Black screen on 18.04 + AMD RX460

This bug is missing log files that will aid in diagnosing the problem. While running an Ubuntu kernel (not a mainline or third-party kernel) please enter the following command in a terminal window:

apport-collect 1761751

and then change the status of the bug to 'Confirmed'.

If, due to the nature of the issue you have encountered, you are unable to run this command, please add a comment stating that fact and change the bug status to 'Confirmed'.

This change has been made by an automated script, maintained by the Ubuntu Kernel Team.

Ok, so I found this message when trying to boot a faulty kernel without workaround, if it helps:

[ 1.797127] [drm] Cannot find any crtc or sizes
[ 1.799000] [drm] Initialized amdgpu 3.23.0 20150101 for 0000:23:00.0 on minor 0

I'd like to perform a bisect to figure out what commit caused this regression. We need to identify the earliest kernel where the issue started happening as well as the latest kernel that did not have this issue.

Can you test the following kernels and report back? We are looking for the first kernel version that exhibits this bug:

v4.14 Final: http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.14/
v4.15-rc1: http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.15-rc1/
v4.15-rc5: http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.15-rc5/
v4.15-rc9: http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.15-rc9/
v4.15 Final: http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.15/
v4.16-rc1: http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.16-rc1/

You don't have to test every kernel, just up until the kernel that first has this bug.

Thanks in advance!

v4.14 Final: http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.14/ - Boots normally
v4.15-rc1: http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.15-rc1/ - Boots normally
v4.15-rc5: http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.15-rc5/ - Black screen
v4.15-rc9: http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.15-rc9/ - Black screen
v4.15 Final: http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.15/ - Black screen
v4.16-rc1: http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.16-rc1/ - Not tested

I'll try rc2, rc3 and rc4 next.

4.15.0-041500rc2-generic - Boots normally
4.15.0-041500rc3-generic - Black screen
4.15.0-041500rc4-generic - Not tested

So something between rc2 and rc3 it is then...

I started a kernel bisect between v4.15-rc2 and v4.15-rc3. The kernel bisect will require testing of about 7-10 test kernels.

I built the first test kernel, up to the following commit:
78d9b048446c3c0a83313444fb706f8f3bccdae7

The test kernel can be downloaded from:
http://kernel.ubuntu.com/~jsalisbury/lp1761751

Can you test that kernel and report back if it has the bug or not? I will build the next test kernel based on your test results.

Thanks in advance

Kernel 4.15rc2+78d9b048446c3c0a83313444fb706f8f3bccdae7: Black screen.

Note: I installed the linux-image-...-generic deb only as the header ones seemed to require libssl1.1...?

Anyway, I did some research on my own as well.

First, I had a look at the commits between 4.15rc2 and 4.15rc3. The only one I could see being relevant is this one:

commit a703c55004e1c5076d57e43771b3e11117796ea0
Author: Daniel Vetter <email address hidden>
Date: Mon Dec 4 21:48:18 2017 +0100

    drm: safely free connectors from connector_iter

...it seems relevant because the error I get is "Cannot find any crtc or sizes", and the code emitting this error message (in drivers/gpu/drm/drm_fb_helper.c) deals with connectors.

This commit is then referenced in other commits later on as being required to fix, but since 4.16 is "black screen" too those fixes seems not have helped here.

Possibly upstream bug found: https://bugs.freedesktop.org/show_bug.cgi?id=105880

I built the first test kernel, up to the following commit:
6a5e05a47b6cb8e59bfad351444322b1e4012326

The test kernel can be downloaded from:
http://kernel.ubuntu.com/~jsalisbury/lp1761751

Can you test that kernel and report back if it has the bug or not? I will build the next test kernel based on your test results.

 > I built the first test kernel, up to the following commit:

You mean the second test kernel, I presume?

Kernel 6a5e05a47b6cb8e59bfad351444322b1e4012326: Black screen.

Whoops, yes, the last kernel was the second.

I built the next test kernel, up to the following commit:
2391f0b4808e3d5af348324d69f5f45c56a26836

The test kernel can be downloaded from:
http://kernel.ubuntu.com/~jsalisbury/lp1761751

Kernel 2391f0b4808e3d5af348324d69f5f45c56a26836: Black screen.

Kernel 4.17-rc1: Black screen.

I built the next test kernel, up to the following commit:
af57b7fffaf74ff337e479abb19026d72065673b

The test kernel can be downloaded from:
http://kernel.ubuntu.com/~jsalisbury/lp1761751

Kernel af57b7fffaf74ff337e479abb19026d72065673b: Black screen.

I built the next test kernel, up to the following commit:
a0da456bbf95d2a9294799bb05c61bfb24736bb7

The test kernel can be downloaded from:
http://kernel.ubuntu.com/~jsalisbury/lp1761751

Kernel a0da456bbf95d2a9294799bb05c61bfb24736bb7: Black screen.

This is starting to feel a bit fishy but maybe git knows what it's doing. :-)

There's nothing graphics related here:

$ git log v4.15-rc2..a0da456bbf95d2a9294799bb --oneline
a0da456 net: phy-micrel: check return code in flp center function
c7799c0 tipc: call tipc_rcv() only if bearer is up in tipc_udp_recv()
cfac7f8 tcp/dccp: block bh before arming time_wait timer
b484d8a Merge branch 'sctp-prsctp-chunk-fixes'
779edd7 sctp: do not abandon the other frags in unsent outq if one msg has outstanding frags
e5f6129 sctp: abandon the whole msg if one part of a fragmented message is abandoned
d30fc51 sctp: only update outstanding_bytes for transmitted queue when doing prsctp_prune
6fef90c net: dsa: bcm_sf2: Set correct CHAIN_ID and slice number mask
f859b4a sit: update frag_off info
3016dad tcp: remove buggy call to tcp_v6_restore_cb()
90a6ec8 act_sample: get rid of tcf_sample_cleanup_rcu()
6c9257a Merge tag 'rxrpc-fixes-20171129' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs
1c08ac0 net: stmmac: dwmac-sun8i: fix allwinner,leds-active-low handling
f8821f9 skbuff: Grammar s/are can/can/, s/change/changes/
a154f8e net: mvpp2: allocate zeroed tx descriptors
282ef47 rxrpc: Fix variable overwrite
5fc62f6 rxrpc: Fix ACK generation from the connection event processor
3d7682a rxrpc: Clean up whitespace

Yet I have double checked that, according /var/log/kern.log,

 - I booted this kernel:

Apr 18 13:18:29 david-desktop kernel: [ 0.000000] Linux version 4.15.0-041500rc1-generic (jsalisbury@kathleen) (gcc version 7.3.0 (Ubuntu 7.3.0-16ubuntu3)) #20180417120

 - I got black screen and the following message:

Apr 18 13:18:29 david-desktop kernel: [ 1.667490] [drm] Cannot find any crtc or sizes

And when I last booted this kernel:

Apr 17 06:17:34 david-desktop kernel: [ 0.000000] Linux version 4.15.0-041500rc2-generic (kernel@gloin) (gcc version 7.2.0 (Ubuntu 7.2.0-8ubuntu3)) #201712031230 SMP Sun Dec 3 17:32:03 UTC 2017

...it works just fine.

There are only 3 kernels left to test, so I hope we hit a good results soon.

I built the next test kernel, up to the following commit:
3016dad75b48279e579117ee3ed566ba90a3b023

The test kernel can be downloaded from:
http://kernel.ubuntu.com/~jsalisbury/lp1761751

Linux version 4.15.0-041500rc1-generic (jsalisbury@kathleen) (gcc version 7.3.0 (Ubuntu 7.3.0-16ubuntu3)) #201804181224: Black screen

I built the next test kernel, up to the following commit:
282ef4729195c8503f7101d574acfb5e7c8a8209

The test kernel can be downloaded from:
http://kernel.ubuntu.com/~jsalisbury/lp1761751

Kernel 282ef4729195c8503f7101d574acfb5e7c8a8209: Black screen.

I built the next test kernel, up to the following commit:
5fc62f6a139a7b06b027bf442cd4205619506f59

The test kernel can be downloaded from:
http://kernel.ubuntu.com/~jsalisbury/lp1761751

Kernel 5fc62f6a139a7b06b027bf442cd4205619506f59: Black screen.

I built the next test kernel, up to the following commit:
3d7682af228fd78dc46bc6bf40e0268ad04521ec

The test kernel can be downloaded from:
http://kernel.ubuntu.com/~jsalisbury/lp1761751

Kernel 3d7682af228fd78dc46bc: Black screen.

Yeah, so I doubt it's a whitespace fix in the rxrpc driver. However, a diff between the configs on v415-rc2 and v415-rc3 has lots of differences, the most interesting being this one:

$ diff -u config-4.15.0-041500rc2-generic config-4.15.0-041500rc3-generic
/* ... snip ... */
@@ -5675,9 +5677,9 @@
 # Display Engine Configuration
 #
 CONFIG_DRM_AMD_DC=y
-# CONFIG_DRM_AMD_DC_PRE_VEGA is not set
+CONFIG_DRM_AMD_DC_PRE_VEGA=y
 # CONFIG_DRM_AMD_DC_FBC is not set
-# CONFIG_DRM_AMD_DC_DCN1_0 is not set
+CONFIG_DRM_AMD_DC_DCN1_0=y
 # CONFIG_DEBUG_KERNEL_DC is not set

I've also been building all of the bisect testing kernel with Bionic configs, which may explain why they are all bad.

I'll build a bionic test kernel with these two configs reverted back.

I built a Bionic test kernel with the following two config options not set:

CONFIG_DRM_AMD_DC_PRE_VEGA
CONFIG_DRM_AMD_DC_DCN1_0

The test kernel can be downloaded from:
http://kernel.ubuntu.com/~jsalisbury/lp1761751

Also, with this test kernel both the linux-image and linux-image-extra .debs need to be installed.

The test kernel (4.15.0-15-generic #16~lp1761751) boots normally.

I built a Bionic test kernel with the following one config option not set:

CONFIG_DRM_AMD_DC_PRE_VEGA

The test kernel can be downloaded from:
http://kernel.ubuntu.com/~jsalisbury/lp1761751

Kernel 4.15.0-15-generic #16~lp1761751v2: Boots normally

Hi Harry,

A kernel bug report was opened against Ubuntu [0].  After a kernel
bisect, it was found the following commit introduced the bug:

commit 4562236b3bc0a28aeb6ee93b2d8a849a4c4e1c7c
Author: Harry Wentland <email address hidden>
Date:   Tue Sep 12 15:58:20 2017 -0400

    drm/amd/dc: Add dc display driver (v2)

The regression was introduced as of v4.15-rc1 and still exists in
current mainline.  The commit does not need to be reverted to resolve
the bug.  Disabling the CONFIG_DRM_AMD_DC_PRE_VEGA option makes the bug
go away.

I was hoping to get your feedback, since you are the patch author.  Do
you think gathering any additional data will help diagnose this issue?

Thanks,

Joe

[0] http://pad.lv/1761751

What display connectors are being used?

If this is about displays with VGA or LVDS connections this is the same bug as upstream bug https://bugs.freedesktop.org/show_bug.cgi?id=105880 mentioned above.

Replied on the ticket.

If this is about non-functioning LVDS or VGA I'm aware of it and trying to find time to find a good solution.

Harry

On 2018-04-30 11:14 AM, Joseph Salisbury wrote:
> Hi Harry,
>
> A kernel bug report was opened against Ubuntu [0].  After a kernel
> bisect, it was found the following commit introduced the bug:
>
>
> commit 4562236b3bc0a28aeb6ee93b2d8a849a4c4e1c7c
> Author: Harry Wentland <email address hidden>
> Date:   Tue Sep 12 15:58:20 2017 -0400
>
>     drm/amd/dc: Add dc display driver (v2)
>
>
> The regression was introduced as of v4.15-rc1 and still exists in
> current mainline.  The commit does not need to be reverted to resolve
> the bug.  Disabling the CONFIG_DRM_AMD_DC_PRE_VEGA option makes the bug
> go away.
>
> I was hoping to get your feedback, since you are the patch author.  Do
> you think gathering any additional data will help diagnose this issue?
>
>
> Thanks,
>
> Joe
>
>
> [0] http://pad.lv/1761751
>

I don't think it is the same bug. I have a card with HDMI and DVI
connectors (XFX RX 460). The HDMI connector is connected to the monitor,
the DVI connector is unplugged.

// David

On 2018-04-30 17:40, Harry Wentland wrote:
> Replied on the ticket.
>
> If this is about non-functioning LVDS or VGA I'm aware of it and trying
> to find time to find a good solution.
>
> Harry
>
> On 2018-04-30 11:14 AM, Joseph Salisbury wrote:
>> Hi Harry,
>>
>> A kernel bug report was opened against Ubuntu [0].  After a kernel
>> bisect, it was found the following commit introduced the bug:
>>
>>
>> commit 4562236b3bc0a28aeb6ee93b2d8a849a4c4e1c7c
>> Author: Harry Wentland <email address hidden>
>> Date:   Tue Sep 12 15:58:20 2017 -0400
>>
>>     drm/amd/dc: Add dc display driver (v2)
>>
>>
>> The regression was introduced as of v4.15-rc1 and still exists in
>> current mainline.  The commit does not need to be reverted to resolve
>> the bug.  Disabling the CONFIG_DRM_AMD_DC_PRE_VEGA option makes the bug
>> go away.
>>
>> I was hoping to get your feedback, since you are the patch author.  Do
>> you think gathering any additional data will help diagnose this issue?
>>
>>
>> Thanks,
>>
>> Joe
>>
>>
>> [0] http://pad.lv/1761751
>>

Is it HDMI from the board to HDMI on the monitor? I'd like to rule out any use of an adapter or HDMI-to-DVI cable.

What monitor is this?

Is the monitor detected?

What do you see in the 'status', 'enabled', 'dpms', and 'modes' files in /sys/class/drm/card0-HDMI-A-1/ ?

I'm not sure if you asked for the contents of these files from a working
or non-working kernel, but I'll take the working kernel since that's
obviously a lot easier.

The monitor, Dell U2415, is connected via its input source "HDMI (MHL)
2". The monitor name "DELL U2415" shows up in the edid file.

status: connected

enabled: enabled

dpms: On

modes:

1920x1200
1920x1080
1920x1080
1920x1080i
1920x1080i
1920x1080
1920x1080i
1920x1080
1920x1080
1920x1080
1920x1080
1920x1080
1600x1200
1280x1024
1280x1024
1152x864
1280x720
1280x720
1280x720
1024x768
1024x768
800x600
800x600
720x576
720x576i
720x480
720x480
720x480i
720x480i
640x480
640x480
640x480
720x400

// David

On 2018-05-02 16:11, Harry Wentland wrote:
> Is it HDMI from the board to HDMI on the monitor? I'd like to rule out
> any use of an adapter or HDMI-to-DVI cable.
>
> What monitor is this?
>
> Is the monitor detected?
>
> What do you see in the 'status', 'enabled', 'dpms', and 'modes' files in
> /sys/class/drm/card0-HDMI-A-1/ ?
>

Thanks, mind posting the same for the non-working kernel? I'm curious if the problem is
1) we don't detect the display
2) we don't enable the display
3) we enable the display but wrong

I booted the 4.17rc1 kernel.

In order to see output I connected an old TV to the DVI output, and that
output was successfully detected (through an DVI-HDMI adapter, if that
matters).

The output of /sys/class/drm/card0-HDMI-A-1 in this case:

dpms: Off

enabled: disabled

modes: (empty)

status: disconnected

I then tried switching which monitor was connected to which port.
Interestingly, in this case both the Dell U2415 and the old TV could
show output successfully.

// David

On 2018-05-02 19:50, Harry Wentland wrote:
> Thanks, mind posting the same for the non-working kernel? I'm curious if the problem is
> 1) we don't detect the display
> 2) we don't enable the display
> 3) we enable the display but wrong
>

Some information from me, using 4.16.8 kernel
System is Lenovo Z50-75 laptop with AMD-A10 cpu (Kaveri) and both HDMI and VGA ports. I use the VGA port to which is connected an Acer X223HQ display, capable of displaying up to 1920x1080. The laptop screen is also 1920x1080

If system is booted with the following kernel options

amdgpu.cik_support=1 radeon.cik_support=0 amdgpu.dc=1

I get a black/blank VGA screen.

If system is booted with

amdgpu.cik_support=1 radeon.cik_support=0 amdgpu.dc=0

I get a working VGA display in addition to the laptop screen.

With amdgpu.dc=1 the contents of /sys/class/drm are:
card0
card0-eDP-1
card0-HDMI-A-1
renderD128
ttm
version

With amdgpu.dc=0 the contents of /sys/class/drm are:
card0
card0-eDP-1
card0-HDMI-A-1
card0-VGA-1
renderD128
ttm
version

With amdgpu.dc=0, dmesg lines containing 'drm' are

[ 1.749135] [drm] radeon kernel modesetting enabled.
[ 1.750430] fb: switching to radeondrmfb from EFI VGA
[ 1.795937] [drm] amdgpu kernel modesetting enabled.
[ 1.804524] [drm] initializing kernel modesetting (KAVERI 0x1002:0x130A 0x17AA:0x3988 0x00).
[ 1.804544] [drm] register mmio base: 0xF0B00000
[ 1.804547] [drm] register mmio size: 262144
[ 1.828764] [drm] BIOS signature incorrect 0 0
[ 1.829561] [drm] vm size is 64 GB, 2 levels, block size is 10-bit, fragment size is 9-bit
[ 1.829586] [drm] Detected VRAM RAM=1024M, BAR=1024M
[ 1.829589] [drm] RAM width 64bits UNKNOWN
[ 1.829727] [drm] amdgpu: 1024M of VRAM memory ready
[ 1.829731] [drm] amdgpu: 3072M of GTT memory ready.
[ 1.829747] [drm] GART: num cpu pages 262144, num gpu pages 262144
[ 1.829797] [drm] PCIE GART of 1024M enabled (table at 0x000000F400040000).
[ 1.829885] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
[ 1.829888] [drm] Driver supports precise vblank timestamp query.
[ 1.829920] [drm] Internal thermal controller without fan control
[ 1.829923] [drm] amdgpu: dpm initialized
[ 1.830640] [drm] amdgpu atom DIG backlight initialized
[ 1.830644] [drm] AMDGPU Display Connectors
[ 1.830647] [drm] Connector 0:
[ 1.830649] [drm] VGA-1
[ 1.830651] [drm] HPD2
[ 1.830654] [drm] DDC: 0x1950 0x1950 0x1951 0x1951 0x1952 0x1952 0x1953 0x1953
[ 1.830657] [drm] Encoders:
[ 1.830659] [drm] CRT1: INTERNAL_UNIPHY2
[ 1.830661] [drm] CRT1: NUTMEG
[ 1.830663] [drm] Connector 1:
[ 1.830665] [drm] HDMI-A-1
[ 1.830667] [drm] HPD3
[ 1.830670] [drm] DDC: 0x1954 0x1954 0x1955 0x1955 0x1956 0x1956 0x1957 0x1957
[ 1.830673] [drm] Encoders:
[ 1.830675] [drm] DFP1: INTERNAL_UNIPHY2
[ 1.830677] [drm] Connector 2:
[ 1.830679] [drm] eDP-1
[ 1.830681] [drm] HPD1
[ 1.830684] [drm] DDC: 0x194c 0x194c 0x194d 0x194d 0x194e 0x194e 0x194f 0x194f
[ 1.830687] [drm] Encoders:
[ 1.830690] [drm] LCD1: INTERNAL_UNIPHY
[ 1.831974] [drm] Found UVD firmware Version: 1.55 Family ID: 9
[ 1.832424] [drm] Found VCE firmware Version: 50.10 Binary ID: 2
[ 1.923930] [drm] UVD initialized successfully.
[ 2.042436] [drm] VCE initialized successfully.
[ 3.097431] [drm] fb mappable at 0xA0428000
[ 3.097436] [drm] vram a...

Problem still exists in 4.16.9 kernel from http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.16.9/
(#uname -r output is 4.16.9-041609-generic )

Problem still exists in 4.16.10 kernel from http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.16.10/
(#uname -r output is 4.16.10-041610-generic )

dmesg output gives same error output as before when amdgpu.dc=1

 [ 1.866312] [drm] amdgpu: 1024M of VRAM memory ready
 [ 1.866316] [drm] amdgpu: 3072M of GTT memory ready.
 [ 1.866332] [drm] GART: num cpu pages 262144, num gpu pages 262144
 [ 1.866382] [drm] PCIE GART of 1024M enabled (table at 0x000000F400040000).
 [ 1.866500] [drm] Internal thermal controller without fan control
 [ 1.866504] [drm] amdgpu: dpm initialized
 [ 1.867685] [drm] Found UVD firmware Version: 1.55 Family ID: 9
 [ 1.868113] [drm] Found VCE firmware Version: 50.10 Binary ID: 2
 [ 1.870125] [drm:construct [amdgpu]] *ERROR* construct: Invalid Connector ObjectID from Adapter Service for connector index:3! type 0 expected 3
 [ 1.870214] [drm:construct [amdgpu]] *ERROR* construct: Invalid Connector ObjectID from Adapter Service for connector index:4! type 0 expected 3
 [ 1.870337] [drm:construct [amdgpu]] *ERROR* construct: Invalid Connector ObjectID from Adapter Service for connector index:5! type 0 expected 3
 [ 1.870445] [drm:construct [amdgpu]] *ERROR* construct: Invalid Connector ObjectID from Adapter Service for connector index:6! type 0 expected 3
 [ 1.880211] [drm] Display Core initialized with v3.1.27!
 [ 1.887329] [drm] SADs count is: -2, don't need to read it
 [ 1.887755] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
 [ 1.887761] [drm] Driver supports precise vblank timestamp query.
 [ 1.906232] [drm] UVD initialized successfully.

Problem persists in kernel 4.16.11

#uname -r
4.16.11-041611-generic

#ls /sys/class/drm
card0 card0-eDP-1 card0-HDMI-A-1 renderD128 ttm version

#dmesg | grep drm
[ 1.783188] [drm] radeon kernel modesetting enabled.
[ 1.784832] fb: switching to radeondrmfb from EFI VGA
[ 1.831813] [drm] amdgpu kernel modesetting enabled.
[ 1.840215] [drm] initializing kernel modesetting (KAVERI 0x1002:0x130A 0x17AA:0x3988 0x00).
[ 1.840237] [drm] register mmio base: 0xF0B00000
[ 1.840239] [drm] register mmio size: 262144
[ 1.865264] [drm] BIOS signature incorrect 0 0
[ 1.865367] [drm] vm size is 64 GB, 2 levels, block size is 10-bit, fragment size is 9-bit
[ 1.865387] [drm] Detected VRAM RAM=1024M, BAR=1024M
[ 1.865390] [drm] RAM width 64bits UNKNOWN
[ 1.865779] [drm] amdgpu: 1024M of VRAM memory ready
[ 1.865783] [drm] amdgpu: 3072M of GTT memory ready.
[ 1.865807] [drm] GART: num cpu pages 262144, num gpu pages 262144
[ 1.865894] [drm] PCIE GART of 1024M enabled (table at 0x000000F400040000).
[ 1.866040] [drm] Internal thermal controller without fan control
[ 1.866046] [drm] amdgpu: dpm initialized
[ 1.867349] [drm] Found UVD firmware Version: 1.55 Family ID: 9
[ 1.867789] [drm] Found VCE firmware Version: 50.10 Binary ID: 2
[ 1.869984] [drm:construct [amdgpu]] *ERROR* construct: Invalid Connector ObjectID from Adapter Service for connector index:3! type 0 expected 3
[ 1.870074] [drm:construct [amdgpu]] *ERROR* construct: Invalid Connector ObjectID from Adapter Service for connector index:4! type 0 expected 3
[ 1.870200] [drm:construct [amdgpu]] *ERROR* construct: Invalid Connector ObjectID from Adapter Service for connector index:5! type 0 expected 3
[ 1.870309] [drm:construct [amdgpu]] *ERROR* construct: Invalid Connector ObjectID from Adapter Service for connector index:6! type 0 expected 3
[ 1.884255] [drm] Display Core initialized with v3.1.27!
[ 1.891240] [drm] SADs count is: -2, don't need to read it
[ 1.891668] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
[ 1.891674] [drm] Driver supports precise vblank timestamp query.
[ 1.910078] [drm] UVD initialized successfully.
[ 2.026666] [drm] VCE initialized successfully.
[ 2.056074] [drm] fb mappable at 0xA042A000
[ 2.056083] [drm] vram apper at 0xA0000000
[ 2.056087] [drm] size 8294400
[ 2.056090] [drm] fb depth is 24
[ 2.056093] [drm] pitch is 7680
[ 2.056218] fbcon: amdgpudrmfb (fb0) is primary device
[ 2.113975] amdgpu 0000:00:01.0: fb0: amdgpudrmfb frame buffer device
[ 2.130285] [drm] Initialized amdgpu 3.23.0 20150101 for 0000:00:01.0 on minor 0

Problem persists in kernel 4.17-rc6 from http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.17-rc6/

#uname -r
4.17.0-041700rc6-generic

#cat /proc/cmdline
BOOT_IMAGE=/vmlinuz-4.17.0-041700rc6-generic <redacted> amdgpu.cik_support=1 radeon.cik_support=0 amdgpu.dc=1

#ls /sys/class/drm
card0 card0-eDP-1 card0-HDMI-A-1 renderD128 ttm version <<--***NOTE no VGA***

#dmesg | grep drm
[ 2.326305] [drm] radeon kernel modesetting enabled.
[ 2.327644] fb: switching to radeondrmfb from EFI VGA
[ 2.372539] [drm] amdgpu kernel modesetting enabled.
[ 2.380846] [drm] initializing kernel modesetting (KAVERI 0x1002:0x130A 0x17AA:0x3988 0x00).
[ 2.380878] [drm] register mmio base: 0xF0B00000
[ 2.380880] [drm] register mmio size: 262144
[ 2.380889] [drm] add ip block number 0 <cik_common>
[ 2.380892] [drm] add ip block number 1 <gmc_v7_0>
[ 2.380894] [drm] add ip block number 2 <cik_ih>
[ 2.380897] [drm] add ip block number 3 <kv_dpm>
[ 2.380900] [drm] add ip block number 4 <dm>
[ 2.380903] [drm] add ip block number 5 <gfx_v7_0>
[ 2.380906] [drm] add ip block number 6 <cik_sdma>
[ 2.380909] [drm] add ip block number 7 <uvd_v4_2>
[ 2.380911] [drm] add ip block number 8 <vce_v2_0>
[ 2.404721] [drm] BIOS signature incorrect 0 0
[ 2.404832] [drm] vm size is 64 GB, 2 levels, block size is 10-bit, fragment size is 9-bit
[ 2.404855] [drm] Detected VRAM RAM=1024M, BAR=1024M
[ 2.404858] [drm] RAM width 64bits UNKNOWN
[ 2.405130] [drm] amdgpu: 1024M of VRAM memory ready
[ 2.405133] [drm] amdgpu: 3072M of GTT memory ready.
[ 2.405145] [drm] GART: num cpu pages 262144, num gpu pages 262144
[ 2.405180] [drm] PCIE GART of 1024M enabled (table at 0x000000F400040000).
[ 2.405266] [drm] Internal thermal controller without fan control
[ 2.405269] [drm] amdgpu: dpm initialized
[ 2.406296] [drm] Found UVD firmware Version: 1.55 Family ID: 9
[ 2.406649] [drm] Found VCE firmware Version: 50.10 Binary ID: 2
[ 2.407995] [drm:construct [amdgpu]] *ERROR* construct: Invalid Connector ObjectID from Adapter Service for connector index:3! type 0 expected 3
[ 2.408115] [drm:construct [amdgpu]] *ERROR* construct: Invalid Connector ObjectID from Adapter Service for connector index:4! type 0 expected 3
[ 2.408204] [drm:construct [amdgpu]] *ERROR* construct: Invalid Connector ObjectID from Adapter Service for connector index:5! type 0 expected 3
[ 2.408324] [drm:construct [amdgpu]] *ERROR* construct: Invalid Connector ObjectID from Adapter Service for connector index:6! type 0 expected 3
[ 2.420247] [drm] Display Core initialized with v3.1.38!
[ 2.426796] [drm] SADs count is: -2, don't need to read it
[ 2.427373] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
[ 2.427379] [drm] Driver supports precise vblank timestamp query.
[ 2.446239] [drm] UVD initialized successfully.
[ 2.561743] [drm] VCE initialized successfully.
[ 2.569100] [drm] fb mappable at 0xA042A000
[ 2.569106] [drm] vram apper at 0xA0000000
[ 2.569109] [drm] size 8294400
[ 2.569111] [drm] fb depth is 24
[ 2.569113] [drm] pitch is 7680
[ 2.569233] fbcon: amdgpudrmfb (fb0) is primary de...

HI guys,

Im trying to use my second Monitor with this videocard, i do not know if i have the same issue but my HDMI monitor just would not work. it worked before the update and only with kernels until 4.13... i don't know if i have to put some info here or somewhere else.

thanks.

I'm getting exactly this issue. An ATI RX 480 on Kubuntu 18.04 with a ASUS MG28U on HDMI-1. It is on a 5m cable so using the displayport wasn't an option. The long HDMI cable is to allow the large system box to be away from my desk. I mostly use the system as a web and data server and it is stuffed with hard drives.

Works perfectly on 4.13 kernels. Fails on the 4.15 kernels before the sddm login screen.

4.15.0-43-generic <-- screen repeats HDMI no signal
4.15.0-42-generic <-- screen repeats HDMI no signal
4.15.0-35-generic <-- screen repeats HDMI no signal
4.13.0-36-generic <-- ok

On 4.13 kernels it runs the screen at 3840x2160 @ 30Hz. On 4.15 kernels it doesn't.

At a guess looking at the syslog, the 4.13 fell back to 30Hz at the hi-res while the 4.15 does not. Over the long cable the screen isn't detecting the screen at 60Hz.

Hope that helps. I may have to look at getting a longer active cable...

Seems to be another AMDGPU DC issue...

Please try mainline kernel v4.20-rc6.
If it doesn't solve the issue, please file an upstream bug at
https://bugs.freedesktop.org/
Product: DRI
Component: DRM/amdgpu

I have previously created an upstream bug report for this:

https://bugs.freedesktop.org/show_bug.cgi?id=107668

Just as an update, in the Freedesktop issue, I found using a LOWER specced/older cable to fix the problem.

I do not know how to debug it further but I suspect the bug is in the support of later version of the HDMI standard.