Re: [PATCH] driver core: platform: Fix wrong comment

[Andy Shevchenko]

On Sat, Dec 25, 2021 at 6:50 PM Randy Dunlap <rdunlap@xxxxxxxxxxxxx> wrote:
>
>
>
> On 12/25/21 08:47, Andy Shevchenko wrote:
> > On Sat, Dec 25, 2021 at 2:02 AM Jiasheng Jiang <jiasheng@xxxxxxxxxxx> wrote:
> >>
> >> I notice that there is a 'WARN(ret == 0, "0 is an invalid IRQ
> >> number\n");' before 'return ret;', which means that it is possible to
> >> return 0 if fails.
> >> Therefore, it might be better to correct the wrong comment.
> >> And also there is reply sent by Damien Le Moal because I submitted a
> >> patch to remove the non-zero check of the platform_get_irq() previously.
> >> Damien Le Moal said that the comment for platform_get_irq() is wrond
> >> because it can actually return 0.
> >> Moreover, platform_get_irq() returns platform_get_irq_optional().
> >> As a conclusion, the comments of the platform_get_irq() and
> >> platform_get_irq_optional() should be fixed.
> >> Not only that, the comments of platform_get_irq_byname() and
> >> platform_get_irq_byname_optional() have the same error.
> >> This time I only submit one as an example.
> >> If the patch is right, I will submit another version to fix all.
> >> But, I also notice that the 'return 0' is removed intentionally in the
> >> fixed tag.
> >> I am not sure which one is right.
> >> Anyway, the success IRQ number should be 'postive' other than
> >> 'non-zero'.
> >> So the comment should be corrected.
> >> Here is the mail from Damien Le Moal.
> >
> > ...
> >
> >> Fixes: c2f3f755f5c7 ("Revert "driver core: platform: Make platform_get_irq_optional() optional"")
> >
> > How can it be a Revert?
>
> That's the "title" of that commit:
>
> commit c2f3f755f5c7
> Author: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> Date:   Wed Apr 7 11:47:56 2021 +0200
>
>     Revert "driver core: platform: Make platform_get_irq_optional() optional"

My point is that the proposed change should fix a revert, how come?


-- 
With Best Regards,
Andy Shevchenko