On 10.12.2021 22:35, Sergei Shtylyov wrote:

[...]

quoted

quoted

quoted

quoted

quoted

quoted

quoted

quoted

quoted

quoted

quoted

platform_get_irq() will print a message when it fails.
No need to repeat this.

While at it, drop redundant check for 0 as platform_get_irq() spills
out a big WARN() in such case.

The reason you should be able to remove the "if (!irq)" test is that
platform_get_irq() never returns 0. At least, that is what the function kdoc
says. But looking at platform_get_irq_optional(), which is called by
platform_get_irq(), the out label is:

	WARN(ret == 0, "0 is an invalid IRQ number\n");
	return ret;

So 0 will be returned as-is. That is rather weird. That should be fixed to
return -ENXIO:

	if (WARN(ret == 0, "0 is an invalid IRQ number\n"))
		return -ENXIO;
	return ret;

    My unmerged patch (https://marc.info/?l=linux-kernel&m=163623041902285) does this
but returns -EINVAL instead.

quoted

Otherwise, I do not think that removing the "if (!irq)" hunk is safe. no ?

    Of course it isn't...

It's unsubstantiated statement. The vIRQ 0 shouldn't be returned by any of
those API calls.

    We do _not_ know what needs to be fixed, that's the problem, and that's why the WARN()
is there...

So, have you seen this warning (being reported) related to libahci_platform?

    No (as if you need to really see this while it's obvious from the code review).

quoted

If no, what we are discussing about then? The workaround is redundant and

    I don't know. :-) Your arguments so far seem bogus (sorry! :-))...

It seems you haven't got them at all. The problems of platform_get_irq() et al
shouldn't be worked around in the callers.

    I have clearly explained to you what I'm working around there. If that wasn't clear
enough, I don't want to continue this talk anymore. Good luck with your patch (not this
one).

Good luck with yours, not the one that touches platform_get_irq_optional() though!

    Mmh, I'm not touching it any way that would break what your patch was trying to do,
unless you've re-thopught that. It also shoudn't matter whose patch gets merged 1st
other than some small adaptation).

    BTW, looking at [1], this comment is wrong:

+ * Return: non-zero IRQ number on success, negative error number on failure.

It doesn't mention 0 which you return from this function.

    Also, your commit log is wrong in the description of how to handle the result:

<<
Now:
	ret = platform_get_irq_optional(...);
	if (ret != -ENXIO)
		return ret; // respect deferred probe
	if (ret > 0)
		...we get an IRQ...
 >>

    The (ret != -ENXIO) check also succeeds on the (positive) IRQ #s, so the 
following code becomes unreachable. :-/

[1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=ed7027fdf4ec41ed6df6814956dc11860232a9d5

MBR, Sergey