[PATCH AUTOSEL 5.4 135/350] RDMA/hns: Fix wrong parameters when initial mtt of srq->idx_que

[PATCH AUTOSEL 5.4 135/350] RDMA/hns: Fix wrong parameters when initial mtt of srq->idx_que

[Sasha Levin]

From: Weihang Li <liweihang@hisilicon.com>

[ Upstream commit e8a07de57ea4ca7c2d604871c52826e66899fc70 ]

The parameters npages used to initial mtt of srq->idx_que shouldn't be
same with srq's. And page_shift should be calculated from idx_buf_pg_sz.
This patch fixes above issues and use field named npage and page_shift
in hns_roce_buf instead of two temporary variables to let us use them
anywhere.

Fixes: 18df508c7970 ("RDMA/hns: Remove if-else judgment statements for creating srq")
Signed-off-by: Weihang Li <liweihang@hisilicon.com>
Link: https://lore.kernel.org/r/1567566885-23088-3-git-send-email-liweihang@hisilicon.com
Signed-off-by: Doug Ledford <dledford@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/hw/hns/hns_roce_srq.c | 24 ++++++++++++++----------
 1 file changed, 14 insertions(+), 10 deletions(-)

diff --git a/drivers/infiniband/hw/hns/hns_roce_srq.c b/drivers/infiniband/hw/hns/hns_roce_srq.c
index 43ea2c13b2122..108667ae6b14c 100644
--- a/drivers/infiniband/hw/hns/hns_roce_srq.c
+++ b/drivers/infiniband/hw/hns/hns_roce_srq.c
@@ -180,8 +180,7 @@ static int create_user_srq(struct hns_roce_srq *srq, struct ib_udata *udata,
 {
 	struct hns_roce_dev *hr_dev = to_hr_dev(srq->ibsrq.device);
 	struct hns_roce_ib_create_srq  ucmd;
-	u32 page_shift;
-	u32 npages;
+	struct hns_roce_buf *buf;
 	int ret;
 
 	if (ib_copy_from_udata(&ucmd, udata, sizeof(ucmd)))
@@ -191,11 +190,13 @@ static int create_user_srq(struct hns_roce_srq *srq, struct ib_udata *udata,
 	if (IS_ERR(srq->umem))
 		return PTR_ERR(srq->umem);
 
-	npages = (ib_umem_page_count(srq->umem) +
-		(1 << hr_dev->caps.srqwqe_buf_pg_sz) - 1) /
-		(1 << hr_dev->caps.srqwqe_buf_pg_sz);
-	page_shift = PAGE_SHIFT + hr_dev->caps.srqwqe_buf_pg_sz;
-	ret = hns_roce_mtt_init(hr_dev, npages, page_shift, &srq->mtt);
+	buf = &srq->buf;
+	buf->npages = (ib_umem_page_count(srq->umem) +
+		       (1 << hr_dev->caps.srqwqe_buf_pg_sz) - 1) /
+		      (1 << hr_dev->caps.srqwqe_buf_pg_sz);
+	buf->page_shift = PAGE_SHIFT + hr_dev->caps.srqwqe_buf_pg_sz;
+	ret = hns_roce_mtt_init(hr_dev, buf->npages, buf->page_shift,
+				&srq->mtt);
 	if (ret)
 		goto err_user_buf;
 
@@ -212,9 +213,12 @@ static int create_user_srq(struct hns_roce_srq *srq, struct ib_udata *udata,
 		goto err_user_srq_mtt;
 	}
 
-	ret = hns_roce_mtt_init(hr_dev, ib_umem_page_count(srq->idx_que.umem),
-				PAGE_SHIFT, &srq->idx_que.mtt);
-
+	buf = &srq->idx_que.idx_buf;
+	buf->npages = DIV_ROUND_UP(ib_umem_page_count(srq->idx_que.umem),
+				   1 << hr_dev->caps.idx_buf_pg_sz);
+	buf->page_shift = PAGE_SHIFT + hr_dev->caps.idx_buf_pg_sz;
+	ret = hns_roce_mtt_init(hr_dev, buf->npages, buf->page_shift,
+				&srq->idx_que.mtt);
 	if (ret) {
 		dev_err(hr_dev->dev, "hns_roce_mtt_init error for idx que\n");
 		goto err_user_idx_mtt;
-- 
2.20.1

[PATCH AUTOSEL 5.4 140/350] RDMA/siw: Fix SQ/RQ drain logic

[Sasha Levin]

From: Bernard Metzler <bmt@zurich.ibm.com>

[ Upstream commit cf049bb31f7101d9672eaf97ade4fdd5171ddf26 ]

Storage ULPs (e.g. iSER & NVMeOF) use ib_drain_qp() to drain
QP/CQ. Current SIW's own drain routines do not properly wait until all
SQ/RQ elements are completed and reaped from the CQ. This may cause touch
after free issues.  New logic relies on generic
__ib_drain_sq()/__ib_drain_rq() posting a final work request, which SIW
immediately flushes to CQ.

Fixes: 303ae1cdfdf7 ("rdma/siw: application interface")
Link: https://lore.kernel.org/r/20191004125356.20673-1-bmt@zurich.ibm.com
Signed-off-by: Krishnamraju Eraparaju <krishna2@chelsio.com>
Signed-off-by: Bernard Metzler <bmt@zurich.ibm.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/sw/siw/siw_main.c  |  20 ----
 drivers/infiniband/sw/siw/siw_verbs.c | 144 ++++++++++++++++++++++----
 2 files changed, 122 insertions(+), 42 deletions(-)

diff --git a/drivers/infiniband/sw/siw/siw_main.c b/drivers/infiniband/sw/siw/siw_main.c
index 05a92f997f603..fb01407a310fa 100644
--- a/drivers/infiniband/sw/siw/siw_main.c
+++ b/drivers/infiniband/sw/siw/siw_main.c
@@ -248,24 +248,6 @@ static struct ib_qp *siw_get_base_qp(struct ib_device *base_dev, int id)
 	return NULL;
 }
 
-static void siw_verbs_sq_flush(struct ib_qp *base_qp)
-{
-	struct siw_qp *qp = to_siw_qp(base_qp);
-
-	down_write(&qp->state_lock);
-	siw_sq_flush(qp);
-	up_write(&qp->state_lock);
-}
-
-static void siw_verbs_rq_flush(struct ib_qp *base_qp)
-{
-	struct siw_qp *qp = to_siw_qp(base_qp);
-
-	down_write(&qp->state_lock);
-	siw_rq_flush(qp);
-	up_write(&qp->state_lock);
-}
-
 static const struct ib_device_ops siw_device_ops = {
 	.owner = THIS_MODULE,
 	.uverbs_abi_ver = SIW_ABI_VERSION,
@@ -284,8 +266,6 @@ static const struct ib_device_ops siw_device_ops = {
 	.destroy_cq = siw_destroy_cq,
 	.destroy_qp = siw_destroy_qp,
 	.destroy_srq = siw_destroy_srq,
-	.drain_rq = siw_verbs_rq_flush,
-	.drain_sq = siw_verbs_sq_flush,
 	.get_dma_mr = siw_get_dma_mr,
 	.get_port_immutable = siw_get_port_immutable,
 	.iw_accept = siw_accept,
diff --git a/drivers/infiniband/sw/siw/siw_verbs.c b/drivers/infiniband/sw/siw/siw_verbs.c
index b18a677832e10..7d0a7edc533d1 100644
--- a/drivers/infiniband/sw/siw/siw_verbs.c
+++ b/drivers/infiniband/sw/siw/siw_verbs.c
@@ -685,6 +685,47 @@ static int siw_copy_inline_sgl(const struct ib_send_wr *core_wr,
 	return bytes;
 }
 
+/* Complete SQ WR's without processing */
+static int siw_sq_flush_wr(struct siw_qp *qp, const struct ib_send_wr *wr,
+			   const struct ib_send_wr **bad_wr)
+{
+	struct siw_sqe sqe = {};
+	int rv = 0;
+
+	while (wr) {
+		sqe.id = wr->wr_id;
+		sqe.opcode = wr->opcode;
+		rv = siw_sqe_complete(qp, &sqe, 0, SIW_WC_WR_FLUSH_ERR);
+		if (rv) {
+			if (bad_wr)
+				*bad_wr = wr;
+			break;
+		}
+		wr = wr->next;
+	}
+	return rv;
+}
+
+/* Complete RQ WR's without processing */
+static int siw_rq_flush_wr(struct siw_qp *qp, const struct ib_recv_wr *wr,
+			   const struct ib_recv_wr **bad_wr)
+{
+	struct siw_rqe rqe = {};
+	int rv = 0;
+
+	while (wr) {
+		rqe.id = wr->wr_id;
+		rv = siw_rqe_complete(qp, &rqe, 0, 0, SIW_WC_WR_FLUSH_ERR);
+		if (rv) {
+			if (bad_wr)
+				*bad_wr = wr;
+			break;
+		}
+		wr = wr->next;
+	}
+	return rv;
+}
+
 /*
  * siw_post_send()
  *
@@ -703,26 +744,54 @@ int siw_post_send(struct ib_qp *base_qp, const struct ib_send_wr *wr,
 	unsigned long flags;
 	int rv = 0;
 
+	if (wr && !qp->kernel_verbs) {
+		siw_dbg_qp(qp, "wr must be empty for user mapped sq\n");
+		*bad_wr = wr;
+		return -EINVAL;
+	}
+
 	/*
 	 * Try to acquire QP state lock. Must be non-blocking
 	 * to accommodate kernel clients needs.
 	 */
 	if (!down_read_trylock(&qp->state_lock)) {
-		*bad_wr = wr;
-		siw_dbg_qp(qp, "QP locked, state %d\n", qp->attrs.state);
-		return -ENOTCONN;
+		if (qp->attrs.state == SIW_QP_STATE_ERROR) {
+			/*
+			 * ERROR state is final, so we can be sure
+			 * this state will not change as long as the QP
+			 * exists.
+			 *
+			 * This handles an ib_drain_sq() call with
+			 * a concurrent request to set the QP state
+			 * to ERROR.
+			 */
+			rv = siw_sq_flush_wr(qp, wr, bad_wr);
+		} else {
+			siw_dbg_qp(qp, "QP locked, state %d\n",
+				   qp->attrs.state);
+			*bad_wr = wr;
+			rv = -ENOTCONN;
+		}
+		return rv;
 	}
 	if (unlikely(qp->attrs.state != SIW_QP_STATE_RTS)) {
+		if (qp->attrs.state == SIW_QP_STATE_ERROR) {
+			/*
+			 * Immediately flush this WR to CQ, if QP
+			 * is in ERROR state. SQ is guaranteed to
+			 * be empty, so WR complets in-order.
+			 *
+			 * Typically triggered by ib_drain_sq().
+			 */
+			rv = siw_sq_flush_wr(qp, wr, bad_wr);
+		} else {
+			siw_dbg_qp(qp, "QP out of state %d\n",
+				   qp->attrs.state);
+			*bad_wr = wr;
+			rv = -ENOTCONN;
+		}
 		up_read(&qp->state_lock);
-		*bad_wr = wr;
-		siw_dbg_qp(qp, "QP out of state %d\n", qp->attrs.state);
-		return -ENOTCONN;
-	}
-	if (wr && !qp->kernel_verbs) {
-		siw_dbg_qp(qp, "wr must be empty for user mapped sq\n");
-		up_read(&qp->state_lock);
-		*bad_wr = wr;
-		return -EINVAL;
+		return rv;
 	}
 	spin_lock_irqsave(&qp->sq_lock, flags);
 
@@ -917,24 +986,55 @@ int siw_post_receive(struct ib_qp *base_qp, const struct ib_recv_wr *wr,
 		*bad_wr = wr;
 		return -EOPNOTSUPP; /* what else from errno.h? */
 	}
-	/*
-	 * Try to acquire QP state lock. Must be non-blocking
-	 * to accommodate kernel clients needs.
-	 */
-	if (!down_read_trylock(&qp->state_lock)) {
-		*bad_wr = wr;
-		return -ENOTCONN;
-	}
 	if (!qp->kernel_verbs) {
 		siw_dbg_qp(qp, "no kernel post_recv for user mapped sq\n");
 		up_read(&qp->state_lock);
 		*bad_wr = wr;
 		return -EINVAL;
 	}
+
+	/*
+	 * Try to acquire QP state lock. Must be non-blocking
+	 * to accommodate kernel clients needs.
+	 */
+	if (!down_read_trylock(&qp->state_lock)) {
+		if (qp->attrs.state == SIW_QP_STATE_ERROR) {
+			/*
+			 * ERROR state is final, so we can be sure
+			 * this state will not change as long as the QP
+			 * exists.
+			 *
+			 * This handles an ib_drain_rq() call with
+			 * a concurrent request to set the QP state
+			 * to ERROR.
+			 */
+			rv = siw_rq_flush_wr(qp, wr, bad_wr);
+		} else {
+			siw_dbg_qp(qp, "QP locked, state %d\n",
+				   qp->attrs.state);
+			*bad_wr = wr;
+			rv = -ENOTCONN;
+		}
+		return rv;
+	}
 	if (qp->attrs.state > SIW_QP_STATE_RTS) {
+		if (qp->attrs.state == SIW_QP_STATE_ERROR) {
+			/*
+			 * Immediately flush this WR to CQ, if QP
+			 * is in ERROR state. RQ is guaranteed to
+			 * be empty, so WR complets in-order.
+			 *
+			 * Typically triggered by ib_drain_rq().
+			 */
+			rv = siw_rq_flush_wr(qp, wr, bad_wr);
+		} else {
+			siw_dbg_qp(qp, "QP out of state %d\n",
+				   qp->attrs.state);
+			*bad_wr = wr;
+			rv = -ENOTCONN;
+		}
 		up_read(&qp->state_lock);
-		*bad_wr = wr;
-		return -EINVAL;
+		return rv;
 	}
 	/*
 	 * Serialize potentially multiple producers.
-- 
2.20.1

[PATCH AUTOSEL 5.4 171/350] RDMA/qedr: Fix memory leak in user qp and mr

[Sasha Levin]

From: Michal Kalderon <michal.kalderon@marvell.com>

[ Upstream commit 24e412c1e00ebfe73619e6b88cbc26c2c7d41b85 ]

User QPs pbl's weren't freed properly.
MR pbls weren't freed properly.

Fixes: e0290cce6ac0 ("qedr: Add support for memory registeration verbs")
Link: https://lore.kernel.org/r/20191027200451.28187-5-michal.kalderon@marvell.com
Signed-off-by: Ariel Elior <ariel.elior@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 | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/drivers/infiniband/hw/qedr/verbs.c b/drivers/infiniband/hw/qedr/verbs.c
index 6f3ce86019b73..a7ccca3c4f894 100644
--- a/drivers/infiniband/hw/qedr/verbs.c
+++ b/drivers/infiniband/hw/qedr/verbs.c
@@ -1577,6 +1577,14 @@ static void qedr_cleanup_user(struct qedr_dev *dev, struct qedr_qp *qp)
 
 	ib_umem_release(qp->urq.umem);
 	qp->urq.umem = NULL;
+
+	if (rdma_protocol_roce(&dev->ibdev, 1)) {
+		qedr_free_pbl(dev, &qp->usq.pbl_info, qp->usq.pbl_tbl);
+		qedr_free_pbl(dev, &qp->urq.pbl_info, qp->urq.pbl_tbl);
+	} else {
+		kfree(qp->usq.pbl_tbl);
+		kfree(qp->urq.pbl_tbl);
+	}
 }
 
 static int qedr_create_user_qp(struct qedr_dev *dev,
@@ -2673,8 +2681,8 @@ int qedr_dereg_mr(struct ib_mr *ib_mr, struct ib_udata *udata)
 
 	dev->ops->rdma_free_tid(dev->rdma_ctx, mr->hw_mr.itid);
 
-	if ((mr->type != QEDR_MR_DMA) && (mr->type != QEDR_MR_FRMR))
-		qedr_free_pbl(dev, &mr->info.pbl_info, mr->info.pbl_table);
+	if (mr->type != QEDR_MR_DMA)
+		free_mr_info(dev, &mr->info);
 
 	/* it could be user registered memory. */
 	ib_umem_release(mr->umem);
-- 
2.20.1

[PATCH AUTOSEL 5.4 172/350] RDMA/hns: Fix memory leak on 'context' on error return path

[Sasha Levin]

From: Colin Ian King <colin.king@canonical.com>

[ Upstream commit 994195e1537074f56df216a9309f6e366cb35b67 ]

Currently, the error return path when the call to function
dev->dfx->query_cqc_info fails will leak object 'context'. Fix this by
making the error return path via 'err' return return codes rather than
-EMSGSIZE, set ret appropriately for all error return paths and for the
memory leak now return via 'err' rather than just returning without
freeing context.

Link: https://lore.kernel.org/r/20191024131034.19989-1-colin.king@canonical.com
Addresses-Coverity: ("Resource leak")
Fixes: e1c9a0dc2939 ("RDMA/hns: Dump detailed driver-specific CQ")
Signed-off-by: Colin Ian King <colin.king@canonical.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/hns/hns_roce_restrack.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/drivers/infiniband/hw/hns/hns_roce_restrack.c b/drivers/infiniband/hw/hns/hns_roce_restrack.c
index 0a31d0a3d657c..06871731ac43a 100644
--- a/drivers/infiniband/hw/hns/hns_roce_restrack.c
+++ b/drivers/infiniband/hw/hns/hns_roce_restrack.c
@@ -98,11 +98,15 @@ static int hns_roce_fill_res_cq_entry(struct sk_buff *msg,
 		goto err;
 
 	table_attr = nla_nest_start(msg, RDMA_NLDEV_ATTR_DRIVER);
-	if (!table_attr)
+	if (!table_attr) {
+		ret = -EMSGSIZE;
 		goto err;
+	}
 
-	if (hns_roce_fill_cq(msg, context))
+	if (hns_roce_fill_cq(msg, context)) {
+		ret = -EMSGSIZE;
 		goto err_cancel_table;
+	}
 
 	nla_nest_end(msg, table_attr);
 	kfree(context);
@@ -113,7 +117,7 @@ static int hns_roce_fill_res_cq_entry(struct sk_buff *msg,
 	nla_nest_cancel(msg, table_attr);
 err:
 	kfree(context);
-	return -EMSGSIZE;
+	return ret;
 }
 
 int hns_roce_fill_res_entry(struct sk_buff *msg,
-- 
2.20.1

[PATCH AUTOSEL 5.4 173/350] RDMA/qedr: Fix srqs xarray initialization

[Sasha Levin]

From: Michal Kalderon <michal.kalderon@marvell.com>

[ Upstream commit 73ab512f720298aabe23b34110e3f6a8545b0ba5 ]

There was a missing initialization for the srqs xarray.
SRQs xarray can also be called from irq context when searching
for an element and uses the xa_XXX_irq apis, therefore should
be initialized with IRQ flags.

Fixes: 9fd15987ed27 ("qedr: Convert srqidr to XArray")
Link: https://lore.kernel.org/r/20191027200451.28187-2-michal.kalderon@marvell.com
Signed-off-by: Ariel Elior <ariel.elior@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/main.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/infiniband/hw/qedr/main.c b/drivers/infiniband/hw/qedr/main.c
index dc71b6e16a07f..b462eaca1ee3c 100644
--- a/drivers/infiniband/hw/qedr/main.c
+++ b/drivers/infiniband/hw/qedr/main.c
@@ -357,6 +357,7 @@ static int qedr_alloc_resources(struct qedr_dev *dev)
 		return -ENOMEM;
 
 	spin_lock_init(&dev->sgid_lock);
+	xa_init_flags(&dev->srqs, XA_FLAGS_LOCK_IRQ);
 
 	if (IS_IWARP(dev)) {
 		xa_init_flags(&dev->qps, XA_FLAGS_LOCK_IRQ);
-- 
2.20.1

[PATCH AUTOSEL 5.4 174/350] RDMA/core: Set DMA parameters correctly

[Sasha Levin]

From: Bart Van Assche <bvanassche@acm.org>

[ Upstream commit c9121262d57b8a3be4f08073546436ba0128ca6a ]

The dma_set_max_seg_size() call in setup_dma_device() does not have any
effect since device->dev.dma_parms is NULL. Fix this by initializing
device->dev.dma_parms first.

Link: https://lore.kernel.org/r/20191025225830.257535-5-bvanassche@acm.org
Fixes: d10bcf947a3e ("RDMA/umem: Combine contiguous PAGE_SIZE regions in SGEs")
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
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/device.c | 16 ++++++++++++++--
 1 file changed, 14 insertions(+), 2 deletions(-)

diff --git a/drivers/infiniband/core/device.c b/drivers/infiniband/core/device.c
index 50a92442c4f7c..e6327d8f5b79a 100644
--- a/drivers/infiniband/core/device.c
+++ b/drivers/infiniband/core/device.c
@@ -1199,9 +1199,21 @@ static void setup_dma_device(struct ib_device *device)
 		WARN_ON_ONCE(!parent);
 		device->dma_device = parent;
 	}
-	/* Setup default max segment size for all IB devices */
-	dma_set_max_seg_size(device->dma_device, SZ_2G);
 
+	if (!device->dev.dma_parms) {
+		if (parent) {
+			/*
+			 * The caller did not provide DMA parameters, so
+			 * 'parent' probably represents a PCI device. The PCI
+			 * core sets the maximum segment size to 64
+			 * KB. Increase this parameter to 2 GB.
+			 */
+			device->dev.dma_parms = parent->dma_parms;
+			dma_set_max_seg_size(device->dma_device, SZ_2G);
+		} else {
+			WARN_ON_ONCE(true);
+		}
+	}
 }
 
 /*
-- 
2.20.1

[PATCH AUTOSEL 5.4 189/350] net/mlx5e: Verify that rule has at least one fwd/drop action

[Sasha Levin]

From: Vlad Buslov <vladbu@mellanox.com>

[ Upstream commit ae2741e2b6ce2bf1b656b1152c4ef147ff35b096 ]

Currently, mlx5 tc layer doesn't verify that rule has at least one forward
or drop action which leads to following firmware syndrome when user tries
to offload such action:

[ 1824.860501] mlx5_core 0000:81:00.0: mlx5_cmd_check:753:(pid 29458): SET_FLOW_TABLE_ENTRY(0x936) op_mod(0x0) failed, status bad parameter(0x3), syndrome (0x144b7a)

Add check at the end of parse_tc_fdb_actions() that verifies that resulting
attribute has action fwd or drop flag set.

Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
Reviewed-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
index f90a9f8e0fc6a..1e8bffebc4cff 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -3443,6 +3443,12 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv,
 		attr->action |= MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
 	}
 
+	if (!(attr->action &
+	      (MLX5_FLOW_CONTEXT_ACTION_FWD_DEST | MLX5_FLOW_CONTEXT_ACTION_DROP))) {
+		NL_SET_ERR_MSG(extack, "Rule must have at least one forward/drop action");
+		return -EOPNOTSUPP;
+	}
+
 	if (attr->split_count > 0 && !mlx5_esw_has_fwd_fdb(priv->mdev)) {
 		NL_SET_ERR_MSG_MOD(extack,
 				   "current firmware doesn't support split rule for port mirroring");
-- 
2.20.1

[PATCH AUTOSEL 5.4 215/350] RDMA/core: Fix return code when modify_port isn't supported

[Sasha Levin]

From: Kamal Heib <kamalheib1@gmail.com>

[ Upstream commit 55bfe905fa97633438c13fb029aed85371d85480 ]

Improve return code from ib_modify_port() by doing the following:
 - Use "-EOPNOTSUPP" instead "-ENOSYS" which is the proper return code

 - Allow only fake IB_PORT_CM_SUP manipulation for RoCE providers that
   didn't implement the modify_port callback, otherwise return
   "-EOPNOTSUPP"

Fixes: 61e0962d5221 ("IB: Avoid ib_modify_port() failure for RoCE devices")
Link: https://lore.kernel.org/r/20191028155931.1114-2-kamalheib1@gmail.com
Signed-off-by: Kamal Heib <kamalheib1@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/core/device.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/infiniband/core/device.c b/drivers/infiniband/core/device.c
index e6327d8f5b79a..2b5bd7206fc6e 100644
--- a/drivers/infiniband/core/device.c
+++ b/drivers/infiniband/core/device.c
@@ -2409,8 +2409,12 @@ int ib_modify_port(struct ib_device *device,
 		rc = device->ops.modify_port(device, port_num,
 					     port_modify_mask,
 					     port_modify);
+	else if (rdma_protocol_roce(device, port_num) &&
+		 ((port_modify->set_port_cap_mask & ~IB_PORT_CM_SUP) == 0 ||
+		  (port_modify->clr_port_cap_mask & ~IB_PORT_CM_SUP) == 0))
+		rc = 0;
 	else
-		rc = rdma_protocol_roce(device, port_num) ? 0 : -ENOSYS;
+		rc = -EOPNOTSUPP;
 	return rc;
 }
 EXPORT_SYMBOL(ib_modify_port);
-- 
2.20.1

[PATCH AUTOSEL 5.4 285/350] RDMA/qib: Validate ->show()/store() callbacks before calling them

[Sasha Levin]

From: Viresh Kumar <viresh.kumar@linaro.org>

[ Upstream commit 7ee23491b39259ae83899dd93b2a29ef0f22f0a7 ]

The permissions of the read-only or write-only sysfs files can be
changed (as root) and the user can then try to read a write-only file or
write to a read-only file which will lead to kernel crash here.

Protect against that by always validating the show/store callbacks.

Link: https://lore.kernel.org/r/d45cc26361a174ae12dbb86c994ef334d257924b.1573096807.git.viresh.kumar@linaro.org
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/hw/qib/qib_sysfs.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/drivers/infiniband/hw/qib/qib_sysfs.c b/drivers/infiniband/hw/qib/qib_sysfs.c
index 3926be78036ee..568b21eb6ea15 100644
--- a/drivers/infiniband/hw/qib/qib_sysfs.c
+++ b/drivers/infiniband/hw/qib/qib_sysfs.c
@@ -301,6 +301,9 @@ static ssize_t qib_portattr_show(struct kobject *kobj,
 	struct qib_pportdata *ppd =
 		container_of(kobj, struct qib_pportdata, pport_kobj);
 
+	if (!pattr->show)
+		return -EIO;
+
 	return pattr->show(ppd, buf);
 }
 
@@ -312,6 +315,9 @@ static ssize_t qib_portattr_store(struct kobject *kobj,
 	struct qib_pportdata *ppd =
 		container_of(kobj, struct qib_pportdata, pport_kobj);
 
+	if (!pattr->store)
+		return -EIO;
+
 	return pattr->store(ppd, buf, len);
 }
 
-- 
2.20.1

[PATCH AUTOSEL 5.4 286/350] RDMA/efa: Clear the admin command buffer prior to its submission

[Sasha Levin]

From: Gal Pressman <galpress@amazon.com>

[ Upstream commit 64c264872b8879e2ab9017eefe9514d4c045c60e ]

We cannot rely on the entry memcpy as we only copy the actual size of the
command, the rest of the bytes must be memset to zero.

Currently providing non-zero memory will not have any user visible impact.
However, since admin commands are extendable (in a backwards compatible
way) everything beyond the size of the command must be cleared to prevent
issues in the future.

Fixes: 0420e542569b ("RDMA/efa: Implement functions that submit and complete admin commands")
Link: https://lore.kernel.org/r/20191112092608.46964-1-galpress@amazon.com
Reviewed-by: Daniel Kranzdorf <dkkranzd@amazon.com>
Reviewed-by: Firas JahJah <firasj@amazon.com>
Signed-off-by: Gal Pressman <galpress@amazon.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/hw/efa/efa_com.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/infiniband/hw/efa/efa_com.c b/drivers/infiniband/hw/efa/efa_com.c
index 3c412bc5b94f1..0778f4f7dccd7 100644
--- a/drivers/infiniband/hw/efa/efa_com.c
+++ b/drivers/infiniband/hw/efa/efa_com.c
@@ -317,6 +317,7 @@ static struct efa_comp_ctx *__efa_com_submit_admin_cmd(struct efa_com_admin_queu
 						       struct efa_admin_acq_entry *comp,
 						       size_t comp_size_in_bytes)
 {
+	struct efa_admin_aq_entry *aqe;
 	struct efa_comp_ctx *comp_ctx;
 	u16 queue_size_mask;
 	u16 cmd_id;
@@ -350,7 +351,9 @@ static struct efa_comp_ctx *__efa_com_submit_admin_cmd(struct efa_com_admin_queu
 
 	reinit_completion(&comp_ctx->wait_event);
 
-	memcpy(&aq->sq.entries[pi], cmd, cmd_size_in_bytes);
+	aqe = &aq->sq.entries[pi];
+	memset(aqe, 0, sizeof(*aqe));
+	memcpy(aqe, cmd, cmd_size_in_bytes);
 
 	aq->sq.pc++;
 	atomic64_inc(&aq->stats.submitted_cmd);
-- 
2.20.1

[PATCH AUTOSEL 5.4 344/350] RDMA/bnxt_re: Fix missing le16_to_cpu

[Sasha Levin]

From: Devesh Sharma <devesh.sharma@broadcom.com>

[ Upstream commit fca5b9dc0986aa49b3f0a7cfe24b6c82422ac1d7 ]

From sparse:

drivers/infiniband/hw/bnxt_re/main.c:1274:18: warning: cast from restricted __le16
drivers/infiniband/hw/bnxt_re/main.c:1275:18: warning: cast from restricted __le16
drivers/infiniband/hw/bnxt_re/main.c:1276:18: warning: cast from restricted __le16
drivers/infiniband/hw/bnxt_re/main.c:1277:21: warning: restricted __le16 degrades to integer

Fixes: 2b827ea1926b ("RDMA/bnxt_re: Query HWRM Interface version from FW")
Link: https://lore.kernel.org/r/1574317343-23300-4-git-send-email-devesh.sharma@broadcom.com
Signed-off-by: Devesh Sharma <devesh.sharma@broadcom.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/hw/bnxt_re/main.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/infiniband/hw/bnxt_re/main.c b/drivers/infiniband/hw/bnxt_re/main.c
index 30a54f8aa42c0..b31e215882004 100644
--- a/drivers/infiniband/hw/bnxt_re/main.c
+++ b/drivers/infiniband/hw/bnxt_re/main.c
@@ -1270,10 +1270,10 @@ static void bnxt_re_query_hwrm_intf_version(struct bnxt_re_dev *rdev)
 		return;
 	}
 	rdev->qplib_ctx.hwrm_intf_ver =
-		(u64)resp.hwrm_intf_major << 48 |
-		(u64)resp.hwrm_intf_minor << 32 |
-		(u64)resp.hwrm_intf_build << 16 |
-		resp.hwrm_intf_patch;
+		(u64)le16_to_cpu(resp.hwrm_intf_major) << 48 |
+		(u64)le16_to_cpu(resp.hwrm_intf_minor) << 32 |
+		(u64)le16_to_cpu(resp.hwrm_intf_build) << 16 |
+		le16_to_cpu(resp.hwrm_intf_patch);
 }
 
 static void bnxt_re_ib_unreg(struct bnxt_re_dev *rdev)
-- 
2.20.1

[PATCH AUTOSEL 5.4 345/350] RDMA/bnxt_re: Fix stat push into dma buffer on gen p5 devices

[Sasha Levin]

From: Devesh Sharma <devesh.sharma@broadcom.com>

[ Upstream commit 98998ffe5216c7fa2c0225bb5b049ca5cdf8d195 ]

Due to recent advances in the firmware for Broadcom's gen p5 series of
adaptors the driver code to report hardware counters has been broken
w.r.t. roce devices.

The new firmware command expects dma length to be specified during stat
dma buffer allocation.

Fixes: 2792b5b95ed5 ("bnxt_en: Update firmware interface spec. to 1.10.0.89.")
Link: https://lore.kernel.org/r/1574317343-23300-3-git-send-email-devesh.sharma@broadcom.com
Signed-off-by: Devesh Sharma <devesh.sharma@broadcom.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/hw/bnxt_re/main.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/infiniband/hw/bnxt_re/main.c b/drivers/infiniband/hw/bnxt_re/main.c
index b31e215882004..27e2df44d043d 100644
--- a/drivers/infiniband/hw/bnxt_re/main.c
+++ b/drivers/infiniband/hw/bnxt_re/main.c
@@ -477,6 +477,7 @@ static int bnxt_re_net_stats_ctx_alloc(struct bnxt_re_dev *rdev,
 	bnxt_re_init_hwrm_hdr(rdev, (void *)&req, HWRM_STAT_CTX_ALLOC, -1, -1);
 	req.update_period_ms = cpu_to_le32(1000);
 	req.stats_dma_addr = cpu_to_le64(dma_map);
+	req.stats_dma_length = cpu_to_le16(sizeof(struct ctx_hw_stats_ext));
 	req.stat_ctx_flags = STAT_CTX_ALLOC_REQ_STAT_CTX_FLAGS_ROCE;
 	bnxt_re_fill_fw_msg(&fw_msg, (void *)&req, sizeof(req), (void *)&resp,
 			    sizeof(resp), DFLT_HWRM_CMD_TIMEOUT);
-- 
2.20.1

[PATCH AUTOSEL 5.4 347/350] RDMA/bnxt_re: Fix chip number validation Broadcom's Gen P5 series

[Sasha Levin]

From: Luke Starrett <luke.starrett@broadcom.com>

[ Upstream commit e284b159c6881c8bec9713daba2653268f4c4948 ]

In the first version of Gen P5 ASIC, chip-id was always set to 0x1750 for
all adaptor port configurations. This has been fixed in the new chip rev.

Due to this missing fix users are not able to use adaptors based on latest
chip rev of Broadcom's Gen P5 adaptors.

Fixes: ae8637e13185 ("RDMA/bnxt_re: Add chip context to identify 57500 series")
Link: https://lore.kernel.org/r/1574317343-23300-2-git-send-email-devesh.sharma@broadcom.com
Signed-off-by: Naresh Kumar PBS <nareshkumar.pbs@broadcom.com>
Signed-off-by: Selvin Xavier <selvin.xavier@broadcom.com>
Signed-off-by: Luke Starrett <luke.starrett@broadcom.com>
Signed-off-by: Devesh Sharma <devesh.sharma@broadcom.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/hw/bnxt_re/qplib_res.h | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/drivers/infiniband/hw/bnxt_re/qplib_res.h b/drivers/infiniband/hw/bnxt_re/qplib_res.h
index fbda11a7ab1aa..aaa76d7921857 100644
--- a/drivers/infiniband/hw/bnxt_re/qplib_res.h
+++ b/drivers/infiniband/hw/bnxt_re/qplib_res.h
@@ -186,7 +186,9 @@ struct bnxt_qplib_chip_ctx {
 	u8	chip_metal;
 };
 
-#define CHIP_NUM_57500          0x1750
+#define CHIP_NUM_57508		0x1750
+#define CHIP_NUM_57504		0x1751
+#define CHIP_NUM_57502		0x1752
 
 struct bnxt_qplib_res {
 	struct pci_dev			*pdev;
@@ -203,7 +205,9 @@ struct bnxt_qplib_res {
 
 static inline bool bnxt_qplib_is_chip_gen_p5(struct bnxt_qplib_chip_ctx *cctx)
 {
-	return (cctx->chip_num == CHIP_NUM_57500);
+	return (cctx->chip_num == CHIP_NUM_57508 ||
+		cctx->chip_num == CHIP_NUM_57504 ||
+		cctx->chip_num == CHIP_NUM_57502);
 }
 
 static inline u8 bnxt_qplib_get_hwq_type(struct bnxt_qplib_res *res)
-- 
2.20.1