[PATCH AUTOSEL 4.19 054/205] IB/rxe: avoid back-to-back retries

[PATCH AUTOSEL 4.19 054/205] IB/rxe: avoid back-to-back retries

[Sasha Levin]

From: Vijay Immanuel <vijayi@attalasystems.com>

[ Upstream commit 4e4c53df567714b3d08b2b5d8ccb1d175fc9be01 ]

Error retries can occur due to timeouts, NAKs or receiving
packets beyond the current read request. Avoid back-to-back
retries due to packet processing, by only retrying the initial
attempt immediately. Subsequent retries must be due to timeouts.

Continue to process completion packets after scheduling a retry.

Signed-off-by: Vijay Immanuel <vijayi@attalasystems.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/sw/rxe/rxe_comp.c  | 18 +++++++++++++++++-
 drivers/infiniband/sw/rxe/rxe_verbs.h |  1 +
 2 files changed, 18 insertions(+), 1 deletion(-)

diff --git a/drivers/infiniband/sw/rxe/rxe_comp.c b/drivers/infiniband/sw/rxe/rxe_comp.c
index 83311dd07019b..ed96441595d81 100644
--- a/drivers/infiniband/sw/rxe/rxe_comp.c
+++ b/drivers/infiniband/sw/rxe/rxe_comp.c
@@ -191,6 +191,7 @@ static inline void reset_retry_counters(struct rxe_qp *qp)
 {
 	qp->comp.retry_cnt = qp->attr.retry_cnt;
 	qp->comp.rnr_retry = qp->attr.rnr_retry;
+	qp->comp.started_retry = 0;
 }
 
 static inline enum comp_state check_psn(struct rxe_qp *qp,
@@ -676,6 +677,20 @@ int rxe_completer(void *arg)
 				goto exit;
 			}
 
+			/* if we've started a retry, don't start another
+			 * retry sequence, unless this is a timeout.
+			 */
+			if (qp->comp.started_retry &&
+			    !qp->comp.timeout_retry) {
+				if (pkt) {
+					rxe_drop_ref(pkt->qp);
+					kfree_skb(skb);
+					skb = NULL;
+				}
+
+				goto done;
+			}
+
 			if (qp->comp.retry_cnt > 0) {
 				if (qp->comp.retry_cnt != 7)
 					qp->comp.retry_cnt--;
@@ -692,6 +707,7 @@ int rxe_completer(void *arg)
 					rxe_counter_inc(rxe,
 							RXE_CNT_COMP_RETRY);
 					qp->req.need_retry = 1;
+					qp->comp.started_retry = 1;
 					rxe_run_task(&qp->req.task, 1);
 				}
 
@@ -701,7 +717,7 @@ int rxe_completer(void *arg)
 					skb = NULL;
 				}
 
-				goto exit;
+				goto done;
 
 			} else {
 				rxe_counter_inc(rxe, RXE_CNT_RETRY_EXCEEDED);
diff --git a/drivers/infiniband/sw/rxe/rxe_verbs.h b/drivers/infiniband/sw/rxe/rxe_verbs.h
index 3b731c7682e5b..a0ec28d2b71a4 100644
--- a/drivers/infiniband/sw/rxe/rxe_verbs.h
+++ b/drivers/infiniband/sw/rxe/rxe_verbs.h
@@ -158,6 +158,7 @@ struct rxe_comp_info {
 	int			opcode;
 	int			timeout;
 	int			timeout_retry;
+	int			started_retry;
 	u32			retry_cnt;
 	u32			rnr_retry;
 	struct rxe_task		task;
-- 
2.20.1

[PATCH AUTOSEL 4.19 055/205] IB/rxe: fixes for rdma read retry

[Sasha Levin]

From: Vijay Immanuel <vijayi@attalasystems.com>

[ Upstream commit 030e46e495af855a13964a0aab9753ea82a96edc ]

When a read request is retried for the remaining partial
data, the response may restart from read response first
or read response only. So support those cases.

Do not advance the comp psn beyond the current wqe's last_psn
as that could skip over an entire read wqe and will cause the
req_retry() logic to set an incorrect req psn.
An example sequence is as follows:
Write        PSN 40 -- this is the current WQE.
Read request PSN 41
Write        PSN 42
Receive ACK  PSN 42 -- this will complete the current WQE
for PSN 40, and set the comp psn to 42 which is a problem
because the read request at PSN 41 has been skipped over.
So when req_retry() tries to retransmit the read request,
it sets the req psn to 42 which is incorrect.

When retrying a read request, calculate the number of psns
completed based on the dma resid instead of the wqe first_psn.
The wqe first_psn could have moved if the read request was
retried multiple times.

Set the reth length to the dma resid to handle read retries for
the remaining partial data.

Signed-off-by: Vijay Immanuel <vijayi@attalasystems.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/sw/rxe/rxe_comp.c | 21 ++++++++++++++++-----
 drivers/infiniband/sw/rxe/rxe_req.c  | 15 +++++++++------
 2 files changed, 25 insertions(+), 11 deletions(-)

diff --git a/drivers/infiniband/sw/rxe/rxe_comp.c b/drivers/infiniband/sw/rxe/rxe_comp.c
index ed96441595d81..ea089cb091ade 100644
--- a/drivers/infiniband/sw/rxe/rxe_comp.c
+++ b/drivers/infiniband/sw/rxe/rxe_comp.c
@@ -254,6 +254,17 @@ static inline enum comp_state check_ack(struct rxe_qp *qp,
 	case IB_OPCODE_RC_RDMA_READ_RESPONSE_MIDDLE:
 		if (pkt->opcode != IB_OPCODE_RC_RDMA_READ_RESPONSE_MIDDLE &&
 		    pkt->opcode != IB_OPCODE_RC_RDMA_READ_RESPONSE_LAST) {
+			/* read retries of partial data may restart from
+			 * read response first or response only.
+			 */
+			if ((pkt->psn == wqe->first_psn &&
+			     pkt->opcode ==
+			     IB_OPCODE_RC_RDMA_READ_RESPONSE_FIRST) ||
+			    (wqe->first_psn == wqe->last_psn &&
+			     pkt->opcode ==
+			     IB_OPCODE_RC_RDMA_READ_RESPONSE_ONLY))
+				break;
+
 			return COMPST_ERROR;
 		}
 		break;
@@ -500,11 +511,11 @@ static inline enum comp_state complete_wqe(struct rxe_qp *qp,
 					   struct rxe_pkt_info *pkt,
 					   struct rxe_send_wqe *wqe)
 {
-	qp->comp.opcode = -1;
-
-	if (pkt) {
-		if (psn_compare(pkt->psn, qp->comp.psn) >= 0)
-			qp->comp.psn = (pkt->psn + 1) & BTH_PSN_MASK;
+	if (pkt && wqe->state == wqe_state_pending) {
+		if (psn_compare(wqe->last_psn, qp->comp.psn) >= 0) {
+			qp->comp.psn = (wqe->last_psn + 1) & BTH_PSN_MASK;
+			qp->comp.opcode = -1;
+		}
 
 		if (qp->req.wait_psn) {
 			qp->req.wait_psn = 0;
diff --git a/drivers/infiniband/sw/rxe/rxe_req.c b/drivers/infiniband/sw/rxe/rxe_req.c
index fa98a52796470..f7dd8de799415 100644
--- a/drivers/infiniband/sw/rxe/rxe_req.c
+++ b/drivers/infiniband/sw/rxe/rxe_req.c
@@ -73,9 +73,6 @@ static void req_retry(struct rxe_qp *qp)
 	int npsn;
 	int first = 1;
 
-	wqe = queue_head(qp->sq.queue);
-	npsn = (qp->comp.psn - wqe->first_psn) & BTH_PSN_MASK;
-
 	qp->req.wqe_index	= consumer_index(qp->sq.queue);
 	qp->req.psn		= qp->comp.psn;
 	qp->req.opcode		= -1;
@@ -107,11 +104,17 @@ static void req_retry(struct rxe_qp *qp)
 		if (first) {
 			first = 0;
 
-			if (mask & WR_WRITE_OR_SEND_MASK)
+			if (mask & WR_WRITE_OR_SEND_MASK) {
+				npsn = (qp->comp.psn - wqe->first_psn) &
+					BTH_PSN_MASK;
 				retry_first_write_send(qp, wqe, mask, npsn);
+			}
 
-			if (mask & WR_READ_MASK)
+			if (mask & WR_READ_MASK) {
+				npsn = (wqe->dma.length - wqe->dma.resid) /
+					qp->mtu;
 				wqe->iova += npsn * qp->mtu;
+			}
 		}
 
 		wqe->state = wqe_state_posted;
@@ -435,7 +438,7 @@ static struct sk_buff *init_req_packet(struct rxe_qp *qp,
 	if (pkt->mask & RXE_RETH_MASK) {
 		reth_set_rkey(pkt, ibwr->wr.rdma.rkey);
 		reth_set_va(pkt, wqe->iova);
-		reth_set_len(pkt, wqe->dma.length);
+		reth_set_len(pkt, wqe->dma.resid);
 	}
 
 	if (pkt->mask & RXE_IMMDT_MASK)
-- 
2.20.1

[PATCH AUTOSEL 4.19 075/205] net/mlx5: Fix atomic_mode enum values

[Sasha Levin]

From: Moni Shoua <monis@mellanox.com>

[ Upstream commit aa7e80b220f3a543eefbe4b7e2c5d2b73e2e2ef7 ]

The field atomic_mode is 4 bits wide and therefore can hold values
from 0x0 to 0xf. Remove the unnecessary 20 bit shift that made the values
be incorrect. While that, remove unused enum values.

Fixes: 57cda166bbe0 ("net/mlx5: Add DCT command interface")
Signed-off-by: Moni Shoua <monis@mellanox.com>
Reviewed-by: Artemy Kovalyov <artemyko@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 include/linux/mlx5/driver.h | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h
index e8b92dee5a726..ae64fced188d1 100644
--- a/include/linux/mlx5/driver.h
+++ b/include/linux/mlx5/driver.h
@@ -163,10 +163,7 @@ enum mlx5_dcbx_oper_mode {
 };
 
 enum mlx5_dct_atomic_mode {
-	MLX5_ATOMIC_MODE_DCT_OFF        = 20,
-	MLX5_ATOMIC_MODE_DCT_NONE       = 0 << MLX5_ATOMIC_MODE_DCT_OFF,
-	MLX5_ATOMIC_MODE_DCT_IB_COMP    = 1 << MLX5_ATOMIC_MODE_DCT_OFF,
-	MLX5_ATOMIC_MODE_DCT_CX         = 2 << MLX5_ATOMIC_MODE_DCT_OFF,
+	MLX5_ATOMIC_MODE_DCT_CX         = 2,
 };
 
 enum {
-- 
2.20.1

[PATCH AUTOSEL 4.19 084/205] IB/mlx5: Change TX affinity assignment in RoCE LAG mode

[Sasha Levin]

From: Majd Dibbiny <majd@mellanox.com>

[ Upstream commit c6a21c3864fc7f5febae7d096cd136f397c791f2 ]

In the current code, the TX affinity is per RoCE device, which can cause
unfairness between different contexts. e.g. if we open two contexts, and
each open 10 QPs concurrently, all of the QPs of the first context might
end up on the first port instead of distributed on the two ports as
expected

To overcome this unfairness between processes, we maintain per device TX
affinity, and per process TX affinity.

The allocation algorithm is as follow:

1. Hold two tx_port_affinity atomic variables, one per RoCE device and one
   per ucontext. Both initialized to 0.

2. In mlx5_ib_alloc_ucontext do:
 2.1. ucontext.tx_port_affinity = device.tx_port_affinity
 2.2. device.tx_port_affinity += 1

3. In modify QP INIT2RST:
 3.1. qp.tx_port_affinity = ucontext.tx_port_affinity % MLX5_PORT_NUM
 3.2. ucontext.tx_port_affinity += 1

Signed-off-by: Majd Dibbiny <majd@mellanox.com>
Reviewed-by: Moni Shoua <monis@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/hw/mlx5/main.c    |  8 ++++++
 drivers/infiniband/hw/mlx5/mlx5_ib.h |  4 ++-
 drivers/infiniband/hw/mlx5/qp.c      | 37 +++++++++++++++++++++++++---
 3 files changed, 44 insertions(+), 5 deletions(-)

diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c
index c05eae93170eb..f4ffdc588ea07 100644
--- a/drivers/infiniband/hw/mlx5/main.c
+++ b/drivers/infiniband/hw/mlx5/main.c
@@ -1823,6 +1823,14 @@ static struct ib_ucontext *mlx5_ib_alloc_ucontext(struct ib_device *ibdev,
 	context->lib_caps = req.lib_caps;
 	print_lib_caps(dev, context->lib_caps);
 
+	if (mlx5_lag_is_active(dev->mdev)) {
+		u8 port = mlx5_core_native_port_num(dev->mdev);
+
+		atomic_set(&context->tx_port_affinity,
+			   atomic_add_return(
+				   1, &dev->roce[port].tx_port_affinity));
+	}
+
 	return &context->ibucontext;
 
 out_mdev:
diff --git a/drivers/infiniband/hw/mlx5/mlx5_ib.h b/drivers/infiniband/hw/mlx5/mlx5_ib.h
index 941d1df54631a..6a060c84598fe 100644
--- a/drivers/infiniband/hw/mlx5/mlx5_ib.h
+++ b/drivers/infiniband/hw/mlx5/mlx5_ib.h
@@ -139,6 +139,8 @@ struct mlx5_ib_ucontext {
 	u64			lib_caps;
 	DECLARE_BITMAP(dm_pages, MLX5_MAX_MEMIC_PAGES);
 	u16			devx_uid;
+	/* For RoCE LAG TX affinity */
+	atomic_t		tx_port_affinity;
 };
 
 static inline struct mlx5_ib_ucontext *to_mucontext(struct ib_ucontext *ibucontext)
@@ -700,7 +702,7 @@ struct mlx5_roce {
 	rwlock_t		netdev_lock;
 	struct net_device	*netdev;
 	struct notifier_block	nb;
-	atomic_t		next_port;
+	atomic_t		tx_port_affinity;
 	enum ib_port_state last_port_state;
 	struct mlx5_ib_dev	*dev;
 	u8			native_port_num;
diff --git a/drivers/infiniband/hw/mlx5/qp.c b/drivers/infiniband/hw/mlx5/qp.c
index 77b1f3fd086ad..69669c3b13d85 100644
--- a/drivers/infiniband/hw/mlx5/qp.c
+++ b/drivers/infiniband/hw/mlx5/qp.c
@@ -2908,6 +2908,37 @@ static int modify_raw_packet_qp(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
 	return 0;
 }
 
+static unsigned int get_tx_affinity(struct mlx5_ib_dev *dev,
+				    struct mlx5_ib_pd *pd,
+				    struct mlx5_ib_qp_base *qp_base,
+				    u8 port_num)
+{
+	struct mlx5_ib_ucontext *ucontext = NULL;
+	unsigned int tx_port_affinity;
+
+	if (pd && pd->ibpd.uobject && pd->ibpd.uobject->context)
+		ucontext = to_mucontext(pd->ibpd.uobject->context);
+
+	if (ucontext) {
+		tx_port_affinity = (unsigned int)atomic_add_return(
+					   1, &ucontext->tx_port_affinity) %
+					   MLX5_MAX_PORTS +
+				   1;
+		mlx5_ib_dbg(dev, "Set tx affinity 0x%x to qpn 0x%x ucontext %p\n",
+				tx_port_affinity, qp_base->mqp.qpn, ucontext);
+	} else {
+		tx_port_affinity =
+			(unsigned int)atomic_add_return(
+				1, &dev->roce[port_num].tx_port_affinity) %
+				MLX5_MAX_PORTS +
+			1;
+		mlx5_ib_dbg(dev, "Set tx affinity 0x%x to qpn 0x%x\n",
+				tx_port_affinity, qp_base->mqp.qpn);
+	}
+
+	return tx_port_affinity;
+}
+
 static int __mlx5_ib_modify_qp(struct ib_qp *ibqp,
 			       const struct ib_qp_attr *attr, int attr_mask,
 			       enum ib_qp_state cur_state, enum ib_qp_state new_state,
@@ -2973,6 +3004,7 @@ static int __mlx5_ib_modify_qp(struct ib_qp *ibqp,
 	if (!context)
 		return -ENOMEM;
 
+	pd = get_pd(qp);
 	context->flags = cpu_to_be32(mlx5_st << 16);
 
 	if (!(attr_mask & IB_QP_PATH_MIG_STATE)) {
@@ -3001,9 +3033,7 @@ static int __mlx5_ib_modify_qp(struct ib_qp *ibqp,
 		    (ibqp->qp_type == IB_QPT_XRC_TGT)) {
 			if (mlx5_lag_is_active(dev->mdev)) {
 				u8 p = mlx5_core_native_port_num(dev->mdev);
-				tx_affinity = (unsigned int)atomic_add_return(1,
-						&dev->roce[p].next_port) %
-						MLX5_MAX_PORTS + 1;
+				tx_affinity = get_tx_affinity(dev, pd, base, p);
 				context->flags |= cpu_to_be32(tx_affinity << 24);
 			}
 		}
@@ -3061,7 +3091,6 @@ static int __mlx5_ib_modify_qp(struct ib_qp *ibqp,
 			goto out;
 	}
 
-	pd = get_pd(qp);
 	get_cqs(qp->ibqp.qp_type, qp->ibqp.send_cq, qp->ibqp.recv_cq,
 		&send_cq, &recv_cq);
 
-- 
2.20.1

[PATCH AUTOSEL 4.19 096/205] IB/mlx5: Don't hold spin lock while checking device state

[Sasha Levin]

From: Parav Pandit <parav@mellanox.com>

[ Upstream commit 6c75520f7e5a6a353f3b332509d205e213d05855 ]

mdev->state device state is not protected by the QP for which WRs are
being processed. Therefore, there is no need to hold spin lock while
checking mdev state.

Given that device fatal error is unlikely situation, wrap the condition
check with unlikely().

Additionally, kernel QP1 is also a kernel ULP for which soft CQEs needs
to be generated. Therefore, check for device fatal error before
processing QP1 work requests.

Fixes: 89ea94a7b6c4 ("IB/mlx5: Reset flow support for IB kernel ULPs")
Signed-off-by: Parav Pandit <parav@mellanox.com>
Reviewed-by: Daniel Jurgens <danielj@mellanox.com>
Reviewed-by: Maor Gottlieb <maorg@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/hw/mlx5/qp.c | 26 ++++++++++++--------------
 1 file changed, 12 insertions(+), 14 deletions(-)

diff --git a/drivers/infiniband/hw/mlx5/qp.c b/drivers/infiniband/hw/mlx5/qp.c
index 69669c3b13d85..3f96f0e4db796 100644
--- a/drivers/infiniband/hw/mlx5/qp.c
+++ b/drivers/infiniband/hw/mlx5/qp.c
@@ -4405,6 +4405,12 @@ static int _mlx5_ib_post_send(struct ib_qp *ibqp, const struct ib_send_wr *wr,
 	u8 next_fence = 0;
 	u8 fence;
 
+	if (unlikely(mdev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR &&
+		     !drain)) {
+		*bad_wr = wr;
+		return -EIO;
+	}
+
 	if (unlikely(ibqp->qp_type == IB_QPT_GSI))
 		return mlx5_ib_gsi_post_send(ibqp, wr, bad_wr);
 
@@ -4414,13 +4420,6 @@ static int _mlx5_ib_post_send(struct ib_qp *ibqp, const struct ib_send_wr *wr,
 
 	spin_lock_irqsave(&qp->sq.lock, flags);
 
-	if (mdev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR && !drain) {
-		err = -EIO;
-		*bad_wr = wr;
-		nreq = 0;
-		goto out;
-	}
-
 	for (nreq = 0; wr; nreq++, wr = wr->next) {
 		if (unlikely(wr->opcode >= ARRAY_SIZE(mlx5_ib_opcode))) {
 			mlx5_ib_warn(dev, "\n");
@@ -4735,18 +4734,17 @@ static int _mlx5_ib_post_recv(struct ib_qp *ibqp, const struct ib_recv_wr *wr,
 	int ind;
 	int i;
 
+	if (unlikely(mdev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR &&
+		     !drain)) {
+		*bad_wr = wr;
+		return -EIO;
+	}
+
 	if (unlikely(ibqp->qp_type == IB_QPT_GSI))
 		return mlx5_ib_gsi_post_recv(ibqp, wr, bad_wr);
 
 	spin_lock_irqsave(&qp->rq.lock, flags);
 
-	if (mdev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR && !drain) {
-		err = -EIO;
-		*bad_wr = wr;
-		nreq = 0;
-		goto out;
-	}
-
 	ind = qp->rq.head & (qp->rq.wqe_cnt - 1);
 
 	for (nreq = 0; wr; nreq++, wr = wr->next) {
-- 
2.20.1

[PATCH AUTOSEL 4.19 097/205] IB/ipoib: Ensure that MTU isn't less than minimum permitted

[Sasha Levin]

From: Muhammad Sammar <muhammads@mellanox.com>

[ Upstream commit 142a9c287613560edf5a03c8d142c8b6ebc1995b ]

It is illegal to change MTU to a value lower than the minimum MTU
stated in ethernet spec. In addition to that we need to add 4 bytes
for encapsulation header (IPOIB_ENCAP_LEN).

Before "ifconfig ib0 mtu 0" command, succeeds while it obviously shouldn't.

Signed-off-by: Muhammad Sammar <muhammads@mellanox.com>
Reviewed-by: Feras Daoud <ferasda@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/ulp/ipoib/ipoib_main.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c
index 78dd36daac00e..d8cb5bbe6eb58 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_main.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c
@@ -243,7 +243,8 @@ static int ipoib_change_mtu(struct net_device *dev, int new_mtu)
 		return 0;
 	}
 
-	if (new_mtu > IPOIB_UD_MTU(priv->max_ib_mtu))
+	if (new_mtu < (ETH_MIN_MTU + IPOIB_ENCAP_LEN) ||
+	    new_mtu > IPOIB_UD_MTU(priv->max_ib_mtu))
 		return -EINVAL;
 
 	priv->admin_mtu = new_mtu;
-- 
2.20.1

[PATCH AUTOSEL 4.19 098/205] RDMA/core: Rate limit MAD error messages

[Sasha Levin]

From: Parav Pandit <parav@mellanox.com>

[ Upstream commit f9d08f1e1939ad4d92e38bd3dee6842512f5bee6 ]

While registering a mad agent, a user space can trigger various errors
and flood the logs.

Therefore, decrease verbosity and rate limit such error messages.
While we are at it, use __func__ to print function name.

Signed-off-by: Parav Pandit <parav@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/core/mad.c | 72 ++++++++++++++++++-----------------
 1 file changed, 37 insertions(+), 35 deletions(-)

diff --git a/drivers/infiniband/core/mad.c b/drivers/infiniband/core/mad.c
index 74aa3e651bc3c..218411282069b 100644
--- a/drivers/infiniband/core/mad.c
+++ b/drivers/infiniband/core/mad.c
@@ -223,30 +223,30 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
 	/* Validate parameters */
 	qpn = get_spl_qp_index(qp_type);
 	if (qpn == -1) {
-		dev_notice(&device->dev,
-			   "ib_register_mad_agent: invalid QP Type %d\n",
-			   qp_type);
+		dev_dbg_ratelimited(&device->dev, "%s: invalid QP Type %d\n",
+				    __func__, qp_type);
 		goto error1;
 	}
 
 	if (rmpp_version && rmpp_version != IB_MGMT_RMPP_VERSION) {
-		dev_notice(&device->dev,
-			   "ib_register_mad_agent: invalid RMPP Version %u\n",
-			   rmpp_version);
+		dev_dbg_ratelimited(&device->dev,
+				    "%s: invalid RMPP Version %u\n",
+				    __func__, rmpp_version);
 		goto error1;
 	}
 
 	/* Validate MAD registration request if supplied */
 	if (mad_reg_req) {
 		if (mad_reg_req->mgmt_class_version >= MAX_MGMT_VERSION) {
-			dev_notice(&device->dev,
-				   "ib_register_mad_agent: invalid Class Version %u\n",
-				   mad_reg_req->mgmt_class_version);
+			dev_dbg_ratelimited(&device->dev,
+					    "%s: invalid Class Version %u\n",
+					    __func__,
+					    mad_reg_req->mgmt_class_version);
 			goto error1;
 		}
 		if (!recv_handler) {
-			dev_notice(&device->dev,
-				   "ib_register_mad_agent: no recv_handler\n");
+			dev_dbg_ratelimited(&device->dev,
+					    "%s: no recv_handler\n", __func__);
 			goto error1;
 		}
 		if (mad_reg_req->mgmt_class >= MAX_MGMT_CLASS) {
@@ -256,9 +256,9 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
 			 */
 			if (mad_reg_req->mgmt_class !=
 			    IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE) {
-				dev_notice(&device->dev,
-					   "ib_register_mad_agent: Invalid Mgmt Class 0x%x\n",
-					   mad_reg_req->mgmt_class);
+				dev_dbg_ratelimited(&device->dev,
+					"%s: Invalid Mgmt Class 0x%x\n",
+					__func__, mad_reg_req->mgmt_class);
 				goto error1;
 			}
 		} else if (mad_reg_req->mgmt_class == 0) {
@@ -266,8 +266,9 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
 			 * Class 0 is reserved in IBA and is used for
 			 * aliasing of IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE
 			 */
-			dev_notice(&device->dev,
-				   "ib_register_mad_agent: Invalid Mgmt Class 0\n");
+			dev_dbg_ratelimited(&device->dev,
+					    "%s: Invalid Mgmt Class 0\n",
+					    __func__);
 			goto error1;
 		} else if (is_vendor_class(mad_reg_req->mgmt_class)) {
 			/*
@@ -275,18 +276,19 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
 			 * ensure supplied OUI is not zero
 			 */
 			if (!is_vendor_oui(mad_reg_req->oui)) {
-				dev_notice(&device->dev,
-					   "ib_register_mad_agent: No OUI specified for class 0x%x\n",
-					   mad_reg_req->mgmt_class);
+				dev_dbg_ratelimited(&device->dev,
+					"%s: No OUI specified for class 0x%x\n",
+					__func__,
+					mad_reg_req->mgmt_class);
 				goto error1;
 			}
 		}
 		/* Make sure class supplied is consistent with RMPP */
 		if (!ib_is_mad_class_rmpp(mad_reg_req->mgmt_class)) {
 			if (rmpp_version) {
-				dev_notice(&device->dev,
-					   "ib_register_mad_agent: RMPP version for non-RMPP class 0x%x\n",
-					   mad_reg_req->mgmt_class);
+				dev_dbg_ratelimited(&device->dev,
+					"%s: RMPP version for non-RMPP class 0x%x\n",
+					__func__, mad_reg_req->mgmt_class);
 				goto error1;
 			}
 		}
@@ -297,9 +299,9 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
 					IB_MGMT_CLASS_SUBN_LID_ROUTED) &&
 			    (mad_reg_req->mgmt_class !=
 					IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE)) {
-				dev_notice(&device->dev,
-					   "ib_register_mad_agent: Invalid SM QP type: class 0x%x\n",
-					   mad_reg_req->mgmt_class);
+				dev_dbg_ratelimited(&device->dev,
+					"%s: Invalid SM QP type: class 0x%x\n",
+					__func__, mad_reg_req->mgmt_class);
 				goto error1;
 			}
 		} else {
@@ -307,9 +309,9 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
 					IB_MGMT_CLASS_SUBN_LID_ROUTED) ||
 			    (mad_reg_req->mgmt_class ==
 					IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE)) {
-				dev_notice(&device->dev,
-					   "ib_register_mad_agent: Invalid GS QP type: class 0x%x\n",
-					   mad_reg_req->mgmt_class);
+				dev_dbg_ratelimited(&device->dev,
+					"%s: Invalid GS QP type: class 0x%x\n",
+					__func__, mad_reg_req->mgmt_class);
 				goto error1;
 			}
 		}
@@ -324,18 +326,18 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
 	/* Validate device and port */
 	port_priv = ib_get_mad_port(device, port_num);
 	if (!port_priv) {
-		dev_notice(&device->dev,
-			   "ib_register_mad_agent: Invalid port %d\n",
-			   port_num);
+		dev_dbg_ratelimited(&device->dev, "%s: Invalid port %d\n",
+				    __func__, port_num);
 		ret = ERR_PTR(-ENODEV);
 		goto error1;
 	}
 
-	/* Verify the QP requested is supported.  For example, Ethernet devices
-	 * will not have QP0 */
+	/* Verify the QP requested is supported. For example, Ethernet devices
+	 * will not have QP0.
+	 */
 	if (!port_priv->qp_info[qpn].qp) {
-		dev_notice(&device->dev,
-			   "ib_register_mad_agent: QP %d not supported\n", qpn);
+		dev_dbg_ratelimited(&device->dev, "%s: QP %d not supported\n",
+				    __func__, qpn);
 		ret = ERR_PTR(-EPROTONOSUPPORT);
 		goto error1;
 	}
-- 
2.20.1

[PATCH AUTOSEL 4.19 099/205] RDMA/core: Follow correct unregister order between sysfs and cgroup

[Sasha Levin]

From: Parav Pandit <parav@mellanox.com>

[ Upstream commit c715a39541bb399eb03d728a996b224d90ce1336 ]

During register_device() init sequence is,
(a) register with rdma cgroup followed by
(b) register with sysfs

Therefore, unregister_device() sequence should follow the reverse order.

Signed-off-by: Parav Pandit <parav@mellanox.com>
Reviewed-by: Daniel Jurgens <danielj@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/core/device.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/infiniband/core/device.c b/drivers/infiniband/core/device.c
index 6d8ac51a39cc0..6a585c3e21923 100644
--- a/drivers/infiniband/core/device.c
+++ b/drivers/infiniband/core/device.c
@@ -599,8 +599,8 @@ void ib_unregister_device(struct ib_device *device)
 	}
 	up_read(&lists_rwsem);
 
-	ib_device_unregister_rdmacg(device);
 	ib_device_unregister_sysfs(device);
+	ib_device_unregister_rdmacg(device);
 
 	mutex_unlock(&device_mutex);
 
-- 
2.20.1

[PATCH AUTOSEL 4.19 129/205] RDMA/hns: Fix an error code in hns_roce_v2_init_eq_table()

[Sasha Levin]

From: Dan Carpenter <dan.carpenter@oracle.com>

[ Upstream commit f1a315420e79fe5c077fa119db9439ffabd2cda2 ]

The error code isn't set on this path.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
index a442b29e76119..72cfd8c56527e 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
@@ -5117,6 +5117,7 @@ static int hns_roce_v2_init_eq_table(struct hns_roce_dev *hr_dev)
 		create_singlethread_workqueue("hns_roce_irq_workqueue");
 	if (!hr_dev->irq_workq) {
 		dev_err(dev, "Create irq workqueue failed!\n");
+		ret = -ENOMEM;
 		goto err_request_irq_fail;
 	}
 
-- 
2.20.1

[PATCH AUTOSEL 4.19 130/205] IB/hfi1: Missing return value in error path for user sdma

[Sasha Levin]

From: "Michael J. Ruhl" <michael.j.ruhl@intel.com>

[ Upstream commit 2bf4b33f83dfe521c4c7c407b6b150aeec04d69c ]

If the set_txreq_header_agh() function returns an error, the exit path
is chosen.

In this path, the code fails to set the return value.  This will cause
the caller to not realize an error has occurred.

Set the return value correctly in the error path.

Signed-off-by: Michael J. Ruhl <michael.j.ruhl@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/hw/hfi1/user_sdma.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/infiniband/hw/hfi1/user_sdma.c b/drivers/infiniband/hw/hfi1/user_sdma.c
index cbff746d9e9de..684a298e15037 100644
--- a/drivers/infiniband/hw/hfi1/user_sdma.c
+++ b/drivers/infiniband/hw/hfi1/user_sdma.c
@@ -856,8 +856,10 @@ static int user_sdma_send_pkts(struct user_sdma_request *req, unsigned maxpkts)
 
 				changes = set_txreq_header_ahg(req, tx,
 							       datalen);
-				if (changes < 0)
+				if (changes < 0) {
+					ret = changes;
 					goto free_tx;
+				}
 			}
 		} else {
 			ret = sdma_txinit(&tx->txreq, 0, sizeof(req->hdr) +
-- 
2.20.1