On Tue, 26 Jun 2007, Daniel de Kok wrote:

> Could you try the attached patch against sys/arch/i386/i386/identcpu.c?
> This seems to be a bug in the vendor documentation.

Thanks Daniel for your help.

I applied the patch and compiles the kernel... No changes in the dmesg,
polaris# diff dmesg.debug dmesg.patch
7c7
< NetBSD 4.99.20 (POLARIS) #9: Tue Jun 26 16:22:19 CDT 2007
---
> NetBSD 4.99.20 (POLARIS) #10: Tue Jun 26 19:01:28 CDT 2007
17c17
< cpu0: VIA C7 M (686-class), 799.97 MHz, id 0x6a9
---
> cpu0: VIA C7 M (686-class), 800.00 MHz, id 0x6a9
82c82
< acpitz0: critical 80.0C passive 29.0C
---
> acpitz0: critical 80.0C passive 45.0C

I am checking the function via_cpu_probe too. It seems such
function never runs but I can be wrong. I can not see the debug
messages in the dmesg... printf and aprint_* do not works for me.

polaris# diff /sys/arch/i386/i386/identcpu.c.orig /sys/arch/i386/i386/identcpu.c
483c483
<                               "C3 Nehemiah", 0, 0, 0, 0, 0, 0,
---
>                               "C3 Nehemiah", "C7 M", 0, 0, 0, 0, 0,
644c644
< via_cpu_probe(struct cpu_info *ci)
---
> via_cpu_probe(struct cpu_info *ci) /* CHECK THIS FUNCTION */
653a654
>       aprint_error("MYTEST: running via_cpu_probe\n");
663a665
>       aprint_verbose("mod=%d step=%d lfunc=%x\n", model, stepping, lfunc);
669c671
<               if (lfunc == 0xc0000001) {
---
>               if (lfunc >= 0xc0000001) {
polaris#

Thanks a lot

Heron Gallegos
Saltillo Coahuila Mexico