[PATCHSET,v2,0/6] blk-mq: per-ctx tag caching

Message ID	20200107163037.31745-1-axboe@kernel.dk (mailing list archive)
Headers	show Return-Path: <SRS0=BJAb=24=vger.kernel.org=linux-block-owner@kernel.org> From: Jens Axboe <axboe@kernel.dk> To: linux-block@vger.kernel.org Subject: [PATCHSET v2 0/6] blk-mq: per-ctx tag caching Date: Tue, 7 Jan 2020 09:30:31 -0700 Message-Id: <20200107163037.31745-1-axboe@kernel.dk> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Sender: linux-block-owner@vger.kernel.org Precedence: bulk
Series	blk-mq: per-ctx tag caching \| expand [PATCHSET,v2,0/6] blk-mq: per-ctx tag caching [1/6] sbitmap: remove cleared bitmask [2/6] sbitmap: add batch tag retrieval [3/6] blk-mq: remove 'clear_ctx_on_error' [4/6] blk-mq: remove ctx->queue [5/6] blk-mq: add struct blk_mq_ctx_type [6/6] blk-mq: allocate tags in batches

Message ID

20200107163037.31745-1-axboe@kernel.dk (mailing list archive)

Headers

From: Jens Axboe <axboe@kernel.dk>
To: linux-block@vger.kernel.org
Subject: [PATCHSET v2 0/6] blk-mq: per-ctx tag caching
Date: Tue,  7 Jan 2020 09:30:31 -0700
Message-Id: <20200107163037.31745-1-axboe@kernel.dk>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Sender: linux-block-owner@vger.kernel.org
Precedence: bulk

Series

blk-mq: per-ctx tag caching | expand

Message

Jens Axboe Jan. 7, 2020, 4:30 p.m. UTC

Here's v2 of this patchset. It should be solid now, the previous version
didn't handle tag flushing correctly, or multiple hardware queue types.

The idea here is that we can reduce the cost of getting a tag for a new
request, if we don't get them piecemeal. Add a per-ctx tag cache, and
grab batches of tags if it's empty. If it's not empty, we can just find
a free bit there.

/sys/kernel/debug/block/<dev>/<hctx>/<cpu>/tag_hit holds some stats
associated with this, so you can check how it's doing.

I've seen nice improvements with this in testing.