There's also plenty of documentation on the north bridge as well. I should say that I suspect that there is not that much that you technically have to implement to make linux happy. I also suggest that you look at bochs to see what they do. to emulate all of this.
On Sun, Nov 23, 2008 at 11:06 PM, Gabe Black <[EMAIL PROTECTED]> wrote: > Last email tonight, I promise. I found documentation for Intel's ICH10 > chipset which appears to have all the information I need. At least with > 840 pages I'd certainly hope so :). > > Gabe > > Gabe Black wrote: >> And I just realized that my Intel system with an nvidia motherboard >> thinks it has an AMD IDE controller. Weird. >> >> Gabe >> >> Gabe Black wrote: >> >>> I think you're right. I looked at some of the entries under /proc on my >>> desktop (it's handy to actually work -in- x86 as well) and it looks like >>> the IO addresses used by the IDE controller are allocated to the AMD >>> south bridge which has a presence on the PCI bus. It looks like I'm >>> going to have to at least partially emulate a south bridge/north bridge >>> pair in order to get boot to progress. I think I saw something sort of >>> like documentation for this at one point, but I think it was only about >>> the on chip northbridge. If anybody (hint hint nudge nudge) can get AMD >>> or Intel to send me or point me to unfettered documentation on something >>> I can emulate that would be great. I'll see what I can find online. At >>> work I have pdfs for the hardware we're emulating, but there's no chance >>> I'm going to use that for M5 unless it happens to be intentionally >>> publicly available as well. >>> >>> Gabe >>> >>> nathan binkert wrote: >>> >>> >>>> Alpha's root bridge is the Tsunami stuff. You're going to need to add >>>> the equivalent for x86. I'm not sure about how the pci devices should >>>> be set up (other than the bios does do that for you if you want to use >>>> what it gives you), but you might want to fix the root bridge problem >>>> to see if that fixes the ide thing. >>>> >>>> Nate >>>> >>>> On Sun, Nov 23, 2008 at 9:11 PM, Gabe Black <[EMAIL PROTECTED]> wrote: >>>> >>>> >>>> >>>>> Does anybody out there have a lot of experience with IDE drives are >>>>> supposed to hook up on an x86 machine? Our controller model looks very >>>>> much like what x86 uses, except that it's set up as a PCI device. It >>>>> looks like the IDE interfaces are expected to be at particular >>>>> addresses, and that seems a lot more like ISA than PCI. Maybe the PCI >>>>> BAR registers are configured to put the IDE control registers in the >>>>> right places? Also, when the PCI subsystem of the kernel tries to >>>>> allocate "resources" for the controller, it fails each of four times. >>>>> Looking at the code, it seems that it expects things to already be set >>>>> up by the BIOS (it's first preference, and maybe how they get to be at >>>>> those legacy addresses), or it tries to allocate some IO/memory space by >>>>> staying within the bounds established as accessible by the configuration >>>>> registers of the parent (in this case root) PCI bus. It doesn't look >>>>> like we actually have one of those which is a little confusing. How does >>>>> Alpha get by with no root bridge? Anyways, I realize everybody is >>>>> probably WAY to busy to worry about this at the moment, but when you get >>>>> a chance a little explanation would be very helpful. >>>>> >>>>> Gabe >>>>> _______________________________________________ >>>>> m5-dev mailing list >>>>> m5-dev@m5sim.org >>>>> http://m5sim.org/mailman/listinfo/m5-dev >>>>> >>>>> >>>>> >>>>> >>>>> >>>> _______________________________________________ >>>> m5-dev mailing list >>>> m5-dev@m5sim.org >>>> http://m5sim.org/mailman/listinfo/m5-dev >>>> >>>> >>>> >>> _______________________________________________ >>> m5-dev mailing list >>> m5-dev@m5sim.org >>> http://m5sim.org/mailman/listinfo/m5-dev >>> >>> >> >> _______________________________________________ >> m5-dev mailing list >> m5-dev@m5sim.org >> http://m5sim.org/mailman/listinfo/m5-dev >> > > _______________________________________________ > m5-dev mailing list > m5-dev@m5sim.org > http://m5sim.org/mailman/listinfo/m5-dev > > _______________________________________________ m5-dev mailing list m5-dev@m5sim.org http://m5sim.org/mailman/listinfo/m5-dev
