Jens Axboe wrote:
> On Mon, Mar 28 2005, Chen, Kenneth W wrote:
> 
>>This patch was posted last year and if I remember correctly, Jens said
>>he is OK with the patch.  In function __generic_unplug_deivce(), kernel
>>can use a cheaper function elv_queue_empty() instead of more expensive
>>elv_next_request to find whether the queue is empty or not. blk_run_queue
>>can also made conditional on whether queue's emptiness before calling
>>request_fn().
>>
>>
>>Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
> 
> 
> Looks good, thanks.
> 
> Signed-off-by: Jens Axboe <axboe@suse.de>
> 

Speaking of which, I've had a few ideas lying around for possible
performance improvement in the block code.

I haven't used a big disk array (or tried any simulation), but I'll
attach the patch if you're looking into that area.

It puts in a few unlikely()s, but the main changes are:
- don't generic_unplug_device unconditionally in get_request_wait,
- removes the relock/retry merge mechanism in __make_request if we
   aren't able to get the GFP_ATOMIC allocation. Just fall through
   and assume the chances of getting a merge will be small (is this
   a valid assumption? Should measure it I guess).
- removes the GFP_ATOMIC allocation. That's always a good thing.