Re: Brainy: bug in x86/cpu_ucode_intel.c

[Robert Elz <kre%munnari.OZ.AU@localhost>]

One more thing I noticed while reading the code - and from what I can tell
it has been like this forever - there's no input validation at all.

The code uses pointer+48 - but nothing has checked that there are
(more than, or even) 48 bytes ... or for that matter, that the uh->
header is even there (the firmware load routines will have just returned
whatever was in the file).

I guess the assumption is that this is a root only facility (I hope!) and
root won't give it a file to load that isn't the correct file,  but it
still seems wrong do do no validity checking at all.

kre