Re: [PATCH RFC] Bluetooth: btintel: Fix WBS setting for Intel legacy ROM products

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Marcel,

On Wed, 2022-01-19 at 18:03 +0100, Marcel Holtmann wrote:
> Hi Tedd,
> 
> > Intel Legacy ROM Products don't support WBS except the SdP(8087:0aa7).
> > But StP2(8087:0a2a) and SdP have the same version information, and
> > btintel cannot distinguish between StP2 and SdP for WBS support.
> > 
> > This patch sets the WBS support flag for SdP based on USB idProduct and
> > uses it in btintel to configure the WBS.
> > 
> > This flag is only applicable for Legacy ROM products.
> > 
> > Fixes: 3df4dfbec0f29 ("Bluetooth: btintel: Move hci quirks to setup routine")
> > Signed-off-by: Tedd Ho-Jeong An <tedd.an@xxxxxxxxx>
> > ---
> > drivers/bluetooth/btintel.c | 10 +++++++---
> > drivers/bluetooth/btintel.h |  1 +
> > drivers/bluetooth/btusb.c   | 12 ++++++++++++
> > 3 files changed, 20 insertions(+), 3 deletions(-)
> > 
> > diff --git a/drivers/bluetooth/btintel.c b/drivers/bluetooth/btintel.c
> > index 1a4f8b227eac..6730c9b2ae33 100644
> > --- a/drivers/bluetooth/btintel.c
> > +++ b/drivers/bluetooth/btintel.c
> > @@ -2428,10 +2428,14 @@ static int btintel_setup_combined(struct hci_dev *hdev)
> > 
> >                         /* Apply the device specific HCI quirks
> >                          *
> > -                        * WBS for SdP - SdP and Stp have a same hw_varaint but
> > -                        * different fw_variant
> > +                        * WBS for SdP - The version information is the same for
> > +                        * both StP2 and SdP, so it cannot be used to
> > +                        * distinguish between StP2 and SdP. Instead, it uses
> > +                        * the flag set by the transport driver(btusb) for
> > +                        * the Legacy ROM SKU and sets the quirk for WBS.
> >                          */
> > -                       if (ver.hw_variant == 0x08 && ver.fw_variant == 0x22)
> > +                       if (btintel_test_flag(hdev,
> > +                                             INTEL_ROM_LEGACY_WBS_SUPPORTED))
> >                                 set_bit(HCI_QUIRK_WIDEBAND_SPEECH_SUPPORTED,
> >                                         &hdev->quirks);
> > 
> > diff --git a/drivers/bluetooth/btintel.h b/drivers/bluetooth/btintel.h
> > index c9b24e9299e2..efdb3d738abf 100644
> > --- a/drivers/bluetooth/btintel.h
> > +++ b/drivers/bluetooth/btintel.h
> > @@ -152,6 +152,7 @@ enum {
> >         INTEL_BROKEN_INITIAL_NCMD,
> >         INTEL_BROKEN_SHUTDOWN_LED,
> >         INTEL_ROM_LEGACY,
> > +       INTEL_ROM_LEGACY_WBS_SUPPORTED,
> > 
> >         __INTEL_NUM_FLAGS,
> > };
> > diff --git a/drivers/bluetooth/btusb.c b/drivers/bluetooth/btusb.c
> > index c30d131da784..286e2fa1ef44 100644
> > --- a/drivers/bluetooth/btusb.c
> > +++ b/drivers/bluetooth/btusb.c
> > @@ -3742,6 +3742,18 @@ static int btusb_probe(struct usb_interface *intf,
> > 
> >                 if (id->driver_info & BTUSB_INTEL_BROKEN_SHUTDOWN_LED)
> >                         btintel_set_flag(hdev, INTEL_BROKEN_SHUTDOWN_LED);
> > +
> > +               /* Intel's Legacy ROM products don't support WBS except
> > +                * the SdP(8087:0aa7). But the StP2(8087:0a2a) and SdP have the
> > +                * same version information, and btintel can't distinguish
> > +                * between StP2 and SdP for the WBS support.
> > +                * It sets the flag here based on the USB PID to enable the WBS
> > +                * support for legacy ROM products.
> > +                * Note that this flag is only applicable to legacy ROM
> > +                * products.
> > +                */
> > +               if (id->idProduct == 0x0aa7)
> > +                       btintel_set_flag(hdev, INTEL_ROM_LEGACY_WBS_SUPPORTED);
> >         }
> > 
> >         if (id->driver_info & BTUSB_MARVELL)
> 
> while this is a total mess from a hardware point of view, I prefer our quirking is kinda the same.
> 
> One way would be to give the idProduct to btintel.c and remove all quirks from btusb.c. Something like this:
> 
> diff --git a/drivers/bluetooth/btintel.h b/drivers/bluetooth/btintel.h
> index c9b24e9299e2..4adb21cf5c20 100644
> --- a/drivers/bluetooth/btintel.h
> +++ b/drivers/bluetooth/btintel.h
> @@ -158,6 +158,7 @@ enum {
>  
>  struct btintel_data {
>         DECLARE_BITMAP(flags, __INTEL_NUM_FLAGS);
> +       u8 usb_pid;
>  };
>  
>  #define btintel_set_flag(hdev, nr)                                     \
> @@ -186,6 +187,12 @@ struct btintel_data {
>  #define btintel_wait_on_flag_timeout(hdev, nr, m, to)                  \
>                 wait_on_bit_timeout(btintel_get_flag(hdev), (nr), m, to)
>  
> +#define btintel_set_usb_pid(hdev, pid)                                 \
> +       do {                                                            \
> +               struct btintel_data *intel = hci_get_priv((hdev));      \
> +               intel->usb_pid = (pid);                                 \
> +       } while (0)
> +
>  #if IS_ENABLED(CONFIG_BT_INTEL)
>  
>  int btintel_check_bdaddr(struct hci_dev *hdev);
> diff --git a/drivers/bluetooth/btusb.c b/drivers/bluetooth/btusb.c
> index c30d131da784..9b5348052421 100644
> --- a/drivers/bluetooth/btusb.c
> +++ b/drivers/bluetooth/btusb.c
> @@ -3737,6 +3737,8 @@ static int btusb_probe(struct usb_interface *intf,
>                 hdev->send = btusb_send_frame_intel;
>                 hdev->cmd_timeout = btusb_intel_cmd_timeout;
>  
> +               btintel_set_usb_pid(hdev, id->idProduct);
> +
>                 if (id->driver_info & BTUSB_INTEL_BROKEN_INITIAL_NCMD)
>                         btintel_set_flag(hdev, INTEL_BROKEN_INITIAL_NCMD);
> 
> Then we need to add a duplicated USB PID table to btintel.c, but everything would be there and all Intel quirks for faulty ROM loader or version information could go to btintel.c.
> 
> Or you create a BTUSB_INTEL_NO_WBS_SUPPORT and add it to 0xa2a and 0x07dc like this:
> 
> diff --git a/drivers/bluetooth/btusb.c b/drivers/bluetooth/btusb.c
> index c30d131da784..a898df89585a 100644
> --- a/drivers/bluetooth/btusb.c
> +++ b/drivers/bluetooth/btusb.c
> @@ -62,6 +62,7 @@ static struct usb_driver btusb_driver;
>  #define BTUSB_QCA_WCN6855      0x1000000
>  #define BTUSB_INTEL_BROKEN_SHUTDOWN_LED        0x2000000
>  #define BTUSB_INTEL_BROKEN_INITIAL_NCMD 0x4000000
> +#define BTUSB_INTEL_NO_WBS_SUPPORT     0x8000000
>  
>  static const struct usb_device_id btusb_table[] = {
>         /* Generic Bluetooth USB device */
> @@ -385,9 +386,11 @@ static const struct usb_device_id blacklist_table[] = {
>         { USB_DEVICE(0x8087, 0x0033), .driver_info = BTUSB_INTEL_COMBINED },
>         { USB_DEVICE(0x8087, 0x07da), .driver_info = BTUSB_CSR },
>         { USB_DEVICE(0x8087, 0x07dc), .driver_info = BTUSB_INTEL_COMBINED |
> +                                                    BTUSB_INTEL_NO_WBS_SUPPORT |
>                                                      BTUSB_INTEL_BROKEN_INITIAL_NCMD |
>                                                      BTUSB_INTEL_BROKEN_SHUTDOWN_LED },
>         { USB_DEVICE(0x8087, 0x0a2a), .driver_info = BTUSB_INTEL_COMBINED |
> +                                                    BTUSB_INTEL_NO_WBS_SUPPORT |
>                                                      BTUSB_INTEL_BROKEN_SHUTDOWN_LED },
>         { USB_DEVICE(0x8087, 0x0a2b), .driver_info = BTUSB_INTEL_COMBINED },
>         { USB_DEVICE(0x8087, 0x0aa7), .driver_info = BTUSB_INTEL_COMBINED |
> @@ -3737,6 +3740,9 @@ static int btusb_probe(struct usb_interface *intf,
>                 hdev->send = btusb_send_frame_intel;
>                 hdev->cmd_timeout = btusb_intel_cmd_timeout;
>  
> +               if (id->driver_info & BTUSB_INTEL_NO_WBS_SUPPORT)
> +                       btintel_set_flag(hdev, INTEL_ROM_LEGACY_NO_WBS);
> +
>                 if (id->driver_info & BTUSB_INTEL_BROKEN_INITIAL_NCMD)
>                         btintel_set_flag(hdev, INTEL_BROKEN_INITIAL_NCMD);
>  
> Frankly, I have no idea what I would favor right now. I do swing a little bit towards the extra btusb.c quirk to keep the btintel.c transport agnostic.
> 

I will go with the second option and will send out the patch.


> Regards
> 
> Marcel
> 





[Index of Archives]     [Bluez Devel]     [Linux Wireless Networking]     [Linux Wireless Personal Area Networking]     [Linux ATH6KL]     [Linux USB Devel]     [Linux Media Drivers]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [Big List of Linux Books]

  Powered by Linux