Received-SPF: pass (google.com: domain of linux-crypto+bounces-2403-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249;
Message-ID: <b749d5ee-c3b8-4cbd-b252-7773e4536e07@linux.ibm.com>
Date: Thu, 29 Feb 2024 13:45:35 -0500
Precedence: bulk
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: [PATCH v3 00/10] Add support for NIST P521 to ecdsa
Content-Language: en-US
To: Lukas Wunner <lukas@wunner.de>
Cc: keyrings@vger.kernel.org, linux-crypto@vger.kernel.org,
        herbert@gondor.apana.org.au, davem@davemloft.net,
        linux-kernel@vger.kernel.org, saulo.alessandre@tse.jus.br
References: <20240223204149.4055630-1-stefanb@linux.ibm.com>
 <20240229093419.GA32424@wunner.de>
From: Stefan Berger <stefanb@linux.ibm.com>
In-Reply-To: <20240229093419.GA32424@wunner.de>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit


On 2/29/24 04:34, Lukas Wunner wrote:
> On Fri, Feb 23, 2024 at 03:41:39PM -0500, Stefan Berger wrote:
>> This series adds support for the NIST P521 curve to the ecdsa module.
>>
>> An issue with the current code in ecdsa is that it assumes that input
>> arrays providing key coordinates for example, are arrays of digits
>> (a 'digit' is a 'u64'). This works well for all currently supported
>> curves, such as NIST P192/256/384, but does not work for NIST P521 where
>> coordinates are 8 digits + 2 bytes long. So some of the changes deal with
>> converting byte arrays to digits and adjusting tests on input byte
>> array lengths to tolerate arrays not providing multiples of 8 bytes.
> 
> Don't you also need to amend software_key_query()?  In the "issig" case,
> it calculates len = crypto_sig_maxsize(sig), which is 72 bytes for P521,
> then further below calculates "info->max_sig_size = 2 * (len + 3) + 2;"
> 
> I believe the ASN.1 encoded integers are just 66 bytes instead of 72,
> so info->max_sig_size is 6 bytes too large.  Am I missing something?

Right! Good catch. While the 'keyctl pkey_verify' interface was already 
working the space was too generous with 72 bytes. So I adjusted 
ecdsa_max_size now to base the size calculations on nbits rather than 
ndigits and we now get 66 bytes.

For so-far supported curves the max_sig_size is:

2 bytes for sequence (0x30) + following length as single byte
Each coordinate may have a 0 prepended to make a possibly negative 
number positive:

  => 2 + 2 * (2 + 1 + len)

In case of NIST P521 the max signature length is calculated as follows:

3 bytes for sequence (0x30) + following length as 2 bytes
The coordinates won't have a preprended 0 byte since only 1 bit is used 
in the highest bit, so only 2 bytes for

  => 3 + 2 * (2 + len)

We would have to adjust the math there as well. The max. signature size 
for NIST P521 is 139 rather than 140 with the first formula.

    Stefan


> 
> Thanks,
> 
> Lukas