Re: [PATCH v2] x86, boot: Fix word-size assumptions in has_eflag ()inline asm

[David Woodhouse]

On Thu, 2014-01-30 at 05:45 -0800, H. Peter Anvin wrote:
> This would seem like a job for <asm/asm.h>.

Except that's all done on CONFIG_X86_32 which isn't useful for what we
are doing here.

We could, potentially, *change* <asm/asm.h> so that it actually makes
its choices based on whether __x86_64__ is defined?

Which would probably work, except for the next horrid corner case I
haven't discovered/realised yet :)

I think I'm inclined to keep this fairly localised, unless we really see
a case for doing it elsewhere in the kernel. Which is unlikely since a
bare 'pushf' should always do the right thing except in the weird
no-mans-land that is building 16-bit code with a 32-bit compiler.

I note that BITS_PER_LONG==64 even when building the 16-bit code as part
of a 64-bit kernel, too. That one is just waiting to bite us some day...

-- 
dwmw2

Attachment: smime.p7s
Description: S/MIME cryptographic signature