I have no news yet; I didn't get MHDD burned to a CD last night, but I have done it now, and I will start it running when I get home this evening.
Searching for g_vfs_done did indeed turn up a lot of mentions of this type of problem, and it doesn't seem to be limited to SATA, the chipset, version of FreeBSD, or any other factor in common with my particular setup. I have, however, gotten some more ideas of things to test.
I would like to avoid wiping this install if I can because of the time I've spent adding and configuring the ports I need. There was a lot of compiling and disk accessing that went on before the freezing started. I would also like to avoid buying more hardware just for testing purposes. This is the only SATA drive I have, although I do have some (smaller) unused IDE drives.
This thread implies it could be memory related:
http://www.daemonforums.org/showthread.php?t=2363
which is interesting, and a possibility, because the memory and CD/DVD are the only parts that were in my previous system. The previous motherboard had a component burn out, literally, which left scorch marks on the back of the mb. It might be that the memory was damaged in a way that wasn't apparent until the freezing occurred.
Another thread somewhere suggested it might be a power supply issue, that is, not enough power, which I guess I can test by removing the 2nd HD. The power supply is new, 460W.
Other ideas my search prompted were:
IRQ conflict? (Not my strong suit, does anyone see anything in dmesg?)
Bad SATA cable? (I have another one that came with the HD)
Bad connector on MB? (move to another connector
Bios update? (there are a couple of newer versions available)
Question about MHDD: if the problem is not the drive itself, but somewhere in the connection between the drive and the processor, will that be shown by the test, or will the test register that the drive is bad when it may not be? After I run the MHDD test, I'll run memtest.