Error: glXChooseFBConfig() failed Error: Error: Couldn't get GL visual config! Error: main: Could not initialize canvas

[honey2006]

Soumy@localhost:~/Downloads$ MESA_LOADER_DRIVER_OVERRIDE=zink glmark2
TU: error: ../src/freedreno/vulkan/tu_device.cc:601: device (chip_id = 7010000, gpu_id = 710) is unsupported (VK_ERROR_INCOMPATIBLE_DRIVER)
MESA: error: ZINK: vkEnumeratePhysicalDevices failed (VK_ERROR_INITIALIZATION_FAILED)
MESA: error: ZINK: failed to choose pdev
glx: failed to create drisw screen
failed to load driver: zink
TU: error: ../src/freedreno/vulkan/tu_device.cc:601: device (chip_id = 7010000, gpu_id = 710) is unsupported (VK_ERROR_INCOMPATIBLE_DRIVER)
MESA: error: ZINK: vkEnumeratePhysicalDevices failed (VK_ERROR_INITIALIZATION_FAILED)
MESA: error: ZINK: failed to choose pdev
glx: failed to create drisw screen
failed to load driver: zink
TU: error: ../src/freedreno/vulkan/tu_device.cc:601: device (chip_id = 7010000, gpu_id = 710) is unsupported (VK_ERROR_INCOMPATIBLE_DRIVER)
MESA: error: ZINK: vkEnumeratePhysicalDevices failed (VK_ERROR_INITIALIZATION_FAILED)
MESA: error: ZINK: failed to choose pdev
glx: failed to create drisw screen
failed to load driver: zink
Error: glXChooseFBConfig() failed
Error: Error: Couldn't get GL visual config!
Error: main: Could not initialize canvas

[honey2006]

That output is the key piece. Your Debian/PRoot environment currently sees only llvmpipe as a Vulkan device. It does not see the Adreno 710 at all.

You have:

GPU0:
deviceType = PHYSICAL_DEVICE_TYPE_CPU
deviceName = llvmpipe
driverName = llvmpipe

There is no:

GPU1:
Turnip
Adreno
Freedreno

So Zink cannot work right now, because Zink needs a Vulkan GPU device underneath it.

What this means

Your stack currently looks like:

glmark2
   ↓
OpenGL
   ↓
Zink        ← you are trying to use this
   ↓
Vulkan
   ↓
llvmpipe    ← only Vulkan device available
   ↓
CPU

Whereas what you want is:

glmark2
   ↓
OpenGL
   ↓
Zink
   ↓
Turnip
   ↓
KGSL
   ↓
Adreno 710
   ↓
GPU

The missing connection is KGSL → Turnip.

Don't focus on glmark2 yet

We need to get Vulkan to report the Adreno first.

The fact that your earlier error contained:

gpu_id = 710

is actually encouraging. It means Turnip was being loaded at some point and was able to identify your GPU, but the particular Turnip build rejected it.

After your upgrade, however, vulkaninfo is only exposing llvmpipe.

Do this from native Termux

Exit your Debian PRoot:

exit

You should be back at the normal Termux shell.

Now run:

ls -l /dev/kgsl-3d0

Then:

ls -l /dev | grep kgsl

Then:

find /sys/class/kgsl -maxdepth 2 -type f 2>/dev/null | head -30

And:

getprop | grep -iE 'gpu|kgsl|vulkan'

Most important test

In native Termux, run:

ls -l /vendor/lib64/hw/vulkan*

and:

ls -l /vendor/lib64/libvulkan.so*

Depending on your Android version, those paths may differ. So also run:

find /vendor /system -iname '*vulkan*' 2>/dev/null | head -50

Then tell me your Termux setup

Specifically, are you using:

• Termux:X11
• PRoot-Distro Debian
• Android version
• phone/device model

You can get Android version with:

getprop ro.build.version.release

and device model:

getprop ro.product.model

---

One thing I want to correct from my previous answer

Your output does not establish that the Adreno device is physically inaccessible from Android. It establishes specifically that the Vulkan loader inside your Debian PRoot currently has no working Turnip GPU driver/device to enumerate.

That's an important distinction.

Your original:

gpu_id = 710

shows that a Turnip component did encounter the Adreno 710. The subsequent installation of Debian's standard:

mesa-vulkan-drivers 25.0.7-2+deb13u1

changed the situation, and now vulkaninfo only sees llvmpipe.

So don't reinstall random Mesa packages yet. Let's inspect the native Termux KGSL/Turnip setup first. Once you give me the outputs above, I can tell you exactly what needs to be bridged into your Debian PRoot environment.