[PATCH AUTOSEL 4.19 327/671] IB/mlx5: Add missing XRC options to QP optional params mask

[PATCH AUTOSEL 4.19 327/671] IB/mlx5: Add missing XRC options to QP optional params mask

[Sasha Levin]

From: Jack Morgenstein <jackm@dev.mellanox.co.il>

[ Upstream commit 8f4426aa19fcdb9326ac44154a117b1a3a5ae126 ]

The QP transition optional parameters for the various transition for XRC
QPs are identical to those for RC QPs.

Many of the XRC QP transition optional parameter bits are missing from the
QP optional mask table.  These omissions caused failures when doing XRC QP
state transitions.

For example, when trying to change the response timer of an XRC receive QP
via the RTS2RTS transition, the new timer value was ignored because
MLX5_QP_OPTPAR_RNR_TIMEOUT bit was missing from the optional params mask
for XRC qps for the RTS2RTS transition.

Fix this by adding the missing XRC optional parameters for all QP
transitions to the opt_mask table.

Fixes: e126ba97dba9 ("mlx5: Add driver for Mellanox Connect-IB adapters")
Fixes: a4774e9095de ("IB/mlx5: Fix opt param mask according to firmware spec")
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
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 | 21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)

diff --git a/drivers/infiniband/hw/mlx5/qp.c b/drivers/infiniband/hw/mlx5/qp.c
index ef0f710587ad..4c0f0ce02d2f 100644
--- a/drivers/infiniband/hw/mlx5/qp.c
+++ b/drivers/infiniband/hw/mlx5/qp.c
@@ -2598,6 +2598,11 @@ static enum mlx5_qp_optpar opt_mask[MLX5_QP_NUM_STATE][MLX5_QP_NUM_STATE][MLX5_Q
 			[MLX5_QP_ST_UD] = MLX5_QP_OPTPAR_PKEY_INDEX	|
 					  MLX5_QP_OPTPAR_Q_KEY		|
 					  MLX5_QP_OPTPAR_PRI_PORT,
+			[MLX5_QP_ST_XRC] = MLX5_QP_OPTPAR_RRE		|
+					  MLX5_QP_OPTPAR_RAE		|
+					  MLX5_QP_OPTPAR_RWE		|
+					  MLX5_QP_OPTPAR_PKEY_INDEX	|
+					  MLX5_QP_OPTPAR_PRI_PORT,
 		},
 		[MLX5_QP_STATE_RTR] = {
 			[MLX5_QP_ST_RC] = MLX5_QP_OPTPAR_ALT_ADDR_PATH  |
@@ -2631,6 +2636,12 @@ static enum mlx5_qp_optpar opt_mask[MLX5_QP_NUM_STATE][MLX5_QP_NUM_STATE][MLX5_Q
 					  MLX5_QP_OPTPAR_RWE		|
 					  MLX5_QP_OPTPAR_PM_STATE,
 			[MLX5_QP_ST_UD] = MLX5_QP_OPTPAR_Q_KEY,
+			[MLX5_QP_ST_XRC] = MLX5_QP_OPTPAR_ALT_ADDR_PATH	|
+					  MLX5_QP_OPTPAR_RRE		|
+					  MLX5_QP_OPTPAR_RAE		|
+					  MLX5_QP_OPTPAR_RWE		|
+					  MLX5_QP_OPTPAR_PM_STATE	|
+					  MLX5_QP_OPTPAR_RNR_TIMEOUT,
 		},
 	},
 	[MLX5_QP_STATE_RTS] = {
@@ -2647,6 +2658,12 @@ static enum mlx5_qp_optpar opt_mask[MLX5_QP_NUM_STATE][MLX5_QP_NUM_STATE][MLX5_Q
 			[MLX5_QP_ST_UD] = MLX5_QP_OPTPAR_Q_KEY		|
 					  MLX5_QP_OPTPAR_SRQN		|
 					  MLX5_QP_OPTPAR_CQN_RCV,
+			[MLX5_QP_ST_XRC] = MLX5_QP_OPTPAR_RRE		|
+					  MLX5_QP_OPTPAR_RAE		|
+					  MLX5_QP_OPTPAR_RWE		|
+					  MLX5_QP_OPTPAR_RNR_TIMEOUT	|
+					  MLX5_QP_OPTPAR_PM_STATE	|
+					  MLX5_QP_OPTPAR_ALT_ADDR_PATH,
 		},
 	},
 	[MLX5_QP_STATE_SQER] = {
@@ -2658,6 +2675,10 @@ static enum mlx5_qp_optpar opt_mask[MLX5_QP_NUM_STATE][MLX5_QP_NUM_STATE][MLX5_Q
 					   MLX5_QP_OPTPAR_RWE		|
 					   MLX5_QP_OPTPAR_RAE		|
 					   MLX5_QP_OPTPAR_RRE,
+			[MLX5_QP_ST_XRC]  = MLX5_QP_OPTPAR_RNR_TIMEOUT	|
+					   MLX5_QP_OPTPAR_RWE		|
+					   MLX5_QP_OPTPAR_RAE		|
+					   MLX5_QP_OPTPAR_RRE,
 		},
 	},
 };
-- 
2.20.1

[PATCH AUTOSEL 4.19 328/671] RDMA/rxe: Consider skb reserve space based on netdev of GID

[Sasha Levin]

From: Parav Pandit <parav@mellanox.com>

[ Upstream commit 3bf3e2b881c1412d0329ce9376dfe1518489b8fc ]

Always consider the skb reserve space based on netdevice of the GID
attribute, regardless of vlan or non vlan netdevice.

Fixes: 43c9fc509fa5 ("rdma_rxe: make rxe work over 802.1q VLAN devices")
Signed-off-by: Parav Pandit <parav@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/sw/rxe/rxe_net.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/infiniband/sw/rxe/rxe_net.c b/drivers/infiniband/sw/rxe/rxe_net.c
index 8094cbaa54a9..54add70c22b5 100644
--- a/drivers/infiniband/sw/rxe/rxe_net.c
+++ b/drivers/infiniband/sw/rxe/rxe_net.c
@@ -533,8 +533,9 @@ struct sk_buff *rxe_init_packet(struct rxe_dev *rxe, struct rxe_av *av,
 	if (unlikely(!skb))
 		goto out;
 
-	skb_reserve(skb, hdr_len + LL_RESERVED_SPACE(rxe->ndev));
+	skb_reserve(skb, hdr_len + LL_RESERVED_SPACE(ndev));
 
+	/* FIXME: hold reference to this netdev until life of this skb. */
 	skb->dev	= ndev;
 	if (av->network_type == RDMA_NETWORK_IPV4)
 		skb->protocol = htons(ETH_P_IP);
-- 
2.20.1

[PATCH AUTOSEL 4.19 369/671] RDMA/qedr: Fix incorrect device rate.

[Sasha Levin]

From: Sagiv Ozeri <sagiv.ozeri@marvell.com>

[ Upstream commit 69054666df0a9b4e8331319f98b6b9a88bc3fcc4 ]

Use the correct enum value introduced in commit 12113a35ada6 ("IB/core:
Add HDR speed enum") Prior to this change a 50Gbps port would show 40Gbps.

This patch also cleaned up the redundant redefiniton of ib speeds for
qedr.

Fixes: 12113a35ada6 ("IB/core: Add HDR speed enum")
Signed-off-by: Sagiv Ozeri <sagiv.ozeri@marvell.com>
Signed-off-by: Michal Kalderon <michal.kalderon@marvell.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/hw/qedr/verbs.c | 25 +++++++++----------------
 1 file changed, 9 insertions(+), 16 deletions(-)

diff --git a/drivers/infiniband/hw/qedr/verbs.c b/drivers/infiniband/hw/qedr/verbs.c
index 8fd8b97348bf..38fe2f741375 100644
--- a/drivers/infiniband/hw/qedr/verbs.c
+++ b/drivers/infiniband/hw/qedr/verbs.c
@@ -158,54 +158,47 @@ int qedr_query_device(struct ib_device *ibdev,
 	return 0;
 }
 
-#define QEDR_SPEED_SDR		(1)
-#define QEDR_SPEED_DDR		(2)
-#define QEDR_SPEED_QDR		(4)
-#define QEDR_SPEED_FDR10	(8)
-#define QEDR_SPEED_FDR		(16)
-#define QEDR_SPEED_EDR		(32)
-
 static inline void get_link_speed_and_width(int speed, u8 *ib_speed,
 					    u8 *ib_width)
 {
 	switch (speed) {
 	case 1000:
-		*ib_speed = QEDR_SPEED_SDR;
+		*ib_speed = IB_SPEED_SDR;
 		*ib_width = IB_WIDTH_1X;
 		break;
 	case 10000:
-		*ib_speed = QEDR_SPEED_QDR;
+		*ib_speed = IB_SPEED_QDR;
 		*ib_width = IB_WIDTH_1X;
 		break;
 
 	case 20000:
-		*ib_speed = QEDR_SPEED_DDR;
+		*ib_speed = IB_SPEED_DDR;
 		*ib_width = IB_WIDTH_4X;
 		break;
 
 	case 25000:
-		*ib_speed = QEDR_SPEED_EDR;
+		*ib_speed = IB_SPEED_EDR;
 		*ib_width = IB_WIDTH_1X;
 		break;
 
 	case 40000:
-		*ib_speed = QEDR_SPEED_QDR;
+		*ib_speed = IB_SPEED_QDR;
 		*ib_width = IB_WIDTH_4X;
 		break;
 
 	case 50000:
-		*ib_speed = QEDR_SPEED_QDR;
-		*ib_width = IB_WIDTH_4X;
+		*ib_speed = IB_SPEED_HDR;
+		*ib_width = IB_WIDTH_1X;
 		break;
 
 	case 100000:
-		*ib_speed = QEDR_SPEED_EDR;
+		*ib_speed = IB_SPEED_EDR;
 		*ib_width = IB_WIDTH_4X;
 		break;
 
 	default:
 		/* Unsupported */
-		*ib_speed = QEDR_SPEED_SDR;
+		*ib_speed = IB_SPEED_SDR;
 		*ib_width = IB_WIDTH_1X;
 	}
 }
-- 
2.20.1

[PATCH AUTOSEL 4.19 391/671] RDMA/uverbs: check for allocation failure in uapi_add_elm()

[Sasha Levin]

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

[ Upstream commit cac2a301c02a9b178842e22df34217da7854e588 ]

If the kzalloc() fails then we should return ERR_PTR(-ENOMEM).  In the
current code it's possible that the kzalloc() fails and the
radix_tree_insert() inserts the NULL pointer successfully and we return
the NULL "elm" pointer to the caller.  That results in a NULL pointer
dereference.

Fixes: 9ed3e5f44772 ("IB/uverbs: Build the specs into a radix tree at runtime")
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/core/uverbs_uapi.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/infiniband/core/uverbs_uapi.c b/drivers/infiniband/core/uverbs_uapi.c
index be854628a7c6..959a3418a192 100644
--- a/drivers/infiniband/core/uverbs_uapi.c
+++ b/drivers/infiniband/core/uverbs_uapi.c
@@ -17,6 +17,8 @@ static void *uapi_add_elm(struct uverbs_api *uapi, u32 key, size_t alloc_size)
 		return ERR_PTR(-EOVERFLOW);
 
 	elm = kzalloc(alloc_size, GFP_KERNEL);
+	if (!elm)
+		return ERR_PTR(-ENOMEM);
 	rc = radix_tree_insert(&uapi->radix, key, elm);
 	if (rc) {
 		kfree(elm);
-- 
2.20.1

[PATCH AUTOSEL 4.19 418/671] IB/hfi1: Handle port down properly in pio

[Sasha Levin]

From: Mike Marciniszyn <mike.marciniszyn@intel.com>

[ Upstream commit 942a899335707fc9cfc97cb382a60734b2ff4e03 ]

The call to sc_buffer_alloc currently returns NULL (no buffer) or
a buffer descriptor.

There is a third case when the port is down.  Currently that
returns NULL and this prevents the caller from properly handling the
sc_buffer_alloc() failure.  A verbs code link test after the call is
racy so the indication needs to come from the state check inside the allocation
routine to be valid.

Fix by encoding the ECOMM failure like SDMA.   IS_ERR_OR_NULL() tests
are added at all call sites.  For verbs send, this needs to treat any
error by returning a completion without any MMIO copy.

Fixes: 7724105686e7 ("IB/hfi1: add driver files")
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/hw/hfi1/pio.c   | 5 +++--
 drivers/infiniband/hw/hfi1/rc.c    | 2 +-
 drivers/infiniband/hw/hfi1/ud.c    | 4 ++--
 drivers/infiniband/hw/hfi1/verbs.c | 4 ++--
 4 files changed, 8 insertions(+), 7 deletions(-)

diff --git a/drivers/infiniband/hw/hfi1/pio.c b/drivers/infiniband/hw/hfi1/pio.c
index 752057647f09..3fcbf56f8be2 100644
--- a/drivers/infiniband/hw/hfi1/pio.c
+++ b/drivers/infiniband/hw/hfi1/pio.c
@@ -1434,7 +1434,8 @@ void sc_stop(struct send_context *sc, int flag)
  * @cb: optional callback to call when the buffer is finished sending
  * @arg: argument for cb
  *
- * Return a pointer to a PIO buffer if successful, NULL if not enough room.
+ * Return a pointer to a PIO buffer, NULL if not enough room, -ECOMM
+ * when link is down.
  */
 struct pio_buf *sc_buffer_alloc(struct send_context *sc, u32 dw_len,
 				pio_release_cb cb, void *arg)
@@ -1450,7 +1451,7 @@ struct pio_buf *sc_buffer_alloc(struct send_context *sc, u32 dw_len,
 	spin_lock_irqsave(&sc->alloc_lock, flags);
 	if (!(sc->flags & SCF_ENABLED)) {
 		spin_unlock_irqrestore(&sc->alloc_lock, flags);
-		goto done;
+		return ERR_PTR(-ECOMM);
 	}
 
 retry:
diff --git a/drivers/infiniband/hw/hfi1/rc.c b/drivers/infiniband/hw/hfi1/rc.c
index 980168a56707..7ed6fb407a68 100644
--- a/drivers/infiniband/hw/hfi1/rc.c
+++ b/drivers/infiniband/hw/hfi1/rc.c
@@ -914,7 +914,7 @@ void hfi1_send_rc_ack(struct hfi1_packet *packet, bool is_fecn)
 	pbc = create_pbc(ppd, pbc_flags, qp->srate_mbps,
 			 sc_to_vlt(ppd->dd, sc5), plen);
 	pbuf = sc_buffer_alloc(rcd->sc, plen, NULL, NULL);
-	if (!pbuf) {
+	if (IS_ERR_OR_NULL(pbuf)) {
 		/*
 		 * We have no room to send at the moment.  Pass
 		 * responsibility for sending the ACK to the send engine
diff --git a/drivers/infiniband/hw/hfi1/ud.c b/drivers/infiniband/hw/hfi1/ud.c
index ef5b3ffd3888..839593641e3f 100644
--- a/drivers/infiniband/hw/hfi1/ud.c
+++ b/drivers/infiniband/hw/hfi1/ud.c
@@ -703,7 +703,7 @@ void return_cnp_16B(struct hfi1_ibport *ibp, struct rvt_qp *qp,
 	pbc = create_pbc(ppd, pbc_flags, qp->srate_mbps, vl, plen);
 	if (ctxt) {
 		pbuf = sc_buffer_alloc(ctxt, plen, NULL, NULL);
-		if (pbuf) {
+		if (!IS_ERR_OR_NULL(pbuf)) {
 			trace_pio_output_ibhdr(ppd->dd, &hdr, sc5);
 			ppd->dd->pio_inline_send(ppd->dd, pbuf, pbc,
 						 &hdr, hwords);
@@ -758,7 +758,7 @@ void return_cnp(struct hfi1_ibport *ibp, struct rvt_qp *qp, u32 remote_qpn,
 	pbc = create_pbc(ppd, pbc_flags, qp->srate_mbps, vl, plen);
 	if (ctxt) {
 		pbuf = sc_buffer_alloc(ctxt, plen, NULL, NULL);
-		if (pbuf) {
+		if (!IS_ERR_OR_NULL(pbuf)) {
 			trace_pio_output_ibhdr(ppd->dd, &hdr, sc5);
 			ppd->dd->pio_inline_send(ppd->dd, pbuf, pbc,
 						 &hdr, hwords);
diff --git a/drivers/infiniband/hw/hfi1/verbs.c b/drivers/infiniband/hw/hfi1/verbs.c
index 4e7b3c027901..90e12f9433a3 100644
--- a/drivers/infiniband/hw/hfi1/verbs.c
+++ b/drivers/infiniband/hw/hfi1/verbs.c
@@ -1096,10 +1096,10 @@ int hfi1_verbs_send_pio(struct rvt_qp *qp, struct hfi1_pkt_state *ps,
 	if (cb)
 		iowait_pio_inc(&priv->s_iowait);
 	pbuf = sc_buffer_alloc(sc, plen, cb, qp);
-	if (unlikely(!pbuf)) {
+	if (unlikely(IS_ERR_OR_NULL(pbuf))) {
 		if (cb)
 			verbs_pio_complete(qp, 0);
-		if (ppd->host_link_state != HLS_UP_ACTIVE) {
+		if (IS_ERR(pbuf)) {
 			/*
 			 * If we have filled the PIO buffers to capacity and are
 			 * not in an active state this request is not going to
-- 
2.20.1

[PATCH AUTOSEL 4.19 453/671] RDMA/hns: Fixs hw access invalid dma memory error

[Sasha Levin]

From: Xi Wang <wangxi11@huawei.com>

[ Upstream commit ec5bc2cc69b4fc494e04d10fc5226f6f9cf67c56 ]

When smmu is enable, if execute the perftest command and then use 'kill
-9' to exit, follow this operation repeatedly, the kernel will have a high
probability to print the following smmu event:

  arm-smmu-v3 arm-smmu-v3.1.auto: event 0x10 received:
  arm-smmu-v3 arm-smmu-v3.1.auto:  0x00007d0000000010
  arm-smmu-v3 arm-smmu-v3.1.auto:  0x0000020900000080
  arm-smmu-v3 arm-smmu-v3.1.auto:  0x00000000f47cf000
  arm-smmu-v3 arm-smmu-v3.1.auto:  0x00000000f47cf000

This is because the hw will periodically refresh the qpc cache until the
next reset.

This patch fixed it by removing the action that release qpc memory in the
'hns_roce_qp_free' function.

Fixes: 9a4435375cd1 ("IB/hns: Add driver files for hns RoCE driver")
Signed-off-by: Xi Wang <wangxi11@huawei.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/hw/hns/hns_roce_qp.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/infiniband/hw/hns/hns_roce_qp.c b/drivers/infiniband/hw/hns/hns_roce_qp.c
index af24698ff226..3012d7eb4ccb 100644
--- a/drivers/infiniband/hw/hns/hns_roce_qp.c
+++ b/drivers/infiniband/hw/hns/hns_roce_qp.c
@@ -262,7 +262,6 @@ void hns_roce_qp_free(struct hns_roce_dev *hr_dev, struct hns_roce_qp *hr_qp)
 			hns_roce_table_put(hr_dev, &qp_table->trrl_table,
 					   hr_qp->qpn);
 		hns_roce_table_put(hr_dev, &qp_table->irrl_table, hr_qp->qpn);
-		hns_roce_table_put(hr_dev, &qp_table->qp_table, hr_qp->qpn);
 	}
 }
 EXPORT_SYMBOL_GPL(hns_roce_qp_free);
-- 
2.20.1

[PATCH AUTOSEL 4.19 480/671] net/mlx5: Fix mlx5_ifc_query_lag_out_bits

[Sasha Levin]

From: Mark Zhang <markz@mellanox.com>

[ Upstream commit ea77388b02270b0af8dc57f668f311235ea068f0 ]

Remove the "reserved_at_40" field to match the device specification.

Fixes: 84df61ebc69b ("net/mlx5: Add HW interfaces used by LAG")
Signed-off-by: Mark Zhang <markz@mellanox.com>
Reviewed-by: Yishai Hadas <yishaih@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 include/linux/mlx5/mlx5_ifc.h | 2 --
 1 file changed, 2 deletions(-)

diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h
index 177f11c96187..76b76b6aa83d 100644
--- a/include/linux/mlx5/mlx5_ifc.h
+++ b/include/linux/mlx5/mlx5_ifc.h
@@ -9053,8 +9053,6 @@ struct mlx5_ifc_query_lag_out_bits {
 
 	u8         syndrome[0x20];
 
-	u8         reserved_at_40[0x40];
-
 	struct mlx5_ifc_lagc_bits ctx;
 };
 
-- 
2.20.1

[PATCH AUTOSEL 4.19 488/671] RDMA/hns: Bugfix for slab-out-of-bounds when unloading hip08 driver

[Sasha Levin]

From: Xi Wang <wangxi11@huawei.com>

[ Upstream commit 9bba3f0cbfc8abf2e1549ea03c0128186081d7a8 ]

kasan will report a BUG when run command 'rmmod hns_roce_hw_v2', the calltrace
is as follows:

==================================================================
BUG: KASAN: slab-out-of-bounds in hns_roce_table_mhop_put+0x584/0x828
[hns_roce]
Read of size 8 at addr ffff802185e08300 by task rmmod/270

Call trace:
dump_backtrace+0x0/0x1e8
show_stack+0x14/0x20
dump_stack+0xc4/0xfc
print_address_description+0x60/0x270
__kasan_report+0x164/0x1b8
kasan_report+0xc/0x18
__asan_load8+0x84/0xa8
hns_roce_table_mhop_put+0x584/0x828 [hns_roce]
hns_roce_table_put+0x174/0x1a0 [hns_roce]
hns_roce_mr_free+0x124/0x210 [hns_roce]
hns_roce_dereg_mr+0x90/0xb8 [hns_roce]
ib_dealloc_pd_user+0x60/0xf0
ib_mad_port_close+0x128/0x1d8
ib_mad_remove_device+0x94/0x118
remove_client_context+0xa0/0xe0
disable_device+0xfc/0x1c0
__ib_unregister_device+0x60/0xe0
ib_unregister_device+0x24/0x38
hns_roce_exit+0x3c/0x138 [hns_roce]
__hns_roce_hw_v2_uninit_instance.isra.30+0x28/0x50 [hns_roce_hw_v2]
hns_roce_hw_v2_uninit_instance+0x44/0x60 [hns_roce_hw_v2]
hclge_uninit_client_instance+0x15c/0x238 [hclge]
hnae3_uninit_client_instance+0x84/0xa8 [hnae3]
hnae3_unregister_client+0x84/0x158 [hnae3]
hns_roce_hw_v2_exit+0x14/0x20 [hns_roce_hw_v2]
__arm64_sys_delete_module+0x20c/0x308
el0_svc_handler+0xbc/0x210
el0_svc+0x8/0xc

Allocated by task 255:
__kasan_kmalloc.isra.0+0xd0/0x180
kasan_kmalloc+0xc/0x18
__kmalloc+0x16c/0x328
hns_roce_init_hem_table+0x20c/0x428 [hns_roce]
hns_roce_init+0x214/0xfe0 [hns_roce]
__hns_roce_hw_v2_init_instance+0x284/0x330 [hns_roce_hw_v2]
hns_roce_hw_v2_init_instance+0xd0/0x1b8 [hns_roce_hw_v2]
hclge_init_roce_client_instance+0x180/0x310 [hclge]
hclge_init_client_instance+0xcc/0x508 [hclge]
hnae3_init_client_instance.part.3+0x3c/0x80 [hnae3]
hnae3_register_client+0x134/0x1a8 [hnae3]
0xffff200009c00014
do_one_initcall+0x9c/0x3e0
do_init_module+0xd4/0x2d8
load_module+0x3284/0x3690
__se_sys_init_module+0x274/0x308
__arm64_sys_init_module+0x40/0x50
el0_svc_handler+0xbc/0x210
el0_svc+0x8/0xc

Freed by task 0:
(stack is not available)

The buggy address belongs to the object at ffff802185e06300
which belongs to the cache kmalloc-8k of size 8192
The buggy address is located 0 bytes to the right of
8192-byte region [ffff802185e06300, ffff802185e08300)
The buggy address belongs to the page:
page:ffff7fe008617800 refcount:1 mapcount:0 mapping:ffff802340020e00 index:0x0
compound_mapcount: 0
flags: 0x5fffe00000010200(slab|head)
raw: 5fffe00000010200 dead000000000100 dead000000000200 ffff802340020e00
raw: 0000000000000000 00000000803e003e 00000001ffffffff 0000000000000000
page dumped because: kasan: bad access detected

Memory state around the buggy address:
ffff802185e08200: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
ffff802185e08280: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>ffff802185e08300: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
^
ffff802185e08380: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
ffff802185e08400: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
==================================================================
Disabling lock debugging due to kernel taint

Fixes: a25d13cbe816 ("RDMA/hns: Add the interfaces to support multi hop addressing for the contexts in hip08")

Signed-off-by: Xi Wang <wangxi11@huawei.com>
Link: https://lore.kernel.org/r/1565343666-73193-6-git-send-email-oulijun@huawei.com
Signed-off-by: Doug Ledford <dledford@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/hw/hns/hns_roce_hem.c | 19 +++++++++++++------
 1 file changed, 13 insertions(+), 6 deletions(-)

diff --git a/drivers/infiniband/hw/hns/hns_roce_hem.c b/drivers/infiniband/hw/hns/hns_roce_hem.c
index a73d388b7093..31b9b99f81cb 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hem.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hem.c
@@ -54,12 +54,13 @@ bool hns_roce_check_whether_mhop(struct hns_roce_dev *hr_dev, u32 type)
 EXPORT_SYMBOL_GPL(hns_roce_check_whether_mhop);
 
 static bool hns_roce_check_hem_null(struct hns_roce_hem **hem, u64 start_idx,
-			    u32 bt_chunk_num)
+			    u32 bt_chunk_num, u64 hem_max_num)
 {
-	int i;
+	u64 check_max_num = start_idx + bt_chunk_num;
+	u64 i;
 
-	for (i = 0; i < bt_chunk_num; i++)
-		if (hem[start_idx + i])
+	for (i = start_idx; (i < check_max_num) && (i < hem_max_num); i++)
+		if (hem[i])
 			return false;
 
 	return true;
@@ -413,6 +414,12 @@ static int hns_roce_table_mhop_get(struct hns_roce_dev *hr_dev,
 		return -EINVAL;
 	}
 
+	if (unlikely(hem_idx >= table->num_hem)) {
+		dev_err(dev, "Table %d exceed hem limt idx = %llu,max = %lu!\n",
+			     table->type, hem_idx, table->num_hem);
+		return -EINVAL;
+	}
+
 	mutex_lock(&table->mutex);
 
 	if (table->hem[hem_idx]) {
@@ -649,7 +656,7 @@ static void hns_roce_table_mhop_put(struct hns_roce_dev *hr_dev,
 	if (check_whether_bt_num_2(table->type, hop_num)) {
 		start_idx = mhop.l0_idx * chunk_ba_num;
 		if (hns_roce_check_hem_null(table->hem, start_idx,
-					    chunk_ba_num)) {
+					    chunk_ba_num, table->num_hem)) {
 			if (table->type < HEM_TYPE_MTT &&
 			    hr_dev->hw->clear_hem(hr_dev, table, obj, 0))
 				dev_warn(dev, "Clear HEM base address failed.\n");
@@ -663,7 +670,7 @@ static void hns_roce_table_mhop_put(struct hns_roce_dev *hr_dev,
 		start_idx = mhop.l0_idx * chunk_ba_num * chunk_ba_num +
 			    mhop.l1_idx * chunk_ba_num;
 		if (hns_roce_check_hem_null(table->hem, start_idx,
-					    chunk_ba_num)) {
+					    chunk_ba_num, table->num_hem)) {
 			if (hr_dev->hw->clear_hem(hr_dev, table, obj, 1))
 				dev_warn(dev, "Clear HEM base address failed.\n");
 
-- 
2.20.1

[PATCH AUTOSEL 4.19 489/671] RDMA/hns: bugfix for slab-out-of-bounds when loading hip08 driver

[Sasha Levin]

From: Xi Wang <wangxi11@huawei.com>

[ Upstream commit bf8c02f961c89e5ccae5987b7ab28f5592a35101 ]

kasan will report a BUG when run command 'insmod hns_roce_hw_v2.ko', the
calltrace is as follows:

==================================================================
BUG: KASAN: slab-out-of-bounds in hns_roce_v2_init_eq_table+0x1324/0x1948
[hns_roce_hw_v2]
Read of size 8 at addr ffff8020e7a10608 by task insmod/256

CPU: 0 PID: 256 Comm: insmod Tainted: G           O      5.2.0-rc4 #1
Hardware name: Huawei D06 /D06, BIOS Hisilicon D06 UEFI RC0
Call trace:
dump_backtrace+0x0/0x1e8
show_stack+0x14/0x20
dump_stack+0xc4/0xfc
print_address_description+0x60/0x270
__kasan_report+0x164/0x1b8
kasan_report+0xc/0x18
__asan_load8+0x84/0xa8
hns_roce_v2_init_eq_table+0x1324/0x1948 [hns_roce_hw_v2]
hns_roce_init+0xf8/0xfe0 [hns_roce]
__hns_roce_hw_v2_init_instance+0x284/0x330 [hns_roce_hw_v2]
hns_roce_hw_v2_init_instance+0xd0/0x1b8 [hns_roce_hw_v2]
hclge_init_roce_client_instance+0x180/0x310 [hclge]
hclge_init_client_instance+0xcc/0x508 [hclge]
hnae3_init_client_instance.part.3+0x3c/0x80 [hnae3]
hnae3_register_client+0x134/0x1a8 [hnae3]
hns_roce_hw_v2_init+0x14/0x10000 [hns_roce_hw_v2]
do_one_initcall+0x9c/0x3e0
do_init_module+0xd4/0x2d8
load_module+0x3284/0x3690
__se_sys_init_module+0x274/0x308
__arm64_sys_init_module+0x40/0x50
el0_svc_handler+0xbc/0x210
el0_svc+0x8/0xc

Allocated by task 256:
__kasan_kmalloc.isra.0+0xd0/0x180
kasan_kmalloc+0xc/0x18
__kmalloc+0x16c/0x328
hns_roce_v2_init_eq_table+0x764/0x1948 [hns_roce_hw_v2]
hns_roce_init+0xf8/0xfe0 [hns_roce]
__hns_roce_hw_v2_init_instance+0x284/0x330 [hns_roce_hw_v2]
hns_roce_hw_v2_init_instance+0xd0/0x1b8 [hns_roce_hw_v2]
hclge_init_roce_client_instance+0x180/0x310 [hclge]
hclge_init_client_instance+0xcc/0x508 [hclge]
hnae3_init_client_instance.part.3+0x3c/0x80 [hnae3]
hnae3_register_client+0x134/0x1a8 [hnae3]
hns_roce_hw_v2_init+0x14/0x10000 [hns_roce_hw_v2]
do_one_initcall+0x9c/0x3e0
do_init_module+0xd4/0x2d8
load_module+0x3284/0x3690
__se_sys_init_module+0x274/0x308
__arm64_sys_init_module+0x40/0x50
el0_svc_handler+0xbc/0x210
el0_svc+0x8/0xc

Freed by task 0:
(stack is not available)

The buggy address belongs to the object at ffff8020e7a10600
which belongs to the cache kmalloc-128 of size 128
The buggy address is located 8 bytes inside of
128-byte region [ffff8020e7a10600, ffff8020e7a10680)
The buggy address belongs to the page:
page:ffff7fe00839e840 refcount:1 mapcount:0 mapping:ffff802340020200 index:0x0
flags: 0x5fffe00000000200(slab)
raw: 5fffe00000000200 dead000000000100 dead000000000200 ffff802340020200
raw: 0000000000000000 0000000081000100 00000001ffffffff 0000000000000000
page dumped because: kasan: bad access detected

Memory state around the buggy address:
ffff8020e7a10500: 00 00 00 00 00 00 00 00 fc fc fc fc fc fc fc fc
ffff8020e7a10580: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
>ffff8020e7a10600: 00 fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
^
ffff8020e7a10680: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
ffff8020e7a10700: 00 fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
==================================================================
Disabling lock debugging due to kernel taint

Fixes: a5073d6054f7 ("RDMA/hns: Add eq support of hip08")

Signed-off-by: Xi Wang <wangxi11@huawei.com>
Link: https://lore.kernel.org/r/1565343666-73193-7-git-send-email-oulijun@huawei.com
Signed-off-by: Doug Ledford <dledford@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
index 7021444f18b4..417de7ac0d5e 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
@@ -4833,7 +4833,8 @@ static int hns_roce_mhop_alloc_eq(struct hns_roce_dev *hr_dev,
 				break;
 		}
 		eq->cur_eqe_ba = eq->buf_dma[0];
-		eq->nxt_eqe_ba = eq->buf_dma[1];
+		if (ba_num > 1)
+			eq->nxt_eqe_ba = eq->buf_dma[1];
 
 	} else if (mhop_num == 2) {
 		/* alloc L1 BT and buf */
@@ -4875,7 +4876,8 @@ static int hns_roce_mhop_alloc_eq(struct hns_roce_dev *hr_dev,
 				break;
 		}
 		eq->cur_eqe_ba = eq->buf_dma[0];
-		eq->nxt_eqe_ba = eq->buf_dma[1];
+		if (ba_num > 1)
+			eq->nxt_eqe_ba = eq->buf_dma[1];
 	}
 
 	eq->l0_last_num = i + 1;
-- 
2.20.1

[PATCH AUTOSEL 4.19 493/671] net/rds: Add a few missing rds_stat_names entries

[Sasha Levin]

From: Gerd Rausch <gerd.rausch@oracle.com>

[ Upstream commit 55c70ca00c982fbc0df4c4d3e31747fb73f4ddb5 ]

In a previous commit, fields were added to "struct rds_statistics"
but array "rds_stat_names" was not updated accordingly.

Please note the inconsistent naming of the string representations
that is done in the name of compatibility
with the Oracle internal code-base.

s_recv_bytes_added_to_socket     -> "recv_bytes_added_to_sock"
s_recv_bytes_removed_from_socket -> "recv_bytes_freed_fromsock"

Fixes: 192a798f5299 ("RDS: add stat for socket recv memory usage")
Signed-off-by: Gerd Rausch <gerd.rausch@oracle.com>
Acked-by: Santosh Shilimkar <santosh.shilimkar@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 net/rds/stats.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/net/rds/stats.c b/net/rds/stats.c
index 73be187d389e..6bbab4d74c4f 100644
--- a/net/rds/stats.c
+++ b/net/rds/stats.c
@@ -76,6 +76,8 @@ static const char *const rds_stat_names[] = {
 	"cong_update_received",
 	"cong_send_error",
 	"cong_send_blocked",
+	"recv_bytes_added_to_sock",
+	"recv_bytes_freed_fromsock",
 };
 
 void rds_stats_info_copy(struct rds_info_iterator *iter,
-- 
2.20.1

[PATCH AUTOSEL 4.19 549/671] RDMA/cma: Fix false error message

[Sasha Levin]

From: Håkon Bugge <haakon.bugge@oracle.com>

[ Upstream commit a6e4d254c19b541a58caced322111084b27a7788 ]

In addr_handler(), assuming status == 0 and the device already has been
acquired (id_priv->cma_dev != NULL), we get the following incorrect
"error" message:

RDMA CM: ADDR_ERROR: failed to resolve IP. status 0

Fixes: 498683c6a7ee ("IB/cma: Add debug messages to error flows")
Link: https://lore.kernel.org/r/20190902092731.1055757-1-haakon.bugge@oracle.com
Signed-off-by: Håkon Bugge <haakon.bugge@oracle.com>
Reviewed-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/core/cma.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c
index 319bfef00a4a..e16872e0724f 100644
--- a/drivers/infiniband/core/cma.c
+++ b/drivers/infiniband/core/cma.c
@@ -2889,7 +2889,7 @@ static void addr_handler(int status, struct sockaddr *src_addr,
 		if (status)
 			pr_debug_ratelimited("RDMA CM: ADDR_ERROR: failed to acquire device. status %d\n",
 					     status);
-	} else {
+	} else if (status) {
 		pr_debug_ratelimited("RDMA CM: ADDR_ERROR: failed to resolve IP. status %d\n", status);
 	}
 
-- 
2.20.1

[PATCH AUTOSEL 4.19 550/671] RDMA: Fix goto target to release the allocated memory

[Sasha Levin]

From: Navid Emamdoost <navid.emamdoost@gmail.com>

[ Upstream commit 4a9d46a9fe14401f21df69cea97c62396d5fb053 ]

In bnxt_re_create_srq(), when ib_copy_to_udata() fails allocated memory
should be released by goto fail.

Fixes: 37cb11acf1f7 ("RDMA/bnxt_re: Add SRQ support for Broadcom adapters")
Link: https://lore.kernel.org/r/20190910222120.16517-1-navid.emamdoost@gmail.com
Signed-off-by: Navid Emamdoost <navid.emamdoost@gmail.com>
Reviewed-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/hw/bnxt_re/ib_verbs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/infiniband/hw/bnxt_re/ib_verbs.c b/drivers/infiniband/hw/bnxt_re/ib_verbs.c
index bea8318e7007..a019ff266f52 100644
--- a/drivers/infiniband/hw/bnxt_re/ib_verbs.c
+++ b/drivers/infiniband/hw/bnxt_re/ib_verbs.c
@@ -1446,7 +1446,7 @@ struct ib_srq *bnxt_re_create_srq(struct ib_pd *ib_pd,
 			dev_err(rdev_to_dev(rdev), "SRQ copy to udata failed!");
 			bnxt_qplib_destroy_srq(&rdev->qplib_res,
 					       &srq->qplib_srq);
-			goto exit;
+			goto fail;
 		}
 	}
 	if (nq)
-- 
2.20.1

[PATCH AUTOSEL 4.19 551/671] net/rds: Fix 'ib_evt_handler_call' element in 'rds_ib_stat_names'

[Sasha Levin]

From: Gerd Rausch <gerd.rausch@oracle.com>

[ Upstream commit 05a82481a3024b94db00b8c816bb3d526b5209e0 ]

All entries in 'rds_ib_stat_names' are stringified versions
of the corresponding "struct rds_ib_statistics" element
without the "s_"-prefix.

Fix entry 'ib_evt_handler_call' to do the same.

Fixes: f4f943c958a2 ("RDS: IB: ack more receive completions to improve performance")
Signed-off-by: Gerd Rausch <gerd.rausch@oracle.com>
Acked-by: Santosh Shilimkar <santosh.shilimkar@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 net/rds/ib_stats.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/net/rds/ib_stats.c b/net/rds/ib_stats.c
index 9252ad126335..ac46d8961b61 100644
--- a/net/rds/ib_stats.c
+++ b/net/rds/ib_stats.c
@@ -42,7 +42,7 @@ DEFINE_PER_CPU_SHARED_ALIGNED(struct rds_ib_statistics, rds_ib_stats);
 static const char *const rds_ib_stat_names[] = {
 	"ib_connect_raced",
 	"ib_listen_closed_stale",
-	"s_ib_evt_handler_call",
+	"ib_evt_handler_call",
 	"ib_tasklet_call",
 	"ib_tx_cq_event",
 	"ib_tx_ring_full",
-- 
2.20.1

[PATCH AUTOSEL 4.19 629/671] RDMA/mlx5: Return proper error value

[Sasha Levin]

From: Leon Romanovsky <leonro@mellanox.com>

[ Upstream commit 546d30099ed204792083f043cd7e016de86016a3 ]

Returned value from mlx5_mr_cache_alloc() is checked to be error or real
pointer. Return proper error code instead of NULL which is not checked
later.

Fixes: 81713d3788d2 ("IB/mlx5: Add implicit MR support")
Link: https://lore.kernel.org/r/20191029055721.7192-1-leon@kernel.org
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Reviewed-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/hw/mlx5/mr.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/infiniband/hw/mlx5/mr.c b/drivers/infiniband/hw/mlx5/mr.c
index bd1fdadf7ba0..18fd9aa6510f 100644
--- a/drivers/infiniband/hw/mlx5/mr.c
+++ b/drivers/infiniband/hw/mlx5/mr.c
@@ -457,7 +457,7 @@ struct mlx5_ib_mr *mlx5_mr_cache_alloc(struct mlx5_ib_dev *dev, int entry)
 
 	if (entry < 0 || entry >= MAX_MR_CACHE_ENTRIES) {
 		mlx5_ib_err(dev, "cache entry %d is out of range\n", entry);
-		return NULL;
+		return ERR_PTR(-EINVAL);
 	}
 
 	ent = &cache->ent[entry];
-- 
2.20.1

[PATCH AUTOSEL 4.19 669/671] IB/iser: Fix dma_nents type definition

[Sasha Levin]

From: Max Gurtovoy <maxg@mellanox.com>

[ Upstream commit c1545f1a200f4adc4ef8dd534bf33e2f1aa22c2f ]

The retured value from ib_dma_map_sg saved in dma_nents variable. To avoid
future mismatch between types, define dma_nents as an integer instead of
unsigned.

Fixes: 57b26497fabe ("IB/iser: Pass the correct number of entries for dma mapped SGL")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Israel Rukshin <israelr@mellanox.com>
Signed-off-by: Max Gurtovoy <maxg@mellanox.com>
Acked-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-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/ulp/iser/iscsi_iser.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.h b/drivers/infiniband/ulp/iser/iscsi_iser.h
index 120b40829560..a7aeaa0c6fbc 100644
--- a/drivers/infiniband/ulp/iser/iscsi_iser.h
+++ b/drivers/infiniband/ulp/iser/iscsi_iser.h
@@ -197,7 +197,7 @@ struct iser_data_buf {
 	struct scatterlist *sg;
 	int                size;
 	unsigned long      data_len;
-	unsigned int       dma_nents;
+	int                dma_nents;
 };
 
 /* fwd declarations */
-- 
2.20.1