From mboxrd@z Thu Jan 1 00:00:00 1970 Date: Sun, 9 Jul 2000 14:11:09 -0300 (BRT) From: Marcelo Tosatti Subject: Swap clustering with new VM In-Reply-To: <20000706142945.A4237@redhat.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-mm@kvack.org Return-Path: To: "Stephen C. Tweedie" Cc: Andrea Arcangeli , Jens Axboe , Alan Cox , Linux Kernel , linux-mm@kvack.org, "David S. Miller" , Rik van Riel List-ID: On Thu, 6 Jul 2000, Stephen C. Tweedie wrote: > For example, our swap clustering relies on allocating > sequential swap addresses to sequentially scanned VM addresses, so > that clustered swapout and swapin work naturally. Switch to > physically-ordered swapping and there's no longer any natural way of > getting the on-disk swap related to VA ordering, so that swapin > clustering breaks completely. To fix this, you need the final swapout > to try to swap nearby pages in VA space at the same time. It's a lot > of work to get it right. AFAIK XFS's pagebuf structure contains a list of contiguous on-disk buffers, so the filesystem can do IO on a pagebuf structure avoiding disk seek time. Do you plan to fix the swap clustering problem with a similar idea? -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux.eu.org/Linux-MM/