Re: 715/50 experience

On Fri, 17 Oct 2003, Otto Moerbeek wrote:

> > sti0: init,0xc8566000(86000000, 0x4af910, 0x4af918, 0xc012412c)
> > panic: trap: uvm_fault(0x40df98, f4380004, 0, 1): 14
> 
> The fault address is f4380004, which is within the last region. So my
> hypothesis is that the init code accesses the last region, which isn't
> mapped correctly for some reason.
> 

With help from miod@ and mickey@ I now have a working sti(4). The patch 
below has the effect that all regions of the sti get mapped individually. 

Also inluded is a patch to make the osiop driver match fail on the 715/50.  
Manually disabling is no longer needed. This has been taken from NetBSD.

I'd appreciate it if people could take a look at both patches.

	-Otto

Index: dev/sti_sgc.c
===================================================================
RCS file: /cvs/src/sys/arch/hppa/dev/sti_sgc.c,v
retrieving revision 1.17
diff -u -r1.17 sti_sgc.c
--- dev/sti_sgc.c	25 Sep 2003 22:17:40 -0000	1.17
+++ dev/sti_sgc.c	26 Oct 2003 12:54:30 -0000
@@ -55,7 +55,7 @@
 
 #include <hppa/dev/cpudevs.h>
 
-#define	STI_MEMSIZE	0x2000000
+#define	STI_MEMSIZE	0x10000
 #define	STI_ROMSIZE	0x8000
 #define	STI_ID_FDDI	0x280b31af	/* Medusa FDDI ROM id */
 
Index: gsc/osiop_gsc.c
===================================================================
RCS file: /cvs/src/sys/arch/hppa/gsc/osiop_gsc.c,v
retrieving revision 1.7
diff -u -r1.7 osiop_gsc.c
--- gsc/osiop_gsc.c	7 Aug 2003 19:47:33 -0000	1.7
+++ gsc/osiop_gsc.c	26 Oct 2003 12:54:30 -0000
@@ -98,8 +98,7 @@
 	struct gsc_attach_args *ga = aux;
 
 	if (ga->ga_type.iodc_type != HPPA_TYPE_FIO ||
-	    (ga->ga_type.iodc_sv_model != HPPA_FIO_GSCSI &&
-	     ga->ga_type.iodc_sv_model != HPPA_FIO_SCSI))
+	    ga->ga_type.iodc_sv_model != HPPA_FIO_GSCSI)
 		return 0;
 
 	return 1;