On Wed, 17 May 2000, Rik van Riel wrote:
>You may want to try Juan Quintela's latest patch (against
>pre9-2) which contains bugfixes for the last known stability
>issues and may well have performance exceeding that of Andrea's
>code...
Waiting I/O completation from shrink_mmap looks broken to me (we just wait
for I/O completation during unplugging and that should be more than
enough). However whatever improvement such patch is supposed to do, it's
_completly_ orthognal with everything I did in the classzone patch
(exactly the same was true with starting I/O from shrink_mmap is been a
change completly orthogonal with the classzone patch). I was starting I/O
from shrink_mmap in early 2.3.x too (I was also coalescing all the dirty
and not locked bh in a bh array of MAX_BH_PER_PAGE) to run only one
ll_rw_block per page (and then I changed ll_rw_block to handle bh with
different major numbers in the same array of bhs of course, that will be
needed also by fs that spawns over multiple blockdevice majors).
As soon as I'll have some more time (early June) I think I'll start
looking into Dave's code for dropping swap_out but that's really not 2.4.x
stuff and I'll for sure it keep it separated.
Andrea
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Tue May 23 2000 - 21:00:16 EST