On May 03 2001, Rob Weir wrote:
> > This memory leak also seems to have evidenced another memory
> > leak: it *seems* that Linux 2.4.4 also has a memory leak
> > (memory which was allocated on swap after the Konqueror thing
> > didn't get back even *after* I went to the console using
> > single user mode!).
> I'm fairly sure that this is a deliberate design decision in all the 2.4.x
> kernels. Once memory is swapped out, it is not removed from the swap, even
> if the page is swapped back in to physical RAM.
> Here's a discussion about it from on kernel traffic:
> http://kt.zork.net/kernel-traffic/kt20010330_113.html#6
I have not tested this extensively (I was so scared that I'm
running the trusted and tried, good ole 2.2.19 right now, with
some lovely patches), but I think that I'll retest it again
(now, I'll have to find another application to stuff the
memory) and report whatever I find (i.e., if memory can be
reclaimed or not).
Anyway, thanks for briging that to my attention. I stopped
following lkml one or two months ago (I need to study much,
much, much more if I ever want to grow up). :-)
[]s, Roger...
--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Rogerio Brito - [EMAIL PROTECTED] - http://www.ime.usp.br/~rbrito/
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
