I tried waiting for !port->write_urb_busy, but it didn't help. Looks like
a small delay is the only way to do it.
Signed-off-by: Pete Zaitcev <zaitcev@xxxxxxxxxx>
---
drivers/usb/serial/mct_u232.c | 10 ++++++++++
1 file changed, 10 insertions(+)
commit 66b10b12c5bee0d65834d7bb7a0620a7e0cd4e12
Author: Master <zaitcev@xxxxxxxxxxxxxxxxxxx>
Date: Sat Dec 5 15:01:05 2009 -0700
Add the delay on close while output drains.
diff --git a/drivers/usb/serial/mct_u232.c b/drivers/usb/serial/mct_u232.c
index f978e08..3486d75 100644
--- a/drivers/usb/serial/mct_u232.c
+++ b/drivers/usb/serial/mct_u232.c
@@ -552,6 +552,16 @@ static void mct_u232_close(struct usb_serial_port *port)
{
dbg("%s port %d", __func__, port->number);
+ /*
+ * Since the driver is reverse-engineered, we don't know how to
+ * ask the device if bytes were actually transmitted. So, when
+ * writes return and our URBs have completed, the bits are still
+ * being sent. When we close and reopen, one data byte is lost.
+ * Nothing to it but sleep a little bit while transmit drains.
+ * RH bz#170490.
+ */
+ msleep(2);
+
if (port->serial->dev) {
/* shutdown our urbs */
usb_kill_urb(port->write_urb);
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
