Bernd Harries wrote:
>
> Hi all,
>
> > If you have multiple host/PCI bridges, I think you can have multiple
> > ISA memory spaces. Cfr. what happens on the PowerMacs with two
> > host/PCI bridges.
>
> Yes, similar to what I thought of. But which one is _the_ ISA space? I
> mean the one, accesses like
>
> request_isa_region(0xA0000,n,name)
>
> go to? I assume the subject of your Idea was to be able to let drivers
> work with 20 or 24 bit addresses. (?) And silently adding the base of
> the ISA area inside the functions to find the actual ISA area.
>
> Are you thinking of 2 sets of functions, one for each ISA space?
>
> Or was the Idea that accesses to ISA 0xA0000 are simply doing NOPs ?
IIRC from past discussions on linux-kernel, request_mem_region and
request_region are tied closely to the X86 port and not really intended
to be portable.
I think that having request_isa_region isn't clean on x86 -- there is no
hardware partition separating ISA and PCI memory space, so there
shouldn't be a request_isa_region in addition to request_mem_region.
Unfortunately following the non-portable nature of request_xxx to its
logical conclusion makes for some ugly ifdef-heavy code..
Jeff
-- Jeff Garzik | ILOVEYOU, Linux. Building 1024 | MandrakeSoft, Inc. |- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Mon May 15 2000 - 21:00:16 EST