[Bugfix] Fix DSA crash under breakable piecewise cudagraphs (#49302)
Signed-off-by: Nick Hill <nickhill123@gmail.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
co-authored by
Claude Opus 4.8
parent
61e10f0116
commit
de6ec294ef
@@ -6,7 +6,6 @@ Unit tests for the breakable cudagraph primitives.
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import os
|
||||
import threading
|
||||
from contextlib import nullcontext
|
||||
from unittest.mock import patch
|
||||
@@ -14,7 +13,21 @@ from unittest.mock import patch
|
||||
import pytest
|
||||
import torch
|
||||
|
||||
os.environ["VLLM_USE_BREAKABLE_CUDAGRAPH"] = "1"
|
||||
|
||||
@pytest.fixture(autouse=True)
|
||||
def _enable_breakable_cudagraph(monkeypatch: pytest.MonkeyPatch):
|
||||
"""Enable breakable cudagraphs for this module's tests only.
|
||||
|
||||
eager_break_during_capture reads the env at decoration time, which
|
||||
happens inside the test bodies, so a per-test fixture suffices.
|
||||
monkeypatch restores the env so other test files running in the same
|
||||
pytest process are unaffected (a module-level os.environ assignment
|
||||
used to leak into test_cudagraph_dispatch.py and break it).
|
||||
"""
|
||||
import vllm.envs as envs
|
||||
|
||||
monkeypatch.setenv("VLLM_USE_BREAKABLE_CUDAGRAPH", "1")
|
||||
envs.disable_envs_cache()
|
||||
|
||||
|
||||
def test_piecewise_capture_builds_fresh_metadata_for_both_passes():
|
||||
@@ -86,10 +99,19 @@ def cuda_capture_stream():
|
||||
"""
|
||||
if not torch.cuda.is_available():
|
||||
pytest.skip("CUDA required")
|
||||
from vllm.utils.torch_utils import _current_stream_tls
|
||||
|
||||
prev_stream = getattr(_current_stream_tls, "value", None)
|
||||
stream = torch.cuda.Stream()
|
||||
with torch.cuda.stream(stream):
|
||||
yield stream
|
||||
torch.cuda.current_stream().wait_stream(stream)
|
||||
# Exiting torch.cuda.stream() records the default stream in vllm's
|
||||
# patched set_stream cache. A later CUDAGraphWrapper capture in the same
|
||||
# process would then run on the default stream, which cannot capture
|
||||
# (this broke test_cudagraph_dispatch.py when run in-process after this
|
||||
# file). Restore the pre-fixture value so this module leaves no trace.
|
||||
_current_stream_tls.value = prev_stream
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
Reference in New Issue
Block a user