On Mon, 25 Apr 2022 22:23:05 +0200
Eric Auger wrote:
> Hi Alex,
>
> On 4/23/22 12:09 AM, Alex Williamson wrote:
> > [Cc +libvirt folks]
> >
> > On Thu, 14 Apr 2022 03:46:52 -0700
> > Yi Liu wrote:
> >
> >> With the introduction of iommufd[1], the linux kernel provides a generic
> >>
On 25/04/2022 09:21, Martin Kletzander wrote:
I wanted to deal with https://bugzilla.redhat.com/2043498 and I got a
suggesstion that removing deprecated features could actually make it
easier to propagate the error. In the end (last patch) it turns out the
error is still just reported with
On 4/25/22 16:49, Peter Krempa wrote:
> On Mon, Apr 25, 2022 at 16:40:17 +0200, Michal Prívozník wrote:
>> On 4/25/22 16:36, Peter Krempa wrote:
>>> On Wed, Apr 20, 2022 at 13:02:12 +0200, Peter Krempa wrote:
Few minor changes in qemu since the last update:
- PIIX4_PM gained
On Mon, Apr 25, 2022 at 16:40:17 +0200, Michal Prívozník wrote:
> On 4/25/22 16:36, Peter Krempa wrote:
> > On Wed, Apr 20, 2022 at 13:02:12 +0200, Peter Krempa wrote:
> >> Few minor changes in qemu since the last update:
> >> - PIIX4_PM gained 'x-not-migrate-acpi-index' property
> >
> >
On 4/25/22 16:36, Peter Krempa wrote:
> On Wed, Apr 20, 2022 at 13:02:12 +0200, Peter Krempa wrote:
>> Few minor changes in qemu since the last update:
>> - PIIX4_PM gained 'x-not-migrate-acpi-index' property
>
> Based on the discussion I can drop this line.
>
>> - 'cocoa' display and
On Mon, 25 Apr 2022 11:10:14 +0100
Daniel P. Berrangé wrote:
> On Fri, Apr 22, 2022 at 04:09:43PM -0600, Alex Williamson wrote:
> > [Cc +libvirt folks]
> >
> > On Thu, 14 Apr 2022 03:46:52 -0700
> > Yi Liu wrote:
> >
> > > With the introduction of iommufd[1], the linux kernel provides a
On Wed, Apr 20, 2022 at 13:02:12 +0200, Peter Krempa wrote:
> Few minor changes in qemu since the last update:
> - PIIX4_PM gained 'x-not-migrate-acpi-index' property
Based on the discussion I can drop this line.
> - 'cocoa' display and corresponding props (not present in this build)
>
On 4/25/22 15:28, Peter Krempa wrote:
> Bugs in the code meant to invalidate the qemu capabilities cache
> resulted in us always re-probing capabilities on startup of
> libvirtd/virtqemud. This got extremely annoying with the modern
> deployments using --timeout.
>
> Apart from issues in
* Daniel P. Berrangé (berra...@redhat.com) wrote:
> On Mon, Apr 25, 2022 at 01:33:41PM +0100, Dr. David Alan Gilbert wrote:
> > * Daniel P. Berrangé (berra...@redhat.com) wrote:
> > > On Mon, Apr 25, 2022 at 12:04:37PM +0100, Dr. David Alan Gilbert wrote:
> > > > * Daniel P. Berrangé
On Mon, Apr 25, 2022 at 15:28:24 +0200, Peter Krempa wrote:
> Bugs in the code meant to invalidate the qemu capabilities cache
> resulted in us always re-probing capabilities on startup of
> libvirtd/virtqemud. This got extremely annoying with the modern
> deployments using --timeout.
>
> Apart
On Mon, Apr 25, 2022 at 10:21:59AM +0200, Martin Kletzander wrote:
> Signed-off-by: Martin Kletzander
> ---
> audio/audio.c | 13 -
> audio/audio.h | 1 -
> audio/audio_legacy.c| 555
> audio/meson.build
On Mon, Apr 25, 2022 at 10:22:00AM +0200, Martin Kletzander wrote:
> Since there is no fallback mechanism and default-guessing this is now
> not used and can be safely removed.
>
> Signed-off-by: Martin Kletzander
> ---
> audio/alsaaudio.c | 1 -
> audio/audio_int.h | 1 -
>
On Mon, Apr 25, 2022 at 10:21:58AM +0200, Martin Kletzander wrote:
> Now that audiodev= is required and audio_init() will not be called
> without and AudioDev we can remove the fallback functionality and error
> out in case audio drivers fail initialisation or when the driver does
> not exist.
>
On Mon, Apr 25, 2022 at 10:21:57AM +0200, Martin Kletzander wrote:
> Since all callers require a valid audiodev this function can now safely
> abort in case of missing AudioState.
>
> Signed-off-by: Martin Kletzander
> ---
> audio/audio.c | 6 ++
> 1 file changed, 2 insertions(+), 4
On Mon, Apr 25, 2022 at 10:21:47AM +0200, Martin Kletzander wrote:
> This will be used in future commit.
>
> Signed-off-by: Martin Kletzander
> ---
> hw/audio/lm4549.c | 3 ++-
> hw/audio/lm4549.h | 3 ++-
> hw/audio/pl041.c | 2 +-
> 3 files changed, 5 insertions(+), 3 deletions(-)
On Mon, Apr 25, 2022 at 02:35:03PM +0100, Daniel P. Berrangé wrote:
> On Mon, Apr 25, 2022 at 10:21:46AM +0200, Martin Kletzander wrote:
> > No return values are used anywhere, so switch the functions to be void
> > and add support for error reporting using errp for use in next patches.
>
>
On Mon, Apr 25, 2022 at 10:21:50AM +0200, Martin Kletzander wrote:
> Many machine types have default audio devices with no way to set the
> underlying
> audiodev. Instead of adding an option for each and every one of them this new
> property can be used as a default during machine initialisation
On Mon, Apr 25, 2022 at 10:21:49AM +0200, Martin Kletzander wrote:
> Signed-off-by: Martin Kletzander
> ---
> ui/vnc.c | 15 +--
> 1 file changed, 9 insertions(+), 6 deletions(-)
>
> diff --git a/ui/vnc.c b/ui/vnc.c
> index badf1d7664fe..2e7af139b030 100644
> --- a/ui/vnc.c
> +++
On Mon, Apr 25, 2022 at 10:21:48AM +0200, Martin Kletzander wrote:
> This will enable removing deprecated default audiodev support.
>
> I did not figure out how to make the audiodev represented as an
> interface node, so this is a workaround. I am not sure what would be
> the proper way.
Not
On Mon, Apr 25, 2022 at 02:34:08PM +0100, Daniel P. Berrangé wrote:
> On Mon, Apr 25, 2022 at 10:21:56AM +0200, Martin Kletzander wrote:
> > Now that all callers support error reporting with errp and all
> > machine-default
> > devices use an explicit audiodev, this can be changed. To make the
On Mon, Apr 25, 2022 at 10:21:46AM +0200, Martin Kletzander wrote:
> No return values are used anywhere, so switch the functions to be void
> and add support for error reporting using errp for use in next patches.
Making it 'void' makes sense, but I'm not sure we need the Error **errp
addition,
On Mon, Apr 25, 2022 at 10:21:56AM +0200, Martin Kletzander wrote:
> Now that all callers support error reporting with errp and all machine-default
> devices use an explicit audiodev, this can be changed. To make the detection
> easier make AUD_register_card() return false on error.
>
>
In case when the 'KVM_GET_SUPPORTED_CPUID' ioctl on /dev/kvm would
fail for other reason than the documented E2BIG, our code would continue
looping and calling it while always increasing the memory buffer even
when that will not help.
Rewrite the function to allow another iteration only with the
AMD cpus report Core (compute unit) identifiers of the cpu running the
instruction under leaf 0x801e. This data is not needed for libvirt
and actually breaks caching of the qemu capabilities where we check that
all of the CPU flags to be identical.
Mask out all of leaf 0x801e.
Move the filtering code into virHostCPUGetCPUIDFilterVolatile.
This also removes a safe but very questionable reuse of 'i' iterator in
the both the top level and nested loop. It's safe for now as the to
level loop will not iterate any more in the current state.
Signed-off-by: Peter Krempa
---
Raise the lower bound to '64' as that seems to currently be the first
size that can fit the CPU data for a modern cpu.
Lower the upper bound to an arbitrary 65536. So many cpu features ougth
to be enough for everyone.
Signed-off-by: Peter Krempa
---
src/util/virhostcpu.c | 2 +-
1 file
The function returns 'virCPUData' but doesn't do two important steps
which other code takes:
1) leaves with all-zero data is stripped from the XML output
2) the data is expected to be sorted in the array
Now the 'virHostCPUGetCPUID' helper returns both all 0 leaves and
doesn't order them as we
Without this it's impossible to debug scenarios when this function
returns a mismatch but the formatted data looks identical.
Signed-off-by: Peter Krempa
---
src/cpu/cpu_x86.c | 16
1 file changed, 12 insertions(+), 4 deletions(-)
diff --git a/src/cpu/cpu_x86.c
The commit adding the code fetching host CPU flags via the
KVM_GET_SUPPORTED_CPUID didn't describe at all why such an alghorithm is
needed.
Add a comment from the documentation outlining how the userspace
function is expected to allocate memory here.
Signed-off-by: Peter Krempa
---
Bugs in the code meant to invalidate the qemu capabilities cache
resulted in us always re-probing capabilities on startup of
libvirtd/virtqemud. This got extremely annoying with the modern
deployments using --timeout.
Apart from issues in inserting the detected features, it also turns out
that
On Mon, Apr 25, 2022 at 10:21:45AM +0200, Martin Kletzander wrote:
> This deduplicates several lines and will make future changes more
> concise.
>
> Signed-off-by: Martin Kletzander
> ---
> hw/input/tsc210x.c | 68 --
> 1 file changed, 24
On Mon, Apr 25, 2022 at 10:21:44AM +0200, Martin Kletzander wrote:
> One thing I am not sure about is whether to keep the aliases of ac97 and
> es1370 in the qdev_alias_table.
I'm not sure we have any prior art in this respect. I feel it is a bit
odd-ball that AC97 and ES1370 are uppercase in
On Mon, Apr 25, 2022 at 10:22:01AM +0200, Martin Kletzander wrote:
> The caller would already fail, but this way the message can better
> express the reason for the failure.
>
> Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2043498
>
> Signed-off-by: Martin Kletzander
> ---
>
On Mon, Apr 25, 2022 at 01:33:41PM +0100, Dr. David Alan Gilbert wrote:
> * Daniel P. Berrangé (berra...@redhat.com) wrote:
> > On Mon, Apr 25, 2022 at 12:04:37PM +0100, Dr. David Alan Gilbert wrote:
> > > * Daniel P. Berrangé (berra...@redhat.com) wrote:
> > > > I'm worried that we could be
* Daniel P. Berrangé (berra...@redhat.com) wrote:
> On Mon, Apr 25, 2022 at 12:04:37PM +0100, Dr. David Alan Gilbert wrote:
> > * Daniel P. Berrangé (berra...@redhat.com) wrote:
> > > I'm worried that we could be taking ourselves down a dead-end by
> > > trying to optimize on the libvirt side,
On Mon, Apr 25, 2022 at 12:04:37PM +0100, Dr. David Alan Gilbert wrote:
> * Daniel P. Berrangé (berra...@redhat.com) wrote:
> > I'm worried that we could be taking ourselves down a dead-end by
> > trying to optimize on the libvirt side, because we've got a
> > mismatch between the QMP APIs we're
* Daniel P. Berrangé (berra...@redhat.com) wrote:
> On Fri, Apr 22, 2022 at 01:40:20PM +0200, Claudio Fontana wrote:
> > On 4/22/22 10:19 AM, Daniel P. Berrangé wrote:
> > > On Thu, Apr 21, 2022 at 08:06:40PM +0200, Claudio Fontana wrote:
> > >> On 4/21/22 7:08 PM, Daniel P. Berrangé wrote:
> >
When constructing mock_libs array it is firstly initialized to a
static set of mocks followed by couple of WITH_* checks to append
driver specific mocks. These checks are then repeated when
filling some other variables (e.g. supplementary helpers,
libraries, tests, etc.). Dissolve the former in
This is an alternative version to:
https://listman.redhat.com/archives/libvir-list/2022-April/230092.html
Michal Prívozník (2):
test: Link qemuhotplugmock with test_utils_qemu and test_utils
tests: Deduplicate some WITH_* checks
tests/meson.build | 94
As of ad81aa8ad07 the qemuhotplugmock.c calls
testQemuPrepareHostBackendChardevOne() which is implemented in
testutilsqemu.c. However, the mock is not linked with
testutilsqemu static library which makes some tools (valgrind
particularly) unhappy because the resulting mock library has
unresolved
On Fri, Apr 22, 2022 at 04:09:43PM -0600, Alex Williamson wrote:
> [Cc +libvirt folks]
>
> On Thu, 14 Apr 2022 03:46:52 -0700
> Yi Liu wrote:
>
> > With the introduction of iommufd[1], the linux kernel provides a generic
> > interface for userspace drivers to propagate their DMA mappings to
On 4/14/22 18:03, Peter Krempa wrote:
> On Thu, Apr 14, 2022 at 16:44:25 +0200, Michal Privoznik wrote:
>> Commit v8.0.0-409-gad81aa8ad0 added another function into
>> qemuhotplugmock.c. However, it did so in a bit clumsy way: the
>> function calls testQemuPrepareHostBackendChardevOne() which is
On 4/25/22 10:37, Peter Krempa wrote:
> Peter Krempa (2):
> virsh: cmdBlockcopy: Add '--print-xml' flag
> virsh: cmdBlockcopy: Fix generator of block copy disk XML
>
> docs/manpages/virsh.rst | 4 +++-
> tools/virsh-domain.c| 24 +---
> 2 files changed, 20
When a SIGHUP is received a thread is spawned that runs
virStateReload(). However, if SIGINT is received while the former
thread is still running then we may get into problematic
situation: the cleanup code in main() sees drivers initialized
and thus calls virStateCleanup(). So now we have two
In a recent commit I've attempted to rewrite the XML generator to use
virXMLFormatElement instead of manual steps. Unfortunately the commit
had multiple problems resulting in a garbled XML:
1) in certain cases the wrong buffer was used resulting in misplaced
snippets
2) the child element
Useful for knowing how to construct the XML and debugging.
Signed-off-by: Peter Krempa
---
docs/manpages/virsh.rst | 4 +++-
tools/virsh-domain.c| 13 -
2 files changed, 15 insertions(+), 2 deletions(-)
diff --git a/docs/manpages/virsh.rst b/docs/manpages/virsh.rst
index
Peter Krempa (2):
virsh: cmdBlockcopy: Add '--print-xml' flag
virsh: cmdBlockcopy: Fix generator of block copy disk XML
docs/manpages/virsh.rst | 4 +++-
tools/virsh-domain.c| 24 +---
2 files changed, 20 insertions(+), 8 deletions(-)
--
2.35.1
Since there is no fallback mechanism and default-guessing this is now
not used and can be safely removed.
Signed-off-by: Martin Kletzander
---
audio/alsaaudio.c | 1 -
audio/audio_int.h | 1 -
audio/coreaudio.m | 1 -
audio/dbusaudio.c | 1 -
audio/dsoundaudio.c | 1 -
audio/jackaudio.c
Since all callers require a valid audiodev this function can now safely
abort in case of missing AudioState.
Signed-off-by: Martin Kletzander
---
audio/audio.c | 6 ++
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/audio/audio.c b/audio/audio.c
index
Now that audiodev= is required and audio_init() will not be called
without and AudioDev we can remove the fallback functionality and error
out in case audio drivers fail initialisation or when the driver does
not exist.
Signed-off-by: Martin Kletzander
---
audio/audio.c | 146
There was no way to set this and we need that for it to be able to properly
initialise.
Signed-off-by: Martin Kletzander
---
hw/display/xlnx_dp.c | 6 ++
1 file changed, 6 insertions(+)
diff --git a/hw/display/xlnx_dp.c b/hw/display/xlnx_dp.c
index 9bb781e31254..b16d6be2b5cc 100644
---
Now that all callers support error reporting with errp and all machine-default
devices use an explicit audiodev, this can be changed. To make the detection
easier make AUD_register_card() return false on error.
Signed-off-by: Martin Kletzander
---
audio/audio.c| 7 +--
The caller would already fail, but this way the message can better
express the reason for the failure.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2043498
Signed-off-by: Martin Kletzander
---
audio/spiceaudio.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git
This will enable removing deprecated default audiodev support.
I did not figure out how to make the audiodev represented as an
interface node, so this is a workaround. I am not sure what would be
the proper way.
Signed-off-by: Martin Kletzander
---
tests/qtest/ac97-test.c | 3
Signed-off-by: Martin Kletzander
---
hw/ppc/prep.c | 4
1 file changed, 4 insertions(+)
diff --git a/hw/ppc/prep.c b/hw/ppc/prep.c
index bf622aa38fab..52d801afb307 100644
--- a/hw/ppc/prep.c
+++ b/hw/ppc/prep.c
@@ -46,6 +46,7 @@
#include "elf.h"
#include "qemu/units.h"
#include
Many machine types have default audio devices with no way to set the underlying
audiodev. Instead of adding an option for each and every one of them this new
property can be used as a default during machine initialisation when creating
such devices.
Signed-off-by: Martin Kletzander
---
Signed-off-by: Martin Kletzander
---
audio/audio.c | 13 -
audio/audio.h | 1 -
audio/audio_legacy.c| 555
audio/meson.build | 1 -
docs/about/deprecated.rst | 7 -
Signed-off-by: Martin Kletzander
---
hw/arm/integratorcp.c | 8 +++-
hw/arm/musicpal.c | 8 +++-
hw/arm/omap2.c| 8
hw/arm/realview.c | 3 +++
hw/arm/spitz.c| 10 +++---
hw/arm/versatilepb.c | 3 +++
hw/arm/vexpress.c | 3 +++
This deduplicates several lines and will make future changes more
concise.
Signed-off-by: Martin Kletzander
---
hw/input/tsc210x.c | 68 --
1 file changed, 24 insertions(+), 44 deletions(-)
diff --git a/hw/input/tsc210x.c b/hw/input/tsc210x.c
index
No return values are used anywhere, so switch the functions to be void
and add support for error reporting using errp for use in next patches.
Signed-off-by: Martin Kletzander
---
hw/audio/hda-codec.c | 32 ++--
hw/audio/intel-hda.c | 4 +---
hw/audio/intel-hda.h |
In case of sound devices which are created by default (some of them even with
-nodefaults) it would not be nice to users if qemu required explicit:
-audiodev driver=none,id=audio0 -machine default-audiodev=audio0
when they just want to run a VM and not care about any audio output. Instead
Signed-off-by: Martin Kletzander
---
hw/input/tsc210x.c | 8
1 file changed, 8 insertions(+)
diff --git a/hw/input/tsc210x.c b/hw/input/tsc210x.c
index f16a8090b7c7..f0b02bc72280 100644
--- a/hw/input/tsc210x.c
+++ b/hw/input/tsc210x.c
@@ -1098,6 +1098,14 @@ static void
Signed-off-by: Martin Kletzander
---
ui/vnc.c | 15 +--
1 file changed, 9 insertions(+), 6 deletions(-)
diff --git a/ui/vnc.c b/ui/vnc.c
index badf1d7664fe..2e7af139b030 100644
--- a/ui/vnc.c
+++ b/ui/vnc.c
@@ -4188,12 +4188,15 @@ void vnc_display_open(const char *id, Error **errp)
I wanted to deal with https://bugzilla.redhat.com/2043498 and I got a
suggesstion that removing deprecated features could actually make it
easier to propagate the error. In the end (last patch) it turns out the
error is still just reported with error_fatal, so it probably is not
really needed,
This will be used in future commit.
Signed-off-by: Martin Kletzander
---
hw/audio/lm4549.c | 3 ++-
hw/audio/lm4549.h | 3 ++-
hw/audio/pl041.c | 2 +-
3 files changed, 5 insertions(+), 3 deletions(-)
diff --git a/hw/audio/lm4549.c b/hw/audio/lm4549.c
index 32b1481b5614..418041bc9c6c 100644
One thing I am not sure about is whether to keep the aliases of ac97 and
es1370 in the qdev_alias_table.
Signed-off-by: Martin Kletzander
---
docs/about/deprecated.rst | 9 -
docs/about/removed-features.rst | 10 +
docs/qdev-device-use.txt
66 matches
Mail list logo
