On Wed, 11 Oct 2000, Joel Votaw wrote:
> I dug around on the mailing list archives for port-netbsd and dug up the
> following patch.  If you're having problems with Promise IDE controllers
> on recent kernels on Alphas, this should work.  (I have not yet tried it
> myself, but will when I get home tonight.)

I just got a Promise Ultra100, and of course, had the same problem...
here's the patch, updated so that it'll apply to 1.5_ALPHA2/1.5_BETA.
It's working great for me so far (with a 45GB IBM DTLA (75GXP)).

--- /usr/src/sys/dev/pci/pciide.c	Thu Oct  5 05:40:37 2000
+++ pciide.c	Wed Oct 18 19:50:42 2000
@@ -3063,12 +3063,29 @@
 	int channel;
 	pcireg_t interface, st, mode;
 	bus_size_t cmdsize, ctlsize;
+	bus_addr_t addr, forceaddr = 0x8000;
+	int i;
 
 	st = pci_conf_read(sc->sc_pc, sc->sc_tag, PDC2xx_STATE);
 	WDCDEBUG_PRINT(("pdc202xx_setup_chip: controller state 0x%x\n", st),
 	    DEBUG_PROBE);
 	if (pciide_chipen(sc, pa) == 0)
 		return;
+
+	/* XXX Force IO addresses if BIOS doesn't set them up for me */
+	for (i = 0x10; i <= 0x20; i += 4) {
+		addr = pci_conf_read (sc->sc_pc, sc->sc_tag, i);
+		printf("%s: Changing IO from 0x%lx to 0x%lx\n",
+		    sc->sc_wdcdev.sc_dev.dv_xname,
+		    addr & PCI_MAPREG_IO_ADDR_MASK,
+		    /* addr & 0xffff & PCI_MAPREG_IO_ADDR_MASK */
+		    forceaddr);
+
+		pci_conf_write (sc->sc_pc, sc->sc_tag,
+		    i, /* addr & 0xffff & PCI_MAPREG_IO_ADDR_MASK */
+		    forceaddr | PCI_MAPREG_TYPE_IO);
+		forceaddr += 16;
+	}
 
 	/* turn off  RAID mode */
 	st &= ~PDC2xx_STATE_IDERAID;