On Fri, Apr 21, 2023 at 10:25 AM Michal Privoznik
wrote:
> This is a leftover from v2.0.0-rc1~300. In v1.2.12-rc1~43 we've
> introduced a code that explicitly sets vm->def->id to -1 to force
> generation of inactive XML. But this was removed in the later
> commit, which forgot to remove the
On Fri, Apr 21, 2023 at 10:25 AM Michal Privoznik
wrote:
> There's a comment in testQemuHotplug() trying to explain why we
> need to unlock the monitor object. Well, while it might have been
> correct when being introduced, it's no long factually correct as
>
*no longer
> just any function
On Fri, Apr 21, 2023 at 10:25 AM Michal Privoznik
wrote:
> After previous cleanups a lot of functions from qemu_hotplug.c
> are called only within the file. Make them static and drop their
> declaration from the header file.
>
*declarations
>
> Signed-off-by: Michal Privoznik
> ---
>
On Fri, Apr 21, 2023 at 10:25 AM Michal Privoznik
wrote:
> There's no reason for qemuhotplugtest to reimplement which device
> update function to call (testQemuHotplugUpdate()) when
> qemuDomainUpdateDeviceLive() already does that. Thus, drop
> testQemuHotplugUpdate() and call
On Fri, Apr 21, 2023 at 10:25 AM Michal Privoznik
wrote:
> There's no reason for qemuhotplugtest to reimplement which device
> attach function to call (testQemuHotplugAttach()) when
> qemuDomainAttachDeviceLive() already does that. Thus, drop
> testQemuHotplugAttach() and call
On Fri, Apr 21, 2023 at 10:25 AM Michal Privoznik
wrote:
> The testQemuHotplugDetach() already does call
> qemuDomainDetachDeviceLive() but only for some device types. For
> the rest it reports an error (but only if running test
> verbosely). This makes no sense. Just call
>
On Fri, Apr 14, 2023 at 03:44:16PM +0200, Michal Privoznik wrote:
From: Zhenzhong Duan
virsh command domxml-to-native failed with below error but start
command succeed for same domain xml.
"internal error: invalid PCI passthrough type 'default'"
If a PCI backend is not set in the XML, the
On Fri, Apr 21, 2023 at 10:25 AM Michal Privoznik
wrote:
> There is no good reason for qemuDomainUpdateDeviceLive() to live
> in (ever growing) qemu_driver.c while we have qemu_hotplug.c
> which already contains the rest of hotplug code. Move the
> function to its new home.
>
> Signed-off-by:
On Thu, Apr 20, 2023 at 02:30:46PM +0200, Michal Prívozník wrote:
On 4/20/23 14:07, Martin Kletzander wrote:
On Thu, Mar 23, 2023 at 05:01:26PM +0100, Michal Privoznik wrote:
In our meson scripts, we use configure_file(copy:true) to copy
files from srcdir into builddir. However, as of
Thank you for your reply and review, I'd appreciate for you to do that.
And I'd also like to confirm that we have the following usages after the
modification.
for non-parallel migration, we can use
--compressed
default to use XBZRLE for migration
--compressed --comp-methods ...
On Fri, Apr 21, 2023 at 10:25 AM Michal Privoznik
wrote:
> There is no good reason for qemuDomainAttachDeviceLive() to live
> in (ever growing) qemu_driver.c while we have qemu_hotplug.c
> which already contains the rest of hotplug code. Move the
> function to its new home.
>
> Signed-off-by:
On Fri, Apr 21, 2023 at 10:25 AM Michal Privoznik
wrote:
> The qemuDomainUpdateDeviceLive() accepts virDomainPtr as one of
> its arguments, but use it only to get QEMU driver out of it.
> Well, the only caller already done that and thus can pass it
>
*does
> instead of virDomainPtr.
>
> This
On 4/20/23 5:34 PM, Michal Prívozník wrote:
On 4/20/23 17:10, Boris Fiuczynski wrote:
With this code in master I am no longer able to build rpms from the
tarball with rpmbuild. Am I the only one?
This is the pipeline that run after this very patch:
There's no reason for qemuhotplugtest to reimplement which device
update function to call (testQemuHotplugUpdate()) when
qemuDomainUpdateDeviceLive() already does that. Thus, drop
testQemuHotplugUpdate() and call qemuDomainUpdateDeviceLive()
directly.
BTW: this also shows why reimplementing
Just like we check the resulting domain XML after ATTACH and
DETACH, we should do the same after UPDATE action. This is as
simple as calling testQemuHotplugCheckResult() and providing
missing XMLs. For those test cases where no change is done, we
can just make the expected XML a symlink to the
This is a leftover from v2.0.0-rc1~300. In v1.2.12-rc1~43 we've
introduced a code that explicitly sets vm->def->id to -1 to force
generation of inactive XML. But this was removed in the later
commit, which forgot to remove the restoration of the original
dom ID.
Signed-off-by: Michal Privoznik
This brings us one step closer to the caller of
qemuDomainAttachDeviceLive()
(qemuDomainAttachDeviceLiveAndConfig()).
Signed-off-by: Michal Privoznik
---
tests/qemuhotplugtest.c | 8 +---
1 file changed, 1 insertion(+), 7 deletions(-)
diff --git a/tests/qemuhotplugtest.c
There is no good reason for qemuDomainUpdateDeviceLive() to live
in (ever growing) qemu_driver.c while we have qemu_hotplug.c
which already contains the rest of hotplug code. Move the
function to its new home.
Signed-off-by: Michal Privoznik
---
src/qemu/qemu_driver.c | 299
The testQemuHotplugDetach() already does call
qemuDomainDetachDeviceLive() but only for some device types. For
the rest it reports an error (but only if running test
verbosely). This makes no sense. Just call
qemuDomainDetachDeviceLive() directly and drop
testQemuHotplugDetach().
Signed-off-by:
There's a comment in testQemuHotplug() trying to explain why we
need to unlock the monitor object. Well, while it might have been
correct when being introduced, it's no long factually correct as
just any function (attach/detach/update) might talk to the
monitor and it expects the monitor to be
There's no reason for qemuhotplugtest to reimplement which device
attach function to call (testQemuHotplugAttach()) when
qemuDomainAttachDeviceLive() already does that. Thus, drop
testQemuHotplugAttach() and call qemuDomainAttachDeviceLive()
directly.
There's one small catch though,
After previous cleanups a lot of functions from qemu_hotplug.c
are called only within the file. Make them static and drop their
declaration from the header file.
Signed-off-by: Michal Privoznik
---
src/qemu/qemu_hotplug.c | 56 +
src/qemu/qemu_hotplug.h | 61
The qemuDomainUpdateDeviceLive() accepts virDomainPtr as one of
its arguments, but use it only to get QEMU driver out of it.
Well, the only caller already done that and thus can pass it
instead of virDomainPtr.
This also makes it look like the rest of device hot(un-)plug
functions:
While working on something else, I've taken a look at qemuhotplugtest
and realized we can do various cleanups and improvements.
The test still covers just live attach/detach/update, but after this
it's prepared better to test also config attach/detach/update.
Michal Prívozník (11):
qemu:
There is no good reason for qemuDomainAttachDeviceLive() to live
in (ever growing) qemu_driver.c while we have qemu_hotplug.c
which already contains the rest of hotplug code. Move the
function to its new home.
Signed-off-by: Michal Privoznik
---
src/qemu/qemu_driver.c | 170
25 matches
Mail list logo