From mboxrd@z Thu Jan 1 00:00:00 1970 From: Chris Mason Subject: Re: [patch 0/2]btrfs: add two ioctls to do metadata readahead Date: Mon, 19 Jul 2010 07:25:56 -0400 Message-ID: <20100719112556.GW8623@think> References: <1279094536.1528.93.camel@sli10-desk.sh.intel.com> <20100719080354.GA12808@infradead.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Shaohua Li , "linux-btrfs@vger.kernel.org" , Arjan van de Ven , "Wu, Fengguang" To: Christoph Hellwig Return-path: In-Reply-To: <20100719080354.GA12808@infradead.org> List-ID: On Mon, Jul 19, 2010 at 04:03:54AM -0400, Christoph Hellwig wrote: > This seems like iteration 66 of the ill fated mobling readahead crap. > > Please go to linux-fsdevel to define a proper interface instead of > your per-filesystem hacks. I had hoped Arjan got it after the last big > flameware, but it seems like you're only moving from one target to the > next. > > NACK in this current form. Honestly I think its good to have some of these features in specific filesystems at first. Sure, it is easy to tell if they help in the lab and during development, but why go through the pain of implementing them in generic form until they have been through a broader test? So the idea for the feature is fine with me, especially for something like readahead which we can support forever by turning it into a noop if there turns out to be a better idea. -chris