Re: Review of Thomas's >2GB ext2fs proposal

[Neal H. Walfield]

At 17 Aug 2004 03:05:43 -0700,
Thomas Bushnell BSG wrote:
> 
> "Neal H. Walfield" <neal@cs.uml.edu> writes:
> 
> > It is not a memory leak as it can still be reclaim.  It will be
> > reclaimed (because it has no references) when the mapping cache is
> > full.  Or, and perhaps this is a better approach, we add a check in
> > the release function to see if the kernel has a copy and if not, drop
> > it at that point.
> >
> > > (I think the solution is to have a proper interlock when you drop the
> > > last reference, but this is very tricky to get right.)
> > 
> > I am not sure what you envision here.
> 
> What I envision is a check when you release the last reference to drop
> it if the kernel doesn't have a copy.  I meant to say that such things
> can be tricky, whether it is depends on the details and what locking
> structure you have.
> 
> What happens if the mapping cache never fills because the kernel
> always pages things out in time?  The problem here is that you still
> have the data around taking up a real memory page; the kernel
> requested a pageout and you haven't freed the page.

Nope, the page is freed.  The only memory that is consumed is the
associations in the two hashes.  However, I have now become convinced
that the better strategy is to drop the association in the release
function if the kernel does not have a copy.

Neal