mirror of
https://github.com/meshtastic/firmware.git
synced 2025-06-17 02:22:05 +00:00
98 lines
3.6 KiB
C++
98 lines
3.6 KiB
C++
#include "CryptoEngine.h"
|
|
#include "target_specific.h"
|
|
#include "PortduinoGPIO.h"
|
|
#include "mesh/RF95Interface.h"
|
|
#include "sleep.h"
|
|
|
|
#include <Utility.h>
|
|
#include <assert.h>
|
|
|
|
// FIXME - move getMacAddr/setBluetoothEnable into a HALPlatform class
|
|
|
|
uint32_t hwId; // fixme move into portduino
|
|
|
|
void getMacAddr(uint8_t *dmac)
|
|
{
|
|
if (!hwId) {
|
|
notImplemented("getMacAddr");
|
|
hwId = random();
|
|
}
|
|
|
|
dmac[0] = 0x80;
|
|
dmac[1] = 0;
|
|
dmac[2] = 0;
|
|
dmac[3] = hwId >> 16;
|
|
dmac[4] = hwId >> 8;
|
|
dmac[5] = hwId & 0xff;
|
|
}
|
|
|
|
void setBluetoothEnable(bool on)
|
|
{
|
|
// not needed
|
|
}
|
|
|
|
void cpuDeepSleep(uint64_t msecs) {
|
|
notImplemented("cpuDeepSleep");
|
|
}
|
|
|
|
void updateBatteryLevel(uint8_t level) NOT_IMPLEMENTED("updateBatteryLevel");
|
|
|
|
/** Dear pinetab hardware geeks!
|
|
*
|
|
* The current pinetab lora module has a slight bug. The ch341 part only provides ISR assertions on edges.
|
|
* This makes sense because USB interrupts happen through fast/repeated special irq urbs that are constantly
|
|
* chattering on the USB bus.
|
|
*
|
|
* But this isn't sufficient for level triggered ISR sources like the sx127x radios. The common way that seems to
|
|
* be addressed by cs341 users is to **always** connect the INT# (pin 26 on the ch341f) signal to one of the GPIO signals
|
|
* on the part. I'd recommend connecting that LORA_DIO0/INT# line to pin 19 (data 4) on the pinetab board. This would
|
|
* provide an efficent mechanism so that the (kernel) code in the cs341 driver that I've slightly hacked up to see the
|
|
* current state of LORA_DIO0. Without that access, I can't know if the interrupt is still pending - which would create
|
|
* race conditions in packet handling.
|
|
*
|
|
* My workaround is to poll the status register internally to the sx127x. Which is expensive because it involves a number of
|
|
* i2c transactions and many trips back and forth between kernel and my userspace app. I think shipping the current version
|
|
* of the pinetab lora device would be fine because I can poll slowly (because lora is slow). But if you ever have cause to
|
|
* rev this board. I highly encourage this small change.
|
|
*
|
|
* Btw - your little "USB lora dongle" is really neat. I encourage you to sell it, because even non pinetab customers could
|
|
* use it to easily add lora to rasberry pi, desktop pcs etc...
|
|
*
|
|
* Porduino helper class to do this i2c based polling:
|
|
*/
|
|
class R595PolledIrqPin : public GPIOPin {
|
|
public:
|
|
R595PolledIrqPin() : GPIOPin(LORA_DIO0, "LORA_DIO0") {}
|
|
|
|
/// Read the low level hardware for this pin
|
|
virtual PinStatus readPinHardware()
|
|
{
|
|
if(isrPinStatus < 0)
|
|
return LOW; // No interrupt handler attached, don't bother polling i2c right now
|
|
else {
|
|
extern RadioInterface *rIf; // FIXME, temporary hack until we know if we need to keep this
|
|
|
|
assert(rIf);
|
|
RF95Interface *rIf95 = static_cast<RF95Interface *>(rIf);
|
|
bool p = rIf95->isIRQPending();
|
|
// log(SysGPIO, LogDebug, "R595PolledIrqPin::readPinHardware(%s, %d, %d)", getName(), getPinNum(), p);
|
|
return p ? HIGH : LOW;
|
|
}
|
|
}
|
|
};
|
|
|
|
|
|
/** apps run under portduino can optionally define a portduinoSetup() to
|
|
* use portduino specific init code (such as gpioBind) to setup portduino on their host machine,
|
|
* before running 'arduino' code.
|
|
*/
|
|
void portduinoSetup() {
|
|
printf("Setting up Meshtastic on Porduino...\n");
|
|
|
|
// FIXME: disable while not testing with real hardware
|
|
// gpioBind(new R595PolledIrqPin());
|
|
|
|
// gpioBind((new SimGPIOPin(LORA_RESET, "LORA_RESET")));
|
|
// gpioBind((new SimGPIOPin(RF95_NSS, "RF95_NSS"))->setSilent());
|
|
}
|