New install. Half of the times Leap boot to a black screen and I need to do a restart
power on
grub menu shows up
I leave as it is
screen goes black and that’s it
sudo journalctl -p 3 --since=today
-- Logs begin at Mon 2019-10-28 07:20:03 PDT, end at Mon 2019-10-28 07:28:19 PDT. --
Oct 28 07:20:04 linux-wy8o kernel: amdgpu 0000:01:00.0: Invalid PCI ROM header signature: expecting 0xaa55, got 0xffff
Oct 28 07:20:11 ethelbert kernel: sp5100_tco: I/O address 0x0cd6 already in use
Oct 28 07:20:18 ethelbert chronyd[1151]: Could not open IPv6 command socket : Address family not supported by protocol
Oct 28 07:20:24 ethelbert NetworkManager[1110]: <error> [1572272424.7807] device (eth0): addrconf6: failed to start neighbor discovery: failure creating libndp socket: Address family not supported by protocol (97)
Oct 28 07:21:16 ethelbert pulseaudio[2575]: [pulseaudio] bluez5-util.c: GetManagedObjects() failed: org.freedesktop.DBus.Error.TimedOut: Failed to activate service 'org.bluez': timed out (service_start_timeout=25000ms)
Strike the E key when the Grub menu appears, move to the end of the wrapped line that began with linu, and append a space and plymouth.enable=0 before proceeding with boot. Does it help?
The black screen after the grub menu happens more often in a cold boot, the computer being turned off for a while. I would say around 1/2 of the time.
The black screen after the grub menu happens a lot less often in a warm boot, the computer being turned off for a short time. I would say around 1/4 of the time.
As anyone should be able to infer reading all the way through that thread, that’s not an option in any recent release of any distro. AMD stopped releasing its proprietary drivers quite some time back, instead investing its effort making the FOSS amdgpu drivers that ship with openSUSE work.
What should work purely for troubleshooting purposes and allowing access to YaST and zypper instead of getting black screened is to boot with nomodeset, to ensure needed packages are installed. The amdgpu drivers depend on modesetting to work, so nomodeset cannot be depended upon. At the basic level you should have xf86-video-ati not installed, and should have xf86-video-amdgpu installed.
Which mesa packages are or cmdline options that might be required with Polaris GPUs I don’t know.
Does the same problem occur if you use a different video cable of same type or different type, or connect the PC to some other display?
For trouble shooting purposes, on your first boot using nomodeset on cmdline, please upload /var/log/Xorg.0.log.old to https://susepaste.org or http://pastebin.com or equivalent for our examination. There should be clues in that file as to what is going wrong, lines containing (EE).
I replied earlier this morning, but I don’t see my reply. I’m sure I clicked on submit reply but the peanut must be becoming mush…
I have removed them all except for the xf86-video-amdgpu
Yes, I took my computer to a neighbour and it also hanged. The screen was an old 19" boob-tube.
I have found what the problem is, but I don’t know how to fix it. It’s not xorg. It dies long before that.
2.843560] fb: switching to amdgpudrmfb from VESA VGA <<<<<<<<<<<<<<<<<<<<<<
2.865051] ehci-pci 0000:00:13.2: irq 17, io mem 0xfeb07000
2.868806] Console: switching to colour dummy device 80x25
2.869138] [drm] initializing kernel modesetting (POLARIS10 0x1002:0x67DF 0x1462:0x341B 0xEF).
2.869149] [drm] register mmio base: 0xFEA00000
2.869153] [drm] register mmio size: 262144
2.869164] [drm] add ip block number 0 <vi_common>
2.869168] [drm] add ip block number 1 <gmc_v8_0>
2.869171] [drm] add ip block number 2 <tonga_ih>
2.869174] [drm] add ip block number 3 <powerplay>
2.869178] [drm] add ip block number 4 <dm>
2.869181] [drm] add ip block number 5 <gfx_v8_0>
2.869184] [drm] add ip block number 6 <sdma_v3_0>
2.869188] [drm] add ip block number 7 <uvd_v6_0>
2.869191] [drm] add ip block number 8 <vce_v3_0>
It dies at switching the frame buffer. What I’m showing is dmesg from a successful start. When the start fails, it dies at the switching to amdgpudrmfb.
Other cmdline parameters to try instead of nomodeset:
iommu=soft
amd_iommu=off
I don’t know much about troubleshooting recent AMD GPUs. My newest AMD GPU is nearly a decade old. If these don’t help you really need one of our AMD experts to weigh in here.
It was more complicated than that, like I said the problem was the switching to the frame buffer, by disabling it, so far it’s working. Here are my boot parameters