[Intel-gfx] ✗ Fi.CI.BAT: failure for series starting with [1/3] drm/i915/DMC/GLK: Load DMC on GLK

[Ander Conselvan De Oliveira]

On Tue, 2016-12-20 at 14:46 +0000, Chris Wilson wrote:
> On Tue, Dec 20, 2016 at 04:34:20PM +0200, Ander Conselvan De Oliveira wrote:
> > 
> > On Tue, 2016-12-20 at 12:53 +0000, Patchwork wrote:
> > > 
> > > == Series Details ==
> > > 
> > > Series: series starting with [1/3] drm/i915/DMC/GLK: Load DMC on GLK
> > > URL   : https://patchwork.freedesktop.org/series/16926/
> > > State : failure
> > > 
> > > == Summary ==
> > > 
> > > Series 16926v1 Series without cover letter
> > > https://patchwork.freedesktop.org/api/1.0/series/16926/revisions/1/mbox/
> > > 
> > > Test drv_module_reload:
> > >         Subgroup basic-reload-inject:
> > >                 pass       -> DMESG-WARN (fi-kbl-7500u)
> > [   33.108496] WARNING: CPU: 2 PID: 6400 at
> > drivers/gpu/drm/i915/i915_gem.c:4254 i915_gem_suspend+0x181/0x190 [i915]
> > [   33.108498] WARN_ON(dev_priv->gt.awake)
> > [   33.108499] Modules linked in:
> > [   33.108501]  i915(-) x86_pkg_temp_thermal intel_powerclamp coretemp
> > crct10dif_pclmul crc32_pclmul ghash_clmulni_intel snd_hda_codec_hdmi
> > snd_hda_codec_realtek snd_hda_codec_generic snd_hda_codec snd_hwdep mei_me
> > snd_hda_core mei snd_pcm e1000e ptp pps_core i2c_hid [last unloaded:
> > snd_hda_intel]
> > [   33.108519] CPU: 2 PID: 6400 Comm: drv_module_relo Tainted:
> > G     U          4.9.0-CI-Patchwork_3338+ #1
> > [   33.108521] Hardware name: GIGABYTE GB-BKi7A-7500/MFLP7AP-00, BIOS F1
> > 07/27/2016
> > [   33.108522]  ffffc90000547d18 ffffffff81435b05 ffffc90000547d68
> > 0000000000000000
> > [   33.108527]  ffffc90000547d58 ffffffff8107e4d6 0000109e00000000
> > ffff88025c620000
> > [   33.108530]  0000000000000000 ffff88025c620068 ffffffffa013f4c0
> > 0000000000000000
> > [   33.108534] Call Trace:
> > [   33.108540]  [<ffffffff81435b05>] dump_stack+0x67/0x92
> > [   33.108543]  [<ffffffff8107e4d6>] __warn+0xc6/0xe0
> > [   33.108547]  [<ffffffff8107e53a>] warn_slowpath_fmt+0x4a/0x50
> > [   33.108572]  [<ffffffffa0051601>] i915_gem_suspend+0x181/0x190 [i915]
> > [   33.108589]  [<ffffffffa000ef4e>] i915_driver_unload+0x1e/0x190 [i915]
> > [   33.108608]  [<ffffffffa0018b24>] i915_pci_remove+0x14/0x20 [i915]
> > [   33.108611]  [<ffffffff81489aa4>] pci_device_remove+0x34/0xb0
> > [   33.108614]  [<ffffffff81588eec>] __device_release_driver+0x9c/0x150
> > [   33.108617]  [<ffffffff81589a66>] driver_detach+0xb6/0xc0
> > [   33.108619]  [<ffffffff81588983>] bus_remove_driver+0x53/0xd0
> > [   33.108621]  [<ffffffff8158a527>] driver_unregister+0x27/0x50
> > [   33.108623]  [<ffffffff81488445>] pci_unregister_driver+0x25/0x70
> > [   33.108652]  [<ffffffffa00fd8a4>] i915_exit+0x1a/0x71 [i915]
> > [   33.108655]  [<ffffffff8111a813>] SyS_delete_module+0x193/0x1e0
> > [   33.108658]  [<ffffffff81827c6e>] entry_SYSCALL_64_fastpath+0x1c/0xb1
> > [   33.108660] ---[ end trace 247dfc1ad299c432 ]---
> > 
> > https://bugs.freedesktop.org/show_bug.cgi?id=98670
> > 
> > > 
> > > Test pm_rpm:
> > >         Subgroup basic-pci-d3-state:
> > >                 pass       -> FAIL       (fi-kbl-7500u)
> > >         Subgroup basic-rte:
> > >                 pass       -> FAIL       (fi-kbl-7500u)
> > Out	
> > IGT-Version: 1.17-g38bffd9 (x86_64) (Linux: 4.9.0-CI-Patchwork_3338+ x86_64)
> > Runtime PM support: 1
> > PC8 residency support: 0
> > Stack trace:
> >   #0 [__igt_fail_assert+0x101]
> >   #1 [main+0x2514]
> >   #2 [__libc_start_main+0xf0]
> >   #3 [_start+0x29]
> >   #4 [<unknown>+0x29]
> > Subtest basic-rte: FAIL (10.053s)
> > Err	
> > (pm_rpm:9532) CRITICAL: Test assertion failure function basic_subtest, file
> > pm_rpm.c:705:
> > (pm_rpm:9532) CRITICAL: Failed assertion: wait_for_suspended()
> > Subtest basic-rte failed.
> > **** DEBUG ****
> > (pm_rpm:9532) CRITICAL: Test assertion failure function basic_subtest, file
> > pm_rpm.c:705:
> > (pm_rpm:9532) CRITICAL: Failed assertion: wait_for_suspended()
> > 
> > Is this related to the gt.awake WARN? Looking at CI history, it seems the
> > three
> > tests always fail together.
> No. The message is from an earlier incarnation of the module (i.e. if it
> does correlates, it is a bug in module loading not sanitizing hw state),
> and https://intel-gfx-ci.01.org/CI/fi-kbl-7500u.html disagrees with you.

I shouldn't have said always. I was looking at that page, but only considered
the fails of basic-reload-inject that had a dmesg-warn type of fail. And that is
very small sample of 2. Anyway, thanks for the explanation.

Ander