linux-kernel - Re: [PATCH v3 0/3] Use update_current_exec

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]

Date:   Mon, 6 Nov 2023 13:35:39 +0100
From:   Peter Zijlstra <peterz@...radead.org>
To:     Yajun Deng <yajun.deng@...ux.dev>
Cc:     mingo@...hat.com, juri.lelli@...hat.com,
        vincent.guittot@...aro.org, dietmar.eggemann@....com,
        rostedt@...dmis.org, bsegall@...gle.com, mgorman@...e.de,
        bristot@...hat.com, vschneid@...hat.com,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 0/3] Use update_current_exec_runtime simplify code

On Tue, Oct 31, 2023 at 08:59:25PM +0800, Yajun Deng wrote:
> update_current_exec_runtime would update execution time for each task,
> we can use update_current_exec_runtime simplify code.
> 
> The 1st and 2nd patch update update_current_exec_runtime() applies to all
> callers.
> The 3rd patch use update_current_exec_runtime simplify update_curr.
> 
> Yajun Deng (3):
>   sched: Don't account execution time for task group
>   sched: Don't trace stat runtime for task group
>   sched/fair: Simplify update_curr()
> 
>  kernel/sched/deadline.c  |  4 +---
>  kernel/sched/fair.c      | 13 +++----------
>  kernel/sched/rt.c        |  5 ++---
>  kernel/sched/sched.h     | 11 +++++++----
>  kernel/sched/stop_task.c |  2 +-
>  5 files changed, 14 insertions(+), 21 deletions(-)

Hurmph, so I'm having conflicts against this:

  https://lkml.kernel.org/r/54d148a144f26d9559698c4dd82d8859038a7380.1699095159.git.bristot@kernel.org

(obviously).. I've resolved the first patch, which also mostly includes
the second patch.

However, your second patch isn't entirely right, it now unconditionally
traces ->vruntime, which isn't the same. Imagine a regular task getting
a PI boost to RT, in that case ->vruntime will be non-zero and the RT
task will now be logging a vruntime.

Anyway, that tracepoint doesn't really make sense to me anyway, that is,
it logs a delta_exec and an absolute vruntime, that's inconsistent.
Also, a delta vruntime can be easily computed because the weight should
be known.

I think I'm going to simply remove the vruntime from that tracepoint and
avoid the whole problem.

This then also makes resolving patch 3 easier.

Let me go squish all this and then I'll post a link to whatever came
out.