fedora core 6 1.2949 + vserver 2.2.0
[linux-2.6.git] / mm / shmem.c
index 7107db1..aa4d837 100644 (file)
  * which makes it a completely usable filesystem.
  */
 
-#include <linux/config.h>
 #include <linux/module.h>
 #include <linux/init.h>
-#include <linux/devfs_fs_kernel.h>
 #include <linux/fs.h>
+#include <linux/xattr.h>
+#include <linux/generic_acl.h>
 #include <linux/mm.h>
 #include <linux/mman.h>
 #include <linux/file.h>
 #include <linux/mempolicy.h>
 #include <linux/namei.h>
 #include <linux/ctype.h>
+#include <linux/migrate.h>
+#include <linux/highmem.h>
+#include <linux/backing-dev.h>
+
 #include <asm/uaccess.h>
 #include <asm/div64.h>
 #include <asm/pgtable.h>
@@ -171,10 +175,11 @@ static inline void shmem_unacct_blocks(unsigned long flags, long pages)
 }
 
 static struct super_operations shmem_ops;
-static struct address_space_operations shmem_aops;
-static struct file_operations shmem_file_operations;
+static const struct address_space_operations shmem_aops;
+static const struct file_operations shmem_file_operations;
 static struct inode_operations shmem_inode_operations;
 static struct inode_operations shmem_dir_inode_operations;
+static struct inode_operations shmem_special_inode_operations;
 static struct vm_operations_struct shmem_vm_ops;
 
 static struct backing_dev_info shmem_backing_dev_info  __read_mostly = {
@@ -396,26 +401,38 @@ static swp_entry_t *shmem_swp_alloc(struct shmem_inode_info *info, unsigned long
 /*
  * shmem_free_swp - free some swap entries in a directory
  *
- * @dir:   pointer to the directory
- * @edir:  pointer after last entry of the directory
+ * @dir:        pointer to the directory
+ * @edir:       pointer after last entry of the directory
+ * @punch_lock: pointer to spinlock when needed for the holepunch case
  */
-static int shmem_free_swp(swp_entry_t *dir, swp_entry_t *edir)
+static int shmem_free_swp(swp_entry_t *dir, swp_entry_t *edir,
+                                               spinlock_t *punch_lock)
 {
+       spinlock_t *punch_unlock = NULL;
        swp_entry_t *ptr;
        int freed = 0;
 
        for (ptr = dir; ptr < edir; ptr++) {
                if (ptr->val) {
+                       if (unlikely(punch_lock)) {
+                               punch_unlock = punch_lock;
+                               punch_lock = NULL;
+                               spin_lock(punch_unlock);
+                               if (!ptr->val)
+                                       continue;
+                       }
                        free_swap_and_cache(*ptr);
                        *ptr = (swp_entry_t){0};
                        freed++;
                }
        }
+       if (punch_unlock)
+               spin_unlock(punch_unlock);
        return freed;
 }
 
-static int shmem_map_and_free_swp(struct page *subdir,
-               int offset, int limit, struct page ***dir)
+static int shmem_map_and_free_swp(struct page *subdir, int offset,
+               int limit, struct page ***dir, spinlock_t *punch_lock)
 {
        swp_entry_t *ptr;
        int freed = 0;
@@ -425,7 +442,8 @@ static int shmem_map_and_free_swp(struct page *subdir,
                int size = limit - offset;
                if (size > LATENCY_LIMIT)
                        size = LATENCY_LIMIT;
-               freed += shmem_free_swp(ptr+offset, ptr+offset+size);
+               freed += shmem_free_swp(ptr+offset, ptr+offset+size,
+                                                       punch_lock);
                if (need_resched()) {
                        shmem_swp_unmap(ptr);
                        if (*dir) {
@@ -475,7 +493,10 @@ static void shmem_truncate_range(struct inode *inode, loff_t start, loff_t end)
        long nr_swaps_freed = 0;
        int offset;
        int freed;
-       int punch_hole = 0;
+       int punch_hole;
+       spinlock_t *needs_lock;
+       spinlock_t *punch_lock;
+       unsigned long upper_limit;
 
        inode->i_ctime = inode->i_mtime = CURRENT_TIME;
        idx = (start + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
@@ -486,11 +507,20 @@ static void shmem_truncate_range(struct inode *inode, loff_t start, loff_t end)
        info->flags |= SHMEM_TRUNCATE;
        if (likely(end == (loff_t) -1)) {
                limit = info->next_index;
+               upper_limit = SHMEM_MAX_INDEX;
                info->next_index = idx;
+               needs_lock = NULL;
+               punch_hole = 0;
        } else {
-               limit = (end + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
-               if (limit > info->next_index)
-                       limit = info->next_index;
+               if (end + 1 >= inode->i_size) { /* we may free a little more */
+                       limit = (inode->i_size + PAGE_CACHE_SIZE - 1) >>
+                                                       PAGE_CACHE_SHIFT;
+                       upper_limit = SHMEM_MAX_INDEX;
+               } else {
+                       limit = (end + 1) >> PAGE_CACHE_SHIFT;
+                       upper_limit = limit;
+               }
+               needs_lock = &info->lock;
                punch_hole = 1;
        }
 
@@ -507,17 +537,30 @@ static void shmem_truncate_range(struct inode *inode, loff_t start, loff_t end)
                size = limit;
                if (size > SHMEM_NR_DIRECT)
                        size = SHMEM_NR_DIRECT;
-               nr_swaps_freed = shmem_free_swp(ptr+idx, ptr+size);
+               nr_swaps_freed = shmem_free_swp(ptr+idx, ptr+size, needs_lock);
        }
 
        /*
         * If there are no indirect blocks or we are punching a hole
         * below indirect blocks, nothing to be done.
         */
-       if (!topdir || (punch_hole && (limit <= SHMEM_NR_DIRECT)))
+       if (!topdir || limit <= SHMEM_NR_DIRECT)
                goto done2;
 
-       BUG_ON(limit <= SHMEM_NR_DIRECT);
+       /*
+        * The truncation case has already dropped info->lock, and we're safe
+        * because i_size and next_index have already been lowered, preventing
+        * access beyond.  But in the punch_hole case, we still need to take
+        * the lock when updating the swap directory, because there might be
+        * racing accesses by shmem_getpage(SGP_CACHE), shmem_unuse_inode or
+        * shmem_writepage.  However, whenever we find we can remove a whole
+        * directory page (not at the misaligned start or end of the range),
+        * we first NULLify its pointer in the level above, and then have no
+        * need to take the lock when updating its contents: needs_lock and
+        * punch_lock (either pointing to info->lock or NULL) manage this.
+        */
+
+       upper_limit -= SHMEM_NR_DIRECT;
        limit -= SHMEM_NR_DIRECT;
        idx = (idx > SHMEM_NR_DIRECT)? (idx - SHMEM_NR_DIRECT): 0;
        offset = idx % ENTRIES_PER_PAGE;
@@ -537,8 +580,14 @@ static void shmem_truncate_range(struct inode *inode, loff_t start, loff_t end)
                if (*dir) {
                        diroff = ((idx - ENTRIES_PER_PAGEPAGE/2) %
                                ENTRIES_PER_PAGEPAGE) / ENTRIES_PER_PAGE;
-                       if (!diroff && !offset) {
-                               *dir = NULL;
+                       if (!diroff && !offset && upper_limit >= stage) {
+                               if (needs_lock) {
+                                       spin_lock(needs_lock);
+                                       *dir = NULL;
+                                       spin_unlock(needs_lock);
+                                       needs_lock = NULL;
+                               } else
+                                       *dir = NULL;
                                nr_pages_to_free++;
                                list_add(&middir->lru, &pages_to_free);
                        }
@@ -564,39 +613,55 @@ static void shmem_truncate_range(struct inode *inode, loff_t start, loff_t end)
                        }
                        stage = idx + ENTRIES_PER_PAGEPAGE;
                        middir = *dir;
-                       *dir = NULL;
-                       nr_pages_to_free++;
-                       list_add(&middir->lru, &pages_to_free);
+                       if (punch_hole)
+                               needs_lock = &info->lock;
+                       if (upper_limit >= stage) {
+                               if (needs_lock) {
+                                       spin_lock(needs_lock);
+                                       *dir = NULL;
+                                       spin_unlock(needs_lock);
+                                       needs_lock = NULL;
+                               } else
+                                       *dir = NULL;
+                               nr_pages_to_free++;
+                               list_add(&middir->lru, &pages_to_free);
+                       }
                        shmem_dir_unmap(dir);
                        cond_resched();
                        dir = shmem_dir_map(middir);
                        diroff = 0;
                }
+               punch_lock = needs_lock;
                subdir = dir[diroff];
-               if (subdir && page_private(subdir)) {
+               if (subdir && !offset && upper_limit-idx >= ENTRIES_PER_PAGE) {
+                       if (needs_lock) {
+                               spin_lock(needs_lock);
+                               dir[diroff] = NULL;
+                               spin_unlock(needs_lock);
+                               punch_lock = NULL;
+                       } else
+                               dir[diroff] = NULL;
+                       nr_pages_to_free++;
+                       list_add(&subdir->lru, &pages_to_free);
+               }
+               if (subdir && page_private(subdir) /* has swap entries */) {
                        size = limit - idx;
                        if (size > ENTRIES_PER_PAGE)
                                size = ENTRIES_PER_PAGE;
                        freed = shmem_map_and_free_swp(subdir,
-                                               offset, size, &dir);
+                                       offset, size, &dir, punch_lock);
                        if (!dir)
                                dir = shmem_dir_map(middir);
                        nr_swaps_freed += freed;
-                       if (offset)
+                       if (offset || punch_lock) {
                                spin_lock(&info->lock);
-                       set_page_private(subdir, page_private(subdir) - freed);
-                       if (offset)
+                               set_page_private(subdir,
+                                       page_private(subdir) - freed);
                                spin_unlock(&info->lock);
-                       if (!punch_hole)
-                               BUG_ON(page_private(subdir) > offset);
-               }
-               if (offset)
-                       offset = 0;
-               else if (subdir && !page_private(subdir)) {
-                       dir[diroff] = NULL;
-                       nr_pages_to_free++;
-                       list_add(&subdir->lru, &pages_to_free);
+                       } else
+                               BUG_ON(page_private(subdir) != freed);
                }
+               offset = 0;
        }
 done1:
        shmem_dir_unmap(dir);
@@ -608,8 +673,16 @@ done2:
                 * generic_delete_inode did it, before we lowered next_index.
                 * Also, though shmem_getpage checks i_size before adding to
                 * cache, no recheck after: so fix the narrow window there too.
+                *
+                * Recalling truncate_inode_pages_range and unmap_mapping_range
+                * every time for punch_hole (which never got a chance to clear
+                * SHMEM_PAGEIN at the start of vmtruncate_range) is expensive,
+                * yet hardly ever necessary: try to optimize them out later.
                 */
                truncate_inode_pages_range(inode->i_mapping, start, end);
+               if (punch_hole)
+                       unmap_mapping_range(inode->i_mapping, start,
+                                                       end - start, 1);
        }
 
        spin_lock(&info->lock);
@@ -640,7 +713,7 @@ static int shmem_notify_change(struct dentry *dentry, struct iattr *attr)
        struct page *page = NULL;
        int error;
 
-       if (attr->ia_valid & ATTR_SIZE) {
+       if (S_ISREG(inode->i_mode) && (attr->ia_valid & ATTR_SIZE)) {
                if (attr->ia_size < inode->i_size) {
                        /*
                         * If truncating down to a partial page, then
@@ -673,6 +746,10 @@ static int shmem_notify_change(struct dentry *dentry, struct iattr *attr)
        error = inode_change_ok(inode, attr);
        if (!error)
                error = inode_setattr(inode, attr);
+#ifdef CONFIG_TMPFS_POSIX_ACL
+       if (!error && (attr->ia_valid & ATTR_MODE))
+               error = generic_acl_chmod(inode, &shmem_acl_ops);
+#endif
        if (page)
                page_cache_release(page);
        return error;
@@ -879,7 +956,7 @@ redirty:
 }
 
 #ifdef CONFIG_NUMA
-static int shmem_parse_mpol(char *value, int *policy, nodemask_t *policy_nodes)
+static inline int shmem_parse_mpol(char *value, int *policy, nodemask_t *policy_nodes)
 {
        char *nodelist = strchr(value, ':');
        int err = 1;
@@ -1048,12 +1125,12 @@ repeat:
                swappage = lookup_swap_cache(swap);
                if (!swappage) {
                        shmem_swp_unmap(entry);
-                       spin_unlock(&info->lock);
                        /* here we actually do the io */
                        if (type && *type == VM_FAULT_MINOR) {
-                               inc_page_state(pgmajfault);
+                               __count_vm_event(PGMAJFAULT);
                                *type = VM_FAULT_MAJOR;
                        }
+                       spin_unlock(&info->lock);
                        swappage = shmem_swapin(info, swap, idx);
                        if (!swappage) {
                                spin_lock(&info->lock);
@@ -1083,14 +1160,6 @@ repeat:
                        page_cache_release(swappage);
                        goto repeat;
                }
-               if (!PageSwapCache(swappage)) {
-                       /* Page migration has occured */
-                       shmem_swp_unmap(entry);
-                       spin_unlock(&info->lock);
-                       unlock_page(swappage);
-                       page_cache_release(swappage);
-                       goto repeat;
-               }
                if (PageWriteback(swappage)) {
                        shmem_swp_unmap(entry);
                        spin_unlock(&info->lock);
@@ -1135,7 +1204,7 @@ repeat:
                        page_cache_release(swappage);
                        if (error == -ENOMEM) {
                                /* let kswapd refresh zone for GFP_ATOMICs */
-                               blk_congestion_wait(WRITE, HZ/50);
+                               congestion_wait(WRITE, HZ/50);
                        }
                        goto repeat;
                }
@@ -1228,7 +1297,7 @@ failed:
 
 struct page *shmem_nopage(struct vm_area_struct *vma, unsigned long address, int *type)
 {
-       struct inode *inode = vma->vm_file->f_dentry->d_inode;
+       struct inode *inode = vma->vm_file->f_path.dentry->d_inode;
        struct page *page = NULL;
        unsigned long idx;
        int error;
@@ -1251,7 +1320,7 @@ static int shmem_populate(struct vm_area_struct *vma,
        unsigned long addr, unsigned long len,
        pgprot_t prot, unsigned long pgoff, int nonblock)
 {
-       struct inode *inode = vma->vm_file->f_dentry->d_inode;
+       struct inode *inode = vma->vm_file->f_path.dentry->d_inode;
        struct mm_struct *mm = vma->vm_mm;
        enum sgp_type sgp = nonblock? SGP_QUICK: SGP_CACHE;
        unsigned long size;
@@ -1296,14 +1365,14 @@ static int shmem_populate(struct vm_area_struct *vma,
 #ifdef CONFIG_NUMA
 int shmem_set_policy(struct vm_area_struct *vma, struct mempolicy *new)
 {
-       struct inode *i = vma->vm_file->f_dentry->d_inode;
+       struct inode *i = vma->vm_file->f_path.dentry->d_inode;
        return mpol_set_shared_policy(&SHMEM_I(i)->policy, vma, new);
 }
 
 struct mempolicy *
 shmem_get_policy(struct vm_area_struct *vma, unsigned long addr)
 {
-       struct inode *i = vma->vm_file->f_dentry->d_inode;
+       struct inode *i = vma->vm_file->f_path.dentry->d_inode;
        unsigned long idx;
 
        idx = ((addr - vma->vm_start) >> PAGE_SHIFT) + vma->vm_pgoff;
@@ -1313,7 +1382,7 @@ shmem_get_policy(struct vm_area_struct *vma, unsigned long addr)
 
 int shmem_lock(struct file *file, int lock, struct user_struct *user)
 {
-       struct inode *inode = file->f_dentry->d_inode;
+       struct inode *inode = file->f_path.dentry->d_inode;
        struct shmem_inode_info *info = SHMEM_I(inode);
        int retval = -ENOMEM;
 
@@ -1362,11 +1431,11 @@ shmem_get_inode(struct super_block *sb, int mode, dev_t dev)
                inode->i_mode = mode;
                inode->i_uid = current->fsuid;
                inode->i_gid = current->fsgid;
-               inode->i_blksize = PAGE_CACHE_SIZE;
                inode->i_blocks = 0;
                inode->i_mapping->a_ops = &shmem_aops;
                inode->i_mapping->backing_dev_info = &shmem_backing_dev_info;
                inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
+               inode->i_generation = get_seconds();
                info = SHMEM_I(inode);
                memset(info, 0, (char *)inode - (char *)info);
                spin_lock_init(&info->lock);
@@ -1374,6 +1443,7 @@ shmem_get_inode(struct super_block *sb, int mode, dev_t dev)
 
                switch (mode & S_IFMT) {
                default:
+                       inode->i_op = &shmem_special_inode_operations;
                        init_special_inode(inode, mode, dev);
                        break;
                case S_IFREG:
@@ -1383,7 +1453,7 @@ shmem_get_inode(struct super_block *sb, int mode, dev_t dev)
                                                        &sbinfo->policy_nodes);
                        break;
                case S_IFDIR:
-                       inode->i_nlink++;
+                       inc_nlink(inode);
                        /* Some things misbehave if size == 0 on a directory */
                        inode->i_size = 2 * BOGO_DIRENT_SIZE;
                        inode->i_op = &shmem_dir_inode_operations;
@@ -1424,7 +1494,7 @@ shmem_prepare_write(struct file *file, struct page *page, unsigned offset, unsig
 static ssize_t
 shmem_file_write(struct file *file, const char __user *buf, size_t count, loff_t *ppos)
 {
-       struct inode    *inode = file->f_dentry->d_inode;
+       struct inode    *inode = file->f_path.dentry->d_inode;
        loff_t          pos;
        unsigned long   written;
        ssize_t         err;
@@ -1444,7 +1514,7 @@ shmem_file_write(struct file *file, const char __user *buf, size_t count, loff_t
        if (err || !count)
                goto out;
 
-       err = remove_suid(file->f_dentry);
+       err = remove_suid(file->f_path.dentry);
        if (err)
                goto out;
 
@@ -1526,7 +1596,7 @@ out:
 
 static void do_shmem_file_read(struct file *filp, loff_t *ppos, read_descriptor_t *desc, read_actor_t actor)
 {
-       struct inode *inode = filp->f_dentry->d_inode;
+       struct inode *inode = filp->f_path.dentry->d_inode;
        struct address_space *mapping = inode->i_mapping;
        unsigned long index, offset;
 
@@ -1656,9 +1726,9 @@ static ssize_t shmem_file_sendfile(struct file *in_file, loff_t *ppos,
        return desc.error;
 }
 
-static int shmem_statfs(struct super_block *sb, struct kstatfs *buf)
+static int shmem_statfs(struct dentry *dentry, struct kstatfs *buf)
 {
-       struct shmem_sb_info *sbinfo = SHMEM_SB(sb);
+       struct shmem_sb_info *sbinfo = SHMEM_SB(dentry->d_sb);
 
        buf->f_type = TMPFS_SUPER_MAGIC;
        buf->f_bsize = PAGE_CACHE_SIZE;
@@ -1694,7 +1764,11 @@ shmem_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t dev)
                                iput(inode);
                                return error;
                        }
-                       error = 0;
+               }
+               error = shmem_acl_init(inode, dir);
+               if (error) {
+                       iput(inode);
+                       return error;
                }
                if (dir->i_mode & S_ISGID) {
                        inode->i_gid = dir->i_gid;
@@ -1715,7 +1789,7 @@ static int shmem_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 
        if ((error = shmem_mknod(dir, dentry, mode | S_IFDIR, 0)))
                return error;
-       dir->i_nlink++;
+       inc_nlink(dir);
        return 0;
 }
 
@@ -1750,7 +1824,7 @@ static int shmem_link(struct dentry *old_dentry, struct inode *dir, struct dentr
 
        dir->i_size += BOGO_DIRENT_SIZE;
        inode->i_ctime = dir->i_ctime = dir->i_mtime = CURRENT_TIME;
-       inode->i_nlink++;
+       inc_nlink(inode);
        atomic_inc(&inode->i_count);    /* New dentry reference */
        dget(dentry);           /* Extra pinning count for the created dentry */
        d_instantiate(dentry, inode);
@@ -1772,7 +1846,7 @@ static int shmem_unlink(struct inode *dir, struct dentry *dentry)
 
        dir->i_size -= BOGO_DIRENT_SIZE;
        inode->i_ctime = dir->i_ctime = dir->i_mtime = CURRENT_TIME;
-       inode->i_nlink--;
+       drop_nlink(inode);
        dput(dentry);   /* Undo the count from "create" - this does all the work */
        return 0;
 }
@@ -1782,7 +1856,8 @@ static int shmem_rmdir(struct inode *dir, struct dentry *dentry)
        if (!simple_empty(dentry))
                return -ENOTEMPTY;
 
-       dir->i_nlink--;
+       drop_nlink(dentry->d_inode);
+       drop_nlink(dir);
        return shmem_unlink(dir, dentry);
 }
 
@@ -1803,10 +1878,10 @@ static int shmem_rename(struct inode *old_dir, struct dentry *old_dentry, struct
        if (new_dentry->d_inode) {
                (void) shmem_unlink(new_dir, new_dentry);
                if (they_are_dirs)
-                       old_dir->i_nlink--;
+                       drop_nlink(old_dir);
        } else if (they_are_dirs) {
-               old_dir->i_nlink--;
-               new_dir->i_nlink++;
+               drop_nlink(old_dir);
+               inc_nlink(new_dir);
        }
 
        old_dir->i_size -= BOGO_DIRENT_SIZE;
@@ -1908,6 +1983,132 @@ static struct inode_operations shmem_symlink_inode_operations = {
        .put_link       = shmem_put_link,
 };
 
+#ifdef CONFIG_TMPFS_POSIX_ACL
+/**
+ * Superblocks without xattr inode operations will get security.* xattr
+ * support from the VFS "for free". As soon as we have any other xattrs
+ * like ACLs, we also need to implement the security.* handlers at
+ * filesystem level, though.
+ */
+
+static size_t shmem_xattr_security_list(struct inode *inode, char *list,
+                                       size_t list_len, const char *name,
+                                       size_t name_len)
+{
+       return security_inode_listsecurity(inode, list, list_len);
+}
+
+static int shmem_xattr_security_get(struct inode *inode, const char *name,
+                                   void *buffer, size_t size)
+{
+       if (strcmp(name, "") == 0)
+               return -EINVAL;
+       return security_inode_getsecurity(inode, name, buffer, size,
+                                         -EOPNOTSUPP);
+}
+
+static int shmem_xattr_security_set(struct inode *inode, const char *name,
+                                   const void *value, size_t size, int flags)
+{
+       if (strcmp(name, "") == 0)
+               return -EINVAL;
+       return security_inode_setsecurity(inode, name, value, size, flags);
+}
+
+static struct xattr_handler shmem_xattr_security_handler = {
+       .prefix = XATTR_SECURITY_PREFIX,
+       .list   = shmem_xattr_security_list,
+       .get    = shmem_xattr_security_get,
+       .set    = shmem_xattr_security_set,
+};
+
+static struct xattr_handler *shmem_xattr_handlers[] = {
+       &shmem_xattr_acl_access_handler,
+       &shmem_xattr_acl_default_handler,
+       &shmem_xattr_security_handler,
+       NULL
+};
+#endif
+
+static struct dentry *shmem_get_parent(struct dentry *child)
+{
+       return ERR_PTR(-ESTALE);
+}
+
+static int shmem_match(struct inode *ino, void *vfh)
+{
+       __u32 *fh = vfh;
+       __u64 inum = fh[2];
+       inum = (inum << 32) | fh[1];
+       return ino->i_ino == inum && fh[0] == ino->i_generation;
+}
+
+static struct dentry *shmem_get_dentry(struct super_block *sb, void *vfh)
+{
+       struct dentry *de = NULL;
+       struct inode *inode;
+       __u32 *fh = vfh;
+       __u64 inum = fh[2];
+       inum = (inum << 32) | fh[1];
+
+       inode = ilookup5(sb, (unsigned long)(inum+fh[0]), shmem_match, vfh);
+       if (inode) {
+               de = d_find_alias(inode);
+               iput(inode);
+       }
+
+       return de? de: ERR_PTR(-ESTALE);
+}
+
+static struct dentry *shmem_decode_fh(struct super_block *sb, __u32 *fh,
+               int len, int type,
+               int (*acceptable)(void *context, struct dentry *de),
+               void *context)
+{
+       if (len < 3)
+               return ERR_PTR(-ESTALE);
+
+       return sb->s_export_op->find_exported_dentry(sb, fh, NULL, acceptable,
+                                                       context);
+}
+
+static int shmem_encode_fh(struct dentry *dentry, __u32 *fh, int *len,
+                               int connectable)
+{
+       struct inode *inode = dentry->d_inode;
+
+       if (*len < 3)
+               return 255;
+
+       if (hlist_unhashed(&inode->i_hash)) {
+               /* Unfortunately insert_inode_hash is not idempotent,
+                * so as we hash inodes here rather than at creation
+                * time, we need a lock to ensure we only try
+                * to do it once
+                */
+               static DEFINE_SPINLOCK(lock);
+               spin_lock(&lock);
+               if (hlist_unhashed(&inode->i_hash))
+                       __insert_inode_hash(inode,
+                                           inode->i_ino + inode->i_generation);
+               spin_unlock(&lock);
+       }
+
+       fh[0] = inode->i_generation;
+       fh[1] = inode->i_ino;
+       fh[2] = ((__u64)inode->i_ino) >> 32;
+
+       *len = 3;
+       return 1;
+}
+
+static struct export_operations shmem_export_ops = {
+       .get_parent     = shmem_get_parent,
+       .get_dentry     = shmem_get_dentry,
+       .encode_fh      = shmem_encode_fh,
+       .decode_fh      = shmem_decode_fh,
+};
+
 static int shmem_parse_options(char *options, int *mode, uid_t *uid,
        gid_t *gid, unsigned long *blocks, unsigned long *inodes,
        int *policy, nodemask_t *policy_nodes)
@@ -2080,6 +2281,7 @@ static int shmem_fill_super(struct super_block *sb,
                                        &inodes, &policy, &policy_nodes))
                        return -EINVAL;
        }
+       sb->s_export_op = &shmem_export_ops;
 #else
        sb->s_flags |= MS_NOUSER;
 #endif
@@ -2105,6 +2307,10 @@ static int shmem_fill_super(struct super_block *sb,
        sb->s_magic = TMPFS_SUPER_MAGIC;
        sb->s_op = &shmem_ops;
        sb->s_time_gran = 1;
+#ifdef CONFIG_TMPFS_POSIX_ACL
+       sb->s_xattr = shmem_xattr_handlers;
+       sb->s_flags |= MS_POSIXACL;
+#endif
 
        inode = shmem_get_inode(sb, S_IFDIR | mode, 0);
        if (!inode)
@@ -2124,12 +2330,12 @@ failed:
        return err;
 }
 
-static kmem_cache_t *shmem_inode_cachep;
+static struct kmem_cache *shmem_inode_cachep;
 
 static struct inode *shmem_alloc_inode(struct super_block *sb)
 {
        struct shmem_inode_info *p;
-       p = (struct shmem_inode_info *)kmem_cache_alloc(shmem_inode_cachep, SLAB_KERNEL);
+       p = (struct shmem_inode_info *)kmem_cache_alloc(shmem_inode_cachep, GFP_KERNEL);
        if (!p)
                return NULL;
        return &p->vfs_inode;
@@ -2141,16 +2347,22 @@ static void shmem_destroy_inode(struct inode *inode)
                /* only struct inode is valid if it's an inline symlink */
                mpol_free_shared_policy(&SHMEM_I(inode)->policy);
        }
+       shmem_acl_destroy_inode(inode);
        kmem_cache_free(shmem_inode_cachep, SHMEM_I(inode));
 }
 
-static void init_once(void *foo, kmem_cache_t *cachep, unsigned long flags)
+static void init_once(void *foo, struct kmem_cache *cachep,
+                     unsigned long flags)
 {
        struct shmem_inode_info *p = (struct shmem_inode_info *) foo;
 
        if ((flags & (SLAB_CTOR_VERIFY|SLAB_CTOR_CONSTRUCTOR)) ==
            SLAB_CTOR_CONSTRUCTOR) {
                inode_init_once(&p->vfs_inode);
+#ifdef CONFIG_TMPFS_POSIX_ACL
+               p->i_acl = NULL;
+               p->i_default_acl = NULL;
+#endif
        }
 }
 
@@ -2166,11 +2378,10 @@ static int init_inodecache(void)
 
 static void destroy_inodecache(void)
 {
-       if (kmem_cache_destroy(shmem_inode_cachep))
-               printk(KERN_INFO "shmem_inode_cache: not all structures were freed\n");
+       kmem_cache_destroy(shmem_inode_cachep);
 }
 
-static struct address_space_operations shmem_aops = {
+static const struct address_space_operations shmem_aops = {
        .writepage      = shmem_writepage,
        .set_page_dirty = __set_page_dirty_nobuffers,
 #ifdef CONFIG_TMPFS
@@ -2180,7 +2391,7 @@ static struct address_space_operations shmem_aops = {
        .migratepage    = migrate_page,
 };
 
-static struct file_operations shmem_file_operations = {
+static const struct file_operations shmem_file_operations = {
        .mmap           = shmem_mmap,
 #ifdef CONFIG_TMPFS
        .llseek         = generic_file_llseek,
@@ -2195,6 +2406,14 @@ static struct inode_operations shmem_inode_operations = {
        .truncate       = shmem_truncate,
        .setattr        = shmem_notify_change,
        .truncate_range = shmem_truncate_range,
+#ifdef CONFIG_TMPFS_POSIX_ACL
+       .setxattr       = generic_setxattr,
+       .getxattr       = generic_getxattr,
+       .listxattr      = generic_listxattr,
+       .removexattr    = generic_removexattr,
+       .permission     = shmem_permission,
+#endif
+
 };
 
 static struct inode_operations shmem_dir_inode_operations = {
@@ -2209,6 +2428,25 @@ static struct inode_operations shmem_dir_inode_operations = {
        .mknod          = shmem_mknod,
        .rename         = shmem_rename,
 #endif
+#ifdef CONFIG_TMPFS_POSIX_ACL
+       .setattr        = shmem_notify_change,
+       .setxattr       = generic_setxattr,
+       .getxattr       = generic_getxattr,
+       .listxattr      = generic_listxattr,
+       .removexattr    = generic_removexattr,
+       .permission     = shmem_permission,
+#endif
+};
+
+static struct inode_operations shmem_special_inode_operations = {
+#ifdef CONFIG_TMPFS_POSIX_ACL
+       .setattr        = shmem_notify_change,
+       .setxattr       = generic_setxattr,
+       .getxattr       = generic_getxattr,
+       .listxattr      = generic_listxattr,
+       .removexattr    = generic_removexattr,
+       .permission     = shmem_permission,
+#endif
 };
 
 static struct super_operations shmem_ops = {
@@ -2233,10 +2471,10 @@ static struct vm_operations_struct shmem_vm_ops = {
 };
 
 
-static struct super_block *shmem_get_sb(struct file_system_type *fs_type,
-       int flags, const char *dev_name, void *data)
+static int shmem_get_sb(struct file_system_type *fs_type,
+       int flags, const char *dev_name, void *data, struct vfsmount *mnt)
 {
-       return get_sb_nodev(fs_type, flags, data, shmem_fill_super);
+       return get_sb_nodev(fs_type, flags, data, shmem_fill_super, mnt);
 }
 
 static struct file_system_type tmpfs_fs_type = {
@@ -2260,10 +2498,8 @@ static int __init init_tmpfs(void)
                printk(KERN_ERR "Could not register tmpfs\n");
                goto out2;
        }
-#ifdef CONFIG_TMPFS
-       devfs_mk_dir("shm");
-#endif
-       shm_mnt = do_kern_mount(tmpfs_fs_type.name, MS_NOUSER,
+
+       shm_mnt = vfs_kern_mount(&tmpfs_fs_type, MS_NOUSER,
                                tmpfs_fs_type.name, NULL);
        if (IS_ERR(shm_mnt)) {
                error = PTR_ERR(shm_mnt);
@@ -2329,8 +2565,8 @@ struct file *shmem_file_setup(char *name, loff_t size, unsigned long flags)
        d_instantiate(dentry, inode);
        inode->i_size = size;
        inode->i_nlink = 0;     /* It is unlinked */
-       file->f_vfsmnt = mntget(shm_mnt);
-       file->f_dentry = dentry;
+       file->f_path.mnt = mntget(shm_mnt);
+       file->f_path.dentry = dentry;
        file->f_mapping = inode->i_mapping;
        file->f_op = &shmem_file_operations;
        file->f_mode = FMODE_WRITE | FMODE_READ;