linux-kernel - Re: [PATCH] crypto: testmgr: don't allocate IV on stack

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: <VI1PR0402MB3342C4C9E798F9973FFDB6A6985E0@VI1PR0402MB3342.eurprd04.prod.outlook.com>
Date:   Tue, 31 Oct 2017 09:05:57 +0000
From:   Horia Geantă <horia.geanta@....com>
To:     Herbert Xu <herbert@...dor.apana.org.au>,
        Gilad Ben-Yossef <gilad@...yossef.com>
CC:     "David S. Miller" <davem@...emloft.net>,
        Ofir Drang <ofir.drang@....com>,
        "linux-crypto@...r.kernel.org" <linux-crypto@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] crypto: testmgr: don't allocate IV on stack

On 10/31/2017 10:00 AM, Herbert Xu wrote:
> On Tue, Oct 31, 2017 at 07:56:26AM +0000, Gilad Ben-Yossef wrote:
>> The IV was allocated on the stack in testmgr skcipher tests.
>> Since HW based tfm providers need to DMA the IV to the HW,
>> this leads to problems and is detected by the DMA-API debug
>> code.
>>
>> Fix it by allocating the IV using kmalloc instead.
>>
>> Signed-off-by: Gilad Ben-Yossef <gilad@...yossef.com>
Gilad, you're not the only one who bumped into this issue:
https://www.mail-archive.com/linux-crypto@vger.kernel.org/msg23074.html
not to mention previous patches that have been accepted:
96692a7305c4 crypto: tcrypt - do not allocate iv on stack for aead speed
tests
9bac019dad80 crypto: testmgr - Fix DMA-API warning
and so on.

> 
> The driver that is mapping the IV directly should be fixed instead.
> Only input that is given in the form of SG lists can be mapped.
> Everything else should be copied if they need to go over DMA.
> 
Herbert, wouldn't it make more sense to follow your previous suggestion:
"Perhaps we should change the API so that it gets passed in as an
SG list."
https://www.mail-archive.com/linux-crypto@vger.kernel.org/msg23082.html

Thanks,
Horia