linux-kernel - Re: [PATCH] ARM: Fix r7/r11 confusion when CONFIG_THUMB2

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: <20130715135420.GG10000@mudshark.cambridge.arm.com>
Date:	Mon, 15 Jul 2013 14:54:20 +0100
From:	Will Deacon <will.deacon@....com>
To:	Jed Davis <jld@...illa.com>
Cc:	Russell King <linux@....linux.org.uk>,
	Peter Zijlstra <a.p.zijlstra@...llo.nl>,
	Paul Mackerras <paulus@...ba.org>,
	Ingo Molnar <mingo@...hat.com>,
	Arnaldo Carvalho de Melo <acme@...stprotocols.net>,
	Robert Richter <rric@...nel.org>,
	"linux-arm-kernel@...ts.infradead.org" 
	<linux-arm-kernel@...ts.infradead.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"oprofile-list@...ts.sf.net" <oprofile-list@...ts.sf.net>
Subject: Re: [PATCH] ARM: Fix r7/r11 confusion when CONFIG_THUMB2_KERNEL=y

Hi Jed,

On Sat, Jul 13, 2013 at 04:18:20AM +0100, Jed Davis wrote:
> There is currently some inconsistency about the "frame pointer" on ARM.
> r11 is the register with assemblers recognize and disassemblers often
> print as "fp", and which is sufficient for stack unwinding when using
> the APCS frame pointer option; but when unwinding with the Exception
> Handling ABI, the register GCC uses when a constant offset won't suffice
> (or when -fno-omit-frame-pointer is used; see kernel/sched/Makefile in
> particular) is r11 on ARM and r7 on Thumb.
> 
> Correspondingly, arch/arm/include/uapi/arm/ptrace.h defines ARM_fp to
> refer to r11, but arch/arm/kernel/unwind.c uses "FP" to mean either r11
> or r7 depending on Thumbness, and it is unclear what other cases such as
> the "fp" in struct stackframe should be doing.
> 
> Effects of this are probably limited to failure of EHABI unwinding when
> starting from a function that uses r7 to restore its stack pointer, but
> the possibility for further breakage (which would be invisible on
> non-Thumb kernels) is worrying.
> 
> With this change, it is hoped, r7 is consistently referred to as "r7",
> and "fp" always means r11; this costs a few extra ifdefs, but it should
> help prevent future issues.

I'm struggling to understand exactly the problem that this patch is trying
to address. If it's just a code consistency issue, I don't think it's worth
it (I actually find it less confusing the way we currently have things) but
if there is a real bug, perhaps you could provide a testcase?

Cheers,

Will
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/