yangsong8-a1 opened a new pull request, #21886:
URL: https://github.com/apache/nuttx/pull/21886

   ## Summary
   
   The watchdog device could only be stopped at runtime (magic 'V' write);
   it could not be started again. Two commits:
   
   - a write whose first byte is 'V' stops the watchdog (or the
     auto-monitor) as before,
   - any other non-empty write starts it again (the auto-monitor, or the
     lower half start()),
   - an empty write is still a keepalive.
   
   The second commit replaces the single-pass loop in wdog_write() with the
   equivalent if/else, removing structurally dead code.
   
   ## Impact
   
   Behavior change for applications. With CONFIG_WATCHDOG_MAGIC_V and no
   CONFIG_WATCHDOG_AUTOMONITOR_TASKS, write(fd, "x", 1) used to feed the
   watchdog, as on Linux. Now it calls start() instead. An empty write is
   not passed to the driver by the VFS, so write() can no longer feed the
   watchdog at all; feeding needs WDIOC_KEEPALIVE. A lower half without
   start() faults on a non-'V' write, because the new path calls
   lower->ops->start() directly instead of WD_START().
   
   ## Testing
   
   Ubuntu 22.04, gcc 13.1. sim:nsh + CONFIG_WATCHDOG, fake lower-half
   test program (full methods / keepalive-only):
   
   | Call | apache/master | This branch |
   |---|---|---|
   | /dev/wdfull: write("x") | keepalive | start |
   | /dev/wdfull: write("", 0) | driver not called | driver not called |
   | /dev/wdfull: write("V") | stop | stop |
   | /dev/wdmin (keepalive only): write("x") | keepalive | segmentation fault 
(start is NULL) |
   


-- 
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]

Reply via email to