Msg: 5972 *Conference*
08-17-95 19:20:28
From: RON WIESEN
To : TRACY ALLEN
Subj: REPLY TO MSG #5971 (UNDELETE -- COLD START RECOVERY)
Thanks for the 8K interval test byte info. I assumed a READ/SAVE/WRITE something other than zero (read unequipped mem space probably yields the low byte of the address due to high/low bus demultiplex) test of this kind. Nice to have your facts on this. The interval I'm concerned with is not related to testing for equipped RAM. What I've seen appears to be remnants of a loop where after N iterations the OS writes a byte in RAM. The address is bumped in each iteration (up/down, matters not) but eventually the OS "discovers" it's in a loop and then cold-starts. It's just that I never noted the modulus of N in the cold-start remnants that I've looked at. They were unintended cold-starts. Still, I'll keep in mind the 8K test but I doubt that the OS ever fails to restore the original byte. After all, once a cold-start begins the OS is "sane" so it would not get "insane again" during the 8K RAM boundary test.