In mu previous commits I've moved internals of
qemuDomainChrDefDropDefaultPath() into a separate function
(qemuDomainChrMatchDefaultPath()) but forgot to remove @buf and
@regexp variables which are now unused.
Signed-off-by: Michal Privoznik
---
Pushed under trivial rule.
On Wed, Aug 16, 2023 at 09:15:20AM +0200, Michal Privoznik wrote:
> This is a resend of:
>
> https://listman.redhat.com/archives/libvir-list/2023-July/240904.html
>
> Diff to previous version:
> - s/9.6.0/9.7.0/ because of missed release
>
> Michal Prívozník (2):
> qemu: Generate shorter
On a Thursday in 2023, Peter Krempa wrote:
There are no more tests depending on '/usr/bin/qemu-system-i386' thus we
don't have to carry the data any more.
Signed-off-by: Peter Krempa
---
tests/testutilsqemu.c | 7 ---
1 file changed, 7 deletions(-)
Reviewed-by: Ján Tomko
Jano
On a Thursday in 2023, Peter Krempa wrote:
Convert the rest of the files using 'qemu-system-i386' to
'qemu-system-x86_64'. The 'cpu*' tests are done separately to emphasisze
*emphasise or *emphasize
that there's no change in the output.
Signed-off-by: Peter Krempa
---
As advertised earlier, now that the _virDomainMemoryDef struct is
cleaned up, we can shorten some names as their placement within
unions define their use.
Signed-off-by: Michal Privoznik
---
src/conf/domain_conf.c | 62
src/conf/domain_conf.h
The qemuDomainChangeMemoryLiveValidateChange() function is called
when a live memory device change is requested (via
virDomainUpdateDeviceFlags()). Currently, the only model that is
allowed to change is VIRTIO_MEM (and the only value that's
allowed to change is requestedsize). The aim of the
*** BLURB HERE ***
Michal Prívozník (9):
virt-aa-helper: Rework setting virDomainMemoryDef labels
virt-aa-helper: Set label on VIRTIO_PMEM device too
qemu_hotplug: validate address on memory device change
qemu_hotplug: Don't validate inaccessible fields in
Conceptually, from host POV there's no difference between NVDIMM
and VIRTIO_PMEM. Both expose a file to the guest (which is used
as a permanent storage). Other secdriver treat NVDIMM and
VIRTIO_PMEM the same. Thus, modify virt-aa-helper so that is
appends virtio-pmem backing path into the domain
Currently, inside of virt-aa-helper code the domain definition is
parsed and then all def->mems are iterated over and for NVDIMM
models corresponding nvdimmPath is set label on. Conceptually,
this code works (except the label should be set for VIRTIO_PMEM
model too, but that is addressed in the
The _virDomainMemoryDef struct is getting a bit messy. It has
various members and only some of them are valid for given model.
Worse, some are re-used for different models. We tried to make
this more bearable by putting a comment next to each member
describing what models the member is valid for,
The _virDomainMemoryDef struct is getting a bit messy. It has
various members and only some of them are valid for given model.
Worse, some are re-used for different models. We tried to make
this more bearable by putting a comment next to each member
describing what models the member is valid for,
When guest acknowledges change in size of virtio-mem (portion
that's exposed to the guest), QEMU emits
MEMORY_DEVICE_SIZE_CHANGE event. We process it in
processMemoryDeviceSizeChange(). So far, QEMU emits the even only
for virtio-mem (as that's the only memory device model that
allows live changes
This is similar to one of my previous commits. Simply speaking,
users can specify address where a memory device is mapped to. And
as such, we should include it when looking up corresponding
device in domain definition (e.g. on device hot unplug).
Signed-off-by: Michal Privoznik
---
As of v7.9.0-rc1~296 users have ability to adjust what portion of
virtio-mem is exposed to the guest. Then, as of v9.4.0-rc2~5 they
have ability to set address where the memory is mapped. But due
to a missing check it was possible to feed
virDomainUpdateDeviceFlags() API with memory device XML
On a Thursday in 2023, Peter Krempa wrote:
Convert tests which use DO_TEST_NOCAPS in both tests and the
qemuxml2xml variant has a symlink back to the qemuxml2argv input file.
This is done to separate the conversion before a patch converts all
DO_TEST_NOCAPS variants in qemuxml2xmltest to use
On a Thursday in 2023, Peter Krempa wrote:
Replace the emulator and architecture to x86_64, for all non-cpu related
test cases.
Signed-off-by: Peter Krempa
---
tests/qemuxml2argvdata/440fx-ide-address-conflict.xml | 4 ++--
tests/qemuxml2argvdata/440fx-wrong-root.xml |
On a Thursday in 2023, Peter Krempa wrote:
Use real latest capabilities.
Signed-off-by: Peter Krempa
---
tests/qemuxml2argvdata/net-many-models.args | 41 --
.../net-many-models.x86_64-latest.args| 43 +++
tests/qemuxml2argvdata/net-many-models.xml|
On a Thursday in 2023, Peter Krempa wrote:
At this point we setup the master key with all VMs, so this specific
test case no longer makes sense.
Signed-off-by: Peter Krempa
---
tests/qemuxml2argvdata/master-key.args | 33 --
tests/qemuxml2argvdata/master-key.xml | 30
Historically we've used QEMU_CAPS_QUERY_HOTPLUGGABLE_CPUS as witness
that the topology must cover the maximum number ov vcpus. qemu started
to enforce this in qemu-2.5, thus we can now always do the check.
This change also requires aligning the topology in certain test files.
Signed-off-by:
Use real capabilities and remove the redundant test case.
Signed-off-by: Peter Krempa
---
.../default-video-type-x86_64-caps-test-0.err | 1 -
.../default-video-type-x86_64-caps-test-1.xml | 16 --
...fault-video-type-x86_64.x86_64-latest.args | 32 +++
...st-0.xml =>
Version-lock the test to qemu-5.0.0 as it's the latest qemu that
supports 'vxhs' and thus the test can't use 'latest'.
Signed-off-by: Peter Krempa
---
...vxhs.xml => disk-network-tlsx509-vxhs.x86_64-5.0.0.xml} | 7 +--
...network-vxhs.xml => disk-network-vxhs.x86_64-5.0.0.xml} | 7 +--
Use latest capabilities for all tests.
Signed-off-by: Peter Krempa
---
.../video-device-pciaddr-default.args | 37 --
...-device-pciaddr-default.x86_64-latest.args | 39 +++
...nvalid-multiple-devices.x86_64-latest.err} | 0
...ideo-multiple-primaries.x86_64-latest.err}
Convert tests which use DO_TEST_NOCAPS in both tests and the
qemuxml2xml variant has a symlink back to the qemuxml2argv input file.
This is done to separate the conversion before a patch converts all
DO_TEST_NOCAPS variants in qemuxml2xmltest to use real capabilities.
Signed-off-by: Peter Krempa
Use latest caps for this rather recent graphics protocol.
Signed-off-by: Peter Krempa
---
... graphics-dbus-address.x86_64-latest.args} | 12 ---
.../graphics-dbus-address.xml | 5 ++-
...=> graphics-dbus-audio.x86_64-latest.args} | 14
There are no more tests depending on '/usr/bin/qemu-system-i386' thus we
don't have to carry the data any more.
Signed-off-by: Peter Krempa
---
tests/testutilsqemu.c | 7 ---
1 file changed, 7 deletions(-)
diff --git a/tests/testutilsqemu.c b/tests/testutilsqemu.c
index
Signed-off-by: Peter Krempa
---
.../net-virtio-teaming-hostdev.xml| 5 -
...th.xml => net-bandwidth.x86_64-latest.xml} | 5 -
...2.xml => net-bandwidth2.x86_64-latest.xml} | 5 -
...xml => net-hostdev-vfio.x86_64-latest.xml} | 5 -
Signed-off-by: Peter Krempa
---
.../qemuxml2argvdata/video-qxl-resolution.xml | 5 ++-
...-device-pciaddr-default.x86_64-latest.xml} | 5 ++-
...ml => video-none-device.x86_64-latest.xml} | 5 ++-
xml => video-qxl-heads.x86_64-latest.xml} | 5 ++-
...ml =>
Convert all tests using fake capabilities to use DO_TEST_CAPS_LATEST.
Note that rename detection in git didn't work too well here and the
files may not correspond.
Signed-off-by: Peter Krempa
---
.../net-client.x86_64-latest.args | 38
...args =>
Upgrade the relevant test cases to use latest capabilities. Note that
the 'shmem' (ivshmem) device is no longer supported and will be dropped
later.
Signed-off-by: Peter Krempa
---
...> shmem-invalid-address.x86_64-latest.err} | 0
.../shmem-invalid-address.xml | 1 +
...r =>
qemu removed the support for the old 'ivshmem' device in 4.0 release.
Signed-off-by: Peter Krempa
---
src/qemu/qemu_capabilities.c | 3 +--
src/qemu/qemu_capabilities.h | 2 +-
2 files changed, 2 insertions(+), 3 deletions(-)
diff --git a/src/qemu/qemu_capabilities.c
The device was removed in qemu-4.0 and is superseded by 'ivshmem-plain'
and 'ivshmem-doorbell'.
Always report error when the old version is used and drop the irrelevant
tests.
Signed-off-by: Peter Krempa
---
src/qemu/qemu_command.c| 40 +---
src/qemu/qemu_hotplug.c
Use latest real capabilities.
Signed-off-by: Peter Krempa
---
...emory-no-numa-topology.x86_64-latest.args} | 16
.../fd-memory-no-numa-topology.xml| 5 ++-
...d-memory-numa-topology.x86_64-latest.args} | 15
.../fd-memory-numa-topology.xml | 5 ++-
Finish the conversion of cases which didn't need any special
capabilities to use real capabilities.
Signed-off-by: Peter Krempa
---
tests/qemuxml2argvdata/net-mtu.xml| 5 +-
tests/qemuxml2argvdata/pages-discard.xml | 5 +-
...table.xml => acpi-table.x86_64-latest.xml} | 5
VIR_WARN of a misconfiguration of a VM is not really going to be seen in
most cases as it's simply logged, and if the VM works users are not
likely to dig through logs.
Signed-off-by: Peter Krempa
---
src/qemu/qemu_validate.c | 8
1 file changed, 8 deletions(-)
diff --git
The output files from 'qemuxml2argvtest' may have the real capability
suffix e.g. 'pci-rom-disabled-invalid.x86_64-latest.xml' which would not
be detected as being invalid and thus causing a test failure.
Change the logic to find '-invalid.' so that we can properly use
'virschematest' with test
The 'disk-cdrom-empty-network-invalid' is a special case were the input
XML is invalid according to the schema, but after processing a valid XML
is produced.
This corner case doesn't play well with 'virschematest' which uses the
file suffix to determine whether the file is invalid.
Upcoming
Convert all tests using the 'DO_TEST_NOCAPS' "fake" capability
invocation to use DO_TEST_CAPS_LATEST and remove the DO_TEST_NOCAPS
macro to prevent further use.
Most of the output file changes are related to default USB controller
type and the CPU becoming defined in the XML.
Signed-off-by:
At this point we setup the master key with all VMs, so this specific
test case no longer makes sense.
Signed-off-by: Peter Krempa
---
tests/qemuxml2argvdata/master-key.args | 33 --
tests/qemuxml2argvdata/master-key.xml | 30 ---
Convert the rest of the files using 'qemu-system-i386' to
'qemu-system-x86_64'. The 'cpu*' tests are done separately to emphasisze
that there's no change in the output.
Signed-off-by: Peter Krempa
---
tests/qemuxml2argvdata/cpu-host-model-features.xml | 4 ++--
Use real latest capabilities.
Signed-off-by: Peter Krempa
---
tests/qemuxml2argvdata/net-many-models.args | 41 --
.../net-many-models.x86_64-latest.args| 43 +++
tests/qemuxml2argvdata/net-many-models.xml| 2 +-
tests/qemuxml2argvtest.c
The series prepares and then converts all tests in qemuxml2xmltest to
real capabilities. Along some tests are converted even in
qemuxml2argvtest as the xml2xmltest output files were symlinks into the
qemuxml2argvdata directory.
Peter Krempa (24):
qemuxml2argvtest: Drop 'master-key' test case
We are getting close to 9.7.0 release of libvirt. To aim for the
release on Friday 01 Sep I suggest entering the freeze on Monday 28
Aug and tagging RC2 on Wednesday 30 Aug.
I hope this works for everyone.
Jirka
On Wed, Aug 09, 2023 at 10:49:42AM +0200, Boris Fiuczynski wrote:
> Adding the value 'optional' to attribute 'enabled' on the element
> async-teardown to provide the optional use of the feature if it is
> provided by QEMU.
This is not something we would typically do. IIUC, this is already
exposed
43 matches
Mail list logo