Bug 1594488 - Blank screen since Kernel 4.17 (AMDGPU.DC) on Amd Radeon RX460 and RX380 on Ultrawidescreen Monitor
Summary: Blank screen since Kernel 4.17 (AMDGPU.DC) on Amd Radeon RX460 and RX380 on U...
Keywords:
Status: CLOSED WORKSFORME
Alias: None
Product: Fedora
Classification: Fedora
Component: kernel
Version: 29
Hardware: x86_64
OS: Linux
unspecified
urgent
Target Milestone: ---
Assignee: Kernel Maintainer List
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2018-06-23 16:12 UTC by You
Modified: 2023-09-14 04:30 UTC (History)
26 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2019-09-17 20:04:52 UTC
Type: Bug
Embargoed:
naheemzaffar: needinfo-


Attachments (Terms of Use)
Systemd Journal Hardware logs from successfully booting Kernel-4.16 (3.59 KB, text/plain)
2018-06-24 12:41 UTC, You
no flags Details
Systemd Journal Hardware logs from Kernel-4.17 (3.55 KB, text/plain)
2018-06-24 12:42 UTC, You
no flags Details
dmesg.log (79.74 KB, text/x-vhdl)
2018-06-25 20:39 UTC, You
no flags Details
new dmesg.log (82.86 KB, text/x-vhdl)
2018-07-01 12:30 UTC, You
no flags Details
dmesg-4.16.log (80.99 KB, text/x-vhdl)
2018-07-15 01:03 UTC, You
no flags Details
Dmesg from boot with Kernel 4.19 (77.35 KB, text/plain)
2018-08-23 22:28 UTC, You
no flags Details
Taken from kernel 4.19 with amdgpu.dc=0 passed (82.27 KB, text/x-vhdl)
2018-08-23 22:29 UTC, You
no flags Details


Links
System ID Private Priority Status Summary Last Updated
Red Hat Bugzilla 1314923 0 unspecified CLOSED Ultra-Widescreen monitor Resolution option not provided for LG25UM57 (2560x1080) 2021-02-22 00:41:40 UTC

Description You 2018-06-23 16:12:43 UTC
Description of problem:

I get a blank screen at boot with kernel-4.17. I tried 4.17 pre releases from the kernel vanilla mainline copr and it had the same problem throughout the Release candidates.

I think it is related to the new display code path code merged into 4.17.

Before 4.17 I had a different problem as the full resolution of my ultrawide monitor would not be used properly (and produce garbled output if the full resolution chosen): https://bugzilla.redhat.com/show_bug.cgi?id=1314923

Not reported earlier as I had mistakenly thought the kernel was hanging at boot

Version-Release number of selected component (if applicable):

Fedora 28 with kernel from updates-testing
kernel version 4.17.2-200.fc28
LG Ultrawide monitor LG25UM57 (2560x1080)


How reproducible:
Always, be booting to new kernel.

Steps to Reproduce:
1. Boot to new kernel 4.17
2.
3.

Actual results:
Blank screen, something a single line of text before moving to blank screen (when an rtc_cmos error used to show).

Expected results:
boot to GUI

Additional info:

Comment 1 You 2018-06-24 12:41:54 UTC
Created attachment 1454132 [details]
Systemd Journal Hardware logs from successfully booting Kernel-4.16

Systemd Journal Hardware logs from successfully booting Kernel-4.16

Comment 2 You 2018-06-24 12:42:41 UTC
Created attachment 1454133 [details]
Systemd Journal Hardware logs from Kernel-4.17

Systemd Journal Hardware logs from Kernel-4.17

Comment 3 Jeremy Cline 2018-06-25 15:39:19 UTC
Hi You,

The system logs look wrong. Can you boot into 4.17, then reboot into 4.16 and run "journalctl -k --no-hostname -b -1 > dmesg.log" and attach that? That should get the full kernel logs from the previous boot. Thanks!

Comment 4 You 2018-06-25 20:39:42 UTC
Created attachment 1454491 [details]
dmesg.log

The result of "journalctl -k --no-hostname -b -1 > dmesg.log"

Comment 5 You 2018-07-01 12:27:20 UTC
I have today tried with the new update kernel with same result on this monitor.

However, I then tried connecting the TV via HDMI and that worked.

So this may be linked to big https://bugzilla.redhat.com/show_bug.cgi?id=1314923 where the video to the monitor was garbled at full resolution on previous kernels.

I will be attaching an updated dmesg.log.

Comment 6 You 2018-07-01 12:30:40 UTC
Created attachment 1455749 [details]
new dmesg.log

Current boot - Kernel 4.16.

Previous boot - 4.17 on LG Ultrawide G25UM57 (2560x1080), followed by connecting TV (1080p) as outpuut and switching back and forth a couple of times.

Comment 7 You 2018-07-15 01:03:13 UTC
Created attachment 1458907 [details]
dmesg-4.16.log

dmesg output from kernel 4.16

Comment 8 Klaus 2018-07-21 08:22:56 UTC
I have a similar problem using Radeon RX560D card.

Only HDMI works, all other outputs of graphic card can not used anymore. Problem persist with 4.17.6

Comment 9 mjnass 2018-07-28 19:34:02 UTC
I've had a similar issue with a Radeon R9 380x video card since the 4.17.x kernels.  One solution is to add "amdgpu.dc=0" to the kernel command line, emulating how the kernel treated the amdgpu prior to the 4.17.x changes (to the best of my understanding anyway).

I'm running Fedora 28.

My /etc/default/grub file:
GRUB_TIMEOUT=5
GRUB_DISTRIBUTOR="$(sed 's, release .*$,,g' /etc/system-release)"
GRUB_DEFAULT=saved
GRUB_DISABLE_SUBMENU=true
GRUB_TERMINAL_OUTPUT="console"
GRUB_CMDLINE_LINUX="rhgb quiet amdgpu.dc=0"
GRUB_DISABLE_RECOVERY="true"

Cat /proc/cmdline:
BOOT_IMAGE=/vmlinuz-4.17.7-200.fc28.x86_64 root=UUID=c65f5bb8-2cbc-4fbc-8853-62841f7ea13a ro rhgb quiet amdgpu.dc=0

My second monitor began working again right after adding the amdgpu.dc=0, and then telling grub to rebuild 'sudo update-grub'.

Hope this helps.

Comment 10 Klaus 2018-07-29 17:19:03 UTC
amdgpu.dc=0 works for me. Thanks for the hint!

Comment 11 You 2018-07-29 20:29:39 UTC
Updated title as blank screen is due to the newly enabled displaycode for previous generations.

Adding amdgpu.dc=0 also shows screen for me (back to previous situation of working up to 1080p on an ultrawidescreen display and being garbled at full resolution - different bug report filed previously on that.)

Comment 12 You 2018-08-02 10:15:28 UTC
For those reporting similar problems, are these also on ultrawidescreen monitors, or are they normal monitors?

For me:

RX460 - HDMI Connection - Ultrawide Monitor LG25UM57 (2560x1080) - Blank screen with AMDGPU.DC, HD Resolution works with AMDGPU.DC=0 (reported separately).

Comment 13 You 2018-08-23 22:28:19 UTC
Created attachment 1478360 [details]
Dmesg from boot with Kernel 4.19

Taken when booting with Vanilla Mainline kernel 4.19.0-0.rc0.git8.1.vanilla.knurd.1.fc28.x86_64

Comment 14 You 2018-08-23 22:29:21 UTC
Created attachment 1478361 [details]
Taken from kernel 4.19 with amdgpu.dc=0 passed

Taken when booting with Vanilla Mainline kernel 4.19.0-0.rc0.git8.1.vanilla.knurd.1.fc28.x86_64 with argument amdgpu.dc=0

Comment 15 You 2018-08-23 22:36:45 UTC
The problem persists with the latest kernel from the vanilla mainline kernel repository that builds a pre-release kernel from 4.19.

Blank screen without amdgpu.dc=0
Cannot get full screen on ultrawidescreen monitory without it.

I have noticed a couple of entries in the dmesg without the amdgpu.dc=0:

kernel: [drm] add ip block number 4 <dm>

with amdgpu.dc=0 it is:

kernel: [drm] add ip block number 4 <dce_v11_0>

The other ip address block entries are identical.

with amdgpu.dc=0 I get this section:

Aug 22 23:46:22 kernel: [drm] PCIE GART of 256M enabled (table at 0x000000F400900000).
Aug 22 23:46:22 kernel: [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
Aug 22 23:46:22 kernel: [drm] Driver supports precise vblank timestamp query.
Aug 22 23:46:22 kernel: [drm] AMDGPU Display Connectors
Aug 22 23:46:22 kernel: [drm] Connector 0:
Aug 22 23:46:22 kernel: [drm]   DP-1
Aug 22 23:46:22 kernel: [drm]   HPD5
Aug 22 23:46:22 kernel: [drm]   DDC: 0x4868 0x4868 0x4869 0x4869 0x486a 0x486a 0x486b 0x486b
Aug 22 23:46:22 kernel: [drm]   Encoders:
Aug 22 23:46:22 kernel: [drm]     DFP1: INTERNAL_UNIPHY1
Aug 22 23:46:22 kernel: [drm] Connector 1:
Aug 22 23:46:22 kernel: [drm]   HDMI-A-1
Aug 22 23:46:22 kernel: [drm]   HPD3
Aug 22 23:46:22 kernel: [drm]   DDC: 0x4874 0x4874 0x4875 0x4875 0x4876 0x4876 0x4877 0x4877
Aug 22 23:46:22 kernel: [drm]   Encoders:
Aug 22 23:46:22 kernel: [drm]     DFP2: INTERNAL_UNIPHY1
Aug 22 23:46:22 kernel: [drm] Connector 2:
Aug 22 23:46:22 kernel: [drm]   DVI-D-1
Aug 22 23:46:22 kernel: [drm]   HPD4
Aug 22 23:46:22 kernel: [drm]   DDC: 0x4878 0x4878 0x4879 0x4879 0x487a 0x487a 0x487b 0x487b
Aug 22 23:46:22 kernel: [drm]   Encoders:
Aug 22 23:46:22 kernel: [drm]     DFP3: INTERNAL_UNIPHY
Aug 22 23:46:22 kernel: [drm] Chained IB support enabled!

but without amdgpu.dc=0 only the first and last line of that section appear.

Later down without the amdgpu.dc=0 I also get the following message:

kernel: [drm] Cannot find any crtc or sizes

Comment 16 Klaus 2018-08-25 08:07:56 UTC
@12:

> For those reporting similar problems, are these also on ultrawidescreen monitors, or are they normal monitors?

Also for for 1920x1080 screen. Only 1 of 3 card outputs are active (HDMI is working). HDMI works as expected, the 2 others are non active, monitor goes to "no input/power down".

Comment 17 You 2018-11-04 22:02:50 UTC
Error still exists with Fedora 29. and with kernel 4.19 and post merge of amdgpu pull on 4.20

I have reporte the bug upstream: https://bugs.freedesktop.org/show_bug.cgi?id=107668

Same problem exists on Ubuntu: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1761751

Comment 18 Jeremy Cline 2018-12-03 17:31:47 UTC
We apologize for the inconvenience.  There is a large number of bugs to go through and several of them have gone stale.  Due to this, we are doing a mass bug update across all of the Fedora 29 kernel bugs.
 
Fedora 29 has now been rebased to 4.19.5-300.fc29.  Please test this kernel update (or newer) and let us know if you issue has been resolved or if it is still present with the newer kernel.
 
If you experience different issues, please open a new bug report for those.

Comment 19 You 2018-12-03 19:32:03 UTC
Problem still persists with kernel 4.19.5-300.fc29.

Comment 20 Joe Wright 2018-12-26 22:55:30 UTC
Can confirm, this is still an issue

Comment 21 Olaf Leidinger 2019-01-08 09:50:06 UTC
I have the very same problem. Not with a ultra wide screen, but a 4K screen from LG. Device is a RX 480.

--- dc=0.txt	2019-01-08 10:45:41.131642231 +0100
+++ dc=1.txt	2019-01-08 10:47:07.048631436 +0100
@@ -10,7 +10,7 @@
 [drm] add ip block number 3 <gfx_v8_0>
 [drm] add ip block number 4 <sdma_v3_0>
 [drm] add ip block number 5 <powerplay>
-[drm] add ip block number 6 <dce_v11_0>
+[drm] add ip block number 6 <dm>
 [drm] add ip block number 7 <uvd_v6_0>
 [drm] add ip block number 8 <vce_v3_0>
 [drm] UVD is enabled in VM mode
@@ -23,43 +23,33 @@
 [drm] amdgpu: 4096M of GTT memory ready.
 [drm] GART: num cpu pages 65536, num gpu pages 65536
 [drm] PCIE GART of 256M enabled (table at 0x000000F400300000).
-[drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
-[drm] Driver supports precise vblank timestamp query.
 [drm] Chained IB support enabled!
-[drm] AMDGPU Display Connectors
-[drm] Connector 0:
-[drm]   DP-1
-[drm]   HPD6
-[drm]   DDC: 0x4868 0x4868 0x4869 0x4869 0x486a 0x486a 0x486b 0x486b
-[drm]   Encoders:
-[drm]     DFP1: INTERNAL_UNIPHY2
-[drm] Connector 1:
-[drm]   DP-2
-[drm]   HPD4
-[drm]   DDC: 0x4870 0x4870 0x4871 0x4871 0x4872 0x4872 0x4873 0x4873
-[drm]   Encoders:
-[drm]     DFP2: INTERNAL_UNIPHY2
-[drm] Connector 2:
-[drm]   HDMI-A-1
-[drm]   HPD1
-[drm]   DDC: 0x486c 0x486c 0x486d 0x486d 0x486e 0x486e 0x486f 0x486f
-[drm]   Encoders:
-[drm]     DFP3: INTERNAL_UNIPHY1
-[drm] Connector 3:
-[drm]   HDMI-A-2
-[drm]   HPD5
-[drm]   DDC: 0x4874 0x4874 0x4875 0x4875 0x4876 0x4876 0x4877 0x4877
-[drm]   Encoders:
-[drm]     DFP4: INTERNAL_UNIPHY1
-[drm] Connector 4:
-[drm]   DVI-D-1
-[drm]   HPD3
-[drm]   DDC: 0x487c 0x487c 0x487d 0x487d 0x487e 0x487e 0x487f 0x487f
-[drm]   Encoders:
-[drm]     DFP5: INTERNAL_UNIPHY
 [drm] Found UVD firmware Version: 1.79 Family ID: 16
 [drm] UVD ENC is disabled
 [drm] Found VCE firmware Version: 52.4 Binary ID: 3
+[drm] DM_PPLIB: values for Engine clock
+[drm] DM_PPLIB:	 300000
+[drm] DM_PPLIB:	 466000
+[drm] DM_PPLIB:	 751000
+[drm] DM_PPLIB:	 1019000
+[drm] DM_PPLIB:	 1074000
+[drm] DM_PPLIB:	 1126000
+[drm] DM_PPLIB:	 1169000
+[drm] DM_PPLIB:	 1260000
+[drm] DM_PPLIB: Validation clocks:
+[drm] DM_PPLIB:    engine_max_clock: 126000
+[drm] DM_PPLIB:    memory_max_clock: 175000
+[drm] DM_PPLIB:    level           : 8
+[drm] DM_PPLIB: values for Memory clock
+[drm] DM_PPLIB:	 300000
+[drm] DM_PPLIB:	 1750000
+[drm] DM_PPLIB: Validation clocks:
+[drm] DM_PPLIB:    engine_max_clock: 126000
+[drm] DM_PPLIB:    memory_max_clock: 175000
+[drm] DM_PPLIB:    level           : 8
+[drm] Display Core initialized with v3.1.68!
+[drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
+[drm] Driver supports precise vblank timestamp query.
 [drm] UVD initialized successfully.
 [drm] VCE initialized successfully.
 [drm] fb mappable at 0xE06B0000
@@ -70,5 +60,4 @@
 fbcon: amdgpudrmfb (fb0) is primary device
 amdgpu 0000:08:00.0: fb0: amdgpudrmfb frame buffer device
 [drm] Initialized amdgpu 3.27.0 20150101 for 0000:08:00.0 on minor 0

Comment 22 You 2019-02-23 16:25:38 UTC
I have just tried using an older 3m long HDMI cable I had at home that I presume cannot support the latest HDMI version and with that I got a display at full resolution.

The problem may be with HDMI handling, there seem to be others experiencing some problems with HDMI 2, which I dont think this card supports.

is there a way to confirm the HDMI versions I was using that failed and the version that succeeded?

Comment 23 Laura Abbott 2019-04-09 20:44:47 UTC
We apologize for the inconvenience.  There is a large number of bugs to go through and several of them have gone stale.  Due to this, we are doing a mass bug update across all of the Fedora 29 kernel bugs.
 
Fedora XX has now been rebased to 5.0.6  Please test this kernel update (or newer) and let us know if you issue has been resolved or if it is still present with the newer kernel.
 
If you have moved on to Fedora 30, and are still experiencing this issue, please change the version to Fedora 30.
 
If you experience different issues, please open a new bug report for those.

Comment 24 mjnass 2019-04-17 22:34:31 UTC
Just verified that this (second monitor for me) is still an issue in Fedora 29 on kernel 5.0.7-200.

The fix of adding "amdgpu.dc=0" to kernel boot parameters still works.

Comment 25 Justin M. Forbes 2019-09-17 20:04:52 UTC
*********** MASS BUG UPDATE **************
This bug is being closed with INSUFFICIENT_DATA as there has not been a response in 3 weeks. If you are still experiencing this issue, please reopen and attach the relevant data from the latest kernel you are running and any data that might have been requested previously.

Comment 26 You 2020-03-29 02:18:23 UTC
I no longer git this with my current HDMI cable.

Comment 27 Red Hat Bugzilla 2023-09-14 04:30:16 UTC
The needinfo request[s] on this closed bug have been removed as they have been unresolved for 1000 days


Note You need to log in before you can comment on or make changes to this bug.