Re: Linux 2.4.19-rc1-ac3

From: Peter Osterlund (petero2@telia.com)
Date: Sun Jul 14 2002 - 10:19:54 EST


Dominik Brodowski <devel@brodo.de> writes:

> On Sun, Jul 14, 2002 at 03:03:08PM +0200, Alan Cox wrote:
> > > 3. The cpu voltage is automatically reduced when the frequency is
> > > reduced.
> >
> > True for some x86 processors, either automatically, or on some
> > controlled by us.
>
> The p4-clockmod driver you seem to using does not scale the voltage.
> In case you own a P4-M and a ICH2-M or ICH3-M southbridge: the
> speedstep.c driver in the 2.5.-cpufreq patchset (backport to 2.4.
> will be available soon) should do the job and adjust the processor
> voltage.

I tried speedstep but it didn't work because of this check in
speedstep_detect_processor:

                /* Intel Pentium 4 Mobile P4-M */
                if (c->x86_model != 2)
                        return 0;

                if (c->x86_mask != 4)
                        return 0; /* all those seem to support Enhanced
                                         SpeedStep */

My cpu has model == 1 and mask == 2. I deleted that code just to see
what would happen, but then I tripped over this code in
pentium4_get_frequency:

        /* Don't trust unseen values yet, except in the MHz field
         */
        if (msr_hi || ((msr_lo & 0x00FFFFFF) != 0x300511)) {
                printk(KERN_INFO "cpufreq: Due to incomplete documentation, please send a mail to devel@brodo.de\n");
                printk(KERN_INFO "with a dmesg of a boot while on ac-power, and one of a boot on battery-power.\n");
                printk(KERN_INFO "Thanks in advance.\n");
                return 0;
        }

I removed the return statement, and then got:

        apm: BIOS version 1.2 Flags 0x03 (Driver version 1.16)
        cpufreq: Intel(R) SpeedStep(TM) support $Revision: 1.34 $
        model:1 mask:2
        cpufreq: chipset 0x3 - processor 0x3
        cpufreq: activating SpeedStep (TM) registers
        cpufreq: read at pmbase 0x1000 + 0x50 returned 0x1
        cpufreq: P4 - MSR_EBC_FREQUENCY_ID: 0x8a01fe00 0x0
        cpufreq: Due to incomplete documentation, please send a mail to devel@brodo.de
        with a dmesg of a boot while on ac-power, and one of a boot on battery-power.
        Thanks in advance.
        cpufreq: read at pmbase 0x1000 + 0x50 returned 0x1
        cpufreq: writing 0x0 to pmbase 0x1000 + 0x50
        cpufreq: read at pmbase 0x1000 + 0x50 returned 0x0
        cpufreq: change to 0 MHz succeded
        cpufreq: read at pmbase 0x1000 + 0x50 returned 0x0
        cpufreq: P4 - MSR_EBC_FREQUENCY_ID: 0x8a01fe00 0x0
        cpufreq: Due to incomplete documentation, please send a mail to devel@brodo.de
        with a dmesg of a boot while on ac-power, and one of a boot on battery-power.
        Thanks in advance.
        cpufreq: read at pmbase 0x1000 + 0x50 returned 0x0
        cpufreq: read at pmbase 0x1000 + 0x50 returned 0x0
        cpufreq: currently at low speed setting - 13800 MHz
        CPU clock: 13800.000 MHz (13800.000-13800.000 MHz)
        Starting kswapd

There was no difference in dmesg output when booting on battery power
instead of ac power.

So what can I do to make speedstep work? According to the notebook
manual, speedstep is supported on this computer.

-- 
Peter Osterlund - petero2@telia.com
http://w1.894.telia.com/~u89404340
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Mon Jul 15 2002 - 22:00:27 EST