From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jens Axboe Subject: Re: unify SCSI_IOCTL_SEND_COMMAND implementations Date: Thu, 23 Mar 2006 16:43:05 +0100 Message-ID: <20060323154304.GA27733@suse.de> References: <20060322165204.GA32287@lst.de> <20060323101819.GV4285@suse.de> <20060323144800.GA27044@lst.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from ns.virtualhost.dk ([195.184.98.160]:15953 "EHLO virtualhost.dk") by vger.kernel.org with ESMTP id S932434AbWCWPnA (ORCPT ); Thu, 23 Mar 2006 10:43:00 -0500 Content-Disposition: inline In-Reply-To: <20060323144800.GA27044@lst.de> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: Christoph Hellwig Cc: jejb@steeleye.com, linux-scsi@vger.kernel.org On Thu, Mar 23 2006, Christoph Hellwig wrote: > On Thu, Mar 23, 2006 at 11:18:19AM +0100, Jens Axboe wrote: > > > the non-S/G I/O path. Jens, what do you think about adding a check > > > for non-S/G I/O in the midlayer? > > > > Looks good to me. Are cdrom/sr sg clean as well? > > there's just one place in cdrom.c that sends a request directly: > cdrom_read_cdda_bpc, which sends a S/G BLOCK_PC one. Everything else > goes through ->packet_command which is implemented as oldstyle-PC > requests in IDE and S/G BLOCK_PC in scsi. all scsi-internal requests > including those from sr are S/G of BLOCK_PC type. Ok thanks for confirming, then lets just add that non-sg check in the mid layer. -- Jens Axboe