pussuw commented on PR #11165: URL: https://github.com/apache/nuttx/pull/11165#issuecomment-1807897534
> > @pkarashchenko @patacongo You might be interested in this. Also, do you mind if I remove the nested cancellation point logic now as well ? > > That is something I had planned to do and is a good idea if you are careful. My primary concern was "sneak" paths that could cause nesting. I don't know if the following is possible but consider this: > > 1. A task sets up a signal handler and calls a cancellation point function, > > 2. The function signals the task, since it is the same task, I think that the signal handler will run immediatedly (true?), > > 3. The signal handler runs and calls another cancellation pointer function, effectively resulting in nesting. > > > If such a think could happen, then that would likely result in a resource leak since the second cancellation point function would not be allowed to do its clean-up. > > One thing that you can do is: > > 1. Retain the nesting count (only) and only if CONFIG_DEBUG_ASSERTIONS is enabled > > 2. Then assert that the nesting count is zero on entry into the cancellation point function. Yes the situation you describe is entirely possible when system calls are disabled. When system calls are enabled the signal delivery gets deferred until the system call is finished. So maybe it is not a good idea for me to fiddle with that part after all.. -- 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: commits-unsubscr...@nuttx.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
