linux-kernel - Re: odd objtool 'unreachable instruction' warning

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: <CAHk-=wjhcrqXoTLMjTF=pH_+Zq4vRdFY3Y4c_A3TemzSvssRzQ@mail.gmail.com>
Date: Wed, 29 Oct 2025 09:17:36 -0700
From: Linus Torvalds <torvalds@...ux-foundation.org>
To: Josh Poimboeuf <jpoimboe@...nel.org>
Cc: Peter Zijlstra <peterz@...radead.org>, 
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>, 
	Alexandre Chartre <alexandre.chartre@...cle.com>
Subject: Re: odd objtool 'unreachable instruction' warning

On Tue, 28 Oct 2025 at 18:50, Josh Poimboeuf <jpoimboe@...nel.org> wrote:
>
> Just to clarify, Alexandre is working on objTOOL, not objDUMP :-)

Bah. Not the first time I misread - or miswrite - one for the other.

I actually have a stupid alias for disassembly, because I can't ever
remember the long argument names:

   alias disassemble='objdump --disassemble --no-show-raw-insn --no-addresses'

so one day when somebody has an improved script around this, I'll just
change the alias.

(Those flags, btw, are really bad flags in general, because you can't
figure out where the jumps go when you don't see addresses. But my
use-case is generally to look at the compiler-generated asm for
"details", and then I use that objdump disassembly to get a filtered
view of what the code looks like without all the section annotations
and without all the other noise. So that alias is more about my
workflow than about any kind of sane use).

Now, if objtool disassembly ever gives the kind of good disassembly
that 'perf report' does - with branches turned into arrows etc - that
would be lovely, and I'd replace my hacky alias in a heartbeat.
> FWIW, this can be slightly improved by adding '-w' (objdump -drw) which
> at least puts the function name on the same line:
>
>      call   34d <delayed_put_task_struct+0x1d>       349: R_X86_64_PLT32     rethook_flush_task-0x4

Yeah, that's better, and I think I could massage it with some
scripting even more.

But I've thought about writing some simple scripting to make objdump
output clearer many times, and I never really end up caring enough.

So I think I'll wait for Alexandre to do that for me ;)

         Linus