From mboxrd@z Thu Jan 1 00:00:00 1970 From: "hch@lst.de" Subject: Re: block: add a error_count field to struct request Date: Wed, 19 Apr 2017 08:15:19 +0200 Message-ID: <20170419061519.GC6257@lst.de> References: <20170418155229.5977-1-hch@bombadil.infradead.org> <20170418155229.5977-18-hch@bombadil.infradead.org> <1492556229.2689.22.camel@sandisk.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <1492556229.2689.22.camel@sandisk.com> Sender: linux-block-owner@vger.kernel.org To: Bart Van Assche Cc: "hch@bombadil.infradead.org" , "axboe@kernel.dk" , "linux-block@vger.kernel.org" , "hch@lst.de" , "konrad.wilk@oracle.com" , "roger.pau@citrix.com" , "linux-scsi@vger.kernel.org" , "linux-nvme@lists.infradead.org" , "jbacik@fb.com" , "james.smart@broadcom.com" , "dm-devel@redhat.com" List-Id: linux-scsi@vger.kernel.org On Tue, Apr 18, 2017 at 10:57:11PM +0000, Bart Van Assche wrote: > Both blk-mq and the traditional block layer support a .cmd_size field to > make the block layer core allocate driver-specific data at the end of struct > request. Could that mechanism have been used for the error_count field? It could, and that's what I did for the modern drivers. It would have been a bit of a pain for these old floppy drivers, though.