Implement end-to-end WiFI functionality #127
Conversation
|
I'm looking through the code still, but some things that stand out from a style perspective:
|
There was a problem hiding this comment.
I've done a first pass review. I didn't focus too much on the python code, mostly on the C++ and C code.
The main class of bugs I found are handling arrays and their sizes. Some APIs need to be tweaked to ether take in an std::array with a known fixed size, or allow passing in both a pointer and a size of the array, and actually check it internally.
Additionally, I don't think malloc is needed in most cases it's being used, unless I'm mising something.
| std::map<int, Module*> req_map; | ||
|
|
||
| /** Store reference to last module */ | ||
| Module* last_module; |
There was a problem hiding this comment.
I'll check once I get to the constructor, but this really needs to be always defined as something (even if just nullptr). Before access it should always be checked as well.
|
|
||
| typedef struct { | ||
| /** Size of receive/request buffer */ | ||
| static const size_t size = Esp32Command_size; |
There was a problem hiding this comment.
Why have a constant in a struct definition? Might as well make it a global constant (looks like Esp32Command_size is already one).
| /** Size of receive/request buffer */ | ||
| static const size_t size = Esp32Command_size; | ||
| /** Buffer to store data */ | ||
| uint8_t data[size] = {}; |
There was a problem hiding this comment.
Might want to use std::array instead of raw arrays.
| #ifdef UNIT_TEST | ||
| static const uint8_t module_handler_buffer_size = 32; | ||
| static uint8_t module_handler_rx_buffer[module_handler_buffer_size] = {}; | ||
| static uint8_t module_handler_tx_buffer[module_handler_buffer_size] = {}; |
| /** Store reference to last module */ | ||
| Module* last_module; | ||
|
|
||
| typedef struct { |
There was a problem hiding this comment.
This is C++, no need to use typedef here.
| } else { | ||
| // create small buffer | ||
| Buffer chunk = {}; | ||
| chunk.data = malloc(g_i2c_buffer_size); |
There was a problem hiding this comment.
Same here. Is dynamic memory needed?
| } Buffer; | ||
|
|
||
| /** @brief Buffer for ControllerTransmit */ | ||
| static Buffer tx = {}; |
There was a problem hiding this comment.
This construct makes a Buffer variable per translation unit (i.e. file) that includes this header. Is this intentional? Or is the intent to have global tx and rx buffers?
Additionally, in C I think = {0} is needed to zero initialize stuff.
| tx.len = 0; | ||
|
|
||
| // allocate rx buffer | ||
| rx.data = (uint8_t*) malloc(buffer_size); |
There was a problem hiding this comment.
Is dynamic memory needed here? Esp32Command_size is known beforehand (right?)
| cmd = DecodeEsp32Command(rx.data, rx.len); | ||
|
|
||
| // copy response | ||
| memcpy(resp, cmd.command.wifi_command.resp.bytes, cmd.command.wifi_command.resp.size); |
There was a problem hiding this comment.
a check to see if it's safe to copy cmd.command.wifi_command.resp.size into resp needs to be done.
There was a problem hiding this comment.
Where is this data structure used? It's introduced in this pull request, but I don't see anything actually using it other than tests.
Name/Affiliation/Title
John, Maintainer
Purpose of the PR
Add end-to-end support for uploading via WiFi. Starting at the stm32, a command should be sent to the esp32 which uploads a measurement to Dirtviz.
Development Environment
Provide the OS, hardware version, Platformio version, and information about any other tools used during the development.
Test Procedure
TBD
Additional Context
Incremental PR for buffer support. This PR adds the
ModuleHandlerclass that handles directing arbitrary command to various "modules" over the i2c interface between the stm32 and esp32. Based on the message type, the data gets "forwarded" to the correct module for processing. The WiFi module is the first implemented one. The commands are encoded via protobuf that tell the esp32 to connect and return the current timestamp as RTC synchronization and an upload command for actual data.Task List
CHANGELOG.mdRelevant Issues
Closes #46