On Tue, Apr 21, 2015 at 12:23:15AM +0800, eJim Lee wrote:
Yes, I have tried gtk_monitor+VGA+seaVGABIOS.
Did you see screen corruption when you tried gtk_monitor+VGA+seaVGABIOS?
The following is my log:
LYJ-UPC:~/qemu$ ./build/x86_64-softmmu/qemu-system-x86_64 -enable-kvm -hda tmp/xp_rcc.append -hdb tmp/data.append -m 1024 -smp 1 -device VGA,romfile=/home/lyj/qemu/seabios/out/vgabios.bin -net none -chardev stdio,id=debugcon -device isa-debugcon,chardev=debugcon,iobase=0x0402 SeaBIOS (version rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org)
[...]
set VGA mode 143 VBE current mode=4143 VBE mode info request: 4143 <---- start wbc set VGA mode 3 set VGA mode 3 <---- exit wbc VBE mode set: 4143 set VGA mode 143 VBE current mode=4143 VBE mode info request: 4143 qemu: terminating on signal 2
Was there screen corruption during the above log?
I debugged both vgabios and qemu, and found the wbc didn't use vgabios interface but changing vgabios caused that the wbc run better. I found the wbc r/w vga register directly to set vga mode when added printf on vga_update_display in qemu. Uncomfortably, gfx_write_char have never been invoked and I deleted it in vgafb_write_char, that bring screen looking better. What happend? I can't understand.
I would guess the defect has some randomness to it and minor changes to code layout can cause or avoid the problem.
Kevin, did you repeat screen mess?
I've tried several different configurations (including the last command line you sent), but I have not been able to reproduce any screen corruption.
-Kevin