chuq@chuq.com writes:
> On Thu, Dec 01, 2005 at 11:37:33PM +0900, ITOH Yasufumi wrote:
> > works, and I found that the HPMC occurs at
> > 
> > hp700/dev/sti_sgc.c:166:	devtype = bus_space_read_1(ca->ca_iot, romh, 3);
> 
> this is due to changes I made a while back to avoid abusing cf_unit.
> we've been adding STI revisions one by one to the list of ones that
> get the ROM pointer from page0 instead of the hpa, but it has become
> obvious that this isn't a great way to handle the issue.  the linux driver
> has some logic that searches for the STI ROM that looks much more

Hmm.

> robust.  to get around the problem in the short term, you can add
> your iodc_revision to the list in sti_sgc_getrom().

OK, it works for me by this change.

diff -uF^[a-zA-Z_][a-z 	A-Z0-9_]*([^;]*$ sti_sgc.c.orig sti_sgc.c
--- sti_sgc.c.orig	Tue May 10 17:05:59 2005
+++ sti_sgc.c	Fri Dec  2 08:03:17 2005
@@ -113,6 +113,7 @@ sti_sgc_getrom(struct confargs *ca)
 	case STI_GOPT5_REV:
 	case STI_GOPT6_REV:
 	case STI_GOPT7_REV:
+	case 0xc0:
 		/* these share the onboard's prom */
 		pagezero_cookie = hp700_pagezero_map();
 		rom = PAGE0->pd_resv2[1];

Thanks,
-- 
ITOH Yasufumi