linux-kernel - Re: [patch V2 28/37] rseq: Switch to fast path processing on exit to user

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: <877bydykef.ffs@tglx>
Date: Thu, 04 Sep 2025 23:31:36 +0200
From: Thomas Gleixner <tglx@...utronix.de>
To: Mathieu Desnoyers <mathieu.desnoyers@...icios.com>, LKML
 <linux-kernel@...r.kernel.org>
Cc: Jens Axboe <axboe@...nel.dk>, Peter Zijlstra <peterz@...radead.org>,
 "Paul E. McKenney" <paulmck@...nel.org>, Boqun Feng
 <boqun.feng@...il.com>, Paolo Bonzini <pbonzini@...hat.com>, Sean
 Christopherson <seanjc@...gle.com>, Wei Liu <wei.liu@...nel.org>, Dexuan
 Cui <decui@...rosoft.com>, x86@...nel.org, Arnd Bergmann <arnd@...db.de>,
 Heiko Carstens <hca@...ux.ibm.com>, Christian Borntraeger
 <borntraeger@...ux.ibm.com>, Sven Schnelle <svens@...ux.ibm.com>, Huacai
 Chen <chenhuacai@...nel.org>, Paul Walmsley <paul.walmsley@...ive.com>,
 Palmer Dabbelt <palmer@...belt.com>
Subject: Re: [patch V2 28/37] rseq: Switch to fast path processing on exit
 to user

On Thu, Sep 04 2025 at 13:54, Mathieu Desnoyers wrote:
> On 2025-09-02 14:36, Thomas Gleixner wrote:
>
> All that being said, I'm perfectly fine with your improvements, but
> I just want to clarify what should be considered as relevant metrics
> that justify future optimization efforts and orient future optimization
> vs code complexity trade offs.

I understand that.

Though my main objective was to optimize for the 'nothing to see here'
case, which is hit both in a kernel compile and also in the stress test
suite as the numbers show.

I definitely was not optizing for the actual handling of critical
sections in the first place. That this turned out to be slightly more
efficient is mostly a byproduct of the main goal as I just integrated
stuff more tightly.

So the actual benchmark code which will only rarely hit that path is not
that interesting. I ran the benchmark script out of curiosity
nevertheless. Here you go:

Before:
      27.883787661 seconds time elapsed
    2983.093796000 seconds user                                                                                                                                      4.227902000 seconds sys                                                                                                                                

After:
      27.908213568 seconds time elapsed
    2994.785114000 seconds user
       2.555690000 seconds sys

The times have quite some variance across multiple runs on both kernels,
but the trend of spending significantly less kernel cycles is very
consistent.

Thanks,

        tglx