On Sat, May 15, 2021 at 11:24:25PM +0200, Vlastimil Babka wrote:

That's a misunderstanding. __kmalloc() is not a dummy function, you
probably found only the header declaration.

Sorry, that was totally my misunderstanding.
I was reading dummy function in arch/alpha/boot/bootpz.c:415.
I wrongly configured the tool.

It appears clang 10.0.1 is mistakenly evaluating __builtin_constant_p()
as true. Probably something to do with LTO, because MAX_OPTINSN_SIZE
seems it could be a "link-time constant".

That is what I was missing. Thank you for kindly explaining it.

Maybe we could extend Marco Elver's followup patch that uses
BUILD_BUG_ON vs BUG() depending on size_is_constant parameter. It could
use BUG() also if the compiler is LLVM < 11 or something. What would be
the proper code for this condition?

Fixing clang's bug in linux kernel doesn't seem to be a solution.
So now I understand why Nathan said we might require LLVM > 11.

I thought I should do something to fix it because I sent the patch.
but I was misunderstanding a lot. Thank you sincerely for letting me know.

Thanks,

Hyeonggon