Re: [gpsd-users] [gpsd-dev] 3.10 release objectives

[Eric S. Raymond]

Gary E. Miller <address@hidden>:
> I still see that PPS is still randomly failing at startup.  That was not
> in 3.9 needs to be fixed in 3.10.

I'll try to reproduce this here. I can now live-test PPS on the serial port
of the TCX0 eval box.

> And, probably related, the PPS thread is now usually dropping root
> permissions before opening /dev/pps0, so kernel PPS is now broken.
> Sometimes the PPS thread does manage to open /dev/pps0, but when it reads
> pps0 it later it gets no data and PPS fails.
> 
> Given the longstanding bug that PPS can only be started when ggpsd
> starts the solution to all three problems may be just to not drop root
> on the PPS thread.

Er, is that even possible?  Don't all threads associated with a process
get the permissions of the process, whatever they are?
-- 
                <a href="http://www.catb.org/~esr/";>Eric S. Raymond</a>

signature.asc
Description: Digital signature