On Sun, Mar 31, 2013 at 9:46 PM, Daniel Veillard veill...@redhat.comwrote:
On Mon, Apr 01, 2013 at 05:18:16AM +0100, Justin Clift wrote:
On 01/04/2013, at 5:10 AM, Daniel Veillard wrote:
As planned I tagged the release in git and pushed the new version
to the usual place as tarball and
On Mon, Apr 01, 2013 at 12:08:56AM -0700, Jason Helfman wrote:
On Sun, Mar 31, 2013 at 9:46 PM, Daniel Veillard veill...@redhat.comwrote:
On Mon, Apr 01, 2013 at 05:18:16AM +0100, Justin Clift wrote:
On 01/04/2013, at 5:10 AM, Daniel Veillard wrote:
As planned I tagged the release in
On Mon, Apr 1, 2013 at 6:18 AM, Justin Clift jcl...@redhat.com wrote:
Libvirt 1.0.4 is now in Homebrew too.
$ brew install libvirt
(how's that for quick turnaround? :)
Regards and best wishes,
Justin Clift
Works on 10.8 as well.
[ruben@odin ~]$ brew install libvirt
== Installing
Hi Cindy,
Thanks for your reply.
Since LXC is an independent driver in libvirt, I guess to migrate LXC domains
doesn't depend on external program.
But it failed with different error messages on source and destination nodes.
1) Migrate the virtual machine to the destination host
This patch series tried to add support for scsi-generic for virtio-scsi.
Changes from v1:
- Mv readonly to virDomainHostdevDef as Osier advises.
- Use scsi adaper definition from storagepool.
- Limit scsi hostdev to add one live guest.
- Create some util function for scsi hostdev.
- Add hotplug
The only parameter in -drive affect scsi-generic is readonly. Introduce
readonly/ to hostdev.
The helper function to look up disk controller model may be used by scsi
hostdev. But it should be changed to use info.
Signed-off-by: Han Cheng hanc.f...@cn.fujitsu.com
---
docs/formatdomain.html.in
As libvirt gives guest minimal cgroup, we need to add sg into guest cgroup
whitelist for scsi hostdev. And we should set and restore selinux label
correctly for scsi hostdev.
Signed-off-by: Han Cheng hanc.f...@cn.fujitsu.com
---
src/qemu/qemu_cgroup.c | 67
Adding two caps to support scsi-generic:
QEMU_CAPS_SCSI_GENERIC
QEMU_CAPS_SCSI_GENERIC_BOOTINDEX
Signed-off-by: Han Cheng hanc.f...@cn.fujitsu.com
---
src/qemu/qemu_capabilities.c | 15 +--
src/qemu/qemu_capabilities.h |2 ++
2 files changed, 15 insertions(+), 2
The definiton of scsi adapter in storagespool.rng can be used by scsi
hostdev. Split storagepool.rng into storagepool.rng and
storagepoolcommon.rng and make domaincomman.rng include the latter one.
PortNumber is both defined in domaincomman.rng and storagespool.rng,
simplify it.
Signed-off-by:
For scsi-generic, the command line will be like:
-drive file=/dev/sg0,if=none,id=drive-hostdev-scsi_host0-0-0-0 \
-device scsi-generic,bus=scsi0.0,channel=0,scsi-id=4,lun=8,\
drive=drive-hostdev-scsi_host0-0-0-0,id=hostdev-scsi_host0-0-0-0
The relationship between the libvirt address attrs
This patch add util functions for scsi hostdev.
Signed-off-by: Han Cheng hanc.f...@cn.fujitsu.com
---
po/POTFILES.in |1 +
src/Makefile.am |1 +
src/libvirt_private.syms | 22 +++
src/util/virscsi.c | 399 ++
Adding scsi hostdev, it should like:
hostdev mode='subsystem' type='scsi'
source
adapter name='scsi_host0'/
address bus='0' target='0' unit='0'/
/source
address type='drive' controller='0' bus='0' target='4' unit='8'/
/hostdev
Signed-off-by: Han Cheng
Although virtio-scsi support SCSI PR, the device in host may do not
support this. To avoid losing data, we only allow one scsi hostdev be
passthrough to one guest.
Signed-off-by: Han Cheng hanc.f...@cn.fujitsu.com
---
src/qemu/qemu_conf.h|2 +
src/qemu/qemu_driver.c |3 +
This patch add hotplug for scsi hostdev.
And user should hotplug a virtio-scsi controller if doesn't exist.
Usb hostdev related codes are in qemuDomainAttachHostDevice, push down to
qemuDomainAttachHostUsbDevice.
Signed-off-by: Han Cheng hanc.f...@cn.fujitsu.com
---
src/qemu/qemu_hotplug.c |
This patch add tests for scsi hostdev.
Signed-off-by: Han Cheng hanc.f...@cn.fujitsu.com
---
tests/qemuhelpdata/qemu-1.0-device | 10 ++
tests/qemuhelpdata/qemu-1.1.0-device | 10 ++
tests/qemuhelpdata/qemu-1.2.0-device |5 +++
Ping, anybody can review this?
On 28/03/13 19:01, Osier Yang wrote:
On 28/03/13 19:36, Osier Yang wrote:
vcpupin and emulatorpin use same code to parse the cpulist, this
abstracts the same code as a helper. Along with various code style
fixes, and error improvement (only error Physical CPU %d
This introduce a new attribute num_queues (same with the good name
QEMU uses) for virtio-scsi controller. An example of the XML:
controller type='scsi' index='0' model='virtio-scsi' num_queues='8'/
The corresponding QEMU command line:
-device
On 04/01/2013 10:04 AM, Osier Yang wrote:
This introduce a new attribute num_queues (same with the good name
QEMU uses) for virtio-scsi controller. An example of the XML:
controller type='scsi' index='0' model='virtio-scsi' num_queues='8'/
Our XML is not consistent on whether we prefer '-'
On 02/04/13 00:21, Eric Blake wrote:
On 04/01/2013 10:04 AM, Osier Yang wrote:
This introduce a new attribute num_queues (same with the good name
QEMU uses) for virtio-scsi controller. An example of the XML:
controller type='scsi' index='0' model='virtio-scsi' num_queues='8'/
Our XML is not
On 04/01/2013 10:31 AM, Osier Yang wrote:
No, multiple queue is not enabled by default. For what happens if it
doesn't match the number of vcpus, honestly, I'm not sure about it,
my understanding is it doesn't have to match. But matching will have
the best performance, and with this thought I
On 03/28/2013 05:36 AM, Osier Yang wrote:
vcpupin and emulatorpin use same code to parse the cpulist, this
abstracts the same code as a helper. Along with various code style
fixes, and error improvement (only error Physical CPU %d doesn't
exist if the specified CPU exceed the range, no
On 03/28/2013 07:36 AM, Osier Yang wrote:
vcpupin and emulatorpin use same code to parse the cpulist, this
How about The 'virsh vcpupin' and 'virsh emulatorpin' commands use the
same code to parse the cpulist. This patch
abstracts the same code as a helper. Along with various code style
On 03/29/2013 01:03 AM, Li Zhang wrote:
Hi all,
I am recently considering about the default setting for different
platforms.
Currently, there is no good way to fix the default setting for PPC64
platform.
For example, we prefer pseries on PPC64 as the default machine type
users don't need
On 03/31/2013 03:00 AM, Michal Privoznik wrote:
This is just a bare Easter Egg. Whenever user run virDomainScreenshot
s/user run/a user runs/
over a domain in test driver, he'll get the Libvirt PNG logo in return.
---
docs/Makefile.am | 1 +
src/test/test_driver.c | 24
On 03/31/2013 10:35 AM, Peter Krempa wrote:
The limits are documented at
http://libvirt.org/formatdomain.html#elementsCPUTuning . Enforce them
when going through XML parsing in addition to being enforced by the API.
---
Notes:
Version 2:
- split out from the conf callback series
On 01/04/2013, at 11:02 AM, Ruben Kerkhof wrote:
snip
Works on 10.8 as well.
[ruben@odin ~]$ brew install libvirt
== Installing libvirt dependency: cmake
Huh, I know it's April 1st, but did we just switch away from autotools? ;-)
Unrelated, but Justin, I noticed that your homebrew
Code added by commit id '523207fe8'
TEST: qemuxml2argvtest
40
80
120
160
On Mon, Apr 1, 2013 at 7:23 PM, Justin Clift jcl...@redhat.com wrote:
I'd really like to hear how it goes with libvirt. If you edit
the libvirt formula so that --without-qemu is --with-qemu:
/usr/local/Library/Formula/libvirt.rb
Then install it again (after brew install qemu), it *might*
On 01/04/2013, at 6:42 PM, Ruben Kerkhof wrote:
snip
Hey Justin,
Just tried that, but it has no effect, probably due to:
if test $with_osx = yes; then
with_qemu=no
fi
in configure.ac
I tried to build the latest libvirt HEAD from source, but I'm bumping into an
(unrelated)
On 04/01/2013 11:38 AM, John Ferlan wrote:
Code added by commit id '523207fe8'
Definitely helpful.
Interesting side note is that running the test singularly via 'make -C tests
check TESTS=qemuxml2argvtest' didn't trip the valgrind error; however,
running during 'make -C tests valgrind' did
On 03/31/2013 03:22 PM, Peter Krempa wrote:
The man page states that with --config the next boot is affected. This
can be understood as if _only_ the next boot was affected. This isn't
true if the machine is running.
This patch adds the full --live, --config, --current infrastructure and
On 03/31/2013 03:22 PM, Peter Krempa wrote:
Use the established approach to improve this function too.
---
Notes:
Version 2:
- same changes as previous patch
tools/virsh-domain.c | 41 -
tools/virsh.pod | 19 +++
2
On 03/31/2013 03:22 PM, Peter Krempa wrote:
Use the established approach to improve this function too.
---
Notes:
Version 2:
- state that legacy API has no well defined semantics
- add note about possible semantics change
- remove mutual exclusivity of --persistent --live
On 03/31/2013 03:22 PM, Peter Krempa wrote:
Use the established approach to improve this function too.
---
Notes:
Version 2:
- same notes as previous patch
tools/virsh-domain.c | 54
+++-
tools/virsh.pod | 19
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 sync with best upstream practices.
* .gnulib: Update to latest.
* bootstrap: Resynchronize.
---
Here's what changed in gnulib:
* .gnulib 819b1c3...076ac82
On 03/21/2013 11:42 AM, Stefan Berger wrote:
[snip]
+
+
+static int
+qemuMonitorJSONGetStringArray(qemuMonitorPtr mon, const char *qmpCmd,
+ char ***array)
+{
+int ret;
+virJSONValuePtr cmd;
+virJSONValuePtr reply = NULL;
+virJSONValuePtr data;
+
As the work to get X86CPU subclasses static properties defined is taking very
long to get reviewed and we risk not getting it finished in 1.5, I am sending
an alternative mechanism to allow libvirt to probe for CPU model information.
It's interesting to note that this alternative solution is
FEAT_7_0_EBX uses ECX as input, so we have to take that into account
when reporting feature word values.
Signed-off-by: Eduardo Habkost ehabk...@redhat.com
---
target-i386/cpu.c | 10 +++---
1 file changed, 7 insertions(+), 3 deletions(-)
diff --git a/target-i386/cpu.c b/target-i386/cpu.c
On 03/31/2013 05:22 PM, Peter Krempa wrote:
The man page states that with --config the next boot is affected. This
can be understood as if _only_ the next boot was affected. This isn't
true if the machine is running.
You should probably change this comment to make it clear that you're
changing
This property will be useful for libvirt, as libvirt already has logic
based on low-level feature bits (not feature names), so it will be
really easy to convert the current libvirt logic to something using the
feature-words property.
The property will have two main use cases:
- Checking host
On 03/21/2013 11:42 AM, Stefan Berger wrote:
Signed-off-by: Stefan Bergerstef...@linux.vnet.ibm.com
---
docs/formatdomain.html.in | 57
++
docs/schemas/domaincommon.rng | 43 +++
2 files changed, 100
The array will be used by the feature-words properties to output the
register enum value.
Signed-off-by: Eduardo Habkost ehabk...@redhat.com
---
target-i386/cpu.c | 25 +
1 file changed, 13 insertions(+), 12 deletions(-)
diff --git a/target-i386/cpu.c b/target-i386/cpu.c
On 04/01/2013 03:45 PM, Corey Bryant wrote:
+if (ret 0)
+goto cleanup;
+
+ret = -1;
+
+if (!(data = virJSONValueObjectGet(reply, return))) {
+virReportError(VIR_ERR_INTERNAL_ERROR,
+ _(%s reply was missing return data),
+
On 03/31/2013 04:48 PM, Peter Krempa wrote:
This patch is the result of running:
for i in $(git ls-files | grep -v html | grep -v \.po$ ); do
sed -i -e s/virDomainXMLConf/virDomainXMLOption/g -e s/xmlconf/xmlopt/g
$i
done
---
Notes:
Version 5:
- new in series
- suggested
On 04/01/2013 01:52 PM, Stefan Berger wrote:
On 04/01/2013 03:45 PM, Corey Bryant wrote:
+if (ret 0)
+goto cleanup;
+
+ret = -1;
+
+if (!(data = virJSONValueObjectGet(reply, return))) {
+virReportError(VIR_ERR_INTERNAL_ERROR,
+ _(%s reply
On 03/21/2013 11:42 AM, Stefan Berger wrote:
Signed-off-by: Stefan Bergerstef...@linux.vnet.ibm.com
---
po/POTFILES.in |1
src/Makefile.am |1
src/libvirt_private.syms |4 +
src/util/virtpm.c| 124
+++
On 03/21/2013 11:42 AM, Stefan Berger wrote:
Signed-off-by: Stefan Bergerstef...@linux.vnet.ibm.com
---
src/security/security_dac.c | 53 ++
src/security/security_selinux.c | 96
2 files changed, 149 insertions(+)
On 03/31/2013 04:48 PM, Peter Krempa wrote:
Move the xmlconf and caps arguments to the end of the argument list.
---
Did you try moving this and 11/11 to the front of the patches? If it
doesn't create too large of a rebase headache, that would make review
much simpler. (If it causes too many
On 04/01/2013 02:29 PM, Eric Blake wrote:
On 04/01/2013 11:38 AM, John Ferlan wrote:
Code added by commit id '523207fe8'
Definitely helpful.
Interesting side note is that running the test singularly via 'make -C tests
check TESTS=qemuxml2argvtest' didn't trip the valgrind error; however,
On 03/31/2013 10:20 AM, Peter Krempa wrote:
From: Viktor Mihajlovski mihaj...@linux.vnet.ibm.com
By adjusting the reference count of the connection object we
prevent races between callback function and virConnectClose.
Signed-off-by: Viktor Mihajlovski mihaj...@linux.vnet.ibm.com
---
Reported by Anthony Messina in
https://bugzilla.redhat.com/show_bug.cgi?id=904692
Present since introduction of smartcard support in commit f5fd9baa
* src/qemu/qemu_command.c (qemuBuildCommandLine): Match qemu spelling.
---
No one has spotted my bug, latent since Feb 2011, until now. I
guess
On 03/31/2013 10:20 AM, Peter Krempa wrote:
---
src/remote/remote_driver.c | 5 +
src/rpc/virnetclient.c | 9 -
2 files changed, 13 insertions(+), 1 deletion(-)
While I agree that this must not be applied upstream, I applied it
locally and was indeed able to more reliably
On 03/31/2013 10:20 AM, Peter Krempa wrote:
From: Viktor Mihajlovski mihaj...@linux.vnet.ibm.com
Before closing the connection we unregister the close callback
to prevent a reference leak.
Further, the messages on virConnectClose != 0 are a bit more specific
now.
Signed-off-by: Viktor
On 03/31/2013 10:20 AM, Peter Krempa wrote:
The function is used to establish connection so it should be in the main
virsh file. This movement also enables further improvements done in next
patches.
---
tools/virsh-host.c | 67
On 03/31/2013 10:20 AM, Peter Krempa wrote:
This patch improves the error message after disconnecting from the
hypervisor and adds the close callback operations required not to leak
the callback reference.
---
tools/virsh.c | 22 --
1 file changed, 16 insertions(+), 6
On 03/31/2013 10:20 AM, Peter Krempa wrote:
The last Viktor's effort to fix the race and memory corruption unfortunately
wasn't complete in the case the close callback was not registered in an
connection. At that time, the trail of event's that I'll describe later could
still happend and
On Mon, Apr 1, 2013 at 10:58 AM, Justin Clift jcl...@redhat.com wrote:
On 01/04/2013, at 6:42 PM, Ruben Kerkhof wrote:
snip
Hey Justin,
Just tried that, but it has no effect, probably due to:
if test $with_osx = yes; then
with_qemu=no
fi
in configure.ac
I tried to build
On Mon, Apr 01, 2013 at 08:00:53PM +0800, Han Cheng wrote:
The only parameter in -drive affect scsi-generic is readonly. Introduce
readonly/ to hostdev.
The helper function to look up disk controller model may be used by scsi
hostdev. But it should be changed to use info.
Signed-off-by: Han
On 04/02/2013 10:15 AM, Hu Tao wrote:
On Mon, Apr 01, 2013 at 08:00:53PM +0800, Han Cheng wrote:
--- a/docs/formatdomain.html.in
+++ b/docs/formatdomain.html.in
@@ -2257,6 +2257,9 @@
codeid/code attribute that specifies the USB vendor and product id.
The ids can be given in
On Mon, Apr 01, 2013 at 08:00:55PM +0800, Han Cheng wrote:
Adding scsi hostdev, it should like:
hostdev mode='subsystem' type='scsi'
source
adapter name='scsi_host0'/
address bus='0' target='0' unit='0'/
/source
address type='drive' controller='0'
On 01/04/13 20:00, Han Cheng wrote:
The definiton of scsi adapter in storagespool.rng can be used by scsi
hostdev. Split storagepool.rng into storagepool.rng and
storagepoolcommon.rng and make domaincomman.rng include the latter one.
Don't think it worth having a separate file, IMHO just
*** NONE ***
Osier Yang (3):
util: Add a helper to check if all bits of a bitmap are clear
qemu: Error out if the bitmap for pinning is all clear
virsh: Prohibit all clear cpumap
src/libvirt_private.syms | 1 +
src/qemu/qemu_driver.c | 12
src/util/virbitmap.c | 30
---
src/libvirt_private.syms | 1 +
src/util/virbitmap.c | 30 ++
src/util/virbitmap.h | 3 +++
3 files changed, 34 insertions(+)
diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms
index 96eea0a..35ac957 100644
--- a/src/libvirt_private.syms
For both live and config changes of vcpupin and emulatorpin, an
all clear bitmap doesn't make sense, and it can just cause corruptions.
E.g (similar for emulatorpin).
% virsh vcpupin hame 0 8,^8 --config
% virsh vcpupin hame
VCPU: CPU Affinity
--
0:
1: 0-63
This prohibits all clear cpumap eariler in virsh, for both vcpupin
and emulatorpin.
---
tools/virsh-domain.c | 35 ++-
1 file changed, 30 insertions(+), 5 deletions(-)
diff --git a/tools/virsh-domain.c b/tools/virsh-domain.c
index 965f92c..8d3d63f 100644
---
On 02/04/13 00:55, John Ferlan wrote:
On 03/28/2013 07:36 AM, Osier Yang wrote:
vcpupin and emulatorpin use same code to parse the cpulist, this
How about The 'virsh vcpupin' and 'virsh emulatorpin' commands use the
same code to parse the cpulist. This patch
Okay.
abstracts the same code
66 matches
Mail list logo