Linux 3.6 compat, kern_path_locked() added

The kern_path_parent() function was removed from Linux 3.6 because
it was observed that all the callers just want the parent dentry.
The simpler kern_path_locked() function replaces kern_path_parent()
and does the lookup while holding the ->i_mutex lock.

This is good news for the vn implementation because it removes the
need for us to handle the locking.  However, it makes it harder to
implement a single readable vn_remove()/vn_rename() function which
is usually what we prefer.

Therefore, we implement a new version of vn_remove()/vn_rename()
for Linux 3.6 and newer kernels.  This allows us to leave the
existing working implementation untouched, and to add a simpler
version for newer kernels.

Long term I would very much like to see all of the vn code removed
since what this code enabled is generally frowned upon in the kernel.
But that can't happen util we either abondon the zpool.cache file
or implement alternate infrastructure to update is correctly in
user space.

Signed-off-by: Yuxuan Shui <[email protected]>
Signed-off-by: Richard Yao <[email protected]>
Signed-off-by: Brian Behlendorf <[email protected]>
Closes #154

3 files changed, 158 insertions, 0 deletions

diff --git a/config/spl-build.m4 b/config/spl-build.m4
index 0c7a03cf1..eb644a188 100644
--- a/config/spl-build.m4
+++ b/config/spl-build.m4
@@ -86,6 +86,7 @@ AC_DEFUN([SPL_AC_CONFIG_KERNEL], [
 	SPL_AC_SHRINK_ICACHE_MEMORY
 	SPL_AC_KERN_PATH_PARENT_HEADER
 	SPL_AC_KERN_PATH_PARENT_SYMBOL
+	SPL_AC_KERN_PATH_LOCKED
 	SPL_AC_2ARGS_ZLIB_DEFLATE_WORKSPACESIZE
 	SPL_AC_SHRINK_CONTROL_STRUCT
 	SPL_AC_RWSEM_SPINLOCK_IS_RAW
@@ -2188,6 +2189,21 @@ AC_DEFUN([SPL_AC_KERN_PATH_PARENT_SYMBOL],
 ])
 
 dnl #
+dnl # 3.6 API compat,
+dnl # The kern_path_parent() function was replaced by the kern_path_locked()
+dnl # function to eliminate all struct nameidata usage outside fs/namei.c.
+dnl #
+AC_DEFUN([SPL_AC_KERN_PATH_LOCKED], [
+	SPL_CHECK_SYMBOL_HEADER(
+		[kern_path_locked],
+		[struct dentry \*kern_path_locked(const char \*, struct path \*)],
+		[include/linux/namei.h],
+		[AC_DEFINE(HAVE_KERN_PATH_LOCKED, 1,
+		[kern_path_locked() is available])],
+		[])
+])
+
+dnl #
 dnl # 2.6.39 API compat,
 dnl # The function zlib_deflate_workspacesize() now take 2 arguments.
 dnl # This was done to avoid always having to allocate the maximum size
diff --git a/include/linux/file_compat.h b/include/linux/file_compat.h
index 2b5b7d225..27819d5e5 100644
--- a/include/linux/file_compat.h
+++ b/include/linux/file_compat.h
@@ -83,6 +83,12 @@ extern kern_path_parent_t kern_path_parent_fn;
 # define spl_kern_path_parent(path, nd)	path_lookup(path, LOOKUP_PARENT, nd)
 #endif /* HAVE_KERN_PATH_PARENT_HEADER */
 
+#ifdef HAVE_KERN_PATH_LOCKED
+typedef struct dentry * (*kern_path_locked_t)(const char *, struct path *);
+extern kern_path_locked_t kern_path_locked_fn;
+# define spl_kern_path_locked(name, path)	kern_path_locked_fn(name, path)
+#endif /* HAVE_KERN_PATH_LOCKED */
+
 #ifndef HAVE_CLEAR_CLOSE_ON_EXEC
 #define __clear_close_on_exec(fd, fdt)	FD_CLR(fd, fdt->close_on_exec)
 #endif
diff --git a/module/spl/spl-vnode.c b/module/spl/spl-vnode.c
index f5fc65d26..a0fed3201 100644
--- a/module/spl/spl-vnode.c
+++ b/module/spl/spl-vnode.c
@@ -50,6 +50,10 @@ EXPORT_SYMBOL(kern_path_parent_fn);
 #endif /* HAVE_KERN_PATH_PARENT_SYMBOL */
 #endif /* HAVE_KERN_PATH_PARENT_HEADER */
 
+#ifdef HAVE_KERN_PATH_LOCKED
+kern_path_locked_t kern_path_locked_fn = SYMBOL_POISON;
+#endif /* HAVE_KERN_PATH_LOCKED */
+
 vtype_t
 vn_mode_to_vtype(mode_t mode)
 {
@@ -298,6 +302,128 @@ vn_seek(vnode_t *vp, offset_t ooff, offset_t *noffp, void *ct)
 }
 EXPORT_SYMBOL(vn_seek);
 
+#ifdef HAVE_KERN_PATH_LOCKED
+/* Based on do_unlinkat() from linux/fs/namei.c */
+int
+vn_remove(const char *path, uio_seg_t seg, int flags)
+{
+	struct dentry *dentry;
+	struct path parent;
+	struct inode *inode = NULL;
+	int rc = 0;
+	SENTRY;
+
+	ASSERT(seg == UIO_SYSSPACE);
+	ASSERT(flags == RMFILE);
+
+	dentry = spl_kern_path_locked(path, &parent);
+	rc = PTR_ERR(dentry);
+	if (!IS_ERR(dentry)) {
+		if (parent.dentry->d_name.name[parent.dentry->d_name.len])
+			SGOTO(slashes, rc = 0);
+
+		inode = dentry->d_inode;
+		if (!inode)
+			SGOTO(slashes, rc = 0);
+
+		if (inode)
+			ihold(inode);
+
+		rc = vfs_unlink(parent.dentry->d_inode, dentry);
+exit1:
+		dput(dentry);
+	}
+
+	spl_inode_unlock(parent.dentry->d_inode);
+	if (inode)
+		iput(inode);    /* truncate the inode here */
+
+	path_put(&parent);
+	SRETURN(-rc);
+
+slashes:
+	rc = !dentry->d_inode ? -ENOENT :
+	    S_ISDIR(dentry->d_inode->i_mode) ? -EISDIR : -ENOTDIR;
+	SGOTO(exit1, rc);
+} /* vn_remove() */
+EXPORT_SYMBOL(vn_remove);
+
+/* Based on do_rename() from linux/fs/namei.c */
+int
+vn_rename(const char *oldname, const char *newname, int x1)
+{
+	struct dentry *old_dir, *new_dir;
+	struct dentry *old_dentry, *new_dentry;
+	struct dentry *trap;
+	struct path old_parent, new_parent;
+	int rc = 0;
+	SENTRY;
+
+	old_dentry = spl_kern_path_locked(oldname, &old_parent);
+	if (IS_ERR(old_dentry))
+		SGOTO(exit, rc = PTR_ERR(old_dentry));
+
+	spl_inode_unlock(old_parent.dentry->d_inode);
+
+	new_dentry = spl_kern_path_locked(newname, &new_parent);
+	if (IS_ERR(new_dentry))
+		SGOTO(exit2, rc = PTR_ERR(new_dentry));
+
+	spl_inode_unlock(new_parent.dentry->d_inode);
+
+	rc = -EXDEV;
+	if (old_parent.mnt != new_parent.mnt)
+		SGOTO(exit3, rc);
+
+	old_dir = old_parent.dentry;
+	new_dir = new_parent.dentry;
+	trap = lock_rename(new_dir, old_dir);
+
+	/* source should not be ancestor of target */
+	rc = -EINVAL;
+	if (old_dentry == trap)
+		SGOTO(exit4, rc);
+
+	/* target should not be an ancestor of source */
+	rc = -ENOTEMPTY;
+	if (new_dentry == trap)
+		SGOTO(exit4, rc);
+
+	/* source must exist */
+	rc = -ENOENT;
+	if (!old_dentry->d_inode)
+		SGOTO(exit4, rc);
+
+	/* unless the source is a directory trailing slashes give -ENOTDIR */
+	if (!S_ISDIR(old_dentry->d_inode->i_mode)) {
+		rc = -ENOTDIR;
+		if (old_dentry->d_name.name[old_dentry->d_name.len])
+			SGOTO(exit4, rc);
+		if (new_dentry->d_name.name[new_dentry->d_name.len])
+			SGOTO(exit4, rc);
+	}
+
+#ifdef HAVE_4ARGS_VFS_RENAME
+	rc = vfs_rename(old_dir->d_inode, old_dentry,
+			new_dir->d_inode, new_dentry);
+#else
+	rc = vfs_rename(old_dir->d_inode, old_dentry, oldnd.nd_mnt,
+			new_dir->d_inode, new_dentry, newnd.nd_mnt);
+#endif /* HAVE_4ARGS_VFS_RENAME */
+exit4:
+	unlock_rename(new_dir, old_dir);
+exit3:
+	dput(new_dentry);
+	path_put(&new_parent);
+exit2:
+	dput(old_dentry);
+	path_put(&old_parent);
+exit:
+	SRETURN(-rc);
+}
+EXPORT_SYMBOL(vn_rename);
+
+#else
 static struct dentry *
 vn_lookup_hash(struct nameidata *nd)
 {
@@ -458,6 +584,7 @@ exit:
         SRETURN(-rc);
 }
 EXPORT_SYMBOL(vn_rename);
+#endif /* HAVE_KERN_PATH_LOCKED */
 
 int
 vn_getattr(vnode_t *vp, vattr_t *vap, int flags, void *x3, void *x4)
@@ -862,6 +989,15 @@ int spl_vn_init_kallsyms_lookup(void)
 #endif /* HAVE_KERN_PATH_PARENT_SYMBOL */
 #endif /* HAVE_KERN_PATH_PARENT_HEADER */
 
+#ifdef HAVE_KERN_PATH_LOCKED
+        kern_path_locked_fn = (kern_path_locked_t)
+                spl_kallsyms_lookup_name("kern_path_locked");
+        if (!kern_path_locked_fn) {
+                printk(KERN_ERR "Error: Unknown symbol kern_path_locked\n");
+                return -EFAULT;
+        }
+#endif
+
 	return (0);
 }