From patchwork Thu Oct 12 01:43:00 2017 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Darrick J. Wong" X-Patchwork-Id: 10000931 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork.web.codeaurora.org (Postfix) with ESMTP id EA04A60216 for ; Thu, 12 Oct 2017 01:43:05 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id E1E7828C1B for ; Thu, 12 Oct 2017 01:43:05 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id D6DF328C1D; Thu, 12 Oct 2017 01:43:05 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-6.9 required=2.0 tests=BAYES_00, RCVD_IN_DNSWL_HI, UNPARSEABLE_RELAY autolearn=ham version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 6876428C1B for ; Thu, 12 Oct 2017 01:43:04 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753134AbdJLBnD (ORCPT ); Wed, 11 Oct 2017 21:43:03 -0400 Received: from aserp1040.oracle.com ([141.146.126.69]:30656 "EHLO aserp1040.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752807AbdJLBnD (ORCPT ); Wed, 11 Oct 2017 21:43:03 -0400 Received: from userv0022.oracle.com (userv0022.oracle.com [156.151.31.74]) by aserp1040.oracle.com (Sentrion-MTA-4.3.2/Sentrion-MTA-4.3.2) with ESMTP id v9C1h1Jf028808 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=OK) for ; Thu, 12 Oct 2017 01:43:02 GMT Received: from aserv0121.oracle.com (aserv0121.oracle.com [141.146.126.235]) by userv0022.oracle.com (8.14.4/8.14.4) with ESMTP id v9C1h1Bn007840 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=OK) for ; Thu, 12 Oct 2017 01:43:01 GMT Received: from abhmp0017.oracle.com (abhmp0017.oracle.com [141.146.116.23]) by aserv0121.oracle.com (8.14.4/8.13.8) with ESMTP id v9C1h1dx029184 for ; Thu, 12 Oct 2017 01:43:01 GMT Received: from localhost (/73.25.142.12) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Wed, 11 Oct 2017 18:43:00 -0700 Subject: [PATCH 21/30] xfs: scrub inodes From: "Darrick J. Wong" To: darrick.wong@oracle.com Cc: linux-xfs@vger.kernel.org Date: Wed, 11 Oct 2017 18:43:00 -0700 Message-ID: <150777258004.1724.14107466871314594061.stgit@magnolia> In-Reply-To: <150777244315.1724.6916081372861799350.stgit@magnolia> References: <150777244315.1724.6916081372861799350.stgit@magnolia> User-Agent: StGit/0.17.1-dirty MIME-Version: 1.0 X-Source-IP: userv0022.oracle.com [156.151.31.74] Sender: linux-xfs-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-xfs@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP From: Darrick J. Wong Scrub the fields within an inode. Signed-off-by: Darrick J. Wong --- fs/xfs/Makefile | 1 fs/xfs/libxfs/xfs_fs.h | 3 fs/xfs/scrub/common.c | 54 ++++ fs/xfs/scrub/common.h | 3 fs/xfs/scrub/inode.c | 607 ++++++++++++++++++++++++++++++++++++++++++++++++ fs/xfs/scrub/scrub.c | 18 + fs/xfs/scrub/scrub.h | 2 7 files changed, 685 insertions(+), 3 deletions(-) create mode 100644 fs/xfs/scrub/inode.c -- To unsubscribe from this list: send the line "unsubscribe linux-xfs" 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/Makefile b/fs/xfs/Makefile index a7c5752..28e14b7 100644 --- a/fs/xfs/Makefile +++ b/fs/xfs/Makefile @@ -151,6 +151,7 @@ xfs-y += $(addprefix scrub/, \ btree.o \ common.o \ ialloc.o \ + inode.o \ refcount.o \ rmap.o \ scrub.o \ diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h index b3f992c..f8463e0 100644 --- a/fs/xfs/libxfs/xfs_fs.h +++ b/fs/xfs/libxfs/xfs_fs.h @@ -494,9 +494,10 @@ struct xfs_scrub_metadata { #define XFS_SCRUB_TYPE_FINOBT 8 /* free inode btree */ #define XFS_SCRUB_TYPE_RMAPBT 9 /* reverse mapping btree */ #define XFS_SCRUB_TYPE_REFCNTBT 10 /* reference count btree */ +#define XFS_SCRUB_TYPE_INODE 11 /* inode record */ /* Number of scrub subcommands. */ -#define XFS_SCRUB_TYPE_NR 11 +#define XFS_SCRUB_TYPE_NR 12 /* i: Repair this metadata. */ #define XFS_SCRUB_IFLAG_REPAIR (1 << 0) diff --git a/fs/xfs/scrub/common.c b/fs/xfs/scrub/common.c index 39165c3..415c6a9 100644 --- a/fs/xfs/scrub/common.c +++ b/fs/xfs/scrub/common.c @@ -30,6 +30,8 @@ #include "xfs_trans.h" #include "xfs_sb.h" #include "xfs_inode.h" +#include "xfs_icache.h" +#include "xfs_itable.h" #include "xfs_alloc.h" #include "xfs_alloc_btree.h" #include "xfs_bmap.h" @@ -488,3 +490,55 @@ xfs_scrub_checkpoint_log( xfs_ail_push_all_sync(mp->m_ail); return 0; } + +/* + * Given an inode and the scrub control structure, grab either the + * inode referenced in the control structure or the inode passed in. + * The inode is not locked. + */ +int +xfs_scrub_get_inode( + struct xfs_scrub_context *sc, + struct xfs_inode *ip_in) +{ + struct xfs_mount *mp = sc->mp; + struct xfs_inode *ip = NULL; + int error; + + /* + * If userspace passed us an AG number or a generation number + * without an inode number, they haven't got a clue so bail out + * immediately. + */ + if (sc->sm->sm_agno || (sc->sm->sm_gen && !sc->sm->sm_ino)) + return -EINVAL; + + /* We want to scan the inode we already had opened. */ + if (sc->sm->sm_ino == 0 || sc->sm->sm_ino == ip_in->i_ino) { + sc->ip = ip_in; + return 0; + } + + /* Look up the inode, see if the generation number matches. */ + if (xfs_internal_inum(mp, sc->sm->sm_ino)) + return -ENOENT; + error = xfs_iget(mp, NULL, sc->sm->sm_ino, + XFS_IGET_UNTRUSTED | XFS_IGET_DONTCACHE, 0, &ip); + if (error == -ENOENT || error == -EINVAL) { + /* inode doesn't exist... */ + return -ENOENT; + } else if (error) { + trace_xfs_scrub_op_error(sc, + XFS_INO_TO_AGNO(mp, sc->sm->sm_ino), + XFS_INO_TO_AGBNO(mp, sc->sm->sm_ino), + error, __return_address); + return error; + } + if (VFS_I(ip)->i_generation != sc->sm->sm_gen) { + iput(VFS_I(ip)); + return -ENOENT; + } + + sc->ip = ip; + return 0; +} diff --git a/fs/xfs/scrub/common.h b/fs/xfs/scrub/common.h index 610e956..fcec11e 100644 --- a/fs/xfs/scrub/common.h +++ b/fs/xfs/scrub/common.h @@ -87,6 +87,8 @@ int xfs_scrub_setup_ag_rmapbt(struct xfs_scrub_context *sc, struct xfs_inode *ip); int xfs_scrub_setup_ag_refcountbt(struct xfs_scrub_context *sc, struct xfs_inode *ip); +int xfs_scrub_setup_inode(struct xfs_scrub_context *sc, + struct xfs_inode *ip); void xfs_scrub_ag_free(struct xfs_scrub_context *sc, struct xfs_scrub_ag *sa); @@ -105,5 +107,6 @@ int xfs_scrub_walk_agfl(struct xfs_scrub_context *sc, int xfs_scrub_setup_ag_btree(struct xfs_scrub_context *sc, struct xfs_inode *ip, bool force_log); +int xfs_scrub_get_inode(struct xfs_scrub_context *sc, struct xfs_inode *ip_in); #endif /* __XFS_SCRUB_COMMON_H__ */ diff --git a/fs/xfs/scrub/inode.c b/fs/xfs/scrub/inode.c new file mode 100644 index 0000000..aa1c549 --- /dev/null +++ b/fs/xfs/scrub/inode.c @@ -0,0 +1,607 @@ +/* + * Copyright (C) 2017 Oracle. All Rights Reserved. + * + * Author: Darrick J. Wong + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version 2 + * of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_trans_resv.h" +#include "xfs_mount.h" +#include "xfs_defer.h" +#include "xfs_btree.h" +#include "xfs_bit.h" +#include "xfs_log_format.h" +#include "xfs_trans.h" +#include "xfs_sb.h" +#include "xfs_inode.h" +#include "xfs_icache.h" +#include "xfs_inode_buf.h" +#include "xfs_inode_fork.h" +#include "xfs_ialloc.h" +#include "xfs_da_format.h" +#include "xfs_reflink.h" +#include "scrub/xfs_scrub.h" +#include "scrub/scrub.h" +#include "scrub/common.h" +#include "scrub/trace.h" + +/* + * Grab total control of the inode metadata. It doesn't matter here if + * the file data is still changing; exclusive access to the metadata is + * the goal. + */ +int +xfs_scrub_setup_inode( + struct xfs_scrub_context *sc, + struct xfs_inode *ip) +{ + struct xfs_mount *mp = sc->mp; + int error; + + /* + * Try to get the inode. If the verifiers fail, we try again + * in raw mode. + */ + error = xfs_scrub_get_inode(sc, ip); + switch (error) { + case 0: + break; + case -EFSCORRUPTED: + case -EFSBADCRC: + return 0; + default: + return error; + } + + /* Got the inode, lock it and we're ready to go. */ + sc->ilock_flags = XFS_IOLOCK_EXCL | XFS_MMAPLOCK_EXCL; + xfs_ilock(sc->ip, sc->ilock_flags); + error = xfs_scrub_trans_alloc(sc->sm, mp, &sc->tp); + if (error) + goto out; + sc->ilock_flags |= XFS_ILOCK_EXCL; + xfs_ilock(sc->ip, XFS_ILOCK_EXCL); + +out: + /* scrub teardown will unlock and release the inode for us */ + return error; +} + +/* Inode core */ + +/* + * di_extsize hint validation is somewhat cumbersome. Rules are: + * + * 1. extent size hint is only valid for directories and regular files + * 2. DIFLAG_EXTSIZE is only valid for regular files + * 3. DIFLAG_EXTSZINHERIT is only valid for directories. + * 4. extsize hint of 0 turns off hints, clears inode flags. + * 5. either flag must be set if extsize != 0 + * 6. Extent size must be a multiple of the appropriate block size. + * 7. extent size hint cannot be longer than maximum extent length + * 8. for non-realtime files, the extent size hint must be limited + * to half the AG size to avoid alignment extending the extent + * beyond the limits of the AG. + */ +STATIC void +xfs_scrub_inode_extsize( + struct xfs_scrub_context *sc, + struct xfs_buf *bp, + struct xfs_dinode *dip, + xfs_ino_t ino, + uint16_t mode, + uint16_t flags) +{ + struct xfs_mount *mp = sc->mp; + bool rt_flag; + bool hint_flag; + bool inherit_flag; + uint32_t extsize; + uint32_t extsize_bytes; + uint32_t blocksize_bytes; + + rt_flag = (flags & XFS_DIFLAG_REALTIME); + hint_flag = (flags & XFS_DIFLAG_EXTSIZE); + inherit_flag = (flags & XFS_DIFLAG_EXTSZINHERIT); + extsize = be32_to_cpu(dip->di_extsize); + extsize_bytes = XFS_FSB_TO_B(sc->mp, extsize); + + if (rt_flag) + blocksize_bytes = mp->m_sb.sb_rextsize << mp->m_sb.sb_blocklog; + else + blocksize_bytes = mp->m_sb.sb_blocksize; + + if ((hint_flag || inherit_flag) && (!S_ISDIR(mode) && !S_ISREG(mode))) + goto bad; + + if (hint_flag && !S_ISREG(mode)) + goto bad; + + if (inherit_flag && !S_ISDIR(mode)) + goto bad; + + if ((hint_flag || inherit_flag) && extsize == 0) + goto bad; + + if (!(hint_flag || inherit_flag) && extsize != 0) + goto bad; + + if (extsize_bytes % blocksize_bytes) + goto bad; + + if (extsize > MAXEXTLEN) + goto bad; + + if (!rt_flag && extsize > mp->m_sb.sb_agblocks / 2) + goto bad; + + return; +bad: + xfs_scrub_ino_set_corrupt(sc, ino, bp); +} + +/* + * di_cowextsize hint validation is somewhat cumbersome. Rules are: + * + * 1. flag requires reflink feature + * 2. cow extent size hint is only valid for directories and regular files + * 3. cow extsize hint of 0 turns off hints, clears inode flags. + * 4. either flag must be set if cow extsize != 0 + * 5. flag cannot be set for rt files + * 6. Extent size must be a multiple of the appropriate block size. + * 7. extent size hint cannot be longer than maximum extent length + * 8. the extent size hint must be limited + * to half the AG size to avoid alignment extending the extent + * beyond the limits of the AG. + */ +STATIC void +xfs_scrub_inode_cowextsize( + struct xfs_scrub_context *sc, + struct xfs_buf *bp, + struct xfs_dinode *dip, + xfs_ino_t ino, + uint16_t mode, + uint16_t flags, + uint64_t flags2) +{ + struct xfs_mount *mp = sc->mp; + bool rt_flag; + bool hint_flag; + uint32_t extsize; + uint32_t extsize_bytes; + + rt_flag = (flags & XFS_DIFLAG_REALTIME); + hint_flag = (flags2 & XFS_DIFLAG2_COWEXTSIZE); + extsize = be32_to_cpu(dip->di_extsize); + extsize_bytes = XFS_FSB_TO_B(sc->mp, extsize); + + if (hint_flag && !xfs_sb_version_hasreflink(&mp->m_sb)) + goto bad; + + if (hint_flag && (!S_ISDIR(mode) && !S_ISREG(mode))) + goto bad; + + if (hint_flag && extsize == 0) + goto bad; + + if (!hint_flag && extsize != 0) + goto bad; + + if (hint_flag && rt_flag) + goto bad; + + if (extsize_bytes % mp->m_sb.sb_blocksize) + goto bad; + + if (extsize > MAXEXTLEN) + goto bad; + + if (extsize > mp->m_sb.sb_agblocks / 2) + goto bad; + + return; +bad: + xfs_scrub_ino_set_corrupt(sc, ino, bp); +} + +/* Make sure the di_flags make sense for the inode. */ +STATIC void +xfs_scrub_inode_flags( + struct xfs_scrub_context *sc, + struct xfs_buf *bp, + struct xfs_dinode *dip, + xfs_ino_t ino, + uint16_t mode, + uint16_t flags) +{ + struct xfs_mount *mp = sc->mp; + + if (flags & ~XFS_DIFLAG_ANY) + goto bad; + + /* rt flags require rt device */ + if ((flags & (XFS_DIFLAG_REALTIME | XFS_DIFLAG_RTINHERIT)) && + !mp->m_rtdev_targp) + goto bad; + + /* new rt bitmap flag only valid for rbmino */ + if ((flags & XFS_DIFLAG_NEWRTBM) && ino != mp->m_sb.sb_rbmino) + goto bad; + + /* directory-only flags */ + if ((flags & (XFS_DIFLAG_RTINHERIT | + XFS_DIFLAG_EXTSZINHERIT | + XFS_DIFLAG_PROJINHERIT | + XFS_DIFLAG_NOSYMLINKS)) && + !S_ISDIR(mode)) + goto bad; + + /* file-only flags */ + if ((flags & (XFS_DIFLAG_REALTIME | FS_XFLAG_EXTSIZE)) && + !S_ISREG(mode)) + goto bad; + + /* filestreams and rt make no sense */ + if ((flags & XFS_DIFLAG_FILESTREAM) && (flags & XFS_DIFLAG_REALTIME)) + goto bad; + + return; +bad: + xfs_scrub_ino_set_corrupt(sc, ino, bp); +} + +/* Make sure the di_flags2 make sense for the inode. */ +STATIC void +xfs_scrub_inode_flags2( + struct xfs_scrub_context *sc, + struct xfs_buf *bp, + struct xfs_dinode *dip, + xfs_ino_t ino, + uint16_t mode, + uint64_t flags2) +{ + struct xfs_mount *mp = sc->mp; + + if (flags2 & ~XFS_DIFLAG2_ANY) + goto bad; + + /* reflink flag requires reflink feature */ + if ((flags2 & XFS_DIFLAG2_REFLINK) && + !xfs_sb_version_hasreflink(&mp->m_sb)) + goto bad; + + /* cowextsize flag is checked w.r.t. mode separately */ + + /* file-only flags */ + if ((flags2 & (XFS_DIFLAG2_DAX | XFS_DIFLAG2_REFLINK)) && + !S_ISREG(mode)) + goto bad; + + /* dax and reflink make no sense, currently */ + if ((flags2 & XFS_DIFLAG2_DAX) && (flags2 & XFS_DIFLAG2_REFLINK)) + goto bad; + + return; +bad: + xfs_scrub_ino_set_corrupt(sc, ino, bp); +} + +/* Scrub all the ondisk inode fields. */ +STATIC void +xfs_scrub_dinode( + struct xfs_scrub_context *sc, + struct xfs_buf *bp, + struct xfs_dinode *dip, + xfs_ino_t ino) +{ + struct xfs_mount *mp = sc->mp; + size_t fork_recs; + unsigned long long isize; + uint64_t flags2; + uint32_t nextents; + uint16_t flags; + uint16_t mode; + + flags = be16_to_cpu(dip->di_flags); + if (dip->di_version >= 3) + flags2 = be64_to_cpu(dip->di_flags2); + else + flags2 = 0; + + /* di_mode */ + mode = be16_to_cpu(dip->di_mode); + if (mode & ~(S_IALLUGO | S_IFMT)) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + + /* v1/v2 fields */ + switch (dip->di_version) { + case 1: + /* + * We autoconvert v1 inodes into v2 inodes on writeout, + * so just mark this inode for preening. + */ + xfs_scrub_ino_set_preen(sc, bp); + break; + case 2: + case 3: + if (dip->di_onlink != 0) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + + if (dip->di_mode == 0 && sc->ip) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + + if (dip->di_projid_hi != 0 && + !xfs_sb_version_hasprojid32bit(&mp->m_sb)) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + break; + default: + xfs_scrub_ino_set_corrupt(sc, ino, bp); + return; + } + + /* + * di_uid/di_gid -- -1 isn't invalid, but there's no way that + * userspace could have created that. + */ + if (dip->di_uid == cpu_to_be32(-1U) || + dip->di_gid == cpu_to_be32(-1U)) + xfs_scrub_ino_set_warning(sc, bp); + + /* di_format */ + switch (dip->di_format) { + case XFS_DINODE_FMT_DEV: + if (!S_ISCHR(mode) && !S_ISBLK(mode) && + !S_ISFIFO(mode) && !S_ISSOCK(mode)) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + break; + case XFS_DINODE_FMT_LOCAL: + if (!S_ISDIR(mode) && !S_ISLNK(mode)) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + break; + case XFS_DINODE_FMT_EXTENTS: + if (!S_ISREG(mode) && !S_ISDIR(mode) && !S_ISLNK(mode)) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + break; + case XFS_DINODE_FMT_BTREE: + if (!S_ISREG(mode) && !S_ISDIR(mode)) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + break; + case XFS_DINODE_FMT_UUID: + default: + xfs_scrub_ino_set_corrupt(sc, ino, bp); + break; + } + + /* + * di_size. xfs_dinode_verify checks for things that screw up + * the VFS such as the upper bit being set and zero-length + * symlinks/directories, but we can do more here. + */ + isize = be64_to_cpu(dip->di_size); + if (isize & (1ULL << 63)) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + + /* Devices, fifos, and sockets must have zero size */ + if (!S_ISDIR(mode) && !S_ISREG(mode) && !S_ISLNK(mode) && isize != 0) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + + /* Directories can't be larger than the data section size (32G) */ + if (S_ISDIR(mode) && (isize == 0 || isize >= XFS_DIR2_SPACE_SIZE)) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + + /* Symlinks can't be larger than SYMLINK_MAXLEN */ + if (S_ISLNK(mode) && (isize == 0 || isize >= XFS_SYMLINK_MAXLEN)) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + + /* + * Warn if the running kernel can't handle the kinds of offsets + * needed to deal with the file size. In other words, if the + * pagecache can't cache all the blocks in this file due to + * overly large offsets, flag the inode for admin review. + */ + if (isize >= mp->m_super->s_maxbytes) + xfs_scrub_ino_set_warning(sc, bp); + + /* di_nblocks */ + if (flags2 & XFS_DIFLAG2_REFLINK) { + ; /* nblocks can exceed dblocks */ + } else if (flags & XFS_DIFLAG_REALTIME) { + /* + * nblocks is the sum of data extents (in the rtdev), + * attr extents (in the datadev), and both forks' bmbt + * blocks (in the datadev). This clumsy check is the + * best we can do without cross-referencing with the + * inode forks. + */ + if (be64_to_cpu(dip->di_nblocks) >= + mp->m_sb.sb_dblocks + mp->m_sb.sb_rblocks) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + } else { + if (be64_to_cpu(dip->di_nblocks) >= mp->m_sb.sb_dblocks) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + } + + xfs_scrub_inode_flags(sc, bp, dip, ino, mode, flags); + + xfs_scrub_inode_extsize(sc, bp, dip, ino, mode, flags); + + /* di_nextents */ + nextents = be32_to_cpu(dip->di_nextents); + fork_recs = XFS_DFORK_DSIZE(dip, mp) / sizeof(struct xfs_bmbt_rec); + switch (dip->di_format) { + case XFS_DINODE_FMT_EXTENTS: + if (nextents > fork_recs) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + break; + case XFS_DINODE_FMT_BTREE: + if (nextents <= fork_recs) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + break; + default: + if (nextents != 0) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + break; + } + + /* di_forkoff */ + if (XFS_DFORK_APTR(dip) >= (char *)dip + mp->m_sb.sb_inodesize) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + if (dip->di_anextents != 0 && dip->di_forkoff == 0) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + if (dip->di_forkoff == 0 && dip->di_aformat != XFS_DINODE_FMT_EXTENTS) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + + /* di_aformat */ + if (dip->di_aformat != XFS_DINODE_FMT_LOCAL && + dip->di_aformat != XFS_DINODE_FMT_EXTENTS && + dip->di_aformat != XFS_DINODE_FMT_BTREE) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + + /* di_anextents */ + nextents = be16_to_cpu(dip->di_anextents); + fork_recs = XFS_DFORK_ASIZE(dip, mp) / sizeof(struct xfs_bmbt_rec); + switch (dip->di_aformat) { + case XFS_DINODE_FMT_EXTENTS: + if (nextents > fork_recs) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + break; + case XFS_DINODE_FMT_BTREE: + if (nextents <= fork_recs) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + break; + default: + if (nextents != 0) + xfs_scrub_ino_set_corrupt(sc, ino, bp); + } + + if (flags2) + xfs_scrub_inode_flags2(sc, bp, dip, ino, mode, flags2); + + xfs_scrub_inode_cowextsize(sc, bp, dip, ino, mode, flags, flags2); +} + +/* Map and read a raw inode. */ +STATIC int +xfs_scrub_inode_map_raw( + struct xfs_scrub_context *sc, + xfs_ino_t ino, + struct xfs_buf **bpp, + struct xfs_dinode **dipp) +{ + struct xfs_imap imap; + struct xfs_mount *mp = sc->mp; + struct xfs_buf *bp; + struct xfs_dinode *dip; + int error; + + error = xfs_imap(mp, sc->tp, ino, &imap, XFS_IGET_UNTRUSTED); + if (error == -EINVAL) { + /* + * Inode could have gotten deleted out from under us; + * just forget about it. + */ + error = -ENOENT; + goto out; + } + if (!xfs_scrub_process_error(sc, XFS_INO_TO_AGNO(mp, ino), + XFS_INO_TO_AGBNO(mp, ino), &error)) + goto out; + + error = xfs_trans_read_buf(mp, sc->tp, mp->m_ddev_targp, + imap.im_blkno, imap.im_len, XBF_UNMAPPED, &bp, + NULL); + if (!xfs_scrub_process_error(sc, XFS_INO_TO_AGNO(mp, ino), + XFS_INO_TO_AGBNO(mp, ino), &error)) + goto out; + + /* Is this really an inode? */ + bp->b_ops = &xfs_inode_buf_ops; + dip = xfs_buf_offset(bp, imap.im_boffset); + if (!xfs_dinode_verify(mp, ino, dip) || + !xfs_dinode_good_version(mp, dip->di_version)) { + xfs_scrub_ino_set_corrupt(sc, ino, bp); + goto out; + } + + /* ...and is it the one we asked for? */ + if (be32_to_cpu(dip->di_gen) != sc->sm->sm_gen) { + error = -ENOENT; + goto out; + } + + *dipp = dip; + *bpp = bp; +out: + return error; +} + +/* Scrub an inode. */ +int +xfs_scrub_inode( + struct xfs_scrub_context *sc) +{ + struct xfs_dinode di; + struct xfs_mount *mp = sc->mp; + struct xfs_buf *bp = NULL; + struct xfs_dinode *dip; + xfs_ino_t ino; + + bool has_shared; + int error = 0; + + /* Did we get the in-core inode, or are we doing this manually? */ + if (sc->ip) { + ino = sc->ip->i_ino; + xfs_inode_to_disk(sc->ip, &di, 0); + dip = &di; + } else { + /* Map & read inode. */ + ino = sc->sm->sm_ino; + error = xfs_scrub_inode_map_raw(sc, ino, &bp, &dip); + if (error) + goto out; + } + + xfs_scrub_dinode(sc, bp, dip, ino); + if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT) + goto out; + + /* Now let's do the things that require a live inode. */ + if (!sc->ip) + goto out; + + /* + * Does this inode have the reflink flag set but no shared extents? + * Set the preening flag if this is the case. + */ + if (xfs_is_reflink_inode(sc->ip)) { + error = xfs_reflink_inode_has_shared_extents(sc->tp, sc->ip, + &has_shared); + if (!xfs_scrub_process_error(sc, XFS_INO_TO_AGNO(mp, ino), + XFS_INO_TO_AGBNO(mp, ino), &error)) + goto out; + if (!has_shared) + xfs_scrub_ino_set_preen(sc, bp); + } + +out: + if (bp) + xfs_trans_brelse(sc->tp, bp); + return error; +} diff --git a/fs/xfs/scrub/scrub.c b/fs/xfs/scrub/scrub.c index 10c9078..ab4209c 100644 --- a/fs/xfs/scrub/scrub.c +++ b/fs/xfs/scrub/scrub.c @@ -30,6 +30,8 @@ #include "xfs_trans.h" #include "xfs_sb.h" #include "xfs_inode.h" +#include "xfs_icache.h" +#include "xfs_itable.h" #include "xfs_alloc.h" #include "xfs_alloc_btree.h" #include "xfs_bmap.h" @@ -131,6 +133,7 @@ xfs_scrub_probe( STATIC int xfs_scrub_teardown( struct xfs_scrub_context *sc, + struct xfs_inode *ip_in, int error) { xfs_scrub_ag_free(sc, &sc->sa); @@ -138,6 +141,13 @@ xfs_scrub_teardown( xfs_trans_cancel(sc->tp); sc->tp = NULL; } + if (sc->ip) { + xfs_iunlock(sc->ip, sc->ilock_flags); + if (sc->ip != ip_in && + !xfs_internal_inum(sc->mp, sc->ip->i_ino)) + iput(VFS_I(sc->ip)); + sc->ip = NULL; + } return error; } @@ -191,6 +201,10 @@ static const struct xfs_scrub_meta_ops meta_scrub_ops[] = { .scrub = xfs_scrub_refcountbt, .has = xfs_sb_version_hasreflink, }, + { /* inode record */ + .setup = xfs_scrub_setup_inode, + .scrub = xfs_scrub_inode, + }, }; /* This isn't a stable feature, warn once per day. */ @@ -290,7 +304,7 @@ xfs_scrub_metadata( * Tear down everything we hold, then set up again with * preparation for worst-case scenarios. */ - error = xfs_scrub_teardown(&sc, 0); + error = xfs_scrub_teardown(&sc, ip, 0); if (error) goto out; try_harder = true; @@ -303,7 +317,7 @@ xfs_scrub_metadata( xfs_alert_ratelimited(mp, "Corruption detected during scrub."); out_teardown: - error = xfs_scrub_teardown(&sc, error); + error = xfs_scrub_teardown(&sc, ip, error); out: trace_xfs_scrub_done(ip, sm, error); return error; diff --git a/fs/xfs/scrub/scrub.h b/fs/xfs/scrub/scrub.h index 1c80bf5..ec635d4 100644 --- a/fs/xfs/scrub/scrub.h +++ b/fs/xfs/scrub/scrub.h @@ -59,6 +59,7 @@ struct xfs_scrub_context { const struct xfs_scrub_meta_ops *ops; struct xfs_trans *tp; struct xfs_inode *ip; + uint ilock_flags; bool try_harder; /* State tracking for single-AG operations. */ @@ -77,5 +78,6 @@ int xfs_scrub_inobt(struct xfs_scrub_context *sc); int xfs_scrub_finobt(struct xfs_scrub_context *sc); int xfs_scrub_rmapbt(struct xfs_scrub_context *sc); int xfs_scrub_refcountbt(struct xfs_scrub_context *sc); +int xfs_scrub_inode(struct xfs_scrub_context *sc); #endif /* __XFS_SCRUB_SCRUB_H__ */