[rkennke]

What does this PR do?:

Makes MallocFootprint::headerBytes() independent of heap state on glibc.

• Before: the header was the smallest address stride between 16 fresh 48-byte blocks, minus their usable size. That is exact only if two blocks are adjacent. With freed same-size chunks scattered between live ones (tcache, fastbins), none are; the derived header exceeds the sanity bound and the probe falls back to 0.
• After: glibc is recognised by malloc_usable_size. A glibc chunk's usable size holds the request and is congruent to size_t modulo 2 * size_t (a free chunk too small to split is handed out whole, so it may be larger). Then the header is sizeof(size_t). Size-class allocators (jemalloc, tcmalloc, mimalloc), sanitizers and glibc's debug modes report other sizes and keep the stride probe, as does any non-glibc build.

Motivation:

PROF-16484.

• Production: when the probe fell back to 0, the NM_* overhead gauges undercounted every small allocation by 8 bytes for the process lifetime (a 96-byte node reads 8.3 % instead of 16.7 %).
• CI: MallocFootprintTest.OverheadFractionCollapsesWithAllocationSize failed on arm64 merge-queue runs of unrelated PRs (Deny direct gradlew calls from agent sessions (PROF-16469) #937 twice, Keep .knowledgebase writes to cairn: file-tool hook and frontmatter check (PROF-16469) #946 twice) with probed per-chunk header = 0.

Additional Notes:

• probeHeaderBytes(), glibcChunks(), glibcUsable() and strideHeaderBytes() are public so tests can call them in a prepared heap state; production code still uses the cached headerBytes(). A gtest friend would pull gtest into a production header.
• The recognition assumes malloc_usable_size belongs to the allocator in force; of() already relied on that.
• Not changed, pre-existing:
  • glibc's own mmap'd large chunks carry a 16-byte header, so of() stays 8 bytes low for them;
  • the test's allocatorExposesOverhead() comment says sanitizers make the probe yield 0; under ASan the stride probe yields 16;
  • the 512 KB < 1 % assertion would fail on arm64 kernels with 16 or 64 KiB pages.

How to test the change?:

Check	Result
gtestRelease_mallocFootprint_ut, gtestDebug_…, gtestAsan_… (x64)	passed
FragmentedHeapStillYieldsGlibcHeader: after freeing every other one of 512 blocks, the stride probe alone misses the header, the full probe returns sizeof(size_t)	passed; the precondition is reached on glibc
Mutation check: drop glibc recognition, glibc header 0, recognition always false, exact-size signature, no alignment check, no size check	each fails a test
Run under preloaded jemalloc (ubuntu:24.04 container)	glibc not recognised, header 0, glibc tests skip

For Datadog employees:

• If this PR touches code that signs or publishes builds or packages, or handles
  credentials of any kind, I've requested a security review (run the dd:platform-security-review
  skill, or file a request via the PSEC review form).
  bewaire also runs automatically on every PR.
• This PR doesn't touch any of that.
• JIRA: PROF-16484

🤖 Generated with Claude Code

[chatgpt-codex-connector]

Codex Review Summary: ✅ Completed

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-11-06T21:10:40.813759Z	3d57e13	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[datadog-datadog-us1-prod]

More details

Usable-size recognition survived a focused glibc fragmentation check, supporting the replacement of heap-dependent stride detection. The Google Test suite could not run because Google Test was unavailable.

Was this helpful? React 👍 or 👎

_{Open Bits AI session}

_{🤖 Bits Code Review · Commit 3d57e13}

[dd-octo-sts]

✅ All 40 integration tests passed

📊 Dashboard · 👷 Pipeline · 📦 3d57e13d

[dd-octo-sts]

CI Test Results

Run: #42294125733 | Commit: 6606344 | Duration: 22m 47s (longest job)

✅ All 32 test jobs passed

Status Overview

JDK	glibc-aarch64/debug	glibc-amd64/debug	musl-aarch64/debug	musl-amd64/debug
8	-	✅	-	-
8-ibm	-	✅	-	-
8-j9	✅	✅	-	-
8-librca	-	-	✅	✅
8-orcl	-	✅	-	-
11	-	✅	-	-
11-j9	✅	✅	-	-
11-librca	-	-	✅	✅
17	✅	✅	-	-
17-graal	✅	✅	-	-
17-j9	✅	✅	-	-
17-librca	-	-	✅	✅
21	✅	✅	-	-
21-graal	✅	✅	-	-
21-librca	-	-	✅	✅
25	✅	✅	-	-
25-graal	✅	✅	-	-
25-librca	-	-	✅	✅

Legend: ✅ passed | ❌ failed | ⚪ skipped | 🚫 cancelled

Summary: Total: 32 | Passed: 32 | Failed: 0

---

Updated: 2026-11-06 21:46:49 UTC

[dd-octo-sts]

Scan-Build Report

User:	runner@runnervm7aw5e
Working Directory:	/home/runner/work/java-profiler/java-profiler/ddprof-lib/src/test/make
Command Line:	make -j4 all
Clang Version:	Ubuntu clang version 18.1.3 (1ubuntu1)
Date:	Fri Nov 6 21:46:47 2026

Bug Summary

Bug Type	Quantity	Display?
All Bugs	2
Logic error
Dereference of null pointer	1
Unused code
Dead initialization	1

Reports

Bug Group	Bug Type ▾	File	Function/Method	Line	Path Length
Unused code	Dead initialization	referenceChains.cpp	applyOomRamp	497	1
Logic error	Dereference of null pointer	perfEvents_linux.cpp	unregisterThread	720	26