> On May 7, 2021, at 12:58 PM, Dave Hansen <dave.hansen@xxxxxxxxx> wrote: > > On 5/7/21 9:52 AM, Paolo Bonzini wrote: >> This can be optimized as well, can't it? This means that the only case >> that needs the rdpkru is in switch_fpu_finish, and __write_pkru can be >> removed completely: >> >> - do the rdpkru+wrpkru in switch_fpu_finish >> >> - just use wrpkru in KVM > > I was going to suggest exactly the same thing. It doesn't require the > compiler to be smart, and wrpkru() is available in the same header as > __write_pkru(). Thanks Paolo, Dave. Good suggestion, I’ll work up a v2 and send it out. > > I also detest the mysterious true/false arguments to functions where you > have no clue what they're doing at the call site without comments. Noted, my apologies. Thats a good point, I won’t make that mistake again.
