Page 1 of 3
My BSOD
Posted: Wed Feb 21, 2007 10:59 am
by AJ
Hi All,
Just wanted to share my nice new BSOD with you! This is after a deliberate page fault at 0xBFFFFFFF (just before my kernel). It takes a Windows user to get excited about this kind of thing!
I think it's more useful than the M$ attempt!
Cheers,
Adam
Posted: Wed Feb 21, 2007 11:53 am
by os64dev
what is usefull about a BSOD
Posted: Wed Feb 21, 2007 1:56 pm
by AJ
Absolutely nothing, which is why it's in General Ramblings! I'd just completed loads of libc-type functions and was just a little over-happy that they worked properly
. I'll keep on taking the tablets....
Posted: Wed Feb 21, 2007 2:31 pm
by Alboin
Grasshopper, you should have a configurable BSOD. So I can have a RSOD and a GSOD. Then, and only then, will your xSOD be truly awesome.
Posted: Thu Feb 22, 2007 9:47 am
by GLneo
Mine is going to be a PSOD, it will have a plasma effect in the background
Posted: Fri Feb 23, 2007 5:31 pm
by pcmattman
I think the best BSODs are the ones that actually tell you what on earth went wrong, instead of giving some cryptic message...
Also, why does everyone default to BSODs? Why not a PSOD? Or even a RSOD?
Alboin wrote:
Grasshopper, you should have a configurable BSOD. So I can have a RSOD and a GSOD. Then, and only then, will your xSOD be truly awesome.
I like your thinking, I'm inspired now...
Posted: Fri Feb 23, 2007 7:15 pm
by GLneo
because the blue is suppose to calm you, but you could always make a flashing flaming SOD with KoRn as background music
Posted: Sat Feb 24, 2007 3:08 am
by pcmattman
No, that's something only the pig-headed freaks over at Apple would try, and even then they might have the common sense (
) to not do it...
Posted: Sat Feb 24, 2007 12:53 pm
by earlz
many times BSOD code actually fails. This isn't just with hardware errors though, what about a kernel/interrupt stack overflow? you would triple fault in an instant...
another thing, what if your page tables or GDT get corrupted? then you make a write to memory and your crashed...
I think a trully fatal BSOD should load a backup GDT(already written in memory and it should be at a constant address that you shouldn't accidentally overwrite..
then you do something load GS with a flat segment, you should then save all registers in some memory area(including segment registers other than GS) and finally, load CS, DS and such and display dump data..
note this is all theoretical
Posted: Sat Feb 24, 2007 2:19 pm
by inflater
I've had some test BSOD lying in PortixOS, but it was "Black Screen of Death". Real mode hasnt much exceptions, no page fault, etc. just a divide overflow, bound errors etc. So I am having a NSOD - No Screen of Death
- INT 00, generated by CPU, points at HLT, INT 06, points at HLT too.
Posted: Wed Feb 28, 2007 9:56 pm
by Andrew275
Here's my RSOD, nothing too fancy. I try to make it somewhat friendly (aside from all the debugging info of course).
Posted: Fri Mar 02, 2007 12:53 am
by Brendan
Hi,
Andrew275 wrote:Here's my RSOD, nothing too fancy. I try to make it somewhat friendly (aside from all the debugging info of course).
I really don't see how people can debug their software without CR1 being displayed in the crash report...
Cheers,
Brendan
Posted: Fri Mar 02, 2007 2:21 am
by Andrew275
Brendan wrote:Hi,
I really don't see how people can debug their software without CR1 being displayed in the crash report...
Cheers,
Brendan
It's pretty difficult, but I manage.
Posted: Fri Mar 02, 2007 3:54 am
by AJ
ok!
. I only put it in becaus it looked like something was missing otherwise!
Posted: Tue Mar 06, 2007 12:21 pm
by oscoder
pcmattman wrote:No, that's something only the pig-headed freaks over at Apple would try, and even then they might have the common sense (
) to not do it...
Shouldn't that be 'pig-headed
phreaks'?