[openib-general] Re: ipoib and core patches for 2.6.17

[Michael S. Tsirkin]

Quoting r. Roland Dreier <rdreier at cisco.com>:
> Subject: Re: ipoib and core patches for 2.6.17
> 
>     Michael> I see. I wish workqueue could simply be changed to do
>     Michael> nothing on flush if I'm running out of the same
>     Michael> workqueue.  I'm not sure setting this as default
>     Michael> behaviour will fly, but maybe add an option for this? 
>     Michael> That would solve this problem, would it not?
> 
> Probably.  I think I just need to really understand the patch to see
> how confusing it is.

The flush flag here has the same rules as the flag in stop_thread:
we set it unless we are running out of the workqueue, and we dont
need the flush if called from workqueue since cancel
done out of the same workqueue does not need a flush.

>     Michael> Actually, since we dont hold any locks when flushing the
>     Michael> queue, we can't deadlock, can we?
> 
> That's a good point.  Yes, it's probably safe.
> 
>  - R.
> 

-- 
Michael S. Tsirkin
Staff Engineer, Mellanox Technologies