Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261491AbVCIESH (ORCPT ); Tue, 8 Mar 2005 23:18:07 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261495AbVCIESH (ORCPT ); Tue, 8 Mar 2005 23:18:07 -0500 Received: from rgminet04.oracle.com ([148.87.122.33]:49810 "EHLO rgminet04.oracle.com") by vger.kernel.org with ESMTP id S261491AbVCIESB (ORCPT ); Tue, 8 Mar 2005 23:18:01 -0500 Date: Tue, 8 Mar 2005 20:07:57 -0800 From: Joel Becker To: Badari Pulavarty Cc: Daniel McNeil , Suparna Bhattacharya , Andrew Morton , S?bastien Dugu? , "linux-aio@kvack.org" , Linux Kernel Mailing List Subject: Re: [PATCH] 2.6.10 - direct-io async short read bug Message-ID: <20050309040757.GY27331@ca-server1.us.oracle.com> Mail-Followup-To: Badari Pulavarty , Daniel McNeil , Suparna Bhattacharya , Andrew Morton , S?bastien Dugu? , "linux-aio@kvack.org" , Linux Kernel Mailing List References: <1110189607.11938.14.camel@frecb000686> <20050307223917.1e800784.akpm@osdl.org> <20050308090946.GA4100@in.ibm.com> <1110302614.24286.61.camel@dyn318077bld.beaverton.ibm.com> <1110309508.24286.74.camel@dyn318077bld.beaverton.ibm.com> <1110324434.6521.23.camel@ibm-c.pdx.osdl.net> <1110326043.24286.134.camel@dyn318077bld.beaverton.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1110326043.24286.134.camel@dyn318077bld.beaverton.ibm.com> X-Burt-Line: Trees are cool. X-Red-Smith: Ninety feet between bases is perhaps as close as man has ever come to perfection. User-Agent: Mutt/1.5.6+20040907i X-Brightmail-Tracker: AAAAAQAAAAI= X-Whitelist: TRUE Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1713 Lines: 44 On Tue, Mar 08, 2005 at 03:54:04PM -0800, Badari Pulavarty wrote: > > 1. return EINVAL if the DIO goes past EOF. > > > > 2. truncate the request to file size (which is what your patch does) > > and if it works, it works. > > > > 3. truncate the request to a size that actually works - like a multiple > > of 512. > > > > 4. Do the full i/o since the user buffer is big enough, truncate the > > result returned to file size (and clear out the user buffer where it > > read past EOF). > > > > Number 4 would make it easy on the user-level code, but AIO DIO might be > > a bit tricky and might be a security hole since the data would be dma'ed > > there and then cleared. I need to look at the code some more. Solaris, which does forcedirectio as a mount option, actually will do buffered I/O on the trailing part. Consider it like a bounce buffer. That way they don't DMA the trailing data and succeed the I/O. The I/O returns actual bytes till EOF, just like read(2) is supposed to. Either this or a fully DMA'd number 4 is really what we should do. If security can only be solved via a bounce buffer, who cares? If the user created themselves a non-aligned file to open O_DIRECT, that's their problem if the last part-sector is negligably slower. Joel -- Life's Little Instruction Book #3 "Watch a sunrise at least once a year." Joel Becker Senior Member of Technical Staff Oracle E-mail: joel.becker@oracle.com Phone: (650) 506-8127 - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/