Setting the DAC write ahead to 1 sample instead of 6

See merge request !636

fix: 18-bit DACs wouldn't set DAC error bit when in HIQTR

See merge request !634

An IOP will set the DAC error bit if FIFO HIQTR is set for six consecutive cycles, but an off-by-two error prevented the IOP from ever counting that high for 18 bit DACs.  20 Bit DACs and 16 bit DACs were not affected.

Adding mfence ops for userspace

See merge request !632

Defining functions for VIO

See merge request !631

Fixing issue where I canged a loop by one, and messed up the DAC delay

See merge request !630

Adding LPTC timing card temp EPICS chan

See merge request !629

Cherry-pick single-threaded awg RPC server into branch 5.1

See merge request !628

This is apparently as tirpc server is meant to be used.  It has unprotected global
state.

Limited testing shows much improved stability (no crashes) when
running basic tests in diaggui.

Adding new LIGO DAC to the RCG/RTS

See merge request !627

Fixing race condition when two models share a DAC, and one is late, the IOP can miss data and won't report anything went wrong.

Switch order so that IOP prefers Ligo Timing Card to old TDS system

See merge request !622

An IOP that sees both a LIGO Timing Card and a Contect 16/16, which an indicator of the old TDS system, will now choose the LIGO Timing Card.

Note that there's still no way to run 1PPS with a contec1616 installed,
and now no way to run either 1PPS or TDS with a LIGO timing card installed.

Add atan2 and phaseunwrap math parts.

See merge request !602

Supress awgtpman "too new" message.

See merge request !597

Also delete send wavform debug string, since this is also sent all the time during an arbitrary stream.

The "too new" error is normal behavior.  Some clients, SIStr in particular, prevent buffer underflows by streaming data until they get a "too new" error message.