[JNI] remove rmm argument to set rw access for fabric handles #17553
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Outdated suggestions cannot be applied.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.
Description
This is a follow up from #17526, where fabric handles can be enabled from RMM. That PR also sets the memory access protection flag (
cudaMemPoolSetAccess
), but I have learned that this second flag is not needed from the owner device. In fact, it causes confusion because the owning device fails to call this function with some of the flags (access none).cudaMemPoolSetAccess
is meant to only be called from peer processes that have imported the pool's handle. In our case, UCX handles this from the peer's side and it does not need to be anywhere in RMM or cuDF.Sorry for the noise. I'd like to get this fix in, and then I am going to fix RMM by removing that API.
Checklist