Re: mainline build failure of powerpc allmodconfig for prom_init_check

[Michael Ellerman]

Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx> writes:
> On Sun, Jul 17, 2022 at 9:41 PM Michael Ellerman <mpe@xxxxxxxxxxxxxx> wrote:
>>
>> >         li 4,254                 #,
>>
>> Here we load 254 into r4, which is the 2nd parameter to memset (c).
>
> I love how even powerpc people know that "4" is bogus, and have to
> make it clear that it means "r4".

I wouldn't say it's bogus, I was just translating from asm to English :)

But I agree it's preferable to use a proper register name rather than a
bare integer. I never write asm using bare integers, I always use r4 or
%r4, because as you say it's too easy to get mixed up otherwise.

When looking at generated code I usually use objdump -d output, which
uses the "r4" syntax.

> It's not even an IBM thing. S390 uses perfectly sane register syntax,
> and calls things '%r4" etc.

as accepts that syntax if you tell it to.

We use that syntax in some of our newer inline asm blocks.

> The human-written asm files have those #define's in headers just to
> make things slightly more legible, because apparently the assembler
> doesn't even *accept* the sane names.

I would like to switch to using %rX everywhere and get rid of those
defines, but it's never seemed like it's worth the churn. We have ~48K
lines of asm in arch/powerpc.

cheers