[pull request][net 0/9] mlx5 fixes 2020-11-17

[pull request][net 0/9] mlx5 fixes 2020-11-17

[Saeed Mahameed]

Hi Jakub,

This series introduces some fixes to mlx5 driver.
Please pull and let me know if there is any problem.

For -stable v4.14
 ('net/mlx5: Disable QoS when min_rates on all VFs are zero')

For -stable v4.20
 ('net/mlx5: Add handling of port type in rule deletion')

For -stable v5.5
 ('net/mlx5: Clear bw_share upon VF disable')

For -stable v5.7
 ('net/mlx5: E-Switch, Fail mlx5_esw_modify_vport_rate if qos disabled')

For -stable v5.8
 ('net/mlx5e: Fix check if netdev is bond slave')

For -stable v5.9
 ('net/mlx5e: Fix refcount leak on kTLS RX resync')

Thanks,
Saeed.

---
The following changes since commit 1b9e2a8c99a5c021041bfb2d512dc3ed92a94ffd:

  tcp: only postpone PROBE_RTT if RTT is < current min_rtt estimate (2020-11-17 11:03:22 -0800)

are available in the Git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux.git tags/mlx5-fixes-2020-11-17

for you to fetch changes up to 68ec32daf7d50a9f7425f8607a7402c13aa0c587:

  net/mlx5: fix error return code in mlx5e_tc_nic_init() (2020-11-17 11:50:54 -0800)

----------------------------------------------------------------
mlx5-fixes-2020-11-17

----------------------------------------------------------------
Eli Cohen (1):
      net/mlx5: E-Switch, Fail mlx5_esw_modify_vport_rate if qos disabled

Huy Nguyen (2):
      net/mlx5e: Set IPsec WAs only in IP's non checksum partial case.
      net/mlx5e: Fix IPsec packet drop by mlx5e_tc_update_skb

Maor Dickman (1):
      net/mlx5e: Fix check if netdev is bond slave

Maxim Mikityanskiy (1):
      net/mlx5e: Fix refcount leak on kTLS RX resync

Michael Guralnik (1):
      net/mlx5: Add handling of port type in rule deletion

Vladyslav Tarasiuk (2):
      net/mlx5: Clear bw_share upon VF disable
      net/mlx5: Disable QoS when min_rates on all VFs are zero

Wang Hai (1):
      net/mlx5: fix error return code in mlx5e_tc_nic_init()

 .../net/ethernet/mellanox/mlx5/core/en/rep/bond.c    |  2 +-
 .../ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c  | 14 +++++++-------
 .../mellanox/mlx5/core/en_accel/ipsec_rxtx.c         |  3 +--
 .../mellanox/mlx5/core/en_accel/ipsec_rxtx.h         |  9 +++++----
 .../ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c   | 13 ++++++++-----
 drivers/net/ethernet/mellanox/mlx5/core/en_tc.c      |  4 +++-
 drivers/net/ethernet/mellanox/mlx5/core/en_tc.h      |  3 +++
 drivers/net/ethernet/mellanox/mlx5/core/en_tx.c      | 13 ++++++-------
 drivers/net/ethernet/mellanox/mlx5/core/eswitch.c    | 20 +++++++++++++-------
 drivers/net/ethernet/mellanox/mlx5/core/fs_core.c    |  7 +++++++
 10 files changed, 54 insertions(+), 34 deletions(-)

[net 1/9] net/mlx5e: Fix refcount leak on kTLS RX resync

[Saeed Mahameed]

From: Maxim Mikityanskiy <maximmi@mellanox.com>

On resync, the driver calls inet_lookup_established
(__inet6_lookup_established) that increases sk_refcnt of the socket. To
decrease it, the driver set skb->destructor to sock_edemux. However, it
didn't work well, because the TCP stack also sets this destructor for
early demux, and the refcount gets decreased only once, while increased
two times (in mlx5e and in the TCP stack). It leads to a socket leak, a
TLS context leak, which in the end leads to calling tls_dev_del twice:
on socket close and on driver unload, which in turn leads to a crash.

This commit fixes the refcount leak by calling sock_gen_put right away
after using the socket, thus fixing all the subsequent issues.

Fixes: 0419d8c9d8f8 ("net/mlx5e: kTLS, Add kTLS RX resync support")
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 .../ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c  | 13 ++++++++-----
 1 file changed, 8 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c
index 7f6221b8b1f7..6a1d82503ef8 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c
@@ -476,19 +476,22 @@ static void resync_update_sn(struct mlx5e_rq *rq, struct sk_buff *skb)
 
 	depth += sizeof(struct tcphdr);
 
-	if (unlikely(!sk || sk->sk_state == TCP_TIME_WAIT))
+	if (unlikely(!sk))
 		return;
 
-	if (unlikely(!resync_queue_get_psv(sk)))
-		return;
+	if (unlikely(sk->sk_state == TCP_TIME_WAIT))
+		goto unref;
 
-	skb->sk = sk;
-	skb->destructor = sock_edemux;
+	if (unlikely(!resync_queue_get_psv(sk)))
+		goto unref;
 
 	seq = th->seq;
 	datalen = skb->len - depth;
 	tls_offload_rx_resync_async_request_start(sk, seq, datalen);
 	rq->stats->tls_resync_req_start++;
+
+unref:
+	sock_gen_put(sk);
 }
 
 void mlx5e_ktls_rx_resync(struct net_device *netdev, struct sock *sk,
-- 
2.26.2

[net 2/9] net/mlx5e: Set IPsec WAs only in IP's non checksum partial case.

[Saeed Mahameed]

From: Huy Nguyen <huyn@mellanox.com>

The IP's checksum partial still requires L4 csum flag on Ethernet WQE.
Make the IPsec WAs only for the IP's non checksum partial case
(for example icmd packet)

Fixes: 5be019040cb7 ("net/mlx5e: IPsec: Add Connect-X IPsec Tx data path offload")
Signed-off-by: Huy Nguyen <huyn@mellanox.com>
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Alaa Hleihel <alaa@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en_tx.c | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
index 82b4419af9d4..6dd3ea3cbbed 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
@@ -144,7 +144,9 @@ static inline void mlx5e_insert_vlan(void *start, struct sk_buff *skb, u16 ihs)
 	memcpy(&vhdr->h_vlan_encapsulated_proto, skb->data + cpy1_sz, cpy2_sz);
 }
 
-/* RM 2311217: no L4 inner checksum for IPsec tunnel type packet */
+/* If packet is not IP's CHECKSUM_PARTIAL (e.g. icmd packet),
+ * need to set L3 checksum flag for IPsec
+ */
 static void
 ipsec_txwqe_build_eseg_csum(struct mlx5e_txqsq *sq, struct sk_buff *skb,
 			    struct mlx5_wqe_eth_seg *eseg)
@@ -154,7 +156,6 @@ ipsec_txwqe_build_eseg_csum(struct mlx5e_txqsq *sq, struct sk_buff *skb,
 		eseg->cs_flags |= MLX5_ETH_WQE_L3_INNER_CSUM;
 		sq->stats->csum_partial_inner++;
 	} else {
-		eseg->cs_flags |= MLX5_ETH_WQE_L4_CSUM;
 		sq->stats->csum_partial++;
 	}
 }
@@ -162,11 +163,6 @@ ipsec_txwqe_build_eseg_csum(struct mlx5e_txqsq *sq, struct sk_buff *skb,
 static inline void
 mlx5e_txwqe_build_eseg_csum(struct mlx5e_txqsq *sq, struct sk_buff *skb, struct mlx5_wqe_eth_seg *eseg)
 {
-	if (unlikely(eseg->flow_table_metadata & cpu_to_be32(MLX5_ETH_WQE_FT_META_IPSEC))) {
-		ipsec_txwqe_build_eseg_csum(sq, skb, eseg);
-		return;
-	}
-
 	if (likely(skb->ip_summed == CHECKSUM_PARTIAL)) {
 		eseg->cs_flags = MLX5_ETH_WQE_L3_CSUM;
 		if (skb->encapsulation) {
@@ -177,6 +173,9 @@ mlx5e_txwqe_build_eseg_csum(struct mlx5e_txqsq *sq, struct sk_buff *skb, struct
 			eseg->cs_flags |= MLX5_ETH_WQE_L4_CSUM;
 			sq->stats->csum_partial++;
 		}
+	} else if (unlikely(eseg->flow_table_metadata & cpu_to_be32(MLX5_ETH_WQE_FT_META_IPSEC))) {
+		ipsec_txwqe_build_eseg_csum(sq, skb, eseg);
+
 	} else
 		sq->stats->csum_none++;
 }
-- 
2.26.2

[net 3/9] net/mlx5e: Fix IPsec packet drop by mlx5e_tc_update_skb

[Saeed Mahameed]

From: Huy Nguyen <huyn@mellanox.com>

Both TC and IPsec crypto offload use metadata_regB to store
private information. Since TC does not use bit 31 of regB, IPsec
will use bit 31 as the IPsec packet marker. The IPsec's regB usage
is changed to:
Bit31: IPsec marker
Bit30-24: IPsec syndrome
Bit23-0: IPsec obj id

Fixes: b2ac7541e377 ("net/mlx5e: IPsec: Add Connect-X IPsec Rx data path offload")
Signed-off-by: Huy Nguyen <huyn@mellanox.com>
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Ariel Levkovich <lariel@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 .../mellanox/mlx5/core/en_accel/ipsec_fs.c         | 14 +++++++-------
 .../mellanox/mlx5/core/en_accel/ipsec_rxtx.c       |  3 +--
 .../mellanox/mlx5/core/en_accel/ipsec_rxtx.h       |  9 +++++----
 drivers/net/ethernet/mellanox/mlx5/core/en_tc.h    |  3 +++
 4 files changed, 16 insertions(+), 13 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c
index 0e45590662a8..381a9c8c9da9 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c
@@ -64,13 +64,13 @@ static int rx_err_add_rule(struct mlx5e_priv *priv,
 	if (!spec)
 		return -ENOMEM;
 
-	/* Action to copy 7 bit ipsec_syndrome to regB[0:6] */
+	/* Action to copy 7 bit ipsec_syndrome to regB[24:30] */
 	MLX5_SET(copy_action_in, action, action_type, MLX5_ACTION_TYPE_COPY);
 	MLX5_SET(copy_action_in, action, src_field, MLX5_ACTION_IN_FIELD_IPSEC_SYNDROME);
 	MLX5_SET(copy_action_in, action, src_offset, 0);
 	MLX5_SET(copy_action_in, action, length, 7);
 	MLX5_SET(copy_action_in, action, dst_field, MLX5_ACTION_IN_FIELD_METADATA_REG_B);
-	MLX5_SET(copy_action_in, action, dst_offset, 0);
+	MLX5_SET(copy_action_in, action, dst_offset, 24);
 
 	modify_hdr = mlx5_modify_header_alloc(mdev, MLX5_FLOW_NAMESPACE_KERNEL,
 					      1, action);
@@ -488,13 +488,13 @@ static int rx_add_rule(struct mlx5e_priv *priv,
 
 	setup_fte_common(attrs, ipsec_obj_id, spec, &flow_act);
 
-	/* Set 1  bit ipsec marker */
-	/* Set 24 bit ipsec_obj_id */
+	/* Set bit[31] ipsec marker */
+	/* Set bit[23-0] ipsec_obj_id */
 	MLX5_SET(set_action_in, action, action_type, MLX5_ACTION_TYPE_SET);
 	MLX5_SET(set_action_in, action, field, MLX5_ACTION_IN_FIELD_METADATA_REG_B);
-	MLX5_SET(set_action_in, action, data, (ipsec_obj_id << 1) | 0x1);
-	MLX5_SET(set_action_in, action, offset, 7);
-	MLX5_SET(set_action_in, action, length, 25);
+	MLX5_SET(set_action_in, action, data, (ipsec_obj_id | BIT(31)));
+	MLX5_SET(set_action_in, action, offset, 0);
+	MLX5_SET(set_action_in, action, length, 32);
 
 	modify_hdr = mlx5_modify_header_alloc(priv->mdev, MLX5_FLOW_NAMESPACE_KERNEL,
 					      1, action);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c
index 11e31a3db2be..a9b45606dbdb 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c
@@ -453,7 +453,6 @@ void mlx5e_ipsec_offload_handle_rx_skb(struct net_device *netdev,
 				       struct mlx5_cqe64 *cqe)
 {
 	u32 ipsec_meta_data = be32_to_cpu(cqe->ft_metadata);
-	u8 ipsec_syndrome = ipsec_meta_data & 0xFF;
 	struct mlx5e_priv *priv;
 	struct xfrm_offload *xo;
 	struct xfrm_state *xs;
@@ -481,7 +480,7 @@ void mlx5e_ipsec_offload_handle_rx_skb(struct net_device *netdev,
 	xo = xfrm_offload(skb);
 	xo->flags = CRYPTO_DONE;
 
-	switch (ipsec_syndrome & MLX5_IPSEC_METADATA_SYNDROM_MASK) {
+	switch (MLX5_IPSEC_METADATA_SYNDROM(ipsec_meta_data)) {
 	case MLX5E_IPSEC_OFFLOAD_RX_SYNDROME_DECRYPTED:
 		xo->status = CRYPTO_SUCCESS;
 		if (WARN_ON_ONCE(priv->ipsec->no_trailer))
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.h
index 056dacb612b0..9df9b9a8e09b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.h
@@ -39,9 +39,10 @@
 #include "en.h"
 #include "en/txrx.h"
 
-#define MLX5_IPSEC_METADATA_MARKER_MASK      (0x80)
-#define MLX5_IPSEC_METADATA_SYNDROM_MASK     (0x7F)
-#define MLX5_IPSEC_METADATA_HANDLE(metadata) (((metadata) >> 8) & 0xFF)
+/* Bit31: IPsec marker, Bit30-24: IPsec syndrome, Bit23-0: IPsec obj id */
+#define MLX5_IPSEC_METADATA_MARKER(metadata)  (((metadata) >> 31) & 0x1)
+#define MLX5_IPSEC_METADATA_SYNDROM(metadata) (((metadata) >> 24) & GENMASK(6, 0))
+#define MLX5_IPSEC_METADATA_HANDLE(metadata)  ((metadata) & GENMASK(23, 0))
 
 struct mlx5e_accel_tx_ipsec_state {
 	struct xfrm_offload *xo;
@@ -78,7 +79,7 @@ static inline unsigned int mlx5e_ipsec_tx_ids_len(struct mlx5e_accel_tx_ipsec_st
 
 static inline bool mlx5_ipsec_is_rx_flow(struct mlx5_cqe64 *cqe)
 {
-	return !!(MLX5_IPSEC_METADATA_MARKER_MASK & be32_to_cpu(cqe->ft_metadata));
+	return MLX5_IPSEC_METADATA_MARKER(be32_to_cpu(cqe->ft_metadata));
 }
 
 static inline bool mlx5e_ipsec_is_tx_flow(struct mlx5e_accel_tx_ipsec_state *ipsec_st)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
index 3b979008143d..4a2ce241522e 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
@@ -283,6 +283,9 @@ static inline bool mlx5e_cqe_regb_chain(struct mlx5_cqe64 *cqe)
 
 	reg_b = be32_to_cpu(cqe->ft_metadata);
 
+	if (reg_b >> (MLX5E_TC_TABLE_CHAIN_TAG_BITS + ZONE_RESTORE_BITS))
+		return false;
+
 	chain = reg_b & MLX5E_TC_TABLE_CHAIN_TAG_MASK;
 	if (chain)
 		return true;
-- 
2.26.2

[net 4/9] net/mlx5e: Fix check if netdev is bond slave

[Saeed Mahameed]

From: Maor Dickman <maord@nvidia.com>

Bond events handler uses bond_slave_get_rtnl to check if net device
is bond slave. bond_slave_get_rtnl return the rcu rx_handler pointer
from the netdev which exists for bond slaves but also exists for
devices that are attached to linux bridge so using it as indication
for bond slave is wrong.

Fix by using netif_is_lag_port instead.

Fixes: 7e51891a237f ("net/mlx5e: Use netdev events to set/del egress acl forward-to-vport rule")
Signed-off-by: Maor Dickman <maord@nvidia.com>
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Ariel Levkovich <lariel@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en/rep/bond.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/rep/bond.c b/drivers/net/ethernet/mellanox/mlx5/core/en/rep/bond.c
index 3e44e4d820c5..95f2b26a3ee3 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/rep/bond.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/rep/bond.c
@@ -187,7 +187,7 @@ static bool mlx5e_rep_is_lag_netdev(struct net_device *netdev)
 	struct mlx5e_priv *priv;
 
 	/* A given netdev is not a representor or not a slave of LAG configuration */
-	if (!mlx5e_eswitch_rep(netdev) || !bond_slave_get_rtnl(netdev))
+	if (!mlx5e_eswitch_rep(netdev) || !netif_is_lag_port(netdev))
 		return false;
 
 	priv = netdev_priv(netdev);
-- 
2.26.2

[net 5/9] net/mlx5: Add handling of port type in rule deletion

[Saeed Mahameed]

From: Michael Guralnik <michaelgur@nvidia.com>

Handle destruction of rules with port destination type to enable
full destruction of flow.

Without this handling of TX rules the deletion of these rules fails.
Dmesg of flow destruction failure:

[  203.714146] mlx5_core 0000:00:0b.0: mlx5_cmd_check:753:(pid 342): SET_FLOW_TABLE_ENTRY(0x936) op_mod(0x0) failed, status bad parameter(0x3), syndrome (0x144b7a)
[  210.547387] ------------[ cut here ]------------
[  210.548663] refcount_t: decrement hit 0; leaking memory.
[  210.550651] WARNING: CPU: 4 PID: 342 at lib/refcount.c:31 refcount_warn_saturate+0x5c/0x110
[  210.550654] Modules linked in: mlx5_ib mlx5_core ib_ipoib rdma_ucm rdma_cm iw_cm ib_cm ib_umad ib_uverbs ib_core
[  210.550675] CPU: 4 PID: 342 Comm: test Not tainted 5.8.0-rc2+ #116
[  210.550678] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.12.1-0-ga5cab58e9a3f-prebuilt.qemu.org 04/01/2014
[  210.550680] RIP: 0010:refcount_warn_saturate+0x5c/0x110
[  210.550685] Code: c6 d1 1b 01 00 0f 84 ad 00 00 00 5b 5d c3 80 3d b5 d1 1b 01 00 75 f4 48 c7 c7 20 d1 15 82 c6 05 a5 d1 1b 01 01 e8 a7 eb af ff <0f> 0b eb dd 80 3d 99 d1 1b 01 00 75 d4 48 c7 c7 c0 cf 15 82 c6 05
[  210.550687] RSP: 0018:ffff8881642e77e8 EFLAGS: 00010282
[  210.550691] RAX: 0000000000000000 RBX: 0000000000000004 RCX: 0000000000000000
[  210.550694] RDX: 0000000000000027 RSI: 0000000000000004 RDI: ffffed102c85ceef
[  210.550696] RBP: ffff888161720428 R08: ffffffff8124c10e R09: ffffed103243beae
[  210.550698] R10: ffff8881921df56b R11: ffffed103243bead R12: ffff8881841b4180
[  210.550701] R13: ffff888161720428 R14: ffff8881616d0000 R15: ffff888161720380
[  210.550704] FS:  00007fc27f025740(0000) GS:ffff888192000000(0000) knlGS:0000000000000000
[  210.550706] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  210.550708] CR2: 0000557e4b41a6a0 CR3: 0000000002415004 CR4: 0000000000360ea0
[  210.550711] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[  210.550713] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[  210.550715] Call Trace:
[  210.550717]  mlx5_del_flow_rules+0x484/0x490 [mlx5_core]
[  210.550720]  ? mlx5_cmd_set_fte+0xa80/0xa80 [mlx5_core]
[  210.550722]  mlx5_ib_destroy_flow+0x17f/0x280 [mlx5_ib]
[  210.550724]  uverbs_free_flow+0x4c/0x90 [ib_uverbs]
[  210.550726]  destroy_hw_idr_uobject+0x41/0xb0 [ib_uverbs]
[  210.550728]  uverbs_destroy_uobject+0xaa/0x390 [ib_uverbs]
[  210.550731]  __uverbs_cleanup_ufile+0x129/0x1b0 [ib_uverbs]
[  210.550733]  ? uverbs_destroy_uobject+0x390/0x390 [ib_uverbs]
[  210.550735]  uverbs_destroy_ufile_hw+0x78/0x190 [ib_uverbs]
[  210.550737]  ib_uverbs_close+0x36/0x140 [ib_uverbs]
[  210.550739]  __fput+0x181/0x380
[  210.550741]  task_work_run+0x88/0xd0
[  210.550743]  do_exit+0x5f6/0x13b0
[  210.550745]  ? sched_clock_cpu+0x30/0x140
[  210.550747]  ? is_current_pgrp_orphaned+0x70/0x70
[  210.550750]  ? lock_downgrade+0x360/0x360
[  210.550752]  ? mark_held_locks+0x1d/0x90
[  210.550754]  do_group_exit+0x8a/0x140
[  210.550756]  get_signal+0x20a/0xf50
[  210.550758]  do_signal+0x8c/0xbe0
[  210.550760]  ? hrtimer_nanosleep+0x1d8/0x200
[  210.550762]  ? nanosleep_copyout+0x50/0x50
[  210.550764]  ? restore_sigcontext+0x320/0x320
[  210.550766]  ? __hrtimer_init+0xf0/0xf0
[  210.550768]  ? timespec64_add_safe+0x150/0x150
[  210.550770]  ? mark_held_locks+0x1d/0x90
[  210.550772]  ? lockdep_hardirqs_on_prepare+0x14c/0x240
[  210.550774]  __prepare_exit_to_usermode+0x119/0x170
[  210.550776]  do_syscall_64+0x65/0x300
[  210.550778]  ? trace_hardirqs_off+0x10/0x120
[  210.550781]  ? mark_held_locks+0x1d/0x90
[  210.550783]  ? asm_sysvec_apic_timer_interrupt+0xa/0x20
[  210.550785]  ? lockdep_hardirqs_on+0x112/0x190
[  210.550787]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[  210.550789] RIP: 0033:0x7fc27f1cd157
[  210.550791] Code: Bad RIP value.
[  210.550793] RSP: 002b:00007ffd4db27ea8 EFLAGS: 00000246 ORIG_RAX: 0000000000000023
[  210.550798] RAX: fffffffffffffdfc RBX: ffffffffffffff80 RCX: 00007fc27f1cd157
[  210.550800] RDX: 00007fc27f025740 RSI: 00007ffd4db27eb0 RDI: 00007ffd4db27eb0
[  210.550803] RBP: 0000000000000016 R08: 0000000000000000 R09: 000000000000000e
[  210.550805] R10: 00007ffd4db27dc7 R11: 0000000000000246 R12: 0000000000400c00
[  210.550808] R13: 00007ffd4db285f0 R14: 0000000000000000 R15: 0000000000000000
[  210.550809] irq event stamp: 49399
[  210.550812] hardirqs last  enabled at (49399): [<ffffffff81172d36>] console_unlock+0x556/0x6f0
[  210.550815] hardirqs last disabled at (49398): [<ffffffff81172897>] console_unlock+0xb7/0x6f0
[  210.550818] softirqs last  enabled at (48706): [<ffffffff81e0037b>] __do_softirq+0x37b/0x60c
[  210.550820] softirqs last disabled at (48697): [<ffffffff81c00e2f>] asm_call_on_stack+0xf/0x20
[  210.550822] ---[ end trace ad18c0e6fa846454 ]---
[  210.581862] mlx5_core 0000:00:0c.0: mlx5_destroy_flow_table:2132:(pid 342): Flow table 262150 wasn't destroyed, refcount > 1

Fixes: a7ee18bdee83 ("RDMA/mlx5: Allow creating a matcher for a NIC TX flow table")
Signed-off-by: Michael Guralnik <michaelgur@nvidia.com>
Reviewed-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 | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
index 325a5b0d6829..9fdd99272e31 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
@@ -534,6 +534,13 @@ 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) {
+		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) {
 		fte->modify_mask |=
-- 
2.26.2

[net 6/9] net/mlx5: Clear bw_share upon VF disable

[Saeed Mahameed]

From: Vladyslav Tarasiuk <vladyslavt@nvidia.com>

Currently, if user disables VFs with some min and max rates configured,
they are cleared. But QoS data is not cleared and restored upon next VF
enable placing limits on minimal rate for given VF, when user expects
none.

To match cleared vport->info struct with QoS-related min and max rates
upon VF disable, clear vport->qos struct too.

Fixes: 556b9d16d3f5 ("net/mlx5: Clear VF's configuration on disabling SRIOV")
Signed-off-by: Vladyslav Tarasiuk <vladyslavt@nvidia.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index e8e6294c7cca..6562f4d484e6 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -1408,6 +1408,7 @@ static void mlx5_eswitch_clear_vf_vports_info(struct mlx5_eswitch *esw)
 	int i;
 
 	mlx5_esw_for_each_vf_vport(esw, i, vport, esw->esw_funcs.num_vfs) {
+		memset(&vport->qos, 0, sizeof(vport->qos));
 		memset(&vport->info, 0, sizeof(vport->info));
 		vport->info.link_state = MLX5_VPORT_ADMIN_STATE_AUTO;
 	}
-- 
2.26.2

[net 7/9] net/mlx5: Disable QoS when min_rates on all VFs are zero

[Saeed Mahameed]

From: Vladyslav Tarasiuk <vladyslavt@nvidia.com>

Currently when QoS is enabled for VF and any min_rate is configured,
the driver sets bw_share value to at least 1 and doesn’t allow to set
it to 0 to make minimal rate unlimited. It means there is always a
minimal rate configured for every VF, even if user tries to remove it.

In order to make QoS disable possible, check whether all vports have
configured min_rate = 0. If this is true, set their bw_share to 0 to
disable min_rate limitations.

Fixes: c9497c98901c ("net/mlx5: Add support for setting VF min rate")
Signed-off-by: Vladyslav Tarasiuk <vladyslavt@nvidia.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index 6562f4d484e6..5ad2308a2a6b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -2222,12 +2222,15 @@ static u32 calculate_vports_min_rate_divider(struct mlx5_eswitch *esw)
 		max_guarantee = evport->info.min_rate;
 	}
 
-	return max_t(u32, max_guarantee / fw_max_bw_share, 1);
+	if (max_guarantee)
+		return max_t(u32, max_guarantee / fw_max_bw_share, 1);
+	return 0;
 }
 
-static int normalize_vports_min_rate(struct mlx5_eswitch *esw, u32 divider)
+static int normalize_vports_min_rate(struct mlx5_eswitch *esw)
 {
 	u32 fw_max_bw_share = MLX5_CAP_QOS(esw->dev, max_tsar_bw_share);
+	u32 divider = calculate_vports_min_rate_divider(esw);
 	struct mlx5_vport *evport;
 	u32 vport_max_rate;
 	u32 vport_min_rate;
@@ -2240,9 +2243,9 @@ static int normalize_vports_min_rate(struct mlx5_eswitch *esw, u32 divider)
 			continue;
 		vport_min_rate = evport->info.min_rate;
 		vport_max_rate = evport->info.max_rate;
-		bw_share = MLX5_MIN_BW_SHARE;
+		bw_share = 0;
 
-		if (vport_min_rate)
+		if (divider)
 			bw_share = MLX5_RATE_TO_BW_SHARE(vport_min_rate,
 							 divider,
 							 fw_max_bw_share);
@@ -2267,7 +2270,6 @@ int mlx5_eswitch_set_vport_rate(struct mlx5_eswitch *esw, u16 vport,
 	struct mlx5_vport *evport = mlx5_eswitch_get_vport(esw, vport);
 	u32 fw_max_bw_share;
 	u32 previous_min_rate;
-	u32 divider;
 	bool min_rate_supported;
 	bool max_rate_supported;
 	int err = 0;
@@ -2292,8 +2294,7 @@ int mlx5_eswitch_set_vport_rate(struct mlx5_eswitch *esw, u16 vport,
 
 	previous_min_rate = evport->info.min_rate;
 	evport->info.min_rate = min_rate;
-	divider = calculate_vports_min_rate_divider(esw);
-	err = normalize_vports_min_rate(esw, divider);
+	err = normalize_vports_min_rate(esw);
 	if (err) {
 		evport->info.min_rate = previous_min_rate;
 		goto unlock;
-- 
2.26.2

[net 8/9] net/mlx5: E-Switch, Fail mlx5_esw_modify_vport_rate if qos disabled

[Saeed Mahameed]

From: Eli Cohen <elic@nvidia.com>

Avoid calling mlx5_esw_modify_vport_rate() if qos is not enabled and
avoid unnecessary syndrome messages from firmware.

Fixes: fcb64c0f5640 ("net/mlx5: E-Switch, add ingress rate support")
Signed-off-by: Eli Cohen <elic@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index 5ad2308a2a6b..d4ee0a9c03db 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -1142,6 +1142,10 @@ int mlx5_esw_modify_vport_rate(struct mlx5_eswitch *esw, u16 vport_num,
 	struct mlx5_vport *vport;
 
 	vport = mlx5_eswitch_get_vport(esw, vport_num);
+
+	if (!vport->qos.enabled)
+		return -EOPNOTSUPP;
+
 	MLX5_SET(scheduling_context, ctx, max_average_bw, rate_mbps);
 
 	return mlx5_modify_scheduling_element_cmd(esw->dev,
-- 
2.26.2

[net 9/9] net/mlx5: fix error return code in mlx5e_tc_nic_init()

[Saeed Mahameed]

From: Wang Hai <wanghai38@huawei.com>

Fix to return a negative error code from the error handling
case instead of 0, as done elsewhere in this function.

Fixes: aedd133d17bc ("net/mlx5e: Support CT offload for tc nic flows")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Wang Hai <wanghai38@huawei.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
index 2e2fa0440032..ce710f22b1ff 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -5229,8 +5229,10 @@ int mlx5e_tc_nic_init(struct mlx5e_priv *priv)
 
 	tc->ct = mlx5_tc_ct_init(priv, tc->chains, &priv->fs.tc.mod_hdr,
 				 MLX5_FLOW_NAMESPACE_KERNEL);
-	if (IS_ERR(tc->ct))
+	if (IS_ERR(tc->ct)) {
+		err = PTR_ERR(tc->ct);
 		goto err_ct;
+	}
 
 	tc->netdevice_nb.notifier_call = mlx5e_tc_netdev_event;
 	err = register_netdevice_notifier_dev_net(priv->netdev,
-- 
2.26.2

Re: [pull request][net 0/9] mlx5 fixes 2020-11-17

[Jakub Kicinski]

On Tue, 17 Nov 2020 11:56:53 -0800 Saeed Mahameed wrote:
> This series introduces some fixes to mlx5 driver.
> Please pull and let me know if there is any problem.
> 
> For -stable v4.14
>  ('net/mlx5: Disable QoS when min_rates on all VFs are zero')
> 
> For -stable v4.20
>  ('net/mlx5: Add handling of port type in rule deletion')
> 
> For -stable v5.5
>  ('net/mlx5: Clear bw_share upon VF disable')
> 
> For -stable v5.7
>  ('net/mlx5: E-Switch, Fail mlx5_esw_modify_vport_rate if qos disabled')
> 
> For -stable v5.8
>  ('net/mlx5e: Fix check if netdev is bond slave')
> 
> For -stable v5.9
>  ('net/mlx5e: Fix refcount leak on kTLS RX resync')

Pulled, thanks!