From mboxrd@z Thu Jan 1 00:00:00 1970 From: Hans Reiser Subject: Re: file as a directory Date: Fri, 17 Dec 2004 08:54:16 -0800 Message-ID: <41C30F38.6090407@namesys.com> References: <200411301631.iAUGVT8h007823@laptop11.inf.utfsm.cl> <41ACA7C9.1070001@namesys.com> <1103043518.21728.159.camel@pear.st-and.ac.uk> <41BF21BC.1020809@namesys.com> <1103059622.2999.17.camel@grape.st-and.ac.uk> <41BFC1C5.1070302@slaphack.com> <1103102854.30601.12.camel@pear.st-and.ac.uk> <41C0CF3B.1030705@slaphack.com> <41C1D870.2020407@namesys.com> <41C30325.4040604@slaphack.com> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Return-path: list-help: list-unsubscribe: list-post: Errors-To: flx@namesys.com In-Reply-To: <41C30325.4040604@slaphack.com> List-Id: Content-Type: text/plain; charset="us-ascii"; format="flowed" To: David Masover Cc: Peter Foldiak , reiserfs-list@namesys.com, linux-kernel@vger.kernel.org David Masover wrote: > Hans Reiser wrote: > | David Masover wrote: > | > |> > |> > |> Speaking of which, how much speed is lost by starting up a process? > |> > |> The idea of caching is that running > |> > |> cat *; cat *; cat *; cat *; cat * > |> > |> is probably slower than > |> > |> cat * > baz; cat baz; cat baz; cat baz; cat baz; cat baz > | > | > | Only for small files where the per file overhead of a read is > significant. > > That's potentially a common problem, and "cat *" is an overly-simplified > example. Either you force the "plugin" to say whether it wants to be > cached or not, or you cache everything, because there are going to be > plugins like "tar -xjp" for which caching is a HUGE increase. Ok, I accept the argument.