SUSE 9.2 and Toshiba A70 ACPI (I think) Problem

gertjan · 02-16-2005, 05:07 PM

Hmm... am not sure if Hyperthreading makes a difference. I believe stock Fedora 3 kernels come with hyperthreading off but ACPI is on - which is all I have ever tried (it gives me the battery display and ties the power on/off button to a gentle shutdown). But I also have the problem of having to manually switch off the lap top after a shutdown -h now so I doubts it's hyperthread related.

As an aside, Pavel M (one of the people involved in swsusp2) mentioned not to use an smp kernel when trying to get
sleep modes (S3/S4) to work.

I briefly looked at the omnibook sourceforge.net page but it doesnt look there is much recent activity. Still - I guess I have to try this too. Sigh...I long for the days when with RedHat6.2, Sleep/Hibernate worked like a charm and the fan rarely ever needed to come on with my Gateway Solo.

moonrover · 02-16-2005, 05:14 PM

Guess what we need is an ACPI module that is HyperSleep tolerant !!
Sorry for the comedy at this point, but we're getting nowhere twice as fast as we were before.

Sounds like with ACPI off, the fan is free-wheeling in some low-level communication with the heatsink to turn off-and-on for all of us, but just not the way we want it to with smooth variable rate speed increases that the ACPI would be capable of managing on demand based on processing in the CPU chip.

Regards,
Moonrover

statguy · 02-16-2005, 06:41 PM

A couple of thoughts. As for 3D acceleration, all I did was switch it on in SaX (is that a SuSE specific utility?) and it appears to work.

Next, let's step back from the problem a bit. There is clearly an ACPI problem. Some evidence is that /proc/acpi/fan is an empty directory on my system. But, if ACPI is the source of my lockups, why do I still get them with ACPI disabled. In fact, so far, nothing I've tried has had any effect on my lockups.

So, let's assume for a moment that ACPI is not the cause of the lockups. What are some other explanations. For example, a friend of mine, who is also using SuSE 9.2, mentioned that he thought there was a bug in the current kernel related to USB. I am using an external USB mouse. Is it within the realm of posibility that my mouse is the problem?

I'm just trying to take a fresh look rather than get tunnel vision on one particular known issue.

Hey, at least I'm learning things.

statguy · 02-16-2005, 07:11 PM

Check out the following URLs

http://wiki.linuxquestions.org/wiki/APIC
http://http://lists.suse.com/archive...-Dec/2716.html
http://seclists.org/lists/linux-kern.../Jan/1137.html

There is another beastie APIC (not ACPI) that deals with interrupts. But it can cause lockups.

Anyway, I've tried booting with the noapic boot parameter. Now, I'm not really clear what the difference is between apic and lapic. Yes, I know the l is for local, but does noapic imply nolapic?

moonrover · 02-16-2005, 10:56 PM

I doubt the mouse is the problem; disconnect it & use the laptop "scratch pad" & I'll bet you still have the same problems. I think we have lockups with or without ACPI, based on all the responses from everyone. It's buggy if enabled (KDE hangs when initializing peripherals for most of us) and we still have random fan & cooling shutdowns without it enabled.

3D graphics acceleration - good for you if that's all it took to enable it. Can you prove to yourself it is actually working?

gertjan · 02-17-2005, 12:23 AM

I think the general conclusion is that the hanging is not related to ACPI - I certainly havent had problems that I could relate to it. Also, I get occaisional freezes in XP as well. Dont use it as much so it's hard to judge how often. I am now using 2.6.10-741 (FC3 2nd to latest update) and my A70 has not hang up at all. I do wonder if the problem is release related (maybe even KDE). What's the latest kernel for Suse ? I also seem to recall sound driver problems at the same time as the hanging and had to add the snd-atiixp-modem on the hotplug blacklist (ask me if this is unclear).

My fans are both going full speed right now. Very annoying. Can someone summarize the results of the omnibook kernel patch ? It wasnt completely clear to me.

As an aside, I manage to do a hibernate (swsusp), at least without X running and removing some kernel modules. One small step...

G

statguy · 02-17-2005, 06:44 AM

The extent of testing on 3D acceleration is pretty limited. While exploring my newly installed system I tried to start the tackballs game and it wouldn't because it require 3D. I clicked it on in SaX and the game started. That's as far as I took it -- I had other things to do.

I ran my system for 3 hours last night after booting with noapic and there were no lockups despite many screensaver activations.

In short, the omnibook module "out of the box" does not solve the ACPI problems in the A70. I would say that all that works is temperature monitoring although it has no effect on fan behaviour. Yes, you could also change the lcd brightness, but those hot keys work on my system.

statguy · 02-17-2005, 03:57 PM

Booting with noapic did not prevent a lockup. I let the computer idle for several hours while I was out this morning. When I keyed my password and pressed enter to unlock the screen, it flashed once and froze.

More googling found this:

http://tiptsoft.com/laptop/linux.html

This guy traced his lockups to the frequency with which some apps were polling the /proc files. When I had turned up the verbosity of powersaved I saw something similar. Here it is:

Feb 12 19:28:46 rho [powersave][5417]: DIAG in Function getCPULoadSMP, line 490: Could not evalutate CPU load, /proc/stat is not updated, yet, please slow down polling this file!

I tried shutting down the kpowersave applet, but still locked up when trying to write this post. I had to start again.

I've also noticed a lot of messages like this one:

Feb 17 16:51:35 localhost kernel: SFW2-INext-DROP-DEFLT IN=eth0 OUT= MAC= SRC=192.168.0.102 DST=192.168.0.255 LEN=249 TOS=0x00 PREC=0x00 TTL=64 ID=9 DF PROTO=UDP SPT=138 DPT=138 LEN=229

I had assumed that these were from the firewall and had ignored them, but the last message before my last crash was this one:

Feb 17 16:33:12 localhost kernel: SFW2-INext-DROP-DEFLT IN=eth0 OUT= MAC=00:0f:b0:38:86:1d:00:10:4b:65:47:1c:08:00 SRC=192.168.0.100 DST=192.168.0.102 LEN=198 TOS=0x00 PREC=0x00 TTL=64 ID=0 DF PROTO=UDP SPT=138 DPT=138 LEN=178

I don't know if it is connected or not.

Right now I'm really really annoyed. Nothing I've tried seems to have any effect. I'm starting to consider exchanging this laptop for something else.

moonrover · 02-18-2005, 12:21 AM

I decided to try the screensaver/powerdown test tonight. I tested with AC power & with the battery and NOACPI in both cases. I set my screensaver to come on at 4 minutes, standby mode at 10 minutes, suspend at 20 minutes, and powerdown at 30 minutes.

I observed the screensaver & standby modes, but never got a suspend to disk or a power off shutdown to take effect. In both cases, as soon as I touched the touchpad, the desktop came back again. I was never dead in the water.

Regards,
Moonrover

Saturday Update:
Performed same tests with HyperThreading ON, same results. Got the screensaver & standby modes, but never the suspend & powerdown.

statguy · 02-18-2005, 07:20 AM

Quote:

Originally posted by gertjan
I think the general conclusion is that the hanging is not related to ACPI - I certainly havent had problems that I could relate to it. Also, I get occaisional freezes in XP as well. Dont use it as much so it's hard to judge how often. I am now using 2.6.10-741 (FC3 2nd to latest update) and my A70 has not hang up at all. I do wonder if the problem is release related (maybe even KDE). What's the latest kernel for Suse ? I also seem to recall sound driver problems at the same time as the hanging and had to add the snd-atiixp-modem on the hotplug blacklist (ask me if this is unclear).

That's interesting. In some instances where I've messed around with boot parameters, the boot process has frozen when doing something with atiixp. I also have these messages from dmesg.

ATIIXP: IDE controller at PCI slot 0000:00:14.1
ATIIXP: chipset revision 0
ATIIXP: not 100% native mode: will probe irqs later
ALSA sound/pci/atiixp.c:522: atiixp: codec reset timeout

The last message about the codec reset timeout also appears in /var/log/messages.

Will putting snd-atiixp-modem in the hotplug blacklist completely disable sound or only some sound component related to the internal modem?

On my system, running lsmod | grep snd gives the following:

Code:

snd_pcm_oss            66728  0
snd_mixer_oss          25216  1 snd_pcm_oss
snd_atiixp             25128  0
snd_ac97_codec         76640  1 snd_atiixp
snd_pcm               113284  3 snd_pcm_oss,snd_atiixp,snd_ac97_codec
snd_timer              31620  1 snd_pcm
snd                    70532  6 snd_pcm_oss,snd_mixer_oss,snd_atiixp,snd_ac97_codec,snd_pcm,snd_timer
soundcore              13792  1 snd
snd_page_alloc         14600  2 snd_atiixp,snd_pcm

I don't know if any of this will be helpful, but figure I'd rather err in giving too much information.

statguy · 02-18-2005, 11:28 AM

Hey. I did some more googling and think I've found the problem. Check out the last post by madmaxter on this URL.

http://notebookforums.com/archive/in...p/t-46919.html

It appears that static discharge is responsible for the freezes. In a matter of minutes I found three links describing the problem.

It is consistent with my most common lockup, walk back to the computer and type my password to unlock it. It explains why gertjan has had lockups in windows as well.

The good news is that the problem does not appear to be related to anything in linux, it's just coincidental that parts of ACPI don't work.

The bad news, how to prevent static discharge. It's a good thing I use an external mouse. I guess resting my hands on the console (especially the left speaker) is a no-no. I guess I have to ground myself before touching anything. Oh yeah, don't use it in an electrical storm.

Maybe some good news. After posting this I looked at some more links. Toshiba has acknowledged the problem and it can (apparently) be fixed by replacing the speakers.

gertjan · 02-18-2005, 01:35 PM

Hi,

Just a brief message (i am at work...) but I presume every one has seen:
http://www.csd.toshiba.com/cgi-bin/t...ceghdgngdgnj.0

(if the link doesnt work, go to Toshiba support and search for A70). The message (from Toshiba) is describing locking problems specifically with the A70.

Re: sound driver - no, blacklisting the modem does not have an effect on the sound card - in fact, I had to put it to get sound to work. SUSE may do this for you - at least I noticed tha the SUSE 9.2 demo CD (i forget the right term for it at the moment...I mean the 'knoppix' like release) got sound working right off the bat. Check your /etc/hotplug black list file. I dont have my laptop with me, or I would paste the entry here.

Gertjan

gertjan · 02-18-2005, 01:39 PM

Oops..reading more carefully, I guess we are both referring to the same issue, although Toshiba doesnt state it. I came across it indepently this morning.

moonrover · 02-18-2005, 02:01 PM

Good Googleing, guys. Guess we'll all need to order grounding straps with 20 foot tethers for use around the house. All my testing was done at a desk in my living room where I have no static discharge problems. I still plan to do the 2nd half of my testing this weekend (with HyperThreading ON; the previous tests were with it OFF) just to see if I see any lockups when the laptop should be hibernating or shutting down.

statguy · 02-18-2005, 02:24 PM

Arrangements have been made to send mine in for repair. They say it takes 2-3 business days.

BTW, one of the pages I read suggested that the lockups don't occur when running on battery. That may be true. I have not had a lockup when using it on the train and my hands are all over the console and speakers.