mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* [PATCH v3 0/2] media: verisilicon: Fix VDPU2 decoder error handling
@ 2026-11-21  6:30 Arun Iyengar via B4 Relay
  2026-11-21  6:30 ` [PATCH v3 1/2] media: verisilicon: Don't report VDPU2 decoder errors as success Arun Iyengar via B4 Relay
  2026-11-21  6:30 ` [PATCH v3 2/2] media: verisilicon: Handle VDPU2 decoder interrupts like the G2 Arun Iyengar via B4 Relay
  0 siblings, 2 replies; 3+ messages in thread
From: Arun Iyengar via B4 Relay @ 2026-11-21  6:30 UTC (permalink / raw)
  To: Nicolas Dufresne, Benjamin Gaignard, Philipp Zabel,
	Mauro Carvalho Chehab, Heiko Stuebner, Jonas Karlman,
	Boris Brezillon, Hans Verkuil
  Cc: linux-media, linux-rockchip, linux-arm-kernel, linux-kernel,
	Arun Iyengar, stable

The VDPU2 decoder reports failed decodes as done, and its interrupt
handler can finish a job the decoder is still working on.

Tested on Orange Pi 5 (RK3588S), v7.3-based KASAN kernel. v2: 72 H.264
fluster suite runs, 50 error-stream runs; no IOMMU faults or timeouts.
v3: one suite run and VP8 61/61, same results. The decoder was still
enabled after both buffer-empty interrupts of that run; a later
interrupt ended the job.
With a TEST-only patch (not in this series), the IRQ_NONE path ran
200 times on real registers; bus error and timeout, simulated 5 times
each by injected status bits, were logged and ended their jobs at once.
VP8 stayed 61/61 bit-exact.
A TEST-only injection of a decoder stopping right after the status clear:
with this order each re-raised interrupt ended its own job (bit-exact);
with the v2 order it ended the next job early.
Build-tested on media next. MPEG-2, RK3399 and RK3568 untested.

Developed with Claude (Anthropic), which wrote the code and the commit
messages. I reviewed every change.

Signed-off-by: Arun Iyengar <aiyengar1990@gmail.com>
---
Changes in v3:
- Patch 2: read DEC_E before clearing the status, as G2 does (Sashiko)
- Link to v2: https://patch.msgid.link/20261120-vdpu2-errreset-v2-0-8f4ef55e6dee@gmail.com

Changes in v2:
- Patch 2: no reset after an error; handle the interrupt like the G2
  handler instead (Nicolas)
- Link to v1: https://patch.msgid.link/20261119-vdpu2-errreset-v1-0-f8fa403b6a3a@gmail.com

---
Arun Iyengar (2):
      media: verisilicon: Don't report VDPU2 decoder errors as success
      media: verisilicon: Handle VDPU2 decoder interrupts like the G2

 .../media/platform/verisilicon/rockchip_vpu_hw.c   | 24 +++++++++++++++++++---
 1 file changed, 21 insertions(+), 3 deletions(-)
---
base-commit: aa0f4c41d4870a18e65772a1817af16427200371
change-id: 20261115-vdpu2-errreset-8688700393a3

Best regards,
--  
Arun Iyengar <aiyengar1990@gmail.com>



^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2026-11-21  6:30 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-11-21  6:30 [PATCH v3 0/2] media: verisilicon: Fix VDPU2 decoder error handling Arun Iyengar via B4 Relay
2026-11-21  6:30 ` [PATCH v3 1/2] media: verisilicon: Don't report VDPU2 decoder errors as success Arun Iyengar via B4 Relay
2026-11-21  6:30 ` [PATCH v3 2/2] media: verisilicon: Handle VDPU2 decoder interrupts like the G2 Arun Iyengar via B4 Relay

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®