Just posted a comment in GitHub, but duplicating it here for visibility. The change might not be needed with https://github.com/apache/nuttx/pull/8951 as it removes access to pid hash table during sem_post()
Best regards, Petro On Wed, Apr 5, 2023, 12:53 AM Gregory Nutt <spudan...@gmail.com> wrote: > > On 4/4/2023 3:43 PM, Fotis Panagiotopoulos wrote: > > I drafted this change here: > > > > https://github.com/apache/nuttx/pull/8965 > > > > Have a look, please. > > > > Although, I don't know if there will be any other places to be fixed like > > that? > This is a unique problem since kmm_free is the only thing that was > called in the unhealthy state. I think this is a pretty robust solution > and I can't think of any broader implications. > > Or if this is the best approach to this issue. This was just a quick > > experiment. > I think it is pretty clean. It is really just an ordering problem, so > changing the ordering seems like a good solution. >