Re: [PATCH v3] media: i2c: ov13858: add regulator, clock and reset GPIO handling

[Sergey Lebedev <lsa.uz@pm.me>]

New evidence for this patch, from media/next rather than from a distro kernel.

I built and booted a kernel from media/next f9536a8065 on the Surface Pro 11
today - for an unrelated series - and the back camera does not come up on it:

  ov13858 i2c-OVTID858:00: failed to find sensor: -5
  ov13858 i2c-OVTID858:00: probe with driver ov13858 failed with error -5

That is the mechanism Hans named on 2026-08-31 - "the ov13858 problem is not a
problem with the INT3472 driver, but rather with the ov13858 driver" - now
visible on current upstream. Reading f9536a8065: ov13858.c contains no
regulator, no supply and no gpiod reference at all, and while probe does
acquire a clock through devm_v4l2_sensor_clk_get() it never enables it. So
nothing asks INT3472 to power the sensor, and the chip-ID read talks to a part
that is still off.

  https://lore.kernel.org/all/df2f3cb1-1bab-432f-a359-59e6bec9580a@kernel.org/

With this patch applied the same machine enumerates the camera and streams from
it; that has been true since 2026-08-31 and is what the original posting
reported. What is new is only that the failure is now demonstrated on the tree
this patch is aimed at, rather than argued from a 7.0.0 kernel.

Both ov13858 patches still apply to f9536a8065 unchanged - checked, not assumed
- so there is nothing to resend and I am not asking for anything here. Recording
it because the next person to read this thread should not have to take the
"camera still dead" claim on trust.

Sergey