From: Claude Code Review Bot <claude-review@example.com>
To: dri-devel-reviews@example.com
Subject: Claude review: drm/v3d: Hold v3d_stats references in each job
Date: Mon, 09 Mar 2026 09:20:15 +1000 [thread overview]
Message-ID: <review-patch4-20260305-v3d-reset-locking-improv-v2-4-fd53c91f7726@igalia.com> (raw)
In-Reply-To: <20260305-v3d-reset-locking-improv-v2-4-fd53c91f7726@igalia.com>
Patch Review
This is the key patch that eliminates the need for `queue_lock` in the stats update path. Each job now holds its own refcounted references to both client and global stats:
```c
job->client_stats = v3d_stats_get(v3d_priv->stats[queue]);
job->global_stats = v3d_stats_get(v3d->queue[queue].stats);
```
The critical simplification in `v3d_job_update_stats` is correct — the old code needed `queue_lock` and the `if (job->file_priv)` NULL check because stats were embedded in `v3d_file_priv` which could be freed. With refcounting, the stats outlive the fd:
```c
void
v3d_job_update_stats(struct v3d_job *job)
{
u64 now = local_clock();
preempt_disable();
v3d_stats_update(job->client_stats, now);
v3d_stats_update(job->global_stats, now);
preempt_enable();
}
```
The refactoring of `v3d_job_start_stats` into `v3d_stats_start` + `v3d_job_start_stats` is clean and makes both functions independent of queue type, which is why the `enum v3d_queue` parameter can be dropped. All callers are updated consistently.
The `v3d_job_free` cleanup correctly puts both stats references before freeing the job. The placement after `v3d_perfmon_put` and before `kfree` is appropriate.
---
Generated by Claude Code Patch Reviewer
next prev parent reply other threads:[~2026-03-08 23:20 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-03-05 23:25 [PATCH v2 0/6] drm/v3d: Improve v3d_stats lifetime and locking Maíra Canal
2026-03-05 23:25 ` [PATCH v2 1/6] drm/v3d: Handle error from drm_sched_entity_init() Maíra Canal
2026-03-08 23:20 ` Claude review: " Claude Code Review Bot
2026-03-05 23:25 ` [PATCH v2 2/6] drm/v3d: Use raw seqcount helpers instead of fighting with lockdep Maíra Canal
2026-03-08 23:20 ` Claude review: " Claude Code Review Bot
2026-03-05 23:25 ` [PATCH v2 3/6] drm/v3d: Refcount v3d_stats Maíra Canal
2026-03-08 23:20 ` Claude review: " Claude Code Review Bot
2026-03-05 23:25 ` [PATCH v2 4/6] drm/v3d: Hold v3d_stats references in each job Maíra Canal
2026-03-08 23:20 ` Claude Code Review Bot [this message]
2026-03-05 23:25 ` [PATCH v2 5/6] drm/v3d: Attach per-fd reset counters to v3d_stats Maíra Canal
2026-03-08 23:20 ` Claude review: " Claude Code Review Bot
2026-03-05 23:25 ` [PATCH v2 6/6] drm/v3d: Remove dedicated fence_lock Maíra Canal
2026-03-08 23:20 ` Claude review: " Claude Code Review Bot
2026-03-08 23:20 ` Claude review: drm/v3d: Improve v3d_stats lifetime and locking Claude Code Review Bot
-- strict thread matches above, loose matches on Subject: below --
2026-03-06 11:30 [PATCH v3 0/6] " Maíra Canal
2026-03-06 11:30 ` [PATCH v3 4/6] drm/v3d: Hold v3d_stats references in each job Maíra Canal
2026-03-08 22:56 ` Claude review: " Claude Code Review Bot
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=review-patch4-20260305-v3d-reset-locking-improv-v2-4-fd53c91f7726@igalia.com \
--to=claude-review@example.com \
--cc=dri-devel-reviews@example.com \
/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