From 9e1f1de02c2275d7172e18dc4e7c2065777611bf Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 3 Jun 2011 18:24:58 -0400 Subject: more conservative S_NOSEC handling Caching "we have already removed suid/caps" was overenthusiastic as merged. On network filesystems we might have had suid/caps set on another client, silently picked by this client on revalidate, all of that *without* clearing the S_NOSEC flag. AFAICS, the only reasonably sane way to deal with that is * new superblock flag; unless set, S_NOSEC is not going to be set. * local block filesystems set it in their ->mount() (more accurately, mount_bdev() does, so does btrfs ->mount(), users of mount_bdev() other than local block ones clear it) * if any network filesystem (or a cluster one) wants to use S_NOSEC, it'll need to set MS_NOSEC in sb->s_flags *AND* take care to clear S_NOSEC when inode attribute changes are picked from other clients. It's not an earth-shattering hole (anybody that can set suid on another client will almost certainly be able to write to the file before doing that anyway), but it's a bug that needs fixing. Signed-off-by: Al Viro --- fs/ocfs2/super.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index cdbaf5e97308..56f61027236b 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -1072,7 +1072,7 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent) sb->s_magic = OCFS2_SUPER_MAGIC; - sb->s_flags = (sb->s_flags & ~MS_POSIXACL) | + sb->s_flags = (sb->s_flags & ~(MS_POSIXACL | MS_NOSEC)) | ((osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0); /* Hard readonly mode only if: bdev_read_only, MS_RDONLY, -- cgit v1.2.3-58-ga151 From 178ea73521d64ba41d7aa5488fb9f549c6d4507d Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 20 Jun 2011 11:31:30 -0400 Subject: kill check_acl callback of generic_permission() its value depends only on inode and does not change; we might as well store it in ->i_op->check_acl and be done with that. Signed-off-by: Al Viro --- fs/afs/security.c | 2 +- fs/btrfs/inode.c | 7 ++++++- fs/ceph/inode.c | 2 +- fs/cifs/cifsfs.c | 2 +- fs/fuse/dir.c | 4 ++-- fs/gfs2/inode.c | 5 ++++- fs/hostfs/hostfs_kern.c | 2 +- fs/hpfs/namei.c | 2 +- fs/namei.c | 17 +++++++---------- fs/nfs/dir.c | 2 +- fs/nilfs2/inode.c | 2 +- fs/ocfs2/file.c | 4 +++- fs/ocfs2/namei.c | 1 + fs/proc/base.c | 2 +- fs/reiserfs/file.c | 1 + fs/reiserfs/namei.c | 4 +++- fs/reiserfs/xattr.c | 18 ++++++++---------- fs/sysfs/inode.c | 2 +- include/linux/fs.h | 3 +-- include/linux/reiserfs_xattr.h | 2 ++ 20 files changed, 47 insertions(+), 37 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/afs/security.c b/fs/afs/security.c index f44b9d355377..745ee654165f 100644 --- a/fs/afs/security.c +++ b/fs/afs/security.c @@ -350,7 +350,7 @@ int afs_permission(struct inode *inode, int mask, unsigned int flags) } key_put(key); - ret = generic_permission(inode, mask, flags, NULL); + ret = generic_permission(inode, mask, flags); _leave(" = %d", ret); return ret; diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 3601f0aebddf..f0bd87371566 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -7339,7 +7339,7 @@ static int btrfs_permission(struct inode *inode, int mask, unsigned int flags) return -EROFS; if ((BTRFS_I(inode)->flags & BTRFS_INODE_READONLY) && (mask & MAY_WRITE)) return -EACCES; - return generic_permission(inode, mask, flags, btrfs_check_acl); + return generic_permission(inode, mask, flags); } static const struct inode_operations btrfs_dir_inode_operations = { @@ -7359,10 +7359,12 @@ static const struct inode_operations btrfs_dir_inode_operations = { .listxattr = btrfs_listxattr, .removexattr = btrfs_removexattr, .permission = btrfs_permission, + .check_acl = btrfs_check_acl, }; static const struct inode_operations btrfs_dir_ro_inode_operations = { .lookup = btrfs_lookup, .permission = btrfs_permission, + .check_acl = btrfs_check_acl, }; static const struct file_operations btrfs_dir_file_operations = { @@ -7431,6 +7433,7 @@ static const struct inode_operations btrfs_file_inode_operations = { .removexattr = btrfs_removexattr, .permission = btrfs_permission, .fiemap = btrfs_fiemap, + .check_acl = btrfs_check_acl, }; static const struct inode_operations btrfs_special_inode_operations = { .getattr = btrfs_getattr, @@ -7440,6 +7443,7 @@ static const struct inode_operations btrfs_special_inode_operations = { .getxattr = btrfs_getxattr, .listxattr = btrfs_listxattr, .removexattr = btrfs_removexattr, + .check_acl = btrfs_check_acl, }; static const struct inode_operations btrfs_symlink_inode_operations = { .readlink = generic_readlink, @@ -7451,6 +7455,7 @@ static const struct inode_operations btrfs_symlink_inode_operations = { .getxattr = btrfs_getxattr, .listxattr = btrfs_listxattr, .removexattr = btrfs_removexattr, + .check_acl = btrfs_check_acl, }; const struct dentry_operations btrfs_dentry_operations = { diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index d8858e96ab18..beb5d55d6fd2 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -1805,7 +1805,7 @@ int ceph_permission(struct inode *inode, int mask, unsigned int flags) err = ceph_do_getattr(inode, CEPH_CAP_AUTH_SHARED); if (!err) - err = generic_permission(inode, mask, flags, NULL); + err = generic_permission(inode, mask, flags); return err; } diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c index bc4b12ca537b..b79804fa410f 100644 --- a/fs/cifs/cifsfs.c +++ b/fs/cifs/cifsfs.c @@ -239,7 +239,7 @@ static int cifs_permission(struct inode *inode, int mask, unsigned int flags) on the client (above and beyond ACL on servers) for servers which do not support setting and viewing mode bits, so allowing client to check permissions is useful */ - return generic_permission(inode, mask, flags, NULL); + return generic_permission(inode, mask, flags); } static struct kmem_cache *cifs_inode_cachep; diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index d50160714595..0a2fcd860ad6 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -1018,7 +1018,7 @@ static int fuse_permission(struct inode *inode, int mask, unsigned int flags) } if (fc->flags & FUSE_DEFAULT_PERMISSIONS) { - err = generic_permission(inode, mask, flags, NULL); + err = generic_permission(inode, mask, flags); /* If permission is denied, try to refresh file attributes. This is also needed, because the root @@ -1027,7 +1027,7 @@ static int fuse_permission(struct inode *inode, int mask, unsigned int flags) err = fuse_perm_getattr(inode, flags); if (!err) err = generic_permission(inode, mask, - flags, NULL); + flags); } /* Note: the opposite of the above test does not diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c index 03e0c529063e..d5f0f4ea25dc 100644 --- a/fs/gfs2/inode.c +++ b/fs/gfs2/inode.c @@ -1564,7 +1564,7 @@ int gfs2_permission(struct inode *inode, int mask, unsigned int flags) if ((mask & MAY_WRITE) && IS_IMMUTABLE(inode)) error = -EACCES; else - error = generic_permission(inode, mask, flags, gfs2_check_acl); + error = generic_permission(inode, mask, flags); if (unlock) gfs2_glock_dq_uninit(&i_gh); @@ -1854,6 +1854,7 @@ const struct inode_operations gfs2_file_iops = { .listxattr = gfs2_listxattr, .removexattr = gfs2_removexattr, .fiemap = gfs2_fiemap, + .check_acl = gfs2_check_acl, }; const struct inode_operations gfs2_dir_iops = { @@ -1874,6 +1875,7 @@ const struct inode_operations gfs2_dir_iops = { .listxattr = gfs2_listxattr, .removexattr = gfs2_removexattr, .fiemap = gfs2_fiemap, + .check_acl = gfs2_check_acl, }; const struct inode_operations gfs2_symlink_iops = { @@ -1888,5 +1890,6 @@ const struct inode_operations gfs2_symlink_iops = { .listxattr = gfs2_listxattr, .removexattr = gfs2_removexattr, .fiemap = gfs2_fiemap, + .check_acl = gfs2_check_acl, }; diff --git a/fs/hostfs/hostfs_kern.c b/fs/hostfs/hostfs_kern.c index 2638c834ed28..a98d0d1aef65 100644 --- a/fs/hostfs/hostfs_kern.c +++ b/fs/hostfs/hostfs_kern.c @@ -770,7 +770,7 @@ int hostfs_permission(struct inode *ino, int desired, unsigned int flags) err = access_file(name, r, w, x); __putname(name); if (!err) - err = generic_permission(ino, desired, flags, NULL); + err = generic_permission(ino, desired, flags); return err; } diff --git a/fs/hpfs/namei.c b/fs/hpfs/namei.c index acf95dab2aac..bd2ce7dd8df3 100644 --- a/fs/hpfs/namei.c +++ b/fs/hpfs/namei.c @@ -398,7 +398,7 @@ again: hpfs_unlock(dir->i_sb); return -ENOSPC; } - if (generic_permission(inode, MAY_WRITE, 0, NULL) || + if (generic_permission(inode, MAY_WRITE, 0) || !S_ISREG(inode->i_mode) || get_write_access(inode)) { d_rehash(dentry); diff --git a/fs/namei.c b/fs/namei.c index d286cbc3f3e5..c5af0f37e67d 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -176,9 +176,9 @@ EXPORT_SYMBOL(putname); /* * This does basic POSIX ACL permission checking */ -static int acl_permission_check(struct inode *inode, int mask, unsigned int flags, - int (*check_acl)(struct inode *inode, int mask, unsigned int flags)) +static int acl_permission_check(struct inode *inode, int mask, unsigned int flags) { + int (*check_acl)(struct inode *inode, int mask, unsigned int flags); unsigned int mode = inode->i_mode; mask &= MAY_READ | MAY_WRITE | MAY_EXEC; @@ -189,6 +189,7 @@ static int acl_permission_check(struct inode *inode, int mask, unsigned int flag if (current_fsuid() == inode->i_uid) mode >>= 6; else { + check_acl = inode->i_op->check_acl; if (IS_POSIXACL(inode) && (mode & S_IRWXG) && check_acl) { int error = check_acl(inode, mask, flags); if (error != -EAGAIN) @@ -212,7 +213,6 @@ other_perms: * generic_permission - check for access rights on a Posix-like filesystem * @inode: inode to check access rights for * @mask: right to check for (%MAY_READ, %MAY_WRITE, %MAY_EXEC) - * @check_acl: optional callback to check for Posix ACLs * @flags: IPERM_FLAG_ flags. * * Used to check for read/write/execute permissions on a file. @@ -224,15 +224,14 @@ other_perms: * request cannot be satisfied (eg. requires blocking or too much complexity). * It would then be called again in ref-walk mode. */ -int generic_permission(struct inode *inode, int mask, unsigned int flags, - int (*check_acl)(struct inode *inode, int mask, unsigned int flags)) +int generic_permission(struct inode *inode, int mask, unsigned int flags) { int ret; /* * Do the basic POSIX ACL permission checks. */ - ret = acl_permission_check(inode, mask, flags, check_acl); + ret = acl_permission_check(inode, mask, flags); if (ret != -EACCES) return ret; @@ -290,8 +289,7 @@ int inode_permission(struct inode *inode, int mask) if (inode->i_op->permission) retval = inode->i_op->permission(inode, mask, 0); else - retval = generic_permission(inode, mask, 0, - inode->i_op->check_acl); + retval = generic_permission(inode, mask, 0); if (retval) return retval; @@ -326,8 +324,7 @@ static inline int exec_permission(struct inode *inode, unsigned int flags) if (likely(!ret)) goto ok; } else { - ret = acl_permission_check(inode, MAY_EXEC, flags, - inode->i_op->check_acl); + ret = acl_permission_check(inode, MAY_EXEC, flags); if (likely(!ret)) goto ok; if (ret != -EACCES) diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index ededdbd0db38..0485dca34fb1 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -2328,7 +2328,7 @@ out: out_notsup: res = nfs_revalidate_inode(NFS_SERVER(inode), inode); if (res == 0) - res = generic_permission(inode, mask, flags, NULL); + res = generic_permission(inode, mask, flags); goto out; } diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c index b9b45fc2903e..650aa7755003 100644 --- a/fs/nilfs2/inode.c +++ b/fs/nilfs2/inode.c @@ -806,7 +806,7 @@ int nilfs_permission(struct inode *inode, int mask, unsigned int flags) root->cno != NILFS_CPTREE_CURRENT_CNO) return -EROFS; /* snapshot is not writable */ - return generic_permission(inode, mask, flags, NULL); + return generic_permission(inode, mask, flags); } int nilfs_load_inode_block(struct inode *inode, struct buffer_head **pbh) diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index b1e35a392ca5..d058cb7e12d4 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c @@ -1293,7 +1293,7 @@ int ocfs2_permission(struct inode *inode, int mask, unsigned int flags) goto out; } - ret = generic_permission(inode, mask, flags, ocfs2_check_acl); + ret = generic_permission(inode, mask, flags); ocfs2_inode_unlock(inode, 0); out: @@ -2593,12 +2593,14 @@ const struct inode_operations ocfs2_file_iops = { .listxattr = ocfs2_listxattr, .removexattr = generic_removexattr, .fiemap = ocfs2_fiemap, + .check_acl = ocfs2_check_acl, }; const struct inode_operations ocfs2_special_file_iops = { .setattr = ocfs2_setattr, .getattr = ocfs2_getattr, .permission = ocfs2_permission, + .check_acl = ocfs2_check_acl, }; /* diff --git a/fs/ocfs2/namei.c b/fs/ocfs2/namei.c index e5d738cd9cc0..33889dc52dd7 100644 --- a/fs/ocfs2/namei.c +++ b/fs/ocfs2/namei.c @@ -2498,4 +2498,5 @@ const struct inode_operations ocfs2_dir_iops = { .listxattr = ocfs2_listxattr, .removexattr = generic_removexattr, .fiemap = ocfs2_fiemap, + .check_acl = ocfs2_check_acl, }; diff --git a/fs/proc/base.c b/fs/proc/base.c index fc5bc2767692..8b8470113576 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -2169,7 +2169,7 @@ static const struct file_operations proc_fd_operations = { */ static int proc_fd_permission(struct inode *inode, int mask, unsigned int flags) { - int rv = generic_permission(inode, mask, flags, NULL); + int rv = generic_permission(inode, mask, flags); if (rv == 0) return 0; if (task_pid(current) == proc_pid(inode)) diff --git a/fs/reiserfs/file.c b/fs/reiserfs/file.c index 91f080cc76c8..bbf31003d308 100644 --- a/fs/reiserfs/file.c +++ b/fs/reiserfs/file.c @@ -312,4 +312,5 @@ const struct inode_operations reiserfs_file_inode_operations = { .listxattr = reiserfs_listxattr, .removexattr = reiserfs_removexattr, .permission = reiserfs_permission, + .check_acl = reiserfs_check_acl, }; diff --git a/fs/reiserfs/namei.c b/fs/reiserfs/namei.c index 118662690cdf..551f1b79dbc4 100644 --- a/fs/reiserfs/namei.c +++ b/fs/reiserfs/namei.c @@ -1529,6 +1529,7 @@ const struct inode_operations reiserfs_dir_inode_operations = { .listxattr = reiserfs_listxattr, .removexattr = reiserfs_removexattr, .permission = reiserfs_permission, + .check_acl = reiserfs_check_acl, }; /* @@ -1545,6 +1546,7 @@ const struct inode_operations reiserfs_symlink_inode_operations = { .listxattr = reiserfs_listxattr, .removexattr = reiserfs_removexattr, .permission = reiserfs_permission, + .check_acl = reiserfs_check_acl, }; @@ -1558,5 +1560,5 @@ const struct inode_operations reiserfs_special_inode_operations = { .listxattr = reiserfs_listxattr, .removexattr = reiserfs_removexattr, .permission = reiserfs_permission, - + .check_acl = reiserfs_check_acl, }; diff --git a/fs/reiserfs/xattr.c b/fs/reiserfs/xattr.c index d78089690965..ddc5301d2986 100644 --- a/fs/reiserfs/xattr.c +++ b/fs/reiserfs/xattr.c @@ -868,11 +868,17 @@ out: return err; } -static int reiserfs_check_acl(struct inode *inode, int mask, unsigned int flags) +int reiserfs_check_acl(struct inode *inode, int mask, unsigned int flags) { struct posix_acl *acl; int error = -EAGAIN; /* do regular unix permission checks by default */ + /* + * Stat data v1 doesn't support ACLs. + */ + if (get_inode_sd_version(inode) == STAT_DATA_V1) + return -EAGAIN; + if (flags & IPERM_FLAG_RCU) return -ECHILD; @@ -961,15 +967,7 @@ int reiserfs_permission(struct inode *inode, int mask, unsigned int flags) if (IS_PRIVATE(inode)) return 0; -#ifdef CONFIG_REISERFS_FS_XATTR - /* - * Stat data v1 doesn't support ACLs. - */ - if (get_inode_sd_version(inode) != STAT_DATA_V1) - return generic_permission(inode, mask, flags, - reiserfs_check_acl); -#endif - return generic_permission(inode, mask, flags, NULL); + return generic_permission(inode, mask, flags); } static int xattr_hide_revalidate(struct dentry *dentry, struct nameidata *nd) diff --git a/fs/sysfs/inode.c b/fs/sysfs/inode.c index 0a12eb89cd32..a37165c64757 100644 --- a/fs/sysfs/inode.c +++ b/fs/sysfs/inode.c @@ -362,5 +362,5 @@ int sysfs_permission(struct inode *inode, int mask, unsigned int flags) sysfs_refresh_inode(sd, inode); mutex_unlock(&sysfs_mutex); - return generic_permission(inode, mask, flags, NULL); + return generic_permission(inode, mask, flags); } diff --git a/include/linux/fs.h b/include/linux/fs.h index 8c84ed930389..0c15d5e459d5 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -2187,8 +2187,7 @@ extern sector_t bmap(struct inode *, sector_t); #endif extern int notify_change(struct dentry *, struct iattr *); extern int inode_permission(struct inode *, int); -extern int generic_permission(struct inode *, int, unsigned int, - int (*check_acl)(struct inode *, int, unsigned int)); +extern int generic_permission(struct inode *, int, unsigned int); static inline bool execute_ok(struct inode *inode) { diff --git a/include/linux/reiserfs_xattr.h b/include/linux/reiserfs_xattr.h index 6deef5dc95fb..1a3ca8f80200 100644 --- a/include/linux/reiserfs_xattr.h +++ b/include/linux/reiserfs_xattr.h @@ -45,6 +45,7 @@ int reiserfs_permission(struct inode *inode, int mask, unsigned int flags); #ifdef CONFIG_REISERFS_FS_XATTR #define has_xattr_dir(inode) (REISERFS_I(inode)->i_flags & i_has_xattr_dir) +int reiserfs_check_acl(struct inode *inode, int mask, unsigned int flags); ssize_t reiserfs_getxattr(struct dentry *dentry, const char *name, void *buffer, size_t size); int reiserfs_setxattr(struct dentry *dentry, const char *name, @@ -122,6 +123,7 @@ static inline void reiserfs_init_xattr_rwsem(struct inode *inode) #define reiserfs_setxattr NULL #define reiserfs_listxattr NULL #define reiserfs_removexattr NULL +#define reiserfs_check_acl NULL static inline void reiserfs_init_xattr_rwsem(struct inode *inode) { -- cgit v1.2.3-58-ga151 From 9c2c703929e4c41210cfa6e3f599514421bab8dc Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 20 Jun 2011 19:06:22 -0400 Subject: ->permission() sanitizing: pass MAY_NOT_BLOCK to ->check_acl() Signed-off-by: Al Viro --- fs/9p/acl.c | 2 +- fs/btrfs/acl.c | 3 +-- fs/ext2/acl.c | 2 +- fs/ext3/acl.c | 2 +- fs/ext4/acl.c | 2 +- fs/generic_acl.c | 2 +- fs/gfs2/acl.c | 2 +- fs/jffs2/acl.c | 2 +- fs/jfs/acl.c | 2 +- fs/namei.c | 4 ++-- fs/ocfs2/acl.c | 2 +- fs/reiserfs/xattr.c | 2 +- fs/xfs/linux-2.6/xfs_acl.c | 2 +- 13 files changed, 14 insertions(+), 15 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/9p/acl.c b/fs/9p/acl.c index 535ab6eccb1a..94af68b092af 100644 --- a/fs/9p/acl.c +++ b/fs/9p/acl.c @@ -101,7 +101,7 @@ int v9fs_check_acl(struct inode *inode, int mask, unsigned int flags) struct posix_acl *acl; struct v9fs_session_info *v9ses; - if (flags & IPERM_FLAG_RCU) + if (mask & MAY_NOT_BLOCK) return -ECHILD; v9ses = v9fs_inode2v9ses(inode); diff --git a/fs/btrfs/acl.c b/fs/btrfs/acl.c index f66fc9959733..a25a4a2e0df2 100644 --- a/fs/btrfs/acl.c +++ b/fs/btrfs/acl.c @@ -199,10 +199,9 @@ int btrfs_check_acl(struct inode *inode, int mask, unsigned int flags) { int error = -EAGAIN; - if (flags & IPERM_FLAG_RCU) { + if (mask & MAY_NOT_BLOCK) { if (!negative_cached_acl(inode, ACL_TYPE_ACCESS)) error = -ECHILD; - } else { struct posix_acl *acl; acl = btrfs_get_acl(inode, ACL_TYPE_ACCESS); diff --git a/fs/ext2/acl.c b/fs/ext2/acl.c index abea5a17c764..6b9442d1be52 100644 --- a/fs/ext2/acl.c +++ b/fs/ext2/acl.c @@ -236,7 +236,7 @@ ext2_check_acl(struct inode *inode, int mask, unsigned int flags) { struct posix_acl *acl; - if (flags & IPERM_FLAG_RCU) { + if (mask & MAY_NOT_BLOCK) { if (!negative_cached_acl(inode, ACL_TYPE_ACCESS)) return -ECHILD; return -EAGAIN; diff --git a/fs/ext3/acl.c b/fs/ext3/acl.c index 9d021c0d472a..0a6940d6c30c 100644 --- a/fs/ext3/acl.c +++ b/fs/ext3/acl.c @@ -244,7 +244,7 @@ ext3_check_acl(struct inode *inode, int mask, unsigned int flags) { struct posix_acl *acl; - if (flags & IPERM_FLAG_RCU) { + if (mask & MAY_NOT_BLOCK) { if (!negative_cached_acl(inode, ACL_TYPE_ACCESS)) return -ECHILD; return -EAGAIN; diff --git a/fs/ext4/acl.c b/fs/ext4/acl.c index 21eacd7b7d79..4f54252e439e 100644 --- a/fs/ext4/acl.c +++ b/fs/ext4/acl.c @@ -242,7 +242,7 @@ ext4_check_acl(struct inode *inode, int mask, unsigned int flags) { struct posix_acl *acl; - if (flags & IPERM_FLAG_RCU) { + if (mask & MAY_NOT_BLOCK) { if (!negative_cached_acl(inode, ACL_TYPE_ACCESS)) return -ECHILD; return -EAGAIN; diff --git a/fs/generic_acl.c b/fs/generic_acl.c index 8f26d1a58912..5976bb1fa4ca 100644 --- a/fs/generic_acl.c +++ b/fs/generic_acl.c @@ -192,7 +192,7 @@ generic_acl_chmod(struct inode *inode) int generic_check_acl(struct inode *inode, int mask, unsigned int flags) { - if (flags & IPERM_FLAG_RCU) { + if (mask & MAY_NOT_BLOCK) { if (!negative_cached_acl(inode, ACL_TYPE_ACCESS)) return -ECHILD; } else { diff --git a/fs/gfs2/acl.c b/fs/gfs2/acl.c index cbc07155b1a0..4d97352d39a1 100644 --- a/fs/gfs2/acl.c +++ b/fs/gfs2/acl.c @@ -80,7 +80,7 @@ int gfs2_check_acl(struct inode *inode, int mask, unsigned int flags) struct posix_acl *acl; int error; - if (flags & IPERM_FLAG_RCU) { + if (mask & MAY_NOT_BLOCK) { if (!negative_cached_acl(inode, ACL_TYPE_ACCESS)) return -ECHILD; return -EAGAIN; diff --git a/fs/jffs2/acl.c b/fs/jffs2/acl.c index 828a0e1ea438..952afb59e6f1 100644 --- a/fs/jffs2/acl.c +++ b/fs/jffs2/acl.c @@ -264,7 +264,7 @@ int jffs2_check_acl(struct inode *inode, int mask, unsigned int flags) struct posix_acl *acl; int rc; - if (flags & IPERM_FLAG_RCU) + if (mask & MAY_NOT_BLOCK) return -ECHILD; acl = jffs2_get_acl(inode, ACL_TYPE_ACCESS); diff --git a/fs/jfs/acl.c b/fs/jfs/acl.c index e5de9422fa32..859ae5a92166 100644 --- a/fs/jfs/acl.c +++ b/fs/jfs/acl.c @@ -118,7 +118,7 @@ int jfs_check_acl(struct inode *inode, int mask, unsigned int flags) { struct posix_acl *acl; - if (flags & IPERM_FLAG_RCU) + if (mask & MAY_NOT_BLOCK) return -ECHILD; acl = jfs_get_acl(inode, ACL_TYPE_ACCESS); diff --git a/fs/namei.c b/fs/namei.c index 723a3fe4bc40..e0624e2f0bba 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -181,7 +181,7 @@ static int acl_permission_check(struct inode *inode, int mask, unsigned int flag int (*check_acl)(struct inode *inode, int mask, unsigned int flags); unsigned int mode = inode->i_mode; - mask &= MAY_READ | MAY_WRITE | MAY_EXEC; + mask &= MAY_READ | MAY_WRITE | MAY_EXEC | MAY_NOT_BLOCK; if (current_user_ns() != inode_userns(inode)) goto other_perms; @@ -204,7 +204,7 @@ other_perms: /* * If the DACs are ok we don't need any capability check. */ - if ((mask & ~mode) == 0) + if ((mask & ~mode & (MAY_READ | MAY_WRITE | MAY_EXEC)) == 0) return 0; return -EACCES; } diff --git a/fs/ocfs2/acl.c b/fs/ocfs2/acl.c index e913ad130fdd..4b683ccc4506 100644 --- a/fs/ocfs2/acl.c +++ b/fs/ocfs2/acl.c @@ -297,7 +297,7 @@ int ocfs2_check_acl(struct inode *inode, int mask, unsigned int flags) struct posix_acl *acl; int ret = -EAGAIN; - if (flags & IPERM_FLAG_RCU) + if (mask & MAY_NOT_BLOCK) return -ECHILD; osb = OCFS2_SB(inode->i_sb); diff --git a/fs/reiserfs/xattr.c b/fs/reiserfs/xattr.c index ddc5301d2986..6747470ec103 100644 --- a/fs/reiserfs/xattr.c +++ b/fs/reiserfs/xattr.c @@ -879,7 +879,7 @@ int reiserfs_check_acl(struct inode *inode, int mask, unsigned int flags) if (get_inode_sd_version(inode) == STAT_DATA_V1) return -EAGAIN; - if (flags & IPERM_FLAG_RCU) + if (mask & MAY_NOT_BLOCK) return -ECHILD; acl = reiserfs_get_acl(inode, ACL_TYPE_ACCESS); diff --git a/fs/xfs/linux-2.6/xfs_acl.c b/fs/xfs/linux-2.6/xfs_acl.c index 39f4f809bb68..278e6736135a 100644 --- a/fs/xfs/linux-2.6/xfs_acl.c +++ b/fs/xfs/linux-2.6/xfs_acl.c @@ -235,7 +235,7 @@ xfs_check_acl(struct inode *inode, int mask, unsigned int flags) if (!XFS_IFORK_Q(ip)) return -EAGAIN; - if (flags & IPERM_FLAG_RCU) { + if (mask & MAY_NOT_BLOCK) { if (!negative_cached_acl(inode, ACL_TYPE_ACCESS)) return -ECHILD; return -EAGAIN; -- cgit v1.2.3-58-ga151 From 7e40145eb111a5192e6d819f764db9d6828d1abb Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 20 Jun 2011 19:12:17 -0400 Subject: ->permission() sanitizing: don't pass flags to ->check_acl() not used in the instances anymore. Signed-off-by: Al Viro --- Documentation/filesystems/Locking | 2 +- Documentation/filesystems/vfs.txt | 2 +- fs/9p/acl.c | 2 +- fs/9p/acl.h | 2 +- fs/btrfs/acl.c | 2 +- fs/btrfs/ctree.h | 2 +- fs/ext2/acl.c | 2 +- fs/ext2/acl.h | 2 +- fs/ext3/acl.c | 2 +- fs/ext3/acl.h | 2 +- fs/ext4/acl.c | 2 +- fs/ext4/acl.h | 2 +- fs/generic_acl.c | 2 +- fs/gfs2/acl.c | 2 +- fs/gfs2/acl.h | 2 +- fs/jffs2/acl.c | 2 +- fs/jffs2/acl.h | 2 +- fs/jfs/acl.c | 2 +- fs/jfs/jfs_acl.h | 2 +- fs/namei.c | 10 +++++----- fs/ocfs2/acl.c | 2 +- fs/ocfs2/acl.h | 2 +- fs/reiserfs/xattr.c | 2 +- fs/xfs/linux-2.6/xfs_acl.c | 2 +- fs/xfs/xfs_acl.h | 2 +- include/linux/fs.h | 2 +- include/linux/generic_acl.h | 2 +- include/linux/reiserfs_xattr.h | 2 +- 28 files changed, 32 insertions(+), 32 deletions(-) (limited to 'fs/ocfs2') diff --git a/Documentation/filesystems/Locking b/Documentation/filesystems/Locking index 57d827d6071d..9b6ed7c9f34f 100644 --- a/Documentation/filesystems/Locking +++ b/Documentation/filesystems/Locking @@ -52,7 +52,7 @@ ata *); void (*put_link) (struct dentry *, struct nameidata *, void *); void (*truncate) (struct inode *); int (*permission) (struct inode *, int, unsigned int); - int (*check_acl)(struct inode *, int, unsigned int); + int (*check_acl)(struct inode *, int); int (*setattr) (struct dentry *, struct iattr *); int (*getattr) (struct vfsmount *, struct dentry *, struct kstat *); int (*setxattr) (struct dentry *, const char *,const void *,size_t,int); diff --git a/Documentation/filesystems/vfs.txt b/Documentation/filesystems/vfs.txt index 88b9f5519af9..8b4c8e04d879 100644 --- a/Documentation/filesystems/vfs.txt +++ b/Documentation/filesystems/vfs.txt @@ -334,7 +334,7 @@ struct inode_operations { void (*put_link) (struct dentry *, struct nameidata *, void *); void (*truncate) (struct inode *); int (*permission) (struct inode *, int, unsigned int); - int (*check_acl)(struct inode *, int, unsigned int); + int (*check_acl)(struct inode *, int); int (*setattr) (struct dentry *, struct iattr *); int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *); int (*setxattr) (struct dentry *, const char *,const void *,size_t,int); diff --git a/fs/9p/acl.c b/fs/9p/acl.c index 94af68b092af..e98f56d3787d 100644 --- a/fs/9p/acl.c +++ b/fs/9p/acl.c @@ -96,7 +96,7 @@ static struct posix_acl *v9fs_get_cached_acl(struct inode *inode, int type) return acl; } -int v9fs_check_acl(struct inode *inode, int mask, unsigned int flags) +int v9fs_check_acl(struct inode *inode, int mask) { struct posix_acl *acl; struct v9fs_session_info *v9ses; diff --git a/fs/9p/acl.h b/fs/9p/acl.h index 7ef3ac9f6d95..59e18c2e8c7e 100644 --- a/fs/9p/acl.h +++ b/fs/9p/acl.h @@ -16,7 +16,7 @@ #ifdef CONFIG_9P_FS_POSIX_ACL extern int v9fs_get_acl(struct inode *, struct p9_fid *); -extern int v9fs_check_acl(struct inode *inode, int mask, unsigned int flags); +extern int v9fs_check_acl(struct inode *inode, int mask); extern int v9fs_acl_chmod(struct dentry *); extern int v9fs_set_create_acl(struct dentry *, struct posix_acl *, struct posix_acl *); diff --git a/fs/btrfs/acl.c b/fs/btrfs/acl.c index a25a4a2e0df2..9f62ab2a7282 100644 --- a/fs/btrfs/acl.c +++ b/fs/btrfs/acl.c @@ -195,7 +195,7 @@ out: return ret; } -int btrfs_check_acl(struct inode *inode, int mask, unsigned int flags) +int btrfs_check_acl(struct inode *inode, int mask) { int error = -EAGAIN; diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 3b859a3e6a0e..9864cec801ed 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -2642,7 +2642,7 @@ do { \ /* acl.c */ #ifdef CONFIG_BTRFS_FS_POSIX_ACL -int btrfs_check_acl(struct inode *inode, int mask, unsigned int flags); +int btrfs_check_acl(struct inode *inode, int mask); #else #define btrfs_check_acl NULL #endif diff --git a/fs/ext2/acl.c b/fs/ext2/acl.c index 6b9442d1be52..bfe651f9ae16 100644 --- a/fs/ext2/acl.c +++ b/fs/ext2/acl.c @@ -232,7 +232,7 @@ ext2_set_acl(struct inode *inode, int type, struct posix_acl *acl) } int -ext2_check_acl(struct inode *inode, int mask, unsigned int flags) +ext2_check_acl(struct inode *inode, int mask) { struct posix_acl *acl; diff --git a/fs/ext2/acl.h b/fs/ext2/acl.h index c939b7b12099..3ff6cbb9ac44 100644 --- a/fs/ext2/acl.h +++ b/fs/ext2/acl.h @@ -54,7 +54,7 @@ static inline int ext2_acl_count(size_t size) #ifdef CONFIG_EXT2_FS_POSIX_ACL /* acl.c */ -extern int ext2_check_acl (struct inode *, int, unsigned int); +extern int ext2_check_acl (struct inode *, int); extern int ext2_acl_chmod (struct inode *); extern int ext2_init_acl (struct inode *, struct inode *); diff --git a/fs/ext3/acl.c b/fs/ext3/acl.c index 0a6940d6c30c..edfeb293d4cb 100644 --- a/fs/ext3/acl.c +++ b/fs/ext3/acl.c @@ -240,7 +240,7 @@ ext3_set_acl(handle_t *handle, struct inode *inode, int type, } int -ext3_check_acl(struct inode *inode, int mask, unsigned int flags) +ext3_check_acl(struct inode *inode, int mask) { struct posix_acl *acl; diff --git a/fs/ext3/acl.h b/fs/ext3/acl.h index 5faf8048e906..597334626de9 100644 --- a/fs/ext3/acl.h +++ b/fs/ext3/acl.h @@ -54,7 +54,7 @@ static inline int ext3_acl_count(size_t size) #ifdef CONFIG_EXT3_FS_POSIX_ACL /* acl.c */ -extern int ext3_check_acl (struct inode *, int, unsigned int); +extern int ext3_check_acl (struct inode *, int); extern int ext3_acl_chmod (struct inode *); extern int ext3_init_acl (handle_t *, struct inode *, struct inode *); diff --git a/fs/ext4/acl.c b/fs/ext4/acl.c index 4f54252e439e..60d900fcc3db 100644 --- a/fs/ext4/acl.c +++ b/fs/ext4/acl.c @@ -238,7 +238,7 @@ ext4_set_acl(handle_t *handle, struct inode *inode, int type, } int -ext4_check_acl(struct inode *inode, int mask, unsigned int flags) +ext4_check_acl(struct inode *inode, int mask) { struct posix_acl *acl; diff --git a/fs/ext4/acl.h b/fs/ext4/acl.h index dec821168fd4..9d843d5deac4 100644 --- a/fs/ext4/acl.h +++ b/fs/ext4/acl.h @@ -54,7 +54,7 @@ static inline int ext4_acl_count(size_t size) #ifdef CONFIG_EXT4_FS_POSIX_ACL /* acl.c */ -extern int ext4_check_acl(struct inode *, int, unsigned int); +extern int ext4_check_acl(struct inode *, int); extern int ext4_acl_chmod(struct inode *); extern int ext4_init_acl(handle_t *, struct inode *, struct inode *); diff --git a/fs/generic_acl.c b/fs/generic_acl.c index 5976bb1fa4ca..70e90b4974ce 100644 --- a/fs/generic_acl.c +++ b/fs/generic_acl.c @@ -190,7 +190,7 @@ generic_acl_chmod(struct inode *inode) } int -generic_check_acl(struct inode *inode, int mask, unsigned int flags) +generic_check_acl(struct inode *inode, int mask) { if (mask & MAY_NOT_BLOCK) { if (!negative_cached_acl(inode, ACL_TYPE_ACCESS)) diff --git a/fs/gfs2/acl.c b/fs/gfs2/acl.c index 4d97352d39a1..8ef1079f1665 100644 --- a/fs/gfs2/acl.c +++ b/fs/gfs2/acl.c @@ -75,7 +75,7 @@ static struct posix_acl *gfs2_acl_get(struct gfs2_inode *ip, int type) * Returns: errno */ -int gfs2_check_acl(struct inode *inode, int mask, unsigned int flags) +int gfs2_check_acl(struct inode *inode, int mask) { struct posix_acl *acl; int error; diff --git a/fs/gfs2/acl.h b/fs/gfs2/acl.h index a93907c8159b..b522b0cb39ea 100644 --- a/fs/gfs2/acl.h +++ b/fs/gfs2/acl.h @@ -16,7 +16,7 @@ #define GFS2_POSIX_ACL_DEFAULT "posix_acl_default" #define GFS2_ACL_MAX_ENTRIES 25 -extern int gfs2_check_acl(struct inode *inode, int mask, unsigned int); +extern int gfs2_check_acl(struct inode *inode, int mask); extern int gfs2_acl_create(struct gfs2_inode *dip, struct inode *inode); extern int gfs2_acl_chmod(struct gfs2_inode *ip, struct iattr *attr); extern const struct xattr_handler gfs2_xattr_system_handler; diff --git a/fs/jffs2/acl.c b/fs/jffs2/acl.c index 952afb59e6f1..3675b3cdee89 100644 --- a/fs/jffs2/acl.c +++ b/fs/jffs2/acl.c @@ -259,7 +259,7 @@ static int jffs2_set_acl(struct inode *inode, int type, struct posix_acl *acl) return rc; } -int jffs2_check_acl(struct inode *inode, int mask, unsigned int flags) +int jffs2_check_acl(struct inode *inode, int mask) { struct posix_acl *acl; int rc; diff --git a/fs/jffs2/acl.h b/fs/jffs2/acl.h index 3119f59253d3..5e42de8d9541 100644 --- a/fs/jffs2/acl.h +++ b/fs/jffs2/acl.h @@ -26,7 +26,7 @@ struct jffs2_acl_header { #ifdef CONFIG_JFFS2_FS_POSIX_ACL -extern int jffs2_check_acl(struct inode *, int, unsigned int); +extern int jffs2_check_acl(struct inode *, int); extern int jffs2_acl_chmod(struct inode *); extern int jffs2_init_acl_pre(struct inode *, struct inode *, int *); extern int jffs2_init_acl_post(struct inode *); diff --git a/fs/jfs/acl.c b/fs/jfs/acl.c index 859ae5a92166..8a0a0666d5a6 100644 --- a/fs/jfs/acl.c +++ b/fs/jfs/acl.c @@ -114,7 +114,7 @@ out: return rc; } -int jfs_check_acl(struct inode *inode, int mask, unsigned int flags) +int jfs_check_acl(struct inode *inode, int mask) { struct posix_acl *acl; diff --git a/fs/jfs/jfs_acl.h b/fs/jfs/jfs_acl.h index f9285c4900fa..54e07559878d 100644 --- a/fs/jfs/jfs_acl.h +++ b/fs/jfs/jfs_acl.h @@ -20,7 +20,7 @@ #ifdef CONFIG_JFS_POSIX_ACL -int jfs_check_acl(struct inode *, int, unsigned int flags); +int jfs_check_acl(struct inode *, int); int jfs_init_acl(tid_t, struct inode *, struct inode *); int jfs_acl_chmod(struct inode *inode); diff --git a/fs/namei.c b/fs/namei.c index e0624e2f0bba..560fd1dff1d0 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -176,9 +176,9 @@ EXPORT_SYMBOL(putname); /* * This does basic POSIX ACL permission checking */ -static int acl_permission_check(struct inode *inode, int mask, unsigned int flags) +static int acl_permission_check(struct inode *inode, int mask) { - int (*check_acl)(struct inode *inode, int mask, unsigned int flags); + int (*check_acl)(struct inode *inode, int mask); unsigned int mode = inode->i_mode; mask &= MAY_READ | MAY_WRITE | MAY_EXEC | MAY_NOT_BLOCK; @@ -191,7 +191,7 @@ static int acl_permission_check(struct inode *inode, int mask, unsigned int flag else { check_acl = inode->i_op->check_acl; if (IS_POSIXACL(inode) && (mode & S_IRWXG) && check_acl) { - int error = check_acl(inode, mask, flags); + int error = check_acl(inode, mask); if (error != -EAGAIN) return error; } @@ -231,7 +231,7 @@ int generic_permission(struct inode *inode, int mask, unsigned int flags) /* * Do the basic POSIX ACL permission checks. */ - ret = acl_permission_check(inode, mask, flags); + ret = acl_permission_check(inode, mask); if (ret != -EACCES) return ret; @@ -327,7 +327,7 @@ static inline int exec_permission(struct inode *inode, unsigned int flags) if (likely(!ret)) goto ok; } else { - ret = acl_permission_check(inode, mask, flags); + ret = acl_permission_check(inode, mask); if (likely(!ret)) goto ok; if (ret != -EACCES) diff --git a/fs/ocfs2/acl.c b/fs/ocfs2/acl.c index 4b683ccc4506..1cee970eb55a 100644 --- a/fs/ocfs2/acl.c +++ b/fs/ocfs2/acl.c @@ -290,7 +290,7 @@ static int ocfs2_set_acl(handle_t *handle, return ret; } -int ocfs2_check_acl(struct inode *inode, int mask, unsigned int flags) +int ocfs2_check_acl(struct inode *inode, int mask) { struct ocfs2_super *osb; struct buffer_head *di_bh = NULL; diff --git a/fs/ocfs2/acl.h b/fs/ocfs2/acl.h index 4fe7c9cf4bfb..5c5d31f05853 100644 --- a/fs/ocfs2/acl.h +++ b/fs/ocfs2/acl.h @@ -26,7 +26,7 @@ struct ocfs2_acl_entry { __le32 e_id; }; -extern int ocfs2_check_acl(struct inode *, int, unsigned int); +extern int ocfs2_check_acl(struct inode *, int); extern int ocfs2_acl_chmod(struct inode *); extern int ocfs2_init_acl(handle_t *, struct inode *, struct inode *, struct buffer_head *, struct buffer_head *, diff --git a/fs/reiserfs/xattr.c b/fs/reiserfs/xattr.c index 6747470ec103..6ee3c11aa8d9 100644 --- a/fs/reiserfs/xattr.c +++ b/fs/reiserfs/xattr.c @@ -868,7 +868,7 @@ out: return err; } -int reiserfs_check_acl(struct inode *inode, int mask, unsigned int flags) +int reiserfs_check_acl(struct inode *inode, int mask) { struct posix_acl *acl; int error = -EAGAIN; /* do regular unix permission checks by default */ diff --git a/fs/xfs/linux-2.6/xfs_acl.c b/fs/xfs/linux-2.6/xfs_acl.c index 278e6736135a..a5dcd6a0f1b5 100644 --- a/fs/xfs/linux-2.6/xfs_acl.c +++ b/fs/xfs/linux-2.6/xfs_acl.c @@ -219,7 +219,7 @@ xfs_set_acl(struct inode *inode, int type, struct posix_acl *acl) } int -xfs_check_acl(struct inode *inode, int mask, unsigned int flags) +xfs_check_acl(struct inode *inode, int mask) { struct xfs_inode *ip; struct posix_acl *acl; diff --git a/fs/xfs/xfs_acl.h b/fs/xfs/xfs_acl.h index 11dd72070cbb..0135e2a669d7 100644 --- a/fs/xfs/xfs_acl.h +++ b/fs/xfs/xfs_acl.h @@ -42,7 +42,7 @@ struct xfs_acl { #define SGI_ACL_DEFAULT_SIZE (sizeof(SGI_ACL_DEFAULT)-1) #ifdef CONFIG_XFS_POSIX_ACL -extern int xfs_check_acl(struct inode *inode, int mask, unsigned int flags); +extern int xfs_check_acl(struct inode *inode, int mask); extern struct posix_acl *xfs_get_acl(struct inode *inode, int type); extern int xfs_inherit_acl(struct inode *inode, struct posix_acl *default_acl); extern int xfs_acl_chmod(struct inode *inode); diff --git a/include/linux/fs.h b/include/linux/fs.h index 60c1fe65bb2d..f218b42718aa 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1579,7 +1579,7 @@ struct inode_operations { struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *); void * (*follow_link) (struct dentry *, struct nameidata *); int (*permission) (struct inode *, int, unsigned int); - int (*check_acl)(struct inode *, int, unsigned int); + int (*check_acl)(struct inode *, int); int (*readlink) (struct dentry *, char __user *,int); void (*put_link) (struct dentry *, struct nameidata *, void *); diff --git a/include/linux/generic_acl.h b/include/linux/generic_acl.h index 0437e377b555..574bea4013b6 100644 --- a/include/linux/generic_acl.h +++ b/include/linux/generic_acl.h @@ -10,6 +10,6 @@ extern const struct xattr_handler generic_acl_default_handler; int generic_acl_init(struct inode *, struct inode *); int generic_acl_chmod(struct inode *); -int generic_check_acl(struct inode *inode, int mask, unsigned int flags); +int generic_check_acl(struct inode *inode, int mask); #endif /* LINUX_GENERIC_ACL_H */ diff --git a/include/linux/reiserfs_xattr.h b/include/linux/reiserfs_xattr.h index 1a3ca8f80200..424ba6416e7c 100644 --- a/include/linux/reiserfs_xattr.h +++ b/include/linux/reiserfs_xattr.h @@ -45,7 +45,7 @@ int reiserfs_permission(struct inode *inode, int mask, unsigned int flags); #ifdef CONFIG_REISERFS_FS_XATTR #define has_xattr_dir(inode) (REISERFS_I(inode)->i_flags & i_has_xattr_dir) -int reiserfs_check_acl(struct inode *inode, int mask, unsigned int flags); +int reiserfs_check_acl(struct inode *inode, int mask); ssize_t reiserfs_getxattr(struct dentry *dentry, const char *name, void *buffer, size_t size); int reiserfs_setxattr(struct dentry *dentry, const char *name, -- cgit v1.2.3-58-ga151 From 2830ba7f34ebb27c4e5b8b6ef408cd6d74860890 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 20 Jun 2011 19:16:29 -0400 Subject: ->permission() sanitizing: don't pass flags to generic_permission() redundant; all callers get it duplicated in mask & MAY_NOT_BLOCK and none of them removes that bit. Signed-off-by: Al Viro --- fs/afs/security.c | 2 +- fs/btrfs/inode.c | 2 +- fs/ceph/inode.c | 2 +- fs/cifs/cifsfs.c | 2 +- fs/fuse/dir.c | 5 ++--- fs/gfs2/inode.c | 2 +- fs/hostfs/hostfs_kern.c | 2 +- fs/hpfs/namei.c | 2 +- fs/namei.c | 4 ++-- fs/nfs/dir.c | 2 +- fs/nilfs2/inode.c | 2 +- fs/ocfs2/file.c | 2 +- fs/proc/base.c | 2 +- fs/reiserfs/xattr.c | 2 +- fs/sysfs/inode.c | 2 +- include/linux/fs.h | 2 +- 16 files changed, 18 insertions(+), 19 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/afs/security.c b/fs/afs/security.c index 745ee654165f..ab6b3147f05f 100644 --- a/fs/afs/security.c +++ b/fs/afs/security.c @@ -350,7 +350,7 @@ int afs_permission(struct inode *inode, int mask, unsigned int flags) } key_put(key); - ret = generic_permission(inode, mask, flags); + ret = generic_permission(inode, mask); _leave(" = %d", ret); return ret; diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index f0bd87371566..b61b0477a8ac 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -7339,7 +7339,7 @@ static int btrfs_permission(struct inode *inode, int mask, unsigned int flags) return -EROFS; if ((BTRFS_I(inode)->flags & BTRFS_INODE_READONLY) && (mask & MAY_WRITE)) return -EACCES; - return generic_permission(inode, mask, flags); + return generic_permission(inode, mask); } static const struct inode_operations btrfs_dir_inode_operations = { diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index beb5d55d6fd2..9c169741b416 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -1805,7 +1805,7 @@ int ceph_permission(struct inode *inode, int mask, unsigned int flags) err = ceph_do_getattr(inode, CEPH_CAP_AUTH_SHARED); if (!err) - err = generic_permission(inode, mask, flags); + err = generic_permission(inode, mask); return err; } diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c index b79804fa410f..b285b5bcf711 100644 --- a/fs/cifs/cifsfs.c +++ b/fs/cifs/cifsfs.c @@ -239,7 +239,7 @@ static int cifs_permission(struct inode *inode, int mask, unsigned int flags) on the client (above and beyond ACL on servers) for servers which do not support setting and viewing mode bits, so allowing client to check permissions is useful */ - return generic_permission(inode, mask, flags); + return generic_permission(inode, mask); } static struct kmem_cache *cifs_inode_cachep; diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index 0a2fcd860ad6..0df56b6c26e2 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -1018,7 +1018,7 @@ static int fuse_permission(struct inode *inode, int mask, unsigned int flags) } if (fc->flags & FUSE_DEFAULT_PERMISSIONS) { - err = generic_permission(inode, mask, flags); + err = generic_permission(inode, mask); /* If permission is denied, try to refresh file attributes. This is also needed, because the root @@ -1026,8 +1026,7 @@ static int fuse_permission(struct inode *inode, int mask, unsigned int flags) if (err == -EACCES && !refreshed) { err = fuse_perm_getattr(inode, flags); if (!err) - err = generic_permission(inode, mask, - flags); + err = generic_permission(inode, mask); } /* Note: the opposite of the above test does not diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c index d5f0f4ea25dc..b776ec8f9c19 100644 --- a/fs/gfs2/inode.c +++ b/fs/gfs2/inode.c @@ -1564,7 +1564,7 @@ int gfs2_permission(struct inode *inode, int mask, unsigned int flags) if ((mask & MAY_WRITE) && IS_IMMUTABLE(inode)) error = -EACCES; else - error = generic_permission(inode, mask, flags); + error = generic_permission(inode, mask); if (unlock) gfs2_glock_dq_uninit(&i_gh); diff --git a/fs/hostfs/hostfs_kern.c b/fs/hostfs/hostfs_kern.c index a98d0d1aef65..b1bc31bde833 100644 --- a/fs/hostfs/hostfs_kern.c +++ b/fs/hostfs/hostfs_kern.c @@ -770,7 +770,7 @@ int hostfs_permission(struct inode *ino, int desired, unsigned int flags) err = access_file(name, r, w, x); __putname(name); if (!err) - err = generic_permission(ino, desired, flags); + err = generic_permission(ino, desired); return err; } diff --git a/fs/hpfs/namei.c b/fs/hpfs/namei.c index bd2ce7dd8df3..2df69e2f07cf 100644 --- a/fs/hpfs/namei.c +++ b/fs/hpfs/namei.c @@ -398,7 +398,7 @@ again: hpfs_unlock(dir->i_sb); return -ENOSPC; } - if (generic_permission(inode, MAY_WRITE, 0) || + if (generic_permission(inode, MAY_WRITE) || !S_ISREG(inode->i_mode) || get_write_access(inode)) { d_rehash(dentry); diff --git a/fs/namei.c b/fs/namei.c index 560fd1dff1d0..684e0f30cf4c 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -224,7 +224,7 @@ other_perms: * request cannot be satisfied (eg. requires blocking or too much complexity). * It would then be called again in ref-walk mode. */ -int generic_permission(struct inode *inode, int mask, unsigned int flags) +int generic_permission(struct inode *inode, int mask) { int ret; @@ -289,7 +289,7 @@ int inode_permission(struct inode *inode, int mask) if (inode->i_op->permission) retval = inode->i_op->permission(inode, mask, 0); else - retval = generic_permission(inode, mask, 0); + retval = generic_permission(inode, mask); if (retval) return retval; diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index 0485dca34fb1..16cf84b4afb9 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -2328,7 +2328,7 @@ out: out_notsup: res = nfs_revalidate_inode(NFS_SERVER(inode), inode); if (res == 0) - res = generic_permission(inode, mask, flags); + res = generic_permission(inode, mask); goto out; } diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c index 650aa7755003..0df6de58bbcb 100644 --- a/fs/nilfs2/inode.c +++ b/fs/nilfs2/inode.c @@ -806,7 +806,7 @@ int nilfs_permission(struct inode *inode, int mask, unsigned int flags) root->cno != NILFS_CPTREE_CURRENT_CNO) return -EROFS; /* snapshot is not writable */ - return generic_permission(inode, mask, flags); + return generic_permission(inode, mask); } int nilfs_load_inode_block(struct inode *inode, struct buffer_head **pbh) diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index d058cb7e12d4..ecb52b028964 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c @@ -1293,7 +1293,7 @@ int ocfs2_permission(struct inode *inode, int mask, unsigned int flags) goto out; } - ret = generic_permission(inode, mask, flags); + ret = generic_permission(inode, mask); ocfs2_inode_unlock(inode, 0); out: diff --git a/fs/proc/base.c b/fs/proc/base.c index 8b8470113576..53a1a961b25f 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -2169,7 +2169,7 @@ static const struct file_operations proc_fd_operations = { */ static int proc_fd_permission(struct inode *inode, int mask, unsigned int flags) { - int rv = generic_permission(inode, mask, flags); + int rv = generic_permission(inode, mask); if (rv == 0) return 0; if (task_pid(current) == proc_pid(inode)) diff --git a/fs/reiserfs/xattr.c b/fs/reiserfs/xattr.c index 6ee3c11aa8d9..f17319613a85 100644 --- a/fs/reiserfs/xattr.c +++ b/fs/reiserfs/xattr.c @@ -967,7 +967,7 @@ int reiserfs_permission(struct inode *inode, int mask, unsigned int flags) if (IS_PRIVATE(inode)) return 0; - return generic_permission(inode, mask, flags); + return generic_permission(inode, mask); } static int xattr_hide_revalidate(struct dentry *dentry, struct nameidata *nd) diff --git a/fs/sysfs/inode.c b/fs/sysfs/inode.c index a37165c64757..04c81e5ba6f7 100644 --- a/fs/sysfs/inode.c +++ b/fs/sysfs/inode.c @@ -362,5 +362,5 @@ int sysfs_permission(struct inode *inode, int mask, unsigned int flags) sysfs_refresh_inode(sd, inode); mutex_unlock(&sysfs_mutex); - return generic_permission(inode, mask, flags); + return generic_permission(inode, mask); } diff --git a/include/linux/fs.h b/include/linux/fs.h index f218b42718aa..a1689c13eb77 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -2188,7 +2188,7 @@ extern sector_t bmap(struct inode *, sector_t); #endif extern int notify_change(struct dentry *, struct iattr *); extern int inode_permission(struct inode *, int); -extern int generic_permission(struct inode *, int, unsigned int); +extern int generic_permission(struct inode *, int); static inline bool execute_ok(struct inode *inode) { -- cgit v1.2.3-58-ga151 From 10556cb21a0d0b24d95f00ea6df16f599a3345b2 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 20 Jun 2011 19:28:19 -0400 Subject: ->permission() sanitizing: don't pass flags to ->permission() not used by the instances anymore. Signed-off-by: Al Viro --- Documentation/filesystems/vfs.txt | 4 ++-- fs/afs/internal.h | 2 +- fs/afs/security.c | 4 ++-- fs/bad_inode.c | 2 +- fs/btrfs/inode.c | 2 +- fs/ceph/inode.c | 4 ++-- fs/ceph/super.h | 2 +- fs/cifs/cifsfs.c | 2 +- fs/coda/coda_linux.h | 2 +- fs/coda/dir.c | 4 ++-- fs/coda/pioctl.c | 4 ++-- fs/ecryptfs/inode.c | 4 ++-- fs/fuse/dir.c | 14 +++++++------- fs/gfs2/file.c | 2 +- fs/gfs2/inode.c | 16 ++++++++-------- fs/gfs2/inode.h | 2 +- fs/hostfs/hostfs_kern.c | 4 ++-- fs/namei.c | 4 ++-- fs/nfs/dir.c | 4 ++-- fs/nilfs2/inode.c | 2 +- fs/nilfs2/nilfs.h | 2 +- fs/ocfs2/file.c | 4 ++-- fs/ocfs2/file.h | 2 +- fs/proc/base.c | 2 +- fs/proc/proc_sysctl.c | 2 +- fs/reiserfs/xattr.c | 2 +- fs/sysfs/inode.c | 4 ++-- fs/sysfs/sysfs.h | 2 +- include/linux/fs.h | 2 +- include/linux/nfs_fs.h | 2 +- include/linux/reiserfs_xattr.h | 2 +- 31 files changed, 55 insertions(+), 55 deletions(-) (limited to 'fs/ocfs2') diff --git a/Documentation/filesystems/vfs.txt b/Documentation/filesystems/vfs.txt index 8b4c8e04d879..d56151fe77dc 100644 --- a/Documentation/filesystems/vfs.txt +++ b/Documentation/filesystems/vfs.txt @@ -333,7 +333,7 @@ struct inode_operations { void * (*follow_link) (struct dentry *, struct nameidata *); void (*put_link) (struct dentry *, struct nameidata *, void *); void (*truncate) (struct inode *); - int (*permission) (struct inode *, int, unsigned int); + int (*permission) (struct inode *, int); int (*check_acl)(struct inode *, int); int (*setattr) (struct dentry *, struct iattr *); int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *); @@ -423,7 +423,7 @@ otherwise noted. permission: called by the VFS to check for access rights on a POSIX-like filesystem. - May be called in rcu-walk mode (flags & IPERM_FLAG_RCU). If in rcu-walk + May be called in rcu-walk mode (mask & MAY_NOT_BLOCK). If in rcu-walk mode, the filesystem must check the permission without blocking or storing to the inode. diff --git a/fs/afs/internal.h b/fs/afs/internal.h index 5a9b6843bac1..f396d337b817 100644 --- a/fs/afs/internal.h +++ b/fs/afs/internal.h @@ -627,7 +627,7 @@ extern void afs_clear_permits(struct afs_vnode *); extern void afs_cache_permit(struct afs_vnode *, struct key *, long); extern void afs_zap_permits(struct rcu_head *); extern struct key *afs_request_key(struct afs_cell *); -extern int afs_permission(struct inode *, int, unsigned int); +extern int afs_permission(struct inode *, int); /* * server.c diff --git a/fs/afs/security.c b/fs/afs/security.c index ab6b3147f05f..8d010422dc89 100644 --- a/fs/afs/security.c +++ b/fs/afs/security.c @@ -285,14 +285,14 @@ static int afs_check_permit(struct afs_vnode *vnode, struct key *key, * - AFS ACLs are attached to directories only, and a file is controlled by its * parent directory's ACL */ -int afs_permission(struct inode *inode, int mask, unsigned int flags) +int afs_permission(struct inode *inode, int mask) { struct afs_vnode *vnode = AFS_FS_I(inode); afs_access_t uninitialized_var(access); struct key *key; int ret; - if (flags & IPERM_FLAG_RCU) + if (mask & MAY_NOT_BLOCK) return -ECHILD; _enter("{{%x:%u},%lx},%x,", diff --git a/fs/bad_inode.c b/fs/bad_inode.c index bfcb18feb1df..f024d8aaddef 100644 --- a/fs/bad_inode.c +++ b/fs/bad_inode.c @@ -229,7 +229,7 @@ static int bad_inode_readlink(struct dentry *dentry, char __user *buffer, return -EIO; } -static int bad_inode_permission(struct inode *inode, int mask, unsigned int flags) +static int bad_inode_permission(struct inode *inode, int mask) { return -EIO; } diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index b61b0477a8ac..cb170ca51c64 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -7331,7 +7331,7 @@ static int btrfs_set_page_dirty(struct page *page) return __set_page_dirty_nobuffers(page); } -static int btrfs_permission(struct inode *inode, int mask, unsigned int flags) +static int btrfs_permission(struct inode *inode, int mask) { struct btrfs_root *root = BTRFS_I(inode)->root; diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index 9c169741b416..dfb2831d8d85 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -1795,11 +1795,11 @@ int ceph_do_getattr(struct inode *inode, int mask) * Check inode permissions. We verify we have a valid value for * the AUTH cap, then call the generic handler. */ -int ceph_permission(struct inode *inode, int mask, unsigned int flags) +int ceph_permission(struct inode *inode, int mask) { int err; - if (flags & IPERM_FLAG_RCU) + if (mask & MAY_NOT_BLOCK) return -ECHILD; err = ceph_do_getattr(inode, CEPH_CAP_AUTH_SHARED); diff --git a/fs/ceph/super.h b/fs/ceph/super.h index f5cabefa98dc..56c41ef47cad 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -692,7 +692,7 @@ extern void ceph_queue_invalidate(struct inode *inode); extern void ceph_queue_writeback(struct inode *inode); extern int ceph_do_getattr(struct inode *inode, int mask); -extern int ceph_permission(struct inode *inode, int mask, unsigned int flags); +extern int ceph_permission(struct inode *inode, int mask); extern int ceph_setattr(struct dentry *dentry, struct iattr *attr); extern int ceph_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat); diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c index b285b5bcf711..cbbb55e781d3 100644 --- a/fs/cifs/cifsfs.c +++ b/fs/cifs/cifsfs.c @@ -224,7 +224,7 @@ cifs_statfs(struct dentry *dentry, struct kstatfs *buf) return 0; } -static int cifs_permission(struct inode *inode, int mask, unsigned int flags) +static int cifs_permission(struct inode *inode, int mask) { struct cifs_sb_info *cifs_sb; diff --git a/fs/coda/coda_linux.h b/fs/coda/coda_linux.h index 9b0c5323890b..44e17e9c21ae 100644 --- a/fs/coda/coda_linux.h +++ b/fs/coda/coda_linux.h @@ -39,7 +39,7 @@ extern const struct file_operations coda_ioctl_operations; /* operations shared over more than one file */ int coda_open(struct inode *i, struct file *f); int coda_release(struct inode *i, struct file *f); -int coda_permission(struct inode *inode, int mask, unsigned int flags); +int coda_permission(struct inode *inode, int mask); int coda_revalidate_inode(struct dentry *); int coda_getattr(struct vfsmount *, struct dentry *, struct kstat *); int coda_setattr(struct dentry *, struct iattr *); diff --git a/fs/coda/dir.c b/fs/coda/dir.c index 2b8dae4d121e..cd5532398c15 100644 --- a/fs/coda/dir.c +++ b/fs/coda/dir.c @@ -132,11 +132,11 @@ exit: } -int coda_permission(struct inode *inode, int mask, unsigned int flags) +int coda_permission(struct inode *inode, int mask) { int error; - if (flags & IPERM_FLAG_RCU) + if (mask & MAY_NOT_BLOCK) return -ECHILD; mask &= MAY_READ | MAY_WRITE | MAY_EXEC; diff --git a/fs/coda/pioctl.c b/fs/coda/pioctl.c index cb140ef293e4..ee0981f1375b 100644 --- a/fs/coda/pioctl.c +++ b/fs/coda/pioctl.c @@ -24,7 +24,7 @@ #include "coda_linux.h" /* pioctl ops */ -static int coda_ioctl_permission(struct inode *inode, int mask, unsigned int flags); +static int coda_ioctl_permission(struct inode *inode, int mask); static long coda_pioctl(struct file *filp, unsigned int cmd, unsigned long user_data); @@ -41,7 +41,7 @@ const struct file_operations coda_ioctl_operations = { }; /* the coda pioctl inode ops */ -static int coda_ioctl_permission(struct inode *inode, int mask, unsigned int flags) +static int coda_ioctl_permission(struct inode *inode, int mask) { return (mask & MAY_EXEC) ? -EACCES : 0; } diff --git a/fs/ecryptfs/inode.c b/fs/ecryptfs/inode.c index 7349ade17de6..bec75f8e91ac 100644 --- a/fs/ecryptfs/inode.c +++ b/fs/ecryptfs/inode.c @@ -942,9 +942,9 @@ int ecryptfs_truncate(struct dentry *dentry, loff_t new_length) } static int -ecryptfs_permission(struct inode *inode, int mask, unsigned int flags) +ecryptfs_permission(struct inode *inode, int mask) { - if (flags & IPERM_FLAG_RCU) + if (mask & MAY_NOT_BLOCK) return -ECHILD; return inode_permission(ecryptfs_inode_to_lower(inode), mask); } diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index 0df56b6c26e2..e2b14001cea5 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -971,9 +971,9 @@ static int fuse_access(struct inode *inode, int mask) return err; } -static int fuse_perm_getattr(struct inode *inode, int flags) +static int fuse_perm_getattr(struct inode *inode, int mask) { - if (flags & IPERM_FLAG_RCU) + if (mask & MAY_NOT_BLOCK) return -ECHILD; return fuse_do_getattr(inode, NULL, NULL); @@ -992,7 +992,7 @@ static int fuse_perm_getattr(struct inode *inode, int flags) * access request is sent. Execute permission is still checked * locally based on file mode. */ -static int fuse_permission(struct inode *inode, int mask, unsigned int flags) +static int fuse_permission(struct inode *inode, int mask) { struct fuse_conn *fc = get_fuse_conn(inode); bool refreshed = false; @@ -1011,7 +1011,7 @@ static int fuse_permission(struct inode *inode, int mask, unsigned int flags) if (fi->i_time < get_jiffies_64()) { refreshed = true; - err = fuse_perm_getattr(inode, flags); + err = fuse_perm_getattr(inode, mask); if (err) return err; } @@ -1024,7 +1024,7 @@ static int fuse_permission(struct inode *inode, int mask, unsigned int flags) attributes. This is also needed, because the root node will at first have no permissions */ if (err == -EACCES && !refreshed) { - err = fuse_perm_getattr(inode, flags); + err = fuse_perm_getattr(inode, mask); if (!err) err = generic_permission(inode, mask); } @@ -1034,7 +1034,7 @@ static int fuse_permission(struct inode *inode, int mask, unsigned int flags) noticed immediately, only after the attribute timeout has expired */ } else if (mask & (MAY_ACCESS | MAY_CHDIR)) { - if (flags & IPERM_FLAG_RCU) + if (mask & MAY_NOT_BLOCK) return -ECHILD; err = fuse_access(inode, mask); @@ -1043,7 +1043,7 @@ static int fuse_permission(struct inode *inode, int mask, unsigned int flags) if (refreshed) return -EACCES; - err = fuse_perm_getattr(inode, flags); + err = fuse_perm_getattr(inode, mask); if (!err && !(inode->i_mode & S_IXUGO)) return -EACCES; } diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c index a9f5cbe45cd9..89c39e53760d 100644 --- a/fs/gfs2/file.c +++ b/fs/gfs2/file.c @@ -243,7 +243,7 @@ static int do_gfs2_set_flags(struct file *filp, u32 reqflags, u32 mask) !capable(CAP_LINUX_IMMUTABLE)) goto out; if (!IS_IMMUTABLE(inode)) { - error = gfs2_permission(inode, MAY_WRITE, 0); + error = gfs2_permission(inode, MAY_WRITE); if (error) goto out; } diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c index b776ec8f9c19..b1090d66a6fd 100644 --- a/fs/gfs2/inode.c +++ b/fs/gfs2/inode.c @@ -307,7 +307,7 @@ struct inode *gfs2_lookupi(struct inode *dir, const struct qstr *name, } if (!is_root) { - error = gfs2_permission(dir, MAY_EXEC, 0); + error = gfs2_permission(dir, MAY_EXEC); if (error) goto out; } @@ -337,7 +337,7 @@ static int create_ok(struct gfs2_inode *dip, const struct qstr *name, { int error; - error = gfs2_permission(&dip->i_inode, MAY_WRITE | MAY_EXEC, 0); + error = gfs2_permission(&dip->i_inode, MAY_WRITE | MAY_EXEC); if (error) return error; @@ -857,7 +857,7 @@ static int gfs2_link(struct dentry *old_dentry, struct inode *dir, if (inode->i_nlink == 0) goto out_gunlock; - error = gfs2_permission(dir, MAY_WRITE | MAY_EXEC, 0); + error = gfs2_permission(dir, MAY_WRITE | MAY_EXEC); if (error) goto out_gunlock; @@ -990,7 +990,7 @@ static int gfs2_unlink_ok(struct gfs2_inode *dip, const struct qstr *name, if (IS_APPEND(&dip->i_inode)) return -EPERM; - error = gfs2_permission(&dip->i_inode, MAY_WRITE | MAY_EXEC, 0); + error = gfs2_permission(&dip->i_inode, MAY_WRITE | MAY_EXEC); if (error) return error; @@ -1336,7 +1336,7 @@ static int gfs2_rename(struct inode *odir, struct dentry *odentry, } } } else { - error = gfs2_permission(ndir, MAY_WRITE | MAY_EXEC, 0); + error = gfs2_permission(ndir, MAY_WRITE | MAY_EXEC); if (error) goto out_gunlock; @@ -1371,7 +1371,7 @@ static int gfs2_rename(struct inode *odir, struct dentry *odentry, /* Check out the dir to be renamed */ if (dir_rename) { - error = gfs2_permission(odentry->d_inode, MAY_WRITE, 0); + error = gfs2_permission(odentry->d_inode, MAY_WRITE); if (error) goto out_gunlock; } @@ -1543,7 +1543,7 @@ static void gfs2_put_link(struct dentry *dentry, struct nameidata *nd, void *p) * Returns: errno */ -int gfs2_permission(struct inode *inode, int mask, unsigned int flags) +int gfs2_permission(struct inode *inode, int mask) { struct gfs2_inode *ip; struct gfs2_holder i_gh; @@ -1553,7 +1553,7 @@ int gfs2_permission(struct inode *inode, int mask, unsigned int flags) ip = GFS2_I(inode); if (gfs2_glock_is_locked_by_me(ip->i_gl) == NULL) { - if (flags & IPERM_FLAG_RCU) + if (mask & MAY_NOT_BLOCK) return -ECHILD; error = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, LM_FLAG_ANY, &i_gh); if (error) diff --git a/fs/gfs2/inode.h b/fs/gfs2/inode.h index 31606076f701..8d90e0c07672 100644 --- a/fs/gfs2/inode.h +++ b/fs/gfs2/inode.h @@ -108,7 +108,7 @@ extern int gfs2_inode_refresh(struct gfs2_inode *ip); extern struct inode *gfs2_lookupi(struct inode *dir, const struct qstr *name, int is_root); -extern int gfs2_permission(struct inode *inode, int mask, unsigned int flags); +extern int gfs2_permission(struct inode *inode, int mask); extern int gfs2_setattr_simple(struct gfs2_inode *ip, struct iattr *attr); extern struct inode *gfs2_lookup_simple(struct inode *dip, const char *name); extern void gfs2_dinode_out(const struct gfs2_inode *ip, void *buf); diff --git a/fs/hostfs/hostfs_kern.c b/fs/hostfs/hostfs_kern.c index b1bc31bde833..6e449c599b9d 100644 --- a/fs/hostfs/hostfs_kern.c +++ b/fs/hostfs/hostfs_kern.c @@ -748,12 +748,12 @@ int hostfs_rename(struct inode *from_ino, struct dentry *from, return err; } -int hostfs_permission(struct inode *ino, int desired, unsigned int flags) +int hostfs_permission(struct inode *ino, int desired) { char *name; int r = 0, w = 0, x = 0, err; - if (flags & IPERM_FLAG_RCU) + if (desired & MAY_NOT_BLOCK) return -ECHILD; if (desired & MAY_READ) r = 1; diff --git a/fs/namei.c b/fs/namei.c index 684e0f30cf4c..c5c382620a86 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -287,7 +287,7 @@ int inode_permission(struct inode *inode, int mask) } if (inode->i_op->permission) - retval = inode->i_op->permission(inode, mask, 0); + retval = inode->i_op->permission(inode, mask); else retval = generic_permission(inode, mask); @@ -323,7 +323,7 @@ static inline int exec_permission(struct inode *inode, unsigned int flags) mask |= MAY_NOT_BLOCK; if (inode->i_op->permission) { - ret = inode->i_op->permission(inode, mask, flags); + ret = inode->i_op->permission(inode, mask); if (likely(!ret)) goto ok; } else { diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index 16cf84b4afb9..a86acd6d0cb4 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -2273,12 +2273,12 @@ int nfs_may_open(struct inode *inode, struct rpc_cred *cred, int openflags) return nfs_do_access(inode, cred, nfs_open_permission_mask(openflags)); } -int nfs_permission(struct inode *inode, int mask, unsigned int flags) +int nfs_permission(struct inode *inode, int mask) { struct rpc_cred *cred; int res = 0; - if (flags & IPERM_FLAG_RCU) + if (mask & MAY_NOT_BLOCK) return -ECHILD; nfs_inc_stats(inode, NFSIOS_VFSACCESS); diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c index 0df6de58bbcb..be8664c6a825 100644 --- a/fs/nilfs2/inode.c +++ b/fs/nilfs2/inode.c @@ -799,7 +799,7 @@ out_err: return err; } -int nilfs_permission(struct inode *inode, int mask, unsigned int flags) +int nilfs_permission(struct inode *inode, int mask) { struct nilfs_root *root = NILFS_I(inode)->i_root; if ((mask & MAY_WRITE) && root && diff --git a/fs/nilfs2/nilfs.h b/fs/nilfs2/nilfs.h index f02b9ad43a21..6fb7511c0328 100644 --- a/fs/nilfs2/nilfs.h +++ b/fs/nilfs2/nilfs.h @@ -264,7 +264,7 @@ extern void nilfs_update_inode(struct inode *, struct buffer_head *); extern void nilfs_truncate(struct inode *); extern void nilfs_evict_inode(struct inode *); extern int nilfs_setattr(struct dentry *, struct iattr *); -int nilfs_permission(struct inode *inode, int mask, unsigned int flags); +int nilfs_permission(struct inode *inode, int mask); int nilfs_load_inode_block(struct inode *inode, struct buffer_head **pbh); extern int nilfs_inode_dirty(struct inode *); int nilfs_set_file_dirty(struct inode *inode, unsigned nr_dirty); diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index ecb52b028964..1406c37a5722 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c @@ -1279,11 +1279,11 @@ bail: return err; } -int ocfs2_permission(struct inode *inode, int mask, unsigned int flags) +int ocfs2_permission(struct inode *inode, int mask) { int ret; - if (flags & IPERM_FLAG_RCU) + if (mask & MAY_NOT_BLOCK) return -ECHILD; ret = ocfs2_inode_lock(inode, NULL, 0); diff --git a/fs/ocfs2/file.h b/fs/ocfs2/file.h index f5afbbef6703..97bf761c9e7c 100644 --- a/fs/ocfs2/file.h +++ b/fs/ocfs2/file.h @@ -61,7 +61,7 @@ int ocfs2_zero_extend(struct inode *inode, struct buffer_head *di_bh, int ocfs2_setattr(struct dentry *dentry, struct iattr *attr); int ocfs2_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat); -int ocfs2_permission(struct inode *inode, int mask, unsigned int flags); +int ocfs2_permission(struct inode *inode, int mask); int ocfs2_should_update_atime(struct inode *inode, struct vfsmount *vfsmnt); diff --git a/fs/proc/base.c b/fs/proc/base.c index 53a1a961b25f..be1ff932033b 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -2167,7 +2167,7 @@ static const struct file_operations proc_fd_operations = { * /proc/pid/fd needs a special permission handler so that a process can still * access /proc/self/fd after it has executed a setuid(). */ -static int proc_fd_permission(struct inode *inode, int mask, unsigned int flags) +static int proc_fd_permission(struct inode *inode, int mask) { int rv = generic_permission(inode, mask); if (rv == 0) diff --git a/fs/proc/proc_sysctl.c b/fs/proc/proc_sysctl.c index 349b22f434d8..1a77dbef226f 100644 --- a/fs/proc/proc_sysctl.c +++ b/fs/proc/proc_sysctl.c @@ -294,7 +294,7 @@ out: return ret; } -static int proc_sys_permission(struct inode *inode, int mask,unsigned int flags) +static int proc_sys_permission(struct inode *inode, int mask) { /* * sysctl entries that are not writeable, diff --git a/fs/reiserfs/xattr.c b/fs/reiserfs/xattr.c index f17319613a85..4ea2ab41fdee 100644 --- a/fs/reiserfs/xattr.c +++ b/fs/reiserfs/xattr.c @@ -958,7 +958,7 @@ static int xattr_mount_check(struct super_block *s) return 0; } -int reiserfs_permission(struct inode *inode, int mask, unsigned int flags) +int reiserfs_permission(struct inode *inode, int mask) { /* * We don't do permission checks on the internal objects. diff --git a/fs/sysfs/inode.c b/fs/sysfs/inode.c index 04c81e5ba6f7..e3f091a81c72 100644 --- a/fs/sysfs/inode.c +++ b/fs/sysfs/inode.c @@ -349,11 +349,11 @@ int sysfs_hash_and_remove(struct sysfs_dirent *dir_sd, const void *ns, const cha return -ENOENT; } -int sysfs_permission(struct inode *inode, int mask, unsigned int flags) +int sysfs_permission(struct inode *inode, int mask) { struct sysfs_dirent *sd; - if (flags & IPERM_FLAG_RCU) + if (mask & MAY_NOT_BLOCK) return -ECHILD; sd = inode->i_private; diff --git a/fs/sysfs/sysfs.h b/fs/sysfs/sysfs.h index 2ed2404f3113..845ab3ad229d 100644 --- a/fs/sysfs/sysfs.h +++ b/fs/sysfs/sysfs.h @@ -201,7 +201,7 @@ static inline void __sysfs_put(struct sysfs_dirent *sd) struct inode *sysfs_get_inode(struct super_block *sb, struct sysfs_dirent *sd); void sysfs_evict_inode(struct inode *inode); int sysfs_sd_setattr(struct sysfs_dirent *sd, struct iattr *iattr); -int sysfs_permission(struct inode *inode, int mask, unsigned int flags); +int sysfs_permission(struct inode *inode, int mask); int sysfs_setattr(struct dentry *dentry, struct iattr *iattr); int sysfs_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat); int sysfs_setxattr(struct dentry *dentry, const char *name, const void *value, diff --git a/include/linux/fs.h b/include/linux/fs.h index a1689c13eb77..1fdefe3995fd 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1578,7 +1578,7 @@ struct file_operations { struct inode_operations { struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *); void * (*follow_link) (struct dentry *, struct nameidata *); - int (*permission) (struct inode *, int, unsigned int); + int (*permission) (struct inode *, int); int (*check_acl)(struct inode *, int); int (*readlink) (struct dentry *, char __user *,int); diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index 1b93b9c60e55..86014811a0fb 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -360,7 +360,7 @@ extern int nfs_refresh_inode(struct inode *, struct nfs_fattr *); extern int nfs_post_op_update_inode(struct inode *inode, struct nfs_fattr *fattr); extern int nfs_post_op_update_inode_force_wcc(struct inode *inode, struct nfs_fattr *fattr); extern int nfs_getattr(struct vfsmount *, struct dentry *, struct kstat *); -extern int nfs_permission(struct inode *, int, unsigned int); +extern int nfs_permission(struct inode *, int); extern int nfs_open(struct inode *, struct file *); extern int nfs_release(struct inode *, struct file *); extern int nfs_attribute_timeout(struct inode *inode); diff --git a/include/linux/reiserfs_xattr.h b/include/linux/reiserfs_xattr.h index 424ba6416e7c..57958c0e1d38 100644 --- a/include/linux/reiserfs_xattr.h +++ b/include/linux/reiserfs_xattr.h @@ -41,7 +41,7 @@ int reiserfs_xattr_init(struct super_block *sb, int mount_flags); int reiserfs_lookup_privroot(struct super_block *sb); int reiserfs_delete_xattrs(struct inode *inode); int reiserfs_chown_xattrs(struct inode *inode, struct iattr *attrs); -int reiserfs_permission(struct inode *inode, int mask, unsigned int flags); +int reiserfs_permission(struct inode *inode, int mask); #ifdef CONFIG_REISERFS_FS_XATTR #define has_xattr_dir(inode) (REISERFS_I(inode)->i_flags & i_has_xattr_dir) -- cgit v1.2.3-58-ga151 From dae6ad8f37529963ae7df52baaccf056b38f210e Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 26 Jun 2011 11:50:15 -0400 Subject: new helpers: kern_path_create/user_path_create combination of kern_path_parent() and lookup_create(). Does *not* expose struct nameidata to caller. Syscalls converted to that... Signed-off-by: Al Viro --- fs/namei.c | 154 ++++++++++++++++++++++++------------------------ fs/ocfs2/refcounttree.c | 49 ++++----------- include/linux/namei.h | 2 + net/unix/af_unix.c | 38 ++++++------ 4 files changed, 106 insertions(+), 137 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/namei.c b/fs/namei.c index f49d6abfa799..b292eb03d9d2 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -2311,6 +2311,35 @@ fail: } EXPORT_SYMBOL_GPL(lookup_create); +struct dentry *kern_path_create(int dfd, const char *pathname, struct path *path, int is_dir) +{ + struct nameidata nd; + struct dentry *res; + int error = do_path_lookup(dfd, pathname, LOOKUP_PARENT, &nd); + if (error) + return ERR_PTR(error); + res = lookup_create(&nd, is_dir); + if (IS_ERR(res)) { + mutex_unlock(&nd.path.dentry->d_inode->i_mutex); + path_put(&nd.path); + } + *path = nd.path; + return res; +} +EXPORT_SYMBOL(kern_path_create); + +struct dentry *user_path_create(int dfd, const char __user *pathname, struct path *path, int is_dir) +{ + char *tmp = getname(pathname); + struct dentry *res; + if (IS_ERR(tmp)) + return ERR_CAST(tmp); + res = kern_path_create(dfd, tmp, path, is_dir); + putname(tmp); + return res; +} +EXPORT_SYMBOL(user_path_create); + int vfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t dev) { int error = may_create(dir, dentry); @@ -2359,54 +2388,46 @@ static int may_mknod(mode_t mode) SYSCALL_DEFINE4(mknodat, int, dfd, const char __user *, filename, int, mode, unsigned, dev) { - int error; - char *tmp; struct dentry *dentry; - struct nameidata nd; + struct path path; + int error; if (S_ISDIR(mode)) return -EPERM; - error = user_path_parent(dfd, filename, &nd, &tmp); - if (error) - return error; + dentry = user_path_create(dfd, filename, &path, 0); + if (IS_ERR(dentry)) + return PTR_ERR(dentry); - dentry = lookup_create(&nd, 0); - if (IS_ERR(dentry)) { - error = PTR_ERR(dentry); - goto out_unlock; - } - if (!IS_POSIXACL(nd.path.dentry->d_inode)) + if (!IS_POSIXACL(path.dentry->d_inode)) mode &= ~current_umask(); error = may_mknod(mode); if (error) goto out_dput; - error = mnt_want_write(nd.path.mnt); + error = mnt_want_write(path.mnt); if (error) goto out_dput; - error = security_path_mknod(&nd.path, dentry, mode, dev); + error = security_path_mknod(&path, dentry, mode, dev); if (error) goto out_drop_write; switch (mode & S_IFMT) { case 0: case S_IFREG: - error = vfs_create(nd.path.dentry->d_inode,dentry,mode,NULL); + error = vfs_create(path.dentry->d_inode,dentry,mode,NULL); break; case S_IFCHR: case S_IFBLK: - error = vfs_mknod(nd.path.dentry->d_inode,dentry,mode, + error = vfs_mknod(path.dentry->d_inode,dentry,mode, new_decode_dev(dev)); break; case S_IFIFO: case S_IFSOCK: - error = vfs_mknod(nd.path.dentry->d_inode,dentry,mode,0); + error = vfs_mknod(path.dentry->d_inode,dentry,mode,0); break; } out_drop_write: - mnt_drop_write(nd.path.mnt); + mnt_drop_write(path.mnt); out_dput: dput(dentry); -out_unlock: - mutex_unlock(&nd.path.dentry->d_inode->i_mutex); - path_put(&nd.path); - putname(tmp); + mutex_unlock(&path.dentry->d_inode->i_mutex); + path_put(&path); return error; } @@ -2439,38 +2460,29 @@ int vfs_mkdir(struct inode *dir, struct dentry *dentry, int mode) SYSCALL_DEFINE3(mkdirat, int, dfd, const char __user *, pathname, int, mode) { - int error = 0; - char * tmp; struct dentry *dentry; - struct nameidata nd; - - error = user_path_parent(dfd, pathname, &nd, &tmp); - if (error) - goto out_err; + struct path path; + int error; - dentry = lookup_create(&nd, 1); - error = PTR_ERR(dentry); + dentry = user_path_create(dfd, pathname, &path, 1); if (IS_ERR(dentry)) - goto out_unlock; + return PTR_ERR(dentry); - if (!IS_POSIXACL(nd.path.dentry->d_inode)) + if (!IS_POSIXACL(path.dentry->d_inode)) mode &= ~current_umask(); - error = mnt_want_write(nd.path.mnt); + error = mnt_want_write(path.mnt); if (error) goto out_dput; - error = security_path_mkdir(&nd.path, dentry, mode); + error = security_path_mkdir(&path, dentry, mode); if (error) goto out_drop_write; - error = vfs_mkdir(nd.path.dentry->d_inode, dentry, mode); + error = vfs_mkdir(path.dentry->d_inode, dentry, mode); out_drop_write: - mnt_drop_write(nd.path.mnt); + mnt_drop_write(path.mnt); out_dput: dput(dentry); -out_unlock: - mutex_unlock(&nd.path.dentry->d_inode->i_mutex); - path_put(&nd.path); - putname(tmp); -out_err: + mutex_unlock(&path.dentry->d_inode->i_mutex); + path_put(&path); return error; } @@ -2730,38 +2742,31 @@ SYSCALL_DEFINE3(symlinkat, const char __user *, oldname, { int error; char *from; - char *to; struct dentry *dentry; - struct nameidata nd; + struct path path; from = getname(oldname); if (IS_ERR(from)) return PTR_ERR(from); - error = user_path_parent(newdfd, newname, &nd, &to); - if (error) - goto out_putname; - - dentry = lookup_create(&nd, 0); + dentry = user_path_create(newdfd, newname, &path, 0); error = PTR_ERR(dentry); if (IS_ERR(dentry)) - goto out_unlock; + goto out_putname; - error = mnt_want_write(nd.path.mnt); + error = mnt_want_write(path.mnt); if (error) goto out_dput; - error = security_path_symlink(&nd.path, dentry, from); + error = security_path_symlink(&path, dentry, from); if (error) goto out_drop_write; - error = vfs_symlink(nd.path.dentry->d_inode, dentry, from); + error = vfs_symlink(path.dentry->d_inode, dentry, from); out_drop_write: - mnt_drop_write(nd.path.mnt); + mnt_drop_write(path.mnt); out_dput: dput(dentry); -out_unlock: - mutex_unlock(&nd.path.dentry->d_inode->i_mutex); - path_put(&nd.path); - putname(to); + mutex_unlock(&path.dentry->d_inode->i_mutex); + path_put(&path); out_putname: putname(from); return error; @@ -2826,11 +2831,9 @@ SYSCALL_DEFINE5(linkat, int, olddfd, const char __user *, oldname, int, newdfd, const char __user *, newname, int, flags) { struct dentry *new_dentry; - struct nameidata nd; - struct path old_path; + struct path old_path, new_path; int how = 0; int error; - char *to; if ((flags & ~(AT_SYMLINK_FOLLOW | AT_EMPTY_PATH)) != 0) return -EINVAL; @@ -2852,32 +2855,27 @@ SYSCALL_DEFINE5(linkat, int, olddfd, const char __user *, oldname, if (error) return error; - error = user_path_parent(newdfd, newname, &nd, &to); - if (error) - goto out; - error = -EXDEV; - if (old_path.mnt != nd.path.mnt) - goto out_release; - new_dentry = lookup_create(&nd, 0); + new_dentry = user_path_create(newdfd, newname, &new_path, 0); error = PTR_ERR(new_dentry); if (IS_ERR(new_dentry)) - goto out_unlock; - error = mnt_want_write(nd.path.mnt); + goto out; + + error = -EXDEV; + if (old_path.mnt != new_path.mnt) + goto out_dput; + error = mnt_want_write(new_path.mnt); if (error) goto out_dput; - error = security_path_link(old_path.dentry, &nd.path, new_dentry); + error = security_path_link(old_path.dentry, &new_path, new_dentry); if (error) goto out_drop_write; - error = vfs_link(old_path.dentry, nd.path.dentry->d_inode, new_dentry); + error = vfs_link(old_path.dentry, new_path.dentry->d_inode, new_dentry); out_drop_write: - mnt_drop_write(nd.path.mnt); + mnt_drop_write(new_path.mnt); out_dput: dput(new_dentry); -out_unlock: - mutex_unlock(&nd.path.dentry->d_inode->i_mutex); -out_release: - path_put(&nd.path); - putname(to); + mutex_unlock(&new_path.dentry->d_inode->i_mutex); + path_put(&new_path); out: path_put(&old_path); diff --git a/fs/ocfs2/refcounttree.c b/fs/ocfs2/refcounttree.c index ebfd3825f12a..cf7823382664 100644 --- a/fs/ocfs2/refcounttree.c +++ b/fs/ocfs2/refcounttree.c @@ -4368,25 +4368,6 @@ static inline int ocfs2_may_create(struct inode *dir, struct dentry *child) return inode_permission(dir, MAY_WRITE | MAY_EXEC); } -/* copied from user_path_parent. */ -static int ocfs2_user_path_parent(const char __user *path, - struct nameidata *nd, char **name) -{ - char *s = getname(path); - int error; - - if (IS_ERR(s)) - return PTR_ERR(s); - - error = kern_path_parent(s, nd); - if (error) - putname(s); - else - *name = s; - - return error; -} - /** * ocfs2_vfs_reflink - Create a reference-counted link * @@ -4460,10 +4441,8 @@ int ocfs2_reflink_ioctl(struct inode *inode, bool preserve) { struct dentry *new_dentry; - struct nameidata nd; - struct path old_path; + struct path old_path, new_path; int error; - char *to = NULL; if (!ocfs2_refcount_tree(OCFS2_SB(inode->i_sb))) return -EOPNOTSUPP; @@ -4474,39 +4453,33 @@ int ocfs2_reflink_ioctl(struct inode *inode, return error; } - error = ocfs2_user_path_parent(newname, &nd, &to); - if (error) { + new_dentry = user_path_create(AT_FDCWD, newname, &new_path, 0); + error = PTR_ERR(new_dentry); + if (IS_ERR(new_dentry)) { mlog_errno(error); goto out; } error = -EXDEV; - if (old_path.mnt != nd.path.mnt) - goto out_release; - new_dentry = lookup_create(&nd, 0); - error = PTR_ERR(new_dentry); - if (IS_ERR(new_dentry)) { + if (old_path.mnt != new_path.mnt) { mlog_errno(error); - goto out_unlock; + goto out_dput; } - error = mnt_want_write(nd.path.mnt); + error = mnt_want_write(new_path.mnt); if (error) { mlog_errno(error); goto out_dput; } error = ocfs2_vfs_reflink(old_path.dentry, - nd.path.dentry->d_inode, + new_path.dentry->d_inode, new_dentry, preserve); - mnt_drop_write(nd.path.mnt); + mnt_drop_write(new_path.mnt); out_dput: dput(new_dentry); -out_unlock: - mutex_unlock(&nd.path.dentry->d_inode->i_mutex); -out_release: - path_put(&nd.path); - putname(to); + mutex_unlock(&new_path.dentry->d_inode->i_mutex); + path_put(&new_path); out: path_put(&old_path); diff --git a/include/linux/namei.h b/include/linux/namei.h index 3439ab862e1d..b8cea804d31a 100644 --- a/include/linux/namei.h +++ b/include/linux/namei.h @@ -74,6 +74,8 @@ extern int user_path_at(int, const char __user *, unsigned, struct path *); extern int kern_path(const char *, unsigned, struct path *); +extern struct dentry *kern_path_create(int, const char *, struct path *, int); +extern struct dentry *user_path_create(int, const char __user *, struct path *, int); extern int kern_path_parent(const char *, struct nameidata *); extern int vfs_path_lookup(struct dentry *, struct vfsmount *, const char *, unsigned int, struct nameidata *); diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c index 0722a25a3a33..ec68e1c05b85 100644 --- a/net/unix/af_unix.c +++ b/net/unix/af_unix.c @@ -808,8 +808,9 @@ static int unix_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len) struct net *net = sock_net(sk); struct unix_sock *u = unix_sk(sk); struct sockaddr_un *sunaddr = (struct sockaddr_un *)uaddr; + char *sun_path = sunaddr->sun_path; struct dentry *dentry = NULL; - struct nameidata nd; + struct path path; int err; unsigned hash; struct unix_address *addr; @@ -845,48 +846,44 @@ static int unix_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len) addr->hash = hash ^ sk->sk_type; atomic_set(&addr->refcnt, 1); - if (sunaddr->sun_path[0]) { + if (sun_path[0]) { unsigned int mode; err = 0; /* * Get the parent directory, calculate the hash for last * component. */ - err = kern_path_parent(sunaddr->sun_path, &nd); - if (err) - goto out_mknod_parent; - - dentry = lookup_create(&nd, 0); + dentry = kern_path_create(AT_FDCWD, sun_path, &path, 0); err = PTR_ERR(dentry); if (IS_ERR(dentry)) - goto out_mknod_unlock; + goto out_mknod_parent; /* * All right, let's create it. */ mode = S_IFSOCK | (SOCK_INODE(sock)->i_mode & ~current_umask()); - err = mnt_want_write(nd.path.mnt); + err = mnt_want_write(path.mnt); if (err) goto out_mknod_dput; - err = security_path_mknod(&nd.path, dentry, mode, 0); + err = security_path_mknod(&path, dentry, mode, 0); if (err) goto out_mknod_drop_write; - err = vfs_mknod(nd.path.dentry->d_inode, dentry, mode, 0); + err = vfs_mknod(path.dentry->d_inode, dentry, mode, 0); out_mknod_drop_write: - mnt_drop_write(nd.path.mnt); + mnt_drop_write(path.mnt); if (err) goto out_mknod_dput; - mutex_unlock(&nd.path.dentry->d_inode->i_mutex); - dput(nd.path.dentry); - nd.path.dentry = dentry; + mutex_unlock(&path.dentry->d_inode->i_mutex); + dput(path.dentry); + path.dentry = dentry; addr->hash = UNIX_HASH_SIZE; } spin_lock(&unix_table_lock); - if (!sunaddr->sun_path[0]) { + if (!sun_path[0]) { err = -EADDRINUSE; if (__unix_find_socket_byname(net, sunaddr, addr_len, sk->sk_type, hash)) { @@ -897,8 +894,8 @@ out_mknod_drop_write: list = &unix_socket_table[addr->hash]; } else { list = &unix_socket_table[dentry->d_inode->i_ino & (UNIX_HASH_SIZE-1)]; - u->dentry = nd.path.dentry; - u->mnt = nd.path.mnt; + u->dentry = path.dentry; + u->mnt = path.mnt; } err = 0; @@ -915,9 +912,8 @@ out: out_mknod_dput: dput(dentry); -out_mknod_unlock: - mutex_unlock(&nd.path.dentry->d_inode->i_mutex); - path_put(&nd.path); + mutex_unlock(&path.dentry->d_inode->i_mutex); + path_put(&path); out_mknod_parent: if (err == -EEXIST) err = -EADDRINUSE; -- cgit v1.2.3-58-ga151 From bd5fe6c5eb9c548d7f07fe8f89a150bb6705e8e3 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 24 Jun 2011 14:29:43 -0400 Subject: fs: kill i_alloc_sem i_alloc_sem is a rather special rw_semaphore. It's the last one that may be released by a non-owner, and it's write side is always mirrored by real exclusion. It's intended use it to wait for all pending direct I/O requests to finish before starting a truncate. Replace it with a hand-grown construct: - exclusion for truncates is already guaranteed by i_mutex, so it can simply fall way - the reader side is replaced by an i_dio_count member in struct inode that counts the number of pending direct I/O requests. Truncate can't proceed as long as it's non-zero - when i_dio_count reaches non-zero we wake up a pending truncate using wake_up_bit on a new bit in i_flags - new references to i_dio_count can't appear while we are waiting for it to read zero because the direct I/O count always needs i_mutex (or an equivalent like XFS's i_iolock) for starting a new operation. This scheme is much simpler, and saves the space of a spinlock_t and a struct list_head in struct inode (typically 160 bits on a non-debug 64-bit system). Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/attr.c | 5 +---- fs/direct-io.c | 65 +++++++++++++++++++++++++++++++++++++++++------------ fs/inode.c | 3 +-- fs/ntfs/file.c | 3 +-- fs/ntfs/inode.c | 10 ++------- fs/ocfs2/aops.c | 7 +++--- fs/ocfs2/file.c | 15 ++++++------- fs/reiserfs/xattr.c | 3 +-- include/linux/fs.h | 11 +++++++-- mm/filemap.c | 3 --- mm/madvise.c | 2 +- mm/rmap.c | 1 - mm/truncate.c | 3 +-- 13 files changed, 78 insertions(+), 53 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/attr.c b/fs/attr.c index caf2aa521e2b..f177ac86fa48 100644 --- a/fs/attr.c +++ b/fs/attr.c @@ -233,16 +233,13 @@ int notify_change(struct dentry * dentry, struct iattr * attr) return error; if (ia_valid & ATTR_SIZE) - down_write(&dentry->d_inode->i_alloc_sem); + inode_dio_wait(inode); if (inode->i_op->setattr) error = inode->i_op->setattr(dentry, attr); else error = simple_setattr(dentry, attr); - if (ia_valid & ATTR_SIZE) - up_write(&dentry->d_inode->i_alloc_sem); - if (!error) fsnotify_change(dentry, ia_valid); diff --git a/fs/direct-io.c b/fs/direct-io.c index 98ce3ac0d94b..354cbdbc14bd 100644 --- a/fs/direct-io.c +++ b/fs/direct-io.c @@ -135,6 +135,50 @@ struct dio { struct page *pages[DIO_PAGES]; /* page buffer */ }; +static void __inode_dio_wait(struct inode *inode) +{ + wait_queue_head_t *wq = bit_waitqueue(&inode->i_state, __I_DIO_WAKEUP); + DEFINE_WAIT_BIT(q, &inode->i_state, __I_DIO_WAKEUP); + + do { + prepare_to_wait(wq, &q.wait, TASK_UNINTERRUPTIBLE); + if (atomic_read(&inode->i_dio_count)) + schedule(); + } while (atomic_read(&inode->i_dio_count)); + finish_wait(wq, &q.wait); +} + +/** + * inode_dio_wait - wait for outstanding DIO requests to finish + * @inode: inode to wait for + * + * Waits for all pending direct I/O requests to finish so that we can + * proceed with a truncate or equivalent operation. + * + * Must be called under a lock that serializes taking new references + * to i_dio_count, usually by inode->i_mutex. + */ +void inode_dio_wait(struct inode *inode) +{ + if (atomic_read(&inode->i_dio_count)) + __inode_dio_wait(inode); +} +EXPORT_SYMBOL_GPL(inode_dio_wait); + +/* + * inode_dio_done - signal finish of a direct I/O requests + * @inode: inode the direct I/O happens on + * + * This is called once we've finished processing a direct I/O request, + * and is used to wake up callers waiting for direct I/O to be quiesced. + */ +void inode_dio_done(struct inode *inode) +{ + if (atomic_dec_and_test(&inode->i_dio_count)) + wake_up_bit(&inode->i_state, __I_DIO_WAKEUP); +} +EXPORT_SYMBOL_GPL(inode_dio_done); + /* * How many pages are in the queue? */ @@ -254,9 +298,7 @@ static ssize_t dio_complete(struct dio *dio, loff_t offset, ssize_t ret, bool is } if (dio->flags & DIO_LOCKING) - /* lockdep: non-owner release */ - up_read_non_owner(&dio->inode->i_alloc_sem); - + inode_dio_done(dio->inode); return ret; } @@ -980,9 +1022,6 @@ out: return ret; } -/* - * Releases both i_mutex and i_alloc_sem - */ static ssize_t direct_io_worker(int rw, struct kiocb *iocb, struct inode *inode, const struct iovec *iov, loff_t offset, unsigned long nr_segs, @@ -1146,15 +1185,14 @@ direct_io_worker(int rw, struct kiocb *iocb, struct inode *inode, * For writes this function is called under i_mutex and returns with * i_mutex held, for reads, i_mutex is not held on entry, but it is * taken and dropped again before returning. - * For reads and writes i_alloc_sem is taken in shared mode and released - * on I/O completion (which may happen asynchronously after returning to - * the caller). + * The i_dio_count counter keeps track of the number of outstanding + * direct I/O requests, and truncate waits for it to reach zero. + * New references to i_dio_count must only be grabbed with i_mutex + * held. * * - if the flags value does NOT contain DIO_LOCKING we don't use any * internal locking but rather rely on the filesystem to synchronize * direct I/O reads/writes versus each other and truncate. - * For reads and writes both i_mutex and i_alloc_sem are not held on - * entry and are never taken. */ ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode, @@ -1234,10 +1272,9 @@ __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode, } /* - * Will be released at I/O completion, possibly in a - * different thread. + * Will be decremented at I/O completion time. */ - down_read_non_owner(&inode->i_alloc_sem); + atomic_inc(&inode->i_dio_count); } /* diff --git a/fs/inode.c b/fs/inode.c index cf81baf1898a..96c77b81167c 100644 --- a/fs/inode.c +++ b/fs/inode.c @@ -168,8 +168,7 @@ int inode_init_always(struct super_block *sb, struct inode *inode) mutex_init(&inode->i_mutex); lockdep_set_class(&inode->i_mutex, &sb->s_type->i_mutex_key); - init_rwsem(&inode->i_alloc_sem); - lockdep_set_class(&inode->i_alloc_sem, &sb->s_type->i_alloc_sem_key); + atomic_set(&inode->i_dio_count, 0); mapping->a_ops = &empty_aops; mapping->host = inode; diff --git a/fs/ntfs/file.c b/fs/ntfs/file.c index f4b1057abdd2..b59f5ac26bef 100644 --- a/fs/ntfs/file.c +++ b/fs/ntfs/file.c @@ -1832,9 +1832,8 @@ static ssize_t ntfs_file_buffered_write(struct kiocb *iocb, * fails again. */ if (unlikely(NInoTruncateFailed(ni))) { - down_write(&vi->i_alloc_sem); + inode_dio_wait(vi); err = ntfs_truncate(vi); - up_write(&vi->i_alloc_sem); if (err || NInoTruncateFailed(ni)) { if (!err) err = -EIO; diff --git a/fs/ntfs/inode.c b/fs/ntfs/inode.c index c05d6dcf77a4..1371487da955 100644 --- a/fs/ntfs/inode.c +++ b/fs/ntfs/inode.c @@ -2357,12 +2357,7 @@ static const char *es = " Leaving inconsistent metadata. Unmount and run " * * Returns 0 on success or -errno on error. * - * Called with ->i_mutex held. In all but one case ->i_alloc_sem is held for - * writing. The only case in the kernel where ->i_alloc_sem is not held is - * mm/filemap.c::generic_file_buffered_write() where vmtruncate() is called - * with the current i_size as the offset. The analogous place in NTFS is in - * fs/ntfs/file.c::ntfs_file_buffered_write() where we call vmtruncate() again - * without holding ->i_alloc_sem. + * Called with ->i_mutex held. */ int ntfs_truncate(struct inode *vi) { @@ -2887,8 +2882,7 @@ void ntfs_truncate_vfs(struct inode *vi) { * We also abort all changes of user, group, and mode as we do not implement * the NTFS ACLs yet. * - * Called with ->i_mutex held. For the ATTR_SIZE (i.e. ->truncate) case, also - * called with ->i_alloc_sem held for writing. + * Called with ->i_mutex held. */ int ntfs_setattr(struct dentry *dentry, struct iattr *attr) { diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c index ac97bca282d2..de1d3953599d 100644 --- a/fs/ocfs2/aops.c +++ b/fs/ocfs2/aops.c @@ -551,9 +551,8 @@ bail: /* * ocfs2_dio_end_io is called by the dio core when a dio is finished. We're - * particularly interested in the aio/dio case. Like the core uses - * i_alloc_sem, we use the rw_lock DLM lock to protect io on one node from - * truncation on another. + * particularly interested in the aio/dio case. We use the rw_lock DLM lock + * to protect io on one node from truncation on another. */ static void ocfs2_dio_end_io(struct kiocb *iocb, loff_t offset, @@ -569,7 +568,7 @@ static void ocfs2_dio_end_io(struct kiocb *iocb, BUG_ON(!ocfs2_iocb_is_rw_locked(iocb)); if (ocfs2_iocb_is_sem_locked(iocb)) { - up_read(&inode->i_alloc_sem); + inode_dio_done(inode); ocfs2_iocb_clear_sem_locked(iocb); } diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index 1406c37a5722..2c3a465514a2 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c @@ -2236,9 +2236,9 @@ static ssize_t ocfs2_file_aio_write(struct kiocb *iocb, ocfs2_iocb_clear_sem_locked(iocb); relock: - /* to match setattr's i_mutex -> i_alloc_sem -> rw_lock ordering */ + /* to match setattr's i_mutex -> rw_lock ordering */ if (direct_io) { - down_read(&inode->i_alloc_sem); + atomic_inc(&inode->i_dio_count); have_alloc_sem = 1; /* communicate with ocfs2_dio_end_io */ ocfs2_iocb_set_sem_locked(iocb); @@ -2290,7 +2290,7 @@ relock: */ if (direct_io && !can_do_direct) { ocfs2_rw_unlock(inode, rw_level); - up_read(&inode->i_alloc_sem); + inode_dio_done(inode); have_alloc_sem = 0; rw_level = -1; @@ -2361,8 +2361,7 @@ out_dio: /* * deep in g_f_a_w_n()->ocfs2_direct_IO we pass in a ocfs2_dio_end_io * function pointer which is called when o_direct io completes so that - * it can unlock our rw lock. (it's the clustered equivalent of - * i_alloc_sem; protects truncate from racing with pending ios). + * it can unlock our rw lock. * Unfortunately there are error cases which call end_io and others * that don't. so we don't have to unlock the rw_lock if either an * async dio is going to do it in the future or an end_io after an @@ -2379,7 +2378,7 @@ out: out_sems: if (have_alloc_sem) { - up_read(&inode->i_alloc_sem); + inode_dio_done(inode); ocfs2_iocb_clear_sem_locked(iocb); } @@ -2531,8 +2530,8 @@ static ssize_t ocfs2_file_aio_read(struct kiocb *iocb, * need locks to protect pending reads from racing with truncate. */ if (filp->f_flags & O_DIRECT) { - down_read(&inode->i_alloc_sem); have_alloc_sem = 1; + atomic_inc(&inode->i_dio_count); ocfs2_iocb_set_sem_locked(iocb); ret = ocfs2_rw_lock(inode, 0); @@ -2575,7 +2574,7 @@ static ssize_t ocfs2_file_aio_read(struct kiocb *iocb, bail: if (have_alloc_sem) { - up_read(&inode->i_alloc_sem); + inode_dio_done(inode); ocfs2_iocb_clear_sem_locked(iocb); } if (rw_level != -1) diff --git a/fs/reiserfs/xattr.c b/fs/reiserfs/xattr.c index 4ea2ab41fdee..6938d8c68d6e 100644 --- a/fs/reiserfs/xattr.c +++ b/fs/reiserfs/xattr.c @@ -555,11 +555,10 @@ reiserfs_xattr_set_handle(struct reiserfs_transaction_handle *th, reiserfs_write_unlock(inode->i_sb); mutex_lock_nested(&dentry->d_inode->i_mutex, I_MUTEX_XATTR); - down_write(&dentry->d_inode->i_alloc_sem); + inode_dio_wait(dentry->d_inode); reiserfs_write_lock(inode->i_sb); err = reiserfs_setattr(dentry, &newattrs); - up_write(&dentry->d_inode->i_alloc_sem); mutex_unlock(&dentry->d_inode->i_mutex); } else update_ctime(inode); diff --git a/include/linux/fs.h b/include/linux/fs.h index 1393742bba9b..2fe920774abf 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -779,7 +779,7 @@ struct inode { struct timespec i_ctime; blkcnt_t i_blocks; unsigned short i_bytes; - struct rw_semaphore i_alloc_sem; + atomic_t i_dio_count; const struct file_operations *i_fop; /* former ->i_op->default_file_ops */ struct file_lock *i_flock; struct address_space *i_mapping; @@ -1705,6 +1705,10 @@ struct super_operations { * set during data writeback, and cleared with a wakeup * on the bit address once it is done. * + * I_REFERENCED Marks the inode as recently references on the LRU list. + * + * I_DIO_WAKEUP Never set. Only used as a key for wait_on_bit(). + * * Q: What is the difference between I_WILL_FREE and I_FREEING? */ #define I_DIRTY_SYNC (1 << 0) @@ -1718,6 +1722,8 @@ struct super_operations { #define __I_SYNC 7 #define I_SYNC (1 << __I_SYNC) #define I_REFERENCED (1 << 8) +#define __I_DIO_WAKEUP 9 +#define I_DIO_WAKEUP (1 << I_DIO_WAKEUP) #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES) @@ -1828,7 +1834,6 @@ struct file_system_type { struct lock_class_key i_lock_key; struct lock_class_key i_mutex_key; struct lock_class_key i_mutex_dir_key; - struct lock_class_key i_alloc_sem_key; }; extern struct dentry *mount_ns(struct file_system_type *fs_type, int flags, @@ -2404,6 +2409,8 @@ enum { }; void dio_end_io(struct bio *bio, int error); +void inode_dio_wait(struct inode *inode); +void inode_dio_done(struct inode *inode); ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode, struct block_device *bdev, const struct iovec *iov, loff_t offset, diff --git a/mm/filemap.c b/mm/filemap.c index a8251a8d3457..f820e600f1ad 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -78,9 +78,6 @@ * ->i_mutex (generic_file_buffered_write) * ->mmap_sem (fault_in_pages_readable->do_page_fault) * - * ->i_mutex - * ->i_alloc_sem (various) - * * inode_wb_list_lock * sb_lock (fs/fs-writeback.c) * ->mapping->tree_lock (__sync_single_inode) diff --git a/mm/madvise.c b/mm/madvise.c index 2221491ed503..74bf193eff04 100644 --- a/mm/madvise.c +++ b/mm/madvise.c @@ -218,7 +218,7 @@ static long madvise_remove(struct vm_area_struct *vma, endoff = (loff_t)(end - vma->vm_start - 1) + ((loff_t)vma->vm_pgoff << PAGE_SHIFT); - /* vmtruncate_range needs to take i_mutex and i_alloc_sem */ + /* vmtruncate_range needs to take i_mutex */ up_read(¤t->mm->mmap_sem); error = vmtruncate_range(mapping->host, offset, endoff); down_read(¤t->mm->mmap_sem); diff --git a/mm/rmap.c b/mm/rmap.c index 23295f65ae43..2540a39eea4a 100644 --- a/mm/rmap.c +++ b/mm/rmap.c @@ -21,7 +21,6 @@ * Lock ordering in mm: * * inode->i_mutex (while writing or truncating, not reading or faulting) - * inode->i_alloc_sem (vmtruncate_range) * mm->mmap_sem * page->flags PG_locked (lock_page) * mapping->i_mmap_mutex diff --git a/mm/truncate.c b/mm/truncate.c index e13f22efaad7..003c6c685fc8 100644 --- a/mm/truncate.c +++ b/mm/truncate.c @@ -622,12 +622,11 @@ int vmtruncate_range(struct inode *inode, loff_t offset, loff_t end) return -ENOSYS; mutex_lock(&inode->i_mutex); - down_write(&inode->i_alloc_sem); + inode_dio_wait(inode); unmap_mapping_range(mapping, offset, (end - offset), 1); inode->i_op->truncate_range(inode, offset, end); /* unmap again to remove racily COWed private pages */ unmap_mapping_range(mapping, offset, (end - offset), 1); - up_write(&inode->i_alloc_sem); mutex_unlock(&inode->i_mutex); return 0; -- cgit v1.2.3-58-ga151 From 562c72aa57c36b178eacc3500a0215651eca9429 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 24 Jun 2011 14:29:45 -0400 Subject: fs: move inode_dio_wait calls into ->setattr Let filesystems handle waiting for direct I/O requests themselves instead of doing it beforehand. This means filesystem-specific locks to prevent new dio referenes from appearing can be held. This is important to allow generalizing i_dio_count to non-DIO_LOCKING filesystems. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/attr.c | 3 --- fs/ext2/inode.c | 2 ++ fs/ext3/inode.c | 3 +++ fs/ext4/inode.c | 2 ++ fs/fat/file.c | 2 ++ fs/gfs2/bmap.c | 2 ++ fs/hfs/inode.c | 2 ++ fs/hfsplus/inode.c | 2 ++ fs/jfs/file.c | 2 ++ fs/nilfs2/inode.c | 2 ++ fs/ocfs2/file.c | 2 ++ fs/reiserfs/inode.c | 3 +++ 12 files changed, 24 insertions(+), 3 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/attr.c b/fs/attr.c index f177ac86fa48..538e27959d3f 100644 --- a/fs/attr.c +++ b/fs/attr.c @@ -232,9 +232,6 @@ int notify_change(struct dentry * dentry, struct iattr * attr) if (error) return error; - if (ia_valid & ATTR_SIZE) - inode_dio_wait(inode); - if (inode->i_op->setattr) error = inode->i_op->setattr(dentry, attr); else diff --git a/fs/ext2/inode.c b/fs/ext2/inode.c index 788e09a07f7e..06e7c767ab35 100644 --- a/fs/ext2/inode.c +++ b/fs/ext2/inode.c @@ -1184,6 +1184,8 @@ static int ext2_setsize(struct inode *inode, loff_t newsize) if (IS_APPEND(inode) || IS_IMMUTABLE(inode)) return -EPERM; + inode_dio_wait(inode); + if (mapping_is_xip(inode->i_mapping)) error = xip_truncate_page(inode->i_mapping, newsize); else if (test_opt(inode->i_sb, NOBH)) diff --git a/fs/ext3/inode.c b/fs/ext3/inode.c index 3451d23c3bae..99c28b246b89 100644 --- a/fs/ext3/inode.c +++ b/fs/ext3/inode.c @@ -3216,6 +3216,9 @@ int ext3_setattr(struct dentry *dentry, struct iattr *attr) ext3_journal_stop(handle); } + if (attr->ia_valid & ATTR_SIZE) + inode_dio_wait(inode); + if (S_ISREG(inode->i_mode) && attr->ia_valid & ATTR_SIZE && attr->ia_size < inode->i_size) { handle_t *handle; diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index bd309764557f..9ec0a2ba2502 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -5351,6 +5351,8 @@ int ext4_setattr(struct dentry *dentry, struct iattr *attr) } if (attr->ia_valid & ATTR_SIZE) { + inode_dio_wait(inode); + if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))) { struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb); diff --git a/fs/fat/file.c b/fs/fat/file.c index a4a3a3c06436..e1587c54d3c1 100644 --- a/fs/fat/file.c +++ b/fs/fat/file.c @@ -397,6 +397,8 @@ int fat_setattr(struct dentry *dentry, struct iattr *attr) * sequence. */ if (attr->ia_valid & ATTR_SIZE) { + inode_dio_wait(inode); + if (attr->ia_size > inode->i_size) { error = fat_cont_expand(inode, attr->ia_size); if (error || attr->ia_valid == ATTR_SIZE) diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index e65493a8ac00..2cd0e56b8893 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -1224,6 +1224,8 @@ int gfs2_setattr_size(struct inode *inode, u64 newsize) if (ret) return ret; + inode_dio_wait(inode); + oldsize = inode->i_size; if (newsize >= oldsize) return do_grow(inode, newsize); diff --git a/fs/hfs/inode.c b/fs/hfs/inode.c index fff16c968e67..48d567cc203e 100644 --- a/fs/hfs/inode.c +++ b/fs/hfs/inode.c @@ -615,6 +615,8 @@ int hfs_inode_setattr(struct dentry *dentry, struct iattr * attr) if ((attr->ia_valid & ATTR_SIZE) && attr->ia_size != i_size_read(inode)) { + inode_dio_wait(inode); + error = vmtruncate(inode, attr->ia_size); if (error) return error; diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c index b248a6cfcad9..b0a0a4b621eb 100644 --- a/fs/hfsplus/inode.c +++ b/fs/hfsplus/inode.c @@ -296,6 +296,8 @@ static int hfsplus_setattr(struct dentry *dentry, struct iattr *attr) if ((attr->ia_valid & ATTR_SIZE) && attr->ia_size != i_size_read(inode)) { + inode_dio_wait(inode); + error = vmtruncate(inode, attr->ia_size); if (error) return error; diff --git a/fs/jfs/file.c b/fs/jfs/file.c index 2f3f531f3606..9f32315acef1 100644 --- a/fs/jfs/file.c +++ b/fs/jfs/file.c @@ -110,6 +110,8 @@ int jfs_setattr(struct dentry *dentry, struct iattr *iattr) if ((iattr->ia_valid & ATTR_SIZE) && iattr->ia_size != i_size_read(inode)) { + inode_dio_wait(inode); + rc = vmtruncate(inode, iattr->ia_size); if (rc) return rc; diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c index be8664c6a825..13f113154a9f 100644 --- a/fs/nilfs2/inode.c +++ b/fs/nilfs2/inode.c @@ -778,6 +778,8 @@ int nilfs_setattr(struct dentry *dentry, struct iattr *iattr) if ((iattr->ia_valid & ATTR_SIZE) && iattr->ia_size != i_size_read(inode)) { + inode_dio_wait(inode); + err = vmtruncate(inode, iattr->ia_size); if (unlikely(err)) goto out_err; diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index 2c3a465514a2..736283ca4a4c 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c @@ -1142,6 +1142,8 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr) if (status) goto bail_unlock; + inode_dio_wait(inode); + if (i_size_read(inode) > attr->ia_size) { if (ocfs2_should_order_data(inode)) { status = ocfs2_begin_ordered_truncate(inode, diff --git a/fs/reiserfs/inode.c b/fs/reiserfs/inode.c index 4fd5bb33dbb5..dcf543d8caf1 100644 --- a/fs/reiserfs/inode.c +++ b/fs/reiserfs/inode.c @@ -3114,6 +3114,9 @@ int reiserfs_setattr(struct dentry *dentry, struct iattr *attr) error = -EFBIG; goto out; } + + inode_dio_wait(inode); + /* fill in hole pointers in the expanding truncate case. */ if (attr->ia_size > inode->i_size) { error = generic_cont_expand_simple(inode, attr->ia_size); -- cgit v1.2.3-58-ga151 From df2d6f26586f12a24f3ae5df4e236dc5c08d6eb4 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 24 Jun 2011 14:29:46 -0400 Subject: fs: always maintain i_dio_count Maintain i_dio_count for all filesystems, not just those using DIO_LOCKING. This these filesystems to also protect truncate against direct I/O requests by using common code. Right now the only non-DIO_LOCKING filesystem that appears to do so is XFS, which uses an opencoded variant of the i_dio_count scheme. Behaviour doesn't change for filesystems never calling inode_dio_wait. For ext4 behaviour changes when using the dioread_nonlock option, which previously was missing any protection between truncate and direct I/O reads. For ocfs2 that handcrafted i_dio_count manipulations are replaced with the common code now enable. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/direct-io.c | 25 +++++++++++++------------ fs/ocfs2/aops.c | 4 +--- fs/ocfs2/file.c | 12 +++--------- 3 files changed, 17 insertions(+), 24 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/direct-io.c b/fs/direct-io.c index 354cbdbc14bd..0a073c7125a6 100644 --- a/fs/direct-io.c +++ b/fs/direct-io.c @@ -297,8 +297,7 @@ static ssize_t dio_complete(struct dio *dio, loff_t offset, ssize_t ret, bool is aio_complete(dio->iocb, ret, 0); } - if (dio->flags & DIO_LOCKING) - inode_dio_done(dio->inode); + inode_dio_done(dio->inode); return ret; } @@ -1185,14 +1184,16 @@ direct_io_worker(int rw, struct kiocb *iocb, struct inode *inode, * For writes this function is called under i_mutex and returns with * i_mutex held, for reads, i_mutex is not held on entry, but it is * taken and dropped again before returning. - * The i_dio_count counter keeps track of the number of outstanding - * direct I/O requests, and truncate waits for it to reach zero. - * New references to i_dio_count must only be grabbed with i_mutex - * held. - * * - if the flags value does NOT contain DIO_LOCKING we don't use any * internal locking but rather rely on the filesystem to synchronize * direct I/O reads/writes versus each other and truncate. + * + * To help with locking against truncate we incremented the i_dio_count + * counter before starting direct I/O, and decrement it once we are done. + * Truncate can wait for it to reach zero to provide exclusion. It is + * expected that filesystem provide exclusion between new direct I/O + * and truncates. For DIO_LOCKING filesystems this is done by i_mutex, + * but other filesystems need to take care of this on their own. */ ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode, @@ -1270,13 +1271,13 @@ __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode, goto out; } } - - /* - * Will be decremented at I/O completion time. - */ - atomic_inc(&inode->i_dio_count); } + /* + * Will be decremented at I/O completion time. + */ + atomic_inc(&inode->i_dio_count); + /* * For file extending writes updating i_size before data * writeouts complete can expose uninitialized blocks. So diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c index de1d3953599d..524d6167fb63 100644 --- a/fs/ocfs2/aops.c +++ b/fs/ocfs2/aops.c @@ -567,10 +567,8 @@ static void ocfs2_dio_end_io(struct kiocb *iocb, /* this io's submitter should not have unlocked this before we could */ BUG_ON(!ocfs2_iocb_is_rw_locked(iocb)); - if (ocfs2_iocb_is_sem_locked(iocb)) { - inode_dio_done(inode); + if (ocfs2_iocb_is_sem_locked(iocb)) ocfs2_iocb_clear_sem_locked(iocb); - } ocfs2_iocb_clear_rw_locked(iocb); diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index 736283ca4a4c..22d604601957 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c @@ -2240,7 +2240,6 @@ static ssize_t ocfs2_file_aio_write(struct kiocb *iocb, relock: /* to match setattr's i_mutex -> rw_lock ordering */ if (direct_io) { - atomic_inc(&inode->i_dio_count); have_alloc_sem = 1; /* communicate with ocfs2_dio_end_io */ ocfs2_iocb_set_sem_locked(iocb); @@ -2292,7 +2291,6 @@ relock: */ if (direct_io && !can_do_direct) { ocfs2_rw_unlock(inode, rw_level); - inode_dio_done(inode); have_alloc_sem = 0; rw_level = -1; @@ -2379,10 +2377,8 @@ out: ocfs2_rw_unlock(inode, rw_level); out_sems: - if (have_alloc_sem) { - inode_dio_done(inode); + if (have_alloc_sem) ocfs2_iocb_clear_sem_locked(iocb); - } mutex_unlock(&inode->i_mutex); @@ -2533,7 +2529,6 @@ static ssize_t ocfs2_file_aio_read(struct kiocb *iocb, */ if (filp->f_flags & O_DIRECT) { have_alloc_sem = 1; - atomic_inc(&inode->i_dio_count); ocfs2_iocb_set_sem_locked(iocb); ret = ocfs2_rw_lock(inode, 0); @@ -2575,10 +2570,9 @@ static ssize_t ocfs2_file_aio_read(struct kiocb *iocb, } bail: - if (have_alloc_sem) { - inode_dio_done(inode); + if (have_alloc_sem) ocfs2_iocb_clear_sem_locked(iocb); - } + if (rw_level != -1) ocfs2_rw_unlock(inode, rw_level); -- cgit v1.2.3-58-ga151 From 72c5052ddc3956d847f21c2b8d55c93664a51b2c Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 24 Jun 2011 14:29:48 -0400 Subject: fs: move inode_dio_done to the end_io handler For filesystems that delay their end_io processing we should keep our i_dio_count until the the processing is done. Enable this by moving the inode_dio_done call to the end_io handler if one exist. Note that the actual move to the workqueue for ext4 and XFS is not done in this patch yet, but left to the filesystem maintainers. At least for XFS it's not needed yet either as XFS has an internal equivalent to i_dio_count. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/direct-io.c | 7 ++++--- fs/ext4/inode.c | 5 +++++ fs/ocfs2/aops.c | 1 + fs/xfs/linux-2.6/xfs_aops.c | 3 +++ 4 files changed, 13 insertions(+), 3 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/direct-io.c b/fs/direct-io.c index 0a073c7125a6..01d2d9ef609c 100644 --- a/fs/direct-io.c +++ b/fs/direct-io.c @@ -293,11 +293,12 @@ static ssize_t dio_complete(struct dio *dio, loff_t offset, ssize_t ret, bool is if (dio->end_io && dio->result) { dio->end_io(dio->iocb, offset, transferred, dio->map_bh.b_private, ret, is_async); - } else if (is_async) { - aio_complete(dio->iocb, ret, 0); + } else { + if (is_async) + aio_complete(dio->iocb, ret, 0); + inode_dio_done(dio->inode); } - inode_dio_done(dio->inode); return ret; } diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 1f35573a34e1..678cde834f19 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -3573,6 +3573,7 @@ static void ext4_end_io_dio(struct kiocb *iocb, loff_t offset, ssize_t size, void *private, int ret, bool is_async) { + struct inode *inode = iocb->ki_filp->f_path.dentry->d_inode; ext4_io_end_t *io_end = iocb->private; struct workqueue_struct *wq; unsigned long flags; @@ -3594,6 +3595,7 @@ static void ext4_end_io_dio(struct kiocb *iocb, loff_t offset, out: if (is_async) aio_complete(iocb, ret, 0); + inode_dio_done(inode); return; } @@ -3614,6 +3616,9 @@ out: /* queue the work to convert unwritten extents to written */ queue_work(wq, &io_end->work); iocb->private = NULL; + + /* XXX: probably should move into the real I/O completion handler */ + inode_dio_done(inode); } static void ext4_end_io_buffer_write(struct buffer_head *bh, int uptodate) diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c index 524d6167fb63..c1efe939c774 100644 --- a/fs/ocfs2/aops.c +++ b/fs/ocfs2/aops.c @@ -577,6 +577,7 @@ static void ocfs2_dio_end_io(struct kiocb *iocb, if (is_async) aio_complete(iocb, ret, 0); + inode_dio_done(inode); } /* diff --git a/fs/xfs/linux-2.6/xfs_aops.c b/fs/xfs/linux-2.6/xfs_aops.c index 79ce38be15a1..b3b418f519f3 100644 --- a/fs/xfs/linux-2.6/xfs_aops.c +++ b/fs/xfs/linux-2.6/xfs_aops.c @@ -1339,6 +1339,9 @@ xfs_end_io_direct_write( } else { xfs_finish_ioend_sync(ioend); } + + /* XXX: probably should move into the real I/O completion handler */ + inode_dio_done(ioend->io_inode); } STATIC ssize_t -- cgit v1.2.3-58-ga151 From 02c24a82187d5a628c68edfe71ae60dc135cd178 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Sat, 16 Jul 2011 20:44:56 -0400 Subject: fs: push i_mutex and filemap_write_and_wait down into ->fsync() handlers Btrfs needs to be able to control how filemap_write_and_wait_range() is called in fsync to make it less of a painful operation, so push down taking i_mutex and the calling of filemap_write_and_wait() down into the ->fsync() handlers. Some file systems can drop taking the i_mutex altogether it seems, like ext3 and ocfs2. For correctness sake I just pushed everything down in all cases to make sure that we keep the current behavior the same for everybody, and then each individual fs maintainer can make up their mind about what to do from there. Thanks, Acked-by: Jan Kara Signed-off-by: Josef Bacik Signed-off-by: Al Viro --- Documentation/filesystems/Locking | 6 ++--- Documentation/filesystems/porting | 7 ++++++ Documentation/filesystems/vfs.txt | 2 +- arch/powerpc/platforms/cell/spufs/file.c | 11 +++++++-- drivers/char/ps3flash.c | 9 ++++++-- drivers/mtd/ubi/cdev.c | 10 ++++++--- drivers/staging/pohmelfs/inode.c | 11 ++++++--- drivers/usb/gadget/printer.c | 5 ++++- drivers/video/fb_defio.c | 11 +++++++-- fs/9p/v9fs_vfs.h | 3 ++- fs/9p/vfs_file.c | 22 ++++++++++++++++-- fs/affs/affs.h | 2 +- fs/affs/file.c | 8 ++++++- fs/afs/internal.h | 2 +- fs/afs/write.c | 18 +++++++++++---- fs/bad_inode.c | 3 ++- fs/block_dev.c | 6 +---- fs/btrfs/ctree.h | 2 +- fs/btrfs/file.c | 21 +++++++++++++----- fs/ceph/caps.c | 6 +++-- fs/ceph/dir.c | 10 ++++++++- fs/ceph/super.h | 3 ++- fs/cifs/cifsfs.h | 4 ++-- fs/cifs/file.c | 18 +++++++++++++-- fs/coda/coda_int.h | 2 +- fs/coda/file.c | 8 ++++++- fs/ecryptfs/file.c | 7 +++--- fs/exofs/file.c | 10 ++++++++- fs/ext2/ext2.h | 3 ++- fs/ext2/file.c | 4 ++-- fs/ext3/fsync.c | 18 +++++++++++++-- fs/ext4/ext4.h | 2 +- fs/ext4/fsync.c | 38 +++++++++++++++++++++++++++++--- fs/fat/fat.h | 3 ++- fs/fat/file.c | 4 ++-- fs/fuse/dir.c | 5 +++-- fs/fuse/file.c | 24 +++++++++++++++----- fs/fuse/fuse_i.h | 3 ++- fs/gfs2/file.c | 17 +++++++++++--- fs/hfs/inode.c | 9 +++++++- fs/hfsplus/hfsplus_fs.h | 3 ++- fs/hfsplus/inode.c | 10 ++++++++- fs/hostfs/hostfs_kern.c | 15 +++++++++++-- fs/hpfs/file.c | 7 +++++- fs/hpfs/hpfs_fn.h | 2 +- fs/hppfs/hppfs.c | 5 +++-- fs/jffs2/file.c | 9 +++++++- fs/jffs2/os-linux.h | 2 +- fs/jfs/file.c | 9 +++++++- fs/jfs/jfs_inode.h | 2 +- fs/libfs.c | 16 ++++++++++---- fs/logfs/file.c | 11 ++++++++- fs/logfs/logfs.h | 2 +- fs/ncpfs/file.c | 4 ++-- fs/nfs/dir.c | 8 +++++-- fs/nfs/file.c | 11 ++++++--- fs/nilfs2/file.c | 12 ++++++++-- fs/nilfs2/nilfs.h | 2 +- fs/ntfs/dir.c | 10 ++++++++- fs/ntfs/file.c | 10 ++++++++- fs/ocfs2/file.c | 14 +++++++++++- fs/reiserfs/dir.c | 13 +++++++++-- fs/reiserfs/file.c | 9 +++++++- fs/sync.c | 25 +++------------------ fs/ubifs/file.c | 21 +++++++++--------- fs/ubifs/ubifs.h | 2 +- fs/xfs/linux-2.6/xfs_file.c | 17 +++++++------- include/linux/ext3_fs.h | 2 +- include/linux/fb.h | 3 ++- include/linux/fs.h | 9 ++++---- ipc/shm.c | 4 ++-- 71 files changed, 462 insertions(+), 164 deletions(-) (limited to 'fs/ocfs2') diff --git a/Documentation/filesystems/Locking b/Documentation/filesystems/Locking index 9b6ed7c9f34f..ca7e25292542 100644 --- a/Documentation/filesystems/Locking +++ b/Documentation/filesystems/Locking @@ -412,7 +412,7 @@ prototypes: int (*open) (struct inode *, struct file *); int (*flush) (struct file *); int (*release) (struct inode *, struct file *); - int (*fsync) (struct file *, int datasync); + int (*fsync) (struct file *, loff_t start, loff_t end, int datasync); int (*aio_fsync) (struct kiocb *, int datasync); int (*fasync) (int, struct file *, int); int (*lock) (struct file *, int, struct file_lock *); @@ -438,9 +438,7 @@ prototypes: locking rules: All may block except for ->setlease. - No VFS locks held on entry except for ->fsync and ->setlease. - -->fsync() has i_mutex on inode. + No VFS locks held on entry except for ->setlease. ->setlease has the file_list_lock held and must not sleep. diff --git a/Documentation/filesystems/porting b/Documentation/filesystems/porting index 6b96773e27cb..7f8861d341ea 100644 --- a/Documentation/filesystems/porting +++ b/Documentation/filesystems/porting @@ -421,3 +421,10 @@ data and there is a virtual hole at the end of the file. So if the provided offset is less than i_size and SEEK_DATA is specified, return the same offset. If the above is true for the offset and you are given SEEK_HOLE, return the end of the file. If the offset is i_size or greater return -ENXIO in either case. + +[mandatory] + If you have your own ->fsync() you must make sure to call +filemap_write_and_wait_range() so that all dirty pages are synced out properly. +You must also keep in mind that ->fsync() is not called with i_mutex held +anymore, so if you require i_mutex locking you must make sure to take it and +release it yourself. diff --git a/Documentation/filesystems/vfs.txt b/Documentation/filesystems/vfs.txt index 6bf85b78cfea..eff6617c9a0f 100644 --- a/Documentation/filesystems/vfs.txt +++ b/Documentation/filesystems/vfs.txt @@ -777,7 +777,7 @@ struct file_operations { int (*open) (struct inode *, struct file *); int (*flush) (struct file *); int (*release) (struct inode *, struct file *); - int (*fsync) (struct file *, int datasync); + int (*fsync) (struct file *, loff_t, loff_t, int datasync); int (*aio_fsync) (struct kiocb *, int datasync); int (*fasync) (int, struct file *, int); int (*lock) (struct file *, int, struct file_lock *); diff --git a/arch/powerpc/platforms/cell/spufs/file.c b/arch/powerpc/platforms/cell/spufs/file.c index 3c7c3f82d842..fb59c46e9e9e 100644 --- a/arch/powerpc/platforms/cell/spufs/file.c +++ b/arch/powerpc/platforms/cell/spufs/file.c @@ -1850,9 +1850,16 @@ out: return ret; } -static int spufs_mfc_fsync(struct file *file, int datasync) +static int spufs_mfc_fsync(struct file *file, loff_t start, loff_t end, int datasync) { - return spufs_mfc_flush(file, NULL); + struct inode *inode = file->f_path.dentry->d_inode; + int err = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (!err) { + mutex_lock(&inode->i_mutex); + err = spufs_mfc_flush(file, NULL); + mutex_unlock(&inode->i_mutex); + } + return err; } static int spufs_mfc_fasync(int fd, struct file *file, int on) diff --git a/drivers/char/ps3flash.c b/drivers/char/ps3flash.c index 5a06787e5be3..d0c57c2e2909 100644 --- a/drivers/char/ps3flash.c +++ b/drivers/char/ps3flash.c @@ -309,9 +309,14 @@ static int ps3flash_flush(struct file *file, fl_owner_t id) return ps3flash_writeback(ps3flash_dev); } -static int ps3flash_fsync(struct file *file, int datasync) +static int ps3flash_fsync(struct file *file, loff_t start, loff_t end, int datasync) { - return ps3flash_writeback(ps3flash_dev); + struct inode *inode = file->f_path.dentry->d_inode; + int err; + mutex_lock(&inode->i_mutex); + err = ps3flash_writeback(ps3flash_dev); + mutex_unlock(&inode->i_mutex); + return err; } static irqreturn_t ps3flash_interrupt(int irq, void *data) diff --git a/drivers/mtd/ubi/cdev.c b/drivers/mtd/ubi/cdev.c index 191f3bb3c41a..3320a50ba4f0 100644 --- a/drivers/mtd/ubi/cdev.c +++ b/drivers/mtd/ubi/cdev.c @@ -189,12 +189,16 @@ static loff_t vol_cdev_llseek(struct file *file, loff_t offset, int origin) return new_offset; } -static int vol_cdev_fsync(struct file *file, int datasync) +static int vol_cdev_fsync(struct file *file, loff_t start, loff_t end, int datasync) { struct ubi_volume_desc *desc = file->private_data; struct ubi_device *ubi = desc->vol->ubi; - - return ubi_sync(ubi->ubi_num); + struct inode *inode = file->f_path.dentry->d_inode; + int err; + mutex_lock(&inode->i_mutex); + err = ubi_sync(ubi->ubi_num); + mutex_unlock(&inode->i_mutex); + return err; } diff --git a/drivers/staging/pohmelfs/inode.c b/drivers/staging/pohmelfs/inode.c index c0f0ac7c1cdb..f3c6060c96b8 100644 --- a/drivers/staging/pohmelfs/inode.c +++ b/drivers/staging/pohmelfs/inode.c @@ -887,11 +887,16 @@ static struct inode *pohmelfs_alloc_inode(struct super_block *sb) /* * We want fsync() to work on POHMELFS. */ -static int pohmelfs_fsync(struct file *file, int datasync) +static int pohmelfs_fsync(struct file *file, loff_t start, loff_t end, int datasync) { struct inode *inode = file->f_mapping->host; - - return sync_inode_metadata(inode, 1); + int err = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (!err) { + mutex_lock(&inode->i_mutex); + err = sync_inode_metadata(inode, 1); + mutex_unlock(&inode->i_mutex); + } + return err; } ssize_t pohmelfs_write(struct file *file, const char __user *buf, diff --git a/drivers/usb/gadget/printer.c b/drivers/usb/gadget/printer.c index 271ef94668e7..978e6a101bf2 100644 --- a/drivers/usb/gadget/printer.c +++ b/drivers/usb/gadget/printer.c @@ -795,12 +795,14 @@ printer_write(struct file *fd, const char __user *buf, size_t len, loff_t *ptr) } static int -printer_fsync(struct file *fd, int datasync) +printer_fsync(struct file *fd, loff_t start, loff_t end, int datasync) { struct printer_dev *dev = fd->private_data; + struct inode *inode = fd->f_path.dentry->d_inode; unsigned long flags; int tx_list_empty; + mutex_lock(&inode->i_mutex); spin_lock_irqsave(&dev->lock, flags); tx_list_empty = (likely(list_empty(&dev->tx_reqs))); spin_unlock_irqrestore(&dev->lock, flags); @@ -810,6 +812,7 @@ printer_fsync(struct file *fd, int datasync) wait_event_interruptible(dev->tx_flush_wait, (likely(list_empty(&dev->tx_reqs_active)))); } + mutex_unlock(&inode->i_mutex); return 0; } diff --git a/drivers/video/fb_defio.c b/drivers/video/fb_defio.c index 804000183c5e..32814e8800e0 100644 --- a/drivers/video/fb_defio.c +++ b/drivers/video/fb_defio.c @@ -66,19 +66,26 @@ static int fb_deferred_io_fault(struct vm_area_struct *vma, return 0; } -int fb_deferred_io_fsync(struct file *file, int datasync) +int fb_deferred_io_fsync(struct file *file, loff_t start, loff_t end, int datasync) { struct fb_info *info = file->private_data; + struct inode *inode = file->f_path.dentry->d_inode; + int err = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (err) + return err; /* Skip if deferred io is compiled-in but disabled on this fbdev */ if (!info->fbdefio) return 0; + mutex_lock(&inode->i_mutex); /* Kill off the delayed work */ cancel_delayed_work_sync(&info->deferred_work); /* Run it immediately */ - return schedule_delayed_work(&info->deferred_work, 0); + err = schedule_delayed_work(&info->deferred_work, 0); + mutex_unlock(&inode->i_mutex); + return err; } EXPORT_SYMBOL_GPL(fb_deferred_io_fsync); diff --git a/fs/9p/v9fs_vfs.h b/fs/9p/v9fs_vfs.h index 4014160903a9..46ce357ca1ab 100644 --- a/fs/9p/v9fs_vfs.h +++ b/fs/9p/v9fs_vfs.h @@ -70,7 +70,8 @@ ssize_t v9fs_file_readn(struct file *, char *, char __user *, u32, u64); ssize_t v9fs_fid_readn(struct p9_fid *, char *, char __user *, u32, u64); void v9fs_blank_wstat(struct p9_wstat *wstat); int v9fs_vfs_setattr_dotl(struct dentry *, struct iattr *); -int v9fs_file_fsync_dotl(struct file *filp, int datasync); +int v9fs_file_fsync_dotl(struct file *filp, loff_t start, loff_t end, + int datasync); ssize_t v9fs_file_write_internal(struct inode *, struct p9_fid *, const char __user *, size_t, loff_t *, int); int v9fs_refresh_inode(struct p9_fid *fid, struct inode *inode); diff --git a/fs/9p/vfs_file.c b/fs/9p/vfs_file.c index ffed55817f0c..3c173fcc2c5a 100644 --- a/fs/9p/vfs_file.c +++ b/fs/9p/vfs_file.c @@ -519,32 +519,50 @@ out: } -static int v9fs_file_fsync(struct file *filp, int datasync) +static int v9fs_file_fsync(struct file *filp, loff_t start, loff_t end, + int datasync) { struct p9_fid *fid; + struct inode *inode = filp->f_mapping->host; struct p9_wstat wstat; int retval; + retval = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (retval) + return retval; + + mutex_lock(&inode->i_mutex); P9_DPRINTK(P9_DEBUG_VFS, "filp %p datasync %x\n", filp, datasync); fid = filp->private_data; v9fs_blank_wstat(&wstat); retval = p9_client_wstat(fid, &wstat); + mutex_unlock(&inode->i_mutex); + return retval; } -int v9fs_file_fsync_dotl(struct file *filp, int datasync) +int v9fs_file_fsync_dotl(struct file *filp, loff_t start, loff_t end, + int datasync) { struct p9_fid *fid; + struct inode *inode = filp->f_mapping->host; int retval; + retval = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (retval) + return retval; + + mutex_lock(&inode->i_mutex); P9_DPRINTK(P9_DEBUG_VFS, "v9fs_file_fsync_dotl: filp %p datasync %x\n", filp, datasync); fid = filp->private_data; retval = p9_client_fsync(fid, datasync); + mutex_unlock(&inode->i_mutex); + return retval; } diff --git a/fs/affs/affs.h b/fs/affs/affs.h index 0e95f73a7023..c2b9c79eb64e 100644 --- a/fs/affs/affs.h +++ b/fs/affs/affs.h @@ -182,7 +182,7 @@ extern int affs_add_entry(struct inode *dir, struct inode *inode, struct dent void affs_free_prealloc(struct inode *inode); extern void affs_truncate(struct inode *); -int affs_file_fsync(struct file *, int); +int affs_file_fsync(struct file *, loff_t, loff_t, int); /* dir.c */ diff --git a/fs/affs/file.c b/fs/affs/file.c index acf321b70fcd..2f4c935cb327 100644 --- a/fs/affs/file.c +++ b/fs/affs/file.c @@ -923,14 +923,20 @@ affs_truncate(struct inode *inode) affs_free_prealloc(inode); } -int affs_file_fsync(struct file *filp, int datasync) +int affs_file_fsync(struct file *filp, loff_t start, loff_t end, int datasync) { struct inode *inode = filp->f_mapping->host; int ret, err; + err = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (err) + return err; + + mutex_lock(&inode->i_mutex); ret = write_inode_now(inode, 0); err = sync_blockdev(inode->i_sb->s_bdev); if (!ret) ret = err; + mutex_unlock(&inode->i_mutex); return ret; } diff --git a/fs/afs/internal.h b/fs/afs/internal.h index f396d337b817..d2b0888126d4 100644 --- a/fs/afs/internal.h +++ b/fs/afs/internal.h @@ -750,7 +750,7 @@ extern void afs_pages_written_back(struct afs_vnode *, struct afs_call *); extern ssize_t afs_file_write(struct kiocb *, const struct iovec *, unsigned long, loff_t); extern int afs_writeback_all(struct afs_vnode *); -extern int afs_fsync(struct file *, int); +extern int afs_fsync(struct file *, loff_t, loff_t, int); /*****************************************************************************/ diff --git a/fs/afs/write.c b/fs/afs/write.c index b806285ff853..9aa52d93c73c 100644 --- a/fs/afs/write.c +++ b/fs/afs/write.c @@ -681,9 +681,10 @@ int afs_writeback_all(struct afs_vnode *vnode) * - the return status from this call provides a reliable indication of * whether any write errors occurred for this process. */ -int afs_fsync(struct file *file, int datasync) +int afs_fsync(struct file *file, loff_t start, loff_t end, int datasync) { struct dentry *dentry = file->f_path.dentry; + struct inode *inode = file->f_mapping->host; struct afs_writeback *wb, *xwb; struct afs_vnode *vnode = AFS_FS_I(dentry->d_inode); int ret; @@ -692,12 +693,19 @@ int afs_fsync(struct file *file, int datasync) vnode->fid.vid, vnode->fid.vnode, dentry->d_name.name, datasync); + ret = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (ret) + return ret; + mutex_lock(&inode->i_mutex); + /* use a writeback record as a marker in the queue - when this reaches * the front of the queue, all the outstanding writes are either * completed or rejected */ wb = kzalloc(sizeof(*wb), GFP_KERNEL); - if (!wb) - return -ENOMEM; + if (!wb) { + ret = -ENOMEM; + goto out; + } wb->vnode = vnode; wb->first = 0; wb->last = -1; @@ -720,7 +728,7 @@ int afs_fsync(struct file *file, int datasync) if (ret < 0) { afs_put_writeback(wb); _leave(" = %d [wb]", ret); - return ret; + goto out; } /* wait for the preceding writes to actually complete */ @@ -729,6 +737,8 @@ int afs_fsync(struct file *file, int datasync) vnode->writebacks.next == &wb->link); afs_put_writeback(wb); _leave(" = %d", ret); +out: + mutex_unlock(&inode->i_mutex); return ret; } diff --git a/fs/bad_inode.c b/fs/bad_inode.c index f024d8aaddef..9205cf25f1c6 100644 --- a/fs/bad_inode.c +++ b/fs/bad_inode.c @@ -87,7 +87,8 @@ static int bad_file_release(struct inode *inode, struct file *filp) return -EIO; } -static int bad_file_fsync(struct file *file, int datasync) +static int bad_file_fsync(struct file *file, loff_t start, loff_t end, + int datasync) { return -EIO; } diff --git a/fs/block_dev.c b/fs/block_dev.c index 966617a422d9..9fb0b15331d3 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -378,7 +378,7 @@ out: return retval; } -int blkdev_fsync(struct file *filp, int datasync) +int blkdev_fsync(struct file *filp, loff_t start, loff_t end, int datasync) { struct inode *bd_inode = filp->f_mapping->host; struct block_device *bdev = I_BDEV(bd_inode); @@ -389,14 +389,10 @@ int blkdev_fsync(struct file *filp, int datasync) * i_mutex and doing so causes performance issues with concurrent * O_SYNC writers to a block device. */ - mutex_unlock(&bd_inode->i_mutex); - error = blkdev_issue_flush(bdev, GFP_KERNEL, NULL); if (error == -EOPNOTSUPP) error = 0; - mutex_lock(&bd_inode->i_mutex); - return error; } EXPORT_SYMBOL(blkdev_fsync); diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index f1ff62bff1b3..82be74efbb26 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -2605,7 +2605,7 @@ int btrfs_defrag_file(struct inode *inode, struct file *file, int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans, struct inode *inode); int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info); -int btrfs_sync_file(struct file *file, int datasync); +int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync); int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end, int skip_pinned); extern const struct file_operations btrfs_file_operations; diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index bd4d061c6e4d..59cbdb120ad0 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -1452,7 +1452,7 @@ int btrfs_release_file(struct inode *inode, struct file *filp) * important optimization for directories because holding the mutex prevents * new operations on the dir while we write to disk. */ -int btrfs_sync_file(struct file *file, int datasync) +int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync) { struct dentry *dentry = file->f_path.dentry; struct inode *inode = dentry->d_inode; @@ -1462,9 +1462,13 @@ int btrfs_sync_file(struct file *file, int datasync) trace_btrfs_sync_file(file, datasync); + ret = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (ret) + return ret; + mutex_lock(&inode->i_mutex); + /* we wait first, since the writeback may change the inode */ root->log_batch++; - /* the VFS called filemap_fdatawrite for us */ btrfs_wait_ordered_range(inode, 0, (u64)-1); root->log_batch++; @@ -1472,8 +1476,10 @@ int btrfs_sync_file(struct file *file, int datasync) * check the transaction that last modified this inode * and see if its already been committed */ - if (!BTRFS_I(inode)->last_trans) + if (!BTRFS_I(inode)->last_trans) { + mutex_unlock(&inode->i_mutex); goto out; + } /* * if the last transaction that changed this file was before @@ -1484,6 +1490,7 @@ int btrfs_sync_file(struct file *file, int datasync) if (BTRFS_I(inode)->last_trans <= root->fs_info->last_trans_committed) { BTRFS_I(inode)->last_trans = 0; + mutex_unlock(&inode->i_mutex); goto out; } @@ -1496,12 +1503,15 @@ int btrfs_sync_file(struct file *file, int datasync) trans = btrfs_start_transaction(root, 0); if (IS_ERR(trans)) { ret = PTR_ERR(trans); + mutex_unlock(&inode->i_mutex); goto out; } ret = btrfs_log_dentry_safe(trans, root, dentry); - if (ret < 0) + if (ret < 0) { + mutex_unlock(&inode->i_mutex); goto out; + } /* we've logged all the items and now have a consistent * version of the file in the log. It is possible that @@ -1513,7 +1523,7 @@ int btrfs_sync_file(struct file *file, int datasync) * file again, but that will end up using the synchronization * inside btrfs_sync_log to keep things safe. */ - mutex_unlock(&dentry->d_inode->i_mutex); + mutex_unlock(&inode->i_mutex); if (ret != BTRFS_NO_LOG_SYNC) { if (ret > 0) { @@ -1528,7 +1538,6 @@ int btrfs_sync_file(struct file *file, int datasync) } else { ret = btrfs_end_transaction(trans, root); } - mutex_lock(&dentry->d_inode->i_mutex); out: return ret > 0 ? -EIO : ret; } diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index f605753c8fe9..8d74ad7ba556 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -1811,7 +1811,7 @@ out: spin_unlock(&ci->i_unsafe_lock); } -int ceph_fsync(struct file *file, int datasync) +int ceph_fsync(struct file *file, loff_t start, loff_t end, int datasync) { struct inode *inode = file->f_mapping->host; struct ceph_inode_info *ci = ceph_inode(inode); @@ -1822,9 +1822,10 @@ int ceph_fsync(struct file *file, int datasync) dout("fsync %p%s\n", inode, datasync ? " datasync" : ""); sync_write_wait(inode); - ret = filemap_write_and_wait(inode->i_mapping); + ret = filemap_write_and_wait_range(inode->i_mapping, start, end); if (ret < 0) return ret; + mutex_lock(&inode->i_mutex); dirty = try_flush_caps(inode, NULL, &flush_tid); dout("fsync dirty caps are %s\n", ceph_cap_string(dirty)); @@ -1841,6 +1842,7 @@ int ceph_fsync(struct file *file, int datasync) } dout("fsync %p%s done\n", inode, datasync ? " datasync" : ""); + mutex_unlock(&inode->i_mutex); return ret; } diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index 0972b457a03f..1065ac779840 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -1118,7 +1118,8 @@ static ssize_t ceph_read_dir(struct file *file, char __user *buf, size_t size, * an fsync() on a dir will wait for any uncommitted directory * operations to commit. */ -static int ceph_dir_fsync(struct file *file, int datasync) +static int ceph_dir_fsync(struct file *file, loff_t start, loff_t end, + int datasync) { struct inode *inode = file->f_path.dentry->d_inode; struct ceph_inode_info *ci = ceph_inode(inode); @@ -1128,6 +1129,11 @@ static int ceph_dir_fsync(struct file *file, int datasync) int ret = 0; dout("dir_fsync %p\n", inode); + ret = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (ret) + return ret; + mutex_lock(&inode->i_mutex); + spin_lock(&ci->i_unsafe_lock); if (list_empty(head)) goto out; @@ -1161,6 +1167,8 @@ static int ceph_dir_fsync(struct file *file, int datasync) } while (req->r_tid < last_tid); out: spin_unlock(&ci->i_unsafe_lock); + mutex_unlock(&inode->i_mutex); + return ret; } diff --git a/fs/ceph/super.h b/fs/ceph/super.h index 56c41ef47cad..30446b144e3d 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -728,7 +728,8 @@ extern void ceph_put_cap(struct ceph_mds_client *mdsc, extern void ceph_queue_caps_release(struct inode *inode); extern int ceph_write_inode(struct inode *inode, struct writeback_control *wbc); -extern int ceph_fsync(struct file *file, int datasync); +extern int ceph_fsync(struct file *file, loff_t start, loff_t end, + int datasync); extern void ceph_kick_flushing_caps(struct ceph_mds_client *mdsc, struct ceph_mds_session *session); extern struct ceph_cap *ceph_get_cap_for_mds(struct ceph_inode_info *ci, diff --git a/fs/cifs/cifsfs.h b/fs/cifs/cifsfs.h index 036ca83e5f46..fbd050c8d52a 100644 --- a/fs/cifs/cifsfs.h +++ b/fs/cifs/cifsfs.h @@ -91,8 +91,8 @@ extern ssize_t cifs_user_writev(struct kiocb *iocb, const struct iovec *iov, extern ssize_t cifs_strict_writev(struct kiocb *iocb, const struct iovec *iov, unsigned long nr_segs, loff_t pos); extern int cifs_lock(struct file *, int, struct file_lock *); -extern int cifs_fsync(struct file *, int); -extern int cifs_strict_fsync(struct file *, int); +extern int cifs_fsync(struct file *, loff_t, loff_t, int); +extern int cifs_strict_fsync(struct file *, loff_t, loff_t, int); extern int cifs_flush(struct file *, fl_owner_t id); extern int cifs_file_mmap(struct file * , struct vm_area_struct *); extern int cifs_file_strict_mmap(struct file * , struct vm_area_struct *); diff --git a/fs/cifs/file.c b/fs/cifs/file.c index bb71471a4d9d..cef584451113 100644 --- a/fs/cifs/file.c +++ b/fs/cifs/file.c @@ -1401,7 +1401,8 @@ static int cifs_write_end(struct file *file, struct address_space *mapping, return rc; } -int cifs_strict_fsync(struct file *file, int datasync) +int cifs_strict_fsync(struct file *file, loff_t start, loff_t end, + int datasync) { int xid; int rc = 0; @@ -1410,6 +1411,11 @@ int cifs_strict_fsync(struct file *file, int datasync) struct inode *inode = file->f_path.dentry->d_inode; struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb); + rc = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (rc) + return rc; + mutex_lock(&inode->i_mutex); + xid = GetXid(); cFYI(1, "Sync file - name: %s datasync: 0x%x", @@ -1428,16 +1434,23 @@ int cifs_strict_fsync(struct file *file, int datasync) rc = CIFSSMBFlush(xid, tcon, smbfile->netfid); FreeXid(xid); + mutex_unlock(&inode->i_mutex); return rc; } -int cifs_fsync(struct file *file, int datasync) +int cifs_fsync(struct file *file, loff_t start, loff_t end, int datasync) { int xid; int rc = 0; struct cifs_tcon *tcon; struct cifsFileInfo *smbfile = file->private_data; struct cifs_sb_info *cifs_sb = CIFS_SB(file->f_path.dentry->d_sb); + struct inode *inode = file->f_mapping->host; + + rc = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (rc) + return rc; + mutex_lock(&inode->i_mutex); xid = GetXid(); @@ -1449,6 +1462,7 @@ int cifs_fsync(struct file *file, int datasync) rc = CIFSSMBFlush(xid, tcon, smbfile->netfid); FreeXid(xid); + mutex_unlock(&inode->i_mutex); return rc; } diff --git a/fs/coda/coda_int.h b/fs/coda/coda_int.h index 6b443ff43a19..b7143cf783ac 100644 --- a/fs/coda/coda_int.h +++ b/fs/coda/coda_int.h @@ -11,7 +11,7 @@ extern int coda_fake_statfs; void coda_destroy_inodecache(void); int coda_init_inodecache(void); -int coda_fsync(struct file *coda_file, int datasync); +int coda_fsync(struct file *coda_file, loff_t start, loff_t end, int datasync); void coda_sysctl_init(void); void coda_sysctl_clean(void); diff --git a/fs/coda/file.c b/fs/coda/file.c index 0433057be330..8edd404e6419 100644 --- a/fs/coda/file.c +++ b/fs/coda/file.c @@ -199,7 +199,7 @@ int coda_release(struct inode *coda_inode, struct file *coda_file) return 0; } -int coda_fsync(struct file *coda_file, int datasync) +int coda_fsync(struct file *coda_file, loff_t start, loff_t end, int datasync) { struct file *host_file; struct inode *coda_inode = coda_file->f_path.dentry->d_inode; @@ -210,6 +210,11 @@ int coda_fsync(struct file *coda_file, int datasync) S_ISLNK(coda_inode->i_mode))) return -EINVAL; + err = filemap_write_and_wait_range(coda_inode->i_mapping, start, end); + if (err) + return err; + mutex_lock(&coda_inode->i_mutex); + cfi = CODA_FTOC(coda_file); BUG_ON(!cfi || cfi->cfi_magic != CODA_MAGIC); host_file = cfi->cfi_container; @@ -217,6 +222,7 @@ int coda_fsync(struct file *coda_file, int datasync) err = vfs_fsync(host_file, datasync); if (!err && !datasync) err = venus_fsync(coda_inode->i_sb, coda_i2f(coda_inode)); + mutex_unlock(&coda_inode->i_mutex); return err; } diff --git a/fs/ecryptfs/file.c b/fs/ecryptfs/file.c index 4ec9eb00a241..c6ac98cf9baa 100644 --- a/fs/ecryptfs/file.c +++ b/fs/ecryptfs/file.c @@ -270,14 +270,15 @@ static int ecryptfs_release(struct inode *inode, struct file *file) } static int -ecryptfs_fsync(struct file *file, int datasync) +ecryptfs_fsync(struct file *file, loff_t start, loff_t end, int datasync) { int rc = 0; - rc = generic_file_fsync(file, datasync); + rc = generic_file_fsync(file, start, end, datasync); if (rc) goto out; - rc = vfs_fsync(ecryptfs_file_to_lower(file), datasync); + rc = vfs_fsync_range(ecryptfs_file_to_lower(file), start, end, + datasync); out: return rc; } diff --git a/fs/exofs/file.c b/fs/exofs/file.c index 45ca323d8363..491c6c078e7f 100644 --- a/fs/exofs/file.c +++ b/fs/exofs/file.c @@ -42,11 +42,19 @@ static int exofs_release_file(struct inode *inode, struct file *filp) * Note, in exofs all metadata is written as part of inode, regardless. * The writeout is synchronous */ -static int exofs_file_fsync(struct file *filp, int datasync) +static int exofs_file_fsync(struct file *filp, loff_t start, loff_t end, + int datasync) { + struct inode *inode = filp->f_mapping->host; int ret; + ret = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (ret) + return ret; + + mutex_lock(&inode->i_mutex); ret = sync_inode_metadata(filp->f_mapping->host, 1); + mutex_unlock(&inode->i_mutex); return ret; } diff --git a/fs/ext2/ext2.h b/fs/ext2/ext2.h index 645be9e7ee47..af9fc89b1b2d 100644 --- a/fs/ext2/ext2.h +++ b/fs/ext2/ext2.h @@ -150,7 +150,8 @@ extern void ext2_write_super (struct super_block *); extern const struct file_operations ext2_dir_operations; /* file.c */ -extern int ext2_fsync(struct file *file, int datasync); +extern int ext2_fsync(struct file *file, loff_t start, loff_t end, + int datasync); extern const struct inode_operations ext2_file_inode_operations; extern const struct file_operations ext2_file_operations; extern const struct file_operations ext2_xip_file_operations; diff --git a/fs/ext2/file.c b/fs/ext2/file.c index 49eec9456c5b..82e06321de35 100644 --- a/fs/ext2/file.c +++ b/fs/ext2/file.c @@ -40,13 +40,13 @@ static int ext2_release_file (struct inode * inode, struct file * filp) return 0; } -int ext2_fsync(struct file *file, int datasync) +int ext2_fsync(struct file *file, loff_t start, loff_t end, int datasync) { int ret; struct super_block *sb = file->f_mapping->host->i_sb; struct address_space *mapping = sb->s_bdev->bd_inode->i_mapping; - ret = generic_file_fsync(file, datasync); + ret = generic_file_fsync(file, start, end, datasync); if (ret == -EIO || test_and_clear_bit(AS_EIO, &mapping->flags)) { /* We don't really know where the IO error happened... */ ext2_error(sb, __func__, diff --git a/fs/ext3/fsync.c b/fs/ext3/fsync.c index 09b13bb34c94..0bcf63adb80a 100644 --- a/fs/ext3/fsync.c +++ b/fs/ext3/fsync.c @@ -43,7 +43,7 @@ * inode to disk. */ -int ext3_sync_file(struct file *file, int datasync) +int ext3_sync_file(struct file *file, loff_t start, loff_t end, int datasync) { struct inode *inode = file->f_mapping->host; struct ext3_inode_info *ei = EXT3_I(inode); @@ -54,6 +54,17 @@ int ext3_sync_file(struct file *file, int datasync) if (inode->i_sb->s_flags & MS_RDONLY) return 0; + ret = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (ret) + return ret; + + /* + * Taking the mutex here just to keep consistent with how fsync was + * called previously, however it looks like we don't need to take + * i_mutex at all. + */ + mutex_lock(&inode->i_mutex); + J_ASSERT(ext3_journal_current_handle() == NULL); /* @@ -70,8 +81,10 @@ int ext3_sync_file(struct file *file, int datasync) * (they were dirtied by commit). But that's OK - the blocks are * safe in-journal, which is all fsync() needs to ensure. */ - if (ext3_should_journal_data(inode)) + if (ext3_should_journal_data(inode)) { + mutex_unlock(&inode->i_mutex); return ext3_force_commit(inode->i_sb); + } if (datasync) commit_tid = atomic_read(&ei->i_datasync_tid); @@ -91,5 +104,6 @@ int ext3_sync_file(struct file *file, int datasync) */ if (needs_barrier) blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL, NULL); + mutex_unlock(&inode->i_mutex); return ret; } diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 1921392cd708..fa44df879711 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -1758,7 +1758,7 @@ extern int ext4_htree_store_dirent(struct file *dir_file, __u32 hash, extern void ext4_htree_free_dir_info(struct dir_private_info *p); /* fsync.c */ -extern int ext4_sync_file(struct file *, int); +extern int ext4_sync_file(struct file *, loff_t, loff_t, int); extern int ext4_flush_completed_IO(struct inode *); /* hash.c */ diff --git a/fs/ext4/fsync.c b/fs/ext4/fsync.c index ce66d2fe826c..da3bed3e0c29 100644 --- a/fs/ext4/fsync.c +++ b/fs/ext4/fsync.c @@ -151,6 +151,32 @@ static int ext4_sync_parent(struct inode *inode) return ret; } +/** + * __sync_file - generic_file_fsync without the locking and filemap_write + * @inode: inode to sync + * @datasync: only sync essential metadata if true + * + * This is just generic_file_fsync without the locking. This is needed for + * nojournal mode to make sure this inodes data/metadata makes it to disk + * properly. The i_mutex should be held already. + */ +static int __sync_inode(struct inode *inode, int datasync) +{ + int err; + int ret; + + ret = sync_mapping_buffers(inode->i_mapping); + if (!(inode->i_state & I_DIRTY)) + return ret; + if (datasync && !(inode->i_state & I_DIRTY_DATASYNC)) + return ret; + + err = sync_inode_metadata(inode, 1); + if (ret == 0) + ret = err; + return ret; +} + /* * akpm: A new design for ext4_sync_file(). * @@ -165,7 +191,7 @@ static int ext4_sync_parent(struct inode *inode) * i_mutex lock is held when entering and exiting this function */ -int ext4_sync_file(struct file *file, int datasync) +int ext4_sync_file(struct file *file, loff_t start, loff_t end, int datasync) { struct inode *inode = file->f_mapping->host; struct ext4_inode_info *ei = EXT4_I(inode); @@ -178,15 +204,20 @@ int ext4_sync_file(struct file *file, int datasync) trace_ext4_sync_file_enter(file, datasync); + ret = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (ret) + return ret; + mutex_lock(&inode->i_mutex); + if (inode->i_sb->s_flags & MS_RDONLY) - return 0; + goto out; ret = ext4_flush_completed_IO(inode); if (ret < 0) goto out; if (!journal) { - ret = generic_file_fsync(file, datasync); + ret = __sync_inode(inode, datasync); if (!ret && !list_empty(&inode->i_dentry)) ret = ext4_sync_parent(inode); goto out; @@ -220,6 +251,7 @@ int ext4_sync_file(struct file *file, int datasync) if (needs_barrier) blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL, NULL); out: + mutex_unlock(&inode->i_mutex); trace_ext4_sync_file_exit(inode, ret); return ret; } diff --git a/fs/fat/fat.h b/fs/fat/fat.h index a975b4147e91..a5d3853822e0 100644 --- a/fs/fat/fat.h +++ b/fs/fat/fat.h @@ -310,7 +310,8 @@ extern int fat_setattr(struct dentry * dentry, struct iattr * attr); extern void fat_truncate_blocks(struct inode *inode, loff_t offset); extern int fat_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat); -extern int fat_file_fsync(struct file *file, int datasync); +extern int fat_file_fsync(struct file *file, loff_t start, loff_t end, + int datasync); /* fat/inode.c */ extern void fat_attach(struct inode *inode, loff_t i_pos); diff --git a/fs/fat/file.c b/fs/fat/file.c index e1587c54d3c1..c118acf16e43 100644 --- a/fs/fat/file.c +++ b/fs/fat/file.c @@ -149,12 +149,12 @@ static int fat_file_release(struct inode *inode, struct file *filp) return 0; } -int fat_file_fsync(struct file *filp, int datasync) +int fat_file_fsync(struct file *filp, loff_t start, loff_t end, int datasync) { struct inode *inode = filp->f_mapping->host; int res, err; - res = generic_file_fsync(filp, datasync); + res = generic_file_fsync(filp, start, end, datasync); err = sync_mapping_buffers(MSDOS_SB(inode->i_sb)->fat_inode->i_mapping); return res ? res : err; diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index 02063dde2728..9f63e493a9b6 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -1176,9 +1176,10 @@ static int fuse_dir_release(struct inode *inode, struct file *file) return 0; } -static int fuse_dir_fsync(struct file *file, int datasync) +static int fuse_dir_fsync(struct file *file, loff_t start, loff_t end, + int datasync) { - return fuse_fsync_common(file, datasync, 1); + return fuse_fsync_common(file, start, end, datasync, 1); } static bool update_mtime(unsigned ivalid) diff --git a/fs/fuse/file.c b/fs/fuse/file.c index 73b89df20851..7bb685cdd00c 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -400,7 +400,8 @@ static void fuse_sync_writes(struct inode *inode) fuse_release_nowrite(inode); } -int fuse_fsync_common(struct file *file, int datasync, int isdir) +int fuse_fsync_common(struct file *file, loff_t start, loff_t end, + int datasync, int isdir) { struct inode *inode = file->f_mapping->host; struct fuse_conn *fc = get_fuse_conn(inode); @@ -412,9 +413,15 @@ int fuse_fsync_common(struct file *file, int datasync, int isdir) if (is_bad_inode(inode)) return -EIO; + err = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (err) + return err; + if ((!isdir && fc->no_fsync) || (isdir && fc->no_fsyncdir)) return 0; + mutex_lock(&inode->i_mutex); + /* * Start writeback against all dirty pages of the inode, then * wait for all outstanding writes, before sending the FSYNC @@ -422,13 +429,15 @@ int fuse_fsync_common(struct file *file, int datasync, int isdir) */ err = write_inode_now(inode, 0); if (err) - return err; + goto out; fuse_sync_writes(inode); req = fuse_get_req(fc); - if (IS_ERR(req)) - return PTR_ERR(req); + if (IS_ERR(req)) { + err = PTR_ERR(req); + goto out; + } memset(&inarg, 0, sizeof(inarg)); inarg.fh = ff->fh; @@ -448,12 +457,15 @@ int fuse_fsync_common(struct file *file, int datasync, int isdir) fc->no_fsync = 1; err = 0; } +out: + mutex_unlock(&inode->i_mutex); return err; } -static int fuse_fsync(struct file *file, int datasync) +static int fuse_fsync(struct file *file, loff_t start, loff_t end, + int datasync) { - return fuse_fsync_common(file, datasync, 0); + return fuse_fsync_common(file, start, end, datasync, 0); } void fuse_read_fill(struct fuse_req *req, struct file *file, loff_t pos, diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h index b788becada76..c6aa2d4b8517 100644 --- a/fs/fuse/fuse_i.h +++ b/fs/fuse/fuse_i.h @@ -589,7 +589,8 @@ void fuse_release_common(struct file *file, int opcode); /** * Send FSYNC or FSYNCDIR request */ -int fuse_fsync_common(struct file *file, int datasync, int isdir); +int fuse_fsync_common(struct file *file, loff_t start, loff_t end, + int datasync, int isdir); /** * Notify poll wakeup diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c index 89c39e53760d..f82cb5e1cb6b 100644 --- a/fs/gfs2/file.c +++ b/fs/gfs2/file.c @@ -544,7 +544,9 @@ static int gfs2_close(struct inode *inode, struct file *file) /** * gfs2_fsync - sync the dirty data for a file (across the cluster) - * @file: the file that points to the dentry (we ignore this) + * @file: the file that points to the dentry + * @start: the start position in the file to sync + * @end: the end position in the file to sync * @datasync: set if we can ignore timestamp changes * * The VFS will flush data for us. We only need to worry @@ -553,23 +555,32 @@ static int gfs2_close(struct inode *inode, struct file *file) * Returns: errno */ -static int gfs2_fsync(struct file *file, int datasync) +static int gfs2_fsync(struct file *file, loff_t start, loff_t end, + int datasync) { struct inode *inode = file->f_mapping->host; int sync_state = inode->i_state & (I_DIRTY_SYNC|I_DIRTY_DATASYNC); struct gfs2_inode *ip = GFS2_I(inode); int ret; + ret = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (ret) + return ret; + mutex_lock(&inode->i_mutex); + if (datasync) sync_state &= ~I_DIRTY_SYNC; if (sync_state) { ret = sync_inode_metadata(inode, 1); - if (ret) + if (ret) { + mutex_unlock(&inode->i_mutex); return ret; + } gfs2_ail_flush(ip->i_gl); } + mutex_unlock(&inode->i_mutex); return 0; } diff --git a/fs/hfs/inode.c b/fs/hfs/inode.c index 5e7c3f309617..96a1b625fc74 100644 --- a/fs/hfs/inode.c +++ b/fs/hfs/inode.c @@ -627,12 +627,18 @@ int hfs_inode_setattr(struct dentry *dentry, struct iattr * attr) return 0; } -static int hfs_file_fsync(struct file *filp, int datasync) +static int hfs_file_fsync(struct file *filp, loff_t start, loff_t end, + int datasync) { struct inode *inode = filp->f_mapping->host; struct super_block * sb; int ret, err; + ret = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (ret) + return ret; + mutex_lock(&inode->i_mutex); + /* sync the inode to buffers */ ret = write_inode_now(inode, 0); @@ -649,6 +655,7 @@ static int hfs_file_fsync(struct file *filp, int datasync) err = sync_blockdev(sb->s_bdev); if (!ret) ret = err; + mutex_unlock(&inode->i_mutex); return ret; } diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h index d6857523336d..38184e360932 100644 --- a/fs/hfsplus/hfsplus_fs.h +++ b/fs/hfsplus/hfsplus_fs.h @@ -392,7 +392,8 @@ int hfsplus_cat_read_inode(struct inode *, struct hfs_find_data *); int hfsplus_cat_write_inode(struct inode *); struct inode *hfsplus_new_inode(struct super_block *, int); void hfsplus_delete_inode(struct inode *); -int hfsplus_file_fsync(struct file *file, int datasync); +int hfsplus_file_fsync(struct file *file, loff_t start, loff_t end, + int datasync); /* ioctl.c */ long hfsplus_ioctl(struct file *filp, unsigned int cmd, unsigned long arg); diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c index 5b1cb98741cc..30486e01d003 100644 --- a/fs/hfsplus/inode.c +++ b/fs/hfsplus/inode.c @@ -308,13 +308,19 @@ static int hfsplus_setattr(struct dentry *dentry, struct iattr *attr) return 0; } -int hfsplus_file_fsync(struct file *file, int datasync) +int hfsplus_file_fsync(struct file *file, loff_t start, loff_t end, + int datasync) { struct inode *inode = file->f_mapping->host; struct hfsplus_inode_info *hip = HFSPLUS_I(inode); struct hfsplus_sb_info *sbi = HFSPLUS_SB(inode->i_sb); int error = 0, error2; + error = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (error) + return error; + mutex_lock(&inode->i_mutex); + /* * Sync inode metadata into the catalog and extent trees. */ @@ -342,6 +348,8 @@ int hfsplus_file_fsync(struct file *file, int datasync) if (!test_bit(HFSPLUS_SB_NOBARRIER, &sbi->flags)) blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL, NULL); + mutex_unlock(&inode->i_mutex); + return error; } diff --git a/fs/hostfs/hostfs_kern.c b/fs/hostfs/hostfs_kern.c index 6e449c599b9d..0d22afdd4611 100644 --- a/fs/hostfs/hostfs_kern.c +++ b/fs/hostfs/hostfs_kern.c @@ -362,9 +362,20 @@ retry: return 0; } -int hostfs_fsync(struct file *file, int datasync) +int hostfs_fsync(struct file *file, loff_t start, loff_t end, int datasync) { - return fsync_file(HOSTFS_I(file->f_mapping->host)->fd, datasync); + struct inode *inode = file->f_mapping->host; + int ret; + + ret = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (ret) + return ret; + + mutex_lock(&inode->i_mutex); + ret = fsync_file(HOSTFS_I(inode)->fd, datasync); + mutex_unlock(&inode->i_mutex); + + return ret; } static const struct file_operations hostfs_file_fops = { diff --git a/fs/hpfs/file.c b/fs/hpfs/file.c index 89c500ee5213..89d2a5803ae3 100644 --- a/fs/hpfs/file.c +++ b/fs/hpfs/file.c @@ -18,9 +18,14 @@ static int hpfs_file_release(struct inode *inode, struct file *file) return 0; } -int hpfs_file_fsync(struct file *file, int datasync) +int hpfs_file_fsync(struct file *file, loff_t start, loff_t end, int datasync) { struct inode *inode = file->f_mapping->host; + int ret; + + ret = filemap_write_and_wait_range(file->f_mapping, start, end); + if (ret) + return ret; return sync_blockdev(inode->i_sb->s_bdev); } diff --git a/fs/hpfs/hpfs_fn.h b/fs/hpfs/hpfs_fn.h index dd552f862c8f..331b5e234ef3 100644 --- a/fs/hpfs/hpfs_fn.h +++ b/fs/hpfs/hpfs_fn.h @@ -258,7 +258,7 @@ void hpfs_set_ea(struct inode *, struct fnode *, const char *, /* file.c */ -int hpfs_file_fsync(struct file *, int); +int hpfs_file_fsync(struct file *, loff_t, loff_t, int); extern const struct file_operations hpfs_file_ops; extern const struct inode_operations hpfs_file_iops; extern const struct address_space_operations hpfs_aops; diff --git a/fs/hppfs/hppfs.c b/fs/hppfs/hppfs.c index 85c098a499f3..8635be5ffd97 100644 --- a/fs/hppfs/hppfs.c +++ b/fs/hppfs/hppfs.c @@ -573,9 +573,10 @@ static int hppfs_readdir(struct file *file, void *ent, filldir_t filldir) return err; } -static int hppfs_fsync(struct file *file, int datasync) +static int hppfs_fsync(struct file *file, loff_t start, loff_t end, + int datasync) { - return 0; + return filemap_write_and_wait_range(file->f_mapping, start, end); } static const struct file_operations hppfs_dir_fops = { diff --git a/fs/jffs2/file.c b/fs/jffs2/file.c index 1c0a08d711aa..3989f7e09f7f 100644 --- a/fs/jffs2/file.c +++ b/fs/jffs2/file.c @@ -27,13 +27,20 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping, struct page **pagep, void **fsdata); static int jffs2_readpage (struct file *filp, struct page *pg); -int jffs2_fsync(struct file *filp, int datasync) +int jffs2_fsync(struct file *filp, loff_t start, loff_t end, int datasync) { struct inode *inode = filp->f_mapping->host; struct jffs2_sb_info *c = JFFS2_SB_INFO(inode->i_sb); + int ret; + + ret = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (ret) + return ret; + mutex_lock(&inode->i_mutex); /* Trigger GC to flush any pending writes for this inode */ jffs2_flush_wbuf_gc(c, inode->i_ino); + mutex_unlock(&inode->i_mutex); return 0; } diff --git a/fs/jffs2/os-linux.h b/fs/jffs2/os-linux.h index 65c6c43ca482..9c252835e8e5 100644 --- a/fs/jffs2/os-linux.h +++ b/fs/jffs2/os-linux.h @@ -158,7 +158,7 @@ extern const struct inode_operations jffs2_dir_inode_operations; extern const struct file_operations jffs2_file_operations; extern const struct inode_operations jffs2_file_inode_operations; extern const struct address_space_operations jffs2_file_address_operations; -int jffs2_fsync(struct file *, int); +int jffs2_fsync(struct file *, loff_t, loff_t, int); int jffs2_do_readpage_unlock (struct inode *inode, struct page *pg); /* ioctl.c */ diff --git a/fs/jfs/file.c b/fs/jfs/file.c index 9f32315acef1..7527855b5cc6 100644 --- a/fs/jfs/file.c +++ b/fs/jfs/file.c @@ -28,19 +28,26 @@ #include "jfs_acl.h" #include "jfs_debug.h" -int jfs_fsync(struct file *file, int datasync) +int jfs_fsync(struct file *file, loff_t start, loff_t end, int datasync) { struct inode *inode = file->f_mapping->host; int rc = 0; + rc = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (rc) + return rc; + + mutex_lock(&inode->i_mutex); if (!(inode->i_state & I_DIRTY) || (datasync && !(inode->i_state & I_DIRTY_DATASYNC))) { /* Make sure committed changes hit the disk */ jfs_flush_journal(JFS_SBI(inode->i_sb)->log, 1); + mutex_unlock(&inode->i_mutex); return rc; } rc |= jfs_commit_inode(inode, 1); + mutex_unlock(&inode->i_mutex); return rc ? -EIO : 0; } diff --git a/fs/jfs/jfs_inode.h b/fs/jfs/jfs_inode.h index ec2fb8b945fc..9271cfe4a149 100644 --- a/fs/jfs/jfs_inode.h +++ b/fs/jfs/jfs_inode.h @@ -21,7 +21,7 @@ struct fid; extern struct inode *ialloc(struct inode *, umode_t); -extern int jfs_fsync(struct file *, int); +extern int jfs_fsync(struct file *, loff_t, loff_t, int); extern long jfs_ioctl(struct file *, unsigned int, unsigned long); extern long jfs_compat_ioctl(struct file *, unsigned int, unsigned long); extern struct inode *jfs_iget(struct super_block *, unsigned long); diff --git a/fs/libfs.c b/fs/libfs.c index bd50b11f92da..8f2271a5df53 100644 --- a/fs/libfs.c +++ b/fs/libfs.c @@ -905,21 +905,29 @@ EXPORT_SYMBOL_GPL(generic_fh_to_parent); * filesystems which track all non-inode metadata in the buffers list * hanging off the address_space structure. */ -int generic_file_fsync(struct file *file, int datasync) +int generic_file_fsync(struct file *file, loff_t start, loff_t end, + int datasync) { struct inode *inode = file->f_mapping->host; int err; int ret; + err = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (err) + return err; + + mutex_lock(&inode->i_mutex); ret = sync_mapping_buffers(inode->i_mapping); if (!(inode->i_state & I_DIRTY)) - return ret; + goto out; if (datasync && !(inode->i_state & I_DIRTY_DATASYNC)) - return ret; + goto out; err = sync_inode_metadata(inode, 1); if (ret == 0) ret = err; +out: + mutex_unlock(&inode->i_mutex); return ret; } EXPORT_SYMBOL(generic_file_fsync); @@ -956,7 +964,7 @@ EXPORT_SYMBOL(generic_check_addressable); /* * No-op implementation of ->fsync for in-memory filesystems. */ -int noop_fsync(struct file *file, int datasync) +int noop_fsync(struct file *file, loff_t start, loff_t end, int datasync) { return 0; } diff --git a/fs/logfs/file.c b/fs/logfs/file.c index c2ad7028def4..b548c87a86f1 100644 --- a/fs/logfs/file.c +++ b/fs/logfs/file.c @@ -219,11 +219,20 @@ long logfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg) } } -int logfs_fsync(struct file *file, int datasync) +int logfs_fsync(struct file *file, loff_t start, loff_t end, int datasync) { struct super_block *sb = file->f_mapping->host->i_sb; + struct inode *inode = file->f_mapping->host; + int ret; + + ret = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (ret) + return ret; + mutex_lock(&inode->i_mutex); logfs_write_anchor(sb); + mutex_unlock(&inode->i_mutex); + return 0; } diff --git a/fs/logfs/logfs.h b/fs/logfs/logfs.h index 57afd4a6fabb..f22d108bfa5d 100644 --- a/fs/logfs/logfs.h +++ b/fs/logfs/logfs.h @@ -506,7 +506,7 @@ extern const struct file_operations logfs_reg_fops; extern const struct address_space_operations logfs_reg_aops; int logfs_readpage(struct file *file, struct page *page); long logfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg); -int logfs_fsync(struct file *file, int datasync); +int logfs_fsync(struct file *file, loff_t start, loff_t end, int datasync); /* gc.c */ u32 get_best_cand(struct super_block *sb, struct candidate_list *list, u32 *ec); diff --git a/fs/ncpfs/file.c b/fs/ncpfs/file.c index 0ed65e0c3dfe..64a326418aa2 100644 --- a/fs/ncpfs/file.c +++ b/fs/ncpfs/file.c @@ -20,9 +20,9 @@ #include "ncp_fs.h" -static int ncp_fsync(struct file *file, int datasync) +static int ncp_fsync(struct file *file, loff_t start, loff_t end, int datasync) { - return 0; + return filemap_write_and_wait_range(file->f_mapping, start, end); } /* diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index 8a45e6d1f6a4..57f578e2560a 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -56,7 +56,7 @@ static int nfs_link(struct dentry *, struct inode *, struct dentry *); static int nfs_mknod(struct inode *, struct dentry *, int, dev_t); static int nfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *); -static int nfs_fsync_dir(struct file *, int); +static int nfs_fsync_dir(struct file *, loff_t, loff_t, int); static loff_t nfs_llseek_dir(struct file *, loff_t, int); static void nfs_readdir_clear_array(struct page*); @@ -945,15 +945,19 @@ out: * All directory operations under NFS are synchronous, so fsync() * is a dummy operation. */ -static int nfs_fsync_dir(struct file *filp, int datasync) +static int nfs_fsync_dir(struct file *filp, loff_t start, loff_t end, + int datasync) { struct dentry *dentry = filp->f_path.dentry; + struct inode *inode = dentry->d_inode; dfprintk(FILE, "NFS: fsync dir(%s/%s) datasync %d\n", dentry->d_parent->d_name.name, dentry->d_name.name, datasync); + mutex_lock(&inode->i_mutex); nfs_inc_stats(dentry->d_inode, NFSIOS_VFSFSYNC); + mutex_unlock(&inode->i_mutex); return 0; } diff --git a/fs/nfs/file.c b/fs/nfs/file.c index 2c1705b6acd7..28b8c3f3cda3 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -55,7 +55,7 @@ static ssize_t nfs_file_splice_write(struct pipe_inode_info *pipe, static ssize_t nfs_file_write(struct kiocb *, const struct iovec *iov, unsigned long nr_segs, loff_t pos); static int nfs_file_flush(struct file *, fl_owner_t id); -static int nfs_file_fsync(struct file *, int datasync); +static int nfs_file_fsync(struct file *, loff_t, loff_t, int datasync); static int nfs_check_flags(int flags); static int nfs_lock(struct file *filp, int cmd, struct file_lock *fl); static int nfs_flock(struct file *filp, int cmd, struct file_lock *fl); @@ -308,7 +308,7 @@ nfs_file_mmap(struct file * file, struct vm_area_struct * vma) * fall back to doing a synchronous write. */ static int -nfs_file_fsync(struct file *file, int datasync) +nfs_file_fsync(struct file *file, loff_t start, loff_t end, int datasync) { struct dentry *dentry = file->f_path.dentry; struct nfs_open_context *ctx = nfs_file_open_context(file); @@ -316,11 +316,15 @@ nfs_file_fsync(struct file *file, int datasync) int have_error, status; int ret = 0; - dprintk("NFS: fsync file(%s/%s) datasync %d\n", dentry->d_parent->d_name.name, dentry->d_name.name, datasync); + ret = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (ret) + return ret; + mutex_lock(&inode->i_mutex); + nfs_inc_stats(inode, NFSIOS_VFSFSYNC); have_error = test_and_clear_bit(NFS_CONTEXT_ERROR_WRITE, &ctx->flags); status = nfs_commit_inode(inode, FLUSH_SYNC); @@ -332,6 +336,7 @@ nfs_file_fsync(struct file *file, int datasync) if (!ret && !datasync) /* application has asked for meta-data sync */ ret = pnfs_layoutcommit_inode(inode, true); + mutex_unlock(&inode->i_mutex); return ret; } diff --git a/fs/nilfs2/file.c b/fs/nilfs2/file.c index d7eeca62febd..26601529dc17 100644 --- a/fs/nilfs2/file.c +++ b/fs/nilfs2/file.c @@ -27,7 +27,7 @@ #include "nilfs.h" #include "segment.h" -int nilfs_sync_file(struct file *file, int datasync) +int nilfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync) { /* * Called from fsync() system call @@ -40,8 +40,15 @@ int nilfs_sync_file(struct file *file, int datasync) struct inode *inode = file->f_mapping->host; int err; - if (!nilfs_inode_dirty(inode)) + err = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (err) + return err; + mutex_lock(&inode->i_mutex); + + if (!nilfs_inode_dirty(inode)) { + mutex_unlock(&inode->i_mutex); return 0; + } if (datasync) err = nilfs_construct_dsync_segment(inode->i_sb, inode, 0, @@ -49,6 +56,7 @@ int nilfs_sync_file(struct file *file, int datasync) else err = nilfs_construct_segment(inode->i_sb); + mutex_unlock(&inode->i_mutex); return err; } diff --git a/fs/nilfs2/nilfs.h b/fs/nilfs2/nilfs.h index 6fb7511c0328..255d5e1c03b7 100644 --- a/fs/nilfs2/nilfs.h +++ b/fs/nilfs2/nilfs.h @@ -235,7 +235,7 @@ extern void nilfs_set_link(struct inode *, struct nilfs_dir_entry *, struct page *, struct inode *); /* file.c */ -extern int nilfs_sync_file(struct file *, int); +extern int nilfs_sync_file(struct file *, loff_t, loff_t, int); /* ioctl.c */ long nilfs_ioctl(struct file *, unsigned int, unsigned long); diff --git a/fs/ntfs/dir.c b/fs/ntfs/dir.c index 0f48e7c5d9e1..99e36107ff60 100644 --- a/fs/ntfs/dir.c +++ b/fs/ntfs/dir.c @@ -1527,13 +1527,20 @@ static int ntfs_dir_open(struct inode *vi, struct file *filp) * this problem for now. We do write the $BITMAP attribute if it is present * which is the important one for a directory so things are not too bad. */ -static int ntfs_dir_fsync(struct file *filp, int datasync) +static int ntfs_dir_fsync(struct file *filp, loff_t start, loff_t end, + int datasync) { struct inode *bmp_vi, *vi = filp->f_mapping->host; int err, ret; ntfs_attr na; ntfs_debug("Entering for inode 0x%lx.", vi->i_ino); + + err = filemap_write_and_wait_range(vi->i_mapping, start, end); + if (err) + return err; + mutex_lock(&vi->i_mutex); + BUG_ON(!S_ISDIR(vi->i_mode)); /* If the bitmap attribute inode is in memory sync it, too. */ na.mft_no = vi->i_ino; @@ -1555,6 +1562,7 @@ static int ntfs_dir_fsync(struct file *filp, int datasync) else ntfs_warning(vi->i_sb, "Failed to f%ssync inode 0x%lx. Error " "%u.", datasync ? "data" : "", vi->i_ino, -ret); + mutex_unlock(&vi->i_mutex); return ret; } diff --git a/fs/ntfs/file.c b/fs/ntfs/file.c index b59f5ac26bef..c587e2d27183 100644 --- a/fs/ntfs/file.c +++ b/fs/ntfs/file.c @@ -2152,12 +2152,19 @@ static ssize_t ntfs_file_aio_write(struct kiocb *iocb, const struct iovec *iov, * with this inode but since we have no simple way of getting to them we ignore * this problem for now. */ -static int ntfs_file_fsync(struct file *filp, int datasync) +static int ntfs_file_fsync(struct file *filp, loff_t start, loff_t end, + int datasync) { struct inode *vi = filp->f_mapping->host; int err, ret = 0; ntfs_debug("Entering for inode 0x%lx.", vi->i_ino); + + err = filemap_write_and_wait_range(vi->i_mapping, start, end); + if (err) + return err; + mutex_lock(&vi->i_mutex); + BUG_ON(S_ISDIR(vi->i_mode)); if (!datasync || !NInoNonResident(NTFS_I(vi))) ret = __ntfs_write_inode(vi, 1); @@ -2175,6 +2182,7 @@ static int ntfs_file_fsync(struct file *filp, int datasync) else ntfs_warning(vi->i_sb, "Failed to f%ssync inode 0x%lx. Error " "%u.", datasync ? "data" : "", vi->i_ino, -ret); + mutex_unlock(&vi->i_mutex); return ret; } diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index 22d604601957..0fc2bd34039d 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c @@ -171,7 +171,8 @@ static int ocfs2_dir_release(struct inode *inode, struct file *file) return 0; } -static int ocfs2_sync_file(struct file *file, int datasync) +static int ocfs2_sync_file(struct file *file, loff_t start, loff_t end, + int datasync) { int err = 0; journal_t *journal; @@ -184,6 +185,16 @@ static int ocfs2_sync_file(struct file *file, int datasync) file->f_path.dentry->d_name.name, (unsigned long long)datasync); + err = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (err) + return err; + + /* + * Probably don't need the i_mutex at all in here, just putting it here + * to be consistent with how fsync used to be called, someone more + * familiar with the fs could possibly remove it. + */ + mutex_lock(&inode->i_mutex); if (datasync && !(inode->i_state & I_DIRTY_DATASYNC)) { /* * We still have to flush drive's caches to get data to the @@ -200,6 +211,7 @@ static int ocfs2_sync_file(struct file *file, int datasync) bail: if (err) mlog_errno(err); + mutex_unlock(&inode->i_mutex); return (err < 0) ? -EIO : 0; } diff --git a/fs/reiserfs/dir.c b/fs/reiserfs/dir.c index 198dabf1b2bb..133e9355dc6f 100644 --- a/fs/reiserfs/dir.c +++ b/fs/reiserfs/dir.c @@ -14,7 +14,8 @@ extern const struct reiserfs_key MIN_KEY; static int reiserfs_readdir(struct file *, void *, filldir_t); -static int reiserfs_dir_fsync(struct file *filp, int datasync); +static int reiserfs_dir_fsync(struct file *filp, loff_t start, loff_t end, + int datasync); const struct file_operations reiserfs_dir_operations = { .llseek = generic_file_llseek, @@ -27,13 +28,21 @@ const struct file_operations reiserfs_dir_operations = { #endif }; -static int reiserfs_dir_fsync(struct file *filp, int datasync) +static int reiserfs_dir_fsync(struct file *filp, loff_t start, loff_t end, + int datasync) { struct inode *inode = filp->f_mapping->host; int err; + + err = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (err) + return err; + + mutex_lock(&inode->i_mutex); reiserfs_write_lock(inode->i_sb); err = reiserfs_commit_for_inode(inode); reiserfs_write_unlock(inode->i_sb); + mutex_unlock(&inode->i_mutex); if (err < 0) return err; return 0; diff --git a/fs/reiserfs/file.c b/fs/reiserfs/file.c index bbf31003d308..c7156dc39ce7 100644 --- a/fs/reiserfs/file.c +++ b/fs/reiserfs/file.c @@ -140,12 +140,18 @@ static void reiserfs_vfs_truncate_file(struct inode *inode) * be removed... */ -static int reiserfs_sync_file(struct file *filp, int datasync) +static int reiserfs_sync_file(struct file *filp, loff_t start, loff_t end, + int datasync) { struct inode *inode = filp->f_mapping->host; int err; int barrier_done; + err = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (err) + return err; + + mutex_lock(&inode->i_mutex); BUG_ON(!S_ISREG(inode->i_mode)); err = sync_mapping_buffers(inode->i_mapping); reiserfs_write_lock(inode->i_sb); @@ -153,6 +159,7 @@ static int reiserfs_sync_file(struct file *filp, int datasync) reiserfs_write_unlock(inode->i_sb); if (barrier_done != 1 && reiserfs_barrier_flush(inode->i_sb)) blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL, NULL); + mutex_unlock(&inode->i_mutex); if (barrier_done < 0) return barrier_done; return (err < 0) ? -EIO : 0; diff --git a/fs/sync.c b/fs/sync.c index c38ec163da6c..c98a7477edfd 100644 --- a/fs/sync.c +++ b/fs/sync.c @@ -165,28 +165,9 @@ SYSCALL_DEFINE1(syncfs, int, fd) */ int vfs_fsync_range(struct file *file, loff_t start, loff_t end, int datasync) { - struct address_space *mapping = file->f_mapping; - int err, ret; - - if (!file->f_op || !file->f_op->fsync) { - ret = -EINVAL; - goto out; - } - - ret = filemap_write_and_wait_range(mapping, start, end); - - /* - * We need to protect against concurrent writers, which could cause - * livelocks in fsync_buffers_list(). - */ - mutex_lock(&mapping->host->i_mutex); - err = file->f_op->fsync(file, datasync); - if (!ret) - ret = err; - mutex_unlock(&mapping->host->i_mutex); - -out: - return ret; + if (!file->f_op || !file->f_op->fsync) + return -EINVAL; + return file->f_op->fsync(file, start, end, datasync); } EXPORT_SYMBOL(vfs_fsync_range); diff --git a/fs/ubifs/file.c b/fs/ubifs/file.c index 5e7fccfc4b29..89ef9a2f7837 100644 --- a/fs/ubifs/file.c +++ b/fs/ubifs/file.c @@ -1304,7 +1304,7 @@ static void *ubifs_follow_link(struct dentry *dentry, struct nameidata *nd) return NULL; } -int ubifs_fsync(struct file *file, int datasync) +int ubifs_fsync(struct file *file, loff_t start, loff_t end, int datasync) { struct inode *inode = file->f_mapping->host; struct ubifs_info *c = inode->i_sb->s_fs_info; @@ -1319,14 +1319,16 @@ int ubifs_fsync(struct file *file, int datasync) */ return 0; - /* - * VFS has already synchronized dirty pages for this inode. Synchronize - * the inode unless this is a 'datasync()' call. - */ + err = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (err) + return err; + mutex_lock(&inode->i_mutex); + + /* Synchronize the inode unless this is a 'datasync()' call. */ if (!datasync || (inode->i_state & I_DIRTY_DATASYNC)) { err = inode->i_sb->s_op->write_inode(inode, NULL); if (err) - return err; + goto out; } /* @@ -1334,10 +1336,9 @@ int ubifs_fsync(struct file *file, int datasync) * them. */ err = ubifs_sync_wbufs_by_inode(c, inode); - if (err) - return err; - - return 0; +out: + mutex_unlock(&inode->i_mutex); + return err; } /** diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h index f79983d6f860..4cd648501fa4 100644 --- a/fs/ubifs/ubifs.h +++ b/fs/ubifs/ubifs.h @@ -1720,7 +1720,7 @@ const struct ubifs_lprops *ubifs_fast_find_frdi_idx(struct ubifs_info *c); int ubifs_calc_dark(const struct ubifs_info *c, int spc); /* file.c */ -int ubifs_fsync(struct file *file, int datasync); +int ubifs_fsync(struct file *file, loff_t start, loff_t end, int datasync); int ubifs_setattr(struct dentry *dentry, struct iattr *attr); /* dir.c */ diff --git a/fs/xfs/linux-2.6/xfs_file.c b/fs/xfs/linux-2.6/xfs_file.c index 7f782af286bf..fbbf657df0cd 100644 --- a/fs/xfs/linux-2.6/xfs_file.c +++ b/fs/xfs/linux-2.6/xfs_file.c @@ -127,6 +127,8 @@ xfs_iozero( STATIC int xfs_file_fsync( struct file *file, + loff_t start, + loff_t end, int datasync) { struct inode *inode = file->f_mapping->host; @@ -138,6 +140,10 @@ xfs_file_fsync( trace_xfs_file_fsync(ip); + error = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (error) + return error; + if (XFS_FORCED_SHUTDOWN(mp)) return -XFS_ERROR(EIO); @@ -875,18 +881,11 @@ xfs_file_aio_write( /* Handle various SYNC-type writes */ if ((file->f_flags & O_DSYNC) || IS_SYNC(inode)) { loff_t end = pos + ret - 1; - int error, error2; xfs_rw_iunlock(ip, iolock); - error = filemap_write_and_wait_range(mapping, pos, end); + ret = -xfs_file_fsync(file, pos, end, + (file->f_flags & __O_SYNC) ? 0 : 1); xfs_rw_ilock(ip, iolock); - - error2 = -xfs_file_fsync(file, - (file->f_flags & __O_SYNC) ? 0 : 1); - if (error) - ret = error; - else if (error2) - ret = error2; } out_unlock: diff --git a/include/linux/ext3_fs.h b/include/linux/ext3_fs.h index 5e06acf95d0f..0c473fd79acb 100644 --- a/include/linux/ext3_fs.h +++ b/include/linux/ext3_fs.h @@ -877,7 +877,7 @@ extern int ext3_htree_store_dirent(struct file *dir_file, __u32 hash, extern void ext3_htree_free_dir_info(struct dir_private_info *p); /* fsync.c */ -extern int ext3_sync_file(struct file *, int); +extern int ext3_sync_file(struct file *, loff_t, loff_t, int); /* hash.c */ extern int ext3fs_dirhash(const char *name, int len, struct diff --git a/include/linux/fb.h b/include/linux/fb.h index 6a8274877171..1d6836c498dd 100644 --- a/include/linux/fb.h +++ b/include/linux/fb.h @@ -1043,7 +1043,8 @@ extern void fb_deferred_io_open(struct fb_info *info, struct inode *inode, struct file *file); extern void fb_deferred_io_cleanup(struct fb_info *info); -extern int fb_deferred_io_fsync(struct file *file, int datasync); +extern int fb_deferred_io_fsync(struct file *file, loff_t start, + loff_t end, int datasync); static inline bool fb_be_math(struct fb_info *info) { diff --git a/include/linux/fs.h b/include/linux/fs.h index 4a61f98823a6..9cd2075c4a39 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1572,7 +1572,7 @@ struct file_operations { int (*open) (struct inode *, struct file *); int (*flush) (struct file *, fl_owner_t id); int (*release) (struct inode *, struct file *); - int (*fsync) (struct file *, int datasync); + int (*fsync) (struct file *, loff_t, loff_t, int datasync); int (*aio_fsync) (struct kiocb *, int datasync); int (*fasync) (int, struct file *, int); int (*lock) (struct file *, int, struct file_lock *); @@ -2360,7 +2360,8 @@ extern int generic_segment_checks(const struct iovec *iov, /* fs/block_dev.c */ extern ssize_t blkdev_aio_write(struct kiocb *iocb, const struct iovec *iov, unsigned long nr_segs, loff_t pos); -extern int blkdev_fsync(struct file *filp, int datasync); +extern int blkdev_fsync(struct file *filp, loff_t start, loff_t end, + int datasync); /* fs/splice.c */ extern ssize_t generic_file_splice_read(struct file *, loff_t *, @@ -2490,7 +2491,7 @@ extern int simple_link(struct dentry *, struct inode *, struct dentry *); extern int simple_unlink(struct inode *, struct dentry *); extern int simple_rmdir(struct inode *, struct dentry *); extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *); -extern int noop_fsync(struct file *, int); +extern int noop_fsync(struct file *, loff_t, loff_t, int); extern int simple_empty(struct dentry *); extern int simple_readpage(struct file *file, struct page *page); extern int simple_write_begin(struct file *file, struct address_space *mapping, @@ -2515,7 +2516,7 @@ extern ssize_t simple_read_from_buffer(void __user *to, size_t count, extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos, const void __user *from, size_t count); -extern int generic_file_fsync(struct file *, int); +extern int generic_file_fsync(struct file *, loff_t, loff_t, int); extern int generic_check_addressable(unsigned, u64); diff --git a/ipc/shm.c b/ipc/shm.c index ab3385a21b27..27884adb1a90 100644 --- a/ipc/shm.c +++ b/ipc/shm.c @@ -277,13 +277,13 @@ static int shm_release(struct inode *ino, struct file *file) return 0; } -static int shm_fsync(struct file *file, int datasync) +static int shm_fsync(struct file *file, loff_t start, loff_t end, int datasync) { struct shm_file_data *sfd = shm_file_data(file); if (!sfd->file->f_op->fsync) return -EINVAL; - return sfd->file->f_op->fsync(sfd->file, datasync); + return sfd->file->f_op->fsync(sfd->file, start, end, datasync); } static unsigned long shm_get_unmapped_area(struct file *file, -- cgit v1.2.3-58-ga151 From d2eece376648d2f7ba0a7d78f3c4d0421e608ac2 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Sun, 24 Jul 2011 10:21:54 -0700 Subject: ocfs2/cluster: Abort heartbeat start on hard-ro devices Currently if the heartbeat device is hard-ro, the o2hb thread keeps chugging along and dumping errors along the way. The user needs to manually stop the heartbeat. The patch addresses this shortcoming by adding a limit to the number of times the hb thread will iterate in an unsteady state. If the hb thread does not ready steady state in that many interation, the start is aborted. Signed-off-by: Sunil Mushran --- fs/ocfs2/cluster/heartbeat.c | 185 +++++++++++++++++++++++++++---------------- 1 file changed, 116 insertions(+), 69 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c index 9a3e6bbff27b..4728a746281d 100644 --- a/fs/ocfs2/cluster/heartbeat.c +++ b/fs/ocfs2/cluster/heartbeat.c @@ -216,6 +216,7 @@ struct o2hb_region { struct list_head hr_all_item; unsigned hr_unclean_stop:1, + hr_aborted_start:1, hr_item_pinned:1, hr_item_dropped:1; @@ -254,6 +255,10 @@ struct o2hb_region { * a more complete api that doesn't lead to this sort of fragility. */ atomic_t hr_steady_iterations; + /* terminate o2hb thread if it does not reach steady state + * (hr_steady_iterations == 0) within hr_unsteady_iterations */ + atomic_t hr_unsteady_iterations; + char hr_dev_name[BDEVNAME_SIZE]; unsigned int hr_timeout_ms; @@ -324,6 +329,10 @@ static void o2hb_write_timeout(struct work_struct *work) static void o2hb_arm_write_timeout(struct o2hb_region *reg) { + /* Arm writeout only after thread reaches steady state */ + if (atomic_read(®->hr_steady_iterations) != 0) + return; + mlog(ML_HEARTBEAT, "Queue write timeout for %u ms\n", O2HB_MAX_WRITE_TIMEOUT_MS); @@ -537,9 +546,14 @@ static int o2hb_verify_crc(struct o2hb_region *reg, return read == computed; } -/* We want to make sure that nobody is heartbeating on top of us -- - * this will help detect an invalid configuration. */ -static void o2hb_check_last_timestamp(struct o2hb_region *reg) +/* + * Compare the slot data with what we wrote in the last iteration. + * If the match fails, print an appropriate error message. This is to + * detect errors like... another node hearting on the same slot, + * flaky device that is losing writes, etc. + * Returns 1 if check succeeds, 0 otherwise. + */ +static int o2hb_check_own_slot(struct o2hb_region *reg) { struct o2hb_disk_slot *slot; struct o2hb_disk_heartbeat_block *hb_block; @@ -548,13 +562,13 @@ static void o2hb_check_last_timestamp(struct o2hb_region *reg) slot = ®->hr_slots[o2nm_this_node()]; /* Don't check on our 1st timestamp */ if (!slot->ds_last_time) - return; + return 0; hb_block = slot->ds_raw_block; if (le64_to_cpu(hb_block->hb_seq) == slot->ds_last_time && le64_to_cpu(hb_block->hb_generation) == slot->ds_last_generation && hb_block->hb_node == slot->ds_node_num) - return; + return 1; #define ERRSTR1 "Another node is heartbeating on device" #define ERRSTR2 "Heartbeat generation mismatch on device" @@ -574,6 +588,8 @@ static void o2hb_check_last_timestamp(struct o2hb_region *reg) (unsigned long long)slot->ds_last_time, hb_block->hb_node, (unsigned long long)le64_to_cpu(hb_block->hb_generation), (unsigned long long)le64_to_cpu(hb_block->hb_seq)); + + return 0; } static inline void o2hb_prepare_block(struct o2hb_region *reg, @@ -719,17 +735,24 @@ static void o2hb_shutdown_slot(struct o2hb_disk_slot *slot) o2nm_node_put(node); } -static void o2hb_set_quorum_device(struct o2hb_region *reg, - struct o2hb_disk_slot *slot) +static void o2hb_set_quorum_device(struct o2hb_region *reg) { - assert_spin_locked(&o2hb_live_lock); - if (!o2hb_global_heartbeat_active()) return; - if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap)) + /* Prevent race with o2hb_heartbeat_group_drop_item() */ + if (kthread_should_stop()) + return; + + /* Tag region as quorum only after thread reaches steady state */ + if (atomic_read(®->hr_steady_iterations) != 0) return; + spin_lock(&o2hb_live_lock); + + if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap)) + goto unlock; + /* * A region can be added to the quorum only when it sees all * live nodes heartbeat on it. In other words, the region has been @@ -737,13 +760,10 @@ static void o2hb_set_quorum_device(struct o2hb_region *reg, */ if (memcmp(reg->hr_live_node_bitmap, o2hb_live_node_bitmap, sizeof(o2hb_live_node_bitmap))) - return; - - if (slot->ds_changed_samples < O2HB_LIVE_THRESHOLD) - return; + goto unlock; - printk(KERN_NOTICE "o2hb: Region %s is now a quorum device\n", - config_item_name(®->hr_item)); + printk(KERN_NOTICE "o2hb: Region %s (%s) is now a quorum device\n", + config_item_name(®->hr_item), reg->hr_dev_name); set_bit(reg->hr_region_num, o2hb_quorum_region_bitmap); @@ -754,6 +774,8 @@ static void o2hb_set_quorum_device(struct o2hb_region *reg, if (o2hb_pop_count(&o2hb_quorum_region_bitmap, O2NM_MAX_REGIONS) > O2HB_PIN_CUT_OFF) o2hb_region_unpin(NULL); +unlock: + spin_unlock(&o2hb_live_lock); } static int o2hb_check_slot(struct o2hb_region *reg, @@ -925,8 +947,6 @@ fire_callbacks: slot->ds_equal_samples = 0; } out: - o2hb_set_quorum_device(reg, slot); - spin_unlock(&o2hb_live_lock); o2hb_run_event_list(&event); @@ -957,7 +977,8 @@ static int o2hb_highest_node(unsigned long *nodes, static int o2hb_do_disk_heartbeat(struct o2hb_region *reg) { - int i, ret, highest_node, change = 0; + int i, ret, highest_node; + int membership_change = 0, own_slot_ok = 0; unsigned long configured_nodes[BITS_TO_LONGS(O2NM_MAX_NODES)]; unsigned long live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)]; struct o2hb_bio_wait_ctxt write_wc; @@ -966,7 +987,7 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg) sizeof(configured_nodes)); if (ret) { mlog_errno(ret); - return ret; + goto bail; } /* @@ -982,8 +1003,9 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg) highest_node = o2hb_highest_node(configured_nodes, O2NM_MAX_NODES); if (highest_node >= O2NM_MAX_NODES) { - mlog(ML_NOTICE, "ocfs2_heartbeat: no configured nodes found!\n"); - return -EINVAL; + mlog(ML_NOTICE, "o2hb: No configured nodes found!\n"); + ret = -EINVAL; + goto bail; } /* No sense in reading the slots of nodes that don't exist @@ -993,29 +1015,27 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg) ret = o2hb_read_slots(reg, highest_node + 1); if (ret < 0) { mlog_errno(ret); - return ret; + goto bail; } /* With an up to date view of the slots, we can check that no * other node has been improperly configured to heartbeat in * our slot. */ - o2hb_check_last_timestamp(reg); + own_slot_ok = o2hb_check_own_slot(reg); /* fill in the proper info for our next heartbeat */ o2hb_prepare_block(reg, reg->hr_generation); - /* And fire off the write. Note that we don't wait on this I/O - * until later. */ ret = o2hb_issue_node_write(reg, &write_wc); if (ret < 0) { mlog_errno(ret); - return ret; + goto bail; } i = -1; while((i = find_next_bit(configured_nodes, O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) { - change |= o2hb_check_slot(reg, ®->hr_slots[i]); + membership_change |= o2hb_check_slot(reg, ®->hr_slots[i]); } /* @@ -1030,18 +1050,39 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg) * disk */ mlog(ML_ERROR, "Write error %d on device \"%s\"\n", write_wc.wc_error, reg->hr_dev_name); - return write_wc.wc_error; + ret = write_wc.wc_error; + goto bail; } - o2hb_arm_write_timeout(reg); + /* Skip disarming the timeout if own slot has stale/bad data */ + if (own_slot_ok) { + o2hb_set_quorum_device(reg); + o2hb_arm_write_timeout(reg); + } +bail: /* let the person who launched us know when things are steady */ - if (!change && (atomic_read(®->hr_steady_iterations) != 0)) { - if (atomic_dec_and_test(®->hr_steady_iterations)) + if (atomic_read(®->hr_steady_iterations) != 0) { + if (!ret && own_slot_ok && !membership_change) { + if (atomic_dec_and_test(®->hr_steady_iterations)) + wake_up(&o2hb_steady_queue); + } + } + + if (atomic_read(®->hr_steady_iterations) != 0) { + if (atomic_dec_and_test(®->hr_unsteady_iterations)) { + printk(KERN_NOTICE "o2hb: Unable to stabilize " + "heartbeart on region %s (%s)\n", + config_item_name(®->hr_item), + reg->hr_dev_name); + atomic_set(®->hr_steady_iterations, 0); + reg->hr_aborted_start = 1; wake_up(&o2hb_steady_queue); + ret = -EIO; + } } - return 0; + return ret; } /* Subtract b from a, storing the result in a. a *must* have a larger @@ -1095,7 +1136,8 @@ static int o2hb_thread(void *data) /* Pin node */ o2nm_depend_this_node(); - while (!kthread_should_stop() && !reg->hr_unclean_stop) { + while (!kthread_should_stop() && + !reg->hr_unclean_stop && !reg->hr_aborted_start) { /* We track the time spent inside * o2hb_do_disk_heartbeat so that we avoid more than * hr_timeout_ms between disk writes. On busy systems @@ -1103,10 +1145,7 @@ static int o2hb_thread(void *data) * likely to time itself out. */ do_gettimeofday(&before_hb); - i = 0; - do { - ret = o2hb_do_disk_heartbeat(reg); - } while (ret && ++i < 2); + ret = o2hb_do_disk_heartbeat(reg); do_gettimeofday(&after_hb); elapsed_msec = o2hb_elapsed_msecs(&before_hb, &after_hb); @@ -1117,7 +1156,8 @@ static int o2hb_thread(void *data) after_hb.tv_sec, (unsigned long) after_hb.tv_usec, elapsed_msec); - if (elapsed_msec < reg->hr_timeout_ms) { + if (!kthread_should_stop() && + elapsed_msec < reg->hr_timeout_ms) { /* the kthread api has blocked signals for us so no * need to record the return value. */ msleep_interruptible(reg->hr_timeout_ms - elapsed_msec); @@ -1134,20 +1174,20 @@ static int o2hb_thread(void *data) * to timeout on this region when we could just as easily * write a clear generation - thus indicating to them that * this node has left this region. - * - * XXX: Should we skip this on unclean_stop? */ - o2hb_prepare_block(reg, 0); - ret = o2hb_issue_node_write(reg, &write_wc); - if (ret == 0) { - o2hb_wait_on_io(reg, &write_wc); - } else { - mlog_errno(ret); + */ + if (!reg->hr_unclean_stop && !reg->hr_aborted_start) { + o2hb_prepare_block(reg, 0); + ret = o2hb_issue_node_write(reg, &write_wc); + if (ret == 0) + o2hb_wait_on_io(reg, &write_wc); + else + mlog_errno(ret); } /* Unpin node */ o2nm_undepend_this_node(); - mlog(ML_HEARTBEAT|ML_KTHREAD, "hb thread exiting\n"); + mlog(ML_HEARTBEAT|ML_KTHREAD, "o2hb thread exiting\n"); return 0; } @@ -1426,6 +1466,8 @@ static void o2hb_region_release(struct config_item *item) struct page *page; struct o2hb_region *reg = to_o2hb_region(item); + mlog(ML_HEARTBEAT, "hb region release (%s)\n", reg->hr_dev_name); + if (reg->hr_tmp_block) kfree(reg->hr_tmp_block); @@ -1792,7 +1834,10 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg, live_threshold <<= 1; spin_unlock(&o2hb_live_lock); } - atomic_set(®->hr_steady_iterations, live_threshold + 1); + ++live_threshold; + atomic_set(®->hr_steady_iterations, live_threshold); + /* unsteady_iterations is double the steady_iterations */ + atomic_set(®->hr_unsteady_iterations, (live_threshold << 1)); hb_task = kthread_run(o2hb_thread, reg, "o2hb-%s", reg->hr_item.ci_name); @@ -1809,14 +1854,12 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg, ret = wait_event_interruptible(o2hb_steady_queue, atomic_read(®->hr_steady_iterations) == 0); if (ret) { - /* We got interrupted (hello ptrace!). Clean up */ - spin_lock(&o2hb_live_lock); - hb_task = reg->hr_task; - reg->hr_task = NULL; - spin_unlock(&o2hb_live_lock); + atomic_set(®->hr_steady_iterations, 0); + reg->hr_aborted_start = 1; + } - if (hb_task) - kthread_stop(hb_task); + if (reg->hr_aborted_start) { + ret = -EIO; goto out; } @@ -1833,8 +1876,8 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg, ret = -EIO; if (hb_task && o2hb_global_heartbeat_active()) - printk(KERN_NOTICE "o2hb: Heartbeat started on region %s\n", - config_item_name(®->hr_item)); + printk(KERN_NOTICE "o2hb: Heartbeat started on region %s (%s)\n", + config_item_name(®->hr_item), reg->hr_dev_name); out: if (filp) @@ -2092,13 +2135,6 @@ static void o2hb_heartbeat_group_drop_item(struct config_group *group, /* stop the thread when the user removes the region dir */ spin_lock(&o2hb_live_lock); - if (o2hb_global_heartbeat_active()) { - clear_bit(reg->hr_region_num, o2hb_region_bitmap); - clear_bit(reg->hr_region_num, o2hb_live_region_bitmap); - if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap)) - quorum_region = 1; - clear_bit(reg->hr_region_num, o2hb_quorum_region_bitmap); - } hb_task = reg->hr_task; reg->hr_task = NULL; reg->hr_item_dropped = 1; @@ -2107,19 +2143,30 @@ static void o2hb_heartbeat_group_drop_item(struct config_group *group, if (hb_task) kthread_stop(hb_task); + if (o2hb_global_heartbeat_active()) { + spin_lock(&o2hb_live_lock); + clear_bit(reg->hr_region_num, o2hb_region_bitmap); + clear_bit(reg->hr_region_num, o2hb_live_region_bitmap); + if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap)) + quorum_region = 1; + clear_bit(reg->hr_region_num, o2hb_quorum_region_bitmap); + spin_unlock(&o2hb_live_lock); + printk(KERN_NOTICE "o2hb: Heartbeat %s on region %s (%s)\n", + ((atomic_read(®->hr_steady_iterations) == 0) ? + "stopped" : "start aborted"), config_item_name(item), + reg->hr_dev_name); + } + /* * If we're racing a dev_write(), we need to wake them. They will * check reg->hr_task */ if (atomic_read(®->hr_steady_iterations) != 0) { + reg->hr_aborted_start = 1; atomic_set(®->hr_steady_iterations, 0); wake_up(&o2hb_steady_queue); } - if (o2hb_global_heartbeat_active()) - printk(KERN_NOTICE "o2hb: Heartbeat stopped on region %s\n", - config_item_name(®->hr_item)); - config_item_put(item); if (!o2hb_global_heartbeat_active() || !quorum_region) -- cgit v1.2.3-58-ga151 From 1dfecf810e0eacb35987905082f23e5c2cd26e91 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Sun, 24 Jul 2011 10:22:54 -0700 Subject: ocfs2/cluster: Clean up messages in o2net o2net messages needed a facelift. Signed-off-by: Sunil Mushran --- fs/ocfs2/cluster/tcp.c | 119 ++++++++++++++++++++++--------------------------- 1 file changed, 53 insertions(+), 66 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c index db5ee4b4f47a..6e97895ecf24 100644 --- a/fs/ocfs2/cluster/tcp.c +++ b/fs/ocfs2/cluster/tcp.c @@ -545,7 +545,7 @@ static void o2net_set_nn_state(struct o2net_node *nn, } if (was_valid && !valid) { - printk(KERN_NOTICE "o2net: no longer connected to " + printk(KERN_NOTICE "o2net: No longer connected to " SC_NODEF_FMT "\n", SC_NODEF_ARGS(old_sc)); o2net_complete_nodes_nsw(nn); } @@ -555,7 +555,7 @@ static void o2net_set_nn_state(struct o2net_node *nn, cancel_delayed_work(&nn->nn_connect_expired); printk(KERN_NOTICE "o2net: %s " SC_NODEF_FMT "\n", o2nm_this_node() > sc->sc_node->nd_num ? - "connected to" : "accepted connection from", + "Connected to" : "Accepted connection from", SC_NODEF_ARGS(sc)); } @@ -643,7 +643,7 @@ static void o2net_state_change(struct sock *sk) o2net_sc_queue_work(sc, &sc->sc_connect_work); break; default: - printk(KERN_INFO "o2net: connection to " SC_NODEF_FMT + printk(KERN_INFO "o2net: Connection to " SC_NODEF_FMT " shutdown, state %d\n", SC_NODEF_ARGS(sc), sk->sk_state); o2net_sc_queue_work(sc, &sc->sc_shutdown_work); @@ -1284,11 +1284,11 @@ static int o2net_check_handshake(struct o2net_sock_container *sc) struct o2net_node *nn = o2net_nn_from_num(sc->sc_node->nd_num); if (hand->protocol_version != cpu_to_be64(O2NET_PROTOCOL_VERSION)) { - mlog(ML_NOTICE, SC_NODEF_FMT " advertised net protocol " - "version %llu but %llu is required, disconnecting\n", - SC_NODEF_ARGS(sc), - (unsigned long long)be64_to_cpu(hand->protocol_version), - O2NET_PROTOCOL_VERSION); + printk(KERN_NOTICE "o2net: " SC_NODEF_FMT " Advertised net " + "protocol version %llu but %llu is required. " + "Disconnecting.\n", SC_NODEF_ARGS(sc), + (unsigned long long)be64_to_cpu(hand->protocol_version), + O2NET_PROTOCOL_VERSION); /* don't bother reconnecting if its the wrong version. */ o2net_ensure_shutdown(nn, sc, -ENOTCONN); @@ -1302,33 +1302,33 @@ static int o2net_check_handshake(struct o2net_sock_container *sc) */ if (be32_to_cpu(hand->o2net_idle_timeout_ms) != o2net_idle_timeout()) { - mlog(ML_NOTICE, SC_NODEF_FMT " uses a network idle timeout of " - "%u ms, but we use %u ms locally. disconnecting\n", - SC_NODEF_ARGS(sc), - be32_to_cpu(hand->o2net_idle_timeout_ms), - o2net_idle_timeout()); + printk(KERN_NOTICE "o2net: " SC_NODEF_FMT " uses a network " + "idle timeout of %u ms, but we use %u ms locally. " + "Disconnecting.\n", SC_NODEF_ARGS(sc), + be32_to_cpu(hand->o2net_idle_timeout_ms), + o2net_idle_timeout()); o2net_ensure_shutdown(nn, sc, -ENOTCONN); return -1; } if (be32_to_cpu(hand->o2net_keepalive_delay_ms) != o2net_keepalive_delay()) { - mlog(ML_NOTICE, SC_NODEF_FMT " uses a keepalive delay of " - "%u ms, but we use %u ms locally. disconnecting\n", - SC_NODEF_ARGS(sc), - be32_to_cpu(hand->o2net_keepalive_delay_ms), - o2net_keepalive_delay()); + printk(KERN_NOTICE "o2net: " SC_NODEF_FMT " uses a keepalive " + "delay of %u ms, but we use %u ms locally. " + "Disconnecting.\n", SC_NODEF_ARGS(sc), + be32_to_cpu(hand->o2net_keepalive_delay_ms), + o2net_keepalive_delay()); o2net_ensure_shutdown(nn, sc, -ENOTCONN); return -1; } if (be32_to_cpu(hand->o2hb_heartbeat_timeout_ms) != O2HB_MAX_WRITE_TIMEOUT_MS) { - mlog(ML_NOTICE, SC_NODEF_FMT " uses a heartbeat timeout of " - "%u ms, but we use %u ms locally. disconnecting\n", - SC_NODEF_ARGS(sc), - be32_to_cpu(hand->o2hb_heartbeat_timeout_ms), - O2HB_MAX_WRITE_TIMEOUT_MS); + printk(KERN_NOTICE "o2net: " SC_NODEF_FMT " uses a heartbeat " + "timeout of %u ms, but we use %u ms locally. " + "Disconnecting.\n", SC_NODEF_ARGS(sc), + be32_to_cpu(hand->o2hb_heartbeat_timeout_ms), + O2HB_MAX_WRITE_TIMEOUT_MS); o2net_ensure_shutdown(nn, sc, -ENOTCONN); return -1; } @@ -1539,28 +1539,16 @@ static void o2net_idle_timer(unsigned long data) { struct o2net_sock_container *sc = (struct o2net_sock_container *)data; struct o2net_node *nn = o2net_nn_from_num(sc->sc_node->nd_num); - #ifdef CONFIG_DEBUG_FS - ktime_t now = ktime_get(); + unsigned long msecs = ktime_to_ms(ktime_get()) - + ktime_to_ms(sc->sc_tv_timer); +#else + unsigned long msecs = o2net_idle_timeout(); #endif - printk(KERN_NOTICE "o2net: connection to " SC_NODEF_FMT " has been idle for %u.%u " - "seconds, shutting it down.\n", SC_NODEF_ARGS(sc), - o2net_idle_timeout() / 1000, - o2net_idle_timeout() % 1000); - -#ifdef CONFIG_DEBUG_FS - mlog(ML_NOTICE, "Here are some times that might help debug the " - "situation: (Timer: %lld, Now %lld, DataReady %lld, Advance %lld-%lld, " - "Key 0x%08x, Func %u, FuncTime %lld-%lld)\n", - (long long)ktime_to_us(sc->sc_tv_timer), (long long)ktime_to_us(now), - (long long)ktime_to_us(sc->sc_tv_data_ready), - (long long)ktime_to_us(sc->sc_tv_advance_start), - (long long)ktime_to_us(sc->sc_tv_advance_stop), - sc->sc_msg_key, sc->sc_msg_type, - (long long)ktime_to_us(sc->sc_tv_func_start), - (long long)ktime_to_us(sc->sc_tv_func_stop)); -#endif + printk(KERN_NOTICE "o2net: Connection to " SC_NODEF_FMT " has been " + "idle for %lu.%lu secs, shutting it down.\n", SC_NODEF_ARGS(sc), + msecs / 1000, msecs % 1000); /* * Initialize the nn_timeout so that the next connection attempt @@ -1693,8 +1681,8 @@ static void o2net_start_connect(struct work_struct *work) out: if (ret) { - mlog(ML_NOTICE, "connect attempt to " SC_NODEF_FMT " failed " - "with errno %d\n", SC_NODEF_ARGS(sc), ret); + printk(KERN_NOTICE "o2net: Connect attempt to " SC_NODEF_FMT + " failed with errno %d\n", SC_NODEF_ARGS(sc), ret); /* 0 err so that another will be queued and attempted * from set_nn_state */ if (sc) @@ -1717,8 +1705,8 @@ static void o2net_connect_expired(struct work_struct *work) spin_lock(&nn->nn_lock); if (!nn->nn_sc_valid) { - mlog(ML_ERROR, "no connection established with node %u after " - "%u.%u seconds, giving up and returning errors.\n", + printk(KERN_NOTICE "o2net: No connection established with " + "node %u after %u.%u seconds, giving up.\n", o2net_num_from_nn(nn), o2net_idle_timeout() / 1000, o2net_idle_timeout() % 1000); @@ -1861,21 +1849,21 @@ static int o2net_accept_one(struct socket *sock) node = o2nm_get_node_by_ip(sin.sin_addr.s_addr); if (node == NULL) { - mlog(ML_NOTICE, "attempt to connect from unknown node at %pI4:%d\n", - &sin.sin_addr.s_addr, ntohs(sin.sin_port)); + printk(KERN_NOTICE "o2net: Attempt to connect from unknown " + "node at %pI4:%d\n", &sin.sin_addr.s_addr, + ntohs(sin.sin_port)); ret = -EINVAL; goto out; } if (o2nm_this_node() >= node->nd_num) { local_node = o2nm_get_node_by_num(o2nm_this_node()); - mlog(ML_NOTICE, "unexpected connect attempt seen at node '%s' (" - "%u, %pI4:%d) from node '%s' (%u, %pI4:%d)\n", - local_node->nd_name, local_node->nd_num, - &(local_node->nd_ipv4_address), - ntohs(local_node->nd_ipv4_port), - node->nd_name, node->nd_num, &sin.sin_addr.s_addr, - ntohs(sin.sin_port)); + printk(KERN_NOTICE "o2net: Unexpected connect attempt seen " + "at node '%s' (%u, %pI4:%d) from node '%s' (%u, " + "%pI4:%d)\n", local_node->nd_name, local_node->nd_num, + &(local_node->nd_ipv4_address), + ntohs(local_node->nd_ipv4_port), node->nd_name, + node->nd_num, &sin.sin_addr.s_addr, ntohs(sin.sin_port)); ret = -EINVAL; goto out; } @@ -1900,10 +1888,10 @@ static int o2net_accept_one(struct socket *sock) ret = 0; spin_unlock(&nn->nn_lock); if (ret) { - mlog(ML_NOTICE, "attempt to connect from node '%s' at " - "%pI4:%d but it already has an open connection\n", - node->nd_name, &sin.sin_addr.s_addr, - ntohs(sin.sin_port)); + printk(KERN_NOTICE "o2net: Attempt to connect from node '%s' " + "at %pI4:%d but it already has an open connection\n", + node->nd_name, &sin.sin_addr.s_addr, + ntohs(sin.sin_port)); goto out; } @@ -1983,7 +1971,7 @@ static int o2net_open_listening_sock(__be32 addr, __be16 port) ret = sock_create(PF_INET, SOCK_STREAM, IPPROTO_TCP, &sock); if (ret < 0) { - mlog(ML_ERROR, "unable to create socket, ret=%d\n", ret); + printk(KERN_ERR "o2net: Error %d while creating socket\n", ret); goto out; } @@ -2000,16 +1988,15 @@ static int o2net_open_listening_sock(__be32 addr, __be16 port) sock->sk->sk_reuse = 1; ret = sock->ops->bind(sock, (struct sockaddr *)&sin, sizeof(sin)); if (ret < 0) { - mlog(ML_ERROR, "unable to bind socket at %pI4:%u, " - "ret=%d\n", &addr, ntohs(port), ret); + printk(KERN_ERR "o2net: Error %d while binding socket at " + "%pI4:%u\n", ret, &addr, ntohs(port)); goto out; } ret = sock->ops->listen(sock, 64); - if (ret < 0) { - mlog(ML_ERROR, "unable to listen on %pI4:%u, ret=%d\n", - &addr, ntohs(port), ret); - } + if (ret < 0) + printk(KERN_ERR "o2net: Error %d while listening on %pI4:%u\n", + ret, &addr, ntohs(port)); out: if (ret) { -- cgit v1.2.3-58-ga151 From 8decab3c8dadcdf4f54ffb30df6e6f67b398b6e0 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Sun, 24 Jul 2011 10:23:54 -0700 Subject: ocfs2/dlm: Clean up messages in o2dlm o2dlm messages needed a facelift. Signed-off-by: Sunil Mushran --- fs/ocfs2/dlm/dlmdomain.c | 18 +++++++++------- fs/ocfs2/dlm/dlmlock.c | 22 ++++++++----------- fs/ocfs2/dlm/dlmmaster.c | 46 ++++++++++++++++++---------------------- fs/ocfs2/dlm/dlmrecovery.c | 53 +++++++++++++++++++++++++++------------------- fs/ocfs2/dlm/dlmthread.c | 2 -- 5 files changed, 71 insertions(+), 70 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/dlm/dlmdomain.c b/fs/ocfs2/dlm/dlmdomain.c index 6ed6b95dcf93..ce225fb0cea7 100644 --- a/fs/ocfs2/dlm/dlmdomain.c +++ b/fs/ocfs2/dlm/dlmdomain.c @@ -539,17 +539,17 @@ again: static void __dlm_print_nodes(struct dlm_ctxt *dlm) { - int node = -1; + int node = -1, num = 0; assert_spin_locked(&dlm->spinlock); - printk(KERN_NOTICE "o2dlm: Nodes in domain %s: ", dlm->name); - + printk("( "); while ((node = find_next_bit(dlm->domain_map, O2NM_MAX_NODES, node + 1)) < O2NM_MAX_NODES) { printk("%d ", node); + ++num; } - printk("\n"); + printk(") %u nodes\n", num); } static int dlm_exit_domain_handler(struct o2net_msg *msg, u32 len, void *data, @@ -566,11 +566,10 @@ static int dlm_exit_domain_handler(struct o2net_msg *msg, u32 len, void *data, node = exit_msg->node_idx; - printk(KERN_NOTICE "o2dlm: Node %u leaves domain %s\n", node, dlm->name); - spin_lock(&dlm->spinlock); clear_bit(node, dlm->domain_map); clear_bit(node, dlm->exit_domain_map); + printk(KERN_NOTICE "o2dlm: Node %u leaves domain %s ", node, dlm->name); __dlm_print_nodes(dlm); /* notify anything attached to the heartbeat events */ @@ -755,6 +754,7 @@ void dlm_unregister_domain(struct dlm_ctxt *dlm) dlm_mark_domain_leaving(dlm); dlm_leave_domain(dlm); + printk(KERN_NOTICE "o2dlm: Leaving domain %s\n", dlm->name); dlm_force_free_mles(dlm); dlm_complete_dlm_shutdown(dlm); } @@ -970,7 +970,7 @@ static int dlm_assert_joined_handler(struct o2net_msg *msg, u32 len, void *data, clear_bit(assert->node_idx, dlm->exit_domain_map); __dlm_set_joining_node(dlm, DLM_LOCK_RES_OWNER_UNKNOWN); - printk(KERN_NOTICE "o2dlm: Node %u joins domain %s\n", + printk(KERN_NOTICE "o2dlm: Node %u joins domain %s ", assert->node_idx, dlm->name); __dlm_print_nodes(dlm); @@ -1701,8 +1701,10 @@ static int dlm_try_to_join_domain(struct dlm_ctxt *dlm) bail: spin_lock(&dlm->spinlock); __dlm_set_joining_node(dlm, DLM_LOCK_RES_OWNER_UNKNOWN); - if (!status) + if (!status) { + printk(KERN_NOTICE "o2dlm: Joining domain %s ", dlm->name); __dlm_print_nodes(dlm); + } spin_unlock(&dlm->spinlock); if (ctxt) { diff --git a/fs/ocfs2/dlm/dlmlock.c b/fs/ocfs2/dlm/dlmlock.c index 8d39e0fd66f7..c7f3e22bda1e 100644 --- a/fs/ocfs2/dlm/dlmlock.c +++ b/fs/ocfs2/dlm/dlmlock.c @@ -319,27 +319,23 @@ static enum dlm_status dlm_send_remote_lock_request(struct dlm_ctxt *dlm, tmpret = o2net_send_message(DLM_CREATE_LOCK_MSG, dlm->key, &create, sizeof(create), res->owner, &status); if (tmpret >= 0) { - // successfully sent and received - ret = status; // this is already a dlm_status + ret = status; if (ret == DLM_REJECTED) { - mlog(ML_ERROR, "%s:%.*s: BUG. this is a stale lockres " - "no longer owned by %u. that node is coming back " - "up currently.\n", dlm->name, create.namelen, + mlog(ML_ERROR, "%s: res %.*s, Stale lockres no longer " + "owned by node %u. That node is coming back up " + "currently.\n", dlm->name, create.namelen, create.name, res->owner); dlm_print_one_lock_resource(res); BUG(); } } else { - mlog(ML_ERROR, "Error %d when sending message %u (key 0x%x) to " - "node %u\n", tmpret, DLM_CREATE_LOCK_MSG, dlm->key, - res->owner); - if (dlm_is_host_down(tmpret)) { + mlog(ML_ERROR, "%s: res %.*s, Error %d send CREATE LOCK to " + "node %u\n", dlm->name, create.namelen, create.name, + tmpret, res->owner); + if (dlm_is_host_down(tmpret)) ret = DLM_RECOVERING; - mlog(0, "node %u died so returning DLM_RECOVERING " - "from lock message!\n", res->owner); - } else { + else ret = dlm_err_to_dlm_status(tmpret); - } } return ret; diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c index 11eefb8c12e9..9f3b093b629a 100644 --- a/fs/ocfs2/dlm/dlmmaster.c +++ b/fs/ocfs2/dlm/dlmmaster.c @@ -829,8 +829,8 @@ lookup: * but they might own this lockres. wait on them. */ bit = find_next_bit(dlm->recovery_map, O2NM_MAX_NODES, 0); if (bit < O2NM_MAX_NODES) { - mlog(ML_NOTICE, "%s:%.*s: at least one node (%d) to " - "recover before lock mastery can begin\n", + mlog(0, "%s: res %.*s, At least one node (%d) " + "to recover before lock mastery can begin\n", dlm->name, namelen, (char *)lockid, bit); wait_on_recovery = 1; } @@ -864,8 +864,8 @@ redo_request: * dlm spinlock would be detectable be a change on the mle, * so we only need to clear out the recovery map once. */ if (dlm_is_recovery_lock(lockid, namelen)) { - mlog(ML_NOTICE, "%s: recovery map is not empty, but " - "must master $RECOVERY lock now\n", dlm->name); + mlog(0, "%s: Recovery map is not empty, but must " + "master $RECOVERY lock now\n", dlm->name); if (!dlm_pre_master_reco_lockres(dlm, res)) wait_on_recovery = 0; else { @@ -883,8 +883,8 @@ redo_request: spin_lock(&dlm->spinlock); bit = find_next_bit(dlm->recovery_map, O2NM_MAX_NODES, 0); if (bit < O2NM_MAX_NODES) { - mlog(ML_NOTICE, "%s:%.*s: at least one node (%d) to " - "recover before lock mastery can begin\n", + mlog(0, "%s: res %.*s, At least one node (%d) " + "to recover before lock mastery can begin\n", dlm->name, namelen, (char *)lockid, bit); wait_on_recovery = 1; } else @@ -913,8 +913,8 @@ redo_request: * yet, keep going until it does. this is how the * master will know that asserts are needed back to * the lower nodes. */ - mlog(0, "%s:%.*s: requests only up to %u but master " - "is %u, keep going\n", dlm->name, namelen, + mlog(0, "%s: res %.*s, Requests only up to %u but " + "master is %u, keep going\n", dlm->name, namelen, lockid, nodenum, mle->master); } } @@ -924,13 +924,12 @@ wait: ret = dlm_wait_for_lock_mastery(dlm, res, mle, &blocked); if (ret < 0) { wait_on_recovery = 1; - mlog(0, "%s:%.*s: node map changed, redo the " - "master request now, blocked=%d\n", - dlm->name, res->lockname.len, + mlog(0, "%s: res %.*s, Node map changed, redo the master " + "request now, blocked=%d\n", dlm->name, res->lockname.len, res->lockname.name, blocked); if (++tries > 20) { - mlog(ML_ERROR, "%s:%.*s: spinning on " - "dlm_wait_for_lock_mastery, blocked=%d\n", + mlog(ML_ERROR, "%s: res %.*s, Spinning on " + "dlm_wait_for_lock_mastery, blocked = %d\n", dlm->name, res->lockname.len, res->lockname.name, blocked); dlm_print_one_lock_resource(res); @@ -940,7 +939,8 @@ wait: goto redo_request; } - mlog(0, "lockres mastered by %u\n", res->owner); + mlog(0, "%s: res %.*s, Mastered by %u\n", dlm->name, res->lockname.len, + res->lockname.name, res->owner); /* make sure we never continue without this */ BUG_ON(res->owner == O2NM_MAX_NODES); @@ -2187,8 +2187,6 @@ int dlm_drop_lockres_ref(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) namelen = res->lockname.len; BUG_ON(namelen > O2NM_MAX_NAME_LEN); - mlog(0, "%s:%.*s: sending deref to %d\n", - dlm->name, namelen, lockname, res->owner); memset(&deref, 0, sizeof(deref)); deref.node_idx = dlm->node_num; deref.namelen = namelen; @@ -2197,14 +2195,12 @@ int dlm_drop_lockres_ref(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) ret = o2net_send_message(DLM_DEREF_LOCKRES_MSG, dlm->key, &deref, sizeof(deref), res->owner, &r); if (ret < 0) - mlog(ML_ERROR, "Error %d when sending message %u (key 0x%x) to " - "node %u\n", ret, DLM_DEREF_LOCKRES_MSG, dlm->key, - res->owner); + mlog(ML_ERROR, "%s: res %.*s, error %d send DEREF to node %u\n", + dlm->name, namelen, lockname, ret, res->owner); else if (r < 0) { /* BAD. other node says I did not have a ref. */ - mlog(ML_ERROR,"while dropping ref on %s:%.*s " - "(master=%u) got %d.\n", dlm->name, namelen, - lockname, res->owner, r); + mlog(ML_ERROR, "%s: res %.*s, DEREF to node %u got %d\n", + dlm->name, namelen, lockname, res->owner, r); dlm_print_one_lock_resource(res); BUG(); } @@ -2916,9 +2912,9 @@ static int dlm_do_migrate_request(struct dlm_ctxt *dlm, &migrate, sizeof(migrate), nodenum, &status); if (ret < 0) { - mlog(ML_ERROR, "Error %d when sending message %u (key " - "0x%x) to node %u\n", ret, DLM_MIGRATE_REQUEST_MSG, - dlm->key, nodenum); + mlog(ML_ERROR, "%s: res %.*s, Error %d send " + "MIGRATE_REQUEST to node %u\n", dlm->name, + migrate.namelen, migrate.name, ret, nodenum); if (!dlm_is_host_down(ret)) { mlog(ML_ERROR, "unhandled error=%d!\n", ret); BUG(); diff --git a/fs/ocfs2/dlm/dlmrecovery.c b/fs/ocfs2/dlm/dlmrecovery.c index 7efab6d28a21..a3c312c43b90 100644 --- a/fs/ocfs2/dlm/dlmrecovery.c +++ b/fs/ocfs2/dlm/dlmrecovery.c @@ -430,6 +430,8 @@ static void dlm_begin_recovery(struct dlm_ctxt *dlm) { spin_lock(&dlm->spinlock); BUG_ON(dlm->reco.state & DLM_RECO_STATE_ACTIVE); + printk(KERN_NOTICE "o2dlm: Begin recovery on domain %s for node %u\n", + dlm->name, dlm->reco.dead_node); dlm->reco.state |= DLM_RECO_STATE_ACTIVE; spin_unlock(&dlm->spinlock); } @@ -440,9 +442,18 @@ static void dlm_end_recovery(struct dlm_ctxt *dlm) BUG_ON(!(dlm->reco.state & DLM_RECO_STATE_ACTIVE)); dlm->reco.state &= ~DLM_RECO_STATE_ACTIVE; spin_unlock(&dlm->spinlock); + printk(KERN_NOTICE "o2dlm: End recovery on domain %s\n", dlm->name); wake_up(&dlm->reco.event); } +static void dlm_print_recovery_master(struct dlm_ctxt *dlm) +{ + printk(KERN_NOTICE "o2dlm: Node %u (%s) is the Recovery Master for the " + "dead node %u in domain %s\n", dlm->reco.new_master, + (dlm->node_num == dlm->reco.new_master ? "me" : "he"), + dlm->reco.dead_node, dlm->name); +} + static int dlm_do_recovery(struct dlm_ctxt *dlm) { int status = 0; @@ -505,9 +516,8 @@ static int dlm_do_recovery(struct dlm_ctxt *dlm) } mlog(0, "another node will master this recovery session.\n"); } - mlog(0, "dlm=%s (%d), new_master=%u, this node=%u, dead_node=%u\n", - dlm->name, task_pid_nr(dlm->dlm_reco_thread_task), dlm->reco.new_master, - dlm->node_num, dlm->reco.dead_node); + + dlm_print_recovery_master(dlm); /* it is safe to start everything back up here * because all of the dead node's lock resources @@ -518,15 +528,13 @@ static int dlm_do_recovery(struct dlm_ctxt *dlm) return 0; master_here: - mlog(ML_NOTICE, "(%d) Node %u is the Recovery Master for the Dead Node " - "%u for Domain %s\n", task_pid_nr(dlm->dlm_reco_thread_task), - dlm->node_num, dlm->reco.dead_node, dlm->name); + dlm_print_recovery_master(dlm); status = dlm_remaster_locks(dlm, dlm->reco.dead_node); if (status < 0) { /* we should never hit this anymore */ - mlog(ML_ERROR, "error %d remastering locks for node %u, " - "retrying.\n", status, dlm->reco.dead_node); + mlog(ML_ERROR, "%s: Error %d remastering locks for node %u, " + "retrying.\n", dlm->name, status, dlm->reco.dead_node); /* yield a bit to allow any final network messages * to get handled on remaining nodes */ msleep(100); @@ -567,7 +575,7 @@ static int dlm_remaster_locks(struct dlm_ctxt *dlm, u8 dead_node) BUG_ON(ndata->state != DLM_RECO_NODE_DATA_INIT); ndata->state = DLM_RECO_NODE_DATA_REQUESTING; - mlog(0, "requesting lock info from node %u\n", + mlog(0, "%s: Requesting lock info from node %u\n", dlm->name, ndata->node_num); if (ndata->node_num == dlm->node_num) { @@ -640,7 +648,7 @@ static int dlm_remaster_locks(struct dlm_ctxt *dlm, u8 dead_node) spin_unlock(&dlm_reco_state_lock); } - mlog(0, "done requesting all lock info\n"); + mlog(0, "%s: Done requesting all lock info\n", dlm->name); /* nodes should be sending reco data now * just need to wait */ @@ -802,10 +810,9 @@ static int dlm_request_all_locks(struct dlm_ctxt *dlm, u8 request_from, /* negative status is handled by caller */ if (ret < 0) - mlog(ML_ERROR, "Error %d when sending message %u (key " - "0x%x) to node %u\n", ret, DLM_LOCK_REQUEST_MSG, - dlm->key, request_from); - + mlog(ML_ERROR, "%s: Error %d send LOCK_REQUEST to node %u " + "to recover dead node %u\n", dlm->name, ret, + request_from, dead_node); // return from here, then // sleep until all received or error return ret; @@ -956,9 +963,9 @@ static int dlm_send_all_done_msg(struct dlm_ctxt *dlm, u8 dead_node, u8 send_to) ret = o2net_send_message(DLM_RECO_DATA_DONE_MSG, dlm->key, &done_msg, sizeof(done_msg), send_to, &tmpret); if (ret < 0) { - mlog(ML_ERROR, "Error %d when sending message %u (key " - "0x%x) to node %u\n", ret, DLM_RECO_DATA_DONE_MSG, - dlm->key, send_to); + mlog(ML_ERROR, "%s: Error %d send RECO_DATA_DONE to node %u " + "to recover dead node %u\n", dlm->name, ret, send_to, + dead_node); if (!dlm_is_host_down(ret)) { BUG(); } @@ -1127,9 +1134,11 @@ static int dlm_send_mig_lockres_msg(struct dlm_ctxt *dlm, if (ret < 0) { /* XXX: negative status is not handled. * this will end up killing this node. */ - mlog(ML_ERROR, "Error %d when sending message %u (key " - "0x%x) to node %u\n", ret, DLM_MIG_LOCKRES_MSG, - dlm->key, send_to); + mlog(ML_ERROR, "%s: res %.*s, Error %d send MIG_LOCKRES to " + "node %u (%s)\n", dlm->name, mres->lockname_len, + mres->lockname, ret, send_to, + (orig_flags & DLM_MRES_MIGRATION ? + "migration" : "recovery")); } else { /* might get an -ENOMEM back here */ ret = status; @@ -2324,9 +2333,9 @@ static void dlm_do_local_recovery_cleanup(struct dlm_ctxt *dlm, u8 dead_node) dlm_revalidate_lvb(dlm, res, dead_node); if (res->owner == dead_node) { if (res->state & DLM_LOCK_RES_DROPPING_REF) { - mlog(ML_NOTICE, "Ignore %.*s for " + mlog(ML_NOTICE, "%s: res %.*s, Skip " "recovery as it is being freed\n", - res->lockname.len, + dlm->name, res->lockname.len, res->lockname.name); } else dlm_move_lockres_to_recovery_list(dlm, diff --git a/fs/ocfs2/dlm/dlmthread.c b/fs/ocfs2/dlm/dlmthread.c index 1d6d1d22c471..46faac20f16c 100644 --- a/fs/ocfs2/dlm/dlmthread.c +++ b/fs/ocfs2/dlm/dlmthread.c @@ -185,8 +185,6 @@ static void dlm_purge_lockres(struct dlm_ctxt *dlm, /* clear our bit from the master's refmap, ignore errors */ ret = dlm_drop_lockres_ref(dlm, res); if (ret < 0) { - mlog(ML_ERROR, "%s: deref %.*s failed %d\n", dlm->name, - res->lockname.len, res->lockname.name, ret); if (!dlm_is_host_down(ret)) BUG(); } -- cgit v1.2.3-58-ga151 From 394eb3d38a3ecc549cc34a3040103a9164be516b Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Sun, 24 Jul 2011 10:24:54 -0700 Subject: ocfs2: Clean up messages in stack_o2cb.c o2cb messages needed a facelift. Signed-off-by: Sunil Mushran --- fs/ocfs2/stack_o2cb.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/stack_o2cb.c b/fs/ocfs2/stack_o2cb.c index 19965b00c43c..be935d89a4e8 100644 --- a/fs/ocfs2/stack_o2cb.c +++ b/fs/ocfs2/stack_o2cb.c @@ -263,8 +263,8 @@ static void o2dlm_eviction_cb(int node_num, void *data) { struct ocfs2_cluster_connection *conn = data; - mlog(ML_NOTICE, "o2dlm has evicted node %d from group %.*s\n", - node_num, conn->cc_namelen, conn->cc_name); + printk(KERN_NOTICE "o2cb: o2dlm has evicted node %d from domain %.*s\n", + node_num, conn->cc_namelen, conn->cc_name); conn->cc_recovery_handler(node_num, conn->cc_recovery_data); } -- cgit v1.2.3-58-ga151 From 0afbba13226fcdbd4327f6b13a42f6efbb8c9caf Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Sun, 24 Jul 2011 10:25:54 -0700 Subject: ocfs2/dlm: Cleanup up dlm_finish_local_lockres_recovery() dlm_finish_local_lockres_recovery() needed a facelift. Signed-off-by: Sunil Mushran --- fs/ocfs2/dlm/dlmrecovery.c | 57 ++++++++++++++++++++-------------------------- 1 file changed, 25 insertions(+), 32 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/dlm/dlmrecovery.c b/fs/ocfs2/dlm/dlmrecovery.c index a3c312c43b90..41e74f0e2d35 100644 --- a/fs/ocfs2/dlm/dlmrecovery.c +++ b/fs/ocfs2/dlm/dlmrecovery.c @@ -2093,6 +2093,9 @@ static void dlm_finish_local_lockres_recovery(struct dlm_ctxt *dlm, list_for_each_entry_safe(res, next, &dlm->reco.resources, recovering) { if (res->owner == dead_node) { + mlog(0, "%s: res %.*s, Changing owner from %u to %u\n", + dlm->name, res->lockname.len, res->lockname.name, + res->owner, new_master); list_del_init(&res->recovering); spin_lock(&res->spinlock); /* new_master has our reference from @@ -2114,40 +2117,30 @@ static void dlm_finish_local_lockres_recovery(struct dlm_ctxt *dlm, for (i = 0; i < DLM_HASH_BUCKETS; i++) { bucket = dlm_lockres_hash(dlm, i); hlist_for_each_entry(res, hash_iter, bucket, hash_node) { - if (res->state & DLM_LOCK_RES_RECOVERING) { - if (res->owner == dead_node) { - mlog(0, "(this=%u) res %.*s owner=%u " - "was not on recovering list, but " - "clearing state anyway\n", - dlm->node_num, res->lockname.len, - res->lockname.name, new_master); - } else if (res->owner == dlm->node_num) { - mlog(0, "(this=%u) res %.*s owner=%u " - "was not on recovering list, " - "owner is THIS node, clearing\n", - dlm->node_num, res->lockname.len, - res->lockname.name, new_master); - } else - continue; + if (!(res->state & DLM_LOCK_RES_RECOVERING)) + continue; - if (!list_empty(&res->recovering)) { - mlog(0, "%s:%.*s: lockres was " - "marked RECOVERING, owner=%u\n", - dlm->name, res->lockname.len, - res->lockname.name, res->owner); - list_del_init(&res->recovering); - dlm_lockres_put(res); - } - spin_lock(&res->spinlock); - /* new_master has our reference from - * the lock state sent during recovery */ - dlm_change_lockres_owner(dlm, res, new_master); - res->state &= ~DLM_LOCK_RES_RECOVERING; - if (__dlm_lockres_has_locks(res)) - __dlm_dirty_lockres(dlm, res); - spin_unlock(&res->spinlock); - wake_up(&res->wq); + if (res->owner != dead_node && + res->owner != dlm->node_num) + continue; + + if (!list_empty(&res->recovering)) { + list_del_init(&res->recovering); + dlm_lockres_put(res); } + + /* new_master has our reference from + * the lock state sent during recovery */ + mlog(0, "%s: res %.*s, Changing owner from %u to %u\n", + dlm->name, res->lockname.len, res->lockname.name, + res->owner, new_master); + spin_lock(&res->spinlock); + dlm_change_lockres_owner(dlm, res, new_master); + res->state &= ~DLM_LOCK_RES_RECOVERING; + if (__dlm_lockres_has_locks(res)) + __dlm_dirty_lockres(dlm, res); + spin_unlock(&res->spinlock); + wake_up(&res->wq); } } } -- cgit v1.2.3-58-ga151 From 8d400b81cc83b171ff872587723a37eb7fae9abe Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Sun, 24 Jul 2011 10:26:54 -0700 Subject: ocfs2/dlm: Clean up refmap helpers Patch cleans up helpers that set/clear refmap bits and grab/drop inflight lock ref counts. Signed-off-by: Sunil Mushran --- fs/ocfs2/dlm/dlmcommon.h | 49 +++++--------------------- fs/ocfs2/dlm/dlmmaster.c | 88 ++++++++++++++++++++++++++++------------------ fs/ocfs2/dlm/dlmrecovery.c | 8 ++--- 3 files changed, 66 insertions(+), 79 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/dlm/dlmcommon.h b/fs/ocfs2/dlm/dlmcommon.h index d602abb51b61..4a5d5804b130 100644 --- a/fs/ocfs2/dlm/dlmcommon.h +++ b/fs/ocfs2/dlm/dlmcommon.h @@ -902,46 +902,15 @@ struct dlm_lock_resource *dlm_new_lockres(struct dlm_ctxt *dlm, const char *name, unsigned int namelen); -#define dlm_lockres_set_refmap_bit(bit,res) \ - __dlm_lockres_set_refmap_bit(bit,res,__FILE__,__LINE__) -#define dlm_lockres_clear_refmap_bit(bit,res) \ - __dlm_lockres_clear_refmap_bit(bit,res,__FILE__,__LINE__) - -static inline void __dlm_lockres_set_refmap_bit(int bit, - struct dlm_lock_resource *res, - const char *file, - int line) -{ - //printk("%s:%d:%.*s: setting bit %d\n", file, line, - // res->lockname.len, res->lockname.name, bit); - set_bit(bit, res->refmap); -} - -static inline void __dlm_lockres_clear_refmap_bit(int bit, - struct dlm_lock_resource *res, - const char *file, - int line) -{ - //printk("%s:%d:%.*s: clearing bit %d\n", file, line, - // res->lockname.len, res->lockname.name, bit); - clear_bit(bit, res->refmap); -} - -void __dlm_lockres_drop_inflight_ref(struct dlm_ctxt *dlm, - struct dlm_lock_resource *res, - const char *file, - int line); -void __dlm_lockres_grab_inflight_ref(struct dlm_ctxt *dlm, - struct dlm_lock_resource *res, - int new_lockres, - const char *file, - int line); -#define dlm_lockres_drop_inflight_ref(d,r) \ - __dlm_lockres_drop_inflight_ref(d,r,__FILE__,__LINE__) -#define dlm_lockres_grab_inflight_ref(d,r) \ - __dlm_lockres_grab_inflight_ref(d,r,0,__FILE__,__LINE__) -#define dlm_lockres_grab_inflight_ref_new(d,r) \ - __dlm_lockres_grab_inflight_ref(d,r,1,__FILE__,__LINE__) +void dlm_lockres_set_refmap_bit(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res, int bit); +void dlm_lockres_clear_refmap_bit(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res, int bit); + +void dlm_lockres_drop_inflight_ref(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res); +void dlm_lockres_grab_inflight_ref(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res); void dlm_queue_ast(struct dlm_ctxt *dlm, struct dlm_lock *lock); void dlm_queue_bast(struct dlm_ctxt *dlm, struct dlm_lock *lock); diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c index 9f3b093b629a..11e446ffb67a 100644 --- a/fs/ocfs2/dlm/dlmmaster.c +++ b/fs/ocfs2/dlm/dlmmaster.c @@ -631,39 +631,58 @@ error: return NULL; } -void __dlm_lockres_grab_inflight_ref(struct dlm_ctxt *dlm, - struct dlm_lock_resource *res, - int new_lockres, - const char *file, - int line) +void dlm_lockres_set_refmap_bit(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res, int bit) { - if (!new_lockres) - assert_spin_locked(&res->spinlock); + assert_spin_locked(&res->spinlock); + + mlog(0, "res %.*s, set node %u, %ps()\n", res->lockname.len, + res->lockname.name, bit, __builtin_return_address(0)); + + set_bit(bit, res->refmap); +} + +void dlm_lockres_clear_refmap_bit(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res, int bit) +{ + assert_spin_locked(&res->spinlock); + + mlog(0, "res %.*s, clr node %u, %ps()\n", res->lockname.len, + res->lockname.name, bit, __builtin_return_address(0)); + + clear_bit(bit, res->refmap); +} + + +void dlm_lockres_grab_inflight_ref(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res) +{ + assert_spin_locked(&res->spinlock); if (!test_bit(dlm->node_num, res->refmap)) { BUG_ON(res->inflight_locks != 0); - dlm_lockres_set_refmap_bit(dlm->node_num, res); + dlm_lockres_set_refmap_bit(dlm, res, dlm->node_num); } res->inflight_locks++; - mlog(0, "%s:%.*s: inflight++: now %u\n", - dlm->name, res->lockname.len, res->lockname.name, - res->inflight_locks); + mlog(0, "%s: res %.*s, inflight++: now %u, %ps()\n", dlm->name, + res->lockname.len, res->lockname.name, res->inflight_locks, + __builtin_return_address(0)); } -void __dlm_lockres_drop_inflight_ref(struct dlm_ctxt *dlm, - struct dlm_lock_resource *res, - const char *file, - int line) +void dlm_lockres_drop_inflight_ref(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res) { assert_spin_locked(&res->spinlock); BUG_ON(res->inflight_locks == 0); + res->inflight_locks--; - mlog(0, "%s:%.*s: inflight--: now %u\n", - dlm->name, res->lockname.len, res->lockname.name, - res->inflight_locks); + mlog(0, "%s: res %.*s, inflight--: now %u, %ps()\n", dlm->name, + res->lockname.len, res->lockname.name, res->inflight_locks, + __builtin_return_address(0)); + if (res->inflight_locks == 0) - dlm_lockres_clear_refmap_bit(dlm->node_num, res); + dlm_lockres_clear_refmap_bit(dlm, res, dlm->node_num); wake_up(&res->wq); } @@ -843,8 +862,10 @@ lookup: /* finally add the lockres to its hash bucket */ __dlm_insert_lockres(dlm, res); - /* since this lockres is new it doesn't not require the spinlock */ - dlm_lockres_grab_inflight_ref_new(dlm, res); + + spin_lock(&res->spinlock); + dlm_lockres_grab_inflight_ref(dlm, res); + spin_unlock(&res->spinlock); /* if this node does not become the master make sure to drop * this inflight reference below */ @@ -1426,9 +1447,7 @@ way_up_top: } if (res->owner == dlm->node_num) { - mlog(0, "%s:%.*s: setting bit %u in refmap\n", - dlm->name, namelen, name, request->node_idx); - dlm_lockres_set_refmap_bit(request->node_idx, res); + dlm_lockres_set_refmap_bit(dlm, res, request->node_idx); spin_unlock(&res->spinlock); response = DLM_MASTER_RESP_YES; if (mle) @@ -1493,10 +1512,8 @@ way_up_top: * go back and clean the mles on any * other nodes */ dispatch_assert = 1; - dlm_lockres_set_refmap_bit(request->node_idx, res); - mlog(0, "%s:%.*s: setting bit %u in refmap\n", - dlm->name, namelen, name, - request->node_idx); + dlm_lockres_set_refmap_bit(dlm, res, + request->node_idx); } else response = DLM_MASTER_RESP_NO; } else { @@ -1702,7 +1719,7 @@ again: "lockres, set the bit in the refmap\n", namelen, lockname, to); spin_lock(&res->spinlock); - dlm_lockres_set_refmap_bit(to, res); + dlm_lockres_set_refmap_bit(dlm, res, to); spin_unlock(&res->spinlock); } } @@ -2256,7 +2273,7 @@ int dlm_deref_lockres_handler(struct o2net_msg *msg, u32 len, void *data, else { BUG_ON(res->state & DLM_LOCK_RES_DROPPING_REF); if (test_bit(node, res->refmap)) { - dlm_lockres_clear_refmap_bit(node, res); + dlm_lockres_clear_refmap_bit(dlm, res, node); cleared = 1; } } @@ -2316,7 +2333,7 @@ static void dlm_deref_lockres_worker(struct dlm_work_item *item, void *data) BUG_ON(res->state & DLM_LOCK_RES_DROPPING_REF); if (test_bit(node, res->refmap)) { __dlm_wait_on_lockres_flags(res, DLM_LOCK_RES_SETREF_INPROG); - dlm_lockres_clear_refmap_bit(node, res); + dlm_lockres_clear_refmap_bit(dlm, res, node); cleared = 1; } spin_unlock(&res->spinlock); @@ -2798,7 +2815,8 @@ static void dlm_remove_nonlocal_locks(struct dlm_ctxt *dlm, BUG_ON(!list_empty(&lock->bast_list)); BUG_ON(lock->ast_pending); BUG_ON(lock->bast_pending); - dlm_lockres_clear_refmap_bit(lock->ml.node, res); + dlm_lockres_clear_refmap_bit(dlm, res, + lock->ml.node); list_del_init(&lock->list); dlm_lock_put(lock); /* In a normal unlock, we would have added a @@ -2819,7 +2837,7 @@ static void dlm_remove_nonlocal_locks(struct dlm_ctxt *dlm, mlog(0, "%s:%.*s: node %u had a ref to this " "migrating lockres, clearing\n", dlm->name, res->lockname.len, res->lockname.name, bit); - dlm_lockres_clear_refmap_bit(bit, res); + dlm_lockres_clear_refmap_bit(dlm, res, bit); } bit++; } @@ -2933,7 +2951,7 @@ static int dlm_do_migrate_request(struct dlm_ctxt *dlm, dlm->name, res->lockname.len, res->lockname.name, nodenum); spin_lock(&res->spinlock); - dlm_lockres_set_refmap_bit(nodenum, res); + dlm_lockres_set_refmap_bit(dlm, res, nodenum); spin_unlock(&res->spinlock); } } @@ -3267,7 +3285,7 @@ int dlm_finish_migration(struct dlm_ctxt *dlm, struct dlm_lock_resource *res, * mastery reference here since old_master will briefly have * a reference after the migration completes */ spin_lock(&res->spinlock); - dlm_lockres_set_refmap_bit(old_master, res); + dlm_lockres_set_refmap_bit(dlm, res, old_master); spin_unlock(&res->spinlock); mlog(0, "now time to do a migrate request to other nodes\n"); diff --git a/fs/ocfs2/dlm/dlmrecovery.c b/fs/ocfs2/dlm/dlmrecovery.c index 41e74f0e2d35..f1a3aa84ad89 100644 --- a/fs/ocfs2/dlm/dlmrecovery.c +++ b/fs/ocfs2/dlm/dlmrecovery.c @@ -1776,7 +1776,7 @@ static int dlm_process_recovery_data(struct dlm_ctxt *dlm, dlm->name, mres->lockname_len, mres->lockname, from); spin_lock(&res->spinlock); - dlm_lockres_set_refmap_bit(from, res); + dlm_lockres_set_refmap_bit(dlm, res, from); spin_unlock(&res->spinlock); added++; break; @@ -1974,7 +1974,7 @@ skip_lvb: mlog(0, "%s:%.*s: added lock for node %u, " "setting refmap bit\n", dlm->name, res->lockname.len, res->lockname.name, ml->node); - dlm_lockres_set_refmap_bit(ml->node, res); + dlm_lockres_set_refmap_bit(dlm, res, ml->node); added++; } spin_unlock(&res->spinlock); @@ -2254,12 +2254,12 @@ static void dlm_free_dead_locks(struct dlm_ctxt *dlm, res->lockname.len, res->lockname.name, freed, dead_node); __dlm_print_one_lock_resource(res); } - dlm_lockres_clear_refmap_bit(dead_node, res); + dlm_lockres_clear_refmap_bit(dlm, res, dead_node); } else if (test_bit(dead_node, res->refmap)) { mlog(0, "%s:%.*s: dead node %u had a ref, but had " "no locks and had not purged before dying\n", dlm->name, res->lockname.len, res->lockname.name, dead_node); - dlm_lockres_clear_refmap_bit(dead_node, res); + dlm_lockres_clear_refmap_bit(dlm, res, dead_node); } /* do not kick thread yet */ -- cgit v1.2.3-58-ga151 From e9f0b6a6233105c188064328c3f15e21942d0241 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Sun, 24 Jul 2011 10:27:54 -0700 Subject: ocfs2/dlm: Trace insert/remove of resource to/from hash Add mlog to trace adding and removing the resource from/to the hash table. Signed-off-by: Sunil Mushran --- fs/ocfs2/dlm/dlmcommon.h | 5 ++--- fs/ocfs2/dlm/dlmdomain.c | 19 ++++++++++++------- fs/ocfs2/dlm/dlmthread.c | 2 +- 3 files changed, 15 insertions(+), 11 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/dlm/dlmcommon.h b/fs/ocfs2/dlm/dlmcommon.h index 4a5d5804b130..b1748ac74751 100644 --- a/fs/ocfs2/dlm/dlmcommon.h +++ b/fs/ocfs2/dlm/dlmcommon.h @@ -877,9 +877,8 @@ static inline void dlm_lockres_get(struct dlm_lock_resource *res) kref_get(&res->refs); } void dlm_lockres_put(struct dlm_lock_resource *res); -void __dlm_unhash_lockres(struct dlm_lock_resource *res); -void __dlm_insert_lockres(struct dlm_ctxt *dlm, - struct dlm_lock_resource *res); +void __dlm_unhash_lockres(struct dlm_ctxt *dlm, struct dlm_lock_resource *res); +void __dlm_insert_lockres(struct dlm_ctxt *dlm, struct dlm_lock_resource *res); struct dlm_lock_resource * __dlm_lookup_lockres_full(struct dlm_ctxt *dlm, const char *name, unsigned int len, diff --git a/fs/ocfs2/dlm/dlmdomain.c b/fs/ocfs2/dlm/dlmdomain.c index ce225fb0cea7..200d9888ffee 100644 --- a/fs/ocfs2/dlm/dlmdomain.c +++ b/fs/ocfs2/dlm/dlmdomain.c @@ -157,16 +157,18 @@ static int dlm_protocol_compare(struct dlm_protocol_version *existing, static void dlm_unregister_domain_handlers(struct dlm_ctxt *dlm); -void __dlm_unhash_lockres(struct dlm_lock_resource *lockres) +void __dlm_unhash_lockres(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) { - if (!hlist_unhashed(&lockres->hash_node)) { - hlist_del_init(&lockres->hash_node); - dlm_lockres_put(lockres); - } + if (hlist_unhashed(&res->hash_node)) + return; + + mlog(0, "%s: Unhash res %.*s\n", dlm->name, res->lockname.len, + res->lockname.name); + hlist_del_init(&res->hash_node); + dlm_lockres_put(res); } -void __dlm_insert_lockres(struct dlm_ctxt *dlm, - struct dlm_lock_resource *res) +void __dlm_insert_lockres(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) { struct hlist_head *bucket; struct qstr *q; @@ -180,6 +182,9 @@ void __dlm_insert_lockres(struct dlm_ctxt *dlm, dlm_lockres_get(res); hlist_add_head(&res->hash_node, bucket); + + mlog(0, "%s: Hash res %.*s\n", dlm->name, res->lockname.len, + res->lockname.name); } struct dlm_lock_resource * __dlm_lookup_lockres_full(struct dlm_ctxt *dlm, diff --git a/fs/ocfs2/dlm/dlmthread.c b/fs/ocfs2/dlm/dlmthread.c index 46faac20f16c..4eff65e7e59e 100644 --- a/fs/ocfs2/dlm/dlmthread.c +++ b/fs/ocfs2/dlm/dlmthread.c @@ -207,7 +207,7 @@ static void dlm_purge_lockres(struct dlm_ctxt *dlm, BUG(); } - __dlm_unhash_lockres(res); + __dlm_unhash_lockres(dlm, res); /* lockres is not in the hash now. drop the flag and wake up * any processes waiting in dlm_get_lock_resource. */ -- cgit v1.2.3-58-ga151 From ed8625c6fb93d750ed022db571a8a7b7a6724b3b Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Sun, 24 Jul 2011 10:28:54 -0700 Subject: ocfs2/dlm: Cleanup dlm_wait_for_node_death() and dlm_wait_for_node_recovery() dlm_wait_for_node_death() and dlm_wait_for_node_recovery() needed a facelift. Signed-off-by: Sunil Mushran --- fs/ocfs2/dlm/dlmcommon.h | 4 ++-- fs/ocfs2/dlm/dlmrecovery.c | 46 ++++++++++++++++++++++------------------------ 2 files changed, 24 insertions(+), 26 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/dlm/dlmcommon.h b/fs/ocfs2/dlm/dlmcommon.h index b1748ac74751..a5952ceecba5 100644 --- a/fs/ocfs2/dlm/dlmcommon.h +++ b/fs/ocfs2/dlm/dlmcommon.h @@ -859,8 +859,8 @@ void dlm_complete_recovery_thread(struct dlm_ctxt *dlm); void dlm_wait_for_recovery(struct dlm_ctxt *dlm); void dlm_kick_recovery_thread(struct dlm_ctxt *dlm); int dlm_is_node_dead(struct dlm_ctxt *dlm, u8 node); -int dlm_wait_for_node_death(struct dlm_ctxt *dlm, u8 node, int timeout); -int dlm_wait_for_node_recovery(struct dlm_ctxt *dlm, u8 node, int timeout); +void dlm_wait_for_node_death(struct dlm_ctxt *dlm, u8 node, int timeout); +void dlm_wait_for_node_recovery(struct dlm_ctxt *dlm, u8 node, int timeout); void dlm_put(struct dlm_ctxt *dlm); struct dlm_ctxt *dlm_grab(struct dlm_ctxt *dlm); diff --git a/fs/ocfs2/dlm/dlmrecovery.c b/fs/ocfs2/dlm/dlmrecovery.c index f1a3aa84ad89..01ebfd0bdad7 100644 --- a/fs/ocfs2/dlm/dlmrecovery.c +++ b/fs/ocfs2/dlm/dlmrecovery.c @@ -362,40 +362,38 @@ static int dlm_is_node_recovered(struct dlm_ctxt *dlm, u8 node) } -int dlm_wait_for_node_death(struct dlm_ctxt *dlm, u8 node, int timeout) +void dlm_wait_for_node_death(struct dlm_ctxt *dlm, u8 node, int timeout) { - if (timeout) { - mlog(ML_NOTICE, "%s: waiting %dms for notification of " - "death of node %u\n", dlm->name, timeout, node); + if (dlm_is_node_dead(dlm, node)) + return; + + printk(KERN_NOTICE "o2dlm: Waiting on the death of node %u in " + "domain %s\n", node, dlm->name); + + if (timeout) wait_event_timeout(dlm->dlm_reco_thread_wq, - dlm_is_node_dead(dlm, node), - msecs_to_jiffies(timeout)); - } else { - mlog(ML_NOTICE, "%s: waiting indefinitely for notification " - "of death of node %u\n", dlm->name, node); + dlm_is_node_dead(dlm, node), + msecs_to_jiffies(timeout)); + else wait_event(dlm->dlm_reco_thread_wq, dlm_is_node_dead(dlm, node)); - } - /* for now, return 0 */ - return 0; } -int dlm_wait_for_node_recovery(struct dlm_ctxt *dlm, u8 node, int timeout) +void dlm_wait_for_node_recovery(struct dlm_ctxt *dlm, u8 node, int timeout) { - if (timeout) { - mlog(0, "%s: waiting %dms for notification of " - "recovery of node %u\n", dlm->name, timeout, node); + if (dlm_is_node_recovered(dlm, node)) + return; + + printk(KERN_NOTICE "o2dlm: Waiting on the recovery of node %u in " + "domain %s\n", node, dlm->name); + + if (timeout) wait_event_timeout(dlm->dlm_reco_thread_wq, - dlm_is_node_recovered(dlm, node), - msecs_to_jiffies(timeout)); - } else { - mlog(0, "%s: waiting indefinitely for notification " - "of recovery of node %u\n", dlm->name, node); + dlm_is_node_recovered(dlm, node), + msecs_to_jiffies(timeout)); + else wait_event(dlm->dlm_reco_thread_wq, dlm_is_node_recovered(dlm, node)); - } - /* for now, return 0 */ - return 0; } /* callers of the top-level api calls (dlmlock/dlmunlock) should -- cgit v1.2.3-58-ga151 From ff0a522e7db79625aa27a433467eb94c5e255718 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Sun, 24 Jul 2011 10:29:54 -0700 Subject: ocfs2/dlm: Take inflight reference count for remotely mastered resources too The inflight reference count, in the lock resource, is taken to pin the resource in memory. We take it when a new resource is created and release it after a lock is attached to it. We do this to prevent the resource from getting purged prematurely. Earlier this reference count was being taken for locally mastered resources only. This patch extends the same functionality for remotely mastered ones. We are doing this because the same premature purging could occur for remotely mastered resources if the remote node were to die before completion of the create lock. Fix for Oracle bug#12405575. Signed-off-by: Sunil Mushran --- fs/ocfs2/dlm/dlmlock.c | 12 ++++++++---- fs/ocfs2/dlm/dlmmaster.c | 47 +++++++++++++++++------------------------------ fs/ocfs2/dlm/dlmthread.c | 12 +++++++----- 3 files changed, 32 insertions(+), 39 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/dlm/dlmlock.c b/fs/ocfs2/dlm/dlmlock.c index c7f3e22bda1e..3ef2c1adfb8f 100644 --- a/fs/ocfs2/dlm/dlmlock.c +++ b/fs/ocfs2/dlm/dlmlock.c @@ -183,10 +183,6 @@ static enum dlm_status dlmlock_master(struct dlm_ctxt *dlm, kick_thread = 1; } } - /* reduce the inflight count, this may result in the lockres - * being purged below during calc_usage */ - if (lock->ml.node == dlm->node_num) - dlm_lockres_drop_inflight_ref(dlm, res); spin_unlock(&res->spinlock); wake_up(&res->wq); @@ -737,6 +733,14 @@ retry_lock: } } + /* Inflight taken in dlm_get_lock_resource() is dropped here */ + spin_lock(&res->spinlock); + dlm_lockres_drop_inflight_ref(dlm, res); + spin_unlock(&res->spinlock); + + dlm_lockres_calc_usage(dlm, res); + dlm_kick_thread(dlm, res); + if (status != DLM_NORMAL) { lock->lksb->flags &= ~DLM_LKSB_GET_LVB; if (status != DLM_NOTQUEUED) diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c index 11e446ffb67a..005261c333b0 100644 --- a/fs/ocfs2/dlm/dlmmaster.c +++ b/fs/ocfs2/dlm/dlmmaster.c @@ -659,11 +659,8 @@ void dlm_lockres_grab_inflight_ref(struct dlm_ctxt *dlm, { assert_spin_locked(&res->spinlock); - if (!test_bit(dlm->node_num, res->refmap)) { - BUG_ON(res->inflight_locks != 0); - dlm_lockres_set_refmap_bit(dlm, res, dlm->node_num); - } res->inflight_locks++; + mlog(0, "%s: res %.*s, inflight++: now %u, %ps()\n", dlm->name, res->lockname.len, res->lockname.name, res->inflight_locks, __builtin_return_address(0)); @@ -677,12 +674,11 @@ void dlm_lockres_drop_inflight_ref(struct dlm_ctxt *dlm, BUG_ON(res->inflight_locks == 0); res->inflight_locks--; + mlog(0, "%s: res %.*s, inflight--: now %u, %ps()\n", dlm->name, res->lockname.len, res->lockname.name, res->inflight_locks, __builtin_return_address(0)); - if (res->inflight_locks == 0) - dlm_lockres_clear_refmap_bit(dlm, res, dlm->node_num); wake_up(&res->wq); } @@ -716,7 +712,6 @@ struct dlm_lock_resource * dlm_get_lock_resource(struct dlm_ctxt *dlm, unsigned int hash; int tries = 0; int bit, wait_on_recovery = 0; - int drop_inflight_if_nonlocal = 0; BUG_ON(!lockid); @@ -728,36 +723,33 @@ lookup: spin_lock(&dlm->spinlock); tmpres = __dlm_lookup_lockres_full(dlm, lockid, namelen, hash); if (tmpres) { - int dropping_ref = 0; - spin_unlock(&dlm->spinlock); - spin_lock(&tmpres->spinlock); - /* We wait for the other thread that is mastering the resource */ + /* Wait on the thread that is mastering the resource */ if (tmpres->owner == DLM_LOCK_RES_OWNER_UNKNOWN) { __dlm_wait_on_lockres(tmpres); BUG_ON(tmpres->owner == DLM_LOCK_RES_OWNER_UNKNOWN); + spin_unlock(&tmpres->spinlock); + dlm_lockres_put(tmpres); + tmpres = NULL; + goto lookup; } - if (tmpres->owner == dlm->node_num) { - BUG_ON(tmpres->state & DLM_LOCK_RES_DROPPING_REF); - dlm_lockres_grab_inflight_ref(dlm, tmpres); - } else if (tmpres->state & DLM_LOCK_RES_DROPPING_REF) - dropping_ref = 1; - spin_unlock(&tmpres->spinlock); - - /* wait until done messaging the master, drop our ref to allow - * the lockres to be purged, start over. */ - if (dropping_ref) { - spin_lock(&tmpres->spinlock); - __dlm_wait_on_lockres_flags(tmpres, DLM_LOCK_RES_DROPPING_REF); + /* Wait on the resource purge to complete before continuing */ + if (tmpres->state & DLM_LOCK_RES_DROPPING_REF) { + BUG_ON(tmpres->owner == dlm->node_num); + __dlm_wait_on_lockres_flags(tmpres, + DLM_LOCK_RES_DROPPING_REF); spin_unlock(&tmpres->spinlock); dlm_lockres_put(tmpres); tmpres = NULL; goto lookup; } - mlog(0, "found in hash!\n"); + /* Grab inflight ref to pin the resource */ + dlm_lockres_grab_inflight_ref(dlm, tmpres); + + spin_unlock(&tmpres->spinlock); if (res) dlm_lockres_put(res); res = tmpres; @@ -863,14 +855,11 @@ lookup: /* finally add the lockres to its hash bucket */ __dlm_insert_lockres(dlm, res); + /* Grab inflight ref to pin the resource */ spin_lock(&res->spinlock); dlm_lockres_grab_inflight_ref(dlm, res); spin_unlock(&res->spinlock); - /* if this node does not become the master make sure to drop - * this inflight reference below */ - drop_inflight_if_nonlocal = 1; - /* get an extra ref on the mle in case this is a BLOCK * if so, the creator of the BLOCK may try to put the last * ref at this time in the assert master handler, so we @@ -973,8 +962,6 @@ wait: wake_waiters: spin_lock(&res->spinlock); - if (res->owner != dlm->node_num && drop_inflight_if_nonlocal) - dlm_lockres_drop_inflight_ref(dlm, res); res->state &= ~DLM_LOCK_RES_IN_PROGRESS; spin_unlock(&res->spinlock); wake_up(&res->wq); diff --git a/fs/ocfs2/dlm/dlmthread.c b/fs/ocfs2/dlm/dlmthread.c index 4eff65e7e59e..e73c833fc2a1 100644 --- a/fs/ocfs2/dlm/dlmthread.c +++ b/fs/ocfs2/dlm/dlmthread.c @@ -94,24 +94,26 @@ int __dlm_lockres_unused(struct dlm_lock_resource *res) { int bit; + assert_spin_locked(&res->spinlock); + if (__dlm_lockres_has_locks(res)) return 0; + /* Locks are in the process of being created */ + if (res->inflight_locks) + return 0; + if (!list_empty(&res->dirty) || res->state & DLM_LOCK_RES_DIRTY) return 0; if (res->state & DLM_LOCK_RES_RECOVERING) return 0; + /* Another node has this resource with this node as the master */ bit = find_next_bit(res->refmap, O2NM_MAX_NODES, 0); if (bit < O2NM_MAX_NODES) return 0; - /* - * since the bit for dlm->node_num is not set, inflight_locks better - * be zero - */ - BUG_ON(res->inflight_locks != 0); return 1; } -- cgit v1.2.3-58-ga151 From a2c0cc1579176bd0808ef7deea456767dfa80217 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Sun, 24 Jul 2011 10:30:54 -0700 Subject: ocfs2/dlm: dlmlock_remote() needs to account for remastery In dlmlock_remote(), we wait for the resource to stop being active before setting the inprogress flag. Active includes recovery, migration, etc. The problem here is that if the resource was being recovered or migrated, the new owner could very well be that node itself (and thus not a remote node). This problem was observed in Oracle bug#12583620. The error messages observed were as follows: dlm_send_remote_lock_request:337 ERROR: Error -40 (ELOOP) when sending message 503 (key 0xd6d8c7) to node 2 dlmlock_remote:271 ERROR: dlm status = DLM_BADARGS dlmlock:751 ERROR: dlm status = DLM_BADARGS Signed-off-by: Sunil Mushran --- fs/ocfs2/dlm/dlmlock.c | 18 ++++++++---------- 1 file changed, 8 insertions(+), 10 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/dlm/dlmlock.c b/fs/ocfs2/dlm/dlmlock.c index 3ef2c1adfb8f..f32fcba04923 100644 --- a/fs/ocfs2/dlm/dlmlock.c +++ b/fs/ocfs2/dlm/dlmlock.c @@ -227,10 +227,16 @@ static enum dlm_status dlmlock_remote(struct dlm_ctxt *dlm, lock->ml.type, res->lockname.len, res->lockname.name, flags); + /* + * Wait if resource is getting recovered, remastered, etc. + * If the resource was remastered and new owner is self, then exit. + */ spin_lock(&res->spinlock); - - /* will exit this call with spinlock held */ __dlm_wait_on_lockres(res); + if (res->owner == dlm->node_num) { + spin_unlock(&res->spinlock); + return DLM_RECOVERING; + } res->state |= DLM_LOCK_RES_IN_PROGRESS; /* add lock to local (secondary) queue */ @@ -710,18 +716,10 @@ retry_lock: if (status == DLM_RECOVERING || status == DLM_MIGRATING || status == DLM_FORWARD) { - mlog(0, "retrying lock with migration/" - "recovery/in progress\n"); msleep(100); - /* no waiting for dlm_reco_thread */ if (recovery) { if (status != DLM_RECOVERING) goto retry_lock; - - mlog(0, "%s: got RECOVERING " - "for $RECOVERY lock, master " - "was %u\n", dlm->name, - res->owner); /* wait to see the node go down, then * drop down and allow the lockres to * get cleaned up. need to remaster. */ -- cgit v1.2.3-58-ga151 From bb570a5d9e74f71d32751823052db4a97d6a5e7c Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Sun, 24 Jul 2011 10:31:54 -0700 Subject: ocfs2/cluster: Fix output in file elapsed_time_in_ms The o2hb debugfs file, elapsed_time_in_ms, should return values only after the timer is armed atleast once. Signed-off-by: Sunil Mushran --- fs/ocfs2/cluster/heartbeat.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c index 4728a746281d..a4e855e3690e 100644 --- a/fs/ocfs2/cluster/heartbeat.c +++ b/fs/ocfs2/cluster/heartbeat.c @@ -1198,6 +1198,7 @@ static int o2hb_debug_open(struct inode *inode, struct file *file) struct o2hb_debug_buf *db = inode->i_private; struct o2hb_region *reg; unsigned long map[BITS_TO_LONGS(O2NM_MAX_NODES)]; + unsigned long lts; char *buf = NULL; int i = -1; int out = 0; @@ -1234,9 +1235,11 @@ static int o2hb_debug_open(struct inode *inode, struct file *file) case O2HB_DB_TYPE_REGION_ELAPSED_TIME: reg = (struct o2hb_region *)db->db_data; - out += snprintf(buf + out, PAGE_SIZE - out, "%u\n", - jiffies_to_msecs(jiffies - - reg->hr_last_timeout_start)); + lts = reg->hr_last_timeout_start; + /* If 0, it has never been set before */ + if (lts) + lts = jiffies_to_msecs(jiffies - lts); + out += snprintf(buf + out, PAGE_SIZE - out, "%lu\n", lts); goto done; case O2HB_DB_TYPE_REGION_PINNED: -- cgit v1.2.3-58-ga151 From 3ba169ccec1c5ad0f678e04fd29b990197fdfe79 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Sun, 24 Jul 2011 10:32:54 -0700 Subject: ocfs2/cluster: Add new function o2net_fill_node_map() Patch adds function o2net_fill_node_map() to return the bitmap of nodes that it is connected to. This bitmap is also accessible by the user via the debugfs file, /sys/kernel/debug/o2net/connected_nodes. Signed-off-by: Sunil Mushran --- fs/ocfs2/cluster/netdebug.c | 102 ++++++++++++++++++++++++++++++-------------- fs/ocfs2/cluster/tcp.c | 19 +++++++++ fs/ocfs2/cluster/tcp.h | 2 + 3 files changed, 90 insertions(+), 33 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/cluster/netdebug.c b/fs/ocfs2/cluster/netdebug.c index 3a5835904b3d..dc45deb19e68 100644 --- a/fs/ocfs2/cluster/netdebug.c +++ b/fs/ocfs2/cluster/netdebug.c @@ -47,6 +47,7 @@ #define SC_DEBUG_NAME "sock_containers" #define NST_DEBUG_NAME "send_tracking" #define STATS_DEBUG_NAME "stats" +#define NODES_DEBUG_NAME "connected_nodes" #define SHOW_SOCK_CONTAINERS 0 #define SHOW_SOCK_STATS 1 @@ -55,6 +56,7 @@ static struct dentry *o2net_dentry; static struct dentry *sc_dentry; static struct dentry *nst_dentry; static struct dentry *stats_dentry; +static struct dentry *nodes_dentry; static DEFINE_SPINLOCK(o2net_debug_lock); @@ -491,53 +493,87 @@ static const struct file_operations sc_seq_fops = { .release = sc_fop_release, }; -int o2net_debugfs_init(void) +static int o2net_fill_bitmap(char *buf, int len) { - o2net_dentry = debugfs_create_dir(O2NET_DEBUG_DIR, NULL); - if (!o2net_dentry) { - mlog_errno(-ENOMEM); - goto bail; - } + unsigned long map[BITS_TO_LONGS(O2NM_MAX_NODES)]; + int i = -1, out = 0; - nst_dentry = debugfs_create_file(NST_DEBUG_NAME, S_IFREG|S_IRUSR, - o2net_dentry, NULL, - &nst_seq_fops); - if (!nst_dentry) { - mlog_errno(-ENOMEM); - goto bail; - } + o2net_fill_node_map(map, sizeof(map)); - sc_dentry = debugfs_create_file(SC_DEBUG_NAME, S_IFREG|S_IRUSR, - o2net_dentry, NULL, - &sc_seq_fops); - if (!sc_dentry) { - mlog_errno(-ENOMEM); - goto bail; - } + while ((i = find_next_bit(map, O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) + out += snprintf(buf + out, PAGE_SIZE - out, "%d ", i); + out += snprintf(buf + out, PAGE_SIZE - out, "\n"); - stats_dentry = debugfs_create_file(STATS_DEBUG_NAME, S_IFREG|S_IRUSR, - o2net_dentry, NULL, - &stats_seq_fops); - if (!stats_dentry) { - mlog_errno(-ENOMEM); - goto bail; - } + return out; +} + +static int nodes_fop_open(struct inode *inode, struct file *file) +{ + char *buf; + + buf = kmalloc(PAGE_SIZE, GFP_KERNEL); + if (!buf) + return -ENOMEM; + + i_size_write(inode, o2net_fill_bitmap(buf, PAGE_SIZE)); + + file->private_data = buf; return 0; -bail: - debugfs_remove(stats_dentry); - debugfs_remove(sc_dentry); - debugfs_remove(nst_dentry); - debugfs_remove(o2net_dentry); - return -ENOMEM; } +static int o2net_debug_release(struct inode *inode, struct file *file) +{ + kfree(file->private_data); + return 0; +} + +static ssize_t o2net_debug_read(struct file *file, char __user *buf, + size_t nbytes, loff_t *ppos) +{ + return simple_read_from_buffer(buf, nbytes, ppos, file->private_data, + i_size_read(file->f_mapping->host)); +} + +static const struct file_operations nodes_fops = { + .open = nodes_fop_open, + .release = o2net_debug_release, + .read = o2net_debug_read, + .llseek = generic_file_llseek, +}; + void o2net_debugfs_exit(void) { + debugfs_remove(nodes_dentry); debugfs_remove(stats_dentry); debugfs_remove(sc_dentry); debugfs_remove(nst_dentry); debugfs_remove(o2net_dentry); } +int o2net_debugfs_init(void) +{ + mode_t mode = S_IFREG|S_IRUSR; + + o2net_dentry = debugfs_create_dir(O2NET_DEBUG_DIR, NULL); + if (o2net_dentry) + nst_dentry = debugfs_create_file(NST_DEBUG_NAME, mode, + o2net_dentry, NULL, &nst_seq_fops); + if (nst_dentry) + sc_dentry = debugfs_create_file(SC_DEBUG_NAME, mode, + o2net_dentry, NULL, &sc_seq_fops); + if (sc_dentry) + stats_dentry = debugfs_create_file(STATS_DEBUG_NAME, mode, + o2net_dentry, NULL, &stats_seq_fops); + if (stats_dentry) + nodes_dentry = debugfs_create_file(NODES_DEBUG_NAME, mode, + o2net_dentry, NULL, &nodes_fops); + if (nodes_dentry) + return 0; + + o2net_debugfs_exit(); + mlog_errno(-ENOMEM); + return -ENOMEM; +} + #endif /* CONFIG_DEBUG_FS */ diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c index 6e97895ecf24..ae13d5ca7908 100644 --- a/fs/ocfs2/cluster/tcp.c +++ b/fs/ocfs2/cluster/tcp.c @@ -1034,6 +1034,25 @@ static int o2net_tx_can_proceed(struct o2net_node *nn, return ret; } +/* Get a map of all nodes to which this node is currently connected to */ +void o2net_fill_node_map(unsigned long *map, unsigned bytes) +{ + struct o2net_sock_container *sc; + int node, ret; + + BUG_ON(bytes < (BITS_TO_LONGS(O2NM_MAX_NODES) * sizeof(unsigned long))); + + memset(map, 0, bytes); + for (node = 0; node < O2NM_MAX_NODES; ++node) { + o2net_tx_can_proceed(o2net_nn_from_num(node), &sc, &ret); + if (!ret) { + set_bit(node, map); + sc_put(sc); + } + } +} +EXPORT_SYMBOL_GPL(o2net_fill_node_map); + int o2net_send_message_vec(u32 msg_type, u32 key, struct kvec *caller_vec, size_t caller_veclen, u8 target_node, int *status) { diff --git a/fs/ocfs2/cluster/tcp.h b/fs/ocfs2/cluster/tcp.h index fd6179eb26d4..5bada2a69b50 100644 --- a/fs/ocfs2/cluster/tcp.h +++ b/fs/ocfs2/cluster/tcp.h @@ -106,6 +106,8 @@ int o2net_register_handler(u32 msg_type, u32 key, u32 max_len, struct list_head *unreg_list); void o2net_unregister_handler_list(struct list_head *list); +void o2net_fill_node_map(unsigned long *map, unsigned bytes); + struct o2nm_node; int o2net_register_hb_callbacks(void); void o2net_unregister_hb_callbacks(void); -- cgit v1.2.3-58-ga151 From 6b27f62fc750d85bc6fc3718b3b38ec60edc2d74 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Sun, 24 Jul 2011 10:33:54 -0700 Subject: ocfs2/cluster: Cluster up now includes network connections too The cluster up check only checks to see if the node is heartbeating or not. If yes it continues assuming that the node is connected to all the nodes. But if that is not the case, the cluster join aborts with a stack of errors that are not easy to comprehend. This patch adds the network connect check upfront and prints the nodes that the node is not yet connected to, before aborting. Signed-off-by: Sunil Mushran --- fs/ocfs2/dlm/dlmdomain.c | 7 ----- fs/ocfs2/stack_o2cb.c | 67 +++++++++++++++++++++++++++++++++++++++++++----- 2 files changed, 61 insertions(+), 13 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/dlm/dlmdomain.c b/fs/ocfs2/dlm/dlmdomain.c index 200d9888ffee..92f2ead0fab6 100644 --- a/fs/ocfs2/dlm/dlmdomain.c +++ b/fs/ocfs2/dlm/dlmdomain.c @@ -2138,13 +2138,6 @@ struct dlm_ctxt * dlm_register_domain(const char *domain, goto leave; } - if (!o2hb_check_local_node_heartbeating()) { - mlog(ML_ERROR, "the local node has not been configured, or is " - "not heartbeating\n"); - ret = -EPROTO; - goto leave; - } - mlog(0, "register called for domain \"%s\"\n", domain); retry: diff --git a/fs/ocfs2/stack_o2cb.c b/fs/ocfs2/stack_o2cb.c index be935d89a4e8..94368017edb3 100644 --- a/fs/ocfs2/stack_o2cb.c +++ b/fs/ocfs2/stack_o2cb.c @@ -28,6 +28,7 @@ #include "cluster/masklog.h" #include "cluster/nodemanager.h" #include "cluster/heartbeat.h" +#include "cluster/tcp.h" #include "stackglue.h" @@ -255,6 +256,61 @@ static void o2cb_dump_lksb(struct ocfs2_dlm_lksb *lksb) dlm_print_one_lock(lksb->lksb_o2dlm.lockid); } +/* + * Check if this node is heartbeating and is connected to all other + * heartbeating nodes. + */ +static int o2cb_cluster_check(void) +{ + u8 node_num; + int i; + unsigned long hbmap[BITS_TO_LONGS(O2NM_MAX_NODES)]; + unsigned long netmap[BITS_TO_LONGS(O2NM_MAX_NODES)]; + + node_num = o2nm_this_node(); + if (node_num == O2NM_MAX_NODES) { + printk(KERN_ERR "o2cb: This node has not been configured.\n"); + return -EINVAL; + } + + /* + * o2dlm expects o2net sockets to be created. If not, then + * dlm_join_domain() fails with a stack of errors which are both cryptic + * and incomplete. The idea here is to detect upfront whether we have + * managed to connect to all nodes or not. If not, then list the nodes + * to allow the user to check the configuration (incorrect IP, firewall, + * etc.) Yes, this is racy. But its not the end of the world. + */ +#define O2CB_MAP_STABILIZE_COUNT 60 + for (i = 0; i < O2CB_MAP_STABILIZE_COUNT; ++i) { + o2hb_fill_node_map(hbmap, sizeof(hbmap)); + if (!test_bit(node_num, hbmap)) { + printk(KERN_ERR "o2cb: %s heartbeat has not been " + "started.\n", (o2hb_global_heartbeat_active() ? + "Global" : "Local")); + return -EINVAL; + } + o2net_fill_node_map(netmap, sizeof(netmap)); + /* Force set the current node to allow easy compare */ + set_bit(node_num, netmap); + if (!memcmp(hbmap, netmap, sizeof(hbmap))) + return 0; + if (i < O2CB_MAP_STABILIZE_COUNT) + msleep(1000); + } + + printk(KERN_ERR "o2cb: This node could not connect to nodes:"); + i = -1; + while ((i = find_next_bit(hbmap, O2NM_MAX_NODES, + i + 1)) < O2NM_MAX_NODES) { + if (!test_bit(i, netmap)) + printk(" %u", i); + } + printk(".\n"); + + return -ENOTCONN; +} + /* * Called from the dlm when it's about to evict a node. This is how the * classic stack signals node death. @@ -280,12 +336,11 @@ static int o2cb_cluster_connect(struct ocfs2_cluster_connection *conn) BUG_ON(conn == NULL); BUG_ON(conn->cc_proto == NULL); - /* for now we only have one cluster/node, make sure we see it - * in the heartbeat universe */ - if (!o2hb_check_local_node_heartbeating()) { - if (o2hb_global_heartbeat_active()) - mlog(ML_ERROR, "Global heartbeat not started\n"); - rc = -EINVAL; + /* Ensure cluster stack is up and all nodes are connected */ + rc = o2cb_cluster_check(); + if (rc) { + printk(KERN_ERR "o2cb: Cluster check failed. Fix errors " + "before retrying.\n"); goto out; } -- cgit v1.2.3-58-ga151 From 619c200de144b44f5e405305241bcd7edbb8c6cf Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Sun, 24 Jul 2011 10:34:54 -0700 Subject: ocfs2: Clean up messages in the fs Convert useful messages from ML_NOTICE to KERN_NOTICE to improve readability. Signed-off-by: Sunil Mushran --- fs/ocfs2/journal.c | 9 ++++++--- fs/ocfs2/quota_local.c | 13 +++++++++---- fs/ocfs2/slot_map.c | 4 ++-- fs/ocfs2/super.c | 10 +++++----- 4 files changed, 22 insertions(+), 14 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c index 295d56454e8b..8eaaa7854409 100644 --- a/fs/ocfs2/journal.c +++ b/fs/ocfs2/journal.c @@ -1544,9 +1544,9 @@ static int ocfs2_replay_journal(struct ocfs2_super *osb, /* we need to run complete recovery for offline orphan slots */ ocfs2_replay_map_set_state(osb, REPLAY_NEEDED); - mlog(ML_NOTICE, "Recovering node %d from slot %d on device (%u,%u)\n", - node_num, slot_num, - MAJOR(osb->sb->s_dev), MINOR(osb->sb->s_dev)); + printk(KERN_NOTICE "ocfs2: Begin replay journal (node %d, slot %d) on "\ + "device (%u,%u)\n", node_num, slot_num, MAJOR(osb->sb->s_dev), + MINOR(osb->sb->s_dev)); OCFS2_I(inode)->ip_clusters = le32_to_cpu(fe->i_clusters); @@ -1601,6 +1601,9 @@ static int ocfs2_replay_journal(struct ocfs2_super *osb, jbd2_journal_destroy(journal); + printk(KERN_NOTICE "ocfs2: End replay journal (node %d, slot %d) on "\ + "device (%u,%u)\n", node_num, slot_num, MAJOR(osb->sb->s_dev), + MINOR(osb->sb->s_dev)); done: /* drop the lock on this nodes journal */ if (got_lock) diff --git a/fs/ocfs2/quota_local.c b/fs/ocfs2/quota_local.c index dc8007fc9247..942fd65bdad3 100644 --- a/fs/ocfs2/quota_local.c +++ b/fs/ocfs2/quota_local.c @@ -404,7 +404,9 @@ struct ocfs2_quota_recovery *ocfs2_begin_quota_recovery( int status = 0; struct ocfs2_quota_recovery *rec; - mlog(ML_NOTICE, "Beginning quota recovery in slot %u\n", slot_num); + printk(KERN_NOTICE "ocfs2: Beginning quota recovery on device (%s) for " + "slot %u\n", osb->dev_str, slot_num); + rec = ocfs2_alloc_quota_recovery(); if (!rec) return ERR_PTR(-ENOMEM); @@ -596,7 +598,9 @@ int ocfs2_finish_quota_recovery(struct ocfs2_super *osb, struct inode *lqinode; unsigned int flags; - mlog(ML_NOTICE, "Finishing quota recovery in slot %u\n", slot_num); + printk(KERN_NOTICE "ocfs2: Finishing quota recovery on device (%s) for " + "slot %u\n", osb->dev_str, slot_num); + mutex_lock(&sb_dqopt(sb)->dqonoff_mutex); for (type = 0; type < MAXQUOTAS; type++) { if (list_empty(&(rec->r_list[type]))) @@ -612,8 +616,9 @@ int ocfs2_finish_quota_recovery(struct ocfs2_super *osb, /* Someone else is holding the lock? Then he must be * doing the recovery. Just skip the file... */ if (status == -EAGAIN) { - mlog(ML_NOTICE, "skipping quota recovery for slot %d " - "because quota file is locked.\n", slot_num); + printk(KERN_NOTICE "ocfs2: Skipping quota recovery on " + "device (%s) for slot %d because quota file is " + "locked.\n", osb->dev_str, slot_num); status = 0; goto out_put; } else if (status < 0) { diff --git a/fs/ocfs2/slot_map.c b/fs/ocfs2/slot_map.c index 26fc0014d509..1424c151cccc 100644 --- a/fs/ocfs2/slot_map.c +++ b/fs/ocfs2/slot_map.c @@ -493,8 +493,8 @@ int ocfs2_find_slot(struct ocfs2_super *osb) goto bail; } } else - mlog(ML_NOTICE, "slot %d is already allocated to this node!\n", - slot); + printk(KERN_INFO "ocfs2: Slot %d on device (%s) was already " + "allocated to this node!\n", slot, osb->dev_str); ocfs2_set_slot(si, slot, osb->node_num); osb->slot_num = slot; diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index 56f61027236b..3e7850380d2b 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -1107,9 +1107,9 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent) ocfs2_set_ro_flag(osb, 1); - printk(KERN_NOTICE "Readonly device detected. No cluster " - "services will be utilized for this mount. Recovery " - "will be skipped.\n"); + printk(KERN_NOTICE "ocfs2: Readonly device (%s) detected. " + "Cluster services will not be used for this mount. " + "Recovery will be skipped.\n", osb->dev_str); } if (!ocfs2_is_hard_readonly(osb)) { @@ -2462,8 +2462,8 @@ static int ocfs2_check_volume(struct ocfs2_super *osb) goto finally; } } else { - mlog(ML_NOTICE, "File system was not unmounted cleanly, " - "recovering volume.\n"); + printk(KERN_NOTICE "ocfs2: File system on device (%s) was not " + "unmounted cleanly, recovering it.\n", osb->dev_str); } local = ocfs2_mount_local(osb); -- cgit v1.2.3-58-ga151 From a035bff6b82aca89c1223e2c614adc2d17ec8aa2 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Sun, 24 Jul 2011 10:35:54 -0700 Subject: ocfs2: Add comment about orphan scanning Add a comment that explains the reason as to why orphan scan scans all the slots. Signed-off-by: Sunil Mushran --- fs/ocfs2/journal.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c index 8eaaa7854409..0a42ae96dca7 100644 --- a/fs/ocfs2/journal.c +++ b/fs/ocfs2/journal.c @@ -1811,6 +1811,20 @@ static inline unsigned long ocfs2_orphan_scan_timeout(void) * every slot, queuing a recovery of the slot on the ocfs2_wq thread. This * is done to catch any orphans that are left over in orphan directories. * + * It scans all slots, even ones that are in use. It does so to handle the + * case described below: + * + * Node 1 has an inode it was using. The dentry went away due to memory + * pressure. Node 1 closes the inode, but it's on the free list. The node + * has the open lock. + * Node 2 unlinks the inode. It grabs the dentry lock to notify others, + * but node 1 has no dentry and doesn't get the message. It trylocks the + * open lock, sees that another node has a PR, and does nothing. + * Later node 2 runs its orphan dir. It igets the inode, trylocks the + * open lock, sees the PR still, and does nothing. + * Basically, we have to trigger an orphan iput on node 1. The only way + * for this to happen is if node 1 runs node 2's orphan dir. + * * ocfs2_queue_orphan_scan gets called every ORPHAN_SCAN_SCHEDULE_TIMEOUT * seconds. It gets an EX lock on os_lockres and checks sequence number * stored in LVB. If the sequence number has changed, it means some other -- cgit v1.2.3-58-ga151 From 5cffff9e29866a3de98c2c25135b3199491f93b0 Mon Sep 17 00:00:00 2001 From: Wengang Wang Date: Sun, 24 Jul 2011 10:36:54 -0700 Subject: ocfs2: Fix ocfs2_page_mkwrite() This patch address two shortcomings in ocfs2_page_mkwrite(): 1. Makes the function return better VM_FAULT_* errors. 2. It handles a error that is triggered when a page is dropped from the mapping due to memory pressure. This patch locks the page to prevent that. [Patch was cleaned up by Sunil Mushran.] Signed-off-by: Wengang Wang Signed-off-by: Sunil Mushran --- fs/ocfs2/aops.c | 51 +++++++++++++++++++++++++++++++++++++++++++-------- fs/ocfs2/mmap.c | 53 ++++++++++++++++++++++++----------------------------- 2 files changed, 67 insertions(+), 37 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c index c1efe939c774..ff98c169b631 100644 --- a/fs/ocfs2/aops.c +++ b/fs/ocfs2/aops.c @@ -862,6 +862,12 @@ struct ocfs2_write_ctxt { struct page *w_pages[OCFS2_MAX_CTXT_PAGES]; struct page *w_target_page; + /* + * w_target_locked is used for page_mkwrite path indicating no unlocking + * against w_target_page in ocfs2_write_end_nolock. + */ + unsigned int w_target_locked:1; + /* * ocfs2_write_end() uses this to know what the real range to * write in the target should be. @@ -895,6 +901,24 @@ void ocfs2_unlock_and_free_pages(struct page **pages, int num_pages) static void ocfs2_free_write_ctxt(struct ocfs2_write_ctxt *wc) { + int i; + + /* + * w_target_locked is only set to true in the page_mkwrite() case. + * The intent is to allow us to lock the target page from write_begin() + * to write_end(). The caller must hold a ref on w_target_page. + */ + if (wc->w_target_locked) { + BUG_ON(!wc->w_target_page); + for (i = 0; i < wc->w_num_pages; i++) { + if (wc->w_target_page == wc->w_pages[i]) { + wc->w_pages[i] = NULL; + break; + } + } + mark_page_accessed(wc->w_target_page); + page_cache_release(wc->w_target_page); + } ocfs2_unlock_and_free_pages(wc->w_pages, wc->w_num_pages); brelse(wc->w_di_bh); @@ -1132,20 +1156,17 @@ static int ocfs2_grab_pages_for_write(struct address_space *mapping, */ lock_page(mmap_page); + /* Exit and let the caller retry */ if (mmap_page->mapping != mapping) { + WARN_ON(mmap_page->mapping); unlock_page(mmap_page); - /* - * Sanity check - the locking in - * ocfs2_pagemkwrite() should ensure - * that this code doesn't trigger. - */ - ret = -EINVAL; - mlog_errno(ret); + ret = -EAGAIN; goto out; } page_cache_get(mmap_page); wc->w_pages[i] = mmap_page; + wc->w_target_locked = true; } else { wc->w_pages[i] = find_or_create_page(mapping, index, GFP_NOFS); @@ -1160,6 +1181,8 @@ static int ocfs2_grab_pages_for_write(struct address_space *mapping, wc->w_target_page = wc->w_pages[i]; } out: + if (ret) + wc->w_target_locked = false; return ret; } @@ -1817,11 +1840,23 @@ try_again: */ ret = ocfs2_grab_pages_for_write(mapping, wc, wc->w_cpos, pos, len, cluster_of_pages, mmap_page); - if (ret) { + if (ret && ret != -EAGAIN) { mlog_errno(ret); goto out_quota; } + /* + * ocfs2_grab_pages_for_write() returns -EAGAIN if it could not lock + * the target page. In this case, we exit with no error and no target + * page. This will trigger the caller, page_mkwrite(), to re-try + * the operation. + */ + if (ret == -EAGAIN) { + BUG_ON(wc->w_target_page); + ret = 0; + goto out_quota; + } + ret = ocfs2_write_cluster_by_desc(mapping, data_ac, meta_ac, wc, pos, len); if (ret) { diff --git a/fs/ocfs2/mmap.c b/fs/ocfs2/mmap.c index 3e9393ca39eb..9cd41083e991 100644 --- a/fs/ocfs2/mmap.c +++ b/fs/ocfs2/mmap.c @@ -61,7 +61,7 @@ static int ocfs2_fault(struct vm_area_struct *area, struct vm_fault *vmf) static int __ocfs2_page_mkwrite(struct file *file, struct buffer_head *di_bh, struct page *page) { - int ret; + int ret = VM_FAULT_NOPAGE; struct inode *inode = file->f_path.dentry->d_inode; struct address_space *mapping = inode->i_mapping; loff_t pos = page_offset(page); @@ -71,32 +71,25 @@ static int __ocfs2_page_mkwrite(struct file *file, struct buffer_head *di_bh, void *fsdata; loff_t size = i_size_read(inode); - /* - * Another node might have truncated while we were waiting on - * cluster locks. - * We don't check size == 0 before the shift. This is borrowed - * from do_generic_file_read. - */ last_index = (size - 1) >> PAGE_CACHE_SHIFT; - if (unlikely(!size || page->index > last_index)) { - ret = -EINVAL; - goto out; - } /* - * The i_size check above doesn't catch the case where nodes - * truncated and then re-extended the file. We'll re-check the - * page mapping after taking the page lock inside of - * ocfs2_write_begin_nolock(). + * There are cases that lead to the page no longer bebongs to the + * mapping. + * 1) pagecache truncates locally due to memory pressure. + * 2) pagecache truncates when another is taking EX lock against + * inode lock. see ocfs2_data_convert_worker. + * + * The i_size check doesn't catch the case where nodes truncated and + * then re-extended the file. We'll re-check the page mapping after + * taking the page lock inside of ocfs2_write_begin_nolock(). + * + * Let VM retry with these cases. */ - if (!PageUptodate(page) || page->mapping != inode->i_mapping) { - /* - * the page has been umapped in ocfs2_data_downconvert_worker. - * So return 0 here and let VFS retry. - */ - ret = 0; + if ((page->mapping != inode->i_mapping) || + (!PageUptodate(page)) || + (page_offset(page) >= size)) goto out; - } /* * Call ocfs2_write_begin() and ocfs2_write_end() to take @@ -116,17 +109,21 @@ static int __ocfs2_page_mkwrite(struct file *file, struct buffer_head *di_bh, if (ret) { if (ret != -ENOSPC) mlog_errno(ret); + if (ret == -ENOMEM) + ret = VM_FAULT_OOM; + else + ret = VM_FAULT_SIGBUS; goto out; } - ret = ocfs2_write_end_nolock(mapping, pos, len, len, locked_page, - fsdata); - if (ret < 0) { - mlog_errno(ret); + if (!locked_page) { + ret = VM_FAULT_NOPAGE; goto out; } + ret = ocfs2_write_end_nolock(mapping, pos, len, len, locked_page, + fsdata); BUG_ON(ret != len); - ret = 0; + ret = VM_FAULT_LOCKED; out: return ret; } @@ -168,8 +165,6 @@ static int ocfs2_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) out: ocfs2_unblock_signals(&oldset); - if (ret) - ret = VM_FAULT_SIGBUS; return ret; } -- cgit v1.2.3-58-ga151 From 93862d5e1ab875664c6cc95254fc365028a48bb1 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Mon, 25 Jul 2011 14:58:15 -0700 Subject: ocfs2: Implement llseek() ocfs2 implements its own llseek() to provide the SEEK_HOLE/SEEK_DATA functionality. SEEK_HOLE sets the file pointer to the start of either a hole or an unwritten (preallocated) extent, that is greater than or equal to the supplied offset. SEEK_DATA sets the file pointer to the start of an allocated extent (not unwritten) that is greater than or equal to the supplied offset. If the supplied offset is on a desired region, then the file pointer is set to it. Offsets greater than or equal to the file size return -ENXIO. Unwritten (preallocated) extents are considered holes because the file system treats reads to such regions in the same way as it does to holes. Signed-off-by: Sunil Mushran --- fs/ocfs2/extent_map.c | 96 +++++++++++++++++++++++++++++++++++++++++++++++++++ fs/ocfs2/extent_map.h | 2 ++ fs/ocfs2/file.c | 55 +++++++++++++++++++++++++++-- 3 files changed, 151 insertions(+), 2 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/extent_map.c b/fs/ocfs2/extent_map.c index 23457b491e8c..2f5b92ef0e53 100644 --- a/fs/ocfs2/extent_map.c +++ b/fs/ocfs2/extent_map.c @@ -832,6 +832,102 @@ out: return ret; } +int ocfs2_seek_data_hole_offset(struct file *file, loff_t *offset, int origin) +{ + struct inode *inode = file->f_mapping->host; + int ret; + unsigned int is_last = 0, is_data = 0; + u16 cs_bits = OCFS2_SB(inode->i_sb)->s_clustersize_bits; + u32 cpos, cend, clen, hole_size; + u64 extoff, extlen; + struct buffer_head *di_bh = NULL; + struct ocfs2_extent_rec rec; + + BUG_ON(origin != SEEK_DATA && origin != SEEK_HOLE); + + ret = ocfs2_inode_lock(inode, &di_bh, 0); + if (ret) { + mlog_errno(ret); + goto out; + } + + down_read(&OCFS2_I(inode)->ip_alloc_sem); + + if (*offset >= inode->i_size) { + ret = -ENXIO; + goto out_unlock; + } + + if (OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL) { + if (origin == SEEK_HOLE) + *offset = inode->i_size; + goto out_unlock; + } + + clen = 0; + cpos = *offset >> cs_bits; + cend = ocfs2_clusters_for_bytes(inode->i_sb, inode->i_size); + + while (cpos < cend && !is_last) { + ret = ocfs2_get_clusters_nocache(inode, di_bh, cpos, &hole_size, + &rec, &is_last); + if (ret) { + mlog_errno(ret); + goto out_unlock; + } + + extoff = cpos; + extoff <<= cs_bits; + + if (rec.e_blkno == 0ULL) { + clen = hole_size; + is_data = 0; + } else { + clen = le16_to_cpu(rec.e_leaf_clusters) - + (cpos - le32_to_cpu(rec.e_cpos)); + is_data = (rec.e_flags & OCFS2_EXT_UNWRITTEN) ? 0 : 1; + } + + if ((!is_data && origin == SEEK_HOLE) || + (is_data && origin == SEEK_DATA)) { + if (extoff > *offset) + *offset = extoff; + goto out_unlock; + } + + if (!is_last) + cpos += clen; + } + + if (origin == SEEK_HOLE) { + extoff = cpos; + extoff <<= cs_bits; + extlen = clen; + extlen <<= cs_bits; + + if ((extoff + extlen) > inode->i_size) + extlen = inode->i_size - extoff; + extoff += extlen; + if (extoff > *offset) + *offset = extoff; + goto out_unlock; + } + + ret = -ENXIO; + +out_unlock: + + brelse(di_bh); + + up_read(&OCFS2_I(inode)->ip_alloc_sem); + + ocfs2_inode_unlock(inode, 0); +out: + if (ret && ret != -ENXIO) + ret = -ENXIO; + return ret; +} + int ocfs2_read_virt_blocks(struct inode *inode, u64 v_block, int nr, struct buffer_head *bhs[], int flags, int (*validate)(struct super_block *sb, diff --git a/fs/ocfs2/extent_map.h b/fs/ocfs2/extent_map.h index e79d41c2c909..67ea57d2fd59 100644 --- a/fs/ocfs2/extent_map.h +++ b/fs/ocfs2/extent_map.h @@ -53,6 +53,8 @@ int ocfs2_extent_map_get_blocks(struct inode *inode, u64 v_blkno, u64 *p_blkno, int ocfs2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, u64 map_start, u64 map_len); +int ocfs2_seek_data_hole_offset(struct file *file, loff_t *offset, int origin); + int ocfs2_xattr_get_clusters(struct inode *inode, u32 v_cluster, u32 *p_cluster, u32 *num_clusters, struct ocfs2_extent_list *el, diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index 0fc2bd34039d..c0f015e11c28 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c @@ -2591,6 +2591,57 @@ bail: return ret; } +/* Refer generic_file_llseek_unlocked() */ +static loff_t ocfs2_file_llseek(struct file *file, loff_t offset, int origin) +{ + struct inode *inode = file->f_mapping->host; + int ret = 0; + + mutex_lock(&inode->i_mutex); + + switch (origin) { + case SEEK_SET: + break; + case SEEK_END: + offset += inode->i_size; + break; + case SEEK_CUR: + if (offset == 0) { + offset = file->f_pos; + goto out; + } + offset += file->f_pos; + break; + case SEEK_DATA: + case SEEK_HOLE: + ret = ocfs2_seek_data_hole_offset(file, &offset, origin); + if (ret) + goto out; + break; + default: + ret = -EINVAL; + goto out; + } + + if (offset < 0 && !(file->f_mode & FMODE_UNSIGNED_OFFSET)) + ret = -EINVAL; + if (!ret && offset > inode->i_sb->s_maxbytes) + ret = -EINVAL; + if (ret) + goto out; + + if (offset != file->f_pos) { + file->f_pos = offset; + file->f_version = 0; + } + +out: + mutex_unlock(&inode->i_mutex); + if (ret) + return ret; + return offset; +} + const struct inode_operations ocfs2_file_iops = { .setattr = ocfs2_setattr, .getattr = ocfs2_getattr, @@ -2615,7 +2666,7 @@ const struct inode_operations ocfs2_special_file_iops = { * ocfs2_fops_no_plocks and ocfs2_dops_no_plocks! */ const struct file_operations ocfs2_fops = { - .llseek = generic_file_llseek, + .llseek = ocfs2_file_llseek, .read = do_sync_read, .write = do_sync_write, .mmap = ocfs2_mmap, @@ -2663,7 +2714,7 @@ const struct file_operations ocfs2_dops = { * the cluster. */ const struct file_operations ocfs2_fops_no_plocks = { - .llseek = generic_file_llseek, + .llseek = ocfs2_file_llseek, .read = do_sync_read, .write = do_sync_write, .mmap = ocfs2_mmap, -- cgit v1.2.3-58-ga151