WIFI-MESH: Node lost connection after Root reboot (IDFGH-11273) #12431
Assignees
Labels
Comments
github-actions
bot
changed the title
3 tasks
|
I noticed that, in the example, |
|
This problem is displayed only in the root. But all devices have the same config: |
|
@yanmartins |
The log means you should open power save when using coexist. |
espressif-bot
added
Status: In Progress
This setting was active. But I had to disable sleep via software
Thanks, I updated to version
@zhangyanjiaoesp Do you have any information about this? |
yes, If you want to delete the RX task, you can call |
|
@zhangyanjiaoesp in my tests, the ip_internal_network example create the RX task in the NODE after each ROOT reboot. Which generated a memory leak that took a while to resolve. It might be interesting to update this example to avoid this type of problem. |
And yes, updating to version |
thanks, we will check it. |
|
@yanmartins
I'm not clear about your memory leak issue, can you describe it in more details? |
|
Thanks, @zhangyanjiaoesp |
|
If your problem has been solved, please close this ticket, thanks. |
espressif-bot
added
Status: Done
|
@zhangyanjiaoesp after modifying the ip_internal_network example to use a fixed root using PPP data. The same error came back in the NODE: I just comment this configs: /* router */
cfg.channel = 0;
cfg.router.ssid_len = strlen((char *) wifi_params->ssid);
memcpy((uint8_t *) &cfg.router.ssid, wifi_params->ssid, cfg.router.ssid_len);
memcpy((uint8_t *) &cfg.router.password, wifi_params->password, sizeof(wifi_params->password));Add this in the init: esp_mesh_fix_root(true);
if(wifi_mesh_config->is_root) esp_mesh_set_type(MESH_ROOT);And this in the PPP GOT IP event: if(!esp_mesh_is_root()) return false;
mesh_netifs_start(true);
esp_mesh_post_toDS_state(true);
return mesh_netif_start_root_ap(true, dns_addr); |
|
@yanmartins Can you provide a complete log of the Node when the reconnection works well? By the way, I think there are two things you can optimize in your code:
|
espressif-bot
added
Status: In Progress
How about you don't comment the router configure lines? |
Yes! I am sending both cases. No reconnection: With reconnection: |
Yes, I already set
I will add this, but it seems that the error is something else, because the NODE stops looking for ROOT
You tell me to add some random value? If I do this, ROOT will be looking for this connection instead of connecting to PPP |
Since you have set the root to fixed, then the root won't try to connect to router |
|
Is there any code difference between scenarios that can and cannot reconnect? (I see they are both fixed-root) |
The same code
Yes, it sometimes can success, and sometimes fail with the same code. |
@zhangyanjiaoesp Root is fixed, but it continues searching for the wifi network |
you can call APIs in this order to stop reconnection: |
|
@yanmartins you can change the code as the above comment and test again. If the problem still exists, can you provide a demo for us to test locally? If you can't provide a demo, please provide the IDF branch commit id, we can try to add some debug log for you to test. |
I changed the code as you asked. Everything is fine in ROOT, but the error remained the same in NODE I can't provide a demo. But the branch commit is the 5.0.4 tag: 8fbf4ba |
|
@yanmartins please use the attachment to replace the esp32s3 wifi libs in the If the replacement succeeds, the wifi firmware is : And please provide the whole log (from start to error happen) of the root and node device, this will help to analyze the cause of the problem. |
|
@zhangyanjiaoesp Thanks, Succeed: Error: I turn off ROOT sometimes in failure scenario to force the error in NODE. With the timestamp I believe it is easy to observe this in both logs When ROOT stops raising the |
|
@yanmartins
Have you ever tried to call the
|
Yes I tried but the error remained the same. Is five seconds a good amount of time? Update: I increased the time to 5 seconds and the error seems to have stopped happening.
I'm using the ip_internal_network example. So, static esp_err_t mesh_netif_transmit_from_node_sta(void *h, void *buffer, size_t len)
{
mesh_data_t data;
ESP_LOGD(TAG, "Sending to root, dest addr: " MACSTR ", size: %d", MAC2STR((uint8_t *) buffer), len);
data.data = buffer;
data.size = len;
data.proto = MESH_PROTO_AP; //! Node's station transmits data to root's AP
data.tos = MESH_TOS_P2P;
esp_err_t err = esp_mesh_send(NULL, &data, MESH_DATA_TODS, NULL, 0);
if(err != ESP_OK) ESP_LOGE(TAG, "Send with err code %d %s", err, esp_err_to_name(err));
return err;
} |
|
sugget set the |
Thanks @zhangyanjiaoesp. It looks like the reconnection problem is gone. Just one more question: what causes so many prints like these? Is there some configuration I may have done wrong? |
|
when the mesh ie of the parent change, there will be |
|
Thank you very much @zhangyanjiaoesp |
espressif-bot
added
Status: Done
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Answers checklist.
General issue report
Hello,
I'm working in a project using ESP32-S3 with 5.0.1 ESP-IDF.
I run the ip_internal_network example with this init:
Following is the log output of the program on a NODE.
ROOT is turned off, after the
MESH_EVENT_PARENT_CONNECTEDevent on the NODE. But when ROOT is turned on again, NODE never gets a connection. Remaining infinitely in this loop. What could be causing this?In other cases, the output is:
The text was updated successfully, but these errors were encountered: