Hi Steven,
> > I deleted your last email by accident. Can you re-send it please.
>
> As you may lose the mail. I'd like to introduce our condition again.
>
> As to the Atheros 3011 chip. yes, it uses DFU mode to download the firmware.
> Instead of downloading the firmware, the new Atheros 3012 chip needs to download patch and radio table.
> The patch and radio table are used to fix those bugs found after chip shipping or for some other specific use.
> And the Atheros 3012 chip itself is a native bluetooth device. Even without downloading patch and radio table, it can work.
>
> So with this design, the Atheros 3012 chip always keeps same PID/VID even after downloading the patch and radio
> table. Due to this limitation, we have problems with current btusb.c implementation.
>
> We have to have a work around way, this is to use the bcdDevice (Device Release Number) to judge
> whether the Atheros 3012 chip has been patched or not.
> The init bcdDevice value of this chip is 0x0001, we can increase this value after patch and radio table downloading.
>
> Do you think it is possible ?
now I understand your problem. So in theory you have a full Bluetooth
device working according to the HCI standard. And you are using vendor
HCI commands to patch the firmware.
So this is similar to what we are doing with a lot of UART based devices
when using hciattach.
For now, I am fine using bcdDevice number. In the end we have to create
a pre-stage for loading further instructions or patches into the device.
We had plans for this pre-stage already, but never implemented them.
> From 2dc0e703ed392f41d91984b0cfd1da56f6f173cb Mon Sep 17 00:00:00 2001
> From: Steven.Li <Steven.Li@xxxxxxxxxxx>
> Date: Thu, 26 May 2011 17:06:52 +0800
> Subject: [PATCH] Use bcdDevice(Device Release Number) to exclude the ath3k and btusb driver.
>
> ---
> drivers/bluetooth/ath3k.c | 5 ++++-
> drivers/bluetooth/btusb.c | 9 ++++++++-
> 2 files changed, 12 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/bluetooth/ath3k.c b/drivers/bluetooth/ath3k.c
> index 6bacef3..c99fb08 100644
> --- a/drivers/bluetooth/ath3k.c
> +++ b/drivers/bluetooth/ath3k.c
> @@ -375,6 +375,9 @@ static int ath3k_probe(struct usb_interface *intf,
>
> /* load patch and sysconfig files for AR3012 */
> if (id->driver_info & BTUSB_ATH3012) {
> + if(le16_to_cpu(udev->descriptor.bcdDevice > 0x0001))
This needs to be like this:
"if (le16_to_cpu(udev->descriptor.bcdDevice) > 0x0001)"
Otherwise you endian swapping is not really doing what it is suppose to
be doing.
> + return -ENODEV;
> +
> ret = ath3k_load_patch(udev);
> if (ret < 0) {
> BT_ERR("Loading patch file failed");
> @@ -390,7 +393,7 @@ static int ath3k_probe(struct usb_interface *intf,
> BT_ERR("Set normal mode failed");
> return ret;
> }
> - ath3k_switch_pid(udev);
> + /* ath3k_switch_pid(udev); */
This is hopefully just a change for testing. Otherwise you break the
other drivers that really do need the PID switching. So please fix this.
> return 0;
> }
>
> diff --git a/drivers/bluetooth/btusb.c b/drivers/bluetooth/btusb.c
> index c2de895..9b67b22 100644
> --- a/drivers/bluetooth/btusb.c
> +++ b/drivers/bluetooth/btusb.c
> @@ -54,6 +54,7 @@ static struct usb_driver btusb_driver;
> #define BTUSB_BCM92035 0x10
> #define BTUSB_BROKEN_ISOC 0x20
> #define BTUSB_WRONG_SCO_MTU 0x40
> +#define BTUSB_ATH3012 0x80
>
> static struct usb_device_id btusb_table[] = {
> /* Generic Bluetooth USB device */
> @@ -110,7 +111,7 @@ static struct usb_device_id blacklist_table[] = {
> { USB_DEVICE(0x03f0, 0x311d), .driver_info = BTUSB_IGNORE },
>
> /* Atheros 3012 with sflash firmware */
> - { USB_DEVICE(0x0cf3, 0x3004), .driver_info = BTUSB_IGNORE },
> + { USB_DEVICE(0x0cf3, 0x3004), .driver_info = BTUSB_ATH3012 },
>
> /* Atheros AR5BBU12 with sflash firmware */
> { USB_DEVICE(0x0489, 0xe02c), .driver_info = BTUSB_IGNORE },
> @@ -914,6 +915,12 @@ static int btusb_probe(struct usb_interface *intf,
> if (ignore_sniffer && id->driver_info & BTUSB_SNIFFER)
> return -ENODEV;
>
> + if (id->driver_info & BTUSB_ATH3012) {
> + struct usb_device *udev = interface_to_usbdev(intf);
> + if(le16_to_cpu(udev->descriptor.bcdDevice <= 0x0001))
See comment above.
> + return -ENODEV;
> + }
> +
> data = kzalloc(sizeof(*data), GFP_KERNEL);
> if (!data)
> return -ENOMEM;
Regards
Marcel
--
To unsubscribe from this list: send the line "unsubscribe linux-bluetooth" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
