FelipeMdeO opened a new pull request, #20100:
URL: https://github.com/apache/nuttx/pull/20100
## Summary
`esp32s3-xiao`'s board bringup never wires up Wi-Fi at all: unlike
`esp32s3-devkit`/`esp32s3-eye`, it has no `#ifdef CONFIG_ESPRESSIF_WIFI`
include of `esp32s3_board_wlan.h` and no call to `board_wlan_init()`.
This PR adds both, mirroring those boards' pattern exactly.
Wiring it up on its own isn't enough for a board that also uses
`CONFIG_PM`, though: Wi-Fi's PHY/RF calibration inside
`board_wlan_init()` cannot tolerate the clock gating of `PM_STANDBY`
(light sleep). If the idle task enters light sleep while `phy_init`'s
calibration is still running -- which the greedy governor is free to
do the moment the CPU goes idle during driver init -- the calibration
hangs forever waiting on a clock that just stopped. This PR also adds
a `pm_stay(PM_IDLE_DOMAIN, PM_IDLE)` / `pm_relax()` guard around
`board_wlan_init()` to prevent that race.
`esp32c3-devkit` and `esp32c6-devkit` also combine `CONFIG_PM` with
Wi-Fi, but don't hit this today: both call `board_wlan_init()` well
before `esp_pmconfigure()` (which is what actually arms the PM
governor) runs, near the very end of bringup -- so PM isn't active yet
during their Wi-Fi init. `esp32s3-xiao`'s PM handling begins earlier in
bringup, which is why the two can race here.
## Impact
- Affects only `esp32s3-xiao`.
- No impact on boards that don't enable `CONFIG_ESPRESSIF_WIFI`, or
that enable it without `CONFIG_PM` -- the new code is compiled out
entirely in both cases.
- No impact on existing behavior for boards that enable both
`CONFIG_PM` and `CONFIG_ESPRESSIF_WIFI` today (none currently do, on
this board -- Wi-Fi wasn't wired up before this PR).
## Testing
Tested on real Seeed XIAO ESP32-S3 hardware (WROOM1-N8R8, 8MB octal
PSRAM), against this branch's `apache/nuttx:master` base, using a
config with `CONFIG_ESPRESSIF_WIFI` + `CONFIG_PM` both enabled (no
tickless scheduling involved -- this race is independent of it).
Without the PM guard (Wi-Fi wired up, `board_wlan_init()` called with
no lock around it): boot hangs 100% of the time, reproducible across
repeated flashes with 50s+ waits, right after the `net80211 rom
version` log line and before `phy_init` ever prints -- never reaches
NSH, and doesn't respond to a UART wakeup keypress either (a genuine
hang, not the board just being asleep).
With the fix in this PR, boot log:
```
I (951) pp: pp rom version: e7ae62f
I (952) net80211: net80211 rom version: e7ae62f
I (980) wifi_init: rx ba win: 6
I (981) wifi_init: LWIP disabled
I (985) phy_init: phy_version 711,97bcf0a2,Aug 25 2025,19:04:10
NuttShell (NSH) NuttX-13.0.1-RC1
nsh>
```
`ps` shows the `wifi` and `netdev-wlan0` kernel threads running
normally alongside the rest of the system:
```
5 0 0 253 RR Kthread - Waiting MQ empty
0000000000000000 0006608 wifi
7 0 0 100 RR Kthread - Waiting Semaphore
0000000000000000 0001960 netdev-wlan0 0x3fcb7168 0
```
`ifup wlan0` and `wapi scan wlan0` both work normally, finding real
access points:
```
nsh> ifup wlan0
ifup wlan0...OK
nsh> wapi scan wlan0
bssid / frequency / signal level / encode / ssid
e8:45:8b:44:a1:90 2437 -68 0800 VIVOFIBRA-WIFI6-A190
00:49:75:a3:55:83 2437 -50 0800 VIVOFIBRA-WIFI6-A190_EXT
94:37:98:2e:f8:fc 2437 -38 0800 repetidor-wifi
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
