Bug 100219 - [SNB] CPU pipe A FIFO underrun
Summary: [SNB] CPU pipe A FIFO underrun
Status: CLOSED DUPLICATE of bug 95736
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: unspecified
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: Intel GFX Bugs mailing list
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard: ReadyForDev
Keywords:
Depends on:
Blocks:
 
Reported: 2017-03-15 18:27 UTC by udo
Modified: 2017-10-06 21:16 UTC (History)
3 users (show)

See Also:
i915 platform: SNB
i915 features: GEM/Other


Attachments
dmesg (53.79 KB, text/plain)
2017-03-15 18:27 UTC, udo
no flags Details
dmesg with drm.debug=0xe (94.62 KB, text/plain)
2017-07-05 16:52 UTC, udo
no flags Details

Description udo 2017-03-15 18:27:01 UTC
Created attachment 130233 [details]
dmesg

The following CPU pipe A underrun can be observed during booting Linux-4.10.X on an Intel SNB processor. It is 100% reproducible during every boot and it occurs before any other software runs.

[    1.071498] [drm] Initialized
[    1.071698] [drm] Disabling PPGTT because VT-d is on
[    1.071714] [drm] Memory usable by graphics device = 2048M
[    1.071717] [drm] VT-d active for gfx access
[    1.071721] [drm] Replacing VGA console driver
[    1.071736] [drm] DMAR active, disabling use of stolen memory
[    1.078093] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
[    1.078099] [drm] Driver supports precise vblank timestamp query.
[    1.151068] [drm] Initialized i915 1.6.0 20161121 for 0000:00:02.0 on minor 0
[    1.365751] fbcon: inteldrmfb (fb0) is primary device
[    2.224399] [drm:intel_cpu_fifo_underrun_irq_handler] *ERROR* CPU pipe A FIFO underrun
[    2.227469] [drm:intel_set_pch_fifo_underrun_reporting] *ERROR* uncleared pch fifo underrun on pch transcoder A
[    2.227478] [drm:cpt_irq_handler] *ERROR* PCH transcoder A FIFO underrun
[    2.414759] i915 0000:00:02.0: fb0: inteldrmfb frame buffer device

Full dmesg attached.
Comment 1 Chris Wilson 2017-03-15 18:38:02 UTC
And does it go away if you disable VT-d?
Comment 2 udo 2017-03-15 19:23:44 UTC
Booted with "intel_iommu=igfx_off" and the pipe underrun goes away.

I'd prefer to keep VT-d enabled though.
Comment 3 Jani Saarinen 2017-05-24 06:23:07 UTC
Chris, what is your suggestion here?
Comment 4 Jani Saarinen 2017-06-08 07:13:25 UTC
Reporter, is this still valid with latest drm-tip?
Comment 5 udo 2017-06-10 18:10:55 UTC
It happens when booting the Linux kernel, so I guess libdrm is not even involved. I tried it with Linux 4.11.3 and drm master b6f450e0.

Result is still the same.

[    1.635652] tsc: Refined TSC clocksource calibration: 2691.258 MHz
[    1.635668] clocksource: tsc: mask: 0xffffffffffffffff max_cycles: 0x26cafc29fa8, max_idle_ns: 440795270837 ns
[    1.886585] [drm:intel_cpu_fifo_underrun_irq_handler] *ERROR* CPU pipe A FIFO underrun
[    1.887064] [drm:intel_set_pch_fifo_underrun_reporting] *ERROR* uncleared pch fifo underrun on pch transcoder A
[    1.887070] [drm:cpt_irq_handler] *ERROR* PCH transcoder A FIFO underrun
Comment 6 Elizabeth 2017-06-27 19:18:21 UTC
(In reply to udo from comment #5)
> It happens when booting the Linux kernel, so I guess libdrm is not even
> involved. I tried it with Linux 4.11.3 and drm master b6f450e0.
> 
> Result is still the same.
> 
> [    1.635652] tsc: Refined TSC clocksource calibration: 2691.258 MHz
> [    1.635668] clocksource: tsc: mask: 0xffffffffffffffff max_cycles:
> 0x26cafc29fa8, max_idle_ns: 440795270837 ns
> [    1.886585] [drm:intel_cpu_fifo_underrun_irq_handler] *ERROR* CPU pipe A
> FIFO underrun
> [    1.887064] [drm:intel_set_pch_fifo_underrun_reporting] *ERROR* uncleared
> pch fifo underrun on pch transcoder A
> [    1.887070] [drm:cpt_irq_handler] *ERROR* PCH transcoder A FIFO underrun

Hello, I'm changing the status to REOPEN since the problem seems to persist. Thank you.
Comment 7 Elizabeth 2017-07-05 15:55:04 UTC
(In reply to udo from comment #5)
> It happens when booting the Linux kernel, so I guess libdrm is not even
> 
Hello,
Could you please attach new dmesg with parameter drm.debug=0xe on grub, with the latest kernel? Thank you.
Comment 8 udo 2017-07-05 16:52:26 UTC
Created attachment 132460 [details]
dmesg with drm.debug=0xe

Attached dmesg with drm.debug=0xe. The problem occurs at time 1.71
Comment 9 Elizabeth 2017-10-06 20:04:36 UTC
Seems same problem had already been reported at bug 95736, please follow this issue in that bug.

*** This bug has been marked as a duplicate of bug 95736 ***


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.