On 10/7/19 11:35 PM, jcfaracco(a)gmail.com wrote:
From: Julio Faracco <jcfaracco(a)gmail.com>
Now, QEMU command line can define 'xres' and 'yres' if XML contains both
properties from video model. This is generetaed if resolution was set.
Code let QEMU handle if video model supports this feature.
Signed-off-by: Julio Faracco <jcfaracco(a)gmail.com>
---
src/qemu/qemu_command.c | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c
index cbf25d5f07..99b43243e3 100644
--- a/src/qemu/qemu_command.c
+++ b/src/qemu/qemu_command.c
@@ -4596,6 +4596,11 @@ qemuBuildDeviceVideoStr(const virDomainDef *def,
virBufferAsprintf(&buf, ",vgamem=%uk", video->vram);
}
+ if (video->res && video->res->x && video->res->y) {
+ /* QEMU accepts resolution xres and yres. */
+ virBufferAsprintf(&buf, ",xres=%u,yres=%u", video->res->x,
video->res->y);
+ }
+
if (qemuBuildDeviceAddressStr(&buf, def, &video->info, qemuCaps) <
0)
return NULL;
If I apply just patch #1 and #2, the code fails to compile. This patch
depends on patch #3. This may break 'git bisect' if people are trying to
locate a regression so it should be avoided. Please ensure every
individual patch in a series is compilable and doesn't break the test
suite, or have syntax-check errors. You can write a git rebase one liner
to do roughly that.
In v2 review I also pointed to this example commit:
https://www.redhat.com/archives/libvir-list/2019-May/msg00820.html
I don't think all these changes need to be lumped together, but adding
docs/ and src/conf and a qemu test case in one commit is nice and self
containers. Then patch #2 can add the qemu_command.c bits and regenerate
the test output, to demonstrate what's actually changing. Generally
though I think it's a good idea to put the conf and docs/ changes in the
same commit, unless there's a large amount of doc additions
- Cole