linux-kernel - Re: [PATCH 00/20] objtool: UACCESS validation v3

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: <20190307171709.dap5hfeof4yo3nsc@treble>
Date:   Thu, 7 Mar 2019 11:17:09 -0600
From:   Josh Poimboeuf <jpoimboe@...hat.com>
To:     Linus Torvalds <torvalds@...ux-foundation.org>
Cc:     Peter Zijlstra <peterz@...radead.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Peter Anvin <hpa@...or.com>,
        Julien Thierry <julien.thierry@....com>,
        Will Deacon <will.deacon@....com>,
        Andy Lutomirski <luto@...capital.net>,
        Ingo Molnar <mingo@...nel.org>,
        Catalin Marinas <catalin.marinas@....com>,
        James Morse <james.morse@....com>, valentin.schneider@....com,
        Brian Gerst <brgerst@...il.com>,
        Andrew Lutomirski <luto@...nel.org>,
        Borislav Petkov <bp@...en8.de>,
        Denys Vlasenko <dvlasenk@...hat.com>,
        Linux List Kernel Mailing <linux-kernel@...r.kernel.org>,
        Dmitry Vyukov <dvyukov@...gle.com>,
        Steven Rostedt <rostedt@...dmis.org>
Subject: Re: [PATCH 00/20] objtool: UACCESS validation v3

On Thu, Mar 07, 2019 at 09:00:49AM -0800, Linus Torvalds wrote:
> On Thu, Mar 7, 2019 at 8:47 AM Josh Poimboeuf <jpoimboe@...hat.com> wrote:
> >
> > This "fixes" it, and also seems to help -Os make much code:
> 
> Yeah, considering that this __trace_if() macro from hell is doing an
> 'if()' on the result of that inner thing, it makes sense to *not*  use
> that "looks simpler and shorter" array sequence, but depend on the
> compiler then noticing that the conditionals are the same and joining
> the two branches together.
> 
> The compiler has to do much more work to either generate the actual
> dynamic array thing, or notice that the _index_ of the array matches
> the _conditional_ on the branch, and undo that thing.

Yeah, agreed.  Now it doesn't have to do the funky xor thing to convert
the conditional to an int.

> But that macro really is the macro from hell regardless.
> 
> Do people really use CONFIG_PROFILE_ALL_BRANCHES?

IIRC, Steven runs it once a year or so...

-- 
Josh