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>
This commit is contained in:
Yang Guo 2019-08-28 11:19:23 -04:00 committed by Theodore Ts'o
parent 7727ae5297
commit 520f897a35
2 changed files with 26 additions and 15 deletions

View File

@ -948,7 +948,7 @@ int ext4_es_lookup_extent(struct inode *inode, ext4_lblk_t lblk,
es->es_pblk = es1->es_pblk;
if (!ext4_es_is_referenced(es1))
ext4_es_set_referenced(es1);
stats->es_stats_cache_hits++;
percpu_counter_inc(&stats->es_stats_cache_hits);
if (next_lblk) {
node = rb_next(&es1->rb_node);
if (node) {
@ -959,7 +959,7 @@ int ext4_es_lookup_extent(struct inode *inode, ext4_lblk_t lblk,
*next_lblk = 0;
}
} else {
stats->es_stats_cache_misses++;
percpu_counter_inc(&stats->es_stats_cache_misses);
}
read_unlock(&EXT4_I(inode)->i_es_lock);
@ -1586,9 +1586,9 @@ int ext4_seq_es_shrinker_info_show(struct seq_file *seq, void *v)
seq_printf(seq, "stats:\n %lld objects\n %lld reclaimable objects\n",
percpu_counter_sum_positive(&es_stats->es_stats_all_cnt),
percpu_counter_sum_positive(&es_stats->es_stats_shk_cnt));
seq_printf(seq, " %lu/%lu cache hits/misses\n",
es_stats->es_stats_cache_hits,
es_stats->es_stats_cache_misses);
seq_printf(seq, " %lld/%lld cache hits/misses\n",
percpu_counter_sum_positive(&es_stats->es_stats_cache_hits),
percpu_counter_sum_positive(&es_stats->es_stats_cache_misses));
if (inode_cnt)
seq_printf(seq, " %d inodes on list\n", inode_cnt);
@ -1615,35 +1615,46 @@ int ext4_es_register_shrinker(struct ext4_sb_info *sbi)
sbi->s_es_nr_inode = 0;
spin_lock_init(&sbi->s_es_lock);
sbi->s_es_stats.es_stats_shrunk = 0;
sbi->s_es_stats.es_stats_cache_hits = 0;
sbi->s_es_stats.es_stats_cache_misses = 0;
err = percpu_counter_init(&sbi->s_es_stats.es_stats_cache_hits, 0,
GFP_KERNEL);
if (err)
return err;
err = percpu_counter_init(&sbi->s_es_stats.es_stats_cache_misses, 0,
GFP_KERNEL);
if (err)
goto err1;
sbi->s_es_stats.es_stats_scan_time = 0;
sbi->s_es_stats.es_stats_max_scan_time = 0;
err = percpu_counter_init(&sbi->s_es_stats.es_stats_all_cnt, 0, GFP_KERNEL);
if (err)
return err;
goto err2;
err = percpu_counter_init(&sbi->s_es_stats.es_stats_shk_cnt, 0, GFP_KERNEL);
if (err)
goto err1;
goto err3;
sbi->s_es_shrinker.scan_objects = ext4_es_scan;
sbi->s_es_shrinker.count_objects = ext4_es_count;
sbi->s_es_shrinker.seeks = DEFAULT_SEEKS;
err = register_shrinker(&sbi->s_es_shrinker);
if (err)
goto err2;
goto err4;
return 0;
err2:
err4:
percpu_counter_destroy(&sbi->s_es_stats.es_stats_shk_cnt);
err1:
err3:
percpu_counter_destroy(&sbi->s_es_stats.es_stats_all_cnt);
err2:
percpu_counter_destroy(&sbi->s_es_stats.es_stats_cache_misses);
err1:
percpu_counter_destroy(&sbi->s_es_stats.es_stats_cache_hits);
return err;
}
void ext4_es_unregister_shrinker(struct ext4_sb_info *sbi)
{
percpu_counter_destroy(&sbi->s_es_stats.es_stats_cache_hits);
percpu_counter_destroy(&sbi->s_es_stats.es_stats_cache_misses);
percpu_counter_destroy(&sbi->s_es_stats.es_stats_all_cnt);
percpu_counter_destroy(&sbi->s_es_stats.es_stats_shk_cnt);
unregister_shrinker(&sbi->s_es_shrinker);

View File

@ -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;