Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1351717 > unrolled thread

Re: Softirq priority inversion from "softirq: reduce latencies"

Started bySebastian Andrzej Siewior <bigeasy@linutronix.de>
First post2016-03-07 16:40 +0100
Last post2016-03-07 18:10 +0100
Articles 2 — 2 participants

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.


Contents

  Re: Softirq priority inversion from "softirq: reduce latencies" Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-03-07 16:40 +0100
    Re: Softirq priority inversion from "softirq: reduce latencies" Mike Galbraith <umgwanakikbuti@gmail.com> - 2016-03-07 18:10 +0100

#1351717 — Re: Softirq priority inversion from "softirq: reduce latencies"

FromSebastian Andrzej Siewior <bigeasy@linutronix.de>
Date2016-03-07 16:40 +0100
SubjectRe: Softirq priority inversion from "softirq: reduce latencies"
Message-ID<ra5qi-4EI-3@gated-at.bofh.it>
On 02/29/2016 05:58 AM, Mike Galbraith wrote:
> WRT -rt: if dma tasklets really do have hard (ish) constraints, -rt
> recently "broke" in the same way.. of all softirqs which are deferred
> to kthread context, due to a recent change, only timer/hrtimer are
> executed at realtime priority by default.

no. All softirqs are invoked in the context of the current process that
triggerd the softirq invocation. If NAPI goes on for too long (or other
softirq can't be executed in this context) it will continue in the
ksoftirqd. And this threads runs at a normal priority like it does in
mainline.
I adjusted it with mainline.

> 	-Mike
> 
Sebastian

[toc] | [next] | [standalone]


#1351805

FromMike Galbraith <umgwanakikbuti@gmail.com>
Date2016-03-07 18:10 +0100
Message-ID<ra6Po-5Ig-7@gated-at.bofh.it>
In reply to#1351717
On Mon, 2016-03-07 at 16:31 +0100, Sebastian Andrzej Siewior wrote:
> On 02/29/2016 05:58 AM, Mike Galbraith wrote:
> > WRT -rt: if dma tasklets really do have hard (ish) constraints, -rt
> > recently "broke" in the same way.. of all softirqs which are deferred
> > to kthread context, due to a recent change, only timer/hrtimer are
> > executed at realtime priority by default.
> 
> no. All softirqs are invoked in the context of the current process that
> triggerd the softirq invocation. If NAPI goes on for too long (or other
> softirq can't be executed in this context) it will continue in the
> ksoftirqd. And this threads runs at a normal priority like it does in
> mainline.
> I adjusted it with mainline.

Yeah, that's what I said, the ksoftirqd case became the same in -rt.

	-Mike

[toc] | [prev] | [standalone]


Back to top | Article view | linux.kernel


csiph-web