MODBUS TCP slave crash on start after stop (IDFGH-6568) #8211
Labels
Comments
github-actions
bot
changed the title
|
Hi @oachkatzalschwoaf , Thank you for reporting this issue. The issue is supposed to be fixed in the commits: I will check this ASAP. Could you check these commits on your side? Thanks. |
|
The issue is confirmed. Please do the following to fix the issue:
The log after the fix is below: Please try the above and let me know the result. Thanks. |
|
now its working but somewehere might be a memory leak |
|
I think what you are measuring here is the cost of heap fragmentation. UPDATE: There are still some leaks in LWIP but after a couple init/destroy cycles the heap size gets stable (log below): Please check this on your side. Thanks. |
|
Thanks @alisitsyn I've tested Heap Tracing feature for the first time last night - wasn't sure about the result, as i just got 7 entries, 6 of them: and a single xMBTCPPortInit but each time 1 missed free with 120byte leaked but after adding I also get the same result. With these changes the issue is fixed for me. |
|
Thank you for update. I guess we use different commits. The leak was in |
alisitsyn
pushed a commit
to alisitsyn/esp-idf
that referenced
this issue
Jan 19, 2022
espressif-bot
added
Status: In Progress
espressif-bot
added
Resolution: NA
|
Thanks for reporting and sorry for slow turnaround, fix on master branch is available at dcaac79. |
espressif-bot
pushed a commit
that referenced
this issue
Mar 9, 2022
espressif-bot
pushed a commit
that referenced
this issue
Mar 13, 2022
|
Thanks for reporting, fixes on release/4.4 and release/4.3 are available, feel free to reopen if the issue still happens. Thanks. |
|
Hello @Alvin1Zhang, @alisitsyn and @oachkatzalschwoaf, i run into the exact same issue, but i am on release/4.4, where this issue normally shouldn't exist anymore?! Looking in the file system, my esp-idf contains the changes of commit dcaac79 linked by @Alvin1Zhang I am not sure if I have to open a new issue or if this issue gets reopened again. Steps to reproduce the issue:
|
|
Hello @dabbinavo, This issue has been corrected. I will check it soon. I will let you know when having some details. |
|
Hello @alisitsyn, ok, thanks for the hint! Now i found the other commit that fixes the issue on the release/v4.4 branch: Switching to the latest commit on release/v4.4 instead of using the tag v4.4 solves the issue. Didn't realize that the issue was fixed after the release of v4.4.. Seems that we have to wait for v4.4.1 to have the fix in an official release. |
|
Hello dabbinavo, I just recently checked this and found the fix in commit ba9c0c2 as well. The example works just fine on my side with release/v4.4 branch. The log:
This was planned to be merged in v4.4 but delays often happen... |
|
Thanks for reporting and sharing the feedback, will update once IDF v4.4.1 releases, thanks. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Environment
Problem Description
Crash/Guru Mediation when trying to initialize modbus-tcp after it got stopped.
Discovered on custom ESP32-S2, reproduced with example code on ESP32
Steps to reproduce
initialize modbus tcp as described in documentation: https://docs.espressif.com/projects/esp-idf/en/latest/esp32/api-reference/protocols/modbus.html?highlight=modbus#modbus-slave-api-overview
start/connect to wifi
mbc_slave_init_tcp()
mbc_slave_setup()
mbc_slave_set_descriptor()
mbc_slave_start()
mbc_slave_destroy()
goto steps 2-5
Code to reproduce this issue
Based on example mb_tcp_slave - changes:
Configured CONFIG_EXAMPLE_WIFI_SSID and CONFIG_EXAMPLE_WIFI_PASSWORD
modified app_main:
void app_main(void) { ESP_ERROR_CHECK(init_services()); // Set UART log level esp_log_level_set(SLAVE_TAG, ESP_LOG_INFO); mb_communication_info_t comm_info = { 0 }; comm_info.ip_addr_type = MB_IPV4; comm_info.ip_mode = MB_MODE_TCP; comm_info.ip_port = MB_TCP_PORT_NUMBER; ESP_ERROR_CHECK(slave_init(&comm_info)); // The Modbus slave logic is located in this function (user handling of Modbus) // slave_operation_func(NULL); while(1){ vTaskDelay(2000); ESP_LOGW(SLAVE_TAG, "Close Modbus..."); slave_destroy(); vTaskDelay(2000); ESP_LOGW(SLAVE_TAG, "Start Modbus..."); slave_init(&comm_info); // < CRASH!!!! } ESP_ERROR_CHECK(slave_destroy()); ESP_ERROR_CHECK(destroy_services()); }Debug Logs
The text was updated successfully, but these errors were encountered: