[OpenAFS] (no subject)
Cheyenne Wills
cwills@sinenomine.net
Thu, 27 Jun 2024 10:13:30 -0600
It appears that the latest CentOS-9 stream has pulled in changes from
the Linux 6.8 kernel, specifically commit 'dentry: switch the lists of
children to hlist' (da549bdd15).
I will double check to see if the latest RHEL9 kernel is pulling this
in (Centos9-Stream is the beta version that eventually feeds into RHEL9)
This specific problem was addressed in the upcoming 1.8.12 by gerrit
15704 "Linux 6.8: use hlist iteration for dentry children". 1.8.12
will have support for linux kernels up to and including 6.9.
I find it odd that 1.8.10 was able to build when 1.8.11 failed. Was
the build using the same updated kernel level?
--
Cheyenne Wills
cwills@sinenomine.net
On Thu, 27 Jun 2024 12:33:51 +0200 (CEST)
Stephan Wonczak <a0033@rrz.uni-koeln.de> wrote:
> Hi everyone,
> nb: For some reason my original report did not make it to
> the list. Maybe because of my attachment; are these forbidden?
> A few days ago I hit a regression while building 1.8.11 for
> CentOS-9. As an rpm was missing from the release, I created a
> source-RPM from the .bz2-tarballs as per documentation and did did a
> rpmbuild. Command line used (yes, we still have need for ka...):
>
> rpmbuild --rebuild --with kauth openafs-1.8.11-1.src.rpm
>
> This worked fine on RHEL-8 (fully updated)
> That very same src.rpm has a problem building the modules on a
> fully updated CentOS-9 Stream (kernel 5.14.0-457).
> Last relevant lines from the build are:
>
> (...)
> ./include/linux/list.h:562:9: note: in expansion of macro 'list_entry'
> 562 | list_entry((pos)->member.next, typeof(*(pos)),
> member) | ^~~~~~~~~~
> ./include/linux/list.h:689:20: note: in expansion of macro
> 'list_next_entry'
> 689 | pos = list_next_entry(pos, member))
> | ^~~~~~~~~~~~~~~
> /root/rpmbuild/BUILD/openafs-1.8.11/src/libafs/MODLOAD-5.14.0-457.el9.x86_64-SP/osi_vcache.c:315:13:
> note: in expansion of macro 'list_for_each_entry'
> 315 | list_for_each_entry(child, &dp->d_subdirs,
> d_child) { | ^~~~~~~~~~~~~~~~~~~
> make[4]: *** [scripts/Makefile.build:268:
> /root/rpmbuild/BUILD/openafs-1.8.11/src/libafs/MODLOAD-5.14.0-457.el9.x86_64-SP/osi_vcache.o]
> Error 1
> make[4]: *** Waiting for unfinished jobs....
> make[3]: *** [Makefile:1942:
> /root/rpmbuild/BUILD/openafs-1.8.11/src/libafs/MODLOAD-5.14.0-457.el9.x86_64-SP]
> Error 2
> make[3]: Leaving directory '/usr/src/kernels/5.14.0-457.el9.x86_64'
> FAILURE: make exit code 2
> make[2]: *** [Makefile.afs:283: openafs.ko] Error 1
> make[2]: Leaving directory
> '/root/rpmbuild/BUILD/openafs-1.8.11/src/libafs/MODLOAD-5.14.0-457.el9.x86_64-SP'
> make[1]: *** [Makefile:188: linux_compdirs] Error 2
> make[1]: Leaving directory
> '/root/rpmbuild/BUILD/openafs-1.8.11/src/libafs'
> make: *** [Makefile:463: libafs] Error 2
> error: Bad exit status from /var/tmp/rpm-tmp.ZUYDKT (%build)
>
> Full build log available on request.
> A few more data points:
>
> - Building with "rpmbuild --rebuild --with kauth --define
> "build_modules 0" works fine and creates RPMs.
> - Builing 1.8.10 on that machine works fine (checked today)
> - Building 1.8.12pre1 on that machine works fine, too.
>
> So only the released version of 1.8.11 is broken on CentOS-9
> stream, and I suspect on RHEL-9, too.
>
> Dipl. Chem. Dr. Stephan Wonczak
>
> Regionales Rechenzentrum der Universitaet zu Koeln (RRZK)
> Universitaet zu Koeln, Weyertal 121, 50931 Koeln
> Tel: +49/(0)221/470-89583, Fax: +49/(0)221/470-89625
> _______________________________________________
> OpenAFS-info mailing list
> OpenAFS-info@openafs.org
> https://lists.openafs.org/mailman/listinfo/openafs-info