From 0533400b7813df6c22a171499434d30bd57e799c Mon Sep 17 00:00:00 2001 From: Stoyan Gaydarov Date: Mon, 7 Jul 2008 07:45:59 -0500 Subject: [JFFS2] Use .unlocked_ioctl This changes the .ioctl to the .unlocked_ioctl version. Signed-off-by: Stoyan Gaydarov Signed-off-by: David Woodhouse --- fs/jffs2/dir.c | 2 +- fs/jffs2/file.c | 2 +- fs/jffs2/ioctl.c | 3 +-- fs/jffs2/os-linux.h | 2 +- 4 files changed, 4 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/jffs2/dir.c b/fs/jffs2/dir.c index c0c141f6fde..cd219ef5525 100644 --- a/fs/jffs2/dir.c +++ b/fs/jffs2/dir.c @@ -38,7 +38,7 @@ const struct file_operations jffs2_dir_operations = { .read = generic_read_dir, .readdir = jffs2_readdir, - .ioctl = jffs2_ioctl, + .unlocked_ioctl=jffs2_ioctl, .fsync = jffs2_fsync }; diff --git a/fs/jffs2/file.c b/fs/jffs2/file.c index 5e920343b2c..5a98aa87c85 100644 --- a/fs/jffs2/file.c +++ b/fs/jffs2/file.c @@ -46,7 +46,7 @@ const struct file_operations jffs2_file_operations = .aio_read = generic_file_aio_read, .write = do_sync_write, .aio_write = generic_file_aio_write, - .ioctl = jffs2_ioctl, + .unlocked_ioctl=jffs2_ioctl, .mmap = generic_file_readonly_mmap, .fsync = jffs2_fsync, .splice_read = generic_file_splice_read, diff --git a/fs/jffs2/ioctl.c b/fs/jffs2/ioctl.c index e2177210f62..9d41f43e47b 100644 --- a/fs/jffs2/ioctl.c +++ b/fs/jffs2/ioctl.c @@ -12,8 +12,7 @@ #include #include "nodelist.h" -int jffs2_ioctl(struct inode *inode, struct file *filp, unsigned int cmd, - unsigned long arg) +long jffs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) { /* Later, this will provide for lsattr.jffs2 and chattr.jffs2, which will include compression support etc. */ diff --git a/fs/jffs2/os-linux.h b/fs/jffs2/os-linux.h index 2cc866cf134..5e194a5c8e2 100644 --- a/fs/jffs2/os-linux.h +++ b/fs/jffs2/os-linux.h @@ -167,7 +167,7 @@ int jffs2_fsync(struct file *, struct dentry *, int); int jffs2_do_readpage_unlock (struct inode *inode, struct page *pg); /* ioctl.c */ -int jffs2_ioctl(struct inode *, struct file *, unsigned int, unsigned long); +long jffs2_ioctl(struct file *, unsigned int, unsigned long); /* symlink.c */ extern const struct inode_operations jffs2_symlink_inode_operations; -- cgit v1.2.3-70-g09d2 From 04e1e0cccade330ab3715ce59234f7e3b087e246 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Tue, 22 Jul 2008 13:04:18 +0000 Subject: [CIFS] Fix compiler warning on 64-bit Signed-off-by: Steve French --- fs/cifs/asn1.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/cifs/asn1.c b/fs/cifs/asn1.c index f58e41d3ba4..42765465701 100644 --- a/fs/cifs/asn1.c +++ b/fs/cifs/asn1.c @@ -400,7 +400,7 @@ asn1_oid_decode(struct asn1_ctx *ctx, size = eoc - ctx->pointer + 1; /* first subid actually encodes first two subids */ - if (size < 2 || size > ULONG_MAX/sizeof(unsigned long)) + if (size < 2 || size > UINT_MAX/sizeof(unsigned long)) return 0; *oid = kmalloc(size * sizeof(unsigned long), GFP_ATOMIC); -- cgit v1.2.3-70-g09d2 From 09e50d55a9490e9b7a6fdfbf8fc078924b25ecb5 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Wed, 23 Jul 2008 10:11:19 -0400 Subject: lockdep: annotate cifs in-kernel sockets Put CIFS sockets in their own class to avoid some lockdep warnings. CIFS sockets are not exposed to user-space, and so are not subject to the same deadlock scenarios. A similar change was made a couple of years ago for RPC sockets in commit ed07536ed6731775219c1df7fa26a7588753e693. This patch should prevent lockdep false-positives like this one: ======================================================= [ INFO: possible circular locking dependency detected ] 2.6.18-98.el5.jtltest.38.bz456320.1debug #1 ------------------------------------------------------- test5/2483 is trying to acquire lock: (sk_lock-AF_INET){--..}, at: [] tcp_sendmsg+0x1c/0xb2f but task is already holding lock: (&inode->i_alloc_sem){--..}, at: [] notify_change+0xf5/0x2e0 which lock already depends on the new lock. the existing dependency chain (in reverse order) is: -> #3 (&inode->i_alloc_sem){--..}: [] __lock_acquire+0x9a9/0xadf [] lock_acquire+0x55/0x70 [] notify_change+0xf5/0x2e0 [] down_write+0x3c/0x68 [] notify_change+0xf5/0x2e0 [] do_truncate+0x50/0x6b [] get_write_access+0x40/0x46 [] may_open+0x1d3/0x22e [] open_namei+0x2c6/0x6dd [] do_filp_open+0x1c/0x38 [] _spin_unlock+0x17/0x20 [] get_unused_fd+0xf9/0x107 [] do_sys_open+0x44/0xbe [] system_call+0x7e/0x83 [] 0xffffffffffffffff -> #2 (&sysfs_inode_imutex_key){--..}: [] __lock_acquire+0x9a9/0xadf [] create_dir+0x26/0x1d7 [] lock_acquire+0x55/0x70 [] create_dir+0x26/0x1d7 [] mutex_lock_nested+0x104/0x29c [] __lock_acquire+0x9ca/0xadf [] create_dir+0x26/0x1d7 [] sysfs_create_dir+0x58/0x76 [] kobject_add+0xdb/0x198 [] class_device_add+0xb2/0x465 [] kobject_get+0x12/0x17 [] register_netdevice+0x270/0x33e [] register_netdev+0x59/0x67 [] net_olddevs_init+0xb/0xac [] init+0x1f9/0x2fc [] _spin_unlock_irq+0x24/0x27 [] trace_hardirqs_on_thunk+0x35/0x37 [] child_rip+0xa/0x11 [] _spin_unlock_irq+0x24/0x27 [] restore_args+0x0/0x30 [] acpi_ds_init_one_object+0x0/0x80 [] init+0x0/0x2fc [] child_rip+0x0/0x11 [] 0xffffffffffffffff -> #1 (rtnl_mutex){--..}: [] __lock_acquire+0x9a9/0xadf [] ip_mc_leave_group+0x23/0xb7 [] lock_acquire+0x55/0x70 [] ip_mc_leave_group+0x23/0xb7 [] mutex_lock_nested+0x104/0x29c [] ip_mc_leave_group+0x23/0xb7 [] do_ip_setsockopt+0x6d1/0x9bf [] lock_release_holdtime+0x27/0x48 [] lock_release_holdtime+0x27/0x48 [] do_page_fault+0x503/0x835 [] socket_has_perm+0x5b/0x68 [] ip_setsockopt+0x22/0x78 [] sys_setsockopt+0x91/0xb7 [] tracesys+0xd5/0xdf [] 0xffffffffffffffff -> #0 (sk_lock-AF_INET){--..}: [] print_stack_trace+0x59/0x68 [] __lock_acquire+0x8bf/0xadf [] lock_acquire+0x55/0x70 [] tcp_sendmsg+0x1c/0xb2f [] lock_sock+0xd4/0xe4 [] _local_bh_enable+0xcb/0xe0 [] restore_args+0x0/0x30 [] tcp_sendmsg+0x1c/0xb2f [] sock_sendmsg+0xf3/0x110 [] autoremove_wake_function+0x0/0x2e [] kernel_text_address+0x1a/0x26 [] dump_trace+0x211/0x23a [] find_usage_backwards+0x5f/0x88 [] MD5Final+0xaf/0xc2 [cifs] [] cifs_calculate_signature+0x55/0x69 [cifs] [] kernel_sendmsg+0x35/0x47 [] smb_send+0xa3/0x151 [cifs] [] SendReceive+0x1a2/0x448 [cifs] [] __lock_acquire+0x95c/0xadf [] CIFSSMBSetEOF+0x20d/0x25b [cifs] [] cifs_set_file_size+0x110/0x3b7 [cifs] [] cifs_setattr+0x3b2/0x6f6 [cifs] [] notify_change+0xf5/0x2e0 [] notify_change+0x145/0x2e0 [] do_truncate+0x50/0x6b [] get_write_access+0x40/0x46 [] may_open+0x1d3/0x22e [] open_namei+0x2c6/0x6dd [] do_filp_open+0x1c/0x38 [] _spin_unlock+0x17/0x20 [] get_unused_fd+0xf9/0x107 [] do_sys_open+0x44/0xbe [] tracesys+0xd5/0xdf [] 0xffffffffffffffff other info that might help us debug this: 2 locks held by test5/2483: #0: (&inode->i_mutex){--..}, at: [] do_truncate+0x45/0x6b #1: (&inode->i_alloc_sem){--..}, at: [] notify_change+0xf5/0x2e0 stack backtrace: Call Trace: [] print_circular_bug_tail+0x65/0x6e [] print_stack_trace+0x59/0x68 [] __lock_acquire+0x8bf/0xadf [] lock_acquire+0x55/0x70 [] tcp_sendmsg+0x1c/0xb2f [] lock_sock+0xd4/0xe4 [] _local_bh_enable+0xcb/0xe0 [] restore_args+0x0/0x30 [] tcp_sendmsg+0x1c/0xb2f [] sock_sendmsg+0xf3/0x110 [] autoremove_wake_function+0x0/0x2e [] kernel_text_address+0x1a/0x26 [] dump_trace+0x211/0x23a [] find_usage_backwards+0x5f/0x88 [] :cifs:MD5Final+0xaf/0xc2 [] :cifs:cifs_calculate_signature+0x55/0x69 [] kernel_sendmsg+0x35/0x47 [] :cifs:smb_send+0xa3/0x151 [] :cifs:SendReceive+0x1a2/0x448 [] __lock_acquire+0x95c/0xadf [] :cifs:CIFSSMBSetEOF+0x20d/0x25b [] :cifs:cifs_set_file_size+0x110/0x3b7 [] :cifs:cifs_setattr+0x3b2/0x6f6 [] notify_change+0xf5/0x2e0 [] notify_change+0x145/0x2e0 [] do_truncate+0x50/0x6b [] get_write_access+0x40/0x46 [] may_open+0x1d3/0x22e [] open_namei+0x2c6/0x6dd [] do_filp_open+0x1c/0x38 [] _spin_unlock+0x17/0x20 [] get_unused_fd+0xf9/0x107 [] do_sys_open+0x44/0xbe [] tracesys+0xd5/0xdf Signed-off-by: Jeff Layton Acked-by: Peter Zijlstra Signed-off-by: Steve French --- fs/cifs/connect.c | 35 +++++++++++++++++++++++++++++++++++ 1 file changed, 35 insertions(+) (limited to 'fs') diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c index e8fa46c7cff..64446272938 100644 --- a/fs/cifs/connect.c +++ b/fs/cifs/connect.c @@ -1461,6 +1461,39 @@ get_dfs_path(int xid, struct cifsSesInfo *pSesInfo, const char *old_path, return rc; } +#ifdef CONFIG_DEBUG_LOCK_ALLOC +static struct lock_class_key cifs_key[2]; +static struct lock_class_key cifs_slock_key[2]; + +static inline void +cifs_reclassify_socket4(struct socket *sock) +{ + struct sock *sk = sock->sk; + BUG_ON(sock_owned_by_user(sk)); + sock_lock_init_class_and_name(sk, "slock-AF_INET-CIFS", + &cifs_slock_key[0], "sk_lock-AF_INET-CIFS", &cifs_key[0]); +} + +static inline void +cifs_reclassify_socket6(struct socket *sock) +{ + struct sock *sk = sock->sk; + BUG_ON(sock_owned_by_user(sk)); + sock_lock_init_class_and_name(sk, "slock-AF_INET6-CIFS", + &cifs_slock_key[1], "sk_lock-AF_INET6-CIFS", &cifs_key[1]); +} +#else +static inline void +cifs_reclassify_socket4(struct socket *sock) +{ +} + +static inline void +cifs_reclassify_socket6(struct socket *sock) +{ +} +#endif + /* See RFC1001 section 14 on representation of Netbios names */ static void rfc1002mangle(char *target, char *source, unsigned int length) { @@ -1495,6 +1528,7 @@ ipv4_connect(struct sockaddr_in *psin_server, struct socket **csocket, /* BB other socket options to set KEEPALIVE, NODELAY? */ cFYI(1, ("Socket created")); (*csocket)->sk->sk_allocation = GFP_NOFS; + cifs_reclassify_socket4(*csocket); } } @@ -1627,6 +1661,7 @@ ipv6_connect(struct sockaddr_in6 *psin_server, struct socket **csocket) /* BB other socket options to set KEEPALIVE, NODELAY? */ cFYI(1, ("ipv6 Socket created")); (*csocket)->sk->sk_allocation = GFP_NOFS; + cifs_reclassify_socket6(*csocket); } } -- cgit v1.2.3-70-g09d2 From 8efdbde647f542ce0d303273df7ad4157caa03d0 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Wed, 23 Jul 2008 21:28:12 +0000 Subject: [CIFS] break ATTR_SIZE changes out into their own function Move the code that handles ATTR_SIZE changes to its own function. This makes for a smaller function and reduces the level of indentation. Signed-off-by: Jeff Layton Signed-off-by: Steve French --- fs/cifs/inode.c | 151 +++++++++++++++++++++++++++++--------------------------- 1 file changed, 78 insertions(+), 73 deletions(-) (limited to 'fs') diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c index 2e904bd111c..46e54d39461 100644 --- a/fs/cifs/inode.c +++ b/fs/cifs/inode.c @@ -1413,6 +1413,82 @@ out_busy: return -ETXTBSY; } +static int +cifs_set_file_size(struct inode *inode, struct iattr *attrs, + int xid, char *full_path) +{ + int rc; + struct cifsFileInfo *open_file; + struct cifsInodeInfo *cifsInode = CIFS_I(inode); + struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb); + struct cifsTconInfo *pTcon = cifs_sb->tcon; + + /* + * To avoid spurious oplock breaks from server, in the case of + * inodes that we already have open, avoid doing path based + * setting of file size if we can do it by handle. + * This keeps our caching token (oplock) and avoids timeouts + * when the local oplock break takes longer to flush + * writebehind data than the SMB timeout for the SetPathInfo + * request would allow + */ + open_file = find_writable_file(cifsInode); + if (open_file) { + __u16 nfid = open_file->netfid; + __u32 npid = open_file->pid; + rc = CIFSSMBSetFileSize(xid, pTcon, attrs->ia_size, nfid, + npid, false); + atomic_dec(&open_file->wrtPending); + cFYI(1, ("SetFSize for attrs rc = %d", rc)); + if ((rc == -EINVAL) || (rc == -EOPNOTSUPP)) { + unsigned int bytes_written; + rc = CIFSSMBWrite(xid, pTcon, nfid, 0, attrs->ia_size, + &bytes_written, NULL, NULL, 1); + cFYI(1, ("Wrt seteof rc %d", rc)); + } + } else + rc = -EINVAL; + + if (rc != 0) { + /* Set file size by pathname rather than by handle + either because no valid, writeable file handle for + it was found or because there was an error setting + it by handle */ + rc = CIFSSMBSetEOF(xid, pTcon, full_path, attrs->ia_size, + false, cifs_sb->local_nls, + cifs_sb->mnt_cifs_flags & + CIFS_MOUNT_MAP_SPECIAL_CHR); + cFYI(1, ("SetEOF by path (setattrs) rc = %d", rc)); + if ((rc == -EINVAL) || (rc == -EOPNOTSUPP)) { + __u16 netfid; + int oplock = 0; + + rc = SMBLegacyOpen(xid, pTcon, full_path, + FILE_OPEN, GENERIC_WRITE, + CREATE_NOT_DIR, &netfid, &oplock, NULL, + cifs_sb->local_nls, + cifs_sb->mnt_cifs_flags & + CIFS_MOUNT_MAP_SPECIAL_CHR); + if (rc == 0) { + unsigned int bytes_written; + rc = CIFSSMBWrite(xid, pTcon, netfid, 0, + attrs->ia_size, + &bytes_written, NULL, + NULL, 1); + cFYI(1, ("wrt seteof rc %d", rc)); + CIFSSMBClose(xid, pTcon, netfid); + } + } + } + + if (rc == 0) { + rc = cifs_vmtruncate(inode, attrs->ia_size); + cifs_truncate_page(inode->i_mapping, inode->i_size); + } + + return rc; +} + int cifs_setattr(struct dentry *direntry, struct iattr *attrs) { int xid; @@ -1420,7 +1496,6 @@ int cifs_setattr(struct dentry *direntry, struct iattr *attrs) struct cifsTconInfo *pTcon; char *full_path = NULL; int rc = -EACCES; - struct cifsFileInfo *open_file = NULL; FILE_BASIC_INFO time_buf; bool set_time = false; bool set_dosattr = false; @@ -1472,78 +1547,8 @@ int cifs_setattr(struct dentry *direntry, struct iattr *attrs) } if (attrs->ia_valid & ATTR_SIZE) { - /* To avoid spurious oplock breaks from server, in the case of - inodes that we already have open, avoid doing path based - setting of file size if we can do it by handle. - This keeps our caching token (oplock) and avoids timeouts - when the local oplock break takes longer to flush - writebehind data than the SMB timeout for the SetPathInfo - request would allow */ - - open_file = find_writable_file(cifsInode); - if (open_file) { - __u16 nfid = open_file->netfid; - __u32 npid = open_file->pid; - rc = CIFSSMBSetFileSize(xid, pTcon, attrs->ia_size, - nfid, npid, false); - atomic_dec(&open_file->wrtPending); - cFYI(1, ("SetFSize for attrs rc = %d", rc)); - if ((rc == -EINVAL) || (rc == -EOPNOTSUPP)) { - unsigned int bytes_written; - rc = CIFSSMBWrite(xid, pTcon, - nfid, 0, attrs->ia_size, - &bytes_written, NULL, NULL, - 1 /* 45 seconds */); - cFYI(1, ("Wrt seteof rc %d", rc)); - } - } else - rc = -EINVAL; - - if (rc != 0) { - /* Set file size by pathname rather than by handle - either because no valid, writeable file handle for - it was found or because there was an error setting - it by handle */ - rc = CIFSSMBSetEOF(xid, pTcon, full_path, - attrs->ia_size, false, - cifs_sb->local_nls, - cifs_sb->mnt_cifs_flags & - CIFS_MOUNT_MAP_SPECIAL_CHR); - cFYI(1, ("SetEOF by path (setattrs) rc = %d", rc)); - if ((rc == -EINVAL) || (rc == -EOPNOTSUPP)) { - __u16 netfid; - int oplock = 0; - - rc = SMBLegacyOpen(xid, pTcon, full_path, - FILE_OPEN, GENERIC_WRITE, - CREATE_NOT_DIR, &netfid, &oplock, - NULL, cifs_sb->local_nls, - cifs_sb->mnt_cifs_flags & - CIFS_MOUNT_MAP_SPECIAL_CHR); - if (rc == 0) { - unsigned int bytes_written; - rc = CIFSSMBWrite(xid, pTcon, - netfid, 0, - attrs->ia_size, - &bytes_written, NULL, - NULL, 1 /* 45 sec */); - cFYI(1, ("wrt seteof rc %d", rc)); - CIFSSMBClose(xid, pTcon, netfid); - } - - } - } - - /* Server is ok setting allocation size implicitly - no need - to call: - CIFSSMBSetEOF(xid, pTcon, full_path, attrs->ia_size, true, - cifs_sb->local_nls); - */ - - if (rc == 0) { - rc = cifs_vmtruncate(inode, attrs->ia_size); - cifs_truncate_page(inode->i_mapping, inode->i_size); - } else + rc = cifs_set_file_size(inode, attrs, xid, full_path); + if (rc != 0) goto cifs_setattr_exit; } -- cgit v1.2.3-70-g09d2 From 5ca33c6ac30596fc1403a092f46ea48c406734e4 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 23 Jul 2008 17:45:58 -0700 Subject: cifs: assorted endian annotations fs/cifs/cifssmb.c:3917:13: warning: incorrect type in assignment (different base types) fs/cifs/cifssmb.c:3917:13: expected bool [unsigned] [usertype] is_unicode fs/cifs/cifssmb.c:3917:13: got restricted __le16 The comment explains why __force is used here. fs/cifs/connect.c:458:16: warning: cast to restricted __be32 fs/cifs/connect.c:458:16: warning: cast to restricted __be32 fs/cifs/connect.c:458:16: warning: cast to restricted __be32 fs/cifs/connect.c:458:16: warning: cast to restricted __be32 fs/cifs/connect.c:458:16: warning: cast to restricted __be32 fs/cifs/connect.c:458:16: warning: cast to restricted __be32 Signed-off-by: Harvey Harrison Signed-off-by: Steve French --- fs/cifs/cifssmb.c | 5 ++++- fs/cifs/connect.c | 2 +- 2 files changed, 5 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifssmb.c b/fs/cifs/cifssmb.c index 4511b708f0f..7e175e1399d 100644 --- a/fs/cifs/cifssmb.c +++ b/fs/cifs/cifssmb.c @@ -3914,7 +3914,10 @@ parse_DFS_referrals(TRANSACTION2_GET_DFS_REFER_RSP *pSMBr, bool is_unicode; struct dfs_referral_level_3 *ref; - is_unicode = pSMBr->hdr.Flags2 & SMBFLG2_UNICODE; + if (pSMBr->hdr.Flags2 & SMBFLG2_UNICODE) + is_unicode = true; + else + is_unicode = false; *num_of_nodes = le16_to_cpu(pSMBr->NumberOfReferrals); if (*num_of_nodes < 1) { diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c index 64446272938..b51d5777cde 100644 --- a/fs/cifs/connect.c +++ b/fs/cifs/connect.c @@ -455,7 +455,7 @@ incomplete_rcv: /* Note that FC 1001 length is big endian on the wire, but we convert it here so it is always manipulated as host byte order */ - pdu_length = ntohl(smb_buffer->smb_buf_length); + pdu_length = be32_to_cpu((__force __be32)smb_buffer->smb_buf_length); smb_buffer->smb_buf_length = pdu_length; cFYI(1, ("rfc1002 length 0x%x", pdu_length+4)); -- cgit v1.2.3-70-g09d2 From f984c7b98253c541ed6e702ee521f6a84c8113d4 Mon Sep 17 00:00:00 2001 From: Alexey Dobriyan Date: Thu, 24 Jul 2008 02:34:24 +0000 Subject: Signed-off-by: Alexey Dobriyan Cc: Steven French Signed-off-by: Andrew Morton Signed-off-by: Steve French --- fs/cifs/cifs_debug.c | 636 +++++++++++++++++++-------------------------------- 1 file changed, 241 insertions(+), 395 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifs_debug.c b/fs/cifs/cifs_debug.c index cc950f69e51..73925a77375 100644 --- a/fs/cifs/cifs_debug.c +++ b/fs/cifs/cifs_debug.c @@ -107,9 +107,7 @@ void cifs_dump_mids(struct TCP_Server_Info *server) #endif /* CONFIG_CIFS_DEBUG2 */ #ifdef CONFIG_PROC_FS -static int -cifs_debug_data_read(char *buf, char **beginBuffer, off_t offset, - int count, int *eof, void *data) +static int cifs_debug_data_proc_show(struct seq_file *m, void *v) { struct list_head *tmp; struct list_head *tmp1; @@ -117,23 +115,13 @@ cifs_debug_data_read(char *buf, char **beginBuffer, off_t offset, struct cifsSesInfo *ses; struct cifsTconInfo *tcon; int i; - int length = 0; - char *original_buf = buf; - *beginBuffer = buf + offset; - - length = - sprintf(buf, + seq_puts(m, "Display Internal CIFS Data Structures for Debugging\n" "---------------------------------------------------\n"); - buf += length; - length = sprintf(buf, "CIFS Version %s\n", CIFS_VERSION); - buf += length; - length = sprintf(buf, - "Active VFS Requests: %d\n", GlobalTotalActiveXid); - buf += length; - length = sprintf(buf, "Servers:"); - buf += length; + seq_printf(m, "CIFS Version %s\n", CIFS_VERSION); + seq_printf(m, "Active VFS Requests: %d\n", GlobalTotalActiveXid); + seq_printf(m, "Servers:"); i = 0; read_lock(&GlobalSMBSeslock); @@ -142,11 +130,10 @@ cifs_debug_data_read(char *buf, char **beginBuffer, off_t offset, ses = list_entry(tmp, struct cifsSesInfo, cifsSessionList); if ((ses->serverDomain == NULL) || (ses->serverOS == NULL) || (ses->serverNOS == NULL)) { - buf += sprintf(buf, "\nentry for %s not fully " + seq_printf(m, "\nentry for %s not fully " "displayed\n\t", ses->serverName); } else { - length = - sprintf(buf, + seq_printf(m, "\n%d) Name: %s Domain: %s Mounts: %d OS:" " %s \n\tNOS: %s\tCapability: 0x%x\n\tSMB" " session status: %d\t", @@ -154,10 +141,9 @@ cifs_debug_data_read(char *buf, char **beginBuffer, off_t offset, atomic_read(&ses->inUse), ses->serverOS, ses->serverNOS, ses->capabilities, ses->status); - buf += length; } if (ses->server) { - buf += sprintf(buf, "TCP status: %d\n\tLocal Users To " + seq_printf(m, "TCP status: %d\n\tLocal Users To " "Server: %d SecMode: 0x%x Req On Wire: %d", ses->server->tcpStatus, atomic_read(&ses->server->socketUseCount), @@ -165,13 +151,12 @@ cifs_debug_data_read(char *buf, char **beginBuffer, off_t offset, atomic_read(&ses->server->inFlight)); #ifdef CONFIG_CIFS_STATS2 - buf += sprintf(buf, " In Send: %d In MaxReq Wait: %d", + seq_printf(m, " In Send: %d In MaxReq Wait: %d", atomic_read(&ses->server->inSend), atomic_read(&ses->server->num_waiters)); #endif - length = sprintf(buf, "\nMIDs:\n"); - buf += length; + seq_puts(m, "\nMIDs:\n"); spin_lock(&GlobalMid_Lock); list_for_each(tmp1, &ses->server->pending_mid_q) { @@ -179,7 +164,7 @@ cifs_debug_data_read(char *buf, char **beginBuffer, off_t offset, mid_q_entry, qhead); if (mid_entry) { - length = sprintf(buf, + seq_printf(m, "State: %d com: %d pid:" " %d tsk: %p mid %d\n", mid_entry->midState, @@ -187,7 +172,6 @@ cifs_debug_data_read(char *buf, char **beginBuffer, off_t offset, mid_entry->pid, mid_entry->tsk, mid_entry->mid); - buf += length; } } spin_unlock(&GlobalMid_Lock); @@ -195,11 +179,9 @@ cifs_debug_data_read(char *buf, char **beginBuffer, off_t offset, } read_unlock(&GlobalSMBSeslock); - sprintf(buf, "\n"); - buf++; + seq_putc(m, '\n'); - length = sprintf(buf, "Shares:"); - buf += length; + seq_puts(m, "Shares:"); i = 0; read_lock(&GlobalSMBSeslock); @@ -208,62 +190,52 @@ cifs_debug_data_read(char *buf, char **beginBuffer, off_t offset, i++; tcon = list_entry(tmp, struct cifsTconInfo, cifsConnectionList); dev_type = le32_to_cpu(tcon->fsDevInfo.DeviceType); - length = sprintf(buf, "\n%d) %s Uses: %d ", i, + seq_printf(m, "\n%d) %s Uses: %d ", i, tcon->treeName, atomic_read(&tcon->useCount)); - buf += length; if (tcon->nativeFileSystem) { - length = sprintf(buf, "Type: %s ", + seq_printf(m, "Type: %s ", tcon->nativeFileSystem); - buf += length; } - length = sprintf(buf, "DevInfo: 0x%x Attributes: 0x%x" + seq_printf(m, "DevInfo: 0x%x Attributes: 0x%x" "\nPathComponentMax: %d Status: %d", le32_to_cpu(tcon->fsDevInfo.DeviceCharacteristics), le32_to_cpu(tcon->fsAttrInfo.Attributes), le32_to_cpu(tcon->fsAttrInfo.MaxPathNameComponentLength), tcon->tidStatus); - buf += length; if (dev_type == FILE_DEVICE_DISK) - length = sprintf(buf, " type: DISK "); + seq_puts(m, " type: DISK "); else if (dev_type == FILE_DEVICE_CD_ROM) - length = sprintf(buf, " type: CDROM "); + seq_puts(m, " type: CDROM "); else - length = - sprintf(buf, " type: %d ", dev_type); - buf += length; + seq_printf(m, " type: %d ", dev_type); if (tcon->tidStatus == CifsNeedReconnect) { - buf += sprintf(buf, "\tDISCONNECTED "); - length += 14; + seq_puts(m, "\tDISCONNECTED "); } } read_unlock(&GlobalSMBSeslock); - length = sprintf(buf, "\n"); - buf += length; + seq_putc(m, '\n'); /* BB add code to dump additional info such as TCP session info now */ - /* Now calculate total size of returned data */ - length = buf - original_buf; - - if (offset + count >= length) - *eof = 1; - if (length < offset) { - *eof = 1; - return 0; - } else { - length = length - offset; - } - if (length > count) - length = count; + return 0; +} - return length; +static int cifs_debug_data_proc_open(struct inode *inode, struct file *file) +{ + return single_open(file, cifs_debug_data_proc_show, NULL); } -#ifdef CONFIG_CIFS_STATS +static const struct file_operations cifs_debug_data_proc_fops = { + .owner = THIS_MODULE, + .open = cifs_debug_data_proc_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, +}; -static int -cifs_stats_write(struct file *file, const char __user *buffer, - unsigned long count, void *data) +#ifdef CONFIG_CIFS_STATS +static ssize_t cifs_stats_proc_write(struct file *file, + const char __user *buffer, size_t count, loff_t *ppos) { char c; int rc; @@ -307,236 +279,127 @@ cifs_stats_write(struct file *file, const char __user *buffer, return count; } -static int -cifs_stats_read(char *buf, char **beginBuffer, off_t offset, - int count, int *eof, void *data) +static int cifs_stats_proc_show(struct seq_file *m, void *v) { - int item_length, i, length; + int i; struct list_head *tmp; struct cifsTconInfo *tcon; - *beginBuffer = buf + offset; - - length = sprintf(buf, + seq_printf(m, "Resources in use\nCIFS Session: %d\n", sesInfoAllocCount.counter); - buf += length; - item_length = - sprintf(buf, "Share (unique mount targets): %d\n", + seq_printf(m, "Share (unique mount targets): %d\n", tconInfoAllocCount.counter); - length += item_length; - buf += item_length; - item_length = - sprintf(buf, "SMB Request/Response Buffer: %d Pool size: %d\n", + seq_printf(m, "SMB Request/Response Buffer: %d Pool size: %d\n", bufAllocCount.counter, cifs_min_rcv + tcpSesAllocCount.counter); - length += item_length; - buf += item_length; - item_length = - sprintf(buf, "SMB Small Req/Resp Buffer: %d Pool size: %d\n", + seq_printf(m, "SMB Small Req/Resp Buffer: %d Pool size: %d\n", smBufAllocCount.counter, cifs_min_small); - length += item_length; - buf += item_length; #ifdef CONFIG_CIFS_STATS2 - item_length = sprintf(buf, "Total Large %d Small %d Allocations\n", + seq_printf(m, "Total Large %d Small %d Allocations\n", atomic_read(&totBufAllocCount), atomic_read(&totSmBufAllocCount)); - length += item_length; - buf += item_length; #endif /* CONFIG_CIFS_STATS2 */ - item_length = - sprintf(buf, "Operations (MIDs): %d\n", - midCount.counter); - length += item_length; - buf += item_length; - item_length = sprintf(buf, + seq_printf(m, "Operations (MIDs): %d\n", midCount.counter); + seq_printf(m, "\n%d session %d share reconnects\n", tcpSesReconnectCount.counter, tconInfoReconnectCount.counter); - length += item_length; - buf += item_length; - item_length = sprintf(buf, + seq_printf(m, "Total vfs operations: %d maximum at one time: %d\n", GlobalCurrentXid, GlobalMaxActiveXid); - length += item_length; - buf += item_length; i = 0; read_lock(&GlobalSMBSeslock); list_for_each(tmp, &GlobalTreeConnectionList) { i++; tcon = list_entry(tmp, struct cifsTconInfo, cifsConnectionList); - item_length = sprintf(buf, "\n%d) %s", i, tcon->treeName); - buf += item_length; - length += item_length; - if (tcon->tidStatus == CifsNeedReconnect) { - buf += sprintf(buf, "\tDISCONNECTED "); - length += 14; - } - item_length = sprintf(buf, "\nSMBs: %d Oplock Breaks: %d", + seq_printf(m, "\n%d) %s", i, tcon->treeName); + if (tcon->tidStatus == CifsNeedReconnect) + seq_puts(m, "\tDISCONNECTED "); + seq_printf(m, "\nSMBs: %d Oplock Breaks: %d", atomic_read(&tcon->num_smbs_sent), atomic_read(&tcon->num_oplock_brks)); - buf += item_length; - length += item_length; - item_length = sprintf(buf, "\nReads: %d Bytes: %lld", + seq_printf(m, "\nReads: %d Bytes: %lld", atomic_read(&tcon->num_reads), (long long)(tcon->bytes_read)); - buf += item_length; - length += item_length; - item_length = sprintf(buf, "\nWrites: %d Bytes: %lld", + seq_printf(m, "\nWrites: %d Bytes: %lld", atomic_read(&tcon->num_writes), (long long)(tcon->bytes_written)); - buf += item_length; - length += item_length; - item_length = sprintf(buf, + seq_printf(m, "\nLocks: %d HardLinks: %d Symlinks: %d", atomic_read(&tcon->num_locks), atomic_read(&tcon->num_hardlinks), atomic_read(&tcon->num_symlinks)); - buf += item_length; - length += item_length; - item_length = sprintf(buf, "\nOpens: %d Closes: %d Deletes: %d", + seq_printf(m, "\nOpens: %d Closes: %d Deletes: %d", atomic_read(&tcon->num_opens), atomic_read(&tcon->num_closes), atomic_read(&tcon->num_deletes)); - buf += item_length; - length += item_length; - item_length = sprintf(buf, "\nMkdirs: %d Rmdirs: %d", + seq_printf(m, "\nMkdirs: %d Rmdirs: %d", atomic_read(&tcon->num_mkdirs), atomic_read(&tcon->num_rmdirs)); - buf += item_length; - length += item_length; - item_length = sprintf(buf, "\nRenames: %d T2 Renames %d", + seq_printf(m, "\nRenames: %d T2 Renames %d", atomic_read(&tcon->num_renames), atomic_read(&tcon->num_t2renames)); - buf += item_length; - length += item_length; - item_length = sprintf(buf, "\nFindFirst: %d FNext %d FClose %d", + seq_printf(m, "\nFindFirst: %d FNext %d FClose %d", atomic_read(&tcon->num_ffirst), atomic_read(&tcon->num_fnext), atomic_read(&tcon->num_fclose)); - buf += item_length; - length += item_length; } read_unlock(&GlobalSMBSeslock); - buf += sprintf(buf, "\n"); - length++; - - if (offset + count >= length) - *eof = 1; - if (length < offset) { - *eof = 1; - return 0; - } else { - length = length - offset; - } - if (length > count) - length = count; + seq_putc(m, '\n'); + return 0; +} - return length; +static int cifs_stats_proc_open(struct inode *inode, struct file *file) +{ + return single_open(file, cifs_stats_proc_show, NULL); } + +static const struct file_operations cifs_stats_proc_fops = { + .owner = THIS_MODULE, + .open = cifs_stats_proc_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, + .write = cifs_stats_proc_write, +}; #endif /* STATS */ static struct proc_dir_entry *proc_fs_cifs; -read_proc_t cifs_txanchor_read; -static read_proc_t cifsFYI_read; -static write_proc_t cifsFYI_write; -static read_proc_t oplockEnabled_read; -static write_proc_t oplockEnabled_write; -static read_proc_t lookupFlag_read; -static write_proc_t lookupFlag_write; -static read_proc_t traceSMB_read; -static write_proc_t traceSMB_write; -static read_proc_t multiuser_mount_read; -static write_proc_t multiuser_mount_write; -static read_proc_t security_flags_read; -static write_proc_t security_flags_write; -/* static read_proc_t ntlmv2_enabled_read; -static write_proc_t ntlmv2_enabled_write; -static read_proc_t packet_signing_enabled_read; -static write_proc_t packet_signing_enabled_write;*/ -static read_proc_t experimEnabled_read; -static write_proc_t experimEnabled_write; -static read_proc_t linuxExtensionsEnabled_read; -static write_proc_t linuxExtensionsEnabled_write; +static const struct file_operations cifsFYI_proc_fops; +static const struct file_operations cifs_oplock_proc_fops; +static const struct file_operations cifs_lookup_cache_proc_fops; +static const struct file_operations traceSMB_proc_fops; +static const struct file_operations cifs_multiuser_mount_proc_fops; +static const struct file_operations cifs_security_flags_proc_fops; +static const struct file_operations cifs_experimental_proc_fops; +static const struct file_operations cifs_linux_ext_proc_fops; void cifs_proc_init(void) { - struct proc_dir_entry *pde; - proc_fs_cifs = proc_mkdir("fs/cifs", NULL); if (proc_fs_cifs == NULL) return; proc_fs_cifs->owner = THIS_MODULE; - create_proc_read_entry("DebugData", 0, proc_fs_cifs, - cifs_debug_data_read, NULL); + proc_create("DebugData", 0, proc_fs_cifs, &cifs_debug_data_proc_fops); #ifdef CONFIG_CIFS_STATS - pde = create_proc_read_entry("Stats", 0, proc_fs_cifs, - cifs_stats_read, NULL); - if (pde) - pde->write_proc = cifs_stats_write; + proc_create("Stats", 0, proc_fs_cifs, &cifs_stats_proc_fops); #endif /* STATS */ - pde = create_proc_read_entry("cifsFYI", 0, proc_fs_cifs, - cifsFYI_read, NULL); - if (pde) - pde->write_proc = cifsFYI_write; - - pde = - create_proc_read_entry("traceSMB", 0, proc_fs_cifs, - traceSMB_read, NULL); - if (pde) - pde->write_proc = traceSMB_write; - - pde = create_proc_read_entry("OplockEnabled", 0, proc_fs_cifs, - oplockEnabled_read, NULL); - if (pde) - pde->write_proc = oplockEnabled_write; - - pde = create_proc_read_entry("Experimental", 0, proc_fs_cifs, - experimEnabled_read, NULL); - if (pde) - pde->write_proc = experimEnabled_write; - - pde = create_proc_read_entry("LinuxExtensionsEnabled", 0, proc_fs_cifs, - linuxExtensionsEnabled_read, NULL); - if (pde) - pde->write_proc = linuxExtensionsEnabled_write; - - pde = - create_proc_read_entry("MultiuserMount", 0, proc_fs_cifs, - multiuser_mount_read, NULL); - if (pde) - pde->write_proc = multiuser_mount_write; - - pde = - create_proc_read_entry("SecurityFlags", 0, proc_fs_cifs, - security_flags_read, NULL); - if (pde) - pde->write_proc = security_flags_write; - - pde = - create_proc_read_entry("LookupCacheEnabled", 0, proc_fs_cifs, - lookupFlag_read, NULL); - if (pde) - pde->write_proc = lookupFlag_write; - -/* pde = - create_proc_read_entry("NTLMV2Enabled", 0, proc_fs_cifs, - ntlmv2_enabled_read, NULL); - if (pde) - pde->write_proc = ntlmv2_enabled_write; - - pde = - create_proc_read_entry("PacketSigningEnabled", 0, proc_fs_cifs, - packet_signing_enabled_read, NULL); - if (pde) - pde->write_proc = packet_signing_enabled_write;*/ + proc_create("cifsFYI", 0, proc_fs_cifs, &cifsFYI_proc_fops); + proc_create("traceSMB", 0, proc_fs_cifs, &traceSMB_proc_fops); + proc_create("OplockEnabled", 0, proc_fs_cifs, &cifs_oplock_proc_fops); + proc_create("Experimental", 0, proc_fs_cifs, &cifs_experimental_proc_fops); + proc_create("LinuxExtensionsEnabled", 0, proc_fs_cifs, &cifs_linux_ext_proc_fops); + proc_create("MultiuserMount", 0, proc_fs_cifs, &cifs_multiuser_mount_proc_fops); + proc_create("SecurityFlags", 0, proc_fs_cifs, &cifs_security_flags_proc_fops); + proc_create("LookupCacheEnabled", 0, proc_fs_cifs, &cifs_lookup_cache_proc_fops); } void @@ -553,39 +416,26 @@ cifs_proc_clean(void) #endif remove_proc_entry("MultiuserMount", proc_fs_cifs); remove_proc_entry("OplockEnabled", proc_fs_cifs); -/* remove_proc_entry("NTLMV2Enabled",proc_fs_cifs); */ remove_proc_entry("SecurityFlags", proc_fs_cifs); -/* remove_proc_entry("PacketSigningEnabled", proc_fs_cifs); */ remove_proc_entry("LinuxExtensionsEnabled", proc_fs_cifs); remove_proc_entry("Experimental", proc_fs_cifs); remove_proc_entry("LookupCacheEnabled", proc_fs_cifs); remove_proc_entry("fs/cifs", NULL); } -static int -cifsFYI_read(char *page, char **start, off_t off, int count, - int *eof, void *data) +static int cifsFYI_proc_show(struct seq_file *m, void *v) { - int len; - - len = sprintf(page, "%d\n", cifsFYI); - - len -= off; - *start = page + off; - - if (len > count) - len = count; - else - *eof = 1; - - if (len < 0) - len = 0; + seq_printf(m, "%d\n", cifsFYI); + return 0; +} - return len; +static int cifsFYI_proc_open(struct inode *inode, struct file *file) +{ + return single_open(file, cifsFYI_proc_show, NULL); } -static int -cifsFYI_write(struct file *file, const char __user *buffer, - unsigned long count, void *data) + +static ssize_t cifsFYI_proc_write(struct file *file, const char __user *buffer, + size_t count, loff_t *ppos) { char c; int rc; @@ -603,30 +453,28 @@ cifsFYI_write(struct file *file, const char __user *buffer, return count; } -static int -oplockEnabled_read(char *page, char **start, off_t off, - int count, int *eof, void *data) -{ - int len; - - len = sprintf(page, "%d\n", oplockEnabled); +static const struct file_operations cifsFYI_proc_fops = { + .owner = THIS_MODULE, + .open = cifsFYI_proc_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, + .write = cifsFYI_proc_write, +}; - len -= off; - *start = page + off; - - if (len > count) - len = count; - else - *eof = 1; - - if (len < 0) - len = 0; +static int cifs_oplock_proc_show(struct seq_file *m, void *v) +{ + seq_printf(m, "%d\n", oplockEnabled); + return 0; +} - return len; +static int cifs_oplock_proc_open(struct inode *inode, struct file *file) +{ + return single_open(file, cifs_oplock_proc_show, NULL); } -static int -oplockEnabled_write(struct file *file, const char __user *buffer, - unsigned long count, void *data) + +static ssize_t cifs_oplock_proc_write(struct file *file, + const char __user *buffer, size_t count, loff_t *ppos) { char c; int rc; @@ -642,30 +490,28 @@ oplockEnabled_write(struct file *file, const char __user *buffer, return count; } -static int -experimEnabled_read(char *page, char **start, off_t off, - int count, int *eof, void *data) -{ - int len; - - len = sprintf(page, "%d\n", experimEnabled); - - len -= off; - *start = page + off; - - if (len > count) - len = count; - else - *eof = 1; +static const struct file_operations cifs_oplock_proc_fops = { + .owner = THIS_MODULE, + .open = cifs_oplock_proc_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, + .write = cifs_oplock_proc_write, +}; - if (len < 0) - len = 0; +static int cifs_experimental_proc_show(struct seq_file *m, void *v) +{ + seq_printf(m, "%d\n", experimEnabled); + return 0; +} - return len; +static int cifs_experimental_proc_open(struct inode *inode, struct file *file) +{ + return single_open(file, cifs_experimental_proc_show, NULL); } -static int -experimEnabled_write(struct file *file, const char __user *buffer, - unsigned long count, void *data) + +static ssize_t cifs_experimental_proc_write(struct file *file, + const char __user *buffer, size_t count, loff_t *ppos) { char c; int rc; @@ -683,29 +529,28 @@ experimEnabled_write(struct file *file, const char __user *buffer, return count; } -static int -linuxExtensionsEnabled_read(char *page, char **start, off_t off, - int count, int *eof, void *data) -{ - int len; - - len = sprintf(page, "%d\n", linuxExtEnabled); - len -= off; - *start = page + off; - - if (len > count) - len = count; - else - *eof = 1; +static const struct file_operations cifs_experimental_proc_fops = { + .owner = THIS_MODULE, + .open = cifs_experimental_proc_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, + .write = cifs_experimental_proc_write, +}; - if (len < 0) - len = 0; +static int cifs_linux_ext_proc_show(struct seq_file *m, void *v) +{ + seq_printf(m, "%d\n", linuxExtEnabled); + return 0; +} - return len; +static int cifs_linux_ext_proc_open(struct inode *inode, struct file *file) +{ + return single_open(file, cifs_linux_ext_proc_show, NULL); } -static int -linuxExtensionsEnabled_write(struct file *file, const char __user *buffer, - unsigned long count, void *data) + +static ssize_t cifs_linux_ext_proc_write(struct file *file, + const char __user *buffer, size_t count, loff_t *ppos) { char c; int rc; @@ -721,31 +566,28 @@ linuxExtensionsEnabled_write(struct file *file, const char __user *buffer, return count; } +static const struct file_operations cifs_linux_ext_proc_fops = { + .owner = THIS_MODULE, + .open = cifs_linux_ext_proc_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, + .write = cifs_linux_ext_proc_write, +}; -static int -lookupFlag_read(char *page, char **start, off_t off, - int count, int *eof, void *data) +static int cifs_lookup_cache_proc_show(struct seq_file *m, void *v) { - int len; - - len = sprintf(page, "%d\n", lookupCacheEnabled); - - len -= off; - *start = page + off; - - if (len > count) - len = count; - else - *eof = 1; - - if (len < 0) - len = 0; + seq_printf(m, "%d\n", lookupCacheEnabled); + return 0; +} - return len; +static int cifs_lookup_cache_proc_open(struct inode *inode, struct file *file) +{ + return single_open(file, cifs_lookup_cache_proc_show, NULL); } -static int -lookupFlag_write(struct file *file, const char __user *buffer, - unsigned long count, void *data) + +static ssize_t cifs_lookup_cache_proc_write(struct file *file, + const char __user *buffer, size_t count, loff_t *ppos) { char c; int rc; @@ -760,30 +602,29 @@ lookupFlag_write(struct file *file, const char __user *buffer, return count; } -static int -traceSMB_read(char *page, char **start, off_t off, int count, - int *eof, void *data) -{ - int len; - len = sprintf(page, "%d\n", traceSMB); +static const struct file_operations cifs_lookup_cache_proc_fops = { + .owner = THIS_MODULE, + .open = cifs_lookup_cache_proc_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, + .write = cifs_lookup_cache_proc_write, +}; - len -= off; - *start = page + off; - - if (len > count) - len = count; - else - *eof = 1; - - if (len < 0) - len = 0; +static int traceSMB_proc_show(struct seq_file *m, void *v) +{ + seq_printf(m, "%d\n", traceSMB); + return 0; +} - return len; +static int traceSMB_proc_open(struct inode *inode, struct file *file) +{ + return single_open(file, traceSMB_proc_show, NULL); } -static int -traceSMB_write(struct file *file, const char __user *buffer, - unsigned long count, void *data) + +static ssize_t traceSMB_proc_write(struct file *file, const char __user *buffer, + size_t count, loff_t *ppos) { char c; int rc; @@ -799,30 +640,28 @@ traceSMB_write(struct file *file, const char __user *buffer, return count; } -static int -multiuser_mount_read(char *page, char **start, off_t off, - int count, int *eof, void *data) -{ - int len; - - len = sprintf(page, "%d\n", multiuser_mount); - - len -= off; - *start = page + off; +static const struct file_operations traceSMB_proc_fops = { + .owner = THIS_MODULE, + .open = traceSMB_proc_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, + .write = traceSMB_proc_write, +}; - if (len > count) - len = count; - else - *eof = 1; - - if (len < 0) - len = 0; +static int cifs_multiuser_mount_proc_show(struct seq_file *m, void *v) +{ + seq_printf(m, "%d\n", multiuser_mount); + return 0; +} - return len; +static int cifs_multiuser_mount_proc_open(struct inode *inode, struct file *file) +{ + return single_open(file, cifs_multiuser_mount_proc_show, NULL); } -static int -multiuser_mount_write(struct file *file, const char __user *buffer, - unsigned long count, void *data) + +static ssize_t cifs_multiuser_mount_proc_write(struct file *file, + const char __user *buffer, size_t count, loff_t *ppos) { char c; int rc; @@ -838,30 +677,28 @@ multiuser_mount_write(struct file *file, const char __user *buffer, return count; } -static int -security_flags_read(char *page, char **start, off_t off, - int count, int *eof, void *data) -{ - int len; - - len = sprintf(page, "0x%x\n", extended_security); - - len -= off; - *start = page + off; +static const struct file_operations cifs_multiuser_mount_proc_fops = { + .owner = THIS_MODULE, + .open = cifs_multiuser_mount_proc_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, + .write = cifs_multiuser_mount_proc_write, +}; - if (len > count) - len = count; - else - *eof = 1; - - if (len < 0) - len = 0; +static int cifs_security_flags_proc_show(struct seq_file *m, void *v) +{ + seq_printf(m, "0x%x\n", extended_security); + return 0; +} - return len; +static int cifs_security_flags_proc_open(struct inode *inode, struct file *file) +{ + return single_open(file, cifs_security_flags_proc_show, NULL); } -static int -security_flags_write(struct file *file, const char __user *buffer, - unsigned long count, void *data) + +static ssize_t cifs_security_flags_proc_write(struct file *file, + const char __user *buffer, size_t count, loff_t *ppos) { unsigned int flags; char flags_string[12]; @@ -917,6 +754,15 @@ security_flags_write(struct file *file, const char __user *buffer, /* BB should we turn on MAY flags for other MUST options? */ return count; } + +static const struct file_operations cifs_security_flags_proc_fops = { + .owner = THIS_MODULE, + .open = cifs_security_flags_proc_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, + .write = cifs_security_flags_proc_write, +}; #else inline void cifs_proc_init(void) { -- cgit v1.2.3-70-g09d2 From 99b1f5b2f6cd2f65cce02c5f63302df5878a5fbc Mon Sep 17 00:00:00 2001 From: Steve French Date: Thu, 24 Jul 2008 02:37:45 +0000 Subject: [CIFS] remove checkpatch warning Signed-off-by: Steve French --- fs/cifs/cifs_debug.c | 19 ++++++++++++------- 1 file changed, 12 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifs_debug.c b/fs/cifs/cifs_debug.c index 73925a77375..c213aa40064 100644 --- a/fs/cifs/cifs_debug.c +++ b/fs/cifs/cifs_debug.c @@ -395,11 +395,16 @@ cifs_proc_init(void) proc_create("cifsFYI", 0, proc_fs_cifs, &cifsFYI_proc_fops); proc_create("traceSMB", 0, proc_fs_cifs, &traceSMB_proc_fops); proc_create("OplockEnabled", 0, proc_fs_cifs, &cifs_oplock_proc_fops); - proc_create("Experimental", 0, proc_fs_cifs, &cifs_experimental_proc_fops); - proc_create("LinuxExtensionsEnabled", 0, proc_fs_cifs, &cifs_linux_ext_proc_fops); - proc_create("MultiuserMount", 0, proc_fs_cifs, &cifs_multiuser_mount_proc_fops); - proc_create("SecurityFlags", 0, proc_fs_cifs, &cifs_security_flags_proc_fops); - proc_create("LookupCacheEnabled", 0, proc_fs_cifs, &cifs_lookup_cache_proc_fops); + proc_create("Experimental", 0, proc_fs_cifs, + &cifs_experimental_proc_fops); + proc_create("LinuxExtensionsEnabled", 0, proc_fs_cifs, + &cifs_linux_ext_proc_fops); + proc_create("MultiuserMount", 0, proc_fs_cifs, + &cifs_multiuser_mount_proc_fops); + proc_create("SecurityFlags", 0, proc_fs_cifs, + &cifs_security_flags_proc_fops); + proc_create("LookupCacheEnabled", 0, proc_fs_cifs, + &cifs_lookup_cache_proc_fops); } void @@ -655,9 +660,9 @@ static int cifs_multiuser_mount_proc_show(struct seq_file *m, void *v) return 0; } -static int cifs_multiuser_mount_proc_open(struct inode *inode, struct file *file) +static int cifs_multiuser_mount_proc_open(struct inode *inode, struct file *fh) { - return single_open(file, cifs_multiuser_mount_proc_show, NULL); + return single_open(fh, cifs_multiuser_mount_proc_show, NULL); } static ssize_t cifs_multiuser_mount_proc_write(struct file *file, -- cgit v1.2.3-70-g09d2 From 76c510ad2e7d56cfe8f2cc7b23783e5c687cf704 Mon Sep 17 00:00:00 2001 From: Shirish Pargaonkar Date: Thu, 24 Jul 2008 14:48:33 +0000 Subject: [CIFS] Fix possible double free if search immediately after search rewind fails Signed-off-by: Shirish Pargaonkar Signed-off-by: Steve French --- fs/cifs/readdir.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/cifs/readdir.c b/fs/cifs/readdir.c index 83f30695488..5f40ed3473f 100644 --- a/fs/cifs/readdir.c +++ b/fs/cifs/readdir.c @@ -690,6 +690,7 @@ static int find_cifs_entry(const int xid, struct cifsTconInfo *pTcon, else cifs_buf_release(cifsFile->srch_inf. ntwrk_buf_start); + cifsFile->srch_inf.ntwrk_buf_start = NULL; } rc = initiate_cifs_search(xid, file); if (rc) { -- cgit v1.2.3-70-g09d2 From b1910ad6222a705650dc991c003af43b94cdb3e1 Mon Sep 17 00:00:00 2001 From: Shirish Pargaonkar Date: Thu, 24 Jul 2008 14:53:20 +0000 Subject: [CIFS] Fix improper endian conversion of ACL subauth field In mode_to_acl when converting a Unix mode to a Windows ACL the subauth fields of the SID in the ACL were translated incorrectly on bigendian architectures Signed-off-by: Steve French --- fs/cifs/cifsacl.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifsacl.c b/fs/cifs/cifsacl.c index 0e9fc2ba90e..18faf65b911 100644 --- a/fs/cifs/cifsacl.c +++ b/fs/cifs/cifsacl.c @@ -169,8 +169,7 @@ static void copy_sec_desc(const struct cifs_ntsd *pntsd, for (i = 0; i < 6; i++) ngroup_sid_ptr->authority[i] = group_sid_ptr->authority[i]; for (i = 0; i < 5; i++) - ngroup_sid_ptr->sub_auth[i] = - cpu_to_le32(group_sid_ptr->sub_auth[i]); + ngroup_sid_ptr->sub_auth[i] = group_sid_ptr->sub_auth[i]; return; } -- cgit v1.2.3-70-g09d2 From ef571cadd516e7ffcdeac6bb8054e5908fcccfcf Mon Sep 17 00:00:00 2001 From: Shirish Pargaonkar Date: Thu, 24 Jul 2008 15:56:05 +0000 Subject: [CIFS] Fix warnings from checkpatch Signed-off-by: Steve French --- fs/cifs/cifs_debug.c | 4 ++-- fs/cifs/cifsacl.c | 38 +++++++++++++++++++------------------- fs/cifs/cifsencrypt.c | 3 +-- fs/cifs/cifsglob.h | 6 +++--- fs/cifs/cifspdu.h | 8 ++++---- fs/cifs/cifssmb.c | 5 ++--- 6 files changed, 31 insertions(+), 33 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifs_debug.c b/fs/cifs/cifs_debug.c index c213aa40064..688a2d42153 100644 --- a/fs/cifs/cifs_debug.c +++ b/fs/cifs/cifs_debug.c @@ -208,9 +208,9 @@ static int cifs_debug_data_proc_show(struct seq_file *m, void *v) seq_puts(m, " type: CDROM "); else seq_printf(m, " type: %d ", dev_type); - if (tcon->tidStatus == CifsNeedReconnect) { + + if (tcon->tidStatus == CifsNeedReconnect) seq_puts(m, "\tDISCONNECTED "); - } } read_unlock(&GlobalSMBSeslock); diff --git a/fs/cifs/cifsacl.c b/fs/cifs/cifsacl.c index 18faf65b911..57ecdc83c26 100644 --- a/fs/cifs/cifsacl.c +++ b/fs/cifs/cifsacl.c @@ -56,7 +56,7 @@ int match_sid(struct cifs_sid *ctsid) struct cifs_sid *cwsid; if (!ctsid) - return (-1); + return -1; for (i = 0; i < NUM_WK_SIDS; ++i) { cwsid = &(wksidarr[i].cifssid); @@ -87,11 +87,11 @@ int match_sid(struct cifs_sid *ctsid) } cFYI(1, ("matching sid: %s\n", wksidarr[i].sidname)); - return (0); /* sids compare/match */ + return 0; /* sids compare/match */ } cFYI(1, ("No matching sid")); - return (-1); + return -1; } /* if the two SIDs (roughly equivalent to a UUID for a user or group) are @@ -102,16 +102,16 @@ int compare_sids(const struct cifs_sid *ctsid, const struct cifs_sid *cwsid) int num_subauth, num_sat, num_saw; if ((!ctsid) || (!cwsid)) - return (0); + return 0; /* compare the revision */ if (ctsid->revision != cwsid->revision) - return (0); + return 0; /* compare all of the six auth values */ for (i = 0; i < 6; ++i) { if (ctsid->authority[i] != cwsid->authority[i]) - return (0); + return 0; } /* compare all of the subauth values if any */ @@ -121,11 +121,11 @@ int compare_sids(const struct cifs_sid *ctsid, const struct cifs_sid *cwsid) if (num_subauth) { for (i = 0; i < num_subauth; ++i) { if (ctsid->sub_auth[i] != cwsid->sub_auth[i]) - return (0); + return 0; } } - return (1); /* sids compare/match */ + return 1; /* sids compare/match */ } @@ -284,7 +284,7 @@ static __u16 fill_ace_for_sid(struct cifs_ace *pntace, size = 1 + 1 + 2 + 4 + 1 + 1 + 6 + (psid->num_subauth * 4); pntace->size = cpu_to_le16(size); - return (size); + return size; } @@ -425,7 +425,7 @@ static int set_chmod_dacl(struct cifs_acl *pndacl, struct cifs_sid *pownersid, pndacl->size = cpu_to_le16(size + sizeof(struct cifs_acl)); pndacl->num_aces = cpu_to_le32(3); - return (0); + return 0; } @@ -509,7 +509,7 @@ static int parse_sec_desc(struct cifs_ntsd *pntsd, int acl_len, sizeof(struct cifs_sid)); */ - return (0); + return 0; } @@ -526,7 +526,7 @@ static int build_sec_desc(struct cifs_ntsd *pntsd, struct cifs_ntsd *pnntsd, struct cifs_acl *ndacl_ptr = NULL; /* no need for SACL ptr */ if ((inode == NULL) || (pntsd == NULL) || (pnntsd == NULL)) - return (-EIO); + return -EIO; owner_sid_ptr = (struct cifs_sid *)((char *)pntsd + le32_to_cpu(pntsd->osidoffset)); @@ -549,7 +549,7 @@ static int build_sec_desc(struct cifs_ntsd *pntsd, struct cifs_ntsd *pnntsd, /* copy security descriptor control portion and owner and group sid */ copy_sec_desc(pntsd, pnntsd, sidsoffset); - return (rc); + return rc; } @@ -628,11 +628,11 @@ static int set_cifs_acl(struct cifs_ntsd *pnntsd, __u32 acllen, cFYI(DBG2, ("set ACL for %s from mode 0x%x", path, inode->i_mode)); if (!inode) - return (rc); + return rc; sb = inode->i_sb; if (sb == NULL) - return (rc); + return rc; cifs_sb = CIFS_SB(sb); xid = GetXid(); @@ -651,7 +651,7 @@ static int set_cifs_acl(struct cifs_ntsd *pnntsd, __u32 acllen, if (rc != 0) { cERROR(1, ("Unable to open file to set ACL")); FreeXid(xid); - return (rc); + return rc; } } @@ -664,7 +664,7 @@ static int set_cifs_acl(struct cifs_ntsd *pnntsd, __u32 acllen, FreeXid(xid); - return (rc); + return rc; } /* Translate the CIFS ACL (simlar to NTFS ACL) for a file into mode bits */ @@ -714,7 +714,7 @@ int mode_to_acl(struct inode *inode, const char *path, __u64 nmode) if (!pnntsd) { cERROR(1, ("Unable to allocate security descriptor")); kfree(pntsd); - return (-ENOMEM); + return -ENOMEM; } rc = build_sec_desc(pntsd, pnntsd, inode, nmode); @@ -731,6 +731,6 @@ int mode_to_acl(struct inode *inode, const char *path, __u64 nmode) kfree(pntsd); } - return (rc); + return rc; } #endif /* CONFIG_CIFS_EXPERIMENTAL */ diff --git a/fs/cifs/cifsencrypt.c b/fs/cifs/cifsencrypt.c index 4ff8939c6cc..83fd40dc1ef 100644 --- a/fs/cifs/cifsencrypt.c +++ b/fs/cifs/cifsencrypt.c @@ -310,9 +310,8 @@ void calc_lanman_hash(struct cifsSesInfo *ses, char *lnm_session_key) utf8 and other multibyte codepages each need their own strupper function since a byte at a time will ont work. */ - for (i = 0; i < CIFS_ENCPWD_SIZE; i++) { + for (i = 0; i < CIFS_ENCPWD_SIZE; i++) password_with_pad[i] = toupper(password_with_pad[i]); - } SMBencrypt(password_with_pad, ses->server->cryptKey, lnm_session_key); /* clear password before we return/free memory */ diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h index 9cfcf326ead..7e1cf262eff 100644 --- a/fs/cifs/cifsglob.h +++ b/fs/cifs/cifsglob.h @@ -27,7 +27,7 @@ #define MAX_SES_INFO 2 #define MAX_TCON_INFO 4 -#define MAX_TREE_SIZE 2 + MAX_SERVER_SIZE + 1 + MAX_SHARE_SIZE + 1 +#define MAX_TREE_SIZE (2 + MAX_SERVER_SIZE + 1 + MAX_SHARE_SIZE + 1) #define MAX_SERVER_SIZE 15 #define MAX_SHARE_SIZE 64 /* used to be 20, this should still be enough */ #define MAX_USERNAME_SIZE 32 /* 32 is to allow for 15 char names + null @@ -537,8 +537,8 @@ require use of the stronger protocol */ #endif /* WEAK_PW_HASH */ #define CIFSSEC_MUST_SEAL 0x40040 /* not supported yet */ -#define CIFSSEC_DEF CIFSSEC_MAY_SIGN | CIFSSEC_MAY_NTLM | CIFSSEC_MAY_NTLMV2 -#define CIFSSEC_MAX CIFSSEC_MUST_SIGN | CIFSSEC_MUST_NTLMV2 +#define CIFSSEC_DEF (CIFSSEC_MAY_SIGN | CIFSSEC_MAY_NTLM | CIFSSEC_MAY_NTLMV2) +#define CIFSSEC_MAX (CIFSSEC_MUST_SIGN | CIFSSEC_MUST_NTLMV2) #define CIFSSEC_AUTH_MASK (CIFSSEC_MAY_NTLM | CIFSSEC_MAY_NTLMV2 | CIFSSEC_MAY_LANMAN | CIFSSEC_MAY_PLNTXT | CIFSSEC_MAY_KRB5) /* ***************************************************************** diff --git a/fs/cifs/cifspdu.h b/fs/cifs/cifspdu.h index 0f327c224da..409abce1273 100644 --- a/fs/cifs/cifspdu.h +++ b/fs/cifs/cifspdu.h @@ -31,7 +31,7 @@ #else #define CIFS_PROT 0 #endif -#define POSIX_PROT CIFS_PROT+1 +#define POSIX_PROT (CIFS_PROT+1) #define BAD_PROT 0xFFFF /* SMB command codes */ @@ -341,7 +341,7 @@ #define CREATE_COMPLETE_IF_OPLK 0x00000100 /* should be zero */ #define CREATE_NO_EA_KNOWLEDGE 0x00000200 #define CREATE_EIGHT_DOT_THREE 0x00000400 /* doc says this is obsolete - "open for recovery" flag - should + "open for recovery" flag should be zero in any case */ #define CREATE_OPEN_FOR_RECOVERY 0x00000400 #define CREATE_RANDOM_ACCESS 0x00000800 @@ -414,8 +414,8 @@ struct smb_hdr { __u8 WordCount; } __attribute__((packed)); /* given a pointer to an smb_hdr retrieve the value of byte count */ -#define BCC(smb_var) ( *(__u16 *)((char *)smb_var + sizeof(struct smb_hdr) + (2 * smb_var->WordCount))) -#define BCC_LE(smb_var) ( *(__le16 *)((char *)smb_var + sizeof(struct smb_hdr) + (2 * smb_var->WordCount))) +#define BCC(smb_var) (*(__u16 *)((char *)smb_var + sizeof(struct smb_hdr) + (2 * smb_var->WordCount))) +#define BCC_LE(smb_var) (*(__le16 *)((char *)smb_var + sizeof(struct smb_hdr) + (2 * smb_var->WordCount))) /* given a pointer to an smb_hdr retrieve the pointer to the byte area */ #define pByteArea(smb_var) ((unsigned char *)smb_var + sizeof(struct smb_hdr) + (2 * smb_var->WordCount) + 2) diff --git a/fs/cifs/cifssmb.c b/fs/cifs/cifssmb.c index 7e175e1399d..c621ffa2ca9 100644 --- a/fs/cifs/cifssmb.c +++ b/fs/cifs/cifssmb.c @@ -686,11 +686,10 @@ CIFSSMBNegotiate(unsigned int xid, struct cifsSesInfo *ses) SecurityBlob, count - 16, &server->secType); - if (rc == 1) { + if (rc == 1) rc = 0; - } else { + else rc = -EINVAL; - } } } else server->capabilities &= ~CAP_EXTENDED_SECURITY; -- cgit v1.2.3-70-g09d2 From c748e1340e0de3fa7fed86f8bdf499be9242afff Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Wed, 23 Jul 2008 21:27:03 -0700 Subject: mm/vmstat.c: proper externs This patch adds proper extern declarations for five variables in include/linux/vmstat.h Signed-off-by: Adrian Bunk Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/proc_misc.c | 4 ---- include/linux/vmstat.h | 6 ++++++ kernel/sysctl.c | 2 +- mm/vmstat.c | 1 + 4 files changed, 8 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/proc/proc_misc.c b/fs/proc/proc_misc.c index c652d469dc0..b14f43d25e9 100644 --- a/fs/proc/proc_misc.c +++ b/fs/proc/proc_misc.c @@ -232,7 +232,6 @@ static int meminfo_read_proc(char *page, char **start, off_t off, #undef K } -extern const struct seq_operations fragmentation_op; static int fragmentation_open(struct inode *inode, struct file *file) { (void)inode; @@ -246,7 +245,6 @@ static const struct file_operations fragmentation_file_operations = { .release = seq_release, }; -extern const struct seq_operations pagetypeinfo_op; static int pagetypeinfo_open(struct inode *inode, struct file *file) { return seq_open(file, &pagetypeinfo_op); @@ -259,7 +257,6 @@ static const struct file_operations pagetypeinfo_file_ops = { .release = seq_release, }; -extern const struct seq_operations zoneinfo_op; static int zoneinfo_open(struct inode *inode, struct file *file) { return seq_open(file, &zoneinfo_op); @@ -356,7 +353,6 @@ static const struct file_operations proc_devinfo_operations = { .release = seq_release, }; -extern const struct seq_operations vmstat_op; static int vmstat_open(struct inode *inode, struct file *file) { return seq_open(file, &vmstat_op); diff --git a/include/linux/vmstat.h b/include/linux/vmstat.h index e83b69346d2..58334d43951 100644 --- a/include/linux/vmstat.h +++ b/include/linux/vmstat.h @@ -44,6 +44,12 @@ enum vm_event_item { PGPGIN, PGPGOUT, PSWPIN, PSWPOUT, NR_VM_EVENT_ITEMS }; +extern const struct seq_operations fragmentation_op; +extern const struct seq_operations pagetypeinfo_op; +extern const struct seq_operations zoneinfo_op; +extern const struct seq_operations vmstat_op; +extern int sysctl_stat_interval; + #ifdef CONFIG_VM_EVENT_COUNTERS /* * Light weight per cpu counter implementation. diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 2a7b9d88706..1f7b3b76a16 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -43,6 +43,7 @@ #include #include #include +#include #include #include #include @@ -80,7 +81,6 @@ extern int sysctl_drop_caches; extern int percpu_pagelist_fraction; extern int compat_log; extern int maps_protect; -extern int sysctl_stat_interval; extern int latencytop_enabled; extern int sysctl_nr_open_min, sysctl_nr_open_max; #ifdef CONFIG_RCU_TORTURE_TEST diff --git a/mm/vmstat.c b/mm/vmstat.c index c3d4a781802..b0d08e667ec 100644 --- a/mm/vmstat.c +++ b/mm/vmstat.c @@ -13,6 +13,7 @@ #include #include #include +#include #include #ifdef CONFIG_VM_EVENT_COUNTERS -- cgit v1.2.3-70-g09d2 From 42b7772812d15b86543a23b82bd6070eef9a08b1 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Wed, 23 Jul 2008 21:27:10 -0700 Subject: mm: remove double indirection on tlb parameter to free_pgd_range() & Co The double indirection here is not needed anywhere and hence (at least) confusing. Signed-off-by: Jan Beulich Cc: Hugh Dickins Cc: Nick Piggin Cc: Christoph Lameter Cc: Benjamin Herrenschmidt Cc: Paul Mackerras Cc: "Luck, Tony" Cc: Paul Mundt Cc: "David S. Miller" Acked-by: Jeremy Fitzhardinge Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/ia64/mm/hugetlbpage.c | 2 +- arch/powerpc/mm/hugetlbpage.c | 8 ++++---- fs/exec.c | 4 ++-- include/asm-ia64/hugetlb.h | 2 +- include/asm-powerpc/hugetlb.h | 2 +- include/asm-sh/hugetlb.h | 2 +- include/asm-sparc/hugetlb.h | 2 +- include/asm-x86/hugetlb.h | 2 +- include/linux/mm.h | 4 +--- mm/internal.h | 3 +++ mm/memory.c | 10 ++++++---- mm/mmap.c | 6 ++++-- 12 files changed, 26 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/arch/ia64/mm/hugetlbpage.c b/arch/ia64/mm/hugetlbpage.c index d3ce8f3bcaa..cd49e2860ee 100644 --- a/arch/ia64/mm/hugetlbpage.c +++ b/arch/ia64/mm/hugetlbpage.c @@ -112,7 +112,7 @@ follow_huge_pmd(struct mm_struct *mm, unsigned long address, pmd_t *pmd, int wri return NULL; } -void hugetlb_free_pgd_range(struct mmu_gather **tlb, +void hugetlb_free_pgd_range(struct mmu_gather *tlb, unsigned long addr, unsigned long end, unsigned long floor, unsigned long ceiling) { diff --git a/arch/powerpc/mm/hugetlbpage.c b/arch/powerpc/mm/hugetlbpage.c index 0d12fba31bc..1a96cc891cf 100644 --- a/arch/powerpc/mm/hugetlbpage.c +++ b/arch/powerpc/mm/hugetlbpage.c @@ -255,7 +255,7 @@ static void hugetlb_free_pud_range(struct mmu_gather *tlb, pgd_t *pgd, * * Must be called with pagetable lock held. */ -void hugetlb_free_pgd_range(struct mmu_gather **tlb, +void hugetlb_free_pgd_range(struct mmu_gather *tlb, unsigned long addr, unsigned long end, unsigned long floor, unsigned long ceiling) { @@ -315,13 +315,13 @@ void hugetlb_free_pgd_range(struct mmu_gather **tlb, return; start = addr; - pgd = pgd_offset((*tlb)->mm, addr); + pgd = pgd_offset(tlb->mm, addr); do { - BUG_ON(get_slice_psize((*tlb)->mm, addr) != mmu_huge_psize); + BUG_ON(get_slice_psize(tlb->mm, addr) != mmu_huge_psize); next = pgd_addr_end(addr, end); if (pgd_none_or_clear_bad(pgd)) continue; - hugetlb_free_pud_range(*tlb, pgd, addr, next, floor, ceiling); + hugetlb_free_pud_range(tlb, pgd, addr, next, floor, ceiling); } while (pgd++, addr = next, addr != end); } diff --git a/fs/exec.c b/fs/exec.c index fd9234379e8..190ed1f9277 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -541,7 +541,7 @@ static int shift_arg_pages(struct vm_area_struct *vma, unsigned long shift) /* * when the old and new regions overlap clear from new_end. */ - free_pgd_range(&tlb, new_end, old_end, new_end, + free_pgd_range(tlb, new_end, old_end, new_end, vma->vm_next ? vma->vm_next->vm_start : 0); } else { /* @@ -550,7 +550,7 @@ static int shift_arg_pages(struct vm_area_struct *vma, unsigned long shift) * have constraints on va-space that make this illegal (IA64) - * for the others its just a little faster. */ - free_pgd_range(&tlb, old_start, old_end, new_end, + free_pgd_range(tlb, old_start, old_end, new_end, vma->vm_next ? vma->vm_next->vm_start : 0); } tlb_finish_mmu(tlb, new_end, old_end); diff --git a/include/asm-ia64/hugetlb.h b/include/asm-ia64/hugetlb.h index f28a9701f1c..e9d1e5e2382 100644 --- a/include/asm-ia64/hugetlb.h +++ b/include/asm-ia64/hugetlb.h @@ -4,7 +4,7 @@ #include -void hugetlb_free_pgd_range(struct mmu_gather **tlb, unsigned long addr, +void hugetlb_free_pgd_range(struct mmu_gather *tlb, unsigned long addr, unsigned long end, unsigned long floor, unsigned long ceiling); diff --git a/include/asm-powerpc/hugetlb.h b/include/asm-powerpc/hugetlb.h index be32ff02f4a..0a37aa5ecaa 100644 --- a/include/asm-powerpc/hugetlb.h +++ b/include/asm-powerpc/hugetlb.h @@ -7,7 +7,7 @@ int is_hugepage_only_range(struct mm_struct *mm, unsigned long addr, unsigned long len); -void hugetlb_free_pgd_range(struct mmu_gather **tlb, unsigned long addr, +void hugetlb_free_pgd_range(struct mmu_gather *tlb, unsigned long addr, unsigned long end, unsigned long floor, unsigned long ceiling); diff --git a/include/asm-sh/hugetlb.h b/include/asm-sh/hugetlb.h index 02402303d89..fb30018938c 100644 --- a/include/asm-sh/hugetlb.h +++ b/include/asm-sh/hugetlb.h @@ -26,7 +26,7 @@ static inline int prepare_hugepage_range(unsigned long addr, unsigned long len) static inline void hugetlb_prefault_arch_hook(struct mm_struct *mm) { } -static inline void hugetlb_free_pgd_range(struct mmu_gather **tlb, +static inline void hugetlb_free_pgd_range(struct mmu_gather *tlb, unsigned long addr, unsigned long end, unsigned long floor, unsigned long ceiling) diff --git a/include/asm-sparc/hugetlb.h b/include/asm-sparc/hugetlb.h index 412af58926a..aeb92374ca3 100644 --- a/include/asm-sparc/hugetlb.h +++ b/include/asm-sparc/hugetlb.h @@ -31,7 +31,7 @@ static inline int prepare_hugepage_range(unsigned long addr, unsigned long len) return 0; } -static inline void hugetlb_free_pgd_range(struct mmu_gather **tlb, +static inline void hugetlb_free_pgd_range(struct mmu_gather *tlb, unsigned long addr, unsigned long end, unsigned long floor, unsigned long ceiling) diff --git a/include/asm-x86/hugetlb.h b/include/asm-x86/hugetlb.h index 14171a4924f..7eed6e0883b 100644 --- a/include/asm-x86/hugetlb.h +++ b/include/asm-x86/hugetlb.h @@ -26,7 +26,7 @@ static inline int prepare_hugepage_range(unsigned long addr, unsigned long len) static inline void hugetlb_prefault_arch_hook(struct mm_struct *mm) { } -static inline void hugetlb_free_pgd_range(struct mmu_gather **tlb, +static inline void hugetlb_free_pgd_range(struct mmu_gather *tlb, unsigned long addr, unsigned long end, unsigned long floor, unsigned long ceiling) diff --git a/include/linux/mm.h b/include/linux/mm.h index 5c7f8f64f70..f8071097302 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -769,10 +769,8 @@ struct mm_walk { int walk_page_range(unsigned long addr, unsigned long end, struct mm_walk *walk); -void free_pgd_range(struct mmu_gather **tlb, unsigned long addr, +void free_pgd_range(struct mmu_gather *tlb, unsigned long addr, unsigned long end, unsigned long floor, unsigned long ceiling); -void free_pgtables(struct mmu_gather **tlb, struct vm_area_struct *start_vma, - unsigned long floor, unsigned long ceiling); int copy_page_range(struct mm_struct *dst, struct mm_struct *src, struct vm_area_struct *vma); void unmap_mapping_range(struct address_space *mapping, diff --git a/mm/internal.h b/mm/internal.h index 50807e12490..858ad01864d 100644 --- a/mm/internal.h +++ b/mm/internal.h @@ -13,6 +13,9 @@ #include +void free_pgtables(struct mmu_gather *tlb, struct vm_area_struct *start_vma, + unsigned long floor, unsigned long ceiling); + static inline void set_page_count(struct page *page, int v) { atomic_set(&page->_count, v); diff --git a/mm/memory.c b/mm/memory.c index 87350321e66..82f3f1c5cf1 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -61,6 +61,8 @@ #include #include +#include "internal.h" + #ifndef CONFIG_NEED_MULTIPLE_NODES /* use the per-pgdat data instead for discontigmem - mbligh */ unsigned long max_mapnr; @@ -211,7 +213,7 @@ static inline void free_pud_range(struct mmu_gather *tlb, pgd_t *pgd, * * Must be called with pagetable lock held. */ -void free_pgd_range(struct mmu_gather **tlb, +void free_pgd_range(struct mmu_gather *tlb, unsigned long addr, unsigned long end, unsigned long floor, unsigned long ceiling) { @@ -262,16 +264,16 @@ void free_pgd_range(struct mmu_gather **tlb, return; start = addr; - pgd = pgd_offset((*tlb)->mm, addr); + pgd = pgd_offset(tlb->mm, addr); do { next = pgd_addr_end(addr, end); if (pgd_none_or_clear_bad(pgd)) continue; - free_pud_range(*tlb, pgd, addr, next, floor, ceiling); + free_pud_range(tlb, pgd, addr, next, floor, ceiling); } while (pgd++, addr = next, addr != end); } -void free_pgtables(struct mmu_gather **tlb, struct vm_area_struct *vma, +void free_pgtables(struct mmu_gather *tlb, struct vm_area_struct *vma, unsigned long floor, unsigned long ceiling) { while (vma) { diff --git a/mm/mmap.c b/mm/mmap.c index 1d102b956fd..75e0d0673d7 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -32,6 +32,8 @@ #include #include +#include "internal.h" + #ifndef arch_mmap_check #define arch_mmap_check(addr, len, flags) (0) #endif @@ -1763,7 +1765,7 @@ static void unmap_region(struct mm_struct *mm, update_hiwater_rss(mm); unmap_vmas(&tlb, vma, start, end, &nr_accounted, NULL); vm_unacct_memory(nr_accounted); - free_pgtables(&tlb, vma, prev? prev->vm_end: FIRST_USER_ADDRESS, + free_pgtables(tlb, vma, prev? prev->vm_end: FIRST_USER_ADDRESS, next? next->vm_start: 0); tlb_finish_mmu(tlb, start, end); } @@ -2063,7 +2065,7 @@ void exit_mmap(struct mm_struct *mm) /* Use -1 here to ensure all VMAs in the mm are unmapped */ end = unmap_vmas(&tlb, vma, 0, -1, &nr_accounted, NULL); vm_unacct_memory(nr_accounted); - free_pgtables(&tlb, vma, FIRST_USER_ADDRESS, 0); + free_pgtables(tlb, vma, FIRST_USER_ADDRESS, 0); tlb_finish_mmu(tlb, 0, end); /* -- cgit v1.2.3-70-g09d2 From da3bbdd4632c0171406b2677e31494afa5bde2f8 Mon Sep 17 00:00:00 2001 From: Kentaro Makita Date: Wed, 23 Jul 2008 21:27:13 -0700 Subject: fix soft lock up at NFS mount via per-SB LRU-list of unused dentries [Summary] Split LRU-list of unused dentries to one per superblock to avoid soft lock up during NFS mounts and remounting of any filesystem. Previously I posted here: http://lkml.org/lkml/2008/3/5/590 [Descriptions] - background dentry_unused is a list of dentries which are not referenced. dentry_unused grows up when references on directories or files are released. This list can be very long if there is huge free memory. - the problem When shrink_dcache_sb() is called, it scans all dentry_unused linearly under spin_lock(), and if dentry->d_sb is differnt from given superblock, scan next dentry. This scan costs very much if there are many entries, and very ineffective if there are many superblocks. IOW, When we need to shrink unused dentries on one dentry, but scans unused dentries on all superblocks in the system. For example, we scan 500 dentries to unmount a filesystem, but scans 1,000,000 or more unused dentries on other superblocks. In our case , At mounting NFS*, shrink_dcache_sb() is called to shrink unused dentries on NFS, but scans 100,000,000 unused dentries on superblocks in the system such as local ext3 filesystems. I hear NFS mounting took 1 min on some system in use. * : NFS uses virtual filesystem in rpc layer, so NFS is affected by this problem. 100,000,000 is possible number on large systems. Per-superblock LRU of unused dentried can reduce the cost in reasonable manner. - How to fix I found this problem is solved by David Chinner's "Per-superblock unused dentry LRU lists V3"(1), so I rebase it and add some fix to reclaim with fairness, which is in Andrew Morton's comments(2). 1) http://lkml.org/lkml/2006/5/25/318 2) http://lkml.org/lkml/2006/5/25/320 Split LRU-list of unused dentries to each superblocks. Then, NFS mounting will check dentries under a superblock instead of all. But this spliting will break LRU of dentry-unused. So, I've attempted to make reclaim unused dentrins with fairness by calculate number of dentries to scan on this sb based on following way number of dentries to scan on this sb = count * (number of dentries on this sb / number of dentries in the machine) - ToDo - I have to measuring performance number and do stress tests. - When unmount occurs during prune_dcache(), scanning on same superblock, It is unable to reach next superblock because it is gone away. We restart scannig superblock from first one, it causes unfairness of reclaim unused dentries on first superblock. But I think this happens very rarely. - Test Results Result on 6GB boxes with excessive unused dentries. Without patch: $ cat /proc/sys/fs/dentry-state 10181835 10180203 45 0 0 0 # mount -t nfs 10.124.60.70:/work/kernel-src nfs real 0m1.830s user 0m0.001s sys 0m1.653s With this patch: $ cat /proc/sys/fs/dentry-state 10236610 10234751 45 0 0 0 # mount -t nfs 10.124.60.70:/work/kernel-src nfs real 0m0.106s user 0m0.002s sys 0m0.032s [akpm@linux-foundation.org: fix comments] Signed-off-by: Kentaro Makita Cc: Neil Brown Cc: Trond Myklebust Cc: David Chinner Cc: "J. Bruce Fields" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/dcache.c | 335 ++++++++++++++++++++++++++++------------------------- fs/super.c | 1 + include/linux/fs.h | 4 + 3 files changed, 185 insertions(+), 155 deletions(-) (limited to 'fs') diff --git a/fs/dcache.c b/fs/dcache.c index 6068c25b393..3818d6ab76c 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -61,7 +61,6 @@ static struct kmem_cache *dentry_cache __read_mostly; static unsigned int d_hash_mask __read_mostly; static unsigned int d_hash_shift __read_mostly; static struct hlist_head *dentry_hashtable __read_mostly; -static LIST_HEAD(dentry_unused); /* Statistics gathering. */ struct dentry_stat_t dentry_stat = { @@ -96,14 +95,6 @@ static void d_free(struct dentry *dentry) call_rcu(&dentry->d_u.d_rcu, d_callback); } -static void dentry_lru_remove(struct dentry *dentry) -{ - if (!list_empty(&dentry->d_lru)) { - list_del_init(&dentry->d_lru); - dentry_stat.nr_unused--; - } -} - /* * Release the dentry's inode, using the filesystem * d_iput() operation if defined. @@ -130,6 +121,41 @@ static void dentry_iput(struct dentry * dentry) } } +/* + * dentry_lru_(add|add_tail|del|del_init) must be called with dcache_lock held. + */ +static void dentry_lru_add(struct dentry *dentry) +{ + list_add(&dentry->d_lru, &dentry->d_sb->s_dentry_lru); + dentry->d_sb->s_nr_dentry_unused++; + dentry_stat.nr_unused++; +} + +static void dentry_lru_add_tail(struct dentry *dentry) +{ + list_add_tail(&dentry->d_lru, &dentry->d_sb->s_dentry_lru); + dentry->d_sb->s_nr_dentry_unused++; + dentry_stat.nr_unused++; +} + +static void dentry_lru_del(struct dentry *dentry) +{ + if (!list_empty(&dentry->d_lru)) { + list_del(&dentry->d_lru); + dentry->d_sb->s_nr_dentry_unused--; + dentry_stat.nr_unused--; + } +} + +static void dentry_lru_del_init(struct dentry *dentry) +{ + if (likely(!list_empty(&dentry->d_lru))) { + list_del_init(&dentry->d_lru); + dentry->d_sb->s_nr_dentry_unused--; + dentry_stat.nr_unused--; + } +} + /** * d_kill - kill dentry and return parent * @dentry: dentry to kill @@ -212,8 +238,7 @@ repeat: goto kill_it; if (list_empty(&dentry->d_lru)) { dentry->d_flags |= DCACHE_REFERENCED; - list_add(&dentry->d_lru, &dentry_unused); - dentry_stat.nr_unused++; + dentry_lru_add(dentry); } spin_unlock(&dentry->d_lock); spin_unlock(&dcache_lock); @@ -222,7 +247,8 @@ repeat: unhash_it: __d_drop(dentry); kill_it: - dentry_lru_remove(dentry); + /* if dentry was on the d_lru list delete it from there */ + dentry_lru_del(dentry); dentry = d_kill(dentry); if (dentry) goto repeat; @@ -290,7 +316,7 @@ int d_invalidate(struct dentry * dentry) static inline struct dentry * __dget_locked(struct dentry *dentry) { atomic_inc(&dentry->d_count); - dentry_lru_remove(dentry); + dentry_lru_del_init(dentry); return dentry; } @@ -406,133 +432,167 @@ static void prune_one_dentry(struct dentry * dentry) if (dentry->d_op && dentry->d_op->d_delete) dentry->d_op->d_delete(dentry); - dentry_lru_remove(dentry); + dentry_lru_del_init(dentry); __d_drop(dentry); dentry = d_kill(dentry); spin_lock(&dcache_lock); } } -/** - * prune_dcache - shrink the dcache - * @count: number of entries to try and free - * @sb: if given, ignore dentries for other superblocks - * which are being unmounted. - * - * Shrink the dcache. This is done when we need - * more memory, or simply when we need to unmount - * something (at which point we need to unuse - * all dentries). - * - * This function may fail to free any resources if - * all the dentries are in use. +/* + * Shrink the dentry LRU on a given superblock. + * @sb : superblock to shrink dentry LRU. + * @count: If count is NULL, we prune all dentries on superblock. + * @flags: If flags is non-zero, we need to do special processing based on + * which flags are set. This means we don't need to maintain multiple + * similar copies of this loop. */ - -static void prune_dcache(int count, struct super_block *sb) +static void __shrink_dcache_sb(struct super_block *sb, int *count, int flags) { - spin_lock(&dcache_lock); - for (; count ; count--) { - struct dentry *dentry; - struct list_head *tmp; - struct rw_semaphore *s_umount; - - cond_resched_lock(&dcache_lock); + LIST_HEAD(referenced); + LIST_HEAD(tmp); + struct dentry *dentry; + int cnt = 0; - tmp = dentry_unused.prev; - if (sb) { - /* Try to find a dentry for this sb, but don't try - * too hard, if they aren't near the tail they will - * be moved down again soon + BUG_ON(!sb); + BUG_ON((flags & DCACHE_REFERENCED) && count == NULL); + spin_lock(&dcache_lock); + if (count != NULL) + /* called from prune_dcache() and shrink_dcache_parent() */ + cnt = *count; +restart: + if (count == NULL) + list_splice_init(&sb->s_dentry_lru, &tmp); + else { + while (!list_empty(&sb->s_dentry_lru)) { + dentry = list_entry(sb->s_dentry_lru.prev, + struct dentry, d_lru); + BUG_ON(dentry->d_sb != sb); + + spin_lock(&dentry->d_lock); + /* + * If we are honouring the DCACHE_REFERENCED flag and + * the dentry has this flag set, don't free it. Clear + * the flag and put it back on the LRU. */ - int skip = count; - while (skip && tmp != &dentry_unused && - list_entry(tmp, struct dentry, d_lru)->d_sb != sb) { - skip--; - tmp = tmp->prev; + if ((flags & DCACHE_REFERENCED) + && (dentry->d_flags & DCACHE_REFERENCED)) { + dentry->d_flags &= ~DCACHE_REFERENCED; + list_move_tail(&dentry->d_lru, &referenced); + spin_unlock(&dentry->d_lock); + } else { + list_move_tail(&dentry->d_lru, &tmp); + spin_unlock(&dentry->d_lock); + cnt--; + if (!cnt) + break; } } - if (tmp == &dentry_unused) - break; - list_del_init(tmp); - prefetch(dentry_unused.prev); - dentry_stat.nr_unused--; - dentry = list_entry(tmp, struct dentry, d_lru); - - spin_lock(&dentry->d_lock); + } + while (!list_empty(&tmp)) { + dentry = list_entry(tmp.prev, struct dentry, d_lru); + dentry_lru_del_init(dentry); + spin_lock(&dentry->d_lock); /* * We found an inuse dentry which was not removed from - * dentry_unused because of laziness during lookup. Do not free - * it - just keep it off the dentry_unused list. + * the LRU because of laziness during lookup. Do not free + * it - just keep it off the LRU list. */ - if (atomic_read(&dentry->d_count)) { - spin_unlock(&dentry->d_lock); + if (atomic_read(&dentry->d_count)) { + spin_unlock(&dentry->d_lock); continue; } - /* If the dentry was recently referenced, don't free it. */ - if (dentry->d_flags & DCACHE_REFERENCED) { - dentry->d_flags &= ~DCACHE_REFERENCED; - list_add(&dentry->d_lru, &dentry_unused); - dentry_stat.nr_unused++; - spin_unlock(&dentry->d_lock); + prune_one_dentry(dentry); + /* dentry->d_lock was dropped in prune_one_dentry() */ + cond_resched_lock(&dcache_lock); + } + if (count == NULL && !list_empty(&sb->s_dentry_lru)) + goto restart; + if (count != NULL) + *count = cnt; + if (!list_empty(&referenced)) + list_splice(&referenced, &sb->s_dentry_lru); + spin_unlock(&dcache_lock); +} + +/** + * prune_dcache - shrink the dcache + * @count: number of entries to try to free + * + * Shrink the dcache. This is done when we need more memory, or simply when we + * need to unmount something (at which point we need to unuse all dentries). + * + * This function may fail to free any resources if all the dentries are in use. + */ +static void prune_dcache(int count) +{ + struct super_block *sb; + int w_count; + int unused = dentry_stat.nr_unused; + int prune_ratio; + int pruned; + + if (unused == 0 || count == 0) + return; + spin_lock(&dcache_lock); +restart: + if (count >= unused) + prune_ratio = 1; + else + prune_ratio = unused / count; + spin_lock(&sb_lock); + list_for_each_entry(sb, &super_blocks, s_list) { + if (sb->s_nr_dentry_unused == 0) continue; - } - /* - * If the dentry is not DCACHED_REFERENCED, it is time - * to remove it from the dcache, provided the super block is - * NULL (which means we are trying to reclaim memory) - * or this dentry belongs to the same super block that - * we want to shrink. - */ - /* - * If this dentry is for "my" filesystem, then I can prune it - * without taking the s_umount lock (I already hold it). + sb->s_count++; + /* Now, we reclaim unused dentrins with fairness. + * We reclaim them same percentage from each superblock. + * We calculate number of dentries to scan on this sb + * as follows, but the implementation is arranged to avoid + * overflows: + * number of dentries to scan on this sb = + * count * (number of dentries on this sb / + * number of dentries in the machine) */ - if (sb && dentry->d_sb == sb) { - prune_one_dentry(dentry); - continue; - } + spin_unlock(&sb_lock); + if (prune_ratio != 1) + w_count = (sb->s_nr_dentry_unused / prune_ratio) + 1; + else + w_count = sb->s_nr_dentry_unused; + pruned = w_count; /* - * ...otherwise we need to be sure this filesystem isn't being - * unmounted, otherwise we could race with - * generic_shutdown_super(), and end up holding a reference to - * an inode while the filesystem is unmounted. - * So we try to get s_umount, and make sure s_root isn't NULL. - * (Take a local copy of s_umount to avoid a use-after-free of - * `dentry'). + * We need to be sure this filesystem isn't being unmounted, + * otherwise we could race with generic_shutdown_super(), and + * end up holding a reference to an inode while the filesystem + * is unmounted. So we try to get s_umount, and make sure + * s_root isn't NULL. */ - s_umount = &dentry->d_sb->s_umount; - if (down_read_trylock(s_umount)) { - if (dentry->d_sb->s_root != NULL) { - prune_one_dentry(dentry); - up_read(s_umount); - continue; + if (down_read_trylock(&sb->s_umount)) { + if ((sb->s_root != NULL) && + (!list_empty(&sb->s_dentry_lru))) { + spin_unlock(&dcache_lock); + __shrink_dcache_sb(sb, &w_count, + DCACHE_REFERENCED); + pruned -= w_count; + spin_lock(&dcache_lock); } - up_read(s_umount); + up_read(&sb->s_umount); } - spin_unlock(&dentry->d_lock); + spin_lock(&sb_lock); + count -= pruned; /* - * Insert dentry at the head of the list as inserting at the - * tail leads to a cycle. + * restart only when sb is no longer on the list and + * we have more work to do. */ - list_add(&dentry->d_lru, &dentry_unused); - dentry_stat.nr_unused++; + if (__put_super_and_need_restart(sb) && count > 0) { + spin_unlock(&sb_lock); + goto restart; + } } + spin_unlock(&sb_lock); spin_unlock(&dcache_lock); } -/* - * Shrink the dcache for the specified super block. - * This allows us to unmount a device without disturbing - * the dcache for the other devices. - * - * This implementation makes just two traversals of the - * unused list. On the first pass we move the selected - * dentries to the most recent end, and on the second - * pass we free them. The second pass must restart after - * each dput(), but since the target dentries are all at - * the end, it's really just a single traversal. - */ - /** * shrink_dcache_sb - shrink dcache for a superblock * @sb: superblock @@ -541,44 +601,9 @@ static void prune_dcache(int count, struct super_block *sb) * is used to free the dcache before unmounting a file * system */ - void shrink_dcache_sb(struct super_block * sb) { - struct list_head *tmp, *next; - struct dentry *dentry; - - /* - * Pass one ... move the dentries for the specified - * superblock to the most recent end of the unused list. - */ - spin_lock(&dcache_lock); - list_for_each_prev_safe(tmp, next, &dentry_unused) { - dentry = list_entry(tmp, struct dentry, d_lru); - if (dentry->d_sb != sb) - continue; - list_move_tail(tmp, &dentry_unused); - } - - /* - * Pass two ... free the dentries for this superblock. - */ -repeat: - list_for_each_prev_safe(tmp, next, &dentry_unused) { - dentry = list_entry(tmp, struct dentry, d_lru); - if (dentry->d_sb != sb) - continue; - dentry_stat.nr_unused--; - list_del_init(tmp); - spin_lock(&dentry->d_lock); - if (atomic_read(&dentry->d_count)) { - spin_unlock(&dentry->d_lock); - continue; - } - prune_one_dentry(dentry); - cond_resched_lock(&dcache_lock); - goto repeat; - } - spin_unlock(&dcache_lock); + __shrink_dcache_sb(sb, NULL, 0); } /* @@ -595,7 +620,7 @@ static void shrink_dcache_for_umount_subtree(struct dentry *dentry) /* detach this root from the system */ spin_lock(&dcache_lock); - dentry_lru_remove(dentry); + dentry_lru_del_init(dentry); __d_drop(dentry); spin_unlock(&dcache_lock); @@ -609,7 +634,7 @@ static void shrink_dcache_for_umount_subtree(struct dentry *dentry) spin_lock(&dcache_lock); list_for_each_entry(loop, &dentry->d_subdirs, d_u.d_child) { - dentry_lru_remove(loop); + dentry_lru_del_init(loop); __d_drop(loop); cond_resched_lock(&dcache_lock); } @@ -791,14 +816,13 @@ resume: struct dentry *dentry = list_entry(tmp, struct dentry, d_u.d_child); next = tmp->next; - dentry_lru_remove(dentry); + dentry_lru_del_init(dentry); /* * move only zero ref count dentries to the end * of the unused list for prune_dcache */ if (!atomic_read(&dentry->d_count)) { - list_add_tail(&dentry->d_lru, &dentry_unused); - dentry_stat.nr_unused++; + dentry_lru_add_tail(dentry); found++; } @@ -840,10 +864,11 @@ out: void shrink_dcache_parent(struct dentry * parent) { + struct super_block *sb = parent->d_sb; int found; while ((found = select_parent(parent)) != 0) - prune_dcache(found, parent->d_sb); + __shrink_dcache_sb(sb, &found, 0); } /* @@ -863,7 +888,7 @@ static int shrink_dcache_memory(int nr, gfp_t gfp_mask) if (nr) { if (!(gfp_mask & __GFP_FS)) return -1; - prune_dcache(nr, NULL); + prune_dcache(nr); } return (dentry_stat.nr_unused / 100) * sysctl_vfs_cache_pressure; } @@ -1215,7 +1240,7 @@ struct dentry *d_splice_alias(struct inode *inode, struct dentry *dentry) * rcu_read_lock() and rcu_read_unlock() are used to disable preemption while * lookup is going on. * - * dentry_unused list is not updated even if lookup finds the required dentry + * The dentry unused LRU is not updated even if lookup finds the required dentry * in there. It is updated in places such as prune_dcache, shrink_dcache_sb, * select_parent and __dget_locked. This laziness saves lookup from dcache_lock * acquisition. diff --git a/fs/super.c b/fs/super.c index 453877c5697..e931ae9511f 100644 --- a/fs/super.c +++ b/fs/super.c @@ -70,6 +70,7 @@ static struct super_block *alloc_super(struct file_system_type *type) INIT_LIST_HEAD(&s->s_instances); INIT_HLIST_HEAD(&s->s_anon); INIT_LIST_HEAD(&s->s_inodes); + INIT_LIST_HEAD(&s->s_dentry_lru); init_rwsem(&s->s_umount); mutex_init(&s->s_lock); lockdep_set_class(&s->s_umount, &type->s_umount_key); diff --git a/include/linux/fs.h b/include/linux/fs.h index ff54ae4933f..e5e6a244096 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1025,6 +1025,7 @@ extern int send_sigurg(struct fown_struct *fown); extern struct list_head super_blocks; extern spinlock_t sb_lock; +#define sb_entry(list) list_entry((list), struct super_block, s_list) #define S_BIAS (1<<30) struct super_block { struct list_head s_list; /* Keep this first */ @@ -1058,6 +1059,9 @@ struct super_block { struct list_head s_more_io; /* parked for more writeback */ struct hlist_head s_anon; /* anonymous dentries for (nfs) exporting */ struct list_head s_files; + /* s_dentry_lru and s_nr_dentry_unused are protected by dcache_lock */ + struct list_head s_dentry_lru; /* unused dentry lru */ + int s_nr_dentry_unused; /* # of dentry on lru */ struct block_device *s_bdev; struct mtd_info *s_mtd; -- cgit v1.2.3-70-g09d2 From a1e78772d72b2616ed20e54896e68e0e7044854e Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Wed, 23 Jul 2008 21:27:23 -0700 Subject: hugetlb: reserve huge pages for reliable MAP_PRIVATE hugetlbfs mappings until fork() This patch reserves huge pages at mmap() time for MAP_PRIVATE mappings in a similar manner to the reservations taken for MAP_SHARED mappings. The reserve count is accounted both globally and on a per-VMA basis for private mappings. This guarantees that a process that successfully calls mmap() will successfully fault all pages in the future unless fork() is called. The characteristics of private mappings of hugetlbfs files behaviour after this patch are; 1. The process calling mmap() is guaranteed to succeed all future faults until it forks(). 2. On fork(), the parent may die due to SIGKILL on writes to the private mapping if enough pages are not available for the COW. For reasonably reliable behaviour in the face of a small huge page pool, children of hugepage-aware processes should not reference the mappings; such as might occur when fork()ing to exec(). 3. On fork(), the child VMAs inherit no reserves. Reads on pages already faulted by the parent will succeed. Successful writes will depend on enough huge pages being free in the pool. 4. Quotas of the hugetlbfs mount are checked at reserve time for the mapper and at fault time otherwise. Before this patch, all reads or writes in the child potentially needs page allocations that can later lead to the death of the parent. This applies to reads and writes of uninstantiated pages as well as COW. After the patch it is only a write to an instantiated page that causes problems. Signed-off-by: Mel Gorman Acked-by: Adam Litke Cc: Andy Whitcroft Cc: William Lee Irwin III Cc: Hugh Dickins Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/hugetlbfs/inode.c | 8 +-- include/linux/hugetlb.h | 9 ++- kernel/fork.c | 9 +++ mm/hugetlb.c | 158 ++++++++++++++++++++++++++++++++++++------------ 4 files changed, 140 insertions(+), 44 deletions(-) (limited to 'fs') diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c index aeabf80f81a..1576bbecd08 100644 --- a/fs/hugetlbfs/inode.c +++ b/fs/hugetlbfs/inode.c @@ -103,9 +103,9 @@ static int hugetlbfs_file_mmap(struct file *file, struct vm_area_struct *vma) ret = -ENOMEM; len = vma_len + ((loff_t)vma->vm_pgoff << PAGE_SHIFT); - if (vma->vm_flags & VM_MAYSHARE && - hugetlb_reserve_pages(inode, vma->vm_pgoff >> (HPAGE_SHIFT-PAGE_SHIFT), - len >> HPAGE_SHIFT)) + if (hugetlb_reserve_pages(inode, + vma->vm_pgoff >> (HPAGE_SHIFT-PAGE_SHIFT), + len >> HPAGE_SHIFT, vma)) goto out; ret = 0; @@ -942,7 +942,7 @@ struct file *hugetlb_file_setup(const char *name, size_t size) goto out_dentry; error = -ENOMEM; - if (hugetlb_reserve_pages(inode, 0, size >> HPAGE_SHIFT)) + if (hugetlb_reserve_pages(inode, 0, size >> HPAGE_SHIFT, NULL)) goto out_inode; d_instantiate(dentry, inode); diff --git a/include/linux/hugetlb.h b/include/linux/hugetlb.h index a79e80b689d..185b14c9f02 100644 --- a/include/linux/hugetlb.h +++ b/include/linux/hugetlb.h @@ -17,6 +17,7 @@ static inline int is_vm_hugetlb_page(struct vm_area_struct *vma) return vma->vm_flags & VM_HUGETLB; } +void reset_vma_resv_huge_pages(struct vm_area_struct *vma); int hugetlb_sysctl_handler(struct ctl_table *, int, struct file *, void __user *, size_t *, loff_t *); int hugetlb_overcommit_handler(struct ctl_table *, int, struct file *, void __user *, size_t *, loff_t *); int hugetlb_treat_movable_handler(struct ctl_table *, int, struct file *, void __user *, size_t *, loff_t *); @@ -30,7 +31,8 @@ int hugetlb_report_node_meminfo(int, char *); unsigned long hugetlb_total_pages(void); int hugetlb_fault(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, int write_access); -int hugetlb_reserve_pages(struct inode *inode, long from, long to); +int hugetlb_reserve_pages(struct inode *inode, long from, long to, + struct vm_area_struct *vma); void hugetlb_unreserve_pages(struct inode *inode, long offset, long freed); extern unsigned long max_huge_pages; @@ -58,6 +60,11 @@ static inline int is_vm_hugetlb_page(struct vm_area_struct *vma) { return 0; } + +static inline void reset_vma_resv_huge_pages(struct vm_area_struct *vma) +{ +} + static inline unsigned long hugetlb_total_pages(void) { return 0; diff --git a/kernel/fork.c b/kernel/fork.c index adefc1131f2..552c8d8e77a 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -33,6 +33,7 @@ #include #include #include +#include #include #include #include @@ -306,6 +307,14 @@ static int dup_mmap(struct mm_struct *mm, struct mm_struct *oldmm) spin_unlock(&file->f_mapping->i_mmap_lock); } + /* + * Clear hugetlb-related page reserves for children. This only + * affects MAP_PRIVATE mappings. Faults generated by the child + * are not guaranteed to succeed, even if read-only + */ + if (is_vm_hugetlb_page(tmp)) + reset_vma_resv_huge_pages(tmp); + /* * Link in the new vma and copy the page table entries. */ diff --git a/mm/hugetlb.c b/mm/hugetlb.c index a4dbba8965f..0af500db363 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -40,6 +40,69 @@ static int hugetlb_next_nid; */ static DEFINE_SPINLOCK(hugetlb_lock); +/* + * These helpers are used to track how many pages are reserved for + * faults in a MAP_PRIVATE mapping. Only the process that called mmap() + * is guaranteed to have their future faults succeed. + * + * With the exception of reset_vma_resv_huge_pages() which is called at fork(), + * the reserve counters are updated with the hugetlb_lock held. It is safe + * to reset the VMA at fork() time as it is not in use yet and there is no + * chance of the global counters getting corrupted as a result of the values. + */ +static unsigned long vma_resv_huge_pages(struct vm_area_struct *vma) +{ + VM_BUG_ON(!is_vm_hugetlb_page(vma)); + if (!(vma->vm_flags & VM_SHARED)) + return (unsigned long)vma->vm_private_data; + return 0; +} + +static void set_vma_resv_huge_pages(struct vm_area_struct *vma, + unsigned long reserve) +{ + VM_BUG_ON(!is_vm_hugetlb_page(vma)); + VM_BUG_ON(vma->vm_flags & VM_SHARED); + + vma->vm_private_data = (void *)reserve; +} + +/* Decrement the reserved pages in the hugepage pool by one */ +static void decrement_hugepage_resv_vma(struct vm_area_struct *vma) +{ + if (vma->vm_flags & VM_SHARED) { + /* Shared mappings always use reserves */ + resv_huge_pages--; + } else { + /* + * Only the process that called mmap() has reserves for + * private mappings. + */ + if (vma_resv_huge_pages(vma)) { + resv_huge_pages--; + reserve = (unsigned long)vma->vm_private_data - 1; + vma->vm_private_data = (void *)reserve; + } + } +} + +void reset_vma_resv_huge_pages(struct vm_area_struct *vma) +{ + VM_BUG_ON(!is_vm_hugetlb_page(vma)); + if (!(vma->vm_flags & VM_SHARED)) + vma->vm_private_data = (void *)0; +} + +/* Returns true if the VMA has associated reserve pages */ +static int vma_has_private_reserves(struct vm_area_struct *vma) +{ + if (vma->vm_flags & VM_SHARED) + return 0; + if (!vma_resv_huge_pages(vma)) + return 0; + return 1; +} + static void clear_huge_page(struct page *page, unsigned long addr) { int i; @@ -101,6 +164,15 @@ static struct page *dequeue_huge_page_vma(struct vm_area_struct *vma, struct zone *zone; struct zoneref *z; + /* + * A child process with MAP_PRIVATE mappings created by their parent + * have no page reserves. This check ensures that reservations are + * not "stolen". The child may still get SIGKILLed + */ + if (!vma_has_private_reserves(vma) && + free_huge_pages - resv_huge_pages == 0) + return NULL; + for_each_zone_zonelist_nodemask(zone, z, zonelist, MAX_NR_ZONES - 1, nodemask) { nid = zone_to_nid(zone); @@ -111,8 +183,8 @@ static struct page *dequeue_huge_page_vma(struct vm_area_struct *vma, list_del(&page->lru); free_huge_pages--; free_huge_pages_node[nid]--; - if (vma && vma->vm_flags & VM_MAYSHARE) - resv_huge_pages--; + decrement_hugepage_resv_vma(vma); + break; } } @@ -461,55 +533,40 @@ static void return_unused_surplus_pages(unsigned long unused_resv_pages) } } - -static struct page *alloc_huge_page_shared(struct vm_area_struct *vma, - unsigned long addr) +static struct page *alloc_huge_page(struct vm_area_struct *vma, + unsigned long addr) { struct page *page; + struct address_space *mapping = vma->vm_file->f_mapping; + struct inode *inode = mapping->host; + unsigned int chg = 0; + + /* + * Processes that did not create the mapping will have no reserves and + * will not have accounted against quota. Check that the quota can be + * made before satisfying the allocation + */ + if (!vma_has_private_reserves(vma)) { + chg = 1; + if (hugetlb_get_quota(inode->i_mapping, chg)) + return ERR_PTR(-ENOSPC); + } spin_lock(&hugetlb_lock); page = dequeue_huge_page_vma(vma, addr); spin_unlock(&hugetlb_lock); - return page ? page : ERR_PTR(-VM_FAULT_OOM); -} -static struct page *alloc_huge_page_private(struct vm_area_struct *vma, - unsigned long addr) -{ - struct page *page = NULL; - - if (hugetlb_get_quota(vma->vm_file->f_mapping, 1)) - return ERR_PTR(-VM_FAULT_SIGBUS); - - spin_lock(&hugetlb_lock); - if (free_huge_pages > resv_huge_pages) - page = dequeue_huge_page_vma(vma, addr); - spin_unlock(&hugetlb_lock); if (!page) { page = alloc_buddy_huge_page(vma, addr); if (!page) { - hugetlb_put_quota(vma->vm_file->f_mapping, 1); + hugetlb_put_quota(inode->i_mapping, chg); return ERR_PTR(-VM_FAULT_OOM); } } - return page; -} -static struct page *alloc_huge_page(struct vm_area_struct *vma, - unsigned long addr) -{ - struct page *page; - struct address_space *mapping = vma->vm_file->f_mapping; - - if (vma->vm_flags & VM_MAYSHARE) - page = alloc_huge_page_shared(vma, addr); - else - page = alloc_huge_page_private(vma, addr); + set_page_refcounted(page); + set_page_private(page, (unsigned long) mapping); - if (!IS_ERR(page)) { - set_page_refcounted(page); - set_page_private(page, (unsigned long) mapping); - } return page; } @@ -757,6 +814,13 @@ out: return ret; } +static void hugetlb_vm_op_close(struct vm_area_struct *vma) +{ + unsigned long reserve = vma_resv_huge_pages(vma); + if (reserve) + hugetlb_acct_memory(-reserve); +} + /* * We cannot handle pagefaults against hugetlb pages at all. They cause * handle_mm_fault() to try to instantiate regular-sized pages in the @@ -771,6 +835,7 @@ static int hugetlb_vm_op_fault(struct vm_area_struct *vma, struct vm_fault *vmf) struct vm_operations_struct hugetlb_vm_ops = { .fault = hugetlb_vm_op_fault, + .close = hugetlb_vm_op_close, }; static pte_t make_huge_pte(struct vm_area_struct *vma, struct page *page, @@ -1289,11 +1354,25 @@ static long region_truncate(struct list_head *head, long end) return chg; } -int hugetlb_reserve_pages(struct inode *inode, long from, long to) +int hugetlb_reserve_pages(struct inode *inode, + long from, long to, + struct vm_area_struct *vma) { long ret, chg; - chg = region_chg(&inode->i_mapping->private_list, from, to); + /* + * Shared mappings base their reservation on the number of pages that + * are already allocated on behalf of the file. Private mappings need + * to reserve the full area even if read-only as mprotect() may be + * called to make the mapping read-write. Assume !vma is a shm mapping + */ + if (!vma || vma->vm_flags & VM_SHARED) + chg = region_chg(&inode->i_mapping->private_list, from, to); + else { + chg = to - from; + set_vma_resv_huge_pages(vma, chg); + } + if (chg < 0) return chg; @@ -1304,7 +1383,8 @@ int hugetlb_reserve_pages(struct inode *inode, long from, long to) hugetlb_put_quota(inode->i_mapping, chg); return ret; } - region_add(&inode->i_mapping->private_list, from, to); + if (!vma || vma->vm_flags & VM_SHARED) + region_add(&inode->i_mapping->private_list, from, to); return 0; } -- cgit v1.2.3-70-g09d2 From 04f2cbe35699d22dbf428373682ead85ca1240f5 Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Wed, 23 Jul 2008 21:27:25 -0700 Subject: hugetlb: guarantee that COW faults for a process that called mmap(MAP_PRIVATE) on hugetlbfs will succeed After patch 2 in this series, a process that successfully calls mmap() for a MAP_PRIVATE mapping will be guaranteed to successfully fault until a process calls fork(). At that point, the next write fault from the parent could fail due to COW if the child still has a reference. We only reserve pages for the parent but a copy must be made to avoid leaking data from the parent to the child after fork(). Reserves could be taken for both parent and child at fork time to guarantee faults but if the mapping is large it is highly likely we will not have sufficient pages for the reservation, and it is common to fork only to exec() immediatly after. A failure here would be very undesirable. Note that the current behaviour of mainline with MAP_PRIVATE pages is pretty bad. The following situation is allowed to occur today. 1. Process calls mmap(MAP_PRIVATE) 2. Process calls mlock() to fault all pages and makes sure it succeeds 3. Process forks() 4. Process writes to MAP_PRIVATE mapping while child still exists 5. If the COW fails at this point, the process gets SIGKILLed even though it had taken care to ensure the pages existed This patch improves the situation by guaranteeing the reliability of the process that successfully calls mmap(). When the parent performs COW, it will try to satisfy the allocation without using reserves. If that fails the parent will steal the page leaving any children without a page. Faults from the child after that point will result in failure. If the child COW happens first, an attempt will be made to allocate the page without reserves and the child will get SIGKILLed on failure. To summarise the new behaviour: 1. If the original mapper performs COW on a private mapping with multiple references, it will attempt to allocate a hugepage from the pool or the buddy allocator without using the existing reserves. On fail, VMAs mapping the same area are traversed and the page being COW'd is unmapped where found. It will then steal the original page as the last mapper in the normal way. 2. The VMAs the pages were unmapped from are flagged to note that pages with data no longer exist. Future no-page faults on those VMAs will terminate the process as otherwise it would appear that data was corrupted. A warning is printed to the console that this situation occured. 2. If the child performs COW first, it will attempt to satisfy the COW from the pool if there are enough pages or via the buddy allocator if overcommit is allowed and the buddy allocator can satisfy the request. If it fails, the child will be killed. If the pool is large enough, existing applications will not notice that the reserves were a factor. Existing applications depending on the no-reserves been set are unlikely to exist as for much of the history of hugetlbfs, pages were prefaulted at mmap(), allocating the pages at that point or failing the mmap(). [npiggin@suse.de: fix CONFIG_HUGETLB=n build] Signed-off-by: Mel Gorman Acked-by: Adam Litke Cc: Andy Whitcroft Cc: William Lee Irwin III Cc: Hugh Dickins Cc: Nick Piggin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/hugetlbfs/inode.c | 2 +- include/linux/hugetlb.h | 8 +- mm/hugetlb.c | 201 +++++++++++++++++++++++++++++++++++++++++++----- mm/memory.c | 2 +- 4 files changed, 190 insertions(+), 23 deletions(-) (limited to 'fs') diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c index 1576bbecd08..428eff5b73f 100644 --- a/fs/hugetlbfs/inode.c +++ b/fs/hugetlbfs/inode.c @@ -441,7 +441,7 @@ hugetlb_vmtruncate_list(struct prio_tree_root *root, pgoff_t pgoff) v_offset = 0; __unmap_hugepage_range(vma, - vma->vm_start + v_offset, vma->vm_end); + vma->vm_start + v_offset, vma->vm_end, NULL); } } diff --git a/include/linux/hugetlb.h b/include/linux/hugetlb.h index 185b14c9f02..abbc187193a 100644 --- a/include/linux/hugetlb.h +++ b/include/linux/hugetlb.h @@ -23,8 +23,10 @@ int hugetlb_overcommit_handler(struct ctl_table *, int, struct file *, void __us int hugetlb_treat_movable_handler(struct ctl_table *, int, struct file *, void __user *, size_t *, loff_t *); int copy_hugetlb_page_range(struct mm_struct *, struct mm_struct *, struct vm_area_struct *); int follow_hugetlb_page(struct mm_struct *, struct vm_area_struct *, struct page **, struct vm_area_struct **, unsigned long *, int *, int, int); -void unmap_hugepage_range(struct vm_area_struct *, unsigned long, unsigned long); -void __unmap_hugepage_range(struct vm_area_struct *, unsigned long, unsigned long); +void unmap_hugepage_range(struct vm_area_struct *, + unsigned long, unsigned long, struct page *); +void __unmap_hugepage_range(struct vm_area_struct *, + unsigned long, unsigned long, struct page *); int hugetlb_prefault(struct address_space *, struct vm_area_struct *); int hugetlb_report_meminfo(char *); int hugetlb_report_node_meminfo(int, char *); @@ -74,7 +76,7 @@ static inline unsigned long hugetlb_total_pages(void) #define follow_huge_addr(mm, addr, write) ERR_PTR(-EINVAL) #define copy_hugetlb_page_range(src, dst, vma) ({ BUG(); 0; }) #define hugetlb_prefault(mapping, vma) ({ BUG(); 0; }) -#define unmap_hugepage_range(vma, start, end) BUG() +#define unmap_hugepage_range(vma, start, end, page) BUG() #define hugetlb_report_meminfo(buf) 0 #define hugetlb_report_node_meminfo(n, buf) 0 #define follow_huge_pmd(mm, addr, pmd, write) NULL diff --git a/mm/hugetlb.c b/mm/hugetlb.c index 0af500db363..a2d29b84501 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -40,6 +40,9 @@ static int hugetlb_next_nid; */ static DEFINE_SPINLOCK(hugetlb_lock); +#define HPAGE_RESV_OWNER (1UL << (BITS_PER_LONG - 1)) +#define HPAGE_RESV_UNMAPPED (1UL << (BITS_PER_LONG - 2)) +#define HPAGE_RESV_MASK (HPAGE_RESV_OWNER | HPAGE_RESV_UNMAPPED) /* * These helpers are used to track how many pages are reserved for * faults in a MAP_PRIVATE mapping. Only the process that called mmap() @@ -54,17 +57,32 @@ static unsigned long vma_resv_huge_pages(struct vm_area_struct *vma) { VM_BUG_ON(!is_vm_hugetlb_page(vma)); if (!(vma->vm_flags & VM_SHARED)) - return (unsigned long)vma->vm_private_data; + return (unsigned long)vma->vm_private_data & ~HPAGE_RESV_MASK; return 0; } static void set_vma_resv_huge_pages(struct vm_area_struct *vma, unsigned long reserve) { + unsigned long flags; VM_BUG_ON(!is_vm_hugetlb_page(vma)); VM_BUG_ON(vma->vm_flags & VM_SHARED); - vma->vm_private_data = (void *)reserve; + flags = (unsigned long)vma->vm_private_data & HPAGE_RESV_MASK; + vma->vm_private_data = (void *)(reserve | flags); +} + +static void set_vma_resv_flags(struct vm_area_struct *vma, unsigned long flags) +{ + unsigned long reserveflags = (unsigned long)vma->vm_private_data; + VM_BUG_ON(!is_vm_hugetlb_page(vma)); + vma->vm_private_data = (void *)(reserveflags | flags); +} + +static int is_vma_resv_set(struct vm_area_struct *vma, unsigned long flag) +{ + VM_BUG_ON(!is_vm_hugetlb_page(vma)); + return ((unsigned long)vma->vm_private_data & flag) != 0; } /* Decrement the reserved pages in the hugepage pool by one */ @@ -78,14 +96,18 @@ static void decrement_hugepage_resv_vma(struct vm_area_struct *vma) * Only the process that called mmap() has reserves for * private mappings. */ - if (vma_resv_huge_pages(vma)) { + if (is_vma_resv_set(vma, HPAGE_RESV_OWNER)) { + unsigned long flags, reserve; resv_huge_pages--; + flags = (unsigned long)vma->vm_private_data & + HPAGE_RESV_MASK; reserve = (unsigned long)vma->vm_private_data - 1; - vma->vm_private_data = (void *)reserve; + vma->vm_private_data = (void *)(reserve | flags); } } } +/* Reset counters to 0 and clear all HPAGE_RESV_* flags */ void reset_vma_resv_huge_pages(struct vm_area_struct *vma) { VM_BUG_ON(!is_vm_hugetlb_page(vma)); @@ -153,7 +175,7 @@ static struct page *dequeue_huge_page(void) } static struct page *dequeue_huge_page_vma(struct vm_area_struct *vma, - unsigned long address) + unsigned long address, int avoid_reserve) { int nid; struct page *page = NULL; @@ -173,6 +195,10 @@ static struct page *dequeue_huge_page_vma(struct vm_area_struct *vma, free_huge_pages - resv_huge_pages == 0) return NULL; + /* If reserves cannot be used, ensure enough pages are in the pool */ + if (avoid_reserve && free_huge_pages - resv_huge_pages == 0) + return NULL; + for_each_zone_zonelist_nodemask(zone, z, zonelist, MAX_NR_ZONES - 1, nodemask) { nid = zone_to_nid(zone); @@ -183,7 +209,9 @@ static struct page *dequeue_huge_page_vma(struct vm_area_struct *vma, list_del(&page->lru); free_huge_pages--; free_huge_pages_node[nid]--; - decrement_hugepage_resv_vma(vma); + + if (!avoid_reserve) + decrement_hugepage_resv_vma(vma); break; } @@ -534,7 +562,7 @@ static void return_unused_surplus_pages(unsigned long unused_resv_pages) } static struct page *alloc_huge_page(struct vm_area_struct *vma, - unsigned long addr) + unsigned long addr, int avoid_reserve) { struct page *page; struct address_space *mapping = vma->vm_file->f_mapping; @@ -546,14 +574,15 @@ static struct page *alloc_huge_page(struct vm_area_struct *vma, * will not have accounted against quota. Check that the quota can be * made before satisfying the allocation */ - if (!vma_has_private_reserves(vma)) { + if (!(vma->vm_flags & VM_SHARED) && + !is_vma_resv_set(vma, HPAGE_RESV_OWNER)) { chg = 1; if (hugetlb_get_quota(inode->i_mapping, chg)) return ERR_PTR(-ENOSPC); } spin_lock(&hugetlb_lock); - page = dequeue_huge_page_vma(vma, addr); + page = dequeue_huge_page_vma(vma, addr, avoid_reserve); spin_unlock(&hugetlb_lock); if (!page) { @@ -909,7 +938,7 @@ nomem: } void __unmap_hugepage_range(struct vm_area_struct *vma, unsigned long start, - unsigned long end) + unsigned long end, struct page *ref_page) { struct mm_struct *mm = vma->vm_mm; unsigned long address; @@ -937,6 +966,27 @@ void __unmap_hugepage_range(struct vm_area_struct *vma, unsigned long start, if (huge_pmd_unshare(mm, &address, ptep)) continue; + /* + * If a reference page is supplied, it is because a specific + * page is being unmapped, not a range. Ensure the page we + * are about to unmap is the actual page of interest. + */ + if (ref_page) { + pte = huge_ptep_get(ptep); + if (huge_pte_none(pte)) + continue; + page = pte_page(pte); + if (page != ref_page) + continue; + + /* + * Mark the VMA as having unmapped its page so that + * future faults in this VMA will fail rather than + * looking like data was lost + */ + set_vma_resv_flags(vma, HPAGE_RESV_UNMAPPED); + } + pte = huge_ptep_get_and_clear(mm, address, ptep); if (huge_pte_none(pte)) continue; @@ -955,7 +1005,7 @@ void __unmap_hugepage_range(struct vm_area_struct *vma, unsigned long start, } void unmap_hugepage_range(struct vm_area_struct *vma, unsigned long start, - unsigned long end) + unsigned long end, struct page *ref_page) { /* * It is undesirable to test vma->vm_file as it should be non-null @@ -967,19 +1017,68 @@ void unmap_hugepage_range(struct vm_area_struct *vma, unsigned long start, */ if (vma->vm_file) { spin_lock(&vma->vm_file->f_mapping->i_mmap_lock); - __unmap_hugepage_range(vma, start, end); + __unmap_hugepage_range(vma, start, end, ref_page); spin_unlock(&vma->vm_file->f_mapping->i_mmap_lock); } } +/* + * This is called when the original mapper is failing to COW a MAP_PRIVATE + * mappping it owns the reserve page for. The intention is to unmap the page + * from other VMAs and let the children be SIGKILLed if they are faulting the + * same region. + */ +int unmap_ref_private(struct mm_struct *mm, + struct vm_area_struct *vma, + struct page *page, + unsigned long address) +{ + struct vm_area_struct *iter_vma; + struct address_space *mapping; + struct prio_tree_iter iter; + pgoff_t pgoff; + + /* + * vm_pgoff is in PAGE_SIZE units, hence the different calculation + * from page cache lookup which is in HPAGE_SIZE units. + */ + address = address & huge_page_mask(hstate_vma(vma)); + pgoff = ((address - vma->vm_start) >> PAGE_SHIFT) + + (vma->vm_pgoff >> PAGE_SHIFT); + mapping = (struct address_space *)page_private(page); + + vma_prio_tree_foreach(iter_vma, &iter, &mapping->i_mmap, pgoff, pgoff) { + /* Do not unmap the current VMA */ + if (iter_vma == vma) + continue; + + /* + * Unmap the page from other VMAs without their own reserves. + * They get marked to be SIGKILLed if they fault in these + * areas. This is because a future no-page fault on this VMA + * could insert a zeroed page instead of the data existing + * from the time of fork. This would look like data corruption + */ + if (!is_vma_resv_set(iter_vma, HPAGE_RESV_OWNER)) + unmap_hugepage_range(iter_vma, + address, address + HPAGE_SIZE, + page); + } + + return 1; +} + static int hugetlb_cow(struct mm_struct *mm, struct vm_area_struct *vma, - unsigned long address, pte_t *ptep, pte_t pte) + unsigned long address, pte_t *ptep, pte_t pte, + struct page *pagecache_page) { struct page *old_page, *new_page; int avoidcopy; + int outside_reserve = 0; old_page = pte_page(pte); +retry_avoidcopy: /* If no-one else is actually using this page, avoid the copy * and just make the page writable */ avoidcopy = (page_count(old_page) == 1); @@ -988,11 +1087,43 @@ static int hugetlb_cow(struct mm_struct *mm, struct vm_area_struct *vma, return 0; } + /* + * If the process that created a MAP_PRIVATE mapping is about to + * perform a COW due to a shared page count, attempt to satisfy + * the allocation without using the existing reserves. The pagecache + * page is used to determine if the reserve at this address was + * consumed or not. If reserves were used, a partial faulted mapping + * at the time of fork() could consume its reserves on COW instead + * of the full address range. + */ + if (!(vma->vm_flags & VM_SHARED) && + is_vma_resv_set(vma, HPAGE_RESV_OWNER) && + old_page != pagecache_page) + outside_reserve = 1; + page_cache_get(old_page); - new_page = alloc_huge_page(vma, address); + new_page = alloc_huge_page(vma, address, outside_reserve); if (IS_ERR(new_page)) { page_cache_release(old_page); + + /* + * If a process owning a MAP_PRIVATE mapping fails to COW, + * it is due to references held by a child and an insufficient + * huge page pool. To guarantee the original mappers + * reliability, unmap the page from child processes. The child + * may get SIGKILLed if it later faults. + */ + if (outside_reserve) { + BUG_ON(huge_pte_none(pte)); + if (unmap_ref_private(mm, vma, old_page, address)) { + BUG_ON(page_count(old_page) != 1); + BUG_ON(huge_pte_none(pte)); + goto retry_avoidcopy; + } + WARN_ON_ONCE(1); + } + return -PTR_ERR(new_page); } @@ -1015,6 +1146,20 @@ static int hugetlb_cow(struct mm_struct *mm, struct vm_area_struct *vma, return 0; } +/* Return the pagecache page at a given address within a VMA */ +static struct page *hugetlbfs_pagecache_page(struct vm_area_struct *vma, + unsigned long address) +{ + struct address_space *mapping; + unsigned long idx; + + mapping = vma->vm_file->f_mapping; + idx = ((address - vma->vm_start) >> HPAGE_SHIFT) + + (vma->vm_pgoff >> (HPAGE_SHIFT - PAGE_SHIFT)); + + return find_lock_page(mapping, idx); +} + static int hugetlb_no_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pte_t *ptep, int write_access) { @@ -1025,6 +1170,18 @@ static int hugetlb_no_page(struct mm_struct *mm, struct vm_area_struct *vma, struct address_space *mapping; pte_t new_pte; + /* + * Currently, we are forced to kill the process in the event the + * original mapper has unmapped pages from the child due to a failed + * COW. Warn that such a situation has occured as it may not be obvious + */ + if (is_vma_resv_set(vma, HPAGE_RESV_UNMAPPED)) { + printk(KERN_WARNING + "PID %d killed due to inadequate hugepage pool\n", + current->pid); + return ret; + } + mapping = vma->vm_file->f_mapping; idx = ((address - vma->vm_start) >> HPAGE_SHIFT) + (vma->vm_pgoff >> (HPAGE_SHIFT - PAGE_SHIFT)); @@ -1039,7 +1196,7 @@ retry: size = i_size_read(mapping->host) >> HPAGE_SHIFT; if (idx >= size) goto out; - page = alloc_huge_page(vma, address); + page = alloc_huge_page(vma, address, 0); if (IS_ERR(page)) { ret = -PTR_ERR(page); goto out; @@ -1081,7 +1238,7 @@ retry: if (write_access && !(vma->vm_flags & VM_SHARED)) { /* Optimization, do the COW without a second fault */ - ret = hugetlb_cow(mm, vma, address, ptep, new_pte); + ret = hugetlb_cow(mm, vma, address, ptep, new_pte, page); } spin_unlock(&mm->page_table_lock); @@ -1126,8 +1283,15 @@ int hugetlb_fault(struct mm_struct *mm, struct vm_area_struct *vma, spin_lock(&mm->page_table_lock); /* Check for a racing update before calling hugetlb_cow */ if (likely(pte_same(entry, huge_ptep_get(ptep)))) - if (write_access && !pte_write(entry)) - ret = hugetlb_cow(mm, vma, address, ptep, entry); + if (write_access && !pte_write(entry)) { + struct page *page; + page = hugetlbfs_pagecache_page(vma, address); + ret = hugetlb_cow(mm, vma, address, ptep, entry, page); + if (page) { + unlock_page(page); + put_page(page); + } + } spin_unlock(&mm->page_table_lock); mutex_unlock(&hugetlb_instantiation_mutex); @@ -1371,6 +1535,7 @@ int hugetlb_reserve_pages(struct inode *inode, else { chg = to - from; set_vma_resv_huge_pages(vma, chg); + set_vma_resv_flags(vma, HPAGE_RESV_OWNER); } if (chg < 0) diff --git a/mm/memory.c b/mm/memory.c index 82f3f1c5cf1..72932489a08 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -901,7 +901,7 @@ unsigned long unmap_vmas(struct mmu_gather **tlbp, } if (unlikely(is_vm_hugetlb_page(vma))) { - unmap_hugepage_range(vma, start, end); + unmap_hugepage_range(vma, start, end, NULL); zap_work -= (end - start) / (HPAGE_SIZE / PAGE_SIZE); start = end; -- cgit v1.2.3-70-g09d2 From cce770815869e9209171d819dfde89bcc738ab62 Mon Sep 17 00:00:00 2001 From: Pavel Machek Date: Wed, 23 Jul 2008 21:27:36 -0700 Subject: SYNC_FILE_RANGE_WRITE may and will block. Document that. [akpm@linux-foundation.org: fix comment text] Signed-off-by: Pavel Machek Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/sync.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/sync.c b/fs/sync.c index 228e17b5e9e..2967562d416 100644 --- a/fs/sync.c +++ b/fs/sync.c @@ -139,7 +139,8 @@ asmlinkage long sys_fdatasync(unsigned int fd) * before performing the write. * * SYNC_FILE_RANGE_WRITE: initiate writeout of all those dirty pages in the - * range which are not presently under writeback. + * range which are not presently under writeback. Note that this may block for + * significant periods due to exhaustion of disk request structures. * * SYNC_FILE_RANGE_WAIT_AFTER: wait upon writeout of all pages in the range * after performing the write. -- cgit v1.2.3-70-g09d2 From a47a126ad5ea072aca3e611ed8f8dc6adad24bab Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Wed, 23 Jul 2008 21:27:38 -0700 Subject: vmallocinfo: add NUMA information Christoph recently added /proc/vmallocinfo file to get information about vmalloc allocations. This patch adds NUMA specific information, giving number of pages allocated on each memory node. This should help to check that vmalloc() is able to respect NUMA policies. Example of output on a four nodes machine (one cpu per node) 1) network hash tables are evenly spreaded on four nodes (OK) (Same point for inodes and dentries hash tables) 2) iptables tables (x_tables) are correctly allocated on each cpu node (OK). 3) sys_swapon() allocates its memory from one node only. 4) each loaded module is using memory on one node. Sysadmins could tune their setup to change points 3) and 4) if necessary. grep "pages=" /proc/vmallocinfo 0xffffc20000000000-0xffffc20000201000 2101248 alloc_large_system_hash+0x204/0x2c0 pages=512 vmalloc N0=128 N1=128 N2=128 N3=128 0xffffc20000201000-0xffffc20000302000 1052672 alloc_large_system_hash+0x204/0x2c0 pages=256 vmalloc N0=64 N1=64 N2=64 N3=64 0xffffc2000031a000-0xffffc2000031d000 12288 alloc_large_system_hash+0x204/0x2c0 pages=2 vmalloc N1=1 N2=1 0xffffc2000031f000-0xffffc2000032b000 49152 cramfs_uncompress_init+0x2e/0x80 pages=11 vmalloc N0=3 N1=3 N2=2 N3=3 0xffffc2000033e000-0xffffc20000341000 12288 sys_swapon+0x640/0xac0 pages=2 vmalloc N0=2 0xffffc20000341000-0xffffc20000344000 12288 xt_alloc_table_info+0xfe/0x130 [x_tables] pages=2 vmalloc N0=2 0xffffc20000344000-0xffffc20000347000 12288 xt_alloc_table_info+0xfe/0x130 [x_tables] pages=2 vmalloc N1=2 0xffffc20000347000-0xffffc2000034a000 12288 xt_alloc_table_info+0xfe/0x130 [x_tables] pages=2 vmalloc N2=2 0xffffc2000034a000-0xffffc2000034d000 12288 xt_alloc_table_info+0xfe/0x130 [x_tables] pages=2 vmalloc N3=2 0xffffc20004381000-0xffffc20004402000 528384 alloc_large_system_hash+0x204/0x2c0 pages=128 vmalloc N0=32 N1=32 N2=32 N3=32 0xffffc20004402000-0xffffc20004803000 4198400 alloc_large_system_hash+0x204/0x2c0 pages=1024 vmalloc vpages N0=256 N1=256 N2=256 N3=256 0xffffc20004803000-0xffffc20004904000 1052672 alloc_large_system_hash+0x204/0x2c0 pages=256 vmalloc N0=64 N1=64 N2=64 N3=64 0xffffc20004904000-0xffffc20004bec000 3047424 sys_swapon+0x640/0xac0 pages=743 vmalloc vpages N0=743 0xffffffffa0000000-0xffffffffa000f000 61440 sys_init_module+0xc27/0x1d00 pages=14 vmalloc N1=14 0xffffffffa000f000-0xffffffffa0014000 20480 sys_init_module+0xc27/0x1d00 pages=4 vmalloc N0=4 0xffffffffa0014000-0xffffffffa0017000 12288 sys_init_module+0xc27/0x1d00 pages=2 vmalloc N0=2 0xffffffffa0017000-0xffffffffa0022000 45056 sys_init_module+0xc27/0x1d00 pages=10 vmalloc N1=10 0xffffffffa0022000-0xffffffffa0028000 24576 sys_init_module+0xc27/0x1d00 pages=5 vmalloc N3=5 0xffffffffa0028000-0xffffffffa0050000 163840 sys_init_module+0xc27/0x1d00 pages=39 vmalloc N1=39 0xffffffffa0050000-0xffffffffa0052000 8192 sys_init_module+0xc27/0x1d00 pages=1 vmalloc N1=1 0xffffffffa0052000-0xffffffffa0056000 16384 sys_init_module+0xc27/0x1d00 pages=3 vmalloc N1=3 0xffffffffa0056000-0xffffffffa0081000 176128 sys_init_module+0xc27/0x1d00 pages=42 vmalloc N3=42 0xffffffffa0081000-0xffffffffa00ae000 184320 sys_init_module+0xc27/0x1d00 pages=44 vmalloc N3=44 0xffffffffa00ae000-0xffffffffa00b1000 12288 sys_init_module+0xc27/0x1d00 pages=2 vmalloc N3=2 0xffffffffa00b1000-0xffffffffa00b9000 32768 sys_init_module+0xc27/0x1d00 pages=7 vmalloc N0=7 0xffffffffa00b9000-0xffffffffa00c4000 45056 sys_init_module+0xc27/0x1d00 pages=10 vmalloc N3=10 0xffffffffa00c6000-0xffffffffa00e0000 106496 sys_init_module+0xc27/0x1d00 pages=25 vmalloc N2=25 0xffffffffa00e0000-0xffffffffa00f1000 69632 sys_init_module+0xc27/0x1d00 pages=16 vmalloc N2=16 0xffffffffa00f1000-0xffffffffa00f4000 12288 sys_init_module+0xc27/0x1d00 pages=2 vmalloc N3=2 0xffffffffa00f4000-0xffffffffa00f7000 12288 sys_init_module+0xc27/0x1d00 pages=2 vmalloc N3=2 [akpm@linux-foundation.org: fix comment] Signed-off-by: Eric Dumazet Cc: Christoph Lameter Cc: Randy Dunlap Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- Documentation/filesystems/proc.txt | 44 ++++++++++++++++++++++++++++++++++++++ fs/proc/proc_misc.c | 15 +++++++++++-- mm/vmalloc.c | 20 +++++++++++++++++ 3 files changed, 77 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/Documentation/filesystems/proc.txt b/Documentation/filesystems/proc.txt index 7f268f327d7..8c6384bdfed 100644 --- a/Documentation/filesystems/proc.txt +++ b/Documentation/filesystems/proc.txt @@ -296,6 +296,7 @@ Table 1-4: Kernel info in /proc uptime System uptime version Kernel version video bttv info of video resources (2.4) + vmallocinfo Show vmalloced areas .............................................................................. You can, for example, check which interrupts are currently in use and what @@ -557,6 +558,49 @@ VmallocTotal: total size of vmalloc memory area VmallocUsed: amount of vmalloc area which is used VmallocChunk: largest contigious block of vmalloc area which is free +.............................................................................. + +vmallocinfo: + +Provides information about vmalloced/vmaped areas. One line per area, +containing the virtual address range of the area, size in bytes, +caller information of the creator, and optional information depending +on the kind of area : + + pages=nr number of pages + phys=addr if a physical address was specified + ioremap I/O mapping (ioremap() and friends) + vmalloc vmalloc() area + vmap vmap()ed pages + user VM_USERMAP area + vpages buffer for pages pointers was vmalloced (huge area) + N=nr (Only on NUMA kernels) + Number of pages allocated on memory node + +> cat /proc/vmallocinfo +0xffffc20000000000-0xffffc20000201000 2101248 alloc_large_system_hash+0x204 ... + /0x2c0 pages=512 vmalloc N0=128 N1=128 N2=128 N3=128 +0xffffc20000201000-0xffffc20000302000 1052672 alloc_large_system_hash+0x204 ... + /0x2c0 pages=256 vmalloc N0=64 N1=64 N2=64 N3=64 +0xffffc20000302000-0xffffc20000304000 8192 acpi_tb_verify_table+0x21/0x4f... + phys=7fee8000 ioremap +0xffffc20000304000-0xffffc20000307000 12288 acpi_tb_verify_table+0x21/0x4f... + phys=7fee7000 ioremap +0xffffc2000031d000-0xffffc2000031f000 8192 init_vdso_vars+0x112/0x210 +0xffffc2000031f000-0xffffc2000032b000 49152 cramfs_uncompress_init+0x2e ... + /0x80 pages=11 vmalloc N0=3 N1=3 N2=2 N3=3 +0xffffc2000033a000-0xffffc2000033d000 12288 sys_swapon+0x640/0xac0 ... + pages=2 vmalloc N1=2 +0xffffc20000347000-0xffffc2000034c000 20480 xt_alloc_table_info+0xfe ... + /0x130 [x_tables] pages=4 vmalloc N0=4 +0xffffffffa0000000-0xffffffffa000f000 61440 sys_init_module+0xc27/0x1d00 ... + pages=14 vmalloc N2=14 +0xffffffffa000f000-0xffffffffa0014000 20480 sys_init_module+0xc27/0x1d00 ... + pages=4 vmalloc N1=4 +0xffffffffa0014000-0xffffffffa0017000 12288 sys_init_module+0xc27/0x1d00 ... + pages=2 vmalloc N1=2 +0xffffffffa0017000-0xffffffffa0022000 45056 sys_init_module+0xc27/0x1d00 ... + pages=10 vmalloc N0=10 1.3 IDE devices in /proc/ide ---------------------------- diff --git a/fs/proc/proc_misc.c b/fs/proc/proc_misc.c index b14f43d25e9..ded96986296 100644 --- a/fs/proc/proc_misc.c +++ b/fs/proc/proc_misc.c @@ -464,14 +464,25 @@ static const struct file_operations proc_slabstats_operations = { #ifdef CONFIG_MMU static int vmalloc_open(struct inode *inode, struct file *file) { - return seq_open(file, &vmalloc_op); + unsigned int *ptr = NULL; + int ret; + + if (NUMA_BUILD) + ptr = kmalloc(nr_node_ids * sizeof(unsigned int), GFP_KERNEL); + ret = seq_open(file, &vmalloc_op); + if (!ret) { + struct seq_file *m = file->private_data; + m->private = ptr; + } else + kfree(ptr); + return ret; } static const struct file_operations proc_vmalloc_operations = { .open = vmalloc_open, .read = seq_read, .llseek = seq_lseek, - .release = seq_release, + .release = seq_release_private, }; #endif diff --git a/mm/vmalloc.c b/mm/vmalloc.c index 6e45b0f3d12..35f29381629 100644 --- a/mm/vmalloc.c +++ b/mm/vmalloc.c @@ -931,6 +931,25 @@ static void s_stop(struct seq_file *m, void *p) read_unlock(&vmlist_lock); } +static void show_numa_info(struct seq_file *m, struct vm_struct *v) +{ + if (NUMA_BUILD) { + unsigned int nr, *counters = m->private; + + if (!counters) + return; + + memset(counters, 0, nr_node_ids * sizeof(unsigned int)); + + for (nr = 0; nr < v->nr_pages; nr++) + counters[page_to_nid(v->pages[nr])]++; + + for_each_node_state(nr, N_HIGH_MEMORY) + if (counters[nr]) + seq_printf(m, " N%u=%u", nr, counters[nr]); + } +} + static int s_show(struct seq_file *m, void *p) { struct vm_struct *v = p; @@ -967,6 +986,7 @@ static int s_show(struct seq_file *m, void *p) if (v->flags & VM_VPAGES) seq_printf(m, " vpages"); + show_numa_info(m, v); seq_putc(m, '\n'); return 0; } -- cgit v1.2.3-70-g09d2 From a5516438959d90b071ff0a484ce4f3f523dc3152 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Wed, 23 Jul 2008 21:27:41 -0700 Subject: hugetlb: modular state for hugetlb page size The goal of this patchset is to support multiple hugetlb page sizes. This is achieved by introducing a new struct hstate structure, which encapsulates the important hugetlb state and constants (eg. huge page size, number of huge pages currently allocated, etc). The hstate structure is then passed around the code which requires these fields, they will do the right thing regardless of the exact hstate they are operating on. This patch adds the hstate structure, with a single global instance of it (default_hstate), and does the basic work of converting hugetlb to use the hstate. Future patches will add more hstate structures to allow for different hugetlbfs mounts to have different page sizes. [akpm@linux-foundation.org: coding-style fixes] Acked-by: Adam Litke Acked-by: Nishanth Aravamudan Signed-off-by: Andi Kleen Signed-off-by: Nick Piggin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/ia64/mm/hugetlbpage.c | 7 +- arch/powerpc/mm/hugetlbpage.c | 3 +- arch/s390/mm/hugetlbpage.c | 3 +- arch/sh/mm/hugetlbpage.c | 3 +- arch/sparc64/mm/hugetlbpage.c | 5 +- arch/x86/mm/hugetlbpage.c | 5 +- fs/hugetlbfs/inode.c | 52 +++--- include/asm-ia64/hugetlb.h | 3 +- include/asm-powerpc/hugetlb.h | 3 +- include/asm-s390/hugetlb.h | 3 +- include/asm-sh/hugetlb.h | 3 +- include/asm-sparc/hugetlb.h | 3 +- include/asm-x86/hugetlb.h | 8 +- include/linux/hugetlb.h | 88 +++++++++- ipc/shm.c | 3 +- mm/hugetlb.c | 368 +++++++++++++++++++++++------------------- mm/memory.c | 2 +- mm/mempolicy.c | 9 +- mm/mmap.c | 3 +- 19 files changed, 356 insertions(+), 218 deletions(-) (limited to 'fs') diff --git a/arch/ia64/mm/hugetlbpage.c b/arch/ia64/mm/hugetlbpage.c index cd49e2860ee..6170f097d25 100644 --- a/arch/ia64/mm/hugetlbpage.c +++ b/arch/ia64/mm/hugetlbpage.c @@ -24,7 +24,7 @@ unsigned int hpage_shift=HPAGE_SHIFT_DEFAULT; pte_t * -huge_pte_alloc (struct mm_struct *mm, unsigned long addr) +huge_pte_alloc(struct mm_struct *mm, unsigned long addr, unsigned long sz) { unsigned long taddr = htlbpage_to_page(addr); pgd_t *pgd; @@ -75,7 +75,8 @@ int huge_pmd_unshare(struct mm_struct *mm, unsigned long *addr, pte_t *ptep) * Don't actually need to do any preparation, but need to make sure * the address is in the right region. */ -int prepare_hugepage_range(unsigned long addr, unsigned long len) +int prepare_hugepage_range(struct file *file, + unsigned long addr, unsigned long len) { if (len & ~HPAGE_MASK) return -EINVAL; @@ -149,7 +150,7 @@ unsigned long hugetlb_get_unmapped_area(struct file *file, unsigned long addr, u /* Handle MAP_FIXED */ if (flags & MAP_FIXED) { - if (prepare_hugepage_range(addr, len)) + if (prepare_hugepage_range(file, addr, len)) return -EINVAL; return addr; } diff --git a/arch/powerpc/mm/hugetlbpage.c b/arch/powerpc/mm/hugetlbpage.c index 1a96cc891cf..c94dc71af98 100644 --- a/arch/powerpc/mm/hugetlbpage.c +++ b/arch/powerpc/mm/hugetlbpage.c @@ -128,7 +128,8 @@ pte_t *huge_pte_offset(struct mm_struct *mm, unsigned long addr) return NULL; } -pte_t *huge_pte_alloc(struct mm_struct *mm, unsigned long addr) +pte_t *huge_pte_alloc(struct mm_struct *mm, + unsigned long addr, unsigned long sz) { pgd_t *pg; pud_t *pu; diff --git a/arch/s390/mm/hugetlbpage.c b/arch/s390/mm/hugetlbpage.c index f4b6124fdb7..9162dc84f77 100644 --- a/arch/s390/mm/hugetlbpage.c +++ b/arch/s390/mm/hugetlbpage.c @@ -72,7 +72,8 @@ void arch_release_hugepage(struct page *page) page[1].index = 0; } -pte_t *huge_pte_alloc(struct mm_struct *mm, unsigned long addr) +pte_t *huge_pte_alloc(struct mm_struct *mm, + unsigned long addr, unsigned long sz) { pgd_t *pgdp; pud_t *pudp; diff --git a/arch/sh/mm/hugetlbpage.c b/arch/sh/mm/hugetlbpage.c index ae8c321d6e2..2f9dbe0ef4a 100644 --- a/arch/sh/mm/hugetlbpage.c +++ b/arch/sh/mm/hugetlbpage.c @@ -22,7 +22,8 @@ #include #include -pte_t *huge_pte_alloc(struct mm_struct *mm, unsigned long addr) +pte_t *huge_pte_alloc(struct mm_struct *mm, + unsigned long addr, unsigned long sz) { pgd_t *pgd; pud_t *pud; diff --git a/arch/sparc64/mm/hugetlbpage.c b/arch/sparc64/mm/hugetlbpage.c index ebefd2a1437..1307b23f6a7 100644 --- a/arch/sparc64/mm/hugetlbpage.c +++ b/arch/sparc64/mm/hugetlbpage.c @@ -175,7 +175,7 @@ hugetlb_get_unmapped_area(struct file *file, unsigned long addr, return -ENOMEM; if (flags & MAP_FIXED) { - if (prepare_hugepage_range(addr, len)) + if (prepare_hugepage_range(file, addr, len)) return -EINVAL; return addr; } @@ -195,7 +195,8 @@ hugetlb_get_unmapped_area(struct file *file, unsigned long addr, pgoff, flags); } -pte_t *huge_pte_alloc(struct mm_struct *mm, unsigned long addr) +pte_t *huge_pte_alloc(struct mm_struct *mm, + unsigned long addr, unsigned long sz) { pgd_t *pgd; pud_t *pud; diff --git a/arch/x86/mm/hugetlbpage.c b/arch/x86/mm/hugetlbpage.c index 0b3d567e686..52476fde899 100644 --- a/arch/x86/mm/hugetlbpage.c +++ b/arch/x86/mm/hugetlbpage.c @@ -124,7 +124,8 @@ int huge_pmd_unshare(struct mm_struct *mm, unsigned long *addr, pte_t *ptep) return 1; } -pte_t *huge_pte_alloc(struct mm_struct *mm, unsigned long addr) +pte_t *huge_pte_alloc(struct mm_struct *mm, + unsigned long addr, unsigned long sz) { pgd_t *pgd; pud_t *pud; @@ -368,7 +369,7 @@ hugetlb_get_unmapped_area(struct file *file, unsigned long addr, return -ENOMEM; if (flags & MAP_FIXED) { - if (prepare_hugepage_range(addr, len)) + if (prepare_hugepage_range(file, addr, len)) return -EINVAL; return addr; } diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c index 428eff5b73f..516c581b537 100644 --- a/fs/hugetlbfs/inode.c +++ b/fs/hugetlbfs/inode.c @@ -80,6 +80,7 @@ static int hugetlbfs_file_mmap(struct file *file, struct vm_area_struct *vma) struct inode *inode = file->f_path.dentry->d_inode; loff_t len, vma_len; int ret; + struct hstate *h = hstate_file(file); /* * vma address alignment (but not the pgoff alignment) has @@ -92,7 +93,7 @@ static int hugetlbfs_file_mmap(struct file *file, struct vm_area_struct *vma) vma->vm_flags |= VM_HUGETLB | VM_RESERVED; vma->vm_ops = &hugetlb_vm_ops; - if (vma->vm_pgoff & ~(HPAGE_MASK >> PAGE_SHIFT)) + if (vma->vm_pgoff & ~(huge_page_mask(h) >> PAGE_SHIFT)) return -EINVAL; vma_len = (loff_t)(vma->vm_end - vma->vm_start); @@ -104,8 +105,8 @@ static int hugetlbfs_file_mmap(struct file *file, struct vm_area_struct *vma) len = vma_len + ((loff_t)vma->vm_pgoff << PAGE_SHIFT); if (hugetlb_reserve_pages(inode, - vma->vm_pgoff >> (HPAGE_SHIFT-PAGE_SHIFT), - len >> HPAGE_SHIFT, vma)) + vma->vm_pgoff >> huge_page_order(h), + len >> huge_page_shift(h), vma)) goto out; ret = 0; @@ -130,20 +131,21 @@ hugetlb_get_unmapped_area(struct file *file, unsigned long addr, struct mm_struct *mm = current->mm; struct vm_area_struct *vma; unsigned long start_addr; + struct hstate *h = hstate_file(file); - if (len & ~HPAGE_MASK) + if (len & ~huge_page_mask(h)) return -EINVAL; if (len > TASK_SIZE) return -ENOMEM; if (flags & MAP_FIXED) { - if (prepare_hugepage_range(addr, len)) + if (prepare_hugepage_range(file, addr, len)) return -EINVAL; return addr; } if (addr) { - addr = ALIGN(addr, HPAGE_SIZE); + addr = ALIGN(addr, huge_page_size(h)); vma = find_vma(mm, addr); if (TASK_SIZE - len >= addr && (!vma || addr + len <= vma->vm_start)) @@ -156,7 +158,7 @@ hugetlb_get_unmapped_area(struct file *file, unsigned long addr, start_addr = TASK_UNMAPPED_BASE; full_search: - addr = ALIGN(start_addr, HPAGE_SIZE); + addr = ALIGN(start_addr, huge_page_size(h)); for (vma = find_vma(mm, addr); ; vma = vma->vm_next) { /* At this point: (!vma || addr < vma->vm_end). */ @@ -174,7 +176,7 @@ full_search: if (!vma || addr + len <= vma->vm_start) return addr; - addr = ALIGN(vma->vm_end, HPAGE_SIZE); + addr = ALIGN(vma->vm_end, huge_page_size(h)); } } #endif @@ -225,10 +227,11 @@ hugetlbfs_read_actor(struct page *page, unsigned long offset, static ssize_t hugetlbfs_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos) { + struct hstate *h = hstate_file(filp); struct address_space *mapping = filp->f_mapping; struct inode *inode = mapping->host; - unsigned long index = *ppos >> HPAGE_SHIFT; - unsigned long offset = *ppos & ~HPAGE_MASK; + unsigned long index = *ppos >> huge_page_shift(h); + unsigned long offset = *ppos & ~huge_page_mask(h); unsigned long end_index; loff_t isize; ssize_t retval = 0; @@ -243,17 +246,17 @@ static ssize_t hugetlbfs_read(struct file *filp, char __user *buf, if (!isize) goto out; - end_index = (isize - 1) >> HPAGE_SHIFT; + end_index = (isize - 1) >> huge_page_shift(h); for (;;) { struct page *page; - int nr, ret; + unsigned long nr, ret; /* nr is the maximum number of bytes to copy from this page */ - nr = HPAGE_SIZE; + nr = huge_page_size(h); if (index >= end_index) { if (index > end_index) goto out; - nr = ((isize - 1) & ~HPAGE_MASK) + 1; + nr = ((isize - 1) & ~huge_page_mask(h)) + 1; if (nr <= offset) { goto out; } @@ -287,8 +290,8 @@ static ssize_t hugetlbfs_read(struct file *filp, char __user *buf, offset += ret; retval += ret; len -= ret; - index += offset >> HPAGE_SHIFT; - offset &= ~HPAGE_MASK; + index += offset >> huge_page_shift(h); + offset &= ~huge_page_mask(h); if (page) page_cache_release(page); @@ -298,7 +301,7 @@ static ssize_t hugetlbfs_read(struct file *filp, char __user *buf, break; } out: - *ppos = ((loff_t)index << HPAGE_SHIFT) + offset; + *ppos = ((loff_t)index << huge_page_shift(h)) + offset; mutex_unlock(&inode->i_mutex); return retval; } @@ -339,8 +342,9 @@ static void truncate_huge_page(struct page *page) static void truncate_hugepages(struct inode *inode, loff_t lstart) { + struct hstate *h = hstate_inode(inode); struct address_space *mapping = &inode->i_data; - const pgoff_t start = lstart >> HPAGE_SHIFT; + const pgoff_t start = lstart >> huge_page_shift(h); struct pagevec pvec; pgoff_t next; int i, freed = 0; @@ -449,8 +453,9 @@ static int hugetlb_vmtruncate(struct inode *inode, loff_t offset) { pgoff_t pgoff; struct address_space *mapping = inode->i_mapping; + struct hstate *h = hstate_inode(inode); - BUG_ON(offset & ~HPAGE_MASK); + BUG_ON(offset & ~huge_page_mask(h)); pgoff = offset >> PAGE_SHIFT; i_size_write(inode, offset); @@ -465,6 +470,7 @@ static int hugetlb_vmtruncate(struct inode *inode, loff_t offset) static int hugetlbfs_setattr(struct dentry *dentry, struct iattr *attr) { struct inode *inode = dentry->d_inode; + struct hstate *h = hstate_inode(inode); int error; unsigned int ia_valid = attr->ia_valid; @@ -476,7 +482,7 @@ static int hugetlbfs_setattr(struct dentry *dentry, struct iattr *attr) if (ia_valid & ATTR_SIZE) { error = -EINVAL; - if (!(attr->ia_size & ~HPAGE_MASK)) + if (!(attr->ia_size & ~huge_page_mask(h))) error = hugetlb_vmtruncate(inode, attr->ia_size); if (error) goto out; @@ -610,9 +616,10 @@ static int hugetlbfs_set_page_dirty(struct page *page) static int hugetlbfs_statfs(struct dentry *dentry, struct kstatfs *buf) { struct hugetlbfs_sb_info *sbinfo = HUGETLBFS_SB(dentry->d_sb); + struct hstate *h = hstate_inode(dentry->d_inode); buf->f_type = HUGETLBFS_MAGIC; - buf->f_bsize = HPAGE_SIZE; + buf->f_bsize = huge_page_size(h); if (sbinfo) { spin_lock(&sbinfo->stat_lock); /* If no limits set, just report 0 for max/free/used @@ -942,7 +949,8 @@ struct file *hugetlb_file_setup(const char *name, size_t size) goto out_dentry; error = -ENOMEM; - if (hugetlb_reserve_pages(inode, 0, size >> HPAGE_SHIFT, NULL)) + if (hugetlb_reserve_pages(inode, 0, + size >> huge_page_shift(hstate_inode(inode)), NULL)) goto out_inode; d_instantiate(dentry, inode); diff --git a/include/asm-ia64/hugetlb.h b/include/asm-ia64/hugetlb.h index e9d1e5e2382..da55c63728e 100644 --- a/include/asm-ia64/hugetlb.h +++ b/include/asm-ia64/hugetlb.h @@ -8,7 +8,8 @@ void hugetlb_free_pgd_range(struct mmu_gather *tlb, unsigned long addr, unsigned long end, unsigned long floor, unsigned long ceiling); -int prepare_hugepage_range(unsigned long addr, unsigned long len); +int prepare_hugepage_range(struct file *file, + unsigned long addr, unsigned long len); static inline int is_hugepage_only_range(struct mm_struct *mm, unsigned long addr, diff --git a/include/asm-powerpc/hugetlb.h b/include/asm-powerpc/hugetlb.h index 0a37aa5ecaa..ca37c4af27b 100644 --- a/include/asm-powerpc/hugetlb.h +++ b/include/asm-powerpc/hugetlb.h @@ -21,7 +21,8 @@ pte_t huge_ptep_get_and_clear(struct mm_struct *mm, unsigned long addr, * If the arch doesn't supply something else, assume that hugepage * size aligned regions are ok without further preparation. */ -static inline int prepare_hugepage_range(unsigned long addr, unsigned long len) +static inline int prepare_hugepage_range(struct file *file, + unsigned long addr, unsigned long len) { if (len & ~HPAGE_MASK) return -EINVAL; diff --git a/include/asm-s390/hugetlb.h b/include/asm-s390/hugetlb.h index 600a776f8f7..670a1d1745d 100644 --- a/include/asm-s390/hugetlb.h +++ b/include/asm-s390/hugetlb.h @@ -22,7 +22,8 @@ void set_huge_pte_at(struct mm_struct *mm, unsigned long addr, * If the arch doesn't supply something else, assume that hugepage * size aligned regions are ok without further preparation. */ -static inline int prepare_hugepage_range(unsigned long addr, unsigned long len) +static inline int prepare_hugepage_range(struct file *file, + unsigned long addr, unsigned long len) { if (len & ~HPAGE_MASK) return -EINVAL; diff --git a/include/asm-sh/hugetlb.h b/include/asm-sh/hugetlb.h index fb30018938c..967068fb79a 100644 --- a/include/asm-sh/hugetlb.h +++ b/include/asm-sh/hugetlb.h @@ -14,7 +14,8 @@ static inline int is_hugepage_only_range(struct mm_struct *mm, * If the arch doesn't supply something else, assume that hugepage * size aligned regions are ok without further preparation. */ -static inline int prepare_hugepage_range(unsigned long addr, unsigned long len) +static inline int prepare_hugepage_range(struct file *file, + unsigned long addr, unsigned long len) { if (len & ~HPAGE_MASK) return -EINVAL; diff --git a/include/asm-sparc/hugetlb.h b/include/asm-sparc/hugetlb.h index aeb92374ca3..177061064ee 100644 --- a/include/asm-sparc/hugetlb.h +++ b/include/asm-sparc/hugetlb.h @@ -22,7 +22,8 @@ static inline int is_hugepage_only_range(struct mm_struct *mm, * If the arch doesn't supply something else, assume that hugepage * size aligned regions are ok without further preparation. */ -static inline int prepare_hugepage_range(unsigned long addr, unsigned long len) +static inline int prepare_hugepage_range(struct file *file, + unsigned long addr, unsigned long len) { if (len & ~HPAGE_MASK) return -EINVAL; diff --git a/include/asm-x86/hugetlb.h b/include/asm-x86/hugetlb.h index 7eed6e0883b..439a9acc132 100644 --- a/include/asm-x86/hugetlb.h +++ b/include/asm-x86/hugetlb.h @@ -14,11 +14,13 @@ static inline int is_hugepage_only_range(struct mm_struct *mm, * If the arch doesn't supply something else, assume that hugepage * size aligned regions are ok without further preparation. */ -static inline int prepare_hugepage_range(unsigned long addr, unsigned long len) +static inline int prepare_hugepage_range(struct file *file, + unsigned long addr, unsigned long len) { - if (len & ~HPAGE_MASK) + struct hstate *h = hstate_file(file); + if (len & ~huge_page_mask(h)) return -EINVAL; - if (addr & ~HPAGE_MASK) + if (addr & ~huge_page_mask(h)) return -EINVAL; return 0; } diff --git a/include/linux/hugetlb.h b/include/linux/hugetlb.h index abbc187193a..ad2271e11f9 100644 --- a/include/linux/hugetlb.h +++ b/include/linux/hugetlb.h @@ -8,7 +8,6 @@ #include #include #include -#include struct ctl_table; @@ -45,7 +44,8 @@ extern int sysctl_hugetlb_shm_group; /* arch callbacks */ -pte_t *huge_pte_alloc(struct mm_struct *mm, unsigned long addr); +pte_t *huge_pte_alloc(struct mm_struct *mm, + unsigned long addr, unsigned long sz); pte_t *huge_pte_offset(struct mm_struct *mm, unsigned long addr); int huge_pmd_unshare(struct mm_struct *mm, unsigned long *addr, pte_t *ptep); struct page *follow_huge_addr(struct mm_struct *mm, unsigned long address, @@ -80,7 +80,7 @@ static inline unsigned long hugetlb_total_pages(void) #define hugetlb_report_meminfo(buf) 0 #define hugetlb_report_node_meminfo(n, buf) 0 #define follow_huge_pmd(mm, addr, pmd, write) NULL -#define prepare_hugepage_range(addr,len) (-EINVAL) +#define prepare_hugepage_range(file, addr, len) (-EINVAL) #define pmd_huge(x) 0 #define is_hugepage_only_range(mm, addr, len) 0 #define hugetlb_free_pgd_range(tlb, addr, end, floor, ceiling) ({BUG(); 0; }) @@ -134,8 +134,6 @@ struct file *hugetlb_file_setup(const char *name, size_t); int hugetlb_get_quota(struct address_space *mapping, long delta); void hugetlb_put_quota(struct address_space *mapping, long delta); -#define BLOCKS_PER_HUGEPAGE (HPAGE_SIZE / 512) - static inline int is_file_hugepages(struct file *file) { if (file->f_op == &hugetlbfs_file_operations) @@ -164,4 +162,84 @@ unsigned long hugetlb_get_unmapped_area(struct file *file, unsigned long addr, unsigned long flags); #endif /* HAVE_ARCH_HUGETLB_UNMAPPED_AREA */ +#ifdef CONFIG_HUGETLB_PAGE + +/* Defines one hugetlb page size */ +struct hstate { + int hugetlb_next_nid; + unsigned int order; + unsigned long mask; + unsigned long max_huge_pages; + unsigned long nr_huge_pages; + unsigned long free_huge_pages; + unsigned long resv_huge_pages; + unsigned long surplus_huge_pages; + unsigned long nr_overcommit_huge_pages; + struct list_head hugepage_freelists[MAX_NUMNODES]; + unsigned int nr_huge_pages_node[MAX_NUMNODES]; + unsigned int free_huge_pages_node[MAX_NUMNODES]; + unsigned int surplus_huge_pages_node[MAX_NUMNODES]; +}; + +extern struct hstate default_hstate; + +static inline struct hstate *hstate_vma(struct vm_area_struct *vma) +{ + return &default_hstate; +} + +static inline struct hstate *hstate_file(struct file *f) +{ + return &default_hstate; +} + +static inline struct hstate *hstate_inode(struct inode *i) +{ + return &default_hstate; +} + +static inline unsigned long huge_page_size(struct hstate *h) +{ + return (unsigned long)PAGE_SIZE << h->order; +} + +static inline unsigned long huge_page_mask(struct hstate *h) +{ + return h->mask; +} + +static inline unsigned int huge_page_order(struct hstate *h) +{ + return h->order; +} + +static inline unsigned huge_page_shift(struct hstate *h) +{ + return h->order + PAGE_SHIFT; +} + +static inline unsigned int pages_per_huge_page(struct hstate *h) +{ + return 1 << h->order; +} + +static inline unsigned int blocks_per_huge_page(struct hstate *h) +{ + return huge_page_size(h) / 512; +} + +#include + +#else +struct hstate {}; +#define hstate_file(f) NULL +#define hstate_vma(v) NULL +#define hstate_inode(i) NULL +#define huge_page_size(h) PAGE_SIZE +#define huge_page_mask(h) PAGE_MASK +#define huge_page_order(h) 0 +#define huge_page_shift(h) PAGE_SHIFT +#define pages_per_huge_page(h) 1 +#endif + #endif /* _LINUX_HUGETLB_H */ diff --git a/ipc/shm.c b/ipc/shm.c index 790240cd067..a726aebce7d 100644 --- a/ipc/shm.c +++ b/ipc/shm.c @@ -577,7 +577,8 @@ static void shm_get_stat(struct ipc_namespace *ns, unsigned long *rss, if (is_file_hugepages(shp->shm_file)) { struct address_space *mapping = inode->i_mapping; - *rss += (HPAGE_SIZE/PAGE_SIZE)*mapping->nrpages; + struct hstate *h = hstate_file(shp->shm_file); + *rss += pages_per_huge_page(h) * mapping->nrpages; } else { struct shmem_inode_info *info = SHMEM_I(inode); spin_lock(&info->lock); diff --git a/mm/hugetlb.c b/mm/hugetlb.c index 32dff4290c6..0d8153e25f0 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -22,18 +22,12 @@ #include "internal.h" const unsigned long hugetlb_zero = 0, hugetlb_infinity = ~0UL; -static unsigned long nr_huge_pages, free_huge_pages, resv_huge_pages; -static unsigned long surplus_huge_pages; -static unsigned long nr_overcommit_huge_pages; unsigned long max_huge_pages; unsigned long sysctl_overcommit_huge_pages; -static struct list_head hugepage_freelists[MAX_NUMNODES]; -static unsigned int nr_huge_pages_node[MAX_NUMNODES]; -static unsigned int free_huge_pages_node[MAX_NUMNODES]; -static unsigned int surplus_huge_pages_node[MAX_NUMNODES]; static gfp_t htlb_alloc_mask = GFP_HIGHUSER; unsigned long hugepages_treat_as_movable; -static int hugetlb_next_nid; + +struct hstate default_hstate; /* * Protects updates to hugepage_freelists, nr_huge_pages, and free_huge_pages @@ -203,11 +197,11 @@ static long region_count(struct list_head *head, long f, long t) * Convert the address within this vma to the page offset within * the mapping, in pagecache page units; huge pages here. */ -static pgoff_t vma_hugecache_offset(struct vm_area_struct *vma, - unsigned long address) +static pgoff_t vma_hugecache_offset(struct hstate *h, + struct vm_area_struct *vma, unsigned long address) { - return ((address - vma->vm_start) >> HPAGE_SHIFT) + - (vma->vm_pgoff >> (HPAGE_SHIFT - PAGE_SHIFT)); + return ((address - vma->vm_start) >> huge_page_shift(h)) + + (vma->vm_pgoff >> huge_page_order(h)); } /* @@ -309,20 +303,21 @@ static int is_vma_resv_set(struct vm_area_struct *vma, unsigned long flag) } /* Decrement the reserved pages in the hugepage pool by one */ -static void decrement_hugepage_resv_vma(struct vm_area_struct *vma) +static void decrement_hugepage_resv_vma(struct hstate *h, + struct vm_area_struct *vma) { if (vma->vm_flags & VM_NORESERVE) return; if (vma->vm_flags & VM_SHARED) { /* Shared mappings always use reserves */ - resv_huge_pages--; + h->resv_huge_pages--; } else if (is_vma_resv_set(vma, HPAGE_RESV_OWNER)) { /* * Only the process that called mmap() has reserves for * private mappings. */ - resv_huge_pages--; + h->resv_huge_pages--; } } @@ -344,12 +339,13 @@ static int vma_has_private_reserves(struct vm_area_struct *vma) return 1; } -static void clear_huge_page(struct page *page, unsigned long addr) +static void clear_huge_page(struct page *page, + unsigned long addr, unsigned long sz) { int i; might_sleep(); - for (i = 0; i < (HPAGE_SIZE/PAGE_SIZE); i++) { + for (i = 0; i < sz/PAGE_SIZE; i++) { cond_resched(); clear_user_highpage(page + i, addr + i * PAGE_SIZE); } @@ -359,41 +355,43 @@ static void copy_huge_page(struct page *dst, struct page *src, unsigned long addr, struct vm_area_struct *vma) { int i; + struct hstate *h = hstate_vma(vma); might_sleep(); - for (i = 0; i < HPAGE_SIZE/PAGE_SIZE; i++) { + for (i = 0; i < pages_per_huge_page(h); i++) { cond_resched(); copy_user_highpage(dst + i, src + i, addr + i*PAGE_SIZE, vma); } } -static void enqueue_huge_page(struct page *page) +static void enqueue_huge_page(struct hstate *h, struct page *page) { int nid = page_to_nid(page); - list_add(&page->lru, &hugepage_freelists[nid]); - free_huge_pages++; - free_huge_pages_node[nid]++; + list_add(&page->lru, &h->hugepage_freelists[nid]); + h->free_huge_pages++; + h->free_huge_pages_node[nid]++; } -static struct page *dequeue_huge_page(void) +static struct page *dequeue_huge_page(struct hstate *h) { int nid; struct page *page = NULL; for (nid = 0; nid < MAX_NUMNODES; ++nid) { - if (!list_empty(&hugepage_freelists[nid])) { - page = list_entry(hugepage_freelists[nid].next, + if (!list_empty(&h->hugepage_freelists[nid])) { + page = list_entry(h->hugepage_freelists[nid].next, struct page, lru); list_del(&page->lru); - free_huge_pages--; - free_huge_pages_node[nid]--; + h->free_huge_pages--; + h->free_huge_pages_node[nid]--; break; } } return page; } -static struct page *dequeue_huge_page_vma(struct vm_area_struct *vma, +static struct page *dequeue_huge_page_vma(struct hstate *h, + struct vm_area_struct *vma, unsigned long address, int avoid_reserve) { int nid; @@ -411,26 +409,26 @@ static struct page *dequeue_huge_page_vma(struct vm_area_struct *vma, * not "stolen". The child may still get SIGKILLed */ if (!vma_has_private_reserves(vma) && - free_huge_pages - resv_huge_pages == 0) + h->free_huge_pages - h->resv_huge_pages == 0) return NULL; /* If reserves cannot be used, ensure enough pages are in the pool */ - if (avoid_reserve && free_huge_pages - resv_huge_pages == 0) + if (avoid_reserve && h->free_huge_pages - h->resv_huge_pages == 0) return NULL; for_each_zone_zonelist_nodemask(zone, z, zonelist, MAX_NR_ZONES - 1, nodemask) { nid = zone_to_nid(zone); if (cpuset_zone_allowed_softwall(zone, htlb_alloc_mask) && - !list_empty(&hugepage_freelists[nid])) { - page = list_entry(hugepage_freelists[nid].next, + !list_empty(&h->hugepage_freelists[nid])) { + page = list_entry(h->hugepage_freelists[nid].next, struct page, lru); list_del(&page->lru); - free_huge_pages--; - free_huge_pages_node[nid]--; + h->free_huge_pages--; + h->free_huge_pages_node[nid]--; if (!avoid_reserve) - decrement_hugepage_resv_vma(vma); + decrement_hugepage_resv_vma(h, vma); break; } @@ -439,12 +437,13 @@ static struct page *dequeue_huge_page_vma(struct vm_area_struct *vma, return page; } -static void update_and_free_page(struct page *page) +static void update_and_free_page(struct hstate *h, struct page *page) { int i; - nr_huge_pages--; - nr_huge_pages_node[page_to_nid(page)]--; - for (i = 0; i < (HPAGE_SIZE / PAGE_SIZE); i++) { + + h->nr_huge_pages--; + h->nr_huge_pages_node[page_to_nid(page)]--; + for (i = 0; i < pages_per_huge_page(h); i++) { page[i].flags &= ~(1 << PG_locked | 1 << PG_error | 1 << PG_referenced | 1 << PG_dirty | 1 << PG_active | 1 << PG_reserved | 1 << PG_private | 1<< PG_writeback); @@ -452,11 +451,16 @@ static void update_and_free_page(struct page *page) set_compound_page_dtor(page, NULL); set_page_refcounted(page); arch_release_hugepage(page); - __free_pages(page, HUGETLB_PAGE_ORDER); + __free_pages(page, huge_page_order(h)); } static void free_huge_page(struct page *page) { + /* + * Can't pass hstate in here because it is called from the + * compound page destructor. + */ + struct hstate *h = &default_hstate; int nid = page_to_nid(page); struct address_space *mapping; @@ -466,12 +470,12 @@ static void free_huge_page(struct page *page) INIT_LIST_HEAD(&page->lru); spin_lock(&hugetlb_lock); - if (surplus_huge_pages_node[nid]) { - update_and_free_page(page); - surplus_huge_pages--; - surplus_huge_pages_node[nid]--; + if (h->surplus_huge_pages_node[nid]) { + update_and_free_page(h, page); + h->surplus_huge_pages--; + h->surplus_huge_pages_node[nid]--; } else { - enqueue_huge_page(page); + enqueue_huge_page(h, page); } spin_unlock(&hugetlb_lock); if (mapping) @@ -483,7 +487,7 @@ static void free_huge_page(struct page *page) * balanced by operating on them in a round-robin fashion. * Returns 1 if an adjustment was made. */ -static int adjust_pool_surplus(int delta) +static int adjust_pool_surplus(struct hstate *h, int delta) { static int prev_nid; int nid = prev_nid; @@ -496,15 +500,15 @@ static int adjust_pool_surplus(int delta) nid = first_node(node_online_map); /* To shrink on this node, there must be a surplus page */ - if (delta < 0 && !surplus_huge_pages_node[nid]) + if (delta < 0 && !h->surplus_huge_pages_node[nid]) continue; /* Surplus cannot exceed the total number of pages */ - if (delta > 0 && surplus_huge_pages_node[nid] >= - nr_huge_pages_node[nid]) + if (delta > 0 && h->surplus_huge_pages_node[nid] >= + h->nr_huge_pages_node[nid]) continue; - surplus_huge_pages += delta; - surplus_huge_pages_node[nid] += delta; + h->surplus_huge_pages += delta; + h->surplus_huge_pages_node[nid] += delta; ret = 1; break; } while (nid != prev_nid); @@ -513,46 +517,46 @@ static int adjust_pool_surplus(int delta) return ret; } -static void prep_new_huge_page(struct page *page, int nid) +static void prep_new_huge_page(struct hstate *h, struct page *page, int nid) { set_compound_page_dtor(page, free_huge_page); spin_lock(&hugetlb_lock); - nr_huge_pages++; - nr_huge_pages_node[nid]++; + h->nr_huge_pages++; + h->nr_huge_pages_node[nid]++; spin_unlock(&hugetlb_lock); put_page(page); /* free it into the hugepage allocator */ } -static struct page *alloc_fresh_huge_page_node(int nid) +static struct page *alloc_fresh_huge_page_node(struct hstate *h, int nid) { struct page *page; page = alloc_pages_node(nid, htlb_alloc_mask|__GFP_COMP|__GFP_THISNODE| __GFP_REPEAT|__GFP_NOWARN, - HUGETLB_PAGE_ORDER); + huge_page_order(h)); if (page) { if (arch_prepare_hugepage(page)) { __free_pages(page, HUGETLB_PAGE_ORDER); return NULL; } - prep_new_huge_page(page, nid); + prep_new_huge_page(h, page, nid); } return page; } -static int alloc_fresh_huge_page(void) +static int alloc_fresh_huge_page(struct hstate *h) { struct page *page; int start_nid; int next_nid; int ret = 0; - start_nid = hugetlb_next_nid; + start_nid = h->hugetlb_next_nid; do { - page = alloc_fresh_huge_page_node(hugetlb_next_nid); + page = alloc_fresh_huge_page_node(h, h->hugetlb_next_nid); if (page) ret = 1; /* @@ -566,11 +570,11 @@ static int alloc_fresh_huge_page(void) * if we just successfully allocated a hugepage so that * the next caller gets hugepages on the next node. */ - next_nid = next_node(hugetlb_next_nid, node_online_map); + next_nid = next_node(h->hugetlb_next_nid, node_online_map); if (next_nid == MAX_NUMNODES) next_nid = first_node(node_online_map); - hugetlb_next_nid = next_nid; - } while (!page && hugetlb_next_nid != start_nid); + h->hugetlb_next_nid = next_nid; + } while (!page && h->hugetlb_next_nid != start_nid); if (ret) count_vm_event(HTLB_BUDDY_PGALLOC); @@ -580,8 +584,8 @@ static int alloc_fresh_huge_page(void) return ret; } -static struct page *alloc_buddy_huge_page(struct vm_area_struct *vma, - unsigned long address) +static struct page *alloc_buddy_huge_page(struct hstate *h, + struct vm_area_struct *vma, unsigned long address) { struct page *page; unsigned int nid; @@ -610,18 +614,18 @@ static struct page *alloc_buddy_huge_page(struct vm_area_struct *vma, * per-node value is checked there. */ spin_lock(&hugetlb_lock); - if (surplus_huge_pages >= nr_overcommit_huge_pages) { + if (h->surplus_huge_pages >= h->nr_overcommit_huge_pages) { spin_unlock(&hugetlb_lock); return NULL; } else { - nr_huge_pages++; - surplus_huge_pages++; + h->nr_huge_pages++; + h->surplus_huge_pages++; } spin_unlock(&hugetlb_lock); page = alloc_pages(htlb_alloc_mask|__GFP_COMP| __GFP_REPEAT|__GFP_NOWARN, - HUGETLB_PAGE_ORDER); + huge_page_order(h)); spin_lock(&hugetlb_lock); if (page) { @@ -636,12 +640,12 @@ static struct page *alloc_buddy_huge_page(struct vm_area_struct *vma, /* * We incremented the global counters already */ - nr_huge_pages_node[nid]++; - surplus_huge_pages_node[nid]++; + h->nr_huge_pages_node[nid]++; + h->surplus_huge_pages_node[nid]++; __count_vm_event(HTLB_BUDDY_PGALLOC); } else { - nr_huge_pages--; - surplus_huge_pages--; + h->nr_huge_pages--; + h->surplus_huge_pages--; __count_vm_event(HTLB_BUDDY_PGALLOC_FAIL); } spin_unlock(&hugetlb_lock); @@ -653,16 +657,16 @@ static struct page *alloc_buddy_huge_page(struct vm_area_struct *vma, * Increase the hugetlb pool such that it can accomodate a reservation * of size 'delta'. */ -static int gather_surplus_pages(int delta) +static int gather_surplus_pages(struct hstate *h, int delta) { struct list_head surplus_list; struct page *page, *tmp; int ret, i; int needed, allocated; - needed = (resv_huge_pages + delta) - free_huge_pages; + needed = (h->resv_huge_pages + delta) - h->free_huge_pages; if (needed <= 0) { - resv_huge_pages += delta; + h->resv_huge_pages += delta; return 0; } @@ -673,7 +677,7 @@ static int gather_surplus_pages(int delta) retry: spin_unlock(&hugetlb_lock); for (i = 0; i < needed; i++) { - page = alloc_buddy_huge_page(NULL, 0); + page = alloc_buddy_huge_page(h, NULL, 0); if (!page) { /* * We were not able to allocate enough pages to @@ -694,7 +698,8 @@ retry: * because either resv_huge_pages or free_huge_pages may have changed. */ spin_lock(&hugetlb_lock); - needed = (resv_huge_pages + delta) - (free_huge_pages + allocated); + needed = (h->resv_huge_pages + delta) - + (h->free_huge_pages + allocated); if (needed > 0) goto retry; @@ -707,7 +712,7 @@ retry: * before they are reserved. */ needed += allocated; - resv_huge_pages += delta; + h->resv_huge_pages += delta; ret = 0; free: /* Free the needed pages to the hugetlb pool */ @@ -715,7 +720,7 @@ free: if ((--needed) < 0) break; list_del(&page->lru); - enqueue_huge_page(page); + enqueue_huge_page(h, page); } /* Free unnecessary surplus pages to the buddy allocator */ @@ -743,7 +748,8 @@ free: * allocated to satisfy the reservation must be explicitly freed if they were * never used. */ -static void return_unused_surplus_pages(unsigned long unused_resv_pages) +static void return_unused_surplus_pages(struct hstate *h, + unsigned long unused_resv_pages) { static int nid = -1; struct page *page; @@ -758,27 +764,27 @@ static void return_unused_surplus_pages(unsigned long unused_resv_pages) unsigned long remaining_iterations = num_online_nodes(); /* Uncommit the reservation */ - resv_huge_pages -= unused_resv_pages; + h->resv_huge_pages -= unused_resv_pages; - nr_pages = min(unused_resv_pages, surplus_huge_pages); + nr_pages = min(unused_resv_pages, h->surplus_huge_pages); while (remaining_iterations-- && nr_pages) { nid = next_node(nid, node_online_map); if (nid == MAX_NUMNODES) nid = first_node(node_online_map); - if (!surplus_huge_pages_node[nid]) + if (!h->surplus_huge_pages_node[nid]) continue; - if (!list_empty(&hugepage_freelists[nid])) { - page = list_entry(hugepage_freelists[nid].next, + if (!list_empty(&h->hugepage_freelists[nid])) { + page = list_entry(h->hugepage_freelists[nid].next, struct page, lru); list_del(&page->lru); - update_and_free_page(page); - free_huge_pages--; - free_huge_pages_node[nid]--; - surplus_huge_pages--; - surplus_huge_pages_node[nid]--; + update_and_free_page(h, page); + h->free_huge_pages--; + h->free_huge_pages_node[nid]--; + h->surplus_huge_pages--; + h->surplus_huge_pages_node[nid]--; nr_pages--; remaining_iterations = num_online_nodes(); } @@ -794,13 +800,14 @@ static void return_unused_surplus_pages(unsigned long unused_resv_pages) * an instantiated the change should be committed via vma_commit_reservation. * No action is required on failure. */ -static int vma_needs_reservation(struct vm_area_struct *vma, unsigned long addr) +static int vma_needs_reservation(struct hstate *h, + struct vm_area_struct *vma, unsigned long addr) { struct address_space *mapping = vma->vm_file->f_mapping; struct inode *inode = mapping->host; if (vma->vm_flags & VM_SHARED) { - pgoff_t idx = vma_hugecache_offset(vma, addr); + pgoff_t idx = vma_hugecache_offset(h, vma, addr); return region_chg(&inode->i_mapping->private_list, idx, idx + 1); @@ -809,7 +816,7 @@ static int vma_needs_reservation(struct vm_area_struct *vma, unsigned long addr) } else { int err; - pgoff_t idx = vma_hugecache_offset(vma, addr); + pgoff_t idx = vma_hugecache_offset(h, vma, addr); struct resv_map *reservations = vma_resv_map(vma); err = region_chg(&reservations->regions, idx, idx + 1); @@ -818,18 +825,18 @@ static int vma_needs_reservation(struct vm_area_struct *vma, unsigned long addr) return 0; } } -static void vma_commit_reservation(struct vm_area_struct *vma, - unsigned long addr) +static void vma_commit_reservation(struct hstate *h, + struct vm_area_struct *vma, unsigned long addr) { struct address_space *mapping = vma->vm_file->f_mapping; struct inode *inode = mapping->host; if (vma->vm_flags & VM_SHARED) { - pgoff_t idx = vma_hugecache_offset(vma, addr); + pgoff_t idx = vma_hugecache_offset(h, vma, addr); region_add(&inode->i_mapping->private_list, idx, idx + 1); } else if (is_vma_resv_set(vma, HPAGE_RESV_OWNER)) { - pgoff_t idx = vma_hugecache_offset(vma, addr); + pgoff_t idx = vma_hugecache_offset(h, vma, addr); struct resv_map *reservations = vma_resv_map(vma); /* Mark this page used in the map. */ @@ -840,6 +847,7 @@ static void vma_commit_reservation(struct vm_area_struct *vma, static struct page *alloc_huge_page(struct vm_area_struct *vma, unsigned long addr, int avoid_reserve) { + struct hstate *h = hstate_vma(vma); struct page *page; struct address_space *mapping = vma->vm_file->f_mapping; struct inode *inode = mapping->host; @@ -852,7 +860,7 @@ static struct page *alloc_huge_page(struct vm_area_struct *vma, * MAP_NORESERVE mappings may also need pages and quota allocated * if no reserve mapping overlaps. */ - chg = vma_needs_reservation(vma, addr); + chg = vma_needs_reservation(h, vma, addr); if (chg < 0) return ERR_PTR(chg); if (chg) @@ -860,11 +868,11 @@ static struct page *alloc_huge_page(struct vm_area_struct *vma, return ERR_PTR(-ENOSPC); spin_lock(&hugetlb_lock); - page = dequeue_huge_page_vma(vma, addr, avoid_reserve); + page = dequeue_huge_page_vma(h, vma, addr, avoid_reserve); spin_unlock(&hugetlb_lock); if (!page) { - page = alloc_buddy_huge_page(vma, addr); + page = alloc_buddy_huge_page(h, vma, addr); if (!page) { hugetlb_put_quota(inode->i_mapping, chg); return ERR_PTR(-VM_FAULT_OOM); @@ -874,7 +882,7 @@ static struct page *alloc_huge_page(struct vm_area_struct *vma, set_page_refcounted(page); set_page_private(page, (unsigned long) mapping); - vma_commit_reservation(vma, addr); + vma_commit_reservation(h, vma, addr); return page; } @@ -882,21 +890,28 @@ static struct page *alloc_huge_page(struct vm_area_struct *vma, static int __init hugetlb_init(void) { unsigned long i; + struct hstate *h = &default_hstate; if (HPAGE_SHIFT == 0) return 0; + if (!h->order) { + h->order = HPAGE_SHIFT - PAGE_SHIFT; + h->mask = HPAGE_MASK; + } + for (i = 0; i < MAX_NUMNODES; ++i) - INIT_LIST_HEAD(&hugepage_freelists[i]); + INIT_LIST_HEAD(&h->hugepage_freelists[i]); - hugetlb_next_nid = first_node(node_online_map); + h->hugetlb_next_nid = first_node(node_online_map); for (i = 0; i < max_huge_pages; ++i) { - if (!alloc_fresh_huge_page()) + if (!alloc_fresh_huge_page(h)) break; } - max_huge_pages = free_huge_pages = nr_huge_pages = i; - printk("Total HugeTLB memory allocated, %ld\n", free_huge_pages); + max_huge_pages = h->free_huge_pages = h->nr_huge_pages = i; + printk(KERN_INFO "Total HugeTLB memory allocated, %ld\n", + h->free_huge_pages); return 0; } module_init(hugetlb_init); @@ -922,34 +937,36 @@ static unsigned int cpuset_mems_nr(unsigned int *array) #ifdef CONFIG_SYSCTL #ifdef CONFIG_HIGHMEM -static void try_to_free_low(unsigned long count) +static void try_to_free_low(struct hstate *h, unsigned long count) { int i; for (i = 0; i < MAX_NUMNODES; ++i) { struct page *page, *next; - list_for_each_entry_safe(page, next, &hugepage_freelists[i], lru) { - if (count >= nr_huge_pages) + struct list_head *freel = &h->hugepage_freelists[i]; + list_for_each_entry_safe(page, next, freel, lru) { + if (count >= h->nr_huge_pages) return; if (PageHighMem(page)) continue; list_del(&page->lru); update_and_free_page(page); - free_huge_pages--; - free_huge_pages_node[page_to_nid(page)]--; + h->free_huge_pages--; + h->free_huge_pages_node[page_to_nid(page)]--; } } } #else -static inline void try_to_free_low(unsigned long count) +static inline void try_to_free_low(struct hstate *h, unsigned long count) { } #endif -#define persistent_huge_pages (nr_huge_pages - surplus_huge_pages) +#define persistent_huge_pages(h) (h->nr_huge_pages - h->surplus_huge_pages) static unsigned long set_max_huge_pages(unsigned long count) { unsigned long min_count, ret; + struct hstate *h = &default_hstate; /* * Increase the pool size @@ -963,19 +980,19 @@ static unsigned long set_max_huge_pages(unsigned long count) * within all the constraints specified by the sysctls. */ spin_lock(&hugetlb_lock); - while (surplus_huge_pages && count > persistent_huge_pages) { - if (!adjust_pool_surplus(-1)) + while (h->surplus_huge_pages && count > persistent_huge_pages(h)) { + if (!adjust_pool_surplus(h, -1)) break; } - while (count > persistent_huge_pages) { + while (count > persistent_huge_pages(h)) { /* * If this allocation races such that we no longer need the * page, free_huge_page will handle it by freeing the page * and reducing the surplus. */ spin_unlock(&hugetlb_lock); - ret = alloc_fresh_huge_page(); + ret = alloc_fresh_huge_page(h); spin_lock(&hugetlb_lock); if (!ret) goto out; @@ -997,21 +1014,21 @@ static unsigned long set_max_huge_pages(unsigned long count) * and won't grow the pool anywhere else. Not until one of the * sysctls are changed, or the surplus pages go out of use. */ - min_count = resv_huge_pages + nr_huge_pages - free_huge_pages; + min_count = h->resv_huge_pages + h->nr_huge_pages - h->free_huge_pages; min_count = max(count, min_count); - try_to_free_low(min_count); - while (min_count < persistent_huge_pages) { - struct page *page = dequeue_huge_page(); + try_to_free_low(h, min_count); + while (min_count < persistent_huge_pages(h)) { + struct page *page = dequeue_huge_page(h); if (!page) break; - update_and_free_page(page); + update_and_free_page(h, page); } - while (count < persistent_huge_pages) { - if (!adjust_pool_surplus(1)) + while (count < persistent_huge_pages(h)) { + if (!adjust_pool_surplus(h, 1)) break; } out: - ret = persistent_huge_pages; + ret = persistent_huge_pages(h); spin_unlock(&hugetlb_lock); return ret; } @@ -1041,9 +1058,10 @@ int hugetlb_overcommit_handler(struct ctl_table *table, int write, struct file *file, void __user *buffer, size_t *length, loff_t *ppos) { + struct hstate *h = &default_hstate; proc_doulongvec_minmax(table, write, file, buffer, length, ppos); spin_lock(&hugetlb_lock); - nr_overcommit_huge_pages = sysctl_overcommit_huge_pages; + h->nr_overcommit_huge_pages = sysctl_overcommit_huge_pages; spin_unlock(&hugetlb_lock); return 0; } @@ -1052,37 +1070,40 @@ int hugetlb_overcommit_handler(struct ctl_table *table, int write, int hugetlb_report_meminfo(char *buf) { + struct hstate *h = &default_hstate; return sprintf(buf, "HugePages_Total: %5lu\n" "HugePages_Free: %5lu\n" "HugePages_Rsvd: %5lu\n" "HugePages_Surp: %5lu\n" "Hugepagesize: %5lu kB\n", - nr_huge_pages, - free_huge_pages, - resv_huge_pages, - surplus_huge_pages, - HPAGE_SIZE/1024); + h->nr_huge_pages, + h->free_huge_pages, + h->resv_huge_pages, + h->surplus_huge_pages, + 1UL << (huge_page_order(h) + PAGE_SHIFT - 10)); } int hugetlb_report_node_meminfo(int nid, char *buf) { + struct hstate *h = &default_hstate; return sprintf(buf, "Node %d HugePages_Total: %5u\n" "Node %d HugePages_Free: %5u\n" "Node %d HugePages_Surp: %5u\n", - nid, nr_huge_pages_node[nid], - nid, free_huge_pages_node[nid], - nid, surplus_huge_pages_node[nid]); + nid, h->nr_huge_pages_node[nid], + nid, h->free_huge_pages_node[nid], + nid, h->surplus_huge_pages_node[nid]); } /* Return the number pages of memory we physically have, in PAGE_SIZE units. */ unsigned long hugetlb_total_pages(void) { - return nr_huge_pages * (HPAGE_SIZE / PAGE_SIZE); + struct hstate *h = &default_hstate; + return h->nr_huge_pages * pages_per_huge_page(h); } -static int hugetlb_acct_memory(long delta) +static int hugetlb_acct_memory(struct hstate *h, long delta) { int ret = -ENOMEM; @@ -1105,18 +1126,18 @@ static int hugetlb_acct_memory(long delta) * semantics that cpuset has. */ if (delta > 0) { - if (gather_surplus_pages(delta) < 0) + if (gather_surplus_pages(h, delta) < 0) goto out; - if (delta > cpuset_mems_nr(free_huge_pages_node)) { - return_unused_surplus_pages(delta); + if (delta > cpuset_mems_nr(h->free_huge_pages_node)) { + return_unused_surplus_pages(h, delta); goto out; } } ret = 0; if (delta < 0) - return_unused_surplus_pages((unsigned long) -delta); + return_unused_surplus_pages(h, (unsigned long) -delta); out: spin_unlock(&hugetlb_lock); @@ -1141,14 +1162,15 @@ static void hugetlb_vm_op_open(struct vm_area_struct *vma) static void hugetlb_vm_op_close(struct vm_area_struct *vma) { + struct hstate *h = hstate_vma(vma); struct resv_map *reservations = vma_resv_map(vma); unsigned long reserve; unsigned long start; unsigned long end; if (reservations) { - start = vma_hugecache_offset(vma, vma->vm_start); - end = vma_hugecache_offset(vma, vma->vm_end); + start = vma_hugecache_offset(h, vma, vma->vm_start); + end = vma_hugecache_offset(h, vma, vma->vm_end); reserve = (end - start) - region_count(&reservations->regions, start, end); @@ -1156,7 +1178,7 @@ static void hugetlb_vm_op_close(struct vm_area_struct *vma) kref_put(&reservations->refs, resv_map_release); if (reserve) - hugetlb_acct_memory(-reserve); + hugetlb_acct_memory(h, -reserve); } } @@ -1214,14 +1236,16 @@ int copy_hugetlb_page_range(struct mm_struct *dst, struct mm_struct *src, struct page *ptepage; unsigned long addr; int cow; + struct hstate *h = hstate_vma(vma); + unsigned long sz = huge_page_size(h); cow = (vma->vm_flags & (VM_SHARED | VM_MAYWRITE)) == VM_MAYWRITE; - for (addr = vma->vm_start; addr < vma->vm_end; addr += HPAGE_SIZE) { + for (addr = vma->vm_start; addr < vma->vm_end; addr += sz) { src_pte = huge_pte_offset(src, addr); if (!src_pte) continue; - dst_pte = huge_pte_alloc(dst, addr); + dst_pte = huge_pte_alloc(dst, addr, sz); if (!dst_pte) goto nomem; @@ -1257,6 +1281,9 @@ void __unmap_hugepage_range(struct vm_area_struct *vma, unsigned long start, pte_t pte; struct page *page; struct page *tmp; + struct hstate *h = hstate_vma(vma); + unsigned long sz = huge_page_size(h); + /* * A page gathering list, protected by per file i_mmap_lock. The * lock is used to avoid list corruption from multiple unmapping @@ -1265,11 +1292,11 @@ void __unmap_hugepage_range(struct vm_area_struct *vma, unsigned long start, LIST_HEAD(page_list); WARN_ON(!is_vm_hugetlb_page(vma)); - BUG_ON(start & ~HPAGE_MASK); - BUG_ON(end & ~HPAGE_MASK); + BUG_ON(start & ~huge_page_mask(h)); + BUG_ON(end & ~huge_page_mask(h)); spin_lock(&mm->page_table_lock); - for (address = start; address < end; address += HPAGE_SIZE) { + for (address = start; address < end; address += sz) { ptep = huge_pte_offset(mm, address); if (!ptep) continue; @@ -1383,6 +1410,7 @@ static int hugetlb_cow(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pte_t *ptep, pte_t pte, struct page *pagecache_page) { + struct hstate *h = hstate_vma(vma); struct page *old_page, *new_page; int avoidcopy; int outside_reserve = 0; @@ -1443,7 +1471,7 @@ retry_avoidcopy: __SetPageUptodate(new_page); spin_lock(&mm->page_table_lock); - ptep = huge_pte_offset(mm, address & HPAGE_MASK); + ptep = huge_pte_offset(mm, address & huge_page_mask(h)); if (likely(pte_same(huge_ptep_get(ptep), pte))) { /* Break COW */ huge_ptep_clear_flush(vma, address, ptep); @@ -1458,14 +1486,14 @@ retry_avoidcopy: } /* Return the pagecache page at a given address within a VMA */ -static struct page *hugetlbfs_pagecache_page(struct vm_area_struct *vma, - unsigned long address) +static struct page *hugetlbfs_pagecache_page(struct hstate *h, + struct vm_area_struct *vma, unsigned long address) { struct address_space *mapping; pgoff_t idx; mapping = vma->vm_file->f_mapping; - idx = vma_hugecache_offset(vma, address); + idx = vma_hugecache_offset(h, vma, address); return find_lock_page(mapping, idx); } @@ -1473,6 +1501,7 @@ static struct page *hugetlbfs_pagecache_page(struct vm_area_struct *vma, static int hugetlb_no_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pte_t *ptep, int write_access) { + struct hstate *h = hstate_vma(vma); int ret = VM_FAULT_SIGBUS; pgoff_t idx; unsigned long size; @@ -1493,7 +1522,7 @@ static int hugetlb_no_page(struct mm_struct *mm, struct vm_area_struct *vma, } mapping = vma->vm_file->f_mapping; - idx = vma_hugecache_offset(vma, address); + idx = vma_hugecache_offset(h, vma, address); /* * Use page lock to guard against racing truncation @@ -1502,7 +1531,7 @@ static int hugetlb_no_page(struct mm_struct *mm, struct vm_area_struct *vma, retry: page = find_lock_page(mapping, idx); if (!page) { - size = i_size_read(mapping->host) >> HPAGE_SHIFT; + size = i_size_read(mapping->host) >> huge_page_shift(h); if (idx >= size) goto out; page = alloc_huge_page(vma, address, 0); @@ -1510,7 +1539,7 @@ retry: ret = -PTR_ERR(page); goto out; } - clear_huge_page(page, address); + clear_huge_page(page, address, huge_page_size(h)); __SetPageUptodate(page); if (vma->vm_flags & VM_SHARED) { @@ -1526,14 +1555,14 @@ retry: } spin_lock(&inode->i_lock); - inode->i_blocks += BLOCKS_PER_HUGEPAGE; + inode->i_blocks += blocks_per_huge_page(h); spin_unlock(&inode->i_lock); } else lock_page(page); } spin_lock(&mm->page_table_lock); - size = i_size_read(mapping->host) >> HPAGE_SHIFT; + size = i_size_read(mapping->host) >> huge_page_shift(h); if (idx >= size) goto backout; @@ -1569,8 +1598,9 @@ int hugetlb_fault(struct mm_struct *mm, struct vm_area_struct *vma, pte_t entry; int ret; static DEFINE_MUTEX(hugetlb_instantiation_mutex); + struct hstate *h = hstate_vma(vma); - ptep = huge_pte_alloc(mm, address); + ptep = huge_pte_alloc(mm, address, huge_page_size(h)); if (!ptep) return VM_FAULT_OOM; @@ -1594,7 +1624,7 @@ int hugetlb_fault(struct mm_struct *mm, struct vm_area_struct *vma, if (likely(pte_same(entry, huge_ptep_get(ptep)))) if (write_access && !pte_write(entry)) { struct page *page; - page = hugetlbfs_pagecache_page(vma, address); + page = hugetlbfs_pagecache_page(h, vma, address); ret = hugetlb_cow(mm, vma, address, ptep, entry, page); if (page) { unlock_page(page); @@ -1615,6 +1645,7 @@ int follow_hugetlb_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long pfn_offset; unsigned long vaddr = *position; int remainder = *length; + struct hstate *h = hstate_vma(vma); spin_lock(&mm->page_table_lock); while (vaddr < vma->vm_end && remainder) { @@ -1626,7 +1657,7 @@ int follow_hugetlb_page(struct mm_struct *mm, struct vm_area_struct *vma, * each hugepage. We have to make * sure we get the * first, for the page indexing below to work. */ - pte = huge_pte_offset(mm, vaddr & HPAGE_MASK); + pte = huge_pte_offset(mm, vaddr & huge_page_mask(h)); if (!pte || huge_pte_none(huge_ptep_get(pte)) || (write && !pte_write(huge_ptep_get(pte)))) { @@ -1644,7 +1675,7 @@ int follow_hugetlb_page(struct mm_struct *mm, struct vm_area_struct *vma, break; } - pfn_offset = (vaddr & ~HPAGE_MASK) >> PAGE_SHIFT; + pfn_offset = (vaddr & ~huge_page_mask(h)) >> PAGE_SHIFT; page = pte_page(huge_ptep_get(pte)); same_page: if (pages) { @@ -1660,7 +1691,7 @@ same_page: --remainder; ++i; if (vaddr < vma->vm_end && remainder && - pfn_offset < HPAGE_SIZE/PAGE_SIZE) { + pfn_offset < pages_per_huge_page(h)) { /* * We use pfn_offset to avoid touching the pageframes * of this compound page. @@ -1682,13 +1713,14 @@ void hugetlb_change_protection(struct vm_area_struct *vma, unsigned long start = address; pte_t *ptep; pte_t pte; + struct hstate *h = hstate_vma(vma); BUG_ON(address >= end); flush_cache_range(vma, address, end); spin_lock(&vma->vm_file->f_mapping->i_mmap_lock); spin_lock(&mm->page_table_lock); - for (; address < end; address += HPAGE_SIZE) { + for (; address < end; address += huge_page_size(h)) { ptep = huge_pte_offset(mm, address); if (!ptep) continue; @@ -1711,6 +1743,7 @@ int hugetlb_reserve_pages(struct inode *inode, struct vm_area_struct *vma) { long ret, chg; + struct hstate *h = hstate_inode(inode); if (vma && vma->vm_flags & VM_NORESERVE) return 0; @@ -1739,7 +1772,7 @@ int hugetlb_reserve_pages(struct inode *inode, if (hugetlb_get_quota(inode->i_mapping, chg)) return -ENOSPC; - ret = hugetlb_acct_memory(chg); + ret = hugetlb_acct_memory(h, chg); if (ret < 0) { hugetlb_put_quota(inode->i_mapping, chg); return ret; @@ -1751,12 +1784,13 @@ int hugetlb_reserve_pages(struct inode *inode, void hugetlb_unreserve_pages(struct inode *inode, long offset, long freed) { + struct hstate *h = hstate_inode(inode); long chg = region_truncate(&inode->i_mapping->private_list, offset); spin_lock(&inode->i_lock); - inode->i_blocks -= BLOCKS_PER_HUGEPAGE * freed; + inode->i_blocks -= blocks_per_huge_page(h); spin_unlock(&inode->i_lock); hugetlb_put_quota(inode->i_mapping, (chg - freed)); - hugetlb_acct_memory(-(chg - freed)); + hugetlb_acct_memory(h, -(chg - freed)); } diff --git a/mm/memory.c b/mm/memory.c index 72932489a08..c1c1d6d8c22 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -903,7 +903,7 @@ unsigned long unmap_vmas(struct mmu_gather **tlbp, if (unlikely(is_vm_hugetlb_page(vma))) { unmap_hugepage_range(vma, start, end, NULL); zap_work -= (end - start) / - (HPAGE_SIZE / PAGE_SIZE); + pages_per_huge_page(hstate_vma(vma)); start = end; } else start = unmap_page_range(*tlbp, vma, diff --git a/mm/mempolicy.c b/mm/mempolicy.c index c94e58b192c..e550bec2058 100644 --- a/mm/mempolicy.c +++ b/mm/mempolicy.c @@ -1481,7 +1481,7 @@ struct zonelist *huge_zonelist(struct vm_area_struct *vma, unsigned long addr, if (unlikely((*mpol)->mode == MPOL_INTERLEAVE)) { zl = node_zonelist(interleave_nid(*mpol, vma, addr, - HPAGE_SHIFT), gfp_flags); + huge_page_shift(hstate_vma(vma))), gfp_flags); } else { zl = policy_zonelist(gfp_flags, *mpol); if ((*mpol)->mode == MPOL_BIND) @@ -2220,9 +2220,12 @@ static void check_huge_range(struct vm_area_struct *vma, { unsigned long addr; struct page *page; + struct hstate *h = hstate_vma(vma); + unsigned long sz = huge_page_size(h); - for (addr = start; addr < end; addr += HPAGE_SIZE) { - pte_t *ptep = huge_pte_offset(vma->vm_mm, addr & HPAGE_MASK); + for (addr = start; addr < end; addr += sz) { + pte_t *ptep = huge_pte_offset(vma->vm_mm, + addr & huge_page_mask(h)); pte_t pte; if (!ptep) diff --git a/mm/mmap.c b/mm/mmap.c index 57d3b6097de..5e0cc99e9cd 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -1812,7 +1812,8 @@ int split_vma(struct mm_struct * mm, struct vm_area_struct * vma, struct mempolicy *pol; struct vm_area_struct *new; - if (is_vm_hugetlb_page(vma) && (addr & ~HPAGE_MASK)) + if (is_vm_hugetlb_page(vma) && (addr & + ~(huge_page_mask(hstate_vma(vma))))) return -EINVAL; if (mm->map_count >= sysctl_max_map_count) -- cgit v1.2.3-70-g09d2 From a137e1cc6d6e7d315fef03962a2a5a113348b13b Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Wed, 23 Jul 2008 21:27:43 -0700 Subject: hugetlbfs: per mount huge page sizes Add the ability to configure the hugetlb hstate used on a per mount basis. - Add a new pagesize= option to the hugetlbfs mount that allows setting the page size - This option causes the mount code to find the hstate corresponding to the specified size, and sets up a pointer to the hstate in the mount's superblock. - Change the hstate accessors to use this information rather than the global_hstate they were using (requires a slight change in mm/memory.c so we don't NULL deref in the error-unmap path -- see comments). [np: take hstate out of hugetlbfs inode and vma->vm_private_data] Acked-by: Adam Litke Acked-by: Nishanth Aravamudan Signed-off-by: Andi Kleen Signed-off-by: Nick Piggin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/hugetlbfs/inode.c | 45 ++++++++++++++++++++++++++++++++++++--------- include/linux/hugetlb.h | 14 +++++++++----- mm/hugetlb.c | 16 +++------------- mm/memory.c | 18 ++++++++++++++++-- 4 files changed, 64 insertions(+), 29 deletions(-) (limited to 'fs') diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c index 516c581b537..dbd01d262ca 100644 --- a/fs/hugetlbfs/inode.c +++ b/fs/hugetlbfs/inode.c @@ -53,6 +53,7 @@ int sysctl_hugetlb_shm_group; enum { Opt_size, Opt_nr_inodes, Opt_mode, Opt_uid, Opt_gid, + Opt_pagesize, Opt_err, }; @@ -62,6 +63,7 @@ static match_table_t tokens = { {Opt_mode, "mode=%o"}, {Opt_uid, "uid=%u"}, {Opt_gid, "gid=%u"}, + {Opt_pagesize, "pagesize=%s"}, {Opt_err, NULL}, }; @@ -750,6 +752,8 @@ hugetlbfs_parse_options(char *options, struct hugetlbfs_config *pconfig) char *p, *rest; substring_t args[MAX_OPT_ARGS]; int option; + unsigned long long size = 0; + enum { NO_SIZE, SIZE_STD, SIZE_PERCENT } setsize = NO_SIZE; if (!options) return 0; @@ -780,17 +784,13 @@ hugetlbfs_parse_options(char *options, struct hugetlbfs_config *pconfig) break; case Opt_size: { - unsigned long long size; /* memparse() will accept a K/M/G without a digit */ if (!isdigit(*args[0].from)) goto bad_val; size = memparse(args[0].from, &rest); - if (*rest == '%') { - size <<= HPAGE_SHIFT; - size *= max_huge_pages; - do_div(size, 100); - } - pconfig->nr_blocks = (size >> HPAGE_SHIFT); + setsize = SIZE_STD; + if (*rest == '%') + setsize = SIZE_PERCENT; break; } @@ -801,6 +801,19 @@ hugetlbfs_parse_options(char *options, struct hugetlbfs_config *pconfig) pconfig->nr_inodes = memparse(args[0].from, &rest); break; + case Opt_pagesize: { + unsigned long ps; + ps = memparse(args[0].from, &rest); + pconfig->hstate = size_to_hstate(ps); + if (!pconfig->hstate) { + printk(KERN_ERR + "hugetlbfs: Unsupported page size %lu MB\n", + ps >> 20); + return -EINVAL; + } + break; + } + default: printk(KERN_ERR "hugetlbfs: Bad mount option: \"%s\"\n", p); @@ -808,6 +821,18 @@ hugetlbfs_parse_options(char *options, struct hugetlbfs_config *pconfig) break; } } + + /* Do size after hstate is set up */ + if (setsize > NO_SIZE) { + struct hstate *h = pconfig->hstate; + if (setsize == SIZE_PERCENT) { + size <<= huge_page_shift(h); + size *= h->max_huge_pages; + do_div(size, 100); + } + pconfig->nr_blocks = (size >> huge_page_shift(h)); + } + return 0; bad_val: @@ -832,6 +857,7 @@ hugetlbfs_fill_super(struct super_block *sb, void *data, int silent) config.uid = current->fsuid; config.gid = current->fsgid; config.mode = 0755; + config.hstate = &default_hstate; ret = hugetlbfs_parse_options(data, &config); if (ret) return ret; @@ -840,14 +866,15 @@ hugetlbfs_fill_super(struct super_block *sb, void *data, int silent) if (!sbinfo) return -ENOMEM; sb->s_fs_info = sbinfo; + sbinfo->hstate = config.hstate; spin_lock_init(&sbinfo->stat_lock); sbinfo->max_blocks = config.nr_blocks; sbinfo->free_blocks = config.nr_blocks; sbinfo->max_inodes = config.nr_inodes; sbinfo->free_inodes = config.nr_inodes; sb->s_maxbytes = MAX_LFS_FILESIZE; - sb->s_blocksize = HPAGE_SIZE; - sb->s_blocksize_bits = HPAGE_SHIFT; + sb->s_blocksize = huge_page_size(config.hstate); + sb->s_blocksize_bits = huge_page_shift(config.hstate); sb->s_magic = HUGETLBFS_MAGIC; sb->s_op = &hugetlbfs_ops; sb->s_time_gran = 1; diff --git a/include/linux/hugetlb.h b/include/linux/hugetlb.h index b75bdb4deba..ba9263e631b 100644 --- a/include/linux/hugetlb.h +++ b/include/linux/hugetlb.h @@ -100,6 +100,7 @@ struct hugetlbfs_config { umode_t mode; long nr_blocks; long nr_inodes; + struct hstate *hstate; }; struct hugetlbfs_sb_info { @@ -108,6 +109,7 @@ struct hugetlbfs_sb_info { long max_inodes; /* inodes allowed */ long free_inodes; /* inodes free */ spinlock_t stat_lock; + struct hstate *hstate; }; @@ -191,19 +193,21 @@ extern unsigned int default_hstate_idx; #define default_hstate (hstates[default_hstate_idx]) -static inline struct hstate *hstate_vma(struct vm_area_struct *vma) +static inline struct hstate *hstate_inode(struct inode *i) { - return &default_hstate; + struct hugetlbfs_sb_info *hsb; + hsb = HUGETLBFS_SB(i->i_sb); + return hsb->hstate; } static inline struct hstate *hstate_file(struct file *f) { - return &default_hstate; + return hstate_inode(f->f_dentry->d_inode); } -static inline struct hstate *hstate_inode(struct inode *i) +static inline struct hstate *hstate_vma(struct vm_area_struct *vma) { - return &default_hstate; + return hstate_file(vma->vm_file); } static inline unsigned long huge_page_size(struct hstate *h) diff --git a/mm/hugetlb.c b/mm/hugetlb.c index 82378d44a0c..4cf7a90e914 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -1439,19 +1439,9 @@ void __unmap_hugepage_range(struct vm_area_struct *vma, unsigned long start, void unmap_hugepage_range(struct vm_area_struct *vma, unsigned long start, unsigned long end, struct page *ref_page) { - /* - * It is undesirable to test vma->vm_file as it should be non-null - * for valid hugetlb area. However, vm_file will be NULL in the error - * cleanup path of do_mmap_pgoff. When hugetlbfs ->mmap method fails, - * do_mmap_pgoff() nullifies vma->vm_file before calling this function - * to clean up. Since no pte has actually been setup, it is safe to - * do nothing in this case. - */ - if (vma->vm_file) { - spin_lock(&vma->vm_file->f_mapping->i_mmap_lock); - __unmap_hugepage_range(vma, start, end, ref_page); - spin_unlock(&vma->vm_file->f_mapping->i_mmap_lock); - } + spin_lock(&vma->vm_file->f_mapping->i_mmap_lock); + __unmap_hugepage_range(vma, start, end, ref_page); + spin_unlock(&vma->vm_file->f_mapping->i_mmap_lock); } /* diff --git a/mm/memory.c b/mm/memory.c index c1c1d6d8c22..02fc6b1047b 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -901,9 +901,23 @@ unsigned long unmap_vmas(struct mmu_gather **tlbp, } if (unlikely(is_vm_hugetlb_page(vma))) { - unmap_hugepage_range(vma, start, end, NULL); - zap_work -= (end - start) / + /* + * It is undesirable to test vma->vm_file as it + * should be non-null for valid hugetlb area. + * However, vm_file will be NULL in the error + * cleanup path of do_mmap_pgoff. When + * hugetlbfs ->mmap method fails, + * do_mmap_pgoff() nullifies vma->vm_file + * before calling this function to clean up. + * Since no pte has actually been setup, it is + * safe to do nothing in this case. + */ + if (vma->vm_file) { + unmap_hugepage_range(vma, start, end, NULL); + zap_work -= (end - start) / pages_per_huge_page(hstate_vma(vma)); + } + start = end; } else start = unmap_page_range(*tlbp, vma, -- cgit v1.2.3-70-g09d2 From f4a67cceee4a6f5ed38011a698c9e34747270ae5 Mon Sep 17 00:00:00 2001 From: Jon Tollefson Date: Wed, 23 Jul 2008 21:27:55 -0700 Subject: fs: check for statfs overflow Adds a check for an overflow in the filesystem size so if someone is checking with statfs() on a 16G blocksize hugetlbfs in a 32bit binary that it will report back EOVERFLOW instead of a size of 0. Acked-by: Nishanth Aravamudan Signed-off-by: Jon Tollefson Signed-off-by: Nick Piggin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/compat.c | 8 ++++---- fs/open.c | 3 ++- 2 files changed, 6 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/compat.c b/fs/compat.c index ed43e17a5dc..b4660428176 100644 --- a/fs/compat.c +++ b/fs/compat.c @@ -197,8 +197,8 @@ static int put_compat_statfs(struct compat_statfs __user *ubuf, struct kstatfs * { if (sizeof ubuf->f_blocks == 4) { - if ((kbuf->f_blocks | kbuf->f_bfree | kbuf->f_bavail) & - 0xffffffff00000000ULL) + if ((kbuf->f_blocks | kbuf->f_bfree | kbuf->f_bavail | + kbuf->f_bsize | kbuf->f_frsize) & 0xffffffff00000000ULL) return -EOVERFLOW; /* f_files and f_ffree may be -1; it's okay * to stuff that into 32 bits */ @@ -271,8 +271,8 @@ out: static int put_compat_statfs64(struct compat_statfs64 __user *ubuf, struct kstatfs *kbuf) { if (sizeof ubuf->f_blocks == 4) { - if ((kbuf->f_blocks | kbuf->f_bfree | kbuf->f_bavail) & - 0xffffffff00000000ULL) + if ((kbuf->f_blocks | kbuf->f_bfree | kbuf->f_bavail | + kbuf->f_bsize | kbuf->f_frsize) & 0xffffffff00000000ULL) return -EOVERFLOW; /* f_files and f_ffree may be -1; it's okay * to stuff that into 32 bits */ diff --git a/fs/open.c b/fs/open.c index a99ad09c319..bb98d2fe809 100644 --- a/fs/open.c +++ b/fs/open.c @@ -64,7 +64,8 @@ static int vfs_statfs_native(struct dentry *dentry, struct statfs *buf) memcpy(buf, &st, sizeof(st)); else { if (sizeof buf->f_blocks == 4) { - if ((st.f_blocks | st.f_bfree | st.f_bavail) & + if ((st.f_blocks | st.f_bfree | st.f_bavail | + st.f_bsize | st.f_frsize) & 0xffffffff00000000ULL) return -EOVERFLOW; /* -- cgit v1.2.3-70-g09d2 From 6e2c10a12a2170856f5582d62d583cbcd1cb5eaf Mon Sep 17 00:00:00 2001 From: Akinobu Mita Date: Wed, 23 Jul 2008 21:29:15 -0700 Subject: binfmt_misc: use simple_read_from_buffer() Signed-off-by: Akinobu Mita Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/binfmt_misc.c | 20 +++----------------- 1 file changed, 3 insertions(+), 17 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_misc.c b/fs/binfmt_misc.c index 7191306367c..756205314c2 100644 --- a/fs/binfmt_misc.c +++ b/fs/binfmt_misc.c @@ -27,6 +27,7 @@ #include #include #include +#include #include @@ -535,31 +536,16 @@ static ssize_t bm_entry_read(struct file * file, char __user * buf, size_t nbytes, loff_t *ppos) { Node *e = file->f_path.dentry->d_inode->i_private; - loff_t pos = *ppos; ssize_t res; char *page; - int len; if (!(page = (char*) __get_free_page(GFP_KERNEL))) return -ENOMEM; entry_status(e, page); - len = strlen(page); - res = -EINVAL; - if (pos < 0) - goto out; - res = 0; - if (pos >= len) - goto out; - if (len < pos + nbytes) - nbytes = len - pos; - res = -EFAULT; - if (copy_to_user(buf, page + pos, nbytes)) - goto out; - *ppos = pos + nbytes; - res = nbytes; -out: + res = simple_read_from_buffer(buf, nbytes, ppos, page, strlen(page)); + free_page((unsigned long) page); return res; } -- cgit v1.2.3-70-g09d2 From 7d9dbca34240ebb6ff88d8a29c6c7bffd098f0c1 Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Wed, 23 Jul 2008 21:29:22 -0700 Subject: flag parameters: anon_inode_getfd extension This patch just extends the anon_inode_getfd interface to take an additional parameter with a flag value. The flag value is passed on to get_unused_fd_flags in anticipation for a use with the O_CLOEXEC flag. No actual semantic changes here, the changed callers all pass 0 for now. [akpm@linux-foundation.org: KVM fix] Signed-off-by: Ulrich Drepper Acked-by: Davide Libenzi Cc: Michael Kerrisk Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/anon_inodes.c | 9 +++++---- fs/eventfd.c | 2 +- fs/eventpoll.c | 2 +- fs/signalfd.c | 3 ++- fs/timerfd.c | 2 +- include/linux/anon_inodes.h | 2 +- virt/kvm/kvm_main.c | 4 ++-- 7 files changed, 13 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/anon_inodes.c b/fs/anon_inodes.c index 977ef208c05..1a4eee620b0 100644 --- a/fs/anon_inodes.c +++ b/fs/anon_inodes.c @@ -58,8 +58,9 @@ static struct dentry_operations anon_inodefs_dentry_operations = { * of the file * * @name: [in] name of the "class" of the new file - * @fops [in] file operations for the new file - * @priv [in] private data for the new file (will be file's private_data) + * @fops: [in] file operations for the new file + * @priv: [in] private data for the new file (will be file's private_data) + * @flags: [in] flags * * Creates a new file by hooking it on a single inode. This is useful for files * that do not need to have a full-fledged inode in order to operate correctly. @@ -68,7 +69,7 @@ static struct dentry_operations anon_inodefs_dentry_operations = { * setup. Returns new descriptor or -error. */ int anon_inode_getfd(const char *name, const struct file_operations *fops, - void *priv) + void *priv, int flags) { struct qstr this; struct dentry *dentry; @@ -78,7 +79,7 @@ int anon_inode_getfd(const char *name, const struct file_operations *fops, if (IS_ERR(anon_inode_inode)) return -ENODEV; - error = get_unused_fd(); + error = get_unused_fd_flags(flags); if (error < 0) return error; fd = error; diff --git a/fs/eventfd.c b/fs/eventfd.c index 343942deeec..6094265ca40 100644 --- a/fs/eventfd.c +++ b/fs/eventfd.c @@ -214,7 +214,7 @@ asmlinkage long sys_eventfd(unsigned int count) * When we call this, the initialization must be complete, since * anon_inode_getfd() will install the fd. */ - fd = anon_inode_getfd("[eventfd]", &eventfd_fops, ctx); + fd = anon_inode_getfd("[eventfd]", &eventfd_fops, ctx, 0); if (fd < 0) kfree(ctx); return fd; diff --git a/fs/eventpoll.c b/fs/eventpoll.c index 990c01d2d66..9392dd96812 100644 --- a/fs/eventpoll.c +++ b/fs/eventpoll.c @@ -1068,7 +1068,7 @@ asmlinkage long sys_epoll_create(int size) * Creates all the items needed to setup an eventpoll file. That is, * a file structure and a free file descriptor. */ - fd = anon_inode_getfd("[eventpoll]", &eventpoll_fops, ep); + fd = anon_inode_getfd("[eventpoll]", &eventpoll_fops, ep, 0); if (fd < 0) ep_free(ep); diff --git a/fs/signalfd.c b/fs/signalfd.c index 619725644c7..ddb328b74bd 100644 --- a/fs/signalfd.c +++ b/fs/signalfd.c @@ -227,7 +227,8 @@ asmlinkage long sys_signalfd(int ufd, sigset_t __user *user_mask, size_t sizemas * When we call this, the initialization must be complete, since * anon_inode_getfd() will install the fd. */ - ufd = anon_inode_getfd("[signalfd]", &signalfd_fops, ctx); + ufd = anon_inode_getfd("[signalfd]", &signalfd_fops, ctx, + 0); if (ufd < 0) kfree(ctx); } else { diff --git a/fs/timerfd.c b/fs/timerfd.c index d87d354ec42..77c2bc92cbe 100644 --- a/fs/timerfd.c +++ b/fs/timerfd.c @@ -198,7 +198,7 @@ asmlinkage long sys_timerfd_create(int clockid, int flags) ctx->clockid = clockid; hrtimer_init(&ctx->tmr, clockid, HRTIMER_MODE_ABS); - ufd = anon_inode_getfd("[timerfd]", &timerfd_fops, ctx); + ufd = anon_inode_getfd("[timerfd]", &timerfd_fops, ctx, 0); if (ufd < 0) kfree(ctx); diff --git a/include/linux/anon_inodes.h b/include/linux/anon_inodes.h index 6129e58ca7c..e0a0cdc2da4 100644 --- a/include/linux/anon_inodes.h +++ b/include/linux/anon_inodes.h @@ -9,7 +9,7 @@ #define _LINUX_ANON_INODES_H int anon_inode_getfd(const char *name, const struct file_operations *fops, - void *priv); + void *priv, int flags); #endif /* _LINUX_ANON_INODES_H */ diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index 904d7b7bd78..a845890b680 100644 --- a/virt/kvm/kvm_main.c +++ b/virt/kvm/kvm_main.c @@ -902,7 +902,7 @@ static const struct file_operations kvm_vcpu_fops = { */ static int create_vcpu_fd(struct kvm_vcpu *vcpu) { - int fd = anon_inode_getfd("kvm-vcpu", &kvm_vcpu_fops, vcpu); + int fd = anon_inode_getfd("kvm-vcpu", &kvm_vcpu_fops, vcpu, 0); if (fd < 0) kvm_put_kvm(vcpu->kvm); return fd; @@ -1261,7 +1261,7 @@ static int kvm_dev_ioctl_create_vm(void) kvm = kvm_create_vm(); if (IS_ERR(kvm)) return PTR_ERR(kvm); - fd = anon_inode_getfd("kvm-vm", &kvm_vm_fops, kvm); + fd = anon_inode_getfd("kvm-vm", &kvm_vm_fops, kvm, 0); if (fd < 0) kvm_put_kvm(kvm); -- cgit v1.2.3-70-g09d2 From 9deb27baedb79759c3ab9435a7d8b841842d56e9 Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Wed, 23 Jul 2008 21:29:24 -0700 Subject: flag parameters: signalfd This patch adds the new signalfd4 syscall. It extends the old signalfd syscall by one parameter which is meant to hold a flag value. In this patch the only flag support is SFD_CLOEXEC which causes the close-on-exec flag for the returned file descriptor to be set. A new name SFD_CLOEXEC is introduced which in this implementation must have the same value as O_CLOEXEC. The following test must be adjusted for architectures other than x86 and x86-64 and in case the syscall numbers changed. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ #include #include #include #include #include #ifndef __NR_signalfd4 # ifdef __x86_64__ # define __NR_signalfd4 289 # elif defined __i386__ # define __NR_signalfd4 327 # else # error "need __NR_signalfd4" # endif #endif #define SFD_CLOEXEC O_CLOEXEC int main (void) { sigset_t ss; sigemptyset (&ss); sigaddset (&ss, SIGUSR1); int fd = syscall (__NR_signalfd4, -1, &ss, 8, 0); if (fd == -1) { puts ("signalfd4(0) failed"); return 1; } int coe = fcntl (fd, F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if (coe & FD_CLOEXEC) { puts ("signalfd4(0) set close-on-exec flag"); return 1; } close (fd); fd = syscall (__NR_signalfd4, -1, &ss, 8, SFD_CLOEXEC); if (fd == -1) { puts ("signalfd4(SFD_CLOEXEC) failed"); return 1; } coe = fcntl (fd, F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if ((coe & FD_CLOEXEC) == 0) { puts ("signalfd4(SFD_CLOEXEC) does not set close-on-exec flag"); return 1; } close (fd); puts ("OK"); return 0; } ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ [akpm@linux-foundation.org: add sys_ni stub] Signed-off-by: Ulrich Drepper Acked-by: Davide Libenzi Cc: Michael Kerrisk Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/x86/ia32/ia32entry.S | 1 + arch/x86/kernel/syscall_table_32.S | 1 + fs/compat.c | 14 ++++++++++---- fs/signalfd.c | 14 ++++++++++++-- include/asm-x86/unistd_32.h | 1 + include/asm-x86/unistd_64.h | 2 ++ include/linux/signalfd.h | 5 +++++ include/linux/syscalls.h | 1 + kernel/sys_ni.c | 1 + 9 files changed, 34 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/arch/x86/ia32/ia32entry.S b/arch/x86/ia32/ia32entry.S index 021d71bc69b..c308128b925 100644 --- a/arch/x86/ia32/ia32entry.S +++ b/arch/x86/ia32/ia32entry.S @@ -826,4 +826,5 @@ ia32_sys_call_table: .quad sys32_fallocate .quad compat_sys_timerfd_settime /* 325 */ .quad compat_sys_timerfd_gettime + .quad compat_sys_signalfd4 ia32_syscall_end: diff --git a/arch/x86/kernel/syscall_table_32.S b/arch/x86/kernel/syscall_table_32.S index adff5562f5f..c12a36c9fd5 100644 --- a/arch/x86/kernel/syscall_table_32.S +++ b/arch/x86/kernel/syscall_table_32.S @@ -326,3 +326,4 @@ ENTRY(sys_call_table) .long sys_fallocate .long sys_timerfd_settime /* 325 */ .long sys_timerfd_gettime + .long sys_signalfd4 diff --git a/fs/compat.c b/fs/compat.c index b4660428176..106eba28ec5 100644 --- a/fs/compat.c +++ b/fs/compat.c @@ -2131,9 +2131,9 @@ asmlinkage long compat_sys_epoll_pwait(int epfd, #ifdef CONFIG_SIGNALFD -asmlinkage long compat_sys_signalfd(int ufd, - const compat_sigset_t __user *sigmask, - compat_size_t sigsetsize) +asmlinkage long compat_sys_signalfd4(int ufd, + const compat_sigset_t __user *sigmask, + compat_size_t sigsetsize, int flags) { compat_sigset_t ss32; sigset_t tmp; @@ -2148,9 +2148,15 @@ asmlinkage long compat_sys_signalfd(int ufd, if (copy_to_user(ksigmask, &tmp, sizeof(sigset_t))) return -EFAULT; - return sys_signalfd(ufd, ksigmask, sizeof(sigset_t)); + return sys_signalfd4(ufd, ksigmask, sizeof(sigset_t), flags); } +asmlinkage long compat_sys_signalfd(int ufd, + const compat_sigset_t __user *sigmask, + compat_size_t sigsetsize) +{ + return compat_sys_signalfd4(ufd, sigmask, sigsetsize, 0); +} #endif /* CONFIG_SIGNALFD */ #ifdef CONFIG_TIMERFD diff --git a/fs/signalfd.c b/fs/signalfd.c index ddb328b74bd..c8609fa51a1 100644 --- a/fs/signalfd.c +++ b/fs/signalfd.c @@ -205,11 +205,15 @@ static const struct file_operations signalfd_fops = { .read = signalfd_read, }; -asmlinkage long sys_signalfd(int ufd, sigset_t __user *user_mask, size_t sizemask) +asmlinkage long sys_signalfd4(int ufd, sigset_t __user *user_mask, + size_t sizemask, int flags) { sigset_t sigmask; struct signalfd_ctx *ctx; + if (flags & ~SFD_CLOEXEC) + return -EINVAL; + if (sizemask != sizeof(sigset_t) || copy_from_user(&sigmask, user_mask, sizeof(sigmask))) return -EINVAL; @@ -228,7 +232,7 @@ asmlinkage long sys_signalfd(int ufd, sigset_t __user *user_mask, size_t sizemas * anon_inode_getfd() will install the fd. */ ufd = anon_inode_getfd("[signalfd]", &signalfd_fops, ctx, - 0); + flags & O_CLOEXEC); if (ufd < 0) kfree(ctx); } else { @@ -250,3 +254,9 @@ asmlinkage long sys_signalfd(int ufd, sigset_t __user *user_mask, size_t sizemas return ufd; } + +asmlinkage long sys_signalfd(int ufd, sigset_t __user *user_mask, + size_t sizemask) +{ + return sys_signalfd4(ufd, user_mask, sizemask, 0); +} diff --git a/include/asm-x86/unistd_32.h b/include/asm-x86/unistd_32.h index 8317d94771d..c310371f561 100644 --- a/include/asm-x86/unistd_32.h +++ b/include/asm-x86/unistd_32.h @@ -332,6 +332,7 @@ #define __NR_fallocate 324 #define __NR_timerfd_settime 325 #define __NR_timerfd_gettime 326 +#define __NR_signalfd4 327 #ifdef __KERNEL__ diff --git a/include/asm-x86/unistd_64.h b/include/asm-x86/unistd_64.h index e323994a370..e0a9b45b234 100644 --- a/include/asm-x86/unistd_64.h +++ b/include/asm-x86/unistd_64.h @@ -641,6 +641,8 @@ __SYSCALL(__NR_timerfd_settime, sys_timerfd_settime) __SYSCALL(__NR_timerfd_gettime, sys_timerfd_gettime) #define __NR_paccept 288 __SYSCALL(__NR_paccept, sys_paccept) +#define __NR_signalfd4 289 +__SYSCALL(__NR_signalfd4, sys_signalfd4) #ifndef __NO_STUBS diff --git a/include/linux/signalfd.h b/include/linux/signalfd.h index ea037f28df9..8b3f7b7420a 100644 --- a/include/linux/signalfd.h +++ b/include/linux/signalfd.h @@ -8,6 +8,11 @@ #ifndef _LINUX_SIGNALFD_H #define _LINUX_SIGNALFD_H +/* For O_CLOEXEC */ +#include + +/* Flags for signalfd4. */ +#define SFD_CLOEXEC O_CLOEXEC struct signalfd_siginfo { __u32 ssi_signo; diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h index 2a2a40af6b2..1c270779784 100644 --- a/include/linux/syscalls.h +++ b/include/linux/syscalls.h @@ -610,6 +610,7 @@ asmlinkage long sys_set_robust_list(struct robust_list_head __user *head, size_t len); asmlinkage long sys_getcpu(unsigned __user *cpu, unsigned __user *node, struct getcpu_cache __user *cache); asmlinkage long sys_signalfd(int ufd, sigset_t __user *user_mask, size_t sizemask); +asmlinkage long sys_signalfd4(int ufd, sigset_t __user *user_mask, size_t sizemask, int flags); asmlinkage long sys_timerfd_create(int clockid, int flags); asmlinkage long sys_timerfd_settime(int ufd, int flags, const struct itimerspec __user *utmr, diff --git a/kernel/sys_ni.c b/kernel/sys_ni.c index 2f0b8a2e600..8627c89ae9e 100644 --- a/kernel/sys_ni.c +++ b/kernel/sys_ni.c @@ -156,6 +156,7 @@ cond_syscall(sys_ioprio_get); /* New file descriptors */ cond_syscall(sys_signalfd); +cond_syscall(sys_signalfd4); cond_syscall(compat_sys_signalfd); cond_syscall(sys_timerfd_create); cond_syscall(sys_timerfd_settime); -- cgit v1.2.3-70-g09d2 From b087498eb5605673b0f260a7620d91818cd72304 Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Wed, 23 Jul 2008 21:29:25 -0700 Subject: flag parameters: eventfd This patch adds the new eventfd2 syscall. It extends the old eventfd syscall by one parameter which is meant to hold a flag value. In this patch the only flag support is EFD_CLOEXEC which causes the close-on-exec flag for the returned file descriptor to be set. A new name EFD_CLOEXEC is introduced which in this implementation must have the same value as O_CLOEXEC. The following test must be adjusted for architectures other than x86 and x86-64 and in case the syscall numbers changed. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ #include #include #include #include #ifndef __NR_eventfd2 # ifdef __x86_64__ # define __NR_eventfd2 290 # elif defined __i386__ # define __NR_eventfd2 328 # else # error "need __NR_eventfd2" # endif #endif #define EFD_CLOEXEC O_CLOEXEC int main (void) { int fd = syscall (__NR_eventfd2, 1, 0); if (fd == -1) { puts ("eventfd2(0) failed"); return 1; } int coe = fcntl (fd, F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if (coe & FD_CLOEXEC) { puts ("eventfd2(0) sets close-on-exec flag"); return 1; } close (fd); fd = syscall (__NR_eventfd2, 1, EFD_CLOEXEC); if (fd == -1) { puts ("eventfd2(EFD_CLOEXEC) failed"); return 1; } coe = fcntl (fd, F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if ((coe & FD_CLOEXEC) == 0) { puts ("eventfd2(EFD_CLOEXEC) does not set close-on-exec flag"); return 1; } close (fd); puts ("OK"); return 0; } ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ [akpm@linux-foundation.org: add sys_ni stub] Signed-off-by: Ulrich Drepper Acked-by: Davide Libenzi Cc: Michael Kerrisk Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/x86/ia32/ia32entry.S | 1 + arch/x86/kernel/syscall_table_32.S | 1 + fs/eventfd.c | 13 +++++++++++-- include/asm-x86/unistd_32.h | 1 + include/asm-x86/unistd_64.h | 2 ++ include/linux/eventfd.h | 6 ++++++ include/linux/syscalls.h | 1 + kernel/sys_ni.c | 1 + 8 files changed, 24 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/arch/x86/ia32/ia32entry.S b/arch/x86/ia32/ia32entry.S index c308128b925..cf0eb31745c 100644 --- a/arch/x86/ia32/ia32entry.S +++ b/arch/x86/ia32/ia32entry.S @@ -827,4 +827,5 @@ ia32_sys_call_table: .quad compat_sys_timerfd_settime /* 325 */ .quad compat_sys_timerfd_gettime .quad compat_sys_signalfd4 + .quad sys_eventfd2 ia32_syscall_end: diff --git a/arch/x86/kernel/syscall_table_32.S b/arch/x86/kernel/syscall_table_32.S index c12a36c9fd5..cf112cb11c3 100644 --- a/arch/x86/kernel/syscall_table_32.S +++ b/arch/x86/kernel/syscall_table_32.S @@ -327,3 +327,4 @@ ENTRY(sys_call_table) .long sys_timerfd_settime /* 325 */ .long sys_timerfd_gettime .long sys_signalfd4 + .long sys_eventfd2 diff --git a/fs/eventfd.c b/fs/eventfd.c index 6094265ca40..bd420e6478a 100644 --- a/fs/eventfd.c +++ b/fs/eventfd.c @@ -198,11 +198,14 @@ struct file *eventfd_fget(int fd) return file; } -asmlinkage long sys_eventfd(unsigned int count) +asmlinkage long sys_eventfd2(unsigned int count, int flags) { int fd; struct eventfd_ctx *ctx; + if (flags & ~EFD_CLOEXEC) + return -EINVAL; + ctx = kmalloc(sizeof(*ctx), GFP_KERNEL); if (!ctx) return -ENOMEM; @@ -214,9 +217,15 @@ asmlinkage long sys_eventfd(unsigned int count) * When we call this, the initialization must be complete, since * anon_inode_getfd() will install the fd. */ - fd = anon_inode_getfd("[eventfd]", &eventfd_fops, ctx, 0); + fd = anon_inode_getfd("[eventfd]", &eventfd_fops, ctx, + flags & O_CLOEXEC); if (fd < 0) kfree(ctx); return fd; } +asmlinkage long sys_eventfd(unsigned int count) +{ + return sys_eventfd2(count, 0); +} + diff --git a/include/asm-x86/unistd_32.h b/include/asm-x86/unistd_32.h index c310371f561..edbd8723c93 100644 --- a/include/asm-x86/unistd_32.h +++ b/include/asm-x86/unistd_32.h @@ -333,6 +333,7 @@ #define __NR_timerfd_settime 325 #define __NR_timerfd_gettime 326 #define __NR_signalfd4 327 +#define __NR_eventfd2 328 #ifdef __KERNEL__ diff --git a/include/asm-x86/unistd_64.h b/include/asm-x86/unistd_64.h index e0a9b45b234..fb059a6feeb 100644 --- a/include/asm-x86/unistd_64.h +++ b/include/asm-x86/unistd_64.h @@ -643,6 +643,8 @@ __SYSCALL(__NR_timerfd_gettime, sys_timerfd_gettime) __SYSCALL(__NR_paccept, sys_paccept) #define __NR_signalfd4 289 __SYSCALL(__NR_signalfd4, sys_signalfd4) +#define __NR_eventfd2 290 +__SYSCALL(__NR_eventfd2, sys_eventfd2) #ifndef __NO_STUBS diff --git a/include/linux/eventfd.h b/include/linux/eventfd.h index a701399b7fe..a6c0eaedb1b 100644 --- a/include/linux/eventfd.h +++ b/include/linux/eventfd.h @@ -10,6 +10,12 @@ #ifdef CONFIG_EVENTFD +/* For O_CLOEXEC */ +#include + +/* Flags for eventfd2. */ +#define EFD_CLOEXEC O_CLOEXEC + struct file *eventfd_fget(int fd); int eventfd_signal(struct file *file, int n); diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h index 1c270779784..9ab09926a7f 100644 --- a/include/linux/syscalls.h +++ b/include/linux/syscalls.h @@ -617,6 +617,7 @@ asmlinkage long sys_timerfd_settime(int ufd, int flags, struct itimerspec __user *otmr); asmlinkage long sys_timerfd_gettime(int ufd, struct itimerspec __user *otmr); asmlinkage long sys_eventfd(unsigned int count); +asmlinkage long sys_eventfd2(unsigned int count, int flags); asmlinkage long sys_fallocate(int fd, int mode, loff_t offset, loff_t len); int kernel_execve(const char *filename, char *const argv[], char *const envp[]); diff --git a/kernel/sys_ni.c b/kernel/sys_ni.c index 8627c89ae9e..2a361ccdc7c 100644 --- a/kernel/sys_ni.c +++ b/kernel/sys_ni.c @@ -164,3 +164,4 @@ cond_syscall(sys_timerfd_gettime); cond_syscall(compat_sys_timerfd_settime); cond_syscall(compat_sys_timerfd_gettime); cond_syscall(sys_eventfd); +cond_syscall(sys_eventfd2); -- cgit v1.2.3-70-g09d2 From 11fcb6c14676023d0bd437841f5dcd670e7990a0 Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Wed, 23 Jul 2008 21:29:26 -0700 Subject: flag parameters: timerfd_create The timerfd_create syscall already has a flags parameter. It just is unused so far. This patch changes this by introducing the TFD_CLOEXEC flag to set the close-on-exec flag for the returned file descriptor. A new name TFD_CLOEXEC is introduced which in this implementation must have the same value as O_CLOEXEC. The following test must be adjusted for architectures other than x86 and x86-64 and in case the syscall numbers changed. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ #include #include #include #include #include #ifndef __NR_timerfd_create # ifdef __x86_64__ # define __NR_timerfd_create 283 # elif defined __i386__ # define __NR_timerfd_create 322 # else # error "need __NR_timerfd_create" # endif #endif #define TFD_CLOEXEC O_CLOEXEC int main (void) { int fd = syscall (__NR_timerfd_create, CLOCK_REALTIME, 0); if (fd == -1) { puts ("timerfd_create(0) failed"); return 1; } int coe = fcntl (fd, F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if (coe & FD_CLOEXEC) { puts ("timerfd_create(0) set close-on-exec flag"); return 1; } close (fd); fd = syscall (__NR_timerfd_create, CLOCK_REALTIME, TFD_CLOEXEC); if (fd == -1) { puts ("timerfd_create(TFD_CLOEXEC) failed"); return 1; } coe = fcntl (fd, F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if ((coe & FD_CLOEXEC) == 0) { puts ("timerfd_create(TFD_CLOEXEC) set close-on-exec flag"); return 1; } close (fd); puts ("OK"); return 0; } ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Signed-off-by: Ulrich Drepper Acked-by: Davide Libenzi Cc: Michael Kerrisk Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/timerfd.c | 5 +++-- include/linux/timerfd.h | 5 +++++ 2 files changed, 8 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/timerfd.c b/fs/timerfd.c index 77c2bc92cbe..c6ef5e33cb3 100644 --- a/fs/timerfd.c +++ b/fs/timerfd.c @@ -184,7 +184,7 @@ asmlinkage long sys_timerfd_create(int clockid, int flags) int ufd; struct timerfd_ctx *ctx; - if (flags) + if (flags & ~TFD_CLOEXEC) return -EINVAL; if (clockid != CLOCK_MONOTONIC && clockid != CLOCK_REALTIME) @@ -198,7 +198,8 @@ asmlinkage long sys_timerfd_create(int clockid, int flags) ctx->clockid = clockid; hrtimer_init(&ctx->tmr, clockid, HRTIMER_MODE_ABS); - ufd = anon_inode_getfd("[timerfd]", &timerfd_fops, ctx, 0); + ufd = anon_inode_getfd("[timerfd]", &timerfd_fops, ctx, + flags & O_CLOEXEC); if (ufd < 0) kfree(ctx); diff --git a/include/linux/timerfd.h b/include/linux/timerfd.h index cf2b10d7573..96ed97dff00 100644 --- a/include/linux/timerfd.h +++ b/include/linux/timerfd.h @@ -8,9 +8,14 @@ #ifndef _LINUX_TIMERFD_H #define _LINUX_TIMERFD_H +/* For O_CLOEXEC */ +#include +/* Flags for timerfd_settime. */ #define TFD_TIMER_ABSTIME (1 << 0) +/* Flags for timerfd_create. */ +#define TFD_CLOEXEC O_CLOEXEC #endif /* _LINUX_TIMERFD_H */ -- cgit v1.2.3-70-g09d2 From a0998b50c3f0b8fdd265c63e0032f86ebe377dbf Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Wed, 23 Jul 2008 21:29:27 -0700 Subject: flag parameters: epoll_create This patch adds the new epoll_create2 syscall. It extends the old epoll_create syscall by one parameter which is meant to hold a flag value. In this patch the only flag support is EPOLL_CLOEXEC which causes the close-on-exec flag for the returned file descriptor to be set. A new name EPOLL_CLOEXEC is introduced which in this implementation must have the same value as O_CLOEXEC. The following test must be adjusted for architectures other than x86 and x86-64 and in case the syscall numbers changed. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ #include #include #include #include #include #ifndef __NR_epoll_create2 # ifdef __x86_64__ # define __NR_epoll_create2 291 # elif defined __i386__ # define __NR_epoll_create2 329 # else # error "need __NR_epoll_create2" # endif #endif #define EPOLL_CLOEXEC O_CLOEXEC int main (void) { int fd = syscall (__NR_epoll_create2, 1, 0); if (fd == -1) { puts ("epoll_create2(0) failed"); return 1; } int coe = fcntl (fd, F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if (coe & FD_CLOEXEC) { puts ("epoll_create2(0) set close-on-exec flag"); return 1; } close (fd); fd = syscall (__NR_epoll_create2, 1, EPOLL_CLOEXEC); if (fd == -1) { puts ("epoll_create2(EPOLL_CLOEXEC) failed"); return 1; } coe = fcntl (fd, F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if ((coe & FD_CLOEXEC) == 0) { puts ("epoll_create2(EPOLL_CLOEXEC) set close-on-exec flag"); return 1; } close (fd); puts ("OK"); return 0; } ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Signed-off-by: Ulrich Drepper Acked-by: Davide Libenzi Cc: Michael Kerrisk Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/x86/ia32/ia32entry.S | 1 + arch/x86/kernel/syscall_table_32.S | 1 + fs/eventpoll.c | 13 +++++++++++-- include/asm-x86/unistd_32.h | 1 + include/asm-x86/unistd_64.h | 2 ++ include/linux/eventpoll.h | 4 ++++ include/linux/syscalls.h | 1 + 7 files changed, 21 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/arch/x86/ia32/ia32entry.S b/arch/x86/ia32/ia32entry.S index cf0eb31745c..04366f08f42 100644 --- a/arch/x86/ia32/ia32entry.S +++ b/arch/x86/ia32/ia32entry.S @@ -828,4 +828,5 @@ ia32_sys_call_table: .quad compat_sys_timerfd_gettime .quad compat_sys_signalfd4 .quad sys_eventfd2 + .quad sys_epoll_create2 ia32_syscall_end: diff --git a/arch/x86/kernel/syscall_table_32.S b/arch/x86/kernel/syscall_table_32.S index cf112cb11c3..4d7007ca263 100644 --- a/arch/x86/kernel/syscall_table_32.S +++ b/arch/x86/kernel/syscall_table_32.S @@ -328,3 +328,4 @@ ENTRY(sys_call_table) .long sys_timerfd_gettime .long sys_signalfd4 .long sys_eventfd2 + .long sys_epoll_create2 diff --git a/fs/eventpoll.c b/fs/eventpoll.c index 9392dd96812..3fd4014f3c5 100644 --- a/fs/eventpoll.c +++ b/fs/eventpoll.c @@ -1046,11 +1046,14 @@ retry: * RB tree. With the current implementation, the "size" parameter is ignored * (besides sanity checks). */ -asmlinkage long sys_epoll_create(int size) +asmlinkage long sys_epoll_create2(int size, int flags) { int error, fd = -1; struct eventpoll *ep; + if (flags & ~EPOLL_CLOEXEC) + return -EINVAL; + DNPRINTK(3, (KERN_INFO "[%p] eventpoll: sys_epoll_create(%d)\n", current, size)); @@ -1068,7 +1071,8 @@ asmlinkage long sys_epoll_create(int size) * Creates all the items needed to setup an eventpoll file. That is, * a file structure and a free file descriptor. */ - fd = anon_inode_getfd("[eventpoll]", &eventpoll_fops, ep, 0); + fd = anon_inode_getfd("[eventpoll]", &eventpoll_fops, ep, + flags & O_CLOEXEC); if (fd < 0) ep_free(ep); @@ -1079,6 +1083,11 @@ error_return: return fd; } +asmlinkage long sys_epoll_create(int size) +{ + return sys_epoll_create2(size, 0); +} + /* * The following function implements the controller interface for * the eventpoll file that enables the insertion/removal/change of diff --git a/include/asm-x86/unistd_32.h b/include/asm-x86/unistd_32.h index edbd8723c93..a37d6b0c4e1 100644 --- a/include/asm-x86/unistd_32.h +++ b/include/asm-x86/unistd_32.h @@ -334,6 +334,7 @@ #define __NR_timerfd_gettime 326 #define __NR_signalfd4 327 #define __NR_eventfd2 328 +#define __NR_epoll_create2 329 #ifdef __KERNEL__ diff --git a/include/asm-x86/unistd_64.h b/include/asm-x86/unistd_64.h index fb059a6feeb..a1a4a5b6e5e 100644 --- a/include/asm-x86/unistd_64.h +++ b/include/asm-x86/unistd_64.h @@ -645,6 +645,8 @@ __SYSCALL(__NR_paccept, sys_paccept) __SYSCALL(__NR_signalfd4, sys_signalfd4) #define __NR_eventfd2 290 __SYSCALL(__NR_eventfd2, sys_eventfd2) +#define __NR_epoll_create2 291 +__SYSCALL(__NR_epoll_create2, sys_epoll_create2) #ifndef __NO_STUBS diff --git a/include/linux/eventpoll.h b/include/linux/eventpoll.h index cf79853967f..1cfaa40059c 100644 --- a/include/linux/eventpoll.h +++ b/include/linux/eventpoll.h @@ -14,8 +14,12 @@ #ifndef _LINUX_EVENTPOLL_H #define _LINUX_EVENTPOLL_H +/* For O_CLOEXEC */ +#include #include +/* Flags for epoll_create2. */ +#define EPOLL_CLOEXEC O_CLOEXEC /* Valid opcodes to issue to sys_epoll_ctl() */ #define EPOLL_CTL_ADD 1 diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h index 9ab09926a7f..85953240f28 100644 --- a/include/linux/syscalls.h +++ b/include/linux/syscalls.h @@ -430,6 +430,7 @@ asmlinkage long sys_poll(struct pollfd __user *ufds, unsigned int nfds, asmlinkage long sys_select(int n, fd_set __user *inp, fd_set __user *outp, fd_set __user *exp, struct timeval __user *tvp); asmlinkage long sys_epoll_create(int size); +asmlinkage long sys_epoll_create2(int size, int flags); asmlinkage long sys_epoll_ctl(int epfd, int op, int fd, struct epoll_event __user *event); asmlinkage long sys_epoll_wait(int epfd, struct epoll_event __user *events, -- cgit v1.2.3-70-g09d2 From 336dd1f70ff62d7dd8655228caed4c5bfc818c56 Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Wed, 23 Jul 2008 21:29:29 -0700 Subject: flag parameters: dup2 This patch adds the new dup3 syscall. It extends the old dup2 syscall by one parameter which is meant to hold a flag value. Support for the O_CLOEXEC flag is added in this patch. The following test must be adjusted for architectures other than x86 and x86-64 and in case the syscall numbers changed. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ #include #include #include #include #include #ifndef __NR_dup3 # ifdef __x86_64__ # define __NR_dup3 292 # elif defined __i386__ # define __NR_dup3 330 # else # error "need __NR_dup3" # endif #endif int main (void) { int fd = syscall (__NR_dup3, 1, 4, 0); if (fd == -1) { puts ("dup3(0) failed"); return 1; } int coe = fcntl (fd, F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if (coe & FD_CLOEXEC) { puts ("dup3(0) set close-on-exec flag"); return 1; } close (fd); fd = syscall (__NR_dup3, 1, 4, O_CLOEXEC); if (fd == -1) { puts ("dup3(O_CLOEXEC) failed"); return 1; } coe = fcntl (fd, F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if ((coe & FD_CLOEXEC) == 0) { puts ("dup3(O_CLOEXEC) set close-on-exec flag"); return 1; } close (fd); puts ("OK"); return 0; } ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Signed-off-by: Ulrich Drepper Acked-by: Davide Libenzi Cc: Michael Kerrisk Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/x86/ia32/ia32entry.S | 1 + arch/x86/kernel/syscall_table_32.S | 1 + fs/fcntl.c | 15 +++++++++++++-- include/asm-x86/unistd_32.h | 1 + include/asm-x86/unistd_64.h | 2 ++ include/linux/syscalls.h | 1 + 6 files changed, 19 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/arch/x86/ia32/ia32entry.S b/arch/x86/ia32/ia32entry.S index 04366f08f42..5614a8f7bed 100644 --- a/arch/x86/ia32/ia32entry.S +++ b/arch/x86/ia32/ia32entry.S @@ -829,4 +829,5 @@ ia32_sys_call_table: .quad compat_sys_signalfd4 .quad sys_eventfd2 .quad sys_epoll_create2 + .quad sys_dup3 /* 330 */ ia32_syscall_end: diff --git a/arch/x86/kernel/syscall_table_32.S b/arch/x86/kernel/syscall_table_32.S index 4d7007ca263..24a3f1ea6a0 100644 --- a/arch/x86/kernel/syscall_table_32.S +++ b/arch/x86/kernel/syscall_table_32.S @@ -329,3 +329,4 @@ ENTRY(sys_call_table) .long sys_signalfd4 .long sys_eventfd2 .long sys_epoll_create2 + .long sys_dup3 /* 330 */ diff --git a/fs/fcntl.c b/fs/fcntl.c index 330a7d78259..9679fcbdeaa 100644 --- a/fs/fcntl.c +++ b/fs/fcntl.c @@ -125,13 +125,16 @@ static int dupfd(struct file *file, unsigned int start, int cloexec) return fd; } -asmlinkage long sys_dup2(unsigned int oldfd, unsigned int newfd) +asmlinkage long sys_dup3(unsigned int oldfd, unsigned int newfd, int flags) { int err = -EBADF; struct file * file, *tofree; struct files_struct * files = current->files; struct fdtable *fdt; + if ((flags & ~O_CLOEXEC) != 0) + return -EINVAL; + spin_lock(&files->file_lock); if (!(file = fcheck(oldfd))) goto out_unlock; @@ -163,7 +166,10 @@ asmlinkage long sys_dup2(unsigned int oldfd, unsigned int newfd) rcu_assign_pointer(fdt->fd[newfd], file); FD_SET(newfd, fdt->open_fds); - FD_CLR(newfd, fdt->close_on_exec); + if (flags & O_CLOEXEC) + FD_SET(newfd, fdt->close_on_exec); + else + FD_CLR(newfd, fdt->close_on_exec); spin_unlock(&files->file_lock); if (tofree) @@ -181,6 +187,11 @@ out_fput: goto out; } +asmlinkage long sys_dup2(unsigned int oldfd, unsigned int newfd) +{ + return sys_dup3(oldfd, newfd, 0); +} + asmlinkage long sys_dup(unsigned int fildes) { int ret = -EBADF; diff --git a/include/asm-x86/unistd_32.h b/include/asm-x86/unistd_32.h index a37d6b0c4e1..a1f6383bf69 100644 --- a/include/asm-x86/unistd_32.h +++ b/include/asm-x86/unistd_32.h @@ -335,6 +335,7 @@ #define __NR_signalfd4 327 #define __NR_eventfd2 328 #define __NR_epoll_create2 329 +#define __NR_dup3 330 #ifdef __KERNEL__ diff --git a/include/asm-x86/unistd_64.h b/include/asm-x86/unistd_64.h index a1a4a5b6e5e..f0fb2bd40cd 100644 --- a/include/asm-x86/unistd_64.h +++ b/include/asm-x86/unistd_64.h @@ -647,6 +647,8 @@ __SYSCALL(__NR_signalfd4, sys_signalfd4) __SYSCALL(__NR_eventfd2, sys_eventfd2) #define __NR_epoll_create2 291 __SYSCALL(__NR_epoll_create2, sys_epoll_create2) +#define __NR_dup3 292 +__SYSCALL(__NR_dup3, sys_dup3) #ifndef __NO_STUBS diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h index 85953240f28..034d3358549 100644 --- a/include/linux/syscalls.h +++ b/include/linux/syscalls.h @@ -305,6 +305,7 @@ asmlinkage long sys_fcntl64(unsigned int fd, #endif asmlinkage long sys_dup(unsigned int fildes); asmlinkage long sys_dup2(unsigned int oldfd, unsigned int newfd); +asmlinkage long sys_dup3(unsigned int oldfd, unsigned int newfd, int flags); asmlinkage long sys_ioperm(unsigned long from, unsigned long num, int on); asmlinkage long sys_ioctl(unsigned int fd, unsigned int cmd, unsigned long arg); -- cgit v1.2.3-70-g09d2 From ed8cae8ba01348bfd83333f4648dd807b04d7f08 Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Wed, 23 Jul 2008 21:29:30 -0700 Subject: flag parameters: pipe This patch introduces the new syscall pipe2 which is like pipe but it also takes an additional parameter which takes a flag value. This patch implements the handling of O_CLOEXEC for the flag. I did not add support for the new syscall for the architectures which have a special sys_pipe implementation. I think the maintainers of those archs have the chance to go with the unified implementation but that's up to them. The implementation introduces do_pipe_flags. I did that instead of changing all callers of do_pipe because some of the callers are written in assembler. I would probably screw up changing the assembly code. To avoid breaking code do_pipe is now a small wrapper around do_pipe_flags. Once all callers are changed over to do_pipe_flags the old do_pipe function can be removed. The following test must be adjusted for architectures other than x86 and x86-64 and in case the syscall numbers changed. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ #include #include #include #include #ifndef __NR_pipe2 # ifdef __x86_64__ # define __NR_pipe2 293 # elif defined __i386__ # define __NR_pipe2 331 # else # error "need __NR_pipe2" # endif #endif int main (void) { int fd[2]; if (syscall (__NR_pipe2, fd, 0) != 0) { puts ("pipe2(0) failed"); return 1; } for (int i = 0; i < 2; ++i) { int coe = fcntl (fd[i], F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if (coe & FD_CLOEXEC) { printf ("pipe2(0) set close-on-exit for fd[%d]\n", i); return 1; } } close (fd[0]); close (fd[1]); if (syscall (__NR_pipe2, fd, O_CLOEXEC) != 0) { puts ("pipe2(O_CLOEXEC) failed"); return 1; } for (int i = 0; i < 2; ++i) { int coe = fcntl (fd[i], F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if ((coe & FD_CLOEXEC) == 0) { printf ("pipe2(O_CLOEXEC) does not set close-on-exit for fd[%d]\n", i); return 1; } } close (fd[0]); close (fd[1]); puts ("OK"); return 0; } ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Signed-off-by: Ulrich Drepper Acked-by: Davide Libenzi Cc: Michael Kerrisk Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/ia64/ia32/sys_ia32.c | 2 +- arch/ia64/kernel/sys_ia64.c | 2 +- arch/mips/kernel/syscall.c | 2 +- arch/parisc/hpux/sys_hpux.c | 2 +- arch/sh/kernel/sys_sh32.c | 2 +- arch/sparc/kernel/sys_sparc.c | 2 +- arch/sparc64/kernel/sys_sparc.c | 2 +- arch/x86/ia32/ia32entry.S | 1 + arch/x86/ia32/sys_ia32.c | 2 +- arch/x86/kernel/syscall_table_32.S | 1 + arch/xtensa/kernel/syscall.c | 2 +- fs/pipe.c | 23 ++++++++++++++++++----- include/asm-x86/unistd_32.h | 1 + include/asm-x86/unistd_64.h | 2 ++ include/linux/fs.h | 1 + 15 files changed, 33 insertions(+), 14 deletions(-) (limited to 'fs') diff --git a/arch/ia64/ia32/sys_ia32.c b/arch/ia64/ia32/sys_ia32.c index 7e028ceb93b..465116aecb8 100644 --- a/arch/ia64/ia32/sys_ia32.c +++ b/arch/ia64/ia32/sys_ia32.c @@ -1139,7 +1139,7 @@ sys32_pipe (int __user *fd) int retval; int fds[2]; - retval = do_pipe(fds); + retval = do_pipe_flags(fds, 0); if (retval) goto out; if (copy_to_user(fd, fds, sizeof(fds))) diff --git a/arch/ia64/kernel/sys_ia64.c b/arch/ia64/kernel/sys_ia64.c index 1eda194b955..bcbb6d8792d 100644 --- a/arch/ia64/kernel/sys_ia64.c +++ b/arch/ia64/kernel/sys_ia64.c @@ -160,7 +160,7 @@ sys_pipe (void) int fd[2]; int retval; - retval = do_pipe(fd); + retval = do_pipe_flags(fd, 0); if (retval) goto out; retval = fd[0]; diff --git a/arch/mips/kernel/syscall.c b/arch/mips/kernel/syscall.c index 3523c8d12ed..343015a2f41 100644 --- a/arch/mips/kernel/syscall.c +++ b/arch/mips/kernel/syscall.c @@ -52,7 +52,7 @@ asmlinkage int sysm_pipe(nabi_no_regargs volatile struct pt_regs regs) int fd[2]; int error, res; - error = do_pipe(fd); + error = do_pipe_flags(fd, 0); if (error) { res = error; goto out; diff --git a/arch/parisc/hpux/sys_hpux.c b/arch/parisc/hpux/sys_hpux.c index 0c5b9dabb47..be255ebb609 100644 --- a/arch/parisc/hpux/sys_hpux.c +++ b/arch/parisc/hpux/sys_hpux.c @@ -448,7 +448,7 @@ int hpux_pipe(int *kstack_fildes) int error; lock_kernel(); - error = do_pipe(kstack_fildes); + error = do_pipe_flags(kstack_fildes, 0); unlock_kernel(); return error; } diff --git a/arch/sh/kernel/sys_sh32.c b/arch/sh/kernel/sys_sh32.c index 125e493ead8..f0aa5c39865 100644 --- a/arch/sh/kernel/sys_sh32.c +++ b/arch/sh/kernel/sys_sh32.c @@ -29,7 +29,7 @@ asmlinkage int sys_pipe(unsigned long r4, unsigned long r5, int fd[2]; int error; - error = do_pipe(fd); + error = do_pipe_flags(fd, 0); if (!error) { regs->regs[1] = fd[1]; return fd[0]; diff --git a/arch/sparc/kernel/sys_sparc.c b/arch/sparc/kernel/sys_sparc.c index 3c6b49a53ae..4d73421559c 100644 --- a/arch/sparc/kernel/sys_sparc.c +++ b/arch/sparc/kernel/sys_sparc.c @@ -97,7 +97,7 @@ asmlinkage int sparc_pipe(struct pt_regs *regs) int fd[2]; int error; - error = do_pipe(fd); + error = do_pipe_flags(fd, 0); if (error) goto out; regs->u_regs[UREG_I1] = fd[1]; diff --git a/arch/sparc64/kernel/sys_sparc.c b/arch/sparc64/kernel/sys_sparc.c index e1f4eba2e57..39749e32dc7 100644 --- a/arch/sparc64/kernel/sys_sparc.c +++ b/arch/sparc64/kernel/sys_sparc.c @@ -418,7 +418,7 @@ asmlinkage long sparc_pipe(struct pt_regs *regs) int fd[2]; int error; - error = do_pipe(fd); + error = do_pipe_flags(fd, 0); if (error) goto out; regs->u_regs[UREG_I1] = fd[1]; diff --git a/arch/x86/ia32/ia32entry.S b/arch/x86/ia32/ia32entry.S index 5614a8f7bed..18808b16457 100644 --- a/arch/x86/ia32/ia32entry.S +++ b/arch/x86/ia32/ia32entry.S @@ -830,4 +830,5 @@ ia32_sys_call_table: .quad sys_eventfd2 .quad sys_epoll_create2 .quad sys_dup3 /* 330 */ + .quad sys_pipe2 ia32_syscall_end: diff --git a/arch/x86/ia32/sys_ia32.c b/arch/x86/ia32/sys_ia32.c index f00afdf61e6..d3c64088b98 100644 --- a/arch/x86/ia32/sys_ia32.c +++ b/arch/x86/ia32/sys_ia32.c @@ -238,7 +238,7 @@ asmlinkage long sys32_pipe(int __user *fd) int retval; int fds[2]; - retval = do_pipe(fds); + retval = do_pipe_flags(fds, 0); if (retval) goto out; if (copy_to_user(fd, fds, sizeof(fds))) diff --git a/arch/x86/kernel/syscall_table_32.S b/arch/x86/kernel/syscall_table_32.S index 24a3f1ea6a0..66154769d52 100644 --- a/arch/x86/kernel/syscall_table_32.S +++ b/arch/x86/kernel/syscall_table_32.S @@ -330,3 +330,4 @@ ENTRY(sys_call_table) .long sys_eventfd2 .long sys_epoll_create2 .long sys_dup3 /* 330 */ + .long sys_pipe2 diff --git a/arch/xtensa/kernel/syscall.c b/arch/xtensa/kernel/syscall.c index f3e16efcd47..ac15ecbdf91 100644 --- a/arch/xtensa/kernel/syscall.c +++ b/arch/xtensa/kernel/syscall.c @@ -49,7 +49,7 @@ asmlinkage long xtensa_pipe(int __user *userfds) int fd[2]; int error; - error = do_pipe(fd); + error = do_pipe_flags(fd, 0); if (!error) { if (copy_to_user(userfds, fd, 2 * sizeof(int))) error = -EFAULT; diff --git a/fs/pipe.c b/fs/pipe.c index 700f4e0d957..68e82061070 100644 --- a/fs/pipe.c +++ b/fs/pipe.c @@ -1027,12 +1027,15 @@ struct file *create_read_pipe(struct file *wrf) return f; } -int do_pipe(int *fd) +int do_pipe_flags(int *fd, int flags) { struct file *fw, *fr; int error; int fdw, fdr; + if (flags & ~O_CLOEXEC) + return -EINVAL; + fw = create_write_pipe(); if (IS_ERR(fw)) return PTR_ERR(fw); @@ -1041,12 +1044,12 @@ int do_pipe(int *fd) if (IS_ERR(fr)) goto err_write_pipe; - error = get_unused_fd(); + error = get_unused_fd_flags(flags); if (error < 0) goto err_read_pipe; fdr = error; - error = get_unused_fd(); + error = get_unused_fd_flags(flags); if (error < 0) goto err_fdr; fdw = error; @@ -1074,16 +1077,21 @@ int do_pipe(int *fd) return error; } +int do_pipe(int *fd) +{ + return do_pipe_flags(fd, 0); +} + /* * sys_pipe() is the normal C calling standard for creating * a pipe. It's not the way Unix traditionally does this, though. */ -asmlinkage long __weak sys_pipe(int __user *fildes) +asmlinkage long __weak sys_pipe2(int __user *fildes, int flags) { int fd[2]; int error; - error = do_pipe(fd); + error = do_pipe_flags(fd, flags); if (!error) { if (copy_to_user(fildes, fd, sizeof(fd))) { sys_close(fd[0]); @@ -1094,6 +1102,11 @@ asmlinkage long __weak sys_pipe(int __user *fildes) return error; } +asmlinkage long __weak sys_pipe(int __user *fildes) +{ + return sys_pipe2(fildes, 0); +} + /* * pipefs should _never_ be mounted by userland - too much of security hassle, * no real gain from having the whole whorehouse mounted. So we don't need diff --git a/include/asm-x86/unistd_32.h b/include/asm-x86/unistd_32.h index a1f6383bf69..748a05c77da 100644 --- a/include/asm-x86/unistd_32.h +++ b/include/asm-x86/unistd_32.h @@ -336,6 +336,7 @@ #define __NR_eventfd2 328 #define __NR_epoll_create2 329 #define __NR_dup3 330 +#define __NR_pipe2 331 #ifdef __KERNEL__ diff --git a/include/asm-x86/unistd_64.h b/include/asm-x86/unistd_64.h index f0fb2bd40cd..d2284b43ad5 100644 --- a/include/asm-x86/unistd_64.h +++ b/include/asm-x86/unistd_64.h @@ -649,6 +649,8 @@ __SYSCALL(__NR_eventfd2, sys_eventfd2) __SYSCALL(__NR_epoll_create2, sys_epoll_create2) #define __NR_dup3 292 __SYSCALL(__NR_dup3, sys_dup3) +#define __NR_pipe2 293 +__SYSCALL(__NR_pipe2, sys_pipe2) #ifndef __NO_STUBS diff --git a/include/linux/fs.h b/include/linux/fs.h index e5e6a244096..0e80cd717d3 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1777,6 +1777,7 @@ static inline void allow_write_access(struct file *file) atomic_inc(&file->f_path.dentry->d_inode->i_writecount); } extern int do_pipe(int *); +extern int do_pipe_flags(int *, int); extern struct file *create_read_pipe(struct file *f); extern struct file *create_write_pipe(void); extern void free_write_pipe(struct file *); -- cgit v1.2.3-70-g09d2 From 4006553b06306b34054529477b06b68a1c66249b Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Wed, 23 Jul 2008 21:29:32 -0700 Subject: flag parameters: inotify_init This patch introduces the new syscall inotify_init1 (note: the 1 stands for the one parameter the syscall takes, as opposed to no parameter before). The values accepted for this parameter are function-specific and defined in the inotify.h header. Here the values must match the O_* flags, though. In this patch CLOEXEC support is introduced. The following test must be adjusted for architectures other than x86 and x86-64 and in case the syscall numbers changed. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ #include #include #include #include #ifndef __NR_inotify_init1 # ifdef __x86_64__ # define __NR_inotify_init1 294 # elif defined __i386__ # define __NR_inotify_init1 332 # else # error "need __NR_inotify_init1" # endif #endif #define IN_CLOEXEC O_CLOEXEC int main (void) { int fd; fd = syscall (__NR_inotify_init1, 0); if (fd == -1) { puts ("inotify_init1(0) failed"); return 1; } int coe = fcntl (fd, F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if (coe & FD_CLOEXEC) { puts ("inotify_init1(0) set close-on-exit"); return 1; } close (fd); fd = syscall (__NR_inotify_init1, IN_CLOEXEC); if (fd == -1) { puts ("inotify_init1(IN_CLOEXEC) failed"); return 1; } coe = fcntl (fd, F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if ((coe & FD_CLOEXEC) == 0) { puts ("inotify_init1(O_CLOEXEC) does not set close-on-exit"); return 1; } close (fd); puts ("OK"); return 0; } ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ [akpm@linux-foundation.org: add sys_ni stub] Signed-off-by: Ulrich Drepper Acked-by: Davide Libenzi Cc: Michael Kerrisk Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/x86/ia32/ia32entry.S | 1 + arch/x86/kernel/syscall_table_32.S | 1 + fs/inotify_user.c | 12 ++++++++++-- include/asm-x86/unistd_32.h | 1 + include/asm-x86/unistd_64.h | 2 ++ include/linux/inotify.h | 5 +++++ include/linux/syscalls.h | 1 + kernel/sys_ni.c | 1 + 8 files changed, 22 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/arch/x86/ia32/ia32entry.S b/arch/x86/ia32/ia32entry.S index 18808b16457..4541073dd83 100644 --- a/arch/x86/ia32/ia32entry.S +++ b/arch/x86/ia32/ia32entry.S @@ -831,4 +831,5 @@ ia32_sys_call_table: .quad sys_epoll_create2 .quad sys_dup3 /* 330 */ .quad sys_pipe2 + .quad sys_inotify_init1 ia32_syscall_end: diff --git a/arch/x86/kernel/syscall_table_32.S b/arch/x86/kernel/syscall_table_32.S index 66154769d52..f59aba5ff0f 100644 --- a/arch/x86/kernel/syscall_table_32.S +++ b/arch/x86/kernel/syscall_table_32.S @@ -331,3 +331,4 @@ ENTRY(sys_call_table) .long sys_epoll_create2 .long sys_dup3 /* 330 */ .long sys_pipe2 + .long sys_inotify_init1 diff --git a/fs/inotify_user.c b/fs/inotify_user.c index 6676c06bb7c..851005998cd 100644 --- a/fs/inotify_user.c +++ b/fs/inotify_user.c @@ -566,7 +566,7 @@ static const struct inotify_operations inotify_user_ops = { .destroy_watch = free_inotify_user_watch, }; -asmlinkage long sys_inotify_init(void) +asmlinkage long sys_inotify_init1(int flags) { struct inotify_device *dev; struct inotify_handle *ih; @@ -574,7 +574,10 @@ asmlinkage long sys_inotify_init(void) struct file *filp; int fd, ret; - fd = get_unused_fd(); + if (flags & ~IN_CLOEXEC) + return -EINVAL; + + fd = get_unused_fd_flags(flags & O_CLOEXEC); if (fd < 0) return fd; @@ -638,6 +641,11 @@ out_put_fd: return ret; } +asmlinkage long sys_inotify_init(void) +{ + return sys_inotify_init1(0); +} + asmlinkage long sys_inotify_add_watch(int fd, const char __user *path, u32 mask) { struct inode *inode; diff --git a/include/asm-x86/unistd_32.h b/include/asm-x86/unistd_32.h index 748a05c77da..b3daf503ab9 100644 --- a/include/asm-x86/unistd_32.h +++ b/include/asm-x86/unistd_32.h @@ -337,6 +337,7 @@ #define __NR_epoll_create2 329 #define __NR_dup3 330 #define __NR_pipe2 331 +#define __NR_inotify_init1 332 #ifdef __KERNEL__ diff --git a/include/asm-x86/unistd_64.h b/include/asm-x86/unistd_64.h index d2284b43ad5..c8cb88d70c6 100644 --- a/include/asm-x86/unistd_64.h +++ b/include/asm-x86/unistd_64.h @@ -651,6 +651,8 @@ __SYSCALL(__NR_epoll_create2, sys_epoll_create2) __SYSCALL(__NR_dup3, sys_dup3) #define __NR_pipe2 293 __SYSCALL(__NR_pipe2, sys_pipe2) +#define __NR_inotify_init1 294 +__SYSCALL(__NR_inotify_init1, sys_inotify_init1) #ifndef __NO_STUBS diff --git a/include/linux/inotify.h b/include/linux/inotify.h index 742b917e7d1..72ef8212051 100644 --- a/include/linux/inotify.h +++ b/include/linux/inotify.h @@ -7,6 +7,8 @@ #ifndef _LINUX_INOTIFY_H #define _LINUX_INOTIFY_H +/* For O_CLOEXEC */ +#include #include /* @@ -63,6 +65,9 @@ struct inotify_event { IN_MOVED_TO | IN_DELETE | IN_CREATE | IN_DELETE_SELF | \ IN_MOVE_SELF) +/* Flags for sys_inotify_init1. */ +#define IN_CLOEXEC O_CLOEXEC + #ifdef __KERNEL__ #include diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h index 034d3358549..93a7e7f017a 100644 --- a/include/linux/syscalls.h +++ b/include/linux/syscalls.h @@ -547,6 +547,7 @@ asmlinkage long sys_get_mempolicy(int __user *policy, unsigned long addr, unsigned long flags); asmlinkage long sys_inotify_init(void); +asmlinkage long sys_inotify_init1(int flags); asmlinkage long sys_inotify_add_watch(int fd, const char __user *path, u32 mask); asmlinkage long sys_inotify_rm_watch(int fd, u32 wd); diff --git a/kernel/sys_ni.c b/kernel/sys_ni.c index 2a361ccdc7c..bd66ac5406f 100644 --- a/kernel/sys_ni.c +++ b/kernel/sys_ni.c @@ -96,6 +96,7 @@ cond_syscall(sys_keyctl); cond_syscall(compat_sys_keyctl); cond_syscall(compat_sys_socketcall); cond_syscall(sys_inotify_init); +cond_syscall(sys_inotify_init1); cond_syscall(sys_inotify_add_watch); cond_syscall(sys_inotify_rm_watch); cond_syscall(sys_migrate_pages); -- cgit v1.2.3-70-g09d2 From 99829b832997d907c30669bfd17da32151e18f04 Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Wed, 23 Jul 2008 21:29:33 -0700 Subject: flag parameters: NONBLOCK in anon_inode_getfd Building on the previous change to anon_inode_getfd, this patch introduces support for handling of O_NONBLOCK in addition to the already supported O_CLOEXEC. Following patches will take advantage of this support. As can be seen, the additional support for supporting this functionality is minimal. Signed-off-by: Ulrich Drepper Acked-by: Davide Libenzi Cc: Michael Kerrisk Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/anon_inodes.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/anon_inodes.c b/fs/anon_inodes.c index 1a4eee620b0..3662dd44896 100644 --- a/fs/anon_inodes.c +++ b/fs/anon_inodes.c @@ -116,7 +116,7 @@ int anon_inode_getfd(const char *name, const struct file_operations *fops, file->f_mapping = anon_inode_inode->i_mapping; file->f_pos = 0; - file->f_flags = O_RDWR; + file->f_flags = O_RDWR | (flags & O_NONBLOCK); file->f_version = 0; file->private_data = priv; -- cgit v1.2.3-70-g09d2 From 5fb5e04926a54bc1c22bba7ca166840f4476196f Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Wed, 23 Jul 2008 21:29:37 -0700 Subject: flag parameters: NONBLOCK in signalfd This patch adds support for the SFD_NONBLOCK flag to signalfd4. The additional changes needed are minimal. The following test must be adjusted for architectures other than x86 and x86-64 and in case the syscall numbers changed. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ #include #include #include #include #include #ifndef __NR_signalfd4 # ifdef __x86_64__ # define __NR_signalfd4 289 # elif defined __i386__ # define __NR_signalfd4 327 # else # error "need __NR_signalfd4" # endif #endif #define SFD_NONBLOCK O_NONBLOCK int main (void) { sigset_t ss; sigemptyset (&ss); sigaddset (&ss, SIGUSR1); int fd = syscall (__NR_signalfd4, -1, &ss, 8, 0); if (fd == -1) { puts ("signalfd4(0) failed"); return 1; } int fl = fcntl (fd, F_GETFL); if (fl == -1) { puts ("fcntl failed"); return 1; } if (fl & O_NONBLOCK) { puts ("signalfd4(0) set non-blocking mode"); return 1; } close (fd); fd = syscall (__NR_signalfd4, -1, &ss, 8, SFD_NONBLOCK); if (fd == -1) { puts ("signalfd4(SFD_NONBLOCK) failed"); return 1; } fl = fcntl (fd, F_GETFL); if (fl == -1) { puts ("fcntl failed"); return 1; } if ((fl & O_NONBLOCK) == 0) { puts ("signalfd4(SFD_NONBLOCK) does not set non-blocking mode"); return 1; } close (fd); puts ("OK"); return 0; } ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Signed-off-by: Ulrich Drepper Acked-by: Davide Libenzi Cc: Michael Kerrisk Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/signalfd.c | 4 ++-- include/linux/signalfd.h | 3 ++- 2 files changed, 4 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/signalfd.c b/fs/signalfd.c index c8609fa51a1..5441a4bca77 100644 --- a/fs/signalfd.c +++ b/fs/signalfd.c @@ -211,7 +211,7 @@ asmlinkage long sys_signalfd4(int ufd, sigset_t __user *user_mask, sigset_t sigmask; struct signalfd_ctx *ctx; - if (flags & ~SFD_CLOEXEC) + if (flags & ~(SFD_CLOEXEC | SFD_NONBLOCK)) return -EINVAL; if (sizemask != sizeof(sigset_t) || @@ -232,7 +232,7 @@ asmlinkage long sys_signalfd4(int ufd, sigset_t __user *user_mask, * anon_inode_getfd() will install the fd. */ ufd = anon_inode_getfd("[signalfd]", &signalfd_fops, ctx, - flags & O_CLOEXEC); + flags & (O_CLOEXEC | O_NONBLOCK)); if (ufd < 0) kfree(ctx); } else { diff --git a/include/linux/signalfd.h b/include/linux/signalfd.h index 8b3f7b7420a..bef0c46d471 100644 --- a/include/linux/signalfd.h +++ b/include/linux/signalfd.h @@ -8,11 +8,12 @@ #ifndef _LINUX_SIGNALFD_H #define _LINUX_SIGNALFD_H -/* For O_CLOEXEC */ +/* For O_CLOEXEC and O_NONBLOCK */ #include /* Flags for signalfd4. */ #define SFD_CLOEXEC O_CLOEXEC +#define SFD_NONBLOCK O_NONBLOCK struct signalfd_siginfo { __u32 ssi_signo; -- cgit v1.2.3-70-g09d2 From e7d476dfdf0bcfed478a207aecfdc84f81efecaf Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Wed, 23 Jul 2008 21:29:38 -0700 Subject: flag parameters: NONBLOCK in eventfd This patch adds support for the EFD_NONBLOCK flag to eventfd2. The additional changes needed are minimal. The following test must be adjusted for architectures other than x86 and x86-64 and in case the syscall numbers changed. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ #include #include #include #include #ifndef __NR_eventfd2 # ifdef __x86_64__ # define __NR_eventfd2 290 # elif defined __i386__ # define __NR_eventfd2 328 # else # error "need __NR_eventfd2" # endif #endif #define EFD_NONBLOCK O_NONBLOCK int main (void) { int fd = syscall (__NR_eventfd2, 1, 0); if (fd == -1) { puts ("eventfd2(0) failed"); return 1; } int fl = fcntl (fd, F_GETFL); if (fl == -1) { puts ("fcntl failed"); return 1; } if (fl & O_NONBLOCK) { puts ("eventfd2(0) sets non-blocking mode"); return 1; } close (fd); fd = syscall (__NR_eventfd2, 1, EFD_NONBLOCK); if (fd == -1) { puts ("eventfd2(EFD_NONBLOCK) failed"); return 1; } fl = fcntl (fd, F_GETFL); if (fl == -1) { puts ("fcntl failed"); return 1; } if ((fl & O_NONBLOCK) == 0) { puts ("eventfd2(EFD_NONBLOCK) does not set non-blocking mode"); return 1; } close (fd); puts ("OK"); return 0; } ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Signed-off-by: Ulrich Drepper Acked-by: Davide Libenzi Cc: Michael Kerrisk Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/eventfd.c | 4 ++-- include/linux/eventfd.h | 3 ++- 2 files changed, 4 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/eventfd.c b/fs/eventfd.c index bd420e6478a..3ed4466177a 100644 --- a/fs/eventfd.c +++ b/fs/eventfd.c @@ -203,7 +203,7 @@ asmlinkage long sys_eventfd2(unsigned int count, int flags) int fd; struct eventfd_ctx *ctx; - if (flags & ~EFD_CLOEXEC) + if (flags & ~(EFD_CLOEXEC | EFD_NONBLOCK)) return -EINVAL; ctx = kmalloc(sizeof(*ctx), GFP_KERNEL); @@ -218,7 +218,7 @@ asmlinkage long sys_eventfd2(unsigned int count, int flags) * anon_inode_getfd() will install the fd. */ fd = anon_inode_getfd("[eventfd]", &eventfd_fops, ctx, - flags & O_CLOEXEC); + flags & (O_CLOEXEC | O_NONBLOCK)); if (fd < 0) kfree(ctx); return fd; diff --git a/include/linux/eventfd.h b/include/linux/eventfd.h index a6c0eaedb1b..a667637b54e 100644 --- a/include/linux/eventfd.h +++ b/include/linux/eventfd.h @@ -10,11 +10,12 @@ #ifdef CONFIG_EVENTFD -/* For O_CLOEXEC */ +/* For O_CLOEXEC and O_NONBLOCK */ #include /* Flags for eventfd2. */ #define EFD_CLOEXEC O_CLOEXEC +#define EFD_NONBLOCK O_NONBLOCK struct file *eventfd_fget(int fd); int eventfd_signal(struct file *file, int n); -- cgit v1.2.3-70-g09d2 From 6b1ef0e60d42f2fdaec26baee8327eb156347b4f Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Wed, 23 Jul 2008 21:29:39 -0700 Subject: flag parameters: NONBLOCK in timerfd_create This patch adds support for the TFD_NONBLOCK flag to timerfd_create. The additional changes needed are minimal. The following test must be adjusted for architectures other than x86 and x86-64 and in case the syscall numbers changed. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ #include #include #include #include #include #ifndef __NR_timerfd_create # ifdef __x86_64__ # define __NR_timerfd_create 283 # elif defined __i386__ # define __NR_timerfd_create 322 # else # error "need __NR_timerfd_create" # endif #endif #define TFD_NONBLOCK O_NONBLOCK int main (void) { int fd = syscall (__NR_timerfd_create, CLOCK_REALTIME, 0); if (fd == -1) { puts ("timerfd_create(0) failed"); return 1; } int fl = fcntl (fd, F_GETFL); if (fl == -1) { puts ("fcntl failed"); return 1; } if (fl & O_NONBLOCK) { puts ("timerfd_create(0) set non-blocking mode"); return 1; } close (fd); fd = syscall (__NR_timerfd_create, CLOCK_REALTIME, TFD_NONBLOCK); if (fd == -1) { puts ("timerfd_create(TFD_NONBLOCK) failed"); return 1; } fl = fcntl (fd, F_GETFL); if (fl == -1) { puts ("fcntl failed"); return 1; } if ((fl & O_NONBLOCK) == 0) { puts ("timerfd_create(TFD_NONBLOCK) set non-blocking mode"); return 1; } close (fd); puts ("OK"); return 0; } ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Signed-off-by: Ulrich Drepper Acked-by: Davide Libenzi Cc: Michael Kerrisk Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/timerfd.c | 4 ++-- include/linux/timerfd.h | 3 ++- 2 files changed, 4 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/timerfd.c b/fs/timerfd.c index c6ef5e33cb3..75d44efe346 100644 --- a/fs/timerfd.c +++ b/fs/timerfd.c @@ -184,7 +184,7 @@ asmlinkage long sys_timerfd_create(int clockid, int flags) int ufd; struct timerfd_ctx *ctx; - if (flags & ~TFD_CLOEXEC) + if (flags & ~(TFD_CLOEXEC | TFD_NONBLOCK)) return -EINVAL; if (clockid != CLOCK_MONOTONIC && clockid != CLOCK_REALTIME) @@ -199,7 +199,7 @@ asmlinkage long sys_timerfd_create(int clockid, int flags) hrtimer_init(&ctx->tmr, clockid, HRTIMER_MODE_ABS); ufd = anon_inode_getfd("[timerfd]", &timerfd_fops, ctx, - flags & O_CLOEXEC); + flags & (O_CLOEXEC | O_NONBLOCK)); if (ufd < 0) kfree(ctx); diff --git a/include/linux/timerfd.h b/include/linux/timerfd.h index 96ed97dff00..86cb0501d3e 100644 --- a/include/linux/timerfd.h +++ b/include/linux/timerfd.h @@ -8,7 +8,7 @@ #ifndef _LINUX_TIMERFD_H #define _LINUX_TIMERFD_H -/* For O_CLOEXEC */ +/* For O_CLOEXEC and O_NONBLOCK */ #include /* Flags for timerfd_settime. */ @@ -16,6 +16,7 @@ /* Flags for timerfd_create. */ #define TFD_CLOEXEC O_CLOEXEC +#define TFD_NONBLOCK O_NONBLOCK #endif /* _LINUX_TIMERFD_H */ -- cgit v1.2.3-70-g09d2 From be61a86d7237dd80510615f38ae21d6e1e98660c Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Wed, 23 Jul 2008 21:29:40 -0700 Subject: flag parameters: NONBLOCK in pipe This patch adds O_NONBLOCK support to pipe2. It is minimally more involved than the patches for eventfd et.al but still trivial. The interfaces of the create_write_pipe and create_read_pipe helper functions were changed and the one other caller as well. The following test must be adjusted for architectures other than x86 and x86-64 and in case the syscall numbers changed. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ #include #include #include #include #ifndef __NR_pipe2 # ifdef __x86_64__ # define __NR_pipe2 293 # elif defined __i386__ # define __NR_pipe2 331 # else # error "need __NR_pipe2" # endif #endif int main (void) { int fds[2]; if (syscall (__NR_pipe2, fds, 0) == -1) { puts ("pipe2(0) failed"); return 1; } for (int i = 0; i < 2; ++i) { int fl = fcntl (fds[i], F_GETFL); if (fl == -1) { puts ("fcntl failed"); return 1; } if (fl & O_NONBLOCK) { printf ("pipe2(0) set non-blocking mode for fds[%d]\n", i); return 1; } close (fds[i]); } if (syscall (__NR_pipe2, fds, O_NONBLOCK) == -1) { puts ("pipe2(O_NONBLOCK) failed"); return 1; } for (int i = 0; i < 2; ++i) { int fl = fcntl (fds[i], F_GETFL); if (fl == -1) { puts ("fcntl failed"); return 1; } if ((fl & O_NONBLOCK) == 0) { printf ("pipe2(O_NONBLOCK) does not set non-blocking mode for fds[%d]\n", i); return 1; } close (fds[i]); } puts ("OK"); return 0; } ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Signed-off-by: Ulrich Drepper Acked-by: Davide Libenzi Cc: Michael Kerrisk Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/pipe.c | 14 +++++++------- include/linux/fs.h | 4 ++-- kernel/kmod.c | 4 ++-- 3 files changed, 11 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/pipe.c b/fs/pipe.c index 68e82061070..10c4e9aa5c4 100644 --- a/fs/pipe.c +++ b/fs/pipe.c @@ -950,7 +950,7 @@ fail_inode: return NULL; } -struct file *create_write_pipe(void) +struct file *create_write_pipe(int flags) { int err; struct inode *inode; @@ -983,7 +983,7 @@ struct file *create_write_pipe(void) goto err_dentry; f->f_mapping = inode->i_mapping; - f->f_flags = O_WRONLY; + f->f_flags = O_WRONLY | (flags & O_NONBLOCK); f->f_version = 0; return f; @@ -1007,7 +1007,7 @@ void free_write_pipe(struct file *f) put_filp(f); } -struct file *create_read_pipe(struct file *wrf) +struct file *create_read_pipe(struct file *wrf, int flags) { struct file *f = get_empty_filp(); if (!f) @@ -1019,7 +1019,7 @@ struct file *create_read_pipe(struct file *wrf) f->f_mapping = wrf->f_path.dentry->d_inode->i_mapping; f->f_pos = 0; - f->f_flags = O_RDONLY; + f->f_flags = O_RDONLY | (flags & O_NONBLOCK); f->f_op = &read_pipe_fops; f->f_mode = FMODE_READ; f->f_version = 0; @@ -1033,13 +1033,13 @@ int do_pipe_flags(int *fd, int flags) int error; int fdw, fdr; - if (flags & ~O_CLOEXEC) + if (flags & ~(O_CLOEXEC | O_NONBLOCK)) return -EINVAL; - fw = create_write_pipe(); + fw = create_write_pipe(flags); if (IS_ERR(fw)) return PTR_ERR(fw); - fr = create_read_pipe(fw); + fr = create_read_pipe(fw, flags); error = PTR_ERR(fr); if (IS_ERR(fr)) goto err_write_pipe; diff --git a/include/linux/fs.h b/include/linux/fs.h index 0e80cd717d3..4b86f806014 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1778,8 +1778,8 @@ static inline void allow_write_access(struct file *file) } extern int do_pipe(int *); extern int do_pipe_flags(int *, int); -extern struct file *create_read_pipe(struct file *f); -extern struct file *create_write_pipe(void); +extern struct file *create_read_pipe(struct file *f, int flags); +extern struct file *create_write_pipe(int flags); extern void free_write_pipe(struct file *); extern struct file *do_filp_open(int dfd, const char *pathname, diff --git a/kernel/kmod.c b/kernel/kmod.c index 90d7af1c165..2989f67c444 100644 --- a/kernel/kmod.c +++ b/kernel/kmod.c @@ -417,12 +417,12 @@ int call_usermodehelper_stdinpipe(struct subprocess_info *sub_info, { struct file *f; - f = create_write_pipe(); + f = create_write_pipe(0); if (IS_ERR(f)) return PTR_ERR(f); *filp = f; - f = create_read_pipe(f); + f = create_read_pipe(f, 0); if (IS_ERR(f)) { free_write_pipe(*filp); return PTR_ERR(f); -- cgit v1.2.3-70-g09d2 From 510df2dd482496083e1c3b1a8c9b6afd5fa4c7d7 Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Wed, 23 Jul 2008 21:29:41 -0700 Subject: flag parameters: NONBLOCK in inotify_init This patch adds non-blocking support for inotify_init1. The additional changes needed are minimal. The following test must be adjusted for architectures other than x86 and x86-64 and in case the syscall numbers changed. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ #include #include #include #include #ifndef __NR_inotify_init1 # ifdef __x86_64__ # define __NR_inotify_init1 294 # elif defined __i386__ # define __NR_inotify_init1 332 # else # error "need __NR_inotify_init1" # endif #endif #define IN_NONBLOCK O_NONBLOCK int main (void) { int fd = syscall (__NR_inotify_init1, 0); if (fd == -1) { puts ("inotify_init1(0) failed"); return 1; } int fl = fcntl (fd, F_GETFL); if (fl == -1) { puts ("fcntl failed"); return 1; } if (fl & O_NONBLOCK) { puts ("inotify_init1(0) set non-blocking mode"); return 1; } close (fd); fd = syscall (__NR_inotify_init1, IN_NONBLOCK); if (fd == -1) { puts ("inotify_init1(IN_NONBLOCK) failed"); return 1; } fl = fcntl (fd, F_GETFL); if (fl == -1) { puts ("fcntl failed"); return 1; } if ((fl & O_NONBLOCK) == 0) { puts ("inotify_init1(IN_NONBLOCK) set non-blocking mode"); return 1; } close (fd); puts ("OK"); return 0; } ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Signed-off-by: Ulrich Drepper Acked-by: Davide Libenzi Cc: Michael Kerrisk Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/inotify_user.c | 4 ++-- include/linux/inotify.h | 3 ++- 2 files changed, 4 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/inotify_user.c b/fs/inotify_user.c index 851005998cd..dc7e1f61974 100644 --- a/fs/inotify_user.c +++ b/fs/inotify_user.c @@ -574,7 +574,7 @@ asmlinkage long sys_inotify_init1(int flags) struct file *filp; int fd, ret; - if (flags & ~IN_CLOEXEC) + if (flags & ~(IN_CLOEXEC | IN_NONBLOCK)) return -EINVAL; fd = get_unused_fd_flags(flags & O_CLOEXEC); @@ -613,7 +613,7 @@ asmlinkage long sys_inotify_init1(int flags) filp->f_path.dentry = dget(inotify_mnt->mnt_root); filp->f_mapping = filp->f_path.dentry->d_inode->i_mapping; filp->f_mode = FMODE_READ; - filp->f_flags = O_RDONLY; + filp->f_flags = O_RDONLY | (flags & O_NONBLOCK); filp->private_data = dev; INIT_LIST_HEAD(&dev->events); diff --git a/include/linux/inotify.h b/include/linux/inotify.h index 72ef8212051..bd578578a8b 100644 --- a/include/linux/inotify.h +++ b/include/linux/inotify.h @@ -7,7 +7,7 @@ #ifndef _LINUX_INOTIFY_H #define _LINUX_INOTIFY_H -/* For O_CLOEXEC */ +/* For O_CLOEXEC and O_NONBLOCK */ #include #include @@ -67,6 +67,7 @@ struct inotify_event { /* Flags for sys_inotify_init1. */ #define IN_CLOEXEC O_CLOEXEC +#define IN_NONBLOCK O_NONBLOCK #ifdef __KERNEL__ -- cgit v1.2.3-70-g09d2 From e38b36f325153eaadd1c2a7abc5762079233e540 Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Wed, 23 Jul 2008 21:29:42 -0700 Subject: flag parameters: check magic constants This patch adds test that ensure the boundary conditions for the various constants introduced in the previous patches is met. No code is generated. [akpm@linux-foundation.org: fix alpha] Signed-off-by: Ulrich Drepper Acked-by: Davide Libenzi Cc: Michael Kerrisk Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/eventfd.c | 4 ++++ fs/eventpoll.c | 3 +++ fs/inotify_user.c | 4 ++++ fs/signalfd.c | 4 ++++ fs/timerfd.c | 4 ++++ net/socket.c | 6 ++++++ 6 files changed, 25 insertions(+) (limited to 'fs') diff --git a/fs/eventfd.c b/fs/eventfd.c index 3ed4466177a..08bf558d040 100644 --- a/fs/eventfd.c +++ b/fs/eventfd.c @@ -203,6 +203,10 @@ asmlinkage long sys_eventfd2(unsigned int count, int flags) int fd; struct eventfd_ctx *ctx; + /* Check the EFD_* constants for consistency. */ + BUILD_BUG_ON(EFD_CLOEXEC != O_CLOEXEC); + BUILD_BUG_ON(EFD_NONBLOCK != O_NONBLOCK); + if (flags & ~(EFD_CLOEXEC | EFD_NONBLOCK)) return -EINVAL; diff --git a/fs/eventpoll.c b/fs/eventpoll.c index 3fd4014f3c5..2fdad420404 100644 --- a/fs/eventpoll.c +++ b/fs/eventpoll.c @@ -1051,6 +1051,9 @@ asmlinkage long sys_epoll_create2(int size, int flags) int error, fd = -1; struct eventpoll *ep; + /* Check the EPOLL_* constant for consistency. */ + BUILD_BUG_ON(EPOLL_CLOEXEC != O_CLOEXEC); + if (flags & ~EPOLL_CLOEXEC) return -EINVAL; diff --git a/fs/inotify_user.c b/fs/inotify_user.c index dc7e1f61974..fe79c25d95d 100644 --- a/fs/inotify_user.c +++ b/fs/inotify_user.c @@ -574,6 +574,10 @@ asmlinkage long sys_inotify_init1(int flags) struct file *filp; int fd, ret; + /* Check the IN_* constants for consistency. */ + BUILD_BUG_ON(IN_CLOEXEC != O_CLOEXEC); + BUILD_BUG_ON(IN_NONBLOCK != O_NONBLOCK); + if (flags & ~(IN_CLOEXEC | IN_NONBLOCK)) return -EINVAL; diff --git a/fs/signalfd.c b/fs/signalfd.c index 5441a4bca77..9c39bc7f843 100644 --- a/fs/signalfd.c +++ b/fs/signalfd.c @@ -211,6 +211,10 @@ asmlinkage long sys_signalfd4(int ufd, sigset_t __user *user_mask, sigset_t sigmask; struct signalfd_ctx *ctx; + /* Check the SFD_* constants for consistency. */ + BUILD_BUG_ON(SFD_CLOEXEC != O_CLOEXEC); + BUILD_BUG_ON(SFD_NONBLOCK != O_NONBLOCK); + if (flags & ~(SFD_CLOEXEC | SFD_NONBLOCK)) return -EINVAL; diff --git a/fs/timerfd.c b/fs/timerfd.c index 75d44efe346..c502c60e4f5 100644 --- a/fs/timerfd.c +++ b/fs/timerfd.c @@ -184,6 +184,10 @@ asmlinkage long sys_timerfd_create(int clockid, int flags) int ufd; struct timerfd_ctx *ctx; + /* Check the TFD_* constants for consistency. */ + BUILD_BUG_ON(TFD_CLOEXEC != O_CLOEXEC); + BUILD_BUG_ON(TFD_NONBLOCK != O_NONBLOCK); + if (flags & ~(TFD_CLOEXEC | TFD_NONBLOCK)) return -EINVAL; if (clockid != CLOCK_MONOTONIC && diff --git a/net/socket.c b/net/socket.c index 31105f9048a..1310a82cbba 100644 --- a/net/socket.c +++ b/net/socket.c @@ -1222,6 +1222,12 @@ asmlinkage long sys_socket(int family, int type, int protocol) struct socket *sock; int flags; + /* Check the SOCK_* constants for consistency. */ + BUILD_BUG_ON(SOCK_CLOEXEC != O_CLOEXEC); + BUILD_BUG_ON((SOCK_MAX | SOCK_TYPE_MASK) != SOCK_TYPE_MASK); + BUILD_BUG_ON(SOCK_CLOEXEC & SOCK_TYPE_MASK); + BUILD_BUG_ON(SOCK_NONBLOCK & SOCK_TYPE_MASK); + flags = type & ~SOCK_TYPE_MASK; if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK)) return -EINVAL; -- cgit v1.2.3-70-g09d2 From 9fe5ad9c8cef9ad5873d8ee55d1cf00d9b607df0 Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Wed, 23 Jul 2008 21:29:43 -0700 Subject: flag parameters add-on: remove epoll_create size param Remove the size parameter from the new epoll_create syscall and renames the syscall itself. The updated test program follows. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ #include #include #include #include #include #ifndef __NR_epoll_create2 # ifdef __x86_64__ # define __NR_epoll_create2 291 # elif defined __i386__ # define __NR_epoll_create2 329 # else # error "need __NR_epoll_create2" # endif #endif #define EPOLL_CLOEXEC O_CLOEXEC int main (void) { int fd = syscall (__NR_epoll_create2, 0); if (fd == -1) { puts ("epoll_create2(0) failed"); return 1; } int coe = fcntl (fd, F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if (coe & FD_CLOEXEC) { puts ("epoll_create2(0) set close-on-exec flag"); return 1; } close (fd); fd = syscall (__NR_epoll_create2, EPOLL_CLOEXEC); if (fd == -1) { puts ("epoll_create2(EPOLL_CLOEXEC) failed"); return 1; } coe = fcntl (fd, F_GETFD); if (coe == -1) { puts ("fcntl failed"); return 1; } if ((coe & FD_CLOEXEC) == 0) { puts ("epoll_create2(EPOLL_CLOEXEC) set close-on-exec flag"); return 1; } close (fd); puts ("OK"); return 0; } ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Signed-off-by: Ulrich Drepper Acked-by: Davide Libenzi Cc: Michael Kerrisk Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/x86/ia32/ia32entry.S | 2 +- arch/x86/kernel/syscall_table_32.S | 2 +- fs/eventpoll.c | 18 ++++++++++-------- include/asm-x86/unistd_32.h | 2 +- include/asm-x86/unistd_64.h | 4 ++-- include/linux/eventpoll.h | 2 +- include/linux/syscalls.h | 2 +- 7 files changed, 17 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/arch/x86/ia32/ia32entry.S b/arch/x86/ia32/ia32entry.S index 4541073dd83..e4bd1793a5e 100644 --- a/arch/x86/ia32/ia32entry.S +++ b/arch/x86/ia32/ia32entry.S @@ -828,7 +828,7 @@ ia32_sys_call_table: .quad compat_sys_timerfd_gettime .quad compat_sys_signalfd4 .quad sys_eventfd2 - .quad sys_epoll_create2 + .quad sys_epoll_create1 .quad sys_dup3 /* 330 */ .quad sys_pipe2 .quad sys_inotify_init1 diff --git a/arch/x86/kernel/syscall_table_32.S b/arch/x86/kernel/syscall_table_32.S index f59aba5ff0f..d44395ff34c 100644 --- a/arch/x86/kernel/syscall_table_32.S +++ b/arch/x86/kernel/syscall_table_32.S @@ -328,7 +328,7 @@ ENTRY(sys_call_table) .long sys_timerfd_gettime .long sys_signalfd4 .long sys_eventfd2 - .long sys_epoll_create2 + .long sys_epoll_create1 .long sys_dup3 /* 330 */ .long sys_pipe2 .long sys_inotify_init1 diff --git a/fs/eventpoll.c b/fs/eventpoll.c index 2fdad420404..0c87474f791 100644 --- a/fs/eventpoll.c +++ b/fs/eventpoll.c @@ -1046,7 +1046,7 @@ retry: * RB tree. With the current implementation, the "size" parameter is ignored * (besides sanity checks). */ -asmlinkage long sys_epoll_create2(int size, int flags) +asmlinkage long sys_epoll_create1(int flags) { int error, fd = -1; struct eventpoll *ep; @@ -1058,14 +1058,13 @@ asmlinkage long sys_epoll_create2(int size, int flags) return -EINVAL; DNPRINTK(3, (KERN_INFO "[%p] eventpoll: sys_epoll_create(%d)\n", - current, size)); + current, flags)); /* - * Sanity check on the size parameter, and create the internal data - * structure ( "struct eventpoll" ). + * Create the internal data structure ( "struct eventpoll" ). */ - error = -EINVAL; - if (size <= 0 || (error = ep_alloc(&ep)) < 0) { + error = ep_alloc(&ep); + if (error < 0) { fd = error; goto error_return; } @@ -1081,14 +1080,17 @@ asmlinkage long sys_epoll_create2(int size, int flags) error_return: DNPRINTK(3, (KERN_INFO "[%p] eventpoll: sys_epoll_create(%d) = %d\n", - current, size, fd)); + current, flags, fd)); return fd; } asmlinkage long sys_epoll_create(int size) { - return sys_epoll_create2(size, 0); + if (size < 0) + return -EINVAL; + + return sys_epoll_create1(0); } /* diff --git a/include/asm-x86/unistd_32.h b/include/asm-x86/unistd_32.h index b3daf503ab9..d7394673b77 100644 --- a/include/asm-x86/unistd_32.h +++ b/include/asm-x86/unistd_32.h @@ -334,7 +334,7 @@ #define __NR_timerfd_gettime 326 #define __NR_signalfd4 327 #define __NR_eventfd2 328 -#define __NR_epoll_create2 329 +#define __NR_epoll_create1 329 #define __NR_dup3 330 #define __NR_pipe2 331 #define __NR_inotify_init1 332 diff --git a/include/asm-x86/unistd_64.h b/include/asm-x86/unistd_64.h index c8cb88d70c6..3a341d79179 100644 --- a/include/asm-x86/unistd_64.h +++ b/include/asm-x86/unistd_64.h @@ -645,8 +645,8 @@ __SYSCALL(__NR_paccept, sys_paccept) __SYSCALL(__NR_signalfd4, sys_signalfd4) #define __NR_eventfd2 290 __SYSCALL(__NR_eventfd2, sys_eventfd2) -#define __NR_epoll_create2 291 -__SYSCALL(__NR_epoll_create2, sys_epoll_create2) +#define __NR_epoll_create1 291 +__SYSCALL(__NR_epoll_create1, sys_epoll_create1) #define __NR_dup3 292 __SYSCALL(__NR_dup3, sys_dup3) #define __NR_pipe2 293 diff --git a/include/linux/eventpoll.h b/include/linux/eventpoll.h index 1cfaa40059c..f1e1d3c4712 100644 --- a/include/linux/eventpoll.h +++ b/include/linux/eventpoll.h @@ -18,7 +18,7 @@ #include #include -/* Flags for epoll_create2. */ +/* Flags for epoll_create1. */ #define EPOLL_CLOEXEC O_CLOEXEC /* Valid opcodes to issue to sys_epoll_ctl() */ diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h index 93a7e7f017a..06f2bf76c03 100644 --- a/include/linux/syscalls.h +++ b/include/linux/syscalls.h @@ -431,7 +431,7 @@ asmlinkage long sys_poll(struct pollfd __user *ufds, unsigned int nfds, asmlinkage long sys_select(int n, fd_set __user *inp, fd_set __user *outp, fd_set __user *exp, struct timeval __user *tvp); asmlinkage long sys_epoll_create(int size); -asmlinkage long sys_epoll_create2(int size, int flags); +asmlinkage long sys_epoll_create1(int flags); asmlinkage long sys_epoll_ctl(int epfd, int op, int fd, struct epoll_event __user *event); asmlinkage long sys_epoll_wait(int epfd, struct epoll_event __user *events, -- cgit v1.2.3-70-g09d2 From 746f1e558bc52b9693c1a1ecdab60f8392e5ff18 Mon Sep 17 00:00:00 2001 From: Michael Halcrow Date: Wed, 23 Jul 2008 21:30:02 -0700 Subject: eCryptfs: Privileged kthread for lower file opens eCryptfs would really like to have read-write access to all files in the lower filesystem. Right now, the persistent lower file may be opened read-only if the attempt to open it read-write fails. One way to keep from having to do that is to have a privileged kthread that can open the lower persistent file on behalf of the user opening the eCryptfs file; this patch implements this functionality. This patch will properly allow a less-privileged user to open the eCryptfs file, followed by a more-privileged user opening the eCryptfs file, with the first user only being able to read and the second user being able to both read and write. eCryptfs currently does this wrong; it will wind up calling vfs_write() on a file that was opened read-only. This is fixed in this patch. Signed-off-by: Michael Halcrow Cc: Dave Kleikamp Cc: Serge Hallyn Cc: Eric Sandeen Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ecryptfs/Makefile | 2 +- fs/ecryptfs/ecryptfs_kernel.h | 19 ++++ fs/ecryptfs/file.c | 7 ++ fs/ecryptfs/kthread.c | 203 ++++++++++++++++++++++++++++++++++++++++++ fs/ecryptfs/main.c | 42 ++++----- 5 files changed, 251 insertions(+), 22 deletions(-) create mode 100644 fs/ecryptfs/kthread.c (limited to 'fs') diff --git a/fs/ecryptfs/Makefile b/fs/ecryptfs/Makefile index 1e34a7fd488..b4755a85996 100644 --- a/fs/ecryptfs/Makefile +++ b/fs/ecryptfs/Makefile @@ -4,4 +4,4 @@ obj-$(CONFIG_ECRYPT_FS) += ecryptfs.o -ecryptfs-objs := dentry.o file.o inode.o main.o super.o mmap.o read_write.o crypto.o keystore.o messaging.o netlink.o miscdev.o debug.o +ecryptfs-objs := dentry.o file.o inode.o main.o super.o mmap.o read_write.o crypto.o keystore.o messaging.o netlink.o miscdev.o kthread.o debug.o diff --git a/fs/ecryptfs/ecryptfs_kernel.h b/fs/ecryptfs/ecryptfs_kernel.h index c15c25745e0..b4a0cccfdd7 100644 --- a/fs/ecryptfs/ecryptfs_kernel.h +++ b/fs/ecryptfs/ecryptfs_kernel.h @@ -559,6 +559,20 @@ extern struct kmem_cache *ecryptfs_key_record_cache; extern struct kmem_cache *ecryptfs_key_sig_cache; extern struct kmem_cache *ecryptfs_global_auth_tok_cache; extern struct kmem_cache *ecryptfs_key_tfm_cache; +extern struct kmem_cache *ecryptfs_open_req_cache; + +struct ecryptfs_open_req { +#define ECRYPTFS_REQ_PROCESSED 0x00000001 +#define ECRYPTFS_REQ_DROPPED 0x00000002 +#define ECRYPTFS_REQ_ZOMBIE 0x00000004 + u32 flags; + struct file **lower_file; + struct dentry *lower_dentry; + struct vfsmount *lower_mnt; + wait_queue_head_t wait; + struct mutex mux; + struct list_head kthread_ctl_list; +}; int ecryptfs_interpose(struct dentry *hidden_dentry, struct dentry *this_dentry, struct super_block *sb, @@ -690,5 +704,10 @@ void ecryptfs_msg_ctx_alloc_to_free(struct ecryptfs_msg_ctx *msg_ctx); int ecryptfs_spawn_daemon(struct ecryptfs_daemon **daemon, uid_t euid, struct user_namespace *user_ns, struct pid *pid); +int ecryptfs_init_kthread(void); +void ecryptfs_destroy_kthread(void); +int ecryptfs_privileged_open(struct file **lower_file, + struct dentry *lower_dentry, + struct vfsmount *lower_mnt); #endif /* #ifndef ECRYPTFS_KERNEL_H */ diff --git a/fs/ecryptfs/file.c b/fs/ecryptfs/file.c index 24749bf0668..f0be2905152 100644 --- a/fs/ecryptfs/file.c +++ b/fs/ecryptfs/file.c @@ -192,6 +192,13 @@ static int ecryptfs_open(struct inode *inode, struct file *file) | ECRYPTFS_ENCRYPTED); } mutex_unlock(&crypt_stat->cs_mutex); + if ((ecryptfs_inode_to_private(inode)->lower_file->f_flags & O_RDONLY) + && !(file->f_flags & O_RDONLY)) { + rc = -EPERM; + printk(KERN_WARNING "%s: Lower persistent file is RO; eCryptfs " + "file must hence be opened RO\n", __func__); + goto out; + } ecryptfs_set_file_lower( file, ecryptfs_inode_to_private(inode)->lower_file); if (S_ISDIR(ecryptfs_dentry->d_inode->i_mode)) { diff --git a/fs/ecryptfs/kthread.c b/fs/ecryptfs/kthread.c new file mode 100644 index 00000000000..c440c6b58b2 --- /dev/null +++ b/fs/ecryptfs/kthread.c @@ -0,0 +1,203 @@ +/** + * eCryptfs: Linux filesystem encryption layer + * + * Copyright (C) 2008 International Business Machines Corp. + * Author(s): Michael A. Halcrow + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation; either version 2 of the + * License, or (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA + * 02111-1307, USA. + */ + +#include +#include +#include +#include +#include "ecryptfs_kernel.h" + +struct kmem_cache *ecryptfs_open_req_cache; + +static struct ecryptfs_kthread_ctl { +#define ECRYPTFS_KTHREAD_ZOMBIE 0x00000001 + u32 flags; + struct mutex mux; + struct list_head req_list; + wait_queue_head_t wait; +} ecryptfs_kthread_ctl; + +static struct task_struct *ecryptfs_kthread; + +/** + * ecryptfs_threadfn + * @ignored: ignored + * + * The eCryptfs kernel thread that has the responsibility of getting + * the lower persistent file with RW permissions. + * + * Returns zero on success; non-zero otherwise + */ +static int ecryptfs_threadfn(void *ignored) +{ + set_freezable(); + while (1) { + struct ecryptfs_open_req *req; + + wait_event_freezable( + ecryptfs_kthread_ctl.wait, + (!list_empty(&ecryptfs_kthread_ctl.req_list) + || kthread_should_stop())); + mutex_lock(&ecryptfs_kthread_ctl.mux); + if (ecryptfs_kthread_ctl.flags & ECRYPTFS_KTHREAD_ZOMBIE) { + mutex_unlock(&ecryptfs_kthread_ctl.mux); + goto out; + } + while (!list_empty(&ecryptfs_kthread_ctl.req_list)) { + req = list_first_entry(&ecryptfs_kthread_ctl.req_list, + struct ecryptfs_open_req, + kthread_ctl_list); + mutex_lock(&req->mux); + list_del(&req->kthread_ctl_list); + if (!(req->flags & ECRYPTFS_REQ_ZOMBIE)) { + dget(req->lower_dentry); + mntget(req->lower_mnt); + (*req->lower_file) = dentry_open( + req->lower_dentry, req->lower_mnt, + (O_RDWR | O_LARGEFILE)); + req->flags |= ECRYPTFS_REQ_PROCESSED; + } + wake_up(&req->wait); + mutex_unlock(&req->mux); + } + mutex_unlock(&ecryptfs_kthread_ctl.mux); + } +out: + return 0; +} + +int ecryptfs_init_kthread(void) +{ + int rc = 0; + + mutex_init(&ecryptfs_kthread_ctl.mux); + init_waitqueue_head(&ecryptfs_kthread_ctl.wait); + INIT_LIST_HEAD(&ecryptfs_kthread_ctl.req_list); + ecryptfs_kthread = kthread_run(&ecryptfs_threadfn, NULL, + "ecryptfs-kthread"); + if (IS_ERR(ecryptfs_kthread)) { + rc = PTR_ERR(ecryptfs_kthread); + printk(KERN_ERR "%s: Failed to create kernel thread; rc = [%d]" + "\n", __func__, rc); + } + return rc; +} + +void ecryptfs_destroy_kthread(void) +{ + struct ecryptfs_open_req *req; + + mutex_lock(&ecryptfs_kthread_ctl.mux); + ecryptfs_kthread_ctl.flags |= ECRYPTFS_KTHREAD_ZOMBIE; + list_for_each_entry(req, &ecryptfs_kthread_ctl.req_list, + kthread_ctl_list) { + mutex_lock(&req->mux); + req->flags |= ECRYPTFS_REQ_ZOMBIE; + wake_up(&req->wait); + mutex_unlock(&req->mux); + } + mutex_unlock(&ecryptfs_kthread_ctl.mux); + kthread_stop(ecryptfs_kthread); + wake_up(&ecryptfs_kthread_ctl.wait); +} + +/** + * ecryptfs_privileged_open + * @lower_file: Result of dentry_open by root on lower dentry + * @lower_dentry: Lower dentry for file to open + * @lower_mnt: Lower vfsmount for file to open + * + * This function gets a r/w file opened againt the lower dentry. + * + * Returns zero on success; non-zero otherwise + */ +int ecryptfs_privileged_open(struct file **lower_file, + struct dentry *lower_dentry, + struct vfsmount *lower_mnt) +{ + struct ecryptfs_open_req *req; + int rc = 0; + + /* Corresponding dput() and mntput() are done when the + * persistent file is fput() when the eCryptfs inode is + * destroyed. */ + dget(lower_dentry); + mntget(lower_mnt); + (*lower_file) = dentry_open(lower_dentry, lower_mnt, + (O_RDWR | O_LARGEFILE)); + if (!IS_ERR(*lower_file)) + goto out; + req = kmem_cache_alloc(ecryptfs_open_req_cache, GFP_KERNEL); + if (!req) { + rc = -ENOMEM; + goto out; + } + mutex_init(&req->mux); + req->lower_file = lower_file; + req->lower_dentry = lower_dentry; + req->lower_mnt = lower_mnt; + init_waitqueue_head(&req->wait); + req->flags = 0; + mutex_lock(&ecryptfs_kthread_ctl.mux); + if (ecryptfs_kthread_ctl.flags & ECRYPTFS_KTHREAD_ZOMBIE) { + rc = -EIO; + mutex_unlock(&ecryptfs_kthread_ctl.mux); + printk(KERN_ERR "%s: We are in the middle of shutting down; " + "aborting privileged request to open lower file\n", + __func__); + goto out_free; + } + list_add_tail(&req->kthread_ctl_list, &ecryptfs_kthread_ctl.req_list); + mutex_unlock(&ecryptfs_kthread_ctl.mux); + wake_up(&ecryptfs_kthread_ctl.wait); + wait_event(req->wait, (req->flags != 0)); + mutex_lock(&req->mux); + BUG_ON(req->flags == 0); + if (req->flags & ECRYPTFS_REQ_DROPPED + || req->flags & ECRYPTFS_REQ_ZOMBIE) { + rc = -EIO; + printk(KERN_WARNING "%s: Privileged open request dropped\n", + __func__); + goto out_unlock; + } + if (IS_ERR(*req->lower_file)) { + rc = PTR_ERR(*req->lower_file); + dget(lower_dentry); + mntget(lower_mnt); + (*lower_file) = dentry_open(lower_dentry, lower_mnt, + (O_RDONLY | O_LARGEFILE)); + if (IS_ERR(*lower_file)) { + rc = PTR_ERR(*req->lower_file); + (*lower_file) = NULL; + printk(KERN_WARNING "%s: Error attempting privileged " + "open of lower file with either RW or RO " + "perms; rc = [%d]. Giving up.\n", + __func__, rc); + } + } +out_unlock: + mutex_unlock(&req->mux); +out_free: + kmem_cache_free(ecryptfs_open_req_cache, req); +out: + return rc; +} diff --git a/fs/ecryptfs/main.c b/fs/ecryptfs/main.c index d603631601e..f36ab2feea2 100644 --- a/fs/ecryptfs/main.c +++ b/fs/ecryptfs/main.c @@ -130,26 +130,12 @@ static int ecryptfs_init_persistent_file(struct dentry *ecryptfs_dentry) ecryptfs_dentry_to_lower_mnt(ecryptfs_dentry); lower_dentry = ecryptfs_dentry_to_lower(ecryptfs_dentry); - /* Corresponding dput() and mntput() are done when the - * persistent file is fput() when the eCryptfs inode - * is destroyed. */ - dget(lower_dentry); - mntget(lower_mnt); - inode_info->lower_file = dentry_open(lower_dentry, - lower_mnt, - (O_RDWR | O_LARGEFILE)); - if (IS_ERR(inode_info->lower_file)) { - dget(lower_dentry); - mntget(lower_mnt); - inode_info->lower_file = dentry_open(lower_dentry, - lower_mnt, - (O_RDONLY - | O_LARGEFILE)); - } - if (IS_ERR(inode_info->lower_file)) { + rc = ecryptfs_privileged_open(&inode_info->lower_file, + lower_dentry, lower_mnt); + if (rc || IS_ERR(inode_info->lower_file)) { printk(KERN_ERR "Error opening lower persistent file " - "for lower_dentry [0x%p] and lower_mnt [0x%p]\n", - lower_dentry, lower_mnt); + "for lower_dentry [0x%p] and lower_mnt [0x%p]; " + "rc = [%d]\n", lower_dentry, lower_mnt, rc); rc = PTR_ERR(inode_info->lower_file); inode_info->lower_file = NULL; } @@ -679,6 +665,11 @@ static struct ecryptfs_cache_info { .name = "ecryptfs_key_tfm_cache", .size = sizeof(struct ecryptfs_key_tfm), }, + { + .cache = &ecryptfs_open_req_cache, + .name = "ecryptfs_open_req_cache", + .size = sizeof(struct ecryptfs_open_req), + }, }; static void ecryptfs_free_kmem_caches(void) @@ -795,11 +786,17 @@ static int __init ecryptfs_init(void) printk(KERN_ERR "sysfs registration failed\n"); goto out_unregister_filesystem; } + rc = ecryptfs_init_kthread(); + if (rc) { + printk(KERN_ERR "%s: kthread initialization failed; " + "rc = [%d]\n", __func__, rc); + goto out_do_sysfs_unregistration; + } rc = ecryptfs_init_messaging(ecryptfs_transport); if (rc) { - ecryptfs_printk(KERN_ERR, "Failure occured while attempting to " + printk(KERN_ERR "Failure occured while attempting to " "initialize the eCryptfs netlink socket\n"); - goto out_do_sysfs_unregistration; + goto out_destroy_kthread; } rc = ecryptfs_init_crypto(); if (rc) { @@ -814,6 +811,8 @@ static int __init ecryptfs_init(void) goto out; out_release_messaging: ecryptfs_release_messaging(ecryptfs_transport); +out_destroy_kthread: + ecryptfs_destroy_kthread(); out_do_sysfs_unregistration: do_sysfs_unregistration(); out_unregister_filesystem: @@ -833,6 +832,7 @@ static void __exit ecryptfs_exit(void) printk(KERN_ERR "Failure whilst attempting to destroy crypto; " "rc = [%d]\n", rc); ecryptfs_release_messaging(ecryptfs_transport); + ecryptfs_destroy_kthread(); do_sysfs_unregistration(); unregister_filesystem(&ecryptfs_fs_type); ecryptfs_free_kmem_caches(); -- cgit v1.2.3-70-g09d2 From 6c4c17b073cd4a5a61bc04329561632870bb21fc Mon Sep 17 00:00:00 2001 From: Tyler Hicks Date: Wed, 23 Jul 2008 21:30:04 -0700 Subject: ecryptfs: discard ecryptfsd registration messages in miscdev The userspace eCryptfs daemon sends HELO and QUIT messages to the kernel for per-user daemon (un)registration. These messages are required when netlink is used as the transport, but (un)registration is handled by opening and closing the device file when miscdev is the transport. These messages should be discarded in the miscdev transport so that a daemon isn't registered twice. Signed-off-by: Tyler Hicks Cc: Michael Halcrow Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ecryptfs/miscdev.c | 59 --------------------------------------------------- 1 file changed, 59 deletions(-) (limited to 'fs') diff --git a/fs/ecryptfs/miscdev.c b/fs/ecryptfs/miscdev.c index 09a4522f65e..b484792a099 100644 --- a/fs/ecryptfs/miscdev.c +++ b/fs/ecryptfs/miscdev.c @@ -357,46 +357,6 @@ out_unlock_daemon: return rc; } -/** - * ecryptfs_miscdev_helo - * @euid: effective user id of miscdevess sending helo packet - * @user_ns: The namespace in which @euid applies - * @pid: miscdevess id of miscdevess sending helo packet - * - * Returns zero on success; non-zero otherwise - */ -static int ecryptfs_miscdev_helo(uid_t euid, struct user_namespace *user_ns, - struct pid *pid) -{ - int rc; - - rc = ecryptfs_process_helo(ECRYPTFS_TRANSPORT_MISCDEV, euid, user_ns, - pid); - if (rc) - printk(KERN_WARNING "Error processing HELO; rc = [%d]\n", rc); - return rc; -} - -/** - * ecryptfs_miscdev_quit - * @euid: effective user id of miscdevess sending quit packet - * @user_ns: The namespace in which @euid applies - * @pid: miscdevess id of miscdevess sending quit packet - * - * Returns zero on success; non-zero otherwise - */ -static int ecryptfs_miscdev_quit(uid_t euid, struct user_namespace *user_ns, - struct pid *pid) -{ - int rc; - - rc = ecryptfs_process_quit(euid, user_ns, pid); - if (rc) - printk(KERN_WARNING - "Error processing QUIT message; rc = [%d]\n", rc); - return rc; -} - /** * ecryptfs_miscdev_response - miscdevess response to message previously sent to daemon * @data: Bytes comprising struct ecryptfs_message @@ -512,26 +472,7 @@ ecryptfs_miscdev_write(struct file *file, const char __user *buf, __func__, rc); break; case ECRYPTFS_MSG_HELO: - rc = ecryptfs_miscdev_helo(current->euid, - current->nsproxy->user_ns, - task_pid(current)); - if (rc) { - printk(KERN_ERR "%s: Error attempting to process " - "helo from pid [0x%p]; rc = [%d]\n", __func__, - task_pid(current), rc); - goto out_free; - } - break; case ECRYPTFS_MSG_QUIT: - rc = ecryptfs_miscdev_quit(current->euid, - current->nsproxy->user_ns, - task_pid(current)); - if (rc) { - printk(KERN_ERR "%s: Error attempting to process " - "quit from pid [0x%p]; rc = [%d]\n", __func__, - task_pid(current), rc); - goto out_free; - } break; default: ecryptfs_printk(KERN_WARNING, "Dropping miscdev " -- cgit v1.2.3-70-g09d2 From 982363c97f8cad7aea4c3d2cfebffc1cc2d2f166 Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Wed, 23 Jul 2008 21:30:04 -0700 Subject: ecryptfs: propagate key errors up at mount time Mounting with invalid key signatures should probably fail, if they were specifically requested but not available. Also fix case checks in process_request_key_err() for the right sign of the errnos, as spotted by Jan Tluka. Signed-off-by: Eric Sandeen Reviewed-by: Jan Tluka Acked-by: Michael Halcrow Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ecryptfs/keystore.c | 9 ++++----- fs/ecryptfs/main.c | 4 ++-- 2 files changed, 6 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/ecryptfs/keystore.c b/fs/ecryptfs/keystore.c index e82b457180b..f5b76a331b9 100644 --- a/fs/ecryptfs/keystore.c +++ b/fs/ecryptfs/keystore.c @@ -44,15 +44,15 @@ static int process_request_key_err(long err_code) int rc = 0; switch (err_code) { - case ENOKEY: + case -ENOKEY: ecryptfs_printk(KERN_WARNING, "No key\n"); rc = -ENOENT; break; - case EKEYEXPIRED: + case -EKEYEXPIRED: ecryptfs_printk(KERN_WARNING, "Key expired\n"); rc = -ETIME; break; - case EKEYREVOKED: + case -EKEYREVOKED: ecryptfs_printk(KERN_WARNING, "Key revoked\n"); rc = -EINVAL; break; @@ -963,8 +963,7 @@ int ecryptfs_keyring_auth_tok_for_sig(struct key **auth_tok_key, if (!(*auth_tok_key) || IS_ERR(*auth_tok_key)) { printk(KERN_ERR "Could not find key with description: [%s]\n", sig); - process_request_key_err(PTR_ERR(*auth_tok_key)); - rc = -EINVAL; + rc = process_request_key_err(PTR_ERR(*auth_tok_key)); goto out; } (*auth_tok) = ecryptfs_get_key_payload_data(*auth_tok_key); diff --git a/fs/ecryptfs/main.c b/fs/ecryptfs/main.c index f36ab2feea2..8876fe7c76e 100644 --- a/fs/ecryptfs/main.c +++ b/fs/ecryptfs/main.c @@ -248,10 +248,11 @@ static int ecryptfs_init_global_auth_toks( "session keyring for sig specified in mount " "option: [%s]\n", global_auth_tok->sig); global_auth_tok->flags |= ECRYPTFS_AUTH_TOK_INVALID; - rc = 0; + goto out; } else global_auth_tok->flags &= ~ECRYPTFS_AUTH_TOK_INVALID; } +out: return rc; } @@ -416,7 +417,6 @@ static int ecryptfs_parse_options(struct super_block *sb, char *options) printk(KERN_WARNING "One or more global auth toks could not " "properly register; rc = [%d]\n", rc); } - rc = 0; out: return rc; } -- cgit v1.2.3-70-g09d2 From 8f2368095e25018838e1bf145041f58270ccd32e Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Wed, 23 Jul 2008 21:30:05 -0700 Subject: ecryptfs: string copy cleanup Clean up overcomplicated string copy, which also gets rid of this bogus warning: fs/ecryptfs/main.c: In function 'ecryptfs_parse_options': include/asm/arch/string_32.h:75: warning: array subscript is above array bounds Signed-off-by: Miklos Szeredi Cc: Michael Halcrow Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ecryptfs/main.c | 18 ++++++------------ 1 file changed, 6 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/ecryptfs/main.c b/fs/ecryptfs/main.c index 8876fe7c76e..10475d93ff5 100644 --- a/fs/ecryptfs/main.c +++ b/fs/ecryptfs/main.c @@ -301,7 +301,6 @@ static int ecryptfs_parse_options(struct super_block *sb, char *options) char *cipher_name_dst; char *cipher_name_src; char *cipher_key_bytes_src; - int cipher_name_len; if (!options) { rc = -EINVAL; @@ -382,17 +381,12 @@ static int ecryptfs_parse_options(struct super_block *sb, char *options) goto out; } if (!cipher_name_set) { - cipher_name_len = strlen(ECRYPTFS_DEFAULT_CIPHER); - if (unlikely(cipher_name_len - >= ECRYPTFS_MAX_CIPHER_NAME_SIZE)) { - rc = -EINVAL; - BUG(); - goto out; - } - memcpy(mount_crypt_stat->global_default_cipher_name, - ECRYPTFS_DEFAULT_CIPHER, cipher_name_len); - mount_crypt_stat->global_default_cipher_name[cipher_name_len] - = '\0'; + int cipher_name_len = strlen(ECRYPTFS_DEFAULT_CIPHER); + + BUG_ON(cipher_name_len >= ECRYPTFS_MAX_CIPHER_NAME_SIZE); + + strcpy(mount_crypt_stat->global_default_cipher_name, + ECRYPTFS_DEFAULT_CIPHER); } if (!cipher_key_bytes_set) { mount_crypt_stat->global_default_cipher_key_size = 0; -- cgit v1.2.3-70-g09d2 From 29335c6a41568d4708d4ec3b9187f9b6d302e5ea Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 23 Jul 2008 21:30:06 -0700 Subject: ecryptfs: crypto.c use unaligned byteorder helpers Fixes the following sparse warnings: fs/ecryptfs/crypto.c:1036:8: warning: cast to restricted __be32 fs/ecryptfs/crypto.c:1038:8: warning: cast to restricted __be32 fs/ecryptfs/crypto.c:1077:10: warning: cast to restricted __be32 fs/ecryptfs/crypto.c:1103:6: warning: incorrect type in assignment (different base types) fs/ecryptfs/crypto.c:1105:6: warning: incorrect type in assignment (different base types) fs/ecryptfs/crypto.c:1124:8: warning: incorrect type in assignment (different base types) fs/ecryptfs/crypto.c:1241:21: warning: incorrect type in assignment (different base types) fs/ecryptfs/crypto.c:1244:30: warning: incorrect type in assignment (different base types) fs/ecryptfs/crypto.c:1414:23: warning: cast to restricted __be32 fs/ecryptfs/crypto.c:1417:32: warning: cast to restricted __be16 Signed-off-by: Harvey Harrison Cc: Michael Halcrow Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ecryptfs/crypto.c | 37 ++++++++++++++----------------------- 1 file changed, 14 insertions(+), 23 deletions(-) (limited to 'fs') diff --git a/fs/ecryptfs/crypto.c b/fs/ecryptfs/crypto.c index e2832bc7869..7b99917ffad 100644 --- a/fs/ecryptfs/crypto.c +++ b/fs/ecryptfs/crypto.c @@ -33,6 +33,7 @@ #include #include #include +#include #include "ecryptfs_kernel.h" static int @@ -1032,10 +1033,8 @@ static int contains_ecryptfs_marker(char *data) { u32 m_1, m_2; - memcpy(&m_1, data, 4); - m_1 = be32_to_cpu(m_1); - memcpy(&m_2, (data + 4), 4); - m_2 = be32_to_cpu(m_2); + m_1 = get_unaligned_be32(data); + m_2 = get_unaligned_be32(data + 4); if ((m_1 ^ MAGIC_ECRYPTFS_MARKER) == m_2) return 1; ecryptfs_printk(KERN_DEBUG, "m_1 = [0x%.8x]; m_2 = [0x%.8x]; " @@ -1073,8 +1072,7 @@ static int ecryptfs_process_flags(struct ecryptfs_crypt_stat *crypt_stat, int i; u32 flags; - memcpy(&flags, page_virt, 4); - flags = be32_to_cpu(flags); + flags = get_unaligned_be32(page_virt); for (i = 0; i < ((sizeof(ecryptfs_flag_map) / sizeof(struct ecryptfs_flag_map_elem))); i++) if (flags & ecryptfs_flag_map[i].file_flag) { @@ -1100,11 +1098,9 @@ static void write_ecryptfs_marker(char *page_virt, size_t *written) get_random_bytes(&m_1, (MAGIC_ECRYPTFS_MARKER_SIZE_BYTES / 2)); m_2 = (m_1 ^ MAGIC_ECRYPTFS_MARKER); - m_1 = cpu_to_be32(m_1); - memcpy(page_virt, &m_1, (MAGIC_ECRYPTFS_MARKER_SIZE_BYTES / 2)); - m_2 = cpu_to_be32(m_2); - memcpy(page_virt + (MAGIC_ECRYPTFS_MARKER_SIZE_BYTES / 2), &m_2, - (MAGIC_ECRYPTFS_MARKER_SIZE_BYTES / 2)); + put_unaligned_be32(m_1, page_virt); + page_virt += (MAGIC_ECRYPTFS_MARKER_SIZE_BYTES / 2); + put_unaligned_be32(m_2, page_virt); (*written) = MAGIC_ECRYPTFS_MARKER_SIZE_BYTES; } @@ -1121,8 +1117,7 @@ write_ecryptfs_flags(char *page_virt, struct ecryptfs_crypt_stat *crypt_stat, flags |= ecryptfs_flag_map[i].file_flag; /* Version is in top 8 bits of the 32-bit flag vector */ flags |= ((((u8)crypt_stat->file_version) << 24) & 0xFF000000); - flags = cpu_to_be32(flags); - memcpy(page_virt, &flags, 4); + put_unaligned_be32(flags, page_virt); (*written) = 4; } @@ -1238,11 +1233,9 @@ ecryptfs_write_header_metadata(char *virt, num_header_extents_at_front = (u16)(crypt_stat->num_header_bytes_at_front / crypt_stat->extent_size); - header_extent_size = cpu_to_be32(header_extent_size); - memcpy(virt, &header_extent_size, 4); + put_unaligned_be32(header_extent_size, virt); virt += 4; - num_header_extents_at_front = cpu_to_be16(num_header_extents_at_front); - memcpy(virt, &num_header_extents_at_front, 2); + put_unaligned_be16(num_header_extents_at_front, virt); (*written) = 6; } @@ -1410,15 +1403,13 @@ static int parse_header_metadata(struct ecryptfs_crypt_stat *crypt_stat, u32 header_extent_size; u16 num_header_extents_at_front; - memcpy(&header_extent_size, virt, sizeof(u32)); - header_extent_size = be32_to_cpu(header_extent_size); - virt += sizeof(u32); - memcpy(&num_header_extents_at_front, virt, sizeof(u16)); - num_header_extents_at_front = be16_to_cpu(num_header_extents_at_front); + header_extent_size = get_unaligned_be32(virt); + virt += sizeof(__be32); + num_header_extents_at_front = get_unaligned_be16(virt); crypt_stat->num_header_bytes_at_front = (((size_t)num_header_extents_at_front * (size_t)header_extent_size)); - (*bytes_read) = (sizeof(u32) + sizeof(u16)); + (*bytes_read) = (sizeof(__be32) + sizeof(__be16)); if ((validate_header_size == ECRYPTFS_VALIDATE_HEADER_SIZE) && (crypt_stat->num_header_bytes_at_front < ECRYPTFS_MINIMUM_HEADER_EXTENT_SIZE)) { -- cgit v1.2.3-70-g09d2 From 0a688ad713949643e201431d3f4a4ceddfeb70ca Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 23 Jul 2008 21:30:07 -0700 Subject: ecryptfs: inode.c mmap.c use unaligned byteorder helpers Fixe sparse warnings: fs/ecryptfs/inode.c:368:15: warning: cast to restricted __be64 fs/ecryptfs/mmap.c:385:12: warning: incorrect type in assignment (different base types) fs/ecryptfs/mmap.c:385:12: expected unsigned long long [unsigned] [assigned] [usertype] file_size fs/ecryptfs/mmap.c:385:12: got restricted __be64 [usertype] fs/ecryptfs/mmap.c:428:12: warning: incorrect type in assignment (different base types) fs/ecryptfs/mmap.c:428:12: expected unsigned long long [unsigned] [assigned] [usertype] file_size fs/ecryptfs/mmap.c:428:12: got restricted __be64 [usertype] Signed-off-by: Harvey Harrison Cc: Michael Halcrow Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ecryptfs/inode.c | 4 ++-- fs/ecryptfs/mmap.c | 11 +++-------- 2 files changed, 5 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/ecryptfs/inode.c b/fs/ecryptfs/inode.c index c92cc1c00aa..7315547193e 100644 --- a/fs/ecryptfs/inode.c +++ b/fs/ecryptfs/inode.c @@ -31,6 +31,7 @@ #include #include #include +#include #include "ecryptfs_kernel.h" static struct dentry *lock_parent(struct dentry *dentry) @@ -364,8 +365,7 @@ static struct dentry *ecryptfs_lookup(struct inode *dir, struct dentry *dentry, else file_size = i_size_read(lower_dentry->d_inode); } else { - memcpy(&file_size, page_virt, sizeof(file_size)); - file_size = be64_to_cpu(file_size); + file_size = get_unaligned_be64(page_virt); } i_size_write(dentry->d_inode, (loff_t)file_size); kmem_cache_free(ecryptfs_header_cache_2, page_virt); diff --git a/fs/ecryptfs/mmap.c b/fs/ecryptfs/mmap.c index 2b6fe1e6e8b..245c2dc02d5 100644 --- a/fs/ecryptfs/mmap.c +++ b/fs/ecryptfs/mmap.c @@ -32,6 +32,7 @@ #include #include #include +#include #include "ecryptfs_kernel.h" /** @@ -372,7 +373,6 @@ out: */ static int ecryptfs_write_inode_size_to_header(struct inode *ecryptfs_inode) { - u64 file_size; char *file_size_virt; int rc; @@ -381,9 +381,7 @@ static int ecryptfs_write_inode_size_to_header(struct inode *ecryptfs_inode) rc = -ENOMEM; goto out; } - file_size = (u64)i_size_read(ecryptfs_inode); - file_size = cpu_to_be64(file_size); - memcpy(file_size_virt, &file_size, sizeof(u64)); + put_unaligned_be64(i_size_read(ecryptfs_inode), file_size_virt); rc = ecryptfs_write_lower(ecryptfs_inode, file_size_virt, 0, sizeof(u64)); kfree(file_size_virt); @@ -403,7 +401,6 @@ static int ecryptfs_write_inode_size_to_xattr(struct inode *ecryptfs_inode) struct dentry *lower_dentry = ecryptfs_inode_to_private(ecryptfs_inode)->lower_file->f_dentry; struct inode *lower_inode = lower_dentry->d_inode; - u64 file_size; int rc; if (!lower_inode->i_op->getxattr || !lower_inode->i_op->setxattr) { @@ -424,9 +421,7 @@ static int ecryptfs_write_inode_size_to_xattr(struct inode *ecryptfs_inode) xattr_virt, PAGE_CACHE_SIZE); if (size < 0) size = 8; - file_size = (u64)i_size_read(ecryptfs_inode); - file_size = cpu_to_be64(file_size); - memcpy(xattr_virt, &file_size, sizeof(u64)); + put_unaligned_be64(i_size_read(ecryptfs_inode), xattr_virt); rc = lower_inode->i_op->setxattr(lower_dentry, ECRYPTFS_XATTR_NAME, xattr_virt, size, 0); mutex_unlock(&lower_inode->i_mutex); -- cgit v1.2.3-70-g09d2 From 72b55fffd631a89e5be6fe1b4f2565bc4cd90deb Mon Sep 17 00:00:00 2001 From: Michael Halcrow Date: Wed, 23 Jul 2008 21:30:07 -0700 Subject: eCryptfs: do not try to open device files on mknod When creating device nodes, eCryptfs needs to delay actually opening the lower persistent file until an application tries to open. Device handles may not be backed by anything when they first come into existence. [Valdis.Kletnieks@vt.edu: build fix] Signed-off-by: Michael Halcrow Cc: Signed-off-by: Linus Torvalds --- fs/ecryptfs/ecryptfs_kernel.h | 6 +++++- fs/ecryptfs/file.c | 14 ++++++++++++++ fs/ecryptfs/inode.c | 6 ++++-- fs/ecryptfs/main.c | 29 +++++++++++++++++++---------- 4 files changed, 42 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/ecryptfs/ecryptfs_kernel.h b/fs/ecryptfs/ecryptfs_kernel.h index b4a0cccfdd7..b0727f91454 100644 --- a/fs/ecryptfs/ecryptfs_kernel.h +++ b/fs/ecryptfs/ecryptfs_kernel.h @@ -235,6 +235,7 @@ struct ecryptfs_crypt_stat { #define ECRYPTFS_METADATA_IN_XATTR 0x00000100 #define ECRYPTFS_VIEW_AS_ENCRYPTED 0x00000200 #define ECRYPTFS_KEY_SET 0x00000400 +#define ECRYPTFS_DELAY_PERSISTENT 0x00000800 u32 flags; unsigned int file_version; size_t iv_bytes; @@ -574,9 +575,11 @@ struct ecryptfs_open_req { struct list_head kthread_ctl_list; }; +#define ECRYPTFS_INTERPOSE_FLAG_D_ADD 0x00000001 +#define ECRYPTFS_INTERPOSE_FLAG_DELAY_PERSISTENT_FILE 0x00000002 int ecryptfs_interpose(struct dentry *hidden_dentry, struct dentry *this_dentry, struct super_block *sb, - int flag); + u32 flags); int ecryptfs_fill_zeros(struct file *file, loff_t new_length); int ecryptfs_decode_filename(struct ecryptfs_crypt_stat *crypt_stat, const char *name, int length, @@ -709,5 +712,6 @@ void ecryptfs_destroy_kthread(void); int ecryptfs_privileged_open(struct file **lower_file, struct dentry *lower_dentry, struct vfsmount *lower_mnt); +int ecryptfs_init_persistent_file(struct dentry *ecryptfs_dentry); #endif /* #ifndef ECRYPTFS_KERNEL_H */ diff --git a/fs/ecryptfs/file.c b/fs/ecryptfs/file.c index f0be2905152..2c2d60df3f6 100644 --- a/fs/ecryptfs/file.c +++ b/fs/ecryptfs/file.c @@ -199,6 +199,20 @@ static int ecryptfs_open(struct inode *inode, struct file *file) "file must hence be opened RO\n", __func__); goto out; } + if (!ecryptfs_inode_to_private(inode)->lower_file) { + BUG_ON(!(crypt_stat->flags & ECRYPTFS_DELAY_PERSISTENT)); + mutex_lock(&crypt_stat->cs_mutex); + crypt_stat->flags &= ~(ECRYPTFS_ENCRYPTED); + mutex_unlock(&crypt_stat->cs_mutex); + rc = ecryptfs_init_persistent_file(ecryptfs_dentry); + if (rc) { + printk(KERN_ERR "%s: Error attempting to initialize " + "the persistent file for the dentry with name " + "[%s]; rc = [%d]\n", __func__, + ecryptfs_dentry->d_name.name, rc); + goto out; + } + } ecryptfs_set_file_lower( file, ecryptfs_inode_to_private(inode)->lower_file); if (S_ISDIR(ecryptfs_dentry->d_inode->i_mode)) { diff --git a/fs/ecryptfs/inode.c b/fs/ecryptfs/inode.c index 7315547193e..26090878c93 100644 --- a/fs/ecryptfs/inode.c +++ b/fs/ecryptfs/inode.c @@ -308,7 +308,8 @@ static struct dentry *ecryptfs_lookup(struct inode *dir, struct dentry *dentry, d_add(dentry, NULL); goto out; } - rc = ecryptfs_interpose(lower_dentry, dentry, dir->i_sb, 1); + rc = ecryptfs_interpose(lower_dentry, dentry, dir->i_sb, + ECRYPTFS_INTERPOSE_FLAG_D_ADD); if (rc) { ecryptfs_printk(KERN_ERR, "Error interposing\n"); goto out_dput; @@ -537,7 +538,8 @@ ecryptfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t dev) rc = vfs_mknod(lower_dir_dentry->d_inode, lower_dentry, mode, dev); if (rc || !lower_dentry->d_inode) goto out; - rc = ecryptfs_interpose(lower_dentry, dentry, dir->i_sb, 0); + rc = ecryptfs_interpose(lower_dentry, dentry, dir->i_sb, + ECRYPTFS_INTERPOSE_FLAG_DELAY_PERSISTENT_FILE); if (rc) goto out; fsstack_copy_attr_times(dir, lower_dir_dentry->d_inode); diff --git a/fs/ecryptfs/main.c b/fs/ecryptfs/main.c index 10475d93ff5..ee4f84b2041 100644 --- a/fs/ecryptfs/main.c +++ b/fs/ecryptfs/main.c @@ -117,7 +117,7 @@ void __ecryptfs_printk(const char *fmt, ...) * * Returns zero on success; non-zero otherwise */ -static int ecryptfs_init_persistent_file(struct dentry *ecryptfs_dentry) +int ecryptfs_init_persistent_file(struct dentry *ecryptfs_dentry) { struct ecryptfs_inode_info *inode_info = ecryptfs_inode_to_private(ecryptfs_dentry->d_inode); @@ -149,14 +149,14 @@ static int ecryptfs_init_persistent_file(struct dentry *ecryptfs_dentry) * @lower_dentry: Existing dentry in the lower filesystem * @dentry: ecryptfs' dentry * @sb: ecryptfs's super_block - * @flag: If set to true, then d_add is called, else d_instantiate is called + * @flags: flags to govern behavior of interpose procedure * * Interposes upper and lower dentries. * * Returns zero on success; non-zero otherwise */ int ecryptfs_interpose(struct dentry *lower_dentry, struct dentry *dentry, - struct super_block *sb, int flag) + struct super_block *sb, u32 flags) { struct inode *lower_inode; struct inode *inode; @@ -193,7 +193,7 @@ int ecryptfs_interpose(struct dentry *lower_dentry, struct dentry *dentry, init_special_inode(inode, lower_inode->i_mode, lower_inode->i_rdev); dentry->d_op = &ecryptfs_dops; - if (flag) + if (flags & ECRYPTFS_INTERPOSE_FLAG_D_ADD) d_add(dentry, inode); else d_instantiate(dentry, inode); @@ -201,12 +201,21 @@ int ecryptfs_interpose(struct dentry *lower_dentry, struct dentry *dentry, /* This size will be overwritten for real files w/ headers and * other metadata */ fsstack_copy_inode_size(inode, lower_inode); - rc = ecryptfs_init_persistent_file(dentry); - if (rc) { - printk(KERN_ERR "%s: Error attempting to initialize the " - "persistent file for the dentry with name [%s]; " - "rc = [%d]\n", __func__, dentry->d_name.name, rc); - goto out; + if (!(flags & ECRYPTFS_INTERPOSE_FLAG_DELAY_PERSISTENT_FILE)) { + rc = ecryptfs_init_persistent_file(dentry); + if (rc) { + printk(KERN_ERR "%s: Error attempting to initialize " + "the persistent file for the dentry with name " + "[%s]; rc = [%d]\n", __func__, + dentry->d_name.name, rc); + goto out; + } + } else { + struct ecryptfs_inode_info *inode_info = + ecryptfs_inode_to_private(dentry->d_inode); + + inode_info->lower_file = NULL; + inode_info->crypt_stat.flags |= ECRYPTFS_DELAY_PERSISTENT; } out: return rc; -- cgit v1.2.3-70-g09d2 From 391b52f98cf2e9bff227dad8bf9ea206fec43fa4 Mon Sep 17 00:00:00 2001 From: Michael Halcrow Date: Wed, 23 Jul 2008 21:30:08 -0700 Subject: eCryptfs: Make all persistent file opens delayed There is no good reason to immediately open the lower file, and that can cause problems with files that the user does not intend to immediately open, such as device nodes. This patch removes the persistent file open from the interpose step and pushes that to the locations where eCryptfs really does need the lower persistent file, such as just before reading or writing the metadata stored in the lower file header. Two functions are jumping to out_dput when they should just be jumping to out on error paths. This patch also fixes these. Signed-off-by: Michael Halcrow Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ecryptfs/ecryptfs_kernel.h | 2 -- fs/ecryptfs/file.c | 4 ---- fs/ecryptfs/inode.c | 27 +++++++++++++++++++++++---- fs/ecryptfs/main.c | 16 ---------------- 4 files changed, 23 insertions(+), 26 deletions(-) (limited to 'fs') diff --git a/fs/ecryptfs/ecryptfs_kernel.h b/fs/ecryptfs/ecryptfs_kernel.h index b0727f91454..b73fb752c5f 100644 --- a/fs/ecryptfs/ecryptfs_kernel.h +++ b/fs/ecryptfs/ecryptfs_kernel.h @@ -235,7 +235,6 @@ struct ecryptfs_crypt_stat { #define ECRYPTFS_METADATA_IN_XATTR 0x00000100 #define ECRYPTFS_VIEW_AS_ENCRYPTED 0x00000200 #define ECRYPTFS_KEY_SET 0x00000400 -#define ECRYPTFS_DELAY_PERSISTENT 0x00000800 u32 flags; unsigned int file_version; size_t iv_bytes; @@ -576,7 +575,6 @@ struct ecryptfs_open_req { }; #define ECRYPTFS_INTERPOSE_FLAG_D_ADD 0x00000001 -#define ECRYPTFS_INTERPOSE_FLAG_DELAY_PERSISTENT_FILE 0x00000002 int ecryptfs_interpose(struct dentry *hidden_dentry, struct dentry *this_dentry, struct super_block *sb, u32 flags); diff --git a/fs/ecryptfs/file.c b/fs/ecryptfs/file.c index 2c2d60df3f6..9244d653743 100644 --- a/fs/ecryptfs/file.c +++ b/fs/ecryptfs/file.c @@ -200,10 +200,6 @@ static int ecryptfs_open(struct inode *inode, struct file *file) goto out; } if (!ecryptfs_inode_to_private(inode)->lower_file) { - BUG_ON(!(crypt_stat->flags & ECRYPTFS_DELAY_PERSISTENT)); - mutex_lock(&crypt_stat->cs_mutex); - crypt_stat->flags &= ~(ECRYPTFS_ENCRYPTED); - mutex_unlock(&crypt_stat->cs_mutex); rc = ecryptfs_init_persistent_file(ecryptfs_dentry); if (rc) { printk(KERN_ERR "%s: Error attempting to initialize " diff --git a/fs/ecryptfs/inode.c b/fs/ecryptfs/inode.c index 26090878c93..d755455e3bf 100644 --- a/fs/ecryptfs/inode.c +++ b/fs/ecryptfs/inode.c @@ -189,6 +189,16 @@ static int ecryptfs_initialize_file(struct dentry *ecryptfs_dentry) "context; rc = [%d]\n", rc); goto out; } + if (!ecryptfs_inode_to_private(ecryptfs_dentry->d_inode)->lower_file) { + rc = ecryptfs_init_persistent_file(ecryptfs_dentry); + if (rc) { + printk(KERN_ERR "%s: Error attempting to initialize " + "the persistent file for the dentry with name " + "[%s]; rc = [%d]\n", __func__, + ecryptfs_dentry->d_name.name, rc); + goto out; + } + } rc = ecryptfs_write_metadata(ecryptfs_dentry); if (rc) { printk(KERN_ERR "Error writing headers; rc = [%d]\n", rc); @@ -312,7 +322,7 @@ static struct dentry *ecryptfs_lookup(struct inode *dir, struct dentry *dentry, ECRYPTFS_INTERPOSE_FLAG_D_ADD); if (rc) { ecryptfs_printk(KERN_ERR, "Error interposing\n"); - goto out_dput; + goto out; } if (S_ISDIR(lower_inode->i_mode)) { ecryptfs_printk(KERN_DEBUG, "Is a directory; returning\n"); @@ -338,11 +348,21 @@ static struct dentry *ecryptfs_lookup(struct inode *dir, struct dentry *dentry, rc = -ENOMEM; ecryptfs_printk(KERN_ERR, "Cannot ecryptfs_kmalloc a page\n"); - goto out_dput; + goto out; } crypt_stat = &ecryptfs_inode_to_private(dentry->d_inode)->crypt_stat; if (!(crypt_stat->flags & ECRYPTFS_POLICY_APPLIED)) ecryptfs_set_default_sizes(crypt_stat); + if (!ecryptfs_inode_to_private(dentry->d_inode)->lower_file) { + rc = ecryptfs_init_persistent_file(dentry); + if (rc) { + printk(KERN_ERR "%s: Error attempting to initialize " + "the persistent file for the dentry with name " + "[%s]; rc = [%d]\n", __func__, + dentry->d_name.name, rc); + goto out; + } + } rc = ecryptfs_read_and_validate_header_region(page_virt, dentry->d_inode); if (rc) { @@ -538,8 +558,7 @@ ecryptfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t dev) rc = vfs_mknod(lower_dir_dentry->d_inode, lower_dentry, mode, dev); if (rc || !lower_dentry->d_inode) goto out; - rc = ecryptfs_interpose(lower_dentry, dentry, dir->i_sb, - ECRYPTFS_INTERPOSE_FLAG_DELAY_PERSISTENT_FILE); + rc = ecryptfs_interpose(lower_dentry, dentry, dir->i_sb, 0); if (rc) goto out; fsstack_copy_attr_times(dir, lower_dir_dentry->d_inode); diff --git a/fs/ecryptfs/main.c b/fs/ecryptfs/main.c index ee4f84b2041..6f403cfba14 100644 --- a/fs/ecryptfs/main.c +++ b/fs/ecryptfs/main.c @@ -201,22 +201,6 @@ int ecryptfs_interpose(struct dentry *lower_dentry, struct dentry *dentry, /* This size will be overwritten for real files w/ headers and * other metadata */ fsstack_copy_inode_size(inode, lower_inode); - if (!(flags & ECRYPTFS_INTERPOSE_FLAG_DELAY_PERSISTENT_FILE)) { - rc = ecryptfs_init_persistent_file(dentry); - if (rc) { - printk(KERN_ERR "%s: Error attempting to initialize " - "the persistent file for the dentry with name " - "[%s]; rc = [%d]\n", __func__, - dentry->d_name.name, rc); - goto out; - } - } else { - struct ecryptfs_inode_info *inode_info = - ecryptfs_inode_to_private(dentry->d_inode); - - inode_info->lower_file = NULL; - inode_info->crypt_stat.flags |= ECRYPTFS_DELAY_PERSISTENT; - } out: return rc; } -- cgit v1.2.3-70-g09d2 From 5f6f4f28b6ba543beef8bad91aa6f69c7ffeee51 Mon Sep 17 00:00:00 2001 From: Ian Kent Date: Wed, 23 Jul 2008 21:30:09 -0700 Subject: autofs4: don't make expiring dentry negative Correct the error of making a positive dentry negative after it has been instantiated. The code that makes this error attempts to re-use the dentry from a concurrent expire and mount to resolve a race and the dentry used for the lookup must be negative for mounts to trigger in the required cases. The fact is that the dentry doesn't need to be re-used because all that is needed is to preserve the flag that indicates an expire is still incomplete at the time of the mount request. This change uses the the dentry to check the flag and wait for the expire to complete then discards it instead of attempting to re-use it. Signed-off-by: Ian Kent Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/autofs_i.h | 6 +-- fs/autofs4/inode.c | 6 +-- fs/autofs4/root.c | 118 ++++++++++++++++++++------------------------------ 3 files changed, 52 insertions(+), 78 deletions(-) (limited to 'fs') diff --git a/fs/autofs4/autofs_i.h b/fs/autofs4/autofs_i.h index c3d352d7fa9..69b1497b002 100644 --- a/fs/autofs4/autofs_i.h +++ b/fs/autofs4/autofs_i.h @@ -52,7 +52,7 @@ struct autofs_info { int flags; - struct list_head rehash; + struct list_head expiring; struct autofs_sb_info *sbi; unsigned long last_used; @@ -112,8 +112,8 @@ struct autofs_sb_info { struct mutex wq_mutex; spinlock_t fs_lock; struct autofs_wait_queue *queues; /* Wait queue pointer */ - spinlock_t rehash_lock; - struct list_head rehash_list; + spinlock_t lookup_lock; + struct list_head expiring_list; }; static inline struct autofs_sb_info *autofs4_sbi(struct super_block *sb) diff --git a/fs/autofs4/inode.c b/fs/autofs4/inode.c index 2fdcf5e1d23..94bfc154d7a 100644 --- a/fs/autofs4/inode.c +++ b/fs/autofs4/inode.c @@ -47,7 +47,7 @@ struct autofs_info *autofs4_init_ino(struct autofs_info *ino, ino->dentry = NULL; ino->size = 0; - INIT_LIST_HEAD(&ino->rehash); + INIT_LIST_HEAD(&ino->expiring); ino->last_used = jiffies; atomic_set(&ino->count, 0); @@ -338,8 +338,8 @@ int autofs4_fill_super(struct super_block *s, void *data, int silent) mutex_init(&sbi->wq_mutex); spin_lock_init(&sbi->fs_lock); sbi->queues = NULL; - spin_lock_init(&sbi->rehash_lock); - INIT_LIST_HEAD(&sbi->rehash_list); + spin_lock_init(&sbi->lookup_lock); + INIT_LIST_HEAD(&sbi->expiring_list); s->s_blocksize = 1024; s->s_blocksize_bits = 10; s->s_magic = AUTOFS_SUPER_MAGIC; diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c index edf5b6bddb5..9ead2279df4 100644 --- a/fs/autofs4/root.c +++ b/fs/autofs4/root.c @@ -493,10 +493,10 @@ void autofs4_dentry_release(struct dentry *de) struct autofs_sb_info *sbi = autofs4_sbi(de->d_sb); if (sbi) { - spin_lock(&sbi->rehash_lock); - if (!list_empty(&inf->rehash)) - list_del(&inf->rehash); - spin_unlock(&sbi->rehash_lock); + spin_lock(&sbi->lookup_lock); + if (!list_empty(&inf->expiring)) + list_del(&inf->expiring); + spin_unlock(&sbi->lookup_lock); } inf->dentry = NULL; @@ -518,7 +518,7 @@ static struct dentry_operations autofs4_dentry_operations = { .d_release = autofs4_dentry_release, }; -static struct dentry *autofs4_lookup_unhashed(struct autofs_sb_info *sbi, struct dentry *parent, struct qstr *name) +static struct dentry *autofs4_lookup_expiring(struct autofs_sb_info *sbi, struct dentry *parent, struct qstr *name) { unsigned int len = name->len; unsigned int hash = name->hash; @@ -526,14 +526,14 @@ static struct dentry *autofs4_lookup_unhashed(struct autofs_sb_info *sbi, struct struct list_head *p, *head; spin_lock(&dcache_lock); - spin_lock(&sbi->rehash_lock); - head = &sbi->rehash_list; + spin_lock(&sbi->lookup_lock); + head = &sbi->expiring_list; list_for_each(p, head) { struct autofs_info *ino; struct dentry *dentry; struct qstr *qstr; - ino = list_entry(p, struct autofs_info, rehash); + ino = list_entry(p, struct autofs_info, expiring); dentry = ino->dentry; spin_lock(&dentry->d_lock); @@ -555,33 +555,16 @@ static struct dentry *autofs4_lookup_unhashed(struct autofs_sb_info *sbi, struct goto next; if (d_unhashed(dentry)) { - struct inode *inode = dentry->d_inode; - - ino = autofs4_dentry_ino(dentry); - list_del_init(&ino->rehash); dget(dentry); - /* - * Make the rehashed dentry negative so the VFS - * behaves as it should. - */ - if (inode) { - dentry->d_inode = NULL; - list_del_init(&dentry->d_alias); - spin_unlock(&dentry->d_lock); - spin_unlock(&sbi->rehash_lock); - spin_unlock(&dcache_lock); - iput(inode); - return dentry; - } spin_unlock(&dentry->d_lock); - spin_unlock(&sbi->rehash_lock); + spin_unlock(&sbi->lookup_lock); spin_unlock(&dcache_lock); return dentry; } next: spin_unlock(&dentry->d_lock); } - spin_unlock(&sbi->rehash_lock); + spin_unlock(&sbi->lookup_lock); spin_unlock(&dcache_lock); return NULL; @@ -591,7 +574,7 @@ next: static struct dentry *autofs4_lookup(struct inode *dir, struct dentry *dentry, struct nameidata *nd) { struct autofs_sb_info *sbi; - struct dentry *unhashed; + struct dentry *expiring; int oz_mode; DPRINTK("name = %.*s", @@ -607,44 +590,44 @@ static struct dentry *autofs4_lookup(struct inode *dir, struct dentry *dentry, s DPRINTK("pid = %u, pgrp = %u, catatonic = %d, oz_mode = %d", current->pid, task_pgrp_nr(current), sbi->catatonic, oz_mode); - unhashed = autofs4_lookup_unhashed(sbi, dentry->d_parent, &dentry->d_name); - if (!unhashed) { - /* - * Mark the dentry incomplete but don't hash it. We do this - * to serialize our inode creation operations (symlink and - * mkdir) which prevents deadlock during the callback to - * the daemon. Subsequent user space lookups for the same - * dentry are placed on the wait queue while the daemon - * itself is allowed passage unresticted so the create - * operation itself can then hash the dentry. Finally, - * we check for the hashed dentry and return the newly - * hashed dentry. - */ - dentry->d_op = &autofs4_root_dentry_operations; - - dentry->d_fsdata = NULL; - d_instantiate(dentry, NULL); - } else { - struct autofs_info *ino = autofs4_dentry_ino(unhashed); - DPRINTK("rehash %p with %p", dentry, unhashed); + expiring = autofs4_lookup_expiring(sbi, dentry->d_parent, &dentry->d_name); + if (expiring) { + struct autofs_info *ino = autofs4_dentry_ino(expiring); /* * If we are racing with expire the request might not * be quite complete but the directory has been removed * so it must have been successful, so just wait for it. - * We need to ensure the AUTOFS_INF_EXPIRING flag is clear - * before continuing as revalidate may fail when calling - * try_to_fill_dentry (returning EAGAIN) if we don't. */ while (ino && (ino->flags & AUTOFS_INF_EXPIRING)) { DPRINTK("wait for incomplete expire %p name=%.*s", - unhashed, unhashed->d_name.len, - unhashed->d_name.name); - autofs4_wait(sbi, unhashed, NFY_NONE); + expiring, expiring->d_name.len, + expiring->d_name.name); + autofs4_wait(sbi, expiring, NFY_NONE); DPRINTK("request completed"); } - dentry = unhashed; + spin_lock(&sbi->lookup_lock); + if (!list_empty(&ino->expiring)) + list_del_init(&ino->expiring); + spin_unlock(&sbi->lookup_lock); + dput(expiring); } + /* + * Mark the dentry incomplete but don't hash it. We do this + * to serialize our inode creation operations (symlink and + * mkdir) which prevents deadlock during the callback to + * the daemon. Subsequent user space lookups for the same + * dentry are placed on the wait queue while the daemon + * itself is allowed passage unresticted so the create + * operation itself can then hash the dentry. Finally, + * we check for the hashed dentry and return the newly + * hashed dentry. + */ + dentry->d_op = &autofs4_root_dentry_operations; + + dentry->d_fsdata = NULL; + d_instantiate(dentry, NULL); + if (!oz_mode) { spin_lock(&dentry->d_lock); dentry->d_flags |= DCACHE_AUTOFS_PENDING; @@ -668,8 +651,6 @@ static struct dentry *autofs4_lookup(struct inode *dir, struct dentry *dentry, s if (sigismember (sigset, SIGKILL) || sigismember (sigset, SIGQUIT) || sigismember (sigset, SIGINT)) { - if (unhashed) - dput(unhashed); return ERR_PTR(-ERESTARTNOINTR); } } @@ -699,15 +680,9 @@ static struct dentry *autofs4_lookup(struct inode *dir, struct dentry *dentry, s else dentry = ERR_PTR(-ENOENT); - if (unhashed) - dput(unhashed); - return dentry; } - if (unhashed) - return dentry; - return NULL; } @@ -769,9 +744,8 @@ static int autofs4_dir_symlink(struct inode *dir, * that the file no longer exists. However, doing that means that the * VFS layer can turn the dentry into a negative dentry. We don't want * this, because the unlink is probably the result of an expire. - * We simply d_drop it and add it to a rehash candidates list in the - * super block, which allows the dentry lookup to reuse it retaining - * the flags, such as expire in progress, in case we're racing with expire. + * We simply d_drop it and add it to a expiring list in the super block, + * which allows the dentry lookup to check for an incomplete expire. * * If a process is blocked on the dentry waiting for the expire to finish, * it will invalidate the dentry and try to mount with a new one. @@ -801,9 +775,9 @@ static int autofs4_dir_unlink(struct inode *dir, struct dentry *dentry) dir->i_mtime = CURRENT_TIME; spin_lock(&dcache_lock); - spin_lock(&sbi->rehash_lock); - list_add(&ino->rehash, &sbi->rehash_list); - spin_unlock(&sbi->rehash_lock); + spin_lock(&sbi->lookup_lock); + list_add(&ino->expiring, &sbi->expiring_list); + spin_unlock(&sbi->lookup_lock); spin_lock(&dentry->d_lock); __d_drop(dentry); spin_unlock(&dentry->d_lock); @@ -829,9 +803,9 @@ static int autofs4_dir_rmdir(struct inode *dir, struct dentry *dentry) spin_unlock(&dcache_lock); return -ENOTEMPTY; } - spin_lock(&sbi->rehash_lock); - list_add(&ino->rehash, &sbi->rehash_list); - spin_unlock(&sbi->rehash_lock); + spin_lock(&sbi->lookup_lock); + list_add(&ino->expiring, &sbi->expiring_list); + spin_unlock(&sbi->lookup_lock); spin_lock(&dentry->d_lock); __d_drop(dentry); spin_unlock(&dentry->d_lock); -- cgit v1.2.3-70-g09d2 From caf7da3d5d4d9dd873eb52d025d8cc63b89f1fdb Mon Sep 17 00:00:00 2001 From: Ian Kent Date: Wed, 23 Jul 2008 21:30:11 -0700 Subject: autofs4: revert - redo lookup in ttfd This patch series enables the use of a single dentry for lookups prior to the dentry being hashed and so we no longer need to redo the lookup. This patch reverts the patch of commit 033790449ba9c4dcf8478a87693d33df625c23b5. Signed-off-by: Ian Kent Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/root.c | 21 --------------------- 1 file changed, 21 deletions(-) (limited to 'fs') diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c index 9ead2279df4..53dabe8d5b8 100644 --- a/fs/autofs4/root.c +++ b/fs/autofs4/root.c @@ -242,7 +242,6 @@ static int try_to_fill_dentry(struct dentry *dentry, int flags) { struct autofs_sb_info *sbi = autofs4_sbi(dentry->d_sb); struct autofs_info *ino = autofs4_dentry_ino(dentry); - struct dentry *new; int status; /* Block on any pending expiry here; invalidate the dentry @@ -320,26 +319,6 @@ static int try_to_fill_dentry(struct dentry *dentry, int flags) dentry->d_flags &= ~DCACHE_AUTOFS_PENDING; spin_unlock(&dentry->d_lock); - /* - * The dentry that is passed in from lookup may not be the one - * we end up using, as mkdir can create a new one. If this - * happens, and another process tries the lookup at the same time, - * it will set the PENDING flag on this new dentry, but add itself - * to our waitq. Then, if after the lookup succeeds, the first - * process that requested the mount performs another lookup of the - * same directory, it will show up as still pending! So, we need - * to redo the lookup here and clear pending on that dentry. - */ - if (d_unhashed(dentry)) { - new = d_lookup(dentry->d_parent, &dentry->d_name); - if (new) { - spin_lock(&new->d_lock); - new->d_flags &= ~DCACHE_AUTOFS_PENDING; - spin_unlock(&new->d_lock); - dput(new); - } - } - return 0; } -- cgit v1.2.3-70-g09d2 From 2576737873dc1d9ea461a5955a5f6779b569a350 Mon Sep 17 00:00:00 2001 From: Ian Kent Date: Wed, 23 Jul 2008 21:30:12 -0700 Subject: autofs4: use look aside list for lookups A while ago a patch to resolve a deadlock during directory creation was merged. This delayed the hashing of lookup dentrys until the ->mkdir() (or ->symlink()) operation completed to ensure we always went through ->lookup() instead of also having processes go through ->revalidate() so our VFS locking remained consistent. Now we are seeing a couple of side affects of that change in situations with heavy mount activity. Two cases have been identified: 1) When a mount request is triggered, due to the delayed hashing, the directory created by user space for the mount point doesn't have the DCACHE_AUTOFS_PENDING flag set. In the case of an autofs multi-mount where a tree of mount point directories are created this can lead to the path walk continuing rather than the dentry being sent to the wait queue to wait for request completion. This is because, if the pending flag isn't set, the criteria for deciding this is a mount in progress fails to hold, namely that the dentry is not a mount point and has no subdirectories. 2) A mount request dentry is initially created negative and unhashed. It remains this way until the ->mkdir() callback completes. Since it is unhashed a fresh dentry is used when the user space mount request creates the mount point directory. This leaves the original dentry negative and unhashed. But revalidate has no way to tell the VFS that the dentry has changed, other than to force another ->lookup() by returning false, which is at best wastefull and at worst not possible. This results in an -ENOENT return from the original path walk when in fact the mount succeeded. To resolve this we need to ensure that the same dentry is used in all calls to ->lookup() during the course of a mount request. This patch achieves that by adding the initial dentry to a look aside list and removes it at ->mkdir() or ->symlink() completion (or when the dentry is released), since these are the only create operations autofs4 supports. Signed-off-by: Ian Kent Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/autofs_i.h | 2 + fs/autofs4/inode.c | 25 +++++--- fs/autofs4/root.c | 169 +++++++++++++++++++++++++++++++++++++++++--------- 3 files changed, 156 insertions(+), 40 deletions(-) (limited to 'fs') diff --git a/fs/autofs4/autofs_i.h b/fs/autofs4/autofs_i.h index 69b1497b002..2dce2334737 100644 --- a/fs/autofs4/autofs_i.h +++ b/fs/autofs4/autofs_i.h @@ -52,6 +52,7 @@ struct autofs_info { int flags; + struct list_head active; struct list_head expiring; struct autofs_sb_info *sbi; @@ -113,6 +114,7 @@ struct autofs_sb_info { spinlock_t fs_lock; struct autofs_wait_queue *queues; /* Wait queue pointer */ spinlock_t lookup_lock; + struct list_head active_list; struct list_head expiring_list; }; diff --git a/fs/autofs4/inode.c b/fs/autofs4/inode.c index 94bfc154d7a..e3e70994ab4 100644 --- a/fs/autofs4/inode.c +++ b/fs/autofs4/inode.c @@ -24,8 +24,10 @@ static void ino_lnkfree(struct autofs_info *ino) { - kfree(ino->u.symlink); - ino->u.symlink = NULL; + if (ino->u.symlink) { + kfree(ino->u.symlink); + ino->u.symlink = NULL; + } } struct autofs_info *autofs4_init_ino(struct autofs_info *ino, @@ -41,16 +43,18 @@ struct autofs_info *autofs4_init_ino(struct autofs_info *ino, if (ino == NULL) return NULL; - ino->flags = 0; - ino->mode = mode; - ino->inode = NULL; - ino->dentry = NULL; - ino->size = 0; - - INIT_LIST_HEAD(&ino->expiring); + if (!reinit) { + ino->flags = 0; + ino->inode = NULL; + ino->dentry = NULL; + ino->size = 0; + INIT_LIST_HEAD(&ino->active); + INIT_LIST_HEAD(&ino->expiring); + atomic_set(&ino->count, 0); + } + ino->mode = mode; ino->last_used = jiffies; - atomic_set(&ino->count, 0); ino->sbi = sbi; @@ -339,6 +343,7 @@ int autofs4_fill_super(struct super_block *s, void *data, int silent) spin_lock_init(&sbi->fs_lock); sbi->queues = NULL; spin_lock_init(&sbi->lookup_lock); + INIT_LIST_HEAD(&sbi->active_list); INIT_LIST_HEAD(&sbi->expiring_list); s->s_blocksize = 1024; s->s_blocksize_bits = 10; diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c index 53dabe8d5b8..dbb70d5a488 100644 --- a/fs/autofs4/root.c +++ b/fs/autofs4/root.c @@ -473,6 +473,8 @@ void autofs4_dentry_release(struct dentry *de) if (sbi) { spin_lock(&sbi->lookup_lock); + if (!list_empty(&inf->active)) + list_del(&inf->active); if (!list_empty(&inf->expiring)) list_del(&inf->expiring); spin_unlock(&sbi->lookup_lock); @@ -497,6 +499,58 @@ static struct dentry_operations autofs4_dentry_operations = { .d_release = autofs4_dentry_release, }; +static struct dentry *autofs4_lookup_active(struct autofs_sb_info *sbi, struct dentry *parent, struct qstr *name) +{ + unsigned int len = name->len; + unsigned int hash = name->hash; + const unsigned char *str = name->name; + struct list_head *p, *head; + + spin_lock(&dcache_lock); + spin_lock(&sbi->lookup_lock); + head = &sbi->active_list; + list_for_each(p, head) { + struct autofs_info *ino; + struct dentry *dentry; + struct qstr *qstr; + + ino = list_entry(p, struct autofs_info, active); + dentry = ino->dentry; + + spin_lock(&dentry->d_lock); + + /* Already gone? */ + if (atomic_read(&dentry->d_count) == 0) + goto next; + + qstr = &dentry->d_name; + + if (dentry->d_name.hash != hash) + goto next; + if (dentry->d_parent != parent) + goto next; + + if (qstr->len != len) + goto next; + if (memcmp(qstr->name, str, len)) + goto next; + + if (d_unhashed(dentry)) { + dget(dentry); + spin_unlock(&dentry->d_lock); + spin_unlock(&sbi->lookup_lock); + spin_unlock(&dcache_lock); + return dentry; + } +next: + spin_unlock(&dentry->d_lock); + } + spin_unlock(&sbi->lookup_lock); + spin_unlock(&dcache_lock); + + return NULL; +} + static struct dentry *autofs4_lookup_expiring(struct autofs_sb_info *sbi, struct dentry *parent, struct qstr *name) { unsigned int len = name->len; @@ -553,7 +607,8 @@ next: static struct dentry *autofs4_lookup(struct inode *dir, struct dentry *dentry, struct nameidata *nd) { struct autofs_sb_info *sbi; - struct dentry *expiring; + struct autofs_info *ino; + struct dentry *expiring, *unhashed; int oz_mode; DPRINTK("name = %.*s", @@ -571,12 +626,12 @@ static struct dentry *autofs4_lookup(struct inode *dir, struct dentry *dentry, s expiring = autofs4_lookup_expiring(sbi, dentry->d_parent, &dentry->d_name); if (expiring) { - struct autofs_info *ino = autofs4_dentry_ino(expiring); /* * If we are racing with expire the request might not * be quite complete but the directory has been removed * so it must have been successful, so just wait for it. */ + ino = autofs4_dentry_ino(expiring); while (ino && (ino->flags & AUTOFS_INF_EXPIRING)) { DPRINTK("wait for incomplete expire %p name=%.*s", expiring, expiring->d_name.len, @@ -591,21 +646,41 @@ static struct dentry *autofs4_lookup(struct inode *dir, struct dentry *dentry, s dput(expiring); } - /* - * Mark the dentry incomplete but don't hash it. We do this - * to serialize our inode creation operations (symlink and - * mkdir) which prevents deadlock during the callback to - * the daemon. Subsequent user space lookups for the same - * dentry are placed on the wait queue while the daemon - * itself is allowed passage unresticted so the create - * operation itself can then hash the dentry. Finally, - * we check for the hashed dentry and return the newly - * hashed dentry. - */ - dentry->d_op = &autofs4_root_dentry_operations; + unhashed = autofs4_lookup_active(sbi, dentry->d_parent, &dentry->d_name); + if (unhashed) + dentry = unhashed; + else { + /* + * Mark the dentry incomplete but don't hash it. We do this + * to serialize our inode creation operations (symlink and + * mkdir) which prevents deadlock during the callback to + * the daemon. Subsequent user space lookups for the same + * dentry are placed on the wait queue while the daemon + * itself is allowed passage unresticted so the create + * operation itself can then hash the dentry. Finally, + * we check for the hashed dentry and return the newly + * hashed dentry. + */ + dentry->d_op = &autofs4_root_dentry_operations; + + /* + * And we need to ensure that the same dentry is used for + * all following lookup calls until it is hashed so that + * the dentry flags are persistent throughout the request. + */ + ino = autofs4_init_ino(NULL, sbi, 0555); + if (!ino) + return ERR_PTR(-ENOMEM); + + dentry->d_fsdata = ino; + ino->dentry = dentry; + + spin_lock(&sbi->lookup_lock); + list_add(&ino->active, &sbi->active_list); + spin_unlock(&sbi->lookup_lock); - dentry->d_fsdata = NULL; - d_instantiate(dentry, NULL); + d_instantiate(dentry, NULL); + } if (!oz_mode) { spin_lock(&dentry->d_lock); @@ -630,12 +705,16 @@ static struct dentry *autofs4_lookup(struct inode *dir, struct dentry *dentry, s if (sigismember (sigset, SIGKILL) || sigismember (sigset, SIGQUIT) || sigismember (sigset, SIGINT)) { + if (unhashed) + dput(unhashed); return ERR_PTR(-ERESTARTNOINTR); } } - spin_lock(&dentry->d_lock); - dentry->d_flags &= ~DCACHE_AUTOFS_PENDING; - spin_unlock(&dentry->d_lock); + if (!oz_mode) { + spin_lock(&dentry->d_lock); + dentry->d_flags &= ~DCACHE_AUTOFS_PENDING; + spin_unlock(&dentry->d_lock); + } } /* @@ -659,9 +738,15 @@ static struct dentry *autofs4_lookup(struct inode *dir, struct dentry *dentry, s else dentry = ERR_PTR(-ENOENT); + if (unhashed) + dput(unhashed); + return dentry; } + if (unhashed) + return unhashed; + return NULL; } @@ -682,20 +767,30 @@ static int autofs4_dir_symlink(struct inode *dir, return -EACCES; ino = autofs4_init_ino(ino, sbi, S_IFLNK | 0555); - if (ino == NULL) - return -ENOSPC; + if (!ino) + return -ENOMEM; - ino->size = strlen(symname); - ino->u.symlink = cp = kmalloc(ino->size + 1, GFP_KERNEL); + spin_lock(&sbi->lookup_lock); + if (!list_empty(&ino->active)) + list_del_init(&ino->active); + spin_unlock(&sbi->lookup_lock); - if (cp == NULL) { - kfree(ino); - return -ENOSPC; + cp = kmalloc(ino->size + 1, GFP_KERNEL); + if (!cp) { + if (!dentry->d_fsdata) + kfree(ino); + return -ENOMEM; } strcpy(cp, symname); inode = autofs4_get_inode(dir->i_sb, ino); + if (!inode) { + kfree(cp); + if (!dentry->d_fsdata) + kfree(ino); + return -ENOMEM; + } d_add(dentry, inode); if (dir == dir->i_sb->s_root->d_inode) @@ -711,6 +806,8 @@ static int autofs4_dir_symlink(struct inode *dir, atomic_inc(&p_ino->count); ino->inode = inode; + ino->size = strlen(symname); + ino->u.symlink = cp; dir->i_mtime = CURRENT_TIME; return 0; @@ -755,7 +852,8 @@ static int autofs4_dir_unlink(struct inode *dir, struct dentry *dentry) spin_lock(&dcache_lock); spin_lock(&sbi->lookup_lock); - list_add(&ino->expiring, &sbi->expiring_list); + if (list_empty(&ino->expiring)) + list_add(&ino->expiring, &sbi->expiring_list); spin_unlock(&sbi->lookup_lock); spin_lock(&dentry->d_lock); __d_drop(dentry); @@ -783,7 +881,8 @@ static int autofs4_dir_rmdir(struct inode *dir, struct dentry *dentry) return -ENOTEMPTY; } spin_lock(&sbi->lookup_lock); - list_add(&ino->expiring, &sbi->expiring_list); + if (list_empty(&ino->expiring)) + list_add(&ino->expiring, &sbi->expiring_list); spin_unlock(&sbi->lookup_lock); spin_lock(&dentry->d_lock); __d_drop(dentry); @@ -819,10 +918,20 @@ static int autofs4_dir_mkdir(struct inode *dir, struct dentry *dentry, int mode) dentry, dentry->d_name.len, dentry->d_name.name); ino = autofs4_init_ino(ino, sbi, S_IFDIR | 0555); - if (ino == NULL) - return -ENOSPC; + if (!ino) + return -ENOMEM; + + spin_lock(&sbi->lookup_lock); + if (!list_empty(&ino->active)) + list_del_init(&ino->active); + spin_unlock(&sbi->lookup_lock); inode = autofs4_get_inode(dir->i_sb, ino); + if (!inode) { + if (!dentry->d_fsdata) + kfree(ino); + return -ENOMEM; + } d_add(dentry, inode); if (dir == dir->i_sb->s_root->d_inode) -- cgit v1.2.3-70-g09d2 From ef581a742874ebc4c28d24b374c78b762144ebdc Mon Sep 17 00:00:00 2001 From: Ian Kent Date: Wed, 23 Jul 2008 21:30:13 -0700 Subject: autofs4: fix symlink name allocation The length of the symlink name has been moved but it needs to be set before allocating space for it in the dentry info struct. This corrects a mistake in a recent patch. Signed-off-by: Ian Kent Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/root.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c index dbb70d5a488..324290c6827 100644 --- a/fs/autofs4/root.c +++ b/fs/autofs4/root.c @@ -775,6 +775,7 @@ static int autofs4_dir_symlink(struct inode *dir, list_del_init(&ino->active); spin_unlock(&sbi->lookup_lock); + ino->size = strlen(symname); cp = kmalloc(ino->size + 1, GFP_KERNEL); if (!cp) { if (!dentry->d_fsdata) @@ -806,7 +807,6 @@ static int autofs4_dir_symlink(struct inode *dir, atomic_inc(&p_ino->count); ino->inode = inode; - ino->size = strlen(symname); ino->u.symlink = cp; dir->i_mtime = CURRENT_TIME; -- cgit v1.2.3-70-g09d2 From c432c2586a0811c7d0030d78f0993568bc889a6f Mon Sep 17 00:00:00 2001 From: Ian Kent Date: Wed, 23 Jul 2008 21:30:14 -0700 Subject: autofs4: don't release directory mutex if called in oz_mode Since we now delay hashing of dentrys until the ->mkdir() call, droping and re-taking the directory mutex within the ->lookup() function when we are being called by user space is not needed. This can lead to a race when other processes are attempting to access the same directory during mount point directory creation. In this case we need to hang onto the mutex to ensure we don't get user processes trying to create a mount request for a newly created dentry after the mount point entry has already been created. This ensures that when we need to check a dentry passed to autofs4_wait(), if it is hashed, it is always the mount point dentry and not a new dentry created by another lookup during directory creation. Signed-off-by: Ian Kent Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/root.c | 11 +++++------ 1 file changed, 5 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c index 324290c6827..1e901e5ea01 100644 --- a/fs/autofs4/root.c +++ b/fs/autofs4/root.c @@ -686,12 +686,11 @@ static struct dentry *autofs4_lookup(struct inode *dir, struct dentry *dentry, s spin_lock(&dentry->d_lock); dentry->d_flags |= DCACHE_AUTOFS_PENDING; spin_unlock(&dentry->d_lock); - } - - if (dentry->d_op && dentry->d_op->d_revalidate) { - mutex_unlock(&dir->i_mutex); - (dentry->d_op->d_revalidate)(dentry, nd); - mutex_lock(&dir->i_mutex); + if (dentry->d_op && dentry->d_op->d_revalidate) { + mutex_unlock(&dir->i_mutex); + (dentry->d_op->d_revalidate)(dentry, nd); + mutex_lock(&dir->i_mutex); + } } /* -- cgit v1.2.3-70-g09d2 From 6d5cb926fa0162b1e62f37c117cc7ce763cfcbb9 Mon Sep 17 00:00:00 2001 From: Ian Kent Date: Wed, 23 Jul 2008 21:30:15 -0700 Subject: autofs4: use lookup intent flags to trigger mounts When an open(2) call is made on an autofs mount point directory that already exists and the O_DIRECTORY flag is not used the needed mount callback to the daemon is not done. This leads to the path walk continuing resulting in a callback to the daemon with an incorrect key. open(2) is called without O_DIRECTORY by the "find" utility but this should be handled properly anyway. This happens because autofs needs to use the lookup flags to decide when to callback to the daemon to perform a mount to prevent mount storms. For example, an autofs indirect mount map that has the "browse" option will have the mount point directories are pre-created and the stat(2) call made by a color ls against each directory will cause all these directories to be mounted. It is unfortunate we need to resort to this but mount maps can be quite large. Additionally, if a user manually umounts an autofs indirect mount the directory isn't removed which also leads to this situation. To resolve this autofs needs to use the lookup intent flags to enable it to make this decision. This patch adds this check and triggers a call back if any of the lookup intent flags are set as all these calls warrant a mount attempt be requested. I know that external VFS code which uses the lookup flags is something that the VFS would like to eliminate but I have no choice as I can't see any other way to do this. A VFS dentry or inode operation callback which returns the lookup "type" (requires a definition) would be sufficient. But this change is needed now and I'm not aware of the form that coming VFS changes will take so I'm not willing to propose anything along these lines. If anyone can provide an alternate method I would be happy to use it. [akpm@linux-foundation.org: fix build for concurrent VFS changes] Signed-off-by: Ian Kent Cc: Al Viro Cc: Jeff Moyer Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/root.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c index 1e901e5ea01..87352654ff4 100644 --- a/fs/autofs4/root.c +++ b/fs/autofs4/root.c @@ -31,6 +31,9 @@ static int autofs4_root_readdir(struct file * filp, void * dirent, filldir_t fil static struct dentry *autofs4_lookup(struct inode *,struct dentry *, struct nameidata *); static void *autofs4_follow_link(struct dentry *, struct nameidata *); +#define TRIGGER_FLAGS (LOOKUP_CONTINUE | LOOKUP_DIRECTORY) +#define TRIGGER_INTENTS (LOOKUP_OPEN | LOOKUP_CREATE) + const struct file_operations autofs4_root_operations = { .open = dcache_dir_open, .release = dcache_dir_close, @@ -291,7 +294,7 @@ static int try_to_fill_dentry(struct dentry *dentry, int flags) return status; } /* Trigger mount for path component or follow link */ - } else if (flags & (LOOKUP_CONTINUE | LOOKUP_DIRECTORY) || + } else if (flags & (TRIGGER_FLAGS | TRIGGER_INTENTS) || current->link_count) { DPRINTK("waiting for mount name=%.*s", dentry->d_name.len, dentry->d_name.name); @@ -336,7 +339,7 @@ static void *autofs4_follow_link(struct dentry *dentry, struct nameidata *nd) nd->flags); /* If it's our master or we shouldn't trigger a mount we're done */ - lookup_type = nd->flags & (LOOKUP_CONTINUE | LOOKUP_DIRECTORY); + lookup_type = nd->flags & (TRIGGER_FLAGS | TRIGGER_INTENTS); if (oz_mode || !lookup_type) goto done; -- cgit v1.2.3-70-g09d2 From 70b52a0a5005ce6a0ceec56e97222437a0ba7506 Mon Sep 17 00:00:00 2001 From: Jeff Moyer Date: Wed, 23 Jul 2008 21:30:16 -0700 Subject: autofs4: use struct qstr in waitq.c The autofs_wait_queue already contains all of the fields of the struct qstr, so change it into a qstr. This patch, from Jeff Moyer, has been modified a liitle by myself. Signed-off-by: Jeff Moyer Signed-off-by: Ian Kent Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/autofs_i.h | 4 +-- fs/autofs4/waitq.c | 86 +++++++++++++++++++++++++++------------------------ 2 files changed, 46 insertions(+), 44 deletions(-) (limited to 'fs') diff --git a/fs/autofs4/autofs_i.h b/fs/autofs4/autofs_i.h index 2dce2334737..da8882ff31e 100644 --- a/fs/autofs4/autofs_i.h +++ b/fs/autofs4/autofs_i.h @@ -75,9 +75,7 @@ struct autofs_wait_queue { struct autofs_wait_queue *next; autofs_wqt_t wait_queue_token; /* We use the following to see what we are waiting for */ - unsigned int hash; - unsigned int len; - char *name; + struct qstr name; u32 dev; u64 ino; uid_t uid; diff --git a/fs/autofs4/waitq.c b/fs/autofs4/waitq.c index 75e5955c3f6..5208cfb1df4 100644 --- a/fs/autofs4/waitq.c +++ b/fs/autofs4/waitq.c @@ -36,8 +36,10 @@ void autofs4_catatonic_mode(struct autofs_sb_info *sbi) while (wq) { nwq = wq->next; wq->status = -ENOENT; /* Magic is gone - report failure */ - kfree(wq->name); - wq->name = NULL; + if (wq->name.name) { + kfree(wq->name.name); + wq->name.name = NULL; + } wake_up_interruptible(&wq->queue); wq = nwq; } @@ -92,7 +94,7 @@ static void autofs4_notify_daemon(struct autofs_sb_info *sbi, size_t pktsz; DPRINTK("wait id = 0x%08lx, name = %.*s, type=%d", - wq->wait_queue_token, wq->len, wq->name, type); + wq->wait_queue_token, wq->name.len, wq->name.name, type); memset(&pkt,0,sizeof pkt); /* For security reasons */ @@ -107,9 +109,9 @@ static void autofs4_notify_daemon(struct autofs_sb_info *sbi, pktsz = sizeof(*mp); mp->wait_queue_token = wq->wait_queue_token; - mp->len = wq->len; - memcpy(mp->name, wq->name, wq->len); - mp->name[wq->len] = '\0'; + mp->len = wq->name.len; + memcpy(mp->name, wq->name.name, wq->name.len); + mp->name[wq->name.len] = '\0'; break; } case autofs_ptype_expire_multi: @@ -119,9 +121,9 @@ static void autofs4_notify_daemon(struct autofs_sb_info *sbi, pktsz = sizeof(*ep); ep->wait_queue_token = wq->wait_queue_token; - ep->len = wq->len; - memcpy(ep->name, wq->name, wq->len); - ep->name[wq->len] = '\0'; + ep->len = wq->name.len; + memcpy(ep->name, wq->name.name, wq->name.len); + ep->name[wq->name.len] = '\0'; break; } /* @@ -138,9 +140,9 @@ static void autofs4_notify_daemon(struct autofs_sb_info *sbi, pktsz = sizeof(*packet); packet->wait_queue_token = wq->wait_queue_token; - packet->len = wq->len; - memcpy(packet->name, wq->name, wq->len); - packet->name[wq->len] = '\0'; + packet->len = wq->name.len; + memcpy(packet->name, wq->name.name, wq->name.len); + packet->name[wq->name.len] = '\0'; packet->dev = wq->dev; packet->ino = wq->ino; packet->uid = wq->uid; @@ -191,15 +193,15 @@ static int autofs4_getpath(struct autofs_sb_info *sbi, } static struct autofs_wait_queue * -autofs4_find_wait(struct autofs_sb_info *sbi, - char *name, unsigned int hash, unsigned int len) +autofs4_find_wait(struct autofs_sb_info *sbi, struct qstr *qstr) { struct autofs_wait_queue *wq; for (wq = sbi->queues; wq; wq = wq->next) { - if (wq->hash == hash && - wq->len == len && - wq->name && !memcmp(wq->name, name, len)) + if (wq->name.hash == qstr->hash && + wq->name.len == qstr->len && + wq->name.name && + !memcmp(wq->name.name, qstr->name, qstr->len)) break; } return wq; @@ -210,9 +212,8 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry, { struct autofs_info *ino; struct autofs_wait_queue *wq; + struct qstr qstr; char *name; - unsigned int len = 0; - unsigned int hash = 0; int status, type; /* In catatonic mode, we don't wait for nobody */ @@ -225,22 +226,23 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry, /* If this is a direct mount request create a dummy name */ if (IS_ROOT(dentry) && (sbi->type & AUTOFS_TYPE_DIRECT)) - len = sprintf(name, "%p", dentry); + qstr.len = sprintf(name, "%p", dentry); else { - len = autofs4_getpath(sbi, dentry, &name); - if (!len) { + qstr.len = autofs4_getpath(sbi, dentry, &name); + if (!qstr.len) { kfree(name); return -ENOENT; } } - hash = full_name_hash(name, len); + qstr.name = name; + qstr.hash = full_name_hash(name, qstr.len); if (mutex_lock_interruptible(&sbi->wq_mutex)) { - kfree(name); + kfree(qstr.name); return -EINTR; } - wq = autofs4_find_wait(sbi, name, hash, len); + wq = autofs4_find_wait(sbi, &qstr); ino = autofs4_dentry_ino(dentry); if (!wq && ino && notify == NFY_NONE) { /* @@ -254,10 +256,10 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry, mutex_unlock(&sbi->wq_mutex); schedule_timeout_interruptible(HZ/10); if (mutex_lock_interruptible(&sbi->wq_mutex)) { - kfree(name); + kfree(qstr.name); return -EINTR; } - wq = autofs4_find_wait(sbi, name, hash, len); + wq = autofs4_find_wait(sbi, &qstr); if (wq) break; } @@ -268,7 +270,7 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry, * return status of the wait. */ if (!wq) { - kfree(name); + kfree(qstr.name); mutex_unlock(&sbi->wq_mutex); return 0; } @@ -278,7 +280,7 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry, /* Create a new wait queue */ wq = kmalloc(sizeof(struct autofs_wait_queue),GFP_KERNEL); if (!wq) { - kfree(name); + kfree(qstr.name); mutex_unlock(&sbi->wq_mutex); return -ENOMEM; } @@ -289,9 +291,7 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry, wq->next = sbi->queues; sbi->queues = wq; init_waitqueue_head(&wq->queue); - wq->hash = hash; - wq->name = name; - wq->len = len; + memcpy(&wq->name, &qstr, sizeof(struct qstr)); wq->dev = autofs4_get_dev(sbi); wq->ino = autofs4_get_ino(sbi); wq->uid = current->uid; @@ -319,16 +319,18 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry, } DPRINTK("new wait id = 0x%08lx, name = %.*s, nfy=%d\n", - (unsigned long) wq->wait_queue_token, wq->len, wq->name, notify); + (unsigned long) wq->wait_queue_token, wq->name.len, + wq->name.name, notify); /* autofs4_notify_daemon() may block */ autofs4_notify_daemon(sbi, wq, type); } else { atomic_inc(&wq->wait_ctr); mutex_unlock(&sbi->wq_mutex); - kfree(name); + kfree(qstr.name); DPRINTK("existing wait id = 0x%08lx, name = %.*s, nfy=%d", - (unsigned long) wq->wait_queue_token, wq->len, wq->name, notify); + (unsigned long) wq->wait_queue_token, wq->name.len, + wq->name.name, notify); } /* wq->name is NULL if and only if the lock is already released */ @@ -336,11 +338,13 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry, if (sbi->catatonic) { /* We might have slept, so check again for catatonic mode */ wq->status = -ENOENT; - kfree(wq->name); - wq->name = NULL; + if (wq->name.name) { + kfree(wq->name.name); + wq->name.name = NULL; + } } - if (wq->name) { + if (wq->name.name) { /* Block all but "shutdown" signals while waiting */ sigset_t oldset; unsigned long irqflags; @@ -351,7 +355,7 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry, recalc_sigpending(); spin_unlock_irqrestore(¤t->sighand->siglock, irqflags); - wait_event_interruptible(wq->queue, wq->name == NULL); + wait_event_interruptible(wq->queue, wq->name.name == NULL); spin_lock_irqsave(¤t->sighand->siglock, irqflags); current->blocked = oldset; @@ -388,8 +392,8 @@ int autofs4_wait_release(struct autofs_sb_info *sbi, autofs_wqt_t wait_queue_tok *wql = wq->next; /* Unlink from chain */ mutex_unlock(&sbi->wq_mutex); - kfree(wq->name); - wq->name = NULL; /* Do not wait on this queue */ + kfree(wq->name.name); + wq->name.name = NULL; /* Do not wait on this queue */ wq->status = status; -- cgit v1.2.3-70-g09d2 From 5a11d4d0ee1ff284271f7265929d07ea4a1168a6 Mon Sep 17 00:00:00 2001 From: Ian Kent Date: Wed, 23 Jul 2008 21:30:17 -0700 Subject: autofs4: fix waitq locking The autofs4_catatonic_mode() function accesses the wait queue without any locking but can be called at any time. This could lead to a possible double free of the name field of the wait and a double fput of the daemon communication pipe or an fput of a NULL file pointer. Signed-off-by: Ian Kent Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/inode.c | 4 ++-- fs/autofs4/waitq.c | 23 ++++++++++++----------- 2 files changed, 14 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/autofs4/inode.c b/fs/autofs4/inode.c index e3e70994ab4..7bb3e5ba053 100644 --- a/fs/autofs4/inode.c +++ b/fs/autofs4/inode.c @@ -163,8 +163,8 @@ void autofs4_kill_sb(struct super_block *sb) if (!sbi) goto out_kill_sb; - if (!sbi->catatonic) - autofs4_catatonic_mode(sbi); /* Free wait queues, close pipe */ + /* Free wait queues, close pipe */ + autofs4_catatonic_mode(sbi); /* Clean up and release dangling references */ autofs4_force_release(sbi); diff --git a/fs/autofs4/waitq.c b/fs/autofs4/waitq.c index 5208cfb1df4..55aac10cf32 100644 --- a/fs/autofs4/waitq.c +++ b/fs/autofs4/waitq.c @@ -28,6 +28,12 @@ void autofs4_catatonic_mode(struct autofs_sb_info *sbi) { struct autofs_wait_queue *wq, *nwq; + mutex_lock(&sbi->wq_mutex); + if (sbi->catatonic) { + mutex_unlock(&sbi->wq_mutex); + return; + } + DPRINTK("entering catatonic mode"); sbi->catatonic = 1; @@ -45,6 +51,8 @@ void autofs4_catatonic_mode(struct autofs_sb_info *sbi) } fput(sbi->pipe); /* Close the pipe */ sbi->pipe = NULL; + sbi->pipefd = -1; + mutex_unlock(&sbi->wq_mutex); } static int autofs4_write(struct file *file, const void *addr, int bytes) @@ -333,17 +341,10 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry, wq->name.name, notify); } - /* wq->name is NULL if and only if the lock is already released */ - - if (sbi->catatonic) { - /* We might have slept, so check again for catatonic mode */ - wq->status = -ENOENT; - if (wq->name.name) { - kfree(wq->name.name); - wq->name.name = NULL; - } - } - + /* + * wq->name.name is NULL iff the lock is already released + * or the mount has been made catatonic. + */ if (wq->name.name) { /* Block all but "shutdown" signals while waiting */ sigset_t oldset; -- cgit v1.2.3-70-g09d2 From a1362fe92f1bde687b3a9e93d6b8d105d0a84f74 Mon Sep 17 00:00:00 2001 From: Ian Kent Date: Wed, 23 Jul 2008 21:30:19 -0700 Subject: autofs4: fix pending mount race Close a race between a pending mount that is about to finish and a new lookup for the same directory. Process P1 triggers a mount of directory foo. It sets DCACHE_AUTOFS_PENDING in the ->lookup routine, creates a waitq entry for 'foo', and calls out to the daemon to perform the mount. The autofs daemon will then create the directory 'foo', using a new dentry that will be hashed in the dcache. Before the mount completes, another process, P2, tries to walk into the 'foo' directory. The vfs path walking code finds an entry for 'foo' and calls the revalidate method. Revalidate finds that the entry is not PENDING (because PENDING was never set on the dentry created by the mkdir), but it does find the directory is empty. Revalidate calls try_to_fill_dentry, which sets the PENDING flag and then calls into the autofs4 wait code to trigger or wait for a mount of 'foo'. The wait code finds the entry for 'foo' and goes to sleep waiting for the completion of the mount. Yet another process, P3, tries to walk into the 'foo' directory. This process again finds a dentry in the dcache for 'foo', and calls into the autofs revalidate code. The revalidate code finds that the PENDING flag is set, and so calls try_to_fill_dentry. a) try_to_fill_dentry sets the PENDING flag redundantly for this dentry, then calls into the autofs4 wait code. b) the autofs4 wait code takes the waitq mutex and searches for an entry for 'foo' Between a and b, P1 is woken up because the mount completed. P1 takes the wait queue mutex, clears the PENDING flag from the dentry, and removes the waitqueue entry for 'foo' from the list. When it releases the waitq mutex, P3 (eventually) acquires it. At this time, it looks for an existing waitq for 'foo', finds none, and so creates a new one and calls out to the daemon to mount the 'foo' directory. Now, the reason that three processes are required to trigger this race is that, because the PENDING flag is not set on the dentry created by mkdir, the window for the race would be way to slim for it to ever occur. Basically, between the testing of d_mountpoint(dentry) and the taking of the waitq mutex, the mount would have to complete and the daemon would have to be woken up, and that in turn would have to wake up P1. This is simply impossible. Add the third process, though, and it becomes slightly more likely. Signed-off-by: Jeff Moyer Signed-off-by: Ian Kent Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/waitq.c | 135 ++++++++++++++++++++++++++++++++++++++--------------- 1 file changed, 97 insertions(+), 38 deletions(-) (limited to 'fs') diff --git a/fs/autofs4/waitq.c b/fs/autofs4/waitq.c index 55aac10cf32..cd3b2a67169 100644 --- a/fs/autofs4/waitq.c +++ b/fs/autofs4/waitq.c @@ -215,19 +215,106 @@ autofs4_find_wait(struct autofs_sb_info *sbi, struct qstr *qstr) return wq; } +/* + * Check if we have a valid request. + * Returns + * 1 if the request should continue. + * In this case we can return an autofs_wait_queue entry if one is + * found or NULL to idicate a new wait needs to be created. + * 0 or a negative errno if the request shouldn't continue. + */ +static int validate_request(struct autofs_wait_queue **wait, + struct autofs_sb_info *sbi, + struct qstr *qstr, + struct dentry*dentry, enum autofs_notify notify) +{ + struct autofs_wait_queue *wq; + struct autofs_info *ino; + + /* Wait in progress, continue; */ + wq = autofs4_find_wait(sbi, qstr); + if (wq) { + *wait = wq; + return 1; + } + + *wait = NULL; + + /* If we don't yet have any info this is a new request */ + ino = autofs4_dentry_ino(dentry); + if (!ino) + return 1; + + /* + * If we've been asked to wait on an existing expire (NFY_NONE) + * but there is no wait in the queue ... + */ + if (notify == NFY_NONE) { + /* + * Either we've betean the pending expire to post it's + * wait or it finished while we waited on the mutex. + * So we need to wait till either, the wait appears + * or the expire finishes. + */ + + while (ino->flags & AUTOFS_INF_EXPIRING) { + mutex_unlock(&sbi->wq_mutex); + schedule_timeout_interruptible(HZ/10); + if (mutex_lock_interruptible(&sbi->wq_mutex)) + return -EINTR; + + wq = autofs4_find_wait(sbi, qstr); + if (wq) { + *wait = wq; + return 1; + } + } + + /* + * Not ideal but the status has already gone. Of the two + * cases where we wait on NFY_NONE neither depend on the + * return status of the wait. + */ + return 0; + } + + /* + * If we've been asked to trigger a mount and the request + * completed while we waited on the mutex ... + */ + if (notify == NFY_MOUNT) { + /* + * If the dentry isn't hashed just go ahead and try the + * mount again with a new wait (not much else we can do). + */ + if (!d_unhashed(dentry)) { + /* + * But if the dentry is hashed, that means that we + * got here through the revalidate path. Thus, we + * need to check if the dentry has been mounted + * while we waited on the wq_mutex. If it has, + * simply return success. + */ + if (d_mountpoint(dentry)) + return 0; + } + } + + return 1; +} + int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry, enum autofs_notify notify) { - struct autofs_info *ino; struct autofs_wait_queue *wq; struct qstr qstr; char *name; - int status, type; + int status, ret, type; /* In catatonic mode, we don't wait for nobody */ if (sbi->catatonic) return -ENOENT; - + name = kmalloc(NAME_MAX + 1, GFP_KERNEL); if (!name) return -ENOMEM; @@ -245,43 +332,15 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry, qstr.name = name; qstr.hash = full_name_hash(name, qstr.len); - if (mutex_lock_interruptible(&sbi->wq_mutex)) { - kfree(qstr.name); + if (mutex_lock_interruptible(&sbi->wq_mutex)) return -EINTR; - } - - wq = autofs4_find_wait(sbi, &qstr); - ino = autofs4_dentry_ino(dentry); - if (!wq && ino && notify == NFY_NONE) { - /* - * Either we've betean the pending expire to post it's - * wait or it finished while we waited on the mutex. - * So we need to wait till either, the wait appears - * or the expire finishes. - */ - while (ino->flags & AUTOFS_INF_EXPIRING) { - mutex_unlock(&sbi->wq_mutex); - schedule_timeout_interruptible(HZ/10); - if (mutex_lock_interruptible(&sbi->wq_mutex)) { - kfree(qstr.name); - return -EINTR; - } - wq = autofs4_find_wait(sbi, &qstr); - if (wq) - break; - } - - /* - * Not ideal but the status has already gone. Of the two - * cases where we wait on NFY_NONE neither depend on the - * return status of the wait. - */ - if (!wq) { - kfree(qstr.name); + ret = validate_request(&wq, sbi, &qstr, dentry, notify); + if (ret <= 0) { + if (ret == 0) mutex_unlock(&sbi->wq_mutex); - return 0; - } + kfree(qstr.name); + return ret; } if (!wq) { @@ -392,9 +451,9 @@ int autofs4_wait_release(struct autofs_sb_info *sbi, autofs_wqt_t wait_queue_tok } *wql = wq->next; /* Unlink from chain */ - mutex_unlock(&sbi->wq_mutex); kfree(wq->name.name); wq->name.name = NULL; /* Do not wait on this queue */ + mutex_unlock(&sbi->wq_mutex); wq->status = status; -- cgit v1.2.3-70-g09d2 From f4c7da02615bebcaf89f15a8d055922f515160b8 Mon Sep 17 00:00:00 2001 From: Ian Kent Date: Wed, 23 Jul 2008 21:30:19 -0700 Subject: autofs4: add missing kfree It see that the patch tittled "autofs4 - fix pending mount race" is missing a change that I had recently made. It's missing a kfree for the case mutex_lock_interruptible() fails to aquire the wait queue mutex. Signed-off-by: Ian Kent Cc: Jeff Moyer Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/waitq.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/autofs4/waitq.c b/fs/autofs4/waitq.c index cd3b2a67169..1132cc2a031 100644 --- a/fs/autofs4/waitq.c +++ b/fs/autofs4/waitq.c @@ -332,8 +332,10 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry, qstr.name = name; qstr.hash = full_name_hash(name, qstr.len); - if (mutex_lock_interruptible(&sbi->wq_mutex)) + if (mutex_lock_interruptible(&sbi->wq_mutex)) { + kfree(qstr.name); return -EINTR; + } ret = validate_request(&wq, sbi, &qstr, dentry, notify); if (ret <= 0) { -- cgit v1.2.3-70-g09d2 From e64be33ccaceaca67c84237dff8805b861398eab Mon Sep 17 00:00:00 2001 From: Ian Kent Date: Wed, 23 Jul 2008 21:30:20 -0700 Subject: autofs4: check kernel communication pipe is valid for write It is possible for an autofs mount to become catatonic (and for the daemon communication pipe to become NULL) after a wait has been initiallized but before the request has been sent to the daemon. We need to check for this before sending the request packet. Signed-off-by: Ian Kent Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/waitq.c | 16 ++++++++++++++-- 1 file changed, 14 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/autofs4/waitq.c b/fs/autofs4/waitq.c index 1132cc2a031..dd2914d7ad7 100644 --- a/fs/autofs4/waitq.c +++ b/fs/autofs4/waitq.c @@ -99,6 +99,7 @@ static void autofs4_notify_daemon(struct autofs_sb_info *sbi, union autofs_packet_union v4_pkt; union autofs_v5_packet_union v5_pkt; } pkt; + struct file *pipe = NULL; size_t pktsz; DPRINTK("wait id = 0x%08lx, name = %.*s, type=%d", @@ -164,8 +165,19 @@ static void autofs4_notify_daemon(struct autofs_sb_info *sbi, return; } - if (autofs4_write(sbi->pipe, &pkt, pktsz)) - autofs4_catatonic_mode(sbi); + /* Check if we have become catatonic */ + mutex_lock(&sbi->wq_mutex); + if (!sbi->catatonic) { + pipe = sbi->pipe; + get_file(pipe); + } + mutex_unlock(&sbi->wq_mutex); + + if (pipe) { + if (autofs4_write(pipe, &pkt, pktsz)) + autofs4_catatonic_mode(sbi); + fput(pipe); + } } static int autofs4_getpath(struct autofs_sb_info *sbi, -- cgit v1.2.3-70-g09d2 From 296f7bf78bc5c7a4d772aea580ce800d14040d1a Mon Sep 17 00:00:00 2001 From: Ian Kent Date: Wed, 23 Jul 2008 21:30:21 -0700 Subject: autofs4: fix waitq memory leak If an autofs mount becomes catatonic before autofs4_wait_release() is called the wait queue counter will not be decremented down to zero and the entry will never be freed. There are also races decrementing the wait counter in the wait release function. To deal with this the counter needs to be updated while holding the wait queue mutex and waiters need to be woken up unconditionally when the wait is removed from the queue to ensure we eventually free the wait. Signed-off-by: Ian Kent Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/autofs_i.h | 2 +- fs/autofs4/waitq.c | 18 +++++++++--------- 2 files changed, 10 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/autofs4/autofs_i.h b/fs/autofs4/autofs_i.h index da8882ff31e..058e1800cae 100644 --- a/fs/autofs4/autofs_i.h +++ b/fs/autofs4/autofs_i.h @@ -84,7 +84,7 @@ struct autofs_wait_queue { pid_t tgid; /* This is for status reporting upon return */ int status; - atomic_t wait_ctr; + unsigned int wait_ctr; }; #define AUTOFS_SBI_MAGIC 0x6d4a556d diff --git a/fs/autofs4/waitq.c b/fs/autofs4/waitq.c index dd2914d7ad7..3458dbc8fff 100644 --- a/fs/autofs4/waitq.c +++ b/fs/autofs4/waitq.c @@ -46,6 +46,7 @@ void autofs4_catatonic_mode(struct autofs_sb_info *sbi) kfree(wq->name.name); wq->name.name = NULL; } + wq->wait_ctr--; wake_up_interruptible(&wq->queue); wq = nwq; } @@ -380,7 +381,7 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry, wq->pid = current->pid; wq->tgid = current->tgid; wq->status = -EINTR; /* Status return if interrupted */ - atomic_set(&wq->wait_ctr, 2); + wq->wait_ctr = 2; mutex_unlock(&sbi->wq_mutex); if (sbi->version < 5) { @@ -406,7 +407,7 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry, /* autofs4_notify_daemon() may block */ autofs4_notify_daemon(sbi, wq, type); } else { - atomic_inc(&wq->wait_ctr); + wq->wait_ctr++; mutex_unlock(&sbi->wq_mutex); kfree(qstr.name); DPRINTK("existing wait id = 0x%08lx, name = %.*s, nfy=%d", @@ -442,8 +443,10 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry, status = wq->status; /* Are we the last process to need status? */ - if (atomic_dec_and_test(&wq->wait_ctr)) + mutex_lock(&sbi->wq_mutex); + if (!--wq->wait_ctr) kfree(wq); + mutex_unlock(&sbi->wq_mutex); return status; } @@ -467,14 +470,11 @@ int autofs4_wait_release(struct autofs_sb_info *sbi, autofs_wqt_t wait_queue_tok *wql = wq->next; /* Unlink from chain */ kfree(wq->name.name); wq->name.name = NULL; /* Do not wait on this queue */ - mutex_unlock(&sbi->wq_mutex); - wq->status = status; - - if (atomic_dec_and_test(&wq->wait_ctr)) /* Is anyone still waiting for this guy? */ + wake_up_interruptible(&wq->queue); + if (!--wq->wait_ctr) kfree(wq); - else - wake_up_interruptible(&wq->queue); + mutex_unlock(&sbi->wq_mutex); return 0; } -- cgit v1.2.3-70-g09d2 From eb3b176796b0e53fd26fce86847231542eb0d198 Mon Sep 17 00:00:00 2001 From: Ian Kent Date: Wed, 23 Jul 2008 21:30:22 -0700 Subject: autofs4: detect invalid direct mount requests autofs v5 direct and offset mounts within an autofs filesystem are triggered by existing autofs triger mounts so the mount point dentry must be positive. If the mount point dentry is negative then the trigger doesn't exist so we can return fail immediately. Signed-off-by: Ian Kent Cc: Jeff Moyer Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/waitq.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'fs') diff --git a/fs/autofs4/waitq.c b/fs/autofs4/waitq.c index 3458dbc8fff..bcb6c526546 100644 --- a/fs/autofs4/waitq.c +++ b/fs/autofs4/waitq.c @@ -328,6 +328,10 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry, if (sbi->catatonic) return -ENOENT; + if (!dentry->d_inode && + (sbi->type & (AUTOFS_TYPE_DIRECT | AUTOFS_TYPE_OFFSET))) + return -ENOENT; + name = kmalloc(NAME_MAX + 1, GFP_KERNEL); if (!name) return -ENOMEM; -- cgit v1.2.3-70-g09d2 From c72305b5472522299bb6f45b736080128eb1c822 Mon Sep 17 00:00:00 2001 From: Ian Kent Date: Wed, 23 Jul 2008 21:30:23 -0700 Subject: autofs4: indirect dentry must almost always be positive We have been seeing mount requests comming to the automount daemon for keys of the form "/" which are lookups for invalid map keys. But we can check for this in the kernel module and return a fail immediately, without having to send a request to the daemon. It is possible to recognise these requests are invalid based on whether the request dentry is negative and its relation to the autofs file system root. For example, given the indirect multi-mount map entry: idm1 \ /mm1 :/ /mm2 :/ For a request to mount idm1, IS_ROOT((idm1)->d_parent) will be always be true and the dentry may be negative. But directories idm1/mm1 and idm1/mm2 will always be created as part of the mount request for idm1. So any mount request within idm1 itself must have a positive dentry otherwise the map key is invalid. In version 4 these multi-mount entries are all mounted and umounted as a single request and in version 5 the directories idm1/mm1 and idm1/mm2 are created and an autofs fs mounted on them to act as a mount trigger so the above is also true. This also holds true for the autofs version 4 pseudo direct mount feature. When this feature is used without the "--ghost" option automount(8) will create internal submounts as we go down the map key paths which are essentially normal indirect mounts for which the above holds. If the "--ghost" option is given the directories for map keys are created at daemon startup so valid map entries correspond to postive dentries in the autofs fs. autofs version 5 direct mount maps are similar except that the IS_ROOT check is not needed. This has been addressed in a previous patch tittled "autofs4 - detect invalid direct mount requests". For example, given the direct multi-mount map entry: /test/dm1 \ /mm1 :/ /mm2 :/ An autofs fs is mounted on /test/dm1 as a trigger mount and when a mount is triggered for /test/dm1, the multi-mount offset directories /test/dm1/mm1 and /test/dm1/mm2 are created and an autofs fs is mounted on them to act as mount triggers. So valid direct mount requests must always have a positive dentry if they correspond to a valid map entry. Signed-off-by: Ian Kent Acked-by: Jeff Moyer Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/waitq.c | 17 ++++++++++++++--- 1 file changed, 14 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/autofs4/waitq.c b/fs/autofs4/waitq.c index bcb6c526546..35216d18d8b 100644 --- a/fs/autofs4/waitq.c +++ b/fs/autofs4/waitq.c @@ -328,9 +328,20 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry, if (sbi->catatonic) return -ENOENT; - if (!dentry->d_inode && - (sbi->type & (AUTOFS_TYPE_DIRECT | AUTOFS_TYPE_OFFSET))) - return -ENOENT; + if (!dentry->d_inode) { + /* + * A wait for a negative dentry is invalid for certain + * cases. A direct or offset mount "always" has its mount + * point directory created and so the request dentry must + * be positive or the map key doesn't exist. The situation + * is very similar for indirect mounts except only dentrys + * in the root of the autofs file system may be negative. + */ + if (sbi->type & (AUTOFS_TYPE_DIRECT|AUTOFS_TYPE_OFFSET)) + return -ENOENT; + else if (!IS_ROOT(dentry->d_parent)) + return -ENOENT; + } name = kmalloc(NAME_MAX + 1, GFP_KERNEL); if (!name) -- cgit v1.2.3-70-g09d2 From ff9cd499d6258952385cb2f12e9a3c0908fd5786 Mon Sep 17 00:00:00 2001 From: Ian Kent Date: Wed, 23 Jul 2008 21:30:24 -0700 Subject: autofs4: cleanup redundant readir code The mount triggering functionality of readdir and related functions is no longer used (and is quite broken as well). The unused portions have been removed. Signed-off-by: Ian Kent Reviewed-by: Jeff Moyer Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/root.c | 149 ++++++------------------------------------------------ 1 file changed, 16 insertions(+), 133 deletions(-) (limited to 'fs') diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c index 87352654ff4..51c873ca8e8 100644 --- a/fs/autofs4/root.c +++ b/fs/autofs4/root.c @@ -25,8 +25,6 @@ static int autofs4_dir_rmdir(struct inode *,struct dentry *); static int autofs4_dir_mkdir(struct inode *,struct dentry *,int); static int autofs4_root_ioctl(struct inode *, struct file *,unsigned int,unsigned long); static int autofs4_dir_open(struct inode *inode, struct file *file); -static int autofs4_dir_close(struct inode *inode, struct file *file); -static int autofs4_dir_readdir(struct file * filp, void * dirent, filldir_t filldir); static int autofs4_root_readdir(struct file * filp, void * dirent, filldir_t filldir); static struct dentry *autofs4_lookup(struct inode *,struct dentry *, struct nameidata *); static void *autofs4_follow_link(struct dentry *, struct nameidata *); @@ -44,9 +42,9 @@ const struct file_operations autofs4_root_operations = { const struct file_operations autofs4_dir_operations = { .open = autofs4_dir_open, - .release = autofs4_dir_close, + .release = dcache_dir_close, .read = generic_read_dir, - .readdir = autofs4_dir_readdir, + .readdir = dcache_readdir, }; const struct inode_operations autofs4_indirect_root_inode_operations = { @@ -98,17 +96,7 @@ static int autofs4_root_readdir(struct file *file, void *dirent, static int autofs4_dir_open(struct inode *inode, struct file *file) { struct dentry *dentry = file->f_path.dentry; - struct vfsmount *mnt = file->f_path.mnt; struct autofs_sb_info *sbi = autofs4_sbi(dentry->d_sb); - struct dentry *cursor; - int status; - - status = dcache_dir_open(inode, file); - if (status) - goto out; - - cursor = file->private_data; - cursor->d_fsdata = NULL; DPRINTK("file=%p dentry=%p %.*s", file, dentry, dentry->d_name.len, dentry->d_name.name); @@ -116,129 +104,24 @@ static int autofs4_dir_open(struct inode *inode, struct file *file) if (autofs4_oz_mode(sbi)) goto out; - if (autofs4_ispending(dentry)) { - DPRINTK("dentry busy"); - dcache_dir_close(inode, file); - status = -EBUSY; - goto out; - } - - status = -ENOENT; - if (!d_mountpoint(dentry) && dentry->d_op && dentry->d_op->d_revalidate) { - struct nameidata nd; - int empty, ret; - - /* In case there are stale directory dentrys from a failed mount */ - spin_lock(&dcache_lock); - empty = list_empty(&dentry->d_subdirs); + /* + * An empty directory in an autofs file system is always a + * mount point. The daemon must have failed to mount this + * during lookup so it doesn't exist. This can happen, for + * example, if user space returns an incorrect status for a + * mount request. Otherwise we're doing a readdir on the + * autofs file system so just let the libfs routines handle + * it. + */ + spin_lock(&dcache_lock); + if (!d_mountpoint(dentry) && __simple_empty(dentry)) { spin_unlock(&dcache_lock); - - if (!empty) - d_invalidate(dentry); - - nd.flags = LOOKUP_DIRECTORY; - ret = (dentry->d_op->d_revalidate)(dentry, &nd); - - if (ret <= 0) { - if (ret < 0) - status = ret; - dcache_dir_close(inode, file); - goto out; - } + return -ENOENT; } + spin_unlock(&dcache_lock); - if (d_mountpoint(dentry)) { - struct file *fp = NULL; - struct path fp_path = { .dentry = dentry, .mnt = mnt }; - - path_get(&fp_path); - - if (!autofs4_follow_mount(&fp_path.mnt, &fp_path.dentry)) { - path_put(&fp_path); - dcache_dir_close(inode, file); - goto out; - } - - fp = dentry_open(fp_path.dentry, fp_path.mnt, file->f_flags); - status = PTR_ERR(fp); - if (IS_ERR(fp)) { - dcache_dir_close(inode, file); - goto out; - } - cursor->d_fsdata = fp; - } - return 0; -out: - return status; -} - -static int autofs4_dir_close(struct inode *inode, struct file *file) -{ - struct dentry *dentry = file->f_path.dentry; - struct autofs_sb_info *sbi = autofs4_sbi(dentry->d_sb); - struct dentry *cursor = file->private_data; - int status = 0; - - DPRINTK("file=%p dentry=%p %.*s", - file, dentry, dentry->d_name.len, dentry->d_name.name); - - if (autofs4_oz_mode(sbi)) - goto out; - - if (autofs4_ispending(dentry)) { - DPRINTK("dentry busy"); - status = -EBUSY; - goto out; - } - - if (d_mountpoint(dentry)) { - struct file *fp = cursor->d_fsdata; - if (!fp) { - status = -ENOENT; - goto out; - } - filp_close(fp, current->files); - } -out: - dcache_dir_close(inode, file); - return status; -} - -static int autofs4_dir_readdir(struct file *file, void *dirent, filldir_t filldir) -{ - struct dentry *dentry = file->f_path.dentry; - struct autofs_sb_info *sbi = autofs4_sbi(dentry->d_sb); - struct dentry *cursor = file->private_data; - int status; - - DPRINTK("file=%p dentry=%p %.*s", - file, dentry, dentry->d_name.len, dentry->d_name.name); - - if (autofs4_oz_mode(sbi)) - goto out; - - if (autofs4_ispending(dentry)) { - DPRINTK("dentry busy"); - return -EBUSY; - } - - if (d_mountpoint(dentry)) { - struct file *fp = cursor->d_fsdata; - - if (!fp) - return -ENOENT; - - if (!fp->f_op || !fp->f_op->readdir) - goto out; - - status = vfs_readdir(fp, filldir, dirent); - file->f_pos = fp->f_pos; - if (status) - autofs4_copy_atime(file, fp); - return status; - } out: - return dcache_readdir(file, dirent, filldir); + return dcache_dir_open(inode, file); } static int try_to_fill_dentry(struct dentry *dentry, int flags) -- cgit v1.2.3-70-g09d2 From 26e81b3142f1ba497d4cd0365c13661684b784ce Mon Sep 17 00:00:00 2001 From: Ian Kent Date: Wed, 23 Jul 2008 21:30:25 -0700 Subject: autofs4: fix pending checks There are two cases for which a dentry that has a pending mount request does not wait for completion. One is via autofs4_revalidate() and the other via autofs4_follow_link(). In revalidate, after the mount point directory is created, but before the mount is done, the check in try_to_fill_dentry() can can fail to send the dentry to the wait queue since the dentry is positive and the lookup flags may contain only LOOKUP_FOLLOW. Although we don't trigger a mount for the LOOKUP_FOLLOW flag, if ther's one pending we might as well wait and use the mounted dentry for the lookup. In autofs4_follow_link() the dentry is not checked to see if it is pending so it may fail to call try_to_fill_dentry() and not wait for mount completion. A dentry that is pending must always be sent to the wait queue. Signed-off-by: Ian Kent Reviewed-by: Jeff Moyer Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/root.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c index 51c873ca8e8..61d1dca1688 100644 --- a/fs/autofs4/root.c +++ b/fs/autofs4/root.c @@ -177,7 +177,8 @@ static int try_to_fill_dentry(struct dentry *dentry, int flags) return status; } /* Trigger mount for path component or follow link */ - } else if (flags & (TRIGGER_FLAGS | TRIGGER_INTENTS) || + } else if (dentry->d_flags & DCACHE_AUTOFS_PENDING || + flags & (TRIGGER_FLAGS | TRIGGER_INTENTS) || current->link_count) { DPRINTK("waiting for mount name=%.*s", dentry->d_name.len, dentry->d_name.name); @@ -223,7 +224,8 @@ static void *autofs4_follow_link(struct dentry *dentry, struct nameidata *nd) /* If it's our master or we shouldn't trigger a mount we're done */ lookup_type = nd->flags & (TRIGGER_FLAGS | TRIGGER_INTENTS); - if (oz_mode || !lookup_type) + if (oz_mode || + !(lookup_type || dentry->d_flags & DCACHE_AUTOFS_PENDING)) goto done; /* If an expire request is pending wait for it. */ @@ -242,7 +244,8 @@ static void *autofs4_follow_link(struct dentry *dentry, struct nameidata *nd) * don't try to mount it again. */ spin_lock(&dcache_lock); - if (!d_mountpoint(dentry) && __simple_empty(dentry)) { + if (dentry->d_flags & DCACHE_AUTOFS_PENDING || + (!d_mountpoint(dentry) && __simple_empty(dentry))) { spin_unlock(&dcache_lock); status = try_to_fill_dentry(dentry, 0); -- cgit v1.2.3-70-g09d2 From 97e7449a7ad883bf9f516fc970778d75999c7843 Mon Sep 17 00:00:00 2001 From: Ian Kent Date: Wed, 23 Jul 2008 21:30:26 -0700 Subject: autofs4: fix indirect mount pending expire race The selection of a dentry for expiration and the setting of the AUTOFS_INF_EXPIRING flag isn't done atomically which can lead to lookups walking into an expiring mount. What happens is that an expire is initiated by the daemon and a dentry is selected for expire but, since there is no lock held between the selection and setting of the expiring flag, a process may find the flag clear and continue walking into the mount tree at the same time the daemon attempts the expire it. Signed-off-by: Ian Kent Reviewed-by: Jeff Moyer Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/autofs_i.h | 10 +++------- fs/autofs4/expire.c | 46 +++++++++++++++++++++++++++++++++++----------- fs/autofs4/root.c | 32 +++++++++++++++++++++++++++----- 3 files changed, 65 insertions(+), 23 deletions(-) (limited to 'fs') diff --git a/fs/autofs4/autofs_i.h b/fs/autofs4/autofs_i.h index 058e1800cae..5d90ed3b4b4 100644 --- a/fs/autofs4/autofs_i.h +++ b/fs/autofs4/autofs_i.h @@ -138,18 +138,14 @@ static inline int autofs4_oz_mode(struct autofs_sb_info *sbi) { static inline int autofs4_ispending(struct dentry *dentry) { struct autofs_info *inf = autofs4_dentry_ino(dentry); - int pending = 0; if (dentry->d_flags & DCACHE_AUTOFS_PENDING) return 1; - if (inf) { - spin_lock(&inf->sbi->fs_lock); - pending = inf->flags & AUTOFS_INF_EXPIRING; - spin_unlock(&inf->sbi->fs_lock); - } + if (inf->flags & AUTOFS_INF_EXPIRING) + return 1; - return pending; + return 0; } static inline void autofs4_copy_atime(struct file *src, struct file *dst) diff --git a/fs/autofs4/expire.c b/fs/autofs4/expire.c index 894fee54d4d..19f5bea2704 100644 --- a/fs/autofs4/expire.c +++ b/fs/autofs4/expire.c @@ -292,6 +292,8 @@ static struct dentry *autofs4_expire_indirect(struct super_block *sb, struct list_head *next; int do_now = how & AUTOFS_EXP_IMMEDIATE; int exp_leaves = how & AUTOFS_EXP_LEAVES; + struct autofs_info *ino; + unsigned int ino_count; if (!root) return NULL; @@ -316,6 +318,9 @@ static struct dentry *autofs4_expire_indirect(struct super_block *sb, dentry = dget(dentry); spin_unlock(&dcache_lock); + spin_lock(&sbi->fs_lock); + ino = autofs4_dentry_ino(dentry); + /* * Case 1: (i) indirect mount or top level pseudo direct mount * (autofs-4.1). @@ -326,6 +331,11 @@ static struct dentry *autofs4_expire_indirect(struct super_block *sb, DPRINTK("checking mountpoint %p %.*s", dentry, (int)dentry->d_name.len, dentry->d_name.name); + /* Path walk currently on this dentry? */ + ino_count = atomic_read(&ino->count) + 2; + if (atomic_read(&dentry->d_count) > ino_count) + goto next; + /* Can we umount this guy */ if (autofs4_mount_busy(mnt, dentry)) goto next; @@ -343,23 +353,25 @@ static struct dentry *autofs4_expire_indirect(struct super_block *sb, /* Case 2: tree mount, expire iff entire tree is not busy */ if (!exp_leaves) { - /* Lock the tree as we must expire as a whole */ - spin_lock(&sbi->fs_lock); - if (!autofs4_tree_busy(mnt, dentry, timeout, do_now)) { - struct autofs_info *inf = autofs4_dentry_ino(dentry); + /* Path walk currently on this dentry? */ + ino_count = atomic_read(&ino->count) + 1; + if (atomic_read(&dentry->d_count) > ino_count) + goto next; - /* Set this flag early to catch sys_chdir and the like */ - inf->flags |= AUTOFS_INF_EXPIRING; - spin_unlock(&sbi->fs_lock); + if (!autofs4_tree_busy(mnt, dentry, timeout, do_now)) { expired = dentry; goto found; } - spin_unlock(&sbi->fs_lock); /* * Case 3: pseudo direct mount, expire individual leaves * (autofs-4.1). */ } else { + /* Path walk currently on this dentry? */ + ino_count = atomic_read(&ino->count) + 1; + if (atomic_read(&dentry->d_count) > ino_count) + goto next; + expired = autofs4_check_leaves(mnt, dentry, timeout, do_now); if (expired) { dput(dentry); @@ -367,6 +379,7 @@ static struct dentry *autofs4_expire_indirect(struct super_block *sb, } } next: + spin_unlock(&sbi->fs_lock); dput(dentry); spin_lock(&dcache_lock); next = next->next; @@ -377,6 +390,9 @@ next: found: DPRINTK("returning %p %.*s", expired, (int)expired->d_name.len, expired->d_name.name); + ino = autofs4_dentry_ino(expired); + ino->flags |= AUTOFS_INF_EXPIRING; + spin_unlock(&sbi->fs_lock); spin_lock(&dcache_lock); list_move(&expired->d_parent->d_subdirs, &expired->d_u.d_child); spin_unlock(&dcache_lock); @@ -390,7 +406,9 @@ int autofs4_expire_run(struct super_block *sb, struct autofs_packet_expire __user *pkt_p) { struct autofs_packet_expire pkt; + struct autofs_info *ino; struct dentry *dentry; + int ret = 0; memset(&pkt,0,sizeof pkt); @@ -406,9 +424,14 @@ int autofs4_expire_run(struct super_block *sb, dput(dentry); if ( copy_to_user(pkt_p, &pkt, sizeof(struct autofs_packet_expire)) ) - return -EFAULT; + ret = -EFAULT; - return 0; + spin_lock(&sbi->fs_lock); + ino = autofs4_dentry_ino(dentry); + ino->flags &= ~AUTOFS_INF_EXPIRING; + spin_unlock(&sbi->fs_lock); + + return ret; } /* Call repeatedly until it returns -EAGAIN, meaning there's nothing @@ -433,9 +456,10 @@ int autofs4_expire_multi(struct super_block *sb, struct vfsmount *mnt, /* This is synchronous because it makes the daemon a little easier */ - ino->flags |= AUTOFS_INF_EXPIRING; ret = autofs4_wait(sbi, dentry, NFY_EXPIRE); + spin_lock(&sbi->fs_lock); ino->flags &= ~AUTOFS_INF_EXPIRING; + spin_unlock(&sbi->fs_lock); dput(dentry); } diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c index 61d1dca1688..1c2579de1f2 100644 --- a/fs/autofs4/root.c +++ b/fs/autofs4/root.c @@ -133,7 +133,10 @@ static int try_to_fill_dentry(struct dentry *dentry, int flags) /* Block on any pending expiry here; invalidate the dentry when expiration is done to trigger mount request with a new dentry */ - if (ino && (ino->flags & AUTOFS_INF_EXPIRING)) { + spin_lock(&sbi->fs_lock); + if (ino->flags & AUTOFS_INF_EXPIRING) { + spin_unlock(&sbi->fs_lock); + DPRINTK("waiting for expire %p name=%.*s", dentry, dentry->d_name.len, dentry->d_name.name); @@ -149,8 +152,11 @@ static int try_to_fill_dentry(struct dentry *dentry, int flags) status = d_invalidate(dentry); if (status != -EBUSY) return -EAGAIN; - } + goto cont; + } + spin_unlock(&sbi->fs_lock); +cont: DPRINTK("dentry=%p %.*s ino=%p", dentry, dentry->d_name.len, dentry->d_name.name, dentry->d_inode); @@ -229,15 +235,21 @@ static void *autofs4_follow_link(struct dentry *dentry, struct nameidata *nd) goto done; /* If an expire request is pending wait for it. */ - if (ino && (ino->flags & AUTOFS_INF_EXPIRING)) { + spin_lock(&sbi->fs_lock); + if (ino->flags & AUTOFS_INF_EXPIRING) { + spin_unlock(&sbi->fs_lock); + DPRINTK("waiting for active request %p name=%.*s", dentry, dentry->d_name.len, dentry->d_name.name); status = autofs4_wait(sbi, dentry, NFY_NONE); DPRINTK("request done status=%d", status); - } + goto cont; + } + spin_unlock(&sbi->fs_lock); +cont: /* * If the dentry contains directories then it is an * autofs multi-mount with no root mount offset. So @@ -292,8 +304,11 @@ static int autofs4_revalidate(struct dentry *dentry, struct nameidata *nd) int status = 1; /* Pending dentry */ + spin_lock(&sbi->fs_lock); if (autofs4_ispending(dentry)) { /* The daemon never causes a mount to trigger */ + spin_unlock(&sbi->fs_lock); + if (oz_mode) return 1; @@ -316,6 +331,7 @@ static int autofs4_revalidate(struct dentry *dentry, struct nameidata *nd) return status; } + spin_unlock(&sbi->fs_lock); /* Negative dentry.. invalidate if "old" */ if (dentry->d_inode == NULL) @@ -329,6 +345,7 @@ static int autofs4_revalidate(struct dentry *dentry, struct nameidata *nd) DPRINTK("dentry=%p %.*s, emptydir", dentry, dentry->d_name.len, dentry->d_name.name); spin_unlock(&dcache_lock); + /* The daemon never causes a mount to trigger */ if (oz_mode) return 1; @@ -521,13 +538,18 @@ static struct dentry *autofs4_lookup(struct inode *dir, struct dentry *dentry, s * so it must have been successful, so just wait for it. */ ino = autofs4_dentry_ino(expiring); - while (ino && (ino->flags & AUTOFS_INF_EXPIRING)) { + spin_lock(&sbi->fs_lock); + if (ino->flags & AUTOFS_INF_EXPIRING) { + spin_unlock(&sbi->fs_lock); DPRINTK("wait for incomplete expire %p name=%.*s", expiring, expiring->d_name.len, expiring->d_name.name); autofs4_wait(sbi, expiring, NFY_NONE); DPRINTK("request completed"); + goto cont; } + spin_unlock(&sbi->fs_lock); +cont: spin_lock(&sbi->lookup_lock); if (!list_empty(&ino->expiring)) list_del_init(&ino->expiring); -- cgit v1.2.3-70-g09d2 From 6e60a9ab5f5d314735467752f623072f5b75157a Mon Sep 17 00:00:00 2001 From: Ian Kent Date: Wed, 23 Jul 2008 21:30:27 -0700 Subject: autofs4: fix direct mount pending expire race For direct and offset type mounts that are covered by another mount we cannot check the AUTOFS_INF_EXPIRING flag during a path walk which leads to lookups walking into an expiring mount while it is being expired. For example, for the direct multi-mount map entry with a couple of offsets: /race/mm1 / :/ /om1 :/ /om2 :/ an autofs trigger mount is mounted on /race/mm1 and when accessed it is over mounted and trigger mounts made for /race/mm1/om1 and /race/mm1/om2. So it isn't possible for path walks to see the expiring flag at all and they happily walk into the file system while it is expiring. When expiring these mounts follow_down() must stop at the autofs mount and all processes must block in the ->follow_link() method (except the daemon) until the expire is complete. This is done by decrementing the d_mounted field of the autofs trigger mount root dentry until the expire is completed. In ->follow_link() all processes wait on the expire and the mount following is completed for the daemon until the expire is complete. Signed-off-by: Ian Kent Cc: Jeff Moyer Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/autofs_i.h | 3 +++ fs/autofs4/expire.c | 16 +++++++++--- fs/autofs4/root.c | 72 +++++++++++++++++++++++++++++++++++---------------- 3 files changed, 65 insertions(+), 26 deletions(-) (limited to 'fs') diff --git a/fs/autofs4/autofs_i.h b/fs/autofs4/autofs_i.h index 5d90ed3b4b4..4b40cbc71e9 100644 --- a/fs/autofs4/autofs_i.h +++ b/fs/autofs4/autofs_i.h @@ -52,6 +52,8 @@ struct autofs_info { int flags; + struct completion expire_complete; + struct list_head active; struct list_head expiring; @@ -69,6 +71,7 @@ struct autofs_info { }; #define AUTOFS_INF_EXPIRING (1<<0) /* dentry is in the process of expiring */ +#define AUTOFS_INF_MOUNTPOINT (1<<1) /* mountpoint status for direct expire */ struct autofs_wait_queue { wait_queue_head_t queue; diff --git a/fs/autofs4/expire.c b/fs/autofs4/expire.c index 19f5bea2704..705b9f057fb 100644 --- a/fs/autofs4/expire.c +++ b/fs/autofs4/expire.c @@ -259,13 +259,15 @@ static struct dentry *autofs4_expire_direct(struct super_block *sb, now = jiffies; timeout = sbi->exp_timeout; - /* Lock the tree as we must expire as a whole */ spin_lock(&sbi->fs_lock); if (!autofs4_direct_busy(mnt, root, timeout, do_now)) { struct autofs_info *ino = autofs4_dentry_ino(root); - - /* Set this flag early to catch sys_chdir and the like */ + if (d_mountpoint(root)) { + ino->flags |= AUTOFS_INF_MOUNTPOINT; + root->d_mounted--; + } ino->flags |= AUTOFS_INF_EXPIRING; + init_completion(&ino->expire_complete); spin_unlock(&sbi->fs_lock); return root; } @@ -392,6 +394,7 @@ found: expired, (int)expired->d_name.len, expired->d_name.name); ino = autofs4_dentry_ino(expired); ino->flags |= AUTOFS_INF_EXPIRING; + init_completion(&ino->expire_complete); spin_unlock(&sbi->fs_lock); spin_lock(&dcache_lock); list_move(&expired->d_parent->d_subdirs, &expired->d_u.d_child); @@ -429,6 +432,7 @@ int autofs4_expire_run(struct super_block *sb, spin_lock(&sbi->fs_lock); ino = autofs4_dentry_ino(dentry); ino->flags &= ~AUTOFS_INF_EXPIRING; + complete_all(&ino->expire_complete); spin_unlock(&sbi->fs_lock); return ret; @@ -457,8 +461,14 @@ int autofs4_expire_multi(struct super_block *sb, struct vfsmount *mnt, /* This is synchronous because it makes the daemon a little easier */ ret = autofs4_wait(sbi, dentry, NFY_EXPIRE); + spin_lock(&sbi->fs_lock); + if (ino->flags & AUTOFS_INF_MOUNTPOINT) { + sb->s_root->d_mounted++; + ino->flags &= ~AUTOFS_INF_MOUNTPOINT; + } ino->flags &= ~AUTOFS_INF_EXPIRING; + complete_all(&ino->expire_complete); spin_unlock(&sbi->fs_lock); dput(dentry); } diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c index 1c2579de1f2..adbd8559e87 100644 --- a/fs/autofs4/root.c +++ b/fs/autofs4/root.c @@ -141,6 +141,7 @@ static int try_to_fill_dentry(struct dentry *dentry, int flags) dentry, dentry->d_name.len, dentry->d_name.name); status = autofs4_wait(sbi, dentry, NFY_NONE); + wait_for_completion(&ino->expire_complete); DPRINTK("expire done status=%d", status); @@ -227,14 +228,32 @@ static void *autofs4_follow_link(struct dentry *dentry, struct nameidata *nd) DPRINTK("dentry=%p %.*s oz_mode=%d nd->flags=%d", dentry, dentry->d_name.len, dentry->d_name.name, oz_mode, nd->flags); - - /* If it's our master or we shouldn't trigger a mount we're done */ - lookup_type = nd->flags & (TRIGGER_FLAGS | TRIGGER_INTENTS); - if (oz_mode || - !(lookup_type || dentry->d_flags & DCACHE_AUTOFS_PENDING)) + /* + * For an expire of a covered direct or offset mount we need + * to beeak out of follow_down() at the autofs mount trigger + * (d_mounted--), so we can see the expiring flag, and manage + * the blocking and following here until the expire is completed. + */ + if (oz_mode) { + spin_lock(&sbi->fs_lock); + if (ino->flags & AUTOFS_INF_EXPIRING) { + spin_unlock(&sbi->fs_lock); + /* Follow down to our covering mount. */ + if (!follow_down(&nd->path.mnt, &nd->path.dentry)) + goto done; + /* + * We shouldn't need to do this but we have no way + * of knowing what may have been done so try a follow + * just in case. + */ + autofs4_follow_mount(&nd->path.mnt, &nd->path.dentry); + goto done; + } + spin_unlock(&sbi->fs_lock); goto done; + } - /* If an expire request is pending wait for it. */ + /* If an expire request is pending everyone must wait. */ spin_lock(&sbi->fs_lock); if (ino->flags & AUTOFS_INF_EXPIRING) { spin_unlock(&sbi->fs_lock); @@ -243,6 +262,7 @@ static void *autofs4_follow_link(struct dentry *dentry, struct nameidata *nd) dentry, dentry->d_name.len, dentry->d_name.name); status = autofs4_wait(sbi, dentry, NFY_NONE); + wait_for_completion(&ino->expire_complete); DPRINTK("request done status=%d", status); @@ -250,10 +270,15 @@ static void *autofs4_follow_link(struct dentry *dentry, struct nameidata *nd) } spin_unlock(&sbi->fs_lock); cont: + /* We trigger a mount for almost all flags */ + lookup_type = nd->flags & (TRIGGER_FLAGS | TRIGGER_INTENTS); + if (!(lookup_type || dentry->d_flags & DCACHE_AUTOFS_PENDING)) + goto done; + /* - * If the dentry contains directories then it is an - * autofs multi-mount with no root mount offset. So - * don't try to mount it again. + * If the dentry contains directories then it is an autofs + * multi-mount with no root mount offset. So don't try to + * mount it again. */ spin_lock(&dcache_lock); if (dentry->d_flags & DCACHE_AUTOFS_PENDING || @@ -264,22 +289,22 @@ cont: if (status) goto out_error; - /* - * The mount succeeded but if there is no root mount - * it must be an autofs multi-mount with no root offset - * so we don't need to follow the mount. - */ - if (d_mountpoint(dentry)) { - if (!autofs4_follow_mount(&nd->path.mnt, - &nd->path.dentry)) { - status = -ENOENT; - goto out_error; - } - } - - goto done; + goto follow; } spin_unlock(&dcache_lock); +follow: + /* + * If there is no root mount it must be an autofs + * multi-mount with no root offset so we don't need + * to follow it. + */ + if (d_mountpoint(dentry)) { + if (!autofs4_follow_mount(&nd->path.mnt, + &nd->path.dentry)) { + status = -ENOENT; + goto out_error; + } + } done: return NULL; @@ -545,6 +570,7 @@ static struct dentry *autofs4_lookup(struct inode *dir, struct dentry *dentry, s expiring, expiring->d_name.len, expiring->d_name.name); autofs4_wait(sbi, expiring, NFY_NONE); + wait_for_completion(&ino->expire_complete); DPRINTK("request completed"); goto cont; } -- cgit v1.2.3-70-g09d2 From ec6e8c7d3f9073336ec7b2eed3fcda6f922087c3 Mon Sep 17 00:00:00 2001 From: Ian Kent Date: Wed, 23 Jul 2008 21:30:28 -0700 Subject: autofs4: fix direct mount pending expire race - correction Appologies, somehow I seem to have sent an out dated version of this patch. Here is an additional patch that brings the patch up to date. Signed-off-by: Ian Kent Cc: Jeff Moyer Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/root.c | 10 ++-------- 1 file changed, 2 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c index adbd8559e87..e062ee5a3ed 100644 --- a/fs/autofs4/root.c +++ b/fs/autofs4/root.c @@ -241,13 +241,7 @@ static void *autofs4_follow_link(struct dentry *dentry, struct nameidata *nd) /* Follow down to our covering mount. */ if (!follow_down(&nd->path.mnt, &nd->path.dentry)) goto done; - /* - * We shouldn't need to do this but we have no way - * of knowing what may have been done so try a follow - * just in case. - */ - autofs4_follow_mount(&nd->path.mnt, &nd->path.dentry); - goto done; + goto follow; } spin_unlock(&sbi->fs_lock); goto done; @@ -273,7 +267,7 @@ cont: /* We trigger a mount for almost all flags */ lookup_type = nd->flags & (TRIGGER_FLAGS | TRIGGER_INTENTS); if (!(lookup_type || dentry->d_flags & DCACHE_AUTOFS_PENDING)) - goto done; + goto follow; /* * If the dentry contains directories then it is an autofs -- cgit v1.2.3-70-g09d2 From 06a3598552dc3b2b30eb18bd53bbac2a901489d7 Mon Sep 17 00:00:00 2001 From: Ian Kent Date: Wed, 23 Jul 2008 21:30:28 -0700 Subject: autofs4: reorganize expire pending wait function calls This patch re-orgnirzes the checking for and waiting on active expires and elininates redundant checks. Signed-off-by: Ian Kent Cc: Jeff Moyer Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/autofs_i.h | 1 + fs/autofs4/expire.c | 29 ++++++++++++++++++++ fs/autofs4/root.c | 75 +++++++-------------------------------------------- 3 files changed, 40 insertions(+), 65 deletions(-) (limited to 'fs') diff --git a/fs/autofs4/autofs_i.h b/fs/autofs4/autofs_i.h index 4b40cbc71e9..69a2f5c9231 100644 --- a/fs/autofs4/autofs_i.h +++ b/fs/autofs4/autofs_i.h @@ -163,6 +163,7 @@ void autofs4_free_ino(struct autofs_info *); /* Expiration */ int is_autofs4_dentry(struct dentry *); +int autofs4_expire_wait(struct dentry *dentry); int autofs4_expire_run(struct super_block *, struct vfsmount *, struct autofs_sb_info *, struct autofs_packet_expire __user *); diff --git a/fs/autofs4/expire.c b/fs/autofs4/expire.c index 705b9f057fb..cdabb796ff0 100644 --- a/fs/autofs4/expire.c +++ b/fs/autofs4/expire.c @@ -402,6 +402,35 @@ found: return expired; } +int autofs4_expire_wait(struct dentry *dentry) +{ + struct autofs_sb_info *sbi = autofs4_sbi(dentry->d_sb); + struct autofs_info *ino = autofs4_dentry_ino(dentry); + int status; + + /* Block on any pending expire */ + spin_lock(&sbi->fs_lock); + if (ino->flags & AUTOFS_INF_EXPIRING) { + spin_unlock(&sbi->fs_lock); + + DPRINTK("waiting for expire %p name=%.*s", + dentry, dentry->d_name.len, dentry->d_name.name); + + status = autofs4_wait(sbi, dentry, NFY_NONE); + wait_for_completion(&ino->expire_complete); + + DPRINTK("expire done status=%d", status); + + if (d_unhashed(dentry)) + return -EAGAIN; + + return status; + } + spin_unlock(&sbi->fs_lock); + + return 0; +} + /* Perform an expiry operation */ int autofs4_expire_run(struct super_block *sb, struct vfsmount *mnt, diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c index e062ee5a3ed..ae22bde0bbd 100644 --- a/fs/autofs4/root.c +++ b/fs/autofs4/root.c @@ -130,34 +130,6 @@ static int try_to_fill_dentry(struct dentry *dentry, int flags) struct autofs_info *ino = autofs4_dentry_ino(dentry); int status; - /* Block on any pending expiry here; invalidate the dentry - when expiration is done to trigger mount request with a new - dentry */ - spin_lock(&sbi->fs_lock); - if (ino->flags & AUTOFS_INF_EXPIRING) { - spin_unlock(&sbi->fs_lock); - - DPRINTK("waiting for expire %p name=%.*s", - dentry, dentry->d_name.len, dentry->d_name.name); - - status = autofs4_wait(sbi, dentry, NFY_NONE); - wait_for_completion(&ino->expire_complete); - - DPRINTK("expire done status=%d", status); - - /* - * If the directory still exists the mount request must - * continue otherwise it can't be followed at the right - * time during the walk. - */ - status = d_invalidate(dentry); - if (status != -EBUSY) - return -EAGAIN; - - goto cont; - } - spin_unlock(&sbi->fs_lock); -cont: DPRINTK("dentry=%p %.*s ino=%p", dentry, dentry->d_name.len, dentry->d_name.name, dentry->d_inode); @@ -248,22 +220,8 @@ static void *autofs4_follow_link(struct dentry *dentry, struct nameidata *nd) } /* If an expire request is pending everyone must wait. */ - spin_lock(&sbi->fs_lock); - if (ino->flags & AUTOFS_INF_EXPIRING) { - spin_unlock(&sbi->fs_lock); - - DPRINTK("waiting for active request %p name=%.*s", - dentry, dentry->d_name.len, dentry->d_name.name); - - status = autofs4_wait(sbi, dentry, NFY_NONE); - wait_for_completion(&ino->expire_complete); + autofs4_expire_wait(dentry); - DPRINTK("request done status=%d", status); - - goto cont; - } - spin_unlock(&sbi->fs_lock); -cont: /* We trigger a mount for almost all flags */ lookup_type = nd->flags & (TRIGGER_FLAGS | TRIGGER_INTENTS); if (!(lookup_type || dentry->d_flags & DCACHE_AUTOFS_PENDING)) @@ -331,6 +289,14 @@ static int autofs4_revalidate(struct dentry *dentry, struct nameidata *nd) if (oz_mode) return 1; + /* + * If the directory has gone away due to an expire + * we have been called as ->d_revalidate() and so + * we need to return false and proceed to ->lookup(). + */ + if (autofs4_expire_wait(dentry) == -EAGAIN) + return 0; + /* * A zero status is success otherwise we have a * negative error code. @@ -339,15 +305,6 @@ static int autofs4_revalidate(struct dentry *dentry, struct nameidata *nd) if (status == 0) return 1; - /* - * A status of EAGAIN here means that the dentry has gone - * away while waiting for an expire to complete. If we are - * racing with expire lookup will wait for it so this must - * be a revalidate and we need to send it to lookup. - */ - if (status == -EAGAIN) - return 0; - return status; } spin_unlock(&sbi->fs_lock); @@ -557,19 +514,7 @@ static struct dentry *autofs4_lookup(struct inode *dir, struct dentry *dentry, s * so it must have been successful, so just wait for it. */ ino = autofs4_dentry_ino(expiring); - spin_lock(&sbi->fs_lock); - if (ino->flags & AUTOFS_INF_EXPIRING) { - spin_unlock(&sbi->fs_lock); - DPRINTK("wait for incomplete expire %p name=%.*s", - expiring, expiring->d_name.len, - expiring->d_name.name); - autofs4_wait(sbi, expiring, NFY_NONE); - wait_for_completion(&ino->expire_complete); - DPRINTK("request completed"); - goto cont; - } - spin_unlock(&sbi->fs_lock); -cont: + autofs4_expire_wait(expiring); spin_lock(&sbi->lookup_lock); if (!list_empty(&ino->expiring)) list_del_init(&ino->expiring); -- cgit v1.2.3-70-g09d2 From aa55ddf340c9fa3f303ee16bbf35887e42c50304 Mon Sep 17 00:00:00 2001 From: Ian Kent Date: Wed, 23 Jul 2008 21:30:29 -0700 Subject: autofs4: remove unused ioctls The ioctls AUTOFS_IOC_TOGGLEREGHOST and AUTOFS_IOC_ASKREGHOST were added several years ago but what they were intended for has never been implemented (as far as I'm aware noone uses them) so remove them. Signed-off-by: Ian Kent Reviewed-by: Jeff Moyer Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/root.c | 68 +----------------------------------------------- fs/compat_ioctl.c | 2 -- include/linux/auto_fs4.h | 2 -- 3 files changed, 1 insertion(+), 71 deletions(-) (limited to 'fs') diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c index ae22bde0bbd..bcfb2dc0a61 100644 --- a/fs/autofs4/root.c +++ b/fs/autofs4/root.c @@ -25,7 +25,6 @@ static int autofs4_dir_rmdir(struct inode *,struct dentry *); static int autofs4_dir_mkdir(struct inode *,struct dentry *,int); static int autofs4_root_ioctl(struct inode *, struct file *,unsigned int,unsigned long); static int autofs4_dir_open(struct inode *inode, struct file *file); -static int autofs4_root_readdir(struct file * filp, void * dirent, filldir_t filldir); static struct dentry *autofs4_lookup(struct inode *,struct dentry *, struct nameidata *); static void *autofs4_follow_link(struct dentry *, struct nameidata *); @@ -36,7 +35,7 @@ const struct file_operations autofs4_root_operations = { .open = dcache_dir_open, .release = dcache_dir_close, .read = generic_read_dir, - .readdir = autofs4_root_readdir, + .readdir = dcache_readdir, .ioctl = autofs4_root_ioctl, }; @@ -71,28 +70,6 @@ const struct inode_operations autofs4_dir_inode_operations = { .rmdir = autofs4_dir_rmdir, }; -static int autofs4_root_readdir(struct file *file, void *dirent, - filldir_t filldir) -{ - struct autofs_sb_info *sbi = autofs4_sbi(file->f_path.dentry->d_sb); - int oz_mode = autofs4_oz_mode(sbi); - - DPRINTK("called, filp->f_pos = %lld", file->f_pos); - - /* - * Don't set reghost flag if: - * 1) f_pos is larger than zero -- we've already been here. - * 2) we haven't even enabled reghosting in the 1st place. - * 3) this is the daemon doing a readdir - */ - if (oz_mode && file->f_pos == 0 && sbi->reghost_enabled) - sbi->needs_reghost = 1; - - DPRINTK("needs_reghost = %d", sbi->needs_reghost); - - return dcache_readdir(file, dirent, filldir); -} - static int autofs4_dir_open(struct inode *inode, struct file *file) { struct dentry *dentry = file->f_path.dentry; @@ -858,44 +835,6 @@ static inline int autofs4_get_protosubver(struct autofs_sb_info *sbi, int __user return put_user(sbi->sub_version, p); } -/* - * Tells the daemon whether we need to reghost or not. Also, clears - * the reghost_needed flag. - */ -static inline int autofs4_ask_reghost(struct autofs_sb_info *sbi, int __user *p) -{ - int status; - - DPRINTK("returning %d", sbi->needs_reghost); - - status = put_user(sbi->needs_reghost, p); - if (status) - return status; - - sbi->needs_reghost = 0; - return 0; -} - -/* - * Enable / Disable reghosting ioctl() operation - */ -static inline int autofs4_toggle_reghost(struct autofs_sb_info *sbi, int __user *p) -{ - int status; - int val; - - status = get_user(val, p); - - DPRINTK("reghost = %d", val); - - if (status) - return status; - - /* turn on/off reghosting, with the val */ - sbi->reghost_enabled = val; - return 0; -} - /* * Tells the daemon whether it can umount the autofs mount. */ @@ -960,11 +899,6 @@ static int autofs4_root_ioctl(struct inode *inode, struct file *filp, case AUTOFS_IOC_SETTIMEOUT: return autofs4_get_set_timeout(sbi, p); - case AUTOFS_IOC_TOGGLEREGHOST: - return autofs4_toggle_reghost(sbi, p); - case AUTOFS_IOC_ASKREGHOST: - return autofs4_ask_reghost(sbi, p); - case AUTOFS_IOC_ASKUMOUNT: return autofs4_ask_umount(filp->f_path.mnt, p); diff --git a/fs/compat_ioctl.c b/fs/compat_ioctl.c index 7b3a03c7c6a..18e2c548161 100644 --- a/fs/compat_ioctl.c +++ b/fs/compat_ioctl.c @@ -2297,8 +2297,6 @@ COMPATIBLE_IOCTL(AUTOFS_IOC_PROTOVER) COMPATIBLE_IOCTL(AUTOFS_IOC_EXPIRE) COMPATIBLE_IOCTL(AUTOFS_IOC_EXPIRE_MULTI) COMPATIBLE_IOCTL(AUTOFS_IOC_PROTOSUBVER) -COMPATIBLE_IOCTL(AUTOFS_IOC_ASKREGHOST) -COMPATIBLE_IOCTL(AUTOFS_IOC_TOGGLEREGHOST) COMPATIBLE_IOCTL(AUTOFS_IOC_ASKUMOUNT) /* Raw devices */ COMPATIBLE_IOCTL(RAW_SETBIND) diff --git a/include/linux/auto_fs4.h b/include/linux/auto_fs4.h index 31a29541b50..b785c6f8644 100644 --- a/include/linux/auto_fs4.h +++ b/include/linux/auto_fs4.h @@ -98,8 +98,6 @@ union autofs_v5_packet_union { #define AUTOFS_IOC_EXPIRE_INDIRECT AUTOFS_IOC_EXPIRE_MULTI #define AUTOFS_IOC_EXPIRE_DIRECT AUTOFS_IOC_EXPIRE_MULTI #define AUTOFS_IOC_PROTOSUBVER _IOR(0x93,0x67,int) -#define AUTOFS_IOC_ASKREGHOST _IOR(0x93,0x68,int) -#define AUTOFS_IOC_TOGGLEREGHOST _IOR(0x93,0x69,int) #define AUTOFS_IOC_ASKUMOUNT _IOR(0x93,0x70,int) -- cgit v1.2.3-70-g09d2 From f9247273cb69ba101877e946d2d83044409cc8c5 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Thu, 24 Jul 2008 17:22:13 +0100 Subject: UFS: add const to parser token table This patch adds a "const" to the parser token table. I've done an allmodconfig build to see if this produces any warnings/failures and the patch includes a fix for the only warning that was produced. Signed-off-by: Steven Whitehouse Acked-by: Alexander Viro Acked-by: Evgeniy Dushistov Signed-off-by: Linus Torvalds --- fs/ufs/super.c | 2 +- include/linux/parser.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ufs/super.c b/fs/ufs/super.c index 85b22b5977f..506f724055c 100644 --- a/fs/ufs/super.c +++ b/fs/ufs/super.c @@ -1232,7 +1232,7 @@ static int ufs_show_options(struct seq_file *seq, struct vfsmount *vfs) { struct ufs_sb_info *sbi = UFS_SB(vfs->mnt_sb); unsigned mval = sbi->s_mount_opt & UFS_MOUNT_UFSTYPE; - struct match_token *tp = tokens; + const struct match_token *tp = tokens; while (tp->token != Opt_onerror_panic && tp->token != mval) ++tp; diff --git a/include/linux/parser.h b/include/linux/parser.h index 7dcd0507575..cc554ca8bc7 100644 --- a/include/linux/parser.h +++ b/include/linux/parser.h @@ -14,7 +14,7 @@ struct match_token { const char *pattern; }; -typedef struct match_token match_table_t[]; +typedef const struct match_token match_table_t[]; /* Maximum number of arguments that match_token will find in a pattern */ enum {MAX_OPT_ARGS = 3}; -- cgit v1.2.3-70-g09d2 From e2d2867ff8700d7431c68c089ff5f5ed7f2d5b40 Mon Sep 17 00:00:00 2001 From: Chris Wright Date: Thu, 24 Jul 2008 20:43:34 +0000 Subject: When verifying the decoded header before decoding the object identifier (expecting a SPNEGO pseudo-mechanism oid), the test to verify it is a primitive encoding is compared against the asn1 class. Primitive is not a class. This brings check in line with similar check for krb/ntlmssp oid. Signed-off-by: Chris Wright Cc: Steven French Signed-off-by: Andrew Morton Signed-off-by: Steve French --- fs/cifs/asn1.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/cifs/asn1.c b/fs/cifs/asn1.c index 42765465701..6bb440b257b 100644 --- a/fs/cifs/asn1.c +++ b/fs/cifs/asn1.c @@ -494,7 +494,7 @@ decode_negTokenInit(unsigned char *security_blob, int length, /* remember to free obj->oid */ rc = asn1_header_decode(&ctx, &end, &cls, &con, &tag); if (rc) { - if ((tag == ASN1_OJI) && (cls == ASN1_PRI)) { + if ((tag == ASN1_OJI) && (con == ASN1_PRI)) { rc = asn1_oid_decode(&ctx, end, &oid, &oidlen); if (rc) { rc = compare_oid(oid, oidlen, -- cgit v1.2.3-70-g09d2 From fb2e405fc1fc8b20d9c78eaa1c7fd5a297efde43 Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Fri, 25 Jul 2008 02:55:49 +0300 Subject: fix fs/nfs/nfsroot.c compilation This fixes the following compile error caused by commit f9247273cb69ba101877e946d2d83044409cc8c5 ("UFS: add const to parser token table"): CC fs/nfs/nfsroot.o /home/bunk/linux/kernel-2.6/git/linux-2.6/fs/nfs/nfsroot.c:130: error: tokens causes a section type conflict make[3]: *** [fs/nfs/nfsroot.o] Error 1 Signed-off-by: Adrian Bunk Signed-off-by: Linus Torvalds --- fs/nfs/nfsroot.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/nfsroot.c b/fs/nfs/nfsroot.c index 46763d1cd39..8478fc25dae 100644 --- a/fs/nfs/nfsroot.c +++ b/fs/nfs/nfsroot.c @@ -127,7 +127,7 @@ enum { Opt_err }; -static match_table_t __initdata tokens = { +static match_table_t __initconst tokens = { {Opt_port, "port=%u"}, {Opt_rsize, "rsize=%u"}, {Opt_wsize, "wsize=%u"}, -- cgit v1.2.3-70-g09d2 From 483fad1c3fa1060d7e6710e84a065ad514571739 Mon Sep 17 00:00:00 2001 From: Nathan Lynch Date: Tue, 22 Jul 2008 04:48:46 +1000 Subject: ELF loader support for auxvec base platform string Some IBM POWER-based platforms have the ability to run in a mode which mostly appears to the OS as a different processor from the actual hardware. For example, a Power6 system may appear to be a Power5+, which makes the AT_PLATFORM value "power5+". This means that programs are restricted to the ISA supported by Power5+; Power6-specific instructions are treated as illegal. However, some applications (virtual machines, optimized libraries) can benefit from knowledge of the underlying CPU model. A new aux vector entry, AT_BASE_PLATFORM, will denote the actual hardware. For example, on a Power6 system in Power5+ compatibility mode, AT_PLATFORM will be "power5+" and AT_BASE_PLATFORM will be "power6". The idea is that AT_PLATFORM indicates the instruction set supported, while AT_BASE_PLATFORM indicates the underlying microarchitecture. If the architecture has defined ELF_BASE_PLATFORM, copy that value to the user stack in the same manner as ELF_PLATFORM. Signed-off-by: Nathan Lynch Acked-by: Andrew Morton Signed-off-by: Benjamin Herrenschmidt --- fs/binfmt_elf.c | 28 ++++++++++++++++++++++++++++ include/linux/auxvec.h | 6 +++++- 2 files changed, 33 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c index 639d2d8b571..742c8f53048 100644 --- a/fs/binfmt_elf.c +++ b/fs/binfmt_elf.c @@ -131,6 +131,15 @@ static int padzero(unsigned long elf_bss) #define STACK_ALLOC(sp, len) ({ sp -= len ; sp; }) #endif +#ifndef ELF_BASE_PLATFORM +/* + * AT_BASE_PLATFORM indicates the "real" hardware/microarchitecture. + * If the arch defines ELF_BASE_PLATFORM (in asm/elf.h), the value + * will be copied to the user stack in the same manner as AT_PLATFORM. + */ +#define ELF_BASE_PLATFORM NULL +#endif + static int create_elf_tables(struct linux_binprm *bprm, struct elfhdr *exec, unsigned long load_addr, unsigned long interp_load_addr) @@ -142,7 +151,9 @@ create_elf_tables(struct linux_binprm *bprm, struct elfhdr *exec, elf_addr_t __user *envp; elf_addr_t __user *sp; elf_addr_t __user *u_platform; + elf_addr_t __user *u_base_platform; const char *k_platform = ELF_PLATFORM; + const char *k_base_platform = ELF_BASE_PLATFORM; int items; elf_addr_t *elf_info; int ei_index = 0; @@ -172,6 +183,19 @@ create_elf_tables(struct linux_binprm *bprm, struct elfhdr *exec, return -EFAULT; } + /* + * If this architecture has a "base" platform capability + * string, copy it to userspace. + */ + u_base_platform = NULL; + if (k_base_platform) { + size_t len = strlen(k_base_platform) + 1; + + u_base_platform = (elf_addr_t __user *)STACK_ALLOC(p, len); + if (__copy_to_user(u_base_platform, k_base_platform, len)) + return -EFAULT; + } + /* Create the ELF interpreter info */ elf_info = (elf_addr_t *)current->mm->saved_auxv; /* update AT_VECTOR_SIZE_BASE if the number of NEW_AUX_ENT() changes */ @@ -209,6 +233,10 @@ create_elf_tables(struct linux_binprm *bprm, struct elfhdr *exec, NEW_AUX_ENT(AT_PLATFORM, (elf_addr_t)(unsigned long)u_platform); } + if (k_base_platform) { + NEW_AUX_ENT(AT_BASE_PLATFORM, + (elf_addr_t)(unsigned long)u_base_platform); + } if (bprm->interp_flags & BINPRM_FLAGS_EXECFD) { NEW_AUX_ENT(AT_EXECFD, bprm->interp_data); } diff --git a/include/linux/auxvec.h b/include/linux/auxvec.h index 0da17d14fd1..d7afa9dd663 100644 --- a/include/linux/auxvec.h +++ b/include/linux/auxvec.h @@ -26,9 +26,13 @@ #define AT_SECURE 23 /* secure mode boolean */ +#define AT_BASE_PLATFORM 24 /* string identifying real platform, may + * differ from AT_PLATFORM. */ + #define AT_EXECFN 31 /* filename of program */ + #ifdef __KERNEL__ -#define AT_VECTOR_SIZE_BASE 17 /* NEW_AUX_ENT entries in auxiliary table */ +#define AT_VECTOR_SIZE_BASE 18 /* NEW_AUX_ENT entries in auxiliary table */ /* number of "#define AT_.*" above, minus {AT_NULL, AT_IGNORE, AT_NOTELF} */ #endif -- cgit v1.2.3-70-g09d2 From b7bbf8fa6ba329b3552b75a0716f5fbc6f839499 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Fri, 25 Jul 2008 01:45:25 -0700 Subject: fs: ldm.[ch] use get_unaligned_* helpers Replace the private BE16/BE32/BE64 macros with direct calls to get_unaligned_be16/32/64. Signed-off-by: Harvey Harrison Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/partitions/ldm.c | 70 ++++++++++++++++++++++++++--------------------------- fs/partitions/ldm.h | 5 ---- 2 files changed, 35 insertions(+), 40 deletions(-) (limited to 'fs') diff --git a/fs/partitions/ldm.c b/fs/partitions/ldm.c index 0fdda2e8a4c..8652fb99e96 100644 --- a/fs/partitions/ldm.c +++ b/fs/partitions/ldm.c @@ -133,17 +133,17 @@ static bool ldm_parse_privhead(const u8 *data, struct privhead *ph) bool is_vista = false; BUG_ON(!data || !ph); - if (MAGIC_PRIVHEAD != BE64(data)) { + if (MAGIC_PRIVHEAD != get_unaligned_be64(data)) { ldm_error("Cannot find PRIVHEAD structure. LDM database is" " corrupt. Aborting."); return false; } - ph->ver_major = BE16(data + 0x000C); - ph->ver_minor = BE16(data + 0x000E); - ph->logical_disk_start = BE64(data + 0x011B); - ph->logical_disk_size = BE64(data + 0x0123); - ph->config_start = BE64(data + 0x012B); - ph->config_size = BE64(data + 0x0133); + ph->ver_major = get_unaligned_be16(data + 0x000C); + ph->ver_minor = get_unaligned_be16(data + 0x000E); + ph->logical_disk_start = get_unaligned_be64(data + 0x011B); + ph->logical_disk_size = get_unaligned_be64(data + 0x0123); + ph->config_start = get_unaligned_be64(data + 0x012B); + ph->config_size = get_unaligned_be64(data + 0x0133); /* Version 2.11 is Win2k/XP and version 2.12 is Vista. */ if (ph->ver_major == 2 && ph->ver_minor == 12) is_vista = true; @@ -191,14 +191,14 @@ static bool ldm_parse_tocblock (const u8 *data, struct tocblock *toc) { BUG_ON (!data || !toc); - if (MAGIC_TOCBLOCK != BE64 (data)) { + if (MAGIC_TOCBLOCK != get_unaligned_be64(data)) { ldm_crit ("Cannot find TOCBLOCK, database may be corrupt."); return false; } strncpy (toc->bitmap1_name, data + 0x24, sizeof (toc->bitmap1_name)); toc->bitmap1_name[sizeof (toc->bitmap1_name) - 1] = 0; - toc->bitmap1_start = BE64 (data + 0x2E); - toc->bitmap1_size = BE64 (data + 0x36); + toc->bitmap1_start = get_unaligned_be64(data + 0x2E); + toc->bitmap1_size = get_unaligned_be64(data + 0x36); if (strncmp (toc->bitmap1_name, TOC_BITMAP1, sizeof (toc->bitmap1_name)) != 0) { @@ -208,8 +208,8 @@ static bool ldm_parse_tocblock (const u8 *data, struct tocblock *toc) } strncpy (toc->bitmap2_name, data + 0x46, sizeof (toc->bitmap2_name)); toc->bitmap2_name[sizeof (toc->bitmap2_name) - 1] = 0; - toc->bitmap2_start = BE64 (data + 0x50); - toc->bitmap2_size = BE64 (data + 0x58); + toc->bitmap2_start = get_unaligned_be64(data + 0x50); + toc->bitmap2_size = get_unaligned_be64(data + 0x58); if (strncmp (toc->bitmap2_name, TOC_BITMAP2, sizeof (toc->bitmap2_name)) != 0) { ldm_crit ("TOCBLOCK's second bitmap is '%s', should be '%s'.", @@ -237,22 +237,22 @@ static bool ldm_parse_vmdb (const u8 *data, struct vmdb *vm) { BUG_ON (!data || !vm); - if (MAGIC_VMDB != BE32 (data)) { + if (MAGIC_VMDB != get_unaligned_be32(data)) { ldm_crit ("Cannot find the VMDB, database may be corrupt."); return false; } - vm->ver_major = BE16 (data + 0x12); - vm->ver_minor = BE16 (data + 0x14); + vm->ver_major = get_unaligned_be16(data + 0x12); + vm->ver_minor = get_unaligned_be16(data + 0x14); if ((vm->ver_major != 4) || (vm->ver_minor != 10)) { ldm_error ("Expected VMDB version %d.%d, got %d.%d. " "Aborting.", 4, 10, vm->ver_major, vm->ver_minor); return false; } - vm->vblk_size = BE32 (data + 0x08); - vm->vblk_offset = BE32 (data + 0x0C); - vm->last_vblk_seq = BE32 (data + 0x04); + vm->vblk_size = get_unaligned_be32(data + 0x08); + vm->vblk_offset = get_unaligned_be32(data + 0x0C); + vm->last_vblk_seq = get_unaligned_be32(data + 0x04); ldm_debug ("Parsed VMDB successfully."); return true; @@ -507,7 +507,7 @@ static bool ldm_validate_vmdb (struct block_device *bdev, unsigned long base, goto out; /* Already logged */ /* Are there uncommitted transactions? */ - if (BE16(data + 0x10) != 0x01) { + if (get_unaligned_be16(data + 0x10) != 0x01) { ldm_crit ("Database is not in a consistent state. Aborting."); goto out; } @@ -802,7 +802,7 @@ static bool ldm_parse_cmp3 (const u8 *buffer, int buflen, struct vblk *vb) return false; len += VBLK_SIZE_CMP3; - if (len != BE32 (buffer + 0x14)) + if (len != get_unaligned_be32(buffer + 0x14)) return false; comp = &vb->vblk.comp; @@ -851,7 +851,7 @@ static int ldm_parse_dgr3 (const u8 *buffer, int buflen, struct vblk *vb) return false; len += VBLK_SIZE_DGR3; - if (len != BE32 (buffer + 0x14)) + if (len != get_unaligned_be32(buffer + 0x14)) return false; dgrp = &vb->vblk.dgrp; @@ -895,7 +895,7 @@ static bool ldm_parse_dgr4 (const u8 *buffer, int buflen, struct vblk *vb) return false; len += VBLK_SIZE_DGR4; - if (len != BE32 (buffer + 0x14)) + if (len != get_unaligned_be32(buffer + 0x14)) return false; dgrp = &vb->vblk.dgrp; @@ -931,7 +931,7 @@ static bool ldm_parse_dsk3 (const u8 *buffer, int buflen, struct vblk *vb) return false; len += VBLK_SIZE_DSK3; - if (len != BE32 (buffer + 0x14)) + if (len != get_unaligned_be32(buffer + 0x14)) return false; disk = &vb->vblk.disk; @@ -968,7 +968,7 @@ static bool ldm_parse_dsk4 (const u8 *buffer, int buflen, struct vblk *vb) return false; len += VBLK_SIZE_DSK4; - if (len != BE32 (buffer + 0x14)) + if (len != get_unaligned_be32(buffer + 0x14)) return false; disk = &vb->vblk.disk; @@ -1034,14 +1034,14 @@ static bool ldm_parse_prt3(const u8 *buffer, int buflen, struct vblk *vb) return false; } len += VBLK_SIZE_PRT3; - if (len > BE32(buffer + 0x14)) { + if (len > get_unaligned_be32(buffer + 0x14)) { ldm_error("len %d > BE32(buffer + 0x14) %d", len, - BE32(buffer + 0x14)); + get_unaligned_be32(buffer + 0x14)); return false; } part = &vb->vblk.part; - part->start = BE64(buffer + 0x24 + r_name); - part->volume_offset = BE64(buffer + 0x2C + r_name); + part->start = get_unaligned_be64(buffer + 0x24 + r_name); + part->volume_offset = get_unaligned_be64(buffer + 0x2C + r_name); part->size = ldm_get_vnum(buffer + 0x34 + r_name); part->parent_id = ldm_get_vnum(buffer + 0x34 + r_size); part->disk_id = ldm_get_vnum(buffer + 0x34 + r_parent); @@ -1139,9 +1139,9 @@ static bool ldm_parse_vol5(const u8 *buffer, int buflen, struct vblk *vb) return false; } len += VBLK_SIZE_VOL5; - if (len > BE32(buffer + 0x14)) { + if (len > get_unaligned_be32(buffer + 0x14)) { ldm_error("len %d > BE32(buffer + 0x14) %d", len, - BE32(buffer + 0x14)); + get_unaligned_be32(buffer + 0x14)); return false; } volu = &vb->vblk.volu; @@ -1294,9 +1294,9 @@ static bool ldm_frag_add (const u8 *data, int size, struct list_head *frags) BUG_ON (!data || !frags); - group = BE32 (data + 0x08); - rec = BE16 (data + 0x0C); - num = BE16 (data + 0x0E); + group = get_unaligned_be32(data + 0x08); + rec = get_unaligned_be16(data + 0x0C); + num = get_unaligned_be16(data + 0x0E); if ((num < 1) || (num > 4)) { ldm_error ("A VBLK claims to have %d parts.", num); return false; @@ -1425,12 +1425,12 @@ static bool ldm_get_vblks (struct block_device *bdev, unsigned long base, } for (v = 0; v < perbuf; v++, data+=size) { /* For each vblk */ - if (MAGIC_VBLK != BE32 (data)) { + if (MAGIC_VBLK != get_unaligned_be32(data)) { ldm_error ("Expected to find a VBLK."); goto out; } - recs = BE16 (data + 0x0E); /* Number of records */ + recs = get_unaligned_be16(data + 0x0E); /* Number of records */ if (recs == 1) { if (!ldm_ldmdb_add (data, size, ldb)) goto out; /* Already logged */ diff --git a/fs/partitions/ldm.h b/fs/partitions/ldm.h index 80f63b5fdd9..30e08e809c1 100644 --- a/fs/partitions/ldm.h +++ b/fs/partitions/ldm.h @@ -98,11 +98,6 @@ struct parsed_partitions; #define TOC_BITMAP1 "config" /* Names of the two defined */ #define TOC_BITMAP2 "log" /* bitmaps in the TOCBLOCK. */ -/* Most numbers we deal with are big-endian and won't be aligned. */ -#define BE16(x) ((u16)be16_to_cpu(get_unaligned((__be16*)(x)))) -#define BE32(x) ((u32)be32_to_cpu(get_unaligned((__be32*)(x)))) -#define BE64(x) ((u64)be64_to_cpu(get_unaligned((__be64*)(x)))) - /* Borrowed from msdos.c */ #define SYS_IND(p) (get_unaligned(&(p)->sys_ind)) -- cgit v1.2.3-70-g09d2 From ba92a43dbaee339cf5915ef766d3d3ffbaaf103c Mon Sep 17 00:00:00 2001 From: Hugh Dickins Date: Fri, 25 Jul 2008 01:45:43 -0700 Subject: exec: remove some includes fs/exec.c used to need mman.h pagemap.h swap.h and rmap.h when it did mm-ish stuff in install_arg_page(); but no need for them after 2.6.22. [akpm@linux-foundation.org: unbreak arm] Signed-off-by: Hugh Dickins Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/exec.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/exec.c b/fs/exec.c index 190ed1f9277..e41aef0fb35 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -25,19 +25,18 @@ #include #include #include -#include +#include #include #include #include +#include #include #include -#include #include #include #include #include #include -#include #include #include #include @@ -47,7 +46,6 @@ #include #include #include -#include #include #include #include -- cgit v1.2.3-70-g09d2 From fb523f32275344282f20ef3352cbf03e599241e6 Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Fri, 25 Jul 2008 01:46:14 -0700 Subject: minix: remove !NO_TRUNCATE code This patch removes the !NO_TRUNCATE code that anyway required a manual editing of the code for being used. Signed-off-by: Adrian Bunk Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/minix/inode.c | 3 --- fs/minix/minix.h | 6 ------ fs/minix/namei.c | 24 ------------------------ 3 files changed, 33 deletions(-) (limited to 'fs') diff --git a/fs/minix/inode.c b/fs/minix/inode.c index 84f6242ba6f..523d7371341 100644 --- a/fs/minix/inode.c +++ b/fs/minix/inode.c @@ -256,9 +256,6 @@ static int minix_fill_super(struct super_block *s, void *data, int silent) if (!s->s_root) goto out_iput; - if (!NO_TRUNCATE) - s->s_root->d_op = &minix_dentry_operations; - if (!(s->s_flags & MS_RDONLY)) { if (sbi->s_version != MINIX_V3) /* s_state is now out from V3 sb */ ms->s_state &= ~MINIX_VALID_FS; diff --git a/fs/minix/minix.h b/fs/minix/minix.h index 326edfe9610..e6a0b193bea 100644 --- a/fs/minix/minix.h +++ b/fs/minix/minix.h @@ -2,11 +2,6 @@ #include #include -/* - * change the define below to 0 if you want names > info->s_namelen chars to be - * truncated. Else they will be disallowed (ENAMETOOLONG). - */ -#define NO_TRUNCATE 1 #define INODE_VERSION(inode) minix_sb(inode->i_sb)->s_version #define MINIX_V1 0x0001 /* original minix fs */ #define MINIX_V2 0x0002 /* minix V2 fs */ @@ -83,7 +78,6 @@ extern const struct inode_operations minix_file_inode_operations; extern const struct inode_operations minix_dir_inode_operations; extern const struct file_operations minix_file_operations; extern const struct file_operations minix_dir_operations; -extern struct dentry_operations minix_dentry_operations; static inline struct minix_sb_info *minix_sb(struct super_block *sb) { diff --git a/fs/minix/namei.c b/fs/minix/namei.c index 102241bc9c7..32b131cd612 100644 --- a/fs/minix/namei.c +++ b/fs/minix/namei.c @@ -18,30 +18,6 @@ static int add_nondir(struct dentry *dentry, struct inode *inode) return err; } -static int minix_hash(struct dentry *dentry, struct qstr *qstr) -{ - unsigned long hash; - int i; - const unsigned char *name; - - i = minix_sb(dentry->d_inode->i_sb)->s_namelen; - if (i >= qstr->len) - return 0; - /* Truncate the name in place, avoids having to define a compare - function. */ - qstr->len = i; - name = qstr->name; - hash = init_name_hash(); - while (i--) - hash = partial_name_hash(*name++, hash); - qstr->hash = end_name_hash(hash); - return 0; -} - -struct dentry_operations minix_dentry_operations = { - .d_hash = minix_hash, -}; - static struct dentry *minix_lookup(struct inode * dir, struct dentry *dentry, struct nameidata *nd) { struct inode * inode = NULL; -- cgit v1.2.3-70-g09d2 From f905f06fca5d3949eca12f5a43e251a404b3470a Mon Sep 17 00:00:00 2001 From: Shen Feng Date: Fri, 25 Jul 2008 01:46:15 -0700 Subject: ext2: remove double definitions of xattr macros remove the definitions of macros: XATTR_TRUSTED_PREFIX XATTR_USER_PREFIX since they are defined in linux/xattr.h Signed-off-by: Shen Feng Signed-off-by: Mingming Cao Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext2/xattr_security.c | 2 +- fs/ext2/xattr_trusted.c | 4 +--- fs/ext2/xattr_user.c | 4 +--- 3 files changed, 3 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/ext2/xattr_security.c b/fs/ext2/xattr_security.c index eaa23d2d521..70c0dbdcdcb 100644 --- a/fs/ext2/xattr_security.c +++ b/fs/ext2/xattr_security.c @@ -14,7 +14,7 @@ static size_t ext2_xattr_security_list(struct inode *inode, char *list, size_t list_size, const char *name, size_t name_len) { - const int prefix_len = sizeof(XATTR_SECURITY_PREFIX)-1; + const int prefix_len = XATTR_SECURITY_PREFIX_LEN; const size_t total_len = prefix_len + name_len + 1; if (list && total_len <= list_size) { diff --git a/fs/ext2/xattr_trusted.c b/fs/ext2/xattr_trusted.c index 83ee149f353..e8219f8eae9 100644 --- a/fs/ext2/xattr_trusted.c +++ b/fs/ext2/xattr_trusted.c @@ -12,13 +12,11 @@ #include #include "xattr.h" -#define XATTR_TRUSTED_PREFIX "trusted." - static size_t ext2_xattr_trusted_list(struct inode *inode, char *list, size_t list_size, const char *name, size_t name_len) { - const int prefix_len = sizeof(XATTR_TRUSTED_PREFIX)-1; + const int prefix_len = XATTR_TRUSTED_PREFIX_LEN; const size_t total_len = prefix_len + name_len + 1; if (!capable(CAP_SYS_ADMIN)) diff --git a/fs/ext2/xattr_user.c b/fs/ext2/xattr_user.c index f383e7c3a7b..92495d28c62 100644 --- a/fs/ext2/xattr_user.c +++ b/fs/ext2/xattr_user.c @@ -11,13 +11,11 @@ #include "ext2.h" #include "xattr.h" -#define XATTR_USER_PREFIX "user." - static size_t ext2_xattr_user_list(struct inode *inode, char *list, size_t list_size, const char *name, size_t name_len) { - const size_t prefix_len = sizeof(XATTR_USER_PREFIX)-1; + const size_t prefix_len = XATTR_USER_PREFIX_LEN; const size_t total_len = prefix_len + name_len + 1; if (!test_opt(inode->i_sb, XATTR_USER)) -- cgit v1.2.3-70-g09d2 From 9cfe7b9010aa66da5f3b2bc33d9e30a4d53bd274 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Fri, 25 Jul 2008 01:46:16 -0700 Subject: ext3: fix synchronization of quota files in journal=data mode In journal=data mode, it is not enough to do write_inode_now as done in vfs_quota_on() to write all data to their final location (which is needed for quota_read to work correctly). Calling journal_flush() does its job. Reported-by: Nick Cc: Signed-off-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/super.c | 35 +++++++++++++++++++++++++++-------- 1 file changed, 27 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/ext3/super.c b/fs/ext3/super.c index 2845425077e..50796e90d07 100644 --- a/fs/ext3/super.c +++ b/fs/ext3/super.c @@ -2759,23 +2759,42 @@ static int ext3_quota_on(struct super_block *sb, int type, int format_id, if (!test_opt(sb, QUOTA)) return -EINVAL; - /* Not journalling quota or remount? */ - if ((!EXT3_SB(sb)->s_qf_names[USRQUOTA] && - !EXT3_SB(sb)->s_qf_names[GRPQUOTA]) || remount) + /* When remounting, no checks are needed and in fact, path is NULL */ + if (remount) return vfs_quota_on(sb, type, format_id, path, remount); + err = path_lookup(path, LOOKUP_FOLLOW, &nd); if (err) return err; + /* Quotafile not on the same filesystem? */ if (nd.path.mnt->mnt_sb != sb) { path_put(&nd.path); return -EXDEV; } - /* Quotafile not in fs root? */ - if (nd.path.dentry->d_parent->d_inode != sb->s_root->d_inode) - printk(KERN_WARNING - "EXT3-fs: Quota file not on filesystem root. " - "Journalled quota will not work.\n"); + /* Journaling quota? */ + if (EXT3_SB(sb)->s_qf_names[type]) { + /* Quotafile not of fs root? */ + if (nd.path.dentry->d_parent->d_inode != sb->s_root->d_inode) + printk(KERN_WARNING + "EXT3-fs: Quota file not on filesystem root. " + "Journaled quota will not work.\n"); + } + + /* + * When we journal data on quota file, we have to flush journal to see + * all updates to the file when we bypass pagecache... + */ + if (ext3_should_journal_data(nd.path.dentry->d_inode)) { + /* + * We don't need to lock updates but journal_flush() could + * otherwise be livelocked... + */ + journal_lock_updates(EXT3_SB(sb)->s_journal); + journal_flush(EXT3_SB(sb)->s_journal); + journal_unlock_updates(EXT3_SB(sb)->s_journal); + } + path_put(&nd.path); return vfs_quota_on(sb, type, format_id, path, remount); } -- cgit v1.2.3-70-g09d2 From 99aeaf639f61ab6be1967e5f92e2e28dafad8383 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Fri, 25 Jul 2008 01:46:17 -0700 Subject: ext3: fix typos in messages and comments (journalled -> journaled) Cc: Signed-off-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/super.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/ext3/super.c b/fs/ext3/super.c index 50796e90d07..0a1bf82845c 100644 --- a/fs/ext3/super.c +++ b/fs/ext3/super.c @@ -1020,7 +1020,7 @@ static int parse_options (char *options, struct super_block *sb, set_qf_name: if (sb_any_quota_enabled(sb)) { printk(KERN_ERR - "EXT3-fs: Cannot change journalled " + "EXT3-fs: Cannot change journaled " "quota options when quota turned on.\n"); return 0; } @@ -1058,7 +1058,7 @@ set_qf_name: clear_qf_name: if (sb_any_quota_enabled(sb)) { printk(KERN_ERR "EXT3-fs: Cannot change " - "journalled quota options when " + "journaled quota options when " "quota turned on.\n"); return 0; } @@ -1169,14 +1169,14 @@ clear_qf_name: } if (!sbi->s_jquota_fmt) { - printk(KERN_ERR "EXT3-fs: journalled quota format " + printk(KERN_ERR "EXT3-fs: journaled quota format " "not specified.\n"); return 0; } } else { if (sbi->s_jquota_fmt) { - printk(KERN_ERR "EXT3-fs: journalled quota format " - "specified with no journalling " + printk(KERN_ERR "EXT3-fs: journaled quota format " + "specified with no journaling " "enabled.\n"); return 0; } @@ -1370,7 +1370,7 @@ static void ext3_orphan_cleanup (struct super_block * sb, int ret = ext3_quota_on_mount(sb, i); if (ret < 0) printk(KERN_ERR - "EXT3-fs: Cannot turn on journalled " + "EXT3-fs: Cannot turn on journaled " "quota: error %d\n", ret); } } @@ -2712,7 +2712,7 @@ static int ext3_release_dquot(struct dquot *dquot) static int ext3_mark_dquot_dirty(struct dquot *dquot) { - /* Are we journalling quotas? */ + /* Are we journaling quotas? */ if (EXT3_SB(dquot->dq_sb)->s_qf_names[USRQUOTA] || EXT3_SB(dquot->dq_sb)->s_qf_names[GRPQUOTA]) { dquot_mark_dquot_dirty(dquot); -- cgit v1.2.3-70-g09d2 From d06bf1d252fe16f5f0d13e04da7a9913420aa1cf Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Fri, 25 Jul 2008 01:46:18 -0700 Subject: ext3: correct mount option parsing to detect when quota options can be changed We should not allow user to change quota mount options when quota is just suspended. I would make mount options and internal quota state inconsistent. Also we should not allow user to change quota format when quota is turned on. On the other hand we can just silently ignore when some option is set to the value it already has (mount does this on remount). Cc: Signed-off-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/super.c | 29 ++++++++++++++++++++++------- 1 file changed, 22 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/ext3/super.c b/fs/ext3/super.c index 0a1bf82845c..615788c6843 100644 --- a/fs/ext3/super.c +++ b/fs/ext3/super.c @@ -842,7 +842,7 @@ static int parse_options (char *options, struct super_block *sb, int data_opt = 0; int option; #ifdef CONFIG_QUOTA - int qtype; + int qtype, qfmt; char *qname; #endif @@ -1018,7 +1018,9 @@ static int parse_options (char *options, struct super_block *sb, case Opt_grpjquota: qtype = GRPQUOTA; set_qf_name: - if (sb_any_quota_enabled(sb)) { + if ((sb_any_quota_enabled(sb) || + sb_any_quota_suspended(sb)) && + !sbi->s_qf_names[qtype]) { printk(KERN_ERR "EXT3-fs: Cannot change journaled " "quota options when quota turned on.\n"); @@ -1056,7 +1058,9 @@ set_qf_name: case Opt_offgrpjquota: qtype = GRPQUOTA; clear_qf_name: - if (sb_any_quota_enabled(sb)) { + if ((sb_any_quota_enabled(sb) || + sb_any_quota_suspended(sb)) && + sbi->s_qf_names[qtype]) { printk(KERN_ERR "EXT3-fs: Cannot change " "journaled quota options when " "quota turned on.\n"); @@ -1069,10 +1073,20 @@ clear_qf_name: sbi->s_qf_names[qtype] = NULL; break; case Opt_jqfmt_vfsold: - sbi->s_jquota_fmt = QFMT_VFS_OLD; - break; + qfmt = QFMT_VFS_OLD; + goto set_qf_format; case Opt_jqfmt_vfsv0: - sbi->s_jquota_fmt = QFMT_VFS_V0; + qfmt = QFMT_VFS_V0; +set_qf_format: + if ((sb_any_quota_enabled(sb) || + sb_any_quota_suspended(sb)) && + sbi->s_jquota_fmt != qfmt) { + printk(KERN_ERR "EXT3-fs: Cannot change " + "journaled quota options when " + "quota turned on.\n"); + return 0; + } + sbi->s_jquota_fmt = qfmt; break; case Opt_quota: case Opt_usrquota: @@ -1084,7 +1098,8 @@ clear_qf_name: set_opt(sbi->s_mount_opt, GRPQUOTA); break; case Opt_noquota: - if (sb_any_quota_enabled(sb)) { + if (sb_any_quota_enabled(sb) || + sb_any_quota_suspended(sb)) { printk(KERN_ERR "EXT3-fs: Cannot change quota " "options when quota turned on.\n"); return 0; -- cgit v1.2.3-70-g09d2 From 3850f7a521dc17659ef6758a219f083418788490 Mon Sep 17 00:00:00 2001 From: Duane Griffin Date: Fri, 25 Jul 2008 01:46:19 -0700 Subject: jbd: replace potentially false assertion with if block If an error occurs during jbd cache initialisation it is possible for the journal_head_cache to be NULL when journal_destroy_journal_head_cache is called. Replace the J_ASSERT with an if block to handle the situation correctly. Note that even with this fix things will break badly if jbd is statically compiled in and cache initialisation fails. Signed-off-by: Duane Griffin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/jbd/journal.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/jbd/journal.c b/fs/jbd/journal.c index b99c3b3654c..15ea16ad866 100644 --- a/fs/jbd/journal.c +++ b/fs/jbd/journal.c @@ -1636,9 +1636,10 @@ static int journal_init_journal_head_cache(void) static void journal_destroy_journal_head_cache(void) { - J_ASSERT(journal_head_cache != NULL); - kmem_cache_destroy(journal_head_cache); - journal_head_cache = NULL; + if (journal_head_cache) { + kmem_cache_destroy(journal_head_cache); + journal_head_cache = NULL; + } } /* -- cgit v1.2.3-70-g09d2 From f4d79ca2fa211cffc07306eeed7013448e77d7ec Mon Sep 17 00:00:00 2001 From: Duane Griffin Date: Fri, 25 Jul 2008 01:46:20 -0700 Subject: jbd: eliminate duplicated code in revocation table init/destroy functions The revocation table initialisation/destruction code is repeated for each of the two revocation tables stored in the journal. Refactoring the duplicated code into functions is tidier, simplifies the logic in initialisation in particular, and slightly reduces the code size. There should not be any functional change. Signed-off-by: Duane Griffin Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/jbd/revoke.c | 127 +++++++++++++++++++++++--------------------------------- 1 file changed, 51 insertions(+), 76 deletions(-) (limited to 'fs') diff --git a/fs/jbd/revoke.c b/fs/jbd/revoke.c index 1bb43e987f4..8ff5a7b89b9 100644 --- a/fs/jbd/revoke.c +++ b/fs/jbd/revoke.c @@ -195,109 +195,84 @@ void journal_destroy_revoke_caches(void) revoke_table_cache = NULL; } -/* Initialise the revoke table for a given journal to a given size. */ - -int journal_init_revoke(journal_t *journal, int hash_size) +static struct jbd_revoke_table_s *journal_init_revoke_table(int hash_size) { - int shift, tmp; + int shift = 0; + int tmp = hash_size; + struct jbd_revoke_table_s *table; - J_ASSERT (journal->j_revoke_table[0] == NULL); + table = kmem_cache_alloc(revoke_table_cache, GFP_KERNEL); + if (!table) + goto out; - shift = 0; - tmp = hash_size; while((tmp >>= 1UL) != 0UL) shift++; - journal->j_revoke_table[0] = kmem_cache_alloc(revoke_table_cache, GFP_KERNEL); - if (!journal->j_revoke_table[0]) - return -ENOMEM; - journal->j_revoke = journal->j_revoke_table[0]; - - /* Check that the hash_size is a power of two */ - J_ASSERT(is_power_of_2(hash_size)); - - journal->j_revoke->hash_size = hash_size; - - journal->j_revoke->hash_shift = shift; - - journal->j_revoke->hash_table = + table->hash_size = hash_size; + table->hash_shift = shift; + table->hash_table = kmalloc(hash_size * sizeof(struct list_head), GFP_KERNEL); - if (!journal->j_revoke->hash_table) { - kmem_cache_free(revoke_table_cache, journal->j_revoke_table[0]); - journal->j_revoke = NULL; - return -ENOMEM; + if (!table->hash_table) { + kmem_cache_free(revoke_table_cache, table); + table = NULL; + goto out; } for (tmp = 0; tmp < hash_size; tmp++) - INIT_LIST_HEAD(&journal->j_revoke->hash_table[tmp]); + INIT_LIST_HEAD(&table->hash_table[tmp]); - journal->j_revoke_table[1] = kmem_cache_alloc(revoke_table_cache, GFP_KERNEL); - if (!journal->j_revoke_table[1]) { - kfree(journal->j_revoke_table[0]->hash_table); - kmem_cache_free(revoke_table_cache, journal->j_revoke_table[0]); - return -ENOMEM; +out: + return table; +} + +static void journal_destroy_revoke_table(struct jbd_revoke_table_s *table) +{ + int i; + struct list_head *hash_list; + + for (i = 0; i < table->hash_size; i++) { + hash_list = &table->hash_table[i]; + J_ASSERT(list_empty(hash_list)); } - journal->j_revoke = journal->j_revoke_table[1]; + kfree(table->hash_table); + kmem_cache_free(revoke_table_cache, table); +} - /* Check that the hash_size is a power of two */ +/* Initialise the revoke table for a given journal to a given size. */ +int journal_init_revoke(journal_t *journal, int hash_size) +{ + J_ASSERT(journal->j_revoke_table[0] == NULL); J_ASSERT(is_power_of_2(hash_size)); - journal->j_revoke->hash_size = hash_size; + journal->j_revoke_table[0] = journal_init_revoke_table(hash_size); + if (!journal->j_revoke_table[0]) + goto fail0; - journal->j_revoke->hash_shift = shift; + journal->j_revoke_table[1] = journal_init_revoke_table(hash_size); + if (!journal->j_revoke_table[1]) + goto fail1; - journal->j_revoke->hash_table = - kmalloc(hash_size * sizeof(struct list_head), GFP_KERNEL); - if (!journal->j_revoke->hash_table) { - kfree(journal->j_revoke_table[0]->hash_table); - kmem_cache_free(revoke_table_cache, journal->j_revoke_table[0]); - kmem_cache_free(revoke_table_cache, journal->j_revoke_table[1]); - journal->j_revoke = NULL; - return -ENOMEM; - } - - for (tmp = 0; tmp < hash_size; tmp++) - INIT_LIST_HEAD(&journal->j_revoke->hash_table[tmp]); + journal->j_revoke = journal->j_revoke_table[1]; spin_lock_init(&journal->j_revoke_lock); return 0; -} -/* Destoy a journal's revoke table. The table must already be empty! */ +fail1: + journal_destroy_revoke_table(journal->j_revoke_table[0]); +fail0: + return -ENOMEM; +} +/* Destroy a journal's revoke table. The table must already be empty! */ void journal_destroy_revoke(journal_t *journal) { - struct jbd_revoke_table_s *table; - struct list_head *hash_list; - int i; - - table = journal->j_revoke_table[0]; - if (!table) - return; - - for (i=0; ihash_size; i++) { - hash_list = &table->hash_table[i]; - J_ASSERT (list_empty(hash_list)); - } - - kfree(table->hash_table); - kmem_cache_free(revoke_table_cache, table); - journal->j_revoke = NULL; - - table = journal->j_revoke_table[1]; - if (!table) - return; - - for (i=0; ihash_size; i++) { - hash_list = &table->hash_table[i]; - J_ASSERT (list_empty(hash_list)); - } - - kfree(table->hash_table); - kmem_cache_free(revoke_table_cache, table); journal->j_revoke = NULL; + if (journal->j_revoke_table[0]) + journal_destroy_revoke_table(journal->j_revoke_table[0]); + if (journal->j_revoke_table[1]) + journal_destroy_revoke_table(journal->j_revoke_table[1]); } -- cgit v1.2.3-70-g09d2 From 1984bb763c2e50d0ebfb0cf56d1b319bd7afe63a Mon Sep 17 00:00:00 2001 From: Duane Griffin Date: Fri, 25 Jul 2008 01:46:21 -0700 Subject: jbd: tidy up revoke cache initialisation and destruction Make revocation cache destruction safe to call if initialisation fails partially or entirely. This allows it to be used to cleanup in the case of initialisation failure, simplifying that code slightly. Signed-off-by: Duane Griffin Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/jbd/revoke.c | 36 +++++++++++++++++++++++------------- 1 file changed, 23 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/jbd/revoke.c b/fs/jbd/revoke.c index 8ff5a7b89b9..c7bd649bbbd 100644 --- a/fs/jbd/revoke.c +++ b/fs/jbd/revoke.c @@ -166,33 +166,43 @@ static struct jbd_revoke_record_s *find_revoke_record(journal_t *journal, return NULL; } +void journal_destroy_revoke_caches(void) +{ + if (revoke_record_cache) { + kmem_cache_destroy(revoke_record_cache); + revoke_record_cache = NULL; + } + if (revoke_table_cache) { + kmem_cache_destroy(revoke_table_cache); + revoke_table_cache = NULL; + } +} + int __init journal_init_revoke_caches(void) { + J_ASSERT(!revoke_record_cache); + J_ASSERT(!revoke_table_cache); + revoke_record_cache = kmem_cache_create("revoke_record", sizeof(struct jbd_revoke_record_s), 0, SLAB_HWCACHE_ALIGN|SLAB_TEMPORARY, NULL); if (!revoke_record_cache) - return -ENOMEM; + goto record_cache_failure; revoke_table_cache = kmem_cache_create("revoke_table", sizeof(struct jbd_revoke_table_s), 0, SLAB_TEMPORARY, NULL); - if (!revoke_table_cache) { - kmem_cache_destroy(revoke_record_cache); - revoke_record_cache = NULL; - return -ENOMEM; - } + if (!revoke_table_cache) + goto table_cache_failure; + return 0; -} -void journal_destroy_revoke_caches(void) -{ - kmem_cache_destroy(revoke_record_cache); - revoke_record_cache = NULL; - kmem_cache_destroy(revoke_table_cache); - revoke_table_cache = NULL; +table_cache_failure: + journal_destroy_revoke_caches(); +record_cache_failure: + return -ENOMEM; } static struct jbd_revoke_table_s *journal_init_revoke_table(int hash_size) -- cgit v1.2.3-70-g09d2 From 9ebfbe9f926553eabc21b4400918d1216b27ed0c Mon Sep 17 00:00:00 2001 From: Shen Feng Date: Fri, 25 Jul 2008 01:46:21 -0700 Subject: ext3: improve some code in rb tree part of dir.c - remove unnecessary code in free_rb_tree_fname - rename free_rb_tree_fname to ext3_htree_create_dir_info since it and ext3_htree_free_dir_info are a pair - replace kmalloc with kzalloc in ext3_htree_free_dir_info Signed-off-by: Shen Feng Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/dir.c | 14 ++++---------- 1 file changed, 4 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/ext3/dir.c b/fs/ext3/dir.c index 8ca3bfd7242..2eea96ec78e 100644 --- a/fs/ext3/dir.c +++ b/fs/ext3/dir.c @@ -272,7 +272,7 @@ static void free_rb_tree_fname(struct rb_root *root) while (n) { /* Do the node's children first */ - if ((n)->rb_left) { + if (n->rb_left) { n = n->rb_left; continue; } @@ -301,24 +301,18 @@ static void free_rb_tree_fname(struct rb_root *root) parent->rb_right = NULL; n = parent; } - root->rb_node = NULL; } -static struct dir_private_info *create_dir_info(loff_t pos) +static struct dir_private_info *ext3_htree_create_dir_info(loff_t pos) { struct dir_private_info *p; - p = kmalloc(sizeof(struct dir_private_info), GFP_KERNEL); + p = kzalloc(sizeof(struct dir_private_info), GFP_KERNEL); if (!p) return NULL; - p->root.rb_node = NULL; - p->curr_node = NULL; - p->extra_fname = NULL; - p->last_pos = 0; p->curr_hash = pos2maj_hash(pos); p->curr_minor_hash = pos2min_hash(pos); - p->next_hash = 0; return p; } @@ -433,7 +427,7 @@ static int ext3_dx_readdir(struct file * filp, int ret; if (!info) { - info = create_dir_info(filp->f_pos); + info = ext3_htree_create_dir_info(filp->f_pos); if (!info) return -ENOMEM; filp->private_data = info; -- cgit v1.2.3-70-g09d2 From 3f31fddfa26b7594b44ff2b34f9a04ba409e0f91 Mon Sep 17 00:00:00 2001 From: Mingming Cao Date: Fri, 25 Jul 2008 01:46:22 -0700 Subject: jbd: fix race between free buffer and commit transaction journal_try_to_free_buffers() could race with jbd commit transaction when the later is holding the buffer reference while waiting for the data buffer to flush to disk. If the caller of journal_try_to_free_buffers() request tries hard to release the buffers, it will treat the failure as error and return back to the caller. We have seen the directo IO failed due to this race. Some of the caller of releasepage() also expecting the buffer to be dropped when passed with GFP_KERNEL mask to the releasepage()->journal_try_to_free_buffers(). With this patch, if the caller is passing the __GFP_WAIT and __GFP_FS to indicating this call could wait, in case of try_to_free_buffers() failed, let's waiting for journal_commit_transaction() to finish commit the current committing transaction, then try to free those buffers again. [akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Mingming Cao Reviewed-by: Badari Pulavarty Acked-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/jbd/transaction.c | 57 ++++++++++++++++++++++++++++++++++++++++++++++++++-- mm/filemap.c | 3 +-- 2 files changed, 56 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/jbd/transaction.c b/fs/jbd/transaction.c index 67ff2024c23..8dee3200750 100644 --- a/fs/jbd/transaction.c +++ b/fs/jbd/transaction.c @@ -1648,12 +1648,42 @@ out: return; } +/* + * journal_try_to_free_buffers() could race with journal_commit_transaction() + * The latter might still hold the a count on buffers when inspecting + * them on t_syncdata_list or t_locked_list. + * + * journal_try_to_free_buffers() will call this function to + * wait for the current transaction to finish syncing data buffers, before + * tryinf to free that buffer. + * + * Called with journal->j_state_lock held. + */ +static void journal_wait_for_transaction_sync_data(journal_t *journal) +{ + transaction_t *transaction = NULL; + tid_t tid; + + spin_lock(&journal->j_state_lock); + transaction = journal->j_committing_transaction; + + if (!transaction) { + spin_unlock(&journal->j_state_lock); + return; + } + + tid = transaction->t_tid; + spin_unlock(&journal->j_state_lock); + log_wait_commit(journal, tid); +} /** * int journal_try_to_free_buffers() - try to free page buffers. * @journal: journal for operation * @page: to try and free - * @unused_gfp_mask: unused + * @gfp_mask: we use the mask to detect how hard should we try to release + * buffers. If __GFP_WAIT and __GFP_FS is set, we wait for commit code to + * release the buffers. * * * For all the buffers on this page, @@ -1682,9 +1712,11 @@ out: * journal_try_to_free_buffer() is changing its state. But that * cannot happen because we never reallocate freed data as metadata * while the data is part of a transaction. Yes? + * + * Return 0 on failure, 1 on success */ int journal_try_to_free_buffers(journal_t *journal, - struct page *page, gfp_t unused_gfp_mask) + struct page *page, gfp_t gfp_mask) { struct buffer_head *head; struct buffer_head *bh; @@ -1713,7 +1745,28 @@ int journal_try_to_free_buffers(journal_t *journal, if (buffer_jbd(bh)) goto busy; } while ((bh = bh->b_this_page) != head); + ret = try_to_free_buffers(page); + + /* + * There are a number of places where journal_try_to_free_buffers() + * could race with journal_commit_transaction(), the later still + * holds the reference to the buffers to free while processing them. + * try_to_free_buffers() failed to free those buffers. Some of the + * caller of releasepage() request page buffers to be dropped, otherwise + * treat the fail-to-free as errors (such as generic_file_direct_IO()) + * + * So, if the caller of try_to_release_page() wants the synchronous + * behaviour(i.e make sure buffers are dropped upon return), + * let's wait for the current transaction to finish flush of + * dirty data buffers, then try to free those buffers again, + * with the journal locked. + */ + if (ret == 0 && (gfp_mask & __GFP_WAIT) && (gfp_mask & __GFP_FS)) { + journal_wait_for_transaction_sync_data(journal); + ret = try_to_free_buffers(page); + } + busy: return ret; } diff --git a/mm/filemap.c b/mm/filemap.c index 7675b91f4f6..5d4c880d7cd 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -2563,9 +2563,8 @@ EXPORT_SYMBOL(generic_file_aio_write); * Otherwise return zero. * * The @gfp_mask argument specifies whether I/O may be performed to release - * this page (__GFP_IO), and whether the call may block (__GFP_WAIT). + * this page (__GFP_IO), and whether the call may block (__GFP_WAIT & __GFP_FS). * - * NOTE: @gfp_mask may go away, and this function may become non-blocking. */ int try_to_release_page(struct page *page, gfp_t gfp_mask) { -- cgit v1.2.3-70-g09d2 From ef1afd39519b74fbe1f63c9ab5a14490effec0e3 Mon Sep 17 00:00:00 2001 From: Shen Feng Date: Fri, 25 Jul 2008 01:46:23 -0700 Subject: ext3: remove double definitions of xattr macros remove the definitions of macros: XATTR_TRUSTED_PREFIX XATTR_USER_PREFIX since they are defined in linux/xattr.h Signed-off-by: Shen Feng Signed-off-by: Mingming Cao Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/xattr_security.c | 2 +- fs/ext3/xattr_trusted.c | 4 +--- fs/ext3/xattr_user.c | 4 +--- 3 files changed, 3 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/ext3/xattr_security.c b/fs/ext3/xattr_security.c index 821efaf2b94..37b81097bdf 100644 --- a/fs/ext3/xattr_security.c +++ b/fs/ext3/xattr_security.c @@ -15,7 +15,7 @@ static size_t ext3_xattr_security_list(struct inode *inode, char *list, size_t list_size, const char *name, size_t name_len) { - const size_t prefix_len = sizeof(XATTR_SECURITY_PREFIX)-1; + const size_t prefix_len = XATTR_SECURITY_PREFIX_LEN; const size_t total_len = prefix_len + name_len + 1; diff --git a/fs/ext3/xattr_trusted.c b/fs/ext3/xattr_trusted.c index 0327497a55c..c7c41a410c4 100644 --- a/fs/ext3/xattr_trusted.c +++ b/fs/ext3/xattr_trusted.c @@ -13,13 +13,11 @@ #include #include "xattr.h" -#define XATTR_TRUSTED_PREFIX "trusted." - static size_t ext3_xattr_trusted_list(struct inode *inode, char *list, size_t list_size, const char *name, size_t name_len) { - const size_t prefix_len = sizeof(XATTR_TRUSTED_PREFIX)-1; + const size_t prefix_len = XATTR_TRUSTED_PREFIX_LEN; const size_t total_len = prefix_len + name_len + 1; if (!capable(CAP_SYS_ADMIN)) diff --git a/fs/ext3/xattr_user.c b/fs/ext3/xattr_user.c index 1abd8f92c44..430fe63b31b 100644 --- a/fs/ext3/xattr_user.c +++ b/fs/ext3/xattr_user.c @@ -12,13 +12,11 @@ #include #include "xattr.h" -#define XATTR_USER_PREFIX "user." - static size_t ext3_xattr_user_list(struct inode *inode, char *list, size_t list_size, const char *name, size_t name_len) { - const size_t prefix_len = sizeof(XATTR_USER_PREFIX)-1; + const size_t prefix_len = XATTR_USER_PREFIX_LEN; const size_t total_len = prefix_len + name_len + 1; if (!test_opt(inode->i_sb, XATTR_USER)) -- cgit v1.2.3-70-g09d2 From ae76dd9a6b5bbe5315fb7028e03f68f75b8538f3 Mon Sep 17 00:00:00 2001 From: Duane Griffin Date: Fri, 25 Jul 2008 01:46:23 -0700 Subject: ext3: handle corrupted orphan list at mount If the orphan node list includes valid, untruncatable nodes with nlink > 0 the ext3_orphan_cleanup loop which attempts to delete them will not do so, causing it to loop forever. Fix by checking for such nodes in the ext3_orphan_get function. This patch fixes the second case (image hdb.20000009.softlockup.gz) reported in http://bugzilla.kernel.org/show_bug.cgi?id=10882. [akpm@linux-foundation.org: coding-style fixes] [akpm@linux-foundation.org: printk warning fix] Signed-off-by: Duane Griffin Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/ialloc.c | 9 +++++++++ fs/ext3/inode.c | 20 ++++++++++++++------ include/linux/ext3_fs.h | 1 + 3 files changed, 24 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/ext3/ialloc.c b/fs/ext3/ialloc.c index 77126821b2e..47b678d73e7 100644 --- a/fs/ext3/ialloc.c +++ b/fs/ext3/ialloc.c @@ -669,6 +669,14 @@ struct inode *ext3_orphan_get(struct super_block *sb, unsigned long ino) if (IS_ERR(inode)) goto iget_failed; + /* + * If the orphans has i_nlinks > 0 then it should be able to be + * truncated, otherwise it won't be removed from the orphan list + * during processing and an infinite loop will result. + */ + if (inode->i_nlink && !ext3_can_truncate(inode)) + goto bad_orphan; + if (NEXT_ORPHAN(inode) > max_ino) goto bad_orphan; brelse(bitmap_bh); @@ -690,6 +698,7 @@ bad_orphan: printk(KERN_NOTICE "NEXT_ORPHAN(inode)=%u\n", NEXT_ORPHAN(inode)); printk(KERN_NOTICE "max_ino=%lu\n", max_ino); + printk(KERN_NOTICE "i_nlink=%u\n", inode->i_nlink); /* Avoid freeing blocks if we got a bad deleted inode */ if (inode->i_nlink == 0) inode->i_blocks = 0; diff --git a/fs/ext3/inode.c b/fs/ext3/inode.c index 6ae4ecf3ce4..74b432fa166 100644 --- a/fs/ext3/inode.c +++ b/fs/ext3/inode.c @@ -2253,6 +2253,19 @@ static void ext3_free_branches(handle_t *handle, struct inode *inode, } } +int ext3_can_truncate(struct inode *inode) +{ + if (IS_APPEND(inode) || IS_IMMUTABLE(inode)) + return 0; + if (S_ISREG(inode->i_mode)) + return 1; + if (S_ISDIR(inode->i_mode)) + return 1; + if (S_ISLNK(inode->i_mode)) + return !ext3_inode_is_fast_symlink(inode); + return 0; +} + /* * ext3_truncate() * @@ -2297,12 +2310,7 @@ void ext3_truncate(struct inode *inode) unsigned blocksize = inode->i_sb->s_blocksize; struct page *page; - if (!(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) || - S_ISLNK(inode->i_mode))) - return; - if (ext3_inode_is_fast_symlink(inode)) - return; - if (IS_APPEND(inode) || IS_IMMUTABLE(inode)) + if (!ext3_can_truncate(inode)) return; /* diff --git a/include/linux/ext3_fs.h b/include/linux/ext3_fs.h index 36c54039637..80171ee89a2 100644 --- a/include/linux/ext3_fs.h +++ b/include/linux/ext3_fs.h @@ -832,6 +832,7 @@ extern void ext3_discard_reservation (struct inode *); extern void ext3_dirty_inode(struct inode *); extern int ext3_change_inode_journal_flag(struct inode *, int); extern int ext3_get_inode_loc(struct inode *, struct ext3_iloc *); +extern int ext3_can_truncate(struct inode *inode); extern void ext3_truncate (struct inode *); extern void ext3_set_inode_flags(struct inode *); extern void ext3_get_inode_flags(struct ext3_inode_info *); -- cgit v1.2.3-70-g09d2 From 95450f5a7e53d5752ce1a0d0b8282e10fe745ae0 Mon Sep 17 00:00:00 2001 From: Hidehiro Kawai Date: Fri, 25 Jul 2008 01:46:24 -0700 Subject: ext3: don't read inode block if the buffer has a write error A transient I/O error can corrupt inode data. Here is the scenario: (1) update inode_A at the block_B (2) pdflush writes out new inode_A to the filesystem, but it results in write I/O error, at this point, BH_Uptodate flag of the buffer for block_B is cleared and BH_Write_EIO is set (3) create new inode_C which located at block_B, and __ext3_get_inode_loc() tries to read on-disk block_B because the buffer is not uptodate (4) if it can read on-disk block_B successfully, inode_A is overwritten by old data This patch makes __ext3_get_inode_loc() not read the inode block if the buffer has BH_Write_EIO flag. In this case, the buffer should have the latest information, so setting the uptodate flag to the buffer (this avoids WARN_ON_ONCE() in mark_buffer_dirty().) According to this change, we would need to test BH_Write_EIO flag for the error checking. Currently nobody checks write I/O errors on metadata buffers, but it will be done in other patches I'm working on. Signed-off-by: Hidehiro Kawai Cc: sugita Cc: Satoshi OSHIMA Cc: Nick Piggin Cc: Jan Kara Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/inode.c | 10 ++++++++++ 1 file changed, 10 insertions(+) (limited to 'fs') diff --git a/fs/ext3/inode.c b/fs/ext3/inode.c index 74b432fa166..36f74f17a11 100644 --- a/fs/ext3/inode.c +++ b/fs/ext3/inode.c @@ -2521,6 +2521,16 @@ static int __ext3_get_inode_loc(struct inode *inode, } if (!buffer_uptodate(bh)) { lock_buffer(bh); + + /* + * If the buffer has the write error flag, we have failed + * to write out another inode in the same block. In this + * case, we don't have to read the block because we may + * read the old inode data successfully. + */ + if (buffer_write_io_error(bh) && !buffer_uptodate(bh)) + set_buffer_uptodate(bh); + if (buffer_uptodate(bh)) { /* someone brought it uptodate while we waited */ unlock_buffer(bh); -- cgit v1.2.3-70-g09d2 From 3ccc3167b0e5d46ab3bf03e22fbdb7616ce038cd Mon Sep 17 00:00:00 2001 From: Duane Griffin Date: Fri, 25 Jul 2008 01:46:26 -0700 Subject: ext3: handle deleting corrupted indirect blocks While freeing indirect blocks we attach a journal head to the parent buffer head, free the blocks, then journal the parent. If the indirect block list is corrupted and points to the parent the journal head will be detached when the block is cleared, causing an OOPS. Check for that explicitly and handle it gracefully. This patch fixes the third case (image hdb.20000057.nullderef.gz) reported in http://bugzilla.kernel.org/show_bug.cgi?id=10882. Immediately above the change, in the ext3_free_data function, we call ext3_clear_blocks to clear the indirect blocks in this parent block. If one of those blocks happens to actually be the parent block it will clear b_private / BH_JBD. I did the check at the end rather than earlier as it seemed more elegant. I don't think there should be much practical difference, although it is possible the FS may not be quite so badly corrupted if we did it the other way (and didn't clear the block at all). To be honest, I'm not convinced there aren't other similar failure modes lurking in this code, although I couldn't find any with a quick review. [akpm@linux-foundation.org: fix printk warning] Signed-off-by: Duane Griffin Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/inode.c | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ext3/inode.c b/fs/ext3/inode.c index 36f74f17a11..3bf07d70b91 100644 --- a/fs/ext3/inode.c +++ b/fs/ext3/inode.c @@ -2127,7 +2127,21 @@ static void ext3_free_data(handle_t *handle, struct inode *inode, if (this_bh) { BUFFER_TRACE(this_bh, "call ext3_journal_dirty_metadata"); - ext3_journal_dirty_metadata(handle, this_bh); + + /* + * The buffer head should have an attached journal head at this + * point. However, if the data is corrupted and an indirect + * block pointed to itself, it would have been detached when + * the block was cleared. Check for this instead of OOPSing. + */ + if (bh2jh(this_bh)) + ext3_journal_dirty_metadata(handle, this_bh); + else + ext3_error(inode->i_sb, "ext3_free_data", + "circular indirect block detected, " + "inode=%lu, block=%llu", + inode->i_ino, + (unsigned long long)this_bh->b_blocknr); } } -- cgit v1.2.3-70-g09d2 From a10320e8f7c4dcfa050aac566092f29b40458d5a Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Fri, 25 Jul 2008 01:46:26 -0700 Subject: jbd: unexport journal_update_superblock Remove the unused EXPORT_SYMBOL(journal_update_superblock). Signed-off-by: Adrian Bunk Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/jbd/journal.c | 1 - 1 file changed, 1 deletion(-) (limited to 'fs') diff --git a/fs/jbd/journal.c b/fs/jbd/journal.c index 15ea16ad866..aa7143a8349 100644 --- a/fs/jbd/journal.c +++ b/fs/jbd/journal.c @@ -68,7 +68,6 @@ EXPORT_SYMBOL(journal_set_features); EXPORT_SYMBOL(journal_create); EXPORT_SYMBOL(journal_load); EXPORT_SYMBOL(journal_destroy); -EXPORT_SYMBOL(journal_update_superblock); EXPORT_SYMBOL(journal_abort); EXPORT_SYMBOL(journal_errno); EXPORT_SYMBOL(journal_ack_err); -- cgit v1.2.3-70-g09d2 From fc80c44277b3c92d808b73e9d40e120229aa4b6a Mon Sep 17 00:00:00 2001 From: Toshiyuki Okajima Date: Fri, 25 Jul 2008 01:46:29 -0700 Subject: jbd: positively dispose the unmapped data buffers in journal_commit_transaction() After ext3-ordered files are truncated, there is a possibility that the pages which cannot be estimated still remain. Remaining pages can be released when the system has really few memory. So, it is not memory leakage. But the resource management software etc. may not work correctly. It is possible that journal_unmap_buffer() cannot release the buffers, and the pages to which they belong because they are attached to a commiting transaction and journal_unmap_buffer() cannot release them. To release such the buffers and the pages later, journal_unmap_buffer() leaves it to journal_commit_transaction(). (journal_unmap_buffer() puts the mark 'BH_Freed' to the buffers so that journal_commit_transaction() can identify whether they can be released or not.) In the journalled mode and the writeback mode, jbd does with only metadata buffers. But in the ordered mode, jbd does with metadata buffers and also data buffers. Actually, journal_commit_transaction() releases only the metadata buffers of which release is demanded by journal_unmap_buffer(), and also releases the pages to which they belong if possible. As a result, the data buffers of which release is demanded by journal_unmap_buffer() remain after a transaction commits. And also the pages to which they belong remain. Such the remained pages don't have mapping any longer. Due to this fact, there is a possibility that the pages which cannot be estimated remain. The metadata buffers marked 'BH_Freed' and the pages to which they belong can be released at 'JBD: commit phase 7'. Therefore, by applying the same code into 'JBD: commit phase 2' (where the data buffers are done with), journal_commit_transaction() can also release the data buffers marked 'BH_Freed' and the pages to which they belong. As a result, all the buffers marked 'BH_Freed' can be released, and also all the pages to which these buffers belong can be released at journal_commit_transaction(). So, the page which cannot be estimated is lost. <> > spin_lock(&journal->j_list_lock); > while (commit_transaction->t_forget) { > transaction_t *cp_transaction; > struct buffer_head *bh; > > jh = commit_transaction->t_forget; >... > if (buffer_freed(bh)) { > ^^^^^^^^^^^^^^^^^^^^^^^^ > clear_buffer_freed(bh); > ^^^^^^^^^^^^^^^^^^^^^^^^ > clear_buffer_jbddirty(bh); > } > > if (buffer_jbddirty(bh)) { > JBUFFER_TRACE(jh, "add to new checkpointing trans"); > __journal_insert_checkpoint(jh, commit_transaction); > JBUFFER_TRACE(jh, "refile for checkpoint writeback"); > __journal_refile_buffer(jh); > jbd_unlock_bh_state(bh); > } else { > J_ASSERT_BH(bh, !buffer_dirty(bh)); > ... > JBUFFER_TRACE(jh, "refile or unfile freed buffer"); > __journal_refile_buffer(jh); > if (!jh->b_transaction) { > jbd_unlock_bh_state(bh); > /* needs a brelse */ > journal_remove_journal_head(bh); > release_buffer_page(bh); > ^^^^^^^^^^^^^^^^^^^^^^^^ > } else > } **************************************************************** * Apply the code of "^^^^^^" lines into 'JBD: commit phase 2' * **************************************************************** At journal_commit_transaction() code, there is one extra message in the series of jbd debug messages. ("JBD: commit phase 2") This patch fixes it, too. Signed-off-by: Toshiyuki Okajima Acked-by: Jan Kara Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/jbd/commit.c | 29 ++++++++++++++++++++--------- 1 file changed, 20 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/jbd/commit.c b/fs/jbd/commit.c index 5a8ca61498c..f943b9b3f20 100644 --- a/fs/jbd/commit.c +++ b/fs/jbd/commit.c @@ -36,7 +36,7 @@ static void journal_end_buffer_io_sync(struct buffer_head *bh, int uptodate) /* * When an ext3-ordered file is truncated, it is possible that many pages are - * not sucessfully freed, because they are attached to a committing transaction. + * not successfully freed, because they are attached to a committing transaction. * After the transaction commits, these pages are left on the LRU, with no * ->mapping, and with attached buffers. These pages are trivially reclaimable * by the VM, but their apparent absence upsets the VM accounting, and it makes @@ -45,8 +45,8 @@ static void journal_end_buffer_io_sync(struct buffer_head *bh, int uptodate) * So here, we have a buffer which has just come off the forget list. Look to * see if we can strip all buffers from the backing page. * - * Called under lock_journal(), and possibly under journal_datalist_lock. The - * caller provided us with a ref against the buffer, and we drop that here. + * Called under journal->j_list_lock. The caller provided us with a ref + * against the buffer, and we drop that here. */ static void release_buffer_page(struct buffer_head *bh) { @@ -77,6 +77,19 @@ nope: __brelse(bh); } +/* + * Decrement reference counter for data buffer. If it has been marked + * 'BH_Freed', release it and the page to which it belongs if possible. + */ +static void release_data_buffer(struct buffer_head *bh) +{ + if (buffer_freed(bh)) { + clear_buffer_freed(bh); + release_buffer_page(bh); + } else + put_bh(bh); +} + /* * Try to acquire jbd_lock_bh_state() against the buffer, when j_list_lock is * held. For ranking reasons we must trylock. If we lose, schedule away and @@ -231,7 +244,7 @@ write_out_data: if (locked) unlock_buffer(bh); BUFFER_TRACE(bh, "already cleaned up"); - put_bh(bh); + release_data_buffer(bh); continue; } if (locked && test_clear_buffer_dirty(bh)) { @@ -258,10 +271,10 @@ write_out_data: if (locked) unlock_buffer(bh); journal_remove_journal_head(bh); - /* Once for our safety reference, once for + /* One for our safety reference, other for * journal_remove_journal_head() */ put_bh(bh); - put_bh(bh); + release_data_buffer(bh); } if (need_resched() || spin_needbreak(&journal->j_list_lock)) { @@ -443,7 +456,7 @@ void journal_commit_transaction(journal_t *journal) } else { jbd_unlock_bh_state(bh); } - put_bh(bh); + release_data_buffer(bh); cond_resched_lock(&journal->j_list_lock); } spin_unlock(&journal->j_list_lock); @@ -453,8 +466,6 @@ void journal_commit_transaction(journal_t *journal) journal_write_revoke_records(journal, commit_transaction); - jbd_debug(3, "JBD: commit phase 2\n"); - /* * If we found any dirty or locked buffers, then we should have * looped back up to the write_out_data label. If there weren't -- cgit v1.2.3-70-g09d2 From 8ef2720397bb813d4985405a5ae7b8ad6474188b Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 25 Jul 2008 01:46:29 -0700 Subject: ext3: kill 2 useless magic numbers dx_root_limit() will never return 20, and I can't figure out what 20 stands for. This function has never changed since htree directory indexing was merged. Similar for dx_node_limit() and the magic 22. Signed-off-by: Li Zefan Acked-by: Andreas Dilger Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/namei.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ext3/namei.c b/fs/ext3/namei.c index 0b8cf80154f..d282ea87008 100644 --- a/fs/ext3/namei.c +++ b/fs/ext3/namei.c @@ -240,13 +240,13 @@ static inline unsigned dx_root_limit (struct inode *dir, unsigned infosize) { unsigned entry_space = dir->i_sb->s_blocksize - EXT3_DIR_REC_LEN(1) - EXT3_DIR_REC_LEN(2) - infosize; - return 0? 20: entry_space / sizeof(struct dx_entry); + return entry_space / sizeof(struct dx_entry); } static inline unsigned dx_node_limit (struct inode *dir) { unsigned entry_space = dir->i_sb->s_blocksize - EXT3_DIR_REC_LEN(0); - return 0? 22: entry_space / sizeof(struct dx_entry); + return entry_space / sizeof(struct dx_entry); } /* -- cgit v1.2.3-70-g09d2 From cbe5f466f6995e10a10c7ae66d6dc8608f08a6b8 Mon Sep 17 00:00:00 2001 From: Hidehiro Kawai Date: Fri, 25 Jul 2008 01:46:30 -0700 Subject: jbd: don't abort if flushing file data failed In ordered mode, the current jbd aborts the journal if a file data buffer has an error. But this behavior is unintended, and we found that it has been adopted accidentally. This patch undoes it and just calls printk() instead of aborting the journal. Additionally, set AS_EIO into the address_space object of the failed buffer which is submitted by journal_do_submit_data() so that fsync() can get -EIO. Missing error checkings are also added to inform errors on file data buffers to the user. The following buffers are targeted. (a) the buffer which has already been written out by pdflush (b) the buffer which has been unlocked before scanned in the t_locked_list loop [akpm@linux-foundation.org: improve grammar in a printk] Signed-off-by: Hidehiro Kawai Acked-by: Jan Kara Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/jbd/commit.c | 35 ++++++++++++++++++++++++++++------- 1 file changed, 28 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/jbd/commit.c b/fs/jbd/commit.c index f943b9b3f20..2eccbfaa1d4 100644 --- a/fs/jbd/commit.c +++ b/fs/jbd/commit.c @@ -185,7 +185,7 @@ static void journal_do_submit_data(struct buffer_head **wbuf, int bufs) /* * Submit all the data buffers to disk */ -static void journal_submit_data_buffers(journal_t *journal, +static int journal_submit_data_buffers(journal_t *journal, transaction_t *commit_transaction) { struct journal_head *jh; @@ -193,6 +193,7 @@ static void journal_submit_data_buffers(journal_t *journal, int locked; int bufs = 0; struct buffer_head **wbuf = journal->j_wbuf; + int err = 0; /* * Whenever we unlock the journal and sleep, things can get added @@ -266,6 +267,8 @@ write_out_data: put_bh(bh); } else { BUFFER_TRACE(bh, "writeout complete: unfile"); + if (unlikely(!buffer_uptodate(bh))) + err = -EIO; __journal_unfile_buffer(jh); jbd_unlock_bh_state(bh); if (locked) @@ -284,6 +287,8 @@ write_out_data: } spin_unlock(&journal->j_list_lock); journal_do_submit_data(wbuf, bufs); + + return err; } /* @@ -423,8 +428,7 @@ void journal_commit_transaction(journal_t *journal) * Now start flushing things to disk, in the order they appear * on the transaction lists. Data blocks go first. */ - err = 0; - journal_submit_data_buffers(journal, commit_transaction); + err = journal_submit_data_buffers(journal, commit_transaction); /* * Wait for all previously submitted IO to complete. @@ -439,10 +443,21 @@ void journal_commit_transaction(journal_t *journal) if (buffer_locked(bh)) { spin_unlock(&journal->j_list_lock); wait_on_buffer(bh); - if (unlikely(!buffer_uptodate(bh))) - err = -EIO; spin_lock(&journal->j_list_lock); } + if (unlikely(!buffer_uptodate(bh))) { + if (TestSetPageLocked(bh->b_page)) { + spin_unlock(&journal->j_list_lock); + lock_page(bh->b_page); + spin_lock(&journal->j_list_lock); + } + if (bh->b_page->mapping) + set_bit(AS_EIO, &bh->b_page->mapping->flags); + + unlock_page(bh->b_page); + SetPageError(bh->b_page); + err = -EIO; + } if (!inverted_lock(journal, bh)) { put_bh(bh); spin_lock(&journal->j_list_lock); @@ -461,8 +476,14 @@ void journal_commit_transaction(journal_t *journal) } spin_unlock(&journal->j_list_lock); - if (err) - journal_abort(journal, err); + if (err) { + char b[BDEVNAME_SIZE]; + + printk(KERN_WARNING + "JBD: Detected IO errors while flushing file data " + "on %s\n", bdevname(journal->j_fs_dev, b)); + err = 0; + } journal_write_revoke_records(journal, commit_transaction); -- cgit v1.2.3-70-g09d2 From 275c0a8f1253a7542ad9726956c918d8a1f694c4 Mon Sep 17 00:00:00 2001 From: Duane Griffin Date: Fri, 25 Jul 2008 01:46:31 -0700 Subject: ext3: validate directory entry data before use ext3_dx_find_entry uses ext3_next_entry without verifying that the entry is valid. If its rec_len == 0 this causes an infinite loop. Refactor the loop to check the validity of entries before checking whether they match and moving onto the next one. There are other uses of ext3_next_entry in this file which also look problematic. They should be reviewed and fixed if/when we have a test-case that triggers them. This patch fixes the first case (image hdb.25.softlockup.gz) reported in http://bugzilla.kernel.org/show_bug.cgi?id=10882. Signed-off-by: Duane Griffin Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/namei.c | 22 ++++++++++++---------- 1 file changed, 12 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/ext3/namei.c b/fs/ext3/namei.c index d282ea87008..de13e919cd8 100644 --- a/fs/ext3/namei.c +++ b/fs/ext3/namei.c @@ -991,19 +991,21 @@ static struct buffer_head * ext3_dx_find_entry(struct dentry *dentry, de = (struct ext3_dir_entry_2 *) bh->b_data; top = (struct ext3_dir_entry_2 *) ((char *) de + sb->s_blocksize - EXT3_DIR_REC_LEN(0)); - for (; de < top; de = ext3_next_entry(de)) - if (ext3_match (namelen, name, de)) { - if (!ext3_check_dir_entry("ext3_find_entry", - dir, de, bh, - (block<b_data))) { - brelse (bh); + for (; de < top; de = ext3_next_entry(de)) { + int off = (block << EXT3_BLOCK_SIZE_BITS(sb)) + + ((char *) de - bh->b_data); + + if (!ext3_check_dir_entry(__func__, dir, de, bh, off)) { + brelse(bh); *err = ERR_BAD_DX_DIR; goto errout; } - *res_dir = de; - dx_release (frames); - return bh; + + if (ext3_match(namelen, name, de)) { + *res_dir = de; + dx_release(frames); + return bh; + } } brelse (bh); /* Check to see if we should continue to search */ -- cgit v1.2.3-70-g09d2 From c0a1633b6201ef79e31b7da464d44fdf5953054d Mon Sep 17 00:00:00 2001 From: Adam Greenblatt Date: Fri, 25 Jul 2008 01:46:32 -0700 Subject: isofs: fix minor filesystem corruption Some iso9660 images contain files with rockridge data that is either incorrect or incompletely parsed. Prior to commit f2966632a134e865db3c819346a1dc7d96e05309 ("[PATCH] rock: handle directory overflows") (included with kernel 2.6.13) the kernel ignored the rockridge data for these files, while still allowing the files to be accessed under their non-rockridge names. That commit inadvertently changed things so that files with invalid rockridge data could not be accessed at all. (I ran across the problem when comparing some old CDs with hard disk copies I had made long ago under kernel 2.4: a few of the files on the hard disk copies were no longer visible on the CDs.) This change reverts to the pre-2.6.13 behavior. Signed-off-by: Adam Greenblatt Reviewed-by: Pekka Enberg Cc: [2.6.25.x, 2.6.26.x] Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/isofs/rock.c | 22 ++++++++++++++++++++-- 1 file changed, 20 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/isofs/rock.c b/fs/isofs/rock.c index 6bd48f0a704..c2fb2dd0131 100644 --- a/fs/isofs/rock.c +++ b/fs/isofs/rock.c @@ -209,6 +209,11 @@ repeat: while (rs.len > 2) { /* There may be one byte for padding somewhere */ rr = (struct rock_ridge *)rs.chr; + /* + * Ignore rock ridge info if rr->len is out of range, but + * don't return -EIO because that would make the file + * invisible. + */ if (rr->len < 3) goto out; /* Something got screwed up here */ sig = isonum_721(rs.chr); @@ -216,8 +221,12 @@ repeat: goto eio; rs.chr += rr->len; rs.len -= rr->len; + /* + * As above, just ignore the rock ridge info if rr->len + * is bogus. + */ if (rs.len < 0) - goto eio; /* corrupted isofs */ + goto out; /* Something got screwed up here */ switch (sig) { case SIG('R', 'R'): @@ -307,6 +316,11 @@ parse_rock_ridge_inode_internal(struct iso_directory_record *de, repeat: while (rs.len > 2) { /* There may be one byte for padding somewhere */ rr = (struct rock_ridge *)rs.chr; + /* + * Ignore rock ridge info if rr->len is out of range, but + * don't return -EIO because that would make the file + * invisible. + */ if (rr->len < 3) goto out; /* Something got screwed up here */ sig = isonum_721(rs.chr); @@ -314,8 +328,12 @@ repeat: goto eio; rs.chr += rr->len; rs.len -= rr->len; + /* + * As above, just ignore the rock ridge info if rr->len + * is bogus. + */ if (rs.len < 0) - goto eio; /* corrupted isofs */ + goto out; /* Something got screwed up here */ switch (sig) { #ifndef CONFIG_ZISOFS /* No flag for SF or ZF */ -- cgit v1.2.3-70-g09d2 From de0ca06a99c33df8333955642843331ab6b6e7ff Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Fri, 25 Jul 2008 01:46:34 -0700 Subject: coda: remove CODA_FS_OLD_API While fixing CONFIG_ leakages to the userspace kernel headers I ran into CODA_FS_OLD_API. After five years, are there still people using the old API left? Especially considering that you have to choose at compile time which API to support in the kernel (and distributions tend to offer the new API for some time). Jan: "The old API can definitely go. Around the time the new interface went in there were some non-Coda userspace file system implementations that took a while longer to convert to the new API, but by now they all switched to the new interface or in some cases to a FUSE-based solution." Signed-off-by: Adrian Bunk Acked-by: Jan Harkes Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/Kconfig | 14 -------------- fs/coda/coda_linux.c | 6 ++---- fs/coda/psdev.c | 4 ---- fs/coda/upcall.c | 15 +-------------- include/linux/coda.h | 43 ------------------------------------------- 5 files changed, 3 insertions(+), 79 deletions(-) (limited to 'fs') diff --git a/fs/Kconfig b/fs/Kconfig index 37db79a2ff9..ed563b9e352 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -2093,20 +2093,6 @@ config CODA_FS To compile the coda client support as a module, choose M here: the module will be called coda. -config CODA_FS_OLD_API - bool "Use 96-bit Coda file identifiers" - depends on CODA_FS - help - A new kernel-userspace API had to be introduced for Coda v6.0 - to support larger 128-bit file identifiers as needed by the - new realms implementation. - - However this new API is not backward compatible with older - clients. If you really need to run the old Coda userspace - cache manager then say Y. - - For most cases you probably want to say N. - config AFS_FS tristate "Andrew File System support (AFS) (EXPERIMENTAL)" depends on INET && EXPERIMENTAL diff --git a/fs/coda/coda_linux.c b/fs/coda/coda_linux.c index e1c854890f9..bf4a3fd3c8e 100644 --- a/fs/coda/coda_linux.c +++ b/fs/coda/coda_linux.c @@ -28,11 +28,9 @@ int coda_fake_statfs; char * coda_f2s(struct CodaFid *f) { static char s[60]; -#ifdef CONFIG_CODA_FS_OLD_API - sprintf(s, "(%08x.%08x.%08x)", f->opaque[0], f->opaque[1], f->opaque[2]); -#else + sprintf(s, "(%08x.%08x.%08x.%08x)", f->opaque[0], f->opaque[1], f->opaque[2], f->opaque[3]); -#endif + return s; } diff --git a/fs/coda/psdev.c b/fs/coda/psdev.c index 40c36f7352a..0d9b80ec689 100644 --- a/fs/coda/psdev.c +++ b/fs/coda/psdev.c @@ -378,11 +378,7 @@ MODULE_AUTHOR("Jan Harkes, Peter J. Braam"); MODULE_DESCRIPTION("Coda Distributed File System VFS interface"); MODULE_ALIAS_CHARDEV_MAJOR(CODA_PSDEV_MAJOR); MODULE_LICENSE("GPL"); -#ifdef CONFIG_CODA_FS_OLD_API -MODULE_VERSION("5.3.21"); -#else MODULE_VERSION("6.6"); -#endif static int __init init_coda(void) { diff --git a/fs/coda/upcall.c b/fs/coda/upcall.c index 359e531094d..ce432bca95d 100644 --- a/fs/coda/upcall.c +++ b/fs/coda/upcall.c @@ -52,12 +52,8 @@ static void *alloc_upcall(int opcode, int size) inp->ih.opcode = opcode; inp->ih.pid = current->pid; inp->ih.pgid = task_pgrp_nr(current); -#ifdef CONFIG_CODA_FS_OLD_API - memset(&inp->ih.cred, 0, sizeof(struct coda_cred)); - inp->ih.cred.cr_fsuid = current->fsuid; -#else inp->ih.uid = current->fsuid; -#endif + return (void*)inp; } @@ -166,20 +162,11 @@ int venus_close(struct super_block *sb, struct CodaFid *fid, int flags, union inputArgs *inp; union outputArgs *outp; int insize, outsize, error; -#ifdef CONFIG_CODA_FS_OLD_API - struct coda_cred cred = { 0, }; - cred.cr_fsuid = uid; -#endif insize = SIZE(release); UPARG(CODA_CLOSE); -#ifdef CONFIG_CODA_FS_OLD_API - memcpy(&(inp->ih.cred), &cred, sizeof(cred)); -#else inp->ih.uid = uid; -#endif - inp->coda_close.VFid = *fid; inp->coda_close.flags = flags; diff --git a/include/linux/coda.h b/include/linux/coda.h index b5cf0780c51..96c87693800 100644 --- a/include/linux/coda.h +++ b/include/linux/coda.h @@ -199,28 +199,6 @@ typedef u_int32_t vuid_t; typedef u_int32_t vgid_t; #endif /*_VUID_T_ */ -#ifdef CONFIG_CODA_FS_OLD_API -struct CodaFid { - u_int32_t opaque[3]; -}; - -static __inline__ ino_t coda_f2i(struct CodaFid *fid) -{ - if ( ! fid ) - return 0; - if (fid->opaque[1] == 0xfffffffe || fid->opaque[1] == 0xffffffff) - return ((fid->opaque[0] << 20) | (fid->opaque[2] & 0xfffff)); - else - return (fid->opaque[2] + (fid->opaque[1]<<10) + (fid->opaque[0]<<20)); -} - -struct coda_cred { - vuid_t cr_uid, cr_euid, cr_suid, cr_fsuid; /* Real, efftve, set, fs uid*/ - vgid_t cr_groupid, cr_egid, cr_sgid, cr_fsgid; /* same for groups */ -}; - -#else /* not defined(CONFIG_CODA_FS_OLD_API) */ - struct CodaFid { u_int32_t opaque[4]; }; @@ -228,8 +206,6 @@ struct CodaFid { #define coda_f2i(fid)\ (fid ? (fid->opaque[3] ^ (fid->opaque[2]<<10) ^ (fid->opaque[1]<<20) ^ fid->opaque[0]) : 0) -#endif - #ifndef _VENUS_VATTR_T_ #define _VENUS_VATTR_T_ /* @@ -313,15 +289,7 @@ struct coda_statfs { #define CIOC_KERNEL_VERSION _IOWR('c', 10, size_t) -#if 0 -#define CODA_KERNEL_VERSION 0 /* don't care about kernel version number */ -#define CODA_KERNEL_VERSION 1 /* The old venus 4.6 compatible interface */ -#endif -#ifdef CONFIG_CODA_FS_OLD_API -#define CODA_KERNEL_VERSION 2 /* venus_lookup got an extra parameter */ -#else #define CODA_KERNEL_VERSION 3 /* 128-bit file identifiers */ -#endif /* * Venus <-> Coda RPC arguments @@ -329,16 +297,9 @@ struct coda_statfs { struct coda_in_hdr { u_int32_t opcode; u_int32_t unique; /* Keep multiple outstanding msgs distinct */ -#ifdef CONFIG_CODA_FS_OLD_API - u_int16_t pid; /* Common to all */ - u_int16_t pgid; /* Common to all */ - u_int16_t sid; /* Common to all */ - struct coda_cred cred; /* Common to all */ -#else pid_t pid; pid_t pgid; vuid_t uid; -#endif }; /* Really important that opcode and unique are 1st two fields! */ @@ -613,11 +574,7 @@ struct coda_vget_out { /* CODA_PURGEUSER is a venus->kernel call */ struct coda_purgeuser_out { struct coda_out_hdr oh; -#ifdef CONFIG_CODA_FS_OLD_API - struct coda_cred cred; -#else vuid_t uid; -#endif }; /* coda_zapfile: */ -- cgit v1.2.3-70-g09d2 From 3084b72de73a6f8af0f16989ffb348068bd066d4 Mon Sep 17 00:00:00 2001 From: Matthias Kaehlcke Date: Fri, 25 Jul 2008 01:46:34 -0700 Subject: hfs: convert bitmap_lock in a mutex Apple Macintosh file system: The semaphore bitmap_lock is used as a mutex. Convert it to the mutex API Signed-off-by: Matthias Kaehlcke Cc: Roman Zippel Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/hfs/bitmap.c | 8 ++++---- fs/hfs/hfs_fs.h | 3 ++- fs/hfs/super.c | 2 +- 3 files changed, 7 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/hfs/bitmap.c b/fs/hfs/bitmap.c index 24e75798ddf..c6e97366e8a 100644 --- a/fs/hfs/bitmap.c +++ b/fs/hfs/bitmap.c @@ -145,7 +145,7 @@ u32 hfs_vbm_search_free(struct super_block *sb, u32 goal, u32 *num_bits) if (!*num_bits) return 0; - down(&HFS_SB(sb)->bitmap_lock); + mutex_lock(&HFS_SB(sb)->bitmap_lock); bitmap = HFS_SB(sb)->bitmap; pos = hfs_find_set_zero_bits(bitmap, HFS_SB(sb)->fs_ablocks, goal, num_bits); @@ -162,7 +162,7 @@ u32 hfs_vbm_search_free(struct super_block *sb, u32 goal, u32 *num_bits) HFS_SB(sb)->free_ablocks -= *num_bits; hfs_bitmap_dirty(sb); out: - up(&HFS_SB(sb)->bitmap_lock); + mutex_unlock(&HFS_SB(sb)->bitmap_lock); return pos; } @@ -205,7 +205,7 @@ int hfs_clear_vbm_bits(struct super_block *sb, u16 start, u16 count) if ((start + count) > HFS_SB(sb)->fs_ablocks) return -2; - down(&HFS_SB(sb)->bitmap_lock); + mutex_lock(&HFS_SB(sb)->bitmap_lock); /* bitmap is always on a 32-bit boundary */ curr = HFS_SB(sb)->bitmap + (start / 32); len = count; @@ -236,7 +236,7 @@ int hfs_clear_vbm_bits(struct super_block *sb, u16 start, u16 count) } out: HFS_SB(sb)->free_ablocks += len; - up(&HFS_SB(sb)->bitmap_lock); + mutex_unlock(&HFS_SB(sb)->bitmap_lock); hfs_bitmap_dirty(sb); return 0; diff --git a/fs/hfs/hfs_fs.h b/fs/hfs/hfs_fs.h index 147374b6f67..ad652881911 100644 --- a/fs/hfs/hfs_fs.h +++ b/fs/hfs/hfs_fs.h @@ -11,6 +11,7 @@ #include #include +#include #include #include @@ -139,7 +140,7 @@ struct hfs_sb_info { struct nls_table *nls_io, *nls_disk; - struct semaphore bitmap_lock; + struct mutex bitmap_lock; unsigned long flags; diff --git a/fs/hfs/super.c b/fs/hfs/super.c index 8cf67974adf..ac2ec5ef66e 100644 --- a/fs/hfs/super.c +++ b/fs/hfs/super.c @@ -372,7 +372,7 @@ static int hfs_fill_super(struct super_block *sb, void *data, int silent) sb->s_op = &hfs_super_operations; sb->s_flags |= MS_NODIRATIME; - init_MUTEX(&sbi->bitmap_lock); + mutex_init(&sbi->bitmap_lock); res = hfs_mdb_get(sb); if (res) { -- cgit v1.2.3-70-g09d2 From 39f8d472f280dee6503a364d1d911b9e20ce3ec9 Mon Sep 17 00:00:00 2001 From: Matthias Kaehlcke Date: Fri, 25 Jul 2008 01:46:35 -0700 Subject: hfs: convert extents_lock in a mutex Apple Macintosh file system: The semaphore extens_lock is used as a mutex. Convert it to the mutex API Signed-off-by: Matthias Kaehlcke Cc: Roman Zippel Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/hfs/btree.c | 2 +- fs/hfs/extent.c | 14 +++++++------- fs/hfs/hfs_fs.h | 2 +- fs/hfs/inode.c | 4 ++-- 4 files changed, 11 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/hfs/btree.c b/fs/hfs/btree.c index f6621a78520..9b9d6395bad 100644 --- a/fs/hfs/btree.c +++ b/fs/hfs/btree.c @@ -40,7 +40,7 @@ struct hfs_btree *hfs_btree_open(struct super_block *sb, u32 id, btree_keycmp ke { struct hfs_mdb *mdb = HFS_SB(sb)->mdb; HFS_I(tree->inode)->flags = 0; - init_MUTEX(&HFS_I(tree->inode)->extents_lock); + mutex_init(&HFS_I(tree->inode)->extents_lock); switch (id) { case HFS_EXT_CNID: hfs_inode_read_fork(tree->inode, mdb->drXTExtRec, mdb->drXTFlSize, diff --git a/fs/hfs/extent.c b/fs/hfs/extent.c index c176f67ba0a..2c16316d291 100644 --- a/fs/hfs/extent.c +++ b/fs/hfs/extent.c @@ -343,16 +343,16 @@ int hfs_get_block(struct inode *inode, sector_t block, goto done; } - down(&HFS_I(inode)->extents_lock); + mutex_lock(&HFS_I(inode)->extents_lock); res = hfs_ext_read_extent(inode, ablock); if (!res) dblock = hfs_ext_find_block(HFS_I(inode)->cached_extents, ablock - HFS_I(inode)->cached_start); else { - up(&HFS_I(inode)->extents_lock); + mutex_unlock(&HFS_I(inode)->extents_lock); return -EIO; } - up(&HFS_I(inode)->extents_lock); + mutex_unlock(&HFS_I(inode)->extents_lock); done: map_bh(bh_result, sb, HFS_SB(sb)->fs_start + @@ -375,7 +375,7 @@ int hfs_extend_file(struct inode *inode) u32 start, len, goal; int res; - down(&HFS_I(inode)->extents_lock); + mutex_lock(&HFS_I(inode)->extents_lock); if (HFS_I(inode)->alloc_blocks == HFS_I(inode)->first_blocks) goal = hfs_ext_lastblock(HFS_I(inode)->first_extents); else { @@ -425,7 +425,7 @@ int hfs_extend_file(struct inode *inode) goto insert_extent; } out: - up(&HFS_I(inode)->extents_lock); + mutex_unlock(&HFS_I(inode)->extents_lock); if (!res) { HFS_I(inode)->alloc_blocks += len; mark_inode_dirty(inode); @@ -487,7 +487,7 @@ void hfs_file_truncate(struct inode *inode) if (blk_cnt == alloc_cnt) goto out; - down(&HFS_I(inode)->extents_lock); + mutex_lock(&HFS_I(inode)->extents_lock); hfs_find_init(HFS_SB(sb)->ext_tree, &fd); while (1) { if (alloc_cnt == HFS_I(inode)->first_blocks) { @@ -514,7 +514,7 @@ void hfs_file_truncate(struct inode *inode) hfs_brec_remove(&fd); } hfs_find_exit(&fd); - up(&HFS_I(inode)->extents_lock); + mutex_unlock(&HFS_I(inode)->extents_lock); HFS_I(inode)->alloc_blocks = blk_cnt; out: diff --git a/fs/hfs/hfs_fs.h b/fs/hfs/hfs_fs.h index ad652881911..9955232fdf8 100644 --- a/fs/hfs/hfs_fs.h +++ b/fs/hfs/hfs_fs.h @@ -54,7 +54,7 @@ struct hfs_inode_info { struct list_head open_dir_list; struct inode *rsrc_inode; - struct semaphore extents_lock; + struct mutex extents_lock; u16 alloc_blocks, clump_blocks; sector_t fs_blocks; diff --git a/fs/hfs/inode.c b/fs/hfs/inode.c index 97f8446c4ff..dc4ec640e87 100644 --- a/fs/hfs/inode.c +++ b/fs/hfs/inode.c @@ -150,7 +150,7 @@ struct inode *hfs_new_inode(struct inode *dir, struct qstr *name, int mode) if (!inode) return NULL; - init_MUTEX(&HFS_I(inode)->extents_lock); + mutex_init(&HFS_I(inode)->extents_lock); INIT_LIST_HEAD(&HFS_I(inode)->open_dir_list); hfs_cat_build_key(sb, (btree_key *)&HFS_I(inode)->cat_key, dir->i_ino, name); inode->i_ino = HFS_SB(sb)->next_id++; @@ -281,7 +281,7 @@ static int hfs_read_inode(struct inode *inode, void *data) HFS_I(inode)->flags = 0; HFS_I(inode)->rsrc_inode = NULL; - init_MUTEX(&HFS_I(inode)->extents_lock); + mutex_init(&HFS_I(inode)->extents_lock); INIT_LIST_HEAD(&HFS_I(inode)->open_dir_list); /* Initialize the inode */ -- cgit v1.2.3-70-g09d2 From 895c23f8c39c0c8d7b536bb2566d4aa968d78be2 Mon Sep 17 00:00:00 2001 From: Matthias Kaehlcke Date: Fri, 25 Jul 2008 01:46:36 -0700 Subject: hfsplus: convert the extents_lock in a mutex Apple Extended HFS file system: The semaphore extents lock is used as a mutex. Convert it to the mutex API. Signed-off-by: Matthias Kaehlcke Cc: Roman Zippel Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/hfsplus/extents.c | 14 +++++++------- fs/hfsplus/hfsplus_fs.h | 3 ++- fs/hfsplus/inode.c | 4 ++-- fs/hfsplus/super.c | 2 +- 4 files changed, 12 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/hfsplus/extents.c b/fs/hfsplus/extents.c index 12e899cd788..fec8f61227f 100644 --- a/fs/hfsplus/extents.c +++ b/fs/hfsplus/extents.c @@ -199,16 +199,16 @@ int hfsplus_get_block(struct inode *inode, sector_t iblock, goto done; } - down(&HFSPLUS_I(inode).extents_lock); + mutex_lock(&HFSPLUS_I(inode).extents_lock); res = hfsplus_ext_read_extent(inode, ablock); if (!res) { dblock = hfsplus_ext_find_block(HFSPLUS_I(inode).cached_extents, ablock - HFSPLUS_I(inode).cached_start); } else { - up(&HFSPLUS_I(inode).extents_lock); + mutex_unlock(&HFSPLUS_I(inode).extents_lock); return -EIO; } - up(&HFSPLUS_I(inode).extents_lock); + mutex_unlock(&HFSPLUS_I(inode).extents_lock); done: dprint(DBG_EXTENT, "get_block(%lu): %llu - %u\n", inode->i_ino, (long long)iblock, dblock); @@ -355,7 +355,7 @@ int hfsplus_file_extend(struct inode *inode) return -ENOSPC; } - down(&HFSPLUS_I(inode).extents_lock); + mutex_lock(&HFSPLUS_I(inode).extents_lock); if (HFSPLUS_I(inode).alloc_blocks == HFSPLUS_I(inode).first_blocks) goal = hfsplus_ext_lastblock(HFSPLUS_I(inode).first_extents); else { @@ -408,7 +408,7 @@ int hfsplus_file_extend(struct inode *inode) goto insert_extent; } out: - up(&HFSPLUS_I(inode).extents_lock); + mutex_unlock(&HFSPLUS_I(inode).extents_lock); if (!res) { HFSPLUS_I(inode).alloc_blocks += len; mark_inode_dirty(inode); @@ -465,7 +465,7 @@ void hfsplus_file_truncate(struct inode *inode) if (blk_cnt == alloc_cnt) goto out; - down(&HFSPLUS_I(inode).extents_lock); + mutex_lock(&HFSPLUS_I(inode).extents_lock); hfs_find_init(HFSPLUS_SB(sb).ext_tree, &fd); while (1) { if (alloc_cnt == HFSPLUS_I(inode).first_blocks) { @@ -492,7 +492,7 @@ void hfsplus_file_truncate(struct inode *inode) hfs_brec_remove(&fd); } hfs_find_exit(&fd); - up(&HFSPLUS_I(inode).extents_lock); + mutex_unlock(&HFSPLUS_I(inode).extents_lock); HFSPLUS_I(inode).alloc_blocks = blk_cnt; out: diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h index 9e59537b43d..f027a905225 100644 --- a/fs/hfsplus/hfsplus_fs.h +++ b/fs/hfsplus/hfsplus_fs.h @@ -11,6 +11,7 @@ #define _LINUX_HFSPLUS_FS_H #include +#include #include #include "hfsplus_raw.h" @@ -154,7 +155,7 @@ struct hfsplus_sb_info { struct hfsplus_inode_info { - struct semaphore extents_lock; + struct mutex extents_lock; u32 clump_blocks, alloc_blocks; sector_t fs_blocks; /* Allocation extents from catalog record or volume header */ diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c index 67e1c8b467c..cc3b5e24339 100644 --- a/fs/hfsplus/inode.c +++ b/fs/hfsplus/inode.c @@ -163,7 +163,7 @@ static struct dentry *hfsplus_file_lookup(struct inode *dir, struct dentry *dent inode->i_ino = dir->i_ino; INIT_LIST_HEAD(&HFSPLUS_I(inode).open_dir_list); - init_MUTEX(&HFSPLUS_I(inode).extents_lock); + mutex_init(&HFSPLUS_I(inode).extents_lock); HFSPLUS_I(inode).flags = HFSPLUS_FLG_RSRC; hfs_find_init(HFSPLUS_SB(sb).cat_tree, &fd); @@ -316,7 +316,7 @@ struct inode *hfsplus_new_inode(struct super_block *sb, int mode) inode->i_nlink = 1; inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME_SEC; INIT_LIST_HEAD(&HFSPLUS_I(inode).open_dir_list); - init_MUTEX(&HFSPLUS_I(inode).extents_lock); + mutex_init(&HFSPLUS_I(inode).extents_lock); atomic_set(&HFSPLUS_I(inode).opencnt, 0); HFSPLUS_I(inode).flags = 0; memset(HFSPLUS_I(inode).first_extents, 0, sizeof(hfsplus_extent_rec)); diff --git a/fs/hfsplus/super.c b/fs/hfsplus/super.c index ce97a54518d..3859118531c 100644 --- a/fs/hfsplus/super.c +++ b/fs/hfsplus/super.c @@ -34,7 +34,7 @@ struct inode *hfsplus_iget(struct super_block *sb, unsigned long ino) return inode; INIT_LIST_HEAD(&HFSPLUS_I(inode).open_dir_list); - init_MUTEX(&HFSPLUS_I(inode).extents_lock); + mutex_init(&HFSPLUS_I(inode).extents_lock); HFSPLUS_I(inode).flags = 0; HFSPLUS_I(inode).rsrc_inode = NULL; atomic_set(&HFSPLUS_I(inode).opencnt, 0); -- cgit v1.2.3-70-g09d2 From 5d4f7fddf8882b214e4aabb3bdb37f90a72b2537 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Fri, 25 Jul 2008 01:46:36 -0700 Subject: reiserfs: fix synchronization of quota files in journal=data mode In journal=data mode, it is not enough to do write_inode_now() as done in vfs_quota_on() to write all data to their final location (which is needed for quota_read to work correctly). Calling journal_end_sync() before calling vfs_quota_on() does it's job because transactions are committed to the journal and data marked as dirty in memory so write_inode_now() writes them to their final locations. Cc: Signed-off-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/reiserfs/super.c | 30 +++++++++++++++++++++--------- 1 file changed, 21 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c index 1d40f2bd197..0cbf4cd1114 100644 --- a/fs/reiserfs/super.c +++ b/fs/reiserfs/super.c @@ -2026,6 +2026,7 @@ static int reiserfs_quota_on(struct super_block *sb, int type, int format_id, int err; struct nameidata nd; struct inode *inode; + struct reiserfs_transaction_handle th; if (!(REISERFS_SB(sb)->s_mount_opt & (1 << REISERFS_QUOTA))) return -EINVAL; @@ -2053,17 +2054,28 @@ static int reiserfs_quota_on(struct super_block *sb, int type, int format_id, } mark_inode_dirty(inode); } - /* Not journalling quota? No more tests needed... */ - if (!REISERFS_SB(sb)->s_qf_names[USRQUOTA] && - !REISERFS_SB(sb)->s_qf_names[GRPQUOTA]) { - path_put(&nd.path); - return vfs_quota_on(sb, type, format_id, path, 0); - } - /* Quotafile not of fs root? */ - if (nd.path.dentry->d_parent->d_inode != sb->s_root->d_inode) - reiserfs_warning(sb, + /* Journaling quota? */ + if (REISERFS_SB(sb)->s_qf_names[type]) { + /* Quotafile not of fs root? */ + if (nd.path.dentry->d_parent->d_inode != sb->s_root->d_inode) + reiserfs_warning(sb, "reiserfs: Quota file not on filesystem root. " "Journalled quota will not work."); + } + + /* + * When we journal data on quota file, we have to flush journal to see + * all updates to the file when we bypass pagecache... + */ + if (reiserfs_file_data_log(inode)) { + /* Just start temporary transaction and finish it */ + err = journal_begin(&th, sb, 1); + if (err) + return err; + err = journal_end_sync(&th, sb, 1); + if (err) + return err; + } path_put(&nd.path); return vfs_quota_on(sb, type, format_id, path, 0); } -- cgit v1.2.3-70-g09d2 From 4506567b24d3ea707e46e8aad64caef539382f4b Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Fri, 25 Jul 2008 01:46:37 -0700 Subject: reiserfs: fix typos in messages and comments (journalled -> journaled) Cc: Signed-off-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/reiserfs/super.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c index 0cbf4cd1114..f723604c5d9 100644 --- a/fs/reiserfs/super.c +++ b/fs/reiserfs/super.c @@ -182,7 +182,7 @@ static int finish_unfinished(struct super_block *s) int ret = reiserfs_quota_on_mount(s, i); if (ret < 0) reiserfs_warning(s, - "reiserfs: cannot turn on journalled quota: error %d", + "reiserfs: cannot turn on journaled quota: error %d", ret); } } @@ -994,7 +994,7 @@ static int reiserfs_parse_options(struct super_block *s, char *options, /* strin if (sb_any_quota_enabled(s)) { reiserfs_warning(s, - "reiserfs_parse_options: cannot change journalled quota options when quota turned on."); + "reiserfs_parse_options: cannot change journaled quota options when quota turned on."); return 0; } if (*arg) { /* Some filename specified? */ @@ -1039,7 +1039,7 @@ static int reiserfs_parse_options(struct super_block *s, char *options, /* strin #else if (c == 'u' || c == 'g' || c == 'f') { reiserfs_warning(s, - "reiserfs_parse_options: journalled quota options not supported."); + "reiserfs_parse_options: journaled quota options not supported."); return 0; } #endif @@ -1050,7 +1050,7 @@ static int reiserfs_parse_options(struct super_block *s, char *options, /* strin && (REISERFS_SB(s)->s_qf_names[USRQUOTA] || REISERFS_SB(s)->s_qf_names[GRPQUOTA])) { reiserfs_warning(s, - "reiserfs_parse_options: journalled quota format not specified."); + "reiserfs_parse_options: journaled quota format not specified."); return 0; } /* This checking is not precise wrt the quota type but for our purposes it is sufficient */ @@ -1980,7 +1980,7 @@ static int reiserfs_release_dquot(struct dquot *dquot) static int reiserfs_mark_dquot_dirty(struct dquot *dquot) { - /* Are we journalling quotas? */ + /* Are we journaling quotas? */ if (REISERFS_SB(dquot->dq_sb)->s_qf_names[USRQUOTA] || REISERFS_SB(dquot->dq_sb)->s_qf_names[GRPQUOTA]) { dquot_mark_dquot_dirty(dquot); -- cgit v1.2.3-70-g09d2 From 00b441970a0ab48185244300ac7d4e4eb76df692 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Fri, 25 Jul 2008 01:46:38 -0700 Subject: reiserfs: correct mount option parsing to detect when quota options can be changed We should not allow user to change quota mount options when quota is just suspended. It would make mount options and internal quota state inconsistent. Also we should not allow user to change quota format when quota is turned on. On the other hand we can just silently ignore when some option is set to the value it already has (some mount versions do this on remount). Finally, we should not discard current quota options if parsing of mount options fails. Cc: Signed-off-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/reiserfs/super.c | 83 ++++++++++++++++++++++++++++++++++++++--------------- 1 file changed, 60 insertions(+), 23 deletions(-) (limited to 'fs') diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c index f723604c5d9..a10a6d2a887 100644 --- a/fs/reiserfs/super.c +++ b/fs/reiserfs/super.c @@ -876,7 +876,9 @@ static int reiserfs_parse_options(struct super_block *s, char *options, /* strin mount options were selected. */ unsigned long *blocks, /* strtol-ed from NNN of resize=NNN */ char **jdev_name, - unsigned int *commit_max_age) + unsigned int *commit_max_age, + char **qf_names, + unsigned int *qfmt) { int c; char *arg = NULL; @@ -992,7 +994,9 @@ static int reiserfs_parse_options(struct super_block *s, char *options, /* strin if (c == 'u' || c == 'g') { int qtype = c == 'u' ? USRQUOTA : GRPQUOTA; - if (sb_any_quota_enabled(s)) { + if ((sb_any_quota_enabled(s) || + sb_any_quota_suspended(s)) && + (!*arg != !REISERFS_SB(s)->s_qf_names[qtype])) { reiserfs_warning(s, "reiserfs_parse_options: cannot change journaled quota options when quota turned on."); return 0; @@ -1011,30 +1015,39 @@ static int reiserfs_parse_options(struct super_block *s, char *options, /* strin "reiserfs_parse_options: quotafile must be on filesystem root."); return 0; } - REISERFS_SB(s)->s_qf_names[qtype] = + qf_names[qtype] = kmalloc(strlen(arg) + 1, GFP_KERNEL); - if (!REISERFS_SB(s)->s_qf_names[qtype]) { + if (!qf_names[qtype]) { reiserfs_warning(s, "reiserfs_parse_options: not enough memory for storing quotafile name."); return 0; } - strcpy(REISERFS_SB(s)->s_qf_names[qtype], arg); + strcpy(qf_names[qtype], arg); *mount_options |= 1 << REISERFS_QUOTA; } else { - kfree(REISERFS_SB(s)->s_qf_names[qtype]); - REISERFS_SB(s)->s_qf_names[qtype] = NULL; + if (qf_names[qtype] != + REISERFS_SB(s)->s_qf_names[qtype]) + kfree(qf_names[qtype]); + qf_names[qtype] = NULL; } } if (c == 'f') { if (!strcmp(arg, "vfsold")) - REISERFS_SB(s)->s_jquota_fmt = QFMT_VFS_OLD; + *qfmt = QFMT_VFS_OLD; else if (!strcmp(arg, "vfsv0")) - REISERFS_SB(s)->s_jquota_fmt = QFMT_VFS_V0; + *qfmt = QFMT_VFS_V0; else { reiserfs_warning(s, "reiserfs_parse_options: unknown quota format specified."); return 0; } + if ((sb_any_quota_enabled(s) || + sb_any_quota_suspended(s)) && + *qfmt != REISERFS_SB(s)->s_jquota_fmt) { + reiserfs_warning(s, + "reiserfs_parse_options: cannot change journaled quota options when quota turned on."); + return 0; + } } #else if (c == 'u' || c == 'g' || c == 'f') { @@ -1046,9 +1059,8 @@ static int reiserfs_parse_options(struct super_block *s, char *options, /* strin } #ifdef CONFIG_QUOTA - if (!REISERFS_SB(s)->s_jquota_fmt - && (REISERFS_SB(s)->s_qf_names[USRQUOTA] - || REISERFS_SB(s)->s_qf_names[GRPQUOTA])) { + if (!REISERFS_SB(s)->s_jquota_fmt && !*qfmt + && (qf_names[USRQUOTA] || qf_names[GRPQUOTA])) { reiserfs_warning(s, "reiserfs_parse_options: journaled quota format not specified."); return 0; @@ -1130,6 +1142,21 @@ static void handle_attrs(struct super_block *s) } } +#ifdef CONFIG_QUOTA +static void handle_quota_files(struct super_block *s, char **qf_names, + unsigned int *qfmt) +{ + int i; + + for (i = 0; i < MAXQUOTAS; i++) { + if (qf_names[i] != REISERFS_SB(s)->s_qf_names[i]) + kfree(REISERFS_SB(s)->s_qf_names[i]); + REISERFS_SB(s)->s_qf_names[i] = qf_names[i]; + } + REISERFS_SB(s)->s_jquota_fmt = *qfmt; +} +#endif + static int reiserfs_remount(struct super_block *s, int *mount_flags, char *arg) { struct reiserfs_super_block *rs; @@ -1141,23 +1168,30 @@ static int reiserfs_remount(struct super_block *s, int *mount_flags, char *arg) struct reiserfs_journal *journal = SB_JOURNAL(s); char *new_opts = kstrdup(arg, GFP_KERNEL); int err; + char *qf_names[MAXQUOTAS]; + unsigned int qfmt = 0; #ifdef CONFIG_QUOTA int i; + + memcpy(qf_names, REISERFS_SB(s)->s_qf_names, sizeof(qf_names)); #endif rs = SB_DISK_SUPER_BLOCK(s); if (!reiserfs_parse_options - (s, arg, &mount_options, &blocks, NULL, &commit_max_age)) { + (s, arg, &mount_options, &blocks, NULL, &commit_max_age, + qf_names, &qfmt)) { #ifdef CONFIG_QUOTA - for (i = 0; i < MAXQUOTAS; i++) { - kfree(REISERFS_SB(s)->s_qf_names[i]); - REISERFS_SB(s)->s_qf_names[i] = NULL; - } + for (i = 0; i < MAXQUOTAS; i++) + if (qf_names[i] != REISERFS_SB(s)->s_qf_names[i]) + kfree(qf_names[i]); #endif err = -EINVAL; goto out_err; } +#ifdef CONFIG_QUOTA + handle_quota_files(s, qf_names, &qfmt); +#endif handle_attrs(s); @@ -1570,6 +1604,8 @@ static int reiserfs_fill_super(struct super_block *s, void *data, int silent) char *jdev_name; struct reiserfs_sb_info *sbi; int errval = -EINVAL; + char *qf_names[MAXQUOTAS] = {}; + unsigned int qfmt = 0; save_mount_options(s, data); @@ -1597,9 +1633,12 @@ static int reiserfs_fill_super(struct super_block *s, void *data, int silent) jdev_name = NULL; if (reiserfs_parse_options (s, (char *)data, &(sbi->s_mount_opt), &blocks, &jdev_name, - &commit_max_age) == 0) { + &commit_max_age, qf_names, &qfmt) == 0) { goto error; } +#ifdef CONFIG_QUOTA + handle_quota_files(s, qf_names, &qfmt); +#endif if (blocks) { SWARN(silent, s, "jmacd-7: reiserfs_fill_super: resize option " @@ -1819,7 +1858,7 @@ static int reiserfs_fill_super(struct super_block *s, void *data, int silent) return (0); - error: +error: if (jinit_done) { /* kill the commit thread, free journal ram */ journal_release_error(NULL, s); } @@ -1830,10 +1869,8 @@ static int reiserfs_fill_super(struct super_block *s, void *data, int silent) #ifdef CONFIG_QUOTA { int j; - for (j = 0; j < MAXQUOTAS; j++) { - kfree(sbi->s_qf_names[j]); - sbi->s_qf_names[j] = NULL; - } + for (j = 0; j < MAXQUOTAS; j++) + kfree(qf_names[j]); } #endif kfree(sbi); -- cgit v1.2.3-70-g09d2 From f68215c4640a38d66429014e524a627bf572d26a Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Fri, 25 Jul 2008 01:46:38 -0700 Subject: reiserfs: convert j_lock to mutex j_lock is a semaphore but uses it as if it were a mutex. This patch converts it to a mutex. Signed-off-by: Jeff Mahoney Cc: Matthew Wilcox Cc: Chris Mason Cc: Edward Shishkin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/reiserfs/journal.c | 6 +++--- include/linux/reiserfs_fs_sb.h | 2 +- 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/reiserfs/journal.c b/fs/reiserfs/journal.c index e396b2fa474..0f7b1e807e6 100644 --- a/fs/reiserfs/journal.c +++ b/fs/reiserfs/journal.c @@ -558,13 +558,13 @@ static inline void insert_journal_hash(struct reiserfs_journal_cnode **table, static inline void lock_journal(struct super_block *p_s_sb) { PROC_INFO_INC(p_s_sb, journal.lock_journal); - down(&SB_JOURNAL(p_s_sb)->j_lock); + mutex_lock(&SB_JOURNAL(p_s_sb)->j_mutex); } /* unlock the current transaction */ static inline void unlock_journal(struct super_block *p_s_sb) { - up(&SB_JOURNAL(p_s_sb)->j_lock); + mutex_unlock(&SB_JOURNAL(p_s_sb)->j_mutex); } static inline void get_journal_list(struct reiserfs_journal_list *jl) @@ -2837,7 +2837,7 @@ int journal_init(struct super_block *p_s_sb, const char *j_dev_name, journal->j_last = NULL; journal->j_first = NULL; init_waitqueue_head(&(journal->j_join_wait)); - sema_init(&journal->j_lock, 1); + mutex_init(&journal->j_mutex); sema_init(&journal->j_flush_sem, 1); journal->j_trans_id = 10; diff --git a/include/linux/reiserfs_fs_sb.h b/include/linux/reiserfs_fs_sb.h index 336ee43ed7d..49b639b88ba 100644 --- a/include/linux/reiserfs_fs_sb.h +++ b/include/linux/reiserfs_fs_sb.h @@ -193,7 +193,7 @@ struct reiserfs_journal { struct buffer_head *j_header_bh; time_t j_trans_start_time; /* time this transaction started */ - struct semaphore j_lock; + struct mutex j_mutex; struct semaphore j_flush_sem; wait_queue_head_t j_join_wait; /* wait for current transaction to finish before starting new one */ atomic_t j_jlock; /* lock for j_join_wait */ -- cgit v1.2.3-70-g09d2 From afe70259076fff0446001eaa1a287f615241a357 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Fri, 25 Jul 2008 01:46:39 -0700 Subject: reiserfs: convert j_flush_sem to mutex j_flush_sem is a semaphore but uses it as if it were a mutex. This patch converts it to a mutex. [akpm@linux-foundation.org: fix mutex_trylock retval treatment] Signed-off-by: Jeff Mahoney Cc: Matthew Wilcox Cc: Chris Mason Cc: Edward Shishkin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/reiserfs/journal.c | 14 +++++++------- include/linux/reiserfs_fs_sb.h | 2 +- 2 files changed, 8 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/reiserfs/journal.c b/fs/reiserfs/journal.c index 0f7b1e807e6..3cb4a562030 100644 --- a/fs/reiserfs/journal.c +++ b/fs/reiserfs/journal.c @@ -1411,8 +1411,8 @@ static int flush_journal_list(struct super_block *s, /* if flushall == 0, the lock is already held */ if (flushall) { - down(&journal->j_flush_sem); - } else if (!down_trylock(&journal->j_flush_sem)) { + mutex_lock(&journal->j_flush_mutex); + } else if (mutex_trylock(&journal->j_flush_mutex)) { BUG(); } @@ -1642,7 +1642,7 @@ static int flush_journal_list(struct super_block *s, jl->j_state = 0; put_journal_list(s, jl); if (flushall) - up(&journal->j_flush_sem); + mutex_unlock(&journal->j_flush_mutex); put_fs_excl(); return err; } @@ -1772,12 +1772,12 @@ static int kupdate_transactions(struct super_block *s, struct reiserfs_journal *journal = SB_JOURNAL(s); chunk.nr = 0; - down(&journal->j_flush_sem); + mutex_lock(&journal->j_flush_mutex); if (!journal_list_still_alive(s, orig_trans_id)) { goto done; } - /* we've got j_flush_sem held, nobody is going to delete any + /* we've got j_flush_mutex held, nobody is going to delete any * of these lists out from underneath us */ while ((num_trans && transactions_flushed < num_trans) || @@ -1812,7 +1812,7 @@ static int kupdate_transactions(struct super_block *s, } done: - up(&journal->j_flush_sem); + mutex_unlock(&journal->j_flush_mutex); return ret; } @@ -2838,7 +2838,7 @@ int journal_init(struct super_block *p_s_sb, const char *j_dev_name, journal->j_first = NULL; init_waitqueue_head(&(journal->j_join_wait)); mutex_init(&journal->j_mutex); - sema_init(&journal->j_flush_sem, 1); + mutex_init(&journal->j_flush_mutex); journal->j_trans_id = 10; journal->j_mount_id = 10; diff --git a/include/linux/reiserfs_fs_sb.h b/include/linux/reiserfs_fs_sb.h index 49b639b88ba..c0751724ee6 100644 --- a/include/linux/reiserfs_fs_sb.h +++ b/include/linux/reiserfs_fs_sb.h @@ -194,7 +194,7 @@ struct reiserfs_journal { time_t j_trans_start_time; /* time this transaction started */ struct mutex j_mutex; - struct semaphore j_flush_sem; + struct mutex j_flush_mutex; wait_queue_head_t j_join_wait; /* wait for current transaction to finish before starting new one */ atomic_t j_jlock; /* lock for j_join_wait */ int j_list_bitmap_index; /* number of next list bitmap to use */ -- cgit v1.2.3-70-g09d2 From 90415deac75a761a25239af6f56381546f8d2201 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Fri, 25 Jul 2008 01:46:40 -0700 Subject: reiserfs: convert j_commit_lock to mutex j_commit_lock is a semaphore but uses it as if it were a mutex. This patch converts it to a mutex. [akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Jeff Mahoney Cc: Matthew Wilcox Cc: Chris Mason Cc: Edward Shishkin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/reiserfs/journal.c | 22 ++++++++++------------ include/linux/reiserfs_fs_sb.h | 2 +- 2 files changed, 11 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/reiserfs/journal.c b/fs/reiserfs/journal.c index 3cb4a562030..c8f60ee183b 100644 --- a/fs/reiserfs/journal.c +++ b/fs/reiserfs/journal.c @@ -34,15 +34,10 @@ ** from within kupdate, it will ignore the immediate flag */ -#include -#include - #include #include - #include #include - #include #include #include @@ -54,6 +49,9 @@ #include #include #include +#include + +#include /* gets a struct reiserfs_journal_list * from a list head */ #define JOURNAL_LIST_ENTRY(h) (list_entry((h), struct reiserfs_journal_list, \ @@ -1045,9 +1043,9 @@ static int flush_commit_list(struct super_block *s, } /* make sure nobody is trying to flush this one at the same time */ - down(&jl->j_commit_lock); + mutex_lock(&jl->j_commit_mutex); if (!journal_list_still_alive(s, trans_id)) { - up(&jl->j_commit_lock); + mutex_unlock(&jl->j_commit_mutex); goto put_jl; } BUG_ON(jl->j_trans_id == 0); @@ -1057,7 +1055,7 @@ static int flush_commit_list(struct super_block *s, if (flushall) { atomic_set(&(jl->j_older_commits_done), 1); } - up(&jl->j_commit_lock); + mutex_unlock(&jl->j_commit_mutex); goto put_jl; } @@ -1181,7 +1179,7 @@ static int flush_commit_list(struct super_block *s, if (flushall) { atomic_set(&(jl->j_older_commits_done), 1); } - up(&jl->j_commit_lock); + mutex_unlock(&jl->j_commit_mutex); put_jl: put_journal_list(s, jl); @@ -2556,7 +2554,7 @@ static struct reiserfs_journal_list *alloc_journal_list(struct super_block *s) INIT_LIST_HEAD(&jl->j_working_list); INIT_LIST_HEAD(&jl->j_tail_bh_list); INIT_LIST_HEAD(&jl->j_bh_list); - sema_init(&jl->j_commit_lock, 1); + mutex_init(&jl->j_commit_mutex); SB_JOURNAL(s)->j_num_lists++; get_journal_list(jl); return jl; @@ -4030,7 +4028,7 @@ static int do_journal_end(struct reiserfs_transaction_handle *th, * the new transaction is fully setup, and we've already flushed the * ordered bh list */ - down(&jl->j_commit_lock); + mutex_lock(&jl->j_commit_mutex); /* save the transaction id in case we need to commit it later */ commit_trans_id = jl->j_trans_id; @@ -4196,7 +4194,7 @@ static int do_journal_end(struct reiserfs_transaction_handle *th, lock_kernel(); } BUG_ON(!list_empty(&jl->j_tail_bh_list)); - up(&jl->j_commit_lock); + mutex_unlock(&jl->j_commit_mutex); /* honor the flush wishes from the caller, simple commits can ** be done outside the journal lock, they are done below diff --git a/include/linux/reiserfs_fs_sb.h b/include/linux/reiserfs_fs_sb.h index c0751724ee6..315517e8bfa 100644 --- a/include/linux/reiserfs_fs_sb.h +++ b/include/linux/reiserfs_fs_sb.h @@ -152,7 +152,7 @@ struct reiserfs_journal_list { atomic_t j_nonzerolen; atomic_t j_commit_left; atomic_t j_older_commits_done; /* all commits older than this on disk */ - struct semaphore j_commit_lock; + struct mutex j_commit_mutex; unsigned long j_trans_id; time_t j_timestamp; struct reiserfs_list_bitmap *j_list_bitmap; -- cgit v1.2.3-70-g09d2 From 3264d4ded4d916d294d776b77b72d477c63ac3be Mon Sep 17 00:00:00 2001 From: Shen Feng Date: Fri, 25 Jul 2008 01:46:41 -0700 Subject: reiserfs: remove double definitions of xattr macros remove the definitions of macros: XATTR_SECURITY_PREFIX XATTR_TRUSTED_PREFIX XATTR_USER_PREFIX since they are defined in linux/xattr.h Signed-off-by: Shen Feng Signed-off-by: Mingming Cao Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/reiserfs/xattr_security.c | 2 -- fs/reiserfs/xattr_trusted.c | 2 -- fs/reiserfs/xattr_user.c | 2 -- 3 files changed, 6 deletions(-) (limited to 'fs') diff --git a/fs/reiserfs/xattr_security.c b/fs/reiserfs/xattr_security.c index 5e90a95ad60..056008db137 100644 --- a/fs/reiserfs/xattr_security.c +++ b/fs/reiserfs/xattr_security.c @@ -6,8 +6,6 @@ #include #include -#define XATTR_SECURITY_PREFIX "security." - static int security_get(struct inode *inode, const char *name, void *buffer, size_t size) { diff --git a/fs/reiserfs/xattr_trusted.c b/fs/reiserfs/xattr_trusted.c index 024a938ca60..60abe2bb1f9 100644 --- a/fs/reiserfs/xattr_trusted.c +++ b/fs/reiserfs/xattr_trusted.c @@ -7,8 +7,6 @@ #include #include -#define XATTR_TRUSTED_PREFIX "trusted." - static int trusted_get(struct inode *inode, const char *name, void *buffer, size_t size) { diff --git a/fs/reiserfs/xattr_user.c b/fs/reiserfs/xattr_user.c index 073f39364b1..1384efcb938 100644 --- a/fs/reiserfs/xattr_user.c +++ b/fs/reiserfs/xattr_user.c @@ -10,8 +10,6 @@ # include #endif -#define XATTR_USER_PREFIX "user." - static int user_get(struct inode *inode, const char *name, void *buffer, size_t size) { -- cgit v1.2.3-70-g09d2 From 8d44d9741f6808c107a144f469fb89e6fe7c55e3 Mon Sep 17 00:00:00 2001 From: OGAWA Hirofumi Date: Fri, 25 Jul 2008 01:46:41 -0700 Subject: fat: fix parse_options() Current parse_options() exits too early. We need to run the code of bottom in this function even if users doesn't specify options. Signed-off-by: OGAWA Hirofumi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fat/inode.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/fat/inode.c b/fs/fat/inode.c index 46a4508ffd2..60deb5fd118 100644 --- a/fs/fat/inode.c +++ b/fs/fat/inode.c @@ -950,7 +950,7 @@ static int parse_options(char *options, int is_vfat, int silent, int *debug, *debug = 0; if (!options) - return 0; + goto out; while ((p = strsep(&options, ",")) != NULL) { int token; @@ -1104,10 +1104,13 @@ static int parse_options(char *options, int is_vfat, int silent, int *debug, return -EINVAL; } } + +out: /* UTF-8 doesn't provide FAT semantics */ if (!strcmp(opts->iocharset, "utf8")) { printk(KERN_ERR "FAT: utf8 is not a recommended IO charset" - " for FAT filesystems, filesystem will be case sensitive!\n"); + " for FAT filesystems, filesystem will be " + "case sensitive!\n"); } /* If user doesn't specify allow_utime, it's initialized from dmask. */ -- cgit v1.2.3-70-g09d2 From 531f710f8e68fd2bad7516a090bff372f5f9cf6d Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Fri, 25 Jul 2008 01:46:43 -0700 Subject: fat/dir.c: switch to struct __fat_dirent struct __fat_dirent is what was formerly the kernel struct dirent (that was different from the userspace struct dirent). Converting all fat users to struct __fat_dirent will allow us to get rid of the conflicting struct dirent definition. Signed-off-by: Adrian Bunk Signed-off-by: OGAWA Hirofumi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fat/dir.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/fat/dir.c b/fs/fat/dir.c index 34541d06e62..b57c4b1db63 100644 --- a/fs/fat/dir.c +++ b/fs/fat/dir.c @@ -17,7 +17,6 @@ #include #include #include -#include #include #include #include @@ -715,7 +714,7 @@ efault: \ return -EFAULT; \ } -FAT_IOCTL_FILLDIR_FUNC(fat_ioctl_filldir, dirent) +FAT_IOCTL_FILLDIR_FUNC(fat_ioctl_filldir, __fat_dirent) static int fat_ioctl_readdir(struct inode *inode, struct file *filp, void __user *dirent, filldir_t filldir, @@ -741,7 +740,7 @@ static int fat_ioctl_readdir(struct inode *inode, struct file *filp, static int fat_dir_ioctl(struct inode *inode, struct file *filp, unsigned int cmd, unsigned long arg) { - struct dirent __user *d1 = (struct dirent __user *)arg; + struct __fat_dirent __user *d1 = (struct __fat_dirent __user *)arg; int short_only, both; switch (cmd) { @@ -757,7 +756,7 @@ static int fat_dir_ioctl(struct inode *inode, struct file *filp, return fat_generic_ioctl(inode, filp, cmd, arg); } - if (!access_ok(VERIFY_WRITE, d1, sizeof(struct dirent[2]))) + if (!access_ok(VERIFY_WRITE, d1, sizeof(struct __fat_dirent[2]))) return -EFAULT; /* * Yes, we don't need this put_user() absolutely. However old -- cgit v1.2.3-70-g09d2 From d688611674cc9c265ee67e89d2ea8bf060c17e8d Mon Sep 17 00:00:00 2001 From: OGAWA Hirofumi Date: Fri, 25 Jul 2008 01:46:43 -0700 Subject: fat: cleanup fs/fat/dir.c This is no logic changes, just cleans fs/fat/dir.c up. Signed-off-by: OGAWA Hirofumi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fat/dir.c | 131 ++++++++++++++++++++++++++++++----------------------------- 1 file changed, 67 insertions(+), 64 deletions(-) (limited to 'fs') diff --git a/fs/fat/dir.c b/fs/fat/dir.c index b57c4b1db63..a4410740627 100644 --- a/fs/fat/dir.c +++ b/fs/fat/dir.c @@ -123,10 +123,11 @@ static inline int fat_get_entry(struct inode *dir, loff_t *pos, * but ignore that right now. * Ahem... Stack smashing in ring 0 isn't fun. Fixed. */ -static int uni16_to_x8(unsigned char *ascii, wchar_t *uni, int len, +static int uni16_to_x8(unsigned char *ascii, const wchar_t *uni, int len, int uni_xlate, struct nls_table *nls) { - wchar_t *ip, ec; + const wchar_t *ip; + wchar_t ec; unsigned char *op, nc; int charlen; int k; @@ -166,6 +167,16 @@ static int uni16_to_x8(unsigned char *ascii, wchar_t *uni, int len, return (op - ascii); } +static inline int fat_uni_to_x8(struct msdos_sb_info *sbi, const wchar_t *uni, + unsigned char *buf, int size) +{ + if (sbi->options.utf8) + return utf8_wcstombs(buf, uni, size); + else + return uni16_to_x8(buf, uni, size, sbi->options.unicode_xlate, + sbi->nls_io); +} + static inline int fat_short2uni(struct nls_table *t, unsigned char *c, int clen, wchar_t *uni) { @@ -226,6 +237,19 @@ fat_shortname2uni(struct nls_table *nls, unsigned char *buf, int buf_size, return len; } +static inline int fat_name_match(struct msdos_sb_info *sbi, + const unsigned char *a, int a_len, + const unsigned char *b, int b_len) +{ + if (a_len != b_len) + return 0; + + if (sbi->options.name_check != 's') + return !nls_strnicmp(sbi->nls_io, a, b, a_len); + else + return !memcmp(a, b, a_len); +} + enum { PARSE_INVALID = 1, PARSE_NOT_LONGNAME, PARSE_EOF, }; /** @@ -311,29 +335,24 @@ int fat_search_long(struct inode *inode, const unsigned char *name, struct msdos_sb_info *sbi = MSDOS_SB(sb); struct buffer_head *bh = NULL; struct msdos_dir_entry *de; - struct nls_table *nls_io = sbi->nls_io; struct nls_table *nls_disk = sbi->nls_disk; - wchar_t bufuname[14]; unsigned char nr_slots; - int xlate_len; + wchar_t bufuname[14]; wchar_t *unicode = NULL; unsigned char work[MSDOS_NAME]; unsigned char *bufname = NULL; - int uni_xlate = sbi->options.unicode_xlate; - int utf8 = sbi->options.utf8; - int anycase = (sbi->options.name_check != 's'); unsigned short opt_shortname = sbi->options.shortname; loff_t cpos = 0; - int chl, i, j, last_u, err; + int chl, i, j, last_u, err, len; bufname = __getname(); if (!bufname) return -ENOMEM; err = -ENOENT; - while(1) { + while (1) { if (fat_get_entry(inode, &cpos, &bh, &de) == -1) - goto EODir; + goto end_of_dir; parse_record: nr_slots = 0; if (de->name[0] == DELETED_FLAG) @@ -352,7 +371,7 @@ parse_record: else if (status == PARSE_NOT_LONGNAME) goto parse_record; else if (status == PARSE_EOF) - goto EODir; + goto end_of_dir; } memcpy(work, de->name, sizeof(de->name)); @@ -393,30 +412,21 @@ parse_record: if (!last_u) continue; + /* Compare shortname */ bufuname[last_u] = 0x0000; - xlate_len = utf8 - ?utf8_wcstombs(bufname, bufuname, PATH_MAX) - :uni16_to_x8(bufname, bufuname, PATH_MAX, uni_xlate, nls_io); - if (xlate_len == name_len) - if ((!anycase && !memcmp(name, bufname, xlate_len)) || - (anycase && !nls_strnicmp(nls_io, name, bufname, - xlate_len))) - goto Found; + len = fat_uni_to_x8(sbi, bufuname, bufname, PATH_MAX); + if (fat_name_match(sbi, name, name_len, bufname, len)) + goto found; if (nr_slots) { - xlate_len = utf8 - ?utf8_wcstombs(bufname, unicode, PATH_MAX) - :uni16_to_x8(bufname, unicode, PATH_MAX, uni_xlate, nls_io); - if (xlate_len != name_len) - continue; - if ((!anycase && !memcmp(name, bufname, xlate_len)) || - (anycase && !nls_strnicmp(nls_io, name, bufname, - xlate_len))) - goto Found; + /* Compare longname */ + len = fat_uni_to_x8(sbi, unicode, bufname, PATH_MAX); + if (fat_name_match(sbi, name, name_len, bufname, len)) + goto found; } } -Found: +found: nr_slots++; /* include the de */ sinfo->slot_off = cpos - nr_slots * sizeof(*de); sinfo->nr_slots = nr_slots; @@ -424,7 +434,7 @@ Found: sinfo->bh = bh; sinfo->i_pos = fat_make_i_pos(sb, sinfo->bh, sinfo->de); err = 0; -EODir: +end_of_dir: if (bufname) __putname(bufname); if (unicode) @@ -452,23 +462,19 @@ static int __fat_readdir(struct inode *inode, struct file *filp, void *dirent, struct msdos_sb_info *sbi = MSDOS_SB(sb); struct buffer_head *bh; struct msdos_dir_entry *de; - struct nls_table *nls_io = sbi->nls_io; struct nls_table *nls_disk = sbi->nls_disk; - unsigned char long_slots; - const char *fill_name; - int fill_len; + unsigned char nr_slots; wchar_t bufuname[14]; wchar_t *unicode = NULL; unsigned char c, work[MSDOS_NAME], bufname[56], *ptname = bufname; - unsigned long lpos, dummy, *furrfu = &lpos; - int uni_xlate = sbi->options.unicode_xlate; + unsigned short opt_shortname = sbi->options.shortname; int isvfat = sbi->options.isvfat; - int utf8 = sbi->options.utf8; int nocase = sbi->options.nocase; - unsigned short opt_shortname = sbi->options.shortname; + const char *fill_name; unsigned long inum; - int chi, chl, i, i2, j, last, last_u, dotoffset = 0; + unsigned long lpos, dummy, *furrfu = &lpos; loff_t cpos; + int chi, chl, i, i2, j, last, last_u, dotoffset = 0, fill_len; int ret = 0; lock_super(sb); @@ -488,43 +494,43 @@ static int __fat_readdir(struct inode *inode, struct file *filp, void *dirent, cpos = 0; } } - if (cpos & (sizeof(struct msdos_dir_entry)-1)) { + if (cpos & (sizeof(struct msdos_dir_entry) - 1)) { ret = -ENOENT; goto out; } bh = NULL; -GetNew: +get_new: if (fat_get_entry(inode, &cpos, &bh, &de) == -1) - goto EODir; + goto end_of_dir; parse_record: - long_slots = 0; + nr_slots = 0; /* Check for long filename entry */ if (isvfat) { if (de->name[0] == DELETED_FLAG) - goto RecEnd; + goto record_end; if (de->attr != ATTR_EXT && (de->attr & ATTR_VOLUME)) - goto RecEnd; + goto record_end; if (de->attr != ATTR_EXT && IS_FREE(de->name)) - goto RecEnd; + goto record_end; } else { if ((de->attr & ATTR_VOLUME) || IS_FREE(de->name)) - goto RecEnd; + goto record_end; } if (isvfat && de->attr == ATTR_EXT) { int status = fat_parse_long(inode, &cpos, &bh, &de, - &unicode, &long_slots); + &unicode, &nr_slots); if (status < 0) { filp->f_pos = cpos; ret = status; goto out; } else if (status == PARSE_INVALID) - goto RecEnd; + goto record_end; else if (status == PARSE_NOT_LONGNAME) goto parse_record; else if (status == PARSE_EOF) - goto EODir; + goto end_of_dir; } if (sbi->options.dotsOK) { @@ -586,12 +592,12 @@ parse_record: } } if (!last) - goto RecEnd; + goto record_end; i = last + dotoffset; j = last_u; - lpos = cpos - (long_slots+1)*sizeof(struct msdos_dir_entry); + lpos = cpos - (nr_slots + 1) * sizeof(struct msdos_dir_entry); if (!memcmp(de->name, MSDOS_DOT, MSDOS_NAME)) inum = inode->i_ino; else if (!memcmp(de->name, MSDOS_DOTDOT, MSDOS_NAME)) { @@ -608,20 +614,17 @@ parse_record: if (isvfat) { bufuname[j] = 0x0000; - i = utf8 ? utf8_wcstombs(bufname, bufuname, sizeof(bufname)) - : uni16_to_x8(bufname, bufuname, sizeof(bufname), uni_xlate, nls_io); + i = fat_uni_to_x8(sbi, bufuname, bufname, sizeof(bufname)); } fill_name = bufname; fill_len = i; - if (!short_only && long_slots) { + if (!short_only && nr_slots) { /* convert the unicode long name. 261 is maximum size * of unicode buffer. (13 * slots + nul) */ void *longname = unicode + 261; int buf_size = PATH_MAX - (261 * sizeof(unicode[0])); - int long_len = utf8 - ? utf8_wcstombs(longname, unicode, buf_size) - : uni16_to_x8(longname, unicode, buf_size, uni_xlate, nls_io); + int long_len = fat_uni_to_x8(sbi, unicode, longname, buf_size); if (!both) { fill_name = longname; @@ -640,15 +643,15 @@ parse_record: } if (filldir(dirent, fill_name, fill_len, *furrfu, inum, (de->attr & ATTR_DIR) ? DT_DIR : DT_REG) < 0) - goto FillFailed; + goto fill_failed; -RecEnd: +record_end: furrfu = &lpos; filp->f_pos = cpos; - goto GetNew; -EODir: + goto get_new; +end_of_dir: filp->f_pos = cpos; -FillFailed: +fill_failed: brelse(bh); if (unicode) __putname(unicode); -- cgit v1.2.3-70-g09d2 From 98a15160049fc1a0f822047f33ff513906a35567 Mon Sep 17 00:00:00 2001 From: OGAWA Hirofumi Date: Fri, 25 Jul 2008 01:46:44 -0700 Subject: fat: use same logic in fat_search_long() and __fat_readdir() This uses uses stack for shortname, and uses __getname() for longname in fat_search_long() and __fat_readdir(). By this, it removes unneeded __getname() for shortname. Signed-off-by: OGAWA Hirofumi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fat/dir.c | 42 ++++++++++++++++++++++++++---------------- 1 file changed, 26 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/fat/dir.c b/fs/fat/dir.c index a4410740627..96a1cad30da 100644 --- a/fs/fat/dir.c +++ b/fs/fat/dir.c @@ -324,6 +324,19 @@ parse_long: return 0; } +/* + * Maximum buffer size of short name. + * [(MSDOS_NAME + '.') * max one char + nul] + * For msdos style, ['.' (hidden) + MSDOS_NAME + '.' + nul] + */ +#define FAT_MAX_SHORT_SIZE ((MSDOS_NAME + 1) * NLS_MAX_CHARSET_SIZE + 1) +/* + * Maximum buffer size of unicode chars from slots. + * [(max longname slots * 13 (size in a slot) + nul) * sizeof(wchar_t)] + */ +#define FAT_MAX_UNI_CHARS ((MSDOS_SLOTS - 1) * 13 + 1) +#define FAT_MAX_UNI_SIZE (FAT_MAX_UNI_CHARS * sizeof(wchar_t)) + /* * Return values: negative -> error, 0 -> not found, positive -> found, * value is the total amount of slots, including the shortname entry. @@ -340,15 +353,11 @@ int fat_search_long(struct inode *inode, const unsigned char *name, wchar_t bufuname[14]; wchar_t *unicode = NULL; unsigned char work[MSDOS_NAME]; - unsigned char *bufname = NULL; + unsigned char bufname[FAT_MAX_SHORT_SIZE]; unsigned short opt_shortname = sbi->options.shortname; loff_t cpos = 0; int chl, i, j, last_u, err, len; - bufname = __getname(); - if (!bufname) - return -ENOMEM; - err = -ENOENT; while (1) { if (fat_get_entry(inode, &cpos, &bh, &de) == -1) @@ -414,14 +423,17 @@ parse_record: /* Compare shortname */ bufuname[last_u] = 0x0000; - len = fat_uni_to_x8(sbi, bufuname, bufname, PATH_MAX); + len = fat_uni_to_x8(sbi, bufuname, bufname, sizeof(bufname)); if (fat_name_match(sbi, name, name_len, bufname, len)) goto found; if (nr_slots) { + void *longname = unicode + FAT_MAX_UNI_CHARS; + int size = PATH_MAX - FAT_MAX_UNI_SIZE; + /* Compare longname */ - len = fat_uni_to_x8(sbi, unicode, bufname, PATH_MAX); - if (fat_name_match(sbi, name, name_len, bufname, len)) + len = fat_uni_to_x8(sbi, unicode, longname, size); + if (fat_name_match(sbi, name, name_len, longname, len)) goto found; } } @@ -435,8 +447,6 @@ found: sinfo->i_pos = fat_make_i_pos(sb, sinfo->bh, sinfo->de); err = 0; end_of_dir: - if (bufname) - __putname(bufname); if (unicode) __putname(unicode); @@ -466,7 +476,8 @@ static int __fat_readdir(struct inode *inode, struct file *filp, void *dirent, unsigned char nr_slots; wchar_t bufuname[14]; wchar_t *unicode = NULL; - unsigned char c, work[MSDOS_NAME], bufname[56], *ptname = bufname; + unsigned char c, work[MSDOS_NAME]; + unsigned char bufname[FAT_MAX_SHORT_SIZE], *ptname = bufname; unsigned short opt_shortname = sbi->options.shortname; int isvfat = sbi->options.isvfat; int nocase = sbi->options.nocase; @@ -620,11 +631,10 @@ parse_record: fill_name = bufname; fill_len = i; if (!short_only && nr_slots) { - /* convert the unicode long name. 261 is maximum size - * of unicode buffer. (13 * slots + nul) */ - void *longname = unicode + 261; - int buf_size = PATH_MAX - (261 * sizeof(unicode[0])); - int long_len = fat_uni_to_x8(sbi, unicode, longname, buf_size); + void *longname = unicode + FAT_MAX_UNI_CHARS; + int long_len, size = PATH_MAX - FAT_MAX_UNI_SIZE; + + long_len = fat_uni_to_x8(sbi, unicode, longname, size); if (!both) { fill_name = longname; -- cgit v1.2.3-70-g09d2 From dcd8c53f13f068ee039589d84fbd0baf686abc41 Mon Sep 17 00:00:00 2001 From: OGAWA Hirofumi Date: Fri, 25 Jul 2008 01:46:44 -0700 Subject: fat: small optimization to __fat_readdir() This removes unnecessary parsing for directory entries. If short_only, we don't need to parse longname. And if !both and it found the longname, we don't need shortname. Signed-off-by: OGAWA Hirofumi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fat/dir.c | 71 +++++++++++++++++++++++++++++++++--------------------------- 1 file changed, 39 insertions(+), 32 deletions(-) (limited to 'fs') diff --git a/fs/fat/dir.c b/fs/fat/dir.c index 96a1cad30da..4c35477bc94 100644 --- a/fs/fat/dir.c +++ b/fs/fat/dir.c @@ -481,11 +481,11 @@ static int __fat_readdir(struct inode *inode, struct file *filp, void *dirent, unsigned short opt_shortname = sbi->options.shortname; int isvfat = sbi->options.isvfat; int nocase = sbi->options.nocase; - const char *fill_name; + const char *fill_name = NULL; unsigned long inum; unsigned long lpos, dummy, *furrfu = &lpos; loff_t cpos; - int chi, chl, i, i2, j, last, last_u, dotoffset = 0, fill_len; + int chi, chl, i, i2, j, last, last_u, dotoffset = 0, fill_len = 0; int ret = 0; lock_super(sb); @@ -516,8 +516,11 @@ get_new: goto end_of_dir; parse_record: nr_slots = 0; - /* Check for long filename entry */ - if (isvfat) { + /* + * Check for long filename entry, but if short_only, we don't + * need to parse long filename. + */ + if (isvfat && !short_only) { if (de->name[0] == DELETED_FLAG) goto record_end; if (de->attr != ATTR_EXT && (de->attr & ATTR_VOLUME)) @@ -542,6 +545,18 @@ parse_record: goto parse_record; else if (status == PARSE_EOF) goto end_of_dir; + + if (nr_slots) { + void *longname = unicode + FAT_MAX_UNI_CHARS; + int size = PATH_MAX - FAT_MAX_UNI_SIZE; + int len = fat_uni_to_x8(sbi, unicode, longname, size); + + fill_name = longname; + fill_len = len; + /* !both && !short_only, so we don't need shortname. */ + if (!both) + goto start_filldir; + } } if (sbi->options.dotsOK) { @@ -608,6 +623,26 @@ parse_record: i = last + dotoffset; j = last_u; + if (isvfat) { + bufuname[j] = 0x0000; + i = fat_uni_to_x8(sbi, bufuname, bufname, sizeof(bufname)); + } + if (nr_slots) { + /* hack for fat_ioctl_filldir() */ + struct fat_ioctl_filldir_callback *p = dirent; + + p->longname = fill_name; + p->long_len = fill_len; + p->shortname = bufname; + p->short_len = i; + fill_name = NULL; + fill_len = 0; + } else { + fill_name = bufname; + fill_len = i; + } + +start_filldir: lpos = cpos - (nr_slots + 1) * sizeof(struct msdos_dir_entry); if (!memcmp(de->name, MSDOS_DOT, MSDOS_NAME)) inum = inode->i_ino; @@ -623,34 +658,6 @@ parse_record: inum = iunique(sb, MSDOS_ROOT_INO); } - if (isvfat) { - bufuname[j] = 0x0000; - i = fat_uni_to_x8(sbi, bufuname, bufname, sizeof(bufname)); - } - - fill_name = bufname; - fill_len = i; - if (!short_only && nr_slots) { - void *longname = unicode + FAT_MAX_UNI_CHARS; - int long_len, size = PATH_MAX - FAT_MAX_UNI_SIZE; - - long_len = fat_uni_to_x8(sbi, unicode, longname, size); - - if (!both) { - fill_name = longname; - fill_len = long_len; - } else { - /* hack for fat_ioctl_filldir() */ - struct fat_ioctl_filldir_callback *p = dirent; - - p->longname = longname; - p->long_len = long_len; - p->shortname = bufname; - p->short_len = i; - fill_name = NULL; - fill_len = 0; - } - } if (filldir(dirent, fill_name, fill_len, *furrfu, inum, (de->attr & ATTR_DIR) ? DT_DIR : DT_REG) < 0) goto fill_failed; -- cgit v1.2.3-70-g09d2 From 7557bc66be629d19a402e752673708bfbb8b5e86 Mon Sep 17 00:00:00 2001 From: Rene Scharfe Date: Fri, 25 Jul 2008 01:46:45 -0700 Subject: msdos fs: remove unsettable atari option It has been impossible to set the option 'atari' of the MSDOS filesystem for several years. Since nobody seems to have missed it, let's remove its remains. Signed-off-by: Rene Scharfe Acked-by: OGAWA Hirofumi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/msdos/namei.c | 18 ++++++------------ include/linux/msdos_fs.h | 1 - 2 files changed, 6 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/msdos/namei.c b/fs/msdos/namei.c index 1f7f2956412..e4ad6c6b753 100644 --- a/fs/msdos/namei.c +++ b/fs/msdos/namei.c @@ -14,12 +14,7 @@ /* Characters that are undesirable in an MS-DOS file name */ static unsigned char bad_chars[] = "*?<>|\""; -static unsigned char bad_if_strict_pc[] = "+=,; "; -/* GEMDOS is less restrictive */ -static unsigned char bad_if_strict_atari[] = " "; - -#define bad_if_strict(opts) \ - ((opts)->atari ? bad_if_strict_atari : bad_if_strict_pc) +static unsigned char bad_if_strict[] = "+=,; "; /***** Formats an MS-DOS file name. Rejects invalid names. */ static int msdos_format_name(const unsigned char *name, int len, @@ -40,21 +35,20 @@ static int msdos_format_name(const unsigned char *name, int len, /* Get rid of dot - test for it elsewhere */ name++; len--; - } else if (!opts->atari) + } else return -EINVAL; } /* - * disallow names that _really_ start with a dot for MS-DOS, - * GEMDOS does not care + * disallow names that _really_ start with a dot */ - space = !opts->atari; + space = 1; c = 0; for (walk = res; len && walk - res < 8; walk++) { c = *name++; len--; if (opts->name_check != 'r' && strchr(bad_chars, c)) return -EINVAL; - if (opts->name_check == 's' && strchr(bad_if_strict(opts), c)) + if (opts->name_check == 's' && strchr(bad_if_strict, c)) return -EINVAL; if (c >= 'A' && c <= 'Z' && opts->name_check == 's') return -EINVAL; @@ -94,7 +88,7 @@ static int msdos_format_name(const unsigned char *name, int len, if (opts->name_check != 'r' && strchr(bad_chars, c)) return -EINVAL; if (opts->name_check == 's' && - strchr(bad_if_strict(opts), c)) + strchr(bad_if_strict, c)) return -EINVAL; if (c < ' ' || c == ':' || c == '\\') return -EINVAL; diff --git a/include/linux/msdos_fs.h b/include/linux/msdos_fs.h index 5161394c789..3346c9c8f17 100644 --- a/include/linux/msdos_fs.h +++ b/include/linux/msdos_fs.h @@ -201,7 +201,6 @@ struct fat_mount_options { utf8:1, /* Use of UTF-8 character set (Default) */ unicode_xlate:1, /* create escape sequences for unhandled Unicode */ numtail:1, /* Does first alias have a numeric '~1' type tail? */ - atari:1, /* Use Atari GEMDOS variation of MS-DOS fs */ flush:1, /* write things quickly */ nocase:1, /* Does this need case conversion? 0=need case conversion*/ usefree:1; /* Use free_clusters for FAT32 */ -- cgit v1.2.3-70-g09d2 From e8938a62a85d1f487e02c3b01955b47c9598f6d2 Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Fri, 25 Jul 2008 01:46:46 -0700 Subject: remove unused #include 's Remove some unused #include 's. Signed-off-by: Adrian Bunk Cc: Ralf Baechle Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/mips/kernel/linux32.c | 1 - fs/compat_ioctl.c | 1 - fs/smbfs/cache.c | 1 - fs/smbfs/proc.c | 1 - include/linux/nfsd/nfsd.h | 1 - 5 files changed, 5 deletions(-) (limited to 'fs') diff --git a/arch/mips/kernel/linux32.c b/arch/mips/kernel/linux32.c index c266211ed65..2fefb14414b 100644 --- a/arch/mips/kernel/linux32.c +++ b/arch/mips/kernel/linux32.c @@ -11,7 +11,6 @@ #include #include #include -#include #include #include #include diff --git a/fs/compat_ioctl.c b/fs/compat_ioctl.c index 18e2c548161..5235c67e759 100644 --- a/fs/compat_ioctl.c +++ b/fs/compat_ioctl.c @@ -25,7 +25,6 @@ #include #include #include -#include #include #include #include diff --git a/fs/smbfs/cache.c b/fs/smbfs/cache.c index 8182f0542a2..8c177eb7e34 100644 --- a/fs/smbfs/cache.c +++ b/fs/smbfs/cache.c @@ -13,7 +13,6 @@ #include #include #include -#include #include #include #include diff --git a/fs/smbfs/proc.c b/fs/smbfs/proc.c index d517a27b7f4..ee536e8a649 100644 --- a/fs/smbfs/proc.c +++ b/fs/smbfs/proc.c @@ -16,7 +16,6 @@ #include #include #include -#include #include #include #include diff --git a/include/linux/nfsd/nfsd.h b/include/linux/nfsd/nfsd.h index a2861d95ecc..108f47e5fd9 100644 --- a/include/linux/nfsd/nfsd.h +++ b/include/linux/nfsd/nfsd.h @@ -12,7 +12,6 @@ #include #include -#include #include #include #include -- cgit v1.2.3-70-g09d2 From b271e067c896ad4082b15e96077675d08db40625 Mon Sep 17 00:00:00 2001 From: Joe Peterson Date: Fri, 25 Jul 2008 01:46:47 -0700 Subject: fatfs: add UTC timestamp option Provide a new mount option ("tz=UTC") for DOS (vfat/msdos) filesystems, allowing timestamps to be in coordinated universal time (UTC) rather than local time in applications where doing this is advantageous. In particular, portable devices that use fat/vfat (such as digital cameras) can benefit from using UTC in their internal clocks, thus avoiding daylight saving time errors and general time ambiguity issues. The user of the device does not have to worry about changing the time when moving from place or when daylight saving changes. The new mount option, when set, disables the counter-adjustment that Linux currently makes to FAT timestamp info in anticipation of the normal userspace time zone correction. When used in this new mode, all daylight saving time and time zone handling is done in userspace as is normal for many other filesystems (like ext3). The default mode, which remains unchanged, is still appropriate when mounting volumes written in Windows (because of its use of local time). I originally based this patch on one submitted last year by Paul Collins, but I updated it to work with current source and changed variable/option naming. Ogawa Hirofumi (who maintains these filesystems) and I discussed this patch at length on lkml, and he suggested using the option name in the attached version of the patch. Barry Bouwsma pointed out a good addition to the patch as well. Signed-off-by: Joe Peterson Signed-off-by: Paul Collins Acked-by: OGAWA Hirofumi Cc: Barry Bouwsma Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fat/dir.c | 2 +- fs/fat/inode.c | 27 ++++++++++++++++++++------- fs/fat/misc.c | 10 ++++++---- fs/msdos/namei.c | 3 ++- fs/vfat/namei.c | 2 +- include/linux/msdos_fs.h | 8 +++++--- 6 files changed, 35 insertions(+), 17 deletions(-) (limited to 'fs') diff --git a/fs/fat/dir.c b/fs/fat/dir.c index 4c35477bc94..cd4a0162e10 100644 --- a/fs/fat/dir.c +++ b/fs/fat/dir.c @@ -1101,7 +1101,7 @@ int fat_alloc_new_dir(struct inode *dir, struct timespec *ts) goto error_free; } - fat_date_unix2dos(ts->tv_sec, &time, &date); + fat_date_unix2dos(ts->tv_sec, &time, &date, sbi->options.tz_utc); de = (struct msdos_dir_entry *)bhs[0]->b_data; /* filling the new directory slots ("." and ".." entries) */ diff --git a/fs/fat/inode.c b/fs/fat/inode.c index 60deb5fd118..23676f9d79c 100644 --- a/fs/fat/inode.c +++ b/fs/fat/inode.c @@ -382,17 +382,20 @@ static int fat_fill_inode(struct inode *inode, struct msdos_dir_entry *de) inode->i_blocks = ((inode->i_size + (sbi->cluster_size - 1)) & ~((loff_t)sbi->cluster_size - 1)) >> 9; inode->i_mtime.tv_sec = - date_dos2unix(le16_to_cpu(de->time), le16_to_cpu(de->date)); + date_dos2unix(le16_to_cpu(de->time), le16_to_cpu(de->date), + sbi->options.tz_utc); inode->i_mtime.tv_nsec = 0; if (sbi->options.isvfat) { int secs = de->ctime_cs / 100; int csecs = de->ctime_cs % 100; inode->i_ctime.tv_sec = date_dos2unix(le16_to_cpu(de->ctime), - le16_to_cpu(de->cdate)) + secs; + le16_to_cpu(de->cdate), + sbi->options.tz_utc) + secs; inode->i_ctime.tv_nsec = csecs * 10000000; inode->i_atime.tv_sec = - date_dos2unix(0, le16_to_cpu(de->adate)); + date_dos2unix(0, le16_to_cpu(de->adate), + sbi->options.tz_utc); inode->i_atime.tv_nsec = 0; } else inode->i_ctime = inode->i_atime = inode->i_mtime; @@ -591,11 +594,14 @@ retry: raw_entry->attr = fat_attr(inode); raw_entry->start = cpu_to_le16(MSDOS_I(inode)->i_logstart); raw_entry->starthi = cpu_to_le16(MSDOS_I(inode)->i_logstart >> 16); - fat_date_unix2dos(inode->i_mtime.tv_sec, &raw_entry->time, &raw_entry->date); + fat_date_unix2dos(inode->i_mtime.tv_sec, &raw_entry->time, + &raw_entry->date, sbi->options.tz_utc); if (sbi->options.isvfat) { __le16 atime; - fat_date_unix2dos(inode->i_ctime.tv_sec,&raw_entry->ctime,&raw_entry->cdate); - fat_date_unix2dos(inode->i_atime.tv_sec,&atime,&raw_entry->adate); + fat_date_unix2dos(inode->i_ctime.tv_sec, &raw_entry->ctime, + &raw_entry->cdate, sbi->options.tz_utc); + fat_date_unix2dos(inode->i_atime.tv_sec, &atime, + &raw_entry->adate, sbi->options.tz_utc); raw_entry->ctime_cs = (inode->i_ctime.tv_sec & 1) * 100 + inode->i_ctime.tv_nsec / 10000000; } @@ -836,6 +842,8 @@ static int fat_show_options(struct seq_file *m, struct vfsmount *mnt) } if (sbi->options.flush) seq_puts(m, ",flush"); + if (opts->tz_utc) + seq_puts(m, ",tz=UTC"); return 0; } @@ -848,7 +856,7 @@ enum { Opt_charset, Opt_shortname_lower, Opt_shortname_win95, Opt_shortname_winnt, Opt_shortname_mixed, Opt_utf8_no, Opt_utf8_yes, Opt_uni_xl_no, Opt_uni_xl_yes, Opt_nonumtail_no, Opt_nonumtail_yes, - Opt_obsolate, Opt_flush, Opt_err, + Opt_obsolate, Opt_flush, Opt_tz_utc, Opt_err, }; static match_table_t fat_tokens = { @@ -883,6 +891,7 @@ static match_table_t fat_tokens = { {Opt_obsolate, "cvf_options=%100s"}, {Opt_obsolate, "posix"}, {Opt_flush, "flush"}, + {Opt_tz_utc, "tz=UTC"}, {Opt_err, NULL}, }; static match_table_t msdos_tokens = { @@ -947,6 +956,7 @@ static int parse_options(char *options, int is_vfat, int silent, int *debug, opts->utf8 = opts->unicode_xlate = 0; opts->numtail = 1; opts->usefree = opts->nocase = 0; + opts->tz_utc = 0; *debug = 0; if (!options) @@ -1036,6 +1046,9 @@ static int parse_options(char *options, int is_vfat, int silent, int *debug, case Opt_flush: opts->flush = 1; break; + case Opt_tz_utc: + opts->tz_utc = 1; + break; /* msdos specific */ case Opt_dots: diff --git a/fs/fat/misc.c b/fs/fat/misc.c index 61f23511eac..79fb98ad36d 100644 --- a/fs/fat/misc.c +++ b/fs/fat/misc.c @@ -142,7 +142,7 @@ static int day_n[] = { }; /* Convert a MS-DOS time/date pair to a UNIX date (seconds since 1 1 70). */ -int date_dos2unix(unsigned short time, unsigned short date) +int date_dos2unix(unsigned short time, unsigned short date, int tz_utc) { int month, year, secs; @@ -156,16 +156,18 @@ int date_dos2unix(unsigned short time, unsigned short date) ((date & 31)-1+day_n[month]+(year/4)+year*365-((year & 3) == 0 && month < 2 ? 1 : 0)+3653); /* days since 1.1.70 plus 80's leap day */ - secs += sys_tz.tz_minuteswest*60; + if (!tz_utc) + secs += sys_tz.tz_minuteswest*60; return secs; } /* Convert linear UNIX date to a MS-DOS time/date pair. */ -void fat_date_unix2dos(int unix_date, __le16 *time, __le16 *date) +void fat_date_unix2dos(int unix_date, __le16 *time, __le16 *date, int tz_utc) { int day, year, nl_day, month; - unix_date -= sys_tz.tz_minuteswest*60; + if (!tz_utc) + unix_date -= sys_tz.tz_minuteswest*60; /* Jan 1 GMT 00:00:00 1980. But what about another time zone? */ if (unix_date < 315532800) diff --git a/fs/msdos/namei.c b/fs/msdos/namei.c index e4ad6c6b753..e844b9809d2 100644 --- a/fs/msdos/namei.c +++ b/fs/msdos/namei.c @@ -237,6 +237,7 @@ static int msdos_add_entry(struct inode *dir, const unsigned char *name, int is_dir, int is_hid, int cluster, struct timespec *ts, struct fat_slot_info *sinfo) { + struct msdos_sb_info *sbi = MSDOS_SB(dir->i_sb); struct msdos_dir_entry de; __le16 time, date; int err; @@ -246,7 +247,7 @@ static int msdos_add_entry(struct inode *dir, const unsigned char *name, if (is_hid) de.attr |= ATTR_HIDDEN; de.lcase = 0; - fat_date_unix2dos(ts->tv_sec, &time, &date); + fat_date_unix2dos(ts->tv_sec, &time, &date, sbi->options.tz_utc); de.cdate = de.adate = 0; de.ctime = 0; de.ctime_cs = 0; diff --git a/fs/vfat/namei.c b/fs/vfat/namei.c index b546ba69be8..155c10b4adb 100644 --- a/fs/vfat/namei.c +++ b/fs/vfat/namei.c @@ -621,7 +621,7 @@ shortname: memcpy(de->name, msdos_name, MSDOS_NAME); de->attr = is_dir ? ATTR_DIR : ATTR_ARCH; de->lcase = lcase; - fat_date_unix2dos(ts->tv_sec, &time, &date); + fat_date_unix2dos(ts->tv_sec, &time, &date, sbi->options.tz_utc); de->time = de->ctime = time; de->date = de->cdate = de->adate = date; de->ctime_cs = 0; diff --git a/include/linux/msdos_fs.h b/include/linux/msdos_fs.h index 3346c9c8f17..ba63858056c 100644 --- a/include/linux/msdos_fs.h +++ b/include/linux/msdos_fs.h @@ -203,7 +203,8 @@ struct fat_mount_options { numtail:1, /* Does first alias have a numeric '~1' type tail? */ flush:1, /* write things quickly */ nocase:1, /* Does this need case conversion? 0=need case conversion*/ - usefree:1; /* Use free_clusters for FAT32 */ + usefree:1, /* Use free_clusters for FAT32 */ + tz_utc:1; /* Filesystem timestamps are in UTC */ }; #define FAT_HASH_BITS 8 @@ -434,8 +435,9 @@ extern int fat_flush_inodes(struct super_block *sb, struct inode *i1, extern void fat_fs_panic(struct super_block *s, const char *fmt, ...); extern void fat_clusters_flush(struct super_block *sb); extern int fat_chain_add(struct inode *inode, int new_dclus, int nr_cluster); -extern int date_dos2unix(unsigned short time, unsigned short date); -extern void fat_date_unix2dos(int unix_date, __le16 *time, __le16 *date); +extern int date_dos2unix(unsigned short time, unsigned short date, int tz_utc); +extern void fat_date_unix2dos(int unix_date, __le16 *time, __le16 *date, + int tz_utc); extern int fat_sync_bhs(struct buffer_head **bhs, int nr_bhs); int fat_cache_init(void); -- cgit v1.2.3-70-g09d2 From b48d380541f634663b71766005838edbb7261685 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Fri, 25 Jul 2008 01:46:49 -0700 Subject: quota: fix possible infinite loop in quota code When quota structure is going to be dropped and it is dirty, quota code tries to write it. If the write fails for some reason (e. g. transaction cannot be started because the journal is aborted), we try writing again and again and again... Fix the problem by clearing the dirty bit even if the write failed. (akpm: for 2.6.27, 2.6.26.x and 2.6.25.x) Signed-off-by: Jan Kara Reviewed-by: dingdinghua Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/dquot.c | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/dquot.c b/fs/dquot.c index 5ac77da1995..ad88cf6fcba 100644 --- a/fs/dquot.c +++ b/fs/dquot.c @@ -562,6 +562,8 @@ static struct shrinker dqcache_shrinker = { */ static void dqput(struct dquot *dquot) { + int ret; + if (!dquot) return; #ifdef __DQUOT_PARANOIA @@ -594,7 +596,19 @@ we_slept: if (test_bit(DQ_ACTIVE_B, &dquot->dq_flags) && dquot_dirty(dquot)) { spin_unlock(&dq_list_lock); /* Commit dquot before releasing */ - dquot->dq_sb->dq_op->write_dquot(dquot); + ret = dquot->dq_sb->dq_op->write_dquot(dquot); + if (ret < 0) { + printk(KERN_ERR "VFS: cannot write quota structure on " + "device %s (error %d). Quota may get out of " + "sync!\n", dquot->dq_sb->s_id, ret); + /* + * We clear dirty bit anyway, so that we avoid + * infinite loop here + */ + spin_lock(&dq_list_lock); + clear_dquot_dirty(dquot); + spin_unlock(&dq_list_lock); + } goto we_slept; } /* Clear flag in case dquot was inactive (something bad happened) */ -- cgit v1.2.3-70-g09d2 From b85f4b87a511bea86dac68c4f0fabaee2cac6c4c Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Fri, 25 Jul 2008 01:46:50 -0700 Subject: quota: rename quota functions from upper case, make bigger ones non-inline Cleanup quotaops.h: Rename functions from uppercase to lowercase (and define backward compatibility macros), move larger functions to dquot.c and make them non-inline. Signed-off-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/dquot.c | 53 +++++++++++ include/linux/quotaops.h | 226 ++++++++++++++++++++++------------------------- 2 files changed, 160 insertions(+), 119 deletions(-) (limited to 'fs') diff --git a/fs/dquot.c b/fs/dquot.c index ad88cf6fcba..0bcaf970bbb 100644 --- a/fs/dquot.c +++ b/fs/dquot.c @@ -1153,6 +1153,28 @@ int dquot_drop(struct inode *inode) return 0; } +/* Wrapper to remove references to quota structures from inode */ +void vfs_dq_drop(struct inode *inode) +{ + /* Here we can get arbitrary inode from clear_inode() so we have + * to be careful. OTOH we don't need locking as quota operations + * are allowed to change only at mount time */ + if (!IS_NOQUOTA(inode) && inode->i_sb && inode->i_sb->dq_op + && inode->i_sb->dq_op->drop) { + int cnt; + /* Test before calling to rule out calls from proc and such + * where we are not allowed to block. Note that this is + * actually reliable test even without the lock - the caller + * must assure that nobody can come after the DQUOT_DROP and + * add quota pointers back anyway */ + for (cnt = 0; cnt < MAXQUOTAS; cnt++) + if (inode->i_dquot[cnt] != NODQUOT) + break; + if (cnt < MAXQUOTAS) + inode->i_sb->dq_op->drop(inode); + } +} + /* * Following four functions update i_blocks+i_bytes fields and * quota information (together with appropriate checks) @@ -1426,6 +1448,18 @@ warn_put_all: return ret; } +/* Wrapper for transferring ownership of an inode */ +int vfs_dq_transfer(struct inode *inode, struct iattr *iattr) +{ + if (sb_any_quota_enabled(inode->i_sb) && !IS_NOQUOTA(inode)) { + vfs_dq_init(inode); + if (inode->i_sb->dq_op->transfer(inode, iattr) == NO_QUOTA) + return 1; + } + return 0; +} + + /* * Write info of quota file to disk */ @@ -1766,6 +1800,22 @@ out: return error; } +/* Wrapper to turn on quotas when remounting rw */ +int vfs_dq_quota_on_remount(struct super_block *sb) +{ + int cnt; + int ret = 0, err; + + if (!sb->s_qcop || !sb->s_qcop->quota_on) + return -ENOSYS; + for (cnt = 0; cnt < MAXQUOTAS; cnt++) { + err = sb->s_qcop->quota_on(sb, cnt, 0, NULL, 1); + if (err < 0 && !ret) + ret = err; + } + return ret; +} + /* Generic routine for getting common part of quota structure */ static void do_get_dqblk(struct dquot *dquot, struct if_dqblk *di) { @@ -2101,8 +2151,11 @@ EXPORT_SYMBOL(dquot_release); EXPORT_SYMBOL(dquot_mark_dquot_dirty); EXPORT_SYMBOL(dquot_initialize); EXPORT_SYMBOL(dquot_drop); +EXPORT_SYMBOL(vfs_dq_drop); EXPORT_SYMBOL(dquot_alloc_space); EXPORT_SYMBOL(dquot_alloc_inode); EXPORT_SYMBOL(dquot_free_space); EXPORT_SYMBOL(dquot_free_inode); EXPORT_SYMBOL(dquot_transfer); +EXPORT_SYMBOL(vfs_dq_transfer); +EXPORT_SYMBOL(vfs_dq_quota_on_remount); diff --git a/include/linux/quotaops.h b/include/linux/quotaops.h index f8670205385..0c8f9fe462a 100644 --- a/include/linux/quotaops.h +++ b/include/linux/quotaops.h @@ -19,34 +19,38 @@ /* * declaration of quota_function calls in kernel. */ -extern void sync_dquots(struct super_block *sb, int type); - -extern int dquot_initialize(struct inode *inode, int type); -extern int dquot_drop(struct inode *inode); - -extern int dquot_alloc_space(struct inode *inode, qsize_t number, int prealloc); -extern int dquot_alloc_inode(const struct inode *inode, unsigned long number); - -extern int dquot_free_space(struct inode *inode, qsize_t number); -extern int dquot_free_inode(const struct inode *inode, unsigned long number); - -extern int dquot_transfer(struct inode *inode, struct iattr *iattr); -extern int dquot_commit(struct dquot *dquot); -extern int dquot_acquire(struct dquot *dquot); -extern int dquot_release(struct dquot *dquot); -extern int dquot_commit_info(struct super_block *sb, int type); -extern int dquot_mark_dquot_dirty(struct dquot *dquot); - -extern int vfs_quota_on(struct super_block *sb, int type, int format_id, - char *path, int remount); -extern int vfs_quota_on_mount(struct super_block *sb, char *qf_name, - int format_id, int type); -extern int vfs_quota_off(struct super_block *sb, int type, int remount); -extern int vfs_quota_sync(struct super_block *sb, int type); -extern int vfs_get_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii); -extern int vfs_set_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii); -extern int vfs_get_dqblk(struct super_block *sb, int type, qid_t id, struct if_dqblk *di); -extern int vfs_set_dqblk(struct super_block *sb, int type, qid_t id, struct if_dqblk *di); +void sync_dquots(struct super_block *sb, int type); + +int dquot_initialize(struct inode *inode, int type); +int dquot_drop(struct inode *inode); + +int dquot_alloc_space(struct inode *inode, qsize_t number, int prealloc); +int dquot_alloc_inode(const struct inode *inode, unsigned long number); + +int dquot_free_space(struct inode *inode, qsize_t number); +int dquot_free_inode(const struct inode *inode, unsigned long number); + +int dquot_transfer(struct inode *inode, struct iattr *iattr); +int dquot_commit(struct dquot *dquot); +int dquot_acquire(struct dquot *dquot); +int dquot_release(struct dquot *dquot); +int dquot_commit_info(struct super_block *sb, int type); +int dquot_mark_dquot_dirty(struct dquot *dquot); + +int vfs_quota_on(struct super_block *sb, int type, int format_id, + char *path, int remount); +int vfs_quota_on_mount(struct super_block *sb, char *qf_name, + int format_id, int type); +int vfs_quota_off(struct super_block *sb, int type, int remount); +int vfs_quota_sync(struct super_block *sb, int type); +int vfs_get_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii); +int vfs_set_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii); +int vfs_get_dqblk(struct super_block *sb, int type, qid_t id, struct if_dqblk *di); +int vfs_set_dqblk(struct super_block *sb, int type, qid_t id, struct if_dqblk *di); + +void vfs_dq_drop(struct inode *inode); +int vfs_dq_transfer(struct inode *inode, struct iattr *iattr); +int vfs_dq_quota_on_remount(struct super_block *sb); /* * Operations supported for diskquotas. @@ -59,38 +63,16 @@ extern struct quotactl_ops vfs_quotactl_ops; /* It is better to call this function outside of any transaction as it might * need a lot of space in journal for dquot structure allocation. */ -static inline void DQUOT_INIT(struct inode *inode) +static inline void vfs_dq_init(struct inode *inode) { BUG_ON(!inode->i_sb); if (sb_any_quota_enabled(inode->i_sb) && !IS_NOQUOTA(inode)) inode->i_sb->dq_op->initialize(inode, -1); } -/* The same as with DQUOT_INIT */ -static inline void DQUOT_DROP(struct inode *inode) -{ - /* Here we can get arbitrary inode from clear_inode() so we have - * to be careful. OTOH we don't need locking as quota operations - * are allowed to change only at mount time */ - if (!IS_NOQUOTA(inode) && inode->i_sb && inode->i_sb->dq_op - && inode->i_sb->dq_op->drop) { - int cnt; - /* Test before calling to rule out calls from proc and such - * where we are not allowed to block. Note that this is - * actually reliable test even without the lock - the caller - * must assure that nobody can come after the DQUOT_DROP and - * add quota pointers back anyway */ - for (cnt = 0; cnt < MAXQUOTAS; cnt++) - if (inode->i_dquot[cnt] != NODQUOT) - break; - if (cnt < MAXQUOTAS) - inode->i_sb->dq_op->drop(inode); - } -} - /* The following allocation/freeing/transfer functions *must* be called inside * a transaction (deadlocks possible otherwise) */ -static inline int DQUOT_PREALLOC_SPACE_NODIRTY(struct inode *inode, qsize_t nr) +static inline int vfs_dq_prealloc_space_nodirty(struct inode *inode, qsize_t nr) { if (sb_any_quota_enabled(inode->i_sb)) { /* Used space is updated in alloc_space() */ @@ -102,15 +84,15 @@ static inline int DQUOT_PREALLOC_SPACE_NODIRTY(struct inode *inode, qsize_t nr) return 0; } -static inline int DQUOT_PREALLOC_SPACE(struct inode *inode, qsize_t nr) +static inline int vfs_dq_prealloc_space(struct inode *inode, qsize_t nr) { int ret; - if (!(ret = DQUOT_PREALLOC_SPACE_NODIRTY(inode, nr))) + if (!(ret = vfs_dq_prealloc_space_nodirty(inode, nr))) mark_inode_dirty(inode); return ret; } -static inline int DQUOT_ALLOC_SPACE_NODIRTY(struct inode *inode, qsize_t nr) +static inline int vfs_dq_alloc_space_nodirty(struct inode *inode, qsize_t nr) { if (sb_any_quota_enabled(inode->i_sb)) { /* Used space is updated in alloc_space() */ @@ -122,25 +104,25 @@ static inline int DQUOT_ALLOC_SPACE_NODIRTY(struct inode *inode, qsize_t nr) return 0; } -static inline int DQUOT_ALLOC_SPACE(struct inode *inode, qsize_t nr) +static inline int vfs_dq_alloc_space(struct inode *inode, qsize_t nr) { int ret; - if (!(ret = DQUOT_ALLOC_SPACE_NODIRTY(inode, nr))) + if (!(ret = vfs_dq_alloc_space_nodirty(inode, nr))) mark_inode_dirty(inode); return ret; } -static inline int DQUOT_ALLOC_INODE(struct inode *inode) +static inline int vfs_dq_alloc_inode(struct inode *inode) { if (sb_any_quota_enabled(inode->i_sb)) { - DQUOT_INIT(inode); + vfs_dq_init(inode); if (inode->i_sb->dq_op->alloc_inode(inode, 1) == NO_QUOTA) return 1; } return 0; } -static inline void DQUOT_FREE_SPACE_NODIRTY(struct inode *inode, qsize_t nr) +static inline void vfs_dq_free_space_nodirty(struct inode *inode, qsize_t nr) { if (sb_any_quota_enabled(inode->i_sb)) inode->i_sb->dq_op->free_space(inode, nr); @@ -148,35 +130,25 @@ static inline void DQUOT_FREE_SPACE_NODIRTY(struct inode *inode, qsize_t nr) inode_sub_bytes(inode, nr); } -static inline void DQUOT_FREE_SPACE(struct inode *inode, qsize_t nr) +static inline void vfs_dq_free_space(struct inode *inode, qsize_t nr) { - DQUOT_FREE_SPACE_NODIRTY(inode, nr); + vfs_dq_free_space_nodirty(inode, nr); mark_inode_dirty(inode); } -static inline void DQUOT_FREE_INODE(struct inode *inode) +static inline void vfs_dq_free_inode(struct inode *inode) { if (sb_any_quota_enabled(inode->i_sb)) inode->i_sb->dq_op->free_inode(inode, 1); } -static inline int DQUOT_TRANSFER(struct inode *inode, struct iattr *iattr) -{ - if (sb_any_quota_enabled(inode->i_sb) && !IS_NOQUOTA(inode)) { - DQUOT_INIT(inode); - if (inode->i_sb->dq_op->transfer(inode, iattr) == NO_QUOTA) - return 1; - } - return 0; -} - /* The following two functions cannot be called inside a transaction */ -static inline void DQUOT_SYNC(struct super_block *sb) +static inline void vfs_dq_sync(struct super_block *sb) { sync_dquots(sb, -1); } -static inline int DQUOT_OFF(struct super_block *sb, int remount) +static inline int vfs_dq_off(struct super_block *sb, int remount) { int ret = -ENOSYS; @@ -185,21 +157,6 @@ static inline int DQUOT_OFF(struct super_block *sb, int remount) return ret; } -static inline int DQUOT_ON_REMOUNT(struct super_block *sb) -{ - int cnt; - int ret = 0, err; - - if (!sb->s_qcop || !sb->s_qcop->quota_on) - return -ENOSYS; - for (cnt = 0; cnt < MAXQUOTAS; cnt++) { - err = sb->s_qcop->quota_on(sb, cnt, 0, NULL, 1); - if (err < 0 && !ret) - ret = err; - } - return ret; -} - #else /* @@ -208,113 +165,144 @@ static inline int DQUOT_ON_REMOUNT(struct super_block *sb) #define sb_dquot_ops (NULL) #define sb_quotactl_ops (NULL) -static inline void DQUOT_INIT(struct inode *inode) +static inline void vfs_dq_init(struct inode *inode) { } -static inline void DQUOT_DROP(struct inode *inode) +static inline void vfs_dq_drop(struct inode *inode) { } -static inline int DQUOT_ALLOC_INODE(struct inode *inode) +static inline int vfs_dq_alloc_inode(struct inode *inode) { return 0; } -static inline void DQUOT_FREE_INODE(struct inode *inode) +static inline void vfs_dq_free_inode(struct inode *inode) { } -static inline void DQUOT_SYNC(struct super_block *sb) +static inline void vfs_dq_sync(struct super_block *sb) { } -static inline int DQUOT_OFF(struct super_block *sb, int remount) +static inline int vfs_dq_off(struct super_block *sb, int remount) { return 0; } -static inline int DQUOT_ON_REMOUNT(struct super_block *sb) +static inline int vfs_dq_quota_on_remount(struct super_block *sb) { return 0; } -static inline int DQUOT_TRANSFER(struct inode *inode, struct iattr *iattr) +static inline int vfs_dq_transfer(struct inode *inode, struct iattr *iattr) { return 0; } -static inline int DQUOT_PREALLOC_SPACE_NODIRTY(struct inode *inode, qsize_t nr) +static inline int vfs_dq_prealloc_space_nodirty(struct inode *inode, qsize_t nr) { inode_add_bytes(inode, nr); return 0; } -static inline int DQUOT_PREALLOC_SPACE(struct inode *inode, qsize_t nr) +static inline int vfs_dq_prealloc_space(struct inode *inode, qsize_t nr) { - DQUOT_PREALLOC_SPACE_NODIRTY(inode, nr); + vfs_dq_prealloc_space_nodirty(inode, nr); mark_inode_dirty(inode); return 0; } -static inline int DQUOT_ALLOC_SPACE_NODIRTY(struct inode *inode, qsize_t nr) +static inline int vfs_dq_alloc_space_nodirty(struct inode *inode, qsize_t nr) { inode_add_bytes(inode, nr); return 0; } -static inline int DQUOT_ALLOC_SPACE(struct inode *inode, qsize_t nr) +static inline int vfs_dq_alloc_space(struct inode *inode, qsize_t nr) { - DQUOT_ALLOC_SPACE_NODIRTY(inode, nr); + vfs_dq_alloc_space_nodirty(inode, nr); mark_inode_dirty(inode); return 0; } -static inline void DQUOT_FREE_SPACE_NODIRTY(struct inode *inode, qsize_t nr) +static inline void vfs_dq_free_space_nodirty(struct inode *inode, qsize_t nr) { inode_sub_bytes(inode, nr); } -static inline void DQUOT_FREE_SPACE(struct inode *inode, qsize_t nr) +static inline void vfs_dq_free_space(struct inode *inode, qsize_t nr) { - DQUOT_FREE_SPACE_NODIRTY(inode, nr); + vfs_dq_free_space_nodirty(inode, nr); mark_inode_dirty(inode); } #endif /* CONFIG_QUOTA */ -static inline int DQUOT_PREALLOC_BLOCK_NODIRTY(struct inode *inode, qsize_t nr) +static inline int vfs_dq_prealloc_block_nodirty(struct inode *inode, qsize_t nr) { - return DQUOT_PREALLOC_SPACE_NODIRTY(inode, + return vfs_dq_prealloc_space_nodirty(inode, nr << inode->i_sb->s_blocksize_bits); } -static inline int DQUOT_PREALLOC_BLOCK(struct inode *inode, qsize_t nr) +static inline int vfs_dq_prealloc_block(struct inode *inode, qsize_t nr) { - return DQUOT_PREALLOC_SPACE(inode, + return vfs_dq_prealloc_space(inode, nr << inode->i_sb->s_blocksize_bits); } -static inline int DQUOT_ALLOC_BLOCK_NODIRTY(struct inode *inode, qsize_t nr) +static inline int vfs_dq_alloc_block_nodirty(struct inode *inode, qsize_t nr) { - return DQUOT_ALLOC_SPACE_NODIRTY(inode, + return vfs_dq_alloc_space_nodirty(inode, nr << inode->i_sb->s_blocksize_bits); } -static inline int DQUOT_ALLOC_BLOCK(struct inode *inode, qsize_t nr) +static inline int vfs_dq_alloc_block(struct inode *inode, qsize_t nr) { - return DQUOT_ALLOC_SPACE(inode, + return vfs_dq_alloc_space(inode, nr << inode->i_sb->s_blocksize_bits); } -static inline void DQUOT_FREE_BLOCK_NODIRTY(struct inode *inode, qsize_t nr) +static inline void vfs_dq_free_block_nodirty(struct inode *inode, qsize_t nr) { - DQUOT_FREE_SPACE_NODIRTY(inode, nr << inode->i_sb->s_blocksize_bits); + vfs_dq_free_space_nodirty(inode, nr << inode->i_sb->s_blocksize_bits); } -static inline void DQUOT_FREE_BLOCK(struct inode *inode, qsize_t nr) +static inline void vfs_dq_free_block(struct inode *inode, qsize_t nr) { - DQUOT_FREE_SPACE(inode, nr << inode->i_sb->s_blocksize_bits); + vfs_dq_free_space(inode, nr << inode->i_sb->s_blocksize_bits); } +/* + * Define uppercase equivalents for compatibility with old function names + * Can go away when we think all users have been converted (15/04/2008) + */ +#define DQUOT_INIT(inode) vfs_dq_init(inode) +#define DQUOT_DROP(inode) vfs_dq_drop(inode) +#define DQUOT_PREALLOC_SPACE_NODIRTY(inode, nr) \ + vfs_dq_prealloc_space_nodirty(inode, nr) +#define DQUOT_PREALLOC_SPACE(inode, nr) vfs_dq_prealloc_space(inode, nr) +#define DQUOT_ALLOC_SPACE_NODIRTY(inode, nr) \ + vfs_dq_alloc_space_nodirty(inode, nr) +#define DQUOT_ALLOC_SPACE(inode, nr) vfs_dq_alloc_space(inode, nr) +#define DQUOT_PREALLOC_BLOCK_NODIRTY(inode, nr) \ + vfs_dq_prealloc_block_nodirty(inode, nr) +#define DQUOT_PREALLOC_BLOCK(inode, nr) vfs_dq_prealloc_block(inode, nr) +#define DQUOT_ALLOC_BLOCK_NODIRTY(inode, nr) \ + vfs_dq_alloc_block_nodirty(inode, nr) +#define DQUOT_ALLOC_BLOCK(inode, nr) vfs_dq_alloc_block(inode, nr) +#define DQUOT_ALLOC_INODE(inode) vfs_dq_alloc_inode(inode) +#define DQUOT_FREE_SPACE_NODIRTY(inode, nr) \ + vfs_dq_free_space_nodirty(inode, nr) +#define DQUOT_FREE_SPACE(inode, nr) vfs_dq_free_space(inode, nr) +#define DQUOT_FREE_BLOCK_NODIRTY(inode, nr) \ + vfs_dq_free_block_nodirty(inode, nr) +#define DQUOT_FREE_BLOCK(inode, nr) vfs_dq_free_block(inode, nr) +#define DQUOT_FREE_INODE(inode) vfs_dq_free_inode(inode) +#define DQUOT_TRANSFER(inode, iattr) vfs_dq_transfer(inode, iattr) +#define DQUOT_SYNC(sb) vfs_dq_sync(sb) +#define DQUOT_OFF(sb, remount) vfs_dq_off(sb, remount) +#define DQUOT_ON_REMOUNT(sb) vfs_dq_quota_on_remount(sb) + #endif /* _LINUX_QUOTAOPS_ */ -- cgit v1.2.3-70-g09d2 From 02a55ca87185e114e5d298a8d00608501dbabf67 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Fri, 25 Jul 2008 01:46:50 -0700 Subject: quota: cleanup loop in sync_dquots() Make loop in sync_dquots() checking whether there's something to write more readable, remove useless variable and macro info_any_dirty() which is used only in this place. Signed-off-by: Jan Kara Cc: "Vegard Nossum" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/quota.c | 18 ++++++++++++------ include/linux/quota.h | 2 -- 2 files changed, 12 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/quota.c b/fs/quota.c index db1cc9f3c7a..7f4386ebc23 100644 --- a/fs/quota.c +++ b/fs/quota.c @@ -186,7 +186,7 @@ static void quota_sync_sb(struct super_block *sb, int type) void sync_dquots(struct super_block *sb, int type) { - int cnt, dirty; + int cnt; if (sb) { if (sb->s_qcop->quota_sync) @@ -198,11 +198,17 @@ void sync_dquots(struct super_block *sb, int type) restart: list_for_each_entry(sb, &super_blocks, s_list) { /* This test just improves performance so it needn't be reliable... */ - for (cnt = 0, dirty = 0; cnt < MAXQUOTAS; cnt++) - if ((type == cnt || type == -1) && sb_has_quota_enabled(sb, cnt) - && info_any_dirty(&sb_dqopt(sb)->info[cnt])) - dirty = 1; - if (!dirty) + for (cnt = 0; cnt < MAXQUOTAS; cnt++) { + if (type != -1 && type != cnt) + continue; + if (!sb_has_quota_enabled(sb, cnt)) + continue; + if (!info_dirty(&sb_dqopt(sb)->info[cnt]) && + list_empty(&sb_dqopt(sb)->info[cnt].dqi_dirty_list)) + continue; + break; + } + if (cnt == MAXQUOTAS) continue; sb->s_count++; spin_unlock(&sb_lock); diff --git a/include/linux/quota.h b/include/linux/quota.h index dcddfb20094..6f1d97ddf82 100644 --- a/include/linux/quota.h +++ b/include/linux/quota.h @@ -224,8 +224,6 @@ struct super_block; extern void mark_info_dirty(struct super_block *sb, int type); #define info_dirty(info) test_bit(DQF_INFO_DIRTY_B, &(info)->dqi_flags) -#define info_any_dquot_dirty(info) (!list_empty(&(info)->dqi_dirty_list)) -#define info_any_dirty(info) (info_dirty(info) || info_any_dquot_dirty(info)) #define sb_dqopt(sb) (&(sb)->s_dquot) #define sb_dqinfo(sb, type) (sb_dqopt(sb)->info+(type)) -- cgit v1.2.3-70-g09d2 From 74abb9890dafb12a50dc140de215ed477beb1b88 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Fri, 25 Jul 2008 01:46:51 -0700 Subject: quota: move function-macros from quota.h to quotaops.h Move declarations of some macros, which should be in fact functions to quotaops.h. This way they can be later converted to inline functions because we can now use declarations from quota.h. Also add necessary includes of quotaops.h to a few files. [akpm@linux-foundation.org: fix JFS build] [akpm@linux-foundation.org: fix UFS build] [vegard.nossum@gmail.com: fix QUOTA=n build] Signed-off-by: Jan Kara Cc: Vegard Nossum Cc: Arjen Pool Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext2/super.c | 1 + fs/jfs/super.c | 1 + fs/quota_v1.c | 1 + fs/quota_v2.c | 1 + fs/reiserfs/super.c | 1 + fs/ufs/super.c | 1 + include/linux/quota.h | 22 +++------------------- include/linux/quotaops.h | 26 ++++++++++++++++++++++++++ 8 files changed, 35 insertions(+), 19 deletions(-) (limited to 'fs') diff --git a/fs/ext2/super.c b/fs/ext2/super.c index ef50cbc792d..31308a3b0b8 100644 --- a/fs/ext2/super.c +++ b/fs/ext2/super.c @@ -31,6 +31,7 @@ #include #include #include +#include #include #include "ext2.h" #include "xattr.h" diff --git a/fs/jfs/super.c b/fs/jfs/super.c index 0288e6d7936..359c091d896 100644 --- a/fs/jfs/super.c +++ b/fs/jfs/super.c @@ -22,6 +22,7 @@ #include #include #include +#include #include #include #include diff --git a/fs/quota_v1.c b/fs/quota_v1.c index a6cf9269105..5ae15b13eeb 100644 --- a/fs/quota_v1.c +++ b/fs/quota_v1.c @@ -1,6 +1,7 @@ #include #include #include +#include #include #include #include diff --git a/fs/quota_v2.c b/fs/quota_v2.c index 234ada90363..b53827dc02d 100644 --- a/fs/quota_v2.c +++ b/fs/quota_v2.c @@ -11,6 +11,7 @@ #include #include #include +#include #include diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c index a10a6d2a887..2ec748ba0bd 100644 --- a/fs/reiserfs/super.c +++ b/fs/reiserfs/super.c @@ -22,6 +22,7 @@ #include #include #include +#include #include #include #include diff --git a/fs/ufs/super.c b/fs/ufs/super.c index 506f724055c..227c9d70004 100644 --- a/fs/ufs/super.c +++ b/fs/ufs/super.c @@ -76,6 +76,7 @@ #include #include +#include #include #include #include diff --git a/include/linux/quota.h b/include/linux/quota.h index 6f1d97ddf82..f9983ea0ff8 100644 --- a/include/linux/quota.h +++ b/include/linux/quota.h @@ -41,9 +41,6 @@ #define __DQUOT_VERSION__ "dquot_6.5.1" #define __DQUOT_NUM_VERSION__ 6*10000+5*100+1 -typedef __kernel_uid32_t qid_t; /* Type in which we store ids in memory */ -typedef __u64 qsize_t; /* Type in which we store sizes */ - /* Size of blocks in which are counted size limits */ #define QUOTABLOCK_BITS 10 #define QUOTABLOCK_SIZE (1 << QUOTABLOCK_BITS) @@ -172,6 +169,9 @@ enum { #include +typedef __kernel_uid32_t qid_t; /* Type in which we store ids in memory */ +typedef __u64 qsize_t; /* Type in which we store sizes */ + extern spinlock_t dq_data_lock; /* Maximal numbers of writes for quota operation (insert/delete/update) @@ -225,9 +225,6 @@ struct super_block; extern void mark_info_dirty(struct super_block *sb, int type); #define info_dirty(info) test_bit(DQF_INFO_DIRTY_B, &(info)->dqi_flags) -#define sb_dqopt(sb) (&(sb)->s_dquot) -#define sb_dqinfo(sb, type) (sb_dqopt(sb)->info+(type)) - struct dqstats { int lookups; int drops; @@ -335,19 +332,6 @@ struct quota_info { struct quota_format_ops *ops[MAXQUOTAS]; /* Operations for each type */ }; -#define sb_has_quota_enabled(sb, type) ((type)==USRQUOTA ? \ - (sb_dqopt(sb)->flags & DQUOT_USR_ENABLED) : (sb_dqopt(sb)->flags & DQUOT_GRP_ENABLED)) - -#define sb_any_quota_enabled(sb) (sb_has_quota_enabled(sb, USRQUOTA) | \ - sb_has_quota_enabled(sb, GRPQUOTA)) - -#define sb_has_quota_suspended(sb, type) \ - ((type) == USRQUOTA ? (sb_dqopt(sb)->flags & DQUOT_USR_SUSPENDED) : \ - (sb_dqopt(sb)->flags & DQUOT_GRP_SUSPENDED)) - -#define sb_any_quota_suspended(sb) (sb_has_quota_suspended(sb, USRQUOTA) | \ - sb_has_quota_suspended(sb, GRPQUOTA)) - int register_quota_format(struct quota_format_type *fmt); void unregister_quota_format(struct quota_format_type *fmt); diff --git a/include/linux/quotaops.h b/include/linux/quotaops.h index 0c8f9fe462a..38218c1334b 100644 --- a/include/linux/quotaops.h +++ b/include/linux/quotaops.h @@ -14,6 +14,8 @@ #include +#define sb_dqopt(sb) (&(sb)->s_dquot) + #if defined(CONFIG_QUOTA) /* @@ -52,6 +54,25 @@ void vfs_dq_drop(struct inode *inode); int vfs_dq_transfer(struct inode *inode, struct iattr *iattr); int vfs_dq_quota_on_remount(struct super_block *sb); +#define sb_dqinfo(sb, type) (sb_dqopt(sb)->info+(type)) + +/* + * Functions for checking status of quota + */ + +#define sb_has_quota_enabled(sb, type) ((type)==USRQUOTA ? \ + (sb_dqopt(sb)->flags & DQUOT_USR_ENABLED) : (sb_dqopt(sb)->flags & DQUOT_GRP_ENABLED)) + +#define sb_any_quota_enabled(sb) (sb_has_quota_enabled(sb, USRQUOTA) | \ + sb_has_quota_enabled(sb, GRPQUOTA)) + +#define sb_has_quota_suspended(sb, type) \ + ((type) == USRQUOTA ? (sb_dqopt(sb)->flags & DQUOT_USR_SUSPENDED) : \ + (sb_dqopt(sb)->flags & DQUOT_GRP_SUSPENDED)) + +#define sb_any_quota_suspended(sb) (sb_has_quota_suspended(sb, USRQUOTA) | \ + sb_has_quota_suspended(sb, GRPQUOTA)) + /* * Operations supported for diskquotas. */ @@ -159,6 +180,11 @@ static inline int vfs_dq_off(struct super_block *sb, int remount) #else +#define sb_has_quota_enabled(sb, type) 0 +#define sb_any_quota_enabled(sb) 0 +#define sb_has_quota_suspended(sb, type) 0 +#define sb_any_quota_suspended(sb) 0 + /* * NO-OP when quota not configured. */ -- cgit v1.2.3-70-g09d2 From 657d3bfa98e542271b449f8cd84c7501ae2b2255 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Fri, 25 Jul 2008 01:46:52 -0700 Subject: quota: implement sending information via netlink about user below quota Sometimes it may be useful for userspace to know (e.g. for some hosting guys) that some user stopped exceeding his hardlimit or softlimit in quotas. Implement sending of such events to userspace via quota netlink protocol so that they don't have to poll for such events. Based on idea and initial implementation by Vladislav Bogdanov. Cc: Vladislav Bogdanov Signed-off-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/dquot.c | 60 +++++++++++++++++++++++++++++++++++++++++++++------ include/linux/quota.h | 4 ++++ 2 files changed, 58 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/dquot.c b/fs/dquot.c index 0bcaf970bbb..1346eebe74c 100644 --- a/fs/dquot.c +++ b/fs/dquot.c @@ -889,7 +889,10 @@ static void print_warning(struct dquot *dquot, const int warntype) char *msg = NULL; struct tty_struct *tty; - if (!need_print_warning(dquot)) + if (warntype == QUOTA_NL_IHARDBELOW || + warntype == QUOTA_NL_ISOFTBELOW || + warntype == QUOTA_NL_BHARDBELOW || + warntype == QUOTA_NL_BSOFTBELOW || !need_print_warning(dquot)) return; mutex_lock(&tty_mutex); @@ -1097,6 +1100,35 @@ static int check_bdq(struct dquot *dquot, qsize_t space, int prealloc, char *war return QUOTA_OK; } +static int info_idq_free(struct dquot *dquot, ulong inodes) +{ + if (test_bit(DQ_FAKE_B, &dquot->dq_flags) || + dquot->dq_dqb.dqb_curinodes <= dquot->dq_dqb.dqb_isoftlimit) + return QUOTA_NL_NOWARN; + + if (dquot->dq_dqb.dqb_curinodes - inodes <= dquot->dq_dqb.dqb_isoftlimit) + return QUOTA_NL_ISOFTBELOW; + if (dquot->dq_dqb.dqb_curinodes >= dquot->dq_dqb.dqb_ihardlimit && + dquot->dq_dqb.dqb_curinodes - inodes < dquot->dq_dqb.dqb_ihardlimit) + return QUOTA_NL_IHARDBELOW; + return QUOTA_NL_NOWARN; +} + +static int info_bdq_free(struct dquot *dquot, qsize_t space) +{ + if (test_bit(DQ_FAKE_B, &dquot->dq_flags) || + toqb(dquot->dq_dqb.dqb_curspace) <= dquot->dq_dqb.dqb_bsoftlimit) + return QUOTA_NL_NOWARN; + + if (toqb(dquot->dq_dqb.dqb_curspace - space) <= + dquot->dq_dqb.dqb_bsoftlimit) + return QUOTA_NL_BSOFTBELOW; + if (toqb(dquot->dq_dqb.dqb_curspace) >= dquot->dq_dqb.dqb_bhardlimit && + toqb(dquot->dq_dqb.dqb_curspace - space) < + dquot->dq_dqb.dqb_bhardlimit) + return QUOTA_NL_BHARDBELOW; + return QUOTA_NL_NOWARN; +} /* * Initialize quota pointers in inode * Transaction must be started at entry @@ -1284,6 +1316,7 @@ warn_put_all: int dquot_free_space(struct inode *inode, qsize_t number) { unsigned int cnt; + char warntype[MAXQUOTAS]; /* First test before acquiring mutex - solves deadlocks when we * re-enter the quota code and are already holding the mutex */ @@ -1292,6 +1325,7 @@ out_sub: inode_sub_bytes(inode, number); return QUOTA_OK; } + down_read(&sb_dqopt(inode->i_sb)->dqptr_sem); /* Now recheck reliably when holding dqptr_sem */ if (IS_NOQUOTA(inode)) { @@ -1302,6 +1336,7 @@ out_sub: for (cnt = 0; cnt < MAXQUOTAS; cnt++) { if (inode->i_dquot[cnt] == NODQUOT) continue; + warntype[cnt] = info_bdq_free(inode->i_dquot[cnt], number); dquot_decr_space(inode->i_dquot[cnt], number); } inode_sub_bytes(inode, number); @@ -1310,6 +1345,7 @@ out_sub: for (cnt = 0; cnt < MAXQUOTAS; cnt++) if (inode->i_dquot[cnt]) mark_dquot_dirty(inode->i_dquot[cnt]); + flush_warnings(inode->i_dquot, warntype); up_read(&sb_dqopt(inode->i_sb)->dqptr_sem); return QUOTA_OK; } @@ -1320,11 +1356,13 @@ out_sub: int dquot_free_inode(const struct inode *inode, unsigned long number) { unsigned int cnt; + char warntype[MAXQUOTAS]; /* First test before acquiring mutex - solves deadlocks when we * re-enter the quota code and are already holding the mutex */ if (IS_NOQUOTA(inode)) return QUOTA_OK; + down_read(&sb_dqopt(inode->i_sb)->dqptr_sem); /* Now recheck reliably when holding dqptr_sem */ if (IS_NOQUOTA(inode)) { @@ -1335,6 +1373,7 @@ int dquot_free_inode(const struct inode *inode, unsigned long number) for (cnt = 0; cnt < MAXQUOTAS; cnt++) { if (inode->i_dquot[cnt] == NODQUOT) continue; + warntype[cnt] = info_idq_free(inode->i_dquot[cnt], number); dquot_decr_inodes(inode->i_dquot[cnt], number); } spin_unlock(&dq_data_lock); @@ -1342,6 +1381,7 @@ int dquot_free_inode(const struct inode *inode, unsigned long number) for (cnt = 0; cnt < MAXQUOTAS; cnt++) if (inode->i_dquot[cnt]) mark_dquot_dirty(inode->i_dquot[cnt]); + flush_warnings(inode->i_dquot, warntype); up_read(&sb_dqopt(inode->i_sb)->dqptr_sem); return QUOTA_OK; } @@ -1359,7 +1399,8 @@ int dquot_transfer(struct inode *inode, struct iattr *iattr) struct dquot *transfer_to[MAXQUOTAS]; int cnt, ret = NO_QUOTA, chuid = (iattr->ia_valid & ATTR_UID) && inode->i_uid != iattr->ia_uid, chgid = (iattr->ia_valid & ATTR_GID) && inode->i_gid != iattr->ia_gid; - char warntype[MAXQUOTAS]; + char warntype_to[MAXQUOTAS]; + char warntype_from_inodes[MAXQUOTAS], warntype_from_space[MAXQUOTAS]; /* First test before acquiring mutex - solves deadlocks when we * re-enter the quota code and are already holding the mutex */ @@ -1368,7 +1409,7 @@ int dquot_transfer(struct inode *inode, struct iattr *iattr) /* Clear the arrays */ for (cnt = 0; cnt < MAXQUOTAS; cnt++) { transfer_to[cnt] = transfer_from[cnt] = NODQUOT; - warntype[cnt] = QUOTA_NL_NOWARN; + warntype_to[cnt] = QUOTA_NL_NOWARN; } down_write(&sb_dqopt(inode->i_sb)->dqptr_sem); /* Now recheck reliably when holding dqptr_sem */ @@ -1400,8 +1441,9 @@ int dquot_transfer(struct inode *inode, struct iattr *iattr) if (transfer_to[cnt] == NODQUOT) continue; transfer_from[cnt] = inode->i_dquot[cnt]; - if (check_idq(transfer_to[cnt], 1, warntype+cnt) == NO_QUOTA || - check_bdq(transfer_to[cnt], space, 0, warntype+cnt) == NO_QUOTA) + if (check_idq(transfer_to[cnt], 1, warntype_to + cnt) == + NO_QUOTA || check_bdq(transfer_to[cnt], space, 0, + warntype_to + cnt) == NO_QUOTA) goto warn_put_all; } @@ -1417,6 +1459,10 @@ int dquot_transfer(struct inode *inode, struct iattr *iattr) /* Due to IO error we might not have transfer_from[] structure */ if (transfer_from[cnt]) { + warntype_from_inodes[cnt] = + info_idq_free(transfer_from[cnt], 1); + warntype_from_space[cnt] = + info_bdq_free(transfer_from[cnt], space); dquot_decr_inodes(transfer_from[cnt], 1); dquot_decr_space(transfer_from[cnt], space); } @@ -1436,7 +1482,9 @@ warn_put_all: if (transfer_to[cnt]) mark_dquot_dirty(transfer_to[cnt]); } - flush_warnings(transfer_to, warntype); + flush_warnings(transfer_to, warntype_to); + flush_warnings(transfer_from, warntype_from_inodes); + flush_warnings(transfer_from, warntype_from_space); for (cnt = 0; cnt < MAXQUOTAS; cnt++) { if (ret == QUOTA_OK && transfer_from[cnt] != NODQUOT) diff --git a/include/linux/quota.h b/include/linux/quota.h index 4e004fef813..376a05048bc 100644 --- a/include/linux/quota.h +++ b/include/linux/quota.h @@ -135,6 +135,10 @@ struct if_dqinfo { #define QUOTA_NL_BHARDWARN 4 /* Block hardlimit reached */ #define QUOTA_NL_BSOFTLONGWARN 5 /* Block grace time expired */ #define QUOTA_NL_BSOFTWARN 6 /* Block softlimit reached */ +#define QUOTA_NL_IHARDBELOW 7 /* Usage got below inode hardlimit */ +#define QUOTA_NL_ISOFTBELOW 8 /* Usage got below inode softlimit */ +#define QUOTA_NL_BHARDBELOW 9 /* Usage got below block hardlimit */ +#define QUOTA_NL_BSOFTBELOW 10 /* Usage got below block softlimit */ enum { QUOTA_NL_C_UNSPEC, -- cgit v1.2.3-70-g09d2 From e4901f92a8dbe843e76651a50f7a2a6dd3d53474 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:31 -0700 Subject: coredump: zap_threads: comments && use while_each_thread() No changes in fs/exec.o The for_each_process() loop in zap_threads() is very subtle, it is not clear why we don't race with fork/exit/exec. Add the fat comment. Also, change the code to use while_each_thread(). Signed-off-by: Oleg Nesterov Acked-by: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/exec.c | 39 ++++++++++++++++++++++++++++++++------- 1 file changed, 32 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/exec.c b/fs/exec.c index e41aef0fb35..af249af4cca 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -1517,7 +1517,7 @@ static void zap_process(struct task_struct *start) sigaddset(&t->pending.signal, SIGKILL); signal_wake_up(t, 1); } - } while ((t = next_thread(t)) != start); + } while_each_thread(start, t); } static inline int zap_threads(struct task_struct *tsk, struct mm_struct *mm, @@ -1539,7 +1539,36 @@ static inline int zap_threads(struct task_struct *tsk, struct mm_struct *mm, if (atomic_read(&mm->mm_users) == mm->core_waiters + 1) goto done; - + /* + * We should find and kill all tasks which use this mm, and we should + * count them correctly into mm->core_waiters. We don't take tasklist + * lock, but this is safe wrt: + * + * fork: + * None of sub-threads can fork after zap_process(leader). All + * processes which were created before this point should be + * visible to zap_threads() because copy_process() adds the new + * process to the tail of init_task.tasks list, and lock/unlock + * of ->siglock provides a memory barrier. + * + * do_exit: + * The caller holds mm->mmap_sem. This means that the task which + * uses this mm can't pass exit_mm(), so it can't exit or clear + * its ->mm. + * + * de_thread: + * It does list_replace_rcu(&leader->tasks, ¤t->tasks), + * we must see either old or new leader, this does not matter. + * However, it can change p->sighand, so lock_task_sighand(p) + * must be used. Since p->mm != NULL and we hold ->mmap_sem + * it can't fail. + * + * Note also that "g" can be the old leader with ->mm == NULL + * and already unhashed and thus removed from ->thread_group. + * This is OK, __unhash_process()->list_del_rcu() does not + * clear the ->next pointer, we will find the new leader via + * next_thread(). + */ rcu_read_lock(); for_each_process(g) { if (g == tsk->group_leader) @@ -1549,17 +1578,13 @@ static inline int zap_threads(struct task_struct *tsk, struct mm_struct *mm, do { if (p->mm) { if (p->mm == mm) { - /* - * p->sighand can't disappear, but - * may be changed by de_thread() - */ lock_task_sighand(p, &flags); zap_process(p); unlock_task_sighand(p, &flags); } break; } - } while ((p = next_thread(p)) != g); + } while_each_thread(g, p); } rcu_read_unlock(); done: -- cgit v1.2.3-70-g09d2 From 7b34e4283c685f5cc6ba6d30e939906eee0d4bcf Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:37 -0700 Subject: introduce PF_KTHREAD flag Introduce the new PF_KTHREAD flag to mark the kernel threads. It is set by INIT_TASK() and copied to the forked childs (we could set it in kthreadd() along with PF_NOFREEZE instead). daemonize() was changed as well. In that case testing of PF_KTHREAD is racy, but daemonize() is hopeless anyway. This flag is cleared in do_execve(), before search_binary_handler(). Probably not the best place, we can do this in exec_mmap() or in start_thread(), or clear it along with PF_FORKNOEXEC. But I think this doesn't matter in practice, and if do_execve() fails kthread should die soon. Signed-off-by: Oleg Nesterov Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/exec.c | 1 + include/linux/init_task.h | 2 +- include/linux/sched.h | 1 + kernel/exit.c | 2 +- 4 files changed, 4 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/exec.c b/fs/exec.c index af249af4cca..cd2e8c9b124 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -1326,6 +1326,7 @@ int do_execve(char * filename, if (retval < 0) goto out; + current->flags &= ~PF_KTHREAD; retval = search_binary_handler(bprm,regs); if (retval >= 0) { /* execve success */ diff --git a/include/linux/init_task.h b/include/linux/init_task.h index 93c45acf249..021d8e720c7 100644 --- a/include/linux/init_task.h +++ b/include/linux/init_task.h @@ -122,7 +122,7 @@ extern struct group_info init_groups; .state = 0, \ .stack = &init_thread_info, \ .usage = ATOMIC_INIT(2), \ - .flags = 0, \ + .flags = PF_KTHREAD, \ .lock_depth = -1, \ .prio = MAX_PRIO-20, \ .static_prio = MAX_PRIO-20, \ diff --git a/include/linux/sched.h b/include/linux/sched.h index 79e749dbf81..eec64a4adb9 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1483,6 +1483,7 @@ static inline void put_task_struct(struct task_struct *t) #define PF_EXITING 0x00000004 /* getting shut down */ #define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */ #define PF_VCPU 0x00000010 /* I'm a virtual CPU */ +#define PF_KTHREAD 0x00000020 /* I am a kernel thread */ #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */ #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */ #define PF_DUMPCORE 0x00000200 /* dumped core */ diff --git a/kernel/exit.c b/kernel/exit.c index a7799d8a640..28a44a2612d 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -430,7 +430,7 @@ void daemonize(const char *name, ...) * We don't want to have TIF_FREEZE set if the system-wide hibernation * or suspend transition begins right now. */ - current->flags |= PF_NOFREEZE; + current->flags |= (PF_NOFREEZE | PF_KTHREAD); if (current->nsproxy != &init_nsproxy) { get_nsproxy(&init_nsproxy); -- cgit v1.2.3-70-g09d2 From 246bb0b1deb29726990620d8b5e55ca29f331362 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:38 -0700 Subject: kill PF_BORROWED_MM in favour of PF_KTHREAD Kill PF_BORROWED_MM. Change use_mm/unuse_mm to not play with ->flags, and do s/PF_BORROWED_MM/PF_KTHREAD/ for a couple of other users. No functional changes yet. But this allows us to do further fixes/cleanups. oom_kill/ptrace/etc often check "p->mm != NULL" to filter out the kthreads, this is wrong because of use_mm(). The problem with PF_BORROWED_MM is that we need task_lock() to avoid races. With this patch we can check PF_KTHREAD directly, or use a simple lockless helper: /* The result must not be dereferenced !!! */ struct mm_struct *__get_task_mm(struct task_struct *tsk) { if (tsk->flags & PF_KTHREAD) return NULL; return tsk->mm; } Note also ecard_task(). It runs with ->mm != NULL, but it's the kernel thread without PF_BORROWED_MM. Signed-off-by: Oleg Nesterov Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/aio.c | 2 -- include/linux/sched.h | 3 +-- kernel/fork.c | 4 ++-- 3 files changed, 3 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/aio.c b/fs/aio.c index 0fb3117ddd9..0051fd94b44 100644 --- a/fs/aio.c +++ b/fs/aio.c @@ -586,7 +586,6 @@ static void use_mm(struct mm_struct *mm) struct task_struct *tsk = current; task_lock(tsk); - tsk->flags |= PF_BORROWED_MM; active_mm = tsk->active_mm; atomic_inc(&mm->mm_count); tsk->mm = mm; @@ -610,7 +609,6 @@ static void unuse_mm(struct mm_struct *mm) struct task_struct *tsk = current; task_lock(tsk); - tsk->flags &= ~PF_BORROWED_MM; tsk->mm = NULL; /* active_mm is still 'mm' */ enter_lazy_tlb(mm, tsk); diff --git a/include/linux/sched.h b/include/linux/sched.h index eec64a4adb9..0560999eb1d 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1483,7 +1483,6 @@ static inline void put_task_struct(struct task_struct *t) #define PF_EXITING 0x00000004 /* getting shut down */ #define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */ #define PF_VCPU 0x00000010 /* I'm a virtual CPU */ -#define PF_KTHREAD 0x00000020 /* I am a kernel thread */ #define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */ #define PF_SUPERPRIV 0x00000100 /* used super-user privileges */ #define PF_DUMPCORE 0x00000200 /* dumped core */ @@ -1497,7 +1496,7 @@ static inline void put_task_struct(struct task_struct *t) #define PF_KSWAPD 0x00040000 /* I am kswapd */ #define PF_SWAPOFF 0x00080000 /* I am in swapoff */ #define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */ -#define PF_BORROWED_MM 0x00200000 /* I am a kthread doing use_mm */ +#define PF_KTHREAD 0x00200000 /* I am a kernel thread */ #define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */ #define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */ #define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */ diff --git a/kernel/fork.c b/kernel/fork.c index 228f80c9155..eeaec6893b0 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -474,7 +474,7 @@ EXPORT_SYMBOL_GPL(mmput); /** * get_task_mm - acquire a reference to the task's mm * - * Returns %NULL if the task has no mm. Checks PF_BORROWED_MM (meaning + * Returns %NULL if the task has no mm. Checks PF_KTHREAD (meaning * this kernel workthread has transiently adopted a user mm with use_mm, * to do its AIO) is not set and if so returns a reference to it, after * bumping up the use count. User must release the mm via mmput() @@ -487,7 +487,7 @@ struct mm_struct *get_task_mm(struct task_struct *task) task_lock(task); mm = task->mm; if (mm) { - if (task->flags & PF_BORROWED_MM) + if (task->flags & PF_KTHREAD) mm = NULL; else atomic_inc(&mm->mm_users); -- cgit v1.2.3-70-g09d2 From 15b9f360c0316c06d37c09b02d85565edbaf9dd3 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:39 -0700 Subject: coredump: zap_threads() must skip kernel threads The main loop in zap_threads() must skip kthreads which may use the same mm. Otherwise we "kill" this thread erroneously (for example, it can not fork or exec after that), and the coredumping task stucks in the TASK_UNINTERRUPTIBLE state forever because of the wrong ->core_waiters count. Signed-off-by: Oleg Nesterov Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/exec.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/exec.c b/fs/exec.c index cd2e8c9b124..e347e6ed161 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -1574,11 +1574,12 @@ static inline int zap_threads(struct task_struct *tsk, struct mm_struct *mm, for_each_process(g) { if (g == tsk->group_leader) continue; - + if (g->flags & PF_KTHREAD) + continue; p = g; do { if (p->mm) { - if (p->mm == mm) { + if (unlikely(p->mm == mm)) { lock_task_sighand(p, &flags); zap_process(p); unlock_task_sighand(p, &flags); -- cgit v1.2.3-70-g09d2 From 24d5288f06ed8b3a368eba967d587cdb012dfdf7 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:40 -0700 Subject: coredump: elf_core_dump: skip kernel threads linux_binfmt->core_dump() runs before the process does exit_aio(), this means that we can hit the kernel thread which shares the same ->mm. Afaics, nothing really bad can happen, but perhaps it makes sense to fix this minor bug. It is sad we have to iterate over all threads in system and use GFP_ATOMIC. Hopefully we can kill theses ugly do_each_thread()s, but this needs some nontrivial changes in mm_struct and do_coredump. Signed-off-by: Oleg Nesterov Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/binfmt_elf.c | 6 ++++++ fs/binfmt_elf_fdpic.c | 3 +++ 2 files changed, 9 insertions(+) (limited to 'fs') diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c index 639d2d8b571..bad7d8770d7 100644 --- a/fs/binfmt_elf.c +++ b/fs/binfmt_elf.c @@ -1520,6 +1520,9 @@ static int fill_note_info(struct elfhdr *elf, int phdrs, rcu_read_lock(); do_each_thread(g, p) if (p->mm == dump_task->mm) { + if (p->flags & PF_KTHREAD) + continue; + t = kzalloc(offsetof(struct elf_thread_core_info, notes[info->thread_notes]), GFP_ATOMIC); @@ -1724,6 +1727,9 @@ static int fill_note_info(struct elfhdr *elf, int phdrs, rcu_read_lock(); do_each_thread(g, p) if (current->mm == p->mm && current != p) { + if (p->flags & PF_KTHREAD) + continue; + ets = kzalloc(sizeof(*ets), GFP_ATOMIC); if (!ets) { rcu_read_unlock(); diff --git a/fs/binfmt_elf_fdpic.c b/fs/binfmt_elf_fdpic.c index d051a32e627..71bcc4b4d08 100644 --- a/fs/binfmt_elf_fdpic.c +++ b/fs/binfmt_elf_fdpic.c @@ -1626,6 +1626,9 @@ static int elf_fdpic_core_dump(long signr, struct pt_regs *regs, rcu_read_lock(); do_each_thread(g,p) if (current->mm == p->mm && current != p) { + if (p->flags & PF_KTHREAD) + continue; + tmp = kzalloc(sizeof(*tmp), GFP_ATOMIC); if (!tmp) { rcu_read_unlock(); -- cgit v1.2.3-70-g09d2 From 32ecb1f26dd50eeaac4e3f4dea4541c97848e459 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:41 -0700 Subject: coredump: turn mm->core_startup_done into the pointer to struct core_state mm->core_startup_done points to "struct completion startup_done" allocated on the coredump_wait()'s stack. Introduce the new structure, core_state, which holds this "struct completion". This way we can add more info visible to the threads participating in coredump without enlarging mm_struct. No changes in affected .o files. Signed-off-by: Oleg Nesterov Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/exec.c | 8 ++++---- include/linux/mm_types.h | 7 ++++++- kernel/exit.c | 2 +- 3 files changed, 11 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/exec.c b/fs/exec.c index e347e6ed161..71734568f01 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -1597,13 +1597,13 @@ static int coredump_wait(int exit_code) { struct task_struct *tsk = current; struct mm_struct *mm = tsk->mm; - struct completion startup_done; + struct core_state core_state; struct completion *vfork_done; int core_waiters; init_completion(&mm->core_done); - init_completion(&startup_done); - mm->core_startup_done = &startup_done; + init_completion(&core_state.startup); + mm->core_state = &core_state; core_waiters = zap_threads(tsk, mm, exit_code); up_write(&mm->mmap_sem); @@ -1622,7 +1622,7 @@ static int coredump_wait(int exit_code) } if (core_waiters) - wait_for_completion(&startup_done); + wait_for_completion(&core_state.startup); fail: BUG_ON(mm->core_waiters); return core_waiters; diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index 02a27ae7853..97819efd233 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -159,6 +159,10 @@ struct vm_area_struct { #endif }; +struct core_state { + struct completion startup; +}; + struct mm_struct { struct vm_area_struct * mmap; /* list of VMAs */ struct rb_root mm_rb; @@ -220,7 +224,8 @@ struct mm_struct { unsigned long flags; /* Must use atomic bitops to access the bits */ /* coredumping support */ - struct completion *core_startup_done, core_done; + struct core_state *core_state; + struct completion core_done; /* aio bits */ rwlock_t ioctx_list_lock; /* aio lock */ diff --git a/kernel/exit.c b/kernel/exit.c index 28a44a2612d..f7fa21dbced 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -680,7 +680,7 @@ static void exit_mm(struct task_struct * tsk) up_read(&mm->mmap_sem); down_write(&mm->mmap_sem); if (!--mm->core_waiters) - complete(mm->core_startup_done); + complete(&mm->core_state->startup); up_write(&mm->mmap_sem); wait_for_completion(&mm->core_done); -- cgit v1.2.3-70-g09d2 From 999d9fc1670bc082928b93b11d1f2e0e417d973c Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:41 -0700 Subject: coredump: move mm->core_waiters into struct core_state Move mm->core_waiters into "struct core_state" allocated on stack. This shrinks mm_struct a little bit and allows further changes. This patch mostly does s/core_waiters/core_state. The only essential change is that coredump_wait() must clear mm->core_state before return. The coredump_wait()'s path is uglified and .text grows by 30 bytes, this is fixed by the next patch. Signed-off-by: Oleg Nesterov Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/exec.c | 21 +++++++++++---------- include/linux/mm_types.h | 2 +- kernel/exit.c | 8 ++++---- kernel/fork.c | 2 +- kernel/signal.c | 4 ++-- 5 files changed, 19 insertions(+), 18 deletions(-) (limited to 'fs') diff --git a/fs/exec.c b/fs/exec.c index 71734568f01..50de3aaff4d 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -722,12 +722,10 @@ static int exec_mmap(struct mm_struct *mm) * Make sure that if there is a core dump in progress * for the old mm, we get out and die instead of going * through with the exec. We must hold mmap_sem around - * checking core_waiters and changing tsk->mm. The - * core-inducing thread will increment core_waiters for - * each thread whose ->mm == old_mm. + * checking core_state and changing tsk->mm. */ down_read(&old_mm->mmap_sem); - if (unlikely(old_mm->core_waiters)) { + if (unlikely(old_mm->core_state)) { up_read(&old_mm->mmap_sem); return -EINTR; } @@ -1514,7 +1512,7 @@ static void zap_process(struct task_struct *start) t = start; do { if (t != current && t->mm) { - t->mm->core_waiters++; + t->mm->core_state->nr_threads++; sigaddset(&t->pending.signal, SIGKILL); signal_wake_up(t, 1); } @@ -1538,11 +1536,11 @@ static inline int zap_threads(struct task_struct *tsk, struct mm_struct *mm, if (err) return err; - if (atomic_read(&mm->mm_users) == mm->core_waiters + 1) + if (atomic_read(&mm->mm_users) == mm->core_state->nr_threads + 1) goto done; /* * We should find and kill all tasks which use this mm, and we should - * count them correctly into mm->core_waiters. We don't take tasklist + * count them correctly into ->nr_threads. We don't take tasklist * lock, but this is safe wrt: * * fork: @@ -1590,7 +1588,7 @@ static inline int zap_threads(struct task_struct *tsk, struct mm_struct *mm, } rcu_read_unlock(); done: - return mm->core_waiters; + return mm->core_state->nr_threads; } static int coredump_wait(int exit_code) @@ -1603,9 +1601,12 @@ static int coredump_wait(int exit_code) init_completion(&mm->core_done); init_completion(&core_state.startup); + core_state.nr_threads = 0; mm->core_state = &core_state; core_waiters = zap_threads(tsk, mm, exit_code); + if (core_waiters < 0) + mm->core_state = NULL; up_write(&mm->mmap_sem); if (unlikely(core_waiters < 0)) @@ -1623,8 +1624,8 @@ static int coredump_wait(int exit_code) if (core_waiters) wait_for_completion(&core_state.startup); + mm->core_state = NULL; fail: - BUG_ON(mm->core_waiters); return core_waiters; } @@ -1702,7 +1703,7 @@ int do_coredump(long signr, int exit_code, struct pt_regs * regs) /* * If another thread got here first, or we are not dumpable, bail out. */ - if (mm->core_waiters || !get_dumpable(mm)) { + if (mm->core_state || !get_dumpable(mm)) { up_write(&mm->mmap_sem); goto fail; } diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index 97819efd233..c0b1747b61a 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -160,6 +160,7 @@ struct vm_area_struct { }; struct core_state { + int nr_threads; struct completion startup; }; @@ -179,7 +180,6 @@ struct mm_struct { atomic_t mm_users; /* How many users with user space? */ atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */ int map_count; /* number of VMAs */ - int core_waiters; struct rw_semaphore mmap_sem; spinlock_t page_table_lock; /* Protects page tables and some counters */ diff --git a/kernel/exit.c b/kernel/exit.c index f7fa21dbced..988e232254e 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -670,16 +670,16 @@ static void exit_mm(struct task_struct * tsk) return; /* * Serialize with any possible pending coredump. - * We must hold mmap_sem around checking core_waiters + * We must hold mmap_sem around checking core_state * and clearing tsk->mm. The core-inducing thread - * will increment core_waiters for each thread in the + * will increment ->nr_threads for each thread in the * group with ->mm != NULL. */ down_read(&mm->mmap_sem); - if (mm->core_waiters) { + if (mm->core_state) { up_read(&mm->mmap_sem); down_write(&mm->mmap_sem); - if (!--mm->core_waiters) + if (!--mm->core_state->nr_threads) complete(&mm->core_state->startup); up_write(&mm->mmap_sem); diff --git a/kernel/fork.c b/kernel/fork.c index eeaec6893b0..813d5c89b9d 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -400,7 +400,7 @@ static struct mm_struct * mm_init(struct mm_struct * mm, struct task_struct *p) INIT_LIST_HEAD(&mm->mmlist); mm->flags = (current->mm) ? current->mm->flags : MMF_DUMP_FILTER_DEFAULT; - mm->core_waiters = 0; + mm->core_state = NULL; mm->nr_ptes = 0; set_mm_counter(mm, file_rss, 0); set_mm_counter(mm, anon_rss, 0); diff --git a/kernel/signal.c b/kernel/signal.c index 39c1706edf0..5c7b7eaa0dc 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -1480,10 +1480,10 @@ static inline int may_ptrace_stop(void) * is a deadlock situation, and pointless because our tracer * is dead so don't allow us to stop. * If SIGKILL was already sent before the caller unlocked - * ->siglock we must see ->core_waiters != 0. Otherwise it + * ->siglock we must see ->core_state != NULL. Otherwise it * is safe to enter schedule(). */ - if (unlikely(current->mm->core_waiters) && + if (unlikely(current->mm->core_state) && unlikely(current->mm == current->parent->mm)) return 0; -- cgit v1.2.3-70-g09d2 From 8cd9c249128a59e8e833d454a784b0cbd338d468 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:42 -0700 Subject: coredump: simplify core_state->nr_threads calculation Change zap_process() to return int instead of incrementing mm->core_state->nr_threads directly. Change zap_threads() to set mm->core_state only on success. This patch restores the original size of .text, and more importantly now ->nr_threads is used in two places only. Signed-off-by: Oleg Nesterov Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/exec.c | 33 ++++++++++++++++----------------- 1 file changed, 16 insertions(+), 17 deletions(-) (limited to 'fs') diff --git a/fs/exec.c b/fs/exec.c index 50de3aaff4d..c74bb34eeef 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -1502,9 +1502,10 @@ out: return ispipe; } -static void zap_process(struct task_struct *start) +static int zap_process(struct task_struct *start) { struct task_struct *t; + int nr = 0; start->signal->flags = SIGNAL_GROUP_EXIT; start->signal->group_stop_count = 0; @@ -1512,31 +1513,33 @@ static void zap_process(struct task_struct *start) t = start; do { if (t != current && t->mm) { - t->mm->core_state->nr_threads++; sigaddset(&t->pending.signal, SIGKILL); signal_wake_up(t, 1); + nr++; } } while_each_thread(start, t); + + return nr; } static inline int zap_threads(struct task_struct *tsk, struct mm_struct *mm, - int exit_code) + struct core_state *core_state, int exit_code) { struct task_struct *g, *p; unsigned long flags; - int err = -EAGAIN; + int nr = -EAGAIN; spin_lock_irq(&tsk->sighand->siglock); if (!signal_group_exit(tsk->signal)) { + mm->core_state = core_state; tsk->signal->group_exit_code = exit_code; - zap_process(tsk); - err = 0; + nr = zap_process(tsk); } spin_unlock_irq(&tsk->sighand->siglock); - if (err) - return err; + if (unlikely(nr < 0)) + return nr; - if (atomic_read(&mm->mm_users) == mm->core_state->nr_threads + 1) + if (atomic_read(&mm->mm_users) == nr + 1) goto done; /* * We should find and kill all tasks which use this mm, and we should @@ -1579,7 +1582,7 @@ static inline int zap_threads(struct task_struct *tsk, struct mm_struct *mm, if (p->mm) { if (unlikely(p->mm == mm)) { lock_task_sighand(p, &flags); - zap_process(p); + nr += zap_process(p); unlock_task_sighand(p, &flags); } break; @@ -1588,7 +1591,8 @@ static inline int zap_threads(struct task_struct *tsk, struct mm_struct *mm, } rcu_read_unlock(); done: - return mm->core_state->nr_threads; + core_state->nr_threads = nr; + return nr; } static int coredump_wait(int exit_code) @@ -1601,12 +1605,7 @@ static int coredump_wait(int exit_code) init_completion(&mm->core_done); init_completion(&core_state.startup); - core_state.nr_threads = 0; - mm->core_state = &core_state; - - core_waiters = zap_threads(tsk, mm, exit_code); - if (core_waiters < 0) - mm->core_state = NULL; + core_waiters = zap_threads(tsk, mm, &core_state, exit_code); up_write(&mm->mmap_sem); if (unlikely(core_waiters < 0)) -- cgit v1.2.3-70-g09d2 From c5f1cc8c1828486a61ab3e575da6e2c62b34d399 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:42 -0700 Subject: coredump: turn core_state->nr_threads into atomic_t Turn core_state->nr_threads into atomic_t and kill now unneeded down_write(&mm->mmap_sem) in exit_mm(). Signed-off-by: Oleg Nesterov Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/exec.c | 2 +- include/linux/mm_types.h | 2 +- kernel/exit.c | 5 ++--- 3 files changed, 4 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/exec.c b/fs/exec.c index c74bb34eeef..15d493fe8aa 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -1591,7 +1591,7 @@ static inline int zap_threads(struct task_struct *tsk, struct mm_struct *mm, } rcu_read_unlock(); done: - core_state->nr_threads = nr; + atomic_set(&core_state->nr_threads, nr); return nr; } diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index c0b1747b61a..ae99a28ba6a 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -160,7 +160,7 @@ struct vm_area_struct { }; struct core_state { - int nr_threads; + atomic_t nr_threads; struct completion startup; }; diff --git a/kernel/exit.c b/kernel/exit.c index 988e232254e..63d82957baa 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -678,10 +678,9 @@ static void exit_mm(struct task_struct * tsk) down_read(&mm->mmap_sem); if (mm->core_state) { up_read(&mm->mmap_sem); - down_write(&mm->mmap_sem); - if (!--mm->core_state->nr_threads) + + if (atomic_dec_and_test(&mm->core_state->nr_threads)) complete(&mm->core_state->startup); - up_write(&mm->mmap_sem); wait_for_completion(&mm->core_done); down_read(&mm->mmap_sem); -- cgit v1.2.3-70-g09d2 From 9d5b327bf198d2720666de958dcc2ae219d86952 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:43 -0700 Subject: coredump: make mm->core_state visible to ->core_dump() Move the "struct core_state core_state" from coredump_wait() to do_coredump(), this makes mm->core_state visible to binfmt->core_dump(). Signed-off-by: Oleg Nesterov Acked-by: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/exec.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/exec.c b/fs/exec.c index 15d493fe8aa..b8ee842d93c 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -1595,17 +1595,16 @@ done: return nr; } -static int coredump_wait(int exit_code) +static int coredump_wait(int exit_code, struct core_state *core_state) { struct task_struct *tsk = current; struct mm_struct *mm = tsk->mm; - struct core_state core_state; struct completion *vfork_done; int core_waiters; init_completion(&mm->core_done); - init_completion(&core_state.startup); - core_waiters = zap_threads(tsk, mm, &core_state, exit_code); + init_completion(&core_state->startup); + core_waiters = zap_threads(tsk, mm, core_state, exit_code); up_write(&mm->mmap_sem); if (unlikely(core_waiters < 0)) @@ -1622,8 +1621,7 @@ static int coredump_wait(int exit_code) } if (core_waiters) - wait_for_completion(&core_state.startup); - mm->core_state = NULL; + wait_for_completion(&core_state->startup); fail: return core_waiters; } @@ -1679,6 +1677,7 @@ int get_dumpable(struct mm_struct *mm) int do_coredump(long signr, int exit_code, struct pt_regs * regs) { + struct core_state core_state; char corename[CORENAME_MAX_SIZE + 1]; struct mm_struct *mm = current->mm; struct linux_binfmt * binfmt; @@ -1717,7 +1716,7 @@ int do_coredump(long signr, int exit_code, struct pt_regs * regs) current->fsuid = 0; /* Dump root private */ } - retval = coredump_wait(exit_code); + retval = coredump_wait(exit_code, &core_state); if (retval < 0) goto fail; @@ -1812,6 +1811,7 @@ fail_unlock: current->fsuid = fsuid; complete_all(&mm->core_done); + mm->core_state = NULL; fail: return retval; } -- cgit v1.2.3-70-g09d2 From b564daf806d492dd4f7afe9b6c83b8d35d137669 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:44 -0700 Subject: coredump: construct the list of coredumping threads at startup time binfmt->core_dump() has to iterate over the all threads in system in order to find the coredumping threads and construct the list using the GFP_ATOMIC allocations. With this patch each thread allocates the list node on exit_mm()'s stack and adds itself to the list. This allows us to do further changes: - simplify ->core_dump() - change exit_mm() to clear ->mm first, then wait for ->core_done. this makes the coredumping process visible to oom_kill - kill mm->core_done Signed-off-by: Oleg Nesterov Acked-by: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/exec.c | 2 ++ include/linux/mm_types.h | 6 ++++++ kernel/exit.c | 15 ++++++++++++--- 3 files changed, 20 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/exec.c b/fs/exec.c index b8ee842d93c..fe2873b8037 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -1604,6 +1604,8 @@ static int coredump_wait(int exit_code, struct core_state *core_state) init_completion(&mm->core_done); init_completion(&core_state->startup); + core_state->dumper.task = tsk; + core_state->dumper.next = NULL; core_waiters = zap_threads(tsk, mm, core_state, exit_code); up_write(&mm->mmap_sem); diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index ae99a28ba6a..4d0d0abc79f 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -159,8 +159,14 @@ struct vm_area_struct { #endif }; +struct core_thread { + struct task_struct *task; + struct core_thread *next; +}; + struct core_state { atomic_t nr_threads; + struct core_thread dumper; struct completion startup; }; diff --git a/kernel/exit.c b/kernel/exit.c index 63d82957baa..b66f0d55c79 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -664,6 +664,7 @@ assign_new_owner: static void exit_mm(struct task_struct * tsk) { struct mm_struct *mm = tsk->mm; + struct core_state *core_state; mm_release(tsk, mm); if (!mm) @@ -676,11 +677,19 @@ static void exit_mm(struct task_struct * tsk) * group with ->mm != NULL. */ down_read(&mm->mmap_sem); - if (mm->core_state) { + core_state = mm->core_state; + if (core_state) { + struct core_thread self; up_read(&mm->mmap_sem); - if (atomic_dec_and_test(&mm->core_state->nr_threads)) - complete(&mm->core_state->startup); + self.task = tsk; + self.next = xchg(&core_state->dumper.next, &self); + /* + * Implies mb(), the result of xchg() must be visible + * to core_state->dumper. + */ + if (atomic_dec_and_test(&core_state->nr_threads)) + complete(&core_state->startup); wait_for_completion(&mm->core_done); down_read(&mm->mmap_sem); -- cgit v1.2.3-70-g09d2 From 83914441f94c6f2cd468ca97365f6c34f418706e Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:45 -0700 Subject: coredump: elf_core_dump: use core_state->dumper list Kill the nasty rcu_read_lock() + do_each_thread() loop, use the list encoded in mm->core_state instead, s/GFP_ATOMIC/GFP_KERNEL/. This patch allows futher cleanups in binfmt_elf.c, in particular we can kill the parallel info->threads list. Signed-off-by: Oleg Nesterov Acked-by: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/binfmt_elf.c | 77 ++++++++++++++++++++++++--------------------------------- 1 file changed, 32 insertions(+), 45 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c index bad7d8770d7..88d180306cf 100644 --- a/fs/binfmt_elf.c +++ b/fs/binfmt_elf.c @@ -1478,7 +1478,7 @@ static int fill_note_info(struct elfhdr *elf, int phdrs, const struct user_regset_view *view = task_user_regset_view(dump_task); struct elf_thread_core_info *t; struct elf_prpsinfo *psinfo; - struct task_struct *g, *p; + struct core_thread *ct; unsigned int i; info->size = 0; @@ -1517,34 +1517,26 @@ static int fill_note_info(struct elfhdr *elf, int phdrs, /* * Allocate a structure for each thread. */ - rcu_read_lock(); - do_each_thread(g, p) - if (p->mm == dump_task->mm) { - if (p->flags & PF_KTHREAD) - continue; - - t = kzalloc(offsetof(struct elf_thread_core_info, - notes[info->thread_notes]), - GFP_ATOMIC); - if (unlikely(!t)) { - rcu_read_unlock(); - return 0; - } - t->task = p; - if (p == dump_task || !info->thread) { - t->next = info->thread; - info->thread = t; - } else { - /* - * Make sure to keep the original task at - * the head of the list. - */ - t->next = info->thread->next; - info->thread->next = t; - } + for (ct = &dump_task->mm->core_state->dumper; ct; ct = ct->next) { + t = kzalloc(offsetof(struct elf_thread_core_info, + notes[info->thread_notes]), + GFP_KERNEL); + if (unlikely(!t)) + return 0; + + t->task = ct->task; + if (ct->task == dump_task || !info->thread) { + t->next = info->thread; + info->thread = t; + } else { + /* + * Make sure to keep the original task at + * the head of the list. + */ + t->next = info->thread->next; + info->thread->next = t; } - while_each_thread(g, p); - rcu_read_unlock(); + } /* * Now fill in each thread's information. @@ -1691,7 +1683,6 @@ static int fill_note_info(struct elfhdr *elf, int phdrs, { #define NUM_NOTES 6 struct list_head *t; - struct task_struct *g, *p; info->notes = NULL; info->prstatus = NULL; @@ -1723,23 +1714,19 @@ static int fill_note_info(struct elfhdr *elf, int phdrs, info->thread_status_size = 0; if (signr) { + struct core_thread *ct; struct elf_thread_status *ets; - rcu_read_lock(); - do_each_thread(g, p) - if (current->mm == p->mm && current != p) { - if (p->flags & PF_KTHREAD) - continue; - - ets = kzalloc(sizeof(*ets), GFP_ATOMIC); - if (!ets) { - rcu_read_unlock(); - return 0; - } - ets->thread = p; - list_add(&ets->list, &info->thread_list); - } - while_each_thread(g, p); - rcu_read_unlock(); + + for (ct = current->mm->core_state->dumper.next; + ct; ct = ct->next) { + ets = kzalloc(sizeof(*ets), GFP_KERNEL); + if (!ets) + return 0; + + ets->thread = ct->task; + list_add(&ets->list, &info->thread_list); + } + list_for_each(t, &info->thread_list) { int sz; -- cgit v1.2.3-70-g09d2 From 182c515fd2a942623aed4e4e0e0b37fe96571b05 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:45 -0700 Subject: coredump: elf_fdpic_core_dump: use core_state->dumper list Kill the nasty rcu_read_lock() + do_each_thread() loop, use the list encoded in mm->core_state instead, s/GFP_ATOMIC/GFP_KERNEL/. This patch allows futher cleanups in binfmt_elf_fdpic.c. Signed-off-by: Oleg Nesterov Acked-by: Roland McGrath Cc: David Howells Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/binfmt_elf_fdpic.c | 29 ++++++++++++----------------- 1 file changed, 12 insertions(+), 17 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_elf_fdpic.c b/fs/binfmt_elf_fdpic.c index 71bcc4b4d08..1b59b1edf26 100644 --- a/fs/binfmt_elf_fdpic.c +++ b/fs/binfmt_elf_fdpic.c @@ -1573,7 +1573,6 @@ static int elf_fdpic_core_dump(long signr, struct pt_regs *regs, struct memelfnote *notes = NULL; struct elf_prstatus *prstatus = NULL; /* NT_PRSTATUS */ struct elf_prpsinfo *psinfo = NULL; /* NT_PRPSINFO */ - struct task_struct *g, *p; LIST_HEAD(thread_list); struct list_head *t; elf_fpregset_t *fpu = NULL; @@ -1622,23 +1621,19 @@ static int elf_fdpic_core_dump(long signr, struct pt_regs *regs, #endif if (signr) { + struct core_thread *ct; struct elf_thread_status *tmp; - rcu_read_lock(); - do_each_thread(g,p) - if (current->mm == p->mm && current != p) { - if (p->flags & PF_KTHREAD) - continue; - - tmp = kzalloc(sizeof(*tmp), GFP_ATOMIC); - if (!tmp) { - rcu_read_unlock(); - goto cleanup; - } - tmp->thread = p; - list_add(&tmp->list, &thread_list); - } - while_each_thread(g,p); - rcu_read_unlock(); + + for (ct = current->mm->core_state->dumper.next; + ct; ct = ct->next) { + tmp = kzalloc(sizeof(*tmp), GFP_KERNEL); + if (!tmp) + goto cleanup; + + tmp->thread = ct->task; + list_add(&tmp->list, &thread_list); + } + list_for_each(t, &thread_list) { struct elf_thread_status *tmp; int sz; -- cgit v1.2.3-70-g09d2 From a94e2d408eaedbd85aae259621d46fafc10479a2 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:46 -0700 Subject: coredump: kill mm->core_done Now that we have core_state->dumper list we can use it to wake up the sub-threads waiting for the coredump completion. This uglifies the code and .text grows by 47 bytes, but otoh mm_struct lessens by sizeof(struct completion). Also, with this change we can decouple exit_mm() from the coredumping code. Signed-off-by: Oleg Nesterov Cc: Roland McGrath Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/exec.c | 25 ++++++++++++++++++++++--- include/linux/mm_types.h | 4 +--- kernel/exit.c | 8 +++++++- 3 files changed, 30 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/exec.c b/fs/exec.c index fe2873b8037..bff43aeb235 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -1602,7 +1602,6 @@ static int coredump_wait(int exit_code, struct core_state *core_state) struct completion *vfork_done; int core_waiters; - init_completion(&mm->core_done); init_completion(&core_state->startup); core_state->dumper.task = tsk; core_state->dumper.next = NULL; @@ -1628,6 +1627,27 @@ fail: return core_waiters; } +static void coredump_finish(struct mm_struct *mm) +{ + struct core_thread *curr, *next; + struct task_struct *task; + + next = mm->core_state->dumper.next; + while ((curr = next) != NULL) { + next = curr->next; + task = curr->task; + /* + * see exit_mm(), curr->task must not see + * ->task == NULL before we read ->next. + */ + smp_mb(); + curr->task = NULL; + wake_up_process(task); + } + + mm->core_state = NULL; +} + /* * set_dumpable converts traditional three-value dumpable to two flags and * stores them into mm->flags. It modifies lower two bits of mm->flags, but @@ -1812,8 +1832,7 @@ fail_unlock: argv_free(helper_argv); current->fsuid = fsuid; - complete_all(&mm->core_done); - mm->core_state = NULL; + coredump_finish(mm); fail: return retval; } diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index 4d0d0abc79f..746f975b58e 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -229,9 +229,7 @@ struct mm_struct { unsigned long flags; /* Must use atomic bitops to access the bits */ - /* coredumping support */ - struct core_state *core_state; - struct completion core_done; + struct core_state *core_state; /* coredumping support */ /* aio bits */ rwlock_t ioctx_list_lock; /* aio lock */ diff --git a/kernel/exit.c b/kernel/exit.c index b66f0d55c79..8a4d4d12e29 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -691,7 +691,13 @@ static void exit_mm(struct task_struct * tsk) if (atomic_dec_and_test(&core_state->nr_threads)) complete(&core_state->startup); - wait_for_completion(&mm->core_done); + for (;;) { + set_task_state(tsk, TASK_UNINTERRUPTIBLE); + if (!self.task) /* see coredump_finish() */ + break; + schedule(); + } + __set_task_state(tsk, TASK_RUNNING); down_read(&mm->mmap_sem); } atomic_inc(&mm->mm_count); -- cgit v1.2.3-70-g09d2 From 565b9b14e7f48131bca58840aa404bbef058fa89 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 25 Jul 2008 01:47:47 -0700 Subject: coredump: format_corename: fix the "core_uses_pid" logic I don't understand why the multi-thread coredump implies the core_uses_pid behaviour, but we shouldn't use mm->mm_users for that. This counter can be incremented by get_task_mm(). Use the valued returned by coredump_wait() instead. Also, remove the "const char *pattern" argument, format_corename() can use core_pattern directly. [akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Oleg Nesterov Cc: Roland McGrath Cc: Alan Cox Cc: Andi Kleen Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/exec.c | 13 +++++-------- 1 file changed, 5 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/exec.c b/fs/exec.c index bff43aeb235..5e559013e30 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -1379,17 +1379,14 @@ EXPORT_SYMBOL(set_binfmt); * name into corename, which must have space for at least * CORENAME_MAX_SIZE bytes plus one byte for the zero terminator. */ -static int format_corename(char *corename, const char *pattern, long signr) +static int format_corename(char *corename, int nr_threads, long signr) { - const char *pat_ptr = pattern; + const char *pat_ptr = core_pattern; + int ispipe = (*pat_ptr == '|'); char *out_ptr = corename; char *const out_end = corename + CORENAME_MAX_SIZE; int rc; int pid_in_pattern = 0; - int ispipe = 0; - - if (*pattern == '|') - ispipe = 1; /* Repeat as long as we have more pattern to process and more output space */ @@ -1490,7 +1487,7 @@ static int format_corename(char *corename, const char *pattern, long signr) * and core_uses_pid is set, then .%pid will be appended to * the filename. Do not do this for piped commands. */ if (!ispipe && !pid_in_pattern - && (core_uses_pid || atomic_read(¤t->mm->mm_users) != 1)) { + && (core_uses_pid || nr_threads)) { rc = snprintf(out_ptr, out_end - out_ptr, ".%d", task_tgid_vnr(current)); if (rc > out_end - out_ptr) @@ -1753,7 +1750,7 @@ int do_coredump(long signr, int exit_code, struct pt_regs * regs) * uses lock_kernel() */ lock_kernel(); - ispipe = format_corename(corename, core_pattern, signr); + ispipe = format_corename(corename, retval, signr); unlock_kernel(); /* * Don't bother to check the RLIMIT_CORE value if core_pattern points -- cgit v1.2.3-70-g09d2 From 79885b227740b9c7d3057f2de556f4098d37cc8f Mon Sep 17 00:00:00 2001 From: "Edgar E. Iglesias" Date: Fri, 25 Jul 2008 01:48:10 -0700 Subject: elf: use ELF_CORE_EFLAGS for kcore ELF header flags ELF_CORE_EFLAGS is already used by the binfmt_elf coredumper to set correct arch specific ELF header flags on coredumps. Use it for kcore dumps as well. At the moment, this affects the CRIS and the H8300 arch. Signed-off-by: Edgar E. Iglesias Cc: Mikael Starvik Cc: Yoshinori Sato Cc: Ralf Baechle Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/kcore.c | 10 +++++----- include/asm-h8300/elf.h | 4 ++-- 2 files changed, 7 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/proc/kcore.c b/fs/proc/kcore.c index e78c81fcf54..c2370c76fb7 100644 --- a/fs/proc/kcore.c +++ b/fs/proc/kcore.c @@ -23,6 +23,10 @@ #define CORE_STR "CORE" +#ifndef ELF_CORE_EFLAGS +#define ELF_CORE_EFLAGS 0 +#endif + static int open_kcore(struct inode * inode, struct file * filp) { return capable(CAP_SYS_RAWIO) ? 0 : -EPERM; @@ -164,11 +168,7 @@ static void elf_kcore_store_hdr(char *bufp, int nphdr, int dataoff) elf->e_entry = 0; elf->e_phoff = sizeof(struct elfhdr); elf->e_shoff = 0; -#if defined(CONFIG_H8300) - elf->e_flags = ELF_FLAGS; -#else - elf->e_flags = 0; -#endif + elf->e_flags = ELF_CORE_EFLAGS; elf->e_ehsize = sizeof(struct elfhdr); elf->e_phentsize= sizeof(struct elf_phdr); elf->e_phnum = nphdr; diff --git a/include/asm-h8300/elf.h b/include/asm-h8300/elf.h index 26bfc7e641d..a8b57d1f412 100644 --- a/include/asm-h8300/elf.h +++ b/include/asm-h8300/elf.h @@ -26,10 +26,10 @@ typedef unsigned long elf_fpregset_t; #define ELF_DATA ELFDATA2MSB #define ELF_ARCH EM_H8_300 #if defined(__H8300H__) -#define ELF_FLAGS 0x810000 +#define ELF_CORE_EFLAGS 0x810000 #endif #if defined(__H8300S__) -#define ELF_FLAGS 0x820000 +#define ELF_CORE_EFLAGS 0x820000 #endif #define ELF_PLAT_INIT(_r) _r->er1 = 0 -- cgit v1.2.3-70-g09d2 From d805dda412346225a50af2d399d958a4bc676c38 Mon Sep 17 00:00:00 2001 From: Abdel Benamrouche Date: Fri, 25 Jul 2008 01:48:25 -0700 Subject: fs/partition/check.c: fix return value warning fs/partitions/check.c:381: warning: ignoring return value of ___device_add___, declared with attribute warn_unused_result [akpm@linux-foundation.org: multiple-return-statements-per-function are evil] Signed-off-by: Abdel Benamrouche Cc: Jens Axboe Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/partitions/check.c | 28 ++++++++++++++++++++++------ include/linux/genhd.h | 2 +- 2 files changed, 23 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/partitions/check.c b/fs/partitions/check.c index efef715135d..2e6413fbd2d 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -344,18 +344,18 @@ static ssize_t whole_disk_show(struct device *dev, static DEVICE_ATTR(whole_disk, S_IRUSR | S_IRGRP | S_IROTH, whole_disk_show, NULL); -void add_partition(struct gendisk *disk, int part, sector_t start, sector_t len, int flags) +int add_partition(struct gendisk *disk, int part, sector_t start, sector_t len, int flags) { struct hd_struct *p; int err; p = kzalloc(sizeof(*p), GFP_KERNEL); if (!p) - return; + return -ENOMEM; if (!init_part_stats(p)) { - kfree(p); - return; + err = -ENOMEM; + goto out0; } p->start_sect = start; p->nr_sects = len; @@ -378,15 +378,31 @@ void add_partition(struct gendisk *disk, int part, sector_t start, sector_t len, /* delay uevent until 'holders' subdir is created */ p->dev.uevent_suppress = 1; - device_add(&p->dev); + err = device_add(&p->dev); + if (err) + goto out1; partition_sysfs_add_subdir(p); p->dev.uevent_suppress = 0; - if (flags & ADDPART_FLAG_WHOLEDISK) + if (flags & ADDPART_FLAG_WHOLEDISK) { err = device_create_file(&p->dev, &dev_attr_whole_disk); + if (err) + goto out2; + } /* suppress uevent if the disk supresses it */ if (!disk->dev.uevent_suppress) kobject_uevent(&p->dev.kobj, KOBJ_ADD); + + return 0; + +out2: + device_del(&p->dev); +out1: + put_device(&p->dev); + free_part_stats(p); +out0: + kfree(p); + return err; } /* Not exported, helper to add_disk(). */ diff --git a/include/linux/genhd.h b/include/linux/genhd.h index e8787417f65..118216f1bd3 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -541,7 +541,7 @@ extern dev_t blk_lookup_devt(const char *name, int part); extern char *disk_name (struct gendisk *hd, int part, char *buf); extern int rescan_partitions(struct gendisk *disk, struct block_device *bdev); -extern void add_partition(struct gendisk *, int, sector_t, sector_t, int); +extern int __must_check add_partition(struct gendisk *, int, sector_t, sector_t, int); extern void delete_partition(struct gendisk *, int); extern void printk_all_partitions(void); -- cgit v1.2.3-70-g09d2 From 04ebd4aee52b06a2c38127d9208546e5b96f3a19 Mon Sep 17 00:00:00 2001 From: Abdel Benamrouche Date: Fri, 25 Jul 2008 01:48:26 -0700 Subject: block/ioctl.c and fs/partition/check.c: check value returned by add_partition() Now that add_partition() has been aught to propagate errors, let's check them. [akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Abdel Benamrouche Cc: Jens Axboe Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- block/ioctl.c | 5 +++-- fs/partitions/check.c | 10 ++++++++-- 2 files changed, 11 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/block/ioctl.c b/block/ioctl.c index 52d6385216a..77185e5c026 100644 --- a/block/ioctl.c +++ b/block/ioctl.c @@ -17,6 +17,7 @@ static int blkpg_ioctl(struct block_device *bdev, struct blkpg_ioctl_arg __user long long start, length; int part; int i; + int err; if (!capable(CAP_SYS_ADMIN)) return -EACCES; @@ -61,9 +62,9 @@ static int blkpg_ioctl(struct block_device *bdev, struct blkpg_ioctl_arg __user } } /* all seems OK */ - add_partition(disk, part, start, length, ADDPART_FLAG_NONE); + err = add_partition(disk, part, start, length, ADDPART_FLAG_NONE); mutex_unlock(&bdev->bd_mutex); - return 0; + return err; case BLKPG_DEL_PARTITION: if (!disk->part[part-1]) return -ENXIO; diff --git a/fs/partitions/check.c b/fs/partitions/check.c index 2e6413fbd2d..7d6b34e201d 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -499,10 +499,16 @@ int rescan_partitions(struct gendisk *disk, struct block_device *bdev) if (!size) continue; if (from + size > get_capacity(disk)) { - printk(" %s: p%d exceeds device capacity\n", + printk(KERN_ERR " %s: p%d exceeds device capacity\n", disk->disk_name, p); + continue; + } + res = add_partition(disk, p, from, size, state->parts[p].flags); + if (res) { + printk(KERN_ERR " %s: p%d could not be added: %d\n", + disk->disk_name, p, -res); + continue; } - add_partition(disk, p, from, size, state->parts[p].flags); #ifdef CONFIG_BLK_DEV_MD if (state->parts[p].flags & ADDPART_FLAG_RAID) md_autodetect_dev(bdev->bd_dev+p); -- cgit v1.2.3-70-g09d2 From d991696263a704be7f41ac186f1a0ed17963c260 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 25 Jul 2008 01:48:26 -0700 Subject: fs/partitions/efi: convert to pr_debug convert the local Dprintk() compile time debug printk wrappers to the generic pr_debug() wrapper. Signed-off-by: Thomas Gleixner Cc: Matt Domsch Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/partitions/efi.c | 42 +++++++++++++++++------------------------- 1 file changed, 17 insertions(+), 25 deletions(-) (limited to 'fs') diff --git a/fs/partitions/efi.c b/fs/partitions/efi.c index e7b07006bc4..038a6022152 100644 --- a/fs/partitions/efi.c +++ b/fs/partitions/efi.c @@ -95,13 +95,6 @@ #include "check.h" #include "efi.h" -#undef EFI_DEBUG -#ifdef EFI_DEBUG -#define Dprintk(x...) printk(KERN_DEBUG x) -#else -#define Dprintk(x...) -#endif - /* This allows a kernel command line option 'gpt' to override * the test for invalid PMBR. Not __initdata because reloading * the partition tables happens after init too. @@ -305,10 +298,10 @@ is_gpt_valid(struct block_device *bdev, u64 lba, /* Check the GUID Partition Table signature */ if (le64_to_cpu((*gpt)->signature) != GPT_HEADER_SIGNATURE) { - Dprintk("GUID Partition Table Header signature is wrong:" - "%lld != %lld\n", - (unsigned long long)le64_to_cpu((*gpt)->signature), - (unsigned long long)GPT_HEADER_SIGNATURE); + pr_debug("GUID Partition Table Header signature is wrong:" + "%lld != %lld\n", + (unsigned long long)le64_to_cpu((*gpt)->signature), + (unsigned long long)GPT_HEADER_SIGNATURE); goto fail; } @@ -318,9 +311,8 @@ is_gpt_valid(struct block_device *bdev, u64 lba, crc = efi_crc32((const unsigned char *) (*gpt), le32_to_cpu((*gpt)->header_size)); if (crc != origcrc) { - Dprintk - ("GUID Partition Table Header CRC is wrong: %x != %x\n", - crc, origcrc); + pr_debug("GUID Partition Table Header CRC is wrong: %x != %x\n", + crc, origcrc); goto fail; } (*gpt)->header_crc32 = cpu_to_le32(origcrc); @@ -328,9 +320,9 @@ is_gpt_valid(struct block_device *bdev, u64 lba, /* Check that the my_lba entry points to the LBA that contains * the GUID Partition Table */ if (le64_to_cpu((*gpt)->my_lba) != lba) { - Dprintk("GPT my_lba incorrect: %lld != %lld\n", - (unsigned long long)le64_to_cpu((*gpt)->my_lba), - (unsigned long long)lba); + pr_debug("GPT my_lba incorrect: %lld != %lld\n", + (unsigned long long)le64_to_cpu((*gpt)->my_lba), + (unsigned long long)lba); goto fail; } @@ -339,15 +331,15 @@ is_gpt_valid(struct block_device *bdev, u64 lba, */ lastlba = last_lba(bdev); if (le64_to_cpu((*gpt)->first_usable_lba) > lastlba) { - Dprintk("GPT: first_usable_lba incorrect: %lld > %lld\n", - (unsigned long long)le64_to_cpu((*gpt)->first_usable_lba), - (unsigned long long)lastlba); + pr_debug("GPT: first_usable_lba incorrect: %lld > %lld\n", + (unsigned long long)le64_to_cpu((*gpt)->first_usable_lba), + (unsigned long long)lastlba); goto fail; } if (le64_to_cpu((*gpt)->last_usable_lba) > lastlba) { - Dprintk("GPT: last_usable_lba incorrect: %lld > %lld\n", - (unsigned long long)le64_to_cpu((*gpt)->last_usable_lba), - (unsigned long long)lastlba); + pr_debug("GPT: last_usable_lba incorrect: %lld > %lld\n", + (unsigned long long)le64_to_cpu((*gpt)->last_usable_lba), + (unsigned long long)lastlba); goto fail; } @@ -360,7 +352,7 @@ is_gpt_valid(struct block_device *bdev, u64 lba, le32_to_cpu((*gpt)->sizeof_partition_entry)); if (crc != le32_to_cpu((*gpt)->partition_entry_array_crc32)) { - Dprintk("GUID Partitition Entry Array CRC check failed.\n"); + pr_debug("GUID Partitition Entry Array CRC check failed.\n"); goto fail_ptes; } @@ -616,7 +608,7 @@ efi_partition(struct parsed_partitions *state, struct block_device *bdev) return 0; } - Dprintk("GUID Partition Table is valid! Yea!\n"); + pr_debug("GUID Partition Table is valid! Yea!\n"); for (i = 0; i < le32_to_cpu(gpt->num_partition_entries) && i < state->limit-1; i++) { if (!is_pte_valid(&ptes[i], last_lba(bdev))) -- cgit v1.2.3-70-g09d2 From 6e644c3126149b65460610fe5a00d8a162092abe Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Fri, 25 Jul 2008 01:48:28 -0700 Subject: move proc_kmsg_operations to fs/proc/internal.h This patch moves the extern of struct proc_kmsg_operations to fs/proc/internal.h and adds an #include "internal.h" to fs/proc/kmsg.c so that the latter sees the former. Signed-off-by: Adrian Bunk Cc: Alexey Dobriyan Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/internal.h | 1 + fs/proc/kmsg.c | 2 ++ include/linux/proc_fs.h | 1 - 3 files changed, 3 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/proc/internal.h b/fs/proc/internal.h index 28cbca80590..8d67616e7bb 100644 --- a/fs/proc/internal.h +++ b/fs/proc/internal.h @@ -63,6 +63,7 @@ extern const struct file_operations proc_smaps_operations; extern const struct file_operations proc_clear_refs_operations; extern const struct file_operations proc_pagemap_operations; extern const struct file_operations proc_net_operations; +extern const struct file_operations proc_kmsg_operations; extern const struct inode_operations proc_net_inode_operations; void free_proc_entry(struct proc_dir_entry *de); diff --git a/fs/proc/kmsg.c b/fs/proc/kmsg.c index ff3b90b56e9..9fd5df3f40c 100644 --- a/fs/proc/kmsg.c +++ b/fs/proc/kmsg.c @@ -15,6 +15,8 @@ #include #include +#include "internal.h" + extern wait_queue_head_t log_wait; extern int do_syslog(int type, char __user *bug, int count); diff --git a/include/linux/proc_fs.h b/include/linux/proc_fs.h index 15a9eaf4a80..cdabc2fc02f 100644 --- a/include/linux/proc_fs.h +++ b/include/linux/proc_fs.h @@ -138,7 +138,6 @@ extern int proc_readdir(struct file *, void *, filldir_t); extern struct dentry *proc_lookup(struct inode *, struct dentry *, struct nameidata *); extern const struct file_operations proc_kcore_operations; -extern const struct file_operations proc_kmsg_operations; extern const struct file_operations ppc_htab_operations; extern int pid_ns_prepare_proc(struct pid_namespace *ns); -- cgit v1.2.3-70-g09d2 From 881adb85358309ea9c6f707394002719982ec607 Mon Sep 17 00:00:00 2001 From: Alexey Dobriyan Date: Fri, 25 Jul 2008 01:48:29 -0700 Subject: proc: always do ->release Current two-stage scheme of removing PDE emphasizes one bug in proc: open rmmod remove_proc_entry close ->release won't be called because ->proc_fops were cleared. In simple cases it's small memory leak. For every ->open, ->release has to be done. List of openers is introduced which is traversed at remove_proc_entry() if neeeded. Discussions with Al long ago (sigh). Signed-off-by: Alexey Dobriyan Cc: Al Viro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/generic.c | 14 ++++++++++ fs/proc/inode.c | 74 ++++++++++++++++++++++++++++++++++++++++++++++--- fs/proc/internal.h | 7 +++++ include/linux/proc_fs.h | 1 + 4 files changed, 92 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/proc/generic.c b/fs/proc/generic.c index 43e54e86cef..bc0a0dd2d84 100644 --- a/fs/proc/generic.c +++ b/fs/proc/generic.c @@ -597,6 +597,7 @@ static struct proc_dir_entry *__proc_create(struct proc_dir_entry **parent, ent->pde_users = 0; spin_lock_init(&ent->pde_unload_lock); ent->pde_unload_completion = NULL; + INIT_LIST_HEAD(&ent->pde_openers); out: return ent; } @@ -789,6 +790,19 @@ void remove_proc_entry(const char *name, struct proc_dir_entry *parent) spin_unlock(&de->pde_unload_lock); continue_removing: + spin_lock(&de->pde_unload_lock); + while (!list_empty(&de->pde_openers)) { + struct pde_opener *pdeo; + + pdeo = list_first_entry(&de->pde_openers, struct pde_opener, lh); + list_del(&pdeo->lh); + spin_unlock(&de->pde_unload_lock); + pdeo->release(pdeo->inode, pdeo->file); + kfree(pdeo); + spin_lock(&de->pde_unload_lock); + } + spin_unlock(&de->pde_unload_lock); + if (S_ISDIR(de->mode)) parent->nlink--; de->nlink = 0; diff --git a/fs/proc/inode.c b/fs/proc/inode.c index b08d1001791..354c0848582 100644 --- a/fs/proc/inode.c +++ b/fs/proc/inode.c @@ -126,12 +126,17 @@ static const struct super_operations proc_sops = { .remount_fs = proc_remount, }; -static void pde_users_dec(struct proc_dir_entry *pde) +static void __pde_users_dec(struct proc_dir_entry *pde) { - spin_lock(&pde->pde_unload_lock); pde->pde_users--; if (pde->pde_unload_completion && pde->pde_users == 0) complete(pde->pde_unload_completion); +} + +static void pde_users_dec(struct proc_dir_entry *pde) +{ + spin_lock(&pde->pde_unload_lock); + __pde_users_dec(pde); spin_unlock(&pde->pde_unload_lock); } @@ -318,36 +323,97 @@ static int proc_reg_open(struct inode *inode, struct file *file) struct proc_dir_entry *pde = PDE(inode); int rv = 0; int (*open)(struct inode *, struct file *); + int (*release)(struct inode *, struct file *); + struct pde_opener *pdeo; + + /* + * What for, you ask? Well, we can have open, rmmod, remove_proc_entry + * sequence. ->release won't be called because ->proc_fops will be + * cleared. Depending on complexity of ->release, consequences vary. + * + * We can't wait for mercy when close will be done for real, it's + * deadlockable: rmmod foo release + * by hand in remove_proc_entry(). For this, save opener's credentials + * for later. + */ + pdeo = kmalloc(sizeof(struct pde_opener), GFP_KERNEL); + if (!pdeo) + return -ENOMEM; spin_lock(&pde->pde_unload_lock); if (!pde->proc_fops) { spin_unlock(&pde->pde_unload_lock); + kfree(pdeo); return rv; } pde->pde_users++; open = pde->proc_fops->open; + release = pde->proc_fops->release; spin_unlock(&pde->pde_unload_lock); if (open) rv = open(inode, file); - pde_users_dec(pde); + spin_lock(&pde->pde_unload_lock); + if (rv == 0 && release) { + /* To know what to release. */ + pdeo->inode = inode; + pdeo->file = file; + /* Strictly for "too late" ->release in proc_reg_release(). */ + pdeo->release = release; + list_add(&pdeo->lh, &pde->pde_openers); + } else + kfree(pdeo); + __pde_users_dec(pde); + spin_unlock(&pde->pde_unload_lock); return rv; } +static struct pde_opener *find_pde_opener(struct proc_dir_entry *pde, + struct inode *inode, struct file *file) +{ + struct pde_opener *pdeo; + + list_for_each_entry(pdeo, &pde->pde_openers, lh) { + if (pdeo->inode == inode && pdeo->file == file) + return pdeo; + } + return NULL; +} + static int proc_reg_release(struct inode *inode, struct file *file) { struct proc_dir_entry *pde = PDE(inode); int rv = 0; int (*release)(struct inode *, struct file *); + struct pde_opener *pdeo; spin_lock(&pde->pde_unload_lock); + pdeo = find_pde_opener(pde, inode, file); if (!pde->proc_fops) { - spin_unlock(&pde->pde_unload_lock); + /* + * Can't simply exit, __fput() will think that everything is OK, + * and move on to freeing struct file. remove_proc_entry() will + * find slacker in opener's list and will try to do non-trivial + * things with struct file. Therefore, remove opener from list. + * + * But if opener is removed from list, who will ->release it? + */ + if (pdeo) { + list_del(&pdeo->lh); + spin_unlock(&pde->pde_unload_lock); + rv = pdeo->release(inode, file); + kfree(pdeo); + } else + spin_unlock(&pde->pde_unload_lock); return rv; } pde->pde_users++; release = pde->proc_fops->release; + if (pdeo) { + list_del(&pdeo->lh); + kfree(pdeo); + } spin_unlock(&pde->pde_unload_lock); if (release) diff --git a/fs/proc/internal.h b/fs/proc/internal.h index 8d67616e7bb..442202314d5 100644 --- a/fs/proc/internal.h +++ b/fs/proc/internal.h @@ -89,3 +89,10 @@ struct dentry *proc_lookup_de(struct proc_dir_entry *de, struct inode *ino, struct dentry *dentry); int proc_readdir_de(struct proc_dir_entry *de, struct file *filp, void *dirent, filldir_t filldir); + +struct pde_opener { + struct inode *inode; + struct file *file; + int (*release)(struct inode *, struct file *); + struct list_head lh; +}; diff --git a/include/linux/proc_fs.h b/include/linux/proc_fs.h index cdabc2fc02f..f560d1705af 100644 --- a/include/linux/proc_fs.h +++ b/include/linux/proc_fs.h @@ -79,6 +79,7 @@ struct proc_dir_entry { int pde_users; /* number of callers into module in progress */ spinlock_t pde_unload_lock; /* proc_fops checks and pde_users bumps */ struct completion *pde_unload_completion; + struct list_head pde_openers; /* who did ->open, but not ->release */ }; struct kcore_list { -- cgit v1.2.3-70-g09d2 From a9bd4a3e070ba7494f154e1a11687a8a957d88dc Mon Sep 17 00:00:00 2001 From: Alexey Dobriyan Date: Fri, 25 Jul 2008 01:48:30 -0700 Subject: proc: remove pathetic remount code MS_RMT_MASK will unmask changes in do_remount_sb() anyway. Signed-off-by: Alexey Dobriyan Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/inode.c | 7 ------- 1 file changed, 7 deletions(-) (limited to 'fs') diff --git a/fs/proc/inode.c b/fs/proc/inode.c index 354c0848582..02eca2ed9dd 100644 --- a/fs/proc/inode.c +++ b/fs/proc/inode.c @@ -111,19 +111,12 @@ int __init proc_init_inodecache(void) return 0; } -static int proc_remount(struct super_block *sb, int *flags, char *data) -{ - *flags |= MS_NODIRATIME; - return 0; -} - static const struct super_operations proc_sops = { .alloc_inode = proc_alloc_inode, .destroy_inode = proc_destroy_inode, .drop_inode = generic_delete_inode, .delete_inode = proc_delete_inode, .statfs = simple_statfs, - .remount_fs = proc_remount, }; static void __pde_users_dec(struct proc_dir_entry *pde) -- cgit v1.2.3-70-g09d2 From 6eedf8d30d2b48e86fbcee1a32fb2fa5f42219ee Mon Sep 17 00:00:00 2001 From: Alexey Dobriyan Date: Fri, 25 Jul 2008 01:48:30 -0700 Subject: proc: move Kconfig to fs/proc/Kconfig Signed-off-by: Alexey Dobriyan Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/Kconfig | 60 +-------------------------------------------------------- fs/proc/Kconfig | 59 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 60 insertions(+), 59 deletions(-) create mode 100644 fs/proc/Kconfig (limited to 'fs') diff --git a/fs/Kconfig b/fs/Kconfig index ed563b9e352..97e3bdedb1e 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -902,65 +902,7 @@ endif # BLOCK menu "Pseudo filesystems" -config PROC_FS - bool "/proc file system support" if EMBEDDED - default y - help - This is a virtual file system providing information about the status - of the system. "Virtual" means that it doesn't take up any space on - your hard disk: the files are created on the fly by the kernel when - you try to access them. Also, you cannot read the files with older - version of the program less: you need to use more or cat. - - It's totally cool; for example, "cat /proc/interrupts" gives - information about what the different IRQs are used for at the moment - (there is a small number of Interrupt ReQuest lines in your computer - that are used by the attached devices to gain the CPU's attention -- - often a source of trouble if two devices are mistakenly configured - to use the same IRQ). The program procinfo to display some - information about your system gathered from the /proc file system. - - Before you can use the /proc file system, it has to be mounted, - meaning it has to be given a location in the directory hierarchy. - That location should be /proc. A command such as "mount -t proc proc - /proc" or the equivalent line in /etc/fstab does the job. - - The /proc file system is explained in the file - and on the proc(5) manpage - ("man 5 proc"). - - This option will enlarge your kernel by about 67 KB. Several - programs depend on this, so everyone should say Y here. - -config PROC_KCORE - bool "/proc/kcore support" if !ARM - depends on PROC_FS && MMU - -config PROC_VMCORE - bool "/proc/vmcore support (EXPERIMENTAL)" - depends on PROC_FS && CRASH_DUMP - default y - help - Exports the dump image of crashed kernel in ELF format. - -config PROC_SYSCTL - bool "Sysctl support (/proc/sys)" if EMBEDDED - depends on PROC_FS - select SYSCTL - default y - ---help--- - The sysctl interface provides a means of dynamically changing - certain kernel parameters and variables on the fly without requiring - a recompile of the kernel or reboot of the system. The primary - interface is through /proc/sys. If you say Y here a tree of - modifiable sysctl entries will be generated beneath the - /proc/sys directory. They are explained in the files - in . Note that enabling this - option will enlarge the kernel by at least 8 KB. - - As it is generally a good thing, you should say Y here unless - building a kernel for install/rescue disks or your system is very - limited in memory. +source "fs/proc/Kconfig" config SYSFS bool "sysfs file system support" if EMBEDDED diff --git a/fs/proc/Kconfig b/fs/proc/Kconfig new file mode 100644 index 00000000000..73cd7a418f0 --- /dev/null +++ b/fs/proc/Kconfig @@ -0,0 +1,59 @@ +config PROC_FS + bool "/proc file system support" if EMBEDDED + default y + help + This is a virtual file system providing information about the status + of the system. "Virtual" means that it doesn't take up any space on + your hard disk: the files are created on the fly by the kernel when + you try to access them. Also, you cannot read the files with older + version of the program less: you need to use more or cat. + + It's totally cool; for example, "cat /proc/interrupts" gives + information about what the different IRQs are used for at the moment + (there is a small number of Interrupt ReQuest lines in your computer + that are used by the attached devices to gain the CPU's attention -- + often a source of trouble if two devices are mistakenly configured + to use the same IRQ). The program procinfo to display some + information about your system gathered from the /proc file system. + + Before you can use the /proc file system, it has to be mounted, + meaning it has to be given a location in the directory hierarchy. + That location should be /proc. A command such as "mount -t proc proc + /proc" or the equivalent line in /etc/fstab does the job. + + The /proc file system is explained in the file + and on the proc(5) manpage + ("man 5 proc"). + + This option will enlarge your kernel by about 67 KB. Several + programs depend on this, so everyone should say Y here. + +config PROC_KCORE + bool "/proc/kcore support" if !ARM + depends on PROC_FS && MMU + +config PROC_VMCORE + bool "/proc/vmcore support (EXPERIMENTAL)" + depends on PROC_FS && CRASH_DUMP + default y + help + Exports the dump image of crashed kernel in ELF format. + +config PROC_SYSCTL + bool "Sysctl support (/proc/sys)" if EMBEDDED + depends on PROC_FS + select SYSCTL + default y + ---help--- + The sysctl interface provides a means of dynamically changing + certain kernel parameters and variables on the fly without requiring + a recompile of the kernel or reboot of the system. The primary + interface is through /proc/sys. If you say Y here a tree of + modifiable sysctl entries will be generated beneath the + /proc/sys directory. They are explained in the files + in . Note that enabling this + option will enlarge the kernel by at least 8 KB. + + As it is generally a good thing, you should say Y here unless + building a kernel for install/rescue disks or your system is very + limited in memory. -- cgit v1.2.3-70-g09d2 From 297c5d92634c809cef23d73e7b2556f2528ff7e2 Mon Sep 17 00:00:00 2001 From: Andrea Righi Date: Fri, 25 Jul 2008 01:48:49 -0700 Subject: task IO accounting: provide distinct tgid/tid I/O statistics Report per-thread I/O statistics in /proc/pid/task/tid/io and aggregate parent I/O statistics in /proc/pid/io. This approach follows the same model used to account per-process and per-thread CPU times. As a practial application, this allows for example to quickly find the top I/O consumer when a process spawns many child threads that perform the actual I/O work, because the aggregated I/O statistics can always be found in /proc/pid/io. [ Oleg Nesterov points out that we should check that the task is still alive before we iterate over the threads, but also says that we can do that fixup on top of this later. - Linus ] Acked-by: Balbir Singh Signed-off-by: Andrea Righi Cc: Matt Heaton Cc: Shailabh Nagar Acked-by-with-comments: Oleg Nesterov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/base.c | 86 ++++++++++++++++++++++++++++++++++++++++++--------- include/linux/sched.h | 4 +++ kernel/exit.c | 27 ++++++++++++++++ kernel/fork.c | 6 ++++ 4 files changed, 108 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/proc/base.c b/fs/proc/base.c index 58c3e6a8e15..a891fe4cb43 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -2376,29 +2376,82 @@ static int proc_base_fill_cache(struct file *filp, void *dirent, } #ifdef CONFIG_TASK_IO_ACCOUNTING -static int proc_pid_io_accounting(struct task_struct *task, char *buffer) -{ +static int do_io_accounting(struct task_struct *task, char *buffer, int whole) +{ + u64 rchar, wchar, syscr, syscw; + struct task_io_accounting ioac; + + if (!whole) { + rchar = task->rchar; + wchar = task->wchar; + syscr = task->syscr; + syscw = task->syscw; + memcpy(&ioac, &task->ioac, sizeof(ioac)); + } else { + unsigned long flags; + struct task_struct *t = task; + rchar = wchar = syscr = syscw = 0; + memset(&ioac, 0, sizeof(ioac)); + + rcu_read_lock(); + do { + rchar += t->rchar; + wchar += t->wchar; + syscr += t->syscr; + syscw += t->syscw; + + ioac.read_bytes += t->ioac.read_bytes; + ioac.write_bytes += t->ioac.write_bytes; + ioac.cancelled_write_bytes += + t->ioac.cancelled_write_bytes; + t = next_thread(t); + } while (t != task); + rcu_read_unlock(); + + if (lock_task_sighand(task, &flags)) { + struct signal_struct *sig = task->signal; + + rchar += sig->rchar; + wchar += sig->wchar; + syscr += sig->syscr; + syscw += sig->syscw; + + ioac.read_bytes += sig->ioac.read_bytes; + ioac.write_bytes += sig->ioac.write_bytes; + ioac.cancelled_write_bytes += + sig->ioac.cancelled_write_bytes; + + unlock_task_sighand(task, &flags); + } + } + return sprintf(buffer, -#ifdef CONFIG_TASK_XACCT "rchar: %llu\n" "wchar: %llu\n" "syscr: %llu\n" "syscw: %llu\n" -#endif "read_bytes: %llu\n" "write_bytes: %llu\n" "cancelled_write_bytes: %llu\n", -#ifdef CONFIG_TASK_XACCT - (unsigned long long)task->rchar, - (unsigned long long)task->wchar, - (unsigned long long)task->syscr, - (unsigned long long)task->syscw, -#endif - (unsigned long long)task->ioac.read_bytes, - (unsigned long long)task->ioac.write_bytes, - (unsigned long long)task->ioac.cancelled_write_bytes); + (unsigned long long)rchar, + (unsigned long long)wchar, + (unsigned long long)syscr, + (unsigned long long)syscw, + (unsigned long long)ioac.read_bytes, + (unsigned long long)ioac.write_bytes, + (unsigned long long)ioac.cancelled_write_bytes); +} + +static int proc_tid_io_accounting(struct task_struct *task, char *buffer) +{ + return do_io_accounting(task, buffer, 0); } -#endif + +static int proc_tgid_io_accounting(struct task_struct *task, char *buffer) +{ + return do_io_accounting(task, buffer, 1); +} +#endif /* CONFIG_TASK_IO_ACCOUNTING */ /* * Thread groups @@ -2470,7 +2523,7 @@ static const struct pid_entry tgid_base_stuff[] = { REG("coredump_filter", S_IRUGO|S_IWUSR, coredump_filter), #endif #ifdef CONFIG_TASK_IO_ACCOUNTING - INF("io", S_IRUGO, pid_io_accounting), + INF("io", S_IRUGO, tgid_io_accounting), #endif }; @@ -2797,6 +2850,9 @@ static const struct pid_entry tid_base_stuff[] = { #ifdef CONFIG_FAULT_INJECTION REG("make-it-fail", S_IRUGO|S_IWUSR, fault_inject), #endif +#ifdef CONFIG_TASK_IO_ACCOUNTING + INF("io", S_IRUGO, tid_io_accounting), +#endif }; static int proc_tid_base_readdir(struct file * filp, diff --git a/include/linux/sched.h b/include/linux/sched.h index af780f299c7..d22ffe06d0e 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -506,6 +506,10 @@ struct signal_struct { unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw; unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt; unsigned long inblock, oublock, cinblock, coublock; +#ifdef CONFIG_TASK_XACCT + u64 rchar, wchar, syscr, syscw; +#endif + struct task_io_accounting ioac; /* * Cumulative ns of scheduled CPU time for dead threads in the diff --git a/kernel/exit.c b/kernel/exit.c index 8a4d4d12e29..ad933bb29ec 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -120,6 +120,18 @@ static void __exit_signal(struct task_struct *tsk) sig->nivcsw += tsk->nivcsw; sig->inblock += task_io_get_inblock(tsk); sig->oublock += task_io_get_oublock(tsk); +#ifdef CONFIG_TASK_XACCT + sig->rchar += tsk->rchar; + sig->wchar += tsk->wchar; + sig->syscr += tsk->syscr; + sig->syscw += tsk->syscw; +#endif /* CONFIG_TASK_XACCT */ +#ifdef CONFIG_TASK_IO_ACCOUNTING + sig->ioac.read_bytes += tsk->ioac.read_bytes; + sig->ioac.write_bytes += tsk->ioac.write_bytes; + sig->ioac.cancelled_write_bytes += + tsk->ioac.cancelled_write_bytes; +#endif /* CONFIG_TASK_IO_ACCOUNTING */ sig->sum_sched_runtime += tsk->se.sum_exec_runtime; sig = NULL; /* Marker for below. */ } @@ -1366,6 +1378,21 @@ static int wait_task_zombie(struct task_struct *p, int options, psig->coublock += task_io_get_oublock(p) + sig->oublock + sig->coublock; +#ifdef CONFIG_TASK_XACCT + psig->rchar += p->rchar + sig->rchar; + psig->wchar += p->wchar + sig->wchar; + psig->syscr += p->syscr + sig->syscr; + psig->syscw += p->syscw + sig->syscw; +#endif /* CONFIG_TASK_XACCT */ +#ifdef CONFIG_TASK_IO_ACCOUNTING + psig->ioac.read_bytes += + p->ioac.read_bytes + sig->ioac.read_bytes; + psig->ioac.write_bytes += + p->ioac.write_bytes + sig->ioac.write_bytes; + psig->ioac.cancelled_write_bytes += + p->ioac.cancelled_write_bytes + + sig->ioac.cancelled_write_bytes; +#endif /* CONFIG_TASK_IO_ACCOUNTING */ spin_unlock_irq(&p->parent->sighand->siglock); } diff --git a/kernel/fork.c b/kernel/fork.c index 813d5c89b9d..b99d73e971a 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -812,6 +812,12 @@ static int copy_signal(unsigned long clone_flags, struct task_struct *tsk) sig->nvcsw = sig->nivcsw = sig->cnvcsw = sig->cnivcsw = 0; sig->min_flt = sig->maj_flt = sig->cmin_flt = sig->cmaj_flt = 0; sig->inblock = sig->oublock = sig->cinblock = sig->coublock = 0; +#ifdef CONFIG_TASK_XACCT + sig->rchar = sig->wchar = sig->syscr = sig->syscw = 0; +#endif +#ifdef CONFIG_TASK_IO_ACCOUNTING + memset(&sig->ioac, 0, sizeof(sig->ioac)); +#endif sig->sum_sched_runtime = 0; INIT_LIST_HEAD(&sig->cpu_timers[0]); INIT_LIST_HEAD(&sig->cpu_timers[1]); -- cgit v1.2.3-70-g09d2 From cc77b1521d06be07c9bb1a4a3e1f775dcaa15093 Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Fri, 25 Jul 2008 01:48:55 -0700 Subject: lockd: dont return EAGAIN for a permanent error Fix nlm_fopen() to return NLM_FAILED (or NLM_LCK_DENIED_NOLOCKS) instead of NLM_LCK_DENIED. The latter means the lock request failed because of a conflicting lock (i.e. a temporary error), which is wrong in this case. Also fix the client to return ENOLCK instead of EAGAIN if a blocking lock request returns with NLM_LOCK_DENIED. Signed-off-by: Miklos Szeredi Cc: Trond Myklebust Cc: "J. Bruce Fields" Cc: Matthew Wilcox Cc: David Teigland Cc: Christoph Hellwig Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/lockd/clntproc.c | 10 +++++++++- fs/nfsd/lockd.c | 13 +++++++++---- 2 files changed, 18 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/lockd/clntproc.c b/fs/lockd/clntproc.c index 1f6dc518505..31668b690e0 100644 --- a/fs/lockd/clntproc.c +++ b/fs/lockd/clntproc.c @@ -582,7 +582,15 @@ again: } if (status < 0) goto out_unlock; - status = nlm_stat_to_errno(resp->status); + /* + * EAGAIN doesn't make sense for sleeping locks, and in some + * cases NLM_LCK_DENIED is returned for a permanent error. So + * turn it into an ENOLCK. + */ + if (resp->status == nlm_lck_denied && (fl_flags & FL_SLEEP)) + status = -ENOLCK; + else + status = nlm_stat_to_errno(resp->status); out_unblock: nlmclnt_finish_block(block); out: diff --git a/fs/nfsd/lockd.c b/fs/nfsd/lockd.c index 6b6225ac492..15c6faeec77 100644 --- a/fs/nfsd/lockd.c +++ b/fs/nfsd/lockd.c @@ -19,6 +19,13 @@ #define NFSDDBG_FACILITY NFSDDBG_LOCKD +#ifdef CONFIG_LOCKD_V4 +#define nlm_stale_fh nlm4_stale_fh +#define nlm_failed nlm4_failed +#else +#define nlm_stale_fh nlm_lck_denied_nolocks +#define nlm_failed nlm_lck_denied_nolocks +#endif /* * Note: we hold the dentry use count while the file is open. */ @@ -47,12 +54,10 @@ nlm_fopen(struct svc_rqst *rqstp, struct nfs_fh *f, struct file **filp) return 0; case nfserr_dropit: return nlm_drop_reply; -#ifdef CONFIG_LOCKD_V4 case nfserr_stale: - return nlm4_stale_fh; -#endif + return nlm_stale_fh; default: - return nlm_lck_denied; + return nlm_failed; } } -- cgit v1.2.3-70-g09d2 From bde74e4bc64415b142e556a34d295a52a1b7da9d Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Fri, 25 Jul 2008 01:48:57 -0700 Subject: locks: add special return value for asynchronous locks Use a special error value FILE_LOCK_DEFERRED to mean that a locking operation returned asynchronously. This is returned by posix_lock_file() for sleeping locks to mean that the lock has been queued on the block list, and will be woken up when it might become available and needs to be retried (either fl_lmops->fl_notify() is called or fl_wait is woken up). f_op->lock() to mean either the above, or that the filesystem will call back with fl_lmops->fl_grant() when the result of the locking operation is known. The filesystem can do this for sleeping as well as non-sleeping locks. This is to make sure, that return values of -EAGAIN and -EINPROGRESS by filesystems are not mistaken to mean an asynchronous locking. This also makes error handling in fs/locks.c and lockd/svclock.c slightly cleaner. Signed-off-by: Miklos Szeredi Cc: Trond Myklebust Cc: "J. Bruce Fields" Cc: Matthew Wilcox Cc: David Teigland Cc: Christoph Hellwig Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/dlm/plock.c | 2 +- fs/lockd/svclock.c | 13 ++++--------- fs/locks.c | 28 ++++++++++++++-------------- include/linux/fs.h | 6 ++++++ 4 files changed, 25 insertions(+), 24 deletions(-) (limited to 'fs') diff --git a/fs/dlm/plock.c b/fs/dlm/plock.c index 78878c5781c..eba87ff3177 100644 --- a/fs/dlm/plock.c +++ b/fs/dlm/plock.c @@ -116,7 +116,7 @@ int dlm_posix_lock(dlm_lockspace_t *lockspace, u64 number, struct file *file, if (xop->callback == NULL) wait_event(recv_wq, (op->done != 0)); else { - rv = -EINPROGRESS; + rv = FILE_LOCK_DEFERRED; goto out; } diff --git a/fs/lockd/svclock.c b/fs/lockd/svclock.c index 821b9acdfb6..cf0d5c2c318 100644 --- a/fs/lockd/svclock.c +++ b/fs/lockd/svclock.c @@ -418,8 +418,8 @@ nlmsvc_lock(struct svc_rqst *rqstp, struct nlm_file *file, goto out; case -EAGAIN: ret = nlm_lck_denied; - break; - case -EINPROGRESS: + goto out; + case FILE_LOCK_DEFERRED: if (wait) break; /* Filesystem lock operation is in progress @@ -434,10 +434,6 @@ nlmsvc_lock(struct svc_rqst *rqstp, struct nlm_file *file, goto out; } - ret = nlm_lck_denied; - if (!wait) - goto out; - ret = nlm_lck_blocked; /* Append to list of blocked */ @@ -507,7 +503,7 @@ nlmsvc_testlock(struct svc_rqst *rqstp, struct nlm_file *file, } error = vfs_test_lock(file->f_file, &lock->fl); - if (error == -EINPROGRESS) { + if (error == FILE_LOCK_DEFERRED) { ret = nlmsvc_defer_lock_rqst(rqstp, block); goto out; } @@ -731,8 +727,7 @@ nlmsvc_grant_blocked(struct nlm_block *block) switch (error) { case 0: break; - case -EAGAIN: - case -EINPROGRESS: + case FILE_LOCK_DEFERRED: dprintk("lockd: lock still blocked error %d\n", error); nlmsvc_insert_block(block, NLM_NEVER); nlmsvc_release_block(block); diff --git a/fs/locks.c b/fs/locks.c index dce8c747371..1ce57b4b362 100644 --- a/fs/locks.c +++ b/fs/locks.c @@ -779,8 +779,10 @@ find_conflict: if (!flock_locks_conflict(request, fl)) continue; error = -EAGAIN; - if (request->fl_flags & FL_SLEEP) - locks_insert_block(fl, request); + if (!(request->fl_flags & FL_SLEEP)) + goto out; + error = FILE_LOCK_DEFERRED; + locks_insert_block(fl, request); goto out; } if (request->fl_flags & FL_ACCESS) @@ -836,7 +838,7 @@ static int __posix_lock_file(struct inode *inode, struct file_lock *request, str error = -EDEADLK; if (posix_locks_deadlock(request, fl)) goto out; - error = -EAGAIN; + error = FILE_LOCK_DEFERRED; locks_insert_block(fl, request); goto out; } @@ -1035,7 +1037,7 @@ int posix_lock_file_wait(struct file *filp, struct file_lock *fl) might_sleep (); for (;;) { error = posix_lock_file(filp, fl, NULL); - if ((error != -EAGAIN) || !(fl->fl_flags & FL_SLEEP)) + if (error != FILE_LOCK_DEFERRED) break; error = wait_event_interruptible(fl->fl_wait, !fl->fl_next); if (!error) @@ -1107,9 +1109,7 @@ int locks_mandatory_area(int read_write, struct inode *inode, for (;;) { error = __posix_lock_file(inode, &fl, NULL); - if (error != -EAGAIN) - break; - if (!(fl.fl_flags & FL_SLEEP)) + if (error != FILE_LOCK_DEFERRED) break; error = wait_event_interruptible(fl.fl_wait, !fl.fl_next); if (!error) { @@ -1531,7 +1531,7 @@ int flock_lock_file_wait(struct file *filp, struct file_lock *fl) might_sleep(); for (;;) { error = flock_lock_file(filp, fl); - if ((error != -EAGAIN) || !(fl->fl_flags & FL_SLEEP)) + if (error != FILE_LOCK_DEFERRED) break; error = wait_event_interruptible(fl->fl_wait, !fl->fl_next); if (!error) @@ -1716,17 +1716,17 @@ out: * fl_grant is set. Callers expecting ->lock() to return asynchronously * will only use F_SETLK, not F_SETLKW; they will set FL_SLEEP if (and only if) * the request is for a blocking lock. When ->lock() does return asynchronously, - * it must return -EINPROGRESS, and call ->fl_grant() when the lock + * it must return FILE_LOCK_DEFERRED, and call ->fl_grant() when the lock * request completes. * If the request is for non-blocking lock the file system should return - * -EINPROGRESS then try to get the lock and call the callback routine with - * the result. If the request timed out the callback routine will return a + * FILE_LOCK_DEFERRED then try to get the lock and call the callback routine + * with the result. If the request timed out the callback routine will return a * nonzero return code and the file system should release the lock. The file * system is also responsible to keep a corresponding posix lock when it * grants a lock so the VFS can find out which locks are locally held and do * the correct lock cleanup when required. * The underlying filesystem must not drop the kernel lock or call - * ->fl_grant() before returning to the caller with a -EINPROGRESS + * ->fl_grant() before returning to the caller with a FILE_LOCK_DEFERRED * return code. */ int vfs_lock_file(struct file *filp, unsigned int cmd, struct file_lock *fl, struct file_lock *conf) @@ -1804,7 +1804,7 @@ again: else { for (;;) { error = posix_lock_file(filp, file_lock, NULL); - if (error != -EAGAIN || cmd == F_SETLK) + if (error != FILE_LOCK_DEFERRED) break; error = wait_event_interruptible(file_lock->fl_wait, !file_lock->fl_next); @@ -1941,7 +1941,7 @@ again: else { for (;;) { error = posix_lock_file(filp, file_lock, NULL); - if (error != -EAGAIN || cmd == F_SETLK64) + if (error != FILE_LOCK_DEFERRED) break; error = wait_event_interruptible(file_lock->fl_wait, !file_lock->fl_next); diff --git a/include/linux/fs.h b/include/linux/fs.h index 4b86f806014..49d8eb7a71b 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -885,6 +885,12 @@ static inline int file_check_writeable(struct file *filp) #define FL_CLOSE 64 /* unlock on close */ #define FL_SLEEP 128 /* A blocking lock */ +/* + * Special return value from posix_lock_file() and vfs_lock_file() for + * asynchronous locking. + */ +#define FILE_LOCK_DEFERRED 1 + /* * The POSIX file lock owner is determined by * the "struct files_struct" in the thread group -- cgit v1.2.3-70-g09d2 From b648a6de00770cc325c22f43bdd4e935f6a2ee55 Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Fri, 25 Jul 2008 01:48:57 -0700 Subject: locks: cleanup code duplication Extract common code into a function. Signed-off-by: Miklos Szeredi Cc: "J. Bruce Fields" Cc: Trond Myklebust Cc: Matthew Wilcox Cc: David Teigland Cc: Christoph Hellwig Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/locks.c | 71 +++++++++++++++++++++++++++----------------------------------- 1 file changed, 31 insertions(+), 40 deletions(-) (limited to 'fs') diff --git a/fs/locks.c b/fs/locks.c index 1ce57b4b362..6222e4b580e 100644 --- a/fs/locks.c +++ b/fs/locks.c @@ -1738,6 +1738,35 @@ int vfs_lock_file(struct file *filp, unsigned int cmd, struct file_lock *fl, str } EXPORT_SYMBOL_GPL(vfs_lock_file); +static int do_lock_file_wait(struct file *filp, unsigned int cmd, + struct file_lock *fl) +{ + int error; + + error = security_file_lock(filp, fl->fl_type); + if (error) + return error; + + if (filp->f_op && filp->f_op->lock != NULL) + error = filp->f_op->lock(filp, cmd, fl); + else { + for (;;) { + error = posix_lock_file(filp, fl, NULL); + if (error != FILE_LOCK_DEFERRED) + break; + error = wait_event_interruptible(fl->fl_wait, + !fl->fl_next); + if (!error) + continue; + + locks_delete_block(fl); + break; + } + } + + return error; +} + /* Apply the lock described by l to an open file descriptor. * This implements both the F_SETLK and F_SETLKW commands of fcntl(). */ @@ -1795,26 +1824,7 @@ again: goto out; } - error = security_file_lock(filp, file_lock->fl_type); - if (error) - goto out; - - if (filp->f_op && filp->f_op->lock != NULL) - error = filp->f_op->lock(filp, cmd, file_lock); - else { - for (;;) { - error = posix_lock_file(filp, file_lock, NULL); - if (error != FILE_LOCK_DEFERRED) - break; - error = wait_event_interruptible(file_lock->fl_wait, - !file_lock->fl_next); - if (!error) - continue; - - locks_delete_block(file_lock); - break; - } - } + error = do_lock_file_wait(filp, cmd, file_lock); /* * Attempt to detect a close/fcntl race and recover by @@ -1932,26 +1942,7 @@ again: goto out; } - error = security_file_lock(filp, file_lock->fl_type); - if (error) - goto out; - - if (filp->f_op && filp->f_op->lock != NULL) - error = filp->f_op->lock(filp, cmd, file_lock); - else { - for (;;) { - error = posix_lock_file(filp, file_lock, NULL); - if (error != FILE_LOCK_DEFERRED) - break; - error = wait_event_interruptible(file_lock->fl_wait, - !file_lock->fl_next); - if (!error) - continue; - - locks_delete_block(file_lock); - break; - } - } + error = do_lock_file_wait(filp, cmd, file_lock); /* * Attempt to detect a close/fcntl race and recover by -- cgit v1.2.3-70-g09d2 From 764c76b371722e0cba5c24d91225f0f954b69d44 Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Fri, 25 Jul 2008 01:48:58 -0700 Subject: locks: allow ->lock() to return FILE_LOCK_DEFERRED Allow filesystem's ->lock() method to call posix_lock_file() instead of posix_lock_file_wait(), and return FILE_LOCK_DEFERRED. This makes it possible to implement a such a ->lock() function, that works with the lock manager, which needs the call to be asynchronous. Now the vfs_lock_file() helper can be used, so this is a cleanup as well. Signed-off-by: Miklos Szeredi Cc: "J. Bruce Fields" Cc: Trond Myklebust Cc: Matthew Wilcox Cc: David Teigland Cc: Christoph Hellwig Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/locks.c | 23 +++++++++-------------- 1 file changed, 9 insertions(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/locks.c b/fs/locks.c index 6222e4b580e..01490300f7c 100644 --- a/fs/locks.c +++ b/fs/locks.c @@ -1747,21 +1747,16 @@ static int do_lock_file_wait(struct file *filp, unsigned int cmd, if (error) return error; - if (filp->f_op && filp->f_op->lock != NULL) - error = filp->f_op->lock(filp, cmd, fl); - else { - for (;;) { - error = posix_lock_file(filp, fl, NULL); - if (error != FILE_LOCK_DEFERRED) - break; - error = wait_event_interruptible(fl->fl_wait, - !fl->fl_next); - if (!error) - continue; - - locks_delete_block(fl); + for (;;) { + error = vfs_lock_file(filp, cmd, fl, NULL); + if (error != FILE_LOCK_DEFERRED) break; - } + error = wait_event_interruptible(fl->fl_wait, !fl->fl_next); + if (!error) + continue; + + locks_delete_block(fl); + break; } return error; -- cgit v1.2.3-70-g09d2 From 0de6256daafa3a97a269995e9b29f956bd419bbf Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Fri, 25 Jul 2008 01:48:59 -0700 Subject: fuse: prepare lookup for nfs export Use d_splice_alias() instead of d_add() in fuse lookup code, to allow NFS exporting. Signed-off-by: Miklos Szeredi Cc: Christoph Hellwig Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fuse/dir.c | 24 ++++++++++++++---------- 1 file changed, 14 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index 2060bf06b90..e5217b213b4 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -239,18 +239,20 @@ int fuse_valid_type(int m) * Add a directory inode to a dentry, ensuring that no other dentry * refers to this inode. Called with fc->inst_mutex. */ -static int fuse_d_add_directory(struct dentry *entry, struct inode *inode) +static struct dentry *fuse_d_add_directory(struct dentry *entry, + struct inode *inode) { struct dentry *alias = d_find_alias(inode); - if (alias) { + if (alias && !(alias->d_flags & DCACHE_DISCONNECTED)) { /* This tries to shrink the subtree below alias */ fuse_invalidate_entry(alias); dput(alias); if (!list_empty(&inode->i_dentry)) - return -EBUSY; + return ERR_PTR(-EBUSY); + } else { + dput(alias); } - d_add(entry, inode); - return 0; + return d_splice_alias(inode, entry); } static struct dentry *fuse_lookup(struct inode *dir, struct dentry *entry, @@ -259,6 +261,7 @@ static struct dentry *fuse_lookup(struct inode *dir, struct dentry *entry, int err; struct fuse_entry_out outarg; struct inode *inode = NULL; + struct dentry *newent; struct fuse_conn *fc = get_fuse_conn(dir); struct fuse_req *req; struct fuse_req *forget_req; @@ -303,21 +306,22 @@ static struct dentry *fuse_lookup(struct inode *dir, struct dentry *entry, if (inode && S_ISDIR(inode->i_mode)) { mutex_lock(&fc->inst_mutex); - err = fuse_d_add_directory(entry, inode); + newent = fuse_d_add_directory(entry, inode); mutex_unlock(&fc->inst_mutex); - if (err) { + if (IS_ERR(newent)) { iput(inode); - return ERR_PTR(err); + return newent; } } else - d_add(entry, inode); + newent = d_splice_alias(inode, entry); + entry = newent ? newent : entry; entry->d_op = &fuse_dentry_operations; if (!err) fuse_change_entry_timeout(entry, &outarg); else fuse_invalidate_entry_cache(entry); - return NULL; + return newent; } /* -- cgit v1.2.3-70-g09d2 From dbd561d236ff16f8143bc727d91758ddd190e8cb Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Fri, 25 Jul 2008 01:49:00 -0700 Subject: fuse: add export operations Implement export_operations, to allow fuse filesystems to be exported to NFS. This feature has been in the out-of-tree fuse module, and is widely used and tested. It has not been originally merged into mainline, because doing the NFS export in userspace was thought to be a cleaner and more efficient way of doing it, than through the kernel. While that is true, it would also have involved a lot of duplicated effort at reimplementing NFS exporting (all the different versions of the protocol). This effort was unfortunately not undertaken by anyone, so we are left with doing it the easy but less efficient way. If this feature goes in, the out-of-tree fuse module can go away, which would have several advantages: - not having to maintain two versions - less confusion for users - no bugs due to kernel API changes Comment from hch: - Use the same fh_type values as XFS, since we use the same fh encoding. Signed-off-by: Miklos Szeredi Cc: Christoph Hellwig Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fuse/dir.c | 4 +- fs/fuse/fuse_i.h | 4 ++ fs/fuse/inode.c | 115 +++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 121 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index e5217b213b4..be5450dd638 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -97,7 +97,7 @@ void fuse_invalidate_attr(struct inode *inode) * timeout is unknown (unlink, rmdir, rename and in some cases * lookup) */ -static void fuse_invalidate_entry_cache(struct dentry *entry) +void fuse_invalidate_entry_cache(struct dentry *entry) { fuse_dentry_settime(entry, 0); } @@ -225,7 +225,7 @@ static int invalid_nodeid(u64 nodeid) return !nodeid || nodeid == FUSE_ROOT_ID; } -static struct dentry_operations fuse_dentry_operations = { +struct dentry_operations fuse_dentry_operations = { .d_revalidate = fuse_dentry_revalidate, }; diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h index bae948657c4..5d3146da64e 100644 --- a/fs/fuse/fuse_i.h +++ b/fs/fuse/fuse_i.h @@ -464,6 +464,8 @@ static inline u64 get_node_id(struct inode *inode) /** Device operations */ extern const struct file_operations fuse_dev_operations; +extern struct dentry_operations fuse_dentry_operations; + /** * Get a filled in inode */ @@ -604,6 +606,8 @@ void fuse_abort_conn(struct fuse_conn *fc); */ void fuse_invalidate_attr(struct inode *inode); +void fuse_invalidate_entry_cache(struct dentry *entry); + /** * Acquire reference to fuse_conn */ diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c index 3141690558c..71fa76a48a3 100644 --- a/fs/fuse/inode.c +++ b/fs/fuse/inode.c @@ -18,6 +18,7 @@ #include #include #include +#include MODULE_AUTHOR("Miklos Szeredi "); MODULE_DESCRIPTION("Filesystem in Userspace"); @@ -552,6 +553,119 @@ static struct inode *get_root_inode(struct super_block *sb, unsigned mode) return fuse_iget(sb, 1, 0, &attr, 0, 0); } +struct fuse_inode_handle +{ + u64 nodeid; + u32 generation; +}; + +static struct dentry *fuse_get_dentry(struct super_block *sb, + struct fuse_inode_handle *handle) +{ + struct inode *inode; + struct dentry *entry; + int err = -ESTALE; + + if (handle->nodeid == 0) + goto out_err; + + inode = ilookup5(sb, handle->nodeid, fuse_inode_eq, &handle->nodeid); + if (!inode) + goto out_err; + err = -ESTALE; + if (inode->i_generation != handle->generation) + goto out_iput; + + entry = d_alloc_anon(inode); + err = -ENOMEM; + if (!entry) + goto out_iput; + + if (get_node_id(inode) != FUSE_ROOT_ID) { + entry->d_op = &fuse_dentry_operations; + fuse_invalidate_entry_cache(entry); + } + + return entry; + + out_iput: + iput(inode); + out_err: + return ERR_PTR(err); +} + +static int fuse_encode_fh(struct dentry *dentry, u32 *fh, int *max_len, + int connectable) +{ + struct inode *inode = dentry->d_inode; + bool encode_parent = connectable && !S_ISDIR(inode->i_mode); + int len = encode_parent ? 6 : 3; + u64 nodeid; + u32 generation; + + if (*max_len < len) + return 255; + + nodeid = get_fuse_inode(inode)->nodeid; + generation = inode->i_generation; + + fh[0] = (u32)(nodeid >> 32); + fh[1] = (u32)(nodeid & 0xffffffff); + fh[2] = generation; + + if (encode_parent) { + struct inode *parent; + + spin_lock(&dentry->d_lock); + parent = dentry->d_parent->d_inode; + nodeid = get_fuse_inode(parent)->nodeid; + generation = parent->i_generation; + spin_unlock(&dentry->d_lock); + + fh[3] = (u32)(nodeid >> 32); + fh[4] = (u32)(nodeid & 0xffffffff); + fh[5] = generation; + } + + *max_len = len; + return encode_parent ? 0x82 : 0x81; +} + +static struct dentry *fuse_fh_to_dentry(struct super_block *sb, + struct fid *fid, int fh_len, int fh_type) +{ + struct fuse_inode_handle handle; + + if ((fh_type != 0x81 && fh_type != 0x82) || fh_len < 3) + return NULL; + + handle.nodeid = (u64) fid->raw[0] << 32; + handle.nodeid |= (u64) fid->raw[1]; + handle.generation = fid->raw[2]; + return fuse_get_dentry(sb, &handle); +} + +static struct dentry *fuse_fh_to_parent(struct super_block *sb, + struct fid *fid, int fh_len, int fh_type) +{ + struct fuse_inode_handle parent; + + if (fh_type != 0x82 || fh_len < 6) + return NULL; + + parent.nodeid = (u64) fid->raw[3] << 32; + parent.nodeid |= (u64) fid->raw[4]; + parent.generation = fid->raw[5]; + return fuse_get_dentry(sb, &parent); +} + + +static const struct export_operations fuse_export_operations = { + .fh_to_dentry = fuse_fh_to_dentry, + .fh_to_parent = fuse_fh_to_parent, + .encode_fh = fuse_encode_fh, +}; + static const struct super_operations fuse_super_operations = { .alloc_inode = fuse_alloc_inode, .destroy_inode = fuse_destroy_inode, @@ -652,6 +766,7 @@ static int fuse_fill_super(struct super_block *sb, void *data, int silent) sb->s_magic = FUSE_SUPER_MAGIC; sb->s_op = &fuse_super_operations; sb->s_maxbytes = MAX_LFS_FILESIZE; + sb->s_export_op = &fuse_export_operations; file = fget(d.fd); if (!file) -- cgit v1.2.3-70-g09d2 From c180eebe1390c2076ead6a9bc95a02efb994edb7 Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Fri, 25 Jul 2008 01:49:01 -0700 Subject: fuse: add fuse_lookup_name() helper Add a new helper function which sends a LOOKUP request with the supplied name. This will be used by the next patch to send special LOOKUP requests with "." and ".." as the name. Signed-off-by: Miklos Szeredi Cc: Christoph Hellwig Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fuse/dir.c | 117 ++++++++++++++++++++++++++++++++++++++-------------------- 1 file changed, 77 insertions(+), 40 deletions(-) (limited to 'fs') diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index be5450dd638..51d0035ff07 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -112,18 +112,16 @@ static void fuse_invalidate_entry(struct dentry *entry) fuse_invalidate_entry_cache(entry); } -static void fuse_lookup_init(struct fuse_req *req, struct inode *dir, - struct dentry *entry, +static void fuse_lookup_init(struct fuse_conn *fc, struct fuse_req *req, + u64 nodeid, struct qstr *name, struct fuse_entry_out *outarg) { - struct fuse_conn *fc = get_fuse_conn(dir); - memset(outarg, 0, sizeof(struct fuse_entry_out)); req->in.h.opcode = FUSE_LOOKUP; - req->in.h.nodeid = get_node_id(dir); + req->in.h.nodeid = nodeid; req->in.numargs = 1; - req->in.args[0].size = entry->d_name.len + 1; - req->in.args[0].value = entry->d_name.name; + req->in.args[0].size = name->len + 1; + req->in.args[0].value = name->name; req->out.numargs = 1; if (fc->minor < 9) req->out.args[0].size = FUSE_COMPAT_ENTRY_OUT_SIZE; @@ -189,7 +187,8 @@ static int fuse_dentry_revalidate(struct dentry *entry, struct nameidata *nd) attr_version = fuse_get_attr_version(fc); parent = dget_parent(entry); - fuse_lookup_init(req, parent->d_inode, entry, &outarg); + fuse_lookup_init(fc, req, get_node_id(parent->d_inode), + &entry->d_name, &outarg); request_send(fc, req); dput(parent); err = req->out.h.error; @@ -255,73 +254,111 @@ static struct dentry *fuse_d_add_directory(struct dentry *entry, return d_splice_alias(inode, entry); } -static struct dentry *fuse_lookup(struct inode *dir, struct dentry *entry, - struct nameidata *nd) +int fuse_lookup_name(struct super_block *sb, u64 nodeid, struct qstr *name, + struct fuse_entry_out *outarg, struct inode **inode) { - int err; - struct fuse_entry_out outarg; - struct inode *inode = NULL; - struct dentry *newent; - struct fuse_conn *fc = get_fuse_conn(dir); + struct fuse_conn *fc = get_fuse_conn_super(sb); struct fuse_req *req; struct fuse_req *forget_req; u64 attr_version; + int err; - if (entry->d_name.len > FUSE_NAME_MAX) - return ERR_PTR(-ENAMETOOLONG); + *inode = NULL; + err = -ENAMETOOLONG; + if (name->len > FUSE_NAME_MAX) + goto out; req = fuse_get_req(fc); + err = PTR_ERR(req); if (IS_ERR(req)) - return ERR_CAST(req); + goto out; forget_req = fuse_get_req(fc); + err = PTR_ERR(forget_req); if (IS_ERR(forget_req)) { fuse_put_request(fc, req); - return ERR_CAST(forget_req); + goto out; } attr_version = fuse_get_attr_version(fc); - fuse_lookup_init(req, dir, entry, &outarg); + fuse_lookup_init(fc, req, nodeid, name, outarg); request_send(fc, req); err = req->out.h.error; fuse_put_request(fc, req); /* Zero nodeid is same as -ENOENT, but with valid timeout */ - if (!err && outarg.nodeid && - (invalid_nodeid(outarg.nodeid) || - !fuse_valid_type(outarg.attr.mode))) - err = -EIO; - if (!err && outarg.nodeid) { - inode = fuse_iget(dir->i_sb, outarg.nodeid, outarg.generation, - &outarg.attr, entry_attr_timeout(&outarg), - attr_version); - if (!inode) { - fuse_send_forget(fc, forget_req, outarg.nodeid, 1); - return ERR_PTR(-ENOMEM); - } + if (err || !outarg->nodeid) + goto out_put_forget; + + err = -EIO; + if (!outarg->nodeid) + goto out_put_forget; + if (!fuse_valid_type(outarg->attr.mode)) + goto out_put_forget; + + *inode = fuse_iget(sb, outarg->nodeid, outarg->generation, + &outarg->attr, entry_attr_timeout(outarg), + attr_version); + err = -ENOMEM; + if (!*inode) { + fuse_send_forget(fc, forget_req, outarg->nodeid, 1); + goto out; } + err = 0; + + out_put_forget: fuse_put_request(fc, forget_req); - if (err && err != -ENOENT) - return ERR_PTR(err); + out: + return err; +} + +static struct dentry *fuse_lookup(struct inode *dir, struct dentry *entry, + struct nameidata *nd) +{ + int err; + struct fuse_entry_out outarg; + struct inode *inode; + struct dentry *newent; + struct fuse_conn *fc = get_fuse_conn(dir); + bool outarg_valid = true; + + err = fuse_lookup_name(dir->i_sb, get_node_id(dir), &entry->d_name, + &outarg, &inode); + if (err == -ENOENT) { + outarg_valid = false; + err = 0; + } + if (err) + goto out_err; + + err = -EIO; + if (inode && get_node_id(inode) == FUSE_ROOT_ID) + goto out_iput; if (inode && S_ISDIR(inode->i_mode)) { mutex_lock(&fc->inst_mutex); newent = fuse_d_add_directory(entry, inode); mutex_unlock(&fc->inst_mutex); - if (IS_ERR(newent)) { - iput(inode); - return newent; - } - } else + err = PTR_ERR(newent); + if (IS_ERR(newent)) + goto out_iput; + } else { newent = d_splice_alias(inode, entry); + } entry = newent ? newent : entry; entry->d_op = &fuse_dentry_operations; - if (!err) + if (outarg_valid) fuse_change_entry_timeout(entry, &outarg); else fuse_invalidate_entry_cache(entry); + return newent; + + out_iput: + iput(inode); + out_err: + return ERR_PTR(err); } /* -- cgit v1.2.3-70-g09d2 From 33670fa296860283f04a7975b8c790f101e43a6e Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Fri, 25 Jul 2008 01:49:02 -0700 Subject: fuse: nfs export special lookups Implement the get_parent export operation by sending a LOOKUP request with ".." as the name. Implement looking up an inode by node ID after it has been evicted from the cache. This is done by seding a LOOKUP request with "." as the name (for all file types, not just directories). The filesystem can set the FUSE_EXPORT_SUPPORT flag in the INIT reply, to indicate that it supports these special lookups. Thanks to John Muir for the original implementation of this feature. Signed-off-by: Miklos Szeredi Cc: "J. Bruce Fields" Cc: Trond Myklebust Cc: Matthew Wilcox Cc: David Teigland Cc: Christoph Hellwig Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fuse/fuse_i.h | 6 +++++ fs/fuse/inode.c | 66 +++++++++++++++++++++++++++++++++++++++++++++++++--- include/linux/fuse.h | 3 +++ 3 files changed, 72 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h index 5d3146da64e..3a876076bdd 100644 --- a/fs/fuse/fuse_i.h +++ b/fs/fuse/fuse_i.h @@ -363,6 +363,9 @@ struct fuse_conn { /** Do not send separate SETATTR request before open(O_TRUNC) */ unsigned atomic_o_trunc : 1; + /** Filesystem supports NFS exporting. Only set in INIT */ + unsigned export_support : 1; + /* * The following bitfields are only for optimization purposes * and hence races in setting them will not cause malfunction @@ -473,6 +476,9 @@ struct inode *fuse_iget(struct super_block *sb, u64 nodeid, int generation, struct fuse_attr *attr, u64 attr_valid, u64 attr_version); +int fuse_lookup_name(struct super_block *sb, u64 nodeid, struct qstr *name, + struct fuse_entry_out *outarg, struct inode **inode); + /** * Send FORGET command */ diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c index 71fa76a48a3..7d2f7d6e22e 100644 --- a/fs/fuse/inode.c +++ b/fs/fuse/inode.c @@ -562,6 +562,7 @@ struct fuse_inode_handle static struct dentry *fuse_get_dentry(struct super_block *sb, struct fuse_inode_handle *handle) { + struct fuse_conn *fc = get_fuse_conn_super(sb); struct inode *inode; struct dentry *entry; int err = -ESTALE; @@ -570,8 +571,27 @@ static struct dentry *fuse_get_dentry(struct super_block *sb, goto out_err; inode = ilookup5(sb, handle->nodeid, fuse_inode_eq, &handle->nodeid); - if (!inode) - goto out_err; + if (!inode) { + struct fuse_entry_out outarg; + struct qstr name; + + if (!fc->export_support) + goto out_err; + + name.len = 1; + name.name = "."; + err = fuse_lookup_name(sb, handle->nodeid, &name, &outarg, + &inode); + if (err && err != -ENOENT) + goto out_err; + if (err || !inode) { + err = -ESTALE; + goto out_err; + } + err = -EIO; + if (get_node_id(inode) != handle->nodeid) + goto out_iput; + } err = -ESTALE; if (inode->i_generation != handle->generation) goto out_iput; @@ -659,11 +679,46 @@ static struct dentry *fuse_fh_to_parent(struct super_block *sb, return fuse_get_dentry(sb, &parent); } +static struct dentry *fuse_get_parent(struct dentry *child) +{ + struct inode *child_inode = child->d_inode; + struct fuse_conn *fc = get_fuse_conn(child_inode); + struct inode *inode; + struct dentry *parent; + struct fuse_entry_out outarg; + struct qstr name; + int err; + + if (!fc->export_support) + return ERR_PTR(-ESTALE); + + name.len = 2; + name.name = ".."; + err = fuse_lookup_name(child_inode->i_sb, get_node_id(child_inode), + &name, &outarg, &inode); + if (err && err != -ENOENT) + return ERR_PTR(err); + if (err || !inode) + return ERR_PTR(-ESTALE); + + parent = d_alloc_anon(inode); + if (!parent) { + iput(inode); + return ERR_PTR(-ENOMEM); + } + if (get_node_id(inode) != FUSE_ROOT_ID) { + parent->d_op = &fuse_dentry_operations; + fuse_invalidate_entry_cache(parent); + } + + return parent; +} static const struct export_operations fuse_export_operations = { .fh_to_dentry = fuse_fh_to_dentry, .fh_to_parent = fuse_fh_to_parent, .encode_fh = fuse_encode_fh, + .get_parent = fuse_get_parent, }; static const struct super_operations fuse_super_operations = { @@ -695,6 +750,11 @@ static void process_init_reply(struct fuse_conn *fc, struct fuse_req *req) fc->no_lock = 1; if (arg->flags & FUSE_ATOMIC_O_TRUNC) fc->atomic_o_trunc = 1; + if (arg->minor >= 9) { + /* LOOKUP has dependency on proto version */ + if (arg->flags & FUSE_EXPORT_SUPPORT) + fc->export_support = 1; + } if (arg->flags & FUSE_BIG_WRITES) fc->big_writes = 1; } else { @@ -721,7 +781,7 @@ static void fuse_send_init(struct fuse_conn *fc, struct fuse_req *req) arg->minor = FUSE_KERNEL_MINOR_VERSION; arg->max_readahead = fc->bdi.ra_pages * PAGE_CACHE_SIZE; arg->flags |= FUSE_ASYNC_READ | FUSE_POSIX_LOCKS | FUSE_ATOMIC_O_TRUNC | - FUSE_BIG_WRITES; + FUSE_EXPORT_SUPPORT | FUSE_BIG_WRITES; req->in.h.opcode = FUSE_INIT; req->in.numargs = 1; req->in.args[0].size = sizeof(*arg); diff --git a/include/linux/fuse.h b/include/linux/fuse.h index d4828219769..265635dc990 100644 --- a/include/linux/fuse.h +++ b/include/linux/fuse.h @@ -104,11 +104,14 @@ struct fuse_file_lock { /** * INIT request/reply flags + * + * FUSE_EXPORT_SUPPORT: filesystem handles lookups of "." and ".." */ #define FUSE_ASYNC_READ (1 << 0) #define FUSE_POSIX_LOCKS (1 << 1) #define FUSE_FILE_OPS (1 << 2) #define FUSE_ATOMIC_O_TRUNC (1 << 3) +#define FUSE_EXPORT_SUPPORT (1 << 4) #define FUSE_BIG_WRITES (1 << 5) /** -- cgit v1.2.3-70-g09d2 From 48e90761b570ff57f58b726229d229729949c5bb Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Fri, 25 Jul 2008 01:49:02 -0700 Subject: fuse: lockd support If fuse filesystem doesn't define it's own lock operations, then allow the lock manager to work with fuse. Adding lockd support for remote locking is also possible, but more rarely used, so leave it till later. Signed-off-by: Miklos Szeredi Cc: "J. Bruce Fields" Cc: Trond Myklebust Cc: Matthew Wilcox Cc: David Teigland Cc: Christoph Hellwig Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fuse/file.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/fuse/file.c b/fs/fuse/file.c index 8092f0d9fd1..67ff2c6a8f6 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -1341,6 +1341,11 @@ static int fuse_setlk(struct file *file, struct file_lock *fl, int flock) pid_t pid = fl->fl_type != F_UNLCK ? current->tgid : 0; int err; + if (fl->fl_lmops && fl->fl_lmops->fl_grant) { + /* NLM needs asynchronous locks, which we don't support yet */ + return -ENOLCK; + } + /* Unlock on close is handled by the flush method */ if (fl->fl_flags & FL_CLOSE) return 0; @@ -1365,7 +1370,9 @@ static int fuse_file_lock(struct file *file, int cmd, struct file_lock *fl) struct fuse_conn *fc = get_fuse_conn(inode); int err; - if (cmd == F_GETLK) { + if (cmd == F_CANCELLK) { + err = 0; + } else if (cmd == F_GETLK) { if (fc->no_lock) { posix_test_lock(file, fl); err = 0; @@ -1373,7 +1380,7 @@ static int fuse_file_lock(struct file *file, int cmd, struct file_lock *fl) err = fuse_getlk(file, fl); } else { if (fc->no_lock) - err = posix_lock_file_wait(file, fl); + err = posix_lock_file(file, fl, NULL); else err = fuse_setlk(file, fl, 0); } -- cgit v1.2.3-70-g09d2 From f3c6ba986ab4527b6dfacf9f3b9e40f72466a8b2 Mon Sep 17 00:00:00 2001 From: Kentaro Makita Date: Fri, 25 Jul 2008 19:44:40 -0700 Subject: vfs: add cond_resched_lock while scanning dentry LRU lists Add cond_resched_lock(&dcache_lock) while scanning LRU lists on superblocks in __shrink_dcache_sb() Signed-off-by: Kentaro Makita Cc: Al Viro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/dcache.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/dcache.c b/fs/dcache.c index 3818d6ab76c..f2584d22cb4 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -487,6 +487,7 @@ restart: if (!cnt) break; } + cond_resched_lock(&dcache_lock); } } while (!list_empty(&tmp)) { -- cgit v1.2.3-70-g09d2 From 7d135a5d50a08bbc53b189d79d8bdb03136f5303 Mon Sep 17 00:00:00 2001 From: Matthias Kaehlcke Date: Fri, 25 Jul 2008 19:44:51 -0700 Subject: affs: convert s_bmlock into a mutex The semaphore s_bmlock is used as a mutex. Convert it to the mutex API. Signed-off-by: Matthias Kaehlcke Cc: Roman Zippel Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/affs/affs.h | 3 ++- fs/affs/bitmap.c | 18 +++++++++--------- fs/affs/super.c | 2 +- 3 files changed, 12 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/affs/affs.h b/fs/affs/affs.h index 223b1917093..e9ec915f755 100644 --- a/fs/affs/affs.h +++ b/fs/affs/affs.h @@ -2,6 +2,7 @@ #include #include #include +#include /* AmigaOS allows file names with up to 30 characters length. * Names longer than that will be silently truncated. If you @@ -98,7 +99,7 @@ struct affs_sb_info { gid_t s_gid; /* gid to override */ umode_t s_mode; /* mode to override */ struct buffer_head *s_root_bh; /* Cached root block. */ - struct semaphore s_bmlock; /* Protects bitmap access. */ + struct mutex s_bmlock; /* Protects bitmap access. */ struct affs_bm_info *s_bitmap; /* Bitmap infos. */ u32 s_bmap_count; /* # of bitmap blocks. */ u32 s_bmap_bits; /* # of bits in one bitmap blocks */ diff --git a/fs/affs/bitmap.c b/fs/affs/bitmap.c index c4a5ad09ddf..dc5ef14bdc1 100644 --- a/fs/affs/bitmap.c +++ b/fs/affs/bitmap.c @@ -45,14 +45,14 @@ affs_count_free_blocks(struct super_block *sb) if (sb->s_flags & MS_RDONLY) return 0; - down(&AFFS_SB(sb)->s_bmlock); + mutex_lock(&AFFS_SB(sb)->s_bmlock); bm = AFFS_SB(sb)->s_bitmap; free = 0; for (i = AFFS_SB(sb)->s_bmap_count; i > 0; bm++, i--) free += bm->bm_free; - up(&AFFS_SB(sb)->s_bmlock); + mutex_unlock(&AFFS_SB(sb)->s_bmlock); return free; } @@ -76,7 +76,7 @@ affs_free_block(struct super_block *sb, u32 block) bit = blk % sbi->s_bmap_bits; bm = &sbi->s_bitmap[bmap]; - down(&sbi->s_bmlock); + mutex_lock(&sbi->s_bmlock); bh = sbi->s_bmap_bh; if (sbi->s_last_bmap != bmap) { @@ -105,19 +105,19 @@ affs_free_block(struct super_block *sb, u32 block) sb->s_dirt = 1; bm->bm_free++; - up(&sbi->s_bmlock); + mutex_unlock(&sbi->s_bmlock); return; err_free: affs_warning(sb,"affs_free_block","Trying to free block %u which is already free", block); - up(&sbi->s_bmlock); + mutex_unlock(&sbi->s_bmlock); return; err_bh_read: affs_error(sb,"affs_free_block","Cannot read bitmap block %u", bm->bm_key); sbi->s_bmap_bh = NULL; sbi->s_last_bmap = ~0; - up(&sbi->s_bmlock); + mutex_unlock(&sbi->s_bmlock); return; err_range: @@ -168,7 +168,7 @@ affs_alloc_block(struct inode *inode, u32 goal) bmap = blk / sbi->s_bmap_bits; bm = &sbi->s_bitmap[bmap]; - down(&sbi->s_bmlock); + mutex_lock(&sbi->s_bmlock); if (bm->bm_free) goto find_bmap_bit; @@ -249,7 +249,7 @@ find_bit: mark_buffer_dirty(bh); sb->s_dirt = 1; - up(&sbi->s_bmlock); + mutex_unlock(&sbi->s_bmlock); pr_debug("%d\n", blk); return blk; @@ -259,7 +259,7 @@ err_bh_read: sbi->s_bmap_bh = NULL; sbi->s_last_bmap = ~0; err_full: - up(&sbi->s_bmlock); + mutex_unlock(&sbi->s_bmlock); pr_debug("failed\n"); return 0; } diff --git a/fs/affs/super.c b/fs/affs/super.c index d214837d5e4..4e030956640 100644 --- a/fs/affs/super.c +++ b/fs/affs/super.c @@ -290,7 +290,7 @@ static int affs_fill_super(struct super_block *sb, void *data, int silent) if (!sbi) return -ENOMEM; sb->s_fs_info = sbi; - init_MUTEX(&sbi->s_bmlock); + mutex_init(&sbi->s_bmlock); if (!parse_options(data,&uid,&gid,&i,&reserved,&root_block, &blocksize,&sbi->s_prefix, -- cgit v1.2.3-70-g09d2 From 75b25b4cabb7ce956c36442bf8225659b1864866 Mon Sep 17 00:00:00 2001 From: Dmitri Vorobiev Date: Fri, 25 Jul 2008 19:44:52 -0700 Subject: bfs: assorted cleanups This patch makes the following cleanups: o removing an unused variable from bfs_fill_super(); o removing unneeded blank spaces from pointer definitions. Signed-off-by: Dmitri Vorobiev Cc: Tigran Aivazian Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/bfs/bfs.h | 4 ++-- fs/bfs/inode.c | 3 +-- 2 files changed, 3 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/bfs/bfs.h b/fs/bfs/bfs.h index 70f5d3a8eed..d1a6cd10afc 100644 --- a/fs/bfs/bfs.h +++ b/fs/bfs/bfs.h @@ -16,8 +16,8 @@ struct bfs_sb_info { unsigned long si_freei; unsigned long si_lf_eblk; unsigned long si_lasti; - unsigned long * si_imap; - struct buffer_head * si_sbh; /* buffer header w/superblock */ + unsigned long *si_imap; + struct buffer_head *si_sbh; /* buffer header w/superblock */ }; /* diff --git a/fs/bfs/inode.c b/fs/bfs/inode.c index 8db623838b5..c8b3982e11c 100644 --- a/fs/bfs/inode.c +++ b/fs/bfs/inode.c @@ -380,7 +380,7 @@ static int bfs_fill_super(struct super_block *s, void *data, int silent) struct bfs_inode *di; int block = (i - BFS_ROOT_INO) / BFS_INODES_PER_BLOCK + 1; int off = (i - BFS_ROOT_INO) % BFS_INODES_PER_BLOCK; - unsigned long sblock, eblock; + unsigned long eblock; if (!off) { brelse(bh); @@ -399,7 +399,6 @@ static int bfs_fill_super(struct super_block *s, void *data, int silent) set_bit(i, info->si_imap); info->si_freeb -= BFS_FILEBLOCKS(di); - sblock = le32_to_cpu(di->i_sblock); eblock = le32_to_cpu(di->i_eblock); if (eblock > info->si_lf_eblk) info->si_lf_eblk = eblock; -- cgit v1.2.3-70-g09d2 From 3f165e4cf2af042af7d2440d688299c0d2a48b1f Mon Sep 17 00:00:00 2001 From: Dmitri Vorobiev Date: Fri, 25 Jul 2008 19:44:54 -0700 Subject: bfs: kill BKL Replace the BKL-based locking scheme used in the bfs driver by a private filesystem-wide mutex. Signed-off-by: Dmitri Vorobiev Cc: Tigran Aivazian Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/bfs/bfs.h | 1 + fs/bfs/dir.c | 46 ++++++++++++++++++++++++++-------------------- fs/bfs/file.c | 4 ++-- fs/bfs/inode.c | 24 +++++++++++++++--------- 4 files changed, 44 insertions(+), 31 deletions(-) (limited to 'fs') diff --git a/fs/bfs/bfs.h b/fs/bfs/bfs.h index d1a6cd10afc..7109e451abf 100644 --- a/fs/bfs/bfs.h +++ b/fs/bfs/bfs.h @@ -18,6 +18,7 @@ struct bfs_sb_info { unsigned long si_lasti; unsigned long *si_imap; struct buffer_head *si_sbh; /* buffer header w/superblock */ + struct mutex bfs_lock; }; /* diff --git a/fs/bfs/dir.c b/fs/bfs/dir.c index 034950cb3cb..87ee5ccee34 100644 --- a/fs/bfs/dir.c +++ b/fs/bfs/dir.c @@ -32,16 +32,17 @@ static int bfs_readdir(struct file *f, void *dirent, filldir_t filldir) struct inode *dir = f->f_path.dentry->d_inode; struct buffer_head *bh; struct bfs_dirent *de; + struct bfs_sb_info *info = BFS_SB(dir->i_sb); unsigned int offset; int block; - lock_kernel(); + mutex_lock(&info->bfs_lock); if (f->f_pos & (BFS_DIRENT_SIZE - 1)) { printf("Bad f_pos=%08lx for %s:%08lx\n", (unsigned long)f->f_pos, dir->i_sb->s_id, dir->i_ino); - unlock_kernel(); + mutex_unlock(&info->bfs_lock); return -EBADF; } @@ -61,7 +62,7 @@ static int bfs_readdir(struct file *f, void *dirent, filldir_t filldir) le16_to_cpu(de->ino), DT_UNKNOWN) < 0) { brelse(bh); - unlock_kernel(); + mutex_unlock(&info->bfs_lock); return 0; } } @@ -71,7 +72,7 @@ static int bfs_readdir(struct file *f, void *dirent, filldir_t filldir) brelse(bh); } - unlock_kernel(); + mutex_unlock(&info->bfs_lock); return 0; } @@ -95,10 +96,10 @@ static int bfs_create(struct inode *dir, struct dentry *dentry, int mode, inode = new_inode(s); if (!inode) return -ENOSPC; - lock_kernel(); + mutex_lock(&info->bfs_lock); ino = find_first_zero_bit(info->si_imap, info->si_lasti); if (ino > info->si_lasti) { - unlock_kernel(); + mutex_unlock(&info->bfs_lock); iput(inode); return -ENOSPC; } @@ -125,10 +126,10 @@ static int bfs_create(struct inode *dir, struct dentry *dentry, int mode, if (err) { inode_dec_link_count(inode); iput(inode); - unlock_kernel(); + mutex_unlock(&info->bfs_lock); return err; } - unlock_kernel(); + mutex_unlock(&info->bfs_lock); d_instantiate(dentry, inode); return 0; } @@ -139,22 +140,23 @@ static struct dentry *bfs_lookup(struct inode *dir, struct dentry *dentry, struct inode *inode = NULL; struct buffer_head *bh; struct bfs_dirent *de; + struct bfs_sb_info *info = BFS_SB(dir->i_sb); if (dentry->d_name.len > BFS_NAMELEN) return ERR_PTR(-ENAMETOOLONG); - lock_kernel(); + mutex_lock(&info->bfs_lock); bh = bfs_find_entry(dir, dentry->d_name.name, dentry->d_name.len, &de); if (bh) { unsigned long ino = (unsigned long)le16_to_cpu(de->ino); brelse(bh); inode = bfs_iget(dir->i_sb, ino); if (IS_ERR(inode)) { - unlock_kernel(); + mutex_unlock(&info->bfs_lock); return ERR_CAST(inode); } } - unlock_kernel(); + mutex_unlock(&info->bfs_lock); d_add(dentry, inode); return NULL; } @@ -163,13 +165,14 @@ static int bfs_link(struct dentry *old, struct inode *dir, struct dentry *new) { struct inode *inode = old->d_inode; + struct bfs_sb_info *info = BFS_SB(inode->i_sb); int err; - lock_kernel(); + mutex_lock(&info->bfs_lock); err = bfs_add_entry(dir, new->d_name.name, new->d_name.len, inode->i_ino); if (err) { - unlock_kernel(); + mutex_unlock(&info->bfs_lock); return err; } inc_nlink(inode); @@ -177,19 +180,19 @@ static int bfs_link(struct dentry *old, struct inode *dir, mark_inode_dirty(inode); atomic_inc(&inode->i_count); d_instantiate(new, inode); - unlock_kernel(); + mutex_unlock(&info->bfs_lock); return 0; } static int bfs_unlink(struct inode *dir, struct dentry *dentry) { int error = -ENOENT; - struct inode *inode; + struct inode *inode = dentry->d_inode; struct buffer_head *bh; struct bfs_dirent *de; + struct bfs_sb_info *info = BFS_SB(inode->i_sb); - inode = dentry->d_inode; - lock_kernel(); + mutex_lock(&info->bfs_lock); bh = bfs_find_entry(dir, dentry->d_name.name, dentry->d_name.len, &de); if (!bh || (le16_to_cpu(de->ino) != inode->i_ino)) goto out_brelse; @@ -210,7 +213,7 @@ static int bfs_unlink(struct inode *dir, struct dentry *dentry) out_brelse: brelse(bh); - unlock_kernel(); + mutex_unlock(&info->bfs_lock); return error; } @@ -220,6 +223,7 @@ static int bfs_rename(struct inode *old_dir, struct dentry *old_dentry, struct inode *old_inode, *new_inode; struct buffer_head *old_bh, *new_bh; struct bfs_dirent *old_de, *new_de; + struct bfs_sb_info *info; int error = -ENOENT; old_bh = new_bh = NULL; @@ -227,7 +231,9 @@ static int bfs_rename(struct inode *old_dir, struct dentry *old_dentry, if (S_ISDIR(old_inode->i_mode)) return -EINVAL; - lock_kernel(); + info = BFS_SB(old_inode->i_sb); + + mutex_lock(&info->bfs_lock); old_bh = bfs_find_entry(old_dir, old_dentry->d_name.name, old_dentry->d_name.len, &old_de); @@ -264,7 +270,7 @@ static int bfs_rename(struct inode *old_dir, struct dentry *old_dentry, error = 0; end_rename: - unlock_kernel(); + mutex_unlock(&info->bfs_lock); brelse(old_bh); brelse(new_bh); return error; diff --git a/fs/bfs/file.c b/fs/bfs/file.c index b11e63e8fbc..6a021265f01 100644 --- a/fs/bfs/file.c +++ b/fs/bfs/file.c @@ -99,7 +99,7 @@ static int bfs_get_block(struct inode *inode, sector_t block, return -ENOSPC; /* The rest has to be protected against itself. */ - lock_kernel(); + mutex_lock(&info->bfs_lock); /* * If the last data block for this file is the last allocated @@ -151,7 +151,7 @@ static int bfs_get_block(struct inode *inode, sector_t block, mark_buffer_dirty(sbh); map_bh(bh_result, sb, phys); out: - unlock_kernel(); + mutex_unlock(&info->bfs_lock); return err; } diff --git a/fs/bfs/inode.c b/fs/bfs/inode.c index c8b3982e11c..053e690ec9e 100644 --- a/fs/bfs/inode.c +++ b/fs/bfs/inode.c @@ -104,6 +104,7 @@ static int bfs_write_inode(struct inode *inode, int unused) struct bfs_inode *di; struct buffer_head *bh; int block, off; + struct bfs_sb_info *info = BFS_SB(inode->i_sb); dprintf("ino=%08x\n", ino); @@ -112,13 +113,13 @@ static int bfs_write_inode(struct inode *inode, int unused) return -EIO; } - lock_kernel(); + mutex_lock(&info->bfs_lock); block = (ino - BFS_ROOT_INO) / BFS_INODES_PER_BLOCK + 1; bh = sb_bread(inode->i_sb, block); if (!bh) { printf("Unable to read inode %s:%08x\n", inode->i_sb->s_id, ino); - unlock_kernel(); + mutex_unlock(&info->bfs_lock); return -EIO; } @@ -145,7 +146,7 @@ static int bfs_write_inode(struct inode *inode, int unused) mark_buffer_dirty(bh); brelse(bh); - unlock_kernel(); + mutex_unlock(&info->bfs_lock); return 0; } @@ -170,7 +171,7 @@ static void bfs_delete_inode(struct inode *inode) inode->i_size = 0; inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME_SEC; - lock_kernel(); + mutex_lock(&info->bfs_lock); mark_inode_dirty(inode); block = (ino - BFS_ROOT_INO) / BFS_INODES_PER_BLOCK + 1; @@ -178,7 +179,7 @@ static void bfs_delete_inode(struct inode *inode) if (!bh) { printf("Unable to read inode %s:%08lx\n", inode->i_sb->s_id, ino); - unlock_kernel(); + mutex_unlock(&info->bfs_lock); return; } off = (ino - BFS_ROOT_INO) % BFS_INODES_PER_BLOCK; @@ -204,14 +205,16 @@ static void bfs_delete_inode(struct inode *inode) info->si_lf_eblk = bi->i_sblock - 1; mark_buffer_dirty(info->si_sbh); } - unlock_kernel(); + mutex_unlock(&info->bfs_lock); clear_inode(inode); } static void bfs_put_super(struct super_block *s) { struct bfs_sb_info *info = BFS_SB(s); + brelse(info->si_sbh); + mutex_destroy(&info->bfs_lock); kfree(info->si_imap); kfree(info); s->s_fs_info = NULL; @@ -236,11 +239,13 @@ static int bfs_statfs(struct dentry *dentry, struct kstatfs *buf) static void bfs_write_super(struct super_block *s) { - lock_kernel(); + struct bfs_sb_info *info = BFS_SB(s); + + mutex_lock(&info->bfs_lock); if (!(s->s_flags & MS_RDONLY)) - mark_buffer_dirty(BFS_SB(s)->si_sbh); + mark_buffer_dirty(info->si_sbh); s->s_dirt = 0; - unlock_kernel(); + mutex_unlock(&info->bfs_lock); } static struct kmem_cache *bfs_inode_cachep; @@ -409,6 +414,7 @@ static int bfs_fill_super(struct super_block *s, void *data, int silent) s->s_dirt = 1; } dump_imap("read_super", s); + mutex_init(&info->bfs_lock); return 0; out: -- cgit v1.2.3-70-g09d2 From 1b002d7b173ae7cc15ed90d3c07f6d106babc510 Mon Sep 17 00:00:00 2001 From: Bob Copeland Date: Fri, 25 Jul 2008 19:45:15 -0700 Subject: omfs: define filesystem structures Add header files containing OMFS on-disk and memory structures. Signed-off-by: Bob Copeland Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/omfs/omfs.h | 67 ++++++++++++++++++++++++++++++++++++++++++++++ fs/omfs/omfs_fs.h | 80 +++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 147 insertions(+) create mode 100644 fs/omfs/omfs.h create mode 100644 fs/omfs/omfs_fs.h (limited to 'fs') diff --git a/fs/omfs/omfs.h b/fs/omfs/omfs.h new file mode 100644 index 00000000000..2bc0f067040 --- /dev/null +++ b/fs/omfs/omfs.h @@ -0,0 +1,67 @@ +#ifndef _OMFS_H +#define _OMFS_H + +#include +#include + +#include "omfs_fs.h" + +/* In-memory structures */ +struct omfs_sb_info { + u64 s_num_blocks; + u64 s_bitmap_ino; + u64 s_root_ino; + u32 s_blocksize; + u32 s_mirrors; + u32 s_sys_blocksize; + u32 s_clustersize; + int s_block_shift; + unsigned long **s_imap; + int s_imap_size; + struct mutex s_bitmap_lock; + int s_uid; + int s_gid; + int s_dmask; + int s_fmask; +}; + +/* convert a cluster number to a scaled block number */ +static inline sector_t clus_to_blk(struct omfs_sb_info *sbi, sector_t block) +{ + return block << sbi->s_block_shift; +} + +static inline struct omfs_sb_info *OMFS_SB(struct super_block *sb) +{ + return sb->s_fs_info; +} + +/* bitmap.c */ +extern unsigned long omfs_count_free(struct super_block *sb); +extern int omfs_allocate_block(struct super_block *sb, u64 block); +extern int omfs_allocate_range(struct super_block *sb, int min_request, + int max_request, u64 *return_block, int *return_size); +extern int omfs_clear_range(struct super_block *sb, u64 block, int count); + +/* dir.c */ +extern struct file_operations omfs_dir_operations; +extern struct inode_operations omfs_dir_inops; +extern int omfs_make_empty(struct inode *inode, struct super_block *sb); +extern int omfs_is_bad(struct omfs_sb_info *sbi, struct omfs_header *header, + u64 fsblock); + +/* file.c */ +extern struct file_operations omfs_file_operations; +extern struct inode_operations omfs_file_inops; +extern struct address_space_operations omfs_aops; +extern void omfs_make_empty_table(struct buffer_head *bh, int offset); +extern int omfs_shrink_inode(struct inode *inode); + +/* inode.c */ +extern struct inode *omfs_iget(struct super_block *sb, ino_t inode); +extern struct inode *omfs_new_inode(struct inode *dir, int mode); +extern int omfs_reserve_block(struct super_block *sb, sector_t block); +extern int omfs_find_empty_block(struct super_block *sb, int mode, ino_t *ino); +extern int omfs_sync_inode(struct inode *inode); + +#endif diff --git a/fs/omfs/omfs_fs.h b/fs/omfs/omfs_fs.h new file mode 100644 index 00000000000..12cca245d6e --- /dev/null +++ b/fs/omfs/omfs_fs.h @@ -0,0 +1,80 @@ +#ifndef _OMFS_FS_H +#define _OMFS_FS_H + +/* OMFS On-disk structures */ + +#define OMFS_MAGIC 0xC2993D87 +#define OMFS_IMAGIC 0xD2 + +#define OMFS_DIR 'D' +#define OMFS_FILE 'F' +#define OMFS_INODE_NORMAL 'e' +#define OMFS_INODE_CONTINUATION 'c' +#define OMFS_INODE_SYSTEM 's' +#define OMFS_NAMELEN 256 +#define OMFS_DIR_START 0x1b8 +#define OMFS_EXTENT_START 0x1d0 +#define OMFS_EXTENT_CONT 0x40 +#define OMFS_XOR_COUNT 19 +#define OMFS_MAX_BLOCK_SIZE 8192 + +struct omfs_super_block { + char s_fill1[256]; + __be64 s_root_block; /* block number of omfs_root_block */ + __be64 s_num_blocks; /* total number of FS blocks */ + __be32 s_magic; /* OMFS_MAGIC */ + __be32 s_blocksize; /* size of a block */ + __be32 s_mirrors; /* # of mirrors of system blocks */ + __be32 s_sys_blocksize; /* size of non-data blocks */ +}; + +struct omfs_header { + __be64 h_self; /* FS block where this is located */ + __be32 h_body_size; /* size of useful data after header */ + __be16 h_crc; /* crc-ccitt of body_size bytes */ + char h_fill1[2]; + u8 h_version; /* version, always 1 */ + char h_type; /* OMFS_INODE_X */ + u8 h_magic; /* OMFS_IMAGIC */ + u8 h_check_xor; /* XOR of header bytes before this */ + __be32 h_fill2; +}; + +struct omfs_root_block { + struct omfs_header r_head; /* header */ + __be64 r_fill1; + __be64 r_num_blocks; /* total number of FS blocks */ + __be64 r_root_dir; /* block # of root directory */ + __be64 r_bitmap; /* block # of free space bitmap */ + __be32 r_blocksize; /* size of a block */ + __be32 r_clustersize; /* size allocated for data blocks */ + __be64 r_mirrors; /* # of mirrors of system blocks */ + char r_name[OMFS_NAMELEN]; /* partition label */ +}; + +struct omfs_inode { + struct omfs_header i_head; /* header */ + __be64 i_parent; /* parent containing this inode */ + __be64 i_sibling; /* next inode in hash bucket */ + __be64 i_ctime; /* ctime, in milliseconds */ + char i_fill1[35]; + char i_type; /* OMFS_[DIR,FILE] */ + __be32 i_fill2; + char i_fill3[64]; + char i_name[OMFS_NAMELEN]; /* filename */ + __be64 i_size; /* size of file, in bytes */ +}; + +struct omfs_extent_entry { + __be64 e_cluster; /* start location of a set of blocks */ + __be64 e_blocks; /* number of blocks after e_cluster */ +}; + +struct omfs_extent { + __be64 e_next; /* next extent table location */ + __be32 e_extent_count; /* total # extents in this table */ + __be32 e_fill; + struct omfs_extent_entry e_entry; /* start of extent entries */ +}; + +#endif -- cgit v1.2.3-70-g09d2 From 555e3775ced1d05203934fc6529bbf0560dd8733 Mon Sep 17 00:00:00 2001 From: Bob Copeland Date: Fri, 25 Jul 2008 19:45:15 -0700 Subject: omfs: add inode routines Add basic superblock and inode handling routines for OMFS Signed-off-by: Bob Copeland Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/omfs/inode.c | 553 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 553 insertions(+) create mode 100644 fs/omfs/inode.c (limited to 'fs') diff --git a/fs/omfs/inode.c b/fs/omfs/inode.c new file mode 100644 index 00000000000..d865f553543 --- /dev/null +++ b/fs/omfs/inode.c @@ -0,0 +1,553 @@ +/* + * Optimized MPEG FS - inode and super operations. + * Copyright (C) 2006 Bob Copeland + * Released under GPL v2. + */ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include "omfs.h" + +MODULE_AUTHOR("Bob Copeland "); +MODULE_DESCRIPTION("OMFS (ReplayTV/Karma) Filesystem for Linux"); +MODULE_LICENSE("GPL"); + +struct inode *omfs_new_inode(struct inode *dir, int mode) +{ + struct inode *inode; + u64 new_block; + int err; + int len; + struct omfs_sb_info *sbi = OMFS_SB(dir->i_sb); + + inode = new_inode(dir->i_sb); + if (!inode) + return ERR_PTR(-ENOMEM); + + err = omfs_allocate_range(dir->i_sb, sbi->s_mirrors, sbi->s_mirrors, + &new_block, &len); + if (err) + goto fail; + + inode->i_ino = new_block; + inode->i_mode = mode; + inode->i_uid = current->fsuid; + inode->i_gid = current->fsgid; + inode->i_blocks = 0; + inode->i_mapping->a_ops = &omfs_aops; + + inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME; + switch (mode & S_IFMT) { + case S_IFDIR: + inode->i_op = &omfs_dir_inops; + inode->i_fop = &omfs_dir_operations; + inode->i_size = sbi->s_sys_blocksize; + inc_nlink(inode); + break; + case S_IFREG: + inode->i_op = &omfs_file_inops; + inode->i_fop = &omfs_file_operations; + inode->i_size = 0; + break; + } + + insert_inode_hash(inode); + mark_inode_dirty(inode); + return inode; +fail: + make_bad_inode(inode); + iput(inode); + return ERR_PTR(err); +} + +/* + * Update the header checksums for a dirty inode based on its contents. + * Caller is expected to hold the buffer head underlying oi and mark it + * dirty. + */ +static void omfs_update_checksums(struct omfs_inode *oi) +{ + int xor, i, ofs = 0, count; + u16 crc = 0; + unsigned char *ptr = (unsigned char *) oi; + + count = be32_to_cpu(oi->i_head.h_body_size); + ofs = sizeof(struct omfs_header); + + crc = crc_itu_t(crc, ptr + ofs, count); + oi->i_head.h_crc = cpu_to_be16(crc); + + xor = ptr[0]; + for (i = 1; i < OMFS_XOR_COUNT; i++) + xor ^= ptr[i]; + + oi->i_head.h_check_xor = xor; +} + +static int omfs_write_inode(struct inode *inode, int wait) +{ + struct omfs_inode *oi; + struct omfs_sb_info *sbi = OMFS_SB(inode->i_sb); + struct buffer_head *bh, *bh2; + unsigned int block; + u64 ctime; + int i; + int ret = -EIO; + int sync_failed = 0; + + /* get current inode since we may have written sibling ptrs etc. */ + block = clus_to_blk(sbi, inode->i_ino); + bh = sb_bread(inode->i_sb, block); + if (!bh) + goto out; + + oi = (struct omfs_inode *) bh->b_data; + + oi->i_head.h_self = cpu_to_be64(inode->i_ino); + if (S_ISDIR(inode->i_mode)) + oi->i_type = OMFS_DIR; + else if (S_ISREG(inode->i_mode)) + oi->i_type = OMFS_FILE; + else { + printk(KERN_WARNING "omfs: unknown file type: %d\n", + inode->i_mode); + goto out_brelse; + } + + oi->i_head.h_body_size = cpu_to_be32(sbi->s_sys_blocksize - + sizeof(struct omfs_header)); + oi->i_head.h_version = 1; + oi->i_head.h_type = OMFS_INODE_NORMAL; + oi->i_head.h_magic = OMFS_IMAGIC; + oi->i_size = cpu_to_be64(inode->i_size); + + ctime = inode->i_ctime.tv_sec * 1000LL + + ((inode->i_ctime.tv_nsec + 999)/1000); + oi->i_ctime = cpu_to_be64(ctime); + + omfs_update_checksums(oi); + + mark_buffer_dirty(bh); + if (wait) { + sync_dirty_buffer(bh); + if (buffer_req(bh) && !buffer_uptodate(bh)) + sync_failed = 1; + } + + /* if mirroring writes, copy to next fsblock */ + for (i = 1; i < sbi->s_mirrors; i++) { + bh2 = sb_bread(inode->i_sb, block + i * + (sbi->s_blocksize / sbi->s_sys_blocksize)); + if (!bh2) + goto out_brelse; + + memcpy(bh2->b_data, bh->b_data, bh->b_size); + mark_buffer_dirty(bh2); + if (wait) { + sync_dirty_buffer(bh2); + if (buffer_req(bh2) && !buffer_uptodate(bh2)) + sync_failed = 1; + } + brelse(bh2); + } + ret = (sync_failed) ? -EIO : 0; +out_brelse: + brelse(bh); +out: + return ret; +} + +int omfs_sync_inode(struct inode *inode) +{ + return omfs_write_inode(inode, 1); +} + +/* + * called when an entry is deleted, need to clear the bits in the + * bitmaps. + */ +static void omfs_delete_inode(struct inode *inode) +{ + truncate_inode_pages(&inode->i_data, 0); + + if (S_ISREG(inode->i_mode)) { + inode->i_size = 0; + omfs_shrink_inode(inode); + } + + omfs_clear_range(inode->i_sb, inode->i_ino, 2); + clear_inode(inode); +} + +struct inode *omfs_iget(struct super_block *sb, ino_t ino) +{ + struct omfs_sb_info *sbi = OMFS_SB(sb); + struct omfs_inode *oi; + struct buffer_head *bh; + unsigned int block; + u64 ctime; + unsigned long nsecs; + struct inode *inode; + + inode = iget_locked(sb, ino); + if (!inode) + return ERR_PTR(-ENOMEM); + if (!(inode->i_state & I_NEW)) + return inode; + + block = clus_to_blk(sbi, ino); + bh = sb_bread(inode->i_sb, block); + if (!bh) + goto iget_failed; + + oi = (struct omfs_inode *)bh->b_data; + + /* check self */ + if (ino != be64_to_cpu(oi->i_head.h_self)) + goto fail_bh; + + inode->i_uid = sbi->s_uid; + inode->i_gid = sbi->s_gid; + + ctime = be64_to_cpu(oi->i_ctime); + nsecs = do_div(ctime, 1000) * 1000L; + + inode->i_atime.tv_sec = ctime; + inode->i_mtime.tv_sec = ctime; + inode->i_ctime.tv_sec = ctime; + inode->i_atime.tv_nsec = nsecs; + inode->i_mtime.tv_nsec = nsecs; + inode->i_ctime.tv_nsec = nsecs; + + inode->i_mapping->a_ops = &omfs_aops; + + switch (oi->i_type) { + case OMFS_DIR: + inode->i_mode = S_IFDIR | (S_IRWXUGO & ~sbi->s_dmask); + inode->i_op = &omfs_dir_inops; + inode->i_fop = &omfs_dir_operations; + inode->i_size = be32_to_cpu(oi->i_head.h_body_size) + + sizeof(struct omfs_header); + inc_nlink(inode); + break; + case OMFS_FILE: + inode->i_mode = S_IFREG | (S_IRWXUGO & ~sbi->s_fmask); + inode->i_fop = &omfs_file_operations; + inode->i_size = be64_to_cpu(oi->i_size); + break; + } + brelse(bh); + unlock_new_inode(inode); + return inode; +fail_bh: + brelse(bh); +iget_failed: + iget_failed(inode); + return ERR_PTR(-EIO); +} + +static void omfs_put_super(struct super_block *sb) +{ + struct omfs_sb_info *sbi = OMFS_SB(sb); + kfree(sbi->s_imap); + kfree(sbi); + sb->s_fs_info = NULL; +} + +static int omfs_statfs(struct dentry *dentry, struct kstatfs *buf) +{ + struct super_block *s = dentry->d_sb; + struct omfs_sb_info *sbi = OMFS_SB(s); + buf->f_type = OMFS_MAGIC; + buf->f_bsize = sbi->s_blocksize; + buf->f_blocks = sbi->s_num_blocks; + buf->f_files = sbi->s_num_blocks; + buf->f_namelen = OMFS_NAMELEN; + + buf->f_bfree = buf->f_bavail = buf->f_ffree = + omfs_count_free(s); + return 0; +} + +static struct super_operations omfs_sops = { + .write_inode = omfs_write_inode, + .delete_inode = omfs_delete_inode, + .put_super = omfs_put_super, + .statfs = omfs_statfs, + .show_options = generic_show_options, +}; + +/* + * For Rio Karma, there is an on-disk free bitmap whose location is + * stored in the root block. For ReplayTV, there is no such free bitmap + * so we have to walk the tree. Both inodes and file data are allocated + * from the same map. This array can be big (300k) so we allocate + * in units of the blocksize. + */ +static int omfs_get_imap(struct super_block *sb) +{ + int bitmap_size; + int array_size; + int count; + struct omfs_sb_info *sbi = OMFS_SB(sb); + struct buffer_head *bh; + unsigned long **ptr; + sector_t block; + + bitmap_size = DIV_ROUND_UP(sbi->s_num_blocks, 8); + array_size = DIV_ROUND_UP(bitmap_size, sb->s_blocksize); + + if (sbi->s_bitmap_ino == ~0ULL) + goto out; + + sbi->s_imap_size = array_size; + sbi->s_imap = kzalloc(array_size * sizeof(unsigned long *), GFP_KERNEL); + if (!sbi->s_imap) + goto nomem; + + block = clus_to_blk(sbi, sbi->s_bitmap_ino); + ptr = sbi->s_imap; + for (count = bitmap_size; count > 0; count -= sb->s_blocksize) { + bh = sb_bread(sb, block++); + if (!bh) + goto nomem_free; + *ptr = kmalloc(sb->s_blocksize, GFP_KERNEL); + if (!*ptr) { + brelse(bh); + goto nomem_free; + } + memcpy(*ptr, bh->b_data, sb->s_blocksize); + if (count < sb->s_blocksize) + memset((void *)*ptr + count, 0xff, + sb->s_blocksize - count); + brelse(bh); + ptr++; + } +out: + return 0; + +nomem_free: + for (count = 0; count < array_size; count++) + kfree(sbi->s_imap[count]); + + kfree(sbi->s_imap); +nomem: + sbi->s_imap = NULL; + sbi->s_imap_size = 0; + return -ENOMEM; +} + +enum { + Opt_uid, Opt_gid, Opt_umask, Opt_dmask, Opt_fmask +}; + +static match_table_t tokens = { + {Opt_uid, "uid=%u"}, + {Opt_gid, "gid=%u"}, + {Opt_umask, "umask=%o"}, + {Opt_dmask, "dmask=%o"}, + {Opt_fmask, "fmask=%o"}, +}; + +static int parse_options(char *options, struct omfs_sb_info *sbi) +{ + char *p; + substring_t args[MAX_OPT_ARGS]; + int option; + + if (!options) + return 1; + + while ((p = strsep(&options, ",")) != NULL) { + int token; + if (!*p) + continue; + + token = match_token(p, tokens, args); + switch (token) { + case Opt_uid: + if (match_int(&args[0], &option)) + return 0; + sbi->s_uid = option; + break; + case Opt_gid: + if (match_int(&args[0], &option)) + return 0; + sbi->s_gid = option; + break; + case Opt_umask: + if (match_octal(&args[0], &option)) + return 0; + sbi->s_fmask = sbi->s_dmask = option; + break; + case Opt_dmask: + if (match_octal(&args[0], &option)) + return 0; + sbi->s_dmask = option; + break; + case Opt_fmask: + if (match_octal(&args[0], &option)) + return 0; + sbi->s_fmask = option; + break; + default: + return 0; + } + } + return 1; +} + +static int omfs_fill_super(struct super_block *sb, void *data, int silent) +{ + struct buffer_head *bh, *bh2; + struct omfs_super_block *omfs_sb; + struct omfs_root_block *omfs_rb; + struct omfs_sb_info *sbi; + struct inode *root; + sector_t start; + int ret = -EINVAL; + + save_mount_options(sb, (char *) data); + + sbi = kzalloc(sizeof(struct omfs_sb_info), GFP_KERNEL); + if (!sbi) + return -ENOMEM; + + sb->s_fs_info = sbi; + + sbi->s_uid = current->uid; + sbi->s_gid = current->gid; + sbi->s_dmask = sbi->s_fmask = current->fs->umask; + + if (!parse_options((char *) data, sbi)) + goto end; + + sb->s_maxbytes = 0xffffffff; + + sb_set_blocksize(sb, 0x200); + + bh = sb_bread(sb, 0); + if (!bh) + goto end; + + omfs_sb = (struct omfs_super_block *)bh->b_data; + + if (omfs_sb->s_magic != cpu_to_be32(OMFS_MAGIC)) { + if (!silent) + printk(KERN_ERR "omfs: Invalid superblock (%x)\n", + omfs_sb->s_magic); + goto out_brelse_bh; + } + sb->s_magic = OMFS_MAGIC; + + sbi->s_num_blocks = be64_to_cpu(omfs_sb->s_num_blocks); + sbi->s_blocksize = be32_to_cpu(omfs_sb->s_blocksize); + sbi->s_mirrors = be32_to_cpu(omfs_sb->s_mirrors); + sbi->s_root_ino = be64_to_cpu(omfs_sb->s_root_block); + sbi->s_sys_blocksize = be32_to_cpu(omfs_sb->s_sys_blocksize); + mutex_init(&sbi->s_bitmap_lock); + + if (sbi->s_sys_blocksize > PAGE_SIZE) { + printk(KERN_ERR "omfs: sysblock size (%d) is out of range\n", + sbi->s_sys_blocksize); + goto out_brelse_bh; + } + + if (sbi->s_blocksize < sbi->s_sys_blocksize || + sbi->s_blocksize > OMFS_MAX_BLOCK_SIZE) { + printk(KERN_ERR "omfs: block size (%d) is out of range\n", + sbi->s_blocksize); + goto out_brelse_bh; + } + + /* + * Use sys_blocksize as the fs block since it is smaller than a + * page while the fs blocksize can be larger. + */ + sb_set_blocksize(sb, sbi->s_sys_blocksize); + + /* + * ...and the difference goes into a shift. sys_blocksize is always + * a power of two factor of blocksize. + */ + sbi->s_block_shift = get_bitmask_order(sbi->s_blocksize) - + get_bitmask_order(sbi->s_sys_blocksize); + + start = clus_to_blk(sbi, be64_to_cpu(omfs_sb->s_root_block)); + bh2 = sb_bread(sb, start); + if (!bh2) + goto out_brelse_bh; + + omfs_rb = (struct omfs_root_block *)bh2->b_data; + + sbi->s_bitmap_ino = be64_to_cpu(omfs_rb->r_bitmap); + sbi->s_clustersize = be32_to_cpu(omfs_rb->r_clustersize); + + if (sbi->s_num_blocks != be64_to_cpu(omfs_rb->r_num_blocks)) { + printk(KERN_ERR "omfs: block count discrepancy between " + "super and root blocks (%llx, %llx)\n", + sbi->s_num_blocks, be64_to_cpu(omfs_rb->r_num_blocks)); + goto out_brelse_bh2; + } + + ret = omfs_get_imap(sb); + if (ret) + goto out_brelse_bh2; + + sb->s_op = &omfs_sops; + + root = omfs_iget(sb, be64_to_cpu(omfs_rb->r_root_dir)); + if (IS_ERR(root)) { + ret = PTR_ERR(root); + goto out_brelse_bh2; + } + + sb->s_root = d_alloc_root(root); + if (!sb->s_root) { + iput(root); + goto out_brelse_bh2; + } + printk(KERN_DEBUG "omfs: Mounted volume %s\n", omfs_rb->r_name); + + ret = 0; +out_brelse_bh2: + brelse(bh2); +out_brelse_bh: + brelse(bh); +end: + return ret; +} + +static int omfs_get_sb(struct file_system_type *fs_type, + int flags, const char *dev_name, + void *data, struct vfsmount *m) +{ + return get_sb_bdev(fs_type, flags, dev_name, data, omfs_fill_super, m); +} + +static struct file_system_type omfs_fs_type = { + .owner = THIS_MODULE, + .name = "omfs", + .get_sb = omfs_get_sb, + .kill_sb = kill_block_super, + .fs_flags = FS_REQUIRES_DEV, +}; + +static int __init init_omfs_fs(void) +{ + return register_filesystem(&omfs_fs_type); +} + +static void __exit exit_omfs_fs(void) +{ + unregister_filesystem(&omfs_fs_type); +} + +module_init(init_omfs_fs); +module_exit(exit_omfs_fs); -- cgit v1.2.3-70-g09d2 From a3ab7155ea21aadc8a4d5687e91b3d876973185e Mon Sep 17 00:00:00 2001 From: Bob Copeland Date: Fri, 25 Jul 2008 19:45:16 -0700 Subject: omfs: add directory routines Add lookup and directory management routines for OMFS. The filesystem uses hashing based on the filename and stores collisions, unordered, in siblings of files' inode structures. To support telldir, the current position in the hash table is encoded in fpos. Signed-off-by: Bob Copeland Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/omfs/dir.c | 504 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 504 insertions(+) create mode 100644 fs/omfs/dir.c (limited to 'fs') diff --git a/fs/omfs/dir.c b/fs/omfs/dir.c new file mode 100644 index 00000000000..05a5bc31e4b --- /dev/null +++ b/fs/omfs/dir.c @@ -0,0 +1,504 @@ +/* + * OMFS (as used by RIO Karma) directory operations. + * Copyright (C) 2005 Bob Copeland + * Released under GPL v2. + */ + +#include +#include +#include +#include "omfs.h" + +static int omfs_hash(const char *name, int namelen, int mod) +{ + int i, hash = 0; + for (i = 0; i < namelen; i++) + hash ^= tolower(name[i]) << (i % 24); + return hash % mod; +} + +/* + * Finds the bucket for a given name and reads the containing block; + * *ofs is set to the offset of the first list entry. + */ +static struct buffer_head *omfs_get_bucket(struct inode *dir, + const char *name, int namelen, int *ofs) +{ + int nbuckets = (dir->i_size - OMFS_DIR_START)/8; + int block = clus_to_blk(OMFS_SB(dir->i_sb), dir->i_ino); + int bucket = omfs_hash(name, namelen, nbuckets); + + *ofs = OMFS_DIR_START + bucket * 8; + return sb_bread(dir->i_sb, block); +} + +static struct buffer_head *omfs_scan_list(struct inode *dir, u64 block, + const char *name, int namelen, + u64 *prev_block) +{ + struct buffer_head *bh; + struct omfs_inode *oi; + int err = -ENOENT; + *prev_block = ~0; + + while (block != ~0) { + bh = sb_bread(dir->i_sb, + clus_to_blk(OMFS_SB(dir->i_sb), block)); + if (!bh) { + err = -EIO; + goto err; + } + + oi = (struct omfs_inode *) bh->b_data; + if (omfs_is_bad(OMFS_SB(dir->i_sb), &oi->i_head, block)) { + brelse(bh); + goto err; + } + + if (strncmp(oi->i_name, name, namelen) == 0) + return bh; + + *prev_block = block; + block = be64_to_cpu(oi->i_sibling); + brelse(bh); + } +err: + return ERR_PTR(err); +} + +static struct buffer_head *omfs_find_entry(struct inode *dir, + const char *name, int namelen) +{ + struct buffer_head *bh; + int ofs; + u64 block, dummy; + + bh = omfs_get_bucket(dir, name, namelen, &ofs); + if (!bh) + return ERR_PTR(-EIO); + + block = be64_to_cpu(*((__be64 *) &bh->b_data[ofs])); + brelse(bh); + + return omfs_scan_list(dir, block, name, namelen, &dummy); +} + +int omfs_make_empty(struct inode *inode, struct super_block *sb) +{ + struct omfs_sb_info *sbi = OMFS_SB(sb); + int block = clus_to_blk(sbi, inode->i_ino); + struct buffer_head *bh; + struct omfs_inode *oi; + + bh = sb_bread(sb, block); + if (!bh) + return -ENOMEM; + + memset(bh->b_data, 0, sizeof(struct omfs_inode)); + + if (inode->i_mode & S_IFDIR) { + memset(&bh->b_data[OMFS_DIR_START], 0xff, + sbi->s_sys_blocksize - OMFS_DIR_START); + } else + omfs_make_empty_table(bh, OMFS_EXTENT_START); + + oi = (struct omfs_inode *) bh->b_data; + oi->i_head.h_self = cpu_to_be64(inode->i_ino); + oi->i_sibling = ~0ULL; + + mark_buffer_dirty(bh); + brelse(bh); + return 0; +} + +static int omfs_add_link(struct dentry *dentry, struct inode *inode) +{ + struct inode *dir = dentry->d_parent->d_inode; + const char *name = dentry->d_name.name; + int namelen = dentry->d_name.len; + struct omfs_inode *oi; + struct buffer_head *bh; + u64 block; + __be64 *entry; + int ofs; + + /* just prepend to head of queue in proper bucket */ + bh = omfs_get_bucket(dir, name, namelen, &ofs); + if (!bh) + goto out; + + entry = (__be64 *) &bh->b_data[ofs]; + block = be64_to_cpu(*entry); + *entry = cpu_to_be64(inode->i_ino); + mark_buffer_dirty(bh); + brelse(bh); + + /* now set the sibling and parent pointers on the new inode */ + bh = sb_bread(dir->i_sb, clus_to_blk(OMFS_SB(dir->i_sb), inode->i_ino)); + if (!bh) + goto out; + + oi = (struct omfs_inode *) bh->b_data; + memcpy(oi->i_name, name, namelen); + memset(oi->i_name + namelen, 0, OMFS_NAMELEN - namelen); + oi->i_sibling = cpu_to_be64(block); + oi->i_parent = cpu_to_be64(dir->i_ino); + mark_buffer_dirty(bh); + brelse(bh); + + dir->i_ctime = CURRENT_TIME_SEC; + + /* mark affected inodes dirty to rebuild checksums */ + mark_inode_dirty(dir); + mark_inode_dirty(inode); + return 0; +out: + return -ENOMEM; +} + +static int omfs_delete_entry(struct dentry *dentry) +{ + struct inode *dir = dentry->d_parent->d_inode; + struct inode *dirty; + const char *name = dentry->d_name.name; + int namelen = dentry->d_name.len; + struct omfs_inode *oi; + struct buffer_head *bh, *bh2; + __be64 *entry, next; + u64 block, prev; + int ofs; + int err = -ENOMEM; + + /* delete the proper node in the bucket's linked list */ + bh = omfs_get_bucket(dir, name, namelen, &ofs); + if (!bh) + goto out; + + entry = (__be64 *) &bh->b_data[ofs]; + block = be64_to_cpu(*entry); + + bh2 = omfs_scan_list(dir, block, name, namelen, &prev); + if (IS_ERR(bh2)) { + err = PTR_ERR(bh2); + goto out_free_bh; + } + + oi = (struct omfs_inode *) bh2->b_data; + next = oi->i_sibling; + brelse(bh2); + + if (prev != ~0) { + /* found in middle of list, get list ptr */ + brelse(bh); + bh = sb_bread(dir->i_sb, + clus_to_blk(OMFS_SB(dir->i_sb), prev)); + if (!bh) + goto out; + + oi = (struct omfs_inode *) bh->b_data; + entry = &oi->i_sibling; + } + + *entry = next; + mark_buffer_dirty(bh); + + if (prev != ~0) { + dirty = omfs_iget(dir->i_sb, prev); + if (!IS_ERR(dirty)) { + mark_inode_dirty(dirty); + iput(dirty); + } + } + + err = 0; +out_free_bh: + brelse(bh); +out: + return err; +} + +static int omfs_dir_is_empty(struct inode *inode) +{ + int nbuckets = (inode->i_size - OMFS_DIR_START) / 8; + struct buffer_head *bh; + u64 *ptr; + int i; + + bh = sb_bread(inode->i_sb, clus_to_blk(OMFS_SB(inode->i_sb), + inode->i_ino)); + + if (!bh) + return 0; + + ptr = (u64 *) &bh->b_data[OMFS_DIR_START]; + + for (i = 0; i < nbuckets; i++, ptr++) + if (*ptr != ~0) + break; + + brelse(bh); + return *ptr != ~0; +} + +static int omfs_unlink(struct inode *dir, struct dentry *dentry) +{ + int ret; + struct inode *inode = dentry->d_inode; + + ret = omfs_delete_entry(dentry); + if (ret) + goto end_unlink; + + inode_dec_link_count(inode); + mark_inode_dirty(dir); + +end_unlink: + return ret; +} + +static int omfs_rmdir(struct inode *dir, struct dentry *dentry) +{ + int err = -ENOTEMPTY; + struct inode *inode = dentry->d_inode; + + if (omfs_dir_is_empty(inode)) { + err = omfs_unlink(dir, dentry); + if (!err) + inode_dec_link_count(inode); + } + return err; +} + +static int omfs_add_node(struct inode *dir, struct dentry *dentry, int mode) +{ + int err; + struct inode *inode = omfs_new_inode(dir, mode); + + if (IS_ERR(inode)) + return PTR_ERR(inode); + + err = omfs_make_empty(inode, dir->i_sb); + if (err) + goto out_free_inode; + + err = omfs_add_link(dentry, inode); + if (err) + goto out_free_inode; + + d_instantiate(dentry, inode); + return 0; + +out_free_inode: + iput(inode); + return err; +} + +static int omfs_mkdir(struct inode *dir, struct dentry *dentry, int mode) +{ + return omfs_add_node(dir, dentry, mode | S_IFDIR); +} + +static int omfs_create(struct inode *dir, struct dentry *dentry, int mode, + struct nameidata *nd) +{ + return omfs_add_node(dir, dentry, mode | S_IFREG); +} + +static struct dentry *omfs_lookup(struct inode *dir, struct dentry *dentry, + struct nameidata *nd) +{ + struct buffer_head *bh; + struct inode *inode = NULL; + + if (dentry->d_name.len > OMFS_NAMELEN) + return ERR_PTR(-ENAMETOOLONG); + + bh = omfs_find_entry(dir, dentry->d_name.name, dentry->d_name.len); + if (!IS_ERR(bh)) { + struct omfs_inode *oi = (struct omfs_inode *)bh->b_data; + ino_t ino = be64_to_cpu(oi->i_head.h_self); + brelse(bh); + inode = omfs_iget(dir->i_sb, ino); + if (IS_ERR(inode)) + return ERR_CAST(inode); + } + d_add(dentry, inode); + return NULL; +} + +/* sanity check block's self pointer */ +int omfs_is_bad(struct omfs_sb_info *sbi, struct omfs_header *header, + u64 fsblock) +{ + int is_bad; + u64 ino = be64_to_cpu(header->h_self); + is_bad = ((ino != fsblock) || (ino < sbi->s_root_ino) || + (ino > sbi->s_num_blocks)); + + if (is_bad) + printk(KERN_WARNING "omfs: bad hash chain detected\n"); + + return is_bad; +} + +static int omfs_fill_chain(struct file *filp, void *dirent, filldir_t filldir, + u64 fsblock, int hindex) +{ + struct inode *dir = filp->f_dentry->d_inode; + struct buffer_head *bh; + struct omfs_inode *oi; + u64 self; + int res = 0; + unsigned char d_type; + + /* follow chain in this bucket */ + while (fsblock != ~0) { + bh = sb_bread(dir->i_sb, clus_to_blk(OMFS_SB(dir->i_sb), + fsblock)); + if (!bh) + goto out; + + oi = (struct omfs_inode *) bh->b_data; + if (omfs_is_bad(OMFS_SB(dir->i_sb), &oi->i_head, fsblock)) { + brelse(bh); + goto out; + } + + self = fsblock; + fsblock = be64_to_cpu(oi->i_sibling); + + /* skip visited nodes */ + if (hindex) { + hindex--; + brelse(bh); + continue; + } + + d_type = (oi->i_type == OMFS_DIR) ? DT_DIR : DT_REG; + + res = filldir(dirent, oi->i_name, strnlen(oi->i_name, + OMFS_NAMELEN), filp->f_pos, self, d_type); + if (res == 0) + filp->f_pos++; + brelse(bh); + } +out: + return res; +} + +static int omfs_rename(struct inode *old_dir, struct dentry *old_dentry, + struct inode *new_dir, struct dentry *new_dentry) +{ + struct inode *new_inode = new_dentry->d_inode; + struct inode *old_inode = old_dentry->d_inode; + struct buffer_head *bh; + int is_dir; + int err; + + is_dir = S_ISDIR(old_inode->i_mode); + + if (new_inode) { + /* overwriting existing file/dir */ + err = -ENOTEMPTY; + if (is_dir && !omfs_dir_is_empty(new_inode)) + goto out; + + err = -ENOENT; + bh = omfs_find_entry(new_dir, new_dentry->d_name.name, + new_dentry->d_name.len); + if (IS_ERR(bh)) + goto out; + brelse(bh); + + err = omfs_unlink(new_dir, new_dentry); + if (err) + goto out; + } + + /* since omfs locates files by name, we need to unlink _before_ + * adding the new link or we won't find the old one */ + inode_inc_link_count(old_inode); + err = omfs_unlink(old_dir, old_dentry); + if (err) { + inode_dec_link_count(old_inode); + goto out; + } + + err = omfs_add_link(new_dentry, old_inode); + if (err) + goto out; + + old_inode->i_ctime = CURRENT_TIME_SEC; +out: + return err; +} + +static int omfs_readdir(struct file *filp, void *dirent, filldir_t filldir) +{ + struct inode *dir = filp->f_dentry->d_inode; + struct buffer_head *bh; + loff_t offset, res; + unsigned int hchain, hindex; + int nbuckets; + u64 fsblock; + int ret = -EINVAL; + + if (filp->f_pos >> 32) + goto success; + + switch ((unsigned long) filp->f_pos) { + case 0: + if (filldir(dirent, ".", 1, 0, dir->i_ino, DT_DIR) < 0) + goto success; + filp->f_pos++; + /* fall through */ + case 1: + if (filldir(dirent, "..", 2, 1, + parent_ino(filp->f_dentry), DT_DIR) < 0) + goto success; + filp->f_pos = 1 << 20; + /* fall through */ + } + + nbuckets = (dir->i_size - OMFS_DIR_START) / 8; + + /* high 12 bits store bucket + 1 and low 20 bits store hash index */ + hchain = (filp->f_pos >> 20) - 1; + hindex = filp->f_pos & 0xfffff; + + bh = sb_bread(dir->i_sb, clus_to_blk(OMFS_SB(dir->i_sb), dir->i_ino)); + if (!bh) + goto out; + + offset = OMFS_DIR_START + hchain * 8; + + for (; hchain < nbuckets; hchain++, offset += 8) { + fsblock = be64_to_cpu(*((__be64 *) &bh->b_data[offset])); + + res = omfs_fill_chain(filp, dirent, filldir, fsblock, hindex); + hindex = 0; + if (res < 0) + break; + + filp->f_pos = (hchain+2) << 20; + } + brelse(bh); +success: + ret = 0; +out: + return ret; +} + +struct inode_operations omfs_dir_inops = { + .lookup = omfs_lookup, + .mkdir = omfs_mkdir, + .rename = omfs_rename, + .create = omfs_create, + .unlink = omfs_unlink, + .rmdir = omfs_rmdir, +}; + +struct file_operations omfs_dir_operations = { + .read = generic_read_dir, + .readdir = omfs_readdir, +}; -- cgit v1.2.3-70-g09d2 From 8f09e98768c17287df076580c4cc72ac358312c6 Mon Sep 17 00:00:00 2001 From: Bob Copeland Date: Fri, 25 Jul 2008 19:45:16 -0700 Subject: omfs: add file routines Add functions for reading and manipulating the storage of file data in the extent-based OMFS. Signed-off-by: Bob Copeland Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/omfs/file.c | 346 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 346 insertions(+) create mode 100644 fs/omfs/file.c (limited to 'fs') diff --git a/fs/omfs/file.c b/fs/omfs/file.c new file mode 100644 index 00000000000..66e01fae438 --- /dev/null +++ b/fs/omfs/file.c @@ -0,0 +1,346 @@ +/* + * OMFS (as used by RIO Karma) file operations. + * Copyright (C) 2005 Bob Copeland + * Released under GPL v2. + */ + +#include +#include +#include +#include +#include +#include "omfs.h" + +static int omfs_sync_file(struct file *file, struct dentry *dentry, + int datasync) +{ + struct inode *inode = dentry->d_inode; + int err; + + err = sync_mapping_buffers(inode->i_mapping); + if (!(inode->i_state & I_DIRTY)) + return err; + if (datasync && !(inode->i_state & I_DIRTY_DATASYNC)) + return err; + err |= omfs_sync_inode(inode); + return err ? -EIO : 0; +} + +void omfs_make_empty_table(struct buffer_head *bh, int offset) +{ + struct omfs_extent *oe = (struct omfs_extent *) &bh->b_data[offset]; + + oe->e_next = ~0ULL; + oe->e_extent_count = cpu_to_be32(1), + oe->e_fill = cpu_to_be32(0x22), + oe->e_entry.e_cluster = ~0ULL; + oe->e_entry.e_blocks = ~0ULL; +} + +int omfs_shrink_inode(struct inode *inode) +{ + struct omfs_sb_info *sbi = OMFS_SB(inode->i_sb); + struct omfs_extent *oe; + struct omfs_extent_entry *entry; + struct buffer_head *bh; + u64 next, last; + u32 extent_count; + int ret; + + /* traverse extent table, freeing each entry that is greater + * than inode->i_size; + */ + next = inode->i_ino; + + /* only support truncate -> 0 for now */ + ret = -EIO; + if (inode->i_size != 0) + goto out; + + bh = sb_bread(inode->i_sb, clus_to_blk(sbi, next)); + if (!bh) + goto out; + + oe = (struct omfs_extent *)(&bh->b_data[OMFS_EXTENT_START]); + + for (;;) { + + if (omfs_is_bad(sbi, (struct omfs_header *) bh->b_data, next)) { + brelse(bh); + goto out; + } + + extent_count = be32_to_cpu(oe->e_extent_count); + last = next; + next = be64_to_cpu(oe->e_next); + entry = &oe->e_entry; + + /* ignore last entry as it is the terminator */ + for (; extent_count > 1; extent_count--) { + u64 start, count; + start = be64_to_cpu(entry->e_cluster); + count = be64_to_cpu(entry->e_blocks); + + omfs_clear_range(inode->i_sb, start, (int) count); + entry++; + } + omfs_make_empty_table(bh, (char *) oe - bh->b_data); + mark_buffer_dirty(bh); + brelse(bh); + + if (last != inode->i_ino) + omfs_clear_range(inode->i_sb, last, sbi->s_mirrors); + + if (next == ~0) + break; + + bh = sb_bread(inode->i_sb, clus_to_blk(sbi, next)); + if (!bh) + goto out; + oe = (struct omfs_extent *) (&bh->b_data[OMFS_EXTENT_CONT]); + } + ret = 0; +out: + return ret; +} + +static void omfs_truncate(struct inode *inode) +{ + omfs_shrink_inode(inode); + mark_inode_dirty(inode); +} + +/* + * Add new blocks to the current extent, or create new entries/continuations + * as necessary. + */ +static int omfs_grow_extent(struct inode *inode, struct omfs_extent *oe, + u64 *ret_block) +{ + struct omfs_extent_entry *terminator; + struct omfs_extent_entry *entry = &oe->e_entry; + struct omfs_sb_info *sbi = OMFS_SB(inode->i_sb); + u32 extent_count = be32_to_cpu(oe->e_extent_count); + u64 new_block = 0; + u32 max_count; + int new_count; + int ret = 0; + + /* reached the end of the extent table with no blocks mapped. + * there are three possibilities for adding: grow last extent, + * add a new extent to the current extent table, and add a + * continuation inode. in last two cases need an allocator for + * sbi->s_cluster_size + */ + + /* TODO: handle holes */ + + /* should always have a terminator */ + if (extent_count < 1) + return -EIO; + + /* trivially grow current extent, if next block is not taken */ + terminator = entry + extent_count - 1; + if (extent_count > 1) { + entry = terminator-1; + new_block = be64_to_cpu(entry->e_cluster) + + be64_to_cpu(entry->e_blocks); + + if (omfs_allocate_block(inode->i_sb, new_block)) { + entry->e_blocks = + cpu_to_be64(be64_to_cpu(entry->e_blocks) + 1); + terminator->e_blocks = ~(cpu_to_be64( + be64_to_cpu(~terminator->e_blocks) + 1)); + goto out; + } + } + max_count = (sbi->s_sys_blocksize - OMFS_EXTENT_START - + sizeof(struct omfs_extent)) / + sizeof(struct omfs_extent_entry) + 1; + + /* TODO: add a continuation block here */ + if (be32_to_cpu(oe->e_extent_count) > max_count-1) + return -EIO; + + /* try to allocate a new cluster */ + ret = omfs_allocate_range(inode->i_sb, 1, sbi->s_clustersize, + &new_block, &new_count); + if (ret) + goto out_fail; + + /* copy terminator down an entry */ + entry = terminator; + terminator++; + memcpy(terminator, entry, sizeof(struct omfs_extent_entry)); + + entry->e_cluster = cpu_to_be64(new_block); + entry->e_blocks = cpu_to_be64((u64) new_count); + + terminator->e_blocks = ~(cpu_to_be64( + be64_to_cpu(~terminator->e_blocks) + (u64) new_count)); + + /* write in new entry */ + oe->e_extent_count = cpu_to_be32(1 + be32_to_cpu(oe->e_extent_count)); + +out: + *ret_block = new_block; +out_fail: + return ret; +} + +/* + * Scans across the directory table for a given file block number. + * If block not found, return 0. + */ +static sector_t find_block(struct inode *inode, struct omfs_extent_entry *ent, + sector_t block, int count, int *left) +{ + /* count > 1 because of terminator */ + sector_t searched = 0; + for (; count > 1; count--) { + int numblocks = clus_to_blk(OMFS_SB(inode->i_sb), + be64_to_cpu(ent->e_blocks)); + + if (block >= searched && + block < searched + numblocks) { + /* + * found it at cluster + (block - searched) + * numblocks - (block - searched) is remainder + */ + *left = numblocks - (block - searched); + return clus_to_blk(OMFS_SB(inode->i_sb), + be64_to_cpu(ent->e_cluster)) + + block - searched; + } + searched += numblocks; + ent++; + } + return 0; +} + +static int omfs_get_block(struct inode *inode, sector_t block, + struct buffer_head *bh_result, int create) +{ + struct buffer_head *bh; + sector_t next, offset; + int ret; + u64 new_block; + int extent_count; + struct omfs_extent *oe; + struct omfs_extent_entry *entry; + struct omfs_sb_info *sbi = OMFS_SB(inode->i_sb); + int max_blocks = bh_result->b_size >> inode->i_blkbits; + int remain; + + ret = -EIO; + bh = sb_bread(inode->i_sb, clus_to_blk(sbi, inode->i_ino)); + if (!bh) + goto out; + + oe = (struct omfs_extent *)(&bh->b_data[OMFS_EXTENT_START]); + next = inode->i_ino; + + for (;;) { + + if (omfs_is_bad(sbi, (struct omfs_header *) bh->b_data, next)) + goto out_brelse; + + extent_count = be32_to_cpu(oe->e_extent_count); + next = be64_to_cpu(oe->e_next); + entry = &oe->e_entry; + + offset = find_block(inode, entry, block, extent_count, &remain); + if (offset > 0) { + ret = 0; + map_bh(bh_result, inode->i_sb, offset); + if (remain > max_blocks) + remain = max_blocks; + bh_result->b_size = (remain << inode->i_blkbits); + goto out_brelse; + } + if (next == ~0) + break; + + brelse(bh); + bh = sb_bread(inode->i_sb, clus_to_blk(sbi, next)); + if (!bh) + goto out; + oe = (struct omfs_extent *) (&bh->b_data[OMFS_EXTENT_CONT]); + } + if (create) { + ret = omfs_grow_extent(inode, oe, &new_block); + if (ret == 0) { + mark_buffer_dirty(bh); + mark_inode_dirty(inode); + map_bh(bh_result, inode->i_sb, + clus_to_blk(sbi, new_block)); + } + } +out_brelse: + brelse(bh); +out: + return ret; +} + +static int omfs_readpage(struct file *file, struct page *page) +{ + return block_read_full_page(page, omfs_get_block); +} + +static int omfs_readpages(struct file *file, struct address_space *mapping, + struct list_head *pages, unsigned nr_pages) +{ + return mpage_readpages(mapping, pages, nr_pages, omfs_get_block); +} + +static int omfs_writepage(struct page *page, struct writeback_control *wbc) +{ + return block_write_full_page(page, omfs_get_block, wbc); +} + +static int +omfs_writepages(struct address_space *mapping, struct writeback_control *wbc) +{ + return mpage_writepages(mapping, wbc, omfs_get_block); +} + +static int omfs_write_begin(struct file *file, struct address_space *mapping, + loff_t pos, unsigned len, unsigned flags, + struct page **pagep, void **fsdata) +{ + *pagep = NULL; + return block_write_begin(file, mapping, pos, len, flags, + pagep, fsdata, omfs_get_block); +} + +static sector_t omfs_bmap(struct address_space *mapping, sector_t block) +{ + return generic_block_bmap(mapping, block, omfs_get_block); +} + +struct file_operations omfs_file_operations = { + .llseek = generic_file_llseek, + .read = do_sync_read, + .write = do_sync_write, + .aio_read = generic_file_aio_read, + .aio_write = generic_file_aio_write, + .mmap = generic_file_mmap, + .fsync = omfs_sync_file, + .splice_read = generic_file_splice_read, +}; + +struct inode_operations omfs_file_inops = { + .truncate = omfs_truncate +}; + +struct address_space_operations omfs_aops = { + .readpage = omfs_readpage, + .readpages = omfs_readpages, + .writepage = omfs_writepage, + .writepages = omfs_writepages, + .sync_page = block_sync_page, + .write_begin = omfs_write_begin, + .write_end = generic_write_end, + .bmap = omfs_bmap, +}; + -- cgit v1.2.3-70-g09d2 From 36cc410a6799a205bfc6ccc38abd9d52f2afba64 Mon Sep 17 00:00:00 2001 From: Bob Copeland Date: Fri, 25 Jul 2008 19:45:17 -0700 Subject: omfs: add bitmap routines Add block allocation and block bitmap management routines for OMFS. Signed-off-by: Bob Copeland Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/omfs/bitmap.c | 192 +++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 192 insertions(+) create mode 100644 fs/omfs/bitmap.c (limited to 'fs') diff --git a/fs/omfs/bitmap.c b/fs/omfs/bitmap.c new file mode 100644 index 00000000000..dc75f22be3f --- /dev/null +++ b/fs/omfs/bitmap.c @@ -0,0 +1,192 @@ +#include +#include +#include +#include +#include "omfs.h" + +unsigned long omfs_count_free(struct super_block *sb) +{ + unsigned int i; + unsigned long sum = 0; + struct omfs_sb_info *sbi = OMFS_SB(sb); + int nbits = sb->s_blocksize * 8; + + for (i = 0; i < sbi->s_imap_size; i++) + sum += nbits - bitmap_weight(sbi->s_imap[i], nbits); + + return sum; +} + +/* + * Counts the run of zero bits starting at bit up to max. + * It handles the case where a run might spill over a buffer. + * Called with bitmap lock. + */ +static int count_run(unsigned long **addr, int nbits, + int addrlen, int bit, int max) +{ + int count = 0; + int x; + + for (; addrlen > 0; addrlen--, addr++) { + x = find_next_bit(*addr, nbits, bit); + count += x - bit; + + if (x < nbits || count > max) + return min(count, max); + + bit = 0; + } + return min(count, max); +} + +/* + * Sets or clears the run of count bits starting with bit. + * Called with bitmap lock. + */ +static int set_run(struct super_block *sb, int map, + int nbits, int bit, int count, int set) +{ + int i; + int err; + struct buffer_head *bh; + struct omfs_sb_info *sbi = OMFS_SB(sb); + + err = -ENOMEM; + bh = sb_bread(sb, clus_to_blk(sbi, sbi->s_bitmap_ino) + map); + if (!bh) + goto out; + + for (i = 0; i < count; i++, bit++) { + if (bit >= nbits) { + bit = 0; + map++; + + mark_buffer_dirty(bh); + brelse(bh); + bh = sb_bread(sb, + clus_to_blk(sbi, sbi->s_bitmap_ino) + map); + if (!bh) + goto out; + } + if (set) { + set_bit(bit, sbi->s_imap[map]); + set_bit(bit, (long *) bh->b_data); + } else { + clear_bit(bit, sbi->s_imap[map]); + clear_bit(bit, (long *) bh->b_data); + } + } + mark_buffer_dirty(bh); + brelse(bh); + err = 0; +out: + return err; +} + +/* + * Tries to allocate exactly one block. Returns true if sucessful. + */ +int omfs_allocate_block(struct super_block *sb, u64 block) +{ + struct buffer_head *bh; + struct omfs_sb_info *sbi = OMFS_SB(sb); + int bits_per_entry = 8 * sb->s_blocksize; + int map, bit; + int ret = 0; + u64 tmp; + + tmp = block; + bit = do_div(tmp, bits_per_entry); + map = tmp; + + mutex_lock(&sbi->s_bitmap_lock); + if (map >= sbi->s_imap_size || test_and_set_bit(bit, sbi->s_imap[map])) + goto out; + + if (sbi->s_bitmap_ino > 0) { + bh = sb_bread(sb, clus_to_blk(sbi, sbi->s_bitmap_ino) + map); + if (!bh) + goto out; + + set_bit(bit, (long *) bh->b_data); + mark_buffer_dirty(bh); + brelse(bh); + } + ret = 1; +out: + mutex_unlock(&sbi->s_bitmap_lock); + return ret; +} + + +/* + * Tries to allocate a set of blocks. The request size depends on the + * type: for inodes, we must allocate sbi->s_mirrors blocks, and for file + * blocks, we try to allocate sbi->s_clustersize, but can always get away + * with just one block. + */ +int omfs_allocate_range(struct super_block *sb, + int min_request, + int max_request, + u64 *return_block, + int *return_size) +{ + struct omfs_sb_info *sbi = OMFS_SB(sb); + int bits_per_entry = 8 * sb->s_blocksize; + int ret = 0; + int i, run, bit; + + mutex_lock(&sbi->s_bitmap_lock); + for (i = 0; i < sbi->s_imap_size; i++) { + bit = 0; + while (bit < bits_per_entry) { + bit = find_next_zero_bit(sbi->s_imap[i], bits_per_entry, + bit); + + if (bit == bits_per_entry) + break; + + run = count_run(&sbi->s_imap[i], bits_per_entry, + sbi->s_imap_size-i, bit, max_request); + + if (run >= min_request) + goto found; + bit += run; + } + } + ret = -ENOSPC; + goto out; + +found: + *return_block = i * bits_per_entry + bit; + *return_size = run; + ret = set_run(sb, i, bits_per_entry, bit, run, 1); + +out: + mutex_unlock(&sbi->s_bitmap_lock); + return ret; +} + +/* + * Clears count bits starting at a given block. + */ +int omfs_clear_range(struct super_block *sb, u64 block, int count) +{ + struct omfs_sb_info *sbi = OMFS_SB(sb); + int bits_per_entry = 8 * sb->s_blocksize; + u64 tmp; + int map, bit, ret; + + tmp = block; + bit = do_div(tmp, bits_per_entry); + map = tmp; + + if (map >= sbi->s_imap_size) + return 0; + + mutex_lock(&sbi->s_bitmap_lock); + ret = set_run(sb, map, bits_per_entry, bit, count, 0); + mutex_unlock(&sbi->s_bitmap_lock); + return ret; +} -- cgit v1.2.3-70-g09d2 From 63ca8ce2a2641f9cb5f0add33ced4591681d1cd7 Mon Sep 17 00:00:00 2001 From: Bob Copeland Date: Fri, 25 Jul 2008 19:45:17 -0700 Subject: omfs: update kbuild to include OMFS Adds OMFS to the fs Kconfig and Makefile Signed-off-by: Bob Copeland Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/Kconfig | 13 +++++++++++++ fs/Makefile | 1 + fs/omfs/Makefile | 4 ++++ 3 files changed, 18 insertions(+) create mode 100644 fs/omfs/Makefile (limited to 'fs') diff --git a/fs/Kconfig b/fs/Kconfig index 97e3bdedb1e..d3873583360 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -1383,6 +1383,19 @@ config MINIX_FS partition (the one containing the directory /) cannot be compiled as a module. +config OMFS_FS + tristate "SonicBlue Optimized MPEG File System support" + depends on BLOCK + select CRC_ITU_T + help + This is the proprietary file system used by the Rio Karma music + player and ReplayTV DVR. Despite the name, this filesystem is not + more efficient than a standard FS for MPEG files, in fact likely + the opposite is true. Say Y if you have either of these devices + and wish to mount its disk. + + To compile this file system support as a module, choose M here: the + module will be called omfs. If unsure, say N. config HPFS_FS tristate "OS/2 HPFS file system support" diff --git a/fs/Makefile b/fs/Makefile index 3b2178b4bb6..a1482a5eff1 100644 --- a/fs/Makefile +++ b/fs/Makefile @@ -111,6 +111,7 @@ obj-$(CONFIG_ADFS_FS) += adfs/ obj-$(CONFIG_FUSE_FS) += fuse/ obj-$(CONFIG_UDF_FS) += udf/ obj-$(CONFIG_SUN_OPENPROMFS) += openpromfs/ +obj-$(CONFIG_OMFS_FS) += omfs/ obj-$(CONFIG_JFS_FS) += jfs/ obj-$(CONFIG_XFS_FS) += xfs/ obj-$(CONFIG_9P_FS) += 9p/ diff --git a/fs/omfs/Makefile b/fs/omfs/Makefile new file mode 100644 index 00000000000..8b82b63f112 --- /dev/null +++ b/fs/omfs/Makefile @@ -0,0 +1,4 @@ + +obj-$(CONFIG_OMFS_FS) += omfs.o + +omfs-y := bitmap.o dir.o file.o inode.o -- cgit v1.2.3-70-g09d2 From f5dd33c494a427b1d1a3b574de5c9e511c888864 Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Fri, 25 Jul 2008 19:45:25 -0700 Subject: dio: use get_user_pages_fast Use get_user_pages_fast in the common/generic block and fs direct IO paths. Signed-off-by: Nick Piggin Cc: Dave Kleikamp Cc: Andy Whitcroft Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Andi Kleen Cc: Dave Kleikamp Cc: Badari Pulavarty Cc: Zach Brown Cc: Jens Axboe Reviewed-by: Peter Zijlstra Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/bio.c | 8 ++------ fs/direct-io.c | 10 ++-------- 2 files changed, 4 insertions(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/bio.c b/fs/bio.c index 88322b066ac..25f1af0d81e 100644 --- a/fs/bio.c +++ b/fs/bio.c @@ -721,12 +721,8 @@ static struct bio *__bio_map_user_iov(struct request_queue *q, const int local_nr_pages = end - start; const int page_limit = cur_page + local_nr_pages; - down_read(¤t->mm->mmap_sem); - ret = get_user_pages(current, current->mm, uaddr, - local_nr_pages, - write_to_vm, 0, &pages[cur_page], NULL); - up_read(¤t->mm->mmap_sem); - + ret = get_user_pages_fast(uaddr, local_nr_pages, + write_to_vm, &pages[cur_page]); if (ret < local_nr_pages) { ret = -EFAULT; goto out_unmap; diff --git a/fs/direct-io.c b/fs/direct-io.c index 9e81addbd6e..9606ee848fd 100644 --- a/fs/direct-io.c +++ b/fs/direct-io.c @@ -150,17 +150,11 @@ static int dio_refill_pages(struct dio *dio) int nr_pages; nr_pages = min(dio->total_pages - dio->curr_page, DIO_PAGES); - down_read(¤t->mm->mmap_sem); - ret = get_user_pages( - current, /* Task for fault acounting */ - current->mm, /* whose pages? */ + ret = get_user_pages_fast( dio->curr_user_address, /* Where from? */ nr_pages, /* How many pages? */ dio->rw == READ, /* Write to memory? */ - 0, /* force (?) */ - &dio->pages[0], - NULL); /* vmas */ - up_read(¤t->mm->mmap_sem); + &dio->pages[0]); /* Put results here */ if (ret < 0 && dio->blocks_available && (dio->rw & WRITE)) { struct page *page = ZERO_PAGE(0); -- cgit v1.2.3-70-g09d2 From bc40d73c950146725e9e768e856a416ec8949065 Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Fri, 25 Jul 2008 19:45:26 -0700 Subject: splice: use get_user_pages_fast Use get_user_pages_fast in splice. This reverts some mmap_sem batching there, however the biggest problem with mmap_sem tends to be hold times blocking out other threads rather than cacheline bouncing. Further: on architectures that implement get_user_pages_fast without locks, mmap_sem can be avoided completely anyway. Signed-off-by: Nick Piggin Cc: Dave Kleikamp Cc: Andy Whitcroft Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Andi Kleen Cc: Dave Kleikamp Cc: Badari Pulavarty Cc: Zach Brown Cc: Jens Axboe Reviewed-by: Peter Zijlstra Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/splice.c | 41 +++-------------------------------------- 1 file changed, 3 insertions(+), 38 deletions(-) (limited to 'fs') diff --git a/fs/splice.c b/fs/splice.c index 399442179d8..47dc1a445d1 100644 --- a/fs/splice.c +++ b/fs/splice.c @@ -1160,36 +1160,6 @@ static long do_splice(struct file *in, loff_t __user *off_in, return -EINVAL; } -/* - * Do a copy-from-user while holding the mmap_semaphore for reading, in a - * manner safe from deadlocking with simultaneous mmap() (grabbing mmap_sem - * for writing) and page faulting on the user memory pointed to by src. - * This assumes that we will very rarely hit the partial != 0 path, or this - * will not be a win. - */ -static int copy_from_user_mmap_sem(void *dst, const void __user *src, size_t n) -{ - int partial; - - if (!access_ok(VERIFY_READ, src, n)) - return -EFAULT; - - pagefault_disable(); - partial = __copy_from_user_inatomic(dst, src, n); - pagefault_enable(); - - /* - * Didn't copy everything, drop the mmap_sem and do a faulting copy - */ - if (unlikely(partial)) { - up_read(¤t->mm->mmap_sem); - partial = copy_from_user(dst, src, n); - down_read(¤t->mm->mmap_sem); - } - - return partial; -} - /* * Map an iov into an array of pages and offset/length tupples. With the * partial_page structure, we can map several non-contiguous ranges into @@ -1203,8 +1173,6 @@ static int get_iovec_page_array(const struct iovec __user *iov, { int buffers = 0, error = 0; - down_read(¤t->mm->mmap_sem); - while (nr_vecs) { unsigned long off, npages; struct iovec entry; @@ -1213,7 +1181,7 @@ static int get_iovec_page_array(const struct iovec __user *iov, int i; error = -EFAULT; - if (copy_from_user_mmap_sem(&entry, iov, sizeof(entry))) + if (copy_from_user(&entry, iov, sizeof(entry))) break; base = entry.iov_base; @@ -1247,9 +1215,8 @@ static int get_iovec_page_array(const struct iovec __user *iov, if (npages > PIPE_BUFFERS - buffers) npages = PIPE_BUFFERS - buffers; - error = get_user_pages(current, current->mm, - (unsigned long) base, npages, 0, 0, - &pages[buffers], NULL); + error = get_user_pages_fast((unsigned long)base, npages, + 0, &pages[buffers]); if (unlikely(error <= 0)) break; @@ -1288,8 +1255,6 @@ static int get_iovec_page_array(const struct iovec __user *iov, iov++; } - up_read(¤t->mm->mmap_sem); - if (buffers) return buffers; -- cgit v1.2.3-70-g09d2 From 19fd6231279be3c3bdd02ed99f9b0eb195978064 Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Fri, 25 Jul 2008 19:45:32 -0700 Subject: mm: spinlock tree_lock mapping->tree_lock has no read lockers. convert the lock from an rwlock to a spinlock. Signed-off-by: Nick Piggin Cc: Benjamin Herrenschmidt Cc: Paul Mackerras Cc: Hugh Dickins Cc: "Paul E. McKenney" Reviewed-by: Peter Zijlstra Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/buffer.c | 4 ++-- fs/inode.c | 2 +- include/asm-arm/cacheflush.h | 4 ++-- include/asm-parisc/cacheflush.h | 4 ++-- include/linux/fs.h | 2 +- mm/filemap.c | 10 +++++----- mm/migrate.c | 11 +++++------ mm/page-writeback.c | 12 ++++++------ mm/swap_state.c | 10 +++++----- mm/swapfile.c | 4 ++-- mm/truncate.c | 6 +++--- mm/vmscan.c | 8 ++++---- 12 files changed, 38 insertions(+), 39 deletions(-) (limited to 'fs') diff --git a/fs/buffer.c b/fs/buffer.c index d48caee12e2..109b261192d 100644 --- a/fs/buffer.c +++ b/fs/buffer.c @@ -706,7 +706,7 @@ static int __set_page_dirty(struct page *page, if (TestSetPageDirty(page)) return 0; - write_lock_irq(&mapping->tree_lock); + spin_lock_irq(&mapping->tree_lock); if (page->mapping) { /* Race with truncate? */ WARN_ON_ONCE(warn && !PageUptodate(page)); @@ -719,7 +719,7 @@ static int __set_page_dirty(struct page *page, radix_tree_tag_set(&mapping->page_tree, page_index(page), PAGECACHE_TAG_DIRTY); } - write_unlock_irq(&mapping->tree_lock); + spin_unlock_irq(&mapping->tree_lock); __mark_inode_dirty(mapping->host, I_DIRTY_PAGES); return 1; diff --git a/fs/inode.c b/fs/inode.c index c36d9480335..35b6414522e 100644 --- a/fs/inode.c +++ b/fs/inode.c @@ -209,7 +209,7 @@ void inode_init_once(struct inode *inode) INIT_LIST_HEAD(&inode->i_dentry); INIT_LIST_HEAD(&inode->i_devices); INIT_RADIX_TREE(&inode->i_data.page_tree, GFP_ATOMIC); - rwlock_init(&inode->i_data.tree_lock); + spin_lock_init(&inode->i_data.tree_lock); spin_lock_init(&inode->i_data.i_mmap_lock); INIT_LIST_HEAD(&inode->i_data.private_list); spin_lock_init(&inode->i_data.private_lock); diff --git a/include/asm-arm/cacheflush.h b/include/asm-arm/cacheflush.h index 70b0fe724b6..03cf1ee977b 100644 --- a/include/asm-arm/cacheflush.h +++ b/include/asm-arm/cacheflush.h @@ -424,9 +424,9 @@ static inline void flush_anon_page(struct vm_area_struct *vma, } #define flush_dcache_mmap_lock(mapping) \ - write_lock_irq(&(mapping)->tree_lock) + spin_lock_irq(&(mapping)->tree_lock) #define flush_dcache_mmap_unlock(mapping) \ - write_unlock_irq(&(mapping)->tree_lock) + spin_unlock_irq(&(mapping)->tree_lock) #define flush_icache_user_range(vma,page,addr,len) \ flush_dcache_page(page) diff --git a/include/asm-parisc/cacheflush.h b/include/asm-parisc/cacheflush.h index 2f1e1b05440..b7ca6dc7fdd 100644 --- a/include/asm-parisc/cacheflush.h +++ b/include/asm-parisc/cacheflush.h @@ -45,9 +45,9 @@ void flush_cache_mm(struct mm_struct *mm); extern void flush_dcache_page(struct page *page); #define flush_dcache_mmap_lock(mapping) \ - write_lock_irq(&(mapping)->tree_lock) + spin_lock_irq(&(mapping)->tree_lock) #define flush_dcache_mmap_unlock(mapping) \ - write_unlock_irq(&(mapping)->tree_lock) + spin_unlock_irq(&(mapping)->tree_lock) #define flush_icache_page(vma,page) do { \ flush_kernel_dcache_page(page); \ diff --git a/include/linux/fs.h b/include/linux/fs.h index 49d8eb7a71b..53d2edb709b 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -499,7 +499,7 @@ struct backing_dev_info; struct address_space { struct inode *host; /* owner: inode, block_device */ struct radix_tree_root page_tree; /* radix tree of all pages */ - rwlock_t tree_lock; /* and rwlock protecting it */ + spinlock_t tree_lock; /* and lock protecting it */ unsigned int i_mmap_writable;/* count VM_SHARED mappings */ struct prio_tree_root i_mmap; /* tree of private and shared mappings */ struct list_head i_mmap_nonlinear;/*list VM_NONLINEAR mappings */ diff --git a/mm/filemap.c b/mm/filemap.c index feb8448d861..2ed8b0389c5 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -109,7 +109,7 @@ /* * Remove a page from the page cache and free it. Caller has to make * sure the page is locked and that nobody else uses it - or that usage - * is safe. The caller must hold a write_lock on the mapping's tree_lock. + * is safe. The caller must hold the mapping's tree_lock. */ void __remove_from_page_cache(struct page *page) { @@ -141,9 +141,9 @@ void remove_from_page_cache(struct page *page) BUG_ON(!PageLocked(page)); - write_lock_irq(&mapping->tree_lock); + spin_lock_irq(&mapping->tree_lock); __remove_from_page_cache(page); - write_unlock_irq(&mapping->tree_lock); + spin_unlock_irq(&mapping->tree_lock); } static int sync_page(void *word) @@ -469,7 +469,7 @@ int add_to_page_cache_locked(struct page *page, struct address_space *mapping, page->mapping = mapping; page->index = offset; - write_lock_irq(&mapping->tree_lock); + spin_lock_irq(&mapping->tree_lock); error = radix_tree_insert(&mapping->page_tree, offset, page); if (likely(!error)) { mapping->nrpages++; @@ -480,7 +480,7 @@ int add_to_page_cache_locked(struct page *page, struct address_space *mapping, page_cache_release(page); } - write_unlock_irq(&mapping->tree_lock); + spin_unlock_irq(&mapping->tree_lock); radix_tree_preload_end(); } else mem_cgroup_uncharge_cache_page(page); diff --git a/mm/migrate.c b/mm/migrate.c index 3ca6392e82c..153572fb60b 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -323,7 +323,7 @@ static int migrate_page_move_mapping(struct address_space *mapping, return 0; } - write_lock_irq(&mapping->tree_lock); + spin_lock_irq(&mapping->tree_lock); pslot = radix_tree_lookup_slot(&mapping->page_tree, page_index(page)); @@ -331,12 +331,12 @@ static int migrate_page_move_mapping(struct address_space *mapping, expected_count = 2 + !!PagePrivate(page); if (page_count(page) != expected_count || (struct page *)radix_tree_deref_slot(pslot) != page) { - write_unlock_irq(&mapping->tree_lock); + spin_unlock_irq(&mapping->tree_lock); return -EAGAIN; } if (!page_freeze_refs(page, expected_count)) { - write_unlock_irq(&mapping->tree_lock); + spin_unlock_irq(&mapping->tree_lock); return -EAGAIN; } @@ -373,10 +373,9 @@ static int migrate_page_move_mapping(struct address_space *mapping, __dec_zone_page_state(page, NR_FILE_PAGES); __inc_zone_page_state(newpage, NR_FILE_PAGES); - write_unlock_irq(&mapping->tree_lock); - if (!PageSwapCache(newpage)) { + spin_unlock_irq(&mapping->tree_lock); + if (!PageSwapCache(newpage)) mem_cgroup_uncharge_cache_page(page); - } return 0; } diff --git a/mm/page-writeback.c b/mm/page-writeback.c index 94c6d8988ab..24de8b65fdb 100644 --- a/mm/page-writeback.c +++ b/mm/page-writeback.c @@ -1088,7 +1088,7 @@ int __set_page_dirty_nobuffers(struct page *page) if (!mapping) return 1; - write_lock_irq(&mapping->tree_lock); + spin_lock_irq(&mapping->tree_lock); mapping2 = page_mapping(page); if (mapping2) { /* Race with truncate? */ BUG_ON(mapping2 != mapping); @@ -1102,7 +1102,7 @@ int __set_page_dirty_nobuffers(struct page *page) radix_tree_tag_set(&mapping->page_tree, page_index(page), PAGECACHE_TAG_DIRTY); } - write_unlock_irq(&mapping->tree_lock); + spin_unlock_irq(&mapping->tree_lock); if (mapping->host) { /* !PageAnon && !swapper_space */ __mark_inode_dirty(mapping->host, I_DIRTY_PAGES); @@ -1258,7 +1258,7 @@ int test_clear_page_writeback(struct page *page) struct backing_dev_info *bdi = mapping->backing_dev_info; unsigned long flags; - write_lock_irqsave(&mapping->tree_lock, flags); + spin_lock_irqsave(&mapping->tree_lock, flags); ret = TestClearPageWriteback(page); if (ret) { radix_tree_tag_clear(&mapping->page_tree, @@ -1269,7 +1269,7 @@ int test_clear_page_writeback(struct page *page) __bdi_writeout_inc(bdi); } } - write_unlock_irqrestore(&mapping->tree_lock, flags); + spin_unlock_irqrestore(&mapping->tree_lock, flags); } else { ret = TestClearPageWriteback(page); } @@ -1287,7 +1287,7 @@ int test_set_page_writeback(struct page *page) struct backing_dev_info *bdi = mapping->backing_dev_info; unsigned long flags; - write_lock_irqsave(&mapping->tree_lock, flags); + spin_lock_irqsave(&mapping->tree_lock, flags); ret = TestSetPageWriteback(page); if (!ret) { radix_tree_tag_set(&mapping->page_tree, @@ -1300,7 +1300,7 @@ int test_set_page_writeback(struct page *page) radix_tree_tag_clear(&mapping->page_tree, page_index(page), PAGECACHE_TAG_DIRTY); - write_unlock_irqrestore(&mapping->tree_lock, flags); + spin_unlock_irqrestore(&mapping->tree_lock, flags); } else { ret = TestSetPageWriteback(page); } diff --git a/mm/swap_state.c b/mm/swap_state.c index 3e3381d6c7e..2c217e33d49 100644 --- a/mm/swap_state.c +++ b/mm/swap_state.c @@ -39,7 +39,7 @@ static struct backing_dev_info swap_backing_dev_info = { struct address_space swapper_space = { .page_tree = RADIX_TREE_INIT(GFP_ATOMIC|__GFP_NOWARN), - .tree_lock = __RW_LOCK_UNLOCKED(swapper_space.tree_lock), + .tree_lock = __SPIN_LOCK_UNLOCKED(swapper_space.tree_lock), .a_ops = &swap_aops, .i_mmap_nonlinear = LIST_HEAD_INIT(swapper_space.i_mmap_nonlinear), .backing_dev_info = &swap_backing_dev_info, @@ -80,7 +80,7 @@ int add_to_swap_cache(struct page *page, swp_entry_t entry, gfp_t gfp_mask) SetPageSwapCache(page); set_page_private(page, entry.val); - write_lock_irq(&swapper_space.tree_lock); + spin_lock_irq(&swapper_space.tree_lock); error = radix_tree_insert(&swapper_space.page_tree, entry.val, page); if (likely(!error)) { @@ -88,7 +88,7 @@ int add_to_swap_cache(struct page *page, swp_entry_t entry, gfp_t gfp_mask) __inc_zone_page_state(page, NR_FILE_PAGES); INC_CACHE_INFO(add_total); } - write_unlock_irq(&swapper_space.tree_lock); + spin_unlock_irq(&swapper_space.tree_lock); radix_tree_preload_end(); if (unlikely(error)) { @@ -182,9 +182,9 @@ void delete_from_swap_cache(struct page *page) entry.val = page_private(page); - write_lock_irq(&swapper_space.tree_lock); + spin_lock_irq(&swapper_space.tree_lock); __delete_from_swap_cache(page); - write_unlock_irq(&swapper_space.tree_lock); + spin_unlock_irq(&swapper_space.tree_lock); swap_free(entry); page_cache_release(page); diff --git a/mm/swapfile.c b/mm/swapfile.c index 2f33edb8bee..af283933c14 100644 --- a/mm/swapfile.c +++ b/mm/swapfile.c @@ -369,13 +369,13 @@ int remove_exclusive_swap_page(struct page *page) retval = 0; if (p->swap_map[swp_offset(entry)] == 1) { /* Recheck the page count with the swapcache lock held.. */ - write_lock_irq(&swapper_space.tree_lock); + spin_lock_irq(&swapper_space.tree_lock); if ((page_count(page) == 2) && !PageWriteback(page)) { __delete_from_swap_cache(page); SetPageDirty(page); retval = 1; } - write_unlock_irq(&swapper_space.tree_lock); + spin_unlock_irq(&swapper_space.tree_lock); } spin_unlock(&swap_lock); diff --git a/mm/truncate.c b/mm/truncate.c index b8961cb6341..e68443d7456 100644 --- a/mm/truncate.c +++ b/mm/truncate.c @@ -349,18 +349,18 @@ invalidate_complete_page2(struct address_space *mapping, struct page *page) if (PagePrivate(page) && !try_to_release_page(page, GFP_KERNEL)) return 0; - write_lock_irq(&mapping->tree_lock); + spin_lock_irq(&mapping->tree_lock); if (PageDirty(page)) goto failed; BUG_ON(PagePrivate(page)); __remove_from_page_cache(page); - write_unlock_irq(&mapping->tree_lock); + spin_unlock_irq(&mapping->tree_lock); ClearPageUptodate(page); page_cache_release(page); /* pagecache ref */ return 1; failed: - write_unlock_irq(&mapping->tree_lock); + spin_unlock_irq(&mapping->tree_lock); return 0; } diff --git a/mm/vmscan.c b/mm/vmscan.c index 0075eac1cd0..8f71761bc4b 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -399,7 +399,7 @@ static int __remove_mapping(struct address_space *mapping, struct page *page) BUG_ON(!PageLocked(page)); BUG_ON(mapping != page_mapping(page)); - write_lock_irq(&mapping->tree_lock); + spin_lock_irq(&mapping->tree_lock); /* * The non racy check for a busy page. * @@ -436,17 +436,17 @@ static int __remove_mapping(struct address_space *mapping, struct page *page) if (PageSwapCache(page)) { swp_entry_t swap = { .val = page_private(page) }; __delete_from_swap_cache(page); - write_unlock_irq(&mapping->tree_lock); + spin_unlock_irq(&mapping->tree_lock); swap_free(swap); } else { __remove_from_page_cache(page); - write_unlock_irq(&mapping->tree_lock); + spin_unlock_irq(&mapping->tree_lock); } return 1; cannot_free: - write_unlock_irq(&mapping->tree_lock); + spin_unlock_irq(&mapping->tree_lock); return 0; } -- cgit v1.2.3-70-g09d2 From 51cc50685a4275c6a02653670af9f108a64e01cf Mon Sep 17 00:00:00 2001 From: Alexey Dobriyan Date: Fri, 25 Jul 2008 19:45:34 -0700 Subject: SL*B: drop kmem cache argument from constructor Kmem cache passed to constructor is only needed for constructors that are themselves multiplexeres. Nobody uses this "feature", nor does anybody uses passed kmem cache in non-trivial way, so pass only pointer to object. Non-trivial places are: arch/powerpc/mm/init_64.c arch/powerpc/mm/hugetlbpage.c This is flag day, yes. Signed-off-by: Alexey Dobriyan Acked-by: Pekka Enberg Acked-by: Christoph Lameter Cc: Jon Tollefson Cc: Nick Piggin Cc: Matt Mackall [akpm@linux-foundation.org: fix arch/powerpc/mm/hugetlbpage.c] [akpm@linux-foundation.org: fix mm/slab.c] [akpm@linux-foundation.org: fix ubifs] Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/arm/plat-s3c24xx/dma.c | 2 +- arch/powerpc/kernel/rtas_flash.c | 2 +- arch/powerpc/mm/hugetlbpage.c | 9 ++------- arch/powerpc/mm/init_64.c | 24 +++++++++--------------- arch/powerpc/platforms/cell/spufs/inode.c | 2 +- arch/sh/mm/pmb.c | 2 +- arch/xtensa/mm/init.c | 2 +- drivers/usb/mon/mon_text.c | 4 ++-- fs/adfs/super.c | 2 +- fs/affs/super.c | 2 +- fs/afs/super.c | 4 ++-- fs/befs/linuxvfs.c | 2 +- fs/bfs/inode.c | 2 +- fs/block_dev.c | 2 +- fs/buffer.c | 2 +- fs/cifs/cifsfs.c | 2 +- fs/coda/inode.c | 2 +- fs/ecryptfs/main.c | 4 ++-- fs/efs/super.c | 2 +- fs/ext2/super.c | 2 +- fs/ext3/super.c | 2 +- fs/ext4/super.c | 2 +- fs/fat/cache.c | 2 +- fs/fat/inode.c | 2 +- fs/fuse/inode.c | 2 +- fs/gfs2/main.c | 4 ++-- fs/hfs/super.c | 2 +- fs/hfsplus/super.c | 2 +- fs/hpfs/super.c | 2 +- fs/hugetlbfs/inode.c | 2 +- fs/inode.c | 2 +- fs/isofs/inode.c | 2 +- fs/jffs2/super.c | 2 +- fs/jfs/jfs_metapage.c | 2 +- fs/jfs/super.c | 2 +- fs/locks.c | 2 +- fs/minix/inode.c | 2 +- fs/ncpfs/inode.c | 2 +- fs/nfs/inode.c | 2 +- fs/ntfs/super.c | 2 +- fs/ocfs2/dlm/dlmfs.c | 3 +-- fs/ocfs2/super.c | 2 +- fs/openpromfs/inode.c | 2 +- fs/proc/inode.c | 2 +- fs/qnx4/inode.c | 2 +- fs/reiserfs/super.c | 2 +- fs/romfs/inode.c | 2 +- fs/smbfs/inode.c | 2 +- fs/sysv/inode.c | 2 +- fs/ubifs/super.c | 2 +- fs/udf/super.c | 2 +- fs/ufs/super.c | 2 +- fs/xfs/linux-2.6/kmem.h | 2 +- fs/xfs/linux-2.6/xfs_super.c | 1 - include/linux/slab.h | 2 +- include/linux/slub_def.h | 2 +- ipc/mqueue.c | 2 +- kernel/fork.c | 2 +- lib/idr.c | 2 +- lib/radix-tree.c | 2 +- mm/rmap.c | 2 +- mm/shmem.c | 2 +- mm/slab.c | 11 +++++------ mm/slob.c | 7 +++---- mm/slub.c | 13 ++++++------- net/socket.c | 2 +- net/sunrpc/rpc_pipe.c | 2 +- 67 files changed, 90 insertions(+), 106 deletions(-) (limited to 'fs') diff --git a/arch/arm/plat-s3c24xx/dma.c b/arch/arm/plat-s3c24xx/dma.c index 60f162dc4fa..8c5e656d5d8 100644 --- a/arch/arm/plat-s3c24xx/dma.c +++ b/arch/arm/plat-s3c24xx/dma.c @@ -1304,7 +1304,7 @@ struct sysdev_class dma_sysclass = { /* kmem cache implementation */ -static void s3c2410_dma_cache_ctor(struct kmem_cache *c, void *p) +static void s3c2410_dma_cache_ctor(void *p) { memset(p, 0, sizeof(struct s3c2410_dma_buf)); } diff --git a/arch/powerpc/kernel/rtas_flash.c b/arch/powerpc/kernel/rtas_flash.c index 09ded5c424a..149cb112cd1 100644 --- a/arch/powerpc/kernel/rtas_flash.c +++ b/arch/powerpc/kernel/rtas_flash.c @@ -286,7 +286,7 @@ static ssize_t rtas_flash_read(struct file *file, char __user *buf, } /* constructor for flash_block_cache */ -void rtas_block_ctor(struct kmem_cache *cache, void *ptr) +void rtas_block_ctor(void *ptr) { memset(ptr, 0, RTAS_BLK_SIZE); } diff --git a/arch/powerpc/mm/hugetlbpage.c b/arch/powerpc/mm/hugetlbpage.c index fb42c4dd321..ed0aab0208a 100644 --- a/arch/powerpc/mm/hugetlbpage.c +++ b/arch/powerpc/mm/hugetlbpage.c @@ -113,7 +113,7 @@ static inline pte_t *hugepte_offset(hugepd_t *hpdp, unsigned long addr, static int __hugepte_alloc(struct mm_struct *mm, hugepd_t *hpdp, unsigned long address, unsigned int psize) { - pte_t *new = kmem_cache_alloc(huge_pgtable_cache(psize), + pte_t *new = kmem_cache_zalloc(huge_pgtable_cache(psize), GFP_KERNEL|__GFP_REPEAT); if (! new) @@ -730,11 +730,6 @@ static int __init hugepage_setup_sz(char *str) } __setup("hugepagesz=", hugepage_setup_sz); -static void zero_ctor(struct kmem_cache *cache, void *addr) -{ - memset(addr, 0, kmem_cache_size(cache)); -} - static int __init hugetlbpage_init(void) { unsigned int psize; @@ -756,7 +751,7 @@ static int __init hugetlbpage_init(void) HUGEPTE_TABLE_SIZE(psize), HUGEPTE_TABLE_SIZE(psize), 0, - zero_ctor); + NULL); if (!huge_pgtable_cache(psize)) panic("hugetlbpage_init(): could not create %s"\ "\n", HUGEPTE_CACHE_NAME(psize)); diff --git a/arch/powerpc/mm/init_64.c b/arch/powerpc/mm/init_64.c index a41bc5aa204..4f7df85129d 100644 --- a/arch/powerpc/mm/init_64.c +++ b/arch/powerpc/mm/init_64.c @@ -136,9 +136,14 @@ static int __init setup_kcore(void) module_init(setup_kcore); #endif -static void zero_ctor(struct kmem_cache *cache, void *addr) +static void pgd_ctor(void *addr) { - memset(addr, 0, kmem_cache_size(cache)); + memset(addr, 0, PGD_TABLE_SIZE); +} + +static void pmd_ctor(void *addr) +{ + memset(addr, 0, PMD_TABLE_SIZE); } static const unsigned int pgtable_cache_size[2] = { @@ -163,19 +168,8 @@ struct kmem_cache *pgtable_cache[ARRAY_SIZE(pgtable_cache_size)]; void pgtable_cache_init(void) { - int i; - - for (i = 0; i < ARRAY_SIZE(pgtable_cache_size); i++) { - int size = pgtable_cache_size[i]; - const char *name = pgtable_cache_name[i]; - - pr_debug("Allocating page table cache %s (#%d) " - "for size: %08x...\n", name, i, size); - pgtable_cache[i] = kmem_cache_create(name, - size, size, - SLAB_PANIC, - zero_ctor); - } + pgtable_cache[0] = kmem_cache_create(pgtable_cache_name[0], PGD_TABLE_SIZE, PGD_TABLE_SIZE, SLAB_PANIC, pgd_ctor); + pgtable_cache[1] = kmem_cache_create(pgtable_cache_name[1], PMD_TABLE_SIZE, PMD_TABLE_SIZE, SLAB_PANIC, pmd_ctor); } #ifdef CONFIG_SPARSEMEM_VMEMMAP diff --git a/arch/powerpc/platforms/cell/spufs/inode.c b/arch/powerpc/platforms/cell/spufs/inode.c index 7123472801d..690ca7b0dcf 100644 --- a/arch/powerpc/platforms/cell/spufs/inode.c +++ b/arch/powerpc/platforms/cell/spufs/inode.c @@ -78,7 +78,7 @@ spufs_destroy_inode(struct inode *inode) } static void -spufs_init_once(struct kmem_cache *cachep, void *p) +spufs_init_once(void *p) { struct spufs_inode_info *ei = p; diff --git a/arch/sh/mm/pmb.c b/arch/sh/mm/pmb.c index 0b0ec6e0475..46911bcbf17 100644 --- a/arch/sh/mm/pmb.c +++ b/arch/sh/mm/pmb.c @@ -293,7 +293,7 @@ void pmb_unmap(unsigned long addr) } while (pmbe); } -static void pmb_cache_ctor(struct kmem_cache *cachep, void *pmb) +static void pmb_cache_ctor(void *pmb) { struct pmb_entry *pmbe = pmb; diff --git a/arch/xtensa/mm/init.c b/arch/xtensa/mm/init.c index 81d0560eaea..ee261005b36 100644 --- a/arch/xtensa/mm/init.c +++ b/arch/xtensa/mm/init.c @@ -309,7 +309,7 @@ void show_mem(void) struct kmem_cache *pgtable_cache __read_mostly; -static void pgd_ctor(struct kmem_cache *cache, void* addr) +static void pgd_ctor(void* addr) { pte_t* ptep = (pte_t*)addr; int i; diff --git a/drivers/usb/mon/mon_text.c b/drivers/usb/mon/mon_text.c index 5e3e4e9b6c7..1f715436d6d 100644 --- a/drivers/usb/mon/mon_text.c +++ b/drivers/usb/mon/mon_text.c @@ -87,7 +87,7 @@ struct mon_reader_text { static struct dentry *mon_dir; /* Usually /sys/kernel/debug/usbmon */ -static void mon_text_ctor(struct kmem_cache *, void *); +static void mon_text_ctor(void *); struct mon_text_ptr { int cnt, limit; @@ -720,7 +720,7 @@ void mon_text_del(struct mon_bus *mbus) /* * Slab interface: constructor. */ -static void mon_text_ctor(struct kmem_cache *slab, void *mem) +static void mon_text_ctor(void *mem) { /* * Nothing to initialize. No, really! diff --git a/fs/adfs/super.c b/fs/adfs/super.c index 9e421eeb672..26f3b43726b 100644 --- a/fs/adfs/super.c +++ b/fs/adfs/super.c @@ -249,7 +249,7 @@ static void adfs_destroy_inode(struct inode *inode) kmem_cache_free(adfs_inode_cachep, ADFS_I(inode)); } -static void init_once(struct kmem_cache *cachep, void *foo) +static void init_once(void *foo) { struct adfs_inode_info *ei = (struct adfs_inode_info *) foo; diff --git a/fs/affs/super.c b/fs/affs/super.c index 4e030956640..3a89094f93d 100644 --- a/fs/affs/super.c +++ b/fs/affs/super.c @@ -90,7 +90,7 @@ static void affs_destroy_inode(struct inode *inode) kmem_cache_free(affs_inode_cachep, AFFS_I(inode)); } -static void init_once(struct kmem_cache *cachep, void *foo) +static void init_once(void *foo) { struct affs_inode_info *ei = (struct affs_inode_info *) foo; diff --git a/fs/afs/super.c b/fs/afs/super.c index 7e3faeef681..250d8c4d66e 100644 --- a/fs/afs/super.c +++ b/fs/afs/super.c @@ -27,7 +27,7 @@ #define AFS_FS_MAGIC 0x6B414653 /* 'kAFS' */ -static void afs_i_init_once(struct kmem_cache *cachep, void *foo); +static void afs_i_init_once(void *foo); static int afs_get_sb(struct file_system_type *fs_type, int flags, const char *dev_name, void *data, struct vfsmount *mnt); @@ -449,7 +449,7 @@ static void afs_put_super(struct super_block *sb) /* * initialise an inode cache slab element prior to any use */ -static void afs_i_init_once(struct kmem_cache *cachep, void *_vnode) +static void afs_i_init_once(void *_vnode) { struct afs_vnode *vnode = _vnode; diff --git a/fs/befs/linuxvfs.c b/fs/befs/linuxvfs.c index e8717de3bab..02c6e62b72f 100644 --- a/fs/befs/linuxvfs.c +++ b/fs/befs/linuxvfs.c @@ -289,7 +289,7 @@ befs_destroy_inode(struct inode *inode) kmem_cache_free(befs_inode_cachep, BEFS_I(inode)); } -static void init_once(struct kmem_cache *cachep, void *foo) +static void init_once(void *foo) { struct befs_inode_info *bi = (struct befs_inode_info *) foo; diff --git a/fs/bfs/inode.c b/fs/bfs/inode.c index 053e690ec9e..0ed57b5ee01 100644 --- a/fs/bfs/inode.c +++ b/fs/bfs/inode.c @@ -264,7 +264,7 @@ static void bfs_destroy_inode(struct inode *inode) kmem_cache_free(bfs_inode_cachep, BFS_I(inode)); } -static void init_once(struct kmem_cache *cachep, void *foo) +static void init_once(void *foo) { struct bfs_inode_info *bi = foo; diff --git a/fs/block_dev.c b/fs/block_dev.c index 10d8a0aa871..dcf37cada36 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -271,7 +271,7 @@ static void bdev_destroy_inode(struct inode *inode) kmem_cache_free(bdev_cachep, bdi); } -static void init_once(struct kmem_cache * cachep, void *foo) +static void init_once(void *foo) { struct bdev_inode *ei = (struct bdev_inode *) foo; struct block_device *bdev = &ei->bdev; diff --git a/fs/buffer.c b/fs/buffer.c index 109b261192d..5fd497cdd6f 100644 --- a/fs/buffer.c +++ b/fs/buffer.c @@ -3272,7 +3272,7 @@ int bh_submit_read(struct buffer_head *bh) EXPORT_SYMBOL(bh_submit_read); static void -init_buffer_head(struct kmem_cache *cachep, void *data) +init_buffer_head(void *data) { struct buffer_head *bh = data; diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c index 22857c639df..fe5f6809cba 100644 --- a/fs/cifs/cifsfs.c +++ b/fs/cifs/cifsfs.c @@ -766,7 +766,7 @@ const struct file_operations cifs_dir_ops = { }; static void -cifs_init_once(struct kmem_cache *cachep, void *inode) +cifs_init_once(void *inode) { struct cifsInodeInfo *cifsi = inode; diff --git a/fs/coda/inode.c b/fs/coda/inode.c index 2f58dfc7008..830f51abb97 100644 --- a/fs/coda/inode.c +++ b/fs/coda/inode.c @@ -58,7 +58,7 @@ static void coda_destroy_inode(struct inode *inode) kmem_cache_free(coda_inode_cachep, ITOC(inode)); } -static void init_once(struct kmem_cache * cachep, void *foo) +static void init_once(void *foo) { struct coda_inode_info *ei = (struct coda_inode_info *) foo; diff --git a/fs/ecryptfs/main.c b/fs/ecryptfs/main.c index 6f403cfba14..448dfd597b5 100644 --- a/fs/ecryptfs/main.c +++ b/fs/ecryptfs/main.c @@ -578,7 +578,7 @@ static struct file_system_type ecryptfs_fs_type = { * Initializes the ecryptfs_inode_info_cache when it is created */ static void -inode_info_init_once(struct kmem_cache *cachep, void *vptr) +inode_info_init_once(void *vptr) { struct ecryptfs_inode_info *ei = (struct ecryptfs_inode_info *)vptr; @@ -589,7 +589,7 @@ static struct ecryptfs_cache_info { struct kmem_cache **cache; const char *name; size_t size; - void (*ctor)(struct kmem_cache *cache, void *obj); + void (*ctor)(void *obj); } ecryptfs_cache_infos[] = { { .cache = &ecryptfs_auth_tok_list_item_cache, diff --git a/fs/efs/super.c b/fs/efs/super.c index d733531b55e..567b134fa1f 100644 --- a/fs/efs/super.c +++ b/fs/efs/super.c @@ -70,7 +70,7 @@ static void efs_destroy_inode(struct inode *inode) kmem_cache_free(efs_inode_cachep, INODE_INFO(inode)); } -static void init_once(struct kmem_cache *cachep, void *foo) +static void init_once(void *foo) { struct efs_inode_info *ei = (struct efs_inode_info *) foo; diff --git a/fs/ext2/super.c b/fs/ext2/super.c index 31308a3b0b8..fd88c7b43e6 100644 --- a/fs/ext2/super.c +++ b/fs/ext2/super.c @@ -159,7 +159,7 @@ static void ext2_destroy_inode(struct inode *inode) kmem_cache_free(ext2_inode_cachep, EXT2_I(inode)); } -static void init_once(struct kmem_cache * cachep, void *foo) +static void init_once(void *foo) { struct ext2_inode_info *ei = (struct ext2_inode_info *) foo; diff --git a/fs/ext3/super.c b/fs/ext3/super.c index 615788c6843..8ddced38467 100644 --- a/fs/ext3/super.c +++ b/fs/ext3/super.c @@ -472,7 +472,7 @@ static void ext3_destroy_inode(struct inode *inode) kmem_cache_free(ext3_inode_cachep, EXT3_I(inode)); } -static void init_once(struct kmem_cache * cachep, void *foo) +static void init_once(void *foo) { struct ext3_inode_info *ei = (struct ext3_inode_info *) foo; diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 1cb371dcd60..b5479b1dff1 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -595,7 +595,7 @@ static void ext4_destroy_inode(struct inode *inode) kmem_cache_free(ext4_inode_cachep, EXT4_I(inode)); } -static void init_once(struct kmem_cache *cachep, void *foo) +static void init_once(void *foo) { struct ext4_inode_info *ei = (struct ext4_inode_info *) foo; diff --git a/fs/fat/cache.c b/fs/fat/cache.c index 3a9ecac8d61..3222f51c41c 100644 --- a/fs/fat/cache.c +++ b/fs/fat/cache.c @@ -36,7 +36,7 @@ static inline int fat_max_cache(struct inode *inode) static struct kmem_cache *fat_cache_cachep; -static void init_once(struct kmem_cache *cachep, void *foo) +static void init_once(void *foo) { struct fat_cache *cache = (struct fat_cache *)foo; diff --git a/fs/fat/inode.c b/fs/fat/inode.c index 23676f9d79c..6d266d793e2 100644 --- a/fs/fat/inode.c +++ b/fs/fat/inode.c @@ -498,7 +498,7 @@ static void fat_destroy_inode(struct inode *inode) kmem_cache_free(fat_inode_cachep, MSDOS_I(inode)); } -static void init_once(struct kmem_cache *cachep, void *foo) +static void init_once(void *foo) { struct msdos_inode_info *ei = (struct msdos_inode_info *)foo; diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c index 7d2f7d6e22e..d2249f174e2 100644 --- a/fs/fuse/inode.c +++ b/fs/fuse/inode.c @@ -956,7 +956,7 @@ static inline void unregister_fuseblk(void) } #endif -static void fuse_inode_init_once(struct kmem_cache *cachep, void *foo) +static void fuse_inode_init_once(void *foo) { struct inode * inode = foo; diff --git a/fs/gfs2/main.c b/fs/gfs2/main.c index bcc668d0fad..bb2cc303ac2 100644 --- a/fs/gfs2/main.c +++ b/fs/gfs2/main.c @@ -24,7 +24,7 @@ #include "util.h" #include "glock.h" -static void gfs2_init_inode_once(struct kmem_cache *cachep, void *foo) +static void gfs2_init_inode_once(void *foo) { struct gfs2_inode *ip = foo; @@ -33,7 +33,7 @@ static void gfs2_init_inode_once(struct kmem_cache *cachep, void *foo) ip->i_alloc = NULL; } -static void gfs2_init_glock_once(struct kmem_cache *cachep, void *foo) +static void gfs2_init_glock_once(void *foo) { struct gfs2_glock *gl = foo; diff --git a/fs/hfs/super.c b/fs/hfs/super.c index ac2ec5ef66e..4abb1047c68 100644 --- a/fs/hfs/super.c +++ b/fs/hfs/super.c @@ -432,7 +432,7 @@ static struct file_system_type hfs_fs_type = { .fs_flags = FS_REQUIRES_DEV, }; -static void hfs_init_once(struct kmem_cache *cachep, void *p) +static void hfs_init_once(void *p) { struct hfs_inode_info *i = p; diff --git a/fs/hfsplus/super.c b/fs/hfsplus/super.c index 3859118531c..e834e578c93 100644 --- a/fs/hfsplus/super.c +++ b/fs/hfsplus/super.c @@ -485,7 +485,7 @@ static struct file_system_type hfsplus_fs_type = { .fs_flags = FS_REQUIRES_DEV, }; -static void hfsplus_init_once(struct kmem_cache *cachep, void *p) +static void hfsplus_init_once(void *p) { struct hfsplus_inode_info *i = p; diff --git a/fs/hpfs/super.c b/fs/hpfs/super.c index f63a699ec65..b8ae9c90ada 100644 --- a/fs/hpfs/super.c +++ b/fs/hpfs/super.c @@ -173,7 +173,7 @@ static void hpfs_destroy_inode(struct inode *inode) kmem_cache_free(hpfs_inode_cachep, hpfs_i(inode)); } -static void init_once(struct kmem_cache *cachep, void *foo) +static void init_once(void *foo) { struct hpfs_inode_info *ei = (struct hpfs_inode_info *) foo; diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c index dbd01d262ca..3f58923fb39 100644 --- a/fs/hugetlbfs/inode.c +++ b/fs/hugetlbfs/inode.c @@ -705,7 +705,7 @@ static const struct address_space_operations hugetlbfs_aops = { }; -static void init_once(struct kmem_cache *cachep, void *foo) +static void init_once(void *foo) { struct hugetlbfs_inode_info *ei = (struct hugetlbfs_inode_info *)foo; diff --git a/fs/inode.c b/fs/inode.c index 35b6414522e..b6726f64453 100644 --- a/fs/inode.c +++ b/fs/inode.c @@ -224,7 +224,7 @@ void inode_init_once(struct inode *inode) EXPORT_SYMBOL(inode_init_once); -static void init_once(struct kmem_cache * cachep, void *foo) +static void init_once(void *foo) { struct inode * inode = (struct inode *) foo; diff --git a/fs/isofs/inode.c b/fs/isofs/inode.c index 044a254d526..26948a6033b 100644 --- a/fs/isofs/inode.c +++ b/fs/isofs/inode.c @@ -73,7 +73,7 @@ static void isofs_destroy_inode(struct inode *inode) kmem_cache_free(isofs_inode_cachep, ISOFS_I(inode)); } -static void init_once(struct kmem_cache *cachep, void *foo) +static void init_once(void *foo) { struct iso_inode_info *ei = foo; diff --git a/fs/jffs2/super.c b/fs/jffs2/super.c index 7da69eae49e..efd401257ed 100644 --- a/fs/jffs2/super.c +++ b/fs/jffs2/super.c @@ -44,7 +44,7 @@ static void jffs2_destroy_inode(struct inode *inode) kmem_cache_free(jffs2_inode_cachep, JFFS2_INODE_INFO(inode)); } -static void jffs2_i_init_once(struct kmem_cache *cachep, void *foo) +static void jffs2_i_init_once(void *foo) { struct jffs2_inode_info *f = foo; diff --git a/fs/jfs/jfs_metapage.c b/fs/jfs/jfs_metapage.c index 854ff0ec574..c350057087d 100644 --- a/fs/jfs/jfs_metapage.c +++ b/fs/jfs/jfs_metapage.c @@ -182,7 +182,7 @@ static inline void remove_metapage(struct page *page, struct metapage *mp) #endif -static void init_once(struct kmem_cache *cachep, void *foo) +static void init_once(void *foo) { struct metapage *mp = (struct metapage *)foo; diff --git a/fs/jfs/super.c b/fs/jfs/super.c index 359c091d896..3630718be39 100644 --- a/fs/jfs/super.c +++ b/fs/jfs/super.c @@ -760,7 +760,7 @@ static struct file_system_type jfs_fs_type = { .fs_flags = FS_REQUIRES_DEV, }; -static void init_once(struct kmem_cache *cachep, void *foo) +static void init_once(void *foo) { struct jfs_inode_info *jfs_ip = (struct jfs_inode_info *) foo; diff --git a/fs/locks.c b/fs/locks.c index 01490300f7c..5eb259e3cd3 100644 --- a/fs/locks.c +++ b/fs/locks.c @@ -201,7 +201,7 @@ EXPORT_SYMBOL(locks_init_lock); * Initialises the fields of the file lock which are invariant for * free file_locks. */ -static void init_once(struct kmem_cache *cache, void *foo) +static void init_once(void *foo) { struct file_lock *lock = (struct file_lock *) foo; diff --git a/fs/minix/inode.c b/fs/minix/inode.c index 523d7371341..d1d1eb84679 100644 --- a/fs/minix/inode.c +++ b/fs/minix/inode.c @@ -68,7 +68,7 @@ static void minix_destroy_inode(struct inode *inode) kmem_cache_free(minix_inode_cachep, minix_i(inode)); } -static void init_once(struct kmem_cache * cachep, void *foo) +static void init_once(void *foo) { struct minix_inode_info *ei = (struct minix_inode_info *) foo; diff --git a/fs/ncpfs/inode.c b/fs/ncpfs/inode.c index 2e5ab1204de..d642f0e5b36 100644 --- a/fs/ncpfs/inode.c +++ b/fs/ncpfs/inode.c @@ -64,7 +64,7 @@ static void ncp_destroy_inode(struct inode *inode) kmem_cache_free(ncp_inode_cachep, NCP_FINFO(inode)); } -static void init_once(struct kmem_cache *cachep, void *foo) +static void init_once(void *foo) { struct ncp_inode_info *ei = (struct ncp_inode_info *) foo; diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index df23f987da6..52daefa2f52 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -1242,7 +1242,7 @@ static inline void nfs4_init_once(struct nfs_inode *nfsi) #endif } -static void init_once(struct kmem_cache * cachep, void *foo) +static void init_once(void *foo) { struct nfs_inode *nfsi = (struct nfs_inode *) foo; diff --git a/fs/ntfs/super.c b/fs/ntfs/super.c index 3e76f3b216b..4a46743b507 100644 --- a/fs/ntfs/super.c +++ b/fs/ntfs/super.c @@ -3080,7 +3080,7 @@ struct kmem_cache *ntfs_inode_cache; struct kmem_cache *ntfs_big_inode_cache; /* Init once constructor for the inode slab cache. */ -static void ntfs_big_inode_init_once(struct kmem_cache *cachep, void *foo) +static void ntfs_big_inode_init_once(void *foo) { ntfs_inode *ni = (ntfs_inode *)foo; diff --git a/fs/ocfs2/dlm/dlmfs.c b/fs/ocfs2/dlm/dlmfs.c index e48aba698b7..533a789c3ef 100644 --- a/fs/ocfs2/dlm/dlmfs.c +++ b/fs/ocfs2/dlm/dlmfs.c @@ -267,8 +267,7 @@ static ssize_t dlmfs_file_write(struct file *filp, return writelen; } -static void dlmfs_init_once(struct kmem_cache *cachep, - void *foo) +static void dlmfs_init_once(void *foo) { struct dlmfs_inode_private *ip = (struct dlmfs_inode_private *) foo; diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index ccecfe5094f..2560b33889a 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -1118,7 +1118,7 @@ bail: return status; } -static void ocfs2_inode_init_once(struct kmem_cache *cachep, void *data) +static void ocfs2_inode_init_once(void *data) { struct ocfs2_inode_info *oi = data; diff --git a/fs/openpromfs/inode.c b/fs/openpromfs/inode.c index d17b4fd204e..9f5b054f06b 100644 --- a/fs/openpromfs/inode.c +++ b/fs/openpromfs/inode.c @@ -430,7 +430,7 @@ static struct file_system_type openprom_fs_type = { .kill_sb = kill_anon_super, }; -static void op_inode_init_once(struct kmem_cache * cachep, void *data) +static void op_inode_init_once(void *data) { struct op_inode_info *oi = (struct op_inode_info *) data; diff --git a/fs/proc/inode.c b/fs/proc/inode.c index 02eca2ed9dd..b37f25dc45a 100644 --- a/fs/proc/inode.c +++ b/fs/proc/inode.c @@ -94,7 +94,7 @@ static void proc_destroy_inode(struct inode *inode) kmem_cache_free(proc_inode_cachep, PROC_I(inode)); } -static void init_once(struct kmem_cache * cachep, void *foo) +static void init_once(void *foo) { struct proc_inode *ei = (struct proc_inode *) foo; diff --git a/fs/qnx4/inode.c b/fs/qnx4/inode.c index b31ab78052b..2aad1044b84 100644 --- a/fs/qnx4/inode.c +++ b/fs/qnx4/inode.c @@ -553,7 +553,7 @@ static void qnx4_destroy_inode(struct inode *inode) kmem_cache_free(qnx4_inode_cachep, qnx4_i(inode)); } -static void init_once(struct kmem_cache *cachep, void *foo) +static void init_once(void *foo) { struct qnx4_inode_info *ei = (struct qnx4_inode_info *) foo; diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c index 2ec748ba0bd..879e54d35c2 100644 --- a/fs/reiserfs/super.c +++ b/fs/reiserfs/super.c @@ -521,7 +521,7 @@ static void reiserfs_destroy_inode(struct inode *inode) kmem_cache_free(reiserfs_inode_cachep, REISERFS_I(inode)); } -static void init_once(struct kmem_cache * cachep, void *foo) +static void init_once(void *foo) { struct reiserfs_inode_info *ei = (struct reiserfs_inode_info *)foo; diff --git a/fs/romfs/inode.c b/fs/romfs/inode.c index 3f13d491c7c..8e51a2aaa97 100644 --- a/fs/romfs/inode.c +++ b/fs/romfs/inode.c @@ -577,7 +577,7 @@ static void romfs_destroy_inode(struct inode *inode) kmem_cache_free(romfs_inode_cachep, ROMFS_I(inode)); } -static void init_once(struct kmem_cache *cachep, void *foo) +static void init_once(void *foo) { struct romfs_inode_info *ei = foo; diff --git a/fs/smbfs/inode.c b/fs/smbfs/inode.c index 376ef3ee6ed..3528f40ffb0 100644 --- a/fs/smbfs/inode.c +++ b/fs/smbfs/inode.c @@ -67,7 +67,7 @@ static void smb_destroy_inode(struct inode *inode) kmem_cache_free(smb_inode_cachep, SMB_I(inode)); } -static void init_once(struct kmem_cache *cachep, void *foo) +static void init_once(void *foo) { struct smb_inode_info *ei = (struct smb_inode_info *) foo; diff --git a/fs/sysv/inode.c b/fs/sysv/inode.c index c5d60de0658..df0d435baa4 100644 --- a/fs/sysv/inode.c +++ b/fs/sysv/inode.c @@ -326,7 +326,7 @@ static void sysv_destroy_inode(struct inode *inode) kmem_cache_free(sysv_inode_cachep, SYSV_I(inode)); } -static void init_once(struct kmem_cache *cachep, void *p) +static void init_once(void *p) { struct sysv_inode_info *si = (struct sysv_inode_info *)p; diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c index 00eb9c68ad0..ca1e2d4e03c 100644 --- a/fs/ubifs/super.c +++ b/fs/ubifs/super.c @@ -1841,7 +1841,7 @@ static struct file_system_type ubifs_fs_type = { /* * Inode slab cache constructor. */ -static void inode_slab_ctor(struct kmem_cache *cachep, void *obj) +static void inode_slab_ctor(void *obj) { struct ubifs_inode *ui = obj; inode_init_once(&ui->vfs_inode); diff --git a/fs/udf/super.c b/fs/udf/super.c index 44cc702f96c..5698bbf83bb 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -148,7 +148,7 @@ static void udf_destroy_inode(struct inode *inode) kmem_cache_free(udf_inode_cachep, UDF_I(inode)); } -static void init_once(struct kmem_cache *cachep, void *foo) +static void init_once(void *foo) { struct udf_inode_info *ei = (struct udf_inode_info *)foo; diff --git a/fs/ufs/super.c b/fs/ufs/super.c index 227c9d70004..3e30e40aa24 100644 --- a/fs/ufs/super.c +++ b/fs/ufs/super.c @@ -1302,7 +1302,7 @@ static void ufs_destroy_inode(struct inode *inode) kmem_cache_free(ufs_inode_cachep, UFS_I(inode)); } -static void init_once(struct kmem_cache * cachep, void *foo) +static void init_once(void *foo) { struct ufs_inode_info *ei = (struct ufs_inode_info *) foo; diff --git a/fs/xfs/linux-2.6/kmem.h b/fs/xfs/linux-2.6/kmem.h index 5e956490297..a20683cf74d 100644 --- a/fs/xfs/linux-2.6/kmem.h +++ b/fs/xfs/linux-2.6/kmem.h @@ -79,7 +79,7 @@ kmem_zone_init(int size, char *zone_name) static inline kmem_zone_t * kmem_zone_init_flags(int size, char *zone_name, unsigned long flags, - void (*construct)(kmem_zone_t *, void *)) + void (*construct)(void *)) { return kmem_cache_create(zone_name, size, 0, flags, construct); } diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c index 742b2c7852c..943381284e2 100644 --- a/fs/xfs/linux-2.6/xfs_super.c +++ b/fs/xfs/linux-2.6/xfs_super.c @@ -843,7 +843,6 @@ xfs_fs_destroy_inode( STATIC void xfs_fs_inode_init_once( - kmem_zone_t *zonep, void *vnode) { inode_init_once(vn_to_inode((bhv_vnode_t *)vnode)); diff --git a/include/linux/slab.h b/include/linux/slab.h index 41103910f8a..9ff8e849940 100644 --- a/include/linux/slab.h +++ b/include/linux/slab.h @@ -58,7 +58,7 @@ int slab_is_available(void); struct kmem_cache *kmem_cache_create(const char *, size_t, size_t, unsigned long, - void (*)(struct kmem_cache *, void *)); + void (*)(void *)); void kmem_cache_destroy(struct kmem_cache *); int kmem_cache_shrink(struct kmem_cache *); void kmem_cache_free(struct kmem_cache *, void *); diff --git a/include/linux/slub_def.h b/include/linux/slub_def.h index d117ea2825a..5bad61a93f6 100644 --- a/include/linux/slub_def.h +++ b/include/linux/slub_def.h @@ -85,7 +85,7 @@ struct kmem_cache { struct kmem_cache_order_objects min; gfp_t allocflags; /* gfp flags to use on each alloc */ int refcount; /* Refcount for slab cache destroy */ - void (*ctor)(struct kmem_cache *, void *); + void (*ctor)(void *); int inuse; /* Offset to metadata */ int align; /* Alignment */ const char *name; /* Name (only for display!) */ diff --git a/ipc/mqueue.c b/ipc/mqueue.c index 1fdc2eb2f6d..474984f9e03 100644 --- a/ipc/mqueue.c +++ b/ipc/mqueue.c @@ -207,7 +207,7 @@ static int mqueue_get_sb(struct file_system_type *fs_type, return get_sb_single(fs_type, flags, data, mqueue_fill_super, mnt); } -static void init_once(struct kmem_cache *cachep, void *foo) +static void init_once(void *foo) { struct mqueue_inode_info *p = (struct mqueue_inode_info *) foo; diff --git a/kernel/fork.c b/kernel/fork.c index b99d73e971a..80e83e459b1 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1442,7 +1442,7 @@ long do_fork(unsigned long clone_flags, #define ARCH_MIN_MMSTRUCT_ALIGN 0 #endif -static void sighand_ctor(struct kmem_cache *cachep, void *data) +static void sighand_ctor(void *data) { struct sighand_struct *sighand = data; diff --git a/lib/idr.c b/lib/idr.c index 3476f8203e9..e728c7fccc4 100644 --- a/lib/idr.c +++ b/lib/idr.c @@ -607,7 +607,7 @@ void *idr_replace(struct idr *idp, void *ptr, int id) } EXPORT_SYMBOL(idr_replace); -static void idr_cache_ctor(struct kmem_cache *idr_layer_cache, void *idr_layer) +static void idr_cache_ctor(void *idr_layer) { memset(idr_layer, 0, sizeof(struct idr_layer)); } diff --git a/lib/radix-tree.c b/lib/radix-tree.c index 9c4f1ffa286..be86b32bc87 100644 --- a/lib/radix-tree.c +++ b/lib/radix-tree.c @@ -1183,7 +1183,7 @@ int radix_tree_tagged(struct radix_tree_root *root, unsigned int tag) EXPORT_SYMBOL(radix_tree_tagged); static void -radix_tree_node_ctor(struct kmem_cache *cachep, void *node) +radix_tree_node_ctor(void *node) { memset(node, 0, sizeof(struct radix_tree_node)); } diff --git a/mm/rmap.c b/mm/rmap.c index abbd29f7c43..39ae5a9bf38 100644 --- a/mm/rmap.c +++ b/mm/rmap.c @@ -138,7 +138,7 @@ void anon_vma_unlink(struct vm_area_struct *vma) anon_vma_free(anon_vma); } -static void anon_vma_ctor(struct kmem_cache *cachep, void *data) +static void anon_vma_ctor(void *data) { struct anon_vma *anon_vma = data; diff --git a/mm/shmem.c b/mm/shmem.c index 1089092aeca..952d361774b 100644 --- a/mm/shmem.c +++ b/mm/shmem.c @@ -2352,7 +2352,7 @@ static void shmem_destroy_inode(struct inode *inode) kmem_cache_free(shmem_inode_cachep, SHMEM_I(inode)); } -static void init_once(struct kmem_cache *cachep, void *foo) +static void init_once(void *foo) { struct shmem_inode_info *p = (struct shmem_inode_info *) foo; diff --git a/mm/slab.c b/mm/slab.c index 052e7d64537..918f04f7fef 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -406,7 +406,7 @@ struct kmem_cache { unsigned int dflags; /* dynamic flags */ /* constructor func */ - void (*ctor)(struct kmem_cache *, void *); + void (*ctor)(void *obj); /* 5) cache creation/removal */ const char *name; @@ -2137,8 +2137,7 @@ static int __init_refok setup_cpu_cache(struct kmem_cache *cachep) */ struct kmem_cache * kmem_cache_create (const char *name, size_t size, size_t align, - unsigned long flags, - void (*ctor)(struct kmem_cache *, void *)) + unsigned long flags, void (*ctor)(void *)) { size_t left_over, slab_size, ralign; struct kmem_cache *cachep = NULL, *pc; @@ -2653,7 +2652,7 @@ static void cache_init_objs(struct kmem_cache *cachep, * They must also be threaded. */ if (cachep->ctor && !(cachep->flags & SLAB_POISON)) - cachep->ctor(cachep, objp + obj_offset(cachep)); + cachep->ctor(objp + obj_offset(cachep)); if (cachep->flags & SLAB_RED_ZONE) { if (*dbg_redzone2(cachep, objp) != RED_INACTIVE) @@ -2669,7 +2668,7 @@ static void cache_init_objs(struct kmem_cache *cachep, cachep->buffer_size / PAGE_SIZE, 0); #else if (cachep->ctor) - cachep->ctor(cachep, objp); + cachep->ctor(objp); #endif slab_bufctl(slabp)[i] = i + 1; } @@ -3093,7 +3092,7 @@ static void *cache_alloc_debugcheck_after(struct kmem_cache *cachep, #endif objp += obj_offset(cachep); if (cachep->ctor && cachep->flags & SLAB_POISON) - cachep->ctor(cachep, objp); + cachep->ctor(objp); #if ARCH_SLAB_MINALIGN if ((u32)objp & (ARCH_SLAB_MINALIGN-1)) { printk(KERN_ERR "0x%p: not aligned to ARCH_SLAB_MINALIGN=%d\n", diff --git a/mm/slob.c b/mm/slob.c index de268eb7ac7..d8fbd4d1bfa 100644 --- a/mm/slob.c +++ b/mm/slob.c @@ -525,12 +525,11 @@ struct kmem_cache { unsigned int size, align; unsigned long flags; const char *name; - void (*ctor)(struct kmem_cache *, void *); + void (*ctor)(void *); }; struct kmem_cache *kmem_cache_create(const char *name, size_t size, - size_t align, unsigned long flags, - void (*ctor)(struct kmem_cache *, void *)) + size_t align, unsigned long flags, void (*ctor)(void *)) { struct kmem_cache *c; @@ -575,7 +574,7 @@ void *kmem_cache_alloc_node(struct kmem_cache *c, gfp_t flags, int node) b = slob_new_page(flags, get_order(c->size), node); if (c->ctor) - c->ctor(c, b); + c->ctor(b); return b; } diff --git a/mm/slub.c b/mm/slub.c index 77c21cf53ff..b7e2cd5d82d 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -1012,7 +1012,7 @@ __setup("slub_debug", setup_slub_debug); static unsigned long kmem_cache_flags(unsigned long objsize, unsigned long flags, const char *name, - void (*ctor)(struct kmem_cache *, void *)) + void (*ctor)(void *)) { /* * Enable debugging if selected on the kernel commandline. @@ -1040,7 +1040,7 @@ static inline int check_object(struct kmem_cache *s, struct page *page, static inline void add_full(struct kmem_cache_node *n, struct page *page) {} static inline unsigned long kmem_cache_flags(unsigned long objsize, unsigned long flags, const char *name, - void (*ctor)(struct kmem_cache *, void *)) + void (*ctor)(void *)) { return flags; } @@ -1103,7 +1103,7 @@ static void setup_object(struct kmem_cache *s, struct page *page, { setup_object_debug(s, page, object); if (unlikely(s->ctor)) - s->ctor(s, object); + s->ctor(object); } static struct page *new_slab(struct kmem_cache *s, gfp_t flags, int node) @@ -2286,7 +2286,7 @@ static int calculate_sizes(struct kmem_cache *s, int forced_order) static int kmem_cache_open(struct kmem_cache *s, gfp_t gfpflags, const char *name, size_t size, size_t align, unsigned long flags, - void (*ctor)(struct kmem_cache *, void *)) + void (*ctor)(void *)) { memset(s, 0, kmem_size); s->name = name; @@ -3042,7 +3042,7 @@ static int slab_unmergeable(struct kmem_cache *s) static struct kmem_cache *find_mergeable(size_t size, size_t align, unsigned long flags, const char *name, - void (*ctor)(struct kmem_cache *, void *)) + void (*ctor)(void *)) { struct kmem_cache *s; @@ -3082,8 +3082,7 @@ static struct kmem_cache *find_mergeable(size_t size, } struct kmem_cache *kmem_cache_create(const char *name, size_t size, - size_t align, unsigned long flags, - void (*ctor)(struct kmem_cache *, void *)) + size_t align, unsigned long flags, void (*ctor)(void *)) { struct kmem_cache *s; diff --git a/net/socket.c b/net/socket.c index 1310a82cbba..8ef8ba81b9e 100644 --- a/net/socket.c +++ b/net/socket.c @@ -265,7 +265,7 @@ static void sock_destroy_inode(struct inode *inode) container_of(inode, struct socket_alloc, vfs_inode)); } -static void init_once(struct kmem_cache *cachep, void *foo) +static void init_once(void *foo) { struct socket_alloc *ei = (struct socket_alloc *)foo; diff --git a/net/sunrpc/rpc_pipe.c b/net/sunrpc/rpc_pipe.c index 5a9b0e7828c..23a2b8f6dc4 100644 --- a/net/sunrpc/rpc_pipe.c +++ b/net/sunrpc/rpc_pipe.c @@ -897,7 +897,7 @@ static struct file_system_type rpc_pipe_fs_type = { }; static void -init_once(struct kmem_cache * cachep, void *foo) +init_once(void *foo) { struct rpc_inode *rpci = (struct rpc_inode *) foo; -- cgit v1.2.3-70-g09d2 From 5c752ad9f35910ff1912b3f3ae82878178ddc432 Mon Sep 17 00:00:00 2001 From: Arjan van de Ven Date: Fri, 25 Jul 2008 19:45:40 -0700 Subject: Use WARN() in fs/ Use WARN() instead of a printk+WARN_ON() pair; this way the message becomes part of the warning section for better reporting/collection. Signed-off-by: Arjan van de Ven Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/buffer.c | 3 +-- fs/namespace.c | 3 +-- 2 files changed, 2 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/buffer.c b/fs/buffer.c index 5fd497cdd6f..f9580501963 100644 --- a/fs/buffer.c +++ b/fs/buffer.c @@ -1214,8 +1214,7 @@ void __brelse(struct buffer_head * buf) put_bh(buf); return; } - printk(KERN_ERR "VFS: brelse: Trying to free free buffer\n"); - WARN_ON(1); + WARN(1, KERN_ERR "VFS: brelse: Trying to free free buffer\n"); } /* diff --git a/fs/namespace.c b/fs/namespace.c index 4f6f7635b59..f30b11e2240 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -309,10 +309,9 @@ static void handle_write_count_underflow(struct vfsmount *mnt) */ if ((atomic_read(&mnt->__mnt_writers) < 0) && !(mnt->mnt_flags & MNT_IMBALANCED_WRITE_COUNT)) { - printk(KERN_DEBUG "leak detected on mount(%p) writers " + WARN(1, KERN_DEBUG "leak detected on mount(%p) writers " "count: %d\n", mnt, atomic_read(&mnt->__mnt_writers)); - WARN_ON(1); /* use the flag to keep the dmesg spam down */ mnt->mnt_flags |= MNT_IMBALANCED_WRITE_COUNT; } -- cgit v1.2.3-70-g09d2 From 99fcd77d15357e8ba51005c25cc750b9c28b2688 Mon Sep 17 00:00:00 2001 From: Arjan van de Ven Date: Fri, 25 Jul 2008 19:45:41 -0700 Subject: Use WARN() in fs/sysfs Use WARN() instead of a printk+WARN_ON() pair; this way the message becomes part of the warning section for better reporting/collection. Also, with this, one fo the if() sections collapses entirely into the WARN(). Signed-off-by: Arjan van de Ven Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/sysfs/dir.c | 5 +---- fs/sysfs/file.c | 3 +-- fs/sysfs/group.c | 3 +-- 3 files changed, 3 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/sysfs/dir.c b/fs/sysfs/dir.c index c1a7efb310b..aedaeba82ae 100644 --- a/fs/sysfs/dir.c +++ b/fs/sysfs/dir.c @@ -459,11 +459,8 @@ int sysfs_add_one(struct sysfs_addrm_cxt *acxt, struct sysfs_dirent *sd) int ret; ret = __sysfs_add_one(acxt, sd); - if (ret == -EEXIST) { - printk(KERN_WARNING "sysfs: duplicate filename '%s' " + WARN(ret == -EEXIST, KERN_WARNING "sysfs: duplicate filename '%s' " "can not be created\n", sd->s_name); - WARN_ON(1); - } return ret; } diff --git a/fs/sysfs/file.c b/fs/sysfs/file.c index 3f07893ff89..c9e4e5091da 100644 --- a/fs/sysfs/file.c +++ b/fs/sysfs/file.c @@ -337,9 +337,8 @@ static int sysfs_open_file(struct inode *inode, struct file *file) if (kobj->ktype && kobj->ktype->sysfs_ops) ops = kobj->ktype->sysfs_ops; else { - printk(KERN_ERR "missing sysfs attribute operations for " + WARN(1, KERN_ERR "missing sysfs attribute operations for " "kobject: %s\n", kobject_name(kobj)); - WARN_ON(1); goto err_out; } diff --git a/fs/sysfs/group.c b/fs/sysfs/group.c index eeba38417b1..fe611949a7f 100644 --- a/fs/sysfs/group.c +++ b/fs/sysfs/group.c @@ -134,9 +134,8 @@ void sysfs_remove_group(struct kobject * kobj, if (grp->name) { sd = sysfs_get_dirent(dir_sd, grp->name); if (!sd) { - printk(KERN_WARNING "sysfs group %p not found for " + WARN(!sd, KERN_WARNING "sysfs group %p not found for " "kobject '%s'\n", grp, kobject_name(kobj)); - WARN_ON(!sd); return; } } else -- cgit v1.2.3-70-g09d2 From 267e2a9c71b8e088ac307f9549f71468e86e26c1 Mon Sep 17 00:00:00 2001 From: Arjan van de Ven Date: Fri, 25 Jul 2008 19:45:41 -0700 Subject: Use WARN() in fs/proc/ Use WARN() instead of a printk+WARN_ON() pair; this way the message becomes part of the warning section for better reporting/collection. This way, the entire if() {} section can collapse into the WARN() as well. Signed-off-by: Arjan van de Ven Acked-by: Alexey Dobriyan Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/generic.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/proc/generic.c b/fs/proc/generic.c index bc0a0dd2d84..cb4096cc3fb 100644 --- a/fs/proc/generic.c +++ b/fs/proc/generic.c @@ -806,12 +806,9 @@ continue_removing: if (S_ISDIR(de->mode)) parent->nlink--; de->nlink = 0; - if (de->subdir) { - printk(KERN_WARNING "%s: removing non-empty directory " + WARN(de->subdir, KERN_WARNING "%s: removing non-empty directory " "'%s/%s', leaking at least '%s'\n", __func__, de->parent->name, de->name, de->subdir->name); - WARN_ON(1); - } if (atomic_dec_and_test(&de->count)) free_proc_entry(de); } -- cgit v1.2.3-70-g09d2 From 6341c393fcc37d58727865f1ee2f65e632e9d4f0 Mon Sep 17 00:00:00 2001 From: Roland McGrath Date: Fri, 25 Jul 2008 19:45:44 -0700 Subject: tracehook: exec This moves all the ptrace hooks related to exec into tracehook.h inlines. This also lifts the calls for tracing out of the binfmt load_binary hooks into search_binary_handler() after it calls into the binfmt module. This change has no effect, since all the binfmt modules' load_binary functions did the call at the end on success, and now search_binary_handler() does it immediately after return if successful. We consolidate the repeated code, and binfmt modules no longer need to import ptrace_notify(). Signed-off-by: Roland McGrath Cc: Oleg Nesterov Reviewed-by: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/x86/ia32/ia32_aout.c | 6 ------ fs/binfmt_aout.c | 6 ------ fs/binfmt_elf.c | 6 ------ fs/binfmt_elf_fdpic.c | 7 ------- fs/binfmt_flat.c | 3 --- fs/binfmt_som.c | 2 -- fs/exec.c | 12 ++++-------- include/linux/tracehook.h | 46 ++++++++++++++++++++++++++++++++++++++++++++++ 8 files changed, 50 insertions(+), 38 deletions(-) (limited to 'fs') diff --git a/arch/x86/ia32/ia32_aout.c b/arch/x86/ia32/ia32_aout.c index 58cccb6483b..a0e1dbe67dc 100644 --- a/arch/x86/ia32/ia32_aout.c +++ b/arch/x86/ia32/ia32_aout.c @@ -441,12 +441,6 @@ beyond_if: regs->r8 = regs->r9 = regs->r10 = regs->r11 = regs->r12 = regs->r13 = regs->r14 = regs->r15 = 0; set_fs(USER_DS); - if (unlikely(current->ptrace & PT_PTRACED)) { - if (current->ptrace & PT_TRACE_EXEC) - ptrace_notify((PTRACE_EVENT_EXEC << 8) | SIGTRAP); - else - send_sig(SIGTRAP, current, 0); - } return 0; } diff --git a/fs/binfmt_aout.c b/fs/binfmt_aout.c index ba4cddb92f1..204cfd1d767 100644 --- a/fs/binfmt_aout.c +++ b/fs/binfmt_aout.c @@ -444,12 +444,6 @@ beyond_if: regs->gp = ex.a_gpvalue; #endif start_thread(regs, ex.a_entry, current->mm->start_stack); - if (unlikely(current->ptrace & PT_PTRACED)) { - if (current->ptrace & PT_TRACE_EXEC) - ptrace_notify ((PTRACE_EVENT_EXEC << 8) | SIGTRAP); - else - send_sig(SIGTRAP, current, 0); - } return 0; } diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c index 3b6ff854d98..655ed8d30a8 100644 --- a/fs/binfmt_elf.c +++ b/fs/binfmt_elf.c @@ -1003,12 +1003,6 @@ static int load_elf_binary(struct linux_binprm *bprm, struct pt_regs *regs) #endif start_thread(regs, elf_entry, bprm->p); - if (unlikely(current->ptrace & PT_PTRACED)) { - if (current->ptrace & PT_TRACE_EXEC) - ptrace_notify ((PTRACE_EVENT_EXEC << 8) | SIGTRAP); - else - send_sig(SIGTRAP, current, 0); - } retval = 0; out: kfree(loc); diff --git a/fs/binfmt_elf_fdpic.c b/fs/binfmt_elf_fdpic.c index 1b59b1edf26..fdeadab2f18 100644 --- a/fs/binfmt_elf_fdpic.c +++ b/fs/binfmt_elf_fdpic.c @@ -433,13 +433,6 @@ static int load_elf_fdpic_binary(struct linux_binprm *bprm, entryaddr = interp_params.entry_addr ?: exec_params.entry_addr; start_thread(regs, entryaddr, current->mm->start_stack); - if (unlikely(current->ptrace & PT_PTRACED)) { - if (current->ptrace & PT_TRACE_EXEC) - ptrace_notify((PTRACE_EVENT_EXEC << 8) | SIGTRAP); - else - send_sig(SIGTRAP, current, 0); - } - retval = 0; error: diff --git a/fs/binfmt_flat.c b/fs/binfmt_flat.c index 2cb1acda3a8..56372ecf169 100644 --- a/fs/binfmt_flat.c +++ b/fs/binfmt_flat.c @@ -920,9 +920,6 @@ static int load_flat_binary(struct linux_binprm * bprm, struct pt_regs * regs) start_thread(regs, start_addr, current->mm->start_stack); - if (current->ptrace & PT_PTRACED) - send_sig(SIGTRAP, current, 0); - return 0; } diff --git a/fs/binfmt_som.c b/fs/binfmt_som.c index fdc36bfd6a7..68be580ba28 100644 --- a/fs/binfmt_som.c +++ b/fs/binfmt_som.c @@ -274,8 +274,6 @@ load_som_binary(struct linux_binprm * bprm, struct pt_regs * regs) map_hpux_gateway_page(current,current->mm); start_thread_som(regs, som_entry, bprm->p); - if (current->ptrace & PT_PTRACED) - send_sig(SIGTRAP, current, 0); return 0; /* error cleanup */ diff --git a/fs/exec.c b/fs/exec.c index 5e559013e30..b8792a13153 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -42,13 +42,13 @@ #include #include #include -#include #include #include #include #include #include #include +#include #include #include @@ -1071,13 +1071,8 @@ EXPORT_SYMBOL(prepare_binprm); static int unsafe_exec(struct task_struct *p) { - int unsafe = 0; - if (p->ptrace & PT_PTRACED) { - if (p->ptrace & PT_PTRACE_CAP) - unsafe |= LSM_UNSAFE_PTRACE_CAP; - else - unsafe |= LSM_UNSAFE_PTRACE; - } + int unsafe = tracehook_unsafe_exec(p); + if (atomic_read(&p->fs->count) > 1 || atomic_read(&p->files->count) > 1 || atomic_read(&p->sighand->count) > 1) @@ -1214,6 +1209,7 @@ int search_binary_handler(struct linux_binprm *bprm,struct pt_regs *regs) read_unlock(&binfmt_lock); retval = fn(bprm, regs); if (retval >= 0) { + tracehook_report_exec(fmt, bprm, regs); put_binfmt(fmt); allow_write_access(bprm->file); if (bprm->file) diff --git a/include/linux/tracehook.h b/include/linux/tracehook.h index bea0f3eeff5..6276353709c 100644 --- a/include/linux/tracehook.h +++ b/include/linux/tracehook.h @@ -48,5 +48,51 @@ #include #include +#include +struct linux_binprm; + +/** + * tracehook_unsafe_exec - check for exec declared unsafe due to tracing + * @task: current task doing exec + * + * Return %LSM_UNSAFE_* bits applied to an exec because of tracing. + * + * Called with task_lock() held on @task. + */ +static inline int tracehook_unsafe_exec(struct task_struct *task) +{ + int unsafe = 0; + int ptrace = task_ptrace(task); + if (ptrace & PT_PTRACED) { + if (ptrace & PT_PTRACE_CAP) + unsafe |= LSM_UNSAFE_PTRACE_CAP; + else + unsafe |= LSM_UNSAFE_PTRACE; + } + return unsafe; +} + +/** + * tracehook_report_exec - a successful exec was completed + * @fmt: &struct linux_binfmt that performed the exec + * @bprm: &struct linux_binprm containing exec details + * @regs: user-mode register state + * + * An exec just completed, we are shortly going to return to user mode. + * The freshly initialized register state can be seen and changed in @regs. + * The name, file and other pointers in @bprm are still on hand to be + * inspected, but will be freed as soon as this returns. + * + * Called with no locks, but with some kernel resources held live + * and a reference on @fmt->module. + */ +static inline void tracehook_report_exec(struct linux_binfmt *fmt, + struct linux_binprm *bprm, + struct pt_regs *regs) +{ + if (!ptrace_event(PT_TRACE_EXEC, PTRACE_EVENT_EXEC, 0) && + unlikely(task_ptrace(current) & PT_PTRACED)) + send_sig(SIGTRAP, current, 0); +} #endif /* */ -- cgit v1.2.3-70-g09d2 From 0d094efeb1e98010c6b99923f1eb7e17bf1e3a74 Mon Sep 17 00:00:00 2001 From: Roland McGrath Date: Fri, 25 Jul 2008 19:45:49 -0700 Subject: tracehook: tracehook_tracer_task This adds the tracehook_tracer_task() hook to consolidate all forms of "Who is using ptrace on me?" logic. This is used for "TracerPid:" in /proc and for permission checks. We also clean up the selinux code the called an identical accessor. Signed-off-by: Roland McGrath Cc: Oleg Nesterov Reviewed-by: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/array.c | 9 +++++++-- fs/proc/base.c | 13 +++++++++---- include/linux/tracehook.h | 18 ++++++++++++++++++ security/selinux/hooks.c | 22 +++------------------- 4 files changed, 37 insertions(+), 25 deletions(-) (limited to 'fs') diff --git a/fs/proc/array.c b/fs/proc/array.c index 797d775e035..0d6eb33597c 100644 --- a/fs/proc/array.c +++ b/fs/proc/array.c @@ -80,6 +80,7 @@ #include #include #include +#include #include #include @@ -168,8 +169,12 @@ static inline void task_state(struct seq_file *m, struct pid_namespace *ns, rcu_read_lock(); ppid = pid_alive(p) ? task_tgid_nr_ns(rcu_dereference(p->real_parent), ns) : 0; - tpid = pid_alive(p) && p->ptrace ? - task_pid_nr_ns(rcu_dereference(p->parent), ns) : 0; + tpid = 0; + if (pid_alive(p)) { + struct task_struct *tracer = tracehook_tracer_task(p); + if (tracer) + tpid = task_pid_nr_ns(tracer, ns); + } seq_printf(m, "State:\t%s\n" "Tgid:\t%d\n" diff --git a/fs/proc/base.c b/fs/proc/base.c index a891fe4cb43..4b74dba69a6 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -69,6 +69,7 @@ #include #include #include +#include #include #include #include @@ -231,10 +232,14 @@ static int check_mem_permission(struct task_struct *task) * If current is actively ptrace'ing, and would also be * permitted to freshly attach with ptrace now, permit it. */ - if (task->parent == current && (task->ptrace & PT_PTRACED) && - task_is_stopped_or_traced(task) && - ptrace_may_access(task, PTRACE_MODE_ATTACH)) - return 0; + if (task_is_stopped_or_traced(task)) { + int match; + rcu_read_lock(); + match = (tracehook_tracer_task(task) == current); + rcu_read_unlock(); + if (match && ptrace_may_access(task, PTRACE_MODE_ATTACH)) + return 0; + } /* * Noone else is allowed. diff --git a/include/linux/tracehook.h b/include/linux/tracehook.h index 9a5b3be2503..6468ca0fe69 100644 --- a/include/linux/tracehook.h +++ b/include/linux/tracehook.h @@ -72,6 +72,24 @@ static inline int tracehook_unsafe_exec(struct task_struct *task) return unsafe; } +/** + * tracehook_tracer_task - return the task that is tracing the given task + * @tsk: task to consider + * + * Returns NULL if noone is tracing @task, or the &struct task_struct + * pointer to its tracer. + * + * Must called under rcu_read_lock(). The pointer returned might be kept + * live only by RCU. During exec, this may be called with task_lock() + * held on @task, still held from when tracehook_unsafe_exec() was called. + */ +static inline struct task_struct *tracehook_tracer_task(struct task_struct *tsk) +{ + if (task_ptrace(tsk) & PT_PTRACED) + return rcu_dereference(tsk->parent); + return NULL; +} + /** * tracehook_report_exec - a successful exec was completed * @fmt: &struct linux_binfmt that performed the exec diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c index 63f131fc42e..3481cde5bf1 100644 --- a/security/selinux/hooks.c +++ b/security/selinux/hooks.c @@ -25,7 +25,7 @@ #include #include -#include +#include #include #include #include @@ -1971,22 +1971,6 @@ static int selinux_vm_enough_memory(struct mm_struct *mm, long pages) return __vm_enough_memory(mm, pages, cap_sys_admin); } -/** - * task_tracer_task - return the task that is tracing the given task - * @task: task to consider - * - * Returns NULL if noone is tracing @task, or the &struct task_struct - * pointer to its tracer. - * - * Must be called under rcu_read_lock(). - */ -static struct task_struct *task_tracer_task(struct task_struct *task) -{ - if (task->ptrace & PT_PTRACED) - return rcu_dereference(task->parent); - return NULL; -} - /* binprm security operations */ static int selinux_bprm_alloc_security(struct linux_binprm *bprm) @@ -2238,7 +2222,7 @@ static void selinux_bprm_apply_creds(struct linux_binprm *bprm, int unsafe) u32 ptsid = 0; rcu_read_lock(); - tracer = task_tracer_task(current); + tracer = tracehook_tracer_task(current); if (likely(tracer != NULL)) { sec = tracer->security; ptsid = sec->sid; @@ -5247,7 +5231,7 @@ static int selinux_setprocattr(struct task_struct *p, Otherwise, leave SID unchanged and fail. */ task_lock(p); rcu_read_lock(); - tracer = task_tracer_task(p); + tracer = tracehook_tracer_task(p); if (tracer != NULL) { struct task_security_struct *ptsec = tracer->security; u32 ptsid = ptsec->sid; -- cgit v1.2.3-70-g09d2 From ebcb67341fee34061430f3367f2e507e52ee051b Mon Sep 17 00:00:00 2001 From: Roland McGrath Date: Fri, 25 Jul 2008 19:46:00 -0700 Subject: /proc/PID/syscall This adds /proc/PID/syscall and /proc/PID/task/TID/syscall magic files. These use task_current_syscall() to show the task's current system call number and argument registers, stack pointer and PC. For a task blocked but not in a syscall, the file shows "-1" in place of the syscall number, followed by only the SP and PC. For a task that's not blocked, it shows "running". Signed-off-by: Roland McGrath Cc: Oleg Nesterov Reviewed-by: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/base.c | 26 ++++++++++++++++++++++++++ 1 file changed, 26 insertions(+) (limited to 'fs') diff --git a/fs/proc/base.c b/fs/proc/base.c index 4b74dba69a6..81bce6791bf 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -509,6 +509,26 @@ static int proc_pid_limits(struct task_struct *task, char *buffer) return count; } +#ifdef CONFIG_HAVE_ARCH_TRACEHOOK +static int proc_pid_syscall(struct task_struct *task, char *buffer) +{ + long nr; + unsigned long args[6], sp, pc; + + if (task_current_syscall(task, &nr, args, 6, &sp, &pc)) + return sprintf(buffer, "running\n"); + + if (nr < 0) + return sprintf(buffer, "%ld 0x%lx 0x%lx\n", nr, sp, pc); + + return sprintf(buffer, + "%ld 0x%lx 0x%lx 0x%lx 0x%lx 0x%lx 0x%lx 0x%lx 0x%lx\n", + nr, + args[0], args[1], args[2], args[3], args[4], args[5], + sp, pc); +} +#endif /* CONFIG_HAVE_ARCH_TRACEHOOK */ + /************************************************************************/ /* Here the fs part begins */ /************************************************************************/ @@ -2477,6 +2497,9 @@ static const struct pid_entry tgid_base_stuff[] = { INF("limits", S_IRUSR, pid_limits), #ifdef CONFIG_SCHED_DEBUG REG("sched", S_IRUGO|S_IWUSR, pid_sched), +#endif +#ifdef CONFIG_HAVE_ARCH_TRACEHOOK + INF("syscall", S_IRUSR, pid_syscall), #endif INF("cmdline", S_IRUGO, pid_cmdline), ONE("stat", S_IRUGO, tgid_stat), @@ -2809,6 +2832,9 @@ static const struct pid_entry tid_base_stuff[] = { INF("limits", S_IRUSR, pid_limits), #ifdef CONFIG_SCHED_DEBUG REG("sched", S_IRUGO|S_IWUSR, pid_sched), +#endif +#ifdef CONFIG_HAVE_ARCH_TRACEHOOK + INF("syscall", S_IRUSR, pid_syscall), #endif INF("cmdline", S_IRUGO, pid_cmdline), ONE("stat", S_IRUGO, tid_stat), -- cgit v1.2.3-70-g09d2 From d70b67c8bc72ee23b55381bd6a884f4796692f77 Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Wed, 2 Jul 2008 21:30:15 +0200 Subject: [patch] vfs: fix lookup on deleted directory Lookup can install a child dentry for a deleted directory. This keeps the directory dentry alive, and the inode pinned in the cache and on disk, even after all external references have gone away. This isn't a big problem normally, since memory pressure or umount will clear out the directory dentry and its children, releasing the inode. But for UBIFS this causes problems because its orphan area can overflow. Fix this by returning ENOENT for all lookups on a S_DEAD directory before creating a child dentry. Thanks to Zoltan Sogor for noticing this while testing UBIFS, and Artem for the excellent analysis of the problem and testing. Reported-by: Artem Bityutskiy Tested-by: Artem Bityutskiy Signed-off-by: Miklos Szeredi Signed-off-by: Al Viro --- fs/namei.c | 19 +++++++++++++++++-- 1 file changed, 17 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/namei.c b/fs/namei.c index 01e67dddcc3..3b26a240ade 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -519,7 +519,14 @@ static struct dentry * real_lookup(struct dentry * parent, struct qstr * name, s */ result = d_lookup(parent, name); if (!result) { - struct dentry * dentry = d_alloc(parent, name); + struct dentry *dentry; + + /* Don't create child dentry for a dead directory. */ + result = ERR_PTR(-ENOENT); + if (IS_DEADDIR(dir)) + goto out_unlock; + + dentry = d_alloc(parent, name); result = ERR_PTR(-ENOMEM); if (dentry) { result = dir->i_op->lookup(dir, dentry, nd); @@ -528,6 +535,7 @@ static struct dentry * real_lookup(struct dentry * parent, struct qstr * name, s else result = dentry; } +out_unlock: mutex_unlock(&dir->i_mutex); return result; } @@ -1317,7 +1325,14 @@ static struct dentry *__lookup_hash(struct qstr *name, dentry = cached_lookup(base, name, nd); if (!dentry) { - struct dentry *new = d_alloc(base, name); + struct dentry *new; + + /* Don't create child dentry for a dead directory. */ + dentry = ERR_PTR(-ENOENT); + if (IS_DEADDIR(inode)) + goto out; + + new = d_alloc(base, name); dentry = ERR_PTR(-ENOMEM); if (!new) goto out; -- cgit v1.2.3-70-g09d2 From d2d9648ec6858e19d16a0b16da62534e85888653 Mon Sep 17 00:00:00 2001 From: Denys Vlasenko Date: Tue, 1 Jul 2008 14:16:09 +0200 Subject: [PATCH] reuse xxx_fifo_fops for xxx_pipe_fops Merge fifo and pipe file_operations. Signed-off-by: Denys Vlasenko Signed-off-by: Al Viro --- fs/fifo.c | 8 ++++---- fs/pipe.c | 51 ++++++++------------------------------------------- include/linux/fs.h | 6 +++--- 3 files changed, 15 insertions(+), 50 deletions(-) (limited to 'fs') diff --git a/fs/fifo.c b/fs/fifo.c index 9785e36f81e..987bf941149 100644 --- a/fs/fifo.c +++ b/fs/fifo.c @@ -57,7 +57,7 @@ static int fifo_open(struct inode *inode, struct file *filp) * POSIX.1 says that O_NONBLOCK means return with the FIFO * opened, even when there is no process writing the FIFO. */ - filp->f_op = &read_fifo_fops; + filp->f_op = &read_pipefifo_fops; pipe->r_counter++; if (pipe->readers++ == 0) wake_up_partner(inode); @@ -86,7 +86,7 @@ static int fifo_open(struct inode *inode, struct file *filp) if ((filp->f_flags & O_NONBLOCK) && !pipe->readers) goto err; - filp->f_op = &write_fifo_fops; + filp->f_op = &write_pipefifo_fops; pipe->w_counter++; if (!pipe->writers++) wake_up_partner(inode); @@ -105,7 +105,7 @@ static int fifo_open(struct inode *inode, struct file *filp) * This implementation will NEVER block on a O_RDWR open, since * the process can at least talk to itself. */ - filp->f_op = &rdwr_fifo_fops; + filp->f_op = &rdwr_pipefifo_fops; pipe->readers++; pipe->writers++; @@ -151,5 +151,5 @@ err_nocleanup: * depending on the access mode of the file... */ const struct file_operations def_fifo_fops = { - .open = fifo_open, /* will set read or write pipe_fops */ + .open = fifo_open, /* will set read_ or write_pipefifo_fops */ }; diff --git a/fs/pipe.c b/fs/pipe.c index 10c4e9aa5c4..fcba6542b8d 100644 --- a/fs/pipe.c +++ b/fs/pipe.c @@ -777,45 +777,10 @@ pipe_rdwr_open(struct inode *inode, struct file *filp) /* * The file_operations structs are not static because they * are also used in linux/fs/fifo.c to do operations on FIFOs. + * + * Pipes reuse fifos' file_operations structs. */ -const struct file_operations read_fifo_fops = { - .llseek = no_llseek, - .read = do_sync_read, - .aio_read = pipe_read, - .write = bad_pipe_w, - .poll = pipe_poll, - .unlocked_ioctl = pipe_ioctl, - .open = pipe_read_open, - .release = pipe_read_release, - .fasync = pipe_read_fasync, -}; - -const struct file_operations write_fifo_fops = { - .llseek = no_llseek, - .read = bad_pipe_r, - .write = do_sync_write, - .aio_write = pipe_write, - .poll = pipe_poll, - .unlocked_ioctl = pipe_ioctl, - .open = pipe_write_open, - .release = pipe_write_release, - .fasync = pipe_write_fasync, -}; - -const struct file_operations rdwr_fifo_fops = { - .llseek = no_llseek, - .read = do_sync_read, - .aio_read = pipe_read, - .write = do_sync_write, - .aio_write = pipe_write, - .poll = pipe_poll, - .unlocked_ioctl = pipe_ioctl, - .open = pipe_rdwr_open, - .release = pipe_rdwr_release, - .fasync = pipe_rdwr_fasync, -}; - -static const struct file_operations read_pipe_fops = { +const struct file_operations read_pipefifo_fops = { .llseek = no_llseek, .read = do_sync_read, .aio_read = pipe_read, @@ -827,7 +792,7 @@ static const struct file_operations read_pipe_fops = { .fasync = pipe_read_fasync, }; -static const struct file_operations write_pipe_fops = { +const struct file_operations write_pipefifo_fops = { .llseek = no_llseek, .read = bad_pipe_r, .write = do_sync_write, @@ -839,7 +804,7 @@ static const struct file_operations write_pipe_fops = { .fasync = pipe_write_fasync, }; -static const struct file_operations rdwr_pipe_fops = { +const struct file_operations rdwr_pipefifo_fops = { .llseek = no_llseek, .read = do_sync_read, .aio_read = pipe_read, @@ -927,7 +892,7 @@ static struct inode * get_pipe_inode(void) inode->i_pipe = pipe; pipe->readers = pipe->writers = 1; - inode->i_fop = &rdwr_pipe_fops; + inode->i_fop = &rdwr_pipefifo_fops; /* * Mark the inode dirty from the very beginning, @@ -978,7 +943,7 @@ struct file *create_write_pipe(int flags) d_instantiate(dentry, inode); err = -ENFILE; - f = alloc_file(pipe_mnt, dentry, FMODE_WRITE, &write_pipe_fops); + f = alloc_file(pipe_mnt, dentry, FMODE_WRITE, &write_pipefifo_fops); if (!f) goto err_dentry; f->f_mapping = inode->i_mapping; @@ -1020,7 +985,7 @@ struct file *create_read_pipe(struct file *wrf, int flags) f->f_pos = 0; f->f_flags = O_RDONLY | (flags & O_NONBLOCK); - f->f_op = &read_pipe_fops; + f->f_op = &read_pipefifo_fops; f->f_mode = FMODE_READ; f->f_version = 0; diff --git a/include/linux/fs.h b/include/linux/fs.h index 53d2edb709b..7721a2ac9c0 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1696,9 +1696,9 @@ extern void init_special_inode(struct inode *, umode_t, dev_t); extern void make_bad_inode(struct inode *); extern int is_bad_inode(struct inode *); -extern const struct file_operations read_fifo_fops; -extern const struct file_operations write_fifo_fops; -extern const struct file_operations rdwr_fifo_fops; +extern const struct file_operations read_pipefifo_fops; +extern const struct file_operations write_pipefifo_fops; +extern const struct file_operations rdwr_pipefifo_fops; extern int fs_may_remount_ro(struct super_block *); -- cgit v1.2.3-70-g09d2 From 7ac6cd653d7c31ad6b7bb5b88c549c4ebf628c34 Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Tue, 1 Jul 2008 23:07:54 +0200 Subject: [patch] hppfs: remove hppfs_permission hppfs_permission() is equivalent to the '.permission == NULL' case. Signed-off-by: Miklos Szeredi Signed-off-by: Al Viro --- fs/hppfs/hppfs.c | 7 ------- 1 file changed, 7 deletions(-) (limited to 'fs') diff --git a/fs/hppfs/hppfs.c b/fs/hppfs/hppfs.c index 65077aa90f0..2b3d1828db9 100644 --- a/fs/hppfs/hppfs.c +++ b/fs/hppfs/hppfs.c @@ -655,20 +655,13 @@ static void *hppfs_follow_link(struct dentry *dentry, struct nameidata *nd) return proc_dentry->d_inode->i_op->follow_link(proc_dentry, nd); } -int hppfs_permission(struct inode *inode, int mask, struct nameidata *nd) -{ - return generic_permission(inode, mask, NULL); -} - static const struct inode_operations hppfs_dir_iops = { .lookup = hppfs_lookup, - .permission = hppfs_permission, }; static const struct inode_operations hppfs_link_iops = { .readlink = hppfs_readlink, .follow_link = hppfs_follow_link, - .permission = hppfs_permission, }; static struct inode *get_inode(struct super_block *sb, struct dentry *dentry) -- cgit v1.2.3-70-g09d2 From 9043476f726802f4b00c96d0c4f418dde48d1304 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 15 Jul 2008 08:54:06 -0400 Subject: [PATCH] sanitize proc_sysctl * keep references to ctl_table_head and ctl_table in /proc/sys inodes * grab the former during operations, use the latter for access to entry if that succeeds * have ->d_compare() check if table should be seen for one who does lookup; that allows us to avoid flipping inodes - if we have the same name resolve to different things, we'll just keep several dentries and ->d_compare() will reject the wrong ones. * have ->lookup() and ->readdir() scan the table of our inode first, then walk all ctl_table_header and scan ->attached_by for those that are attached to our directory. * implement ->getattr(). * get rid of insane amounts of tree-walking * get rid of the need to know dentry in ->permission() and of the contortions induced by that. Signed-off-by: Al Viro --- fs/proc/inode.c | 5 + fs/proc/proc_sysctl.c | 427 ++++++++++++++++++++++-------------------------- include/linux/proc_fs.h | 5 + include/linux/sysctl.h | 1 + kernel/sysctl.c | 15 ++ 5 files changed, 218 insertions(+), 235 deletions(-) (limited to 'fs') diff --git a/fs/proc/inode.c b/fs/proc/inode.c index b37f25dc45a..8bb03f056c2 100644 --- a/fs/proc/inode.c +++ b/fs/proc/inode.c @@ -17,6 +17,7 @@ #include #include #include +#include #include #include @@ -65,6 +66,8 @@ static void proc_delete_inode(struct inode *inode) module_put(de->owner); de_put(de); } + if (PROC_I(inode)->sysctl) + sysctl_head_put(PROC_I(inode)->sysctl); clear_inode(inode); } @@ -84,6 +87,8 @@ static struct inode *proc_alloc_inode(struct super_block *sb) ei->fd = 0; ei->op.proc_get_link = NULL; ei->pde = NULL; + ei->sysctl = NULL; + ei->sysctl_entry = NULL; inode = &ei->vfs_inode; inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME; return inode; diff --git a/fs/proc/proc_sysctl.c b/fs/proc/proc_sysctl.c index 5acc001d49f..fa1ec2433e4 100644 --- a/fs/proc/proc_sysctl.c +++ b/fs/proc/proc_sysctl.c @@ -10,149 +10,110 @@ static struct dentry_operations proc_sys_dentry_operations; static const struct file_operations proc_sys_file_operations; static const struct inode_operations proc_sys_inode_operations; +static const struct file_operations proc_sys_dir_file_operations; +static const struct inode_operations proc_sys_dir_operations; -static void proc_sys_refresh_inode(struct inode *inode, struct ctl_table *table) -{ - /* Refresh the cached information bits in the inode */ - if (table) { - inode->i_uid = 0; - inode->i_gid = 0; - inode->i_mode = table->mode; - if (table->proc_handler) { - inode->i_mode |= S_IFREG; - inode->i_nlink = 1; - } else { - inode->i_mode |= S_IFDIR; - inode->i_nlink = 0; /* It is too hard to figure out */ - } - } -} - -static struct inode *proc_sys_make_inode(struct inode *dir, struct ctl_table *table) +static struct inode *proc_sys_make_inode(struct super_block *sb, + struct ctl_table_header *head, struct ctl_table *table) { struct inode *inode; - struct proc_inode *dir_ei, *ei; - int depth; + struct proc_inode *ei; - inode = new_inode(dir->i_sb); + inode = new_inode(sb); if (!inode) goto out; - /* A directory is always one deeper than it's parent */ - dir_ei = PROC_I(dir); - depth = dir_ei->fd + 1; - + sysctl_head_get(head); ei = PROC_I(inode); - ei->fd = depth; + ei->sysctl = head; + ei->sysctl_entry = table; + inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME; - inode->i_op = &proc_sys_inode_operations; - inode->i_fop = &proc_sys_file_operations; inode->i_flags |= S_PRIVATE; /* tell selinux to ignore this inode */ - proc_sys_refresh_inode(inode, table); + inode->i_mode = table->mode; + if (!table->child) { + inode->i_mode |= S_IFREG; + inode->i_op = &proc_sys_inode_operations; + inode->i_fop = &proc_sys_file_operations; + } else { + inode->i_mode |= S_IFDIR; + inode->i_nlink = 0; + inode->i_op = &proc_sys_dir_operations; + inode->i_fop = &proc_sys_dir_file_operations; + } out: return inode; } -static struct dentry *proc_sys_ancestor(struct dentry *dentry, int depth) -{ - for (;;) { - struct proc_inode *ei; - - ei = PROC_I(dentry->d_inode); - if (ei->fd == depth) - break; /* found */ - - dentry = dentry->d_parent; - } - return dentry; -} - -static struct ctl_table *proc_sys_lookup_table_one(struct ctl_table *table, - struct qstr *name) +static struct ctl_table *find_in_table(struct ctl_table *p, struct qstr *name) { int len; - for ( ; table->ctl_name || table->procname; table++) { + for ( ; p->ctl_name || p->procname; p++) { - if (!table->procname) + if (!p->procname) continue; - len = strlen(table->procname); + len = strlen(p->procname); if (len != name->len) continue; - if (memcmp(table->procname, name->name, len) != 0) + if (memcmp(p->procname, name->name, len) != 0) continue; /* I have a match */ - return table; + return p; } return NULL; } -static struct ctl_table *proc_sys_lookup_table(struct dentry *dentry, - struct ctl_table *table) +struct ctl_table_header *grab_header(struct inode *inode) { - struct dentry *ancestor; - struct proc_inode *ei; - int depth, i; + if (PROC_I(inode)->sysctl) + return sysctl_head_grab(PROC_I(inode)->sysctl); + else + return sysctl_head_next(NULL); +} - ei = PROC_I(dentry->d_inode); - depth = ei->fd; +static struct dentry *proc_sys_lookup(struct inode *dir, struct dentry *dentry, + struct nameidata *nd) +{ + struct ctl_table_header *head = grab_header(dir); + struct ctl_table *table = PROC_I(dir)->sysctl_entry; + struct ctl_table_header *h = NULL; + struct qstr *name = &dentry->d_name; + struct ctl_table *p; + struct inode *inode; + struct dentry *err = ERR_PTR(-ENOENT); - if (depth == 0) - return table; + if (IS_ERR(head)) + return ERR_CAST(head); - for (i = 1; table && (i <= depth); i++) { - ancestor = proc_sys_ancestor(dentry, i); - table = proc_sys_lookup_table_one(table, &ancestor->d_name); - if (table) - table = table->child; + if (table && !table->child) { + WARN_ON(1); + goto out; } - return table; - -} -static struct ctl_table *proc_sys_lookup_entry(struct dentry *dparent, - struct qstr *name, - struct ctl_table *table) -{ - table = proc_sys_lookup_table(dparent, table); - if (table) - table = proc_sys_lookup_table_one(table, name); - return table; -} -static struct ctl_table *do_proc_sys_lookup(struct dentry *parent, - struct qstr *name, - struct ctl_table_header **ptr) -{ - struct ctl_table_header *head; - struct ctl_table *table = NULL; + table = table ? table->child : head->ctl_table; - for (head = sysctl_head_next(NULL); head; - head = sysctl_head_next(head)) { - table = proc_sys_lookup_entry(parent, name, head->ctl_table); - if (table) - break; + p = find_in_table(table, name); + if (!p) { + for (h = sysctl_head_next(NULL); h; h = sysctl_head_next(h)) { + if (h->attached_to != table) + continue; + p = find_in_table(h->attached_by, name); + if (p) + break; + } } - *ptr = head; - return table; -} - -static struct dentry *proc_sys_lookup(struct inode *dir, struct dentry *dentry, - struct nameidata *nd) -{ - struct ctl_table_header *head; - struct inode *inode; - struct dentry *err; - struct ctl_table *table; - err = ERR_PTR(-ENOENT); - table = do_proc_sys_lookup(dentry->d_parent, &dentry->d_name, &head); - if (!table) + if (!p) goto out; err = ERR_PTR(-ENOMEM); - inode = proc_sys_make_inode(dir, table); + inode = proc_sys_make_inode(dir->i_sb, h ? h : head, p); + if (h) + sysctl_head_finish(h); + if (!inode) goto out; @@ -168,22 +129,14 @@ out: static ssize_t proc_sys_call_handler(struct file *filp, void __user *buf, size_t count, loff_t *ppos, int write) { - struct dentry *dentry = filp->f_dentry; - struct ctl_table_header *head; - struct ctl_table *table; + struct inode *inode = filp->f_path.dentry->d_inode; + struct ctl_table_header *head = grab_header(inode); + struct ctl_table *table = PROC_I(inode)->sysctl_entry; ssize_t error; size_t res; - table = do_proc_sys_lookup(dentry->d_parent, &dentry->d_name, &head); - /* Has the sysctl entry disappeared on us? */ - error = -ENOENT; - if (!table) - goto out; - - /* Has the sysctl entry been replaced by a directory? */ - error = -EISDIR; - if (!table->proc_handler) - goto out; + if (IS_ERR(head)) + return PTR_ERR(head); /* * At this point we know that the sysctl was not unregistered @@ -193,6 +146,11 @@ static ssize_t proc_sys_call_handler(struct file *filp, void __user *buf, if (sysctl_perm(head->root, table, write ? MAY_WRITE : MAY_READ)) goto out; + /* if that can happen at all, it should be -EINVAL, not -EISDIR */ + error = -EINVAL; + if (!table->proc_handler) + goto out; + /* careful: calling conventions are nasty here */ res = count; error = table->proc_handler(table, write, filp, buf, &res, ppos); @@ -218,82 +176,86 @@ static ssize_t proc_sys_write(struct file *filp, const char __user *buf, static int proc_sys_fill_cache(struct file *filp, void *dirent, - filldir_t filldir, struct ctl_table *table) + filldir_t filldir, + struct ctl_table_header *head, + struct ctl_table *table) { - struct ctl_table_header *head; - struct ctl_table *child_table = NULL; struct dentry *child, *dir = filp->f_path.dentry; struct inode *inode; struct qstr qname; ino_t ino = 0; unsigned type = DT_UNKNOWN; - int ret; qname.name = table->procname; qname.len = strlen(table->procname); qname.hash = full_name_hash(qname.name, qname.len); - /* Suppress duplicates. - * Only fill a directory entry if it is the value that - * an ordinary lookup of that name returns. Hide all - * others. - * - * If we ever cache this translation in the dcache - * I should do a dcache lookup first. But for now - * it is just simpler not to. - */ - ret = 0; - child_table = do_proc_sys_lookup(dir, &qname, &head); - sysctl_head_finish(head); - if (child_table != table) - return 0; - child = d_lookup(dir, &qname); if (!child) { - struct dentry *new; - new = d_alloc(dir, &qname); - if (new) { - inode = proc_sys_make_inode(dir->d_inode, table); - if (!inode) - child = ERR_PTR(-ENOMEM); - else { - new->d_op = &proc_sys_dentry_operations; - d_add(new, inode); + child = d_alloc(dir, &qname); + if (child) { + inode = proc_sys_make_inode(dir->d_sb, head, table); + if (!inode) { + dput(child); + return -ENOMEM; + } else { + child->d_op = &proc_sys_dentry_operations; + d_add(child, inode); } - if (child) - dput(new); - else - child = new; + } else { + return -ENOMEM; } } - if (!child || IS_ERR(child) || !child->d_inode) - goto end_instantiate; inode = child->d_inode; - if (inode) { - ino = inode->i_ino; - type = inode->i_mode >> 12; - } + ino = inode->i_ino; + type = inode->i_mode >> 12; dput(child); -end_instantiate: - if (!ino) - ino= find_inode_number(dir, &qname); - if (!ino) - ino = 1; - return filldir(dirent, qname.name, qname.len, filp->f_pos, ino, type); + return !!filldir(dirent, qname.name, qname.len, filp->f_pos, ino, type); +} + +static int scan(struct ctl_table_header *head, ctl_table *table, + unsigned long *pos, struct file *file, + void *dirent, filldir_t filldir) +{ + + for (; table->ctl_name || table->procname; table++, (*pos)++) { + int res; + + /* Can't do anything without a proc name */ + if (!table->procname) + continue; + + if (*pos < file->f_pos) + continue; + + res = proc_sys_fill_cache(file, dirent, filldir, head, table); + if (res) + return res; + + file->f_pos = *pos + 1; + } + return 0; } static int proc_sys_readdir(struct file *filp, void *dirent, filldir_t filldir) { - struct dentry *dentry = filp->f_dentry; + struct dentry *dentry = filp->f_path.dentry; struct inode *inode = dentry->d_inode; - struct ctl_table_header *head = NULL; - struct ctl_table *table; + struct ctl_table_header *head = grab_header(inode); + struct ctl_table *table = PROC_I(inode)->sysctl_entry; + struct ctl_table_header *h = NULL; unsigned long pos; - int ret; + int ret = -EINVAL; + + if (IS_ERR(head)) + return PTR_ERR(head); - ret = -ENOTDIR; - if (!S_ISDIR(inode->i_mode)) + if (table && !table->child) { + WARN_ON(1); goto out; + } + + table = table ? table->child : head->ctl_table; ret = 0; /* Avoid a switch here: arm builds fail with missing __cmpdi2 */ @@ -311,30 +273,17 @@ static int proc_sys_readdir(struct file *filp, void *dirent, filldir_t filldir) } pos = 2; - /* - Find each instance of the directory - * - Read all entries in each instance - * - Before returning an entry to user space lookup the entry - * by name and if I find a different entry don't return - * this one because it means it is a buried dup. - * For sysctl this should only happen for directory entries. - */ - for (head = sysctl_head_next(NULL); head; head = sysctl_head_next(head)) { - table = proc_sys_lookup_table(dentry, head->ctl_table); + ret = scan(head, table, &pos, filp, dirent, filldir); + if (ret) + goto out; - if (!table) + for (h = sysctl_head_next(NULL); h; h = sysctl_head_next(h)) { + if (h->attached_to != table) continue; - - for (; table->ctl_name || table->procname; table++, pos++) { - /* Can't do anything without a proc name */ - if (!table->procname) - continue; - - if (pos < filp->f_pos) - continue; - - if (proc_sys_fill_cache(filp, dirent, filldir, table) < 0) - goto out; - filp->f_pos = pos + 1; + ret = scan(h, h->attached_by, &pos, filp, dirent, filldir); + if (ret) { + sysctl_head_finish(h); + break; } } ret = 1; @@ -349,47 +298,18 @@ static int proc_sys_permission(struct inode *inode, int mask, struct nameidata * * sysctl entries that are not writeable, * are _NOT_ writeable, capabilities or not. */ - struct ctl_table_header *head; - struct ctl_table *table; - struct dentry *dentry; - int mode; - int depth; + struct ctl_table_header *head = grab_header(inode); + struct ctl_table *table = PROC_I(inode)->sysctl_entry; int error; - head = NULL; - depth = PROC_I(inode)->fd; - - /* First check the cached permissions, in case we don't have - * enough information to lookup the sysctl table entry. - */ - error = -EACCES; - mode = inode->i_mode; - - if (current->euid == 0) - mode >>= 6; - else if (in_group_p(0)) - mode >>= 3; - - if ((mode & mask & (MAY_READ|MAY_WRITE|MAY_EXEC)) == mask) - error = 0; - - /* If we can't get a sysctl table entry the permission - * checks on the cached mode will have to be enough. - */ - if (!nd || !depth) - goto out; + if (IS_ERR(head)) + return PTR_ERR(head); - dentry = nd->path.dentry; - table = do_proc_sys_lookup(dentry->d_parent, &dentry->d_name, &head); + if (!table) /* global root - r-xr-xr-x */ + error = mask & MAY_WRITE ? -EACCES : 0; + else /* Use the permissions on the sysctl table entry */ + error = sysctl_perm(head->root, table, mask); - /* If the entry does not exist deny permission */ - error = -EACCES; - if (!table) - goto out; - - /* Use the permissions on the sysctl table entry */ - error = sysctl_perm(head->root, table, mask); -out: sysctl_head_finish(head); return error; } @@ -409,33 +329,70 @@ static int proc_sys_setattr(struct dentry *dentry, struct iattr *attr) return error; } -/* I'm lazy and don't distinguish between files and directories, - * until access time. - */ +static int proc_sys_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat) +{ + struct inode *inode = dentry->d_inode; + struct ctl_table_header *head = grab_header(inode); + struct ctl_table *table = PROC_I(inode)->sysctl_entry; + + if (IS_ERR(head)) + return PTR_ERR(head); + + generic_fillattr(inode, stat); + if (table) + stat->mode = (stat->mode & S_IFMT) | table->mode; + + sysctl_head_finish(head); + return 0; +} + static const struct file_operations proc_sys_file_operations = { .read = proc_sys_read, .write = proc_sys_write, +}; + +static const struct file_operations proc_sys_dir_file_operations = { .readdir = proc_sys_readdir, }; static const struct inode_operations proc_sys_inode_operations = { + .permission = proc_sys_permission, + .setattr = proc_sys_setattr, + .getattr = proc_sys_getattr, +}; + +static const struct inode_operations proc_sys_dir_operations = { .lookup = proc_sys_lookup, .permission = proc_sys_permission, .setattr = proc_sys_setattr, + .getattr = proc_sys_getattr, }; static int proc_sys_revalidate(struct dentry *dentry, struct nameidata *nd) { - struct ctl_table_header *head; - struct ctl_table *table; - table = do_proc_sys_lookup(dentry->d_parent, &dentry->d_name, &head); - proc_sys_refresh_inode(dentry->d_inode, table); - sysctl_head_finish(head); - return !!table; + return !PROC_I(dentry->d_inode)->sysctl->unregistering; +} + +static int proc_sys_delete(struct dentry *dentry) +{ + return !!PROC_I(dentry->d_inode)->sysctl->unregistering; +} + +static int proc_sys_compare(struct dentry *dir, struct qstr *qstr, + struct qstr *name) +{ + struct dentry *dentry = container_of(qstr, struct dentry, d_name); + if (qstr->len != name->len) + return 1; + if (memcmp(qstr->name, name->name, name->len)) + return 1; + return !sysctl_is_seen(PROC_I(dentry->d_inode)->sysctl); } static struct dentry_operations proc_sys_dentry_operations = { .d_revalidate = proc_sys_revalidate, + .d_delete = proc_sys_delete, + .d_compare = proc_sys_compare, }; static struct proc_dir_entry *proc_sys_root; @@ -443,8 +400,8 @@ static struct proc_dir_entry *proc_sys_root; int proc_sys_init(void) { proc_sys_root = proc_mkdir("sys", NULL); - proc_sys_root->proc_iops = &proc_sys_inode_operations; - proc_sys_root->proc_fops = &proc_sys_file_operations; + proc_sys_root->proc_iops = &proc_sys_dir_operations; + proc_sys_root->proc_fops = &proc_sys_dir_file_operations; proc_sys_root->nlink = 0; return 0; } diff --git a/include/linux/proc_fs.h b/include/linux/proc_fs.h index f560d1705af..fb61850d1cf 100644 --- a/include/linux/proc_fs.h +++ b/include/linux/proc_fs.h @@ -282,11 +282,16 @@ union proc_op { struct task_struct *task); }; +struct ctl_table_header; +struct ctl_table; + struct proc_inode { struct pid *pid; int fd; union proc_op op; struct proc_dir_entry *pde; + struct ctl_table_header *sysctl; + struct ctl_table *sysctl_entry; struct inode vfs_inode; }; diff --git a/include/linux/sysctl.h b/include/linux/sysctl.h index 3f6599aeb0d..d0437f36921 100644 --- a/include/linux/sysctl.h +++ b/include/linux/sysctl.h @@ -961,6 +961,7 @@ struct ctl_table_header; extern void sysctl_head_get(struct ctl_table_header *); extern void sysctl_head_put(struct ctl_table_header *); +extern int sysctl_is_seen(struct ctl_table_header *); extern struct ctl_table_header *sysctl_head_grab(struct ctl_table_header *); extern struct ctl_table_header *sysctl_head_next(struct ctl_table_header *prev); extern struct ctl_table_header *__sysctl_head_next(struct nsproxy *namespaces, diff --git a/kernel/sysctl.c b/kernel/sysctl.c index c9a0af88703..ff5abcca5dd 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -1932,6 +1932,21 @@ void unregister_sysctl_table(struct ctl_table_header * header) spin_unlock(&sysctl_lock); } +int sysctl_is_seen(struct ctl_table_header *p) +{ + struct ctl_table_set *set = p->set; + int res; + spin_lock(&sysctl_lock); + if (p->unregistering) + res = 0; + else if (!set->is_seen) + res = 1; + else + res = set->is_seen(set); + spin_unlock(&sysctl_lock); + return res; +} + void setup_sysctl_set(struct ctl_table_set *p, struct ctl_table_set *parent, int (*is_seen)(struct ctl_table_set *)) -- cgit v1.2.3-70-g09d2 From 1bd5191d9f5d1928c4efdf604c4164b04bb88dbe Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Wed, 21 May 2008 19:15:03 +0200 Subject: [patch 05/14] hpfs: dont call permission() hpfs_unlink() calls permission() prior to truncating the file. HPFS doesn't define a .permission method, so replace with explicit call to generic_permission(). This is equivalent, except that devcgroup_inode_permission() and security_inode_permission() are not called. The truncation is just an implementation detail of the unlink, so these security checks are unnecessary. I suspect that even calling generic_permission() is unnecessary, since we shouldn't mind if the file isn't writable. But I leave that to the maintainer to decide. Signed-off-by: Miklos Szeredi CC: Mikulas Patocka --- fs/hpfs/namei.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/hpfs/namei.c b/fs/hpfs/namei.c index d256559b410..d9c59a77544 100644 --- a/fs/hpfs/namei.c +++ b/fs/hpfs/namei.c @@ -415,7 +415,7 @@ again: d_drop(dentry); spin_lock(&dentry->d_lock); if (atomic_read(&dentry->d_count) > 1 || - permission(inode, MAY_WRITE, NULL) || + generic_permission(inode, MAY_WRITE, NULL) || !S_ISREG(inode->i_mode) || get_write_access(inode)) { spin_unlock(&dentry->d_lock); -- cgit v1.2.3-70-g09d2 From e6305c43eda10ebfd2ad9e35d6e172ccc7bb3695 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 15 Jul 2008 21:03:57 -0400 Subject: [PATCH] sanitize ->permission() prototype * kill nameidata * argument; map the 3 bits in ->flags anybody cares about to new MAY_... ones and pass with the mask. * kill redundant gfs2_iop_permission() * sanitize ecryptfs_permission() * fix remaining places where ->permission() instances might barf on new MAY_... found in mask. The obvious next target in that direction is permission(9) folded fix for nfs_permission() breakage from Miklos Szeredi Signed-off-by: Al Viro --- fs/afs/internal.h | 4 +--- fs/afs/security.c | 2 +- fs/bad_inode.c | 3 +-- fs/cifs/cifsfs.c | 2 +- fs/coda/dir.c | 4 +++- fs/coda/pioctl.c | 6 ++---- fs/ecryptfs/inode.c | 17 ++--------------- 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/fuse/dir.c | 6 +++--- fs/gfs2/ops_inode.c | 12 +++--------- fs/hfs/inode.c | 3 +-- fs/hfsplus/inode.c | 2 +- fs/hostfs/hostfs_kern.c | 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 | 23 +++++++++++++++++------ fs/nfs/dir.c | 11 +++++------ fs/ocfs2/file.c | 2 +- fs/ocfs2/file.h | 3 +-- fs/proc/base.c | 3 +-- fs/proc/proc_sysctl.c | 2 +- fs/reiserfs/xattr.c | 2 +- fs/smbfs/file.c | 4 ++-- fs/xfs/linux-2.6/xfs_iops.c | 3 +-- include/linux/coda_linux.h | 2 +- include/linux/fs.h | 5 ++++- include/linux/nfs_fs.h | 2 +- include/linux/reiserfs_xattr.h | 2 +- include/linux/shmem_fs.h | 2 +- kernel/sysctl.c | 10 +++++----- mm/shmem_acl.c | 2 +- 38 files changed, 74 insertions(+), 87 deletions(-) (limited to 'fs') diff --git a/fs/afs/internal.h b/fs/afs/internal.h index 7102824ba84..3cb6920ff30 100644 --- a/fs/afs/internal.h +++ b/fs/afs/internal.h @@ -469,8 +469,6 @@ extern bool afs_cm_incoming_call(struct afs_call *); extern const struct inode_operations afs_dir_inode_operations; extern const struct file_operations afs_dir_file_operations; -extern int afs_permission(struct inode *, int, struct nameidata *); - /* * file.c */ @@ -605,7 +603,7 @@ 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, struct nameidata *); +extern int afs_permission(struct inode *, int); /* * server.c diff --git a/fs/afs/security.c b/fs/afs/security.c index 3bcbeceba1b..3ef50437003 100644 --- a/fs/afs/security.c +++ b/fs/afs/security.c @@ -284,7 +284,7 @@ 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, struct nameidata *nd) +int afs_permission(struct inode *inode, int mask) { struct afs_vnode *vnode = AFS_FS_I(inode); afs_access_t uninitialized_var(access); diff --git a/fs/bad_inode.c b/fs/bad_inode.c index f1c2ea8342f..5f1538c03b1 100644 --- a/fs/bad_inode.c +++ b/fs/bad_inode.c @@ -243,8 +243,7 @@ static int bad_inode_readlink(struct dentry *dentry, char __user *buffer, return -EIO; } -static int bad_inode_permission(struct inode *inode, int mask, - struct nameidata *nd) +static int bad_inode_permission(struct inode *inode, int mask) { return -EIO; } diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c index fe5f6809cba..1ec7076f7b2 100644 --- a/fs/cifs/cifsfs.c +++ b/fs/cifs/cifsfs.c @@ -267,7 +267,7 @@ cifs_statfs(struct dentry *dentry, struct kstatfs *buf) return 0; } -static int cifs_permission(struct inode *inode, int mask, struct nameidata *nd) +static int cifs_permission(struct inode *inode, int mask) { struct cifs_sb_info *cifs_sb; diff --git a/fs/coda/dir.c b/fs/coda/dir.c index 3d2580e00a3..c5916228243 100644 --- a/fs/coda/dir.c +++ b/fs/coda/dir.c @@ -137,9 +137,11 @@ exit: } -int coda_permission(struct inode *inode, int mask, struct nameidata *nd) +int coda_permission(struct inode *inode, int mask) { int error = 0; + + mask &= MAY_READ | MAY_WRITE | MAY_EXEC; if (!mask) return 0; diff --git a/fs/coda/pioctl.c b/fs/coda/pioctl.c index c21a1f552a6..c38a98974fb 100644 --- a/fs/coda/pioctl.c +++ b/fs/coda/pioctl.c @@ -24,8 +24,7 @@ #include /* pioctl ops */ -static int coda_ioctl_permission(struct inode *inode, int mask, - struct nameidata *nd); +static int coda_ioctl_permission(struct inode *inode, int mask); static int coda_pioctl(struct inode * inode, struct file * filp, unsigned int cmd, unsigned long user_data); @@ -42,8 +41,7 @@ const struct file_operations coda_ioctl_operations = { }; /* the coda pioctl inode ops */ -static int coda_ioctl_permission(struct inode *inode, int mask, - struct nameidata *nd) +static int coda_ioctl_permission(struct inode *inode, int mask) { return 0; } diff --git a/fs/ecryptfs/inode.c b/fs/ecryptfs/inode.c index d755455e3bf..32f4228efcd 100644 --- a/fs/ecryptfs/inode.c +++ b/fs/ecryptfs/inode.c @@ -830,22 +830,9 @@ out: } static int -ecryptfs_permission(struct inode *inode, int mask, struct nameidata *nd) +ecryptfs_permission(struct inode *inode, int mask) { - int rc; - - if (nd) { - struct vfsmount *vfsmnt_save = nd->path.mnt; - struct dentry *dentry_save = nd->path.dentry; - - nd->path.mnt = ecryptfs_dentry_to_lower_mnt(nd->path.dentry); - nd->path.dentry = ecryptfs_dentry_to_lower(nd->path.dentry); - rc = permission(ecryptfs_inode_to_lower(inode), mask, nd); - nd->path.mnt = vfsmnt_save; - nd->path.dentry = dentry_save; - } else - rc = permission(ecryptfs_inode_to_lower(inode), mask, NULL); - return rc; + return permission(ecryptfs_inode_to_lower(inode), mask, NULL); } /** diff --git a/fs/ext2/acl.c b/fs/ext2/acl.c index e58669e1b87..ae8c4f850b2 100644 --- a/fs/ext2/acl.c +++ b/fs/ext2/acl.c @@ -294,7 +294,7 @@ ext2_check_acl(struct inode *inode, int mask) } int -ext2_permission(struct inode *inode, int mask, struct nameidata *nd) +ext2_permission(struct inode *inode, int mask) { return generic_permission(inode, mask, ext2_check_acl); } diff --git a/fs/ext2/acl.h b/fs/ext2/acl.h index 0bde85bafe3..b42cf578554 100644 --- a/fs/ext2/acl.h +++ b/fs/ext2/acl.h @@ -58,7 +58,7 @@ static inline int ext2_acl_count(size_t size) #define EXT2_ACL_NOT_CACHED ((void *)-1) /* acl.c */ -extern int ext2_permission (struct inode *, int, struct nameidata *); +extern int ext2_permission (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 a754d184817..b60bb241880 100644 --- a/fs/ext3/acl.c +++ b/fs/ext3/acl.c @@ -299,7 +299,7 @@ ext3_check_acl(struct inode *inode, int mask) } int -ext3_permission(struct inode *inode, int mask, struct nameidata *nd) +ext3_permission(struct inode *inode, int mask) { return generic_permission(inode, mask, ext3_check_acl); } diff --git a/fs/ext3/acl.h b/fs/ext3/acl.h index 0d1e6279cbf..42da16b8cac 100644 --- a/fs/ext3/acl.h +++ b/fs/ext3/acl.h @@ -58,7 +58,7 @@ static inline int ext3_acl_count(size_t size) #define EXT3_ACL_NOT_CACHED ((void *)-1) /* acl.c */ -extern int ext3_permission (struct inode *, int, struct nameidata *); +extern int ext3_permission (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 3c8dab880d9..c7d04e16544 100644 --- a/fs/ext4/acl.c +++ b/fs/ext4/acl.c @@ -299,7 +299,7 @@ ext4_check_acl(struct inode *inode, int mask) } int -ext4_permission(struct inode *inode, int mask, struct nameidata *nd) +ext4_permission(struct inode *inode, int mask) { return generic_permission(inode, mask, ext4_check_acl); } diff --git a/fs/ext4/acl.h b/fs/ext4/acl.h index 26a5c1abf14..cd2b855a07d 100644 --- a/fs/ext4/acl.h +++ b/fs/ext4/acl.h @@ -58,7 +58,7 @@ static inline int ext4_acl_count(size_t size) #define EXT4_ACL_NOT_CACHED ((void *)-1) /* acl.c */ -extern int ext4_permission (struct inode *, int, struct nameidata *); +extern int ext4_permission (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/fuse/dir.c b/fs/fuse/dir.c index 51d0035ff07..48a7934cb95 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -898,7 +898,7 @@ static int fuse_access(struct inode *inode, int mask) return PTR_ERR(req); memset(&inarg, 0, sizeof(inarg)); - inarg.mask = mask; + inarg.mask = mask & (MAY_READ | MAY_WRITE | MAY_EXEC); req->in.h.opcode = FUSE_ACCESS; req->in.h.nodeid = get_node_id(inode); req->in.numargs = 1; @@ -927,7 +927,7 @@ static int fuse_access(struct inode *inode, int mask) * access request is sent. Execute permission is still checked * locally based on file mode. */ -static int fuse_permission(struct inode *inode, int mask, struct nameidata *nd) +static int fuse_permission(struct inode *inode, int mask) { struct fuse_conn *fc = get_fuse_conn(inode); bool refreshed = false; @@ -962,7 +962,7 @@ static int fuse_permission(struct inode *inode, int mask, struct nameidata *nd) exist. So if permissions are revoked this won't be noticed immediately, only after the attribute timeout has expired */ - } else if (nd && (nd->flags & (LOOKUP_ACCESS | LOOKUP_CHDIR))) { + } else if (mask & (MAY_ACCESS | MAY_CHDIR)) { err = fuse_access(inode, mask); } else if ((mask & MAY_EXEC) && S_ISREG(inode->i_mode)) { if (!(inode->i_mode & S_IXUGO)) { diff --git a/fs/gfs2/ops_inode.c b/fs/gfs2/ops_inode.c index 1e252dfc529..4e982532f08 100644 --- a/fs/gfs2/ops_inode.c +++ b/fs/gfs2/ops_inode.c @@ -915,12 +915,6 @@ int gfs2_permission(struct inode *inode, int mask) return error; } -static int gfs2_iop_permission(struct inode *inode, int mask, - struct nameidata *nd) -{ - return gfs2_permission(inode, mask); -} - static int setattr_size(struct inode *inode, struct iattr *attr) { struct gfs2_inode *ip = GFS2_I(inode); @@ -1150,7 +1144,7 @@ static int gfs2_removexattr(struct dentry *dentry, const char *name) } const struct inode_operations gfs2_file_iops = { - .permission = gfs2_iop_permission, + .permission = gfs2_permission, .setattr = gfs2_setattr, .getattr = gfs2_getattr, .setxattr = gfs2_setxattr, @@ -1169,7 +1163,7 @@ const struct inode_operations gfs2_dir_iops = { .rmdir = gfs2_rmdir, .mknod = gfs2_mknod, .rename = gfs2_rename, - .permission = gfs2_iop_permission, + .permission = gfs2_permission, .setattr = gfs2_setattr, .getattr = gfs2_getattr, .setxattr = gfs2_setxattr, @@ -1181,7 +1175,7 @@ const struct inode_operations gfs2_dir_iops = { const struct inode_operations gfs2_symlink_iops = { .readlink = gfs2_readlink, .follow_link = gfs2_follow_link, - .permission = gfs2_iop_permission, + .permission = gfs2_permission, .setattr = gfs2_setattr, .getattr = gfs2_getattr, .setxattr = gfs2_setxattr, diff --git a/fs/hfs/inode.c b/fs/hfs/inode.c index dc4ec640e87..aa73f3fd5dd 100644 --- a/fs/hfs/inode.c +++ b/fs/hfs/inode.c @@ -511,8 +511,7 @@ void hfs_clear_inode(struct inode *inode) } } -static int hfs_permission(struct inode *inode, int mask, - struct nameidata *nd) +static int hfs_permission(struct inode *inode, int mask) { if (S_ISREG(inode->i_mode) && mask & MAY_EXEC) return 0; diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c index cc3b5e24339..d4014e3044d 100644 --- a/fs/hfsplus/inode.c +++ b/fs/hfsplus/inode.c @@ -238,7 +238,7 @@ static void hfsplus_set_perms(struct inode *inode, struct hfsplus_perm *perms) perms->dev = cpu_to_be32(HFSPLUS_I(inode).dev); } -static int hfsplus_permission(struct inode *inode, int mask, struct nameidata *nd) +static int hfsplus_permission(struct inode *inode, int mask) { /* MAY_EXEC is also used for lookup, if no x bit is set allow lookup, * open_exec has the same test, so it's still not executable, if a x bit diff --git a/fs/hostfs/hostfs_kern.c b/fs/hostfs/hostfs_kern.c index 5222345ddcc..d6ecabf4d23 100644 --- a/fs/hostfs/hostfs_kern.c +++ b/fs/hostfs/hostfs_kern.c @@ -822,7 +822,7 @@ int hostfs_rename(struct inode *from_ino, struct dentry *from, return err; } -int hostfs_permission(struct inode *ino, int desired, struct nameidata *nd) +int hostfs_permission(struct inode *ino, int desired) { char *name; int r = 0, w = 0, x = 0, err; diff --git a/fs/jffs2/acl.c b/fs/jffs2/acl.c index 4c80404a9ab..d98713777a1 100644 --- a/fs/jffs2/acl.c +++ b/fs/jffs2/acl.c @@ -314,7 +314,7 @@ static int jffs2_check_acl(struct inode *inode, int mask) return -EAGAIN; } -int jffs2_permission(struct inode *inode, int mask, struct nameidata *nd) +int jffs2_permission(struct inode *inode, int mask) { return generic_permission(inode, mask, jffs2_check_acl); } diff --git a/fs/jffs2/acl.h b/fs/jffs2/acl.h index 0bb7f003fd8..8ca058aed38 100644 --- a/fs/jffs2/acl.h +++ b/fs/jffs2/acl.h @@ -28,7 +28,7 @@ struct jffs2_acl_header { #define JFFS2_ACL_NOT_CACHED ((void *)-1) -extern int jffs2_permission(struct inode *, int, struct nameidata *); +extern int jffs2_permission(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 4d84bdc8829..d3e5c33665d 100644 --- a/fs/jfs/acl.c +++ b/fs/jfs/acl.c @@ -140,7 +140,7 @@ static int jfs_check_acl(struct inode *inode, int mask) return -EAGAIN; } -int jfs_permission(struct inode *inode, int mask, struct nameidata *nd) +int jfs_permission(struct inode *inode, int mask) { return generic_permission(inode, mask, jfs_check_acl); } diff --git a/fs/jfs/jfs_acl.h b/fs/jfs/jfs_acl.h index 455fa429204..88475f10a38 100644 --- a/fs/jfs/jfs_acl.h +++ b/fs/jfs/jfs_acl.h @@ -20,7 +20,7 @@ #ifdef CONFIG_JFS_POSIX_ACL -int jfs_permission(struct inode *, int, struct nameidata *); +int jfs_permission(struct inode *, int); int jfs_init_acl(tid_t, struct inode *, struct inode *); int jfs_setattr(struct dentry *, struct iattr *); diff --git a/fs/namei.c b/fs/namei.c index 3b26a240ade..46af98ed136 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -185,6 +185,8 @@ int generic_permission(struct inode *inode, int mask, { umode_t mode = inode->i_mode; + mask &= MAY_READ | MAY_WRITE | MAY_EXEC; + if (current->fsuid == inode->i_uid) mode >>= 6; else { @@ -203,7 +205,7 @@ int generic_permission(struct inode *inode, int mask, /* * If the DACs are ok we don't need any capability check. */ - if (((mode & mask & (MAY_READ|MAY_WRITE|MAY_EXEC)) == mask)) + if ((mask & ~mode) == 0) return 0; check_capabilities: @@ -228,7 +230,7 @@ int generic_permission(struct inode *inode, int mask, int permission(struct inode *inode, int mask, struct nameidata *nd) { - int retval, submask; + int retval; struct vfsmount *mnt = NULL; if (nd) @@ -261,9 +263,17 @@ int permission(struct inode *inode, int mask, struct nameidata *nd) } /* Ordinary permission routines do not understand MAY_APPEND. */ - submask = mask & ~MAY_APPEND; if (inode->i_op && inode->i_op->permission) { - retval = inode->i_op->permission(inode, submask, nd); + int extra = 0; + if (nd) { + if (nd->flags & LOOKUP_ACCESS) + extra |= MAY_ACCESS; + if (nd->flags & LOOKUP_CHDIR) + extra |= MAY_CHDIR; + if (nd->flags & LOOKUP_OPEN) + extra |= MAY_OPEN; + } + retval = inode->i_op->permission(inode, mask | extra); if (!retval) { /* * Exec permission on a regular file is denied if none @@ -277,7 +287,7 @@ int permission(struct inode *inode, int mask, struct nameidata *nd) return -EACCES; } } else { - retval = generic_permission(inode, submask, NULL); + retval = generic_permission(inode, mask, NULL); } if (retval) return retval; @@ -286,7 +296,8 @@ int permission(struct inode *inode, int mask, struct nameidata *nd) if (retval) return retval; - return security_inode_permission(inode, mask, nd); + return security_inode_permission(inode, + mask & (MAY_READ|MAY_WRITE|MAY_EXEC), nd); } /** diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index 28a238dab23..74f92b717f7 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -1884,7 +1884,7 @@ static int nfs_do_access(struct inode *inode, struct rpc_cred *cred, int mask) return status; nfs_access_add_cache(inode, &cache); out: - if ((cache.mask & mask) == mask) + if ((mask & ~cache.mask & (MAY_READ | MAY_WRITE | MAY_EXEC)) == 0) return 0; return -EACCES; } @@ -1907,17 +1907,17 @@ 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, struct nameidata *nd) +int nfs_permission(struct inode *inode, int mask) { struct rpc_cred *cred; int res = 0; nfs_inc_stats(inode, NFSIOS_VFSACCESS); - if (mask == 0) + if ((mask & (MAY_READ | MAY_WRITE | MAY_EXEC)) == 0) goto out; /* Is this sys_access() ? */ - if (nd != NULL && (nd->flags & LOOKUP_ACCESS)) + if (mask & MAY_ACCESS) goto force_lookup; switch (inode->i_mode & S_IFMT) { @@ -1926,8 +1926,7 @@ int nfs_permission(struct inode *inode, int mask, struct nameidata *nd) case S_IFREG: /* NFSv4 has atomic_open... */ if (nfs_server_capable(inode, NFS_CAP_ATOMIC_OPEN) - && nd != NULL - && (nd->flags & LOOKUP_OPEN)) + && (mask & MAY_OPEN)) goto out; break; case S_IFDIR: diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index e8514e8b6ce..be2dd95d3a1 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c @@ -1176,7 +1176,7 @@ bail: return err; } -int ocfs2_permission(struct inode *inode, int mask, struct nameidata *nd) +int ocfs2_permission(struct inode *inode, int mask) { int ret; diff --git a/fs/ocfs2/file.h b/fs/ocfs2/file.h index 048ddcaf5c8..1e27b4d017e 100644 --- a/fs/ocfs2/file.h +++ b/fs/ocfs2/file.h @@ -62,8 +62,7 @@ int ocfs2_lock_allocators(struct inode *inode, struct ocfs2_dinode *di, 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, - struct nameidata *nd); +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 81bce6791bf..d82d800389f 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -1859,8 +1859,7 @@ 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, - struct nameidata *nd) +static int proc_fd_permission(struct inode *inode, int mask) { int rv; diff --git a/fs/proc/proc_sysctl.c b/fs/proc/proc_sysctl.c index fa1ec2433e4..f9a8b892718 100644 --- a/fs/proc/proc_sysctl.c +++ b/fs/proc/proc_sysctl.c @@ -292,7 +292,7 @@ out: return ret; } -static int proc_sys_permission(struct inode *inode, int mask, struct nameidata *nd) +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 d7c4935c103..bb3cb5b7cdb 100644 --- a/fs/reiserfs/xattr.c +++ b/fs/reiserfs/xattr.c @@ -1250,7 +1250,7 @@ static int reiserfs_check_acl(struct inode *inode, int mask) return error; } -int reiserfs_permission(struct inode *inode, int mask, struct nameidata *nd) +int reiserfs_permission(struct inode *inode, int mask) { /* * We don't do permission checks on the internal objects. diff --git a/fs/smbfs/file.c b/fs/smbfs/file.c index 2294783320c..e4f8d51a555 100644 --- a/fs/smbfs/file.c +++ b/fs/smbfs/file.c @@ -408,7 +408,7 @@ smb_file_release(struct inode *inode, struct file * file) * privileges, so we need our own check for this. */ static int -smb_file_permission(struct inode *inode, int mask, struct nameidata *nd) +smb_file_permission(struct inode *inode, int mask) { int mode = inode->i_mode; int error = 0; @@ -417,7 +417,7 @@ smb_file_permission(struct inode *inode, int mask, struct nameidata *nd) /* Look at user permissions */ mode >>= 6; - if ((mode & 7 & mask) != mask) + if (mask & ~mode & (MAY_READ | MAY_WRITE | MAY_EXEC)) error = -EACCES; return error; } diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c index 2bf287ef548..5fc61c824bb 100644 --- a/fs/xfs/linux-2.6/xfs_iops.c +++ b/fs/xfs/linux-2.6/xfs_iops.c @@ -589,8 +589,7 @@ xfs_check_acl( STATIC int xfs_vn_permission( struct inode *inode, - int mask, - struct nameidata *nd) + int mask) { return generic_permission(inode, mask, xfs_check_acl); } diff --git a/include/linux/coda_linux.h b/include/linux/coda_linux.h index 31b75311e2c..dcc228aa335 100644 --- a/include/linux/coda_linux.h +++ b/include/linux/coda_linux.h @@ -37,7 +37,7 @@ 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, struct nameidata *nd); +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/include/linux/fs.h b/include/linux/fs.h index 7721a2ac9c0..6c923c9b79b 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -60,6 +60,9 @@ extern int dir_notify_enable; #define MAY_WRITE 2 #define MAY_READ 4 #define MAY_APPEND 8 +#define MAY_ACCESS 16 +#define MAY_CHDIR 32 +#define MAY_OPEN 64 #define FMODE_READ 1 #define FMODE_WRITE 2 @@ -1272,7 +1275,7 @@ 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, struct nameidata *); + int (*permission) (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/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index 29d26191873..f08f9ca602a 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -332,7 +332,7 @@ 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, struct nameidata *); +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 66a96814d61..af135ae895d 100644 --- a/include/linux/reiserfs_xattr.h +++ b/include/linux/reiserfs_xattr.h @@ -55,7 +55,7 @@ int reiserfs_removexattr(struct dentry *dentry, const char *name); int reiserfs_delete_xattrs(struct inode *inode); int reiserfs_chown_xattrs(struct inode *inode, struct iattr *attrs); int reiserfs_xattr_init(struct super_block *sb, int mount_flags); -int reiserfs_permission(struct inode *inode, int mask, struct nameidata *nd); +int reiserfs_permission(struct inode *inode, int mask); int reiserfs_xattr_del(struct inode *, const char *); int reiserfs_xattr_get(const struct inode *, const char *, void *, size_t); diff --git a/include/linux/shmem_fs.h b/include/linux/shmem_fs.h index f2d12d5a21b..fd83f2584b1 100644 --- a/include/linux/shmem_fs.h +++ b/include/linux/shmem_fs.h @@ -43,7 +43,7 @@ static inline struct shmem_inode_info *SHMEM_I(struct inode *inode) } #ifdef CONFIG_TMPFS_POSIX_ACL -int shmem_permission(struct inode *, int, struct nameidata *); +int shmem_permission(struct inode *, int); int shmem_acl_init(struct inode *, struct inode *); void shmem_acl_destroy_inode(struct inode *); diff --git a/kernel/sysctl.c b/kernel/sysctl.c index ff5abcca5dd..911d846f050 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -1516,9 +1516,9 @@ static int do_sysctl_strategy(struct ctl_table_root *root, int op = 0, rc; if (oldval) - op |= 004; + op |= MAY_READ; if (newval) - op |= 002; + op |= MAY_WRITE; if (sysctl_perm(root, table, op)) return -EPERM; @@ -1560,7 +1560,7 @@ repeat: if (n == table->ctl_name) { int error; if (table->child) { - if (sysctl_perm(root, table, 001)) + if (sysctl_perm(root, table, MAY_EXEC)) return -EPERM; name++; nlen--; @@ -1635,7 +1635,7 @@ static int test_perm(int mode, int op) mode >>= 6; else if (in_egroup_p(0)) mode >>= 3; - if ((mode & op & 0007) == op) + if ((op & ~mode & (MAY_READ|MAY_WRITE|MAY_EXEC)) == 0) return 0; return -EACCES; } @@ -1645,7 +1645,7 @@ int sysctl_perm(struct ctl_table_root *root, struct ctl_table *table, int op) int error; int mode; - error = security_sysctl(table, op); + error = security_sysctl(table, op & (MAY_READ | MAY_WRITE | MAY_EXEC)); if (error) return error; diff --git a/mm/shmem_acl.c b/mm/shmem_acl.c index f5664c5b9eb..8e5aadd7dcd 100644 --- a/mm/shmem_acl.c +++ b/mm/shmem_acl.c @@ -191,7 +191,7 @@ shmem_check_acl(struct inode *inode, int mask) * shmem_permission - permission() inode operation */ int -shmem_permission(struct inode *inode, int mask, struct nameidata *nd) +shmem_permission(struct inode *inode, int mask) { return generic_permission(inode, mask, shmem_check_acl); } -- cgit v1.2.3-70-g09d2 From c82e42da8a6b2f3a85dc4d4278cb8238702f8f64 Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Tue, 24 Jun 2008 16:50:12 +0200 Subject: [patch 1/5] vfs: truncate: dont check immutable twice vfs_permission(MAY_WRITE) already checked for the inode being immutable, so no need to repeat it. Signed-off-by: Miklos Szeredi Acked-by: Christoph Hellwig --- fs/open.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/open.c b/fs/open.c index bb98d2fe809..b2e4c93aed0 100644 --- a/fs/open.c +++ b/fs/open.c @@ -256,7 +256,7 @@ static long do_sys_truncate(const char __user * path, loff_t length) goto mnt_drop_write_and_out; error = -EPERM; - if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) + if (IS_APPEND(inode)) goto mnt_drop_write_and_out; error = get_write_access(inode); -- cgit v1.2.3-70-g09d2 From 2f1936b87783a3a56c9441b27b9ba7a747f11e8e Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Tue, 24 Jun 2008 16:50:14 +0200 Subject: [patch 3/5] vfs: change remove_suid() to file_remove_suid() All calls to remove_suid() are made with a file pointer, because (similarly to file_update_time) it is called when the file is written. Clean up callers by passing in a file instead of a dentry. Signed-off-by: Miklos Szeredi --- fs/fuse/file.c | 2 +- fs/ntfs/file.c | 2 +- fs/splice.c | 4 ++-- fs/xfs/linux-2.6/xfs_lrw.c | 2 +- include/linux/fs.h | 2 +- mm/filemap.c | 7 ++++--- mm/filemap_xip.c | 2 +- 7 files changed, 11 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/fuse/file.c b/fs/fuse/file.c index 67ff2c6a8f6..2bada6bbc31 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -893,7 +893,7 @@ static ssize_t fuse_file_aio_write(struct kiocb *iocb, const struct iovec *iov, if (count == 0) goto out; - err = remove_suid(file->f_path.dentry); + err = file_remove_suid(file); if (err) goto out; diff --git a/fs/ntfs/file.c b/fs/ntfs/file.c index 3c5550cd11d..d020866d423 100644 --- a/fs/ntfs/file.c +++ b/fs/ntfs/file.c @@ -2118,7 +2118,7 @@ static ssize_t ntfs_file_aio_write_nolock(struct kiocb *iocb, goto out; if (!count) goto out; - err = remove_suid(file->f_path.dentry); + err = file_remove_suid(file); if (err) goto out; file_update_time(file); diff --git a/fs/splice.c b/fs/splice.c index 47dc1a445d1..b30311ba8af 100644 --- a/fs/splice.c +++ b/fs/splice.c @@ -772,7 +772,7 @@ generic_file_splice_write_nolock(struct pipe_inode_info *pipe, struct file *out, ssize_t ret; int err; - err = remove_suid(out->f_path.dentry); + err = file_remove_suid(out); if (unlikely(err)) return err; @@ -830,7 +830,7 @@ generic_file_splice_write(struct pipe_inode_info *pipe, struct file *out, ssize_t ret; inode_double_lock(inode, pipe->inode); - ret = remove_suid(out->f_path.dentry); + ret = file_remove_suid(out); if (likely(!ret)) ret = __splice_from_pipe(pipe, &sd, pipe_to_file); inode_double_unlock(inode, pipe->inode); diff --git a/fs/xfs/linux-2.6/xfs_lrw.c b/fs/xfs/linux-2.6/xfs_lrw.c index 5e3b57516ec..82333b3e118 100644 --- a/fs/xfs/linux-2.6/xfs_lrw.c +++ b/fs/xfs/linux-2.6/xfs_lrw.c @@ -711,7 +711,7 @@ start: !capable(CAP_FSETID)) { error = xfs_write_clear_setuid(xip); if (likely(!error)) - error = -remove_suid(file->f_path.dentry); + error = -file_remove_suid(file); if (unlikely(error)) { goto out_unlock_internal; } diff --git a/include/linux/fs.h b/include/linux/fs.h index 6c923c9b79b..1a3546e69f9 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1834,7 +1834,7 @@ extern void clear_inode(struct inode *); extern void destroy_inode(struct inode *); extern struct inode *new_inode(struct super_block *); extern int should_remove_suid(struct dentry *); -extern int remove_suid(struct dentry *); +extern int file_remove_suid(struct file *); extern void __insert_inode_hash(struct inode *, unsigned long hashval); extern void remove_inode_hash(struct inode *); diff --git a/mm/filemap.c b/mm/filemap.c index 2ed8b0389c5..5de7633e1db 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -1758,8 +1758,9 @@ static int __remove_suid(struct dentry *dentry, int kill) return notify_change(dentry, &newattrs); } -int remove_suid(struct dentry *dentry) +int file_remove_suid(struct file *file) { + struct dentry *dentry = file->f_path.dentry; int killsuid = should_remove_suid(dentry); int killpriv = security_inode_need_killpriv(dentry); int error = 0; @@ -1773,7 +1774,7 @@ int remove_suid(struct dentry *dentry) return error; } -EXPORT_SYMBOL(remove_suid); +EXPORT_SYMBOL(file_remove_suid); static size_t __iovec_copy_from_user_inatomic(char *vaddr, const struct iovec *iov, size_t base, size_t bytes) @@ -2529,7 +2530,7 @@ __generic_file_aio_write_nolock(struct kiocb *iocb, const struct iovec *iov, if (count == 0) goto out; - err = remove_suid(file->f_path.dentry); + err = file_remove_suid(file); if (err) goto out; diff --git a/mm/filemap_xip.c b/mm/filemap_xip.c index 3e744abcce9..98a3f31ccd6 100644 --- a/mm/filemap_xip.c +++ b/mm/filemap_xip.c @@ -380,7 +380,7 @@ xip_file_write(struct file *filp, const char __user *buf, size_t len, if (count == 0) goto out_backing; - ret = remove_suid(filp->f_path.dentry); + ret = file_remove_suid(filp); if (ret) goto out_backing; -- cgit v1.2.3-70-g09d2 From 7e79eedb3b22200cc8b774baea3a7bf28d766101 Mon Sep 17 00:00:00 2001 From: Tetsuo Handa Date: Tue, 24 Jun 2008 16:50:15 +0200 Subject: [patch 4/5] vfs: reuse local variable in vfs_link() Why not reuse "inode" which is assigned as struct inode *inode = old_dentry->d_inode; in the beginning of vfs_link() ? Signed-off-by: Tetsuo Handa Signed-off-by: Miklos Szeredi --- fs/namei.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/namei.c b/fs/namei.c index 46af98ed136..3b67be7631d 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -2524,19 +2524,19 @@ int vfs_link(struct dentry *old_dentry, struct inode *dir, struct dentry *new_de return -EPERM; if (!dir->i_op || !dir->i_op->link) return -EPERM; - if (S_ISDIR(old_dentry->d_inode->i_mode)) + if (S_ISDIR(inode->i_mode)) return -EPERM; error = security_inode_link(old_dentry, dir, new_dentry); if (error) return error; - mutex_lock(&old_dentry->d_inode->i_mutex); + mutex_lock(&inode->i_mutex); DQUOT_INIT(dir); error = dir->i_op->link(old_dentry, dir, new_dentry); - mutex_unlock(&old_dentry->d_inode->i_mutex); + mutex_unlock(&inode->i_mutex); if (!error) - fsnotify_link(dir, old_dentry->d_inode, new_dentry); + fsnotify_link(dir, inode, new_dentry); return error; } -- cgit v1.2.3-70-g09d2 From db2e747b14991a4c6a5c98b0e5f552a193237c03 Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Tue, 24 Jun 2008 16:50:16 +0200 Subject: [patch 5/5] vfs: remove mode parameter from vfs_symlink() Remove the unused mode parameter from vfs_symlink and callers. Thanks to Tetsuo Handa for noticing. CC: Tetsuo Handa Signed-off-by: Miklos Szeredi --- fs/ecryptfs/inode.c | 4 +--- fs/namei.c | 4 ++-- fs/nfsd/vfs.c | 10 ++-------- include/linux/fs.h | 2 +- 4 files changed, 6 insertions(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/ecryptfs/inode.c b/fs/ecryptfs/inode.c index 32f4228efcd..f25caf2b088 100644 --- a/fs/ecryptfs/inode.c +++ b/fs/ecryptfs/inode.c @@ -465,7 +465,6 @@ static int ecryptfs_symlink(struct inode *dir, struct dentry *dentry, int rc; struct dentry *lower_dentry; struct dentry *lower_dir_dentry; - umode_t mode; char *encoded_symname; int encoded_symlen; struct ecryptfs_crypt_stat *crypt_stat = NULL; @@ -473,7 +472,6 @@ static int ecryptfs_symlink(struct inode *dir, struct dentry *dentry, lower_dentry = ecryptfs_dentry_to_lower(dentry); dget(lower_dentry); lower_dir_dentry = lock_parent(lower_dentry); - mode = S_IALLUGO; encoded_symlen = ecryptfs_encode_filename(crypt_stat, symname, strlen(symname), &encoded_symname); @@ -482,7 +480,7 @@ static int ecryptfs_symlink(struct inode *dir, struct dentry *dentry, goto out_lock; } rc = vfs_symlink(lower_dir_dentry->d_inode, lower_dentry, - encoded_symname, mode); + encoded_symname); kfree(encoded_symname); if (rc || !lower_dentry->d_inode) goto out_lock; diff --git a/fs/namei.c b/fs/namei.c index 3b67be7631d..ae0e56fdb74 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -2434,7 +2434,7 @@ asmlinkage long sys_unlink(const char __user *pathname) return do_unlinkat(AT_FDCWD, pathname); } -int vfs_symlink(struct inode *dir, struct dentry *dentry, const char *oldname, int mode) +int vfs_symlink(struct inode *dir, struct dentry *dentry, const char *oldname) { int error = may_create(dir, dentry, NULL); @@ -2483,7 +2483,7 @@ asmlinkage long sys_symlinkat(const char __user *oldname, error = mnt_want_write(nd.path.mnt); if (error) goto out_dput; - error = vfs_symlink(nd.path.dentry->d_inode, dentry, from, S_IALLUGO); + error = vfs_symlink(nd.path.dentry->d_inode, dentry, from); mnt_drop_write(nd.path.mnt); out_dput: dput(dentry); diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index 0f4481e0502..ad1ad59e374 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -1516,7 +1516,6 @@ nfsd_symlink(struct svc_rqst *rqstp, struct svc_fh *fhp, struct dentry *dentry, *dnew; __be32 err, cerr; int host_err; - umode_t mode; err = nfserr_noent; if (!flen || !plen) @@ -1535,11 +1534,6 @@ nfsd_symlink(struct svc_rqst *rqstp, struct svc_fh *fhp, if (IS_ERR(dnew)) goto out_nfserr; - mode = S_IALLUGO; - /* Only the MODE ATTRibute is even vaguely meaningful */ - if (iap && (iap->ia_valid & ATTR_MODE)) - mode = iap->ia_mode & S_IALLUGO; - host_err = mnt_want_write(fhp->fh_export->ex_path.mnt); if (host_err) goto out_nfserr; @@ -1551,11 +1545,11 @@ nfsd_symlink(struct svc_rqst *rqstp, struct svc_fh *fhp, else { strncpy(path_alloced, path, plen); path_alloced[plen] = 0; - host_err = vfs_symlink(dentry->d_inode, dnew, path_alloced, mode); + host_err = vfs_symlink(dentry->d_inode, dnew, path_alloced); kfree(path_alloced); } } else - host_err = vfs_symlink(dentry->d_inode, dnew, path, mode); + host_err = vfs_symlink(dentry->d_inode, dnew, path); if (!host_err) { if (EX_ISSYNC(fhp->fh_export)) diff --git a/include/linux/fs.h b/include/linux/fs.h index 1a3546e69f9..25998e803fc 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1139,7 +1139,7 @@ extern int vfs_permission(struct nameidata *, int); extern int vfs_create(struct inode *, struct dentry *, int, struct nameidata *); extern int vfs_mkdir(struct inode *, struct dentry *, int); extern int vfs_mknod(struct inode *, struct dentry *, int, dev_t); -extern int vfs_symlink(struct inode *, struct dentry *, const char *, int); +extern int vfs_symlink(struct inode *, struct dentry *, const char *); extern int vfs_link(struct dentry *, struct inode *, struct dentry *); extern int vfs_rmdir(struct inode *, struct dentry *); extern int vfs_unlink(struct inode *, struct dentry *); -- cgit v1.2.3-70-g09d2 From 8bb79224b87aab92071e94d46e70bd160d89bf34 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 16 Jul 2008 09:51:03 -0400 Subject: [PATCH] permission checks for chdir need special treatment only on the last step ... so we ought to pass MAY_CHDIR to vfs_permission() instead of having it triggered on every step of preceding pathname resolution. LOOKUP_CHDIR is killed by that. Signed-off-by: Al Viro --- fs/namei.c | 2 -- fs/open.c | 5 ++--- include/linux/namei.h | 1 - 3 files changed, 2 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/namei.c b/fs/namei.c index ae0e56fdb74..6c76e1ee9c4 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -268,8 +268,6 @@ int permission(struct inode *inode, int mask, struct nameidata *nd) if (nd) { if (nd->flags & LOOKUP_ACCESS) extra |= MAY_ACCESS; - if (nd->flags & LOOKUP_CHDIR) - extra |= MAY_CHDIR; if (nd->flags & LOOKUP_OPEN) extra |= MAY_OPEN; } diff --git a/fs/open.c b/fs/open.c index b2e4c93aed0..8e02d42bfe4 100644 --- a/fs/open.c +++ b/fs/open.c @@ -501,12 +501,11 @@ asmlinkage long sys_chdir(const char __user * filename) struct nameidata nd; int error; - error = __user_walk(filename, - LOOKUP_FOLLOW|LOOKUP_DIRECTORY|LOOKUP_CHDIR, &nd); + error = __user_walk(filename, LOOKUP_FOLLOW|LOOKUP_DIRECTORY, &nd); if (error) goto out; - error = vfs_permission(&nd, MAY_EXEC); + error = vfs_permission(&nd, MAY_EXEC | MAY_CHDIR); if (error) goto dput_and_out; diff --git a/include/linux/namei.h b/include/linux/namei.h index 24d88e98a62..3cf62d26d49 100644 --- a/include/linux/namei.h +++ b/include/linux/namei.h @@ -55,7 +55,6 @@ enum {LAST_NORM, LAST_ROOT, LAST_DOT, LAST_DOTDOT, LAST_BIND}; #define LOOKUP_OPEN (0x0100) #define LOOKUP_CREATE (0x0200) #define LOOKUP_ACCESS (0x0400) -#define LOOKUP_CHDIR (0x0800) extern int __user_walk(const char __user *, unsigned, struct nameidata *); extern int __user_walk_fd(int dfd, const char __user *, unsigned, struct nameidata *); -- cgit v1.2.3-70-g09d2 From 7f2da1e7d0330395e5e9e350b879b98a1ea495df Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 10 May 2008 20:44:54 -0400 Subject: [PATCH] kill altroot long overdue... Signed-off-by: Al Viro --- fs/namei.c | 89 +------------------------------------------ fs/namespace.c | 8 +--- fs/open.c | 3 +- include/asm-alpha/namei.h | 17 --------- include/asm-arm/namei.h | 25 ------------ include/asm-avr32/namei.h | 7 ---- include/asm-blackfin/namei.h | 19 --------- include/asm-cris/namei.h | 17 --------- include/asm-frv/namei.h | 18 --------- include/asm-h8300/namei.h | 17 --------- include/asm-ia64/namei.h | 25 ------------ include/asm-m32r/namei.h | 17 --------- include/asm-m68k/namei.h | 17 --------- include/asm-m68knommu/namei.h | 1 - include/asm-mips/namei.h | 11 ------ include/asm-mn10300/namei.h | 22 ----------- include/asm-parisc/namei.h | 17 --------- include/asm-powerpc/namei.h | 20 ---------- include/asm-s390/namei.h | 21 ---------- include/asm-sh/namei.h | 17 --------- include/asm-sparc/namei.h | 8 ---- include/asm-sparc64/namei.h | 1 - include/asm-um/namei.h | 6 --- include/asm-v850/namei.h | 17 --------- include/asm-x86/namei.h | 11 ------ include/asm-xtensa/namei.h | 26 ------------- include/linux/fs_struct.h | 3 +- include/linux/namei.h | 1 - kernel/exec_domain.c | 1 - kernel/exit.c | 2 - kernel/fork.c | 7 ---- 31 files changed, 5 insertions(+), 466 deletions(-) delete mode 100644 include/asm-alpha/namei.h delete mode 100644 include/asm-arm/namei.h delete mode 100644 include/asm-avr32/namei.h delete mode 100644 include/asm-blackfin/namei.h delete mode 100644 include/asm-cris/namei.h delete mode 100644 include/asm-frv/namei.h delete mode 100644 include/asm-h8300/namei.h delete mode 100644 include/asm-ia64/namei.h delete mode 100644 include/asm-m32r/namei.h delete mode 100644 include/asm-m68k/namei.h delete mode 100644 include/asm-m68knommu/namei.h delete mode 100644 include/asm-mips/namei.h delete mode 100644 include/asm-mn10300/namei.h delete mode 100644 include/asm-parisc/namei.h delete mode 100644 include/asm-powerpc/namei.h delete mode 100644 include/asm-s390/namei.h delete mode 100644 include/asm-sh/namei.h delete mode 100644 include/asm-sparc/namei.h delete mode 100644 include/asm-sparc64/namei.h delete mode 100644 include/asm-um/namei.h delete mode 100644 include/asm-v850/namei.h delete mode 100644 include/asm-x86/namei.h delete mode 100644 include/asm-xtensa/namei.h (limited to 'fs') diff --git a/fs/namei.c b/fs/namei.c index 6c76e1ee9c4..095818089ac 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -31,7 +31,6 @@ #include #include #include -#include #include #define ACC_MODE(x) ("\000\004\002\006"[(x)&O_ACCMODE]) @@ -562,27 +561,16 @@ out_unlock: return result; } -static int __emul_lookup_dentry(const char *, struct nameidata *); - /* SMP-safe */ -static __always_inline int +static __always_inline void walk_init_root(const char *name, struct nameidata *nd) { struct fs_struct *fs = current->fs; read_lock(&fs->lock); - if (fs->altroot.dentry && !(nd->flags & LOOKUP_NOALT)) { - nd->path = fs->altroot; - path_get(&fs->altroot); - read_unlock(&fs->lock); - if (__emul_lookup_dentry(name,nd)) - return 0; - read_lock(&fs->lock); - } nd->path = fs->root; path_get(&fs->root); read_unlock(&fs->lock); - return 1; } /* @@ -623,12 +611,9 @@ static __always_inline int __vfs_follow_link(struct nameidata *nd, const char *l if (*link == '/') { path_put(&nd->path); - if (!walk_init_root(link, nd)) - /* weird __emul_prefix() stuff did it */ - goto out; + walk_init_root(link, nd); } res = link_path_walk(link, nd); -out: if (nd->depth || res || nd->last_type!=LAST_NORM) return res; /* @@ -1077,67 +1062,6 @@ static int path_walk(const char *name, struct nameidata *nd) return link_path_walk(name, nd); } -/* - * SMP-safe: Returns 1 and nd will have valid dentry and mnt, if - * everything is done. Returns 0 and drops input nd, if lookup failed; - */ -static int __emul_lookup_dentry(const char *name, struct nameidata *nd) -{ - if (path_walk(name, nd)) - return 0; /* something went wrong... */ - - if (!nd->path.dentry->d_inode || - S_ISDIR(nd->path.dentry->d_inode->i_mode)) { - struct path old_path = nd->path; - struct qstr last = nd->last; - int last_type = nd->last_type; - struct fs_struct *fs = current->fs; - - /* - * NAME was not found in alternate root or it's a directory. - * Try to find it in the normal root: - */ - nd->last_type = LAST_ROOT; - read_lock(&fs->lock); - nd->path = fs->root; - path_get(&fs->root); - read_unlock(&fs->lock); - if (path_walk(name, nd) == 0) { - if (nd->path.dentry->d_inode) { - path_put(&old_path); - return 1; - } - path_put(&nd->path); - } - nd->path = old_path; - nd->last = last; - nd->last_type = last_type; - } - return 1; -} - -void set_fs_altroot(void) -{ - char *emul = __emul_prefix(); - struct nameidata nd; - struct path path = {}, old_path; - int err; - struct fs_struct *fs = current->fs; - - if (!emul) - goto set_it; - err = path_lookup(emul, LOOKUP_FOLLOW|LOOKUP_DIRECTORY|LOOKUP_NOALT, &nd); - if (!err) - path = nd.path; -set_it: - write_lock(&fs->lock); - old_path = fs->altroot; - fs->altroot = path; - write_unlock(&fs->lock); - if (old_path.dentry) - path_put(&old_path); -} - /* Returns 0 and nd will be valid on success; Retuns error, otherwise. */ static int do_path_lookup(int dfd, const char *name, unsigned int flags, struct nameidata *nd) @@ -1153,14 +1077,6 @@ static int do_path_lookup(int dfd, const char *name, if (*name=='/') { read_lock(&fs->lock); - if (fs->altroot.dentry && !(nd->flags & LOOKUP_NOALT)) { - nd->path = fs->altroot; - path_get(&fs->altroot); - read_unlock(&fs->lock); - if (__emul_lookup_dentry(name,nd)) - goto out; /* found in altroot */ - read_lock(&fs->lock); - } nd->path = fs->root; path_get(&fs->root); read_unlock(&fs->lock); @@ -1194,7 +1110,6 @@ static int do_path_lookup(int dfd, const char *name, } retval = path_walk(name, nd); -out: if (unlikely(!retval && !audit_dummy_context() && nd->path.dentry && nd->path.dentry->d_inode)) audit_inode(name, nd->path.dentry); diff --git a/fs/namespace.c b/fs/namespace.c index f30b11e2240..c4fcf48acef 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -1972,7 +1972,7 @@ static struct mnt_namespace *dup_mnt_ns(struct mnt_namespace *mnt_ns, struct fs_struct *fs) { struct mnt_namespace *new_ns; - struct vfsmount *rootmnt = NULL, *pwdmnt = NULL, *altrootmnt = NULL; + struct vfsmount *rootmnt = NULL, *pwdmnt = NULL; struct vfsmount *p, *q; new_ns = kmalloc(sizeof(struct mnt_namespace), GFP_KERNEL); @@ -2015,10 +2015,6 @@ static struct mnt_namespace *dup_mnt_ns(struct mnt_namespace *mnt_ns, pwdmnt = p; fs->pwd.mnt = mntget(q); } - if (p == fs->altroot.mnt) { - altrootmnt = p; - fs->altroot.mnt = mntget(q); - } } p = next_mnt(p, mnt_ns->root); q = next_mnt(q, new_ns->root); @@ -2029,8 +2025,6 @@ static struct mnt_namespace *dup_mnt_ns(struct mnt_namespace *mnt_ns, mntput(rootmnt); if (pwdmnt) mntput(pwdmnt); - if (altrootmnt) - mntput(altrootmnt); return new_ns; } diff --git a/fs/open.c b/fs/open.c index 8e02d42bfe4..d3a2a00f52d 100644 --- a/fs/open.c +++ b/fs/open.c @@ -548,7 +548,7 @@ asmlinkage long sys_chroot(const char __user * filename) struct nameidata nd; int error; - error = __user_walk(filename, LOOKUP_FOLLOW | LOOKUP_DIRECTORY | LOOKUP_NOALT, &nd); + error = __user_walk(filename, LOOKUP_FOLLOW | LOOKUP_DIRECTORY, &nd); if (error) goto out; @@ -561,7 +561,6 @@ asmlinkage long sys_chroot(const char __user * filename) goto dput_and_out; set_fs_root(current->fs, &nd.path); - set_fs_altroot(); error = 0; dput_and_out: path_put(&nd.path); diff --git a/include/asm-alpha/namei.h b/include/asm-alpha/namei.h deleted file mode 100644 index 5cc9bb39499..00000000000 --- a/include/asm-alpha/namei.h +++ /dev/null @@ -1,17 +0,0 @@ -/* $Id: namei.h,v 1.1 1996/12/13 14:48:21 jj Exp $ - * linux/include/asm-alpha/namei.h - * - * Included from linux/fs/namei.c - */ - -#ifndef __ALPHA_NAMEI_H -#define __ALPHA_NAMEI_H - -/* This dummy routine maybe changed to something useful - * for /usr/gnemul/ emulation stuff. - * Look at asm-sparc/namei.h for details. - */ - -#define __emul_prefix() NULL - -#endif /* __ALPHA_NAMEI_H */ diff --git a/include/asm-arm/namei.h b/include/asm-arm/namei.h deleted file mode 100644 index a402d3b9d0f..00000000000 --- a/include/asm-arm/namei.h +++ /dev/null @@ -1,25 +0,0 @@ -/* - * linux/include/asm-arm/namei.h - * - * Routines to handle famous /usr/gnemul - * Derived from the Sparc version of this file - * - * Included from linux/fs/namei.c - */ - -#ifndef __ASMARM_NAMEI_H -#define __ASMARM_NAMEI_H - -#define ARM_BSD_EMUL "usr/gnemul/bsd/" - -static inline char *__emul_prefix(void) -{ - switch (current->personality) { - case PER_BSD: - return ARM_BSD_EMUL; - default: - return NULL; - } -} - -#endif /* __ASMARM_NAMEI_H */ diff --git a/include/asm-avr32/namei.h b/include/asm-avr32/namei.h deleted file mode 100644 index f0a26de06ca..00000000000 --- a/include/asm-avr32/namei.h +++ /dev/null @@ -1,7 +0,0 @@ -#ifndef __ASM_AVR32_NAMEI_H -#define __ASM_AVR32_NAMEI_H - -/* This dummy routine may be changed to something useful */ -#define __emul_prefix() NULL - -#endif /* __ASM_AVR32_NAMEI_H */ diff --git a/include/asm-blackfin/namei.h b/include/asm-blackfin/namei.h deleted file mode 100644 index 8b89a2d65cb..00000000000 --- a/include/asm-blackfin/namei.h +++ /dev/null @@ -1,19 +0,0 @@ -/* - * linux/include/asm/namei.h - * - * Included from linux/fs/namei.c - * - * Changes made by Lineo Inc. May 2001 - */ - -#ifndef __BFIN_NAMEI_H -#define __BFIN_NAMEI_H - -/* This dummy routine maybe changed to something useful - * for /usr/gnemul/ emulation stuff. - * Look at asm-sparc/namei.h for details. - */ - -#define __emul_prefix() NULL - -#endif diff --git a/include/asm-cris/namei.h b/include/asm-cris/namei.h deleted file mode 100644 index 8a3be7a6d9f..00000000000 --- a/include/asm-cris/namei.h +++ /dev/null @@ -1,17 +0,0 @@ -/* $Id: namei.h,v 1.1 2000/07/10 16:32:31 bjornw Exp $ - * linux/include/asm-cris/namei.h - * - * Included from linux/fs/namei.c - */ - -#ifndef __CRIS_NAMEI_H -#define __CRIS_NAMEI_H - -/* used to find file-system prefixes for doing emulations - * see for example asm-sparc/namei.h - * we don't use it... - */ - -#define __emul_prefix() NULL - -#endif /* __CRIS_NAMEI_H */ diff --git a/include/asm-frv/namei.h b/include/asm-frv/namei.h deleted file mode 100644 index 4ea57171d95..00000000000 --- a/include/asm-frv/namei.h +++ /dev/null @@ -1,18 +0,0 @@ -/* - * include/asm-frv/namei.h - * - * Included from linux/fs/namei.c - */ - -#ifndef __ASM_NAMEI_H -#define __ASM_NAMEI_H - -/* This dummy routine maybe changed to something useful - * for /usr/gnemul/ emulation stuff. - * Look at asm-sparc/namei.h for details. - */ - -#define __emul_prefix() NULL - -#endif - diff --git a/include/asm-h8300/namei.h b/include/asm-h8300/namei.h deleted file mode 100644 index ab6f196db6e..00000000000 --- a/include/asm-h8300/namei.h +++ /dev/null @@ -1,17 +0,0 @@ -/* - * linux/include/asm-h8300/namei.h - * - * Included from linux/fs/namei.c - */ - -#ifndef __H8300_NAMEI_H -#define __H8300_NAMEI_H - -/* This dummy routine maybe changed to something useful - * for /usr/gnemul/ emulation stuff. - * Look at asm-sparc/namei.h for details. - */ - -#define __emul_prefix() NULL - -#endif diff --git a/include/asm-ia64/namei.h b/include/asm-ia64/namei.h deleted file mode 100644 index 78e76807908..00000000000 --- a/include/asm-ia64/namei.h +++ /dev/null @@ -1,25 +0,0 @@ -#ifndef _ASM_IA64_NAMEI_H -#define _ASM_IA64_NAMEI_H - -/* - * Modified 1998, 1999, 2001 - * David Mosberger-Tang , Hewlett-Packard Co - */ - -#include -#include - -#define EMUL_PREFIX_LINUX_IA32 "/emul/ia32-linux/" - -static inline char * -__emul_prefix (void) -{ - switch (current->personality) { - case PER_LINUX32: - return EMUL_PREFIX_LINUX_IA32; - default: - return NULL; - } -} - -#endif /* _ASM_IA64_NAMEI_H */ diff --git a/include/asm-m32r/namei.h b/include/asm-m32r/namei.h deleted file mode 100644 index 210f8056b80..00000000000 --- a/include/asm-m32r/namei.h +++ /dev/null @@ -1,17 +0,0 @@ -#ifndef _ASM_M32R_NAMEI_H -#define _ASM_M32R_NAMEI_H - -/* - * linux/include/asm-m32r/namei.h - * - * Included from linux/fs/namei.c - */ - -/* This dummy routine maybe changed to something useful - * for /usr/gnemul/ emulation stuff. - * Look at asm-sparc/namei.h for details. - */ - -#define __emul_prefix() NULL - -#endif /* _ASM_M32R_NAMEI_H */ diff --git a/include/asm-m68k/namei.h b/include/asm-m68k/namei.h deleted file mode 100644 index f33f243b644..00000000000 --- a/include/asm-m68k/namei.h +++ /dev/null @@ -1,17 +0,0 @@ -/* - * linux/include/asm-m68k/namei.h - * - * Included from linux/fs/namei.c - */ - -#ifndef __M68K_NAMEI_H -#define __M68K_NAMEI_H - -/* This dummy routine maybe changed to something useful - * for /usr/gnemul/ emulation stuff. - * Look at asm-sparc/namei.h for details. - */ - -#define __emul_prefix() NULL - -#endif diff --git a/include/asm-m68knommu/namei.h b/include/asm-m68knommu/namei.h deleted file mode 100644 index 31a85d27b93..00000000000 --- a/include/asm-m68knommu/namei.h +++ /dev/null @@ -1 +0,0 @@ -#include diff --git a/include/asm-mips/namei.h b/include/asm-mips/namei.h deleted file mode 100644 index a6605a75246..00000000000 --- a/include/asm-mips/namei.h +++ /dev/null @@ -1,11 +0,0 @@ -#ifndef _ASM_NAMEI_H -#define _ASM_NAMEI_H - -/* - * This dummy routine maybe changed to something useful - * for /usr/gnemul/ emulation stuff. - */ - -#define __emul_prefix() NULL - -#endif /* _ASM_NAMEI_H */ diff --git a/include/asm-mn10300/namei.h b/include/asm-mn10300/namei.h deleted file mode 100644 index bd9ce94aeb6..00000000000 --- a/include/asm-mn10300/namei.h +++ /dev/null @@ -1,22 +0,0 @@ -/* Emulation stuff - * - * Copyright (C) 2007 Red Hat, Inc. All Rights Reserved. - * Written by David Howells (dhowells@redhat.com) - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public Licence - * as published by the Free Software Foundation; either version - * 2 of the Licence, or (at your option) any later version. - */ - -#ifndef _ASM_NAMEI_H -#define _ASM_NAMEI_H - -/* This dummy routine maybe changed to something useful - * for /usr/gnemul/ emulation stuff. - * Look at asm-sparc/namei.h for details. - */ - -#define __emul_prefix() NULL - -#endif /* _ASM_NAMEI_H */ diff --git a/include/asm-parisc/namei.h b/include/asm-parisc/namei.h deleted file mode 100644 index 8d29b3d9fb3..00000000000 --- a/include/asm-parisc/namei.h +++ /dev/null @@ -1,17 +0,0 @@ -/* $Id: namei.h,v 1.1 1996/12/13 14:48:21 jj Exp $ - * linux/include/asm-parisc/namei.h - * - * Included from linux/fs/namei.c - */ - -#ifndef __PARISC_NAMEI_H -#define __PARISC_NAMEI_H - -/* This dummy routine maybe changed to something useful - * for /usr/gnemul/ emulation stuff. - * Look at asm-sparc/namei.h for details. - */ - -#define __emul_prefix() NULL - -#endif /* __PARISC_NAMEI_H */ diff --git a/include/asm-powerpc/namei.h b/include/asm-powerpc/namei.h deleted file mode 100644 index 657443474a6..00000000000 --- a/include/asm-powerpc/namei.h +++ /dev/null @@ -1,20 +0,0 @@ -#ifndef _ASM_POWERPC_NAMEI_H -#define _ASM_POWERPC_NAMEI_H - -#ifdef __KERNEL__ - -/* - * Adapted from include/asm-alpha/namei.h - * - * Included from fs/namei.c - */ - -/* This dummy routine maybe changed to something useful - * for /usr/gnemul/ emulation stuff. - * Look at asm-sparc/namei.h for details. - */ - -#define __emul_prefix() NULL - -#endif /* __KERNEL__ */ -#endif /* _ASM_POWERPC_NAMEI_H */ diff --git a/include/asm-s390/namei.h b/include/asm-s390/namei.h deleted file mode 100644 index 3e286bdde4b..00000000000 --- a/include/asm-s390/namei.h +++ /dev/null @@ -1,21 +0,0 @@ -/* - * include/asm-s390/namei.h - * - * S390 version - * - * Derived from "include/asm-i386/namei.h" - * - * Included from linux/fs/namei.c - */ - -#ifndef __S390_NAMEI_H -#define __S390_NAMEI_H - -/* This dummy routine maybe changed to something useful - * for /usr/gnemul/ emulation stuff. - * Look at asm-sparc/namei.h for details. - */ - -#define __emul_prefix() NULL - -#endif /* __S390_NAMEI_H */ diff --git a/include/asm-sh/namei.h b/include/asm-sh/namei.h deleted file mode 100644 index 338a5d94714..00000000000 --- a/include/asm-sh/namei.h +++ /dev/null @@ -1,17 +0,0 @@ -/* $Id: namei.h,v 1.3 2000/07/04 06:24:49 gniibe Exp $ - * linux/include/asm-sh/namei.h - * - * Included from linux/fs/namei.c - */ - -#ifndef __ASM_SH_NAMEI_H -#define __ASM_SH_NAMEI_H - -/* This dummy routine maybe changed to something useful - * for /usr/gnemul/ emulation stuff. - * Look at asm-sparc/namei.h for details. - */ - -#define __emul_prefix() NULL - -#endif /* __ASM_SH_NAMEI_H */ diff --git a/include/asm-sparc/namei.h b/include/asm-sparc/namei.h deleted file mode 100644 index eff944b8e32..00000000000 --- a/include/asm-sparc/namei.h +++ /dev/null @@ -1,8 +0,0 @@ -#ifndef ___ASM_SPARC_NAMEI_H -#define ___ASM_SPARC_NAMEI_H -#if defined(__sparc__) && defined(__arch64__) -#include -#else -#include -#endif -#endif diff --git a/include/asm-sparc64/namei.h b/include/asm-sparc64/namei.h deleted file mode 100644 index 1344a910ba2..00000000000 --- a/include/asm-sparc64/namei.h +++ /dev/null @@ -1 +0,0 @@ -#include diff --git a/include/asm-um/namei.h b/include/asm-um/namei.h deleted file mode 100644 index 002984d5bc8..00000000000 --- a/include/asm-um/namei.h +++ /dev/null @@ -1,6 +0,0 @@ -#ifndef __UM_NAMEI_H -#define __UM_NAMEI_H - -#include "asm/arch/namei.h" - -#endif diff --git a/include/asm-v850/namei.h b/include/asm-v850/namei.h deleted file mode 100644 index ee8339b2384..00000000000 --- a/include/asm-v850/namei.h +++ /dev/null @@ -1,17 +0,0 @@ -/* - * linux/include/asm-v850/namei.h - * - * Included from linux/fs/namei.c - */ - -#ifndef __V850_NAMEI_H__ -#define __V850_NAMEI_H__ - -/* This dummy routine maybe changed to something useful - * for /usr/gnemul/ emulation stuff. - * Look at asm-sparc/namei.h for details. - */ - -#define __emul_prefix() NULL - -#endif /* __V850_NAMEI_H__ */ diff --git a/include/asm-x86/namei.h b/include/asm-x86/namei.h deleted file mode 100644 index 415ef5d9550..00000000000 --- a/include/asm-x86/namei.h +++ /dev/null @@ -1,11 +0,0 @@ -#ifndef _ASM_X86_NAMEI_H -#define _ASM_X86_NAMEI_H - -/* This dummy routine maybe changed to something useful - * for /usr/gnemul/ emulation stuff. - * Look at asm-sparc/namei.h for details. - */ - -#define __emul_prefix() NULL - -#endif /* _ASM_X86_NAMEI_H */ diff --git a/include/asm-xtensa/namei.h b/include/asm-xtensa/namei.h deleted file mode 100644 index 3fdff039d27..00000000000 --- a/include/asm-xtensa/namei.h +++ /dev/null @@ -1,26 +0,0 @@ -/* - * include/asm-xtensa/namei.h - * - * Included from linux/fs/namei.c - * - * This file is subject to the terms and conditions of the GNU General Public - * License. See the file "COPYING" in the main directory of this archive - * for more details. - * - * Copyright (C) 2001 - 2005 Tensilica Inc. - */ - -#ifndef _XTENSA_NAMEI_H -#define _XTENSA_NAMEI_H - -#ifdef __KERNEL__ - -/* This dummy routine maybe changed to something useful - * for /usr/gnemul/ emulation stuff. - * Look at asm-sparc/namei.h for details. - */ - -#define __emul_prefix() NULL - -#endif /* __KERNEL__ */ -#endif /* _XTENSA_NAMEI_H */ diff --git a/include/linux/fs_struct.h b/include/linux/fs_struct.h index 282f5421912..9e5a06e78d0 100644 --- a/include/linux/fs_struct.h +++ b/include/linux/fs_struct.h @@ -7,7 +7,7 @@ struct fs_struct { atomic_t count; rwlock_t lock; int umask; - struct path root, pwd, altroot; + struct path root, pwd; }; #define INIT_FS { \ @@ -19,7 +19,6 @@ struct fs_struct { extern struct kmem_cache *fs_cachep; extern void exit_fs(struct task_struct *); -extern void set_fs_altroot(void); extern void set_fs_root(struct fs_struct *, struct path *); extern void set_fs_pwd(struct fs_struct *, struct path *); extern struct fs_struct *copy_fs_struct(struct fs_struct *); diff --git a/include/linux/namei.h b/include/linux/namei.h index 3cf62d26d49..768773d5785 100644 --- a/include/linux/namei.h +++ b/include/linux/namei.h @@ -47,7 +47,6 @@ enum {LAST_NORM, LAST_ROOT, LAST_DOT, LAST_DOTDOT, LAST_BIND}; #define LOOKUP_DIRECTORY 2 #define LOOKUP_CONTINUE 4 #define LOOKUP_PARENT 16 -#define LOOKUP_NOALT 32 #define LOOKUP_REVAL 64 /* * Intent data diff --git a/kernel/exec_domain.c b/kernel/exec_domain.c index c1ef192aa65..0d407e88673 100644 --- a/kernel/exec_domain.c +++ b/kernel/exec_domain.c @@ -168,7 +168,6 @@ __set_personality(u_long personality) current->personality = personality; oep = current_thread_info()->exec_domain; current_thread_info()->exec_domain = ep; - set_fs_altroot(); module_put(oep->module); return 0; diff --git a/kernel/exit.c b/kernel/exit.c index 6cdf60712bd..0caf590548a 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -565,8 +565,6 @@ void put_fs_struct(struct fs_struct *fs) if (atomic_dec_and_test(&fs->count)) { path_put(&fs->root); path_put(&fs->pwd); - if (fs->altroot.dentry) - path_put(&fs->altroot); kmem_cache_free(fs_cachep, fs); } } diff --git a/kernel/fork.c b/kernel/fork.c index abb3ed6298f..5e050c1317c 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -657,13 +657,6 @@ static struct fs_struct *__copy_fs_struct(struct fs_struct *old) path_get(&old->root); fs->pwd = old->pwd; path_get(&old->pwd); - if (old->altroot.dentry) { - fs->altroot = old->altroot; - path_get(&old->altroot); - } else { - fs->altroot.mnt = NULL; - fs->altroot.dentry = NULL; - } read_unlock(&old->lock); } return fs; -- cgit v1.2.3-70-g09d2 From a110343f0d6d41f68b7cf8c00b57a3172c67f816 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Thu, 17 Jul 2008 09:19:08 -0400 Subject: [PATCH] fix MAY_CHDIR/MAY_ACCESS/LOOKUP_ACCESS mess * MAY_CHDIR is redundant - it's an equivalent of MAY_ACCESS * MAY_ACCESS on fuse should affect only the last step of pathname resolution * fchdir() and chroot() should pass MAY_ACCESS, for the same reason why chdir() needs that. * now that we pass MAY_ACCESS explicitly in all cases, LOOKUP_ACCESS can be removed; it has no business being in nameidata. Signed-off-by: Al Viro --- fs/fuse/dir.c | 2 +- fs/namei.c | 2 -- fs/open.c | 10 +++++----- include/linux/fs.h | 3 +-- include/linux/namei.h | 1 - 5 files changed, 7 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index 48a7934cb95..fd03330cade 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -962,7 +962,7 @@ static int fuse_permission(struct inode *inode, int mask) exist. So if permissions are revoked this won't be noticed immediately, only after the attribute timeout has expired */ - } else if (mask & (MAY_ACCESS | MAY_CHDIR)) { + } else if (mask & MAY_ACCESS) { err = fuse_access(inode, mask); } else if ((mask & MAY_EXEC) && S_ISREG(inode->i_mode)) { if (!(inode->i_mode & S_IXUGO)) { diff --git a/fs/namei.c b/fs/namei.c index 095818089ac..33dcaf025c4 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -265,8 +265,6 @@ int permission(struct inode *inode, int mask, struct nameidata *nd) if (inode->i_op && inode->i_op->permission) { int extra = 0; if (nd) { - if (nd->flags & LOOKUP_ACCESS) - extra |= MAY_ACCESS; if (nd->flags & LOOKUP_OPEN) extra |= MAY_OPEN; } diff --git a/fs/open.c b/fs/open.c index d3a2a00f52d..3317e1909b2 100644 --- a/fs/open.c +++ b/fs/open.c @@ -457,11 +457,11 @@ asmlinkage long sys_faccessat(int dfd, const char __user *filename, int mode) old_cap = cap_set_effective(current->cap_permitted); } - res = __user_walk_fd(dfd, filename, LOOKUP_FOLLOW|LOOKUP_ACCESS, &nd); + res = __user_walk_fd(dfd, filename, LOOKUP_FOLLOW, &nd); if (res) goto out; - res = vfs_permission(&nd, mode); + res = vfs_permission(&nd, mode | MAY_ACCESS); /* SuS v2 requires we report a read only fs too */ if(res || !(mode & S_IWOTH) || special_file(nd.path.dentry->d_inode->i_mode)) @@ -505,7 +505,7 @@ asmlinkage long sys_chdir(const char __user * filename) if (error) goto out; - error = vfs_permission(&nd, MAY_EXEC | MAY_CHDIR); + error = vfs_permission(&nd, MAY_EXEC | MAY_ACCESS); if (error) goto dput_and_out; @@ -534,7 +534,7 @@ asmlinkage long sys_fchdir(unsigned int fd) if (!S_ISDIR(inode->i_mode)) goto out_putf; - error = file_permission(file, MAY_EXEC); + error = file_permission(file, MAY_EXEC | MAY_ACCESS); if (!error) set_fs_pwd(current->fs, &file->f_path); out_putf: @@ -552,7 +552,7 @@ asmlinkage long sys_chroot(const char __user * filename) if (error) goto out; - error = vfs_permission(&nd, MAY_EXEC); + error = vfs_permission(&nd, MAY_EXEC | MAY_ACCESS); if (error) goto dput_and_out; diff --git a/include/linux/fs.h b/include/linux/fs.h index 25998e803fc..d8721e818b4 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -61,8 +61,7 @@ extern int dir_notify_enable; #define MAY_READ 4 #define MAY_APPEND 8 #define MAY_ACCESS 16 -#define MAY_CHDIR 32 -#define MAY_OPEN 64 +#define MAY_OPEN 32 #define FMODE_READ 1 #define FMODE_WRITE 2 diff --git a/include/linux/namei.h b/include/linux/namei.h index 768773d5785..60e35a02f6c 100644 --- a/include/linux/namei.h +++ b/include/linux/namei.h @@ -53,7 +53,6 @@ enum {LAST_NORM, LAST_ROOT, LAST_DOT, LAST_DOTDOT, LAST_BIND}; */ #define LOOKUP_OPEN (0x0100) #define LOOKUP_CREATE (0x0200) -#define LOOKUP_ACCESS (0x0400) extern int __user_walk(const char __user *, unsigned, struct nameidata *); extern int __user_walk_fd(int dfd, const char __user *, unsigned, struct nameidata *); -- cgit v1.2.3-70-g09d2 From b77b0646ef4efe31a7449bb3d9360fd00f95433d Mon Sep 17 00:00:00 2001 From: Al Viro Date: Thu, 17 Jul 2008 09:37:02 -0400 Subject: [PATCH] pass MAY_OPEN to vfs_permission() explicitly ... and get rid of the last "let's deduce mask from nameidata->flags" bit. Signed-off-by: Al Viro --- fs/exec.c | 4 ++-- fs/namei.c | 13 ++++--------- include/linux/security.h | 7 +++---- security/capability.c | 3 +-- security/security.c | 4 ++-- security/selinux/hooks.c | 5 ++--- security/smack/smack_lsm.c | 3 +-- 7 files changed, 15 insertions(+), 24 deletions(-) (limited to 'fs') diff --git a/fs/exec.c b/fs/exec.c index b8792a13153..0ba5d355c5a 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -118,7 +118,7 @@ asmlinkage long sys_uselib(const char __user * library) if (!S_ISREG(nd.path.dentry->d_inode->i_mode)) goto exit; - error = vfs_permission(&nd, MAY_READ | MAY_EXEC); + error = vfs_permission(&nd, MAY_READ | MAY_EXEC | MAY_OPEN); if (error) goto exit; @@ -666,7 +666,7 @@ struct file *open_exec(const char *name) struct inode *inode = nd.path.dentry->d_inode; file = ERR_PTR(-EACCES); if (S_ISREG(inode->i_mode)) { - int err = vfs_permission(&nd, MAY_EXEC); + int err = vfs_permission(&nd, MAY_EXEC | MAY_OPEN); file = ERR_PTR(err); if (!err) { file = nameidata_to_filp(&nd, diff --git a/fs/namei.c b/fs/namei.c index 33dcaf025c4..6b0e8e5e079 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -263,12 +263,7 @@ int permission(struct inode *inode, int mask, struct nameidata *nd) /* Ordinary permission routines do not understand MAY_APPEND. */ if (inode->i_op && inode->i_op->permission) { - int extra = 0; - if (nd) { - if (nd->flags & LOOKUP_OPEN) - extra |= MAY_OPEN; - } - retval = inode->i_op->permission(inode, mask | extra); + retval = inode->i_op->permission(inode, mask); if (!retval) { /* * Exec permission on a regular file is denied if none @@ -292,7 +287,7 @@ int permission(struct inode *inode, int mask, struct nameidata *nd) return retval; return security_inode_permission(inode, - mask & (MAY_READ|MAY_WRITE|MAY_EXEC), nd); + mask & (MAY_READ|MAY_WRITE|MAY_EXEC)); } /** @@ -492,7 +487,7 @@ static int exec_permission_lite(struct inode *inode, return -EACCES; ok: - return security_inode_permission(inode, MAY_EXEC, nd); + return security_inode_permission(inode, MAY_EXEC); } /* @@ -1692,7 +1687,7 @@ struct file *do_filp_open(int dfd, const char *pathname, int will_write; int flag = open_to_namei_flags(open_flag); - acc_mode = ACC_MODE(flag); + acc_mode = MAY_OPEN | ACC_MODE(flag); /* O_TRUNC implies we need access checks for write permissions */ if (flag & O_TRUNC) diff --git a/include/linux/security.h b/include/linux/security.h index f0e9adb22ac..fd96e7f8a6f 100644 --- a/include/linux/security.h +++ b/include/linux/security.h @@ -1362,7 +1362,7 @@ struct security_operations { struct inode *new_dir, struct dentry *new_dentry); int (*inode_readlink) (struct dentry *dentry); int (*inode_follow_link) (struct dentry *dentry, struct nameidata *nd); - int (*inode_permission) (struct inode *inode, int mask, struct nameidata *nd); + int (*inode_permission) (struct inode *inode, int mask); int (*inode_setattr) (struct dentry *dentry, struct iattr *attr); int (*inode_getattr) (struct vfsmount *mnt, struct dentry *dentry); void (*inode_delete) (struct inode *inode); @@ -1628,7 +1628,7 @@ int security_inode_rename(struct inode *old_dir, struct dentry *old_dentry, struct inode *new_dir, struct dentry *new_dentry); int security_inode_readlink(struct dentry *dentry); int security_inode_follow_link(struct dentry *dentry, struct nameidata *nd); -int security_inode_permission(struct inode *inode, int mask, struct nameidata *nd); +int security_inode_permission(struct inode *inode, int mask); int security_inode_setattr(struct dentry *dentry, struct iattr *attr); int security_inode_getattr(struct vfsmount *mnt, struct dentry *dentry); void security_inode_delete(struct inode *inode); @@ -2021,8 +2021,7 @@ static inline int security_inode_follow_link(struct dentry *dentry, return 0; } -static inline int security_inode_permission(struct inode *inode, int mask, - struct nameidata *nd) +static inline int security_inode_permission(struct inode *inode, int mask) { return 0; } diff --git a/security/capability.c b/security/capability.c index 5b01c0b0242..63d10da515a 100644 --- a/security/capability.c +++ b/security/capability.c @@ -211,8 +211,7 @@ static int cap_inode_follow_link(struct dentry *dentry, return 0; } -static int cap_inode_permission(struct inode *inode, int mask, - struct nameidata *nd) +static int cap_inode_permission(struct inode *inode, int mask) { return 0; } diff --git a/security/security.c b/security/security.c index 59f23b5918b..78ed3ffde24 100644 --- a/security/security.c +++ b/security/security.c @@ -429,11 +429,11 @@ int security_inode_follow_link(struct dentry *dentry, struct nameidata *nd) return security_ops->inode_follow_link(dentry, nd); } -int security_inode_permission(struct inode *inode, int mask, struct nameidata *nd) +int security_inode_permission(struct inode *inode, int mask) { if (unlikely(IS_PRIVATE(inode))) return 0; - return security_ops->inode_permission(inode, mask, nd); + return security_ops->inode_permission(inode, mask); } int security_inode_setattr(struct dentry *dentry, struct iattr *attr) diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c index 3481cde5bf1..5ba13908b5b 100644 --- a/security/selinux/hooks.c +++ b/security/selinux/hooks.c @@ -2624,12 +2624,11 @@ static int selinux_inode_follow_link(struct dentry *dentry, struct nameidata *na return dentry_has_perm(current, NULL, dentry, FILE__READ); } -static int selinux_inode_permission(struct inode *inode, int mask, - struct nameidata *nd) +static int selinux_inode_permission(struct inode *inode, int mask) { int rc; - rc = secondary_ops->inode_permission(inode, mask, nd); + rc = secondary_ops->inode_permission(inode, mask); if (rc) return rc; diff --git a/security/smack/smack_lsm.c b/security/smack/smack_lsm.c index ee5a51cbc5e..1b40e558f98 100644 --- a/security/smack/smack_lsm.c +++ b/security/smack/smack_lsm.c @@ -522,8 +522,7 @@ static int smack_inode_rename(struct inode *old_inode, * * Returns 0 if access is permitted, -EACCES otherwise */ -static int smack_inode_permission(struct inode *inode, int mask, - struct nameidata *nd) +static int smack_inode_permission(struct inode *inode, int mask) { /* * No permission to check. Existence test. Yup, it's there. -- cgit v1.2.3-70-g09d2 From 672b16b2f66c149888bd876a4f92342112205fe1 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Thu, 17 Jul 2008 09:45:01 -0400 Subject: [PATCH] more nameidata removal: exec_permission_lite() doesn't need it Signed-off-by: Al Viro --- fs/namei.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/namei.c b/fs/namei.c index 6b0e8e5e079..6d75430358a 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -460,8 +460,7 @@ static struct dentry * cached_lookup(struct dentry * parent, struct qstr * name, * short-cut DAC fails, then call permission() to do more * complete permission check. */ -static int exec_permission_lite(struct inode *inode, - struct nameidata *nd) +static int exec_permission_lite(struct inode *inode) { umode_t mode = inode->i_mode; @@ -884,7 +883,7 @@ static int __link_path_walk(const char *name, struct nameidata *nd) unsigned int c; nd->flags |= LOOKUP_CONTINUE; - err = exec_permission_lite(inode, nd); + err = exec_permission_lite(inode); if (err == -EAGAIN) err = vfs_permission(nd, MAY_EXEC); if (err) -- cgit v1.2.3-70-g09d2 From 88b387824fdaecb6ba0f471acf0aadf7d24739fd Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Mon, 21 Jul 2008 18:06:36 +0800 Subject: [PATCH] vfs: use kstrdup() and check failing allocation - use kstrdup() instead of kmalloc() + memcpy() - return NULL if allocating ->mnt_devname failed - mnt_devname should be const Signed-off-by: Li Zefan Acked-by: Cyrill Gorcunov Signed-off-by: Al Viro --- fs/namespace.c | 24 +++++++++++++----------- include/linux/mount.h | 2 +- 2 files changed, 14 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/namespace.c b/fs/namespace.c index c4fcf48acef..26380f59953 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -112,9 +112,13 @@ struct vfsmount *alloc_vfsmnt(const char *name) int err; err = mnt_alloc_id(mnt); - if (err) { - kmem_cache_free(mnt_cache, mnt); - return NULL; + if (err) + goto out_free_cache; + + if (name) { + mnt->mnt_devname = kstrdup(name, GFP_KERNEL); + if (!mnt->mnt_devname) + goto out_free_id; } atomic_set(&mnt->mnt_count, 1); @@ -127,16 +131,14 @@ struct vfsmount *alloc_vfsmnt(const char *name) INIT_LIST_HEAD(&mnt->mnt_slave_list); INIT_LIST_HEAD(&mnt->mnt_slave); atomic_set(&mnt->__mnt_writers, 0); - if (name) { - int size = strlen(name) + 1; - char *newname = kmalloc(size, GFP_KERNEL); - if (newname) { - memcpy(newname, name, size); - mnt->mnt_devname = newname; - } - } } return mnt; + +out_free_id: + mnt_free_id(mnt); +out_free_cache: + kmem_cache_free(mnt_cache, mnt); + return NULL; } /* diff --git a/include/linux/mount.h b/include/linux/mount.h index 4374d1adeb4..b5efaa2132a 100644 --- a/include/linux/mount.h +++ b/include/linux/mount.h @@ -47,7 +47,7 @@ struct vfsmount { struct list_head mnt_child; /* and going through their mnt_child */ int mnt_flags; /* 4 bytes hole on 64bits arches */ - char *mnt_devname; /* Name of device e.g. /dev/dsk/hda1 */ + const char *mnt_devname; /* Name of device e.g. /dev/dsk/hda1 */ struct list_head mnt_list; struct list_head mnt_expire; /* link in fs-specific expiry list */ struct list_head mnt_share; /* circular list of shared mounts */ -- cgit v1.2.3-70-g09d2 From 9767d74957450da6365c363d69e3d02d605d7375 Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Tue, 1 Jul 2008 15:01:26 +0200 Subject: [patch 1/4] vfs: utimes: move owner check into inode_change_ok() Add a new ia_valid flag: ATTR_TIMES_SET, to handle the UTIMES_OMIT/UTIMES_NOW and UTIMES_NOW/UTIMES_OMIT cases. In these cases neither ATTR_MTIME_SET nor ATTR_ATIME_SET is in the flags, yet the POSIX draft specifies that permission checking is performed the same way as if one or both of the times was explicitly set to a timestamp. See the path "vfs: utimensat(): fix error checking for {UTIME_NOW,UTIME_OMIT} case" by Michael Kerrisk for the patch introducing this behavior. This is a cleanup, as well as allowing filesystems (NFS/fuse/...) to perform their own permission checking instead of the default. CC: Ulrich Drepper CC: Michael Kerrisk Signed-off-by: Miklos Szeredi Signed-off-by: Al Viro --- fs/attr.c | 2 +- fs/utimes.c | 17 ++++------------- include/linux/fs.h | 33 +++++++++++++++++---------------- 3 files changed, 22 insertions(+), 30 deletions(-) (limited to 'fs') diff --git a/fs/attr.c b/fs/attr.c index 966b73e25f8..765fc75fab3 100644 --- a/fs/attr.c +++ b/fs/attr.c @@ -51,7 +51,7 @@ int inode_change_ok(struct inode *inode, struct iattr *attr) } /* Check for setting the inode time. */ - if (ia_valid & (ATTR_MTIME_SET | ATTR_ATIME_SET)) { + if (ia_valid & (ATTR_MTIME_SET | ATTR_ATIME_SET | ATTR_TIMES_SET)) { if (!is_owner_or_cap(inode)) goto error; } diff --git a/fs/utimes.c b/fs/utimes.c index b6b664e7145..ecf8941ba34 100644 --- a/fs/utimes.c +++ b/fs/utimes.c @@ -101,7 +101,6 @@ long do_utimes(int dfd, char __user *filename, struct timespec *times, int flags times[1].tv_nsec == UTIME_NOW) times = NULL; - /* In most cases, the checks are done in inode_change_ok() */ newattrs.ia_valid = ATTR_CTIME | ATTR_MTIME | ATTR_ATIME; if (times) { error = -EPERM; @@ -123,21 +122,13 @@ long do_utimes(int dfd, char __user *filename, struct timespec *times, int flags newattrs.ia_mtime.tv_nsec = times[1].tv_nsec; newattrs.ia_valid |= ATTR_MTIME_SET; } - /* - * For the UTIME_OMIT/UTIME_NOW and UTIME_NOW/UTIME_OMIT - * cases, we need to make an extra check that is not done by - * inode_change_ok(). + * Tell inode_change_ok(), that this is an explicit time + * update, even if neither ATTR_ATIME_SET nor ATTR_MTIME_SET + * were used. */ - if (((times[0].tv_nsec == UTIME_NOW && - times[1].tv_nsec == UTIME_OMIT) - || - (times[0].tv_nsec == UTIME_OMIT && - times[1].tv_nsec == UTIME_NOW)) - && !is_owner_or_cap(inode)) - goto mnt_drop_write_and_out; + newattrs.ia_valid |= ATTR_TIMES_SET; } else { - /* * If times is NULL (or both times are UTIME_NOW), * then we need to check permissions, because diff --git a/include/linux/fs.h b/include/linux/fs.h index d8721e818b4..527b9e482f9 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -320,22 +320,23 @@ typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset, * Attribute flags. These should be or-ed together to figure out what * has been changed! */ -#define ATTR_MODE 1 -#define ATTR_UID 2 -#define ATTR_GID 4 -#define ATTR_SIZE 8 -#define ATTR_ATIME 16 -#define ATTR_MTIME 32 -#define ATTR_CTIME 64 -#define ATTR_ATIME_SET 128 -#define ATTR_MTIME_SET 256 -#define ATTR_FORCE 512 /* Not a change, but a change it */ -#define ATTR_ATTR_FLAG 1024 -#define ATTR_KILL_SUID 2048 -#define ATTR_KILL_SGID 4096 -#define ATTR_FILE 8192 -#define ATTR_KILL_PRIV 16384 -#define ATTR_OPEN 32768 /* Truncating from open(O_TRUNC) */ +#define ATTR_MODE (1 << 0) +#define ATTR_UID (1 << 1) +#define ATTR_GID (1 << 2) +#define ATTR_SIZE (1 << 3) +#define ATTR_ATIME (1 << 4) +#define ATTR_MTIME (1 << 5) +#define ATTR_CTIME (1 << 6) +#define ATTR_ATIME_SET (1 << 7) +#define ATTR_MTIME_SET (1 << 8) +#define ATTR_FORCE (1 << 9) /* Not a change, but a change it */ +#define ATTR_ATTR_FLAG (1 << 10) +#define ATTR_KILL_SUID (1 << 11) +#define ATTR_KILL_SGID (1 << 12) +#define ATTR_FILE (1 << 13) +#define ATTR_KILL_PRIV (1 << 14) +#define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */ +#define ATTR_TIMES_SET (1 << 16) /* * This is the Inode Attributes structure, used for notify_change(). It -- cgit v1.2.3-70-g09d2 From e9b76fedc61235da80b6b7f81dfd67ec224dfb49 Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Tue, 1 Jul 2008 15:01:27 +0200 Subject: [patch 2/4] vfs: utimes cleanup Untange the mess that is do_utimes(). Add kerneldoc comment to do_utimes(). CC: Ulrich Drepper CC: Michael Kerrisk Signed-off-by: Miklos Szeredi Signed-off-by: Al Viro --- fs/utimes.c | 114 ++++++++++++++++++++++++++++++++++-------------------------- 1 file changed, 65 insertions(+), 49 deletions(-) (limited to 'fs') diff --git a/fs/utimes.c b/fs/utimes.c index ecf8941ba34..8e09dbdfd7f 100644 --- a/fs/utimes.c +++ b/fs/utimes.c @@ -48,54 +48,15 @@ static bool nsec_valid(long nsec) return nsec >= 0 && nsec <= 999999999; } -/* If times==NULL, set access and modification to current time, - * must be owner or have write permission. - * Else, update from *times, must be owner or super user. - */ -long do_utimes(int dfd, char __user *filename, struct timespec *times, int flags) +static int utimes_common(struct path *path, struct timespec *times) { int error; - struct nameidata nd; - struct dentry *dentry; - struct inode *inode; struct iattr newattrs; - struct file *f = NULL; - struct vfsmount *mnt; - - error = -EINVAL; - if (times && (!nsec_valid(times[0].tv_nsec) || - !nsec_valid(times[1].tv_nsec))) { - goto out; - } - - if (flags & ~AT_SYMLINK_NOFOLLOW) - goto out; - - if (filename == NULL && dfd != AT_FDCWD) { - error = -EINVAL; - if (flags & AT_SYMLINK_NOFOLLOW) - goto out; - - error = -EBADF; - f = fget(dfd); - if (!f) - goto out; - dentry = f->f_path.dentry; - mnt = f->f_path.mnt; - } else { - error = __user_walk_fd(dfd, filename, (flags & AT_SYMLINK_NOFOLLOW) ? 0 : LOOKUP_FOLLOW, &nd); - if (error) - goto out; - - dentry = nd.path.dentry; - mnt = nd.path.mnt; - } - - inode = dentry->d_inode; + struct inode *inode = path->dentry->d_inode; - error = mnt_want_write(mnt); + error = mnt_want_write(path->mnt); if (error) - goto dput_and_out; + goto out; if (times && times[0].tv_nsec == UTIME_NOW && times[1].tv_nsec == UTIME_NOW) @@ -145,15 +106,70 @@ long do_utimes(int dfd, char __user *filename, struct timespec *times, int flags } } mutex_lock(&inode->i_mutex); - error = notify_change(dentry, &newattrs); + error = notify_change(path->dentry, &newattrs); mutex_unlock(&inode->i_mutex); + mnt_drop_write_and_out: - mnt_drop_write(mnt); -dput_and_out: - if (f) - fput(f); - else + mnt_drop_write(path->mnt); +out: + return error; +} + +/* + * do_utimes - change times on filename or file descriptor + * @dfd: open file descriptor, -1 or AT_FDCWD + * @filename: path name or NULL + * @times: new times or NULL + * @flags: zero or more flags (only AT_SYMLINK_NOFOLLOW for the moment) + * + * If filename is NULL and dfd refers to an open file, then operate on + * the file. Otherwise look up filename, possibly using dfd as a + * starting point. + * + * If times==NULL, set access and modification to current time, + * must be owner or have write permission. + * Else, update from *times, must be owner or super user. + */ +long do_utimes(int dfd, char __user *filename, struct timespec *times, int flags) +{ + int error = -EINVAL; + + if (times && (!nsec_valid(times[0].tv_nsec) || + !nsec_valid(times[1].tv_nsec))) { + goto out; + } + + if (flags & ~AT_SYMLINK_NOFOLLOW) + goto out; + + if (filename == NULL && dfd != AT_FDCWD) { + struct file *file; + + if (flags & AT_SYMLINK_NOFOLLOW) + goto out; + + file = fget(dfd); + error = -EBADF; + if (!file) + goto out; + + error = utimes_common(&file->f_path, times); + fput(file); + } else { + struct nameidata nd; + int lookup_flags = 0; + + if (!(flags & AT_SYMLINK_NOFOLLOW)) + lookup_flags |= LOOKUP_FOLLOW; + + error = __user_walk_fd(dfd, filename, lookup_flags, &nd); + if (error) + goto out; + + error = utimes_common(&nd.path, times); path_put(&nd.path); + } + out: return error; } -- cgit v1.2.3-70-g09d2 From b1da47e29e467f1ec36dc78d009bfb109fd533c7 Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Tue, 1 Jul 2008 15:01:28 +0200 Subject: [patch 3/4] fat: dont call notify_change The FAT_IOCTL_SET_ATTRIBUTES ioctl() calls notify_change() to change the file mode before changing the inode attributes. Replace with explicit calls to security_inode_setattr(), fat_setattr() and fsnotify_change(). This is equivalent to the original. The reason it is needed, is that later in the series we move the immutable check into notify_change(). That would break the FAT_IOCTL_SET_ATTRIBUTES ioctl, as it needs to perform the mode change regardless of the immutability of the file. [Fix error if fat is built as a module. Thanks to OGAWA Hirofumi for noticing.] Signed-off-by: Miklos Szeredi Acked-by: OGAWA Hirofumi Signed-off-by: Al Viro --- fs/fat/file.c | 15 ++++++++++++++- security/security.c | 1 + 2 files changed, 15 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/fat/file.c b/fs/fat/file.c index c672df4036e..8707a8cfa02 100644 --- a/fs/fat/file.c +++ b/fs/fat/file.c @@ -15,6 +15,8 @@ #include #include #include +#include +#include int fat_generic_ioctl(struct inode *inode, struct file *filp, unsigned int cmd, unsigned long arg) @@ -64,6 +66,7 @@ int fat_generic_ioctl(struct inode *inode, struct file *filp, /* Equivalent to a chmod() */ ia.ia_valid = ATTR_MODE | ATTR_CTIME; + ia.ia_ctime = current_fs_time(inode->i_sb); if (is_dir) { ia.ia_mode = MSDOS_MKMODE(attr, S_IRWXUGO & ~sbi->options.fs_dmask) @@ -90,11 +93,21 @@ int fat_generic_ioctl(struct inode *inode, struct file *filp, } } + /* + * The security check is questionable... We single + * out the RO attribute for checking by the security + * module, just because it maps to a file mode. + */ + err = security_inode_setattr(filp->f_path.dentry, &ia); + if (err) + goto up; + /* This MUST be done before doing anything irreversible... */ - err = notify_change(filp->f_path.dentry, &ia); + err = fat_setattr(filp->f_path.dentry, &ia); if (err) goto up; + fsnotify_change(filp->f_path.dentry, ia.ia_valid); if (sbi->options.sys_immutable) { if (attr & ATTR_SYS) inode->i_flags |= S_IMMUTABLE; diff --git a/security/security.c b/security/security.c index 78ed3ffde24..ff706872775 100644 --- a/security/security.c +++ b/security/security.c @@ -442,6 +442,7 @@ int security_inode_setattr(struct dentry *dentry, struct iattr *attr) return 0; return security_ops->inode_setattr(dentry, attr); } +EXPORT_SYMBOL_GPL(security_inode_setattr); int security_inode_getattr(struct vfsmount *mnt, struct dentry *dentry) { -- cgit v1.2.3-70-g09d2 From beb29e058c35ab69e96e455a12ccf7505f6de425 Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Tue, 1 Jul 2008 15:01:29 +0200 Subject: [patch 4/4] vfs: immutable inode checking cleanup Move the immutable and append-only checks from chmod, chown and utimes into notify_change(). Checks for immutable and append-only files are always performed by the VFS and not by the filesystem (see permission() and may_...() in namei.c), so these belong in notify_change(), and not in inode_change_ok(). This should be completely equivalent. CC: Ulrich Drepper CC: Michael Kerrisk Signed-off-by: Miklos Szeredi Signed-off-by: Al Viro --- fs/attr.c | 5 +++++ fs/open.c | 24 ++---------------------- fs/utimes.c | 4 ---- 3 files changed, 7 insertions(+), 26 deletions(-) (limited to 'fs') diff --git a/fs/attr.c b/fs/attr.c index 765fc75fab3..26c71ba1eed 100644 --- a/fs/attr.c +++ b/fs/attr.c @@ -108,6 +108,11 @@ int notify_change(struct dentry * dentry, struct iattr * attr) struct timespec now; unsigned int ia_valid = attr->ia_valid; + if (ia_valid & (ATTR_MODE | ATTR_UID | ATTR_GID | ATTR_TIMES_SET)) { + if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) + return -EPERM; + } + now = current_fs_time(inode->i_sb); attr->ia_ctime = now; diff --git a/fs/open.c b/fs/open.c index 3317e1909b2..3b3c43674be 100644 --- a/fs/open.c +++ b/fs/open.c @@ -588,9 +588,6 @@ asmlinkage long sys_fchmod(unsigned int fd, mode_t mode) err = mnt_want_write(file->f_path.mnt); if (err) goto out_putf; - err = -EPERM; - if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) - goto out_drop_write; mutex_lock(&inode->i_mutex); if (mode == (mode_t) -1) mode = inode->i_mode; @@ -598,8 +595,6 @@ asmlinkage long sys_fchmod(unsigned int fd, mode_t mode) newattrs.ia_valid = ATTR_MODE | ATTR_CTIME; err = notify_change(dentry, &newattrs); mutex_unlock(&inode->i_mutex); - -out_drop_write: mnt_drop_write(file->f_path.mnt); out_putf: fput(file); @@ -623,11 +618,6 @@ asmlinkage long sys_fchmodat(int dfd, const char __user *filename, error = mnt_want_write(nd.path.mnt); if (error) goto dput_and_out; - - error = -EPERM; - if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) - goto out_drop_write; - mutex_lock(&inode->i_mutex); if (mode == (mode_t) -1) mode = inode->i_mode; @@ -635,8 +625,6 @@ asmlinkage long sys_fchmodat(int dfd, const char __user *filename, newattrs.ia_valid = ATTR_MODE | ATTR_CTIME; error = notify_change(nd.path.dentry, &newattrs); mutex_unlock(&inode->i_mutex); - -out_drop_write: mnt_drop_write(nd.path.mnt); dput_and_out: path_put(&nd.path); @@ -651,18 +639,10 @@ asmlinkage long sys_chmod(const char __user *filename, mode_t mode) static int chown_common(struct dentry * dentry, uid_t user, gid_t group) { - struct inode * inode; + struct inode *inode = dentry->d_inode; int error; struct iattr newattrs; - error = -ENOENT; - if (!(inode = dentry->d_inode)) { - printk(KERN_ERR "chown_common: NULL inode\n"); - goto out; - } - error = -EPERM; - if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) - goto out; newattrs.ia_valid = ATTR_CTIME; if (user != (uid_t) -1) { newattrs.ia_valid |= ATTR_UID; @@ -678,7 +658,7 @@ static int chown_common(struct dentry * dentry, uid_t user, gid_t group) mutex_lock(&inode->i_mutex); error = notify_change(dentry, &newattrs); mutex_unlock(&inode->i_mutex); -out: + return error; } diff --git a/fs/utimes.c b/fs/utimes.c index 8e09dbdfd7f..dad679d3a15 100644 --- a/fs/utimes.c +++ b/fs/utimes.c @@ -64,10 +64,6 @@ static int utimes_common(struct path *path, struct timespec *times) newattrs.ia_valid = ATTR_CTIME | ATTR_MTIME | ATTR_ATIME; if (times) { - error = -EPERM; - if (IS_APPEND(inode) || IS_IMMUTABLE(inode)) - goto mnt_drop_write_and_out; - if (times[0].tv_nsec == UTIME_OMIT) newattrs.ia_valid &= ~ATTR_ATIME; else if (times[0].tv_nsec != UTIME_NOW) { -- cgit v1.2.3-70-g09d2 From e56b6a5dda1a36ffaa532df6f975ea324298fa4d Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 19 May 2008 07:53:34 +0200 Subject: Re: [PATCH 3/6] vfs: open_exec cleanup On Mon, May 19, 2008 at 12:01:49AM +0200, Marcin Slusarz wrote: > open_exec is needlessly indented, calls ERR_PTR with 0 argument > (which is not valid errno) and jumps into middle of function > just to return value. > So clean it up a bit. Still looks rather messy. See below for a better version. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/exec.c | 58 ++++++++++++++++++++++++++++++---------------------------- 1 file changed, 30 insertions(+), 28 deletions(-) (limited to 'fs') diff --git a/fs/exec.c b/fs/exec.c index 0ba5d355c5a..346e3f69c6e 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -656,38 +656,40 @@ EXPORT_SYMBOL(setup_arg_pages); struct file *open_exec(const char *name) { struct nameidata nd; - int err; struct file *file; + int err; - err = path_lookup_open(AT_FDCWD, name, LOOKUP_FOLLOW, &nd, FMODE_READ|FMODE_EXEC); - file = ERR_PTR(err); - - if (!err) { - struct inode *inode = nd.path.dentry->d_inode; - file = ERR_PTR(-EACCES); - if (S_ISREG(inode->i_mode)) { - int err = vfs_permission(&nd, MAY_EXEC | MAY_OPEN); - file = ERR_PTR(err); - if (!err) { - file = nameidata_to_filp(&nd, - O_RDONLY|O_LARGEFILE); - if (!IS_ERR(file)) { - err = deny_write_access(file); - if (err) { - fput(file); - file = ERR_PTR(err); - } - } -out: - return file; - } - } - release_open_intent(&nd); - path_put(&nd.path); + err = path_lookup_open(AT_FDCWD, name, LOOKUP_FOLLOW, &nd, + FMODE_READ|FMODE_EXEC); + if (err) + goto out; + + err = -EACCES; + if (!S_ISREG(nd.path.dentry->d_inode->i_mode)) + goto out_path_put; + + err = vfs_permission(&nd, MAY_EXEC | MAY_OPEN); + if (err) + goto out_path_put; + + file = nameidata_to_filp(&nd, O_RDONLY|O_LARGEFILE); + if (IS_ERR(file)) + return file; + + err = deny_write_access(file); + if (err) { + fput(file); + goto out; } - goto out; -} + return file; + + out_path_put: + release_open_intent(&nd); + path_put(&nd.path); + out: + return ERR_PTR(err); +} EXPORT_SYMBOL(open_exec); int kernel_read(struct file *file, unsigned long offset, -- cgit v1.2.3-70-g09d2 From 30524472c2f728c20d6bf35191042a5d455c0a64 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 22 Jul 2008 00:02:33 -0400 Subject: [PATCH] take noexec checks to very few callers that care Signed-off-by: Al Viro --- fs/exec.c | 7 +++++++ fs/namei.c | 9 --------- fs/open.c | 10 ++++++++++ 3 files changed, 17 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/exec.c b/fs/exec.c index 346e3f69c6e..eca58c29ede 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -118,6 +118,10 @@ asmlinkage long sys_uselib(const char __user * library) if (!S_ISREG(nd.path.dentry->d_inode->i_mode)) goto exit; + error = -EACCES; + if (nd.path.mnt->mnt_flags & MNT_NOEXEC) + goto exit; + error = vfs_permission(&nd, MAY_READ | MAY_EXEC | MAY_OPEN); if (error) goto exit; @@ -668,6 +672,9 @@ struct file *open_exec(const char *name) if (!S_ISREG(nd.path.dentry->d_inode->i_mode)) goto out_path_put; + if (nd.path.mnt->mnt_flags & MNT_NOEXEC) + goto out_path_put; + err = vfs_permission(&nd, MAY_EXEC | MAY_OPEN); if (err) goto out_path_put; diff --git a/fs/namei.c b/fs/namei.c index 6d75430358a..396cb3e5c36 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -252,15 +252,6 @@ int permission(struct inode *inode, int mask, struct nameidata *nd) return -EACCES; } - if ((mask & MAY_EXEC) && S_ISREG(inode->i_mode)) { - /* - * MAY_EXEC on regular files is denied if the fs is mounted - * with the "noexec" flag. - */ - if (mnt && (mnt->mnt_flags & MNT_NOEXEC)) - return -EACCES; - } - /* Ordinary permission routines do not understand MAY_APPEND. */ if (inode->i_op && inode->i_op->permission) { retval = inode->i_op->permission(inode, mask); diff --git a/fs/open.c b/fs/open.c index 3b3c43674be..d5e421ad0cf 100644 --- a/fs/open.c +++ b/fs/open.c @@ -461,6 +461,16 @@ asmlinkage long sys_faccessat(int dfd, const char __user *filename, int mode) if (res) goto out; + if ((mode & MAY_EXEC) && S_ISREG(nd.path.dentry->d_inode->i_mode)) { + /* + * MAY_EXEC on regular files is denied if the fs is mounted + * with the "noexec" flag. + */ + res = -EACCES; + if (nd.path.mnt->mnt_flags & MNT_NOEXEC) + goto out_path_release; + } + res = vfs_permission(&nd, mode | MAY_ACCESS); /* SuS v2 requires we report a read only fs too */ if(res || !(mode & S_IWOTH) || -- cgit v1.2.3-70-g09d2 From f419a2e3b64def707e1384ee38abb77f99af5f6d Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 22 Jul 2008 00:07:17 -0400 Subject: [PATCH] kill nameidata passing to permission(), rename to inode_permission() Incidentally, the name that gives hundreds of false positives on grep is not a good idea... Signed-off-by: Al Viro --- fs/ecryptfs/inode.c | 2 +- fs/namei.c | 22 +++++++++------------- fs/nfsd/nfsfh.c | 2 +- fs/nfsd/vfs.c | 4 ++-- fs/utimes.c | 2 +- fs/xattr.c | 2 +- include/linux/fs.h | 2 +- ipc/mqueue.c | 2 +- 8 files changed, 17 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/fs/ecryptfs/inode.c b/fs/ecryptfs/inode.c index f25caf2b088..89209f00f9c 100644 --- a/fs/ecryptfs/inode.c +++ b/fs/ecryptfs/inode.c @@ -830,7 +830,7 @@ out: static int ecryptfs_permission(struct inode *inode, int mask) { - return permission(ecryptfs_inode_to_lower(inode), mask, NULL); + return inode_permission(ecryptfs_inode_to_lower(inode), mask); } /** diff --git a/fs/namei.c b/fs/namei.c index 396cb3e5c36..5029b93ebbd 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -227,13 +227,9 @@ int generic_permission(struct inode *inode, int mask, return -EACCES; } -int permission(struct inode *inode, int mask, struct nameidata *nd) +int inode_permission(struct inode *inode, int mask) { int retval; - struct vfsmount *mnt = NULL; - - if (nd) - mnt = nd->path.mnt; if (mask & MAY_WRITE) { umode_t mode = inode->i_mode; @@ -293,7 +289,7 @@ int permission(struct inode *inode, int mask, struct nameidata *nd) */ int vfs_permission(struct nameidata *nd, int mask) { - return permission(nd->path.dentry->d_inode, mask, nd); + return inode_permission(nd->path.dentry->d_inode, mask); } /** @@ -310,7 +306,7 @@ int vfs_permission(struct nameidata *nd, int mask) */ int file_permission(struct file *file, int mask) { - return permission(file->f_path.dentry->d_inode, mask, NULL); + return inode_permission(file->f_path.dentry->d_inode, mask); } /* @@ -1262,7 +1258,7 @@ static struct dentry *lookup_hash(struct nameidata *nd) { int err; - err = permission(nd->path.dentry->d_inode, MAY_EXEC, nd); + err = inode_permission(nd->path.dentry->d_inode, MAY_EXEC); if (err) return ERR_PTR(err); return __lookup_hash(&nd->last, nd->path.dentry, nd); @@ -1310,7 +1306,7 @@ struct dentry *lookup_one_len(const char *name, struct dentry *base, int len) if (err) return ERR_PTR(err); - err = permission(base->d_inode, MAY_EXEC, NULL); + err = inode_permission(base->d_inode, MAY_EXEC); if (err) return ERR_PTR(err); return __lookup_hash(&this, base, NULL); @@ -1400,7 +1396,7 @@ static int may_delete(struct inode *dir,struct dentry *victim,int isdir) BUG_ON(victim->d_parent->d_inode != dir); audit_inode_child(victim->d_name.name, victim, dir); - error = permission(dir,MAY_WRITE | MAY_EXEC, NULL); + error = inode_permission(dir, MAY_WRITE | MAY_EXEC); if (error) return error; if (IS_APPEND(dir)) @@ -1437,7 +1433,7 @@ static inline int may_create(struct inode *dir, struct dentry *child, return -EEXIST; if (IS_DEADDIR(dir)) return -ENOENT; - return permission(dir,MAY_WRITE | MAY_EXEC, nd); + return inode_permission(dir, MAY_WRITE | MAY_EXEC); } /* @@ -2543,7 +2539,7 @@ static int vfs_rename_dir(struct inode *old_dir, struct dentry *old_dentry, * we'll need to flip '..'. */ if (new_dir != old_dir) { - error = permission(old_dentry->d_inode, MAY_WRITE, NULL); + error = inode_permission(old_dentry->d_inode, MAY_WRITE); if (error) return error; } @@ -2897,7 +2893,7 @@ EXPORT_SYMBOL(page_symlink); EXPORT_SYMBOL(page_symlink_inode_operations); EXPORT_SYMBOL(path_lookup); EXPORT_SYMBOL(vfs_path_lookup); -EXPORT_SYMBOL(permission); +EXPORT_SYMBOL(inode_permission); EXPORT_SYMBOL(vfs_permission); EXPORT_SYMBOL(file_permission); EXPORT_SYMBOL(unlock_rename); diff --git a/fs/nfsd/nfsfh.c b/fs/nfsd/nfsfh.c index f45451eb1e3..ea37c96f044 100644 --- a/fs/nfsd/nfsfh.c +++ b/fs/nfsd/nfsfh.c @@ -51,7 +51,7 @@ static int nfsd_acceptable(void *expv, struct dentry *dentry) /* make sure parents give x permission to user */ int err; parent = dget_parent(tdentry); - err = permission(parent->d_inode, MAY_EXEC, NULL); + err = inode_permission(parent->d_inode, MAY_EXEC); if (err < 0) { dput(parent); break; diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index ad1ad59e374..18060bed526 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -1953,12 +1953,12 @@ nfsd_permission(struct svc_rqst *rqstp, struct svc_export *exp, return 0; /* This assumes NFSD_MAY_{READ,WRITE,EXEC} == MAY_{READ,WRITE,EXEC} */ - err = permission(inode, acc & (MAY_READ|MAY_WRITE|MAY_EXEC), NULL); + err = inode_permission(inode, acc & (MAY_READ|MAY_WRITE|MAY_EXEC)); /* Allow read access to binaries even when mode 111 */ if (err == -EACCES && S_ISREG(inode->i_mode) && acc == (NFSD_MAY_READ | NFSD_MAY_OWNER_OVERRIDE)) - err = permission(inode, MAY_EXEC, NULL); + err = inode_permission(inode, MAY_EXEC); return err? nfserrno(err) : 0; } diff --git a/fs/utimes.c b/fs/utimes.c index dad679d3a15..dc28b782625 100644 --- a/fs/utimes.c +++ b/fs/utimes.c @@ -96,7 +96,7 @@ static int utimes_common(struct path *path, struct timespec *times) goto mnt_drop_write_and_out; if (!is_owner_or_cap(inode)) { - error = permission(inode, MAY_WRITE, NULL); + error = inode_permission(inode, MAY_WRITE); if (error) goto mnt_drop_write_and_out; } diff --git a/fs/xattr.c b/fs/xattr.c index 4706a8b1f49..b96222e05ba 100644 --- a/fs/xattr.c +++ b/fs/xattr.c @@ -63,7 +63,7 @@ xattr_permission(struct inode *inode, const char *name, int mask) return -EPERM; } - return permission(inode, mask, NULL); + return inode_permission(inode, mask); } int diff --git a/include/linux/fs.h b/include/linux/fs.h index 527b9e482f9..9d2de4cadab 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1770,7 +1770,7 @@ extern int do_remount_sb(struct super_block *sb, int flags, extern sector_t bmap(struct inode *, sector_t); #endif extern int notify_change(struct dentry *, struct iattr *); -extern int permission(struct inode *, int, struct nameidata *); +extern int inode_permission(struct inode *, int); extern int generic_permission(struct inode *, int, int (*check_acl)(struct inode *, int)); diff --git a/ipc/mqueue.c b/ipc/mqueue.c index 474984f9e03..96fb36cd987 100644 --- a/ipc/mqueue.c +++ b/ipc/mqueue.c @@ -638,7 +638,7 @@ static int oflag2acc[O_ACCMODE] = { MAY_READ, MAY_WRITE, return ERR_PTR(-EINVAL); } - if (permission(dentry->d_inode, oflag2acc[oflag & O_ACCMODE], NULL)) { + if (inode_permission(dentry->d_inode, oflag2acc[oflag & O_ACCMODE])) { dput(dentry); mntput(mqueue_mnt); return ERR_PTR(-EACCES); -- cgit v1.2.3-70-g09d2 From 256984a83880ff7ac78055cb87baea48137f0b77 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 22 Jul 2008 08:09:30 -0400 Subject: [PATCH] preparation to __user_walk_fd cleanup Almost all users __user_walk_fd() and friends care only about struct path. Get rid of the few that do not. Signed-off-by: Al Viro --- fs/inotify_user.c | 2 +- fs/open.c | 18 ++++++++++-------- 2 files changed, 11 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/inotify_user.c b/fs/inotify_user.c index fe79c25d95d..9b99ebf2888 100644 --- a/fs/inotify_user.c +++ b/fs/inotify_user.c @@ -365,7 +365,7 @@ static int find_inode(const char __user *dirname, struct nameidata *nd, if (error) return error; /* you can only watch an inode if you have read permissions on it */ - error = vfs_permission(nd, MAY_READ); + error = inode_permission(nd->path.dentry->d_inode, MAY_READ); if (error) path_put(&nd->path); return error; diff --git a/fs/open.c b/fs/open.c index d5e421ad0cf..e94266700ed 100644 --- a/fs/open.c +++ b/fs/open.c @@ -251,7 +251,7 @@ static long do_sys_truncate(const char __user * path, loff_t length) if (error) goto dput_and_out; - error = vfs_permission(&nd, MAY_WRITE); + error = inode_permission(inode, MAY_WRITE); if (error) goto mnt_drop_write_and_out; @@ -426,6 +426,7 @@ out: asmlinkage long sys_faccessat(int dfd, const char __user *filename, int mode) { struct nameidata nd; + struct inode *inode; int old_fsuid, old_fsgid; kernel_cap_t uninitialized_var(old_cap); /* !SECURE_NO_SETUID_FIXUP */ int res; @@ -461,7 +462,9 @@ asmlinkage long sys_faccessat(int dfd, const char __user *filename, int mode) if (res) goto out; - if ((mode & MAY_EXEC) && S_ISREG(nd.path.dentry->d_inode->i_mode)) { + inode = nd.path.dentry->d_inode; + + if ((mode & MAY_EXEC) && S_ISREG(inode->i_mode)) { /* * MAY_EXEC on regular files is denied if the fs is mounted * with the "noexec" flag. @@ -471,10 +474,9 @@ asmlinkage long sys_faccessat(int dfd, const char __user *filename, int mode) goto out_path_release; } - res = vfs_permission(&nd, mode | MAY_ACCESS); + res = inode_permission(inode, mode | MAY_ACCESS); /* SuS v2 requires we report a read only fs too */ - if(res || !(mode & S_IWOTH) || - special_file(nd.path.dentry->d_inode->i_mode)) + if (res || !(mode & S_IWOTH) || special_file(inode->i_mode)) goto out_path_release; /* * This is a rare case where using __mnt_is_readonly() @@ -515,7 +517,7 @@ asmlinkage long sys_chdir(const char __user * filename) if (error) goto out; - error = vfs_permission(&nd, MAY_EXEC | MAY_ACCESS); + error = inode_permission(nd.path.dentry->d_inode, MAY_EXEC | MAY_ACCESS); if (error) goto dput_and_out; @@ -544,7 +546,7 @@ asmlinkage long sys_fchdir(unsigned int fd) if (!S_ISDIR(inode->i_mode)) goto out_putf; - error = file_permission(file, MAY_EXEC | MAY_ACCESS); + error = inode_permission(inode, MAY_EXEC | MAY_ACCESS); if (!error) set_fs_pwd(current->fs, &file->f_path); out_putf: @@ -562,7 +564,7 @@ asmlinkage long sys_chroot(const char __user * filename) if (error) goto out; - error = vfs_permission(&nd, MAY_EXEC | MAY_ACCESS); + error = inode_permission(nd.path.dentry->d_inode, MAY_EXEC | MAY_ACCESS); if (error) goto dput_and_out; -- cgit v1.2.3-70-g09d2 From 2d8f30380ab8c706f4e0a8f1aaa22b5886e9ac8a Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 22 Jul 2008 09:59:21 -0400 Subject: [PATCH] sanitize __user_walk_fd() et.al. * do not pass nameidata; struct path is all the callers want. * switch to new helpers: user_path_at(dfd, pathname, flags, &path) user_path(pathname, &path) user_lpath(pathname, &path) user_path_dir(pathname, &path) (fail if not a directory) The last 3 are trivial macro wrappers for the first one. * remove nameidata in callers. Signed-off-by: Al Viro --- arch/alpha/kernel/osf_sys.c | 10 ++-- arch/parisc/hpux/sys_hpux.c | 10 ++-- fs/coda/pioctl.c | 14 ++--- fs/compat.c | 20 +++---- fs/inotify_user.c | 22 ++++---- fs/namei.c | 36 ++++++------- fs/namespace.c | 74 +++++++++++++------------- fs/open.c | 124 +++++++++++++++++++++---------------------- fs/stat.c | 32 +++++------ fs/utimes.c | 8 +-- fs/xattr.c | 96 ++++++++++++++++----------------- fs/xfs/linux-2.6/xfs_ioctl.c | 14 +++-- include/linux/namei.h | 13 ++--- 13 files changed, 235 insertions(+), 238 deletions(-) (limited to 'fs') diff --git a/arch/alpha/kernel/osf_sys.c b/arch/alpha/kernel/osf_sys.c index 32ca1b92730..6e943135f0e 100644 --- a/arch/alpha/kernel/osf_sys.c +++ b/arch/alpha/kernel/osf_sys.c @@ -253,15 +253,15 @@ do_osf_statfs(struct dentry * dentry, struct osf_statfs __user *buffer, } asmlinkage int -osf_statfs(char __user *path, struct osf_statfs __user *buffer, unsigned long bufsiz) +osf_statfs(char __user *pathname, struct osf_statfs __user *buffer, unsigned long bufsiz) { - struct nameidata nd; + struct path path; int retval; - retval = user_path_walk(path, &nd); + retval = user_path(pathname, &path); if (!retval) { - retval = do_osf_statfs(nd.path.dentry, buffer, bufsiz); - path_put(&nd.path); + retval = do_osf_statfs(path.dentry, buffer, bufsiz); + path_put(&path); } return retval; } diff --git a/arch/parisc/hpux/sys_hpux.c b/arch/parisc/hpux/sys_hpux.c index be255ebb609..18072e03a01 100644 --- a/arch/parisc/hpux/sys_hpux.c +++ b/arch/parisc/hpux/sys_hpux.c @@ -210,19 +210,19 @@ static int vfs_statfs_hpux(struct dentry *dentry, struct hpux_statfs *buf) } /* hpux statfs */ -asmlinkage long hpux_statfs(const char __user *path, +asmlinkage long hpux_statfs(const char __user *pathname, struct hpux_statfs __user *buf) { - struct nameidata nd; + struct path path; int error; - error = user_path_walk(path, &nd); + error = user_path(pathname, &path); if (!error) { struct hpux_statfs tmp; - error = vfs_statfs_hpux(nd.path.dentry, &tmp); + error = vfs_statfs_hpux(path.dentry, &tmp); if (!error && copy_to_user(buf, &tmp, sizeof(tmp))) error = -EFAULT; - path_put(&nd.path); + path_put(&path); } return error; } diff --git a/fs/coda/pioctl.c b/fs/coda/pioctl.c index c38a98974fb..c51365422aa 100644 --- a/fs/coda/pioctl.c +++ b/fs/coda/pioctl.c @@ -49,7 +49,7 @@ static int coda_ioctl_permission(struct inode *inode, int mask) static int coda_pioctl(struct inode * inode, struct file * filp, unsigned int cmd, unsigned long user_data) { - struct nameidata nd; + struct path path; int error; struct PioctlData data; struct inode *target_inode = NULL; @@ -64,21 +64,21 @@ static int coda_pioctl(struct inode * inode, struct file * filp, * Look up the pathname. Note that the pathname is in * user memory, and namei takes care of this */ - if ( data.follow ) { - error = user_path_walk(data.path, &nd); + if (data.follow) { + error = user_path(data.path, &path); } else { - error = user_path_walk_link(data.path, &nd); + error = user_lpath(data.path, &path); } if ( error ) { return error; } else { - target_inode = nd.path.dentry->d_inode; + target_inode = path.dentry->d_inode; } /* return if it is not a Coda inode */ if ( target_inode->i_sb != inode->i_sb ) { - path_put(&nd.path); + path_put(&path); return -EINVAL; } @@ -87,7 +87,7 @@ static int coda_pioctl(struct inode * inode, struct file * filp, error = venus_pioctl(inode->i_sb, &(cnp->c_fid), cmd, &data); - path_put(&nd.path); + path_put(&path); return error; } diff --git a/fs/compat.c b/fs/compat.c index 106eba28ec5..c9d1472e65c 100644 --- a/fs/compat.c +++ b/fs/compat.c @@ -234,18 +234,18 @@ static int put_compat_statfs(struct compat_statfs __user *ubuf, struct kstatfs * * The following statfs calls are copies of code from fs/open.c and * should be checked against those from time to time */ -asmlinkage long compat_sys_statfs(const char __user *path, struct compat_statfs __user *buf) +asmlinkage long compat_sys_statfs(const char __user *pathname, struct compat_statfs __user *buf) { - struct nameidata nd; + struct path path; int error; - error = user_path_walk(path, &nd); + error = user_path(pathname, &path); if (!error) { struct kstatfs tmp; - error = vfs_statfs(nd.path.dentry, &tmp); + error = vfs_statfs(path.dentry, &tmp); if (!error) error = put_compat_statfs(buf, &tmp); - path_put(&nd.path); + path_put(&path); } return error; } @@ -299,21 +299,21 @@ static int put_compat_statfs64(struct compat_statfs64 __user *ubuf, struct kstat return 0; } -asmlinkage long compat_sys_statfs64(const char __user *path, compat_size_t sz, struct compat_statfs64 __user *buf) +asmlinkage long compat_sys_statfs64(const char __user *pathname, compat_size_t sz, struct compat_statfs64 __user *buf) { - struct nameidata nd; + struct path path; int error; if (sz != sizeof(*buf)) return -EINVAL; - error = user_path_walk(path, &nd); + error = user_path(pathname, &path); if (!error) { struct kstatfs tmp; - error = vfs_statfs(nd.path.dentry, &tmp); + error = vfs_statfs(path.dentry, &tmp); if (!error) error = put_compat_statfs64(buf, &tmp); - path_put(&nd.path); + path_put(&path); } return error; } diff --git a/fs/inotify_user.c b/fs/inotify_user.c index 9b99ebf2888..60249429a25 100644 --- a/fs/inotify_user.c +++ b/fs/inotify_user.c @@ -354,20 +354,20 @@ static void inotify_dev_event_dequeue(struct inotify_device *dev) } /* - * find_inode - resolve a user-given path to a specific inode and return a nd + * find_inode - resolve a user-given path to a specific inode */ -static int find_inode(const char __user *dirname, struct nameidata *nd, +static int find_inode(const char __user *dirname, struct path *path, unsigned flags) { int error; - error = __user_walk(dirname, flags, nd); + error = user_path_at(AT_FDCWD, dirname, flags, path); if (error) return error; /* you can only watch an inode if you have read permissions on it */ - error = inode_permission(nd->path.dentry->d_inode, MAY_READ); + error = inode_permission(path->dentry->d_inode, MAY_READ); if (error) - path_put(&nd->path); + path_put(path); return error; } @@ -650,11 +650,11 @@ asmlinkage long sys_inotify_init(void) return sys_inotify_init1(0); } -asmlinkage long sys_inotify_add_watch(int fd, const char __user *path, u32 mask) +asmlinkage long sys_inotify_add_watch(int fd, const char __user *pathname, u32 mask) { struct inode *inode; struct inotify_device *dev; - struct nameidata nd; + struct path path; struct file *filp; int ret, fput_needed; unsigned flags = 0; @@ -674,12 +674,12 @@ asmlinkage long sys_inotify_add_watch(int fd, const char __user *path, u32 mask) if (mask & IN_ONLYDIR) flags |= LOOKUP_DIRECTORY; - ret = find_inode(path, &nd, flags); + ret = find_inode(pathname, &path, flags); if (unlikely(ret)) goto fput_and_out; - /* inode held in place by reference to nd; dev by fget on fd */ - inode = nd.path.dentry->d_inode; + /* inode held in place by reference to path; dev by fget on fd */ + inode = path.dentry->d_inode; dev = filp->private_data; mutex_lock(&dev->up_mutex); @@ -688,7 +688,7 @@ asmlinkage long sys_inotify_add_watch(int fd, const char __user *path, u32 mask) ret = create_watch(dev, inode, mask); mutex_unlock(&dev->up_mutex); - path_put(&nd.path); + path_put(&path); fput_and_out: fput_light(filp, fput_needed); return ret; diff --git a/fs/namei.c b/fs/namei.c index 5029b93ebbd..edb5e973f9b 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -1334,24 +1334,24 @@ struct dentry *lookup_one_noperm(const char *name, struct dentry *base) return __lookup_hash(&this, base, NULL); } -int __user_walk_fd(int dfd, const char __user *name, unsigned flags, - struct nameidata *nd) +int user_path_at(int dfd, const char __user *name, unsigned flags, + struct path *path) { + struct nameidata nd; char *tmp = getname(name); int err = PTR_ERR(tmp); - if (!IS_ERR(tmp)) { - err = do_path_lookup(dfd, tmp, flags, nd); + + BUG_ON(flags & LOOKUP_PARENT); + + err = do_path_lookup(dfd, tmp, flags, &nd); putname(tmp); + if (!err) + *path = nd.path; } return err; } -int __user_walk(const char __user *name, unsigned flags, struct nameidata *nd) -{ - return __user_walk_fd(AT_FDCWD, name, flags, nd); -} - /* * It's inline, so penalty for filesystems that don't use sticky bit is * minimal. @@ -2446,7 +2446,8 @@ asmlinkage long sys_linkat(int olddfd, const char __user *oldname, int flags) { struct dentry *new_dentry; - struct nameidata nd, old_nd; + struct nameidata nd; + struct path old_path; int error; char * to; @@ -2457,16 +2458,16 @@ asmlinkage long sys_linkat(int olddfd, const char __user *oldname, if (IS_ERR(to)) return PTR_ERR(to); - error = __user_walk_fd(olddfd, oldname, - flags & AT_SYMLINK_FOLLOW ? LOOKUP_FOLLOW : 0, - &old_nd); + error = user_path_at(olddfd, oldname, + flags & AT_SYMLINK_FOLLOW ? LOOKUP_FOLLOW : 0, + &old_path); if (error) goto exit; error = do_path_lookup(newdfd, to, LOOKUP_PARENT, &nd); if (error) goto out; error = -EXDEV; - if (old_nd.path.mnt != nd.path.mnt) + if (old_path.mnt != nd.path.mnt) goto out_release; new_dentry = lookup_create(&nd, 0); error = PTR_ERR(new_dentry); @@ -2475,7 +2476,7 @@ asmlinkage long sys_linkat(int olddfd, const char __user *oldname, error = mnt_want_write(nd.path.mnt); if (error) goto out_dput; - error = vfs_link(old_nd.path.dentry, nd.path.dentry->d_inode, new_dentry); + error = vfs_link(old_path.dentry, nd.path.dentry->d_inode, new_dentry); mnt_drop_write(nd.path.mnt); out_dput: dput(new_dentry); @@ -2484,7 +2485,7 @@ out_unlock: out_release: path_put(&nd.path); out: - path_put(&old_nd.path); + path_put(&old_path); exit: putname(to); @@ -2877,8 +2878,7 @@ const struct inode_operations page_symlink_inode_operations = { .put_link = page_put_link, }; -EXPORT_SYMBOL(__user_walk); -EXPORT_SYMBOL(__user_walk_fd); +EXPORT_SYMBOL(user_path_at); EXPORT_SYMBOL(follow_down); EXPORT_SYMBOL(follow_up); EXPORT_SYMBOL(get_write_access); /* binfmt_aout */ diff --git a/fs/namespace.c b/fs/namespace.c index 26380f59953..411728c0c8b 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -1130,27 +1130,27 @@ static int do_umount(struct vfsmount *mnt, int flags) asmlinkage long sys_umount(char __user * name, int flags) { - struct nameidata nd; + struct path path; int retval; - retval = __user_walk(name, LOOKUP_FOLLOW, &nd); + retval = user_path(name, &path); if (retval) goto out; retval = -EINVAL; - if (nd.path.dentry != nd.path.mnt->mnt_root) + if (path.dentry != path.mnt->mnt_root) goto dput_and_out; - if (!check_mnt(nd.path.mnt)) + if (!check_mnt(path.mnt)) goto dput_and_out; retval = -EPERM; if (!capable(CAP_SYS_ADMIN)) goto dput_and_out; - retval = do_umount(nd.path.mnt, flags); + retval = do_umount(path.mnt, flags); dput_and_out: /* we mustn't call path_put() as that would clear mnt_expiry_mark */ - dput(nd.path.dentry); - mntput_no_expire(nd.path.mnt); + dput(path.dentry); + mntput_no_expire(path.mnt); out: return retval; } @@ -2179,28 +2179,26 @@ asmlinkage long sys_pivot_root(const char __user * new_root, const char __user * put_old) { struct vfsmount *tmp; - struct nameidata new_nd, old_nd; - struct path parent_path, root_parent, root; + struct path new, old, parent_path, root_parent, root; int error; if (!capable(CAP_SYS_ADMIN)) return -EPERM; - error = __user_walk(new_root, LOOKUP_FOLLOW | LOOKUP_DIRECTORY, - &new_nd); + error = user_path_dir(new_root, &new); if (error) goto out0; error = -EINVAL; - if (!check_mnt(new_nd.path.mnt)) + if (!check_mnt(new.mnt)) goto out1; - error = __user_walk(put_old, LOOKUP_FOLLOW | LOOKUP_DIRECTORY, &old_nd); + error = user_path_dir(put_old, &old); if (error) goto out1; - error = security_sb_pivotroot(&old_nd.path, &new_nd.path); + error = security_sb_pivotroot(&old, &new); if (error) { - path_put(&old_nd.path); + path_put(&old); goto out1; } @@ -2209,69 +2207,69 @@ asmlinkage long sys_pivot_root(const char __user * new_root, path_get(¤t->fs->root); read_unlock(¤t->fs->lock); down_write(&namespace_sem); - mutex_lock(&old_nd.path.dentry->d_inode->i_mutex); + mutex_lock(&old.dentry->d_inode->i_mutex); error = -EINVAL; - if (IS_MNT_SHARED(old_nd.path.mnt) || - IS_MNT_SHARED(new_nd.path.mnt->mnt_parent) || + if (IS_MNT_SHARED(old.mnt) || + IS_MNT_SHARED(new.mnt->mnt_parent) || IS_MNT_SHARED(root.mnt->mnt_parent)) goto out2; if (!check_mnt(root.mnt)) goto out2; error = -ENOENT; - if (IS_DEADDIR(new_nd.path.dentry->d_inode)) + if (IS_DEADDIR(new.dentry->d_inode)) goto out2; - if (d_unhashed(new_nd.path.dentry) && !IS_ROOT(new_nd.path.dentry)) + if (d_unhashed(new.dentry) && !IS_ROOT(new.dentry)) goto out2; - if (d_unhashed(old_nd.path.dentry) && !IS_ROOT(old_nd.path.dentry)) + if (d_unhashed(old.dentry) && !IS_ROOT(old.dentry)) goto out2; error = -EBUSY; - if (new_nd.path.mnt == root.mnt || - old_nd.path.mnt == root.mnt) + if (new.mnt == root.mnt || + old.mnt == root.mnt) goto out2; /* loop, on the same file system */ error = -EINVAL; if (root.mnt->mnt_root != root.dentry) goto out2; /* not a mountpoint */ if (root.mnt->mnt_parent == root.mnt) goto out2; /* not attached */ - if (new_nd.path.mnt->mnt_root != new_nd.path.dentry) + if (new.mnt->mnt_root != new.dentry) goto out2; /* not a mountpoint */ - if (new_nd.path.mnt->mnt_parent == new_nd.path.mnt) + if (new.mnt->mnt_parent == new.mnt) goto out2; /* not attached */ /* make sure we can reach put_old from new_root */ - tmp = old_nd.path.mnt; + tmp = old.mnt; spin_lock(&vfsmount_lock); - if (tmp != new_nd.path.mnt) { + if (tmp != new.mnt) { for (;;) { if (tmp->mnt_parent == tmp) goto out3; /* already mounted on put_old */ - if (tmp->mnt_parent == new_nd.path.mnt) + if (tmp->mnt_parent == new.mnt) break; tmp = tmp->mnt_parent; } - if (!is_subdir(tmp->mnt_mountpoint, new_nd.path.dentry)) + if (!is_subdir(tmp->mnt_mountpoint, new.dentry)) goto out3; - } else if (!is_subdir(old_nd.path.dentry, new_nd.path.dentry)) + } else if (!is_subdir(old.dentry, new.dentry)) goto out3; - detach_mnt(new_nd.path.mnt, &parent_path); + detach_mnt(new.mnt, &parent_path); detach_mnt(root.mnt, &root_parent); /* mount old root on put_old */ - attach_mnt(root.mnt, &old_nd.path); + attach_mnt(root.mnt, &old); /* mount new_root on / */ - attach_mnt(new_nd.path.mnt, &root_parent); + attach_mnt(new.mnt, &root_parent); touch_mnt_namespace(current->nsproxy->mnt_ns); spin_unlock(&vfsmount_lock); - chroot_fs_refs(&root, &new_nd.path); - security_sb_post_pivotroot(&root, &new_nd.path); + chroot_fs_refs(&root, &new); + security_sb_post_pivotroot(&root, &new); error = 0; path_put(&root_parent); path_put(&parent_path); out2: - mutex_unlock(&old_nd.path.dentry->d_inode->i_mutex); + mutex_unlock(&old.dentry->d_inode->i_mutex); up_write(&namespace_sem); path_put(&root); - path_put(&old_nd.path); + path_put(&old); out1: - path_put(&new_nd.path); + path_put(&new); out0: return error; out3: diff --git a/fs/open.c b/fs/open.c index e94266700ed..3fe1a6857c7 100644 --- a/fs/open.c +++ b/fs/open.c @@ -122,37 +122,37 @@ static int vfs_statfs64(struct dentry *dentry, struct statfs64 *buf) return 0; } -asmlinkage long sys_statfs(const char __user * path, struct statfs __user * buf) +asmlinkage long sys_statfs(const char __user *pathname, struct statfs __user * buf) { - struct nameidata nd; + struct path path; int error; - error = user_path_walk(path, &nd); + error = user_path(pathname, &path); if (!error) { struct statfs tmp; - error = vfs_statfs_native(nd.path.dentry, &tmp); + error = vfs_statfs_native(path.dentry, &tmp); if (!error && copy_to_user(buf, &tmp, sizeof(tmp))) error = -EFAULT; - path_put(&nd.path); + path_put(&path); } return error; } -asmlinkage long sys_statfs64(const char __user *path, size_t sz, struct statfs64 __user *buf) +asmlinkage long sys_statfs64(const char __user *pathname, size_t sz, struct statfs64 __user *buf) { - struct nameidata nd; + struct path path; long error; if (sz != sizeof(*buf)) return -EINVAL; - error = user_path_walk(path, &nd); + error = user_path(pathname, &path); if (!error) { struct statfs64 tmp; - error = vfs_statfs64(nd.path.dentry, &tmp); + error = vfs_statfs64(path.dentry, &tmp); if (!error && copy_to_user(buf, &tmp, sizeof(tmp))) error = -EFAULT; - path_put(&nd.path); + path_put(&path); } return error; } @@ -223,20 +223,20 @@ int do_truncate(struct dentry *dentry, loff_t length, unsigned int time_attrs, return err; } -static long do_sys_truncate(const char __user * path, loff_t length) +static long do_sys_truncate(const char __user *pathname, loff_t length) { - struct nameidata nd; - struct inode * inode; + struct path path; + struct inode *inode; int error; error = -EINVAL; if (length < 0) /* sorry, but loff_t says... */ goto out; - error = user_path_walk(path, &nd); + error = user_path(pathname, &path); if (error) goto out; - inode = nd.path.dentry->d_inode; + inode = path.dentry->d_inode; /* For directories it's -EISDIR, for other non-regulars - -EINVAL */ error = -EISDIR; @@ -247,7 +247,7 @@ static long do_sys_truncate(const char __user * path, loff_t length) if (!S_ISREG(inode->i_mode)) goto dput_and_out; - error = mnt_want_write(nd.path.mnt); + error = mnt_want_write(path.mnt); if (error) goto dput_and_out; @@ -274,15 +274,15 @@ static long do_sys_truncate(const char __user * path, loff_t length) error = locks_verify_truncate(inode, NULL, length); if (!error) { DQUOT_INIT(inode); - error = do_truncate(nd.path.dentry, length, 0, NULL); + error = do_truncate(path.dentry, length, 0, NULL); } put_write_and_out: put_write_access(inode); mnt_drop_write_and_out: - mnt_drop_write(nd.path.mnt); + mnt_drop_write(path.mnt); dput_and_out: - path_put(&nd.path); + path_put(&path); out: return error; } @@ -425,7 +425,7 @@ out: */ asmlinkage long sys_faccessat(int dfd, const char __user *filename, int mode) { - struct nameidata nd; + struct path path; struct inode *inode; int old_fsuid, old_fsgid; kernel_cap_t uninitialized_var(old_cap); /* !SECURE_NO_SETUID_FIXUP */ @@ -449,7 +449,7 @@ asmlinkage long sys_faccessat(int dfd, const char __user *filename, int mode) * FIXME: There is a race here against sys_capset. The * capabilities can change yet we will restore the old * value below. We should hold task_capabilities_lock, - * but we cannot because user_path_walk can sleep. + * but we cannot because user_path_at can sleep. */ #endif /* ndef CONFIG_SECURITY_FILE_CAPABILITIES */ if (current->uid) @@ -458,11 +458,11 @@ asmlinkage long sys_faccessat(int dfd, const char __user *filename, int mode) old_cap = cap_set_effective(current->cap_permitted); } - res = __user_walk_fd(dfd, filename, LOOKUP_FOLLOW, &nd); + res = user_path_at(dfd, filename, LOOKUP_FOLLOW, &path); if (res) goto out; - inode = nd.path.dentry->d_inode; + inode = path.dentry->d_inode; if ((mode & MAY_EXEC) && S_ISREG(inode->i_mode)) { /* @@ -470,7 +470,7 @@ asmlinkage long sys_faccessat(int dfd, const char __user *filename, int mode) * with the "noexec" flag. */ res = -EACCES; - if (nd.path.mnt->mnt_flags & MNT_NOEXEC) + if (path.mnt->mnt_flags & MNT_NOEXEC) goto out_path_release; } @@ -488,11 +488,11 @@ asmlinkage long sys_faccessat(int dfd, const char __user *filename, int mode) * inherently racy and know that the fs may change * state before we even see this result. */ - if (__mnt_is_readonly(nd.path.mnt)) + if (__mnt_is_readonly(path.mnt)) res = -EROFS; out_path_release: - path_put(&nd.path); + path_put(&path); out: current->fsuid = old_fsuid; current->fsgid = old_fsgid; @@ -510,21 +510,21 @@ asmlinkage long sys_access(const char __user *filename, int mode) asmlinkage long sys_chdir(const char __user * filename) { - struct nameidata nd; + struct path path; int error; - error = __user_walk(filename, LOOKUP_FOLLOW|LOOKUP_DIRECTORY, &nd); + error = user_path_dir(filename, &path); if (error) goto out; - error = inode_permission(nd.path.dentry->d_inode, MAY_EXEC | MAY_ACCESS); + error = inode_permission(path.dentry->d_inode, MAY_EXEC | MAY_ACCESS); if (error) goto dput_and_out; - set_fs_pwd(current->fs, &nd.path); + set_fs_pwd(current->fs, &path); dput_and_out: - path_put(&nd.path); + path_put(&path); out: return error; } @@ -557,14 +557,14 @@ out: asmlinkage long sys_chroot(const char __user * filename) { - struct nameidata nd; + struct path path; int error; - error = __user_walk(filename, LOOKUP_FOLLOW | LOOKUP_DIRECTORY, &nd); + error = user_path_dir(filename, &path); if (error) goto out; - error = inode_permission(nd.path.dentry->d_inode, MAY_EXEC | MAY_ACCESS); + error = inode_permission(path.dentry->d_inode, MAY_EXEC | MAY_ACCESS); if (error) goto dput_and_out; @@ -572,10 +572,10 @@ asmlinkage long sys_chroot(const char __user * filename) if (!capable(CAP_SYS_CHROOT)) goto dput_and_out; - set_fs_root(current->fs, &nd.path); + set_fs_root(current->fs, &path); error = 0; dput_and_out: - path_put(&nd.path); + path_put(&path); out: return error; } @@ -617,17 +617,17 @@ out: asmlinkage long sys_fchmodat(int dfd, const char __user *filename, mode_t mode) { - struct nameidata nd; - struct inode * inode; + struct path path; + struct inode *inode; int error; struct iattr newattrs; - error = __user_walk_fd(dfd, filename, LOOKUP_FOLLOW, &nd); + error = user_path_at(dfd, filename, LOOKUP_FOLLOW, &path); if (error) goto out; - inode = nd.path.dentry->d_inode; + inode = path.dentry->d_inode; - error = mnt_want_write(nd.path.mnt); + error = mnt_want_write(path.mnt); if (error) goto dput_and_out; mutex_lock(&inode->i_mutex); @@ -635,11 +635,11 @@ asmlinkage long sys_fchmodat(int dfd, const char __user *filename, mode = inode->i_mode; newattrs.ia_mode = (mode & S_IALLUGO) | (inode->i_mode & ~S_IALLUGO); newattrs.ia_valid = ATTR_MODE | ATTR_CTIME; - error = notify_change(nd.path.dentry, &newattrs); + error = notify_change(path.dentry, &newattrs); mutex_unlock(&inode->i_mutex); - mnt_drop_write(nd.path.mnt); + mnt_drop_write(path.mnt); dput_and_out: - path_put(&nd.path); + path_put(&path); out: return error; } @@ -676,19 +676,19 @@ static int chown_common(struct dentry * dentry, uid_t user, gid_t group) asmlinkage long sys_chown(const char __user * filename, uid_t user, gid_t group) { - struct nameidata nd; + struct path path; int error; - error = user_path_walk(filename, &nd); + error = user_path(filename, &path); if (error) goto out; - error = mnt_want_write(nd.path.mnt); + error = mnt_want_write(path.mnt); if (error) goto out_release; - error = chown_common(nd.path.dentry, user, group); - mnt_drop_write(nd.path.mnt); + error = chown_common(path.dentry, user, group); + mnt_drop_write(path.mnt); out_release: - path_put(&nd.path); + path_put(&path); out: return error; } @@ -696,7 +696,7 @@ out: asmlinkage long sys_fchownat(int dfd, const char __user *filename, uid_t user, gid_t group, int flag) { - struct nameidata nd; + struct path path; int error = -EINVAL; int follow; @@ -704,35 +704,35 @@ asmlinkage long sys_fchownat(int dfd, const char __user *filename, uid_t user, goto out; follow = (flag & AT_SYMLINK_NOFOLLOW) ? 0 : LOOKUP_FOLLOW; - error = __user_walk_fd(dfd, filename, follow, &nd); + error = user_path_at(dfd, filename, follow, &path); if (error) goto out; - error = mnt_want_write(nd.path.mnt); + error = mnt_want_write(path.mnt); if (error) goto out_release; - error = chown_common(nd.path.dentry, user, group); - mnt_drop_write(nd.path.mnt); + error = chown_common(path.dentry, user, group); + mnt_drop_write(path.mnt); out_release: - path_put(&nd.path); + path_put(&path); out: return error; } asmlinkage long sys_lchown(const char __user * filename, uid_t user, gid_t group) { - struct nameidata nd; + struct path path; int error; - error = user_path_walk_link(filename, &nd); + error = user_lpath(filename, &path); if (error) goto out; - error = mnt_want_write(nd.path.mnt); + error = mnt_want_write(path.mnt); if (error) goto out_release; - error = chown_common(nd.path.dentry, user, group); - mnt_drop_write(nd.path.mnt); + error = chown_common(path.dentry, user, group); + mnt_drop_write(path.mnt); out_release: - path_put(&nd.path); + path_put(&path); out: return error; } diff --git a/fs/stat.c b/fs/stat.c index 9cf41f719d5..7c46fbeb8b7 100644 --- a/fs/stat.c +++ b/fs/stat.c @@ -57,13 +57,13 @@ EXPORT_SYMBOL(vfs_getattr); int vfs_stat_fd(int dfd, char __user *name, struct kstat *stat) { - struct nameidata nd; + struct path path; int error; - error = __user_walk_fd(dfd, name, LOOKUP_FOLLOW, &nd); + error = user_path_at(dfd, name, LOOKUP_FOLLOW, &path); if (!error) { - error = vfs_getattr(nd.path.mnt, nd.path.dentry, stat); - path_put(&nd.path); + error = vfs_getattr(path.mnt, path.dentry, stat); + path_put(&path); } return error; } @@ -77,13 +77,13 @@ EXPORT_SYMBOL(vfs_stat); int vfs_lstat_fd(int dfd, char __user *name, struct kstat *stat) { - struct nameidata nd; + struct path path; int error; - error = __user_walk_fd(dfd, name, 0, &nd); + error = user_path_at(dfd, name, 0, &path); if (!error) { - error = vfs_getattr(nd.path.mnt, nd.path.dentry, stat); - path_put(&nd.path); + error = vfs_getattr(path.mnt, path.dentry, stat); + path_put(&path); } return error; } @@ -291,29 +291,29 @@ asmlinkage long sys_newfstat(unsigned int fd, struct stat __user *statbuf) return error; } -asmlinkage long sys_readlinkat(int dfd, const char __user *path, +asmlinkage long sys_readlinkat(int dfd, const char __user *pathname, char __user *buf, int bufsiz) { - struct nameidata nd; + struct path path; int error; if (bufsiz <= 0) return -EINVAL; - error = __user_walk_fd(dfd, path, 0, &nd); + error = user_path_at(dfd, pathname, 0, &path); if (!error) { - struct inode *inode = nd.path.dentry->d_inode; + struct inode *inode = path.dentry->d_inode; error = -EINVAL; if (inode->i_op && inode->i_op->readlink) { - error = security_inode_readlink(nd.path.dentry); + error = security_inode_readlink(path.dentry); if (!error) { - touch_atime(nd.path.mnt, nd.path.dentry); - error = inode->i_op->readlink(nd.path.dentry, + touch_atime(path.mnt, path.dentry); + error = inode->i_op->readlink(path.dentry, buf, bufsiz); } } - path_put(&nd.path); + path_put(&path); } return error; } diff --git a/fs/utimes.c b/fs/utimes.c index dc28b782625..6929e3e91d0 100644 --- a/fs/utimes.c +++ b/fs/utimes.c @@ -152,18 +152,18 @@ long do_utimes(int dfd, char __user *filename, struct timespec *times, int flags error = utimes_common(&file->f_path, times); fput(file); } else { - struct nameidata nd; + struct path path; int lookup_flags = 0; if (!(flags & AT_SYMLINK_NOFOLLOW)) lookup_flags |= LOOKUP_FOLLOW; - error = __user_walk_fd(dfd, filename, lookup_flags, &nd); + error = user_path_at(dfd, filename, lookup_flags, &path); if (error) goto out; - error = utimes_common(&nd.path, times); - path_put(&nd.path); + error = utimes_common(&path, times); + path_put(&path); } out: diff --git a/fs/xattr.c b/fs/xattr.c index b96222e05ba..468377e6653 100644 --- a/fs/xattr.c +++ b/fs/xattr.c @@ -252,40 +252,40 @@ setxattr(struct dentry *d, const char __user *name, const void __user *value, } asmlinkage long -sys_setxattr(const char __user *path, const char __user *name, +sys_setxattr(const char __user *pathname, const char __user *name, const void __user *value, size_t size, int flags) { - struct nameidata nd; + struct path path; int error; - error = user_path_walk(path, &nd); + error = user_path(pathname, &path); if (error) return error; - error = mnt_want_write(nd.path.mnt); + error = mnt_want_write(path.mnt); if (!error) { - error = setxattr(nd.path.dentry, name, value, size, flags); - mnt_drop_write(nd.path.mnt); + error = setxattr(path.dentry, name, value, size, flags); + mnt_drop_write(path.mnt); } - path_put(&nd.path); + path_put(&path); return error; } asmlinkage long -sys_lsetxattr(const char __user *path, const char __user *name, +sys_lsetxattr(const char __user *pathname, const char __user *name, const void __user *value, size_t size, int flags) { - struct nameidata nd; + struct path path; int error; - error = user_path_walk_link(path, &nd); + error = user_lpath(pathname, &path); if (error) return error; - error = mnt_want_write(nd.path.mnt); + error = mnt_want_write(path.mnt); if (!error) { - error = setxattr(nd.path.dentry, name, value, size, flags); - mnt_drop_write(nd.path.mnt); + error = setxattr(path.dentry, name, value, size, flags); + mnt_drop_write(path.mnt); } - path_put(&nd.path); + path_put(&path); return error; } @@ -350,32 +350,32 @@ getxattr(struct dentry *d, const char __user *name, void __user *value, } asmlinkage ssize_t -sys_getxattr(const char __user *path, const char __user *name, +sys_getxattr(const char __user *pathname, const char __user *name, void __user *value, size_t size) { - struct nameidata nd; + struct path path; ssize_t error; - error = user_path_walk(path, &nd); + error = user_path(pathname, &path); if (error) return error; - error = getxattr(nd.path.dentry, name, value, size); - path_put(&nd.path); + error = getxattr(path.dentry, name, value, size); + path_put(&path); return error; } asmlinkage ssize_t -sys_lgetxattr(const char __user *path, const char __user *name, void __user *value, +sys_lgetxattr(const char __user *pathname, const char __user *name, void __user *value, size_t size) { - struct nameidata nd; + struct path path; ssize_t error; - error = user_path_walk_link(path, &nd); + error = user_lpath(pathname, &path); if (error) return error; - error = getxattr(nd.path.dentry, name, value, size); - path_put(&nd.path); + error = getxattr(path.dentry, name, value, size); + path_put(&path); return error; } @@ -425,30 +425,30 @@ listxattr(struct dentry *d, char __user *list, size_t size) } asmlinkage ssize_t -sys_listxattr(const char __user *path, char __user *list, size_t size) +sys_listxattr(const char __user *pathname, char __user *list, size_t size) { - struct nameidata nd; + struct path path; ssize_t error; - error = user_path_walk(path, &nd); + error = user_path(pathname, &path); if (error) return error; - error = listxattr(nd.path.dentry, list, size); - path_put(&nd.path); + error = listxattr(path.dentry, list, size); + path_put(&path); return error; } asmlinkage ssize_t -sys_llistxattr(const char __user *path, char __user *list, size_t size) +sys_llistxattr(const char __user *pathname, char __user *list, size_t size) { - struct nameidata nd; + struct path path; ssize_t error; - error = user_path_walk_link(path, &nd); + error = user_lpath(pathname, &path); if (error) return error; - error = listxattr(nd.path.dentry, list, size); - path_put(&nd.path); + error = listxattr(path.dentry, list, size); + path_put(&path); return error; } @@ -486,38 +486,38 @@ removexattr(struct dentry *d, const char __user *name) } asmlinkage long -sys_removexattr(const char __user *path, const char __user *name) +sys_removexattr(const char __user *pathname, const char __user *name) { - struct nameidata nd; + struct path path; int error; - error = user_path_walk(path, &nd); + error = user_path(pathname, &path); if (error) return error; - error = mnt_want_write(nd.path.mnt); + error = mnt_want_write(path.mnt); if (!error) { - error = removexattr(nd.path.dentry, name); - mnt_drop_write(nd.path.mnt); + error = removexattr(path.dentry, name); + mnt_drop_write(path.mnt); } - path_put(&nd.path); + path_put(&path); return error; } asmlinkage long -sys_lremovexattr(const char __user *path, const char __user *name) +sys_lremovexattr(const char __user *pathname, const char __user *name) { - struct nameidata nd; + struct path path; int error; - error = user_path_walk_link(path, &nd); + error = user_lpath(pathname, &path); if (error) return error; - error = mnt_want_write(nd.path.mnt); + error = mnt_want_write(path.mnt); if (!error) { - error = removexattr(nd.path.dentry, name); - mnt_drop_write(nd.path.mnt); + error = removexattr(path.dentry, name); + mnt_drop_write(path.mnt); } - path_put(&nd.path); + path_put(&path); return error; } diff --git a/fs/xfs/linux-2.6/xfs_ioctl.c b/fs/xfs/linux-2.6/xfs_ioctl.c index a42ba9d7115..01939ba2d8d 100644 --- a/fs/xfs/linux-2.6/xfs_ioctl.c +++ b/fs/xfs/linux-2.6/xfs_ioctl.c @@ -84,17 +84,15 @@ xfs_find_handle( switch (cmd) { case XFS_IOC_PATH_TO_FSHANDLE: case XFS_IOC_PATH_TO_HANDLE: { - struct nameidata nd; - int error; - - error = user_path_walk_link((const char __user *)hreq.path, &nd); + struct path path; + int error = user_lpath((const char __user *)hreq.path, &path); if (error) return error; - ASSERT(nd.path.dentry); - ASSERT(nd.path.dentry->d_inode); - inode = igrab(nd.path.dentry->d_inode); - path_put(&nd.path); + ASSERT(path.dentry); + ASSERT(path.dentry->d_inode); + inode = igrab(path.dentry->d_inode); + path_put(&path); break; } diff --git a/include/linux/namei.h b/include/linux/namei.h index 60e35a02f6c..00888ff6950 100644 --- a/include/linux/namei.h +++ b/include/linux/namei.h @@ -54,12 +54,13 @@ enum {LAST_NORM, LAST_ROOT, LAST_DOT, LAST_DOTDOT, LAST_BIND}; #define LOOKUP_OPEN (0x0100) #define LOOKUP_CREATE (0x0200) -extern int __user_walk(const char __user *, unsigned, struct nameidata *); -extern int __user_walk_fd(int dfd, const char __user *, unsigned, struct nameidata *); -#define user_path_walk(name,nd) \ - __user_walk_fd(AT_FDCWD, name, LOOKUP_FOLLOW, nd) -#define user_path_walk_link(name,nd) \ - __user_walk_fd(AT_FDCWD, name, 0, nd) +extern int user_path_at(int, const char __user *, unsigned, struct path *); + +#define user_path(name, path) user_path_at(AT_FDCWD, name, LOOKUP_FOLLOW, path) +#define user_lpath(name, path) user_path_at(AT_FDCWD, name, 0, path) +#define user_path_dir(name, path) \ + user_path_at(AT_FDCWD, name, LOOKUP_FOLLOW | LOOKUP_DIRECTORY, path) + extern int path_lookup(const char *, unsigned, struct nameidata *); extern int vfs_path_lookup(struct dentry *, struct vfsmount *, const char *, unsigned int, struct nameidata *); -- cgit v1.2.3-70-g09d2 From 2ad94ae654f5eb72fd3260b706aea645cf4a7791 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 21 Jul 2008 09:32:51 -0400 Subject: [PATCH] new (local) helper: user_path_parent() Preparation to untangling intents mess: reduce the number of do_path_lookup() callers. Signed-off-by: Al Viro --- fs/namei.c | 137 +++++++++++++++++++++++++------------------------------------ 1 file changed, 57 insertions(+), 80 deletions(-) (limited to 'fs') diff --git a/fs/namei.c b/fs/namei.c index edb5e973f9b..38ceb6e06eb 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -1352,6 +1352,24 @@ int user_path_at(int dfd, const char __user *name, unsigned flags, return err; } +static int user_path_parent(int dfd, const char __user *path, + struct nameidata *nd, char **name) +{ + char *s = getname(path); + int error; + + if (IS_ERR(s)) + return PTR_ERR(s); + + error = do_path_lookup(dfd, s, LOOKUP_PARENT, nd); + if (error) + putname(s); + else + *name = s; + + return error; +} + /* * It's inline, so penalty for filesystems that don't use sticky bit is * minimal. @@ -1989,20 +2007,18 @@ static int may_mknod(mode_t mode) asmlinkage long sys_mknodat(int dfd, const char __user *filename, int mode, unsigned dev) { - int error = 0; - char * tmp; - struct dentry * dentry; + int error; + char *tmp; + struct dentry *dentry; struct nameidata nd; if (S_ISDIR(mode)) return -EPERM; - tmp = getname(filename); - if (IS_ERR(tmp)) - return PTR_ERR(tmp); - error = do_path_lookup(dfd, tmp, LOOKUP_PARENT, &nd); + error = user_path_parent(dfd, filename, &nd, &tmp); if (error) - goto out; + return error; + dentry = lookup_create(&nd, 0); if (IS_ERR(dentry)) { error = PTR_ERR(dentry); @@ -2034,7 +2050,6 @@ out_dput: out_unlock: mutex_unlock(&nd.path.dentry->d_inode->i_mutex); path_put(&nd.path); -out: putname(tmp); return error; @@ -2074,14 +2089,10 @@ asmlinkage long sys_mkdirat(int dfd, const char __user *pathname, int mode) struct dentry *dentry; struct nameidata nd; - tmp = getname(pathname); - error = PTR_ERR(tmp); - if (IS_ERR(tmp)) + error = user_path_parent(dfd, pathname, &nd, &tmp); + if (error) goto out_err; - error = do_path_lookup(dfd, tmp, LOOKUP_PARENT, &nd); - if (error) - goto out; dentry = lookup_create(&nd, 1); error = PTR_ERR(dentry); if (IS_ERR(dentry)) @@ -2099,7 +2110,6 @@ out_dput: out_unlock: mutex_unlock(&nd.path.dentry->d_inode->i_mutex); path_put(&nd.path); -out: putname(tmp); out_err: return error; @@ -2177,13 +2187,9 @@ static long do_rmdir(int dfd, const char __user *pathname) struct dentry *dentry; struct nameidata nd; - name = getname(pathname); - if(IS_ERR(name)) - return PTR_ERR(name); - - error = do_path_lookup(dfd, name, LOOKUP_PARENT, &nd); + error = user_path_parent(dfd, pathname, &nd, &name); if (error) - goto exit; + return error; switch(nd.last_type) { case LAST_DOTDOT: @@ -2212,7 +2218,6 @@ exit2: mutex_unlock(&nd.path.dentry->d_inode->i_mutex); exit1: path_put(&nd.path); -exit: putname(name); return error; } @@ -2261,19 +2266,16 @@ int vfs_unlink(struct inode *dir, struct dentry *dentry) */ static long do_unlinkat(int dfd, const char __user *pathname) { - int error = 0; - char * name; + int error; + char *name; struct dentry *dentry; struct nameidata nd; struct inode *inode = NULL; - name = getname(pathname); - if(IS_ERR(name)) - return PTR_ERR(name); - - error = do_path_lookup(dfd, name, LOOKUP_PARENT, &nd); + error = user_path_parent(dfd, pathname, &nd, &name); if (error) - goto exit; + return error; + error = -EISDIR; if (nd.last_type != LAST_NORM) goto exit1; @@ -2300,7 +2302,6 @@ static long do_unlinkat(int dfd, const char __user *pathname) iput(inode); /* truncate the inode here */ exit1: path_put(&nd.path); -exit: putname(name); return error; @@ -2350,23 +2351,20 @@ int vfs_symlink(struct inode *dir, struct dentry *dentry, const char *oldname) asmlinkage long sys_symlinkat(const char __user *oldname, int newdfd, const char __user *newname) { - int error = 0; - char * from; - char * to; + int error; + char *from; + char *to; struct dentry *dentry; struct nameidata nd; from = getname(oldname); - if(IS_ERR(from)) + if (IS_ERR(from)) return PTR_ERR(from); - to = getname(newname); - error = PTR_ERR(to); - if (IS_ERR(to)) - goto out_putname; - error = do_path_lookup(newdfd, to, LOOKUP_PARENT, &nd); + error = user_path_parent(newdfd, newname, &nd, &to); if (error) - goto out; + goto out_putname; + dentry = lookup_create(&nd, 0); error = PTR_ERR(dentry); if (IS_ERR(dentry)) @@ -2382,7 +2380,6 @@ out_dput: out_unlock: mutex_unlock(&nd.path.dentry->d_inode->i_mutex); path_put(&nd.path); -out: putname(to); out_putname: putname(from); @@ -2449,21 +2446,18 @@ asmlinkage long sys_linkat(int olddfd, const char __user *oldname, struct nameidata nd; struct path old_path; int error; - char * to; + char *to; if ((flags & ~AT_SYMLINK_FOLLOW) != 0) return -EINVAL; - to = getname(newname); - if (IS_ERR(to)) - return PTR_ERR(to); - error = user_path_at(olddfd, oldname, flags & AT_SYMLINK_FOLLOW ? LOOKUP_FOLLOW : 0, &old_path); if (error) - goto exit; - error = do_path_lookup(newdfd, to, LOOKUP_PARENT, &nd); + return error; + + error = user_path_parent(newdfd, newname, &nd, &to); if (error) goto out; error = -EXDEV; @@ -2484,10 +2478,9 @@ out_unlock: mutex_unlock(&nd.path.dentry->d_inode->i_mutex); out_release: path_put(&nd.path); + putname(to); out: path_put(&old_path); -exit: - putname(to); return error; } @@ -2643,20 +2636,22 @@ int vfs_rename(struct inode *old_dir, struct dentry *old_dentry, return error; } -static int do_rename(int olddfd, const char *oldname, - int newdfd, const char *newname) +asmlinkage long sys_renameat(int olddfd, const char __user *oldname, + int newdfd, const char __user *newname) { - int error = 0; - struct dentry * old_dir, * new_dir; - struct dentry * old_dentry, *new_dentry; - struct dentry * trap; + struct dentry *old_dir, *new_dir; + struct dentry *old_dentry, *new_dentry; + struct dentry *trap; struct nameidata oldnd, newnd; + char *from; + char *to; + int error; - error = do_path_lookup(olddfd, oldname, LOOKUP_PARENT, &oldnd); + error = user_path_parent(olddfd, oldname, &oldnd, &from); if (error) goto exit; - error = do_path_lookup(newdfd, newname, LOOKUP_PARENT, &newnd); + error = user_path_parent(newdfd, newname, &newnd, &to); if (error) goto exit1; @@ -2718,29 +2713,11 @@ exit3: unlock_rename(new_dir, old_dir); exit2: path_put(&newnd.path); + putname(to); exit1: path_put(&oldnd.path); -exit: - return error; -} - -asmlinkage long sys_renameat(int olddfd, const char __user *oldname, - int newdfd, const char __user *newname) -{ - int error; - char * from; - char * to; - - from = getname(oldname); - if(IS_ERR(from)) - return PTR_ERR(from); - to = getname(newname); - error = PTR_ERR(to); - if (!IS_ERR(to)) { - error = do_rename(olddfd, from, newdfd, to); - putname(to); - } putname(from); +exit: return error; } -- cgit v1.2.3-70-g09d2 From a569c711f63995ad80c23918525111e0cdb0bc73 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 23 Jul 2008 14:42:05 -0400 Subject: [PATCH] don't pass nameidata to gfs2_lookupi() Signed-off-by: Al Viro --- fs/gfs2/inode.c | 6 +++--- fs/gfs2/inode.h | 2 +- fs/gfs2/ops_export.c | 2 +- fs/gfs2/ops_inode.c | 4 ++-- fs/gfs2/super.c | 2 +- 5 files changed, 8 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c index 6da0ab355b8..8b0806a3294 100644 --- a/fs/gfs2/inode.c +++ b/fs/gfs2/inode.c @@ -448,7 +448,7 @@ struct inode *gfs2_lookup_simple(struct inode *dip, const char *name) struct qstr qstr; struct inode *inode; gfs2_str2qstr(&qstr, name); - inode = gfs2_lookupi(dip, &qstr, 1, NULL); + inode = gfs2_lookupi(dip, &qstr, 1); /* gfs2_lookupi has inconsistent callers: vfs * related routines expect NULL for no entry found, * gfs2_lookup_simple callers expect ENOENT @@ -477,7 +477,7 @@ struct inode *gfs2_lookup_simple(struct inode *dip, const char *name) */ struct inode *gfs2_lookupi(struct inode *dir, const struct qstr *name, - int is_root, struct nameidata *nd) + int is_root) { struct super_block *sb = dir->i_sb; struct gfs2_inode *dip = GFS2_I(dir); @@ -1173,7 +1173,7 @@ int gfs2_ok_to_move(struct gfs2_inode *this, struct gfs2_inode *to) break; } - tmp = gfs2_lookupi(dir, &dotdot, 1, NULL); + tmp = gfs2_lookupi(dir, &dotdot, 1); if (IS_ERR(tmp)) { error = PTR_ERR(tmp); break; diff --git a/fs/gfs2/inode.h b/fs/gfs2/inode.h index 6074c2506f7..58f9607d6a8 100644 --- a/fs/gfs2/inode.h +++ b/fs/gfs2/inode.h @@ -83,7 +83,7 @@ int gfs2_inode_refresh(struct gfs2_inode *ip); int gfs2_dinode_dealloc(struct gfs2_inode *inode); int gfs2_change_nlink(struct gfs2_inode *ip, int diff); struct inode *gfs2_lookupi(struct inode *dir, const struct qstr *name, - int is_root, struct nameidata *nd); + int is_root); struct inode *gfs2_createi(struct gfs2_holder *ghs, const struct qstr *name, unsigned int mode, dev_t dev); int gfs2_rmdiri(struct gfs2_inode *dip, const struct qstr *name, diff --git a/fs/gfs2/ops_export.c b/fs/gfs2/ops_export.c index 990d9f4bc46..9cda8536530 100644 --- a/fs/gfs2/ops_export.c +++ b/fs/gfs2/ops_export.c @@ -134,7 +134,7 @@ static struct dentry *gfs2_get_parent(struct dentry *child) struct dentry *dentry; gfs2_str2qstr(&dotdot, ".."); - inode = gfs2_lookupi(child->d_inode, &dotdot, 1, NULL); + inode = gfs2_lookupi(child->d_inode, &dotdot, 1); if (!inode) return ERR_PTR(-ENOENT); diff --git a/fs/gfs2/ops_inode.c b/fs/gfs2/ops_inode.c index 4e982532f08..e2c62f73a77 100644 --- a/fs/gfs2/ops_inode.c +++ b/fs/gfs2/ops_inode.c @@ -74,7 +74,7 @@ static int gfs2_create(struct inode *dir, struct dentry *dentry, return PTR_ERR(inode); } - inode = gfs2_lookupi(dir, &dentry->d_name, 0, nd); + inode = gfs2_lookupi(dir, &dentry->d_name, 0); if (inode) { if (!IS_ERR(inode)) { gfs2_holder_uninit(ghs); @@ -109,7 +109,7 @@ static struct dentry *gfs2_lookup(struct inode *dir, struct dentry *dentry, dentry->d_op = &gfs2_dops; - inode = gfs2_lookupi(dir, &dentry->d_name, 0, nd); + inode = gfs2_lookupi(dir, &dentry->d_name, 0); if (inode && IS_ERR(inode)) return ERR_CAST(inode); diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c index 63a8a902d9d..ca831991cbc 100644 --- a/fs/gfs2/super.c +++ b/fs/gfs2/super.c @@ -389,7 +389,7 @@ int gfs2_jindex_hold(struct gfs2_sbd *sdp, struct gfs2_holder *ji_gh) break; INIT_LIST_HEAD(&jd->extent_list); - jd->jd_inode = gfs2_lookupi(sdp->sd_jindex, &name, 1, NULL); + jd->jd_inode = gfs2_lookupi(sdp->sd_jindex, &name, 1); if (!jd->jd_inode || IS_ERR(jd->jd_inode)) { if (!jd->jd_inode) error = -ENOENT; -- cgit v1.2.3-70-g09d2 From 58ec42b061bf5dad8fa0370a19966cfd96eaf80c Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 23 Jul 2008 14:45:55 -0400 Subject: [PATCH] don't pass nameidata to __ncp_lookup_validate() Signed-off-by: Al Viro --- fs/ncpfs/dir.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ncpfs/dir.c b/fs/ncpfs/dir.c index 011ef0b6d2d..07e9715b865 100644 --- a/fs/ncpfs/dir.c +++ b/fs/ncpfs/dir.c @@ -266,7 +266,7 @@ leave_me:; static int -__ncp_lookup_validate(struct dentry * dentry, struct nameidata *nd) +__ncp_lookup_validate(struct dentry *dentry) { struct ncp_server *server; struct dentry *parent; @@ -340,7 +340,7 @@ ncp_lookup_validate(struct dentry * dentry, struct nameidata *nd) { int res; lock_kernel(); - res = __ncp_lookup_validate(dentry, nd); + res = __ncp_lookup_validate(dentry); unlock_kernel(); return res; } -- cgit v1.2.3-70-g09d2 From 3c333937ee3be114b181c4861188cfe8f6a59697 Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Fri, 25 Jul 2008 22:32:13 -0400 Subject: [PATCH] dup3 fix Al Viro notice one cornercase that the new dup3() code. The dup2() function, as a special case, handles dup-ing to the same file descriptor. In this case the current dup3() code does nothing at all. I.e., it ingnores the flags parameter. This shouldn't happen, the close-on-exec flag should be set if requested. In case the O_CLOEXEC bit in the flags parameter is not set the dup3() function should behave in this respect identical to dup2(). This means dup3(fd, fd, 0) should not actively reset the c-o-e flag. The patch below implements this minor change. [AV: credits to Artur Grabowski for bringing that up as potential subtle point in dup2() behaviour] Signed-off-by: Ulrich Drepper Signed-off-by: Al Viro --- fs/fcntl.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/fcntl.c b/fs/fcntl.c index 9679fcbdeaa..ce12a688511 100644 --- a/fs/fcntl.c +++ b/fs/fcntl.c @@ -139,8 +139,13 @@ asmlinkage long sys_dup3(unsigned int oldfd, unsigned int newfd, int flags) if (!(file = fcheck(oldfd))) goto out_unlock; err = newfd; - if (newfd == oldfd) + if (unlikely(newfd == oldfd)) { + if (flags & O_CLOEXEC) { + fdt = files_fdtable(files); + FD_SET(newfd, fdt->close_on_exec); + } goto out_unlock; + } err = -EBADF; if (newfd >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur) goto out_unlock; -- cgit v1.2.3-70-g09d2 From 516e0cc5646f377ab80fcc2ee639892eccb99853 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 26 Jul 2008 00:39:17 -0400 Subject: [PATCH] f_count may wrap around make it atomic_long_t; while we are at it, get rid of useless checks in affs, hfs and hpfs - ->open() always has it equal to 1, ->release() - to 0. Signed-off-by: Al Viro --- drivers/net/ppp_generic.c | 6 +++--- fs/affs/file.c | 4 ---- fs/aio.c | 6 +++--- fs/file_table.c | 10 +++++----- fs/hfs/inode.c | 4 ---- fs/hfsplus/inode.c | 4 ---- include/linux/fs.h | 6 +++--- include/net/af_unix.h | 2 +- net/sched/sch_atm.c | 4 ++-- net/unix/af_unix.c | 2 +- net/unix/garbage.c | 18 +++++++++--------- 11 files changed, 27 insertions(+), 39 deletions(-) (limited to 'fs') diff --git a/drivers/net/ppp_generic.c b/drivers/net/ppp_generic.c index 739b3ab7bcc..ddccc074a76 100644 --- a/drivers/net/ppp_generic.c +++ b/drivers/net/ppp_generic.c @@ -581,12 +581,12 @@ static long ppp_ioctl(struct file *file, unsigned int cmd, unsigned long arg) if (file == ppp->owner) ppp_shutdown_interface(ppp); } - if (atomic_read(&file->f_count) <= 2) { + if (atomic_long_read(&file->f_count) <= 2) { ppp_release(NULL, file); err = 0; } else - printk(KERN_DEBUG "PPPIOCDETACH file->f_count=%d\n", - atomic_read(&file->f_count)); + printk(KERN_DEBUG "PPPIOCDETACH file->f_count=%ld\n", + atomic_long_read(&file->f_count)); unlock_kernel(); return err; } diff --git a/fs/affs/file.c b/fs/affs/file.c index 6eac7bdeec9..1377b1240b6 100644 --- a/fs/affs/file.c +++ b/fs/affs/file.c @@ -46,8 +46,6 @@ const struct inode_operations affs_file_inode_operations = { static int affs_file_open(struct inode *inode, struct file *filp) { - if (atomic_read(&filp->f_count) != 1) - return 0; pr_debug("AFFS: open(%lu,%d)\n", inode->i_ino, atomic_read(&AFFS_I(inode)->i_opencnt)); atomic_inc(&AFFS_I(inode)->i_opencnt); @@ -57,8 +55,6 @@ affs_file_open(struct inode *inode, struct file *filp) static int affs_file_release(struct inode *inode, struct file *filp) { - if (atomic_read(&filp->f_count) != 0) - return 0; pr_debug("AFFS: release(%lu, %d)\n", inode->i_ino, atomic_read(&AFFS_I(inode)->i_opencnt)); diff --git a/fs/aio.c b/fs/aio.c index 0051fd94b44..f658441d566 100644 --- a/fs/aio.c +++ b/fs/aio.c @@ -512,8 +512,8 @@ static void aio_fput_routine(struct work_struct *data) */ static int __aio_put_req(struct kioctx *ctx, struct kiocb *req) { - dprintk(KERN_DEBUG "aio_put(%p): f_count=%d\n", - req, atomic_read(&req->ki_filp->f_count)); + dprintk(KERN_DEBUG "aio_put(%p): f_count=%ld\n", + req, atomic_long_read(&req->ki_filp->f_count)); assert_spin_locked(&ctx->ctx_lock); @@ -528,7 +528,7 @@ static int __aio_put_req(struct kioctx *ctx, struct kiocb *req) /* Must be done under the lock to serialise against cancellation. * Call this aio_fput as it duplicates fput via the fput_work. */ - if (unlikely(atomic_dec_and_test(&req->ki_filp->f_count))) { + if (unlikely(atomic_long_dec_and_test(&req->ki_filp->f_count))) { get_ioctx(ctx); spin_lock(&fput_lock); list_add(&req->ki_list, &fput_head); diff --git a/fs/file_table.c b/fs/file_table.c index 83084225b4c..f45a4493f9e 100644 --- a/fs/file_table.c +++ b/fs/file_table.c @@ -120,7 +120,7 @@ struct file *get_empty_filp(void) tsk = current; INIT_LIST_HEAD(&f->f_u.fu_list); - atomic_set(&f->f_count, 1); + atomic_long_set(&f->f_count, 1); rwlock_init(&f->f_owner.lock); f->f_uid = tsk->fsuid; f->f_gid = tsk->fsgid; @@ -219,7 +219,7 @@ EXPORT_SYMBOL(init_file); void fput(struct file *file) { - if (atomic_dec_and_test(&file->f_count)) + if (atomic_long_dec_and_test(&file->f_count)) __fput(file); } @@ -294,7 +294,7 @@ struct file *fget(unsigned int fd) rcu_read_lock(); file = fcheck_files(files, fd); if (file) { - if (!atomic_inc_not_zero(&file->f_count)) { + if (!atomic_long_inc_not_zero(&file->f_count)) { /* File object ref couldn't be taken */ rcu_read_unlock(); return NULL; @@ -326,7 +326,7 @@ struct file *fget_light(unsigned int fd, int *fput_needed) rcu_read_lock(); file = fcheck_files(files, fd); if (file) { - if (atomic_inc_not_zero(&file->f_count)) + if (atomic_long_inc_not_zero(&file->f_count)) *fput_needed = 1; else /* Didn't get the reference, someone's freed */ @@ -341,7 +341,7 @@ struct file *fget_light(unsigned int fd, int *fput_needed) void put_filp(struct file *file) { - if (atomic_dec_and_test(&file->f_count)) { + if (atomic_long_dec_and_test(&file->f_count)) { security_file_free(file); file_kill(file); file_free(file); diff --git a/fs/hfs/inode.c b/fs/hfs/inode.c index aa73f3fd5dd..7e19835efa2 100644 --- a/fs/hfs/inode.c +++ b/fs/hfs/inode.c @@ -522,8 +522,6 @@ static int hfs_file_open(struct inode *inode, struct file *file) { if (HFS_IS_RSRC(inode)) inode = HFS_I(inode)->rsrc_inode; - if (atomic_read(&file->f_count) != 1) - return 0; atomic_inc(&HFS_I(inode)->opencnt); return 0; } @@ -534,8 +532,6 @@ static int hfs_file_release(struct inode *inode, struct file *file) if (HFS_IS_RSRC(inode)) inode = HFS_I(inode)->rsrc_inode; - if (atomic_read(&file->f_count) != 0) - return 0; if (atomic_dec_and_test(&HFS_I(inode)->opencnt)) { mutex_lock(&inode->i_mutex); hfs_file_truncate(inode); diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c index d4014e3044d..b085d64a2b6 100644 --- a/fs/hfsplus/inode.c +++ b/fs/hfsplus/inode.c @@ -254,8 +254,6 @@ static int hfsplus_file_open(struct inode *inode, struct file *file) { if (HFSPLUS_IS_RSRC(inode)) inode = HFSPLUS_I(inode).rsrc_inode; - if (atomic_read(&file->f_count) != 1) - return 0; atomic_inc(&HFSPLUS_I(inode).opencnt); return 0; } @@ -266,8 +264,6 @@ static int hfsplus_file_release(struct inode *inode, struct file *file) if (HFSPLUS_IS_RSRC(inode)) inode = HFSPLUS_I(inode).rsrc_inode; - if (atomic_read(&file->f_count) != 0) - return 0; if (atomic_dec_and_test(&HFSPLUS_I(inode).opencnt)) { mutex_lock(&inode->i_mutex); hfsplus_file_truncate(inode); diff --git a/include/linux/fs.h b/include/linux/fs.h index 9d2de4cadab..7676fa1c20a 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -795,7 +795,7 @@ struct file { #define f_dentry f_path.dentry #define f_vfsmnt f_path.mnt const struct file_operations *f_op; - atomic_t f_count; + atomic_long_t f_count; unsigned int f_flags; mode_t f_mode; loff_t f_pos; @@ -824,8 +824,8 @@ extern spinlock_t files_lock; #define file_list_lock() spin_lock(&files_lock); #define file_list_unlock() spin_unlock(&files_lock); -#define get_file(x) atomic_inc(&(x)->f_count) -#define file_count(x) atomic_read(&(x)->f_count) +#define get_file(x) atomic_long_inc(&(x)->f_count) +#define file_count(x) atomic_long_read(&(x)->f_count) #ifdef CONFIG_DEBUG_WRITECOUNT static inline void file_take_write(struct file *f) diff --git a/include/net/af_unix.h b/include/net/af_unix.h index 2dfa96b0575..7dd29b7e461 100644 --- a/include/net/af_unix.h +++ b/include/net/af_unix.h @@ -51,7 +51,7 @@ struct unix_sock { struct sock *peer; struct sock *other; struct list_head link; - atomic_t inflight; + atomic_long_t inflight; spinlock_t lock; unsigned int gc_candidate : 1; wait_queue_head_t peer_wait; diff --git a/net/sched/sch_atm.c b/net/sched/sch_atm.c index 04faa835be1..6b517b9dac5 100644 --- a/net/sched/sch_atm.c +++ b/net/sched/sch_atm.c @@ -162,7 +162,7 @@ static void atm_tc_put(struct Qdisc *sch, unsigned long cl) qdisc_destroy(flow->q); tcf_destroy_chain(&flow->filter_list); if (flow->sock) { - pr_debug("atm_tc_put: f_count %d\n", + pr_debug("atm_tc_put: f_count %ld\n", file_count(flow->sock->file)); flow->vcc->pop = flow->old_pop; sockfd_put(flow->sock); @@ -259,7 +259,7 @@ static int atm_tc_change(struct Qdisc *sch, u32 classid, u32 parent, sock = sockfd_lookup(fd, &error); if (!sock) return error; /* f_count++ */ - pr_debug("atm_tc_change: f_count %d\n", file_count(sock->file)); + pr_debug("atm_tc_change: f_count %ld\n", file_count(sock->file)); if (sock->ops->family != PF_ATMSVC && sock->ops->family != PF_ATMPVC) { error = -EPROTOTYPE; goto err_out; diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c index 70ceb1604ad..6e7fec74bdb 100644 --- a/net/unix/af_unix.c +++ b/net/unix/af_unix.c @@ -603,7 +603,7 @@ static struct sock * unix_create1(struct net *net, struct socket *sock) u->dentry = NULL; u->mnt = NULL; spin_lock_init(&u->lock); - atomic_set(&u->inflight, 0); + atomic_long_set(&u->inflight, 0); INIT_LIST_HEAD(&u->link); mutex_init(&u->readlock); /* single task reading lock */ init_waitqueue_head(&u->peer_wait); diff --git a/net/unix/garbage.c b/net/unix/garbage.c index ebdff3d877a..2a27b84f740 100644 --- a/net/unix/garbage.c +++ b/net/unix/garbage.c @@ -127,7 +127,7 @@ void unix_inflight(struct file *fp) if(s) { struct unix_sock *u = unix_sk(s); spin_lock(&unix_gc_lock); - if (atomic_inc_return(&u->inflight) == 1) { + if (atomic_long_inc_return(&u->inflight) == 1) { BUG_ON(!list_empty(&u->link)); list_add_tail(&u->link, &gc_inflight_list); } else { @@ -145,7 +145,7 @@ void unix_notinflight(struct file *fp) struct unix_sock *u = unix_sk(s); spin_lock(&unix_gc_lock); BUG_ON(list_empty(&u->link)); - if (atomic_dec_and_test(&u->inflight)) + if (atomic_long_dec_and_test(&u->inflight)) list_del_init(&u->link); unix_tot_inflight--; spin_unlock(&unix_gc_lock); @@ -237,17 +237,17 @@ static void scan_children(struct sock *x, void (*func)(struct unix_sock *), static void dec_inflight(struct unix_sock *usk) { - atomic_dec(&usk->inflight); + atomic_long_dec(&usk->inflight); } static void inc_inflight(struct unix_sock *usk) { - atomic_inc(&usk->inflight); + atomic_long_inc(&usk->inflight); } static void inc_inflight_move_tail(struct unix_sock *u) { - atomic_inc(&u->inflight); + atomic_long_inc(&u->inflight); /* * If this is still a candidate, move it to the end of the * list, so that it's checked even if it was already passed @@ -288,11 +288,11 @@ void unix_gc(void) * before the detach without atomicity guarantees. */ list_for_each_entry_safe(u, next, &gc_inflight_list, link) { - int total_refs; - int inflight_refs; + long total_refs; + long inflight_refs; total_refs = file_count(u->sk.sk_socket->file); - inflight_refs = atomic_read(&u->inflight); + inflight_refs = atomic_long_read(&u->inflight); BUG_ON(inflight_refs < 1); BUG_ON(total_refs < inflight_refs); @@ -324,7 +324,7 @@ void unix_gc(void) /* Move cursor to after the current position. */ list_move(&cursor, &u->link); - if (atomic_read(&u->inflight) > 0) { + if (atomic_long_read(&u->inflight) > 0) { list_move_tail(&u->link, &gc_inflight_list); u->gc_candidate = 0; scan_children(&u->sk, inc_inflight_move_tail, NULL); -- cgit v1.2.3-70-g09d2 From 964bd183624c03680796b63b4ab97ee3905a806a Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 26 Jul 2008 03:33:14 -0400 Subject: [PATCH] get rid of __user_path_lookup_open Signed-off-by: Al Viro --- fs/exec.c | 14 ++++++++++---- fs/namei.c | 13 ------------- include/linux/namei.h | 1 - 3 files changed, 10 insertions(+), 18 deletions(-) (limited to 'fs') diff --git a/fs/exec.c b/fs/exec.c index eca58c29ede..9696bbf0f0b 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -106,11 +106,17 @@ static inline void put_binfmt(struct linux_binfmt * fmt) */ asmlinkage long sys_uselib(const char __user * library) { - struct file * file; + struct file *file; struct nameidata nd; - int error; - - error = __user_path_lookup_open(library, LOOKUP_FOLLOW, &nd, FMODE_READ|FMODE_EXEC); + char *tmp = getname(library); + int error = PTR_ERR(tmp); + + if (!IS_ERR(tmp)) { + error = path_lookup_open(AT_FDCWD, tmp, + LOOKUP_FOLLOW, &nd, + FMODE_READ|FMODE_EXEC); + putname(tmp); + } if (error) goto out; diff --git a/fs/namei.c b/fs/namei.c index 38ceb6e06eb..a7b0a0b8012 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -1193,19 +1193,6 @@ static int path_lookup_create(int dfd, const char *name, nd, open_flags, create_mode); } -int __user_path_lookup_open(const char __user *name, unsigned int lookup_flags, - struct nameidata *nd, int open_flags) -{ - char *tmp = getname(name); - int err = PTR_ERR(tmp); - - if (!IS_ERR(tmp)) { - err = __path_lookup_intent_open(AT_FDCWD, tmp, lookup_flags, nd, open_flags, 0); - putname(tmp); - } - return err; -} - static struct dentry *__lookup_hash(struct qstr *name, struct dentry *base, struct nameidata *nd) { diff --git a/include/linux/namei.h b/include/linux/namei.h index 00888ff6950..68f8c3203c8 100644 --- a/include/linux/namei.h +++ b/include/linux/namei.h @@ -65,7 +65,6 @@ extern int path_lookup(const char *, unsigned, struct nameidata *); extern int vfs_path_lookup(struct dentry *, struct vfsmount *, const char *, unsigned int, struct nameidata *); -extern int __user_path_lookup_open(const char __user *, unsigned lookup_flags, struct nameidata *nd, int open_flags); extern int path_lookup_open(int dfd, const char *name, unsigned lookup_flags, struct nameidata *, int open_flags); extern struct file *lookup_instantiate_filp(struct nameidata *nd, struct dentry *dentry, int (*open)(struct inode *, struct file *)); -- cgit v1.2.3-70-g09d2 From 3f8206d496e9e9495afb1d4e70d29712b4d403c9 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 26 Jul 2008 03:46:43 -0400 Subject: [PATCH] get rid of indirect users of namei.h fs.h needs path.h, not namei.h; nfs_fs.h doesn't need it at all. Several places in the tree needed direct include. Signed-off-by: Al Viro --- fs/nfsd/nfsctl.c | 1 + fs/ubifs/file.c | 1 + include/linux/fs.h | 2 +- include/linux/nfs_fs.h | 1 - kernel/cgroup.c | 1 + 5 files changed, 4 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index 1955a2702e6..c53e65f8f3a 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -12,6 +12,7 @@ #include #include #include +#include #include #include #include diff --git a/fs/ubifs/file.c b/fs/ubifs/file.c index 005a3b854d9..8565e586e53 100644 --- a/fs/ubifs/file.c +++ b/fs/ubifs/file.c @@ -53,6 +53,7 @@ #include "ubifs.h" #include +#include static int read_block(struct inode *inode, void *addr, unsigned int block, struct ubifs_data_node *dn) diff --git a/include/linux/fs.h b/include/linux/fs.h index 7676fa1c20a..8252b045e62 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -279,7 +279,7 @@ extern int dir_notify_enable; #include #include #include -#include +#include #include #include #include diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index f08f9ca602a..78a5922a2f1 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -42,7 +42,6 @@ #include #include #include -#include #include #include #include diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 89bd6fb7894..657f8f8d93a 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -45,6 +45,7 @@ #include #include #include +#include #include -- cgit v1.2.3-70-g09d2 From 6c5d0512a091480c9f981162227fdb1c9d70e555 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 26 Jul 2008 13:38:19 -0400 Subject: [PATCH] get rid of corner case in dup3() entirely Since Ulrich is OK with getting rid of dup3(fd, fd, flags) completely, to hell the damn thing goes. Corner case for dup2() is handled in sys_dup2() (complete with -EBADF if dup2(fd, fd) is called with fd that is not open), the rest is done in dup3(). Signed-off-by: Al Viro --- fs/fcntl.c | 20 +++++++++++--------- 1 file changed, 11 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/fcntl.c b/fs/fcntl.c index ce12a688511..3deec988708 100644 --- a/fs/fcntl.c +++ b/fs/fcntl.c @@ -135,18 +135,12 @@ asmlinkage long sys_dup3(unsigned int oldfd, unsigned int newfd, int flags) if ((flags & ~O_CLOEXEC) != 0) return -EINVAL; + if (unlikely(oldfd == newfd)) + return -EINVAL; + spin_lock(&files->file_lock); if (!(file = fcheck(oldfd))) goto out_unlock; - err = newfd; - if (unlikely(newfd == oldfd)) { - if (flags & O_CLOEXEC) { - fdt = files_fdtable(files); - FD_SET(newfd, fdt->close_on_exec); - } - goto out_unlock; - } - err = -EBADF; if (newfd >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur) goto out_unlock; get_file(file); /* We are now finished with oldfd */ @@ -194,6 +188,14 @@ out_fput: asmlinkage long sys_dup2(unsigned int oldfd, unsigned int newfd) { + if (unlikely(newfd == oldfd)) { /* corner case */ + struct files_struct *files = current->files; + rcu_read_lock(); + if (!fcheck_files(files, oldfd)) + oldfd = -EBADF; + rcu_read_unlock(); + return oldfd; + } return sys_dup3(oldfd, newfd, 0); } -- cgit v1.2.3-70-g09d2 From 4e1e018ecc6f7bfd10fc75b3ff9715cc8164e0a2 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 26 Jul 2008 16:01:20 -0400 Subject: [PATCH] fix RLIM_NOFILE handling * dup2() should return -EBADF on exceeded sysctl_nr_open * dup() should *not* return -EINVAL even if you have rlimit set to 0; it should get -EMFILE instead. Check for orig_start exceeding rlimit taken to sys_fcntl(). Failing expand_files() in dup{2,3}() now gets -EMFILE remapped to -EBADF. Consequently, remaining checks for rlimit are taken to expand_files(). Signed-off-by: Al Viro --- fs/fcntl.c | 18 ++++++------------ fs/file.c | 9 +++++++++ fs/open.c | 9 --------- 3 files changed, 15 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/fs/fcntl.c b/fs/fcntl.c index 3deec988708..61d62513681 100644 --- a/fs/fcntl.c +++ b/fs/fcntl.c @@ -64,11 +64,6 @@ static int locate_fd(unsigned int orig_start, int cloexec) struct fdtable *fdt; spin_lock(&files->file_lock); - - error = -EINVAL; - if (orig_start >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur) - goto out; - repeat: fdt = files_fdtable(files); /* @@ -83,10 +78,6 @@ repeat: if (start < fdt->max_fds) newfd = find_next_zero_bit(fdt->open_fds->fds_bits, fdt->max_fds, start); - - error = -EMFILE; - if (newfd >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur) - goto out; error = expand_files(files, newfd); if (error < 0) @@ -141,13 +132,14 @@ asmlinkage long sys_dup3(unsigned int oldfd, unsigned int newfd, int flags) spin_lock(&files->file_lock); if (!(file = fcheck(oldfd))) goto out_unlock; - if (newfd >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur) - goto out_unlock; get_file(file); /* We are now finished with oldfd */ err = expand_files(files, newfd); - if (err < 0) + if (unlikely(err < 0)) { + if (err == -EMFILE) + err = -EBADF; goto out_fput; + } /* To avoid races with open() and dup(), we will mark the fd as * in-use in the open-file bitmap throughout the entire dup2() @@ -328,6 +320,8 @@ static long do_fcntl(int fd, unsigned int cmd, unsigned long arg, switch (cmd) { case F_DUPFD: case F_DUPFD_CLOEXEC: + if (arg >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur) + break; get_file(filp); err = dupfd(filp, arg, cmd == F_DUPFD_CLOEXEC); break; diff --git a/fs/file.c b/fs/file.c index 7b3887e054d..d8773b19fe4 100644 --- a/fs/file.c +++ b/fs/file.c @@ -250,9 +250,18 @@ int expand_files(struct files_struct *files, int nr) struct fdtable *fdt; fdt = files_fdtable(files); + + /* + * N.B. For clone tasks sharing a files structure, this test + * will limit the total number of files that can be opened. + */ + if (nr >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur) + return -EMFILE; + /* Do we need to expand? */ if (nr < fdt->max_fds) return 0; + /* Can we expand? */ if (nr >= sysctl_nr_open) return -EMFILE; diff --git a/fs/open.c b/fs/open.c index 3fe1a6857c7..52647be277a 100644 --- a/fs/open.c +++ b/fs/open.c @@ -972,7 +972,6 @@ int get_unused_fd_flags(int flags) int fd, error; struct fdtable *fdt; - error = -EMFILE; spin_lock(&files->file_lock); repeat: @@ -980,13 +979,6 @@ repeat: fd = find_next_zero_bit(fdt->open_fds->fds_bits, fdt->max_fds, files->next_fd); - /* - * N.B. For clone tasks sharing a files structure, this test - * will limit the total number of files that can be opened. - */ - if (fd >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur) - goto out; - /* Do we need to expand the fd array or fd set? */ error = expand_files(files, fd); if (error < 0) @@ -997,7 +989,6 @@ repeat: * If we needed to expand the fs array we * might have blocked - try again. */ - error = -EMFILE; goto repeat; } -- cgit v1.2.3-70-g09d2 From b2d002dba5a8a4c0c3ec96fd1ff3c9def6bd71a1 Mon Sep 17 00:00:00 2001 From: Andrea Righi Date: Sat, 26 Jul 2008 15:22:27 -0700 Subject: task IO accounting: correctly account threads IO statistics Oleg Nesterov points out that we should check that the task is still alive before we iterate over the threads. This patch includes a fixup for this. Also simplify do_io_accounting() implementation. Signed-off-by: Andrea Righi Signed-off-by: Oleg Nesterov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/base.c | 56 ++++++++++++++++++++++---------------------------------- 1 file changed, 22 insertions(+), 34 deletions(-) (limited to 'fs') diff --git a/fs/proc/base.c b/fs/proc/base.c index 81bce6791bf..d744aa3c9f7 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -2406,35 +2406,18 @@ static int do_io_accounting(struct task_struct *task, char *buffer, int whole) u64 rchar, wchar, syscr, syscw; struct task_io_accounting ioac; - if (!whole) { - rchar = task->rchar; - wchar = task->wchar; - syscr = task->syscr; - syscw = task->syscw; - memcpy(&ioac, &task->ioac, sizeof(ioac)); - } else { - unsigned long flags; - struct task_struct *t = task; - rchar = wchar = syscr = syscw = 0; - memset(&ioac, 0, sizeof(ioac)); + rchar = task->rchar; + wchar = task->wchar; + syscr = task->syscr; + syscw = task->syscw; + memcpy(&ioac, &task->ioac, sizeof(ioac)); - rcu_read_lock(); - do { - rchar += t->rchar; - wchar += t->wchar; - syscr += t->syscr; - syscw += t->syscw; - - ioac.read_bytes += t->ioac.read_bytes; - ioac.write_bytes += t->ioac.write_bytes; - ioac.cancelled_write_bytes += - t->ioac.cancelled_write_bytes; - t = next_thread(t); - } while (t != task); - rcu_read_unlock(); + if (whole) { + unsigned long flags; if (lock_task_sighand(task, &flags)) { struct signal_struct *sig = task->signal; + struct task_struct *t = task; rchar += sig->rchar; wchar += sig->wchar; @@ -2445,11 +2428,20 @@ static int do_io_accounting(struct task_struct *task, char *buffer, int whole) ioac.write_bytes += sig->ioac.write_bytes; ioac.cancelled_write_bytes += sig->ioac.cancelled_write_bytes; - + while_each_thread(task, t) { + rchar += t->rchar; + wchar += t->wchar; + syscr += t->syscr; + syscw += t->syscw; + + ioac.read_bytes += t->ioac.read_bytes; + ioac.write_bytes += t->ioac.write_bytes; + ioac.cancelled_write_bytes += + t->ioac.cancelled_write_bytes; + } unlock_task_sighand(task, &flags); } } - return sprintf(buffer, "rchar: %llu\n" "wchar: %llu\n" @@ -2458,13 +2450,9 @@ static int do_io_accounting(struct task_struct *task, char *buffer, int whole) "read_bytes: %llu\n" "write_bytes: %llu\n" "cancelled_write_bytes: %llu\n", - (unsigned long long)rchar, - (unsigned long long)wchar, - (unsigned long long)syscr, - (unsigned long long)syscw, - (unsigned long long)ioac.read_bytes, - (unsigned long long)ioac.write_bytes, - (unsigned long long)ioac.cancelled_write_bytes); + rchar, wchar, syscr, syscw, + ioac.read_bytes, ioac.write_bytes, + ioac.cancelled_write_bytes); } static int proc_tid_io_accounting(struct task_struct *task, char *buffer) -- cgit v1.2.3-70-g09d2