------ =_NextPart_001_01BED9C5.D0012F40
Content-Type: text/plain
I work on a big-endian PowerPC box. A pci bus and a PC104 (ISA) is
available on this
system. I can try every device made for an intel PC. A lot of drivers
are broken in this configuration: On linux 2.2.9, the classic network
board PCI NE2000
don't work (very small changes are needed to fix it), 3Com Etherlink III
is also broken etc ...
Other works well: Matrox FB, adaptec SCSI...
The main problems on theses drivers are the lacks of little-endian to
big-endian
conversions. The problem can also be related to physical memory space
and
pci memory space in the case of bus master PCI devices.
These space are the same on a PC but are different on a PowerPC
architecture.
The linux kernel own a method to handle properly theses differences, but
a lot
of intel drivers bypass it.
Some works will be necessary to have all the PCI devices working
properly on all
differents linux platform. Having a little-endian architecture help a
little...
Patrick LERDA
------ =_NextPart_001_01BED9C5.D0012F40
Content-Type: text/html
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
I work on a big-endian PowerPC box. A pci bus and a PC104 (ISA) is available on this
system. I can try every device made for an intel PC. A lot of drivers
are broken in this configuration: On linux 2.2.9, the classic network board PCI NE2000
don't work (very small changes are needed to fix it), 3Com Etherlink III is also broken etc ...
Other works well: Matrox FB, adaptec SCSI...
The main problems on theses drivers are the lacks of little-endian to big-endian
conversions. The problem can also be related to physical memory space and
pci memory space in the case of bus master PCI devices.
These space are the same on a PC but are different on a PowerPC architecture.
The linux kernel own a method to handle properly theses differences, but a lot
of intel drivers bypass it.
Some works will be necessary to have all the PCI devices working properly on all
differents linux platform. Having a little-endian architecture help a little...
Patrick LERDA
------ =_NextPart_001_01BED9C5.D0012F40--
-
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/