linux-kernel - Re: [PATCH v2 04/11] smp: replace smp_call_function_single() with smp

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, 25 Apr 2022 11:33:17 +0200
From:   Peter Zijlstra <peterz@...radead.org>
To:     Donghai Qiao <dqiao@...hat.com>
Cc:     akpm@...ux-foundation.org, sfr@...b.auug.org.au, arnd@...db.de,
        heying24@...wei.com, andriy.shevchenko@...ux.intel.com,
        axboe@...nel.dk, rdunlap@...radead.org, tglx@...utronix.de,
        gor@...ux.ibm.com, donghai.w.qiao@...il.com,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 04/11] smp: replace smp_call_function_single() with
 smp_call()

On Fri, Apr 22, 2022 at 04:00:33PM -0400, Donghai Qiao wrote:
> Eliminated the percpu global csd_data and temporarily hook up
> to smp_call().
> 
> There is no obvious reason or evidence that the differentiation

Using the on-stack csd seems like an obvious benefit to me. Stack is
often cache-hot, while the percpu stuff is likely a cache-miss. Worse,
the call_function_data as used by the mask functions is likely a double
miss due to the O(n^2) nature of the thing.