[lkml]   [1999]   [Dec]   [18]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [patch] read[bwl] and ioremap problem
>>>>> "Linus" == Linus Torvalds <> writes:

Linus> On 18 Dec 1999, Jes Sorensen wrote:
>> This patch is clearly wrong, since the Mac doesn't have IDE on an
>> ISA bus, thus it shouldn't use those macros. However it's not on a
>> PCI bus either so it shouldn't use readb/writeb in the first place
>> ;-)

Linus> readb/writeb do not have to be PCI.

Linus> Any bus that looks from a sw perspective enough like it might
Linus> as well be considered "PCI-like". Some of them may or may not
Linus> need ioremap(). It becomes a problem only if there is a machine
Linus> that has BOTH a real PCI bus _and_ something else, but with an
Linus> old m68k mac that doesn't end up being a problem.

Agreed, however we have m68k boxes that have an ISA bus and another
bus, where the native bus behaves pretty much like PCI in a PC and the
ISA bus requiring special treatment on those machines, with weird byte
order issues, special areas to read from than write from etc. I
therefore prefer to keep the native busses from using the inb/outb and
readb/writeb macros and reserve those macros to to do the magic for
ISA and PCI.

The Macintosh case is not problematic since I don't think there are
m68k based with anything but NUBUS.


To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
Please read the FAQ at

 \ /
  Last update: 2005-03-22 13:55    [from the cache]
©2003-2014 Jasper Spaans. hosted at Digital Ocean