[SOLVED] hp Spectre x360 freezing without warning

cellisten · 2015-12-22 11:59:27

Hello all,
I have been trying to troubleshoot this for some time now. On my laptop, the screen freezes (no mouse cursor movement, no graphical updates, no other TTYs available and no ssh server available). This happens immediately/after a few minutes/after a few hours. I thought it might be the noAccel option that was required, didn't help. I have tried forcing the fan to be always on, didn't help.
To check if it was heat related, I have booted up Windows directly after getting lots of crashes, cannot reproduce it there.
When it happens, it seems that the screen freezes and the mouse cursor starts moving very sluggishly.

Haven't been able to find anything with journalctl -xe for the boots in question.

It seems that the crash does not occur in gdm but only when gnome is started. Lately it almost happens every time...

Found some more info... If I set i915.semaphores=1 I get crashes at every boot including this log message:

dec 22 13:17:33 bb-8 kernel: BUG: unable to handle kernel NULL pointer dereference at           (null)
dec 22 13:17:33 bb-8 kernel: IP: [<          (null)>]           (null)
dec 22 13:17:33 bb-8 kernel: PGD 7d3e7067 PUD 7d3c6067 PMD 0 
dec 22 13:17:33 bb-8 kernel: Oops: 0010 [#1] PREEMPT SMP 
dec 22 13:17:33 bb-8 kernel: Modules linked in: sha256_ssse3 sha256_generic hmac drbg ansi_cprng ctr ccm fuse hid_sensor_gyro_3d hid_sensor_magn_3d hid_sensor_accel_3d hid_sensor_incl_3d hid_sensor_rotation hid_se
dec 22 13:17:33 bb-8 kernel:  soundcore glue_helper ablk_helper cryptd input_leds serio_raw rtsx_pci_ms memstick rfkill shpchp i2c_i801 hp_accel led_class lis3lv02d wmi i2c_hid thermal battery input_polldev pinctr
dec 22 13:17:33 bb-8 kernel: CPU: 0 PID: 762 Comm: gnome-shell Tainted: G     U  W       4.2.5-1-ARCH #1
dec 22 13:17:33 bb-8 kernel: Hardware name: HP HP Spectre x360 Convertible /804E, BIOS F.2A 09/18/2015
dec 22 13:17:33 bb-8 kernel: task: ffff88007d360000 ti: ffff88025200c000 task.ti: ffff88025200c000
dec 22 13:17:33 bb-8 kernel: RIP: 0010:[<0000000000000000>]  [<          (null)>]           (null)
dec 22 13:17:33 bb-8 kernel: RSP: 0018:ffff88025200fac0  EFLAGS: 00010282
dec 22 13:17:33 bb-8 kernel: RAX: ffff8800859b43a0 RBX: 0000000000000001 RCX: 00000000fffff116
dec 22 13:17:33 bb-8 kernel: RDX: 00000000fffff116 RSI: ffff8800859b4398 RDI: ffff8800859b1f28
dec 22 13:17:33 bb-8 kernel: RBP: ffff88025200fb48 R08: 0000000000000000 R09: ffff880072172600
dec 22 13:17:33 bb-8 kernel: R10: ffff880255d90800 R11: 00000000000000c0 R12: ffff8800859b1f28
dec 22 13:17:33 bb-8 kernel: R13: ffff8800859b4398 R14: 0000000000000000 R15: ffff880259b1e9c0
dec 22 13:17:33 bb-8 kernel: FS:  00007f505a2e39c0(0000) GS:ffff880273c00000(0000) knlGS:0000000000000000
dec 22 13:17:33 bb-8 kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
dec 22 13:17:33 bb-8 kernel: CR2: 0000000000000000 CR3: 000000007d3da000 CR4: 00000000003406f0
dec 22 13:17:33 bb-8 kernel: Stack:
dec 22 13:17:33 bb-8 kernel:  ffffffffa00e4725 000000000000038c ffff8800859b43a0 fffff11600000002
dec 22 13:17:33 bb-8 kernel:  ffff880268c05440 ffff880259b1e9c0 0000000002691000 0000000000000004
dec 22 13:17:33 bb-8 kernel:  0000000000000019 00000000000005a4 000000009e006ef4 0000000200000002
dec 22 13:17:33 bb-8 kernel: Call Trace:
dec 22 13:17:33 bb-8 kernel:  [<ffffffffa00e4725>] ? i915_gem_object_sync+0x125/0x270 [i915]
dec 22 13:17:33 bb-8 kernel:  [<ffffffffa0100316>] intel_execlists_submission+0x286/0x530 [i915]
dec 22 13:17:33 bb-8 kernel:  [<ffffffffa00d9c89>] i915_gem_do_execbuffer.isra.13+0x9e9/0x1190 [i915]
dec 22 13:17:33 bb-8 kernel:  [<ffffffff813df58c>] ? __pm_runtime_resume+0x4c/0x60
dec 22 13:17:33 bb-8 kernel:  [<ffffffff811b309e>] ? __kmalloc+0x2e/0x270
dec 22 13:17:33 bb-8 kernel:  [<ffffffffa00db587>] i915_gem_execbuffer2+0xd7/0x230 [i915]
dec 22 13:17:33 bb-8 kernel:  [<ffffffffa0005575>] drm_ioctl+0x125/0x610 [drm]
dec 22 13:17:33 bb-8 kernel:  [<ffffffffa00db4b0>] ? i915_gem_execbuffer+0x320/0x320 [i915]
dec 22 13:17:33 bb-8 kernel:  [<ffffffff811ecc67>] ? __fget+0x77/0xb0
dec 22 13:17:33 bb-8 kernel:  [<ffffffff811e2b65>] do_vfs_ioctl+0x295/0x480
dec 22 13:17:33 bb-8 kernel:  [<ffffffff811ecc67>] ? __fget+0x77/0xb0
dec 22 13:17:33 bb-8 kernel:  [<ffffffff811e2dc9>] SyS_ioctl+0x79/0x90
dec 22 13:17:33 bb-8 kernel:  [<ffffffff8157626e>] entry_SYSCALL_64_fastpath+0x12/0x71
dec 22 13:17:33 bb-8 kernel: Code:  Bad RIP value.
dec 22 13:17:33 bb-8 kernel: RIP  [<          (null)>]           (null)

Without the semaphore, I get a more stable experience but no log message on crash.

Any ideas?
Best regards
/Balder

Last edited by cellisten (2016-01-21 07:19:43)

ewaller · 2015-12-22 15:45:55

Can you ping the box?
Are your keyboard LEDs flashing?
Have you tried enabling Magic SysRq keys? Do they work?

You sound like you are technically competent., so you might try one of my favorite tricks. You stated that other consoles are not available. You might use the technique givern here for forwarding the journal to tt12, but instead of forwarding it to /dev/tty12, send it to a serial port, like /dev/ttyS0 or /dev/ttyUSB0. Connect that serial port to another computer and capture the output in a serial terminal emulator (minicom, screen, emacs serial-term mode). When your computer crashes, things that have been sent to the journal will be safely captured on the other machine. Sometimes, at the crash, journal messages may be enqueued waiting to be written to disk -- a task that never happens, but those data may be sent over the serial link before the system goes Tango-Uniform.

cellisten · 2015-12-23 13:56:31

Hello,
thank you for your tip. I will see if I can send it to ttyUSB0 and look at the logs from somewhere else. I have a chromebook available so maybe I can figure out something from there.
I thought of another thing to try, I will enable the intel microcode upgrades. (done that, no difference) If you think of anything else to try, please give me a shout.

Today I managed to get a crash while watching youtube, the sound kept repeating in the background but everything else was frozen. Maybe that can be a clue?!?

I will update the thread as I go along.

Last edited by cellisten (2015-12-23 14:13:04)

ewaller · 2015-12-23 14:52:42

cellisten wrote:

I thought of another thing to try, I will enable the intel microcode upgrades.

*facepalm*
I should have mentioned that. I am surprised it made no difference. Are you sure it is configured correctly?
What flavor processor is it?

TheSgtBilko · 2015-12-23 17:21:40

I have an HP Spectre x360 (Intel i5-6200U).

To be able to start X, I had to add: "i915.preliminary_hw_support=1"
To get rid of the random freezes, I had to add: "intel_pstate=disable"

These issues will hopefully be fixed with newer kernels/microcode, making these two temporary params redundant.

Depending on your bootloader, look up how to add these there.

I'm using GRUB (edit /etc/default/grub and then rebuild grub)

...
GRUB_CMDLINE_LINUX="intel_pstate=disable i915.preliminary_hw_support=1"
...

cellisten · 2015-12-24 09:50:14

Thank you for the tip TheSgtBilko!
The computer is at work during the holidays but I will make sure to add intel_pstate=disable and hopefully it will work.
How did you figure it out?

Best regards
/Balder

cellisten · 2015-12-29 07:16:32

I have now tried with intel_pstate=disable and first I thought it worked but then the crash reoccured. Is it sensitive to order between the preliminary hw support and the pstate?
I am also getting several drm:check_crtc_state errors.

[drm:check_crtc_state [i915]] *ERROR* mismatch in base.adjusted_mode.crtc_clock (expected 138780, found 92519)
kernel: ------------[ cut here ]------------
kernel: WARNING: CPU: 0 PID: 705 at drivers/gpu/drm/i915/intel_display.c:12324 check_crtc_state+0x8df/0xf80 [i915]()
kernel: pipe state doesn't match!
kernel: Modules linked in: sha256_ssse3 sha256_generic hmac drbg ansi_cprng ctr ccm fuse arc4 intel_rapl iosf_mbi x86_pkg_temp_thermal coretemp iwlmvm hi
kernel:  hp_accel led_class pinctrl_intel lis3lv02d i2c_hid usbhid input_polldev tpm_crb hp_wireless tpm hid acpi_als ac evdev kfifo_buf acpi_pad mac_hid
kernel: CPU: 0 PID: 705 Comm: Xorg Tainted: G        W       4.2.5-1-ARCH #1
kernel: Hardware name: HP HP Spectre x360 Convertible /804E, BIOS F.2A 09/18/2015
kernel:  0000000000000000 00000000d40ae571 ffff880256883788 ffffffff81570d0a
kernel:  0000000000000000 ffff8802568837e0 ffff8802568837c8 ffffffff810748a6
kernel:  ffff880256883870 ffff880268694b50 ffff880269fe6000 ffff880268694800
kernel: Call Trace:
kernel:  [<ffffffff81570d0a>] dump_stack+0x4c/0x6e
kernel:  [<ffffffff810748a6>] warn_slowpath_common+0x86/0xc0

Last edited by cellisten (2015-12-29 07:31:06)

cellisten · 2016-01-21 07:18:46

Since the last couple of kernel updates I don't seem to be getting the crashes anymore. I have some other mysterious phenomena but I will probably post about them some other day.

Arch Linux

#1 2015-12-22 11:59:27

[SOLVED] hp Spectre x360 freezing without warning

#2 2015-12-22 15:45:55

Re: [SOLVED] hp Spectre x360 freezing without warning

#3 2015-12-23 13:56:31

Re: [SOLVED] hp Spectre x360 freezing without warning

#4 2015-12-23 14:52:42

Re: [SOLVED] hp Spectre x360 freezing without warning

#5 2015-12-23 17:21:40

Re: [SOLVED] hp Spectre x360 freezing without warning

#6 2015-12-24 09:50:14

Re: [SOLVED] hp Spectre x360 freezing without warning

#7 2015-12-29 07:16:32

Re: [SOLVED] hp Spectre x360 freezing without warning

#8 2016-01-21 07:18:46

Re: [SOLVED] hp Spectre x360 freezing without warning

Board footer