Re: [PATCH net-next 04/15] batman-adv: tt: use dedicated key type for hashes

[Sven Eckelmann <sven@narfation.org>]

[-- Attachment #1: Type: text/plain, Size: 1388 bytes --]

On Sunday, 15 November 2026 10:24:43 CET sashiko-bot@kernel.org wrote:
[...]
> [Severity: Medium]
> This problem wasn't introduced by this patch, but can batadv_tt_local_crc()
> include a BATADV_TT_CLIENT_PENDING entry in the local CRC calculation
> [1]?
> 
> batadv_tt_local_commit_changes_nolock() runs
> batadv_tt_local_purge_pending_clients() and batadv_tt_local_update_crc()
> sequentially under per-bucket locks, and batadv_tt_local_crc() only skips
> BATADV_TT_CLIENT_NEW:
> 
> CPU0                                    CPU1
> batadv_tt_local_commit_changes_nolock()
>   batadv_tt_local_purge_pending_clients()
>                                         batadv_tt_purge()
>                                           batadv_tt_local_purge()
>                                             ...
>                                               common->flags |=
>                                                 BATADV_TT_CLIENT_PENDING
>   batadv_tt_local_update_crc()
>     batadv_tt_local_crc()
>       includes pending entry in crc     <-- CRC mismatch
> 
> The local CRC then disagrees with the committed translation table state and
> triggers unnecessary full TT synchronizations across the mesh.


Not related to this patch. And the patchset to handle this better is pending [1].

Regards,
	Sven

[1] https://git.open-mesh.org/pub/ecsv/batman-adv.git/log/?h=b4/tt-avoid-unhashed-event

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 228 bytes --]