copy_file_range support in GlusterFS

    * libglusterfs changes to add new fop

    * Fuse changes:
      - Changes in fuse bridge xlator to receive and send responses

    * posix changes to perform the op on the backend filesystem

    * protocol and rpc changes for sending and receiving the fop

    * gfapi changes for performing the fop

    * tools: glfs-copy-file-range tool for testing copy_file_range fop

      - Although, copy_file_range support has been added to the upstream
	    fuse kernel module, no release has been made yet of a kernel
        which contains the support. It is expected to come in the
        upcoming release of linux-4.20

        So, as of now, executing copy_file_range fop on a fused based
        filesystem results in fuse kernel module sending read on the
	    source fd and write on the destination fd.

	    Therefore a small gfapi based tool has been written to be able
        test the copy_file_range fop. This tool is similar (in functionality)
	    to the example program given in copy_file_range man page.

	    So, running regular copy_file_range on a fuse mount point and
	    running gfapi based glfs-copy-file-range tool gives some idea about
	    how fast, the copy_file_range (or reflink) can be.

	    On the local machine this was the result obtained.

	    mount -t glusterfs workstation:new /mnt/glusterfs
	    [root@workstation ~]# cd /mnt/glusterfs/
	    [root@workstation glusterfs]# ls
	    file
	    [root@workstation glusterfs]# cd
	    [root@workstation ~]# time /tmp/a.out /mnt/glusterfs/file /mnt/glusterfs/new
	    real  0m6.495s
	    user  0m0.000s
	    sys   0m1.439s
	    [root@workstation ~]# time glfs-copy-file-range $(hostname) new /tmp/glfs.log /file /rrr
	    OPEN_SRC: opening /file is success
	    OPEN_DST: opening /rrr is success
	    FSTAT_SRC: fstat on /rrr is success
	    copy_file_range successful

        real  0m0.309s
        user  0m0.039s
        sys   0m0.017s

        This tool needs following arguments
         1) hostname
         2) volume name
         3) log file path
         4) source file path (relative to the gluster volume root)
         5) destination file path (relative to the gluster volume root)

        "glfs-copy-file-range <hostname> <volume> <log file path> <source> <destination>"

      - Added a testcase as well to run glfs-copy-file-range tool

    * io-stats changes to capture the fop for profiling

    * NOTE:

      - Added conditional check to see whether the copy_file_range syscall
        is available or not. If not, then return ENOSYS.

      - Added conditional check for kernel minor version in fuse_kernel.h
        and fuse-bridge while referring to copy_file_range. And the kernel
        minor version is kept as it is. i.e. 24. Increment it in future
        when there is a kernel release which contains the support for
        copy_file_range fop in fuse kernel module.

    * The document which contains a writeup on this enhancement can be found at
      https://docs.google.com/document/d/1BSILbXr_knynNwxSyyu503JoTz5QFM_4suNIh2WwrSc/edit

Change-Id: I280069c814dd21ce6ec3be00a884fc24ab692367
updates: #536
Signed-off-by: Raghavendra Bhat <raghavendra@redhat.com>

6 files changed, 199 insertions, 7 deletions

diff --git a/xlators/protocol/server/src/server-common.c b/xlators/protocol/server/src/server-common.c
index 25b36155065..0639ac3feb3 100644
--- a/xlators/protocol/server/src/server-common.c
+++ b/xlators/protocol/server/src/server-common.c
@@ -541,6 +541,16 @@ server4_post_common_3iatt(server_state_t *state, gfx_common_3iatt_rsp *rsp,
 }
 
 void
+server4_post_common_3iatt_noinode(gfx_common_3iatt_rsp *rsp, struct iatt *stbuf,
+                                  struct iatt *prebuf_dst,
+                                  struct iatt *postbuf_dst)
+{
+    gfx_stat_from_iattx(&rsp->stat, stbuf);
+    gfx_stat_from_iattx(&rsp->preparent, prebuf_dst);
+    gfx_stat_from_iattx(&rsp->postparent, postbuf_dst);
+}
+
+void
 server4_post_common_2iatt(gfx_common_2iatt_rsp *rsp, struct iatt *prebuf,
                           struct iatt *postbuf)
 {
diff --git a/xlators/protocol/server/src/server-common.h b/xlators/protocol/server/src/server-common.h
index 2844ee95756..6200415e304 100644
--- a/xlators/protocol/server/src/server-common.h
+++ b/xlators/protocol/server/src/server-common.h
@@ -192,3 +192,8 @@ void
 server4_post_link(server_state_t *state, gfx_common_3iatt_rsp *rsp,
                   inode_t *inode, struct iatt *stbuf, struct iatt *pre,
                   struct iatt *post);
+
+void
+server4_post_common_3iatt_noinode(gfx_common_3iatt_rsp *rsp, struct iatt *stbuf,
+                                  struct iatt *prebuf_dst,
+                                  struct iatt *postbuf_dst);
diff --git a/xlators/protocol/server/src/server-helpers.c b/xlators/protocol/server/src/server-helpers.c
index c55a422679d..8ad2d8492ed 100644
--- a/xlators/protocol/server/src/server-helpers.c
+++ b/xlators/protocol/server/src/server-helpers.c
@@ -4948,6 +4948,8 @@ server_populate_compound_response_v2(xlator_t *this, gfx_compound_rsp *rsp,
             rsp_args->op_errno = gf_errno_to_error(this_args_cbk->op_errno);
             break;
         }
+        case GF_FOP_COPY_FILE_RANGE:
+            /* Not handling this fop. */
         default:
             return ENOTSUP;
     }
@@ -5380,6 +5382,12 @@ server_get_compound_resolve_v2(server_state_t *state, gfx_compound_req *req)
             memcpy(state->resolve.gfid, this_req.gfid, 16);
             break;
         }
+        case GF_FOP_COPY_FILE_RANGE:
+            /*
+             * Compound operations is not being used anymore and
+             * planned for subsequent removal. Hence not handling
+             * this fop here.
+             */
         default:
             return ENOTSUP;
     }
diff --git a/xlators/protocol/server/src/server-resolve.c b/xlators/protocol/server/src/server-resolve.c
index 26260a5ee2c..ec768acba44 100644
--- a/xlators/protocol/server/src/server-resolve.c
+++ b/xlators/protocol/server/src/server-resolve.c
@@ -545,14 +545,39 @@ server_resolve_fd(call_frame_t *frame)
         return 0;
     }
 
-    state->fd = gf_fd_fdptr_get(serv_ctx->fdtable, fd_no);
-
+    /*
+     * With copy_file_range, there will be 2 fds to resolve.
+     * This same function is called to resolve both the source
+     * fd and the destination fd. As of now, this function does
+     * not have any mechanism to distinguish between the 2 fds
+     * being resolved except for checking the value of state->fd.
+     * The assumption is that, if source fd the one which is
+     * being resolved here, then state->fd would be NULL. If it
+     * is not NULL, then it is the destination fd which is being
+     * resolved.
+     * This  method (provided the above assumption is true) is
+     * to achieve the ability to distinguish between 2 fds with
+     * minimum changes being done to this function. If this way
+     * is not correct, then more changes might be needed.
+     */
     if (!state->fd) {
-        gf_msg("", GF_LOG_INFO, EBADF, PS_MSG_FD_NOT_FOUND,
-               "fd not "
-               "found in context");
-        resolve->op_ret = -1;
-        resolve->op_errno = EBADF;
+        state->fd = gf_fd_fdptr_get(serv_ctx->fdtable, fd_no);
+        if (!state->fd) {
+            gf_msg("", GF_LOG_INFO, EBADF, PS_MSG_FD_NOT_FOUND,
+                   "fd not "
+                   "found in context");
+            resolve->op_ret = -1;
+            resolve->op_errno = EBADF;
+        }
+    } else {
+        state->fd_out = gf_fd_fdptr_get(serv_ctx->fdtable, fd_no);
+        if (!state->fd_out) {
+            gf_msg("", GF_LOG_INFO, EBADF, PS_MSG_FD_NOT_FOUND,
+                   "fd not "
+                   "found in context");
+            resolve->op_ret = -1;
+            resolve->op_errno = EBADF;
+        }
     }
 
     server_resolve_all(frame);
diff --git a/xlators/protocol/server/src/server-rpc-fops_v2.c b/xlators/protocol/server/src/server-rpc-fops_v2.c
index c5a8e482621..16570294f6d 100644
--- a/xlators/protocol/server/src/server-rpc-fops_v2.c
+++ b/xlators/protocol/server/src/server-rpc-fops_v2.c
@@ -2259,6 +2259,64 @@ out:
     return 0;
 }
 
+int
+server4_copy_file_range_cbk(call_frame_t *frame, void *cookie, xlator_t *this,
+                            int32_t op_ret, int32_t op_errno,
+                            struct iatt *stbuf, struct iatt *prebuf_dst,
+                            struct iatt *postbuf_dst, dict_t *xdata)
+{
+    gfx_common_3iatt_rsp rsp = {
+        0,
+    };
+    server_state_t *state = NULL;
+    rpcsvc_request_t *req = NULL;
+    char in_gfid[GF_UUID_BUF_SIZE] = {0};
+    char out_gfid[GF_UUID_BUF_SIZE] = {0};
+
+    dict_to_xdr(xdata, &rsp.xdata);
+
+    if (op_ret < 0) {
+        state = CALL_STATE(frame);
+
+        uuid_utoa_r(state->resolve.gfid, in_gfid);
+        uuid_utoa_r(state->resolve2.gfid, out_gfid);
+
+        gf_msg(this->name, fop_log_level(GF_FOP_COPY_FILE_RANGE, op_errno),
+               op_errno, PS_MSG_WRITE_INFO,
+               "%" PRId64 ": COPY_FILE_RANGE %" PRId64 " (%s), %" PRId64
+               " (%s) client: %s, "
+               "error-xlator: %s",
+               frame->root->unique, state->resolve.fd_no, in_gfid,
+               state->resolve2.fd_no, out_gfid, STACK_CLIENT_NAME(frame->root),
+               STACK_ERR_XL_NAME(frame->root));
+        goto out;
+    }
+
+    /*
+     * server4_post_common_3iatt (ex: used by server4_put_cbk and some
+     * other cbks) also performs inode linking along with copying of 3
+     * iatt structures to the response. But, for copy_file_range, linking
+     * of inode is not needed. Therefore a new function is used to
+     * construct the response using 3 iatt structures.
+     * @stbuf: iatt or stat of the source file (or fd)
+     * @prebuf_dst: iatt or stat of destination file (or fd) before the fop
+     * @postbuf_dst: iatt or stat of destination file (or fd) after the fop
+     */
+    server4_post_common_3iatt_noinode(&rsp, stbuf, prebuf_dst, postbuf_dst);
+
+out:
+    rsp.op_ret = op_ret;
+    rsp.op_errno = gf_errno_to_error(op_errno);
+
+    req = frame->local;
+    server_submit_reply(frame, req, &rsp, NULL, 0, NULL,
+                        (xdrproc_t)xdr_gfx_common_3iatt_rsp);
+
+    GF_FREE(rsp.xdata.pairs.pairs_val);
+
+    return 0;
+}
+
 /* Resume function section */
 
 int
@@ -3448,6 +3506,29 @@ err:
 }
 
 int
+server4_copy_file_range_resume(call_frame_t *frame, xlator_t *bound_xl)
+{
+    server_state_t *state = NULL;
+
+    state = CALL_STATE(frame);
+
+    if (state->resolve.op_ret != 0)
+        goto err;
+
+    STACK_WIND(frame, server4_copy_file_range_cbk, bound_xl,
+               bound_xl->fops->copy_file_range, state->fd, state->off_in,
+               state->fd_out, state->off_out, state->size, state->flags,
+               state->xdata);
+
+    return 0;
+err:
+    server4_copy_file_range_cbk(frame, NULL, frame->this, state->resolve.op_ret,
+                                state->resolve.op_errno, NULL, NULL, NULL,
+                                NULL);
+    return 0;
+}
+
+int
 server4_0_stat(rpcsvc_request_t *req)
 {
     server_state_t *state = NULL;
@@ -6104,6 +6185,53 @@ out:
     return ret;
 }
 
+int
+server4_0_copy_file_range(rpcsvc_request_t *req)
+{
+    server_state_t *state = NULL;
+    call_frame_t *frame = NULL;
+    gfx_copy_file_range_req args = {
+        {
+            0,
+        },
+    };
+    ssize_t len = 0;
+    int ret = -1;
+    int op_errno = 0;
+
+    if (!req)
+        return ret;
+
+    ret = rpc_receive_common(req, &frame, &state, &len, &args,
+                             xdr_gfx_copy_file_range_req,
+                             GF_FOP_COPY_FILE_RANGE);
+    if (ret != 0) {
+        goto out;
+    }
+
+    state->resolve.type = RESOLVE_MUST;
+    state->resolve.fd_no = args.fd_in;
+    state->resolve2.type = RESOLVE_MUST; /*making this resolve must */
+    state->resolve2.fd_no = args.fd_out;
+    state->off_in = args.off_in;
+    state->off_out = args.off_out;
+    state->size = args.size;
+    state->flags = args.flag;
+    memcpy(state->resolve.gfid, args.gfid1, 16);
+    memcpy(state->resolve2.gfid, args.gfid2, 16);
+
+    xdr_to_dict(&args.xdata, &state->xdata);
+
+    ret = 0;
+    resolve_and_resume(frame, server4_copy_file_range_resume);
+out:
+
+    if (op_errno)
+        SERVER_REQ_SET_ERROR(req, ret);
+
+    return ret;
+}
+
 rpcsvc_actor_t glusterfs4_0_fop_actors[] = {
     [GFS3_OP_NULL] = {"NULL", GFS3_OP_NULL, server_null, NULL, 0},
     [GFS3_OP_STAT] = {"STAT", GFS3_OP_STAT, server4_0_stat, NULL, 0},
@@ -6195,6 +6323,8 @@ rpcsvc_actor_t glusterfs4_0_fop_actors[] = {
                          DRC_NA},
     [GFS3_OP_NAMELINK] = {"NAMELINK", GFS3_OP_NAMELINK, server4_0_namelink,
                           NULL, 0, DRC_NA},
+    [GFS3_OP_COPY_FILE_RANGE] = {"COPY-FILE-RANGE", GFS3_OP_COPY_FILE_RANGE,
+                                 server4_0_copy_file_range, NULL, 0, DRC_NA},
 };
 
 struct rpcsvc_program glusterfs4_0_fop_prog = {
diff --git a/xlators/protocol/server/src/server.h b/xlators/protocol/server/src/server.h
index 2a77aba1f3c..bdf98c96f1c 100644
--- a/xlators/protocol/server/src/server.h
+++ b/xlators/protocol/server/src/server.h
@@ -180,7 +180,12 @@ struct _server_state {
     struct iatt stbuf;
     int valid;
 
+    /*
+     * this fd is used in all the fd based operations PLUS
+     * as a source fd in copy_file_range
+     */
     fd_t *fd;
+    fd_t *fd_out; /* destination fd in copy_file_range */
     dict_t *params;
     int32_t flags;
     int wbflags;
@@ -191,6 +196,15 @@ struct _server_state {
 
     size_t size;
     off_t offset;
+    /*
+     * According to the man page of copy_file_range,
+     * the offsets for source and destination file
+     * are of type loff_t. But the type loff_t is
+     * linux specific and is actual a typedef of
+     * off64_t.
+     */
+    off64_t off_in;  /* source offset in copy_file_range */
+    off64_t off_out; /* destination offset in copy_file_range */
     mode_t mode;
     dev_t dev;
     size_t nr_count;