This forums is for OS project announcements including project openings, new releases, update notices, test requests, and job openings (both paying and volunteer).
This release has (as usual) been delayed. I have started a new very cool job that takes almost all my time. Instead of just fixing bugs I have also refactored the entire driver and file system layer. It is normally a big mistake to change code in a beta phase but it made it much easier for me to fix bugs and it will be much easier to add features in the future. I just hope I have not introduced too many new bugs
internal kernel error in src/kernel/memory/memory.c at line 104:
failed expression: '0 != address_space->cr3'
Tried it with a VFD image, with the same result. Tried it again in Sun xVM VirtualBox 2.1.4, with the same result. Too bad, the screen shots look really promising.
EDIT: in Bochs 2.3.7, at first it boots fine, but hangs at "Interface 81%" when loading the GUI. Conway's Game of Life is great though!
Roel
Last edited by qw on Mon Oct 05, 2009 5:57 am, edited 2 times in total.
Hi,
i have tested it on bochs( running fine), and on a real computer , running very good.
the gui works greate, especial on the real, (aproximent 200 FPS). But i can not chose color mode ? not even in bochs.
Keep up the good work
KMT dk
well, what to say, to much to do in too little space.
when it goes up hill, increase work, when it goes straight, test yourself but when going down, slow down.
internal kernel error in src/kernel/memory/memory.c at line 104:
failed expression: '0 != address_space->cr3'
Tried it with a VFD image, with the same result. Tried it again in Sun xVM VirtualBox 2.1.4, with the same result. Too bad, the screen shots look really promising.
Mojo needs at least 32 Mb memory to boot. I assume that this is the cause. I have adjusted the memory check during boot to reflect this fact. Remember to use the '--vm=virtual-pc' kernel parameter when testing the GUI under Virtual PC.
Hobbes wrote:EDIT: in Bochs 2.3.7, at first it boots fine, but hangs at "Interface 81%" when loading the GUI.
I forgot to add an image resource to the floppy image. I have corrected the error and updated my release procedure to catch this mistake in the future.
Tommy wrote:Tested on Virtual PC 2007, it boots fine.
But when a run the GUI, I obtain a General Protection Fault at EIP=0x01004db5
You need to use the kernel parameter '--vm=virtual-pc' when testing the GUI under Virtual PC. My SSE detection code doesn't work under VPC. I haven't investigated the cause yet but just provided a workaround.
But since I aim to deliver robust software I have added this as a bug. I will also add the kernel parameter workaround to the FAQ I have started writing.
kmtdk wrote:i have tested it on bochs( running fine), and on a real computer , running very good.
the gui works greate, especial on the real, (aproximent 200 FPS). But i can not chose color mode ? not even in bochs.
I have not yet implemented support for anything but 16 color VGA mode on anything but Bochs and QEMU. Which version of Bochs are you using? I just tested and it works with version 2.4.1 running under Windows.
This is a little off topic but doesn't it bug you that you need hacks and workarounds for every different virtualization program? Like it's your fault that SSE detection does not work under Virtual PC.
clange wrote:
You need to use the kernel parameter '--vm=virtual-pc' when testing the GUI under Virtual PC. My SSE detection code doesn't work under VPC. I haven't investigated the cause yet but just provided a workaround.
But since I aim to deliver robust software I have added this as a bug. I will also add the kernel parameter workaround to the FAQ I have started writing.
Tested with "--vm=virtual-pc" (I have also removed "--disablegui" parameter), but I still have an exception at the same address ... with a sexy girl ASCII image!
I think I should test it in another emulator...
Hobbes wrote:This is a little off topic but doesn't it bug you that you need hacks and workarounds for every different virtualization program? Like it's your fault that SSE detection does not work under Virtual PC.
Hmmm, a bit of a philosophical question It does annoys me that I have to make workarounds and hack - but at the same time it is part of software development. I don't mind that some of my code is messy and ugly if that is what is needed - as long as I am able to confine the mess and put a pretty interface on it.
Tommy wrote:Tested with "--vm=virtual-pc" (I have also removed "--disablegui" parameter), but I still have an exception at the same address ... with a sexy girl ASCII image!
I think I should test it in another emulator...
Hmm, it seems that I have some problems with the caching of the content of my homepage. I just checked that it is actaully the newest release that I uploaded and it is. But when I try downloading it I get the old version. I will solve it later today and let you know. I sincerely apologize for waisting your time. Thanks for testing again.
I have tested it on QEmu... Wow!
It's in black & white, but it is fast! And the GUI is really good!
Very nice work!
EDIT: I think your GUI use double-buffering (I saw a FPS value on the bottom of the screen). Do you use SSE instructions to
copy the backbuffer in the framebuffer?
It seems like the new uploads have propagated through the caches.
Tommy wrote:I have tested it on QEmu... Wow!
It's in black & white, but it is fast! And the GUI is really good!
Very nice work!
Thanks You need to use the '-vga std' to be able to use the BGA driver.
Tommy wrote:EDIT: I think your GUI use double-buffering (I saw a FPS value on the bottom of the screen). Do you use SSE instructions to
copy the backbuffer in the framebuffer?
Yes, I use double buffering and I also use SSE2 or MMX to do the copying if the instructions are available. It is the only major optimization I have made and it actually made quite a difference. Just using simple rep movsd instead of a for or while loop also gave a respectable speedup. I didn't do any accurate measurements of the copying alone but the frame rate rose about a factor 2-4 all in all.