lekshith
commented
2 days ago Chipset: ESP32S3
IDF version: v5.1.4
SPI uses the spi_device_polling_transmit in all cases from the ESP-IDF v5..1.4
Custom PCB with U12-W5500 module and ethernet port placed less than 2 cm away from the chip.
Ethernet initialisation is the basic initialisation example. I've been using the same ethernet port to host a local web server which works without any issue. Now I have disabled the web server and integrated a new feature i.e., MODBUS poll over TCP/IP stack for which I'm using the freemodbus library. Please have a look at the ethernet initialisation function.
`void init_ethernet(void)
{
esp_netif_config_t netif_cfg_DHCP_client = ESP_NETIF_DEFAULT_ETH();
esp_netif_t *eth_netif = esp_netif_new(&netif_cfg_DHCP_client);
spi_bus_config_t buscfg = {
.miso_io_num = SPI_MISO_GPIO,
.mosi_io_num = SPI_MOSI_GPIO,
.sclk_io_num = SPI_SCLK_GPIO,
.quadwp_io_num = -1,
.quadhd_io_num = -1,
};
ESP_ERROR_CHECK(spi_bus_initialize(ETH_SPI_HOST, &buscfg, SPI_DMA_CH_AUTO));
/* w5500 ethernet driver is based on spi driver */
spi_device_interface_config_t spi_devcfg = {
.mode = 0,
.clock_speed_hz = ETH_SPI_CLOCK_MHZ * 1000 * 1000,
.spics_io_num = ETH_SPI_CS_GPIO,
.queue_size = 20,
.cs_ena_posttrans = 2,
};
eth_w5500_config_t w5500_config = ETH_W5500_DEFAULT_CONFIG(ETH_SPI_HOST, &spi_devcfg);
w5500_config.int_gpio_num = ETH_SPI_INT_GPIO; // Set GPIO number for ethernet interrupt
eth_mac_config_t mac_config = ETH_MAC_DEFAULT_CONFIG();
mac_config.rx_task_stack_size = KILOBYTE * 16; // Set task size
esp_eth_mac_t *mac = esp_eth_mac_new_w5500(&w5500_config, &mac_config);
eth_phy_config_t phy_config = ETH_PHY_DEFAULT_CONFIG();
phy_config.reset_gpio_num = -1;
esp_eth_phy_t *phy = esp_eth_phy_new_w5500(&phy_config);
esp_eth_config_t eth_config = ETH_DEFAULT_CONFIG(mac, phy);
esp_err_t err = esp_eth_driver_install(ð_config, ð_handle);
if (err != ESP_OK)
{
ESP_LOGE(TAG, "Failed to install ethernet w5500 driver %u", err);
}
else
{
uint8_t mac_addr[6];
esp_read_mac(mac_addr, ESP_MAC_ETH);
mac->set_addr(mac, mac_addr); /* Set mac address */
ESP_LOGI(TAG, "--->> Set mac" MACSTR, MAC2STR(mac_addr));
/* Attach Ethernet driver to TCP/IP stack */
ESP_ERROR_CHECK(esp_netif_attach(eth_netif, esp_eth_new_netif_glue(eth_handle)));
/* Register user defined event handers */
ESP_ERROR_CHECK(esp_event_handler_register(ETH_EVENT, ESP_EVENT_ANY_ID, ð_event_handler, NULL));
ESP_ERROR_CHECK(esp_event_handler_register(IP_EVENT, IP_EVENT_ETH_GOT_IP, &got_ip_event_handler, NULL));
/* Start Ethernet driver state machine */
ESP_ERROR_CHECK(esp_eth_start(eth_handle));
}
}`
Answers checklist.
IDF version.
5.1v
Espressif SoC revision.
ESP-IDF v5.1.4, chip revision: v0.1
Operating System used.
Windows
How did you build your project?
VS Code IDE
If you are using Windows, please specify command line type.
None
Development Kit.
ESP32S3
Power Supply used.
USB
What is the expected behavior?
In my application, I have initialised Ethernet in a standard mode with clock speed of 30MHZ. I also use the SPIRAM. Both the application share the same bus config and hosted using SPI_HOST 2. I was previously hosting a web server along side using PSRAM, which never had any problem.
What is the actual behavior?
However, now I have disabled the web server to use ethernet over MODBUS TCP/IP stack. I use the standard freemodbus library and create a continuous polling loop to monitor the incoming packets at the port. Here is the RTOS task created:
And the bus configuration for SPI initialisation:
Steps to reproduce.
The task mostly uses the standard library function and works as expected. polling interval is 500ms. Ethernet is currently just hosting the TCP stack, while the webserver is disabled. PSRAM is used. However, it keeps triggering a unrecoverable error at random times (usually after 30 mins of runtime), Debug logs are shown below:
Debug Logs.
More Information.
I have tried initialising different clock speed for the ethernet. A similar error was specified as bug in this issue https://github.com/espressif/esp-idf/issues/11845 and has been fixed for ESP-IDF 5.0v and above, but I'm still facing this issue. Either way, I still integrated the commit shown, didn't help the issue. Any suggestions are welcome. Thanks.