Reorg raid5 block xor routines

From: Richard Henderson (rth@twiddle.net)
Date: Sun Sep 10 2000 - 01:02:07 EST


There are two main purposes to this reorg:

 * Split up the tremndously huge xor.c.

 * Make it easier to write pure assembly routines without having
   to interface with C structure offsets. You can see how nasty
   the Alpha and Sparc64 routines were; it promised to be even
   worse for IA-64.

The bulk of the patch is just moving code around. But additionally,

 * Bug fixed in the SSE code ("8*(x*2)" isn't the same as "16*(x)"
   when x expands to "4+1").

 * Alpha code doesn't run into problems with gcc 2.95.

 * New IA-64 routines.

Jakub, I've inspected the restructured VIS routines by eye, and
run the result through a cross-assembler, but couldn't test it.
All the others have been run through the attached user-land test
harness, which someone else might find useful.

r~





-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Fri Sep 15 2000 - 21:00:13 EST