I have isolated the problem I was having with my AlphaStation 500 hanging
once reentering the SRM on halt or reboot.

Because of a logic bug (pointed out to me by Ross Harvey; thanks!), the
"de" driver was explicitly disabling the mem space where the Tulip
registers were mapped, and using i/o space, instead.

If other drivers had this error, they would simply use the other space,
not explictly disable one of them; the "de" driver is the only one
that does that.

In any case, the SRM console was expecting the Tulip to be memory-mapped,
and since that space was disabled, Bad Things were happening; my AlphaStation
was hanging.  I suspect that the machine checks experienced by Curt
and Matthias are related to this, as well.

As soon as I test a few more things, I plan on committing a fix for this.

Jason R. Thorpe                                       thorpej@nas.nasa.gov
NASA Ames Research Center                            Home: +1 408 866 1912
NAS: M/S 258-6                                       Work: +1 415 604 0935
Moffett Field, CA 94035                             Pager: +1 415 428 6939