Extra IPsec cleanup

PATCH In review Leon Romanovsky 2 participants 24 messages Apr 10, 2022 – Apr 11, 2022

Subsystems (7)

Networking Network drivers InfiniBand Mellanox mlx5 core +3 more subsystems

lore

Share & tools

lore

Leon Romanovsky

author

2022-04-10 08:28 UTC #0 lore raw

cover Extra IPsec cleanup

After FPGA IPsec removal, we can go further and make sure that flow steering logic is aligned to mlx5_core standard together with deep cleaning of whole IPsec path.

Thanks

Leon Romanovsky (17):
net/mlx5: Simplify IPsec flow steering init/cleanup functions net/mlx5: Check IPsec TX flow steering namespace in advance net/mlx5: Don't hide fallback to software IPsec in FS code net/mlx5: Reduce useless indirection in IPsec FS add/delete flows net/mlx5: Store IPsec ESN update work in XFRM state net/mlx5: Remove useless validity check
net/mlx5: Merge various control path IPsec headers into one file net/mlx5: Remove accel notations and indirections from esp functions net/mlx5: Simplify HW context interfaces by using SA entry net/mlx5: Clean IPsec FS add/delete rules
net/mlx5: Make sure that no dangling IPsec FS pointers exist net/mlx5: Don't advertise IPsec netdev support for non-IPsec device net/mlx5: Simplify IPsec capabilities logic
net/mlx5: Remove not-supported ICV length
net/mlx5: Cleanup XFRM attributes struct
net/mlx5: Allow future addition of IPsec object modifiers net/mlx5: Don't perform lookup after already known sec_path

 .../net/ethernet/mellanox/mlx5/core/en/fs.h   |   1 -
 .../ethernet/mellanox/mlx5/core/en/params.c   |   2 +-
 .../mellanox/mlx5/core/en_accel/ipsec.c       | 172 +++------
 .../mellanox/mlx5/core/en_accel/ipsec.h       |  85 +++-
 .../mellanox/mlx5/core/en_accel/ipsec_fs.c    | 362 ++++++------------
 .../mellanox/mlx5/core/en_accel/ipsec_fs.h    |  21 -
 .../mlx5/core/en_accel/ipsec_offload.c        | 332 +++-------------
 .../mlx5/core/en_accel/ipsec_offload.h        |  14 -
 .../mellanox/mlx5/core/en_accel/ipsec_rxtx.c  |   6 +-
 .../mellanox/mlx5/core/en_accel/ipsec_stats.c |   4 +-
 .../net/ethernet/mellanox/mlx5/core/en_main.c |   1 -
 .../net/ethernet/mellanox/mlx5/core/en_rx.c   |   2 +-
 .../net/ethernet/mellanox/mlx5/core/main.c    |   2 +-
 include/linux/mlx5/accel.h                    | 153 --------
 include/linux/mlx5/mlx5_ifc.h                 |   2 -
 15 files changed, 320 insertions(+), 839 deletions(-)
 delete mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.h
 delete mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_offload.h
 delete mode 100644 include/linux/mlx5/accel.h

Signature · 1 line

2.35.1

Leon Romanovsky

author

2022-04-10 08:28 UTC #1 lore raw

PATCH 1/17 net/mlx5: Simplify IPsec flow steering init/cleanup functions

From: Leon Romanovsky <[email protected]>

Cleanup IPsec FS initialization and cleanup functions.

Reviewed-by Raed Salem <[email protected]>

Signed-off-by Leon Romanovsky <[email protected]>

---
 .../mellanox/mlx5/core/en_accel/ipsec.c       |  4 +-
 .../mellanox/mlx5/core/en_accel/ipsec_fs.c    | 73 ++++++-------------
 .../mellanox/mlx5/core/en_accel/ipsec_fs.h    |  4 +-
 3 files changed, 27 insertions(+), 54 deletions(-)

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c +2 -2

index c280a18ff002..5a10755dd4f1 100644
@@ -424,7 +424,7 @@ int mlx5e_ipsec_init(struct mlx5e_priv *priv)
 	}
 
 	priv->ipsec = ipsec;
-	mlx5e_accel_ipsec_fs_init(priv);
+	mlx5e_ipsec_fs_init(ipsec);
 	netdev_dbg(priv->netdev, "IPSec attached to netdevice\n");
 	return 0;
 }

@@ -436,7 +436,7 @@ void mlx5e_ipsec_cleanup(struct mlx5e_priv *priv)
 	if (!ipsec)
 		return;
 
-	mlx5e_accel_ipsec_fs_cleanup(priv);
+	mlx5e_ipsec_fs_cleanup(ipsec);
 	destroy_workqueue(ipsec->wq);
 
 	kfree(ipsec);

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c +23 ~~-50~~

104 lines not on this page · raw message

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.h +2 -2

index b70953979709..8e0e829ab58f 100644
@@ -9,8 +9,8 @@
 #include "ipsec_offload.h"
 #include "en/fs.h"
 
-void mlx5e_accel_ipsec_fs_cleanup(struct mlx5e_priv *priv);
-int mlx5e_accel_ipsec_fs_init(struct mlx5e_priv *priv);
+void mlx5e_ipsec_fs_cleanup(struct mlx5e_ipsec *ipsec);
+int mlx5e_ipsec_fs_init(struct mlx5e_ipsec *ipsec);
 int mlx5e_accel_ipsec_fs_add_rule(struct mlx5e_priv *priv,
 				  struct mlx5_accel_esp_xfrm_attrs *attrs,
 				  u32 ipsec_obj_id,

Signature · 1 line

2.35.1

Saeed Mahameed

maintainer

2022-04-10 16:46 UTC #2 lore raw

Re: [PATCH mlx5-next 01/17] net/mlx5: Simplify IPsec flow steering init/cleanup functions

Quoted 3 lines · Leon Romanovsky

From: Leon Romanovsky <[email protected]>
Cleanup IPsec FS initialization and cleanup functions.

Can you be more clear about what are you cleaning up ?

unfolding/joining static functions shouldn't be considered as cleanup.

Also i don't agree these patches should go to mlx5-next, we need to avoid bloating mlx5-next.
Many of these patches are purely ipsec, yes i understand you are heavily modifying include/linux/mlx5/accel.h but from what I can tell, it's not affecting rdma.

Please give me a chance to review the whole series until next week as i am out of office this week.

Quoted 158 lines

Reviewed-by Raed Salem <[email protected]>

Signed-off-by Leon Romanovsky <[email protected]>

---

.../mellanox/mlx5/core/en_accel/ipsec.c       |  4 +-
.../mellanox/mlx5/core/en_accel/ipsec_fs.c    | 73 ++++++-------------
.../mellanox/mlx5/core/en_accel/ipsec_fs.h    |  4 +-
3 files changed, 27 insertions(+), 54 deletions(-)

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c +2 -2

index c280a18ff002..5a10755dd4f1 100644
@@ -424,7 +424,7 @@ int mlx5e_ipsec_init(struct mlx5e_priv *priv)
 	}
 
 	priv->ipsec = ipsec;
-	mlx5e_accel_ipsec_fs_init(priv);
+	mlx5e_ipsec_fs_init(ipsec);
 	netdev_dbg(priv->netdev, "IPSec attached to netdevice\n");
 	return 0;
 }

@@ -436,7 +436,7 @@ void mlx5e_ipsec_cleanup(struct mlx5e_priv *priv)
 	if (!ipsec)
 		return;
 
-	mlx5e_accel_ipsec_fs_cleanup(priv);
+	mlx5e_ipsec_fs_cleanup(ipsec);
 	destroy_workqueue(ipsec->wq);
 
 	kfree(ipsec);

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c +23 ~~-50~~

index 66b529e36ea1..869b5692e9b9 100644
@@ -632,81 +632,54 @@ void mlx5e_accel_ipsec_fs_del_rule(struct mlx5e_priv *priv,
 		tx_del_rule(priv, ipsec_rule);
 }
 
-static void fs_cleanup_tx(struct mlx5e_priv *priv)
-{
-	mutex_destroy(&priv->ipsec->tx_fs->mutex);
-	WARN_ON(priv->ipsec->tx_fs->refcnt);
-	kfree(priv->ipsec->tx_fs);
-	priv->ipsec->tx_fs = NULL;
-}
-
-static void fs_cleanup_rx(struct mlx5e_priv *priv)
+void mlx5e_ipsec_fs_cleanup(struct mlx5e_ipsec *ipsec)
 {
 	struct mlx5e_accel_fs_esp_prot *fs_prot;
 	struct mlx5e_accel_fs_esp *accel_esp;
 	enum accel_fs_esp_type i;
 
-	accel_esp = priv->ipsec->rx_fs;
+	if (!ipsec->rx_fs)
+		return;
+
+	mutex_destroy(&ipsec->tx_fs->mutex);
+	WARN_ON(ipsec->tx_fs->refcnt);
+	kfree(ipsec->tx_fs);
+
+	accel_esp = ipsec->rx_fs;
 	for (i = 0; i < ACCEL_FS_ESP_NUM_TYPES; i++) {
 		fs_prot = &accel_esp->fs_prot[i];
 		mutex_destroy(&fs_prot->prot_mutex);
 		WARN_ON(fs_prot->refcnt);
 	}
-	kfree(priv->ipsec->rx_fs);
-	priv->ipsec->rx_fs = NULL;
-}
-
-static int fs_init_tx(struct mlx5e_priv *priv)
-{
-	priv->ipsec->tx_fs =
-		kzalloc(sizeof(struct mlx5e_ipsec_tx), GFP_KERNEL);
-	if (!priv->ipsec->tx_fs)
-		return -ENOMEM;
-
-	mutex_init(&priv->ipsec->tx_fs->mutex);
-	return 0;
+	kfree(ipsec->rx_fs);
 }
 
-static int fs_init_rx(struct mlx5e_priv *priv)
+int mlx5e_ipsec_fs_init(struct mlx5e_ipsec *ipsec)
 {
 	struct mlx5e_accel_fs_esp_prot *fs_prot;
 	struct mlx5e_accel_fs_esp *accel_esp;
 	enum accel_fs_esp_type i;
+	int err = -ENOMEM;
 
-	priv->ipsec->rx_fs =
-		kzalloc(sizeof(struct mlx5e_accel_fs_esp), GFP_KERNEL);
-	if (!priv->ipsec->rx_fs)
+	ipsec->tx_fs = kzalloc(sizeof(*ipsec->tx_fs), GFP_KERNEL);
+	if (!ipsec->tx_fs)
 		return -ENOMEM;
 
-	accel_esp = priv->ipsec->rx_fs;
+	ipsec->rx_fs = kzalloc(sizeof(*ipsec->rx_fs), GFP_KERNEL);
+	if (!ipsec->rx_fs)
+		goto err_rx;
+
+	mutex_init(&ipsec->tx_fs->mutex);
+
+	accel_esp = ipsec->rx_fs;
 	for (i = 0; i < ACCEL_FS_ESP_NUM_TYPES; i++) {
 		fs_prot = &accel_esp->fs_prot[i];
 		mutex_init(&fs_prot->prot_mutex);
 	}
 
 	return 0;
-}
-
-void mlx5e_accel_ipsec_fs_cleanup(struct mlx5e_priv *priv)
-{
-	if (!priv->ipsec->rx_fs)
-		return;
-
-	fs_cleanup_tx(priv);
-	fs_cleanup_rx(priv);
-}
-
-int mlx5e_accel_ipsec_fs_init(struct mlx5e_priv *priv)
-{
-	int err;
-
-	err = fs_init_tx(priv);
-	if (err)
-		return err;
-
-	err = fs_init_rx(priv);
-	if (err)
-		fs_cleanup_tx(priv);
 
+err_rx:
+	kfree(ipsec->tx_fs);
 	return err;
 }

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.h +2 -2

index b70953979709..8e0e829ab58f 100644
@@ -9,8 +9,8 @@
 #include "ipsec_offload.h"
 #include "en/fs.h"
 
-void mlx5e_accel_ipsec_fs_cleanup(struct mlx5e_priv *priv);
-int mlx5e_accel_ipsec_fs_init(struct mlx5e_priv *priv);
+void mlx5e_ipsec_fs_cleanup(struct mlx5e_ipsec *ipsec);
+int mlx5e_ipsec_fs_init(struct mlx5e_ipsec *ipsec);
 int mlx5e_accel_ipsec_fs_add_rule(struct mlx5e_priv *priv,
 				  struct mlx5_accel_esp_xfrm_attrs *attrs,
 				  u32 ipsec_obj_id,

Signature · 1 line

2.35.1

Leon Romanovsky

author

2022-04-10 17:21 UTC #3 lore raw

Re: [PATCH mlx5-next 01/17] net/mlx5: Simplify IPsec flow steering init/cleanup functions

Quoted 8 lines · Saeed Mahameed

Quoted 3 lines · Leon Romanovsky
From: Leon Romanovsky <[email protected]>
Cleanup IPsec FS initialization and cleanup functions.
Can you be more clear about what are you cleaning up ?
unfolding/joining static functions shouldn't be considered as cleanup.

And how would you describe extensive usage of one time called functions that have no use as standalone ones?

This patch makes sure that all flow steering initialized and cleaned at one place and allows me to present coherent picture of what is needed for IPsec FS.

You should focus on the end result of this series rather on single patch.

 15 files changed, 320 insertions(+), 839 deletions(-)

Quoted 6 lines

Also i don't agree these patches should go to mlx5-next, we need to avoid bloating mlx5-next.
Many of these patches are purely ipsec, yes i understand you are heavily modifying include/linux/mlx5/accel.h but from what I can tell, it's not affecting rdma.

I'm deleting include/linux/mlx5/accel.h, it is not needed. But I don't care about the target, it can be net-next and not mlx5-next.

Quoted 3 lines

Please give me a chance to review the whole series until next week as i am out of office this week.

I see this problematic as next week will be combination of my Passover vacation and paternity leave at the same time.

Thanks

Quoted 159 lines

Reviewed-by Raed Salem <[email protected]>

Signed-off-by Leon Romanovsky <[email protected]>

---

.../mellanox/mlx5/core/en_accel/ipsec.c       |  4 +-
.../mellanox/mlx5/core/en_accel/ipsec_fs.c    | 73 ++++++-------------
.../mellanox/mlx5/core/en_accel/ipsec_fs.h    |  4 +-
3 files changed, 27 insertions(+), 54 deletions(-)

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c +2 -2

index c280a18ff002..5a10755dd4f1 100644
@@ -424,7 +424,7 @@ int mlx5e_ipsec_init(struct mlx5e_priv *priv)
 	}
 
 	priv->ipsec = ipsec;
-	mlx5e_accel_ipsec_fs_init(priv);
+	mlx5e_ipsec_fs_init(ipsec);
 	netdev_dbg(priv->netdev, "IPSec attached to netdevice\n");
 	return 0;
 }

@@ -436,7 +436,7 @@ void mlx5e_ipsec_cleanup(struct mlx5e_priv *priv)
 	if (!ipsec)
 		return;
 
-	mlx5e_accel_ipsec_fs_cleanup(priv);
+	mlx5e_ipsec_fs_cleanup(ipsec);
 	destroy_workqueue(ipsec->wq);
 
 	kfree(ipsec);

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c +23 ~~-50~~

index 66b529e36ea1..869b5692e9b9 100644
@@ -632,81 +632,54 @@ void mlx5e_accel_ipsec_fs_del_rule(struct mlx5e_priv *priv,
 		tx_del_rule(priv, ipsec_rule);
 }
 
-static void fs_cleanup_tx(struct mlx5e_priv *priv)
-{
-	mutex_destroy(&priv->ipsec->tx_fs->mutex);
-	WARN_ON(priv->ipsec->tx_fs->refcnt);
-	kfree(priv->ipsec->tx_fs);
-	priv->ipsec->tx_fs = NULL;
-}
-
-static void fs_cleanup_rx(struct mlx5e_priv *priv)
+void mlx5e_ipsec_fs_cleanup(struct mlx5e_ipsec *ipsec)
 {
 	struct mlx5e_accel_fs_esp_prot *fs_prot;
 	struct mlx5e_accel_fs_esp *accel_esp;
 	enum accel_fs_esp_type i;
 
-	accel_esp = priv->ipsec->rx_fs;
+	if (!ipsec->rx_fs)
+		return;
+
+	mutex_destroy(&ipsec->tx_fs->mutex);
+	WARN_ON(ipsec->tx_fs->refcnt);
+	kfree(ipsec->tx_fs);
+
+	accel_esp = ipsec->rx_fs;
 	for (i = 0; i < ACCEL_FS_ESP_NUM_TYPES; i++) {
 		fs_prot = &accel_esp->fs_prot[i];
 		mutex_destroy(&fs_prot->prot_mutex);
 		WARN_ON(fs_prot->refcnt);
 	}
-	kfree(priv->ipsec->rx_fs);
-	priv->ipsec->rx_fs = NULL;
-}
-
-static int fs_init_tx(struct mlx5e_priv *priv)
-{
-	priv->ipsec->tx_fs =
-		kzalloc(sizeof(struct mlx5e_ipsec_tx), GFP_KERNEL);
-	if (!priv->ipsec->tx_fs)
-		return -ENOMEM;
-
-	mutex_init(&priv->ipsec->tx_fs->mutex);
-	return 0;
+	kfree(ipsec->rx_fs);
 }
 
-static int fs_init_rx(struct mlx5e_priv *priv)
+int mlx5e_ipsec_fs_init(struct mlx5e_ipsec *ipsec)
 {
 	struct mlx5e_accel_fs_esp_prot *fs_prot;
 	struct mlx5e_accel_fs_esp *accel_esp;
 	enum accel_fs_esp_type i;
+	int err = -ENOMEM;
 
-	priv->ipsec->rx_fs =
-		kzalloc(sizeof(struct mlx5e_accel_fs_esp), GFP_KERNEL);
-	if (!priv->ipsec->rx_fs)
+	ipsec->tx_fs = kzalloc(sizeof(*ipsec->tx_fs), GFP_KERNEL);
+	if (!ipsec->tx_fs)
 		return -ENOMEM;
 
-	accel_esp = priv->ipsec->rx_fs;
+	ipsec->rx_fs = kzalloc(sizeof(*ipsec->rx_fs), GFP_KERNEL);
+	if (!ipsec->rx_fs)
+		goto err_rx;
+
+	mutex_init(&ipsec->tx_fs->mutex);
+
+	accel_esp = ipsec->rx_fs;
 	for (i = 0; i < ACCEL_FS_ESP_NUM_TYPES; i++) {
 		fs_prot = &accel_esp->fs_prot[i];
 		mutex_init(&fs_prot->prot_mutex);
 	}
 
 	return 0;
-}
-
-void mlx5e_accel_ipsec_fs_cleanup(struct mlx5e_priv *priv)
-{
-	if (!priv->ipsec->rx_fs)
-		return;
-
-	fs_cleanup_tx(priv);
-	fs_cleanup_rx(priv);
-}
-
-int mlx5e_accel_ipsec_fs_init(struct mlx5e_priv *priv)
-{
-	int err;
-
-	err = fs_init_tx(priv);
-	if (err)
-		return err;
-
-	err = fs_init_rx(priv);
-	if (err)
-		fs_cleanup_tx(priv);
 
+err_rx:
+	kfree(ipsec->tx_fs);
 	return err;
 }

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.h +2 -2

index b70953979709..8e0e829ab58f 100644
@@ -9,8 +9,8 @@
 #include "ipsec_offload.h"
 #include "en/fs.h"
 
-void mlx5e_accel_ipsec_fs_cleanup(struct mlx5e_priv *priv);
-int mlx5e_accel_ipsec_fs_init(struct mlx5e_priv *priv);
+void mlx5e_ipsec_fs_cleanup(struct mlx5e_ipsec *ipsec);
+int mlx5e_ipsec_fs_init(struct mlx5e_ipsec *ipsec);
 int mlx5e_accel_ipsec_fs_add_rule(struct mlx5e_priv *priv,
 				  struct mlx5_accel_esp_xfrm_attrs *attrs,
 				  u32 ipsec_obj_id,

Signature · 1 line

2.35.1

Saeed Mahameed

maintainer

2022-04-10 21:58 UTC #4 lore raw

Re: [PATCH mlx5-next 01/17] net/mlx5: Simplify IPsec flow steering init/cleanup functions

Quoted 4 lines · Leon Romanovsky

Quoted 3 lines · Saeed Mahameed
Quoted 2 lines · Leon Romanovsky
From: Leon Romanovsky <[email protected]>

Quoted 9 lines

Quoted 5 lines
Quoted 1 line
Cleanup IPsec FS initialization and cleanup functions.
Can you be more clear about what are you cleaning up ?
unfolding/joining static functions shouldn't be considered as cleanup.
And how would you describe extensive usage of one time called functions that have no use as standalone ones?

Functional programming.

Quoted 4 lines

This patch makes sure that all flow steering initialized and cleaned at one place and allows me to present coherent picture of what is needed for IPsec FS.

This is already the case before this patch.

Quoted 2 lines

You should focus on the end result of this series rather on single patch.
15 files changed, 320 insertions(+), 839 deletions(-)

Overall the series is fine, this patch in particular is unnecessary cancelation of others previous decisions, which i personally like and might as well have suggested myself, so let's avoid such clutter.

Leon Romanovsky

author

2022-04-11 06:38 UTC #5 lore raw

Re: [PATCH mlx5-next 01/17] net/mlx5: Simplify IPsec flow steering init/cleanup functions

Quoted 9 lines · Saeed Mahameed

Quoted 8 lines · Leon Romanovsky
Quoted 7 lines · Saeed Mahameed
Quoted 3 lines · Leon Romanovsky
From: Leon Romanovsky <[email protected]>
Cleanup IPsec FS initialization and cleanup functions.
Can you be more clear about what are you cleaning up ?

Quoted 7 lines

Quoted 5 lines
Quoted 1 line
unfolding/joining static functions shouldn't be considered as cleanup.
And how would you describe extensive usage of one time called functions that have no use as standalone ones?
Functional programming.

The separation between various modules and their functions is function programming, but wrapper in .c file over basic kernel primitive (kzalloc) is obfuscation.

Quoted 7 lines

Quoted 5 lines
This patch makes sure that all flow steering initialized and cleaned at one place and allows me to present coherent picture of what is needed for IPsec FS.
This is already the case before this patch.

With two main differences:
First, it is is less code to achieve the same and second, it is easy to read (I read this code a lot lately).

 3 files changed, 27 insertions(+), 54 deletions(-)

Quoted 7 lines

Quoted 3 lines
You should focus on the end result of this series rather on single patch.
15 files changed, 320 insertions(+), 839 deletions(-)
Overall the series is fine, this patch in particular is unnecessary cancelation of others previous decisions, which i personally like and might as well have suggested myself, so let's avoid such clutter.

Sorry, but I disagree that removal of useless indirection that hurts readability is clutter. It is refactoring.

Please focus on end goal of this series.

Thanks

Quoted 2 lines

Leon Romanovsky

author

2022-04-10 08:29 UTC #6 lore raw

PATCH 2/17 net/mlx5: Check IPsec TX flow steering namespace in advance

In reply to Leon Romanovsky

From: Leon Romanovsky <[email protected]>

Ensure that flow steering is usable as early as possible, to understand if crypto IPsec is supported or not.

Reviewed-by Raed Salem <[email protected]>

Signed-off-by Leon Romanovsky <[email protected]>

---
 drivers/net/ethernet/mellanox/mlx5/core/en/fs.h  |  1 -
 .../ethernet/mellanox/mlx5/core/en_accel/ipsec.c |  1 +
 .../ethernet/mellanox/mlx5/core/en_accel/ipsec.h |  1 +
 .../mellanox/mlx5/core/en_accel/ipsec_fs.c       | 16 +++++++++-------
 4 files changed, 11 insertions(+), 8 deletions(-)

drivers/net/ethernet/mellanox/mlx5/core/en/fs.h +0 -1

index 678ffbb48a25..4130a871de61 100644
@@ -164,7 +164,6 @@ struct mlx5e_ptp_fs;
 
 struct mlx5e_flow_steering {
 	struct mlx5_flow_namespace      *ns;
-	struct mlx5_flow_namespace      *egress_ns;
 #ifdef CONFIG_MLX5_EN_RXNFC
 	struct mlx5e_ethtool_steering   ethtool;
 #endif

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c +1 -0

index 5a10755dd4f1..285ccb773de6 100644
@@ -415,6 +415,7 @@ int mlx5e_ipsec_init(struct mlx5e_priv *priv)
 
 	hash_init(ipsec->sadb_rx);
 	spin_lock_init(&ipsec->sadb_rx_lock);
+	ipsec->mdev = priv->mdev;
 	ipsec->en_priv = priv;
 	ipsec->wq = alloc_ordered_workqueue("mlx5e_ipsec: %s", 0,
 					    priv->netdev->name);

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h +1 -0

index a0e9dade09e9..bbf48d4616f9 100644
@@ -61,6 +61,7 @@ struct mlx5e_accel_fs_esp;
 struct mlx5e_ipsec_tx;
 
 struct mlx5e_ipsec {
+	struct mlx5_core_dev *mdev;
 	struct mlx5e_priv *en_priv;
 	DECLARE_HASHTABLE(sadb_rx, MLX5E_IPSEC_SADB_RX_BITS);
 	spinlock_t sadb_rx_lock; /* Protects sadb_rx */

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c +9 -7

index 869b5692e9b9..66b8ead8b579 100644
@@ -35,6 +35,7 @@ struct mlx5e_accel_fs_esp {
 };
 
 struct mlx5e_ipsec_tx {
+	struct mlx5_flow_namespace *ns;
 	struct mlx5_flow_table *ft;
 	struct mutex mutex; /* Protect IPsec TX steering */
 	u32 refcnt;

@@ -338,15 +339,9 @@ static int tx_create(struct mlx5e_priv *priv)
 	struct mlx5_flow_table *ft;
 	int err;
 
-	priv->fs.egress_ns =
-		mlx5_get_flow_namespace(priv->mdev,
-					MLX5_FLOW_NAMESPACE_EGRESS_KERNEL);
-	if (!priv->fs.egress_ns)
-		return -EOPNOTSUPP;
-
 	ft_attr.max_fte = NUM_IPSEC_FTE;
 	ft_attr.autogroup.max_num_groups = 1;
-	ft = mlx5_create_auto_grouped_flow_table(priv->fs.egress_ns, &ft_attr);
+	ft = mlx5_create_auto_grouped_flow_table(ipsec->tx_fs->ns, &ft_attr);
 	if (IS_ERR(ft)) {
 		err = PTR_ERR(ft);
 		netdev_err(priv->netdev, "fail to create ipsec tx ft err=%d\n", err);

@@ -658,9 +653,15 @@ int mlx5e_ipsec_fs_init(struct mlx5e_ipsec *ipsec)
 {
 	struct mlx5e_accel_fs_esp_prot *fs_prot;
 	struct mlx5e_accel_fs_esp *accel_esp;
+	struct mlx5_flow_namespace *ns;
 	enum accel_fs_esp_type i;
 	int err = -ENOMEM;
 
+	ns = mlx5_get_flow_namespace(ipsec->mdev,
+				     MLX5_FLOW_NAMESPACE_EGRESS_KERNEL);
+	if (!ns)
+		return -EOPNOTSUPP;
+
 	ipsec->tx_fs = kzalloc(sizeof(*ipsec->tx_fs), GFP_KERNEL);
 	if (!ipsec->tx_fs)
 		return -ENOMEM;

@@ -670,6 +671,7 @@ int mlx5e_ipsec_fs_init(struct mlx5e_ipsec *ipsec)
 		goto err_rx;
 
 	mutex_init(&ipsec->tx_fs->mutex);
+	ipsec->tx_fs->ns = ns;
 
 	accel_esp = ipsec->rx_fs;
 	for (i = 0; i < ACCEL_FS_ESP_NUM_TYPES; i++) {

Signature · 1 line

2.35.1

Saeed Mahameed

maintainer Reviewed-by on v1 2/17

2022-04-10 23:46 UTC #7 lore raw

Re: [PATCH mlx5-next 02/17] net/mlx5: Check IPsec TX flow steering namespace in advance

Quoted 43 lines · Leon Romanovsky

From: Leon Romanovsky <[email protected]>

Ensure that flow steering is usable as early as possible, to understand if crypto IPsec is supported or not.

Reviewed-by Raed Salem <[email protected]>

Signed-off-by Leon Romanovsky <[email protected]>

---

drivers/net/ethernet/mellanox/mlx5/core/en/fs.h  |  1 -
.../ethernet/mellanox/mlx5/core/en_accel/ipsec.c |  1 +
.../ethernet/mellanox/mlx5/core/en_accel/ipsec.h |  1 +
.../mellanox/mlx5/core/en_accel/ipsec_fs.c       | 16 +++++++++-------
4 files changed, 11 insertions(+), 8 deletions(-)

drivers/net/ethernet/mellanox/mlx5/core/en/fs.h +0 -1

index 678ffbb48a25..4130a871de61 100644
@@ -164,7 +164,6 @@ struct mlx5e_ptp_fs;
 
 struct mlx5e_flow_steering {
 	struct mlx5_flow_namespace      *ns;
-	struct mlx5_flow_namespace      *egress_ns;
 #ifdef CONFIG_MLX5_EN_RXNFC
 	struct mlx5e_ethtool_steering   ethtool;
 #endif

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c +1 -0

index 5a10755dd4f1..285ccb773de6 100644
@@ -415,6 +415,7 @@ int mlx5e_ipsec_init(struct mlx5e_priv *priv)
 
 	hash_init(ipsec->sadb_rx);
 	spin_lock_init(&ipsec->sadb_rx_lock);
+	ipsec->mdev = priv->mdev;
 	ipsec->en_priv = priv;
 	ipsec->wq = alloc_ordered_workqueue("mlx5e_ipsec: %s", 0,
 					    priv->netdev->name);

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h +0 -0

index a0e9dade09e9..bbf48d4616f9 100644

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h struct mlx5e_accel_fs_esp; · line 65

Quoted 5 lines

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h +1 -0

@@ -61,6 +61,7 @@ struct mlx5e_accel_fs_esp;
 struct mlx5e_ipsec_tx;
 
 struct mlx5e_ipsec {
+	struct mlx5_core_dev *mdev;
 	struct mlx5e_priv *en_priv;

Reviewed-by Saeed Mahameed <[email protected]>

we could probably remove en_priv, I already sent you a patch, please try to include it in the next version.

Leon Romanovsky

author

2022-04-11 06:21 UTC #8 lore raw

Re: [PATCH mlx5-next 02/17] net/mlx5: Check IPsec TX flow steering namespace in advance

Quoted 46 lines · Saeed Mahameed

Quoted 45 lines · Leon Romanovsky

From: Leon Romanovsky <[email protected]>

Ensure that flow steering is usable as early as possible, to understand if crypto IPsec is supported or not.

Reviewed-by Raed Salem <[email protected]>

Signed-off-by Leon Romanovsky <[email protected]>

---

drivers/net/ethernet/mellanox/mlx5/core/en/fs.h  |  1 -
.../ethernet/mellanox/mlx5/core/en_accel/ipsec.c |  1 +
.../ethernet/mellanox/mlx5/core/en_accel/ipsec.h |  1 +
.../mellanox/mlx5/core/en_accel/ipsec_fs.c       | 16 +++++++++-------
4 files changed, 11 insertions(+), 8 deletions(-)

drivers/net/ethernet/mellanox/mlx5/core/en/fs.h +0 -1

index 678ffbb48a25..4130a871de61 100644
@@ -164,7 +164,6 @@ struct mlx5e_ptp_fs;
 
 struct mlx5e_flow_steering {
 	struct mlx5_flow_namespace      *ns;
-	struct mlx5_flow_namespace      *egress_ns;
 #ifdef CONFIG_MLX5_EN_RXNFC
 	struct mlx5e_ethtool_steering   ethtool;
 #endif

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c +1 -0

index 5a10755dd4f1..285ccb773de6 100644
@@ -415,6 +415,7 @@ int mlx5e_ipsec_init(struct mlx5e_priv *priv)
 
 	hash_init(ipsec->sadb_rx);
 	spin_lock_init(&ipsec->sadb_rx_lock);
+	ipsec->mdev = priv->mdev;
 	ipsec->en_priv = priv;
 	ipsec->wq = alloc_ordered_workqueue("mlx5e_ipsec: %s", 0,
 					    priv->netdev->name);

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h +0 -0

index a0e9dade09e9..bbf48d4616f9 100644
@@ -61,6 +61,7 @@ struct mlx5e_accel_fs_esp;
 struct mlx5e_ipsec_tx;

Quoted 8 lines

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h struct mlx5e_accel_fs_esp; · line 65
Quoted 3 lines
drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h +1 -0
 struct mlx5e_ipsec {
+	struct mlx5_core_dev *mdev;
 	struct mlx5e_priv *en_priv;
Reviewed-by Saeed Mahameed <[email protected]>
we could probably remove en_priv, I already sent you a patch, please try to include it in the next version.

I removed this en_priv in patch #6 "net/mlx5: Store IPsec ESN update work in XFRM state"
https://lore.kernel.org/netdev/1d965d2697dda0bd2c34fc1ccbbb45efaf03b0de.1649578827.git.leonro@nvidia.com/T/#u

Thanks

Leon Romanovsky

author

2022-04-10 08:28 UTC #9 lore raw

PATCH 3/17 net/mlx5: Don't hide fallback to software IPsec in FS code

In reply to Leon Romanovsky

From: Leon Romanovsky <[email protected]>

The XFRM code performs fallback to software IPsec if .xdo_dev_state_add() returns -EOPNOTSUPP. This is what mlx5 did very deep in its stack trace, despite have all the knowledge that IPsec is not going to work in very early stage.

This is achieved by making sure that priv->ipsec pointer is valid for fully working and supported hardware crypto IPsec engine.

In case, the hardware IPsec is not supported, the XFRM code will set NULL to xso->dev and it will prevent from calls to various .xdo_dev_state_*() callbacks.

Reviewed-by Raed Salem <[email protected]>

Signed-off-by Leon Romanovsky <[email protected]>

---
 .../mellanox/mlx5/core/en_accel/ipsec.c       | 41 ++++++++-----------
 .../mellanox/mlx5/core/en_accel/ipsec_fs.c    |  6 ---
 2 files changed, 17 insertions(+), 30 deletions(-)

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c +17 ~~-24~~

index 285ccb773de6..be30b6e2a00f 100644
@@ -43,17 +43,7 @@
 
 static struct mlx5e_ipsec_sa_entry *to_ipsec_sa_entry(struct xfrm_state *x)
 {
-	struct mlx5e_ipsec_sa_entry *sa;
-
-	if (!x)
-		return NULL;
-
-	sa = (struct mlx5e_ipsec_sa_entry *)x->xso.offload_handle;
-	if (!sa)
-		return NULL;
-
-	WARN_ON(sa->x != x);
-	return sa;
+	return (struct mlx5e_ipsec_sa_entry *)x->xso.offload_handle;
 }
 
 struct xfrm_state *mlx5e_ipsec_sadb_rx_lookup(struct mlx5e_ipsec *ipsec,

@@ -306,6 +296,8 @@ static int mlx5e_xfrm_add_state(struct xfrm_state *x)
 	int err;
 
 	priv = netdev_priv(netdev);
+	if (!priv->ipsec)
+		return -EOPNOTSUPP;
 
 	err = mlx5e_xfrm_validate_state(x);
 	if (err)

@@ -375,9 +367,6 @@ static void mlx5e_xfrm_del_state(struct xfrm_state *x)
 {
 	struct mlx5e_ipsec_sa_entry *sa_entry = to_ipsec_sa_entry(x);
 
-	if (!sa_entry)
-		return;
-
 	if (x->xso.flags & XFRM_OFFLOAD_INBOUND)
 		mlx5e_ipsec_sadb_rx_del(sa_entry);
 }

@@ -387,9 +376,6 @@ static void mlx5e_xfrm_free_state(struct xfrm_state *x)
 	struct mlx5e_ipsec_sa_entry *sa_entry = to_ipsec_sa_entry(x);
 	struct mlx5e_priv *priv = netdev_priv(x->xso.dev);
 
-	if (!sa_entry)
-		return;
-
 	if (sa_entry->hw_context) {
 		flush_workqueue(sa_entry->ipsec->wq);
 		mlx5e_xfrm_fs_del_rule(priv, sa_entry);

@@ -402,7 +388,8 @@ static void mlx5e_xfrm_free_state(struct xfrm_state *x)
 
 int mlx5e_ipsec_init(struct mlx5e_priv *priv)
 {
-	struct mlx5e_ipsec *ipsec = NULL;
+	struct mlx5e_ipsec *ipsec;
+	int ret;
 
 	if (!mlx5_ipsec_device_caps(priv->mdev)) {
 		netdev_dbg(priv->netdev, "Not an IPSec offload device\n");

@@ -420,14 +407,23 @@ int mlx5e_ipsec_init(struct mlx5e_priv *priv)
 	ipsec->wq = alloc_ordered_workqueue("mlx5e_ipsec: %s", 0,
 					    priv->netdev->name);
 	if (!ipsec->wq) {
-		kfree(ipsec);
-		return -ENOMEM;
+		ret = -ENOMEM;
+		goto err_wq;
 	}
 
+	ret = mlx5e_ipsec_fs_init(ipsec);
+	if (ret)
+		goto err_fs_init;
+
 	priv->ipsec = ipsec;
-	mlx5e_ipsec_fs_init(ipsec);
 	netdev_dbg(priv->netdev, "IPSec attached to netdevice\n");
 	return 0;
+
+err_fs_init:
+	destroy_workqueue(ipsec->wq);
+err_wq:
+	kfree(ipsec);
+	return (ret != -EOPNOTSUPP) ? ret : 0;
 }
 
 void mlx5e_ipsec_cleanup(struct mlx5e_priv *priv)

@@ -487,9 +483,6 @@ static void mlx5e_xfrm_advance_esn_state(struct xfrm_state *x)
 	struct mlx5e_ipsec_modify_state_work *modify_work;
 	bool need_update;
 
-	if (!sa_entry)
-		return;
-
 	need_update = mlx5e_ipsec_update_esn_state(sa_entry);
 	if (!need_update)
 		return;

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c +0 -6

index 66b8ead8b579..dcc6ff0fc521 100644
@@ -605,9 +605,6 @@ int mlx5e_accel_ipsec_fs_add_rule(struct mlx5e_priv *priv,
 				  u32 ipsec_obj_id,
 				  struct mlx5e_ipsec_rule *ipsec_rule)
 {
-	if (!priv->ipsec->rx_fs)
-		return -EOPNOTSUPP;
-
 	if (attrs->action == MLX5_ACCEL_ESP_ACTION_DECRYPT)
 		return rx_add_rule(priv, attrs, ipsec_obj_id, ipsec_rule);
 	else

@@ -618,9 +615,6 @@ void mlx5e_accel_ipsec_fs_del_rule(struct mlx5e_priv *priv,
 				   struct mlx5_accel_esp_xfrm_attrs *attrs,
 				   struct mlx5e_ipsec_rule *ipsec_rule)
 {
-	if (!priv->ipsec->rx_fs)
-		return;
-
 	if (attrs->action == MLX5_ACCEL_ESP_ACTION_DECRYPT)
 		rx_del_rule(priv, attrs, ipsec_rule);
 	else

Signature · 1 line

2.35.1

Leon Romanovsky

author

2022-04-10 08:29 UTC #10 lore raw

PATCH 4/17 net/mlx5: Reduce useless indirection in IPsec FS add/delete flows

In reply to Leon Romanovsky

From: Leon Romanovsky <[email protected]>

There is no need in one-liners wrappers to call internal functions. Let's remove them, together with rename of functions to remove _accel_ notation from their names.

Reviewed-by Raed Salem <[email protected]>

Signed-off-by Leon Romanovsky <[email protected]>

---
 .../mellanox/mlx5/core/en_accel/ipsec.c       | 25 ++++++-------------
 .../mellanox/mlx5/core/en_accel/ipsec_fs.c    | 14 +++++------
 .../mellanox/mlx5/core/en_accel/ipsec_fs.h    | 14 +++++------
 3 files changed, 21 insertions(+), 32 deletions(-)

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c +7 ~~-18~~

index be30b6e2a00f..7f8aad4cad18 100644
@@ -271,21 +271,6 @@ static inline int mlx5e_xfrm_validate_state(struct xfrm_state *x)
 	return 0;
 }
 
-static int mlx5e_xfrm_fs_add_rule(struct mlx5e_priv *priv,
-				  struct mlx5e_ipsec_sa_entry *sa_entry)
-{
-	return mlx5e_accel_ipsec_fs_add_rule(priv, &sa_entry->xfrm->attrs,
-					     sa_entry->ipsec_obj_id,
-					     &sa_entry->ipsec_rule);
-}
-
-static void mlx5e_xfrm_fs_del_rule(struct mlx5e_priv *priv,
-				   struct mlx5e_ipsec_sa_entry *sa_entry)
-{
-	mlx5e_accel_ipsec_fs_del_rule(priv, &sa_entry->xfrm->attrs,
-				      &sa_entry->ipsec_rule);
-}
-
 static int mlx5e_xfrm_add_state(struct xfrm_state *x)
 {
 	struct mlx5e_ipsec_sa_entry *sa_entry = NULL;

@@ -334,7 +319,9 @@ static int mlx5e_xfrm_add_state(struct xfrm_state *x)
 	}
 
 	sa_entry->ipsec_obj_id = sa_handle;
-	err = mlx5e_xfrm_fs_add_rule(priv, sa_entry);
+	err = mlx5e_ipsec_fs_add_rule(priv, &sa_entry->xfrm->attrs,
+				      sa_entry->ipsec_obj_id,
+				      &sa_entry->ipsec_rule);
 	if (err)
 		goto err_hw_ctx;

@@ -351,7 +338,8 @@ static int mlx5e_xfrm_add_state(struct xfrm_state *x)
 	goto out;
 
 err_add_rule:
-	mlx5e_xfrm_fs_del_rule(priv, sa_entry);
+	mlx5e_ipsec_fs_del_rule(priv, &sa_entry->xfrm->attrs,
+				&sa_entry->ipsec_rule);
 err_hw_ctx:
 	mlx5_accel_esp_free_hw_context(priv->mdev, sa_entry->hw_context);
 err_xfrm:

@@ -378,7 +366,8 @@ static void mlx5e_xfrm_free_state(struct xfrm_state *x)
 
 	if (sa_entry->hw_context) {
 		flush_workqueue(sa_entry->ipsec->wq);
-		mlx5e_xfrm_fs_del_rule(priv, sa_entry);
+		mlx5e_ipsec_fs_del_rule(priv, &sa_entry->xfrm->attrs,
+					&sa_entry->ipsec_rule);
 		mlx5_accel_esp_free_hw_context(sa_entry->xfrm->mdev, sa_entry->hw_context);
 		mlx5_accel_esp_destroy_xfrm(sa_entry->xfrm);
 	}

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c +7 -7

index dcc6ff0fc521..3b715da86e6d 100644
@@ -600,10 +600,10 @@ static void tx_del_rule(struct mlx5e_priv *priv,
 	tx_ft_put(priv);
 }
 
-int mlx5e_accel_ipsec_fs_add_rule(struct mlx5e_priv *priv,
-				  struct mlx5_accel_esp_xfrm_attrs *attrs,
-				  u32 ipsec_obj_id,
-				  struct mlx5e_ipsec_rule *ipsec_rule)
+int mlx5e_ipsec_fs_add_rule(struct mlx5e_priv *priv,
+			    struct mlx5_accel_esp_xfrm_attrs *attrs,
+			    u32 ipsec_obj_id,
+			    struct mlx5e_ipsec_rule *ipsec_rule)
 {
 	if (attrs->action == MLX5_ACCEL_ESP_ACTION_DECRYPT)
 		return rx_add_rule(priv, attrs, ipsec_obj_id, ipsec_rule);

@@ -611,9 +611,9 @@ int mlx5e_accel_ipsec_fs_add_rule(struct mlx5e_priv *priv,
 		return tx_add_rule(priv, attrs, ipsec_obj_id, ipsec_rule);
 }
 
-void mlx5e_accel_ipsec_fs_del_rule(struct mlx5e_priv *priv,
-				   struct mlx5_accel_esp_xfrm_attrs *attrs,
-				   struct mlx5e_ipsec_rule *ipsec_rule)
+void mlx5e_ipsec_fs_del_rule(struct mlx5e_priv *priv,
+			     struct mlx5_accel_esp_xfrm_attrs *attrs,
+			     struct mlx5e_ipsec_rule *ipsec_rule)
 {
 	if (attrs->action == MLX5_ACCEL_ESP_ACTION_DECRYPT)
 		rx_del_rule(priv, attrs, ipsec_rule);

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.h +7 -7

index 8e0e829ab58f..011af408ee03 100644
@@ -11,11 +11,11 @@
 
 void mlx5e_ipsec_fs_cleanup(struct mlx5e_ipsec *ipsec);
 int mlx5e_ipsec_fs_init(struct mlx5e_ipsec *ipsec);
-int mlx5e_accel_ipsec_fs_add_rule(struct mlx5e_priv *priv,
-				  struct mlx5_accel_esp_xfrm_attrs *attrs,
-				  u32 ipsec_obj_id,
-				  struct mlx5e_ipsec_rule *ipsec_rule);
-void mlx5e_accel_ipsec_fs_del_rule(struct mlx5e_priv *priv,
-				   struct mlx5_accel_esp_xfrm_attrs *attrs,
-				   struct mlx5e_ipsec_rule *ipsec_rule);
+int mlx5e_ipsec_fs_add_rule(struct mlx5e_priv *priv,
+			    struct mlx5_accel_esp_xfrm_attrs *attrs,
+			    u32 ipsec_obj_id,
+			    struct mlx5e_ipsec_rule *ipsec_rule);
+void mlx5e_ipsec_fs_del_rule(struct mlx5e_priv *priv,
+			     struct mlx5_accel_esp_xfrm_attrs *attrs,
+			     struct mlx5e_ipsec_rule *ipsec_rule);
 #endif /* __MLX5_IPSEC_STEERING_H__ */

Signature · 1 line

2.35.1

Leon Romanovsky

author

2022-04-10 08:29 UTC #11 lore raw

PATCH 5/17 net/mlx5: Store IPsec ESN update work in XFRM state

In reply to Leon Romanovsky

From: Leon Romanovsky <[email protected]>

mlx5 IPsec code updated ESN through workqueue with allocation calls in the data path, which can be saved easily if the work is created during XFRM state initialization routine.

The locking used later in the work didn't protect from anything because change of HW context is possible during XFRM state add or delete only, which can cancel work and make sure that it is not running.

Reviewed-by Raed Salem <[email protected]>

Signed-off-by Leon Romanovsky <[email protected]>

---
 .../mellanox/mlx5/core/en_accel/ipsec.c       | 49 ++++++-------------
 .../mellanox/mlx5/core/en_accel/ipsec.h       |  7 ++-
 .../mlx5/core/en_accel/ipsec_offload.c        | 39 +++------------
 include/linux/mlx5/accel.h                    | 10 ++--
 4 files changed, 33 insertions(+), 72 deletions(-)

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c +15 ~~-34~~

index 7f8aad4cad18..142a2e8313f1 100644
@@ -271,6 +271,16 @@ static inline int mlx5e_xfrm_validate_state(struct xfrm_state *x)
 	return 0;
 }
 
+static void _update_xfrm_state(struct work_struct *work)
+{
+	struct mlx5e_ipsec_modify_state_work *modify_work =
+		container_of(work, struct mlx5e_ipsec_modify_state_work, work);
+	struct mlx5e_ipsec_sa_entry *sa_entry = container_of(
+		modify_work, struct mlx5e_ipsec_sa_entry, modify_work);
+
+	mlx5_accel_esp_modify_xfrm(sa_entry->xfrm, &modify_work->attrs);
+}
+
 static int mlx5e_xfrm_add_state(struct xfrm_state *x)
 {
 	struct mlx5e_ipsec_sa_entry *sa_entry = NULL;

@@ -334,6 +344,7 @@ static int mlx5e_xfrm_add_state(struct xfrm_state *x)
 				mlx5e_ipsec_set_iv_esn : mlx5e_ipsec_set_iv;
 	}
 
+	INIT_WORK(&sa_entry->modify_work.work, _update_xfrm_state);
 	x->xso.offload_handle = (unsigned long)sa_entry;
 	goto out;

@@ -365,7 +376,7 @@ static void mlx5e_xfrm_free_state(struct xfrm_state *x)
 	struct mlx5e_priv *priv = netdev_priv(x->xso.dev);
 
 	if (sa_entry->hw_context) {
-		flush_workqueue(sa_entry->ipsec->wq);
+		cancel_work_sync(&sa_entry->modify_work.work);
 		mlx5e_ipsec_fs_del_rule(priv, &sa_entry->xfrm->attrs,
 					&sa_entry->ipsec_rule);
 		mlx5_accel_esp_free_hw_context(sa_entry->xfrm->mdev, sa_entry->hw_context);

@@ -392,7 +403,6 @@ int mlx5e_ipsec_init(struct mlx5e_priv *priv)
 	hash_init(ipsec->sadb_rx);
 	spin_lock_init(&ipsec->sadb_rx_lock);
 	ipsec->mdev = priv->mdev;
-	ipsec->en_priv = priv;
 	ipsec->wq = alloc_ordered_workqueue("mlx5e_ipsec: %s", 0,
 					    priv->netdev->name);
 	if (!ipsec->wq) {

@@ -424,7 +434,6 @@ void mlx5e_ipsec_cleanup(struct mlx5e_priv *priv)
 
 	mlx5e_ipsec_fs_cleanup(ipsec);
 	destroy_workqueue(ipsec->wq);
-
 	kfree(ipsec);
 	priv->ipsec = NULL;
 }

@@ -444,47 +453,19 @@ static bool mlx5e_ipsec_offload_ok(struct sk_buff *skb, struct xfrm_state *x)
 	return true;
 }
 
-struct mlx5e_ipsec_modify_state_work {
-	struct work_struct		work;
-	struct mlx5_accel_esp_xfrm_attrs attrs;
-	struct mlx5e_ipsec_sa_entry	*sa_entry;
-};
-
-static void _update_xfrm_state(struct work_struct *work)
-{
-	int ret;
-	struct mlx5e_ipsec_modify_state_work *modify_work =
-		container_of(work, struct mlx5e_ipsec_modify_state_work, work);
-	struct mlx5e_ipsec_sa_entry *sa_entry = modify_work->sa_entry;
-
-	ret = mlx5_accel_esp_modify_xfrm(sa_entry->xfrm,
-					 &modify_work->attrs);
-	if (ret)
-		netdev_warn(sa_entry->ipsec->en_priv->netdev,
-			    "Not an IPSec offload device\n");
-
-	kfree(modify_work);
-}
-
 static void mlx5e_xfrm_advance_esn_state(struct xfrm_state *x)
 {
 	struct mlx5e_ipsec_sa_entry *sa_entry = to_ipsec_sa_entry(x);
-	struct mlx5e_ipsec_modify_state_work *modify_work;
+	struct mlx5e_ipsec_modify_state_work *modify_work =
+		&sa_entry->modify_work;
 	bool need_update;
 
 	need_update = mlx5e_ipsec_update_esn_state(sa_entry);
 	if (!need_update)
 		return;
 
-	modify_work = kzalloc(sizeof(*modify_work), GFP_ATOMIC);
-	if (!modify_work)
-		return;
-
 	mlx5e_ipsec_build_accel_xfrm_attrs(sa_entry, &modify_work->attrs);
-	modify_work->sa_entry = sa_entry;
-
-	INIT_WORK(&modify_work->work, _update_xfrm_state);
-	WARN_ON(!queue_work(sa_entry->ipsec->wq, &modify_work->work));
+	queue_work(sa_entry->ipsec->wq, &modify_work->work);
 }
 
 static const struct xfrmdev_ops mlx5e_ipsec_xfrmdev_ops = {

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h +6 -1

index bbf48d4616f9..35a751faeb33 100644
@@ -62,7 +62,6 @@ struct mlx5e_ipsec_tx;
 
 struct mlx5e_ipsec {
 	struct mlx5_core_dev *mdev;
-	struct mlx5e_priv *en_priv;
 	DECLARE_HASHTABLE(sadb_rx, MLX5E_IPSEC_SADB_RX_BITS);
 	spinlock_t sadb_rx_lock; /* Protects sadb_rx */
 	struct mlx5e_ipsec_sw_stats sw_stats;

@@ -82,6 +81,11 @@ struct mlx5e_ipsec_rule {
 	struct mlx5_modify_hdr *set_modify_hdr;
 };
 
+struct mlx5e_ipsec_modify_state_work {
+	struct work_struct		work;
+	struct mlx5_accel_esp_xfrm_attrs attrs;
+};
+
 struct mlx5e_ipsec_sa_entry {
 	struct hlist_node hlist; /* Item in SADB_RX hashtable */
 	struct mlx5e_ipsec_esn_state esn_state;

@@ -94,6 +98,7 @@ struct mlx5e_ipsec_sa_entry {
 			  struct xfrm_offload *xo);
 	u32 ipsec_obj_id;
 	struct mlx5e_ipsec_rule ipsec_rule;
+	struct mlx5e_ipsec_modify_state_work modify_work;
 };
 
 int mlx5e_ipsec_init(struct mlx5e_priv *priv);

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_offload.c +6 ~~-33~~

index 37c9880719cf..bbfb6643ed80 100644
@@ -18,7 +18,6 @@ struct mlx5_ipsec_sa_ctx {
 struct mlx5_ipsec_esp_xfrm {
 	/* reference counter of SA ctx */
 	struct mlx5_ipsec_sa_ctx *sa_ctx;
-	struct mutex lock; /* protects mlx5_ipsec_esp_xfrm */
 	struct mlx5_accel_esp_xfrm accel_xfrm;
 };

@@ -117,7 +116,6 @@ mlx5_ipsec_offload_esp_create_xfrm(struct mlx5_core_dev *mdev,
 	if (!mxfrm)
 		return ERR_PTR(-ENOMEM);
 
-	mutex_init(&mxfrm->lock);
 	memcpy(&mxfrm->accel_xfrm.attrs, attrs,
 	       sizeof(mxfrm->accel_xfrm.attrs));

@@ -129,8 +127,6 @@ static void mlx5_ipsec_offload_esp_destroy_xfrm(struct mlx5_accel_esp_xfrm *xfrm
 	struct mlx5_ipsec_esp_xfrm *mxfrm = container_of(xfrm, struct mlx5_ipsec_esp_xfrm,
 							 accel_xfrm);
 
-	/* assuming no sa_ctx are connected to this xfrm_ctx */
-	WARN_ON(mxfrm->sa_ctx);
 	kfree(mxfrm);
 }

@@ -232,7 +228,6 @@ static void *mlx5_ipsec_offload_create_sa_ctx(struct mlx5_core_dev *mdev,
 	sa_ctx->dev = mdev;
 
 	mxfrm = container_of(accel_xfrm, struct mlx5_ipsec_esp_xfrm, accel_xfrm);
-	mutex_lock(&mxfrm->lock);
 	sa_ctx->mxfrm = mxfrm;
 
 	/* key */

@@ -258,14 +253,12 @@ static void *mlx5_ipsec_offload_create_sa_ctx(struct mlx5_core_dev *mdev,
 
 	*hw_handle = sa_ctx->ipsec_obj_id;
 	mxfrm->sa_ctx = sa_ctx;
-	mutex_unlock(&mxfrm->lock);
 
 	return sa_ctx;
 
 err_enc_key:
 	mlx5_destroy_encryption_key(mdev, sa_ctx->enc_key_id);
 err_sa_ctx:
-	mutex_unlock(&mxfrm->lock);
 	kfree(sa_ctx);
 	return ERR_PTR(err);
 }

@@ -273,14 +266,10 @@ static void *mlx5_ipsec_offload_create_sa_ctx(struct mlx5_core_dev *mdev,
 static void mlx5_ipsec_offload_delete_sa_ctx(void *context)
 {
 	struct mlx5_ipsec_sa_ctx *sa_ctx = (struct mlx5_ipsec_sa_ctx *)context;
-	struct mlx5_ipsec_esp_xfrm *mxfrm = sa_ctx->mxfrm;
 
-	mutex_lock(&mxfrm->lock);
 	mlx5_destroy_ipsec_obj(sa_ctx->dev, sa_ctx->ipsec_obj_id);
 	mlx5_destroy_encryption_key(sa_ctx->dev, sa_ctx->enc_key_id);
 	kfree(sa_ctx);
-	mxfrm->sa_ctx = NULL;
-	mutex_unlock(&mxfrm->lock);
 }
 
 static int mlx5_modify_ipsec_obj(struct mlx5_core_dev *mdev,

@@ -331,29 +320,17 @@ static int mlx5_modify_ipsec_obj(struct mlx5_core_dev *mdev,
 	return mlx5_cmd_exec(mdev, in, sizeof(in), out, sizeof(out));
 }
 
-static int mlx5_ipsec_offload_esp_modify_xfrm(struct mlx5_accel_esp_xfrm *xfrm,
-					      const struct mlx5_accel_esp_xfrm_attrs *attrs)
+static void mlx5_ipsec_offload_esp_modify_xfrm(
+	struct mlx5_accel_esp_xfrm *xfrm,
+	const struct mlx5_accel_esp_xfrm_attrs *attrs)
 {
 	struct mlx5_ipsec_obj_attrs ipsec_attrs = {};
 	struct mlx5_core_dev *mdev = xfrm->mdev;
 	struct mlx5_ipsec_esp_xfrm *mxfrm;
-
 	int err = 0;
 
-	if (!memcmp(&xfrm->attrs, attrs, sizeof(xfrm->attrs)))
-		return 0;
-
-	if (mlx5_ipsec_offload_esp_validate_xfrm_attrs(mdev, attrs))
-		return -EOPNOTSUPP;
-
 	mxfrm = container_of(xfrm, struct mlx5_ipsec_esp_xfrm, accel_xfrm);
 
-	mutex_lock(&mxfrm->lock);
-
-	if (!mxfrm->sa_ctx)
-		/* Not bound xfrm, change only sw attrs */
-		goto change_sw_xfrm_attrs;
-
 	/* need to add find and replace in ipsec_rhash_sa the sa_ctx */
 	/* modify device with new hw_sa */
 	ipsec_attrs.accel_flags = attrs->flags;

@@ -362,12 +339,8 @@ static int mlx5_ipsec_offload_esp_modify_xfrm(struct mlx5_accel_esp_xfrm *xfrm,
 				    &ipsec_attrs,
 				    mxfrm->sa_ctx->ipsec_obj_id);
 
-change_sw_xfrm_attrs:
 	if (!err)
 		memcpy(&xfrm->attrs, attrs, sizeof(xfrm->attrs));
-
-	mutex_unlock(&mxfrm->lock);
-	return err;
 }
 
 void *mlx5_accel_esp_create_hw_context(struct mlx5_core_dev *mdev,

@@ -413,8 +386,8 @@ void mlx5_accel_esp_destroy_xfrm(struct mlx5_accel_esp_xfrm *xfrm)
 	mlx5_ipsec_offload_esp_destroy_xfrm(xfrm);
 }
 
-int mlx5_accel_esp_modify_xfrm(struct mlx5_accel_esp_xfrm *xfrm,
-			       const struct mlx5_accel_esp_xfrm_attrs *attrs)
+void mlx5_accel_esp_modify_xfrm(struct mlx5_accel_esp_xfrm *xfrm,
+				const struct mlx5_accel_esp_xfrm_attrs *attrs)
 {
-	return mlx5_ipsec_offload_esp_modify_xfrm(xfrm, attrs);
+	mlx5_ipsec_offload_esp_modify_xfrm(xfrm, attrs);
 }

include/linux/mlx5/accel.h +6 -4

index 0f2596297f6a..a2720ebbb9fd 100644
@@ -127,8 +127,8 @@ struct mlx5_accel_esp_xfrm *
 mlx5_accel_esp_create_xfrm(struct mlx5_core_dev *mdev,
 			   const struct mlx5_accel_esp_xfrm_attrs *attrs);
 void mlx5_accel_esp_destroy_xfrm(struct mlx5_accel_esp_xfrm *xfrm);
-int mlx5_accel_esp_modify_xfrm(struct mlx5_accel_esp_xfrm *xfrm,
-			       const struct mlx5_accel_esp_xfrm_attrs *attrs);
+void mlx5_accel_esp_modify_xfrm(struct mlx5_accel_esp_xfrm *xfrm,
+				const struct mlx5_accel_esp_xfrm_attrs *attrs);
 
 #else

@@ -145,9 +145,11 @@ mlx5_accel_esp_create_xfrm(struct mlx5_core_dev *mdev,
 }
 static inline void
 mlx5_accel_esp_destroy_xfrm(struct mlx5_accel_esp_xfrm *xfrm) {}
-static inline int
+static inline void
 mlx5_accel_esp_modify_xfrm(struct mlx5_accel_esp_xfrm *xfrm,
-			   const struct mlx5_accel_esp_xfrm_attrs *attrs) { return -EOPNOTSUPP; }
+			   const struct mlx5_accel_esp_xfrm_attrs *attrs)
+{
+}
 
 #endif /* CONFIG_MLX5_EN_IPSEC */
 #endif /* __MLX5_ACCEL_H__ */

Signature · 1 line

2.35.1

Leon Romanovsky

author

2022-04-10 08:29 UTC #12 lore raw

PATCH 6/17 net/mlx5: Remove useless validity check

In reply to Leon Romanovsky

From: Leon Romanovsky <[email protected]>

All callers build xfrm attributes with help of mlx5e_ipsec_build_accel_xfrm_attrs() function that ensure validity of attributes. There is no need to recheck them again.

Reviewed-by Raed Salem <[email protected]>

Signed-off-by Leon Romanovsky <[email protected]>

---
 .../mlx5/core/en_accel/ipsec_offload.c        | 44 -------------------
 include/linux/mlx5/accel.h                    | 10 -----
 2 files changed, 54 deletions(-)

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_offload.c +0 ~~-44~~

index bbfb6643ed80..9d2932cf12f1 100644
@@ -62,55 +62,11 @@ u32 mlx5_ipsec_device_caps(struct mlx5_core_dev *mdev)
 }
 EXPORT_SYMBOL_GPL(mlx5_ipsec_device_caps);
 
-static int
-mlx5_ipsec_offload_esp_validate_xfrm_attrs(struct mlx5_core_dev *mdev,
-					   const struct mlx5_accel_esp_xfrm_attrs *attrs)
-{
-	if (attrs->replay_type != MLX5_ACCEL_ESP_REPLAY_NONE) {
-		mlx5_core_err(mdev, "Cannot offload xfrm states with anti replay (replay_type = %d)\n",
-			      attrs->replay_type);
-		return -EOPNOTSUPP;
-	}
-
-	if (attrs->keymat_type != MLX5_ACCEL_ESP_KEYMAT_AES_GCM) {
-		mlx5_core_err(mdev, "Only aes gcm keymat is supported (keymat_type = %d)\n",
-			      attrs->keymat_type);
-		return -EOPNOTSUPP;
-	}
-
-	if (attrs->keymat.aes_gcm.iv_algo !=
-	    MLX5_ACCEL_ESP_AES_GCM_IV_ALGO_SEQ) {
-		mlx5_core_err(mdev, "Only iv sequence algo is supported (iv_algo = %d)\n",
-			      attrs->keymat.aes_gcm.iv_algo);
-		return -EOPNOTSUPP;
-	}
-
-	if (attrs->keymat.aes_gcm.key_len != 128 &&
-	    attrs->keymat.aes_gcm.key_len != 256) {
-		mlx5_core_err(mdev, "Cannot offload xfrm states with key length other than 128/256 bit (key length = %d)\n",
-			      attrs->keymat.aes_gcm.key_len);
-		return -EOPNOTSUPP;
-	}
-
-	if ((attrs->flags & MLX5_ACCEL_ESP_FLAGS_ESN_TRIGGERED) &&
-	    !MLX5_CAP_IPSEC(mdev, ipsec_esn)) {
-		mlx5_core_err(mdev, "Cannot offload xfrm states with ESN triggered\n");
-		return -EOPNOTSUPP;
-	}
-
-	return 0;
-}
-
 static struct mlx5_accel_esp_xfrm *
 mlx5_ipsec_offload_esp_create_xfrm(struct mlx5_core_dev *mdev,
 				   const struct mlx5_accel_esp_xfrm_attrs *attrs)
 {
 	struct mlx5_ipsec_esp_xfrm *mxfrm;
-	int err = 0;
-
-	err = mlx5_ipsec_offload_esp_validate_xfrm_attrs(mdev, attrs);
-	if (err)
-		return ERR_PTR(err);
 
 	mxfrm = kzalloc(sizeof(*mxfrm), GFP_KERNEL);
 	if (!mxfrm)

include/linux/mlx5/accel.h +0 ~~-10~~

index a2720ebbb9fd..9c511d466e55 100644
@@ -36,10 +36,6 @@
 
 #include <linux/mlx5/driver.h>
 
-enum mlx5_accel_esp_aes_gcm_keymat_iv_algo {
-	MLX5_ACCEL_ESP_AES_GCM_IV_ALGO_SEQ,
-};
-
 enum mlx5_accel_esp_flags {
 	MLX5_ACCEL_ESP_FLAGS_TUNNEL            = 0,    /* Default */
 	MLX5_ACCEL_ESP_FLAGS_TRANSPORT         = 1UL << 0,

@@ -57,14 +53,9 @@ enum mlx5_accel_esp_keymats {
 	MLX5_ACCEL_ESP_KEYMAT_AES_GCM,
 };
 
-enum mlx5_accel_esp_replay {
-	MLX5_ACCEL_ESP_REPLAY_NONE,
-	MLX5_ACCEL_ESP_REPLAY_BMP,
-};
 
 struct aes_gcm_keymat {
 	u64   seq_iv;
-	enum mlx5_accel_esp_aes_gcm_keymat_iv_algo iv_algo;
 
 	u32   salt;
 	u32   icv_len;

@@ -81,7 +72,6 @@ struct mlx5_accel_esp_xfrm_attrs {
 	u32   tfc_pad;
 	u32   flags;
 	u32   sa_handle;
-	enum mlx5_accel_esp_replay replay_type;
 	union {
 		struct {
 			u32 size;

Signature · 1 line

2.35.1

Leon Romanovsky

author

2022-04-10 08:29 UTC #13 lore raw

PATCH 7/17 net/mlx5: Merge various control path IPsec headers into one file

In reply to Leon Romanovsky

From: Leon Romanovsky <[email protected]>

The mlx5 IPsec code has logical separation between code that operates with XFRM objects (ipsec.c), HW objects (ipsec_offload.c), flow steering logic (ipsec_fs.c) and data path (ipsec_rxtx.c).

Such separation makes sense for C-files, but isn't needed at all for H-files as they are included in batch anyway.

Reviewed-by Raed Salem <[email protected]>

Signed-off-by Leon Romanovsky <[email protected]>

---
 .../ethernet/mellanox/mlx5/core/en/params.c   |   2 +-
 .../mellanox/mlx5/core/en_accel/ipsec.c       |   5 +-
 .../mellanox/mlx5/core/en_accel/ipsec.h       | 101 +++++++++++-
 .../mellanox/mlx5/core/en_accel/ipsec_fs.c    |  23 +--
 .../mellanox/mlx5/core/en_accel/ipsec_fs.h    |  21 ---
 .../mlx5/core/en_accel/ipsec_offload.c        |   3 +-
 .../mlx5/core/en_accel/ipsec_offload.h        |  14 --
 .../mellanox/mlx5/core/en_accel/ipsec_rxtx.c  |   5 +-
 .../mellanox/mlx5/core/en_accel/ipsec_stats.c |   4 +-
 .../net/ethernet/mellanox/mlx5/core/en_main.c |   1 -
 .../net/ethernet/mellanox/mlx5/core/en_rx.c   |   2 +-
 .../net/ethernet/mellanox/mlx5/core/main.c    |   2 +-
 include/linux/mlx5/accel.h                    | 145 ------------------
 13 files changed, 120 insertions(+), 208 deletions(-)
 delete mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.h
 delete mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_offload.h
 delete mode 100644 include/linux/mlx5/accel.h

drivers/net/ethernet/mellanox/mlx5/core/en/params.c +1 -1

index 1e8700957280..3c1edfa33aa7 100644
@@ -5,7 +5,7 @@
 #include "en/txrx.h"
 #include "en/port.h"
 #include "en_accel/en_accel.h"
-#include "en_accel/ipsec_offload.h"
+#include "en_accel/ipsec.h"
 
 static bool mlx5e_rx_is_xdp(struct mlx5e_params *params,
 			    struct mlx5e_xsk_param *xsk)

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c +2 -3

index 142a2e8313f1..512a5ab9291d 100644
@@ -37,9 +37,8 @@
 #include <linux/netdevice.h>
 
 #include "en.h"
-#include "en_accel/ipsec.h"
-#include "en_accel/ipsec_rxtx.h"
-#include "en_accel/ipsec_fs.h"
+#include "ipsec.h"
+#include "ipsec_rxtx.h"
 
 static struct mlx5e_ipsec_sa_entry *to_ipsec_sa_entry(struct xfrm_state *x)
 {

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h +99 -2

122 lines not on this page · raw message

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c +12 ~~-11~~

index 3b715da86e6d..e278a8fd43ae 100644
@@ -2,8 +2,9 @@
 /* Copyright (c) 2020, Mellanox Technologies inc. All rights reserved. */
 
 #include <linux/netdevice.h>
-#include "ipsec_offload.h"
-#include "ipsec_fs.h"
+#include "en.h"
+#include "en/fs.h"
+#include "ipsec.h"
 #include "fs_core.h"
 
 #define NUM_IPSEC_FTE BIT(15)

@@ -565,7 +566,7 @@ static int tx_add_rule(struct mlx5e_priv *priv,
 	if (IS_ERR(rule)) {
 		err = PTR_ERR(rule);
 		netdev_err(priv->netdev, "fail to add ipsec rule attrs->action=0x%x, err=%d\n",
-			   attrs->action, err);
+				attrs->action, err);
 		goto out;
 	}

@@ -579,8 +580,8 @@ static int tx_add_rule(struct mlx5e_priv *priv,
 }
 
 static void rx_del_rule(struct mlx5e_priv *priv,
-			struct mlx5_accel_esp_xfrm_attrs *attrs,
-			struct mlx5e_ipsec_rule *ipsec_rule)
+		struct mlx5_accel_esp_xfrm_attrs *attrs,
+		struct mlx5e_ipsec_rule *ipsec_rule)
 {
 	mlx5_del_flow_rules(ipsec_rule->rule);
 	ipsec_rule->rule = NULL;

@@ -592,7 +593,7 @@ static void rx_del_rule(struct mlx5e_priv *priv,
 }
 
 static void tx_del_rule(struct mlx5e_priv *priv,
-			struct mlx5e_ipsec_rule *ipsec_rule)
+		struct mlx5e_ipsec_rule *ipsec_rule)
 {
 	mlx5_del_flow_rules(ipsec_rule->rule);
 	ipsec_rule->rule = NULL;

@@ -601,9 +602,9 @@ static void tx_del_rule(struct mlx5e_priv *priv,
 }
 
 int mlx5e_ipsec_fs_add_rule(struct mlx5e_priv *priv,
-			    struct mlx5_accel_esp_xfrm_attrs *attrs,
-			    u32 ipsec_obj_id,
-			    struct mlx5e_ipsec_rule *ipsec_rule)
+		struct mlx5_accel_esp_xfrm_attrs *attrs,
+		u32 ipsec_obj_id,
+		struct mlx5e_ipsec_rule *ipsec_rule)
 {
 	if (attrs->action == MLX5_ACCEL_ESP_ACTION_DECRYPT)
 		return rx_add_rule(priv, attrs, ipsec_obj_id, ipsec_rule);

@@ -612,8 +613,8 @@ int mlx5e_ipsec_fs_add_rule(struct mlx5e_priv *priv,
 }
 
 void mlx5e_ipsec_fs_del_rule(struct mlx5e_priv *priv,
-			     struct mlx5_accel_esp_xfrm_attrs *attrs,
-			     struct mlx5e_ipsec_rule *ipsec_rule)
+		struct mlx5_accel_esp_xfrm_attrs *attrs,
+		struct mlx5e_ipsec_rule *ipsec_rule)
 {
 	if (attrs->action == MLX5_ACCEL_ESP_ACTION_DECRYPT)
 		rx_del_rule(priv, attrs, ipsec_rule);

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.h deleted +0 ~~-21~~

deleted file mode 100644
index 011af408ee03..000000000000
@@ -1,21 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */
-/* Copyright (c) 2020, Mellanox Technologies inc. All rights reserved. */
-
-#ifndef __MLX5_IPSEC_STEERING_H__
-#define __MLX5_IPSEC_STEERING_H__
-
-#include "en.h"
-#include "ipsec.h"
-#include "ipsec_offload.h"
-#include "en/fs.h"
-
-void mlx5e_ipsec_fs_cleanup(struct mlx5e_ipsec *ipsec);
-int mlx5e_ipsec_fs_init(struct mlx5e_ipsec *ipsec);
-int mlx5e_ipsec_fs_add_rule(struct mlx5e_priv *priv,
-			    struct mlx5_accel_esp_xfrm_attrs *attrs,
-			    u32 ipsec_obj_id,
-			    struct mlx5e_ipsec_rule *ipsec_rule);
-void mlx5e_ipsec_fs_del_rule(struct mlx5e_priv *priv,
-			     struct mlx5_accel_esp_xfrm_attrs *attrs,
-			     struct mlx5e_ipsec_rule *ipsec_rule);
-#endif /* __MLX5_IPSEC_STEERING_H__ */

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_offload.c +1 -2

index 9d2932cf12f1..6c03ce8aba92 100644
@@ -2,9 +2,8 @@
 /* Copyright (c) 2017, Mellanox Technologies inc. All rights reserved. */
 
 #include "mlx5_core.h"
-#include "ipsec_offload.h"
+#include "ipsec.h"
 #include "lib/mlx5.h"
-#include "en_accel/ipsec_fs.h"
 
 struct mlx5_ipsec_sa_ctx {
 	struct rhash_head hash;

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_offload.h deleted +0 ~~-14~~

deleted file mode 100644
index 7dac104e6ef1..000000000000
@@ -1,14 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */
-/* Copyright (c) 2020, Mellanox Technologies inc. All rights reserved. */
-
-#ifndef __MLX5_IPSEC_OFFLOAD_H__
-#define __MLX5_IPSEC_OFFLOAD_H__
-
-#include <linux/mlx5/driver.h>
-#include <linux/mlx5/accel.h>
-
-void *mlx5_accel_esp_create_hw_context(struct mlx5_core_dev *mdev,
-				       struct mlx5_accel_esp_xfrm *xfrm,
-				       u32 *sa_handle);
-void mlx5_accel_esp_free_hw_context(struct mlx5_core_dev *mdev, void *context);
-#endif /* __MLX5_IPSEC_OFFLOAD_H__ */

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c +2 -3

index 9b65c765cbd9..d30922e1b60f 100644
@@ -34,9 +34,8 @@
 #include <crypto/aead.h>
 #include <net/xfrm.h>
 #include <net/esp.h>
-#include "ipsec_offload.h"
-#include "en_accel/ipsec_rxtx.h"
-#include "en_accel/ipsec.h"
+#include "ipsec.h"
+#include "ipsec_rxtx.h"
 #include "en.h"
 
 enum {

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_stats.c +1 -3

index 3aace1c2a763..9de84821dafb 100644
@@ -35,9 +35,7 @@
 #include <net/sock.h>
 
 #include "en.h"
-#include "ipsec_offload.h"
-#include "fpga/sdk.h"
-#include "en_accel/ipsec.h"
+#include "ipsec.h"
 
 static const struct counter_desc mlx5e_ipsec_sw_stats_desc[] = {
 	{ MLX5E_DECLARE_STAT(struct mlx5e_ipsec_sw_stats, ipsec_rx_drop_sp_alloc) },

drivers/net/ethernet/mellanox/mlx5/core/en_main.c +0 -1

index 12b72a0bcb1a..d27986869b8b 100644
@@ -48,7 +48,6 @@
 #include "en_accel/ipsec.h"
 #include "en_accel/en_accel.h"
 #include "en_accel/ktls.h"
-#include "en_accel/ipsec_offload.h"
 #include "lib/vxlan.h"
 #include "lib/clock.h"
 #include "en/port.h"

drivers/net/ethernet/mellanox/mlx5/core/en_rx.c +1 -1

index a5f6fd16b665..faf195b6deb8 100644
@@ -48,7 +48,7 @@
 #include "en_rep.h"
 #include "en/rep/tc.h"
 #include "ipoib/ipoib.h"
-#include "en_accel/ipsec_offload.h"
+#include "en_accel/ipsec.h"
 #include "en_accel/ipsec_rxtx.h"
 #include "en_accel/ktls_txrx.h"
 #include "en/xdp.h"

drivers/net/ethernet/mellanox/mlx5/core/main.c +1 -1

index d504c8cb8f96..27f3597a4964 100644
@@ -62,7 +62,7 @@
 #include "lib/mlx5.h"
 #include "lib/tout.h"
 #include "fpga/core.h"
-#include "en_accel/ipsec_offload.h"
+#include "en_accel/ipsec.h"
 #include "lib/clock.h"
 #include "lib/vxlan.h"
 #include "lib/geneve.h"

include/linux/mlx5/accel.h deleted +0 ~~-145~~

145 lines not on this page · raw message

Signature · 1 line

2.35.1

Leon Romanovsky

author

2022-04-10 08:29 UTC #14 lore raw

PATCH 8/17 net/mlx5: Remove accel notations and indirections from esp functions

In reply to Leon Romanovsky

From: Leon Romanovsky <[email protected]>

This change cleanups the mlx5 esp interface.

Reviewed-by Raed Salem <[email protected]>

Signed-off-by Leon Romanovsky <[email protected]>

---
 .../mellanox/mlx5/core/en_accel/ipsec.c       |  8 ++--
 .../mellanox/mlx5/core/en_accel/ipsec.h       |  6 +--
 .../mlx5/core/en_accel/ipsec_offload.c        | 47 +++++--------------
 3 files changed, 19 insertions(+), 42 deletions(-)

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c +4 -4

index 512a5ab9291d..792f01264196 100644
@@ -277,7 +277,7 @@ static void _update_xfrm_state(struct work_struct *work)
 	struct mlx5e_ipsec_sa_entry *sa_entry = container_of(
 		modify_work, struct mlx5e_ipsec_sa_entry, modify_work);
 
-	mlx5_accel_esp_modify_xfrm(sa_entry->xfrm, &modify_work->attrs);
+	mlx5_ipsec_esp_modify_xfrm(sa_entry->xfrm, &modify_work->attrs);
 }
 
 static int mlx5e_xfrm_add_state(struct xfrm_state *x)

@@ -311,7 +311,7 @@ static int mlx5e_xfrm_add_state(struct xfrm_state *x)
 
 	/* create xfrm */
 	mlx5e_ipsec_build_accel_xfrm_attrs(sa_entry, &attrs);
-	sa_entry->xfrm = mlx5_accel_esp_create_xfrm(priv->mdev, &attrs);
+	sa_entry->xfrm = mlx5_ipsec_esp_create_xfrm(priv->mdev, &attrs);
 	if (IS_ERR(sa_entry->xfrm)) {
 		err = PTR_ERR(sa_entry->xfrm);
 		goto err_sa_entry;

@@ -353,7 +353,7 @@ static int mlx5e_xfrm_add_state(struct xfrm_state *x)
 err_hw_ctx:
 	mlx5_accel_esp_free_hw_context(priv->mdev, sa_entry->hw_context);
 err_xfrm:
-	mlx5_accel_esp_destroy_xfrm(sa_entry->xfrm);
+	mlx5_ipsec_esp_destroy_xfrm(sa_entry->xfrm);
 err_sa_entry:
 	kfree(sa_entry);

@@ -379,7 +379,7 @@ static void mlx5e_xfrm_free_state(struct xfrm_state *x)
 		mlx5e_ipsec_fs_del_rule(priv, &sa_entry->xfrm->attrs,
 					&sa_entry->ipsec_rule);
 		mlx5_accel_esp_free_hw_context(sa_entry->xfrm->mdev, sa_entry->hw_context);
-		mlx5_accel_esp_destroy_xfrm(sa_entry->xfrm);
+		mlx5_ipsec_esp_destroy_xfrm(sa_entry->xfrm);
 	}
 
 	kfree(sa_entry);

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h +3 -3

index 71c7a7355473..84c8d835a26b 100644
@@ -196,10 +196,10 @@ void mlx5_accel_esp_free_hw_context(struct mlx5_core_dev *mdev, void *context);
 u32 mlx5_ipsec_device_caps(struct mlx5_core_dev *mdev);
 
 struct mlx5_accel_esp_xfrm *
-mlx5_accel_esp_create_xfrm(struct mlx5_core_dev *mdev,
+mlx5_ipsec_esp_create_xfrm(struct mlx5_core_dev *mdev,
 			   const struct mlx5_accel_esp_xfrm_attrs *attrs);
-void mlx5_accel_esp_destroy_xfrm(struct mlx5_accel_esp_xfrm *xfrm);
-void mlx5_accel_esp_modify_xfrm(struct mlx5_accel_esp_xfrm *xfrm,
+void mlx5_ipsec_esp_destroy_xfrm(struct mlx5_accel_esp_xfrm *xfrm);
+void mlx5_ipsec_esp_modify_xfrm(struct mlx5_accel_esp_xfrm *xfrm,
 				const struct mlx5_accel_esp_xfrm_attrs *attrs);
 #else
 static inline int mlx5e_ipsec_init(struct mlx5e_priv *priv)

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_offload.c +12 ~~-35~~

81 lines not on this page · raw message

Signature · 1 line

2.35.1

Leon Romanovsky

author

2022-04-10 08:29 UTC #15 lore raw

PATCH 9/17 net/mlx5: Simplify HW context interfaces by using SA entry

In reply to Leon Romanovsky

From: Leon Romanovsky <[email protected]>

SA context logic used multiple structures to store same data over and over. By simplifying the SA context interfaces, we can remove extra structs.

Reviewed-by Raed Salem <[email protected]>

Signed-off-by Leon Romanovsky <[email protected]>

---
 .../mellanox/mlx5/core/en_accel/ipsec.c       |  50 ++---
 .../mellanox/mlx5/core/en_accel/ipsec.h       |  27 ++-
 .../mlx5/core/en_accel/ipsec_offload.c        | 182 ++++--------------
 3 files changed, 61 insertions(+), 198 deletions(-)

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c +13 ~~-37~~

index 792f01264196..1b65b0c3d02b 100644
@@ -63,9 +63,9 @@ struct xfrm_state *mlx5e_ipsec_sadb_rx_lookup(struct mlx5e_ipsec *ipsec,
 	return ret;
 }
 
-static int  mlx5e_ipsec_sadb_rx_add(struct mlx5e_ipsec_sa_entry *sa_entry,
-				    unsigned int handle)
+static int mlx5e_ipsec_sadb_rx_add(struct mlx5e_ipsec_sa_entry *sa_entry)
 {
+	unsigned int handle = sa_entry->ipsec_obj_id;
 	struct mlx5e_ipsec *ipsec = sa_entry->ipsec;
 	struct mlx5e_ipsec_sa_entry *_sa_entry;
 	unsigned long flags;

@@ -277,16 +277,14 @@ static void _update_xfrm_state(struct work_struct *work)
 	struct mlx5e_ipsec_sa_entry *sa_entry = container_of(
 		modify_work, struct mlx5e_ipsec_sa_entry, modify_work);
 
-	mlx5_ipsec_esp_modify_xfrm(sa_entry->xfrm, &modify_work->attrs);
+	mlx5_ipsec_esp_modify_xfrm(sa_entry, &modify_work->attrs);
 }
 
 static int mlx5e_xfrm_add_state(struct xfrm_state *x)
 {
 	struct mlx5e_ipsec_sa_entry *sa_entry = NULL;
 	struct net_device *netdev = x->xso.real_dev;
-	struct mlx5_accel_esp_xfrm_attrs attrs;
 	struct mlx5e_priv *priv;
-	unsigned int sa_handle;
 	int err;
 
 	priv = netdev_priv(netdev);

@@ -309,33 +307,20 @@ static int mlx5e_xfrm_add_state(struct xfrm_state *x)
 	/* check esn */
 	mlx5e_ipsec_update_esn_state(sa_entry);
 
-	/* create xfrm */
-	mlx5e_ipsec_build_accel_xfrm_attrs(sa_entry, &attrs);
-	sa_entry->xfrm = mlx5_ipsec_esp_create_xfrm(priv->mdev, &attrs);
-	if (IS_ERR(sa_entry->xfrm)) {
-		err = PTR_ERR(sa_entry->xfrm);
-		goto err_sa_entry;
-	}
-
+	mlx5e_ipsec_build_accel_xfrm_attrs(sa_entry, &sa_entry->attrs);
 	/* create hw context */
-	sa_entry->hw_context =
-			mlx5_accel_esp_create_hw_context(priv->mdev,
-							 sa_entry->xfrm,
-							 &sa_handle);
-	if (IS_ERR(sa_entry->hw_context)) {
-		err = PTR_ERR(sa_entry->hw_context);
+	err = mlx5_ipsec_create_sa_ctx(sa_entry);
+	if (err)
 		goto err_xfrm;
-	}
 
-	sa_entry->ipsec_obj_id = sa_handle;
-	err = mlx5e_ipsec_fs_add_rule(priv, &sa_entry->xfrm->attrs,
+	err = mlx5e_ipsec_fs_add_rule(priv, &sa_entry->attrs,
 				      sa_entry->ipsec_obj_id,
 				      &sa_entry->ipsec_rule);
 	if (err)
 		goto err_hw_ctx;
 
 	if (x->xso.flags & XFRM_OFFLOAD_INBOUND) {
-		err = mlx5e_ipsec_sadb_rx_add(sa_entry, sa_handle);
+		err = mlx5e_ipsec_sadb_rx_add(sa_entry);
 		if (err)
 			goto err_add_rule;
 	} else {

@@ -348,15 +333,11 @@ static int mlx5e_xfrm_add_state(struct xfrm_state *x)
 	goto out;
 
 err_add_rule:
-	mlx5e_ipsec_fs_del_rule(priv, &sa_entry->xfrm->attrs,
-				&sa_entry->ipsec_rule);
+	mlx5e_ipsec_fs_del_rule(priv, &sa_entry->attrs, &sa_entry->ipsec_rule);
 err_hw_ctx:
-	mlx5_accel_esp_free_hw_context(priv->mdev, sa_entry->hw_context);
+	mlx5_ipsec_free_sa_ctx(sa_entry);
 err_xfrm:
-	mlx5_ipsec_esp_destroy_xfrm(sa_entry->xfrm);
-err_sa_entry:
 	kfree(sa_entry);
-
 out:
 	return err;
 }

@@ -374,14 +355,9 @@ static void mlx5e_xfrm_free_state(struct xfrm_state *x)
 	struct mlx5e_ipsec_sa_entry *sa_entry = to_ipsec_sa_entry(x);
 	struct mlx5e_priv *priv = netdev_priv(x->xso.dev);
 
-	if (sa_entry->hw_context) {
-		cancel_work_sync(&sa_entry->modify_work.work);
-		mlx5e_ipsec_fs_del_rule(priv, &sa_entry->xfrm->attrs,
-					&sa_entry->ipsec_rule);
-		mlx5_accel_esp_free_hw_context(sa_entry->xfrm->mdev, sa_entry->hw_context);
-		mlx5_ipsec_esp_destroy_xfrm(sa_entry->xfrm);
-	}
-
+	cancel_work_sync(&sa_entry->modify_work.work);
+	mlx5e_ipsec_fs_del_rule(priv, &sa_entry->attrs, &sa_entry->ipsec_rule);
+	mlx5_ipsec_free_sa_ctx(sa_entry);
 	kfree(sa_entry);
 }

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h +11 ~~-16~~

index 84c8d835a26b..284b907549d9 100644
@@ -102,11 +102,6 @@ struct mlx5_accel_esp_xfrm_attrs {
 	u8 is_ipv6;
 };
 
-struct mlx5_accel_esp_xfrm {
-	struct mlx5_core_dev  *mdev;
-	struct mlx5_accel_esp_xfrm_attrs attrs;
-};
-
 enum mlx5_accel_ipsec_cap {
 	MLX5_ACCEL_IPSEC_CAP_DEVICE		= 1 << 0,
 	MLX5_ACCEL_IPSEC_CAP_ESP		= 1 << 1,

@@ -162,11 +157,11 @@ struct mlx5e_ipsec_sa_entry {
 	unsigned int handle; /* Handle in SADB_RX */
 	struct xfrm_state *x;
 	struct mlx5e_ipsec *ipsec;
-	struct mlx5_accel_esp_xfrm *xfrm;
-	void *hw_context;
+	struct mlx5_accel_esp_xfrm_attrs attrs;
 	void (*set_iv_op)(struct sk_buff *skb, struct xfrm_state *x,
 			  struct xfrm_offload *xo);
 	u32 ipsec_obj_id;
+	u32 enc_key_id;
 	struct mlx5e_ipsec_rule ipsec_rule;
 	struct mlx5e_ipsec_modify_state_work modify_work;
 };

@@ -188,19 +183,19 @@ void mlx5e_ipsec_fs_del_rule(struct mlx5e_priv *priv,
 			     struct mlx5_accel_esp_xfrm_attrs *attrs,
 			     struct mlx5e_ipsec_rule *ipsec_rule);
 
-void *mlx5_accel_esp_create_hw_context(struct mlx5_core_dev *mdev,
-				       struct mlx5_accel_esp_xfrm *xfrm,
-				       u32 *sa_handle);
-void mlx5_accel_esp_free_hw_context(struct mlx5_core_dev *mdev, void *context);
+int mlx5_ipsec_create_sa_ctx(struct mlx5e_ipsec_sa_entry *sa_entry);
+void mlx5_ipsec_free_sa_ctx(struct mlx5e_ipsec_sa_entry *sa_entry);
 
 u32 mlx5_ipsec_device_caps(struct mlx5_core_dev *mdev);
 
-struct mlx5_accel_esp_xfrm *
-mlx5_ipsec_esp_create_xfrm(struct mlx5_core_dev *mdev,
-			   const struct mlx5_accel_esp_xfrm_attrs *attrs);
-void mlx5_ipsec_esp_destroy_xfrm(struct mlx5_accel_esp_xfrm *xfrm);
-void mlx5_ipsec_esp_modify_xfrm(struct mlx5_accel_esp_xfrm *xfrm,
+void mlx5_ipsec_esp_modify_xfrm(struct mlx5e_ipsec_sa_entry *sa_entry,
 				const struct mlx5_accel_esp_xfrm_attrs *attrs);
+
+static inline struct mlx5_core_dev *
+mlx5e_ipsec_sa2dev(struct mlx5e_ipsec_sa_entry *sa_entry)
+{
+	return sa_entry->ipsec->mdev;
+}
 #else
 static inline int mlx5e_ipsec_init(struct mlx5e_priv *priv)
 {

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_offload.c +37 ~~-145~~

283 lines not on this page · raw message

Signature · 1 line

2.35.1

Leon Romanovsky

author

2022-04-10 08:29 UTC #16 lore raw

PATCH 10/17 net/mlx5: Clean IPsec FS add/delete rules

In reply to Leon Romanovsky

From: Leon Romanovsky <[email protected]>

Reuse existing struct to pass parameters instead of open code them.

Reviewed-by Raed Salem <[email protected]>

Signed-off-by Leon Romanovsky <[email protected]>

---
 .../mellanox/mlx5/core/en_accel/ipsec.c       |  9 +--
 .../mellanox/mlx5/core/en_accel/ipsec.h       |  7 +--
 .../mellanox/mlx5/core/en_accel/ipsec_fs.c    | 55 ++++++++++---------
 3 files changed, 34 insertions(+), 37 deletions(-)

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c +3 -6

index 1b65b0c3d02b..815c77953a1a 100644
@@ -312,10 +312,7 @@ static int mlx5e_xfrm_add_state(struct xfrm_state *x)
 	err = mlx5_ipsec_create_sa_ctx(sa_entry);
 	if (err)
 		goto err_xfrm;
-
-	err = mlx5e_ipsec_fs_add_rule(priv, &sa_entry->attrs,
-				      sa_entry->ipsec_obj_id,
-				      &sa_entry->ipsec_rule);
+	err = mlx5e_ipsec_fs_add_rule(priv, sa_entry);
 	if (err)
 		goto err_hw_ctx;

@@ -333,7 +330,7 @@ static int mlx5e_xfrm_add_state(struct xfrm_state *x)
 	goto out;
 
 err_add_rule:
-	mlx5e_ipsec_fs_del_rule(priv, &sa_entry->attrs, &sa_entry->ipsec_rule);
+	mlx5e_ipsec_fs_del_rule(priv, sa_entry);
 err_hw_ctx:
 	mlx5_ipsec_free_sa_ctx(sa_entry);
 err_xfrm:

@@ -356,7 +353,7 @@ static void mlx5e_xfrm_free_state(struct xfrm_state *x)
 	struct mlx5e_priv *priv = netdev_priv(x->xso.dev);
 
 	cancel_work_sync(&sa_entry->modify_work.work);
-	mlx5e_ipsec_fs_del_rule(priv, &sa_entry->attrs, &sa_entry->ipsec_rule);
+	mlx5e_ipsec_fs_del_rule(priv, sa_entry);
 	mlx5_ipsec_free_sa_ctx(sa_entry);
 	kfree(sa_entry);
 }

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h +2 -5

index 284b907549d9..eb605a0bc5bb 100644
@@ -176,12 +176,9 @@ struct xfrm_state *mlx5e_ipsec_sadb_rx_lookup(struct mlx5e_ipsec *dev,
 void mlx5e_ipsec_fs_cleanup(struct mlx5e_ipsec *ipsec);
 int mlx5e_ipsec_fs_init(struct mlx5e_ipsec *ipsec);
 int mlx5e_ipsec_fs_add_rule(struct mlx5e_priv *priv,
-			    struct mlx5_accel_esp_xfrm_attrs *attrs,
-			    u32 ipsec_obj_id,
-			    struct mlx5e_ipsec_rule *ipsec_rule);
+			    struct mlx5e_ipsec_sa_entry *sa_entry);
 void mlx5e_ipsec_fs_del_rule(struct mlx5e_priv *priv,
-			     struct mlx5_accel_esp_xfrm_attrs *attrs,
-			     struct mlx5e_ipsec_rule *ipsec_rule);
+			     struct mlx5e_ipsec_sa_entry *sa_entry);
 
 int mlx5_ipsec_create_sa_ctx(struct mlx5e_ipsec_sa_entry *sa_entry);
 void mlx5_ipsec_free_sa_ctx(struct mlx5e_ipsec_sa_entry *sa_entry);

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c +29 ~~-26~~

112 lines not on this page · raw message

Signature · 1 line

2.35.1

Leon Romanovsky

author

2022-04-10 08:29 UTC #17 lore raw

PATCH 11/17 net/mlx5: Make sure that no dangling IPsec FS pointers exist

In reply to Leon Romanovsky

From: Leon Romanovsky <[email protected]>

The IPsec FS code was implemented with anti-pattern there failures in create functions left the system with dangling pointers that were cleaned in global routines.

The less error prone approach is to make sure that failed function cleans everything internally.

As part of this change, we remove the batch of one liners and rewrite get/put functions to remove ambiguity.

Reviewed-by Raed Salem <[email protected]>

Signed-off-by Leon Romanovsky <[email protected]>

---
 .../mellanox/mlx5/core/en_accel/ipsec_fs.c    | 229 ++++++------------
 1 file changed, 73 insertions(+), 156 deletions(-)

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c +73 ~~-156~~

index 3e8972bab085..2975788292ec 100644
@@ -60,7 +60,7 @@ static int rx_err_add_rule(struct mlx5e_priv *priv,
 	struct mlx5_modify_hdr *modify_hdr;
 	struct mlx5_flow_handle *fte;
 	struct mlx5_flow_spec *spec;
-	int err = 0;
+	int err;
 
 	spec = kvzalloc(sizeof(*spec), GFP_KERNEL);
 	if (!spec)

@@ -96,101 +96,27 @@ static int rx_err_add_rule(struct mlx5e_priv *priv,
 		goto out;
 	}
 
+	kvfree(spec);
 	rx_err->rule = fte;
 	rx_err->copy_modify_hdr = modify_hdr;
+	return 0;
 
 out:
-	if (err)
-		mlx5_modify_header_dealloc(mdev, modify_hdr);
+	mlx5_modify_header_dealloc(mdev, modify_hdr);
 out_spec:
 	kvfree(spec);
 	return err;
 }
 
-static void rx_err_del_rule(struct mlx5e_priv *priv,
-			    struct mlx5e_ipsec_rx_err *rx_err)
-{
-	if (rx_err->rule) {
-		mlx5_del_flow_rules(rx_err->rule);
-		rx_err->rule = NULL;
-	}
-
-	if (rx_err->copy_modify_hdr) {
-		mlx5_modify_header_dealloc(priv->mdev, rx_err->copy_modify_hdr);
-		rx_err->copy_modify_hdr = NULL;
-	}
-}
-
-static void rx_err_destroy_ft(struct mlx5e_priv *priv, struct mlx5e_ipsec_rx_err *rx_err)
-{
-	rx_err_del_rule(priv, rx_err);
-
-	if (rx_err->ft) {
-		mlx5_destroy_flow_table(rx_err->ft);
-		rx_err->ft = NULL;
-	}
-}
-
-static int rx_err_create_ft(struct mlx5e_priv *priv,
-			    struct mlx5e_accel_fs_esp_prot *fs_prot,
-			    struct mlx5e_ipsec_rx_err *rx_err)
-{
-	struct mlx5_flow_table_attr ft_attr = {};
-	struct mlx5_flow_table *ft;
-	int err;
-
-	ft_attr.max_fte = 1;
-	ft_attr.autogroup.max_num_groups = 1;
-	ft_attr.level = MLX5E_ACCEL_FS_ESP_FT_ERR_LEVEL;
-	ft_attr.prio = MLX5E_NIC_PRIO;
-	ft = mlx5_create_auto_grouped_flow_table(priv->fs.ns, &ft_attr);
-	if (IS_ERR(ft)) {
-		err = PTR_ERR(ft);
-		netdev_err(priv->netdev, "fail to create ipsec rx inline ft err=%d\n", err);
-		return err;
-	}
-
-	rx_err->ft = ft;
-	err = rx_err_add_rule(priv, fs_prot, rx_err);
-	if (err)
-		goto out_err;
-
-	return 0;
-
-out_err:
-	mlx5_destroy_flow_table(ft);
-	rx_err->ft = NULL;
-	return err;
-}
-
-static void rx_fs_destroy(struct mlx5e_accel_fs_esp_prot *fs_prot)
-{
-	if (fs_prot->miss_rule) {
-		mlx5_del_flow_rules(fs_prot->miss_rule);
-		fs_prot->miss_rule = NULL;
-	}
-
-	if (fs_prot->miss_group) {
-		mlx5_destroy_flow_group(fs_prot->miss_group);
-		fs_prot->miss_group = NULL;
-	}
-
-	if (fs_prot->ft) {
-		mlx5_destroy_flow_table(fs_prot->ft);
-		fs_prot->ft = NULL;
-	}
-}
-
 static int rx_fs_create(struct mlx5e_priv *priv,
 			struct mlx5e_accel_fs_esp_prot *fs_prot)
 {
 	int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
-	struct mlx5_flow_table_attr ft_attr = {};
+	struct mlx5_flow_table *ft = fs_prot->ft;
 	struct mlx5_flow_group *miss_group;
 	struct mlx5_flow_handle *miss_rule;
 	MLX5_DECLARE_FLOW_ACT(flow_act);
 	struct mlx5_flow_spec *spec;
-	struct mlx5_flow_table *ft;
 	u32 *flow_group_in;
 	int err = 0;

@@ -201,20 +127,6 @@ static int rx_fs_create(struct mlx5e_priv *priv,
 		goto out;
 	}
 
-	/* Create FT */
-	ft_attr.max_fte = NUM_IPSEC_FTE;
-	ft_attr.level = MLX5E_ACCEL_FS_ESP_FT_LEVEL;
-	ft_attr.prio = MLX5E_NIC_PRIO;
-	ft_attr.autogroup.num_reserved_entries = 1;
-	ft_attr.autogroup.max_num_groups = 1;
-	ft = mlx5_create_auto_grouped_flow_table(priv->fs.ns, &ft_attr);
-	if (IS_ERR(ft)) {
-		err = PTR_ERR(ft);
-		netdev_err(priv->netdev, "fail to create ipsec rx ft err=%d\n", err);
-		goto out;
-	}
-	fs_prot->ft = ft;
-
 	/* Create miss_group */
 	MLX5_SET(create_flow_group_in, flow_group_in, start_flow_index, ft->max_fte - 1);
 	MLX5_SET(create_flow_group_in, flow_group_in, end_flow_index, ft->max_fte - 1);

@@ -229,19 +141,19 @@ static int rx_fs_create(struct mlx5e_priv *priv,
 	/* Create miss rule */
 	miss_rule = mlx5_add_flow_rules(ft, spec, &flow_act, &fs_prot->default_dest, 1);
 	if (IS_ERR(miss_rule)) {
+		mlx5_destroy_flow_group(fs_prot->miss_group);
 		err = PTR_ERR(miss_rule);
 		netdev_err(priv->netdev, "fail to create ipsec rx miss_rule err=%d\n", err);
 		goto out;
 	}
 	fs_prot->miss_rule = miss_rule;
-
 out:
 	kvfree(flow_group_in);
 	kvfree(spec);
 	return err;
 }
 
-static int rx_destroy(struct mlx5e_priv *priv, enum accel_fs_esp_type type)
+static void rx_destroy(struct mlx5e_priv *priv, enum accel_fs_esp_type type)
 {
 	struct mlx5e_accel_fs_esp_prot *fs_prot;
 	struct mlx5e_accel_fs_esp *accel_esp;

@@ -251,17 +163,21 @@ static int rx_destroy(struct mlx5e_priv *priv, enum accel_fs_esp_type type)
 	/* The netdev unreg already happened, so all offloaded rule are already removed */
 	fs_prot = &accel_esp->fs_prot[type];
 
-	rx_fs_destroy(fs_prot);
-
-	rx_err_destroy_ft(priv, &fs_prot->rx_err);
+	mlx5_del_flow_rules(fs_prot->miss_rule);
+	mlx5_destroy_flow_group(fs_prot->miss_group);
+	mlx5_destroy_flow_table(fs_prot->ft);
 
-	return 0;
+	mlx5_del_flow_rules(fs_prot->rx_err.rule);
+	mlx5_modify_header_dealloc(priv->mdev, fs_prot->rx_err.copy_modify_hdr);
+	mlx5_destroy_flow_table(fs_prot->rx_err.ft);
 }
 
 static int rx_create(struct mlx5e_priv *priv, enum accel_fs_esp_type type)
 {
+	struct mlx5_flow_table_attr ft_attr = {};
 	struct mlx5e_accel_fs_esp_prot *fs_prot;
 	struct mlx5e_accel_fs_esp *accel_esp;
+	struct mlx5_flow_table *ft;
 	int err;
 
 	accel_esp = priv->ipsec->rx_fs;

@@ -270,14 +186,45 @@ static int rx_create(struct mlx5e_priv *priv, enum accel_fs_esp_type type)
 	fs_prot->default_dest =
 		mlx5_ttc_get_default_dest(priv->fs.ttc, fs_esp2tt(type));
 
-	err = rx_err_create_ft(priv, fs_prot, &fs_prot->rx_err);
+	ft_attr.max_fte = 1;
+	ft_attr.autogroup.max_num_groups = 1;
+	ft_attr.level = MLX5E_ACCEL_FS_ESP_FT_ERR_LEVEL;
+	ft_attr.prio = MLX5E_NIC_PRIO;
+	ft = mlx5_create_auto_grouped_flow_table(priv->fs.ns, &ft_attr);
+	if (IS_ERR(ft))
+		return PTR_ERR(ft);
+
+	fs_prot->rx_err.ft = ft;
+	err = rx_err_add_rule(priv, fs_prot, &fs_prot->rx_err);
 	if (err)
-		return err;
+		goto err_add;
+
+	/* Create FT */
+	ft_attr.max_fte = NUM_IPSEC_FTE;
+	ft_attr.level = MLX5E_ACCEL_FS_ESP_FT_LEVEL;
+	ft_attr.prio = MLX5E_NIC_PRIO;
+	ft_attr.autogroup.num_reserved_entries = 1;
+	ft_attr.autogroup.max_num_groups = 1;
+	ft = mlx5_create_auto_grouped_flow_table(priv->fs.ns, &ft_attr);
+	if (IS_ERR(ft)) {
+		err = PTR_ERR(ft);
+		goto err_fs_ft;
+	}
+	fs_prot->ft = ft;
 
 	err = rx_fs_create(priv, fs_prot);
 	if (err)
-		rx_destroy(priv, type);
+		goto err_fs;
+
+	return 0;
 
+err_fs:
+	mlx5_destroy_flow_table(fs_prot->ft);
+err_fs_ft:
+	mlx5_del_flow_rules(fs_prot->rx_err.rule);
+	mlx5_modify_header_dealloc(priv->mdev, fs_prot->rx_err.copy_modify_hdr);
+err_add:
+	mlx5_destroy_flow_table(fs_prot->rx_err.ft);
 	return err;
 }

@@ -291,21 +238,21 @@ static int rx_ft_get(struct mlx5e_priv *priv, enum accel_fs_esp_type type)
 	accel_esp = priv->ipsec->rx_fs;
 	fs_prot = &accel_esp->fs_prot[type];
 	mutex_lock(&fs_prot->prot_mutex);
-	if (fs_prot->refcnt++)
-		goto out;
+	if (fs_prot->refcnt)
+		goto skip;
 
 	/* create FT */
 	err = rx_create(priv, type);
-	if (err) {
-		fs_prot->refcnt--;
+	if (err)
 		goto out;
-	}
 
 	/* connect */
 	dest.type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
 	dest.ft = fs_prot->ft;
 	mlx5_ttc_fwd_dest(priv->fs.ttc, fs_esp2tt(type), &dest);
 
+skip:
+	fs_prot->refcnt++;
 out:
 	mutex_unlock(&fs_prot->prot_mutex);
 	return err;

@@ -319,7 +266,8 @@ static void rx_ft_put(struct mlx5e_priv *priv, enum accel_fs_esp_type type)
 	accel_esp = priv->ipsec->rx_fs;
 	fs_prot = &accel_esp->fs_prot[type];
 	mutex_lock(&fs_prot->prot_mutex);
-	if (--fs_prot->refcnt)
+	fs_prot->refcnt--;
+	if (fs_prot->refcnt)
 		goto out;
 
 	/* disconnect */

@@ -352,32 +300,20 @@ static int tx_create(struct mlx5e_priv *priv)
 	return 0;
 }
 
-static void tx_destroy(struct mlx5e_priv *priv)
-{
-	struct mlx5e_ipsec *ipsec = priv->ipsec;
-
-	if (IS_ERR_OR_NULL(ipsec->tx_fs->ft))
-		return;
-
-	mlx5_destroy_flow_table(ipsec->tx_fs->ft);
-	ipsec->tx_fs->ft = NULL;
-}
-
 static int tx_ft_get(struct mlx5e_priv *priv)
 {
 	struct mlx5e_ipsec_tx *tx_fs = priv->ipsec->tx_fs;
 	int err = 0;
 
 	mutex_lock(&tx_fs->mutex);
-	if (tx_fs->refcnt++)
-		goto out;
+	if (tx_fs->refcnt)
+		goto skip;
 
 	err = tx_create(priv);
-	if (err) {
-		tx_fs->refcnt--;
+	if (err)
 		goto out;
-	}
-
+skip:
+	tx_fs->refcnt++;
 out:
 	mutex_unlock(&tx_fs->mutex);
 	return err;

@@ -388,11 +324,11 @@ static void tx_ft_put(struct mlx5e_priv *priv)
 	struct mlx5e_ipsec_tx *tx_fs = priv->ipsec->tx_fs;
 
 	mutex_lock(&tx_fs->mutex);
-	if (--tx_fs->refcnt)
+	tx_fs->refcnt--;
+	if (tx_fs->refcnt)
 		goto out;
 
-	tx_destroy(priv);
-
+	mlx5_destroy_flow_table(tx_fs->ft);
 out:
 	mutex_unlock(&tx_fs->mutex);
 }

@@ -579,32 +515,6 @@ static int tx_add_rule(struct mlx5e_priv *priv,
 	return err;
 }
 
-static void rx_del_rule(struct mlx5e_priv *priv,
-			struct mlx5e_ipsec_sa_entry *sa_entry)
-{
-	struct mlx5e_ipsec_rule *ipsec_rule = &sa_entry->ipsec_rule;
-
-	mlx5_del_flow_rules(ipsec_rule->rule);
-	ipsec_rule->rule = NULL;
-
-	mlx5_modify_header_dealloc(priv->mdev, ipsec_rule->set_modify_hdr);
-	ipsec_rule->set_modify_hdr = NULL;
-
-	rx_ft_put(priv,
-		  sa_entry->attrs.is_ipv6 ? ACCEL_FS_ESP6 : ACCEL_FS_ESP4);
-}
-
-static void tx_del_rule(struct mlx5e_priv *priv,
-			struct mlx5e_ipsec_sa_entry *sa_entry)
-{
-	struct mlx5e_ipsec_rule *ipsec_rule = &sa_entry->ipsec_rule;
-
-	mlx5_del_flow_rules(ipsec_rule->rule);
-	ipsec_rule->rule = NULL;
-
-	tx_ft_put(priv);
-}
-
 int mlx5e_ipsec_fs_add_rule(struct mlx5e_priv *priv,
 			    struct mlx5e_ipsec_sa_entry *sa_entry)
 {

@@ -617,12 +527,19 @@ int mlx5e_ipsec_fs_add_rule(struct mlx5e_priv *priv,
 void mlx5e_ipsec_fs_del_rule(struct mlx5e_priv *priv,
 			     struct mlx5e_ipsec_sa_entry *sa_entry)
 {
+	struct mlx5e_ipsec_rule *ipsec_rule = &sa_entry->ipsec_rule;
+	struct mlx5_core_dev *mdev = mlx5e_ipsec_sa2dev(sa_entry);
+
+	mlx5_del_flow_rules(ipsec_rule->rule);
+
 	if (sa_entry->attrs.action == MLX5_ACCEL_ESP_ACTION_ENCRYPT) {
-		tx_del_rule(priv, sa_entry);
+		tx_ft_put(priv);
 		return;
 	}
 
-	rx_del_rule(priv, sa_entry);
+	mlx5_modify_header_dealloc(mdev, ipsec_rule->set_modify_hdr);
+	rx_ft_put(priv,
+		  sa_entry->attrs.is_ipv6 ? ACCEL_FS_ESP6 : ACCEL_FS_ESP4);
 }
 
 void mlx5e_ipsec_fs_cleanup(struct mlx5e_ipsec *ipsec)

Signature · 1 line

2.35.1

Leon Romanovsky

author

2022-04-10 08:29 UTC #18 lore raw

PATCH 12/17 net/mlx5: Don't advertise IPsec netdev support for non-IPsec device

In reply to Leon Romanovsky

From: Leon Romanovsky <[email protected]>

Device that lacks proper IPsec capabilities won't pass mlx5e_ipsec_init() later, so no need to advertise HW netdev offload support for something that isn't going to work anyway.

Fixes 8ad893e516a7 ("net/mlx5e: Remove dependency in IPsec initialization flows")

Reviewed-by Raed Salem <[email protected]>

Signed-off-by Leon Romanovsky <[email protected]>

---
 drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c | 3 +++
 1 file changed, 3 insertions(+)

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c +3 -0

index 815c77953a1a..6f93d749b21a 100644
@@ -453,6 +453,9 @@ void mlx5e_ipsec_build_netdev(struct mlx5e_priv *priv)
 	struct mlx5_core_dev *mdev = priv->mdev;
 	struct net_device *netdev = priv->netdev;
 
+	if (!mlx5_ipsec_device_caps(mdev))
+		return;
+
 	if (!(mlx5_ipsec_device_caps(mdev) & MLX5_ACCEL_IPSEC_CAP_ESP) ||
 	    !MLX5_CAP_ETH(mdev, swp)) {
 		mlx5_core_dbg(mdev, "mlx5e: ESP and SWP offload not supported\n");

Signature · 1 line

2.35.1

Leon Romanovsky

author

2022-04-10 08:29 UTC #19 lore raw

PATCH 13/17 net/mlx5: Simplify IPsec capabilities logic

In reply to Leon Romanovsky

From: Leon Romanovsky <[email protected]>

Reduce number of hard-coded IPsec capabilities by making sure that mlx5_ipsec_device_caps() sets only supported bits.

As part of this change, remove _accel_ notations from the names and prepare the code to IPsec full offload mode.

Reviewed-by Raed Salem <[email protected]>

Signed-off-by Leon Romanovsky <[email protected]>

---
 .../mellanox/mlx5/core/en_accel/ipsec.c       | 13 +++--------
 .../mellanox/mlx5/core/en_accel/ipsec.h       |  9 +++-----
 .../mlx5/core/en_accel/ipsec_offload.c        | 23 ++++++++++---------
 3 files changed, 18 insertions(+), 27 deletions(-)

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c +3 ~~-10~~

index 6f93d749b21a..e138a4d1a9c0 100644
@@ -215,7 +215,7 @@ static inline int mlx5e_xfrm_validate_state(struct xfrm_state *x)
 		return -EINVAL;
 	}
 	if (x->props.flags & XFRM_STATE_ESN &&
-	    !(mlx5_ipsec_device_caps(priv->mdev) & MLX5_ACCEL_IPSEC_CAP_ESN)) {
+	    !(mlx5_ipsec_device_caps(priv->mdev) & MLX5_IPSEC_CAP_ESN)) {
 		netdev_info(netdev, "Cannot offload ESN xfrm states\n");
 		return -EINVAL;
 	}

@@ -263,7 +263,7 @@ static inline int mlx5e_xfrm_validate_state(struct xfrm_state *x)
 		return -EINVAL;
 	}
 	if (x->props.family == AF_INET6 &&
-	    !(mlx5_ipsec_device_caps(priv->mdev) & MLX5_ACCEL_IPSEC_CAP_IPV6)) {
+	    !mlx5_ipsec_device_caps(priv->mdev)) {
 		netdev_info(netdev, "IPv6 xfrm state offload is not supported by this device\n");
 		return -EINVAL;
 	}

@@ -456,12 +456,6 @@ void mlx5e_ipsec_build_netdev(struct mlx5e_priv *priv)
 	if (!mlx5_ipsec_device_caps(mdev))
 		return;
 
-	if (!(mlx5_ipsec_device_caps(mdev) & MLX5_ACCEL_IPSEC_CAP_ESP) ||
-	    !MLX5_CAP_ETH(mdev, swp)) {
-		mlx5_core_dbg(mdev, "mlx5e: ESP and SWP offload not supported\n");
-		return;
-	}
-
 	mlx5_core_info(mdev, "mlx5e: IPSec ESP acceleration enabled\n");
 	netdev->xfrmdev_ops = &mlx5e_ipsec_xfrmdev_ops;
 	netdev->features |= NETIF_F_HW_ESP;

@@ -475,8 +469,7 @@ void mlx5e_ipsec_build_netdev(struct mlx5e_priv *priv)
 	netdev->features |= NETIF_F_HW_ESP_TX_CSUM;
 	netdev->hw_enc_features |= NETIF_F_HW_ESP_TX_CSUM;
 
-	if (!(mlx5_ipsec_device_caps(mdev) & MLX5_ACCEL_IPSEC_CAP_LSO) ||
-	    !MLX5_CAP_ETH(mdev, swp_lso)) {
+	if (!MLX5_CAP_ETH(mdev, swp_lso)) {
 		mlx5_core_dbg(mdev, "mlx5e: ESP LSO not supported\n");
 		return;
 	}

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h +3 -6

index eb605a0bc5bb..998a812fbd15 100644
@@ -102,12 +102,9 @@ struct mlx5_accel_esp_xfrm_attrs {
 	u8 is_ipv6;
 };
 
-enum mlx5_accel_ipsec_cap {
-	MLX5_ACCEL_IPSEC_CAP_DEVICE		= 1 << 0,
-	MLX5_ACCEL_IPSEC_CAP_ESP		= 1 << 1,
-	MLX5_ACCEL_IPSEC_CAP_IPV6		= 1 << 2,
-	MLX5_ACCEL_IPSEC_CAP_LSO		= 1 << 3,
-	MLX5_ACCEL_IPSEC_CAP_ESN		= 1 << 4,
+enum mlx5_ipsec_cap {
+	MLX5_IPSEC_CAP_CRYPTO		= 1 << 0,
+	MLX5_IPSEC_CAP_ESN		= 1 << 1,
 };
 
 struct mlx5e_priv;

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_offload.c +12 ~~-11~~

index d3585c8c876b..1dc8588635e9 100644
@@ -7,7 +7,7 @@
 
 u32 mlx5_ipsec_device_caps(struct mlx5_core_dev *mdev)
 {
-	u32 caps;
+	u32 caps = 0;
 
 	if (!MLX5_CAP_GEN(mdev, ipsec_offload))
 		return 0;

@@ -19,23 +19,24 @@ u32 mlx5_ipsec_device_caps(struct mlx5_core_dev *mdev)
 	    MLX5_HCA_CAP_GENERAL_OBJECT_TYPES_IPSEC))
 		return 0;
 
-	if (!MLX5_CAP_IPSEC(mdev, ipsec_crypto_offload) ||
-	    !MLX5_CAP_ETH(mdev, insert_trailer))
-		return 0;
-
 	if (!MLX5_CAP_FLOWTABLE_NIC_TX(mdev, ipsec_encrypt) ||
 	    !MLX5_CAP_FLOWTABLE_NIC_RX(mdev, ipsec_decrypt))
 		return 0;
 
-	caps = MLX5_ACCEL_IPSEC_CAP_DEVICE | MLX5_ACCEL_IPSEC_CAP_IPV6 |
-	       MLX5_ACCEL_IPSEC_CAP_LSO;
+	if (!MLX5_CAP_IPSEC(mdev, ipsec_crypto_esp_aes_gcm_128_encrypt) ||
+	    !MLX5_CAP_IPSEC(mdev, ipsec_crypto_esp_aes_gcm_128_decrypt) ||
+	    !MLX5_CAP_ETH(mdev, swp))
+		return 0;
 
-	if (MLX5_CAP_IPSEC(mdev, ipsec_crypto_esp_aes_gcm_128_encrypt) &&
-	    MLX5_CAP_IPSEC(mdev, ipsec_crypto_esp_aes_gcm_128_decrypt))
-		caps |= MLX5_ACCEL_IPSEC_CAP_ESP;
+	if (MLX5_CAP_IPSEC(mdev, ipsec_crypto_offload) &&
+	    MLX5_CAP_ETH(mdev, insert_trailer))
+		caps |= MLX5_IPSEC_CAP_CRYPTO;
+
+	if (!caps)
+		return 0;
 
 	if (MLX5_CAP_IPSEC(mdev, ipsec_esn))
-		caps |= MLX5_ACCEL_IPSEC_CAP_ESN;
+		caps |= MLX5_IPSEC_CAP_ESN;
 
 	/* We can accommodate up to 2^24 different IPsec objects
 	 * because we use up to 24 bit in flow table metadata

Signature · 1 line

2.35.1

Leon Romanovsky

author

2022-04-10 08:29 UTC #20 lore raw

PATCH 14/17 net/mlx5: Remove not-supported ICV length

In reply to Leon Romanovsky

From: Leon Romanovsky <[email protected]>

mlx5 doesn't allow to configure any AEAD ICV length other than 128, so remove the logic that configures other unsupported values.

Signed-off-by Leon Romanovsky <[email protected]>

---
 .../mellanox/mlx5/core/en_accel/ipsec_offload.c | 17 +----------------
 include/linux/mlx5/mlx5_ifc.h                   |  2 --
 2 files changed, 1 insertion(+), 18 deletions(-)

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_offload.c +1 ~~-16~~

index 1dc8588635e9..4f3ffe5a09fa 100644
@@ -63,22 +63,7 @@ static int mlx5_create_ipsec_obj(struct mlx5e_ipsec_sa_entry *sa_entry)
 	salt_p = MLX5_ADDR_OF(ipsec_obj, obj, salt);
 	memcpy(salt_p, &aes_gcm->salt, sizeof(aes_gcm->salt));
 
-	switch (aes_gcm->icv_len) {
-	case 64:
-		MLX5_SET(ipsec_obj, obj, icv_length,
-			 MLX5_IPSEC_OBJECT_ICV_LEN_8B);
-		break;
-	case 96:
-		MLX5_SET(ipsec_obj, obj, icv_length,
-			 MLX5_IPSEC_OBJECT_ICV_LEN_12B);
-		break;
-	case 128:
-		MLX5_SET(ipsec_obj, obj, icv_length,
-			 MLX5_IPSEC_OBJECT_ICV_LEN_16B);
-		break;
-	default:
-		return -EINVAL;
-	}
+	MLX5_SET(ipsec_obj, obj, icv_length, MLX5_IPSEC_OBJECT_ICV_LEN_16B);
 	salt_iv_p = MLX5_ADDR_OF(ipsec_obj, obj, implicit_iv);
 	memcpy(salt_iv_p, &aes_gcm->seq_iv, sizeof(aes_gcm->seq_iv));
 	/* esn */

include/linux/mlx5/mlx5_ifc.h +0 -2

index 7d2d0ba82144..1fa4ade54c76 100644
@@ -11383,8 +11383,6 @@ enum {
 
 enum {
 	MLX5_IPSEC_OBJECT_ICV_LEN_16B,
-	MLX5_IPSEC_OBJECT_ICV_LEN_12B,
-	MLX5_IPSEC_OBJECT_ICV_LEN_8B,
 };
 
 struct mlx5_ifc_ipsec_obj_bits {

Signature · 1 line

2.35.1

Leon Romanovsky

author

2022-04-10 08:30 UTC #21 lore raw

PATCH 15/17 net/mlx5: Cleanup XFRM attributes struct

In reply to Leon Romanovsky

From: Leon Romanovsky <[email protected]>

Remove everything that is not used or from mlx5_accel_esp_xfrm_attrs, together with change type of spi to store proper type from the beginning.

Signed-off-by Leon Romanovsky <[email protected]>

---
 .../mellanox/mlx5/core/en_accel/ipsec.c       | 10 ++-------
 .../mellanox/mlx5/core/en_accel/ipsec.h       | 21 ++-----------------
 .../mellanox/mlx5/core/en_accel/ipsec_fs.c    |  4 ++--
 .../mlx5/core/en_accel/ipsec_offload.c        |  4 ++--
 4 files changed, 8 insertions(+), 31 deletions(-)

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c +2 -8

index e138a4d1a9c0..4476397d3d22 100644
@@ -137,7 +137,7 @@ mlx5e_ipsec_build_accel_xfrm_attrs(struct mlx5e_ipsec_sa_entry *sa_entry,
 				   struct mlx5_accel_esp_xfrm_attrs *attrs)
 {
 	struct xfrm_state *x = sa_entry->x;
-	struct aes_gcm_keymat *aes_gcm = &attrs->keymat.aes_gcm;
+	struct aes_gcm_keymat *aes_gcm = &attrs->aes_gcm;
 	struct aead_geniv_ctx *geniv_ctx;
 	struct crypto_aead *aead;
 	unsigned int crypto_data_len, key_len;

@@ -171,12 +171,6 @@ mlx5e_ipsec_build_accel_xfrm_attrs(struct mlx5e_ipsec_sa_entry *sa_entry,
 			attrs->flags |= MLX5_ACCEL_ESP_FLAGS_ESN_STATE_OVERLAP;
 	}
 
-	/* rx handle */
-	attrs->sa_handle = sa_entry->handle;
-
-	/* algo type */
-	attrs->keymat_type = MLX5_ACCEL_ESP_KEYMAT_AES_GCM;
-
 	/* action */
 	attrs->action = (!(x->xso.flags & XFRM_OFFLOAD_INBOUND)) ?
 			MLX5_ACCEL_ESP_ACTION_ENCRYPT :

@@ -187,7 +181,7 @@ mlx5e_ipsec_build_accel_xfrm_attrs(struct mlx5e_ipsec_sa_entry *sa_entry,
 			MLX5_ACCEL_ESP_FLAGS_TUNNEL;
 
 	/* spi */
-	attrs->spi = x->id.spi;
+	attrs->spi = be32_to_cpu(x->id.spi);
 
 	/* source , destination ips */
 	memcpy(&attrs->saddr, x->props.saddr.a6, sizeof(attrs->saddr));

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h +2 ~~-19~~

index 998a812fbd15..1e31acc04a44 100644
@@ -55,11 +55,6 @@ enum mlx5_accel_esp_action {
 	MLX5_ACCEL_ESP_ACTION_ENCRYPT,
 };
 
-enum mlx5_accel_esp_keymats {
-	MLX5_ACCEL_ESP_KEYMAT_AES_NONE,
-	MLX5_ACCEL_ESP_KEYMAT_AES_GCM,
-};
-
 struct aes_gcm_keymat {
 	u64   seq_iv;

@@ -73,21 +68,9 @@ struct aes_gcm_keymat {
 struct mlx5_accel_esp_xfrm_attrs {
 	enum mlx5_accel_esp_action action;
 	u32   esn;
-	__be32 spi;
-	u32   seq;
-	u32   tfc_pad;
+	u32   spi;
 	u32   flags;
-	u32   sa_handle;
-	union {
-		struct {
-			u32 size;
-
-		} bmp;
-	} replay;
-	enum mlx5_accel_esp_keymats keymat_type;
-	union {
-		struct aes_gcm_keymat aes_gcm;
-	} keymat;
+	struct aes_gcm_keymat aes_gcm;
 
 	union {
 		__be32 a4;

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c +2 -2

index 2975788292ec..b6a389ca8ce3 100644
@@ -356,8 +356,8 @@ static void setup_fte_common(struct mlx5_accel_esp_xfrm_attrs *attrs,
 
 	/* SPI number */
 	MLX5_SET_TO_ONES(fte_match_param, spec->match_criteria, misc_parameters.outer_esp_spi);
-	MLX5_SET(fte_match_param, spec->match_value, misc_parameters.outer_esp_spi,
-		 be32_to_cpu(attrs->spi));
+	MLX5_SET(fte_match_param, spec->match_value,
+		 misc_parameters.outer_esp_spi, attrs->spi);
 
 	if (ip_version == 4) {
 		memcpy(MLX5_ADDR_OF(fte_match_param, spec->match_value,

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_offload.c +2 -2

index 4f3ffe5a09fa..2f32d53761b1 100644
@@ -51,7 +51,7 @@ static int mlx5_create_ipsec_obj(struct mlx5e_ipsec_sa_entry *sa_entry)
 {
 	struct mlx5_accel_esp_xfrm_attrs *attrs = &sa_entry->attrs;
 	struct mlx5_core_dev *mdev = mlx5e_ipsec_sa2dev(sa_entry);
-	struct aes_gcm_keymat *aes_gcm = &attrs->keymat.aes_gcm;
+	struct aes_gcm_keymat *aes_gcm = &attrs->aes_gcm;
 	u32 out[MLX5_ST_SZ_DW(general_obj_out_cmd_hdr)];
 	u32 in[MLX5_ST_SZ_DW(create_ipsec_obj_in)] = {};
 	void *obj, *salt_p, *salt_iv_p;

@@ -107,7 +107,7 @@ static void mlx5_destroy_ipsec_obj(struct mlx5e_ipsec_sa_entry *sa_entry)
 
 int mlx5_ipsec_create_sa_ctx(struct mlx5e_ipsec_sa_entry *sa_entry)
 {
-	struct aes_gcm_keymat *aes_gcm = &sa_entry->attrs.keymat.aes_gcm;
+	struct aes_gcm_keymat *aes_gcm = &sa_entry->attrs.aes_gcm;
 	struct mlx5_core_dev *mdev = mlx5e_ipsec_sa2dev(sa_entry);
 	int err;

Signature · 1 line

2.35.1

Leon Romanovsky

author

2022-04-10 08:30 UTC #22 lore raw

PATCH 16/17 net/mlx5: Allow future addition of IPsec object modifiers

In reply to Leon Romanovsky

From: Leon Romanovsky <[email protected]>

Currently, all released FW versions support only two IPsec object modifiers, and modify_field_select get and set same value with proper bits.

However, it is not future compatible, as new FW can have more modifiers and "default" will cause to overwrite not-changed fields.

Fix it by setting explicitly fields that need to be overwritten.

Signed-off-by Huy Nguyen <[email protected]>

Reviewed-by Raed Salem <[email protected]>

Signed-off-by Leon Romanovsky <[email protected]>

---
 .../net/ethernet/mellanox/mlx5/core/en_accel/ipsec_offload.c   | 3 +++
 1 file changed, 3 insertions(+)

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_offload.c +3 -0

index 2f32d53761b1..e079f86fae33 100644
@@ -180,6 +180,9 @@ static int mlx5_modify_ipsec_obj(struct mlx5e_ipsec_sa_entry *sa_entry,
 		return -EOPNOTSUPP;
 
 	obj = MLX5_ADDR_OF(modify_ipsec_obj_in, in, ipsec_object);
+	MLX5_SET64(ipsec_obj, obj, modify_field_select,
+		   MLX5_MODIFY_IPSEC_BITMASK_ESN_OVERLAP |
+			   MLX5_MODIFY_IPSEC_BITMASK_ESN_MSB);
 	MLX5_SET(ipsec_obj, obj, esn_msb, attrs->esn);
 	if (attrs->flags & MLX5_ACCEL_ESP_FLAGS_ESN_STATE_OVERLAP)
 		MLX5_SET(ipsec_obj, obj, esn_overlap, 1);

Signature · 1 line

2.35.1

Leon Romanovsky

author

2022-04-10 08:30 UTC #23 lore raw

PATCH 17/17 net/mlx5: Don't perform lookup after already known sec_path

In reply to Leon Romanovsky

From: Leon Romanovsky <[email protected]>

There is no need to perform extra lookup in order to get already known sec_path that was set a couple of lines above. Simply reuse it.

Reviewed-by Raed Salem <[email protected]>

Signed-off-by Leon Romanovsky <[email protected]>

---
 drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c | 1 -
 1 file changed, 1 deletion(-)

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c +0 -1

index d30922e1b60f..6859f1c1a831 100644
@@ -332,7 +332,6 @@ void mlx5e_ipsec_offload_handle_rx_skb(struct net_device *netdev,
 		return;
 	}
 
-	sp = skb_sec_path(skb);
 	sp->xvec[sp->len++] = xs;
 	sp->olen++;

Signature · 1 line

2.35.1