Page 1 of 1

vesa checksum

Posted: Thu Aug 07, 2008 1:09 pm
by Adek336
Why my Vesa PM Info Block has the cheksum field filled with a 0 while the sum of the fields is 0xF8 :D Do your computers have the PM Info Block checksum properly filled?

Re: vesa checksum

Posted: Thu Aug 07, 2008 1:33 pm
by Adek336
Do Bochs/VirtualBox support Vesa? :(
Am I right I have to scan between 0xC0000 and 0xD0000 for the pm info block? The specs don't say that the block has to be 2-byte,4-byte,8-byte aligned etc. Does it mean I shouldn't assume it's 2-byte/etc aligned?

Re: vesa checksum

Posted: Thu Aug 07, 2008 7:18 pm
by -m32
Adek336 wrote:Do Bochs/VirtualBox support Vesa? :(
Am I right I have to scan between 0xC0000 and 0xD0000 for the pm info block? The specs don't say that the block has to be 2-byte,4-byte,8-byte aligned etc. Does it mean I shouldn't assume it's 2-byte/etc aligned?
I think they support Vesa, but not the VBE3 Protected Mode interface. I wouldn't count on the PMID signature being aligned at all. I've found it at many different places.
Adek336 wrote:Why my Vesa PM Info Block has the cheksum field filled with a 0 while the sum of the fields is 0xF8 :D Do your computers have the PM Info Block checksum properly filled?
I find that the checksum value is wrong on many boxes - especially ones by Dell, yet they do _have_ a PMI Block. Not sure why they'd provide one but not give a checksum...

Re: vesa checksum

Posted: Thu Aug 07, 2008 7:26 pm
by Adek336
I think they support Vesa, but not the VBE3 Protected Mode interface.
That's too bad for them, they won't have the opportunity to run panaLiX :D It seems my Dell Latitude laptop with broken vbe3 checksum gets all the fun!

Re: vesa checksum

Posted: Fri Aug 08, 2008 12:53 am
by Adek336
Vesa problem: I was aware my interrupt handling wasn't able to service interrupts while the kernel was executing the 16bit vbe-dedicated code segment descriptor selector, but only after 5 hours of stepping with BIOS code without a debugger I found out that two instructions before retf, BIOS does "sti" and a random timer interrupt unleashes hell :D I was so sure it wouldn't sti I didn't even check :D

Re: vesa checksum

Posted: Fri Aug 08, 2008 12:53 am
by Adek336
I wish crashes would unleash shell :D

Re: vesa checksum

Posted: Fri Aug 08, 2008 3:23 pm
by Adek336
I am having some troubles with VBE on an ATI card.
I choose to use a mode, I get the mode-information block all right, resolution, physical base address is like 0xE8000000, I switch to the mode, the screen goes blank, I map the phys address in, fill it with data, nothing appears:(
Do I need some init before it draws to the screen like selecting the memory address of the first pixel which I think am doing with a 0x4F07 vbe call.

Re: vesa checksum

Posted: Fri Aug 08, 2008 3:45 pm
by Laksen
Adek336 wrote: Do I need some init before it draws to the screen like selecting the memory address of the first pixel which I think am doing with a 0x4F07 vbe call.
I don't think that is needed but I don't actually know. You could try to set the start offset to 0 just to check.