linux-kernel - Re: [patch 2/3] x86/process: Optimize TIF

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] [day] [month] [year] [list]

Message-ID: <CALCETrX7rXvp0NeCgBWaBzHYhAJq0OPwGOU4x1thTag45LLTOA@mail.gmail.com>
Date:   Fri, 16 Dec 2016 11:22:55 -0800
From:   Andy Lutomirski <luto@...capital.net>
To:     Thomas Gleixner <tglx@...utronix.de>
Cc:     LKML <linux-kernel@...r.kernel.org>, X86 ML <x86@...nel.org>,
        Peter Zijlstra <peterz@...radead.org>,
        Kyle Huey <khuey@...ehuey.com>,
        Andy Lutomirski <luto@...nel.org>
Subject: Re: [patch 2/3] x86/process: Optimize TIF_BLOCKSTEP switch

On Fri, Dec 16, 2016 at 12:47 AM, Thomas Gleixner <tglx@...utronix.de> wrote:
> On Thu, 15 Dec 2016, Andy Lutomirski wrote:
>> On Thu, Dec 15, 2016 at 8:44 AM, Thomas Gleixner <tglx@...utronix.de> wrote:
>> > +static inline void toggle_debugctlmsr(unsigned long mask)
>> > +{
>> > +       unsigned long msrval;
>> > +
>> > +#ifndef CONFIG_X86_DEBUGCTLMSR
>> > +       if (boot_cpu_data.x86 < 6)
>> > +               return;
>> > +#endif
>> > +       rdmsrl(MSR_IA32_DEBUGCTLMSR, msrval);
>> > +       wrmsrl(MSR_IA32_DEBUGCTLMSR, msrval ^ mask);
>> > +}
>> > +
>>
>> This scares me.  If the MSR ever gets out of sync with the TI flag,
>> this will malfunction.  And IIRC the MSR is highly magical and the CPU
>> clears it all by itself under a variety of not-so-well documented
>> circumstances.
>
> If that is true, then the code today is broken as well, when the flag has
> been cleared and both prev and next have the flag set. Then it won't be
> updated for the next task.
>
> The we should not use the TIF flag and store a debugmask in thread info and
> do:
>
>         if (prev->debugmask || next->debugmask) {
>                 if (static_cpu_has(X86_FEATURE_BLOCKSTEP)) {
>                         rdmsrl(MSR_IA32_DEBUGCTLMSR, msrval);
>                         msrval &= DEBUGCTLMSR_BTF;
>                         msrval |= next->debugmask;
>                 }
>         }

Seems reasonable to me.  Although keeping it in flags might simplify
the logic a bit.  FWIW, I doubt we care about performance much when
either prev or next has the bit set.

--Andy