fixes shmem SIGSEGV#479
Merged
Merged
Conversation
yanboshao
force-pushed
the
yashao/fix_shmem
branch
from
56bee9c to
9e34cc9
Compare
…) ABI to fix shmem SIGSEGV
yanboshao
force-pushed
the
yashao/fix_shmem
branch
from
9e34cc9 to
3efbbf8
Compare
Collaborator
|
LGTM. merged @yanboshao |
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
2 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Motivation
FlyDSL's <#fly.explicit_module> declared its GPU-module loader helpers as single-arg void(void*) that hand-unpacked the outputs, but MLIR's ExecutionEngine already wraps every JIT entry in its own packed-args trampoline, so two layers of "packing" stacked up. Python packed the args once, the trampoline dereferenced once more, the real loader received NULL pointers and store-to-NULL segfaulted on the very first GPU module init.
Technical Details
Change init/load's LLVM signature to a plain N-arg void(void** module_out, int32_t* err) so MLIR's wrapper sees a normal one-slot-per-arg function. Rewrite _pack_ciface_args to the canonical packed-args layout — one holder cell per argument — so the C++ and Python sides finally agree on a single, MLIR-standard ABI.
Test Plan
Test Result
Submission Checklist