Do not use buffered io channel reads for chunked device input

The wakelock avare evdev drivers hold a wakelock only until all
clients have read all events. If buffered reads are used, we might
hit automatic suspend before having chance to obtain user space
wakelock for the duration of actually processing the input.

mce-io.c

@@ -1221,6 +1221,13 @@ gconstpointer mce_register_io_monitor_chunk(const gint fd,
 	/* We only read this file in binary form */
 	(void)g_io_channel_set_encoding(iomon->iochan, NULL, &error);
 
+	/* No buffering since we're using this for reading data from
+	 * device drivers and need to keep the i/o state in sync
+	 * between kernel and user space for the automatic suspend
+	 * prevention via wakelocks to work
+	 */
+	g_io_channel_set_buffered(iomon->iochan, FALSE);
+
 	/* Reset errno,
 	 * to avoid false positives down the line
 	 */