Re: [PATCH v1] Bluetooth: btusb: Add Qualcomm Bluetooth SoC WCN6855 support

[Rocky Liao]

Hi Marcel,

在 2020-09-14 21:25，Marcel Holtmann 写道：
> Hi Rocky,
>
> > > > This patch add support for WCN6855 i.e. patch and nvm download
> > > > support.
> > > > Signed-off-by: Rocky Liao <rjliao@xxxxxxxxxxxxxx>
> > > > ---
> > > > drivers/bluetooth/btusb.c | 42 
> > > > +++++++++++++++++++++++++++++++++++----
> > > > 1 file changed, 38 insertions(+), 4 deletions(-)
> > > > diff --git a/drivers/bluetooth/btusb.c b/drivers/bluetooth/btusb.c
> > > > index fe80588c7bd3..e51e754ca9b8 100644
> > > > --- a/drivers/bluetooth/btusb.c
> > > > +++ b/drivers/bluetooth/btusb.c
> > > > @@ -59,6 +59,7 @@ static struct usb_driver btusb_driver;
> > > > #define BTUSB_MEDIATEK		0x200000
> > > > #define BTUSB_WIDEBAND_SPEECH	0x400000
> > > > #define BTUSB_VALID_LE_STATES   0x800000
> > > > +#define BTUSB_QCA_WCN6855	0x1000000
> > > > static const struct usb_device_id btusb_table[] = {
> > > > 	/* Generic Bluetooth USB device */
> > > > @@ -273,6 +274,10 @@ static const struct usb_device_id 
> > > > blacklist_table[] = {
> > > > 	{ USB_DEVICE(0x13d3, 0x3496), .driver_info = BTUSB_QCA_ROME },
> > > > 	{ USB_DEVICE(0x13d3, 0x3501), .driver_info = BTUSB_QCA_ROME },
> > > > +	/* QCA WCN6855 chipset */
> > > > +	{ USB_DEVICE(0x0cf3, 0xe600), .driver_info = BTUSB_QCA_WCN6855 |
> > > > +						     BTUSB_WIDEBAND_SPEECH },
> > > > +
> > > > 	/* Broadcom BCM2035 */
> > > > 	{ USB_DEVICE(0x0a5c, 0x2009), .driver_info = BTUSB_BCM92035 },
> > > > 	{ USB_DEVICE(0x0a5c, 0x200a), .driver_info = BTUSB_WRONG_SCO_MTU },
> > > > @@ -3391,6 +3396,26 @@ static int btusb_set_bdaddr_ath3012(struct 
> > > > hci_dev *hdev,
> > > > 	return 0;
> > > > }
> > > > +static int btusb_set_bdaddr_wcn6855(struct hci_dev *hdev,
> > > > +				const bdaddr_t *bdaddr)
> > > > +{
> > > > +	struct sk_buff *skb;
> > > > +	u8 buf[6];
> > > > +	long ret;
> > > > +
> > > > +	memcpy(buf, bdaddr, sizeof(bdaddr_t));
> > > > +
> > > > +	skb = __hci_cmd_sync(hdev, 0xfc14, sizeof(buf), buf, 
> > > > HCI_INIT_TIMEOUT);
> > > > +	if (IS_ERR(skb)) {
> > > > +		ret = PTR_ERR(skb);
> > > > +		bt_dev_err(hdev, "Change address command failed (%ld)", ret);
> > > > +		return ret;
> > > > +	}
> > > > +	kfree_skb(skb);
> > > > +
> > > > +	return 0;
> > > > +}
> > > What is wrong with using qca_set_bdaddr() function.
> > WCN6855 is using different VSC to set the bt addr
>
> int qca_set_bdaddr(struct hci_dev *hdev, const bdaddr_t *bdaddr)
>
> {
>
>         struct sk_buff *skb;
>
>         int err;
>
>
>
>         skb = __hci_cmd_sync_ev(hdev, EDL_WRITE_BD_ADDR_OPCODE, 6,
> bdaddr,
>                                 HCI_EV_VENDOR, HCI_INIT_TIMEOUT);
>
>         if (IS_ERR(skb)) {
>
>                 err = PTR_ERR(skb);
>
>                 bt_dev_err(hdev, "QCA Change address cmd failed (%d)",
> err);
>                 return err;
>
>         }
>
>
>
>         kfree_skb(skb);
>
>
>
>         return 0;
>
> }
>
> EXPORT_SYMBOL_GPL(qca_set_bdaddr);
>
> I see that the other command is using HCI_EV_VENDOR, but is that on
> purpose or an accident? Might want to confirm with the btmon trace.

You are right this is an accident, this command for WCN6855 have command 
complete event return. I will modify this in next patch update. Below is 
the btmon log:

Bluetooth monitor ver 5.48
= Note: Linux version 5.8.0-rc6-hsp-upstream+ (x86_64)                 
0.729933
= Note: Bluetooth subsystem version 2.22                               
0.729934
= New Index: 00:00:00:00:5A:AD (Primary,USB,hci1)               [hci1] 
0.729935
= Open Index: 00:00:00:00:5A:AD                                 [hci1] 
0.729935
= Index Info: 00:00:00:00:5A:AD (Qualcomm)                      [hci1] 
0.729935
= New Index: 00:00:00:00:00:00 (Primary,UART,hci0)              [hci0] 
0.729936
@ MGMT Open: bluetoothd (privileged) version 1.18             {0x0002} 
0.729936
@ MGMT Open: bluetoothd (privileged) version 1.18             {0x0001} 
0.729936
@ MGMT Open: btmon (privileged) version 1.18                  {0x0003} 
0.729945
@ RAW Open: hcitool (privileged) version 2.22               {0x0004} 
122.556176
@ RAW Close: hcitool                                        {0x0004} 
122.556200
@ RAW Open: hcitool (privileged) version 2.22               {0x0004} 
122.556219
@ RAW Close: hcitool                                        {0x0004} 
122.556223
@ RAW Open: hcitool (privileged) version 2.22        {0x0004} [hci1] 
122.556242
< HCI Command: Vendor (0x3f|0x0014) plen 6                 #1 [hci1] 
122.556643
        01 02 03 04 05 06                                ......
> HCI Event: Command Complete (0x0e) plen 4                #2 [hci1] 
> 122.675312
      Vendor (0x3f|0x0014) ncmd 1
        Status: Success (0x00)
@ RAW Close: hcitool                                 {0x0004} [hci1] 
122.675545
@ RAW Open: hcitool (privileged) version 2.22               {0x0004} 
124.528658
@ RAW Close: hcitool                                        {0x0004} 
124.528683
@ RAW Open: hcitool (privileged) version 2.22               {0x0004} 
124.528703
@ RAW Close: hcitool                                        {0x0004} 
124.528708
@ RAW Open: hcitool (privileged) version 2.22        {0x0004} [hci1] 
124.528760
< HCI Command: Read BD ADDR (0x04|0x0009) plen 0           #3 [hci1] 
124.529024
> HCI Event: Command Complete (0x0e) plen 10               #4 [hci1] 
> 124.530311
      Read BD ADDR (0x04|0x0009) ncmd 1
        Status: Success (0x00)
        Address: 01:02:03:04:05:06 (OUI 01-02-03)
@ RAW Close: hcitool                                 {0x0004} [hci1] 
124.530509



> Regards
>
> Marcel