View previous topic :: View next topic |
Author |
Message |
akolechko n00b
Joined: 12 May 2022 Posts: 7
|
Posted: Thu May 12, 2022 5:55 pm Post subject: [SOLVED] Unable to turn on TearFree on amdgpu |
|
|
I have ryzen 7 5700g, amdgpu kernel driver, unsuccessfully trying to enable TearFree
Trying to do it in two ways: through xorg.conf.d and through xrandr
First approach - config
/etc/X11/xorg.conf.d/20-amdgpu.conf:
Code: | Section "Device"
Identifier "AMD Graphics"
Driver "amdgpu"
Option "TearFree" "true"
EndSection
|
Looks like config does nothing. Xorg actually reads it (it's visible from logs), but I still see tearing and xrandr states that TearFree is disabled. But I'm able to successfully enter desktop env, no craches
Second approach - xrandr
Execute command Code: | xrandr --output HDMI-A-0 --set TearFree on |
After execution of this command xorg server craches with backtrace:
Code: | [ 185.136] failed to create CRTC scanout FB
[ 185.136] failed to create CRTC scanout FB
[ 185.203] (EE)
[ 185.203] (EE) Backtrace:
[ 185.204] (EE) 0: /usr/bin/X (xorg_backtrace+0x5b) [0x562665df9b0b]
[ 185.204] (EE) 1: /usr/bin/X (0x562665cbb000+0x1428d5) [0x562665dfd8d5]
[ 185.204] (EE) 2: /lib64/libc.so.6 (0x7eff61f31000+0x3daa0) [0x7eff61f6eaa0]
[ 185.204] (EE) 3: /usr/lib64/xorg/modules/drivers/amdgpu_drv.so (0x7eff61b25000+0xfddc) [0x7eff61b34ddc]
[ 185.204] (EE) 4: /usr/bin/X (BlockHandler+0xa4) [0x562665d337c4]
[ 185.204] (EE) 5: /usr/bin/X (WaitForSomething+0x121) [0x562665df7121]
[ 185.204] (EE) 6: /usr/bin/X (0x562665cbb000+0x73cc3) [0x562665d2ecc3]
[ 185.204] (EE) 7: /usr/bin/X (0x562665cbb000+0x77e13) [0x562665d32e13]
[ 185.204] (EE) 8: /lib64/libc.so.6 (0x7eff61f31000+0x2931a) [0x7eff61f5a31a]
[ 185.204] (EE) 9: /lib64/libc.so.6 (__libc_start_main+0x7c) [0x7eff61f5a3cc]
[ 185.204] (EE) 10: /usr/bin/X (_start+0x21) [0x562665cf7051]
[ 185.204] (EE)
[ 185.204] (EE) Segmentation fault at address 0x10
[ 185.204] (EE)
Fatal server error:
[ 185.204] (EE) Caught signal 11 (Segmentation fault). Server aborting
[ 185.204] (EE)
[ 185.204] (EE)
|
Does anyone have thoughts on how to deal with such problem?
x.log https://pastebin.com/pWhZ4jRC
Last edited by akolechko on Fri May 13, 2022 5:00 pm; edited 1 time in total |
|
Back to top |
|
|
NeddySeagoon Administrator
Joined: 05 Jul 2003 Posts: 54234 Location: 56N 3W
|
Posted: Thu May 12, 2022 6:22 pm Post subject: |
|
|
akolechko,
Welcome to Gentoo
Code: | [ 172.470] (==) AMDGPU(0): DRI3 disabled
[ 172.470] (==) AMDGPU(0): Backing store enabled
[ 172.470] (WW) AMDGPU(0): Direct rendering disabled
[ 172.470] (II) AMDGPU(0): 2D and 3D acceleration disabled | tells that all your hardware acceleration is turned off.
Code: | [ 172.435] (--) PCI:*(8@0:0:0) 1002:1638: ... | That's a ("Advanced Micro Devices, Inc. [AMD/ATI]"), device: 1638 ("Cezanne") GPU.
The line in the log Code: | [ 172.438] (--) AMDGPU(0): Chipset: "Unknown AMD Radeon GPU" (ChipID = 0x1638) | suggests that your 5.15.32-gentoo-r1 kernel may be too old for that GPU or your x11-drivers/xf86-video-amdgpu may be too old or or both.
dmesg may have a few pointers tto.
Is your normal user in the video group?
Code: | $ groups
tty lp wheel uucp audio cdrom video kvm cdrw usb users plugdev vboxusers scanner users |
Not being in the video group will prevent hardware acceleration from working. _________________ Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
Back to top |
|
|
akolechko n00b
Joined: 12 May 2022 Posts: 7
|
Posted: Fri May 13, 2022 8:40 am Post subject: |
|
|
Hey Neddy, thanks for the response
Here's dmesg log: https://pastebin.com/kDtVc2uC
.config, just in case: https://pastebin.com/7VCAatDt
Regarding hardware acceleration, glxinfo says that direct rendering is enabled, I don't understand why info differs here (plus I followed https://wiki.gentoo.org/wiki/Xorg/Hardware_3D_acceleration_guide guide)
Regarding "Unknown AMD Radeon GPU", looks like I use the latest stable kernel and xf86-video-amdgpu. Plus I did xorg installation on ubuntu server on thic PC, which likely have older kernel, and I didn't experience such problem
Yes, my normal user is in the video group
One thing I want to add. I have two outputs - HDMI and Display port. When , trough xrandr, I enable TearFree on HDMI - it craches, but on Display port it doesn't, and xrandr says that it's enabled.
I don't have display port cable though so I can't check that with monitor, but maybe I'm lacking some HDMI driver or something?
Update: found display port cable, tried the same - segmentation fault. So it's something else |
|
Back to top |
|
|
NeddySeagoon Administrator
Joined: 05 Jul 2003 Posts: 54234 Location: 56N 3W
|
Posted: Fri May 13, 2022 4:14 pm Post subject: |
|
|
akolechko,
dmesg tells a few things but nothing video related as far as I can see.
Code: | [ 1.676899] Loading firmware: regulatory.db
[ 1.677237] platform regulatory.0: Direct firmware load for regulatory.db failed with error -2
[ 1.677572] cfg80211: failed to load regulatory.db | means that your wifi will be suboptimally configured.
That database lists permitted channels and wifi transmit powers for the various regions of the world.
Code: | [ 4.819624] drm_panel_orientation_quirks: exports duplicate symbol drm_get_panel_orientation_quirk (owned by kernel) | tells that your kernel and module don't match.
Its not possible to have the same symbol owned by both the kernel and a module at the same time.
Your running kernel was build at Code: | Thu May 12 15:52:27 EEST 2022 | If that fits with your memory, did you skip the Code: | make modules_install | step?
Code: | CONFIG_DRM_AMDGPU=m
# CONFIG_DRM_AMDGPU_SI is not set
# CONFIG_DRM_AMDGPU_CIK is not set
CONFIG_DRM_AMDGPU_USERPTR=y | is correct.
With DRM_AMDGPU=m, you won't have a console until the module loads, that's OK.
To get an earlier console, turn on
Code: | # CONFIG_FB_VESA is not set
# CONFIG_FB_EFI is not set
# CONFIG_FB_SIMPLE is not set | The kernel will pick one, then hand over to amdgpu when its ready. _________________ Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
Back to top |
|
|
akolechko n00b
Joined: 12 May 2022 Posts: 7
|
Posted: Fri May 13, 2022 4:59 pm Post subject: |
|
|
Hey Neddy, I just fixed the issue
Upon another xrandr trial after crash I noticed in report that mesa failed to load radeonsi driver. So I added radeonsi to VIDEO_CARDS, emerged, rebooted, and tearfree now applied correctly.
Ps: I just noticed that amdgpu wiki page mentions multiple times that both amdgpu and radeonsi are necessary, especially for TearFree.... Well, at least now I know what happens if it's not specified
Thanks for help Neddy! |
|
Back to top |
|
|
NeddySeagoon Administrator
Joined: 05 Jul 2003 Posts: 54234 Location: 56N 3W
|
Posted: Fri May 13, 2022 7:43 pm Post subject: |
|
|
akolechko,
Once upon a time, life was simple. VIDEO_CARDS= listed only the names for the Xorg video drivers you wanted to build.
Indeed, if it was empty, none were installed.
There was a one to one correspondence between entries and drivers.
Then mesa began to use VIDEO_CARDS= to know what acceleration to provide. That broke the one to one correspondence.
amdgpu is for Xorg
radeonsi is for mesa. When that was missing, you didn't have any acceleration, as we saw in your logs. _________________ Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
Back to top |
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
|