From: Claude Code Review Bot <claude-review@example.com>
To: dri-devel-reviews@example.com
Subject: Claude review: drm/v3d: Ensure atomic submissions in v3d_submit_jobs()
Date: Thu, 04 Jun 2026 11:24:43 +1000 [thread overview]
Message-ID: <review-patch14-20260603-v3d-sched-misc-fixes-v3-14-d7114bba55a0@igalia.com> (raw)
In-Reply-To: <20260603-v3d-sched-misc-fixes-v3-14-d7114bba55a0@igalia.com>
Patch Review
This is the capstone fix. The submission is restructured into three phases:
1. **Arm** all jobs (get `done_fence`, take refcount)
2. **Wire** inter-job dependencies (can fail)
3. **Push** all jobs to scheduler (unconditional)
If phase 2 fails, all armed jobs have their finished fence marked with an error:
```c
for (int i = 0; i < submit->job_count; i++)
dma_fence_set_error(&submit->jobs[i]->base.s_fence->finished, ret);
```
Then all jobs are pushed anyway, relying on the `run_job()` callbacks' existing `s_fence->finished.error` checks to skip execution.
The patch also adds the error check to `v3d_cpu_job_run()` and `v3d_cache_clean_job_run()`, which previously lacked them. BIN, RENDER, TFU, and CSD already had these checks (added/consolidated in patch 2).
The error path in `v3d_submit_jobs()` now handles reservation unlock and job put internally, so callers don't need `fail_unreserve` labels. The callers' error paths are updated accordingly with a new `fail_submit` label that skips job cleanup (since `v3d_submit_jobs` already did it on error).
**One concern with the error path**: On failure, `v3d_submit_jobs()` calls `v3d_submit_unlock_reservations(submit)` and `v3d_submit_put_jobs(submit)`, but does NOT call `v3d_submit_attach_object_fences()` or `v3d_submit_process_post_deps()`. This is correct -- on error, we don't want to attach fences to BOs or signal syncobjs to userspace. The jobs are pushed with error-marked fences, the scheduler will run them (no-op via the error check), and they'll be freed by the scheduler completion callback.
**Also correct**: The caller's `fail_submit` path still calls `v3d_submit_put_post_deps(sync_out, &se)` to release the sync_out reference and any multi-sync output references, which is the right thing since those were never consumed.
No issues. Well-designed error handling.
---
**Summary**: The series is well-decomposed and addresses real issues. The one actionable finding is:
1. **Patch 8**: `v3d_submit_add_job()` should check `submit->job_count < V3D_MAX_JOBS_PER_SUBMISSION` before writing to `submit->jobs[]` to prevent a potential out-of-bounds array write.
---
Generated by Claude Code Patch Reviewer
next prev parent reply other threads:[~2026-06-04 1:24 UTC|newest]
Thread overview: 31+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-06-03 22:25 [PATCH v3 00/14] drm/v3d: Scheduler and submission fixes and refactoring Maíra Canal
2026-06-03 22:25 ` [PATCH v3 01/14] drm/v3d: Drop unused drm_encoder.h include from v3d_drv.h Maíra Canal
2026-06-04 1:24 ` Claude review: " Claude Code Review Bot
2026-06-03 22:25 ` [PATCH v3 02/14] drm/v3d: Clear queue->active_job when v3d_fence_create() fails Maíra Canal
2026-06-04 1:24 ` Claude review: " Claude Code Review Bot
2026-06-03 22:25 ` [PATCH v3 03/14] drm/v3d: Use inline lock for dma fence initialization Maíra Canal
2026-06-04 1:24 ` Claude review: " Claude Code Review Bot
2026-06-03 22:25 ` [PATCH v3 04/14] drm/v3d: Replace spin_lock_irqsave() with spin_lock() Maíra Canal
2026-06-04 1:24 ` Claude review: " Claude Code Review Bot
2026-06-03 22:25 ` [PATCH v3 05/14] drm/v3d: Extract v3d_job_add_syncobjs() helper Maíra Canal
2026-06-04 1:24 ` Claude review: " Claude Code Review Bot
2026-06-03 22:25 ` [PATCH v3 06/14] drm/v3d: Reject invalid syncobj handles in submit ioctls Maíra Canal
2026-06-04 1:24 ` Claude review: " Claude Code Review Bot
2026-06-03 22:25 ` [PATCH v3 07/14] drm/v3d: Migrate BO reservation locking to DRM exec Maíra Canal
2026-06-04 1:24 ` Claude review: " Claude Code Review Bot
2026-06-03 22:25 ` [PATCH v3 08/14] drm/v3d: Introduce struct v3d_submit and convert CL/TFU/CSD ioctls Maíra Canal
2026-06-04 1:24 ` Claude review: " Claude Code Review Bot
2026-06-03 22:25 ` [PATCH v3 09/14] drm/v3d: Make v3d_get_cpu_indirect_csd_params() a pure parser Maíra Canal
2026-06-04 1:24 ` Claude review: " Claude Code Review Bot
2026-06-03 22:25 ` [PATCH v3 10/14] drm/v3d: Convert submit helpers to operate on struct v3d_submit Maíra Canal
2026-06-04 1:24 ` Claude review: " Claude Code Review Bot
2026-06-03 22:25 ` [PATCH v3 11/14] drm/v3d: Refactor CPU ioctl into unified submission chain Maíra Canal
2026-06-04 1:24 ` Claude review: " Claude Code Review Bot
2026-06-03 22:25 ` [PATCH v3 12/14] drm/v3d: Split BO fence attach from syncobj output handling Maíra Canal
2026-06-04 1:24 ` Claude review: " Claude Code Review Bot
2026-06-03 22:25 ` [PATCH v3 13/14] drm/v3d: Reject invalid out_sync handles in submit ioctls Maíra Canal
2026-06-04 1:24 ` Claude review: " Claude Code Review Bot
2026-06-03 22:25 ` [PATCH v3 14/14] drm/v3d: Ensure atomic submissions in v3d_submit_jobs() Maíra Canal
2026-06-04 1:24 ` Claude Code Review Bot [this message]
2026-06-04 1:24 ` Claude review: drm/v3d: Scheduler and submission fixes and refactoring Claude Code Review Bot
-- strict thread matches above, loose matches on Subject: below --
2026-05-10 22:11 [PATCH v2 00/14] " Maíra Canal
2026-05-10 22:12 ` [PATCH v2 14/14] drm/v3d: Ensure atomic submissions in v3d_submit_jobs() Maíra Canal
2026-05-16 5:59 ` 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-patch14-20260603-v3d-sched-misc-fixes-v3-14-d7114bba55a0@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