Re: [PATCH v4 0/6] media: Add the ST VD55G0 global shutter sensor

[Emfox Zhou <emfox@debian.org>]

Hi Sergey,

A Surface Pro 8 (Tiger Lake, IPU6) is also SMO55F0:

    _SB.PC00.I2C3.CAM3
    I2cSerialBusV2 0x60, 400 kHz
    _DEP ICL2

I built 7.4-rc2 with these patches applied and have it running on this machine
now, so the notes below are measurements rather than an ACPI reading. Its ICL2
reports four lines:

    pin 0x83  type 0x0B  POWER_ENABLE
    pin 0x0A  type 0x10  DOVDD
    pin 0x4A  type 0x00  RESET
    pin 0x90  type 0x0D  PRIVACY_LED

(_DSM's pin field for the DOVDD entry says 0x0F; the _CRS pin table says 0x0A,
and 0x0A is what the driver maps and what is live in the running system.)

3/6 covers it as written. vd55g0 probes on 3-0060, and the two supplies it maps
are enabled while streaming, vddio and vana both at use count 1, on a device
that has four lines rather than five.

6/6 needs one more entry, and gpios = { 1 } is right for this machine. With
DMI_PRODUCT_NAME "Surface Pro 8" added to sensor_leds_dmi_ids[] the flash
control appears and the illuminator strobes: at a fixed exposure (600, AE off)
the Y8 frames average 15.8 with led_mode 0 and 40.1 with led_mode 1, peak values
17 and 75, unchanged over three runs. Without an entry the driver does not
register V4L2_CID_FLASH_LED_MODE at all, because nb_leds comes only from st,leds
and nothing else on this machine provides it.

The clock wants 0 here too. The SSDB asks for 24 MHz, INT3472:02-clk reports
rate 24000000 with prepare count 1 while streaming, and the stream runs with it.

One caveat that has nothing to do with these patches: on this machine frames
only arrive after working around a 1-lane receive FIFO problem in ipu6-isys (the
i-wake watermark). I am reporting that separately; it is not the sensor.

I am happy to give a Tested-by for 3/6, 4/6 and 6/6, the last one with the DMI
entry added, if that helps.

Thanks,
Emfox Zhou