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Outdated suggestions cannot be applied.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.
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add API for CDI --devices flag in Docker and Podman for mapping GPUs #3290
base: main
Are you sure you want to change the base?
Add API for CDI --devices flag in Docker and Podman for mapping GPUs #3290
Changes from 1 commit
04862545ff36e4830a0dd0d202356e4314db125982d2a13baFile filter
Filter by extension
Conversations
Jump to
There are no files selected for viewing
Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
If we use
docker infowe can also extract information as to whetherexperimentalis enabled.Does changing this function to
assertCDISupported()or something similar make sense here? We could then check the combination of server version and experimental to see whether it is supported.There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Instead of having to remember to call
strings.TrimSpacewhereever the elements ofnode.CDIDevicesare used, can we do that somewhere once?There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
You should be able to use
"github.com/container-orchestrated-devices/container-device-interface/pkg/parser"and callparser.IsQualifiedName(device)here. This is used in docker/cli#4084 and significantly reduces the dependencies pulled in.It should also be fine to allow
podmanand / ordockerto perform this validation though.There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'll give that a try.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Please update to what Evan suggests to reduce the dependencies pulled in.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@BenTheElder What do you think?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@klueska I am using parser.ParseQualifiedName(device) instead of parser.ParseQualifiedName(device). The results are the same except you get the additional error message about why your device string is invalid. I think this is preferable. ParseQualifiedName() is called in IsQualifiedName().
This doesn't change the dependencies at all. There are tests on github.com/container-orchestrated-devices/container-device-interface/pkg/parser pulling in the additional packages. Maybe we should put the tests in a different package? @elezar wdyt?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I can update the tests in that package to use a different import for comparison or to be in the
parser_testpackage if require. As a matter of interest, which package that is equivalent torequirewould be preferred?There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I don't personally have a preference - I haven't found any with fewer dependencies. I was thinking use the
parser_testpackage to keep things very lightweight.There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I have created cncf-tags/container-device-interface#149 with an update to the test package. This should prevent
github.com/stretchr/testify/requirefrom getting pulled in.There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Sorry, a few things came up back in the real world and here with #3277 etc., the dependency discussion wound up forked in https://github.com/kubernetes-sigs/kind/pull/3290/files#r1304873501
I'm going to try to get #3335 after which when this PR is rebased and
go mod tidyit should be clearer what we're actually talking about for the new dependencies.See linked comment for stance on deps.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Currently for example we have a lot of CRI inspired types with in-tree dependency-free code instead, to avoid creating dep hell for our users. That may not be so reasonable here.