linux-kernel - Re: [PATCH v2 1/4] mm: define MAX_PTRS_PER

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: <80f340f2-0323-8092-7e6d-c93b26fb7cf7@c-s.fr>
Date:   Tue, 10 Dec 2019 07:35:15 +0100
From:   Christophe Leroy <christophe.leroy@....fr>
To:     Daniel Axtens <dja@...ens.net>, linux-kernel@...r.kernel.org,
        linux-mm@...ck.org, linuxppc-dev@...ts.ozlabs.org,
        linux-s390@...r.kernel.org, linux-xtensa@...ux-xtensa.org,
        linux-arch@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
        kasan-dev@...glegroups.com, aneesh.kumar@...ux.ibm.com,
        bsingharora@...il.com
Subject: Re: [PATCH v2 1/4] mm: define MAX_PTRS_PER_{PTE,PMD,PUD}



Le 10/12/2019 à 05:47, Daniel Axtens a écrit :
> powerpc has boot-time configurable PTRS_PER_PTE, PMD and PUD. The
> values are selected based on the MMU under which the kernel is
> booted. This is much like how 4 vs 5-level paging on x86_64 leads to
> boot-time configurable PTRS_PER_P4D.
> 
> So far, this hasn't leaked out of arch/powerpc. But with KASAN, we
> have static arrays based on PTRS_PER_*, so for powerpc support must
> provide constant upper bounds for generic code.
> 
> Define MAX_PTRS_PER_{PTE,PMD,PUD} for this purpose.
> 
> I have configured these constants:
>   - in asm-generic headers
>   - on arches that implement KASAN: x86, s390, arm64, xtensa and powerpc

I think we shoud avoid spreading default values all over the place when 
all arches but one uses the default.

I would drop this patch 1, squash the powerpc part of it in the last 
patch, and define defaults in patch 2, see my comments there.

> 
> I haven't wired up any other arches just yet - there is no user of
> the constants outside of the KASAN code I add in the next patch, so
> missing the constants on arches that don't support KASAN shouldn't
> break anything.
> 
> Suggested-by: Christophe Leroy <christophe.leroy@....fr>
> Signed-off-by: Daniel Axtens <dja@...ens.net>
> ---
>   arch/arm64/include/asm/pgtable-hwdef.h       | 3 +++
>   arch/powerpc/include/asm/book3s/64/hash.h    | 4 ++++
>   arch/powerpc/include/asm/book3s/64/pgtable.h | 7 +++++++
>   arch/powerpc/include/asm/book3s/64/radix.h   | 5 +++++
>   arch/s390/include/asm/pgtable.h              | 3 +++
>   arch/x86/include/asm/pgtable_types.h         | 5 +++++
>   arch/xtensa/include/asm/pgtable.h            | 1 +
>   include/asm-generic/pgtable-nop4d-hack.h     | 9 +++++----
>   include/asm-generic/pgtable-nopmd.h          | 9 +++++----
>   include/asm-generic/pgtable-nopud.h          | 9 +++++----
>   10 files changed, 43 insertions(+), 12 deletions(-)
> 

Christophe