linux-kernel - Re: [PATCH for 5.9 v2 1/4] futex: introduce FUTEX

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]

Date:   Tue, 11 Aug 2020 11:26:20 -0400
From:   Rik van Riel <riel@...riel.com>
To:     peterz@...radead.org, Peter Oskolkov <posk@...k.io>
Cc:     Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>, Ingo Molnar <mingo@...nel.org>,
        Darren Hart <dvhart@...radead.org>,
        Vincent Guittot <vincent.guittot@...aro.org>,
        Peter Oskolkov <posk@...gle.com>,
        Andrei Vagin <avagin@...gle.com>, Paul Turner <pjt@...gle.com>,
        Ben Segall <bsegall@...gle.com>, Aaron Lu <aaron.lwe@...il.com>
Subject: Re: [PATCH for 5.9 v2 1/4] futex: introduce FUTEX_SWAP operation

On Tue, 2020-08-04 at 14:31 +0200, peterz@...radead.org wrote:
> On Mon, Aug 03, 2020 at 03:15:07PM -0700, Peter Oskolkov wrote:
> > A simplified/idealized use case: imagine a multi-user service
> > application
> > (e.g. a DBMS) that has to implement the following user CPU quota
> > policy:
> 
> So the last posting made hackernews; and there a bunch expressed far
> more interest in coroutines, which, if I'm not mistaken, can also be
> implemented using all this.
> 
> Would that not make for a far simpler and more convincing use-case?

Also just worker threads in general. Instead of waking up
an idle worker thread every time a new request comes in,
why not allow worker threads to indicate "I'm almost done",
and the main/receiving thread to enqueue the new request,
to be handled when the worker thread is done with the current
request?

> I really think we want to have block/resume detection sorted before
> this
> goes anywhere, I also strongly feel BPF should not be used for
> functional interfaces like that.
> 
> That is, I want to see a complete interface before I want to commit
> to
> an ABI that we're stuck with.

The work case above also needs to have that figured out,
for the (slow path, but still common) case of the worker
thread actually having gone to sleep while the work got
enqueued, and then needing to be woken up anyway.

Not sure that is a direction you are interested in, given
the description, but it could make coroutines and worker
threads more efficient :)

-- 
All Rights Reversed.

Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)