From: Ivan Malov <ivan.malov@arknetworks.am>
To: Khadem Ullah <14pwcse1224@uetpeshawar.edu.pk>
Cc: dev@dpdk.org, jasvinder.singh@intel.com,
bruce.richardson@intel.com, thomas@monjalon.net,
andrew.rybchenko@oktetlabs.ru, stable@dpdk.org
Subject: Re: [PATCH] net/rte_net: fix inner L2 length for tunneled Ethernet packets
Date: Mon, 28 Jul 2025 17:28:47 +0400 (+04) [thread overview]
Message-ID: <e7c3c2b2-5f7a-eec4-cbd8-cd17ecec4c46@arknetworks.am> (raw)
In-Reply-To: <20250728131222.50441-1-14pwcse1224@uetpeshawar.edu.pk>
Hi,
On Mon, 28 Jul 2025, Khadem Ullah wrote:
> Fix incorrect inner_l2_len values for VXLAN, VXLAN-GPE, GTPU, and Geneve.
> These protocols carry full Ethernet frames, so inner_l2_len should be
> set to the size of an Ethernet header (14 bytes), not include tunnel or
> UDP headers.
Does the bug break 'testpmd' (see 'inner_l2_len' over there) or what? How can
this be reproduced?
Also, how well does the patch agree with what is currently said in 'rte_net.h'?
> /* Outer_L4_len + ... + inner L2_len for tunneling pkt. */
> uint8_t inner_l2_len;
Is the comment also wrong or do we have a fundamental confusion in DPDK?
I know, for example, that in mbufs, in the case of a tunneled (VXLAN) packet,
fields 'outer_l2_len' and 'outer_l3_len' are Ethernet and IP sizes respectively,
whilst 'l2_len' is a sum of the UDP size, VXLAN size and inner Ethernet header.
But that's in mbuf, which does not have a separate 'tunnel_len' field.
So either there's some confusion, or the patch also needs to fix other places,
such as the header file mentioned ('rte_net.h').
Thank you.
>
> Fixes: 64ed7f854c ('net: add tunnel packet type parsing')
> Cc: stable@dpdk.org
>
> Signed-off-by: Khadem Ullah <14pwcse1224@uetpeshawar.edu.pk>
> ---
> lib/net/rte_net.c | 14 +++++++-------
> 1 file changed, 7 insertions(+), 7 deletions(-)
>
> diff --git a/lib/net/rte_net.c b/lib/net/rte_net.c
> index 44fb6c0f51..adcddeb670 100644
> --- a/lib/net/rte_net.c
> +++ b/lib/net/rte_net.c
> @@ -198,7 +198,7 @@ ptype_tunnel_with_udp(uint16_t *proto, const struct rte_mbuf *m,
> case RTE_VXLAN_DEFAULT_PORT: {
> *off += sizeof(struct rte_vxlan_hdr);
> hdr_lens->tunnel_len = sizeof(struct rte_vxlan_hdr);
> - hdr_lens->inner_l2_len = RTE_ETHER_VXLAN_HLEN;
> + hdr_lens->inner_l2_len = sizeof(struct rte_ether_hdr);
> *proto = RTE_VXLAN_GPE_TYPE_ETH; /* just for eth header parse. */
> return RTE_PTYPE_TUNNEL_VXLAN;
> }
> @@ -210,7 +210,7 @@ ptype_tunnel_with_udp(uint16_t *proto, const struct rte_mbuf *m,
> return 0;
> *off += sizeof(struct rte_vxlan_gpe_hdr);
> hdr_lens->tunnel_len = sizeof(struct rte_vxlan_gpe_hdr);
> - hdr_lens->inner_l2_len = RTE_ETHER_VXLAN_GPE_HLEN;
> + hdr_lens->inner_l2_len = sizeof(struct rte_ether_hdr);
> *proto = vgh->proto;
>
> return RTE_PTYPE_TUNNEL_VXLAN_GPE;
> @@ -244,7 +244,7 @@ ptype_tunnel_with_udp(uint16_t *proto, const struct rte_mbuf *m,
> *proto = 0;
> }
> *off += gtp_len;
> - hdr_lens->inner_l2_len = gtp_len + sizeof(struct rte_udp_hdr);
> + hdr_lens->inner_l2_len = sizeof(struct rte_ether_hdr);
> hdr_lens->tunnel_len = gtp_len;
> if (port_no == RTE_GTPC_UDP_PORT)
> return RTE_PTYPE_TUNNEL_GTPC;
> @@ -262,7 +262,7 @@ ptype_tunnel_with_udp(uint16_t *proto, const struct rte_mbuf *m,
> geneve_len = sizeof(*gnh) + gnh->opt_len * 4;
> *off += geneve_len;
> hdr_lens->tunnel_len = geneve_len;
> - hdr_lens->inner_l2_len = sizeof(struct rte_udp_hdr) + geneve_len;
> + hdr_lens->inner_l2_len = sizeof(struct rte_ether_hdr);
> *proto = gnh->proto;
> if (gnh->proto == 0)
> *proto = rte_cpu_to_be_16(RTE_ETHER_TYPE_IPV4);
> @@ -498,7 +498,7 @@ uint32_t rte_net_get_ptype(const struct rte_mbuf *m,
> pkt_type |= RTE_PTYPE_INNER_L2_ETHER;
> proto = eh->ether_type;
> off += sizeof(*eh);
> - hdr_lens->inner_l2_len += sizeof(*eh);
> + hdr_lens->inner_l2_len = sizeof(struct rte_ether_hdr);
> }
>
> if (proto == rte_cpu_to_be_16(RTE_ETHER_TYPE_VLAN)) {
> @@ -511,7 +511,7 @@ uint32_t rte_net_get_ptype(const struct rte_mbuf *m,
> if (unlikely(vh == NULL))
> return pkt_type;
> off += sizeof(*vh);
> - hdr_lens->inner_l2_len += sizeof(*vh);
> + hdr_lens->inner_l2_len += sizeof(struct rte_vlan_hdr);
> proto = vh->eth_proto;
> } else if (proto == rte_cpu_to_be_16(RTE_ETHER_TYPE_QINQ)) {
> const struct rte_vlan_hdr *vh;
> @@ -524,7 +524,7 @@ uint32_t rte_net_get_ptype(const struct rte_mbuf *m,
> if (unlikely(vh == NULL))
> return pkt_type;
> off += 2 * sizeof(*vh);
> - hdr_lens->inner_l2_len += 2 * sizeof(*vh);
> + hdr_lens->inner_l2_len += 2 * sizeof(struct rte_vlan_hdr);
> proto = vh->eth_proto;
> }
>
> --
> 2.47.3
>
>
next prev parent reply other threads:[~2025-07-28 13:28 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-07-28 13:12 Khadem Ullah
2025-07-28 13:28 ` Ivan Malov [this message]
2025-07-29 5:06 ` Khadem Ullah
-- strict thread matches above, loose matches on Subject: below --
2025-07-28 12:30 Khadem Ullah
2025-07-28 12:21 Khadem Ullah
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=e7c3c2b2-5f7a-eec4-cbd8-cd17ecec4c46@arknetworks.am \
--to=ivan.malov@arknetworks.am \
--cc=14pwcse1224@uetpeshawar.edu.pk \
--cc=andrew.rybchenko@oktetlabs.ru \
--cc=bruce.richardson@intel.com \
--cc=dev@dpdk.org \
--cc=jasvinder.singh@intel.com \
--cc=stable@dpdk.org \
--cc=thomas@monjalon.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).