Adrian> But with multiple PCI buses, the base_address cannot be a
Adrian> simple bus address, because the PCI bus number must be coded
Adrian> in there. For example, on alpha the PCI_HANDLE from
Adrian> include/asm-alpha/pci.h is or'ed into the base_address.
Adrian> As far as I can tell, the base_address[] is always correct for
Adrian> ioremap. But are the following two tricks valid:
Thats my assumption and at least thats what I have been telling people
so far ;-)
Adrian> 1) Using the contents of base_address[] as an offset when
Adrian> mmapping /dev/mem from userspace?
I would probably go for using virt_to_bus(ioremap(base_address[]))
here.
Adrian> 2) Using the contents of base_address[] as a target address
Adrian> for peer to peer DMA on PCI?
Hmmm, interesting, I would say same answer as for 1).
By using this approach it should be safe for the boot sequence code to
mangle base_address[] when you have multiple PCI busses and it will
still be possible for ioremap() to do the right thing.
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/