Re: [OE-core] [PATCH v3 00/11] systemd based initrd and modular kernel support

[Mikko Rapeli <mikko.rapeli@linaro.org>]

Hi,

On Tue, Apr 08, 2025 at 01:26:21PM +0200, Mathieu Dubois-Briand wrote:
> On Mon Apr 7, 2025 at 1:13 PM CEST, Mikko Rapeli wrote:
> > Hi,
> >
> > On Mon, Apr 07, 2025 at 01:26:06PM +0300, Mikko Rapeli via lists.openembedded.org wrote:
> > Sent this out as "oeqa wic.py: clean image build dir before rebuild in test_permissions()"
> > separately. I think this test failure is unrelated to the changes but related to the
> > way tests are executed. I think tests behave different depending if sstate has the
> > image recipe output or not.
> >
> > Cheers,
> >
> > -Mikko
> 
> Ok, it seems this patch indeed fixes the oe-selftest on the autobuilder:
> I had some successful builds this morning on a separate branch. Not sure
> about what is really the link between the error we saw locally and the
> one on the autobuilder.
> 
> I will pick both this series and the said patch in my -next branch later
> today.

Thanks. I've also been running wic selftests and they seem to be fine
with this change. I had one additional failure, very similar
pseudo problem, with wic.Wic.test_wic_sector_size test and I was
testing fix/workaround:

--- a/meta/lib/oeqa/selftest/cases/wic.py
+++ b/meta/lib/oeqa/selftest/cases/wic.py
@@ -852,6 +852,7 @@ bootloader --ptable gpt""")
             config = 'WIC_SECTOR_SIZE = "4096"\n'\
                      'WICVARS:append = " WIC_SECTOR_SIZE"\n'
             self.append_config(config)
+            bitbake('-c clean core-image-minimal')
             bitbake('core-image-minimal')
 
             # Check WIC_SECTOR_SIZE apply to bitbake variable

but I can't reproduce the problem anymore. There is something happening
with sstate cache here. I think builds of core-image-minimal and possibly
world before running selftests hide these problems. I can't explain
the details either.

Cheers,

-Mikko