bnxt: fix receive VLAN offload flags

Message ID 20190325212754.13786-1-stephen@networkplumber.org
State Accepted, archived
Delegated to: Ferruh Yigit
Headers show
Series
  • bnxt: fix receive VLAN offload flags
Related show

Checks

Context Check Description
ci/mellanox-Performance-Testing success Performance Testing PASS
ci/intel-Performance-Testing success Performance Testing PASS
ci/Intel-compilation success Compilation OK
ci/checkpatch success coding style OK

Commit Message

Stephen Hemminger March 25, 2019, 9:27 p.m.
From: Stephen Hemminger <sthemmin@microsoft.com>

The bnxt driver is not correctly setting the receive VLAN offload
flags. When VLAN is offloaded the driver must set the PKT_RX_VLAN_STRIPPED
flag.

Actually, several drivers have the same bug, only most of the
Intel drivers look right. Any driver that sets vlan_tci is probably
stripping the tag, and should be setting RX_VLAN_STRIPPED.

To quote rte_mbuf.h:

/**
 * The RX packet is a 802.1q VLAN packet, and the tci has been
 * saved in in mbuf->vlan_tci.
 * If the flag PKT_RX_VLAN_STRIPPED is also present, the VLAN
 * header has been stripped from mbuf data, else it is still
 * present.
 */

Fixes: 2eb53b134aae ("net/bnxt: add initial Rx code")
Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>
---
 drivers/net/bnxt/bnxt_rxr.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

Comments

Ajit Khaparde March 25, 2019, 9:34 p.m. | #1
On Mon, Mar 25, 2019 at 2:28 PM Stephen Hemminger <
stephen@networkplumber.org> wrote:

> From: Stephen Hemminger <sthemmin@microsoft.com>
>
> The bnxt driver is not correctly setting the receive VLAN offload
> flags. When VLAN is offloaded the driver must set the PKT_RX_VLAN_STRIPPED
> flag.
>
> Actually, several drivers have the same bug, only most of the
> Intel drivers look right. Any driver that sets vlan_tci is probably
> stripping the tag, and should be setting RX_VLAN_STRIPPED.
>
> To quote rte_mbuf.h:
>
> /**
>  * The RX packet is a 802.1q VLAN packet, and the tci has been
>  * saved in in mbuf->vlan_tci.
>  * If the flag PKT_RX_VLAN_STRIPPED is also present, the VLAN
>  * header has been stripped from mbuf data, else it is still
>  * present.
>  */
>
> Fixes: 2eb53b134aae ("net/bnxt: add initial Rx code")
> Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>
>
Acked-by: Ajit Khaparde <ajit.khaparde@broadcom.com>


> ---
>  drivers/net/bnxt/bnxt_rxr.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/net/bnxt/bnxt_rxr.c b/drivers/net/bnxt/bnxt_rxr.c
> index 1bfc63d9304c..dc695e17771b 100644
> --- a/drivers/net/bnxt/bnxt_rxr.c
> +++ b/drivers/net/bnxt/bnxt_rxr.c
> @@ -154,7 +154,7 @@ static void bnxt_tpa_start(struct bnxt_rx_queue *rxq,
>         if (tpa_start1->flags2 &
>             rte_cpu_to_le_32(RX_TPA_START_CMPL_FLAGS2_META_FORMAT_VLAN)) {
>                 mbuf->vlan_tci = rte_le_to_cpu_32(tpa_start1->metadata);
> -               mbuf->ol_flags |= PKT_RX_VLAN;
> +               mbuf->ol_flags |= PKT_RX_VLAN | PKT_RX_VLAN_STRIPPED;
>         }
>         if (likely(tpa_start1->flags2 &
>                    rte_cpu_to_le_32(RX_TPA_START_CMPL_FLAGS2_L4_CS_CALC)))
> @@ -437,7 +437,7 @@ static int bnxt_rx_pkt(struct rte_mbuf **rx_pkt,
>                         (RX_PKT_CMPL_METADATA_VID_MASK |
>                         RX_PKT_CMPL_METADATA_DE |
>                         RX_PKT_CMPL_METADATA_PRI_MASK);
> -               mbuf->ol_flags |= PKT_RX_VLAN;
> +               mbuf->ol_flags |= PKT_RX_VLAN | PKT_RX_VLAN_STRIPPED;
>         }
>
>         if (likely(RX_CMP_IP_CS_OK(rxcmp1)))
> --
> 2.17.1
>
>
Ferruh Yigit March 26, 2019, 5:37 p.m. | #2
On 3/25/2019 9:34 PM, Ajit Khaparde wrote:
> On Mon, Mar 25, 2019 at 2:28 PM Stephen Hemminger <
> stephen@networkplumber.org> wrote:
> 
>> From: Stephen Hemminger <sthemmin@microsoft.com>
>>
>> The bnxt driver is not correctly setting the receive VLAN offload
>> flags. When VLAN is offloaded the driver must set the PKT_RX_VLAN_STRIPPED
>> flag.
>>
>> Actually, several drivers have the same bug, only most of the
>> Intel drivers look right. Any driver that sets vlan_tci is probably
>> stripping the tag, and should be setting RX_VLAN_STRIPPED.
>>
>> To quote rte_mbuf.h:
>>
>> /**
>>  * The RX packet is a 802.1q VLAN packet, and the tci has been
>>  * saved in in mbuf->vlan_tci.
>>  * If the flag PKT_RX_VLAN_STRIPPED is also present, the VLAN
>>  * header has been stripped from mbuf data, else it is still
>>  * present.
>>  */
>>
>> Fixes: 2eb53b134aae ("net/bnxt: add initial Rx code")
>> Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>
>>
> Acked-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

Applied to dpdk-next-net/master, thanks.


In a quick glance, following PMDs as well set "PKT_RX_VLAN" but not
"PKT_RX_VLAN_STRIPPED", their maintainers are cc'ed, can you please double check:

- bnx2x
- qede (it doesn't set 'vlan_tci' if not stripped, should be fixed)
- dpaa2
- dpaa (is 'vlan_tci' set?)
- avp
- fm10k

Patch

diff --git a/drivers/net/bnxt/bnxt_rxr.c b/drivers/net/bnxt/bnxt_rxr.c
index 1bfc63d9304c..dc695e17771b 100644
--- a/drivers/net/bnxt/bnxt_rxr.c
+++ b/drivers/net/bnxt/bnxt_rxr.c
@@ -154,7 +154,7 @@  static void bnxt_tpa_start(struct bnxt_rx_queue *rxq,
 	if (tpa_start1->flags2 &
 	    rte_cpu_to_le_32(RX_TPA_START_CMPL_FLAGS2_META_FORMAT_VLAN)) {
 		mbuf->vlan_tci = rte_le_to_cpu_32(tpa_start1->metadata);
-		mbuf->ol_flags |= PKT_RX_VLAN;
+		mbuf->ol_flags |= PKT_RX_VLAN | PKT_RX_VLAN_STRIPPED;
 	}
 	if (likely(tpa_start1->flags2 &
 		   rte_cpu_to_le_32(RX_TPA_START_CMPL_FLAGS2_L4_CS_CALC)))
@@ -437,7 +437,7 @@  static int bnxt_rx_pkt(struct rte_mbuf **rx_pkt,
 			(RX_PKT_CMPL_METADATA_VID_MASK |
 			RX_PKT_CMPL_METADATA_DE |
 			RX_PKT_CMPL_METADATA_PRI_MASK);
-		mbuf->ol_flags |= PKT_RX_VLAN;
+		mbuf->ol_flags |= PKT_RX_VLAN | PKT_RX_VLAN_STRIPPED;
 	}
 
 	if (likely(RX_CMP_IP_CS_OK(rxcmp1)))