Re: [PATCH] commands/crc32: add compare 2 files crc

[Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>]

On 18:54 Wed 22 Sep     , Sascha Hauer wrote:
> On Wed, Sep 22, 2010 at 06:12:18PM +0200, Jean-Christophe PLAGNIOL-VILLARD wrote:
> > On 18:08 Wed 22 Sep     , Sascha Hauer wrote:
> > > On Wed, Sep 22, 2010 at 05:25:49PM +0200, Peter Korsgaard wrote:
> > > > >>>>> "Sascha" == Sascha Hauer <s.hauer@pengutronix.de> writes:
> > > > 
> > > >  Sascha> On Wed, Sep 22, 2010 at 12:36:25PM +0200, Jean-Christophe PLAGNIOL-VILLARD wrote:
> > > >  >> add -F options to compare to file crc
> > > >  >> 
> > > >  >> it's usefull to compare what you flash in a partition
> > > > 
> > > >  Sascha> Why don't you use memcmp?
> > > > 
> > > >  Sascha> memcmp -s uImage -d /dev/nand0.kernel.bb 0 0
> > > > 
> > > > It's sometimes interesting to be able to compare a checksum with a file
> > > > on the host (E.G. output of cksum/md5sum/sha1sum/..)
> > > 
> > > crc32 can output the crc checksum of a file using the -f option. What
> > > Jean-Christophe added is an option to calculate the checksum of two
> > > files (one given with -f and the other with -F). I just fail to see why
> > > this must be done in a single step instead of using a crc32 followed by
> > > a memcmp.
> > it's faster btw
> 
> Yes, that's the remaining argument, so this should be the selling point
> for this patch?
yes
> 
> btw, does the patch handle the case when the partition is bigger than
> the image file?
yes

> 
> I would also write the image file directly to the flash instead of
> transfering it to RAM and then copy it to flash. That should be even
> faster.
here we just do a crc32 but we could integrate it in the tftp & co to as check
in a second step

Best Regards,
J.

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox