media: i2c: imx477: Add vsync trigger_mode parameter

[neocortex-vision]

trigger_mode == 0 (default) => no effect / no registers written
trigger_mode == 1 => source
trigger_mode == 2 => sink

This can be set e.g. in /boot/cmdline.txt as imx477.trigger_mode=N
Setting it to 0 or not setting it at all has the same effect.

Tested on HQ cameras (XVS <-> XVS, GND <-> GND) with libcamera.

Credits to 6by9.

[6by9]

The content looks OK, although we get a couple of warnings from scripts/checkpatch.pl --strict --codespell

No codespell typos will be found - file '/usr/share/codespell/dictionary.txt': No such file or directory
ERROR: do not initialise statics to 0
#25: FILE: drivers/media/i2c/imx477.c:28:
+static int trigger_mode = 0;

CHECK: Alignment should match open parenthesis
#55: FILE: drivers/media/i2c/imx477.c:1738:
+		imx477_write_reg(imx477, IMX477_REG_MC_MODE,
+			IMX477_REG_VALUE_08BIT, 1);

CHECK: Alignment should match open parenthesis
#57: FILE: drivers/media/i2c/imx477.c:1740:
+		imx477_write_reg(imx477, IMX477_REG_MS_SEL,
+			IMX477_REG_VALUE_08BIT, val);

CHECK: Alignment should match open parenthesis
#59: FILE: drivers/media/i2c/imx477.c:1742:
+		imx477_write_reg(imx477, IMX477_REG_XVS_IO_CTRL,
+			IMX477_REG_VALUE_08BIT, val);

CHECK: Alignment should match open parenthesis
#61: FILE: drivers/media/i2c/imx477.c:1744:
+		imx477_write_reg(imx477, IMX477_REG_EXTOUT_EN,
+			IMX477_REG_VALUE_08BIT, val);

ERROR: Missing Signed-off-by: line(s)

total: 2 errors, 0 warnings, 4 checks, 43 lines checked

As currently imx477.c is totally clean of checkpatch warnings, it would be nice to fix all of those up, even though --strict is not strictly required (those on linux-media seem to expect it though).

"Signed-off By" is required as we are going to upstream this driver, so all contributions need to be appropriately tagged - see https://www.kernel.org/doc/html/latest/process/submitting-patches.html#sign-your-work-the-developer-s-certificate-of-origin. Please note "sorry, no pseudonyms or anonymous contributions".

@naushir We were talking about this change being needed earlier in the week - one fewer job for you then.

[naushir]

Once the warning are cleared, LGTM.

[neocortex-vision]

Thanks, done. Regarding initialisation, default (ineffective) value is now -1.

[6by9]

Thanks for the update.

Could you also squash your commits together. We apply them exactly as presented in the PR, so don't need your development history.

[6by9]

It was fine using a value of 0, but a static variable is automatically zeroed and so didn't need it explicitly.

static int trigger_mode;
would have been fine.

[pelwell]

If the aim is to squash everything into one commit, especially if the commit message on the first commit doesn't require modification, then we can squash during the merge, but for more complicated combinations it's necessary for the submitter to squash before pushing.

[6by9]

Upstream may point out that
if (trigger_mode != 0)
can be simplified to
if (trigger_mode)
but otherwise that looks fine to me. Thanks for your efforts.

[neocortex-vision]

Thanks for the feedback and excuse the explicit initialisation/checking (it's habitual, coming from a MISRA background).
I was planning to add hsync, vsync and frstobe signal parameters but I see this may have already been planned @naushir.