diff options
Diffstat (limited to 'fs/xfs/xfs_inode.c')
| -rw-r--r-- | fs/xfs/xfs_inode.c | 1713 |
1 files changed, 447 insertions, 1266 deletions
diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index d55b42b2480d..7dc6f326936c 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -16,6 +16,7 @@ #include "xfs_inode.h" #include "xfs_dir2.h" #include "xfs_attr.h" +#include "xfs_bit.h" #include "xfs_trans_space.h" #include "xfs_trans.h" #include "xfs_buf_item.h" @@ -38,56 +39,13 @@ #include "xfs_ag.h" #include "xfs_log_priv.h" #include "xfs_health.h" +#include "xfs_pnfs.h" +#include "xfs_parent.h" +#include "xfs_xattr.h" +#include "xfs_inode_util.h" struct kmem_cache *xfs_inode_cache; -STATIC int xfs_iunlink(struct xfs_trans *, struct xfs_inode *); -STATIC int xfs_iunlink_remove(struct xfs_trans *tp, struct xfs_perag *pag, - struct xfs_inode *); - -/* - * helper function to extract extent size hint from inode - */ -xfs_extlen_t -xfs_get_extsz_hint( - struct xfs_inode *ip) -{ - /* - * No point in aligning allocations if we need to COW to actually - * write to them. - */ - if (xfs_is_always_cow_inode(ip)) - return 0; - if ((ip->i_diflags & XFS_DIFLAG_EXTSIZE) && ip->i_extsize) - return ip->i_extsize; - if (XFS_IS_REALTIME_INODE(ip)) - return ip->i_mount->m_sb.sb_rextsize; - return 0; -} - -/* - * Helper function to extract CoW extent size hint from inode. - * Between the extent size hint and the CoW extent size hint, we - * return the greater of the two. If the value is zero (automatic), - * use the default size. - */ -xfs_extlen_t -xfs_get_cowextsz_hint( - struct xfs_inode *ip) -{ - xfs_extlen_t a, b; - - a = 0; - if (ip->i_diflags2 & XFS_DIFLAG2_COWEXTSIZE) - a = ip->i_cowextsize; - b = xfs_get_extsz_hint(ip); - - a = max(a, b); - if (a == 0) - return XFS_DEFAULT_COWEXTSZ_HINT; - return a; -} - /* * These two are wrapper routines around the xfs_ilock() routine used to * centralize some grungy code. They are used in places that wish to lock the @@ -420,7 +378,7 @@ xfs_lock_inumorder( * lock more than one at a time, lockdep will report false positives saying we * have violated locking orders. */ -static void +void xfs_lock_inodes( struct xfs_inode **ips, int inodes, @@ -565,55 +523,6 @@ xfs_lock_two_inodes( } } -uint -xfs_ip2xflags( - struct xfs_inode *ip) -{ - uint flags = 0; - - if (ip->i_diflags & XFS_DIFLAG_ANY) { - if (ip->i_diflags & XFS_DIFLAG_REALTIME) - flags |= FS_XFLAG_REALTIME; - if (ip->i_diflags & XFS_DIFLAG_PREALLOC) - flags |= FS_XFLAG_PREALLOC; - if (ip->i_diflags & XFS_DIFLAG_IMMUTABLE) - flags |= FS_XFLAG_IMMUTABLE; - if (ip->i_diflags & XFS_DIFLAG_APPEND) - flags |= FS_XFLAG_APPEND; - if (ip->i_diflags & XFS_DIFLAG_SYNC) - flags |= FS_XFLAG_SYNC; - if (ip->i_diflags & XFS_DIFLAG_NOATIME) - flags |= FS_XFLAG_NOATIME; - if (ip->i_diflags & XFS_DIFLAG_NODUMP) - flags |= FS_XFLAG_NODUMP; - if (ip->i_diflags & XFS_DIFLAG_RTINHERIT) - flags |= FS_XFLAG_RTINHERIT; - if (ip->i_diflags & XFS_DIFLAG_PROJINHERIT) - flags |= FS_XFLAG_PROJINHERIT; - if (ip->i_diflags & XFS_DIFLAG_NOSYMLINKS) - flags |= FS_XFLAG_NOSYMLINKS; - if (ip->i_diflags & XFS_DIFLAG_EXTSIZE) - flags |= FS_XFLAG_EXTSIZE; - if (ip->i_diflags & XFS_DIFLAG_EXTSZINHERIT) - flags |= FS_XFLAG_EXTSZINHERIT; - if (ip->i_diflags & XFS_DIFLAG_NODEFRAG) - flags |= FS_XFLAG_NODEFRAG; - if (ip->i_diflags & XFS_DIFLAG_FILESTREAM) - flags |= FS_XFLAG_FILESTREAM; - } - - if (ip->i_diflags2 & XFS_DIFLAG2_ANY) { - if (ip->i_diflags2 & XFS_DIFLAG2_DAX) - flags |= FS_XFLAG_DAX; - if (ip->i_diflags2 & XFS_DIFLAG2_COWEXTSIZE) - flags |= FS_XFLAG_COWEXTSIZE; - } - - if (xfs_inode_has_attr_fork(ip)) - flags |= FS_XFLAG_HASATTR; - return flags; -} - /* * Lookups up an inode from "name". If ci_name is not NULL, then a CI match * is allowed, otherwise it has to be an exact match. If a CI match is found, @@ -655,135 +564,22 @@ out_unlock: return error; } -/* Propagate di_flags from a parent inode to a child inode. */ -static void -xfs_inode_inherit_flags( - struct xfs_inode *ip, - const struct xfs_inode *pip) -{ - unsigned int di_flags = 0; - xfs_failaddr_t failaddr; - umode_t mode = VFS_I(ip)->i_mode; - - if (S_ISDIR(mode)) { - if (pip->i_diflags & XFS_DIFLAG_RTINHERIT) - di_flags |= XFS_DIFLAG_RTINHERIT; - if (pip->i_diflags & XFS_DIFLAG_EXTSZINHERIT) { - di_flags |= XFS_DIFLAG_EXTSZINHERIT; - ip->i_extsize = pip->i_extsize; - } - if (pip->i_diflags & XFS_DIFLAG_PROJINHERIT) - di_flags |= XFS_DIFLAG_PROJINHERIT; - } else if (S_ISREG(mode)) { - if ((pip->i_diflags & XFS_DIFLAG_RTINHERIT) && - xfs_has_realtime(ip->i_mount)) - di_flags |= XFS_DIFLAG_REALTIME; - if (pip->i_diflags & XFS_DIFLAG_EXTSZINHERIT) { - di_flags |= XFS_DIFLAG_EXTSIZE; - ip->i_extsize = pip->i_extsize; - } - } - if ((pip->i_diflags & XFS_DIFLAG_NOATIME) && - xfs_inherit_noatime) - di_flags |= XFS_DIFLAG_NOATIME; - if ((pip->i_diflags & XFS_DIFLAG_NODUMP) && - xfs_inherit_nodump) - di_flags |= XFS_DIFLAG_NODUMP; - if ((pip->i_diflags & XFS_DIFLAG_SYNC) && - xfs_inherit_sync) - di_flags |= XFS_DIFLAG_SYNC; - if ((pip->i_diflags & XFS_DIFLAG_NOSYMLINKS) && - xfs_inherit_nosymlinks) - di_flags |= XFS_DIFLAG_NOSYMLINKS; - if ((pip->i_diflags & XFS_DIFLAG_NODEFRAG) && - xfs_inherit_nodefrag) - di_flags |= XFS_DIFLAG_NODEFRAG; - if (pip->i_diflags & XFS_DIFLAG_FILESTREAM) - di_flags |= XFS_DIFLAG_FILESTREAM; - - ip->i_diflags |= di_flags; - - /* - * Inode verifiers on older kernels only check that the extent size - * hint is an integer multiple of the rt extent size on realtime files. - * They did not check the hint alignment on a directory with both - * rtinherit and extszinherit flags set. If the misaligned hint is - * propagated from a directory into a new realtime file, new file - * allocations will fail due to math errors in the rt allocator and/or - * trip the verifiers. Validate the hint settings in the new file so - * that we don't let broken hints propagate. - */ - failaddr = xfs_inode_validate_extsize(ip->i_mount, ip->i_extsize, - VFS_I(ip)->i_mode, ip->i_diflags); - if (failaddr) { - ip->i_diflags &= ~(XFS_DIFLAG_EXTSIZE | - XFS_DIFLAG_EXTSZINHERIT); - ip->i_extsize = 0; - } -} - -/* Propagate di_flags2 from a parent inode to a child inode. */ -static void -xfs_inode_inherit_flags2( - struct xfs_inode *ip, - const struct xfs_inode *pip) -{ - xfs_failaddr_t failaddr; - - if (pip->i_diflags2 & XFS_DIFLAG2_COWEXTSIZE) { - ip->i_diflags2 |= XFS_DIFLAG2_COWEXTSIZE; - ip->i_cowextsize = pip->i_cowextsize; - } - if (pip->i_diflags2 & XFS_DIFLAG2_DAX) - ip->i_diflags2 |= XFS_DIFLAG2_DAX; - - /* Don't let invalid cowextsize hints propagate. */ - failaddr = xfs_inode_validate_cowextsize(ip->i_mount, ip->i_cowextsize, - VFS_I(ip)->i_mode, ip->i_diflags, ip->i_diflags2); - if (failaddr) { - ip->i_diflags2 &= ~XFS_DIFLAG2_COWEXTSIZE; - ip->i_cowextsize = 0; - } -} - /* * Initialise a newly allocated inode and return the in-core inode to the * caller locked exclusively. + * + * Caller is responsible for unlocking the inode manually upon return */ int -xfs_init_new_inode( - struct mnt_idmap *idmap, +xfs_icreate( struct xfs_trans *tp, - struct xfs_inode *pip, xfs_ino_t ino, - umode_t mode, - xfs_nlink_t nlink, - dev_t rdev, - prid_t prid, - bool init_xattrs, + const struct xfs_icreate_args *args, struct xfs_inode **ipp) { - struct inode *dir = pip ? VFS_I(pip) : NULL; struct xfs_mount *mp = tp->t_mountp; - struct xfs_inode *ip; - unsigned int flags; + struct xfs_inode *ip = NULL; int error; - struct timespec64 tv; - struct inode *inode; - - /* - * Protect against obviously corrupt allocation btree records. Later - * xfs_iget checks will catch re-allocation of other active in-memory - * and on-disk inodes. If we don't catch reallocating the parent inode - * here we will deadlock in xfs_iget() so we have to do these checks - * first. - */ - if ((pip && ino == pip->i_ino) || !xfs_verify_dir_ino(mp, ino)) { - xfs_alert(mp, "Allocated a known in-use inode 0x%llx!", ino); - xfs_agno_mark_sick(mp, XFS_INO_TO_AGNO(mp, ino), - XFS_SICK_AG_INOBT); - return -EFSCORRUPTED; - } /* * Get the in-core inode with the lock held exclusively to prevent @@ -794,89 +590,8 @@ xfs_init_new_inode( return error; ASSERT(ip != NULL); - inode = VFS_I(ip); - set_nlink(inode, nlink); - inode->i_rdev = rdev; - ip->i_projid = prid; - - if (dir && !(dir->i_mode & S_ISGID) && xfs_has_grpid(mp)) { - inode_fsuid_set(inode, idmap); - inode->i_gid = dir->i_gid; - inode->i_mode = mode; - } else { - inode_init_owner(idmap, inode, dir, mode); - } - - /* - * If the group ID of the new file does not match the effective group - * ID or one of the supplementary group IDs, the S_ISGID bit is cleared - * (and only if the irix_sgid_inherit compatibility variable is set). - */ - if (irix_sgid_inherit && (inode->i_mode & S_ISGID) && - !vfsgid_in_group_p(i_gid_into_vfsgid(idmap, inode))) - inode->i_mode &= ~S_ISGID; - - ip->i_disk_size = 0; - ip->i_df.if_nextents = 0; - ASSERT(ip->i_nblocks == 0); - - tv = inode_set_ctime_current(inode); - inode_set_mtime_to_ts(inode, tv); - inode_set_atime_to_ts(inode, tv); - - ip->i_extsize = 0; - ip->i_diflags = 0; - - if (xfs_has_v3inodes(mp)) { - inode_set_iversion(inode, 1); - ip->i_cowextsize = 0; - ip->i_crtime = tv; - } - - flags = XFS_ILOG_CORE; - switch (mode & S_IFMT) { - case S_IFIFO: - case S_IFCHR: - case S_IFBLK: - case S_IFSOCK: - ip->i_df.if_format = XFS_DINODE_FMT_DEV; - flags |= XFS_ILOG_DEV; - break; - case S_IFREG: - case S_IFDIR: - if (pip && (pip->i_diflags & XFS_DIFLAG_ANY)) - xfs_inode_inherit_flags(ip, pip); - if (pip && (pip->i_diflags2 & XFS_DIFLAG2_ANY)) - xfs_inode_inherit_flags2(ip, pip); - fallthrough; - case S_IFLNK: - ip->i_df.if_format = XFS_DINODE_FMT_EXTENTS; - ip->i_df.if_bytes = 0; - ip->i_df.if_data = NULL; - break; - default: - ASSERT(0); - } - - /* - * If we need to create attributes immediately after allocating the - * inode, initialise an empty attribute fork right now. We use the - * default fork offset for attributes here as we don't know exactly what - * size or how many attributes we might be adding. We can do this - * safely here because we know the data fork is completely empty and - * this saves us from needing to run a separate transaction to set the - * fork offset in the immediate future. - */ - if (init_xattrs && xfs_has_attr(mp)) { - ip->i_forkoff = xfs_default_attroffset(ip) >> 3; - xfs_ifork_init_attr(ip, XFS_DINODE_FMT_EXTENTS, 0); - } - - /* - * Log the new values stuffed into the inode. - */ - xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); - xfs_trans_log_inode(tp, ip, flags); + xfs_trans_ijoin(tp, ip, 0); + xfs_inode_init(tp, args, ip); /* now that we have an i_mode we can setup the inode structure */ xfs_setup_inode(ip); @@ -885,146 +600,60 @@ xfs_init_new_inode( return 0; } -/* - * Decrement the link count on an inode & log the change. If this causes the - * link count to go to zero, move the inode to AGI unlinked list so that it can - * be freed when the last active reference goes away via xfs_inactive(). - */ -static int /* error */ -xfs_droplink( - xfs_trans_t *tp, - xfs_inode_t *ip) -{ - if (VFS_I(ip)->i_nlink == 0) { - xfs_alert(ip->i_mount, - "%s: Attempt to drop inode (%llu) with nlink zero.", - __func__, ip->i_ino); - return -EFSCORRUPTED; - } - - xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_CHG); - - drop_nlink(VFS_I(ip)); - xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); - - if (VFS_I(ip)->i_nlink) - return 0; - - return xfs_iunlink(tp, ip); -} - -/* - * Increment the link count on an inode & log the change. - */ -static void -xfs_bumplink( - xfs_trans_t *tp, - xfs_inode_t *ip) -{ - xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_CHG); - - inc_nlink(VFS_I(ip)); - xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); -} - -#ifdef CONFIG_XFS_LIVE_HOOKS -/* - * Use a static key here to reduce the overhead of directory live update hooks. - * If the compiler supports jump labels, the static branch will be replaced by - * a nop sled when there are no hook users. Online fsck is currently the only - * caller, so this is a reasonable tradeoff. - * - * Note: Patching the kernel code requires taking the cpu hotplug lock. Other - * parts of the kernel allocate memory with that lock held, which means that - * XFS callers cannot hold any locks that might be used by memory reclaim or - * writeback when calling the static_branch_{inc,dec} functions. - */ -DEFINE_STATIC_XFS_HOOK_SWITCH(xfs_dir_hooks_switch); - -void -xfs_dir_hook_disable(void) -{ - xfs_hooks_switch_off(&xfs_dir_hooks_switch); -} - -void -xfs_dir_hook_enable(void) -{ - xfs_hooks_switch_on(&xfs_dir_hooks_switch); -} - -/* Call hooks for a directory update relating to a child dirent update. */ -inline void -xfs_dir_update_hook( - struct xfs_inode *dp, - struct xfs_inode *ip, - int delta, - const struct xfs_name *name) -{ - if (xfs_hooks_switched_on(&xfs_dir_hooks_switch)) { - struct xfs_dir_update_params p = { - .dp = dp, - .ip = ip, - .delta = delta, - .name = name, - }; - struct xfs_mount *mp = ip->i_mount; - - xfs_hooks_call(&mp->m_dir_update_hooks, 0, &p); - } -} - -/* Call the specified function during a directory update. */ +/* Return dquots for the ids that will be assigned to a new file. */ int -xfs_dir_hook_add( - struct xfs_mount *mp, - struct xfs_dir_hook *hook) -{ - return xfs_hooks_add(&mp->m_dir_update_hooks, &hook->dirent_hook); -} +xfs_icreate_dqalloc( + const struct xfs_icreate_args *args, + struct xfs_dquot **udqpp, + struct xfs_dquot **gdqpp, + struct xfs_dquot **pdqpp) +{ + struct inode *dir = VFS_I(args->pip); + kuid_t uid = GLOBAL_ROOT_UID; + kgid_t gid = GLOBAL_ROOT_GID; + prid_t prid = 0; + unsigned int flags = XFS_QMOPT_QUOTALL; + + if (args->idmap) { + /* + * The uid/gid computation code must match what the VFS uses to + * assign i_[ug]id. INHERIT adjusts the gid computation for + * setgid/grpid systems. + */ + uid = mapped_fsuid(args->idmap, i_user_ns(dir)); + gid = mapped_fsgid(args->idmap, i_user_ns(dir)); + prid = xfs_get_initial_prid(args->pip); + flags |= XFS_QMOPT_INHERIT; + } -/* Stop calling the specified function during a directory update. */ -void -xfs_dir_hook_del( - struct xfs_mount *mp, - struct xfs_dir_hook *hook) -{ - xfs_hooks_del(&mp->m_dir_update_hooks, &hook->dirent_hook); -} + *udqpp = *gdqpp = *pdqpp = NULL; -/* Configure directory update hook functions. */ -void -xfs_dir_hook_setup( - struct xfs_dir_hook *hook, - notifier_fn_t mod_fn) -{ - xfs_hook_setup(&hook->dirent_hook, mod_fn); + return xfs_qm_vop_dqalloc(args->pip, uid, gid, prid, flags, udqpp, + gdqpp, pdqpp); } -#endif /* CONFIG_XFS_LIVE_HOOKS */ int xfs_create( - struct mnt_idmap *idmap, - xfs_inode_t *dp, + const struct xfs_icreate_args *args, struct xfs_name *name, - umode_t mode, - dev_t rdev, - bool init_xattrs, - xfs_inode_t **ipp) + struct xfs_inode **ipp) { - int is_dir = S_ISDIR(mode); + struct xfs_inode *dp = args->pip; + struct xfs_dir_update du = { + .dp = dp, + .name = name, + }; struct xfs_mount *mp = dp->i_mount; - struct xfs_inode *ip = NULL; struct xfs_trans *tp = NULL; - int error; - bool unlock_dp_on_error = false; - prid_t prid; - struct xfs_dquot *udqp = NULL; - struct xfs_dquot *gdqp = NULL; - struct xfs_dquot *pdqp = NULL; + struct xfs_dquot *udqp; + struct xfs_dquot *gdqp; + struct xfs_dquot *pdqp; struct xfs_trans_res *tres; - uint resblks; xfs_ino_t ino; + bool unlock_dp_on_error = false; + bool is_dir = S_ISDIR(args->mode); + uint resblks; + int error; trace_xfs_create(dp, name); @@ -1033,26 +662,23 @@ xfs_create( if (xfs_ifork_zapped(dp, XFS_DATA_FORK)) return -EIO; - prid = xfs_get_initial_prid(dp); - - /* - * Make sure that we have allocated dquot(s) on disk. - */ - error = xfs_qm_vop_dqalloc(dp, mapped_fsuid(idmap, &init_user_ns), - mapped_fsgid(idmap, &init_user_ns), prid, - XFS_QMOPT_QUOTALL | XFS_QMOPT_INHERIT, - &udqp, &gdqp, &pdqp); + /* Make sure that we have allocated dquot(s) on disk. */ + error = xfs_icreate_dqalloc(args, &udqp, &gdqp, &pdqp); if (error) return error; if (is_dir) { - resblks = XFS_MKDIR_SPACE_RES(mp, name->len); + resblks = xfs_mkdir_space_res(mp, name->len); tres = &M_RES(mp)->tr_mkdir; } else { - resblks = XFS_CREATE_SPACE_RES(mp, name->len); + resblks = xfs_create_space_res(mp, name->len); tres = &M_RES(mp)->tr_create; } + error = xfs_parent_start(mp, &du.ppargs); + if (error) + goto out_release_dquots; + /* * Initially assume that the file does not exist and * reserve the resources for that case. If that is not @@ -1068,7 +694,7 @@ xfs_create( resblks, &tp); } if (error) - goto out_release_dquots; + goto out_parent; xfs_ilock(dp, XFS_ILOCK_EXCL | XFS_ILOCK_PARENT); unlock_dp_on_error = true; @@ -1078,10 +704,9 @@ xfs_create( * entry pointing to them, but a directory also the "." entry * pointing to itself. */ - error = xfs_dialloc(&tp, dp->i_ino, mode, &ino); + error = xfs_dialloc(&tp, dp->i_ino, args->mode, &ino); if (!error) - error = xfs_init_new_inode(idmap, tp, dp, ino, mode, - is_dir ? 2 : 1, rdev, prid, init_xattrs, &ip); + error = xfs_icreate(tp, ino, args, &du.ip); if (error) goto out_trans_cancel; @@ -1092,31 +717,11 @@ xfs_create( * the transaction cancel unlocking dp so don't do it explicitly in the * error path. */ - xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL); - unlock_dp_on_error = false; + xfs_trans_ijoin(tp, dp, 0); - error = xfs_dir_createname(tp, dp, name, ip->i_ino, - resblks - XFS_IALLOC_SPACE_RES(mp)); - if (error) { - ASSERT(error != -ENOSPC); + error = xfs_dir_create_child(tp, resblks, &du); + if (error) goto out_trans_cancel; - } - xfs_trans_ichgtime(tp, dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); - xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE); - - if (is_dir) { - error = xfs_dir_init(tp, ip, dp); - if (error) - goto out_trans_cancel; - - xfs_bumplink(tp, dp); - } - - /* - * Create ip with a reference from dp, and add '.' and '..' references - * if it's a directory. - */ - xfs_dir_update_hook(dp, ip, 1, name); /* * If this is a synchronous mount, make sure that the @@ -1131,7 +736,7 @@ xfs_create( * 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); + xfs_qm_vop_create_dqattach(tp, du.ip, udqp, gdqp, pdqp); error = xfs_trans_commit(tp); if (error) @@ -1141,7 +746,10 @@ xfs_create( xfs_qm_dqrele(gdqp); xfs_qm_dqrele(pdqp); - *ipp = ip; + *ipp = du.ip; + xfs_iunlock(du.ip, XFS_ILOCK_EXCL); + xfs_iunlock(dp, XFS_ILOCK_EXCL); + xfs_parent_finish(mp, du.ppargs); return 0; out_trans_cancel: @@ -1152,10 +760,13 @@ xfs_create( * setup of the inode and release the inode. This prevents recursive * transactions and deadlocks from xfs_inactive. */ - if (ip) { - xfs_finish_inode_setup(ip); - xfs_irele(ip); + if (du.ip) { + xfs_iunlock(du.ip, XFS_ILOCK_EXCL); + xfs_finish_inode_setup(du.ip); + xfs_irele(du.ip); } + out_parent: + xfs_parent_finish(mp, du.ppargs); out_release_dquots: xfs_qm_dqrele(udqp); xfs_qm_dqrele(gdqp); @@ -1168,35 +779,28 @@ xfs_create( int xfs_create_tmpfile( - struct mnt_idmap *idmap, - struct xfs_inode *dp, - umode_t mode, + const struct xfs_icreate_args *args, struct xfs_inode **ipp) { + struct xfs_inode *dp = args->pip; struct xfs_mount *mp = dp->i_mount; struct xfs_inode *ip = NULL; struct xfs_trans *tp = NULL; - int error; - prid_t prid; - struct xfs_dquot *udqp = NULL; - struct xfs_dquot *gdqp = NULL; - struct xfs_dquot *pdqp = NULL; + struct xfs_dquot *udqp; + struct xfs_dquot *gdqp; + struct xfs_dquot *pdqp; struct xfs_trans_res *tres; - uint resblks; xfs_ino_t ino; + uint resblks; + int error; + + ASSERT(args->flags & XFS_ICREATE_TMPFILE); if (xfs_is_shutdown(mp)) return -EIO; - prid = xfs_get_initial_prid(dp); - - /* - * Make sure that we have allocated dquot(s) on disk. - */ - error = xfs_qm_vop_dqalloc(dp, mapped_fsuid(idmap, &init_user_ns), - mapped_fsgid(idmap, &init_user_ns), prid, - XFS_QMOPT_QUOTALL | XFS_QMOPT_INHERIT, - &udqp, &gdqp, &pdqp); + /* Make sure that we have allocated dquot(s) on disk. */ + error = xfs_icreate_dqalloc(args, &udqp, &gdqp, &pdqp); if (error) return error; @@ -1208,10 +812,9 @@ xfs_create_tmpfile( if (error) goto out_release_dquots; - error = xfs_dialloc(&tp, dp->i_ino, mode, &ino); + error = xfs_dialloc(&tp, dp->i_ino, args->mode, &ino); if (!error) - error = xfs_init_new_inode(idmap, tp, dp, ino, mode, - 0, 0, prid, false, &ip); + error = xfs_icreate(tp, ino, args, &ip); if (error) goto out_trans_cancel; @@ -1238,6 +841,7 @@ xfs_create_tmpfile( xfs_qm_dqrele(pdqp); *ipp = ip; + xfs_iunlock(ip, XFS_ILOCK_EXCL); return 0; out_trans_cancel: @@ -1249,6 +853,7 @@ xfs_create_tmpfile( * transactions and deadlocks from xfs_inactive. */ if (ip) { + xfs_iunlock(ip, XFS_ILOCK_EXCL); xfs_finish_inode_setup(ip); xfs_irele(ip); } @@ -1262,12 +867,17 @@ xfs_create_tmpfile( int xfs_link( - xfs_inode_t *tdp, - xfs_inode_t *sip, + struct xfs_inode *tdp, + struct xfs_inode *sip, struct xfs_name *target_name) { - xfs_mount_t *mp = tdp->i_mount; - xfs_trans_t *tp; + struct xfs_dir_update du = { + .dp = tdp, + .name = target_name, + .ip = sip, + }; + struct xfs_mount *mp = tdp->i_mount; + struct xfs_trans *tp; int error, nospace_error = 0; int resblks; @@ -1288,11 +898,25 @@ xfs_link( if (error) goto std_return; - resblks = XFS_LINK_SPACE_RES(mp, target_name->len); + error = xfs_parent_start(mp, &du.ppargs); + if (error) + goto std_return; + + resblks = xfs_link_space_res(mp, target_name->len); error = xfs_trans_alloc_dir(tdp, &M_RES(mp)->tr_link, sip, &resblks, &tp, &nospace_error); if (error) - goto std_return; + goto out_parent; + + /* + * We don't allow reservationless or quotaless hardlinking when parent + * pointers are enabled because we can't back out if the xattrs must + * grow. + */ + if (du.ppargs && nospace_error) { + error = nospace_error; + goto error_return; + } /* * If we are using project inheritance, we only allow hard link @@ -1316,34 +940,9 @@ xfs_link( } } - if (!resblks) { - error = xfs_dir_canenter(tp, tdp, target_name); - if (error) - goto error_return; - } - - /* - * Handle initial link state of O_TMPFILE inode - */ - if (VFS_I(sip)->i_nlink == 0) { - struct xfs_perag *pag; - - pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, sip->i_ino)); - error = xfs_iunlink_remove(tp, pag, sip); - xfs_perag_put(pag); - if (error) - goto error_return; - } - - error = xfs_dir_createname(tp, tdp, target_name, sip->i_ino, - resblks); + error = xfs_dir_add_child(tp, resblks, &du); if (error) goto error_return; - xfs_trans_ichgtime(tp, tdp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); - xfs_trans_log_inode(tp, tdp, XFS_ILOG_CORE); - - xfs_bumplink(tp, sip); - xfs_dir_update_hook(tdp, sip, 1, target_name); /* * If this is a synchronous mount, make sure that the @@ -1353,10 +952,18 @@ xfs_link( if (xfs_has_wsync(mp) || xfs_has_dirsync(mp)) xfs_trans_set_sync(tp); - return xfs_trans_commit(tp); + error = xfs_trans_commit(tp); + xfs_iunlock(tdp, XFS_ILOCK_EXCL); + xfs_iunlock(sip, XFS_ILOCK_EXCL); + xfs_parent_finish(mp, du.ppargs); + return error; error_return: xfs_trans_cancel(tp); + xfs_iunlock(tdp, XFS_ILOCK_EXCL); + xfs_iunlock(sip, XFS_ILOCK_EXCL); + out_parent: + xfs_parent_finish(mp, du.ppargs); std_return: if (error == -ENOSPC && nospace_error) error = nospace_error; @@ -1522,7 +1129,7 @@ xfs_release( if (!xfs_ilock_nowait(ip, XFS_IOLOCK_EXCL)) return 0; - if (xfs_can_free_eofblocks(ip, false)) { + if (xfs_can_free_eofblocks(ip)) { /* * Check if the inode is being opened, written and closed * frequently and we have delayed allocation blocks outstanding @@ -1555,6 +1162,51 @@ out_unlock: } /* + * Mark all the buffers attached to this directory stale. In theory we should + * never be freeing a directory with any blocks at all, but this covers the + * case where we've recovered a directory swap with a "temporary" directory + * created by online repair and now need to dump it. + */ +STATIC void +xfs_inactive_dir( + struct xfs_inode *dp) +{ + struct xfs_iext_cursor icur; + struct xfs_bmbt_irec got; + struct xfs_mount *mp = dp->i_mount; + struct xfs_da_geometry *geo = mp->m_dir_geo; + struct xfs_ifork *ifp = xfs_ifork_ptr(dp, XFS_DATA_FORK); + xfs_fileoff_t off; + + /* + * Invalidate each directory block. All directory blocks are of + * fsbcount length and alignment, so we only need to walk those same + * offsets. We hold the only reference to this inode, so we must wait + * for the buffer locks. + */ + for_each_xfs_iext(ifp, &icur, &got) { + for (off = round_up(got.br_startoff, geo->fsbcount); + off < got.br_startoff + got.br_blockcount; + off += geo->fsbcount) { + struct xfs_buf *bp = NULL; + xfs_fsblock_t fsbno; + int error; + + fsbno = (off - got.br_startoff) + got.br_startblock; + error = xfs_buf_incore(mp->m_ddev_targp, + XFS_FSB_TO_DADDR(mp, fsbno), + XFS_FSB_TO_BB(mp, geo->fsbcount), + XBF_LIVESCAN, &bp); + if (error) + continue; + + xfs_buf_stale(bp); + xfs_buf_relse(bp); + } + } +} + +/* * xfs_inactive_truncate * * Called to perform a truncate when an inode becomes unlinked. @@ -1738,15 +1390,13 @@ xfs_inode_needs_inactive( /* * This file isn't being freed, so check if there are post-eof blocks - * to free. @force is true because we are evicting an inode from the - * cache. Post-eof blocks must be freed, lest we end up with broken - * free space accounting. + * to free. * * Note: don't bother with iolock here since lockdep complains about * acquiring it in reclaim context. We have the only reference to the * inode at this point anyways. */ - return xfs_can_free_eofblocks(ip, true); + return xfs_can_free_eofblocks(ip); } /* @@ -1829,15 +1479,11 @@ xfs_inactive( if (VFS_I(ip)->i_nlink != 0) { /* - * force is true because we are evicting an inode from the - * cache. Post-eof blocks must be freed, lest we end up with - * broken free space accounting. - * * Note: don't bother with iolock here since lockdep complains * about acquiring it in reclaim context. We have the only * reference to the inode at this point anyways. */ - if (xfs_can_free_eofblocks(ip, true)) + if (xfs_can_free_eofblocks(ip)) error = xfs_free_eofblocks(ip); goto out; @@ -1864,6 +1510,11 @@ xfs_inactive( goto out; } + if (S_ISDIR(VFS_I(ip)->i_mode) && ip->i_df.if_nextents > 0) { + xfs_inactive_dir(ip); + truncate = 1; + } + if (S_ISLNK(VFS_I(ip)->i_mode)) error = xfs_inactive_symlink(ip); else if (truncate) @@ -1899,45 +1550,12 @@ out: } /* - * In-Core Unlinked List Lookups - * ============================= - * - * Every inode is supposed to be reachable from some other piece of metadata - * with the exception of the root directory. Inodes with a connection to a - * file descriptor but not linked from anywhere in the on-disk directory tree - * are collectively known as unlinked inodes, though the filesystem itself - * maintains links to these inodes so that on-disk metadata are consistent. - * - * XFS implements a per-AG on-disk hash table of unlinked inodes. The AGI - * header contains a number of buckets that point to an inode, and each inode - * record has a pointer to the next inode in the hash chain. This - * singly-linked list causes scaling problems in the iunlink remove function - * because we must walk that list to find the inode that points to the inode - * being removed from the unlinked hash bucket list. - * - * Hence we keep an in-memory double linked list to link each inode on an - * unlinked list. Because there are 64 unlinked lists per AGI, keeping pointer - * based lists would require having 64 list heads in the perag, one for each - * list. This is expensive in terms of memory (think millions of AGs) and cache - * misses on lookups. Instead, use the fact that inodes on the unlinked list - * must be referenced at the VFS level to keep them on the list and hence we - * have an existence guarantee for inodes on the unlinked list. - * - * Given we have an existence guarantee, we can use lockless inode cache lookups - * to resolve aginos to xfs inodes. This means we only need 8 bytes per inode - * for the double linked unlinked list, and we don't need any extra locking to - * keep the list safe as all manipulations are done under the AGI buffer lock. - * Keeping the list up to date does not require memory allocation, just finding - * the XFS inode and updating the next/prev unlinked list aginos. - */ - -/* * Find an inode on the unlinked list. This does not take references to the * inode as we have existence guarantees by holding the AGI buffer lock and that * only unlinked, referenced inodes can be on the unlinked inode list. If we * don't find the inode in cache, then let the caller handle the situation. */ -static struct xfs_inode * +struct xfs_inode * xfs_iunlink_lookup( struct xfs_perag *pag, xfs_agino_t agino) @@ -1966,75 +1584,11 @@ xfs_iunlink_lookup( } /* - * Update the prev pointer of the next agino. Returns -ENOLINK if the inode - * is not in cache. - */ -static int -xfs_iunlink_update_backref( - struct xfs_perag *pag, - xfs_agino_t prev_agino, - xfs_agino_t next_agino) -{ - struct xfs_inode *ip; - - /* No update necessary if we are at the end of the list. */ - if (next_agino == NULLAGINO) - return 0; - - ip = xfs_iunlink_lookup(pag, next_agino); - if (!ip) - return -ENOLINK; - - ip->i_prev_unlinked = prev_agino; - return 0; -} - -/* - * Point the AGI unlinked bucket at an inode and log the results. The caller - * is responsible for validating the old value. - */ -STATIC int -xfs_iunlink_update_bucket( - struct xfs_trans *tp, - struct xfs_perag *pag, - struct xfs_buf *agibp, - unsigned int bucket_index, - xfs_agino_t new_agino) -{ - struct xfs_agi *agi = agibp->b_addr; - xfs_agino_t old_value; - int offset; - - ASSERT(xfs_verify_agino_or_null(pag, new_agino)); - - old_value = be32_to_cpu(agi->agi_unlinked[bucket_index]); - trace_xfs_iunlink_update_bucket(tp->t_mountp, pag->pag_agno, bucket_index, - old_value, new_agino); - - /* - * We should never find the head of the list already set to the value - * passed in because either we're adding or removing ourselves from the - * head of the list. - */ - if (old_value == new_agino) { - xfs_buf_mark_corrupt(agibp); - xfs_ag_mark_sick(pag, XFS_SICK_AG_AGI); - return -EFSCORRUPTED; - } - - agi->agi_unlinked[bucket_index] = cpu_to_be32(new_agino); - offset = offsetof(struct xfs_agi, agi_unlinked) + - (sizeof(xfs_agino_t) * bucket_index); - xfs_trans_log_buf(tp, agibp, offset, offset + sizeof(xfs_agino_t) - 1); - return 0; -} - -/* * Load the inode @next_agino into the cache and set its prev_unlinked pointer * to @prev_agino. Caller must hold the AGI to synchronize with other changes * to the unlinked list. */ -STATIC int +int xfs_iunlink_reload_next( struct xfs_trans *tp, struct xfs_buf *agibp, @@ -2090,187 +1644,6 @@ rele: return error; } -static int -xfs_iunlink_insert_inode( - struct xfs_trans *tp, - struct xfs_perag *pag, - struct xfs_buf *agibp, - struct xfs_inode *ip) -{ - struct xfs_mount *mp = tp->t_mountp; - struct xfs_agi *agi = agibp->b_addr; - xfs_agino_t next_agino; - xfs_agino_t agino = XFS_INO_TO_AGINO(mp, ip->i_ino); - short bucket_index = agino % XFS_AGI_UNLINKED_BUCKETS; - int error; - - /* - * Get the index into the agi hash table for the list this inode will - * go on. Make sure the pointer isn't garbage and that this inode - * isn't already on the list. - */ - next_agino = be32_to_cpu(agi->agi_unlinked[bucket_index]); - if (next_agino == agino || - !xfs_verify_agino_or_null(pag, next_agino)) { - xfs_buf_mark_corrupt(agibp); - xfs_ag_mark_sick(pag, XFS_SICK_AG_AGI); - return -EFSCORRUPTED; - } - - /* - * Update the prev pointer in the next inode to point back to this - * inode. - */ - error = xfs_iunlink_update_backref(pag, agino, next_agino); - if (error == -ENOLINK) - error = xfs_iunlink_reload_next(tp, agibp, agino, next_agino); - if (error) - return error; - - if (next_agino != NULLAGINO) { - /* - * There is already another inode in the bucket, so point this - * inode to the current head of the list. - */ - error = xfs_iunlink_log_inode(tp, ip, pag, next_agino); - if (error) - return error; - ip->i_next_unlinked = next_agino; - } - - /* Point the head of the list to point to this inode. */ - ip->i_prev_unlinked = NULLAGINO; - return xfs_iunlink_update_bucket(tp, pag, agibp, bucket_index, agino); -} - -/* - * This is called when the inode's link count has gone to 0 or we are creating - * a tmpfile via O_TMPFILE. The inode @ip must have nlink == 0. - * - * We place the on-disk inode on a list in the AGI. It will be pulled from this - * list when the inode is freed. - */ -STATIC int -xfs_iunlink( - struct xfs_trans *tp, - struct xfs_inode *ip) -{ - struct xfs_mount *mp = tp->t_mountp; - struct xfs_perag *pag; - struct xfs_buf *agibp; - int error; - - ASSERT(VFS_I(ip)->i_nlink == 0); - ASSERT(VFS_I(ip)->i_mode != 0); - trace_xfs_iunlink(ip); - - pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino)); - - /* Get the agi buffer first. It ensures lock ordering on the list. */ - error = xfs_read_agi(pag, tp, &agibp); - if (error) - goto out; - - error = xfs_iunlink_insert_inode(tp, pag, agibp, ip); -out: - xfs_perag_put(pag); - return error; -} - -static int -xfs_iunlink_remove_inode( - struct xfs_trans *tp, - struct xfs_perag *pag, - struct xfs_buf *agibp, - struct xfs_inode *ip) -{ - struct xfs_mount *mp = tp->t_mountp; - struct xfs_agi *agi = agibp->b_addr; - xfs_agino_t agino = XFS_INO_TO_AGINO(mp, ip->i_ino); - xfs_agino_t head_agino; - short bucket_index = agino % XFS_AGI_UNLINKED_BUCKETS; - int error; - - trace_xfs_iunlink_remove(ip); - - /* - * Get the index into the agi hash table for the list this inode will - * go on. Make sure the head pointer isn't garbage. - */ - head_agino = be32_to_cpu(agi->agi_unlinked[bucket_index]); - if (!xfs_verify_agino(pag, head_agino)) { - XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, - agi, sizeof(*agi)); - xfs_ag_mark_sick(pag, XFS_SICK_AG_AGI); - return -EFSCORRUPTED; - } - - /* - * Set our inode's next_unlinked pointer to NULL and then return - * the old pointer value so that we can update whatever was previous - * to us in the list to point to whatever was next in the list. - */ - error = xfs_iunlink_log_inode(tp, ip, pag, NULLAGINO); - if (error) - return error; - - /* - * Update the prev pointer in the next inode to point back to previous - * inode in the chain. - */ - error = xfs_iunlink_update_backref(pag, ip->i_prev_unlinked, - ip->i_next_unlinked); - if (error == -ENOLINK) - error = xfs_iunlink_reload_next(tp, agibp, ip->i_prev_unlinked, - ip->i_next_unlinked); - if (error) - return error; - - if (head_agino != agino) { - struct xfs_inode *prev_ip; - - prev_ip = xfs_iunlink_lookup(pag, ip->i_prev_unlinked); - if (!prev_ip) { - xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE); - return -EFSCORRUPTED; - } - - error = xfs_iunlink_log_inode(tp, prev_ip, pag, - ip->i_next_unlinked); - prev_ip->i_next_unlinked = ip->i_next_unlinked; - } else { - /* Point the head of the list to the next unlinked inode. */ - error = xfs_iunlink_update_bucket(tp, pag, agibp, bucket_index, - ip->i_next_unlinked); - } - - ip->i_next_unlinked = NULLAGINO; - ip->i_prev_unlinked = 0; - return error; -} - -/* - * Pull the on-disk inode from the AGI unlinked list. - */ -STATIC int -xfs_iunlink_remove( - struct xfs_trans *tp, - struct xfs_perag *pag, - struct xfs_inode *ip) -{ - struct xfs_buf *agibp; - int error; - - trace_xfs_iunlink_remove(ip); - - /* Get the agi buffer first. It ensures lock ordering on the list. */ - error = xfs_read_agi(pag, tp, &agibp); - if (error) - return error; - - return xfs_iunlink_remove_inode(tp, pag, agibp, ip); -} - /* * Look up the inode number specified and if it is not already marked XFS_ISTALE * mark it stale. We should only find clean inodes in this lookup that aren't @@ -2425,11 +1798,26 @@ xfs_ifree_cluster( * This buffer may not have been correctly initialised as we * didn't read it from disk. That's not important because we are * only using to mark the buffer as stale in the log, and to - * attach stale cached inodes on it. That means it will never be - * dispatched for IO. If it is, we want to know about it, and we - * want it to fail. We can acheive this by adding a write - * verifier to the buffer. + * attach stale cached inodes on it. + * + * For the inode that triggered the cluster freeing, this + * attachment may occur in xfs_inode_item_precommit() after we + * have marked this buffer stale. If this buffer was not in + * memory before xfs_ifree_cluster() started, it will not be + * marked XBF_DONE and this will cause problems later in + * xfs_inode_item_precommit() when we trip over a (stale, !done) + * buffer to attached to the transaction. + * + * Hence we have to mark the buffer as XFS_DONE here. This is + * safe because we are also marking the buffer as XBF_STALE and + * XFS_BLI_STALE. That means it will never be dispatched for + * IO and it won't be unlocked until the cluster freeing has + * been committed to the journal and the buffer unpinned. If it + * is written, we want to know about it, and we want it to + * fail. We can acheive this by adding a write verifier to the + * buffer. */ + bp->b_flags |= XBF_DONE; bp->b_ops = &xfs_inode_buf_ops; /* @@ -2474,36 +1862,10 @@ xfs_ifree( pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino)); - /* - * Free the inode first so that we guarantee that the AGI lock is going - * to be taken before we remove the inode from the unlinked list. This - * makes the AGI lock -> unlinked list modification order the same as - * used in O_TMPFILE creation. - */ - error = xfs_difree(tp, pag, ip->i_ino, &xic); - if (error) - goto out; - - error = xfs_iunlink_remove(tp, pag, ip); + error = xfs_inode_uninit(tp, pag, ip, &xic); if (error) goto out; - /* - * Free any local-format data sitting around before we reset the - * data fork to extents format. Note that the attr fork data has - * already been freed by xfs_attr_inactive. - */ - if (ip->i_df.if_format == XFS_DINODE_FMT_LOCAL) { - kfree(ip->i_df.if_data); - ip->i_df.if_data = NULL; - ip->i_df.if_bytes = 0; - } - - VFS_I(ip)->i_mode = 0; /* mark incore inode as free */ - ip->i_diflags = 0; - ip->i_diflags2 = mp->m_ino_geo.new_diflags2; - ip->i_forkoff = 0; /* mark the attr fork not in use */ - ip->i_df.if_format = XFS_DINODE_FMT_EXTENTS; if (xfs_iflags_test(ip, XFS_IPRESERVE_DM_FIELDS)) xfs_iflags_clear(ip, XFS_IPRESERVE_DM_FIELDS); @@ -2512,13 +1874,6 @@ xfs_ifree( iip->ili_fields &= ~(XFS_ILOG_AOWNER | XFS_ILOG_DOWNER); spin_unlock(&iip->ili_lock); - /* - * Bump the generation count so no one will be confused - * by reincarnations of this inode. - */ - VFS_I(ip)->i_generation++; - xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); - if (xic.deleted) error = xfs_ifree_cluster(tp, pag, ip, &xic); out: @@ -2598,12 +1953,17 @@ xfs_iunpin_wait( */ int xfs_remove( - xfs_inode_t *dp, + struct xfs_inode *dp, struct xfs_name *name, - xfs_inode_t *ip) + struct xfs_inode *ip) { - xfs_mount_t *mp = dp->i_mount; - xfs_trans_t *tp = NULL; + struct xfs_dir_update du = { + .dp = dp, + .name = name, + .ip = ip, + }; + struct xfs_mount *mp = dp->i_mount; + struct xfs_trans *tp = NULL; int is_dir = S_ISDIR(VFS_I(ip)->i_mode); int dontcare; int error = 0; @@ -2624,6 +1984,10 @@ xfs_remove( if (error) goto std_return; + error = xfs_parent_start(mp, &du.ppargs); + if (error) + goto std_return; + /* * We try to get the real space reservation first, allowing for * directory btree deletion(s) implying possible bmap insert(s). If we @@ -2635,77 +1999,18 @@ xfs_remove( * the directory code can handle a reservationless update and we don't * want to prevent a user from trying to free space by deleting things. */ - resblks = XFS_REMOVE_SPACE_RES(mp); + resblks = xfs_remove_space_res(mp, name->len); error = xfs_trans_alloc_dir(dp, &M_RES(mp)->tr_remove, ip, &resblks, &tp, &dontcare); if (error) { ASSERT(error != -ENOSPC); - goto std_return; + goto out_parent; } - /* - * If we're removing a directory perform some additional validation. - */ - if (is_dir) { - ASSERT(VFS_I(ip)->i_nlink >= 2); - if (VFS_I(ip)->i_nlink != 2) { - error = -ENOTEMPTY; - goto out_trans_cancel; - } - if (!xfs_dir_isempty(ip)) { - error = -ENOTEMPTY; - goto out_trans_cancel; - } - - /* Drop the link from ip's "..". */ - error = xfs_droplink(tp, dp); - if (error) - goto out_trans_cancel; - - /* Drop the "." link from ip to self. */ - error = xfs_droplink(tp, ip); - if (error) - goto out_trans_cancel; - - /* - * Point the unlinked child directory's ".." entry to the root - * directory to eliminate back-references to inodes that may - * get freed before the child directory is closed. If the fs - * gets shrunk, this can lead to dirent inode validation errors. - */ - if (dp->i_ino != tp->t_mountp->m_sb.sb_rootino) { - error = xfs_dir_replace(tp, ip, &xfs_name_dotdot, - tp->t_mountp->m_sb.sb_rootino, 0); - if (error) - goto out_trans_cancel; - } - } else { - /* - * When removing a non-directory we need to log the parent - * inode here. For a directory this is done implicitly - * by the xfs_droplink call for the ".." entry. - */ - xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE); - } - xfs_trans_ichgtime(tp, dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); - - /* Drop the link from dp to ip. */ - error = xfs_droplink(tp, ip); + error = xfs_dir_remove_child(tp, resblks, &du); if (error) goto out_trans_cancel; - error = xfs_dir_removename(tp, dp, name, ip->i_ino, resblks); - if (error) { - ASSERT(error != -ENOENT); - goto out_trans_cancel; - } - - /* - * Drop the link from dp to ip, and if ip was a directory, remove the - * '.' and '..' references since we freed the directory. - */ - xfs_dir_update_hook(dp, ip, -1, name); - /* * If this is a synchronous mount, make sure that the * remove transaction goes to disk before returning to @@ -2716,19 +2021,42 @@ xfs_remove( error = xfs_trans_commit(tp); if (error) - goto std_return; + goto out_unlock; if (is_dir && xfs_inode_is_filestream(ip)) xfs_filestream_deassociate(ip); + xfs_iunlock(ip, XFS_ILOCK_EXCL); + xfs_iunlock(dp, XFS_ILOCK_EXCL); + xfs_parent_finish(mp, du.ppargs); return 0; out_trans_cancel: xfs_trans_cancel(tp); + out_unlock: + xfs_iunlock(ip, XFS_ILOCK_EXCL); + xfs_iunlock(dp, XFS_ILOCK_EXCL); + out_parent: + xfs_parent_finish(mp, du.ppargs); std_return: return error; } +static inline void +xfs_iunlock_rename( + struct xfs_inode **i_tab, + int num_inodes) +{ + int i; + + for (i = num_inodes - 1; i >= 0; i--) { + /* Skip duplicate inodes if src and target dps are the same */ + if (!i_tab[i] || (i > 0 && i_tab[i] == i_tab[i - 1])) + continue; + xfs_iunlock(i_tab[i], XFS_ILOCK_EXCL); + } +} + /* * Enter all inodes for a rename transaction into a sorted array. */ @@ -2743,7 +2071,7 @@ xfs_sort_for_rename( struct xfs_inode **i_tab,/* out: sorted array of inodes */ int *num_inodes) /* in/out: inodes in array */ { - int i, j; + int i; ASSERT(*num_inodes == __XFS_SORT_INODES); memset(i_tab, 0, *num_inodes * sizeof(struct xfs_inode *)); @@ -2765,156 +2093,28 @@ xfs_sort_for_rename( i_tab[i++] = wip; *num_inodes = i; - /* - * Sort the elements via bubble sort. (Remember, there are at - * most 5 elements to sort, so this is adequate.) - */ - for (i = 0; i < *num_inodes; i++) { - for (j = 1; j < *num_inodes; j++) { - if (i_tab[j]->i_ino < i_tab[j-1]->i_ino) { - struct xfs_inode *temp = i_tab[j]; - i_tab[j] = i_tab[j-1]; - i_tab[j-1] = temp; - } - } - } + xfs_sort_inodes(i_tab, *num_inodes); } -static int -xfs_finish_rename( - struct xfs_trans *tp) -{ - /* - * If this is a synchronous mount, make sure that the rename transaction - * goes to disk before returning to the user. - */ - if (xfs_has_wsync(tp->t_mountp) || xfs_has_dirsync(tp->t_mountp)) - xfs_trans_set_sync(tp); - - return xfs_trans_commit(tp); -} - -/* - * xfs_cross_rename() - * - * responsible for handling RENAME_EXCHANGE flag in renameat2() syscall - */ -STATIC int -xfs_cross_rename( - struct xfs_trans *tp, - struct xfs_inode *dp1, - struct xfs_name *name1, - struct xfs_inode *ip1, - struct xfs_inode *dp2, - struct xfs_name *name2, - struct xfs_inode *ip2, - int spaceres) +void +xfs_sort_inodes( + struct xfs_inode **i_tab, + unsigned int num_inodes) { - int error = 0; - int ip1_flags = 0; - int ip2_flags = 0; - int dp2_flags = 0; - - /* Swap inode number for dirent in first parent */ - error = xfs_dir_replace(tp, dp1, name1, ip2->i_ino, spaceres); - if (error) - goto out_trans_abort; + int i, j; - /* Swap inode number for dirent in second parent */ - error = xfs_dir_replace(tp, dp2, name2, ip1->i_ino, spaceres); - if (error) - goto out_trans_abort; + ASSERT(num_inodes <= __XFS_SORT_INODES); /* - * If we're renaming one or more directories across different parents, - * update the respective ".." entries (and link counts) to match the new - * parents. + * Sort the elements via bubble sort. (Remember, there are at + * most 5 elements to sort, so this is adequate.) */ - if (dp1 != dp2) { - dp2_flags = XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG; - - if (S_ISDIR(VFS_I(ip2)->i_mode)) { - error = xfs_dir_replace(tp, ip2, &xfs_name_dotdot, - dp1->i_ino, spaceres); - if (error) - goto out_trans_abort; - - /* transfer ip2 ".." reference to dp1 */ - if (!S_ISDIR(VFS_I(ip1)->i_mode)) { - error = xfs_droplink(tp, dp2); - if (error) - goto out_trans_abort; - xfs_bumplink(tp, dp1); - } - - /* - * Although ip1 isn't changed here, userspace needs - * to be warned about the change, so that applications - * relying on it (like backup ones), will properly - * notify the change - */ - ip1_flags |= XFS_ICHGTIME_CHG; - ip2_flags |= XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG; + for (i = 0; i < num_inodes; i++) { + for (j = 1; j < num_inodes; j++) { + if (i_tab[j]->i_ino < i_tab[j-1]->i_ino) + swap(i_tab[j], i_tab[j - 1]); } - - if (S_ISDIR(VFS_I(ip1)->i_mode)) { - error = xfs_dir_replace(tp, ip1, &xfs_name_dotdot, - dp2->i_ino, spaceres); - if (error) - goto out_trans_abort; - - /* transfer ip1 ".." reference to dp2 */ - if (!S_ISDIR(VFS_I(ip2)->i_mode)) { - error = xfs_droplink(tp, dp1); - if (error) - goto out_trans_abort; - xfs_bumplink(tp, dp2); - } - - /* - * Although ip2 isn't changed here, userspace needs - * to be warned about the change, so that applications - * relying on it (like backup ones), will properly - * notify the change - */ - ip1_flags |= XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG; - ip2_flags |= XFS_ICHGTIME_CHG; - } - } - - if (ip1_flags) { - xfs_trans_ichgtime(tp, ip1, ip1_flags); - xfs_trans_log_inode(tp, ip1, XFS_ILOG_CORE); - } - if (ip2_flags) { - xfs_trans_ichgtime(tp, ip2, ip2_flags); - xfs_trans_log_inode(tp, ip2, XFS_ILOG_CORE); } - if (dp2_flags) { - xfs_trans_ichgtime(tp, dp2, dp2_flags); - xfs_trans_log_inode(tp, dp2, XFS_ILOG_CORE); - } - xfs_trans_ichgtime(tp, dp1, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); - xfs_trans_log_inode(tp, dp1, XFS_ILOG_CORE); - - /* - * Inform our hook clients that we've finished an exchange operation as - * follows: removed the source and target files from their directories; - * added the target to the source directory; and added the source to - * the target directory. All inodes are locked, so it's ok to model a - * rename this way so long as we say we deleted entries before we add - * new ones. - */ - xfs_dir_update_hook(dp1, ip1, -1, name1); - xfs_dir_update_hook(dp2, ip2, -1, name2); - xfs_dir_update_hook(dp1, ip2, 1, name1); - xfs_dir_update_hook(dp2, ip1, 1, name2); - - return xfs_finish_rename(tp); - -out_trans_abort: - xfs_trans_cancel(tp); - return error; } /* @@ -2932,12 +2132,17 @@ xfs_rename_alloc_whiteout( struct xfs_inode *dp, struct xfs_inode **wip) { + struct xfs_icreate_args args = { + .idmap = idmap, + .pip = dp, + .mode = S_IFCHR | WHITEOUT_MODE, + .flags = XFS_ICREATE_TMPFILE, + }; struct xfs_inode *tmpfile; struct qstr name; int error; - error = xfs_create_tmpfile(idmap, dp, S_IFCHR | WHITEOUT_MODE, - &tmpfile); + error = xfs_create_tmpfile(&args, &tmpfile); if (error) return error; @@ -2977,9 +2182,19 @@ xfs_rename( struct xfs_inode *target_ip, unsigned int flags) { + struct xfs_dir_update du_src = { + .dp = src_dp, + .name = src_name, + .ip = src_ip, + }; + struct xfs_dir_update du_tgt = { + .dp = target_dp, + .name = target_name, + .ip = target_ip, + }; + struct xfs_dir_update du_wip = { }; struct xfs_mount *mp = src_dp->i_mount; struct xfs_trans *tp; - struct xfs_inode *wip = NULL; /* whiteout inode */ struct xfs_inode *inodes[__XFS_SORT_INODES]; int i; int num_inodes = __XFS_SORT_INODES; @@ -3000,8 +2215,8 @@ xfs_rename( * appropriately. */ if (flags & RENAME_WHITEOUT) { - error = xfs_rename_alloc_whiteout(idmap, src_name, - target_dp, &wip); + error = xfs_rename_alloc_whiteout(idmap, src_name, target_dp, + &du_wip.ip); if (error) return error; @@ -3009,12 +2224,29 @@ xfs_rename( src_name->type = XFS_DIR3_FT_CHRDEV; } - xfs_sort_for_rename(src_dp, target_dp, src_ip, target_ip, wip, - inodes, &num_inodes); + xfs_sort_for_rename(src_dp, target_dp, src_ip, target_ip, du_wip.ip, + inodes, &num_inodes); + + error = xfs_parent_start(mp, &du_src.ppargs); + if (error) + goto out_release_wip; + + if (du_wip.ip) { + error = xfs_parent_start(mp, &du_wip.ppargs); + if (error) + goto out_src_ppargs; + } + + if (target_ip) { + error = xfs_parent_start(mp, &du_tgt.ppargs); + if (error) + goto out_wip_ppargs; + } retry: nospace_error = 0; - spaceres = XFS_RENAME_SPACE_RES(mp, target_name->len); + spaceres = xfs_rename_space_res(mp, src_name->len, target_ip != NULL, + target_name->len, du_wip.ip != NULL); error = xfs_trans_alloc(mp, &M_RES(mp)->tr_rename, spaceres, 0, 0, &tp); if (error == -ENOSPC) { nospace_error = error; @@ -3023,14 +2255,26 @@ retry: &tp); } if (error) - goto out_release_wip; + goto out_tgt_ppargs; + + /* + * We don't allow reservationless renaming when parent pointers are + * enabled because we can't back out if the xattrs must grow. + */ + if (du_src.ppargs && nospace_error) { + error = nospace_error; + xfs_trans_cancel(tp); + goto out_tgt_ppargs; + } /* * Attach the dquots to the inodes */ error = xfs_qm_vop_rename_dqattach(inodes); - if (error) - goto out_trans_cancel; + if (error) { + xfs_trans_cancel(tp); + goto out_tgt_ppargs; + } /* * Lock all the participating inodes. Depending upon whether @@ -3041,18 +2285,16 @@ retry: xfs_lock_inodes(inodes, num_inodes, XFS_ILOCK_EXCL); /* - * Join all the inodes to the transaction. From this point on, - * we can rely on either trans_commit or trans_cancel to unlock - * them. + * Join all the inodes to the transaction. */ - xfs_trans_ijoin(tp, src_dp, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, src_dp, 0); if (new_parent) - xfs_trans_ijoin(tp, target_dp, XFS_ILOCK_EXCL); - xfs_trans_ijoin(tp, src_ip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, target_dp, 0); + xfs_trans_ijoin(tp, src_ip, 0); if (target_ip) - xfs_trans_ijoin(tp, target_ip, XFS_ILOCK_EXCL); - if (wip) - xfs_trans_ijoin(tp, wip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, target_ip, 0); + if (du_wip.ip) + xfs_trans_ijoin(tp, du_wip.ip, 0); /* * If we are using project inheritance, we only allow renames @@ -3066,10 +2308,13 @@ retry: } /* RENAME_EXCHANGE is unique from here on. */ - if (flags & RENAME_EXCHANGE) - return xfs_cross_rename(tp, src_dp, src_name, src_ip, - target_dp, target_name, target_ip, - spaceres); + if (flags & RENAME_EXCHANGE) { + error = xfs_dir_exchange_children(tp, &du_src, &du_tgt, + spaceres); + if (error) + goto out_trans_cancel; + goto out_commit; + } /* * Try to reserve quota to handle an expansion of the target directory. @@ -3083,6 +2328,7 @@ retry: if (error == -EDQUOT || error == -ENOSPC) { if (!retried) { xfs_trans_cancel(tp); + xfs_iunlock_rename(inodes, num_inodes); xfs_blockgc_free_quota(target_dp, 0); retried = true; goto retry; @@ -3097,30 +2343,12 @@ retry: } /* - * Check for expected errors before we dirty the transaction - * so we can return an error without a transaction abort. + * We don't allow quotaless renaming when parent pointers are enabled + * because we can't back out if the xattrs must grow. */ - if (target_ip == NULL) { - /* - * If there's no space reservation, check the entry will - * fit before actually inserting it. - */ - if (!spaceres) { - error = xfs_dir_canenter(tp, target_dp, target_name); - if (error) - goto out_trans_cancel; - } - } else { - /* - * If target exists and it's a directory, check that whether - * it can be destroyed. - */ - if (S_ISDIR(VFS_I(target_ip)->i_mode) && - (!xfs_dir_isempty(target_ip) || - (VFS_I(target_ip)->i_nlink > 2))) { - error = -EEXIST; - goto out_trans_cancel; - } + if (du_src.ppargs && nospace_error) { + error = nospace_error; + goto out_trans_cancel; } /* @@ -3134,7 +2362,7 @@ retry: * target_ip is either null or an empty directory. */ for (i = 0; i < num_inodes && inodes[i] != NULL; i++) { - if (inodes[i] == wip || + if (inodes[i] == du_wip.ip || (inodes[i] == target_ip && (VFS_I(target_ip)->i_nlink == 1 || src_is_directory))) { struct xfs_perag *pag; @@ -3142,182 +2370,52 @@ retry: pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, inodes[i]->i_ino)); - error = xfs_read_agi(pag, tp, &bp); + error = xfs_read_agi(pag, tp, 0, &bp); xfs_perag_put(pag); if (error) goto out_trans_cancel; } } - /* - * Directory entry creation below may acquire the AGF. Remove - * the whiteout from the unlinked list first to preserve correct - * AGI/AGF locking order. This dirties the transaction so failures - * after this point will abort and log recovery will clean up the - * mess. - * - * For whiteouts, we need to bump the link count on the whiteout - * inode. After this point, we have a real link, clear the tmpfile - * state flag from the inode so it doesn't accidentally get misused - * in future. - */ - if (wip) { - struct xfs_perag *pag; - - ASSERT(VFS_I(wip)->i_nlink == 0); - - pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, wip->i_ino)); - error = xfs_iunlink_remove(tp, pag, wip); - xfs_perag_put(pag); - if (error) - goto out_trans_cancel; - - xfs_bumplink(tp, wip); - VFS_I(wip)->i_state &= ~I_LINKABLE; - } - - /* - * Set up the target. - */ - if (target_ip == NULL) { - /* - * If target does not exist and the rename crosses - * directories, adjust the target directory link count - * to account for the ".." reference from the new entry. - */ - error = xfs_dir_createname(tp, target_dp, target_name, - src_ip->i_ino, spaceres); - if (error) - goto out_trans_cancel; - - xfs_trans_ichgtime(tp, target_dp, - XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); - - if (new_parent && src_is_directory) { - xfs_bumplink(tp, target_dp); - } - } else { /* target_ip != NULL */ - /* - * Link the source inode under the target name. - * If the source inode is a directory and we are moving - * it across directories, its ".." entry will be - * inconsistent until we replace that down below. - * - * In case there is already an entry with the same - * name at the destination directory, remove it first. - */ - error = xfs_dir_replace(tp, target_dp, target_name, - src_ip->i_ino, spaceres); - if (error) - goto out_trans_cancel; - - xfs_trans_ichgtime(tp, target_dp, - XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); - - /* - * Decrement the link count on the target since the target - * dir no longer points to it. - */ - error = xfs_droplink(tp, target_ip); - if (error) - goto out_trans_cancel; - - if (src_is_directory) { - /* - * Drop the link from the old "." entry. - */ - error = xfs_droplink(tp, target_ip); - if (error) - goto out_trans_cancel; - } - } /* target_ip != NULL */ - - /* - * Remove the source. - */ - if (new_parent && src_is_directory) { - /* - * Rewrite the ".." entry to point to the new - * directory. - */ - error = xfs_dir_replace(tp, src_ip, &xfs_name_dotdot, - target_dp->i_ino, spaceres); - ASSERT(error != -EEXIST); - if (error) - goto out_trans_cancel; - } - - /* - * We always want to hit the ctime on the source inode. - * - * This isn't strictly required by the standards since the source - * inode isn't really being changed, but old unix file systems did - * it and some incremental backup programs won't work without it. - */ - xfs_trans_ichgtime(tp, src_ip, XFS_ICHGTIME_CHG); - xfs_trans_log_inode(tp, src_ip, XFS_ILOG_CORE); - - /* - * Adjust the link count on src_dp. This is necessary when - * renaming a directory, either within one parent when - * the target existed, or across two parent directories. - */ - if (src_is_directory && (new_parent || target_ip != NULL)) { + error = xfs_dir_rename_children(tp, &du_src, &du_tgt, spaceres, + &du_wip); + if (error) + goto out_trans_cancel; + if (du_wip.ip) { /* - * Decrement link count on src_directory since the - * entry that's moved no longer points to it. + * Now we have a real link, clear the "I'm a tmpfile" state + * flag from the inode so it doesn't accidentally get misused in + * future. */ - error = xfs_droplink(tp, src_dp); - if (error) - goto out_trans_cancel; + VFS_I(du_wip.ip)->i_state &= ~I_LINKABLE; } +out_commit: /* - * For whiteouts, we only need to update the source dirent with the - * inode number of the whiteout inode rather than removing it - * altogether. - */ - if (wip) - error = xfs_dir_replace(tp, src_dp, src_name, wip->i_ino, - spaceres); - else - error = xfs_dir_removename(tp, src_dp, src_name, src_ip->i_ino, - spaceres); - - if (error) - goto out_trans_cancel; - - xfs_trans_ichgtime(tp, src_dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); - xfs_trans_log_inode(tp, src_dp, XFS_ILOG_CORE); - if (new_parent) - xfs_trans_log_inode(tp, target_dp, XFS_ILOG_CORE); - - /* - * Inform our hook clients that we've finished a rename operation as - * follows: removed the source and target files from their directories; - * that we've added the source to the target directory; and finally - * that we've added the whiteout, if there was one. All inodes are - * locked, so it's ok to model a rename this way so long as we say we - * deleted entries before we add new ones. + * If this is a synchronous mount, make sure that the rename + * transaction goes to disk before returning to the user. */ - if (target_ip) - xfs_dir_update_hook(target_dp, target_ip, -1, target_name); - xfs_dir_update_hook(src_dp, src_ip, -1, src_name); - xfs_dir_update_hook(target_dp, src_ip, 1, target_name); - if (wip) - xfs_dir_update_hook(src_dp, wip, 1, src_name); + if (xfs_has_wsync(tp->t_mountp) || xfs_has_dirsync(tp->t_mountp)) + xfs_trans_set_sync(tp); - error = xfs_finish_rename(tp); - if (wip) - xfs_irele(wip); - return error; + error = xfs_trans_commit(tp); + nospace_error = 0; + goto out_unlock; out_trans_cancel: xfs_trans_cancel(tp); +out_unlock: + xfs_iunlock_rename(inodes, num_inodes); +out_tgt_ppargs: + xfs_parent_finish(mp, du_tgt.ppargs); +out_wip_ppargs: + xfs_parent_finish(mp, du_wip.ppargs); +out_src_ppargs: + xfs_parent_finish(mp, du_src.ppargs); out_release_wip: - if (wip) - xfs_irele(wip); + if (du_wip.ip) + xfs_irele(du_wip.ip); if (error == -ENOSPC && nospace_error) error = nospace_error; return error; @@ -3457,6 +2555,7 @@ flush_out: iip->ili_last_fields = iip->ili_fields; iip->ili_fields = 0; iip->ili_fsync_fields = 0; + set_bit(XFS_LI_FLUSHING, &iip->ili_item.li_flags); spin_unlock(&iip->ili_lock); /* @@ -3814,7 +2913,7 @@ xfs_inode_reload_unlinked_bucket( /* Grab the first inode in the list */ pag = xfs_perag_get(mp, agno); - error = xfs_ialloc_read_agi(pag, tp, &agibp); + error = xfs_ialloc_read_agi(pag, tp, 0, &agibp); xfs_perag_put(pag); if (error) return error; @@ -3946,3 +3045,85 @@ xfs_inode_count_blocks( xfs_bmap_count_leaves(ifp, rblocks); *dblocks = ip->i_nblocks - *rblocks; } + +static void +xfs_wait_dax_page( + struct inode *inode) +{ + struct xfs_inode *ip = XFS_I(inode); + + xfs_iunlock(ip, XFS_MMAPLOCK_EXCL); + schedule(); + xfs_ilock(ip, XFS_MMAPLOCK_EXCL); +} + +int +xfs_break_dax_layouts( + struct inode *inode, + bool *retry) +{ + struct page *page; + + xfs_assert_ilocked(XFS_I(inode), XFS_MMAPLOCK_EXCL); + + page = dax_layout_busy_page(inode->i_mapping); + if (!page) + return 0; + + *retry = true; + return ___wait_var_event(&page->_refcount, + atomic_read(&page->_refcount) == 1, TASK_INTERRUPTIBLE, + 0, 0, xfs_wait_dax_page(inode)); +} + +int +xfs_break_layouts( + struct inode *inode, + uint *iolock, + enum layout_break_reason reason) +{ + bool retry; + int error; + + xfs_assert_ilocked(XFS_I(inode), XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL); + + do { + retry = false; + switch (reason) { + case BREAK_UNMAP: + error = xfs_break_dax_layouts(inode, &retry); + if (error || retry) + break; + fallthrough; + case BREAK_WRITE: + error = xfs_break_leased_layouts(inode, iolock, &retry); + break; + default: + WARN_ON_ONCE(1); + error = -EINVAL; + } + } while (error == 0 && retry); + + return error; +} + +/* Returns the size of fundamental allocation unit for a file, in bytes. */ +unsigned int +xfs_inode_alloc_unitsize( + struct xfs_inode *ip) +{ + unsigned int blocks = 1; + + if (XFS_IS_REALTIME_INODE(ip)) + blocks = ip->i_mount->m_sb.sb_rextsize; + + return XFS_FSB_TO_B(ip->i_mount, blocks); +} + +/* Should we always be using copy on write for file writes? */ +bool +xfs_is_always_cow_inode( + struct xfs_inode *ip) +{ + return ip->i_mount->m_always_cow && xfs_has_reflink(ip->i_mount); +} |