summaryrefslogtreecommitdiffstats
path: root/xlators/features/qemu-block
Commit message (Collapse)AuthorAgeFilesLines
* quota/marker: use smaller stacksize in synctask for marker updationvmallika2015-07-091-1/+1
| | | | | | | | | | | | | | | | | Default stacksize that synctask uses is 2M. For marker we set it to 16k Also move market xlator close to io-threads to have smaller stack Change-Id: I8730132a6365cc9e242a3564a1e615d94ef2c651 BUG: 1207735 Signed-off-by: vmallika <vmallika@redhat.com> Reviewed-on: http://review.gluster.org/11499 Tested-by: NetBSD Build System <jenkins@build.gluster.org> Tested-by: Gluster Build System <jenkins@build.gluster.com> Reviewed-by: Krishnan Parthasarathi <kparthas@redhat.com> Reviewed-by: Raghavendra G <rgowdapp@redhat.com>
* features/qemu-block: Add comment about root_inode handlingPranith Kumar K2015-06-121-1/+2
| | | | | | | | | | | | | This is handling the comment on 3.7 backport http://review.gluster.com/11047 Change-Id: I253a8e0bfebde63f25c839efd55ed3cc30e6cd9b BUG: 1226276 Signed-off-by: Pranith Kumar K <pkarampu@redhat.com> Reviewed-on: http://review.gluster.org/11076 Tested-by: NetBSD Build System <jenkins@build.gluster.org> Tested-by: Gluster Build System <jenkins@build.gluster.com> Reviewed-by: Poornima G <pgurusid@redhat.com>
* features/qemu-block: Don't unref root inodePranith Kumar K2015-05-301-2/+1
| | | | | | | | | | | | Root inode doesn't participate in ref/unref. Don't do it in fini as by the time fini is called itable would be destroyed. BUG: 1226276 Change-Id: I704d0a3c0813cb8f6c3f1f7d613c89aca8f4f9ad Signed-off-by: Pranith Kumar K <pkarampu@redhat.com> Reviewed-on: http://review.gluster.org/11002 Tested-by: Gluster Build System <jenkins@build.gluster.com> Reviewed-by: Niels de Vos <ndevos@redhat.com>
* build: do not #include "config.h" in each fileNiels de Vos2015-05-297-35/+0
| | | | | | | | | | | | | | | | | | Instead of including config.h in each file, and have the additional config.h included from the compiler commandline (-include option). When a .c file tests for a certain #define, and config.h was not included, incorrect assumtions were made. With this change, it can not happen again. BUG: 1222319 Change-Id: I4f9097b8740b81ecfe8b218d52ca50361f74cb64 Signed-off-by: Niels de Vos <ndevos@redhat.com> Reviewed-on: http://review.gluster.org/10808 Tested-by: Gluster Build System <jenkins@build.gluster.com> Tested-by: NetBSD Build System Reviewed-by: Kaleb KEITHLEY <kkeithle@redhat.com> Reviewed-by: Pranith Kumar Karampuri <pkarampu@redhat.com>
* libglusterfs/syncop: Add xdata to all syncop callsRaghavendra Talur2015-04-082-17/+19
| | | | | | | | | | | | | | | | | | | | | | | | | | | This patch adds support for xdata in both the request and response path of syncops. Few calls like lookup already had the support; have renamed variables in few places to maintain uniformity. xdata passed downwards is known as xdata_in and xdata passed upwards is known as xdata_out. There is an old patch by Jeff Darcy at http://review.gluster.org/#/c/8769/3 which does the same for some selected calls. It also brings in xdata support at gfapi level. xdata support at gfapi level would be introduced in subsequent patches. Change-Id: I340e94ebaf2a38e160e65bc30732e8fe1c532dcc BUG: 1158621 Signed-off-by: Raghavendra Talur <rtalur@redhat.com> Reviewed-on: http://review.gluster.org/9859 Tested-by: Gluster Build System <jenkins@build.gluster.com> Reviewed-by: Vijay Bellur <vbellur@redhat.com>
* Avoid conflict between contrib/uuid and system uuidEmmanuel Dreyfus2015-04-043-12/+12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | glusterfs relies on Linux uuid implementation, which API is incompatible with most other systems's uuid. As a result, libglusterfs has to embed contrib/uuid, which is the Linux implementation, on non Linux systems. This implementation is incompatible with systtem's built in, but the symbols have the same names. Usually this is not a problem because when we link with -lglusterfs, libc's symbols are trumped. However there is a problem when a program not linked with -lglusterfs will dlopen() glusterfs component. In such a case, libc's uuid implementation is already loaded in the calling program, and it will be used instead of libglusterfs's implementation, causing crashes. A possible workaround is to use pre-load libglusterfs in the calling program (using LD_PRELOAD on NetBSD for instance), but such a mechanism is not portable, nor is it flexible. A much better approach is to rename libglusterfs's uuid_* functions to gf_uuid_* to avoid any possible conflict. This is what this change attempts. BUG: 1206587 Change-Id: I9ccd3e13afed1c7fc18508e92c7beb0f5d49f31a Signed-off-by: Emmanuel Dreyfus <manu@netbsd.org> Reviewed-on: http://review.gluster.org/10017 Tested-by: Gluster Build System <jenkins@build.gluster.com> Reviewed-by: Niels de Vos <ndevos@redhat.com>
* qemu-block: Fixing coverity issue(Unused value)Nandaja Varma2015-03-301-5/+4
| | | | | | | | | | | | | Coverity ID: 1124889 1124889 1124889 1124889 1124889 1124889 1124889 1124889 1124889 1124889 1124889 Change-Id: I82d57022c4a06fb70573b4e9436f934185f8fb84 BUG: 789278 Signed-off-by: Nandaja Varma <nandaja.varma@gmail.com> Reviewed-on: http://review.gluster.org/9642 Tested-by: Gluster Build System <jenkins@build.gluster.com> Reviewed-by: Humble Devassy Chirammal <humble.devassy@gmail.com> Reviewed-by: Kaleb KEITHLEY <kkeithle@redhat.com>
* Do not assume sizeof(size_t)Emmanuel Dreyfus2014-09-162-2/+2
| | | | | | | | | | | | | | | This fixes an assumption that sizeof(size_t) == sizeof(uint64_t), which is not guaranteed. At least on NetBSD/i386, size_t is 32 bit long. Caught by tests/basics/file-snapshot.t BUG: 764655 Change-Id: Ib7620a2ffe8758521886af37bc280101a040d860 Signed-off-by: Emmanuel Dreyfus <manu@netbsd.org> Reviewed-on: http://review.gluster.org/8441 Reviewed-by: Humble Devassy Chirammal <humble.devassy@gmail.com> Tested-by: Gluster Build System <jenkins@build.gluster.com> Reviewed-by: Dan Lambright <dlambrig@redhat.com>
* Build fix: -lz for qemu-block xlatorEmmanuel Dreyfus2014-09-051-1/+1
| | | | | | | | | | | | | | | LIBZ_LIBS has been changed into ZLIB_LIBS everywhere in the sources, except in qemu-block xlator. Fix it so that it can find -lz and link correctly. BUG: 1129939 Change-Id: I34ded175f056d1a0898804fe602e3d2d2cba27f5 Signed-off-by: Emmanuel Dreyfus <manu@netbsd.org> Reviewed-on: http://review.gluster.org/8623 Reviewed-by: Kaleb KEITHLEY <kkeithle@redhat.com> Tested-by: Gluster Build System <jenkins@build.gluster.com> Reviewed-by: Harshavardhana <harsha@harshavardhana.net> Tested-by: Harshavardhana <harsha@harshavardhana.net>
* osx: Enable qemu feature build on DarwinHarshavardhana2014-08-111-1/+2
| | | | | | | | | Change-Id: Ie5d437aa6c52c180fd8d54680c5f882e75c0bf7e BUG: 1089172 Signed-off-by: Harshavardhana <harsha@harshavardhana.net> Reviewed-on: http://review.gluster.org/8448 Tested-by: Gluster Build System <jenkins@build.gluster.com> Reviewed-by: Kaleb KEITHLEY <kkeithle@redhat.com>
* features/qemu-block: Remove unref of freed iobufPoornima2014-01-291-1/+0
| | | | | | | | | | Change-Id: I507a6504b379eef54be77b54d6e2beee63975ebf BUG: 789278 Signed-off-by: Poornima <pgurusid@redhat.com> Reviewed-on: http://review.gluster.org/6824 Tested-by: Gluster Build System <jenkins@build.gluster.com> Reviewed-by: Raghavendra Talur <rtalur@redhat.com> Reviewed-by: Anand Avati <avati@redhat.com>
* syncop: Change return value of syncopPranith Kumar K2014-01-192-16/+14
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Problem: We found a day-1 bug when syncop_xxx() infra is used inside a synctask with compilation optimization (CFLAGS -O2). Detailed explanation of the Root cause: We found the bug in 'gf_defrag_migrate_data' in rebalance operation: Lets look at interesting parts of the function: int gf_defrag_migrate_data (xlator_t *this, gf_defrag_info_t *defrag, loc_t *loc, dict_t *migrate_data) { ..... code section - [ Loop ] while ((ret = syncop_readdirp (this, fd, 131072, offset, NULL, &entries)) != 0) { ..... code section - [ ERRNO-1 ] (errno of readdirp is stored in readdir_operrno by a thread) /* Need to keep track of ENOENT errno, that means, there is no need to send more readdirp() */ readdir_operrno = errno; ..... code section - [ SYNCOP-1 ] (syncop_getxattr is called by a thread) ret = syncop_getxattr (this, &entry_loc, &dict, GF_XATTR_LINKINFO_KEY); code section - [ ERRNO-2] (checking for failures of syncop_getxattr(). This may not always be executed in same thread which executed [SYNCOP-1]) if (ret < 0) { if (errno != ENODATA) { loglevel = GF_LOG_ERROR; defrag->total_failures += 1; ..... } the function above could be executed by thread(t1) till [SYNCOP-1] and code from [ERRNO-2] can be executed by a different thread(t2) because of the way syncop-infra schedules the tasks. when the code is compiled with -O2 optimization this is the assembly code that is generated: [ERRNO-1] 1165 readdir_operrno = errno; <<---- errno gets expanded as *(__errno_location()) 0x00007fd149d48b60 <+496>: callq 0x7fd149d410c0 <address@hidden> 0x00007fd149d48b72 <+514>: mov %rax,0x50(%rsp) <<------ Address returned by __errno_location() is stored in a special location in stack for later use. 0x00007fd149d48b77 <+519>: mov (%rax),%eax 0x00007fd149d48b79 <+521>: mov %eax,0x78(%rsp) .... [ERRNO-2] 1281 if (errno != ENODATA) { 0x00007fd149d492ae <+2366>: mov 0x50(%rsp),%rax <<----- Because it already stored the address returned by __errno_location(), it just dereferences the address to get the errno value. BUT THIS CODE NEED NOT BE EXECUTED BY SAME THREAD!!! 0x00007fd149d492b3 <+2371>: mov $0x9,%ebp 0x00007fd149d492b8 <+2376>: mov (%rax),%edi 0x00007fd149d492ba <+2378>: cmp $0x3d,%edi The problem is that __errno_location() value of t1 and t2 are different. So [ERRNO-2] ends up reading errno of t1 instead of errno of t2 even though t2 is executing [ERRNO-2] code section. When code is compiled without any optimization for [ERRNO-2]: 1281 if (errno != ENODATA) { 0x00007fd58e7a326f <+2237>: callq 0x7fd58e797300 <address@hidden><<--- As it is calling __errno_location() again it gets the location from t2 so it works as intended. 0x00007fd58e7a3274 <+2242>: mov (%rax),%eax 0x00007fd58e7a3276 <+2244>: cmp $0x3d,%eax 0x00007fd58e7a3279 <+2247>: je 0x7fd58e7a32a1 <gf_defrag_migrate_data+2287> Fix: Make syncop_xxx() return (-errno) value as the return value in case of errors and all the functions which make syncop_xxx() will need to use (-ret) to figure out the reason for failure in case of syncop_xxx() failures. Change-Id: I314d20dabe55d3e62ff66f3b4adb1cac2eaebb57 BUG: 1040356 Signed-off-by: Pranith Kumar K <pkarampu@redhat.com> Reviewed-on: http://review.gluster.org/6475 Tested-by: Gluster Build System <jenkins@build.gluster.com> Reviewed-by: Anand Avati <avati@redhat.com>
* features/qemu-block: simplify coroutine model to use single synctask, ucontextBrian Foster2013-11-107-248/+72
| | | | | | | | | | | | | | | | | | | | | | | | The current coroutine model, mapping synctasks 1-1 with qemu internal Coroutines, has some unresolved raciness issues. This problem usually manifests as lifecycle mismatches between top-level (gluster created) synctasks and the subsequently created internal coroutines from that context. Qemu's internal queueing (and locking) can cause situations where the top-level synctask is destroyed before the internal scheduler has released references to memory, leading to use after free crashes and asserts. Simplify the coroutine model to use a single synctask as a coroutine processor and rely on the existing native ucontext coroutine implementation. The syncenv thread is donated to qemu and ensures a single top-level coroutine is processed at a time. Qemu now has complete control over coroutine scheduling. BUG: 986775 Change-Id: I38223479a608d80353128e390f243933fc946fd6 Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-on: http://review.gluster.org/6110 Tested-by: Gluster Build System <jenkins@build.gluster.com> Reviewed-by: Anand Avati <avati@redhat.com>
* features/qemu-block: add qemu backing image support (clone)Brian Foster2013-11-104-23/+166
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Add basic backing image support to the block-format mechanism. This is a functionality checkpoint that enables the raw mechanism required to support client driven "snapshot" and "clone" requests. This change enhances the block-format setxattr command to support an additional and optional backing image reference. For example: setxattr -n trusted.glusterfs.block-format -v "qcow2:10GB:<bimg>" ./newimage ... where <bimg> refers to the backing image for unallocated blocks in newimage. <bimg> can be provided in one of two formats: - a gfid string in the following format (assuming a valid gfid): <gfid:00000000-0000-0000-0000-000000000000> - or a filename that must be resident in the same directory as the new clone file being formatted. E.g., setxattr -n trusted.glusterfs.block-format -v "qcow2:10GB:baseimg" ./newimage This latter format is more restrictive, simply provided for convenience or until something more refined is available. This change makes no assumptions about the backing image file and affords no additional protection. It is up to the user/client to recognize the relationship between the files and manage them appropriately (i.e., no writes to the backing image, etc.). BUG: 986775 Change-Id: I7aff7bdc59b85a6459001a6bfeae4db6bf74f703 Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-on: http://review.gluster.org/5967 Tested-by: Gluster Build System <jenkins@build.gluster.com> Reviewed-by: Anand Avati <avati@redhat.com>
* features/qemu-block: invoke bdrv_init() only onceBrian Foster2013-10-091-1/+5
| | | | | | | | | | | | | | | | bdrv_init() is intended to be invoked only once. If invoked again after initialization (i.e., due to graph changes), the block driver registration code can reinsert entries into bdrv_drivers, effectively corrupting the NULL terminated linked list. This manifests as an infinite loop for callers attempt to traverse the list (to locate a driver). BUG: 986775 Change-Id: I2f95bda3c753dece4648cdad009d0e2a2d16f644 Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-on: http://review.gluster.org/6021 Tested-by: Gluster Build System <jenkins@build.gluster.com> Reviewed-by: Anand Avati <avati@redhat.com>
* qemu-block: support readdirpBrian Foster2013-09-171-0/+58
| | | | | | | | | | | | | | Support the readdirp fop in qemu-block to ensure that image files are handled correctly when readdirp is enabled. E.g., without readdirp support, incorrect stat data for formatted files can be reported back (and cached) via the client. BUG: 986775 Change-Id: Ibc4bd0b42548953ebe30832f3d853bb68095f0ac Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-on: http://review.gluster.org/5946 Tested-by: Gluster Build System <jenkins@build.gluster.com> Reviewed-by: Anand Avati <avati@redhat.com>
* qemu-block: fix building from distribution tarball when glib2-devel is installedNiels de Vos2013-09-121-0/+8
| | | | | | | | | | | | | | | | | | | | | | Building RPMs from a 'make dist' tarball fails when qemu-block is enabled. Enabling is done automatically when the glib2 development files are available (enabled by ./configure). Manual building with: $ ./autogen.sh && ./configure && make dist && rpmbuild -ta *.tar.gz Building in mock works fine, glib2-devel is not installed by default so the qemu-block xlator gets disabled. This change also adds glib2-devel to the BuildRequires in the glusterfs.spec file, causing the qemu-block xlator to be built by default, and included in the glusterfs RPM. Change-Id: Ibb73628772586d9e07bbfde7a8ff2fc973489086 BUG: 986775 Signed-off-by: Niels de Vos <ndevos@redhat.com> Reviewed-on: http://review.gluster.org/5896 Tested-by: Gluster Build System <jenkins@build.gluster.com> Reviewed-by: Anand Avati <avati@redhat.com>
* features/qemu-block: support for QCOW2 and QED formatsAnand Avati2013-09-0313-0/+2756
This patch adds support for internals snapshots using QCOW2 and general framework for external snapshots (next patch) with QCOW2 and QED. For internal snapshots, the file must be "initialized" or "formatted" into QCOW2 format, and specify a file size. Snapshots can be created, deleted, and applied ("goto"). e.g: // Format and Initialize sh# setfattr -n trusted.glusterfs.block-format -v qcow2:10GB /mnt/imgfile sh# ls -l /mnt/imgfile -rw-r--r-- 1 root root 10G Jul 18 21:20 imgfile // Create a snapshot sh# setfattr -n trusted.glusterfs.block-snapshot-create -v name1 imgfile // Apply a snapshot sh# setfattr -n trusted.gluterfs.block-snapshot-goto -v name1 imgfile Change-Id: If993e057a9455967ba3fa9dcabb7f74b8b2cf4c3 BUG: 986775 Signed-off-by: Anand Avati <avati@redhat.com> Reviewed-on: http://review.gluster.org/5367 Tested-by: Gluster Build System <jenkins@build.gluster.com> Reviewed-by: Brian Foster <bfoster@redhat.com>