Warning: Broadwell/i915: Unclaimed register detected before reading register 0x130040
Hey,
this just popped up in the dmesg of my Dell XPS 13 -- the system seems to run well, but still, it asks about being cut and sent, so here it is. It is on 4.1.0-rc4+ (Linus' tree as of May 24th, around 3pm UTC -- don't have the git commit ID anymore).
------------[ cut here ]------------ WARNING: CPU: 0 PID: 1167 at /home/brodo/local/kernel/git/linux/drivers/gpu/drm/i915/intel_uncore.c:566 hsw_unclaimed_reg_debug+0x7d/0xa0() Unclaimed register detected before reading register 0x130040 Modules linked in: CPU: 0 PID: 1167 Comm: Xorg Not tainted 4.1.0-rc4+ #1 Hardware name: Dell Inc. XPS 13 9343/0TM99H, BIOS A03 03/25/2015 ffffffffb39694d8 ffff88020e14b998 ffffffffb322906f 0000000080000001 ffff88020e14b9e8 ffff88020e14b9d8 ffffffffb272ca8a ffff88020e14b9a8 ffff880215190000 0000000000130040 ffff880215bfd800 0000000000130040 Call Trace: [<ffffffffb322906f>] dump_stack+0x4f/0x7b [<ffffffffb272ca8a>] warn_slowpath_common+0x8a/0xc0 [<ffffffffb272cb06>] warn_slowpath_fmt+0x46/0x50 [<ffffffffb2be63cd>] hsw_unclaimed_reg_debug+0x7d/0xa0 [<ffffffffb2be84b6>] gen6_read32+0x56/0x1c0 [<ffffffffb2c096e6>] hsw_disable_pc8+0x36/0x2c0 [<ffffffffb2b98e60>] intel_runtime_resume+0x160/0x1b0 [<ffffffffb2a94bdf>] pci_pm_runtime_resume+0x7f/0xc0 [<ffffffffb2a94b60>] ? pci_restore_standard_config+0x50/0x50 [<ffffffffb2c5c8f6>] __rpm_callback+0x36/0x90 [<ffffffffb2c5c976>] rpm_callback+0x26/0xa0 [<ffffffffb2c5dcc6>] rpm_resume+0x496/0x670 [<ffffffffb2756209>] ? preempt_count_add+0x79/0xc0 [<ffffffffb2c5dee0>] __pm_runtime_resume+0x40/0x60 [<ffffffffb2ba7fd5>] intel_runtime_pm_get+0x55/0x80 [<ffffffffb2bc6989>] i915_gem_free_object+0x29/0x360 [<ffffffffb2b6b417>] drm_gem_object_free+0x27/0x40 [<ffffffffb2b6b548>] drm_gem_object_handle_unreference_unlocked+0x118/0x130 [<ffffffffb2b6b6d4>] drm_gem_handle_delete+0xb4/0x100 [<ffffffffb2fde8ab>] ? sock_recvmsg+0x3b/0x50 [<ffffffffb2b6ba95>] drm_gem_close_ioctl+0x25/0x30 [<ffffffffb2b6c4a5>] drm_ioctl+0x1a5/0x6a0 [<ffffffffb2869672>] ? do_readv_writev+0x1b2/0x260 [<ffffffffb323187a>] ? _raw_spin_unlock_irqrestore+0x2a/0x60 [<ffffffffb2a5b804>] ? timerqueue_del+0x24/0x70 [<ffffffffb278b2fe>] ? __remove_hrtimer+0x4e/0xe0 [<ffffffffb287c110>] do_vfs_ioctl+0x2e0/0x4e0 [<ffffffffb2886482>] ? __fget+0x72/0xb0 [<ffffffffb287c391>] SyS_ioctl+0x81/0xa0 [<ffffffffb3232017>] system_call_fastpath+0x12/0x6a ---[ end trace 8ff5864d3e587195 ]---
Best,
Dominik
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 6 lines · Dominik Brodowski
Hey,
this just popped up in the dmesg of my Dell XPS 13 -- the system seems to run well, but still, it asks about being cut and sent, so here it is. It is on 4.1.0-rc4+ (Linus' tree as of May 24th, around 3pm UTC -- don't have the git commit ID anymore).
Can you please boot with i915.mmio_debug=10 and try to reproduce the issue? That should light more onto this issue. And yup it's indeed a bug since this means a register write has been dropped by the hardware, which often tends to be somewhat harmless except when it doesn't ;-) -Daniel
Quoted 52 lines
------------[ cut here ]------------ WARNING: CPU: 0 PID: 1167 at /home/brodo/local/kernel/git/linux/drivers/gpu/drm/i915/intel_uncore.c:566 hsw_unclaimed_reg_debug+0x7d/0xa0() Unclaimed register detected before reading register 0x130040 Modules linked in: CPU: 0 PID: 1167 Comm: Xorg Not tainted 4.1.0-rc4+ #1 Hardware name: Dell Inc. XPS 13 9343/0TM99H, BIOS A03 03/25/2015 ffffffffb39694d8 ffff88020e14b998 ffffffffb322906f 0000000080000001 ffff88020e14b9e8 ffff88020e14b9d8 ffffffffb272ca8a ffff88020e14b9a8 ffff880215190000 0000000000130040 ffff880215bfd800 0000000000130040 Call Trace: [<ffffffffb322906f>] dump_stack+0x4f/0x7b [<ffffffffb272ca8a>] warn_slowpath_common+0x8a/0xc0 [<ffffffffb272cb06>] warn_slowpath_fmt+0x46/0x50 [<ffffffffb2be63cd>] hsw_unclaimed_reg_debug+0x7d/0xa0 [<ffffffffb2be84b6>] gen6_read32+0x56/0x1c0 [<ffffffffb2c096e6>] hsw_disable_pc8+0x36/0x2c0 [<ffffffffb2b98e60>] intel_runtime_resume+0x160/0x1b0 [<ffffffffb2a94bdf>] pci_pm_runtime_resume+0x7f/0xc0 [<ffffffffb2a94b60>] ? pci_restore_standard_config+0x50/0x50 [<ffffffffb2c5c8f6>] __rpm_callback+0x36/0x90 [<ffffffffb2c5c976>] rpm_callback+0x26/0xa0 [<ffffffffb2c5dcc6>] rpm_resume+0x496/0x670 [<ffffffffb2756209>] ? preempt_count_add+0x79/0xc0 [<ffffffffb2c5dee0>] __pm_runtime_resume+0x40/0x60 [<ffffffffb2ba7fd5>] intel_runtime_pm_get+0x55/0x80 [<ffffffffb2bc6989>] i915_gem_free_object+0x29/0x360 [<ffffffffb2b6b417>] drm_gem_object_free+0x27/0x40 [<ffffffffb2b6b548>] drm_gem_object_handle_unreference_unlocked+0x118/0x130 [<ffffffffb2b6b6d4>] drm_gem_handle_delete+0xb4/0x100 [<ffffffffb2fde8ab>] ? sock_recvmsg+0x3b/0x50 [<ffffffffb2b6ba95>] drm_gem_close_ioctl+0x25/0x30 [<ffffffffb2b6c4a5>] drm_ioctl+0x1a5/0x6a0 [<ffffffffb2869672>] ? do_readv_writev+0x1b2/0x260 [<ffffffffb323187a>] ? _raw_spin_unlock_irqrestore+0x2a/0x60 [<ffffffffb2a5b804>] ? timerqueue_del+0x24/0x70 [<ffffffffb278b2fe>] ? __remove_hrtimer+0x4e/0xe0 [<ffffffffb287c110>] do_vfs_ioctl+0x2e0/0x4e0 [<ffffffffb2886482>] ? __fget+0x72/0xb0 [<ffffffffb287c391>] SyS_ioctl+0x81/0xa0 [<ffffffffb3232017>] system_call_fastpath+0x12/0x6a ---[ end trace 8ff5864d3e587195 ]---Best,
Dominik_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 7 lines
Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch _______________________________________________ Intel-gfx mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 10 lines · Daniel Vetter
Quoted 6 lines · Dominik Brodowski
Hey,
this just popped up in the dmesg of my Dell XPS 13 -- the system seems to run well, but still, it asks about being cut and sent, so here it is. It is on 4.1.0-rc4+ (Linus' tree as of May 24th, around 3pm UTC -- don't have the git commit ID anymore).
Can you please boot with i915.mmio_debug=10 and try to reproduce the issue?
It won't help. To see this WARN implies that hsw_unclaimed_reg_detect() fired and mmio debugging is enabled, and the WARN then fires before a subsequent register access. So it implies the unclaimed register is not being accessed by the display driver -- presuming that we have wrapped all register access appropriately.
<invalid register access>hsw_unclaimed_reg_detect -> fires, clears the invalid flag
...
(
hsw_unclaimed_reg_debug before
mmio
hsw_unclaimed_reg_debug after
) -> nothing x N
...
<invalid register access>hsw_unclaimed_reg_debug -> fires before, ergo no more useful information
Or else I nerfed the automatic debugging too much. -Chris
Signature · 5 lines
Chris Wilson, Intel Open Source Technology Centre _______________________________________________ Intel-gfx mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 30 lines · Chris Wilson
Quoted 10 lines · Daniel Vetter
Quoted 6 lines · Dominik Brodowski
Hey,
this just popped up in the dmesg of my Dell XPS 13 -- the system seems to run well, but still, it asks about being cut and sent, so here it is. It is on 4.1.0-rc4+ (Linus' tree as of May 24th, around 3pm UTC -- don't have the git commit ID anymore).
Can you please boot with i915.mmio_debug=10 and try to reproduce the issue?
It won't help. To see this WARN implies that
hsw_unclaimed_reg_detect()fired and mmio debugging is enabled, and the WARN then fires before a subsequent register access. So it implies the unclaimed register is not being accessed by the display driver -- presuming that we have wrapped all register access appropriately.<invalid register access>
hsw_unclaimed_reg_detect-> fires, clears the invalid flag
...
(
hsw_unclaimed_reg_debugbefore
mmio
hsw_unclaimed_reg_debugafter
) -> nothing x N
...
<invalid register access>hsw_unclaimed_reg_debug-> fires before, ergo no more useful information
Quoted 1 line
Or else I nerfed the automatic debugging too much.
Hm I thought it's also gating some of the interrupt frobbery, but that doesn't exist on bdw anyway. It would be still good to figure out who's stomping all over the gpu though.
-Daniel
Signature · 7 lines
Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch _______________________________________________ Intel-gfx mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/intel-gfx