lkml.org 
[lkml]   [2018]   [Apr]   [23]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH tip/core/rcu 06/22] rcu: Rename cond_resched_rcu_qs() to cond_resched_tasks_rcu_qs()
On Mon, Apr 23, 2018 at 02:48:55PM +0200, Peter Zijlstra wrote:
> On Mon, Apr 23, 2018 at 05:32:23AM -0700, Paul E. McKenney wrote:
> > On Mon, Apr 23, 2018 at 10:53:47AM +0200, Peter Zijlstra wrote:
> > > On Sun, Apr 22, 2018 at 07:32:11PM -0700, Paul E. McKenney wrote:
> > > > Commit e31d28b6ab8f ("trace: Eliminate cond_resched_rcu_qs() in favor
> > > > of cond_resched()") substituted cond_resched() for the earlier call
> > > > to cond_resched_rcu_qs(). However, the new-age cond_resched() does
> > > > not do anything to help RCU-tasks grace periods because (1) RCU-tasks
> > > > is only enabled when CONFIG_PREEMPT=y and (2) cond_resched() is a
> > > > complete no-op when preemption is enabled. This situation results
> > > > in hangs when running the trace benchmarks.
> > >
> > > But a few patches earlier you tried to address that..
> >
> > Yes. And all this patch does is change the name to better match
> > what this primitive actually does.
> >
> > I could merge the two patches, if that would help.
>
> I understood the earlier patch was actually changing cond_resched() to
> also work for task-rcu; which resulted in it acually not being a no-op
> for PREEMPT anymore.
>
> But then here you say it doesn't because it's a no-op.
>
> It's not about the rename.. It is about one patch changing PREEMPT
> cond_resched to not be a no-op and you here justifying things because
> PREEMPT cond_resched it a no-op.

So would it help if I instead said in the commit log that it is a
no-op in production builds of the Linux kernel?

Thanx, Paul

\
 
 \ /
  Last update: 2018-04-23 15:12    [W:0.050 / U:0.308 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site