From: Guixin Liu <kanie@linux.alibaba.com>
To: Tao Cui <cui.tao@linux.dev>, Jens Axboe <axboe@kernel.dk>,
Tang Yizhou <tangyeechou@gmail.com>
Cc: tj@kernel.org, josef@toxicpanda.com, linux-block@vger.kernel.org,
cgroups@vger.kernel.org, linux-kernel@vger.kernel.org,
cuitao@kylinos.cn
Subject: Re: [PATCH v3] block/blk-iolatency: fix always-zero cur_stat mean on non-SSD
Date: Fri, 20 Nov 2026 10:19:09 +0800 [thread overview]
Message-ID: <856407d2-db33-4c1d-a5c9-2ba7e0053ad4@linux.alibaba.com> (raw)
In-Reply-To: <20260904065426.1281914-1-cui.tao@linux.dev>
在 2026/9/4 14:54, Tao Cui 写道:
> From: Tao Cui <cuitao@kylinos.cn>
>
> 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: d70675121546 ("block: introduce blk-iolatency io controller")
Ti Tao:
Would it be better to change the fix tag to "451bb7c33197 ("blk-iolatency: keep track of previous windows stats")"?
Best Regards,
Guixin Liu
> Signed-off-by: Tao Cui <cuitao@kylinos.cn>
>
> ---
> Changes in v3:
> - Rebase onto current linux-next head (no code change).
> - Verified with a two-kernel QEMU run (virtio-blk forced rotational,
> cgroup v2 io.latency target set below the achievable latency so the
> scale path runs every window): on the unpatched kernel cur_stat->mean
> stayed 0 across all windows while the per-window mean was 120-250us;
> with this patch cur_stat->mean tracked the per-window mean.
>
> Changes in v2:
> - Move the fix from blk-stat.c (blk_rq_stat_sum) into blk-iolatency.c,
> per Tang Yizhou's review: blk_rq_stat_sum() is contracted on a raw
> per-cpu src, so the aggregated-stat merge belongs in the iolatency
> caller, not the shared helper.
>
> v1: https://lore.kernel.org/all/20260720133813.45150-1-cui.tao@linux.dev/
> v2: https://lore.kernel.org/all/20260721030312.85655-1-cui.tao@linux.dev/
> ---
> block/blk-iolatency.c | 27 ++++++++++++++++++++++++++-
> 1 file changed, 26 insertions(+), 1 deletion(-)
>
> diff --git a/block/blk-iolatency.c b/block/blk-iolatency.c
> index 2caa79a008ad..853608a4f187 100644
> --- a/block/blk-iolatency.c
> +++ b/block/blk-iolatency.c
> @@ -216,6 +216,31 @@ static inline void latency_stat_sum(struct iolatency_grp *iolat,
> blk_rq_stat_sum(&sum->rqs, &stat->rqs);
> }
>
> +/*
> + * Merge an already-aggregated @stat into @sum by reconstructed totals
> + * (mean * nr_samples); blk_rq_stat_sum() reads src->batch, which is 0
> + * here. The overflow check also rejects the 0 samples case.
> + */
> +static inline void latency_stat_merge(struct iolatency_grp *iolat,
> + struct latency_stat *sum,
> + struct latency_stat *stat)
> +{
> + if (iolat->ssd) {
> + sum->ps.total += stat->ps.total;
> + sum->ps.missed += stat->ps.missed;
> + } else {
> + if (sum->rqs.nr_samples + stat->rqs.nr_samples <=
> + sum->rqs.nr_samples)
> + return;
> + sum->rqs.mean = div_u64(sum->rqs.mean * sum->rqs.nr_samples +
> + stat->rqs.mean * stat->rqs.nr_samples,
> + sum->rqs.nr_samples + stat->rqs.nr_samples);
> + sum->rqs.min = min(sum->rqs.min, stat->rqs.min);
> + sum->rqs.max = max(sum->rqs.max, stat->rqs.max);
> + sum->rqs.nr_samples += stat->rqs.nr_samples;
> + }
> +}
> +
> static inline void latency_stat_record_time(struct iolatency_grp *iolat,
> u64 req_time)
> {
> @@ -547,7 +572,7 @@ static void iolatency_check_latencies(struct iolatency_grp *iolat, u64 now)
> /* Somebody beat us to the punch, just bail. */
> spin_lock_irqsave(&lat_info->lock, flags);
>
> - latency_stat_sum(iolat, &iolat->cur_stat, &stat);
> + latency_stat_merge(iolat, &iolat->cur_stat, &stat);
> lat_info->nr_samples -= iolat->nr_samples;
> lat_info->nr_samples += latency_stat_samples(iolat, &iolat->cur_stat);
> iolat->nr_samples = latency_stat_samples(iolat, &iolat->cur_stat);
prev parent reply other threads:[~2026-11-20 2:19 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-09-04 6:54 [PATCH v3] block/blk-iolatency: fix always-zero cur_stat mean on non-SSD Tao Cui
2026-11-20 2:19 ` Guixin Liu [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=856407d2-db33-4c1d-a5c9-2ba7e0053ad4@linux.alibaba.com \
--to=kanie@linux.alibaba.com \
--cc=axboe@kernel.dk \
--cc=cgroups@vger.kernel.org \
--cc=cui.tao@linux.dev \
--cc=cuitao@kylinos.cn \
--cc=josef@toxicpanda.com \
--cc=linux-block@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=tangyeechou@gmail.com \
--cc=tj@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox