Swatantra-66 opened a new pull request, #20769: URL: https://github.com/apache/nuttx/pull/20769
## Summary Resolves #14594. In `esp32_txint()` (`arch/xtensa/src/esp32/esp32_serial.c`), when `CONFIG_SUPPRESS_SERIAL_INTS` was active, `spin_unlock_irqrestore(&priv->lock, flags)` was called inside the `if (enable)` block before invoking `uart_xmitchars(dev)`. The execution then exited the `if (enable)` block and hit an unconditional `spin_unlock_irqrestore(&priv->lock, flags)` at the bottom of the function, causing consecutive unlocks on the same lock (issue #14594). This PR restructures `esp32_txint()` so that: 1. The spinlock is held across register manipulation in both `if (enable)` and `else` branches and released exactly once at the end of the critical section. 2. Under `CONFIG_SUPPRESS_SERIAL_INTS`, `uart_xmitchars(dev)` is called safely after `priv->lock` is released, preventing recursive spinlock deadlocks when `uart_xmitchars` calls back into `txint(dev, false)`. 3. Updated the comment to accurately document that `uart_xmitchars()` runs after the spinlock is released. ## Impact - **ESP32 SMP & Debugging**: Prevents kernel panics and CPU interrupt state corruption caused by unlocking an already-unlocked spinlock. - **Concurrency**: Eliminates deadlock risk during recursive upper-half TX character processing. ## Testing - Verified with `./tools/checkpatch.sh -u HEAD` (0 errors, 0 warnings). - Verified with `./tools/nxstyle` on `arch/xtensa/src/esp32/esp32_serial.c` (0 errors, 0 warnings). -- 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]
