On Mon, Apr 08, 2013 at 03:32:07PM -0400, Laine Stump wrote:
On 04/08/2013 12:48 PM, Daniel P. Berrange wrote:
On Mon, Apr 08, 2013 at 12:37:49PM -0400, Laine Stump wrote:
I think we're starting to get closer to the concrete problem that's
bothering me. As I understand it (and again - what
On Mon, Apr 08, 2013 at 10:37:45PM +0300, Michael S. Tsirkin wrote:
On Mon, Apr 08, 2013 at 03:32:07PM -0400, Laine Stump wrote:
On 04/08/2013 12:48 PM, Daniel P. Berrange wrote:
On Mon, Apr 08, 2013 at 12:37:49PM -0400, Laine Stump wrote:
I think we're starting to get closer to the
On Mon, Apr 08, 2013 at 09:15:28PM -0600, Eric Blake wrote:
On 02/11/2013 09:46 AM, Daniel P. Berrange wrote:
From: Daniel P. Berrange berra...@redhat.com
When removing a VM from the virDomainObjListPtr, we must not
be holding the VM lock while acquiring the list lock. Re-order
code to
On Mon, Apr 08, 2013 at 04:26:19PM -0600, Alex Williamson wrote:
On Mon, 2013-04-08 at 12:37 -0400, Laine Stump wrote:
On 04/05/2013 03:26 PM, Alex Williamson wrote:
Sure. I just mean pci bus 0 (hmm, but actually this does point out a
problem with my logic - the same namespace (well,
On 04/09/13 11:08, Daniel P. Berrange wrote:
On Mon, Apr 08, 2013 at 09:15:28PM -0600, Eric Blake wrote:
On 02/11/2013 09:46 AM, Daniel P. Berrange wrote:
From: Daniel P. Berrange berra...@redhat.com
When removing a VM from the virDomainObjListPtr, we must not
be holding the VM lock while
On Tue, Apr 09, 2013 at 10:08:31AM +0100, Daniel P. Berrange wrote:
On Mon, Apr 08, 2013 at 09:15:28PM -0600, Eric Blake wrote:
On 02/11/2013 09:46 AM, Daniel P. Berrange wrote:
From: Daniel P. Berrange berra...@redhat.com
When removing a VM from the virDomainObjListPtr, we must not
On Tue, Apr 09, 2013 at 12:22:38PM +0200, Peter Krempa wrote:
On 04/09/13 11:08, Daniel P. Berrange wrote:
On Mon, Apr 08, 2013 at 09:15:28PM -0600, Eric Blake wrote:
On 02/11/2013 09:46 AM, Daniel P. Berrange wrote:
From: Daniel P. Berrange berra...@redhat.com
When removing a VM from the
On 04/09/13 12:26, Daniel P. Berrange wrote:
On Tue, Apr 09, 2013 at 10:08:31AM +0100, Daniel P. Berrange wrote:
On Mon, Apr 08, 2013 at 09:15:28PM -0600, Eric Blake wrote:
On 02/11/2013 09:46 AM, Daniel P. Berrange wrote:
From: Daniel P. Berrange berra...@redhat.com
When removing a VM from
On Tue, Apr 09, 2013 at 01:19:05PM +0200, Peter Krempa wrote:
On 04/09/13 12:26, Daniel P. Berrange wrote:
On Tue, Apr 09, 2013 at 10:08:31AM +0100, Daniel P. Berrange wrote:
On Mon, Apr 08, 2013 at 09:15:28PM -0600, Eric Blake wrote:
On 02/11/2013 09:46 AM, Daniel P. Berrange wrote:
From:
On 04/09/13 13:22, Daniel P. Berrange wrote:
On Tue, Apr 09, 2013 at 01:19:05PM +0200, Peter Krempa wrote:
On 04/09/13 12:26, Daniel P. Berrange wrote:
On Tue, Apr 09, 2013 at 10:08:31AM +0100, Daniel P. Berrange wrote:
On Mon, Apr 08, 2013 at 09:15:28PM -0600, Eric Blake wrote:
On
From: Daniel P. Berrange berra...@redhat.com
Typically when you get EOF on a stream, poll will return
POLLIN|POLLHUP at the same time. Thus when we deal with
stream reads, if we see EOF during the read, we can then
clear the VIR_STREAM_EVENT_HANGUP VIR_STREAM_EVENT_ERROR
event bits.
please help me, i just installed libvirt-git and qemu-git to my gentoo os.
when i start libvirtd, then type virsh list , it has no response. hangs
forever.
following is my typed cmd:
$LIBVIRT_DEBUG=1 libvirtd
..
2013-04-09 08:39:03.649+: 17021: debug : virFileClose:72 :
This series fixes the race between freeing and accessing of domain objects and
provides a reproducer case to test it.
Peter Krempa (2):
conf: Fix race between looking up a domain object and freeing it
DO NOT APPLY UPSTREAM: Reproducer of domain free crash
src/conf/domain_conf.c | 8
This patch fixes crash of the daemon that happens due to the following race
condition:
Let's have two threads in the libvirtd daemon's qemu driver:
A - thread executing a API call to get information about a domain
B - thread executing undefine on the same domain
Assume following serialization of
The issue is triggered using the following shell oneliner with this patch
applied:
virsh undefine domain sleep .1; virsh dominfo domain
---
src/conf/domain_conf.c | 4
src/qemu/qemu_driver.c | 6 ++
2 files changed, 10 insertions(+)
diff --git a/src/conf/domain_conf.c
On 04/09/2013 03:08 AM, Daniel P. Berrange wrote:
I'm trying to understand what the behavior was before this patch went in.
Well this was just fixing a deadlock introduced in a previous patch.
You need to look further back than just this patch. Originally the
global QEMU driver lock would
On 04/09/2013 03:02 PM, Peter Krempa wrote:
1 file changed, 4 insertions(+)
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index 03e5740..cafef0c 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -2238,10 +2238,14 @@ void
On 04/05/2013 12:12 PM, Peter Krempa wrote:
This series fixes the default network model for s390 that was broken by the
recent addition of post parse close callbacks and cleans up code obsoleted by
that addition.
Peter Krempa (2):
qemu: Clean up network device CLI generator
qemu:
On Tue, Apr 09, 2013 at 07:57:11AM -0600, Eric Blake wrote:
On 04/09/2013 06:25 AM, Daniel P. Berrange wrote:
From: Daniel P. Berrange berra...@redhat.com
Typically when you get EOF on a stream, poll will return
POLLIN|POLLHUP at the same time. Thus when we deal with
stream reads, if
On 04/09/13 15:41, Ján Tomko wrote:
On 04/05/2013 12:12 PM, Peter Krempa wrote:
This series fixes the default network model for s390 that was broken by the
recent addition of post parse close callbacks and cleans up code obsoleted by
that addition.
Peter Krempa (2):
qemu: Clean up network
On 09/04/13 10:32, Han Cheng wrote:
The helper function to look up disk controller model may be used by scsi
hostdev. But it should be changed to use device info.
Signed-off-by: Han Cheng hanc.f...@cn.fujitsu.com
---
src/conf/domain_conf.c |8
src/conf/domain_conf.h |6
https://bugzilla.redhat.com/show_bug.cgi?id=948821
---
src/libvirt.c | 24
tools/virsh.pod | 4 ++--
2 files changed, 14 insertions(+), 14 deletions(-)
diff --git a/src/libvirt.c b/src/libvirt.c
index cef9f1f..c5221f5 100644
--- a/src/libvirt.c
+++ b/src/libvirt.c
@@
On 09/04/13 10:32, Han Cheng wrote:
The definiton of scsi adapter in storagespool.rng (sourceinfoadapter)
can be used by scsi hostdev. Move it to basictypes.rng.
PortNumber is defined in both domaincommon.rng and storagespool.rng,
simplify it by moving it to basictypes.rng.
Signed-off-by: Han
On 04/09/2013 08:29 AM, Ján Tomko wrote:
https://bugzilla.redhat.com/show_bug.cgi?id=948821
---
src/libvirt.c | 24
tools/virsh.pod | 4 ++--
2 files changed, 14 insertions(+), 14 deletions(-)
ACK.
--
Eric Blake eblake redhat com+1-919-301-3266
Libvirt
On 04/09/2013 04:44 PM, Eric Blake wrote:
On 04/09/2013 08:29 AM, Ján Tomko wrote:
https://bugzilla.redhat.com/show_bug.cgi?id=948821
---
src/libvirt.c | 24
tools/virsh.pod | 4 ++--
2 files changed, 14 insertions(+), 14 deletions(-)
ACK.
Thank you, pushed
It's not desired to force users imagine path for a socket they
are not even supposed to connect to. On the other hand, we
already have a release where the qemu agent socket path is
exposed to XML, so we cannot silently drop it from there.
The new path is generated in form:
Commit 9a3ff01d7f16cc280ce3176620c0714f55511a65 (which was ACKed at
the end of January, but for some reason didn't get pushed until during
the 1.0.4 freeze) fixed the logic in virPCIGetVirtualFunctions().
Unfortunately, a typo in the fix (replacing VIR_REALLOC_N with
VIR_ALLOC_N during code
On 04/09/2013 12:15 PM, Laine Stump wrote:
Commit 9a3ff01d7f16cc280ce3176620c0714f55511a65 (which was ACKed at
the end of January, but for some reason didn't get pushed until during
the 1.0.4 freeze) fixed the logic in virPCIGetVirtualFunctions().
Unfortunately, a typo in the fix (replacing
I believe this set of patches should be applied instead of those from
https://www.redhat.com/archives/libvirt-cim/2013-April/msg6.html
I suppose we could consider combining the two functions; however, there
are a couple place where the return version string is compared more than
once, so I
---
suites/libvirt-cim/lib/XenKvmLib/vxml.py | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/suites/libvirt-cim/lib/XenKvmLib/vxml.py
b/suites/libvirt-cim/lib/XenKvmLib/vxml.py
index 51a4166..82ab501 100644
--- a/suites/libvirt-cim/lib/XenKvmLib/vxml.py
+++
The string version comparisons failed when determining if 0.10.2 was
a greater version than 0.4.1 (and other similar checks). Needed to create
a version comparison method that did the right thing.
Also, not change to 12_create_netfs_storagevolume_errs.py to use or logic
rather than and logic on
On 04/09/2013 02:21 PM, Eric Blake wrote:
On 04/09/2013 12:15 PM, Laine Stump wrote:
Commit 9a3ff01d7f16cc280ce3176620c0714f55511a65 (which was ACKed at
the end of January, but for some reason didn't get pushed until during
the 1.0.4 freeze) fixed the logic in virPCIGetVirtualFunctions().
On 04/09/2013 04:58 AM, Daniel P. Berrange wrote:
On Mon, Apr 08, 2013 at 03:32:07PM -0400, Laine Stump wrote:
On 04/08/2013 12:48 PM, Daniel P. Berrange wrote:
On Mon, Apr 08, 2013 at 12:37:49PM -0400, Laine Stump wrote:
I think we're starting to get closer to the concrete problem that's
On 04/09/2013 12:41 PM, Laine Stump wrote:
(I noticed there is a v1.0.4-maint branch so (although nothing else had
been pushed there since the release of 1.0.4) I pushed this patch to
that branch as well.)
I pushed a few more patches into that branch. I don't know if Fedora 19
will end up on
On 04/04/2013 12:08 PM, Laine Stump wrote:
On 04/03/2013 04:02 PM, Gene Czarcinski wrote:
On 04/02/2013 03:31 PM, Laine Stump wrote:
On 03/15/2013 02:10 PM, Gene Czarcinski wrote:
This patch adds support for adding a static route for
a network. The via specifies the gateway's IP
address.
On 04/09/2013 04:08 PM, Eric Blake wrote:
When pushing to a maint branch, remember to use 'cherry-pick -x', so
that it is obvious which mainline patch you are backporting.
Yeah, I always seem to remember that just *after* I push instead of just
before :-(
--
libvir-list mailing list
On 04/05/2013 02:34 AM, Daniel P. Berrange wrote:
On Thu, Apr 04, 2013 at 08:33:16PM -0600, Eric Blake wrote:
On 04/01/2013 12:55 PM, Eric Blake wrote:
While this update doesn't address any reported problems in libvirt,
doing a post-release update to latest gnulib makes it easier to
stay in
I wanted to check if event loop integration was supported even for remote
applications using libvirt. My application would like to make non-blocking
calls and also recv asynchronous notification remotely. Can some one point
out if its supported.
thanks
--
libvir-list mailing list
On 04/09/2013 12:38 AM, Tal Kain wrote:
On Mon, Apr 8, 2013 at 4:32 PM, Daniel P. Berrange berra...@redhat.comwrote:
On Mon, Apr 08, 2013 at 03:33:07PM +0300, Tal Kain wrote:
Signed-off-by: Tal Kain tal.k...@ravellosystems.com
---
src/qemu/qemu_capabilities.c | 13 +++--
1 file
On 04/09/2013 07:13 AM, Eric Blake wrote:
On 04/09/2013 03:08 AM, Daniel P. Berrange wrote:
I'm trying to understand what the behavior was before this patch went in.
Well this was just fixing a deadlock introduced in a previous patch.
You need to look further back than just this patch.
On 04/09/13 20:33, John Ferlan wrote:
I believe this set of patches should be applied instead of those from
https://www.redhat.com/archives/libvirt-cim/2013-April/msg6.html
I suppose we could consider combining the two functions; however, there
are a couple place where the return version
Kerberos uses 'primary' or 'key' files (principals), not 'abstract
ideal' (principles). Reported by Jason Meinzer.
Reflow a paragraph to fit in 80 columns in the process.
* docs/auth.html.in: Fix spelling.
---
Pushing under the trivial rule.
docs/auth.html.in | 18 ++
1 file
On 04/09/2013 04:35 PM, Peter Krempa wrote:
On 04/09/13 20:33, John Ferlan wrote:
I believe this set of patches should be applied instead of those from
https://www.redhat.com/archives/libvirt-cim/2013-April/msg6.html
I suppose we could consider combining the two functions; however, there
libxl uses some xenstore entries for hints in memory management
(especially when starting new domain). This includes dom0 memory limit
and Xen free memory margin, based on current system state. Entries are
created at first usage, so force such usage at daemon startup, which most
likely will be
For now only for PCI devices. Mostly copy-paste from old xen driver.
---
src/libxl/libxl_driver.c | 193 +++
1 file changed, 193 insertions(+)
diff --git a/src/libxl/libxl_driver.c b/src/libxl/libxl_driver.c
index 40a7a6b..011edf8 100644
---
This are some additional features to libxl driver. Some of them require change
in domain config structures/syntax. Details described with each patch.
The last two patches are bugfix for deadlock during daemon startup.
PS Please keep me in To/Cc - I'm not subscribed to the list.
Marek
This can be useful for route or NAT networks, or any other custom
network setup. Especially configuration example in documentation uses
script/ tag with type 'ethernet'.
---
src/libxl/libxl_conf.c | 17 +
1 file changed, 5 insertions(+), 12 deletions(-)
diff --git
This implement handling of domain name=''/ parameter introduced in previous
patch.
Lookup on domain name (to get domain ID) requires libxlDriverPrivate
object, so it must be passed down to libxlMakeDisk and libxlMakeNet from
top level callers.
---
src/libxl/libxl_conf.c | 62
While iterating with virDomainObjListForEach it is safe to remove
current element. But while iterating, 'doms' lock is already taken, so
can't use standard virDomainObjListRemove. So introduce
virDomainObjListRemoveLocked for this purpose. This is required to fix deadlock
in libxl driver.
---
Use virDomainObjListRemoveLocked instead of virDomainObjListRemove, as
driver-domains is already taken by virDomainObjListForEach.
Above deadlock can be triggered when libvirtd is started after some
domain have been started by hand (in which case driver will not find
libvirt-xml domain config).
At least Xen supports backend drivers in another domain (aka driver
domain). This patch introduces XML config option for such setting as
'domain' element with 'name' attribute. Verification its content is left
for the driver.
In the future some option will be needed for USB devices (hostdev
Xen have feature of having device model in separate domain (called stub
domain). It used to be enabled by special 'stubdom-dm' device model path.
Recent xl have separate config option for this feature
(device_model_stubdomain_override), but I'm not sure if it worth introducing
another xen-specific
Script to be called to prepare custom device for domain. Done with Xen
in mind, it maps to libxl_device_disk.script.
XML configuration would be:
disk type='block' device='disk'
source dev='/dev/mapper/custom-device'/
script path='/script/to/setup/custom-device'/
target dev='xvdc'/
Implement handling of previously introduced script/ element for disk config.
This can be used for custom backend configuration like non-standard
device-mapper nodes, or to prepare device in other domain (see the next patch).
---
src/libxl/libxl_conf.c | 5 +
1 file changed, 5 insertions(+)
Vfb entries in domain config are used only by PV drivers. Qemu parameters are
build based on b_info struct. So fill it with the same data as vfb entries
(actually the first one).
This will additionally allow graphic-less domain, when no graphics/ entries
are present in domain XML (previously VNC
On 04/09/2013 07:53 PM, Marek Marczykowski wrote:
This are some additional features to libxl driver. Some of them require change
in domain config structures/syntax. Details described with each patch.
The last two patches are bugfix for deadlock during daemon startup.
PS Please keep me in
This can be useful for route or NAT networks, or any other custom
network setup. Especially configuration example in documentation uses
script/ tag with type 'ethernet'.
---
src/libxl/libxl_conf.c | 17 +
1 file changed, 5 insertions(+), 12 deletions(-)
diff --git
---
src/libxl/libxl_conf.c | 72 ++
src/libxl/libxl_conf.h | 2 ++
2 files changed, 74 insertions(+)
diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c
index ffc7bbb..7668305 100644
--- a/src/libxl/libxl_conf.c
+++
Script to be called to prepare custom device for domain. Done with Xen
in mind, it maps to libxl_device_disk.script.
XML configuration would be:
disk type='block' device='disk'
source dev='/dev/mapper/custom-device'/
script path='/script/to/setup/custom-device'/
target dev='xvdc'/
Implement handling of previously introduced script/ element for disk
config.
This can be used for custom backend configuration like non-standard
device-mapper nodes, or to prepare device in other domain (see the next
patch).
---
src/libxl/libxl_conf.c | 5 +
1 file changed, 5 insertions(+)
At least Xen supports backend drivers in another domain (aka driver
domain). This patch introduces XML config option for such setting as
'domain' element with 'name' attribute. Verification its content is left
for the driver.
In the future some option will be needed for USB devices (hostdev
Vfb entries in domain config are used only by PV drivers. Qemu
parameters are build based on b_info struct. So fill it with the same
data as vfb entries (actually the first one).
This will additionally allow graphic-less domain, when no graphics/
entries are present in domain XML (previously VNC
Xen have feature of having device model in separate domain (called stub
domain). It used to be enabled by special 'stubdom-dm' device model
path. Recent xl have separate config option for this feature
(device_model_stubdomain_override), but I'm not sure if it worth
introducing another
While iterating with virDomainObjListForEach it is safe to remove
current element. But while iterating, 'doms' lock is already taken, so
can't use standard virDomainObjListRemove. So introduce
virDomainObjListRemoveLocked for this purpose.
---
src/conf/domain_conf.c | 17 +
This are some additional features to libxl driver. Some of them require change
in domain config structures/syntax. Details described with each patch.
The last two patches are bugfix for deadlock during daemon startup.
Marek Marczykowski (12):
libxl: allow script for any network interface, not
For now only for PCI devices. Mostly copy-paste from old xen driver.
---
src/libxl/libxl_driver.c | 193 +++
1 file changed, 193 insertions(+)
diff --git a/src/libxl/libxl_driver.c b/src/libxl/libxl_driver.c
index 40a7a6b..011edf8 100644
---
On 10.04.2013 04:34, Eric Blake wrote:
On 04/09/2013 07:53 PM, Marek Marczykowski wrote:
This are some additional features to libxl driver. Some of them require
change
in domain config structures/syntax. Details described with each patch.
The last two patches are bugfix for deadlock during
Use virDomainObjListRemoveLocked instead of virDomainObjListRemove, as
driver-domains is already taken by virDomainObjListForEach.
Above deadlock can be triggered when libvirtd is started after some
domain have been started by hand (in which case driver will not find
libvirt-xml domain config).
Hi Daniel,
Would you help review this ?
I have added the test case in [PTACH v4 2/2].
Thanks. :)
On 2013年04月07日 16:31, Li Zhang wrote:
From: Li Zhang zhlci...@linux.vnet.ibm.com
To avoid the collision for creating USB controllers in machine-init()
and -device xx command line, it needs to set
Hi Daniel,
Would you please help review this patch?
Thanks. :)
On 2013年04月07日 16:31, Li Zhang wrote:
From: Li Zhang zhlci...@linux.vnet.ibm.com
Add one test case to test CLI with -machine xxx,usb=off.
Signed-off-by: Li Zhang zhlci...@linux.vnet.ibm.com
---
70 matches
Mail list logo