Skip to content

fix(agent-feed): an AtomicIntegerFieldUpdater publishes under its field's own slot (#1317, part) - #1455

Merged
PIsberg merged 10 commits into
mainfrom
fix/p4-int-updater-publication
Nov 21, 2026
Merged

PIsberg merged 10 commits into
mainfrom
fix/p4-int-updater-publication

Conversation

@PIsberg

@PIsberg PIsberg commented Nov 21, 2026

Copy link
Copy Markdown
Owner

Part of #1317. Stacked on #1444, #1446, #1447, #1448, #1452, #1453 and #1454, review the last commit.

Why

With the agent attached, AtomicityValidator reported an unordered publication on correct code in two cases:

  • a field published through an AtomicIntegerFieldUpdater's set, lazySet, a won compareAndSet, getAndSet, or its arithmetic and update forms, then read after the updater's get (or a direct volatile read of the field) saw the value;
  • a direct volatile write of the field, read through the updater.

Updater calls released and acquired nothing.

Change

Still open on #1317: AtomicLongFieldUpdater, VarHandle and array forms of a primitive.

Verified

  • 9 new AtomicFlagPublicationWeavingTest rows (7 updater forms, plus updater and direct access in both directions) were red before the change and are green after. The negative, a set made before the update, still reports the race.
  • A woven trace of all 13 updater forms checks that each returns the JDK's value. The library's JDK-equivalence test runs without the agent, where the field never resolves, so it does not reach the new loop.
  • Agent -P e2e: 639 tests green, the spinlock weaving tests included. Library -P fast: 4341 tests green. Full corpus green after a fresh install. Static gates, the javadoc jar and the guardrail drift check are clean.

Docs

  • docs/agent/attaching.md: how the updater is modelled, and what still orders nothing.
  • docs/CHANGELOG.md.

No guardrail note changes: the weaver's substitutions are unchanged, and only what the hooks do with them changed.

🤖 Generated with Claude Code

PIsberg and others added 10 commits November 20, 2026 18:59
…#1317, part)

#1315 made set, getAndSet, compareAndSet and the arithmetic forms of
AtomicBoolean/AtomicInteger/AtomicLong release and get acquire. A field
published through lazySet, setRelease, updateAndGet/getAndUpdate/
accumulateAndGet/getAndAccumulate, a won compareAndExchange(Release) or
a won weakCompareAndSetVolatile/Release, and read after a get that saw
the value, was still reported by AtomicityValidator as an unordered
publication on correct code. And a reader whose read was getAndSet's
previous value, a compareAndExchange witness, or what an add or update
returned acquired nothing.

Those hooks now call primitiveSlotStoring (before the store; after a
won swap for the compare forms) and HappensBefore.acquireVolatile on
the value read. The update and accumulate forms run as a CAS loop like
addPublishing, so the value to be stored is known before the store.
setPlain, setOpaque and weakCompareAndSetPlain stay unordered, as the
JMM says; AtomicFlagPublicationWeavingTest pins setPlain.

Verified failing first: 23 new AtomicFlagPublicationWeavingTest rows
(one per form) failed before the hook change and pass after; the
setPlain negative passed throughout. Library atomic, spinlock,
happens-before and atomicity tests plus MinTrustGateTest under -P e2e,
the full agent suite under -P e2e and the corpus eval are green;
static gates and the javadoc jar are clean.

Still open on #1317: getAcquire, intValue and longValue reads, the
AtomicLong forms the weaver does not substitute, and the updater,
VarHandle and array forms.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The previous commit made the value an increment, add, update or
accumulate returns acquire. The corpus idioms lane then filled a 2 GB
heap within seconds on CI (Java 26 OOM; 21 and 25 hung past 25 min),
reproduced locally under -Xmx2g. A heap histogram of the dead fork
showed 1.38 GB of long[] and 690 MB of int[] averaging 32 KB and 14 KB:
HappensBefore stamps near MAX_ENTRIES (4096). The add releases before
its store, so a counter every thread bumps (CorpusRecorder's static
BODY_EXECUTIONS, bumped once per body execution) collects every
bumper's clock; acquiring on the read handed that whole clock to each
bumper, and every access it recorded afterwards pinned a copy. A user
suite with a global id counter would hit the same OOM.

The add and update forms now release only, as on main before #1317.
getAndSet and the compareAndExchange witness still acquire, and a get
of the counter still acquires what it published.

Verified failing first: AtomicCounterClockTest (11 forms, 200 earlier
bumpers) failed with a clock of 201 threads on every form and passes
after. AtomicFlagPublicationWeavingTest's two add/update-result rows
are inverted to assertRaced, pinning the decision. Idioms lane under
-Xmx2g: OOM before, 220/220 in 37 s after; full corpus (6 lanes)
green; library -P fast 4336 tests green; static gates and javadoc jar
clean.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The woven addAll released every element before the call and, outside
queues, withdrew nothing. An element a concurrent set refused (it held
an equal one) or one a call throwing part way never added stayed
published, so a thread that reached it another way was ordered after
the producer by a hand-off that never happened, hiding a race.
HappensBefore.retract could not be reused: it keeps only the thread's
last release and forgets it at the next stamped access, which the
element's own equals/hashCode makes inside addAll.

HappensBefore.releaseWithdrawable hands each release back and withdraw
takes it back exactly (CAS of the element's clock from what the release
installed to what it held before; a later release by another thread
keeps it). collectionAddAll withdraws every element's release when the
call returns false, and those of elements the collection does not hold
by identity when it throws.

Verified failing first: AgentHappensBeforeFeedTest's two new tests
(newKeySet already holding equal tags; newKeySet.addAll stopping at a
null) failed before and pass after; the twin (the element that went in
stays published) passes throughout. Library -P fast 4338 green, full
corpus green after a fresh install, static gates and javadoc jar clean.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…se (#1316, part)

The woven increment, add and update forms of AtomicInteger/AtomicLong
release the value they are about to store before each compare-and-set
attempt (#1315). An attempt that lost to another thread storing the
same value left its release as the atomic's latest, so a get that saw
the winner's value was ordered after the loser's earlier writes too:
an edge the program does not have, which can hide a race on them.

HappensBefore.releaseVolatile now remembers, per thread and without
allocating, the field clock and the release it installed;
withdrawVolatile puts the field back to the release before it while it
is still the latest. addPublishing (int, long) and updatePublishing
call it after a lost compareAndSet. When the winner released after the
loser, its cumulative release already holds the loser's clock and
nothing is withdrawn; #1316 stays open for that order.

Verified failing first: AgentHappensBeforeFeedTest
.aLostAttemptWithdrawsItsRelease (the update function lets another
thread win the first swap, then a reader gets the winner's value) was
red before and is green after; its twin (the attempt stored the value)
stays silent throughout. Library -P fast 4339, agent
AtomicFlagPublicationWeavingTest 35/35 under -P e2e, full corpus after
a fresh install, static gates and javadoc jar all green.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
… function are ordered (#1336, part)

With the agent attached, an AtomicReferenceArray element replaced by a
successor built in updateAndGet (or getAndUpdate, accumulateAndGet,
getAndAccumulate, compareAndExchange, weakCompareAndSetVolatile), or a
field replaced through an AtomicReferenceFieldUpdater's update or
accumulate forms, was reported by AtomicityValidator at PROMPT when
another thread read the successor's state through the slot. These forms
stored with a volatile write that released nothing; #1334 fixed only
AtomicReference's.

Each form now has a hook that runs it as the woven get and the woven
compareAndSet of its slot kind, the loop the JDK runs inside it, and
FieldAccessWeaver's REFERENCE_ARRAY_FORMS and REFERENCE_UPDATER_FORMS
substitute them. weakCompareAndSetPlain and the deprecated
weakCompareAndSet stay unwoven (SpinLockHookTableTest pins it). The
AsyncTestAgent @aicore note names the new substitutions; the guardrail
files are regenerated from it.

Verified failing first: the two new must-stay-silent corpus idiom rows
(idiom_arrayUpdateAndGet_/idiom_updaterUpdateAndGet_buildsTheSuccessor
InTheFunction) were both reported before the hooks; with them the lane
passes 224/224 and the full corpus is green. TelemetryRegistryHooksMatch
TheJdkTest compares every new hook with the JDK call. Library -P fast
4339, agent -P e2e 610, static gates green.

Still open on #1336: the reference VarHandle compareAndExchange and
weak swaps, and the acquire/release flavours on every slot kind.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ompareAndSet publish (#1336, part)

With the agent attached, an object published through a reference
VarHandle compareAndExchange loop on a volatile field was reported by
AtomicityValidator at PROMPT when another thread read its state through
the handle: compareAndExchange and weakCompareAndSet (volatile on a
handle) stored with a volatile write that released nothing.

Each now has a hook for an instance field, a static field and an array
element, run as the woven compareAndSet and, for a lost exchange, the
woven getVolatile witness. FieldAccessWeaver.referenceHandleHook matches
them and the hook descriptor returns Object (narrowed back by the
existing CHECKCAST) or boolean. weakCompareAndSetPlain and the acquire
and release flavours stay unwoven; SpinLockHookTableTest pins both
sides. The AsyncTestAgent @aicore note and the generated guardrails
name the new substitutions.

Verified failing first: the new must-stay-silent idiom row
idiom_varHandleCompareAndExchange_retriesFromTheWitness was reported
against the jar without these hooks (lane 226, 1 failure) and passes
with them; the full corpus is green. TelemetryRegistryHooksMatchTheJdk
Test compares each hook with the handle's own call. Agent -P e2e 610,
library -P fast (JavadocDescribesRatherThanRestates caught two
restating @PARAM lines, fixed), static gates and javadoc jar green.

Still open on #1336: the acquire and release flavours on every slot
kind, which need a decision on what a one-sided swap publishes.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…omic acquire (#1317, part)

With the agent attached, a field published through a store into an
AtomicBoolean, AtomicInteger or AtomicLong and read after getAcquire(),
intValue() or longValue() saw the stored value was reported by
AtomicityValidator as an unordered publication on correct code: the
weaver substituted only get, so these reads acquired nothing.

Seven hooks now weave them: getAcquire on all three, intValue and
longValue on AtomicInteger and AtomicLong. Each acquires what the store
of the value it returned published, and the AtomicInteger/AtomicLong
ones are the check half of a check-then-act, as get is (#1249). The
AsyncTestAgent @aicore note and the generated guardrails name them.

Verified failing first: seven new AtomicFlagPublicationWeavingTest rows
(one per read, through a woven bean) were red before the hooks and are
green after. TelemetryRegistryHooksMatchTheJdkTest compares each hook
with the JDK call (AtomicLong.intValue on a value above 2^32 included).
Agent -P e2e 617, library -P fast 4339, full corpus, static gates and
javadoc jar green.

Still open on #1317: the AtomicLong forms the weaver does not substitute,
and the updater, VarHandle and array forms of a primitive.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…orm AtomicInteger does (#1317, part)

With the agent attached, a field published through AtomicLong.getAndSet,
a won compareAndExchange or compareAndExchangeRelease, a won
weakCompareAndSetRelease, or updateAndGet and the other update and
accumulate forms, then read after a get that saw the value, was reported
by AtomicityValidator as an unordered publication on correct code: the
weaver did not substitute those forms on AtomicLong, so they released
nothing. A getAndSet's previous value and a compareAndExchange or
compareAndExchangeAcquire witness acquired nothing either.

Ten hooks, modelled on AtomicInteger's without the spinlock parts (an
AtomicLong is never treated as a lock): releases before the store or
after a won swap per flavour, acquires on a returned previous value or
witness per flavour, and the update forms as the CAS loop that releases
before each attempt and withdraws a lost one (#1316). The plain and
opaque forms still order nothing. The @aicore note and the generated
guardrails name the new substitutions.

Verified failing first: 11 new AtomicFlagPublicationWeavingTest rows
(8 publish, 3 acquire) were red before the hooks and green after (53).
TelemetryRegistryHooksMatchTheJdkTest compares each hook with the JDK
call on a value above 2^32. Agent -P e2e 628, library -P fast 4339, full
corpus, static gates and javadoc jar green.

Still open on #1317: the updater, VarHandle and array forms of a
primitive, which need a slot identity that matches direct access.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…Updater is reported (#1275, part)

if (UPDATER.get(o) < LIMIT) UPDATER.incrementAndGet(o) admits past the
limit when another thread's update lands between the check and the
increment, as the same code on an AtomicInteger does, but only the
latter was reported: AtomicNonAtomicUpdateDetector keys its rule by the
atomic object, and an updater names a field of an owner. The updater's
get, incrementAndGet and getAndIncrement were not woven at all.

Design call: the slot is keyed by HappensBefore.fieldKey(receiver,
field), the field clock the model already keeps once per object and
field and holds weakly, so the key is canonical and keeps nothing alive;
the finding is named by the field (DeclaringClass.field, as
SpinLocks.fieldOf reads it), so no string is built on the record path.
The detector gains (key, name, receiver) overloads; the receiver is the
object whose monitor can guard the field. Every woven IntUpdater hook
reports its role (read, increment, swap, store), and get,
incrementAndGet and getAndIncrement are now substituted.

Verified failing first: corpus idiom pair idiom_updaterAdmission_*;
the check-then-increment row was SILENT but MUST_FIRE before, fires
after, and the get/compareAndSet loop twin stays silent (lane 228).
AtomicNonAtomicUpdateDetectorTest pins the field-named finding and that
two fields of one object are two slots. Agent -P e2e 628, library -P
fast 4341, full corpus, static gates, javadoc jar green.
SubstitutionHooksArePointsTest lists the four new recording entry
points with the existing atomic* ones.

Still open on #1275: AtomicLongFieldUpdater and VarHandle.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ld's own slot (#1317, part)

With the agent attached, a field published through an
AtomicIntegerFieldUpdater's set, lazySet, won compareAndSet, getAndSet
or arithmetic and update forms, then read after the updater's get (or a
direct volatile read of the field), was reported by AtomicityValidator
as an unordered publication on correct code, and so was a direct
volatile write read through the updater: updater calls released and
acquired nothing.

The updater's stores now release, and its get and getAndSet's previous
value acquire, under the field's own name (DeclaringClass.field, as
SpinLocks.fieldOf resolves it), which is the slot FieldAccessWeaver
gives a direct volatile access of the field, so the two kinds of access
order each other. The arithmetic and update forms run as a releasing
compare-and-set loop that withdraws a lost attempt (#1316); the value
they return acquires nothing, for AtomicCounterClockTest's reason. An
updater whose field cannot be resolved keeps the JDK's own call.

Verified failing first: 9 new AtomicFlagPublicationWeavingTest rows
(7 updater forms, and updater/direct access in both directions) were red
before and are green after; the "set before the update" negative stays
raced. A woven trace of all 13 forms checks each returns the JDK's
value. Agent -P e2e 639 (spinlock weaving tests included), library -P
fast 4341, full corpus, static gates and javadoc jar green.

Still open on #1317: AtomicLongFieldUpdater, VarHandle and array forms.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@PIsberg
PIsberg merged commit b14f3be into main Nov 21, 2026
71 of 73 checks passed
@PIsberg
PIsberg deleted the fix/p4-int-updater-publication branch November 21, 2026 07:37
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant