linux-kernel - Re: [PATCH v2] sched/debug: Add new tracepoint to track cpu

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: <20210111140842.hwl4qojw3qymzw34@e107158-lin>
Date:   Mon, 11 Jan 2021 14:08:42 +0000
From:   Qais Yousef <qais.yousef@....com>
To:     Peter Zijlstra <peterz@...radead.org>
Cc:     Phil Auld <pauld@...hat.com>,
        Dietmar Eggemann <dietmar.eggemann@....com>,
        vincent.donnefort@....com, mingo@...hat.com,
        vincent.guittot@...aro.org, linux-kernel@...r.kernel.org,
        valentin.schneider@....com
Subject: Re: [PATCH v2] sched/debug: Add new tracepoint to track cpu_capacity

On 01/11/21 15:04, Peter Zijlstra wrote:
> On Mon, Jan 04, 2021 at 06:26:42PM +0000, Qais Yousef wrote:
> 
> > So I have a proper patch for that now, that actually turned out to be really
> > tiny once you untangle exactly what is missing.
> > 
> > Peter, bpf programs aren't considered ABIs AFAIK, do you have concerns about
> > that?
> 
> In order to use these you need to rely on BTF to get anything useful
> done right? And anything that relies on BTF cannot be ABI.

Yes. To decode struct rq for instance one has to either hardcode it in their
program or use BTF to get the definition dynamically.

The worry is if we modify the function signature of the tracepoint. Alexei did
confirm this can't be an ABI and I'm adding additional documentation to make
this crystal clear.

Thanks

--
Qais Yousef