Re: [patch 2.5] VGA IO on systems with multiple PCI IO domains

[Ivan Kokshaysky <ink@jurassic.park.msu.ru>]

On Mon, Jan 27, 2003 at 10:46:45AM +0100, Martin Mares wrote:
> > What about some kind of ioport_remap() that would take a pci_bus and an
> > port range as arguments ? If pci_bus is NULL, that would match a
> > "legacy" ISA bus (non-PCI machine or default ISA bus for machines where
> > that makes sense).
> > 
> > What do you think ?
> 
> Looks good, but maybe we should use some other functions than iob() et al.
> to do I/O on the remapped addresses.

What about this (what we'd have on alpha):

int
legacy_ioport_remap(struct resource *res)
{
	switch (res->start) {
	case 0x3c0:	/* VGA */
		res->start += pci_vga_hose->io_space->start;
		res->end += pci_vga_hose->io_space->start;
	case ???
		...
	default:
		return -ENODEV;
	}
	return request_resource(pci_vga_hose->io_space, res);
}

void
legacy_ioport_unmap(struct resource *res)
{
	release_resource(res);
}

Then vgacon.c would be changed like this:

...
-	request_resource(&ioport_resource, &vga_console_resource);
+	if (legacy_ioport_remap(&vga_console_resource) < 0)
+		goto failure;
...

And all in/out port calls would use respective resource.start+offset:
...
-	outb_p(6, 0x3ce)
+	outb_p(6, vga_console_resource.start + 0xe);

No need for other special IO functions then.

Ivan.