linux-kernel - Re: KVM: s390/mm: Clarification for two return value checks in gmap

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:   Sat, 2 May 2020 19:27:31 +0200
From:   David Hildenbrand <david@...hat.com>
To:     Markus Elfring <Markus.Elfring@....de>, linux-s390@...r.kernel.org,
        kvm@...r.kernel.org,
        Christian Bornträger <borntraeger@...ibm.com>,
        Claudio Imbrenda <imbrenda@...ux.ibm.com>,
        Cornelia Huck <cohuck@...hat.com>,
        Heiko Carstens <heiko.carstens@...ibm.com>,
        Janosch Frank <frankja@...ux.ibm.com>,
        Vasily Gorbik <gor@...ux.ibm.com>
Cc:     LKML <linux-kernel@...r.kernel.org>
Subject: Re: KVM: s390/mm: Clarification for two return value checks in
 gmap_shadow()

On 02.05.20 17:43, Markus Elfring wrote:
> Hello,
> 
> I have tried another small script out for the semantic patch language.
> This source code analysis approach points out that the function “gmap_find_shadow”
> is called two times by the function “gmap_shadow”.
> https://elixir.bootlin.com/linux/v5.7-rc3/source/arch/s390/mm/gmap.c#L1628
> https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/log/arch/s390/mm/gmap.c
> 
> Null pointer checks are performed at these places.

Right, in case we have already a shadow, we return it. In case we are
just concurrently creating/initializing another one, we return -EINVAL
so the caller will retry (and find the fully initialized one). In case
we get NULL, we have to create a new one.

> The function “gmap_find_shadow” is documented in the same source file
> that the pointer “ERR_PTR(-EAGAIN)” can eventually be returned.
> Are the referenced gmap data structures always initialised here?

-EAGAIN makes sure that we are not touching partially initialized one.
In case we find a valid gmap shadow, it is fully initialized. That's
what we have the ->initialized field for.

Hope that answers your question.

-- 
Thanks,

David / dhildenb