Page 1 of 1
Quirks
Posted: Fri Oct 10, 2014 5:46 am
by elderK
Hey peeps,
Just playing with an idea here but would anyone find any use in a patch that adds options to Bochs so to enable "quirky" behavior?
I've been perusing the forums and studying Linux's source a little and a quirk that I can use as an example is when E820 wipes %ebx even though it's not signalling an error or the completion of the memory map.
I personally feel it'd be useful to be able to test against various quirks.
If enough people are interested, I'll do some work to add those options
Re: Quirks
Posted: Fri Oct 10, 2014 6:08 am
by max
I think it would be quite useful, for example to see how your kernel behaves in such situations. Voting for it
Re: Quirks
Posted: Fri Oct 10, 2014 6:41 am
by SpyderTL
Agreed.
But you may not need to patch Bochs itself. All you would really need is a modified BIOS. I'm pretty sure that bochs will let you swap out the BIOS in the configuration settings.
Good idea tho.
Re: Quirks
Posted: Fri Oct 10, 2014 7:09 am
by max
You could maybe even
get involved and add it as an activatable standard feature.
Re: Quirks
Posted: Fri Oct 10, 2014 7:21 am
by xenos
Re: Quirks
Posted: Sat Oct 11, 2014 12:50 am
by elderK
It's true, SpyderTL, that I don't necessarily need to make alterations to Bochs itself in order to have some kind of kirks mode.
But! I feel that doing so and presenting some kind of option to disable support for some BIOS call would be easier to use.
That and messing with SeaBIOS directly scares the hell out of me
XenOS, how far along is ReBochs?
~K
Re: Quirks
Posted: Sat Oct 11, 2014 12:57 am
by elderK
Actually, now that I think about it... would it not be simple just to implement a small proxy that hooks the interrupt that you want to mess with? TSRs and stuff did it all the time back in the day.
Like, create a vector that checks whether it's time to do something stupid otherwise just pass the call directly to the "real" E820?
Edit: Or rather, real int 0x15?
~K
Re: Quirks
Posted: Sat Oct 11, 2014 1:46 am
by xenos
elderK wrote:XenOS, how far along is ReBochs?
In fact, I don't know. I just remembered that bewing (one of the Bochs developers) was working on it, so I did a [url=forum.osdev.org/search.php?st=0&sk=t&sd=d&sr=posts&keywords=rebochs]search[/url], but the last post is already more than a year old.