1 files changed, 279 insertions, 90 deletions

diff --git a/fs/file_table.c b/fs/file_table.c
index dd88701e54a9..cd4a3db4659a 100644
--- a/fs/file_table.c
+++ b/fs/file_table.c
@@ -9,10 +9,10 @@
 #include <linux/string.h>
 #include <linux/slab.h>
 #include <linux/file.h>
-#include <linux/fdtable.h>
 #include <linux/init.h>
 #include <linux/module.h>
 #include <linux/fs.h>
+#include <linux/filelock.h>
 #include <linux/security.h>
 #include <linux/cred.h>
 #include <linux/eventpoll.h>
@@ -25,7 +25,6 @@
 #include <linux/percpu_counter.h>
 #include <linux/percpu.h>
 #include <linux/task_work.h>
-#include <linux/ima.h>
 #include <linux/swap.h>
 #include <linux/kmemleak.h>
 
@@ -39,24 +38,46 @@ static struct files_stat_struct files_stat = {
 };
 
 /* SLAB cache for file structures */
-static struct kmem_cache *filp_cachep __read_mostly;
+static struct kmem_cache *filp_cachep __ro_after_init;
+static struct kmem_cache *bfilp_cachep __ro_after_init;
 
 static struct percpu_counter nr_files __cacheline_aligned_in_smp;
 
-static void file_free_rcu(struct rcu_head *head)
+/* Container for backing file with optional user path */
+struct backing_file {
+	struct file file;
+	union {
+		struct path user_path;
+		freeptr_t bf_freeptr;
+	};
+};
+
+#define backing_file(f) container_of(f, struct backing_file, file)
+
+const struct path *backing_file_user_path(const struct file *f)
 {
-	struct file *f = container_of(head, struct file, f_rcuhead);
+	return &backing_file(f)->user_path;
+}
+EXPORT_SYMBOL_GPL(backing_file_user_path);
 
-	put_cred(f->f_cred);
-	kmem_cache_free(filp_cachep, f);
+void backing_file_set_user_path(struct file *f, const struct path *path)
+{
+	backing_file(f)->user_path = *path;
 }
+EXPORT_SYMBOL_GPL(backing_file_set_user_path);
 
 static inline void file_free(struct file *f)
 {
 	security_file_free(f);
-	if (!(f->f_mode & FMODE_NOACCOUNT))
+	if (likely(!(f->f_mode & FMODE_NOACCOUNT)))
 		percpu_counter_dec(&nr_files);
-	call_rcu(&f->f_rcuhead, file_free_rcu);
+	put_cred(f->f_cred);
+	if (unlikely(f->f_mode & FMODE_BACKING)) {
+		path_put(backing_file_user_path(f));
+		kmem_cache_free(bfilp_cachep, backing_file(f));
+	} else {
+		kmem_cache_free(filp_cachep, f);
+	}
 }
 
 /*
@@ -81,14 +102,14 @@ EXPORT_SYMBOL_GPL(get_max_files);
 /*
  * Handle nr_files sysctl
  */
-static int proc_nr_files(struct ctl_table *table, int write, void *buffer,
+static int proc_nr_files(const struct ctl_table *table, int write, void *buffer,
 			 size_t *lenp, loff_t *ppos)
 {
-	files_stat.nr_files = get_nr_files();
+	files_stat.nr_files = percpu_counter_sum_positive(&nr_files);
 	return proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
 }
 
-static struct ctl_table fs_stat_sysctls[] = {
+static const struct ctl_table fs_stat_sysctls[] = {
 	{
 		.procname	= "file-nr",
 		.data		= &files_stat,
@@ -110,11 +131,10 @@ static struct ctl_table fs_stat_sysctls[] = {
 		.data		= &sysctl_nr_open,
 		.maxlen		= sizeof(unsigned int),
 		.mode		= 0644,
-		.proc_handler	= proc_dointvec_minmax,
+		.proc_handler	= proc_douintvec_minmax,
 		.extra1		= &sysctl_nr_open_min,
 		.extra2		= &sysctl_nr_open_max,
 	},
-	{ }
 };
 
 static int __init init_fs_stat_sysctls(void)
@@ -122,6 +142,7 @@ static int __init init_fs_stat_sysctls(void)
 	register_sysctl_init("fs", fs_stat_sysctls);
 	if (IS_ENABLED(CONFIG_BINFMT_MISC)) {
 		struct ctl_table_header *hdr;
+
 		hdr = register_sysctl_mount_point("fs/binfmt_misc");
 		kmemleak_not_leak(hdr);
 	}
@@ -130,31 +151,58 @@ static int __init init_fs_stat_sysctls(void)
 fs_initcall(init_fs_stat_sysctls);
 #endif
 
-static struct file *__alloc_file(int flags, const struct cred *cred)
+static int init_file(struct file *f, int flags, const struct cred *cred)
 {
-	struct file *f;
 	int error;
 
-	f = kmem_cache_zalloc(filp_cachep, GFP_KERNEL);
-	if (unlikely(!f))
-		return ERR_PTR(-ENOMEM);
-
 	f->f_cred = get_cred(cred);
 	error = security_file_alloc(f);
 	if (unlikely(error)) {
-		file_free_rcu(&f->f_rcuhead);
-		return ERR_PTR(error);
+		put_cred(f->f_cred);
+		return error;
 	}
 
-	atomic_long_set(&f->f_count, 1);
-	rwlock_init(&f->f_owner.lock);
 	spin_lock_init(&f->f_lock);
+	/*
+	 * Note that f_pos_lock is only used for files raising
+	 * FMODE_ATOMIC_POS and directories. Other files such as pipes
+	 * don't need it and since f_pos_lock is in a union may reuse
+	 * the space for other purposes. They are expected to initialize
+	 * the respective member when opening the file.
+	 */
 	mutex_init(&f->f_pos_lock);
-	f->f_flags = flags;
-	f->f_mode = OPEN_FMODE(flags);
-	/* f->f_version: 0 */
+	memset(&f->__f_path, 0, sizeof(f->f_path));
+	memset(&f->f_ra, 0, sizeof(f->f_ra));
+
+	f->f_flags	= flags;
+	f->f_mode	= OPEN_FMODE(flags);
+
+	f->f_op		= NULL;
+	f->f_mapping	= NULL;
+	f->private_data = NULL;
+	f->f_inode	= NULL;
+	f->f_owner	= NULL;
+#ifdef CONFIG_EPOLL
+	f->f_ep		= NULL;
+#endif
 
-	return f;
+	f->f_iocb_flags = 0;
+	f->f_pos	= 0;
+	f->f_wb_err	= 0;
+	f->f_sb_err	= 0;
+
+	/*
+	 * We're SLAB_TYPESAFE_BY_RCU so initialize f_ref last. While
+	 * fget-rcu pattern users need to be able to handle spurious
+	 * refcount bumps we should reinitialize the reused file first.
+	 */
+	file_ref_init(&f->f_ref, 1);
+	/*
+	 * Disable permission and pre-content events for all files by default.
+	 * They may be enabled later by fsnotify_open_perm_and_set_mode().
+	 */
+	file_set_fsnotify_mode(f, FMODE_NONOTIFY_PERM);
+	return 0;
 }
 
 /* Find an unused file structure and return a pointer to it.
@@ -171,11 +219,13 @@ struct file *alloc_empty_file(int flags, const struct cred *cred)
 {
 	static long old_max;
 	struct file *f;
+	int error;
 
 	/*
 	 * Privileged users can go above max_files
 	 */
-	if (get_nr_files() >= files_stat.max_files && !capable(CAP_SYS_ADMIN)) {
+	if (unlikely(get_nr_files() >= files_stat.max_files) &&
+	    !capable(CAP_SYS_ADMIN)) {
 		/*
 		 * percpu_counters are inaccurate.  Do an expensive check before
 		 * we go and fail.
@@ -184,9 +234,17 @@ struct file *alloc_empty_file(int flags, const struct cred *cred)
 			goto over;
 	}
 
-	f = __alloc_file(flags, cred);
-	if (!IS_ERR(f))
-		percpu_counter_inc(&nr_files);
+	f = kmem_cache_alloc(filp_cachep, GFP_KERNEL);
+	if (unlikely(!f))
+		return ERR_PTR(-ENOMEM);
+
+	error = init_file(f, flags, cred);
+	if (unlikely(error)) {
+		kmem_cache_free(filp_cachep, f);
+		return ERR_PTR(error);
+	}
+
+	percpu_counter_inc(&nr_files);
 
 	return f;
 
@@ -202,35 +260,66 @@ over:
 /*
  * Variant of alloc_empty_file() that doesn't check and modify nr_files.
  *
- * Should not be used unless there's a very good reason to do so.
+ * This is only for kernel internal use, and the allocate file must not be
+ * installed into file tables or such.
  */
 struct file *alloc_empty_file_noaccount(int flags, const struct cred *cred)
 {
-	struct file *f = __alloc_file(flags, cred);
+	struct file *f;
+	int error;
 
-	if (!IS_ERR(f))
-		f->f_mode |= FMODE_NOACCOUNT;
+	f = kmem_cache_alloc(filp_cachep, GFP_KERNEL);
+	if (unlikely(!f))
+		return ERR_PTR(-ENOMEM);
+
+	error = init_file(f, flags, cred);
+	if (unlikely(error)) {
+		kmem_cache_free(filp_cachep, f);
+		return ERR_PTR(error);
+	}
+
+	f->f_mode |= FMODE_NOACCOUNT;
 
 	return f;
 }
 
+/*
+ * Variant of alloc_empty_file() that allocates a backing_file container
+ * and doesn't check and modify nr_files.
+ *
+ * This is only for kernel internal use, and the allocate file must not be
+ * installed into file tables or such.
+ */
+struct file *alloc_empty_backing_file(int flags, const struct cred *cred)
+{
+	struct backing_file *ff;
+	int error;
+
+	ff = kmem_cache_alloc(bfilp_cachep, GFP_KERNEL);
+	if (unlikely(!ff))
+		return ERR_PTR(-ENOMEM);
+
+	error = init_file(&ff->file, flags, cred);
+	if (unlikely(error)) {
+		kmem_cache_free(bfilp_cachep, ff);
+		return ERR_PTR(error);
+	}
+
+	ff->file.f_mode |= FMODE_BACKING | FMODE_NOACCOUNT;
+	return &ff->file;
+}
+
 /**
- * alloc_file - allocate and initialize a 'struct file'
+ * file_init_path - initialize a 'struct file' based on path
  *
+ * @file: the file to set up
  * @path: the (dentry, vfsmount) pair for the new file
- * @flags: O_... flags with which the new file will be opened
  * @fop: the 'struct file_operations' for the new file
  */
-static struct file *alloc_file(const struct path *path, int flags,
-		const struct file_operations *fop)
+static void file_init_path(struct file *file, const struct path *path,
+			   const struct file_operations *fop)
 {
-	struct file *file;
-
-	file = alloc_empty_file(flags, current_cred());
-	if (IS_ERR(file))
-		return file;
-
-	file->f_path = *path;
+	file->__f_path = *path;
 	file->f_inode = path->dentry->d_inode;
 	file->f_mapping = path->dentry->d_inode->i_mapping;
 	file->f_wb_err = filemap_sample_wb_err(file->f_mapping);
@@ -248,40 +337,99 @@ static struct file *alloc_file(const struct path *path, int flags,
 	file->f_op = fop;
 	if ((file->f_mode & (FMODE_READ | FMODE_WRITE)) == FMODE_READ)
 		i_readcount_inc(path->dentry->d_inode);
+}
+
+/**
+ * alloc_file - allocate and initialize a 'struct file'
+ *
+ * @path: the (dentry, vfsmount) pair for the new file
+ * @flags: O_... flags with which the new file will be opened
+ * @fop: the 'struct file_operations' for the new file
+ */
+static struct file *alloc_file(const struct path *path, int flags,
+		const struct file_operations *fop)
+{
+	struct file *file;
+
+	file = alloc_empty_file(flags, current_cred());
+	if (!IS_ERR(file))
+		file_init_path(file, path, fop);
 	return file;
 }
 
+static inline int alloc_path_pseudo(const char *name, struct inode *inode,
+				    struct vfsmount *mnt, struct path *path)
+{
+	path->dentry = d_alloc_pseudo(mnt->mnt_sb, &QSTR(name));
+	if (!path->dentry)
+		return -ENOMEM;
+	path->mnt = mntget(mnt);
+	d_instantiate(path->dentry, inode);
+	return 0;
+}
+
 struct file *alloc_file_pseudo(struct inode *inode, struct vfsmount *mnt,
-				const char *name, int flags,
-				const struct file_operations *fops)
+			       const char *name, int flags,
+			       const struct file_operations *fops)
 {
-	static const struct dentry_operations anon_ops = {
-		.d_dname = simple_dname
-	};
-	struct qstr this = QSTR_INIT(name, strlen(name));
+	int ret;
 	struct path path;
 	struct file *file;
 
-	path.dentry = d_alloc_pseudo(mnt->mnt_sb, &this);
-	if (!path.dentry)
-		return ERR_PTR(-ENOMEM);
-	if (!mnt->mnt_sb->s_d_op)
-		d_set_d_op(path.dentry, &anon_ops);
-	path.mnt = mntget(mnt);
-	d_instantiate(path.dentry, inode);
+	ret = alloc_path_pseudo(name, inode, mnt, &path);
+	if (ret)
+		return ERR_PTR(ret);
+
 	file = alloc_file(&path, flags, fops);
 	if (IS_ERR(file)) {
 		ihold(inode);
 		path_put(&path);
+		return file;
 	}
+	/*
+	 * Disable all fsnotify events for pseudo files by default.
+	 * They may be enabled by caller with file_set_fsnotify_mode().
+	 */
+	file_set_fsnotify_mode(file, FMODE_NONOTIFY);
 	return file;
 }
 EXPORT_SYMBOL(alloc_file_pseudo);
 
+struct file *alloc_file_pseudo_noaccount(struct inode *inode,
+					 struct vfsmount *mnt, const char *name,
+					 int flags,
+					 const struct file_operations *fops)
+{
+	int ret;
+	struct path path;
+	struct file *file;
+
+	ret = alloc_path_pseudo(name, inode, mnt, &path);
+	if (ret)
+		return ERR_PTR(ret);
+
+	file = alloc_empty_file_noaccount(flags, current_cred());
+	if (IS_ERR(file)) {
+		ihold(inode);
+		path_put(&path);
+		return file;
+	}
+	file_init_path(file, &path, fops);
+	/*
+	 * Disable all fsnotify events for pseudo files by default.
+	 * They may be enabled by caller with file_set_fsnotify_mode().
+	 */
+	file_set_fsnotify_mode(file, FMODE_NONOTIFY);
+	return file;
+}
+EXPORT_SYMBOL_GPL(alloc_file_pseudo_noaccount);
+
 struct file *alloc_file_clone(struct file *base, int flags,
 				const struct file_operations *fops)
 {
-	struct file *f = alloc_file(&base->f_path, flags, fops);
+	struct file *f;
+
+	f = alloc_file(&base->f_path, flags, fops);
 	if (!IS_ERR(f)) {
 		path_get(&f->f_path);
 		f->f_mapping = base->f_mapping;
@@ -311,7 +459,7 @@ static void __fput(struct file *file)
 	eventpoll_release(file);
 	locks_remove_file(file);
 
-	ima_file_free(file);
+	security_file_release(file);
 	if (unlikely(file->f_flags & FASYNC)) {
 		if (file->f_op->fasync)
 			file->f_op->fasync(-1, file, 0);
@@ -323,7 +471,7 @@ static void __fput(struct file *file)
 		cdev_put(inode->i_cdev);
 	}
 	fops_put(file->f_op);
-	put_pid(file->f_owner.pid);
+	file_f_owner_release(file);
 	put_file_access(file);
 	dput(dentry);
 	if (unlikely(mode & FMODE_NEED_UNMOUNT))
@@ -345,9 +493,11 @@ static void delayed_fput(struct work_struct *unused)
 
 static void ____fput(struct callback_head *work)
 {
-	__fput(container_of(work, struct file, f_rcuhead));
+	__fput(container_of(work, struct file, f_task_work));
 }
 
+static DECLARE_DELAYED_WORK(delayed_fput_work, delayed_fput);
+
 /*
  * If kernel thread really needs to have the final fput() it has done
  * to complete, call this.  The only user right now is the boot - we
@@ -361,31 +511,40 @@ static void ____fput(struct callback_head *work)
 void flush_delayed_fput(void)
 {
 	delayed_fput(NULL);
+	flush_delayed_work(&delayed_fput_work);
 }
 EXPORT_SYMBOL_GPL(flush_delayed_fput);
 
-static DECLARE_DELAYED_WORK(delayed_fput_work, delayed_fput);
+static void __fput_deferred(struct file *file)
+{
+	struct task_struct *task = current;
+
+	if (unlikely(!(file->f_mode & (FMODE_BACKING | FMODE_OPENED)))) {
+		file_free(file);
+		return;
+	}
+
+	if (likely(!in_interrupt() && !(task->flags & PF_KTHREAD))) {
+		init_task_work(&file->f_task_work, ____fput);
+		if (!task_work_add(task, &file->f_task_work, TWA_RESUME))
+			return;
+		/*
+		 * After this task has run exit_task_work(),
+		 * task_work_add() will fail.  Fall through to delayed
+		 * fput to avoid leaking *file.
+		 */
+	}
+
+	if (llist_add(&file->f_llist, &delayed_fput_list))
+		schedule_delayed_work(&delayed_fput_work, 1);
+}
 
 void fput(struct file *file)
 {
-	if (atomic_long_dec_and_test(&file->f_count)) {
-		struct task_struct *task = current;
-
-		if (likely(!in_interrupt() && !(task->flags & PF_KTHREAD))) {
-			init_task_work(&file->f_rcuhead, ____fput);
-			if (!task_work_add(task, &file->f_rcuhead, TWA_RESUME))
-				return;
-			/*
-			 * After this task has run exit_task_work(),
-			 * task_work_add() will fail.  Fall through to delayed
-			 * fput to avoid leaking *file.
-			 */
-		}
-
-		if (llist_add(&file->f_llist, &delayed_fput_list))
-			schedule_delayed_work(&delayed_fput_work, 1);
-	}
+	if (unlikely(file_ref_put(&file->f_ref)))
+		__fput_deferred(file);
 }
+EXPORT_SYMBOL(fput);
 
 /*
  * synchronous analog of fput(); for kernel threads that might be needed
@@ -397,20 +556,50 @@ void fput(struct file *file)
  */
 void __fput_sync(struct file *file)
 {
-	if (atomic_long_dec_and_test(&file->f_count)) {
-		struct task_struct *task = current;
-		BUG_ON(!(task->flags & PF_KTHREAD));
+	if (file_ref_put(&file->f_ref))
 		__fput(file);
-	}
 }
-
-EXPORT_SYMBOL(fput);
 EXPORT_SYMBOL(__fput_sync);
 
+/*
+ * Equivalent to __fput_sync(), but optimized for being called with the last
+ * reference.
+ *
+ * See file_ref_put_close() for details.
+ */
+void fput_close_sync(struct file *file)
+{
+	if (likely(file_ref_put_close(&file->f_ref)))
+		__fput(file);
+}
+
+/*
+ * Equivalent to fput(), but optimized for being called with the last
+ * reference.
+ *
+ * See file_ref_put_close() for details.
+ */
+void fput_close(struct file *file)
+{
+	if (file_ref_put_close(&file->f_ref))
+		__fput_deferred(file);
+}
+
 void __init files_init(void)
 {
-	filp_cachep = kmem_cache_create("filp", sizeof(struct file), 0,
-			SLAB_HWCACHE_ALIGN | SLAB_PANIC | SLAB_ACCOUNT, NULL);
+	struct kmem_cache_args args = {
+		.use_freeptr_offset = true,
+		.freeptr_offset = offsetof(struct file, f_freeptr),
+	};
+
+	filp_cachep = kmem_cache_create("filp", sizeof(struct file), &args,
+				SLAB_HWCACHE_ALIGN | SLAB_PANIC |
+				SLAB_ACCOUNT | SLAB_TYPESAFE_BY_RCU);
+
+	args.freeptr_offset = offsetof(struct backing_file, bf_freeptr);
+	bfilp_cachep = kmem_cache_create("bfilp", sizeof(struct backing_file),
+				&args, SLAB_HWCACHE_ALIGN | SLAB_PANIC |
+				SLAB_ACCOUNT | SLAB_TYPESAFE_BY_RCU);
 	percpu_counter_init(&nr_files, 0, GFP_KERNEL);
 }