Description
Description
- Type: Bug
- Priority: Major
Bug
The LoRaMac.cpp resets the parameters _params.sys_params.max_eirp and _params.sys_params.antenna_gain in the wrong way. In one part it uses phy_param.value (uint32_t interpretation of the union) and in another place it uses the phy_param.f_value interpretation (float interpretation)
LoRaMac.cpp:1740 (initialize())
get_phy.attribute = PHY_DEF_MAX_EIRP;
phy_param = lora_phy->get_phy_params(&get_phy);
_params.sys_params.max_eirp = phy_param.f_value;
get_phy.attribute = PHY_DEF_ANTENNA_GAIN;
phy_param = lora_phy->get_phy_params(&get_phy);
_params.sys_params.antenna_gain = phy_param.f_value;LoRaMac.cpp:1333 (reset_mac_parameters())
get_phy.attribute = PHY_DEF_MAX_EIRP;
phy_param = lora_phy->get_phy_params(&get_phy);
_params.sys_params.max_eirp = phy_param.value;
get_phy.attribute = PHY_DEF_ANTENNA_GAIN;
phy_param = lora_phy->get_phy_params(&get_phy);
_params.sys_params.antenna_gain = phy_param.value;Actual behavior
In my case (EU868 PHY selection) it causes the LoRaMac::initialize() function to first initialize the antenna gain correctly to 2.15f and then reset it to the abnormally large uint32_t representation of 1.07*10^9, which consequently causes it to compute phy_tx_power to -128dB in LoRaPHY::tx_config.
Expected behaviour
The interpretation should in both cases be float, i.e. the code should use phy_param.f_value in all cases in antenna gain and max EIRP.
When I manually apply the fix, phy_tx_power gets a value of +13dB.
Steps to reproduce
Use the https://github.com/ARMmbed/mbed-os-example-lorawan/ code and step through the LoRaMac::initialize function and watch the global _params object.
(Debugger screenshot: https://i.imgur.com/uQVJvRN.png)
