linux-kernel - Re: [PATCH v4 2/2] lib/vsprintf: Initialize vsprintf's pointer hash once the random core is ready.

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:   Thu, 29 Sep 2022 10:52:02 +0200
From:   Petr Mladek <pmladek@...e.com>
To:     "Jason A. Donenfeld" <Jason@...c4.com>
Cc:     Sebastian Andrzej Siewior <bigeasy@...utronix.de>,
        linux-kernel@...r.kernel.org, Theodore Ts'o <tytso@....edu>,
        Andy Shevchenko <andriy.shevchenko@...ux.intel.com>,
        John Ogness <john.ogness@...utronix.de>,
        Mike Galbraith <efault@....de>,
        Peter Zijlstra <peterz@...radead.org>,
        Rasmus Villemoes <linux@...musvillemoes.dk>,
        Sergey Senozhatsky <senozhatsky@...omium.org>,
        Steven Rostedt <rostedt@...dmis.org>,
        Thomas Gleixner <tglx@...utronix.de>, Tejun Heo <tj@...nel.org>
Subject: Re: [PATCH v4 2/2] lib/vsprintf: Initialize vsprintf's pointer hash
 once the random core is ready.

On Wed 2022-09-28 11:21:05, Jason A. Donenfeld wrote:
> On 9/28/22, Sebastian Andrzej Siewior <bigeasy@...utronix.de> wrote:
> >
> >> I could replace "system_unbound_wq" with "system_wq" when
> >> pushing. Is anybody against it, please?
> >
> > so schedule_delayed_work() then?

yup.

> > I don't mind at all. I used that one just because serialisation is not
> > needed and neither is the CPU important.
> 
> Indeed, given that this very much is unbound, I think Sebastian's
> original patch makes most sense.

Yes, the work does not need any specific CPU. The thing is that the
normal system_wq is the preferred one. Any other workqueues should
be used only when there is a particular reason for it.

The unbound_wq should be used only when:

    + the work needs a lot of CPU time.

    + there are waves on related (sleeping) work items that might be
      triggered from different CPUs.

In our case, the work is only one and short. The preferred
system_wq is perfectly fine.

Best Regards,
Petr

PS: It is not obvious. Tejun told me this when I converted a kthread
    into the workqueue API. Also I spent quite some time understanding
    the workqueue code recently.