Skip to content

openssl 3.6.4 (11 CVEs, 2 Critical) - #710

Merged
bryan-minimal merged 3 commits into
mainfrom
fix/openssl-3.6.4-pcre2-10.48
Sep 18, 2026
Merged

bryan-minimal merged 3 commits into
mainfrom
fix/openssl-3.6.4-pcre2-10.48

Conversation

@bryan-minimal

@bryan-minimal bryan-minimal commented Sep 17, 2026 •

Copy link
Copy Markdown
Member

Why

openssl 3.6.3 matches 11 advisories — 2 Critical, 8 High, 1 Medium — and 3.6.4 (released 2026-08-25) fixes every one.

Checked rather than assumed: the 11 CVEs pkgmgr vulns --package openssl reports against 3.6.3 and the 11 CVEs named in the 3.6.4 release notes are the same set, with nothing left over on either side.

Critical CVE-2026-63073 (9.8), CVE-2026-75803 (9.1)
High CVE-2026-14456, -14457, -18798, -54874, -54876, -63072, -63075, -63076
Medium CVE-2026-63074

⚠️ OpenSSL's own release note says "the most severe CVE fixed in this release is Moderate". That is their severity scale; NVD rates two of these 9.8 and 9.1. Same vendor-vs-NVD understatement as the chromium work — don't let the upstream label set the urgency.

This is the second half of pkgs#709

These are the same upstream vulnerabilities, reaching us through two packaging paths:

  • pkgs#709 — the copy inside the Alpine minirootfs (libcrypto3/libssl3 3.5.7-r0 → 3.5.8-r0)
  • this PR — our own openssl package (3.6.3 → 3.6.4)

One upstream advisory set, two places to bump. Found via the transitive view on minimal.dev/pkgs/hunk — hunk itself is clean, but inherits all of these through git.

No version inversion, and no jump to 4.x. OpenSSL 4.0.2 (and 4.1.0-alpha1) exist, but 3.6.4 is the fix for our line. Alpine's 3.5.8 is the LTS-branch backport of the same CVE set — both lines got patched; we were one patch behind on ours.

pcre2 is bundled for the rebuild, not the yield

Stating this plainly so nobody reads it as a security win it isn't. Of pcre2's 7 advisories, 10.48 clears exactly one — the LOW GHSA-q7rw-r7qq-2hx6:

  • OSV-2026-343 (HIGH) is commit-ranged, and its fix commit e19e76ad is not an ancestor of the pcre2-10.48 tag (compare reports 10.48 behind it), so 10.48 does not clear it.
  • The other 5 have no upstream fix at all.

The reason to bundle is the cascade: openssl and pcre2 share git, nginx, wget and rizin, so this is one rebuild instead of two.

Verification

Both tarballs verified byte-identical against upstream before mirroring:

openssl  c6b94124bb76ac5f8aa80de121e0f922a93d3bbf1be478bd50ebc36a30dc1db1
pcre2    b6c68fdf6f3ac31388b50aa89ff0fc49c00c987c16e7b5146491d12003f2c8ed

Diff is minimal — version + sha256 only, no license churn (askalono misreads OpenSSL's license as "Pixar"; it did not get written).

Not built locally. min session activate on this worktree hangs on every exec — the symptom gominimal/inbox#693 / #690 describe. Left to CI's minimal-check, which does a real build.

Reviewer note: the cascade is wide

openssl has 38 direct importers, pcre2 has 7:

actions-runner atuin codex cmake ca-certificates cups curl dnsutils cloud-hypervisor ffmpeg foundationdb git grpc gurk lean libevent libkrunfw libssh2 nghttp2 nginx ngtcp2 node-lts node nushell openssh postgres python rsync rizin ruby rust sccache socat typst virtio-linux virtio-linux-detonation wget zellij

That is a big rebuild and deserves a reviewer's eye on the cascade rather than mine alone.

Why this didn't auto-update (worth its own issue)

openssl is not base-soup (no replace_on_cycle), so that isn't the reason. The nightly auto-fix triage has logged opened=0 every night, with skip_no_vuln_cleared=117 as its dominant bucket — and every one of openssl's advisories carries 324 commit-SHA fixes and zero version-shaped fixes (all GIT ranges), so a version-based "would this clear?" test cannot prove 3.6.4 helps.

I can't confirm openssl's bucket from the logs, because skips are only ever counted in aggregate, never named — which is the other half of the problem.

🤖 Generated with Claude Code

Summary by CodeRabbit

  • Updates
    • Updated the underlying OpenSSL and PCRE2 components to newer versions.
    • Includes the latest available fixes and improvements from both components, supporting continued security and reliability.

openssl 3.6.3 matches 11 advisories — 2 Critical, 8 High, 1 Medium — and
3.6.4 (2026-08-25) fixes every one of them. Checked rather than assumed: the
11 CVEs the scan reports against 3.6.3 and the 11 CVEs named in the 3.6.4
release notes are the SAME SET, with nothing left over on either side.

  CVE-2026-63073 (9.8) CVE-2026-75803 (9.1)  — Critical
  CVE-2026-14456 -14457 -18798 -54874 -54876 -63072 -63075 -63076 — High
  CVE-2026-63074 — Medium

These are the same upstream vulnerabilities as #709, reaching us
through a second packaging path: that PR fixes the copy inside the Alpine
minirootfs, this fixes our own openssl package. One upstream advisory set, two
places to bump.

No version inversion and no jump to 4.x. OpenSSL 4.0.2 exists, but 3.6.4 is
the fix for our line, and Alpine's 3.5.8 is the LTS-branch backport of the
same CVEs — both lines were patched, we were one patch behind on ours.

pcre2 10.47 -> 10.48 is bundled deliberately, not opportunistically. Its
advisory yield is small and stated honestly: of pcre2's 7 advisories 10.48
clears exactly ONE, the LOW GHSA-q7rw-r7qq-2hx6. The HIGH OSV-2026-343 is
commit-ranged and its fix commit e19e76ad is NOT an ancestor of the 10.48 tag
(checked via compare: 10.48 is behind it), so 10.48 does not clear it. The
remaining 5 have no upstream fix at all. The reason to bundle is the REBUILD,
not the yield: openssl and pcre2 share git, nginx, wget and rizin, so one
cascade instead of two.

Both tarballs verified byte-identical against upstream before mirroring:

  openssl c6b94124bb76ac5f8aa80de121e0f922a93d3bbf1be478bd50ebc36a30dc1db1
  pcre2   b6c68fdf6f3ac31388b50aa89ff0fc49c00c987c16e7b5146491d12003f2c8ed

Not built locally: `min session activate` on this worktree hangs on exec, which
is the symptom gominimal/inbox#693 / #690 describe. Left to CI's minimal-check,
which does a real build. openssl has 38 direct importers and pcre2 7, so the
rebuild is wide — worth a reviewer's eye on the cascade rather than mine alone.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
@coderabbitai

coderabbitai Bot commented Sep 17, 2026 •

Copy link
Copy Markdown
Contributor

Review Change StackReview Change Stack

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info
⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Essentials

Run ID: fa33277f-f029-4f49-948f-fe0ec72f50c3

📥 Commits

Reviewing files that changed from the base of the PR and between 6e7f7ea and 9222ea7.

📒 Files selected for processing (2)
  • packages/openssl/build.ncl
  • packages/pcre2/build.ncl

Included review availability: 0 reviews are currently available. Your included PR review attempts over the past 7 days set your current allowance at 1 review per hour.


📝 Walkthrough

Walkthrough

The OpenSSL package moves to version 3.6.4. The PCRE2 package moves to version 10.48. Both definitions update their source archive SHA-256 checksums.

Changes

Package source refreshes

Layer / File(s) Summary
Source version and checksum updates
packages/openssl/build.ncl, packages/pcre2/build.ncl
The build definitions reference newer source versions and matching SHA-256 checksums.

Priority: ⬆️ High

Estimated code review effort: 1 (Trivial) | ~5 minutes

Change: Bug fix

Merge Risk: ⚪ Minimal · up to 6ee4c

The dependency refresh introduces no identified merge-blocking behavior or risk.

🚥 Pre-merge checks | ✅ 5
✅ Passed checks (5 passed)
Check name Status Explanation
Docstring Coverage ✅ Passed No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check. Docstring coverage is scoped to functions touched by this diff. Analyzed 0 functions across 0…
Linked Issues check ✅ Passed Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check ✅ Passed Check skipped because no linked issues were found for this pull request.
Description Check ✅ Passed Check skipped - CodeRabbit’s high-level summary is enabled.
Title check ✅ Passed The title clearly identifies the primary change: updating OpenSSL to 3.6.4. It also highlights the security advisories addressed by the update.
✨ Finishing Touches
🧪 Generate unit tests (beta)
  • Create PR with unit tests
  • Commit unit tests in branch fix/openssl-3.6.4-pcre2-10.48

Comment @coderabbitai help to get the list of available commands.

@bryan-minimal

Copy link
Copy Markdown
Member Author

Correcting myself: pcre2 is worth far more than I claimed

I wrote above that 10.48 "clears exactly ONE, the LOW GHSA-q7rw-r7qq-2hx6". That is wrong, and it understates this PR.

Measured by actually running the scanner against both trees rather than reading the raw advisory JSON:

pcre2 10.47 (main)            8 vulns
  GHSA-3r4p-g7gg-ppmf (HIGH)   OSV-2026-343 (HIGH)
  GHSA-fmgr-6ggq-9859 (MED)    GHSA-q8g2-wprr-34m9 (MED)
  CVE-2026-89162 (LOW)         GHSA-2p8c-ff85-vh9x (LOW)
  GHSA-9qww-pwc4-77qq (LOW)    GHSA-q7rw-r7qq-2hx6 (LOW)

pcre2 10.48 (this PR)         1 vuln
  OSV-2026-343 (HIGH)

7 of 8 clear. Only OSV-2026-343 survives, which matches the ancestry check in the description — its fix commit is not in the 10.48 tag.

Why I got it wrong: I queried affected[].ranges[].events[].fixed in SQL and concluded five advisories had "no upstream fix". The scanner resolves more than that raw field — commit ancestry plus GHSA/NVD fix-version enrichment — and at 10.48 seven of them no longer match at all. The SQL was a weaker instrument than the tool whose behaviour I was predicting, and I should have run the tool first.

Combined effect

openssl 3.6.3  11 vulns  ->  3.6.4   0 vulns   (all 11 clear via commit ancestry)
pcre2   10.47   8 vulns  ->  10.48   1 vuln

For hunk, whose page prompted this and whose advisories are entirely transitive through git: 18 → 1.

So the bundling argument is no longer only about amortising the rebuild — pcre2 carries real security value on its own.

#698 (open since 2026-09-15, green, mergeable) sets pcre2 to 10.48 with the
identical sha256 b6c68fdf, and also carries xz 5.8.3 -> 5.8.4. I cut this PR
without checking the open queue first, which duplicated it and would have put
two PRs on packages/pcre2/build.ncl.

This PR is now openssl-only, which is its unique contribution. The pcre2 half
of the hunk fix rides #698.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
@bryan-minimal bryan-minimal changed the title openssl 3.6.4, pcre2 10.48 (11 CVEs, 2 Critical) openssl 3.6.4 (11 CVEs, 2 Critical) Sep 17, 2026
@bryan-minimal

Copy link
Copy Markdown
Member Author

Scope reduced: this PR is now openssl-only

I cut this without checking the open PR queue first. #698 (open since 2026-09-15, green, mergeable) already bumps pcre2 to 10.48 with the identical sha256 b6c68fdf…, and also carries xz 5.8.3 → 5.8.4.

Duplicating it would have put two PRs on packages/pcre2/build.ncl for no gain, so I have reverted the pcre2 change here. #698 should merge as-is — it is older, green, and carries more.

That does not change the hunk outcome, only which PR delivers which half:

clears PR
openssl 3.6.3 → 3.6.4 11 (2 Critical) this PR
pcre2 10.47 → 10.48 7 of 8 #698

hunk still goes 18 → 1 once both land; only OSV-2026-343 (HIGH, pcre2) survives, and its fix commit genuinely is not in the 10.48 tag.

The "one cascade instead of two" argument in the description no longer applies to this PR — it now rebuilds openssl's 38 importers only.

@bryan-minimal
bryan-minimal added this pull request to the merge queue Sep 17, 2026
@github-merge-queue
github-merge-queue Bot removed this pull request from the merge queue due to no response for status checks Sep 17, 2026
@bryan-minimal
bryan-minimal added this pull request to the merge queue Sep 18, 2026
Merged via the queue into main with commit 43e2f0a Sep 18, 2026
11 checks passed
@bryan-minimal
bryan-minimal deleted the fix/openssl-3.6.4-pcre2-10.48 branch September 18, 2026 01:51
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.

2 participants