vfs: fix copy_file_range() averts filesystem freeze protection

Commit 868f9f2f ("vfs: fix copy_file_range() regression in cross-fs
copies") removed fallback to generic_copy_file_range() for cross-fs
cases inside vfs_copy_file_range().

To preserve behavior of nfsd and ksmbd server-side-copy, the fallback to
generic_copy_file_range() was added in nfsd and ksmbd code, but that
call is missing sb_start_write(), fsnotify hooks and more.

Ideally, nfsd and ksmbd would pass a flag to vfs_copy_file_range() that
will take care of the fallback, but that code would be subtle and we got
vfs_copy_file_range() logic wrong too many times already.

Instead, add a flag to explicitly request vfs_copy_file_range() to
perform only generic_copy_file_range() and let nfsd and ksmbd use this
flag only in the fallback path.

This choise keeps the logic changes to minimum in the non-nfsd/ksmbd code
paths to reduce the risk of further regressions.

Fixes: 868f9f2f ("vfs: fix copy_file_range() regression in cross-fs copies")
Tested-by: Namjae Jeon <linkinjeon@kernel.org>
Tested-by: Luis Henriques <lhenriques@suse.de>
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>

fs/ksmbd/vfs.c

@@ -1794,9 +1794,9 @@ int ksmbd_vfs_copy_file_ranges(struct ksmbd_work *work,
 		ret = vfs_copy_file_range(src_fp->filp, src_off,
 					  dst_fp->filp, dst_off, len, 0);
 		if (ret == -EOPNOTSUPP || ret == -EXDEV)
-			ret = generic_copy_file_range(src_fp->filp, src_off,
-						      dst_fp->filp, dst_off,
-						      len, 0);
+			ret = vfs_copy_file_range(src_fp->filp, src_off,
+						  dst_fp->filp, dst_off, len,
+						  COPY_FILE_SPLICE);
 		if (ret < 0)
 			return ret;
 

fs/nfsd/vfs.c

@@ -596,8 +596,8 @@ ssize_t nfsd_copy_file_range(struct file *src, u64 src_pos, struct file *dst,
 	ret = vfs_copy_file_range(src, src_pos, dst, dst_pos, count, 0);
 
 	if (ret == -EOPNOTSUPP || ret == -EXDEV)
-		ret = generic_copy_file_range(src, src_pos, dst, dst_pos,
-					      count, 0);
+		ret = vfs_copy_file_range(src, src_pos, dst, dst_pos, count,
+					  COPY_FILE_SPLICE);
 	return ret;
 }
 

fs/read_write.c

@@ -1388,6 +1388,8 @@ ssize_t generic_copy_file_range(struct file *file_in, loff_t pos_in,
 				struct file *file_out, loff_t pos_out,
 				size_t len, unsigned int flags)
 {
+	lockdep_assert(sb_write_started(file_inode(file_out)->i_sb));
+
 	return do_splice_direct(file_in, &pos_in, file_out, &pos_out,
 				len > MAX_RW_COUNT ? MAX_RW_COUNT : len, 0);
 }
@@ -1424,7 +1426,9 @@ static int generic_copy_file_checks(struct file *file_in, loff_t pos_in,
 	 * and several different sets of file_operations, but they all end up
 	 * using the same ->copy_file_range() function pointer.
 	 */
-	if (file_out->f_op->copy_file_range) {
+	if (flags & COPY_FILE_SPLICE) {
+		/* cross sb splice is allowed */
+	} else if (file_out->f_op->copy_file_range) {
 		if (file_in->f_op->copy_file_range !=
 		    file_out->f_op->copy_file_range)
 			return -EXDEV;
@@ -1474,8 +1478,9 @@ ssize_t vfs_copy_file_range(struct file *file_in, loff_t pos_in,
 			    size_t len, unsigned int flags)
 {
 	ssize_t ret;
+	bool splice = flags & COPY_FILE_SPLICE;
 
-	if (flags != 0)
+	if (flags & ~COPY_FILE_SPLICE)
 		return -EINVAL;
 
 	ret = generic_copy_file_checks(file_in, pos_in, file_out, pos_out, &len,
@@ -1501,14 +1506,14 @@ ssize_t vfs_copy_file_range(struct file *file_in, loff_t pos_in,
 	 * same sb using clone, but for filesystems where both clone and copy
 	 * are supported (e.g. nfs,cifs), we only call the copy method.
 	 */
-	if (file_out->f_op->copy_file_range) {
+	if (!splice && file_out->f_op->copy_file_range) {
 		ret = file_out->f_op->copy_file_range(file_in, pos_in,
 						      file_out, pos_out,
 						      len, flags);
 		goto done;
 	}
 
-	if (file_in->f_op->remap_file_range &&
+	if (!splice && file_in->f_op->remap_file_range &&
 	    file_inode(file_in)->i_sb == file_inode(file_out)->i_sb) {
 		ret = file_in->f_op->remap_file_range(file_in, pos_in,
 				file_out, pos_out,
@@ -1528,6 +1533,8 @@ ssize_t vfs_copy_file_range(struct file *file_in, loff_t pos_in,
 	 * consistent story about which filesystems support copy_file_range()
 	 * and which filesystems do not, that will allow userspace tools to
 	 * make consistent desicions w.r.t using copy_file_range().
+	 *
+	 * We also get here if caller (e.g. nfsd) requested COPY_FILE_SPLICE.
 	 */
 	ret = generic_copy_file_range(file_in, pos_in, file_out, pos_out, len,
 				      flags);
@@ -1582,6 +1589,10 @@ SYSCALL_DEFINE6(copy_file_range, int, fd_in, loff_t __user *, off_in,
 		pos_out = f_out.file->f_pos;
 	}
 
+	ret = -EINVAL;
+	if (flags != 0)
+		goto out;
+
 	ret = vfs_copy_file_range(f_in.file, pos_in, f_out.file, pos_out, len,
 				  flags);
 	if (ret > 0) {

include/linux/fs.h

@@ -2089,6 +2089,14 @@ struct dir_context {
  */
 #define REMAP_FILE_ADVISORY		(REMAP_FILE_CAN_SHORTEN)
 
+/*
+ * These flags control the behavior of vfs_copy_file_range().
+ * They are not available to the user via syscall.
+ *
+ * COPY_FILE_SPLICE: call splice direct instead of fs clone/copy ops
+ */
+#define COPY_FILE_SPLICE		(1 << 0)
+
 struct iov_iter;
 struct io_uring_cmd;