linux-kernel - Re: [PATCH v4 1/4] crypto: ti - Add support for AES-XTS in DTHEv2 driver

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: <c76843e9-e644-4b9b-803f-b9eb9a31b89c@ti.com>
Date: Tue, 21 Oct 2025 20:40:40 +0530
From: T Pratham <t-pratham@...com>
To: Herbert Xu <herbert@...dor.apana.org.au>
CC: "David S. Miller" <davem@...emloft.net>,
        Manorit Chawdhry
	<m-chawdhry@...com>,
        Kamlesh Gurudasani <kamlesh@...com>,
        Shiva Tripathi
	<s-tripathi1@...com>,
        Kavitha Malarvizhi <k-malarvizhi@...com>,
        Vishal
 Mahaveer <vishalm@...com>, Praneeth Bajjuri <praneeth@...com>,
        <linux-crypto@...r.kernel.org>, <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v4 1/4] crypto: ti - Add support for AES-XTS in DTHEv2
 driver

Hi Herbert,
On 17-10-2025 22:13, T Pratham wrote:
> On 17-10-2025 11:11, Herbert Xu wrote:
>> On Thu, Oct 09, 2025 at 04:11:31PM +0530, T Pratham wrote:
>>>
>>
>> You should allocate a fallback that is synchronous only.  Then you
>> can store the sub-request on the stack with SYNC_SKCIPHER_REQUEST_ON_STACK.
>> Otherwise the sub-request reqsize may overflow your request object.
> 
> Understood. Will correct this.

Since a similar mechanism is not available for AEADs, I was wondering if
doing the following in the tfm_init functions of AEAD algs (PATCH 3/4
and 4/4 of this series) is acceptable, esp. since setting reqsize is
moved to the algorithm struct field cra_reqsize.

+       unsigned int reqsize = crypto_aead_reqsize(tfm);
+       crypto_aead_set_reqsize(tfm, reqsize +
crypto_aead_reqsize(ctx->aead_fb));


I noticed all older existing drivers using a fallback do something
similar, and was wondering if you'll still accept new drivers which use
crypto_*_set_reqsize() to modify reqsize for fallback use-cases?

-- 
Regards
T Prathm <t-pratham@...com>