new helper: generic_file_read_iter()

iov_iter-using variant of generic_file_aio_read().  Some callers
converted.  Note that it's still not quite there for use as ->read_iter() -
we depend on having zero iter->iov_offset in O_DIRECT case.  Fortunately,
that's true for all converted callers (and for generic_file_aio_read() itself).
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>

fs/ceph/file.c

@@ -833,24 +833,11 @@ static ssize_t ceph_aio_read(struct kiocb *iocb, const struct iovec *iov,
 		/* hmm, this isn't really async... */
 		ret = ceph_sync_read(iocb, &i, &checkeof);
 	} else {
-		/*
-		 * We can't modify the content of iov,
-		 * so we only read from beginning.
-		 *
-		 * When we switch generic_file_aio_read() to iov_iter, the
-		 * if () below will be removed -- AV
-		 */
-		if (read) {
-			iocb->ki_pos = pos;
-			len = iocb->ki_nbytes;
-			read = 0;
-			iov_iter_init(&i, iov, nr_segs, len, 0);
-		}
 		dout("aio_read %p %llx.%llx %llu~%u got cap refs on %s\n",
 		     inode, ceph_vinop(inode), pos, (unsigned)len,
 		     ceph_cap_string(got));
 
-		ret = generic_file_aio_read(iocb, iov, nr_segs, pos);
+		ret = generic_file_read_iter(iocb, &i);
 	}
 	dout("aio_read %p %llx.%llx dropping cap refs on %s = %d\n",
 	     inode, ceph_vinop(inode), ceph_cap_string(got), (int)ret);

fs/nfs/file.c

@@ -184,7 +184,7 @@ nfs_file_read(struct kiocb *iocb, const struct iovec *iov,
 
 	result = nfs_revalidate_mapping(inode, iocb->ki_filp->f_mapping);
 	if (!result) {
-		result = generic_file_aio_read(iocb, iov, nr_segs, pos);
+		result = generic_file_read_iter(iocb, &to);
 		if (result > 0)
 			nfs_add_stats(inode, NFSIOS_NORMALREADBYTES, result);
 	}

include/linux/fs.h

@@ -2404,6 +2404,7 @@ extern int generic_file_remap_pages(struct vm_area_struct *, unsigned long addr,
 		unsigned long size, pgoff_t pgoff);
 int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk);
 extern ssize_t generic_file_aio_read(struct kiocb *, const struct iovec *, unsigned long, loff_t);
+extern ssize_t generic_file_read_iter(struct kiocb *, struct iov_iter *);
 extern ssize_t __generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long);
 extern ssize_t generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long, loff_t);
 extern ssize_t generic_file_direct_write(struct kiocb *, struct iov_iter *,

mm/filemap.c

@@ -1663,55 +1663,34 @@ static ssize_t do_generic_file_read(struct file *filp, loff_t *ppos,
 	return written ? written : error;
 }
 
-/**
- * generic_file_aio_read - generic filesystem read routine
- * @iocb:	kernel I/O control block
- * @iov:	io vector request
- * @nr_segs:	number of segments in the iovec
- * @pos:	current file position
- *
- * This is the "read()" routine for all filesystems
- * that can use the page cache directly.
- */
 ssize_t
-generic_file_aio_read(struct kiocb *iocb, const struct iovec *iov,
-		unsigned long nr_segs, loff_t pos)
+generic_file_read_iter(struct kiocb *iocb, struct iov_iter *iter)
 {
-	struct file *filp = iocb->ki_filp;
+	struct file *file = iocb->ki_filp;
 	ssize_t retval = 0;
-	size_t count;
 	loff_t *ppos = &iocb->ki_pos;
-	struct iov_iter i;
-
-	count = iov_length(iov, nr_segs);
-	iov_iter_init(&i, iov, nr_segs, count, 0);
+	loff_t pos = *ppos;
 
 	/* coalesce the iovecs and go direct-to-BIO for O_DIRECT */
-	if (filp->f_flags & O_DIRECT) {
+	if (file->f_flags & O_DIRECT) {
+		struct address_space *mapping = file->f_mapping;
+		struct inode *inode = mapping->host;
+		size_t count = iov_iter_count(iter);
 		loff_t size;
-		struct address_space *mapping;
-		struct inode *inode;
 
-		mapping = filp->f_mapping;
-		inode = mapping->host;
 		if (!count)
 			goto out; /* skip atime */
 		size = i_size_read(inode);
 		retval = filemap_write_and_wait_range(mapping, pos,
 					pos + count - 1);
 		if (!retval) {
-			struct iov_iter data = i;
+			struct iov_iter data = *iter;
 			retval = mapping->a_ops->direct_IO(READ, iocb, &data, pos);
 		}
 
 		if (retval > 0) {
 			*ppos = pos + retval;
-			count -= retval;
-			/*
-			 * If we did a short DIO read we need to skip the
-			 * section of the iov that we've already read data into.
-			 */
-			iov_iter_advance(&i, retval);
+			iov_iter_advance(iter, retval);
 		}
 
 		/*
@@ -1722,16 +1701,38 @@ generic_file_aio_read(struct kiocb *iocb, const struct iovec *iov,
 		 * and return.  Otherwise fallthrough to buffered io for
 		 * the rest of the read.
 		 */
-		if (retval < 0 || !count || *ppos >= size) {
-			file_accessed(filp);
+		if (retval < 0 || !iov_iter_count(iter) || *ppos >= size) {
+			file_accessed(file);
 			goto out;
 		}
 	}
 
-	retval = do_generic_file_read(filp, ppos, &i, retval);
+	retval = do_generic_file_read(file, ppos, iter, retval);
 out:
 	return retval;
 }
+EXPORT_SYMBOL(generic_file_read_iter);
+
+/**
+ * generic_file_aio_read - generic filesystem read routine
+ * @iocb:	kernel I/O control block
+ * @iov:	io vector request
+ * @nr_segs:	number of segments in the iovec
+ * @pos:	current file position
+ *
+ * This is the "read()" routine for all filesystems
+ * that can use the page cache directly.
+ */
+ssize_t
+generic_file_aio_read(struct kiocb *iocb, const struct iovec *iov,
+		unsigned long nr_segs, loff_t pos)
+{
+	size_t count = iov_length(iov, nr_segs);
+	struct iov_iter i;
+
+	iov_iter_init(&i, iov, nr_segs, count, 0);
+	return generic_file_read_iter(iocb, &i);
+}
 EXPORT_SYMBOL(generic_file_aio_read);
 
 #ifdef CONFIG_MMU