i've rewritten the 2.5.1 RAID-1 code to handle the new block layer
properly. The patch, against 2.5.1-pre10, can be found at:
http://redhat.com/~mingo/raid-patches/raid-2.5.1-H2
Changes:
- rewrote raid1.c/raid1.h to use bio's
- large cleanups all around raid1.c/raid1.h. Streamlined/simplified code,
more standard constructs used such as list.h. The RAID-1 code now uses
sector_t all around the place, which enables the switching to 64-bit
sector numbers later in 2.5.
- got rid of the private bh/buffer pools within raid1.c, it's using the
generic mempool interface now. This not only brings in a bit more
performance, but is also much more maintainable and uses the same
generic piece of code that is used by the IO layer's bio allocation
code as well.
- resync feature: it uses big IO requests to do resync now. This resulted
in significant resync speedups on faster SCSI drives. resync thread CPU
usage is much lower now as well.
- a number of generic RAID layer cleanups. (md.c, include files.)
obviously, besides 2.5.1-pre10 being an extremely development kernel, this
patch is large, intrusive, and is work in progress as well. It is working
just fine on all systems tested so far (including Jens Axobe's own RAID1
array), but take care nevertheless.
(i'm now working on adopting raid5.c and multipath.c to the 2.5 block IO
architecture.)
Ingo
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sat Dec 15 2001 - 21:00:25 EST