driver: bluetooth: HCI: add Ambiq Apollo3 BLE support

[aaronyegx]

This PR implements the SPI based HCI driver for Ambiq Apollo3 Blue/Apollo3 Plus Blue. Apollo3 Blue/Apollo3 Plus Blue BLE integrates one core for running BLE controller only. The MCU can work as BLE host and talk to the integrated BLE controller via SPI based BLEIF module. The SPI BLEIF pins are not exposed from MCU and the new implemented SPI BLEIF driver is added for HCI.

Test passed on the following samples:
-samples/bluetooth/peripheral
-samples/bluetooth/peripheral_hr
-samples/bluetooth/peripheral_ht
-samples/bluetooth/peripheral_esp
-samples/bluetooth/peripheral_dis
-samples/bluetooth/peripheral_gatt_write
-samples/bluetooth/peripheral_sc_only
-samples/bluetooth/central
-samples/bluetooth/central_hr
-samples/bluetooth/central_ht
-samples/bluetooth/central_gatt_write
-samples/bluetooth/broadcaster
-samples/bluetooth/broadcaster_multiple
-samples/bluetooth/beacon
-samples/bluetooth/observer
-samples/bluetooth/scan_adv
-samples/bluetooth/mtu_update
-samples/bluetooth/eddystone
-samples/bluetooth/ibeacon
-samples/bluetooth/hci_uart
-tests/bluetooth/init

[zephyrbot]

The following west manifest projects have been modified in this Pull Request:

Name	Old Revision	New Revision	Diff
hal_ambiq	zephyrproject-rtos/hal_ambiq@94dd874	zephyrproject-rtos/hal_ambiq@fcbbd99 (main)	zephyrproject-rtos/hal_ambiq@94dd874c..fcbbd99e

Note: This message is automatically posted and updated by the Manifest GitHub Action.

[RichardSWheatley]

As far as the Ambiq stuff goes, this looks good.
I will let the other BLE people discuss that section.

[aaronyegx]

@jhedberg it seems you have another PR for HCI driver API change #72323 and HCI Ambiq driver will also be modified (not update yet). So should this PR for Apollo3 support go first or after that PR getting done?

[jhedberg]

@jhedberg it seems you have another PR for HCI driver API change #72323 and HCI Ambiq driver will also be modified (not update yet). So should this PR for Apollo3 support go first or after that PR getting done?

It's likely that the HCI API rewrite will only get merged after the next release, so in that sense any new drivers (using the old API) should be able to be merged before that. For now, I also have #if DT_HAS_CHOSEN(zephyr_bt_hci) blocks in my PR in order to support drivers using the old API, but those might go away before my PR is merged.

[aaronyegx]

@jhedberg Would you mind taking a look at this PR when you are available? Thank you.

[jhedberg]

@aaronyegx approved from my side, but I'll likely need some guidance on how to best convert this to the new HCI API (especially the DTS parts of it).

[aaronyegx]

@aaronyegx approved from my side, but I'll likely need some guidance on how to best convert this to the new HCI API (especially the DTS parts of it).

It is likely we need to have a new "zephyr,bt-hci-spi" binding. And the bt-hci dts node may need to adapt for Apollo4 Blue from:

bt-hci@0 {
	compatible = "ambiq,bt-hci-spi";
	reg = <0>;
	irq-gpios = <&gpio32_63 21 GPIO_ACTIVE_HIGH>;
	reset-gpios = <&gpio32_63 23 GPIO_ACTIVE_LOW>;
	clkreq-gpios = <&gpio32_63 20 GPIO_ACTIVE_HIGH>;
};

to:

bt_hci_spi: bt_hci_spi {
	compatible = "zephyr,bt-hci-spi";
	status = "okay";

	bt-hci@0 {
		compatible = "ambiq,bt-hci-spi";
		reg = <0>;
		irq-gpios = <&gpio32_63 21 GPIO_ACTIVE_HIGH>;
		reset-gpios = <&gpio32_63 23 GPIO_ACTIVE_LOW>;
		clkreq-gpios = <&gpio32_63 20 GPIO_ACTIVE_HIGH>;
	};
};

chosen {
	zephyr,bt-hci = &bt_hci_spi;
};

In the driver, I think we also need a hci_spi_config struct like:

struct hci_spi_config {
	struct bt_hci_info info;
	const struct device *hci_spi;
	k_thread_stack_t *rx_thread_stack;
	size_t rx_thread_stack_size;
	struct k_thread *rx_thread;
};

but not sure if we must include all global variables to one new hci_spi_data structure.
For the API adaption I think you already have a good implementation for drivers/bluetooth/hci/h4.c in #72323. The hci_ambiq.c should be similar.

Anyway, I will test the changes once you finish the upgrading on the supported Apollo3/4 EVB.

[aaronyegx]

@RichardSWheatley Please help take another look, thanks.