Description of problem: Fedora kernel causes traceback when run in Qemu with kqemu. Following line was used: qemu -cdrom Fedora-8-Test-1-Live-i686.iso -m 512 -net nic,model=rtl8139 -net user -soundhw sb16 -kernel-kqemu Version-Release number of selected component (if applicable): qemu-0.9.0 kqemu-1.3.0_pre11 How reproducible: Run Fedora Test1 ISO image in Qemu, kqemu turned on. Steps to Reproduce: 1. 2. 3. Actual results: traceback Expected results: no traceback :) Additional info: Boot menu -> Uncompressing Linux ... -> Lock dependency validator -> SELinux initializing -> Checking 'hlt' instruction... OK -> invalid opcode: 0000 [#1] -> SMP -> Module linked in: -> CPU: 0 (Traceback follows) -> ... -> Kernel panic ...
So, where is the message it printed?
I guess, I don't get what you think. The message is printed in Qemu's window, hard to get it from it. If you know how to get the information to stdout (i.e. console window) via builtin "seriali line", write me how.
even a jpeg screen capture of the window would be better than nothing.
Created attachment 161362 [details] Captured Qemu screen
Created attachment 161363 [details] Captured Qemu screen
Created attachment 161364 [details] Captured Qemu screen
Created attachment 161365 [details] Captured Qemu screen
Attached four shots of Qemu's screen.
This looks like a bug in kqemu -- it is not supporting the clflush instruction. Nothing we can do about that. 3b: 0f ae 38 clflush (%eax) <== oops here 3e: 58 pop %eax 3f: 5a pop %edx 40: 5b pop %ebx