KA630 emulator: minor milestone!

Perhaps it does; perhaps it did, even. When I first started, I _think_ I went and looked and came to the conclusion that simh was identifiably different from all real hardware VAXen (different SID value or some such). Perhaps I was mistaken, or perhaps simh has changed in this respect in the meantime - it's certainly been long enough. As for why the KA630, it's the VAX implementation I have the most hardware copies of and the one I know best. It's also a relatively simple target; for example, the timer hardware (ICCS et al) is about as rudimentary as it can get and still deserve the name. As for why do my own implementation, partly because of the above impression of simh and partly because I have a strong tendency to roll my own in most cases (a tendency with both positive and negative aspects, of course). My memory is telling me that simh was difficult to build for some reason, but that memory is so fuzzy it could very well be mis-attaching a memory from thing B to (inapproprate) thing A. I also specifically wanted to do a KA630 simulator accurate enough to pass the ROM selftests from a real KA630. My impression was that simh was not faithful at the bug-for-bug level. One of the first things I ran into was that the KA630 has an instruction prefetch buffer - and the ROM code depends on being able to execute an instruction out of the prefetch buffer after turning on MAPEN. I've had to do a number of such contrivances, things which are (as far as I can tell) not documented anywhere but which the selftest code depends on. Besides the prefetch buffer, I've found that the selftest code depends on sending and receiving characters taking time - and depends on the apparent fact that, when in loopback, TXCS_RDY is set after, or only a few instructions before, RXCS_DONE. And that, if both receive and transmit interrupt conditions are met, the receive interrupt is the one that's taken. And that the SLU has at least two but no more than four characters worth of receive FIFO. And that MSER_WRWPAR actually works. And...you get the idea. Unfortunately I can't seem to check the simh stuff now. The URL given on the trailing-edge.com webpage (sigh!) for modern simh is https://github.com/simh/simh, which AIUI would lead me to git clone git://simh.github.com/simh, but simh.github.com appears to broken, or behind broken firewalling, or some such: it answers pings but doesn't answer port-9418 SYNs. Whether it's my misunderstanding of how to derive the git clone argument or their misbehaving host, I am unable to clone it at the moment and thus can't look at it. There's that, too. I greatly enjoy creating software and I've had great fun hacking on the KA630 simulator so far and expect to enjoy the further work. /~\ The ASCII Mouse \ / Ribbon Campaign X Against HTML mouse at rodents-montreal.org / \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B