fan@CS.UCLA.EDU (04/22/87)
------------------- On the 286's selector: bit 0-1 is RPL, bit 2 is TI, and bit 3-15 are the index into the descriptor table. Now we are going to use protected mode to run some programs for the IBM PC. Could it be that some programs have to be rewritten because they use the fact that on the 8086/8088, effective address = (SEGMENT << 4 + OFFSET). Furthermore, would SideKick or any other TSR (Terminate but Stay Resident) utility work? If the TSR program redirects some interrupt vectors to itself by modifying the interrupt vector, wouldn't that generate a trap on the 286's protected mode? Besides, 286 protected mode doesn't use interrupt vectors any more, since IDT is now used. On the 386, there is the virtual 8086 mode. Presumably programs could run unmodified on the 386. Now the question is: should we buy an AT now or should we wait and buy the 386 machines? Roy Fan
dan@prairie.UUCP (Daniel M. Frank) (04/25/87)
In article <5634@shemp.UCLA.EDU> fan@CS.UCLA.EDU (Roy Fan) writes: >------------------- >Now we are going to use >protected mode to run some programs for the IBM PC. Could it be that some >programs have to be rewritten because they use the fact that on the >8086/8088, effective address = (SEGMENT << 4 + OFFSET). Yes. > Furthermore, would SideKick or any other TSR (Terminate but >Stay Resident) utility work? It is unlikely that any software that plays fast and loose with the I/O and memory of the PC is likely to work in protected mode. > On the 386, there is the virtual 8086 mode. Presumably >programs could run unmodified on the 386. Now the question is: >should we buy an AT now or should we wait and buy the 386 machines? How about the third option: don't wait, and buy the 386 machine? There are lots of good ones out there right now. I have a 286 box running System V Unix, so I am no knee-jerk 286 detractor, but I absolutely would not buy a 286 box now that the 386 is available. Not only is the 386 a vastly superior chip for Unix and C programming, but, as you have pointed out, the availability of the virtual-8086 mode means that lots of nice DOS-style control software is going to run only on the new chip. It's kind of funny. The good old 8088 and 8086 family (80188, 80186 included) are still perfectly viable for a range of applications, but I think the 80286 is effectively a dead product. Its feature set and performance have "interim" stamped all over them, in light of the 80386. -- Dan Frank (w9nk) ARPA: dan@db.wisc.edu ATT: (608) 255-0002 (home) UUCP: ... uwvax!prairie!dan (608) 262-4196 (office) SNAILMAIL: 1802 Keyes Ave. Madison, WI 53711-2006