1 /* refcount size is 2 bytes */
+
+#define MIN_CLUSTER_BITS 9
+#define MAX_CLUSTER_BITS 21
+
+#define L2_CACHE_SIZE 16
+
+/* Must be at least 4 to cover all cases of refcount table growth */
+#define REFCOUNT_CACHE_SIZE 4
+
+#define DEFAULT_CLUSTER_SIZE 65536
+
+
+#define QCOW2_OPT_LAZY_REFCOUNTS "lazy_refcounts"
+#define QCOW2_OPT_DISCARD_REQUEST "pass_discard_request"
+#define QCOW2_OPT_DISCARD_SNAPSHOT "pass_discard_snapshot"
+#define QCOW2_OPT_DISCARD_OTHER "pass_discard_other"
+
+typedef struct QCowHeader {
+    uint32_t magic;
+    uint32_t version;
+    uint64_t backing_file_offset;
+    uint32_t backing_file_size;
+    uint32_t cluster_bits;
+    uint64_t size; /* in bytes */
+    uint32_t crypt_method;
+    uint32_t l1_size; /* XXX: save number of clusters instead ? */
+    uint64_t l1_table_offset;
+    uint64_t refcount_table_offset;
+    uint32_t refcount_table_clusters;
+    uint32_t nb_snapshots;
+    uint64_t snapshots_offset;
+
+    /* The following fields are only valid for version >= 3 */
+    uint64_t incompatible_features;
+    uint64_t compatible_features;
+    uint64_t autoclear_features;
+
+    uint32_t refcount_order;
+    uint32_t header_length;
+} QCowHeader;
+
+typedef struct QCowSnapshot {
+    uint64_t l1_table_offset;
+    uint32_t l1_size;
+    char *id_str;
+    char *name;
+    uint64_t disk_size;
+    uint64_t vm_state_size;
+    uint32_t date_sec;
+    uint32_t date_nsec;
+    uint64_t vm_clock_nsec;
+} QCowSnapshot;
+
+struct Qcow2Cache;
+typedef struct Qcow2Cache Qcow2Cache;
+
+typedef struct Qcow2UnknownHeaderExtension {
+    uint32_t magic;
+    uint32_t len;
+    QLIST_ENTRY(Qcow2UnknownHeaderExtension) next;
+    uint8_t data[];
+} Qcow2UnknownHeaderExtension;
+
+enum {
+    QCOW2_FEAT_TYPE_INCOMPATIBLE    = 0,
+    QCOW2_FEAT_TYPE_COMPATIBLE      = 1,
+    QCOW2_FEAT_TYPE_AUTOCLEAR       = 2,
+};
+
+/* Incompatible feature bits */
+enum {
+    QCOW2_INCOMPAT_DIRTY_BITNR   = 0,
+    QCOW2_INCOMPAT_DIRTY         = 1 << QCOW2_INCOMPAT_DIRTY_BITNR,
+
+    QCOW2_INCOMPAT_MASK          = QCOW2_INCOMPAT_DIRTY,
+};
+
+/* Compatible feature bits */
+enum {
+    QCOW2_COMPAT_LAZY_REFCOUNTS_BITNR = 0,
+    QCOW2_COMPAT_LAZY_REFCOUNTS       = 1 << QCOW2_COMPAT_LAZY_REFCOUNTS_BITNR,
+
+    QCOW2_COMPAT_FEAT_MASK            = QCOW2_COMPAT_LAZY_REFCOUNTS,
+};
+
+enum qcow2_discard_type {
+    QCOW2_DISCARD_NEVER = 0,
+    QCOW2_DISCARD_ALWAYS,
+    QCOW2_DISCARD_REQUEST,
+    QCOW2_DISCARD_SNAPSHOT,
+    QCOW2_DISCARD_OTHER,
+    QCOW2_DISCARD_MAX
+};
+
+typedef struct Qcow2Feature {
+    uint8_t type;
+    uint8_t bit;
+    char    name[46];
+} QEMU_PACKED Qcow2Feature;
+
+typedef struct Qcow2DiscardRegion {
+    BlockDriverState *bs;
+    uint64_t offset;
+    uint64_t bytes;
+    QTAILQ_ENTRY(Qcow2DiscardRegion) next;
+} Qcow2DiscardRegion;
+
+typedef struct BDRVQcowState {
+    int cluster_bits;
+    int cluster_size;
+    int cluster_sectors;
+    int l2_bits;
+    int l2_size;
+    int l1_size;
+    int l1_vm_state_index;
+    int csize_shift;
+    int csize_mask;
+    uint64_t cluster_offset_mask;
+    uint64_t l1_table_offset;
+    uint64_t *l1_table;
+
+    Qcow2Cache* l2_table_cache;
+    Qcow2Cache* refcount_block_cache;
+
+    uint8_t *cluster_cache;
+    uint8_t *cluster_data;
+    uint64_t cluster_cache_offset;
+    QLIST_HEAD(QCowClusterAlloc, QCowL2Meta) cluster_allocs;
+
+    uint64_t *refcount_table;
+    uint64_t refcount_table_offset;
+    uint32_t refcount_table_size;
+    int64_t free_cluster_index;
+    int64_t free_byte_offset;
+
+    CoMutex lock;
+
+    uint32_t crypt_method; /* current crypt method, 0 if no key yet */
+    uint32_t crypt_method_header;
+    AES_KEY aes_encrypt_key;
+    AES_KEY aes_decrypt_key;
+    uint64_t snapshots_offset;
+    int snapshots_size;
+    int nb_snapshots;
+    QCowSnapshot *snapshots;
+
+    int flags;
+    int qcow_version;
+    bool use_lazy_refcounts;
+
+    bool discard_passthrough[QCOW2_DISCARD_MAX];
+
+    uint64_t incompatible_features;
+    uint64_t compatible_features;
+    uint64_t autoclear_features;
+
+    size_t unknown_header_fields_size;
+    void* unknown_header_fields;
+    QLIST_HEAD(, Qcow2UnknownHeaderExtension) unknown_header_ext;
+    QTAILQ_HEAD (, Qcow2DiscardRegion) discards;
+    bool cache_discards;
+} BDRVQcowState;
+
+/* XXX: use std qcow open function ? */
+typedef struct QCowCreateState {
+    int cluster_size;
+    int cluster_bits;
+    uint16_t *refcount_block;
+    uint64_t *refcount_table;
+    int64_t l1_table_offset;
+    int64_t refcount_table_offset;
+    int64_t refcount_block_offset;
+} QCowCreateState;
+
+struct QCowAIOCB;
+
+typedef struct Qcow2COWRegion {
+    /**
+     * Offset of the COW region in bytes from the start of the first cluster
+     * touched by the request.
+     */
+    uint64_t    offset;
+
+    /** Number of sectors to copy */
+    int         nb_sectors;
+} Qcow2COWRegion;
+
+/**
+ * Describes an in-flight (part of a) write request that writes to clusters
+ * that are not referenced in their L2 table yet.
+ */
+typedef struct QCowL2Meta
+{
+    /** Guest offset of the first newly allocated cluster */
+    uint64_t offset;
+
+    /** Host offset of the first newly allocated cluster */
+    uint64_t alloc_offset;
+
+    /**
+     * Number of sectors from the start of the first allocated cluster to
+     * the end of the (possibly shortened) request
+     */
+    int nb_available;
+
+    /** Number of newly allocated clusters */
+    int nb_clusters;
+
+    /**
+     * Requests that overlap with this allocation and wait to be restarted
+     * when the allocating request has completed.
+     */
+    CoQueue dependent_requests;
+
+    /**
+     * The COW Region between the start of the first allocated cluster and the
+     * area the guest actually writes to.
+     */
+    Qcow2COWRegion cow_start;
+
+    /**
+     * The COW Region between the area the guest actually writes to and the
+     * end of the last allocated cluster.
+     */
+    Qcow2COWRegion cow_end;
+
+    /** Pointer to next L2Meta of the same write request */
+    struct QCowL2Meta *next;
+
+    QLIST_ENTRY(QCowL2Meta) next_in_flight;
+} QCowL2Meta;
+
+enum {
+    QCOW2_CLUSTER_UNALLOCATED,
+    QCOW2_CLUSTER_NORMAL,
+    QCOW2_CLUSTER_COMPRESSED,
+    QCOW2_CLUSTER_ZERO
+};
+
+#define L1E_OFFSET_MASK 0x00ffffffffffff00ULL
+#define L2E_OFFSET_MASK 0x00ffffffffffff00ULL
+#define L2E_COMPRESSED_OFFSET_SIZE_MASK 0x3fffffffffffffffULL
+
+#define REFT_OFFSET_MASK 0xffffffffffffff00ULL
+
+static inline int64_t start_of_cluster(BDRVQcowState *s, int64_t offset)
+{
+    return offset & ~(s->cluster_size - 1);
+}
+
+static inline int64_t offset_into_cluster(BDRVQcowState *s, int64_t offset)
+{
+    return offset & (s->cluster_size - 1);
+}
+
+static inline int size_to_clusters(BDRVQcowState *s, int64_t size)
+{
+    return (size + (s->cluster_size - 1)) >> s->cluster_bits;
+}
+
+static inline int64_t size_to_l1(BDRVQcowState *s, int64_t size)
+{
+    int shift = s->cluster_bits + s->l2_bits;
+    return (size + (1ULL << shift) - 1) >> shift;
+}
+
+static inline int offset_to_l2_index(BDRVQcowState *s, int64_t offset)
+{
+    return (offset >> s->cluster_bits) & (s->l2_size - 1);
+}
+
+static inline int64_t align_offset(int64_t offset, int n)
+{
+    offset = (offset + n - 1) & ~(n - 1);
+    return offset;
+}
+
+static inline int qcow2_get_cluster_type(uint64_t l2_entry)
+{
+    if (l2_entry & QCOW_OFLAG_COMPRESSED) {
+        return QCOW2_CLUSTER_COMPRESSED;
+    } else if (l2_entry & QCOW_OFLAG_ZERO) {
+        return QCOW2_CLUSTER_ZERO;
+    } else if (!(l2_entry & L2E_OFFSET_MASK)) {
+        return QCOW2_CLUSTER_UNALLOCATED;
+    } else {
+        return QCOW2_CLUSTER_NORMAL;
+    }
+}
+
+/* Check whether refcounts are eager or lazy */
+static inline bool qcow2_need_accurate_refcounts(BDRVQcowState *s)
+{
+    return !(s->incompatible_features & QCOW2_INCOMPAT_DIRTY);
+}
+
+static inline uint64_t l2meta_cow_start(QCowL2Meta *m)
+{
+    return m->offset + m->cow_start.offset;
+}
+
+static inline uint64_t l2meta_cow_end(QCowL2Meta *m)
+{
+    return m->offset + m->cow_end.offset
+        + (m->cow_end.nb_sectors << BDRV_SECTOR_BITS);
+}
+
+// FIXME Need qcow2_ prefix to global functions
+
+/* qcow2.c functions */
+int qcow2_backing_read1(BlockDriverState *bs, QEMUIOVector *qiov,
+                  int64_t sector_num, int nb_sectors);
+
+int qcow2_mark_dirty(BlockDriverState *bs);
+int qcow2_update_header(BlockDriverState *bs);
+
+/* qcow2-refcount.c functions */
+int qcow2_refcount_init(BlockDriverState *bs);
+void qcow2_refcount_close(BlockDriverState *bs);
+
+int64_t qcow2_alloc_clusters(BlockDriverState *bs, int64_t size);
+int qcow2_alloc_clusters_at(BlockDriverState *bs, uint64_t offset,
+    int nb_clusters);
+int64_t qcow2_alloc_bytes(BlockDriverState *bs, int size);
+void qcow2_free_clusters(BlockDriverState *bs,
+                          int64_t offset, int64_t size,
+                          enum qcow2_discard_type type);
+void qcow2_free_any_clusters(BlockDriverState *bs, uint64_t l2_entry,
+                             int nb_clusters, enum qcow2_discard_type type);
+
+int qcow2_update_snapshot_refcount(BlockDriverState *bs,
+    int64_t l1_table_offset, int l1_size, int addend);
+
+int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
+                          BdrvCheckMode fix);
+
+void qcow2_process_discards(BlockDriverState *bs, int ret);
+
+/* qcow2-cluster.c functions */
+int qcow2_grow_l1_table(BlockDriverState *bs, uint64_t min_size,
+                        bool exact_size);
+void qcow2_l2_cache_reset(BlockDriverState *bs);
+int qcow2_decompress_cluster(BlockDriverState *bs, uint64_t cluster_offset);
+void qcow2_encrypt_sectors(BDRVQcowState *s, int64_t sector_num,
+                     uint8_t *out_buf, const uint8_t *in_buf,
+                     int nb_sectors, int enc,
+                     const AES_KEY *key);
+
+int qcow2_get_cluster_offset(BlockDriverState *bs, uint64_t offset,
+    int *num, uint64_t *cluster_offset);
+int qcow2_alloc_cluster_offset(BlockDriverState *bs, uint64_t offset,
+    int n_start, int n_end, int *num, uint64_t *host_offset, QCowL2Meta **m);
+uint64_t qcow2_alloc_compressed_cluster_offset(BlockDriverState *bs,
+                                         uint64_t offset,
+                                         int compressed_size);
+
+int qcow2_alloc_cluster_link_l2(BlockDriverState *bs, QCowL2Meta *m);
+int qcow2_discard_clusters(BlockDriverState *bs, uint64_t offset,
+    int nb_sectors);
+int qcow2_zero_clusters(BlockDriverState *bs, uint64_t offset, int nb_sectors);
+
+/* qcow2-snapshot.c functions */
+int qcow2_snapshot_create(BlockDriverState *bs, QEMUSnapshotInfo *sn_info);
+int qcow2_snapshot_goto(BlockDriverState *bs, const char *snapshot_id);
+int qcow2_snapshot_delete(BlockDriverState *bs, const char *snapshot_id);
+int qcow2_snapshot_list(BlockDriverState *bs, QEMUSnapshotInfo **psn_tab);
+int qcow2_snapshot_load_tmp(BlockDriverState *bs, const char *snapshot_name);
+
+void qcow2_free_snapshots(BlockDriverState *bs);
+int qcow2_read_snapshots(BlockDriverState *bs);
+
+/* qcow2-cache.c functions */
+Qcow2Cache *qcow2_cache_create(BlockDriverState *bs, int num_tables);
+int qcow2_cache_destroy(BlockDriverState* bs, Qcow2Cache *c);
+
+void qcow2_cache_entry_mark_dirty(Qcow2Cache *c, void *table);
+int qcow2_cache_flush(BlockDriverState *bs, Qcow2Cache *c);
+int qcow2_cache_set_dependency(BlockDriverState *bs, Qcow2Cache *c,
+    Qcow2Cache *dependency);
+void qcow2_cache_depends_on_flush(Qcow2Cache *c);
+
+int qcow2_cache_get(BlockDriverState *bs, Qcow2Cache *c, uint64_t offset,
+    void **table);
+int qcow2_cache_get_empty(BlockDriverState *bs, Qcow2Cache *c, uint64_t offset,
+    void **table);
+int qcow2_cache_put(BlockDriverState *bs, Qcow2Cache *c, void **table);
+
+#endif
diff --git a/contrib/qemu/block/qed-check.c b/contrib/qemu/block/qed-check.c
new file mode 100644
index 000000000..b473dcd61
--- /dev/null
+++ b/contrib/qemu/block/qed-check.c
@@ -0,0 +1,248 @@
+/*
+ * QEMU Enhanced Disk Format Consistency Check
+ *
+ * Copyright IBM, Corp. 2010
+ *
+ * Authors:
+ *  Stefan Hajnoczi   <stefanha@linux.vnet.ibm.com>
+ *
+ * This work is licensed under the terms of the GNU LGPL, version 2 or later.
+ * See the COPYING.LIB file in the top-level directory.
+ *
+ */
+
+#include "qed.h"
+
+typedef struct {
+    BDRVQEDState *s;
+    BdrvCheckResult *result;
+    bool fix;                           /* whether to fix invalid offsets */
+
+    uint64_t nclusters;
+    uint32_t *used_clusters;            /* referenced cluster bitmap */
+
+    QEDRequest request;
+} QEDCheck;
+
+static bool qed_test_bit(uint32_t *bitmap, uint64_t n) {
+    return !!(bitmap[n / 32] & (1 << (n % 32)));
+}
+
+static void qed_set_bit(uint32_t *bitmap, uint64_t n) {
+    bitmap[n / 32] |= 1 << (n % 32);
+}
+
+/**
+ * Set bitmap bits for clusters
+ *
+ * @check:          Check structure
+ * @offset:         Starting offset in bytes
+ * @n:              Number of clusters
+ */
+static bool qed_set_used_clusters(QEDCheck *check, uint64_t offset,
+                                  unsigned int n)
+{
+    uint64_t cluster = qed_bytes_to_clusters(check->s, offset);
+    unsigned int corruptions = 0;
+
+    while (n-- != 0) {
+        /* Clusters should only be referenced once */
+        if (qed_test_bit(check->used_clusters, cluster)) {
+            corruptions++;
+        }
+
+        qed_set_bit(check->used_clusters, cluster);
+        cluster++;
+    }
+
+    check->result->corruptions += corruptions;
+    return corruptions == 0;
+}
+
+/**
+ * Check an L2 table
+ *
+ * @ret:            Number of invalid cluster offsets
+ */
+static unsigned int qed_check_l2_table(QEDCheck *check, QEDTable *table)
+{
+    BDRVQEDState *s = check->s;
+    unsigned int i, num_invalid = 0;
+    uint64_t last_offset = 0;
+
+    for (i = 0; i < s->table_nelems; i++) {
+        uint64_t offset = table->offsets[i];
+
+        if (qed_offset_is_unalloc_cluster(offset) ||
+            qed_offset_is_zero_cluster(offset)) {
+            continue;
+        }
+        check->result->bfi.allocated_clusters++;
+        if (last_offset && (last_offset + s->header.cluster_size != offset)) {
+            check->result->bfi.fragmented_clusters++;
+        }
+        last_offset = offset;
+
+        /* Detect invalid cluster offset */
+        if (!qed_check_cluster_offset(s, offset)) {
+            if (check->fix) {
+                table->offsets[i] = 0;
+                check->result->corruptions_fixed++;
+            } else {
+                check->result->corruptions++;
+            }
+
+            num_invalid++;
+            continue;
+        }
+
+        qed_set_used_clusters(check, offset, 1);
+    }
+
+    return num_invalid;
+}
+
+/**
+ * Descend tables and check each cluster is referenced once only
+ */
+static int qed_check_l1_table(QEDCheck *check, QEDTable *table)
+{
+    BDRVQEDState *s = check->s;
+    unsigned int i, num_invalid_l1 = 0;
+    int ret, last_error = 0;
+
+    /* Mark L1 table clusters used */
+    qed_set_used_clusters(check, s->header.l1_table_offset,
+                          s->header.table_size);
+
+    for (i = 0; i < s->table_nelems; i++) {
+        unsigned int num_invalid_l2;
+        uint64_t offset = table->offsets[i];
+
+        if (qed_offset_is_unalloc_cluster(offset)) {
+            continue;
+        }
+
+        /* Detect invalid L2 offset */
+        if (!qed_check_table_offset(s, offset)) {
+            /* Clear invalid offset */
+            if (check->fix) {
+                table->offsets[i] = 0;
+                check->result->corruptions_fixed++;
+            } else {
+                check->result->corruptions++;
+            }
+
+            num_invalid_l1++;
+            continue;
+        }
+
+        if (!qed_set_used_clusters(check, offset, s->header.table_size)) {
+            continue; /* skip an invalid table */
+        }
+
+        ret = qed_read_l2_table_sync(s, &check->request, offset);
+        if (ret) {
+            check->result->check_errors++;
+            last_error = ret;
+            continue;
+        }
+
+        num_invalid_l2 = qed_check_l2_table(check,
+                                            check->request.l2_table->table);
+
+        /* Write out fixed L2 table */
+        if (num_invalid_l2 > 0 && check->fix) {
+            ret = qed_write_l2_table_sync(s, &check->request, 0,
+                                          s->table_nelems, false);
+            if (ret) {
+                check->result->check_errors++;
+                last_error = ret;
+                continue;
+            }
+        }
+    }
+
+    /* Drop reference to final table */
+    qed_unref_l2_cache_entry(check->request.l2_table);
+    check->request.l2_table = NULL;
+
+    /* Write out fixed L1 table */
+    if (num_invalid_l1 > 0 && check->fix) {
+        ret = qed_write_l1_table_sync(s, 0, s->table_nelems);
+        if (ret) {
+            check->result->check_errors++;
+            last_error = ret;
+        }
+    }
+
+    return last_error;
+}
+
+/**
+ * Check for unreferenced (leaked) clusters
+ */
+static void qed_check_for_leaks(QEDCheck *check)
+{
+    BDRVQEDState *s = check->s;
+    uint64_t i;
+
+    for (i = s->header.header_size; i < check->nclusters; i++) {
+        if (!qed_test_bit(check->used_clusters, i)) {
+            check->result->leaks++;
+        }
+    }
+}
+
+/**
+ * Mark an image clean once it passes check or has been repaired
+ */
+static void qed_check_mark_clean(BDRVQEDState *s, BdrvCheckResult *result)
+{
+    /* Skip if there were unfixable corruptions or I/O errors */
+    if (result->corruptions > 0 || result->check_errors > 0) {
+        return;
+    }
+
+    /* Skip if image is already marked clean */
+    if (!(s->header.features & QED_F_NEED_CHECK)) {
+        return;
+    }
+
+    /* Ensure fixes reach storage before clearing check bit */
+    bdrv_flush(s->bs);
+
+    s->header.features &= ~QED_F_NEED_CHECK;
+    qed_write_header_sync(s);
+}
+
+int qed_check(BDRVQEDState *s, BdrvCheckResult *result, bool fix)
+{
+    QEDCheck check = {
+        .s = s,
+        .result = result,
+        .nclusters = qed_bytes_to_clusters(s, s->file_size),
+        .request = { .l2_table = NULL },
+        .fix = fix,
+    };
+    int ret;
+
+    check.used_clusters = g_malloc0(((check.nclusters + 31) / 32) *
+                                       sizeof(check.used_clusters[0]));
+
+    check.result->bfi.total_clusters =
+        (s->header.image_size + s->header.cluster_size - 1) /
+            s->header.cluster_size;
+    ret = qed_check_l1_table(&check, s->l1_table);
+    if (ret == 0) {
+        /* Only check for leaks if entire image was scanned successfully */
+        qed_check_for_leaks(&check);
+
+        if (fix) {
+            qed_check_mark_clean(s, result);
+        }
+    }
+
+    g_free(check.used_clusters);
+    return ret;
+}
diff --git a/contrib/qemu/block/qed-cluster.c b/contrib/qemu/block/qed-cluster.c
new file mode 100644
index 000000000..f64b2af8f
--- /dev/null
+++ b/contrib/qemu/block/qed-cluster.c
@@ -0,0 +1,165 @@
+/*
+ * QEMU Enhanced Disk Format Cluster functions
+ *
+ * Copyright IBM, Corp. 2010
+ *
+ * Authors:
+ *  Stefan Hajnoczi   <stefanha@linux.vnet.ibm.com>
+ *  Anthony Liguori   <aliguori@us.ibm.com>
+ *
+ * This work is licensed under the terms of the GNU LGPL, version 2 or later.
+ * See the COPYING.LIB file in the top-level directory.
+ *
+ */
+
+#include "qed.h"
+
+/**
+ * Count the number of contiguous data clusters
+ *
+ * @s:              QED state
+ * @table:          L2 table
+ * @index:          First cluster index
+ * @n:              Maximum number of clusters
+ * @offset:         Set to first cluster offset
+ *
+ * This function scans tables for contiguous clusters.  A contiguous run of
+ * clusters may be allocated, unallocated, or zero.
+ */
+static unsigned int qed_count_contiguous_clusters(BDRVQEDState *s,
+                                                  QEDTable *table,
+                                                  unsigned int index,
+                                                  unsigned int n,
+                                                  uint64_t *offset)
+{
+    unsigned int end = MIN(index + n, s->table_nelems);
+    uint64_t last = table->offsets[index];
+    unsigned int i;
+
+    *offset = last;
+
+    for (i = index + 1; i < end; i++) {
+        if (qed_offset_is_unalloc_cluster(last)) {
+            /* Counting unallocated clusters */
+            if (!qed_offset_is_unalloc_cluster(table->offsets[i])) {
+                break;
+            }
+        } else if (qed_offset_is_zero_cluster(last)) {
+            /* Counting zero clusters */
+            if (!qed_offset_is_zero_cluster(table->offsets[i])) {
+                break;
+            }
+        } else {
+            /* Counting allocated clusters */
+            if (table->offsets[i] != last + s->header.cluster_size) {
+                break;
+            }
+            last = table->offsets[i];
+        }
+    }
+    return i - index;
+}
+
+typedef struct {
+    BDRVQEDState *s;
+    uint64_t pos;
+    size_t len;
+
+    QEDRequest *request;
+
+    /* User callback */
+    QEDFindClusterFunc *cb;
+    void *opaque;
+} QEDFindClusterCB;
+
+static void qed_find_cluster_cb(void *opaque, int ret)
+{
+    QEDFindClusterCB *find_cluster_cb = opaque;
+    BDRVQEDState *s = find_cluster_cb->s;
+    QEDRequest *request = find_cluster_cb->request;
+    uint64_t offset = 0;
+    size_t len = 0;
+    unsigned int index;
+    unsigned int n;
+
+    if (ret) {
+        goto out;
+    }
+
+    index = qed_l2_index(s, find_cluster_cb->pos);
+    n = qed_bytes_to_clusters(s,
+                              qed_offset_into_cluster(s, find_cluster_cb->pos) +
+                              find_cluster_cb->len);
+    n = qed_count_contiguous_clusters(s, request->l2_table->table,
+                                      index, n, &offset);
+
+    if (qed_offset_is_unalloc_cluster(offset)) {
+        ret = QED_CLUSTER_L2;
+    } else if (qed_offset_is_zero_cluster(offset)) {
+        ret = QED_CLUSTER_ZERO;
+    } else if (qed_check_cluster_offset(s, offset)) {
+        ret = QED_CLUSTER_FOUND;
+    } else {
+        ret = -EINVAL;
+    }
+
+    len = MIN(find_cluster_cb->len, n * s->header.cluster_size -
+              qed_offset_into_cluster(s, find_cluster_cb->pos));
+
+out:
+    find_cluster_cb->cb(find_cluster_cb->opaque, ret, offset, len);
+    g_free(find_cluster_cb);
+}
+
+/**
+ * Find the offset of a data cluster
+ *
+ * @s:          QED state
+ * @request:    L2 cache entry
+ * @pos:        Byte position in device
+ * @len:        Number of bytes
+ * @cb:         Completion function
+ * @opaque:     User data for completion function
+ *
+ * This function translates a position in the block device to an offset in the
+ * image file.  It invokes the cb completion callback to report back the
+ * translated offset or unallocated range in the image file.
+ *
+ * If the L2 table exists, request->l2_table points to the L2 table cache entry
+ * and the caller must free the reference when they are finished.  The cache
+ * entry is exposed in this way to avoid callers having to read the L2 table
+ * again later during request processing.  If request->l2_table is non-NULL it
+ * will be unreferenced before taking on the new cache entry.
+ */
+void qed_find_cluster(BDRVQEDState *s, QEDRequest *request, uint64_t pos,
+                      size_t len, QEDFindClusterFunc *cb, void *opaque)
+{
+    QEDFindClusterCB *find_cluster_cb;
+    uint64_t l2_offset;
+
+    /* Limit length to L2 boundary.  Requests are broken up at the L2 boundary
+     * so that a request acts on one L2 table at a time.
+     */
+    len = MIN(len, (((pos >> s->l1_shift) + 1) << s->l1_shift) - pos);
+
+    l2_offset = s->l1_table->offsets[qed_l1_index(s, pos)];
+    if (qed_offset_is_unalloc_cluster(l2_offset)) {
+        cb(opaque, QED_CLUSTER_L1, 0, len);
+        return;
+    }
+    if (!qed_check_table_offset(s, l2_offset)) {
+        cb(opaque, -EINVAL, 0, 0);
+        return;
+    }
+
+    find_cluster_cb = g_malloc(sizeof(*find_cluster_cb));
+    find_cluster_cb->s = s;
+    find_cluster_cb->pos = pos;
+    find_cluster_cb->len = len;
+    find_cluster_cb->cb = cb;
+    find_cluster_cb->opaque = opaque;
+    find_cluster_cb->request = request;
+
+    qed_read_l2_table(s, request, l2_offset,
+                      qed_find_cluster_cb, find_cluster_cb);
+}
diff --git a/contrib/qemu/block/qed-gencb.c b/contrib/qemu/block/qed-gencb.c
new file mode 100644
index 000000000..7d7ac1ffc
--- /dev/null
+++ b/contrib/qemu/block/qed-gencb.c
@@ -0,0 +1,32 @@
+/*
+ * QEMU Enhanced Disk Format
+ *
+ * Copyright IBM, Corp. 2010
+ *
+ * Authors:
+ *  Stefan Hajnoczi   <stefanha@linux.vnet.ibm.com>
+ *
+ * This work is licensed under the terms of the GNU LGPL, version 2 or later.
+ * See the COPYING.LIB file in the top-level directory.
+ *
+ */
+
+#include "qed.h"
+
+void *gencb_alloc(size_t len, BlockDriverCompletionFunc *cb, void *opaque)
+{
+    GenericCB *gencb = g_malloc(len);
+    gencb->cb = cb;
+    gencb->opaque = opaque;
+    return gencb;
+}
+
+void gencb_complete(void *opaque, int ret)
+{
+    GenericCB *gencb = opaque;
+    BlockDriverCompletionFunc *cb = gencb->cb;
+    void *user_opaque = gencb->opaque;
+
+    g_free(gencb);
+    cb(user_opaque, ret);
+}
diff --git a/contrib/qemu/block/qed-l2-cache.c b/contrib/qemu/block/qed-l2-cache.c
new file mode 100644
index 000000000..e9b2aae44
--- /dev/null
+++ b/contrib/qemu/block/qed-l2-cache.c
@@ -0,0 +1,187 @@
+/*
+ * QEMU Enhanced Disk Format L2 Cache
+ *
+ * Copyright IBM, Corp. 2010
+ *
+ * Authors:
+ *  Anthony Liguori   <aliguori@us.ibm.com>
+ *
+ * This work is licensed under the terms of the GNU LGPL, version 2 or later.
+ * See the COPYING.LIB file in the top-level directory.
+ *
+ */
+
+/*
+ * L2 table cache usage is as follows:
+ *
+ * An open image has one L2 table cache that is used to avoid accessing the
+ * image file for recently referenced L2 tables.
+ *
+ * Cluster offset lookup translates the logical offset within the block device
+ * to a cluster offset within the image file.  This is done by indexing into
+ * the L1 and L2 tables which store cluster offsets.  It is here where the L2
+ * table cache serves up recently referenced L2 tables.
+ *
+ * If there is a cache miss, that L2 table is read from the image file and
+ * committed to the cache.  Subsequent accesses to that L2 table will be served
+ * from the cache until the table is evicted from the cache.
+ *
+ * L2 tables are also committed to the cache when new L2 tables are allocated
+ * in the image file.  Since the L2 table cache is write-through, the new L2
+ * table is first written out to the image file and then committed to the
+ * cache.
+ *
+ * Multiple I/O requests may be using an L2 table cache entry at any given
+ * time.  That means an entry may be in use across several requests and
+ * reference counting is needed to free the entry at the correct time.  In
+ * particular, an entry evicted from the cache will only be freed once all
+ * references are dropped.
+ *
+ * An in-flight I/O request will hold a reference to a L2 table cache entry for
+ * the period during which it needs to access the L2 table.  This includes
+ * cluster offset lookup, L2 table allocation, and L2 table update when a new
+ * data cluster has been allocated.
+ *
+ * An interesting case occurs when two requests need to access an L2 table that
+ * is not in the cache.  Since the operation to read the table from the image
+ * file takes some time to complete, both requests may see a cache miss and
+ * start reading the L2 table from the image file.  The first to finish will
+ * commit its L2 table into the cache.  When the second tries to commit its
+ * table will be deleted in favor of the existing cache entry.
+ */
+
+#include "trace.h"
+#include "qed.h"
+
+/* Each L2 holds 2GB so this let's us fully cache a 100GB disk */
+#define MAX_L2_CACHE_SIZE 50
+
+/**
+ * Initialize the L2 cache
+ */
+void qed_init_l2_cache(L2TableCache *l2_cache)
+{
+    QTAILQ_INIT(&l2_cache->entries);
+    l2_cache->n_entries = 0;
+}
+
+/**
+ * Free the L2 cache
+ */
+void qed_free_l2_cache(L2TableCache *l2_cache)
+{
+    CachedL2Table *entry, *next_entry;
+
+    QTAILQ_FOREACH_SAFE(entry, &l2_cache->entries, node, next_entry) {
+        qemu_vfree(entry->table);
+        g_free(entry);
+    }
+}
+
+/**
+ * Allocate an uninitialized entry from the cache
+ *
+ * The returned entry has a reference count of 1 and is owned by the caller.
+ * The caller must allocate the actual table field for this entry and it must
+ * be freeable using qemu_vfree().
+ */
+CachedL2Table *qed_alloc_l2_cache_entry(L2TableCache *l2_cache)
+{
+    CachedL2Table *entry;
+
+    entry = g_malloc0(sizeof(*entry));
+    entry->ref++;
+
+    trace_qed_alloc_l2_cache_entry(l2_cache, entry);
+
+    return entry;
+}
+
+/**
+ * Decrease an entry's reference count and free if necessary when the reference
+ * count drops to zero.
+ */
+void qed_unref_l2_cache_entry(CachedL2Table *entry)
+{
+    if (!entry) {
+        return;
+    }
+
+    entry->ref--;
+    trace_qed_unref_l2_cache_entry(entry, entry->ref);
+    if (entry->ref == 0) {
+        qemu_vfree(entry->table);
+        g_free(entry);
+    }
+}
+
+/**
+ * Find an entry in the L2 cache.  This may return NULL and it's up to the
+ * caller to satisfy the cache miss.
+ *
+ * For a cached entry, this function increases the reference count and returns
+ * the entry.
+ */
+CachedL2Table *qed_find_l2_cache_entry(L2TableCache *l2_cache, uint64_t offset)
+{
+    CachedL2Table *entry;
+
+    QTAILQ_FOREACH(entry, &l2_cache->entries, node) {
+        if (entry->offset == offset) {
+            trace_qed_find_l2_cache_entry(l2_cache, entry, offset, entry->ref);
+            entry->ref++;
+            return entry;
+        }
+    }
+    return NULL;
+}
+
+/**
+ * Commit an L2 cache entry into the cache.  This is meant to be used as part of
+ * the process to satisfy a cache miss.  A caller would allocate an entry which
+ * is not actually in the L2 cache and then once the entry was valid and
+ * present on disk, the entry can be committed into the cache.
+ *
+ * Since the cache is write-through, it's important that this function is not
+ * called until the entry is present on disk and the L1 has been updated to
+ * point to the entry.
+ *
+ * N.B. This function steals a reference to the l2_table from the caller so the
+ * caller must obtain a new reference by issuing a call to
+ * qed_find_l2_cache_entry().
+ */
+void qed_commit_l2_cache_entry(L2TableCache *l2_cache, CachedL2Table *l2_table)
+{
+    CachedL2Table *entry;
+
+    entry = qed_find_l2_cache_entry(l2_cache, l2_table->offset);
+    if (entry) {
+        qed_unref_l2_cache_entry(entry);
+        qed_unref_l2_cache_entry(l2_table);
+        return;
+    }
+
+    /* Evict an unused cache entry so we have space.  If all entries are in use
+     * we can grow the cache temporarily and we try to shrink back down later.
+     */
+    if (l2_cache->n_entries >= MAX_L2_CACHE_SIZE) {
+        CachedL2Table *next;
+        QTAILQ_FOREACH_SAFE(entry, &l2_cache->entries, node, next) {
+            if (entry->ref > 1) {
+                continue;
+            }
+
+            QTAILQ_REMOVE(&l2_cache->entries, entry, node);
+            l2_cache->n_entries--;
+            qed_unref_l2_cache_entry(entry);
+
+            /* Stop evicting when we've shrunk back to max size */
+            if (l2_cache->n_entries < MAX_L2_CACHE_SIZE) {
+                break;
+            }
+        }
+    }
+
+    l2_cache->n_entries++;
+    QTAILQ_INSERT_TAIL(&l2_cache->entries, l2_table, node);
+}
diff --git a/contrib/qemu/block/qed-table.c b/contrib/qemu/block/qed-table.c
new file mode 100644
index 000000000..76d2dcccf
--- /dev/null
+++ b/contrib/qemu/block/qed-table.c
@@ -0,0 +1,296 @@
+/*
+ * QEMU Enhanced Disk Format Table I/O
+ *
+ * Copyright IBM, Corp. 2010
+ *
+ * Authors:
+ *  Stefan Hajnoczi   <stefanha@linux.vnet.ibm.com>
+ *  Anthony Liguori   <aliguori@us.ibm.com>
+ *
+ * This work is licensed under the terms of the GNU LGPL, version 2 or later.
+ * See the COPYING.LIB file in the top-level directory.
+ *
+ */
+
+#include "trace.h"
+#include "qemu/sockets.h" /* for EINPROGRESS on Windows */
+#include "qed.h"
+
+typedef struct {
+    GenericCB gencb;
+    BDRVQEDState *s;
+    QEDTable *table;
+
+    struct iovec iov;
+    QEMUIOVector qiov;
+} QEDReadTableCB;
+
+static void qed_read_table_cb(void *opaque, int ret)
+{
+    QEDReadTableCB *read_table_cb = opaque;
+    QEDTable *table = read_table_cb->table;
+    int noffsets = read_table_cb->qiov.size / sizeof(uint64_t);
+    int i;
+
+    /* Handle I/O error */
+    if (ret) {
+        goto out;
+    }
+
+    /* Byteswap offsets */
+    for (i = 0; i < noffsets; i++) {
+        table->offsets[i] = le64_to_cpu(table->offsets[i]);
+    }
+
+out:
+    /* Completion */
+    trace_qed_read_table_cb(read_table_cb->s, read_table_cb->table, ret);
+    gencb_complete(&read_table_cb->gencb, ret);
+}
+
+static void qed_read_table(BDRVQEDState *s, uint64_t offset, QEDTable *table,
+                           BlockDriverCompletionFunc *cb, void *opaque)
+{
+    QEDReadTableCB *read_table_cb = gencb_alloc(sizeof(*read_table_cb),
+                                                cb, opaque);
+    QEMUIOVector *qiov = &read_table_cb->qiov;
+
+    trace_qed_read_table(s, offset, table);
+
+    read_table_cb->s = s;
+    read_table_cb->table = table;
+    read_table_cb->iov.iov_base = table->offsets,
+    read_table_cb->iov.iov_len = s->header.cluster_size * s->header.table_size,
+
+    qemu_iovec_init_external(qiov, &read_table_cb->iov, 1);
+    bdrv_aio_readv(s->bs->file, offset / BDRV_SECTOR_SIZE, qiov,
+                   qiov->size / BDRV_SECTOR_SIZE,
+                   qed_read_table_cb, read_table_cb);
+}
+
+typedef struct {
+    GenericCB gencb;
+    BDRVQEDState *s;
+    QEDTable *orig_table;
+    QEDTable *table;
+    bool flush;             /* flush after write? */
+
+    struct iovec iov;
+    QEMUIOVector qiov;
+} QEDWriteTableCB;
+
+static void qed_write_table_cb(void *opaque, int ret)
+{
+    QEDWriteTableCB *write_table_cb = opaque;
+
+    trace_qed_write_table_cb(write_table_cb->s,
+                             write_table_cb->orig_table,
+                             write_table_cb->flush,
+                             ret);
+
+    if (ret) {
+        goto out;
+    }
+
+    if (write_table_cb->flush) {
+        /* We still need to flush first */
+        write_table_cb->flush = false;
+        bdrv_aio_flush(write_table_cb->s->bs, qed_write_table_cb,
+                       write_table_cb);
+        return;
+    }
+
+out:
+    qemu_vfree(write_table_cb->table);
+    gencb_complete(&write_table_cb->gencb, ret);
+}
+
+/**
+ * Write out an updated part or all of a table
+ *
+ * @s:          QED state
+ * @offset:     Offset of table in image file, in bytes
+ * @table:      Table
+ * @index:      Index of first element
+ * @n:          Number of elements
+ * @flush:      Whether or not to sync to disk
+ * @cb:         Completion function
+ * @opaque:     Argument for completion function
+ */
+static void qed_write_table(BDRVQEDState *s, uint64_t offset, QEDTable *table,
+                            unsigned int index, unsigned int n, bool flush,
+                            BlockDriverCompletionFunc *cb, void *opaque)
+{
+    QEDWriteTableCB *write_table_cb;
+    unsigned int sector_mask = BDRV_SECTOR_SIZE / sizeof(uint64_t) - 1;
+    unsigned int start, end, i;
+    size_t len_bytes;
+
+    trace_qed_write_table(s, offset, table, index, n);
+
+    /* Calculate indices of the first and one after last elements */
+    start = index & ~sector_mask;
+    end = (index + n + sector_mask) & ~sector_mask;
+
+    len_bytes = (end - start) * sizeof(uint64_t);
+
+    write_table_cb = gencb_alloc(sizeof(*write_table_cb), cb, opaque);
+    write_table_cb->s = s;
+    write_table_cb->orig_table = table;
+    write_table_cb->flush = flush;
+    write_table_cb->table = qemu_blockalign(s->bs, len_bytes);
+    write_table_cb->iov.iov_base = write_table_cb->table->offsets;
+    write_table_cb->iov.iov_len = len_bytes;
+    qemu_iovec_init_external(&write_table_cb->qiov, &write_table_cb->iov, 1);
+
+    /* Byteswap table */
+    for (i = start; i < end; i++) {
+        uint64_t le_offset = cpu_to_le64(table->offsets[i]);
+        write_table_cb->table->offsets[i - start] = le_offset;
+    }
+
+    /* Adjust for offset into table */
+    offset += start * sizeof(uint64_t);
+
+    bdrv_aio_writev(s->bs->file, offset / BDRV_SECTOR_SIZE,
+                    &write_table_cb->qiov,
+                    write_table_cb->qiov.size / BDRV_SECTOR_SIZE,
+                    qed_write_table_cb, write_table_cb);
+}
+
+/**
+ * Propagate return value from async callback
+ */
+static void qed_sync_cb(void *opaque, int ret)
+{
+    *(int *)opaque = ret;
+}
+
+int qed_read_l1_table_sync(BDRVQEDState *s)
+{
+    int ret = -EINPROGRESS;
+
+    qed_read_table(s, s->header.l1_table_offset,
+                   s->l1_table, qed_sync_cb, &ret);
+    while (ret == -EINPROGRESS) {
+        qemu_aio_wait();
+    }
+
+    return ret;
+}
+
+void qed_write_l1_table(BDRVQEDState *s, unsigned int index, unsigned int n,
+                        BlockDriverCompletionFunc *cb, void *opaque)
+{
+    BLKDBG_EVENT(s->bs->file, BLKDBG_L1_UPDATE);
+    qed_write_table(s, s->header.l1_table_offset,
+                    s->l1_table, index, n, false, cb, opaque);
+}
+
+int qed_write_l1_table_sync(BDRVQEDState *s, unsigned int index,
+                            unsigned int n)
+{
+    int ret = -EINPROGRESS;
+
+    qed_write_l1_table(s, index, n, qed_sync_cb, &ret);
+    while (ret == -EINPROGRESS) {
+        qemu_aio_wait();
+    }
+
+    return ret;
+}
+
+typedef struct {
+    GenericCB gencb;
+    BDRVQEDState *s;
+    uint64_t l2_offset;
+    QEDRequest *request;
+} QEDReadL2TableCB;
+
+static void qed_read_l2_table_cb(void *opaque, int ret)
+{
+    QEDReadL2TableCB *read_l2_table_cb = opaque;
+    QEDRequest *request = read_l2_table_cb->request;
+    BDRVQEDState *s = read_l2_table_cb->s;
+    CachedL2Table *l2_table = request->l2_table;
+    uint64_t l2_offset = read_l2_table_cb->l2_offset;
+
+    if (ret) {
+        /* can't trust loaded L2 table anymore */
+        qed_unref_l2_cache_entry(l2_table);
+        request->l2_table = NULL;
+    } else {
+        l2_table->offset = l2_offset;
+
+        qed_commit_l2_cache_entry(&s->l2_cache, l2_table);
+
+        /* This is guaranteed to succeed because we just committed the entry
+         * to the cache.
+         */
+        request->l2_table = qed_find_l2_cache_entry(&s->l2_cache, l2_offset);
+        assert(request->l2_table != NULL);
+    }
+
+    gencb_complete(&read_l2_table_cb->gencb, ret);
+}
+
+void qed_read_l2_table(BDRVQEDState *s, QEDRequest *request, uint64_t offset,
+                       BlockDriverCompletionFunc *cb, void *opaque)
+{
+    QEDReadL2TableCB *read_l2_table_cb;
+
+    qed_unref_l2_cache_entry(request->l2_table);
+
+    /* Check for cached L2 entry */
+    request->l2_table = qed_find_l2_cache_entry(&s->l2_cache, offset);
+    if (request->l2_table) {
+        cb(opaque, 0);
+        return;
+    }
+
+    request->l2_table = qed_alloc_l2_cache_entry(&s->l2_cache);
+    request->l2_table->table = qed_alloc_table(s);
+
+    read_l2_table_cb = gencb_alloc(sizeof(*read_l2_table_cb), cb, opaque);
+    read_l2_table_cb->s = s;
+    read_l2_table_cb->l2_offset = offset;
+    read_l2_table_cb->request = request;
+
+    BLKDBG_EVENT(s->bs->file, BLKDBG_L2_LOAD);
+    qed_read_table(s, offset, request->l2_table->table,
+                   qed_read_l2_table_cb, read_l2_table_cb);
+}
+
+int qed_read_l2_table_sync(BDRVQEDState *s, QEDRequest *request, uint64_t offset)
+{
+    int ret = -EINPROGRESS;
+
+    qed_read_l2_table(s, request, offset, qed_sync_cb, &ret);
+    while (ret == -EINPROGRESS) {
+        qemu_aio_wait();
+    }
+
+    return ret;
+}
+
+void qed_write_l2_table(BDRVQEDState *s, QEDRequest *request,
+                        unsigned int index, unsigned int n, bool flush,
+                        BlockDriverCompletionFunc *cb, void *opaque)
+{
+    BLKDBG_EVENT(s->bs->file, BLKDBG_L2_UPDATE);
+    qed_write_table(s, request->l2_table->offset,
+                    request->l2_table->table, index, n, flush, cb, opaque);
+}
+
+int qed_write_l2_table_sync(BDRVQEDState *s, QEDRequest *request,
+                            unsigned int index, unsigned int n, bool flush)
+{
+    int ret = -EINPROGRESS;
+
+    qed_write_l2_table(s, request, index, n, flush, qed_sync_cb, &ret);
+    while (ret == -EINPROGRESS) {
+        qemu_aio_wait();
+    }
+
+    return ret;
+}
diff --git a/contrib/qemu/block/qed.c b/contrib/qemu/block/qed.c
new file mode 100644
index 000000000..f767b0528
--- /dev/null
+++ b/contrib/qemu/block/qed.c
@@ -0,0 +1,1596 @@
+/*
+ * QEMU Enhanced Disk Format
+ *
+ * Copyright IBM, Corp. 2010
+ *
+ * Authors:
+ *  Stefan Hajnoczi   <stefanha@linux.vnet.ibm.com>
+ *  Anthony Liguori   <aliguori@us.ibm.com>
+ *
+ * This work is licensed under the terms of the GNU LGPL, version 2 or later.
+ * See the COPYING.LIB file in the top-level directory.
+ *
+ */
+
+#include "qemu/timer.h"
+#include "trace.h"
+#include "qed.h"
+#include "qapi/qmp/qerror.h"
+#include "migration/migration.h"
+
+static void qed_aio_cancel(BlockDriverAIOCB *blockacb)
+{
+    QEDAIOCB *acb = (QEDAIOCB *)blockacb;
+    bool finished = false;
+
+    /* Wait for the request to finish */
+    acb->finished = &finished;
+    while (!finished) {
+        qemu_aio_wait();
+    }
+}
+
+static const AIOCBInfo qed_aiocb_info = {
+    .aiocb_size         = sizeof(QEDAIOCB),
+    .cancel             = qed_aio_cancel,
+};
+
+static int bdrv_qed_probe(const uint8_t *buf, int buf_size,
+                          const char *filename)
+{
+    const QEDHeader *header = (const QEDHeader *)buf;
+
+    if (buf_size < sizeof(*header)) {
+        return 0;
+    }
+    if (le32_to_cpu(header->magic) != QED_MAGIC) {
+        return 0;
+    }
+    return 100;
+}
+
+/**
+ * Check whether an image format is raw
+ *
+ * @fmt:    Backing file format, may be NULL
+ */
+static bool qed_fmt_is_raw(const char *fmt)
+{
+    return fmt && strcmp(fmt, "raw") == 0;
+}
+
+static void qed_header_le_to_cpu(const QEDHeader *le, QEDHeader *cpu)
+{
+    cpu->magic = le32_to_cpu(le->magic);
+    cpu->cluster_size = le32_to_cpu(le->cluster_size);
+    cpu->table_size = le32_to_cpu(le->table_size);
+    cpu->header_size = le32_to_cpu(le->header_size);
+    cpu->features = le64_to_cpu(le->features);
+    cpu->compat_features = le64_to_cpu(le->compat_features);
+    cpu->autoclear_features = le64_to_cpu(le->autoclear_features);
+    cpu->l1_table_offset = le64_to_cpu(le->l1_table_offset);
+    cpu->image_size = le64_to_cpu(le->image_size);
+    cpu->backing_filename_offset = le32_to_cpu(le->backing_filename_offset);
+    cpu->backing_filename_size = le32_to_cpu(le->backing_filename_size);
+}
+
+static void qed_header_cpu_to_le(const QEDHeader *cpu, QEDHeader *le)
+{
+    le->magic = cpu_to_le32(cpu->magic);
+    le->cluster_size = cpu_to_le32(cpu->cluster_size);
+    le->table_size = cpu_to_le32(cpu->table_size);
+    le->header_size = cpu_to_le32(cpu->header_size);
+    le->features = cpu_to_le64(cpu->features);
+    le->compat_features = cpu_to_le64(cpu->compat_features);
+    le->autoclear_features = cpu_to_le64(cpu->autoclear_features);
+    le->l1_table_offset = cpu_to_le64(cpu->l1_table_offset);
+    le->image_size = cpu_to_le64(cpu->image_size);
+    le->backing_filename_offset = cpu_to_le32(cpu->backing_filename_offset);
+    le->backing_filename_size = cpu_to_le32(cpu->backing_filename_size);
+}
+
+int qed_write_header_sync(BDRVQEDState *s)
+{
+    QEDHeader le;
+    int ret;
+
+    qed_header_cpu_to_le(&s->header, &le);
+    ret = bdrv_pwrite(s->bs->file, 0, &le, sizeof(le));
+    if (ret != sizeof(le)) {
+        return ret;
+    }
+    return 0;
+}
+
+typedef struct {
+    GenericCB gencb;
+    BDRVQEDState *s;
+    struct iovec iov;
+    QEMUIOVector qiov;
+    int nsectors;
+    uint8_t *buf;
+} QEDWriteHeaderCB;
+
+static void qed_write_header_cb(void *opaque, int ret)
+{
+    QEDWriteHeaderCB *write_header_cb = opaque;
+
+    qemu_vfree(write_header_cb->buf);
+    gencb_complete(write_header_cb, ret);
+}
+
+static void qed_write_header_read_cb(void *opaque, int ret)
+{
+    QEDWriteHeaderCB *write_header_cb = opaque;
+    BDRVQEDState *s = write_header_cb->s;
+
+    if (ret) {
+        qed_write_header_cb(write_header_cb, ret);
+        return;
+    }
+
+    /* Update header */
+    qed_header_cpu_to_le(&s->header, (QEDHeader *)write_header_cb->buf);
+
+    bdrv_aio_writev(s->bs->file, 0, &write_header_cb->qiov,
+                    write_header_cb->nsectors, qed_write_header_cb,
+                    write_header_cb);
+}
+
+/**
+ * Update header in-place (does not rewrite backing filename or other strings)
+ *
+ * This function only updates known header fields in-place and does not affect
+ * extra data after the QED header.
+ */
+static void qed_write_header(BDRVQEDState *s, BlockDriverCompletionFunc cb,
+                             void *opaque)
+{
+    /* We must write full sectors for O_DIRECT but cannot necessarily generate
+     * the data following the header if an unrecognized compat feature is
+     * active.  Therefore, first read the sectors containing the header, update
+     * them, and write back.
+     */
+
+    int nsectors = (sizeof(QEDHeader) + BDRV_SECTOR_SIZE - 1) /
+                   BDRV_SECTOR_SIZE;
+    size_t len = nsectors * BDRV_SECTOR_SIZE;
+    QEDWriteHeaderCB *write_header_cb = gencb_alloc(sizeof(*write_header_cb),
+                                                    cb, opaque);
+
+    write_header_cb->s = s;
+    write_header_cb->nsectors = nsectors;
+    write_header_cb->buf = qemu_blockalign(s->bs, len);
+    write_header_cb->iov.iov_base = write_header_cb->buf;
+    write_header_cb->iov.iov_len = len;
+    qemu_iovec_init_external(&write_header_cb->qiov, &write_header_cb->iov, 1);
+
+    bdrv_aio_readv(s->bs->file, 0, &write_header_cb->qiov, nsectors,
+                   qed_write_header_read_cb, write_header_cb);
+}
+
+static uint64_t qed_max_image_size(uint32_t cluster_size, uint32_t table_size)
+{
+    uint64_t table_entries;
+    uint64_t l2_size;
+
+    table_entries = (table_size * cluster_size) / sizeof(uint64_t);
+    l2_size = table_entries * cluster_size;
+
+    return l2_size * table_entries;
+}
+
+static bool qed_is_cluster_size_valid(uint32_t cluster_size)
+{
+    if (cluster_size < QED_MIN_CLUSTER_SIZE ||
+        cluster_size > QED_MAX_CLUSTER_SIZE) {
+        return false;
+    }
+    if (cluster_size & (cluster_size - 1)) {
+        return false; /* not power of 2 */
+    }
+    return true;
+}
+
+static bool qed_is_table_size_valid(uint32_t table_size)
+{
+    if (table_size < QED_MIN_TABLE_SIZE ||
+        table_size > QED_MAX_TABLE_SIZE) {
+        return false;
+    }
+    if (table_size & (table_size - 1)) {
+        return false; /* not power of 2 */
+    }
+    return true;
+}
+
+static bool qed_is_image_size_valid(uint64_t image_size, uint32_t cluster_size,
+                                    uint32_t table_size)
+{
+    if (image_size % BDRV_SECTOR_SIZE != 0) {
+        return false; /* not multiple of sector size */
+    }
+    if (image_size > qed_max_image_size(cluster_size, table_size)) {
+        return false; /* image is too large */
+    }
+    return true;
+}
+
+/**
+ * Read a string of known length from the image file
+ *
+ * @file:       Image file
+ * @offset:     File offset to start of string, in bytes
+ * @n:          String length in bytes
+ * @buf:        Destination buffer
+ * @buflen:     Destination buffer length in bytes
+ * @ret:        0 on success, -errno on failure
+ *
+ * The string is NUL-terminated.
+ */
+static int qed_read_string(BlockDriverState *file, uint64_t offset, size_t n,
+                           char *buf, size_t buflen)
+{
+    int ret;
+    if (n >= buflen) {
+        return -EINVAL;
+    }
+    ret = bdrv_pread(file, offset, buf, n);
+    if (ret < 0) {
+        return ret;
+    }
+    buf[n] = '\0';
+    return 0;
+}
+
+/**
+ * Allocate new clusters
+ *
+ * @s:          QED state
+ * @n:          Number of contiguous clusters to allocate
+ * @ret:        Offset of first allocated cluster
+ *
+ * This function only produces the offset where the new clusters should be
+ * written.  It updates BDRVQEDState but does not make any changes to the image
+ * file.
+ */
+static uint64_t qed_alloc_clusters(BDRVQEDState *s, unsigned int n)
+{
+    uint64_t offset = s->file_size;
+    s->file_size += n * s->header.cluster_size;
+    return offset;
+}
+
+QEDTable *qed_alloc_table(BDRVQEDState *s)
+{
+    /* Honor O_DIRECT memory alignment requirements */
+    return qemu_blockalign(s->bs,
+                           s->header.cluster_size * s->header.table_size);
+}
+
+/**
+ * Allocate a new zeroed L2 table
+ */
+static CachedL2Table *qed_new_l2_table(BDRVQEDState *s)
+{
+    CachedL2Table *l2_table = qed_alloc_l2_cache_entry(&s->l2_cache);
+
+    l2_table->table = qed_alloc_table(s);
+    l2_table->offset = qed_alloc_clusters(s, s->header.table_size);
+
+    memset(l2_table->table->offsets, 0,
+           s->header.cluster_size * s->header.table_size);
+    return l2_table;
+}
+
+static void qed_aio_next_io(void *opaque, int ret);
+
+static void qed_plug_allocating_write_reqs(BDRVQEDState *s)
+{
+    assert(!s->allocating_write_reqs_plugged);
+
+    s->allocating_write_reqs_plugged = true;
+}
+
+static void qed_unplug_allocating_write_reqs(BDRVQEDState *s)
+{
+    QEDAIOCB *acb;
+
+    assert(s->allocating_write_reqs_plugged);
+
+    s->allocating_write_reqs_plugged = false;
+
+    acb = QSIMPLEQ_FIRST(&s->allocating_write_reqs);
+    if (acb) {
+        qed_aio_next_io(acb, 0);
+    }
+}
+
+static void qed_finish_clear_need_check(void *opaque, int ret)
+{
+    /* Do nothing */
+}
+
+static void qed_flush_after_clear_need_check(void *opaque, int ret)
+{
+    BDRVQEDState *s = opaque;
+
+    bdrv_aio_flush(s->bs, qed_finish_clear_need_check, s);
+
+    /* No need to wait until flush completes */
+    qed_unplug_allocating_write_reqs(s);
+}
+
+static void qed_clear_need_check(void *opaque, int ret)
+{
+    BDRVQEDState *s = opaque;
+
+    if (ret) {
+        qed_unplug_allocating_write_reqs(s);
+        return;
+    }
+
+    s->header.features &= ~QED_F_NEED_CHECK;
+    qed_write_header(s, qed_flush_after_clear_need_check, s);
+}
+
+static void qed_need_check_timer_cb(void *opaque)
+{
+    BDRVQEDState *s = opaque;
+
+    /* The timer should only fire when allocating writes have drained */
+    assert(!QSIMPLEQ_FIRST(&s->allocating_write_reqs));
+
+    trace_qed_need_check_timer_cb(s);
+
+    qed_plug_allocating_write_reqs(s);
+
+    /* Ensure writes are on disk before clearing flag */
+    bdrv_aio_flush(s->bs, qed_clear_need_check, s);
+}
+
+static void qed_start_need_check_timer(BDRVQEDState *s)
+{
+    trace_qed_start_need_check_timer(s);
+
+    /* Use vm_clock so we don't alter the image file while suspended for
+     * migration.
+     */
+    qemu_mod_timer(s->need_check_timer, qemu_get_clock_ns(vm_clock) +
+                   get_ticks_per_sec() * QED_NEED_CHECK_TIMEOUT);
+}
+
+/* It's okay to call this multiple times or when no timer is started */
+static void qed_cancel_need_check_timer(BDRVQEDState *s)
+{
+    trace_qed_cancel_need_check_timer(s);
+    qemu_del_timer(s->need_check_timer);
+}
+
+static void bdrv_qed_rebind(BlockDriverState *bs)
+{
+    BDRVQEDState *s = bs->opaque;
+    s->bs = bs;
+}
+
+static int bdrv_qed_open(BlockDriverState *bs, QDict *options, int flags)
+{
+    BDRVQEDState *s = bs->opaque;
+    QEDHeader le_header;
+    int64_t file_size;
+    int ret;
+
+    s->bs = bs;
+    QSIMPLEQ_INIT(&s->allocating_write_reqs);
+
+    ret = bdrv_pread(bs->file, 0, &le_header, sizeof(le_header));
+    if (ret < 0) {
+        return ret;
+    }
+    qed_header_le_to_cpu(&le_header, &s->header);
+
+    if (s->header.magic != QED_MAGIC) {
+        return -EMEDIUMTYPE;
+    }
+    if (s->header.features & ~QED_FEATURE_MASK) {
+        /* image uses unsupported feature bits */
+        char buf[64];
+        snprintf(buf, sizeof(buf), "%" PRIx64,
+            s->header.features & ~QED_FEATURE_MASK);
+        qerror_report(QERR_UNKNOWN_BLOCK_FORMAT_FEATURE,
+            bs->device_name, "QED", buf);
+        return -ENOTSUP;
+    }
+    if (!qed_is_cluster_size_valid(s->header.cluster_size)) {
+        return -EINVAL;
+    }
+
+    /* Round down file size to the last cluster */
+    file_size = bdrv_getlength(bs->file);
+    if (file_size < 0) {
+        return file_size;
+    }
+    s->file_size = qed_start_of_cluster(s, file_size);
+
+    if (!qed_is_table_size_valid(s->header.table_size)) {
+        return -EINVAL;
+    }
+    if (!qed_is_image_size_valid(s->header.image_size,
+                                 s->header.cluster_size,
+                                 s->header.table_size)) {
+        return -EINVAL;
+    }
+    if (!qed_check_table_offset(s, s->header.l1_table_offset)) {
+        return -EINVAL;
+    }
+
+    s->table_nelems = (s->header.cluster_size * s->header.table_size) /
+                      sizeof(uint64_t);
+    s->l2_shift = ffs(s->header.cluster_size) - 1;
+    s->l2_mask = s->table_nelems - 1;
+    s->l1_shift = s->l2_shift + ffs(s->table_nelems) - 1;
+
+    if ((s->header.features & QED_F_BACKING_FILE)) {
+        if ((uint64_t)s->header.backing_filename_offset +
+            s->header.backing_filename_size >
+            s->header.cluster_size * s->header.header_size) {
+            return -EINVAL;
+        }
+
+        ret = qed_read_string(bs->file, s->header.backing_filename_offset,
+                              s->header.backing_filename_size, bs->backing_file,
+                              sizeof(bs->backing_file));
+        if (ret < 0) {
+            return ret;
+        }
+
+        if (s->header.features & QED_F_BACKING_FORMAT_NO_PROBE) {
+            pstrcpy(bs->backing_format, sizeof(bs->backing_format), "raw");
+        }
+    }
+
+    /* Reset unknown autoclear feature bits.  This is a backwards
+     * compatibility mechanism that allows images to be opened by older
+     * programs, which "knock out" unknown feature bits.  When an image is
+     * opened by a newer program again it can detect that the autoclear
+     * feature is no longer valid.
+     */
+    if ((s->header.autoclear_features & ~QED_AUTOCLEAR_FEATURE_MASK) != 0 &&
+        !bdrv_is_read_only(bs->file) && !(flags & BDRV_O_INCOMING)) {
+        s->header.autoclear_features &= QED_AUTOCLEAR_FEATURE_MASK;
+
+        ret = qed_write_header_sync(s);
+        if (ret) {
+            return ret;
+        }
+
+        /* From here on only known autoclear feature bits are valid */
+        bdrv_flush(bs->file);
+    }
+
+    s->l1_table = qed_alloc_table(s);
+    qed_init_l2_cache(&s->l2_cache);
+
+    ret = qed_read_l1_table_sync(s);
+    if (ret) {
+        goto out;
+    }
+
+    /* If image was not closed cleanly, check consistency */
+    if (!(flags & BDRV_O_CHECK) && (s->header.features & QED_F_NEED_CHECK)) {
+        /* Read-only images cannot be fixed.  There is no risk of corruption
+         * since write operations are not possible.  Therefore, allow
+         * potentially inconsistent images to be opened read-only.  This can
+         * aid data recovery from an otherwise inconsistent image.
+         */
+        if (!bdrv_is_read_only(bs->file) &&
+            !(flags & BDRV_O_INCOMING)) {
+            BdrvCheckResult result = {0};
+
+            ret = qed_check(s, &result, true);
+            if (ret) {
+                goto out;
+            }
+        }
+    }
+
+    s->need_check_timer = qemu_new_timer_ns(vm_clock,
+                                            qed_need_check_timer_cb, s);
+
+out:
+    if (ret) {
+        qed_free_l2_cache(&s->l2_cache);
+        qemu_vfree(s->l1_table);
+    }
+    return ret;
+}
+
+/* We have nothing to do for QED reopen, stubs just return
+ * success */