[PATCH net] net: airoha: Fix max TX packet length configuration

[PATCH net] net: airoha: Fix max TX packet length configuration

[Lorenzo Bianconi]

According to the Airoha documentation, REG_GDM_LEN_CFG() register does not
include FCS length. Fix MTU configuration removing ETH_FCS_LEN from
maximum TX packet length configuration.

Fixes: 54d989d58d2ac ("net: airoha: Move min/max packet len configuration in airoha_dev_open()")
Fixes: 03b1b69f0662c ("net: airoha: Introduce airoha_dev_change_mtu callback")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
---
 drivers/net/ethernet/airoha/airoha_eth.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/airoha/airoha_eth.c b/drivers/net/ethernet/airoha/airoha_eth.c
index c14cdce588a7..a81ffda72b39 100644
--- a/drivers/net/ethernet/airoha/airoha_eth.c
+++ b/drivers/net/ethernet/airoha/airoha_eth.c
@@ -1629,9 +1629,9 @@ static void airoha_update_hw_stats(struct airoha_gdm_port *port)
 
 static int airoha_dev_open(struct net_device *dev)
 {
-	int err, len = ETH_HLEN + dev->mtu + ETH_FCS_LEN;
 	struct airoha_gdm_port *port = netdev_priv(dev);
 	struct airoha_qdma *qdma = port->qdma;
+	int err, len = ETH_HLEN + dev->mtu;
 
 	netif_tx_start_all_queues(dev);
 	err = airoha_set_vip_for_gdm_port(port, true);
@@ -1833,7 +1833,7 @@ static int airoha_dev_change_mtu(struct net_device *dev, int mtu)
 {
 	struct airoha_gdm_port *port = netdev_priv(dev);
 	struct airoha_eth *eth = port->qdma->eth;
-	u32 len = ETH_HLEN + mtu + ETH_FCS_LEN;
+	u32 len = ETH_HLEN + mtu;
 
 	airoha_fe_rmw(eth, REG_GDM_LEN_CFG(port->id),
 		      GDM_LONG_LEN_MASK,

---
base-commit: 02f72964395911e7a09bb2ea2fe6f79eda4ea2c2
change-id: 20260412-airoha-fix-max-mtu-f9c7823ce2a2

Best regards,
-- 
Lorenzo Bianconi <lorenzo@kernel.org>

Re: [PATCH net] net: airoha: Fix max TX packet length configuration

[Simon Horman]

From: 'Simon Horman' <horms@kernel.org>

This is an AI-generated review of your patch. The human sending this
email has considered the AI review valid, or at least plausible.
Full review at: https://sashiko.dev
---
net: airoha: Fix max TX packet length configuration

This commit fixes the MTU configuration for the Airoha driver by removing
ETH_FCS_LEN from the maximum TX packet length, matching the hardware
documentation which states the register does not include the FCS length.

> diff --git a/drivers/net/ethernet/airoha/airoha_eth.c b/drivers/net/ethernet/airoha/airoha_eth.c
> index c14cdce588a7c..a81ffda72b39f 100644
> --- a/drivers/net/ethernet/airoha/airoha_eth.c
> +++ b/drivers/net/ethernet/airoha/airoha_eth.c
> @@ -1629,9 +1629,9 @@ static void airoha_update_hw_stats(struct airoha_gdm_port *port)
>  
>  static int airoha_dev_open(struct net_device *dev)
>  {
> -	int err, len = ETH_HLEN + dev->mtu + ETH_FCS_LEN;
>  	struct airoha_gdm_port *port = netdev_priv(dev);
>  	struct airoha_qdma *qdma = port->qdma;
> +	int err, len = ETH_HLEN + dev->mtu;

Does this regression cause MTU-sized VLAN-tagged frames to be dropped by the
hardware?

By limiting the frame size to ETH_HLEN + dev->mtu, the maximum allowed length
becomes 1514 bytes for a standard 1500 MTU. An Ethernet frame containing a
standard 802.1Q VLAN tag and a maximum-sized 1500-byte IP payload requires an
18-byte L2 header, making the total frame size 1518 bytes (excluding FCS).

Previously, the inclusion of ETH_FCS_LEN inadvertently provided the 4-byte
headroom needed to accommodate a single VLAN tag. 

Could the configuration explicitly account for VLAN headers using VLAN_ETH_HLEN
instead?

[ ... ]

Re: [PATCH net] net: airoha: Fix max TX packet length configuration

[Lorenzo Bianconi]

[-- Attachment #1: Type: text/plain, Size: 1888 bytes --]

> From: 'Simon Horman' <horms@kernel.org>
> 
> This is an AI-generated review of your patch. The human sending this
> email has considered the AI review valid, or at least plausible.
> Full review at: https://sashiko.dev
> ---
> net: airoha: Fix max TX packet length configuration
> 
> This commit fixes the MTU configuration for the Airoha driver by removing
> ETH_FCS_LEN from the maximum TX packet length, matching the hardware
> documentation which states the register does not include the FCS length.
> 
> > diff --git a/drivers/net/ethernet/airoha/airoha_eth.c b/drivers/net/ethernet/airoha/airoha_eth.c
> > index c14cdce588a7c..a81ffda72b39f 100644
> > --- a/drivers/net/ethernet/airoha/airoha_eth.c
> > +++ b/drivers/net/ethernet/airoha/airoha_eth.c
> > @@ -1629,9 +1629,9 @@ static void airoha_update_hw_stats(struct airoha_gdm_port *port)
> >  
> >  static int airoha_dev_open(struct net_device *dev)
> >  {
> > -	int err, len = ETH_HLEN + dev->mtu + ETH_FCS_LEN;
> >  	struct airoha_gdm_port *port = netdev_priv(dev);
> >  	struct airoha_qdma *qdma = port->qdma;
> > +	int err, len = ETH_HLEN + dev->mtu;
> 
> Does this regression cause MTU-sized VLAN-tagged frames to be dropped by the
> hardware?
> 
> By limiting the frame size to ETH_HLEN + dev->mtu, the maximum allowed length
> becomes 1514 bytes for a standard 1500 MTU. An Ethernet frame containing a
> standard 802.1Q VLAN tag and a maximum-sized 1500-byte IP payload requires an
> 18-byte L2 header, making the total frame size 1518 bytes (excluding FCS).
> 
> Previously, the inclusion of ETH_FCS_LEN inadvertently provided the 4-byte
> headroom needed to accommodate a single VLAN tag. 
> 
> Could the configuration explicitly account for VLAN headers using VLAN_ETH_HLEN
> instead?

ack, thx for pointing this out. I will fix it in v2.

Regards,
Lorenzo

> 
> [ ... ]

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: [PATCH net] net: airoha: Fix max TX packet length configuration

[Paolo Abeni]

On 4/12/26 10:09 AM, Lorenzo Bianconi wrote:
> According to the Airoha documentation, REG_GDM_LEN_CFG() register does not
> include FCS length. Fix MTU configuration removing ETH_FCS_LEN from
> maximum TX packet length configuration.
> 
> Fixes: 54d989d58d2ac ("net: airoha: Move min/max packet len configuration in airoha_dev_open()")
> Fixes: 03b1b69f0662c ("net: airoha: Introduce airoha_dev_change_mtu callback")
> Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
> ---
>  drivers/net/ethernet/airoha/airoha_eth.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/net/ethernet/airoha/airoha_eth.c b/drivers/net/ethernet/airoha/airoha_eth.c
> index c14cdce588a7..a81ffda72b39 100644
> --- a/drivers/net/ethernet/airoha/airoha_eth.c
> +++ b/drivers/net/ethernet/airoha/airoha_eth.c
> @@ -1629,9 +1629,9 @@ static void airoha_update_hw_stats(struct airoha_gdm_port *port)
>  
>  static int airoha_dev_open(struct net_device *dev)
>  {
> -	int err, len = ETH_HLEN + dev->mtu + ETH_FCS_LEN;
>  	struct airoha_gdm_port *port = netdev_priv(dev);
>  	struct airoha_qdma *qdma = port->qdma;
> +	int err, len = ETH_HLEN + dev->mtu;

Sashiko noted that the above may cause regressions, dropping max MTU
vlan frames.

/P