perf(runtime): let the GC header pick the side-registry probe on dynamic dispatch (#7850) #1732

Workflow file for this run

.github/workflows/gc-root-dominance.yml at a18286c

	name: GC Root Dominance

	# Static gate for the invariant a GC-managed value's root store must DOMINATE
	# every subsequent site that can trigger a collection (#7154).
	#
	# Two bugs of this class shipped before there was an instrument for it. #7184:
	# the root store was emitted but its shadow-slot index fell outside the pushed
	# frame, so `js_shadow_slot_bind` bounds-checked it into a silent no-op. #7192:
	# the store was emitted in-frame but AFTER a call that allocates. Both present
	# identically — a rooted slot holding a dangling pointer, surfacing cycles
	# later as "TypeError: value is not a function" — and neither is visible to any
	# runtime GC probe, because at the moment of the collection there is nothing
	# for the collector to find. A static pass over the emitted IR is the only
	# instrument that sees them before they crash, which is why this is a gate and
	# not a benchmark.
	#
	# THIS JOB IS DESIGNED TO BE ABLE TO FAIL, and is checked against all four ways
	# a gate can be unable to (CLAUDE.md):
	#
	# 1. no `continue-on-error`, no `\|\| true`, no pipe between the checker and
	# the shell's exit status;
	# 2. NOT yet in branch protection's required contexts. This was deliberate at
	# first — a new gate has never been green, so promoting it immediately
	# blocks every open PR — but the second step was never taken, and in the
	# meantime the job WAS red on `main` and blocked nothing: the five
	# ClassExprFresh hits now tracked in #7211 have been reported on every run
	# since #7198 and went unread. That is hazard 2 doing exactly what the
	# corollary in CLAUDE.md warns about.
	#
	# **ACTION REQUIRED, and it is not something this workflow can do to
	# itself**: a repo admin must add `gc-root-dominance` to branch
	# protection's required contexts. Until that happens this file is
	# documentation, not a gate.
	#
	# Both conditions #7198 named are met as of #7236: the dominance check is
	# green with an EMPTY allowlist, and `--unrooted-allocas --moving-only` --
	# which was 98 before #7235 and 2 after -- now reads 0 and is a step
	# below. Promote after this job's first green run on `main` WITH that
	# step, not before: a gate that has never been green in its current shape
	# blocks every open PR the day it becomes required, which is the corollary
	# that produced this whole paragraph.
	# See docs/src/internals/gc-rooting-invariant.md, "Promoting this gate".
	# 3. `concurrency` cancels pull-request runs only, never `main` runs;
	# 4. the subject is ASSERTED live, not assumed, at three levels.
	# `--self-test` proves the checker still reports a planted violation in
	# hand-written IR and still clears the control. `--seeded-violations`
	# goes further and plants collection points into the REAL corpus,
	# requiring every one to be reported -- that is what catches the case
	# where perry's emitted IR drifts to a shape the parser can no longer
	# read, which frozen fixtures cannot detect. And `--min-files` /
	# `--min-binds` / `--min-funcs` refuse a clean verdict over a corpus with
	# too few modules, root stores or functions to have exercised anything.
	# An empty `.perry-trace/llvm` is a routine outcome of a failed compile,
	# so "0 violations" over 0 files must be an error rather than a pass.
	#
	# Known-remaining violations live in scripts/gc_root_dominance_allowlist.json,
	# one named entry each with an issue and a written justification -- NOT a
	# numeric threshold, which cannot tell a new violation from an old one. An
	# entry that matches nothing fails the build, so a fixed bug's entry must be
	# deleted rather than left to widen coverage later.

	on:
	pull_request:
	push:
	branches: [main]
	workflow_dispatch:

	permissions:
	contents: read

	concurrency:
	# One group per main COMMIT, cancelling PR runs only. A shared group starves
	# `main` on a deep runner queue, and a gate that is always cancelled never
	# fails. `cancel-in-progress: false` is not enough on its own: GitHub allows
	# at most one PENDING run per group and cancels the previously pending one
	# when a new run enters, regardless of that setting (#7205 — measured on
	# gc-ratchet, whose shape this file copied). Keying push runs on the SHA gives
	# every merged commit a group of its own, so no two main runs can contend.
	# Same reasoning as gc-ratchet.yml, which carries the full writeup.
	group: gc-root-dominance-${{ github.event_name }}-${{ github.event_name == 'push' && github.sha \|\| github.ref }}
	cancel-in-progress: ${{ github.event_name == 'pull_request' }}

	env:
	CARGO_TERM_COLOR: always
	MACOSX_DEPLOYMENT_TARGET: "13.0"

	jobs:
	gc-root-dominance:
	runs-on: macos-14
	timeout-minutes: 90
	steps:
	- uses: actions/checkout@v7
	with:
	persist-credentials: false

	# Fast structural failure first: prove the checker can still fail before
	# spending a compiler build on it. This is the arm that would have caught
	# `PERRY_GC_FORCE_EVACUATE` being inert for every test that "exercised"
	# it (#6942/#6946).
	- name: Checker self-test (can this gate still fail?)
	run: python3 scripts/gc_root_dominance_check.py --self-test

	# The checker only reports a stale register when it recognises what
	# MATERIALIZED the value, and that recognition is one regex. An
	# alternative in it that matches no real symbol reads as coverage and is
	# not -- the gate-can't-fail pattern in regex form. It has shipped nine
	# times across two rounds: four (`regexp_alloc\w`, `promise_alloc\w`,
	# `bigint_alloc\w`, `typed_array_alloc\w`) found only because
	# `js_regexp_new` cost #7154 a whole investigation round, and five more
	# introduced by the change that removed those four. A prose audit does not
	# survive its own next edit, so it is a gate now: every alternative must
	# match at least one `extern "C" fn js_*` the runtime actually exports.
	#
	# Static and instant — no toolchain, no corpus — so it runs before the
	# build with the self-test.
	- name: ALLOC_RE alternatives must match real runtime symbols
	run: python3 scripts/gc_root_dominance_check.py --audit-alloc-re

	# The same hazard on the other axis, and a sharper one. `ALLOC_RE`
	# decides whether a register has a heap-value source; `POLL_CAPABLE_RUNTIME`
	# decides whether the window around it is MOVING -- which is the arm the
	# two `--moving-only` steps below actually gate on. An entry naming a
	# symbol that does not exist can never classify a window, so it reads as
	# coverage of an operation the gate is in fact blind to, and unlike a dead
	# regex alternative it survives review because the misspelling is a
	# PLAUSIBLE name for a real operation.
	#
	# Ten of twenty-eight entries were phantoms when this step was added,
	# including four separate spellings of "call a JS closure" (`js_call_closure`,
	# `js_invoke_closure`, `js_function_call`, `js_apply_function`) where the
	# real symbols are `js_closure_callN`. The consequence was measured at
	# zod's `clone`: the from-space protector faulted deterministically on a
	# window that `--stale-registers` classified `MOVING: no`, so the
	# `--moving-only` arm could not see it.
	#
	# Static and instant, like the audit above, so it runs before the build.
	- name: POLL_CAPABLE_RUNTIME entries must name real runtime symbols
	run: python3 scripts/gc_root_dominance_check.py --audit-poll-capable

	# ★ The THIRD audit, and the direction the other two structurally cannot
	# look (#7616). Both of those hunt a NAME WITH NO REFERENT: a dead
	# ALLOC_RE alternative, a phantom POLL_CAPABLE_RUNTIME entry. Nothing
	# hunted a REFERENT WITH NO NAME -- a real allocating symbol that is
	# missing from POLL_CAPABLE_RUNTIME -- and a missing entry suppresses
	# findings exactly as silently as a phantom one.
	#
	# #7453 is the worked example and it cost a shipped use-after-free.
	# `new URL(input, base)` held a raw `*mut StringHeader` across the
	# lowering of `base`; the fix added `url_coerce_string` to ALLOC_RE,
	# said in its own comment "that gap is why the checker did not flag
	# #7453", and stopped one list short. Re-planting that exact code and
	# running EVERY mode this workflow has (#7616's measurement):
	#
	# --moving-only (dominance) 0 clean -> 0 sabotaged
	# --unrooted-allocas --moving-only 0 -> 0
	# --stale-registers --moving-only 2 -> 2
	# --statepoints --moving-only 2 -> 2 <- the mode
	# that reads the lowering that SHIPS, added in #7663
	# --stale-registers (unfiltered) 24 -> 35
	# --statepoints (unfiltered) 15 -> 21
	#
	# Every GATED arm blind, both UNFILTERED arms not. The single missing
	# list entry was all of it: adding `js_url_coerce_string` takes the
	# sabotaged arms to 13 and 8 and leaves the clean arms at 2 and 2.
	#
	# Static and instant, like the two above, so it runs before the build.
	- name: ALLOC_RE symbols that reach a poll-capable one must be listed
	run: python3 scripts/gc_root_dominance_check.py --audit-poll-reach

	# `--unrooted-allocas` exempts two heap-value sources whose objects can
	# neither move nor be reclaimed (#7210): the old-arena class-keys array,
	# and `js_box_alloc*`'s Box, which is not a GC-heap allocation at all.
	# Each exemption SUPPRESSES reports, and each is only as sound as
	# premises that live in the runtime crates and can change without anyone
	# touching this checker — the class-keys array staying `_longlived`,
	# old-page defrag staying off by default, the keys global staying a
	# registered root, and boxes staying `std::alloc::alloc` with no free
	# path.
	#
	# An exemption whose premise has quietly lapsed is strictly WORSE than no
	# exemption: it reads as a triaged false positive and is a live hazard.
	# So the premises are a gate rather than a paragraph. Static and instant,
	# like the audit above, so it runs before the build.
	- name: Immovable-source exemptions must still hold
	run: python3 scripts/gc_root_dominance_check.py --audit-immovable-sources

	- name: Install Rust toolchain
	uses: dtolnay/rust-toolchain@stable
	- uses: ./.github/actions/setup-llvm22

	# ★ The dependency-scale corpus needs the dependency.
	#
	# `zod` is this repo's own package.json devDependency, pinned by
	# package-lock.json and governed by the same soak window as everything
	# else in that file -- not a fixture invented for this job.
	# `--ignore-scripts` because nothing here needs a lifecycle script to
	# run, and a corpus generator is a bad place to execute one.
	- uses: actions/setup-node@v6
	with:
	node-version-file: .node-version
	- name: Install the npm dependencies the dep corpus compiles
	run: npm ci --ignore-scripts --no-audit --no-fund

	- name: Cache cargo
	uses: actions/cache@v6
	with:
	path: \|
	~/.cargo/registry
	~/.cargo/git
	target
	key: ${{ runner.os }}-cargo-gcdom-${{ hashFiles('**/Cargo.lock') }}
	restore-keys: ${{ runner.os }}-cargo-

	- name: Build perry and the runtime archives
	run: \|
	set -euo pipefail
	# perry-runtime and perry-stdlib are rlib-only; the .a files come from
	# the -static wrapper crates. The package set is fixed so cargo
	# feature unification matches every other job that builds the
	# compiler.
	cargo build --release -p perry -p perry-runtime-static -p perry-stdlib-static
	for artifact in perry libperry_runtime.a libperry_stdlib.a; do
	test -s "target/release/$artifact" \
	\|\| { echo "::error::target/release/$artifact was not produced"; exit 1; }
	done

	# The source list and the env knobs live in the script, not here, so that
	# reproducing a CI failure is one command rather than a re-read of this
	# YAML. A retyped invocation that drops PERRY_GC_MOVING_LOOP_POLLS
	# produces IR in which the bug is not expressible at all, and the local
	# run then "cannot reproduce" a real finding.
	- name: Emit the IR corpus
	run: ./scripts/gc_root_dominance_corpus.sh ir-corpus

	- name: Check root-store dominance
	run: \|
	set -euo pipefail
	# No pipe: the checker's own exit status is the job's.
	#
	# The floors are asserted, not hoped for. On the corpus as of this
	# commit the run reports ~1993 functions / 117 modules / 2501 root
	# stores, so these sit below that with room for churn and well above
	# "something compiled". Raise them when the corpus grows; never lower
	# one to make a run pass -- a shrinking corpus is the finding.
	#
	# --seeded-violations plants 40 collection points into this very IR
	# and requires all 40 to be reported. That is the arm that fails if
	# the checker has stopped understanding perry's output, which is the
	# only way a green verdict here could be a lie.
	python3 scripts/gc_root_dominance_check.py ir-corpus \
	--moving-only \
	--min-files 90 --min-binds 1500 --min-funcs 1200 \
	--allowlist scripts/gc_root_dominance_allowlist.json \
	--seeded-violations 40 \
	-v

	# ★ The OTHER mode, and until #7236 it had never been at zero.
	#
	# `--unrooted-allocas` asks a different question from the step above: not
	# "is the root store late" but "is there a root store AT ALL". #7235 split
	# its heap-source predicate by movability and got the count from 98 to 2,
	# and both residuals were one bug -- `Type::Symbol` classified as an
	# immediate, so a `Symbol` local got no shadow slot. That is now fixed and
	# the corpus reads 0, which is the condition #7198 named for promoting
	# this job to a required context.
	#
	# A number that is the acceptance criterion for a promotion and is checked
	# by nothing will regress silently -- CLAUDE.md hazard 4, applied to the
	# measurement rather than to the job. So it is a step. It shares the
	# corpus, the exemptions and the allowlist file with the step above, and
	# carries the same liveness floors: "0 violations" over a corpus that did
	# not compile means nothing.
	#
	# If this arm ever goes red for a hazard that is real but deliberately
	# deferred, the answer is an entry in gc_root_dominance_allowlist.json
	# with an issue number, NOT deleting the step -- rule 1 in that file
	# (an entry matching nothing fails the build) is what makes the tolerance
	# temporary.
	- name: Check that every GC value in an alloca has a root store
	run: \|
	set -euo pipefail
	python3 scripts/gc_root_dominance_check.py ir-corpus \
	--unrooted-allocas \
	--moving-only \
	--min-files 90 --min-binds 1500 --min-funcs 1200 \
	--allowlist scripts/gc_root_dominance_allowlist.json \
	-v

	# ★★ The DEPENDENCY-SCALE corpus (#7280).
	#
	# Everything above this line runs over ~124 hand-written `test-files/`
	# sources. That corpus read ZERO in both gated modes while twenty lines of
	# stock `zod` faulted deterministically under the from-space protector,
	# and #7280 records the gap in one sentence: 25 curated files pass while
	# 20 lines of stock zod fail.
	#
	# It is not a size problem, it is a distribution problem. Measured with
	# `--stale-registers --moving-only` on the same compiler:
	#
	# curated (124 sources, 144 modules): 116 stale uses, and what
	# dominates is property-GET helper windows and js_number_coerce
	# dependency-scale (81 modules, 62 MB): 370 stale uses, and what
	# dominates is js_object_assign_one (object spread, 137) and
	# js_new_function_construct (102) -- populations the curated corpus
	# produces 12 and 1 of
	#
	# Nothing is sampled away: all 81 modules and all 62 MB are checked. The
	# cost is ~8s to emit and ~4s for the two gated arms below, because those
	# arms are linear in instruction count. The `--stale-registers` ratchet is
	# the expensive one (~5 min) and says so where it runs.
	- name: Emit the dependency-scale IR corpus
	run: ./scripts/gc_root_dominance_dep_corpus.sh ir-corpus-dep

	- name: Check root-store dominance (dependency-scale)
	run: \|
	set -euo pipefail
	# Floors from the corpus as of this commit (81 modules, ~12900
	# functions, ~7700 root stores), set below that with room for the
	# dependency's own churn. `zod` growing is fine; `zod` no longer
	# compiling natively is the finding, and these are what make it one.
	python3 scripts/gc_root_dominance_check.py ir-corpus-dep \
	--moving-only \
	--min-files 60 --min-binds 4000 --min-funcs 6000 \
	--allowlist scripts/gc_root_dominance_allowlist.json \
	--seeded-violations 40 \
	-v

	- name: Check that every GC value in an alloca has a root store (dependency-scale)
	run: \|
	set -euo pipefail
	python3 scripts/gc_root_dominance_check.py ir-corpus-dep \
	--unrooted-allocas \
	--moving-only \
	--min-files 60 --min-binds 4000 --min-funcs 6000 \
	--allowlist scripts/gc_root_dominance_allowlist.json \
	-v

	# ★ The stale-register RATCHET, on both corpora.
	#
	# `--stale-registers` asks the third question: not "is the root store
	# late" and not "is there a root store at all", but "is a register holding
	# a rooted value used below a collection point". It is the mode that found
	# #7206's two bugs and the mode #7280's zod fault lives in, and until now
	# it ran only by hand -- so its number could move in either direction
	# between one investigation and the next with nothing to say so.
	#
	# It is a BUDGET rather than an allowlist because the residual is a
	# population, not a list of triaged sites: the remaining uses are the ones
	# whose slot the program itself reassigns inside the window, which need a
	# temp root rather than a re-read (see crate::root_reload). A budget can
	# only be lowered, and lowering it is the ratchet.
	#
	# This step is minutes, not seconds -- the scan is superlinear in
	# instruction count and the dependency corpus is 62 MB. That is the price
	# of checking the population that actually breaks.
	- name: Stale-register budget (curated)
	run: \|
	set -euo pipefail
	python3 scripts/gc_root_dominance_check.py ir-corpus \
	--stale-registers --moving-only \
	--min-files 90 --min-binds 1500 --min-funcs 1200 \
	--max-stale 39

	# 86 -> 104 at #7616: widening POLL_CAPABLE_RUNTIME by 77 symbols makes
	# windows MOVING that `--moving-only` previously dropped. Still inside
	# the pinned 118, so the budget is deliberately NOT raised — a ratchet
	# you loosen every time it gets closer is not a ratchet.
	- name: Stale-register budget (dependency-scale)
	run: \|
	set -euo pipefail
	python3 scripts/gc_root_dominance_check.py ir-corpus-dep \
	--stale-registers --moving-only \
	--min-files 60 --min-binds 4000 --min-funcs 6000 \
	--max-stale 118

	- name: Upload the IR corpus on failure
	if: failure()
	uses: actions/upload-artifact@v7
	with:
	name: gc-root-dominance-ir
	path: \|
	ir-corpus
	ir-corpus-dep
	retention-days: 7

	# ★★★ THE LOWERING THAT ACTUALLY SHIPS (#7663).
	#
	# Everything in the job above compiles its corpus under `PERRY_RS4GC=0` — the
	# SHADOW-STACK lowering — and since #7370 that is not the default on any
	# target whose frames the runtime can walk. The reason was sound and is
	# stated inline in `gc_root_dominance_corpus.sh`: the checker anchored on
	# `@js_shadow_slot_bind`, the native lowering emits zero of them, and
	# `--min-binds` therefore failed the job. So the gate was pointed at the only
	# lowering it could read, and `docs/src/internals/gc-rooting-invariant.md`
	# records the consequence as a blind spot: **a green `gc-root-dominance` was
	# evidence about a lowering that does not ship.**
	#
	# This job closes that. It is a SEPARATE job, not more steps in the one
	# above, for three reasons: it needs a differently-generated corpus, its
	# floors are about safepoints rather than root stores, and — the one that
	# matters — a separate job is a separate branch-protection context, so this
	# can be promoted to required on its own schedule without dragging the
	# shadow arms with it. The price is a second compiler build; `needs:` would
	# avoid it and was rejected, because it would make this job unrunnable
	# whenever the shadow job is red — precisely when someone wants to know
	# whether the OTHER lowering is affected.
	#
	# NOT REQUIRED, deliberately, and this is the second step nobody took last
	# time. A gate that has never been green blocks every open PR the day it
	# becomes required (CLAUDE.md's corollary, which this repo has already paid
	# for once — see hazard 2 in the header of this file). Let it run on `main`
	# first, then promote. Promotion is an admin action and is tracked separately
	# from the PR that adds this.
	gc-root-dominance-statepoints:
	runs-on: macos-14
	timeout-minutes: 90
	steps:
	- uses: actions/checkout@v7
	with:
	persist-credentials: false

	# Same fast structural failure first. `--self-test` now covers BOTH
	# lowerings: the statepoint half asserts the bundle parser reads the
	# wrapped callee and a non-empty live set (the two mistakes the Rust
	# reader of this construct already made and caught), that the two verdict
	# classes are distinguished by the bundle alone, that the relocated-value
	# control reports zero, and that `--min-statepoints` refuses the shadow
	# corpus while `--min-binds` refuses this one.
	- name: Checker self-test (can this gate still fail?)
	run: python3 scripts/gc_root_dominance_check.py --self-test

	# The FOUR static audits are shared vocabulary, not shadow-mode
	# details: `--statepoints` classifies its sources with the same ALLOC_RE,
	# decides MOVING with the same POLL_CAPABLE_RUNTIME, and honours the same
	# #7210 exemptions. A phantom entry in any of them is a hole in THIS gate
	# too, so all four run here as well rather than being inherited by
	# assumption from a job that may not have run.
	#
	# `--audit-poll-reach` matters most HERE, because #7616's measurement was
	# taken in this job's mode: the re-planted #7453 shape reported 2 clean
	# and 2 sabotaged under `--statepoints --moving-only`, i.e. the newest
	# mode was as blind as the three it was added to cover for, and for the
	# same one-line reason. The full census is in the shadow job above.
	- name: ALLOC_RE alternatives must match real runtime symbols
	run: python3 scripts/gc_root_dominance_check.py --audit-alloc-re
	- name: POLL_CAPABLE_RUNTIME entries must name real runtime symbols
	run: python3 scripts/gc_root_dominance_check.py --audit-poll-capable
	- name: ALLOC_RE symbols that reach a poll-capable one must be listed
	run: python3 scripts/gc_root_dominance_check.py --audit-poll-reach
	- name: Immovable-source exemptions must still hold
	run: python3 scripts/gc_root_dominance_check.py --audit-immovable-sources

	- name: Install Rust toolchain
	uses: dtolnay/rust-toolchain@stable
	# This also puts a matched LLVM 22 `opt` on disk and exports
	# LLVM_SYS_221_PREFIX, which is how the corpus script finds it. The
	# native corpus is `--trace llvm` output PLUS the production statepoint
	# rewrite, because codegen does not emit statepoints — LLVM does, later,
	# in the linker step.
	- uses: ./.github/actions/setup-llvm22

	- name: Cache cargo
	uses: actions/cache@v6
	with:
	path: \|
	~/.cargo/registry
	~/.cargo/git
	target
	key: ${{ runner.os }}-cargo-gcdom-${{ hashFiles('**/Cargo.lock') }}
	restore-keys: ${{ runner.os }}-cargo-

	- name: Build perry and the runtime archives
	run: \|
	set -euo pipefail
	cargo build --release -p perry -p perry-runtime-static -p perry-stdlib-static
	for artifact in perry libperry_runtime.a libperry_stdlib.a; do
	test -s "target/release/$artifact" \
	\|\| { echo "::error::target/release/$artifact was not produced"; exit 1; }
	done

	# The script single-sources the pass string from
	# `STATEPOINT_REWRITE_PASSES` in perry-codegen and refuses to run if it
	# cannot read it, so this corpus cannot drift away from the pipeline
	# production uses. It also asserts its own subject: zero statepoints or
	# zero live bundles is an error at generation time, not a clean verdict
	# downstream.
	- name: Emit the NATIVE (statepoint) IR corpus
	run: ./scripts/gc_root_dominance_corpus.sh ir-corpus-native --lowering native

	- name: Check GC values across safepoints (native roots)
	run: \|
	set -euo pipefail
	# No pipe: the checker's own exit status is the job's.
	#
	# FLOORS, from the corpus as of this commit — 149 modules, 2452
	# functions, 30033 safepoints, 17478 with a live bundle, 40759
	# relocations — set below that with room for churn and far above
	# "something compiled". Three separate floors because they fail for
	# three different reasons and the message has to say which:
	# --min-statepoints this is the shadow corpus, or codegen stopped
	# marking functions gc "statepoint-example"
	# --min-relocates the corpus was copied through un-rewritten
	# (`opt` exits 0 on a module with nothing to do)
	# --min-live-bundles safepoints that record NO roots, which is what
	# an unrooted build looks like
	#
	# BUDGETS. `unrooted` (the object is in NO live bundle at all, so
	# nothing marks or rewrites it) is the serious class, and it is a
	# RATCHET rather than a calibrated zero: this is a young instrument
	# pointed at a lowering nothing had ever checked statically, and the
	# remainder is a population under triage rather than a number anyone
	# has adjudicated. It is a budget rather than allowlist entries for
	# the reason `--stale-registers` records — tombstones with no issue
	# numbers would be worse documentation than one number that can only
	# go down. Lower it as sites are fixed.
	#
	# 21 at #7663. #7664 fixed 14 of them in `root_reload.rs` and
	# `lower_call/new.rs` — the whole `strhandle` population (10), the
	# `js_new_target_get` save/restore (1), and 3 of the 7 unmasked
	# receivers — so the budget is 7. #7616 (widening POLL_CAPABLE_RUNTIME)
	# and #7690 (rooting the spread argument-bundle accumulator) moved it
	# 7 -> 11 -> 8, tracked in the two changelog fragments for those PRs.
	#
	# ★ RE-VERIFIED AT #7664's SECOND PASS: the 8 above was already stale
	# by the time this ran. A fresh corpus read 9, not 8 — a static-method
	# receiver hazard in `test_gap_static_method_value_name_collision`
	# (added after the "8" snapshot, by the #7689/#7691 fix that armed a
	# receiver-sensitive static `this`) had joined the population without
	# anyone lowering — er, RAISING — the budget to match. That in itself
	# is the thing CLAUDE.md's hazard-4 corollary warns about: a budget
	# nobody re-measures silently absorbs the next hazard.
	#
	# Of the 9: 4 were CHECKER FALSE POSITIVES, not real hits. `chain`
	# (the untracked cast-closure a stale use is searched in) treated
	# `phi` as unconditionally transparent — one tainted incoming edge
	# blanket-tainted the phi's RESULT, and a downstream use of that
	# result was checked against ANY CFG path from source to use
	# (`between_blocks` is deliberately path-insensitive, which is sound
	# for an ordinary register but not a phi, whose dynamic value depends
	# on which edge was actually taken). All four were the identical
	# shape, e.g. `readCtx`:
	#
	# entry.0: %r2 = <unmask of the receiver>
	# br i1 %r4, label %then, label %merge
	# merge: %r87 = phi double [ %r2, %entry.0 ],
	# [ %r86, %pget.recv_merge ]
	# ret double %r87
	#
	# The safepoints are all on the `%then` path, where the phi selects
	# `%r86`. On the edge that carries `%r2` nothing collects between its
	# definition and the join — the checker was reporting the OTHER
	# edge's safepoint against THIS edge's value. Verified register-by-
	# register on all four (`readCtx`, `__closure_5`, `Readable`,
	# `__obj_method_toLocaleString_3` — every one a `logical.merge` join
	# of an `&&`).
	#
	# Fixed: `_cast_closure` gained `phi_all_edges` — a phi joins `chain`
	# only once EVERY incoming edge is independently in it, closing the
	# false positive. That deliberately gives up the case of a SINGLE
	# tainted edge with its own intervening safepoint before its own
	# predecessor's terminator; `_phi_edge_hazard` covers that separately,
	# checking each edge's window on its own. Two sabotage-tested
	# self-test fixtures (`phi_safe_edge` / `phi_hazard_edge`) pin both
	# directions.
	#
	# The remaining 5 were real, and split two ways:
	#
	# 3 unrooted:global. `test_gap_arraybuffer_transfer::main` (2,
	# `new DataView`/`new Uint8Array` reading a module-global
	# receiver, then holding it across a sibling `{ valueOf() {...} }`
	# argument's allocation) — fixed upstream by #7719, which covers
	# the identical shape across all 30 `lower_call/builtin.rs` ctor
	# arms via a shared `RootedGroup`. `test_gap_static_method_value_
	# name_collision::main` (1, `(Lexer as any).lex(...)`'s receiver
	# held across arg-bundling that always allocates for a rest
	# param) — fixed here, in
	# `lower_call/property_get/static_dispatch.rs`, the same
	# `RootedGroup::adopt`/`reread` shape on the receiver instead of
	# a constructor argument.
	# 2 unrooted:capture. `js_closure_get_capture_bits` returns a raw
	# `i64` that may be a NaN-boxed heap value, and unlike
	# `%this_closure` itself (which codegen already re-enters into
	# the `ptr addrspace(1)` tracked domain), its generic "read a
	# captured value" call sites never re-enter it into either that
	# domain or a temp root.
	# `test_gap_class_expr_dynamic_parent_ctor::__closure_21`: a
	# captured dynamic-parent-class reference read at the top of the
	# synthesized implicit ctor closure, used ~60 lines later as
	# `js_new_function_construct`'s callee, across
	# `js_object_alloc_class_inline_keys` AND the class's own user
	# constructor.
	# `test_gap_computed_key_method_nested_this::__closure_9`: a
	# captured numeric local read, then `total + x`-style `fadd`
	# after `js_number_coerce` (which can run a user
	# `Symbol.toPrimitive`) intervenes. NOT reloadable the way a
	# strhandle global is (re-deriving from `%this_closure` reads
	# the pre-move closure — RS4GC does not relocate a raw `i64`
	# parameter) — but IS reloadable the way OTHER root-derived
	# values are: re-calling `js_closure_get_capture_bits` with the
	# same closure pointer and index re-reads the same slot, sound
	# under the identical store side-condition
	# (`js_closure_set_capture_bits` to that index) `root_reload.rs`
	# already tracks for shadow slots and handle globals.
	#
	# ★ #7725 CLOSED both `unrooted:capture` hits, and the budget is
	# deleted rather than set to 0 — `--max-unrooted` already defaults
	# to 0, and a budget nobody re-measures is exactly the silently-
	# absorbing-the-next-hazard failure mode CLAUDE.md's hazard-4
	# corollary warns about (the "8 was already stale, a fresh corpus
	# read 9" paragraph above is that failure mode, once). The fix:
	# `root_reload.rs`'s `Facts` gained `capture_get_key`, so a
	# `js_closure_get_capture_bits` call whose ptr operand already
	# belongs to a reloadable recipe (the closure-ptr shadow-slot load
	# chain #7055 already protects) extends that recipe like a
	# transparent bit op — the whole chain (slot load, mask, a FRESH
	# call to `js_closure_get_capture_bits` with the same index, and
	# any further cast) re-materialises at every reload point. The
	# store side-condition switches from "was the closure-ptr slot
	# reassigned" (never, in generated code) to "was THIS INDEX set" —
	# a synthetic per-index key a `js_closure_set_capture_bits` call
	# populates via the same `stores_to` field a plain store uses,
	# which needed the `recipe[0]`-only group key widened to
	# `(recipe[0], root_ptr)` so the two conditions don't get
	# conflated under one reachability walk. Verified: native corpus
	# 2 -> 0 (`stale` held at 0), `--lowering shadow` corpus clean,
	# `--self-test` still proves the checker can fail.
	#
	# --seeded-violations plants 40 safepoints into this very IR, between
	# a `ptrtoint ptr addrspace(1)` and its use, and requires all 40 to be
	# reported. It runs even when the budgets are already red, because
	# the run whose "can it still fail" arm you most want is the one that
	# is already telling you something. That is the arm that fails if the
	# mode has stopped understanding LLVM's output — the only way a green
	# verdict here could be a lie.
	python3 scripts/gc_root_dominance_check.py ir-corpus-native \
	--statepoints \
	--moving-only \
	--min-files 90 --min-funcs 1200 \
	--min-statepoints 15000 \
	--min-live-bundles 8000 \
	--min-relocates 20000 \
	--max-stale 0 \
	--allowlist scripts/gc_root_dominance_allowlist.json \
	--seeded-violations 40 \
	-v

	# ★ The unfiltered arm, as a DIAGNOSTIC rather than a gate.
	#
	# `--moving-only` keeps only hazards whose window reaches a safepoint
	# that can run an evacuating minor. The rest are still hazards under the
	# RECLAIM half of the invariant (a value held only in an untracked
	# register can be swept, not just moved), and the count is two orders of
	# magnitude larger: 1444 (1123 unrooted, 321 stale) against the gated
	# arm's 21. Printing it keeps the filtered number honest — if the gated
	# arm ever reads zero, this line says whether that is the corpus or the
	# filter. `\|\| true` is deliberate and is the ONLY one in this file —
	# this step is not a gate and must not be read as one.
	- name: Unfiltered hazard census (diagnostic, not a gate)
	# The explicit opt-out `gc_gate_wiring_check.py` looks for. Without it
	# that gate reads the `\|\| true` below as a swallowed failure -- which
	# is exactly right for a gating step and exactly wrong for this one, so
	# the difference has to be declared rather than argued in a comment.
	continue-on-error: true
	run: \|
	python3 scripts/gc_root_dominance_check.py ir-corpus-native \
	--statepoints \
	--min-files 90 --min-funcs 1200 \
	--min-statepoints 15000 --min-live-bundles 8000 \
	--min-relocates 20000 \
	--max-unrooted 100000 --max-stale 100000 \|\| true

	- name: Upload the IR corpus on failure
	if: failure()
	uses: actions/upload-artifact@v7
	with:
	name: gc-root-dominance-statepoint-ir
	path: ir-corpus-native
	retention-days: 7

Navigation Menu

Uh oh!

Uh oh!

perf(runtime): let the GC header pick the side-registry probe on dynamic dispatch (#7850) #1732

Workflow file

perf(runtime): let the GC header pick the side-registry probe on dynamic dispatch (#7850) #1732

Uh oh!

Workflow file for this run