[PATCH] nvmet: make nvmet_wq unbound

[Sagi Grimberg]

On 06/05/2024 8:33, Christoph Hellwig wrote:
> On Sun, May 05, 2024 at 01:39:51PM +0300, Sagi Grimberg wrote:
>> From: Sagi Grimberg <sagi.grimberg at vastdata.com>
>>
>> When deleting many controllers one-by-one, it takes a very
>> long time as these work elements may serialize as they are
>> scheduled on the executing cpu instead of spreading. In general
>> nvmet_wq can definitely be used for long standing work elements
>> so its better to make it unbound regardless.
> This looks sensible, thanks.
>
>> While we are at it, expose it via sysfs to allow cpumask
>> control of this workqueue if a user wants to modify it.
> This on the other looks very questionable to me.  It needs a way better
> justification and should be split into a separate patch if we really
> need it.

It doesn't have any better justification. I never really saw any other 
justification
to add WQ_SYSFS to a (usually unbound) workqueue, than simply have 
visibility
or ability to control it.

What makes you think this is not harmless and useful?