Subject: Re: NEC RiscServer 2200
To: bob meader <>
From: Noriyuki Soda <>
List: port-arc
Date: 08/02/2000 22:40:26
>>>>> On Tue, 01 Aug 2000 14:14:58 -0700,
	bob meader <> said:

bob> I am curious about the statement above when you say it boots
bob> do you mean it successfully mounts a remote nfs root file system
bob> ie you a get a NetBSD login prompt ?

>>>>> On Tue, 01 Aug 2000 21:13:01 -0400,
	Ross Alexander <> said:

ross> No, I mean it boots and I get a "root device:" prompt.

NetBSD/arc kernel reached the "root device:" prompt at April with
Bob's machine. But since interrupt related code didn't work,
didn't reach single user shell.
(and perhaps the problem may be fixed now, but I'm not sure.)

bob> The reason I am asking is I looked at the 2000-06-17 source code
bob> listing ( We will assume for now that is the source for above binary)
bob> and the code for the clock chip is wrong .  From my hazy memory,
bob> (it's been 6 months since I last looked )the code for the NFS client
bob> needs to read the clock chip to get the current date to check
bob> that file system doesn't have a 'stale' date..

bob> Any way the file that needs to be fixed is 'sys\arch\arc\arc\clock_mc.c'
bob> the clock used in NEC RiscStation 2200 is like 'ACER_PICA_61'
bob> not like 'NEC_R94' as currently coded...

Hmm, that means the patch attached below is needed?
I've made new kernel with this patch, and put it on the following URL:

Could anyone who have RISCserver 2200 test this?

--- arch/arc/arc/clock_mc.c-	Fri Jun  9 14:41:57 2000
+++ arch/arc/arc/clock_mc.c	Wed Aug  2 16:30:12 2000
@@ -130,6 +130,7 @@
         case ACER_PICA_61:
 	case MAGNUM:
+	case NEC_R96:
 		csc->sc_init = mcclock_init_pica;
 		csc->sc_data = &mcclockdata_pica;
 		mc146818_write(csc, MC_REGB, MC_REGB_BINARY | MC_REGB_24HR);
@@ -138,7 +139,6 @@
 	case NEC_R94:
 	case NEC_RAx94:
 	case NEC_RD94:
-	case NEC_R96:
 		csc->sc_init = mcclock_init_rd94;
 		csc->sc_data = &mcclockdata_rd94;
 		mc146818_write(csc, MC_REGB, MC_REGB_BINARY | MC_REGB_24HR);