Date: Fri, 08 Jun 2001 00:44:49 -0700 From: Terry Lambert <tlambert2@mindspring.com> To: Ian Dowse <iedowse@maths.tcd.ie> Cc: Matt Dillon <dillon@earth.backplane.com>, freebsd-hackers@FreeBSD.ORG Subject: Re: UFS large directory performance Message-ID: <3B208271.9F1E93B1@mindspring.com> References: <200106031747.aa54373@salmon.maths.tcd.ie>
next in thread | previous in thread | raw e-mail | index | archive | help
Ian Dowse wrote: > Nice idea, but I'm not sure I see the benefit of partially reclaiming > second-level arrays. Because it is a hash array, there isn't really > the concept of a working set; a directory that is `in use' will > rarely see many create/rename/delete operations on a small fixed > set of filenames. The lookup case is already cached elsewhere. I > think an all-or-nothing approach is likely to perform better and > be simpler to implement. Even the lazy allocation of second-level > arrays is unlikely to help a lot if the hash function does its job > well. From this perspective, it seems to me that there could be significant benefit in getting rid of the ihash cache; but then I've always hated the thing, since there's no way to reassociate a vnode with valid, clean data hanging off it with an ihash entry, when you get an ihash cache hit. Totally off the dirpref topic, cut of the same philosophical bent that your statement above seems to advocate... -- Terry To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3B208271.9F1E93B1>