network/tasks/main.yml: WARN if WiFi firmware crashed since boot

Author: holta

roles/firmware/tasks/warn_fw_crash.yml

@@ -0,0 +1,38 @@
+# 2025-03-31: WiFi firmware was crashing OFTEN in roles/network/tasks/NM-debian.yml
+# so let's help everyone better monitor in future, on RPi or elsewhere.
+# Extensive March 2025 testing summarized in roles/network/tasks/main.yml
+# (just below NM-debian.yml) and at https://github.com/iiab/iiab/pull/3965
+
+
+- name: Did WiFi firmware crash since boot? Run 'ls /sys/class/ieee80211/ | head -1' to check
+  shell: ls /sys/class/ieee80211/ | head -1
+  register: phyname
+  failed_when: False    # Stronger than 'ignore_errors: yes' (defer red errors until below!)
+
+- name: Did WiFi firmware crash since boot? Run '! dmesg | grep "Firmware has halted or crashed"' to check
+  shell: '! dmesg | grep "Firmware has halted or crashed"'
+  register: dmesg_fw_crash
+  failed_when: False
+
+
+- name: WARN if WiFi firmware appears to have crashed since boot (intentionally show red error)
+  fail:
+    msg: "WARNING: /sys/class/ieee80211/{{ phyname.stdout }} (normally phy0) SUGGESTS YOUR WIFI FIRMWARE CRASHED SINCE BOOT."
+  when: phyname.stdout != "phy0"
+  ignore_errors: yes
+
+- name: WARN if WiFi firmware appears to have crashed since boot (intentionally show red error)
+  fail:
+    msg: "WARNING: dmesg SHOWS WIFI FIRMWARE CRASH(ES) SINCE BOOT... {{ dmesg_fw_crash.stdout }}"
+  when: dmesg_fw_crash.rc != 0
+  ignore_errors: yes
+
+
+- fail:    # Also intentionally red, with a 1 minute pause below, for extra emphasis
+    msg: "PLEASE CONSIDER: (1) adding 'brcmfmac.debug=0x100000' to /boot/firmware/cmdline.txt (and then reboot) to provide much more logging, (2) running 'sudo iiab-diagnostics' to submit a bug report, (3) reading https://github.com/iiab/iiab/pull/3965 to learn more."
+  when: phyname.stdout != "phy0" or dmesg_fw_crash.rc != 0
+  ignore_errors: yes
+
+- pause:
+    minutes: 1
+  when: phyname.stdout != "phy0" or dmesg_fw_crash.rc != 0

roles/network/tasks/main.yml

@@ -110,6 +110,9 @@
     name: firmware
   when: rpi_model != "none"
 
+- name: "2025-03-31: WiFi firmware was crashing OFTEN in NM-debian.yml above, so let's help everyone better monitor in future, on RPi or elsewhere"
+  include_tasks: roles/firmware/tasks/warn_fw_crash.yml
+
 
 - name: Create {{ iiab_etc_path }}/install-flags/iiab-network-complete on second pass of network role
   file: