[src/trunk]: src/sys/arch/x86/x86 Some core i7 CPUs report model 0c. In this ...

[jym <jym%NetBSD.org@localhost>]

details:   https://anonhg.NetBSD.org/src/rev/453129f2c911
branches:  trunk
changeset: 756935:453129f2c911
user:      jym <jym%NetBSD.org@localhost>
date:      Sun Aug 08 00:48:46 2010 +0000

description:
Some core i7 CPUs report model 0c. In this case, check for the extended
model value.

Required to avoid faulting on rdmsr(MSR_FSB_FREQ) early during boot.

Will ask for a pull-up. This affects GENERIC, and most likely, install iso
too.

XXX quick hack. Obtaining FSB through ACPI should be cleaner.

diffstat:

 sys/arch/x86/x86/intel_busclock.c |  20 ++++++++++++++++----
 1 files changed, 16 insertions(+), 4 deletions(-)

diffs (48 lines):

diff -r b8c06de33f6a -r 453129f2c911 sys/arch/x86/x86/intel_busclock.c
--- a/sys/arch/x86/x86/intel_busclock.c Sun Aug 08 00:44:38 2010 +0000
+++ b/sys/arch/x86/x86/intel_busclock.c Sun Aug 08 00:48:46 2010 +0000
@@ -1,4 +1,4 @@
-/*     $NetBSD: intel_busclock.c,v 1.10 2010/05/23 19:00:28 christos Exp $     */
+/*     $NetBSD: intel_busclock.c,v 1.11 2010/08/08 00:48:46 jym Exp $  */
 
 /*-
  * Copyright (c) 1999, 2000, 2001 The NetBSD Foundation, Inc.
@@ -30,7 +30,7 @@
  */
 
 #include <sys/cdefs.h>
-__KERNEL_RCSID(0, "$NetBSD: intel_busclock.c,v 1.10 2010/05/23 19:00:28 christos Exp $");
+__KERNEL_RCSID(0, "$NetBSD: intel_busclock.c,v 1.11 2010/08/08 00:48:46 jym Exp $");
 
 #include <sys/param.h>
 #include <sys/systm.h>
@@ -97,7 +97,19 @@
        case 0x9: /* Pentium M (130 nm, Banias) */
                bus_clock = 10000;
                break;
-       case 0xc: /* Atom, model 1 */
+       case 0xc: /* Core i7, Atom, model 1 */
+               /*
+                * XXX (See also case 0xe)
+                * Some core i7 CPUs can report model 0xc.
+                * Newer CPUs will GP when attemping to access MSR_FSB_FREQ.
+                * In the long-term, use ACPI instead of all this.
+                */
+               switch (CPUID2EXTMODEL(ci->ci_signature)) {
+               case 0x2:
+                       aprint_debug("%s: unable to determine bus speed",
+                           device_xname(ci->ci_dev));
+                       goto print_msr;
+               }
                msr = rdmsr(MSR_FSB_FREQ);
                bus = (msr >> 0) & 0x7;
                switch (bus) {
@@ -128,7 +140,7 @@
                break;
        case 0xe: /* Core Duo/Solo */
                /*
-                * XXX
+                * XXX (See also case 0xc)
                 * Newer CPUs will GP when attemping to access MSR_FSB_FREQ.
                 * In the long-term, use ACPI instead of all this.
                 */