Normalize NVIDIA library paths to /usr/lib/#919
Open
maherthomsi wants to merge 2 commits into
Open
Conversation
maherthomsi
changed the title
maherthomsi
force-pushed
the
nvidia-normalization
branch
2 times, most recently
from
cf2022c to
7614bfc
Compare
Contributor
Author
|
Added |
piyush-jena
reviewed
May 7, 2026
piyush-jena
reviewed
May 7, 2026
piyush-jena
reviewed
May 7, 2026
maherthomsi
force-pushed
the
nvidia-normalization
branch
2 times, most recently
from
64e28c4 to
d3f12f0
Compare
maherthomsi
force-pushed
the
nvidia-normalization
branch
from
d3f12f0 to
b16ac81
Compare
piyush-jena
approved these changes
May 18, 2026
arnaldo2792
reviewed
May 21, 2026
| {{/if}} | ||
| deviceIDStrategy: {{default "index" settings.kubelet-device-plugins.nvidia.device-id-strategy}} | ||
| containerDriverRoot: "/" | ||
| containerDriverRoot: "/x86_64-bottlerocket-linux-gnu/sys-root" |
Contributor
There was a problem hiding this comment.
This is not correct, you are hardcoding the path for both arches. If this is the way to go from now on, you will have to make this file an "input" templated file. Similar to what was done in the files you removed from the kernel kit.
maherthomsi
force-pushed
the
nvidia-normalization
branch
from
b16ac81 to
270b737
Compare
maherthomsi
force-pushed
the
nvidia-normalization
branch
from
270b737 to
8be6943
Compare
maherthomsi
force-pushed
the
nvidia-normalization
branch
from
8be6943 to
d73ab01
Compare
Add --additional-symlinks flag to nvidia-ctk cdi generate that creates symlinks in a specified directory pointing to each discovered library. Configure generate-cdi-specs.service with: - --driver-root /x86_64-bottlerocket-linux-gnu/sys-root - --dev-root / - --additional-symlinks /usr/lib/nvidia/tesla This ensures libraries appear at /usr/lib/ in containers with backwards-compat symlinks at /usr/lib/nvidia/tesla/. Signed-off-by: Maher Homsi <maherhom@amazon.com>
maherthomsi
force-pushed
the
nvidia-normalization
branch
from
d73ab01 to
11e6c8d
Compare
mgsharm
approved these changes
Jun 8, 2026
vigh-m
approved these changes
Jun 9, 2026
…rmalization Set containerDriverRoot to the Bottlerocket sysroot path so the device plugin discovers libraries correctly and generates CDI specs with normalized /usr/lib/ container paths. Add --additional-symlinks support patches (1002, 1003) to the device plugin vendored nvidia-container-toolkit code. Signed-off-by: Maher Homsi <maherhom@amazon.com>
maherthomsi
force-pushed
the
nvidia-normalization
branch
from
11e6c8d to
f5e65aa
Compare
Contributor
Author
|
Force-push: Added --cdi-enabled-hooks create-lib-symlinks to exec-start template |
arnaldo2792
reviewed
Jun 16, 2026
| install -d %{buildroot}%{_cross_unitdir}/nvidia-k8s-device-plugin.service.d | ||
| install -d %{buildroot}%{_cross_unitdir}/nvidia-mps-control-daemon.service.d | ||
| install -D -m 0644 %{S:2} %{buildroot}%{_cross_templatedir}/nvidia-k8s-device-plugin-conf | ||
| sed -e 's|__PREFIX__|/%{_target_cpu}-bottlerocket-linux-gnu/sys-root|g' %{S:2} > nvidia-k8s-device-plugin-conf |
Contributor
There was a problem hiding this comment.
%{__target_cpu} is not the correct macro, you should use %{_cross_arch}
Contributor
|
Nit: Please call out in the future variant releases, we will remove this symlink. This is a change for existing variants 👍 |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
6 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Merge with bottlerocket-os/bottlerocket-kernel-kit#425
Description of changes:
Normalize NVIDIA library paths in containers so libraries appear at
/usr/lib/instead of the Bottlerocket cross-compilation sysroot path (/x86_64-bottlerocket-linux-gnu/sys-root/usr/lib/).--additional-symlinksflag patch tonvidia-ctk cdi generatethat creates backwards-compatibility symlinks in a specified directory pointing to each discovered library. Configuregenerate-cdi-specs.servicewith--additional-symlinks /usr/lib/nvidia/tesla.containerDriverRootto the Bottlerocket sysroot path so the device plugin discovers libraries correctly and generates CDI specs with normalized/usr/lib/container paths. EnableCreateLibSymlinksHookvia patch 1002 so the device plugin's CDI spec includes backwards-compat symlinks. Add--cdi-enabled-hooks create-lib-symlinksto the exec-start template to pass the hook enablement through the systemd drop-in. FixcontainerDriverRootmacro expansion from%{_cross_sysroot}(undefined) to/%{_target_cpu}-bottlerocket-linux-gnu/sys-root.Result: containers see libraries at
/usr/lib/libcuda.so.580.159.03with backwards-compat symlinks at/usr/lib/nvidia/tesla/libcuda.so.580.159.03pointing to/usr/lib/libcuda.so.580.159.03.Note: Backwards-compat symlinks are supported in
cdi-crimode (the default device-list-strategy). The legacyvolume-mountsmode does not use CDI hooks and therefore does not create symlinks in containers.Testing done:
containerDriverRootresolves to/x86_64-bottlerocket-linux-gnu/sys-rootnvidia-smiworks in container/usr/lib/inside container/usr/lib/nvidia/tesla/point to/usr/lib/generate-cdi-specs.servicepasses on boot/var/run/cdi/includes symlink hooksldconfigresolveslibcuda.so.1on hostld.so.conf.dentry needed (libs in standard path)nvidia-smiworks on host (dlopen via both paths)--cdi-enabled-hooks, symlinks do not appear in containersTerms of contribution:
By submitting this pull request, I agree that this contribution is dual-licensed under the terms of both the Apache License, version 2.0, and the MIT license.