On 12/6/21 11:20 AM, Paolo Abeni wrote:

On Fri, 2021-12-03 at 23:04 +0100, Daniel Borkmann wrote:

quoted

Hi Paolo,

Changes look good to me as well, we can route the series via bpf-next after tree
resync, or alternatively ask David/Jakub to take it directly into net-next with our
Ack given in bpf-next there is no drivers/net/ethernet/microsoft/mana/mana_bpf.c yet.

On 11/30/21 11:08 AM, Paolo Abeni wrote:
[...]> diff --git a/net/core/filter.c b/net/core/filter.c

quoted

index 5631acf3f10c..392838fa7652 100644

--- a/net/core/filter.c
+++ b/net/core/filter.c

@@ -8181,13 +8181,13 @@ static bool xdp_is_valid_access(int off, int size,
   	return __is_valid_xdp_access(off, size);
   }
   
-void bpf_warn_invalid_xdp_action(u32 act)
+void bpf_warn_invalid_xdp_action(struct net_device *dev, struct bpf_prog *prog, u32 act)
   {
   	const u32 act_max = XDP_REDIRECT;
   
-	pr_warn_once("%s XDP return value %u, expect packet loss!\n",
+	pr_warn_once("%s XDP return value %u on prog %s (id %d) dev %s, expect packet loss!\n",
   		     act > act_max ? "Illegal" : "Driver unsupported",
-		     act);
+		     act, prog->aux->name, prog->aux->id, dev ? dev->name : "");

One tiny nit, but we could fix it up while applying I'd have is that for !dev case
we should probably dump a "<n/a>" or so just to avoid a kernel log message like
"dev , expect packet loss".

Yep, that would probably be better. Pleas let me know it you prefer a
formal new version for the patch.

Ok, I think no need, we can take care of it when applying.

Thanks,
Daniel