From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Hellwig Subject: Re: [FSAIO][PATCH 7/8] Filesystem AIO read Date: Tue, 2 Jan 2007 14:29:01 +0000 Message-ID: <20070102142901.GB14954@infradead.org> References: <20061227153855.GA25898@in.ibm.com> <20061228082308.GA4476@in.ibm.com> <20061228084252.GG6971@in.ibm.com> <20061228115747.GB25644@infradead.org> <20061228151830.GB10156@in.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Christoph Hellwig , linux-aio@kvack.org, akpm@osdl.org, drepper@redhat.com, linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org, jakub@redhat.com, mingo@elte.hu Return-path: Received: from pentafluge.infradead.org ([213.146.154.40]:41155 "EHLO pentafluge.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755280AbXABO3J (ORCPT ); Tue, 2 Jan 2007 09:29:09 -0500 To: Suparna Bhattacharya Content-Disposition: inline In-Reply-To: <20061228151830.GB10156@in.ibm.com> Sender: linux-fsdevel-owner@vger.kernel.org List-Id: linux-fsdevel.vger.kernel.org On Thu, Dec 28, 2006 at 08:48:30PM +0530, Suparna Bhattacharya wrote: > Yes, we can do that -- how about aio_restarted() as an alternate name ? Sounds fine to me. > > Pluse possible naming updates discussed in the last mail. Also do we > > really need to pass current->io_wait here? Isn't the waitqueue in > > the kiocb always guaranteed to be the same? Now that all pagecache > > We don't have have the kiocb available to this routine. Using current->io_wait > avoids the need to pass the iocb down to deeper levels just for the sync vs > async checks, also allowing such routines to be shared by other code which > does not use iocbs (e.g. generic_file_sendfile->do_generic_file_read > ->do_generic_mapping_read) without having to set up dummy iocbs. We really want to switch senfile to kiocbs btw, - for one thing to allow an aio_sendfile implementation and second to make it more common to all the other I/O path code so we can avoid special cases in the fs code So I'm not convinced by that argument. But again we don't need to put the io_wait removal into your patchkit. I'll try to hack on it once I'll get a little spare time.