Re: [PATCH 2/2] cfq-iosched: get rid of the need for __GFP_FAIL in cfq_find_alloc_queue()

[Jens Axboe <jens.axboe@oracle.com>]

On Thu, Jul 02 2009, Shan Wei wrote:
> Jens Axboe said:
> > On Wed, Jul 01 2009, Shan Wei wrote:
> >> Jens Axboe said:
> >>> Setup an emergency fallback cfqq that we allocate at IO scheduler init
> >>> time. If the slab allocation fails in cfq_find_alloc_queue(), we'll just
> >>> punt IO to that cfqq instead. This ensures that cfq_find_alloc_queue()
> >>> never fails without having to ensure free memory.
> >>>
> >>> Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
> >>> ---
> >>>  
> >>> @@ -1740,11 +1745,8 @@ cfq_get_queue(struct cfq_data *cfqd, int is_sync, struct io_context *ioc,
> >>>  		cfqq = *async_cfqq;
> >>>  	}
> >>>  
> >>> -	if (!cfqq) {
> >>> +	if (!cfqq)
> >>>  		cfqq = cfq_find_alloc_queue(cfqd, is_sync, ioc, gfp_mask);
> >>> -		if (!cfqq)
> >>> -			return NULL;
> >>> -	}
> >> I jsut reviewed the code and found that the check of cfqq is also redundant
> >> after doing cfq_get_queue() in cfq_set_request.
> >>
> >> The patch is based on Linus's main tree. 
> > 
> > It's not redundant in Linus' tree, cfq_get_queue() can return NULL for
> > != __GFP_WAIT.
> > 
> 
> Yes. So, the patch is only for "for-linus" branch of your tree, not for Linus's tree.
> 
> I noticed the patch is in your tree now, thanks.

Sorry I should have been more clear as well, I did merge it for
for-linus. It's just your original wording said if was against Linus'
main tree, where it didn't apply (the patch would apply, but it would be
wrong :-)

It's in upstream now.

-- 
Jens Axboe