Re: [PATCH v7 07/11] arch/x86: enable task isolation functionality

[Christoph Lameter]

On Thu, 1 Oct 2015, Thomas Gleixner wrote:

> And I really want to see a proper engineering for that isolation
> stuff, which can be done with an out of tree patch set in the first
> place. But sure, it's more convenient to push crap into mainline and
> let everyone else deal with the fallouts.

Yes lets keep the half baked stuff out please. Firing a timer that signals
the app via a signal causes an event that is not unlike the OS noise that
we are trying to avoid. Its an asynchrononous event that may interrupt at
random locations in the code. In that case I would say its perfectly fine
to use the tick and other timer processing on the processor that requested
it. If you really want low latency and be OS intervention free then please
do not set up timers. In fact any signal should bring on full OS services
on a processor.

AFAICT one would communicate via shared memory structures rather than IPIs
within the threads of an app that requires low latency and the OS noise to
be minimal.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/