| PLEX86 | ||
What ever happened to Tandem and NonStop OS 1981I realized that. I'll try. Since I didn't do any of the coding, anything I say will be second-hand. For whom? The monitor itself or the user it's currently servicing? There is big difference. If the monitor forks itself, there isn't any way to unwind itself....gag...I just thought about debugging. Just as a baby user, I'm finding that Unix is terribly sloppy. What ever happened to Tandem and NonStop OS 1983 Which, IME, dooms it because there is no long range plan for extending the feature included in the design... If the unix fork is getting called for the monitor itself, are these parallel or is there a strict bottom-up heirarchy? Correction: once context-PUSH-POP pair minus a few bits. PUSH-POP were user commands that would tell the monitor to save the current context while the user did something that would save his butt. But this is not at exec level; it is a user level context preservation. Why a monitor would have to PUSH-POP its context is way behind my level of extrapolation. And async I-O is not a good reason for this. All I can see is mbuttive headaches in the future with no, absolutely none, side benefits. I suspect (note that this is a stretch of my imagination) that those threads were implemented to allow disk code to be reentrant. This was the problem that JMF never solved when he did a flavor of Unix SMP. BAH What ever happened to Tandem and NonStop OS 1982 The current threads on *n*x is just like c++, a problematic implementation of a pretty good idea. *n*x, in the clbuttic implementation that is done in the BSDs, Linux, SVR4 etc. has...
|
||||
What ever happened to Tandem and NonStop OS 1982 Alt Folklore Computers from Newsgroups The #1 Usenet Provider on the Internet
|
||||