Gentoo Archives: gentoo-laptop

From: Beso <givemesugarr@×××××.com>
To: gentoo-laptop@l.g.o
Subject: Re: [gentoo-laptop] Re: [gentoo-user] Problem with CPU temperature (Santa Rosa CPU)
Date: Sat, 13 Oct 2007 17:19:48
Message-Id: d257c3560710131008m6b8e370ala1b5af2416b4f294@mail.gmail.com
In Reply to: Re: [gentoo-laptop] Re: [gentoo-user] Problem with CPU temperature (Santa Rosa CPU) by Marco Calviani
1 i think that you'll have to wait a little more and always stay aware when
2 using the pc till acpi supports it.
3 i'd recommend to stay on powersave when using the processor extensively so
4 that you'd not incur in hardware errors and failures. and it the fan is not
5 starting try to modify it by command prompt via echo "on" >
6 /proc/acpi/fan/..
7 this will work at least the thermal doesn't reach the state when it would
8 stop the fan, but since you don't have that point and that you cannot read
9 temperature starting it via echo "on" should always stay on. you'd have some
10 noise maybe, but you'd be sure that the processor would not run overheat.
11 and hope that the acpi people would fix that in the near future.
12 i'm sorry for not being able to help you more.
13 and as an advice for the future: before getting a notebook in the future
14 have some surf on the web to see if it's fully supported by linux (acer
15 sells linux only notebooks actually but only from taiwan). i had my linux
16 notebook not working with linux for 5-6 months and yet i had to change the
17 wireless since it wasn't supported after almost 2 years. for what i know
18 dell, hp and compaq are quite well supported.
19
20 2007/10/13, Marco Calviani <marco.calviani@×××××.com>:
21 >
22 > Just to give a short update:
23 > 1) no luck with the acer_acpi module: it does not load in the system,
24 > as the developer is having a look at it. However he told me that the
25 > module can do practically nothing on the thermal zone since this last
26 > point is controlled fully by ACPI
27 > 2) i've tried to modify the DSDT table: it had errors, but also after
28 > correction the system behave in practically the same way. I've mailed
29 > to the acpi-linux mailing list, hoping that the acpi gurus can try to
30 > solve the issue.
31 > As far as i understand now it is just a matter of ACPI implementation.
32 >
33 > Regards,
34 > m
35 >
36 > On 10/11/07, Beso <givemesugarr@×××××.com> wrote:
37 > > normally all acer have dsdt problems and don't work unless you load
38 > > acer_acpi.
39 > > now, in the portage tree there's only the 0.8.2 version, so you have to
40 > go
41 > > here:
42 > >
43 > > http://code.google.com/p/aceracpi/
44 > >
45 > > download the latest version and install it. there's a 5720 product
46 > listed
47 > > (under travelmate and not aspire, but it may have been an error). simply
48 > > search for acer_acpi in the /lib/modules/2.6.22-gentoo-r8 and delete it
49 > then
50 > > do a depmod -a and update the modules installed and retry loading the
51 > newly
52 > > compiled driver. if you manage to get it working in this way ok, if not
53 > try
54 > > contacting the guy maintaing the aceracpi module and tell him about you
55 > > problems. he may be of help.
56 > >
57 > >
58 > > 2007/10/11, Marco Calviani <marco.calviani@×××××.com>:
59 > > > Hi Beso,
60 > > > i have an Acer Aspire 5720. I've tried with the acer_acpi, it
61 > > > compiles well but when it comes to loading it fails, saying
62 > > > "No or unsupported WMI interface, unable to load".
63 > > >
64 > > > However as far as i can understood, this module deals with issues of
65 > > > buttons and hotkey not with ACPI issues.
66 > > >
67 > > > Yes, my DSDT failed during recompiling, but i have managed to solve
68 > > > the issues, and now it compiled well (with warning but i'm not
69 > > > caring). I'll try to recompile the new DSDT in the kernel, maybe it
70 > > > could help (but i'm not sure).
71 > > >
72 > > > Regards,
73 > > > m
74 > > >
75 > > > On 10/11/07, Beso <givemesugarr@×××××.com> wrote:
76 > > > > yep you're right.... i cannot modify my trippoints so this file
77 > cannot
78 > > be
79 > > > > modified....
80 > > > > i don't think that it's random, since it is 40 c.... try doing some
81 > > > > compiling or so and see if it goes up and then stop compiling and do
82 > > nothing
83 > > > > and see if it goes down.... this will tell you if thermal is
84 > working....
85 > > > > for dsdt problem you have follow this guide:
86 > > > > http://gentoo-wiki.com/HOWTO_Fix_Common_ACPI_Problems
87 > > > > if it gives you errors when recompiling dsdt then it may be a dsdt
88 > > problem,
89 > > > > but if it gives you no problem recompiling it then it may be that
90 > you
91 > > need
92 > > > > some additional modules like asus_acpi or ibm or toshiba ones based
93 > on
94 > > your
95 > > > > pc model. i don't know what you're using (i've managed to see some
96 > > lenovo,
97 > > > > asus toshiba and acer models around and for what i know every one of
98 > > then
99 > > > > needs an additional acpi module to have it work correctly). that was
100 > why
101 > > > > i've asked you for your brand and model name.
102 > > > >
103 > > > > 2007/10/11, Marco Calviani <marco.calviani@×××××.com>:
104 > > > > > Hi Beso,
105 > > > > > i've tried with your trip_points modification but it gives this
106 > > error:
107 > > > > >
108 > > > > > bash: echo: write error: Invalid argument
109 > > > > >
110 > > > > > like i'm not able to write on that file.
111 > > > > >
112 > > > > > And of course acpitool gives me a random ACPI temperature:
113 > > > > >
114 > > > > > Battery #1 : charged
115 > > > > > AC adapter : on-line
116 > > > > > Thermal zone 1 : ok, 40 C
117 > > > > >
118 > > > > > PS: i followed all your suggestions concerning the microcode and
119 > fan
120 > > > > > option in the kernel.
121 > > > > > Could this be a problem of DSDT?
122 > > > > >
123 > > > > > regards,
124 > > > > > m
125 > > > > >
126 > > > > > On 10/10/07, Beso <givemesugarr@×××××.com > wrote:
127 > > > > > >
128 > > > > > >
129 > > > > > > 2007/10/10, Marco Calviani < marco.calviani@×××××.com>:
130 > > > > > > > Hi Beso,
131 > > > > > > > sorry i misunderstood your suggestion. I did what you
132 > suggest
133 > > and
134 > > > > > > > this is the result of the trip_points:
135 > > > > > > >
136 > > > > > > > critical (S5): 100C
137 > > > > > >
138 > > > > > > wow.... you don't have anything that says to the cpu to slow
139 > down
140 > > when
141 > > > > it
142 > > > > > > reaches some point....
143 > > > > > > now, to add some other trip points you have to copy these in a
144 > > konsole
145 > > > > with
146 > > > > > > root priviledges:
147 > > > > > > echo "passive: 78 C: tc1=3 tc2=1 tsp=150
148 > > devices=CPU0 \
149 > > > > > > active[0]: 68 C: devices= FN1 \
150 > > > > > > active[1]: 58 C: devices= FN2" >>
151 > > > > > > /proc/acpi/thermal_zone/TZ01/trip_points
152 > > > > > >
153 > > > > > > then do a cat on the thermal_zone/TZ0/trip_points to see if
154 > you've
155 > > added
156 > > > > the
157 > > > > > > lines for passive and active lines.
158 > > > > > > that means that when your thermal reaches 78 degrees it will
159 > slow
160 > > down
161 > > > > the
162 > > > > > > processor. from 58 to 68 it will turn on the fan but don't turn
163 > down
164 > > the
165 > > > > > > speed of the processor, below 58 it will turn off the fan.
166 > > > > > > >
167 > > > > > > >
168 > > > > > >
169 > > > > > > > which i suppose is the reason why at that temperature the
170 > laptop
171 > > > > switch
172 > > > > > > off.
173 > > > > > > > So, nothing except for the critical state. Should i have to
174 > add
175 > > there
176 > > > > > > something?
177 > > > > > >
178 > > > > > > if the pc turns down then it can read from somewhere the actual
179 > > thermal
180 > > > > > > point. you try to see after actually setting the things i've
181 > just
182 > > said,
183 > > > > if
184 > > > > > > your pc is behaving as it should. remember to also turn on the
185 > > polling
186 > > > > > > frequency. without it it will not look for thermal changes. and
187 > > remember
188 > > > > to
189 > > > > > > actually compile the mce and speedstep features in the kernel
190 > and
191 > > not as
192 > > > > > > module and reboot and then set the things i've mentioned. after
193 > that
194 > > > > type
195 > > > > > > acpitool (it should be installed by default with the acpi
196 > package)
197 > > and
198 > > > > see
199 > > > > > > what it says. it should give something like this:
200 > > > > > >
201 > > > > > > > Battery #1 : charging, 46.00%, 01:17:04
202 > > > > > > > AC adapter : on-line
203 > > > > > > > Thermal zone 1 : activ, 58 C
204 > > > > > >
205 > > > > > > it indicates, as you can see not only the battery and ac status
206 > but
207 > > also
208 > > > > the
209 > > > > > > current processor mode (active) and the current thermal
210 > > temperature....
211 > > > > if
212 > > > > > > you don't have acpitool try acpi -t (you'll surely have either
213 > one
214 > > or
215 > > > > the
216 > > > > > > other) and it should indicate the thermal state and temperature.
217 > if
218 > > this
219 > > > > > > command don't give you these infos then you'll have to be very
220 > > careful
221 > > > > using
222 > > > > > > your pc since acpi probably don't support santarosa well. you
223 > should
224 > > > > then
225 > > > > > > unmask newer acpi in portage (adding acpi in
226 > /etc/package.keywords)
227 > > and
228 > > > > try
229 > > > > > > with the new acpi ( 1.0.6).
230 > > > > > > you may need some additional acpi modules, like ibm_acpi for
231 > > example,
232 > > > > but
233 > > > > > > that depends on your's pc brand. i'm looking around to see if
234 > there
235 > > were
236 > > > > > > someone that had problems with santarosa and linux acpi, but for
237 > the
238 > > > > moment
239 > > > > > > i couldn't find something useful. try what i've said and see if
240 > the
241 > > > > things
242 > > > > > > work. if they work then append the tweak in some script that
243 > starts
244 > > at
245 > > > > boot
246 > > > > > > like the top of xdm script and you'll have a functional system.
247 > let
248 > > me
249 > > > > know
250 > > > > > > if you were succesful on that.
251 > > > > > >
252 > > > > > >
253 > > > > > > > An additional problem is this:
254 > > > > > > > doing a
255 > > > > > > > $ cat /proc/acpi/thermal_zone/TZ01/temperature
256 > > > > > > > it gives only a:
257 > > > > > > >
258 > > > > > > > temperature: 0C
259 > > > > > > >
260 > > > > > > > My question is:
261 > > > > > > > even if i change the polling frequency, how the fan can start
262 > if
263 > > the
264 > > > > > > > temperature gives 0??
265 > > > > > > > Do you know if it's possible to link the fan start with the
266 > core
267 > > > > > > > temperature instead of the ACPI thermal zone?
268 > > > > > >
269 > > > > > > probably the fan don't start since you have it as a module.
270 > i've
271 > > > > curently
272 > > > > > > had the same problem which solved by compiling it integrated in
273 > the
274 > > > > kernel.
275 > > > > > > the fan in your case should always be on, not always be off....
276 > > > > > >
277 > > > > > > >
278 > > > > > > > Regards,
279 > > > > > > > m
280 > > > > > > >
281 > > > > > > > >
282 > > > > > > > > thats why i told you to do this commanda:
283 > > > > > > > > > echo "2 seconds" >
284 > > > > > > > > /proc/acpi/thermal_zone/TZ01/polling:frequency
285 > > > > > > > > >
286 > > > > > > > > this enables the polling of your thermal every 2 seconds.
287 > this
288 > > > > should
289 > > > > > > be
290 > > > > > > > > enough.
291 > > > > > > > > do you have the other file that i mentioned:
292 > > > > > > > > > /proc/acpi/thermal_zone/TZ01/trip_points ?!
293 > > > > > > > > this sets the trip points for your processor. whitout it you
294 > > > > governor
295 > > > > > > cannot
296 > > > > > > > > understand what to do even if it polls right.
297 > > > > > > > >
298 > > > > > > > > as for the kernel thigs, set these options:
299 > > > > > > > > select processor type: intel core2 instead of normal x86
300 > > > > > > > > CONFIG_X86_SPEEDSTEP_CENTRINO=y instead of m
301 > > > > > > > > CONFIG_ACPI_FAN=y instead of m
302 > > > > > > > > CONFIG_MICROCODE=y instead of m (for what i know cpu
303 > micocode is
304 > > > > needed
305 > > > > > > on
306 > > > > > > > > intels)
307 > > > > > > > > CONFIG_K8_NUMA=n instead of y (this should be the amdk8
308 > numa,
309 > > that
310 > > > > you
311 > > > > > > > > should not need. if it's not then let him be)
312 > > > > > > > > CONFIG_X86_MCE_AMD=n instead of y (you don't need amd mce
313 > > features
314 > > > > since
315 > > > > > > > > they are not included into intel cpus)
316 > > > > > > > > CONFIG_NR_CPUS=32 <-- this leaves me a little dazzled: do
317 > you
318 > > really
319 > > > > > > have 32
320 > > > > > > > > cpus in your core?! for what i know this sets the real
321 > number of
322 > > > > cpus
323 > > > > > > inside
324 > > > > > > > > the kernel, but i might be wrong. so if this is really what
325 > i
326 > > think
327 > > > > it
328 > > > > > > is,
329 > > > > > > > > ie the real nr of cpus (not virtual ones) set this to 2 or 4
330 > > based
331 > > > > on
332 > > > > > > your
333 > > > > > > > > cpu cores.
334 > > > > > > > > CONFIG_HOTPLUG_CPU=n instead of y (i don't really think that
335 > > you'll
336 > > > > > > unplug
337 > > > > > > > > your cpu from your laptop when the laptop is still running
338 > > > > considering
339 > > > > > > that
340 > > > > > > > > you don't use multi cpus but a single multicore cpu. the
341 > same
342 > > goes
343 > > > > for
344 > > > > > > > > memory hotplug: i don't think that your laptop supports it,
345 > so
346 > > just
347 > > > > > > disable
348 > > > > > > > > it.)
349 > > > > > > > > for what i have seen the acpi problems may be due to a
350 > failure
351 > > in
352 > > > > > > loading
353 > > > > > > > > the intel speedstep module. if you look into the modules
354 > loaded
355 > > > > (lsmod)
356 > > > > > > you
357 > > > > > > > > should not see it. so it's better to insert it directly in
358 > the
359 > > > > kernel,
360 > > > > > > since
361 > > > > > > > > it is one of the first modules called (if you use it as a
362 > module
363 > > you
364 > > > > > > should
365 > > > > > > > > be loading it with initramdisk before loading acpi to have a
366 > > full
367 > > > > acpi
368 > > > > > > > > configuration).
369 > > > > > > > > try setting these options and recompile and install the new
370 > > kernel
371 > > > > and
372 > > > > > > > > modules and reboot (kexec is not working on my amd turion
373 > with
374 > > 64bit
375 > > > > > > enabled
376 > > > > > > > > and so may also be for your core2duo).
377 > > > > > > > >
378 > > > > > > > > 2007/10/10, Marco Calviani < marco.calviani@×××××.com>:
379 > > > > > > > > > Sorry i missed the attachment.
380 > > > > > > > > >
381 > > > > > > > > > regards,
382 > > > > > > > > > m
383 > > > > > > > > >
384 > > > > > > > > >
385 > > > > > > > >
386 > > > > > > > >
387 > > > > > > > >
388 > > > > > > > > --
389 > > > > > > > > dott. ing. beso
390 > > > > > > > --
391 > > > > > > > gentoo-laptop@g.o mailing list
392 > > > > > > >
393 > > > > > > >
394 > > > > > >
395 > > > > > >
396 > > > > > >
397 > > > > > > --
398 > > > > > > dott. ing. beso
399 > > > > > --
400 > > > > > gentoo-laptop@g.o mailing list
401 > > > > >
402 > > > > >
403 > > > >
404 > > > >
405 > > > >
406 > > > > --
407 > > > > dott. ing. beso
408 > > > --
409 > > > gentoo-laptop@g.o mailing list
410 > > >
411 > > >
412 > >
413 > >
414 > >
415 > > --
416 > > dott. ing. beso
417 > --
418 > gentoo-laptop@g.o mailing list
419 >
420 >
421
422
423 --
424 dott. ing. beso

Replies

Subject Author
Re: [gentoo-laptop] Re: [gentoo-user] Problem with CPU temperature (Santa Rosa CPU) Marco Calviani <marco.calviani@×××××.com>