@@ -4375,7 +4375,7 @@ int skb_gro_receive(struct sk_buff *p, struct sk_buff *skb)memcpy(frag+1,skbinfo->frags,sizeof(*frag)*skbinfo->nr_frags);/* We dont need to clear skbinfo->nr_frags here */-new_truesize=SKB_TRUESIZE(sizeof(structsk_buff));+new_truesize=SKB_DATA_ALIGN(sizeof(structsk_buff));delta_truesize=skb->truesize-new_truesize;skb->truesize=new_truesize;NAPI_GRO_CB(skb)->free=NAPI_GRO_FREE_STOLEN_HEAD;
From: Mat Martineau <hidden> Date: 2021-08-04 19:29:23
On Wed, 4 Aug 2021, Paolo Abeni wrote:
commit 5e10da5385d2 ("skbuff: allow 'slow_gro' for skb carring sock
reference") introduces a serious regression at the GRO layer setting
the wrong truesize for stolen-head skbs.
Restore the correct truesize: SKB_DATA_ALIGN(...) instead of
SKB_TRUESIZE(...)
Reported-by: Mat Martineau <redacted>
Fixes: 5e10da5385d2 ("skbuff: allow 'slow_gro' for skb carring sock reference")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
With this patch I'm no longer seeing the WARN_ON_ONCE() fire in
skb_try_coalesce(). Thanks Paolo!
Tested-by: Mat Martineau <redacted>
Hello:
This patch was applied to netdev/net-next.git (refs/heads/master):
On Wed, 4 Aug 2021 21:07:00 +0200 you wrote:
commit 5e10da5385d2 ("skbuff: allow 'slow_gro' for skb carring sock
reference") introduces a serious regression at the GRO layer setting
the wrong truesize for stolen-head skbs.
Restore the correct truesize: SKB_DATA_ALIGN(...) instead of
SKB_TRUESIZE(...)
[...]