linux-kernel - Re: [patch 09/41] x86/kvm: Avoid looking up PKRU in XSAVE buffer

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <YMh8PAomWPzHx3Nt@zn.tnic>
Date:   Tue, 15 Jun 2021 12:09:00 +0200
From:   Borislav Petkov <bp@...e.de>
To:     Dave Hansen <dave.hansen@...el.com>
Cc:     Thomas Gleixner <tglx@...utronix.de>,
        LKML <linux-kernel@...r.kernel.org>,
        Andy Lutomirski <luto@...nel.org>,
        Dave Hansen <dave.hansen@...ux.intel.com>,
        Fenghua Yu <fenghua.yu@...el.com>,
        Tony Luck <tony.luck@...el.com>,
        Yu-cheng Yu <yu-cheng.yu@...el.com>,
        Sebastian Andrzej Siewior <bigeasy@...utronix.de>,
        Peter Zijlstra <peterz@...radead.org>,
        Kan Liang <kan.liang@...ux.intel.com>
Subject: Re: [patch 09/41] x86/kvm: Avoid looking up PKRU in XSAVE buffer

On Mon, Jun 14, 2021 at 12:34:31PM -0700, Dave Hansen wrote:
> I gave that a shot.  Two wrinkles: The PKRU memcpy() needs 'offset' from
> cpuid_count() and the PKRU case also needs the 'valid -=' manipulation.
>  The result is attached, and while it makes the diff look better, I
> don't think the resulting code is an improvement.

Bah, that was too much wishful and faulty thinking on my part, forget
what I said.

> I *think* these are already stored in xfeature_uncompacted_offset[].  It
> would be a pretty simple matter to export it.  I just assumed that this
> is a slow enough path that the KVM folks don't care.

I guess. Yeah, let's cleanup the FPU mess first and then see what makes
sense or not.

> I'm happy to change it, but I usually like to separate declarations from
> pure code.  Although, I guess that's a bit inconsistent in that file.

No, this is what I mean:

+			src = get_xsave_addr(xsave, xfeature_nr);
+			if (src)
+				memcpy(dest + offset, src, size);

vs

+			void *dest = get_xsave_addr(xsave, xfeature_nr);
+
+			if (dest)
 				memcpy(dest, src + offset, size);

both in your patch.

It is a lot easier when reading the code to have the error handling
glued together with the previous function call.

Thx.

-- 
Regards/Gruss,
    Boris.

SUSE Software Solutions Germany GmbH, GF: Felix Imendörffer, HRB 36809, AG Nürnberg