Re: [PATCH net-next 07/15] batman-adv: never consume skb in BLA rx and DAT incoming snooping

[Sven Eckelmann <sven@narfation.org>]

[-- Attachment #1: Type: text/plain, Size: 1657 bytes --]

On Sunday, 15 November 2026 10:24:44 CET sashiko-bot@kernel.org wrote:
> [Severity: Medium]
> This problem wasn't introduced by this patch, but can batadv_bla_rx() ->
> batadv_handle_claim() register the local node's own backbone gateway with
> own_backbone set to false [1]?
> 
> When a frame from an unclaimed source MAC arrives on a VLAN before a local
> entry exists in bat_priv->bla.backbone_hash, batadv_bla_rx() calls
> batadv_handle_claim() with primary_if->net_dev->dev_addr before reaching
> batadv_bla_update_own_backbone_gw() at the allow label:
> 
> net/batman-adv/bridge_loop_avoidance.c:batadv_handle_claim() {
> 	...
> 	backbone_gw = batadv_bla_get_backbone_gw(bat_priv, backbone_addr, vid,
> 						 false);
> 	...
> 	if (batadv_compare_eth(backbone_addr, primary_if->net_dev->dev_addr))
> 		batadv_bla_send_claim(bat_priv, claim_addr, vid,
> 				      BATADV_CLAIM_TYPE_CLAIM);
> 	...
> }
> 
> Because batadv_bla_get_backbone_gw() inserts the new entry with
> own_backbone as false and batadv_bla_update_own_backbone_gw() then finds the
> already-inserted entry, the local gateway skips the initial
> batadv_bla_send_announce() call and never enters the
> BATADV_BLA_BACKBONE_GW_UNSYNCED startup grace period (entry->wait_periods
> and bat_priv->bla.num_requests).
> 
> [1] https://sashiko.dev/#/bug/linux-dba9d95d-5eeb-40fd-abf4-e054aee1f735
> 
> 

Not related to this patch. And the patchset to handle this better is 
pending [1] ("bla: register own backbone gw when claiming for local node" 
might be submitted outside this patchset).

Regards,
        Sven

[1] https://git.open-mesh.org/pub/ecsv/batman-adv.git/log/?h=b4/resource-limit

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