[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

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

From: Thomas Bushnell BSG
Subject: Re: Review of Thomas's >2GB ext2fs proposal
Date: 17 Aug 2004 01:48:05 -0700
User-agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3

"Neal H. Walfield" <neal@cs.uml.edu> writes:

> If you didn't have this eviction strategy in mind for draining the
> mapping cache, I am curious what you were going to do.  It seems to me
> that anything else would be in far greater contention with the
> kernel's eviction strategy.  This is my analysis from the I sent on
> the 16th entitled "Review of Thomas's >2GB ext2fs proposal:"

Nothing is necessary.  You simply de-map, with vm_unmap.  Doing so
does not cause the page to be paged out by the kernel, and has
nothing whatsoever to do with the kernel's pageout strategy.  Unlike
paging, vm_map and vm_unmap are purely local operations where no
global optimization is relevant, so there is no reason to prefer
kernel control over them.

So I would simply advocate a straightforward LRU of such things, with
a reference count for the number of users of a mapping.  

The reference count, incidentally, you still need for your version,
don't you?  

If you drop mappings whenever the kernel evicts a page, what do you do
for pageouts of currently active pages?  How do you distinguish this
from a pageout of an idle page?


reply via email to

[Prev in Thread] Current Thread [Next in Thread]