[Gc] GC_mem_found gets minus?

Boehm, Hans hans.boehm at hp.com
Wed Oct 24 10:29:24 PDT 2007


That's intentional, though could probably be clearer.  I added a comment
to the declaration.  (It's GC_bytes_found in 7.0)

Since remaining free lists are dropped at the beginning of a collection,
we are essentially throwing away free space, which will usually be added
back during the sweep phase.  We shouldn't really be getting credit for
free lists we drop and then rebuild.  Hence the dropped space is
subtracted initially.  Since the swwep is done lazily, it might not
actually be added back in some cases.

However, I do now believe that the accounting bug in
http://article.gmane.org/gmane.comp.programming.garbage-collection.boehm
gc/1868 is potentially serious for applications using explicit
deallocation, and has been around approximately forever.  I did check in
essentially the patch provided there.  If you're worried about this sort
of thing, you might want to fix that one.

Hans

> -----Original Message-----
> From: gc-bounces at napali.hpl.hp.com 
> [mailto:gc-bounces at napali.hpl.hp.com] On Behalf Of Bruce Hoult
> Sent: Tuesday, October 23, 2007 11:26 PM
> To: Jiwon Seo
> Cc: gc at napali.hpl.hp.com
> Subject: Re: [Gc] GC_mem_found gets minus?
> 
> On 10/24/07, Jiwon Seo <jiwon at stanford.edu> wrote:
> > I'm using boehm gc for my research. With my experiment, I found out 
> > recently that GC_mem_found sometimes becomes minus. I've 
> looked at it 
> > a while, and think it might be an actual bug.
> 
> I have noticed this as well, but without finding the reason.
> _______________________________________________
> Gc mailing list
> Gc at linux.hpl.hp.com
> http://www.hpl.hp.com/hosted/linux/mail-archives/gc/
> 



More information about the Gc mailing list