https://bugzilla.novell.com/show_bug.cgi?id=231104 ------- Comment #37 from snwint@novell.com 2007-03-01 07:50 MST ------- Ok, here the story so far. The problem with those VIA cpus is that the real to protected mode switch does not work properly. In particular it *looks* like it does not flush its prefetch queues which results in interpreting 32 bit opcodes as 16 bit codes and vice versa. The numbers are checkpoints. Each just a plain 'mov' instruction into display memory. So not much could go wrong there. '3' was the first after the real->pm switch and the fact that it's not shown indicates the instruction is not decoded properly. I *know* the program flow comes across this instruction as a result of test5, where I put just a 'hlt' there (halting the processor). After some time the prefetch queues are empty and the processor works normally (around checkpoints '4' or '5'). Unfortunately all crashes in the end because it has missed vital internal init functions on the way. :-( The various tests were attempts to trick the cpu into flushing its buffers. With only some success ('a'-'c' shown), but not '3'. Now, to test13: It will prompt for a number at the beginning. Only '0' to '3' are meaningful (all other keys are equal to '0'). With '0' it should print numbers as before with a delay after some of them. Which are shown? In particular: what about '2'? It will reboot/crash around number '9' at latest. By selecting '1' to '3' it will *probably* stop earlier (around '1', I think). The questions are: does it stop? And where? Please try '1' to '3' even if it reboots right away with '0'. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.