v4l2: apply the control values userspace set before STREAMON

[vrilutza]

fthd_start_channel() forces brightness and contrast to 0x80, so every VIDIOC_STREAMON throws away whatever the application had set. VIDIOC_G_CTRL keeps reporting the value that was asked for, which is what makes it hard to see: the control reads back correctly while the ISP runs with the default.

Drop the two hardcoded commands and call v4l2_ctrl_handler_setup() once the channel is up. That covers saturation, hue and auto white balance too, which had the same problem.

Measured on a MacBookPro14,1 against a module built from clean master. One open(), four frames from the same file descriptor, 20 frames dropped between each so the exposure settles. Mean luma of the Y plane at 640x360:

brightness      set when           master r1   master r2   patched
40              before STREAMON        37.38       34.36      0.02
210             while streaming       119.36      116.77    116.37
40              while streaming         0.01        0.01      0.01
128 (default)   while streaming        37.46       33.61     34.29

Before the change, brightness 40 set before STREAMON gives the same picture as the default — not as 40. The third row is the negative control taken in the same run: the setting does work, it is only the pre-STREAMON value that is lost. The two master columns are the same module measured twice, interleaved with the patched one; absolute values move with the ambient light, the relation does not.

After the change, the first row matches the third rather than the fourth.

---

The series

Seven PRs on this driver, all on master (364b1c6), all found while debugging a camera problem on a MacBookPro14,1:

#328	two small cleanups: missing break, hardcoded buffer count
#329 (this one)	control values discarded at every VIDIOC_STREAMON
#330	cmd.y1 never assigned in the crop command — one line
#331	ALIGN(width, 7) aligns nothing; second commit re-does 545cb18 and should wait
#332	one firmware timeout leaves the driver with no usable buffers
#333	the camera writes in front of any buffer that is not page aligned
#334	the auto-exposure wait: 1000 ms → 200 ms

They can be taken in any order, with one exception: the second commit of #331 should not be taken yet — reasons in that PR.

(Until 23 August there was a second exception: #328 and #334 both carried the same mdelay → msleep commit, so whichever merged first made the other conflict. I have dropped it from #328; it lives in #334. The seven are now independent of each other.)

There is one more patch not sent yet: the channel-start crop returns the top left corner of the frame at any size below the sensor. It is measured and ready, and it sits on top of #330. I am holding it back rather than adding an eighth PR to the pile.

[patjak]

There is no need to explain what the code used to do here. Just removing the brightness and contrast commands is enough.