linux-kernel - Re: [PATCH v6 02/27] x86: Use symbol name in jump table for PIE support

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]

Date:   Thu, 7 Feb 2019 09:04:45 -0800
From:   Thomas Garnier <thgarnie@...omium.org>
To:     Borislav Petkov <bp@...en8.de>
Cc:     Kernel Hardening <kernel-hardening@...ts.openwall.com>,
        Kristen Carlson Accardi <kristen@...ux.intel.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>,
        "H. Peter Anvin" <hpa@...or.com>,
        "the arch/x86 maintainers" <x86@...nel.org>,
        "Peter Zijlstra (Intel)" <peterz@...radead.org>,
        Ard Biesheuvel <ard.biesheuvel@...aro.org>,
        Nadav Amit <namit@...are.com>,
        Masahiro Yamada <yamada.masahiro@...ionext.com>,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v6 02/27] x86: Use symbol name in jump table for PIE support

On Thu, Feb 7, 2019 at 4:17 AM Borislav Petkov <bp@...en8.de> wrote:
>
> On Thu, Jan 31, 2019 at 11:24:09AM -0800, Thomas Garnier wrote:
> > Replace the %c constraint with %P. The %c is incompatible with PIE
> > because it implies an immediate value whereas %P reference a symbol.
>
> How so?
>
> AFAIK, %c requires a constant operand and if %P is used to print a
> constant, it simply drops syntax-specific prefixes and does a bare
> constant.
>
> I guess that here
>
> https://gcc.gnu.org/onlinedocs/gcc/Extended-Asm.html#x86Operandmodifiers
>
> is not entirely correct as it should not say "If used for a constant"
> for %P but say "symbol or constant".
>
> But before/after asm doesn't show any difference. So what gives?

I assume that's an optimisation done by gcc later. The P modifier in
the documentation does state that it is used to generate PIC code.

>
> before:
> # 39 "./arch/x86/include/asm/jump_label.h" 1
>         1:
>         .byte 0xe9
>          .long .L241 - 2f       #
>         2:
>         .pushsection __jump_table,  "aw"
>          .balign 8
>         .long 1b - ., .L241 - .         #
>          .quad __use_tsc + 1 - .        #,
>         .popsection
>
> after:
> # 39 "./arch/x86/include/asm/jump_label.h" 1
>         1:
>         .byte 0xe9
>          .long .L241 - 2f       #
>         2:
>         .pushsection __jump_table,  "aw"
>          .balign 8
>         .long 1b - ., .L241 - .         #
>          .quad __use_tsc+1 - .  #
>         .popsection
>
> --
> Regards/Gruss,
>     Boris.
>
> Good mailing practices for 400: avoid top-posting and trim the reply.