Random freezes - Intel

Purgator · 2016-12-15 11:43:12

TL;DR : Solve freeze issue on Intel baytrail

If you are experiencing freezing issues on a baytrail CPU you are probably affected by the problem described in this topic.
The issue seems to be about energy management when the system is running idle.

There are 3 way to help you at the moment :

Disable c6-state. If you have grub, edit /etc/default/grub and add the parameter intel_idle.max_cstate=1 to your GRUB_CMDLINE_LINUX_DEFAULT. Then run grub-mkconfig -o /boot/grub/grub.cfg as root.
Patch the kernel with this : https://bugzilla.kernel.org/attachment. … ction=edit. According to https://bugzilla.kernel.org/show_bug.cgi?id=109051
Use the baytrail custom kernel from aur : https://aur.archlinux.org/packages/linux-baytrail48/

It seems WiFi can cause issues. Try using only a wired connection (to do a pacman -Syu for example).

------------------------

Hello everyone,

I installed Arch on my laptop a year ago, and since that day I got the same problem : My system freez to death randomly. I have to push the power button 5 sec to turn off.

While I have tryed everything to solve the issue, I will try to describe everything that I have done, and I hope someone gonna give me some help. I'm honestly bored about this.

First, I downgraded xf86-video-intel to the version release on march or something like that and blocked the update of this package.

Then, I tried to force uxa and such following this doc : https://wiki.archlinux.org/index.php/in … figuration
I tried everything, uxa, sna, DRI 2-3 etc nothing really changed anything.

Something quite useful (can't really prove it) was setting intel_idle.max_cstate=1 in grub.cfg.

With all these settings, i was having the freez like 1 time a week.

I wanna add that it seems to happend more frequently while watching videos on firefox. But i can't proove it.

After getting bored of freezing again, i decided to update xf86-video-intel, i did again some tests by changing parameters in /etc/X11/xorg.conf.d/20-intel.con but that was crappy.

I finally decided to uninstall that intel driver, as it seemed to be the source of the problem, as the wiki and many people said.

I removed xf86-video-intel and the conf file /etc/X11/xorg.conf.d/20-intel.con. I also removed the intel_idle.max_cstate=1 in grub.cfg.

It seems my system is more reactive, but now I have the impression that i get more freezes while watching a video on firefox, and no freeze when doing common things like editing a file.
After weeks of testing, it appear that the system still can freeze everywhere randomly. For exemple, today, I got 2 freezes with only a terminal emulator open. It was during an update (pacman -Syu). This is totally not related while it's the first time i got a freez during an update.

I don't remember everything I tryed but that a summary.

I really lost in blue guys, what i can do ?

Last edited by Purgator (2017-10-12 10:18:25)

Awebb · 2016-12-15 12:11:28

Look up "microcode" on the wiki, if you haven't already.

Purgator · 2016-12-15 12:20:46

I already took a look at this but it seems I never did anything about that.
Thank you btw, microcode wasnt installed. I installed it and runned grub-mkconfig. Let's see if something changed

EDIT : After a reboot, i launched FIrefox. I openned a terminal and ran pacman -Syu and the computer freez :'(
I Stoped the computer by pushing the on/off button, started again, oppened a terminal, ran pacman -SYu, it worked.
Now i oppened firefox, i'm here but the freez can happend any time.

What's next ?

Last edited by Purgator (2016-12-15 12:28:00)

Awebb · 2016-12-15 12:59:07

Next time the thing freezes, try switching to a TTY (Ctrl-Alt-F#), I didn't see you mention that, so maybe the machine isn't frozen and just X is. This is just basic stuff you didn't mention, like the microcode.

The next step would be to filter the journal (journalctl), dmesg and the xorg log file for the times before and after your freezes and see if it tells you anything.

seth · 2016-12-15 13:01:58

Figure what "freeze" means (see whether the system echoes to pings; ideally you can ssh into it)
And inspect dmesg (ensure the journal doesn't log to volatile memory and "journalctl _TRANSPORT=kernel") after the freeze.

Purgator · 2016-12-15 13:09:35

I got the problem during a year, as i said i have to shutdown the computer by pressing the on/off button 5 seconds. It's frozen, there is no way i can't do anything. The first thing i always try when i have a problem with X is ctrl alt f# of course. It would be too easy. Plus, to be helpful there is stricly no log cause the system can't log cause of the freez.
When the freez appear, it's dead.

Note that i also tryed to connect with ssh during the freez. That didnt worked.

EDIT : @Seth, by freez i really mean that a freez. The screen display the last display correctly, but nothing moves, the last bit of song output is looping, and all peripherals are down.

Last edited by Purgator (2016-12-15 13:12:52)

Awebb · 2016-12-15 15:38:25

I'd start running memtest86+ from memtest.org, this smells a bit fishy.

seth · 2016-12-15 15:41:05

It's a halted kernel - the inability to cause it at will makes it incredibly hard to debug (since you'll have to watch kernel messages while it happens)
Memory errors are always a good candidate, but also record the loaded modules (if there's something "odd") ie. dump "lsmod"

Purgator · 2016-12-15 16:53:07

I can't record it cause it can happend anytime but also it can not happend during days.
It cannot be a memory issue cause there is no problem on Windows, and i don't think a new hardware can be already broken.

[spoiler]

Module                  Size  Used by
ctr                    16384  4
ccm                    20480  2
fuse                   98304  3
joydev                 20480  0
mousedev               20480  0
uvcvideo               86016  0
videobuf2_vmalloc      16384  1 uvcvideo
videobuf2_memops       16384  1 videobuf2_vmalloc
videobuf2_v4l2         20480  1 uvcvideo
videobuf2_core         36864  2 uvcvideo,videobuf2_v4l2
videodev              151552  3 uvcvideo,videobuf2_core,videobuf2_v4l2
snd_hda_codec_hdmi     45056  1
media                  32768  2 uvcvideo,videodev
hid_generic            16384  0
snd_hda_codec_via      24576  1
snd_hda_codec_generic    69632  1 snd_hda_codec_via
btusb                  40960  0
btrtl                  16384  1 btusb
btbcm                  16384  1 btusb
btintel                16384  1 btusb
bluetooth             487424  5 btrtl,btintel,btbcm,btusb
usbhid                 49152  0
iTCO_wdt               16384  0
arc4                   16384  2
hid                   114688  2 hid_generic,usbhid
iTCO_vendor_support    16384  1 iTCO_wdt
intel_rapl             20480  0
intel_soc_dts_iosf     16384  0
intel_powerclamp       16384  0
coretemp               16384  0
kvm_intel             188416  0
i915                 1241088  4
kvm                   516096  1 kvm_intel
irqbypass              16384  1 kvm
crct10dif_pclmul       16384  0
crc32_pclmul           16384  0
crc32c_intel           24576  0
ghash_clmulni_intel    16384  0
cryptd                 20480  1 ghash_clmulni_intel
intel_cstate           16384  0
drm_kms_helper        126976  1 i915
evdev                  24576  17
snd_intel_sst_acpi     16384  0
snd_intel_sst_core     69632  1 snd_intel_sst_acpi
input_leds             16384  0
mac_hid                16384  0
pcspkr                 16384  0
psmouse               126976  0
snd_soc_sst_mfld_platform    77824  1 snd_intel_sst_core
snd_soc_sst_match      16384  1 snd_intel_sst_acpi
iwlmvm                278528  0
snd_hda_intel          32768  1
i2c_i801               24576  0
drm                   294912  5 i915,drm_kms_helper
snd_soc_core          188416  1 snd_soc_sst_mfld_platform
i2c_smbus              16384  1 i2c_i801
snd_hda_codec         106496  4 snd_hda_intel,snd_hda_codec_hdmi,snd_hda_codec_generic,snd_hda_codec_via
iwlwifi               188416  1 iwlmvm
rtsx_pci_ms            20480  0
intel_gtt              20480  1 i915
memstick               16384  1 rtsx_pci_ms
syscopyarea            16384  1 drm_kms_helper
sysfillrect            16384  1 drm_kms_helper
sysimgblt              16384  1 drm_kms_helper
snd_hda_core           65536  5 snd_hda_intel,snd_hda_codec,snd_hda_codec_hdmi,snd_hda_codec_generic,snd_hda_codec_via
fb_sys_fops            16384  1 drm_kms_helper
i2c_algo_bit           16384  1 i915
r8169                  77824  0
snd_compress           20480  1 snd_soc_core
snd_hwdep              16384  1 snd_hda_codec
mii                    16384  1 r8169
lpc_ich                24576  0
snd_pcm_dmaengine      16384  1 snd_soc_core
mei_txe                20480  0
ac97_bus               16384  1 snd_soc_core
mei                    86016  1 mei_txe
snd_pcm                90112  7 snd_hda_intel,snd_hda_codec,snd_pcm_dmaengine,snd_hda_core,snd_hda_codec_hdmi,snd_soc_sst_mfld_platform,snd_soc_core
shpchp                 32768  0
snd_timer              28672  1 snd_pcm
thermal                20480  0
wmi                    16384  0
snd                    69632  13 snd_compress,snd_hda_intel,snd_hwdep,snd_hda_codec,snd_timer,snd_hda_codec_hdmi,snd_hda_codec_generic,snd_soc_sst_mfld_platform,snd_soc_core,snd_pcm,snd_hda_codec_via
soundcore              16384  1 snd
fjes                   28672  0
video                  36864  1 i915
battery                20480  0
ac                     16384  0
tpm_tis                16384  0
tpm_tis_core           20480  1 tpm_tis
tpm                    36864  2 tpm_tis,tpm_tis_core
button                 16384  1 i915
sch_fq_codel           20480  6
rtl8187                40960  0
led_class              16384  3 iwlmvm,input_leds,rtl8187
mac80211              667648  2 iwlmvm,rtl8187
cfg80211              495616  4 iwlmvm,iwlwifi,mac80211,rtl8187
eeprom_93cx6           16384  1 rtl8187
rfkill                 20480  5 bluetooth,cfg80211
cdc_acm                32768  0
ip_tables              28672  0
x_tables               28672  1 ip_tables
ext4                  524288  2
crc16                  16384  2 bluetooth,ext4
jbd2                   90112  1 ext4
fscrypto               24576  1 ext4
mbcache                16384  3 ext4
sr_mod                 24576  0
cdrom                  53248  1 sr_mod
sd_mod                 36864  4
rtsx_pci_sdmmc         24576  0
mmc_core              122880  1 rtsx_pci_sdmmc
serio_raw              16384  0
atkbd                  24576  0
libps2                 16384  2 atkbd,psmouse
ahci                   36864  3
libahci                28672  1 ahci
xhci_pci               16384  0
xhci_hcd              172032  1 xhci_pci
usbcore               208896  7 uvcvideo,usbhid,cdc_acm,xhci_pci,btusb,rtl8187,xhci_hcd
usb_common             16384  1 usbcore
rtsx_pci               49152  2 rtsx_pci_sdmmc,rtsx_pci_ms
libata                208896  2 ahci,libahci
scsi_mod              159744  3 sd_mod,libata,sr_mod
i8042                  28672  0
serio                  20480  9 serio_raw,atkbd,psmouse,i8042

[/spoiler]

Slithery · 2016-12-15 16:57:44

Purgator wrote:

It cannot be a memory issue cause there is no problem on Windows, and i don't think a new hardware can be already broken.

Neither of these statements are true.

You should run memtest for several passes to make sure (I usually let it run overnight).

Last edited by Slithery (2016-12-15 16:59:54)

Purgator · 2016-12-16 13:31:15

Sure yeah I will do a memtest. I'm looking for new ram anyway, that would be so great if it solve the issue

Awebb · 2016-12-16 13:58:23

slithery wrote:

Purgator wrote:
It cannot be a memory issue cause there is no problem on Windows, and i don't think a new hardware can be already broken.
Neither of these statements are true.

Indeed. My laptop was broken from the start, but I didn't notice. Some 3D applications in Windows didn't work, but I thought it's just a quirk with the Optimus setup. The performance had its limits, but I had no frame of reference, I just thought this was as much as this machine can do. Same performance on Linux and no problems at all. A year later, the video device gave up entirely. As it turns out, half the customers complained about the exact same problem with those cards from the same batch. The problems went away after they replaced the card, the overall 3D performance almost doubled. RAM problems are hard to catch, because memory management is quite flexible and many processes won't do the exact same thing twice.

Doesn't mean it's the RAM. I just spent so many hours chasing freezes myself, when a RAM check revealed the culprit over night.

Purgator · 2016-12-19 13:27:54

Yeah sure that possible but no one got this problem with the computer as i seen (or not reported).
The fact is here, the RAM is the component that can break the less.

I still didn't take time to do the memtest.

Btw i'm getting the freez again while doing an update, during pacman download packages. Coincidence ? Dunno.

Do you know any way to dump something interesting from my memory ? /dev/mem isnt really relevant.
Is that not pointless ? Ram is too verbose, but maybe, dumping about some GB of memory that is recorded every second, and saved to the disk, would maybe detect something wrong in the logs.

Awebb · 2016-12-19 14:16:59

You're wrong on RAM being a robust component. I've had my own PC repair service company for many years. The statistics on actual hardware failure looked like HDD > RAM > GFX card > PSU > mainboard > CPU > other for internal components.

Some heavy dd or rsync to a tmpfs and observing the dmesg and other logs might be a fit first fix, but an extensive memtest86+ is easier, more reliable and produces the least trouble for inexperienced users. What you could do, is run the computer without X for a while. Do console stuff. Move files around. Download stuff. This might tell you something about the nature of the problem.

Purgator · 2016-12-19 15:45:36

I don't know for the past but now ram is one of the most robust component, HD and Graphics card pretend to fail so much more. According to one of the biggest IT retailer in France. Anyway that just stats, can't prove anything.
The only thing that we should consider is ram guaranties time is the biggest one.
Regardless of any stat, ram is just some memory on a chip, hard to break compared to others components.

I would like to work without X, but no time for that

Gonna try memtest first then.

Purgator · 2016-12-23 10:27:53

Memtest86+ 14 hours, 7 pass 0 error.

Ram is strong.

I have noticed that Gmail is very slow these times, I may gonna try install back xf86-video-intel

R00KIE · 2016-12-23 15:31:17

TL;DR, after installing intel-ucode and running grub-mkconfig did you check that it is actually loaded by grub? Do check the generated grub.cfg and make sure it is configured to be loaded, also check dmesg and see if the microcode is actually being updated early.

Purgator · 2016-12-26 10:49:22

In my grub.cfg it is loaded :
initrd /boot/intel-ucode.img /boot/initramfs-linux.img

From dmesg :

[  +0,001901] microcode: sig=0x30678, pf=0x8, revision=0x809
[  +0,000698] microcode: Microcode Update Driver: v2.01 <tigran@aivazian.fsnet.co.uk>, Peter Oruba

But I remember the first time I installed microcode I saw an update from dmesg.

Then I Installed also iucode-tool from AUR

and ran (as in the wiki)

sudo modprobe cpuid
bsdtar -Oxf /boot/intel-ucode.img | iucode_tool -tb -lS -

That return this (no update available)

iucode_tool: system has processor(s) with signature 0x00030678
microcode bundle 1: (stdin)
selected microcodes:

More verbosity :

$ grep microcode /proc/cpuinfo
microcode	: 0x809
microcode	: 0x809
microcode	: 0x809
microcode	: 0x809

collector1871 · 2016-12-26 11:08:24

I had the same problem (random freezing) and it was not issue with processor but with graphic card(s).

My issue was like this:
- notebook with integrated graphic card (intel) and external radeon.
- random freezing, only "hard power off" was working
- working radeon was providing hardware to the heating problem (high temperature).

What was helpful:
0. check temperature, sensors
1. clean notebook inside (assembly it first...)
2. turn off radeon graphic card, look here:
https://help.ubuntu.com/community/HybridGraphics
3. blacklist in kernel modules "radeon"
4. turn off battery save mode
5. turn on Intel graphic , set is as default
6. install correct drivers for it, turn off everything with name "radeon"

After that - no more "freezing" (in my case).

Last edited by collector1871 (2016-12-26 11:13:23)

Purgator · 2016-12-26 12:18:54

Hey thank you for answering.

As I don't have any dedicated GC, I'm already running on intel chip.

But i would like to know what do you use as graphics drivers and configuration files for this please.

For anyone who read this I accept any kind of advice or suggestion.

Last edited by Purgator (2016-12-27 15:33:00)

Purgator · 2016-12-30 09:31:37

I'm still up for any suggestion.

I'm back to intel driver btw.

adamlau · 2016-12-30 10:49:27

So fifteen days in and you have not found a suitable resolutiuon. I would reinstall Arch as that should only take a few hours vs the potential of waiting longer for additional forum insight. My suggestion may be bad practice and unpopular, but that is what I would do were I completely baffled for weeks...

Purgator · 2016-12-30 17:12:18

Lol no I will not install again, it won't change anything, why it would change ? I will just follow the same steps.

BTW it's not been fifteen days but near a year i got this issue.

Maybe my computer just can't run Linux.

seth · 2016-12-30 20:01:57

Sure it's "linux" - more like X11? Maybe the xorg driver? Or did you try to trigger this in a VT only session?
In case it's "more like X11", maybe kick the intel driver and try the (auto-fallback) modesetting driver?
If that's relevant, what kind of session do you usually run?

Other branch: since cryptd et al. are loaded: is this an encrypted installation? Do you also get it on a non-encrypted installation?
Does this also happen with other (live) linux'es, eg. grml or knoppix?

...?

Purgator · 2017-01-04 14:44:14

Hey !

I already tryed to kick the intel drive and use modesetting. The freezes happened differently, but that was not better. I tryed both during a long time. It's better on intel driver with specifics settings.

What do you mean by VT session ? You mean no xorg on ? Only on terminal like we got thru 'Ctrl + Alt + <FX>' ? I can try but It may need a lot of time then I will see.

I already think about trying on an other distro, I'm gonna try that, that can be a hudge help.

I don't know anything about an encrypted installation

Arch Linux

#1 2016-12-15 11:43:12

Random freezes - Intel

TL;DR : Solve freeze issue on Intel baytrail

#2 2016-12-15 12:11:28

Re: Random freezes - Intel

#3 2016-12-15 12:20:46

Re: Random freezes - Intel

#4 2016-12-15 12:59:07

Re: Random freezes - Intel

#5 2016-12-15 13:01:58

Re: Random freezes - Intel

#6 2016-12-15 13:09:35

Re: Random freezes - Intel

#7 2016-12-15 15:38:25

Re: Random freezes - Intel

#8 2016-12-15 15:41:05

Re: Random freezes - Intel

#9 2016-12-15 16:53:07

Re: Random freezes - Intel

#10 2016-12-15 16:57:44

Re: Random freezes - Intel

#11 2016-12-16 13:31:15

Re: Random freezes - Intel

#12 2016-12-16 13:58:23

Re: Random freezes - Intel

#13 2016-12-19 13:27:54

Re: Random freezes - Intel

#14 2016-12-19 14:16:59

Re: Random freezes - Intel

#15 2016-12-19 15:45:36

Re: Random freezes - Intel

#16 2016-12-23 10:27:53

Re: Random freezes - Intel

#17 2016-12-23 15:31:17

Re: Random freezes - Intel

#18 2016-12-26 10:49:22

Re: Random freezes - Intel

#19 2016-12-26 11:08:24

Re: Random freezes - Intel

#20 2016-12-26 12:18:54

Re: Random freezes - Intel

#21 2016-12-30 09:31:37

Re: Random freezes - Intel

#22 2016-12-30 10:49:27

Re: Random freezes - Intel

#23 2016-12-30 17:12:18

Re: Random freezes - Intel

#24 2016-12-30 20:01:57

Re: Random freezes - Intel

#25 2017-01-04 14:44:14

Re: Random freezes - Intel

Board footer