Re: [PATCH v5] HID: corsair-void: Add Corsair Void headset family driver

[Jiri Slaby <jirislaby@kernel.org>]

On 11. 02. 25, 12:30, Stuart wrote:
>> And what about having one work for all three (reg, up, unreg) as I
>> suggested. As I wrote, IMO, you would need no lock then. Only flags to
>> see what to do there.
> 
> I wasn't super keen on that because it felt like forcing three otherwise
> separate things into one, but it makes sense if I look at it as corsair_void's
> work function.
> 
> Is there a guarantee that the same work item won't run twice at the same time?

It's a single work, it cannot run twice.

> Do I need to make a work queue for that? I'm also not sure how to pass
> arguments to the work function. I'm sure all of that is answered by some
> documentation somewhere, but I couldn't find it myself when I looked previously.

E.g. set_bit() (one bit per OP) to something like 'unsigned long flags' 
in 'struct corsair_void_drvdata'. But it depends if you want to preserve 
order of actions... (Which was not guaranteed with separate queues 
anyway, so I assume not?)

thanks,
-- 
js
suse labs