linux-kernel - Re: [POC][RFC][PATCH] build: Make weak functions visible in kallsyms

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-=wiL8B2=fPaRwDPGgQhVs=3G=8Gy=QyR59L85L0GF67Gbg@mail.gmail.com>
Date: Thu, 26 Dec 2024 19:35:18 -0800
From: Linus Torvalds <torvalds@...ux-foundation.org>
To: Steven Rostedt <rostedt@...dmis.org>
Cc: LKML <linux-kernel@...r.kernel.org>, 
	Linux Trace Kernel <linux-trace-kernel@...r.kernel.org>, bpf <bpf@...r.kernel.org>, 
	linux-kbuild@...r.kernel.org, Masami Hiramatsu <mhiramat@...nel.org>, 
	Mathieu Desnoyers <mathieu.desnoyers@...icios.com>, Peter Zijlstra <peterz@...radead.org>, 
	Masahiro Yamada <masahiroy@...nel.org>, Nathan Chancellor <nathan@...nel.org>, 
	Nicolas Schier <nicolas@...sle.eu>, Zheng Yejian <zhengyejian1@...wei.com>, 
	Martin Kelly <martin.kelly@...wdstrike.com>, 
	Christophe Leroy <christophe.leroy@...roup.eu>, Josh Poimboeuf <jpoimboe@...hat.com>, 
	Mark Rutland <mark.rutland@....com>
Subject: Re: [POC][RFC][PATCH] build: Make weak functions visible in kallsyms

On Thu, 26 Dec 2024 at 13:49, Steven Rostedt <rostedt@...dmis.org> wrote:
>
> But then, when the linker removes these functions because they were
> overridden, the code does not disappear, leaving the pointers in the
> __mcount_loc locations.

Btw, does this actually happen when the compiler does the mcount thing for us?

It *feels* like this might be a bug in the FTRACE_MCOUNT_USE_OBJTOOL
logic interacting with --gc-setions.

Is the problem that --gc-sections removed the function section that
contained the weak function that was then never used, but the objtool
thing with create_mcount_loc_sections() would generate that
mcount_loc_list and nothing realized that it's no longer there?

Or does it happen even with the compiler-generated case (ie with the
-mrecord-mcount and FTRACE_MCOUNT_USE_CC)?

We can disable LD_DEAD_CODE_DATA_ELIMINATION, if that's what triggers it.

It's marked as experimental, and it does smell like either
--gc-sections is buggy, or we're doing something wrong to trigger it
(and I could easily see objtool rewriting object files being that
trigger...)

                  Linus