On 11/11/25 1:24 PM, Wei Deng wrote:
Hi Konrad,
Thanks for your comments.
On 11/10/2025 7:49 PM, Konrad Dybcio wrote:
On 11/10/25 6:57 AM, Wei Deng wrote:
There's a WCN6855 WiFi/Bluetooth module on an M.2 card. To make Bluetooth work, we need to define the necessary device tree nodes, including UART configuration and power supplies.
Since there is no standard M.2 binding in the device tree at present, the PMU is described using dedicated PMU nodes to represent the internal regulators required by the module.
The 3.3V supply for the module is assumed to come directly from the main board supply, which is 12V. To model this in the device tree, we add a fixed 12V regulator node as the DC-IN source and connect it to the 3.3V regulator node.
Signed-off-by: Wei Deng wei.deng@oss.qualcomm.com
[...]
&apps_rsc { @@ -627,6 +708,22 @@ &qupv3_id_2 { status = "okay"; }; +&qup_uart17_cts {
- bias-disable;
+};
+&qup_uart17_rts {
- bias-pull-down;
+};
+&qup_uart17_tx {
- bias-pull-up;
+};
+&qup_uart17_rx {
- bias-pull-down;
+};
This is notably different than all other platforms' bluetooth pin settings - for example pulling down RX sounds odd, since UART signal is supposed to be high at idle
see hamoa.dtsi : qup_uart14_default as an example
I followed the qup_uart17 settings from lemans-ride-common.dtsi. Since these configurations are not required for Bluetooth functionality. I will remove this configuration in the next patch.
This feels like you're essentially saying you don't know/care why you did this before and don't know why you're changing it again. This doesn't give me a lot of confidence. Are you testing your changes on real hw, running an upstream kernel with some distro userland?
Konrad