On 12/17/20 9:43 AM, Sebastian Andrzej Siewior wrote:

On 2020-12-08 13:13:19 [+0000], Christoph Hellwig wrote:

quoted

On Mon, Dec 07, 2020 at 04:52:57PM -0700, Jens Axboe wrote:

quoted

On 12/4/20 12:13 PM, Sebastian Andrzej Siewior wrote:

quoted

Controllers with multiple queues have their IRQ-handelers pinned to a
CPU. The core shouldn't need to complete the request on a remote CPU.

Remove this case and always raise the softirq to complete the request.

I don't like this one at all, it'll add a softirq jump for the fast path
for eg nvme devices.

For the real fast path, that is either a polled queue or irq driven
queues that only map to a single CPU we are never reaching this code,
so I'm not too worried.  Not that I'd complain about numbers, preferably
in the commit log.

Did Daniel provide all the numbers you/Jens were looking for or do we
still wait for some?

Yeah, I think we're good at this point. I'll queue this up for 5.11.

-- 
Jens Axboe