socket/netlink: implement NETLINK_RDMA to report copy-on-fork - #14042
Open
ayushr2 wants to merge 1 commit into
Open
socket/netlink: implement NETLINK_RDMA to report copy-on-fork#14042ayushr2 wants to merge 1 commit into
ayushr2 wants to merge 1 commit into
Conversation
libfabric/rdma-core learn whether the host kernel does copy-on-fork for pinned pages by querying NETLINK_RDMA's RDMA_NLDEV_CMD_SYS_GET for the RDMA_NLDEV_SYS_ATTR_COPY_ON_FORK attribute. When it is unset, providers such as AWS EFA arm an abort-on-fork handler and SIGABRT the process on the first fork() that has a registered MR (e.g. Megatron's async checkpoint save). gVisor did not implement NETLINK_RDMA, so the query failed and every such fork aborted. Add a NETLINK_RDMA provider that answers RDMA_NLDEV_CMD_SYS_GET with COPY_ON_FORK=1, mirroring drivers/infiniband/core/nldev.c. This answer is truthful because the sentry implements copy-on-fork for pinned pages (PR google#14031): rdmaproxy-pinned DMA pages stay stable across fork() while the host driver keeps DMAing the original.
ayushr2
marked this pull request as ready for review
August 7, 2026 21:14
Collaborator
Author
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
libfabric/rdma-core learn whether the host kernel does copy-on-fork for
pinned pages by querying NETLINK_RDMA's RDMA_NLDEV_CMD_SYS_GET for the
RDMA_NLDEV_SYS_ATTR_COPY_ON_FORK attribute. When it is unset, providers such
as AWS EFA arm an abort-on-fork handler and SIGABRT the process on the first
fork() that has a registered MR (e.g. Megatron's async checkpoint save).
gVisor did not implement NETLINK_RDMA, so the query failed and every such
fork aborted. Add a NETLINK_RDMA provider that answers RDMA_NLDEV_CMD_SYS_GET
with COPY_ON_FORK=1, mirroring drivers/infiniband/core/nldev.c.
This answer is truthful because the sentry implements copy-on-fork for pinned
pages (#14031): rdmaproxy-pinned DMA pages stay stable across fork() while the
host driver keeps DMAing the original.
Depends on #14031; should be merged after it.