linux-kernel - Re: [PATCH v2 6/7] sched: Shard per-LLC shared runqueues

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]

Message-ID: <20230712122252.GG3100107@hirez.programming.kicks-ass.net>
Date:   Wed, 12 Jul 2023 14:22:52 +0200
From:   Peter Zijlstra <peterz@...radead.org>
To:     "Gautham R. Shenoy" <gautham.shenoy@....com>
Cc:     David Vernet <void@...ifault.com>, linux-kernel@...r.kernel.org,
        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, kprateek.nayak@....com,
        aaron.lu@...el.com, clm@...a.com, tj@...nel.org,
        roman.gushchin@...ux.dev, kernel-team@...a.com
Subject: Re: [PATCH v2 6/7] sched: Shard per-LLC shared runqueues

On Wed, Jul 12, 2023 at 03:36:27PM +0530, Gautham R. Shenoy wrote:

> On some Intel servers, it is possible that the CPU numbers are
> interleaved across the two sockets. On my 2 socket, 32Cores per socket
> Ice Lake Server, all the even numbered CPUs are in one socket and all
> the odd numbered CPUs in the other socket.
> 
> The SMT siblings are {0,64}, {2, 66}, .... on one socket and {1, 65},
> {3, 67}, .. on the other.

Yeah, Intel SMT enumeration is a mess. There's a random mix of {n,n+1}
and {0..n-1} {n..2n-1}. And then there's the fun hybrid stuff.  Those
appear to do {n,n+1} for the big cores and then continue with the small
cores in a dense set. My 8+8 ADL, has:

{0,1} {2,3} {4,5} {6,7} {8,9} {10,11} {12,13} {14,15} {16} {17} {18} {19} {20} {21} {22} {23}

I suspect it might be easier to re-number the whole show at boot to fit
a sane pattern rather than trying to match the various random garbage
gifted to us by the BIOS.


I wouldn't worry about it too much at this point.