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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20201211164915.GA2414@hirez.programming.kicks-ass.net>
Date:   Fri, 11 Dec 2020 17:49:15 +0100
From:   Peter Zijlstra <peterz@...radead.org>
To:     Josh Poimboeuf <jpoimboe@...hat.com>
Cc:     Nick Desaulniers <ndesaulniers@...gle.com>,
        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 10:37:48AM -0600, Josh Poimboeuf wrote:
> On Fri, Dec 11, 2020 at 10:32:05AM +0100, Peter Zijlstra wrote:
> > Looking at elf.c, it seems we're missing an STT_SECTION symbol for
> > .text.
> > 
> > And indeed, when I add -fno-asynchronous-unwind-tables to clang-11, that
> > goes missing from the readelf .symtab listing. Help ?!
> 
> I had a similar problem with ORC relocations:
> 
>   e81e07244325 ("objtool: Support Clang non-section symbols in ORC generation")
> 
> If Clang strips the section symbol then we have to find the function
> symbol instead.

Argh, I knew I'd seen something like that before, but I couldn't find it :-/

> diff --git a/tools/objtool/check.c b/tools/objtool/check.c
> index c6ab44543c92..9bc18864154f 100644
> --- a/tools/objtool/check.c
> +++ b/tools/objtool/check.c
> @@ -472,8 +472,25 @@ static int create_static_call_sections(struct objtool_file *file)
>  			return -1;
>  		}
>  		memset(reloc, 0, sizeof(*reloc));
> -		reloc->sym = insn->sec->sym;
> -		reloc->addend = insn->offset;
> +
> +		if (insn->sec->sym) {
> +			reloc->sym = insn->sec->sym;
> +			reloc->addend = insn->offset;
> +		} else {
> +			reloc->sym = find_symbol_containing(insn->sec, insn->offset);
> +			if (!reloc->sym) {
> +				WARN_FUNC("can't create static call: missing containing symbol",
> +					  insn->sec, insn->offset);
> +				return -1;
> +			}
> +
> +			reloc->addend = insn->offset - reloc->sym->offset;
> +		}
> +

Do we want to capture all that gunk in something like
elf_reloc_to_insn(reloc, insn) instead of duplicating the magic?