Messages in this thread | | | Date | Mon, 27 Apr 2015 20:47:30 +0200 | From | Borislav Petkov <> | Subject | Re: [PATCH] x86_64, asm: Work around AMD SYSRET SS descriptor attribute issue |
| |
On Mon, Apr 27, 2015 at 11:12:05AM -0700, Linus Torvalds wrote: > So if one or two cycles in this code doesn't matter, then why are we > adding alternate instructions just to avoid a few ALU instructions and > a conditional branch that predicts perfectly? And if it does matter, > then the 6-byte option looks clearly better..
You know what? I haven't even measured the tree *without* Denys' stricter RCX canonical-ness patch. All numbers so far are from 4.0+ with tip/master ontop which has Denys' patch.
And I *should* measure once with plain 4.1-rc1 and then with Denys' patch ontop to see whether this all jumping-thru-alternatives-hoops is even worth it.
If nothing else, it was a nice exercise. :-)
-- Regards/Gruss, Boris.
ECO tip #101: Trim your mails when you reply. --
| |