diff options
Diffstat (limited to 'fs/cifs/cifsfs.c')
| -rw-r--r-- | fs/cifs/cifsfs.c | 1449 |
1 files changed, 877 insertions, 572 deletions
diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c index 877095a1192..88839806742 100644 --- a/fs/cifs/cifsfs.c +++ b/fs/cifs/cifsfs.c @@ -1,7 +1,7 @@ /* * fs/cifs/cifsfs.c * - * Copyright (C) International Business Machines Corp., 2002,2004 + * Copyright (C) International Business Machines Corp., 2002,2008 * Author(s): Steve French (sfrench@us.ibm.com) * * Common Internet FileSystem (CIFS) client @@ -32,6 +32,12 @@ #include <linux/seq_file.h> #include <linux/vfs.h> #include <linux/mempool.h> +#include <linux/delay.h> +#include <linux/kthread.h> +#include <linux/freezer.h> +#include <linux/namei.h> +#include <linux/random.h> +#include <net/ipv6.h> #include "cifsfs.h" #include "cifspdu.h" #define DECLARE_GLOBALS_HERE @@ -40,205 +46,189 @@ #include "cifs_debug.h" #include "cifs_fs_sb.h" #include <linux/mm.h> -#define CIFS_MAGIC_NUMBER 0xFF534D42 /* the first four bytes of SMB PDUs */ - -#ifdef CONFIG_CIFS_QUOTA -static struct quotactl_ops cifs_quotactl_ops; +#include <linux/key-type.h> +#include "cifs_spnego.h" +#include "fscache.h" +#ifdef CONFIG_CIFS_SMB2 +#include "smb2pdu.h" #endif int cifsFYI = 0; -int cifsERROR = 1; int traceSMB = 0; -unsigned int oplockEnabled = 1; -unsigned int experimEnabled = 0; +bool enable_oplocks = true; unsigned int linuxExtEnabled = 1; unsigned int lookupCacheEnabled = 1; -unsigned int multiuser_mount = 0; -unsigned int extended_security = 0; -unsigned int ntlmv2_support = 0; +unsigned int global_secflags = CIFSSEC_DEF; +/* unsigned int ntlmv2_support = 0; */ unsigned int sign_CIFS_PDUs = 1; -extern struct task_struct * oplockThread; /* remove sparse warning */ -struct task_struct * oplockThread = NULL; -extern struct task_struct * dnotifyThread; /* remove sparse warning */ -struct task_struct * dnotifyThread = NULL; +static const struct super_operations cifs_super_ops; unsigned int CIFSMaxBufSize = CIFS_MAX_MSGSIZE; -module_param(CIFSMaxBufSize, int, 0); -MODULE_PARM_DESC(CIFSMaxBufSize,"Network buffer size (not including header). Default: 16384 Range: 8192 to 130048"); +module_param(CIFSMaxBufSize, uint, 0); +MODULE_PARM_DESC(CIFSMaxBufSize, "Network buffer size (not including header). " + "Default: 16384 Range: 8192 to 130048"); unsigned int cifs_min_rcv = CIFS_MIN_RCV_POOL; -module_param(cifs_min_rcv, int, 0); -MODULE_PARM_DESC(cifs_min_rcv,"Network buffers in pool. Default: 4 Range: 1 to 64"); +module_param(cifs_min_rcv, uint, 0); +MODULE_PARM_DESC(cifs_min_rcv, "Network buffers in pool. Default: 4 Range: " + "1 to 64"); unsigned int cifs_min_small = 30; -module_param(cifs_min_small, int, 0); -MODULE_PARM_DESC(cifs_min_small,"Small network buffers in pool. Default: 30 Range: 2 to 256"); +module_param(cifs_min_small, uint, 0); +MODULE_PARM_DESC(cifs_min_small, "Small network buffers in pool. Default: 30 " + "Range: 2 to 256"); unsigned int cifs_max_pending = CIFS_MAX_REQ; -module_param(cifs_max_pending, int, 0); -MODULE_PARM_DESC(cifs_max_pending,"Simultaneous requests to server. Default: 50 Range: 2 to 256"); - -static DECLARE_COMPLETION(cifs_oplock_exited); -static DECLARE_COMPLETION(cifs_dnotify_exited); +module_param(cifs_max_pending, uint, 0444); +MODULE_PARM_DESC(cifs_max_pending, "Simultaneous requests to server. " + "Default: 32767 Range: 2 to 32767."); +module_param(enable_oplocks, bool, 0644); +MODULE_PARM_DESC(enable_oplocks, "Enable or disable oplocks. Default: y/Y/1"); extern mempool_t *cifs_sm_req_poolp; extern mempool_t *cifs_req_poolp; extern mempool_t *cifs_mid_poolp; -extern kmem_cache_t *cifs_oplock_cachep; +struct workqueue_struct *cifsiod_wq; + +/* + * Bumps refcount for cifs super block. + * Note that it should be only called if a referece to VFS super block is + * already held, e.g. in open-type syscalls context. Otherwise it can race with + * atomic_dec_and_test in deactivate_locked_super. + */ +void +cifs_sb_active(struct super_block *sb) +{ + struct cifs_sb_info *server = CIFS_SB(sb); + + if (atomic_inc_return(&server->active) == 1) + atomic_inc(&sb->s_active); +} + +void +cifs_sb_deactive(struct super_block *sb) +{ + struct cifs_sb_info *server = CIFS_SB(sb); + + if (atomic_dec_and_test(&server->active)) + deactivate_super(sb); +} static int -cifs_read_super(struct super_block *sb, void *data, - const char *devname, int silent) +cifs_read_super(struct super_block *sb) { struct inode *inode; struct cifs_sb_info *cifs_sb; + struct cifs_tcon *tcon; int rc = 0; - sb->s_flags |= MS_NODIRATIME; /* and probably even noatime */ - sb->s_fs_info = kmalloc(sizeof(struct cifs_sb_info),GFP_KERNEL); cifs_sb = CIFS_SB(sb); - if(cifs_sb == NULL) - return -ENOMEM; - else - memset(cifs_sb,0,sizeof(struct cifs_sb_info)); - + tcon = cifs_sb_master_tcon(cifs_sb); - rc = cifs_mount(sb, cifs_sb, data, devname); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_POSIXACL) + sb->s_flags |= MS_POSIXACL; - if (rc) { - if (!silent) - cERROR(1, - ("cifs_mount failed w/return code = %d", rc)); - goto out_mount_failed; - } + if (tcon->ses->capabilities & tcon->ses->server->vals->cap_large_files) + sb->s_maxbytes = MAX_LFS_FILESIZE; + else + sb->s_maxbytes = MAX_NON_LFS; + + /* BB FIXME fix time_gran to be larger for LANMAN sessions */ + sb->s_time_gran = 100; sb->s_magic = CIFS_MAGIC_NUMBER; sb->s_op = &cifs_super_ops; -/* if(cifs_sb->tcon->ses->server->maxBuf > MAX_CIFS_HDR_SIZE + 512) - sb->s_blocksize = cifs_sb->tcon->ses->server->maxBuf - MAX_CIFS_HDR_SIZE; */ -#ifdef CONFIG_CIFS_QUOTA - sb->s_qcop = &cifs_quotactl_ops; -#endif + sb->s_bdi = &cifs_sb->bdi; sb->s_blocksize = CIFS_MAX_MSGSIZE; sb->s_blocksize_bits = 14; /* default 2**14 = CIFS_MAX_MSGSIZE */ - inode = iget(sb, ROOT_I); + inode = cifs_root_iget(sb); - if (!inode) { - rc = -ENOMEM; + if (IS_ERR(inode)) { + rc = PTR_ERR(inode); goto out_no_root; } - sb->s_root = d_alloc_root(inode); + if (tcon->nocase) + sb->s_d_op = &cifs_ci_dentry_ops; + else + sb->s_d_op = &cifs_dentry_ops; + sb->s_root = d_make_root(inode); if (!sb->s_root) { rc = -ENOMEM; goto out_no_root; } +#ifdef CONFIG_CIFS_NFSD_EXPORT + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_SERVER_INUM) { + cifs_dbg(FYI, "export ops supported\n"); + sb->s_export_op = &cifs_export_ops; + } +#endif /* CONFIG_CIFS_NFSD_EXPORT */ + return 0; out_no_root: - cERROR(1, ("cifs_read_super: get root inode failed")); - if (inode) - iput(inode); - -out_mount_failed: - if(cifs_sb) { - if(cifs_sb->local_nls) - unload_nls(cifs_sb->local_nls); - kfree(cifs_sb); - } + cifs_dbg(VFS, "%s: get root inode failed\n", __func__); return rc; } -static void -cifs_put_super(struct super_block *sb) +static void cifs_kill_sb(struct super_block *sb) { - int rc = 0; - struct cifs_sb_info *cifs_sb; - - cFYI(1, ("In cifs_put_super")); - cifs_sb = CIFS_SB(sb); - if(cifs_sb == NULL) { - cFYI(1,("Empty cifs superblock info passed to unmount")); - return; - } - rc = cifs_umount(sb, cifs_sb); - if (rc) { - cERROR(1, ("cifs_umount failed with return code %d", rc)); - } - unload_nls(cifs_sb->local_nls); - kfree(cifs_sb); - return; + struct cifs_sb_info *cifs_sb = CIFS_SB(sb); + kill_anon_super(sb); + cifs_umount(cifs_sb); } static int -cifs_statfs(struct super_block *sb, struct kstatfs *buf) +cifs_statfs(struct dentry *dentry, struct kstatfs *buf) { - int xid; - int rc = -EOPNOTSUPP; - struct cifs_sb_info *cifs_sb; - struct cifsTconInfo *pTcon; - - xid = GetXid(); - - cifs_sb = CIFS_SB(sb); - pTcon = cifs_sb->tcon; + struct super_block *sb = dentry->d_sb; + struct cifs_sb_info *cifs_sb = CIFS_SB(sb); + struct cifs_tcon *tcon = cifs_sb_master_tcon(cifs_sb); + struct TCP_Server_Info *server = tcon->ses->server; + unsigned int xid; + int rc = 0; - buf->f_type = CIFS_MAGIC_NUMBER; + xid = get_xid(); - /* instead could get the real value via SMB_QUERY_FS_ATTRIBUTE_INFO */ - buf->f_namelen = PATH_MAX; /* PATH_MAX may be too long - it would - presumably be total path, but note - that some servers (includinng Samba 3) - have a shorter maximum path */ + /* + * PATH_MAX may be too long - it would presumably be total path, + * but note that some servers (includinng Samba 3) have a shorter + * maximum path. + * + * Instead could get the real value via SMB_QUERY_FS_ATTRIBUTE_INFO. + */ + buf->f_namelen = PATH_MAX; buf->f_files = 0; /* undefined */ buf->f_ffree = 0; /* unlimited */ -#ifdef CONFIG_CIFS_EXPERIMENTAL -/* BB we could add a second check for a QFS Unix capability bit */ -/* BB FIXME check CIFS_POSIX_EXTENSIONS Unix cap first FIXME BB */ - if ((pTcon->ses->capabilities & CAP_UNIX) && (CIFS_POSIX_EXTENSIONS & - le64_to_cpu(pTcon->fsUnixInfo.Capability))) - rc = CIFSSMBQFSPosixInfo(xid, pTcon, buf); - - /* Only need to call the old QFSInfo if failed - on newer one */ - if(rc) -#endif /* CIFS_EXPERIMENTAL */ - rc = CIFSSMBQFSInfo(xid, pTcon, buf); - - /* Old Windows servers do not support level 103, retry with level - one if old server failed the previous call */ - if(rc) - rc = SMBOldQFSInfo(xid, pTcon, buf); - /* - int f_type; - __fsid_t f_fsid; - int f_namelen; */ - /* BB get from info in tcon struct at mount time call to QFSAttrInfo */ - FreeXid(xid); - return 0; /* always return success? what if volume is no - longer available? */ -} - -static int cifs_permission(struct inode * inode, int mask, struct nameidata *nd) + if (server->ops->queryfs) + rc = server->ops->queryfs(xid, tcon, buf); + + free_xid(xid); + return 0; +} + +static int cifs_permission(struct inode *inode, int mask) { struct cifs_sb_info *cifs_sb; cifs_sb = CIFS_SB(inode->i_sb); if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NO_PERM) { - return 0; - } else /* file mode might have been restricted at mount time - on the client (above and beyond ACL on servers) for + if ((mask & MAY_EXEC) && !execute_ok(inode)) + return -EACCES; + else + return 0; + } else /* file mode might have been restricted at mount time + on the client (above and beyond ACL on servers) for servers which do not support setting and viewing mode bits, - so allowing client to check permissions is useful */ - return generic_permission(inode, mask, NULL); + so allowing client to check permissions is useful */ + return generic_permission(inode, mask); } -static kmem_cache_t *cifs_inode_cachep; -static kmem_cache_t *cifs_req_cachep; -static kmem_cache_t *cifs_mid_cachep; -kmem_cache_t *cifs_oplock_cachep; -static kmem_cache_t *cifs_sm_req_cachep; +static struct kmem_cache *cifs_inode_cachep; +static struct kmem_cache *cifs_req_cachep; +static struct kmem_cache *cifs_mid_cachep; +static struct kmem_cache *cifs_sm_req_cachep; mempool_t *cifs_sm_req_poolp; mempool_t *cifs_req_poolp; mempool_t *cifs_mid_poolp; @@ -247,297 +237,611 @@ static struct inode * cifs_alloc_inode(struct super_block *sb) { struct cifsInodeInfo *cifs_inode; - cifs_inode = kmem_cache_alloc(cifs_inode_cachep, SLAB_KERNEL); + cifs_inode = kmem_cache_alloc(cifs_inode_cachep, GFP_KERNEL); if (!cifs_inode) return NULL; cifs_inode->cifsAttrs = 0x20; /* default */ - atomic_set(&cifs_inode->inUse, 0); cifs_inode->time = 0; - /* Until the file is open and we have gotten oplock - info back from the server, can not assume caching of - file data or metadata */ - cifs_inode->clientCanCacheRead = FALSE; - cifs_inode->clientCanCacheAll = FALSE; - cifs_inode->vfs_inode.i_blksize = CIFS_MAX_MSGSIZE; + /* + * Until the file is open and we have gotten oplock info back from the + * server, can not assume caching of file data or metadata. + */ + cifs_set_oplock_level(cifs_inode, 0); + cifs_inode->flags = 0; + spin_lock_init(&cifs_inode->writers_lock); + cifs_inode->writers = 0; cifs_inode->vfs_inode.i_blkbits = 14; /* 2**14 = CIFS_MAX_MSGSIZE */ - cifs_inode->vfs_inode.i_flags = S_NOATIME | S_NOCMTIME; + cifs_inode->server_eof = 0; + cifs_inode->uniqueid = 0; + cifs_inode->createtime = 0; + cifs_inode->epoch = 0; +#ifdef CONFIG_CIFS_SMB2 + get_random_bytes(cifs_inode->lease_key, SMB2_LEASE_KEY_SIZE); +#endif + /* + * Can not set i_flags here - they get immediately overwritten to zero + * by the VFS. + */ + /* cifs_inode->vfs_inode.i_flags = S_NOATIME | S_NOCMTIME; */ INIT_LIST_HEAD(&cifs_inode->openFileList); + INIT_LIST_HEAD(&cifs_inode->llist); return &cifs_inode->vfs_inode; } +static void cifs_i_callback(struct rcu_head *head) +{ + struct inode *inode = container_of(head, struct inode, i_rcu); + kmem_cache_free(cifs_inode_cachep, CIFS_I(inode)); +} + static void cifs_destroy_inode(struct inode *inode) { - kmem_cache_free(cifs_inode_cachep, CIFS_I(inode)); + call_rcu(&inode->i_rcu, cifs_i_callback); } -/* - * cifs_show_options() is for displaying mount options in /proc/mounts. - * Not all settable options are displayed but most of the important - * ones are. - */ -static int -cifs_show_options(struct seq_file *s, struct vfsmount *m) +static void +cifs_evict_inode(struct inode *inode) { - struct cifs_sb_info *cifs_sb; + truncate_inode_pages_final(&inode->i_data); + clear_inode(inode); + cifs_fscache_release_inode_cookie(inode); +} - cifs_sb = CIFS_SB(m->mnt_sb); - - if (cifs_sb) { - if (cifs_sb->tcon) { - seq_printf(s, ",unc=%s", cifs_sb->tcon->treeName); - if (cifs_sb->tcon->ses) { - if (cifs_sb->tcon->ses->userName) - seq_printf(s, ",username=%s", - cifs_sb->tcon->ses->userName); - if(cifs_sb->tcon->ses->domainName) - seq_printf(s, ",domain=%s", - cifs_sb->tcon->ses->domainName); - } - } - seq_printf(s, ",rsize=%d",cifs_sb->rsize); - seq_printf(s, ",wsize=%d",cifs_sb->wsize); +static void +cifs_show_address(struct seq_file *s, struct TCP_Server_Info *server) +{ + struct sockaddr_in *sa = (struct sockaddr_in *) &server->dstaddr; + struct sockaddr_in6 *sa6 = (struct sockaddr_in6 *) &server->dstaddr; + + seq_puts(s, ",addr="); + + switch (server->dstaddr.ss_family) { + case AF_INET: + seq_printf(s, "%pI4", &sa->sin_addr.s_addr); + break; + case AF_INET6: + seq_printf(s, "%pI6", &sa6->sin6_addr.s6_addr); + if (sa6->sin6_scope_id) + seq_printf(s, "%%%u", sa6->sin6_scope_id); + break; + default: + seq_puts(s, "(unknown)"); } - return 0; } -#ifdef CONFIG_CIFS_QUOTA -int cifs_xquota_set(struct super_block * sb, int quota_type, qid_t qid, - struct fs_disk_quota * pdquota) +static void +cifs_show_security(struct seq_file *s, struct cifs_ses *ses) { - int xid; - int rc = 0; - struct cifs_sb_info *cifs_sb = CIFS_SB(sb); - struct cifsTconInfo *pTcon; - - if(cifs_sb) - pTcon = cifs_sb->tcon; - else - return -EIO; - + if (ses->sectype == Unspecified) + return; - xid = GetXid(); - if(pTcon) { - cFYI(1,("set type: 0x%x id: %d",quota_type,qid)); - } else { - return -EIO; + seq_puts(s, ",sec="); + + switch (ses->sectype) { + case LANMAN: + seq_puts(s, "lanman"); + break; + case NTLMv2: + seq_puts(s, "ntlmv2"); + break; + case NTLM: + seq_puts(s, "ntlm"); + break; + case Kerberos: + seq_puts(s, "krb5"); + break; + case RawNTLMSSP: + seq_puts(s, "ntlmssp"); + break; + default: + /* shouldn't ever happen */ + seq_puts(s, "unknown"); + break; } - FreeXid(xid); - return rc; + if (ses->sign) + seq_puts(s, "i"); } -int cifs_xquota_get(struct super_block * sb, int quota_type, qid_t qid, - struct fs_disk_quota * pdquota) +static void +cifs_show_cache_flavor(struct seq_file *s, struct cifs_sb_info *cifs_sb) { - int xid; - int rc = 0; - struct cifs_sb_info *cifs_sb = CIFS_SB(sb); - struct cifsTconInfo *pTcon; + seq_puts(s, ",cache="); - if(cifs_sb) - pTcon = cifs_sb->tcon; + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_STRICT_IO) + seq_puts(s, "strict"); + else if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_DIRECT_IO) + seq_puts(s, "none"); else - return -EIO; + seq_puts(s, "loose"); +} - xid = GetXid(); - if(pTcon) { - cFYI(1,("set type: 0x%x id: %d",quota_type,qid)); - } else { - rc = -EIO; - } +static void +cifs_show_nls(struct seq_file *s, struct nls_table *cur) +{ + struct nls_table *def; - FreeXid(xid); - return rc; + /* Display iocharset= option if it's not default charset */ + def = load_nls_default(); + if (def != cur) + seq_printf(s, ",iocharset=%s", cur->charset); + unload_nls(def); } -int cifs_xstate_set(struct super_block * sb, unsigned int flags, int operation) +/* + * cifs_show_options() is for displaying mount options in /proc/mounts. + * Not all settable options are displayed but most of the important + * ones are. + */ +static int +cifs_show_options(struct seq_file *s, struct dentry *root) { - int xid; - int rc = 0; - struct cifs_sb_info *cifs_sb = CIFS_SB(sb); - struct cifsTconInfo *pTcon; + struct cifs_sb_info *cifs_sb = CIFS_SB(root->d_sb); + struct cifs_tcon *tcon = cifs_sb_master_tcon(cifs_sb); + struct sockaddr *srcaddr; + srcaddr = (struct sockaddr *)&tcon->ses->server->srcaddr; + + seq_printf(s, ",vers=%s", tcon->ses->server->vals->version_string); + cifs_show_security(s, tcon->ses); + cifs_show_cache_flavor(s, cifs_sb); + + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MULTIUSER) + seq_puts(s, ",multiuser"); + else if (tcon->ses->user_name) + seq_printf(s, ",username=%s", tcon->ses->user_name); + + if (tcon->ses->domainName) + seq_printf(s, ",domain=%s", tcon->ses->domainName); + + if (srcaddr->sa_family != AF_UNSPEC) { + struct sockaddr_in *saddr4; + struct sockaddr_in6 *saddr6; + saddr4 = (struct sockaddr_in *)srcaddr; + saddr6 = (struct sockaddr_in6 *)srcaddr; + if (srcaddr->sa_family == AF_INET6) + seq_printf(s, ",srcaddr=%pI6c", + &saddr6->sin6_addr); + else if (srcaddr->sa_family == AF_INET) + seq_printf(s, ",srcaddr=%pI4", + &saddr4->sin_addr.s_addr); + else + seq_printf(s, ",srcaddr=BAD-AF:%i", + (int)(srcaddr->sa_family)); + } - if(cifs_sb) - pTcon = cifs_sb->tcon; + seq_printf(s, ",uid=%u", + from_kuid_munged(&init_user_ns, cifs_sb->mnt_uid)); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_OVERR_UID) + seq_puts(s, ",forceuid"); else - return -EIO; + seq_puts(s, ",noforceuid"); - xid = GetXid(); - if(pTcon) { - cFYI(1,("flags: 0x%x operation: 0x%x",flags,operation)); - } else { - rc = -EIO; - } + seq_printf(s, ",gid=%u", + from_kgid_munged(&init_user_ns, cifs_sb->mnt_gid)); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_OVERR_GID) + seq_puts(s, ",forcegid"); + else + seq_puts(s, ",noforcegid"); - FreeXid(xid); - return rc; -} + cifs_show_address(s, tcon->ses->server); -int cifs_xstate_get(struct super_block * sb, struct fs_quota_stat *qstats) -{ - int xid; - int rc = 0; - struct cifs_sb_info *cifs_sb = CIFS_SB(sb); - struct cifsTconInfo *pTcon; + if (!tcon->unix_ext) + seq_printf(s, ",file_mode=0%ho,dir_mode=0%ho", + cifs_sb->mnt_file_mode, + cifs_sb->mnt_dir_mode); - if(cifs_sb) { - pTcon = cifs_sb->tcon; - } else { - return -EIO; - } - xid = GetXid(); - if(pTcon) { - cFYI(1,("pqstats %p",qstats)); - } else { - rc = -EIO; - } + cifs_show_nls(s, cifs_sb->local_nls); - FreeXid(xid); - return rc; -} + if (tcon->seal) + seq_puts(s, ",seal"); + if (tcon->nocase) + seq_puts(s, ",nocase"); + if (tcon->retry) + seq_puts(s, ",hard"); + if (tcon->unix_ext) + seq_puts(s, ",unix"); + else + seq_puts(s, ",nounix"); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_POSIX_PATHS) + seq_puts(s, ",posixpaths"); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_SET_UID) + seq_puts(s, ",setuids"); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_SERVER_INUM) + seq_puts(s, ",serverino"); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_RWPIDFORWARD) + seq_puts(s, ",rwpidforward"); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NOPOSIXBRL) + seq_puts(s, ",forcemand"); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NO_XATTR) + seq_puts(s, ",nouser_xattr"); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MAP_SPECIAL_CHR) + seq_puts(s, ",mapchars"); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_UNX_EMUL) + seq_puts(s, ",sfu"); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NO_BRL) + seq_puts(s, ",nobrl"); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_CIFS_ACL) + seq_puts(s, ",cifsacl"); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_DYNPERM) + seq_puts(s, ",dynperm"); + if (root->d_sb->s_flags & MS_POSIXACL) + seq_puts(s, ",acl"); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MF_SYMLINKS) + seq_puts(s, ",mfsymlinks"); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_FSCACHE) + seq_puts(s, ",fsc"); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NOSSYNC) + seq_puts(s, ",nostrictsync"); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NO_PERM) + seq_puts(s, ",noperm"); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_CIFS_BACKUPUID) + seq_printf(s, ",backupuid=%u", + from_kuid_munged(&init_user_ns, + cifs_sb->mnt_backupuid)); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_CIFS_BACKUPGID) + seq_printf(s, ",backupgid=%u", + from_kgid_munged(&init_user_ns, + cifs_sb->mnt_backupgid)); + + seq_printf(s, ",rsize=%u", cifs_sb->rsize); + seq_printf(s, ",wsize=%u", cifs_sb->wsize); + /* convert actimeo and display it in seconds */ + seq_printf(s, ",actimeo=%lu", cifs_sb->actimeo / HZ); -static struct quotactl_ops cifs_quotactl_ops = { - .set_xquota = cifs_xquota_set, - .get_xquota = cifs_xquota_set, - .set_xstate = cifs_xstate_set, - .get_xstate = cifs_xstate_get, -}; -#endif + return 0; +} -static void cifs_umount_begin(struct super_block * sblock) +static void cifs_umount_begin(struct super_block *sb) { - struct cifs_sb_info *cifs_sb; - struct cifsTconInfo * tcon; + struct cifs_sb_info *cifs_sb = CIFS_SB(sb); + struct cifs_tcon *tcon; - cifs_sb = CIFS_SB(sblock); - if(cifs_sb == NULL) + if (cifs_sb == NULL) return; - tcon = cifs_sb->tcon; - if(tcon == NULL) + tcon = cifs_sb_master_tcon(cifs_sb); + + spin_lock(&cifs_tcp_ses_lock); + if ((tcon->tc_count > 1) || (tcon->tidStatus == CifsExiting)) { + /* we have other mounts to same share or we have + already tried to force umount this and woken up + all waiting network requests, nothing to do */ + spin_unlock(&cifs_tcp_ses_lock); return; - down(&tcon->tconSem); - if (atomic_read(&tcon->useCount) == 1) + } else if (tcon->tc_count == 1) tcon->tidStatus = CifsExiting; - up(&tcon->tconSem); + spin_unlock(&cifs_tcp_ses_lock); - if(tcon->ses && tcon->ses->server) - { - cERROR(1,("wake up tasks now - umount begin not complete")); + /* cancel_brl_requests(tcon); */ /* BB mark all brl mids as exiting */ + /* cancel_notify_requests(tcon); */ + if (tcon->ses && tcon->ses->server) { + cifs_dbg(FYI, "wake up tasks now - umount begin not complete\n"); wake_up_all(&tcon->ses->server->request_q); + wake_up_all(&tcon->ses->server->response_q); + msleep(1); /* yield */ + /* we have to kick the requests once more */ + wake_up_all(&tcon->ses->server->response_q); + msleep(1); } -/* BB FIXME - finish add checks for tidStatus BB */ return; } - + +#ifdef CONFIG_CIFS_STATS2 +static int cifs_show_stats(struct seq_file *s, struct dentry *root) +{ + /* BB FIXME */ + return 0; +} +#endif static int cifs_remount(struct super_block *sb, int *flags, char *data) { + sync_filesystem(sb); *flags |= MS_NODIRATIME; return 0; } -struct super_operations cifs_super_ops = { - .read_inode = cifs_read_inode, - .put_super = cifs_put_super, +static int cifs_drop_inode(struct inode *inode) +{ + struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb); + + /* no serverino => unconditional eviction */ + return !(cifs_sb->mnt_cifs_flags & CIFS_MOUNT_SERVER_INUM) || + generic_drop_inode(inode); +} + +static const struct super_operations cifs_super_ops = { .statfs = cifs_statfs, .alloc_inode = cifs_alloc_inode, .destroy_inode = cifs_destroy_inode, -/* .drop_inode = generic_delete_inode, - .delete_inode = cifs_delete_inode, *//* Do not need the above two functions - unless later we add lazy close of inodes or unless the kernel forgets to call - us with the same number of releases (closes) as opens */ + .drop_inode = cifs_drop_inode, + .evict_inode = cifs_evict_inode, +/* .delete_inode = cifs_delete_inode, */ /* Do not need above + function unless later we add lazy close of inodes or unless the + kernel forgets to call us with the same number of releases (closes) + as opens */ .show_options = cifs_show_options, -/* .umount_begin = cifs_umount_begin, */ /* BB finish in the future */ + .umount_begin = cifs_umount_begin, .remount_fs = cifs_remount, +#ifdef CONFIG_CIFS_STATS2 + .show_stats = cifs_show_stats, +#endif }; -static struct super_block * -cifs_get_sb(struct file_system_type *fs_type, - int flags, const char *dev_name, void *data) +/* + * Get root dentry from superblock according to prefix path mount option. + * Return dentry with refcount + 1 on success and NULL otherwise. + */ +static struct dentry * +cifs_get_root(struct smb_vol *vol, struct super_block *sb) +{ + struct dentry *dentry; + struct cifs_sb_info *cifs_sb = CIFS_SB(sb); + char *full_path = NULL; + char *s, *p; + char sep; + + full_path = cifs_build_path_to_root(vol, cifs_sb, + cifs_sb_master_tcon(cifs_sb)); + if (full_path == NULL) + return ERR_PTR(-ENOMEM); + + cifs_dbg(FYI, "Get root dentry for %s\n", full_path); + + sep = CIFS_DIR_SEP(cifs_sb); + dentry = dget(sb->s_root); + p = s = full_path; + + do { + struct inode *dir = dentry->d_inode; + struct dentry *child; + + if (!dir) { + dput(dentry); + dentry = ERR_PTR(-ENOENT); + break; + } + if (!S_ISDIR(dir->i_mode)) { + dput(dentry); + dentry = ERR_PTR(-ENOTDIR); + break; + } + + /* skip separators */ + while (*s == sep) + s++; + if (!*s) + break; + p = s++; + /* next separator */ + while (*s && *s != sep) + s++; + + mutex_lock(&dir->i_mutex); + child = lookup_one_len(p, dentry, s - p); + mutex_unlock(&dir->i_mutex); + dput(dentry); + dentry = child; + } while (!IS_ERR(dentry)); + kfree(full_path); + return dentry; +} + +static int cifs_set_super(struct super_block *sb, void *data) +{ + struct cifs_mnt_data *mnt_data = data; + sb->s_fs_info = mnt_data->cifs_sb; + return set_anon_super(sb, NULL); +} + +static struct dentry * +cifs_do_mount(struct file_system_type *fs_type, + int flags, const char *dev_name, void *data) { int rc; - struct super_block *sb = sget(fs_type, NULL, set_anon_super, NULL); + struct super_block *sb; + struct cifs_sb_info *cifs_sb; + struct smb_vol *volume_info; + struct cifs_mnt_data mnt_data; + struct dentry *root; - cFYI(1, ("Devname: %s flags: %d ", dev_name, flags)); + cifs_dbg(FYI, "Devname: %s flags: %d\n", dev_name, flags); - if (IS_ERR(sb)) - return sb; + volume_info = cifs_get_volume_info((char *)data, dev_name); + if (IS_ERR(volume_info)) + return ERR_CAST(volume_info); - sb->s_flags = flags; + cifs_sb = kzalloc(sizeof(struct cifs_sb_info), GFP_KERNEL); + if (cifs_sb == NULL) { + root = ERR_PTR(-ENOMEM); + goto out_nls; + } + + cifs_sb->mountdata = kstrndup(data, PAGE_SIZE, GFP_KERNEL); + if (cifs_sb->mountdata == NULL) { + root = ERR_PTR(-ENOMEM); + goto out_cifs_sb; + } + + cifs_setup_cifs_sb(volume_info, cifs_sb); - rc = cifs_read_super(sb, data, dev_name, flags & MS_VERBOSE ? 1 : 0); + rc = cifs_mount(cifs_sb, volume_info); if (rc) { - up_write(&sb->s_umount); - deactivate_super(sb); - return ERR_PTR(rc); + if (!(flags & MS_SILENT)) + cifs_dbg(VFS, "cifs_mount failed w/return code = %d\n", + rc); + root = ERR_PTR(rc); + goto out_mountdata; } - sb->s_flags |= MS_ACTIVE; - return sb; -} -static ssize_t -cifs_read_wrapper(struct file * file, char __user *read_data, size_t read_size, - loff_t * poffset) -{ - if(file->f_dentry == NULL) - return -EIO; - else if(file->f_dentry->d_inode == NULL) - return -EIO; + mnt_data.vol = volume_info; + mnt_data.cifs_sb = cifs_sb; + mnt_data.flags = flags; - cFYI(1,("In read_wrapper size %zd at %lld",read_size,*poffset)); + /* BB should we make this contingent on mount parm? */ + flags |= MS_NODIRATIME | MS_NOATIME; - if(CIFS_I(file->f_dentry->d_inode)->clientCanCacheRead) { - return generic_file_read(file,read_data,read_size,poffset); + sb = sget(fs_type, cifs_match_super, cifs_set_super, flags, &mnt_data); + if (IS_ERR(sb)) { + root = ERR_CAST(sb); + cifs_umount(cifs_sb); + goto out; + } + + if (sb->s_root) { + cifs_dbg(FYI, "Use existing superblock\n"); + cifs_umount(cifs_sb); } else { - /* BB do we need to lock inode from here until after invalidate? */ -/* if(file->f_dentry->d_inode->i_mapping) { - filemap_fdatawrite(file->f_dentry->d_inode->i_mapping); - filemap_fdatawait(file->f_dentry->d_inode->i_mapping); - }*/ -/* cifs_revalidate(file->f_dentry);*/ /* BB fixme */ - - /* BB we should make timer configurable - perhaps - by simply calling cifs_revalidate here */ - /* invalidate_remote_inode(file->f_dentry->d_inode);*/ - return generic_file_read(file,read_data,read_size,poffset); + rc = cifs_read_super(sb); + if (rc) { + root = ERR_PTR(rc); + goto out_super; + } + + sb->s_flags |= MS_ACTIVE; } + + root = cifs_get_root(volume_info, sb); + if (IS_ERR(root)) + goto out_super; + + cifs_dbg(FYI, "dentry root is: %p\n", root); + goto out; + +out_super: + deactivate_locked_super(sb); +out: + cifs_cleanup_volume_info(volume_info); + return root; + +out_mountdata: + kfree(cifs_sb->mountdata); +out_cifs_sb: + kfree(cifs_sb); +out_nls: + unload_nls(volume_info->local_nls); + goto out; } static ssize_t -cifs_write_wrapper(struct file * file, const char __user *write_data, - size_t write_size, loff_t * poffset) +cifs_loose_read_iter(struct kiocb *iocb, struct iov_iter *iter) { + ssize_t rc; + struct inode *inode = file_inode(iocb->ki_filp); + + rc = cifs_revalidate_mapping(inode); + if (rc) + return rc; + + return generic_file_read_iter(iocb, iter); +} + +static ssize_t cifs_file_write_iter(struct kiocb *iocb, struct iov_iter *from) +{ + struct inode *inode = file_inode(iocb->ki_filp); + struct cifsInodeInfo *cinode = CIFS_I(inode); ssize_t written; + int rc; + + written = cifs_get_writer(cinode); + if (written) + return written; + + written = generic_file_write_iter(iocb, from); - if(file->f_dentry == NULL) - return -EIO; - else if(file->f_dentry->d_inode == NULL) - return -EIO; + if (CIFS_CACHE_WRITE(CIFS_I(inode))) + goto out; - cFYI(1,("In write_wrapper size %zd at %lld",write_size,*poffset)); + rc = filemap_fdatawrite(inode->i_mapping); + if (rc) + cifs_dbg(FYI, "cifs_file_write_iter: %d rc on %p inode\n", + rc, inode); - written = generic_file_write(file,write_data,write_size,poffset); - if(!CIFS_I(file->f_dentry->d_inode)->clientCanCacheAll) { - if(file->f_dentry->d_inode->i_mapping) { - filemap_fdatawrite(file->f_dentry->d_inode->i_mapping); +out: + cifs_put_writer(cinode); + return written; +} + +static loff_t cifs_llseek(struct file *file, loff_t offset, int whence) +{ + /* + * whence == SEEK_END || SEEK_DATA || SEEK_HOLE => we must revalidate + * the cached file length + */ + if (whence != SEEK_SET && whence != SEEK_CUR) { + int rc; + struct inode *inode = file_inode(file); + + /* + * We need to be sure that all dirty pages are written and the + * server has the newest file length. + */ + if (!CIFS_CACHE_READ(CIFS_I(inode)) && inode->i_mapping && + inode->i_mapping->nrpages != 0) { + rc = filemap_fdatawait(inode->i_mapping); + if (rc) { + mapping_set_error(inode->i_mapping, rc); + return rc; + } } + /* + * Some applications poll for the file length in this strange + * way so we must seek to end on non-oplocked files by + * setting the revalidate time to zero. + */ + CIFS_I(inode)->time = 0; + + rc = cifs_revalidate_file_attr(file); + if (rc < 0) + return (loff_t)rc; } - return written; + return generic_file_llseek(file, offset, whence); } +static int cifs_setlease(struct file *file, long arg, struct file_lock **lease) +{ + /* + * Note that this is called by vfs setlease with i_lock held to + * protect *lease from going away. + */ + struct inode *inode = file_inode(file); + struct cifsFileInfo *cfile = file->private_data; + + if (!(S_ISREG(inode->i_mode))) + return -EINVAL; + + /* check if file is oplocked */ + if (((arg == F_RDLCK) && CIFS_CACHE_READ(CIFS_I(inode))) || + ((arg == F_WRLCK) && CIFS_CACHE_WRITE(CIFS_I(inode)))) + return generic_setlease(file, arg, lease); + else if (tlink_tcon(cfile->tlink)->local_lease && + !CIFS_CACHE_READ(CIFS_I(inode))) + /* + * If the server claims to support oplock on this file, then we + * still need to check oplock even if the local_lease mount + * option is set, but there are servers which do not support + * oplock for which this mount option may be useful if the user + * knows that the file won't be changed on the server by anyone + * else. + */ + return generic_setlease(file, arg, lease); + else + return -EAGAIN; +} -static struct file_system_type cifs_fs_type = { +struct file_system_type cifs_fs_type = { .owner = THIS_MODULE, .name = "cifs", - .get_sb = cifs_get_sb, - .kill_sb = kill_anon_super, + .mount = cifs_do_mount, + .kill_sb = cifs_kill_sb, /* .fs_flags */ }; -struct inode_operations cifs_dir_inode_ops = { +MODULE_ALIAS_FS("cifs"); +const struct inode_operations cifs_dir_inode_ops = { .create = cifs_create, + .atomic_open = cifs_atomic_open, .lookup = cifs_lookup, .getattr = cifs_getattr, .unlink = cifs_unlink, @@ -558,24 +862,23 @@ struct inode_operations cifs_dir_inode_ops = { #endif }; -struct inode_operations cifs_file_inode_ops = { +const struct inode_operations cifs_file_inode_ops = { /* revalidate:cifs_revalidate, */ .setattr = cifs_setattr, .getattr = cifs_getattr, /* do we need this anymore? */ - .rename = cifs_rename, .permission = cifs_permission, #ifdef CONFIG_CIFS_XATTR .setxattr = cifs_setxattr, .getxattr = cifs_getxattr, .listxattr = cifs_listxattr, .removexattr = cifs_removexattr, -#endif +#endif }; -struct inode_operations cifs_symlink_inode_ops = { - .readlink = generic_readlink, +const struct inode_operations cifs_symlink_inode_ops = { + .readlink = generic_readlink, .follow_link = cifs_follow_link, - .put_link = cifs_put_link, + .put_link = kfree_put_link, .permission = cifs_permission, /* BB add the following two eventually */ /* revalidate: cifs_revalidate, @@ -585,81 +888,147 @@ struct inode_operations cifs_symlink_inode_ops = { .getxattr = cifs_getxattr, .listxattr = cifs_listxattr, .removexattr = cifs_removexattr, -#endif +#endif }; -struct file_operations cifs_file_ops = { - .read = cifs_read_wrapper, - .write = cifs_write_wrapper, +const struct file_operations cifs_file_ops = { + .read = new_sync_read, + .write = new_sync_write, + .read_iter = cifs_loose_read_iter, + .write_iter = cifs_file_write_iter, .open = cifs_open, .release = cifs_close, .lock = cifs_lock, .fsync = cifs_fsync, .flush = cifs_flush, .mmap = cifs_file_mmap, - .sendfile = generic_file_sendfile, + .splice_read = generic_file_splice_read, + .llseek = cifs_llseek, #ifdef CONFIG_CIFS_POSIX - .ioctl = cifs_ioctl, + .unlocked_ioctl = cifs_ioctl, #endif /* CONFIG_CIFS_POSIX */ + .setlease = cifs_setlease, +}; -#ifdef CONFIG_CIFS_EXPERIMENTAL - .readv = generic_file_readv, - .writev = generic_file_writev, - .aio_read = generic_file_aio_read, - .aio_write = generic_file_aio_write, - .dir_notify = cifs_dir_notify, -#endif /* CONFIG_CIFS_EXPERIMENTAL */ +const struct file_operations cifs_file_strict_ops = { + .read = new_sync_read, + .write = new_sync_write, + .read_iter = cifs_strict_readv, + .write_iter = cifs_strict_writev, + .open = cifs_open, + .release = cifs_close, + .lock = cifs_lock, + .fsync = cifs_strict_fsync, + .flush = cifs_flush, + .mmap = cifs_file_strict_mmap, + .splice_read = generic_file_splice_read, + .llseek = cifs_llseek, +#ifdef CONFIG_CIFS_POSIX + .unlocked_ioctl = cifs_ioctl, +#endif /* CONFIG_CIFS_POSIX */ + .setlease = cifs_setlease, }; -struct file_operations cifs_file_direct_ops = { - /* no mmap, no aio, no readv - - BB reevaluate whether they can be done with directio, no cache */ - .read = cifs_user_read, - .write = cifs_user_write, +const struct file_operations cifs_file_direct_ops = { + /* BB reevaluate whether they can be done with directio, no cache */ + .read = new_sync_read, + .write = new_sync_write, + .read_iter = cifs_user_readv, + .write_iter = cifs_user_writev, .open = cifs_open, .release = cifs_close, .lock = cifs_lock, .fsync = cifs_fsync, .flush = cifs_flush, - .sendfile = generic_file_sendfile, /* BB removeme BB */ + .mmap = cifs_file_mmap, + .splice_read = generic_file_splice_read, +#ifdef CONFIG_CIFS_POSIX + .unlocked_ioctl = cifs_ioctl, +#endif /* CONFIG_CIFS_POSIX */ + .llseek = cifs_llseek, + .setlease = cifs_setlease, +}; + +const struct file_operations cifs_file_nobrl_ops = { + .read = new_sync_read, + .write = new_sync_write, + .read_iter = cifs_loose_read_iter, + .write_iter = cifs_file_write_iter, + .open = cifs_open, + .release = cifs_close, + .fsync = cifs_fsync, + .flush = cifs_flush, + .mmap = cifs_file_mmap, + .splice_read = generic_file_splice_read, + .llseek = cifs_llseek, +#ifdef CONFIG_CIFS_POSIX + .unlocked_ioctl = cifs_ioctl, +#endif /* CONFIG_CIFS_POSIX */ + .setlease = cifs_setlease, +}; + +const struct file_operations cifs_file_strict_nobrl_ops = { + .read = new_sync_read, + .write = new_sync_write, + .read_iter = cifs_strict_readv, + .write_iter = cifs_strict_writev, + .open = cifs_open, + .release = cifs_close, + .fsync = cifs_strict_fsync, + .flush = cifs_flush, + .mmap = cifs_file_strict_mmap, + .splice_read = generic_file_splice_read, + .llseek = cifs_llseek, #ifdef CONFIG_CIFS_POSIX - .ioctl = cifs_ioctl, + .unlocked_ioctl = cifs_ioctl, #endif /* CONFIG_CIFS_POSIX */ + .setlease = cifs_setlease, +}; -#ifdef CONFIG_CIFS_EXPERIMENTAL - .dir_notify = cifs_dir_notify, -#endif /* CONFIG_CIFS_EXPERIMENTAL */ +const struct file_operations cifs_file_direct_nobrl_ops = { + /* BB reevaluate whether they can be done with directio, no cache */ + .read = new_sync_read, + .write = new_sync_write, + .read_iter = cifs_user_readv, + .write_iter = cifs_user_writev, + .open = cifs_open, + .release = cifs_close, + .fsync = cifs_fsync, + .flush = cifs_flush, + .mmap = cifs_file_mmap, + .splice_read = generic_file_splice_read, +#ifdef CONFIG_CIFS_POSIX + .unlocked_ioctl = cifs_ioctl, +#endif /* CONFIG_CIFS_POSIX */ + .llseek = cifs_llseek, + .setlease = cifs_setlease, }; -struct file_operations cifs_dir_ops = { - .readdir = cifs_readdir, +const struct file_operations cifs_dir_ops = { + .iterate = cifs_readdir, .release = cifs_closedir, .read = generic_read_dir, -#ifdef CONFIG_CIFS_EXPERIMENTAL - .dir_notify = cifs_dir_notify, -#endif /* CONFIG_CIFS_EXPERIMENTAL */ - .ioctl = cifs_ioctl, + .unlocked_ioctl = cifs_ioctl, + .llseek = generic_file_llseek, }; static void -cifs_init_once(void *inode, kmem_cache_t * cachep, unsigned long flags) +cifs_init_once(void *inode) { struct cifsInodeInfo *cifsi = inode; - if ((flags & (SLAB_CTOR_VERIFY | SLAB_CTOR_CONSTRUCTOR)) == - SLAB_CTOR_CONSTRUCTOR) { - inode_init_once(&cifsi->vfs_inode); - INIT_LIST_HEAD(&cifsi->lockList); - } + inode_init_once(&cifsi->vfs_inode); + init_rwsem(&cifsi->lock_sem); } -static int +static int __init cifs_init_inodecache(void) { cifs_inode_cachep = kmem_cache_create("cifs_inode_cache", - sizeof (struct cifsInodeInfo), - 0, SLAB_RECLAIM_ACCOUNT, - cifs_init_once, NULL); + sizeof(struct cifsInodeInfo), + 0, (SLAB_RECLAIM_ACCOUNT| + SLAB_MEM_SPREAD), + cifs_init_once); if (cifs_inode_cachep == NULL) return -ENOMEM; @@ -669,14 +1038,26 @@ cifs_init_inodecache(void) static void cifs_destroy_inodecache(void) { - if (kmem_cache_destroy(cifs_inode_cachep)) - printk(KERN_WARNING "cifs_inode_cache: error freeing\n"); + /* + * Make sure all delayed rcu free inodes are flushed before we + * destroy cache. + */ + rcu_barrier(); + kmem_cache_destroy(cifs_inode_cachep); } static int cifs_init_request_bufs(void) { - if(CIFSMaxBufSize < 8192) { + size_t max_hdr_size = MAX_CIFS_HDR_SIZE; +#ifdef CONFIG_CIFS_SMB2 + /* + * SMB2 maximum header size is bigger than CIFS one - no problems to + * allocate some more bytes for CIFS. + */ + max_hdr_size = MAX_SMB2_HDR_SIZE; +#endif + if (CIFSMaxBufSize < 8192) { /* Buffer size can not be smaller than 2 * PATH_MAX since maximum Unicode path name has to fit in any SMB/CIFS path based frames */ CIFSMaxBufSize = 8192; @@ -685,59 +1066,58 @@ cifs_init_request_bufs(void) } else { CIFSMaxBufSize &= 0x1FE00; /* Round size to even 512 byte mult*/ } -/* cERROR(1,("CIFSMaxBufSize %d 0x%x",CIFSMaxBufSize,CIFSMaxBufSize)); */ +/* + cifs_dbg(VFS, "CIFSMaxBufSize %d 0x%x\n", + CIFSMaxBufSize, CIFSMaxBufSize); +*/ cifs_req_cachep = kmem_cache_create("cifs_request", - CIFSMaxBufSize + - MAX_CIFS_HDR_SIZE, 0, - SLAB_HWCACHE_ALIGN, NULL, NULL); + CIFSMaxBufSize + max_hdr_size, 0, + SLAB_HWCACHE_ALIGN, NULL); if (cifs_req_cachep == NULL) return -ENOMEM; - if(cifs_min_rcv < 1) + if (cifs_min_rcv < 1) cifs_min_rcv = 1; else if (cifs_min_rcv > 64) { cifs_min_rcv = 64; - cERROR(1,("cifs_min_rcv set to maximum (64)")); + cifs_dbg(VFS, "cifs_min_rcv set to maximum (64)\n"); } - cifs_req_poolp = mempool_create(cifs_min_rcv, - mempool_alloc_slab, - mempool_free_slab, - cifs_req_cachep); + cifs_req_poolp = mempool_create_slab_pool(cifs_min_rcv, + cifs_req_cachep); - if(cifs_req_poolp == NULL) { + if (cifs_req_poolp == NULL) { kmem_cache_destroy(cifs_req_cachep); return -ENOMEM; } - /* 256 (MAX_CIFS_HDR_SIZE bytes is enough for most SMB responses and + /* MAX_CIFS_SMALL_BUFFER_SIZE bytes is enough for most SMB responses and almost all handle based requests (but not write response, nor is it sufficient for path based requests). A smaller size would have - been more efficient (compacting multiple slab items on one 4k page) + been more efficient (compacting multiple slab items on one 4k page) for the case in which debug was on, but this larger size allows more SMBs to use small buffer alloc and is still much more - efficient to alloc 1 per page off the slab compared to 17K (5page) + efficient to alloc 1 per page off the slab compared to 17K (5page) alloc of large cifs buffers even when page debugging is on */ cifs_sm_req_cachep = kmem_cache_create("cifs_small_rq", - MAX_CIFS_HDR_SIZE, 0, SLAB_HWCACHE_ALIGN, NULL, NULL); + MAX_CIFS_SMALL_BUFFER_SIZE, 0, SLAB_HWCACHE_ALIGN, + NULL); if (cifs_sm_req_cachep == NULL) { mempool_destroy(cifs_req_poolp); kmem_cache_destroy(cifs_req_cachep); - return -ENOMEM; + return -ENOMEM; } - if(cifs_min_small < 2) + if (cifs_min_small < 2) cifs_min_small = 2; else if (cifs_min_small > 256) { cifs_min_small = 256; - cFYI(1,("cifs_min_small set to maximum (256)")); + cifs_dbg(FYI, "cifs_min_small set to maximum (256)\n"); } - cifs_sm_req_poolp = mempool_create(cifs_min_small, - mempool_alloc_slab, - mempool_free_slab, - cifs_sm_req_cachep); + cifs_sm_req_poolp = mempool_create_slab_pool(cifs_min_small, + cifs_sm_req_cachep); - if(cifs_sm_req_poolp == NULL) { + if (cifs_sm_req_poolp == NULL) { mempool_destroy(cifs_req_poolp); kmem_cache_destroy(cifs_req_cachep); kmem_cache_destroy(cifs_sm_req_cachep); @@ -751,39 +1131,24 @@ static void cifs_destroy_request_bufs(void) { mempool_destroy(cifs_req_poolp); - if (kmem_cache_destroy(cifs_req_cachep)) - printk(KERN_WARNING - "cifs_destroy_request_cache: error not all structures were freed\n"); + kmem_cache_destroy(cifs_req_cachep); mempool_destroy(cifs_sm_req_poolp); - if (kmem_cache_destroy(cifs_sm_req_cachep)) - printk(KERN_WARNING - "cifs_destroy_request_cache: cifs_small_rq free error\n"); + kmem_cache_destroy(cifs_sm_req_cachep); } static int cifs_init_mids(void) { cifs_mid_cachep = kmem_cache_create("cifs_mpx_ids", - sizeof (struct mid_q_entry), 0, - SLAB_HWCACHE_ALIGN, NULL, NULL); + sizeof(struct mid_q_entry), 0, + SLAB_HWCACHE_ALIGN, NULL); if (cifs_mid_cachep == NULL) return -ENOMEM; - cifs_mid_poolp = mempool_create(3 /* a reasonable min simultan opers */, - mempool_alloc_slab, - mempool_free_slab, - cifs_mid_cachep); - if(cifs_mid_poolp == NULL) { - kmem_cache_destroy(cifs_mid_cachep); - return -ENOMEM; - } - - cifs_oplock_cachep = kmem_cache_create("cifs_oplock_structs", - sizeof (struct oplock_q_entry), 0, - SLAB_HWCACHE_ALIGN, NULL, NULL); - if (cifs_oplock_cachep == NULL) { + /* 3 is a reasonable minimum number of simultaneous operations */ + cifs_mid_poolp = mempool_create_slab_pool(3, cifs_mid_cachep); + if (cifs_mid_poolp == NULL) { kmem_cache_destroy(cifs_mid_cachep); - mempool_destroy(cifs_mid_poolp); return -ENOMEM; } @@ -794,200 +1159,140 @@ static void cifs_destroy_mids(void) { mempool_destroy(cifs_mid_poolp); - if (kmem_cache_destroy(cifs_mid_cachep)) - printk(KERN_WARNING - "cifs_destroy_mids: error not all structures were freed\n"); - - if (kmem_cache_destroy(cifs_oplock_cachep)) - printk(KERN_WARNING - "error not all oplock structures were freed\n"); -} - -static int cifs_oplock_thread(void * dummyarg) -{ - struct oplock_q_entry * oplock_item; - struct cifsTconInfo *pTcon; - struct inode * inode; - __u16 netfid; - int rc; - - daemonize("cifsoplockd"); - allow_signal(SIGTERM); - - oplockThread = current; - do { - if (try_to_freeze()) - continue; - - spin_lock(&GlobalMid_Lock); - if(list_empty(&GlobalOplock_Q)) { - spin_unlock(&GlobalMid_Lock); - set_current_state(TASK_INTERRUPTIBLE); - schedule_timeout(39*HZ); - } else { - oplock_item = list_entry(GlobalOplock_Q.next, - struct oplock_q_entry, qhead); - if(oplock_item) { - cFYI(1,("found oplock item to write out")); - pTcon = oplock_item->tcon; - inode = oplock_item->pinode; - netfid = oplock_item->netfid; - spin_unlock(&GlobalMid_Lock); - DeleteOplockQEntry(oplock_item); - /* can not grab inode sem here since it would - deadlock when oplock received on delete - since vfs_unlink holds the i_sem across - the call */ - /* down(&inode->i_sem);*/ - if (S_ISREG(inode->i_mode)) { - rc = filemap_fdatawrite(inode->i_mapping); - if(CIFS_I(inode)->clientCanCacheRead == 0) { - filemap_fdatawait(inode->i_mapping); - invalidate_remote_inode(inode); - } - } else - rc = 0; - /* up(&inode->i_sem);*/ - if (rc) - CIFS_I(inode)->write_behind_rc = rc; - cFYI(1,("Oplock flush inode %p rc %d",inode,rc)); - - /* releasing a stale oplock after recent reconnection - of smb session using a now incorrect file - handle is not a data integrity issue but do - not bother sending an oplock release if session - to server still is disconnected since oplock - already released by the server in that case */ - if(pTcon->tidStatus != CifsNeedReconnect) { - rc = CIFSSMBLock(0, pTcon, netfid, - 0 /* len */ , 0 /* offset */, 0, - 0, LOCKING_ANDX_OPLOCK_RELEASE, - 0 /* wait flag */); - cFYI(1,("Oplock release rc = %d ",rc)); - } - } else - spin_unlock(&GlobalMid_Lock); - set_current_state(TASK_INTERRUPTIBLE); - schedule_timeout(1); /* yield in case q were corrupt */ - } - } while(!signal_pending(current)); - oplockThread = NULL; - complete_and_exit (&cifs_oplock_exited, 0); -} - -static int cifs_dnotify_thread(void * dummyarg) -{ - daemonize("cifsdnotifyd"); - allow_signal(SIGTERM); - - dnotifyThread = current; - do { - if(try_to_freeze()) - continue; - set_current_state(TASK_INTERRUPTIBLE); - schedule_timeout(39*HZ); - } while(!signal_pending(current)); - complete_and_exit (&cifs_dnotify_exited, 0); + kmem_cache_destroy(cifs_mid_cachep); } static int __init init_cifs(void) { int rc = 0; -#ifdef CONFIG_PROC_FS cifs_proc_init(); -#endif - INIT_LIST_HEAD(&GlobalServerList); /* BB not implemented yet */ - INIT_LIST_HEAD(&GlobalSMBSessionList); - INIT_LIST_HEAD(&GlobalTreeConnectionList); - INIT_LIST_HEAD(&GlobalOplock_Q); -#ifdef CONFIG_CIFS_EXPERIMENTAL + INIT_LIST_HEAD(&cifs_tcp_ses_list); +#ifdef CONFIG_CIFS_DNOTIFY_EXPERIMENTAL /* unused temporarily */ INIT_LIST_HEAD(&GlobalDnotifyReqList); INIT_LIST_HEAD(&GlobalDnotifyRsp_Q); -#endif +#endif /* was needed for dnotify, and will be needed for inotify when VFS fix */ /* * Initialize Global counters */ atomic_set(&sesInfoAllocCount, 0); atomic_set(&tconInfoAllocCount, 0); - atomic_set(&tcpSesAllocCount,0); + atomic_set(&tcpSesAllocCount, 0); atomic_set(&tcpSesReconnectCount, 0); atomic_set(&tconInfoReconnectCount, 0); atomic_set(&bufAllocCount, 0); + atomic_set(&smBufAllocCount, 0); +#ifdef CONFIG_CIFS_STATS2 + atomic_set(&totBufAllocCount, 0); + atomic_set(&totSmBufAllocCount, 0); +#endif /* CONFIG_CIFS_STATS2 */ + atomic_set(&midCount, 0); GlobalCurrentXid = 0; GlobalTotalActiveXid = 0; GlobalMaxActiveXid = 0; - rwlock_init(&GlobalSMBSeslock); + spin_lock_init(&cifs_tcp_ses_lock); + spin_lock_init(&cifs_file_list_lock); spin_lock_init(&GlobalMid_Lock); - if(cifs_max_pending < 2) { + if (cifs_max_pending < 2) { cifs_max_pending = 2; - cFYI(1,("cifs_max_pending set to min of 2")); - } else if(cifs_max_pending > 256) { - cifs_max_pending = 256; - cFYI(1,("cifs_max_pending set to max of 256")); + cifs_dbg(FYI, "cifs_max_pending set to min of 2\n"); + } else if (cifs_max_pending > CIFS_MAX_REQ) { + cifs_max_pending = CIFS_MAX_REQ; + cifs_dbg(FYI, "cifs_max_pending set to max of %u\n", + CIFS_MAX_REQ); } - rc = cifs_init_inodecache(); - if (!rc) { - rc = cifs_init_mids(); - if (!rc) { - rc = cifs_init_request_bufs(); - if (!rc) { - rc = register_filesystem(&cifs_fs_type); - if (!rc) { - rc = (int)kernel_thread(cifs_oplock_thread, NULL, - CLONE_FS | CLONE_FILES | CLONE_VM); - if(rc > 0) { - rc = (int)kernel_thread(cifs_dnotify_thread, NULL, - CLONE_FS | CLONE_FILES | CLONE_VM); - if(rc > 0) - return 0; - else - cERROR(1,("error %d create dnotify thread", rc)); - } else { - cERROR(1,("error %d create oplock thread",rc)); - } - } - cifs_destroy_request_bufs(); - } - cifs_destroy_mids(); - } - cifs_destroy_inodecache(); + cifsiod_wq = alloc_workqueue("cifsiod", WQ_FREEZABLE|WQ_MEM_RECLAIM, 0); + if (!cifsiod_wq) { + rc = -ENOMEM; + goto out_clean_proc; } -#ifdef CONFIG_PROC_FS - cifs_proc_clean(); + + rc = cifs_fscache_register(); + if (rc) + goto out_destroy_wq; + + rc = cifs_init_inodecache(); + if (rc) + goto out_unreg_fscache; + + rc = cifs_init_mids(); + if (rc) + goto out_destroy_inodecache; + + rc = cifs_init_request_bufs(); + if (rc) + goto out_destroy_mids; + +#ifdef CONFIG_CIFS_UPCALL + rc = register_key_type(&cifs_spnego_key_type); + if (rc) + goto out_destroy_request_bufs; +#endif /* CONFIG_CIFS_UPCALL */ + +#ifdef CONFIG_CIFS_ACL + rc = init_cifs_idmap(); + if (rc) + goto out_register_key_type; +#endif /* CONFIG_CIFS_ACL */ + + rc = register_filesystem(&cifs_fs_type); + if (rc) + goto out_init_cifs_idmap; + + return 0; + +out_init_cifs_idmap: +#ifdef CONFIG_CIFS_ACL + exit_cifs_idmap(); +out_register_key_type: #endif +#ifdef CONFIG_CIFS_UPCALL + unregister_key_type(&cifs_spnego_key_type); +out_destroy_request_bufs: +#endif + cifs_destroy_request_bufs(); +out_destroy_mids: + cifs_destroy_mids(); +out_destroy_inodecache: + cifs_destroy_inodecache(); +out_unreg_fscache: + cifs_fscache_unregister(); +out_destroy_wq: + destroy_workqueue(cifsiod_wq); +out_clean_proc: + cifs_proc_clean(); return rc; } static void __exit exit_cifs(void) { - cFYI(0, ("In unregister ie exit_cifs")); -#ifdef CONFIG_PROC_FS - cifs_proc_clean(); -#endif + cifs_dbg(NOISY, "exit_cifs\n"); unregister_filesystem(&cifs_fs_type); - cifs_destroy_inodecache(); - cifs_destroy_mids(); + cifs_dfs_release_automount_timer(); +#ifdef CONFIG_CIFS_ACL + exit_cifs_idmap(); +#endif +#ifdef CONFIG_CIFS_UPCALL + unregister_key_type(&cifs_spnego_key_type); +#endif cifs_destroy_request_bufs(); - if(oplockThread) { - send_sig(SIGTERM, oplockThread, 1); - wait_for_completion(&cifs_oplock_exited); - } - if(dnotifyThread) { - send_sig(SIGTERM, dnotifyThread, 1); - wait_for_completion(&cifs_dnotify_exited); - } + cifs_destroy_mids(); + cifs_destroy_inodecache(); + cifs_fscache_unregister(); + destroy_workqueue(cifsiod_wq); + cifs_proc_clean(); } MODULE_AUTHOR("Steve French <sfrench@us.ibm.com>"); -MODULE_LICENSE("GPL"); /* combination of LGPL + GPL source behaves as GPL */ +MODULE_LICENSE("GPL"); /* combination of LGPL + GPL source behaves as GPL */ MODULE_DESCRIPTION - ("VFS to access servers complying with the SNIA CIFS Specification e.g. Samba and Windows"); + ("VFS to access servers complying with the SNIA CIFS Specification " + "e.g. Samba and Windows"); MODULE_VERSION(CIFS_VERSION); module_init(init_cifs) module_exit(exit_cifs) |