I don't know if I'm hunting a bug or I have to add support for my
hardware somewhere.
I succeded in making C2 work adjusting the DSDT table, I'm uploading the
new one on acpi.sf.net under Asus/L3100D
Now for cpufreq: I'm using the powernow-k7 driver, since the acpi
P-state is not supported.
Whenever I try to change the governor by echoing "powersave" in
/sys/devices/system/cpu/cpu0/cpufreq/scaling_governor everything stops
and a hard reset is required. It seemes to fall in some tight loop since
after a few seconds the fan goes on as if there is heavy processor
usage.
With some printk I have verified that it halts inside wrmsrl() at line
199, file powernow-k7.c (2.6.0-test4).
The call trace is:
powernow_target/change_speed/change_FID/wrmsrl
For what is my knowledge (very, very little) the frequency passed to
powernow_target is right, but I can't say nothing on all other
parameters...
Full dmesg and .config available here[2] and here[3], /proc/cpuinfo
and /proc/acpi/processor/CPU0/{power,info} follows.
/proc/cpuinfo
processor : 0
vendor_id : AuthenticAMD
cpu family : 6
model : 8
model name : mobile AMD Athlon(tm) XP-M 2000+
stepping : 1
cpu MHz : 1673.782
cache size : 256 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 mmx fxsr sse syscall mp mmxext 3dnowext 3dnow
bogomips : 3309.56
/proc/acpi/processor/CPU0/power
active state: C2
default state: C1
bus master activity: 00000000
states:
C1: promotion[C2] demotion[--] latency[000] usage[00653320]
*C2: promotion[--] demotion[C1] latency[090] usage[04958506]
C3: <not supported>
/proc/acpi/processor/CPU0/info
processor id: 0
acpi id: 1
bus mastering control: no
power management: yes
throttling control: yes
performance management: no
limit interface: yes
Using the acpi p state driver I get a message saying something about
'unsupported address space' (see [1] for details) and the machine
stops working also by echoing 'performance' (that is already there) in
sysfs.
I don't know what to do next, someone here can give me any hint ?
Thanks for any help on this matter.
[1] http://www.ussg.lu.edu/hypermail/linux/kernel/0308.2/1892.html
[2] http://digilander.libero.it/webvenza/download/kpatches/dmesg-cpufreq
[3] http://digilander.libero.it/webvenza/download/kpatches/config-cpufreq
--
----------------------------------------
Daniele Venzano
Web: http://digilander.iol.it/webvenza/
On Sun, Aug 24, 2003 at 06:48:29PM +0200, Daniele Venzano wrote:
> I don't know if I'm hunting a bug or I have to add support for my
> hardware somewhere.
> I succeded in making C2 work adjusting the DSDT table, I'm uploading the
> new one on acpi.sf.net under Asus/L3100D
>
> Now for cpufreq: I'm using the powernow-k7 driver, since the acpi
> P-state is not supported.
> Whenever I try to change the governor by echoing "powersave" in
> /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor everything stops
> and a hard reset is required. It seemes to fall in some tight loop since
> after a few seconds the fan goes on as if there is heavy processor
> usage.
>
> With some printk I have verified that it halts inside wrmsrl() at line
> 199, file powernow-k7.c (2.6.0-test4).
> The call trace is:
> powernow_target/change_speed/change_FID/wrmsrl
>
> For what is my knowledge (very, very little) the frequency passed to
> powernow_target is right, but I can't say nothing on all other
> parameters...
>
This is a known issue (at least for me :). When the difference
between 2 vid and/or 2 fid are too big, that hang.
For now, as a quick (and mostly dirty though) solution is to go
with the userspace governor only and to go to the lowest frequency with
one or two step in-between.
I'm also wondering if those athlons have the same kind of stuff than
the opteron (Dave)?
Cheers,
--
Ducrot Bruno
-- Which is worse: ignorance or apathy?
-- Don't know. Don't care.
Hi!
> This is a known issue (at least for me :). When the difference
> between 2 vid and/or 2 fid are too big, that hang.
>
> For now, as a quick (and mostly dirty though) solution is to go
> with the userspace governor only and to go to the lowest frequency with
> one or two step in-between.
>
> I'm also wondering if those athlons have the same kind of stuff than
> the opteron (Dave)?
k8 has similar restriction to small steps, only.
Pavel
--
Pavel
Written on sharp zaurus, because my Velo1 broke. If you have Velo you don't need...
On Mon, Sep 01, 2003 at 10:09:42PM +0200, Pavel Machek wrote:
> > I'm also wondering if those athlons have the same kind of stuff than
> > the opteron (Dave)?
>
> k8 has similar restriction to small steps, only.
The k8 docs actually specify this though, the k7 ones don't iirc.
Dave
--
Dave Jones http://www.codemonkey.org.uk
On Tue, Sep 02, 2003 at 01:56:25PM +0100, Dave Jones wrote:
> On Mon, Sep 01, 2003 at 10:09:42PM +0200, Pavel Machek wrote:
> > > I'm also wondering if those athlons have the same kind of stuff than
> > > the opteron (Dave)?
> >
> > k8 has similar restriction to small steps, only.
>
> The k8 docs actually specify this though, the k7 ones don't iirc.
>
Hem, some of vid/fid combinations cant be used . But
there is no verifications of such in powernow-k7 though.
Since there is a need to change only one at a time...
--
Ducrot Bruno
-- Which is worse: ignorance or apathy?
-- Don't know. Don't care.