Revision bd8350b3475b0a7c6feace151ebe2752820efa8a authored by Keno Fischer on 19 July 2023, 19:36:59 UTC, committed by GitHub on 19 July 2023, 19:36:59 UTC
In #48919, the tid selection logic inside `enq_task` gained a `!GC.in_finalizer()` condition. However, this made it possible for `workqueue_at` to be reached with `tid==0`, which would attempt and out-of-bounds write under `@inbounds`, corrupting memory. This was not caught in the test suite despite `--check-bounds=yes`, because our `--check-bounds=yes` is currently best effort. That would be fixed by #50239, which exposed this bug. This PR attempts to fix this by marking any tasks launched inside a finalizer as not sticky. Finalizers don't have any thread they run on semantically, so i don't think there's a meaningful sense in which tasks launched inside finalizers could be sticky.
1 parent 077efd4
| File | Mode | Size |
|---|---|---|
| .devcontainer | ||
| .github | ||
| base | ||
| cli | ||
| contrib | ||
| deps | ||
| doc | ||
| etc | ||
| src | ||
| stdlib | ||
| test | ||
| .buildkite-external-version | -rw-r--r-- | 5 bytes |
| .clang-format | -rw-r--r-- | 3.3 KB |
| .clangd | -rw-r--r-- | 114 bytes |
| .codecov.yml | -rw-r--r-- | 52 bytes |
| .git-blame-ignore-revs | -rw-r--r-- | 371 bytes |
| .gitattributes | -rw-r--r-- | 65 bytes |
| .gitignore | -rw-r--r-- | 523 bytes |
| .mailmap | -rw-r--r-- | 12.7 KB |
| CITATION.bib | -rw-r--r-- | 513 bytes |
| CITATION.cff | -rw-r--r-- | 940 bytes |
| CONTRIBUTING.md | -rw-r--r-- | 23.4 KB |
| HISTORY.md | -rw-r--r-- | 370.0 KB |
| LICENSE.md | -rw-r--r-- | 1.3 KB |
| Make.inc | -rw-r--r-- | 55.3 KB |
| Makefile | -rw-r--r-- | 30.1 KB |
| NEWS.md | -rw-r--r-- | 1.1 KB |
| README.md | -rw-r--r-- | 7.3 KB |
| THIRDPARTY.md | -rw-r--r-- | 3.8 KB |
| VERSION | -rw-r--r-- | 11 bytes |
| julia.spdx.json | -rw-r--r-- | 35.8 KB |
| pkgimage.mk | -rw-r--r-- | 6.0 KB |
| sysimage.mk | -rw-r--r-- | 4.3 KB |
Computing file changes ...
