On Thu, Aug 20, 2020 at 03:40:01PM +0200, pet...@infradead.org wrote:
> Anyway, tl;dr, what do we actually want? Do we favour the embedded
> irq_work variant over smp_call_function_single_asyn() ?
For blk-mq we really like an embedded structure that allows to bounce
the work to another CPU using I
On Thu, Aug 20, 2020 at 08:19:27AM +0200, Christoph Hellwig wrote:
> On Tue, Aug 18, 2020 at 12:51:10PM +0200, Peter Zijlstra wrote:
> > if (blk_mq_complete_need_ipi(rq)) {
> > - INIT_CSD(&rq->csd, __blk_mq_complete_request_remote, rq);
> > - smp_call_function_single_async(r
On Wed, Aug 19, 2020 at 03:04:56PM -0700, Linus Torvalds wrote:
> On Wed, Aug 19, 2020 at 12:41 PM wrote:
> >
> > I'm not sure I get the "expensive irq_work queues" argument, I fully
> > agree with you that adding the atomic op is fairly crap.
>
> There's an atomic op on the actual runing side to
On Wed, Aug 19, 2020 at 11:50:55AM -0700, Linus Torvalds wrote:
> On Wed, Aug 19, 2020 at 12:22 AM wrote:
> >
> > That is, the external serialization comes from the non-atomic
> > test-and-set they both have. This works nicely when there is external
> > state that already serializes things, but to
On Tue, Aug 18, 2020 at 12:51:10PM +0200, Peter Zijlstra wrote:
> if (blk_mq_complete_need_ipi(rq)) {
> - INIT_CSD(&rq->csd, __blk_mq_complete_request_remote, rq);
> - smp_call_function_single_async(rq->mq_ctx->cpu, &rq->csd);
> + rq->work = IRQ_WORK_INIT_H
On Wed, Aug 19, 2020 at 12:41 PM wrote:
>
> I'm not sure I get the "expensive irq_work queues" argument, I fully
> agree with you that adding the atomic op is fairly crap.
There's an atomic op on the actual runing side too, because of the
whole IRQ_WORK_PENDING thing.
So you get that double hit.
On Wed, Aug 19, 2020 at 11:50:55AM -0700, Linus Torvalds wrote:
> On Wed, Aug 19, 2020 at 12:22 AM wrote:
> >
> > That is, the external serialization comes from the non-atomic
> > test-and-set they both have. This works nicely when there is external
> > state that already serializes things, but to
On Wed, Aug 19, 2020 at 12:22 AM wrote:
>
> That is, the external serialization comes from the non-atomic
> test-and-set they both have. This works nicely when there is external
> state that already serializes things, but totally comes apart (and
> causes trivial list corruption) when you get it w
On Tue, Aug 18, 2020 at 06:25:42PM +0200, Christoph Hellwig wrote:
> On Tue, Aug 18, 2020 at 12:51:10PM +0200, Peter Zijlstra wrote:
> > Convert the performance sensitive users of
> > smp_call_single_function_async() over to the new
> > irq_work_queue_remote_static().
> >
> > The new API is margin
On Tue, Aug 18, 2020 at 12:51:10PM +0200, Peter Zijlstra wrote:
> Convert the performance sensitive users of
> smp_call_single_function_async() over to the new
> irq_work_queue_remote_static().
>
> The new API is marginally less crap but taking these users away allows
> fixing up smp_call_single_f
Convert the performance sensitive users of
smp_call_single_function_async() over to the new
irq_work_queue_remote_static().
The new API is marginally less crap but taking these users away allows
fixing up smp_call_single_function_async() without risk of performance
regressions.
Signed-off-by: Pet
11 matches
Mail list logo
