Skip to content

kernel: Decode HEVC in hardware on the A9 and A10 iPads - #121

Merged
henrikklee merged 2 commits into
OrchardLinux:mainfrom
brkosebay:j71t/vxd-hevc
Nov 12, 2026
Merged

kernel: Decode HEVC in hardware on the A9 and A10 iPads#121
henrikklee merged 2 commits into
OrchardLinux:mainfrom
brkosebay:j71t/vxd-hevc

Conversation

@brkosebay

@brkosebay brkosebay commented Nov 12, 2026

Copy link
Copy Markdown
Contributor

Part 15 of 23 of the iPad 5 (J71t) series in this repository; merge after #120.

Problem

The VXD decoder only exposed H.264, so HEVC video was decoded on the CPU.

Change

Prepare the driver for a second coded format, add the HEVC command-stream builder and the stateless HEVC decoder, enable it on the S8000 and T8010 variants, and require a proven firmware stop before kexec. Adds the video_decode_hevc feature (working on J71t, untested on the other boards that have the driver, missing elsewhere).

Testing

  • Hardware: iPad 5 Wi-Fi (J71t). Hands-on testing on 2026-11-11 with an integrated build of this series, made before the final failure-path fixes, covered HEVC and H.264 decoding, bit-exact against software decoding.
  • 2026-11-12: HEVC decoding was tested by hand on J71t with the release build and works.
  • Host: the Linux test suite passes at this branch's head, and the kernel and m1n1 patch series apply on apple-16k and apple-4k. At the top of the series the applied kernel tree is identical to the one in the built release image.
  • Other boards that share these drivers or compatibles have not been tested.

Dependencies

Depends on #120.

This branch is stacked on the PRs it depends on and carries their commits; the change of this PR is its last commit, "media: apple-vxd: Stateless HEVC decoding on S8000 and T8010".

🤖 Generated with Claude Code

@brkosebay brkosebay mentioned this pull request Nov 12, 2026
Merged
@brkosebay @henrikklee
kernel: Decode HEVC in hardware on the A9 and A10 iPads
6d181d5 
The VXD video decoder of the S8000/S8003 (A9) and T8010 (A10) also
decodes HEVC (H.265). Add a stateless HEVC decoder beside the H.264
one, with the per-SoC values Apple's driver uses. H.264 stays the
first and default format; HEVC shows up as a second one.

From orchard PR OrchardLinux#121 (commit 4016f5e): patches 0008-0011 there, the
last two adapted to main, which does not carry the VXD sleep changes of
PR OrchardLinux#117. The kexec firmware-stop patch (0012 there) belongs to the warm
restart topic and is not part of this one.
@brkosebay @henrikklee
catalog: Add hardware HEVC decoding as a feature
88c3c83 
J71t decodes HEVC bit-exact against software decoding (tested by hand
on 2026-11-11 and 2026-11-12). The other iPads with the decoder enabled
(J71s, J72s, J72t, and the iPad 6 J71b/J72b) are untested; every other
board has no decoder node and is missing it.

From orchard PR OrchardLinux#121 (commit 4016f5e), device data part.
@henrikklee henrikklee changed the title media: apple-vxd: Stateless HEVC decoding on S8000 and T8010 kernel: Decode HEVC in hardware on the A9 and A10 iPads Nov 12, 2026
@henrikklee

henrikklee commented Nov 12, 2026

Copy link
Copy Markdown
Contributor

Landing this as your commits. Patches 0008-0011 are in, the last two adapted to main (it does not carry the VXD sleep changes of #117, so the S8000 variant has no sleep_stop). The kexec firmware-stop patch (0012) is not part of this topic; it goes with warm restart. The device data is yours unchanged: HEVC working on J71t, untested on the other iPads with the decoder, missing elsewhere. Your previous head is kept as tag contrib/brkosebay-j71t-vxd-hevc. Please check HEVC decoding on the next nightly.

@henrikklee
henrikklee merged commit 18f78d7 into OrchardLinux:main Nov 12, 2026
1 of 3 checks passed
brkosebay added a commit to brkosebay/hardware that referenced this pull request Nov 13, 2026
@brkosebay
Merge upstream main into j71t/psci
8c737f8 
Bring this branch onto current main. Main has since merged OrchardLinux#124, OrchardLinux#122,
OrchardLinux#118, OrchardLinux#121, OrchardLinux#110, OrchardLinux#127, OrchardLinux#101, OrchardLinux#126, OrchardLinux#108 and OrchardLinux#111, several in part or
reworked, and OrchardLinux#123 and OrchardLinux#128 were closed. Where main carries a topic,
main's version is kept; the left-out parts that warm restart and
suspend need come back in OrchardLinux#107 and OrchardLinux#117.
brkosebay added a commit to brkosebay/hardware that referenced this pull request Nov 13, 2026
@brkosebay
Merge the j71t/psci branch into j71t/pmu-rtc
72ca165 
Bring this branch onto current main. Main has since merged OrchardLinux#124, OrchardLinux#122,
OrchardLinux#118, OrchardLinux#121, OrchardLinux#110, OrchardLinux#127, OrchardLinux#101, OrchardLinux#126, OrchardLinux#108 and OrchardLinux#111, several in part or
reworked, and OrchardLinux#123 and OrchardLinux#128 were closed. Where main carries a topic,
main's version is kept; the left-out parts that warm restart and
suspend need come back in OrchardLinux#107 and OrchardLinux#117.

This branch's own patches in drivers/usb are renumbered to follow
main's.
brkosebay added a commit to brkosebay/hardware that referenced this pull request Nov 13, 2026
@brkosebay
Merge the j71t/pmu-rtc branch into j71t/warm-storage
bfb58aa 
Bring this branch onto current main. Main has since merged OrchardLinux#124, OrchardLinux#122,
OrchardLinux#118, OrchardLinux#121, OrchardLinux#110, OrchardLinux#127, OrchardLinux#101, OrchardLinux#126, OrchardLinux#108 and OrchardLinux#111, several in part or
reworked, and OrchardLinux#123 and OrchardLinux#128 were closed. Where main carries a topic,
main's version is kept; the left-out parts that warm restart and
suspend need come back in OrchardLinux#107 and OrchardLinux#117.

This branch's own patches in devices/s8000, drivers/base and drivers/usb
are renumbered to follow main's.
brkosebay added a commit to brkosebay/hardware that referenced this pull request Nov 13, 2026
@brkosebay
Merge the j71t/warm-storage branch into j71t/warm-peripherals
86c47ee 
Bring this branch onto current main. Main has since merged OrchardLinux#124, OrchardLinux#122,
OrchardLinux#118, OrchardLinux#121, OrchardLinux#110, OrchardLinux#127, OrchardLinux#101, OrchardLinux#126, OrchardLinux#108 and OrchardLinux#111, several in part or
reworked, and OrchardLinux#123 and OrchardLinux#128 were closed. Where main carries a topic,
main's version is kept; the left-out parts that warm restart and
suspend need come back in OrchardLinux#107 and OrchardLinux#117.

This branch's own patches in devices/s8000, drivers/base and drivers/usb
are renumbered to follow main's.
brkosebay added a commit to brkosebay/hardware that referenced this pull request Nov 13, 2026
@brkosebay
Merge the j71t/warm-peripherals branch into j71t/warm-kernel
2de73b0 
Bring this branch onto current main. Main has since merged OrchardLinux#124, OrchardLinux#122,
OrchardLinux#118, OrchardLinux#121, OrchardLinux#110, OrchardLinux#127, OrchardLinux#101, OrchardLinux#126, OrchardLinux#108 and OrchardLinux#111, several in part or
reworked, and OrchardLinux#123 and OrchardLinux#128 were closed. Where main carries a topic,
main's version is kept; the left-out parts that warm restart and
suspend need come back in OrchardLinux#107 and OrchardLinux#117.

This branch's own patches in devices/s8000, drivers/base,
drivers/pmdomain and drivers/usb are renumbered to follow main's.
brkosebay added a commit to brkosebay/hardware that referenced this pull request Nov 13, 2026
@brkosebay
Merge the j71t/warm-kernel branch into j71t/reset-guard
172c287 
Bring this branch onto current main. Main has since merged OrchardLinux#124, OrchardLinux#122,
OrchardLinux#118, OrchardLinux#121, OrchardLinux#110, OrchardLinux#127, OrchardLinux#101, OrchardLinux#126, OrchardLinux#108 and OrchardLinux#111, several in part or
reworked, and OrchardLinux#123 and OrchardLinux#128 were closed. Where main carries a topic,
main's version is kept; the left-out parts that warm restart and
suspend need come back in OrchardLinux#107 and OrchardLinux#117.

This branch's own patches in devices/s8000, drivers/base,
drivers/pmdomain and drivers/usb are renumbered to follow main's.
brkosebay added a commit to brkosebay/hardware that referenced this pull request Nov 13, 2026
@brkosebay
Merge the j71t/reset-guard branch into j71t/suspend
934d6c5 
Bring this branch onto current main. Main has since merged OrchardLinux#124, OrchardLinux#122,
OrchardLinux#118, OrchardLinux#121, OrchardLinux#110, OrchardLinux#127, OrchardLinux#101, OrchardLinux#126, OrchardLinux#108 and OrchardLinux#111, several in part or
reworked, and OrchardLinux#123 and OrchardLinux#128 were closed. Where main carries a topic,
main's version is kept; the left-out parts that warm restart and
suspend need come back in OrchardLinux#107 and OrchardLinux#117.

This branch's own patches in devices/s8000, devices/t8010, drivers/base,
drivers/media, drivers/pmdomain, drivers/power and drivers/usb are
renumbered to follow main's.

Main's versions apply first now, so 0011 in drivers/media keep their
change with a diff regenerated for the new context.
brkosebay added a commit to brkosebay/hardware that referenced this pull request Nov 13, 2026
@brkosebay
kernel: Require a proven VXD firmware stop before kexec
ce7d3b2 
From OrchardLinux#121, the patch main did not take with HEVC. It uses the VXD
sleep stop and sleep gate this branch adds, so it sits here rather
than with the earlier warm restart branch: a failed firmware stop keeps
the firmware, command buffers and power, and kexec takes the cold-reset
path before those pages become the next kernel's RAM.

Assisted-by: LLM
brkosebay added a commit to brkosebay/hardware that referenced this pull request Nov 13, 2026
@brkosebay
Merge the j71t/suspend branch into apple-g5p-render
e040874 
Bring this branch onto current main. Main has since merged OrchardLinux#124, OrchardLinux#122,
OrchardLinux#118, OrchardLinux#121, OrchardLinux#110, OrchardLinux#127, OrchardLinux#101, OrchardLinux#126, OrchardLinux#108 and OrchardLinux#111, several in part or
reworked, and OrchardLinux#123 and OrchardLinux#128 were closed. Where main carries a topic,
main's version is kept; the left-out parts that warm restart and
suspend need come back in OrchardLinux#107 and OrchardLinux#117.

This branch's own patches in devices/s8000, devices/s8003,
devices/t8010, drivers/base, drivers/media, drivers/pmdomain,
drivers/power and drivers/usb are renumbered to follow main's.

Main's versions apply first now, so 0011 in drivers/media keep their
change with a diff regenerated for the new context.
brkosebay added a commit to brkosebay/hardware that referenced this pull request Nov 13, 2026
@brkosebay
Merge the apple-g5p-render branch into j71t/g5p-pipeline
a8e4e7a 
Bring this branch onto current main. Main has since merged OrchardLinux#124, OrchardLinux#122,
OrchardLinux#118, OrchardLinux#121, OrchardLinux#110, OrchardLinux#127, OrchardLinux#101, OrchardLinux#126, OrchardLinux#108 and OrchardLinux#111, several in part or
reworked, and OrchardLinux#123 and OrchardLinux#128 were closed. Where main carries a topic,
main's version is kept; the left-out parts that warm restart and
suspend need come back in OrchardLinux#107 and OrchardLinux#117.

This branch's own patches in devices/s8000, devices/s8003,
devices/t8010, drivers/base, drivers/media, drivers/pmdomain,
drivers/power and drivers/usb are renumbered to follow main's.

Main's versions apply first now, so 0011 in drivers/media keep their
change with a diff regenerated for the new context.
brkosebay added a commit to brkosebay/hardware that referenced this pull request Nov 13, 2026
@brkosebay
Merge the j71t/g5p-pipeline branch into j71t/g5p-suspend
7a6370c 
Bring this branch onto current main. Main has since merged OrchardLinux#124, OrchardLinux#122,
OrchardLinux#118, OrchardLinux#121, OrchardLinux#110, OrchardLinux#127, OrchardLinux#101, OrchardLinux#126, OrchardLinux#108 and OrchardLinux#111, several in part or
reworked, and OrchardLinux#123 and OrchardLinux#128 were closed. Where main carries a topic,
main's version is kept; the left-out parts that warm restart and
suspend need come back in OrchardLinux#107 and OrchardLinux#117.

This branch's own patches in devices/s8000, devices/s8003,
devices/t8010, drivers/base, drivers/media, drivers/pmdomain,
drivers/power, drivers/soc-apple and drivers/usb are renumbered to
follow main's.

Main's versions apply first now, so 0011 in drivers/media keep their
change with a diff regenerated for the new context.
brkosebay added a commit to brkosebay/hardware that referenced this pull request Nov 13, 2026
@brkosebay
Merge the j71t/g5p-suspend branch into j71t/gpu-handover
f97a7d0 
Bring this branch onto current main. Main has since merged OrchardLinux#124, OrchardLinux#122,
OrchardLinux#118, OrchardLinux#121, OrchardLinux#110, OrchardLinux#127, OrchardLinux#101, OrchardLinux#126, OrchardLinux#108 and OrchardLinux#111, several in part or
reworked, and OrchardLinux#123 and OrchardLinux#128 were closed. Where main carries a topic,
main's version is kept; the left-out parts that warm restart and
suspend need come back in OrchardLinux#107 and OrchardLinux#117.

This branch's own patches in devices/s8000, devices/s8003,
devices/t8010, drivers/base, drivers/media, drivers/pmdomain,
drivers/power, drivers/soc-apple and drivers/usb are renumbered to
follow main's.

Main's versions apply first now, so 0011 in drivers/media keep their
change with a diff regenerated for the new context.
This was referenced Nov 13, 2026
Open
Open
@brkosebay

brkosebay commented Nov 13, 2026

Copy link
Copy Markdown
Contributor Author

Checked on an iPad 5 (J71t) on 2026-11-13 with a user install image built from main 8ea3182 plus the open 0.4 PRs (top j71t/gpu-handover), installed with orchard install from DFU, then a cold boot from DFU and a warm restart (kexec) from the desktop.

Confirmed: HEVC decodes on the hardware. 4/4 HEVC streams (IPB 176p, 352p with dependent slices, 64p, 1080p) match the software decode frame for frame (framemd5), through ffmpeg's v4l2-request hwaccel. The warm-restart decode was not run separately.

Side note, not this PR: 1080p buffers can fail with ENOMEM while Firefox is open, because the CMA area fragments. With Firefox closed they decode bit-exact.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@brkosebay @henrikklee