linux-kernel - Re: [PATCH 7/8] perf tools: Add fallback for exclude

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] [thread-next>] [day] [month] [year] [list]
Message-ID: <9cdc97eb-a194-454a-916c-2acf64f0c726@linaro.org>
Date: Wed, 4 Sep 2024 14:29:17 +0100
From: James Clark <james.clark@...aro.org>
To: Namhyung Kim <namhyung@...nel.org>,
 Arnaldo Carvalho de Melo <acme@...nel.org>
Cc: Jiri Olsa <jolsa@...nel.org>, Adrian Hunter <adrian.hunter@...el.com>,
 Peter Zijlstra <peterz@...radead.org>, Ingo Molnar <mingo@...nel.org>,
 LKML <linux-kernel@...r.kernel.org>, linux-perf-users@...r.kernel.org,
 Ravi Bangoria <ravi.bangoria@....com>, Mark Rutland <mark.rutland@....com>,
 James Clark <james.clark@....com>,
 Athira Rajeev <atrajeev@...ux.vnet.ibm.com>, Kajol Jain
 <kjain@...ux.ibm.com>, Thomas Richter <tmricht@...ux.ibm.com>,
 Atish Patra <atishp@...shpatra.org>, Palmer Dabbelt <palmer@...osinc.com>,
 Mingwei Zhang <mizhang@...gle.com>, Ian Rogers <irogers@...gle.com>,
 Kan Liang <kan.liang@...ux.intel.com>
Subject: Re: [PATCH 7/8] perf tools: Add fallback for exclude_guest



On 04/09/2024 2:28 pm, James Clark wrote:
> 
> 
> On 04/09/2024 7:41 am, Namhyung Kim wrote:
>> It seems Apple M1 PMU requires exclude_guest set and returns EOPNOTSUPP
>> if not.  Let's add a fallback so that it can work with default events.
>>
>> Cc: Mark Rutland <mark.rutland@....com>
>> Cc: James Clark <james.clark@...aro.org>
>> Signed-off-by: Namhyung Kim <namhyung@...nel.org>
>> ---
>>   tools/perf/util/evsel.c | 21 +++++++++++++++++++++
>>   1 file changed, 21 insertions(+)
>>
>> diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
>> index 0de0a72947db3f10..8c4d70f7b2f5b880 100644
>> --- a/tools/perf/util/evsel.c
>> +++ b/tools/perf/util/evsel.c
>> @@ -3400,6 +3400,27 @@ bool evsel__fallback(struct evsel *evsel, 
>> struct target *target, int err,
>>                 "to fall back to excluding hypervisor samples", 
>> paranoid);
>>           evsel->core.attr.exclude_hv = 1;
>> +        return true;
>> +    } else if (err == EOPNOTSUPP && !evsel->core.attr.exclude_guest &&
>> +           !evsel->exclude_GH) {
>> +        const char *name = evsel__name(evsel);
>> +        char *new_name;
>> +        const char *sep = ":";
>> +
>> +        /* Is there already the separator in the name. */
>> +        if (strchr(name, '/') ||
>> +            (strchr(name, ':') && !evsel->is_libpfm_event))
>> +            sep = "";
>> +
>> +        if (asprintf(&new_name, "%s%su", name, sep) < 0)
>> +            return false;
>> +
>> +        free(evsel->name);
>> +        evsel->name = new_name;
>> +        /* Apple M1 requires exclude_guest */
>> +        scnprintf(msg, msgsize, "trying to fall back to excluding 
>> guest samples");
>> +        evsel->core.attr.exclude_guest = 1;
>> +
>>           return true;
>>       }
> 
> Not sure if this is working, for some reason it doesn't try the 
> fallback. With exclude guest made mandatory in the Arm PMU, then:
> 
>   $ perf stat -e cycles -vvv -- true
> 
>    Control descriptor is not initialized
>    Opening: cycles
>    ------------------------------------------------------------
>    perf_event_attr:
>      type                             0 (PERF_TYPE_HARDWARE)
>      size                             136
>      config                           0xb00000000
>      sample_type                      IDENTIFIER
>      read_format                      TOTAL_TIME_ENABLED|TOTAL_TIME_RUNNING
>      disabled                         1
>      inherit                          1
>      enable_on_exec                   1
>    ------------------------------------------------------------
>    sys_perf_event_open: pid 698  cpu -1  group_fd -1  flags 0x8
>    sys_perf_event_open failed, error -95
>    Warning:
>    cycles event is not supported by the kernel.
>    Opening: cycles
>    ------------------------------------------------------------
>    perf_event_attr:
>      type                             0 (PERF_TYPE_HARDWARE)
>      size                             136
>      config                           0xa00000000
>      sample_type                      IDENTIFIER
>      read_format                      TOTAL_TIME_ENABLED|TOTAL_TIME_RUNNING
>      disabled                         1
>      inherit                          1
>      enable_on_exec                   1
>    ------------------------------------------------------------
>    sys_perf_event_open: pid 698  cpu -1  group_fd -1  flags 0x8
>    sys_perf_event_open failed, error -95
>    Warning:
>    cycles event is not supported by the kernel.
>    failed to read counter cycles
>    failed to read counter cycles
> 
>     Performance counter stats for 'true':
> 
>       <not supported>      armv8_cortex_a53/cycles/
>       <not supported>      armv8_cortex_a57/cycles/
> 
> 
> 
> Other than that, all the tests are passing on Juno (without the 
> exclude_guest requirement).

Sorry one other thing, I think this commit would also need to come 
before the change to the exclude_guest default to keep bisect working.