[SOLVED] Black screen at boot

dnl09_ · 2022-06-02 08:31:16

Hey everyone,

as this is my first post on this board, please be patient. :-)

I seem to have the same problem as vejtics. However, I don't use dual-boot with Refind, Arch is installed directly on my disk using the latest image (2022.06.01) with archinstall. I've just selected desktop profile with kde and open-sources-drivers.

First, some hardware info:
I am using an HP x360 435 G7 with an AMD Ryzen 7 4700U with Radeon Graphics.

Output of lspci | grep VGA:

04:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Renoir (rev c2)

What i did so far, was enabling early KMS with amdgpu-Module (as also mentioned in https://wiki.archlinux.org/title/Laptop/HP), regenerated my initramfs-Image. I still encounter the blackscreen-issue.

Here are all outputs any commands you requested to vejtics:

sudo journalctl -b -1
http://0x0.st/oBkf.txt

/etc/mkinitcpio.conf

 
MODULES=(amdgpu)
BINARIES=()
FILES=()
HOOKS=(base udev autodetect keyboard keymap modconf block filesystems fsck)

mkinitcpio -P

==> Building image from preset: /etc/mkinitcpio.d/linux.preset: 'default'
  -> -k /boot/vmlinuz-linux -c /etc/mkinitcpio.conf -g /boot/initramfs-linux.img
==> Starting build: 5.18.1-arch1-1
  -> Running build hook: [base]
  -> Running build hook: [udev]
  -> Running build hook: [autodetect]
  -> Running build hook: [keyboard]
==> WARNING: Possibly missing firmware for module: xhci_pci
  -> Running build hook: [keymap]
  -> Running build hook: [modconf]
  -> Running build hook: [block]
  -> Running build hook: [filesystems]
  -> Running build hook: [fsck]
==> Generating module dependencies
==> Creating zstd-compressed initcpio image: /boot/initramfs-linux.img
==> Image generation successful
==> Building image from preset: /etc/mkinitcpio.d/linux.preset: 'fallback'
  -> -k /boot/vmlinuz-linux -c /etc/mkinitcpio.conf -g /boot/initramfs-linux-fallback.img -S autodetect
==> Starting build: 5.18.1-arch1-1
  -> Running build hook: [base]
  -> Running build hook: [udev]
  -> Running build hook: [keyboard]
==> WARNING: Possibly missing firmware for module: xhci_pci
  -> Running build hook: [keymap]
  -> Running build hook: [modconf]
  -> Running build hook: [block]
==> WARNING: Possibly missing firmware for module: qla2xxx
==> WARNING: Possibly missing firmware for module: bfa
==> WARNING: Possibly missing firmware for module: qla1280
==> WARNING: Possibly missing firmware for module: wd719x
==> WARNING: Possibly missing firmware for module: qed
==> WARNING: Possibly missing firmware for module: aic94xx
  -> Running build hook: [filesystems]
  -> Running build hook: [fsck]
==> Generating module dependencies
==> Creating zstd-compressed initcpio image: /boot/initramfs-linux-fallback.img
==> Image generation successful

Shift+Alt+F12 does nothing while Blackscreen.

/var/log/Xorg.0.log
http://0x0.st/oBkW.json

lsblk -f

NAME        FSTYPE FSVER LABEL UUID                                 FSAVAIL FSUSE% MOUNTPOINTS
zram0                                                                              [SWAP]
nvme0n1                                                                            
├─nvme0n1p1 vfat   FAT32       CEEA-B001                             421.4M    17% /boot
├─nvme0n1p2 ext4   1.0         0749fae1-dd20-4acd-af94-8a9e5fd136c8   13.4G    25% /
└─nvme0n1p3 ext4   1.0         3ae99cbf-68c6-4de5-b1c0-66d23a328316  425.8G     0% /home

lsinitcpio /boot/initramfs-linux.img | grep amdgpu
http://0x0.st/oBk4.txt

Additionally my dmesg output:
http://0x0.st/oBkJ.json

Fast Boot is disabled in bios.

Hopefully you can help me; i'm struggeling on this issue since a month, i think (lately with Arch-Image 2022.05.01). I hoped it's fixed with the new kernel, but i'm encountering the same error.

Best regards,
Daniel

Solution (or at least a workaround):
It seems like the laptop can't get the right VBIOS while booting. As it's working sometimes, sometimes not, there is a workaround here: https://bugzilla.kernel.org/show_bug.cgi?id=188301#c15
The code posted by neoe needs to be merged with the kernel's amdgpu_bios.c -File.

When booting works next time, the code saves the working VBIOS on the disk and uses it as fallback, when no VBIOS is recognized.
But take care: This is overwritten with each kernel upgrade and needs to get patched after.

Last edited by dnl09_ (2022-06-09 12:09:33)

seth · 2022-06-02 14:39:43

@dnl09_ I suspect you're facing a different problem.

[     8.029] (--) AMDGPU(0): Chipset: "Unknown AMD Radeon GPU" (ChipID = 0x1636)
…
[     8.069] (II) AMDGPU(0): glamor X acceleration enabled on AMD RENOIR (LLVM 13.0.1, DRM 3.46, 5.18.1-arch1-1)
[     8.069] (II) AMDGPU(0): glamor detected, initialising EGL layer.
[     8.069] (==) AMDGPU(0): TearFree property default: auto
[     8.069] (==) AMDGPU(0): VariableRefresh: disabled
[     8.069] (==) AMDGPU(0): AsyncFlipSecondaries: disabled
[     8.069] (II) AMDGPU(0): KMS Pageflipping: enabled
=====> [     8.070] (WW) AMDGPU(0): No outputs definitely connected, trying again...
=====> [     8.070] (WW) AMDGPU(0): Unable to find connected outputs - setting 1024x768 initial framebuffer
[     8.070] (II) AMDGPU(0): mem size init: gart size :bffff5ef000 vram size: s:1e716000 visible:1e716000
[     8.070] (==) AMDGPU(0): DPI set to (96, 96)
[     8.070] (==) AMDGPU(0): Using gamma correction (1.0, 1.0, 1.0)
…
=====> [     8.070] (EE) AMDGPU(0): No modes.

Jun 02 09:06:26 ds-archlap kernel: amdgpu 0000:04:00.0: amdgpu: RAS: optional ras ta ucode is not available
Jun 02 09:06:26 ds-archlap kernel: amdgpu 0000:04:00.0: amdgpu: RAP: optional rap ta ucode is not available
Jun 02 09:06:26 ds-archlap kernel: amdgpu 0000:04:00.0: amdgpu: SECUREDISPLAY: securedisplay ta ucode is not available
Jun 02 09:06:26 ds-archlap kernel: amdgpu 0000:04:00.0: amdgpu: SMU is initialized successfully!
Jun 02 09:06:26 ds-archlap kernel: [drm] Unsupported Connector type:21!
Jun 02 09:06:26 ds-archlap kernel: [drm] Unsupported Connector type:21!
Jun 02 09:06:26 ds-archlap kernel: [drm] Unsupported Connector type:21!
Jun 02 09:06:26 ds-archlap kernel: [drm] Unsupported Connector type:21!
Jun 02 09:06:26 ds-archlap kernel: [drm] Unsupported Connector type:21!

What kind of output(s) are supposed to be attached?
You can try to use the modesetting driver (drop xf86-video-amdgpu)

There's also https://lkml.org/lkml/2020/8/10/1173 leading to https://bbs.archlinux.org/viewtopic.php?id=248273 - which seems https://gitlab.freedesktop.org/drm/amd/-/issues/1256
maybe he has ultimately resolved it.

You also have

Jun 02 09:06:26 ds-archlap kernel: [drm] RAM width 128bits DDR4

Make sure to check whether there're BIOS updates available for your device.

Fast Boot is disabled in bios.

ftr, when it comes to "hibernating windows brokes my hardsware" that's the wrong kind of fast boot to worry about (and is usually safe to have enabled)

dnl09_ · 2022-06-02 20:15:41

Thanks for your answer!

Dropping the xf86-video-amdgpu Driver didn't helped. :-( Updating BIOS is not that easy on HP devices. As i can see, i can update it by plugging in ethernet - but for ethernet, i first need any usb-c ethernet-adapter. :-/ But i'll do next time.

For now, i just want to have the build-in-screen working. When it works, i'll plug in 2 Screens, 1 by build-in HDMI and 1 by usb-c adapter.

As described on the links you've posted, the screen sometimes works, most times not. For now it's working and here are my dmesg & Xorg-Log-Outputs:
dmesg
http://0x0.st/oBR4.json

Xorg.0.log
http://0x0.st/oBRJ.json

May any other suggestions to make it work reliable?

seth · 2022-06-03 05:39:27

The working dmesg follows the pattern

[    1.583042] [drm] RAM width 64bits DDR4

As well as Alex explanation on the situation (from the bad boot)

[    1.589633] [drm] BIOS signature incorrect 0 0
[    1.589638] resource sanity check: requesting [mem 0x000c0000-0x000dffff], which spans more than PCI Bus 0000:00 [mem 0x000c0000-0x000cbfff window]
[    1.589640] caller pci_map_rom+0x69/0x180 mapping multiple BARs
[    1.589651] amdgpu 0000:04:00.0: Invalid PCI ROM header signature: expecting 0xaa55, got 0xffff
[    1.592709] [drm] BIOS signature incorrect 0 0
[    1.594316] amdgpu 0000:04:00.0: amdgpu: Fetched VBIOS from platform
[    1.594318] amdgpu: ATOM BIOS: BRT55908.001

"Invalid PCI ROM header signature: expecting 0xaa55, got 0xffff" is far more interesting on google and lead to https://bugzilla.kernel.org/show_bug.cgi?id=188301#c15

Because of the racy nature: is there somehow another OS (windows) running on that system at all?

dnl09_ · 2022-06-03 06:44:06

I'm pretty sure, it can't. There was Windows 11 (i think) running, but i've wiped the complete disk multiple times while installing Arch...

WorMzy · 2022-06-04 11:17:05

Mod note: split from https://bbs.archlinux.org/viewtopic.php?id=276121

dnl09_ · 2022-06-08 14:10:58

Thanks for spliting both threads!

Back to my topic:
The only operating system installed is Arch. So far I have not found any other solution either.

Any other suggestions?

seth · 2022-06-08 14:20:17

https://bugzilla.kernel.org/show_bug.cgi?id=188301#c15

dnl09_ · 2022-06-09 07:14:26

Sorry, I missed that.

At first glance, not much has changed; I still get the blackscreen and can do nothing. However, the blackscreen now comes after the "Reached Target Graphical Interface" message.

What have I done?

I downloaded the 5.18.1 kernel, made the adjustments to the file according to the diff, recompiled the module and replaced it with the old amdgpu module. The module is also loaded according to lsmod; however, there is a "0" at the "used by" column.

Here's the dmesg-Output:
http://0x0.st/oMxw.json

Edit:
I think it worked. As i understand the Code correctly, it caches the VBIOS on a succesful boot and uses it, if no VBIOS is found on the "normal" way? So, then it fixed it. After 5 reboots, it worked and for now, it's still working after 3 reboots.

Last edited by dnl09_ (2022-06-09 07:33:09)

dnl09_ · 2022-06-09 07:25:10

// see edit in https://bbs.archlinux.org/viewtopic.php … 4#p2040174

Last edited by dnl09_ (2022-06-09 07:33:23)

seth · 2022-06-09 09:22:19

As i understand the Code correctly, it caches the VBIOS on a succesful boot and uses it, if no VBIOS is found on the "normal" way?

Yes. It's a workaround for the weird condition where the bad RAM width gets you a broken VBIOS.

Obviously this is gonna break w/ the next kernel update, but the on-disk cached VBIOS should™ work when patching that kernel module.

dnl09_ · 2022-06-09 12:01:16

You're right, it broke at the kernel update, but worked patching the driver. A bit annoying, but ok. Now the screen works, touchscreen too, but touchpad not. This notebook is just annoying... But that's another topic.

I'll mark this thread as solved and add the workaround to my first post. Hope i explained it right.
Thank you very much @seth!

Last edited by dnl09_ (2022-06-09 12:10:05)

Arch Linux

#1 2022-06-02 08:31:16

[SOLVED] Black screen at boot

#2 2022-06-02 14:39:43

Re: [SOLVED] Black screen at boot

#3 2022-06-02 20:15:41

Re: [SOLVED] Black screen at boot

#4 2022-06-03 05:39:27

Re: [SOLVED] Black screen at boot

#5 2022-06-03 06:44:06

Re: [SOLVED] Black screen at boot

#6 2022-06-04 11:17:05

Re: [SOLVED] Black screen at boot

#7 2022-06-08 14:10:58

Re: [SOLVED] Black screen at boot

#8 2022-06-08 14:20:17

Re: [SOLVED] Black screen at boot

#9 2022-06-09 07:14:26

Re: [SOLVED] Black screen at boot

#10 2022-06-09 07:25:10

Re: [SOLVED] Black screen at boot

#11 2022-06-09 09:22:19

Re: [SOLVED] Black screen at boot

#12 2022-06-09 12:01:16

Re: [SOLVED] Black screen at boot

Board footer