Bug 62931

Summary:	drm/i915: regression in LONGTERM kernel gpu driver > kernel 3.2.46
Product:	Drivers	Reporter:	tom schorpp (thomas.schorpp)
Component:	Video(DRI - Intel)	Assignee:	intel-gfx-bugs (intel-gfx-bugs)
Status:	CLOSED UNREPRODUCIBLE
Severity:	low	CC:	daniel, intel-gfx-bugs, thomas.schorpp
Priority:	P1
Hardware:	x86-64
OS:	Linux
Kernel Version:	3.2.51	Subsystem:
Regression:	Yes	Bisected commit-id:

Description tom schorpp 2013-10-13 12:31:52 UTC

Drivers in kernel versions 3.2.46...3.2.51 are corrupting video memory and crashing GPU on 

00:02.0 VGA compatible controller: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 0c)
00:02.1 Display controller: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 0c)

and

00:02.0 VGA compatible controller: Intel Corporation Mobile 915GM/GMS/910GML Express Graphics Controller (rev 04)
00:02.1 Display controller: Intel Corporation Mobile 915GM/GMS/910GML Express Graphics Controller (rev 04)

after some time intermittently with blank screen on GM965 and kernel stalls on 915GM, 

must be one of those changes:

2013-09-10	drm/i915: ivb: fix edp voltage swing reg val	Imre Deak	1	-1/+1
2013-09-10	drm/i915: Invalidate TLBs for the rings after a reset	Chris Wilson	2	-0/+15
2013-09-10	drm/i915/lvds: ditch ->prepare special case	Daniel Vetter	1	-7/+1
2013-09-10	drm/i915: quirk no PCH_PWM_ENABLE for Dell XPS13 backlight	Kamal Mostafa	3	-1/+19
2013-08-02	ipv6: ip6_append_data_mtu did not care about pmtudisc and frag_size	Hannes Frederic Sowa	1	-6/+10
2013-06-29	Revert "drm/i915: GFX_MODE Flush TLB Invalidate Mode must be '1' for scanline waits"	Ben Hutchings	1	-5/+0
2013-06-19	drm/i915: prefer VBT modes for SVDO-LVDS over EDID	Daniel Vetter	1	-4/+6
2013-06-19	drm/i915/sdvo: Use &intel_sdvo->ddc instead of intel_sdvo->i2c for DDC.	Egbert Eich	1	-1/+1
2013-06-19	drm/i915: no lvds quirk for hp t5740	Ben Mesman	1	-2/+2

2013-08-02	Linux 3.2.50v3.2.50	Ben Hutchings	1	-1/+1
2013-07-27	Linux 3.2.49v3.2.49	Ben Hutchings	1	-1/+1
2013-06-29	Linux 3.2.48v3.2.48	Ben Hutchings	1	-1/+1
2013-06-19	Linux 3.2.47v3.2.47	Ben Hutchings	1	-1/+1
2013-05-30	Linux 3.2.46v3.2.46	Ben Hutchings	1	-1/+1

dmesg: zcat /var/log/*.gz |grep -i i915

Sep 27 20:45:36 tom3 kernel: [1305472.172086] [drm:i915_hangcheck_elapsed] *ERROR* Hangcheck timer elapsed... GPU hung
Sep 27 20:45:36 tom3 kernel: [1305472.172372] [drm:i915_wait_request] *ERROR* i915_wait_request returns -11 (awaiting 50676045 at 50676030, next 50676046)
Sep 27 20:45:36 tom3 kernel: [1305472.261541] WARNING: at drivers/gpu/drm/i915/intel_display.c:812 intel_enable_pipe+0x136/0x160()
Sep 27 20:45:36 tom3 kernel: [1305472.261931]  [<ffffffff812fdc7f>] i915_reset+0x35f/0x430
Sep 27 20:45:36 tom3 kernel: [1305472.261937]  [<ffffffff81303217>] i915_error_work_func+0xd7/0x120
Sep 27 20:45:36 tom3 kernel: [1305472.366211] WARNING: at drivers/gpu/drm/i915/intel_display.c:949 assert_pipe+0x77/0x80()
Sep 27 20:45:36 tom3 kernel: [1305472.366866]  [<ffffffff812fdc7f>] i915_reset+0x35f/0x430
Sep 27 20:45:36 tom3 kernel: [1305472.366879]  [<ffffffff81303217>] i915_error_work_func+0xd7/0x120
Sep 27 20:45:36 tom3 kernel: [1305472.470546] WARNING: at drivers/gpu/drm/i915/intel_display.c:812 intel_enable_pipe+0x136/0x160()
Sep 27 20:45:36 tom3 kernel: [1305472.470930]  [<ffffffff812fdc7f>] i915_reset+0x35f/0x430
Sep 27 20:45:36 tom3 kernel: [1305472.470939]  [<ffffffff81303217>] i915_error_work_func+0xd7/0x120
Sep 27 20:45:36 tom3 kernel: [1305472.523139] WARNING: at drivers/gpu/drm/i915/intel_display.c:949 assert_pipe+0x77/0x80()
Sep 27 20:45:36 tom3 kernel: [1305472.523767]  [<ffffffff812fdc7f>] i915_reset+0x35f/0x430
Sep 27 20:45:36 tom3 kernel: [1305472.523780]  [<ffffffff81303217>] i915_error_work_func+0xd7/0x120
Sep 27 20:45:37 tom3 kernel: [1305473.710557] [drm:i915_wait_request] *ERROR* something (likely vbetool) disabled interrupts, re-enabling
Sep 27 20:45:37 tom3 kernel: [1305473.730182] [drm:i915_wait_request] *ERROR* something (likely vbetool) disabled interrupts, re-enabling
Sep 27 20:45:37 tom3 kernel: [1305473.750185] [drm:i915_wait_request] *ERROR* something (likely vbetool) disabled interrupts, re-enabling
Sep 27 20:45:39 tom3 kernel: [1305475.210262] [drm:i915_wait_request] *ERROR* something (likely vbetool) disabled interrupts, re-enabling
Sep 27 20:45:39 tom3 kernel: [1305475.212060] [drm:i915_hangcheck_elapsed] *ERROR* Hangcheck timer elapsed... GPU hung
Sep 27 20:45:39 tom3 kernel: [1305475.212115] [drm:i915_wait_request] *ERROR* i915_wait_request returns -11 (awaiting 50676072 at 50676030, next 50676073)
Sep 27 20:45:39 tom3 kernel: [1305475.212590] [drm:i915_reset] *ERROR* GPU hanging too fast, declaring wedged!
Sep 27 20:45:39 tom3 kernel: [1305475.212603] [drm:i915_reset] *ERROR* Failed to reset chip.
Sep 27 21:02:48 tom3 kernel: [    6.556398] i915 0000:00:02.0: power state changed by ACPI to D0
Sep 27 21:02:48 tom3 kernel: [    6.556468] i915 0000:00:02.0: power state changed by ACPI to D0
Sep 27 21:02:48 tom3 kernel: [    6.556921] i915 0000:00:02.0: setting latency timer to 64
Sep 27 21:02:48 tom3 kernel: [    6.600819] i915 0000:00:02.0: irq 48 for MSI/MSI-X
Sep 27 21:02:48 tom3 kernel: [    7.279273] [drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0

Sep 27 20:45:28 tom3 kernel: [1305464.696387] [drm] capturing error event; look for more information in /debug/dri/0/i915_error_state
Sep 27 20:45:28 tom3 kernel: [1305464.783588] WARNING: at drivers/gpu/drm/i915/intel_display.c:812 intel_enable_pipe+0x136/0x160()
Sep 27 20:45:28 tom3 kernel: [1305464.783853]  [<ffffffff812fdc7f>] i915_reset+0x35f/0x430
Sep 27 20:45:28 tom3 kernel: [1305464.783858]  [<ffffffff81303217>] i915_error_work_func+0xd7/0x120
Sep 27 20:45:28 tom3 kernel: [1305464.886119] WARNING: at drivers/gpu/drm/i915/intel_display.c:949 assert_pipe+0x77/0x80()
Sep 27 20:45:28 tom3 kernel: [1305464.886437]  [<ffffffff812fdc7f>] i915_reset+0x35f/0x430
Sep 27 20:45:28 tom3 kernel: [1305464.886443]  [<ffffffff81303217>] i915_error_work_func+0xd7/0x120
Sep 27 20:45:29 tom3 kernel: [1305464.989594] WARNING: at drivers/gpu/drm/i915/intel_display.c:812 intel_enable_pipe+0x136/0x160()
Sep 27 20:45:29 tom3 kernel: [1305464.990214]  [<ffffffff812fdc7f>] i915_reset+0x35f/0x430
Sep 27 20:45:29 tom3 kernel: [1305464.990227]  [<ffffffff81303217>] i915_error_work_func+0xd7/0x120
Sep 27 20:45:29 tom3 kernel: [1305465.041088] WARNING: at drivers/gpu/drm/i915/intel_display.c:949 assert_pipe+0x77/0x80()
Sep 27 20:45:29 tom3 kernel: [1305465.041409]  [<ffffffff812fdc7f>] i915_reset+0x35f/0x430
Sep 27 20:45:29 tom3 kernel: [1305465.041414]  [<ffffffff81303217>] i915_error_work_func+0xd7/0x120
Sep 27 20:45:36 tom3 kernel: [1305472.261541] WARNING: at drivers/gpu/drm/i915/intel_display.c:812 intel_enable_pipe+0x136/0x160()
Sep 27 20:45:36 tom3 kernel: [1305472.261931]  [<ffffffff812fdc7f>] i915_reset+0x35f/0x430
Sep 27 20:45:36 tom3 kernel: [1305472.261937]  [<ffffffff81303217>] i915_error_work_func+0xd7/0x120
Sep 27 20:45:36 tom3 kernel: [1305472.366211] WARNING: at drivers/gpu/drm/i915/intel_display.c:949 assert_pipe+0x77/0x80()
Sep 27 20:45:36 tom3 kernel: [1305472.366866]  [<ffffffff812fdc7f>] i915_reset+0x35f/0x430
Sep 27 20:45:36 tom3 kernel: [1305472.366879]  [<ffffffff81303217>] i915_error_work_func+0xd7/0x120
Sep 27 20:45:36 tom3 kernel: [1305472.470546] WARNING: at drivers/gpu/drm/i915/intel_display.c:812 intel_enable_pipe+0x136/0x160()
Sep 27 20:45:36 tom3 kernel: [1305472.470930]  [<ffffffff812fdc7f>] i915_reset+0x35f/0x430
Sep 27 20:45:36 tom3 kernel: [1305472.470939]  [<ffffffff81303217>] i915_error_work_func+0xd7/0x120
Sep 27 20:45:36 tom3 kernel: [1305472.523139] WARNING: at drivers/gpu/drm/i915/intel_display.c:949 assert_pipe+0x77/0x80()
Sep 27 20:45:36 tom3 kernel: [1305472.523767]  [<ffffffff812fdc7f>] i915_reset+0x35f/0x430
Sep 27 20:45:36 tom3 kernel: [1305472.523780]  [<ffffffff81303217>] i915_error_work_func+0xd7/0x120
Sep 27 21:02:48 tom3 kernel: [    6.556398] i915 0000:00:02.0: power state changed by ACPI to D0
Sep 27 21:02:48 tom3 kernel: [    6.556468] i915 0000:00:02.0: power state changed by ACPI to D0
Sep 27 21:02:48 tom3 kernel: [    7.279273] [drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0

[    2.288741] [drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0
Sep 27 20:45:28 tom3 kernel: [1305464.696052] [drm:i915_hangcheck_elapsed] *ERROR* Hangcheck timer elapsed... GPU hung
Sep 27 20:45:28 tom3 kernel: [1305464.696387] [drm] capturing error event; look for more information in /debug/dri/0/i915_error_state
Sep 27 20:45:28 tom3 kernel: [1305464.697970] [drm:i915_wait_request] *ERROR* i915_wait_request returns -11 (awaiting 50676031 at 50676030, next 50676039)
Sep 27 20:45:28 tom3 kernel: [1305464.783588] WARNING: at drivers/gpu/drm/i915/intel_display.c:812 intel_enable_pipe+0x136/0x160()
Sep 27 20:45:28 tom3 kernel: [1305464.783853]  [<ffffffff812fdc7f>] i915_reset+0x35f/0x430
Sep 27 20:45:28 tom3 kernel: [1305464.783858]  [<ffffffff81303217>] i915_error_work_func+0xd7/0x120
Sep 27 20:45:28 tom3 kernel: [1305464.886119] WARNING: at drivers/gpu/drm/i915/intel_display.c:949 assert_pipe+0x77/0x80()
Sep 27 20:45:28 tom3 kernel: [1305464.886437]  [<ffffffff812fdc7f>] i915_reset+0x35f/0x430
Sep 27 20:45:28 tom3 kernel: [1305464.886443]  [<ffffffff81303217>] i915_error_work_func+0xd7/0x120
Sep 27 20:45:29 tom3 kernel: [1305464.989594] WARNING: at drivers/gpu/drm/i915/intel_display.c:812 intel_enable_pipe+0x136/0x160()
Sep 27 20:45:29 tom3 kernel: [1305464.990214]  [<ffffffff812fdc7f>] i915_reset+0x35f/0x430
Sep 27 20:45:29 tom3 kernel: [1305464.990227]  [<ffffffff81303217>] i915_error_work_func+0xd7/0x120
Sep 27 20:45:29 tom3 kernel: [1305465.041088] WARNING: at drivers/gpu/drm/i915/intel_display.c:949 assert_pipe+0x77/0x80()
Sep 27 20:45:29 tom3 kernel: [1305465.041409]  [<ffffffff812fdc7f>] i915_reset+0x35f/0x430
Sep 27 20:45:29 tom3 kernel: [1305465.041414]  [<ffffffff81303217>] i915_error_work_func+0xd7/0x120
Sep 27 20:45:30 tom3 kernel: [1305466.157807] [drm:i915_wait_request] *ERROR* something (likely vbetool) disabled interrupts, re-enabling
Sep 27 20:45:30 tom3 kernel: [1305466.177150] [drm:i915_wait_request] *ERROR* something (likely vbetool) disabled interrupts, re-enabling
Sep 27 20:45:30 tom3 kernel: [1305466.197164] [drm:i915_wait_request] *ERROR* something (likely vbetool) disabled interrupts, re-enabling

# apt-cache rdepends vbetool
vbetool
Reverse Depends:
  pm-utils
  pm-utils
  hibernate
  acpi-support

Bug not occured during ACPI/PM state changes or screensaver kick in.

# grep -irsn i915 /etc/modprobe.d
/etc/modprobe.d/i915-kms.conf:1:options i915 modeset=1
root@tom3:~# 

915GM machine out of logs, will follow, just trying to trigger the bug, not occured yet after upgrading firmware from debian stable to testing:

ii  firmware-linux                    0.40                  all                   Binary firmware for various drivers in the Linux kernel (meta-package)
ii  firmware-linux-free               3.2                   all                   Binary firmware for various drivers in the Linux kernel
ii  firmware-linux-nonfree            0.40                  all                   Binary firmware for various drivers in the Linux kernel

y
tom

Comment 1 Daniel Vetter 2013-10-14 07:06:32 UTC

A few things ...
- Can you please attach a full dmesg exhibiting these issues?
- You have gpu hangs - to analyze those we need the error state from /sys/kernel/debug/dri/0/i915_error_state. Also gpu hangs are often due to bugs in userspace.
- Can you please confirm the issue on latest released kernels?
- Also a bisect of the few candidate commits might be useful.

Comment 2 tom schorpp 2013-10-14 10:45:13 UTC

I know. But let me reproduce it, first, that may take some days.