Bug 111894 - [i915][CFL][3E91] kernel-view: HDMI-A-1 detected -> connected -> pipe_off wait timed out; external-display-view: HDMI connected -> no signal
Summary: [i915][CFL][3E91] kernel-view: HDMI-A-1 detected -> connected -> pipe_off wai...
Status: NEW
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: unspecified
Hardware: x86-64 (AMD64) Linux (All)
: high critical
Assignee: Intel GFX Bugs mailing list
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard: Triaged, ReadyForDev
Keywords:
Depends on:
Blocks:
 
Reported: 2019-10-03 12:03 UTC by amike2019
Modified: 2019-10-06 07:56 UTC (History)
3 users (show)

See Also:
i915 platform: CFL
i915 features: display/HDMI


Attachments
complete kernel(5.4.0-rc1-drm-tip-git-ga6af6b11a) log with few tries to connect/disconnect hdmi display (362.02 KB, text/plain)
2019-10-03 23:00 UTC, amike2019
no flags Details
/sys/kernel/debug/dri/0/i915_display_info (3.13 KB, text/plain)
2019-10-04 12:32 UTC, amike2019
no flags Details
complete kernel(5.4.0-rc1-drm-tip-git-ga6af6b11a) log with args(drm.debug=0xe log_buf_len=1M 3) and few tries to connect/disconnect hdmi display (272.91 KB, text/plain)
2019-10-04 12:34 UTC, amike2019
no flags Details
kernel(5.4.0-rc1-drm-tip-git-ga6af6b11a) log with args(drm.debug=0xe log_buf_len=1M 3) for i915_live_test.sh (284.65 KB, text/plain)
2019-10-04 16:23 UTC, amike2019
no flags Details
i915_live_test.sh (387 bytes, text/plain)
2019-10-04 16:24 UTC, amike2019
no flags Details
i915_live_test.txt (89.55 KB, text/plain)
2019-10-04 16:25 UTC, amike2019
no flags Details
part of kernel(5.3.1-arch1-1-ARCH) log and xrandr output from another device(hp laptop with radeon driver) with that display (14.22 KB, text/plain)
2019-10-06 07:50 UTC, amike2019
no flags Details

Note You need to log in before you can comment on or make changes to this bug.
Description amike2019 2019-10-03 12:03:21 UTC
HDMI connection works for at least virtual terminal with module_blacklist=i915

HDMI connection doesn't work at all without module_blacklist=i915.

kernel: ------------[ cut here ]------------
kernel: pipe_off wait timed out
kernel: WARNING: CPU: 2 PID: 339059 at drivers/gpu/drm/i915/display/intel_display.c:1082 intel_disable_pipe+0x1c4/0x1d0 [i915]
kernel: Modules linked in: msr 8021q garp mrp stp llc snd_hda_codec_hdmi snd_hda_codec_realtek snd_hda_codec_generic ledtrig_audio ipt_REJECT nf_reject_ipv4 xt_conntrack iptable_filter xt_REDIRECT xt_tcpudp xt_owner iptable_nat nf_nat nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 libcrc32c sof_pci_dev snd_sof_intel_hda_common snd_sof_intel_hda snd_sof_intel_byt snd_sof_intel_ipc snd_sof snd_sof_xtensa_dsp snd_soc_skl intel_rapl_msr intel_rapl_common snd_soc_hdac_hda snd_hda_ext_core snd_soc_skl_ipc snd_soc_sst_ipc snd_soc_sst_dsp snd_soc_acpi_intel_match i915 x86_pkg_temp_thermal snd_soc_acpi intel_powerclamp snd_soc_core coretemp iTCO_wdt snd_compress ac97_bus iTCO_vendor_support mei_hdcp snd_pcm_dmaengine kvm_intel wmi_bmof snd_hda_intel intel_wmi_thunderbolt iwlmvm snd_hda_codec kvm mac80211 i2c_algo_bit irqbypass intel_cstate libarc4 drm_kms_helper snd_hda_core intel_uncore btusb btrtl btbcm intel_rapl_perf iwlwifi pcspkr snd_hwdep btintel snd_pcm drm snd_timer snd intel_gtt cfg80211
kernel:  e1000e bluetooth i2c_i801 soundcore agpgart mei_me mousedev joydev ecdh_generic syscopyarea input_leds rfkill sysfillrect ecc mei sysimgblt fb_sys_fops ie31200_edac tpm_crb wmi tpm_tis tpm_tis_core tpm rng_core evdev mac_hid ip_tables x_tables ext4 crc32c_generic crc16 mbcache jbd2 serpent_avx2 serpent_avx_x86_64 serpent_sse2_x86_64 serpent_generic dm_crypt dm_mod sd_mod hid_generic usbhid hid uas usb_storage crct10dif_pclmul crc32_pclmul crc32c_intel ghash_clmulni_intel aesni_intel ahci libahci libata aes_x86_64 crypto_simd xhci_pci cryptd glue_helper scsi_mod xhci_hcd
kernel: CPU: 2 PID: 339059 Comm: kworker/2:0 Not tainted 5.3.1-arch1-1-ARCH #1
kernel: Workqueue: events i915_hotplug_work_func [i915]
kernel: RIP: 0010:intel_disable_pipe+0x1c4/0x1d0 [i915]
kernel: Code: ff ff e8 ef a8 d5 cc e9 aa fe ff ff 5b 31 f6 5d 4c 89 c7 41 5c 41 5d 41 5e e9 68 d1 ff ff 48 c7 c7 14 14 f7 c0 e8 6b 45 de cc <0f> 0b e9 5e ff ff ff 0f 1f 44 00 00 0f 1f 44 00 00 41 57 4c 8d 7e
kernel: RSP: 0018:ffffb240c070fb90 EFLAGS: 00010286
kernel: RAX: 0000000000000000 RBX: 00000000fffffffc RCX: 0000000000000000
kernel: RDX: 0000000000000001 RSI: 0000000000000096 RDI: 00000000ffffffff
kernel: RBP: ffff968f0ff20000 R08: 0000000000000759 R09: 0000000000000004
kernel: R10: 0000000000000000 R11: 0000000000000001 R12: ffff968f0ff20760
kernel: R13: ffff968f15348000 R14: 0000000000070008 R15: ffff968f13fc6828
kernel: FS:  0000000000000000(0000) GS:ffff968f1a300000(0000) knlGS:0000000000000000
kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
kernel: CR2: 00007f870d15b3c0 CR3: 0000000006e0a006 CR4: 00000000003606e0
kernel: Call Trace:
kernel:  haswell_crtc_disable+0xd1/0x120 [i915]
kernel:  intel_atomic_commit_tail+0x24b/0x1360 [i915]
kernel:  ? flush_workqueue+0x1b8/0x440
kernel:  intel_atomic_commit+0x298/0x2d0 [i915]
kernel:  drm_client_modeset_commit_atomic+0x1e8/0x200 [drm]
kernel:  drm_client_modeset_commit_force+0x50/0x150 [drm]
kernel:  drm_fb_helper_restore_fbdev_mode_unlocked+0x49/0xa0 [drm_kms_helper]
kernel:  drm_fb_helper_set_par+0x2c/0x50 [drm_kms_helper]
kernel:  drm_fb_helper_hotplug_event.part.0+0x97/0xb0 [drm_kms_helper]
kernel:  drm_kms_helper_hotplug_event+0x26/0x30 [drm_kms_helper]
kernel:  i915_hotplug_work_func+0x1dd/0x1f0 [i915]
kernel:  process_one_work+0x1d1/0x3a0
kernel:  worker_thread+0x4a/0x3d0
kernel:  kthread+0xfb/0x130
kernel:  ? process_one_work+0x3a0/0x3a0
kernel:  ? kthread_park+0x80/0x80
kernel:  ret_from_fork+0x35/0x40
kernel: ---[ end trace dc2e8cfe65397b44 ]---
Comment 1 amike2019 2019-10-03 17:39:43 UTC
Please, feel free to request any details.
Comment 2 amike2019 2019-10-03 23:00:11 UTC
Created attachment 145633 [details]
complete kernel(5.4.0-rc1-drm-tip-git-ga6af6b11a) log with few tries to connect/disconnect hdmi display
Comment 3 Jani Saarinen 2019-10-04 10:41:35 UTC
Can you add also output from  /sys/kernel/debug/dri/0/i915_display_info when HDMI monitor is connected.
Comment 4 Jani Saarinen 2019-10-04 11:38:39 UTC
Also can you boot with drm.debug=0xe log_buf_len=1M 3 parameters in grub to see if you boot to run level 3 if that is i915 issue.
Comment 5 amike2019 2019-10-04 12:32:06 UTC
Created attachment 145641 [details]
/sys/kernel/debug/dri/0/i915_display_info
Comment 6 amike2019 2019-10-04 12:34:38 UTC
Created attachment 145642 [details]
complete kernel(5.4.0-rc1-drm-tip-git-ga6af6b11a) log with args(drm.debug=0xe log_buf_len=1M 3) and few tries to connect/disconnect hdmi display
Comment 7 Jani Saarinen 2019-10-04 14:30:44 UTC
I see that according display_info to HDMI-A-2, status: connected
Comment 8 shashank.sharma@intel.com 2019-10-04 15:14:44 UTC
From the logs, I can see that:
HDMI-A-2 gets detected during the boot-up
HDMI-A-2 gets modeset of 1360x768
Then it gets a hot-unplug ( I guess this is the manual unplug)
While disabling the PIPE, the PIPECONF register times out.

Can you please provide this register's value (0x70008), when you see this error?
./intel_reg read 0x70008

As this error has shown up on HDMI hot-unplg, Does the display ever come-up, in the consecutive hot-plus, after we see this timeout error?
Comment 9 Lakshmi 2019-10-04 15:21:08 UTC
(In reply to amike2019 from comment #6)
> Created attachment 145642 [details]
> complete kernel(5.4.0-rc1-drm-tip-git-ga6af6b11a) log with
> args(drm.debug=0xe log_buf_len=1M 3) and few tries to connect/disconnect
> hdmi display

Have you done any changes recently so this issue appeared?
Comment 10 amike2019 2019-10-04 15:36:27 UTC
(In reply to Lakshmi from comment #9) 
> Have you done any changes recently so this issue appeared?
No. I've added kernels logs for drm-tip branch to prove that this kernel version has the same problem as 5.3.1 and 4.19.76.

(In reply to shashank.sharma@intel.com from comment #8)
> As this error has shown up on HDMI hot-unplg, Does the display ever come-up,
> in the consecutive hot-plus, after we see this timeout error?
No, it always reports "no signal" as soon as i can start ssh session.


(In reply to shashank.sharma@intel.com from comment #8)
> From the logs, I can see that:
> HDMI-A-2 gets detected during the boot-up
> HDMI-A-2 gets modeset of 1360x768
> Then it gets a hot-unplug ( I guess this is the manual unplug)
> While disabling the PIPE, the PIPECONF register times out.

I've booted mini-pc with connected display and only after established ssh connection did the following two times:
 turned off power of display -> turned on power of display
Comment 11 amike2019 2019-10-04 16:23:52 UTC
Created attachment 145646 [details]
kernel(5.4.0-rc1-drm-tip-git-ga6af6b11a) log with args(drm.debug=0xe log_buf_len=1M 3) for i915_live_test.sh
Comment 12 amike2019 2019-10-04 16:24:32 UTC
Created attachment 145647 [details]
i915_live_test.sh
Comment 13 amike2019 2019-10-04 16:25:10 UTC
Created attachment 145648 [details]
i915_live_test.txt
Comment 14 amike2019 2019-10-04 16:28:47 UTC
(In reply to shashank.sharma@intel.com from comment #8)
> Can you please provide this register's value (0x70008), when you see this
> error?
> ./intel_reg read 0x70008

See the following attached files: 
i915_live_test.txt
i915_live_test.sh
kernel(5.4.0-rc1-drm-tip-git-ga6af6b11a) log with args(drm.debug=0xe log_buf_len=1M 3) for i915_live_test.sh
Comment 15 shashank.sharma@intel.com 2019-10-06 03:41:56 UTC
Can you reproduce this issue, with another monitor ? I am slightly skeptical about all the 1080i modes getting pruned, and falling to the resolution of 1360x768 (instead of 1366x768).

- Shashank
Comment 16 amike2019 2019-10-06 07:50:12 UTC
Created attachment 145663 [details]
part of kernel(5.3.1-arch1-1-ARCH) log and xrandr output from another device(hp laptop with radeon driver) with that display
Comment 17 amike2019 2019-10-06 07:53:12 UTC
(In reply to shashank.sharma@intel.com from comment #15)
> Can you reproduce this issue, with another monitor ? I am slightly skeptical
> about all the 1080i modes getting pruned, and falling to the resolution of
> 1360x768 (instead of 1366x768).
> 
> - Shashank

No, it's the only available display here.

But this display works with another device, see attachment 145663 [details].
Comment 18 amike2019 2019-10-06 07:56:05 UTC
(In reply to shashank.sharma@intel.com from comment #15)
> Can you reproduce this issue, with another monitor ? I am slightly skeptical
> about all the 1080i modes getting pruned, and falling to the resolution of
> 1360x768 (instead of 1366x768).
> 
> - Shashank

It should not be a problem that this display has 1360x768 mode in edid instead of 1366x768.


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.