[OE-core] [PATCH 3/4] qemu: switch to '-vga std' emulated hardware from vmware/cirrus for x86/mips

[Alexander Kanavin]

On Fri, 30 Aug 2019 at 03:55, Mittal, Anuj <anuj.mittal at intel.com> wrote:

> > -QB_OPT_APPEND = "-show-cursor -device VGA,edid=on"
> > +QB_OPT_APPEND = "-show-cursor"
>
> Was this intentional? This is reverting an earlier fix:
>
>
> https://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/meta/conf/machine/qemuarm64.conf?h=master-next&id=b181bf92e2bc51a808478c12fe62177d966f2a85
>
>
> https://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/meta/conf/machine/qemuarm64.conf?h=master-next&id=ff35b6c57a380ca4dc158c5104a6137f8faea5cd
>
> I'm not sure if this is the reason but master-next build is also
> showing problems when running X with qemuarm:
>
>
> https://autobuilder.yoctoproject.org/typhoon/#/builders/38/builds/978/steps/8/logs/step1c
>

Thanks, I wasn't aware that qemu's list of supported graphical hardware
(and what the default selection is) differs between architectures and
machines.
The documentation erroneously claims that '-vga std' is the default
everywhere where in reality it is the default only on x86.

Patch will be reworked to not touch arm bits.


...
>
> > diff --git a/meta/lib/oeqa/runtime/cases/parselogs.py
> > b/meta/lib/oeqa/runtime/cases/parselogs.py
> > index 19c9c52a0a9..0002730d46f 100644
> > --- a/meta/lib/oeqa/runtime/cases/parselogs.py
> > +++ b/meta/lib/oeqa/runtime/cases/parselogs.py
> > @@ -55,7 +55,11 @@ common_errors = [
> >      "Failed to read /var/lib/nfs/statd/state: Success",
> >      "error retry time-out =",
> >      "logind: cannot setup systemd-logind helper (-61), using legacy
> > fallback",
> > -    "Error changing net interface name 'eth0' to "
> > +    "Error changing net interface name 'eth0' to ",
> > +    "modeset(0): Failed to initialize the DRI2 extension",
> > +    "uvesafb: cannot reserve video memory at",
> > +    "uvesafb: probe of uvesafb.0 failed with error",
> > +    "glamor initialization failed"
> >      ]
> >
>
> parselogs test is used not only for qemu but for other MACHINES as well
> including genericx86 on actual hardware where glamor initialization
> error might need looking into. Perhaps this should be added to qemu
> specific parts of this test only?
>

 That's right as well. Let's add the messages to qemux86 parts, and see if
other qemu targets need this as well.

Alex
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openembedded.org/pipermail/openembedded-core/attachments/20190830/a314427c/attachment.html>