Bug 107588

Revealed by

commit 08ea70a417baa1c0e5faa580070cd2ffd04e6285
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Sun Aug 12 23:36:31 2018 +0100

    drm/i915: Disable runtime-pm using lowlevel functions if !HAS_RC6

it's a very old bug (since we introduced runtime-pm tracking many, many years ago) that I've sent fixes for. Will send them again when we have the current wakeref bugs under wraps.

Also seen on igt@debugfs_test@read_all_entries_display_off.

https://intel-gfx-ci.01.org/tree/drm-tip/drmtip_99/fi-ilk-650/igt@debugfs_test@read_all_entries_display_off.html

[  227.282219] ------------[ cut here ]------------
[  227.282283] RPM wakelock ref not held during HW access
[  227.282375] WARNING: CPU: 0 PID: 1509 at drivers/gpu/drm/i915/intel_drv.h:2009 gen5_read32+0x16b/0x1a0 [i915]
[  227.282378] Modules linked in: vgem snd_hda_codec_hdmi snd_hda_codec_realtek snd_hda_codec_generic i915 coretemp crct10dif_pclmul crc32_pclmul ghash_clmulni_intel lpc_ich snd_hda_intel snd_hda_codec snd_hwdep snd_hda_core snd_pcm mei_me e1000e mei prime_numbers
[  227.282453] CPU: 0 PID: 1509 Comm: debugfs_test Tainted: G     U  W         4.19.0-rc1-gb4059d5fdbfd-drmtip_99+ #1
[  227.282458] Hardware name: Hewlett-Packard HP Compaq 8100 Elite SFF PC/304Ah, BIOS 786H1 v01.13 07/14/2011
[  227.282522] RIP: 0010:gen5_read32+0x16b/0x1a0 [i915]
[  227.282527] Code: 3e cd f0 0f 0b e9 d2 fe ff ff 80 3d f8 1d 19 00 00 0f 85 d3 fe ff ff 48 c7 c7 58 20 4b c0 c6 05 e4 1d 19 00 01 e8 65 3e cd f0 <0f> 0b e9 b9 fe ff ff e8 79 57 d5 f0 85 c0 75 92 48 c7 c2 88 20 4b
[  227.282531] RSP: 0018:ffffb25480d67d38 EFLAGS: 00010282
[  227.282538] RAX: 0000000000000000 RBX: ffff8de1f7e00000 RCX: 0000000000000001
[  227.282542] RDX: 0000000080000001 RSI: ffffffffb2080d20 RDI: 00000000ffffffff
[  227.282546] RBP: 0000000000011020 R08: 00000000b653baf4 R09: 0000000000000000
[  227.282549] R10: ffffb25480d67d80 R11: 0000000000000000 R12: 0000000000000001
[  227.282553] R13: ffff8de1f7e00068 R14: 0000000000000001 R15: ffff8de20e137488
[  227.282558] FS:  00007f6a919b6980(0000) GS:ffff8de213a00000(0000) knlGS:0000000000000000
[  227.282562] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  227.282566] CR2: 000055d4953490e8 CR3: 00000001a1df4004 CR4: 00000000000206f0
[  227.282569] Call Trace:
[  227.282633]  i915_mch_val+0x1b/0x60 [i915]
[  227.282684]  i915_emon_status+0x45/0xd0 [i915]
[  227.282697]  seq_read+0xdb/0x3c0
[  227.282707]  ? lockdep_hardirqs_off+0x94/0xd0
[  227.282714]  ? __slab_free+0x253/0x4f0
[  227.282726]  full_proxy_read+0x51/0x80
[  227.282735]  __vfs_read+0x31/0x170
[  227.282748]  ? rcu_lockdep_current_cpu_online+0x8f/0xd0
[  227.282755]  ? do_sys_open+0x13b/0x240
[  227.282760]  ? rcu_read_lock_sched_held+0x6f/0x80
[  227.282767]  vfs_read+0x9e/0x140
[  227.282775]  ksys_read+0x50/0xc0
[  227.282787]  do_syscall_64+0x55/0x190
[  227.282795]  entry_SYSCALL_64_after_hwframe+0x49/0xbe
[  227.282800] RIP: 0033:0x7f6a9134b34e
[  227.282805] Code: 00 00 00 00 48 8b 15 71 8c 20 00 f7 d8 64 89 02 48 c7 c0 ff ff ff ff c3 0f 1f 40 00 8b 05 ba d0 20 00 85 c0 75 16 31 c0 0f 05 <48> 3d 00 f0 ff ff 77 5a f3 c3 0f 1f 84 00 00 00 00 00 41 54 55 49
[  227.282808] RSP: 002b:00007ffd37113318 EFLAGS: 00000246 ORIG_RAX: 0000000000000000
[  227.282816] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f6a9134b34e
[  227.282820] RDX: 0000000000000200 RSI: 00007ffd37113360 RDI: 0000000000000006
[  227.282825] RBP: 00007ffd37113570 R08: 0000000000000000 R09: 0000000000000020
[  227.282828] R10: 0000000000000000 R11: 0000000000000246 R12: 000055d49510a2c0
[  227.282832] R13: 00007ffd37113a60 R14: 0000000000000000 R15: 0000000000000000
[  227.282852] irq event stamp: 54538
[  227.282859] hardirqs last  enabled at (54537): [<ffffffffb10fbb46>] console_unlock+0x1b6/0x5e0
[  227.282864] hardirqs last disabled at (54538): [<ffffffffb1001910>] trace_hardirqs_off_thunk+0x1a/0x1c
[  227.282870] softirqs last  enabled at (54368): [<ffffffffb1c0031d>] __do_softirq+0x31d/0x483
[  227.282875] softirqs last disabled at (54349): [<ffffffffb1090099>] irq_exit+0xa9/0xc0
[  227.282936] WARNING: CPU: 0 PID: 1509 at drivers/gpu/drm/i915/intel_drv.h:2009 gen5_read32+0x16b/0x1a0 [i915]
[  227.282940] ---[ end trace 5e2fef4de3145efe ]---

Increasing the priority of the bug as it blocks ILK testing on BAT as IGT_4715:

https://intel-gfx-ci.01.org/tree/drm-tip/IGT_4715/fi-ilk-m540/igt@runner@aborted.html
	
Aborting.
Previous test: gem_exec_suspend (basic-s3)
Next test: gem_exec_suspend (basic-s4-devices)

Kernel tainted (0x240 -- 200)

I'm sure this patch is linked to elsewhere in bugs.fdo...

https://cgit.freedesktop.org/~ickle/linux-2.6/commit/?h=prescheduler&id=13ab88965bb597e1b228cc50fad76da09a5b8da0

A CI Bug Log filter associated to this bug has been updated:

{- ILK: random tests - dmesg-warn - RPM wakelock ref not held during HW access -}
{+ ILK: suspend tests + random ones - dmesg-warn - RPM wakelock ref not held during HW access +}

New failures caught by the filter:

* https://intel-gfx-ci.01.org/tree/drm-tip/IGT_4768/fi-ilk-m540/igt@kms_pipe_crc_basic@suspend-read-crc-pipe-a.html

commit 4a8ab5ea0cde753b03bfefe4c98a8c4c61f46550
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Mon Jan 14 14:21:29 2019 +0000

    drm/i915: Mark up Ironlake ips with rpm wakerefs
    
    Currently Ironlake operates under the assumption that rpm awake (and its
    error checking is disabled). As such, we have missed a few places where we
    access registers without taking the rpm wakeref and thus trigger
    warnings. intel_ips being one culprit.
    
    As this involved adding a potentially sleeping rpm_get, we have to
    rearrange the spinlocks slightly and so switch to acquiring a device-ref
    under the spinlock rather than hold the spinlock for the whole
    operation. To be consistent, we make the change in pattern common to the
    intel_ips interface even though this adds a few more atomic operations
    than necessary in a few cases.
    
    v2: Sagar noted the mb around setting mch_dev were overkill as we only
    need ordering there, and that i915_emon_status was still using
    struct_mutex for no reason, but lacked rpm.
    
    Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
    Cc: Jani Nikula <jani.nikula@intel.com>
    Reviewed-by: John Harrison <John.C.Harrison@Intel.com>
    Link: https://patchwork.freedesktop.org/patch/msgid/20190114142129.24398-21-chris@chris-wilson.co.uk

(In reply to Chris Wilson from comment #6)
> commit 4a8ab5ea0cde753b03bfefe4c98a8c4c61f46550
> Author: Chris Wilson <chris@chris-wilson.co.uk>
> Date:   Mon Jan 14 14:21:29 2019 +0000
> 
>     drm/i915: Mark up Ironlake ips with rpm wakerefs
>     
>     Currently Ironlake operates under the assumption that rpm awake (and its
>     error checking is disabled). As such, we have missed a few places where
> we
>     access registers without taking the rpm wakeref and thus trigger
>     warnings. intel_ips being one culprit.
>     
>     As this involved adding a potentially sleeping rpm_get, we have to
>     rearrange the spinlocks slightly and so switch to acquiring a device-ref
>     under the spinlock rather than hold the spinlock for the whole
>     operation. To be consistent, we make the change in pattern common to the
>     intel_ips interface even though this adds a few more atomic operations
>     than necessary in a few cases.
>     
>     v2: Sagar noted the mb around setting mch_dev were overkill as we only
>     need ordering there, and that i915_emon_status was still using
>     struct_mutex for no reason, but lacked rpm.
>     
>     Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
>     Cc: Jani Nikula <jani.nikula@intel.com>
>     Reviewed-by: John Harrison <John.C.Harrison@Intel.com>
>     Link:
> https://patchwork.freedesktop.org/patch/msgid/20190114142129.24398-21-
> chris@chris-wilson.co.uk

Thanks, it did the trick!