[PATCH v6 0/9] Fix DSI host idx detection on HW revision clash

[Konrad Dybcio]

v5 -> v6:
- Squash both fixes that concerned the deprecated QCM2290 compatible to
  avoid warnings

v5: https://lore.kernel.org/r/20230307-topic-dsi_qcm-v5-0-9d4235b77f4f@linaro.org

v4 -> v5:
- Drop superfluous items: level in [8/10]
- Remove the header define for the qcm2290 config in [6/10] instead of
  [7/10]
- Pick up tags

v4: https://lore.kernel.org/r/20230307-topic-dsi_qcm-v4-0-54b4898189cb@linaro.org

v3 -> v4:
- Use the shiny new compatible in the 6115 bindings example [9/10]
- Remove the leftover include and header definition [6, 7/10]
- Deduplicate the qcm2290 clks/regs in the common deduplication commit
  instead of doing it separately
- Pick up tags
- Rebase on next-20230314 (nothing seems to have changed fwiw)

v3: https://lore.kernel.org/r/20230307-topic-dsi_qcm-v3-0-8bd7e1add38a@linaro.org

v2 -> v3:
- Merge with [1], I should have done that earlier..
  - Squash 6115 compatible patches into one
- Pick up tags (except Rob's ack in 6115 compatible addition, as it was changed)
- Use b4 (sorry if you got an incomplete set of messages before..)

[1] https://lore.kernel.org/linux-arm-msm/145066db-5723-6baa-237d-7c2b8fd476d9@linaro.org/
v2: https://lore.kernel.org/linux-arm-msm/20230213121012.1768296-1-konrad.dybcio@linaro.org/

v1 -> v2:
- squash the 2d-array-ification and fixing up the logic into one patch
- drop num_variants, loop over VARIANTS_MAX*DSI_MAX unconditionally
- drop inadequate Fixes: tags
- pick up rbs

v1: https://lore.kernel.org/linux-arm-msm/20230211115110.1462920-1-konrad.dybcio@linaro.org/

Some DSI host versions are implemented on multiple SoCs which use
vastly different register maps. This messes with our current
assumptions of being able to map {dsi0, dsi1} to {reg0, reg1}.
Solve that by adding a way of specifying multiple sets of base
registers and try comparing them against the register specified in DT
until we find a match.

This removes the need for the QCM2290-specific compatible which was
used in the SM6115 DT (which uses DSIv2.4.1, just like SC7180).
The series also takes care of that.

Tested on SM6115P Lenovo Tab P11 and SM8350 PDX215

Signed-off-by: Konrad Dybcio <konrad.dybcio at linaro.org>
---
Konrad Dybcio (9):
      dt-bindings: display/msm: dsi-controller-main: Fix deprecated QCM2290 compatible
      drm/msm/dsi: Get rid of msm_dsi_config::num_dsi
      drm/msm/dsi: Fix DSI index detection when version clash occurs
      drm/msm/dsi: dsi_cfg: Deduplicate identical structs
      drm/msm/dsi: dsi_cfg: Merge SC7180 config into SDM845
      drm/msm/dsi: Switch the QCM2290-specific compatible to index autodetection
      drm/msm/dsi: Remove custom DSI config handling
      dt-bindings: display/msm: dsi-controller-main: Add SM6115
      arm64: dts: qcom: sm6115: Use the correct DSI compatible

 .../bindings/display/msm/dsi-controller-main.yaml  |   9 +-
 .../bindings/display/msm/qcom,sm6115-mdss.yaml     |  10 +-
 arch/arm64/boot/dts/qcom/sm6115.dtsi               |   2 +-
 drivers/gpu/drm/msm/dsi/dsi.c                      |   7 +-
 drivers/gpu/drm/msm/dsi/dsi_cfg.c                  | 161 ++++++++-------------
 drivers/gpu/drm/msm/dsi/dsi_cfg.h                  |   9 +-
 drivers/gpu/drm/msm/dsi/dsi_host.c                 |  14 +-
 7 files changed, 84 insertions(+), 128 deletions(-)
---
base-commit: 6f08c1de13a9403341c18b66638a05588b2663ce
change-id: 20230307-topic-dsi_qcm-5cd03c230f8f

Best regards,
-- 
Konrad Dybcio <konrad.dybcio at linaro.org>