Re: [PATCH -v2 1/5] sched: Fix ttwu() race

[Chris Wilson]

Quoting Peter Zijlstra (2020-07-23 19:28:41)
> On Wed, Jul 22, 2020 at 10:57:56AM +0100, Chris Wilson wrote:
> 
> > Perhaps more damning is that I can replace WF_ON_CPU with p->on_cpu to
> > suppress the warning:
> 
> *argh*, I'm starting to go mad...
> 
> Chris, could you please try the below patch?

 ttwu-IPI-self: 1==1, p->on_cpu=0;0, task_cpu(p)=1;1
 ttwu-IPI-self: 1==1, p->on_cpu=0;0, task_cpu(p)=1;1
 ttwu-IPI-self: 0==0, p->on_cpu=0;0, task_cpu(p)=0;0
 ttwu-IPI-self: 3==3, p->on_cpu=0;0, task_cpu(p)=3;3
 ttwu-IPI-self: 2==2, p->on_cpu=0;0, task_cpu(p)=2;2
 ttwu-IPI-self: 1==1, p->on_cpu=0;0, task_cpu(p)=1;1
 ttwu-IPI-self: 2==2, p->on_cpu=0;0, task_cpu(p)=2;2
 ttwu-IPI-self: 2==2, p->on_cpu=0;0, task_cpu(p)=2;2
 ttwu-IPI-self: 2==2, p->on_cpu=0;0, task_cpu(p)=2;2

> Can you also confirm that if you do:
> 
> $ echo NO_TTWU_QUEUE_ON_CPU > /debug/sched_features

With,

 sched_feat_disable(10):TTWU_QUEUE_ON_CPU

the pr_alert is still being hit

 ttwu-IPI-self: 3==3, p->on_cpu=0;0, task_cpu(p)=3;3

At which point, it darns on me. Mea culpa, stray bits being passed into
default_wake_function.

I am very sorry for the wild goose chase.
-Chris