Bug 93190

Summary: [BAT HSW] hsw-brixbox fails fdi link training
Product: DRI Reporter: Daniel Vetter <daniel>
Component: DRM/IntelAssignee: cprigent <christophe.prigent>
Status: CLOSED FIXED QA Contact: Intel GFX Bugs mailing list <intel-gfx-bugs>
Severity: critical    
Priority: highest CC: intel-gfx-bugs
Version: XOrg git   
Hardware: Other   
OS: All   
Whiteboard:
i915 platform: HSW i915 features:

Description Daniel Vetter 2015-12-01 09:19:02 UTC 
Relevant snippet from dmesg:

 	

[  134.706727] [drm:hsw_fdi_link_train [i915]] *ERROR* FDI link training failed!
[  134.706905] ------------[ cut here ]------------
[  134.706916] WARNING: CPU: 0 PID: 5945 at drivers/gpu/drm/i915/intel_display.c:1249 assert_fdi_rx+0x68/0xa0 [i915]()
[  134.706917] FDI RX state assertion failure (expected on, current off)
[  134.706918] Modules linked in: snd_hda_codec_hdmi x86_pkg_temp_thermal intel_powerclamp coretemp crct10dif_pclmul crc32_pclmul snd_hda_codec_realtek snd_hda_codec_generic i915 snd_hda_intel snd_hda_codec snd_hwdep snd_hda_core i2c_hid mei_me snd_pcm mei i2c_designware_platform i2c_designware_core lpc_ich r8169 mii sdhci_acpi sdhci mmc_core
[  134.706932] CPU: 0 PID: 5945 Comm: kms_flip Tainted: G     U  W       4.4.0-rc2-gfxbench+ #1
[  134.706932] Hardware name: GIGABYTE M4HM87P-00/M4HM87P-00, BIOS F6 12/10/2014
[  134.706933]  ffffffffa02ac6d0 ffff880212d83ab0 ffffffff813df18c ffff880212d83af8
[  134.706935]  ffff880212d83ae8 ffffffff810747b1 0000000000000001 ffff880210620000
[  134.706937]  ffff8800da5c4520 ffff880210364520 0000000000000000 ffff880212d83b48
[  134.706939] Call Trace:
[  134.706943]  [<ffffffff813df18c>] dump_stack+0x4e/0x82
[  134.706945]  [<ffffffff810747b1>] warn_slowpath_common+0x81/0xc0
[  134.706947]  [<ffffffff81074837>] warn_slowpath_fmt+0x47/0x50
[  134.706956]  [<ffffffffa022c0a8>] assert_fdi_rx+0x68/0xa0 [i915]
[  134.706966]  [<ffffffffa0243acb>] haswell_crtc_enable+0x67b/0x8b0 [i915]
[  134.706975]  [<ffffffffa023f075>] intel_atomic_commit+0x695/0x1810 [i915]
[  134.706977]  [<ffffffff8150d042>] ? drm_atomic_check_only+0x142/0x5e0
[  134.706979]  [<ffffffff8150c92b>] ? drm_atomic_set_crtc_for_connector+0x1b/0x90
[  134.706980]  [<ffffffff8150d512>] drm_atomic_commit+0x32/0x50
[  134.706982]  [<ffffffff814ea305>] drm_atomic_helper_set_config+0x75/0xb0
[  134.706984]  [<ffffffff814fc110>] drm_mode_set_config_internal+0x60/0x110
[  134.706985]  [<ffffffff81500d26>] drm_mode_setcrtc+0x186/0x4e0
[  134.706987]  [<ffffffff814f306d>] drm_ioctl+0x13d/0x590
[  134.706988]  [<ffffffff81500ba0>] ? drm_mode_setplane+0x1b0/0x1b0
[  134.706990]  [<ffffffff811d4d9c>] do_vfs_ioctl+0x2fc/0x550
[  134.706992]  [<ffffffff8118d60a>] ? vm_munmap+0x4a/0x60
[  134.706993]  [<ffffffff811e080a>] ? __fget_light+0x6a/0x90
[  134.706995]  [<ffffffff811d502c>] SyS_ioctl+0x3c/0x70
[  134.706997]  [<ffffffff81797e1b>] entry_SYSCALL_64_fastpath+0x16/0x73
[  134.706998] ---[ end trace 08eedc206766a1be ]---
[  134.708255] [drm:intel_set_pch_fifo_underrun_reporting [i915]] *ERROR* uncleared pch fifo underrun on pch transcoder A
[  134.708268] [drm:intel_pch_fifo_underrun_irq_handler [i915]] *ERROR* PCH transcoder A FIFO underrun

End result is that the pipe is dead, resulting in tons of testcases to fail. So until this FDI issue is addressed there's probably no point in trying to look at specific testcase results.

Testcases from the BAT run which seem to be affected by this are:

igt@kms_flip@basic-*
igt@kms_pipe_crc_basic@*
Comment 1 Daniel Vetter 2015-12-01 09:19:27 UTC 
High severity since it prevents a bunch of testcases from running.
Comment 2 Ville Syrjala 2015-12-01 12:20:00 UTC 
I have a fix (not registering the crt connector).
Comment 3 Kimmo Nikkanen 2015-12-04 12:34:33 UTC 
Fix for this in nightly. Assigning to QA for validation.
Comment 4 Daniel Vetter 2015-12-08 09:32:18 UTC 
General process is to mark a bug as resolved when the fixed landed. QA then either moves to "CLOSED" or reopens.

We need to get this bug handling bkms straigthened out again ...
Comment 5 cprigent 2016-02-01 11:12:24 UTC 
kms_flip@basic-flip-vs-* and kms_pipe_crc_basic@* are all Pass.
I don't see such FDI link training error in kernel log.

Tested with:
Hardware: 
Motherboard: SawTooth Peak 
cpu model name : Intel(R) Core(TM) i7-4550U CPU @ 1.50GHz 
cpu model : 69 
cpu family : 6 
Graphic card: Haswell-ULT Integrated Graphics Controller (rev 09) 
Software: 
Kernel: 4.4.0-nightly 9e561f8 from http://cgit.freedesktop.org/drm-intel/
  commit 9e561f8adcedd2812675f6209fc16d94501333b2
  Author: Daniel Vetter <daniel.vetter@ffwll.ch>
  Date:   Sun Jan 24 22:49:47 2016 +0100
  drm-intel-nightly: 2016y-01m-24d-21h-49m-20s UTC integration manifest
Bios: HSWLPTU1.86C.0135.R01.1311020052
Libdrm: 2.4.66 from http://cgit.freedesktop.org/mesa/drm/

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.