On 08/17/2011 09:49 AM, Yang Ruirui R wrote:
On 08/16/2011 06:38 PM, Sebastian Andrzej Siewior wrote:
* Yang Ruirui | 2011-08-16 18:01:54 [+0800]:
diff -uprN a/drivers/usb/gadget/composite.c b/drivers/usb/gadget/composite.c
--- a/drivers/usb/gadget/composite.c 2011-08-16 17:49:38.813038094 +0800
+++ b/drivers/usb/gadget/composite.c 2011-08-16 17:47:03.716372207 +0800
@@ -603,6 +603,7 @@ static int set_config(struct usb_composi
int result = -EINVAL;
unsigned power = gadget_is_otg(gadget) ? 8 : 100;
int tmp;
+ enum usb_device_speed s = USB_SPEED_LOW;
if (number) {
list_for_each_entry(c,&cdev->configs, list) {
@@ -626,6 +627,14 @@ static int set_config(struct usb_composi
result = 0;
}
+ if (c->fullspeed&& (gadget->speed == USB_SPEED_FULL))
+ s = USB_SPEED_FULL;
+ if (c->highspeed&& (gadget->speed == USB_SPEED_HIGH))
+ s = USB_SPEED_HIGH;
+ if (c->superspeed&& (gadget->speed == USB_SPEED_SUPER))
+ s = USB_SPEED_SUPER;
+ gadget->speed = s;
+
I'm not sure if this is the right way. After all the speed is still HS
we just use FS descriptors. And I'm not sure if changing the speed is a
wise thing to do.
Is this a regression from recent changes in the framework?
Indeed, 3.0-rc7 works. I tried to bisect this issue, but kernel stall after
"Uncompressing Linux... done, booting the kernel" in the middle commits.
Do you know some suspicious commits I can revert to debug?
What about config_ep_by_speed(), config_buf(), set_config() not assigned
a null pointer in the first place?
It is in composite.c:set_config
Maybe I misunderstood your points, I did not see the null assign yet.
I will continue to debug this issue.
Sebastian
--
Thanks
Yang Ruirui
--
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
