From mboxrd@z Thu Jan  1 00:00:00 1970
From: Marek Vasut <marex@denx.de>
Date: Mon, 31 Mar 2014 11:29:17 +0200
Subject: [U-Boot] [PATCH 3/3] dfu: Introduction of the
	"dfu_checksum_method" env variable for checksum method setting
In-Reply-To: <20140331112431.1157e1c8@amdc2363>
References: <1396255729-6573-1-git-send-email-l.majewski@samsung.com>
	<201403311104.13887.marex@denx.de>
	<20140331112431.1157e1c8@amdc2363>
Message-ID: <201403311129.17258.marex@denx.de>
List-Id: <u-boot.lists.denx.de>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: u-boot@lists.denx.de

On Monday, March 31, 2014 at 11:24:31 AM, Lukasz Majewski wrote:
> Hi Marek,
> 
> > On Monday, March 31, 2014 at 10:48:49 AM, Lukasz Majewski wrote:
> > > Up till now the CRC32 of received data was calculated
> > > unconditionally. The standard crc32 implementation causes long
> > > delays when large images were uploaded.
> > 
> > You might want to check common/cmd_hash.c and include/hash.h for the
> > hash_command() call. It does the resolution of the hash algorithm
> > from it's name and you can operate also SHA1 and SHA256 with it. It
> > would be nice if you could just extend it a bit and use that instead
> > of adding another ad-hoc mechanism.
> > 
> > Do you think it'd be possible to reuse it please ?
> 
> I think, that crc32 shall be calculated when needed. That is why I've
> added a dfu_ckecksum_method variable.
> 
> With its help it is now possible to use different algorithms for
> checking - not only crc32 (which in u-boot is the default and painfully
> slow implementation).
> 
> In the future the code:
> if (dfu_checksum_method == DFU_CRC32)
> 	crc32 calculation;
> 
> will be changed to:
> 
> switch (dfu_checksum_method) {
> 	case CRC32:
> 		crc32 calculation;
> 		break;
> 	case SHA1:
> 		sha1 calculation;
> 		break;
> 	case MD5:
> 		md5 calculation;
> 		break;
> }
> 
> Moreover it is possible to dynamically change the checksum method
> (between invoking dfu command) via adjusting "dfu_checksum_method"
> variable.
> 
> The default approach is to not calculate anything.

I get it, but the direct calling of crc32() function can be abstracted already 
with the hash_command() now. You won't need to the above switch in such case. 
Also, you can implement a NULL hash algo, which would effectivelly model the 
case where you want to disable hashing.

Best regards,
Marek Vasut