Re: [PATCH v4 1/1] use crc32 instead of md5 for hibernation e820 integrity check

[Chris von Recklinghausen]

On 4/9/21 12:56 PM, David Laight wrote:
> From: Chris von Recklinghausen
> > Sent: 08 April 2021 11:46
> >
> > Suspend fails on a system in fips mode because md5 is used for the e820
> > integrity check and is not available. Use crc32 instead.
> >
> > Prior to this patch, MD5 is used only to create a digest to ensure integrity of
> > the region, no actual encryption is done. This patch set changes the integrity
> > check to use crc32 instead of md5 since crc32 is available in both FIPS and
> > non-FIPS modes.
> >
> > Note that the digest is only used as an integrity check. No actual encryption
> > is done.
> If crc32 is good enough, would a 1's compliment sum be good enough?
> It is likely to be faster to calculate and not need special
> functions be built into the kernel at all.


Eric Biggers <ebiggers@xxxxxxxxxx> suggested using crc32_le() which is 
in the library interface (lib/crc32.c) and will always be available 
without any special ifdefs. That's what my next version will be based on.

Thanks,

Chris


> 	David
>
> -
> Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
> Registration No: 1397386 (Wales)