Date: Wed, 11 Aug 1999 13:51:58 -0700 (PDT) From: Matthew Dillon <dillon@apollo.backplane.com> To: John-Mark Gurney <gurney_j@efn.org> Cc: Jeff Aitken <jaitken@aitken.com>, des@flood.ping.uio.no, freebsd-security@FreeBSD.ORG Subject: Re: 4 Swap partitions limit (was Re: RE: Little question (offtopic)) Message-ID: <199908112051.NAA77694@apollo.backplane.com> References: <14257.50009.162402.381699@trooper.velocet.ca> <199908111851.OAA21052@eagle.aitken.com> <19990811120248.23702@hydrogen.fircrest.net>
next in thread | previous in thread | raw e-mail | index | archive | help
:this should also be updated to say that it doesn't dump and zero the
:pages... it just puts them on disk so if there is an immediate need
:for a large chunk of memory that it can dump pages that have been swapped
:out w/o having to wait for them to be swapped out..
:
:so the more use of swap actually helps performance because you can get
:large chunks of memory faster...
:
:--
: John-Mark Gurney Voice: +1 541 684 8449
: Cu Networking P.O. Box 5693, 97405
Most systems work this way. You have dirty pages, clean pages, and
free pages. The act of allocating swap and writing a dirty page to
the swap block simply changes the page from dirty to clean and does not
remove it. If the page is touched again prior to being reused, the
underlying swap allocation is simply thrown away. If the page is accessed
prior to being reused its LRU position is reset but the swap remains
allocated. If the page is reused and then later touched or accessed,
the page is loaded in from its swap backing store and (if the page is
being touched), the swap backing store is deallocated.
-Matt
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-security" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199908112051.NAA77694>