From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Darrick J. Wong" Subject: Re: [PATCH v7 6/6] loop: Add support for REQ_ALLOCATE Date: Thu, 13 Feb 2020 10:11:41 -0800 Message-ID: <20200213181141.GT6874@magnolia> References: <158157930219.111879.12072477040351921368.stgit@localhost.localdomain> <158157957565.111879.5952051034259419400.stgit@localhost.localdomain> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <158157957565.111879.5952051034259419400.stgit@localhost.localdomain> Sender: linux-block-owner@vger.kernel.org To: Kirill Tkhai Cc: martin.petersen@oracle.com, bob.liu@oracle.com, axboe@kernel.dk, agk@redhat.com, snitzer@redhat.com, dm-devel@redhat.com, song@kernel.org, tytso@mit.edu, adilger.kernel@dilger.ca, Chaitanya.Kulkarni@wdc.com, ming.lei@redhat.com, osandov@fb.com, jthumshirn@suse.de, minwoo.im.dev@gmail.com, damien.lemoal@wdc.com, andrea.parri@amarulasolutions.com, hare@suse.com, tj@kernel.org, ajay.joshi@wdc.com, sagi@grimberg.me, dsterba@suse.com, bvanassche@acm.org, dhowells@redhat.com, asml.silence@gmail.com, linux-block@vger.kernel.org, linux-kernel@vger.kernel.org List-Id: dm-devel.ids On Thu, Feb 13, 2020 at 10:39:35AM +0300, Kirill Tkhai wrote: > Support for new modifier of REQ_OP_WRITE_ZEROES command. > This results in allocation extents in backing file instead > of actual blocks zeroing. > > Signed-off-by: Kirill Tkhai > Reviewed-by: Bob Liu > Reviewed-by: Darrick J. Wong > --- > drivers/block/loop.c | 20 +++++++++++++------- > 1 file changed, 13 insertions(+), 7 deletions(-) > > diff --git a/drivers/block/loop.c b/drivers/block/loop.c > index 739b372a5112..0704167a5aaa 100644 > --- a/drivers/block/loop.c > +++ b/drivers/block/loop.c > @@ -581,6 +581,16 @@ static int lo_rw_aio(struct loop_device *lo, struct loop_cmd *cmd, > return 0; > } > Urgh, I meant "copy the comment directly to here", e.g. /* * Convert REQ_OP_WRITE_ZEROES modifiers into fallocate mode * * If the caller requires allocation, select that mode. If the caller * doesn't want deallocation, call zeroout to write zeroes the range. * Otherwise, punch out the blocks. */ The goal here is to reinforce the notion that FL_ZERO_RANGE is how we achieve nounmap zeroing... --D > +static unsigned int write_zeroes_to_fallocate_mode(unsigned int flags) > +{ > + if (flags & REQ_ALLOCATE) > + return 0; > + if (flags & REQ_NOUNMAP) > + return FALLOC_FL_ZERO_RANGE; > + return FALLOC_FL_PUNCH_HOLE; > +} > + > static int do_req_filebacked(struct loop_device *lo, struct request *rq) > { > struct loop_cmd *cmd = blk_mq_rq_to_pdu(rq); > @@ -599,14 +609,8 @@ static int do_req_filebacked(struct loop_device *lo, struct request *rq) > case REQ_OP_FLUSH: > return lo_req_flush(lo, rq); > case REQ_OP_WRITE_ZEROES: > - /* > - * If the caller doesn't want deallocation, call zeroout to > - * write zeroes the range. Otherwise, punch them out. > - */ > return lo_fallocate(lo, rq, pos, > - (rq->cmd_flags & REQ_NOUNMAP) ? > - FALLOC_FL_ZERO_RANGE : > - FALLOC_FL_PUNCH_HOLE); > + write_zeroes_to_fallocate_mode(rq->cmd_flags)); > case REQ_OP_DISCARD: > return lo_fallocate(lo, rq, pos, FALLOC_FL_PUNCH_HOLE); > case REQ_OP_WRITE: > @@ -877,6 +881,7 @@ static void loop_config_discard(struct loop_device *lo) > q->limits.discard_alignment = 0; > blk_queue_max_discard_sectors(q, 0); > blk_queue_max_write_zeroes_sectors(q, 0); > + blk_queue_max_allocate_sectors(q, 0); > blk_queue_flag_clear(QUEUE_FLAG_DISCARD, q); > return; > } > @@ -886,6 +891,7 @@ static void loop_config_discard(struct loop_device *lo) > > blk_queue_max_discard_sectors(q, UINT_MAX >> 9); > blk_queue_max_write_zeroes_sectors(q, UINT_MAX >> 9); > + blk_queue_max_allocate_sectors(q, UINT_MAX >> 9); > blk_queue_flag_set(QUEUE_FLAG_DISCARD, q); > } > > >