[PATCH] BCM1480: Init pci controller io_map_base

[PATCH] BCM1480: Init pci controller io_map_base

[Thomas Bogendoerfer]

BCM1480: Init pci controller io_map_base

Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
---

 arch/mips/pci/pci-bcm1480.c |    5 +++--
 1 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/arch/mips/pci/pci-bcm1480.c b/arch/mips/pci/pci-bcm1480.c
index 30ed361..ab68c43 100644
--- a/arch/mips/pci/pci-bcm1480.c
+++ b/arch/mips/pci/pci-bcm1480.c
@@ -249,8 +249,9 @@ static int __init bcm1480_pcibios_init(void)
 	 * XXX ehs: Should this happen in PCI Device mode?
 	 */
 
-	set_io_port_base((unsigned long)
-		ioremap(A_BCM1480_PHYS_PCI_IO_MATCH_BYTES, 65536));
+	bcm1480_controller.io_map_base = (unsigned long)
+		ioremap(A_BCM1480_PHYS_PCI_IO_MATCH_BYTES, 65536);
+	set_io_port_base(bcm1480_controller.io_map_base);
 	isa_slot_offset = (unsigned long)
 		ioremap(A_BCM1480_PHYS_PCI_MEM_MATCH_BYTES, 1024*1024);
 

Re: [PATCH] BCM1480: Init pci controller io_map_base

[Ralf Baechle]

On Sat, Mar 08, 2008 at 07:51:55PM +0100, Thomas Bogendoerfer wrote:

> BCM1480: Init pci controller io_map_base

Thanks, applied.  But this patch only solves the problem for the BCM1480's
native PCI bus.  Support for PCI behind HT still needs to be fixed and I'm
not quite sure where the ports of such devices and busses are getting
mapped to.

  Ralf

Re: [PATCH] BCM1480: Init pci controller io_map_base

[Maciej W. Rozycki]

On Mon, 10 Mar 2008, Ralf Baechle wrote:

> Thanks, applied.  But this patch only solves the problem for the BCM1480's
> native PCI bus.  Support for PCI behind HT still needs to be fixed and I'm
> not quite sure where the ports of such devices and busses are getting
> mapped to.

 Hmm, A_BCM1480_PHYS_HT_IO_MATCH_BYTES?

  Maciej

Re: [PATCH] BCM1480: Init pci controller io_map_base

[Thomas Bogendoerfer]

On Wed, Mar 19, 2008 at 11:12:23AM +0000, Maciej W. Rozycki wrote:
> On Mon, 10 Mar 2008, Ralf Baechle wrote:
> 
> > Thanks, applied.  But this patch only solves the problem for the BCM1480's
> > native PCI bus.  Support for PCI behind HT still needs to be fixed and I'm
> > not quite sure where the ports of such devices and busses are getting
> > mapped to.
> 
>  Hmm, A_BCM1480_PHYS_HT_IO_MATCH_BYTES?

we could switch to IO_MATCH_BITS, but then we need to disable software
swapping and use address mangling. I personaly prefer to do swapping
in software.

Thomas.

-- 
Crap can work. Given enough thrust pigs will fly, but it's not necessary a
good idea.                                                [ RFC1925, 2.3 ]