diff options
| author | Yang Guo <guoyang2@huawei.com> | 2019-08-28 11:19:23 -0400 |
|---|---|---|
| committer | Theodore Ts'o <tytso@mit.edu> | 2019-08-28 11:19:23 -0400 |
| commit | 520f897a3554b0665af1ae5d5ba286f290cecf5c (patch) | |
| tree | 362c3faf7e1f8f026276be06334b25249d2a262f /fs/ext4/extents_status.h | |
| parent | 7727ae52975d4f4ef7ff69ed8e6e25f6a4168158 (diff) | |
ext4: use percpu_counters for extent_status cache hits/misses
@es_stats_cache_hits and @es_stats_cache_misses are accessed frequently in
ext4_es_lookup_extent function, it would influence the ext4 read/write
performance in NUMA system. Let's optimize it using percpu_counter,
it is profitable for the performance.
The test command is as below:
fio -name=randwrite -numjobs=8 -filename=/mnt/test1 -rw=randwrite
-ioengine=libaio -direct=1 -iodepth=64 -sync=0 -norandommap
-group_reporting -runtime=120 -time_based -bs=4k -size=5G
And the result is better 10% than the initial implement:
without the patch，IOPS=197k, BW=770MiB/s (808MB/s)(90.3GiB/120002msec)
with the patch, IOPS=218k, BW=852MiB/s (894MB/s)(99.9GiB/120002msec)
Cc: "Theodore Ts'o" <tytso@mit.edu>
Cc: Andreas Dilger <adilger.kernel@dilger.ca>
Cc: Eric Biggers <ebiggers@kernel.org>
Signed-off-by: Yang Guo <guoyang2@huawei.com>
Signed-off-by: Shaokun Zhang <zhangshaokun@hisilicon.com>
Diffstat (limited to 'fs/ext4/extents_status.h')
| -rw-r--r-- | fs/ext4/extents_status.h | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/fs/ext4/extents_status.h b/fs/ext4/extents_status.h index 5e5c4a40d863..825313c59752 100644 --- a/fs/ext4/extents_status.h +++ b/fs/ext4/extents_status.h @@ -70,8 +70,8 @@ struct ext4_es_tree { struct ext4_es_stats { unsigned long es_stats_shrunk; - unsigned long es_stats_cache_hits; - unsigned long es_stats_cache_misses; + struct percpu_counter es_stats_cache_hits; + struct percpu_counter es_stats_cache_misses; u64 es_stats_scan_time; u64 es_stats_max_scan_time; struct percpu_counter es_stats_all_cnt; |