iomap: Fix error handling in iomap_zero_iter()

iomap_write_end() does not return a negative errno to indicate an
error, but the number of bytes successfully copied.  It cannot return
an error today, so include a debugging assertion like the one in
iomap_unshare_iter().

Fixes: c6f4046865 ("fsdax: decouple zeroing from the iomap buffered I/O code")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20211221044450.517558-1-willy@infradead.org
Signed-off-by: Dan Williams <dan.j.williams@intel.com>

fs/iomap/buffered-io.c

@@ -895,8 +895,8 @@ static loff_t iomap_zero_iter(struct iomap_iter *iter, bool *did_zero)
 		mark_page_accessed(page);
 
 		bytes = iomap_write_end(iter, pos, bytes, bytes, page);
-		if (bytes < 0)
-			return bytes;
+		if (WARN_ON_ONCE(bytes == 0))
+			return -EIO;
 
 		pos += bytes;
 		length -= bytes;