linux-kernel - Re: [PATCH V4 0/2] powercap: Introduce TPMI RAPL PMU support

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: <CAJZ5v0jpaa2tp_NF9N=ViSiz3xmY_abREPZQxK1Hk57Wv7ggHg@mail.gmail.com>
Date: Tue, 30 Apr 2024 21:13:28 +0200
From: "Rafael J. Wysocki" <rafael@...nel.org>
To: Zhang Rui <rui.zhang@...el.com>
Cc: rafael.j.wysocki@...el.com, linux-kernel@...r.kernel.org, 
	linux-pm@...r.kernel.org, srinivas.pandruvada@...el.com
Subject: Re: [PATCH V4 0/2] powercap: Introduce TPMI RAPL PMU support

On Sun, Apr 28, 2024 at 11:24 AM Zhang Rui <rui.zhang@...el.com> wrote:
>
> RAPL energy counter MSRs are exposed via perf PMU. But this is done by
> separate code which is not part of RAPL framework, and it cannot be
> reused by other RAPL Interface drivers like TPMI RAPL.
>
> Introduce two new APIs for PMU support in RAPL framework. This allows
> TPMI RAPL PMU support and also makes it possible for future cleanups of
> MSR RAPL PMU code.
>
> Changes since V3:
> - Fix is_rp_pmu_cpu() return value. (Rafael)
> - Following the conversion in other PMU drivers to use
>   local64_try_cmpxchg() instead of local64_cmpxchg(). E.g. bcc6ec3d954b
>   ("perf/x86/rapl: Use local64_try_cmpxchg in rapl_event_update()"), and
>   update with a comment.
>
> Changes since V2:
> - drop patch 1/3 and rebase on PM tree bleeding-edge branch.
> - fix a regression introduced in patch V2 that RAPL PMU is not
>   successfully unregisted when unbinding the device. This causes kernel
>   calltrace about duplicate PMU name in sysfs when reloading the module.
>   Our unit test cases reproduced this but the test framework failed to
>   catch it, and now that problem is also fixed.
> - add comment for RAPL PMU implementation details.
> - add comment/document for struct rapl_pmu and rapl_package_pmu_data.
> - remove unneeded global RAPL PMU cpumask.
> - use an array to map RAPL PMU events to RAPL Domains.
> - Other minor improvements.
>
> Changes since V1:
> - remove the MSR RAPL PMU conversion because it is a separate work that
>   can be done later.
> - instead of using a flag to indicate the need of PMU support, introduce
>   two APIs for the RAPL Interface driver to invoke explicitly.
> - minor code/comments/changelog improvements.
>
> thanks,
> rui
>
> ----------------------------------------------------------------
> Zhang Rui (2):
>       powercap: intel_rapl: Introduce APIs for PMU support
>       powercap: intel_rapl_tpmi: Enable PMU support
>
>  drivers/powercap/intel_rapl_common.c | 582 +++++++++++++++++++++++++++++++++++
>  drivers/powercap/intel_rapl_tpmi.c   |   3 +
>  include/linux/intel_rapl.h           |  32 ++
>  3 files changed, 617 insertions(+)

Both patches applied as 6.10 material, thanks!