Return-Path: Message-ID: Date: Sun, 21 Jun 2015 12:23:07 +0800 From: Xi Ruoyao MIME-Version: 1.0 To: Larry Finger , Marcel Holtmann CC: "Gustavo F. Padovan" , Johan Hedberg , Linux Bluetooth Maillist , Linux Kernel Maillist , Daniel Drake , Xi Ruoyao Subject: Re: [PATCH] Bluetooth: btusb: Add Realtek devices into module device table References: <55860F8C.4050205@lwfinger.net> In-Reply-To: Content-Type: text/plain; charset="windows-1252"; format=flowed List-ID: On 06/21/2015 12:11 PM, Xi Ruoyao wrote: > On 06/21/2015 09:12 AM, Larry Finger wrote: >> On 06/20/2015 04:58 PM, Marcel Holtmann wrote: >>> Hi Xi, >>> >>>> In 'commit a2698a9bf9b0 ("Bluetooth: btusb: Add Realtek >>>> 8723A/8723B/8761A/8821A support"), support of some Realtek >>>> devices was added to the Generic Bluetooth USB driver in kernel. >>>> However, these devices are not in the module device table of >>>> btusb.ko, so the kernel wouldn't probe them at all. >>>> >>>> To fix this, add four entries in the device table btusb_table, >>>> based on code from >>>> ('new' branch). >>>> >>>> This enables bluetooth support in the Lenovo Ideapad Yoga 13 >>>> which has RTL8723AU USB device, with product ID 0bda:1724. >>>> >>>> Signed-off-by: Xi Ruoyao >>>> --- >>>> drivers/bluetooth/btusb.c | 6 ++++++ >>>> 1 file changed, 6 insertions(+) >>>> >>>> diff --git a/drivers/bluetooth/btusb.c b/drivers/bluetooth/btusb.c >>>> index 3c10d4d..dd87623 100644 >>>> --- a/drivers/bluetooth/btusb.c >>>> +++ b/drivers/bluetooth/btusb.c >>>> @@ -148,6 +148,12 @@ static const struct usb_device_id >>>> btusb_table[] = { >>>> { USB_DEVICE(0x8087, 0x0a5a), >>>> .driver_info = BTUSB_INTEL_BOOT | BTUSB_BROKEN_ISOC }, >>>> >>>> + /* Realtek Bluetooth */ >>>> + { USB_VENDOR_AND_INTERFACE_INFO(0x0bda, 0xe0, 0x01, 0x01) }, >>>> + { USB_VENDOR_AND_INTERFACE_INFO(0x0bd5, 0xe0, 0x01, 0x01) }, >>>> + { USB_VENDOR_AND_INTERFACE_INFO(0x13d3, 0xe0, 0x01, 0x01) }, >>>> + { USB_VENDOR_AND_INTERFACE_INFO(0x0489, 0xe0, 0x01, 0x01) }, >>>> + >>> >>> I highly doubt these are needed. There is a generic entry for an USB >>> transport that is already covering these. >> >> Something changed in recent kernels, and the generic USB transport >> entries do not seem to be recognized in the case where a wifi and BT >> device share a single USB ID. As the OP's tablet has such a case, I >> suggest that 0bda line be accepted, but not the others. If I had a >> computer with an RTL8723AU, I would have a chance at debugging what >> changed, but sadly, that is not the case. >> >> Larry >> > > I viewed the output of 'lsusb -d 0bda:1724 -v | less', and I found > that my RTL8723AU has 'bDeviceClass' 239 (0xef, means 'Miscellaneous > Device') > but 'bInterfaceClass' 224 (0xe0, means 'Wireless'). > > So maybe the Generic Bluetooth USB device entry should be > '{ USB_INTERFACE_INFO(0xe0, 0x01, 0x01) },' > instead of > '{ USB_DEVICE_INFO(0xe0, 0x01, 0x01) },' > > Now I'm compiling this. > This approach works well on my tablet. So, which way is better, edit Generic Bluetooth USB device entry, or add a new 0bda entry? And below the Generic Bluetooth USB device entry there is Generic Bluetooth AMP device entry. Should we use USB_INTERFACE_INFO on this too? -- Xi Ruoyao School of Aerospace Science and Technology Xidian University, Xi'an, China