Re: [PATCH 0/3] KASAN support for 64-bit Book 3S powerpc

[Christophe Leroy <christophe.leroy@csgroup.eu>]

Le 11/05/2022 à 09:25, Paul Mackerras a écrit :
> [Vous ne recevez pas souvent de courriers de la part de paulus@ozlabs.org. D?couvrez pourquoi cela peut ?tre important ? l'adresse https://aka.ms/LearnAboutSenderIdentification.]
> 
> This patch series implements KASAN on 64-bit POWER with radix MMU,
> such as POWER9 or POWER10.  Daniel Axtens posted previous versions of
> these patches, but is no longer working on KASAN, and I have been
> asked to get them ready for inclusion.
> 
> Because of various technical difficulties, mostly around the need to
> allow for code that runs in real mode, we only support "outline" mode
> (as opposed to "inline" mode), where the compiler adds a call to
> a checking procedure before every store to memory.
> 
> The main patch is still marked [RFC] because of known deficiencies,
> specifically that the kernel will crash on boot on a HPT system, and
> that out-of-bounds accesses to module global data are not caught
> (which leads to one of the KASAN tests failing).
> 
> Comments welcome.  I may not have addressed all the comments about
> previous versions of these patches; if so, then please just repeat
> your comments.

Last patch is quite big. I think it would be worth splitting in several 
patches. A made a few suggestions while reviewing it. I think it could 
also be worth putting all the adding of no_instr and KASAN_SANITIZE := n 
in a preparatory patch.

Also, your last patch includes v1 of "powerpc: make stack walking 
KASAN-safe" 
(https://patchwork.ozlabs.org/project/linuxppc-dev/patch/20210528074806.1311297-1-dja@axtens.net/) 
while v2 was merged 
(https://patchwork.ozlabs.org/project/linuxppc-dev/patch/20210614120907.1952321-1-dja@axtens.net/)

Christophe