| Message ID | 20250407110946.47038-1-mikko.rapeli@linaro.org |
|---|---|
| State | New |
| Headers | show |
| Series | oeqa wic.py: clean image build dir before rebuild in test_permissions() | expand |
On Mon 2025-04-07 @ 02:09:45 PM, Mikko Rapeli via lists.openembedded.org wrote: > The test builds images using wic and then with bitbake > and compares results. This fails at least on aarch64 > build machine and multiple target machines due to > pseudo errors. Fix by cleaning the build recipe > build directory before rebuild. > > Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org> > --- > meta/lib/oeqa/selftest/cases/wic.py | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/meta/lib/oeqa/selftest/cases/wic.py b/meta/lib/oeqa/selftest/cases/wic.py > index 4f5d43b5ee..55c655cd6d 100644 > --- a/meta/lib/oeqa/selftest/cases/wic.py > +++ b/meta/lib/oeqa/selftest/cases/wic.py > @@ -674,6 +674,7 @@ part /etc --source rootfs --fstype=ext4 --change-directory=etc > > config = 'IMAGE_FSTYPES += "wic"\nWKS_FILE = "%s"\n' % wks_file > self.append_config(config) > + bitbake('-c clean core-image-minimal') > bitbake('core-image-minimal') > tmpdir = os.path.join(get_bb_var('WORKDIR', 'core-image-minimal'),'build-wic') > I have seen this issue too, but it is intermittent for me; is this an intermittent issue with you as well? This feels a lot more like a hack rather than a solution.
Hi, On Mon, Apr 14, 2025 at 04:56:14PM -0400, Trevor Woerner wrote: > On Mon 2025-04-07 @ 02:09:45 PM, Mikko Rapeli via lists.openembedded.org wrote: > > The test builds images using wic and then with bitbake > > and compares results. This fails at least on aarch64 > > build machine and multiple target machines due to > > pseudo errors. Fix by cleaning the build recipe > > build directory before rebuild. > > > > Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org> > > --- > > meta/lib/oeqa/selftest/cases/wic.py | 1 + > > 1 file changed, 1 insertion(+) > > > > diff --git a/meta/lib/oeqa/selftest/cases/wic.py b/meta/lib/oeqa/selftest/cases/wic.py > > index 4f5d43b5ee..55c655cd6d 100644 > > --- a/meta/lib/oeqa/selftest/cases/wic.py > > +++ b/meta/lib/oeqa/selftest/cases/wic.py > > @@ -674,6 +674,7 @@ part /etc --source rootfs --fstype=ext4 --change-directory=etc > > > > config = 'IMAGE_FSTYPES += "wic"\nWKS_FILE = "%s"\n' % wks_file > > self.append_config(config) > > + bitbake('-c clean core-image-minimal') > > bitbake('core-image-minimal') > > tmpdir = os.path.join(get_bb_var('WORKDIR', 'core-image-minimal'),'build-wic') > > > > I have seen this issue too, but it is intermittent for me; is this an > intermittent issue with you as well? This feels a lot more like a hack rather > than a solution. This is somewhat reproducible. I don't fully understand what is happening but it has something to do with pseudo fakeroot database and creating files with wic directly vs wic via bitbake build. Hack, possibly, but it fixes the issue. I don't expect pseudo to work correctly if tools are run in and outside of bitbake build environment. Cheers, -Mikko
On Mon, 2025-04-07 at 14:09 +0300, Mikko Rapeli via lists.openembedded.org wrote: > The test builds images using wic and then with bitbake > and compares results. This fails at least on aarch64 > build machine and multiple target machines due to > pseudo errors. Fix by cleaning the build recipe > build directory before rebuild. > > Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org> > --- > meta/lib/oeqa/selftest/cases/wic.py | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/meta/lib/oeqa/selftest/cases/wic.py b/meta/lib/oeqa/selftest/cases/wic.py > index 4f5d43b5ee..55c655cd6d 100644 > --- a/meta/lib/oeqa/selftest/cases/wic.py > +++ b/meta/lib/oeqa/selftest/cases/wic.py > @@ -674,6 +674,7 @@ part /etc --source rootfs --fstype=ext4 --change-directory=etc > > config = 'IMAGE_FSTYPES += "wic"\nWKS_FILE = "%s"\n' % wks_file > self.append_config(config) > + bitbake('-c clean core-image-minimal') > bitbake('core-image-minimal') > tmpdir = os.path.join(get_bb_var('WORKDIR', 'core-image-minimal'),'build-wic') > > Just to follow up on this, I'm reluctant to merge this as it will "live forever" without people understanding why it is there. We really need to get to the bottom of the underlying issue. Cheers, Richard
diff --git a/meta/lib/oeqa/selftest/cases/wic.py b/meta/lib/oeqa/selftest/cases/wic.py index 4f5d43b5ee..55c655cd6d 100644 --- a/meta/lib/oeqa/selftest/cases/wic.py +++ b/meta/lib/oeqa/selftest/cases/wic.py @@ -674,6 +674,7 @@ part /etc --source rootfs --fstype=ext4 --change-directory=etc config = 'IMAGE_FSTYPES += "wic"\nWKS_FILE = "%s"\n' % wks_file self.append_config(config) + bitbake('-c clean core-image-minimal') bitbake('core-image-minimal') tmpdir = os.path.join(get_bb_var('WORKDIR', 'core-image-minimal'),'build-wic')
The test builds images using wic and then with bitbake and compares results. This fails at least on aarch64 build machine and multiple target machines due to pseudo errors. Fix by cleaning the build recipe build directory before rebuild. Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org> --- meta/lib/oeqa/selftest/cases/wic.py | 1 + 1 file changed, 1 insertion(+)