From 163467d3753e77e1d77da75727975cc3803a1dbc Mon Sep 17 00:00:00 2001 From: Zhi Yong Wu Date: Wed, 18 Dec 2013 08:22:39 +0800 Subject: xfs: factor prid related codes into xfs_get_initial_prid() It will be reused by the O_TMPFILE creation function. Reviewed-by: Christoph Hellwig Signed-off-by: Zhi Yong Wu Signed-off-by: Ben Myers diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 001aa89..c79b875 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -42,7 +42,6 @@ #include "xfs_bmap_util.h" #include "xfs_error.h" #include "xfs_quota.h" -#include "xfs_dinode.h" #include "xfs_filestream.h" #include "xfs_cksum.h" #include "xfs_trace.h" @@ -1169,10 +1168,7 @@ xfs_create( if (XFS_FORCED_SHUTDOWN(mp)) return XFS_ERROR(EIO); - if (dp->i_d.di_flags & XFS_DIFLAG_PROJINHERIT) - prid = xfs_get_projid(dp); - else - prid = XFS_PROJID_DEFAULT; + prid = xfs_get_initial_prid(dp); /* * Make sure that we have allocated dquot(s) on disk. diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h index 9e6efccb..6c58349 100644 --- a/fs/xfs/xfs_inode.h +++ b/fs/xfs/xfs_inode.h @@ -20,6 +20,7 @@ #include "xfs_inode_buf.h" #include "xfs_inode_fork.h" +#include "xfs_dinode.h" /* * Kernel only inode definitions @@ -192,6 +193,15 @@ xfs_set_projid(struct xfs_inode *ip, ip->i_d.di_projid_lo = (__uint16_t) (projid & 0xffff); } +static inline prid_t +xfs_get_initial_prid(struct xfs_inode *dp) +{ + if (dp->i_d.di_flags & XFS_DIFLAG_PROJINHERIT) + return xfs_get_projid(dp); + + return XFS_PROJID_DEFAULT; +} + /* * In-core inode flags. */ diff --git a/fs/xfs/xfs_symlink.c b/fs/xfs/xfs_symlink.c index 14e58f2..13140c7 100644 --- a/fs/xfs/xfs_symlink.c +++ b/fs/xfs/xfs_symlink.c @@ -208,10 +208,7 @@ xfs_symlink( return XFS_ERROR(ENAMETOOLONG); udqp = gdqp = NULL; - if (dp->i_d.di_flags & XFS_DIFLAG_PROJINHERIT) - prid = xfs_get_projid(dp); - else - prid = XFS_PROJID_DEFAULT; + prid = xfs_get_initial_prid(dp); /* * Make sure that we have allocated dquot(s) on disk. -- cgit v0.10.2 From 99b6436bc29e4f10e4388c27a3e4810191cc4788 Mon Sep 17 00:00:00 2001 From: Zhi Yong Wu Date: Wed, 18 Dec 2013 08:22:40 +0800 Subject: xfs: add O_TMPFILE support Add two functions xfs_create_tmpfile() and xfs_vn_tmpfile() to support O_TMPFILE file creation. In contrast to xfs_create(), xfs_create_tmpfile() has a different log reservation to the regular file creation because there is no directory modification, and doesn't check if an entry can be added to the directory, but the reservation quotas is required appropriately, and finally its inode is added to the unlinked list. xfs_vn_tmpfile() add one O_TMPFILE method to VFS interface and directly invoke xfs_create_tmpfile(). Signed-off-by: Zhi Yong Wu Reviewed-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Ben Myers diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index c79b875..ac133ea 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -1333,6 +1333,113 @@ xfs_create( } int +xfs_create_tmpfile( + struct xfs_inode *dp, + struct dentry *dentry, + umode_t mode) +{ + struct xfs_mount *mp = dp->i_mount; + struct xfs_inode *ip = NULL; + struct xfs_trans *tp = NULL; + int error; + uint cancel_flags = XFS_TRANS_RELEASE_LOG_RES; + prid_t prid; + struct xfs_dquot *udqp = NULL; + struct xfs_dquot *gdqp = NULL; + struct xfs_dquot *pdqp = NULL; + struct xfs_trans_res *tres; + uint resblks; + + if (XFS_FORCED_SHUTDOWN(mp)) + return XFS_ERROR(EIO); + + prid = xfs_get_initial_prid(dp); + + /* + * Make sure that we have allocated dquot(s) on disk. + */ + error = xfs_qm_vop_dqalloc(dp, xfs_kuid_to_uid(current_fsuid()), + xfs_kgid_to_gid(current_fsgid()), prid, + XFS_QMOPT_QUOTALL | XFS_QMOPT_INHERIT, + &udqp, &gdqp, &pdqp); + if (error) + return error; + + resblks = XFS_IALLOC_SPACE_RES(mp); + tp = xfs_trans_alloc(mp, XFS_TRANS_CREATE_TMPFILE); + + tres = &M_RES(mp)->tr_create_tmpfile; + error = xfs_trans_reserve(tp, tres, resblks, 0); + if (error == ENOSPC) { + /* No space at all so try a "no-allocation" reservation */ + resblks = 0; + error = xfs_trans_reserve(tp, tres, 0, 0); + } + if (error) { + cancel_flags = 0; + goto out_trans_cancel; + } + + error = xfs_trans_reserve_quota(tp, mp, udqp, gdqp, + pdqp, resblks, 1, 0); + if (error) + goto out_trans_cancel; + + error = xfs_dir_ialloc(&tp, dp, mode, 1, 0, + prid, resblks > 0, &ip, NULL); + if (error) { + if (error == ENOSPC) + goto out_trans_cancel; + goto out_trans_abort; + } + + if (mp->m_flags & XFS_MOUNT_WSYNC) + xfs_trans_set_sync(tp); + + /* + * Attach the dquot(s) to the inodes and modify them incore. + * These ids of the inode couldn't have changed since the new + * inode has been locked ever since it was created. + */ + xfs_qm_vop_create_dqattach(tp, ip, udqp, gdqp, pdqp); + + ip->i_d.di_nlink--; + d_tmpfile(dentry, VFS_I(ip)); + error = xfs_iunlink(tp, ip); + if (error) + goto out_trans_abort; + + error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES); + if (error) + goto out_release_inode; + + xfs_qm_dqrele(udqp); + xfs_qm_dqrele(gdqp); + xfs_qm_dqrele(pdqp); + + return 0; + + out_trans_abort: + cancel_flags |= XFS_TRANS_ABORT; + out_trans_cancel: + xfs_trans_cancel(tp, cancel_flags); + out_release_inode: + /* + * Wait until after the current transaction is aborted to + * release the inode. This prevents recursive transactions + * and deadlocks from xfs_inactive. + */ + if (ip) + IRELE(ip); + + xfs_qm_dqrele(udqp); + xfs_qm_dqrele(gdqp); + xfs_qm_dqrele(pdqp); + + return error; +} + +int xfs_link( xfs_inode_t *tdp, xfs_inode_t *sip, diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h index 6c58349..3a97820 100644 --- a/fs/xfs/xfs_inode.h +++ b/fs/xfs/xfs_inode.h @@ -333,6 +333,8 @@ int xfs_lookup(struct xfs_inode *dp, struct xfs_name *name, struct xfs_inode **ipp, struct xfs_name *ci_name); int xfs_create(struct xfs_inode *dp, struct xfs_name *name, umode_t mode, xfs_dev_t rdev, struct xfs_inode **ipp); +int xfs_create_tmpfile(struct xfs_inode *dp, struct dentry *dentry, + umode_t mode); int xfs_remove(struct xfs_inode *dp, struct xfs_name *name, struct xfs_inode *ip); int xfs_link(struct xfs_inode *tdp, struct xfs_inode *sip, diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c index 0ce1d75..cb06cc4 100644 --- a/fs/xfs/xfs_iops.c +++ b/fs/xfs/xfs_iops.c @@ -39,6 +39,7 @@ #include "xfs_da_btree.h" #include "xfs_dir2_priv.h" #include "xfs_dinode.h" +#include "xfs_trans_space.h" #include #include @@ -1043,6 +1044,19 @@ xfs_vn_fiemap( return 0; } +STATIC int +xfs_vn_tmpfile( + struct inode *dir, + struct dentry *dentry, + umode_t mode) +{ + int error; + + error = xfs_create_tmpfile(XFS_I(dir), dentry, mode); + + return -error; +} + static const struct inode_operations xfs_inode_operations = { .get_acl = xfs_get_acl, .getattr = xfs_vn_getattr, @@ -1079,6 +1093,7 @@ static const struct inode_operations xfs_dir_inode_operations = { .removexattr = generic_removexattr, .listxattr = xfs_vn_listxattr, .update_time = xfs_vn_update_time, + .tmpfile = xfs_vn_tmpfile, }; static const struct inode_operations xfs_dir_ci_inode_operations = { @@ -1105,6 +1120,7 @@ static const struct inode_operations xfs_dir_ci_inode_operations = { .removexattr = generic_removexattr, .listxattr = xfs_vn_listxattr, .update_time = xfs_vn_update_time, + .tmpfile = xfs_vn_tmpfile, }; static const struct inode_operations xfs_symlink_inode_operations = { diff --git a/fs/xfs/xfs_shared.h b/fs/xfs/xfs_shared.h index 8c5035a1..4484e51 100644 --- a/fs/xfs/xfs_shared.h +++ b/fs/xfs/xfs_shared.h @@ -104,7 +104,8 @@ extern const struct xfs_buf_ops xfs_symlink_buf_ops; #define XFS_TRANS_SB_COUNT 41 #define XFS_TRANS_CHECKPOINT 42 #define XFS_TRANS_ICREATE 43 -#define XFS_TRANS_TYPE_MAX 43 +#define XFS_TRANS_CREATE_TMPFILE 44 +#define XFS_TRANS_TYPE_MAX 44 /* new transaction types need to be reflected in xfs_logprint(8) */ #define XFS_TRANS_TYPES \ @@ -112,6 +113,7 @@ extern const struct xfs_buf_ops xfs_symlink_buf_ops; { XFS_TRANS_SETATTR_SIZE, "SETATTR_SIZE" }, \ { XFS_TRANS_INACTIVE, "INACTIVE" }, \ { XFS_TRANS_CREATE, "CREATE" }, \ + { XFS_TRANS_CREATE_TMPFILE, "CREATE_TMPFILE" }, \ { XFS_TRANS_CREATE_TRUNC, "CREATE_TRUNC" }, \ { XFS_TRANS_TRUNCATE_FILE, "TRUNCATE_FILE" }, \ { XFS_TRANS_REMOVE, "REMOVE" }, \ diff --git a/fs/xfs/xfs_trans_resv.c b/fs/xfs/xfs_trans_resv.c index 2fd59c0..bd3b4b7 100644 --- a/fs/xfs/xfs_trans_resv.c +++ b/fs/xfs/xfs_trans_resv.c @@ -229,6 +229,18 @@ xfs_calc_link_reservation( } /* + * For adding an inode to unlinked list we can modify: + * the agi hash list: sector size + * the unlinked inode: inode size + */ +STATIC uint +xfs_calc_iunlink_add_reservation(xfs_mount_t *mp) +{ + return xfs_calc_buf_res(1, mp->m_sb.sb_sectsize) + + xfs_calc_inode_res(mp, 1); +} + +/* * For removing a directory entry we can modify: * the parent directory inode: inode size * the removed inode: inode size @@ -245,10 +257,11 @@ xfs_calc_remove_reservation( struct xfs_mount *mp) { return XFS_DQUOT_LOGRES(mp) + - MAX((xfs_calc_inode_res(mp, 2) + + xfs_calc_iunlink_add_reservation(mp) + + MAX((xfs_calc_inode_res(mp, 1) + xfs_calc_buf_res(XFS_DIROP_LOG_COUNT(mp), XFS_FSB_TO_B(mp, 1))), - (xfs_calc_buf_res(5, mp->m_sb.sb_sectsize) + + (xfs_calc_buf_res(4, mp->m_sb.sb_sectsize) + xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 2), XFS_FSB_TO_B(mp, 1)))); } @@ -343,6 +356,20 @@ xfs_calc_create_reservation( } +STATIC uint +xfs_calc_create_tmpfile_reservation( + struct xfs_mount *mp) +{ + uint res = XFS_DQUOT_LOGRES(mp); + + if (xfs_sb_version_hascrc(&mp->m_sb)) + res += xfs_calc_icreate_resv_alloc(mp); + else + res += xfs_calc_create_resv_alloc(mp); + + return res + xfs_calc_iunlink_add_reservation(mp); +} + /* * Making a new directory is the same as creating a new file. */ @@ -729,6 +756,11 @@ xfs_trans_resv_calc( resp->tr_create.tr_logcount = XFS_CREATE_LOG_COUNT; resp->tr_create.tr_logflags |= XFS_TRANS_PERM_LOG_RES; + resp->tr_create_tmpfile.tr_logres = + xfs_calc_create_tmpfile_reservation(mp); + resp->tr_create_tmpfile.tr_logcount = XFS_CREATE_TMPFILE_LOG_COUNT; + resp->tr_create_tmpfile.tr_logflags |= XFS_TRANS_PERM_LOG_RES; + resp->tr_mkdir.tr_logres = xfs_calc_mkdir_reservation(mp); resp->tr_mkdir.tr_logcount = XFS_MKDIR_LOG_COUNT; resp->tr_mkdir.tr_logflags |= XFS_TRANS_PERM_LOG_RES; diff --git a/fs/xfs/xfs_trans_resv.h b/fs/xfs/xfs_trans_resv.h index de7de9a..285621d 100644 --- a/fs/xfs/xfs_trans_resv.h +++ b/fs/xfs/xfs_trans_resv.h @@ -38,6 +38,7 @@ struct xfs_trans_resv { struct xfs_trans_res tr_remove; /* unlink trans */ struct xfs_trans_res tr_symlink; /* symlink trans */ struct xfs_trans_res tr_create; /* create trans */ + struct xfs_trans_res tr_create_tmpfile; /* create O_TMPFILE trans */ struct xfs_trans_res tr_mkdir; /* mkdir trans */ struct xfs_trans_res tr_ifree; /* inode free trans */ struct xfs_trans_res tr_ichange; /* inode update trans */ @@ -100,6 +101,7 @@ struct xfs_trans_resv { #define XFS_ITRUNCATE_LOG_COUNT 2 #define XFS_INACTIVE_LOG_COUNT 2 #define XFS_CREATE_LOG_COUNT 2 +#define XFS_CREATE_TMPFILE_LOG_COUNT 2 #define XFS_MKDIR_LOG_COUNT 3 #define XFS_SYMLINK_LOG_COUNT 3 #define XFS_REMOVE_LOG_COUNT 2 -- cgit v0.10.2 From ab29743117f9f4c22ac44c13c1647fb24fb2bafe Mon Sep 17 00:00:00 2001 From: Zhi Yong Wu Date: Wed, 18 Dec 2013 08:22:41 +0800 Subject: xfs: allow linkat() on O_TMPFILE files The VFS allows an anonymous temporary file to be named at a later time via a linkat() syscall. The inodes for O_TMPFILE files are are marked with a special flag I_LINKABLE and have a zero link count. To support this in XFS, xfs_link() detects if this flag I_LINKABLE is set and behaves appropriately when detected. So in this case, its transaciton reservation takes into account the additional overhead of removing the inode from the unlinked list. Then the inode is removed from the unlinked list and the directory entry is added. Finally its link count is bumped accordingly. Signed-off-by: Zhi Yong Wu Reviewed-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Ben Myers diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index ac133ea..b08b5a8 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -61,6 +61,8 @@ kmem_zone_t *xfs_inode_zone; STATIC int xfs_iflush_int(xfs_inode_t *, xfs_buf_t *); +STATIC int xfs_iunlink_remove(xfs_trans_t *, xfs_inode_t *); + /* * helper function to extract extent size hint from inode */ @@ -1118,7 +1120,7 @@ xfs_bumplink( { xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_CHG); - ASSERT(ip->i_d.di_nlink > 0); + ASSERT(ip->i_d.di_nlink > 0 || (VFS_I(ip)->i_state & I_LINKABLE)); ip->i_d.di_nlink++; inc_nlink(VFS_I(ip)); if ((ip->i_d.di_version == 1) && @@ -1504,6 +1506,12 @@ xfs_link( xfs_bmap_init(&free_list, &first_block); + if (sip->i_d.di_nlink == 0) { + error = xfs_iunlink_remove(tp, sip); + if (error) + goto abort_return; + } + error = xfs_dir_createname(tp, tdp, target_name, sip->i_ino, &first_block, &free_list, resblks); if (error) diff --git a/fs/xfs/xfs_trans_resv.c b/fs/xfs/xfs_trans_resv.c index bd3b4b7..76f9a02 100644 --- a/fs/xfs/xfs_trans_resv.c +++ b/fs/xfs/xfs_trans_resv.c @@ -204,6 +204,20 @@ xfs_calc_rename_reservation( } /* + * For removing an inode from unlinked list at first, we can modify: + * the agi hash list and counters: sector size + * the on disk inode before ours in the agi hash list: inode cluster size + */ +STATIC uint +xfs_calc_iunlink_remove_reservation( + struct xfs_mount *mp) +{ + return xfs_calc_buf_res(1, mp->m_sb.sb_sectsize) + + MAX((__uint16_t)XFS_FSB_TO_B(mp, 1), + (__uint16_t)XFS_INODE_CLUSTER_SIZE(mp)); +} + +/* * For creating a link to an inode: * the parent directory inode: inode size * the linked inode: inode size @@ -220,6 +234,7 @@ xfs_calc_link_reservation( struct xfs_mount *mp) { return XFS_DQUOT_LOGRES(mp) + + xfs_calc_iunlink_remove_reservation(mp) + MAX((xfs_calc_inode_res(mp, 2) + xfs_calc_buf_res(XFS_DIROP_LOG_COUNT(mp), XFS_FSB_TO_B(mp, 1))), @@ -410,9 +425,9 @@ xfs_calc_ifree_reservation( { return XFS_DQUOT_LOGRES(mp) + xfs_calc_inode_res(mp, 1) + - xfs_calc_buf_res(2, mp->m_sb.sb_sectsize) + + xfs_calc_buf_res(1, mp->m_sb.sb_sectsize) + xfs_calc_buf_res(1, XFS_FSB_TO_B(mp, 1)) + - max_t(uint, XFS_FSB_TO_B(mp, 1), XFS_INODE_CLUSTER_SIZE(mp)) + + xfs_calc_iunlink_remove_reservation(mp) + xfs_calc_buf_res(1, 0) + xfs_calc_buf_res(2 + XFS_IALLOC_BLOCKS(mp) + mp->m_in_maxlevels, 0) + -- cgit v0.10.2 From 70bbca07766091be699736163a07ee016ed72482 Mon Sep 17 00:00:00 2001 From: Brian Foster Date: Fri, 7 Feb 2014 14:53:50 +1100 Subject: xfs: use tr_growrtalloc for growing rt files This is a regression from the following commit: 3d3c8b5222b9 xfs: refactor xfs_trans_reserve() interface Use the tr_growrtalloc log reservation for growing the bitmap/summary files. Signed-off-by: Brian Foster Reviewed-by: Eric Sandeen Reviewed-by: Jie Liu Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c index a6a76b2..ec5ca65 100644 --- a/fs/xfs/xfs_rtalloc.c +++ b/fs/xfs/xfs_rtalloc.c @@ -842,7 +842,7 @@ xfs_growfs_rt_alloc( /* * Reserve space & log for one extent added to the file. */ - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_growdata, + error = xfs_trans_reserve(tp, &M_RES(mp)->tr_growrtalloc, resblks, 0); if (error) goto error_cancel; -- cgit v0.10.2 From c19ec235352c2a001c9dc7e86acdfd9f2b62150d Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Fri, 7 Feb 2014 14:54:22 +1100 Subject: xfs: remove unused tr_swrite tr_swrite is never used, remove it. From a very quick look, I think the usage of it (and its ancestor XFS_SWRITE_LOG_RES) went away in commit 13e6d5cd "xfs: merge fsync and O_SYNC handling" back in 2009. Signed-off-by: Eric Sandeen Reviewed-by: Jie Liu Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_trans_resv.c b/fs/xfs/xfs_trans_resv.c index 2ffd3e33..c9d2708 100644 --- a/fs/xfs/xfs_trans_resv.c +++ b/fs/xfs/xfs_trans_resv.c @@ -784,7 +784,6 @@ xfs_trans_resv_calc( /* The following transaction are logged in logical format */ resp->tr_ichange.tr_logres = xfs_calc_ichange_reservation(mp); resp->tr_growdata.tr_logres = xfs_calc_growdata_reservation(mp); - resp->tr_swrite.tr_logres = xfs_calc_swrite_reservation(mp); resp->tr_fsyncts.tr_logres = xfs_calc_swrite_reservation(mp); resp->tr_writeid.tr_logres = xfs_calc_writeid_reservation(mp); resp->tr_attrsetrt.tr_logres = xfs_calc_attrsetrt_reservation(mp); diff --git a/fs/xfs/xfs_trans_resv.h b/fs/xfs/xfs_trans_resv.h index de7de9a..f76c129 100644 --- a/fs/xfs/xfs_trans_resv.h +++ b/fs/xfs/xfs_trans_resv.h @@ -42,7 +42,6 @@ struct xfs_trans_resv { struct xfs_trans_res tr_ifree; /* inode free trans */ struct xfs_trans_res tr_ichange; /* inode update trans */ struct xfs_trans_res tr_growdata; /* fs data section grow trans */ - struct xfs_trans_res tr_swrite; /* sync write inode trans */ struct xfs_trans_res tr_addafork; /* add inode attr fork trans */ struct xfs_trans_res tr_writeid; /* write setuid/setgid file */ struct xfs_trans_res tr_attrinval; /* attr fork buffer -- cgit v0.10.2 From 410b11a675dca827e893f07c3155691eda3b5887 Mon Sep 17 00:00:00 2001 From: Brian Foster Date: Fri, 7 Feb 2014 14:55:54 +1100 Subject: xfs: use tr_qm_dqalloc log reservation for dquot alloc The dquot allocation path in xfs_qm_dqread() currently uses the attribute set log reservation, which appears to be incorrect. We have reports of transaction reservation overruns with the current code. E.g., a repeated run of xfstests test generic/270 on a 512b block size fs occassionally produces the following in dmesg: XFS (sdN): xlog_write: reservation summary: trans type = QM_DQALLOC (30) unit res = 7080 bytes current res = -632 bytes total reg = 0 bytes (o/flow = 0 bytes) ophdrs = 0 (ophdr space = 0 bytes) ophdr + reg = 0 bytes num regions = 0 XFS (sdN): xlog_write: reservation ran out. Need to up reservation The dquot allocation case should consist of a write reservation (i.e., we are allocating a range of the internal quota file) plus the size of the actual dquots. We already have a log reservation definition for this operation (tr_qm_dqalloc). Use it in xfs_qm_dqread() and update the log reservation calculation function to use the write res. calculation function rather than reading the assumed to be pre-calculated value directly. Signed-off-by: Brian Foster Reviewed-by: Jie Liu Reviewed-by: Ben Myers Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c index 7aeb4c8..868b19f 100644 --- a/fs/xfs/xfs_dquot.c +++ b/fs/xfs/xfs_dquot.c @@ -615,7 +615,7 @@ xfs_qm_dqread( if (flags & XFS_QMOPT_DQALLOC) { tp = xfs_trans_alloc(mp, XFS_TRANS_QM_DQALLOC); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_attrsetm, + error = xfs_trans_reserve(tp, &M_RES(mp)->tr_qm_dqalloc, XFS_QM_DQALLOC_SPACE_RES(mp), 0); if (error) goto error1; diff --git a/fs/xfs/xfs_trans_resv.c b/fs/xfs/xfs_trans_resv.c index c9d2708..8515b04 100644 --- a/fs/xfs/xfs_trans_resv.c +++ b/fs/xfs/xfs_trans_resv.c @@ -644,15 +644,14 @@ xfs_calc_qm_setqlim_reservation( /* * Allocating quota on disk if needed. - * the write transaction log space: M_RES(mp)->tr_write.tr_logres + * the write transaction log space for quota file extent allocation * the unit of quota allocation: one system block size */ STATIC uint xfs_calc_qm_dqalloc_reservation( struct xfs_mount *mp) { - ASSERT(M_RES(mp)->tr_write.tr_logres); - return M_RES(mp)->tr_write.tr_logres + + return xfs_calc_write_reservation(mp) + xfs_calc_buf_res(1, XFS_FSB_TO_B(mp, XFS_DQUOT_CLUSTER_SIZE_FSB) - 1); } -- cgit v0.10.2 From c6f9726444c8f8c7df24950864bf1a4cb2c61b3e Mon Sep 17 00:00:00 2001 From: Jie Liu Date: Fri, 7 Feb 2014 15:26:07 +1100 Subject: xfs: convert xfs_log_commit_cil() to void Convert xfs_log_commit_cil() to a void function since it return nothing but 0 in any case, after that we can simplify the relative code logic in xfs_trans_commit() accordingly. Signed-off-by: Jie Liu Reviewed-by: Brian Foster Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_log.h b/fs/xfs/xfs_log.h index b0f4ef7..2c40044 100644 --- a/fs/xfs/xfs_log.h +++ b/fs/xfs/xfs_log.h @@ -175,7 +175,7 @@ void xlog_iodone(struct xfs_buf *); struct xlog_ticket *xfs_log_ticket_get(struct xlog_ticket *ticket); void xfs_log_ticket_put(struct xlog_ticket *ticket); -int xfs_log_commit_cil(struct xfs_mount *mp, struct xfs_trans *tp, +void xfs_log_commit_cil(struct xfs_mount *mp, struct xfs_trans *tp, xfs_lsn_t *commit_lsn, int flags); bool xfs_log_item_in_current_chkpt(struct xfs_log_item *lip); diff --git a/fs/xfs/xfs_log_cil.c b/fs/xfs/xfs_log_cil.c index cdebd83..f034cbd 100644 --- a/fs/xfs/xfs_log_cil.c +++ b/fs/xfs/xfs_log_cil.c @@ -710,7 +710,7 @@ xlog_cil_empty( * background commit, returns without it held once background commits are * allowed again. */ -int +void xfs_log_commit_cil( struct xfs_mount *mp, struct xfs_trans *tp, @@ -756,7 +756,6 @@ xfs_log_commit_cil( xlog_cil_push_background(log); up_read(&cil->xc_ctx_lock); - return 0; } /* diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c index c812c5c0..54a5732 100644 --- a/fs/xfs/xfs_trans.c +++ b/fs/xfs/xfs_trans.c @@ -887,12 +887,7 @@ xfs_trans_commit( xfs_trans_apply_sb_deltas(tp); xfs_trans_apply_dquot_deltas(tp); - error = xfs_log_commit_cil(mp, tp, &commit_lsn, flags); - if (error == ENOMEM) { - xfs_force_shutdown(mp, SHUTDOWN_LOG_IO_ERROR); - error = XFS_ERROR(EIO); - goto out_unreserve; - } + xfs_log_commit_cil(mp, tp, &commit_lsn, flags); current_restore_flags_nested(&tp->t_pflags, PF_FSTRANS); xfs_trans_free(tp); @@ -902,10 +897,7 @@ xfs_trans_commit( * log out now and wait for it. */ if (sync) { - if (!error) { - error = _xfs_log_force_lsn(mp, commit_lsn, - XFS_LOG_SYNC, NULL); - } + error = _xfs_log_force_lsn(mp, commit_lsn, XFS_LOG_SYNC, NULL); XFS_STATS_INC(xs_trans_sync); } else { XFS_STATS_INC(xs_trans_async); -- cgit v0.10.2 From 392c6de98af1fd7e2fc9c7bf5e52be16286f7b42 Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Fri, 7 Feb 2014 15:26:11 +1100 Subject: xfs: sanitize sb_inopblock in xfs_mount_validate_sb xfs_mount_validate_sb doesn't check sb_inopblock for sanity (as does its xfs_repair counterpart, FWIW). If it's out of bounds, we can go off the rails in i.e. xfs_inode_buf_verify(), which uses sb_inopblock as a loop limit when stepping through a metadata buffer. The problem can be demonstrated easily by corrupting sb_inopblock with xfs_db and trying to mount the result: # mkfs.xfs -dfile,name=fsfile,size=1g # xfs_db -x fsfile xfs_db> sb 0 xfs_db> write inopblock 512 inopblock = 512 xfs_db> quit # mount -o loop fsfile mnt and we blow up in xfs_inode_buf_verify(). With this patch, we get a (very noisy) corruption error, and fail the mount as we should. Signed-off-by: Eric Sandeen Reviewed-by: Jie Liu Reviewed-by: Brian Foster Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_sb.c b/fs/xfs/xfs_sb.c index b7c9aea..511cce9 100644 --- a/fs/xfs/xfs_sb.c +++ b/fs/xfs/xfs_sb.c @@ -288,6 +288,7 @@ xfs_mount_validate_sb( sbp->sb_inodelog < XFS_DINODE_MIN_LOG || sbp->sb_inodelog > XFS_DINODE_MAX_LOG || sbp->sb_inodesize != (1 << sbp->sb_inodelog) || + sbp->sb_inopblock != howmany(sbp->sb_blocksize,sbp->sb_inodesize) || (sbp->sb_blocklog - sbp->sb_inodelog != sbp->sb_inopblog) || (sbp->sb_rextsize * sbp->sb_blocksize > XFS_MAX_RTEXTSIZE) || (sbp->sb_rextsize * sbp->sb_blocksize < XFS_MIN_RTEXTSIZE) || -- cgit v0.10.2 From 4ae69fea588148360d470ce604714b6d619ea749 Mon Sep 17 00:00:00 2001 From: Jie Liu Date: Fri, 7 Feb 2014 15:26:11 +1100 Subject: xfs: return -E2BIG if hit the maximum size limits of ACLs We should return -E2BIG rather than -EINVAL if hit the maximum size limits of ACLS, as the former is consistent with VFS xattr syscalls. Signed-off-by: Jie Liu Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_acl.c b/fs/xfs/xfs_acl.c index 0ecec18..6888ad8 100644 --- a/fs/xfs/xfs_acl.c +++ b/fs/xfs/xfs_acl.c @@ -281,7 +281,7 @@ xfs_set_acl(struct inode *inode, struct posix_acl *acl, int type) if (!acl) goto set_acl; - error = -EINVAL; + error = -E2BIG; if (acl->a_count > XFS_ACL_MAX_ENTRIES(XFS_M(inode->i_sb))) return error; -- cgit v0.10.2 From 492185ef1dd261768203a6c3accfd445cde8c503 Mon Sep 17 00:00:00 2001 From: Jie Liu Date: Fri, 7 Feb 2014 15:26:11 +1100 Subject: xfs: remove XFS_TRANS_DEBUG dead code Remove the leftover XFS_TRANS_DEBUG dead code following the previous cleaning up of it in commits ec47eb6b0b450. Signed-off-by: Jie Liu Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c index 3314911..8752821 100644 --- a/fs/xfs/xfs_buf_item.c +++ b/fs/xfs/xfs_buf_item.c @@ -796,20 +796,6 @@ xfs_buf_item_init( bip->bli_formats[i].blf_map_size = map_size; } -#ifdef XFS_TRANS_DEBUG - /* - * Allocate the arrays for tracking what needs to be logged - * and what our callers request to be logged. bli_orig - * holds a copy of the original, clean buffer for comparison - * against, and bli_logged keeps a 1 bit flag per byte in - * the buffer to indicate which bytes the callers have asked - * to have logged. - */ - bip->bli_orig = kmem_alloc(BBTOB(bp->b_length), KM_SLEEP); - memcpy(bip->bli_orig, bp->b_addr, BBTOB(bp->b_length)); - bip->bli_logged = kmem_zalloc(BBTOB(bp->b_length) / NBBY, KM_SLEEP); -#endif - /* * Put the buf item into the list of items attached to the * buffer at the front. @@ -957,11 +943,6 @@ STATIC void xfs_buf_item_free( xfs_buf_log_item_t *bip) { -#ifdef XFS_TRANS_DEBUG - kmem_free(bip->bli_orig); - kmem_free(bip->bli_logged); -#endif /* XFS_TRANS_DEBUG */ - xfs_buf_item_free_format(bip); kmem_zone_free(xfs_buf_item_zone, bip); } -- cgit v0.10.2 From 6039257378e4c84da06e68230b14fef955508ce6 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 10 Feb 2014 10:27:11 +1100 Subject: direct-io: add flag to allow aio writes beyond i_size Some filesystems can handle direct I/O writes beyond i_size safely, so allow them to opt into receiving them. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner diff --git a/fs/direct-io.c b/fs/direct-io.c index 160a548..a701752 100644 --- a/fs/direct-io.c +++ b/fs/direct-io.c @@ -1194,13 +1194,19 @@ do_blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode, } /* - * For file extending writes updating i_size before data - * writeouts complete can expose uninitialized blocks. So - * even for AIO, we need to wait for i/o to complete before - * returning in this case. + * For file extending writes updating i_size before data writeouts + * complete can expose uninitialized blocks in dumb filesystems. + * In that case we need to wait for I/O completion even if asked + * for an asynchronous write. */ - dio->is_async = !is_sync_kiocb(iocb) && !((rw & WRITE) && - (end > i_size_read(inode))); + if (is_sync_kiocb(iocb)) + dio->is_async = false; + else if (!(dio->flags & DIO_ASYNC_EXTEND) && + (rw & WRITE) && end > i_size_read(inode)) + dio->is_async = false; + else + dio->is_async = true; + dio->inode = inode; dio->rw = rw; diff --git a/include/linux/fs.h b/include/linux/fs.h index 09f553c..f7faefc 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -2527,6 +2527,9 @@ enum { /* filesystem does not support filling holes */ DIO_SKIP_HOLES = 0x02, + + /* filesystem can handle aio writes beyond i_size */ + DIO_ASYNC_EXTEND = 0x04, }; void dio_end_io(struct bio *bio, int error); -- cgit v0.10.2 From d531d91d69902e55633ed834f531aa0b48d618cc Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 10 Feb 2014 10:27:43 +1100 Subject: xfs: always use unwritten extents for direct I/O writes To allow aio writes beyond i_size we need to create unwritten extents for newly allocated blocks, similar to how we already do inside i_size. Instead of adding another special case we now use unwritten extents unconditionally. This also marks the end of directly allocation data extents in all of XFS - we now always use either delalloc or unwritten extents. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c index 22d1cbe..3b80eba 100644 --- a/fs/xfs/xfs_iomap.c +++ b/fs/xfs/xfs_iomap.c @@ -128,7 +128,6 @@ xfs_iomap_write_direct( xfs_fsblock_t firstfsb; xfs_extlen_t extsz, temp; int nimaps; - int bmapi_flag; int quota_flag; int rt; xfs_trans_t *tp; @@ -200,18 +199,15 @@ xfs_iomap_write_direct( xfs_trans_ijoin(tp, ip, 0); - bmapi_flag = 0; - if (offset < XFS_ISIZE(ip) || extsz) - bmapi_flag |= XFS_BMAPI_PREALLOC; - /* * From this point onwards we overwrite the imap pointer that the * caller gave to us. */ xfs_bmap_init(&free_list, &firstfsb); nimaps = 1; - error = xfs_bmapi_write(tp, ip, offset_fsb, count_fsb, bmapi_flag, - &firstfsb, 0, imap, &nimaps, &free_list); + error = xfs_bmapi_write(tp, ip, offset_fsb, count_fsb, + XFS_BMAPI_PREALLOC, &firstfsb, 0, + imap, &nimaps, &free_list); if (error) goto out_bmap_cancel; -- cgit v0.10.2 From 9862f62faba8c279ac07415a6f610041116fbdc0 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 10 Feb 2014 10:28:04 +1100 Subject: xfs: allow appending aio writes XFS can easily support appending aio writes by ensuring we always allocate blocks as unwritten extents when performing direct I/O writes and only converting them to written extents at I/O completion. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c index db2cfb0..ef62c6b 100644 --- a/fs/xfs/xfs_aops.c +++ b/fs/xfs/xfs_aops.c @@ -1441,7 +1441,8 @@ xfs_vm_direct_IO( ret = __blockdev_direct_IO(rw, iocb, inode, bdev, iov, offset, nr_segs, xfs_get_blocks_direct, - xfs_end_io_direct_write, NULL, 0); + xfs_end_io_direct_write, NULL, + DIO_ASYNC_EXTEND); if (ret != -EIOCBQUEUED && iocb->private) goto out_destroy_ioend; } else { -- cgit v0.10.2 From 00f5e61998dd17f5375d9dfc01331f104b83f841 Mon Sep 17 00:00:00 2001 From: Namjae Jeon Date: Mon, 24 Feb 2014 10:58:15 +1100 Subject: fs: Add new flag(FALLOC_FL_COLLAPSE_RANGE) for fallocate This patch is in response of the following post: http://lwn.net/Articles/556136/ "ext4: introduce two new ioctls" Dave chinner suggested that truncate_block_range (which was one of the ioctls name) should be a fallocate operation and not any fs specific ioctl, hence we add this functionality to new flags of fallocate. This new functionality of collapsing range could be used by media editing tools which does non linear editing to quickly purge and edit parts of a media file. This will immensely improve the performance of these operations. The limitation of fs block size aligned offsets can be easily handled by media codecs which are encapsulated in a conatiner as they have to just change the offset to next keyframe value to match the proper alignment. Signed-off-by: Namjae Jeon Signed-off-by: Ashish Sangwan Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner diff --git a/fs/open.c b/fs/open.c index 4b3e1ed..4a923a5 100644 --- a/fs/open.c +++ b/fs/open.c @@ -231,7 +231,8 @@ int do_fallocate(struct file *file, int mode, loff_t offset, loff_t len) return -EINVAL; /* Return error if mode is not supported */ - if (mode & ~(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE)) + if (mode & ~(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE | + FALLOC_FL_COLLAPSE_RANGE)) return -EOPNOTSUPP; /* Punch hole must have keep size set */ @@ -239,11 +240,20 @@ int do_fallocate(struct file *file, int mode, loff_t offset, loff_t len) !(mode & FALLOC_FL_KEEP_SIZE)) return -EOPNOTSUPP; + /* Collapse range should only be used exclusively. */ + if ((mode & FALLOC_FL_COLLAPSE_RANGE) && + (mode & ~FALLOC_FL_COLLAPSE_RANGE)) + return -EINVAL; + if (!(file->f_mode & FMODE_WRITE)) return -EBADF; - /* It's not possible punch hole on append only file */ - if (mode & FALLOC_FL_PUNCH_HOLE && IS_APPEND(inode)) + /* + * It's not possible to punch hole or perform collapse range + * on append only file + */ + if (mode & (FALLOC_FL_PUNCH_HOLE | FALLOC_FL_COLLAPSE_RANGE) + && IS_APPEND(inode)) return -EPERM; if (IS_IMMUTABLE(inode)) @@ -271,6 +281,14 @@ int do_fallocate(struct file *file, int mode, loff_t offset, loff_t len) if (((offset + len) > inode->i_sb->s_maxbytes) || ((offset + len) < 0)) return -EFBIG; + /* + * There is no need to overlap collapse range with EOF, in which case + * it is effectively a truncate operation + */ + if ((mode & FALLOC_FL_COLLAPSE_RANGE) && + (offset + len >= i_size_read(inode))) + return -EINVAL; + if (!file->f_op->fallocate) return -EOPNOTSUPP; diff --git a/include/uapi/linux/falloc.h b/include/uapi/linux/falloc.h index 990c4cc..5ff562d 100644 --- a/include/uapi/linux/falloc.h +++ b/include/uapi/linux/falloc.h @@ -5,5 +5,26 @@ #define FALLOC_FL_PUNCH_HOLE 0x02 /* de-allocates range */ #define FALLOC_FL_NO_HIDE_STALE 0x04 /* reserved codepoint */ +/* + * FALLOC_FL_COLLAPSE_RANGE is used to remove a range of a file + * without leaving a hole in the file. The contents of the file beyond + * the range being removed is appended to the start offset of the range + * being removed (i.e. the hole that was punched is "collapsed"), + * resulting in a file layout that looks like the range that was + * removed never existed. As such collapsing a range of a file changes + * the size of the file, reducing it by the same length of the range + * that has been removed by the operation. + * + * Different filesystems may implement different limitations on the + * granularity of the operation. Most will limit operations to + * filesystem block size boundaries, but this boundary may be larger or + * smaller depending on the filesystem and/or the configuration of the + * filesystem or file. + * + * Attempting to collapse a range that crosses the end of the file is + * considered an illegal operation - just use ftruncate(2) if you need + * to collapse a range that crosses EOF. + */ +#define FALLOC_FL_COLLAPSE_RANGE 0x08 #endif /* _UAPI_FALLOC_H_ */ -- cgit v0.10.2 From e1d8fb88a64c1f8094b9f6c3b6d2d9e6719c970d Mon Sep 17 00:00:00 2001 From: Namjae Jeon Date: Mon, 24 Feb 2014 10:58:19 +1100 Subject: xfs: Add support FALLOC_FL_COLLAPSE_RANGE for fallocate This patch implements fallocate's FALLOC_FL_COLLAPSE_RANGE for XFS. The semantics of this flag are following: 1) It collapses the range lying between offset and length by removing any data blocks which are present in this range and than updates all the logical offsets of extents beyond "offset + len" to nullify the hole created by removing blocks. In short, it does not leave a hole. 2) It should be used exclusively. No other fallocate flag in combination. 3) Offset and length supplied to fallocate should be fs block size aligned in case of xfs and ext4. 4) Collaspe range does not work beyond i_size. Signed-off-by: Namjae Jeon Signed-off-by: Ashish Sangwan Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 152543c..5b6092e 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -5378,3 +5378,196 @@ error0: } return error; } + +/* + * Shift extent records to the left to cover a hole. + * + * The maximum number of extents to be shifted in a single operation + * is @num_exts, and @current_ext keeps track of the current extent + * index we have shifted. @offset_shift_fsb is the length by which each + * extent is shifted. If there is no hole to shift the extents + * into, this will be considered invalid operation and we abort immediately. + */ +int +xfs_bmap_shift_extents( + struct xfs_trans *tp, + struct xfs_inode *ip, + int *done, + xfs_fileoff_t start_fsb, + xfs_fileoff_t offset_shift_fsb, + xfs_extnum_t *current_ext, + xfs_fsblock_t *firstblock, + struct xfs_bmap_free *flist, + int num_exts) +{ + struct xfs_btree_cur *cur; + struct xfs_bmbt_rec_host *gotp; + struct xfs_bmbt_irec got; + struct xfs_bmbt_irec left; + struct xfs_mount *mp = ip->i_mount; + struct xfs_ifork *ifp; + xfs_extnum_t nexts = 0; + xfs_fileoff_t startoff; + int error = 0; + int i; + int whichfork = XFS_DATA_FORK; + int logflags; + xfs_filblks_t blockcount = 0; + + if (unlikely(XFS_TEST_ERROR( + (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS && + XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE), + mp, XFS_ERRTAG_BMAPIFORMAT, XFS_RANDOM_BMAPIFORMAT))) { + XFS_ERROR_REPORT("xfs_bmap_shift_extents", + XFS_ERRLEVEL_LOW, mp); + return XFS_ERROR(EFSCORRUPTED); + } + + if (XFS_FORCED_SHUTDOWN(mp)) + return XFS_ERROR(EIO); + + ASSERT(current_ext != NULL); + + ifp = XFS_IFORK_PTR(ip, whichfork); + + if (!(ifp->if_flags & XFS_IFEXTENTS)) { + /* Read in all the extents */ + error = xfs_iread_extents(tp, ip, whichfork); + if (error) + return error; + } + + /* + * If *current_ext is 0, we would need to lookup the extent + * from where we would start shifting and store it in gotp. + */ + if (!*current_ext) { + gotp = xfs_iext_bno_to_ext(ifp, start_fsb, current_ext); + /* + * gotp can be null in 2 cases: 1) if there are no extents + * or 2) start_fsb lies in a hole beyond which there are + * no extents. Either way, we are done. + */ + if (!gotp) { + *done = 1; + return 0; + } + } + + /* We are going to change core inode */ + logflags = XFS_ILOG_CORE; + + if (ifp->if_flags & XFS_IFBROOT) { + cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork); + cur->bc_private.b.firstblock = *firstblock; + cur->bc_private.b.flist = flist; + cur->bc_private.b.flags = 0; + } else { + cur = NULL; + logflags |= XFS_ILOG_DEXT; + } + + while (nexts++ < num_exts && + *current_ext < XFS_IFORK_NEXTENTS(ip, whichfork)) { + + gotp = xfs_iext_get_ext(ifp, *current_ext); + xfs_bmbt_get_all(gotp, &got); + startoff = got.br_startoff - offset_shift_fsb; + + /* + * Before shifting extent into hole, make sure that the hole + * is large enough to accomodate the shift. + */ + if (*current_ext) { + xfs_bmbt_get_all(xfs_iext_get_ext(ifp, + *current_ext - 1), &left); + + if (startoff < left.br_startoff + left.br_blockcount) + error = XFS_ERROR(EINVAL); + } else if (offset_shift_fsb > got.br_startoff) { + /* + * When first extent is shifted, offset_shift_fsb + * should be less than the stating offset of + * the first extent. + */ + error = XFS_ERROR(EINVAL); + } + + if (error) + goto del_cursor; + + if (cur) { + error = xfs_bmbt_lookup_eq(cur, got.br_startoff, + got.br_startblock, + got.br_blockcount, + &i); + if (error) + goto del_cursor; + XFS_WANT_CORRUPTED_GOTO(i == 1, del_cursor); + } + + /* Check if we can merge 2 adjacent extents */ + if (*current_ext && + left.br_startoff + left.br_blockcount == startoff && + left.br_startblock + left.br_blockcount == + got.br_startblock && + left.br_state == got.br_state && + left.br_blockcount + got.br_blockcount <= MAXEXTLEN) { + blockcount = left.br_blockcount + + got.br_blockcount; + xfs_iext_remove(ip, *current_ext, 1, 0); + if (cur) { + error = xfs_btree_delete(cur, &i); + if (error) + goto del_cursor; + XFS_WANT_CORRUPTED_GOTO(i == 1, del_cursor); + } + XFS_IFORK_NEXT_SET(ip, whichfork, + XFS_IFORK_NEXTENTS(ip, whichfork) - 1); + gotp = xfs_iext_get_ext(ifp, --*current_ext); + xfs_bmbt_get_all(gotp, &got); + + /* Make cursor point to the extent we will update */ + if (cur) { + error = xfs_bmbt_lookup_eq(cur, got.br_startoff, + got.br_startblock, + got.br_blockcount, + &i); + if (error) + goto del_cursor; + XFS_WANT_CORRUPTED_GOTO(i == 1, del_cursor); + } + + xfs_bmbt_set_blockcount(gotp, blockcount); + got.br_blockcount = blockcount; + } else { + /* We have to update the startoff */ + xfs_bmbt_set_startoff(gotp, startoff); + got.br_startoff = startoff; + } + + if (cur) { + error = xfs_bmbt_update(cur, got.br_startoff, + got.br_startblock, + got.br_blockcount, + got.br_state); + if (error) + goto del_cursor; + } + + (*current_ext)++; + } + + /* Check if we are done */ + if (*current_ext == XFS_IFORK_NEXTENTS(ip, whichfork)) + *done = 1; + +del_cursor: + if (cur) + xfs_btree_del_cursor(cur, + error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR); + + xfs_trans_log_inode(tp, ip, logflags); + + return error; +} diff --git a/fs/xfs/xfs_bmap.h b/fs/xfs/xfs_bmap.h index 33b41f3..f84bd7a 100644 --- a/fs/xfs/xfs_bmap.h +++ b/fs/xfs/xfs_bmap.h @@ -127,6 +127,16 @@ static inline void xfs_bmap_init(xfs_bmap_free_t *flp, xfs_fsblock_t *fbp) { BMAP_RIGHT_FILLING, "RF" }, \ { BMAP_ATTRFORK, "ATTR" } + +/* + * This macro is used to determine how many extents will be shifted + * in one write transaction. We could require two splits, + * an extent move on the first and an extent merge on the second, + * So it is proper that one extent is shifted inside write transaction + * at a time. + */ +#define XFS_BMAP_MAX_SHIFT_EXTENTS 1 + #ifdef DEBUG void xfs_bmap_trace_exlist(struct xfs_inode *ip, xfs_extnum_t cnt, int whichfork, unsigned long caller_ip); @@ -169,5 +179,10 @@ int xfs_bunmapi(struct xfs_trans *tp, struct xfs_inode *ip, int xfs_check_nostate_extents(struct xfs_ifork *ifp, xfs_extnum_t idx, xfs_extnum_t num); uint xfs_default_attroffset(struct xfs_inode *ip); +int xfs_bmap_shift_extents(struct xfs_trans *tp, struct xfs_inode *ip, + int *done, xfs_fileoff_t start_fsb, + xfs_fileoff_t offset_shift_fsb, xfs_extnum_t *current_ext, + xfs_fsblock_t *firstblock, struct xfs_bmap_free *flist, + int num_exts); #endif /* __XFS_BMAP_H__ */ diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c index f264616..01f6a64 100644 --- a/fs/xfs/xfs_bmap_util.c +++ b/fs/xfs/xfs_bmap_util.c @@ -1349,7 +1349,6 @@ xfs_free_file_space( * the freeing of the space succeeds at ENOSPC. */ tp = xfs_trans_alloc(mp, XFS_TRANS_DIOSTRAT); - tp->t_flags |= XFS_TRANS_RESERVE; error = xfs_trans_reserve(tp, &M_RES(mp)->tr_write, resblks, 0); /* @@ -1468,6 +1467,102 @@ out: } /* + * xfs_collapse_file_space() + * This routine frees disk space and shift extent for the given file. + * The first thing we do is to free data blocks in the specified range + * by calling xfs_free_file_space(). It would also sync dirty data + * and invalidate page cache over the region on which collapse range + * is working. And Shift extent records to the left to cover a hole. + * RETURNS: + * 0 on success + * errno on error + * + */ +int +xfs_collapse_file_space( + struct xfs_inode *ip, + xfs_off_t offset, + xfs_off_t len) +{ + int done = 0; + struct xfs_mount *mp = ip->i_mount; + struct xfs_trans *tp; + int error; + xfs_extnum_t current_ext = 0; + struct xfs_bmap_free free_list; + xfs_fsblock_t first_block; + int committed; + xfs_fileoff_t start_fsb; + xfs_fileoff_t shift_fsb; + + ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL)); + + trace_xfs_collapse_file_space(ip); + + start_fsb = XFS_B_TO_FSB(mp, offset + len); + shift_fsb = XFS_B_TO_FSB(mp, len); + + error = xfs_free_file_space(ip, offset, len); + if (error) + return error; + + while (!error && !done) { + tp = xfs_trans_alloc(mp, XFS_TRANS_DIOSTRAT); + tp->t_flags |= XFS_TRANS_RESERVE; + /* + * We would need to reserve permanent block for transaction. + * This will come into picture when after shifting extent into + * hole we found that adjacent extents can be merged which + * may lead to freeing of a block during record update. + */ + error = xfs_trans_reserve(tp, &M_RES(mp)->tr_write, + XFS_DIOSTRAT_SPACE_RES(mp, 0), 0); + if (error) { + ASSERT(error == ENOSPC || XFS_FORCED_SHUTDOWN(mp)); + xfs_trans_cancel(tp, 0); + break; + } + + xfs_ilock(ip, XFS_ILOCK_EXCL); + error = xfs_trans_reserve_quota(tp, mp, ip->i_udquot, + ip->i_gdquot, ip->i_pdquot, + XFS_DIOSTRAT_SPACE_RES(mp, 0), 0, + XFS_QMOPT_RES_REGBLKS); + if (error) + goto out; + + xfs_trans_ijoin(tp, ip, 0); + + xfs_bmap_init(&free_list, &first_block); + + /* + * We are using the write transaction in which max 2 bmbt + * updates are allowed + */ + error = xfs_bmap_shift_extents(tp, ip, &done, start_fsb, + shift_fsb, ¤t_ext, + &first_block, &free_list, + XFS_BMAP_MAX_SHIFT_EXTENTS); + if (error) + goto out; + + error = xfs_bmap_finish(&tp, &free_list, &committed); + if (error) + goto out; + + error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES); + xfs_iunlock(ip, XFS_ILOCK_EXCL); + } + + return error; + +out: + xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES | XFS_TRANS_ABORT); + xfs_iunlock(ip, XFS_ILOCK_EXCL); + return error; +} + +/* * We need to check that the format of the data fork in the temporary inode is * valid for the target inode before doing the swap. This is not a problem with * attr1 because of the fixed fork offset, but attr2 has a dynamically sized diff --git a/fs/xfs/xfs_bmap_util.h b/fs/xfs/xfs_bmap_util.h index 900747b..935ed2b 100644 --- a/fs/xfs/xfs_bmap_util.h +++ b/fs/xfs/xfs_bmap_util.h @@ -99,6 +99,8 @@ int xfs_free_file_space(struct xfs_inode *ip, xfs_off_t offset, xfs_off_t len); int xfs_zero_file_space(struct xfs_inode *ip, xfs_off_t offset, xfs_off_t len); +int xfs_collapse_file_space(struct xfs_inode *, xfs_off_t offset, + xfs_off_t len); /* EOF block manipulation functions */ bool xfs_can_free_eofblocks(struct xfs_inode *ip, bool force); diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c index 2e7989e..52f96e1 100644 --- a/fs/xfs/xfs_file.c +++ b/fs/xfs/xfs_file.c @@ -823,7 +823,8 @@ xfs_file_fallocate( if (!S_ISREG(inode->i_mode)) return -EINVAL; - if (mode & ~(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE)) + if (mode & ~(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE | + FALLOC_FL_COLLAPSE_RANGE)) return -EOPNOTSUPP; xfs_ilock(ip, XFS_IOLOCK_EXCL); @@ -831,6 +832,20 @@ xfs_file_fallocate( error = xfs_free_file_space(ip, offset, len); if (error) goto out_unlock; + } else if (mode & FALLOC_FL_COLLAPSE_RANGE) { + unsigned blksize_mask = (1 << inode->i_blkbits) - 1; + + if (offset & blksize_mask || len & blksize_mask) { + error = -EINVAL; + goto out_unlock; + } + + ASSERT(offset + len < i_size_read(inode)); + new_size = i_size_read(inode) - len; + + error = xfs_collapse_file_space(ip, offset, len); + if (error) + goto out_unlock; } else { if (!(mode & FALLOC_FL_KEEP_SIZE) && offset + len > i_size_read(inode)) { @@ -859,7 +874,7 @@ xfs_file_fallocate( if (ip->i_d.di_mode & S_IXGRP) ip->i_d.di_mode &= ~S_ISGID; - if (!(mode & FALLOC_FL_PUNCH_HOLE)) + if (!(mode & (FALLOC_FL_PUNCH_HOLE | FALLOC_FL_COLLAPSE_RANGE))) ip->i_d.di_flags |= XFS_DIFLAG_PREALLOC; xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index 425dfa4..a4ae41c 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -603,6 +603,7 @@ DEFINE_INODE_EVENT(xfs_readlink); DEFINE_INODE_EVENT(xfs_inactive_symlink); DEFINE_INODE_EVENT(xfs_alloc_file_space); DEFINE_INODE_EVENT(xfs_free_file_space); +DEFINE_INODE_EVENT(xfs_collapse_file_space); DEFINE_INODE_EVENT(xfs_readdir); #ifdef CONFIG_XFS_POSIX_ACL DEFINE_INODE_EVENT(xfs_get_acl); -- cgit v0.10.2 From e0d2c23a253149b4f8a6100f94ca62ddc4b2ae84 Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Thu, 27 Feb 2014 15:14:31 +1100 Subject: xfs: skip pointless CRC updates after verifier failures Most write verifiers don't update CRCs after the verifier has failed and the buffer has been marked in error. These two didn't, but should. Add returns to the verifier failure block, since the buffer won't be written anyway. Signed-off-by: Eric Sandeen Reviewed-by: Jie Liu Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_alloc_btree.c b/fs/xfs/xfs_alloc_btree.c index 1308542..144d3b0 100644 --- a/fs/xfs/xfs_alloc_btree.c +++ b/fs/xfs/xfs_alloc_btree.c @@ -373,6 +373,7 @@ xfs_allocbt_write_verify( XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, bp->b_target->bt_mount, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); + return; } xfs_btree_sblock_calc_crc(bp); diff --git a/fs/xfs/xfs_ialloc_btree.c b/fs/xfs/xfs_ialloc_btree.c index c8fa5bb..0028c50 100644 --- a/fs/xfs/xfs_ialloc_btree.c +++ b/fs/xfs/xfs_ialloc_btree.c @@ -261,6 +261,7 @@ xfs_inobt_write_verify( XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, bp->b_target->bt_mount, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); + return; } xfs_btree_sblock_calc_crc(bp); -- cgit v0.10.2 From 533b81c875589ad0a2fc116991534b4601195253 Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Thu, 27 Feb 2014 15:15:27 +1100 Subject: xfs: Use defines for CRC offsets in all cases Some calls to crc functions used useful #defines, others used awkward offsetof() constructs. Switch them all to #define to make things a bit cleaner. Signed-off-by: Eric Sandeen Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_ag.h b/fs/xfs/xfs_ag.h index 3fc1098..0fdd410 100644 --- a/fs/xfs/xfs_ag.h +++ b/fs/xfs/xfs_ag.h @@ -89,6 +89,8 @@ typedef struct xfs_agf { /* structure must be padded to 64 bit alignment */ } xfs_agf_t; +#define XFS_AGF_CRC_OFF offsetof(struct xfs_agf, agf_crc) + #define XFS_AGF_MAGICNUM 0x00000001 #define XFS_AGF_VERSIONNUM 0x00000002 #define XFS_AGF_SEQNO 0x00000004 @@ -167,6 +169,8 @@ typedef struct xfs_agi { /* structure must be padded to 64 bit alignment */ } xfs_agi_t; +#define XFS_AGI_CRC_OFF offsetof(struct xfs_agi, agi_crc) + #define XFS_AGI_MAGICNUM 0x00000001 #define XFS_AGI_VERSIONNUM 0x00000002 #define XFS_AGI_SEQNO 0x00000004 @@ -222,6 +226,8 @@ typedef struct xfs_agfl { __be32 agfl_bno[]; /* actually XFS_AGFL_SIZE(mp) */ } xfs_agfl_t; +#define XFS_AGFL_CRC_OFF offsetof(struct xfs_agfl, agfl_crc) + /* * tags for inode radix tree */ diff --git a/fs/xfs/xfs_alloc.c b/fs/xfs/xfs_alloc.c index 9eab2df..72ea855 100644 --- a/fs/xfs/xfs_alloc.c +++ b/fs/xfs/xfs_alloc.c @@ -486,7 +486,7 @@ xfs_agfl_read_verify( return; agfl_ok = xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length), - offsetof(struct xfs_agfl, agfl_crc)); + XFS_AGFL_CRC_OFF); agfl_ok = agfl_ok && xfs_agfl_verify(bp); @@ -516,8 +516,7 @@ xfs_agfl_write_verify( if (bip) XFS_BUF_TO_AGFL(bp)->agfl_lsn = cpu_to_be64(bip->bli_item.li_lsn); - xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), - offsetof(struct xfs_agfl, agfl_crc)); + xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), XFS_AGFL_CRC_OFF); } const struct xfs_buf_ops xfs_agfl_buf_ops = { @@ -2242,7 +2241,7 @@ xfs_agf_read_verify( if (xfs_sb_version_hascrc(&mp->m_sb)) agf_ok = xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length), - offsetof(struct xfs_agf, agf_crc)); + XFS_AGF_CRC_OFF); agf_ok = agf_ok && xfs_agf_verify(mp, bp); @@ -2272,8 +2271,7 @@ xfs_agf_write_verify( if (bip) XFS_BUF_TO_AGF(bp)->agf_lsn = cpu_to_be64(bip->bli_item.li_lsn); - xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), - offsetof(struct xfs_agf, agf_crc)); + xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), XFS_AGF_CRC_OFF); } const struct xfs_buf_ops xfs_agf_buf_ops = { diff --git a/fs/xfs/xfs_dinode.h b/fs/xfs/xfs_dinode.h index e5869b5..623bbe8 100644 --- a/fs/xfs/xfs_dinode.h +++ b/fs/xfs/xfs_dinode.h @@ -89,6 +89,8 @@ typedef struct xfs_dinode { /* structure must be padded to 64 bit alignment */ } xfs_dinode_t; +#define XFS_DINODE_CRC_OFF offsetof(struct xfs_dinode, di_crc) + #define DI_MAX_FLUSH 0xffff /* diff --git a/fs/xfs/xfs_format.h b/fs/xfs/xfs_format.h index b6ab5a3..9898f31 100644 --- a/fs/xfs/xfs_format.h +++ b/fs/xfs/xfs_format.h @@ -145,6 +145,8 @@ struct xfs_dsymlink_hdr { __be64 sl_lsn; }; +#define XFS_SYMLINK_CRC_OFF offsetof(struct xfs_dsymlink_hdr, sl_crc) + /* * The maximum pathlen is 1024 bytes. Since the minimum file system * blocksize is 512 bytes, we can get a max of 3 extents back from diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c index 5d7f105..d79210b 100644 --- a/fs/xfs/xfs_ialloc.c +++ b/fs/xfs/xfs_ialloc.c @@ -1572,7 +1572,7 @@ xfs_agi_read_verify( if (xfs_sb_version_hascrc(&mp->m_sb)) agi_ok = xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length), - offsetof(struct xfs_agi, agi_crc)); + XFS_AGI_CRC_OFF); agi_ok = agi_ok && xfs_agi_verify(bp); if (unlikely(XFS_TEST_ERROR(!agi_ok, mp, XFS_ERRTAG_IALLOC_READ_AGI, @@ -1600,8 +1600,7 @@ xfs_agi_write_verify( if (bip) XFS_BUF_TO_AGI(bp)->agi_lsn = cpu_to_be64(bip->bli_item.li_lsn); - xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), - offsetof(struct xfs_agi, agi_crc)); + xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), XFS_AGI_CRC_OFF); } const struct xfs_buf_ops xfs_agi_buf_ops = { diff --git a/fs/xfs/xfs_inode_buf.c b/fs/xfs/xfs_inode_buf.c index 4fc9f39..606b43a 100644 --- a/fs/xfs/xfs_inode_buf.c +++ b/fs/xfs/xfs_inode_buf.c @@ -306,7 +306,7 @@ xfs_dinode_verify( if (!xfs_sb_version_hascrc(&mp->m_sb)) return false; if (!xfs_verify_cksum((char *)dip, mp->m_sb.sb_inodesize, - offsetof(struct xfs_dinode, di_crc))) + XFS_DINODE_CRC_OFF)) return false; if (be64_to_cpu(dip->di_ino) != ip->i_ino) return false; @@ -327,7 +327,7 @@ xfs_dinode_calc_crc( ASSERT(xfs_sb_version_hascrc(&mp->m_sb)); crc = xfs_start_cksum((char *)dip, mp->m_sb.sb_inodesize, - offsetof(struct xfs_dinode, di_crc)); + XFS_DINODE_CRC_OFF); dip->di_crc = xfs_end_cksum(crc); } diff --git a/fs/xfs/xfs_sb.c b/fs/xfs/xfs_sb.c index 1e11679..1ea7c86 100644 --- a/fs/xfs/xfs_sb.c +++ b/fs/xfs/xfs_sb.c @@ -611,7 +611,7 @@ xfs_sb_read_verify( dsb->sb_crc != 0)) { if (!xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length), - offsetof(struct xfs_sb, sb_crc))) { + XFS_SB_CRC_OFF)) { /* Only fail bad secondaries on a known V5 filesystem */ if (bp->b_bn == XFS_SB_DADDR || xfs_sb_version_hascrc(&mp->m_sb)) { @@ -674,8 +674,7 @@ xfs_sb_write_verify( if (bip) XFS_BUF_TO_SBP(bp)->sb_lsn = cpu_to_be64(bip->bli_item.li_lsn); - xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), - offsetof(struct xfs_sb, sb_crc)); + xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), XFS_SB_CRC_OFF); } const struct xfs_buf_ops xfs_sb_buf_ops = { diff --git a/fs/xfs/xfs_sb.h b/fs/xfs/xfs_sb.h index 35061d4..f7b2fe7 100644 --- a/fs/xfs/xfs_sb.h +++ b/fs/xfs/xfs_sb.h @@ -182,6 +182,8 @@ typedef struct xfs_sb { /* must be padded to 64 bit alignment */ } xfs_sb_t; +#define XFS_SB_CRC_OFF offsetof(struct xfs_sb, sb_crc) + /* * Superblock - on disk version. Must match the in core version above. * Must be padded to 64 bit alignment. diff --git a/fs/xfs/xfs_symlink_remote.c b/fs/xfs/xfs_symlink_remote.c index bf59a2b..7a705a4 100644 --- a/fs/xfs/xfs_symlink_remote.c +++ b/fs/xfs/xfs_symlink_remote.c @@ -134,7 +134,7 @@ xfs_symlink_read_verify( return; if (!xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length), - offsetof(struct xfs_dsymlink_hdr, sl_crc)) || + XFS_SYMLINK_CRC_OFF) || !xfs_symlink_verify(bp)) { XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); @@ -162,8 +162,7 @@ xfs_symlink_write_verify( struct xfs_dsymlink_hdr *dsl = bp->b_addr; dsl->sl_lsn = cpu_to_be64(bip->bli_item.li_lsn); } - xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), - offsetof(struct xfs_dsymlink_hdr, sl_crc)); + xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), XFS_SYMLINK_CRC_OFF); } const struct xfs_buf_ops xfs_symlink_buf_ops = { -- cgit v0.10.2 From 5158217058fc2cdb92e05b9bb3c1a350d2a51ed9 Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Thu, 27 Feb 2014 15:17:27 +1100 Subject: xfs: add helper for verifying checksums on xfs_bufs Many/most callers of xfs_verify_cksum() pass bp->b_addr and BBTOB(bp->b_length) as the first 2 args. Add a helper which can just accept the bp and the crc offset, and work it out on its own, for brevity. Signed-off-by: Eric Sandeen Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_alloc.c b/fs/xfs/xfs_alloc.c index 72ea855..5050c9a 100644 --- a/fs/xfs/xfs_alloc.c +++ b/fs/xfs/xfs_alloc.c @@ -485,8 +485,7 @@ xfs_agfl_read_verify( if (!xfs_sb_version_hascrc(&mp->m_sb)) return; - agfl_ok = xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length), - XFS_AGFL_CRC_OFF); + agfl_ok = xfs_buf_verify_cksum(bp, XFS_AGFL_CRC_OFF); agfl_ok = agfl_ok && xfs_agfl_verify(bp); @@ -2240,8 +2239,7 @@ xfs_agf_read_verify( int agf_ok = 1; if (xfs_sb_version_hascrc(&mp->m_sb)) - agf_ok = xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length), - XFS_AGF_CRC_OFF); + agf_ok = xfs_buf_verify_cksum(bp, XFS_AGF_CRC_OFF); agf_ok = agf_ok && xfs_agf_verify(mp, bp); diff --git a/fs/xfs/xfs_attr_leaf.c b/fs/xfs/xfs_attr_leaf.c index 7b126f4..a19a023 100644 --- a/fs/xfs/xfs_attr_leaf.c +++ b/fs/xfs/xfs_attr_leaf.c @@ -240,8 +240,7 @@ xfs_attr3_leaf_read_verify( struct xfs_mount *mp = bp->b_target->bt_mount; if ((xfs_sb_version_hascrc(&mp->m_sb) && - !xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length), - XFS_ATTR3_LEAF_CRC_OFF)) || + !xfs_buf_verify_cksum(bp, XFS_ATTR3_LEAF_CRC_OFF)) || !xfs_attr3_leaf_verify(bp)) { XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); diff --git a/fs/xfs/xfs_btree.c b/fs/xfs/xfs_btree.c index 9adaae4..4e8524d 100644 --- a/fs/xfs/xfs_btree.c +++ b/fs/xfs/xfs_btree.c @@ -243,8 +243,8 @@ xfs_btree_lblock_verify_crc( struct xfs_buf *bp) { if (xfs_sb_version_hascrc(&bp->b_target->bt_mount->m_sb)) - return xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length), - XFS_BTREE_LBLOCK_CRC_OFF); + return xfs_buf_verify_cksum(bp, XFS_BTREE_LBLOCK_CRC_OFF); + return true; } @@ -276,8 +276,8 @@ xfs_btree_sblock_verify_crc( struct xfs_buf *bp) { if (xfs_sb_version_hascrc(&bp->b_target->bt_mount->m_sb)) - return xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length), - XFS_BTREE_SBLOCK_CRC_OFF); + return xfs_buf_verify_cksum(bp, XFS_BTREE_SBLOCK_CRC_OFF); + return true; } diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h index 9953395..5edcfba 100644 --- a/fs/xfs/xfs_buf.h +++ b/fs/xfs/xfs_buf.h @@ -369,6 +369,13 @@ static inline void xfs_buf_relse(xfs_buf_t *bp) xfs_buf_rele(bp); } +static inline int +xfs_buf_verify_cksum(struct xfs_buf *bp, unsigned long cksum_offset) +{ + return xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length), + cksum_offset); +} + /* * Handling of buftargs. */ diff --git a/fs/xfs/xfs_da_btree.c b/fs/xfs/xfs_da_btree.c index 796272a..6cece55 100644 --- a/fs/xfs/xfs_da_btree.c +++ b/fs/xfs/xfs_da_btree.c @@ -214,8 +214,7 @@ xfs_da3_node_read_verify( switch (be16_to_cpu(info->magic)) { case XFS_DA3_NODE_MAGIC: - if (!xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length), - XFS_DA3_NODE_CRC_OFF)) + if (!xfs_buf_verify_cksum(bp, XFS_DA3_NODE_CRC_OFF)) break; /* fall through */ case XFS_DA_NODE_MAGIC: diff --git a/fs/xfs/xfs_dir2_block.c b/fs/xfs/xfs_dir2_block.c index 90cdbf4..948dc39 100644 --- a/fs/xfs/xfs_dir2_block.c +++ b/fs/xfs/xfs_dir2_block.c @@ -90,8 +90,7 @@ xfs_dir3_block_read_verify( struct xfs_mount *mp = bp->b_target->bt_mount; if ((xfs_sb_version_hascrc(&mp->m_sb) && - !xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length), - XFS_DIR3_DATA_CRC_OFF)) || + !xfs_buf_verify_cksum(bp, XFS_DIR3_DATA_CRC_OFF)) || !xfs_dir3_block_verify(bp)) { XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); diff --git a/fs/xfs/xfs_dir2_data.c b/fs/xfs/xfs_dir2_data.c index 70acff4..1952f00 100644 --- a/fs/xfs/xfs_dir2_data.c +++ b/fs/xfs/xfs_dir2_data.c @@ -268,8 +268,7 @@ xfs_dir3_data_read_verify( struct xfs_mount *mp = bp->b_target->bt_mount; if ((xfs_sb_version_hascrc(&mp->m_sb) && - !xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length), - XFS_DIR3_DATA_CRC_OFF)) || + !xfs_buf_verify_cksum(bp, XFS_DIR3_DATA_CRC_OFF)) || !xfs_dir3_data_verify(bp)) { XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); diff --git a/fs/xfs/xfs_dir2_leaf.c b/fs/xfs/xfs_dir2_leaf.c index ae47ec6..1a412eb 100644 --- a/fs/xfs/xfs_dir2_leaf.c +++ b/fs/xfs/xfs_dir2_leaf.c @@ -180,8 +180,7 @@ __read_verify( struct xfs_mount *mp = bp->b_target->bt_mount; if ((xfs_sb_version_hascrc(&mp->m_sb) && - !xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length), - XFS_DIR3_LEAF_CRC_OFF)) || + !xfs_buf_verify_cksum(bp, XFS_DIR3_LEAF_CRC_OFF)) || !xfs_dir3_leaf_verify(bp, magic)) { XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); diff --git a/fs/xfs/xfs_dir2_node.c b/fs/xfs/xfs_dir2_node.c index 48c7d18..875e7c0 100644 --- a/fs/xfs/xfs_dir2_node.c +++ b/fs/xfs/xfs_dir2_node.c @@ -116,8 +116,7 @@ xfs_dir3_free_read_verify( struct xfs_mount *mp = bp->b_target->bt_mount; if ((xfs_sb_version_hascrc(&mp->m_sb) && - !xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length), - XFS_DIR3_FREE_CRC_OFF)) || + !xfs_buf_verify_cksum(bp, XFS_DIR3_FREE_CRC_OFF)) || !xfs_dir3_free_verify(bp)) { XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c index d79210b..d6a879d 100644 --- a/fs/xfs/xfs_ialloc.c +++ b/fs/xfs/xfs_ialloc.c @@ -1571,8 +1571,8 @@ xfs_agi_read_verify( int agi_ok = 1; if (xfs_sb_version_hascrc(&mp->m_sb)) - agi_ok = xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length), - XFS_AGI_CRC_OFF); + agi_ok = xfs_buf_verify_cksum(bp, XFS_AGI_CRC_OFF); + agi_ok = agi_ok && xfs_agi_verify(bp); if (unlikely(XFS_TEST_ERROR(!agi_ok, mp, XFS_ERRTAG_IALLOC_READ_AGI, diff --git a/fs/xfs/xfs_linux.h b/fs/xfs/xfs_linux.h index f9bb590..e8fed74 100644 --- a/fs/xfs/xfs_linux.h +++ b/fs/xfs/xfs_linux.h @@ -119,6 +119,7 @@ typedef __uint64_t __psunsigned_t; #include "xfs_iops.h" #include "xfs_aops.h" #include "xfs_super.h" +#include "xfs_cksum.h" #include "xfs_buf.h" #include "xfs_message.h" diff --git a/fs/xfs/xfs_sb.c b/fs/xfs/xfs_sb.c index 1ea7c86..36f287fa 100644 --- a/fs/xfs/xfs_sb.c +++ b/fs/xfs/xfs_sb.c @@ -610,8 +610,7 @@ xfs_sb_read_verify( XFS_SB_VERSION_5) || dsb->sb_crc != 0)) { - if (!xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length), - XFS_SB_CRC_OFF)) { + if (!xfs_buf_verify_cksum(bp, XFS_SB_CRC_OFF)) { /* Only fail bad secondaries on a known V5 filesystem */ if (bp->b_bn == XFS_SB_DADDR || xfs_sb_version_hascrc(&mp->m_sb)) { diff --git a/fs/xfs/xfs_symlink_remote.c b/fs/xfs/xfs_symlink_remote.c index 7a705a4..b172955 100644 --- a/fs/xfs/xfs_symlink_remote.c +++ b/fs/xfs/xfs_symlink_remote.c @@ -133,8 +133,7 @@ xfs_symlink_read_verify( if (!xfs_sb_version_hascrc(&mp->m_sb)) return; - if (!xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length), - XFS_SYMLINK_CRC_OFF) || + if (!xfs_buf_verify_cksum(bp, XFS_SYMLINK_CRC_OFF) || !xfs_symlink_verify(bp)) { XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); -- cgit v0.10.2 From f1dbcd7e38c80c2165516b4432231b46f1adc76c Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Thu, 27 Feb 2014 15:18:23 +1100 Subject: xfs: add helper for updating checksums on xfs_bufs Many/most callers of xfs_update_cksum() pass bp->b_addr and BBTOB(bp->b_length) as the first 2 args. Add a helper which can just accept the bp and the crc offset, and work it out on its own, for brevity. Signed-off-by: Eric Sandeen Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_alloc.c b/fs/xfs/xfs_alloc.c index 5050c9a..9c7cf3d 100644 --- a/fs/xfs/xfs_alloc.c +++ b/fs/xfs/xfs_alloc.c @@ -515,7 +515,7 @@ xfs_agfl_write_verify( if (bip) XFS_BUF_TO_AGFL(bp)->agfl_lsn = cpu_to_be64(bip->bli_item.li_lsn); - xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), XFS_AGFL_CRC_OFF); + xfs_buf_update_cksum(bp, XFS_AGFL_CRC_OFF); } const struct xfs_buf_ops xfs_agfl_buf_ops = { @@ -2269,7 +2269,7 @@ xfs_agf_write_verify( if (bip) XFS_BUF_TO_AGF(bp)->agf_lsn = cpu_to_be64(bip->bli_item.li_lsn); - xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), XFS_AGF_CRC_OFF); + xfs_buf_update_cksum(bp, XFS_AGF_CRC_OFF); } const struct xfs_buf_ops xfs_agf_buf_ops = { diff --git a/fs/xfs/xfs_attr_leaf.c b/fs/xfs/xfs_attr_leaf.c index a19a023..b552378 100644 --- a/fs/xfs/xfs_attr_leaf.c +++ b/fs/xfs/xfs_attr_leaf.c @@ -224,7 +224,7 @@ xfs_attr3_leaf_write_verify( if (bip) hdr3->info.lsn = cpu_to_be64(bip->bli_item.li_lsn); - xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), XFS_ATTR3_LEAF_CRC_OFF); + xfs_buf_update_cksum(bp, XFS_ATTR3_LEAF_CRC_OFF); } /* diff --git a/fs/xfs/xfs_btree.c b/fs/xfs/xfs_btree.c index 4e8524d..e80d59f 100644 --- a/fs/xfs/xfs_btree.c +++ b/fs/xfs/xfs_btree.c @@ -234,8 +234,7 @@ xfs_btree_lblock_calc_crc( return; if (bip) block->bb_u.l.bb_lsn = cpu_to_be64(bip->bli_item.li_lsn); - xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), - XFS_BTREE_LBLOCK_CRC_OFF); + xfs_buf_update_cksum(bp, XFS_BTREE_LBLOCK_CRC_OFF); } bool @@ -267,8 +266,7 @@ xfs_btree_sblock_calc_crc( return; if (bip) block->bb_u.s.bb_lsn = cpu_to_be64(bip->bli_item.li_lsn); - xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), - XFS_BTREE_SBLOCK_CRC_OFF); + xfs_buf_update_cksum(bp, XFS_BTREE_SBLOCK_CRC_OFF); } bool diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h index 5edcfba..b8a3abf 100644 --- a/fs/xfs/xfs_buf.h +++ b/fs/xfs/xfs_buf.h @@ -376,6 +376,13 @@ xfs_buf_verify_cksum(struct xfs_buf *bp, unsigned long cksum_offset) cksum_offset); } +static inline void +xfs_buf_update_cksum(struct xfs_buf *bp, unsigned long cksum_offset) +{ + xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), + cksum_offset); +} + /* * Handling of buftargs. */ diff --git a/fs/xfs/xfs_da_btree.c b/fs/xfs/xfs_da_btree.c index 6cece55..75ef990 100644 --- a/fs/xfs/xfs_da_btree.c +++ b/fs/xfs/xfs_da_btree.c @@ -196,7 +196,7 @@ xfs_da3_node_write_verify( if (bip) hdr3->info.lsn = cpu_to_be64(bip->bli_item.li_lsn); - xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), XFS_DA3_NODE_CRC_OFF); + xfs_buf_update_cksum(bp, XFS_DA3_NODE_CRC_OFF); } /* diff --git a/fs/xfs/xfs_dir2_block.c b/fs/xfs/xfs_dir2_block.c index 948dc39..724377e 100644 --- a/fs/xfs/xfs_dir2_block.c +++ b/fs/xfs/xfs_dir2_block.c @@ -117,7 +117,7 @@ xfs_dir3_block_write_verify( if (bip) hdr3->lsn = cpu_to_be64(bip->bli_item.li_lsn); - xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), XFS_DIR3_DATA_CRC_OFF); + xfs_buf_update_cksum(bp, XFS_DIR3_DATA_CRC_OFF); } const struct xfs_buf_ops xfs_dir3_block_buf_ops = { diff --git a/fs/xfs/xfs_dir2_data.c b/fs/xfs/xfs_dir2_data.c index 1952f00..74ae85e 100644 --- a/fs/xfs/xfs_dir2_data.c +++ b/fs/xfs/xfs_dir2_data.c @@ -295,7 +295,7 @@ xfs_dir3_data_write_verify( if (bip) hdr3->lsn = cpu_to_be64(bip->bli_item.li_lsn); - xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), XFS_DIR3_DATA_CRC_OFF); + xfs_buf_update_cksum(bp, XFS_DIR3_DATA_CRC_OFF); } const struct xfs_buf_ops xfs_dir3_data_buf_ops = { diff --git a/fs/xfs/xfs_dir2_leaf.c b/fs/xfs/xfs_dir2_leaf.c index 1a412eb..dffb61b 100644 --- a/fs/xfs/xfs_dir2_leaf.c +++ b/fs/xfs/xfs_dir2_leaf.c @@ -208,7 +208,7 @@ __write_verify( if (bip) hdr3->info.lsn = cpu_to_be64(bip->bli_item.li_lsn); - xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), XFS_DIR3_LEAF_CRC_OFF); + xfs_buf_update_cksum(bp, XFS_DIR3_LEAF_CRC_OFF); } static void diff --git a/fs/xfs/xfs_dir2_node.c b/fs/xfs/xfs_dir2_node.c index 875e7c0..0904b20 100644 --- a/fs/xfs/xfs_dir2_node.c +++ b/fs/xfs/xfs_dir2_node.c @@ -143,7 +143,7 @@ xfs_dir3_free_write_verify( if (bip) hdr3->lsn = cpu_to_be64(bip->bli_item.li_lsn); - xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), XFS_DIR3_FREE_CRC_OFF); + xfs_buf_update_cksum(bp, XFS_DIR3_FREE_CRC_OFF); } const struct xfs_buf_ops xfs_dir3_free_buf_ops = { diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c index d6a879d..4657586 100644 --- a/fs/xfs/xfs_ialloc.c +++ b/fs/xfs/xfs_ialloc.c @@ -1600,7 +1600,7 @@ xfs_agi_write_verify( if (bip) XFS_BUF_TO_AGI(bp)->agi_lsn = cpu_to_be64(bip->bli_item.li_lsn); - xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), XFS_AGI_CRC_OFF); + xfs_buf_update_cksum(bp, XFS_AGI_CRC_OFF); } const struct xfs_buf_ops xfs_agi_buf_ops = { diff --git a/fs/xfs/xfs_sb.c b/fs/xfs/xfs_sb.c index 36f287fa..818359f 100644 --- a/fs/xfs/xfs_sb.c +++ b/fs/xfs/xfs_sb.c @@ -673,7 +673,7 @@ xfs_sb_write_verify( if (bip) XFS_BUF_TO_SBP(bp)->sb_lsn = cpu_to_be64(bip->bli_item.li_lsn); - xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), XFS_SB_CRC_OFF); + xfs_buf_update_cksum(bp, XFS_SB_CRC_OFF); } const struct xfs_buf_ops xfs_sb_buf_ops = { diff --git a/fs/xfs/xfs_symlink_remote.c b/fs/xfs/xfs_symlink_remote.c index b172955..defa09f 100644 --- a/fs/xfs/xfs_symlink_remote.c +++ b/fs/xfs/xfs_symlink_remote.c @@ -161,7 +161,7 @@ xfs_symlink_write_verify( struct xfs_dsymlink_hdr *dsl = bp->b_addr; dsl->sl_lsn = cpu_to_be64(bip->bli_item.li_lsn); } - xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), XFS_SYMLINK_CRC_OFF); + xfs_buf_update_cksum(bp, XFS_SYMLINK_CRC_OFF); } const struct xfs_buf_ops xfs_symlink_buf_ops = { -- cgit v0.10.2 From ca23f8fdd6a0dd37b3909ce7709c491f0c26399d Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Thu, 27 Feb 2014 15:21:07 +1100 Subject: xfs: add xfs_verifier_error() We want to distinguish between corruption, CRC errors, etc. In addition, the full stack trace on verifier errors seems less than helpful; it looks more like an oops than corruption. Create a new function to specifically alert the user to verifier errors, which can differentiate between EFSCORRUPTED and CRC mismatches. It doesn't dump stack unless the xfs error level is turned up high. Define a new error message (EFSBADCRC) to clearly identify CRC errors. (Defined to EBADMSG, bad message) Signed-off-by: Eric Sandeen Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_error.c b/fs/xfs/xfs_error.c index 9995b80..a8b2ecb 100644 --- a/fs/xfs/xfs_error.c +++ b/fs/xfs/xfs_error.c @@ -178,3 +178,28 @@ xfs_corruption_error( xfs_error_report(tag, level, mp, filename, linenum, ra); xfs_alert(mp, "Corruption detected. Unmount and run xfs_repair"); } + +/* + * Warnings specifically for verifier errors. Differentiate CRC vs. invalid + * values, and omit the stack trace unless the error level is tuned high. + */ +void +xfs_verifier_error( + struct xfs_buf *bp) +{ + struct xfs_mount *mp = bp->b_target->bt_mount; + + xfs_alert(mp, "Metadata %s detected at %pF, block 0x%llx", + bp->b_error == EFSBADCRC ? "CRC error" : "corruption", + __return_address, bp->b_bn); + + xfs_alert(mp, "Unmount and run xfs_repair"); + + if (xfs_error_level >= XFS_ERRLEVEL_LOW) { + xfs_alert(mp, "First 64 bytes of corrupted metadata buffer:"); + xfs_hex_dump(xfs_buf_offset(bp, 0), 64); + } + + if (xfs_error_level >= XFS_ERRLEVEL_HIGH) + xfs_stack_trace(); +} diff --git a/fs/xfs/xfs_error.h b/fs/xfs/xfs_error.h index 079a367..c1c57d4 100644 --- a/fs/xfs/xfs_error.h +++ b/fs/xfs/xfs_error.h @@ -34,6 +34,7 @@ extern void xfs_error_report(const char *tag, int level, struct xfs_mount *mp, extern void xfs_corruption_error(const char *tag, int level, struct xfs_mount *mp, void *p, const char *filename, int linenum, inst_t *ra); +extern void xfs_verifier_error(struct xfs_buf *bp); #define XFS_ERROR_REPORT(e, lvl, mp) \ xfs_error_report(e, lvl, mp, __FILE__, __LINE__, __return_address) diff --git a/fs/xfs/xfs_linux.h b/fs/xfs/xfs_linux.h index e8fed74..825249d 100644 --- a/fs/xfs/xfs_linux.h +++ b/fs/xfs/xfs_linux.h @@ -179,6 +179,7 @@ typedef __uint64_t __psunsigned_t; #define ENOATTR ENODATA /* Attribute not found */ #define EWRONGFS EINVAL /* Mount with wrong filesystem type */ #define EFSCORRUPTED EUCLEAN /* Filesystem is corrupted */ +#define EFSBADCRC EBADMSG /* Bad CRC detected */ #define SYNCHRONIZE() barrier() #define __return_address __builtin_return_address(0) -- cgit v0.10.2 From db9355c296eb71271bb3807ad4a9d43f6b3c35d3 Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Thu, 27 Feb 2014 15:21:37 +1100 Subject: xfs: print useful caller information in xfs_error_report xfs_error_report used to just print the hex address of the caller; %pF will give us something more human-readable. Signed-off-by: Eric Sandeen Reviewed-by: Jie Liu Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_error.c b/fs/xfs/xfs_error.c index a8b2ecb..edac5b0 100644 --- a/fs/xfs/xfs_error.c +++ b/fs/xfs/xfs_error.c @@ -156,7 +156,7 @@ xfs_error_report( { if (level <= xfs_error_level) { xfs_alert_tag(mp, XFS_PTAG_ERROR_REPORT, - "Internal error %s at line %d of file %s. Caller 0x%p", + "Internal error %s at line %d of file %s. Caller %pF", tag, linenum, filename, ra); xfs_stack_trace(); -- cgit v0.10.2 From ce5028cfe3ca48695b6a128638fe224426d37ebe Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Thu, 27 Feb 2014 15:23:10 +1100 Subject: xfs: modify verifiers to differentiate CRC from other errors Modify all read & write verifiers to differentiate between CRC errors and other inconsistencies. This sets the appropriate error number on bp->b_error, and then calls xfs_verifier_error() if something went wrong. That function will issue the appropriate message to the user. Signed-off-by: Eric Sandeen Reviewed-by: Brian Foster Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_alloc.c b/fs/xfs/xfs_alloc.c index 9c7cf3d..c1cf6a3 100644 --- a/fs/xfs/xfs_alloc.c +++ b/fs/xfs/xfs_alloc.c @@ -474,7 +474,6 @@ xfs_agfl_read_verify( struct xfs_buf *bp) { struct xfs_mount *mp = bp->b_target->bt_mount; - int agfl_ok = 1; /* * There is no verification of non-crc AGFLs because mkfs does not @@ -485,14 +484,13 @@ xfs_agfl_read_verify( if (!xfs_sb_version_hascrc(&mp->m_sb)) return; - agfl_ok = xfs_buf_verify_cksum(bp, XFS_AGFL_CRC_OFF); - - agfl_ok = agfl_ok && xfs_agfl_verify(bp); - - if (!agfl_ok) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); + if (!xfs_buf_verify_cksum(bp, XFS_AGFL_CRC_OFF)) + xfs_buf_ioerror(bp, EFSBADCRC); + else if (!xfs_agfl_verify(bp)) xfs_buf_ioerror(bp, EFSCORRUPTED); - } + + if (bp->b_error) + xfs_verifier_error(bp); } static void @@ -507,8 +505,8 @@ xfs_agfl_write_verify( return; if (!xfs_agfl_verify(bp)) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); + xfs_verifier_error(bp); return; } @@ -2236,18 +2234,17 @@ xfs_agf_read_verify( struct xfs_buf *bp) { struct xfs_mount *mp = bp->b_target->bt_mount; - int agf_ok = 1; - - if (xfs_sb_version_hascrc(&mp->m_sb)) - agf_ok = xfs_buf_verify_cksum(bp, XFS_AGF_CRC_OFF); - agf_ok = agf_ok && xfs_agf_verify(mp, bp); - - if (unlikely(XFS_TEST_ERROR(!agf_ok, mp, XFS_ERRTAG_ALLOC_READ_AGF, - XFS_RANDOM_ALLOC_READ_AGF))) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); + if (xfs_sb_version_hascrc(&mp->m_sb) && + !xfs_buf_verify_cksum(bp, XFS_AGF_CRC_OFF)) + xfs_buf_ioerror(bp, EFSBADCRC); + else if (XFS_TEST_ERROR(!xfs_agf_verify(mp, bp), mp, + XFS_ERRTAG_ALLOC_READ_AGF, + XFS_RANDOM_ALLOC_READ_AGF)) xfs_buf_ioerror(bp, EFSCORRUPTED); - } + + if (bp->b_error) + xfs_verifier_error(bp); } static void @@ -2258,8 +2255,8 @@ xfs_agf_write_verify( struct xfs_buf_log_item *bip = bp->b_fspriv; if (!xfs_agf_verify(mp, bp)) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); + xfs_verifier_error(bp); return; } diff --git a/fs/xfs/xfs_alloc_btree.c b/fs/xfs/xfs_alloc_btree.c index 144d3b0..cc1eadc 100644 --- a/fs/xfs/xfs_alloc_btree.c +++ b/fs/xfs/xfs_alloc_btree.c @@ -355,12 +355,14 @@ static void xfs_allocbt_read_verify( struct xfs_buf *bp) { - if (!(xfs_btree_sblock_verify_crc(bp) && - xfs_allocbt_verify(bp))) { - trace_xfs_btree_corrupt(bp, _RET_IP_); - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, - bp->b_target->bt_mount, bp->b_addr); + if (!xfs_btree_sblock_verify_crc(bp)) + xfs_buf_ioerror(bp, EFSBADCRC); + else if (!xfs_allocbt_verify(bp)) xfs_buf_ioerror(bp, EFSCORRUPTED); + + if (bp->b_error) { + trace_xfs_btree_corrupt(bp, _RET_IP_); + xfs_verifier_error(bp); } } @@ -370,9 +372,8 @@ xfs_allocbt_write_verify( { if (!xfs_allocbt_verify(bp)) { trace_xfs_btree_corrupt(bp, _RET_IP_); - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, - bp->b_target->bt_mount, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); + xfs_verifier_error(bp); return; } xfs_btree_sblock_calc_crc(bp); diff --git a/fs/xfs/xfs_attr_leaf.c b/fs/xfs/xfs_attr_leaf.c index b552378..fe9587f 100644 --- a/fs/xfs/xfs_attr_leaf.c +++ b/fs/xfs/xfs_attr_leaf.c @@ -213,8 +213,8 @@ xfs_attr3_leaf_write_verify( struct xfs_attr3_leaf_hdr *hdr3 = bp->b_addr; if (!xfs_attr3_leaf_verify(bp)) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); + xfs_verifier_error(bp); return; } @@ -239,12 +239,14 @@ xfs_attr3_leaf_read_verify( { struct xfs_mount *mp = bp->b_target->bt_mount; - if ((xfs_sb_version_hascrc(&mp->m_sb) && - !xfs_buf_verify_cksum(bp, XFS_ATTR3_LEAF_CRC_OFF)) || - !xfs_attr3_leaf_verify(bp)) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); + if (xfs_sb_version_hascrc(&mp->m_sb) && + !xfs_buf_verify_cksum(bp, XFS_ATTR3_LEAF_CRC_OFF)) + xfs_buf_ioerror(bp, EFSBADCRC); + else if (!xfs_attr3_leaf_verify(bp)) xfs_buf_ioerror(bp, EFSCORRUPTED); - } + + if (bp->b_error) + xfs_verifier_error(bp); } const struct xfs_buf_ops xfs_attr3_leaf_buf_ops = { diff --git a/fs/xfs/xfs_attr_remote.c b/fs/xfs/xfs_attr_remote.c index 5549d69..6e37823 100644 --- a/fs/xfs/xfs_attr_remote.c +++ b/fs/xfs/xfs_attr_remote.c @@ -125,7 +125,6 @@ xfs_attr3_rmt_read_verify( struct xfs_mount *mp = bp->b_target->bt_mount; char *ptr; int len; - bool corrupt = false; xfs_daddr_t bno; /* no verification of non-crc buffers */ @@ -140,11 +139,11 @@ xfs_attr3_rmt_read_verify( while (len > 0) { if (!xfs_verify_cksum(ptr, XFS_LBSIZE(mp), XFS_ATTR3_RMT_CRC_OFF)) { - corrupt = true; + xfs_buf_ioerror(bp, EFSBADCRC); break; } if (!xfs_attr3_rmt_verify(mp, ptr, XFS_LBSIZE(mp), bno)) { - corrupt = true; + xfs_buf_ioerror(bp, EFSCORRUPTED); break; } len -= XFS_LBSIZE(mp); @@ -152,10 +151,9 @@ xfs_attr3_rmt_read_verify( bno += mp->m_bsize; } - if (corrupt) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); - xfs_buf_ioerror(bp, EFSCORRUPTED); - } else + if (bp->b_error) + xfs_verifier_error(bp); + else ASSERT(len == 0); } @@ -180,9 +178,8 @@ xfs_attr3_rmt_write_verify( while (len > 0) { if (!xfs_attr3_rmt_verify(mp, ptr, XFS_LBSIZE(mp), bno)) { - XFS_CORRUPTION_ERROR(__func__, - XFS_ERRLEVEL_LOW, mp, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); + xfs_verifier_error(bp); return; } if (bip) { diff --git a/fs/xfs/xfs_bmap_btree.c b/fs/xfs/xfs_bmap_btree.c index 706bc3f..818d546 100644 --- a/fs/xfs/xfs_bmap_btree.c +++ b/fs/xfs/xfs_bmap_btree.c @@ -780,12 +780,14 @@ static void xfs_bmbt_read_verify( struct xfs_buf *bp) { - if (!(xfs_btree_lblock_verify_crc(bp) && - xfs_bmbt_verify(bp))) { - trace_xfs_btree_corrupt(bp, _RET_IP_); - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, - bp->b_target->bt_mount, bp->b_addr); + if (!xfs_btree_lblock_verify_crc(bp)) + xfs_buf_ioerror(bp, EFSBADCRC); + else if (!xfs_bmbt_verify(bp)) xfs_buf_ioerror(bp, EFSCORRUPTED); + + if (bp->b_error) { + trace_xfs_btree_corrupt(bp, _RET_IP_); + xfs_verifier_error(bp); } } @@ -794,11 +796,9 @@ xfs_bmbt_write_verify( struct xfs_buf *bp) { if (!xfs_bmbt_verify(bp)) { - xfs_warn(bp->b_target->bt_mount, "bmbt daddr 0x%llx failed", bp->b_bn); trace_xfs_btree_corrupt(bp, _RET_IP_); - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, - bp->b_target->bt_mount, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); + xfs_verifier_error(bp); return; } xfs_btree_lblock_calc_crc(bp); diff --git a/fs/xfs/xfs_da_btree.c b/fs/xfs/xfs_da_btree.c index 75ef990..1f5af79 100644 --- a/fs/xfs/xfs_da_btree.c +++ b/fs/xfs/xfs_da_btree.c @@ -185,8 +185,8 @@ xfs_da3_node_write_verify( struct xfs_da3_node_hdr *hdr3 = bp->b_addr; if (!xfs_da3_node_verify(bp)) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); + xfs_verifier_error(bp); return; } @@ -209,17 +209,20 @@ static void xfs_da3_node_read_verify( struct xfs_buf *bp) { - struct xfs_mount *mp = bp->b_target->bt_mount; struct xfs_da_blkinfo *info = bp->b_addr; switch (be16_to_cpu(info->magic)) { case XFS_DA3_NODE_MAGIC: - if (!xfs_buf_verify_cksum(bp, XFS_DA3_NODE_CRC_OFF)) + if (!xfs_buf_verify_cksum(bp, XFS_DA3_NODE_CRC_OFF)) { + xfs_buf_ioerror(bp, EFSBADCRC); break; + } /* fall through */ case XFS_DA_NODE_MAGIC: - if (!xfs_da3_node_verify(bp)) + if (!xfs_da3_node_verify(bp)) { + xfs_buf_ioerror(bp, EFSCORRUPTED); break; + } return; case XFS_ATTR_LEAF_MAGIC: case XFS_ATTR3_LEAF_MAGIC: @@ -236,8 +239,7 @@ xfs_da3_node_read_verify( } /* corrupt block */ - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); - xfs_buf_ioerror(bp, EFSCORRUPTED); + xfs_verifier_error(bp); } const struct xfs_buf_ops xfs_da3_node_buf_ops = { diff --git a/fs/xfs/xfs_dir2_block.c b/fs/xfs/xfs_dir2_block.c index 724377e..4f6a38c 100644 --- a/fs/xfs/xfs_dir2_block.c +++ b/fs/xfs/xfs_dir2_block.c @@ -89,12 +89,14 @@ xfs_dir3_block_read_verify( { struct xfs_mount *mp = bp->b_target->bt_mount; - if ((xfs_sb_version_hascrc(&mp->m_sb) && - !xfs_buf_verify_cksum(bp, XFS_DIR3_DATA_CRC_OFF)) || - !xfs_dir3_block_verify(bp)) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); + if (xfs_sb_version_hascrc(&mp->m_sb) && + !xfs_buf_verify_cksum(bp, XFS_DIR3_DATA_CRC_OFF)) + xfs_buf_ioerror(bp, EFSBADCRC); + else if (!xfs_dir3_block_verify(bp)) xfs_buf_ioerror(bp, EFSCORRUPTED); - } + + if (bp->b_error) + xfs_verifier_error(bp); } static void @@ -106,8 +108,8 @@ xfs_dir3_block_write_verify( struct xfs_dir3_blk_hdr *hdr3 = bp->b_addr; if (!xfs_dir3_block_verify(bp)) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); + xfs_verifier_error(bp); return; } diff --git a/fs/xfs/xfs_dir2_data.c b/fs/xfs/xfs_dir2_data.c index 74ae85e..afa4ad5 100644 --- a/fs/xfs/xfs_dir2_data.c +++ b/fs/xfs/xfs_dir2_data.c @@ -241,7 +241,6 @@ static void xfs_dir3_data_reada_verify( struct xfs_buf *bp) { - struct xfs_mount *mp = bp->b_target->bt_mount; struct xfs_dir2_data_hdr *hdr = bp->b_addr; switch (hdr->magic) { @@ -255,8 +254,8 @@ xfs_dir3_data_reada_verify( xfs_dir3_data_verify(bp); return; default: - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, hdr); xfs_buf_ioerror(bp, EFSCORRUPTED); + xfs_verifier_error(bp); break; } } @@ -267,12 +266,14 @@ xfs_dir3_data_read_verify( { struct xfs_mount *mp = bp->b_target->bt_mount; - if ((xfs_sb_version_hascrc(&mp->m_sb) && - !xfs_buf_verify_cksum(bp, XFS_DIR3_DATA_CRC_OFF)) || - !xfs_dir3_data_verify(bp)) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); + if (xfs_sb_version_hascrc(&mp->m_sb) && + !xfs_buf_verify_cksum(bp, XFS_DIR3_DATA_CRC_OFF)) + xfs_buf_ioerror(bp, EFSBADCRC); + else if (!xfs_dir3_data_verify(bp)) xfs_buf_ioerror(bp, EFSCORRUPTED); - } + + if (bp->b_error) + xfs_verifier_error(bp); } static void @@ -284,8 +285,8 @@ xfs_dir3_data_write_verify( struct xfs_dir3_blk_hdr *hdr3 = bp->b_addr; if (!xfs_dir3_data_verify(bp)) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); + xfs_verifier_error(bp); return; } diff --git a/fs/xfs/xfs_dir2_leaf.c b/fs/xfs/xfs_dir2_leaf.c index dffb61b..d36e97d 100644 --- a/fs/xfs/xfs_dir2_leaf.c +++ b/fs/xfs/xfs_dir2_leaf.c @@ -179,12 +179,14 @@ __read_verify( { struct xfs_mount *mp = bp->b_target->bt_mount; - if ((xfs_sb_version_hascrc(&mp->m_sb) && - !xfs_buf_verify_cksum(bp, XFS_DIR3_LEAF_CRC_OFF)) || - !xfs_dir3_leaf_verify(bp, magic)) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); + if (xfs_sb_version_hascrc(&mp->m_sb) && + !xfs_buf_verify_cksum(bp, XFS_DIR3_LEAF_CRC_OFF)) + xfs_buf_ioerror(bp, EFSBADCRC); + else if (!xfs_dir3_leaf_verify(bp, magic)) xfs_buf_ioerror(bp, EFSCORRUPTED); - } + + if (bp->b_error) + xfs_verifier_error(bp); } static void @@ -197,8 +199,8 @@ __write_verify( struct xfs_dir3_leaf_hdr *hdr3 = bp->b_addr; if (!xfs_dir3_leaf_verify(bp, magic)) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); + xfs_verifier_error(bp); return; } diff --git a/fs/xfs/xfs_dir2_node.c b/fs/xfs/xfs_dir2_node.c index 0904b20..cb434d7 100644 --- a/fs/xfs/xfs_dir2_node.c +++ b/fs/xfs/xfs_dir2_node.c @@ -115,12 +115,14 @@ xfs_dir3_free_read_verify( { struct xfs_mount *mp = bp->b_target->bt_mount; - if ((xfs_sb_version_hascrc(&mp->m_sb) && - !xfs_buf_verify_cksum(bp, XFS_DIR3_FREE_CRC_OFF)) || - !xfs_dir3_free_verify(bp)) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); + if (xfs_sb_version_hascrc(&mp->m_sb) && + !xfs_buf_verify_cksum(bp, XFS_DIR3_FREE_CRC_OFF)) + xfs_buf_ioerror(bp, EFSBADCRC); + else if (!xfs_dir3_free_verify(bp)) xfs_buf_ioerror(bp, EFSCORRUPTED); - } + + if (bp->b_error) + xfs_verifier_error(bp); } static void @@ -132,8 +134,8 @@ xfs_dir3_free_write_verify( struct xfs_dir3_blk_hdr *hdr3 = bp->b_addr; if (!xfs_dir3_free_verify(bp)) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); + xfs_verifier_error(bp); return; } diff --git a/fs/xfs/xfs_dquot_buf.c b/fs/xfs/xfs_dquot_buf.c index d401457..610da81 100644 --- a/fs/xfs/xfs_dquot_buf.c +++ b/fs/xfs/xfs_dquot_buf.c @@ -257,10 +257,13 @@ xfs_dquot_buf_read_verify( { struct xfs_mount *mp = bp->b_target->bt_mount; - if (!xfs_dquot_buf_verify_crc(mp, bp) || !xfs_dquot_buf_verify(mp, bp)) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); + if (!xfs_dquot_buf_verify_crc(mp, bp)) + xfs_buf_ioerror(bp, EFSBADCRC); + else if (!xfs_dquot_buf_verify(mp, bp)) xfs_buf_ioerror(bp, EFSCORRUPTED); - } + + if (bp->b_error) + xfs_verifier_error(bp); } /* @@ -275,8 +278,8 @@ xfs_dquot_buf_write_verify( struct xfs_mount *mp = bp->b_target->bt_mount; if (!xfs_dquot_buf_verify(mp, bp)) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); + xfs_verifier_error(bp); return; } } diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c index 4657586..5959b3b 100644 --- a/fs/xfs/xfs_ialloc.c +++ b/fs/xfs/xfs_ialloc.c @@ -1568,18 +1568,17 @@ xfs_agi_read_verify( struct xfs_buf *bp) { struct xfs_mount *mp = bp->b_target->bt_mount; - int agi_ok = 1; - if (xfs_sb_version_hascrc(&mp->m_sb)) - agi_ok = xfs_buf_verify_cksum(bp, XFS_AGI_CRC_OFF); - - agi_ok = agi_ok && xfs_agi_verify(bp); - - if (unlikely(XFS_TEST_ERROR(!agi_ok, mp, XFS_ERRTAG_IALLOC_READ_AGI, - XFS_RANDOM_IALLOC_READ_AGI))) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); + if (xfs_sb_version_hascrc(&mp->m_sb) && + !xfs_buf_verify_cksum(bp, XFS_AGI_CRC_OFF)) + xfs_buf_ioerror(bp, EFSBADCRC); + else if (XFS_TEST_ERROR(!xfs_agi_verify(bp), mp, + XFS_ERRTAG_IALLOC_READ_AGI, + XFS_RANDOM_IALLOC_READ_AGI)) xfs_buf_ioerror(bp, EFSCORRUPTED); - } + + if (bp->b_error) + xfs_verifier_error(bp); } static void @@ -1590,8 +1589,8 @@ xfs_agi_write_verify( struct xfs_buf_log_item *bip = bp->b_fspriv; if (!xfs_agi_verify(bp)) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); + xfs_verifier_error(bp); return; } diff --git a/fs/xfs/xfs_ialloc_btree.c b/fs/xfs/xfs_ialloc_btree.c index 0028c50..7e309b1 100644 --- a/fs/xfs/xfs_ialloc_btree.c +++ b/fs/xfs/xfs_ialloc_btree.c @@ -243,12 +243,14 @@ static void xfs_inobt_read_verify( struct xfs_buf *bp) { - if (!(xfs_btree_sblock_verify_crc(bp) && - xfs_inobt_verify(bp))) { - trace_xfs_btree_corrupt(bp, _RET_IP_); - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, - bp->b_target->bt_mount, bp->b_addr); + if (!xfs_btree_sblock_verify_crc(bp)) + xfs_buf_ioerror(bp, EFSBADCRC); + else if (!xfs_inobt_verify(bp)) xfs_buf_ioerror(bp, EFSCORRUPTED); + + if (bp->b_error) { + trace_xfs_btree_corrupt(bp, _RET_IP_); + xfs_verifier_error(bp); } } @@ -258,9 +260,8 @@ xfs_inobt_write_verify( { if (!xfs_inobt_verify(bp)) { trace_xfs_btree_corrupt(bp, _RET_IP_); - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, - bp->b_target->bt_mount, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); + xfs_verifier_error(bp); return; } xfs_btree_sblock_calc_crc(bp); diff --git a/fs/xfs/xfs_inode_buf.c b/fs/xfs/xfs_inode_buf.c index 606b43a..24e9939 100644 --- a/fs/xfs/xfs_inode_buf.c +++ b/fs/xfs/xfs_inode_buf.c @@ -102,8 +102,7 @@ xfs_inode_buf_verify( } xfs_buf_ioerror(bp, EFSCORRUPTED); - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_HIGH, - mp, dip); + xfs_verifier_error(bp); #ifdef DEBUG xfs_alert(mp, "bad inode magic/vsn daddr %lld #%d (magic=%x)", diff --git a/fs/xfs/xfs_sb.c b/fs/xfs/xfs_sb.c index 818359f..b134aa8 100644 --- a/fs/xfs/xfs_sb.c +++ b/fs/xfs/xfs_sb.c @@ -614,7 +614,7 @@ xfs_sb_read_verify( /* Only fail bad secondaries on a known V5 filesystem */ if (bp->b_bn == XFS_SB_DADDR || xfs_sb_version_hascrc(&mp->m_sb)) { - error = EFSCORRUPTED; + error = EFSBADCRC; goto out_error; } } @@ -623,10 +623,9 @@ xfs_sb_read_verify( out_error: if (error) { - if (error == EFSCORRUPTED) - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, - mp, bp->b_addr); xfs_buf_ioerror(bp, error); + if (error == EFSCORRUPTED || error == EFSBADCRC) + xfs_verifier_error(bp); } } @@ -661,9 +660,8 @@ xfs_sb_write_verify( error = xfs_sb_verify(bp, false); if (error) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, - mp, bp->b_addr); xfs_buf_ioerror(bp, error); + xfs_verifier_error(bp); return; } diff --git a/fs/xfs/xfs_symlink_remote.c b/fs/xfs/xfs_symlink_remote.c index defa09f..9b32052 100644 --- a/fs/xfs/xfs_symlink_remote.c +++ b/fs/xfs/xfs_symlink_remote.c @@ -133,11 +133,13 @@ xfs_symlink_read_verify( if (!xfs_sb_version_hascrc(&mp->m_sb)) return; - if (!xfs_buf_verify_cksum(bp, XFS_SYMLINK_CRC_OFF) || - !xfs_symlink_verify(bp)) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); + if (!xfs_buf_verify_cksum(bp, XFS_SYMLINK_CRC_OFF)) + xfs_buf_ioerror(bp, EFSBADCRC); + else if (!xfs_symlink_verify(bp)) xfs_buf_ioerror(bp, EFSCORRUPTED); - } + + if (bp->b_error) + xfs_verifier_error(bp); } static void @@ -152,8 +154,8 @@ xfs_symlink_write_verify( return; if (!xfs_symlink_verify(bp)) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr); xfs_buf_ioerror(bp, EFSCORRUPTED); + xfs_verifier_error(bp); return; } -- cgit v0.10.2 From f876e44603ad091c840a5fae5b0753bbb421c037 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 27 Feb 2014 16:40:42 +1100 Subject: xfs: always do log forces via the workqueue Log forces can occur deep in the call chain when we have relatively little stack free. Log forces can also happen at close to the call chain leaves (e.g. xfs_buf_lock()) and hence we can trigger IO from places where we really don't want to add more stack overhead. This stack overhead occurs because log forces do foreground CIL pushes (xlog_cil_push_foreground()) rather than waking the background push wq and waiting for the for the push to complete. This foreground push was done to avoid confusing the CFQ Io scheduler when fsync()s were issued, as it has trouble dealing with dependent IOs being issued from different process contexts. Avoiding blowing the stack is much more critical than performance optimisations for CFQ, especially as we've been recommending against the use of CFQ for XFS since 3.2 kernels were release because of it's problems with multi-threaded IO workloads. Hence convert xlog_cil_push_foreground() to move the push work to the CIL workqueue. We already do the waiting for the push to complete in xlog_cil_force_lsn(), so there's nothing else we need to modify to make this work. Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_log_cil.c b/fs/xfs/xfs_log_cil.c index cdebd83..1270ded 100644 --- a/fs/xfs/xfs_log_cil.c +++ b/fs/xfs/xfs_log_cil.c @@ -488,13 +488,6 @@ xlog_cil_push( cil->xc_ctx = new_ctx; /* - * mirror the new sequence into the cil structure so that we can do - * unlocked checks against the current sequence in log forces without - * risking deferencing a freed context pointer. - */ - cil->xc_current_sequence = new_ctx->sequence; - - /* * The switch is now done, so we can drop the context lock and move out * of a shared context. We can't just go straight to the commit record, * though - we need to synchronise with previous and future commits so @@ -512,8 +505,15 @@ xlog_cil_push( * Hence we need to add this context to the committing context list so * that higher sequences will wait for us to write out a commit record * before they do. + * + * xfs_log_force_lsn requires us to mirror the new sequence into the cil + * structure atomically with the addition of this sequence to the + * committing list. This also ensures that we can do unlocked checks + * against the current sequence in log forces without risking + * deferencing a freed context pointer. */ spin_lock(&cil->xc_push_lock); + cil->xc_current_sequence = new_ctx->sequence; list_add(&ctx->committing, &cil->xc_committing); spin_unlock(&cil->xc_push_lock); up_write(&cil->xc_ctx_lock); @@ -651,8 +651,14 @@ xlog_cil_push_background( } +/* + * xlog_cil_push_now() is used to trigger an immediate CIL push to the sequence + * number that is passed. When it returns, the work will be queued for + * @push_seq, but it won't be completed. The caller is expected to do any + * waiting for push_seq to complete if it is required. + */ static void -xlog_cil_push_foreground( +xlog_cil_push_now( struct xlog *log, xfs_lsn_t push_seq) { @@ -677,10 +683,8 @@ xlog_cil_push_foreground( } cil->xc_push_seq = push_seq; + queue_work(log->l_mp->m_cil_workqueue, &cil->xc_push_work); spin_unlock(&cil->xc_push_lock); - - /* do the push now */ - xlog_cil_push(log); } bool @@ -785,7 +789,8 @@ xlog_cil_force_lsn( * xlog_cil_push() handles racing pushes for the same sequence, * so no need to deal with it here. */ - xlog_cil_push_foreground(log, sequence); +restart: + xlog_cil_push_now(log, sequence); /* * See if we can find a previous sequence still committing. @@ -793,7 +798,6 @@ xlog_cil_force_lsn( * before allowing the force of push_seq to go ahead. Hence block * on commits for those as well. */ -restart: spin_lock(&cil->xc_push_lock); list_for_each_entry(ctx, &cil->xc_committing, committing) { if (ctx->sequence > sequence) @@ -811,6 +815,28 @@ restart: /* found it! */ commit_lsn = ctx->commit_lsn; } + + /* + * The call to xlog_cil_push_now() executes the push in the background. + * Hence by the time we have got here it our sequence may not have been + * pushed yet. This is true if the current sequence still matches the + * push sequence after the above wait loop and the CIL still contains + * dirty objects. + * + * When the push occurs, it will empty the CIL and + * atomically increment the currect sequence past the push sequence and + * move it into the committing list. Of course, if the CIL is clean at + * the time of the push, it won't have pushed the CIL at all, so in that + * case we should try the push for this sequence again from the start + * just in case. + */ + + if (sequence == cil->xc_current_sequence && + !list_empty(&cil->xc_cil)) { + spin_unlock(&cil->xc_push_lock); + goto restart; + } + spin_unlock(&cil->xc_push_lock); return commit_lsn; } -- cgit v0.10.2 From a1358aa3d39251c16c6d8b20945e40bdfc2aad68 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 27 Feb 2014 16:51:26 +1100 Subject: xfs: allocate xfs_da_args to reduce stack footprint The struct xfs_da_args used to pass directory/attribute operation information to the lower layers is 128 bytes in size and is allocated on the stack. Dynamically allocate them to reduce the stack footprint of directory operations. Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_dir2.c b/fs/xfs/xfs_dir2.c index ce16ef0..fda4625 100644 --- a/fs/xfs/xfs_dir2.c +++ b/fs/xfs/xfs_dir2.c @@ -180,16 +180,23 @@ xfs_dir_init( xfs_inode_t *dp, xfs_inode_t *pdp) { - xfs_da_args_t args; + struct xfs_da_args *args; int error; - memset((char *)&args, 0, sizeof(args)); - args.dp = dp; - args.trans = tp; ASSERT(S_ISDIR(dp->i_d.di_mode)); - if ((error = xfs_dir_ino_validate(tp->t_mountp, pdp->i_ino))) + error = xfs_dir_ino_validate(tp->t_mountp, pdp->i_ino); + if (error) return error; - return xfs_dir2_sf_create(&args, pdp->i_ino); + + args = kmem_zalloc(sizeof(*args), KM_SLEEP | KM_NOFS); + if (!args) + return ENOMEM; + + args->dp = dp; + args->trans = tp; + error = xfs_dir2_sf_create(args, pdp->i_ino); + kmem_free(args); + return error; } /* @@ -205,41 +212,56 @@ xfs_dir_createname( xfs_bmap_free_t *flist, /* bmap's freeblock list */ xfs_extlen_t total) /* bmap's total block count */ { - xfs_da_args_t args; + struct xfs_da_args *args; int rval; int v; /* type-checking value */ ASSERT(S_ISDIR(dp->i_d.di_mode)); - if ((rval = xfs_dir_ino_validate(tp->t_mountp, inum))) + rval = xfs_dir_ino_validate(tp->t_mountp, inum); + if (rval) return rval; XFS_STATS_INC(xs_dir_create); - memset(&args, 0, sizeof(xfs_da_args_t)); - args.name = name->name; - args.namelen = name->len; - args.filetype = name->type; - args.hashval = dp->i_mount->m_dirnameops->hashname(name); - args.inumber = inum; - args.dp = dp; - args.firstblock = first; - args.flist = flist; - args.total = total; - args.whichfork = XFS_DATA_FORK; - args.trans = tp; - args.op_flags = XFS_DA_OP_ADDNAME | XFS_DA_OP_OKNOENT; - - if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL) - rval = xfs_dir2_sf_addname(&args); - else if ((rval = xfs_dir2_isblock(tp, dp, &v))) - return rval; - else if (v) - rval = xfs_dir2_block_addname(&args); - else if ((rval = xfs_dir2_isleaf(tp, dp, &v))) - return rval; - else if (v) - rval = xfs_dir2_leaf_addname(&args); + args = kmem_zalloc(sizeof(*args), KM_SLEEP | KM_NOFS); + if (!args) + return ENOMEM; + + args->name = name->name; + args->namelen = name->len; + args->filetype = name->type; + args->hashval = dp->i_mount->m_dirnameops->hashname(name); + args->inumber = inum; + args->dp = dp; + args->firstblock = first; + args->flist = flist; + args->total = total; + args->whichfork = XFS_DATA_FORK; + args->trans = tp; + args->op_flags = XFS_DA_OP_ADDNAME | XFS_DA_OP_OKNOENT; + + if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL) { + rval = xfs_dir2_sf_addname(args); + goto out_free; + } + + rval = xfs_dir2_isblock(tp, dp, &v); + if (rval) + goto out_free; + if (v) { + rval = xfs_dir2_block_addname(args); + goto out_free; + } + + rval = xfs_dir2_isleaf(tp, dp, &v); + if (rval) + goto out_free; + if (v) + rval = xfs_dir2_leaf_addname(args); else - rval = xfs_dir2_node_addname(&args); + rval = xfs_dir2_node_addname(args); + +out_free: + kmem_free(args); return rval; } @@ -282,46 +304,66 @@ xfs_dir_lookup( xfs_ino_t *inum, /* out: inode number */ struct xfs_name *ci_name) /* out: actual name if CI match */ { - xfs_da_args_t args; + struct xfs_da_args *args; int rval; int v; /* type-checking value */ ASSERT(S_ISDIR(dp->i_d.di_mode)); XFS_STATS_INC(xs_dir_lookup); - memset(&args, 0, sizeof(xfs_da_args_t)); - args.name = name->name; - args.namelen = name->len; - args.filetype = name->type; - args.hashval = dp->i_mount->m_dirnameops->hashname(name); - args.dp = dp; - args.whichfork = XFS_DATA_FORK; - args.trans = tp; - args.op_flags = XFS_DA_OP_OKNOENT; + /* + * We need to use KM_NOFS here so that lockdep will not throw false + * positive deadlock warnings on a non-transactional lookup path. It is + * safe to recurse into inode recalim in that case, but lockdep can't + * easily be taught about it. Hence KM_NOFS avoids having to add more + * lockdep Doing this avoids having to add a bunch of lockdep class + * annotations into the reclaim path for the ilock. + */ + args = kmem_zalloc(sizeof(*args), KM_SLEEP | KM_NOFS); + args->name = name->name; + args->namelen = name->len; + args->filetype = name->type; + args->hashval = dp->i_mount->m_dirnameops->hashname(name); + args->dp = dp; + args->whichfork = XFS_DATA_FORK; + args->trans = tp; + args->op_flags = XFS_DA_OP_OKNOENT; if (ci_name) - args.op_flags |= XFS_DA_OP_CILOOKUP; + args->op_flags |= XFS_DA_OP_CILOOKUP; - if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL) - rval = xfs_dir2_sf_lookup(&args); - else if ((rval = xfs_dir2_isblock(tp, dp, &v))) - return rval; - else if (v) - rval = xfs_dir2_block_lookup(&args); - else if ((rval = xfs_dir2_isleaf(tp, dp, &v))) - return rval; - else if (v) - rval = xfs_dir2_leaf_lookup(&args); + if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL) { + rval = xfs_dir2_sf_lookup(args); + goto out_check_rval; + } + + rval = xfs_dir2_isblock(tp, dp, &v); + if (rval) + goto out_free; + if (v) { + rval = xfs_dir2_block_lookup(args); + goto out_check_rval; + } + + rval = xfs_dir2_isleaf(tp, dp, &v); + if (rval) + goto out_free; + if (v) + rval = xfs_dir2_leaf_lookup(args); else - rval = xfs_dir2_node_lookup(&args); + rval = xfs_dir2_node_lookup(args); + +out_check_rval: if (rval == EEXIST) rval = 0; if (!rval) { - *inum = args.inumber; + *inum = args->inumber; if (ci_name) { - ci_name->name = args.value; - ci_name->len = args.valuelen; + ci_name->name = args->value; + ci_name->len = args->valuelen; } } +out_free: + kmem_free(args); return rval; } @@ -338,38 +380,51 @@ xfs_dir_removename( xfs_bmap_free_t *flist, /* bmap's freeblock list */ xfs_extlen_t total) /* bmap's total block count */ { - xfs_da_args_t args; + struct xfs_da_args *args; int rval; int v; /* type-checking value */ ASSERT(S_ISDIR(dp->i_d.di_mode)); XFS_STATS_INC(xs_dir_remove); - memset(&args, 0, sizeof(xfs_da_args_t)); - args.name = name->name; - args.namelen = name->len; - args.filetype = name->type; - args.hashval = dp->i_mount->m_dirnameops->hashname(name); - args.inumber = ino; - args.dp = dp; - args.firstblock = first; - args.flist = flist; - args.total = total; - args.whichfork = XFS_DATA_FORK; - args.trans = tp; - - if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL) - rval = xfs_dir2_sf_removename(&args); - else if ((rval = xfs_dir2_isblock(tp, dp, &v))) - return rval; - else if (v) - rval = xfs_dir2_block_removename(&args); - else if ((rval = xfs_dir2_isleaf(tp, dp, &v))) - return rval; - else if (v) - rval = xfs_dir2_leaf_removename(&args); + args = kmem_zalloc(sizeof(*args), KM_SLEEP | KM_NOFS); + if (!args) + return ENOMEM; + + args->name = name->name; + args->namelen = name->len; + args->filetype = name->type; + args->hashval = dp->i_mount->m_dirnameops->hashname(name); + args->inumber = ino; + args->dp = dp; + args->firstblock = first; + args->flist = flist; + args->total = total; + args->whichfork = XFS_DATA_FORK; + args->trans = tp; + + if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL) { + rval = xfs_dir2_sf_removename(args); + goto out_free; + } + + rval = xfs_dir2_isblock(tp, dp, &v); + if (rval) + goto out_free; + if (v) { + rval = xfs_dir2_block_removename(args); + goto out_free; + } + + rval = xfs_dir2_isleaf(tp, dp, &v); + if (rval) + goto out_free; + if (v) + rval = xfs_dir2_leaf_removename(args); else - rval = xfs_dir2_node_removename(&args); + rval = xfs_dir2_node_removename(args); +out_free: + kmem_free(args); return rval; } @@ -386,40 +441,54 @@ xfs_dir_replace( xfs_bmap_free_t *flist, /* bmap's freeblock list */ xfs_extlen_t total) /* bmap's total block count */ { - xfs_da_args_t args; + struct xfs_da_args *args; int rval; int v; /* type-checking value */ ASSERT(S_ISDIR(dp->i_d.di_mode)); - if ((rval = xfs_dir_ino_validate(tp->t_mountp, inum))) + rval = xfs_dir_ino_validate(tp->t_mountp, inum); + if (rval) return rval; - memset(&args, 0, sizeof(xfs_da_args_t)); - args.name = name->name; - args.namelen = name->len; - args.filetype = name->type; - args.hashval = dp->i_mount->m_dirnameops->hashname(name); - args.inumber = inum; - args.dp = dp; - args.firstblock = first; - args.flist = flist; - args.total = total; - args.whichfork = XFS_DATA_FORK; - args.trans = tp; - - if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL) - rval = xfs_dir2_sf_replace(&args); - else if ((rval = xfs_dir2_isblock(tp, dp, &v))) - return rval; - else if (v) - rval = xfs_dir2_block_replace(&args); - else if ((rval = xfs_dir2_isleaf(tp, dp, &v))) - return rval; - else if (v) - rval = xfs_dir2_leaf_replace(&args); + args = kmem_zalloc(sizeof(*args), KM_SLEEP | KM_NOFS); + if (!args) + return ENOMEM; + + args->name = name->name; + args->namelen = name->len; + args->filetype = name->type; + args->hashval = dp->i_mount->m_dirnameops->hashname(name); + args->inumber = inum; + args->dp = dp; + args->firstblock = first; + args->flist = flist; + args->total = total; + args->whichfork = XFS_DATA_FORK; + args->trans = tp; + + if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL) { + rval = xfs_dir2_sf_replace(args); + goto out_free; + } + + rval = xfs_dir2_isblock(tp, dp, &v); + if (rval) + goto out_free; + if (v) { + rval = xfs_dir2_block_replace(args); + goto out_free; + } + + rval = xfs_dir2_isleaf(tp, dp, &v); + if (rval) + goto out_free; + if (v) + rval = xfs_dir2_leaf_replace(args); else - rval = xfs_dir2_node_replace(&args); + rval = xfs_dir2_node_replace(args); +out_free: + kmem_free(args); return rval; } @@ -434,7 +503,7 @@ xfs_dir_canenter( struct xfs_name *name, /* name of entry to add */ uint resblks) { - xfs_da_args_t args; + struct xfs_da_args *args; int rval; int v; /* type-checking value */ @@ -443,29 +512,42 @@ xfs_dir_canenter( ASSERT(S_ISDIR(dp->i_d.di_mode)); - memset(&args, 0, sizeof(xfs_da_args_t)); - args.name = name->name; - args.namelen = name->len; - args.filetype = name->type; - args.hashval = dp->i_mount->m_dirnameops->hashname(name); - args.dp = dp; - args.whichfork = XFS_DATA_FORK; - args.trans = tp; - args.op_flags = XFS_DA_OP_JUSTCHECK | XFS_DA_OP_ADDNAME | + args = kmem_zalloc(sizeof(*args), KM_SLEEP | KM_NOFS); + if (!args) + return ENOMEM; + + args->name = name->name; + args->namelen = name->len; + args->filetype = name->type; + args->hashval = dp->i_mount->m_dirnameops->hashname(name); + args->dp = dp; + args->whichfork = XFS_DATA_FORK; + args->trans = tp; + args->op_flags = XFS_DA_OP_JUSTCHECK | XFS_DA_OP_ADDNAME | XFS_DA_OP_OKNOENT; - if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL) - rval = xfs_dir2_sf_addname(&args); - else if ((rval = xfs_dir2_isblock(tp, dp, &v))) - return rval; - else if (v) - rval = xfs_dir2_block_addname(&args); - else if ((rval = xfs_dir2_isleaf(tp, dp, &v))) - return rval; - else if (v) - rval = xfs_dir2_leaf_addname(&args); + if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL) { + rval = xfs_dir2_sf_addname(args); + goto out_free; + } + + rval = xfs_dir2_isblock(tp, dp, &v); + if (rval) + goto out_free; + if (v) { + rval = xfs_dir2_block_addname(args); + goto out_free; + } + + rval = xfs_dir2_isleaf(tp, dp, &v); + if (rval) + goto out_free; + if (v) + rval = xfs_dir2_leaf_addname(args); else - rval = xfs_dir2_node_addname(&args); + rval = xfs_dir2_node_addname(args); +out_free: + kmem_free(args); return rval; } -- cgit v0.10.2 From 93a8614e3a4dccd526aca34e892ac0b27f64b506 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 27 Feb 2014 16:51:39 +1100 Subject: xfs: fix directory inode iolock lockdep false positive The change to add the IO lock to protect the directory extent map during readdir operations has cause lockdep to have a heart attack as it now sees a different locking order on inodes w.r.t. the mmap_sem because readdir has a different ordering to write(). Add a new lockdep class for directory inodes to avoid this false positive. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c index f35d5c9..234e843 100644 --- a/fs/xfs/xfs_iops.c +++ b/fs/xfs/xfs_iops.c @@ -48,6 +48,18 @@ #include #include +/* + * Directories have different lock order w.r.t. mmap_sem compared to regular + * files. This is due to readdir potentially triggering page faults on a user + * buffer inside filldir(), and this happens with the ilock on the directory + * held. For regular files, the lock order is the other way around - the + * mmap_sem is taken during the page fault, and then we lock the ilock to do + * block mapping. Hence we need a different class for the directory ilock so + * that lockdep can tell them apart. + */ +static struct lock_class_key xfs_nondir_ilock_class; +static struct lock_class_key xfs_dir_ilock_class; + static int xfs_initxattrs( struct inode *inode, @@ -1191,6 +1203,7 @@ xfs_setup_inode( xfs_diflags_to_iflags(inode, ip); ip->d_ops = ip->i_mount->m_nondir_inode_ops; + lockdep_set_class(&ip->i_lock.mr_lock, &xfs_nondir_ilock_class); switch (inode->i_mode & S_IFMT) { case S_IFREG: inode->i_op = &xfs_inode_operations; @@ -1198,6 +1211,7 @@ xfs_setup_inode( inode->i_mapping->a_ops = &xfs_address_space_operations; break; case S_IFDIR: + lockdep_set_class(&ip->i_lock.mr_lock, &xfs_dir_ilock_class); if (xfs_sb_version_hasasciici(&XFS_M(inode->i_sb)->m_sb)) inode->i_op = &xfs_dir_ci_inode_operations; else -- cgit v0.10.2 From ac75a1f7a4af4dddcc1ac3c0778f0e3f75dc8f32 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Fri, 7 Mar 2014 16:19:14 +1100 Subject: xfs: don't leak EFSBADCRC to userspace While the verifier routines may return EFSBADCRC when a buffer has a bad CRC, we need to translate that to EFSCORRUPTED so that the higher layers treat the error appropriately and we return a consistent error to userspace. This fixes a xfs/005 regression. Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index 02df7b4..5c670f5 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -307,6 +307,9 @@ reread: error = bp->b_error; if (loud) xfs_warn(mp, "SB validate failed with error %d.", error); + /* bad CRC means corrupted metadata */ + if (error == EFSBADCRC) + error = EFSCORRUPTED; goto release_buf; } diff --git a/fs/xfs/xfs_symlink.c b/fs/xfs/xfs_symlink.c index 14e58f2..5fda189 100644 --- a/fs/xfs/xfs_symlink.c +++ b/fs/xfs/xfs_symlink.c @@ -80,6 +80,10 @@ xfs_readlink_bmap( if (error) { xfs_buf_ioerror_alert(bp, __func__); xfs_buf_relse(bp); + + /* bad CRC means corrupted metadata */ + if (error == EFSBADCRC) + error = EFSCORRUPTED; goto out; } byte_cnt = XFS_SYMLINK_BUF_SPACE(mp, byte_cnt); diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c index 647b6f1..b8eef05 100644 --- a/fs/xfs/xfs_trans_buf.c +++ b/fs/xfs/xfs_trans_buf.c @@ -275,6 +275,10 @@ xfs_trans_read_buf_map( XFS_BUF_UNDONE(bp); xfs_buf_stale(bp); xfs_buf_relse(bp); + + /* bad CRC means corrupted metadata */ + if (error == EFSBADCRC) + error = EFSCORRUPTED; return error; } #ifdef DEBUG @@ -338,6 +342,9 @@ xfs_trans_read_buf_map( if (tp->t_flags & XFS_TRANS_DIRTY) xfs_force_shutdown(tp->t_mountp, SHUTDOWN_META_IO_ERROR); + /* bad CRC means corrupted metadata */ + if (error == EFSBADCRC) + error = EFSCORRUPTED; return error; } } @@ -375,6 +382,10 @@ xfs_trans_read_buf_map( if (tp->t_flags & XFS_TRANS_DIRTY) xfs_force_shutdown(tp->t_mountp, SHUTDOWN_META_IO_ERROR); xfs_buf_relse(bp); + + /* bad CRC means corrupted metadata */ + if (error == EFSBADCRC) + error = EFSCORRUPTED; return error; } #ifdef DEBUG -- cgit v0.10.2 From ae687e58b3f09b1b3c0faf2cac8c27fbbefb5a48 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Fri, 7 Mar 2014 16:19:14 +1100 Subject: xfs: use NOIO contexts for vm_map_ram When we map pages in the buffer cache, we can do so in GFP_NOFS contexts. However, the vmap interfaces do not provide any method of communicating this information to memory reclaim, and hence we get lockdep complaining about it regularly and occassionally see hangs that may be vmap related reclaim deadlocks. We can also see these same problems from anywhere where we use vmalloc for a large buffer (e.g. attribute code) inside a transaction context. A typical lockdep report shows up as a reclaim state warning like so: [14046.101458] ================================= [14046.102850] [ INFO: inconsistent lock state ] [14046.102850] 3.14.0-rc4+ #2 Not tainted [14046.102850] --------------------------------- [14046.102850] inconsistent {RECLAIM_FS-ON-W} -> {IN-RECLAIM_FS-W} usage. [14046.102850] kswapd0/14 [HC0[0]:SC0[0]:HE1:SE1] takes: [14046.102850] (&xfs_dir_ilock_class){++++?+}, at: [<791a04bb>] xfs_ilock+0xff/0x16a [14046.102850] {RECLAIM_FS-ON-W} state was registered at: [14046.102850] [<7904cdb1>] mark_held_locks+0x81/0xe7 [14046.102850] [<7904d390>] lockdep_trace_alloc+0x5c/0xb4 [14046.102850] [<790c2c28>] kmem_cache_alloc_trace+0x2b/0x11e [14046.102850] [<790ba7f4>] vm_map_ram+0x119/0x3e6 [14046.102850] [<7914e124>] _xfs_buf_map_pages+0x5b/0xcf [14046.102850] [<7914ed74>] xfs_buf_get_map+0x67/0x13f [14046.102850] [<7917506f>] xfs_attr_rmtval_set+0x396/0x4d5 [14046.102850] [<7916e8bb>] xfs_attr_leaf_addname+0x18f/0x37d [14046.102850] [<7916ed9e>] xfs_attr_set_int+0x2f5/0x3e8 [14046.102850] [<7916eefc>] xfs_attr_set+0x6b/0x74 [14046.102850] [<79168355>] xfs_xattr_set+0x61/0x81 [14046.102850] [<790e5b10>] generic_setxattr+0x59/0x68 [14046.102850] [<790e4c06>] __vfs_setxattr_noperm+0x58/0xce [14046.102850] [<790e4d0a>] vfs_setxattr+0x8e/0x92 [14046.102850] [<790e4ddd>] setxattr+0xcf/0x159 [14046.102850] [<790e5423>] SyS_lsetxattr+0x88/0xbb [14046.102850] [<79268438>] sysenter_do_call+0x12/0x36 Now, we can't completely remove these traces - mainly because vm_map_ram() will do GFP_KERNEL allocation and that generates the above warning before we get into the reclaim code, but we can turn them all into false positive warnings. To do that, use the method that DM and other IO context code uses to avoid this problem: there is a process flag to tell memory reclaim not to do IO that we can set appropriately. That prevents GFP_KERNEL context reclaim being done from deep inside the vmalloc code in places we can't directly pass a GFP_NOFS context to. That interface has a pair of wrapper functions: memalloc_noio_save() and memalloc_noio_restore(). Adding them around vm_map_ram and the vzalloc call in kmem_alloc_large() will prevent deadlocks and most lockdep reports for this issue. Also, convert the vzalloc() call in kmem_alloc_large() to use __vmalloc() so that we can pass the correct gfp context to the data page allocation routine inside __vmalloc() so that it is clear that GFP_NOFS context is important to this vmalloc call. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner diff --git a/fs/xfs/kmem.c b/fs/xfs/kmem.c index 66a36be..844e288 100644 --- a/fs/xfs/kmem.c +++ b/fs/xfs/kmem.c @@ -65,12 +65,31 @@ kmem_alloc(size_t size, xfs_km_flags_t flags) void * kmem_zalloc_large(size_t size, xfs_km_flags_t flags) { + unsigned noio_flag = 0; void *ptr; + gfp_t lflags; ptr = kmem_zalloc(size, flags | KM_MAYFAIL); if (ptr) return ptr; - return vzalloc(size); + + /* + * __vmalloc() will allocate data pages and auxillary structures (e.g. + * pagetables) with GFP_KERNEL, yet we may be under GFP_NOFS context + * here. Hence we need to tell memory reclaim that we are in such a + * context via PF_MEMALLOC_NOIO to prevent memory reclaim re-entering + * the filesystem here and potentially deadlocking. + */ + if ((current->flags & PF_FSTRANS) || (flags & KM_NOFS)) + noio_flag = memalloc_noio_save(); + + lflags = kmem_flags_convert(flags); + ptr = __vmalloc(size, lflags | __GFP_HIGHMEM | __GFP_ZERO, PAGE_KERNEL); + + if ((current->flags & PF_FSTRANS) || (flags & KM_NOFS)) + memalloc_noio_restore(noio_flag); + + return ptr; } void diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index 9c061ef..107f2fd 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -396,7 +396,17 @@ _xfs_buf_map_pages( bp->b_addr = NULL; } else { int retried = 0; + unsigned noio_flag; + /* + * vm_map_ram() will allocate auxillary structures (e.g. + * pagetables) with GFP_KERNEL, yet we are likely to be under + * GFP_NOFS context here. Hence we need to tell memory reclaim + * that we are in such a context via PF_MEMALLOC_NOIO to prevent + * memory reclaim re-entering the filesystem here and + * potentially deadlocking. + */ + noio_flag = memalloc_noio_save(); do { bp->b_addr = vm_map_ram(bp->b_pages, bp->b_page_count, -1, PAGE_KERNEL); @@ -404,6 +414,7 @@ _xfs_buf_map_pages( break; vm_unmap_aliases(); } while (retried++ <= 1); + memalloc_noio_restore(noio_flag); if (!bp->b_addr) return -ENOMEM; -- cgit v0.10.2 From e480a7239723afe579060239564884d1fa4c9325 Mon Sep 17 00:00:00 2001 From: Brian Foster Date: Fri, 7 Mar 2014 16:19:14 +1100 Subject: xfs: avoid AGI/AGF deadlock scenario for inode chunk allocation The inode chunk allocation path can lead to deadlock conditions if a transaction is dirtied with an AGF (to fix up the freelist) for an AG that cannot satisfy the actual allocation request. This code path is written to try and avoid this scenario, but it can be reproduced by running xfstests generic/270 in a loop on a 512b fs. An example situation is: - process A attempts an inode allocation on AG 3, modifies the freelist, fails the allocation and ultimately moves on to AG 0 with the AG 3 AGF held - process B is doing a free space operation (i.e., truncate) and acquires the AG 0 AGF, waits on the AG 3 AGF - process A acquires the AG 0 AGI, waits on the AG 0 AGF (deadlock) The problem here is that process A acquired the AG 3 AGF while moving on to AG 0 (and releasing the AG 3 AGI with the AG 3 AGF held). xfs_dialloc() makes one pass through each of the AGs when attempting to allocate an inode chunk. The expectation is a clean transaction if a particular AG cannot satisfy the allocation request. xfs_ialloc_ag_alloc() is written to support this through use of the minalignslop allocation args field. When using the agi->agi_newino optimization, we attempt an exact bno allocation request based on the location of the previously allocated chunk. minalignslop is set to inform the allocator that we will require alignment on this chunk, and thus to not allow the request for this AG if the extra space is not available. Suppose that the AG in question has just enough space for this request, but not at the requested bno. xfs_alloc_fix_freelist() will proceed as normal as it determines the request should succeed, and thus it is allowed to modify the agf. xfs_alloc_ag_vextent() ultimately fails because the requested bno is not available. In response, the caller moves on to a NEAR_BNO allocation request for the same AG. The alignment is set, but the minalignslop field is never reset. This increases the overall requirement of the request from the first attempt. If this delta is the difference between allocation success and failure for the AG, xfs_alloc_fix_freelist() rejects this request outright the second time around and causes the allocation request to unnecessarily fail for this AG. To address this situation, reset the minalignslop field immediately after use and prevent it from leaking into subsequent requests. Signed-off-by: Brian Foster Reviewed-by: Mark Tinguely Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c index 5d7f105..283a76d 100644 --- a/fs/xfs/xfs_ialloc.c +++ b/fs/xfs/xfs_ialloc.c @@ -363,6 +363,18 @@ xfs_ialloc_ag_alloc( args.minleft = args.mp->m_in_maxlevels - 1; if ((error = xfs_alloc_vextent(&args))) return error; + + /* + * This request might have dirtied the transaction if the AG can + * satisfy the request, but the exact block was not available. + * If the allocation did fail, subsequent requests will relax + * the exact agbno requirement and increase the alignment + * instead. It is critical that the total size of the request + * (len + alignment + slop) does not increase from this point + * on, so reset minalignslop to ensure it is not included in + * subsequent requests. + */ + args.minalignslop = 0; } else args.fsbno = NULLFSBLOCK; -- cgit v0.10.2 From a49935f200e24e95fffcc705014c4b60ad78ff1f Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Fri, 7 Mar 2014 16:19:14 +1100 Subject: xfs: xfs_check_page_type buffer checks need help xfs_aops_discard_page() was introduced in the following commit: xfs: truncate delalloc extents when IO fails in writeback ... to clean up left over delalloc ranges after I/O failure in ->writepage(). generic/224 tests for this scenario and occasionally reproduces panics on sub-4k blocksize filesystems. The cause of this is failure to clean up the delalloc range on a page where the first buffer does not match one of the expected states of xfs_check_page_type(). If a buffer is not unwritten, delayed or dirty&mapped, xfs_check_page_type() stops and immediately returns 0. The stress test of generic/224 creates a scenario where the first several buffers of a page with delayed buffers are mapped & uptodate and some subsequent buffer is delayed. If the ->writepage() happens to fail for this page, xfs_aops_discard_page() incorrectly skips the entire page. This then causes later failures either when direct IO maps the range and finds the stale delayed buffer, or we evict the inode and find that the inode still has a delayed block reservation accounted to it. We can easily fix this xfs_aops_discard_page() failure by making xfs_check_page_type() check all buffers, but this breaks xfs_convert_page() more than it is already broken. Indeed, xfs_convert_page() wants xfs_check_page_type() to tell it if the first buffers on the pages are of a type that can be aggregated into the contiguous IO that is already being built. xfs_convert_page() should not be writing random buffers out of a page, but the current behaviour will cause it to do so if there are buffers that don't match the current specification on the page. Hence for xfs_convert_page() we need to: a) return "not ok" if the first buffer on the page does not match the specification provided to we don't write anything; and b) abort it's buffer-add-to-io loop the moment we come across a buffer that does not match the specification. Hence we need to fix both xfs_check_page_type() and xfs_convert_page() to work correctly with pages that have mixed buffer types, whilst allowing xfs_aops_discard_page() to scan all buffers on the page for a type match. Reported-by: Brian Foster Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c index db2cfb0..5935cce 100644 --- a/fs/xfs/xfs_aops.c +++ b/fs/xfs/xfs_aops.c @@ -632,38 +632,46 @@ xfs_map_at_offset( } /* - * Test if a given page is suitable for writing as part of an unwritten - * or delayed allocate extent. + * Test if a given page contains at least one buffer of a given @type. + * If @check_all_buffers is true, then we walk all the buffers in the page to + * try to find one of the type passed in. If it is not set, then the caller only + * needs to check the first buffer on the page for a match. */ -STATIC int +STATIC bool xfs_check_page_type( struct page *page, - unsigned int type) + unsigned int type, + bool check_all_buffers) { - if (PageWriteback(page)) - return 0; + struct buffer_head *bh; + struct buffer_head *head; - if (page->mapping && page_has_buffers(page)) { - struct buffer_head *bh, *head; - int acceptable = 0; + if (PageWriteback(page)) + return false; + if (!page->mapping) + return false; + if (!page_has_buffers(page)) + return false; - bh = head = page_buffers(page); - do { - if (buffer_unwritten(bh)) - acceptable += (type == XFS_IO_UNWRITTEN); - else if (buffer_delay(bh)) - acceptable += (type == XFS_IO_DELALLOC); - else if (buffer_dirty(bh) && buffer_mapped(bh)) - acceptable += (type == XFS_IO_OVERWRITE); - else - break; - } while ((bh = bh->b_this_page) != head); + bh = head = page_buffers(page); + do { + if (buffer_unwritten(bh)) { + if (type == XFS_IO_UNWRITTEN) + return true; + } else if (buffer_delay(bh)) { + if (type == XFS_IO_DELALLOC); + return true; + } else if (buffer_dirty(bh) && buffer_mapped(bh)) { + if (type == XFS_IO_OVERWRITE); + return true; + } - if (acceptable) - return 1; - } + /* If we are only checking the first buffer, we are done now. */ + if (!check_all_buffers) + break; + } while ((bh = bh->b_this_page) != head); - return 0; + return false; } /* @@ -697,7 +705,7 @@ xfs_convert_page( goto fail_unlock_page; if (page->mapping != inode->i_mapping) goto fail_unlock_page; - if (!xfs_check_page_type(page, (*ioendp)->io_type)) + if (!xfs_check_page_type(page, (*ioendp)->io_type, false)) goto fail_unlock_page; /* @@ -742,6 +750,15 @@ xfs_convert_page( p_offset = p_offset ? roundup(p_offset, len) : PAGE_CACHE_SIZE; page_dirty = p_offset / len; + /* + * The moment we find a buffer that doesn't match our current type + * specification or can't be written, abort the loop and start + * writeback. As per the above xfs_imap_valid() check, only + * xfs_vm_writepage() can handle partial page writeback fully - we are + * limited here to the buffers that are contiguous with the current + * ioend, and hence a buffer we can't write breaks that contiguity and + * we have to defer the rest of the IO to xfs_vm_writepage(). + */ bh = head = page_buffers(page); do { if (offset >= end_offset) @@ -750,7 +767,7 @@ xfs_convert_page( uptodate = 0; if (!(PageUptodate(page) || buffer_uptodate(bh))) { done = 1; - continue; + break; } if (buffer_unwritten(bh) || buffer_delay(bh) || @@ -762,10 +779,11 @@ xfs_convert_page( else type = XFS_IO_OVERWRITE; - if (!xfs_imap_valid(inode, imap, offset)) { - done = 1; - continue; - } + /* + * imap should always be valid because of the above + * partial page end_offset check on the imap. + */ + ASSERT(xfs_imap_valid(inode, imap, offset)); lock_buffer(bh); if (type != XFS_IO_OVERWRITE) @@ -777,6 +795,7 @@ xfs_convert_page( count++; } else { done = 1; + break; } } while (offset += len, (bh = bh->b_this_page) != head); @@ -868,7 +887,7 @@ xfs_aops_discard_page( struct buffer_head *bh, *head; loff_t offset = page_offset(page); - if (!xfs_check_page_type(page, XFS_IO_DELALLOC)) + if (!xfs_check_page_type(page, XFS_IO_DELALLOC, true)) goto out_invalidate; if (XFS_FORCED_SHUTDOWN(ip->i_mount)) -- cgit v0.10.2 From fe4c224aa1ffa4352849ac5f452de7132739bee2 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Fri, 7 Mar 2014 16:19:14 +1100 Subject: xfs: inode log reservations are still too small Back in commit 23956703 ("xfs: inode log reservations are too small"), the reservation size was increased to take into account the difference in size between the in-memory BMBT block headers and the on-disk BMDR headers. This solved a transaction overrun when logging the inode size. Recently, however, we've seen a number of these same overruns on kernels with the above fix in it. All of them have been by 4 bytes, so we must still not be accounting for something correctly. Through inspection it turns out the above commit didn't take into account everything it should have. That is, it only accounts for a single log op_hdr structure, when it can actually require up to four op_hdrs - one for each region (log iovec) that is formatted. These regions are the inode log format header, the inode core, and the two forks that can be held in the literal area of the inode. This means we are not accounting for 36 bytes of log space that the transaction can use, and hence when we get inodes in certain formats with particular fragmentation patterns we can overrun the transaction. Fix this by adding the correct accounting for log op_headers in the transaction. Tested-by: Brian Foster Signed-off-by: Dave Chinner Reviewed-by: Eric Sandeen Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_trans_resv.c b/fs/xfs/xfs_trans_resv.c index 2ffd3e33..c486e4b 100644 --- a/fs/xfs/xfs_trans_resv.c +++ b/fs/xfs/xfs_trans_resv.c @@ -81,20 +81,28 @@ xfs_calc_buf_res( * on disk. Hence we need an inode reservation function that calculates all this * correctly. So, we log: * - * - log op headers for object + * - 4 log op headers for object + * - for the ilf, the inode core and 2 forks * - inode log format object - * - the entire inode contents (core + 2 forks) - * - two bmap btree block headers + * - the inode core + * - two inode forks containing bmap btree root blocks. + * - the btree data contained by both forks will fit into the inode size, + * hence when combined with the inode core above, we have a total of the + * actual inode size. + * - the BMBT headers need to be accounted separately, as they are + * additional to the records and pointers that fit inside the inode + * forks. */ STATIC uint xfs_calc_inode_res( struct xfs_mount *mp, uint ninodes) { - return ninodes * (sizeof(struct xlog_op_header) + - sizeof(struct xfs_inode_log_format) + - mp->m_sb.sb_inodesize + - 2 * XFS_BMBT_BLOCK_LEN(mp)); + return ninodes * + (4 * sizeof(struct xlog_op_header) + + sizeof(struct xfs_inode_log_format) + + mp->m_sb.sb_inodesize + + 2 * XFS_BMBT_BLOCK_LEN(mp)); } /* -- cgit v0.10.2 From 409332b65d3ed8cfa7a8030f1e9d52f372219642 Mon Sep 17 00:00:00 2001 From: Lukas Czerner Date: Thu, 13 Mar 2014 19:07:42 +1100 Subject: fs: Introduce FALLOC_FL_ZERO_RANGE flag for fallocate Introduce new FALLOC_FL_ZERO_RANGE flag for fallocate. This has the same functionality as xfs ioctl XFS_IOC_ZERO_RANGE. It can be used to convert a range of file to zeros preferably without issuing data IO. Blocks should be preallocated for the regions that span holes in the file, and the entire range is preferable converted to unwritten extents - even though file system may choose to zero out the extent or do whatever which will result in reading zeros from the range while the range remains allocated for the file. This can be also used to preallocate blocks past EOF in the same way as with fallocate. Flag FALLOC_FL_KEEP_SIZE which should cause the inode size to remain the same. Signed-off-by: Lukas Czerner Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner diff --git a/fs/open.c b/fs/open.c index 4a923a5..c4465b2 100644 --- a/fs/open.c +++ b/fs/open.c @@ -232,7 +232,12 @@ int do_fallocate(struct file *file, int mode, loff_t offset, loff_t len) /* Return error if mode is not supported */ if (mode & ~(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE | - FALLOC_FL_COLLAPSE_RANGE)) + FALLOC_FL_COLLAPSE_RANGE | FALLOC_FL_ZERO_RANGE)) + return -EOPNOTSUPP; + + /* Punch hole and zero range are mutually exclusive */ + if ((mode & (FALLOC_FL_PUNCH_HOLE | FALLOC_FL_ZERO_RANGE)) == + (FALLOC_FL_PUNCH_HOLE | FALLOC_FL_ZERO_RANGE)) return -EOPNOTSUPP; /* Punch hole must have keep size set */ diff --git a/include/uapi/linux/falloc.h b/include/uapi/linux/falloc.h index 5ff562d..d1197ae 100644 --- a/include/uapi/linux/falloc.h +++ b/include/uapi/linux/falloc.h @@ -27,4 +27,18 @@ */ #define FALLOC_FL_COLLAPSE_RANGE 0x08 +/* + * FALLOC_FL_ZERO_RANGE is used to convert a range of file to zeros preferably + * without issuing data IO. Blocks should be preallocated for the regions that + * span holes in the file, and the entire range is preferable converted to + * unwritten extents - even though file system may choose to zero out the + * extent or do whatever which will result in reading zeros from the range + * while the range remains allocated for the file. + * + * This can be also used to preallocate blocks past EOF in the same way as + * with fallocate. Flag FALLOC_FL_KEEP_SIZE should cause the inode + * size to remain the same. + */ +#define FALLOC_FL_ZERO_RANGE 0x10 + #endif /* _UAPI_FALLOC_H_ */ -- cgit v0.10.2 From 376ba313147b4172f3e8cf620b9fb591f3e8cdfa Mon Sep 17 00:00:00 2001 From: Lukas Czerner Date: Thu, 13 Mar 2014 19:07:58 +1100 Subject: xfs: Add support for FALLOC_FL_ZERO_RANGE Introduce new FALLOC_FL_ZERO_RANGE flag for fallocate. This has the same functionality as xfs ioctl XFS_IOC_ZERO_RANGE. We can also preallocate blocks past EOF in the same was as with fallocate. Flag FALLOC_FL_KEEP_SIZE will cause the inode size to remain the same even if we preallocate blocks past EOF. It uses the same code to zero range as it is used by the XFS_IOC_ZERO_RANGE ioctl. Signed-off-by: Lukas Czerner Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c index 52f96e1..8fb97a6 100644 --- a/fs/xfs/xfs_file.c +++ b/fs/xfs/xfs_file.c @@ -824,7 +824,7 @@ xfs_file_fallocate( if (!S_ISREG(inode->i_mode)) return -EINVAL; if (mode & ~(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE | - FALLOC_FL_COLLAPSE_RANGE)) + FALLOC_FL_COLLAPSE_RANGE | FALLOC_FL_ZERO_RANGE)) return -EOPNOTSUPP; xfs_ilock(ip, XFS_IOLOCK_EXCL); @@ -855,8 +855,11 @@ xfs_file_fallocate( goto out_unlock; } - error = xfs_alloc_file_space(ip, offset, len, - XFS_BMAPI_PREALLOC); + if (mode & FALLOC_FL_ZERO_RANGE) + error = xfs_zero_file_space(ip, offset, len); + else + error = xfs_alloc_file_space(ip, offset, len, + XFS_BMAPI_PREALLOC); if (error) goto out_unlock; } -- cgit v0.10.2 From 805eeb8e04706a16cb0b23fd4c4abbb0bc7df82d Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Fri, 4 Apr 2014 06:56:30 +1100 Subject: xfs: extra semi-colon breaks a condition There were some extra semi-colons here which mean that we return true unintentionally. Fixes: a49935f200e2 ('xfs: xfs_check_page_type buffer checks need help') Signed-off-by: Dan Carpenter Reviewed-by: Brian Foster Reviewed-by: Eric Sandeen Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c index 5935cce..dba4885 100644 --- a/fs/xfs/xfs_aops.c +++ b/fs/xfs/xfs_aops.c @@ -659,10 +659,10 @@ xfs_check_page_type( if (type == XFS_IO_UNWRITTEN) return true; } else if (buffer_delay(bh)) { - if (type == XFS_IO_DELALLOC); + if (type == XFS_IO_DELALLOC) return true; } else if (buffer_dirty(bh) && buffer_mapped(bh)) { - if (type == XFS_IO_OVERWRITE); + if (type == XFS_IO_OVERWRITE) return true; } -- cgit v0.10.2 From c88547a8119e3b581318ab65e9b72f27f23e641d Mon Sep 17 00:00:00 2001 From: Mark Tinguely Date: Fri, 4 Apr 2014 07:10:49 +1100 Subject: xfs: fix directory hash ordering bug Commit f5ea1100 ("xfs: add CRCs to dir2/da node blocks") introduced in 3.10 incorrectly converted the btree hash index array pointer in xfs_da3_fixhashpath(). It resulted in the the current hash always being compared against the first entry in the btree rather than the current block index into the btree block's hash entry array. As a result, it was comparing the wrong hashes, and so could misorder the entries in the btree. For most cases, this doesn't cause any problems as it requires hash collisions to expose the ordering problem. However, when there are hash collisions within a directory there is a very good probability that the entries will be ordered incorrectly and that actually matters when duplicate hashes are placed into or removed from the btree block hash entry array. This bug results in an on-disk directory corruption and that results in directory verifier functions throwing corruption warnings into the logs. While no data or directory entries are lost, access to them may be compromised, and attempts to remove entries from a directory that has suffered from this corruption may result in a filesystem shutdown. xfs_repair will fix the directory hash ordering without data loss occuring. [dchinner: wrote useful a commit message] cc: Reported-by: Hannes Frederic Sowa Signed-off-by: Mark Tinguely Reviewed-by: Ben Myers Signed-off-by: Dave Chinner diff --git a/fs/xfs/xfs_da_btree.c b/fs/xfs/xfs_da_btree.c index 796272a..e69d57b 100644 --- a/fs/xfs/xfs_da_btree.c +++ b/fs/xfs/xfs_da_btree.c @@ -1295,7 +1295,7 @@ xfs_da3_fixhashpath( node = blk->bp->b_addr; dp->d_ops->node_hdr_from_disk(&nodehdr, node); btree = dp->d_ops->node_tree_p(node); - if (be32_to_cpu(btree->hashval) == lasthash) + if (be32_to_cpu(btree[blk->index].hashval) == lasthash) break; blk->hashval = lasthash; btree[blk->index].hashval = cpu_to_be32(lasthash); -- cgit v0.10.2