On 27.01.2025 16:48, Alejandro Vallejo wrote: > On Mon Jan 27, 2025 at 11:05 AM GMT, Jan Beulich wrote: >> On 10.01.2025 14:28, Alejandro Vallejo wrote: >>> No functional change. >>> >>> Signed-off-by: Alejandro Vallejo <alejandro.vall...@cloud.com> >> >> Acked-by: Jan Beulich <jbeul...@suse.com> >> >>> --- >>> v2->v3: >>> * const-ified v in fpu_xrstor() >>> * Removed v in fpu_fxrstor() >> >> On this basis the parameter could also be removed from fpu_fxsave(), by >> passing in fip_width instead. > > Could be, but there's not a whole lot of gain to be had? The access must be > done either way before or after the fpu_fxsave() call, and a parameter must be > passed (be it fip_width or v). Passing the vCPU encapsulates the access of > fip_width where its actually used, which seems more desirable, I'd say.
Not much of a gain indeed, largely for symmetry between the two sibling functions. Jan
