[GPU Driver Container Resiliency] Implement userspace-only fast path when kernel modules already loaded with matching config #534
+265
−125
Conversation
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
karthikvetrivel
force-pushed
the
feat/operator-digest-driver-config
branch
2 times, most recently
from
7bf634d to
f959234
Compare
This was referenced Jan 8, 2026
karthikvetrivel
force-pushed
the
feat/operator-digest-driver-config
branch
from
f959234 to
a9bde48
Compare
cdesiniotis
reviewed
Jan 12, 2026
rhel9/nvidia-driver
Outdated
| } | ||
|
|
||
| _userspace_only_install() { | ||
| echo "Detected matching loaded driver & config (${DRIVER_VERSION}); performing userspace-only install" |
Contributor
There was a problem hiding this comment.
nit:
Suggested change
| echo "Detected matching loaded driver & config (${DRIVER_VERSION}); performing userspace-only install" | |
| echo "The NVIDIA driver is already loaded with the desired configuration, performing userspace-only install" |
rhel9/ocp_dtk_entrypoint
Outdated
| # Check if fast path is being used - if so, skip building and signal completion | ||
| if _should_use_fast_path; then | ||
| echo "Fast path detected in DTK container: driver already loaded with matching config, skipping build" | ||
| echo "Signaling driver_built and sleeping forever..." |
Contributor
There was a problem hiding this comment.
Suggested change
| echo "Signaling driver_built and sleeping forever..." | |
| echo "Signaling driver_built to the main container and sleeping forever..." |
ubuntu22.04/nvidia-driver
Outdated
Comment on lines
754
to
758
| if [ -f /sys/module/nvidia/refcnt ] && [ -f /run/nvidia/nvidia-driver.state ]; then | ||
| current_digest=$(_get_current_digest) | ||
| stored_digest=$(_read_stored_digest) | ||
|
|
||
| if [ -n "${current_digest}" ] && [ "${current_digest}" = "${stored_digest}" ]; then |
Contributor
There was a problem hiding this comment.
Similar to the changes you made in rhel9, can we create a helper function to check if the fast-path install path should be taken or not? That will simplify the call-site here.
Member
Author
There was a problem hiding this comment.
Updated the call-site, should look a bit cleaner here now.
karthikvetrivel
force-pushed
the
feat/operator-digest-driver-config
branch
from
a9bde48 to
82123af
Compare
…fig digest matches Signed-off-by: Karthik Vetrivel <[email protected]>
tariq1890
reviewed
Jan 12, 2026
| exit 0 | ||
| } | ||
|
|
||
| _userspace_only_install() { |
Contributor
There was a problem hiding this comment.
I believe this name would imply that it is "userspace only"
Suggested change
| _userspace_only_install() { | |
| _userspace_install() { |
|
|
||
| # Check if fast path should be used (driver already loaded with matching config) | ||
| # Compares current digest from DRIVER_CONFIG_DIGEST env var with stored digest | ||
| _should_use_fast_path() { |
Contributor
There was a problem hiding this comment.
How about this as the method name?
Suggested change
| _should_use_fast_path() { | |
| _should_skip_kernel_module_reload() { |
OR invert the conditional statement and use the name below
Suggested change
| _should_use_fast_path() { | |
| _should_rebuild_kernel_module() { |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 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.
Relevant PRs:
Description
/sys/module/nvidia/refcnt) and compares configuration digest