Hi,
I have sent the wrong version of the patch. I realized on the same issue when I 
created the wrynose one and I had this version in my scarthgap build.

It should have been the same that I sent for wrynose and master. Those are 
correct. It is the same libssh2 version.

I'll resend a v2 on Monday.

I really thought I sent the same patch to the three releases.

In the future, if it is the same patch to multiple branches, how should it be 
marked?

Thanks
Daniel
________________________________
From: Yoann Congal <[email protected]>
Sent: Saturday, June 27, 2026 1:06 AM
To: Daniel Turull <[email protected]>; 
[email protected] 
<[email protected]>
Cc: [email protected] <[email protected]>
Subject: Re: [OE-core] [scarthgap] [patch] libssh2: fix CVE-2026-55200

On Mon Jun 22, 2026 at 11:33 AM CEST, Daniel Turull via lists.openembedded.org 
wrote:
> From: Daniel Turull <[email protected]>
>
> Backport patch to fix CVE-2026-55200.
> https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fnvd.nist.gov%2Fvuln%2Fdetail%2FCVE-2026-55200&data=05%7C02%7Cdaniel.turull%40ericsson.com%7C2f3364dd9a284c28d67508ded3d79154%7C92e84cebfbfd47abbe52080c6b87953f%7C0%7C0%7C639181119971761508%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=tmWBrZABG8OQtWK6Cds18odLAc%2FOcm4ql3en4e0Wgio%3D&reserved=0<https://nvd.nist.gov/vuln/detail/CVE-2026-55200>
>
> Upstream fix:
>   
> https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Flibssh2%2Flibssh2%2Fcommit%2F97acf3dfda80c91c3a8c9f2372546301d4a1a7a8&data=05%7C02%7Cdaniel.turull%40ericsson.com%7C2f3364dd9a284c28d67508ded3d79154%7C92e84cebfbfd47abbe52080c6b87953f%7C0%7C0%7C639181119971778602%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=ct5f7AASg%2FMtVlbo6YFwhO4C1xrUbTzkI3J7pTZpCRM%3D&reserved=0<https://github.com/libssh2/libssh2/commit/97acf3dfda80c91c3a8c9f2372546301d4a1a7a8>
>
> Tested with ptest:
> Before: PASSED: 3, FAILED: 0, SKIPPED: 0
> After: PASSED: 3, FAILED: 0, SKIPPED: 0
>
> Reviewed-by: Anders Heimer <[email protected]
> Signed-off-by: Daniel Turull <[email protected]>
> ---
>  .../libssh2/libssh2/CVE-2026-55200.patch      | 51 +++++++++++++++++++
>  .../recipes-support/libssh2/libssh2_1.11.1.bb |  1 +
>  2 files changed, 52 insertions(+)
>  create mode 100644 meta/recipes-support/libssh2/libssh2/CVE-2026-55200.patch
>
> diff --git a/meta/recipes-support/libssh2/libssh2/CVE-2026-55200.patch 
> b/meta/recipes-support/libssh2/libssh2/CVE-2026-55200.patch
> new file mode 100644
> index 00000000000..f5ab9b9f204
> --- /dev/null
> +++ b/meta/recipes-support/libssh2/libssh2/CVE-2026-55200.patch
> @@ -0,0 +1,51 @@
> +From df0b03ee5ef12f3a46fccc0fc688ebfb91702972 Mon Sep 17 00:00:00 2001
> +From: Will Cosgrove <[email protected]>
> +Date: Fri, 12 Jun 2026 15:57:44 -0700
> +Subject: [PATCH] transport.c: Additional boundary checks for packet length
> + (#2052)
> +MIME-Version: 1.0
> +Content-Type: text/plain; charset=UTF-8
> +Content-Transfer-Encoding: 8bit
> +
> +Backport Resolution: Add additional bounds checking on packet length to
> +prevent OOB write — checks that packet_length is not below 1 and not
> +above LIBSSH2_PACKET_MAXPAYLOAD before proceeding.
> +
> +Conflicts Resolved:
> +
> +src/transport.c (1 conflict):
> +- Upstream uses renamed API ssh2_ntohu32(); stable branch uses
> +  _libssh2_ntohu32(). Kept stable function name while applying the
> +  new upper-bound check (LIBSSH2_PACKET_MAXPAYLOAD) unchanged.
> +
> +Assisted-by: kiro:claude-sonnet-4.6
> +
> +Changes from upstream commit 97acf3dfda80:
> +  - src/transport.c: adapted from upstream

While resolving the conflict, your AI completely rewrote the patch
message and notably removed the Credit: line.

Please keep the original patch message and put the backporting
information below the "Upstream-Status:" line. While you changing this,
the conflict resolution is trivial, no need to be so verbose (a line
referencing ssh2_ntohu32 should be enough).

Thanks!

> +
> +CVE: CVE-2026-55200
> +Upstream-Status: Backport 
> [https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Flibssh2%2Flibssh2%2Fcommit%2F97acf3dfda80c91c3a8c9f2372546301d4a1a7a8&data=05%7C02%7Cdaniel.turull%40ericsson.com%7C2f3364dd9a284c28d67508ded3d79154%7C92e84cebfbfd47abbe52080c6b87953f%7C0%7C0%7C639181119971788230%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=2qunEVqQYFYAKjgr3pUGogxQYBa9GYLKBT%2FQslEJb9o%3D&reserved=0<https://github.com/libssh2/libssh2/commit/97acf3dfda80c91c3a8c9f2372546301d4a1a7a8>]
> +
> +Signed-off-by: Daniel Turull <[email protected]>
> +---
> + src/transport.c | 6 +++++-
> + 1 file changed, 5 insertions(+), 1 deletion(-)
> +
> +diff --git a/src/transport.c b/src/transport.c
> +index e1120656..d147505b 100644
> +--- a/src/transport.c
> ++++ b/src/transport.c
> +@@ -639,8 +639,12 @@ int _libssh2_transport_read(LIBSSH2_SESSION * session)
> +                 total_num = 4;
> +
> +                 p->packet_length = _libssh2_ntohu32(block);
> +-                if(p->packet_length < 1)
> ++                if(p->packet_length < 1) {
> +                     return LIBSSH2_ERROR_DECRYPT;
> ++                }
> ++                else if(p->packet_length > LIBSSH2_PACKET_MAXPAYLOAD) {
> ++                    return LIBSSH2_ERROR_OUT_OF_BOUNDARY;
> ++                }
> +
> +                 /* total_num may include size field, however due to existing
> +                  * logic it needs to be removed after the entire packet is 
> read
> diff --git a/meta/recipes-support/libssh2/libssh2_1.11.1.bb 
> b/meta/recipes-support/libssh2/libssh2_1.11.1.bb
> index 2284d054b10..d6ee97f7ed0 100644
> --- a/meta/recipes-support/libssh2/libssh2_1.11.1.bb
> +++ b/meta/recipes-support/libssh2/libssh2_1.11.1.bb
> @@ -11,6 +11,7 @@ SRC_URI = 
> "https://eur02.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.libssh2.org%2Fdownload%2F%24&data=05%7C02%7Cdaniel.turull%40ericsson.com%7C2f3364dd9a284c28d67508ded3d79154%7C92e84cebfbfd47abbe52080c6b87953f%7C0%7C0%7C639181119971797263%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=b6xuzsvaWoBiC%2BiylM5H5sHW7Boo2xTyzR7gquZrO1g%3D&reserved=0{BP}.tar.gz
>  \
>             file://run-ptest \
>             file://0001-Return-error-if-user-KEX-methods-are-invalid.patch \
>             file://CVE-2026-7598.patch \
> +           file://CVE-2026-55200.patch \
>             "
>
>  SRC_URI[sha256sum] = 
> "d9ec76cbe34db98eec3539fe2c899d26b0c837cb3eb466a56b0f109cabf658f7"


--
Yoann Congal
Smile ECS

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#239684): 
https://lists.openembedded.org/g/openembedded-core/message/239684
Mute This Topic: https://lists.openembedded.org/mt/119921124/21656
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub 
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-

Reply via email to