[PATCH] driver core: platform: Rename platform_get_irq_optional() to platform_get_irq_silent()

Sergey Shtylyov s.shtylyov at omp.ru
Wed Jan 19 11:44:06 PST 2022


Hello!

On 1/19/22 9:36 PM, Andy Shevchenko wrote:

[...]
>> So you oppose to the name chosen, but not the renaming as such.
> 
> I oppose the name change. The unneeded churn right now since it won't fix
> the issues with the underneath API (platform_get_irq() in this case) and
> will require one more iteration over callers again.
> 
> The main issue that platform_get_irq*() returns magic error code while
> treating 0 in a very special way (issuing WARN() and considering it as
> a successful cookie) and this all is quite confusing.

   I have a patch for that -- to which you were hostile for some reason I still
can't understand. :-)

> If you are going to fix the underlying issue, welcome! Now I see only
> the step to somewhere. I.o.w. this change _standalone_ makes no sense
> to me.

   We already have a fix, no? It just hasn't been applied still... :-)
   Without it the 2 patches dealing with *_optional() don't have much sense.

MBR, Sergey



More information about the Linux-mediatek mailing list