RE: crc32 cleanups

[Jeff Garzik]

On Fri, 12 Oct 2001 [e-mail] wrote:
> > That leaves (a) unconditionally building 
> > it into the kernel, or (b) Makefile and Config.in rules.
> 
> (a) is simple, but needs a 1KB malloc (or alternately, a 1KB static const
> array - I've taken the approach that the malloc is better)
> (b) isn't that much harder, but requires drivers to be sure to call
> init_crc32 and cleanup_crc32.  If somehow they manage not to do that, Oops.
> I don't want to add a runtime check for the existance of the array in
> crc32().

You are talking about the data; I was talking about the code.

I do not think kernels need the data table, kmalloc'd or statically
built, unless it will be used.  That implies a refcounting scheme.
[WRT "Oops", that is a driver bug, not a case to be considered.  In
Linuxland we do not write code to protect us from rogue code.]

I was pondering whether it was ok to unconditionally include the
lib/crc32.c code, regardless of need.  I am leaning towards "no," which
implies Makefile and Config.in rules which must be updated for each
driver that uses crc32.

	Jeff