Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6:
Remove two unneeded exports and make two symbols static in fs/mpage.c
Cleanup after commit 585d3bc06f4ca57f975a5a1f698f65a45ea66225
Trim includes of fdtable.h
Don't crap into descriptor table in binfmt_som
Trim includes in binfmt_elf
Don't mess with descriptor table in load_elf_binary()
Get rid of indirect include of fs_struct.h
New helper - current_umask()
check_unsafe_exec() doesn't care about signal handlers sharing
New locking/refcounting for fs_struct
Take fs_struct handling to new file (fs/fs_struct.c)
Get rid of bumping fs_struct refcount in pivot_root(2)
Kill unsharing fs_struct in __set_personality()

arch/cris/kernel/process.c

···
 #include <asm/system.h>
 #include <linux/module.h>
 #include <linux/spinlock.h>
-#include <linux/fs_struct.h>
 #include <linux/init_task.h>
 #include <linux/sched.h>
 #include <linux/fs.h>

arch/powerpc/platforms/cell/spufs/inode.c

···
 	if (dentry->d_inode)
 		goto out_dput;
 
-	mode &= ~current->fs->umask;
+	mode &= ~current_umask();
 
 	if (flags & SPU_CREATE_GANG)
 		ret = spufs_create_gang(nd->path.dentry->d_inode,

drivers/char/tty_audit.c

···
  */
 
 #include <linux/audit.h>
-#include <linux/file.h>
-#include <linux/fdtable.h>
 #include <linux/tty.h>
 
 struct tty_audit_buf {

drivers/char/tty_ldisc.c

···
 #include <linux/tty_flip.h>
 #include <linux/devpts_fs.h>
 #include <linux/file.h>
-#include <linux/fdtable.h>
 #include <linux/console.h>
 #include <linux/timer.h>
 #include <linux/ctype.h>

fs/Makefile

···
 		attr.o bad_inode.o file.o filesystems.o namespace.o \
 		seq_file.o xattr.o libfs.o fs-writeback.o \
 		pnode.o drop_caches.o splice.o sync.o utimes.o \
-		stack.o
+		stack.o fs_struct.o
 
 ifeq ($(CONFIG_BLOCK),y)
 obj-y +=	buffer.o bio.o block_dev.o direct-io.o mpage.o ioprio.o

fs/binfmt_elf.c

···
 #include <linux/module.h>
 #include <linux/kernel.h>
 #include <linux/fs.h>
-#include <linux/stat.h>
-#include <linux/time.h>
 #include <linux/mm.h>
 #include <linux/mman.h>
 #include <linux/errno.h>
···
 #include <linux/binfmts.h>
 #include <linux/string.h>
 #include <linux/file.h>
-#include <linux/fcntl.h>
-#include <linux/ptrace.h>
 #include <linux/slab.h>
-#include <linux/shm.h>
 #include <linux/personality.h>
 #include <linux/elfcore.h>
 #include <linux/init.h>
 #include <linux/highuid.h>
-#include <linux/smp.h>
 #include <linux/compiler.h>
 #include <linux/highmem.h>
 #include <linux/pagemap.h>
 #include <linux/security.h>
-#include <linux/syscalls.h>
 #include <linux/random.h>
 #include <linux/elf.h>
 #include <linux/utsname.h>
···
 	unsigned long error;
 	struct elf_phdr *elf_ppnt, *elf_phdata;
 	unsigned long elf_bss, elf_brk;
-	int elf_exec_fileno;
 	int retval, i;
 	unsigned int size;
 	unsigned long elf_entry;
···
 		goto out_free_ph;
 	}
 
-	retval = get_unused_fd();
-	if (retval < 0)
-		goto out_free_ph;
-	get_file(bprm->file);
-	fd_install(elf_exec_fileno = retval, bprm->file);
-
 	elf_ppnt = elf_phdata;
 	elf_bss = 0;
 	elf_brk = 0;
···
 			retval = -ENOEXEC;
 			if (elf_ppnt->p_filesz > PATH_MAX || 
 			    elf_ppnt->p_filesz < 2)
-				goto out_free_file;
+				goto out_free_ph;
 
 			retval = -ENOMEM;
 			elf_interpreter = kmalloc(elf_ppnt->p_filesz,
 						  GFP_KERNEL);
 			if (!elf_interpreter)
-				goto out_free_file;
+				goto out_free_ph;
 
 			retval = kernel_read(bprm->file, elf_ppnt->p_offset,
 					     elf_interpreter,
···
 
 	kfree(elf_phdata);
 
-	sys_close(elf_exec_fileno);
-
 	set_binfmt(&elf_format);
 
 #ifdef ARCH_HAS_SETUP_ADDITIONAL_PAGES
···
 		fput(interpreter);
 out_free_interp:
 	kfree(elf_interpreter);
-out_free_file:
-	sys_close(elf_exec_fileno);
 out_free_ph:
 	kfree(elf_phdata);
 	goto out;

fs/binfmt_som.c

···
 static int
 load_som_binary(struct linux_binprm * bprm, struct pt_regs * regs)
 {
-	int som_exec_fileno;
 	int retval;
 	unsigned int size;
 	unsigned long som_entry;
···
 			retval = -EIO;
 		goto out_free;
 	}
-
-	retval = get_unused_fd();
-	if (retval < 0)
-		goto out_free;
-	get_file(bprm->file);
-	fd_install(som_exec_fileno = retval, bprm->file);
 
 	/* Flush all traces of the currently running executable */
 	retval = flush_old_exec(bprm);

fs/block_dev.c

···
 	}
 	return sync_blockdev(bdev);
 }
+EXPORT_SYMBOL(fsync_bdev);
 
 /**
  * freeze_bdev  --  lock a filesystem and force it into a consistent state

fs/btrfs/acl.c

···
 		}
 
 		if (!acl)
-			inode->i_mode &= ~current->fs->umask;
+			inode->i_mode &= ~current_umask();
 	}
 
 	if (IS_POSIXACL(dir) && acl) {

fs/btrfs/ioctl.c

···
 		goto out_dput;
 
 	if (!IS_POSIXACL(parent->dentry->d_inode))
-		mode &= ~current->fs->umask;
+		mode &= ~current_umask();
 
 	error = mnt_want_write(parent->mnt);
 	if (error)

fs/buffer.c

···
 EXPORT_SYMBOL(end_buffer_read_sync);
 EXPORT_SYMBOL(end_buffer_write_sync);
 EXPORT_SYMBOL(file_fsync);
-EXPORT_SYMBOL(fsync_bdev);
 EXPORT_SYMBOL(generic_block_bmap);
 EXPORT_SYMBOL(generic_cont_expand_simple);
 EXPORT_SYMBOL(init_buffer);

fs/cifs/dir.c

···
 		return -ENOMEM;
 	}
 
-	mode &= ~current->fs->umask;
+	mode &= ~current_umask();
 	if (oplockEnabled)
 		oplock = REQ_OPLOCK;
 
···
 		rc = -ENOMEM;
 	else if (pTcon->unix_ext) {
 		struct cifs_unix_set_info_args args = {
-			.mode	= mode & ~current->fs->umask,
+			.mode	= mode & ~current_umask(),
 			.ctime	= NO_CHANGE_64,
 			.atime	= NO_CHANGE_64,
 			.mtime	= NO_CHANGE_64,

fs/cifs/inode.c

···
 			goto mkdir_out;
 		}
 
-		mode &= ~current->fs->umask;
+		mode &= ~current_umask();
 		rc = CIFSPOSIXCreate(xid, pTcon, SMB_O_DIRECTORY | SMB_O_CREAT,
 				mode, NULL /* netfid */, pInfo, &oplock,
 				full_path, cifs_sb->local_nls,
···
 		if ((direntry->d_inode) && (direntry->d_inode->i_nlink < 2))
 				direntry->d_inode->i_nlink = 2;
 
-		mode &= ~current->fs->umask;
+		mode &= ~current_umask();
 		/* must turn on setgid bit if parent dir has it */
 		if (inode->i_mode & S_ISGID)
 			mode |= S_ISGID;

fs/compat.c

···
 #include <linux/poll.h>
 #include <linux/mm.h>
 #include <linux/eventpoll.h>
+#include <linux/fs_struct.h>
 
 #include <asm/uaccess.h>
 #include <asm/mmu_context.h>
···
 	bprm->cred = prepare_exec_creds();
 	if (!bprm->cred)
 		goto out_unlock;
-	check_unsafe_exec(bprm);
+
+	retval = check_unsafe_exec(bprm);
+	if (retval)
+		goto out_unlock;
 
 	file = open_exec(filename);
 	retval = PTR_ERR(file);
 	if (IS_ERR(file))
-		goto out_unlock;
+		goto out_unmark;
 
 	sched_exec();
 
···
 		goto out;
 
 	/* execve succeeded */
+	write_lock(&current->fs->lock);
+	current->fs->in_exec = 0;
+	write_unlock(&current->fs->lock);
 	current->in_execve = 0;
 	mutex_unlock(&current->cred_exec_mutex);
 	acct_update_integrals(current);
···
 		allow_write_access(bprm->file);
 		fput(bprm->file);
 	}
+
+out_unmark:
+	write_lock(&current->fs->lock);
+	current->fs->in_exec = 0;
+	write_unlock(&current->fs->lock);
 
 out_unlock:
 	current->in_execve = 0;

fs/dcache.c

···
 #include <linux/syscalls.h>
 #include <linux/string.h>
 #include <linux/mm.h>
-#include <linux/fdtable.h>
 #include <linux/fs.h>
 #include <linux/fsnotify.h>
 #include <linux/slab.h>
···
 #include <linux/seqlock.h>
 #include <linux/swap.h>
 #include <linux/bootmem.h>
+#include <linux/fs_struct.h>
 #include "internal.h"
 
 int sysctl_vfs_cache_pressure __read_mostly = 100;

fs/exec.c

···
 #include <linux/tracehook.h>
 #include <linux/kmod.h>
 #include <linux/fsnotify.h>
+#include <linux/fs_struct.h>
 
 #include <asm/uaccess.h>
 #include <asm/mmu_context.h>
···
  * - the caller must hold current->cred_exec_mutex to protect against
  *   PTRACE_ATTACH
  */
-void check_unsafe_exec(struct linux_binprm *bprm)
+int check_unsafe_exec(struct linux_binprm *bprm)
 {
 	struct task_struct *p = current, *t;
 	unsigned long flags;
-	unsigned n_fs, n_sighand;
+	unsigned n_fs;
+	int res = 0;
 
 	bprm->unsafe = tracehook_unsafe_exec(p);
 
 	n_fs = 1;
-	n_sighand = 1;
+	write_lock(&p->fs->lock);
 	lock_task_sighand(p, &flags);
 	for (t = next_thread(p); t != p; t = next_thread(t)) {
 		if (t->fs == p->fs)
 			n_fs++;
-		n_sighand++;
 	}
 
-	if (atomic_read(&p->fs->count) > n_fs ||
-	    atomic_read(&p->sighand->count) > n_sighand)
+	if (p->fs->users > n_fs) {
 		bprm->unsafe |= LSM_UNSAFE_SHARE;
+	} else {
+		if (p->fs->in_exec)
+			res = -EAGAIN;
+		p->fs->in_exec = 1;
+	}
 
 	unlock_task_sighand(p, &flags);
+	write_unlock(&p->fs->lock);
+
+	return res;
 }
 
 /* 
···
 	bprm->cred = prepare_exec_creds();
 	if (!bprm->cred)
 		goto out_unlock;
-	check_unsafe_exec(bprm);
+
+	retval = check_unsafe_exec(bprm);
+	if (retval)
+		goto out_unlock;
 
 	file = open_exec(filename);
 	retval = PTR_ERR(file);
 	if (IS_ERR(file))
-		goto out_unlock;
+		goto out_unmark;
 
 	sched_exec();
 
···
 		goto out;
 
 	/* execve succeeded */
+	write_lock(&current->fs->lock);
+	current->fs->in_exec = 0;
+	write_unlock(&current->fs->lock);
 	current->in_execve = 0;
 	mutex_unlock(&current->cred_exec_mutex);
 	acct_update_integrals(current);
···
 		allow_write_access(bprm->file);
 		fput(bprm->file);
 	}
+
+out_unmark:
+	write_lock(&current->fs->lock);
+	current->fs->in_exec = 0;
+	write_unlock(&current->fs->lock);
 
 out_unlock:
 	current->in_execve = 0;

fs/ext2/acl.c

···
 				return PTR_ERR(acl);
 		}
 		if (!acl)
-			inode->i_mode &= ~current->fs->umask;
+			inode->i_mode &= ~current_umask();
 	}
 	if (test_opt(inode->i_sb, POSIX_ACL) && acl) {
                struct posix_acl *clone;

fs/ext3/acl.c

···
 				return PTR_ERR(acl);
 		}
 		if (!acl)
-			inode->i_mode &= ~current->fs->umask;
+			inode->i_mode &= ~current_umask();
 	}
 	if (test_opt(inode->i_sb, POSIX_ACL) && acl) {
 		struct posix_acl *clone;

fs/ext4/acl.c

···
 				return PTR_ERR(acl);
 		}
 		if (!acl)
-			inode->i_mode &= ~current->fs->umask;
+			inode->i_mode &= ~current_umask();
 	}
 	if (test_opt(inode->i_sb, POSIX_ACL) && acl) {
 		struct posix_acl *clone;

fs/fat/inode.c

···
 
 	opts->fs_uid = current_uid();
 	opts->fs_gid = current_gid();
-	opts->fs_fmask = opts->fs_dmask = current->fs->umask;
+	opts->fs_fmask = current_umask();
 	opts->allow_utime = -1;
 	opts->codepage = fat_default_codepage;
 	opts->iocharset = fat_default_iocharset;

fs/fs_struct.c

···
+#include <linux/module.h>
+#include <linux/sched.h>
+#include <linux/fs.h>
+#include <linux/path.h>
+#include <linux/slab.h>
+#include <linux/fs_struct.h>
+
+/*
+ * Replace the fs->{rootmnt,root} with {mnt,dentry}. Put the old values.
+ * It can block.
+ */
+void set_fs_root(struct fs_struct *fs, struct path *path)
+{
+	struct path old_root;
+
+	write_lock(&fs->lock);
+	old_root = fs->root;
+	fs->root = *path;
+	path_get(path);
+	write_unlock(&fs->lock);
+	if (old_root.dentry)
+		path_put(&old_root);
+}
+
+/*
+ * Replace the fs->{pwdmnt,pwd} with {mnt,dentry}. Put the old values.
+ * It can block.
+ */
+void set_fs_pwd(struct fs_struct *fs, struct path *path)
+{
+	struct path old_pwd;
+
+	write_lock(&fs->lock);
+	old_pwd = fs->pwd;
+	fs->pwd = *path;
+	path_get(path);
+	write_unlock(&fs->lock);
+
+	if (old_pwd.dentry)
+		path_put(&old_pwd);
+}
+
+void chroot_fs_refs(struct path *old_root, struct path *new_root)
+{
+	struct task_struct *g, *p;
+	struct fs_struct *fs;
+	int count = 0;
+
+	read_lock(&tasklist_lock);
+	do_each_thread(g, p) {
+		task_lock(p);
+		fs = p->fs;
+		if (fs) {
+			write_lock(&fs->lock);
+			if (fs->root.dentry == old_root->dentry
+			    && fs->root.mnt == old_root->mnt) {
+				path_get(new_root);
+				fs->root = *new_root;
+				count++;
+			}
+			if (fs->pwd.dentry == old_root->dentry
+			    && fs->pwd.mnt == old_root->mnt) {
+				path_get(new_root);
+				fs->pwd = *new_root;
+				count++;
+			}
+			write_unlock(&fs->lock);
+		}
+		task_unlock(p);
+	} while_each_thread(g, p);
+	read_unlock(&tasklist_lock);
+	while (count--)
+		path_put(old_root);
+}
+
+void free_fs_struct(struct fs_struct *fs)
+{
+	path_put(&fs->root);
+	path_put(&fs->pwd);
+	kmem_cache_free(fs_cachep, fs);
+}
+
+void exit_fs(struct task_struct *tsk)
+{
+	struct fs_struct *fs = tsk->fs;
+
+	if (fs) {
+		int kill;
+		task_lock(tsk);
+		write_lock(&fs->lock);
+		tsk->fs = NULL;
+		kill = !--fs->users;
+		write_unlock(&fs->lock);
+		task_unlock(tsk);
+		if (kill)
+			free_fs_struct(fs);
+	}
+}
+
+struct fs_struct *copy_fs_struct(struct fs_struct *old)
+{
+	struct fs_struct *fs = kmem_cache_alloc(fs_cachep, GFP_KERNEL);
+	/* We don't need to lock fs - think why ;-) */
+	if (fs) {
+		fs->users = 1;
+		fs->in_exec = 0;
+		rwlock_init(&fs->lock);
+		fs->umask = old->umask;
+		read_lock(&old->lock);
+		fs->root = old->root;
+		path_get(&old->root);
+		fs->pwd = old->pwd;
+		path_get(&old->pwd);
+		read_unlock(&old->lock);
+	}
+	return fs;
+}
+
+int unshare_fs_struct(void)
+{
+	struct fs_struct *fs = current->fs;
+	struct fs_struct *new_fs = copy_fs_struct(fs);
+	int kill;
+
+	if (!new_fs)
+		return -ENOMEM;
+
+	task_lock(current);
+	write_lock(&fs->lock);
+	kill = !--fs->users;
+	current->fs = new_fs;
+	write_unlock(&fs->lock);
+	task_unlock(current);
+
+	if (kill)
+		free_fs_struct(fs);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(unshare_fs_struct);
+
+int current_umask(void)
+{
+	return current->fs->umask;
+}
+EXPORT_SYMBOL(current_umask);
+
+/* to be mentioned only in INIT_TASK */
+struct fs_struct init_fs = {
+	.users		= 1,
+	.lock		= __RW_LOCK_UNLOCKED(init_fs.lock),
+	.umask		= 0022,
+};
+
+void daemonize_fs_struct(void)
+{
+	struct fs_struct *fs = current->fs;
+
+	if (fs) {
+		int kill;
+
+		task_lock(current);
+
+		write_lock(&init_fs.lock);
+		init_fs.users++;
+		write_unlock(&init_fs.lock);
+
+		write_lock(&fs->lock);
+		current->fs = &init_fs;
+		kill = !--fs->users;
+		write_unlock(&fs->lock);
+
+		task_unlock(current);
+		if (kill)
+			free_fs_struct(fs);
+	}
+}

fs/generic_acl.c

···
 	mode_t mode = inode->i_mode;
 	int error;
 
-	inode->i_mode = mode & ~current->fs->umask;
+	inode->i_mode = mode & ~current_umask();
 	if (!S_ISLNK(inode->i_mode))
 		acl = ops->getacl(dir, ACL_TYPE_DEFAULT);
 	if (acl) {

fs/gfs2/acl.c

···
 	if (error)
 		return error;
 	if (!acl) {
-		mode &= ~current->fs->umask;
+		mode &= ~current_umask();
 		if (mode != ip->i_inode.i_mode)
 			error = munge_mode(ip, mode);
 		return error;

fs/hfsplus/options.c

···
 
 	opts->creator = HFSPLUS_DEF_CR_TYPE;
 	opts->type = HFSPLUS_DEF_CR_TYPE;
-	opts->umask = current->fs->umask;
+	opts->umask = current_umask();
 	opts->uid = current_uid();
 	opts->gid = current_gid();
 	opts->part = -1;

fs/hpfs/super.c

···
 
 	uid = current_uid();
 	gid = current_gid();
-	umask = current->fs->umask;
+	umask = current_umask();
 	lowercase = 0;
 	conv = CONV_BINARY;
 	eas = 2;

fs/internal.h

···
 
 struct super_block;
 struct linux_binprm;
+struct path;
 
 /*
  * block_dev.c
···
 /*
  * exec.c
  */
-extern void check_unsafe_exec(struct linux_binprm *);
+extern int check_unsafe_exec(struct linux_binprm *);
 
 /*
  * namespace.c
···
 extern struct vfsmount *copy_tree(struct vfsmount *, struct dentry *, int);
 
 extern void __init mnt_init(void);
+
+/*
+ * fs_struct.c
+ */
+extern void chroot_fs_refs(struct path *, struct path *);

fs/jffs2/acl.c

···
 		return PTR_ERR(acl);
 
 	if (!acl) {
-		*i_mode &= ~current->fs->umask;
+		*i_mode &= ~current_umask();
 	} else {
 		if (S_ISDIR(*i_mode))
 			jffs2_iset_acl(inode, &f->i_acl_default, acl);

fs/jfs/acl.c

···
 cleanup:
 		posix_acl_release(acl);
 	} else
-		inode->i_mode &= ~current->fs->umask;
+		inode->i_mode &= ~current_umask();
 
 	JFS_IP(inode)->mode2 = (JFS_IP(inode)->mode2 & 0xffff0000) |
 			       inode->i_mode;

fs/mpage.c

···
 	bio_put(bio);
 }
 
-struct bio *mpage_bio_submit(int rw, struct bio *bio)
+static struct bio *mpage_bio_submit(int rw, struct bio *bio)
 {
 	bio->bi_end_io = mpage_end_io_read;
 	if (rw == WRITE)
···
 	submit_bio(rw, bio);
 	return NULL;
 }
-EXPORT_SYMBOL(mpage_bio_submit);
 
 static struct bio *
 mpage_alloc(struct block_device *bdev,
···
  * just allocate full-size (16-page) BIOs.
  */
 
-int __mpage_writepage(struct page *page, struct writeback_control *wbc,
+struct mpage_data {
+	struct bio *bio;
+	sector_t last_block_in_bio;
+	get_block_t *get_block;
+	unsigned use_writepage;
+};
+
+static int __mpage_writepage(struct page *page, struct writeback_control *wbc,
 		      void *data)
 {
 	struct mpage_data *mpd = data;
···
 	mpd->bio = bio;
 	return ret;
 }
-EXPORT_SYMBOL(__mpage_writepage);
 
 /**
  * mpage_writepages - walk the list of dirty pages of the given address space & writepage() all of them

fs/namei.c

···
 #include <linux/file.h>
 #include <linux/fcntl.h>
 #include <linux/device_cgroup.h>
+#include <linux/fs_struct.h>
 #include <asm/uaccess.h>
 
 #define ACC_MODE(x) ("\000\004\002\006"[(x)&O_ACCMODE])
···
 	struct dentry *dir = nd->path.dentry;
 
 	if (!IS_POSIXACL(dir->d_inode))
-		mode &= ~current->fs->umask;
+		mode &= ~current_umask();
 	error = security_path_mknod(&nd->path, path->dentry, mode, 0);
 	if (error)
 		goto out_unlock;
···
 		goto out_unlock;
 	}
 	if (!IS_POSIXACL(nd.path.dentry->d_inode))
-		mode &= ~current->fs->umask;
+		mode &= ~current_umask();
 	error = may_mknod(mode);
 	if (error)
 		goto out_dput;
···
 		goto out_unlock;
 
 	if (!IS_POSIXACL(nd.path.dentry->d_inode))
-		mode &= ~current->fs->umask;
+		mode &= ~current_umask();
 	error = mnt_want_write(nd.path.mnt);
 	if (error)
 		goto out_dput;
···
 EXPORT_SYMBOL(vfs_unlink);
 EXPORT_SYMBOL(dentry_unhash);
 EXPORT_SYMBOL(generic_readlink);
-
-/* to be mentioned only in INIT_TASK */
-struct fs_struct init_fs = {
-	.count		= ATOMIC_INIT(1),
-	.lock		= __RW_LOCK_UNLOCKED(init_fs.lock),
-	.umask		= 0022,
-};

fs/namespace.c

···
 #include <linux/ramfs.h>
 #include <linux/log2.h>
 #include <linux/idr.h>
+#include <linux/fs_struct.h>
 #include <asm/uaccess.h>
 #include <asm/unistd.h>
 #include "pnode.h"
···
 out1:
 	free_page(type_page);
 	return retval;
-}
-
-/*
- * Replace the fs->{rootmnt,root} with {mnt,dentry}. Put the old values.
- * It can block. Requires the big lock held.
- */
-void set_fs_root(struct fs_struct *fs, struct path *path)
-{
-	struct path old_root;
-
-	write_lock(&fs->lock);
-	old_root = fs->root;
-	fs->root = *path;
-	path_get(path);
-	write_unlock(&fs->lock);
-	if (old_root.dentry)
-		path_put(&old_root);
-}
-
-/*
- * Replace the fs->{pwdmnt,pwd} with {mnt,dentry}. Put the old values.
- * It can block. Requires the big lock held.
- */
-void set_fs_pwd(struct fs_struct *fs, struct path *path)
-{
-	struct path old_pwd;
-
-	write_lock(&fs->lock);
-	old_pwd = fs->pwd;
-	fs->pwd = *path;
-	path_get(path);
-	write_unlock(&fs->lock);
-
-	if (old_pwd.dentry)
-		path_put(&old_pwd);
-}
-
-static void chroot_fs_refs(struct path *old_root, struct path *new_root)
-{
-	struct task_struct *g, *p;
-	struct fs_struct *fs;
-
-	read_lock(&tasklist_lock);
-	do_each_thread(g, p) {
-		task_lock(p);
-		fs = p->fs;
-		if (fs) {
-			atomic_inc(&fs->count);
-			task_unlock(p);
-			if (fs->root.dentry == old_root->dentry
-			    && fs->root.mnt == old_root->mnt)
-				set_fs_root(fs, new_root);
-			if (fs->pwd.dentry == old_root->dentry
-			    && fs->pwd.mnt == old_root->mnt)
-				set_fs_pwd(fs, new_root);
-			put_fs_struct(fs);
-		} else
-			task_unlock(p);
-	} while_each_thread(g, p);
-	read_unlock(&tasklist_lock);
 }
 
 /*

fs/nfs/nfs3proc.c

···
 		data->arg.create.verifier[1] = current->pid;
 	}
 
-	sattr->ia_mode &= ~current->fs->umask;
+	sattr->ia_mode &= ~current_umask();
 
 	for (;;) {
 		status = nfs3_do_create(dir, dentry, data);
···
 
 	dprintk("NFS call  mkdir %s\n", dentry->d_name.name);
 
-	sattr->ia_mode &= ~current->fs->umask;
+	sattr->ia_mode &= ~current_umask();
 
 	data = nfs3_alloc_createdata();
 	if (data == NULL)
···
 	dprintk("NFS call  mknod %s %u:%u\n", dentry->d_name.name,
 			MAJOR(rdev), MINOR(rdev));
 
-	sattr->ia_mode &= ~current->fs->umask;
+	sattr->ia_mode &= ~current_umask();
 
 	data = nfs3_alloc_createdata();
 	if (data == NULL)

fs/nfs/nfs4proc.c

···
 		attr.ia_mode = nd->intent.open.create_mode;
 		attr.ia_valid = ATTR_MODE;
 		if (!IS_POSIXACL(dir))
-			attr.ia_mode &= ~current->fs->umask;
+			attr.ia_mode &= ~current_umask();
 	} else {
 		attr.ia_valid = 0;
 		BUG_ON(nd->intent.open.flags & O_CREAT);

fs/nfsd/nfssvc.c

···
 nfsd(void *vrqstp)
 {
 	struct svc_rqst *rqstp = (struct svc_rqst *) vrqstp;
-	struct fs_struct *fsp;
 	int err, preverr = 0;
 
 	/* Lock module and set up kernel thread */
···
 	/* At this point, the thread shares current->fs
 	 * with the init process. We need to create files with a
 	 * umask of 0 instead of init's umask. */
-	fsp = copy_fs_struct(current->fs);
-	if (!fsp) {
+	if (unshare_fs_struct() < 0) {
 		printk("Unable to start nfsd thread: out of memory\n");
 		goto out;
 	}
-	exit_fs(current);
-	current->fs = fsp;
+
 	current->fs->umask = 0;
 
 	/*

fs/ocfs2/acl.c

···
 				return PTR_ERR(acl);
 		}
 		if (!acl)
-			inode->i_mode &= ~current->fs->umask;
+			inode->i_mode &= ~current_umask();
 	}
 	if ((osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL) && acl) {
 		struct posix_acl *clone;

fs/omfs/inode.c

···
 
 	sbi->s_uid = current_uid();
 	sbi->s_gid = current_gid();
-	sbi->s_dmask = sbi->s_fmask = current->fs->umask;
+	sbi->s_dmask = sbi->s_fmask = current_umask();
 
 	if (!parse_options((char *) data, sbi))
 		goto end;

fs/open.c

···
 #include <linux/rcupdate.h>
 #include <linux/audit.h>
 #include <linux/falloc.h>
+#include <linux/fs_struct.h>
 
 int vfs_statfs(struct dentry *dentry, struct kstatfs *buf)
 {

fs/proc/base.c

···
 #include <linux/oom.h>
 #include <linux/elf.h>
 #include <linux/pid_namespace.h>
+#include <linux/fs_struct.h>
 #include "internal.h"
 
 /* NOTE:

fs/proc/task_nommu.c

···
 #include <linux/mm.h>
 #include <linux/file.h>
 #include <linux/fdtable.h>
+#include <linux/fs_struct.h>
 #include <linux/mount.h>
 #include <linux/ptrace.h>
 #include <linux/seq_file.h>
···
 	else
 		bytes += kobjsize(mm);
 	
-	if (current->fs && atomic_read(&current->fs->count) > 1)
+	if (current->fs && current->fs->users > 1)
 		sbytes += kobjsize(current->fs);
 	else
 		bytes += kobjsize(current->fs);

fs/reiserfs/xattr_acl.c

···
 	} else {
 	      apply_umask:
 		/* no ACL, apply umask */
-		inode->i_mode &= ~current->fs->umask;
+		inode->i_mode &= ~current_umask();
 	}
 
 	return err;

fs/xfs/linux-2.6/xfs_iops.c

···
 	xfs_dentry_to_name(&name, dentry);
 
 	if (IS_POSIXACL(dir) && !default_acl)
-		mode &= ~current->fs->umask;
+		mode &= ~current_umask();
 
 	switch (mode & S_IFMT) {
 	case S_IFCHR:
···
 	mode_t		mode;
 
 	mode = S_IFLNK |
-		(irix_symlink_mode ? 0777 & ~current->fs->umask : S_IRWXUGO);
+		(irix_symlink_mode ? 0777 & ~current_umask() : S_IRWXUGO);
 	xfs_dentry_to_name(&name, dentry);
 
 	error = xfs_symlink(XFS_I(dir), &name, symname, mode, &cip, NULL);

include/linux/buffer_head.h

···
 
 static inline void buffer_init(void) {}
 static inline int try_to_free_buffers(struct page *page) { return 1; }
-static inline int sync_blockdev(struct block_device *bdev) { return 0; }
 static inline int inode_has_buffers(struct inode *inode) { return 0; }
 static inline void invalidate_inode_buffers(struct inode *inode) {}
 static inline int remove_inode_buffers(struct inode *inode) { return 1; }
 static inline int sync_mapping_buffers(struct address_space *mapping) { return 0; }
-static inline void invalidate_bdev(struct block_device *bdev) {}
-
-static inline struct super_block *freeze_bdev(struct block_device *sb)
-{
-	return NULL;
-}
-
-static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb)
-{
-	return 0;
-}
 
 #endif /* CONFIG_BLOCK */
 #endif /* _LINUX_BUFFER_HEAD_H */

include/linux/fs.h

···
 
 extern int vfs_statfs(struct dentry *, struct kstatfs *);
 
+extern int current_umask(void);
+
 /* /sys/fs */
 extern struct kobject *fs_kobj;
 
···
 extern int fsync_no_super(struct block_device *);
 #else
 static inline void bd_forget(struct inode *inode) {}
+static inline int sync_blockdev(struct block_device *bdev) { return 0; }
+static inline void invalidate_bdev(struct block_device *bdev) {}
+
+static inline struct super_block *freeze_bdev(struct block_device *sb)
+{
+	return NULL;
+}
+
+static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb)
+{
+	return 0;
+}
 #endif
 extern const struct file_operations def_blk_fops;
 extern const struct file_operations def_chr_fops;

include/linux/fs_struct.h

···
 #include <linux/path.h>
 
 struct fs_struct {
-	atomic_t count;	/* This usage count is used by check_unsafe_exec() for
-			 * security checking purposes - therefore it may not be
-			 * incremented, except by clone(CLONE_FS).
-			 */
+	int users;
 	rwlock_t lock;
 	int umask;
+	int in_exec;
 	struct path root, pwd;
 };
 
···
 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 *);
-extern void put_fs_struct(struct fs_struct *);
+extern void free_fs_struct(struct fs_struct *);
+extern void daemonize_fs_struct(void);
+extern int unshare_fs_struct(void);
 
 #endif /* _LINUX_FS_STRUCT_H */

include/linux/mnt_namespace.h

···
 	int event;
 };
 
+struct fs_struct;
+
 extern struct mnt_namespace *copy_mnt_ns(unsigned long, struct mnt_namespace *,
 		struct fs_struct *);
 extern void __put_mnt_ns(struct mnt_namespace *ns);

include/linux/mpage.h

···
  */
 #ifdef CONFIG_BLOCK
 
-struct mpage_data {
-	struct bio *bio;
-	sector_t last_block_in_bio;
-	get_block_t *get_block;
-	unsigned use_writepage;
-};
-
 struct writeback_control;
 
-struct bio *mpage_bio_submit(int rw, struct bio *bio);
 int mpage_readpages(struct address_space *mapping, struct list_head *pages,
 				unsigned nr_pages, get_block_t get_block);
 int mpage_readpage(struct page *page, get_block_t get_block);
-int __mpage_writepage(struct page *page, struct writeback_control *wbc,
-		      void *data);
 int mpage_writepages(struct address_space *mapping,
 		struct writeback_control *wbc, get_block_t get_block);
 int mpage_writepage(struct page *page, get_block_t *get_block,

include/linux/nsproxy.h

···
 struct uts_namespace;
 struct ipc_namespace;
 struct pid_namespace;
+struct fs_struct;
 
 /*
  * A structure to contain pointers to all per-process

include/linux/sched.h

···
 #include <linux/smp.h>
 #include <linux/sem.h>
 #include <linux/signal.h>
-#include <linux/fs_struct.h>
+#include <linux/path.h>
 #include <linux/compiler.h>
 #include <linux/completion.h>
 #include <linux/pid.h>
···
 struct robust_list_head;
 struct bio;
 struct bts_tracer;
+struct fs_struct;
 
 /*
  * List of flags we want to share for kernel threads,

init/do_mounts.c

···
 #include <linux/fs.h>
 #include <linux/initrd.h>
 #include <linux/async.h>
+#include <linux/fs_struct.h>
 
 #include <linux/nfs_fs.h>
 #include <linux/nfs_fs_sb.h>

ipc/mqueue.c

···
 		dentry->d_fsdata = attr;
 	}
 
-	mode &= ~current->fs->umask;
+	mode &= ~current_umask();
 	ret = mnt_want_write(mqueue_mnt);
 	if (ret)
 		goto out;

kernel/auditsc.c

···
 #include <linux/syscalls.h>
 #include <linux/inotify.h>
 #include <linux/capability.h>
+#include <linux/fs_struct.h>
 
 #include "audit.h"
 

kernel/exec_domain.c

···
 #include <linux/syscalls.h>
 #include <linux/sysctl.h>
 #include <linux/types.h>
+#include <linux/fs_struct.h>
 
 
 static void default_handler(int, struct pt_regs *);
···
 		module_put(ep->module);
 		return 0;
 	}
-
-	if (atomic_read(&current->fs->count) != 1) {
-		struct fs_struct *fsp, *ofsp;
-
-		fsp = copy_fs_struct(current->fs);
-		if (fsp == NULL) {
-			module_put(ep->module);
-			return -ENOMEM;
-		}
-
-		task_lock(current);
-		ofsp = current->fs;
-		current->fs = fsp;
-		task_unlock(current);
-
-		put_fs_struct(ofsp);
-	}
-
-	/*
-	 * At that point we are guaranteed to be the sole owner of
-	 * current->fs.
-	 */
 
 	current->personality = personality;
 	oep = current_thread_info()->exec_domain;

kernel/exit.c

···
 #include <linux/blkdev.h>
 #include <linux/task_io_accounting_ops.h>
 #include <linux/tracehook.h>
+#include <linux/fs_struct.h>
 #include <linux/init_task.h>
 #include <trace/sched.h>
 
···
 void daemonize(const char *name, ...)
 {
 	va_list args;
-	struct fs_struct *fs;
 	sigset_t blocked;
 
 	va_start(args, name);
···
 
 	/* Become as one with the init task */
 
-	exit_fs(current);	/* current->fs->count--; */
-	fs = init_task.fs;
-	current->fs = fs;
-	atomic_inc(&fs->count);
-
+	daemonize_fs_struct();
 	exit_files(current);
 	current->files = init_task.files;
 	atomic_inc(&current->files->count);
···
 		put_files_struct(files);
 	}
 }
-
-void put_fs_struct(struct fs_struct *fs)
-{
-	/* No need to hold fs->lock if we are killing it */
-	if (atomic_dec_and_test(&fs->count)) {
-		path_put(&fs->root);
-		path_put(&fs->pwd);
-		kmem_cache_free(fs_cachep, fs);
-	}
-}
-
-void exit_fs(struct task_struct *tsk)
-{
-	struct fs_struct * fs = tsk->fs;
-
-	if (fs) {
-		task_lock(tsk);
-		tsk->fs = NULL;
-		task_unlock(tsk);
-		put_fs_struct(fs);
-	}
-}
-
-EXPORT_SYMBOL_GPL(exit_fs);
 
 #ifdef CONFIG_MM_OWNER
 /*

kernel/fork.c

···
 #include <linux/tty.h>
 #include <linux/proc_fs.h>
 #include <linux/blkdev.h>
+#include <linux/fs_struct.h>
 #include <trace/sched.h>
 #include <linux/magic.h>
 
···
 	return retval;
 }
 
-static struct fs_struct *__copy_fs_struct(struct fs_struct *old)
-{
-	struct fs_struct *fs = kmem_cache_alloc(fs_cachep, GFP_KERNEL);
-	/* We don't need to lock fs - think why ;-) */
-	if (fs) {
-		atomic_set(&fs->count, 1);
-		rwlock_init(&fs->lock);
-		fs->umask = old->umask;
-		read_lock(&old->lock);
-		fs->root = old->root;
-		path_get(&old->root);
-		fs->pwd = old->pwd;
-		path_get(&old->pwd);
-		read_unlock(&old->lock);
-	}
-	return fs;
-}
-
-struct fs_struct *copy_fs_struct(struct fs_struct *old)
-{
-	return __copy_fs_struct(old);
-}
-
-EXPORT_SYMBOL_GPL(copy_fs_struct);
-
 static int copy_fs(unsigned long clone_flags, struct task_struct *tsk)
 {
+	struct fs_struct *fs = current->fs;
 	if (clone_flags & CLONE_FS) {
-		atomic_inc(&current->fs->count);
+		/* tsk->fs is already what we want */
+		write_lock(&fs->lock);
+		if (fs->in_exec) {
+			write_unlock(&fs->lock);
+			return -EAGAIN;
+		}
+		fs->users++;
+		write_unlock(&fs->lock);
 		return 0;
 	}
-	tsk->fs = __copy_fs_struct(current->fs);
+	tsk->fs = copy_fs_struct(fs);
 	if (!tsk->fs)
 		return -ENOMEM;
 	return 0;
···
 {
 	struct fs_struct *fs = current->fs;
 
-	if ((unshare_flags & CLONE_FS) &&
-	    (fs && atomic_read(&fs->count) > 1)) {
-		*new_fsp = __copy_fs_struct(current->fs);
-		if (!*new_fsp)
-			return -ENOMEM;
-	}
+	if (!(unshare_flags & CLONE_FS) || !fs)
+		return 0;
+
+	/* don't need lock here; in the worst case we'll do useless copy */
+	if (fs->users == 1)
+		return 0;
+
+	*new_fsp = copy_fs_struct(fs);
+	if (!*new_fsp)
+		return -ENOMEM;
 
 	return 0;
 }
···
 
 		if (new_fs) {
 			fs = current->fs;
+			write_lock(&fs->lock);
 			current->fs = new_fs;
-			new_fs = fs;
+			if (--fs->users)
+				new_fs = NULL;
+			else
+				new_fs = fs;
+			write_unlock(&fs->lock);
 		}
 
 		if (new_mm) {
···
 
 bad_unshare_cleanup_fs:
 	if (new_fs)
-		put_fs_struct(new_fs);
+		free_fs_struct(new_fs);
 
 bad_unshare_cleanup_thread:
 bad_unshare_out:

kernel/sys.c

···
 #include <linux/seccomp.h>
 #include <linux/cpu.h>
 #include <linux/ptrace.h>
+#include <linux/fs_struct.h>
 
 #include <linux/compat.h>
 #include <linux/syscalls.h>

net/unix/af_unix.c

···
 		 * All right, let's create it.
 		 */
 		mode = S_IFSOCK |
-		       (SOCK_INODE(sock)->i_mode & ~current->fs->umask);
+		       (SOCK_INODE(sock)->i_mode & ~current_umask());
 		err = mnt_want_write(nd.path.mnt);
 		if (err)
 			goto out_mknod_dput;

security/tomoyo/realpath.c

···
 #include <linux/types.h>
 #include <linux/mount.h>
 #include <linux/mnt_namespace.h>
+#include <linux/fs_struct.h>
 #include "common.h"
 #include "realpath.h"