Re: [OE-core] [PATCH v9 00/11] systemd uki support

[Richard Purdie <richard.purdie@linuxfoundation.org>]

On Tue, 2024-10-22 at 22:34 +0300, Mikko Rapeli via lists.openembedded.org wrote:
> Hi,
> 
> On Tue, Oct 22, 2024 at 10:26:52AM +0000, Ross Burton wrote:
> > On 17 Oct 2024, at 09:45, Mikko Rapeli via lists.openembedded.org <mikko.rapeli=linaro.org@lists.openembedded.org> wrote:
> > > v9: Fixed unrelated ext4 rawcopy wic test on genericarm64. Enabled new aarch64 target
> > >    wic test on x86_64 build host only since boot with qemu is currently
> > >    failing on aarch64 build host.
> > 
> > Sorry, but I don’t think disabling tests based on the build machine is the right thing to do here.
> 
> Well I'm running out of options to try.
> 
> > My gut reaction is that there is another test that is running with kvm enabled, thus breaking uboot.  I’ll do a build on my own machine shortly and see if I can replicate.
> 
> IMO this is a generic problem. The selftest code is not fully in control of
> the bitbake build config, including runqemu. Snippets get enabled here and
> there. Difficult to know what the effective config in CI is. This also not visible
> in the AB build logs, or at least I could not find it.
> 
> Thanks for helping, I hope you can reproduce this.

Ross asked for a failing build log, there is one here:

https://valkyrie.yoctoproject.org/#/builders/23/builds/306/steps/14/logs/stdio

and the config for this is clearly visible here in the step before it:

https://valkyrie.yoctoproject.org/#/builders/23/builds/306/steps/13/logs/stdio

There isn't anything more being set or any other magic, that is the
config and it runs an oe-selftest command (the one at the top of the
first link above).

Cheers,

Richard