Messages in this thread | | | Date | Wed, 3 Feb 2021 16:54:08 +0100 | From | Borislav Petkov <> | Subject | Re: [PATCH] x86/fpu: Use consistent test for X86_FEATURE_XSAVES |
| |
On Wed, Feb 03, 2021 at 07:40:07AM -0800, Dave Hansen wrote: > On 2/3/21 3:23 AM, Borislav Petkov wrote: > >> -/* > >> - * 'XSAVES' implies two different things: > >> - * 1. saving of supervisor/system state > >> - * 2. using the compacted format > >> - * > >> - * Use this function when dealing with the compacted format so > >> - * that it is obvious which aspect of 'XSAVES' is being handled > >> - * by the calling code. > > @dhansen, are you still hung up on that "obvious aspect" or can we kill > > this? > > I still want the compacted-format handling code to be marked. You can > do that with new comments: > > /* Note: XSAVES always uses compacted format: */ > if (boot_cpu_has(X86_FEATURE_XSAVES)) { > > or, leave it as-is: > > if (using_compacted_format()) { > ... > > Otherwise, we assume that every human being that looks at this code > *KNOWS* that XSAVES==compacted. That's not a great assumption.
Well, the reason why I reacted to this is because I was looking at using_compacted_format() - aha this, checks X86_FEATURE_XSAVES - but then other code paths in fpu/ check X86_FEATURE_XSAVES directly. And this is confusing, making me wonder why is that special oneliner there. Sure, the comment above it says why...
I guess if you wanna keep it, then we need another oneliner for 1. or really do comments at each call site.
Hmm.
-- Regards/Gruss, Boris.
SUSE Software Solutions Germany GmbH, GF: Felix Imendörffer, HRB 36809, AG Nürnberg
| |