Skip to content

[WIP] [bp/1.39] deps: Bump envoy_toolshed -> 0.4.13.post5, CVE exclude, an RBE pool pin and ext_proc fuzzer fix - #49393

Draft
phlax wants to merge 4 commits into
envoyproxy:release/v1.39from
phlax:bp-1.39-toolshed-cve
Draft

phlax wants to merge 4 commits into
envoyproxy:release/v1.39from
phlax:bp-1.39-toolshed-cve

Conversation

@phlax

@phlax phlax commented Nov 18, 2026 •

Copy link
Copy Markdown
Member

The release/v1.39 counterpart of #49374 (release/v1.40), ahead of the same toolshed bump on release/v1.38.

From #49374, the registry bump is this branch's toolshed bump, and the rules_python and zstd pin alignments are bzlmod only.

The CVE scan of this branch will still report the eight OpenSSL 3.5.9 CVEs, which are fixed in 3.5.10 (not yet released upstream).

phlax and others added 3 commits November 18, 2026 14:51
…yproxy#49375)

CVE-2026-130065 is a heap buffer overflow in
`msgpack_vrefbuffer_migrate()` (`src/vrefbuffer.c`) of msgpack-c through
7.0.2: the capacity check compares against the source array's end
pointer, so migrating between vrefbuffers of unequal size can skip the
realloc and write past the destination array.

Envoy uses the header-only C++ library (`msgpack-cxx` 7.0.0), so the C
file is not built. The C++ `msgpack::vrefbuffer::migrate()` has the same
check (`include/msgpack/v1/vrefbuffer.hpp`), but it is not reachable:
the only msgpack users are the fluentd access logger and tracer, which
pack into `msgpack::sbuffer`
(`source/extensions/common/fluentd/fluentd_base.h`). Nothing in
`source/` or `contrib/` uses a `vrefbuffer`.

The CVE scan on `main` reports it (alongside the OpenSSL 3.5.9 ones,
which this does not address). As for CVE-2026-72854 (envoyproxy#49052).

Risk Level: Low (CVE scan config only)
Testing: N/A
Docs Changes: N/A
Release Notes: N/A

Signed-off-by: Ryan Northey <ryan@synca.io>
…roxy#49360)

Change-Id: Ic669a99370269f92a16bd621b402408536e937a5

<!--
!!!ATTENTION!!!

If you are fixing **any** crash or **any** potential security issue,
**do not
open a pull request**.

Instead, please
[open a GitHub Security
Advisory](https://github.com/envoyproxy/envoy/security/advisories/new)
(preferred). Alternatively, you may email
envoy-security@googlegroups.com.

Thank you in advance for helping to keep Envoy secure.

!!!ATTENTION!!!

For an explanation of how to fill out the fields, please see the
relevant section
in
[PULL_REQUESTS.md](https://github.com/envoyproxy/envoy/blob/main/PULL_REQUESTS.md)

!!!ATTENTION!!!

Please check the [use of generative AI
policy](https://github.com/envoyproxy/envoy/blob/main/CONTRIBUTING.md?plain=1#L41).

You may use generative AI only if you fully understand the code. You
need to disclose
this usage in the PR description to ensure transparency.
-->

Commit Message:
Additional Description:
Risk Level:
Testing:
Docs Changes:
Release Notes:
Platform Specific Features:
[Optional Runtime guard:]
[Optional Fixes #Issue]
[Optional Fixes commit #PR or SHA]
[Optional Deprecated:]
[Optional [API
Considerations](https://github.com/envoyproxy/envoy/blob/main/api/review_checklist.md):]

Signed-off-by: Ethan Truong <ethantruong@google.com>
Signed-off-by: Ryan Northey <ryan@synca.io>
0.4.13.post5 has, backported to the WORKSPACE line
(envoyproxy/toolshed#5623), a `cve_scan` that does not need the
scanner's `RunEnvironmentInfo`. With Bazel 7 an `sh_binary` does not
provide it, so the CVE scan failed at analysis there (as it does on
`release/v1.38`); with Bazel 8 nothing changes.

Signed-off-by: Ryan Northey <ryan@synca.io>
@repokitteh-read-only

Copy link
Copy Markdown

As a reminder, PRs marked as draft will not be automatically assigned reviewers,
or be handled by maintainer-oncall triage.

Please mark your PR as ready when you want it to be reviewed!

🐱

Caused by: #49393 was opened by phlax.

see: more, trace.

…small RBE pool

envoyproxy#49321's pins came from a measurement in which the GCC jobs had been left on
the default pool: they have their own RBE platform, which the measurement did
not put on the large one, so their actions were killed and retried rather than
measured there. Measured again on the large pool, on this branch as landed
(2026-11-17, every GCC action on the pool it requested, none killed), the GCC
pins are the same but for this one: the compile peaks at 2928MB, which is over
the default pool (3GiB) with the 5% headroom the pins use.

Every measured peak was then checked against the pool this branch assigns it:
none is over its pool, and none is on a larger pool than it needs.

Signed-off-by: Ryan Northey <ryan@synca.io>
@phlax phlax changed the title [WIP] [bp/1.39] deps: Bump envoy_toolshed -> 0.4.13.post5, CVE exclude and ext_proc fuzzer fix [WIP] [bp/1.39] deps: Bump envoy_toolshed -> 0.4.13.post5, CVE exclude, an RBE pool pin and ext_proc fuzzer fix Nov 18, 2026
@phlax

phlax commented Nov 18, 2026

Copy link
Copy Markdown
Member Author

/retest verify

1 similar comment
@phlax

phlax commented Nov 18, 2026

Copy link
Copy Markdown
Member Author

/retest verify

This branch has not been deployed

No deployments
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