block: Make rq_affinity = 1 work as expected
authorTao Ma <boyu.mt@taobao.com>
Fri, 5 Aug 2011 07:37:10 +0000 (09:37 +0200)
committerJens Axboe <jaxboe@fusionio.com>
Fri, 5 Aug 2011 07:37:47 +0000 (09:37 +0200)
commit35ae66e0a09ab70ed588e65f26b4c725cd1656b6
tree10f866f13a32109f7fa23ab8ab392fab759c9f6d
parentf41c53a569c4cf0556893ec9cfcf697d069799e1
block: Make rq_affinity = 1 work as expected

Commit 5757a6d76c introduced a new rq_affinity = 2 so as to make
the request completed in the __make_request cpu. But it makes the
old rq_affinity = 1 not work any more. The root cause is that
if the 'cpu' and 'req->cpu' is in the same group and cpu != req->cpu,
ccpu will be the same as group_cpu, so the completion will be
excuted in the 'cpu' not 'group_cpu'.

This patch fix problem by simpling removing group_cpu and the codes
are more explicit now. If ccpu == cpu, we complete in cpu, otherwise
we raise_blk_irq to ccpu.

Cc: Christoph Hellwig <hch@infradead.org>
Cc: Roland Dreier <roland@purestorage.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Jens Axboe <jaxboe@fusionio.com>
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Reviewed-by: Shaohua Li <shaohua.li@intel.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
block/blk-softirq.c