at91udc: startup fix

PATCH In review Cerrato Renaud 1 participant 1 message Mar 13, 2013

Subsystems (2)

USB Arm/Microchip (AT91) SoC support

lore

Share & tools

lore

Cerrato Renaud

author

2013-03-13 13:40 UTC #0 lore raw

PATCH 2/5 at91udc: startup fix

I noticed that depending on boot timings, the USB gadget weren't detected at all by hosts. After few hours of debugging and using the at91clk output, I found that the UDP clock wasn't enabled from time to time.

This patch fix the startup of UDC by adding a check into the at91_update_vbus().

Signed-off-by Cerrato Renaud <[email protected]>

---
 drivers/usb/gadget/at91_udc.c |   10 ++++++++--
 1 files changed, 8 insertions(+), 2 deletions(-)

drivers/usb/gadget/at91_udc.c +8 -2

index 3899db2..0654038 100644
@@ -1327,6 +1327,12 @@ static void at91_update_vbus(struct at91_udc *udc, u32 value)
         dev_set_param(udc->dev, "vbus", "0");
 
     udc->gpio_vbus_val = value;
+    udc->vbus = value != 0;
+
+    if(udc->driver)
+        pullup(udc, udc->vbus);
+    else
+        pullup(udc, 0);
 }
 
 int usb_gadget_poll(void)

@@ -1508,7 +1514,7 @@ static int __init at91udc_probe(struct device_d *dev)
          * Get the initial state of VBUS - we cannot expect
          * a pending interrupt.
          */
-        udc->vbus = gpio_get_value(udc->board.vbus_pin);
+        udc->vbus = gpio_get_value(udc->board.vbus_pin) ^ udc->board.vbus_active_low;
         DBG(udc, "VBUS detection: host:%s \n",
             udc->vbus ? "present":"absent");
     } else {

@@ -1517,7 +1523,7 @@ static int __init at91udc_probe(struct device_d *dev)
     }
 
     dev_add_param(dev, "vbus", NULL, NULL, 0);
-    dev_set_param(dev, "vbus", "0");
+    dev_set_param(dev, "vbus", udc->vbus ? "1" : "0");
 
     poller_register(&poller);

Signature · 7 lines

1.7.2.5


_______________________________________________
barebox mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/barebox