Help with hashcat on Debian: "Unstable OpenCL driver detected!"

Help with hashcat on Debian: "Unstable OpenCL driver detected!" - Printable Version

+- hashcat Forum (https://hashcat.net/forum)
+-- Forum: Support (https://hashcat.net/forum/forum-3.html)
+--- Forum: hashcat (https://hashcat.net/forum/forum-45.html)
+--- Thread: Help with hashcat on Debian: "Unstable OpenCL driver detected!" (/thread-13362.html)

Help with hashcat on Debian: "Unstable OpenCL driver detected!" - erjoalgo - 09-01-2025

I'm trying to get my AMD card working with hashcat on Debian. I've installed the latest version of hashcat 7.1.2. When I try a benchmark, I get the warning "Unstable OpenCL driver detected!", and a segfault when trying to --force it.

Code:
█[asrock][hashcat][255]$ sudo sudo hashcat -b -d1 --force

hashcat (v7.1.2-65-g5e615188b) starting in benchmark mode

Benchmarking uses hand-optimized kernel code by default.

You can use it in your cracking session by setting the -O option.

Note: Using optimized kernel code limits the maximum supported password length.

To disable the optimized kernel code in benchmark mode, use the -w option.

You have enabled --force to bypass dangerous warnings and errors!

This can hide serious problems and should only be done when debugging.

Do not report hashcat issues encountered when using --force.

OpenCL API (OpenCL 1.1 Mesa 22.3.6) - Platform #1 [Mesa]

========================================================

* Device #01: PITCAIRN (, LLVM 15.0.6, DRM 2.50, 6.1.0-38-amd64), 2048/2048 MB (512 MB allocatable), 20MCU

OpenCL API (OpenCL 3.0 PoCL 3.1+debian  Linux, None+Asserts, RELOC, SPIR, LLVM 15.0.6, SLEEF, DISTRO, POCL_DEBUG) - Platform #2 [The pocl project]

==================================================================================================================================================

* Device #02: pthread-haswell-Intel(R) Core(TM) i7-6700 CPU @ 3.40GHz, skipped

Benchmark relevant options:

===========================

* --force

* --backend-devices=1

* --backend-devices-virtmulti=1

* --backend-devices-virthost=1

* --optimized-kernel-enable

---------------------

* Hash-Mode 900 (MD4)

---------------------

Segmentation fault

█[asrock][hashcat][139]$

I'm willing to install whatever proprietary driver is needed. I tried installing these nonfree packages, but nothing seems to change.

Code:
sudo apt-get install firmware-linux firmware-linux-nonfree libdrm-amdgpu1 xserver-xorg-video-amdgpu

What change should I expect to see after installing these non-free drivers? How do I verify that the new driver is being used, or how do I tell the OS to use the newly installed driver? Do I need to reboot?

My inxi -G output appears unchanged:

Code:
█[asrock][hashcat][0]$ inxi -G

Graphics:

 Device-1: AMD Curacao PRO [Radeon R7 370 / R9 270/370 OEM] driver: radeon v: kernel

 Display: server: X.org v: 1.21.1.7 driver: X: loaded: amdgpu unloaded: fbdev,modesetting,vesa

   dri: radeonsi gpu: radeon tty: 112x25

 API: OpenGL Message: GL data unavailable in console. Try -G --display

█[asrock][hashcat][0]$

And most importantly, hashcat still complains about "unstable opengl" driver and still segfaults.

I went to the official AMD website to try to download drivers, but they only appear to support RHEL, Ubuntu. Should I try the Ubuntu drivers? Or should the debian firmware-nonfree drivers work with hashcat?

[Image: PaQ7jkafKLzO.png?o=1]

RE: Help with hashcat on Debian: "Unstable OpenCL driver detected!" - atom - 09-01-2025

AMD GPUs on Linux require "AMD Radeon Software for Linux" with "ROCm". Not sure if ROCm drivers support your hardware.

RE: Help with hashcat on Debian: "Unstable OpenCL driver detected!" - erjoalgo - 09-02-2025

I followed the instructions in https://wiki.debian.org/ROCm, but I ran into several errors:

Code:
CMake Error at /usr/share/cmake-3.25/Modules/CMakeDetermineHIPCompiler.cmake:106 (message):

  The ROCm root directory:

   /usr

  does not contain the HIP runtime CMake package, expected at:

   /usr/lib/cmake/hip-lang/hip-lang-config.cmake

I noticed the file is provided by libamdhip64-dev and located at a different path:

Code:
libamdhip64-dev: /usr/lib/x86_64-linux-gnu/cmake/hip-lang/hip-lang-config.cmake

I saw this issue being referenced here: https://gitlab.kitware.com/cmake/cmake/-/merge_requests/8534, I guess in a newer version of cmake than the default in my system.

I tried to work around the issue by creating a symlink:

Code:
sudo ln -s /usr/lib/x86_64-linux-gnu/cmake/hip-lang /usr/lib/cmake/

But obviously this only resulted in another error down the line:

Code:
CMake Error at /usr/share/cmake-3.25/Modules/CMakeDetermineHIPCompiler.cmake:183 (message):

  Failed to find a default HIP architecture.

Call Stack (most recent call first):

  CMakeLists.txt:2 (project)

-- Configuring incomplete, errors occurred!

See also "/home/ealfonso/git/dotfiles/installs/rocm/build/CMakeFiles/CMakeOutput.log".

See also "/home/ealfonso/git/dotfiles/installs/rocm/build/CMakeFiles/CMakeError.log".

I guess I'm going to have to try building the latest cmake, and trying the build again, maybe only to find that ROCm drivers don't support my hardware after all.

RE: Help with hashcat on Debian: "Unstable OpenCL driver detected!" - erjoalgo - 09-02-2025

I manually built the latest cmake 4.1.* and re-ran the build script, it still failed at the "HIP compiler identification" step, but I guess at least it found the ROCm root directory:

Code:
+ rm -rf ./build

+ cmake -version

cmake version 4.1.20250901-gbc0774c

CMake suite maintained and supported by Kitware (kitware.com/cmake).

+ CXX=hipcc

+ cmake -S. -Bbuild

-- The HIP compiler identification is unknown

CMake Error at /usr/local/share/cmake-4.1/Modules/CMakeDetermineHIPCompiler.cmake:348 (message):

  Failed to find a default HIP architecture.

Call Stack (most recent call first):

  CMakeLists.txt:2 (project)

-- Configuring incomplete, errors occurred!

I imagine I'll need to try building the latest rocm instead of copy-pasting some code from a debian wiki page.

RE: Help with hashcat on Debian: "Unstable OpenCL driver detected!" - erjoalgo - 09-02-2025

I tried following the ROCm installation instructions from AMD for Debian:

https://rocm.docs.amd.com/projects/install-on-linux/en/latest/install/quick-start.html

This failed at the step:

Code:
sudo apt install -y rocm

There appear to be some package dependency/version issues.

Code:
-*- mode: compilation; default-directory: "~/git/dotfiles/installs/rocm/" -*-

Compilation started at Mon Sep  1 21:03:52

./adm-rocm.sh

+ set -euo pipefail

+ wget https://repo.radeon.com/amdgpu-install/6.4.3/ubuntu/jammy/amdgpu-install_6.4.60403-1_all.deb

--2025-09-01 21:03:52--  https://repo.radeon.com/amdgpu-install/6.4.3/ubuntu/jammy/amdgpu-install_6.4.60403-1_all.deb

Resolving repo.radeon.com (repo.radeon.com)... 92.123.236.96, 92.123.236.57, 2a02:26f0:9500:27::17d7:2de, ...

Connecting to repo.radeon.com (repo.radeon.com)|92.123.236.96|:443... connected.

HTTP request sent, awaiting response... 200 OK

Length: 17060 (17K) [application/octet-stream]

Saving to: ‘amdgpu-install_6.4.60403-1_all.deb.3’

amdgpu-install_6.4.60403-1_all.d 100%[========================================================>]  16.66K  --.-KB/s    in 0.007s  

2025-09-01 21:03:56 (2.18 MB/s) - ‘amdgpu-install_6.4.60403-1_all.deb.3’ saved [17060/17060]

+ sudo apt install -y ./amdgpu-install_6.4.60403-1_all.deb

Reading package lists... Done

Building dependency tree... Done

Reading state information... Done

Note, selecting 'amdgpu-install' instead of './amdgpu-install_6.4.60403-1_all.deb'

amdgpu-install is already the newest version (6.4.60403-2194681.22.04).

The following packages were automatically installed and are no longer required:

  libpulsedsp pulseaudio-utils

Use 'sudo apt autoremove' to remove them.

0 upgraded, 0 newly installed, 0 to remove and 19 not upgraded.

+ sudo apt update

Hit:1 https://deb.debian.org/debian bookworm InRelease

Hit:2 https://repo.radeon.com/amdgpu/6.4.3/ubuntu jammy InRelease

Hit:3 https://download.docker.com/linux/debian bookworm InRelease

Hit:4 https://deb.debian.org/debian-security bookworm-security InRelease

Hit:5 https://deb.debian.org/debian bookworm-updates InRelease

Hit:6 https://repo.radeon.com/rocm/apt/6.4.3 jammy InRelease

Hit:7 https://developer.download.nvidia.com/compute/cuda/repos/debian12/x86_64  InRelease

Reading package lists... Done

Building dependency tree... Done

Reading state information... Done

19 packages can be upgraded. Run 'apt list --upgradable' to see them.

+ sudo apt install -y python3-setuptools python3-wheel

Reading package lists... Done

Building dependency tree... Done

Reading state information... Done

python3-setuptools is already the newest version (66.1.1-1+deb12u1).

python3-wheel is already the newest version (0.38.4-2).

The following packages were automatically installed and are no longer required:

  libpulsedsp pulseaudio-utils

Use 'sudo apt autoremove' to remove them.

0 upgraded, 0 newly installed, 0 to remove and 19 not upgraded.

+ sudo usermod -a -G render,video ealfonso

+ sudo apt install -y rocm

Reading package lists... Done

Building dependency tree... Done

Reading state information... Done

Some packages could not be installed. This may mean that you have

requested an impossible situation or if you are using the unstable

distribution that some required packages have not yet been created

or been moved out of Incoming.

The following information may help to resolve the situation:

The following packages have unmet dependencies:

 rocm-hip-runtime : Depends: rocminfo (= 1.0.0.60403-128~22.04) but 5.2.3-2 is to be installed

 rocm-hip-runtime-dev : Depends: rocm-device-libs (= 1.0.0.60403-128~22.04) but 5.2.3-2 is to be installed

                        Depends: hipcc (= 1.1.1.60403-128~22.04) but 5.2.3-8 is to be installed

 rocm-openmp-sdk : Depends: rocm-device-libs (= 1.0.0.60403-128~22.04) but 5.2.3-2 is to be installed

 rocm-utils : Depends: rocminfo (= 1.0.0.60403-128~22.04) but 5.2.3-2 is to be installed

E: Unable to correct problems, you have held broken packages.

Compilation exited abnormally with code 100 at Mon Sep  1 21:04:07

I'm not sure if it's worth fighting what seems like an uphill battle. Do I try to manually install the expected versions for each of those dependencies? Or just give up on Debian and install Ubuntu, which AMD appears to support better?

These are the 3 packages that appear to have the wrong versions:

Code:
Depends: rocminfo (= 1.0.0.60403-128~22.04) but 5.2.3-2 is to be installed

Depends: rocm-device-libs (= 1.0.0.60403-128~22.04) but 5.2.3-2 is to be installed

Depends: hipcc (= 1.1.1.60403-128~22.04) but 5.2.3-8 is to be installed

How should I go about fixing the dependency issues, where should I get the latest versions of these pacakges?
Has anyone on debian gotten hashcat to work with AMD hardware? Help.

RE: Help with hashcat on Debian: "Unstable OpenCL driver detected!" - atom - 09-02-2025

Yes, my AMD hardware works fine. I just followed the official instructions. It even works under Docker:

https://github.com/hashcat/hashcat/blob/master/docker/runtime.amd.ubuntu24.release

You might want to have a look into the docker base image rocm/rocm-terminal