Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [BUG] cc5b114dcf bpf, i40e: add meta data support
To:     Keith Busch <keith.busch@linux.intel.com>,
        intel-gfx@lists.freedesktop.org,
        John Fastabend <john.fastabend@gmail.com>,
        Jeff Kirsher <jeffrey.t.kirsher@intel.com>,
        linux-kernel@vger.kernel.org
References: <20180612161739.GA3303@localhost.localdomain>
From:   Daniel Borkmann <daniel@iogearbox.net>
Message-ID: <dc3176ae-5b16-ffa8-c8c8-595a0ede57f4@iogearbox.net>
Date:   Tue, 12 Jun 2018 22:24:52 +0200
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101
 Thunderbird/52.3.0
MIME-Version: 1.0
In-Reply-To: <20180612161739.GA3303@localhost.localdomain>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 06/12/2018 06:17 PM, Keith Busch wrote:
> My server's i40e no longer obtains an IP address on linux
> mainline. Bisected to the following:
> 
> commit cc5b114dcf986bfd8e4c37bf65d1b7b1e5290ac6
> Author: Daniel Borkmann <daniel@iogearbox.net>
> Date:   Mon May 28 11:07:20 2018 +0200
> 
>     bpf, i40e: add meta data support
> 
> Reverting on mainline resolves the issue.
> 
> Is there something wrong with my i40e adapter, or is the patch possibly
> doing something wrong? Or any other information I can get to help
> understand why it's stopped working with this feature?
> 
> An excert from "journalctl -xe" on on the failing network adapter
> is below.

Thanks for the report & sorry for the issue, Keith! Instead of revert,
could you give the below fix a try?

diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
index 8ffb745..ed6dbcf 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
@@ -2103,9 +2103,8 @@ static struct sk_buff *i40e_build_skb(struct i40e_ring *rx_ring,
 	unsigned int truesize = i40e_rx_pg_size(rx_ring) / 2;
 #else
 	unsigned int truesize = SKB_DATA_ALIGN(sizeof(struct skb_shared_info)) +
-				SKB_DATA_ALIGN(I40E_SKB_PAD +
-					       (xdp->data_end -
-						xdp->data_hard_start));
+				SKB_DATA_ALIGN(xdp->data_end -
+					       xdp->data_hard_start);
 #endif
 	struct sk_buff *skb;

@@ -2124,7 +2123,7 @@ static struct sk_buff *i40e_build_skb(struct i40e_ring *rx_ring,
 		return NULL;

 	/* update pointers within the skb to store the data */
-	skb_reserve(skb, I40E_SKB_PAD + (xdp->data - xdp->data_hard_start));
+	skb_reserve(skb, xdp->data - xdp->data_hard_start);
 	__skb_put(skb, xdp->data_end - xdp->data);
 	if (metasize)
 		skb_metadata_set(skb, metasize);

Thanks,
Daniel