RE: [PATCH] mxcsr patch for i386 & x86-64

[Kamble, Nitin A]

>Ahh. Have we verified that the new semantics of that MXCSR_MASK field
>works on non-intel CPU's too?
Yes, this semantics is same for AMD and Intel. And we have verified that
it also works AMD.

>
>It would also (in my opinion) make sense to just export the
>"common_mxcsr_mask" (and probably just rename it as
"mxcsr_feature_mask"
>or something - where does that "common" come from? Is it just to imply
>that it's the bits that all CPU's support "in common", or what?

Yes, The "common" word is for the common set of bit-mask between all the
SMP processors, in case they are not same. Renaming the
mxcsr_common_mask to mxcsr_feature_mask is ok.

>
>Right now you export a function that does just a simple mask operation,
>and quite frankly, that just seems to make it less clear what the code
is
>doing. So who not get rid of that "set_fpu_mxcsr()" function, and just
>replace all the "0xffbf" uses with "mxcsr_feature_mask"?
>
>Hmm?
I agree that it will be cleaner to read. The idea of the set_fpu_mxcsr()
was to export access to the static mxcsr_common_mask variable. We can
get rid of the function by making the mask a global variable, or through
a new function exporting the static variable.
>
>		Linus
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/