[PATCH net-next v3] igc: offload queue max SDU from tc-taprio

[PATCH net-next v3] igc: offload queue max SDU from tc-taprio

[Muhammad Husaini Zulkifli]

From: Tan Tee Min <tee.min.tan@linux.intel.com>

Add support for configuring the max SDU for each Tx queue.
If not specified, keep the default.

Signed-off-by: Tan Tee Min <tee.min.tan@linux.intel.com>
Signed-off-by: Muhammad Husaini Zulkifli <muhammad.husaini.zulkifli@intel.com>
---
 drivers/net/ethernet/intel/igc/igc.h      |  1 +
 drivers/net/ethernet/intel/igc/igc_main.c | 26 +++++++++++++++++++++++
 2 files changed, 27 insertions(+)

diff --git a/drivers/net/ethernet/intel/igc/igc.h b/drivers/net/ethernet/intel/igc/igc.h
index 9db93c1f97679..34aebf00a5123 100644
--- a/drivers/net/ethernet/intel/igc/igc.h
+++ b/drivers/net/ethernet/intel/igc/igc.h
@@ -99,6 +99,7 @@ struct igc_ring {
 
 	u32 start_time;
 	u32 end_time;
+	u32 max_sdu;
 
 	/* CBS parameters */
 	bool cbs_enable;                /* indicates if CBS is enabled */
diff --git a/drivers/net/ethernet/intel/igc/igc_main.c b/drivers/net/ethernet/intel/igc/igc_main.c
index 0cc327294dfb5..38ad437957ada 100644
--- a/drivers/net/ethernet/intel/igc/igc_main.c
+++ b/drivers/net/ethernet/intel/igc/igc_main.c
@@ -1508,6 +1508,7 @@ static netdev_tx_t igc_xmit_frame_ring(struct sk_buff *skb,
 	__le32 launch_time = 0;
 	u32 tx_flags = 0;
 	unsigned short f;
+	u32 max_sdu = 0;
 	ktime_t txtime;
 	u8 hdr_len = 0;
 	int tso = 0;
@@ -1527,6 +1528,14 @@ static netdev_tx_t igc_xmit_frame_ring(struct sk_buff *skb,
 		return NETDEV_TX_BUSY;
 	}
 
+	if (tx_ring->max_sdu > 0) {
+		max_sdu = tx_ring->max_sdu +
+			  (skb_vlan_tagged(skb) ? VLAN_HLEN : 0);
+
+		if (skb->len > max_sdu)
+			goto skb_drop;
+	}
+
 	if (!tx_ring->launchtime_enable)
 		goto done;
 
@@ -1606,6 +1615,11 @@ static netdev_tx_t igc_xmit_frame_ring(struct sk_buff *skb,
 	dev_kfree_skb_any(first->skb);
 	first->skb = NULL;
 
+	return NETDEV_TX_OK;
+
+skb_drop:
+	dev_kfree_skb_any(skb);
+
 	return NETDEV_TX_OK;
 }
 
@@ -6039,6 +6053,7 @@ static int igc_tsn_clear_schedule(struct igc_adapter *adapter)
 
 		ring->start_time = 0;
 		ring->end_time = NSEC_PER_SEC;
+		ring->max_sdu = 0;
 	}
 
 	return 0;
@@ -6122,6 +6137,16 @@ static int igc_save_qbv_schedule(struct igc_adapter *adapter,
 		}
 	}
 
+	for (i = 0; i < adapter->num_tx_queues; i++) {
+		struct igc_ring *ring = adapter->tx_ring[i];
+		struct net_device *dev = adapter->netdev;
+
+		if (qopt->max_sdu[i])
+			ring->max_sdu = qopt->max_sdu[i] + dev->hard_header_len;
+		else
+			ring->max_sdu = 0;
+	}
+
 	return 0;
 }
 
@@ -6221,6 +6246,7 @@ static int igc_tc_query_caps(struct igc_adapter *adapter,
 		if (hw->mac.type != igc_i225)
 			return -EOPNOTSUPP;
 
+		caps->supports_queue_max_sdu = true;
 		caps->gate_mask_per_txq = true;
 
 		return 0;
-- 
2.17.1

Re: [PATCH net-next v3] igc: offload queue max SDU from tc-taprio

[Jakub Kicinski]

On Wed,  8 Feb 2023 08:33:27 +0800 Muhammad Husaini Zulkifli wrote:
> From: Tan Tee Min <tee.min.tan@linux.intel.com>
> 
> Add support for configuring the max SDU for each Tx queue.
> If not specified, keep the default.

> diff --git a/drivers/net/ethernet/intel/igc/igc_main.c b/drivers/net/ethernet/intel/igc/igc_main.c
> index 0cc327294dfb5..38ad437957ada 100644
> --- a/drivers/net/ethernet/intel/igc/igc_main.c
> +++ b/drivers/net/ethernet/intel/igc/igc_main.c
> @@ -1508,6 +1508,7 @@ static netdev_tx_t igc_xmit_frame_ring(struct sk_buff *skb,
>  	__le32 launch_time = 0;
>  	u32 tx_flags = 0;
>  	unsigned short f;
> +	u32 max_sdu = 0;

This variable can be moved to the scope of the if() ?

>  	ktime_t txtime;
>  	u8 hdr_len = 0;
>  	int tso = 0;
> @@ -1527,6 +1528,14 @@ static netdev_tx_t igc_xmit_frame_ring(struct sk_buff *skb,
>  		return NETDEV_TX_BUSY;
>  	}
>  
> +	if (tx_ring->max_sdu > 0) {
> +		max_sdu = tx_ring->max_sdu +
> +			  (skb_vlan_tagged(skb) ? VLAN_HLEN : 0);
> +
> +		if (skb->len > max_sdu)

You should increment some counter here. Otherwise it's a silent discard.

> +			goto skb_drop;
> +	}
> +
>  	if (!tx_ring->launchtime_enable)
>  		goto done;
>  
> @@ -1606,6 +1615,11 @@ static netdev_tx_t igc_xmit_frame_ring(struct sk_buff *skb,
>  	dev_kfree_skb_any(first->skb);

first->skb is skb, as far as I can tell, you can reshuffle this code to
avoid adding the new return flow.

>  	first->skb = NULL;
>  
> +	return NETDEV_TX_OK;
> +
> +skb_drop:
> +	dev_kfree_skb_any(skb);
> +
>  	return NETDEV_TX_OK;
>  }

> @@ -6122,6 +6137,16 @@ static int igc_save_qbv_schedule(struct igc_adapter *adapter,
>  		}
>  	}
>  
> +	for (i = 0; i < adapter->num_tx_queues; i++) {
> +		struct igc_ring *ring = adapter->tx_ring[i];
> +		struct net_device *dev = adapter->netdev;
> +
> +		if (qopt->max_sdu[i])
> +			ring->max_sdu = qopt->max_sdu[i] + dev->hard_header_len;

why hard_header_len? Isn't it always ETH_HLEN?

> +		else
> +			ring->max_sdu = 0;

RE: [PATCH net-next v3] igc: offload queue max SDU from tc-taprio

[Zulkifli, Muhammad Husaini]

Hi Jakub,

> -----Original Message-----
> From: Jakub Kicinski <kuba@kernel.org>
> Sent: Thursday, 9 February, 2023 1:30 PM
> To: Zulkifli, Muhammad Husaini <muhammad.husaini.zulkifli@intel.com>
> Cc: intel-wired-lan@osuosl.org; Gomes, Vinicius
> <vinicius.gomes@intel.com>; naamax.meir@linux.intel.com; Nguyen,
> Anthony L <anthony.l.nguyen@intel.com>; leon@kernel.org;
> davem@davemloft.net; pabeni@redhat.com; edumazet@google.com;
> tee.min.tan@linux.intel.com; netdev@vger.kernel.org; Neftin, Sasha
> <sasha.neftin@intel.com>
> Subject: Re: [PATCH net-next v3] igc: offload queue max SDU from tc-taprio
> 
> On Wed,  8 Feb 2023 08:33:27 +0800 Muhammad Husaini Zulkifli wrote:
> > From: Tan Tee Min <tee.min.tan@linux.intel.com>
> >
> > Add support for configuring the max SDU for each Tx queue.
> > If not specified, keep the default.
> 
> > diff --git a/drivers/net/ethernet/intel/igc/igc_main.c
> b/drivers/net/ethernet/intel/igc/igc_main.c
> > index 0cc327294dfb5..38ad437957ada 100644
> > --- a/drivers/net/ethernet/intel/igc/igc_main.c
> > +++ b/drivers/net/ethernet/intel/igc/igc_main.c
> > @@ -1508,6 +1508,7 @@ static netdev_tx_t igc_xmit_frame_ring(struct
> sk_buff *skb,
> >  	__le32 launch_time = 0;
> >  	u32 tx_flags = 0;
> >  	unsigned short f;
> > +	u32 max_sdu = 0;
> 
> This variable can be moved to the scope of the if() ?

Sure. Let me move to the if() scope in v5. 

> 
> >  	ktime_t txtime;
> >  	u8 hdr_len = 0;
> >  	int tso = 0;
> > @@ -1527,6 +1528,14 @@ static netdev_tx_t igc_xmit_frame_ring(struct
> sk_buff *skb,
> >  		return NETDEV_TX_BUSY;
> >  	}
> >
> > +	if (tx_ring->max_sdu > 0) {
> > +		max_sdu = tx_ring->max_sdu +
> > +			  (skb_vlan_tagged(skb) ? VLAN_HLEN : 0);
> > +
> > +		if (skb->len > max_sdu)
> 
> You should increment some counter here. Otherwise it's a silent discard.

I am thinking to use tx_dropped counters for this. Is it ok?

> 
> > +			goto skb_drop;
> > +	}
> > +
> >  	if (!tx_ring->launchtime_enable)
> >  		goto done;
> >
> > @@ -1606,6 +1615,11 @@ static netdev_tx_t igc_xmit_frame_ring(struct
> sk_buff *skb,
> >  	dev_kfree_skb_any(first->skb);
> 
> first->skb is skb, as far as I can tell, you can reshuffle this code to
> avoid adding the new return flow.

What we try to do is to check the current max_sdu size at the
beginning stage of the func() and drop it quickly.

> 
> >  	first->skb = NULL;
> >
> > +	return NETDEV_TX_OK;
> > +
> > +skb_drop:
> > +	dev_kfree_skb_any(skb);
> > +
> >  	return NETDEV_TX_OK;
> >  }
> 
> > @@ -6122,6 +6137,16 @@ static int igc_save_qbv_schedule(struct
> igc_adapter *adapter,
> >  		}
> >  	}
> >
> > +	for (i = 0; i < adapter->num_tx_queues; i++) {
> > +		struct igc_ring *ring = adapter->tx_ring[i];
> > +		struct net_device *dev = adapter->netdev;
> > +
> > +		if (qopt->max_sdu[i])
> > +			ring->max_sdu = qopt->max_sdu[i] + dev-
> >hard_header_len;
> 
> why hard_header_len? Isn't it always ETH_HLEN?

We followed the taprio_parse_tc_entries() implementation for this.
But the hard_header_len should be same value as ETH_HLEN.

> 
> > +		else
> > +			ring->max_sdu = 0;

Re: [PATCH net-next v3] igc: offload queue max SDU from tc-taprio

[Jakub Kicinski]

On Tue, 14 Feb 2023 06:27:17 +0000 Zulkifli, Muhammad Husaini wrote:
> > > +	if (tx_ring->max_sdu > 0) {
> > > +		max_sdu = tx_ring->max_sdu +
> > > +			  (skb_vlan_tagged(skb) ? VLAN_HLEN : 0);
> > > +
> > > +		if (skb->len > max_sdu)  
> > 
> > You should increment some counter here. Otherwise it's a silent discard.  
> 
> I am thinking to use tx_dropped counters for this. Is it ok?

Yup!

> > > +			goto skb_drop;
> > > +	}
> > > +
> > >  	if (!tx_ring->launchtime_enable)
> > >  		goto done;
> > >
> > > @@ -1606,6 +1615,11 @@ static netdev_tx_t igc_xmit_frame_ring(struct sk_buff *skb,  
> > >  	dev_kfree_skb_any(first->skb);  
> > 
> > first->skb is skb, as far as I can tell, you can reshuffle this code to
> > avoid adding the new return flow.  
> 
> What we try to do is to check the current max_sdu size at the
> beginning stage of the func() and drop it quickly.

I understand, what I'm saying is that the code which is already here
can be reused, it currently operates on first->skb but it can as well
use skb. And then you'll be able to jump to the same statement, rather
than have to create a separate return.

RE: [PATCH net-next v3] igc: offload queue max SDU from tc-taprio

[Zulkifli, Muhammad Husaini]

> -----Original Message-----
> From: Jakub Kicinski <kuba@kernel.org>
> Sent: Tuesday, 14 February, 2023 2:40 PM
> To: Zulkifli, Muhammad Husaini <muhammad.husaini.zulkifli@intel.com>
> Cc: intel-wired-lan@osuosl.org; Gomes, Vinicius <vinicius.gomes@intel.com>;
> naamax.meir@linux.intel.com; Nguyen, Anthony L
> <anthony.l.nguyen@intel.com>; leon@kernel.org; davem@davemloft.net;
> pabeni@redhat.com; edumazet@google.com; tee.min.tan@linux.intel.com;
> netdev@vger.kernel.org; Neftin, Sasha <sasha.neftin@intel.com>
> Subject: Re: [PATCH net-next v3] igc: offload queue max SDU from tc-taprio
> 
> On Tue, 14 Feb 2023 06:27:17 +0000 Zulkifli, Muhammad Husaini wrote:
> > > > +	if (tx_ring->max_sdu > 0) {
> > > > +		max_sdu = tx_ring->max_sdu +
> > > > +			  (skb_vlan_tagged(skb) ? VLAN_HLEN : 0);
> > > > +
> > > > +		if (skb->len > max_sdu)
> > >
> > > You should increment some counter here. Otherwise it's a silent discard.
> >
> > I am thinking to use tx_dropped counters for this. Is it ok?
> 
> Yup!
> 
> > > > +			goto skb_drop;
> > > > +	}
> > > > +
> > > >  	if (!tx_ring->launchtime_enable)
> > > >  		goto done;
> > > >
> > > > @@ -1606,6 +1615,11 @@ static netdev_tx_t
> igc_xmit_frame_ring(struct sk_buff *skb,
> > > >  	dev_kfree_skb_any(first->skb);
> > >
> > > first->skb is skb, as far as I can tell, you can reshuffle this code
> > > first->to
> > > avoid adding the new return flow.
> >
> > What we try to do is to check the current max_sdu size at the
> > beginning stage of the func() and drop it quickly.
> 
> I understand, what I'm saying is that the code which is already here can be
> reused, it currently operates on first->skb but it can as well use skb. And then
> you'll be able to jump to the same statement, rather than have to create a
> separate return.

Sure. Thanks