linux-kernel - Re: [RFC PATCH] kernel: sched: Provide a pointer to the valid CPU mask

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, 6 Apr 2017 13:56:12 +0200 (CEST)
From:   Thomas Gleixner <tglx@...utronix.de>
To:     Peter Zijlstra <peterz@...radead.org>
cc:     Sebastian Andrzej Siewior <bigeasy@...utronix.de>,
        linux-kernel@...r.kernel.org, Mike Galbraith <efault@....de>,
        Ingo Molnar <mingo@...e.hu>,
        "Rafael J . Wysocki" <rjw@...ysocki.net>
Subject: Re: [RFC PATCH] kernel: sched: Provide a pointer to the valid CPU
 mask

On Thu, 6 Apr 2017, Peter Zijlstra wrote:
> On Thu, Apr 06, 2017 at 01:03:16PM +0200, Thomas Gleixner wrote:
> 
> > Fair enough, but I prefer to have the ability for a temporary restriction
> > of the user space visible cpus allowed mask in the core code, which can be
> > used for both migrate_disable() and things like that MIPS FPU stuff rather
> > than all those home brewn hackeries which are prone to bitrot, security
> > issues and subtle wreckage. cpus_allowed should be solely under sched core
> > control and not accessible from anything outside.
> 
> So the big difference between what MIPS does and what migrate_disable()
> needs here is the amount of storage.
> 
> For migrate_disable() we don't need a whole second cpumask bitmap. We
> can simply use one of the static cpumask_of() things.
> 
> Given that cpumask bitmaps can be quite large, adding another one to
> task_struct just for $very_rare_occasion, doesn't seem like a good
> thing.

A pointer to the effective mask is definitely enough. And that's what we
need for migrate_disable() as well. That still leaves the storage
requirement to MIPS.

Thanks,

	tglx