[Asterisk-bsd] Stack size in pthreads

[Wolfgang S. Rupprecht]

kprokazov@svr.kiev.ua (Konstantin Prokazoff) writes:
> Welcome!

Thanks!

>     EAGAIN error in *BSD distribution (I think) reliable to:
> 1. Invalid timestamp calc./scheduling time/different sheduling in *BSD.
> 2. Packets aren't loss. They are passthrough by cache in *BSD.
> Why it works under linux platforms, and takes too much attention in *BSD?
> Still working.

The EAGAIN shouldn't really be generated by the BSD kernel either.  It
is only something that the kernel passes up the user as a last resort
when it can't sleep waiting for more data (because the user won't
allow it), and it has no data to pass to the user.

There was some confusion about linux udp checksum code where linux
also makes use of the EAGAIN bail-out to return to userland when it
has no packets for the user and the user has requested that system
calls don't sleep.  Someone thought that they could test for UDP
checksum errors that way.

-wolfgang
-- 
Wolfgang S. Rupprecht                http://www.wsrcc.com/wolfgang/
openbsd amd64 http://www.wsrcc.com/wolfgang/ftp/asterisk-openbsd35.patch