xf86-video-amdgpu issueshttps://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues2021-12-02T15:04:30Zhttps://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/1Segfault with amdgpu dri3 and 2d acceleration disabled2021-12-02T15:04:30ZBugzilla Migration UserSegfault with amdgpu dri3 and 2d acceleration disabled## Submitted by csa..@..il.com
Assigned to **xf86-video-ati maintainers**
**[Link to original bug (#94214)](https://bugs.freedesktop.org/show_bug.cgi?id=94214)**
## Description
Running X with option AccelMethod "none" (possibly al...## Submitted by csa..@..il.com
Assigned to **xf86-video-ati maintainers**
**[Link to original bug (#94214)](https://bugs.freedesktop.org/show_bug.cgi?id=94214)**
## Description
Running X with option AccelMethod "none" (possibly also option noaccel) causes segfault if dri3 is enabled:<br>
<br>
```<br>
Program received signal SIGSEGV, Segmentation fault.<br>
0x0000000000000000 in ?? ()<br>
(gdb) bt<br>
#0 0x0000000000000000 in ?? ()<br>
#1 0x00007fc9a1c87db8 in amdgpu_dri3_pixmap_from_fd (screen=0x561f9d7c1e40, fd=26, width=16, height=<optimized out>,<br>
stride=128, depth=<optimized out>, bpp=32 ' ') at ../../src/amdgpu_dri3.c:149<br>
#2 0x0000561f9c32d289 in dri3_pixmap_from_fd (ppixmap=ppixmap@entry=0x7ffc1c145628, screen=<optimized out>, fd=fd@entry=26,<br>
width=<optimized out>, height=<optimized out>, stride=128, depth=24 '\030', bpp=32 ' ') at ../../dri3/dri3_screen.c:73<br>
#3 0x0000561f9c32cdfa in proc_dri3_pixmap_from_buffer (client=0x561f9dc17e50) at ../../dri3/dri3_request.c:162<br>
#4 0x0000561f9c1fe12e in Dispatch () at ../../dix/dispatch.c:430<br>
#5 0x0000561f9c201ecb in dix_main (argc=5, argv=0x7ffc1c145808, envp=<optimized out>) at ../../dix/main.c:300<br>
#6 0x00007fc9a50baead in __libc_start_main (main=<optimized out>, argc=<optimized out>, ubp_av=<optimized out>,<br>
init=<optimized out>, fini=<optimized out>, rtld_fini=<optimized out>, stack_end=0x7ffc1c1457f8) at libc-start.c:244<br>
#7 0x0000561f9c1ec6b1 in _start ()<br>
```<br>
<br>
The offending line seems to be<br>
<br>
if (screen->SetSharedPixmapBacking(pixmap, (void*)(intptr_t)fd))<br>
return pixmap;<br>
<br>
Version info: Xorg 1.18, amdgpu git a3eac85dhttps://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/2AccelMethod "none" breaks DRI2 and VDPAU on amdgpu2021-12-02T15:04:30ZBugzilla Migration UserAccelMethod "none" breaks DRI2 and VDPAU on amdgpu## Submitted by csa..@..il.com
Assigned to **xf86-video-ati maintainers**
**[Link to original bug (#94220)](https://bugs.freedesktop.org/show_bug.cgi?id=94220)**
## Description
Created attachment 121847<br>
Xorg log<br>
<br>
As fa...## Submitted by csa..@..il.com
Assigned to **xf86-video-ati maintainers**
**[Link to original bug (#94220)](https://bugs.freedesktop.org/show_bug.cgi?id=94220)**
## Description
Created attachment 121847<br>
Xorg log<br>
<br>
As far as I understand accelmethod none should just disable 2D (glamor). The man page says:<br>
<br>
"Setting this option to none disables use of the glamor acceleration architecture. In that case, all 2D rendering is done by the CPU, but 3D and video hardware acceleration can still work."<br>
<br>
Also Xorg log seems to confirm this:<br>
<br>
[ 55521.565] (II) AMDGPU(0): [DRI2] Setup complete<br>
[ 55521.565] (II) AMDGPU(0): [DRI2] DRI driver: radeonsi<br>
[ 55521.565] (II) AMDGPU(0): [DRI2] VDPAU driver: radeonsi<br>
[ 55521.565] (II) AMDGPU(0): Front buffer pitch: 15360 bytes<br>
[ 55521.566] (==) AMDGPU(0): DRI3 disabled<br>
[ 55521.566] (==) AMDGPU(0): Backing store enabled<br>
[ 55521.566] (II) AMDGPU(0): Direct rendering enabled<br>
[ 55521.566] (II) AMDGPU(0): 2D acceleration disabled<br>
<br>
Nevertheless running glxgears produces the following error:<br>
<br>
X Error of failed request: BadDrawable (invalid Pixmap or Window parameter)<br>
Major opcode of failed request: 152 (DRI2)<br>
Minor opcode of failed request: 8 (DRI2SwapBuffers )<br>
Resource id in failed request: 0xe00002<br>
Serial number of failed request: 48<br>
Current serial number in output stream: 49<br>
<br>
With this in the Xorg log:<br>
<br>
[ 55718.984] (EE) AMDGPU(0): [DRI2] DRI2SwapBuffers: drawable has no back or front?<br>
<br>
Running mplayer with -vo vdpau shows errors:<br>
<br>
[vdpau] Error when calling vdp_presentation_queue_display: An invalid handle value was provided. Either the handle does not exist at all, or refers to an object of an incorrect type.<br>
X11 error: BadAlloc (insufficient resources for operation)<br>
<br>
Version info: Xorg 1.18, amdgpu git a3eac85d<br>
<br>
**Attachment 121847**, "Xorg log": <br>
[Xorg.0.log.accelnone](/uploads/9a36796fea2d2294131c7e4c4be1bd8e/Xorg.0.log.accelnone)<br>https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/3FreeSync / VRR stops working after window switching or minimizing2021-12-02T15:04:30ZBugzilla Migration UserFreeSync / VRR stops working after window switching or minimizing## Submitted by clapbr `@clapbr`
Assigned to **xf86-video-ati maintainers**
**[Link to original bug (#110473)](https://bugs.freedesktop.org/show_bug.cgi?id=110473)**
## Description
As title says, FreeSync stops working for me afte...## Submitted by clapbr `@clapbr`
Assigned to **xf86-video-ati maintainers**
**[Link to original bug (#110473)](https://bugs.freedesktop.org/show_bug.cgi?id=110473)**
## Description
As title says, FreeSync stops working for me after I alt+tab or minimize the fullscreened game window.<br>
<br>
I mentioned the issue in RADV VRR merge request on Mesa gitlab -https://gitlab.freedesktop.org/mesa/mesa/merge_requests/672<br>
- because I can only reproduce it with Vulkan games, but M. Danzer mentioned other people had it with radeonsi/OpenGL so the problem likely is in xf86-video-amdgpu.<br>
<br>
All drivers (kernel, mesa, xf86) are latest from git.https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/8Support building with `-z now`2022-05-30T19:11:33ZYurii Kolesnykovroot@yurikoles.comSupport building with `-z now`I had built [xserver](https://aur.archlinux.org/pkgbase/xorg-server-git/) and [amdgpu](https://aur.archlinux.org/packages/xf86-video-amdgpu-git/) from master via AUR and see this in
[Xorg.0.log](https://git.io/Jvn6n):
```
Failed to loa...I had built [xserver](https://aur.archlinux.org/pkgbase/xorg-server-git/) and [amdgpu](https://aur.archlinux.org/packages/xf86-video-amdgpu-git/) from master via AUR and see this in
[Xorg.0.log](https://git.io/Jvn6n):
```
Failed to load /usr/lib/xorg/modules/drivers/amdgpu_drv.so: /usr/lib/xorg/modules/drivers/amdgpu_drv.so: undefined symbol: glamor_egl_init_textured_pixmap
```https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/11Prime Sync not working with amdgpu2022-01-20T08:50:09ZAndrew Nowa Ammerlaanandrewammerlaan@gentoo.orgPrime Sync not working with amdgpuHi,
Prime Synchronization seems to be non-functional on my RX 590, and the tearing is driving me insane:
```
andrew@andrew-gentoo-pc ~ % xrandr --properties | grep PRIME
PRIME Synchronization: 1
PRIME Synchronization: 1
PRIME Synchroni...Hi,
Prime Synchronization seems to be non-functional on my RX 590, and the tearing is driving me insane:
```
andrew@andrew-gentoo-pc ~ % xrandr --properties | grep PRIME
PRIME Synchronization: 1
PRIME Synchronization: 1
PRIME Synchronization: 0
PRIME Synchronization: 0
PRIME Synchronization: 1
```
The 0's are from the connected monitors, the 1's are disconnected outputs.
Using `xrandr --output DP-1-5 --set "PRIME Synchronization" 1` does not change anything.
I also tried toggling the amdgpu.dc boot paramter to 0 and 1, without effect.
What does work, is to force usage of the modesetting driver:
```
Section "Device"
Identifier "modesetting"
Driver "modesetting"
EndSection
```
This results in working Prime Sync and a tear-free display. However the modesetting driver gives me problems with one of the two monitors connected to the iGPU. I get massive screen corruption on that monitor, while the other one is nice and tear-free. (This might be a different unrelated issue though)
```
andrew@andrew-gentoo-pc ~ % glxinfo | grep "OpenGL renderer"
OpenGL renderer string: Radeon RX 590 Series (POLARIS10, DRM 3.36.0, 5.6.7-gentoo, LLVM 9.0.1)
```
This whole setup, same monitors, same configuration, same everything, was working just fine with working Prime Sync before I switched out my Nvidia GTX 1060 with the RX 590.
[Xorg.0.log](/uploads/5893b6596891a96bf0007b12dabad8c8/Xorg.0.log)
[dmesg.txt](/uploads/b6d85e990025462932a90cbdd1cec76a/dmesg.txt)
xorg-server-1.20.8
mesa-20.0.5
xf86-video-amdgpu-19.1.0
What can I do to further debug this?
To summarize: Prime Sync does not work with amdgpu, but does work with modesetting.https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/17debian/sid: amdgpu crash on undock2023-07-12T20:32:12ZTim Foersterdebian/sid: amdgpu crash on undockI currently trigger the following backtrace on my T495 always on undocking. The crash only happens with opened lid.
```
[108137.178] (II) AMDGPU(0): EDID vendor "AUO", prod id 22333
[108137.179] (II) AMDGPU(0): Printing DDC gathered Mod...I currently trigger the following backtrace on my T495 always on undocking. The crash only happens with opened lid.
```
[108137.178] (II) AMDGPU(0): EDID vendor "AUO", prod id 22333
[108137.179] (II) AMDGPU(0): Printing DDC gathered Modelines:
[108137.179] (II) AMDGPU(0): Modeline "1920x1080"x0.0 142.60 1920 1936 1952 2080 1080 1083 1088 1142 -hsync -vsync (68.6 kHz eP)
[108137.199] (II) AMDGPU(0): EDID vendor "AUO", prod id 22333
[108137.199] (II) AMDGPU(0): Printing DDC gathered Modelines:
[108137.199] (II) AMDGPU(0): Modeline "1920x1080"x0.0 142.60 1920 1936 1952 2080 1080 1083 1088 1142 -hsync -vsync (68.6 kHz eP)
[108137.339] (EE)
[108137.339] (EE) Backtrace:
[108137.340] (EE) 0: /usr/lib/xorg/Xorg (OsLookupColor+0x138) [0x5587e6e21e88]
[108137.344] (EE) 1: /lib/x86_64-linux-gnu/libpthread.so.0 (funlockfile+0x50) [0x7fdb068bd15f]
[108137.346] (EE) unw_get_proc_name failed: no unwind info found [-10]
[108137.347] (EE) 2: /usr/lib/xorg/modules/drivers/amdgpu_drv.so (?+0x0) [0x7fdb05e5aff0]
[108137.348] (EE) unw_get_proc_name failed: no unwind info found [-10]
[108137.348] (EE) 3: /usr/lib/xorg/modules/drivers/amdgpu_drv.so (?+0x0) [0x7fdb05e5b290]
[108137.349] (EE) 4: /usr/lib/xorg/Xorg (xf86DPMSSet+0x63) [0x5587e6d33db3]
[108137.350] (EE) 5: /usr/lib/xorg/Xorg (xf86SetDGAMode+0x886) [0x5587e6d01b06]
[108137.350] (EE) 6: /usr/lib/xorg/Xorg (XaceIsLocal+0x5b16) [0x5587e6d7ac86]
[108137.351] (EE) 7: /usr/lib/xorg/Xorg (mieqProcessInputEvents+0x165) [0x5587e6e01b95]
[108137.352] (EE) 8: /usr/lib/xorg/Xorg (ProcessInputEvents+0x19) [0x5587e6d015f9]
[108137.352] (EE) 9: /usr/lib/xorg/Xorg (SendErrorToClient+0x2c9) [0x5587e6cc48d9]
[108137.353] (EE) 10: /usr/lib/xorg/Xorg (InitFonts+0x3b4) [0x5587e6cc8914]
[108137.356] (EE) 11: /lib/x86_64-linux-gnu/libc.so.6 (__libc_start_main+0xeb) [0x7fdb0670ae0b]
[108137.356] (EE) 12: /usr/lib/xorg/Xorg (_start+0x2a) [0x5587e6cb273a]
[108137.356] (EE)
[108137.356] (EE) Segmentation fault at address 0x0
[108137.356] (EE)
Fatal server error:
[108137.356] (EE) Caught signal 11 (Segmentation fault). Server aborting
[108137.357] (EE)
[108137.357] (EE)
```
```
xserver-xorg-core:
Installed: 2:1.20.8-2
xserver-xorg-video-amdgpu:
Installed: 19.1.0-1
```
```
>>> uname -a
Linux tonobo-top 5.7.0-1-amd64 #1 SMP Debian 5.7.6-1 (2020-06-24) x86_64 GNU/Linux
```
[Xorg.0.log.old](/uploads/70b453e1340e778578e31c0506c5b497/Xorg.0.log.old)https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/24amdgpu (RAVEN) X server crash: "Failed to make 1x1x32bpp GBM bo", bo conflict...2020-09-18T17:21:48ZNick Alcockamdgpu (RAVEN) X server crash: "Failed to make 1x1x32bpp GBM bo", bo conflict (probably a kernel or Mesa bug?)This happened with an Athlon 3000G APU (RAVEN), with kernel 5.8.9, X.org X server 1.20.9, xf86-video-amdgpu commit 442efe73dd579dc36445a3b232937abbed9d2fbb. Dual monitors, one DVI, one HDMI, identical resolutions (1080x1050), effective s...This happened with an Athlon 3000G APU (RAVEN), with kernel 5.8.9, X.org X server 1.20.9, xf86-video-amdgpu commit 442efe73dd579dc36445a3b232937abbed9d2fbb. Dual monitors, one DVI, one HDMI, identical resolutions (1080x1050), effective screen 0 res 3360x1050. After a few days of normal usage (almost no 3D other than the usual glamor usage), I popped up calibre's ebook-viewer, and got an instant black screen, though the machine was still responsive over the network so I could shut down cleanly. There was a spray of tens of thousands of lines of this in the X server log:
[419221.702] (EE) AMDGPU(0): Failed to make 1x1x32bpp GBM bo
At the same time, the kernel log filled up with fifty thousand lines or so of this:
[418985.683944] amdgpu 0000:0a:00.0: amdgpu: bo 00000000a83d33d2 va 0x080010176e-0x080010176f conflict with 0x080010176e-0x0800101770
[418985.684129] amdgpu 0000:0a:00.0: amdgpu: bo 00000000a83d33d2 va 0x080010176e-0x080010176f conflict with 0x080010176e-0x0800101770
[418985.690683] amdgpu 0000:0a:00.0: amdgpu: bo 00000000c4a6faf4 va 0x080010176e-0x080010176f conflict with 0x080010176e-0x0800101770
(different bo address every time, the other two addresses were unchanged.)
This is probably a Mesa or amdgpu kernel bug, but I have no idea which it might be, nor really any idea how to replicate this. (I mean, possibly just rerunning the ebook-viewer would do it, but if I try now and it works I lose this bug report :) )https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/28--scale does not compress or zoom on amdgpu2020-12-10T09:37:35ZRich--scale does not compress or zoom on amdgpuWhen using `--scale`, the framebuffer size is adjusted but the output is not scaled to the display. e.g. on a 1920x1080 display, a scale of 2 produces 3840x2160 output, but the display only shows the upper left 1920x1080 quadrant.
Dis...When using `--scale`, the framebuffer size is adjusted but the output is not scaled to the display. e.g. on a 1920x1080 display, a scale of 2 produces 3840x2160 output, but the display only shows the upper left 1920x1080 quadrant.
Displays connect to the intel gpu scale as expected.
xrandr 1.5.1
xorg 1.20.10
RX570https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/32AMD rx560 could not boot into desktop.2021-03-04T20:50:05Znetroby1AMD rx560 could not boot into desktop.Hello, i fill this issue, because xorg amdgpu driver broken for AMD rx560 ,
I tried install many linux distro, both them failed to boot into desktop, (xorg seems not working, i think it was a amd gpu driver problem)
The affected distro...Hello, i fill this issue, because xorg amdgpu driver broken for AMD rx560 ,
I tried install many linux distro, both them failed to boot into desktop, (xorg seems not working, i think it was a amd gpu driver problem)
The affected distro:
Ubuntu 18.04, to Ubuntu 20.10
Linut mint 19.3, 20.1 (With mate , linuxmint cinnamon) LMDE 4,
Fedora linux 33
x86_64 arch.
The product link: https://item.m.jd.com/product/8413442.html?wxa_abtest=b&utm_user=plusmember&gx=RnE2wTEKbDWLmNRL-sA1AgOz&ad_od=share&utm_source=androidapp&utm_medium=appshare&utm_campaign=t_335139774&utm_term=CopyURL
华擎(ASRock)RX560 4G 幻影国度 钻石版 Phantom Gaming 1176-1223 MHz/7000 MHz 4GB GDDR5 显卡
I can only login into linux rescure mode.
and i also see the syslog,
looks the xsesssion was terminated due to some segmentfault.
The same hardware, which running windows 10 well.
so the hardware has no problem i guess.https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/40[nouveau + amdgpu] external screen response very slow when internal display s...2021-08-16T15:45:32Zaaditya[nouveau + amdgpu] external screen response very slow when internal display switched offHi,
If both internal display and external display are connected, things work fine. However if I disable the internal laptop display, the external display responds very slowly - windows take a lot of time to open up and all graphical upd...Hi,
If both internal display and external display are connected, things work fine. However if I disable the internal laptop display, the external display responds very slowly - windows take a lot of time to open up and all graphical updates are slow. Mouse cursor acts normal though.
Could not find anything useful in dmesg when this happens.
My system details:
Hp Omen 15 laptop, Amd Ryzen 4600H (amdgpu) with Nvidia 1660Ti (nouveau) running Debian Bullseye.
~~~~
aaditya@debian-ext:~$ inxi -Fxz
System: Kernel: 5.10.56-xanmod1 x86_64 bits: 64 compiler: gcc v: 10.2.1 Desktop: GNOME 3.38.4
Distro: Debian GNU/Linux 11 (bullseye)
Machine: Type: Laptop System: HP product: OMEN Laptop 15-en0xxx v: N/A serial: <filter>
Mobo: HP model: 8786 v: 22.58 serial: <filter> UEFI: AMI v: F.13 date: 03/04/2021
Battery: ID-1: BAT0 charge: 66.2 Wh condition: 66.2/66.2 Wh (100%) model: Hewlett-Packard Primary status: Full
CPU: Info: 6-Core model: AMD Ryzen 5 4600H with Radeon Graphics bits: 64 type: MT MCP arch: Zen 2 rev: 1 L2 cache: 3 MiB
flags: avx avx2 lm nx pae sse sse2 sse3 sse4_1 sse4_2 sse4a ssse3 svm bogomips: 71865
Speed: 1399 MHz min/max: 1400/3000 MHz boost: enabled Core speeds (MHz): 1: 1399 2: 1397 3: 1399 4: 1323 5: 1683
6: 1397 7: 1397 8: 1396 9: 1397 10: 1396 11: 1397 12: 1396
Graphics: Device-1: NVIDIA TU116M [GeForce GTX 1660 Ti Mobile] vendor: Hewlett-Packard driver: nouveau v: kernel
bus ID: 01:00.0
Device-2: Advanced Micro Devices [AMD/ATI] Renoir vendor: Hewlett-Packard driver: amdgpu v: kernel bus ID: 07:00.0
Device-3: Cheng Uei Precision Industry (Foxlink) HP Wide Vision HD Camera type: USB driver: uvcvideo bus ID: 3-3:3
Display: x11 server: X.Org 1.20.11 driver: loaded: amdgpu,ati,modesetting unloaded: fbdev,vesa resolution:
1: 1280x720~60Hz 2: 1600x900
OpenGL: renderer: AMD RENOIR (DRM 3.40.0 5.10.56-xanmod1 LLVM 11.0.1) v: 4.6 Mesa 20.3.5 direct render: Yes
Audio: Device-1: NVIDIA TU116 High Definition Audio vendor: Hewlett-Packard driver: snd_hda_intel v: kernel
bus ID: 01:00.1
Device-2: Advanced Micro Devices [AMD] Raven/Raven2/FireFlight/Renoir Audio Processor vendor: Hewlett-Packard
driver: snd_rn_pci_acp3x v: kernel bus ID: 07:00.5
Device-3: Advanced Micro Devices [AMD] Family 17h HD Audio vendor: Hewlett-Packard driver: snd_hda_intel v: kernel
bus ID: 07:00.6
Sound Server: ALSA v: k5.10.56-xanmod1
Network: Device-1: Realtek RTL8111/8168/8411 PCI Express Gigabit Ethernet vendor: Hewlett-Packard driver: r8169 v: kernel
port: e000 bus ID: 02:00.0
IF: eno1 state: down mac: <filter>
Device-2: Intel Wi-Fi 6 AX200 driver: iwlwifi v: kernel port: e000 bus ID: 03:00.0
IF: wlo1 state: up mac: <filter>
Bluetooth: Device-1: Intel AX200 Bluetooth type: USB driver: btusb v: 0.8 bus ID: 3-4:5
Report: ID: hci0 state: down address: <filter>
Drives: Local Storage: total: 2.29 TiB used: 166.68 GiB (7.1%)
ID-1: /dev/nvme0n1 vendor: Kingston model: SA2000M81000G size: 931.51 GiB temp: 30.9 C
ID-2: /dev/nvme1n1 vendor: Samsung model: MZVLB512HBJQ-000H1 size: 476.94 GiB temp: 39.9 C
ID-3: /dev/sda type: USB vendor: Crucial model: CT1000MX 500SSD1 size: 931.51 GiB
Partition: ID-1: / size: 156.49 GiB used: 21.75 GiB (13.9%) fs: ext4 dev: /dev/sda2
Swap: ID-1: swap-1 type: file size: 8 GiB used: 0 KiB (0.0%) file: /swapfile
ID-2: swap-2 type: zram size: 3.77 GiB used: 0 KiB (0.0%) dev: /dev/zram0
Sensors: System Temperatures: cpu: 54.9 C mobo: N/A
Fan Speeds (RPM): N/A
GPU: device: nouveau temp: 42.0 C fan: 0 device: amdgpu temp: 47.0 C
Info: Processes: 348 Uptime: 7m Memory: 15.06 GiB used: 1.3 GiB (8.6%) Init: systemd runlevel: 5 Compilers: gcc: 10.2.1
Packages: 2355 Shell: Bash v: 5.1.4 inxi: 3.3.01
~~~~
Tested with openbox and GNOME. If I turn off compositor in openbox it works as expected, but there are lots of visual artifacts (small black boxes) when activating windows and when hovering over videos in youtube.
Thanks for your help!https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/46segfault when using DRI_PRIME=12022-04-25T19:34:10ZTom St Denissegfault when using DRI_PRIME=1While using DRI_PRIME=1 on my Fedora 35 stock install with an APU and dGPU piglit (and other) programs can cause dri2 to segfault reliably.
The backtrace always points to the same dri2 functions:
```
Thread 1 "Xorg" received signal SIG...While using DRI_PRIME=1 on my Fedora 35 stock install with an APU and dGPU piglit (and other) programs can cause dri2 to segfault reliably.
The backtrace always points to the same dri2 functions:
```
Thread 1 "Xorg" received signal SIGSEGV, Segmentation fault.
0x0000000000000000 in ?? ()
(gdb) backtrace
#0 0x0000000000000000 in ()
#1 0x000055ae6f19c884 in ShmDestroyPixmap (pPixmap=<optimized out>) at ../../Xext/shm.c:260
#2 0x00007f5c6123d285 in amdgpu_dri2_destroy_buffer2 (pScreen=<optimized out>, drawable=<optimized out>, buffers=0x55ae71844a20)
at /usr/src/debug/xorg-x11-drv-amdgpu-21.0.0-1.fc35.x86_64/src/amdgpu_dri2.c:227
#3 0x000055ae6f226763 in do_get_buffers
(pDraw=pDraw@entry=0x55ae72c1f530, width=0x7ffcfbbc2a80, height=0x7ffcfbbc2a88, attachments=<optimized out>, count=1, out_count=out_count@entry=0x7ffcfbbc2a90, has_format=1) at dri2/dri2.c:683
#4 0x000055ae6f22822f in DRI2GetBuffersWithFormat (out_count=<optimized out>, count=<optimized out>, attachments=<optimized out>, height=<optimized out>, width=<optimized out>, pDraw=<optimized out>)
at dri2/dri2.c:708
#5 ProcDRI2GetBuffersWithFormat (client=<optimized out>) at dri2/dri2ext.c:317
#6 ProcDRI2Dispatch (client=0x55ae71be87b0) at dri2/dri2ext.c:619
#7 0x000055ae6f0ec2d7 in Dispatch () at ../../dix/dispatch.c:478
#8 dix_main (envp=<optimized out>, argv=0x7ffcfbbc2ca8, argc=6) at ../../dix/main.c:276
#9 main (argc=6, argv=0x7ffcfbbc2ca8, envp=<optimized out>) at ../../dix/stubmain.c:34
```https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/49vt switch between X servers broken after mst hub reinitalized2022-10-03T08:41:02ZChristoph Haagvt switch between X servers broken after mst hub reinitalizedI have two monitors connected to an MST hub.
To reproduce:
Start two X sessions. While in one X session, reconnect the MST hub. Mine has a physical "reset" button, that's enough to trigger it too. After the monitors are back, switch to ...I have two monitors connected to an MST hub.
To reproduce:
Start two X sessions. While in one X session, reconnect the MST hub. Mine has a physical "reset" button, that's enough to trigger it too. After the monitors are back, switch to the other X session. The switch will fail and the other X session is completely killed.
```
[ 30811.548] (EE) AMDGPU(0): failed to set mode: Invalid argument
[ 30811.548] (WW) AMDGPU(0): Failed to set mode on CRTC 0
[ 30811.548] (EE) AMDGPU(0): drmmode_do_crtc_dpms cannot get last vblank counter
[ 30811.549] (EE) AMDGPU(0): failed to set mode: Invalid argument
[ 30811.549] (WW) AMDGPU(0): Failed to set mode on CRTC 1
[ 30811.549] (EE) AMDGPU(0): drmmode_do_crtc_dpms cannot get last vblank counter
[ 30811.550] (EE) AMDGPU(0): Failed to enable any CRTC
[ 30811.550] (EE)
Fatal server error:
[ 30811.550] (EE) EnterVT failed for screen 0
```
Full log: [Xorg.1.log](/uploads/1d8b23ce6f220cff7e538c228f86c015/Xorg.1.log)
With modesetting the switch works, but after that switch modesetting is sometimes very confused about which monitor is which and controlling position or on/off state might control the wrong one. Also sometimes one monitor is just black until disabling and enabling it again via randr.
Possibly related: https://gitlab.freedesktop.org/drm/amd/-/issues/1336https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/52Instability after installing the xf86-video-amdgpu package2022-09-08T12:49:26ZPQCraftInstability after installing the xf86-video-amdgpu packageOn installing the `xf86-video-amdgpu` package, graphical glitches and kernel panics have been occurring. I removed the package for a few weeks and my system became stable again but due to a problem with the kernel driver and my display, ...On installing the `xf86-video-amdgpu` package, graphical glitches and kernel panics have been occurring. I removed the package for a few weeks and my system became stable again but due to a problem with the kernel driver and my display, I had to reinstall. After reinstalling, my system became unstable again. On one of these frequent unstable blips, graphical glitches appeared on my browser where it appeared as though tiles would not update and then it caused another app to crash until I switched back to the browser which caused it to die with SIGILL (illegal instruction). I have no logs from the kernel panics as either my system or kernel is somehow configured not to save them. My hardware is a Ryzen 5 3500U. The GPU memory is shared with the system memory (I have 12GB installed and only 10GB is visible to the OS while the other 2GB appears as VRAM).
Neofetch lines:
```
CPU: AMD Ryzen 5 3500U with Radeon Vega Mobile Gfx (8) @ 2.100GHz
GPU: AMD ATI Radeon Vega Series / Radeon Vega Mobile Series
```
Side note: I suspect (but have little evidence to prove) that there is little to no protection on VRAM overflows as programs that cause the VRAM to overflow seem to crash or cause the system to become unstable. I can verify that the example above did not happen due to a VRAM overflow however as I had the system monitor open and could see that only 1GB of VRAM was in use.https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/56base kernel driver installs with WARNING but will not load on Dell R750 serve...2023-04-01T19:03:21Zwkaileybase kernel driver installs with WARNING but will not load on Dell R750 server running OpenSuse Leap 15.3 with Linux kernel 5.3.18I am installing the ROCm all open driver stack 5.2.1 with the --usecase=openclsdk, but the same thing happens if I don't use this option and install everything.
I am running on a Dell R750 server with Intel Ice Lake CPUs and Linux kernel...I am installing the ROCm all open driver stack 5.2.1 with the --usecase=openclsdk, but the same thing happens if I don't use this option and install everything.
I am running on a Dell R750 server with Intel Ice Lake CPUs and Linux kernel 5.3.18-15300.59.63-default.
I downloaded the ROCm 5.2.1 version of amdgpu-install and installed it using zypper on OpenSuse Leap 15.3.
The installation succeeds, but it gives the warning:
WARNING: nomodeset detected in kernel parameters, amdgpu requires KMS
The rocminfo app is not included in the installation for some reason, so we used clinfo to check if everything is working. **clinfo reports**
```
Number of platforms: 1
Platform Profile: FULL_PROFILE
Platform Version: OpenCL 2.1 AMD-APP (3452.0)
Platform Name: AMD Accelerated Parallel Processing
Platform Vendor: Advanced Micro Devices, Inc.
Platform Extensions: cl_khr_icd cl_amd_event_callback
Platform Name: AMD Accelerated Parallel Processing
Number of devices: 0
```
**dkms status reports**
```
amdgpu, 5.16.9.22.20-1447096, 5.3.18-150300.59.63-default, x86_64: installed (original_module exists)
```
When we try to edit he boot script using grub to remove 'nomodeswitch', the server will not boot at all, and we have to use grub to reinsert 'nomodeswitch' before we can get it to boot.
Is this a known incompatibility with Dell servers, since they seem to require 'nomodeswitch', and is there any way around it?
I do not understand why we need nomodeswitch or KMS, since we are running this as a headless server, except when we log onto the network KVM system console. Since KMS is associated with setting display driver modes during boot up, as I understand it, is there some way we can inhibit the need for this, inasmuch as we don't want the GPU to display anything. We are using it for computation.
Cordially,
Walter F. Kailey, Ph.D.
Performance Engineer
Senior Software Development Engineer
Ricoh USAhttps://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/57Can't set any resolution/mode on some screen2022-11-16T21:18:15ZChristopher CourtoisCan't set any resolution/mode on some screenCan't get any display on a screen except the initial boot messages.
After boot it won't accept any resolution/mode.
Tried every mode returned by the EDID but none works, here is the error returned by xrandr
```
❯ xrandr --output HDMI-A...Can't get any display on a screen except the initial boot messages.
After boot it won't accept any resolution/mode.
Tried every mode returned by the EDID but none works, here is the error returned by xrandr
```
❯ xrandr --output HDMI-A-0 --preferred
xrandr: Configure crtc 0 failed
```
```
❯ xrandr
Screen 0: minimum 320 x 200, current 1440 x 2560, maximum 16384 x 16384
HDMI-A-0 connected primary (normal left inverted right x axis y axis)
1440x2560 50.16 +
1280x1024 50.16
1440x900 50.16
1280x800 50.16
1280x720 50.16
1024x768 50.16
800x600 50.16
640x480 50.16
HDMI-A-1 disconnected (normal left inverted right x axis y axis)
DisplayPort-0 disconnected (normal left inverted right x axis y axis)
DisplayPort-1 disconnected (normal left inverted right x axis y axis)
```
Looking at the Xorg log, this:
```
[ 5.070] (EE) AMDGPU(0): failed to set mode: Invalid argument
[ 5.070] (WW) AMDGPU(0): Failed to set mode on CRTC 0
[ 5.070] (EE) AMDGPU(0): Failed to enable any CRTC
```
Tested the same screen on other machines with either Intel HD Graphics or NVidia GPU and its works good.
Also tried with wayland instead of Xorg but no chances here too.
[Xorg.0.log](/uploads/a856e65983679b9c729c60c8d2793468/Xorg.0.log)
[dmesg.log](/uploads/5008e03f83c0709a81ce60137c2acbae/dmesg.log)https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/59Output source runs at 1FPS when laptop lid is closed - issue is back2023-11-26T20:28:44ZMarcin SzewczykOutput source runs at 1FPS when laptop lid is closed - issue is backWith recent nvidia drivers version 525.60 the issue is back, same as [here](https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/37). The fix was working for few months and with new nvidia drivers is back. Downgrading nv...With recent nvidia drivers version 525.60 the issue is back, same as [here](https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/37). The fix was working for few months and with new nvidia drivers is back. Downgrading nvidia drivers to 520.56 solves the problem.
My specs:
```
System:
Host: vampirhon Kernel: 6.0.12-arch1-1 arch: x86_64 bits: 64 Desktop: GNOME
v: 43.2 Distro: Arch Linux
Machine:
Type: Laptop System: ASUSTeK product: ROG Zephyrus G14 GA401IV_GA401IV
v: 1.0 serial: <superuser required>
Mobo: ASUSTeK model: GA401IV v: 1.0 serial: <superuser required>
UEFI: American Megatrends v: GA401IV.219 date: 12/30/2020
Graphics:
Device-1: NVIDIA TU106M [GeForce RTX 2060 Max-Q] driver: nvidia v: 525.60.11
Device-2: AMD Renoir driver: amdgpu v: kernel
driver: snd-usb-audio,uvcvideo
Display: x11 server: X.org v: 1.21.1.4 with: Xwayland v: 22.1.5 driver: X:
loaded: amdgpu,nvidia unloaded: modesetting dri: radeonsi
gpu: amdgpu,nvidia,nvidia-nvswitch resolution: 1: 1920x1080 2: 1920x1080
3: 1920x1080~120Hz
API: OpenGL Message: Unable to show GL data. Required tool glxinfo
missing.
```
xorg packages:
- xorg-fonts-encodings 1.0.6-1
- xorg-server 21.1.4-1
- xorg-server-common 21.1.4-1
- xorg-setxkbmap 1.3.3-1
- xorg-xev 1.2.5-1
- xorg-xhost 1.0.9-1
- xorg-xinput 1.6.3-3
- xorg-xkbcomp 1.4.6-1
- xorg-xprop 1.2.6-1
- xorg-xrandr 1.5.2-1
- xorg-xrdb 1.2.1-1
- xorg-xset 1.2.5-1
- xorg-xwayland 22.1.5-1
- xorgproto 2022.2-1https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/60Cursor is now affected by redshift gamma ramps2023-10-24T02:13:08ZTasos SahanidisCursor is now affected by redshift gamma ramps## Brief summary of the problem:
After updating the kernel, under X11, the cursor is now noticeably affected by gamma ramps as applied by [redshift](https://github.com/jonls/redshift). This behaviour was introduced by [drm/amd/display: ...## Brief summary of the problem:
After updating the kernel, under X11, the cursor is now noticeably affected by gamma ramps as applied by [redshift](https://github.com/jonls/redshift). This behaviour was introduced by [drm/amd/display: don't enable DRM CRTC degamma property for DCE (0e444a4de6b3)](https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=0e444a4de6b38c4593a07e4cfb5bf54c40cc79b6).
At first I thought this occurred because hardware cursors were no longer being used, but comparing drm_info with and without the commit reverted, assuming I'm understanding it correctly, it seems that they are now used with that commit.
It's really difficult to tell by looking at a screen if it was applied before, but maybe after this change, the gamma ramps are now somehow applied multiple times, intensifying the effect.
When the cursor is moved to the display attached to the card driven by nouveau, it always remains unaffected.
## Hardware description:
- GPU:
```
*-display
description: VGA compatible controller
product: Bonaire XTX [Radeon R7 260X/360] [1002:6658]
vendor: Advanced Micro Devices, Inc. [AMD/ATI] [1002]
physical id: 0
bus info: pci@0000:08:00.0
version: 00
width: 64 bits
clock: 33MHz
capabilities: pm pciexpress msi vga_controller bus_master cap_list rom
configuration: driver=amdgpu latency=0
resources: irq:33 memory:d0000000-dfffffff memory:ef000000-ef7fffff ioport:c800(size=256) memory:feac0000-feafffff memory:c0000-dffff
*-display
description: VGA compatible controller
product: G86 [GeForce 8400 GS] [10DE:422]
vendor: NVIDIA Corporation [10DE]
physical id: 0
bus info: pci@0000:07:00.0
version: a1
width: 64 bits
clock: 33MHz
capabilities: pm msi pciexpress vga_controller bus_master cap_list rom
configuration: driver=nouveau latency=0
resources: irq:30 memory:fd000000-fdffffff memory:b0000000-bfffffff memory:fa000000-fbffffff ioport:bc00(size=128) memory:fe9e0000-fe9fffff
```
## System information:
- Distro name and Version: Ubuntu 20.04.5 LTS
- Kernel version: torvalds/linux.git (b6bb9676f2165d518b35ba3bea5f1fcfc0d969bf)
- Custom kernel: custom config
- Mesa: 22.3.1 - kisak-mesa
## How to reproduce the issue:
Run a recent kernel (after the commit that introduced this behaviour) and enable redshift. For example, to force a visible change `redshift -t 4000K:4600K`.
## Attached files:
### Screenshots/video files
![cursors](/uploads/0ba670523a466063d6fea6994338f02c/cursors.jpg)
### Log files
- drm_info with the commit reverted: [drm_info_reverted.txt](/uploads/f1f649eef6c8b26150193b55f05612b1/drm_info_reverted.txt)
- drm_info without revert: [drm_info_not_reverted.txt](/uploads/be111283641b84e5955f7ea5caeb2fbe/drm_info_not_reverted.txt)https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/62color calibration curve is not being applied to absolute black on RX 66002023-02-04T09:36:38ZNicolas Carranzacolor calibration curve is not being applied to absolute black on RX 6600Recently upgraded my graphics card from AMD RX 460 to an RX 6600 and I'm noticing a bug on the way the RX 6600 is applying the color calibration curves. The card is ignoring the calibration correction for black (RGB 0/0/0). This bug is r...Recently upgraded my graphics card from AMD RX 460 to an RX 6600 and I'm noticing a bug on the way the RX 6600 is applying the color calibration curves. The card is ignoring the calibration correction for black (RGB 0/0/0). This bug is reproducible by applying any calibration that modifies the RGB 0/0/0 value: the modification is ignored and it will always display the black color as if no calibration curve has been applied on black (and black only) 0/0/0. An easy way to reproduce this bug is running "xcalib -alter -gammacor 0.5" (https://github.com/OpenICC/xcalib) and noticing that all the colors get displayed lighter as expected except black 0/0/0 (it is not affected at all). This bug was not present on the RX 460.
I'm running linux kernel 5.15.91 , with x11-drivers/xf86-video-amdgpu version 22.0.0. Let me know if you need more info, I guess this is a software bug somewhere on xorg or the amdgpu driver/kernel mod. Thank you.
Update: Another way I can reproduce this bug is by doing "xrandr --output HDMI-A-0 --gamma 10". The bug manifests the same as by using the xcalib command previously noted.https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/65NetBSD: bswap* re-definitions2023-02-27T10:23:00ZThomas KlausnerNetBSD: bswap* re-definitionsThe configure script mentions that it knows that bswap16 etc. exist on NetBSD:
```
checking for sys/endian.h... yes
checking for __swap16 variant of <sys/endian....The configure script mentions that it knows that bswap16 etc. exist on NetBSD:
```
checking for sys/endian.h... yes
checking for __swap16 variant of <sys/endian.h> byteswapping macros... no
checking for bswap16 variant of <sys/endian.h> byteswapping macros... yes
checking that generated files are newer than configure... done
```
but re-defines them anyway, leading to:
```
In file included from /usr/include/stdlib.h:37,
from amdgpu_drv.h:40,
from amdgpu_dri2.c:32:
../config.h:138:18: error: expected declaration specifiers or '...' before '__builtin_constant_p'
138 | #define bswap_64 bswap64
| ^~~~~~~
/scratch/x11/xf86-video-amdgpu/work/.buildlink/include/xorg/misc.h:302:1: note: in expansion of macro 'bswap_64'
302 | bswap_64(uint64_t x)
| ^~~~~~~~
../config.h:135:18: error: expected declaration specifiers or '...' before '__builtin_constant_p'
135 | #define bswap_32 bswap32
| ^~~~~~~
/scratch/x11/xf86-video-amdgpu/work/.buildlink/include/xorg/misc.h:321:1: note: in expansion of macro 'bswap_32'
321 | bswap_32(uint32_t x)
| ^~~~~~~~
../config.h:132:18: error: expected declaration specifiers or '...' before '__builtin_constant_p'
132 | #define bswap_16 bswap16
| ^~~~~~~
/scratch/x11/xf86-video-amdgpu/work/.buildlink/include/xorg/misc.h:365:1: note: in expansion of macro 'bswap_16'
365 | bswap_16(uint16_t x)
| ^~~~~~~~
```https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/issues/67Fix conditions for enabling non-legacy/atomic DDX color mgmt support2024-03-07T04:32:14ZMelissa WenFix conditions for enabling non-legacy/atomic DDX color mgmt supportThe conditions for CRTC color mgmt support are not defined correctly. It only enables atomic (non-legacy) color mgmt support if all DRM CRTC color properties are enabled (https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/blo...The conditions for CRTC color mgmt support are not defined correctly. It only enables atomic (non-legacy) color mgmt support if all DRM CRTC color properties are enabled (https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/-/blob/master/src/drmmode_display.c#L3342)
```
/* cm is enabled only if all prop ids are found */
```
But according to DRM documentation, each DRM color mgmt property is optional and the driver should enable only the supported properties (in the set of degamma/ctm/gamma props). KMD uses gamma LUT size to decide between [legacy and atomic path](https://gitlab.freedesktop.org/agd5f/linux/-/blob/amd-staging-drm-next/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_color.c#L420) and enables each property [according to the HW caps](https://gitlab.freedesktop.org/agd5f/linux/-/blob/amd-staging-drm-next/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c#L467).