It's not uncommon for the particular device to support only a subset of
HDMI InfoFrames. Currently it's mostly ignored by the framework: it
calls write_infoframe() / clear_infoframe() callbacks for all frames and
expects them to return success even if the InfoFrame is not supported.

Sort that out, making sure that all interfaces are consistent:
- split function interfaces, having a pair of callbacks per each
  InfoFrame type.
- write_infoframe() / clear_infoframe() functions return -EOPNOTSUPP
  for unsupported InfoFrames.
- don't create and enable unsupported InfoFrames.
- limit debugfs to provide supported InfoFrames only.

Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
---
Changes in v4:
- Rebased on top of drm-misc-next, moving inno-hdmi to the bridge patch
- Expanded "normal" test description (Maxime)
- Added EDEADLK handling to the tests (Maxime)
- Moved hdmi_update_failures to drm_atomic_helper_connector_hdmi_priv
  (Maxime)
- Link to v3: https://lore.kernel.org/r/20251224-limit-infoframes-2-v3-0-7fd3cacfefed@oss.qualcomm.com

Changes in v3:
- Added DRM KUnit tests verifying InfoFrames behaviour (Maxime)
- Reworked the patchset, having per-infoframe callbacks rather than
  extra flags (Maxime)
- Link to v2: https://lore.kernel.org/r/20250928-limit-infoframes-2-v2-0-6f8f5fd04214@oss.qualcomm.com

Changes in v2:
- Fixed build issue in common code
- Fixed comments regarding HDR / audio frames (Diedrik)
- In adv7511 actually check for the HDR frame (the comment was correct)
  rather than the audio infoframe (c&p error).
- Link to v1: https://lore.kernel.org/r/20250927-limit-infoframes-2-v1-0-697511bd050b@oss.qualcomm.com

---
Dmitry Baryshkov (10):
      drm/tests: hdmi: check the infoframes behaviour
      drm/vc4: hdmi: implement clear_infoframe
      drm/sun4i: hdmi_enc: implement clear_infoframe stub
      drm/connector: make clear_infoframe callback mandatory for HDMI connectors
      drm/bridge: refactor HDMI InfoFrame callbacks
      drm/display: hdmi_state_helper: split InfoFrame functions per type
      drm/display: hdmi_state_helper: reject Audio IF updates if it's not supported
      drm/display: hdmi_state_helper: don't generate unsupported InfoFrames
      drm/display: bridge_connector: dynamically generate HDMI callbacks
      drm/debug: don't register files for unsupported HDMI InfoFrames

 drivers/gpu/drm/bridge/adv7511/adv7511_drv.c       | 180 +++---
 drivers/gpu/drm/bridge/inno-hdmi.c                 |  41 +-
 drivers/gpu/drm/bridge/ite-it6263.c                |  95 +--
 drivers/gpu/drm/bridge/lontium-lt9611.c            | 143 +++--
 drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c       | 110 ++--
 drivers/gpu/drm/display/drm_bridge_connector.c     | 190 +++++-
 drivers/gpu/drm/display/drm_hdmi_state_helper.c    |  94 +--
 drivers/gpu/drm/drm_connector.c                    |   6 +
 drivers/gpu/drm/drm_debugfs.c                      |   7 +
 drivers/gpu/drm/mediatek/mtk_hdmi_common.c         |   8 +-
 drivers/gpu/drm/mediatek/mtk_hdmi_v2.c             | 110 ++--
 drivers/gpu/drm/msm/hdmi/hdmi_bridge.c             | 195 +++---
 drivers/gpu/drm/rockchip/rk3066_hdmi.c             |  47 +-
 drivers/gpu/drm/sun4i/sun4i_hdmi_enc.c             |  42 +-
 drivers/gpu/drm/tests/drm_client_modeset_test.c    |   3 +
 drivers/gpu/drm/tests/drm_connector_test.c         |  19 +
 drivers/gpu/drm/tests/drm_hdmi_state_helper_test.c | 667 +++++++++++++++++++++
 drivers/gpu/drm/tests/drm_kunit_edid.h             | 119 ++++
 drivers/gpu/drm/vc4/vc4_hdmi.c                     | 105 +++-
 include/drm/drm_bridge.h                           | 127 +++-
 include/drm/drm_connector.h                        | 105 +++-
 21 files changed, 1900 insertions(+), 513 deletions(-)
---
base-commit: 349d4efadc1f831ebc0b872ba1e3a2b7dd58b72b
change-id: 20250927-limit-infoframes-2-6b93e599e79a

Best regards,
-- 
With best wishes
Dmitry

On Wed, 07 Jan 2026 20:14:57 +0200, Dmitry Baryshkov wrote:
> It's not uncommon for the particular device to support only a subset of
> HDMI InfoFrames. Currently it's mostly ignored by the framework: it
> calls write_infoframe() / clear_infoframe() callbacks for all frames and
> expects them to return success even if the InfoFrame is not supported.
> 
> Sort that out, making sure that all interfaces are consistent:
> - split function interfaces, having a pair of callbacks per each
>   InfoFrame type.
> - write_infoframe() / clear_infoframe() functions return -EOPNOTSUPP
>   for unsupported InfoFrames.
> - don't create and enable unsupported InfoFrames.
> - limit debugfs to provide supported InfoFrames only.
> 
> [...]

Applied to drm-misc-next, thanks!

[01/10] drm/tests: hdmi: check the infoframes behaviour
        commit: 7436a87db99d57196c49d10de35f41531993d5f1
[02/10] drm/vc4: hdmi: implement clear_infoframe
        commit: 3a4ceb89a9723088615ea1c960fc589f87caddb7
[03/10] drm/sun4i: hdmi_enc: implement clear_infoframe stub
        commit: 638409979c5f7d3155afcded67532003e07a7d0e
[04/10] drm/connector: make clear_infoframe callback mandatory for HDMI connectors
        commit: afc399f7a5ea7bf405b2ef85c7470529b1a9e47c
[05/10] drm/bridge: refactor HDMI InfoFrame callbacks
        commit: b626b1a1c9ccadd8861870a2a450f02e0c61ab88
[06/10] drm/display: hdmi_state_helper: split InfoFrame functions per type
        commit: e802c783be94bf71541a7e2ac8b1b5486aad10db
[07/10] drm/display: hdmi_state_helper: reject Audio IF updates if it's not supported
        commit: 1d8847f457648ed4932019dcd3081bc27bcea936
[08/10] drm/display: hdmi_state_helper: don't generate unsupported InfoFrames
        commit: 4fc30c2c5c61db88e3d6644bd3dd2032ec4bee06
[09/10] drm/display: bridge_connector: dynamically generate HDMI callbacks
        commit: ae219fdc952c315182b471f5aa71b379584b70ca
[10/10] drm/debug: don't register files for unsupported HDMI InfoFrames
        commit: 5a4e4e30f6dc4d2a68eec08257128906572f3346

Best regards,
-- 
With best wishes
Dmitry