MMapDirectory applies a slice's read advice to the wrong bytes when the slice starts inside an mmap chunk and crosses into the next one (the #14109 prefetch fix missed this caller of advise())
Finding LFV-126Area: store / MMapDirectory read advice (MemorySegmentIndexInput slices, compound files)Priority: lowImpact: minorNovelty: new
AI-assisted finding, confirmed by automated checks, not yet triaged by a human Lucene committer. This site files nothing upstream.
What it shows
You take a slice of an MMapDirectory file with IndexInput#slice(name, offset, length, IOContext) and a context that asks for RANDOM or SEQUENTIAL read advice, as the compound-file reader does for each inner file, and you expect the madvise hint to land on the slice's own pages. When the slice starts inside an mmap chunk and crosses into the next chunk, the hint is applied from the start of that chunk instead: in the repro, a slice of [512 KiB, 1.5 MiB) with 1 MiB chunks puts the RANDOM hint on [0, 1 MiB), so other inner files of a compound file get it, and a short slice that straddles a chunk boundary can get none of its own pages advised. Reads and query results stay correct; only the operating system's readahead and caching behaviour on those pages changes.
The test shows that IndexInput#slice(name, 512 KiB, 1 MiB, RANDOM context) on an MMapDirectory with 1 MiB chunks (ADVISE_BY_CONTEXT) puts the MADV_RANDOM hint on the wrong bytes, because MultiSegmentImpl does not add its slice offset in advise() (Linux only; read from the rr flag in /proc/self/smaps).
Expected vs actual
Expected: RANDOM advice only inside the slice [524288, 1572864); actual (-ea and -da): file bytes [0, 1048576) get it, half of them before the slice.
Files
Run it
- Tested on Lucene main e1faeba08bde (2026-11-02): still reproduces.
- Tested on Lucene main 645d4fe14a20 (2026-10-28): reproduced.
- Recorded outcome of run 1 on Lucene main 645d4fe14a20: 1 test(s) run, 1 failed.
- How to read the outcome: a failing test is the finding reproducing (the failure message states expected against actual).
- Seeds: a rerun with the same seed replays the same random choices. Assertions off (-da -dsa -Dtests.asserts=false) is the production-like setting.
# Lucene checkout at the tested commit, built with: ./gradlew --no-daemon assemble (JDK 25)
LUCENE=/path/to/lucene; V=12.0.0-SNAPSHOT
CP=$LUCENE/lucene/core/build/libs/lucene-core-$V.jar
CP=$CP:$LUCENE/lucene/codecs/build/libs/lucene-codecs-$V.jar
CP=$CP:$LUCENE/lucene/test-framework/build/libs/lucene-test-framework-$V.jar
CP=$CP:randomizedtesting-runner-2.9.2.jar:junit-4.13.2.jar:hamcrest-3.0.jar # test-framework dependencies (Maven Central)
CP=$CP:junit-platform-console-standalone-6.1.3.jar # JUnit Platform console launcher (Maven Central)
javac -d classes -cp "$CP" $(find src -name '*.java')
# run 1: assertions on, seed DEADBEEF
java -ea -esa --add-modules jdk.incubator.vector --enable-native-access=ALL-UNNAMED -Dtests.seed=DEADBEEF \
-cp "classes:$CP" org.junit.platform.console.ConsoleLauncher \
execute --select-class org.apache.lucene.store.TestMultiChunkSliceReadAdviceGradle alternative: copy the test into the matching module's src/test/ tree of the checkout and run ./gradlew -p lucene/<module> test --tests <class> -Ptests.seed=<seed> (add -Ptests.asserts=false for a run without assertions).
Feedback
Open an issue in lucene-ai-bug-hunt. Please do not open issues or pull requests in apache/lucene about this finding unless a Lucene committer agrees it should be filed.
Snapshot of 2026-11-04. AI-assisted; not yet triaged by a human Lucene committer. Feedback: open an issue in lucene-ai-bug-hunt.