Re: classzone-VM + mapped pages out of lru_cache

From: Craig Kulesa (
Date: Wed May 03 2000 - 22:54:50 EST

In initial tests with this patch against pre7-3 on a 128 MB box, the
kswapd problem seems to be eliminated. I used large 'dd' tests and
general application loads, and the general paging behavior (which was
seemingly very unbalanced in pre5) seems reasonable IMHO -- i.e. the
cache shrinks properly. I haven't tried it out on lower-end hardware yet.


... the classzone-18 patch has caused _innumerable_ lockups of my
machine -- all of them relating to mail programs so far. Both 'pine'
(ver 4.21) and 'mutt' cause very reproduceable and _instantaneous_ hard
lockups if I attempt to expunge messages from the mail spool or compose a
new message. A simple test, editing a file from pine's editor 'pico'
(ver 3.7) caused a lockup upon writing the output file.

No oopses, no messages, just instant death[TM].

Not pretty. Has anyone else been able to reproduce this? This is
unfortunate, since performance while the machine is actually alive
seems quite good.

I reverted to 'stock' pre7-3 (editing out the #error line in
mm/vmscan.c) and there are no such problems (other than excessively
CPU-intensive VM and I/O behavior).

Configuration info is at
Email for additional details or testing against new patches.


To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
Please read the FAQ at

This archive was generated by hypermail 2b29 : Sun May 07 2000 - 21:00:13 EST