Re: mbr gpt crc32 calculation

["Jonathan A. Kollasch" <jakllsch%kollasch.net@localhost>]

On Tue, Dec 19, 2017 at 12:55:28PM +0000, Taylor R Campbell wrote:
> dave0 in #NetBSD on Freenode suggested the attached replacement for
> the crc32 subroutine in src/sys/arch/i386/stand/mbr/gpt.S, to save a
> bit of space where speed is unimportant.
> 
> I don't have the capacity to test it and make sure it's the right
> thing here, but it looks plausible.  Could someone who is more
> i386-wizardly than I am take a look at it?


I've tested it the best I could (in 32-bit userland) and found it to
produced identical results for the one test I subjected it to.  I have
committed this patch.

But this is all sort of moot as we're using/shipping the code from
gptmbr.S and not gpt.S.

	Jonathan Kollasch