[radeon] Pictures including wallpaper not showing in kernel 5.14.21 and later (fixed with intel_iommu=off)

[Mícheál Na bPíob]

I can't see pictures in emails or on some pictures on websites. Sometimes it just shows the bottom 20mm of pictures. The wallpaper is just a blank white screen.

ProblemType: Bug
DistroRelease: Ubuntu 22.04
Package: xorg 1:7.7+23ubuntu2
ProcVersionSignature: Ubuntu 5.15.0-27.28-generic 5.15.30
Uname: Linux 5.15.0-27-generic x86_64
ApportVersion: 2.20.11-0ubuntu82
Architecture: amd64
BootLog: Error: [Errno 13] Permission denied: '/var/log/boot.log'
CasperMD5CheckResult: unknown
CompositorRunning: None
CurrentDesktop: ubuntu:GNOME
Date: Mon May 2 13:12:31 2022
DistUpgraded: 2022-04-24 19:05:38,365 DEBUG Running PostInstallScript: '/usr/lib/ubuntu-advantage/upgrade_lts_contract.py'
DistroCodename: jammy
DistroVariant: ubuntu
DkmsStatus:
 virtualbox/6.1.32, 5.15.0-25-generic, x86_64: installed
 virtualbox/6.1.32, 5.15.0-27-generic, x86_64: installed
ExtraDebuggingInterest: Yes
GraphicsCard:
 Advanced Micro Devices, Inc. [AMD/ATI] Venus XT [Radeon HD 8870M / R9 M270X/M370X] [1002:6821] (rev 83) (prog-if 00 [VGA controller])
   Subsystem: Apple Inc. Radeon R9 M370X Mac Edition [106b:0149]
InstallationDate: Installed on 2021-04-26 (371 days ago)
InstallationMedia: Ubuntu 20.10 "Groovy Gorilla" - Release amd64 (20201022)
Lsusb:
 Bus 002 Device 003: ID 05ac:8406 Apple, Inc. Internal Memory Card Reader
 Bus 002 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
 Bus 001 Device 002: ID 05ac:8290 Apple, Inc. Bluetooth Host Controller
 Bus 001 Device 003: ID 05ac:0274 Apple, Inc. Apple Internal Keyboard / Trackpad
 Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
MachineType: Apple Inc. MacBookPro11,5
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-5.15.0-27-generic root=UUID=2234e6d3-3ebc-44e6-b0a5-bef132fd1f72 ro quiet splash vt.handoff=7
SourcePackage: xorg
Symptom: display
UpgradeStatus: Upgraded to jammy on 2022-04-24 (7 days ago)
dmi.bios.date: 06/11/2020
dmi.bios.release: 0.1
dmi.bios.vendor: Apple Inc.
dmi.bios.version: 199.0.0.0.0
dmi.board.name: Mac-06F11F11946D27C5
dmi.board.vendor: Apple Inc.
dmi.board.version: MacBookPro11,5
dmi.chassis.type: 9
dmi.chassis.vendor: Apple Inc.
dmi.chassis.version: Mac-06F11F11946D27C5
dmi.modalias: dmi:bvnAppleInc.:bvr199.0.0.0.0:bd06/11/2020:br0.1:svnAppleInc.:pnMacBookPro11,5:pvr1.0:rvnAppleInc.:rnMac-06F11F11946D27C5:rvrMacBookPro11,5:cvnAppleInc.:ct9:cvrMac-06F11F11946D27C5:sku:
dmi.product.family: MacBook Pro
dmi.product.name: MacBookPro11,5
dmi.product.version: 1.0
dmi.sys.vendor: Apple Inc.
version.compiz: compiz N/A
version.libdrm2: libdrm2 2.4.110-1ubuntu1
version.libgl1-mesa-dri: libgl1-mesa-dri 22.0.1-1ubuntu2
version.libgl1-mesa-glx: libgl1-mesa-glx 22.0.1-1ubuntu2
version.xserver-xorg-core: xserver-xorg-core 2:21.1.3-2ubuntu2
version.xserver-xorg-input-evdev: xserver-xorg-input-evdev N/A
version.xserver-xorg-video-ati: xserver-xorg-video-ati 1:19.1.0-2build3
version.xserver-xorg-video-intel: xserver-xorg-video-intel 2:2.99.917+git20210115-1
version.xserver-xorg-video-nouveau: xserver-xorg-video-nouveau 1:1.0.17-2build1

[Daniel van Vugt (vanvugt)]

Thanks for the bug report.

Please try selecting 'Ubuntu' on the login screen instead of 'Ubuntu on Xorg'. That way you can try a Wayland session instead.

[Daniel van Vugt (vanvugt)]

Maybe related to bug 1953507 or bug 1969815

[Mícheál Na bPíob (mi0hoz)]

Tried it, no difference. Also if I log in remotely with anydesk I can't
see the desktop.

regards

Mick

On 03/05/2022 03:55, Daniel van Vugt wrote:
> Thanks for the bug report.
>
> Please try selecting 'Ubuntu' on the login screen instead of 'Ubuntu on
> Xorg'. That way you can try a Wayland session instead.
>
> ** Summary changed:
>
> - Pictures including wallpaper not showing
> + [radeon] Pictures including wallpaper not showing
>
> ** Tags added: radeon
>
> ** Package changed: xorg (Ubuntu) => mesa (Ubuntu)
>
> ** Also affects: mutter (Ubuntu)
> Importance: Undecided
> Status: New
>
> ** Package changed: mutter (Ubuntu) => xserver-xorg-video-ati (Ubuntu)
>
> ** Changed in: xserver-xorg-video-ati (Ubuntu)
> Status: New => Incomplete
>
> ** Changed in: mesa (Ubuntu)
> Status: New => Incomplete
>

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

This change was made by a bot.

[Daniel van Vugt (vanvugt)]

Can you provide a photo of the problem?

[Mícheál Na bPíob (mi0hoz)]

On 03/05/2022 04:41, Daniel van Vugt wrote:
> Can you provide a photo of the problem?
>

[Daniel van Vugt (vanvugt)]

Presumably Ubuntu worked before the upgrade. Please try booting some older/newer kernels to see which if any are free from the bug:

https://kernel.ubuntu.com/~kernel-ppa/mainline/?C=M;O=D

[Mícheál Na bPíob (mi0hoz)]

Ubuntu has been working fine for years.

I'm now using kernel 5.17.5-051705-generic and it works, but only
sometimes. If I reboot the problem will return. It's working properly
now, is there a command or log file I can generate to see if anything is
different?

On 03/05/2022 05:07, Daniel van Vugt wrote:
> Presumably Ubuntu worked before the upgrade. Please try booting some
> older/newer kernels to see which if any are free from the bug:
>
> https://kernel.ubuntu.com/~kernel-ppa/mainline/?C=M;O=D
>

[Daniel van Vugt (vanvugt)]

Great to hear.

Additional logs certainly won't help if it's working now. But if you have the patience then finding out what version number exactly fixed it would help. Or what kernel version was the last to have the bug, keeping in mind the broken kernel was (roughly) upstream version 5.15.30

[Mícheál Na bPíob (mi0hoz)]

uname -r

5.17.5-051705-generic

On 03/05/2022 06:30, Daniel van Vugt wrote:
> Great to hear.
>
> Additional logs certainly won't help if it's working now. But if you
> have the patience then finding out what version number exactly fixed it
> would help. Or what kernel version was the last to have the bug, keeping
> in mind the broken kernel was (roughly) upstream version 5.15.30
>
> ** No longer affects: mutter (Ubuntu)
>

[Mícheál Na bPíob (mi0hoz)]

As expected a reboot and the bug is back :(

On 03/05/2022 06:30, Daniel van Vugt wrote:
> Great to hear.
>
> Additional logs certainly won't help if it's working now. But if you
> have the patience then finding out what version number exactly fixed it
> would help. Or what kernel version was the last to have the bug, keeping
> in mind the broken kernel was (roughly) upstream version 5.15.30
>
> ** No longer affects: mutter (Ubuntu)
>

[Mícheál Na bPíob (mi0hoz)]

If I boot into recovery mode then click normal boot the desktop appears,
every time. Kernel 5.17.5-051705-generic

If I just do a normal boot the desktop appears maybe 1 in out of every
20 reboots, same kernel.

Regards,

Mick

On 03/05/2022 06:30, Daniel van Vugt wrote:
> Great to hear.
>
> Additional logs certainly won't help if it's working now. But if you
> have the patience then finding out what version number exactly fixed it
> would help. Or what kernel version was the last to have the bug, keeping
> in mind the broken kernel was (roughly) upstream version 5.15.30
>
> ** No longer affects: mutter (Ubuntu)
>

[Daniel van Vugt (vanvugt)]

Not solved then. Recovery mode is a workaround but not a fix.

[Mícheál Na bPíob (mi0hoz)]

Yes but I think it points to the graphics driver maybe

On 03/05/2022 07:26, Daniel van Vugt wrote:
> Not solved then. Recovery mode is a workaround but not a fix.
>

[Daniel van Vugt (vanvugt)]

Yes it seems to be specific to 'radeon' which is both a kernel driver and a Mesa driver.

[Mícheál Na bPíob (mi0hoz)]

No problem with kernel: 5.13.0-051300-generic

On 03/05/2022 08:21, Daniel van Vugt wrote:
> Yes it seems to be specific to 'radeon' which is both a kernel driver
> and a Mesa driver.
>

[Daniel van Vugt (vanvugt)]

Great. Next would be to try 5.14/ and 5.13.9/ (because it is halfway between 5.13 and 5.14)

[Mícheál Na bPíob (mi0hoz)]

Works 5.14/ but not 5.15.0-25

On 03/05/2022 09:09, Daniel van Vugt wrote:
> Great. Next would be to try 5.14/ and 5.13.9/ (because it is halfway
> between 5.13 and 5.14)
>
> ** No longer affects: mesa (Ubuntu)
>
> ** Tags added: regression-release
>

[Daniel van Vugt (vanvugt)]

Next would be to try 5.14.10, and then go up or down depending on the result.

[Mícheál Na bPíob (mi0hoz)]

Ok, I've installed many kernels and can confirm 5.14.20-051420-generic
works and 5.14.21-051421 produces the bug. I hope this helps.

Regards,

Mick

On 03/05/2022 09:43, Daniel van Vugt wrote:
> Next would be to try 5.14.10, and then go up or down depending on the
> result.
>

[Daniel van Vugt (vanvugt)]

Nice work, thanks!

The only remaining thing to do is to try the latest drm-tip kernel:

  https://kernel.ubuntu.com/~kernel-ppa/mainline/drm-tip/current/amd64/

That will tell us if the bug needs to be reported to the kernel developers, or if they already have a fix on the way.

[Daniel van Vugt (vanvugt)]

Another DRM regression in v5.14.21: bug 1965882

[Timo Aaltonen (tjaalton)]

git log1 v5.14.20..v5.14.21
545728d9e085937 (tag: v5.14.21, stable/linux-5.14.y) Linux 5.14.21
999805aaf925594 Revert "ACPI: scan: Release PM resources blocked by unused objects"
0750f769b95841b thermal: Fix NULL pointer dereferences in of_thermal_ functions
50812034d105b81 perf/core: Avoid put_page() when GUP fails
ae2afdf82a00049 PCI: Add MSI masking quirk for Nvidia ION AHCI
18c55aed17b2d9b PCI/MSI: Deal with devices lying about their MSI mask capability
f64af173573104d PCI/MSI: Destroy sysfs before freeing entries
f41389affe96f14 parisc/entry: fix trace test in syscall exit path
56249b445477c1a Bluetooth: btusb: Add support for TP-Link UB500 Adapter
dbaa485bc5a1273 bootconfig: init: Fix memblock leak in xbc_make_cmdline()
6bb162d35d74ea8 loop: Use blk_validate_block_size() to validate block size
e2d865b3109c5da block: Add a helper to validate the block size
5b9f31a3be76f14 fortify: Explicitly disable Clang support
87304542f264d99 KVM: Fix steal time asm constraints
02eafa8f1558417 Revert "drm: fb_helper: fix CONFIG_FB dependency"
bd462930f7e5342 Revert "drm: fb_helper: improve CONFIG_FB dependency"

nothing radeon'y in there that I can see

[Daniel van Vugt (vanvugt)]

I know, same for bug 1965882. But still a regression in 5.14.21.

[Daniel van Vugt (vanvugt)]

Is there an easy way to diff the configs used between:

https://kernel.ubuntu.com/~kernel-ppa/mainline/v5.14.20/
https://kernel.ubuntu.com/~kernel-ppa/mainline/v5.14.21/

?

[Mícheál Na bPíob (mi0hoz)]

I can confim the bug is in the latest drm-tip kernel

Regards,

Mick

On 04/05/2022 06:08, Daniel van Vugt wrote:
> Nice work, thanks!
>
> The only remaining thing to do is to try the latest drm-tip kernel:
>
> https://kernel.ubuntu.com/~kernel-ppa/mainline/drm-tip/current/amd64/
>
> That will tell us if the bug needs to be reported to the kernel
> developers, or if they already have a fix on the way.
>
> ** Summary changed:
>
> - [radeon] Pictures including wallpaper not showing
> + [radeon] Pictures including wallpaper not showing in kernel 5.14.21 and later (but 5.14.20 and earlier works)
>
> ** Changed in: linux (Ubuntu)
> Importance: Undecided => High
>
> ** Changed in: linux (Ubuntu)
> Status: Confirmed => Triaged
>

[Timo Aaltonen (tjaalton)]

extract the kernel, diff ../boot/config-*

[Timo Aaltonen (tjaalton)]

--- config-5.14.20-051420-generic 2021-11-18 16:31:49.000000000 +0200
+++ config-5.14.21-051421-generic 2021-11-21 15:31:59.000000000 +0200
@@ -1,8 +1,8 @@
 #
 # Automatically generated file; DO NOT EDIT.
-# Linux/x86 5.14.20-051420-generic Kernel Configuration
+# Linux/x86 5.14.21-051421-generic Kernel Configuration
 #
-CONFIG_CC_VERSION_TEXT="gcc (Ubuntu 11.2.0-10ubuntu1) 11.2.0"
+CONFIG_CC_VERSION_TEXT="gcc (Ubuntu 11.2.0-12ubuntu1) 11.2.0"
 CONFIG_CC_IS_GCC=y
 CONFIG_GCC_VERSION=110200
 CONFIG_CLANG_VERSION=0
@@ -9036,9 +9036,9 @@
 CONFIG_DMAR_TABLE=y
 CONFIG_INTEL_IOMMU=y
 CONFIG_INTEL_IOMMU_SVM=y
-# CONFIG_INTEL_IOMMU_DEFAULT_ON is not set
+CONFIG_INTEL_IOMMU_DEFAULT_ON=y
 CONFIG_INTEL_IOMMU_FLOPPY_WA=y
-# CONFIG_INTEL_IOMMU_SCALABLE_MODE_DEFAULT_ON is not set
+CONFIG_INTEL_IOMMU_SCALABLE_MODE_DEFAULT_ON=y
 CONFIG_IRQ_REMAP=y
 CONFIG_HYPERV_IOMMU=y
 CONFIG_VIRTIO_IOMMU=y

[Timo Aaltonen (tjaalton)]

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1951440

[Daniel van Vugt (vanvugt)]

Mícheál, please try adding kernel parameter:

  intel_iommu=off

to /etc/default/grub (in GRUB_CMDLINE_LINUX_DEFAULT) and then run:

  sudo update-grub

and reboot.

[Daniel van Vugt (vanvugt)]

That should read:

  sudo update-grub

[Mícheál Na bPíob (mi0hoz)]

Using kernel 5.18.0-051800rc5drmtip20220504-generic

That works ok.

On 06/05/2022 09:07, Daniel van Vugt wrote:
> Mícheál, please try adding kernel parameter:
>
> intel_iommu=off
>
> to /etc/default/grub (in GRUB_CMDLINE_LINUX_DEFAULT) and then run:
>
> sudo upgrade-grub
>
> and reboot.
>

[Daniel van Vugt (vanvugt)]

Please try using the original Ubuntu kernel again (5.15.0-27), verify the bug occurs, and then try comment #31. We need to verify what the exact fix was.

[Mícheál Na bPíob (mi0hoz)]

I can confirm the bug exists on the Ubuntu kernel 5.15.0-27 and by
adding intel_iommu=off to /etc/default/grub solves the issue.

Regards, Mick

On 09/05/2022 02:52, Daniel van Vugt wrote:
> Please try using the original Ubuntu kernel again (5.15.0-27), verify
> the bug occurs, and then try comment #31. We need to verify what the
> exact fix was.
>

[koba (kobako)]

There's dmar error, i thought Device[01:00.0] may access a out-of-bound address.
[ 843.730480] DMAR: DRHD: handling fault status reg 3
[ 843.730488] DMAR: [DMA Read NO_PASID] Request device [01:00.0] fault addr 0x6e000000 [fault reason 0x0c] non-zero reserved fields in PTE
[ 843.730494] DMAR: DRHD: handling fault status reg 3
[ 843.730496] DMAR: [DMA Read NO_PASID] Request device [01:00.0] fault addr 0x6e016000 [fault reason 0x0c] non-zero reserved fields in PTE
[ 843.730499] DMAR: DRHD: handling fault status reg 3
[ 843.730501] DMAR: [DMA Read NO_PASID] Request device [01:00.0] fault addr 0x6e01e000 [fault reason 0x0c] non-zero reserved fields in PTE
[ 843.730505] DMAR: DRHD: handling fault status reg 3

Could someone dump
sudo cat /sys/bus/pci/devices/0000\:01\:00.0/iommu/intel-iommu/*
sudo cat /sys/bus/pci/devices/0000\:01\:00.0/iommu_group/reserved_regions
sudo cat /sys/bus/pci/devices/0000\:01\:00.0/iommu_group/type

[koba (kobako)]

could someone help to verify this kernel on target w/o intel_iommu=off?thanks
https://drive.google.com/drive/folders/1c1p8y265T9urgOwKoJLXaiGUrlKpJlKO?usp=sharing

[Daniel van Vugt (vanvugt)]

Fix released in jammy https://launchpad.net/ubuntu/+source/linux/5.15.0-40.43 according to bug 1971699.

[Daniel van Vugt (vanvugt)]

Is this an issue in kinetic still?

[Heinrich Schuchardt (xypron)]

intel_iommu=off may result in a security risk as PCIe devices can access all of the RAM. intel_iommu=igfx_off should be enough to resolve issues with GPUs.

[koba (kobako)]

@Mario, would you please have some comment in this? thanks

[Mario Limonciello (superm1)]

> intel_iommu=igfx_off should be enough to resolve issues with GPUs.

Maybe this design dGPU should have IOMMU off.

> @Mario, would you please have some comment in this? thanks

Maybe a new parameter is needed for Intel IOMMU driver to allow filtering dGPU from IOMMU too?

But this type of error seems to happen on a few kinds of other devices too though.
1) TBT controller: https://bugzilla.kernel.org/show_bug.cgi?id=209321
2) eSATA: https://bugzilla.kernel.org/show_bug.cgi?id=98451
3) audio: https://bugzilla.kernel.org/show_bug.cgi?id=215919

[Mario Limonciello (superm1)]

Looking at the code though, I notice that igfx=off actually filters all GFX. Maybe it's just badly named:

#define IS_GFX_DEVICE(pdev) ((pdev->class >> 16) == PCI_BASE_CLASS_DISPLAY)

So yeah; have a try with "intel_iommu=igfx_off" on kernel command line.

[Kai-Heng Feng (kaihengfeng)]

Please test the kernel, which adds the affected GFX to IOMMU quirk list:
https://people.canonical.com/~khfeng/gfx-no-dmar/

[Utkarsh Gupta (utkarsh)]

Ubuntu 22.10 (Kinetic Kudu) has reached end of life, so this bug will not be fixed for that specific release.