On Fri, Aug 24, 2012 at 2:33 PM, H. Peter Anvin wrote:
> I have applied this to tip:x86/fpu, but I have also asked Suresh to
> prepare a followon patch to decouple eager save from the existence of
> the XSAVE instruction. It seems pretty clear that eager save is a net
> benefit in the presence of
I have applied this to tip:x86/fpu, but I have also asked Suresh to
prepare a followon patch to decouple eager save from the existence of
the XSAVE instruction. It seems pretty clear that eager save is a net
benefit in the presence of the XSAVEOPT, but it isn't as clear for only
having XSAVE, as f
These patches are against tip/x86/fpu. Few cleanups and more improtantly
this series introduces the non-lazy FPU restore mechanism for processors
supporting xsave feature. More details in the individual patch changelogs.
Thanks.
Suresh Siddha (6):
x86, fpu: drop_fpu() before restoring new state
3 matches
Mail list logo
