[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20200726151916.GC25328@infradead.org>
Date: Sun, 26 Jul 2020 16:19:16 +0100
From: Christoph Hellwig <hch@...radead.org>
To: Kanchan Joshi <joshi.k@...sung.com>
Cc: axboe@...nel.dk, viro@...iv.linux.org.uk, bcrl@...ck.org,
willy@...radead.org, hch@...radead.org, Damien.LeMoal@....com,
asml.silence@...il.com, linux-fsdevel@...r.kernel.org,
linux-kernel@...r.kernel.org, linux-aio@...ck.org,
io-uring@...r.kernel.org, linux-block@...r.kernel.org,
linux-api@...r.kernel.org, Selvakumar S <selvakuma.s1@...sung.com>,
Nitesh Shetty <nj.shetty@...sung.com>,
Arnav Dawn <a.dawn@...sung.com>,
Javier Gonzalez <javier.gonz@...sung.com>
Subject: Re: [PATCH v4 4/6] block: add zone append handling for direct I/O
path
On Fri, Jul 24, 2020 at 09:19:20PM +0530, Kanchan Joshi wrote:
> For zoned block device, opt in for zone-append by setting
> FMODE_ZONE_APPEND during open. Make direct IO submission path use
> IOCB_ZONE_APPEND to send bio with append op. Make direct IO completion
> return written-offset, in bytes, to upper layer via ret2 of
> kiocb->ki_complete interface.
> Write with the flag IOCB_ZONE_APPEND are ensured not be be short.
> Prevent short write and instead return failure if appending write spans
> beyond end of device.
> Return failure if write is larger than max_append_limit and therefore
> requires formation of multiple bios.
We should support reporting the append offset for all block devices
and all file systems support by iomap at least. There is nothing that
requires actual zone append support here.
Powered by blists - more mailing lists