On Sat, 13 Jun 2020 10:05:19 +0200, Martin Pieuchot wrote: > Diff below extends the existing kqfilter handlers to be able to set > POLLHUP in the new poll(2) implementation. > > __EV_HUP is introduced and now set for this purpose. A new kqfilter > for deadfs is introduced to match the existing poll handler.
OK. I wonder if POLLERR is actually more appropriate for this but I understand you are just matching existing behavior at this point. > __EV_HUP is not exactly like EV_EOF. Very few handlers set POLLHUP so > this approach is conservative. Right, there are some subtle differences. > __EV_HUP reuses the EV_FLAG1 flag which isn't used for `file_filtops'. > We could decided to set it only if __EV_POLL is set which should ensure > it is not exported to userland. Is it desirable? I think that would be safest. Once this is complete we may want to compare our poll support to other systems. I suspect we do things differently for some of the corner cases. For example, how we handle the write end of a pipe that has had its read end closed looks difference thant, say, Linux. Of course that is out of scope for now. - todd
