Hi everyone,
I just wanted to report on my progress on getting Linux going on my
still unnameable industrial handheld, and thank you all for your
suggestions.
Just a quick recap:
* The device is based on a Marvell pxa270 SoC, which has had excellent
community / commercial support in the kernel.
* Originally, I was very happy to get a console up on the Standard
UART (STUART), which is NOT the usual UART used for early printk
messages (I believe that's the Full UART or FUART). That is, until I
realized that I could only read bytes from the STUART and not write to
it because it was spliced in hardware with an RFID transceiver..
* Then, I was able to get the LCD up and running, and saw our
favourite penguin show up (yaay!)
* Needed console access somehow, other than the UART, so opted for
using it as a USB Ethernet Gadget.
* Although the device posted several messages on boot saying that the
handheld usb0 was up and running, my workstation still did not receive
any notification that a USB network device was present, even after
disconnecting / reconnecting.
Which leads us to today.
I just tried working with linux-2.6.git instead of sticking with the
2.6.29 kernel, and I guess there may have been some more debug / error
message code in the pxa27x-udc driver, which led me to the problem.
The error message I finally saw was this:
=========================================
pxa27x-udc pxa27x-udc: Couldn't request gpio 57 : -16
pxa27x-udc: probe of pxa27x-udc failed with error -16
=========================================
Error -16 corresponds to -EBUSY, which I thought was odd. Peaking into
pxa27x-udc, showed me that the driver uses the gpio_request function
to claim the appropriate GPIO. Actually, gpio_request would only
report this error if that particular gpio descriptor already had the
REQUESTED flag set.
The problem was that I had set the GPIO statically in my pin_config[]
array, like so:
========================================
static unsigned long xxxxx_pin_config[] = {
...
/*
* UDC
*/
MFP_CFG_OUT(GPIO57_nCL_USB_ENABLE,AF0,DRIVE_HIGH),
MFP_CFG_IN(GPIO79_nCL_USB_DETECT,AF0),
...
}
========================================
Removing the declarations from my static pin_config[], while ensuring
that the following code (slightly modified from Robert Jarzmik's
mioa701.c) in my UDC section fixed the problem.
========================================
static int is_usb_connected(void)
{
return !gpio_get_value(GPIO79_nCL_USB_DETECT);
}
static void udc_command( int cmd ) {
switch(cmd) {
case PXA2XX_UDC_CMD_CONNECT:
gpio_set_value(GPIO57_nCL_USB_ENABLE,0);
break;
case PXA2XX_UDC_CMD_DISCONNECT:
gpio_set_value(GPIO57_nCL_USB_ENABLE,1);
break;
default:
break;
}
}
static struct pxa2xx_udc_mach_info xxxxxx_udc_info = {
.udc_is_connected = is_usb_connected,
.udc_command = udc_command,
.gpio_pullup = GPIO57_nCL_USB_ENABLE,
.gpio_pullup_inverted = 1,
.gpio_vbus = GPIO79_nCL_USB_DETECT,
.gpio_vbus_inverted = 1,
};
static struct gpio_vbus_mach_info gpio_vbus_data = {
.gpio_vbus = GPIO79_nCL_USB_DETECT,
.gpio_vbus_inverted = 1,
// .gpio_pullup = GPIO57_nCL_USB_ENABLE,
// .gpio_pullup_inverted = 1,
};
static struct platform_device gpio_vbus = {
.name = "gpio-vbus",
.id = -1,
.dev = {
.platform_data = &gpio_vbus_data,
.parent = NULL,
},
};
static void __init udc_init() {
pxa_set_udc_info(&xxxxxx_udc_info);
}
========================================
Also note, that the gpio_vbus platform_device should be registered as
part of an array of platform_devices, with
platform_add_devices(platform_devices, ARRAY_SIZE(platform_devices));
So to whoever is responsible for inserting those two lines of
debugging / error messages in pxa27x-udc.c: Thanks!!!
Cheers,
Chris
--
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
