Re: [OE-core] Reproducible build issues with buildpaths - take 2

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

On Fri, 2022-07-08 at 09:37 -0400, Bruce Ashfield wrote:
> On Fri, Jul 8, 2022 at 7:07 AM Richard Purdie
> <richard.purdie@linuxfoundation.org> wrote:
> > 
> > Hi All,
> > 
> > We've made good progress and I appreciate the help from people. I ran a
> > fresh build on the autobuilder and I forced it to rerun the QA checks
> > for all recipes. The tricky bit about warnings is that they don't show
> > up for objects from sstate and the addition of the buildpaths QA check
> > doesn't cause the checks to rerun. Due to that a few more previsouly
> > unseen issues cropped up.
> > 
> > The list of issues now stands are:
> > 
> > qemuppc:
> > https://autobuilder.yoctoproject.org/typhoon/#/builders/63/builds/5428
> > WARNING: kernel-devsrc-1.0-r0 do_package_qa: QA Issue: File /lib/modules/5.15.48-yocto-standard/build/include/generated/.vdso32-offsets.h.cmd in package kernel-devsrc contains reference to TMPDIR [buildpaths]
> 
> I'll take this one. My fix was only for arch/arm, but I'll expand it
> to cover ppc32.

Thanks.

I built qemuppc locally and got a different failure, I'd really like to
know why I saw that and the autobuilder did not. The failure was:

WARNING: linux-yocto-5.15.48+gitAUTOINC+b35d1d1e83_5aa3abf07b-r0 do_package_qa: QA Issue: File /usr/src/debug/linux-yocto/5.15.48+gitAUTOINC+b35d1d1e83_5aa3abf07b-r0/linux-qemuppc-standard-build/drivers/tty/vt/consolemap_deftbl.c in package linux-yocto-src contains reference to TMPDIR
File /usr/src/debug/linux-yocto/5.15.48+gitAUTOINC+b35d1d1e83_5aa3abf07b-r0/linux-qemuppc-standard-build/lib/oid_registry_data.c in package linux-yocto-src contains reference to TMPDIR [buildpaths]

sure enough, tmp/work/qemuppc-poky-linux/linux-yocto/5.15.48+gitAUTOINC+b35d1d1e83_5aa3abf07b-r0$ head package/usr/src/debug/linux-yocto/5.15.48+gitAUTOINC+b35d1d1e83_5aa3abf07b-r0/linux-qemuppc-standard-build/drivers/tty/vt/consolemap_deftbl.c 
/*
 * Do not edit this file; it was automatically generated by
 *
 * conmakehash /media/build1/poky/build/tmp/work-shared/qemuppc/kernel-source/drivers/tty/vt/cp437.uni > [this file]
 *
 */


I think the warning is saying there is a second issue in linux-qemuppc-
standard-build/lib/oid_registry_data.c too.

Sorry Bruce, not sure why several roads lead to the kernel :/

Cheers,

Richard