Hi Yoann, Paul,

Thanks for the feedback.

I have opened a discussion [1] with the rsync upstream project to clarify
the status and maintenance model of the v3.4.1-sec-patches3 and
v3.2.7-sec-patches3 branches, including whether they will remain available
and whether future security patch branches will be published.

Regarding the 3.4.4 upgrade, I noticed that upgrading from 3.4.1 to 3.4.4
would still leave us with around 33 CVEs to address. The rsync 3.5.0
release [2] addresses these 33 CVEs.

I will continue to investigate the available options

[1] https://github.com/RsyncProject/rsync/discussions/1095
[2] https://download.samba.org/pub/rsync/NEWS#3.5.0

Thanks & Regards,

Vijay


On Thu, Sep 17, 2026 at 6:21 PM Paul Barker <[email protected]> wrote:

> On Thu, 2026-09-17 at 14:31 +0200, Yoann Congal wrote:
> > On Thu Sep 17, 2026 at 1:38 PM CEST, Yoann Congal wrote:
> > > On Thu Sep 17, 2026 at 12:55 PM CEST, Vijay Anusuri via
> lists.openembedded.org wrote:
> > > > Backport the security fixes from the upstream v3.4.1-sec-patches3
> > > > branch to address the known rsync security vulnerabilities.
> > > >
> > > > The v3.4.1-sec-patches3 branch contains 239 commits. The GitHub
> > > > workflow commits (037, 038, and 160), which only modify
> > > > .github/workflows files, and the testsuite-only commits (238 and
> > > > 239) are excluded because they are not applicable to the Yocto
> > > > build.
> > > >
> > > > The remaining 237 security fixes are combined into a single patch
> > > > series and applied on top of the rsync 3.4.1 source.
> > > >
> > > > SUSE has also backported these security fixes to
> rsync-3.4.1-160000.6.1
> > > > to address the corresponding CVEs.
> > > >
> > > > References:
> > > >
> > > > https://rsync.samba.org/security.html
> > > > https://github.com/RsyncProject/rsync/tree/v3.4.1-sec-patches3
> > > > https://bugzilla.suse.com/show_bug.cgi?id=CVE-2026-53802
> > > >
> > > > This fix handles CVE-2025-10158 CVE-2026-29518 CVE-2026-43617
> CVE-2026-43618 CVE-2026-43619 CVE-2026-43620 CVE-2026-45232 CVE-2026-44507
> CVE-2026-44508 CVE-2026-44509 CVE-2026-44510 CVE-2026-53783 CVE-2026-53784
> CVE-2026-53785 CVE-2026-53786 CVE-2026-53788 CVE-2026-53789 CVE-2026-53790
> CVE-2026-53791 CVE-2026-53792 CVE-2026-53793 CVE-2026-53794 CVE-2026-53795
> CVE-2026-53796 CVE-2026-53797 CVE-2026-53798 CVE-2026-53799 CVE-2026-53800
> CVE-2026-53801 CVE-2026-53802 CVE-2026-53803 CVE-2026-70452 CVE-2026-70453
> CVE-2026-70454 CVE-2026-70455 CVE-2026-70456 CVE-2026-70457 CVE-2026-70458
> CVE-2026-70459 CVE-2026-70460 CVE-2026-70461 CVE-2026-70462 CVE-2026-70463
> CVE-2026-70464
> > > >
> > > > Dropped CVE-2025-10158.patch
> > > > Refreshed the patch
> 0001-Add-missing-prototypes-to-function-declarations.patch
> > > >
> > > > Signed-off-by: Vijay Anusuri <[email protected]>
> > > > ---
> > > >  ...-prototypes-to-function-declarations.patch |    81 +-
> > > >  .../rsync/files/CVE-2025-10158.patch          |    36 -
> > > >  .../files/rsync-3.4.1-sec-patches3.patch      | 34052
> ++++++++++++++++
> > > >  meta/recipes-devtools/rsync/rsync_3.4.1.bb    |     2 +-
> > > >  4 files changed, 34071 insertions(+), 100 deletions(-)
> > > >  delete mode 100644
> meta/recipes-devtools/rsync/files/CVE-2025-10158.patch
> > > >  create mode 100644
> meta/recipes-devtools/rsync/files/rsync-3.4.1-sec-patches3.patch
> > >
> > > Hello,
> > >
> > > I don't think I want to carry a 34000 lines patch.
> > >
> > > Does the rsync project released a v3.4.1-sec-patches3 archive?
> > >
> > > If not, can we try to switch the recipe to git and point SRCREV to the
> > > "rsync-3.4.1-sec-patches" branch?
> > > Since the recipe is not git-based now, we'll need to switch to git
> > > first, then upgrade.
> >
> > Paul asked a good question about this idea though: How official is this
> > branch?
> > Can you ask upstream the status of it? Will it stay published? Will we
> > see v3.4.1-sec-patches4,5... branches someday?
>
> Hi Yoann, Vijay,
>
> Some thoughts here...
>
> The upgrade to v3.4.3 was rejected [1] due to a few minor feature
> additions. In this case it may be lower risk to take an upgrade rather
> than backporting a 34 kLOC patch. I think we should avoid v3.5.0 due to
> the number of regressions reported [2]. v3.4.4 is an option. We can then
> see how many CVEs remain open and decide what to do about them.
>
> The onus here is on contributors, not on Yoann as stable maintainer. To
> go ahead we would need some investigation, testing and a proposal to the
> TSC to approve the update as an exception to our usual stable policy.
>
> [1]: https://lore.kernel.org/all/[email protected]/
> [2]:
> https://github.com/RsyncProject/rsync/issues?q=is%3Aissue%20%223.5.0%22
>
> Best regards,
>
> --
> Paul Barker
>
>
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#246066): 
https://lists.openembedded.org/g/openembedded-core/message/246066
Mute This Topic: https://lists.openembedded.org/mt/121293944/21656
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub 
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-

Reply via email to