[Reason found]Freeze, need help to trace reason.

freakyzoidberg · 2009-06-29 00:04:29

Hi,

for a few weeks now, i have a lots of crash when my laptop is on heavy load
(compiling, playing games (usually spring, etc)) not when watching movies (vlc))

heres the last lines of log before it crash

Jun 29 11:31:50 ZoidBox kernel: CE: hpet increasing min_delta_ns to 75936 nsec

i downgraded the nvidia driver from 185 to 173 after reading a post with similar symptoms

but it just crash again with

Jun 29 11:31:50 ZoidBox kernel: CE: hpet increasing min_delta_ns to 75936 nsec
Jun 29 11:45:17 ZoidBox kernel: ACPI Exception (evregion-0422): AE_TIME, Returned by Handler for [EmbeddedControl] [20090320]
Jun 29 11:45:17 ZoidBox kernel: ACPI Error (psparse-0537): Method parse/execution failed [\_TZ_.THR1._TMP] (Node f7013600), AE_TIME

any idea what could be the reason, or how to know it

thanks

Last edited by freakyzoidberg (2009-06-30 03:40:36)

arkham · 2009-06-29 00:15:34

What does google say?
Try noapic.

freakyzoidberg · 2009-06-29 00:25:16

thanks will try this,

i let you know

edit:

i havent applied yet noapic at boot but here my last log before crash on spring rts

Jun 29 12:04:40 ZoidBox kernel: hda-intel: IRQ timing workaround is activated for card #0. Suggest a bigger bdl_pos_adj.
Jun 29 12:10:21 ZoidBox kernel: CE: hpet increasing min_delta_ns to 15000 nsec
Jun 29 12:12:08 ZoidBox kernel: CE: hpet increasing min_delta_ns to 22500 nsec
Jun 29 12:13:16 ZoidBox kernel: CE: hpet increasing min_delta_ns to 33750 nsec
Jun 29 12:13:43 ZoidBox kernel: CE: hpet increasing min_delta_ns to 50624 nsec
Jun 29 12:16:35 ZoidBox kernel: CE: hpet increasing min_delta_ns to 75936 nsec

zoidberg@ZoidBox ~ $ uname -vr
2.6.30-ARCH #1 SMP PREEMPT Fri Jun 19 21:25:17 UTC 2009

Last edited by freakyzoidberg (2009-06-29 00:29:02)

freakyzoidberg · 2009-06-29 01:17:42

nop freezed again even with noapic in grub menu.lst

Sharpsy · 2009-06-29 01:47:29

Try adding hpet=disable in menu.lst...

eldragon · 2009-06-29 03:44:02

heavy load + freeze ?

usually thats temperature related....

ralvez · 2009-06-29 13:23:04

It could also be related to bad RAM. Do you have access to another stick of RAM (that you know to be good, of course) to test it?

R.

windtalker · 2009-06-29 15:04:44

A bug report with some possible fixes:

https://bugs.launchpad.net/ubuntu/+sour … bug/267913

I got this message in my dmesg:
hda-intel: IRQ timing workaround is activated for card #0. Suggest a bigger bdl_pos_adj.

You (Nil) might try doing a "cat /proc/interrupts" at the console and see if snd-hda-intel is using the same interrupt as your SD card reader. If it is, you may be able to work around the SD card problem by adding the line "options snd-hda-intel enable_msi=1" to the bottom of your /etc/modprobe.d/alsa-base file.

I would like to add information that may shed light on this.
I too started getting this warning in the log after the upgrade from hardy to intrepid
could it have something to do with changing the time on the system?
For example, if I run hwclock --systohc + adjtimexconfig, I see the warning after a few seconds in the log. Here's the sequence:
Feb 22 01:07:45 myhost sudo: username : TTY=pts/9 ; PWD=/home/username ; USER=root
; COMMAND=/sbin/hwclock --systohc
Feb 22 01:07:47 myhost sudo: username : TTY=pts/9 ; PWD=/home/username ; USER=root
; COMMAND=/usr/sbin/adjtimexconfig
Feb 22 01:07:58 myhost kernel: [439231.900129] hda-intel: IRQ timing workaround is
activated for card #0. Suggest a bigger bdl_pos_adj.
This makes me suspect that changing the hardware clock, causes the hda-intel driver to wrongly conclude that something might be wrong and activate the presumably inferior "IRQ timing workaround". A pity.

freakyzoidberg · 2009-06-29 19:58:16

wow loads of answer:)

after setting noapic and disabling hpet, i didnt noticed any more freeze
but i wasnt able to read a simple divx with vlc too much frames drop.
so removed those.

the laptop has a few years and wasnt acting like that before (used debian on it during the 2 last years)

about the temperature
cpu is about 80 degrees when freezing and hddtemp says something laround 55-60 degrees

for the memory, i have a laptop which reduce the chance to find a stick quickly. but i ll follow your idea and will try a memcheck

cat /proc/interrupts says

           CPU0       CPU1       
  0:    5596141       4652   IO-APIC-edge      timer
  1:       1929          0   IO-APIC-edge      i8042
  8:        445          0   IO-APIC-edge      rtc0
  9:    1563358      29192   IO-APIC-fasteoi   acpi
 12:        123          0   IO-APIC-edge      i8042
 14:          0          0   IO-APIC-edge      ata_piix
 15:          0          0   IO-APIC-edge      ata_piix
 16:     384113         19   IO-APIC-fasteoi   uhci_hcd:usb5, ohci1394, nvidia
 17:        256          0   IO-APIC-fasteoi   mmc0
 18:          0          0   IO-APIC-fasteoi   uhci_hcd:usb4
 19:          0          0   IO-APIC-fasteoi   uhci_hcd:usb3
 23:     161730          0   IO-APIC-fasteoi   ehci_hcd:usb1, uhci_hcd:usb2
 28:      61009          1   PCI-MSI-edge      ahci
 29:     391261          5   PCI-MSI-edge      HDA Intel
 30:     147922         40   PCI-MSI-edge      eth0
 31:          0          0   PCI-MSI-edge      iwl3945
NMI:          0          0   Non-maskable interrupts
LOC:    1214800    3724442   Local timer interrupts
SPU:          0          0   Spurious interrupts
RES:     847061    1067025   Rescheduling interrupts
CAL:        429       2672   Function call interrupts
TLB:      57055      49347   TLB shootdowns
TRM:          0          0   Thermal event interrupts
ERR:          0
MIS:          0

so i guess there is no IRQ mistake

what i dont understand is why an error about sound caused by a possible hdd overheat when compiling result in computer freeze ^^

i ll be away today, but will keep the post update later this week, if i find anything to fix or to narrow the problem

thanks all of you anyway for your time and answers.

Last edited by freakyzoidberg (2009-06-29 20:00:35)

eldragon · 2009-06-29 20:21:04

freakyzoidberg wrote:

about the temperature
cpu is about 80 degrees when freezing and hddtemp says something laround 55-60 degrees

i suppose you mean Farenheit, otherwise, theres your problem: 80C is too hot.

freakyzoidberg · 2009-06-29 21:32:32

nop celsius

CPU
40*C on idle, or coding/surfing
50*C while watching videos
70-75*C on 1080p videos
never saw more than 80

but looking at the spec of a core2duo it says it can go up to 90-95* without damages

what scares me the most is the hard drive which go up to 60*C but in the laptop there is no 'cooling corridor' towards the hdd

on idle the hdd is at 46*C i cleaned it like last week, could it be possible that it s system related like never caching r/w or anything ?

arkham · 2009-06-29 21:41:02

Mmm ouch, install cpufreq and do your hardware a favor.

freakyzoidberg · 2009-06-29 21:43:20

already have ondemand :s

arkham · 2009-06-29 23:58:59

I suppose ondemand is not the right one since it uses the max frequency when the system is on heavy load; try limiting the maximum frequency in /etc/conf.d/cpufreq or setting the 'userspace' governor.

Have you tried booting the kernel just disabling hpet?

uwinkelvos · 2009-06-30 03:06:23

Your temps are ways to high. (all of them) Your CPU will not be damaged below the spec'ed temperatures, and somewhere before reaching that level, it will enventually simpy shutdown. even if that is not your problem, I have not seen a Core2Duo working reliably on temps higher than about 70°C. You might try mprime (http://aur.archlinux.org/packages.php?ID=6975). I have not used it on Linux, but pretty much on Windows for stability testing. If this shows any errors before your system freezes, your problem will be CPU/Ram-Overheating related. last but not least: As you mentioned, you are scared about your HDD-Temps, you really should be. I learned the hard way, that temperatures as high as that, will almost certainly result in a final headcrash. If there already are sectors damaged, you might have damaged binarys on your system, that cause the freeze. But i think this should result in a segfault and happen anytime the binary is accesed. Anyhow i have seen some bad things what dust can do to the cooling system of a notebook. You might try to open and clean it. (I recommend to google for an disassembly guide though )

freakyzoidberg · 2009-06-30 03:11:25

i m currently booting with :

kernel /boot/vmlinuz26 root=/dev/sda2 ro vga=791 irqpoll hpet=disable

irqpoll because noapic was causing way too much perf decrease
and to fix this in my log

irq 11: nobody cared (try booting with the "irqpoll" option)

i still have those stuff during boot and have already been reported by someone, and dont know if its related to my problem

Jun 30 14:50:14 ZoidBox load-modules.sh: 'acpi:PNP0C04:' is not a valid module or alias name
Jun 30 14:50:14 ZoidBox load-modules.sh: 'acpi:PNP0303:' is not a valid module or alias name
Jun 30 14:50:14 ZoidBox load-modules.sh: 'acpi:PNP0100:' is not a valid module or alias name
Jun 30 14:50:14 ZoidBox load-modules.sh: 'acpi:PNP0C02:' is not a valid module or alias name
Jun 30 14:50:14 ZoidBox load-modules.sh: 'acpi:PNP0C02:' is not a valid module or alias name

i tried also only using hpet=disable but crashed again.

i set cpufreq on userspace and change /etc/conf.d/cpufreq max to 1.2GHz (real freq is 1.6Ghz)

freakyzoidberg · 2009-06-30 03:39:38

success to finish a party in spring

at the end temperature are as follow

/dev/sda: HTS541010G9SA00: 49°C
temperature:             52 C

the begining of the end of my trouble ?

i guess the userspace + cpu hard limit did the trick
so clearly heat related

the only downside is that i now have a 1.2Ghz core 2 duo

ngoonee · 2009-06-30 04:32:45

Well, you could always experiment with cooling options. I use one of those pads which you can place your laptop on, with fans running below.

Or maybe just a huge block of ice (wrapped in plastic, of course) =p

freakyzoidberg · 2009-06-30 04:48:27

is it possible to tweak sysctl to less use the harddrive ? i mean more caching or delay writting ?
could not find any suitable article on how doing that

after idling the cpu is at 34*C but the HDD remains at 45*C

freakyzoidberg · 2009-06-30 05:11:53

ok i feel quite like a fool now,

while looking at my rc.conf i noticed that laptop-mode was not set.

i honnestly think this is the reason why the laptop overheated (maybe helped by some dust of course )

sorry guys if i make you loose your time, but thank you anyway for your help and support
glad to have join arch for a reactive community like you were on my problem

i ll mark this post as fix tomorrow if the temperature stay under an acceptable level.

Last edited by freakyzoidberg (2009-06-30 05:12:20)

eldragon · 2009-06-30 12:10:52

shouldnt there be anything related to laptop-mode and cpufreq in the begginers guide? its not too obvious people should set them up after install.

freakyzoidberg · 2009-07-01 02:38:42

i had it at a time, and then removed it because of i dont remember what, and apparently forgot to came back to that issue

Arch Linux

#1 2009-06-29 00:04:29

[Reason found]Freeze, need help to trace reason.

#2 2009-06-29 00:15:34

Re: [Reason found]Freeze, need help to trace reason.

#3 2009-06-29 00:25:16

Re: [Reason found]Freeze, need help to trace reason.

#4 2009-06-29 01:17:42

Re: [Reason found]Freeze, need help to trace reason.

#5 2009-06-29 01:47:29

Re: [Reason found]Freeze, need help to trace reason.

#6 2009-06-29 03:44:02

Re: [Reason found]Freeze, need help to trace reason.

#7 2009-06-29 13:23:04

Re: [Reason found]Freeze, need help to trace reason.

#8 2009-06-29 15:04:44

Re: [Reason found]Freeze, need help to trace reason.

#9 2009-06-29 19:58:16

Re: [Reason found]Freeze, need help to trace reason.

#10 2009-06-29 20:21:04

Re: [Reason found]Freeze, need help to trace reason.

#11 2009-06-29 21:32:32

Re: [Reason found]Freeze, need help to trace reason.

#12 2009-06-29 21:41:02

Re: [Reason found]Freeze, need help to trace reason.

#13 2009-06-29 21:43:20

Re: [Reason found]Freeze, need help to trace reason.

#14 2009-06-29 23:58:59

Re: [Reason found]Freeze, need help to trace reason.

#15 2009-06-30 03:06:23

Re: [Reason found]Freeze, need help to trace reason.

#16 2009-06-30 03:11:25

Re: [Reason found]Freeze, need help to trace reason.

#17 2009-06-30 03:39:38

Re: [Reason found]Freeze, need help to trace reason.

#18 2009-06-30 04:32:45

Re: [Reason found]Freeze, need help to trace reason.

#19 2009-06-30 04:48:27

Re: [Reason found]Freeze, need help to trace reason.

#20 2009-06-30 05:11:53

Re: [Reason found]Freeze, need help to trace reason.

#21 2009-06-30 12:10:52

Re: [Reason found]Freeze, need help to trace reason.

#22 2009-07-01 02:38:42

Re: [Reason found]Freeze, need help to trace reason.

Board footer