[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Y8Ds9JbWNGlWnoj9@infradead.org>
Date: Thu, 12 Jan 2023 21:32:36 -0800
From: Christoph Hellwig <hch@...radead.org>
To: Bart Van Assche <bart.vanassche@...il.com>
Cc: Al Viro <viro@...iv.linux.org.uk>,
Christoph Hellwig <hch@...radead.org>,
David Howells <dhowells@...hat.com>,
Matthew Wilcox <willy@...radead.org>,
Jens Axboe <axboe@...nel.dk>, Jan Kara <jack@...e.cz>,
Jeff Layton <jlayton@...nel.org>,
Logan Gunthorpe <logang@...tatee.com>,
linux-fsdevel@...r.kernel.org, linux-block@...r.kernel.org,
linux-kernel@...r.kernel.org,
"Martin K. Petersen" <martin.petersen@...cle.com>,
Douglas Gilbert <dgilbert@...erlog.com>
Subject: Re: [PATCH v5 3/9] iov_iter: Use IOCB/IOMAP_WRITE if available
rather than iterator direction
On Thu, Jan 12, 2023 at 01:49:14PM -0800, Bart Van Assche wrote:
> I'm not sure that we still need the double copy in the sg driver. It seems
> obscure to me that there is user space software that relies on finding
> "0xec" in bytes not originating from a SCSI device. Additionally, SCSI
> drivers that do not support residuals should be something from the past.
>
> Others may be better qualified to comment on this topic.
Yeah. And that weird (ab)use of blk_rq_map_user_iov into preallocated
pages in sg has been a constant source of pain. I'd be very happy to
make it match the generic SG_IO implementation and either do
get_user_pages or the normal bounce buffering implemented in the
common code.
Powered by blists - more mailing lists