On Wed, Dec 20 2006, Chen, Kenneth W wrote:
> Kiyoshi Ueda wrote on Wednesday, December 20, 2006 9:50 AM
> > On Wed, 20 Dec 2006 14:48:49 +0100, Jens Axboe <jens.axboe@xxxxxxxxxx> wrote:
> > > Big NACK on this - it's not only really ugly, it's also buggy to pass
> > > interrupt flags as function arguments. As you also mention in the 0/1
> > > mail, this also breaks CFQ.
> > >
> > > Why do you need in-interrupt request allocation?
> >
> > Because I'd like to use blk_get_request() in q->request_fn()
> > which can be called from interrupt context like below:
> > scsi_io_completion -> scsi_end_request -> scsi_next_command
> > -> scsi_run_queue -> blk_run_queue -> q->request_fn
> >
> > [ ...]
> >
> > Do you think creating another function like blk_get_request_nowait()
> > is acceptable?
>
> You don't need to create another function. blk_get_request already
> have both wait and nowait semantics via gfp_mask argument. If you can
> not block, then clear __GFP_WAIT bit in the mask before calling
> blk_get_request.
Doesn't work, get_request() assumes that the caller grabbed the queue
lock and disabled interrupts, and does an unconditionaly
spin_unlock_irq()
inside it. So you can NEVER use get_request() for even GFP_ATOMIC
allocations, as it assumes the original context was a process context.
--
Jens Axboe
--
dm-devel mailing list
dm-devel@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/dm-devel
