From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from bombadil.infradead.org ([198.137.202.133]:33102 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1727332AbeKTAKQ (ORCPT ); Mon, 19 Nov 2018 19:10:16 -0500 Received: from 089144201193.atnat0010.highway.a1.net ([89.144.201.193] helo=localhost) by bombadil.infradead.org with esmtpsa (Exim 4.90_1 #2 (Red Hat Linux)) id 1gOjsq-0004uA-50 for linux-xfs@vger.kernel.org; Mon, 19 Nov 2018 13:46:36 +0000 From: Christoph Hellwig Subject: [PATCH 5/9] xfs: don't use delalloc extents for COW on files with extsize hints Date: Mon, 19 Nov 2018 14:46:15 +0100 Message-Id: <20181119134619.16812-6-hch@lst.de> In-Reply-To: <20181119134619.16812-1-hch@lst.de> References: <20181119134619.16812-1-hch@lst.de> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Sender: linux-xfs-owner@vger.kernel.org List-ID: List-Id: xfs To: linux-xfs@vger.kernel.org While using delalloc for extsize hints is generally a good idea, the current code that does so only for COW doesn't help us much and creates a lot of special cases. Switch it to use real allocations like we do for direct I/O. Signed-off-by: Christoph Hellwig --- fs/xfs/xfs_iomap.c | 28 +++++++++++++++++----------- fs/xfs/xfs_reflink.c | 5 ++++- fs/xfs/xfs_reflink.h | 5 ++--- 3 files changed, 23 insertions(+), 15 deletions(-) diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c index 2ece2363a2a0..f64155c1fac8 100644 --- a/fs/xfs/xfs_iomap.c +++ b/fs/xfs/xfs_iomap.c @@ -1047,22 +1047,28 @@ xfs_file_iomap_begin( * been done up front, so we don't need to do them here. */ if (xfs_is_reflink_inode(ip)) { + struct xfs_bmbt_irec orig = imap; + /* if zeroing doesn't need COW allocation, then we are done. */ if ((flags & IOMAP_ZERO) && !needs_cow_for_zeroing(&imap, nimaps)) goto out_found; - if (flags & IOMAP_DIRECT) { - /* may drop and re-acquire the ilock */ - error = xfs_reflink_allocate_cow(ip, &imap, &shared, - &lockmode); - if (error) - goto out_unlock; - } else { - error = xfs_reflink_reserve_cow(ip, &imap); - if (error) - goto out_unlock; - } + error = xfs_reflink_allocate_cow(ip, &imap, &shared, &lockmode, + flags); + if (error) + goto out_unlock; + + /* + * For buffered writes we need to report the address of the + * previous block (if there was any) so that the higher level + * write code can perform read-modify-write operations. For + * direct I/O code, which must be block aligned we need to + * report the newly allocated address. + */ + if (!(flags & IOMAP_DIRECT) && + orig.br_startblock != HOLESTARTBLOCK) + imap = orig; end_fsb = imap.br_startoff + imap.br_blockcount; length = XFS_FSB_TO_B(mp, end_fsb) - offset; diff --git a/fs/xfs/xfs_reflink.c b/fs/xfs/xfs_reflink.c index bfa489c8799f..5fe7126b2a3a 100644 --- a/fs/xfs/xfs_reflink.c +++ b/fs/xfs/xfs_reflink.c @@ -397,7 +397,8 @@ xfs_reflink_allocate_cow( struct xfs_inode *ip, struct xfs_bmbt_irec *imap, bool *shared, - uint *lockmode) + uint *lockmode, + unsigned flags) { struct xfs_mount *mp = ip->i_mount; xfs_fileoff_t offset_fsb = imap->br_startoff; @@ -471,6 +472,8 @@ xfs_reflink_allocate_cow( if (nimaps == 0) return -ENOSPC; convert: + if (!(flags & IOMAP_DIRECT)) + return 0; return xfs_reflink_convert_cow_extent(ip, imap, offset_fsb, count_fsb); out_unreserve: diff --git a/fs/xfs/xfs_reflink.h b/fs/xfs/xfs_reflink.h index 6d73daef1f13..d76fc520cac8 100644 --- a/fs/xfs/xfs_reflink.h +++ b/fs/xfs/xfs_reflink.h @@ -12,10 +12,9 @@ extern int xfs_reflink_find_shared(struct xfs_mount *mp, struct xfs_trans *tp, extern int xfs_reflink_trim_around_shared(struct xfs_inode *ip, struct xfs_bmbt_irec *irec, bool *shared); -extern int xfs_reflink_reserve_cow(struct xfs_inode *ip, - struct xfs_bmbt_irec *imap); extern int xfs_reflink_allocate_cow(struct xfs_inode *ip, - struct xfs_bmbt_irec *imap, bool *shared, uint *lockmode); + struct xfs_bmbt_irec *imap, bool *shared, uint *lockmode, + unsigned flags); extern int xfs_reflink_convert_cow(struct xfs_inode *ip, xfs_off_t offset, xfs_off_t count); -- 2.19.1