Re: ncr53c8xx.c and 2.1.49

Gerard Roudier (groudier@club-internet.fr)
Sun, 17 Aug 1997 18:01:05 +0200 (MET DST)


On Tue, 12 Aug 1997, a sun wrote:

>
> following is a small fix for the ncr53c8xx driver. there's also a
> small cosmetic printk bug on line 8794, but i'm not sure what's
> supposed to be there. here's the line:
> printf("ncr53c8xx_setup: unexpected boot option '%.*s' ignored\n",
> pc-cur+1, cur);

I've directly written this expression and it seemed to work as expected.
Did I magically use some extra-features of printk?

> --- linux/drivers/scsi/ncr53c8xx.c.save Tue Aug 12 09:08:46 1997
> +++ linux/drivers/scsi/ncr53c8xx.c Tue Aug 12 09:08:59 1997
> @@ -9283,7 +9283,7 @@
> */
> #ifdef NCR_IOMAPPED
> request_region(io_port, 128, "ncr53c8xx");
> - device->slot.port = ioport;
> + device->slot.port = io_port;
> #else
> device->slot.reg = (struct ncr_reg *) remap_pci_mem((ulong) base, 128);
> if (!device->slot.reg)

Thanks for the fix. Except for Alpha, you should use memory mapped IOs.
It is the recommended PCI IO method. Normal IOs are not supported by
all processors and look like some historical compatibility (ISA, ...).

Gerard.