> Date: Wed, 1 Sep 1999 12:31:51 +0100 (BST)
> From: Alan Cox <alan@lxorguk.ukuu.org.uk>
> To: Robert Dinse <nanook@eskimo.com>
> Cc: ericj@eskimo.com, evol@eskimo.com, linux-kernel@vger.rutgers.edu
> Subject: Re: 2.2.12 Memory Leak?
>
> > I ended up going back to 2.2.10, 2.2.12 seems to have a memory leak
> > similiar to the TCP memory leak in 2.2.11 that existed without the TCP memory
> > leak patch.
>
> Did you apply the tcp memory leak patch to 2.2.11 and then apply 2.2.12 over
> the top of it. If so you'll end up with a 2.2.12 with the memory leak bug
> re-inserted
Nope, I got a stock 2.2.12 distribution which kept doing the spin_lock
thing, applied a patch from David Miller that changed the sparc assembly code
slightly (added a few nop's), with that patch neither machine did the spin_lock
thing again, but after a few hours it would eat up all the memory and die.
Then I tried applying Andrea's patch since I also had that applied to
2.2.10 (schedualer patch). Same behavior (except the machine was quicker and
CPU load less).
> Buffers or network buffers ?
The buffers that vmstat reports:
procs memory swap io system cpu
r b w swpd free buff cache si so bi bo in cs us sy id
0 0 0 0 130664 155928 56940 0 0 55 1 179 101 2 5 93
^^^^^^
That number.. With 2.2.10, that value would start out high but as people
retrieved web pages and the cache grew, first free mem would shrink, but then
when free mem got down to a certain value, buff would start to shrink, and
eventually some sort of equalibrium would be reached between buff and cache
with free mem hanging around that minimal value.
But with 2.2.12, free mem just keeps decreasing, then swapd starts to
increase, eventually it gets to a point where anything you try to do says
"killed" and all you can do is reboot and start over.
-
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/