linux-kernel - Re: objtool crashes with some clang produced .o files

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:   Fri, 11 Dec 2020 16:42:01 -0800
From:   Nick Desaulniers <ndesaulniers@...gle.com>
To:     Josh Poimboeuf <jpoimboe@...hat.com>,
        Peter Zijlstra <peterz@...radead.org>
Cc:     Arnd Bergmann <arnd@...nel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        clang-built-linux <clang-built-linux@...glegroups.com>
Subject: Re: objtool crashes with some clang produced .o files

On Fri, Dec 11, 2020 at 12:57 PM Nick Desaulniers
<ndesaulniers@...gle.com> wrote:
>
> Thanks for the patch!
>
> Reviewed-by: Nick Desaulniers <ndesaulniers@...gle.com>
> Link: https://github.com/ClangBuiltLinux/linux/issues/1207

Arnd reported another objtool warning/error from another randconfig in
https://github.com/ClangBuiltLinux/linux/issues/1209 and CrOS just hit
this as well.

I haven't been able to isolate the configs yet (Arnd has posted the
full config: https://pastebin.com/wwwhUL8L

$ ./tools/objtool/objtool orc generate  --no-fp --no-unreachable
--retpoline arch/x86/entry/thunk_64.o
arch/x86/entry/thunk_64.o: warning: objtool: missing symbol for insn
at offset 0x3e

Is the offset 0x3e referring to the final `ret` instruction in
preempt_schedule_notrace_thunk?  Observing insn_to_reloc_sym_addend()
(with your patch applied), it looks like both calls to
find_symbol_containing() with offset and offset-1 returns NULL.  I'm
curious if there's another quirk going on here, or possibly a config
from the randconfig that's messing up the special case? I don't follow
the comment about:
119        * Hack alert.  This happens when we need to reference
120        * the NOP pad insn immediately after the function.
121        */

Attached the object file FWIW.
-- 
Thanks,
~Nick Desaulniers

Download attachment "thunk_64.o" of type "application/octet-stream" (744 bytes)