gpio: stm32: support disabling and reenabling interrupts on gpio pin

[mniestroj]

Up to now interrupts could be only configured once, with no way to disable them in runtime.

Allow interrupts to be disabled in runtime and then properly reenabled on user request. This allows to ignore interrupts when software is not expecting them.

The improvement over previously reverted patch [1] is that we disable interrupts only when we configure port for which interrupt line was previously selected. This for example prevents to disable interrupts line 2 in case PA2 was previously configured as interrupt source, but we are currently configuring PB2 as output.

Besides we make a small improvement by declaring gpio_stm32_enable_int as static and splitting into smaller functions in order to achieve better readability.

Those patches are a result of previously noticed bug #19177.

[1] 0951ce2 ("gpio: stm32: support disabling and reenabling interrupts on pin")

[erwango]

@mniestroj, btw I didn't mentioned it previously but since this change is going to have much more impact than initially, you should be aware of imminent driver rework due to new GPIO API introduction, merged yesterday actually (#16648).
So please have a look and check if it fits your need.

[mniestroj]

So please have a look and check if it fits your need.

@erwango I don't see any problems, as those changes look unrelated. But thanks for pointing that, I'll try to follow it.

[mnkp]

Maybe it would be better if this PR targeted topic-gpio branch?

[erwango]

@mniestroj, this PR is of particular interest for driver adaptation for GPIO API.
I've reused it on #19607 (and hence validated it as part of my tetsing).
So to make things a little bit more simple, would be nice to have it on gpio-topic branch.
Also, as mentioned by @mnkp, some changes on gpio_stm32_config function would be needed, but are not particularly linked to the purpose of this PR.
So, I've pushed #19646 on topic-gpio branch. It would be required that you base your change on that as well.

To sum up, after updating this PR wrt requested changes, can you rebase on top of #19646 and then push it on topic-gpio branch?

Thanks for your understanding on this unusualy complex PR handling.

[zephyrbot]

All checks are passing now.

checkpatch (informational only, not a failure)

-:92: WARNING:PREFER_KERNEL_TYPES: Prefer kernel type 'u32_t' over 'uint32_t'
#92: FILE: drivers/gpio/gpio_stm32.c:244:
+	uint32_t line;

- total: 0 errors, 1 warnings, 174 lines checked

NOTE: For some of the reported defects, checkpatch may be able to
      mechanically convert to the typical style using --fix or --fix-inplace.

Your patch has style problems, please review.

NOTE: Ignored message types: AVOID_EXTERNS BRACES CONFIG_EXPERIMENTAL CONST_STRUCT DATE_TIME FILE_PATH_CHANGES MINMAX NETWORKING_BLOCK_COMMENT_STYLE PRINTK_WITHOUT_KERN_LEVEL SPLIT_STRING VOLATILE

NOTE: If any of the errors are false positives, please report
      them to the maintainers.

Tip: The bot edits this comment instead of posting a new one, so you can check the comment's history to see earlier messages.

[mniestroj]

@erwango I've changed code according to review comments and rebased on top of topic-gpio branch.

[erwango]

@mniestroj, thanks!
I'm a little bit confused though, since I've finally done the changes myself directly in #19579 last week since we wanted to make it move fastly. I should have informed you to spare you the effort, sorry for this.
Anyway, your commits are present in #19579 and will be merged form this branch.
Thanks for this contribution that has been useful to initiate the driver re-spawn.

[mniestroj]

@erwango Ahh, I see that you fixed those issues in separate commit :) Thanks for reviewing and pulling those changes!

[mniestroj]

Main changes from this PR has been merged part of #20443 PR.