Gentoo Archives: gentoo-amd64

From: "Ângelo Miguel Arrifano" <miknix@×××××.com>
To: gentoo-amd64@l.g.o
Subject: Re: [gentoo-amd64] Turion64 X2 k8temp temperatures.
Date: Sat, 03 Nov 2007 21:02:13
Message-Id: 20071103210751.e8a1ed8b.miknix@gmail.com
In Reply to: Re: [gentoo-amd64] Turion64 X2 k8temp temperatures. by Beso
1 On Sat, 3 Nov 2007 20:38:16 +0100
2 Beso <givemesugarr@×××××.com> wrote:
3
4 >> This is a laptop with a tiny fan and 53º C while processor is idle is
5 >> understandable since cooler is on passive mode
6 >> most of time. But 73º C looks way too much for me.
7 > this for me seems the thermal zone, and not the core temperature...
8 No, they are core temperatures:
9
10 #ls /sys/bus/pci/drivers/k8temp
11 0000:00:18.3 bind new_id unbind
12 /\
13
14 Also if you read k8temp doc in kernel src:
15 "Mapping of /sys files is as follows:
16
17 temp1_input - temperature of Core 0 and "place" 0
18 temp2_input - temperature of Core 0 and "place" 1
19 temp3_input - temperature of Core 1 and "place" 0
20 temp4_input - temperature of Core 1 and "place" 1"
21
22 So I'm sure /sys/devices/pci0000:00/0000:00:18.3/temp1_input is a core temperature.
23
24 > anyway, install a monitoring daemon (kima for kde is very light/good and
25 > integrates into the kicker) or any other like ksensors or others (i don't
26 > know the gnome ones since i completely dislike gnome but there are a lot of
27 > superkaramba widgets to do this) and see there your effective temp.
28 > there you should see:
29 > the core temp. mine turion 64 single core go from 30-35° idle to 60-65° C
30 > when at max speed and max load.
31 > the thermal goes from 58° when the fan starts to 75° when the processor goes
32 > in passive mode decreasing speed for avoiding overheating problems. at 105
33 > it shuts down.
34
35 They are just a frontend to lm_sensors which, in place, for k8temp sensor gets values from sysfs.
36 Values provided by lm_sensors are basically the same as /sys/devices/pci0000:00/0000:00:18.3/tempX_input
37
38 > now, i've heard that dualcores can have a temp of 5 to 10° higher than the
39 > single cores, but i don't know if this is really true.
40 > anyway, how do you know that that 2 devices are effectively the processor?!
41
42 # lspci -v -s 00:18.3
43 00:18.3 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron] Miscellaneous Control
44 /\
45
46 > your problem is that acpi doesn't get the processor states. if it doesn't
47 > provide you with trip points in /proc/acpi/thermal/thrmx/trip_points then
48 > you have an acpi problem. this may be caused by bios or by wrong kernel
49 > config.
50 > make sure that your thermal is compiled and started and that your bios is ok
51 > reading and doing this:
52 > http://gentoo-wiki.com/HOWTO_Fix_Common_ACPI_Problems
53 > the steps needed are the ones till you recompile the dsdt and see if the
54 > compiler raises errors. if it does then you might have a bios problem. try
55 > to fix it according the infos there and recompile.
56
57 I have processor states working and thermal_acpi is built into kernel.
58 I've already fixed some dsdt problems but they were not related to thermal management,
59 maybe this laptop ACPI doesn't have thermal management at all.
60
61 > if the bios is ok or the fixed one doesn't make the processor work with the
62 > latest kernel then search on the kernel.org the acpi bugzilla or the mailing
63 > list and report there your problem. they could help you fix it or they could
64 > patch the actual acpi to make this work.
65 > if you have an acer notebook, this might be the problem, since i've
66 > encountered many acer laptops that don't read the thermal correctly.
67
68 The strange thing is that this laptop ACPI do have an OSI specification for Linux, it should
69 be fairly compatible with it.
70
71 > --
72 > dott. ing. beso
73 >
74
75 --
76 Angelo Arrifano AKA MiKNiX
77 CSE Student at UBI, Portugal
78 Gentoo Linux AMD64 Arch Tester
79 miknix@×××××.com
80 http://miknix.homelinux.com
81 PGP Pubkey online
82
83 - -
84 People of privilege will always risk their complete destruction
85 rather than surrender any material part of their advantage.
86 -- John Kenneth Galbraith
87 --
88 gentoo-amd64@g.o mailing list