Re: [PATCH RT 21/23] sched: migrate_enable: Busy loop until the migration request is completed

[Tom Zanussi]

Hi David,

On Thu, 2020-03-05 at 13:38 +0000, David Laight wrote:
> From: zanussi@xxxxxxxxxx
> > Sent: 27 February 2020 14:34
> > [ Upstream commit 140d7f54a5fff02898d2ca9802b39548bf7455f1 ]
> > 
> > If user task changes the CPU affinity mask of a running task it
> > will
> > dispatch migration request if the current CPU is no longer allowed.
> > This
> > might happen shortly before a task enters a migrate_disable()
> > section.
> > Upon leaving the migrate_disable() section, the task will notice
> > that
> > the current CPU is no longer allowed and will will dispatch its own
> > migration request to move it off the current CPU.
> > While invoking __schedule() the first migration request will be
> > processed and the task returns on the "new" CPU with "arg.done =
> > 0". Its
> > own migration request will be processed shortly after and will
> > result in
> > memory corruption if the stack memory, designed for request, was
> > used
> > otherwise in the meantime.
> > 
> > Spin until the migration request has been processed if it was
> > accepted.
> 
> What happens if the process changing the affinity mask is running
> at a higher RT priority than that of the task being changed and
> the new mask requires it run on the same cpu?
> 
> 	David
> 

This patch, 'sched: migrate_enable: Use per-cpu cpu_stop_work', removes
the busy wait and is queued for the next update:

https://lore.kernel.org/linux-rt-users/20200203173732.ldbgbpwao7xm23mm@xxxxxxxxxxxxx/T/#mf19a8af38ac4ea0cc01775835e9d715f175f0b7b

Thanks,

Tom


> -
> Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes,
> MK1 1PT, UK
> Registration No: 1397386 (Wales)
>