Re: [PATCH] ARM64: dts: meson-gxbb-wetek: add ttyAML1 uart for Bluetooth

[Christian Hewitt]

> On 4 Sep 2018, at 11:05 pm, Martin Blumenstingl <martin.blumenstingl@xxxxxxxxxxxxxx> wrote:
> 
> Hi Christian,
> 
> On Tue, Sep 4, 2018 at 4:47 PM chewitt <christianshewitt@xxxxxxxxx> wrote:
>> 
>> This change adds the ttyAML1 uart used by the brmcfmac sdio module in
>> the WeTek Hub and WeTek Play 2 devices.
> do you know which Broadcom chip this is exactly?

Itâs an AMPAK 6335 module. WiFi shows as BCM4339, BT shows as BCM4335(rev C0). Iâve submitted a patch to add the BT id earlier, see:
(https://marc.info/?l=linux-bluetooth&m=153608351032693&w=2)

> I assume you want to use the "patchram" userspace program (or
> something similar) to initialize the HCI controller?
> it's been a while since a similar patch was rejected, but things have
> gotten better since then. more details below
> 
>> Signed-off-by: chewitt <christianshewitt@xxxxxxxxx>
>> ---
>> arch/arm64/boot/dts/amlogic/meson-gxbb-wetek.dtsi | 8 ++++++++
>> 1 file changed, 8 insertions(+)
>> 
>> diff --git a/arch/arm64/boot/dts/amlogic/meson-gxbb-wetek.dtsi b/arch/arm64/boot/dts/amlogic/meson-gxbb-wetek.dtsi
>> index 70325b2..9476868 100644
>> --- a/arch/arm64/boot/dts/amlogic/meson-gxbb-wetek.dtsi
>> +++ b/arch/arm64/boot/dts/amlogic/meson-gxbb-wetek.dtsi
>> @@ -10,6 +10,7 @@
>> / {
>>        aliases {
>>                serial0 = &uart_AO;
>> +               serial1 = &uart_A;
> this can be dropped if you read on
> 
>>                ethernet0 = &ethmac;
>>        };
>> 
>> @@ -239,6 +240,13 @@
>>        vqmmc-supply = <&vddio_boot>;
>> };
>> 
>> +/* This is connected to the Bluetooth module: */
>> +&uart_A {
>> +       status = "okay";
>> +       pinctrl-0 = <&uart_a_pins>, <&uart_a_cts_rts_pins>;
>> +       pinctrl-names = "default";
> insert bluetooth child-node from [0] here (with the correct max-speed,
> *-gpios and clock* properties for this board)
> 
> some background: with Rob's serdev framework some Bluetooth controller
> drivers recently gained the ability to get the controller into a
> working state without any userspace utility. if you add the bluetooth
> child-node here the kernel will:
> - not create a /dev/ttyAMLn
> - load the matching Bluetooth controller driver (the Broadcom one in your case)
> - do the "initialize this Bluetooth controller with all the vendor
> specific magic" dance (toggling GPIOs, loading firmware from userspace
> via standard request_firmware, etc.)
> - provide you with a ready-to-use hci0 interface

No problem, iâve done that for another device. Iâll investigate.

> 
> Regards
> Martin
> 
> 
> [0] https://elixir.bootlin.com/linux/v4.18/source/Documentation/devicetree/bindings/net/broadcom-bluetooth.txt