Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Liu Song <liusong@linux.alibaba.com>
To:     hch@lst.de
Cc:     axboe@kernel.dk, kbusch@kernel.org, linux-kernel@vger.kernel.org,
        linux-nvme@lists.infradead.org, liusong@linux.alibaba.com,
        sagi@grimberg.me
Subject: [RFC PATCH v2] blk-mq: hctx has only one ctx mapping is no need to redirect the completion
Date:   Thu, 22 Sep 2022 16:27:54 +0800
Message-Id: <1663835274-3703-1-git-send-email-liusong@linux.alibaba.com>
In-Reply-To: <20220922062010.GA27946@lst.de>
References: <20220922062010.GA27946@lst.de>
Precedence: bulk

From: Liu Song <liusong@linux.alibaba.com>

High-performance NVMe devices usually support a larger number of hw queue,
which ensures a 1:1 mapping of hctx and ctx. In this case there will be no
remote request, so we don't need to care about it.

Signed-off-by: Liu Song <liusong@linux.alibaba.com>
---
 block/blk-mq.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/block/blk-mq.c b/block/blk-mq.c
index c11949d..7f9be13 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -1093,10 +1093,11 @@ bool blk_mq_complete_request_remote(struct request *rq)
 	WRITE_ONCE(rq->state, MQ_RQ_COMPLETE);
 
 	/*
-	 * For a polled request, always complete locally, it's pointless
-	 * to redirect the completion.
+	 * If the request's hctx has only one ctx mapping,
+	 * or is a polled request, both always complete locally,
+	 * it's pointless to redirect the completion.
 	 */
-	if (rq->cmd_flags & REQ_POLLED)
+	if (rq->mq_hctx->nr_ctx == 1 || rq->cmd_flags & REQ_POLLED)
 		return false;
 
 	if (blk_mq_complete_need_ipi(rq)) {
-- 
1.8.3.1