On 27/11/13 00:48, Peter Krempa wrote:
Tweak the existing file so that it can be tested for command line
corectness.
It's actually lots of change, should provide detailed commit log.
---
src/conf/domain_conf.h | 1 +
src/libvirt_private.syms
qemu removes the builtin pvpanic device for all qemu versions since 1.7,
in order to support on_crash, '-device pvpanic' has to be added to
qemu command line.
Signed-off-by: Hu Tao hu...@cn.fujitsu.com
---
src/qemu/qemu_capabilities.c | 8
src/qemu/qemu_capabilities.h | 2 ++
On Wed, Nov 27, 2013 at 04:01:45AM +0530, Nehal J Wani wrote:
On running the command make -C tests valgrind, there used to be a bunch of
memory leaks shown by valgrind. Specifically, one can check it by running:
libtool --mode=execute valgrind --quiet --leak-check=full
On Tue, Nov 26, 2013 at 07:08:07PM -0600, Jesse Cook wrote:
It does not currently appear to be possible to attach/detach a filesystem
with kvm through the C API or virsh. Looking at
src/qemu/qemu_driver.c:qemuDomain{Attach,Detach,Update}Config,
VIR_DOMAIN_DEVICE_FS is not a case within the
On Wed, Nov 27, 2013 at 03:04:26PM +0800, Osier Yang wrote:
The second parameter for virNodeDeviceLookupSCSIHostByWWN is wwnn
instead of wwpn, and the API supports flags.
---
AUTHORS| 1 +
lib/Sys/Virt/NodeDevice.pm | 5 +++--
2 files changed, 4 insertions(+), 2
The 'alias' attribute should *not* be parsed from the XML provided by
the user. It should only be parsed in the live state XML. In the latter
case no codepath should take us to qemuAssignDeviceAliases. So this is
certainly not the right fix. It sounds like the test case is flawed to
me.
Does
On 11/27/13 09:29, Osier Yang wrote:
On 27/11/13 00:48, Peter Krempa wrote:
Tweak the existing file so that it can be tested for command line
corectness.
It's actually lots of change, should provide detailed commit log.
Hmmm, this patch actually should only touch the test files, but I
On 11/27/13 08:47, Osier Yang wrote:
On 27/11/13 00:48, Peter Krempa wrote:
To support testing of volume disk backing, we need to implement a few
disk driver backend functions.
The fake storage driver uses files in storagepoolxml2xmlout/POOLNAME.xml
as XML files for pool definitions and
On Wed, Nov 27, 2013 at 03:08:18PM +0530, Nehal J Wani wrote:
The 'alias' attribute should *not* be parsed from the XML provided by
the user. It should only be parsed in the live state XML. In the latter
case no codepath should take us to qemuAssignDeviceAliases. So this is
certainly not
On 11/27/13 09:41, Hu Tao wrote:
qemu removes the builtin pvpanic device for all qemu versions since 1.7,
in order to support on_crash, '-device pvpanic' has to be added to
qemu command line.
Signed-off-by: Hu Tao hu...@cn.fujitsu.com
---
src/qemu/qemu_capabilities.c | 8
From: Bing Bu Cao m...@linux.vnet.ibm.com
virsh command 'domxml-to-native' did not copy the
MAC address which parsed from a domain xml, in the commandline
the MAC addr of network device always was 00:00:00:00:00:00.
This patch fix it.
Signed-off-by: Bing Bu Cao m...@linux.vnet.ibm.com
---
On Tue, Nov 26, 2013 at 01:00:38PM -0700, Eric Blake wrote:
On 11/26/2013 10:00 AM, Daniel P. Berrange wrote:
From: Daniel P. Berrange berra...@redhat.com
The python code generator tries to figure out what class a
method should be in by looking at the list of arguments for
any which
On Tue, Nov 26, 2013 at 01:17:16PM -0700, Eric Blake wrote:
On 11/26/2013 11:32 AM, Daniel P. Berrange wrote:
From: Daniel P. Berrange berra...@redhat.com
The virNodeGetSecurityModel, virDomainGetSecurityLabel and
virDomainGetSecurityLabelList methods were disabled in the
python
On 26.11.2013 17:49, Peter Krempa wrote:
Currently the snapshot code did not check if it actually supports
snapshots on various disk backends for domains. To avoid future problems
add checkers that whitelist the supported configurations.
---
src/qemu/qemu_driver.c | 264
On 26.11.2013 17:49, Peter Krempa wrote:
Add virDomainDiskAuthClear to help cleaning out the struct in other
places too.
---
src/conf/domain_conf.c | 17 ++---
src/conf/domain_conf.h | 1 +
src/libvirt_private.syms | 1 +
3 files changed, 16 insertions(+), 3 deletions(-)
On 26.11.2013 17:49, Peter Krempa wrote:
---
src/qemu/qemu_driver.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c
index d2dbaf5..96bc87e 100644
--- a/src/qemu/qemu_driver.c
+++ b/src/qemu/qemu_driver.c
@@ -11858,8
On 26.11.2013 17:48, Peter Krempa wrote:
Disk source elements for snapshots were using separate code from our
config parser. As snapshots can be stored on more than just regular
files, we will need the universal parser to allow us to expose a variety
of snapshot disk targets. This patch reuses
From: Daniel P. Berrange berra...@redhat.com
If the libvirt-override-vir.py file has methods which call
C APIs that don't exist in the version of libvirt built against
we need to skip copying their code.
eg for 0.9.13 libvirt we should not copy the 'listAllDomains'
method.
The way this
On 26.11.2013 17:49, Peter Krempa wrote:
To avoid future pain, add placeholder functions to get the actual
snapshot disk type.
---
src/qemu/qemu_conf.c | 23 +++
src/qemu/qemu_conf.h | 6 ++
2 files changed, 29 insertions(+)
diff --git a/src/qemu/qemu_conf.c
On 26.11.2013 17:49, Peter Krempa wrote:
To simplify operations on virDomainDiskHostDef arrays we will need deep
copy and freeing functions. Add and properly export them.
---
src/conf/domain_conf.c | 55
+---
src/conf/domain_conf.h | 3 +++
On 26.11.2013 17:48, Peter Krempa wrote:
The code is common for all the various disk types. Split it out to a
common function.
---
src/conf/domain_conf.c | 62
--
1 file changed, 30 insertions(+), 32 deletions(-)
ACK
Michal
--
On 26.11.2013 17:48, Peter Krempa wrote:
The source element formatting function was expecting a
virDomainDiskDefPtr to store the data. As snapshots are not using this
data structure to hold the data, we need to add an internal function
which splits out individual fields separately.
---
From: Daniel P. Berrange berra...@redhat.com
Older libvirt has files named 'events' and 'virterror'
rather than 'virevent' and 'virerror'. This is visible
in the API XML files. We must look for both names to
ensure we don't loose generation of methods with older
versions of libvirt.
On 26.11.2013 17:49, Peter Krempa wrote:
Clear the old data to avoid leaking it when attempting to re-translate a
pool on the same domain object.
---
src/qemu/qemu_conf.c | 4
1 file changed, 4 insertions(+)
diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c
index
On 26.11.2013 17:48, Peter Krempa wrote:
This code will be reused in the snapshot disk definition parser.
---
src/conf/domain_conf.c | 4 ++--
src/conf/domain_conf.h | 20
2 files changed, 22 insertions(+), 2 deletions(-)
diff --git a/src/conf/domain_conf.c
On 26.11.2013 17:48, Peter Krempa wrote:
Consider the following valid snapshot XML as the driver element is
allowed to be empty in the domainsnapshot.rng schema:
$ cat snap.xml
domainsnapshot
disks
disk name='vda' snapshot='external'
source file='/tmp/foo'/
driver/
On 26.11.2013 17:48, Peter Krempa wrote:
---
src/qemu/qemu_command.c | 117
+---
1 file changed, 61 insertions(+), 56 deletions(-)
diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c
index 6ba8df9..15a6e9b 100644
---
On 26.11.2013 17:48, Peter Krempa wrote:
---
src/qemu/qemu_command.c | 155
+++-
1 file changed, 61 insertions(+), 94 deletions(-)
diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c
index 15a6e9b..799209d 100644
---
On 26.11.2013 17:49, Peter Krempa wrote:
When doing an internal snapshot on a VM with sheepdog or RBD disks we
would not set a flag to mark the domain is using internal snapshots and
might end up creating a mixed snapshot. Move the setting of the variable
to avoid this problem.
---
From: Daniel P. Berrange berra...@redhat.com
When reading/writing a global variable from inside a method
it must be declared as a global, otherwise a local variable
by the same name will be used.
Special case the virConnectListDomainsID method which is
bizarrely renamed for no obvious reason.
On 26.11.2013 17:48, Peter Krempa wrote:
Avoid if statements when used with virBufferEscapeString which
automaticaly omits the whole string. Also add some line breaks to
visualy separate the code.
---
src/conf/domain_conf.c | 48
1 file
On 26.11.2013 17:48, Peter Krempa wrote:
---
src/qemu/qemu_command.c | 128
1 file changed, 86 insertions(+), 42 deletions(-)
ACK
Michal
--
libvir-list mailing list
libvir-list@redhat.com
On 26.11.2013 17:48, Peter Krempa wrote:
---
src/qemu/qemu_command.c | 37 -
1 file changed, 24 insertions(+), 13 deletions(-)
diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c
index 09ebd00..6ba8df9 100644
--- a/src/qemu/qemu_command.c
On Tue, Nov 26, 2013 at 01:16:24PM -0600, Doug Goldstein wrote:
On Tue, Nov 26, 2013 at 12:32 PM, Daniel P. Berrange
berra...@redhat.com wrote:
From: Daniel P. Berrange berra...@redhat.com
Validate that every public API method is mapped into the python
and that every python method has a
From: Daniel P. Berrange berra...@redhat.com
Fix the RPM summary line, add placeholder %changelog tag,
make %setup quiet, add Url: tag and filter out bogus
provides and add example programs as docs.
Signed-off-by: Daniel P. Berrange berra...@redhat.com
---
libvirt-python.spec.in | 13
On 11/27/2013 12:52 PM, m...@linux.vnet.ibm.com wrote:
From: Bing Bu Cao m...@linux.vnet.ibm.com
virsh command 'domxml-to-native' did not copy the
MAC address which parsed from a domain xml, in the commandline
the MAC addr of network device always was 00:00:00:00:00:00.
This patch fix it.
Awhile back a bug was filed against libvirt about the inability to
completely exclude a disk from the boot order:
https://bugzilla.redhat.com/show_bug.cgi?id=888635
In short, you can't have a domain that used PXE to boot, but also has an
un-bootable disk device *even if that disk isn't listed
On 11/27/2013 03:39 AM, Peter Krempa wrote:
On 11/27/13 09:41, Hu Tao wrote:
qemu removes the builtin pvpanic device for all qemu versions since 1.7,
in order to support on_crash, '-device pvpanic' has to be added to
qemu command line.
Signed-off-by: Hu Tao hu...@cn.fujitsu.com
---
I
On 11/27/2013 04:14 AM, Daniel P. Berrange wrote:
From: Daniel P. Berrange berra...@redhat.com
Older libvirt has files named 'events' and 'virterror'
rather than 'virevent' and 'virerror'. This is visible
in the API XML files. We must look for both names to
ensure we don't loose generation
On 11/27/2013 04:14 AM, Daniel P. Berrange wrote:
From: Daniel P. Berrange berra...@redhat.com
When reading/writing a global variable from inside a method
it must be declared as a global, otherwise a local variable
by the same name will be used.
Special case the virConnectListDomainsID
On Wed, Nov 27, 2013 at 14:37:02 +0200, Laine Stump wrote:
Awhile back a bug was filed against libvirt about the inability to
completely exclude a disk from the boot order:
https://bugzilla.redhat.com/show_bug.cgi?id=888635
In short, you can't have a domain that used PXE to boot, but
On 11/27/2013 04:58 AM, Daniel P. Berrange wrote:
From: Daniel P. Berrange berra...@redhat.com
Fix the RPM summary line, add placeholder %changelog tag,
make %setup quiet, add Url: tag and filter out bogus
provides and add example programs as docs.
Signed-off-by: Daniel P. Berrange
On Wed, Nov 27, 2013 at 02:37:02PM +0200, Laine Stump wrote:
Awhile back a bug was filed against libvirt about the inability to
completely exclude a disk from the boot order:
https://bugzilla.redhat.com/show_bug.cgi?id=888635
In short, you can't have a domain that used PXE to boot, but
On 11/26/2013 11:43 PM, Osier Yang wrote:
On 27/11/13 12:00, Eric Blake wrote:
When attempting to backport gluster pools to an older version
where there is no VIR_STRDUP, I got a crash from calling
strdup(,NULL). Rather than relying on the current else branch
safely doing nothing when there
On 11/26/13 19:31, Michal Privoznik wrote:
On 26.11.2013 17:48, Peter Krempa wrote:
Export string conversion from and to the virStorageVolType enum.
---
src/libvirt_private.syms | 2 ++
1 file changed, 2 insertions(+)
diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms
index
A subset of the big gluster snapshot series. I accidentaly squashed two patches
together so I'm reposting the subset that was broken for re-review.
Peter Krempa (4):
test: Implement fake storage pool driver in qemuxml2argv test
qemuxml2argv: Add test to verify correct usage of disk
Before this patch, the translation function still needs a second ugly
helper function to actually format the command line for qemu. But if we
do the right stuff in the translation functio, we don't have to bother
with the second function any more.
This patch removes the messy
To support testing of volume disk backing, we need to implement a few
disk driver backend functions.
The fake storage driver uses files in storagepoolxml2xmlout/POOLNAME.xml
as XML files for pool definitions and volume names are in format
VOL_TYPE+VOL_PATH. By default type block is assumed (for
Tweak the existing file to test command line generator too.
---
tests/qemuxml2argvdata/qemuxml2argv-disk-source-pool.args | 8
tests/qemuxml2argvdata/qemuxml2argv-disk-source-pool.xml | 2 +-
tests/qemuxml2argvtest.c | 2 ++
3 files changed, 11
On Nov 27, 2013, at 6:14 AM, Daniel P. Berrange berra...@redhat.com wrote:
From: Daniel P. Berrange berra...@redhat.com
Older libvirt has files named 'events' and 'virterror'
rather than 'virevent' and 'virerror'. This is visible
in the API XML files. We must look for both names to
Tweak the existing file so that it can be tested for command line
corectness.
---
tests/qemuxml2argvdata/qemuxml2argv-disk-source-pool-mode.args | 10 ++
tests/qemuxml2argvdata/qemuxml2argv-disk-source-pool-mode.xml | 4 ++--
tests/qemuxml2argvtest.c
On Nov 27, 2013, at 6:19 AM, Daniel P. Berrange berra...@redhat.com wrote:
On Tue, Nov 26, 2013 at 01:16:24PM -0600, Doug Goldstein wrote:
On Tue, Nov 26, 2013 at 12:32 PM, Daniel P. Berrange
berra...@redhat.com wrote:
From: Daniel P. Berrange berra...@redhat.com
Validate that every
On 11/27/2013 08:34 AM, Nehal J Wani wrote:
The 'alias' attribute should *not* be parsed from the XML provided by
the user. It should only be parsed in the live state XML. In the latter
case no codepath should take us to qemuAssignDeviceAliases. So this is
certainly not the right fix. It
On Wed, Nov 27, 2013 at 09:04:17PM +0530, Nehal J Wani wrote:
The 'alias' attribute should *not* be parsed from the XML provided by
the user. It should only be parsed in the live state XML. In the latter
case no codepath should take us to qemuAssignDeviceAliases. So this is
certainly not
On 11/26/13 19:30, Michal Privoznik wrote:
On 26.11.2013 17:48, Peter Krempa wrote:
Automatically assign secret type from the disk source definition and
pull in adding of the comma. Then update callers to keep generated
output the same.
---
src/qemu/qemu_command.c | 34
From: Daniel P. Berrange berra...@redhat.com
Move the code for lxcContainerGetSubtree into the virfile
module creating 2 new functions
int virFileGetMountSubtree(const char *mtabpath,
const char *prefix,
char ***mountsret,
From: Daniel P. Berrange berra...@redhat.com
Define 'abs_srcdir' in AM_CFLAGS, just as we do for 'abs_builddir'.
This lets test files write code like
abs_srcdir /some/file
and thus avoid the tedium of virAsprintf() to build paths in some
places.
Signed-off-by: Daniel P. Berrange
From: Daniel P. Berrange berra...@redhat.com
Dan Walsh identified that if you ask LXC to mount both
/etc/aliases and /etc/aliases.db in the container you
only end up with one of them mounted. This was due to
a bogus subpath check.
It is a one line fix which turned into a much bigger
series so
From: Daniel P. Berrange berra...@redhat.com
The code for extracting sub-mounts would just do a STRPREFIX
check on the mount. This was flawed because if there were
the following mounts
/etc/aliases
/etc/aliases.db
and '/etc/aliases' was asked for, it would return both even
though the latter
This patch resolves:
https://bugzilla.redhat.com/show_bug.cgi?id=1035336
The basic problem is that during a network update, the required
iptables rules sometimes change, and this was being handled by simply
removing and re-adding the rules. However, the removal of the old
rules was done based
While running valgrind on the qemuhotplugtest, one of the many memory leaks
detected are:
==3915== 12 bytes in 1 blocks are definitely lost in loss record 36 of 129
==3915==at 0x4A0887C: malloc (vg_replace_malloc.c:270)
==3915==by 0x34268AF275: xmlStrndup (in /usr/lib64/libxml2.so.2.9.1)
On Thu, Nov 28, 2013 at 12:17:31AM +0530, Nehal J Wani wrote:
While running valgrind on the qemuhotplugtest, one of the many memory leaks
detected are:
==3915== 12 bytes in 1 blocks are definitely lost in loss record 36 of 129
==3915==at 0x4A0887C: malloc (vg_replace_malloc.c:270)
NACK, this is completely wrong. I said the *XML parsing* was likely
to be wrong. This is the QEMU monitor code you've deleted.
Oh! Apologies for misunderstanding.
If I am not wrong (again :'() , is this the code you are talking
about? (inside virDomainChrSourceDefParseXML())
While trying to debug a failure of virpcitest during 'make distcheck',
I noticed that with a VPATH build, 'cd tests; ./virpcitest' fails for
an entirely different reason. To reproduce the distcheck failure, I
had to run 'cd tests; abs_srcdir=/path/to/src ./virpcitest'. But we
document in HACKING
On 11/27/2013 02:39 PM, Eric Blake wrote:
While trying to debug a failure of virpcitest during 'make distcheck',
I noticed that with a VPATH build, 'cd tests; ./virpcitest' fails for
an entirely different reason. To reproduce the distcheck failure, I
had to run 'cd tests;
'make distcheck' has been broken since commit 21685c9; basically,
it emulates the case of a read-only $(srcdir) (such as building
from a tarball exploded onto a CD-ROM), but we were creating our
fake pci device as a symlink into $(srcdir) and failing when that
requires opening the config file for
On 11/27/2013 11:03 AM, Laine Stump wrote:
This patch resolves:
https://bugzilla.redhat.com/show_bug.cgi?id=1035336
The basic problem is that during a network update, the required
iptables rules sometimes change, and this was being handled by simply
removing and re-adding the rules.
On 11/27/2013 09:31 AM, Daniel P. Berrange wrote:
From: Daniel P. Berrange berra...@redhat.com
Define 'abs_srcdir' in AM_CFLAGS, just as we do for 'abs_builddir'.
This lets test files write code like
abs_srcdir /some/file
and thus avoid the tedium of virAsprintf() to build paths in
On 11/27/2013 09:31 AM, Daniel P. Berrange wrote:
From: Daniel P. Berrange berra...@redhat.com
Move the code for lxcContainerGetSubtree into the virfile
module creating 2 new functions
int virFileGetMountSubtree(const char *mtabpath,
const char *prefix,
On 11/25/2013 08:20 AM, Eric Blake wrote:
On 11/25/2013 07:24 AM, Daniel P. Berrange wrote:
object-locking-files.txt
+
+# Some tests tend to write into files. Notably, the virpcitest, which
detach
+# and reset a pci device (achieved byt writing into a pci config file).
+check-local:
On 11/27/2013 09:31 AM, Daniel P. Berrange wrote:
From: Daniel P. Berrange berra...@redhat.com
The code for extracting sub-mounts would just do a STRPREFIX
check on the mount. This was flawed because if there were
the following mounts
/etc/aliases
/etc/aliases.db
and '/etc/aliases'
On Wed, Nov 27, 2013 at 02:37:02PM +0200, Laine Stump wrote:
Awhile back a bug was filed against libvirt about the inability to completely
exclude a disk from the boot order:
https://bugzilla.redhat.com/show_bug.cgi?id=888635
In short, you can't have a domain that used PXE to boot, but
I have tagged rc2 in gits and pushed the tarballs and rpms to
ftp://libvirt.org/libvirt/
Seems to build and work fine for me, but please continue testing.
I'm very likely to make 1.2.0 final early on Monday. Beside the
usual testing it would be good if people who asked to separate
On 2013年11月27日 10:56, Jincheng Miao wrote:
* repos/domain/blkstatsflags.py
* repos/domain/block_iotune.py
* repos/domain/block_peek.py
* repos/domain/block_resize.py
* repos/domain/domain_blkio.py
* cases/basic_blockjob.conf
V1-V2:
Removed diskpath params for
On 2013年11月28日 13:25, Guannan Ren wrote:
On 2013年11月27日 10:56, Jincheng Miao wrote:
* repos/domain/blkstatsflags.py
* repos/domain/block_iotune.py
* repos/domain/block_peek.py
* repos/domain/block_resize.py
* repos/domain/domain_blkio.py
* cases/basic_blockjob.conf
V1-V2:
- Original Message -
I know the reason, the fix to the issue can be found in another patch.
One thing you should notice, every patch should not break the running of
existed testcases.
If this patch is based on other patch. you should send the two patches
in a set.
OK, I will send a
On 27/11/13 17:29, Daniel P. Berrange wrote:
On Wed, Nov 27, 2013 at 03:04:26PM +0800, Osier Yang wrote:
The second parameter for virNodeDeviceLookupSCSIHostByWWN is wwnn
instead of wwpn, and the API supports flags.
---
AUTHORS| 1 +
lib/Sys/Virt/NodeDevice.pm | 5 +++--
77 matches
Mail list logo