linux-kernel - Re: [PATCH v6 1/3] x86/fpu: Measure the Latency of XSAVES and XRSTORS

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: <ZQQp5LbMOHFJITkn@gmail.com>
Date:   Fri, 15 Sep 2023 11:54:44 +0200
From:   Ingo Molnar <mingo@...nel.org>
To:     Yi Sun <yi.sun@...el.com>
Cc:     dave.hansen@...el.com, tglx@...utronix.de,
        linux-kernel@...r.kernel.org, x86@...nel.org,
        sohil.mehta@...el.com, ak@...ux.intel.com,
        ilpo.jarvinen@...ux.intel.com, heng.su@...el.com,
        tony.luck@...el.com, dave.hansen@...ux.intel.com,
        yi.sun@...el.intel.com
Subject: Re: [PATCH v6 1/3] x86/fpu: Measure the Latency of XSAVES and XRSTORS


* Yi Sun <yi.sun@...el.com> wrote:

> > Instead of adding overhead to the regular FPU context saving/restoring 
> > code paths, could you add a helper function that has tracing code 
> > included, but which isn't otherwise used - and leave the regular code 
> > with no tracing overhead?

> Furthermore, according doc static-keys.txt, the condition 
> xrstor_tracing_enabled() would introduce only a minimal overhead when the 
> trace is disabled. I believe it is a negligible impact on the performance 
> when the trace is disabled.

Why introduce *any* extra overhead if it's possible to test the 
functionality separately? The stated goal of the series is only to measure 
FPU context switch performance, which doesn't require extra added overhead 
to the actual context switch path.

[ Or if you want to convince reviewers that the overhead is indeed minimal, 
  please provide before/after generated assembly of the affected code that 
  demonstrates minimal impact. ]

Thanks,

	Ingo