diff options
| author | Jens Axboe <axboe@fb.com> | 2014-05-19 19:23:55 +0400 |
|---|---|---|
| committer | Jens Axboe <axboe@fb.com> | 2014-05-19 21:02:47 +0400 |
| commit | 1429d7c9467e1e3de0b0ff91d7e4d67c1a92f8a3 (patch) | |
| tree | 3b15abb587392becc4ba37d0869d25a4d9420d1d /include/linux | |
| parent | e93ecf602beb8439f0bdcc1fa2cbc1f31fdfb8e2 (diff) | |
| download | linux-1429d7c9467e1e3de0b0ff91d7e4d67c1a92f8a3.tar.xz | |
blk-mq: switch ctx pending map to the sparser blk_align_bitmap
Each hardware queue has a bitmap of software queues with pending
requests. When new IO is queued on a software queue, the bit is
set, and when IO is pruned on a hardware queue run, the bit is
cleared. This causes a lot of traffic. Switch this from the regular
BITS_PER_LONG bitmap to a sparser layout, similarly to what was
done for blk-mq tagging.
20% performance increase was observed for single threaded IO, and
about 15% performanc increase on multiple threads driving the
same device.
Signed-off-by: Jens Axboe <axboe@fb.com>
Diffstat (limited to 'include/linux')
| -rw-r--r-- | include/linux/blk-mq.h | 10 |
1 files changed, 8 insertions, 2 deletions
diff --git a/include/linux/blk-mq.h b/include/linux/blk-mq.h index f83d15f6e1c1..952e558ee598 100644 --- a/include/linux/blk-mq.h +++ b/include/linux/blk-mq.h @@ -11,6 +11,12 @@ struct blk_mq_cpu_notifier { void (*notify)(void *data, unsigned long action, unsigned int cpu); }; +struct blk_mq_ctxmap { + unsigned int map_size; + unsigned int bits_per_word; + struct blk_align_bitmap *map; +}; + struct blk_mq_hw_ctx { struct { spinlock_t lock; @@ -31,8 +37,8 @@ struct blk_mq_hw_ctx { void *driver_data; - unsigned int nr_ctx_map; - unsigned long *ctx_map; + struct blk_mq_ctxmap ctx_map; + unsigned int nr_ctx; struct blk_mq_ctx **ctxs; |