Commit Graph

12342 Commits

Author SHA1 Message Date
John Ferlan
1d8b4f0592 valgrind: Adjust the suppression file 2013-02-07 14:08:14 -05:00
John Ferlan
2792ad143e virnetttlcontext: Resolve issues found by Valgrind
Need to initialize 'usage' and 'critical' since the VIR_DEBUG will
attempt to use them.
2013-02-07 14:08:14 -05:00
John Ferlan
9442d03c48 qemumonitortestutils: Resolve resource leaks found by Valgrind
When Valgrind runs the 'qemumonitorjsontest' it would claim that the
thread created is leaked. That's because the virThreadJoin won't get
called due to the 'running' flag being cleared.  In order to avoid that,
call virThreadJoin unconditionally at cleanup time.  Also noted that the
qemuMonitorTestWorker() didn't get the test mutex lock on the failure path.

The incoming and outgoing buffers allocated by qemuMonitorTestIO() and
qemuMonitorTestAddReponse() were never VIR_FREE()'d in qemuMonitorTestFree().
2013-02-07 14:08:14 -05:00
John Ferlan
3f5250ff8e qemumonitorjsontest: Resolve resource leaks found by Valgrind
The 'package' string returned by qemuMonitorGetVersion() needs to
be VIR_FREE()'d.

testQemuMonitorJSONGetMachines(), testQemuMonitorJSONGetCPUDefinitions(),
and testQemuMonitorJSONGetCommands() did not VIR_FREE() the array and
array elements allocated by their respective qemuMonitorGet* routines.
2013-02-07 14:08:14 -05:00
John Ferlan
890b6b351f qemu_command: Resolve resource leaks found by Valgrind
The qemuParseGlusterString() replaced dst->src without a VIR_FREE() of
what was in there before.

The qemuBuildCommandLine() did not properly free the boot_buf depending
on various usages.

The qemuParseCommandLineDisk() had numerous paths that didn't clean up
the virDomainDiskDefPtr def properly. Adjust the logic to go through an
error: label before cleanup in order to free the resource.
2013-02-07 14:08:14 -05:00
John Ferlan
7af7c42d05 domain_conf: Resolve resource leaks found by Valgrind
Fix various resource leaks discovered while parsing through Valgrind output
2013-02-07 14:08:14 -05:00
John Ferlan
cbd5dc09e8 vport_profile_conf: Resolve memory leak found by Valgrind
The 'virtPortInterfaceID' was not VIR_FREE()'d
2013-02-07 14:08:14 -05:00
John Ferlan
ba9c7a3026 netdev_vlan_conf: Resolve memory leak found by Valgrind.
The 'trunk' is filled in with virXPathString() value, but was
never VIR_FREE()'d.
2013-02-07 14:08:14 -05:00
John Ferlan
785d6bc58a qemuxml2argvtest: Resolve resource leaks found by Valgrind
Valgrind deterimined that fakeSecretGetValue() was using the secret
value without checking validity.  Returning NULL causes the caller
to emit a message and results in failure.

Additionally commit 'b090aa7d' changes leaked vncSASLdir and vncTLSx509certdir
2013-02-07 14:08:14 -05:00
John Ferlan
0eb62708a7 virnettlscontexttest: Resolve memory leak found by Valgrind
testTLSDerEncode() will allocate memory for der.data, it wasn't VIR_FREE()'d.
also don't initialized der to use static buffer.
2013-02-07 14:08:13 -05:00
Guido Günther
7f5c285082 Add syntax-check to make sure Python files don't contain trailing semicolon
We allow for a trailing semicolon in full line comments since
docs/index.py has some SQL statements in it.
2013-02-07 19:52:49 +01:00
Guido Günther
3591f79b7e Remove more trailing semicolons in Python files 2013-02-07 19:52:44 +01:00
Peter Krempa
2198dadc2f syntax-check: Don't check non-reentrant functions in docs
Otherwise constructions like "random (used by default)" end up breaking
syntax check by apparently using non-reentrant functions.
2013-02-07 18:00:47 +01:00
Guido Günther
8da9464b1c Cosmetics: Remove semicolons
It's Python, not C
2013-02-07 07:49:57 +01:00
Eric Blake
542fbfb076 build: fix build of HAL node backend
Commit 2025356 missed uses of PCI functions in the older HAL-related
code, probably because hal-devel is no longer available in latest Fedora.

* src/node_device/node_device_hal.c (gather_pci_cap): Reflect
function rename.
2013-02-05 16:32:02 -07:00
Eric Blake
98fc0137f1 bitmap: add way to find next clear bit
We had an easy way to iterate set bits, but not for iterating
cleared bits.

* src/util/virbitmap.h (virBitmapNextClearBit): New prototype.
* src/util/virbitmap.c (virBitmapNextClearBit): Implement it.
* src/libvirt_private.syms (bitmap.h): Export it.
* tests/virbitmaptest.c (test4): Test it.
2013-02-05 16:23:14 -07:00
John Ferlan
75fabbdf3f qemu_hotplug: Need to call virUSBDeviceFree() 2013-02-05 17:11:06 -05:00
John Ferlan
0ea43d177c rpc: Need to virCommandFree on error path 2013-02-05 16:51:07 -05:00
John Ferlan
ce6de782e9 uml: Resolve leak if need to requery in umlIdentifyOneChrPTY()
Coverity noted that in the retry logic loop if res had been set, then
it could be leaked so add a VIR_FREE(res) prior to retry.
2013-02-05 16:51:07 -05:00
John Ferlan
40ef77cd86 openvz: Need to error check openvzDomainSetVcpusFlagsInternal() 2013-02-05 16:51:07 -05:00
John Ferlan
7f94521703 parallels: Need to free memory on error path 2013-02-05 16:51:07 -05:00
John Ferlan
3fa23653ee network: Remove conditional settings to resolve resource leak
The conditional setting of cmdout in networkBuildDhcpDaemonCommandLine()
caused Coverity to complain that 'cmd' could be leaked if !cmdout.  Since
the function is local and only called with cmdout being passed those checks
have been removed.
2013-02-05 16:51:07 -05:00
John Ferlan
ede91bde83 storage: Resolve resource leaks with cmd processing 2013-02-05 16:51:06 -05:00
John Ferlan
1c8c819028 keepalive: Resolve Coverity complaint
The Coverity analysis emitted a BAD_SIZEOF error when doing the math
within the TRACE macro. Doing the math outside the macro keeps Coverity quiet.
2013-02-05 16:51:06 -05:00
John Ferlan
ffd61edd87 tlscontext: Make sure to get proper pointer to 'dname'
Keep Coverity happy by passing a pointer to 'dname' rather than the
array itself.  The PROBE expansion would cause a BAD_SIZEOF.
2013-02-05 16:51:06 -05:00
Jiri Denemark
90873ab968 sanitytest.py: Do not rely on system libvirt
When running sanitytest.py we should not rely on libvirt library
installed on the system. And since we generate a nice wrapper called
"run" that sets both PYTHON_PATH and LD_LIBRARY_PATH, we should just use
it rather than trying to duplicate it in the Makefile.
2013-02-05 21:01:15 +01:00
Daniel P. Berrange
0f5e3f136f Initialize qemuImageBinary path at startup 2013-02-05 19:22:26 +00:00
Daniel P. Berrange
34589575bd Introduce annotations for virQEMUDriverPtr fields
Annotate the fields in virQEMUDriverPtr to indicate the locking
rules for their use

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-02-05 19:22:26 +00:00
Daniel P. Berrange
011cf7ad10 Protect USB/PCI device list access in QEMU with dedicated locks
Currently the activePciHostdevs, inactivePciHostdevsd and
activeUsbHostdevs lists are all implicitly protected by the
QEMU driver lock. Now that the lists all inherit from the
virObjectLockable, we can make the locking explicit, removing
the dependency on the QEMU driver lock for correctness.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-02-05 19:22:26 +00:00
Daniel P. Berrange
0f9ef55814 Convert virPCIDeviceList and virUSBDeviceList into virObjectLockable
To allow modifications to the lists to be synchronized, convert
virPCIDeviceList and virUSBDeviceList into virObjectLockable
classes. The locking, however, will not be self-contained. The
users of these classes will have to call virObjectLock/Unlock
in the critical regions.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-02-05 19:22:26 +00:00
Daniel P. Berrange
77c3015f9c Rename all USB device functions to have a standard name prefix
Rename all the usbDeviceXXX and usbXXXDevice APIs to have a
fixed virUSBDevice name prefix
2013-02-05 19:22:25 +00:00
Daniel P. Berrange
3e86e8f327 Fix leak of usbDevice struct when initializing cgroups
When iterating over USB host devices to setup cgroups, the
usbDevice object was leaked in both LXC and QEMU driers

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-02-05 19:22:25 +00:00
Daniel P. Berrange
202535601c Rename all PCI device functions to have a standard name prefix
Rename all the pciDeviceXXX and pciXXXDevice APIs to have a
fixed virPCIDevice name prefix
2013-02-05 19:22:25 +00:00
Daniel P. Berrange
b46f7f4a0b Remove pointless 'qemuVersion' field from virQEMUDriverPtr
The QEMU driver struct has a 'qemuVersion' field that was previously
used to cache the version lookup from capabilities. With the recent
QEMU capabilities rewrite the caching happens at a lower level so
this field is pointless. Removing it avoids worries about locking
when updating it.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-02-05 19:22:25 +00:00
Daniel P. Berrange
6ffcab65c9 Use atomic ops to increment nextvmid
Use atomic ops to increment nextvmid and encapsulate it in a
method to prevent accidental non-atomic access
2013-02-05 19:22:25 +00:00
Daniel P. Berrange
582c445a96 Make virDomainObjList self-locking via virObjectLockable
Switch virDomainObjList to inherit from virObjectLockable and
make all the APIs acquire/release the mutex when running. This
makes virDomainObjList completely self-locking and no longer
reliant on the hypervisor driver locks
2013-02-05 19:22:25 +00:00
Daniel P. Berrange
eea87129f1 Merge virDomainObjListIsDuplicate into virDomainObjListAdd
The duplicate VM checking should be done atomically with
virDomainObjListAdd, so shoud not be a separate function.
Instead just use flags to indicate what kind of checks are
required.

This pair, used in virDomainCreateXML:

   if (virDomainObjListIsDuplicate(privconn->domains, def, 1) < 0)
     goto cleanup;
   if (!(dom = virDomainObjListAdd(privconn->domains,
                                   privconn->caps,
                                   def, false)))
     goto cleanup;

Changes to

   if (!(dom = virDomainObjListAdd(privconn->domains,
                                   privconn->caps,
                                   def,
                                   VIR_DOMAIN_OBJ_LIST_ADD_CHECK_LIVE,
                                   NULL)))
     goto cleanup;

This pair, used in virDomainRestoreFlags:

   if (virDomainObjListIsDuplicate(privconn->domains, def, 1) < 0)
     goto cleanup;
   if (!(dom = virDomainObjListAdd(privconn->domains,
                                   privconn->caps,
                                   def, true)))
     goto cleanup;

Changes to

   if (!(dom = virDomainObjListAdd(privconn->domains,
                                   privconn->caps,
                                   def,
                                   VIR_DOMAIN_OBJ_LIST_ADD_LIVE |
                                   VIR_DOMAIN_OBJ_LIST_ADD_CHECK_LIVE,
                                   NULL)))
     goto cleanup;

This pair, used in virDomainDefineXML:

   if (virDomainObjListIsDuplicate(privconn->domains, def, 0) < 0)
     goto cleanup;
   if (!(dom = virDomainObjListAdd(privconn->domains,
                                   privconn->caps,
                                   def, false)))
     goto cleanup;

Changes to

   if (!(dom = virDomainObjListAdd(privconn->domains,
                                   privconn->caps,
                                   def,
                                   0, NULL)))
     goto cleanup;
2013-02-05 19:22:25 +00:00
Daniel P. Berrange
9fe43021f3 Fix typo in configure.ac causing $LIBS to gain a copy of $CFLAGS
The virt-dbus.m4 check for DBus was preserving $LIBS before
modifying it. Except it wasn't. It was preserving another
copy of $CFLAGS. The result was that after the check completed,
$LIBS got polluted with $CFLAGS

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-02-05 18:04:16 +00:00
Eric Blake
3a9382ca2a tests: reserve more fds for commandtest
Commit 39c77fe triggered random failures, depending on the platform
and what other fds leak into the testsuite (for me, it passed on
RHEL 6 but failed on Fedora 18).  The reason was that we were
expecting an fd that fell outside of our reserved range.  By reserving
a larger range, the test once again passes on all platforms.

* tests/commandtest.c (mymain): Reserve enough fds.
2013-02-05 10:47:00 -07:00
Eric Blake
753020dc2c qemu: don't log failure during QMP add-fd probe
Otherwise, we get a lot of scary (but harmless) noise in the logs:

2013-02-05 15:35:48.555+0000: 8637: error : qemuMonitorJSONCheckError:353 : internal error unable to execute QEMU command 'add-fd': Parameter 'fdset-id' expects an existing fdset-id

one for every qemu 1.2 binary that we probe.

* src/qemu/qemu_monitor_json.c (qemuMonitorJSONAddFd): During
probe, avoid logging failures.
2013-02-05 10:46:12 -07:00
John Ferlan
1eb7a909c4 Add John Ferlan to the committers list 2013-02-05 10:59:32 -05:00
Daniel P. Berrange
37abd47165 Turn virDomainObjList into an opaque virObject
As a step towards making virDomainObjList thread-safe turn it
into an opaque virObject, preventing any direct access to its
internals.

As part of this a new method virDomainObjListForEach is
introduced to replace all existing usage of virHashForEach
2013-02-05 15:49:25 +00:00
Daniel P. Berrange
4f6ed6c33a Rename all domain list APIs to have virDomainObjList prefix
The APIs names for accessing the domain list object are
very inconsistent. Rename them all to have a standard
virDomainObjList prefix.
2013-02-05 15:49:25 +00:00
Daniel P. Berrange
b090aa7d55 Introduce a virQEMUDriverConfigPtr object
Currently the virQEMUDriverPtr struct contains an wide variety
of data with varying access needs. Move all the static config
data into a dedicated virQEMUDriverConfigPtr object. The only
locking requirement is to hold the driver lock, while obtaining
an instance of virQEMUDriverConfigPtr. Once a reference is held
on the config object, it can be used completely lockless since
it is immutable.

NB, not all APIs correctly hold the driver lock while getting
a reference to the config object in this patch. This is safe
for now since the config is never updated on the fly. Later
patches will address this fully.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-02-05 15:49:25 +00:00
Daniel P. Berrange
25ea8e47e7 Fix missing error constants in libvirt python module
The previous change to the generator, changed too much - only
the functions are in 'virerror.c', the constants remained in
'virerror.h' which could not be renamed for API compat reasons.

Add a test case to sanity check the generated python bindings

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-02-05 15:40:20 +00:00
Michal Privoznik
137229bf4a qemu: Catch stderr of image compression binary
If a compression binary prints something to stderr, currently
it is discarded. However, it can contain useful data from
debugging POV, so we should catch it.
2013-02-05 15:45:21 +01:00
Michal Privoznik
cc6c425f94 qemu: Catch stderr of image decompression binary
If a decompression binary prints something to stderr, currently
it is discarded. However, it can contain useful data from
debugging POV, so we should catch it.
2013-02-05 15:45:21 +01:00
Michal Privoznik
1f25194ad1 virFileWrapperFd: Switch to new virCommandDoAsyncIO
Commit 34e8f63a32 introduced support for catching errors from
libvirt iohelper. However, at those times there wasn't such fancy
API as virCommandDoAsyncIO(), so everything has to be implemented
on our own. But since we do have the API now, we can use it and
drop our implementation then.
2013-02-05 15:45:21 +01:00
Michal Privoznik
f0154959b3 tests: Create test for virCommandDoAsyncIO
This is just a basic test, so we don't break virCommand in the
future. A "Hello world\n" string is written to commanhelper,
which copies input to stdout and stderr where we read it from.
Then the read strings are compared with expected values.
2013-02-05 15:45:21 +01:00
Michal Privoznik
39c77fe586 Introduce event loop to commandtest
This is just preparing environment for the next patch, which is
going to need an event loop.
2013-02-05 15:45:21 +01:00