linux-kernel - Re: [PATCH 00/82] overflow: Refactor open-coded arithmetic wrap-around

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: Tue, 23 Jan 2024 13:56:34 -0800
From: Kees Cook <keescook@...omium.org>
To: Mark Rutland <mark.rutland@....com>
Cc: linux-hardening@...r.kernel.org,
	"Gustavo A. R. Silva" <gustavoars@...nel.org>,
	Bill Wendling <morbo@...gle.com>,
	Justin Stitt <justinstitt@...gle.com>, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 00/82] overflow: Refactor open-coded arithmetic
 wrap-around

On Tue, Jan 23, 2024 at 09:46:35AM +0000, Mark Rutland wrote:
> With that in mind, I note that this patch primarily modifies addition
> operations, but leaves subtraction operations unchanged (even though those
> permit the value to go below the minimum, or above the maximum if a negative
> value is used as the subtrahend).

Right, this was kind of a "first pass" on what I'd found so far.

> Shouldn't we address both at the same time? I'll note that in many places the
> same logic is used for both the add and sub, and can legitimately overflow or
> underflow; I hope that whatever we use to suppress overflow warnings also
> ignores underflow.
> 
> [...]
> 
> Looking at the diffstat, I think you've missed a few places:
> 
> [...]
> 
> This misses the include/asm-generic/{atomic,atomic64}.h implementations.
> 
> This also misses the include/linux/atomic/atomic-arch-fallback.h
> implementations. Those are generated from the scripts/atomic/fallbacks/*
> templates, and you'll need to adjust at least fetch_add_unless and
> inc_unless_negative. As noted on other patches, my preference is to use
> add_wrap() in those.
> [...]
> This misses lib/atomic64.c.

Thanks! I'll take a closer look at places we can use the helpers on the
atomics.

-Kees

-- 
Kees Cook