Subject: Re: PCI Physical Address allocation (i386)
To: Hayakawa Koichi <>
From: Michael Richardson <>
List: tech-kern
Date: 01/19/2000 13:37:49
>>>>> "Hayakawa" == Hayakawa Koichi <> writes:
    Hayakawa> Hello,

    Hayakawa> From: Michael Richardson <>
    Hayakawa> Subject: PCI Physical Address allocation (i386)
    Hayakawa> Date: Tue, 18 Jan 2000 21:26:53 -0500
    Hayakawa> Message-ID: <>

    >> I want allocate a bunch of physical addresses on a PCI bus.
    Hayakawa> [snip]
    >> The question is, how can I, as a NetBSD driver writer, find out what this
    >> physical address map looks like? Or better yet, just have the PCI BIOS
    >> allocate me a chunk of address space.

    Hayakawa> I think there are no perfect solutions.

  Okay... are there any imperfect solutions?

    Hayakawa> According to CardBus-PCI bridge, the attach routine for the
    Hayakawa> bridge uses config_defer() in order to obtain physical
    Hayakawa> memory space after all PCI devices are attached.  This
    Hayakawa> method prevents the illigal use of bus space kept for other
    Hayakawa> PCI devices, but it has no effect for the deives on the
    Hayakawa> other buses.  Please refer to the function pccbbattach()
    Hayakawa> defined at sys/dev/pci/pccbb.c.

  Yes, okay, I worked through this. Is there a function that I'm missing
that would live in dev/pci or in arch/foo/pci that would permit simply return
a free extent from the point of view of PCI...

    Hayakawa> Rbus is used for all ports, because some ports (i.e. macppc)
    Hayakawa> do not manage their bus space.

  What does this mean? You mean their BIOS does not do it?

    >> Advice?

    Hayakawa> Is this what you need?

  I guess I want to know if allocation of a PCI address for my use is 
even possible...

   :!mcr!:            |  Solidum Systems Corporation,
   Michael Richardson |For a better connected world,where data flows faster<tm>