Repository navigation
Conversation
|
Upstream branch: bc35965 |
8242bf1 to
5b28f57
Compare
|
Upstream branch: df29080 |
b571268 to
7e18d47
Compare
|
Upstream branch: 28924df |
7e18d47 to
0452dcb
Compare
4ddd216 to
00cc4ca
Compare
|
Upstream branch: 893e117 |
0452dcb to
7e96de5
Compare
00cc4ca to
7efd8cd
Compare
|
Upstream branch: 50d05c7 |
7e96de5 to
3073b40
Compare
7efd8cd to
a0aeca9
Compare
|
Upstream branch: 5225b8e |
3073b40 to
23e907c
Compare
a0aeca9 to
772381e
Compare
|
Upstream branch: 2f0c1cf |
23e907c to
236a5ef
Compare
772381e to
0224dee
Compare
|
Upstream branch: 2f0c1cf |
236a5ef to
5a88a45
Compare
0224dee to
f14340f
Compare
|
Upstream branch: 5878583 |
5a88a45 to
4f0a6c7
Compare
c65e2dc to
2414e80
Compare
|
Upstream branch: f010036 |
6eebdbd to
b3b0590
Compare
2414e80 to
38f4bbb
Compare
|
Upstream branch: f49a343 |
b3b0590 to
b812d38
Compare
38f4bbb to
372fcdc
Compare
|
Upstream branch: 6812ce4 |
b812d38 to
c067119
Compare
372fcdc to
c6537fe
Compare
|
Upstream branch: 72d3fcf |
c067119 to
904b78a
Compare
c6537fe to
eebb5bc
Compare
|
Upstream branch: 551c722 |
904b78a to
118d7df
Compare
eebb5bc to
bbd3af0
Compare
|
Upstream branch: ce1e022 |
118d7df to
07f1099
Compare
bbd3af0 to
a9c0b46
Compare
|
Upstream branch: e767a4e |
07f1099 to
47daea2
Compare
a9c0b46 to
72d0f5e
Compare
|
Upstream branch: a74306e |
iolatency_check_latencies() accumulates the per-window latency stat into iolat->cur_stat via latency_stat_sum(), which for non-SSD devices calls blk_rq_stat_sum(). But blk_rq_stat_sum() is contracted on a raw per-cpu src: it folds src->batch into the new mean. By the time the per-window stat reaches cur_stat it is already aggregated, with batch left at 0. Each window therefore contributes zero latency to cur_stat->mean, which on non-SSD devices stays at 0; the latency_sum_ok(&cur_stat) check that gates scaling up is always true, so the scale-up hysteresis never engages. SSD devices use the percentile path and are unaffected. Add latency_stat_merge(), which merges two already-aggregated stats by their reconstructed totals (mean * nr_samples) instead of src->batch, and use it only for the cur_stat accumulation; the earlier per-cpu sum keeps latency_stat_sum(), whose src is still raw. Fixes: d706751 ("block: introduce blk-iolatency io controller") Signed-off-by: Tao Cui <cuitao@kylinos.cn>
47daea2 to
80c2fd9
Compare
|
At least one diff in series https://patchwork.kernel.org/series/1157639/ irrelevant now for [{'archived': False, 'project': 241}] search patterns |
Pull request for series with
subject: block/blk-iolatency: fix always-zero cur_stat mean on non-SSD
version: 3
url: https://patchwork.kernel.org/project/linux-block/list/?series=1157639