On Tue, Sep 13, 2016 at 05:08:02PM +0200, Peter Krempa wrote:
See patch 2.
Peter Krempa (2):
util: numa: Remove impossible error handling
numa: Rename virNumaGetHostNodeset and make it return only nodes with
memory
Looks sane, ACK.
src/libvirt_private.syms | 2 +-
src/qemu/qemu_cgroup.
The automatic nodeset would creep into the persistent definition for live VMs
with automatic pinning which is incorrect.
https://bugzilla.redhat.com/show_bug.cgi?id=1365779
Peter Krempa (4):
conf: Introduce virDomainObjGetOneDefState
qemu: domain: Add macro to simplify access to vm private da
Calling virDomainGetEmulatorPinInfo on a live VM with automatic NUMA
pinning and VIR_DOMAIN_AFFECT_CONFIG would return the automatic pinning
data in some cases which is bogus. Use the autoCpuset property only when
called on a live definition.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1
Sometimes adding a separate variable to access vm->privateData is not
necessary. Add a macro that will do the typecasting rather than having
to add a temp variable to force the compiler to typecast it.
---
src/qemu/qemu_domain.h | 3 +++
1 file changed, 3 insertions(+)
diff --git a/src/qemu/qemu_
Return whether the live or persistent definition was returned. Sometimes
it's necessary to base the decisions on this.
---
src/conf/domain_conf.c | 40
src/conf/domain_conf.h | 3 +++
src/libvirt_private.syms | 1 +
3 files changed, 40 insertions(+),
Calling virDomainGetVcpuPinInfo on a live VM with automatic NUMA pinning
and VIR_DOMAIN_AFFECT_CONFIG would return the automatic pinning data
in some cases which is bogus. Use the autoCpuset property only when
called on a live definition.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=13657
At 09/14/2016 11:35 AM, Peter Krempa wrote:
On Wed, Sep 14, 2016 at 10:26:51 +0800, Dou Liyang wrote:
Hi Peter,
At 09/14/2016 12:27 AM, Peter Krempa wrote:
Return whether a vcpu entry is hotpluggable or online so that upper
layers don't have to infer the information from other data.
Advanta
On Wed, Sep 14, 2016 at 10:26:51 +0800, Dou Liyang wrote:
> Hi Peter,
>
> At 09/14/2016 12:27 AM, Peter Krempa wrote:
> > Return whether a vcpu entry is hotpluggable or online so that upper
> > layers don't have to infer the information from other data.
> >
> > Advantage is that this code can be t
Hi Peter,
At 09/14/2016 12:27 AM, Peter Krempa wrote:
Return whether a vcpu entry is hotpluggable or online so that upper
layers don't have to infer the information from other data.
Advantage is that this code can be tested by unit tests.
---
src/qemu/qemu_monitor.c
On 09/13/2016 05:31 PM, Eric Blake wrote:
[hit send too soon]
> Okay, I have confirmed that this prime-the-cache idea DOES work, using
> libvirt.git commit 419bc8cf (one commit prior to d53fa838 which
> installed a hack into libvirt to force the use of -Werror for the
> duration of AC_HEADER_MAJO
[adding libvirt list, as this is what sparked my investigations so far
today]
On 09/13/2016 04:36 PM, Eric Blake wrote:
> One other possibility that distros can do is to prime a config.site
> file, with $ac_cv_header_sys_types_h_makedev=no, to forcefully bypass
> the configure check that is norma
something weird with your email client and adding extra 's
everywhere... Plaintext is your friend as much as 'git send-email'...
On 08/01/2016 10:20 PM, weifuqiang wrote:
> [PATCH] qemu: fix libvirtd crash in migration after vm shutdown
>
>
>
>
>
> If we shutdown a guest, then migrate it w
On 09/13/2016 04:49 PM, John Ferlan wrote:
[...]
I took a peek at the v1, but suffice to say didn't follow it that
closely since there's a number of changes in v2 related to patch
ordering. You could get 3 different opinions on that matter too!
Hopefully a consensus before too long. :)
ye
[...]
>>> diff --git a/src/util/virscsi.c b/src/util/virscsi.c
>>> index 4843367..290b692 100644
>>> --- a/src/util/virscsi.c
>>> +++ b/src/util/virscsi.c
>>> @@ -105,6 +105,32 @@ virSCSIDeviceGetAdapterId(const char *adapter,
>>> return -1;
>>> }
>>> +int
>>> +virSCSIOpenVhost(int *vhos
[...]
>>>
>> I took a peek at the v1, but suffice to say didn't follow it that
>> closely since there's a number of changes in v2 related to patch
>> ordering. You could get 3 different opinions on that matter too!
>
> Hopefully a consensus before too long. :)
>
yeah - v1 was lots of little st
On Mon, Aug 29, 2016 at 12:45:03 -0400, John Ferlan wrote:
>
>
> On 08/12/2016 09:33 AM, Jiri Denemark wrote:
> > Changing a host architecture or a CPU is not as easy as assigning a new
> > value to the appropriate element in virCaps since there is a relation
> > between the CPU and host architec
On 09/05/2016 11:45 PM, Chen Hanxiao wrote:
> From: Chen Hanxiao
>
> We try to find a snapshot that had no different between
> the current state of RBD image.
> If we failed in rbd_diff_iterate, just continue for the
> next search iteration.
>
> Signed-off-by: Chen Hanxiao
> ---
> src/storag
On Mon, Aug 29, 2016 at 12:32:34 -0400, John Ferlan wrote:
...
> > -size_t virQEMUCapsGetCPUDefinitions(virQEMUCapsPtr qemuCaps,
> > -char ***names)
> > +int
> > +virQEMUCapsGetCPUDefinitions(virQEMUCapsPtr qemuCaps,
> > + char ***name
On 09/12/2016 06:10 PM, John Ferlan wrote:
On 09/06/2016 08:58 AM, Eric Farman wrote:
Open /dev/vhost-scsi, and record the resulting file descriptor, so that
the guest has access to the host device outside of the libvirt daemon.
Pass this information, along with data parsed from the XML file,
On Mon, Aug 29, 2016 at 12:31:44 -0400, John Ferlan wrote:
> > +int
> > +virDomainCapsCPUModelsAddSteal(virDomainCapsCPUModelsPtr cpuModels,
> > + char **name)
> > +{
> > +if (VIR_RESIZE_N(cpuModels->models, cpuModels->alloc,
> > + cpuModels->co
On 09/12/2016 05:34 PM, John Ferlan wrote:
On 09/06/2016 08:58 AM, Eric Farman wrote:
We already have a "scsi" hostdev type, which refers to a single LUN
that is passed through to a guest. But what of things where multiple
LUNs are passed through via a single SCSI HBA, such as with the
vhost-
On 09/13/2016 06:02 AM, Michal Privoznik wrote:
>> So, just to make it crystal clear, is this what you're saying?
>>
>> ready | initial C/R |fudged C/R
>
> Oh, This should have been C/E instead of C/R. Current/End. 'e' and 'r'
> keys are just too close to each other :-)
>
>> --+-
The algorithm that matches data from query-cpus and
query-hotpluggable-cpus is quite complex. Start using descriptive
iterator names to avoid confusion.
---
src/qemu/qemu_monitor.c | 25 +
1 file changed, 13 insertions(+), 12 deletions(-)
diff --git a/src/qemu/qemu_monitor
Return whether a vcpu entry is hotpluggable or online so that upper
layers don't have to infer the information from other data.
Advantage is that this code can be tested by unit tests.
---
src/qemu/qemu_monitor.c| 11 +
src/qemu/qemu_monitor.h
See patches 3 and 4 for explanation.
Peter Krempa (4):
qemu: monitor: Use a more obvious iterator name
qemu: monitor: qemuMonitorGetCPUInfoHotplug: Add iterator 'anycpu'
qemu: monitor: Add vcpu state information to monitor data
qemu: domain: Don't infer vcpu state
src/qemu/qemu_domain.c
Add separate iterator for iterating all the entries
---
src/qemu/qemu_monitor.c | 17 +
1 file changed, 9 insertions(+), 8 deletions(-)
diff --git a/src/qemu/qemu_monitor.c b/src/qemu/qemu_monitor.c
index b8da637..4489997 100644
--- a/src/qemu/qemu_monitor.c
+++ b/src/qemu/qemu_mo
Use the state information (online, hotpluggable) provided by the monitor
code rather than trying to infer it. This fixes a issue where on
architectures that require hotplug of multiple threads at once the
sub-cores would get updated as offline on daemon restart thus creating
an invalid configuratio
On 13.09.2016 16:11, Erik Skultety wrote:
> On 13/09/16 14:30, Michal Privoznik wrote:
>> On 12.09.2016 10:20, Erik Skultety wrote:
>>> Change the logic in a way, so that VSH_CMD_FLAG_ALIAS behaves similarly to
>>> how VSH_OT_ALIAS for command options, i.e. there is no need for code
>>> duplicatio
On 08/27/2016 03:13 AM, Chen Hanxiao wrote:
> From: Chen Hanxiao
>
> We check compress prog in qemuCompressProgramAvailable,
> then check it again in virExec.
>
> This path will pass compress prog's path directly.
>
> Signed-off-by: Chen Hanxiao
> ---
> src/qemu/qemu_driver.c | 42 +
On 09/13/2016 10:43 AM, Cornelia Huck wrote:
[I've browsed through the thread a bit, but as I'm not a libvirt
developer I may be missing some basic things]
On Wed, 7 Sep 2016 16:34:17 -0400
Laine Stump wrote:
On 09/07/2016 03:38 PM, Sascha Silbe wrote:
Dear Laine,
Laine Stump writes:
On
See patch 2.
Peter Krempa (2):
util: numa: Remove impossible error handling
numa: Rename virNumaGetHostNodeset and make it return only nodes with
memory
src/libvirt_private.syms | 2 +-
src/qemu/qemu_cgroup.c | 4 ++--
src/util/virnuma.c | 18 +++---
src/util/virnum
Make use of the new recently introduced alias handling for virt-admin srv-*
commands.
Signed-off-by: Erik Skultety
---
tools/virt-admin.c | 24
1 file changed, 24 insertions(+)
diff --git a/tools/virt-admin.c b/tools/virt-admin.c
index 12ec057..5d6beda 100644
--- a/tool
Change the logic in a way, so that VSH_CMD_FLAG_ALIAS behaves similarly to
how VSH_OT_ALIAS for command options, i.e. there is no need for code duplication
for the alias and the aliased command structures. Along with that change,
switch any existing VSH_CMD_FLAG_ALIAS occurrences to this new format
Name it virNumaGetHostMemoryNodeset and return only NUMA nodes which
have memory installed. This is necessary as the kernel is not very happy
to set the memory cgroup setting for nodes which do not have any memory.
This would break vcpu hotplug with following message on such
configruation:
Inva
the original version:
https://www.redhat.com/archives/libvir-list/2016-September/msg00312.html
since v1:
- tweaked the virsh-self-test so that it also checks the aliased commands
instead of skipping them (since there was a good reason for that before the
changes this series introduces)
Since the old command names are being shadowed by the new ones in the code as
well as in the help messages, update the man page accordingly.
Signed-off-by: Erik Skultety
---
tools/virt-admin.pod | 30 +++---
1 file changed, 15 insertions(+), 15 deletions(-)
diff --git a/
The code guarantees that virBitmapSetBit won't be called with out of
range values. Just ignore the return value and remove dead error
handling.
---
src/util/virnuma.c | 7 +--
1 file changed, 1 insertion(+), 6 deletions(-)
diff --git a/src/util/virnuma.c b/src/util/virnuma.c
index fc25051..c4
[I've browsed through the thread a bit, but as I'm not a libvirt
developer I may be missing some basic things]
On Wed, 7 Sep 2016 16:34:17 -0400
Laine Stump wrote:
> On 09/07/2016 03:38 PM, Sascha Silbe wrote:
> > Dear Laine,
> >
> > Laine Stump writes:
> >
> >> On 09/07/2016 02:35 PM, Sascha S
On Tue, Sep 13, 2016 at 04:31:45PM +0200, Michal Privoznik wrote:
> Thing is, in f3f15cc24 I'm trying to adapt libvirt to the newest
> glibc where major()/minor()/makedev() are moved from sys/types.h
> to sys/sysmacros.h. However, my commit back then expect autoconf
> to be fixed too as we already
Thing is, in f3f15cc24 I'm trying to adapt libvirt to the newest
glibc where major()/minor()/makedev() are moved from sys/types.h
to sys/sysmacros.h. However, my commit back then expect autoconf
to be fixed too as we already use AC_HEADER_MAJOR to determine
which header file the functions are in, b
On 13/09/16 14:30, Michal Privoznik wrote:
> On 12.09.2016 10:20, Erik Skultety wrote:
>> Change the logic in a way, so that VSH_CMD_FLAG_ALIAS behaves similarly to
>> how VSH_OT_ALIAS for command options, i.e. there is no need for code
>> duplication
>> for the alias and the aliased command struc
On 12.09.2016 10:20, Erik Skultety wrote:
> The virsh-self-test script compared the test's return code with 1 and only if
> the return code matched this value then the test was marked as failed. Problem
> is that SIGSEGV returns 139 (or 11 to be precise, since shell reserves the MSB
> for abnormal
On 12.09.2016 10:20, Erik Skultety wrote:
> Change the logic in a way, so that VSH_CMD_FLAG_ALIAS behaves similarly to
> how VSH_OT_ALIAS for command options, i.e. there is no need for code
> duplication
> for the alias and the aliased command structures. Along with that change,
> switch any exist
On 12.09.2016 10:20, Erik Skultety wrote:
> Since the old command names are being shadowed by the new ones in the code as
> well as in the help messages, update the man page accordingly.
>
> Signed-off-by: Erik Skultety
> ---
> tools/virt-admin.pod | 30 +++---
> 1 file c
Ping
As REBOOT is concerned with both qemu and libvirtd, it's not atomic job, thus,
it maybe not just qemu's bug.
Maybe libvirtd should also send RESET qmp command after migration, what do you
think of that?
We've tried to send fakeReboot to dest side to fix this problem, it seems to
have fi
On 13.09.2016 11:52, Michal Privoznik wrote:
> On 12.09.2016 21:19, Eric Blake wrote:
>> On 09/09/2016 04:30 PM, John Ferlan wrote:
>>
+/* Fix job completeness reporting. If cur == end mgmt
+ * applications think job is completed. Except when both cur
+ * and end a
---
Pushed as trivial.
src/qemu/qemu_cgroup.c | 6 +++---
src/qemu/qemu_cgroup.h | 2 +-
src/qemu/qemu_driver.c | 2 +-
src/qemu/qemu_process.c | 2 +-
4 files changed, 6 insertions(+), 6 deletions(-)
diff --git a/src/qemu/qemu_cgroup.c b/src/qemu/qemu_cgroup.c
index e2b5bab..fe94613 100644
On Mon, Sep 12, 2016 at 05:15:53PM +0200, Michal Privoznik wrote:
> On 12.09.2016 16:54, Daniel P. Berrange wrote:
> > On Tue, Sep 06, 2016 at 03:46:59PM +0200, Michal Privoznik wrote:
> >> In the latest glibc, major() and minor() functions are marked as
> >> deprecated (glibc commit dbab6577):
> >
On 12.09.2016 21:19, Eric Blake wrote:
> On 09/09/2016 04:30 PM, John Ferlan wrote:
>
>>>
>>> +/* Fix job completeness reporting. If cur == end mgmt
>>> + * applications think job is completed. Except when both cur
>>> + * and end are zero, in which case qemu hasn't started the
>>> +
On 09/07/2016 02:04 PM, Peter Krempa wrote:
When hot-adding vcpus qemu needs to allocate some structures in the DMA
zone which may be outside of the numa pinning. Extract the code doing
this in a set of helpers so that it can be reused.
---
src/qemu/qemu_cgroup.c | 90 +++
On 12.09.2016 10:20, Michal Privoznik wrote:
> On 27.05.2016 10:05, Nikolay Shirokovskiy wrote:
>> There is already a patch [1] on this topic with a different approach - keep
>> nvram file by default. There is also some discussion there. To sum up keeping
>> nvram on undefine could be useful in s
51 matches
Mail list logo