Skip to content
Snippets Groups Projects
New look: Off

*ERROR* IB test failed on gfx (-110) on P14s Gen2 AMD when waking up from s0ix

    • Closed Issue created by Björn Daase @bdaase

    Before submitting your bug report:

    Brief summary of the problem:

    *ERROR* IB test failed on gfx (-110) occasionally appears on my P14s Gen2 AMD when waking up from s0ix. The chance is ~1:1. When this happens, the display is flickering (interestingly, a GNOME screenshot does only capture a transparent rectangle). When this happens, I have to reboot the laptop to fully recover.

    This does never happen with s3 sleep.

    Hardware description:

    • CPU: 5850u
    • GPU: Cezanne

    System information:

    • Distro name and Version: Fedora 35
    • Kernel version: 5.15.6
    • DE Version: GNOME 41.1 (Wayland)

    How to reproduce the issue:

    1. echo s2idle | sudo tee -a /sys/power/mem_sleep
    2. Close the lid
    3. Open the lid

    Attached files:

    Screenshots/video files

    [ TODO: For rendering errors, attach screenshots of the problem and (if possible) of how it should look. For freezes, it may be useful to provide a screenshot of the affected game scene. Prefer screenshots over videos. ]

    Log files (for system lockups / game freezes / crashes)

    dmesg.txt

    Edited

    Designs

      Child items ...

      2. Activity

      • Björn Daase
        Björn Daase @bdaase ·
        Author

        Looks like there is actually quite more interesting stuff just before the failed suspend

        Dez 09 22:47:26.643666 p14s-gen2-amd kernel: [drm:dc_dmub_srv_wait_idle [amdgpu]] *ERROR* Error waiting for DMUB idle: status=3
Dez 09 22:47:27.680875 p14s-gen2-amd kernel: PM: suspend entry (s2idle)
Dez 09 22:47:28.088659 p14s-gen2-amd kernel: Filesystems sync: 0.408 seconds
Dez 09 22:51:46.287208 p14s-gen2-amd kernel: Freezing user space processes ... (elapsed 0.002 seconds) done.
Dez 09 22:51:46.287658 p14s-gen2-amd kernel: OOM killer disabled.
Dez 09 22:51:46.287936 p14s-gen2-amd kernel: Freezing remaining freezable tasks ... (elapsed 0.001 seconds) done.
Dez 09 22:51:46.288189 p14s-gen2-amd kernel: printk: Suspending console(s) (use no_console_suspend to debug)
Dez 09 22:51:46.288295 p14s-gen2-amd kernel: [drm] free PSP TMR buffer
Dez 09 22:51:46.288397 p14s-gen2-amd kernel: PM: suspend devices took 0.819 seconds
Dez 09 22:51:46.288464 p14s-gen2-amd kernel: ACPI: EC: interrupt blocked
Dez 09 22:51:46.288514 p14s-gen2-amd kernel: ACPI: EC: interrupt unblocked
Dez 09 22:51:46.288580 p14s-gen2-amd kernel: [drm] PCIE GART of 1024M enabled.
Dez 09 22:51:46.288642 p14s-gen2-amd kernel: [drm] PTB located at 0x000000F400900000
Dez 09 22:51:46.288715 p14s-gen2-amd kernel: [drm] PSP is resuming...
Dez 09 22:51:46.288795 p14s-gen2-amd kernel: pci 0000:00:00.2: can't derive routing for PCI INT A
Dez 09 22:51:46.289433 p14s-gen2-amd kernel: pci 0000:00:00.2: PCI INT A: no GSI
Dez 09 22:51:46.290139 p14s-gen2-amd kernel: usb usb1: root hub lost power or was reset
Dez 09 22:51:46.290781 p14s-gen2-amd kernel: usb usb2: root hub lost power or was reset
Dez 09 22:51:46.291388 p14s-gen2-amd kernel: xhci_hcd 0000:06:00.0: Zeroing 64bit base registers, expecting fault
Dez 09 22:51:46.291899 p14s-gen2-amd kernel: [drm] reserve 0x400000 from 0xf4ff800000 for PSP TMR
Dez 09 22:51:46.292041 p14s-gen2-amd kernel: nvme nvme0: Shutdown timeout set to 8 seconds
Dez 09 22:51:46.292562 p14s-gen2-amd kernel: nvme nvme0: 16/0/0 default/read/poll queues
Dez 09 22:51:46.292997 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: RAS: optional ras ta ucode is not available
Dez 09 22:51:46.293503 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: RAP: optional rap ta ucode is not available
Dez 09 22:51:46.294010 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: SECUREDISPLAY: securedisplay ta ucode is not available
Dez 09 22:51:46.294691 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: SMU is resuming...
Dez 09 22:51:46.295192 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: dpm has been disabled
Dez 09 22:51:46.295628 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: SMU is resumed successfully!
Dez 09 22:51:46.296235 p14s-gen2-amd kernel: [drm] kiq ring mec 2 pipe 1 q 0
Dez 09 22:51:46.296361 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: [gfxhub0] retry page fault (src_id:0 ring:0 vmid:5 pasid:32769, for process gnome-shell pid 2007 thread gnome-shel:cs0 pid 2052)
Dez 09 22:51:46.296746 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:   in page starting at address 0x0000000000000000 from IH client 0x1b (UTCL2)
Dez 09 22:51:46.297206 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: VM_L2_PROTECTION_FAULT_STATUS:0x005C0071
Dez 09 22:51:46.297717 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          Faulty UTCL2 client ID: CB (0x0)
Dez 09 22:51:46.298178 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          MORE_FAULTS: 0x1
Dez 09 22:51:46.298663 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          WALKER_ERROR: 0x0
Dez 09 22:51:46.299109 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          PERMISSION_FAULTS: 0x7
Dez 09 22:51:46.299628 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          MAPPING_ERROR: 0x0
Dez 09 22:51:46.300100 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          RW: 0x1
Dez 09 22:51:46.300750 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: [gfxhub0] retry page fault (src_id:0 ring:0 vmid:5 pasid:32769, for process gnome-shell pid 2007 thread gnome-shel:cs0 pid 2052)
Dez 09 22:51:46.301160 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:   in page starting at address 0x0000000000000000 from IH client 0x1b (UTCL2)
Dez 09 22:51:46.301794 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: VM_L2_PROTECTION_FAULT_STATUS:0x005C0071
Dez 09 22:51:46.302210 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          Faulty UTCL2 client ID: CB (0x0)
Dez 09 22:51:46.302647 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          MORE_FAULTS: 0x1
Dez 09 22:51:46.302980 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          WALKER_ERROR: 0x0
Dez 09 22:51:46.303291 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          PERMISSION_FAULTS: 0x7
Dez 09 22:51:46.303796 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          MAPPING_ERROR: 0x0
Dez 09 22:51:46.304186 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          RW: 0x1
Dez 09 22:51:46.304508 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: [gfxhub0] retry page fault (src_id:0 ring:0 vmid:5 pasid:32769, for process gnome-shell pid 2007 thread gnome-shel:cs0 pid 2052)
Dez 09 22:51:46.304775 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:   in page starting at address 0x0000000000000000 from IH client 0x1b (UTCL2)
Dez 09 22:51:46.305086 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: VM_L2_PROTECTION_FAULT_STATUS:0x005C0071
Dez 09 22:51:46.305425 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          Faulty UTCL2 client ID: CB (0x0)
Dez 09 22:51:46.305786 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          MORE_FAULTS: 0x1
Dez 09 22:51:46.306093 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          WALKER_ERROR: 0x0
Dez 09 22:51:46.306583 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          PERMISSION_FAULTS: 0x7
Dez 09 22:51:46.307010 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          MAPPING_ERROR: 0x0
Dez 09 22:51:46.307477 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          RW: 0x1
Dez 09 22:51:46.307822 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: [gfxhub0] retry page fault (src_id:0 ring:0 vmid:5 pasid:32769, for process gnome-shell pid 2007 thread gnome-shel:cs0 pid 2052)
Dez 09 22:51:46.308166 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:   in page starting at address 0x0000000000000000 from IH client 0x1b (UTCL2)
Dez 09 22:51:46.308608 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: VM_L2_PROTECTION_FAULT_STATUS:0x005C0071
Dez 09 22:51:46.309122 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          Faulty UTCL2 client ID: CB (0x0)
Dez 09 22:51:46.309467 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          MORE_FAULTS: 0x1
Dez 09 22:51:46.309716 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          WALKER_ERROR: 0x0
Dez 09 22:51:46.310182 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          PERMISSION_FAULTS: 0x7
Dez 09 22:51:46.310928 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          MAPPING_ERROR: 0x0
Dez 09 22:51:46.311356 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          RW: 0x1
Dez 09 22:51:46.311735 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: [gfxhub0] retry page fault (src_id:0 ring:0 vmid:5 pasid:32769, for process gnome-shell pid 2007 thread gnome-shel:cs0 pid 2052)
Dez 09 22:51:46.312068 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:   in page starting at address 0x0000000000000000 from IH client 0x1b (UTCL2)
Dez 09 22:51:46.312402 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: VM_L2_PROTECTION_FAULT_STATUS:0x005C0071
Dez 09 22:51:46.312693 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          Faulty UTCL2 client ID: CB (0x0)
Dez 09 22:51:46.313007 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          MORE_FAULTS: 0x1
Dez 09 22:51:46.313293 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          WALKER_ERROR: 0x0
Dez 09 22:51:46.313599 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          PERMISSION_FAULTS: 0x7
Dez 09 22:51:46.313874 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          MAPPING_ERROR: 0x0
Dez 09 22:51:46.314451 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          RW: 0x1
Dez 09 22:51:46.314848 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: [gfxhub0] retry page fault (src_id:0 ring:0 vmid:5 pasid:32769, for process gnome-shell pid 2007 thread gnome-shel:cs0 pid 2052)
Dez 09 22:51:46.315140 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:   in page starting at address 0x0000000000000000 from IH client 0x1b (UTCL2)
Dez 09 22:51:46.315429 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: VM_L2_PROTECTION_FAULT_STATUS:0x005C0071
Dez 09 22:51:46.315669 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          Faulty UTCL2 client ID: CB (0x0)
Dez 09 22:51:46.315920 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          MORE_FAULTS: 0x1
Dez 09 22:51:46.316189 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          WALKER_ERROR: 0x0
Dez 09 22:51:46.316488 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          PERMISSION_FAULTS: 0x7
Dez 09 22:51:46.316786 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          MAPPING_ERROR: 0x0
Dez 09 22:51:46.317074 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          RW: 0x1
Dez 09 22:51:46.317283 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: [gfxhub0] retry page fault (src_id:0 ring:0 vmid:5 pasid:32769, for process gnome-shell pid 2007 thread gnome-shel:cs0 pid 2052)
Dez 09 22:51:46.317834 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:   in page starting at address 0x0000000000000000 from IH client 0x1b (UTCL2)
Dez 09 22:51:46.318022 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: VM_L2_PROTECTION_FAULT_STATUS:0x005C0071
Dez 09 22:51:46.318204 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          Faulty UTCL2 client ID: CB (0x0)
Dez 09 22:51:46.318530 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          MORE_FAULTS: 0x1
Dez 09 22:51:46.318731 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          WALKER_ERROR: 0x0
Dez 09 22:51:46.318899 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          PERMISSION_FAULTS: 0x7
Dez 09 22:51:46.319064 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          MAPPING_ERROR: 0x0
Dez 09 22:51:46.319237 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          RW: 0x1
Dez 09 22:51:46.319480 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: [gfxhub0] retry page fault (src_id:0 ring:0 vmid:5 pasid:32769, for process gnome-shell pid 2007 thread gnome-shel:cs0 pid 2052)
Dez 09 22:51:46.319675 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:   in page starting at address 0x0000000000000000 from IH client 0x1b (UTCL2)
Dez 09 22:51:46.319874 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: VM_L2_PROTECTION_FAULT_STATUS:0x005C0071
Dez 09 22:51:46.320064 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          Faulty UTCL2 client ID: CB (0x0)
Dez 09 22:51:46.320267 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          MORE_FAULTS: 0x1
Dez 09 22:51:46.320472 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          WALKER_ERROR: 0x0
Dez 09 22:51:46.320652 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          PERMISSION_FAULTS: 0x7
Dez 09 22:51:46.320840 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          MAPPING_ERROR: 0x0
Dez 09 22:51:46.321022 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          RW: 0x1
Dez 09 22:51:46.321203 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: [gfxhub0] retry page fault (src_id:0 ring:0 vmid:5 pasid:32769, for process gnome-shell pid 2007 thread gnome-shel:cs0 pid 2052)
Dez 09 22:51:46.322000 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:   in page starting at address 0x0000000000000000 from IH client 0x1b (UTCL2)
Dez 09 22:51:46.322175 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: VM_L2_PROTECTION_FAULT_STATUS:0x005C0071
Dez 09 22:51:46.322431 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          Faulty UTCL2 client ID: CB (0x0)
Dez 09 22:51:46.322652 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          MORE_FAULTS: 0x1
Dez 09 22:51:46.322802 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          WALKER_ERROR: 0x0
Dez 09 22:51:46.322953 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          PERMISSION_FAULTS: 0x7
Dez 09 22:51:46.323098 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          MAPPING_ERROR: 0x0
Dez 09 22:51:46.323254 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          RW: 0x1
Dez 09 22:51:46.323549 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: [gfxhub0] retry page fault (src_id:0 ring:0 vmid:5 pasid:32769, for process gnome-shell pid 2007 thread gnome-shel:cs0 pid 2052)
Dez 09 22:51:46.323719 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:   in page starting at address 0x0000000000000000 from IH client 0x1b (UTCL2)
Dez 09 22:51:46.323889 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu: VM_L2_PROTECTION_FAULT_STATUS:0x005C0071
Dez 09 22:51:46.324063 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          Faulty UTCL2 client ID: CB (0x0)
Dez 09 22:51:46.324231 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          MORE_FAULTS: 0x1
Dez 09 22:51:46.324744 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          WALKER_ERROR: 0x0
Dez 09 22:51:46.324946 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          PERMISSION_FAULTS: 0x7
Dez 09 22:51:46.325121 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          MAPPING_ERROR: 0x0
Dez 09 22:51:46.325288 p14s-gen2-amd kernel: amdgpu 0000:07:00.0: amdgpu:          RW: 0x1
Dez 09 22:51:46.325564 p14s-gen2-amd kernel: [drm] DMUB hardware initialized: version=0x01010019
        Edited by Björn Daase
      • Björn Daase changed the description

        changed the description

        • Mario Limonciello
          Mario Limonciello @superm1 ·
          Developer

          Dez 09 22:47:26.643666 p14s-gen2-amd kernel: [drm:dc_dmub_srv_wait_idle [amdgpu]] *ERROR* Error waiting for DMUB idle: status=3

          That seems most interesting to me.

          DMUB hardware initialized: version=0x01010019

          Can you potentially try with the newer firmware at linux-firmware.git 0x101001c and see if it persists? This was uploaded as commit fbdf20e

        • Björn Daase
          Björn Daase @bdaase ·
          Author

          That seems most interesting to me.

          Alright. I actually have no clue what all the errors are about tbh.

          Can you potentially try with the newer firmware at linux-firmware.git 0x101001c and see if it persists? This was uploaded as commit fbdf20e

          This didn't hit Fedora repos yet, right? If there isn't an easy way to build an rpm out of it, I am afraid we have to wait till it is in the F350 repo...

        • Michel Dänzer
          Michel Dänzer @daenzer ·
          Developer

          This didn't hit Fedora repos yet, right?

          F35 has linux-firmware 20211027, but the referenced commit landed after that upstream.

          If there isn't an easy way to build an rpm out of it, I am afraid we have to wait till it is in the F350 repo...

          FWIW, you could manually put the files from upstream Git into /lib/firmware/amdgpu/ and re-generate /boot/initramfs-$(uname -r).img. If something goes wrong, in the worst case you can boot another kernel and reinstall the Fedora linux-firmware package, which should re-generate the initrd as well.

        • Björn Daase
          Björn Daase @bdaase ·
          Author

          FWIW, you could manually put the files from upstream Git into /lib/firmware/amdgpu/ and re-generate /boot/initramfs-$(uname -r).img. If something goes wrong, in the worst case you can boot another kernel and reinstall the Fedora linux-firmware package, which should re-generate the initrd as well.

          /lib/firmware/amdgpu/ seems to contain compressed .bin.xz files, while the upstream files are raw .bin files. Do I still have to transform the files somehow before placing them there?

        • Björn Daase
          Björn Daase @bdaase ·
          Author

          In the mean time, I got another s0ix resume fail, again showing the *ERROR* IB test failed on gfx (-110) warning but this time not including the *ERROR* Error waiting for DMUB idle: status=3 warning.

          Here is the full log: dmesg.txt

        • Michel Dänzer
          Michel Dänzer @daenzer ·
          Developer

          /lib/firmware/amdgpu/ seems to contain compressed .bin.xz files, while the upstream files are raw .bin files. Do I still have to transform the files somehow before placing them there?

          Good question. I guess safest would be to compress the new files as well with xz.

        • Björn Daase
          Björn Daase @bdaase ·
          Author

          FWIW, you could manually put the files from upstream Git into /lib/firmware/amdgpu/ and re-generate /boot/initramfs-$(uname -r).img. If something goes wrong, in the worst case you can boot another kernel and reinstall the Fedora linux-firmware package, which should re-generate the initrd as well.

          Turns out, that didn't work (the kernel didn't boot, showed the Lenovo logo forever). Went back to an older kernel then, which still booted. However, reinstalling linux-firmware and regenerating the initramfs still didn't give me a booting kernel. I then reinstalled kernel-*, which gave me a booting kernel again.

          TLDR; I wasn't able to test the new firmware.

        • Julian Sikorski
          Julian Sikorski @belegdol ·

          Do you know how to use fedora developer tools? If so, you might be able to generate an updated rpm and test like this. In brief, you should fedpkg clone linux-firmware and slip in the updated linux-firmware tarball created by make dist or equivalent.

        • Björn Daase
          Björn Daase @bdaase ·
          Author

          Thanks for the idea, but I unfortunately don't have any experience.

          I have cloned the existing linux-firmware packaging spec now and this builds just fine. But tbh, I have no idea how to force it to use for example the linux-firmware git directly or a local tar.gz.

          Edited by Björn Daase
        • Alex Deucher
          Alex Deucher @agd5f ·
          Owner

          If you are having trouble updating your initrd, you can just load the driver after the system has booted. put the new firmware in /lib/firmware/amdgpu, then append modprobe.blacklist=amdgpu 3 to the kernel command line in grub. Once the system has booted to a console, run the following:

          sudo modprobe amdgpu
sudo telinit 5
        • Julian Sikorski
          Julian Sikorski @belegdol ·

          @bdaase, if you can rebuild the RPM from the spec you are most of the way there:

          1. clone linux-firmware git: git clone git://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git
          2. make a tarball: git archive --output=linux-firmware-20211210.tar.xz --prefix=linux-firmware-20211210/ HEAD
          3. move the tarball to wherever you checked out the fedora linux-firmware RPM source
          4. edit the Fedora spec version and change it to 20211210.

          It is crude but should be enough to let you test the updated firmware. Once you are done, you can revert with dnf downgrade linux-firmware.

          Edited by Julian Sikorski
        • Björn Daase
          Björn Daase @bdaase ·
          Author

          Thanks for the help Julian, I actually got it working in the mean time (actually pretty much the same what you recommended) :)

          However, I unfortunately have to confirm that the problem persists after s0ix suspend :/

          Edited by Björn Daase
        • Björn Daase
          Björn Daase @bdaase ·
          Author

          FWIW, here is the log with new firmware: dmesg.txt

          However, for me it looks basically identical.

        • Björn Daase
          Björn Daase @bdaase ·
          Author

          Disabling secure boot as well as dpm also doesn't fix the issue.

        • Mario Limonciello
          Mario Limonciello @superm1 ·
          Developer

          FWIW, here is the log with new firmware: dmesg.txt However, for me it looks basically identical.

          It's missing that error waiting for DMUB idle message. On your log that appears it might have been fixed by the newer firmware.

        • Björn Daase
          Björn Daase @bdaase ·
          Author

          But it was also sometimes missing before the firmware update. It might be that it was fixed however.

        • Please register or sign in to reply
      • Mario Limonciello added s0ix label

        added s0ix label

      • Björn Daase changed title from ERROR IB test failed on gfx (-110) occasionally on P14s Gen2 AMD when waking up from s0ix to ERROR IB test failed on gfx (-110) on P14s Gen2 AMD when waking up from s0ix

        changed title from ERROR IB test failed on gfx (-110) occasionally on P14s Gen2 AMD when waking up from s0ix to ERROR IB test failed on gfx (-110) on P14s Gen2 AMD when waking up from s0ix

      • Björn Daase changed the description

        changed the description

        • Mario Limonciello
          Mario Limonciello @superm1 ·
          Developer

          This might help with the problems after resume: https://patchwork.freedesktop.org/patch/466321/. Can you try it?

          FYI I don't think it helps the underlying issue you see though "Error waiting for DMUB idle: status=3", it just might help you recover. For the actual underlying issue, this is actually a regression right? Do you think you could bisect back to what caused it?

        • Björn Daase
          Björn Daase @bdaase ·
          Author

          This might help with the problems after resume: https://patchwork.freedesktop.org/patch/466321/. Can you try it?

          Currently building a kernel with it.

          For the actual underlying issue, this is actually a regression right? Do you think you could bisect back to what caused it?

          I don't know whether it is a regression. Prior to https://www.spinics.net/lists/stable/msg517737.html, it wasn't resuming it all so I was using s3.

        • Mario Limonciello
          Mario Limonciello @superm1 ·
          Developer

          I don't know whether it is a regression. Prior to https://www.spinics.net/lists/stable/msg517737.html, it wasn't resuming it all so I was using s3.

          Ah right... that is needed specifically w/ newer mutter in GNOME 41 at least w/ Wayland. You might be able to try with X on older kernel.

        • Björn Daase
          Björn Daase @bdaase ·
          Author

          Hmm, unfortunately none of my custom build kernels (even a non-modified 5.15.7, using the fedora kernel.spec, but also the one with your suggested patch) boot past fb0: switching to amdgpudrmfb from efi vga. I have no idea what I am doing wrong here...

          Edited by Björn Daase
        • Julian Sikorski
          Julian Sikorski @belegdol ·

          May I ask how are you building your kernel? For official sources, you should use kernel-ark gitlab and run make dist-srpm and then rebuild the resulting srpm. This should give you pretty much the same kernel as the official fedora one. You will need to disable secure boot, or jump through very many hoops.

        • Björn Daase
          Björn Daase @bdaase ·
          Author

          Sure, I did

          fedpkg clone --anonymous kernel
cd kernel
Uncomment %define buildid .local
fedpkg local

          I disabled secure boot, because that was failing with some shim error earlier.

        • Michel Dänzer
          Michel Dänzer @daenzer ·
          Developer

          Hmm, unfortunately none of my custom build kernels (even a non-modified 5.15.7, using the fedora kernel.spec, but also the one with your suggested patch) boot past fb0: switching to amdgpudrmfb from efi vga.

          How long have you waited? E.g. if some needed firmware file is missing from the initrd, there's a 3 minute timeout IIRC.

          A general troubleshooting trick is to boot with modprobe.blacklist=amdgpu on the kernel command line, then after boot-up log in via SSH and run sudo modprobe amdgpu. That should allow getting more information from dmesg at least.

        • Björn Daase
          Björn Daase @bdaase ·
          Author

          How long have you waited? E.g. if some needed firmware file is missing from the initrd, there's a 3 minute timeout IIRC.

          ~8 hours, I gave it some time overnight. But it was still hanging there in the morning.

          A general troubleshooting trick is to boot with modprobe.blacklist=amdgpu on the kernel command line, then after boot-up log in via SSH and run sudo modprobe amdgpu. That should allow getting more information from dmesg at least.

          I might be able to try this that weekend. But I am still wondering what I am doing wrong, because it should be a quite well tested scenario...

        • Julian Sikorski
          Julian Sikorski @belegdol ·

          Do you still have modprobe.blacklist=amdgpu in your grub command line? In Fedora the boot line of newly installed kernels gets based on the most recent one of installed kernels. Maybe your self-built kernels ended up inheriting some changes you did. You can check this by checking the files in /boot/loader/entries.

        • Björn Daase
          Björn Daase @bdaase ·
          Author

          Doesn't look like it AFAICS

          title Fedora Linux (5.15.7-200.local.fc35.x86_64) 35 (Workstation Edition)
version 5.15.7-200.local.fc35.x86_64
linux /vmlinuz-5.15.7-200.local.fc35.x86_64
initrd /initramfs-5.15.7-200.local.fc35.x86_64.img
options root=UUID=9641d196-a942-4f26-b1ff-61c0473b9984 ro rootflags=subvol=@ rhgb quiet
grub_users $grub_users
grub_arg --unrestricted
grub_class fedora
        • Björn Daase
          Björn Daase @bdaase ·
          Author

          Thank you all, I finally got the patched custom kernel (including https://patchwork.freedesktop.org/patch/466321/) working!

          Unfortunately, the error persists. Here is the dmesg: dmesg.txt

        • Giulio Fidente
          Giulio Fidente @gfidente ·

          FWIW I used to see [amdgpu]] *ERROR* Error waiting for DMUB idle: status=3 in kernel logs with a 6600XT and kernel 5.15.10-arch1-1 ; GPU wouldn't come back from sleep (screen black) and GPU sensors appeared to be stuck

          I am now testing on same kernel 5.15.10-arch1-1 patched using https://patchwork.freedesktop.org/patch/466321/ and I am not seeing the issue from a couple of days now

        • Giulio Fidente
          Giulio Fidente @gfidente ·

          I have to amend my previous comment; for whatever reason it worked for a couple of days then it happened again despite the patch

          Edited by Giulio Fidente
        • Please register or sign in to reply
        • Björn Daase
          Björn Daase @bdaase ·
          Author

          Alright guys, it looks like I found the problem.

          Until yesterday, I thought the situation is as follows. On the P14s Gen2a there are three configurations available:

          • "Windows 10" mode selected in BIOS, "s2idle" selected in /sys/power/mem_sleep
          • "Linux" mode selected in BIOS, "s2idle" selected in /sys/power/mem_sleep
          • "Linux" mode selected in BIOS, "deep" selected in /sys/power/mem_sleep

          It seemed to me, that the first and second case (so "s2idle", no matter which BIOS mode selected), should be the same.

          It turns out, however, that it DOES make a difference whether it's "Windows 10" or "Linux" mode selected in BIOS. I was testing with "Linux" all the time and got the crashes seen in this thread. However, when selecting "Windows 10", the laptop suspends and resumes just fine.

          What do you guys think is happening here? Is this an AMD or Lenovo issue? It's really weird that this behaves differently even though you can't see a difference when using the OS.

          Edited by Björn Daase
        • Mario Limonciello
          Mario Limonciello @superm1 ·
          Developer

          This is a kernel bug in my opinion - it shouldn't be offering s2idle on x86 when the firmware is set to deep.

          At least on AMD platforms there is firmware configuration caused by changing that option in bios setup beyond the visible ASL difference.

        • Björn Daase
          Björn Daase @bdaase ·
          Author

          Alright, I opened https://bugzilla.kernel.org/show_bug.cgi?id=215387.

          I think we should then close this issue here.

        • Please register or sign in to reply
      • Björn Daase closed

        closed

      • Mario Limonciello reopened

        reopened

      • Mario Limonciello
        Mario Limonciello @superm1 ·
        Developer

        Re-opening this issue as the patch series to drop s2idle from AMD systems when configured to S3 was rejected. I've sent up https://patchwork.freedesktop.org/patch/469143/ as another idea to try out here.

      • Mario Limonciello
        Mario Limonciello @superm1 ·
        Developer

        I think that https://github.com/torvalds/linux/commit/6dc8265f9803ccb7e5da804e01601f0c14f270e0 actually will help fix the problem you had. I'll still send up a v2 to improve the warnings though.

      • Mario Limonciello closed

        closed

      • Alex Deucher mentioned in commit agd5f/linux@55335bcd

        mentioned in commit agd5f/linux@55335bcd

      • Alex Deucher mentioned in commit agd5f/linux@f588a1bb

        mentioned in commit agd5f/linux@f588a1bb

      • Alex Deucher mentioned in commit agd5f/linux@a6ed2035

        mentioned in commit agd5f/linux@a6ed2035

      Please register or sign in to reply