linux-kernel - Re: [PATCH RT] Defer migrate_enable migration while task state != TASK

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: <20180326155451.GA16545@zipoli.concurrent-rt.com>
Date:   Mon, 26 Mar 2018 11:54:51 -0400
From:   joe.korty@...current-rt.com
To:     Steven Rostedt <rostedt@...dmis.org>
Cc:     Julia Cartwright <julia@...com>, <bigeasy@...utronix.de>,
        <mingo@...hat.com>, <tglx@...utronix.de>,
        <linux-rt-users@...r.kernel.org>, <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH RT] Defer migrate_enable migration while task state !=
 TASK_RUNNING

Oh well.  Makes me wonder why might_sleep is testing for
!TASK_RUNNABLE though.

Thanks for the correction,
Joe


On Mon, Mar 26, 2018 at 11:35:15AM -0400, Steven Rostedt wrote:
> On Fri, 23 Mar 2018 13:21:31 -0400
> joe.korty@...current-rt.com wrote:
> 
> > My understanding is, in standard Linux and in rt, setting
> > task state to anything other than TASK_RUNNING in of itself
> > blocks preemption.
> 
> That is clearly false. The only thing that blocks preemption with a
> CONFIG_PREEMPT kernel is preempt_disable() and local_irq*() disabling.
> 
> (Note spin_locks call preempt_disable in non RT).
> 
> Otherwise, nothing will stop preemption.
> 
> >  A preemption is not really needed here
> > as it is expected that there is a schedule() written in that
> > will shortly be executed.  And if a 'involuntary schedule'
> > (ie, preemption) were allowed to occur between the task
> > state set and the schedule(), that would change the task
> > state back to TASK_RUNNING, which would cause the schedule
> > to NOP.  Thus we risk not having paused long enough here
> > for the condition we were waiting for to become true.
> 
> That is also incorrect. As Julia mentioned, a preemption keeps the
> state of the task.