Commit Graph

8204 Commits

Author SHA1 Message Date
ajia@redhat.com
d02c4a1877 util: avoid memory leak
Leak in pciGetVirtualFunctionIndex present since commit 17d64ca.

* src/util/pci.c: fix memory leak.

Signed-off-by: Alex Jia <ajia@redhat.com>
2011-09-19 13:40:21 -06:00
ajia@redhat.com
809999b5a3 remote: avoid memory leak
Leak present since introduction of remoteDomainBuildEventGraphics
in commit 987e31e.

* src/remote/remote_driver.c: fix memory leak.

Signed-off-by: Alex Jia <ajia@redhat.com>
2011-09-19 13:34:52 -06:00
ajia@redhat.com
db3b32c4e1 qemu: avoid memory leak
Leak introduced in commit 036ad50, affects only error case.

* src/qemu/qemu_command.c: fix memory leak.

Signed-off-by: Alex Jia <ajia@redhat.com>
2011-09-19 13:23:53 -06:00
ajia@redhat.com
2fdd441a4a daemon: avoid memory leak
Introduced in commit efa7fc9f.

* daemon/remote.c: fix memory leak in remoteDispatchDomainBlockStatsFlags

Signed-off-by: Alex Jia <ajia@redhat.com>
2011-09-19 13:03:18 -06:00
Eric Blake
ecd8725c1a spec: silence warnings when installing in F16
https://bugzilla.redhat.com/show_bug.cgi?id=738725 documents that
'yum install libvirt' in Fedora 16 is rather noisy.  This fixes
the problems.

* libvirt.spec.in (%post client): Silence chkconfig warning about
SysV services.
(%post) [with_cgconfig]: Drop for Fedora 15 and newer, where
systemd does this automatically.
2011-09-17 06:27:00 -06:00
Eric Blake
43c2641c18 snapshot: allow disk snapshots of qcow2 disks
For all types of disks other than qcow2, we were requesting that
SELinux labeling visit the new file as if it were qcow2, which
means labeling would try to find the backing files of an empty file.
And for a pre-existing qcow2 disk, we were passing NULL, which meant
that labelling tried to probe the file type (and if probing is
disabled, per the default qemu.conf, this made snapshots fail).
What we really want is to make SELinux labeling visit the new
file as raw; it will later be converted to qcow2 if qemu successfully
made the snapshot.

* src/qemu/qemu_driver.c
(qemuDomainSnapshotCreateSingleDiskActive): Force SELinux labeling
to avoid probe of new file.
2011-09-17 06:16:08 -06:00
Eric Blake
2895905a0b snapshot: affect persistent xml after disk snapshot
For external snapshots to be useful on persistent domains, we must
alter the persistent definition alongside the running definition.
Thanks to the possibility of disk hotplug as well as of edits that
only affect the persistent xml, we can't assume that vm->def and
vm->newDef have the same disk at the same index, so we can only
update the persistent copy if the device destination matches up.

* src/qemu/qemu_driver.c (qemuDomainSnapshotCreateDiskActive)
(qemuDomainSnapshotCreateSingleDiskActive): Also affect newDef, if
present.
2011-09-17 05:57:23 -06:00
Eric Blake
1726a736fb build: work around lack of MacOS fdatasync
Mingw lacks fsync, but gnulib provides that.  Meanwhile, gnulib does
not (yet) provide fdatasync, so this is a quick hack to fake that
function on MacOS X; we can revert this configure change once gnulib
gives us a real module.

We have been implicitly relying on gnulib's largefile module being
pulled in by other modules, but it's better to make that explicit.

* bootstrap.conf (gnulib_modules): Add fsync.  Make largefile use
explicit.
* configure.ac (AC_CHECK_FUNCS_ONCE): Check for fdatasync, and
fake it with fsync when not present.
2011-09-16 17:42:28 -06:00
Daniel P. Berrange
8e44e5593e Prevent crash from dlclose() of libvirt.so
When libvirt calls virInitialize it creates a thread local
for the virErrorPtr storage, and registers a callback to
cleanup memory when a thread exits. When libvirt is dlclose()d
or otherwise made non-resident, the callback function is
removed from memory, but the thread local may still exist
and if a thread later exists, it will invoke the callback
and SEGV. There may also be other thread locals with callbacks
pointing to libvirt code, so it is in general never safe to
unload libvirt.so from memory once initialized.

To allow dlclose() to succeed, but keep libvirt.so resident
in memory, link with '-z nodelete'. This issue was first
found with the libvirt CIM provider, but can potentially
hit many of the dynamic language bindings which all ultimately
involve dlopen() in some way, either on libvirt.so itself,
or on the glue code for the binding which in turns links
to libvirt

* configure.ac, src/Makefile.am: Ensure libvirt.so is linked
  with -z nodelete
* cfg.mk, .gitignore, tests/Makefile.am, tests/shunloadhelper.c,
  tests/shunloadtest.c: A test case to unload libvirt while
  a thread is still running.
2011-09-16 15:51:31 -06:00
Alex Jia
14c22b3b64 qemu: add return value check
* src/qemu/qemu_command.c: missing return value check.

Signed-off-by: Alex Jia <ajia@redhat.com>
2011-09-16 14:20:11 -06:00
Jiri Denemark
d2d6776342 qemu: Introduce shutdown reason for paused state
Qemu sends STOP event as part of the shutdown process. Detect such STOP
event and consider shutdown to be reason of emitting such event. That's
the best we can do until qemu provides us the reason directly in STOP
event. This allows us to report shutdown reason for paused state so that
apps can detect domains that failed to finish the shutdown process
(e.g., because qemu is buggy and doesn't exit on SIGTERM or it is
blocked in flushing disk buffers).
2011-09-16 17:25:55 +02:00
Jiri Denemark
96fc478417 qemu: Prevent disk corruption on domain shutdown
Ever since we introduced fake reboot, we call qemuProcessKill as a
reaction to SHUTDOWN event. Unfortunately, qemu doesn't guarantee it
flushed all internal buffers before sending SHUTDOWN, in which case
killing the process forcibly may result in (virtual) disk corruption.

By sending just SIGTERM without SIGKILL we give qemu time to to flush
all buffers and exit. Once qemu exits, we will see an EOF on monitor
connection and tear down the domain. In case qemu ignores SIGTERM or
just hangs there, the process stays running but that's not any different
from a possible hang anytime during the shutdown process so I think it's
just fine.

Also qemu (since 0.14 until it's fixed) has a bug in SIGTERM processing
which causes it not to exit but instead send new SHUTDOWN event and keep
waiting. I think the best we can do is to ignore duplicate SHUTDOWN
events to avoid a SHUTDOWN-SIGTERM loop and leave the domain in paused
state.
2011-09-16 17:21:10 +02:00
Jiri Denemark
c2e9fab273 qemu: Properly detect crash of a rebooted domain
When a domain is rebooted using libvirt API, we use fake reboot
consisting of shutting down and resetting the domain. Thus we see a
SHUTDOWN event and set gotShutdown flag. But we never reset it back and
if the domain crashes after it was rebooted this way, we consider it was
a normal shutdown and not a crash.
2011-09-16 17:18:20 +02:00
Jiri Denemark
3acb664c56 qemu: Fix shutoff reason when domain crashes
Commit 4454a9efc7 changed shutoff reason
from VIR_DOMAIN_SHUTOFF_CRASHED to VIR_DOMAIN_SHUTOFF_FAILED in case we
see an unexpected EOF on monitor connection. But FAILED reason is
dedicated for domains that fail to start. CRASHED reason is the right
one to use in this situation.
2011-09-16 17:14:39 +02:00
Jiri Denemark
018f2e5c9f python: Fix bindings generated in VPATH build 2011-09-16 17:07:57 +02:00
Eric Blake
e9f55e4688 Revert "virsh: Use old API if remote libvirtd does not support new"
This reverts commit 799912fa05b8c3aa37bd04c57b196755f3f70552; now
that the rpc regression is fixed, virsh no longer needs the special
case here.
2011-09-16 08:20:32 -06:00
Eric Blake
4a075f7e7f rpc: convert unknown procedures to VIR_ERR_NO_SUPPORT
Libvirt special-cases a specific VIR_ERR_RPC from the remote driver
back into VIR_ERR_NO_SUPPORT on the client, so that clients can
handle missing rpc functions the same whether the hypervisor driver
is local or remote.  However, commit c1b22644 introduced a regression:
VIR_FROM_THIS changed from VIR_FROM_REMOTE to VIR_FROM_RPC, so the
special casing no longer works if the server uses the newer error
domain.

* src/rpc/virnetclientprogram.c
(virNetClientProgramDispatchError): Also cater to 0.9.3 and newer.
2011-09-16 08:20:32 -06:00
Stefan Berger
3f2cb3ab59 Fix buzzilla 738778
This patch fixes the bug shown in bugzilla 738778. It's not an nwfilter problem but a connection sharing / closure issue.

https://bugzilla.redhat.com/show_bug.cgi?id=738778

Depending on the speed / #CPUs of the machine you are using you may not see this bug all the time.
2011-09-16 09:44:43 -04:00
Alex Jia
b996110285 conf: avoid memory leak on virDomainDefParseXML
* conf/domain_conf.c: allocate memory to def->redirdevs in
  virDomainDefParseXML such as VIR_ALLOC_N(def->redirdevs, n),
  however, virDomainDefFree(def) hasn't released these memory.

* Detected in valgrind run:

==19820== 209 (16 direct, 193 indirect) bytes in 1 blocks are definitely lost in loss record 25 of 26
==19820==    at 0x4A04A28: calloc (vg_replace_malloc.c:467)
==19820==    by 0x4A13AF: virAllocN (memory.c:129)
==19820==    by 0x4D4A0E: virDomainDefParseXML (domain_conf.c:7258)
==19820==    by 0x4D4C93: virDomainDefParseNode (domain_conf.c:7512)
==19820==    by 0x4D562F: virDomainDefParse (domain_conf.c:7465)
==19820==    by 0x415863: testCompareXMLToXMLFiles (qemuxml2xmltest.c:35)
==19820==    by 0x415982: testCompareXMLToXMLHelper (qemuxml2xmltest.c:80)
==19820==    by 0x416D31: virtTestRun (testutils.c:140)
==19820==    by 0x415604: mymain (qemuxml2xmltest.c:192)
==19820==    by 0x416437: virtTestMain (testutils.c:689)
==19820==    by 0x3CA7A1ECDC: (below main) (in /lib64/libc-2.12.so)
==19820==
==19820== LEAK SUMMARY:
==19820==    definitely lost: 16 bytes in 1 blocks
==19820==    indirectly lost: 193 bytes in 5 blocks
==19820==      possibly lost: 0 bytes in 0 blocks
==19820==    still reachable: 1,054 bytes in 21 blocks

* How to reproduce?
% valgrind -v --leak-check=full ./tests/qemuxml2xmltest

Signed-off-by: Alex Jia <ajia@redhat.com>
2011-09-16 21:15:05 +08:00
Peter Krempa
bc35f12a45 build: storage: Macro 'MKFS' is undefined on some platforms.
Mac OS X 10.6. Snow Leopard and probably other do not provide a mkfs
command to create filesystems. Macro MKFS then remained undefined and
did not provide any substitute, so that build failed on a missing
argument.
2011-09-16 21:07:02 +08:00
Peter Krempa
1ce3b61fa5 build: storage: Conditionaly compiled structure caused build fail on OSX
Struct virStoragePoolProbeResult was compiled in conditionaly, but
virStorageBackendFileSystemProbe used it unconditionaly. This patch
exempts the struct from conditional include.
2011-09-16 21:03:50 +08:00
Peter Krempa
79cf07af7c Avoid using "devname" as an identifier.
/usr/lib/stdlib.h in Mac OS X and probably also in BSD's
exports this symbol :(
2011-09-16 20:49:04 +08:00
Peter Krempa
4d4430e125 doc: Add statment about permissions needed to do a core dump
Documentation did not specify, that some permissions are required on
target path for coredump for the user running the hypervisor.

Diff to v1:
- reword statements
2011-09-16 20:40:21 +08:00
Eric Blake
85d2810823 snapshot: tweak snapshot-create-as diskspec docs
With this patch, it is hopefully a bit more obvious that for
snapshot-create-as, a literal '--diskspec' is mandatory if name
or description was omitted, but optional if all earlier options
were provided.

These all denote two diskspecs and a description:
virsh snapshot-create-as dom name desc vda vdb
virsh snapshot-create-as dom name desc --diskspec vda --diskspec vdb
virsh snapshot-create-as dom name desc --diskspec vda vdb
virsh snapshot-create-as dom name desc vda --diskspec vdb
virsh snapshot-create-as dom --diskspec vda --diskspec vdb name desc

This gives two diskspecs but no description:
virsh snapshot-create-as dom name --diskspec vda --diskspec vdb

And this treats 'vda' as the description, with only one diskspec:
virsh snapshot-create-as dom name vda vdb

The help output now shows:
    snapshot-create-as <domain> [<name>] [<description>] [--print-xml] [--no-metadata] [--halt] [--disk-only] [[--diskspec] <string>]...

I also checked the help output for echo and send-key, which are two
other variants of argv commands.

* tools/virsh.pod (snapshot-create-as): Document when a literal
--diskspec must preceed a diskspec argument.
* tools/virsh.c (vshCmddefHelp): Update help output for argv when
naming the option is useful.
(vshCmddefGetData): Fix logic on when argv was seen.
* tests/virsh-optparse: Add tests to avoid regressions.
2011-09-15 16:18:12 -06:00
Cole Robinson
9ed46a3ed7 python: Fix libvirt.py generation to include virterror info
Recent generator refactoring broke libvirt.py. With this patch, libvirt.py
is generated exactly the same as before offending commit 9eba0d25.
2011-09-15 16:44:50 -04:00
Eric Blake
69d8c75333 qemu_api: doc improvements
The new doc text had a few readability issues.  Also, the
monitor command text copied a bit too much from the attach case.

* src/libvirt-qemu.c (virDomainQemuMonitorCommand)
(virDomainQemuAttach): Fix typos and grammar.
2011-09-15 13:56:38 -06:00
Peter Krempa
b4a01cecd6 doc: virsh: Fix command name in man page
Fix cut&paste error having command named domif-setlink instead of
domif-getlink.
2011-09-15 13:37:13 -06:00
Alex Jia
e99f01ecb6 sanlock: add missing test command in virt-sanlock-cleanup.in
* tools/virt-sanlock-cleanup.in: fix missing test command when judging
  second condition.

https://bugzilla.redhat.com/show_bug.cgi?id=738534

Signed-off-by: Alex Jia <ajia@redhat.com>
2011-09-15 09:44:48 -06:00
Michal Privoznik
a73e92ba08 spec: Require augeas for sanlock
Even though we BuildRequire augeas in some cases, we need to require
it even after if we build with sanlock. virt-sanlock-cleanup use it.
2011-09-15 17:08:59 +02:00
Jim Fehlig
1c90642f85 Fix build after commit 829bce17
Pushing under build-breaker rule.
2011-09-14 17:57:55 -06:00
Jim Fehlig
0257ba8f9f Use max bandwidth from qemuDomainObjPrivate struct when migrating
Adjust qemuMigrationRun() to use migMaxBandwidth in qemuDomainObjPrivate
structure when setting qemu migration speed.  Caller-specified 'resource'
parameter overrides migMaxBandwidth.
2011-09-14 16:42:11 -06:00
Jim Fehlig
ef1065cf5a Set qemu migration speed unlimited when migrating to file
The qemu migration speed default is 32MiB/s as defined in migration.c

/* Migration speed throttling */
static int64_t max_throttle = (32 << 20);

There's no need to throttle migration when targeting a file, so set migration
speed to unlimited prior to migration, and restore to libvirt default value
after migration.

Default units is MB for migrate_set_speed monitor command, so
(INT64_MAX / (1024 * 1024)) is used for unlimited migration speed.

Tested with both json and text monitors.
2011-09-14 16:42:11 -06:00
Jim Fehlig
8fc40c511c Save migration speed in qemuDomainMigrateSetMaxSpeed
Now that migration speed is stored in qemuDomainObjPrivate structure,
save the new value when invoking qemuDomainMigrateSetMaxSpeed().

Allow setting migration speed on inactive domain too.
2011-09-14 16:42:10 -06:00
Jim Fehlig
829bce174c Impl virDomainMigrateGetMaxSpeed in qemu driver 2011-09-14 16:42:10 -06:00
Jim Fehlig
6f84e110d6 Store max migration bandwidth in qemuDomainObjPrivate struct
The maximum bandwidth that can be consumed when migrating a domain
is better classified as an operational vs configuration parameter of
the dommain.  As such, store this parameter in qemuDomainObjPrivate
structure.
2011-09-14 16:42:10 -06:00
Marc-André Lureau
dd428d4798 conf: add missing break in virDomainAuditRedirdev
Also initialize to NULL a few variables that might get
free before being set.
2011-09-14 15:30:32 -06:00
Eric Blake
247726bf38 network: add missing exports
Commit c246b025 added new functions, but forgot to export them,
resulting in a build failure when using modules.

* src/libvirt_private.syms (network.h): Export new functions.
2011-09-14 11:55:17 -06:00
Eric Blake
f2fc1eee4c snapshot: ABI stability must include memory sizing
Commit 973fcd8f introduced the ability for qemu to reject snapshot
reversion on an ABI incompatibility; but the very example that was
first proposed on-list[1] as a demonstration of an ABI incompatibility,
namely that of changing the max memory allocation, was not being
checked for, resulting in a cryptic failure when running with larger
max mem than what the snapshot was created with:
error: operation failed: Error -22 while loading VM state

This commit merely protects the three variables within mem that are
referenced by qemu_command.c, rather than all 7 (the other 4 variables
affect cgroup handling, but as far as I can tell, have no visible effect
to the qemu guest).  This also affects migration and save file handling,
which are other places where we perform ABI compatibility checks.

[1] https://www.redhat.com/archives/libvir-list/2010-December/msg00331.html

* src/conf/domain_conf.c (virDomainDefCheckABIStability): Add
memory sizing checks.
2011-09-14 09:56:30 -06:00
Eric Blake
72f865dd23 maint: ignore generated files
* .gitignore: Exempt recently added generated files.
2011-09-14 09:30:33 -06:00
Peter Krempa
b998f1f77c xml: Clean up rest of virtual XML document names for XML strings
Commit 498d783 cleans up some of virtual file names for parsing strings
in memory. This patch cleans up (hopefuly) the rest forgotten by the
first patch.

This patch also changes all of the previously modified "filenames" to
valid URI's replacing spaces for underscores.

Changes to v1:
- Replace all spaces for underscores, so that the strings form valid
  URI's
- Replace spaces in places changed by commit 498d783
2011-09-14 09:09:04 -06:00
Daniel Veillard
9d0ae85088 Remove two references to files not generated
This was breaking "make dist"
2011-09-14 22:57:46 +08:00
Peter Krempa
49ce282a6d make: Fix 'make syntax-check' failing.
Commit 2a0d75e5 added file python/libvirt-qemu-override.c that contains
code that does not pass "make syntax-check". This patch adds an
exception for this file and the check.

prohibit_always_true_header_tests
python/libvirt-qemu-override.c:17:#undef HAVE_PTHREAD_H
maint.mk: do not test the above HAVE_<header>_H symbol(s);
  with the corresponding gnulib module, they are always true
make: *** [sc_prohibit_always_true_header_tests] Error 1
2011-09-14 08:35:25 -06:00
Osier Yang
8918fc8edf qemu_api: Update libvirt spec file 2011-09-14 11:39:02 +08:00
Osier Yang
477d240778 qemu_api: Update Makefile to generate libvirtmod_qemu lib 2011-09-14 11:38:59 +08:00
Osier Yang
9eba0d2591 qemu_api: Update Py binding generator to generate files for QEMU APIs
It will generate:
  libvirt-qemu.py
  libvirt-qemu.h
  libvirt-qemu.c
  libvirt-qemu-export.c
2011-09-14 11:38:55 +08:00
Osier Yang
2a0d75e5ee qemu_api: Add override XML and C files for QEMU APIs
There is only one function (virDomainQemuMonitorCommand) need to
be hand-craft.
2011-09-14 11:38:51 +08:00
Osier Yang
4e53546911 qemu_api: Add comments for API virDomainQemuMonitorCommand
And fix argument @pid's type of virDomainQemuAttach.
2011-09-14 11:38:47 +08:00
Osier Yang
60290c84ad qemu_api: Update Makefile for subdir docs 2011-09-14 11:38:39 +08:00
Osier Yang
c5edc92071 qemu_api: Modify apibuild.py to generate docs for QEMU APIs
The generated docs are: libvirt-qemu-api.xml, libvirt-qemu-refs.xml
2011-09-14 11:36:10 +08:00
Peter Krempa
799912fa05 virsh: Use old API if remote libvirtd does not support new
Commit ffe28ab74b introduced regression
while communicating with older libvirtd command 'domblkstat' used the new
API and did not check for VIR_ERR_RPC error code signalling the remote
server does not support this API and did not fall back to older API.
Thereafter 'domblkstat' ended with "error: unknown procedure: 243".
2011-09-13 16:55:04 -06:00