linux-kernel - Re: [PATCH 23/27] x86/fpu: Defer FPU state load until return to userspace

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: <20190412164827.GG19808@zn.tnic>
Date:   Fri, 12 Apr 2019 18:48:28 +0200
From:   Borislav Petkov <bp@...en8.de>
To:     Sebastian Andrzej Siewior <bigeasy@...utronix.de>
Cc:     linux-kernel@...r.kernel.org, x86@...nel.org,
        Andy Lutomirski <luto@...nel.org>,
        Paolo Bonzini <pbonzini@...hat.com>,
        Radim Krčmář <rkrcmar@...hat.com>,
        kvm@...r.kernel.org, "Jason A. Donenfeld" <Jason@...c4.com>,
        Rik van Riel <riel@...riel.com>,
        Dave Hansen <dave.hansen@...ux.intel.com>
Subject: Re: [PATCH 23/27] x86/fpu: Defer FPU state load until return to
 userspace

On Fri, Apr 12, 2019 at 06:37:41PM +0200, Sebastian Andrzej Siewior wrote:
> (as you mentioned) so we would always record both trace points.
> Therefore I would suggest to remove it.

Remove which one?

Recording both TPs seems to make sense unless it doesn't make a whole
lotta sense to have:

fpregs_mark_activate
|-> trace_x86_fpu_activate_state		<-- TP1
|-> fpregs_activate
    |-> trace_x86_fpu_regs_activated		<-- TP2


Yeah, looks like the two are too much and too close for no good
reason. There's nothing particularly spectacular in-between in
fpregs_activate().

> Maybe we could add a new one to __fpregs_load_activate() one in case we
> avoid loading registers because of fpregs_state_valid(). This might make
> sense.

But that's only the switch_fpu_return() path. Is fpregs_mark_activate()
is going to use only the trace_x86_fpu_regs_activated() one? Note the
"d" at the end.

  [ Btw, those two names need adjusting too: who came up with such close,
     confusing names?!
     ]

-- 
Regards/Gruss,
    Boris.

Good mailing practices for 400: avoid top-posting and trim the reply.