linux-kernel - Re: [PATCH v1] tools lib api cpu: Remove unused file

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]

Message-ID: <691c84d7-3b3b-4896-a607-09ba7a2263d7@linux.intel.com>
Date: Tue, 7 Jan 2025 14:12:00 -0500
From: "Liang, Kan" <kan.liang@...ux.intel.com>
To: Ian Rogers <irogers@...gle.com>
Cc: Arnaldo Carvalho de Melo <acme@...hat.com>,
 Jesper Juhl <jesperjuhl76@...il.com>, Sam James <sam@...too.org>,
 Namhyung Kim <namhyung@...nel.org>, linux-kernel@...r.kernel.org,
 Kan Liang <kan.liang@...el.com>
Subject: Re: [PATCH v1] tools lib api cpu: Remove unused file



On 2025-01-07 11:01 a.m., Ian Rogers wrote:
> On Tue, Jan 7, 2025 at 6:22 AM Liang, Kan <kan.liang@...ux.intel.com> wrote:
>>
>> On 2025-01-06 5:32 p.m., Ian Rogers wrote:
>>> On Thu, Dec 19, 2024 at 12:54 PM Ian Rogers <irogers@...gle.com> wrote:
>>>>
>>>> No use in tools could be found, remove to simplify the code base.
>>>>
>>
>> This was to support the patch which generate per-sample
>> freq/CPU%/CORE_BUSY%.
>> https://lore.kernel.org/lkml/1442413316-33518-1-git-send-email-kan.liang@intel.com/
>>
>> However, the patch set was not merged. I don't remember the exact reason.
>>
>> It looks like there is no one interested in it. There should be no
>> reason to keep it.
>>
>> Acked-by: Kan Liang <kan.liang@...ux.intel.com>
> 
> Thanks Kan! That series looks nice. There is an overlap with computing
> frequencies and metrics like cpu_operating_frequency:
> https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git/tree/tools/perf/pmu-events/arch/x86/sapphirerapids/spr-metrics.json?h=perf-tools-next#n45
> 
> I wonder two things:
> 1) perhaps we can add the --perf-freq flag but make the implementation
> use metrics rather than hard coded events. Outside of perf stat the
> json metrics aren't really plumbed up, so this would be a bunch of
> work.

At that time, the metrics was not well supported. Everything is
hardcoded. Yes, now, we should utilize the json metrics for it.

> 2) the metrics compute the TSC frequency from cpuid, and fall back to
> /proc/cpuinfo:
> https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git/tree/tools/perf/arch/x86/util/tsc.c?h=perf-tools-next
> Perhaps there should be another fallback to
> /sys/devices/system/cpu/cpu%d/cpufreq/cpuinfo_max_freq, as in this
> series.

As my understanding, the TSC frequency should be a non-turbo frequency.
The cpuinfo_max_freq should be a turbo frequency. It may not be used as
a fallback.

Thanks,
Kan
> 
> Thanks,
> Ian