More stuff runs now without "VM killing process" problem.
I've tried: bonnie, aim, dbench, doio and kernel
compilation, all of which didn't show any problems.
There are a couple of water-marks which control
the amount of delayed allocation pages ...
these need to be a function of total memory
and/or amount of buffer memory. For now,
these water-marks are arbitrarily set.
ananth.
Modid: 2.3.99pre2-xfs:slinx:62763a
Date: Thu May 25 12:47:52 PDT 2000
Workarea: bonnie.engr.sgi.com:/build2/ananth/slinx23-xfs-tot
Author: ananth
The following file(s) were checked into:
bonnie:/isms/slinx/2.3.99pre2-xfs
linux/fs/page_buf.c - 1.99
http://oss.sgi.com/cgi-bin/cvsweb.cgi/linux-2.4-xfs/linux/fs/page_buf.c.diff?r1=text&tr1=1.99&r2=text&tr2=1.98&f=h
- Fixes to delayed allocation code and better balancing of
memory usage. Turn ON delay_alloc.
|