David> Byte swapping happens transparently on Sparc for PCI devices,
David> the normal read{b,w,l}() etc. interfaces are used for
David> everything and it just works.
David> It works because we architected Sparc to map PCI areas in the
David> proper endianness.
David> The end result is that you should not need "big endian"
David> supporting PCI devices, and in fact I consider such devices
David> silly. Little endian is the native PCI byte ordering, and any
David> worthwhile architecture we care about has hardware facilities
David> to make them accessible at zero cost (either via MMU mappings
David> of device areas, or endianness attributes in the load/store
David> instructions themselves).
How do you expect to handle say video memory on 24 bit cards where it
doesn't fill up and use 32 bits per pixel using this scheme on big
endian machines?
If the hardware can swap your data it is still cheaper on some systems
to let it do so.
Jes
-
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/