Skip to contentSkip to conversation

i40e: Fix the inability to attach XDP program on downed interface

PATCH Superseded Mateusz Palczewski 4 participants 4 messages Dec 5, 2022 – Dec 6, 2022

Subsystems (2)

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 i40e: Fix the inability to attach XDP program on downed interface

From: "Staszewski, BartoszX" <[email protected]>

Whenever interface was down, function
i40e_xdp was passing vsi->rx_buf_len field
to i40e_xdp_setup() which was equal 0.
i40e_open() calls i40e_vsi_configure_rx()
which configures that field, but that only
happens when interface is up. When it is
down, i40e_open() is not being called, thus
vsi->rx_buf_len is not set.

Solution for this is calculate buffer length
in newly created function - i40e_calculate_vsi_rx_buf_len() that return actual buffer length. Buffer length is being calculated based on the same rules applied previously in i40e_vsi_configure_rx() function.

Fixes 613142b0bb88 ("i40e: Log error for oversized MTU on device")
Fixes 0c8493d90b6b ("i40e: add XDP support for pass and drop actions")

---
v2: Fixed commit title and message, this patch is only for a case of
fresh start so I believe there is no need for rx_buf_len clearance

---
 drivers/net/ethernet/intel/i40e/i40e_main.c | 42 +++++++++++++++------
 1 file changed, 30 insertions(+), 12 deletions(-)
drivers/net/ethernet/intel/i40e/i40e_main.c +30 -12
index b5dcd15ced36..2fec0cff282c 100644
@@ -3693,6 +3693,30 @@ static int i40e_vsi_configure_tx(struct i40e_vsi *vsi)
 	return err;
 }
 
+/**
+ * i40e_calculate_vsi_rx_buf_len - Calculates buffer length
+ *
+ * @vsi: VSI to calculate rx_buf_len from
+ */
+static u16 i40e_calculate_vsi_rx_buf_len(struct i40e_vsi *vsi)
+{
+	u16 ret;
+
+	if (!vsi->netdev || (vsi->back->flags & I40E_FLAG_LEGACY_RX)) {
+		ret = I40E_RXBUFFER_2048;
+#if (PAGE_SIZE < 8192)
+	} else if (!I40E_2K_TOO_SMALL_WITH_PADDING &&
+		   (vsi->netdev->mtu <= ETH_DATA_LEN)) {
+		ret = I40E_RXBUFFER_1536 - NET_IP_ALIGN;
+#endif
+	} else {
+		ret = (PAGE_SIZE < 8192) ? I40E_RXBUFFER_3072 :
+					   I40E_RXBUFFER_2048;
+	}
+
+	return ret;
+}
+
 /**
  * i40e_vsi_configure_rx - Configure the VSI for Rx
  * @vsi: the VSI being configured
@@ -3704,20 +3728,14 @@ static int i40e_vsi_configure_rx(struct i40e_vsi *vsi)
 	int err = 0;
 	u16 i;
 
-	if (!vsi->netdev || (vsi->back->flags & I40E_FLAG_LEGACY_RX)) {
-		vsi->max_frame = I40E_MAX_RXBUFFER;
-		vsi->rx_buf_len = I40E_RXBUFFER_2048;
+	vsi->max_frame = I40E_MAX_RXBUFFER;
+	vsi->rx_buf_len = i40e_calculate_vsi_rx_buf_len(vsi);
+
 #if (PAGE_SIZE < 8192)
-	} else if (!I40E_2K_TOO_SMALL_WITH_PADDING &&
-		   (vsi->netdev->mtu <= ETH_DATA_LEN)) {
+	if (vsi->netdev && !I40E_2K_TOO_SMALL_WITH_PADDING &&
+	    vsi->netdev->mtu <= ETH_DATA_LEN)
 		vsi->max_frame = I40E_RXBUFFER_1536 - NET_IP_ALIGN;
-		vsi->rx_buf_len = I40E_RXBUFFER_1536 - NET_IP_ALIGN;
 #endif
-	} else {
-		vsi->max_frame = I40E_MAX_RXBUFFER;
-		vsi->rx_buf_len = (PAGE_SIZE < 8192) ? I40E_RXBUFFER_3072 :
-						       I40E_RXBUFFER_2048;
-	}
 
 	/* set up individual rings */
 	for (i = 0; i < vsi->num_queue_pairs && !err; i++)
@@ -13265,7 +13283,7 @@ static int i40e_xdp_setup(struct i40e_vsi *vsi, struct bpf_prog *prog,
 	int i;
 
 	/* Don't allow frames that span over multiple buffers */
-	if (frame_size > vsi->rx_buf_len) {
+	if (frame_size > i40e_calculate_vsi_rx_buf_len(vsi)) {
 		NL_SET_ERR_MSG_MOD(extack, "MTU too large to enable XDP");
 		return -EINVAL;
 	}
Signature · 6 lines
2.31.1

_______________________________________________
Intel-wired-lan mailing list
[email protected]
https://lists.osuosl.org/mailman/listinfo/intel-wired-lan
maintainer
#1 lore raw

+ Maciej, Bjorn, and Magnus as I presume this is the update to this patch

https://lore.kernel.org/netdev/Y3OFmgLa56rwVQ4j@boxer/

Quoted 1 line · Mateusz Palczewski

From: "Staszewski, BartoszX" <[email protected]>

This is incorrect; first name then last name, no 'X'.

Quoted 88 lines

Whenever interface was down, function
i40e_xdp was passing vsi->rx_buf_len field
to i40e_xdp_setup() which was equal 0.
i40e_open() calls i40e_vsi_configure_rx()
which configures that field, but that only
happens when interface is up. When it is
down, i40e_open() is not being called, thus
vsi->rx_buf_len is not set.

Solution for this is calculate buffer length
in newly created function - i40e_calculate_vsi_rx_buf_len() that return actual buffer length. Buffer length is being calculated based on the same rules applied previously in i40e_vsi_configure_rx() function.

Fixes 613142b0bb88 ("i40e: Log error for oversized MTU on device")
Fixes 0c8493d90b6b ("i40e: add XDP support for pass and drop actions")

---
v2: Fixed commit title and message, this patch is only for a case of
fresh start so I believe there is no need for rx_buf_len clearance

---
  drivers/net/ethernet/intel/i40e/i40e_main.c | 42 +++++++++++++++------
  1 file changed, 30 insertions(+), 12 deletions(-)
drivers/net/ethernet/intel/i40e/i40e_main.c +29 -11
index b5dcd15ced36..2fec0cff282c 100644
@@ -3693,6 +3693,30 @@ static int i40e_vsi_configure_tx(struct i40e_vsi *vsi)
  	return err;
  }
 
+/**
+ * i40e_calculate_vsi_rx_buf_len - Calculates buffer length
+ *
+ * @vsi: VSI to calculate rx_buf_len from
+ */
+static u16 i40e_calculate_vsi_rx_buf_len(struct i40e_vsi *vsi)
+{
+	u16 ret;
+
+	if (!vsi->netdev || (vsi->back->flags & I40E_FLAG_LEGACY_RX)) {
+		ret = I40E_RXBUFFER_2048;
+#if (PAGE_SIZE < 8192)
+	} else if (!I40E_2K_TOO_SMALL_WITH_PADDING &&
+		   (vsi->netdev->mtu <= ETH_DATA_LEN)) {
+		ret = I40E_RXBUFFER_1536 - NET_IP_ALIGN;
+#endif
+	} else {
+		ret = (PAGE_SIZE < 8192) ? I40E_RXBUFFER_3072 :
+					   I40E_RXBUFFER_2048;
+	}
+
+	return ret;
+}
+
  /**
   * i40e_vsi_configure_rx - Configure the VSI for Rx
   * @vsi: the VSI being configured
@@ -3704,20 +3728,14 @@ static int i40e_vsi_configure_rx(struct i40e_vsi *vsi)
  	int err = 0;
  	u16 i;
 
-	if (!vsi->netdev || (vsi->back->flags & I40E_FLAG_LEGACY_RX)) {
-		vsi->max_frame = I40E_MAX_RXBUFFER;
-		vsi->rx_buf_len = I40E_RXBUFFER_2048;
+	vsi->max_frame = I40E_MAX_RXBUFFER;
+	vsi->rx_buf_len = i40e_calculate_vsi_rx_buf_len(vsi);
+
  #if (PAGE_SIZE < 8192)
-	} else if (!I40E_2K_TOO_SMALL_WITH_PADDING &&
-		   (vsi->netdev->mtu <= ETH_DATA_LEN)) {
+	if (vsi->netdev && !I40E_2K_TOO_SMALL_WITH_PADDING &&
+	    vsi->netdev->mtu <= ETH_DATA_LEN)
  		vsi->max_frame = I40E_RXBUFFER_1536 - NET_IP_ALIGN;
-		vsi->rx_buf_len = I40E_RXBUFFER_1536 - NET_IP_ALIGN;
  #endif
-	} else {
-		vsi->max_frame = I40E_MAX_RXBUFFER;
-		vsi->rx_buf_len = (PAGE_SIZE < 8192) ? I40E_RXBUFFER_3072 :
-						       I40E_RXBUFFER_2048;
-	}
 
  	/* set up individual rings */
  	for (i = 0; i < vsi->num_queue_pairs && !err; i++)

drivers/net/ethernet/intel/i40e/i40e_main.c static int i40e_xdp_setup(struct i40e_vsi *vsi, struct bpf_prog *prog, · line 13289

Quoted 8 lines
drivers/net/ethernet/intel/i40e/i40e_main.c +1 -1
@@ -13265,7 +13283,7 @@ static int i40e_xdp_setup(struct i40e_vsi *vsi, struct bpf_prog *prog,
  	int i;
 
  	/* Don't allow frames that span over multiple buffers */
-	if (frame_size > vsi->rx_buf_len) {
+	if (frame_size > i40e_calculate_vsi_rx_buf_len(vsi)) {
  		NL_SET_ERR_MSG_MOD(extack, "MTU too large to enable XDP");
  		return -EINVAL;
  	}

_______________________________________________
Intel-wired-lan mailing list
[email protected]
https://lists.osuosl.org/mailman/listinfo/intel-wired-lan

Reviewed-by on v2 1/1
#2 lore raw
In reply to Mateusz Palczewski
Quoted 3 lines · Mateusz Palczewski

From: "Staszewski, BartoszX" <[email protected]>

Whenever interface was down, function

Can you say explicitly that you were trying to load XDP prog on downed iface?

Quoted 7 lines

i40e_xdp was passing vsi->rx_buf_len field
to i40e_xdp_setup() which was equal 0.
i40e_open() calls i40e_vsi_configure_rx()
which configures that field, but that only
happens when interface is up. When it is
down, i40e_open() is not being called, thus
vsi->rx_buf_len is not set.

looks odd, can you set your editor to have 72 chars per line in the commit msg? Also, would be good to include the mention that you were getting "MTU too large to enable XDP" like you had in v1.

Quoted 6 lines

Solution for this is calculate buffer length
in newly created function - i40e_calculate_vsi_rx_buf_len() that return actual buffer length. Buffer length is being calculated based on the same rules applied previously in i40e_vsi_configure_rx() function.

Contents of the patch looks ok to me, but still I would improve commit msg. Would be good if you could take a look at ixgbe and ice if they have similar issue - eazy commitz.

Quoted 89 lines
Fixes 613142b0bb88 ("i40e: Log error for oversized MTU on device")
Fixes 0c8493d90b6b ("i40e: add XDP support for pass and drop actions")

---
v2: Fixed commit title and message, this patch is only for a case of
fresh start so I believe there is no need for rx_buf_len clearance

---
 drivers/net/ethernet/intel/i40e/i40e_main.c | 42 +++++++++++++++------
 1 file changed, 30 insertions(+), 12 deletions(-)
drivers/net/ethernet/intel/i40e/i40e_main.c +30 -12
index b5dcd15ced36..2fec0cff282c 100644
@@ -3693,6 +3693,30 @@ static int i40e_vsi_configure_tx(struct i40e_vsi *vsi)
 	return err;
 }
 
+/**
+ * i40e_calculate_vsi_rx_buf_len - Calculates buffer length
+ *
+ * @vsi: VSI to calculate rx_buf_len from
+ */
+static u16 i40e_calculate_vsi_rx_buf_len(struct i40e_vsi *vsi)
+{
+	u16 ret;
+
+	if (!vsi->netdev || (vsi->back->flags & I40E_FLAG_LEGACY_RX)) {
+		ret = I40E_RXBUFFER_2048;
+#if (PAGE_SIZE < 8192)
+	} else if (!I40E_2K_TOO_SMALL_WITH_PADDING &&
+		   (vsi->netdev->mtu <= ETH_DATA_LEN)) {
+		ret = I40E_RXBUFFER_1536 - NET_IP_ALIGN;
+#endif
+	} else {
+		ret = (PAGE_SIZE < 8192) ? I40E_RXBUFFER_3072 :
+					   I40E_RXBUFFER_2048;
+	}
+
+	return ret;
+}
+
 /**
  * i40e_vsi_configure_rx - Configure the VSI for Rx
  * @vsi: the VSI being configured
@@ -3704,20 +3728,14 @@ static int i40e_vsi_configure_rx(struct i40e_vsi *vsi)
 	int err = 0;
 	u16 i;
 
-	if (!vsi->netdev || (vsi->back->flags & I40E_FLAG_LEGACY_RX)) {
-		vsi->max_frame = I40E_MAX_RXBUFFER;
-		vsi->rx_buf_len = I40E_RXBUFFER_2048;
+	vsi->max_frame = I40E_MAX_RXBUFFER;
+	vsi->rx_buf_len = i40e_calculate_vsi_rx_buf_len(vsi);
+
 #if (PAGE_SIZE < 8192)
-	} else if (!I40E_2K_TOO_SMALL_WITH_PADDING &&
-		   (vsi->netdev->mtu <= ETH_DATA_LEN)) {
+	if (vsi->netdev && !I40E_2K_TOO_SMALL_WITH_PADDING &&
+	    vsi->netdev->mtu <= ETH_DATA_LEN)
 		vsi->max_frame = I40E_RXBUFFER_1536 - NET_IP_ALIGN;
-		vsi->rx_buf_len = I40E_RXBUFFER_1536 - NET_IP_ALIGN;
 #endif
-	} else {
-		vsi->max_frame = I40E_MAX_RXBUFFER;
-		vsi->rx_buf_len = (PAGE_SIZE < 8192) ? I40E_RXBUFFER_3072 :
-						       I40E_RXBUFFER_2048;
-	}
 
 	/* set up individual rings */
 	for (i = 0; i < vsi->num_queue_pairs && !err; i++)
@@ -13265,7 +13283,7 @@ static int i40e_xdp_setup(struct i40e_vsi *vsi, struct bpf_prog *prog,
 	int i;
 
 	/* Don't allow frames that span over multiple buffers */
-	if (frame_size > vsi->rx_buf_len) {
+	if (frame_size > i40e_calculate_vsi_rx_buf_len(vsi)) {
 		NL_SET_ERR_MSG_MOD(extack, "MTU too large to enable XDP");
 		return -EINVAL;
 	}
Signature · 6 lines
2.31.1

_______________________________________________
Intel-wired-lan mailing list
[email protected]
https://lists.osuosl.org/mailman/listinfo/intel-wired-lan

_______________________________________________
Intel-wired-lan mailing list
[email protected]
https://lists.osuosl.org/mailman/listinfo/intel-wired-lan

#3 lore raw

Dear Mateusz,

Thank you for the patch.

Am 06.12.22 um 12:55 schrieb Maciej Fijalkowski:

Quoted 1 line · Mateusz Palczewski

[…]

Quoted 5 lines
Quoted 5 lines

Solution for this is calculate buffer length
in newly created function - i40e_calculate_vsi_rx_buf_len() that return actual buffer length. Buffer length is being calculated based on the same rules applied previously in i40e_vsi_configure_rx() function.

s/that return/that returns/

Quoted 2 lines

Contents of the patch looks ok to me, but still I would improve commit msg.

I agree with Maciej. Please also improve the git commit message summary/title:

Quoted 1 line

Allow to attach XDP programs on downed interface

[…]

Kind regards,

Paul
_______________________________________________
Intel-wired-lan mailing list
[email protected]
https://lists.osuosl.org/mailman/listinfo/intel-wired-lan