linux-kernel - RE: [PATCH] sched: topology: make cache topology separate from cpu topology

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: <SL2PR06MB3082047700A8440BD0DE80DEBD0F9@SL2PR06MB3082.apcprd06.prod.outlook.com>
Date:   Mon, 14 Mar 2022 02:37:48 +0000
From:   王擎 <wangqing@...o.com>
To:     Peter Zijlstra <peterz@...radead.org>
CC:     Catalin Marinas <catalin.marinas@....com>,
        Will Deacon <will@...nel.org>,
        Sudeep Holla <sudeep.holla@....com>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        "Rafael J. Wysocki" <rafael@...nel.org>,
        Ingo Molnar <mingo@...hat.com>,
        Juri Lelli <juri.lelli@...hat.com>,
        Vincent Guittot <vincent.guittot@...aro.org>,
        Dietmar Eggemann <dietmar.eggemann@....com>,
        Steven Rostedt <rostedt@...dmis.org>,
        Ben Segall <bsegall@...gle.com>, Mel Gorman <mgorman@...e.de>,
        Daniel Bristot de Oliveira <bristot@...hat.com>,
        "linux-arm-kernel@...ts.infradead.org" 
        <linux-arm-kernel@...ts.infradead.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: RE: [PATCH] sched: topology: make cache topology separate from cpu
 topology

 
> > From: Wang Qing <wangqing@...o.com>
> > 
> > Some architectures(e.g. ARM64), caches are implemented below:
> > cluster:              ****** cluster 0 *****      ****** cluster 1 *****
> > core:                 0      1      2      3      4      5      6      7
> > cache(Leveln):        **cache0**   **cache1**  **cache2**   **cache3**
> > sd_llc_id(current):   0      0      0      0      4      4      4      4
> > sd_llc_id(should be): 0      0      2      2      4      4      6      6
> > 
> > Caches and cpus have different topology, this causes cpus_share_cache()
> > return the wrong value, which will affect the CPU load balance.
> > 
> > Cache topology should be separated with CPU topology, it can be obtained
> > from "next-level-cache" in DTS preferentially.
> 
> If your clusters do not have cache, then you're currently setting
> SD_SHARE_PKG_RESOURCES wrong, if they do, things are correct.

If there is a shared cache(L3) between clusters(cls 0 and cls 1) for all cores, 
but not within the cluster like above, should we set SD_SHARE_PKG_RESOURCES 
for MC0(cls 0), or just set SD_SHARE_PKG_RESOURCES for CLS?
> 
> If you want to represent L2, use the new fangled cluster level or
> something, that's what it's there for.
> 
> That is, you can represent the above like:
> 
>        DIE:    0-7
>        MC:     0-3,            4-7
>         CLS:    0-1,1-2,        4-5,6-7
> 
> But if there is cache at MC, LLC is what it is.

There is no CLS support for LTS now, any plans to backport?

Thanks,
Wang