From patchwork Wed Oct 7 04:58:25 2015 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Darrick J. Wong" X-Patchwork-Id: 7342151 Return-Path: X-Original-To: patchwork-linux-fsdevel@patchwork.kernel.org Delivered-To: patchwork-parsemail@patchwork2.web.kernel.org Received: from mail.kernel.org (mail.kernel.org [198.145.29.136]) by patchwork2.web.kernel.org (Postfix) with ESMTP id ABEA6BEEA4 for ; Wed, 7 Oct 2015 04:58:33 +0000 (UTC) Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id 949C52066D for ; Wed, 7 Oct 2015 04:58:32 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 47A4A20504 for ; Wed, 7 Oct 2015 04:58:31 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752426AbbJGE6a (ORCPT ); Wed, 7 Oct 2015 00:58:30 -0400 Received: from aserp1040.oracle.com ([141.146.126.69]:36176 "EHLO aserp1040.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752375AbbJGE63 (ORCPT ); Wed, 7 Oct 2015 00:58:29 -0400 Received: from aserv0022.oracle.com (aserv0022.oracle.com [141.146.126.234]) by aserp1040.oracle.com (Sentrion-MTA-4.3.2/Sentrion-MTA-4.3.2) with ESMTP id t974wRXV007384 (version=TLSv1 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Wed, 7 Oct 2015 04:58:27 GMT Received: from aserv0121.oracle.com (aserv0121.oracle.com [141.146.126.235]) by aserv0022.oracle.com (8.13.8/8.13.8) with ESMTP id t974wR4r001726 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=FAIL); Wed, 7 Oct 2015 04:58:27 GMT Received: from abhmp0012.oracle.com (abhmp0012.oracle.com [141.146.116.18]) by aserv0121.oracle.com (8.13.8/8.13.8) with ESMTP id t974wRkj003435; Wed, 7 Oct 2015 04:58:27 GMT Received: from localhost (/24.21.154.84) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Tue, 06 Oct 2015 21:58:26 -0700 Subject: [PATCH 30/58] xfs: implement XFS_IOC_SWAPEXT when rmap btree is enabled From: "Darrick J. Wong" To: david@fromorbit.com, darrick.wong@oracle.com Cc: linux-fsdevel@vger.kernel.org, xfs@oss.sgi.com Date: Tue, 06 Oct 2015 21:58:25 -0700 Message-ID: <20151007045825.30457.37098.stgit@birch.djwong.org> In-Reply-To: <20151007045443.30457.47038.stgit@birch.djwong.org> References: <20151007045443.30457.47038.stgit@birch.djwong.org> User-Agent: StGit/0.17.1-dirty MIME-Version: 1.0 X-Source-IP: aserv0022.oracle.com [141.146.126.234] Sender: linux-fsdevel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-fsdevel@vger.kernel.org X-Spam-Status: No, score=-6.9 required=5.0 tests=BAYES_00, RCVD_IN_DNSWL_HI, T_RP_MATCHES_RCVD, UNPARSEABLE_RELAY autolearn=unavailable version=3.3.1 X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on mail.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP Implement extent swapping when reverse-mapping is enabled. Signed-off-by: Darrick J. Wong --- fs/xfs/libxfs/xfs_btree.c | 17 ++++++ fs/xfs/libxfs/xfs_rmap.c | 112 ++++++++++++++++++++++++++++++++++++++++ fs/xfs/libxfs/xfs_rmap_btree.h | 8 +++ fs/xfs/xfs_bmap_util.c | 38 +++++++++----- 4 files changed, 161 insertions(+), 14 deletions(-) -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c index 13971c6..2d7dc8c 100644 --- a/fs/xfs/libxfs/xfs_btree.c +++ b/fs/xfs/libxfs/xfs_btree.c @@ -32,6 +32,7 @@ #include "xfs_trace.h" #include "xfs_cksum.h" #include "xfs_alloc.h" +#include "xfs_rmap_btree.h" /* * Cursor allocation zone. @@ -3988,6 +3989,8 @@ xfs_btree_block_change_owner( struct xfs_btree_block *block; struct xfs_buf *bp; union xfs_btree_ptr rptr; + struct xfs_owner_info old_oinfo, new_oinfo; + int error; /* do right sibling readahead */ xfs_btree_readahead(cur, level, XFS_BTCUR_RIGHTRA); @@ -3999,6 +4002,20 @@ xfs_btree_block_change_owner( else block->bb_u.s.bb_owner = cpu_to_be32(new_owner); + /* change rmap owners (bmbt blocks only) */ + if (cur->bc_flags & XFS_BTREE_LONG_PTRS) { + XFS_RMAP_INO_BMBT_OWNER(&old_oinfo, + cur->bc_private.b.ip->i_ino, + cur->bc_private.b.whichfork); + XFS_RMAP_INO_BMBT_OWNER(&new_oinfo, + new_owner, + cur->bc_private.b.whichfork); + error = xfs_rmap_change_bmbt_owner(cur, bp, &old_oinfo, + &new_oinfo); + if (error) + return error; + } + /* * If the block is a root block hosted in an inode, we might not have a * buffer pointer here and we shouldn't attempt to log the change as the diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c index d821b1a..b2781f5 100644 --- a/fs/xfs/libxfs/xfs_rmap.c +++ b/fs/xfs/libxfs/xfs_rmap.c @@ -35,6 +35,8 @@ #include "xfs_trace.h" #include "xfs_error.h" #include "xfs_extent_busy.h" +#include "xfs_bmap.h" +#include "xfs_bmap_btree.h" /* * Lookup the first record less than or equal to [bno, len, owner, offset] @@ -859,3 +861,113 @@ xfs_rmap_resize( done: return error; } + +/** + * Change ownership of a file's BMBT block reverse-mappings. + */ +int +xfs_rmap_change_bmbt_owner( + struct xfs_btree_cur *bcur, + struct xfs_buf *bp, + struct xfs_owner_info *old_owner, + struct xfs_owner_info *new_owner) +{ + struct xfs_buf *agfbp; + xfs_fsblock_t fsbno; + xfs_agnumber_t agno; + xfs_agblock_t agbno; + int error; + + if (!xfs_sb_version_hasrmapbt(&bcur->bc_mp->m_sb) || !bp) + return 0; + + fsbno = XFS_DADDR_TO_FSB(bcur->bc_mp, XFS_BUF_ADDR(bp)); + agno = XFS_FSB_TO_AGNO(bcur->bc_mp, fsbno); + agbno = XFS_FSB_TO_AGBNO(bcur->bc_mp, fsbno); + + error = xfs_read_agf(bcur->bc_mp, bcur->bc_tp, agno, 0, &agfbp); + + error = xfs_rmap_free(bcur->bc_tp, agfbp, agno, agbno, 1, old_owner); + if (error) + goto err; + + error = xfs_rmap_alloc(bcur->bc_tp, agfbp, agno, agbno, 1, new_owner); + if (error) + goto err; + +err: + xfs_trans_brelse(bcur->bc_tp, agfbp); + return error; +} + +/** + * Change the ownership on a file's extent's reverse-mappings. + */ +int +xfs_rmap_change_extent_owner( + struct xfs_mount *mp, + struct xfs_inode *ip, + xfs_ino_t ino, + xfs_fileoff_t isize, + struct xfs_trans *tp, + int whichfork, + xfs_ino_t new_owner) +{ + struct xfs_bmbt_irec imap; + struct xfs_btree_cur *cur = NULL; + struct xfs_buf *agfbp = NULL; + int nimaps; + xfs_fileoff_t offset; + xfs_filblks_t len; + xfs_agnumber_t agno; + int flags = 0; + int error; + + if (!xfs_sb_version_hasrmapbt(&mp->m_sb)) + return 0; + + if (whichfork == XFS_ATTR_FORK) + flags |= XFS_BMAPI_ATTRFORK; + + offset = 0; + len = XFS_B_TO_FSB(mp, isize); + nimaps = 1; + error = xfs_bmapi_read(ip, offset, len, &imap, &nimaps, flags); + while (error == 0 && nimaps > 0) { + if (imap.br_startblock == HOLESTARTBLOCK || + imap.br_startblock == DELAYSTARTBLOCK) + goto advloop; + + agno = XFS_FSB_TO_AGNO(mp, imap.br_startblock); + + error = xfs_read_agf(mp, tp, agno, 0, &agfbp); + if (error) + break; + + cur = xfs_rmapbt_init_cursor(mp, tp, agfbp, agno); + + error = xfs_rmap_delete(cur, ino, whichfork, &imap); + if (error) + break; + error = xfs_rmap_insert(cur, new_owner, whichfork, &imap); + if (error) + break; + + xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR); + cur = NULL; + xfs_trans_brelse(tp, agfbp); + agfbp = NULL; +advloop: + offset += imap.br_blockcount; + len -= imap.br_blockcount; + nimaps = 1; + error = xfs_bmapi_read(ip, offset, len, &imap, &nimaps, flags); + } + + if (cur) + xfs_btree_del_cursor(cur, error ? XFS_BTREE_ERROR : + XFS_BTREE_NOERROR); + if (agfbp) + xfs_trans_brelse(tp, agfbp); + return error; +} diff --git a/fs/xfs/libxfs/xfs_rmap_btree.h b/fs/xfs/libxfs/xfs_rmap_btree.h index 0131d9a..5d248b5 100644 --- a/fs/xfs/libxfs/xfs_rmap_btree.h +++ b/fs/xfs/libxfs/xfs_rmap_btree.h @@ -88,4 +88,12 @@ int xfs_rmap_slide(struct xfs_btree_cur *rcur, xfs_ino_t ino, int whichfork, int xfs_rmap_resize(struct xfs_btree_cur *rcur, xfs_ino_t ino, int whichfork, struct xfs_bmbt_irec *PREV, long size_adj); +/* functions for changing rmap ownership */ +int xfs_rmap_change_extent_owner(struct xfs_mount *mp, struct xfs_inode *ip, + xfs_ino_t ino, xfs_fileoff_t isize, struct xfs_trans *tp, + int whichfork, xfs_ino_t new_owner); +int xfs_rmap_change_bmbt_owner(struct xfs_btree_cur *bcur, struct xfs_buf *bp, + struct xfs_owner_info *old_owner, + struct xfs_owner_info *new_owner); + #endif /* __XFS_RMAP_BTREE_H__ */ diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c index f41a6f7..245a34a 100644 --- a/fs/xfs/xfs_bmap_util.c +++ b/fs/xfs/xfs_bmap_util.c @@ -40,6 +40,7 @@ #include "xfs_trace.h" #include "xfs_icache.h" #include "xfs_log.h" +#include "xfs_rmap_btree.h" /* Kernel only BMAP related definitions and functions */ @@ -1668,6 +1669,17 @@ xfs_swap_extent_flush( return 0; } +static int +change_extent_owner( + struct xfs_inode *ip, + struct xfs_trans *tp, + int whichfork, + struct xfs_inode *tip) +{ + return xfs_rmap_change_extent_owner(ip->i_mount, ip, ip->i_ino, + ip->i_d.di_size, tp, whichfork, tip->i_ino); +} + int xfs_swap_extents( xfs_inode_t *ip, /* target inode */ @@ -1684,19 +1696,7 @@ xfs_swap_extents( int taforkblks = 0; __uint64_t tmp; int lock_flags; - - /* - * We can't swap extents on rmap btree enabled filesystems yet - * as there is no mechanism to update the owner of extents in - * the rmap tree yet. Hence, for the moment, just reject attempts - * to swap extents with EINVAL after emitting a warning once to remind - * us this needs fixing. - */ - if (xfs_sb_version_hasrmapbt(&mp->m_sb)) { - WARN_ONCE(1, - "XFS: XFS_IOC_SWAPEXT not supported on RMAP enabled filesystems\n"); - return -EINVAL; - } + struct xfs_trans_res tres = {.tr_logres = 262144, .tr_logcount = 1, .tr_logflags = 0}; tempifp = kmem_alloc(sizeof(xfs_ifork_t), KM_MAYFAIL); if (!tempifp) { @@ -1734,7 +1734,9 @@ xfs_swap_extents( goto out_unlock; tp = xfs_trans_alloc(mp, XFS_TRANS_SWAPEXT); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_ichange, 0, 0); + /* XXX How do we create a potentially huge transaction here? */ + /* error = xfs_trans_reserve(tp, &M_RES(mp)->tr_ichange, 0, 0); */ + error = xfs_trans_reserve(tp, &tres, 0, 0); if (error) { xfs_trans_cancel(tp); goto out_unlock; @@ -1834,6 +1836,14 @@ xfs_swap_extents( goto out_trans_cancel; } + /* Change owners in the extent rmaps */ + error = change_extent_owner(ip, tp, XFS_DATA_FORK, tip); + if (error) + goto out_trans_cancel; + error = change_extent_owner(tip, tp, XFS_DATA_FORK, ip); + if (error) + goto out_trans_cancel; + /* * Swap the data forks of the inodes */