Re: Strong Crypto Support for GRUB2

[Patrick Georgi <patrick@georgi-clan.de>]

Robert Millan schrieb:
>> +/*
>> + * Forward S-box
>> + */
>> +static const uint8 FSb[256] =
>> +{
>> +    0x63, 0x7C, 0x77, 0x7B, 0xF2, 0x6B, 0x6F, 0xC5,
>> +    0x30, 0x01, 0x67, 0x2B, 0xFE, 0xD7, 0xAB, 0x76,
>> +    0xCA, 0x82, 0xC9, 0x7D, 0xFA, 0x59, 0x47, 0xF0,
...
>> +    0x9B, 0x1E, 0x87, 0xE9, 0xCE, 0x55, 0x28, 0xDF,
>> +    0x8C, 0xA1, 0x89, 0x0D, 0xBF, 0xE6, 0x42, 0x68,
>> +    0x41, 0x99, 0x2D, 0x0F, 0xB0, 0x54, 0xBB, 0x16
>> +};
> 
> This (and similar binary blobs) seems like a problem.  What do these numbers
> mean?  If the spec defines a simple algorithm to generate them, a comment
> with its implementation would be nice to have (e.g. like in md5sum.c from
> coreutils).
It's the common distribution form, and not supposed to be edited, as the 
algorithm relies on that particular table layout.

from the header of that file:
 >> + *  FIPS-197 compliant AES implementation
reasonably states where to get it from - maybe extend it with
http://csrc.nist.gov/publications/fips/fips197/fips-197.pdf as an actual 
pointer, and that table with "taken from page 16 of the spec"

The original AES paper defined some criteria on s-box design and a bunch 
of equations on how they do it - but that's not enough information to 
create _this_ particular (ie. the standard's) table.


Regards,
Patrick Georgi