Linux 3.11 compat: fops->iterate()

Commit torvalds/linux@2233f31aad
replaced ->readdir() with ->iterate() in struct file_operations.
All filesystems must now use the new ->iterate method.

To handle this the code was reworked to use the new ->iterate
interface.  Care was taken to keep the majority of changes
confined to the ZPL layer which is already Linux specific.
However, minor changes were required to the common zfs_readdir()
function.

Compatibility with older kernels was accomplished by adding
versions of the trivial dir_emit* helper functions.  Also the
various *_readdir() functions were reworked in to wrappers
which create a dir_context structure to pass to the new
*_iterate() functions.

Unfortunately, the new dir_emit* functions prevent us from
passing a private pointer to the filldir function.  The xattr
directory code leveraged this ability through zfs_readdir()
to generate the list of xattr names.  Since we can no longer
use zfs_readdir() a simplified zpl_xattr_readdir() function
was added to perform the same task.

Signed-off-by: Richard Yao <ryao@cs.stonybrook.edu>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes #1653
Issue #1591

config/kernel-vfs-iterate.m4

@@ -0,0 +1,43 @@
+dnl #
+dnl # 3.11 API change
+dnl #
+AC_DEFUN([ZFS_AC_KERNEL_VFS_ITERATE], [
+	AC_MSG_CHECKING([whether fops->iterate() is available])
+	ZFS_LINUX_TRY_COMPILE([
+		#include <linux/fs.h>
+		int iterate(struct file *filp, struct dir_context * context)
+		    { return 0; }
+
+		static const struct file_operations fops
+		    __attribute__ ((unused)) = {
+			.iterate	 = iterate,
+		};
+	],[
+	],[
+		AC_MSG_RESULT(yes)
+		AC_DEFINE(HAVE_VFS_ITERATE, 1,
+		          [fops->iterate() is available])
+	],[
+		AC_MSG_RESULT(no)
+
+		AC_MSG_CHECKING([whether fops->readdir() is available])
+		ZFS_LINUX_TRY_COMPILE([
+			#include <linux/fs.h>
+			int readdir(struct file *filp, void *entry, filldir_t func)
+			    { return 0; }
+
+			static const struct file_operations fops
+			    __attribute__ ((unused)) = {
+				.readdir = readdir,
+			};
+		],[
+		],[
+			AC_MSG_RESULT(yes)
+			AC_DEFINE(HAVE_VFS_READDIR, 1,
+				  [fops->readdir() is available])
+		],[
+			AC_MSG_ERROR(no; file a bug report with ZFSOnLinux)
+		])
+
+	])
+])

config/kernel.m4

@@ -80,6 +80,7 @@ AC_DEFUN([ZFS_AC_CONFIG_KERNEL], [
 	ZFS_AC_KERNEL_ELEVATOR_CHANGE
 	ZFS_AC_KERNEL_5ARG_SGET
 	ZFS_AC_KERNEL_LSEEK_EXECUTE
+	ZFS_AC_KERNEL_VFS_ITERATE
 
 	AS_IF([test "$LINUX_OBJ" != "$LINUX"], [
 		KERNELMAKE_PARAMS="$KERNELMAKE_PARAMS O=$LINUX_OBJ"

include/sys/zfs_vnops.h

@@ -31,6 +31,7 @@
 #include <sys/cred.h>
 #include <sys/fcntl.h>
 #include <sys/pathname.h>
+#include <sys/zpl.h>
 
 #ifdef	__cplusplus
 extern "C" {
@@ -51,8 +52,7 @@ extern int zfs_mkdir(struct inode *dip, char *dirname, vattr_t *vap,
     struct inode **ipp, cred_t *cr, int flags, vsecattr_t *vsecp);
 extern int zfs_rmdir(struct inode *dip, char *name, struct inode *cwd,
     cred_t *cr, int flags);
-extern int zfs_readdir(struct inode *ip, void *dirent, filldir_t filldir,
+extern int zfs_readdir(struct inode *ip, struct dir_context *ctx, cred_t *cr);
-    loff_t *pos, cred_t *cr);
 extern int zfs_fsync(struct inode *ip, int syncflag, cred_t *cr);
 extern int zfs_getattr(struct inode *ip, vattr_t *vap, int flag, cred_t *cr);
 extern int zfs_getattr_fast(struct inode *ip, struct kstat *sp);

include/sys/zpl.h

@@ -89,4 +89,63 @@ extern const struct inode_operations zpl_ops_snapdirs;
 extern const struct file_operations zpl_fops_shares;
 extern const struct inode_operations zpl_ops_shares;
 
+#ifdef HAVE_VFS_ITERATE
+
+#define DIR_CONTEXT_INIT(_dirent, _actor, _pos) {	\
+	.actor = _actor,				\
+	.pos = _pos,					\
+}
+
+#else
+
+typedef struct dir_context {
+	void *dirent;
+	const filldir_t actor;
+	loff_t pos;
+} dir_context_t;
+
+#define DIR_CONTEXT_INIT(_dirent, _actor, _pos) {	\
+	.dirent = _dirent,				\
+	.actor = _actor,				\
+	.pos = _pos,					\
+}
+
+static inline bool
+dir_emit(struct dir_context *ctx, const char *name, int namelen,
+    uint64_t ino, unsigned type)
+{
+	return ctx->actor(ctx->dirent, name, namelen, ctx->pos, ino, type) == 0;
+}
+
+static inline bool
+dir_emit_dot(struct file *file, struct dir_context *ctx)
+{
+	return ctx->actor(ctx->dirent, ".", 1, ctx->pos,
+	    file->f_path.dentry->d_inode->i_ino, DT_DIR) == 0;
+}
+
+static inline bool
+dir_emit_dotdot(struct file *file, struct dir_context *ctx)
+{
+	return ctx->actor(ctx->dirent, "..", 2, ctx->pos,
+	    parent_ino(file->f_path.dentry), DT_DIR) == 0;
+}
+
+static inline bool
+dir_emit_dots(struct file *file, struct dir_context *ctx)
+{
+	if (ctx->pos == 0) {
+		if (!dir_emit_dot(file, ctx))
+			return false;
+		ctx->pos = 1;
+	}
+	if (ctx->pos == 1) {
+		if (!dir_emit_dotdot(file, ctx))
+			return false;
+		ctx->pos = 2;
+	}
+	return true;
+}
+#endif /* HAVE_VFS_ITERATE */
+
 #endif	/* _SYS_ZPL_H */

module/zfs/zfs_vnops.c

@@ -1997,8 +1997,7 @@ EXPORT_SYMBOL(zfs_rmdir);
  */
 /* ARGSUSED */
 int
-zfs_readdir(struct inode *ip, void *dirent, filldir_t filldir,
+zfs_readdir(struct inode *ip, struct dir_context *ctx, cred_t *cr)
-    loff_t *pos, cred_t *cr)
 {
 	znode_t		*zp = ITOZ(ip);
 	zfs_sb_t	*zsb = ITOZSB(ip);
@@ -2010,6 +2009,7 @@ zfs_readdir(struct inode *ip, void *dirent, filldir_t filldir,
 	uint8_t		prefetch;
 	int		done = 0;
 	uint64_t	parent;
+	loff_t		*pos = &(ctx->pos);
 
 	ZFS_ENTER(zsb);
 	ZFS_VERIFY_ZP(zp);
@@ -2098,11 +2098,11 @@ zfs_readdir(struct inode *ip, void *dirent, filldir_t filldir,
 
 			objnum = ZFS_DIRENT_OBJ(zap.za_first_integer);
 		}
-		done = filldir(dirent, zap.za_name, strlen(zap.za_name),
+
-			       *pos, objnum, ZFS_DIRENT_TYPE(zap.za_first_integer));
+		done = !dir_emit(ctx, zap.za_name, strlen(zap.za_name),
-		if (done) {
+		    objnum, ZFS_DIRENT_TYPE(zap.za_first_integer));
+		if (done)
 			break;
-		}
 
 		/* Prefetch znode */
 		if (prefetch) {

module/zfs/zpl_ctldir.c

@@ -46,79 +46,34 @@ zpl_common_open(struct inode *ip, struct file *filp)
 	return generic_file_open(ip, filp);
 }
 
-static int
-zpl_common_readdir(struct file *filp, void *dirent, filldir_t filldir)
-{
-	struct dentry *dentry = filp->f_path.dentry;
-	struct inode *ip = dentry->d_inode;
-	int error = 0;
-
-	switch (filp->f_pos) {
-	case 0:
-		error = filldir(dirent, ".", 1, 0, ip->i_ino, DT_DIR);
-		if (error)
-			break;
-
-		filp->f_pos++;
-		/* fall-thru */
-	case 1:
-		error = filldir(dirent, "..", 2, 1, parent_ino(dentry), DT_DIR);
-		if (error)
-			break;
-
-		filp->f_pos++;
-		/* fall-thru */
-	default:
-		break;
-	}
-
-	return (error);
-}
-
 /*
  * Get root directory contents.
  */
 static int
-zpl_root_readdir(struct file *filp, void *dirent, filldir_t filldir)
+zpl_root_iterate(struct file *filp, struct dir_context *ctx)
 {
-	struct dentry *dentry = filp->f_path.dentry;
+	zfs_sb_t *zsb = ITOZSB(filp->f_path.dentry->d_inode);
-	struct inode *ip = dentry->d_inode;
-	zfs_sb_t *zsb = ITOZSB(ip);
 	int error = 0;
 
 	ZFS_ENTER(zsb);
 
-	switch (filp->f_pos) {
+	if (!dir_emit_dots(filp, ctx))
-	case 0:
-		error = filldir(dirent, ".", 1, 0, ip->i_ino, DT_DIR);
-		if (error)
 		goto out;
 
-		filp->f_pos++;
+	if (ctx->pos == 2) {
-		/* fall-thru */
+		if (!dir_emit(ctx, ZFS_SNAPDIR_NAME, strlen(ZFS_SNAPDIR_NAME),
-	case 1:
+		    ZFSCTL_INO_SNAPDIR, DT_DIR))
-		error = filldir(dirent, "..", 2, 1, parent_ino(dentry), DT_DIR);
-		if (error)
 			goto out;
 
-		filp->f_pos++;
+		ctx->pos++;
-		/* fall-thru */
+	}
-	case 2:
+
-		error = filldir(dirent, ZFS_SNAPDIR_NAME,
+	if (ctx->pos == 3) {
-		    strlen(ZFS_SNAPDIR_NAME), 2, ZFSCTL_INO_SNAPDIR, DT_DIR);
+		if (!dir_emit(ctx, ZFS_SHAREDIR_NAME, strlen(ZFS_SHAREDIR_NAME),
-		if (error)
+		    ZFSCTL_INO_SHARES, DT_DIR))
 			goto out;
 
-		filp->f_pos++;
+		ctx->pos++;
-		/* fall-thru */
-	case 3:
-		error = filldir(dirent, ZFS_SHAREDIR_NAME,
-		    strlen(ZFS_SHAREDIR_NAME), 3, ZFSCTL_INO_SHARES, DT_DIR);
-		if (error)
-			goto out;
-
-		filp->f_pos++;
-		/* fall-thru */
 	}
 out:
 	ZFS_EXIT(zsb);
@@ -126,6 +81,20 @@ out:
 	return (error);
 }
 
+#if !defined(HAVE_VFS_ITERATE)
+static int
+zpl_root_readdir(struct file *filp, void *dirent, filldir_t filldir)
+{
+	struct dir_context ctx = DIR_CONTEXT_INIT(dirent, filldir, filp->f_pos);
+	int error;
+
+	error = zpl_root_iterate(filp, &ctx);
+	filp->f_pos = ctx.pos;
+
+	return (error);
+}
+#endif /* HAVE_VFS_ITERATE */
+
 /*
  * Get root directory attributes.
  */
@@ -175,7 +144,11 @@ const struct file_operations zpl_fops_root = {
 	.open		= zpl_common_open,
 	.llseek		= generic_file_llseek,
 	.read		= generic_read_dir,
+#ifdef HAVE_VFS_ITERATE
+	.iterate	= zpl_root_iterate,
+#else
 	.readdir	= zpl_root_readdir,
+#endif
 };
 
 const struct inode_operations zpl_ops_root = {
@@ -273,50 +246,29 @@ zpl_snapdir_lookup(struct inode *dip, struct dentry *dentry,
 	return d_splice_alias(ip, dentry);
 }
 
-/* ARGSUSED */
 static int
-zpl_snapdir_readdir(struct file *filp, void *dirent, filldir_t filldir)
+zpl_snapdir_iterate(struct file *filp, struct dir_context *ctx)
 {
-	struct dentry *dentry = filp->f_path.dentry;
+	zfs_sb_t *zsb = ITOZSB(filp->f_path.dentry->d_inode);
-	struct inode *dip = dentry->d_inode;
-	zfs_sb_t *zsb = ITOZSB(dip);
 	char snapname[MAXNAMELEN];
-	uint64_t id, cookie;
 	boolean_t case_conflict;
+	uint64_t id;
 	int error = 0;
 
 	ZFS_ENTER(zsb);
 
-	cookie = filp->f_pos;
+	if (!dir_emit_dots(filp, ctx))
-	switch (filp->f_pos) {
-	case 0:
-		error = filldir(dirent, ".", 1, 0, dip->i_ino, DT_DIR);
-		if (error)
 		goto out;
 
-		filp->f_pos++;
-		/* fall-thru */
-	case 1:
-		error = filldir(dirent, "..", 2, 1, parent_ino(dentry), DT_DIR);
-		if (error)
-			goto out;
-
-		filp->f_pos++;
-		/* fall-thru */
-	default:
 	while (error == 0) {
 		error = -dmu_snapshot_list_next(zsb->z_os, MAXNAMELEN,
-			    snapname, &id, &cookie, &case_conflict);
+		    snapname, &id, &(ctx->pos), &case_conflict);
 		if (error)
 			goto out;
 
-			error = filldir(dirent, snapname, strlen(snapname),
+		if (!dir_emit(ctx, snapname, strlen(snapname),
-			    filp->f_pos, ZFSCTL_INO_SHARES - id, DT_DIR);
+		    ZFSCTL_INO_SHARES - id, DT_DIR))
-			if (error)
 			goto out;
-
-			filp->f_pos = cookie;
-		}
 	}
 out:
 	ZFS_EXIT(zsb);
@@ -327,6 +279,20 @@ out:
 	return (error);
 }
 
+#if !defined(HAVE_VFS_ITERATE)
+static int
+zpl_snapdir_readdir(struct file *filp, void *dirent, filldir_t filldir)
+{
+	struct dir_context ctx = DIR_CONTEXT_INIT(dirent, filldir, filp->f_pos);
+	int error;
+
+	error = zpl_snapdir_iterate(filp, &ctx);
+	filp->f_pos = ctx.pos;
+
+	return (error);
+}
+#endif /* HAVE_VFS_ITERATE */
+
 int
 zpl_snapdir_rename(struct inode *sdip, struct dentry *sdentry,
     struct inode *tdip, struct dentry *tdentry)
@@ -413,7 +379,12 @@ const struct file_operations zpl_fops_snapdir = {
 	.open		= zpl_common_open,
 	.llseek		= generic_file_llseek,
 	.read		= generic_read_dir,
+#ifdef HAVE_VFS_ITERATE
+	.iterate	= zpl_snapdir_iterate,
+#else
 	.readdir	= zpl_snapdir_readdir,
+#endif
+
 };
 
 /*
@@ -458,42 +429,51 @@ zpl_shares_lookup(struct inode *dip, struct dentry *dentry,
 	return d_splice_alias(ip, dentry);
 }
 
-/* ARGSUSED */
 static int
-zpl_shares_readdir(struct file *filp, void *dirent, filldir_t filldir)
+zpl_shares_iterate(struct file *filp, struct dir_context *ctx)
 {
 	cred_t *cr = CRED();
-	struct dentry *dentry = filp->f_path.dentry;
+	zfs_sb_t *zsb = ITOZSB(filp->f_path.dentry->d_inode);
-	struct inode *ip = dentry->d_inode;
-	zfs_sb_t *zsb = ITOZSB(ip);
 	znode_t *dzp;
-	int error;
+	int error = 0;
 
 	ZFS_ENTER(zsb);
 
 	if (zsb->z_shares_dir == 0) {
-		error = zpl_common_readdir(filp, dirent, filldir);
+		dir_emit_dots(filp, ctx);
-		ZFS_EXIT(zsb);
+		goto out;
-		return (error);
 	}
 
 	error = -zfs_zget(zsb, zsb->z_shares_dir, &dzp);
-	if (error) {
+	if (error)
-		ZFS_EXIT(zsb);
+		goto out;
-		return (error);
-	}
 
 	crhold(cr);
-	error = -zfs_readdir(ZTOI(dzp), dirent, filldir, &filp->f_pos, cr);
+	error = -zfs_readdir(ZTOI(dzp), ctx, cr);
 	crfree(cr);
 
 	iput(ZTOI(dzp));
+out:
 	ZFS_EXIT(zsb);
 	ASSERT3S(error, <=, 0);
 
 	return (error);
 }
 
+#if !defined(HAVE_VFS_ITERATE)
+static int
+zpl_shares_readdir(struct file *filp, void *dirent, filldir_t filldir)
+{
+	struct dir_context ctx = DIR_CONTEXT_INIT(dirent, filldir, filp->f_pos);
+	int error;
+
+	error = zpl_shares_iterate(filp, &ctx);
+	filp->f_pos = ctx.pos;
+
+	return (error);
+}
+#endif /* HAVE_VFS_ITERATE */
+
 /* ARGSUSED */
 static int
 zpl_shares_getattr(struct vfsmount *mnt, struct dentry *dentry,
@@ -532,7 +512,12 @@ const struct file_operations zpl_fops_shares = {
 	.open		= zpl_common_open,
 	.llseek		= generic_file_llseek,
 	.read		= generic_read_dir,
+#ifdef HAVE_VFS_ITERATE
+	.iterate	= zpl_shares_iterate,
+#else
 	.readdir	= zpl_shares_readdir,
+#endif
+
 };
 
 /*

module/zfs/zpl_file.c

@@ -64,21 +64,34 @@ zpl_release(struct inode *ip, struct file *filp)
 }
 
 static int
-zpl_readdir(struct file *filp, void *dirent, filldir_t filldir)
+zpl_iterate(struct file *filp, struct dir_context *ctx)
 {
 	struct dentry *dentry = filp->f_path.dentry;
 	cred_t *cr = CRED();
 	int error;
 
 	crhold(cr);
-	error = -zfs_readdir(dentry->d_inode, dirent, filldir,
+	error = -zfs_readdir(dentry->d_inode, ctx, cr);
-	    &filp->f_pos, cr);
 	crfree(cr);
 	ASSERT3S(error, <=, 0);
 
 	return (error);
 }
 
+#if !defined(HAVE_VFS_ITERATE)
+static int
+zpl_readdir(struct file *filp, void *dirent, filldir_t filldir)
+{
+	struct dir_context ctx = DIR_CONTEXT_INIT(dirent, filldir, filp->f_pos);
+	int error;
+
+	error = zpl_iterate(filp, &ctx);
+	filp->f_pos = ctx.pos;
+
+	return (error);
+}
+#endif /* HAVE_VFS_ITERATE */
+
 #if defined(HAVE_FSYNC_WITH_DENTRY)
 /*
  * Linux 2.6.x - 2.6.34 API,
@@ -506,7 +519,11 @@ const struct file_operations zpl_file_operations = {
 const struct file_operations zpl_dir_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= generic_read_dir,
+#ifdef HAVE_VFS_ITERATE
+	.iterate	= zpl_iterate,
+#else
 	.readdir	= zpl_readdir,
+#endif
 	.fsync		= zpl_fsync,
 	.unlocked_ioctl = zpl_ioctl,
 #ifdef CONFIG_COMPAT

module/zfs/zpl_xattr.c

@@ -80,6 +80,7 @@
 #include <sys/zfs_vfsops.h>
 #include <sys/zfs_vnops.h>
 #include <sys/zfs_znode.h>
+#include <sys/zap.h>
 #include <sys/vfs.h>
 #include <sys/zpl.h>
 
@@ -91,11 +92,8 @@ typedef struct xattr_filldir {
 } xattr_filldir_t;
 
 static int
-zpl_xattr_filldir(void *arg, const char *name, int name_len,
+zpl_xattr_filldir(xattr_filldir_t *xf, const char *name, int name_len)
-    loff_t offset, uint64_t objnum, unsigned int d_type)
 {
-	xattr_filldir_t *xf = arg;
-
 	if (!strncmp(name, XATTR_USER_PREFIX, XATTR_USER_PREFIX_LEN))
 		if (!(ITOZSB(xf->inode)->z_flags & ZSB_XATTR))
 			return (0);
@@ -118,12 +116,46 @@ zpl_xattr_filldir(void *arg, const char *name, int name_len,
 	return (0);
 }
 
+/*
+ * Read as many directory entry names as will fit in to the provided buffer,
+ * or when no buffer is provided calculate the required buffer size.
+ */
+int
+zpl_xattr_readdir(struct inode *dxip, xattr_filldir_t *xf)
+{
+	zap_cursor_t zc;
+	zap_attribute_t	zap;
+	int error;
+
+	zap_cursor_init(&zc, ITOZSB(dxip)->z_os, ITOZ(dxip)->z_id);
+
+	while ((error = -zap_cursor_retrieve(&zc, &zap)) == 0) {
+
+		if (zap.za_integer_length != 8 || zap.za_num_integers != 1) {
+			error = -ENXIO;
+			break;
+		}
+
+		error = zpl_xattr_filldir(xf, zap.za_name, strlen(zap.za_name));
+		if (error)
+			break;
+
+		zap_cursor_advance(&zc);
+	}
+
+	zap_cursor_fini(&zc);
+
+	if (error == -ENOENT)
+		error = 0;
+
+	return (error);
+}
+
 static ssize_t
 zpl_xattr_list_dir(xattr_filldir_t *xf, cred_t *cr)
 {
 	struct inode *ip = xf->inode;
 	struct inode *dxip = NULL;
-	loff_t pos = 3;  /* skip '.', '..', and '.zfs' entries. */
 	int error;
 
 	/* Lookup the xattr directory */
@@ -135,8 +167,7 @@ zpl_xattr_list_dir(xattr_filldir_t *xf, cred_t *cr)
 		return (error);
 	}
 
-	/* Fill provided buffer via zpl_zattr_filldir helper */
+	error = zpl_xattr_readdir(dxip, xf);
-	error = -zfs_readdir(dxip, (void *)xf, zpl_xattr_filldir, &pos, cr);
 	iput(dxip);
 
 	return (error);
@@ -162,8 +193,8 @@ zpl_xattr_list_sa(xattr_filldir_t *xf)
 	while ((nvp = nvlist_next_nvpair(zp->z_xattr_cached, nvp)) != NULL) {
 		ASSERT3U(nvpair_type(nvp), ==, DATA_TYPE_BYTE_ARRAY);
 
-		error = zpl_xattr_filldir((void *)xf, nvpair_name(nvp),
+		error = zpl_xattr_filldir(xf, nvpair_name(nvp),
-		     strlen(nvpair_name(nvp)), 0, 0, 0);
+		     strlen(nvpair_name(nvp)));
 		if (error)
 			return (error);
 	}