Repository navigation
Conversation
resolveFn looked a registered handler's name up across every language, so a C table could link to a same-named Python or Rust function, and such a function could take the place of the real C handler.
colbymchenry
pushed a commit
that referenced
this pull request
Oct 27, 2026
…ions resolveFn in the C function-pointer pass looked a registered handler's name up across every language, filtered only to function and method kinds. In a repository that also holds Rust or Python, a C table could link to a same-named Rust or Python function, and when that function sorted first it also took the place of the real C handler. The fix keeps candidates in C and C++ files only, using the pass's own C_CPP_EXT, the file set it scans for tables and dispatch. Headers the language detector labels Objective-C still count. Objective-C `.m` and `.mm` files stay out, like they are out of the scan: a table in a `.m` file is never read, and allowing them would let an Objective-C method named `init` or `read` stand in for a C handler of that name. On GLFW, which mixes C and Objective-C, main links 354 table entries, none into a .m file; its one C table naming a .m function (_glfwConnectCocoa) is an anonymous-struct table the pass doesn't bridge on main either. Validation, main b6379e6 vs this: - c-fnptr-synthesizer.test.ts: the new case fails on main (impl_read links to aaa/gen.py) and passes here; 19/19. - Linux d135ab45 subtree (rust/, drivers/android, drivers/base, drivers/net/dsa, fs, tools/net, include/linux; 6,580 files): byte-identical, 12,244 fn-pointer-dispatch edges, all into C or C++ files on both arms. Main links none of this tree into Rust or Python, so the PR's 193 (a whole-kernel index on its fork) don't show here; a subtree holding its three example sites (pmcraid's `->reset_cmd = cmd`, afs's `->update_time = new`, map.c's `p->lock = lock`) has none either, since a plain assignment of a variable doesn't register on main. - C and C++ controls byte-identical (dump-graph, every table): redis 7f440372 (863 fn-pointer-dispatch edges), GLFW 92dcf4ce (354), jq 49fbca12 and leveldb 7ee830d0. - The resolver change is TS-side; the kernel ports only the per-file sweep, so both engines take the same edges. kernel-parity on the kernel's drivers/base, drivers/net/dsa and drivers/android (279 C files): 240 at parity, 0 diffs, 39 deferred to wasm by policy. Ours on top: the CHANGELOG bullet moved from the released 1.6.1 section to Unreleased and reworded. Original PR #1989 by @bompus. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Owner
|
Thanks @bompus! Your C/C++-only filter landed unchanged in #2935, with you as the commit author; we only moved the CHANGELOG line from the released 1.6.1 section into Unreleased. Pure C and C++ graphs come out byte-identical (redis, GLFW, jq, leveldb, and a 6,580-file Linux subtree), so it costs nothing where there's no other language, and your test pins the mixed-language case. We kept Objective-C |
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
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.
The C function-pointer pass resolves a registered handler by name with
ctx.getNodesByName(name), filtered only to function and method kinds. In a repository with other languages, a C table can therefore link to a Python or Rust function that shares the handler's name. When that function sorts first, it also takes the place of the real C handler.On a Linux kernel index, 193
fn-pointer-dispatchedges pointed into Rust (171) or Python (22). Examples:touch_atime→newindrivers/android/binder/allocation.rskobj_lookup→lockinrust/kernel/gpu/buddy.rssja1105_static_config_upload→cmdintools/net/ynl/pyynl/lib/ynl.pyThe fix limits candidates to functions in C/C++ files, using the pass's existing
C_CPP_EXT. That is the file set the pass scans, so headers the language detector labels Objective-C still count. We measured this change in our fork on the same index: the 193 cross-language edges go away and 133 C targets they had displaced come back.Checks:
c-fnptr-synthesizer.test.ts: a C table registeringimpl_read(defined inimpl.cand also in a Python script) andgen_probe(defined only in Python). The table links toimpl.c, and nothing links to Python. Onmainit linksimpl_readto the Python file.tsc --noEmitis clean.