ACK: [SRU][Q/J][PATCH 0/2] CVE-2026-31705

[Jian Hui Lee]

Acked-by: Jian Hui Lee <jianhui.lee at canonical.com>

On Tue, Jun 23, 2026 at 11:46 AM Cengiz Can via kernel-team
<kernel-team at lists.ubuntu.com> wrote:
>
> https://ubuntu.com/security/CVE-2026-31705
>
> [ Impact ]
>
> smb2_get_ea() in ksmbd applies 4-byte alignment padding via memset() after
> writing each EA entry, but the bounds check on buf_free_len is performed before
> the value memcpy while the alignment memset fires unconditionally with no check
> on remaining space. When an EA value exactly fills the remaining buffer, the
> alignment memset writes 1-3 NUL bytes past the buf_free_len boundary, and in
> compound requests where the response buffer is shared across commands this
> overwrites past the physical kvmalloc allocation into adjacent kernel heap
> memory. With a CVSS score of 9.8, this out-of-bounds write is remotely
> triggerable and can corrupt kernel memory, potentially leading to denial of
> service or further compromise.
>
> [ Fix ]
>
> questing: clean cherry-pick
> jammy: clean cherry-pick
>
> [ Test Plan ]
>
> Boot tested.
>
> [ Where Problems Could Occur ]
>
> A regression in this fix could affect the ksmbd SMB2 server's QUERY_INFO EA
> response handling, potentially causing valid EA queries to be truncated or
> rejected. Any error in the added bounds check could affect clients retrieving
> extended attributes over SMB.
>