APPLIED[F]/NAK[B]/CMT: [SRU][B][F][PATCH 0/1] kernel: s390/cpum_cf, perf: change DFLT_CCERROR counter name (LP: 1891454)

[Kelsey Skunberg]

Hi Frank! This has been applied to focal master-next. 

For Bionic, the patch does not apply since the commit "extended counters
for IBM z15" is not currently in bionic master-next. I see that patch
was SRU'd for focal, though not bionic. May you please send a v2 for
just bionic which includes the needed patches? Thanks so much! 

-Kelsey

On 2020-08-25 15:53:33 , frank.heimes at canonical.com wrote:
> Buglink: https://bugs.launchpad.net/bugs/1891454
> 
> SRU Justification:
> 
> [Impact]
> 
> * Change the counter name DLFT_CCERROR to DLFT_CCFINISH on IBM z15 to avoid confusion.
> 
> * This counter counts completed DEFLATE instructions with exit code 0, 1 or 2.
> 
> * And since exit code 0 means success and exit code 1 or 2 indicate errors the name can be ambiguous.
> 
> [Fix]
> 
> * Rename s390 z15 counter named DFLT_CCERROR to DFLT_CCFINISH.
> 
> * Backport Focal: https://launchpadlibrarian.net/493272507/0001-s390-cpum_cf-perf-change-DFLT_CCERROR-counter-name.patch
> 
> * Backport Bionic: https://launchpadlibrarian.net/493439681/u18-0001-s390-cpum_cf-Add-new-extended-counters-for-IBM-z15.patch
> 
> [Test Case]
> 
> * Enable hardware counters in the activation profile of a z15 LPAR.
> 
> * Just check the countername in sysfs with: ls /sys/devices/cpum_cf/events/{DFLT_CCERROR,DFLT_CCFINISH}
> 
> [Regression Potential] 
> 
> * The regression potential can be considered as low, since:
> 
> * Counters like these are usually not activated by default and need to be explicitly enabled.
> 
> * No code functionality is changed with that patch, only a renaming is done at 3 places in one code file arch/s390/kernel/perf_cpum_cf_events.c as well as in /tools/perf/pmu-events/arch/s390/cf_z15/extended.json.
> 
> * The fix got already upstream accepted in 5.8-rc7, hence it got upstream reviewed,
> 
> * and the modifications are limited to s390x, are only relevant for the z15 / LinuxONE III hw generation that are relatively new.
> 
> * Issue with the renaming I can think of are that people don't know about the new name and may look for the old one and don't find it anymore,
> 
> * and with that also if the old counter name is used in scripts or applications, these may fails due to an unknown counter name.
> 
> [Other Info]
> 
> * The backports are based on commit 3d3af181d370069861a3be94608464e2ff3682e2 ("s390/cpum_cf,perf: change DFLT_CCERROR counter name") that is upstream accepted since v5.8-rc7.
> 
> * Hence this SRU is not requested for groovy, since it's already in groovys kernel 5.8 in proposed, but only requested for focal and bionic.
> 
> Thomas Richter (1):
>   s390/cpum_cf,perf: change DFLT_CCERROR counter name
> 
>  arch/s390/kernel/perf_cpum_cf_events.c               | 4 ++--
>  tools/perf/pmu-events/arch/s390/cf_z15/extended.json | 2 +-
>  2 files changed, 3 insertions(+), 3 deletions(-)
> 
> -- 
> 2.25.1
> 
> 
> -- 
> kernel-team mailing list
> kernel-team at lists.ubuntu.com
> https://lists.ubuntu.com/mailman/listinfo/kernel-team