linux-kernel - Re: [PATCH AUTOSEL 4.9 003/141] soc: fsl: qe: change return type of cpm_muram

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] [day] [month] [year] [list]

Date:   Fri, 14 Feb 2020 21:49:19 +0100
From:   Rasmus Villemoes <linux@...musvillemoes.dk>
To:     Sasha Levin <sashal@...nel.org>, linux-kernel@...r.kernel.org,
        stable@...r.kernel.org
Cc:     Timur Tabi <timur@...nel.org>, Li Yang <leoyang.li@....com>,
        linuxppc-dev@...ts.ozlabs.org,
        linux-arm-kernel@...ts.infradead.org,
        Christophe Leroy <christophe.leroy@....fr>
Subject: Re: [PATCH AUTOSEL 4.9 003/141] soc: fsl: qe: change return type of
 cpm_muram_alloc() to s32

On 14/02/2020 17.19, Sasha Levin wrote:
> From: Rasmus Villemoes <linux@...musvillemoes.dk>
> 
> [ Upstream commit 800cd6fb76f0ec7711deb72a86c924db1ae42648 ]

Hmm. Please note that these two autosel patches were part of a giant
48-patch series. While not all depending on each other, there are
definitely some dependencies, and between 800cd6fb76f0 and 148587a59f6b
there is e.g. be2e9415f8b3 which changes the type used to store the
return value from cpm_muram_alloc(), and a whole lot of other
cpm_muram_alloc() refactorings and cleanups - one of which
(b6231ea2b3c6) caused a regression on ppc 8xx.

So I think taking just these two might not work as expected, but taking
even more from that series is quite error-prone. Unless someone speaks
up and explicitly points out and verifies some specific subset of the
patches for a specific stable tree, I think they should not be added to
any -stable kernel.

[FWIW, we use the whole series backported to 4.19.y on both arm and ppc
platforms, but as the b6231ea2b3c6 case showed, that doesn't really
prove there are no problems cherry-picking these].

Rasmus