linux-kernel - Re: [perf] 64a7f64478: perf-event-tests.check_reset

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Thu, 26 Mar 2020 23:13:41 +0000
From:   Song Liu <songliubraving@...com>
To:     kernel test robot <rong.a.chen@...el.com>,
        Peter Zijlstra <peterz@...radead.org>
CC:     linux-kernel <linux-kernel@...r.kernel.org>,
        Kernel Team <Kernel-team@...com>,
        Arnaldo Carvalho de Melo <arnaldo.melo@...il.com>,
        "Jiri Olsa" <jolsa@...nel.org>,
        Arnaldo Carvalho de Melo <acme@...hat.com>,
        "Alexey Budankov" <alexey.budankov@...ux.intel.com>,
        Namhyung Kim <namhyung@...nel.org>, Tejun Heo <tj@...nel.org>,
        LKP <lkp@...ts.01.org>
Subject: Re: [perf] 64a7f64478: perf-event-tests.check_reset_mpx.failed

Thanks for the report. 

> On Mar 18, 2020, at 5:49 AM, kernel test robot <rong.a.chen@...el.com> wrote:

[...]

> 
> 
> If you fix the issue, kindly add following tag
> Reported-by: kernel test robot <rong.a.chen@...el.com>
> 
> 
> * Checking for corner-cases in the ABI (not necessarily bugs)
>  + tests/corner_cases/multiple_active
>    Testing multiple simultaneous measurements...              PASSED
>  + tests/corner_cases/max_num_events
>    Testing max events that can be created...                  PASSED
>  + tests/corner_cases/max_multiplex
>    Testing limit of multiplexing...                           PASSED
>  + tests/corner_cases/reset_leader
>    Testing reset on group leader...                           PASSED
>  + tests/corner_cases/check_reset_mpx
>    Testing if reset clears multiplex fields...                FAILED

This is an easy fix. 

>  + tests/corner_cases/huge_events_start
>    Testing start of max events...                             PASSED
>  + tests/corner_cases/huge_group_start
>    Testing start of max event group...                        PASSED
>  + tests/corner_cases/signal_after_close
>    Testing signal after close...                              PASSED
> 
> * Checking for fast RDPMC support
>  + tests/rdpmc/rdpmc_support
>    Testing if userspace rdpmc reads are supported...          PASSED
>  + tests/rdpmc/rdpmc_validation
>    Testing if userspace rdpmc reads give expected results...  PASSED
>  + tests/rdpmc/rdpmc_multiplexing
>    Testing if userspace rdpmc multiplexing works...           PASSED
>  + tests/rdpmc/rdpmc_reset
>    Testing if resetting while using rdpmc works...            FAILED
>  + tests/rdpmc/rdpmc_group
>    Testing if rdpmc works with event groups...                FAILED
>  + tests/rdpmc/rdpmc_attach
>    Testing if rdpmc attach works...                           PASSED
>  + tests/rdpmc/rdpmc_attach_cpu
>    Testing if rdpmc behavior on attach CPU...                 FAILED
>  + tests/rdpmc/rdpmc_attach_global_cpu
>    Running on CPU 4
> Testing if rdpmc behavior on attach all procs on other CPU... FAILED
>  + tests/rdpmc/rdpmc_attach_other_cpu
>    Testing if rdpmc behavior on attach other CPU...           FAILED
>  + tests/rdpmc/rdpmc_multiattach
>    Testing if rdpmc multi-attach works...                     PASSED
>  + tests/rdpmc/rdpmc_multiattach_papi
>    Testing if rdpmc papi-multi-attach works...                PASSED
>  + tests/rdpmc/rdpmc_pthreads
>    Testing if rdpmc with pthreads works...                    PASSED
>  + tests/rdpmc/rdpmc_pthreads_group
>    Testing if rdpmc with pthreads works...                    PASSED
>  + tests/rdpmc/rdpmc_attach_multi_enable
>    Testing if minimized rdpmc papi-multi-attach works...      PASSED
>  + tests/rdpmc/rdpmc_exec
>    Testing if we can rdpmc in execed process...               PASSED
>  + tests/rdpmc/rdpmc_exec_papi
>    Testing if we can rdpmc in execed process (PAPI) ...       PASSED

The RDPMC break is trickier. 

IIUC, there is a critical call of perf_event_update_userpage() from 
x86_pmu_enable(), which is not easy to fix.

Peter, do you have suggestions on how to fix this? 

Thanks,
Song