| Message ID | 20190326131925.60255-1-yuyufen@huawei.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | blk-mq: add trace block plug and unplug for multiple queues | expand |
On Tue, Mar 26, 2019 at 09:19:25PM +0800, Yufen Yu wrote: > For now, we just trace plug for single queue device or drivers > provide .commit_rqs, and have not trace plug for multiple queues > device. But, unplug events will be recorded when call > blk_mq_flush_plug_list(). Then, trace events will be asymmetrical, > just have unplug and without plug. > > This patch add trace plug and unplug for multiple queues device in > blk_mq_make_request(). After that, we can accurately trace plug and > unplug for multiple queues. > > Signed-off-by: Yufen Yu <yuyufen@huawei.com> Looks good, Reviewed-by: Christoph Hellwig <hch@lst.de>
ping ? On 2019/3/27 16:13, Christoph Hellwig wrote: > On Tue, Mar 26, 2019 at 09:19:25PM +0800, Yufen Yu wrote: >> For now, we just trace plug for single queue device or drivers >> provide .commit_rqs, and have not trace plug for multiple queues >> device. But, unplug events will be recorded when call >> blk_mq_flush_plug_list(). Then, trace events will be asymmetrical, >> just have unplug and without plug. >> >> This patch add trace plug and unplug for multiple queues device in >> blk_mq_make_request(). After that, we can accurately trace plug and >> unplug for multiple queues. >> >> Signed-off-by: Yufen Yu <yuyufen@huawei.com> > Looks good, > > Reviewed-by: Christoph Hellwig <hch@lst.de> > > . >
On 3/26/19 7:19 AM, Yufen Yu wrote: > For now, we just trace plug for single queue device or drivers > provide .commit_rqs, and have not trace plug for multiple queues > device. But, unplug events will be recorded when call > blk_mq_flush_plug_list(). Then, trace events will be asymmetrical, > just have unplug and without plug. > > This patch add trace plug and unplug for multiple queues device in > blk_mq_make_request(). After that, we can accurately trace plug and > unplug for multiple queues. Applied, thanks.
diff --git a/block/blk-mq.c b/block/blk-mq.c index 5486c7d8b408..4aa673a6d285 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -1987,11 +1987,13 @@ static blk_qc_t blk_mq_make_request(struct request_queue *q, struct bio *bio) plug->rq_count--; } blk_add_rq_to_plug(plug, rq); + trace_block_plug(q); blk_mq_put_ctx(data.ctx); if (same_queue_rq) { data.hctx = same_queue_rq->mq_hctx; + trace_block_unplug(q, 1, true); blk_mq_try_issue_directly(data.hctx, same_queue_rq, &cookie, false, true); }
For now, we just trace plug for single queue device or drivers provide .commit_rqs, and have not trace plug for multiple queues device. But, unplug events will be recorded when call blk_mq_flush_plug_list(). Then, trace events will be asymmetrical, just have unplug and without plug. This patch add trace plug and unplug for multiple queues device in blk_mq_make_request(). After that, we can accurately trace plug and unplug for multiple queues. Signed-off-by: Yufen Yu <yuyufen@huawei.com> --- block/blk-mq.c | 2 ++ 1 file changed, 2 insertions(+)