Re: [PATCH 4/9] media: camss: csid: add the 980 generation

[Shuai Liu]

Hi Qianhui,

The csid-980 register map and RDI configuration itself look sane to me
(the vc=0 stream programming matches what gen2/gen3 already do, and the
split RUP/AUP handling for the full variant is clearly commented).

> Only the register map, the RDI stream configuration and the
> buffer-done reporting are described here; frame start reporting is
> added on top of it by a separate change.

The commit message does not match the patch content, though.  As
posted, this patch touches six files, not three:

1. drivers/media/platform/qcom/camss/camss-csiphy-3ph-1-0.c is modified
   to flip the four SKEW_CAL data bytes (0x005c and per-lane copies)
   from 0x54 to 0x00.  That is a correction to patch 2/9 and belongs
   there - patch 2 should simply program the intended value, otherwise
   the series carries a value its own commit message calls wrong.

2. camss-vfe.c and camss-vfe.h carry groundwork for patches 5/9 and
   8/9, none of it described by this commit message:

   - the v4l2-event.h include and the hoisting of `line` plus the new
     `state` variable in vfe_buf_done() (needed by 5/9);
   - the new reg_update_after_csid_config flag in struct
     vfe_subdev_resources and the corresponding gating of the
     reg_update() call in vfe_enable_output_v2();
   - the CAMSS_8750 cases in vfe_src_pad_code() and
     vfe_bpl_align_rdi(), and the extern declaration of
     vfe_ops_gen4 (all needed by 8/9).

Please fold the csiphy fix into patch 2, the vfe_buf_done/event
groundwork into patch 5, and the gen4 enablement into patch 8 (or
describe the split here if you prefer to keep it), so that each patch
stands on its own.

Thanks,
Shuai Liu

(This review was prepared with the assistance of an AI coding tool;
all findings were verified against the patched source.)