[pull request][net-next 00/15] mlx5 updates 2022-05-02

[pull request][net-next 00/15] mlx5 updates 2022-05-02

[Saeed Mahameed]

This series provides some misc updates to mlx5 driver and a refactoring 
series for flow steering core logic.
For more information please see tag log below.

Please pull and let me know if there is any problem.

Thanks,
Saeed.


The following changes since commit 0530a683fc858aa641d88ad83315ea53c27bce10:

  Merge branch 'vsock-virtio-add-support-for-device-suspend-resume' (2022-05-02 16:04:37 -0700)

are available in the Git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux.git tags/mlx5-updates-2022-05-02

for you to fetch changes up to 3a09fae035c879c7ae8e5e154d7b03ddf0de5f20:

  net/mlx5: fs, an FTE should have no dests when deleted (2022-05-02 21:21:16 -0700)

----------------------------------------------------------------
mlx5-updates-2022-05-02

1) Trivial Misc updates to mlx5 driver

2) From Mark Bloch: Flow steering, general steering refactoring/cleaning

An issue with flow steering deletion flow (when creating a rule without
dests) turned out to be easy to fix but during the fix some issue
with the flow steering creation/deletion flows have been found.

The following patch series tries to fix long standing issues with flow
steering code and hopefully preventing silly future bugs.

  A) Fix an issue where a proper dest type wasn't assigned.
  B) Refactor and fix dests enums values, refactor deletion
     function and do proper bookkeeping of dests.
  C) Change mlx5_del_flow_rules() to delete rules when there are no
     no more rules attached associated with an FTE.
  D) Don't call hard coded deletion function but use the node's
     defined one.
  E) Add a WARN_ON() to catch future bugs when an FTE with dests
     is deleted.

----------------------------------------------------------------
Gal Pressman (1):
      net/mlx5e: Remove unused mlx5e_dcbnl_build_rep_netdev function

Haowen Bai (1):
      net/mlx5: Remove useless kfree

Mark Bloch (9):
      net/mlx5e: TC, set proper dest type
      net/mlx5: fs, split software and IFC flow destination definitions
      net/mlx5: fs, refactor software deletion rule
      net/mlx5: fs, jump to exit point and don't fall through
      net/mlx5: fs, add unused destination type
      net/mlx5: fs, do proper bookkeeping for forward destinations
      net/mlx5: fs, delete the FTE when there are no rules attached to it
      net/mlx5: fs, call the deletion function of the node
      net/mlx5: fs, an FTE should have no dests when deleted

Maxim Mikityanskiy (1):
      net/mlx5e: Drop error CQE handling from the XSK RX handler

Shay Drory (2):
      net/mlx5: Delete redundant default assignment of runtime devlink params
      net/mlx5: Print initializing field in case of timeout

Ziyang Xuan (1):
      net/mlx5: use kvfree() for kvzalloc() in mlx5_ct_fs_smfs_matcher_create

 drivers/net/ethernet/mellanox/mlx5/core/devlink.c  | 20 ----------
 .../mellanox/mlx5/core/diag/fs_tracepoint.c        |  3 ++
 drivers/net/ethernet/mellanox/mlx5/core/en.h       |  4 +-
 drivers/net/ethernet/mellanox/mlx5/core/en/dcbnl.h |  2 -
 .../ethernet/mellanox/mlx5/core/en/tc/ct_fs_smfs.c |  2 +-
 .../net/ethernet/mellanox/mlx5/core/en/tc/sample.c |  1 +
 drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c |  1 -
 .../net/ethernet/mellanox/mlx5/core/en/xsk/rx.c    |  6 ---
 .../net/ethernet/mellanox/mlx5/core/en/xsk/rx.h    |  1 -
 drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c |  9 -----
 drivers/net/ethernet/mellanox/mlx5/core/en_rx.c    | 16 ++++----
 drivers/net/ethernet/mellanox/mlx5/core/eswitch.c  |  3 ++
 drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c   | 18 +++++++--
 drivers/net/ethernet/mellanox/mlx5/core/fs_core.c  | 46 ++++++++++++++++------
 drivers/net/ethernet/mellanox/mlx5/core/fs_core.h  |  1 +
 drivers/net/ethernet/mellanox/mlx5/core/main.c     | 17 ++++----
 .../ethernet/mellanox/mlx5/core/steering/dr_cmd.c  | 21 +++++++---
 include/linux/mlx5/fs.h                            | 12 ++++++
 include/linux/mlx5/mlx5_ifc.h                      | 16 +++-----
 19 files changed, 109 insertions(+), 90 deletions(-)

[net-next 01/15] net/mlx5: use kvfree() for kvzalloc() in mlx5_ct_fs_smfs_matcher_create

[Saeed Mahameed]

From: Ziyang Xuan <william.xuanziyang@huawei.com>

The memory of spec is allocated with kvzalloc(), the corresponding
release function should not be kfree(), use kvfree() instead.

Generated by: scripts/coccinelle/api/kfree_mismatch.cocci

Signed-off-by: Ziyang Xuan <william.xuanziyang@huawei.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en/tc/ct_fs_smfs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/ct_fs_smfs.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/ct_fs_smfs.c
index 59988e24b704..b979826f3f6c 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/ct_fs_smfs.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/ct_fs_smfs.c
@@ -100,7 +100,7 @@ mlx5_ct_fs_smfs_matcher_create(struct mlx5_ct_fs *fs, struct mlx5dr_table *tbl,
 	spec->match_criteria_enable = MLX5_MATCH_MISC_PARAMETERS_2 | MLX5_MATCH_OUTER_HEADERS;
 
 	dr_matcher = mlx5_smfs_matcher_create(tbl, priority, spec);
-	kfree(spec);
+	kvfree(spec);
 	if (!dr_matcher)
 		return ERR_PTR(-EINVAL);
 
-- 
2.35.1

[net-next 02/15] net/mlx5: Remove useless kfree

[Saeed Mahameed]

From: Haowen Bai <baihaowen@meizu.com>

After alloc fail, we do not need to kfree.

Signed-off-by: Haowen Bai <baihaowen@meizu.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c
index e49f51124c74..89aa0208b5d2 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c
@@ -1812,7 +1812,6 @@ __mlx5_tc_ct_flow_offload(struct mlx5_tc_ct_priv *ct_priv,
 
 	ct_flow = kzalloc(sizeof(*ct_flow), GFP_KERNEL);
 	if (!ct_flow) {
-		kfree(ct_flow);
 		return ERR_PTR(-ENOMEM);
 	}
 
-- 
2.35.1

[net-next 03/15] net/mlx5: Delete redundant default assignment of runtime devlink params

[Saeed Mahameed]

From: Shay Drory <shayd@nvidia.com>

Runtime devlink params always read their values from the get() callbacks.
Also, it is an error to set driverinit_value for params which don't
support driverinit cmode. Delete such assignments.

In addition, move the set of default matching mode inside eswitch code.

Signed-off-by: Shay Drory <shayd@nvidia.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 .../net/ethernet/mellanox/mlx5/core/devlink.c | 20 -------------------
 .../net/ethernet/mellanox/mlx5/core/eswitch.c |  3 +++
 2 files changed, 3 insertions(+), 20 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/devlink.c b/drivers/net/ethernet/mellanox/mlx5/core/devlink.c
index 057dde6f4417..e8789e6d7e7b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/devlink.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/devlink.c
@@ -584,14 +584,6 @@ static void mlx5_devlink_set_params_init_values(struct devlink *devlink)
 	struct mlx5_core_dev *dev = devlink_priv(devlink);
 	union devlink_param_value value;
 
-	if (dev->priv.steering->mode == MLX5_FLOW_STEERING_MODE_DMFS)
-		strcpy(value.vstr, "dmfs");
-	else
-		strcpy(value.vstr, "smfs");
-	devlink_param_driverinit_value_set(devlink,
-					   MLX5_DEVLINK_PARAM_ID_FLOW_STEERING_MODE,
-					   value);
-
 	value.vbool = MLX5_CAP_GEN(dev, roce);
 	devlink_param_driverinit_value_set(devlink,
 					   DEVLINK_PARAM_GENERIC_ID_ENABLE_ROCE,
@@ -602,18 +594,6 @@ static void mlx5_devlink_set_params_init_values(struct devlink *devlink)
 	devlink_param_driverinit_value_set(devlink,
 					   MLX5_DEVLINK_PARAM_ID_ESW_LARGE_GROUP_NUM,
 					   value);
-
-	if (MLX5_ESWITCH_MANAGER(dev)) {
-		if (mlx5_esw_vport_match_metadata_supported(dev->priv.eswitch)) {
-			dev->priv.eswitch->flags |= MLX5_ESWITCH_VPORT_MATCH_METADATA;
-			value.vbool = true;
-		} else {
-			value.vbool = false;
-		}
-		devlink_param_driverinit_value_set(devlink,
-						   MLX5_DEVLINK_PARAM_ID_ESW_PORT_METADATA,
-						   value);
-	}
 #endif
 
 	value.vu32 = MLX5_COMP_EQ_SIZE;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index 458ec0bca1b8..25f2d2717aaa 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -1582,6 +1582,9 @@ int mlx5_eswitch_init(struct mlx5_core_dev *dev)
 		esw->offloads.encap = DEVLINK_ESWITCH_ENCAP_MODE_BASIC;
 	else
 		esw->offloads.encap = DEVLINK_ESWITCH_ENCAP_MODE_NONE;
+	if (MLX5_ESWITCH_MANAGER(dev) &&
+	    mlx5_esw_vport_match_metadata_supported(esw))
+		esw->flags |= MLX5_ESWITCH_VPORT_MATCH_METADATA;
 
 	dev->priv.eswitch = esw;
 	BLOCKING_INIT_NOTIFIER_HEAD(&esw->n_head);
-- 
2.35.1

[net-next 04/15] net/mlx5: Print initializing field in case of timeout

[Saeed Mahameed]

From: Shay Drory <shayd@nvidia.com>

Print the initializing field in case of FW couldn't initialize before
timeout. This will help to better understand the root cause in some
cases.

Signed-off-by: Shay Drory <shayd@nvidia.com>
Reviewed-by: Maor Gottlieb <maorg@nvidia.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/main.c | 17 ++++++++---------
 1 file changed, 8 insertions(+), 9 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c
index d504c8cb8f96..95d7712c2d9a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c
@@ -177,30 +177,29 @@ static struct mlx5_profile profile[] = {
 	},
 };
 
-static int fw_initializing(struct mlx5_core_dev *dev)
-{
-	return ioread32be(&dev->iseg->initializing) >> 31;
-}
-
 static int wait_fw_init(struct mlx5_core_dev *dev, u32 max_wait_mili,
 			u32 warn_time_mili)
 {
 	unsigned long warn = jiffies + msecs_to_jiffies(warn_time_mili);
 	unsigned long end = jiffies + msecs_to_jiffies(max_wait_mili);
+	u32 fw_initializing;
 	int err = 0;
 
-	while (fw_initializing(dev)) {
+	do {
+		fw_initializing = ioread32be(&dev->iseg->initializing);
+		if (!(fw_initializing >> 31))
+			break;
 		if (time_after(jiffies, end)) {
 			err = -EBUSY;
 			break;
 		}
 		if (warn_time_mili && time_after(jiffies, warn)) {
-			mlx5_core_warn(dev, "Waiting for FW initialization, timeout abort in %ds\n",
-				       jiffies_to_msecs(end - warn) / 1000);
+			mlx5_core_warn(dev, "Waiting for FW initialization, timeout abort in %ds (0x%x)\n",
+				       jiffies_to_msecs(end - warn) / 1000, fw_initializing);
 			warn = jiffies + msecs_to_jiffies(warn_time_mili);
 		}
 		msleep(mlx5_tout_ms(dev, FW_PRE_INIT_WAIT));
-	}
+	} while (true);
 
 	return err;
 }
-- 
2.35.1

[net-next 05/15] net/mlx5e: Drop error CQE handling from the XSK RX handler

[Saeed Mahameed]

From: Maxim Mikityanskiy <maximmi@nvidia.com>

This commit removes the redundant check and removes the unused cqe parameter
of skb_from_cqe handlers.

Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Gal Pressman <gal@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en.h     |  4 ++--
 .../net/ethernet/mellanox/mlx5/core/en/xsk/rx.c  |  6 ------
 .../net/ethernet/mellanox/mlx5/core/en/xsk/rx.h  |  1 -
 drivers/net/ethernet/mellanox/mlx5/core/en_rx.c  | 16 ++++++++--------
 4 files changed, 10 insertions(+), 17 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index 50818081bdc0..b90902db7819 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -648,8 +648,8 @@ typedef struct sk_buff *
 (*mlx5e_fp_skb_from_cqe_mpwrq)(struct mlx5e_rq *rq, struct mlx5e_mpw_info *wi,
 			       u16 cqe_bcnt, u32 head_offset, u32 page_idx);
 typedef struct sk_buff *
-(*mlx5e_fp_skb_from_cqe)(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
-			 struct mlx5e_wqe_frag_info *wi, u32 cqe_bcnt);
+(*mlx5e_fp_skb_from_cqe)(struct mlx5e_rq *rq, struct mlx5e_wqe_frag_info *wi,
+			 u32 cqe_bcnt);
 typedef bool (*mlx5e_fp_post_rx_wqes)(struct mlx5e_rq *rq);
 typedef void (*mlx5e_fp_dealloc_wqe)(struct mlx5e_rq*, u16);
 typedef void (*mlx5e_fp_shampo_dealloc_hd)(struct mlx5e_rq*, u16, u16, bool);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/rx.c
index 021da085e603..9a1553598a7c 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/rx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/rx.c
@@ -80,7 +80,6 @@ struct sk_buff *mlx5e_xsk_skb_from_cqe_mpwrq_linear(struct mlx5e_rq *rq,
 }
 
 struct sk_buff *mlx5e_xsk_skb_from_cqe_linear(struct mlx5e_rq *rq,
-					      struct mlx5_cqe64 *cqe,
 					      struct mlx5e_wqe_frag_info *wi,
 					      u32 cqe_bcnt)
 {
@@ -99,11 +98,6 @@ struct sk_buff *mlx5e_xsk_skb_from_cqe_linear(struct mlx5e_rq *rq,
 	xsk_buff_dma_sync_for_cpu(xdp, rq->xsk_pool);
 	net_prefetch(xdp->data);
 
-	if (unlikely(get_cqe_opcode(cqe) != MLX5_CQE_RESP_SEND)) {
-		rq->stats->wqe_err++;
-		return NULL;
-	}
-
 	prog = rcu_dereference(rq->xdp_prog);
 	if (likely(prog && mlx5e_xdp_handle(rq, NULL, prog, xdp)))
 		return NULL; /* page/packet was consumed by XDP */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/rx.h b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/rx.h
index 7f88ccf67fdd..a8cfab4a393c 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/rx.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/rx.h
@@ -15,7 +15,6 @@ struct sk_buff *mlx5e_xsk_skb_from_cqe_mpwrq_linear(struct mlx5e_rq *rq,
 						    u32 head_offset,
 						    u32 page_idx);
 struct sk_buff *mlx5e_xsk_skb_from_cqe_linear(struct mlx5e_rq *rq,
-					      struct mlx5_cqe64 *cqe,
 					      struct mlx5e_wqe_frag_info *wi,
 					      u32 cqe_bcnt);
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
index a5f6fd16b665..2dea9e4649a6 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
@@ -1521,8 +1521,8 @@ static void mlx5e_fill_xdp_buff(struct mlx5e_rq *rq, void *va, u16 headroom,
 }
 
 static struct sk_buff *
-mlx5e_skb_from_cqe_linear(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
-			  struct mlx5e_wqe_frag_info *wi, u32 cqe_bcnt)
+mlx5e_skb_from_cqe_linear(struct mlx5e_rq *rq, struct mlx5e_wqe_frag_info *wi,
+			  u32 cqe_bcnt)
 {
 	struct mlx5e_dma_info *di = wi->di;
 	u16 rx_headroom = rq->buff.headroom;
@@ -1565,8 +1565,8 @@ mlx5e_skb_from_cqe_linear(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
 }
 
 static struct sk_buff *
-mlx5e_skb_from_cqe_nonlinear(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
-			     struct mlx5e_wqe_frag_info *wi, u32 cqe_bcnt)
+mlx5e_skb_from_cqe_nonlinear(struct mlx5e_rq *rq, struct mlx5e_wqe_frag_info *wi,
+			     u32 cqe_bcnt)
 {
 	struct mlx5e_rq_frag_info *frag_info = &rq->wqe.info.arr[0];
 	struct mlx5e_wqe_frag_info *head_wi = wi;
@@ -1709,7 +1709,7 @@ static void mlx5e_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe)
 	skb = INDIRECT_CALL_2(rq->wqe.skb_from_cqe,
 			      mlx5e_skb_from_cqe_linear,
 			      mlx5e_skb_from_cqe_nonlinear,
-			      rq, cqe, wi, cqe_bcnt);
+			      rq, wi, cqe_bcnt);
 	if (!skb) {
 		/* probably for XDP */
 		if (__test_and_clear_bit(MLX5E_RQ_FLAG_XDP_XMIT, rq->flags)) {
@@ -1762,7 +1762,7 @@ static void mlx5e_handle_rx_cqe_rep(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe)
 	skb = INDIRECT_CALL_2(rq->wqe.skb_from_cqe,
 			      mlx5e_skb_from_cqe_linear,
 			      mlx5e_skb_from_cqe_nonlinear,
-			      rq, cqe, wi, cqe_bcnt);
+			      rq, wi, cqe_bcnt);
 	if (!skb) {
 		/* probably for XDP */
 		if (__test_and_clear_bit(MLX5E_RQ_FLAG_XDP_XMIT, rq->flags)) {
@@ -2361,7 +2361,7 @@ static void mlx5i_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe)
 	skb = INDIRECT_CALL_2(rq->wqe.skb_from_cqe,
 			      mlx5e_skb_from_cqe_linear,
 			      mlx5e_skb_from_cqe_nonlinear,
-			      rq, cqe, wi, cqe_bcnt);
+			      rq, wi, cqe_bcnt);
 	if (!skb)
 		goto wq_free_wqe;
 
@@ -2453,7 +2453,7 @@ static void mlx5e_trap_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe
 		goto free_wqe;
 	}
 
-	skb = mlx5e_skb_from_cqe_nonlinear(rq, cqe, wi, cqe_bcnt);
+	skb = mlx5e_skb_from_cqe_nonlinear(rq, wi, cqe_bcnt);
 	if (!skb)
 		goto free_wqe;
 
-- 
2.35.1

[net-next 06/15] net/mlx5e: Remove unused mlx5e_dcbnl_build_rep_netdev function

[Saeed Mahameed]

From: Gal Pressman <gal@nvidia.com>

Commit
7a9fb35e8c3a ("net/mlx5e: Do not reload ethernet ports when changing eswitch mode")
removed the usage of mlx5e_dcbnl_build_rep_netdev() from the driver,
delete the function.

Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en/dcbnl.h | 2 --
 drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c | 9 ---------
 2 files changed, 11 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/dcbnl.h b/drivers/net/ethernet/mellanox/mlx5/core/en/dcbnl.h
index 9976de8b9047..b59aee75de94 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/dcbnl.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/dcbnl.h
@@ -40,13 +40,11 @@ struct mlx5e_dcbx_dp {
 };
 
 void mlx5e_dcbnl_build_netdev(struct net_device *netdev);
-void mlx5e_dcbnl_build_rep_netdev(struct net_device *netdev);
 void mlx5e_dcbnl_initialize(struct mlx5e_priv *priv);
 void mlx5e_dcbnl_init_app(struct mlx5e_priv *priv);
 void mlx5e_dcbnl_delete_app(struct mlx5e_priv *priv);
 #else
 static inline void mlx5e_dcbnl_build_netdev(struct net_device *netdev) {}
-static inline void mlx5e_dcbnl_build_rep_netdev(struct net_device *netdev) {}
 static inline void mlx5e_dcbnl_initialize(struct mlx5e_priv *priv) {}
 static inline void mlx5e_dcbnl_init_app(struct mlx5e_priv *priv) {}
 static inline void mlx5e_dcbnl_delete_app(struct mlx5e_priv *priv) {}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c b/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c
index d659fe07d464..6435517c0bee 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c
@@ -1026,15 +1026,6 @@ void mlx5e_dcbnl_build_netdev(struct net_device *netdev)
 		netdev->dcbnl_ops = &mlx5e_dcbnl_ops;
 }
 
-void mlx5e_dcbnl_build_rep_netdev(struct net_device *netdev)
-{
-	struct mlx5e_priv *priv = netdev_priv(netdev);
-	struct mlx5_core_dev *mdev = priv->mdev;
-
-	if (MLX5_CAP_GEN(mdev, qos))
-		netdev->dcbnl_ops = &mlx5e_dcbnl_ops;
-}
-
 static void mlx5e_dcbnl_query_dcbx_mode(struct mlx5e_priv *priv,
 					enum mlx5_dcbx_oper_mode *mode)
 {
-- 
2.35.1

[net-next 07/15] net/mlx5e: TC, set proper dest type

[Saeed Mahameed]

From: Mark Bloch <mbloch@nvidia.com>

Dest type isn't set, this works only because
MLX5_FLOW_DESTINATION_TYPE_VPORT is zero. Set the proper type.

Signed-off-by: Mark Bloch <mbloch@nvidia.com>
Reviewed-by: Maor Gottlieb <maorg@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en/tc/sample.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/sample.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/sample.c
index fd4504518578..1cbd2eb9d04f 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/sample.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/sample.c
@@ -93,6 +93,7 @@ sampler_termtbl_create(struct mlx5e_tc_psample *tc_psample)
 
 	act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
 	dest.vport.num = esw->manager_vport;
+	dest.type = MLX5_FLOW_DESTINATION_TYPE_VPORT;
 	tc_psample->termtbl_rule = mlx5_add_flow_rules(tc_psample->termtbl, NULL, &act, &dest, 1);
 	if (IS_ERR(tc_psample->termtbl_rule)) {
 		err = PTR_ERR(tc_psample->termtbl_rule);
-- 
2.35.1

[net-next 08/15] net/mlx5: fs, split software and IFC flow destination definitions

[Saeed Mahameed]

From: Mark Bloch <mbloch@nvidia.com>

Separate flow destinations between software and IFC.
Flow destination type passed by callers was used as the input in
firmware commands and over the years software only types were added
which resulted in mixing between the two.

Create an IFC enum that contains only the flow destinations defined
when talking to the firmware.

Now that there is a proper software only enum for flow destinations
the hardcoded values can be removed as the values are no longer used
in firmware commands.

Signed-off-by: Mark Bloch <mbloch@nvidia.com>
Reviewed-by: Maor Gottlieb <maorg@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c | 13 ++++++++++---
 .../mellanox/mlx5/core/steering/dr_cmd.c         | 16 ++++++++++++----
 include/linux/mlx5/fs.h                          | 11 +++++++++++
 include/linux/mlx5/mlx5_ifc.h                    | 16 ++++++----------
 4 files changed, 39 insertions(+), 17 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c
index 33e9f86cf7d4..a5662cb46660 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c
@@ -571,7 +571,9 @@ static int mlx5_cmd_set_fte(struct mlx5_core_dev *dev,
 		int list_size = 0;
 
 		list_for_each_entry(dst, &fte->node.children, node.list) {
-			unsigned int id, type = dst->dest_attr.type;
+			enum mlx5_flow_destination_type type = dst->dest_attr.type;
+			enum mlx5_ifc_flow_destination_type ifc_type;
+			unsigned int id;
 
 			if (type == MLX5_FLOW_DESTINATION_TYPE_COUNTER)
 				continue;
@@ -579,10 +581,11 @@ static int mlx5_cmd_set_fte(struct mlx5_core_dev *dev,
 			switch (type) {
 			case MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE_NUM:
 				id = dst->dest_attr.ft_num;
-				type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+				ifc_type = MLX5_IFC_FLOW_DESTINATION_TYPE_FLOW_TABLE;
 				break;
 			case MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE:
 				id = dst->dest_attr.ft->id;
+				ifc_type = MLX5_IFC_FLOW_DESTINATION_TYPE_FLOW_TABLE;
 				break;
 			case MLX5_FLOW_DESTINATION_TYPE_UPLINK:
 			case MLX5_FLOW_DESTINATION_TYPE_VPORT:
@@ -596,8 +599,10 @@ static int mlx5_cmd_set_fte(struct mlx5_core_dev *dev,
 				if (type == MLX5_FLOW_DESTINATION_TYPE_UPLINK) {
 					/* destination_id is reserved */
 					id = 0;
+					ifc_type = MLX5_IFC_FLOW_DESTINATION_TYPE_UPLINK;
 					break;
 				}
+				ifc_type = MLX5_IFC_FLOW_DESTINATION_TYPE_VPORT;
 				id = dst->dest_attr.vport.num;
 				if (extended_dest &&
 				    dst->dest_attr.vport.pkt_reformat) {
@@ -612,13 +617,15 @@ static int mlx5_cmd_set_fte(struct mlx5_core_dev *dev,
 				break;
 			case MLX5_FLOW_DESTINATION_TYPE_FLOW_SAMPLER:
 				id = dst->dest_attr.sampler_id;
+				ifc_type = MLX5_IFC_FLOW_DESTINATION_TYPE_FLOW_SAMPLER;
 				break;
 			default:
 				id = dst->dest_attr.tir_num;
+				ifc_type = MLX5_IFC_FLOW_DESTINATION_TYPE_TIR;
 			}
 
 			MLX5_SET(dest_format_struct, in_dests, destination_type,
-				 type);
+				 ifc_type);
 			MLX5_SET(dest_format_struct, in_dests, destination_id, id);
 			in_dests += dst_cnt_size;
 			list_size++;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_cmd.c
index 4dd619d238cc..728ccb950fec 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_cmd.c
@@ -311,7 +311,7 @@ int mlx5dr_cmd_set_fte_modify_and_vport(struct mlx5_core_dev *mdev,
 
 	in_dests = MLX5_ADDR_OF(flow_context, in_flow_context, destination);
 	MLX5_SET(dest_format_struct, in_dests, destination_type,
-		 MLX5_FLOW_DESTINATION_TYPE_VPORT);
+		 MLX5_IFC_FLOW_DESTINATION_TYPE_VPORT);
 	MLX5_SET(dest_format_struct, in_dests, destination_id, vport);
 
 	err = mlx5_cmd_exec(mdev, in, inlen, out, sizeof(out));
@@ -719,7 +719,9 @@ int mlx5dr_cmd_set_fte(struct mlx5_core_dev *dev,
 		int list_size = 0;
 
 		for (i = 0; i < fte->dests_size; i++) {
-			unsigned int id, type = fte->dest_arr[i].type;
+			enum mlx5_flow_destination_type type = fte->dest_arr[i].type;
+			enum mlx5_ifc_flow_destination_type ifc_type;
+			unsigned int id;
 
 			if (type == MLX5_FLOW_DESTINATION_TYPE_COUNTER)
 				continue;
@@ -727,10 +729,12 @@ int mlx5dr_cmd_set_fte(struct mlx5_core_dev *dev,
 			switch (type) {
 			case MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE_NUM:
 				id = fte->dest_arr[i].ft_num;
-				type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+				ifc_type = MLX5_IFC_FLOW_DESTINATION_TYPE_FLOW_TABLE;
 				break;
 			case MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE:
 				id = fte->dest_arr[i].ft_id;
+				ifc_type = MLX5_IFC_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+
 				break;
 			case MLX5_FLOW_DESTINATION_TYPE_UPLINK:
 			case MLX5_FLOW_DESTINATION_TYPE_VPORT:
@@ -740,8 +744,10 @@ int mlx5dr_cmd_set_fte(struct mlx5_core_dev *dev,
 						 destination_eswitch_owner_vhca_id_valid,
 						 !!(fte->dest_arr[i].vport.flags &
 						    MLX5_FLOW_DEST_VPORT_VHCA_ID));
+					ifc_type = MLX5_IFC_FLOW_DESTINATION_TYPE_VPORT;
 				} else {
 					id = 0;
+					ifc_type = MLX5_IFC_FLOW_DESTINATION_TYPE_UPLINK;
 					MLX5_SET(dest_format_struct, in_dests,
 						 destination_eswitch_owner_vhca_id_valid, 1);
 				}
@@ -761,13 +767,15 @@ int mlx5dr_cmd_set_fte(struct mlx5_core_dev *dev,
 				break;
 			case MLX5_FLOW_DESTINATION_TYPE_FLOW_SAMPLER:
 				id = fte->dest_arr[i].sampler_id;
+				ifc_type = MLX5_IFC_FLOW_DESTINATION_TYPE_FLOW_SAMPLER;
 				break;
 			default:
 				id = fte->dest_arr[i].tir_num;
+				ifc_type = MLX5_IFC_FLOW_DESTINATION_TYPE_TIR;
 			}
 
 			MLX5_SET(dest_format_struct, in_dests, destination_type,
-				 type);
+				 ifc_type);
 			MLX5_SET(dest_format_struct, in_dests, destination_id, id);
 			in_dests += dst_cnt_size;
 			list_size++;
diff --git a/include/linux/mlx5/fs.h b/include/linux/mlx5/fs.h
index e3bfed68b08a..9da9df9ae751 100644
--- a/include/linux/mlx5/fs.h
+++ b/include/linux/mlx5/fs.h
@@ -40,6 +40,17 @@
 
 #define MLX5_SET_CFG(p, f, v) MLX5_SET(create_flow_group_in, p, f, v)
 
+enum mlx5_flow_destination_type {
+	MLX5_FLOW_DESTINATION_TYPE_VPORT,
+	MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE,
+	MLX5_FLOW_DESTINATION_TYPE_TIR,
+	MLX5_FLOW_DESTINATION_TYPE_FLOW_SAMPLER,
+	MLX5_FLOW_DESTINATION_TYPE_UPLINK,
+	MLX5_FLOW_DESTINATION_TYPE_PORT,
+	MLX5_FLOW_DESTINATION_TYPE_COUNTER,
+	MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE_NUM,
+};
+
 enum {
 	MLX5_FLOW_CONTEXT_ACTION_FWD_NEXT_PRIO	= 1 << 16,
 	MLX5_FLOW_CONTEXT_ACTION_ENCRYPT	= 1 << 17,
diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h
index 7d2d0ba82144..7f4ec9faa180 100644
--- a/include/linux/mlx5/mlx5_ifc.h
+++ b/include/linux/mlx5/mlx5_ifc.h
@@ -1806,16 +1806,12 @@ struct mlx5_ifc_cmd_hca_cap_2_bits {
 	u8	   reserved_at_c0[0x740];
 };
 
-enum mlx5_flow_destination_type {
-	MLX5_FLOW_DESTINATION_TYPE_VPORT        = 0x0,
-	MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE   = 0x1,
-	MLX5_FLOW_DESTINATION_TYPE_TIR          = 0x2,
-	MLX5_FLOW_DESTINATION_TYPE_FLOW_SAMPLER = 0x6,
-	MLX5_FLOW_DESTINATION_TYPE_UPLINK       = 0x8,
-
-	MLX5_FLOW_DESTINATION_TYPE_PORT         = 0x99,
-	MLX5_FLOW_DESTINATION_TYPE_COUNTER      = 0x100,
-	MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE_NUM = 0x101,
+enum mlx5_ifc_flow_destination_type {
+	MLX5_IFC_FLOW_DESTINATION_TYPE_VPORT        = 0x0,
+	MLX5_IFC_FLOW_DESTINATION_TYPE_FLOW_TABLE   = 0x1,
+	MLX5_IFC_FLOW_DESTINATION_TYPE_TIR          = 0x2,
+	MLX5_IFC_FLOW_DESTINATION_TYPE_FLOW_SAMPLER = 0x6,
+	MLX5_IFC_FLOW_DESTINATION_TYPE_UPLINK       = 0x8,
 };
 
 enum mlx5_flow_table_miss_action {
-- 
2.35.1

[net-next 09/15] net/mlx5: fs, refactor software deletion rule

[Saeed Mahameed]

From: Mark Bloch <mbloch@nvidia.com>

When deleting a rule make sure that for every type dests_size is
decreased only once and no other logic is executed.

Without this dests_size might be decreased twice when dests_size == 1
so the if for that type won't be entered and if action has
MLX5_FLOW_CONTEXT_ACTION_FWD_DEST set dests_size will be decreased again.

Signed-off-by: Mark Bloch <mbloch@nvidia.com>
Reviewed-by: Maor Gottlieb <maorg@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/fs_core.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
index 297e6a468a3e..ae83962fc5fc 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
@@ -550,8 +550,8 @@ static void del_sw_hw_rule(struct fs_node *node)
 		mutex_unlock(&rule->dest_attr.ft->lock);
 	}
 
-	if (rule->dest_attr.type == MLX5_FLOW_DESTINATION_TYPE_COUNTER  &&
-	    --fte->dests_size) {
+	if (rule->dest_attr.type == MLX5_FLOW_DESTINATION_TYPE_COUNTER) {
+		--fte->dests_size;
 		fte->modify_mask |=
 			BIT(MLX5_SET_FTE_MODIFY_ENABLE_MASK_ACTION) |
 			BIT(MLX5_SET_FTE_MODIFY_ENABLE_MASK_FLOW_COUNTERS);
@@ -559,15 +559,15 @@ static void del_sw_hw_rule(struct fs_node *node)
 		goto out;
 	}
 
-	if (rule->dest_attr.type == MLX5_FLOW_DESTINATION_TYPE_PORT &&
-	    --fte->dests_size) {
+	if (rule->dest_attr.type == MLX5_FLOW_DESTINATION_TYPE_PORT) {
+		--fte->dests_size;
 		fte->modify_mask |= BIT(MLX5_SET_FTE_MODIFY_ENABLE_MASK_ACTION);
 		fte->action.action &= ~MLX5_FLOW_CONTEXT_ACTION_ALLOW;
 		goto out;
 	}
 
-	if ((fte->action.action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST) &&
-	    --fte->dests_size) {
+	if (fte->action.action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST) {
+		--fte->dests_size;
 		fte->modify_mask |=
 			BIT(MLX5_SET_FTE_MODIFY_ENABLE_MASK_DESTINATION_LIST);
 	}
-- 
2.35.1

[net-next 10/15] net/mlx5: fs, jump to exit point and don't fall through

[Saeed Mahameed]

From: Mark Bloch <mbloch@nvidia.com>

For code clarity and to prevent future bugs make sure to jump
to the exit point once done handling that specific type.
This aligns the code with the rest logic in the function.

Signed-off-by: Mark Bloch <mbloch@nvidia.com>
Reviewed-by: Maor Gottlieb <maorg@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/fs_core.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
index ae83962fc5fc..e282d80f1fd2 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
@@ -570,6 +570,7 @@ static void del_sw_hw_rule(struct fs_node *node)
 		--fte->dests_size;
 		fte->modify_mask |=
 			BIT(MLX5_SET_FTE_MODIFY_ENABLE_MASK_DESTINATION_LIST);
+		goto out;
 	}
 out:
 	kfree(rule);
-- 
2.35.1

[net-next 11/15] net/mlx5: fs, add unused destination type

[Saeed Mahameed]

From: Mark Bloch <mbloch@nvidia.com>

When the caller doesn't pass a destination fs_core will create a unused
rule just so a context can be returned. This unused rule
is zeroed out and its type is 0 which can be mixed up with
MLX5_FLOW_DESTINATION_TYPE_VPORT.

Create a dedicated type to differentiate between the two
named MLX5_FLOW_DESTINATION_TYPE_NONE.

Signed-off-by: Mark Bloch <mbloch@nvidia.com>
Reviewed-by: Maor Gottlieb <maorg@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/diag/fs_tracepoint.c | 3 +++
 drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c             | 5 ++++-
 drivers/net/ethernet/mellanox/mlx5/core/fs_core.c            | 2 ++
 drivers/net/ethernet/mellanox/mlx5/core/steering/dr_cmd.c    | 5 ++++-
 include/linux/mlx5/fs.h                                      | 1 +
 5 files changed, 14 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/diag/fs_tracepoint.c b/drivers/net/ethernet/mellanox/mlx5/core/diag/fs_tracepoint.c
index 7841ef6c193c..c5bb79a4fa57 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/diag/fs_tracepoint.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/diag/fs_tracepoint.c
@@ -259,6 +259,9 @@ const char *parse_fs_dst(struct trace_seq *p,
 	case MLX5_FLOW_DESTINATION_TYPE_PORT:
 		trace_seq_printf(p, "port\n");
 		break;
+	case MLX5_FLOW_DESTINATION_TYPE_NONE:
+		trace_seq_printf(p, "none\n");
+		break;
 	}
 
 	trace_seq_putc(p, 0);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c
index a5662cb46660..2ccf7bef9b05 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c
@@ -455,7 +455,8 @@ static int mlx5_set_extended_dest(struct mlx5_core_dev *dev,
 		return 0;
 
 	list_for_each_entry(dst, &fte->node.children, node.list) {
-		if (dst->dest_attr.type == MLX5_FLOW_DESTINATION_TYPE_COUNTER)
+		if (dst->dest_attr.type == MLX5_FLOW_DESTINATION_TYPE_COUNTER ||
+		    dst->dest_attr.type == MLX5_FLOW_DESTINATION_TYPE_NONE)
 			continue;
 		if ((dst->dest_attr.type == MLX5_FLOW_DESTINATION_TYPE_VPORT ||
 		     dst->dest_attr.type == MLX5_FLOW_DESTINATION_TYPE_UPLINK) &&
@@ -579,6 +580,8 @@ static int mlx5_cmd_set_fte(struct mlx5_core_dev *dev,
 				continue;
 
 			switch (type) {
+			case MLX5_FLOW_DESTINATION_TYPE_NONE:
+				continue;
 			case MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE_NUM:
 				id = dst->dest_attr.ft_num;
 				ifc_type = MLX5_IFC_FLOW_DESTINATION_TYPE_FLOW_TABLE;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
index e282d80f1fd2..f9d6ddd865e0 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
@@ -1289,6 +1289,8 @@ static struct mlx5_flow_rule *alloc_rule(struct mlx5_flow_destination *dest)
 	rule->node.type = FS_TYPE_FLOW_DEST;
 	if (dest)
 		memcpy(&rule->dest_attr, dest, sizeof(*dest));
+	else
+		rule->dest_attr.type = MLX5_FLOW_DESTINATION_TYPE_NONE;
 
 	return rule;
 }
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_cmd.c
index 728ccb950fec..223c8741b7ae 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_cmd.c
@@ -604,7 +604,8 @@ static int mlx5dr_cmd_set_extended_dest(struct mlx5_core_dev *dev,
 	if (!(fte->action.action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST))
 		return 0;
 	for (i = 0; i < fte->dests_size; i++) {
-		if (fte->dest_arr[i].type == MLX5_FLOW_DESTINATION_TYPE_COUNTER)
+		if (fte->dest_arr[i].type == MLX5_FLOW_DESTINATION_TYPE_COUNTER ||
+		    fte->dest_arr[i].type == MLX5_FLOW_DESTINATION_TYPE_NONE)
 			continue;
 		if ((fte->dest_arr[i].type == MLX5_FLOW_DESTINATION_TYPE_VPORT ||
 		     fte->dest_arr[i].type == MLX5_FLOW_DESTINATION_TYPE_UPLINK) &&
@@ -727,6 +728,8 @@ int mlx5dr_cmd_set_fte(struct mlx5_core_dev *dev,
 				continue;
 
 			switch (type) {
+			case MLX5_FLOW_DESTINATION_TYPE_NONE:
+				continue;
 			case MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE_NUM:
 				id = fte->dest_arr[i].ft_num;
 				ifc_type = MLX5_IFC_FLOW_DESTINATION_TYPE_FLOW_TABLE;
diff --git a/include/linux/mlx5/fs.h b/include/linux/mlx5/fs.h
index 9da9df9ae751..8135713b0d2d 100644
--- a/include/linux/mlx5/fs.h
+++ b/include/linux/mlx5/fs.h
@@ -41,6 +41,7 @@
 #define MLX5_SET_CFG(p, f, v) MLX5_SET(create_flow_group_in, p, f, v)
 
 enum mlx5_flow_destination_type {
+	MLX5_FLOW_DESTINATION_TYPE_NONE,
 	MLX5_FLOW_DESTINATION_TYPE_VPORT,
 	MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE,
 	MLX5_FLOW_DESTINATION_TYPE_TIR,
-- 
2.35.1

[net-next 12/15] net/mlx5: fs, do proper bookkeeping for forward destinations

[Saeed Mahameed]

From: Mark Bloch <mbloch@nvidia.com>

Keep track after destinations that are forward destinations.
When a forward destinations is removed from an FTE check if
the actions bits need to be updated.

Signed-off-by: Mark Bloch <mbloch@nvidia.com>
Reviewed-by: Maor Gottlieb <maorg@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 .../net/ethernet/mellanox/mlx5/core/fs_core.c | 20 ++++++++++++++++++-
 .../net/ethernet/mellanox/mlx5/core/fs_core.h |  1 +
 2 files changed, 20 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
index f9d6ddd865e0..ec91727eee2a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
@@ -424,6 +424,16 @@ static bool is_fwd_next_action(u32 action)
 			 MLX5_FLOW_CONTEXT_ACTION_FWD_NEXT_NS);
 }
 
+static bool is_fwd_dest_type(enum mlx5_flow_destination_type type)
+{
+	return type == MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE_NUM ||
+		type == MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE ||
+		type == MLX5_FLOW_DESTINATION_TYPE_UPLINK ||
+		type == MLX5_FLOW_DESTINATION_TYPE_VPORT ||
+		type == MLX5_FLOW_DESTINATION_TYPE_FLOW_SAMPLER ||
+		type == MLX5_FLOW_DESTINATION_TYPE_TIR;
+}
+
 static bool check_valid_spec(const struct mlx5_flow_spec *spec)
 {
 	int i;
@@ -566,8 +576,13 @@ static void del_sw_hw_rule(struct fs_node *node)
 		goto out;
 	}
 
-	if (fte->action.action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST) {
+	if (is_fwd_dest_type(rule->dest_attr.type)) {
 		--fte->dests_size;
+		--fte->fwd_dests;
+
+		if (!fte->fwd_dests)
+			fte->action.action &=
+				~MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
 		fte->modify_mask |=
 			BIT(MLX5_SET_FTE_MODIFY_ENABLE_MASK_DESTINATION_LIST);
 		goto out;
@@ -1367,6 +1382,9 @@ create_flow_handle(struct fs_fte *fte,
 		if (dest) {
 			fte->dests_size++;
 
+			if (is_fwd_dest_type(dest[i].type))
+				fte->fwd_dests++;
+
 			type = dest[i].type ==
 				MLX5_FLOW_DESTINATION_TYPE_COUNTER;
 			*modify_mask |= type ? count : dst;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h
index c488a7c5b07e..67cad7a6d836 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h
@@ -226,6 +226,7 @@ struct fs_fte {
 	struct mlx5_fs_dr_rule		fs_dr_rule;
 	u32				val[MLX5_ST_SZ_DW_MATCH_PARAM];
 	u32				dests_size;
+	u32				fwd_dests;
 	u32				index;
 	struct mlx5_flow_context	flow_context;
 	struct mlx5_flow_act		action;
-- 
2.35.1

[net-next 13/15] net/mlx5: fs, delete the FTE when there are no rules attached to it

[Saeed Mahameed]

From: Mark Bloch <mbloch@nvidia.com>

When an FTE has no children is means all the rules where removed
and the FTE can be deleted regardless of the dests_size value.
While dests_size should be 0 when there are no children
be extra careful not to leak memory or get firmware syndrome
if the proper bookkeeping of dests_size wasn't done.

Signed-off-by: Mark Bloch <mbloch@nvidia.com>
Reviewed-by: Maor Gottlieb <maorg@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/fs_core.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
index ec91727eee2a..572237f262e0 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
@@ -2084,16 +2084,16 @@ void mlx5_del_flow_rules(struct mlx5_flow_handle *handle)
 	down_write_ref_node(&fte->node, false);
 	for (i = handle->num_rules - 1; i >= 0; i--)
 		tree_remove_node(&handle->rule[i]->node, true);
-	if (fte->dests_size) {
-		if (fte->modify_mask)
-			modify_fte(fte);
-		up_write_ref_node(&fte->node, false);
-	} else if (list_empty(&fte->node.children)) {
+	if (list_empty(&fte->node.children)) {
 		del_hw_fte(&fte->node);
 		/* Avoid double call to del_hw_fte */
 		fte->node.del_hw_func = NULL;
 		up_write_ref_node(&fte->node, false);
 		tree_put_node(&fte->node, false);
+	} else if (fte->dests_size) {
+		if (fte->modify_mask)
+			modify_fte(fte);
+		up_write_ref_node(&fte->node, false);
 	} else {
 		up_write_ref_node(&fte->node, false);
 	}
-- 
2.35.1

[net-next 14/15] net/mlx5: fs, call the deletion function of the node

[Saeed Mahameed]

From: Mark Bloch <mbloch@nvidia.com>

Don't call del_hw_fte() directly, instead use the hardware deletion
function set. This is just a small cleanup and doesn't change anything
as for an FTE the deletion function is already set to del_hw_fte().

Signed-off-by: Mark Bloch <mbloch@nvidia.com>
Reviewed-by: Maor Gottlieb <maorg@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/fs_core.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
index 572237f262e0..d512445c7627 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
@@ -2085,7 +2085,7 @@ void mlx5_del_flow_rules(struct mlx5_flow_handle *handle)
 	for (i = handle->num_rules - 1; i >= 0; i--)
 		tree_remove_node(&handle->rule[i]->node, true);
 	if (list_empty(&fte->node.children)) {
-		del_hw_fte(&fte->node);
+		fte->node.del_hw_func(&fte->node);
 		/* Avoid double call to del_hw_fte */
 		fte->node.del_hw_func = NULL;
 		up_write_ref_node(&fte->node, false);
-- 
2.35.1

[net-next 15/15] net/mlx5: fs, an FTE should have no dests when deleted

[Saeed Mahameed]

From: Mark Bloch <mbloch@nvidia.com>

When deleting an FTE it should have no dests, which means
fte->dests_size should be 0. Add a WARN_ON() to catch bugs
where the proper tracking wasn't done.

Signed-off-by: Mark Bloch <mbloch@nvidia.com>
Reviewed-by: Maor Gottlieb <maorg@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/fs_core.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
index d512445c7627..fb8175672478 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
@@ -605,6 +605,7 @@ static void del_hw_fte(struct fs_node *node)
 	fs_get_obj(ft, fg->node.parent);
 
 	trace_mlx5_fs_del_fte(fte);
+	WARN_ON(fte->dests_size);
 	dev = get_dev(&ft->node);
 	root = find_root(&ft->node);
 	if (node->active) {
-- 
2.35.1

Re: [net-next 01/15] net/mlx5: use kvfree() for kvzalloc() in mlx5_ct_fs_smfs_matcher_create

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net-next.git (master)
by Saeed Mahameed <saeedm@nvidia.com>:

On Mon,  2 May 2022 21:41:55 -0700 you wrote:
> From: Ziyang Xuan <william.xuanziyang@huawei.com>
> 
> The memory of spec is allocated with kvzalloc(), the corresponding
> release function should not be kfree(), use kvfree() instead.
> 
> Generated by: scripts/coccinelle/api/kfree_mismatch.cocci
> 
> [...]

Here is the summary with links:
  - [net-next,01/15] net/mlx5: use kvfree() for kvzalloc() in mlx5_ct_fs_smfs_matcher_create
    https://git.kernel.org/netdev/net-next/c/c389362096be
  - [net-next,02/15] net/mlx5: Remove useless kfree
    https://git.kernel.org/netdev/net-next/c/7134c602812a
  - [net-next,03/15] net/mlx5: Delete redundant default assignment of runtime devlink params
    https://git.kernel.org/netdev/net-next/c/b5235a9979f9
  - [net-next,04/15] net/mlx5: Print initializing field in case of timeout
    https://git.kernel.org/netdev/net-next/c/cdfc6ffbfb39
  - [net-next,05/15] net/mlx5e: Drop error CQE handling from the XSK RX handler
    https://git.kernel.org/netdev/net-next/c/84a137f051a5
  - [net-next,06/15] net/mlx5e: Remove unused mlx5e_dcbnl_build_rep_netdev function
    https://git.kernel.org/netdev/net-next/c/a90889b4e8bd
  - [net-next,07/15] net/mlx5e: TC, set proper dest type
    https://git.kernel.org/netdev/net-next/c/c70c3336a63e
  - [net-next,08/15] net/mlx5: fs, split software and IFC flow destination definitions
    https://git.kernel.org/netdev/net-next/c/d639af621600
  - [net-next,09/15] net/mlx5: fs, refactor software deletion rule
    https://git.kernel.org/netdev/net-next/c/d49d63075e0f
  - [net-next,10/15] net/mlx5: fs, jump to exit point and don't fall through
    https://git.kernel.org/netdev/net-next/c/c3ae3a9cfe2f
  - [net-next,11/15] net/mlx5: fs, add unused destination type
    https://git.kernel.org/netdev/net-next/c/6510bc0d7cb4
  - [net-next,12/15] net/mlx5: fs, do proper bookkeeping for forward destinations
    https://git.kernel.org/netdev/net-next/c/a30c8b9025db
  - [net-next,13/15] net/mlx5: fs, delete the FTE when there are no rules attached to it
    https://git.kernel.org/netdev/net-next/c/7b0c63385976
  - [net-next,14/15] net/mlx5: fs, call the deletion function of the node
    https://git.kernel.org/netdev/net-next/c/72191a4cd525
  - [net-next,15/15] net/mlx5: fs, an FTE should have no dests when deleted
    https://git.kernel.org/netdev/net-next/c/3a09fae035c8

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html