On Tue, Aug 16, 2016 at 13:45:10 -0400, John Ferlan wrote:
> Modify virDomainDefGetVcpuSched to emit an error message if
> virDomainDefGetVcpu returns NULL meaning the vcpu could not
> be found. Prior to commit id '9cc931f0b' the error message
> would have been issued in virDomainDefGetVcpu.
>
> S
On 17.08.2016 02:36, Luyao Huang wrote:
> Hi Michal,
>
> I remember i have used vhost-user-bridge to test vhost-user before.
> And you can find it in qemu source code tests/vhost-user-bridge.c, but
> i am not sure if you can avoid this crash with vhost-user-bridge :)
Hey, that's is awesome! Exact
Hi all,
I've found that if I create a QEMU guest that:
(a) is backed by host hugepages and
(b) specifies a host NUMA memory placement policy
... then the NUMA policy (b) is ignored. It's even possible to start the guest
if the policy specifies that memory must come from non-existant host nodes.
On 08/16/2016 02:59 PM, John Ferlan wrote:
> https://bugzilla.redhat.com/show_bug.cgi?id=1182074
>
> Since libvirt still uses a legacy qemu arg format to add a disk, the
> manner in which the 'password-secret' argument is passed to qemu needs
> to change to prepend a 'file.' If in the future, usage
xen-xm doesn't support mult serial devices at all, so these tests are
meaningless.
Signed-off-by: Bob Liu
---
.../test-fullvirt-serial-dev-2-ports.cfg | 27 -
.../test-fullvirt-serial-dev-2-ports.xml | 65 --
.../test-fullvirt-serial-dev-2nd-port.c
Add support for multi serial devices, after this patch virsh can be used to
connect different serial devices of running domains. E.g.
vish # console --devname serial
Note:
This depends on a xen/libxl bug fix to have libxl_console_get_tty(...) correctly
returning the tty path (as opposed to always
xen-xm doesn't support multi serial at all, this patch drop the
domXML <-> xl.cfg conversions.
Signed-off-by: Bob Liu
---
src/xenconfig/xen_common.c | 21 +
1 file changed, 17 insertions(+), 4 deletions(-)
diff --git a/src/xenconfig/xen_common.c b/src/xenconfig/xen_common.c
Adding tests for domXML <-> xl.cfg conversions containing multiple
serial devices.
Signed-off-by: Bob Liu
---
tests/xlconfigdata/test-fullvirt-multiserial.cfg | 25 +
tests/xlconfigdata/test-fullvirt-multiserial.xml | 64
tests/xlconfigtest.c
Hi Michal,
I remember i have used vhost-user-bridge to test vhost-user before.
And you can find it in qemu source code tests/vhost-user-bridge.c, but
i am not sure if you can avoid this crash with vhost-user-bridge :)
BR,
Luyao
- Original Message -
From: "Michal Privoznik"
To: libvir-li
On 08/15/2016 07:28 AM, Pavel Hrdina wrote:
> Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1364843
>
Similar to 4/5 a few extra words here would help so one doesn't have to
chase into the bz to attempt to figure out what's being fixed and why
moving the Setup earlier does what you want
Could use a little bit of explanation here! Details as to why we can
ignore those glisten types...
John
On 08/15/2016 07:28 AM, Pavel Hrdina wrote:
> Signed-off-by: Pavel Hrdina
> ---
> src/qemu/qemu_process.c | 11 +++
> 1 file changed, 11 insertions(+)
>
> diff --git a/src/qemu/qem
On 08/15/2016 07:28 AM, Pavel Hrdina wrote:
> Signed-off-by: Pavel Hrdina ---
> src/qemu/qemu_process.c | 44 +---
> 1 file changed, 25 insertions(+), 19 deletions(-)
>
ACK - couple of nits below to consider...
John
> diff --git a/src/qemu/qemu_process
https://bugzilla.redhat.com/show_bug.cgi?id=1182074
Since libvirt still uses a legacy qemu arg format to add a disk, the
manner in which the 'password-secret' argument is passed to qemu needs
to change to prepend a 'file.' If in the future, usage of the more
modern disk format, then the prepended
On Tue, Aug 16, 2016 at 07:55:11PM +0200, Andrea Bolognani wrote:
On Tue, 2016-08-16 at 18:49 +0200, Michal Privoznik wrote:
On 16.08.2016 13:40, Andrea Bolognani wrote:
>
> The first argument should be const char ** instead of
> char **, because this is a search function and as such it
> doesn
On 08/16/2016 01:29 PM, Andrea Bolognani wrote:
On Mon, 2016-08-15 at 01:50 -0400, Laine Stump wrote:
libvirt previously assigned nearly all devices to a hotpluggable
legacy PCI slot even on machines with a PCIe root complex. Doing this
means that the domain will need a dmi-to-pci-bridge (to con
On Tue, 2016-08-16 at 17:45 +0200, Michal Privoznik wrote:
> Dear list,
>
> I've set up a small script running on my Raspberry that produces
> following stats:
>
> https://mprivozn.fedorapeople.org/libvirt.gitstats/index.html
>
> It's polling git every 3 hours (which is about time it takes to ge
On Tue, 2016-08-16 at 18:49 +0200, Michal Privoznik wrote:
> On 16.08.2016 13:40, Andrea Bolognani wrote:
> >
> > The first argument should be const char ** instead of
> > char **, because this is a search function and as such it
> > doesn't, and shouldn't, alter the haystack in any way.
> >
> >
Modify virDomainDefGetVcpuSched to emit an error message if
virDomainDefGetVcpu returns NULL meaning the vcpu could not
be found. Prior to commit id '9cc931f0b' the error message
would have been issued in virDomainDefGetVcpu.
Signed-off-by: John Ferlan
---
v1: http://www.redhat.com/archives/libv
On Mon, 2016-08-15 at 01:50 -0400, Laine Stump wrote:
> libvirt previously assigned nearly all devices to a hotpluggable
> legacy PCI slot even on machines with a PCIe root complex. Doing this
> means that the domain will need a dmi-to-pci-bridge (to convert from
> PCIe to legacy PCI) and a pci-bri
On 16.08.2016 13:40, Andrea Bolognani wrote:
> The first argument should be const char ** instead of
> char **, because this is a search function and as such it
> doesn't, and shouldn't, alter the haystack in any way.
>
> This change means we no longer have to cast arrays of
> immutable strings to
Dear list,
I've set up a small script running on my Raspberry that produces
following stats:
https://mprivozn.fedorapeople.org/libvirt.gitstats/index.html
It's polling git every 3 hours (which is about time it takes to generate
the stats there :D). Don't expect interactive stats regeneration, bu
Currently, what we do for vhost-user network is generate the
following part of command line:
-netdev type=vhost-user,id=hostnet0,chardev=charnet0
There's no need for 'type=' it is the default. Drop it.
Signed-off-by: Michal Privoznik
---
src/qemu/qemu_command.c
https://bugzilla.redhat.com/show_bug.cgi?id=1366505
So far, this function lacked support for
VIR_DOMAIN_NET_TYPE_VHOSTUSER leaving callers to hack around the
problem by constructing the command line on their own. This is
not ideal as it blocks hot plug support.
Signed-off-by: Michal Privoznik
--
Instead of blindly claim support for hot-plugging of every
interface type out there we should copy approach we have for
device types: white listing supported types and explicitly error
out on unsupported ones.
For instance, trying to hotplug vhostuser interface results in
nothing usable from guest
We tend to prevent using 'default' in switches. And it is for a
good reason - control may end up in paths we wouldn't want for
new values. In this specific case, if qemuBuildHostNetStr is
called over VIR_DOMAIN_NET_TYPE_VHOSTUSER it would produce
meaningless output. Fortunately, there no such call
Signed-off-by: Michal Privoznik
---
src/qemu/qemu_command.c | 64 -
1 file changed, 32 insertions(+), 32 deletions(-)
diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c
index 59fcb4c..e6b3d9d 100644
--- a/src/qemu/qemu_command.c
+++ b/s
This alone makes not much sense. But the aim is to reuse this
function in qemuBuildVhostuserCommandLine() where 'nowait' is not
supported for vhost-user devices.
Signed-off-by: Michal Privoznik
---
src/qemu/qemu_command.c | 34 ++
1 file changed, 18 insertions(+),
Depending on domain OS type, and interface address type we might
not want to use -netdev even though qemu has the capability. We
should use more advanced check implemented in
qemuDomainSupportsNetdev() function.
Signed-off-by: Michal Privoznik
---
src/qemu/qemu_hotplug.c | 12 ++--
1 fil
https://bugzilla.redhat.com/show_bug.cgi?id=1366108
So far, the hotplug of vhost-user "worked" by pure chance. Well,
qemu would error on our commands, but nevertheless. Now that we
have everything prepare, We should support hotplugging og
vhost-user. Firstly, we need to plug in chardev (through wh
There's no need to reinvent the wheel here. We already have a
function to format virDomainChrSourceDefPtr. It's called
qemuBuildChrChardevStr(). Use that instead of some dummy
virBufferAsprintf().
Signed-off-by: Michal Privoznik
---
src/qemu/qemu_command.c | 16 +---
1 file changed,
While working on this, I've discovered a qemu crasher [1], so
there are not tested properly. If you have pointers how to test
vhost-user without need to install openvswitch, I'm all ears.
1: http://lists.nongnu.org/archive/html/qemu-devel/2016-08/msg02933.html
Michal Privoznik (9):
qemuDomainAt
https://bugzilla.redhat.com/show_bug.cgi?id=1367259
Crash occurs because 'secrets' is being dereferenced in call:
if (qemuDomainSecretSetup(conn, priv, secinfo, disk->info.alias,
VIR_SECRET_USAGE_TYPE_VOLUME, NULL,
&src->
On Mon, Aug 15, 2016 at 12:03:53 -0400, John Ferlan wrote:
>
>
> On 08/15/2016 11:07 AM, Peter Krempa wrote:
> > On Mon, Aug 15, 2016 at 10:55:33 -0400, John Ferlan wrote:
> >> When commit id '6dfb4507' refactored where the iothreadsched data was
> >> stored, the error message for when the virDom
On Tue, Aug 16, 2016 at 02:35:12PM +0200, Peter Krempa wrote:
> As requested in the review for the vcpu hotplug series, split the API code
> even more.
Nice cleanup!
ACK series
Pavel
--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list
On 08/16/2016 10:17 AM, Andrea Bolognani wrote:
On Mon, 2016-08-15 at 01:50 -0400, Laine Stump wrote:
A pci-bridge has *almost* the same rules as a legacy PCI endpoint
device for where it can be automatically connected, and until now both
had been considered identical. There is one pairing that
On Mon, 2016-08-15 at 01:50 -0400, Laine Stump wrote:
> A pci-bridge has *almost* the same rules as a legacy PCI endpoint
> device for where it can be automatically connected, and until now both
> had been considered identical. There is one pairing that is okay when
> specifically requested by the
First, make function logPrlEventErrorHelper be void and only
print information (if any) from an event.
Second, don't rewrite original error with any errors we get
during parsing event info.
Third, ignore PRL_ERR_NO_DATA at all.
Signed-off-by: Maxim Nestratov
---
src/vz/vz_sdk.c | 33
Signed-off-by: Maxim Nestratov
---
src/vz/vz_driver.c | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/src/vz/vz_driver.c b/src/vz/vz_driver.c
index b1b6d14..cf4b9e8 100644
--- a/src/vz/vz_driver.c
+++ b/src/vz/vz_driver.c
@@ -600,6 +600,7 @@ static int
vzDomainGetInfo
We should not be too strict regarding device indices check
because it is not really necessary and in fact it breaks
some common scenarios when CD-ROMs are added as the only deivce
on IDE bus with /dev/hdd device name, which corresponds to
ide0-1-1 device indices, rather than ide0-0-0(hda).
Signed-
There is a possibility that qemu driver frees by unreferencing its
closeCallbacks pointer as it has the only reference to the object,
while in fact not all users of CloseCallbacks called thier
virCloseCallbacksUnset.
Backtrace is the following:
Thread #1:
0 in pthread_cond_wait@@GLIBC_2.3.2 () fr
Somehow we lost this during recent refactoring
Signed-off-by: Maxim Nestratov
---
src/vz/vz_sdk.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/src/vz/vz_sdk.c b/src/vz/vz_sdk.c
index 999fce1..2363f5c 100644
--- a/src/vz/vz_sdk.c
+++ b/src/vz/vz_sdk.c
@@ -978,6 +978,7 @@ prlsdkGetNetInfo
From: Yuri Pudgorodskiy
A separate error code will help recognize real failures from
necessity to try again
Signed-off-by: Maxim Nestratov
---
include/libvirt/virterror.h | 2 ++
src/qemu/qemu_agent.c | 6 +++---
src/util/virerror.c | 6 ++
3 files changed, 11 insertions(+),
The live code does ugly things. Contain it in a separate function.
---
src/qemu/qemu_driver.c | 121 +++--
1 file changed, 68 insertions(+), 53 deletions(-)
diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c
index b81baa5..95a7838 100644
--- a
All other modes of qemuDomainSetVcpusFlags have helpers so finish the
work by splitting the regular code into a new function.
This patch also touches up the coding (spacing) style.
---
src/qemu/qemu_driver.c | 100 ++---
1 file changed, 53 insertions(+)
Setting of the maximum vcpu count is slightly semantically different
thus split it into a self-contained func.
---
src/qemu/qemu_driver.c | 67 +-
1 file changed, 44 insertions(+), 23 deletions(-)
diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_
As requested in the review for the vcpu hotplug series, split the API code even
more.
Peter Krempa (3):
qemu: setvcpus: Extract setting of maximum vcpu count
qemu: driver: Extract setting of live vcpu count
qemu: driver: Split out regular vcpu hotplug code into a function
src/qemu/qemu_dr
v1-v2 changes:
- fixed "vz: reset errors after ignoring return values"
- added "vz: relax disk bus controller and device indices check"
Maxim Nestratov (6):
vz: get additional error information from job correctly
util: fix crash in virClassIsDerivedFrom for CloseCallbacks objects
vz: don't f
10-Aug-16 20:44, Maxim Nestratov пишет:
Maxim Nestratov (5):
vz: get additional error information from job correctly
util: fix crash in virClassIsDerivedFrom for CloseCallbacks objects
vz: don't fail query domain info in case we don't have valid stats
handle
vz: reset errors aft
If we are going to ignore return value of a functions
that can raise an error, it's not enough to use ignore_value
construction. We should explicitly call virResetLastError
Signed-off-by: Maxim Nestratov
---
src/vz/vz_driver.c | 5 +
src/vz/vz_sdk.c| 7 +++
2 files changed, 12 insert
On Sun, Aug 14, 2016 at 04:14:55AM +0200, Tomasz Flendrich wrote:
We can as well generate the number of buses inside that function. Fewer
parameters will be useful when recalculating the pci address set on
demand.
---
src/qemu/qemu_domain_address.c | 17 +
1 file changed, 13 insert
On Thu, 2016-08-11 at 13:57 +0200, Ján Tomko wrote:
> Check whether the disable-legacy property is present on the following
> devices:
> virtio-balloon-pci
> virtio-blk-pci
> virtio-scsi-pci
> virtio-serial-pci
> virtio-9p-pci
> virtio-net-pci
> virtio-rng-pci
> virtio-gpu-pci
> v
[...]
>
> It appears that the libvirt code for CURL is also using the legacy syntax
> so probably needs the same fix too. It looks like ISCSI takes the same
> codepath in libvirt as CURL, so probably nedes the same fix.
>
> I can't tell for sure though, since it appears we never added any unit
The first argument should be const char ** instead of
char **, because this is a search function and as such it
doesn't, and shouldn't, alter the haystack in any way.
This change means we no longer have to cast arrays of
immutable strings to arrays of mutable strings; we still
have to do the oppos
On 08/16/2016 07:25 AM, Daniel P. Berrange wrote:
> On Tue, Aug 16, 2016 at 12:14:50PM +0100, Daniel P. Berrange wrote:
>> On Wed, Aug 10, 2016 at 04:01:11PM -0600, Jim Fehlig wrote:
>>> Hi John,
>>>
>>> I've been having problems with rbd auth since the change to using qemu's
>>> secret
>>> obje
On Tue, Aug 16, 2016 at 13:25:55 +0200, Ján Tomko wrote:
> On Tue, Aug 16, 2016 at 09:47:52AM +0200, Peter Krempa wrote:
> >qemu uses 'url' instead of 'uri'. They unfortunately look very similar.
> >
> >Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1367260
> >---
> > src/util/virstoragefile
On Mon, Aug 15, 2016 at 05:35:56PM -0600, Jim Fehlig wrote:
> On 08/15/2016 03:38 PM, John Ferlan wrote:
> >
> > On 08/10/2016 06:01 PM, Jim Fehlig wrote:
> >> Hi John,
> >>
> >> I've been having problems with rbd auth since the change to using qemu's
> >> secret
> >> objects. E.g. when hotpluggin
On Tue, Aug 16, 2016 at 09:47:52AM +0200, Peter Krempa wrote:
qemu uses 'url' instead of 'uri'. They unfortunately look very similar.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1367260
---
src/util/virstoragefile.c | 2 +-
tests/virstoragetest.c| 6 +++---
2 files changed, 4 inserti
On Tue, Aug 16, 2016 at 12:14:50PM +0100, Daniel P. Berrange wrote:
> On Wed, Aug 10, 2016 at 04:01:11PM -0600, Jim Fehlig wrote:
> > Hi John,
> >
> > I've been having problems with rbd auth since the change to using qemu's
> > secret
> > objects. E.g. when hotplugging disk config
> >
> >
> >
On Sun, Aug 14, 2016 at 04:14:54AM +0200, Tomasz Flendrich wrote:
With a parameter, we can decide whether we want
qemuDomainValidateDevicePCISlotsChipsets to assign new addresses, or
simply verify the existing ones and reserve slots for potential future
devices.
The PCI address set will be recalc
On Thu, Aug 11, 2016 at 10:02:55 +0200, Pavel Hrdina wrote:
> On Fri, Aug 05, 2016 at 03:56:13PM +0200, Peter Krempa wrote:
> > Setting of the maximum vcpu count is slightly semantically different
> > thus split it into a self-contained func.
> > ---
> > src/qemu/qemu_driver.c | 47 +++
On Wed, Aug 10, 2016 at 04:01:11PM -0600, Jim Fehlig wrote:
> Hi John,
>
> I've been having problems with rbd auth since the change to using qemu's
> secret
> objects. E.g. when hotplugging disk config
>
>
>
>name="volumes/volume-f9c33a0a-5313-44fc-9624-c3b09ed21a57">
>
>
>
>
On Mon, Aug 15, 2016 at 11:01:22AM +0200, Boris Fiuczynski wrote:
On 08/12/2016 04:59 PM, Ján Tomko wrote:
On Fri, Aug 12, 2016 at 10:47:13AM +0200, Cornelia Huck wrote:
On Thu, 11 Aug 2016 16:17:10 +0200
Ján Tomko wrote:
On Thu, Aug 11, 2016 at 02:31:55PM +0100, Daniel P. Berrange wrote:
>O
On Tue, Aug 16, 2016 at 11:01:23AM +0200, Michal Privoznik wrote:
> Signed-off-by: Michal Privoznik
> ---
>
> I figured that there's no need to update Changelog here as it
> already mentions nodedev events implementation.
>
> Virt.xs| 1 +
> lib/Sys/Virt/NodeDevice.pm | 4 ++
Signed-off-by: Michal Privoznik
---
I figured that there's no need to update Changelog here as it
already mentions nodedev events implementation.
Virt.xs| 1 +
lib/Sys/Virt/NodeDevice.pm | 4
2 files changed, 5 insertions(+)
diff --git a/Virt.xs b/Virt.xs
index 5025f2e
On Tue, Aug 16, 2016 at 10:11:18 +0200, Peter Krempa wrote:
> Error message in qemuDomainSetVcpusFlags uses %d for the two arguments
> but both are unsigned. The resulting error message may be weird when
> attempting to set a very large vcpu count.
> ---
> src/qemu/qemu_driver.c | 2 +-
> 1 file c
On Mon, 2016-08-15 at 18:14 +0200, Ján Tomko wrote:
> If any of the devices referenced a USB hub that does not exist,
> defining the domain would either fail with:
> error: An error occurred, but the cause is unknown
> (if only the last hub in the path is missing)
> or crash.
>
> Return a proper e
Error message in qemuDomainSetVcpusFlags uses %d for the two arguments
but both are unsigned. The resulting error message may be weird when
attempting to set a very large vcpu count.
---
src/qemu/qemu_driver.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/qemu/qemu_driver
qemu uses 'url' instead of 'uri'. They unfortunately look very similar.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1367260
---
src/util/virstoragefile.c | 2 +-
tests/virstoragetest.c| 6 +++---
2 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/src/util/virstoragefile.
On Mon, 2016-08-15 at 18:14 +0200, Ján Tomko wrote:
> The array needs to be freed too, not just its members.
>
> https://bugzilla.redhat.com/show_bug.cgi?id=1366097
> ---
> src/conf/domain_addr.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/src/conf/domain_addr.c b/src/conf/domain_ad
Because of change in caaa1bd357a99ad this macro is no under
#ifdef block. That means it needs to be re-intended correctly.
Signed-off-by: Michal Privoznik
---
Pushed under build-breaker rule.
src/libxl/libxl_driver.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/libxl
70 matches
Mail list logo