[pull request][rdma-next 0/8] Export device and port properties

[pull request][rdma-next 0/8] Export device and port properties

[Leon Romanovsky]

Hi,

This is the third part of RDMAtool infrastructure changes. In this series,
I'm exporting various device and port properties which are parsed in
RDMAtool.

Due to the fact, that the "IB/hfi1: Document phys port state bits not
used in IB", I embed its content into the actual patches.

As part of this series, I performed small code cleanup to improve our
static checker situation.

Thanks

Cc: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Cc: Jason Gunthorpe <jgunthorpe-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
----------------------------------------------------------------
The following changes since commit 30307f91febf1b99947307a479c0a82c3af72f93:

  Documentation: Hardware tag matching (2017-08-15 11:46:45 +0300)

are available in the git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/leon/linux-rdma.git tags/rdma-next-2017-08-15-2

for you to fetch changes up to 9335303fee9bac70b5d5c143126b846d20df56d6:

  RDMA: Use defines instead of hard coded value for phys_state (2017-08-15 11:46:53 +0300)

----------------------------------------------------------------
Export port properties, physical state and node types to the UAPI

----------------------------------------------------------------
Leon Romanovsky (8):
      RDMA/mlx4: Properly annotate link layer variable
      RDMA/core: Export device capabilities flags
      RDMA/core: Provide node type flags
      RDMA/core: Move port capabilities to UAPI
      RDMA/core: Move port state to UAPI and rename it to be link
      RDMA/core: Add physical link state information to the UAPI
      RDMA/nes: Remove zeroed parameter from port query callback
      RDMA: Use defines instead of hard coded value for phys_state

 drivers/infiniband/core/sysfs.c              |  24 ++--
 drivers/infiniband/hw/bnxt_re/ib_verbs.c     |   4 +-
 drivers/infiniband/hw/hfi1/opa_compat.h      |  30 ++--
 drivers/infiniband/hw/hns/hns_roce_main.c    |   4 +-
 drivers/infiniband/hw/mlx4/main.c            |   4 +-
 drivers/infiniband/hw/mlx4/qp.c              |   6 +-
 drivers/infiniband/hw/mlx5/main.c            |   4 +-
 drivers/infiniband/hw/nes/nes_verbs.c        |   5 -
 drivers/infiniband/hw/ocrdma/ocrdma_verbs.c  |   4 +-
 drivers/infiniband/hw/qedr/verbs.c           |   4 +-
 drivers/infiniband/hw/usnic/usnic_ib_verbs.c |   6 +-
 drivers/infiniband/sw/rxe/rxe_param.h        |   2 +-
 include/rdma/ib_verbs.h                      | 135 +++++++++---------
 include/uapi/rdma/ib_user_verbs.h            |   1 +
 include/uapi/rdma/rdma.h                     | 196 +++++++++++++++++++++++++++
 15 files changed, 314 insertions(+), 115 deletions(-)
 create mode 100644 include/uapi/rdma/rdma.h
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[rdma-next 1/8] RDMA/mlx4: Properly annotate link layer variable

[Leon Romanovsky]

From: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

The rdma_port_get_link_layer() returns enum rdma_link_layer as
a return value, hence it is better to store the return value in
specially annotated variable and not in int.

Signed-off-by: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
 drivers/infiniband/hw/mlx4/qp.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/infiniband/hw/mlx4/qp.c b/drivers/infiniband/hw/mlx4/qp.c
index 2874d8790570..8c6ccc392d87 100644
--- a/drivers/infiniband/hw/mlx4/qp.c
+++ b/drivers/infiniband/hw/mlx4/qp.c
@@ -2732,19 +2732,17 @@ enum {
 static int _mlx4_ib_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
 			      int attr_mask, struct ib_udata *udata)
 {
+	enum rdma_link_layer ll = IB_LINK_LAYER_UNSPECIFIED;
 	struct mlx4_ib_dev *dev = to_mdev(ibqp->device);
 	struct mlx4_ib_qp *qp = to_mqp(ibqp);
 	enum ib_qp_state cur_state, new_state;
 	int err = -EINVAL;
-	int ll;
 	mutex_lock(&qp->mutex);
 
 	cur_state = attr_mask & IB_QP_CUR_STATE ? attr->cur_qp_state : qp->state;
 	new_state = attr_mask & IB_QP_STATE ? attr->qp_state : cur_state;
 
-	if (cur_state == new_state && cur_state == IB_QPS_RESET) {
-		ll = IB_LINK_LAYER_UNSPECIFIED;
-	} else {
+	if (cur_state != new_state || cur_state != IB_QPS_RESET) {
 		int port = attr_mask & IB_QP_PORT ? attr->port_num : qp->port;
 		ll = rdma_port_get_link_layer(&dev->ib_dev, port);
 	}
-- 
2.14.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[rdma-next 2/8] RDMA/core: Export device capabilities flags

[Leon Romanovsky]

From: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

Provide to the user space application the ability
to understand device capabilities.

Signed-off-by: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
 include/rdma/ib_verbs.h           |  68 ++++++++++++------------
 include/uapi/rdma/ib_user_verbs.h |   1 +
 include/uapi/rdma/rdma.h          | 106 ++++++++++++++++++++++++++++++++++++++
 3 files changed, 140 insertions(+), 35 deletions(-)
 create mode 100644 include/uapi/rdma/rdma.h

diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index 80ede01d2237..bfedcda04ef7 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -157,22 +157,21 @@ enum rdma_link_layer {
 };
 
 enum ib_device_cap_flags {
-	IB_DEVICE_RESIZE_MAX_WR			= (1 << 0),
-	IB_DEVICE_BAD_PKEY_CNTR			= (1 << 1),
-	IB_DEVICE_BAD_QKEY_CNTR			= (1 << 2),
-	IB_DEVICE_RAW_MULTI			= (1 << 3),
-	IB_DEVICE_AUTO_PATH_MIG			= (1 << 4),
-	IB_DEVICE_CHANGE_PHY_PORT		= (1 << 5),
-	IB_DEVICE_UD_AV_PORT_ENFORCE		= (1 << 6),
-	IB_DEVICE_CURR_QP_STATE_MOD		= (1 << 7),
-	IB_DEVICE_SHUTDOWN_PORT			= (1 << 8),
+	IB_DEVICE_RESIZE_MAX_WR			= RDMA_DEV_RESIZE_MAX_WR,
+	IB_DEVICE_BAD_PKEY_CNTR			= RDMA_DEV_BAD_PKEY_CNTR,
+	IB_DEVICE_BAD_QKEY_CNTR			= RDMA_DEV_BAD_QKEY_CNTR,
+	IB_DEVICE_RAW_MULTI			= RDMA_DEV_RAW_MULTI,
+	IB_DEVICE_AUTO_PATH_MIG			= RDMA_DEV_AUTO_PATH_MIG,
+	IB_DEVICE_CHANGE_PHY_PORT		= RDMA_DEV_CHANGE_PHY_PORT,
+	IB_DEVICE_UD_AV_PORT_ENFORCE		= RDMA_DEV_UD_AV_PORT_ENFORCE,
+	IB_DEVICE_CURR_QP_STATE_MOD		= RDMA_DEV_CURR_QP_STATE_MOD,
+	IB_DEVICE_SHUTDOWN_PORT			= RDMA_DEV_SHUTDOWN_PORT,
 	/* Not in use, former INIT_TYPE		= (1 << 9),*/
-	IB_DEVICE_PORT_ACTIVE_EVENT		= (1 << 10),
-	IB_DEVICE_SYS_IMAGE_GUID		= (1 << 11),
-	IB_DEVICE_RC_RNR_NAK_GEN		= (1 << 12),
-	IB_DEVICE_SRQ_RESIZE			= (1 << 13),
-	IB_DEVICE_N_NOTIFY_CQ			= (1 << 14),
-
+	IB_DEVICE_PORT_ACTIVE_EVENT		= RDMA_DEV_PORT_ACTIVE_EVENT,
+	IB_DEVICE_SYS_IMAGE_GUID		= RDMA_DEV_SYS_IMAGE_GUID,
+	IB_DEVICE_RC_RNR_NAK_GEN		= RDMA_DEV_RC_RNR_NAK_GEN,
+	IB_DEVICE_SRQ_RESIZE			= RDMA_DEV_SRQ_RESIZE,
+	IB_DEVICE_N_NOTIFY_CQ			= RDMA_DEV_N_NOTIFY_CQ,
 	/*
 	 * This device supports a per-device lkey or stag that can be
 	 * used without performing a memory registration for the local
@@ -180,9 +179,9 @@ enum ib_device_cap_flags {
 	 * instead of use the local_dma_lkey flag in the ib_pd structure,
 	 * which will always contain a usable lkey.
 	 */
-	IB_DEVICE_LOCAL_DMA_LKEY		= (1 << 15),
+	IB_DEVICE_LOCAL_DMA_LKEY		= RDMA_DEV_LOCAL_DMA_LKEY,
 	/* Reserved, old SEND_W_INV		= (1 << 16),*/
-	IB_DEVICE_MEM_WINDOW			= (1 << 17),
+	IB_DEVICE_MEM_WINDOW			= RDMA_DEV_MEM_WINDOW,
 	/*
 	 * Devices should set IB_DEVICE_UD_IP_SUM if they support
 	 * insertion of UDP and TCP checksum on outgoing UD IPoIB
@@ -190,10 +189,9 @@ enum ib_device_cap_flags {
 	 * incoming messages.  Setting this flag implies that the
 	 * IPoIB driver may set NETIF_F_IP_CSUM for datagram mode.
 	 */
-	IB_DEVICE_UD_IP_CSUM			= (1 << 18),
-	IB_DEVICE_UD_TSO			= (1 << 19),
-	IB_DEVICE_XRC				= (1 << 20),
-
+	IB_DEVICE_UD_IP_CSUM			= RDMA_DEV_UD_IP_CSUM,
+	IB_DEVICE_UD_TSO			= RDMA_DEV_UD_TSO,
+	IB_DEVICE_XRC				= RDMA_DEV_XRC,
 	/*
 	 * This device supports the IB "base memory management extension",
 	 * which includes support for fast registrations (IB_WR_REG_MR,
@@ -203,28 +201,28 @@ enum ib_device_cap_flags {
 	 * IB_WR_RDMA_READ_WITH_INV verb for RDMA READs that invalidate the
 	 * stag.
 	 */
-	IB_DEVICE_MEM_MGT_EXTENSIONS		= (1 << 21),
-	IB_DEVICE_BLOCK_MULTICAST_LOOPBACK	= (1 << 22),
-	IB_DEVICE_MEM_WINDOW_TYPE_2A		= (1 << 23),
-	IB_DEVICE_MEM_WINDOW_TYPE_2B		= (1 << 24),
-	IB_DEVICE_RC_IP_CSUM			= (1 << 25),
+	IB_DEVICE_MEM_MGT_EXTENSIONS		= RDMA_DEV_MEM_MGT_EXTENSIONS,
+	IB_DEVICE_BLOCK_MULTICAST_LOOPBACK	= RDMA_DEV_BLOCK_MULTICAST_LOOPBACK,
+	IB_DEVICE_MEM_WINDOW_TYPE_2A		= RDMA_DEV_MEM_WINDOW_TYPE_2A,
+	IB_DEVICE_MEM_WINDOW_TYPE_2B		= RDMA_DEV_MEM_WINDOW_TYPE_2B,
+	IB_DEVICE_RC_IP_CSUM			= RDMA_DEV_RC_IP_CSUM,
 	/* Deprecated. Please use IB_RAW_PACKET_CAP_IP_CSUM. */
-	IB_DEVICE_RAW_IP_CSUM			= (1 << 26),
+	IB_DEVICE_RAW_IP_CSUM			= RDMA_DEV_RAW_IP_CSUM,
 	/*
 	 * Devices should set IB_DEVICE_CROSS_CHANNEL if they
 	 * support execution of WQEs that involve synchronization
 	 * of I/O operations with single completion queue managed
 	 * by hardware.
 	 */
-	IB_DEVICE_CROSS_CHANNEL			= (1 << 27),
-	IB_DEVICE_MANAGED_FLOW_STEERING		= (1 << 29),
-	IB_DEVICE_SIGNATURE_HANDOVER		= (1 << 30),
-	IB_DEVICE_ON_DEMAND_PAGING		= (1ULL << 31),
-	IB_DEVICE_SG_GAPS_REG			= (1ULL << 32),
-	IB_DEVICE_VIRTUAL_FUNCTION		= (1ULL << 33),
+	IB_DEVICE_CROSS_CHANNEL			= RDMA_DEV_CROSS_CHANNEL,
+	IB_DEVICE_MANAGED_FLOW_STEERING		= RDMA_DEV_MANAGED_FLOW_STEERING,
+	IB_DEVICE_SIGNATURE_HANDOVER		= RDMA_DEV_SIGNATURE_HANDOVER,
+	IB_DEVICE_ON_DEMAND_PAGING		= RDMA_DEV_ON_DEMAND_PAGING,
+	IB_DEVICE_SG_GAPS_REG			= RDMA_DEV_SG_GAPS_REG,
+	IB_DEVICE_VIRTUAL_FUNCTION		= RDMA_DEV_VIRTUAL_FUNCTION,
 	/* Deprecated. Please use IB_RAW_PACKET_CAP_SCATTER_FCS. */
-	IB_DEVICE_RAW_SCATTER_FCS		= (1ULL << 34),
-	IB_DEVICE_RDMA_NETDEV_OPA_VNIC		= (1ULL << 35),
+	IB_DEVICE_RAW_SCATTER_FCS		= RDMA_DEV_RAW_SCATTER_FCS,
+	IB_DEVICE_RDMA_NETDEV_OPA_VNIC		= RDMA_DEV_RDMA_NETDEV_OPA_VNIC,
 };
 
 enum ib_signature_prot_cap {
diff --git a/include/uapi/rdma/ib_user_verbs.h b/include/uapi/rdma/ib_user_verbs.h
index 9a0b6479fe0c..803106ffa93c 100644
--- a/include/uapi/rdma/ib_user_verbs.h
+++ b/include/uapi/rdma/ib_user_verbs.h
@@ -37,6 +37,7 @@
 #define IB_USER_VERBS_H
 
 #include <linux/types.h>
+#include <rdma/rdma.h>
 
 /*
  * Increment this value if any changes that break userspace ABI
diff --git a/include/uapi/rdma/rdma.h b/include/uapi/rdma/rdma.h
new file mode 100644
index 000000000000..dcdf61474118
--- /dev/null
+++ b/include/uapi/rdma/rdma.h
@@ -0,0 +1,106 @@
+/*
+ * Copyright (c) 2017 Mellanox Technologies.  All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef _RDMA_H
+#define _RDMA_H
+
+enum rdma_dev_cap {
+	RDMA_DEV_RESIZE_MAX_WR			= (1 << 0),
+	RDMA_DEV_BAD_PKEY_CNTR			= (1 << 1),
+	RDMA_DEV_BAD_QKEY_CNTR			= (1 << 2),
+	RDMA_DEV_RAW_MULTI			= (1 << 3),
+	RDMA_DEV_AUTO_PATH_MIG			= (1 << 4),
+	RDMA_DEV_CHANGE_PHY_PORT		= (1 << 5),
+	RDMA_DEV_UD_AV_PORT_ENFORCE		= (1 << 6),
+	RDMA_DEV_CURR_QP_STATE_MOD		= (1 << 7),
+	RDMA_DEV_SHUTDOWN_PORT			= (1 << 8),
+	/* Not in use, former INIT_TYPE		= (1 << 9),*/
+	RDMA_DEV_PORT_ACTIVE_EVENT		= (1 << 10),
+	RDMA_DEV_SYS_IMAGE_GUID			= (1 << 11),
+	RDMA_DEV_RC_RNR_NAK_GEN			= (1 << 12),
+	RDMA_DEV_SRQ_RESIZE			= (1 << 13),
+	RDMA_DEV_N_NOTIFY_CQ			= (1 << 14),
+
+	/*
+	 * This device supports a per-device lkey or stag that can be
+	 * used without performing a memory registration for the local
+	 * memory.  Note that ULPs should never check this flag, but
+	 * instead of use the local_dma_lkey flag in the ib_pd structure,
+	 * which will always contain a usable lkey.
+	 */
+	RDMA_DEV_LOCAL_DMA_LKEY			= (1 << 15),
+	/* Reserved, old SEND_W_INV		= (1 << 16),*/
+	RDMA_DEV_MEM_WINDOW			= (1 << 17),
+	/*
+	 * Devices should set RDMA_DEV_UD_IP_SUM if they support
+	 * insertion of UDP and TCP checksum on outgoing UD IPoIB
+	 * messages and can verify the validity of checksum for
+	 * incoming messages.  Setting this flag implies that the
+	 * IPoIB driver may set NETIF_F_IP_CSUM for datagram mode.
+	 */
+	RDMA_DEV_UD_IP_CSUM			= (1 << 18),
+	RDMA_DEV_UD_TSO				= (1 << 19),
+	RDMA_DEV_XRC				= (1 << 20),
+
+	/*
+	 * This device supports the IB "base memory management extension",
+	 * which includes support for fast registrations (IB_WR_REG_MR,
+	 * IB_WR_LOCAL_INV and IB_WR_SEND_WITH_INV verbs).  This flag should
+	 * also be set by any iWarp device which must support FRs to comply
+	 * to the iWarp verbs spec.  iWarp devices also support the
+	 * IB_WR_RDMA_READ_WITH_INV verb for RDMA READs that invalidate the
+	 * stag.
+	 */
+	RDMA_DEV_MEM_MGT_EXTENSIONS		= (1 << 21),
+	RDMA_DEV_BLOCK_MULTICAST_LOOPBACK	= (1 << 22),
+	RDMA_DEV_MEM_WINDOW_TYPE_2A		= (1 << 23),
+	RDMA_DEV_MEM_WINDOW_TYPE_2B		= (1 << 24),
+	RDMA_DEV_RC_IP_CSUM			= (1 << 25),
+	/* Deprecated. Please use IB_RAW_PACKET_CAP_IP_CSUM. */
+	RDMA_DEV_RAW_IP_CSUM			= (1 << 26),
+	/*
+	 * Devices should set RDMA_DEV_CROSS_CHANNEL if they
+	 * support execution of WQEs that involve synchronization
+	 * of I/O operations with single completion queue managed
+	 * by hardware.
+	 */
+	RDMA_DEV_CROSS_CHANNEL			= (1 << 27),
+	RDMA_DEV_MANAGED_FLOW_STEERING		= (1 << 29),
+	RDMA_DEV_SIGNATURE_HANDOVER		= (1 << 30),
+	RDMA_DEV_ON_DEMAND_PAGING		= (1ULL << 31),
+	RDMA_DEV_SG_GAPS_REG			= (1ULL << 32),
+	RDMA_DEV_VIRTUAL_FUNCTION		= (1ULL << 33),
+	/* Deprecated. Please use IB_RAW_PACKET_CAP_SCATTER_FCS. */
+	RDMA_DEV_RAW_SCATTER_FCS		= (1ULL << 34),
+	RDMA_DEV_RDMA_NETDEV_OPA_VNIC		= (1ULL << 35),
+};
+#endif /* _RDMA_H */
-- 
2.14.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[rdma-next 3/8] RDMA/core: Provide node type flags

[Leon Romanovsky]

From: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

Expose node types for the user space applications.

Signed-off-by: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
 include/rdma/ib_verbs.h  | 7 -------
 include/uapi/rdma/rdma.h | 7 +++++++
 2 files changed, 7 insertions(+), 7 deletions(-)

diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index bfedcda04ef7..0aced267cc53 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -93,13 +93,6 @@ struct ib_gid_attr {
 	struct net_device	*ndev;
 };
 
-enum rdma_node_type {
-	/* IB values map to NodeInfo:NodeType. */
-	RDMA_NODE_IB_CA		= 1,
-	RDMA_NODE_RNIC		= 4,
-	RDMA_NODE_USNIC_UDP	= 6,
-};
-
 enum {
 	/* set the local administered indication */
 	IB_SA_WELL_KNOWN_GUID	= BIT_ULL(57) | 2,
diff --git a/include/uapi/rdma/rdma.h b/include/uapi/rdma/rdma.h
index dcdf61474118..75c1baad313b 100644
--- a/include/uapi/rdma/rdma.h
+++ b/include/uapi/rdma/rdma.h
@@ -103,4 +103,11 @@ enum rdma_dev_cap {
 	RDMA_DEV_RAW_SCATTER_FCS		= (1ULL << 34),
 	RDMA_DEV_RDMA_NETDEV_OPA_VNIC		= (1ULL << 35),
 };
+
+enum rdma_node_type {
+	/* IB values map to NodeInfo:NodeType. */
+	RDMA_NODE_IB_CA		= 1,
+	RDMA_NODE_RNIC		= 4,
+	RDMA_NODE_USNIC_UDP	= 6,
+};
 #endif /* _RDMA_H */
-- 
2.14.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[rdma-next 4/8] RDMA/core: Move port capabilities to UAPI

[Leon Romanovsky]

From: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

Provide to user space applications ability to reuse kernel
port capabilities enum.

Signed-off-by: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
 include/rdma/ib_verbs.h  | 48 ++++++++++++++++++++++++------------------------
 include/uapi/rdma/rdma.h | 33 +++++++++++++++++++++++++++++++++
 2 files changed, 57 insertions(+), 24 deletions(-)

diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index 0aced267cc53..a7cd54061f48 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -392,30 +392,30 @@ enum ib_port_state {
 };
 
 enum ib_port_cap_flags {
-	IB_PORT_SM				= 1 <<  1,
-	IB_PORT_NOTICE_SUP			= 1 <<  2,
-	IB_PORT_TRAP_SUP			= 1 <<  3,
-	IB_PORT_OPT_IPD_SUP                     = 1 <<  4,
-	IB_PORT_AUTO_MIGR_SUP			= 1 <<  5,
-	IB_PORT_SL_MAP_SUP			= 1 <<  6,
-	IB_PORT_MKEY_NVRAM			= 1 <<  7,
-	IB_PORT_PKEY_NVRAM			= 1 <<  8,
-	IB_PORT_LED_INFO_SUP			= 1 <<  9,
-	IB_PORT_SM_DISABLED			= 1 << 10,
-	IB_PORT_SYS_IMAGE_GUID_SUP		= 1 << 11,
-	IB_PORT_PKEY_SW_EXT_PORT_TRAP_SUP	= 1 << 12,
-	IB_PORT_EXTENDED_SPEEDS_SUP             = 1 << 14,
-	IB_PORT_CM_SUP				= 1 << 16,
-	IB_PORT_SNMP_TUNNEL_SUP			= 1 << 17,
-	IB_PORT_REINIT_SUP			= 1 << 18,
-	IB_PORT_DEVICE_MGMT_SUP			= 1 << 19,
-	IB_PORT_VENDOR_CLASS_SUP		= 1 << 20,
-	IB_PORT_DR_NOTICE_SUP			= 1 << 21,
-	IB_PORT_CAP_MASK_NOTICE_SUP		= 1 << 22,
-	IB_PORT_BOOT_MGMT_SUP			= 1 << 23,
-	IB_PORT_LINK_LATENCY_SUP		= 1 << 24,
-	IB_PORT_CLIENT_REG_SUP			= 1 << 25,
-	IB_PORT_IP_BASED_GIDS			= 1 << 26,
+	IB_PORT_SM				= RDMA_PORT_SM,
+	IB_PORT_NOTICE_SUP			= RDMA_PORT_NOTICE,
+	IB_PORT_TRAP_SUP			= RDMA_PORT_TRAP,
+	IB_PORT_OPT_IPD_SUP                     = RDMA_PORT_OPT_IPD,
+	IB_PORT_AUTO_MIGR_SUP			= RDMA_PORT_AUTO_MIGR,
+	IB_PORT_SL_MAP_SUP			= RDMA_PORT_SL_MAP,
+	IB_PORT_MKEY_NVRAM			= RDMA_PORT_MKEY_NVRAM,
+	IB_PORT_PKEY_NVRAM			= RDMA_PORT_PKEY_NVRAM,
+	IB_PORT_LED_INFO_SUP			= RDMA_PORT_LED_INFO,
+	IB_PORT_SM_DISABLED			= RDMA_PORT_SM_DISABLED,
+	IB_PORT_SYS_IMAGE_GUID_SUP		= RDMA_PORT_SYS_IMAGE_GUID,
+	IB_PORT_PKEY_SW_EXT_PORT_TRAP_SUP	= RDMA_PORT_PKEY_SW_EXT_PORT_TRAP,
+	IB_PORT_EXTENDED_SPEEDS_SUP             = RDMA_PORT_EXTENDED_SPEEDS,
+	IB_PORT_CM_SUP				= RDMA_PORT_CM,
+	IB_PORT_SNMP_TUNNEL_SUP			= RDMA_PORT_SNMP_TUNNEL,
+	IB_PORT_REINIT_SUP			= RDMA_PORT_REINIT,
+	IB_PORT_DEVICE_MGMT_SUP			= RDMA_PORT_DEVICE_MGMT,
+	IB_PORT_VENDOR_CLASS_SUP		= RDMA_PORT_VENDOR_CLASS,
+	IB_PORT_DR_NOTICE_SUP			= RDMA_PORT_DR_NOTICE,
+	IB_PORT_CAP_MASK_NOTICE_SUP		= RDMA_PORT_CAP_MASK_NOTICE,
+	IB_PORT_BOOT_MGMT_SUP			= RDMA_PORT_BOOT_MGMT,
+	IB_PORT_LINK_LATENCY_SUP		= RDMA_PORT_LINK_LATENCY,
+	IB_PORT_CLIENT_REG_SUP			= RDMA_PORT_CLIENT_REG,
+	IB_PORT_IP_BASED_GIDS			= RDMA_PORT_IP_BASED_GIDS,
 };
 
 enum ib_port_width {
diff --git a/include/uapi/rdma/rdma.h b/include/uapi/rdma/rdma.h
index 75c1baad313b..397796e63a3a 100644
--- a/include/uapi/rdma/rdma.h
+++ b/include/uapi/rdma/rdma.h
@@ -110,4 +110,37 @@ enum rdma_node_type {
 	RDMA_NODE_RNIC		= 4,
 	RDMA_NODE_USNIC_UDP	= 6,
 };
+
+/*
+ * This capability flags are taken from
+ * InfiniBandTM Architecture Specification Volume 1, Revision 1.3
+ * 14.2.5.6 PORTINFO - CapabilityMask
+ *
+ */
+enum rdma_port_cap {
+	RDMA_PORT_SM				= 1 <<  1,
+	RDMA_PORT_NOTICE			= 1 <<  2,
+	RDMA_PORT_TRAP				= 1 <<  3,
+	RDMA_PORT_OPT_IPD			= 1 <<  4,
+	RDMA_PORT_AUTO_MIGR			= 1 <<  5,
+	RDMA_PORT_SL_MAP			= 1 <<  6,
+	RDMA_PORT_MKEY_NVRAM			= 1 <<  7,
+	RDMA_PORT_PKEY_NVRAM			= 1 <<  8,
+	RDMA_PORT_LED_INFO			= 1 <<  9,
+	RDMA_PORT_SM_DISABLED			= 1 << 10,
+	RDMA_PORT_SYS_IMAGE_GUID		= 1 << 11,
+	RDMA_PORT_PKEY_SW_EXT_PORT_TRAP		= 1 << 12,
+	RDMA_PORT_EXTENDED_SPEEDS		= 1 << 14,
+	RDMA_PORT_CM				= 1 << 16,
+	RDMA_PORT_SNMP_TUNNEL			= 1 << 17,
+	RDMA_PORT_REINIT			= 1 << 18,
+	RDMA_PORT_DEVICE_MGMT			= 1 << 19,
+	RDMA_PORT_VENDOR_CLASS			= 1 << 20,
+	RDMA_PORT_DR_NOTICE			= 1 << 21,
+	RDMA_PORT_CAP_MASK_NOTICE		= 1 << 22,
+	RDMA_PORT_BOOT_MGMT			= 1 << 23,
+	RDMA_PORT_LINK_LATENCY			= 1 << 24,
+	RDMA_PORT_CLIENT_REG			= 1 << 25,
+	RDMA_PORT_IP_BASED_GIDS			= 1 << 26,
+};
 #endif /* _RDMA_H */
-- 
2.14.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[rdma-next 5/8] RDMA/core: Move port state to UAPI and rename it to be link

[Leon Romanovsky]

From: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

Provide to user space applications ability to reuse kernel
link state enum.

Signed-off-by: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
 include/rdma/ib_verbs.h  | 12 ++++++------
 include/uapi/rdma/rdma.h | 10 ++++++++++
 2 files changed, 16 insertions(+), 6 deletions(-)

diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index a7cd54061f48..7e9940244e38 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -383,12 +383,12 @@ static inline enum ib_mtu ib_mtu_int_to_enum(int mtu)
 }
 
 enum ib_port_state {
-	IB_PORT_NOP		= 0,
-	IB_PORT_DOWN		= 1,
-	IB_PORT_INIT		= 2,
-	IB_PORT_ARMED		= 3,
-	IB_PORT_ACTIVE		= 4,
-	IB_PORT_ACTIVE_DEFER	= 5
+	IB_PORT_NOP		= RDMA_LINK_STATE_NOP,
+	IB_PORT_DOWN		= RDMA_LINK_STATE_DOWN,
+	IB_PORT_INIT		= RDMA_LINK_STATE_INIT,
+	IB_PORT_ARMED		= RDMA_LINK_STATE_ARMED,
+	IB_PORT_ACTIVE		= RDMA_LINK_STATE_ACTIVE,
+	IB_PORT_ACTIVE_DEFER	= RDMA_LINK_STATE_ACTIVE_DEFER
 };
 
 enum ib_port_cap_flags {
diff --git a/include/uapi/rdma/rdma.h b/include/uapi/rdma/rdma.h
index 397796e63a3a..5f4019e2c5d4 100644
--- a/include/uapi/rdma/rdma.h
+++ b/include/uapi/rdma/rdma.h
@@ -143,4 +143,14 @@ enum rdma_port_cap {
 	RDMA_PORT_CLIENT_REG			= 1 << 25,
 	RDMA_PORT_IP_BASED_GIDS			= 1 << 26,
 };
+
+enum rdma_link_state {
+	RDMA_LINK_STATE_NOP,
+	RDMA_LINK_STATE_DOWN,
+	RDMA_LINK_STATE_INIT,
+	RDMA_LINK_STATE_ARMED,
+	RDMA_LINK_STATE_ACTIVE,
+	RDMA_LINK_STATE_ACTIVE_DEFER,
+};
+};
 #endif /* _RDMA_H */
-- 
2.14.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[rdma-next 6/8] RDMA/core: Add physical link state information to the UAPI

[Leon Romanovsky]

From: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

Add physical link state information to the UAPI and convert hardcoded
values to use the exported enum.

Signed-off-by: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
 drivers/infiniband/core/sysfs.c | 24 ++++++++++++++++--------
 include/uapi/rdma/rdma.h        | 40 ++++++++++++++++++++++++++++++++++++++++
 2 files changed, 56 insertions(+), 8 deletions(-)

diff --git a/drivers/infiniband/core/sysfs.c b/drivers/infiniband/core/sysfs.c
index c43280f8d5b3..913919c6822b 100644
--- a/drivers/infiniband/core/sysfs.c
+++ b/drivers/infiniband/core/sysfs.c
@@ -284,14 +284,22 @@ static ssize_t phys_state_show(struct ib_port *p, struct port_attribute *unused,
 		return ret;
 
 	switch (attr.phys_state) {
-	case 1:  return sprintf(buf, "1: Sleep\n");
-	case 2:  return sprintf(buf, "2: Polling\n");
-	case 3:  return sprintf(buf, "3: Disabled\n");
-	case 4:  return sprintf(buf, "4: PortConfigurationTraining\n");
-	case 5:  return sprintf(buf, "5: LinkUp\n");
-	case 6:  return sprintf(buf, "6: LinkErrorRecovery\n");
-	case 7:  return sprintf(buf, "7: Phy Test\n");
-	default: return sprintf(buf, "%d: <unknown>\n", attr.phys_state);
+	case RDMA_LINK_PHYS_STATE_SLEEP:
+		return sprintf(buf, "1: Sleep\n");
+	case RDMA_LINK_PHYS_STATE_POLLING:
+		return sprintf(buf, "2: Polling\n");
+	case RDMA_LINK_PHYS_STATE_DISABLED:
+		return sprintf(buf, "3: Disabled\n");
+	case RDMA_LINK_PHYS_STATE_PORT_CONFIGURATION_TRAINING:
+		return sprintf(buf, "4: PortConfigurationTraining\n");
+	case RDMA_LINK_PHYS_STATE_LINK_UP:
+		return sprintf(buf, "5: LinkUp\n");
+	case RDMA_LINK_PHYS_STATE_LINK_ERROR_RECOVER:
+		return sprintf(buf, "6: LinkErrorRecovery\n");
+	case RDMA_LINK_PHYS_STATE_LINK_PHY_TEST:
+		return sprintf(buf, "7: Phy Test\n");
+	default:
+		return sprintf(buf, "%d: <unknown>\n", attr.phys_state);
 	}
 }
 
diff --git a/include/uapi/rdma/rdma.h b/include/uapi/rdma/rdma.h
index 5f4019e2c5d4..42af3fe57b5a 100644
--- a/include/uapi/rdma/rdma.h
+++ b/include/uapi/rdma/rdma.h
@@ -152,5 +152,45 @@ enum rdma_link_state {
 	RDMA_LINK_STATE_ACTIVE,
 	RDMA_LINK_STATE_ACTIVE_DEFER,
 };
+
+/*
+ * When writing this field, only values 0, 1, 2, and 3
+ * are valid. Other values are ignored. See InfiniBand Architecture
+ * Specification Volume 2, Link/Phy Interface chapter.
+ */
+enum rdma_link_phys_state {
+	/*
+	 * 0 - 7 are according to the IB specification
+	 * 8 - 15 -  Reserved and ignored in IB, but partially
+	 *	     in use by OmniPath
+	 */
+	RDMA_LINK_PHYS_STATE_NOP,
+	RDMA_LINK_PHYS_STATE_SLEEP,
+	RDMA_LINK_PHYS_STATE_POLLING,
+	RDMA_LINK_PHYS_STATE_DISABLED,
+	RDMA_LINK_PHYS_STATE_PORT_CONFIGURATION_TRAINING,
+	RDMA_LINK_PHYS_STATE_LINK_UP,
+	RDMA_LINK_PHYS_STATE_LINK_ERROR_RECOVER,
+	RDMA_LINK_PHYS_STATE_LINK_PHY_TEST,
+
+	/* 8 is reserved in OPA */
+
+	/*
+	 * Offline: Port is quiet (transmitters disabled) due to lack of
+	 * physical media, unsupported media, or transition between link up
+	 * and next link up attempt
+	 */
+	RDMA_LINK_PHYS_STATE_OFFLINE = 9,
+
+	/* 10 is reserved in OPA */
+
+	/*
+	 * Phy_Test: Specific test patterns are transmitted, and receiver BER
+	 * can be monitored. This facilitates signal integrity testing for the
+	 * physical layer of the port.
+	 */
+	RDMA_LINK_PHYS_STATE_TEST = 11,
+
+	/* values 12-15 are reserved/ignored */
 };
 #endif /* _RDMA_H */
-- 
2.14.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[rdma-next 7/8] RDMA/nes: Remove zeroed parameter from port query callback

[Leon Romanovsky]

From: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

There is no need to explicitly zero parameters, because
the structure requested to be filled already initialized to zeros.

Signed-off-by: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
 drivers/infiniband/hw/nes/nes_verbs.c | 5 -----
 1 file changed, 5 deletions(-)

diff --git a/drivers/infiniband/hw/nes/nes_verbs.c b/drivers/infiniband/hw/nes/nes_verbs.c
index c2943e39d2f9..f0dc5f4aa177 100644
--- a/drivers/infiniband/hw/nes/nes_verbs.c
+++ b/drivers/infiniband/hw/nes/nes_verbs.c
@@ -481,21 +481,16 @@ static int nes_query_port(struct ib_device *ibdev, u8 port, struct ib_port_attr
 	props->active_mtu = ib_mtu_int_to_enum(netdev->mtu);
 
 	props->lid = 1;
-	props->lmc = 0;
-	props->sm_lid = 0;
-	props->sm_sl = 0;
 	if (netif_queue_stopped(netdev))
 		props->state = IB_PORT_DOWN;
 	else if (nesvnic->linkup)
 		props->state = IB_PORT_ACTIVE;
 	else
 		props->state = IB_PORT_DOWN;
-	props->phys_state = 0;
 	props->port_cap_flags = IB_PORT_CM_SUP | IB_PORT_REINIT_SUP |
 			IB_PORT_VENDOR_CLASS_SUP | IB_PORT_BOOT_MGMT_SUP;
 	props->gid_tbl_len = 1;
 	props->pkey_tbl_len = 1;
-	props->qkey_viol_cntr = 0;
 	props->active_width = IB_WIDTH_4X;
 	props->active_speed = IB_SPEED_SDR;
 	props->max_msg_sz = 0x80000000;
-- 
2.14.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[rdma-next 8/8] RDMA: Use defines instead of hard coded value for phys_state

[Leon Romanovsky]

From: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

There are new defines to express the physical state of the link, so
let's reuse them instead hard-coded values.

Signed-off-by: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
 drivers/infiniband/hw/bnxt_re/ib_verbs.c     |  4 ++--
 drivers/infiniband/hw/hfi1/opa_compat.h      | 30 +++++++++++++++++-----------
 drivers/infiniband/hw/hns/hns_roce_main.c    |  4 +++-
 drivers/infiniband/hw/mlx4/main.c            |  4 +++-
 drivers/infiniband/hw/mlx5/main.c            |  4 ++--
 drivers/infiniband/hw/ocrdma/ocrdma_verbs.c  |  4 ++--
 drivers/infiniband/hw/qedr/verbs.c           |  4 ++--
 drivers/infiniband/hw/usnic/usnic_ib_verbs.c |  6 +++---
 drivers/infiniband/sw/rxe/rxe_param.h        |  2 +-
 9 files changed, 36 insertions(+), 26 deletions(-)

diff --git a/drivers/infiniband/hw/bnxt_re/ib_verbs.c b/drivers/infiniband/hw/bnxt_re/ib_verbs.c
index d78fedc654d0..9bc28c756210 100644
--- a/drivers/infiniband/hw/bnxt_re/ib_verbs.c
+++ b/drivers/infiniband/hw/bnxt_re/ib_verbs.c
@@ -234,10 +234,10 @@ int bnxt_re_query_port(struct ib_device *ibdev, u8 port_num,
 
 	if (netif_running(rdev->netdev) && netif_carrier_ok(rdev->netdev)) {
 		port_attr->state = IB_PORT_ACTIVE;
-		port_attr->phys_state = 5;
+		port_attr->phys_state = RDMA_LINK_PHYS_STATE_LINK_UP;
 	} else {
 		port_attr->state = IB_PORT_DOWN;
-		port_attr->phys_state = 3;
+		port_attr->phys_state = RDMA_LINK_PHYS_STATE_DISABLED;
 	}
 	port_attr->max_mtu = IB_MTU_4096;
 	port_attr->active_mtu = iboe_get_mtu(rdev->netdev->mtu);
diff --git a/drivers/infiniband/hw/hfi1/opa_compat.h b/drivers/infiniband/hw/hfi1/opa_compat.h
index 6ef3c1cbdcd7..011ea5c89e25 100644
--- a/drivers/infiniband/hw/hfi1/opa_compat.h
+++ b/drivers/infiniband/hw/hfi1/opa_compat.h
@@ -92,20 +92,26 @@ static inline u8 port_states_to_phys_state(struct opa_port_states *ps)
  * Returned by the ibphys_portstate() routine.
  */
 enum opa_port_phys_state {
-	IB_PORTPHYSSTATE_NOP = 0,
+	IB_PORTPHYSSTATE_NOP = RDMA_LINK_PHYS_STATE_NOP,
+
 	/* 1 is reserved */
-	IB_PORTPHYSSTATE_POLLING = 2,
-	IB_PORTPHYSSTATE_DISABLED = 3,
-	IB_PORTPHYSSTATE_TRAINING = 4,
-	IB_PORTPHYSSTATE_LINKUP = 5,
-	IB_PORTPHYSSTATE_LINK_ERROR_RECOVERY = 6,
-	IB_PORTPHYSSTATE_PHY_TEST = 7,
+
+	IB_PORTPHYSSTATE_POLLING = RDMA_LINK_PHYS_STATE_POLLING,
+	IB_PORTPHYSSTATE_DISABLED = RDMA_LINK_PHYS_STATE_DISABLED,
+	IB_PORTPHYSSTATE_TRAINING = RDMA_LINK_PHYS_STATE_PORT_CONFIGURATION_TRAINING,
+	IB_PORTPHYSSTATE_LINKUP = RDMA_LINK_PHYS_STATE_LINK_UP,
+	IB_PORTPHYSSTATE_LINK_ERROR_RECOVERY = RDMA_LINK_PHYS_STATE_LINK_ERROR_RECOVER,
+	IB_PORTPHYSSTATE_PHY_TEST = RDMA_LINK_PHYS_STATE_LINK_PHY_TEST,
+
 	/* 8 is reserved */
-	OPA_PORTPHYSSTATE_OFFLINE = 9,
-	OPA_PORTPHYSSTATE_GANGED = 10,
-	OPA_PORTPHYSSTATE_TEST = 11,
-	OPA_PORTPHYSSTATE_MAX = 11,
+
+	OPA_PORTPHYSSTATE_OFFLINE = RDMA_LINK_PHYS_STATE_OFFLINE,
+
+	/* 10 is reserved */
+
+	OPA_PORTPHYSSTATE_TEST = RDMA_LINK_PHYS_STATE_TEST,
+	OPA_PORTPHYSSTATE_MAX = OPA_PORTPHYSSTATE_TEST,
+
 	/* values 12-15 are reserved/ignored */
 };
-
 #endif /* _LINUX_H */
diff --git a/drivers/infiniband/hw/hns/hns_roce_main.c b/drivers/infiniband/hw/hns/hns_roce_main.c
index d9777b662eba..48e88357d1c4 100644
--- a/drivers/infiniband/hw/hns/hns_roce_main.c
+++ b/drivers/infiniband/hw/hns/hns_roce_main.c
@@ -272,7 +272,9 @@ static int hns_roce_query_port(struct ib_device *ib_dev, u8 port_num,
 	props->active_mtu = mtu ? min(props->max_mtu, mtu) : IB_MTU_256;
 	props->state = (netif_running(net_dev) && netif_carrier_ok(net_dev)) ?
 			IB_PORT_ACTIVE : IB_PORT_DOWN;
-	props->phys_state = (props->state == IB_PORT_ACTIVE) ? 5 : 3;
+	props->phys_state = (props->state == IB_PORT_ACTIVE) ?
+			    RDMA_LINK_PHYS_STATE_LINK_UP :
+			    RDMA_LINK_PHYS_STATE_DISABLED;
 
 	spin_unlock_irqrestore(&hr_dev->iboe.lock, flags);
 
diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c
index c636842c5be0..d1598ef15406 100644
--- a/drivers/infiniband/hw/mlx4/main.c
+++ b/drivers/infiniband/hw/mlx4/main.c
@@ -694,7 +694,9 @@ static int ib_link_query_port(struct ib_device *ibdev, u8 port,
 
 static u8 state_to_phys_state(enum ib_port_state state)
 {
-	return state == IB_PORT_ACTIVE ? 5 : 3;
+	return state == IB_PORT_ACTIVE ?
+		RDMA_LINK_PHYS_STATE_LINK_UP :
+		RDMA_LINK_PHYS_STATE_DISABLED;
 }
 
 static int eth_link_query_port(struct ib_device *ibdev, u8 port,
diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c
index 894aec4a7c9d..8b3f19060a67 100644
--- a/drivers/infiniband/hw/mlx5/main.c
+++ b/drivers/infiniband/hw/mlx5/main.c
@@ -282,7 +282,7 @@ static int mlx5_query_port_roce(struct ib_device *device, u8 port_num,
 	props->max_msg_sz       = 1 << MLX5_CAP_GEN(dev->mdev, log_max_msg);
 	props->pkey_tbl_len     = 1;
 	props->state            = IB_PORT_DOWN;
-	props->phys_state       = 3;
+	props->phys_state       = RDMA_LINK_PHYS_STATE_DISABLED;
 
 	mlx5_query_nic_vport_qkey_viol_cntr(dev->mdev, &qkey_viol_cntr);
 	props->qkey_viol_cntr = qkey_viol_cntr;
@@ -304,7 +304,7 @@ static int mlx5_query_port_roce(struct ib_device *device, u8 port_num,
 
 	if (netif_running(ndev) && netif_carrier_ok(ndev)) {
 		props->state      = IB_PORT_ACTIVE;
-		props->phys_state = 5;
+		props->phys_state = RDMA_LINK_PHYS_STATE_LINK_UP;
 	}
 
 	ndev_ib_mtu = iboe_get_mtu(ndev->mtu);
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c b/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c
index 27d5e8d9f08d..32e7a04b027f 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c
@@ -220,10 +220,10 @@ int ocrdma_query_port(struct ib_device *ibdev,
 	netdev = dev->nic_info.netdev;
 	if (netif_running(netdev) && netif_oper_up(netdev)) {
 		port_state = IB_PORT_ACTIVE;
-		props->phys_state = 5;
+		props->phys_state = RDMA_LINK_PHYS_STATE_LINK_UP;
 	} else {
 		port_state = IB_PORT_DOWN;
-		props->phys_state = 3;
+		props->phys_state = RDMA_LINK_PHYS_STATE_DISABLED;
 	}
 	props->max_mtu = IB_MTU_4096;
 	props->active_mtu = iboe_get_mtu(netdev->mtu);
diff --git a/drivers/infiniband/hw/qedr/verbs.c b/drivers/infiniband/hw/qedr/verbs.c
index 9ee2dce3e5bb..724921aad51b 100644
--- a/drivers/infiniband/hw/qedr/verbs.c
+++ b/drivers/infiniband/hw/qedr/verbs.c
@@ -251,10 +251,10 @@ int qedr_query_port(struct ib_device *ibdev, u8 port, struct ib_port_attr *attr)
 	/* *attr being zeroed by the caller, avoid zeroing it here */
 	if (rdma_port->port_state == QED_RDMA_PORT_UP) {
 		attr->state = IB_PORT_ACTIVE;
-		attr->phys_state = 5;
+		attr->phys_state = RDMA_LINK_PHYS_STATE_LINK_UP;
 	} else {
 		attr->state = IB_PORT_DOWN;
-		attr->phys_state = 3;
+		attr->phys_state = RDMA_LINK_PHYS_STATE_DISABLED;
 	}
 	attr->max_mtu = IB_MTU_4096;
 	attr->active_mtu = iboe_get_mtu(dev->ndev->mtu);
diff --git a/drivers/infiniband/hw/usnic/usnic_ib_verbs.c b/drivers/infiniband/hw/usnic/usnic_ib_verbs.c
index e5f57dd49980..8889b549a6bc 100644
--- a/drivers/infiniband/hw/usnic/usnic_ib_verbs.c
+++ b/drivers/infiniband/hw/usnic/usnic_ib_verbs.c
@@ -324,13 +324,13 @@ int usnic_ib_query_port(struct ib_device *ibdev, u8 port,
 
 	if (!us_ibdev->ufdev->link_up) {
 		props->state = IB_PORT_DOWN;
-		props->phys_state = 3;
+		props->phys_state = RDMA_LINK_PHYS_STATE_DISABLED;
 	} else if (!us_ibdev->ufdev->inaddr) {
 		props->state = IB_PORT_INIT;
-		props->phys_state = 4;
+		props->phys_state = RDMA_LINK_PHYS_STATE_PORT_CONFIGURATION_TRAINING;
 	} else {
 		props->state = IB_PORT_ACTIVE;
-		props->phys_state = 5;
+		props->phys_state = RDMA_LINK_PHYS_STATE_LINK_UP;
 	}
 
 	props->port_cap_flags = 0;
diff --git a/drivers/infiniband/sw/rxe/rxe_param.h b/drivers/infiniband/sw/rxe/rxe_param.h
index 1b596fbbe251..475767262102 100644
--- a/drivers/infiniband/sw/rxe/rxe_param.h
+++ b/drivers/infiniband/sw/rxe/rxe_param.h
@@ -157,7 +157,7 @@ enum rxe_port_param {
 	RXE_PORT_ACTIVE_WIDTH		= IB_WIDTH_1X,
 	RXE_PORT_ACTIVE_SPEED		= 1,
 	RXE_PORT_PKEY_TBL_LEN		= 64,
-	RXE_PORT_PHYS_STATE		= 2,
+	RXE_PORT_PHYS_STATE		= RDMA_LINK_PHYS_STATE_POLLING,
 	RXE_PORT_SUBNET_PREFIX		= 0xfe80000000000000ULL,
 };
 
-- 
2.14.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [rdma-next 8/8] RDMA: Use defines instead of hard coded value for phys_state

[Boyer, Andrew]

On 8/15/17, 5:14 AM, "linux-rdma-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org on behalf of Leon
Romanovsky" <linux-rdma-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org on behalf of
leon-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote:

>From: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
>
>There are new defines to express the physical state of the link, so
>let's reuse them instead hard-coded values.
>
>Signed-off-by: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
>---
>diff --git a/drivers/infiniband/sw/rxe/rxe_param.h
>b/drivers/infiniband/sw/rxe/rxe_param.h
>index 1b596fbbe251..475767262102 100644
>--- a/drivers/infiniband/sw/rxe/rxe_param.h
>+++ b/drivers/infiniband/sw/rxe/rxe_param.h
>@@ -157,7 +157,7 @@ enum rxe_port_param {
> 	RXE_PORT_ACTIVE_WIDTH		= IB_WIDTH_1X,
> 	RXE_PORT_ACTIVE_SPEED		= 1,
> 	RXE_PORT_PKEY_TBL_LEN		= 64,
>-	RXE_PORT_PHYS_STATE		= 2,
>+	RXE_PORT_PHYS_STATE		= RDMA_LINK_PHYS_STATE_POLLING,
> 	RXE_PORT_SUBNET_PREFIX		= 0xfe80000000000000ULL,
> };

You can also remove IB_PHYS_STATE_* from rxe.h and the places where
they¹re used.

> 
>-- 
>2.14.0
>
>--
>To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
>the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
>More majordomo info at  http://vger.kernel.org/majordomo-info.html

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [rdma-next 8/8] RDMA: Use defines instead of hard coded value for phys_state

[Leon Romanovsky]

[-- Attachment #1: Type: text/plain, Size: 1625 bytes --]

On Wed, Aug 23, 2017 at 02:59:52PM +0000, Boyer, Andrew wrote:
>
>
> On 8/15/17, 5:14 AM, "linux-rdma-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org on behalf of Leon
> Romanovsky" <linux-rdma-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org on behalf of
> leon-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote:
>
> >From: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
> >
> >There are new defines to express the physical state of the link, so
> >let's reuse them instead hard-coded values.
> >
> >Signed-off-by: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
> >---
> >diff --git a/drivers/infiniband/sw/rxe/rxe_param.h
> >b/drivers/infiniband/sw/rxe/rxe_param.h
> >index 1b596fbbe251..475767262102 100644
> >--- a/drivers/infiniband/sw/rxe/rxe_param.h
> >+++ b/drivers/infiniband/sw/rxe/rxe_param.h
> >@@ -157,7 +157,7 @@ enum rxe_port_param {
> > 	RXE_PORT_ACTIVE_WIDTH		= IB_WIDTH_1X,
> > 	RXE_PORT_ACTIVE_SPEED		= 1,
> > 	RXE_PORT_PKEY_TBL_LEN		= 64,
> >-	RXE_PORT_PHYS_STATE		= 2,
> >+	RXE_PORT_PHYS_STATE		= RDMA_LINK_PHYS_STATE_POLLING,
> > 	RXE_PORT_SUBNET_PREFIX		= 0xfe80000000000000ULL,
> > };
>
> You can also remove IB_PHYS_STATE_* from rxe.h and the places where
> they¹re used.

At the end, I found the way to avoid the whole this series.
This series is dropped.

Thanks


>
> >
> >--
> >2.14.0
> >
> >--
> >To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
> >the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> >More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [pull request][rdma-next 0/8] Export device and port properties

[Leon Romanovsky]

[-- Attachment #1: Type: text/plain, Size: 772 bytes --]

On Tue, Aug 15, 2017 at 12:14:23PM +0300, Leon Romanovsky wrote:
> Hi,
>
> This is the third part of RDMAtool infrastructure changes. In this series,
> I'm exporting various device and port properties which are parsed in
> RDMAtool.
>
> Due to the fact, that the "IB/hfi1: Document phys port state bits not
> used in IB", I embed its content into the actual patches.
>
> As part of this series, I performed small code cleanup to improve our
> static checker situation.
>
> Thanks
>
> Cc: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
> Cc: Jason Gunthorpe <jgunthorpe-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
> ----------------------------------------------------------------

Doug,
To be clear, this PR is not needed.

Thanks,

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [rdma-next 8/8] RDMA: Use defines instead of hard coded value for phys_state

[Boyer, Andrew]

On 8/23/17, 11:29 AM, "Leon Romanovsky" <leon@kernel.org> wrote:

>On Wed, Aug 23, 2017 at 02:59:52PM +0000, Boyer, Andrew wrote:
>>
>>
>> On 8/15/17, 5:14 AM, "linux-rdma-owner@vger.kernel.org on behalf of Leon
>> Romanovsky" <linux-rdma-owner@vger.kernel.org on behalf of
>> leon@kernel.org> wrote:
>>
>> >From: Leon Romanovsky <leonro@mellanox.com>
>> >
>> >There are new defines to express the physical state of the link, so
>> >let's reuse them instead hard-coded values.
>> >
>> >Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
>> >---
>> >diff --git a/drivers/infiniband/sw/rxe/rxe_param.h
>> >b/drivers/infiniband/sw/rxe/rxe_param.h
>> >index 1b596fbbe251..475767262102 100644
>> >--- a/drivers/infiniband/sw/rxe/rxe_param.h
>> >+++ b/drivers/infiniband/sw/rxe/rxe_param.h
>> >@@ -157,7 +157,7 @@ enum rxe_port_param {
>> > 	RXE_PORT_ACTIVE_WIDTH		= IB_WIDTH_1X,
>> > 	RXE_PORT_ACTIVE_SPEED		= 1,
>> > 	RXE_PORT_PKEY_TBL_LEN		= 64,
>> >-	RXE_PORT_PHYS_STATE		= 2,
>> >+	RXE_PORT_PHYS_STATE		= RDMA_LINK_PHYS_STATE_POLLING,
>> > 	RXE_PORT_SUBNET_PREFIX		= 0xfe80000000000000ULL,
>> > };
>>
>> You can also remove IB_PHYS_STATE_* from rxe.h and the places where
>> they¹re used.
>
>At the end, I found the way to avoid the whole this series.
>This series is dropped.
>
>Thanks

Would you object if I added similar / matching defines in ib_verbs.h?

The phys_state field allows userspace to distinguish between a link that’s
down and a link that’s turned off, but only a few drivers seem to support
it.

-Andrew

Re: [rdma-next 8/8] RDMA: Use defines instead of hard coded value for phys_state

[Leon Romanovsky]

[-- Attachment #1: Type: text/plain, Size: 2022 bytes --]

On Thu, Aug 24, 2017 at 12:56:29PM +0000, Boyer, Andrew wrote:
> On 8/23/17, 11:29 AM, "Leon Romanovsky" <leon-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote:
>
> >On Wed, Aug 23, 2017 at 02:59:52PM +0000, Boyer, Andrew wrote:
> >>
> >>
> >> On 8/15/17, 5:14 AM, "linux-rdma-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org on behalf of Leon
> >> Romanovsky" <linux-rdma-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org on behalf of
> >> leon-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote:
> >>
> >> >From: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
> >> >
> >> >There are new defines to express the physical state of the link, so
> >> >let's reuse them instead hard-coded values.
> >> >
> >> >Signed-off-by: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
> >> >---
> >> >diff --git a/drivers/infiniband/sw/rxe/rxe_param.h
> >> >b/drivers/infiniband/sw/rxe/rxe_param.h
> >> >index 1b596fbbe251..475767262102 100644
> >> >--- a/drivers/infiniband/sw/rxe/rxe_param.h
> >> >+++ b/drivers/infiniband/sw/rxe/rxe_param.h
> >> >@@ -157,7 +157,7 @@ enum rxe_port_param {
> >> > 	RXE_PORT_ACTIVE_WIDTH		= IB_WIDTH_1X,
> >> > 	RXE_PORT_ACTIVE_SPEED		= 1,
> >> > 	RXE_PORT_PKEY_TBL_LEN		= 64,
> >> >-	RXE_PORT_PHYS_STATE		= 2,
> >> >+	RXE_PORT_PHYS_STATE		= RDMA_LINK_PHYS_STATE_POLLING,
> >> > 	RXE_PORT_SUBNET_PREFIX		= 0xfe80000000000000ULL,
> >> > };
> >>
> >> You can also remove IB_PHYS_STATE_* from rxe.h and the places where
> >> they¹re used.
> >
> >At the end, I found the way to avoid the whole this series.
> >This series is dropped.
> >
> >Thanks
>
> Would you object if I added similar / matching defines in ib_verbs.h?

Of course not, it will be good to see most of the cleanups from this
pull request, so feel free to resend/base your work on it.

>
> The phys_state field allows userspace to distinguish between a link that’s
> down and a link that’s turned off, but only a few drivers seem to support
> it.
>
> -Andrew
>
>

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [pull request][rdma-next 0/8] Export device and port properties

[Doug Ledford]

On Wed, 2017-08-23 at 19:27 +0300, Leon Romanovsky wrote:
> On Tue, Aug 15, 2017 at 12:14:23PM +0300, Leon Romanovsky wrote:
> > Hi,
> > 
> > This is the third part of RDMAtool infrastructure changes. In this
> > series,
> > I'm exporting various device and port properties which are parsed
> > in
> > RDMAtool.
> > 
> > Due to the fact, that the "IB/hfi1: Document phys port state bits
> > not
> > used in IB", I embed its content into the actual patches.
> > 
> > As part of this series, I performed small code cleanup to improve
> > our
> > static checker situation.
> > 
> > Thanks
> > 
> > Cc: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
> > Cc: Jason Gunthorpe <jgunthorpe-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
> > ----------------------------------------------------------------
> 
> Doug,
> To be clear, this PR is not needed.

Dropped.

-- 
Doug Ledford <dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
    GPG KeyID: B826A3330E572FDD
    Key fingerprint = AE6B 1BDA 122B 23B4 265B  1274 B826 A333 0E57 2FDD

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html