20.04 Upgraded graphics card, now freeze after login

[Agent24]

I have an installation of Ubuntu Studio 20.04 that has seen graphics cards upgrades from Radeon HD5770, then R9 270X, then RX570 (all without issue) until now.

I just swapped my RX570 for an RX 5700XT and now I can't login, unless I boot with NOMODESET enabled.

I can see the login screen just fine, but once it accepts my password, there is some HDD activity before the mouse freezes, keyboard won't respond, and I can't recover.

Has anyone an idea of where to start looking?

[Agent24]

I tried installing the 5.8 kernel through the HWE package.

Now I can login without freezing, but the display is corrupted, chopped up into strips, and skewed, like shredded paper, horizontally. I can see enough to logoff/reboot, but it's nowhere near usable.

I then tried a trick that helped me in the past with video cards, in 18.04, downloading a newer version of the Linux-firmware package from a newer distro (in this case I tried 190.3 from 20.10) but this made no difference.

Perhaps an old config file or such is causing problems? If so, which one might it be?

[ajgreeny]

It may help us if you can show us the output of command
inxi -Fzx
You may have to install inxi package., but I believe it is a default application in 20.04.

[Agent24]

Here is the output:
I am at the desktop but with nomodeset enabled.



System: Kernel: 5.8.0-43-lowlatency x86_64 bits: 64 compiler: N/A Desktop: Xfce 4.14.2
Distro: Ubuntu 20.04.2 LTS (Focal Fossa)
Machine: Type: Desktop Mobo: Gigabyte model: GA-78LMT-USB3 6.0 v: x.x serial: <filter> BIOS: Award v: F2 date: 11/25/2014
CPU: Topology: 8-Core model: AMD FX-8370 bits: 64 type: MCP arch: Bulldozer L2 cache: 2048 KiB
flags: avx lm nx pae sse sse2 sse3 sse4_1 sse4_2 sse4a ssse3 svm bogomips: 64444
Speed: 1434 MHz min/max: 1400/4000 MHz Core speeds (MHz): 1: 1518 2: 1409 3: 1408 4: 1414 5: 1406 6: 1405 7: 1527
8: 1443
Graphics: Device-1: Advanced Micro Devices [AMD/ATI] Navi 10 [Radeon RX 5600 OEM/5600 XT / 5700/5700 XT]
vendor: Sapphire Limited driver: N/A bus ID: 03:00.0
Display: x11 server: X.Org 1.20.9 driver: ati,fbdev unloaded: modesetting,radeon,vesa resolution: 1680x1050~77Hz
OpenGL: renderer: llvmpipe (LLVM 11.0.0 256 bits) v: 4.5 Mesa 20.2.6 direct render: Yes
Audio: Device-1: Advanced Micro Devices [AMD/ATI] SBx00 Azalia vendor: Gigabyte driver: snd_hda_intel v: kernel
bus ID: 00:14.2
Device-2: Advanced Micro Devices [AMD/ATI] Navi 10 HDMI Audio driver: snd_hda_intel v: kernel bus ID: 03:00.1
Device-3: C-Media Audio Device type: USB driver: snd-usb-audio bus ID: 8-1.1:3
Device-4: Generic ICT Camera type: USB driver: snd-usb-audio,uvcvideo bus ID: 1-3:2
Sound Server: ALSA v: k5.8.0-43-lowlatency
Network: Device-1: Realtek RTL8111/8168/8411 PCI Express Gigabit Ethernet vendor: Gigabyte driver: r8169 v: kernel
port: ce00 bus ID: 05:00.0
IF: enp5s0 state: up speed: 1000 Mbps duplex: full mac: <filter>
Drives: Local Storage: total: 1.82 TiB used: 484.94 GiB (26.0%)
ID-1: /dev/sda vendor: Western Digital model: WD1003FZEX-00K3CA0 size: 931.51 GiB
ID-2: /dev/sdb vendor: Western Digital model: WD10EZEX-22MFCA0 size: 931.51 GiB
Partition: ID-1: / size: 913.74 GiB used: 484.54 GiB (53.0%) fs: ext4 dev: /dev/dm-1
ID-2: /boot size: 703.5 MiB used: 406.5 MiB (57.8%) fs: ext4 dev: /dev/sda5
ID-3: swap-1 size: 976.0 MiB used: 0 KiB (0.0%) fs: swap dev: /dev/dm-2
Sensors: System Temperatures: cpu: 27.0 C mobo: 24.0 C
Fan Speeds (RPM): cpu: 2265 fan-2: 1033 fan-3: 0
Voltages: 12v: N/A 5v: 3.67 3.3v: N/A vbat: 2.90
Info: Processes: 278 Uptime: 5m Memory: 7.76 GiB used: 1.21 GiB (15.5%) Init: systemd runlevel: 5 Compilers: gcc: 9.3.0
Shell: bash v: 5.0.17 inxi: 3.0.38

[MartyBuntu]

What make and model of power supply are you running this system on?

[Agent24]

A brand new Silverstone Strider Gold ST75F-GS (750 Watt) so no problem there. Also I forgot to mention, but the card does work in Windows 7, which I also have installed.

In fact, it's looking like this may be related to my monitor...

I loaded the latest Ubuntu Studio 20.04.2 Live DVD. The same problem happened.
Then I tried swapping the monitor and suddenly the display came up normally.

Now of course the strange thing is, this same monitor works fine in Windows 7 and in Ubuntu with NOMODESET. I just don't know why it doesn't work with the amdgpu driver enabled.
I guess it's not a leftover config file issue otherwise the LiveCD would probably have worked.

At this point I'm even more confused as to why this is happening.
Some bug in the driver causing wrong display timings etc?

[rbmorse]

Are you using an DVD-I, HDMI or DisplayPort connection to the monitor? Same for both monitors tested?

[Agent24]

I tried both monitors using the same HDMI cable in the same port on the graphics card.

But the monitor that doesn't work properly is using an HDMI-DVI adapter, as it does not have an HDMI input.

I doubt it is a hardware incompatibility though, as the same setup works fine in Windows 7.

[Agent24]

I have experimented some more and I found some things but no solution yet.

I tried generating a new modeline with cvt and adding with xrandr in case somehow it was generating an incorrect one, but this made no difference.

I have noticed that there is no corruption if I run the display in 640x480 resolution. Anything higher than that (even 800x600) causes problems.

I also noticed that after running the commands to add and select the modeline with xrandr, I saw these errors in dmesg:


[ 183.620958] [drm] Mode Validation Warning: Pitch support failed validation.
[ 183.620977] [drm] Mode Validation Warning: Pitch support failed validation.
[ 183.621035] ------------[ cut here ]------------
[ 183.621196] WARNING: CPU: 4 PID: 1190 at drivers/gpu/drm/amd/amdgpu/../display/dc/dcn20/dcn20_resource.c:3197 dcn20_validate_bandwidth_fp+0x9d/0xf0 [amdgpu]
[ 183.621197] Modules linked in: uas usb_storage binfmt_misc nls_iso8859_1 uvcvideo snd_hda_codec_realtek videobuf2_vmalloc videobuf2_memops videobuf2_v4l2 snd_hda_codec_generic videobuf2_common ledtrig_audio snd_hda_codec_hdmi videodev snd_hda_intel snd_usb_audio snd_usbmidi_lib snd_intel_dspcfg snd_hda_codec mc snd_hda_core snd_hwdep snd_pcm joydev edac_mce_amd snd_seq_midi snd_seq_midi_event snd_rawmidi snd_seq snd_seq_device snd_timer snd k10temp soundcore fam15h_power ccp kvm wmi_bmof input_leds serio_raw mac_hid sch_fq_codel cuse it87 hwmon_vid parport_pc ppdev lp parport ip_tables x_tables autofs4 dm_crypt hid_generic usbhid hid amdgpu iommu_v2 gpu_sched i2c_algo_bit ttm drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops cec rc_core pata_acpi crct10dif_pclmul crc32_pclmul ghash_clmulni_intel aesni_intel crypto_simd cryptd drm glue_helper r8169 ahci xhci_pci i2c_piix4 pata_atiixp realtek libahci xhci_pci_renesas wmi
[ 183.621229] CPU: 4 PID: 1190 Comm: Xorg Not tainted 5.8.0-43-lowlatency #49~20.04.1-Ubuntu
[ 183.621230] Hardware name: Gigabyte Technology Co., Ltd. GA-78LMT-USB3 6.0/GA-78LMT-USB3 6.0, BIOS F2 11/25/2014
[ 183.621363] RIP: 0010:dcn20_validate_bandwidth_fp+0x9d/0xf0 [amdgpu]
[ 183.621387] Code: 24 a8 1e 00 00 75 29 31 d2 4c 89 e6 4c 89 ef f2 41 0f 11 84 24 d8 25 00 00 e8 df fc ff ff 89 c2 41 22 94 24 a8 1e 00 00 75 39 <0f> 0b eb 02 75 cd f2 0f 10 55 e8 f2 41 0f 11 94 24 d8 25 00 00 48
[ 183.621388] RSP: 0018:ffffbf2cc0c6faf0 EFLAGS: 00010246
[ 183.621390] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 00000000050d8004
[ 183.621390] RDX: 0000000000000000 RSI: bd1592668edb836d RDI: 00000000000311a0
[ 183.621391] RBP: ffffbf2cc0c6fb08 R08: ffff9db8b6d18cd0 R09: 0000000000000487
[ 183.621392] R10: 000000000001898c R11: 0000000000000004 R12: ffff9db8a8560000
[ 183.621393] R13: ffff9db8a20c0000 R14: ffff9db8a2c42800 R15: 0000000000000006
[ 183.621395] FS: 00007fbbdde35a40(0000) GS:ffff9db8b6d00000(0000) knlGS:0000000000000000
[ 183.621396] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 183.621397] CR2: 000055f29851d4f0 CR3: 0000000222238000 CR4: 00000000000406e0
[ 183.621398] Call Trace:
[ 183.621519] dcn20_validate_bandwidth+0x29/0x40 [amdgpu]
[ 183.621639] dc_validate_global_state+0x27c/0x300 [amdgpu]
[ 183.621763] amdgpu_dm_atomic_check+0x91a/0x1060 [amdgpu]
[ 183.621801] drm_atomic_check_only+0x2c7/0x450 [drm]
[ 183.621828] drm_atomic_commit+0x18/0x50 [drm]
[ 183.621850] drm_mode_obj_set_property_ioctl+0x150/0x3b0 [drm]
[ 183.621878] ? drm_mode_obj_find_prop_id+0x40/0x40 [drm]
[ 183.621897] drm_ioctl_kernel+0xae/0xf0 [drm]
[ 183.621917] drm_ioctl+0x234/0x3d0 [drm]
[ 183.621944] ? drm_mode_obj_find_prop_id+0x40/0x40 [drm]
[ 183.622037] amdgpu_drm_ioctl+0x4e/0x80 [amdgpu]
[ 183.622062] ksys_ioctl+0x9d/0xd0
[ 183.622064] __x64_sys_ioctl+0x1a/0x20
[ 183.622067] do_syscall_64+0x49/0xc0
[ 183.622069] entry_SYSCALL_64_after_hwframe+0x44/0xa9
[ 183.622071] RIP: 0033:0x7fbbdd44f50b
[ 183.622073] Code: 0f 1e fa 48 8b 05 85 39 0d 00 64 c7 00 26 00 00 00 48 c7 c0 ff ff ff ff c3 66 0f 1f 44 00 00 f3 0f 1e fa b8 10 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 55 39 0d 00 f7 d8 64 89 01 48
[ 183.622074] RSP: 002b:00007fffb54a1418 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
[ 183.622076] RAX: ffffffffffffffda RBX: 00007fffb54a1450 RCX: 00007fbbdd44f50b
[ 183.622077] RDX: 00007fffb54a1450 RSI: 00000000c01864ba RDI: 000000000000000d
[ 183.622077] RBP: 00000000c01864ba R08: 0000000000000071 R09: 00000000cccccccc
[ 183.622078] R10: 0000559c54172054 R11: 0000000000000246 R12: 0000559c5325be80
[ 183.622079] R13: 000000000000000d R14: 0000000000000fff R15: 0000000000000003
[ 183.622082] ---[ end trace c194dc0e6cea89e2 ]---
[ 183.636535] [drm] Mode Validation Warning: Pitch support failed validation.
[ 183.636560] [drm] Mode Validation Warning: Pitch support failed validation.


I don't know what this means, if it maybe related to the problem, or not.

For fun I will try copying the modeline that Windows 7 uses and see if it works in Linux...

[Agent24]

I almost got my hopes up there - I found a modeline in Windows, which turned out different from the one generated in cvt, but upon trying it out, I get the same issue.


Modeline from cvt:
1680x1050_60.00 146.25 1680 1784 1960 2240 1050 1053 1059 1089 -hsync +vsync
Modeline from Windows:
1680x1050_60.00 147.136 1680 1786 1967 2256 1050 1051 1054 1087 -hsync -vsync

I wonder if there is a bug in the amdgpu driver, given that there is that crash dump in dmesg, and every time I use xrandr to do anything with the display, I get

[drm] Mode Validation Warning: Pitch support failed validation.

I shall test the latest Ubuntu 21.04 Daily Live DVD and see what result that has.

Also, I don't know if it's relevant, but when adding the modelines in xrandr, they seem to add fine, but I also get this error:

X Error of failed request: BadName (named color or font does not exist)
Major opcode of failed request: 140 (RANDR)
Minor opcode of failed request: 16 (RRCreateMode)
Serial number of failed request: 41
Current serial number in output stream: 41

[Agent24]

I think I can safely say this is a software and/or driver bug.

I booted the latest Ubuntu 21.04 Daily Live DVD and there is no issue at all. I get full resolution with no display corruption and the amdgpu module is active and loaded.

I can't seem to install any utils like glxinfo or inxi but the behaviour of the desktop does not suggest it is running without a driver - all compositing effects are smooth, YouTube 4K works, etc.


So the question is - what exactly is causing the problem? How can I figure it out.... and more importantly, fix it....

[Agent24]

After much annoyance and updating packages relating to video drivers, drm etc, to no avail, it tuns out the problem goes away if I disable desktop compositing in xfce window manager settings.....

So maybe it is a bug in xfce? For now I have compositing disabled so I can use my computer, but would like to fix that if possible.


Success - I found this thread: https://forum.xfce.org/viewtopic.php?id=13233

In which someone else described the exact same "shredded" screen issue, and solved it by changing xfwm4 vblank_mode.

Now I can have my compositing and see it too...!