linux-kernel - Re: [PATCH v3 05/17] arm64, hibernate: check pgd table allocation

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: <ddd81093-89fc-5146-0b33-ad3bd9a1c10c@arm.com>
Date:   Fri, 6 Sep 2019 16:17:53 +0100
From:   James Morse <james.morse@....com>
To:     Pavel Tatashin <pasha.tatashin@...een.com>
Cc:     jmorris@...ei.org, sashal@...nel.org, ebiederm@...ssion.com,
        kexec@...ts.infradead.org, linux-kernel@...r.kernel.org,
        corbet@....net, catalin.marinas@....com, will@...nel.org,
        linux-arm-kernel@...ts.infradead.org, marc.zyngier@....com,
        vladimir.murzin@....com, matthias.bgg@...il.com,
        bhsharma@...hat.com, linux-mm@...ck.org, mark.rutland@....com
Subject: Re: [PATCH v3 05/17] arm64, hibernate: check pgd table allocation

Hi Pavel,

On 21/08/2019 19:31, Pavel Tatashin wrote:
> There is a bug in create_safe_exec_page(), when page table is allocated
> it is not checked that table is allocated successfully:
> 
> But it is dereferenced in: pgd_none(READ_ONCE(*pgdp)).

If there is a bug, it shouldn't be fixed part way through a series. This makes it
difficult to backport the fix.

Please split this out as an independent patch with a 'Fixes:' tag for the commit that
introduced the bug.

> Another issue,

So this patch does two things? That is rarely a good idea. Again, this makes it difficult
to backport the fix.

> is that phys_to_ttbr() uses an offset in page table instead
> of pgd directly.

If you were going to reuse this, that would be a bug. But because the only page that is
being mapped, is mapped to PAGE_SIZE, all the top bits will be 0. The offset calls are
boiler-plate. It doesn't look intentional, but its harmless.

Please separate out the potential NULL-dereference bits so there is a clean stand-alone
fix that can be sent to the stable trees.

Thanks,

James