CODA timeout fix & assorted changes
Get series with b4
b4 am [email protected]Versions newest first · linked by author and title
- v2 Latest6 patches 8 repliesIn review
- v16 patches 6 repliesSuperseded
Changes in v2 from the cover letter
* Keep the error MB message, but move it to coda_dbg(1, ctx).
* Add per-device rate limitting for the error MB message.
* Rename V4L2_CID_CODA_ERR_MB description.
* s/__coda_decoder_drop_used_metas/coda_decoder_drop_used_metas
Patches in v2 No applies recorded yet
| # | Subject and files | Size | Replies | Applied |
|---|---|---|---|---|
| 1/6 | coda: Remove redundant ctx->initialized settingdrivers/ | +2 | 0 replies | |
| 2/6 | coda: Simplify H.264 small buffer padding logicdrivers/ | +4 | 0 replies | |
| 3/6 | coda: Clarify device registered logdrivers/ | +5 | 0 replies | |
| 4/6 | coda: Clarify interrupt registered namedrivers/ | +2 | 0 replies | |
| 5/6 | coda: coda_buffer_meta housekeeping fixdrivers/ drivers/ | +39 | 0 replies | |
| 6/6 | coda: Add a V4L2 user for control error macroblocks countMAINTAINERS drivers/ drivers/ +3 more | +70 | 1 reply |
Reviews trailers per patch
5 of 6 patches have a Reviewed-by or Acked-by.
| Patch | Philipp Zabel |
|---|---|
| 1/6 | R-b |
| 2/6 | R-b |
| 3/6 | A-b |
| 4/6 | A-b |
| 5/6 | R-b |
| 6/6 |
Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes.
Cover letter [PATCH v2 0/6] CODA timeout fix & assorted changes
Hi all,
The main motivation for this series is to address a PIC_RUN timeout, which we managed to link with a hardware bitstream buffer underrun condition.
Upon further investigation we discovered that the underrun was produced by a subtle issue in the way buffer_meta's were being tracked.
The issue is fixed by patch "5/6 coda: coda_buffer_meta housekeeping fix".
Patches 1 to 4 are mostly cleanups and minor cosmetic changes.
Finally, while testing with corrupted bitstreams we realized the driver was logging too verbosely, so patch 6 addresses this by introducing a private control to read an macroblock-error counter.
These patches have been tested against media's upstream and v5.4-based, on i.MX6 (Wandboard) with H.264 and MPEG-2 streams.
In particular, this series manages to fix the PIC_RUN timeout reported by Benjamin Bara [1].
I believe this timeout can be systematically reproduced with a video containing small black frames, which can be generated with:
gst-launch-1.0 videotestsrc pattern=black num-buffers=300 ! \
video/x-raw,format=I420,width=176,height=120 ! avenc_mpeg2video ! \
mpegvideoparse ! mpegtsmux ! filesink location=black-qcif-10s.ts
Reviews and feedback are appreciated, as always.
[1] https://lkml.org/lkml/2020/8/21/495
Changelog
---------
v2:
* Keep the error MB message, but move it to coda_dbg(1, ctx).
* Add per-device rate limitting for the error MB message.
* Rename V4L2_CID_CODA_ERR_MB description.
* s/__coda_decoder_drop_used_metas/coda_decoder_drop_used_metas
Ezequiel Garcia (6):
coda: Remove redundant ctx->initialized setting
coda: Simplify H.264 small buffer padding logic
coda: Clarify device registered log
coda: Clarify interrupt registered name
coda: coda_buffer_meta housekeeping fix
coda: Add a V4L2 user for control error macroblocks count
MAINTAINERS | 1 + drivers/media/platform/coda/coda-bit.c | 72 ++++++++++++++++------- drivers/media/platform/coda/coda-common.c | 57 +++++++++++++++--- drivers/media/platform/coda/coda.h | 5 ++ include/media/drv-intf/coda.h | 13 ++++ include/uapi/linux/v4l2-controls.h | 4 ++ 6 files changed, 122 insertions(+), 30 deletions(-) create mode 100644 include/media/drv-intf/coda.h
Signature · 1 line
2.27.0