[BUG] Wrongly GPU detection?

[AngryPenguinPL]

General description of bug:

• What happened: After upgrade to 2.8.2 my GPU is not detected correctly but instead it it detect whole series.
  As example before 2.8.2 my GPU was detected as: AMD Radeon RX580 (that correct) and now is: MD Radeon RX 470/480/570/570X/580/580X/590 - so right now whole Polaris series.
• What should happen: GPU should be detected correctly.
• Did it work in an older version: Yes, 2.7.1
• Where did you get the binary: Compiled by myself.
• Does this issue still occurs in the latest dev build? Not tested.

Often helpful information:

Screenshot:

On left side old fastfetch 2.7.1 and on right side new fastfetcg 2.8.2

The content of the configuration file you use (if any):

//paste here

Output of fastfetch -c ci.jsonc --format json:

// https://gist.github.com/AngryPenguinPL/1a444cc553b5acf7468e0a80e51a02ed

Output of fastfetch --list-features:

//fastfetch --list-features
threads
vulkan
wayland
xcb-randr
xcb
xrandr
x11
drm
gio
dconf
dbus
imagemagick7
zlib
sqlite3
rpm
egl
glx
osmesa
opencl
libpulse
libnm
videodev2
Proprietary GPU driver API

If fastfatch crashed

Paste the stacktrace here. You may get it with:

$ gdb /path/to/fastfetch
$ run
$ bt

If you are able to identify which module crashed, the strace can be helpful too

$ strace /path/to/fastfetch --multithreading false --structure {MODULE} --pipe

If you cannot do the instructions above, please upload the core dump file:

If my image logo didn't show / work

• The image protocol you used:
• The terminal you used:
• Upload the image file here, or paste the image URL:
• Does it work with --logo-width {WIDTH} --logo-height {HEIGHT}?

If fastfetch behaves incorrectly on shell starting

• The bug is reproduceable with fresh / clean shell configuration (i.e. fastfetch is the single line of .zshrc or ~/.config/fish/config.fish):
• Does sleep 1 before running fastfetch work?

[CarterLi]

Thanks for reporting. What does cat /sys/class/drm/card{0,1,2}/device/uevent print?

[AngryPenguinPL]

cat /sys/class/drm/card{0,1,2}/device/uevent
DRIVER=amdgpu
PCI_CLASS=30000
PCI_ID=1002:67DF
PCI_SUBSYS_ID=1DA2:E353
PCI_SLOT_NAME=0000:01:00.0
MODALIAS=pci:v00001002d000067DFsv00001DA2sd0000E353bc03sc00i00
cat: /sys/class/drm/card1/device/uevent: No such file or directory
cat: /sys/class/drm/card2/device/uevent: No such file or directory

[SuperMewio]

I have come here with the same issue.

My RX 6800 XT now states "GPU: AMD Radeon RX 6800/6800 XT / 6900 XT" instead of it being an RX 6800 XT as it did prior to this update.

I do not use a config but a line in my fish configuration file and bashrc.

#Fastfetch
fastfetch --color green -s Seperator:Kernel:Packages:CPU:GPU:Memory:Seperator --logo Arch_small  --cpu-format "{1} ({5}) @ 4.15GHz"  --packages-format "{2} (P), {15} (F-U)"  --set-keyless Seperator=❰·───────────────────────────────────·❱

My output of cat /sys/class/drm/card{0,1,2}/device/uevent

❯ cat /sys/class/drm/card{0,1,2}/device/uevent
DRIVER=amdgpu
PCI_CLASS=30000
PCI_ID=1002:73BF
PCI_SUBSYS_ID=1EAE:6701
PCI_SLOT_NAME=0000:09:00.0
MODALIAS=pci:v00001002d000073BFsv00001EAEsd00006701bc03sc00i00
cat: /sys/class/drm/card1/device/uevent: No such file or directory
cat: /sys/class/drm/card2/device/uevent: No such file or directory

[CarterLi]

Please test if the latest dev build works for you

[AngryPenguinPL]

Please test if the latest dev build works for you

Almost. Now my GPU is visible as Radeon RX 570 4GB, while my GPU is RX580 and 8GB.

[CarterLi]

Seems that RX570 and RX580 use the same device ID and sub device ID. The only difference is the revision number.

The same issue exists in neofetch: dylanaraps/neofetch#1985

[CarterLi]

Please pull the dev branch and test it again

[AngryPenguinPL]

Yes, looks like it is fixed. Thanks.

[SuperMewio]

Looks like the dev build fixed it for me as well

[CarterLi]

Thanks for testing