Luigi B.

I am building bochs again after almost 2 years. I saw that it has been improved in github. Is this the right place to ask for issues in the github version? For example, I have problems with the save/restore feature with the new memory stuff. Does it work for you?

I saw that bochs is still alive in git. It seems that Ruppert is trying to perfect the voodo masterpiece, maybe to get the Nobel prize in emulation. I also noticed other fixes in cpu and other stuff. Niceeee!

It's too bad if this project is going to die, but I see no hope if Volker and Stanislav go away. Anyway It still is a unique and remarkable piece of software. There are no alternatives. Yes, there is qemu, but it is not easy portable like bochs. You can build bochs and have a very functional PC emulator on any device which supports a C++ compiler and a port of SDL. The CPU emulation of bochs (Stanislav) is very accurate and I think that one of the latest works of Volker, Voodoo emulation, is a masterpiece....

I'm very impressed by the improvements in the banshee. I can run the demo of 3DMark2000 in XP. The benchmark seems not to work correctly but I'm not tested it well. But I have a problem with the save/restore. If I save the bochs session while the demo is running, then after a restore I have a screen flickering with the last image, as if the video processing has not been restarted correctly. It does not happen with voodoo2 (tested on a Win95 guest).

Ah. Ok. what do you think about my other 2 topics? 1) Find a way to have a msr default that avoids the annoying message and does not require a configuration. 2) The ram missing probably due to PCI memory mapping. They are not bugs, just things that could be made more functional I think.

Why do you say that "Timer activation after restore is not necessary." ? If so, what is the meaning in saving the active state? It is set to false in the constructor when the timer is created. Then I suppose it is reset to true by the restore mechanism and... what? Maybe the cpu works in a way that it does not need that because after a restore it start in some clean way and restart the instruction, I don know the details of this. But, if so, why to save the active state of the timer?

Why do you say that "Timer activation after restore is not necessary." ? If so, what is the meaning in saving the active state? It is set to false in the constructor when the timer is created. Then I suppose it is reset to true by the restore mechanism and... what?

Ah. That could explain some issues after restore. I am experimenting with some recent linux systems. I also noticed that sometimes for some systems the mwait makes the system hang or to become very slow. In some cases that could be solved with the option mwait_is_nop. Maybe it it not a bochs issue but the linux distro. I hope that the issue you mentioned will be fixed soon.