From bf77cc34f3e6905bc8841489dd628bbffe37d743 Mon Sep 17 00:00:00 2001
From: Jianxin Xiong <jianxin.xiong@intel.com>
Date: Thu, 12 May 2016 10:23:53 -0700
Subject: ib_pack.h: Add opcode definition for send with invalidate

The opcode for "SEND Last with Invalidate" and "SEND Only with
Invalidate" have been defined for RC in IBA Specification Vol 1
since Release 1.2. Add the definition to the header file in
preparation of supporting these opcodes in rdmavt based drivers.

Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Jianxin Xiong <jianxin.xiong@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
---
 include/rdma/ib_pack.h | 5 +++++
 1 file changed, 5 insertions(+)

(limited to 'include')

diff --git a/include/rdma/ib_pack.h b/include/rdma/ib_pack.h
index 0f3daae44bf9..b13419ce99ff 100644
--- a/include/rdma/ib_pack.h
+++ b/include/rdma/ib_pack.h
@@ -103,6 +103,9 @@ enum {
 	IB_OPCODE_ATOMIC_ACKNOWLEDGE                = 0x12,
 	IB_OPCODE_COMPARE_SWAP                      = 0x13,
 	IB_OPCODE_FETCH_ADD                         = 0x14,
+	/* opcode 0x15 is reserved */
+	IB_OPCODE_SEND_LAST_WITH_INVALIDATE         = 0x16,
+	IB_OPCODE_SEND_ONLY_WITH_INVALIDATE         = 0x17,
 
 	/* real constants follow -- see comment about above IB_OPCODE()
 	   macro for more details */
@@ -129,6 +132,8 @@ enum {
 	IB_OPCODE(RC, ATOMIC_ACKNOWLEDGE),
 	IB_OPCODE(RC, COMPARE_SWAP),
 	IB_OPCODE(RC, FETCH_ADD),
+	IB_OPCODE(RC, SEND_LAST_WITH_INVALIDATE),
+	IB_OPCODE(RC, SEND_ONLY_WITH_INVALIDATE),
 
 	/* UC */
 	IB_OPCODE(UC, SEND_FIRST),
-- 
cgit v1.2.3


From c34d3761872a2784d7bae2a0bb6205e8ee07a7f7 Mon Sep 17 00:00:00 2001
From: Mark Bloch <markb@mellanox.com>
Date: Thu, 19 May 2016 17:12:34 +0300
Subject: IB/netlink: Add a new local service operation

This commits adds a new RDMA local service operation:
- IP to GID resolution.

The client request would include the ifindex of the outgoing interface
and would place in an attribute (LS_NLA_TYPE_IPV4 or LS_NLA_TYPE_IPV6)
the destnation IP.

The local service would answer with a message that has the attribute:
- LS_NLA_TYPE_DGID - The destination GID.

Signed-off-by: Mark Bloch <markb@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
---
 include/uapi/rdma/rdma_netlink.h | 10 ++++++++++
 1 file changed, 10 insertions(+)

(limited to 'include')

diff --git a/include/uapi/rdma/rdma_netlink.h b/include/uapi/rdma/rdma_netlink.h
index 6e373d151cad..02fe8390c18f 100644
--- a/include/uapi/rdma/rdma_netlink.h
+++ b/include/uapi/rdma/rdma_netlink.h
@@ -135,10 +135,12 @@ enum {
  * Local service operations:
  *   RESOLVE - The client requests the local service to resolve a path.
  *   SET_TIMEOUT - The local service requests the client to set the timeout.
+ *   IP_RESOLVE - The client requests the local service to resolve an IP to GID.
  */
 enum {
 	RDMA_NL_LS_OP_RESOLVE = 0,
 	RDMA_NL_LS_OP_SET_TIMEOUT,
+	RDMA_NL_LS_OP_IP_RESOLVE,
 	RDMA_NL_LS_NUM_OPS
 };
 
@@ -176,6 +178,10 @@ struct rdma_ls_resolve_header {
 	__u8 path_use;
 };
 
+struct rdma_ls_ip_resolve_header {
+	__u32 ifindex;
+};
+
 /* Local service attribute type */
 #define RDMA_NLA_F_MANDATORY	(1 << 13)
 #define RDMA_NLA_TYPE_MASK	(~(NLA_F_NESTED | NLA_F_NET_BYTEORDER | \
@@ -193,6 +199,8 @@ struct rdma_ls_resolve_header {
  *   TCLASS          u8
  *   PKEY            u16                        cpu
  *   QOS_CLASS       u16                        cpu
+ *   IPV4            u32                        BE
+ *   IPV6            u8[16]                     BE
  */
 enum {
 	LS_NLA_TYPE_UNSPEC = 0,
@@ -204,6 +212,8 @@ enum {
 	LS_NLA_TYPE_TCLASS,
 	LS_NLA_TYPE_PKEY,
 	LS_NLA_TYPE_QOS_CLASS,
+	LS_NLA_TYPE_IPV4,
+	LS_NLA_TYPE_IPV6,
 	LS_NLA_TYPE_MAX
 };
 
-- 
cgit v1.2.3


From 507f6afa3ab6c9473d81c8091be0d1b859509a6d Mon Sep 17 00:00:00 2001
From: Erez Shitrit <erezsh@mellanox.com>
Date: Wed, 25 May 2016 22:02:04 +0300
Subject: IB/core: Introduce capabilitymask2 field in ClassPortInfo mad

Change struct ib_class_port_info to conform to IB Spec 1.3
That in order to get specific capability mask from ClassPortInfo mad.

>From the IB Spec, ClassPortInfo section:
        "CapabilityMask2 Bits 0-26: Additional class-specific capabilities...
         RespTimeValue the rest 5 bits"

The new struct now has one field for capabilitymask2 (previously was the
reserved field) and the resp_time field.

And it fixes up qib and srpt, use of the field repurposed to be used as
capabilitymask2:
IB/qib: Change pma_get_classportinfo
IB/srpt: Adjust the use of ib_class_port_info

Signed-off-by: Erez Shitrit <erezsh@mellanox.com>
Reviewed-by: Leon Romanovsky <leonro@mellanox.com>
Reviewed-by: Hal Rosenstock <hal@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
---
 drivers/infiniband/hw/qib/qib_mad.c   |  6 ++--
 drivers/infiniband/ulp/srpt/ib_srpt.c |  2 +-
 include/rdma/ib_mad.h                 | 60 +++++++++++++++++++++++++++++++++--
 3 files changed, 63 insertions(+), 5 deletions(-)

(limited to 'include')

diff --git a/drivers/infiniband/hw/qib/qib_mad.c b/drivers/infiniband/hw/qib/qib_mad.c
index 0bd18375d7df..d2ac29861af5 100644
--- a/drivers/infiniband/hw/qib/qib_mad.c
+++ b/drivers/infiniband/hw/qib/qib_mad.c
@@ -1172,11 +1172,13 @@ static int pma_get_classportinfo(struct ib_pma_mad *pmp,
 	 * Set the most significant bit of CM2 to indicate support for
 	 * congestion statistics
 	 */
-	p->reserved[0] = dd->psxmitwait_supported << 7;
+	ib_set_cpi_capmask2(p,
+			    dd->psxmitwait_supported <<
+			    (31 - IB_CLASS_PORT_INFO_RESP_TIME_FIELD_SIZE));
 	/*
 	 * Expected response time is 4.096 usec. * 2^18 == 1.073741824 sec.
 	 */
-	p->resp_time_value = 18;
+	ib_set_cpi_resp_time(p, 18);
 
 	return reply((struct ib_smp *) pmp);
 }
diff --git a/drivers/infiniband/ulp/srpt/ib_srpt.c b/drivers/infiniband/ulp/srpt/ib_srpt.c
index 2843f1ae75bd..887ebadd4774 100644
--- a/drivers/infiniband/ulp/srpt/ib_srpt.c
+++ b/drivers/infiniband/ulp/srpt/ib_srpt.c
@@ -254,8 +254,8 @@ static void srpt_get_class_port_info(struct ib_dm_mad *mad)
 	memset(cif, 0, sizeof(*cif));
 	cif->base_version = 1;
 	cif->class_version = 1;
-	cif->resp_time_value = 20;
 
+	ib_set_cpi_resp_time(cif, 20);
 	mad->mad_hdr.status = 0;
 }
 
diff --git a/include/rdma/ib_mad.h b/include/rdma/ib_mad.h
index 37dd534cbeab..c8a773ffe23b 100644
--- a/include/rdma/ib_mad.h
+++ b/include/rdma/ib_mad.h
@@ -239,12 +239,15 @@ struct ib_vendor_mad {
 
 #define IB_MGMT_CLASSPORTINFO_ATTR_ID	cpu_to_be16(0x0001)
 
+#define IB_CLASS_PORT_INFO_RESP_TIME_MASK	0x1F
+#define IB_CLASS_PORT_INFO_RESP_TIME_FIELD_SIZE 5
+
 struct ib_class_port_info {
 	u8			base_version;
 	u8			class_version;
 	__be16			capability_mask;
-	u8			reserved[3];
-	u8			resp_time_value;
+	  /* 27 bits for cap_mask2, 5 bits for resp_time */
+	__be32			cap_mask2_resp_time;
 	u8			redirect_gid[16];
 	__be32			redirect_tcslfl;
 	__be16			redirect_lid;
@@ -259,6 +262,59 @@ struct ib_class_port_info {
 	__be32			trap_qkey;
 };
 
+/**
+ * ib_get_cpi_resp_time - Returns the resp_time value from
+ * cap_mask2_resp_time in ib_class_port_info.
+ * @cpi: A struct ib_class_port_info mad.
+ */
+static inline u8 ib_get_cpi_resp_time(struct ib_class_port_info *cpi)
+{
+	return (u8)(be32_to_cpu(cpi->cap_mask2_resp_time) &
+		    IB_CLASS_PORT_INFO_RESP_TIME_MASK);
+}
+
+/**
+ * ib_set_cpi_resptime - Sets the response time in an
+ * ib_class_port_info mad.
+ * @cpi: A struct ib_class_port_info.
+ * @rtime: The response time to set.
+ */
+static inline void ib_set_cpi_resp_time(struct ib_class_port_info *cpi,
+					u8 rtime)
+{
+	cpi->cap_mask2_resp_time =
+		(cpi->cap_mask2_resp_time &
+		 cpu_to_be32(~IB_CLASS_PORT_INFO_RESP_TIME_MASK)) |
+		cpu_to_be32(rtime & IB_CLASS_PORT_INFO_RESP_TIME_MASK);
+}
+
+/**
+ * ib_get_cpi_capmask2 - Returns the capmask2 value from
+ * cap_mask2_resp_time in ib_class_port_info.
+ * @cpi: A struct ib_class_port_info mad.
+ */
+static inline u32 ib_get_cpi_capmask2(struct ib_class_port_info *cpi)
+{
+	return (be32_to_cpu(cpi->cap_mask2_resp_time) >>
+		IB_CLASS_PORT_INFO_RESP_TIME_FIELD_SIZE);
+}
+
+/**
+ * ib_set_cpi_capmask2 - Sets the capmask2 in an
+ * ib_class_port_info mad.
+ * @cpi: A struct ib_class_port_info.
+ * @capmask2: The capmask2 to set.
+ */
+static inline void ib_set_cpi_capmask2(struct ib_class_port_info *cpi,
+				       u32 capmask2)
+{
+	cpi->cap_mask2_resp_time =
+		(cpi->cap_mask2_resp_time &
+		 cpu_to_be32(IB_CLASS_PORT_INFO_RESP_TIME_MASK)) |
+		cpu_to_be32(capmask2 <<
+			    IB_CLASS_PORT_INFO_RESP_TIME_FIELD_SIZE);
+}
+
 struct ib_mad_notice_attr {
 	u8 generic_type;
 	u8 prod_type_msb;
-- 
cgit v1.2.3


From 628e6f75156eac6979eacc2e9383e41f6432beb4 Mon Sep 17 00:00:00 2001
From: Erez Shitrit <erezsh@mellanox.com>
Date: Wed, 25 May 2016 22:02:05 +0300
Subject: IB/SA Agent: Add support for SA agent get ClassPortInfo

New SA query function to return the ClassPortInfo struct from the SA.
If the SM supports FullMemberSendOnly mode for MCG's, it sets a
capability bit in the capability_mask2 field of the response.

Signed-off-by: Erez Shitrit <erezsh@mellanox.com>
Reviewed-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
---
 drivers/infiniband/core/sa_query.c | 173 +++++++++++++++++++++++++++++++++++++
 include/rdma/ib_sa.h               |  12 +++
 2 files changed, 185 insertions(+)

(limited to 'include')

diff --git a/drivers/infiniband/core/sa_query.c b/drivers/infiniband/core/sa_query.c
index 3ebd108bcc5f..5c37f2f56d3f 100644
--- a/drivers/infiniband/core/sa_query.c
+++ b/drivers/infiniband/core/sa_query.c
@@ -119,6 +119,12 @@ struct ib_sa_guidinfo_query {
 	struct ib_sa_query sa_query;
 };
 
+struct ib_sa_classport_info_query {
+	void (*callback)(int, struct ib_class_port_info *, void *);
+	void *context;
+	struct ib_sa_query sa_query;
+};
+
 struct ib_sa_mcmember_query {
 	void (*callback)(int, struct ib_sa_mcmember_rec *, void *);
 	void *context;
@@ -392,6 +398,82 @@ static const struct ib_field service_rec_table[] = {
 	  .size_bits    = 2*64 },
 };
 
+#define CLASSPORTINFO_REC_FIELD(field) \
+	.struct_offset_bytes = offsetof(struct ib_class_port_info, field),	\
+	.struct_size_bytes   = sizeof((struct ib_class_port_info *)0)->field,	\
+	.field_name          = "ib_class_port_info:" #field
+
+static const struct ib_field classport_info_rec_table[] = {
+	{ CLASSPORTINFO_REC_FIELD(base_version),
+	  .offset_words = 0,
+	  .offset_bits  = 0,
+	  .size_bits    = 8 },
+	{ CLASSPORTINFO_REC_FIELD(class_version),
+	  .offset_words = 0,
+	  .offset_bits  = 8,
+	  .size_bits    = 8 },
+	{ CLASSPORTINFO_REC_FIELD(capability_mask),
+	  .offset_words = 0,
+	  .offset_bits  = 16,
+	  .size_bits    = 16 },
+	{ CLASSPORTINFO_REC_FIELD(cap_mask2_resp_time),
+	  .offset_words = 1,
+	  .offset_bits  = 0,
+	  .size_bits    = 32 },
+	{ CLASSPORTINFO_REC_FIELD(redirect_gid),
+	  .offset_words = 2,
+	  .offset_bits  = 0,
+	  .size_bits    = 128 },
+	{ CLASSPORTINFO_REC_FIELD(redirect_tcslfl),
+	  .offset_words = 6,
+	  .offset_bits  = 0,
+	  .size_bits    = 32 },
+	{ CLASSPORTINFO_REC_FIELD(redirect_lid),
+	  .offset_words = 7,
+	  .offset_bits  = 0,
+	  .size_bits    = 16 },
+	{ CLASSPORTINFO_REC_FIELD(redirect_pkey),
+	  .offset_words = 7,
+	  .offset_bits  = 16,
+	  .size_bits    = 16 },
+
+	{ CLASSPORTINFO_REC_FIELD(redirect_qp),
+	  .offset_words = 8,
+	  .offset_bits  = 0,
+	  .size_bits    = 32 },
+	{ CLASSPORTINFO_REC_FIELD(redirect_qkey),
+	  .offset_words = 9,
+	  .offset_bits  = 0,
+	  .size_bits    = 32 },
+
+	{ CLASSPORTINFO_REC_FIELD(trap_gid),
+	  .offset_words = 10,
+	  .offset_bits  = 0,
+	  .size_bits    = 128 },
+	{ CLASSPORTINFO_REC_FIELD(trap_tcslfl),
+	  .offset_words = 14,
+	  .offset_bits  = 0,
+	  .size_bits    = 32 },
+
+	{ CLASSPORTINFO_REC_FIELD(trap_lid),
+	  .offset_words = 15,
+	  .offset_bits  = 0,
+	  .size_bits    = 16 },
+	{ CLASSPORTINFO_REC_FIELD(trap_pkey),
+	  .offset_words = 15,
+	  .offset_bits  = 16,
+	  .size_bits    = 16 },
+
+	{ CLASSPORTINFO_REC_FIELD(trap_hlqp),
+	  .offset_words = 16,
+	  .offset_bits  = 0,
+	  .size_bits    = 32 },
+	{ CLASSPORTINFO_REC_FIELD(trap_qkey),
+	  .offset_words = 17,
+	  .offset_bits  = 0,
+	  .size_bits    = 32 },
+};
+
 #define GUIDINFO_REC_FIELD(field) \
 	.struct_offset_bytes = offsetof(struct ib_sa_guidinfo_rec, field),	\
 	.struct_size_bytes   = sizeof((struct ib_sa_guidinfo_rec *) 0)->field,	\
@@ -1645,6 +1727,97 @@ err1:
 }
 EXPORT_SYMBOL(ib_sa_guid_info_rec_query);
 
+/* Support get SA ClassPortInfo */
+static void ib_sa_classport_info_rec_callback(struct ib_sa_query *sa_query,
+					      int status,
+					      struct ib_sa_mad *mad)
+{
+	struct ib_sa_classport_info_query *query =
+		container_of(sa_query, struct ib_sa_classport_info_query, sa_query);
+
+	if (mad) {
+		struct ib_class_port_info rec;
+
+		ib_unpack(classport_info_rec_table,
+			  ARRAY_SIZE(classport_info_rec_table),
+			  mad->data, &rec);
+		query->callback(status, &rec, query->context);
+	} else {
+		query->callback(status, NULL, query->context);
+	}
+}
+
+static void ib_sa_portclass_info_rec_release(struct ib_sa_query *sa_query)
+{
+	kfree(container_of(sa_query, struct ib_sa_classport_info_query,
+			   sa_query));
+}
+
+int ib_sa_classport_info_rec_query(struct ib_sa_client *client,
+				   struct ib_device *device, u8 port_num,
+				   int timeout_ms, gfp_t gfp_mask,
+				   void (*callback)(int status,
+						    struct ib_class_port_info *resp,
+						    void *context),
+				   void *context,
+				   struct ib_sa_query **sa_query)
+{
+	struct ib_sa_classport_info_query *query;
+	struct ib_sa_device *sa_dev = ib_get_client_data(device, &sa_client);
+	struct ib_sa_port *port;
+	struct ib_mad_agent *agent;
+	struct ib_sa_mad *mad;
+	int ret;
+
+	if (!sa_dev)
+		return -ENODEV;
+
+	port  = &sa_dev->port[port_num - sa_dev->start_port];
+	agent = port->agent;
+
+	query = kzalloc(sizeof(*query), gfp_mask);
+	if (!query)
+		return -ENOMEM;
+
+	query->sa_query.port = port;
+	ret = alloc_mad(&query->sa_query, gfp_mask);
+	if (ret)
+		goto err1;
+
+	ib_sa_client_get(client);
+	query->sa_query.client = client;
+	query->callback        = callback;
+	query->context         = context;
+
+	mad = query->sa_query.mad_buf->mad;
+	init_mad(mad, agent);
+
+	query->sa_query.callback = callback ? ib_sa_classport_info_rec_callback : NULL;
+
+	query->sa_query.release  = ib_sa_portclass_info_rec_release;
+	/* support GET only */
+	mad->mad_hdr.method	 = IB_MGMT_METHOD_GET;
+	mad->mad_hdr.attr_id	 = cpu_to_be16(IB_SA_ATTR_CLASS_PORTINFO);
+	mad->sa_hdr.comp_mask	 = 0;
+	*sa_query = &query->sa_query;
+
+	ret = send_mad(&query->sa_query, timeout_ms, gfp_mask);
+	if (ret < 0)
+		goto err2;
+
+	return ret;
+
+err2:
+	*sa_query = NULL;
+	ib_sa_client_put(query->sa_query.client);
+	free_mad(&query->sa_query);
+
+err1:
+	kfree(query);
+	return ret;
+}
+EXPORT_SYMBOL(ib_sa_classport_info_rec_query);
+
 static void send_handler(struct ib_mad_agent *agent,
 			 struct ib_mad_send_wc *mad_send_wc)
 {
diff --git a/include/rdma/ib_sa.h b/include/rdma/ib_sa.h
index cdc1c81aa275..384041669489 100644
--- a/include/rdma/ib_sa.h
+++ b/include/rdma/ib_sa.h
@@ -94,6 +94,8 @@ enum ib_sa_selector {
 	IB_SA_BEST = 3
 };
 
+#define IB_SA_CAP_MASK2_SENDONLY_FULL_MEM_SUPPORT	BIT(12)
+
 /*
  * Structures for SA records are named "struct ib_sa_xxx_rec."  No
  * attempt is made to pack structures to match the physical layout of
@@ -439,4 +441,14 @@ int ib_sa_guid_info_rec_query(struct ib_sa_client *client,
 			      void *context,
 			      struct ib_sa_query **sa_query);
 
+/* Support get SA ClassPortInfo */
+int ib_sa_classport_info_rec_query(struct ib_sa_client *client,
+				   struct ib_device *device, u8 port_num,
+				   int timeout_ms, gfp_t gfp_mask,
+				   void (*callback)(int status,
+						    struct ib_class_port_info *resp,
+						    void *context),
+				   void *context,
+				   struct ib_sa_query **sa_query);
+
 #endif /* IB_SA_H */
-- 
cgit v1.2.3


From 0eb626590dcf1280c6d01a784e9d53a3de6d5e8e Mon Sep 17 00:00:00 2001
From: Dennis Dalessandro <dennis.dalessandro@intel.com>
Date: Thu, 19 May 2016 05:25:50 -0700
Subject: IB/hfi1: Remove multiple device cdev

hfi1 current exports a cdev that can be used to target all of the hfi's
in the system. However there is a problem with this approach in
that the devices could be on different subnets. This is a problem that
user space can figure out and explicitly tell the driver on which device
to create a context.

Remove the multi-purpose cdev leaving a dedicated cdev for each port.
Also remove the striping capability that is dependent upon the user
choosing the multi-purpose cdev. It is now up to user space to determine
how to stripe contexts.

Reviewed-by: Dean Luick <dean.luick@intel.com>
Reviewed-by: Mitko Haralanov <mitko.haralanov@intel.com>
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
---
 drivers/staging/rdma/hfi1/file_ops.c | 102 +++++++----------------------------
 include/uapi/rdma/hfi/hfi1_user.h    |  19 +------
 2 files changed, 21 insertions(+), 100 deletions(-)

(limited to 'include')

diff --git a/drivers/staging/rdma/hfi1/file_ops.c b/drivers/staging/rdma/hfi1/file_ops.c
index c1c5bf82addb..518cd891b63d 100644
--- a/drivers/staging/rdma/hfi1/file_ops.c
+++ b/drivers/staging/rdma/hfi1/file_ops.c
@@ -86,8 +86,7 @@ static int get_ctxt_info(struct file *, void __user *, __u32);
 static int get_base_info(struct file *, void __user *, __u32);
 static int setup_ctxt(struct file *);
 static int setup_subctxt(struct hfi1_ctxtdata *);
-static int get_user_context(struct file *, struct hfi1_user_info *,
-			    int, unsigned);
+static int get_user_context(struct file *, struct hfi1_user_info *, int);
 static int find_shared_ctxt(struct file *, const struct hfi1_user_info *);
 static int allocate_ctxt(struct file *, struct hfi1_devdata *,
 			 struct hfi1_user_info *);
@@ -836,7 +835,7 @@ static u64 kvirt_to_phys(void *addr)
 static int assign_ctxt(struct file *fp, struct hfi1_user_info *uinfo)
 {
 	int i_minor, ret = 0;
-	unsigned swmajor, swminor, alg = HFI1_ALG_ACROSS;
+	unsigned int swmajor, swminor;
 
 	swmajor = uinfo->userversion >> 16;
 	if (swmajor != HFI1_USER_SWMAJOR) {
@@ -846,9 +845,6 @@ static int assign_ctxt(struct file *fp, struct hfi1_user_info *uinfo)
 
 	swminor = uinfo->userversion & 0xffff;
 
-	if (uinfo->hfi1_alg < HFI1_ALG_COUNT)
-		alg = uinfo->hfi1_alg;
-
 	mutex_lock(&hfi1_mutex);
 	/* First, lets check if we need to setup a shared context? */
 	if (uinfo->subctxt_cnt) {
@@ -868,7 +864,7 @@ static int assign_ctxt(struct file *fp, struct hfi1_user_info *uinfo)
 	 */
 	if (!ret) {
 		i_minor = iminor(file_inode(fp)) - HFI1_USER_MINOR_BASE;
-		ret = get_user_context(fp, uinfo, i_minor - 1, alg);
+		ret = get_user_context(fp, uinfo, i_minor);
 	}
 done_unlock:
 	mutex_unlock(&hfi1_mutex);
@@ -876,71 +872,26 @@ done:
 	return ret;
 }
 
-/* return true if the device available for general use */
-static int usable_device(struct hfi1_devdata *dd)
-{
-	struct hfi1_pportdata *ppd = dd->pport;
-
-	return driver_lstate(ppd) == IB_PORT_ACTIVE;
-}
-
 static int get_user_context(struct file *fp, struct hfi1_user_info *uinfo,
-			    int devno, unsigned alg)
+			    int devno)
 {
 	struct hfi1_devdata *dd = NULL;
-	int ret = 0, devmax, npresent, nup, dev;
+	int devmax, npresent, nup;
 
 	devmax = hfi1_count_units(&npresent, &nup);
-	if (!npresent) {
-		ret = -ENXIO;
-		goto done;
-	}
-	if (!nup) {
-		ret = -ENETDOWN;
-		goto done;
-	}
-	if (devno >= 0) {
-		dd = hfi1_lookup(devno);
-		if (!dd)
-			ret = -ENODEV;
-		else if (!dd->freectxts)
-			ret = -EBUSY;
-	} else {
-		struct hfi1_devdata *pdd;
-
-		if (alg == HFI1_ALG_ACROSS) {
-			unsigned free = 0U;
-
-			for (dev = 0; dev < devmax; dev++) {
-				pdd = hfi1_lookup(dev);
-				if (!pdd)
-					continue;
-				if (!usable_device(pdd))
-					continue;
-				if (pdd->freectxts &&
-				    pdd->freectxts > free) {
-					dd = pdd;
-					free = pdd->freectxts;
-				}
-			}
-		} else {
-			for (dev = 0; dev < devmax; dev++) {
-				pdd = hfi1_lookup(dev);
-				if (!pdd)
-					continue;
-				if (!usable_device(pdd))
-					continue;
-				if (pdd->freectxts) {
-					dd = pdd;
-					break;
-				}
-			}
-		}
-		if (!dd)
-			ret = -EBUSY;
-	}
-done:
-	return ret ? ret : allocate_ctxt(fp, dd, uinfo);
+	if (!npresent)
+		return -ENXIO;
+
+	if (!nup)
+		return -ENETDOWN;
+
+	dd = hfi1_lookup(devno);
+	if (!dd)
+		return -ENODEV;
+	else if (!dd->freectxts)
+		return -EBUSY;
+
+	return allocate_ctxt(fp, dd, uinfo);
 }
 
 static int find_shared_ctxt(struct file *fp,
@@ -1698,15 +1649,8 @@ static const struct file_operations ui_file_ops = {
 #define UI_OFFSET 192	/* device minor offset for UI devices */
 static int create_ui = 1;
 
-static struct cdev wildcard_cdev;
-static struct device *wildcard_device;
-
-static atomic_t user_count = ATOMIC_INIT(0);
-
 static void user_remove(struct hfi1_devdata *dd)
 {
-	if (atomic_dec_return(&user_count) == 0)
-		hfi1_cdev_cleanup(&wildcard_cdev, &wildcard_device);
 
 	hfi1_cdev_cleanup(&dd->user_cdev, &dd->user_device);
 	hfi1_cdev_cleanup(&dd->ui_cdev, &dd->ui_device);
@@ -1717,16 +1661,8 @@ static int user_add(struct hfi1_devdata *dd)
 	char name[10];
 	int ret;
 
-	if (atomic_inc_return(&user_count) == 1) {
-		ret = hfi1_cdev_init(0, class_name(), &hfi1_file_ops,
-				     &wildcard_cdev, &wildcard_device,
-				     true);
-		if (ret)
-			goto done;
-	}
-
 	snprintf(name, sizeof(name), "%s_%d", class_name(), dd->unit);
-	ret = hfi1_cdev_init(dd->unit + 1, name, &hfi1_file_ops,
+	ret = hfi1_cdev_init(dd->unit, name, &hfi1_file_ops,
 			     &dd->user_cdev, &dd->user_device,
 			     true);
 	if (ret)
diff --git a/include/uapi/rdma/hfi/hfi1_user.h b/include/uapi/rdma/hfi/hfi1_user.h
index a533cecab14f..09558999ca1d 100644
--- a/include/uapi/rdma/hfi/hfi1_user.h
+++ b/include/uapi/rdma/hfi/hfi1_user.h
@@ -75,7 +75,7 @@
  * may not be implemented; the user code must deal with this if it
  * cares, or it must abort after initialization reports the difference.
  */
-#define HFI1_USER_SWMINOR 0
+#define HFI1_USER_SWMINOR 1
 
 /*
  * Set of HW and driver capability/feature bits.
@@ -107,19 +107,6 @@
 #define HFI1_RCVHDR_ENTSIZE_16   (1UL << 1)
 #define HFI1_RCVDHR_ENTSIZE_32   (1UL << 2)
 
-/*
- * If the unit is specified via open, HFI choice is fixed.  If port is
- * specified, it's also fixed.  Otherwise we try to spread contexts
- * across ports and HFIs, using different algorithms.  WITHIN is
- * the old default, prior to this mechanism.
- */
-#define HFI1_ALG_ACROSS 0 /* round robin contexts across HFIs, then
-			  * ports; this is the default */
-#define HFI1_ALG_WITHIN 1 /* use all contexts on an HFI (round robin
-			  * active ports within), then next HFI */
-#define HFI1_ALG_COUNT  2 /* number of algorithm choices */
-
-
 /* User commands. */
 #define HFI1_CMD_ASSIGN_CTXT     1	/* allocate HFI and context */
 #define HFI1_CMD_CTXT_INFO       2	/* find out what resources we got */
@@ -199,9 +186,7 @@ struct hfi1_user_info {
 	 * Should be set to HFI1_USER_SWVERSION.
 	 */
 	__u32 userversion;
-	__u16 pad;
-	/* HFI selection algorithm, if unit has not selected */
-	__u16 hfi1_alg;
+	__u32 pad;
 	/*
 	 * If two or more processes wish to share a context, each process
 	 * must set the subcontext_cnt and subcontext_id to the same
-- 
cgit v1.2.3


From d079031742023a00e1deda0fa847d403b4b91c76 Mon Sep 17 00:00:00 2001
From: Dennis Dalessandro <dennis.dalessandro@intel.com>
Date: Thu, 19 May 2016 05:26:04 -0700
Subject: IB/hfi1: Remove EPROM functionality from data device

Remove EPROM handling from the cdev which is used for user application
data traffic.

Reviewed-by: Dean Luick <dean.luick@intel.com>
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
---
 drivers/staging/rdma/hfi1/eprom.c    | 369 -----------------------------------
 drivers/staging/rdma/hfi1/file_ops.c |  23 ---
 include/uapi/rdma/hfi/hfi1_user.h    |   7 -
 3 files changed, 399 deletions(-)

(limited to 'include')

diff --git a/drivers/staging/rdma/hfi1/eprom.c b/drivers/staging/rdma/hfi1/eprom.c
index bd8771570f81..36b77943cbfd 100644
--- a/drivers/staging/rdma/hfi1/eprom.c
+++ b/drivers/staging/rdma/hfi1/eprom.c
@@ -49,387 +49,18 @@
 #include "common.h"
 #include "eprom.h"
 
-/*
- * The EPROM is logically divided into three partitions:
- *	partition 0: the first 128K, visible from PCI ROM BAR
- *	partition 1: 4K config file (sector size)
- *	partition 2: the rest
- */
-#define P0_SIZE (128 * 1024)
-#define P1_SIZE   (4 * 1024)
-#define P1_START P0_SIZE
-#define P2_START (P0_SIZE + P1_SIZE)
-
-/* erase sizes supported by the controller */
-#define SIZE_4KB (4 * 1024)
-#define MASK_4KB (SIZE_4KB - 1)
-
-#define SIZE_32KB (32 * 1024)
-#define MASK_32KB (SIZE_32KB - 1)
-
-#define SIZE_64KB (64 * 1024)
-#define MASK_64KB (SIZE_64KB - 1)
-
-/* controller page size, in bytes */
-#define EP_PAGE_SIZE 256
-#define EEP_PAGE_MASK (EP_PAGE_SIZE - 1)
-
-/* controller commands */
 #define CMD_SHIFT 24
-#define CMD_NOP			    (0)
-#define CMD_PAGE_PROGRAM(addr)	    ((0x02 << CMD_SHIFT) | addr)
-#define CMD_READ_DATA(addr)	    ((0x03 << CMD_SHIFT) | addr)
-#define CMD_READ_SR1		    ((0x05 << CMD_SHIFT))
-#define CMD_WRITE_ENABLE	    ((0x06 << CMD_SHIFT))
-#define CMD_SECTOR_ERASE_4KB(addr)  ((0x20 << CMD_SHIFT) | addr)
-#define CMD_SECTOR_ERASE_32KB(addr) ((0x52 << CMD_SHIFT) | addr)
-#define CMD_CHIP_ERASE		    ((0x60 << CMD_SHIFT))
-#define CMD_READ_MANUF_DEV_ID	    ((0x90 << CMD_SHIFT))
 #define CMD_RELEASE_POWERDOWN_NOID  ((0xab << CMD_SHIFT))
-#define CMD_SECTOR_ERASE_64KB(addr) ((0xd8 << CMD_SHIFT) | addr)
 
 /* controller interface speeds */
 #define EP_SPEED_FULL 0x2	/* full speed */
 
-/* controller status register 1 bits */
-#define SR1_BUSY 0x1ull		/* the BUSY bit in SR1 */
-
-/* sleep length while waiting for controller */
-#define WAIT_SLEEP_US 100	/* must be larger than 5 (see usage) */
-#define COUNT_DELAY_SEC(n) ((n) * (1000000 / WAIT_SLEEP_US))
-
-/* GPIO pins */
-#define EPROM_WP_N BIT_ULL(14)	/* EPROM write line */
-
 /*
  * How long to wait for the EPROM to become available, in ms.
  * The spec 32 Mb EPROM takes around 40s to erase then write.
  * Double it for safety.
  */
 #define EPROM_TIMEOUT 80000 /* ms */
-
-/*
- * Turn on external enable line that allows writing on the flash.
- */
-static void write_enable(struct hfi1_devdata *dd)
-{
-	/* raise signal */
-	write_csr(dd, ASIC_GPIO_OUT, read_csr(dd, ASIC_GPIO_OUT) | EPROM_WP_N);
-	/* raise enable */
-	write_csr(dd, ASIC_GPIO_OE, read_csr(dd, ASIC_GPIO_OE) | EPROM_WP_N);
-}
-
-/*
- * Turn off external enable line that allows writing on the flash.
- */
-static void write_disable(struct hfi1_devdata *dd)
-{
-	/* lower signal */
-	write_csr(dd, ASIC_GPIO_OUT, read_csr(dd, ASIC_GPIO_OUT) & ~EPROM_WP_N);
-	/* lower enable */
-	write_csr(dd, ASIC_GPIO_OE, read_csr(dd, ASIC_GPIO_OE) & ~EPROM_WP_N);
-}
-
-/*
- * Wait for the device to become not busy.  Must be called after all
- * write or erase operations.
- */
-static int wait_for_not_busy(struct hfi1_devdata *dd)
-{
-	unsigned long count = 0;
-	u64 reg;
-	int ret = 0;
-
-	/* starts page mode */
-	write_csr(dd, ASIC_EEP_ADDR_CMD, CMD_READ_SR1);
-	while (1) {
-		udelay(WAIT_SLEEP_US);
-		usleep_range(WAIT_SLEEP_US - 5, WAIT_SLEEP_US + 5);
-		count++;
-		reg = read_csr(dd, ASIC_EEP_DATA);
-		if ((reg & SR1_BUSY) == 0)
-			break;
-		/* 200s is the largest time for a 128Mb device */
-		if (count > COUNT_DELAY_SEC(200)) {
-			dd_dev_err(dd, "waited too long for SPI FLASH busy to clear - failing\n");
-			ret = -ETIMEDOUT;
-			break; /* break, not goto - must stop page mode */
-		}
-	}
-
-	/* stop page mode with a NOP */
-	write_csr(dd, ASIC_EEP_ADDR_CMD, CMD_NOP);
-
-	return ret;
-}
-
-/*
- * Read the device ID from the SPI controller.
- */
-static u32 read_device_id(struct hfi1_devdata *dd)
-{
-	/* read the Manufacture Device ID */
-	write_csr(dd, ASIC_EEP_ADDR_CMD, CMD_READ_MANUF_DEV_ID);
-	return (u32)read_csr(dd, ASIC_EEP_DATA);
-}
-
-/*
- * Erase the whole flash.
- */
-static int erase_chip(struct hfi1_devdata *dd)
-{
-	int ret;
-
-	write_enable(dd);
-
-	write_csr(dd, ASIC_EEP_ADDR_CMD, CMD_WRITE_ENABLE);
-	write_csr(dd, ASIC_EEP_ADDR_CMD, CMD_CHIP_ERASE);
-	ret = wait_for_not_busy(dd);
-
-	write_disable(dd);
-
-	return ret;
-}
-
-/*
- * Erase a range.
- */
-static int erase_range(struct hfi1_devdata *dd, u32 start, u32 len)
-{
-	u32 end = start + len;
-	int ret = 0;
-
-	if (end < start)
-		return -EINVAL;
-
-	/* check the end points for the minimum erase */
-	if ((start & MASK_4KB) || (end & MASK_4KB)) {
-		dd_dev_err(dd,
-			   "%s: non-aligned range (0x%x,0x%x) for a 4KB erase\n",
-			   __func__, start, end);
-		return -EINVAL;
-	}
-
-	write_enable(dd);
-
-	while (start < end) {
-		write_csr(dd, ASIC_EEP_ADDR_CMD, CMD_WRITE_ENABLE);
-		/* check in order of largest to smallest */
-		if (((start & MASK_64KB) == 0) && (start + SIZE_64KB <= end)) {
-			write_csr(dd, ASIC_EEP_ADDR_CMD,
-				  CMD_SECTOR_ERASE_64KB(start));
-			start += SIZE_64KB;
-		} else if (((start & MASK_32KB) == 0) &&
-			   (start + SIZE_32KB <= end)) {
-			write_csr(dd, ASIC_EEP_ADDR_CMD,
-				  CMD_SECTOR_ERASE_32KB(start));
-			start += SIZE_32KB;
-		} else {	/* 4KB will work */
-			write_csr(dd, ASIC_EEP_ADDR_CMD,
-				  CMD_SECTOR_ERASE_4KB(start));
-			start += SIZE_4KB;
-		}
-		ret = wait_for_not_busy(dd);
-		if (ret)
-			goto done;
-	}
-
-done:
-	write_disable(dd);
-
-	return ret;
-}
-
-/*
- * Read a 256 byte (64 dword) EPROM page.
- * All callers have verified the offset is at a page boundary.
- */
-static void read_page(struct hfi1_devdata *dd, u32 offset, u32 *result)
-{
-	int i;
-
-	write_csr(dd, ASIC_EEP_ADDR_CMD, CMD_READ_DATA(offset));
-	for (i = 0; i < EP_PAGE_SIZE / sizeof(u32); i++)
-		result[i] = (u32)read_csr(dd, ASIC_EEP_DATA);
-	write_csr(dd, ASIC_EEP_ADDR_CMD, CMD_NOP); /* close open page */
-}
-
-/*
- * Read length bytes starting at offset.  Copy to user address addr.
- */
-static int read_length(struct hfi1_devdata *dd, u32 start, u32 len, u64 addr)
-{
-	u32 offset;
-	u32 buffer[EP_PAGE_SIZE / sizeof(u32)];
-	int ret = 0;
-
-	/* reject anything not on an EPROM page boundary */
-	if ((start & EEP_PAGE_MASK) || (len & EEP_PAGE_MASK))
-		return -EINVAL;
-
-	for (offset = 0; offset < len; offset += EP_PAGE_SIZE) {
-		read_page(dd, start + offset, buffer);
-		if (copy_to_user((void __user *)(addr + offset),
-				 buffer, EP_PAGE_SIZE)) {
-			ret = -EFAULT;
-			goto done;
-		}
-	}
-
-done:
-	return ret;
-}
-
-/*
- * Write a 256 byte (64 dword) EPROM page.
- * All callers have verified the offset is at a page boundary.
- */
-static int write_page(struct hfi1_devdata *dd, u32 offset, u32 *data)
-{
-	int i;
-
-	write_csr(dd, ASIC_EEP_ADDR_CMD, CMD_WRITE_ENABLE);
-	write_csr(dd, ASIC_EEP_DATA, data[0]);
-	write_csr(dd, ASIC_EEP_ADDR_CMD, CMD_PAGE_PROGRAM(offset));
-	for (i = 1; i < EP_PAGE_SIZE / sizeof(u32); i++)
-		write_csr(dd, ASIC_EEP_DATA, data[i]);
-	/* will close the open page */
-	return wait_for_not_busy(dd);
-}
-
-/*
- * Write length bytes starting at offset.  Read from user address addr.
- */
-static int write_length(struct hfi1_devdata *dd, u32 start, u32 len, u64 addr)
-{
-	u32 offset;
-	u32 buffer[EP_PAGE_SIZE / sizeof(u32)];
-	int ret = 0;
-
-	/* reject anything not on an EPROM page boundary */
-	if ((start & EEP_PAGE_MASK) || (len & EEP_PAGE_MASK))
-		return -EINVAL;
-
-	write_enable(dd);
-
-	for (offset = 0; offset < len; offset += EP_PAGE_SIZE) {
-		if (copy_from_user(buffer, (void __user *)(addr + offset),
-				   EP_PAGE_SIZE)) {
-			ret = -EFAULT;
-			goto done;
-		}
-		ret = write_page(dd, start + offset, buffer);
-		if (ret)
-			goto done;
-	}
-
-done:
-	write_disable(dd);
-	return ret;
-}
-
-/* convert an range composite to a length, in bytes */
-static inline u32 extract_rlen(u32 composite)
-{
-	return (composite & 0xffff) * EP_PAGE_SIZE;
-}
-
-/* convert an range composite to a start, in bytes */
-static inline u32 extract_rstart(u32 composite)
-{
-	return (composite >> 16) * EP_PAGE_SIZE;
-}
-
-/*
- * Perform the given operation on the EPROM.  Called from user space.  The
- * user credentials have already been checked.
- *
- * Return 0 on success, -ERRNO on error
- */
-int handle_eprom_command(struct file *fp, const struct hfi1_cmd *cmd)
-{
-	struct hfi1_devdata *dd;
-	u32 dev_id;
-	u32 rlen;	/* range length */
-	u32 rstart;	/* range start */
-	int i_minor;
-	int ret = 0;
-
-	/*
-	 * Map the device file to device data using the relative minor.
-	 * The device file minor number is the unit number + 1.  0 is
-	 * the generic device file - reject it.
-	 */
-	i_minor = iminor(file_inode(fp)) - HFI1_USER_MINOR_BASE;
-	if (i_minor <= 0)
-		return -EINVAL;
-	dd = hfi1_lookup(i_minor - 1);
-	if (!dd) {
-		pr_err("%s: cannot find unit %d!\n", __func__, i_minor);
-		return -EINVAL;
-	}
-
-	/* some devices do not have an EPROM */
-	if (!dd->eprom_available)
-		return -EOPNOTSUPP;
-
-	ret = acquire_chip_resource(dd, CR_EPROM, EPROM_TIMEOUT);
-	if (ret) {
-		dd_dev_err(dd, "%s: unable to acquire EPROM resource\n",
-			   __func__);
-		goto done_asic;
-	}
-
-	dd_dev_info(dd, "%s: cmd: type %d, len 0x%x, addr 0x%016llx\n",
-		    __func__, cmd->type, cmd->len, cmd->addr);
-
-	switch (cmd->type) {
-	case HFI1_CMD_EP_INFO:
-		if (cmd->len != sizeof(u32)) {
-			ret = -ERANGE;
-			break;
-		}
-		dev_id = read_device_id(dd);
-		/* addr points to a u32 user buffer */
-		if (copy_to_user((void __user *)cmd->addr, &dev_id,
-				 sizeof(u32)))
-			ret = -EFAULT;
-		break;
-
-	case HFI1_CMD_EP_ERASE_CHIP:
-		ret = erase_chip(dd);
-		break;
-
-	case HFI1_CMD_EP_ERASE_RANGE:
-		rlen = extract_rlen(cmd->len);
-		rstart = extract_rstart(cmd->len);
-		ret = erase_range(dd, rstart, rlen);
-		break;
-
-	case HFI1_CMD_EP_READ_RANGE:
-		rlen = extract_rlen(cmd->len);
-		rstart = extract_rstart(cmd->len);
-		ret = read_length(dd, rstart, rlen, cmd->addr);
-		break;
-
-	case HFI1_CMD_EP_WRITE_RANGE:
-		rlen = extract_rlen(cmd->len);
-		rstart = extract_rstart(cmd->len);
-		ret = write_length(dd, rstart, rlen, cmd->addr);
-		break;
-
-	default:
-		dd_dev_err(dd, "%s: unexpected command %d\n",
-			   __func__, cmd->type);
-		ret = -EINVAL;
-		break;
-	}
-
-	release_chip_resource(dd, CR_EPROM);
-done_asic:
-	return ret;
-}
-
 /*
  * Initialize the EPROM handler.
  */
diff --git a/drivers/staging/rdma/hfi1/file_ops.c b/drivers/staging/rdma/hfi1/file_ops.c
index eb66befd51fc..2eddd3305c6a 100644
--- a/drivers/staging/rdma/hfi1/file_ops.c
+++ b/drivers/staging/rdma/hfi1/file_ops.c
@@ -189,7 +189,6 @@ static ssize_t hfi1_file_write(struct file *fp, const char __user *data,
 	void *dest = NULL;
 	__u64 user_val = 0;
 	int uctxt_required = 1;
-	int must_be_root = 0;
 
 	/* FIXME: This interface cannot continue out of staging */
 	if (WARN_ON_ONCE(!ib_safe_file_access(fp)))
@@ -234,15 +233,6 @@ static ssize_t hfi1_file_write(struct file *fp, const char __user *data,
 		copy = 0;
 		user_val = cmd.addr;
 		break;
-	case HFI1_CMD_EP_INFO:
-	case HFI1_CMD_EP_ERASE_CHIP:
-	case HFI1_CMD_EP_ERASE_RANGE:
-	case HFI1_CMD_EP_READ_RANGE:
-	case HFI1_CMD_EP_WRITE_RANGE:
-		uctxt_required = 0;	/* assigned user context not required */
-		must_be_root = 1;	/* validate user */
-		copy = 0;
-		break;
 	default:
 		ret = -EINVAL;
 		goto bail;
@@ -265,12 +255,6 @@ static ssize_t hfi1_file_write(struct file *fp, const char __user *data,
 		goto bail;
 	}
 
-	/* only root can do these operations */
-	if (must_be_root && !capable(CAP_SYS_ADMIN)) {
-		ret = -EPERM;
-		goto bail;
-	}
-
 	switch (cmd.type) {
 	case HFI1_CMD_ASSIGN_CTXT:
 		ret = assign_ctxt(fp, &uinfo);
@@ -409,13 +393,6 @@ static ssize_t hfi1_file_write(struct file *fp, const char __user *data,
 			sc_return_credits(sc);
 		break;
 	}
-	case HFI1_CMD_EP_INFO:
-	case HFI1_CMD_EP_ERASE_CHIP:
-	case HFI1_CMD_EP_ERASE_RANGE:
-	case HFI1_CMD_EP_READ_RANGE:
-	case HFI1_CMD_EP_WRITE_RANGE:
-		ret = handle_eprom_command(fp, &cmd);
-		break;
 	}
 
 	if (ret >= 0)
diff --git a/include/uapi/rdma/hfi/hfi1_user.h b/include/uapi/rdma/hfi/hfi1_user.h
index 09558999ca1d..3e3680d11b25 100644
--- a/include/uapi/rdma/hfi/hfi1_user.h
+++ b/include/uapi/rdma/hfi/hfi1_user.h
@@ -122,13 +122,6 @@
 #define HFI1_CMD_SET_PKEY        11     /* set context's pkey */
 #define HFI1_CMD_CTXT_RESET      12     /* reset context's HW send context */
 #define HFI1_CMD_TID_INVAL_READ  13     /* read TID cache invalidations */
-/* separate EPROM commands from normal PSM commands */
-#define HFI1_CMD_EP_INFO         64      /* read EPROM device ID */
-#define HFI1_CMD_EP_ERASE_CHIP   65      /* erase whole EPROM */
-/* range 66-74 no longer used */
-#define HFI1_CMD_EP_ERASE_RANGE  75      /* erase EPROM range */
-#define HFI1_CMD_EP_READ_RANGE   76      /* read EPROM range */
-#define HFI1_CMD_EP_WRITE_RANGE  77      /* write EPROM range */
 
 #define _HFI1_EVENT_FROZEN_BIT         0
 #define _HFI1_EVENT_LINKDOWN_BIT       1
-- 
cgit v1.2.3


From ac56f162d4586885238da05131c4e6bbdb536d4e Mon Sep 17 00:00:00 2001
From: Dennis Dalessandro <dennis.dalessandro@intel.com>
Date: Thu, 19 May 2016 05:26:17 -0700
Subject: IB/hfi1: Remove unused user command

The HFI1_CMD_SDMA_STATUS_UPD command was never implemented it has no
reason to live in the driver. Remove it.

Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Mitko Haralanov <mitko.haralanov@intel.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
---
 drivers/staging/rdma/hfi1/file_ops.c | 3 ---
 include/uapi/rdma/hfi/hfi1_user.h    | 1 -
 2 files changed, 4 deletions(-)

(limited to 'include')

diff --git a/drivers/staging/rdma/hfi1/file_ops.c b/drivers/staging/rdma/hfi1/file_ops.c
index 4287c806642e..322e55f7b4b4 100644
--- a/drivers/staging/rdma/hfi1/file_ops.c
+++ b/drivers/staging/rdma/hfi1/file_ops.c
@@ -213,7 +213,6 @@ static ssize_t hfi1_file_write(struct file *fp, const char __user *data,
 		copy = sizeof(uinfo);
 		dest = &uinfo;
 		break;
-	case HFI1_CMD_SDMA_STATUS_UPD:
 	case HFI1_CMD_CREDIT_UPD:
 		copy = 0;
 		break;
@@ -273,8 +272,6 @@ static ssize_t hfi1_file_write(struct file *fp, const char __user *data,
 		ret = get_base_info(fp, (void __user *)(unsigned long)
 				    user_val, cmd.len);
 		break;
-	case HFI1_CMD_SDMA_STATUS_UPD:
-		break;
 	case HFI1_CMD_CREDIT_UPD:
 		if (uctxt && uctxt->sc)
 			sc_return_credits(uctxt->sc);
diff --git a/include/uapi/rdma/hfi/hfi1_user.h b/include/uapi/rdma/hfi/hfi1_user.h
index 3e3680d11b25..aa48fbe016af 100644
--- a/include/uapi/rdma/hfi/hfi1_user.h
+++ b/include/uapi/rdma/hfi/hfi1_user.h
@@ -114,7 +114,6 @@
 #define HFI1_CMD_TID_UPDATE      4	/* update expected TID entries */
 #define HFI1_CMD_TID_FREE        5	/* free expected TID entries */
 #define HFI1_CMD_CREDIT_UPD      6	/* force an update of PIO credit */
-#define HFI1_CMD_SDMA_STATUS_UPD 7      /* force update of SDMA status ring */
 
 #define HFI1_CMD_RECV_CTRL       8	/* control receipt of packets */
 #define HFI1_CMD_POLL_TYPE       9	/* set the kind of polling we want */
-- 
cgit v1.2.3


From 8d970cf991a6c38a5566572979487b906d643740 Mon Sep 17 00:00:00 2001
From: Dennis Dalessandro <dennis.dalessandro@intel.com>
Date: Thu, 19 May 2016 05:26:24 -0700
Subject: IB/hfi1: Add ioctl() interface for user commands

IOCTL is more suited to what user space commands need to do than the
write() interface. Add IOCTL definitions for all existing write commands
and the handling for those. The write() interface will be removed in a
follow on patch.

Reviewed-by: Mitko Haralanov <mitko.haralanov@intel.com>
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
---
 drivers/staging/rdma/hfi1/common.h   |   5 +-
 drivers/staging/rdma/hfi1/file_ops.c | 204 +++++++++++++++++++++++++++++++++++
 include/uapi/rdma/hfi/hfi1_user.h    |  45 ++++++++
 3 files changed, 253 insertions(+), 1 deletion(-)

(limited to 'include')

diff --git a/drivers/staging/rdma/hfi1/common.h b/drivers/staging/rdma/hfi1/common.h
index e9b6bb322025..fcc9c217a97a 100644
--- a/drivers/staging/rdma/hfi1/common.h
+++ b/drivers/staging/rdma/hfi1/common.h
@@ -178,7 +178,8 @@
 		     HFI1_CAP_PKEY_CHECK |		\
 		     HFI1_CAP_NO_INTEGRITY)
 
-#define HFI1_USER_SWVERSION ((HFI1_USER_SWMAJOR << 16) | HFI1_USER_SWMINOR)
+#define HFI1_USER_SWVERSION ((HFI1_USER_SWMAJOR << HFI1_SWMAJOR_SHIFT) | \
+			     HFI1_USER_SWMINOR)
 
 #ifndef HFI1_KERN_TYPE
 #define HFI1_KERN_TYPE 0
@@ -349,6 +350,8 @@ struct hfi1_message_header {
 #define HFI1_BECN_MASK 1
 #define HFI1_BECN_SMASK BIT(HFI1_BECN_SHIFT)
 
+#define HFI1_PSM_IOC_BASE_SEQ 0x0
+
 static inline __u64 rhf_to_cpu(const __le32 *rbuf)
 {
 	return __le64_to_cpu(*((__le64 *)rbuf));
diff --git a/drivers/staging/rdma/hfi1/file_ops.c b/drivers/staging/rdma/hfi1/file_ops.c
index 322e55f7b4b4..a338238bef92 100644
--- a/drivers/staging/rdma/hfi1/file_ops.c
+++ b/drivers/staging/rdma/hfi1/file_ops.c
@@ -96,6 +96,8 @@ static int user_event_ack(struct hfi1_ctxtdata *, int, unsigned long);
 static int set_ctxt_pkey(struct hfi1_ctxtdata *, unsigned, u16);
 static int manage_rcvq(struct hfi1_ctxtdata *, unsigned, int);
 static int vma_fault(struct vm_area_struct *, struct vm_fault *);
+static long hfi1_file_ioctl(struct file *fp, unsigned int cmd,
+			    unsigned long arg);
 
 static const struct file_operations hfi1_file_ops = {
 	.owner = THIS_MODULE,
@@ -103,6 +105,7 @@ static const struct file_operations hfi1_file_ops = {
 	.write_iter = hfi1_write_iter,
 	.open = hfi1_file_open,
 	.release = hfi1_file_close,
+	.unlocked_ioctl = hfi1_file_ioctl,
 	.poll = hfi1_poll,
 	.mmap = hfi1_file_mmap,
 	.llseek = noop_llseek,
@@ -175,6 +178,207 @@ static int hfi1_file_open(struct inode *inode, struct file *fp)
 	return fp->private_data ? 0 : -ENOMEM;
 }
 
+static long hfi1_file_ioctl(struct file *fp, unsigned int cmd,
+			    unsigned long arg)
+{
+	struct hfi1_filedata *fd = fp->private_data;
+	struct hfi1_ctxtdata *uctxt = fd->uctxt;
+	struct hfi1_user_info uinfo;
+	struct hfi1_tid_info tinfo;
+	int ret = 0;
+	unsigned long addr;
+	int uval = 0;
+	unsigned long ul_uval = 0;
+	u16 uval16 = 0;
+
+	if (cmd != HFI1_IOCTL_ASSIGN_CTXT &&
+	    cmd != HFI1_IOCTL_GET_VERS &&
+	    !uctxt)
+		return -EINVAL;
+
+	switch (cmd) {
+	case HFI1_IOCTL_ASSIGN_CTXT:
+		if (copy_from_user(&uinfo,
+				   (struct hfi1_user_info __user *)arg,
+				   sizeof(uinfo)))
+			return -EFAULT;
+
+		ret = assign_ctxt(fp, &uinfo);
+		if (ret < 0)
+			return ret;
+		setup_ctxt(fp);
+		if (ret)
+			return ret;
+		ret = user_init(fp);
+		break;
+	case HFI1_IOCTL_CTXT_INFO:
+		ret = get_ctxt_info(fp, (void __user *)(unsigned long)arg,
+				    sizeof(struct hfi1_ctxt_info));
+		break;
+	case HFI1_IOCTL_USER_INFO:
+		ret = get_base_info(fp, (void __user *)(unsigned long)arg,
+				    sizeof(struct hfi1_base_info));
+		break;
+	case HFI1_IOCTL_CREDIT_UPD:
+		if (uctxt && uctxt->sc)
+			sc_return_credits(uctxt->sc);
+		break;
+
+	case HFI1_IOCTL_TID_UPDATE:
+		if (copy_from_user(&tinfo,
+				   (struct hfi11_tid_info __user *)arg,
+				   sizeof(tinfo)))
+			return -EFAULT;
+
+		ret = hfi1_user_exp_rcv_setup(fp, &tinfo);
+		if (!ret) {
+			/*
+			 * Copy the number of tidlist entries we used
+			 * and the length of the buffer we registered.
+			 * These fields are adjacent in the structure so
+			 * we can copy them at the same time.
+			 */
+			addr = arg + offsetof(struct hfi1_tid_info, tidcnt);
+			if (copy_to_user((void __user *)addr, &tinfo.tidcnt,
+					 sizeof(tinfo.tidcnt) +
+					 sizeof(tinfo.length)))
+				ret = -EFAULT;
+		}
+		break;
+
+	case HFI1_IOCTL_TID_FREE:
+		if (copy_from_user(&tinfo,
+				   (struct hfi11_tid_info __user *)arg,
+				   sizeof(tinfo)))
+			return -EFAULT;
+
+		ret = hfi1_user_exp_rcv_clear(fp, &tinfo);
+		if (ret)
+			break;
+		addr = arg + offsetof(struct hfi1_tid_info, tidcnt);
+		if (copy_to_user((void __user *)addr, &tinfo.tidcnt,
+				 sizeof(tinfo.tidcnt)))
+			ret = -EFAULT;
+		break;
+
+	case HFI1_IOCTL_TID_INVAL_READ:
+		if (copy_from_user(&tinfo,
+				   (struct hfi11_tid_info __user *)arg,
+				   sizeof(tinfo)))
+			return -EFAULT;
+
+		ret = hfi1_user_exp_rcv_invalid(fp, &tinfo);
+		if (ret)
+			break;
+		addr = arg + offsetof(struct hfi1_tid_info, tidcnt);
+		if (copy_to_user((void __user *)addr, &tinfo.tidcnt,
+				 sizeof(tinfo.tidcnt)))
+			ret = -EFAULT;
+		break;
+
+	case HFI1_IOCTL_RECV_CTRL:
+		ret = get_user(uval, (int __user *)arg);
+		if (ret != 0)
+			return -EFAULT;
+		ret = manage_rcvq(uctxt, fd->subctxt, uval);
+		break;
+
+	case HFI1_IOCTL_POLL_TYPE:
+		ret = get_user(uval, (int __user *)arg);
+		if (ret != 0)
+			return -EFAULT;
+		uctxt->poll_type = (typeof(uctxt->poll_type))uval;
+		break;
+
+	case HFI1_IOCTL_ACK_EVENT:
+		ret = get_user(ul_uval, (unsigned long __user *)arg);
+		if (ret != 0)
+			return -EFAULT;
+		ret = user_event_ack(uctxt, fd->subctxt, ul_uval);
+		break;
+
+	case HFI1_IOCTL_SET_PKEY:
+		ret = get_user(uval16, (u16 __user *)arg);
+		if (ret != 0)
+			return -EFAULT;
+		if (HFI1_CAP_IS_USET(PKEY_CHECK))
+			ret = set_ctxt_pkey(uctxt, fd->subctxt, uval16);
+		else
+			return -EPERM;
+		break;
+
+	case HFI1_IOCTL_CTXT_RESET: {
+		struct send_context *sc;
+		struct hfi1_devdata *dd;
+
+		if (!uctxt || !uctxt->dd || !uctxt->sc)
+			return -EINVAL;
+
+		/*
+		 * There is no protection here. User level has to
+		 * guarantee that no one will be writing to the send
+		 * context while it is being re-initialized.
+		 * If user level breaks that guarantee, it will break
+		 * it's own context and no one else's.
+		 */
+		dd = uctxt->dd;
+		sc = uctxt->sc;
+		/*
+		 * Wait until the interrupt handler has marked the
+		 * context as halted or frozen. Report error if we time
+		 * out.
+		 */
+		wait_event_interruptible_timeout(
+			sc->halt_wait, (sc->flags & SCF_HALTED),
+			msecs_to_jiffies(SEND_CTXT_HALT_TIMEOUT));
+		if (!(sc->flags & SCF_HALTED))
+			return -ENOLCK;
+
+		/*
+		 * If the send context was halted due to a Freeze,
+		 * wait until the device has been "unfrozen" before
+		 * resetting the context.
+		 */
+		if (sc->flags & SCF_FROZEN) {
+			wait_event_interruptible_timeout(
+				dd->event_queue,
+				!(ACCESS_ONCE(dd->flags) & HFI1_FROZEN),
+				msecs_to_jiffies(SEND_CTXT_HALT_TIMEOUT));
+			if (dd->flags & HFI1_FROZEN)
+				return -ENOLCK;
+
+			if (dd->flags & HFI1_FORCED_FREEZE)
+				/*
+				 * Don't allow context reset if we are into
+				 * forced freeze
+				 */
+				return -ENODEV;
+
+			sc_disable(sc);
+			ret = sc_enable(sc);
+			hfi1_rcvctrl(dd, HFI1_RCVCTRL_CTXT_ENB,
+				     uctxt->ctxt);
+		} else {
+			ret = sc_restart(sc);
+		}
+		if (!ret)
+			sc_return_credits(sc);
+		break;
+	}
+
+	case HFI1_IOCTL_GET_VERS:
+		uval = HFI1_USER_SWVERSION;
+		if (put_user(uval, (int __user *)arg))
+			return -EFAULT;
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	return ret;
+}
+
 static ssize_t hfi1_file_write(struct file *fp, const char __user *data,
 			       size_t count, loff_t *offset)
 {
diff --git a/include/uapi/rdma/hfi/hfi1_user.h b/include/uapi/rdma/hfi/hfi1_user.h
index aa48fbe016af..9784159c9d52 100644
--- a/include/uapi/rdma/hfi/hfi1_user.h
+++ b/include/uapi/rdma/hfi/hfi1_user.h
@@ -77,6 +77,11 @@
  */
 #define HFI1_USER_SWMINOR 1
 
+/*
+ * We will encode the major/minor inside a single 32bit version number.
+ */
+#define HFI1_SWMAJOR_SHIFT 16
+
 /*
  * Set of HW and driver capability/feature bits.
  * These bit values are used to configure enabled/disabled HW and
@@ -121,6 +126,46 @@
 #define HFI1_CMD_SET_PKEY        11     /* set context's pkey */
 #define HFI1_CMD_CTXT_RESET      12     /* reset context's HW send context */
 #define HFI1_CMD_TID_INVAL_READ  13     /* read TID cache invalidations */
+#define HFI1_CMD_GET_VERS	 14	/* get the version of the user cdev */
+
+/*
+ * User IOCTLs can not go above 128 if they do then see common.h and change the
+ * base for the snoop ioctl
+ */
+#define IB_IOCTL_MAGIC 0x1b /* See Documentation/ioctl/ioctl-number.txt */
+
+/*
+ * Make the ioctls occupy the last 0xf0-0xff portion of the IB range
+ */
+#define __NUM(cmd) (HFI1_CMD_##cmd + 0xe0)
+
+struct hfi1_cmd;
+#define HFI1_IOCTL_ASSIGN_CTXT \
+	_IOWR(IB_IOCTL_MAGIC, __NUM(ASSIGN_CTXT), struct hfi1_user_info)
+#define HFI1_IOCTL_CTXT_INFO \
+	_IOW(IB_IOCTL_MAGIC, __NUM(CTXT_INFO), struct hfi1_ctxt_info)
+#define HFI1_IOCTL_USER_INFO \
+	_IOW(IB_IOCTL_MAGIC, __NUM(USER_INFO), struct hfi1_base_info)
+#define HFI1_IOCTL_TID_UPDATE \
+	_IOWR(IB_IOCTL_MAGIC, __NUM(TID_UPDATE), struct hfi1_tid_info)
+#define HFI1_IOCTL_TID_FREE \
+	_IOWR(IB_IOCTL_MAGIC, __NUM(TID_FREE), struct hfi1_tid_info)
+#define HFI1_IOCTL_CREDIT_UPD \
+	_IO(IB_IOCTL_MAGIC, __NUM(CREDIT_UPD))
+#define HFI1_IOCTL_RECV_CTRL \
+	_IOW(IB_IOCTL_MAGIC, __NUM(RECV_CTRL), int)
+#define HFI1_IOCTL_POLL_TYPE \
+	_IOW(IB_IOCTL_MAGIC, __NUM(POLL_TYPE), int)
+#define HFI1_IOCTL_ACK_EVENT \
+	_IOW(IB_IOCTL_MAGIC, __NUM(ACK_EVENT), unsigned long)
+#define HFI1_IOCTL_SET_PKEY \
+	_IOW(IB_IOCTL_MAGIC, __NUM(SET_PKEY), __u16)
+#define HFI1_IOCTL_CTXT_RESET \
+	_IO(IB_IOCTL_MAGIC, __NUM(CTXT_RESET))
+#define HFI1_IOCTL_TID_INVAL_READ \
+	_IOWR(IB_IOCTL_MAGIC, __NUM(TID_INVAL_READ), struct hfi1_tid_info)
+#define HFI1_IOCTL_GET_VERS \
+	_IOR(IB_IOCTL_MAGIC, __NUM(GET_VERS), int)
 
 #define _HFI1_EVENT_FROZEN_BIT         0
 #define _HFI1_EVENT_LINKDOWN_BIT       1
-- 
cgit v1.2.3


From 380fb942888e7afc3420ce195a5188ff73b5a782 Mon Sep 17 00:00:00 2001
From: Dennis Dalessandro <dennis.dalessandro@intel.com>
Date: Thu, 19 May 2016 05:26:31 -0700
Subject: IB/hfi1: Remove write(), use ioctl() for user cmds

Remove the write() handler for user space commands now that ioctl
handling is available. User apps will need to change to use ioctl from
this point forward.

Reviewed-by: Mitko Haralanov <mitko.haralanov@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
---
 drivers/staging/rdma/hfi1/file_ops.c | 226 -----------------------------------
 include/uapi/rdma/hfi/hfi1_user.h    |   8 +-
 2 files changed, 1 insertion(+), 233 deletions(-)

(limited to 'include')

diff --git a/drivers/staging/rdma/hfi1/file_ops.c b/drivers/staging/rdma/hfi1/file_ops.c
index a338238bef92..c46eadadf3bf 100644
--- a/drivers/staging/rdma/hfi1/file_ops.c
+++ b/drivers/staging/rdma/hfi1/file_ops.c
@@ -72,8 +72,6 @@
  */
 static int hfi1_file_open(struct inode *, struct file *);
 static int hfi1_file_close(struct inode *, struct file *);
-static ssize_t hfi1_file_write(struct file *, const char __user *,
-			       size_t, loff_t *);
 static ssize_t hfi1_write_iter(struct kiocb *, struct iov_iter *);
 static unsigned int hfi1_poll(struct file *, struct poll_table_struct *);
 static int hfi1_file_mmap(struct file *, struct vm_area_struct *);
@@ -101,7 +99,6 @@ static long hfi1_file_ioctl(struct file *fp, unsigned int cmd,
 
 static const struct file_operations hfi1_file_ops = {
 	.owner = THIS_MODULE,
-	.write = hfi1_file_write,
 	.write_iter = hfi1_write_iter,
 	.open = hfi1_file_open,
 	.release = hfi1_file_close,
@@ -379,229 +376,6 @@ static long hfi1_file_ioctl(struct file *fp, unsigned int cmd,
 	return ret;
 }
 
-static ssize_t hfi1_file_write(struct file *fp, const char __user *data,
-			       size_t count, loff_t *offset)
-{
-	const struct hfi1_cmd __user *ucmd;
-	struct hfi1_filedata *fd = fp->private_data;
-	struct hfi1_ctxtdata *uctxt = fd->uctxt;
-	struct hfi1_cmd cmd;
-	struct hfi1_user_info uinfo;
-	struct hfi1_tid_info tinfo;
-	unsigned long addr;
-	ssize_t consumed = 0, copy = 0, ret = 0;
-	void *dest = NULL;
-	__u64 user_val = 0;
-	int uctxt_required = 1;
-
-	/* FIXME: This interface cannot continue out of staging */
-	if (WARN_ON_ONCE(!ib_safe_file_access(fp)))
-		return -EACCES;
-
-	if (count < sizeof(cmd)) {
-		ret = -EINVAL;
-		goto bail;
-	}
-
-	ucmd = (const struct hfi1_cmd __user *)data;
-	if (copy_from_user(&cmd, ucmd, sizeof(cmd))) {
-		ret = -EFAULT;
-		goto bail;
-	}
-
-	consumed = sizeof(cmd);
-
-	switch (cmd.type) {
-	case HFI1_CMD_ASSIGN_CTXT:
-		uctxt_required = 0;	/* assigned user context not required */
-		copy = sizeof(uinfo);
-		dest = &uinfo;
-		break;
-	case HFI1_CMD_CREDIT_UPD:
-		copy = 0;
-		break;
-	case HFI1_CMD_TID_UPDATE:
-	case HFI1_CMD_TID_FREE:
-	case HFI1_CMD_TID_INVAL_READ:
-		copy = sizeof(tinfo);
-		dest = &tinfo;
-		break;
-	case HFI1_CMD_USER_INFO:
-	case HFI1_CMD_RECV_CTRL:
-	case HFI1_CMD_POLL_TYPE:
-	case HFI1_CMD_ACK_EVENT:
-	case HFI1_CMD_CTXT_INFO:
-	case HFI1_CMD_SET_PKEY:
-	case HFI1_CMD_CTXT_RESET:
-		copy = 0;
-		user_val = cmd.addr;
-		break;
-	default:
-		ret = -EINVAL;
-		goto bail;
-	}
-
-	/* If the command comes with user data, copy it. */
-	if (copy) {
-		if (copy_from_user(dest, (void __user *)cmd.addr, copy)) {
-			ret = -EFAULT;
-			goto bail;
-		}
-		consumed += copy;
-	}
-
-	/*
-	 * Make sure there is a uctxt when needed.
-	 */
-	if (uctxt_required && !uctxt) {
-		ret = -EINVAL;
-		goto bail;
-	}
-
-	switch (cmd.type) {
-	case HFI1_CMD_ASSIGN_CTXT:
-		ret = assign_ctxt(fp, &uinfo);
-		if (ret < 0)
-			goto bail;
-		ret = setup_ctxt(fp);
-		if (ret)
-			goto bail;
-		ret = user_init(fp);
-		break;
-	case HFI1_CMD_CTXT_INFO:
-		ret = get_ctxt_info(fp, (void __user *)(unsigned long)
-				    user_val, cmd.len);
-		break;
-	case HFI1_CMD_USER_INFO:
-		ret = get_base_info(fp, (void __user *)(unsigned long)
-				    user_val, cmd.len);
-		break;
-	case HFI1_CMD_CREDIT_UPD:
-		if (uctxt && uctxt->sc)
-			sc_return_credits(uctxt->sc);
-		break;
-	case HFI1_CMD_TID_UPDATE:
-		ret = hfi1_user_exp_rcv_setup(fp, &tinfo);
-		if (!ret) {
-			/*
-			 * Copy the number of tidlist entries we used
-			 * and the length of the buffer we registered.
-			 * These fields are adjacent in the structure so
-			 * we can copy them at the same time.
-			 */
-			addr = (unsigned long)cmd.addr +
-				offsetof(struct hfi1_tid_info, tidcnt);
-			if (copy_to_user((void __user *)addr, &tinfo.tidcnt,
-					 sizeof(tinfo.tidcnt) +
-					 sizeof(tinfo.length)))
-				ret = -EFAULT;
-		}
-		break;
-	case HFI1_CMD_TID_INVAL_READ:
-		ret = hfi1_user_exp_rcv_invalid(fp, &tinfo);
-		if (ret)
-			break;
-		addr = (unsigned long)cmd.addr +
-			offsetof(struct hfi1_tid_info, tidcnt);
-		if (copy_to_user((void __user *)addr, &tinfo.tidcnt,
-				 sizeof(tinfo.tidcnt)))
-			ret = -EFAULT;
-		break;
-	case HFI1_CMD_TID_FREE:
-		ret = hfi1_user_exp_rcv_clear(fp, &tinfo);
-		if (ret)
-			break;
-		addr = (unsigned long)cmd.addr +
-			offsetof(struct hfi1_tid_info, tidcnt);
-		if (copy_to_user((void __user *)addr, &tinfo.tidcnt,
-				 sizeof(tinfo.tidcnt)))
-			ret = -EFAULT;
-		break;
-	case HFI1_CMD_RECV_CTRL:
-		ret = manage_rcvq(uctxt, fd->subctxt, (int)user_val);
-		break;
-	case HFI1_CMD_POLL_TYPE:
-		uctxt->poll_type = (typeof(uctxt->poll_type))user_val;
-		break;
-	case HFI1_CMD_ACK_EVENT:
-		ret = user_event_ack(uctxt, fd->subctxt, user_val);
-		break;
-	case HFI1_CMD_SET_PKEY:
-		if (HFI1_CAP_IS_USET(PKEY_CHECK))
-			ret = set_ctxt_pkey(uctxt, fd->subctxt, user_val);
-		else
-			ret = -EPERM;
-		break;
-	case HFI1_CMD_CTXT_RESET: {
-		struct send_context *sc;
-		struct hfi1_devdata *dd;
-
-		if (!uctxt || !uctxt->dd || !uctxt->sc) {
-			ret = -EINVAL;
-			break;
-		}
-		/*
-		 * There is no protection here. User level has to
-		 * guarantee that no one will be writing to the send
-		 * context while it is being re-initialized.
-		 * If user level breaks that guarantee, it will break
-		 * it's own context and no one else's.
-		 */
-		dd = uctxt->dd;
-		sc = uctxt->sc;
-		/*
-		 * Wait until the interrupt handler has marked the
-		 * context as halted or frozen. Report error if we time
-		 * out.
-		 */
-		wait_event_interruptible_timeout(
-			sc->halt_wait, (sc->flags & SCF_HALTED),
-			msecs_to_jiffies(SEND_CTXT_HALT_TIMEOUT));
-		if (!(sc->flags & SCF_HALTED)) {
-			ret = -ENOLCK;
-			break;
-		}
-		/*
-		 * If the send context was halted due to a Freeze,
-		 * wait until the device has been "unfrozen" before
-		 * resetting the context.
-		 */
-		if (sc->flags & SCF_FROZEN) {
-			wait_event_interruptible_timeout(
-				dd->event_queue,
-				!(ACCESS_ONCE(dd->flags) & HFI1_FROZEN),
-				msecs_to_jiffies(SEND_CTXT_HALT_TIMEOUT));
-			if (dd->flags & HFI1_FROZEN) {
-				ret = -ENOLCK;
-				break;
-			}
-			if (dd->flags & HFI1_FORCED_FREEZE) {
-				/*
-				 * Don't allow context reset if we are into
-				 * forced freeze
-				 */
-				ret = -ENODEV;
-				break;
-			}
-			sc_disable(sc);
-			ret = sc_enable(sc);
-			hfi1_rcvctrl(dd, HFI1_RCVCTRL_CTXT_ENB,
-				     uctxt->ctxt);
-		} else {
-			ret = sc_restart(sc);
-		}
-		if (!ret)
-			sc_return_credits(sc);
-		break;
-	}
-	}
-
-	if (ret >= 0)
-		ret = consumed;
-bail:
-	return ret;
-}
-
 static ssize_t hfi1_write_iter(struct kiocb *kiocb, struct iov_iter *from)
 {
 	struct hfi1_filedata *fd = kiocb->ki_filp->private_data;
diff --git a/include/uapi/rdma/hfi/hfi1_user.h b/include/uapi/rdma/hfi/hfi1_user.h
index 9784159c9d52..98bebf8bef55 100644
--- a/include/uapi/rdma/hfi/hfi1_user.h
+++ b/include/uapi/rdma/hfi/hfi1_user.h
@@ -66,7 +66,7 @@
  * The major version changes when data structures change in an incompatible
  * way. The driver must be the same for initialization to succeed.
  */
-#define HFI1_USER_SWMAJOR 5
+#define HFI1_USER_SWMAJOR 6
 
 /*
  * Minor version differences are always compatible
@@ -265,12 +265,6 @@ struct hfi1_tid_info {
 	__u32 length;
 };
 
-struct hfi1_cmd {
-	__u32 type;        /* command type */
-	__u32 len;         /* length of struct pointed to by add */
-	__u64 addr;        /* pointer to user structure */
-};
-
 enum hfi1_sdma_comp_state {
 	FREE = 0,
 	QUEUED,
-- 
cgit v1.2.3


From 4c0b653335bbdfe62a5f4483c98cb5d581432917 Mon Sep 17 00:00:00 2001
From: Mike Marciniszyn <mike.marciniszyn@intel.com>
Date: Tue, 24 May 2016 12:50:34 -0700
Subject: IB/rdmavt: Max atomic value should be a u8

This matches the ib_qp_attr size and
avoids a extremely large value when the lower level
driver registers.

As part of the patch, the u8 ordinals are moved to the
end of the struct to reduce pahole noted excesses.

Reviewed-by: Mitko Haralanov <mitko.haralanov@intel.com>
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
---
 include/rdma/rdma_vt.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'include')

diff --git a/include/rdma/rdma_vt.h b/include/rdma/rdma_vt.h
index d57ceee90d26..3f12da9384a1 100644
--- a/include/rdma/rdma_vt.h
+++ b/include/rdma/rdma_vt.h
@@ -149,15 +149,15 @@ struct rvt_driver_params {
 	int qpn_res_end;
 	int nports;
 	int npkeys;
-	u8 qos_shift;
 	char cq_name[RVT_CQN_MAX];
 	int node;
-	int max_rdma_atomic;
 	int psn_mask;
 	int psn_shift;
 	int psn_modify_mask;
 	u32 core_cap_flags;
 	u32 max_mad_size;
+	u8 qos_shift;
+	u8 max_rdma_atomic;
 };
 
 /* Protection domain */
-- 
cgit v1.2.3


From 8b103e9cdee5f381bc20a8a9f9bb5be11de8e68f Mon Sep 17 00:00:00 2001
From: Mike Marciniszyn <mike.marciniszyn@intel.com>
Date: Tue, 24 May 2016 12:50:40 -0700
Subject: IB/rdamvt: Fix rdmavt s_ack_queue sizing

rdmavt allows the driver to specify the size of the ack queue, but
only uses it for the modify QP limit testing for setting the atomic
limit value.

The driver dependent size is now used to size the s_ack_queue ring
dynamicially.

Since the driver knows its size, the driver will use its define
for any ring size dependent code.

Reviewed-by: Mitko Haralanov <mitko.haralanov@intel.com>
Signed-off-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
---
 drivers/infiniband/sw/rdmavt/qp.c | 22 ++++++++++++++++++++--
 include/rdma/rdma_vt.h            |  9 +++++++++
 include/rdma/rdmavt_qp.h          |  5 +----
 3 files changed, 30 insertions(+), 6 deletions(-)

(limited to 'include')

diff --git a/drivers/infiniband/sw/rdmavt/qp.c b/drivers/infiniband/sw/rdmavt/qp.c
index dd03b23a6a7c..5fa4d4d81ee0 100644
--- a/drivers/infiniband/sw/rdmavt/qp.c
+++ b/drivers/infiniband/sw/rdmavt/qp.c
@@ -397,6 +397,7 @@ static void free_qpn(struct rvt_qpn_table *qpt, u32 qpn)
 static void rvt_clear_mr_refs(struct rvt_qp *qp, int clr_sends)
 {
 	unsigned n;
+	struct rvt_dev_info *rdi = ib_to_rvt(qp->ibqp.device);
 
 	if (test_and_clear_bit(RVT_R_REWIND_SGE, &qp->r_aflags))
 		rvt_put_ss(&qp->s_rdma_read_sge);
@@ -431,7 +432,7 @@ static void rvt_clear_mr_refs(struct rvt_qp *qp, int clr_sends)
 	if (qp->ibqp.qp_type != IB_QPT_RC)
 		return;
 
-	for (n = 0; n < ARRAY_SIZE(qp->s_ack_queue); n++) {
+	for (n = 0; n < rvt_max_atomic(rdi); n++) {
 		struct rvt_ack_entry *e = &qp->s_ack_queue[n];
 
 		if (e->opcode == IB_OPCODE_RC_RDMA_READ_REQUEST &&
@@ -569,7 +570,12 @@ static void rvt_reset_qp(struct rvt_dev_info *rdi, struct rvt_qp *qp,
 	qp->s_ssn = 1;
 	qp->s_lsn = 0;
 	qp->s_mig_state = IB_MIG_MIGRATED;
-	memset(qp->s_ack_queue, 0, sizeof(qp->s_ack_queue));
+	if (qp->s_ack_queue)
+		memset(
+			qp->s_ack_queue,
+			0,
+			rvt_max_atomic(rdi) *
+				sizeof(*qp->s_ack_queue));
 	qp->r_head_ack_queue = 0;
 	qp->s_tail_ack_queue = 0;
 	qp->s_num_rd_atomic = 0;
@@ -677,6 +683,16 @@ struct ib_qp *rvt_create_qp(struct ib_pd *ibpd,
 			goto bail_swq;
 
 		RCU_INIT_POINTER(qp->next, NULL);
+		if (init_attr->qp_type == IB_QPT_RC) {
+			qp->s_ack_queue =
+				kzalloc_node(
+					sizeof(*qp->s_ack_queue) *
+					 rvt_max_atomic(rdi),
+					gfp,
+					rdi->dparms.node);
+			if (!qp->s_ack_queue)
+				goto bail_qp;
+		}
 
 		/*
 		 * Driver needs to set up it's private QP structure and do any
@@ -857,6 +873,7 @@ bail_driver_priv:
 	rdi->driver_f.qp_priv_free(rdi, qp);
 
 bail_qp:
+	kfree(qp->s_ack_queue);
 	kfree(qp);
 
 bail_swq:
@@ -1284,6 +1301,7 @@ int rvt_destroy_qp(struct ib_qp *ibqp)
 		vfree(qp->r_rq.wq);
 	vfree(qp->s_wq);
 	rdi->driver_f.qp_priv_free(rdi, qp);
+	kfree(qp->s_ack_queue);
 	kfree(qp);
 	return 0;
 }
diff --git a/include/rdma/rdma_vt.h b/include/rdma/rdma_vt.h
index 3f12da9384a1..16274e2133cd 100644
--- a/include/rdma/rdma_vt.h
+++ b/include/rdma/rdma_vt.h
@@ -425,6 +425,15 @@ static inline unsigned rvt_get_npkeys(struct rvt_dev_info *rdi)
 	return rdi->dparms.npkeys;
 }
 
+/*
+ * Return the max atomic suitable for determining
+ * the size of the ack ring buffer in a QP.
+ */
+static inline unsigned int rvt_max_atomic(struct rvt_dev_info *rdi)
+{
+	return rdi->dparms.max_rdma_atomic + 1;
+}
+
 /*
  * Return the indexed PKEY from the port PKEY table.
  */
diff --git a/include/rdma/rdmavt_qp.h b/include/rdma/rdmavt_qp.h
index 0e1ff2abfe92..6d23b879416a 100644
--- a/include/rdma/rdmavt_qp.h
+++ b/include/rdma/rdmavt_qp.h
@@ -211,8 +211,6 @@ struct rvt_mmap_info {
 	unsigned size;
 };
 
-#define RVT_MAX_RDMA_ATOMIC	16
-
 /*
  * This structure holds the information that the send tasklet needs
  * to send a RDMA read response or atomic operation.
@@ -282,8 +280,7 @@ struct rvt_qp {
 	atomic_t refcount ____cacheline_aligned_in_smp;
 	wait_queue_head_t wait;
 
-	struct rvt_ack_entry s_ack_queue[RVT_MAX_RDMA_ATOMIC + 1]
-		____cacheline_aligned_in_smp;
+	struct rvt_ack_entry *s_ack_queue;
 	struct rvt_sge_state s_rdma_read_sge;
 
 	spinlock_t r_lock ____cacheline_aligned_in_smp;      /* used for APM */
-- 
cgit v1.2.3


From b40f4757daa1b28e586fddad76638c98e2edfc34 Mon Sep 17 00:00:00 2001
From: Christoph Lameter <cl@linux.com>
Date: Mon, 16 May 2016 12:49:33 -0500
Subject: IB/core: Make device counter infrastructure dynamic

In practice, each RDMA device has a unique set of counters that the
hardware implements.  Having a central set of counters that they must
all adhere to is limiting and causes many useful counters to not be
available.

Therefore we create a dynamic counter registration infrastructure.

The driver must implement a stats structure allocation routine, in
which the driver must place the directory name it wants, a list of
names for all of the counters, an array of u64 counters themselves,
plus a few generic configuration options.

We then implement a core routine to create a sysfs file for each
of the named stats elements, and a core routine to retrieve the
stats when any of the sysfs attribute files are read.

To avoid excessive beating on the stats generation routine in the
drivers, the core code also caches the stats for a short period of
time so that someone attempting to read all of the stats in a
given device's directory will not result in a stats generation
call per file read.

Future work will attempt to standardize just the shared stats
elements, and possibly add a method to get the stats via netlink
in addition to sysfs.

Signed-off-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Mark Bloch <markb@mellanox.com>
Reviewed-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
[ Add caching, make structure names more informative, add i40iw support,
  other significant rewrites from the original patch ]
---
 Documentation/infiniband/sysfs.txt          |  12 +
 drivers/infiniband/core/sysfs.c             | 366 ++++++++++++++++++----------
 drivers/infiniband/hw/cxgb3/iwch_provider.c | 147 +++++++----
 drivers/infiniband/hw/cxgb4/provider.c      |  58 ++++-
 drivers/infiniband/hw/i40iw/i40iw_verbs.c   | 145 ++++++++---
 include/rdma/ib_verbs.h                     | 126 ++++++----
 6 files changed, 590 insertions(+), 264 deletions(-)

(limited to 'include')

diff --git a/Documentation/infiniband/sysfs.txt b/Documentation/infiniband/sysfs.txt
index 3ecf0c3a133f..45bcafe6ff8a 100644
--- a/Documentation/infiniband/sysfs.txt
+++ b/Documentation/infiniband/sysfs.txt
@@ -56,6 +56,18 @@ SYSFS FILES
   ports/1/pkeys/10 contains the value at index 10 in port 1's P_Key
   table.
 
+  There is an optional "hw_counters" subdirectory that may be under either
+  the parent device or the port subdirectories or both.  If present,
+  there are a list of counters provided by the hardware.  They may match
+  some of the counters in the counters directory, but they often include
+  many other counters.  In addition to the various counters, there will
+  be a file named "lifespan" that configures how frequently the core
+  should update the counters when they are being accessed (counters are
+  not updated if they are not being accessed).  The lifespan is in milli-
+  seconds and defaults to 10 unless set to something else by the driver.
+  Users may echo a value between 0 - 10000 to the lifespan file to set
+  the length of time between updates in milliseconds.
+
 MTHCA
 
   The Mellanox HCA driver also creates the files:
diff --git a/drivers/infiniband/core/sysfs.c b/drivers/infiniband/core/sysfs.c
index 14606afbfaa8..5e573bb18660 100644
--- a/drivers/infiniband/core/sysfs.c
+++ b/drivers/infiniband/core/sysfs.c
@@ -56,8 +56,10 @@ struct ib_port {
 	struct gid_attr_group *gid_attr_group;
 	struct attribute_group gid_group;
 	struct attribute_group pkey_group;
-	u8                     port_num;
 	struct attribute_group *pma_table;
+	struct attribute_group *hw_stats_ag;
+	struct rdma_hw_stats   *hw_stats;
+	u8                     port_num;
 };
 
 struct port_attribute {
@@ -80,6 +82,18 @@ struct port_table_attribute {
 	__be16			attr_id;
 };
 
+struct hw_stats_attribute {
+	struct attribute	attr;
+	ssize_t			(*show)(struct kobject *kobj,
+					struct attribute *attr, char *buf);
+	ssize_t			(*store)(struct kobject *kobj,
+					 struct attribute *attr,
+					 const char *buf,
+					 size_t count);
+	int			index;
+	u8			port_num;
+};
+
 static ssize_t port_attr_show(struct kobject *kobj,
 			      struct attribute *attr, char *buf)
 {
@@ -733,6 +747,212 @@ static struct attribute_group *get_counter_table(struct ib_device *dev,
 	return &pma_group;
 }
 
+static int update_hw_stats(struct ib_device *dev, struct rdma_hw_stats *stats,
+			   u8 port_num, int index)
+{
+	int ret;
+
+	if (time_is_after_eq_jiffies(stats->timestamp + stats->lifespan))
+		return 0;
+	ret = dev->get_hw_stats(dev, stats, port_num, index);
+	if (ret < 0)
+		return ret;
+	if (ret == stats->num_counters)
+		stats->timestamp = jiffies;
+
+	return 0;
+}
+
+static ssize_t print_hw_stat(struct rdma_hw_stats *stats, int index, char *buf)
+{
+	return sprintf(buf, "%llu\n", stats->value[index]);
+}
+
+static ssize_t show_hw_stats(struct kobject *kobj, struct attribute *attr,
+			     char *buf)
+{
+	struct ib_device *dev;
+	struct ib_port *port;
+	struct hw_stats_attribute *hsa;
+	struct rdma_hw_stats *stats;
+	int ret;
+
+	hsa = container_of(attr, struct hw_stats_attribute, attr);
+	if (!hsa->port_num) {
+		dev = container_of((struct device *)kobj,
+				   struct ib_device, dev);
+		stats = dev->hw_stats;
+	} else {
+		port = container_of(kobj, struct ib_port, kobj);
+		dev = port->ibdev;
+		stats = port->hw_stats;
+	}
+	ret = update_hw_stats(dev, stats, hsa->port_num, hsa->index);
+	if (ret)
+		return ret;
+	return print_hw_stat(stats, hsa->index, buf);
+}
+
+static ssize_t show_stats_lifespan(struct kobject *kobj,
+				   struct attribute *attr,
+				   char *buf)
+{
+	struct hw_stats_attribute *hsa;
+	int msecs;
+
+	hsa = container_of(attr, struct hw_stats_attribute, attr);
+	if (!hsa->port_num) {
+		struct ib_device *dev = container_of((struct device *)kobj,
+						     struct ib_device, dev);
+		msecs = jiffies_to_msecs(dev->hw_stats->lifespan);
+	} else {
+		struct ib_port *p = container_of(kobj, struct ib_port, kobj);
+		msecs = jiffies_to_msecs(p->hw_stats->lifespan);
+	}
+	return sprintf(buf, "%d\n", msecs);
+}
+
+static ssize_t set_stats_lifespan(struct kobject *kobj,
+				  struct attribute *attr,
+				  const char *buf, size_t count)
+{
+	struct hw_stats_attribute *hsa;
+	int msecs;
+	int jiffies;
+	int ret;
+
+	ret = kstrtoint(buf, 10, &msecs);
+	if (ret)
+		return ret;
+	if (msecs < 0 || msecs > 10000)
+		return -EINVAL;
+	jiffies = msecs_to_jiffies(msecs);
+	hsa = container_of(attr, struct hw_stats_attribute, attr);
+	if (!hsa->port_num) {
+		struct ib_device *dev = container_of((struct device *)kobj,
+						     struct ib_device, dev);
+		dev->hw_stats->lifespan = jiffies;
+	} else {
+		struct ib_port *p = container_of(kobj, struct ib_port, kobj);
+		p->hw_stats->lifespan = jiffies;
+	}
+	return count;
+}
+
+static void free_hsag(struct kobject *kobj, struct attribute_group *attr_group)
+{
+	struct attribute **attr;
+
+	sysfs_remove_group(kobj, attr_group);
+
+	for (attr = attr_group->attrs; *attr; attr++)
+		kfree(*attr);
+	kfree(attr_group);
+}
+
+static struct attribute *alloc_hsa(int index, u8 port_num, const char *name)
+{
+	struct hw_stats_attribute *hsa;
+
+	hsa = kmalloc(sizeof(*hsa), GFP_KERNEL);
+	if (!hsa)
+		return NULL;
+
+	hsa->attr.name = (char *)name;
+	hsa->attr.mode = S_IRUGO;
+	hsa->show = show_hw_stats;
+	hsa->store = NULL;
+	hsa->index = index;
+	hsa->port_num = port_num;
+
+	return &hsa->attr;
+}
+
+static struct attribute *alloc_hsa_lifespan(char *name, u8 port_num)
+{
+	struct hw_stats_attribute *hsa;
+
+	hsa = kmalloc(sizeof(*hsa), GFP_KERNEL);
+	if (!hsa)
+		return NULL;
+
+	hsa->attr.name = name;
+	hsa->attr.mode = S_IWUSR | S_IRUGO;
+	hsa->show = show_stats_lifespan;
+	hsa->store = set_stats_lifespan;
+	hsa->index = 0;
+	hsa->port_num = port_num;
+
+	return &hsa->attr;
+}
+
+static void setup_hw_stats(struct ib_device *device, struct ib_port *port,
+			   u8 port_num)
+{
+	struct attribute_group *hsag = NULL;
+	struct rdma_hw_stats *stats;
+	int i = 0, ret;
+
+	stats = device->alloc_hw_stats(device, port_num);
+
+	if (!stats)
+		return;
+
+	if (!stats->names || stats->num_counters <= 0)
+		goto err;
+
+	hsag = kzalloc(sizeof(*hsag) +
+		       // 1 extra for the lifespan config entry
+		       sizeof(void *) * (stats->num_counters + 1),
+		       GFP_KERNEL);
+	if (!hsag)
+		return;
+
+	ret = device->get_hw_stats(device, stats, port_num,
+				   stats->num_counters);
+	if (ret != stats->num_counters)
+		goto err;
+
+	stats->timestamp = jiffies;
+
+	hsag->name = "hw_counters";
+	hsag->attrs = (void *)hsag + sizeof(*hsag);
+
+	for (i = 0; i < stats->num_counters; i++) {
+		hsag->attrs[i] = alloc_hsa(i, port_num, stats->names[i]);
+		if (!hsag->attrs[i])
+			goto err;
+	}
+
+	/* treat an error here as non-fatal */
+	hsag->attrs[i] = alloc_hsa_lifespan("lifespan", port_num);
+
+	if (port) {
+		struct kobject *kobj = &port->kobj;
+		ret = sysfs_create_group(kobj, hsag);
+		if (ret)
+			goto err;
+		port->hw_stats_ag = hsag;
+		port->hw_stats = stats;
+	} else {
+		struct kobject *kobj = &device->dev.kobj;
+		ret = sysfs_create_group(kobj, hsag);
+		if (ret)
+			goto err;
+		device->hw_stats_ag = hsag;
+		device->hw_stats = stats;
+	}
+
+	return;
+
+err:
+	kfree(stats);
+	for (; i >= 0; i--)
+		kfree(hsag->attrs[i]);
+	kfree(hsag);
+	return;
+}
+
 static int add_port(struct ib_device *device, int port_num,
 		    int (*port_callback)(struct ib_device *,
 					 u8, struct kobject *))
@@ -835,6 +1055,14 @@ static int add_port(struct ib_device *device, int port_num,
 			goto err_remove_pkey;
 	}
 
+	/*
+	 * If port == 0, it means we have only one port and the parent
+	 * device, not this port device, should be the holder of the
+	 * hw_counters
+	 */
+	if (device->alloc_hw_stats && port_num)
+		setup_hw_stats(device, p, port_num);
+
 	list_add_tail(&p->kobj.entry, &device->port_list);
 
 	kobject_uevent(&p->kobj, KOBJ_ADD);
@@ -972,120 +1200,6 @@ static struct device_attribute *ib_class_attributes[] = {
 	&dev_attr_node_desc
 };
 
-/* Show a given an attribute in the statistics group */
-static ssize_t show_protocol_stat(const struct device *device,
-			    struct device_attribute *attr, char *buf,
-			    unsigned offset)
-{
-	struct ib_device *dev = container_of(device, struct ib_device, dev);
-	union rdma_protocol_stats stats;
-	ssize_t ret;
-
-	ret = dev->get_protocol_stats(dev, &stats);
-	if (ret)
-		return ret;
-
-	return sprintf(buf, "%llu\n",
-		       (unsigned long long) ((u64 *) &stats)[offset]);
-}
-
-/* generate a read-only iwarp statistics attribute */
-#define IW_STATS_ENTRY(name)						\
-static ssize_t show_##name(struct device *device,			\
-			   struct device_attribute *attr, char *buf)	\
-{									\
-	return show_protocol_stat(device, attr, buf,			\
-				  offsetof(struct iw_protocol_stats, name) / \
-				  sizeof (u64));			\
-}									\
-static DEVICE_ATTR(name, S_IRUGO, show_##name, NULL)
-
-IW_STATS_ENTRY(ipInReceives);
-IW_STATS_ENTRY(ipInHdrErrors);
-IW_STATS_ENTRY(ipInTooBigErrors);
-IW_STATS_ENTRY(ipInNoRoutes);
-IW_STATS_ENTRY(ipInAddrErrors);
-IW_STATS_ENTRY(ipInUnknownProtos);
-IW_STATS_ENTRY(ipInTruncatedPkts);
-IW_STATS_ENTRY(ipInDiscards);
-IW_STATS_ENTRY(ipInDelivers);
-IW_STATS_ENTRY(ipOutForwDatagrams);
-IW_STATS_ENTRY(ipOutRequests);
-IW_STATS_ENTRY(ipOutDiscards);
-IW_STATS_ENTRY(ipOutNoRoutes);
-IW_STATS_ENTRY(ipReasmTimeout);
-IW_STATS_ENTRY(ipReasmReqds);
-IW_STATS_ENTRY(ipReasmOKs);
-IW_STATS_ENTRY(ipReasmFails);
-IW_STATS_ENTRY(ipFragOKs);
-IW_STATS_ENTRY(ipFragFails);
-IW_STATS_ENTRY(ipFragCreates);
-IW_STATS_ENTRY(ipInMcastPkts);
-IW_STATS_ENTRY(ipOutMcastPkts);
-IW_STATS_ENTRY(ipInBcastPkts);
-IW_STATS_ENTRY(ipOutBcastPkts);
-IW_STATS_ENTRY(tcpRtoAlgorithm);
-IW_STATS_ENTRY(tcpRtoMin);
-IW_STATS_ENTRY(tcpRtoMax);
-IW_STATS_ENTRY(tcpMaxConn);
-IW_STATS_ENTRY(tcpActiveOpens);
-IW_STATS_ENTRY(tcpPassiveOpens);
-IW_STATS_ENTRY(tcpAttemptFails);
-IW_STATS_ENTRY(tcpEstabResets);
-IW_STATS_ENTRY(tcpCurrEstab);
-IW_STATS_ENTRY(tcpInSegs);
-IW_STATS_ENTRY(tcpOutSegs);
-IW_STATS_ENTRY(tcpRetransSegs);
-IW_STATS_ENTRY(tcpInErrs);
-IW_STATS_ENTRY(tcpOutRsts);
-
-static struct attribute *iw_proto_stats_attrs[] = {
-	&dev_attr_ipInReceives.attr,
-	&dev_attr_ipInHdrErrors.attr,
-	&dev_attr_ipInTooBigErrors.attr,
-	&dev_attr_ipInNoRoutes.attr,
-	&dev_attr_ipInAddrErrors.attr,
-	&dev_attr_ipInUnknownProtos.attr,
-	&dev_attr_ipInTruncatedPkts.attr,
-	&dev_attr_ipInDiscards.attr,
-	&dev_attr_ipInDelivers.attr,
-	&dev_attr_ipOutForwDatagrams.attr,
-	&dev_attr_ipOutRequests.attr,
-	&dev_attr_ipOutDiscards.attr,
-	&dev_attr_ipOutNoRoutes.attr,
-	&dev_attr_ipReasmTimeout.attr,
-	&dev_attr_ipReasmReqds.attr,
-	&dev_attr_ipReasmOKs.attr,
-	&dev_attr_ipReasmFails.attr,
-	&dev_attr_ipFragOKs.attr,
-	&dev_attr_ipFragFails.attr,
-	&dev_attr_ipFragCreates.attr,
-	&dev_attr_ipInMcastPkts.attr,
-	&dev_attr_ipOutMcastPkts.attr,
-	&dev_attr_ipInBcastPkts.attr,
-	&dev_attr_ipOutBcastPkts.attr,
-	&dev_attr_tcpRtoAlgorithm.attr,
-	&dev_attr_tcpRtoMin.attr,
-	&dev_attr_tcpRtoMax.attr,
-	&dev_attr_tcpMaxConn.attr,
-	&dev_attr_tcpActiveOpens.attr,
-	&dev_attr_tcpPassiveOpens.attr,
-	&dev_attr_tcpAttemptFails.attr,
-	&dev_attr_tcpEstabResets.attr,
-	&dev_attr_tcpCurrEstab.attr,
-	&dev_attr_tcpInSegs.attr,
-	&dev_attr_tcpOutSegs.attr,
-	&dev_attr_tcpRetransSegs.attr,
-	&dev_attr_tcpInErrs.attr,
-	&dev_attr_tcpOutRsts.attr,
-	NULL
-};
-
-static struct attribute_group iw_stats_group = {
-	.name	= "proto_stats",
-	.attrs	= iw_proto_stats_attrs,
-};
-
 static void free_port_list_attributes(struct ib_device *device)
 {
 	struct kobject *p, *t;
@@ -1093,6 +1207,10 @@ static void free_port_list_attributes(struct ib_device *device)
 	list_for_each_entry_safe(p, t, &device->port_list, entry) {
 		struct ib_port *port = container_of(p, struct ib_port, kobj);
 		list_del(&p->entry);
+		if (port->hw_stats) {
+			kfree(port->hw_stats);
+			free_hsag(&port->kobj, port->hw_stats_ag);
+		}
 		sysfs_remove_group(p, port->pma_table);
 		sysfs_remove_group(p, &port->pkey_group);
 		sysfs_remove_group(p, &port->gid_group);
@@ -1149,11 +1267,8 @@ int ib_device_register_sysfs(struct ib_device *device,
 		}
 	}
 
-	if (device->node_type == RDMA_NODE_RNIC && device->get_protocol_stats) {
-		ret = sysfs_create_group(&class_dev->kobj, &iw_stats_group);
-		if (ret)
-			goto err_put;
-	}
+	if (device->alloc_hw_stats)
+		setup_hw_stats(device, NULL, 0);
 
 	return 0;
 
@@ -1169,15 +1284,18 @@ err:
 
 void ib_device_unregister_sysfs(struct ib_device *device)
 {
-	/* Hold kobject until ib_dealloc_device() */
-	struct kobject *kobj_dev = kobject_get(&device->dev.kobj);
 	int i;
 
-	if (device->node_type == RDMA_NODE_RNIC && device->get_protocol_stats)
-		sysfs_remove_group(kobj_dev, &iw_stats_group);
+	/* Hold kobject until ib_dealloc_device() */
+	kobject_get(&device->dev.kobj);
 
 	free_port_list_attributes(device);
 
+	if (device->hw_stats) {
+		kfree(device->hw_stats);
+		free_hsag(&device->dev.kobj, device->hw_stats_ag);
+	}
+
 	for (i = 0; i < ARRAY_SIZE(ib_class_attributes); ++i)
 		device_remove_file(&device->dev, ib_class_attributes[i]);
 
diff --git a/drivers/infiniband/hw/cxgb3/iwch_provider.c b/drivers/infiniband/hw/cxgb3/iwch_provider.c
index 47cb927a0dd6..bb1a839d4d6d 100644
--- a/drivers/infiniband/hw/cxgb3/iwch_provider.c
+++ b/drivers/infiniband/hw/cxgb3/iwch_provider.c
@@ -1218,59 +1218,119 @@ static ssize_t show_board(struct device *dev, struct device_attribute *attr,
 		       iwch_dev->rdev.rnic_info.pdev->device);
 }
 
-static int iwch_get_mib(struct ib_device *ibdev,
-			union rdma_protocol_stats *stats)
+enum counters {
+	IPINRECEIVES,
+	IPINHDRERRORS,
+	IPINADDRERRORS,
+	IPINUNKNOWNPROTOS,
+	IPINDISCARDS,
+	IPINDELIVERS,
+	IPOUTREQUESTS,
+	IPOUTDISCARDS,
+	IPOUTNOROUTES,
+	IPREASMTIMEOUT,
+	IPREASMREQDS,
+	IPREASMOKS,
+	IPREASMFAILS,
+	TCPACTIVEOPENS,
+	TCPPASSIVEOPENS,
+	TCPATTEMPTFAILS,
+	TCPESTABRESETS,
+	TCPCURRESTAB,
+	TCPINSEGS,
+	TCPOUTSEGS,
+	TCPRETRANSSEGS,
+	TCPINERRS,
+	TCPOUTRSTS,
+	TCPRTOMIN,
+	TCPRTOMAX,
+	NR_COUNTERS
+};
+
+static const char * const names[] = {
+	[IPINRECEIVES] = "ipInReceives",
+	[IPINHDRERRORS] = "ipInHdrErrors",
+	[IPINADDRERRORS] = "ipInAddrErrors",
+	[IPINUNKNOWNPROTOS] = "ipInUnknownProtos",
+	[IPINDISCARDS] = "ipInDiscards",
+	[IPINDELIVERS] = "ipInDelivers",
+	[IPOUTREQUESTS] = "ipOutRequests",
+	[IPOUTDISCARDS] = "ipOutDiscards",
+	[IPOUTNOROUTES] = "ipOutNoRoutes",
+	[IPREASMTIMEOUT] = "ipReasmTimeout",
+	[IPREASMREQDS] = "ipReasmReqds",
+	[IPREASMOKS] = "ipReasmOKs",
+	[IPREASMFAILS] = "ipReasmFails",
+	[TCPACTIVEOPENS] = "tcpActiveOpens",
+	[TCPPASSIVEOPENS] = "tcpPassiveOpens",
+	[TCPATTEMPTFAILS] = "tcpAttemptFails",
+	[TCPESTABRESETS] = "tcpEstabResets",
+	[TCPCURRESTAB] = "tcpCurrEstab",
+	[TCPINSEGS] = "tcpInSegs",
+	[TCPOUTSEGS] = "tcpOutSegs",
+	[TCPRETRANSSEGS] = "tcpRetransSegs",
+	[TCPINERRS] = "tcpInErrs",
+	[TCPOUTRSTS] = "tcpOutRsts",
+	[TCPRTOMIN] = "tcpRtoMin",
+	[TCPRTOMAX] = "tcpRtoMax",
+};
+
+static struct rdma_hw_stats *iwch_alloc_stats(struct ib_device *ibdev,
+					      u8 port_num)
+{
+	BUILD_BUG_ON(ARRAY_SIZE(names) != NR_COUNTERS);
+
+	/* Our driver only supports device level stats */
+	if (port_num != 0)
+		return NULL;
+
+	return rdma_alloc_hw_stats_struct(names, NR_COUNTERS,
+					  RDMA_HW_STATS_DEFAULT_LIFESPAN);
+}
+
+static int iwch_get_mib(struct ib_device *ibdev, struct rdma_hw_stats *stats,
+			u8 port, int index)
 {
 	struct iwch_dev *dev;
 	struct tp_mib_stats m;
 	int ret;
 
+	if (port != 0 || !stats)
+		return -ENOSYS;
+
 	PDBG("%s ibdev %p\n", __func__, ibdev);
 	dev = to_iwch_dev(ibdev);
 	ret = dev->rdev.t3cdev_p->ctl(dev->rdev.t3cdev_p, RDMA_GET_MIB, &m);
 	if (ret)
 		return -ENOSYS;
 
-	memset(stats, 0, sizeof *stats);
-	stats->iw.ipInReceives = ((u64) m.ipInReceive_hi << 32) +
-				m.ipInReceive_lo;
-	stats->iw.ipInHdrErrors = ((u64) m.ipInHdrErrors_hi << 32) +
-				  m.ipInHdrErrors_lo;
-	stats->iw.ipInAddrErrors = ((u64) m.ipInAddrErrors_hi << 32) +
-				   m.ipInAddrErrors_lo;
-	stats->iw.ipInUnknownProtos = ((u64) m.ipInUnknownProtos_hi << 32) +
-				      m.ipInUnknownProtos_lo;
-	stats->iw.ipInDiscards = ((u64) m.ipInDiscards_hi << 32) +
-				 m.ipInDiscards_lo;
-	stats->iw.ipInDelivers = ((u64) m.ipInDelivers_hi << 32) +
-				 m.ipInDelivers_lo;
-	stats->iw.ipOutRequests = ((u64) m.ipOutRequests_hi << 32) +
-				  m.ipOutRequests_lo;
-	stats->iw.ipOutDiscards = ((u64) m.ipOutDiscards_hi << 32) +
-				  m.ipOutDiscards_lo;
-	stats->iw.ipOutNoRoutes = ((u64) m.ipOutNoRoutes_hi << 32) +
-				  m.ipOutNoRoutes_lo;
-	stats->iw.ipReasmTimeout = (u64) m.ipReasmTimeout;
-	stats->iw.ipReasmReqds = (u64) m.ipReasmReqds;
-	stats->iw.ipReasmOKs = (u64) m.ipReasmOKs;
-	stats->iw.ipReasmFails = (u64) m.ipReasmFails;
-	stats->iw.tcpActiveOpens = (u64) m.tcpActiveOpens;
-	stats->iw.tcpPassiveOpens = (u64) m.tcpPassiveOpens;
-	stats->iw.tcpAttemptFails = (u64) m.tcpAttemptFails;
-	stats->iw.tcpEstabResets = (u64) m.tcpEstabResets;
-	stats->iw.tcpOutRsts = (u64) m.tcpOutRsts;
-	stats->iw.tcpCurrEstab = (u64) m.tcpCurrEstab;
-	stats->iw.tcpInSegs = ((u64) m.tcpInSegs_hi << 32) +
-			      m.tcpInSegs_lo;
-	stats->iw.tcpOutSegs = ((u64) m.tcpOutSegs_hi << 32) +
-			       m.tcpOutSegs_lo;
-	stats->iw.tcpRetransSegs = ((u64) m.tcpRetransSeg_hi << 32) +
-				  m.tcpRetransSeg_lo;
-	stats->iw.tcpInErrs = ((u64) m.tcpInErrs_hi << 32) +
-			      m.tcpInErrs_lo;
-	stats->iw.tcpRtoMin = (u64) m.tcpRtoMin;
-	stats->iw.tcpRtoMax = (u64) m.tcpRtoMax;
-	return 0;
+	stats->value[IPINRECEIVES] = ((u64)m.ipInReceive_hi << 32) +	m.ipInReceive_lo;
+	stats->value[IPINHDRERRORS] = ((u64)m.ipInHdrErrors_hi << 32) + m.ipInHdrErrors_lo;
+	stats->value[IPINADDRERRORS] = ((u64)m.ipInAddrErrors_hi << 32) + m.ipInAddrErrors_lo;
+	stats->value[IPINUNKNOWNPROTOS] = ((u64)m.ipInUnknownProtos_hi << 32) + m.ipInUnknownProtos_lo;
+	stats->value[IPINDISCARDS] = ((u64)m.ipInDiscards_hi << 32) + m.ipInDiscards_lo;
+	stats->value[IPINDELIVERS] = ((u64)m.ipInDelivers_hi << 32) + m.ipInDelivers_lo;
+	stats->value[IPOUTREQUESTS] = ((u64)m.ipOutRequests_hi << 32) + m.ipOutRequests_lo;
+	stats->value[IPOUTDISCARDS] = ((u64)m.ipOutDiscards_hi << 32) + m.ipOutDiscards_lo;
+	stats->value[IPOUTNOROUTES] = ((u64)m.ipOutNoRoutes_hi << 32) + m.ipOutNoRoutes_lo;
+	stats->value[IPREASMTIMEOUT] = 	m.ipReasmTimeout;
+	stats->value[IPREASMREQDS] = m.ipReasmReqds;
+	stats->value[IPREASMOKS] = m.ipReasmOKs;
+	stats->value[IPREASMFAILS] = m.ipReasmFails;
+	stats->value[TCPACTIVEOPENS] =	m.tcpActiveOpens;
+	stats->value[TCPPASSIVEOPENS] =	m.tcpPassiveOpens;
+	stats->value[TCPATTEMPTFAILS] = m.tcpAttemptFails;
+	stats->value[TCPESTABRESETS] = m.tcpEstabResets;
+	stats->value[TCPCURRESTAB] = m.tcpOutRsts;
+	stats->value[TCPINSEGS] = m.tcpCurrEstab;
+	stats->value[TCPOUTSEGS] = ((u64)m.tcpInSegs_hi << 32) + m.tcpInSegs_lo;
+	stats->value[TCPRETRANSSEGS] = ((u64)m.tcpOutSegs_hi << 32) + m.tcpOutSegs_lo;
+	stats->value[TCPINERRS] = ((u64)m.tcpRetransSeg_hi << 32) + m.tcpRetransSeg_lo,
+	stats->value[TCPOUTRSTS] = ((u64)m.tcpInErrs_hi << 32) + m.tcpInErrs_lo;
+	stats->value[TCPRTOMIN] = m.tcpRtoMin;
+	stats->value[TCPRTOMAX] = m.tcpRtoMax;
+
+	return stats->num_counters;
 }
 
 static DEVICE_ATTR(hw_rev, S_IRUGO, show_rev, NULL);
@@ -1373,7 +1433,8 @@ int iwch_register_device(struct iwch_dev *dev)
 	dev->ibdev.req_notify_cq = iwch_arm_cq;
 	dev->ibdev.post_send = iwch_post_send;
 	dev->ibdev.post_recv = iwch_post_receive;
-	dev->ibdev.get_protocol_stats = iwch_get_mib;
+	dev->ibdev.alloc_hw_stats = iwch_alloc_stats;
+	dev->ibdev.get_hw_stats = iwch_get_mib;
 	dev->ibdev.uverbs_abi_ver = IWCH_UVERBS_ABI_VERSION;
 	dev->ibdev.get_port_immutable = iwch_port_immutable;
 
diff --git a/drivers/infiniband/hw/cxgb4/provider.c b/drivers/infiniband/hw/cxgb4/provider.c
index 7574f394fdac..dd8a86b726d2 100644
--- a/drivers/infiniband/hw/cxgb4/provider.c
+++ b/drivers/infiniband/hw/cxgb4/provider.c
@@ -446,20 +446,59 @@ static ssize_t show_board(struct device *dev, struct device_attribute *attr,
 		       c4iw_dev->rdev.lldi.pdev->device);
 }
 
+enum counters {
+	IP4INSEGS,
+	IP4OUTSEGS,
+	IP4RETRANSSEGS,
+	IP4OUTRSTS,
+	IP6INSEGS,
+	IP6OUTSEGS,
+	IP6RETRANSSEGS,
+	IP6OUTRSTS,
+	NR_COUNTERS
+};
+
+static const char * const names[] = {
+	[IP4INSEGS] = "ip4InSegs",
+	[IP4OUTSEGS] = "ip4OutSegs",
+	[IP4RETRANSSEGS] = "ip4RetransSegs",
+	[IP4OUTRSTS] = "ip4OutRsts",
+	[IP6INSEGS] = "ip6InSegs",
+	[IP6OUTSEGS] = "ip6OutSegs",
+	[IP6RETRANSSEGS] = "ip6RetransSegs",
+	[IP6OUTRSTS] = "ip6OutRsts"
+};
+
+static struct rdma_hw_stats *c4iw_alloc_stats(struct ib_device *ibdev,
+					      u8 port_num)
+{
+	BUILD_BUG_ON(ARRAY_SIZE(names) != NR_COUNTERS);
+
+	if (port_num != 0)
+		return NULL;
+
+	return rdma_alloc_hw_stats_struct(names, NR_COUNTERS,
+					  RDMA_HW_STATS_DEFAULT_LIFESPAN);
+}
+
 static int c4iw_get_mib(struct ib_device *ibdev,
-			union rdma_protocol_stats *stats)
+			struct rdma_hw_stats *stats,
+			u8 port, int index)
 {
 	struct tp_tcp_stats v4, v6;
 	struct c4iw_dev *c4iw_dev = to_c4iw_dev(ibdev);
 
 	cxgb4_get_tcp_stats(c4iw_dev->rdev.lldi.pdev, &v4, &v6);
-	memset(stats, 0, sizeof *stats);
-	stats->iw.tcpInSegs = v4.tcp_in_segs + v6.tcp_in_segs;
-	stats->iw.tcpOutSegs = v4.tcp_out_segs + v6.tcp_out_segs;
-	stats->iw.tcpRetransSegs = v4.tcp_retrans_segs + v6.tcp_retrans_segs;
-	stats->iw.tcpOutRsts = v4.tcp_out_rsts + v6.tcp_out_rsts;
-
-	return 0;
+	stats->value[IP4INSEGS] = v4.tcp_in_segs;
+	stats->value[IP4OUTSEGS] = v4.tcp_out_segs;
+	stats->value[IP4RETRANSSEGS] = v4.tcp_retrans_segs;
+	stats->value[IP4OUTRSTS] = v4.tcp_out_rsts;
+	stats->value[IP6INSEGS] = v6.tcp_in_segs;
+	stats->value[IP6OUTSEGS] = v6.tcp_out_segs;
+	stats->value[IP6RETRANSSEGS] = v6.tcp_retrans_segs;
+	stats->value[IP6OUTRSTS] = v6.tcp_out_rsts;
+
+	return stats->num_counters;
 }
 
 static DEVICE_ATTR(hw_rev, S_IRUGO, show_rev, NULL);
@@ -562,7 +601,8 @@ int c4iw_register_device(struct c4iw_dev *dev)
 	dev->ibdev.req_notify_cq = c4iw_arm_cq;
 	dev->ibdev.post_send = c4iw_post_send;
 	dev->ibdev.post_recv = c4iw_post_receive;
-	dev->ibdev.get_protocol_stats = c4iw_get_mib;
+	dev->ibdev.alloc_hw_stats = c4iw_alloc_stats;
+	dev->ibdev.get_hw_stats = c4iw_get_mib;
 	dev->ibdev.uverbs_abi_ver = C4IW_UVERBS_ABI_VERSION;
 	dev->ibdev.get_port_immutable = c4iw_port_immutable;
 	dev->ibdev.drain_sq = c4iw_drain_sq;
diff --git a/drivers/infiniband/hw/i40iw/i40iw_verbs.c b/drivers/infiniband/hw/i40iw/i40iw_verbs.c
index 4a740f7a0519..02a735b64208 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_verbs.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_verbs.c
@@ -2361,58 +2361,130 @@ static int i40iw_port_immutable(struct ib_device *ibdev, u8 port_num,
 	return 0;
 }
 
+static const char * const i40iw_hw_stat_names[] = {
+	// 32bit names
+	[I40IW_HW_STAT_INDEX_IP4RXDISCARD] = "ip4InDiscards",
+	[I40IW_HW_STAT_INDEX_IP4RXTRUNC] = "ip4InTruncatedPkts",
+	[I40IW_HW_STAT_INDEX_IP4TXNOROUTE] = "ip4OutNoRoutes",
+	[I40IW_HW_STAT_INDEX_IP6RXDISCARD] = "ip6InDiscards",
+	[I40IW_HW_STAT_INDEX_IP6RXTRUNC] = "ip6InTruncatedPkts",
+	[I40IW_HW_STAT_INDEX_IP6TXNOROUTE] = "ip6OutNoRoutes",
+	[I40IW_HW_STAT_INDEX_TCPRTXSEG] = "tcpRetransSegs",
+	[I40IW_HW_STAT_INDEX_TCPRXOPTERR] = "tcpInOptErrors",
+	[I40IW_HW_STAT_INDEX_TCPRXPROTOERR] = "tcpInProtoErrors",
+	// 64bit names
+	[I40IW_HW_STAT_INDEX_IP4RXOCTS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"ip4InOctets",
+	[I40IW_HW_STAT_INDEX_IP4RXPKTS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"ip4InPkts",
+	[I40IW_HW_STAT_INDEX_IP4RXFRAGS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"ip4InReasmRqd",
+	[I40IW_HW_STAT_INDEX_IP4RXMCPKTS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"ip4InMcastPkts",
+	[I40IW_HW_STAT_INDEX_IP4TXOCTS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"ip4OutOctets",
+	[I40IW_HW_STAT_INDEX_IP4TXPKTS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"ip4OutPkts",
+	[I40IW_HW_STAT_INDEX_IP4TXFRAGS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"ip4OutSegRqd",
+	[I40IW_HW_STAT_INDEX_IP4TXMCPKTS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"ip4OutMcastPkts",
+	[I40IW_HW_STAT_INDEX_IP6RXOCTS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"ip6InOctets",
+	[I40IW_HW_STAT_INDEX_IP6RXPKTS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"ip6InPkts",
+	[I40IW_HW_STAT_INDEX_IP6RXFRAGS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"ip6InReasmRqd",
+	[I40IW_HW_STAT_INDEX_IP6RXMCPKTS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"ip6InMcastPkts",
+	[I40IW_HW_STAT_INDEX_IP6TXOCTS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"ip6OutOctets",
+	[I40IW_HW_STAT_INDEX_IP6TXPKTS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"ip6OutPkts",
+	[I40IW_HW_STAT_INDEX_IP6TXFRAGS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"ip6OutSegRqd",
+	[I40IW_HW_STAT_INDEX_IP6TXMCPKTS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"ip6OutMcastPkts",
+	[I40IW_HW_STAT_INDEX_TCPRXSEGS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"tcpInSegs",
+	[I40IW_HW_STAT_INDEX_TCPTXSEG + I40IW_HW_STAT_INDEX_MAX_32] =
+		"tcpOutSegs",
+	[I40IW_HW_STAT_INDEX_RDMARXRDS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"iwInRdmaReads",
+	[I40IW_HW_STAT_INDEX_RDMARXSNDS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"iwInRdmaSends",
+	[I40IW_HW_STAT_INDEX_RDMARXWRS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"iwInRdmaWrites",
+	[I40IW_HW_STAT_INDEX_RDMATXRDS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"iwOutRdmaReads",
+	[I40IW_HW_STAT_INDEX_RDMATXSNDS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"iwOutRdmaSends",
+	[I40IW_HW_STAT_INDEX_RDMATXWRS + I40IW_HW_STAT_INDEX_MAX_32] =
+		"iwOutRdmaWrites",
+	[I40IW_HW_STAT_INDEX_RDMAVBND + I40IW_HW_STAT_INDEX_MAX_32] =
+		"iwRdmaBnd",
+	[I40IW_HW_STAT_INDEX_RDMAVINV + I40IW_HW_STAT_INDEX_MAX_32] =
+		"iwRdmaInv"
+};
+
 /**
- * i40iw_get_protocol_stats - Populates the rdma_stats structure
- * @ibdev: ib dev struct
- * @stats: iw protocol stats struct
+ * i40iw_alloc_hw_stats - Allocate a hw stats structure
+ * @ibdev: device pointer from stack
+ * @port_num: port number
  */
-static int i40iw_get_protocol_stats(struct ib_device *ibdev,
-				    union rdma_protocol_stats *stats)
+static struct rdma_hw_stats *i40iw_alloc_hw_stats(struct ib_device *ibdev,
+						  u8 port_num)
+{
+	struct i40iw_device *iwdev = to_iwdev(ibdev);
+	struct i40iw_sc_dev *dev = &iwdev->sc_dev;
+	int num_counters = I40IW_HW_STAT_INDEX_MAX_32 +
+		I40IW_HW_STAT_INDEX_MAX_64;
+	unsigned long lifespan = RDMA_HW_STATS_DEFAULT_LIFESPAN;
+
+	BUILD_BUG_ON(ARRAY_SIZE(i40iw_hw_stat_names) !=
+		     (I40IW_HW_STAT_INDEX_MAX_32 +
+		      I40IW_HW_STAT_INDEX_MAX_64));
+
+	/*
+	 * PFs get the default update lifespan, but VFs only update once
+	 * per second
+	 */
+	if (!dev->is_pf)
+		lifespan = 1000;
+	return rdma_alloc_hw_stats_struct(i40iw_hw_stat_names, num_counters,
+					  lifespan);
+}
+
+/**
+ * i40iw_get_hw_stats - Populates the rdma_hw_stats structure
+ * @ibdev: device pointer from stack
+ * @stats: stats pointer from stack
+ * @port_num: port number
+ * @index: which hw counter the stack is requesting we update
+ */
+static int i40iw_get_hw_stats(struct ib_device *ibdev,
+			      struct rdma_hw_stats *stats,
+			      u8 port_num, int index)
 {
 	struct i40iw_device *iwdev = to_iwdev(ibdev);
 	struct i40iw_sc_dev *dev = &iwdev->sc_dev;
 	struct i40iw_dev_pestat *devstat = &dev->dev_pestat;
 	struct i40iw_dev_hw_stats *hw_stats = &devstat->hw_stats;
-	struct timespec curr_time;
-	static struct timespec last_rd_time = {0, 0};
 	unsigned long flags;
 
-	curr_time = current_kernel_time();
-	memset(stats, 0, sizeof(*stats));
-
 	if (dev->is_pf) {
 		spin_lock_irqsave(&devstat->stats_lock, flags);
 		devstat->ops.iw_hw_stat_read_all(devstat,
 			&devstat->hw_stats);
 		spin_unlock_irqrestore(&devstat->stats_lock, flags);
 	} else {
-		if (((u64)curr_time.tv_sec - (u64)last_rd_time.tv_sec) > 1)
-			if (i40iw_vchnl_vf_get_pe_stats(dev, &devstat->hw_stats))
-				return -ENOSYS;
+		if (i40iw_vchnl_vf_get_pe_stats(dev, &devstat->hw_stats))
+			return -ENOSYS;
 	}
 
-	stats->iw.ipInReceives = hw_stats->stat_value_64[I40IW_HW_STAT_INDEX_IP4RXPKTS] +
-				 hw_stats->stat_value_64[I40IW_HW_STAT_INDEX_IP6RXPKTS];
-	stats->iw.ipInTruncatedPkts = hw_stats->stat_value_32[I40IW_HW_STAT_INDEX_IP4RXTRUNC] +
-				      hw_stats->stat_value_32[I40IW_HW_STAT_INDEX_IP6RXTRUNC];
-	stats->iw.ipInDiscards = hw_stats->stat_value_32[I40IW_HW_STAT_INDEX_IP4RXDISCARD] +
-				 hw_stats->stat_value_32[I40IW_HW_STAT_INDEX_IP6RXDISCARD];
-	stats->iw.ipOutNoRoutes = hw_stats->stat_value_32[I40IW_HW_STAT_INDEX_IP4TXNOROUTE] +
-				  hw_stats->stat_value_32[I40IW_HW_STAT_INDEX_IP6TXNOROUTE];
-	stats->iw.ipReasmReqds = hw_stats->stat_value_64[I40IW_HW_STAT_INDEX_IP4RXFRAGS] +
-				 hw_stats->stat_value_64[I40IW_HW_STAT_INDEX_IP6RXFRAGS];
-	stats->iw.ipFragCreates = hw_stats->stat_value_64[I40IW_HW_STAT_INDEX_IP4TXFRAGS] +
-				  hw_stats->stat_value_64[I40IW_HW_STAT_INDEX_IP6TXFRAGS];
-	stats->iw.ipInMcastPkts = hw_stats->stat_value_64[I40IW_HW_STAT_INDEX_IP4RXMCPKTS] +
-				  hw_stats->stat_value_64[I40IW_HW_STAT_INDEX_IP6RXMCPKTS];
-	stats->iw.ipOutMcastPkts = hw_stats->stat_value_64[I40IW_HW_STAT_INDEX_IP4TXMCPKTS] +
-				   hw_stats->stat_value_64[I40IW_HW_STAT_INDEX_IP6TXMCPKTS];
-	stats->iw.tcpOutSegs = hw_stats->stat_value_64[I40IW_HW_STAT_INDEX_TCPTXSEG];
-	stats->iw.tcpInSegs = hw_stats->stat_value_64[I40IW_HW_STAT_INDEX_TCPRXSEGS];
-	stats->iw.tcpRetransSegs = hw_stats->stat_value_32[I40IW_HW_STAT_INDEX_TCPRTXSEG];
-
-	last_rd_time = curr_time;
-	return 0;
+	memcpy(&stats->value[0], &hw_stats, sizeof(*hw_stats));
+
+	return stats->num_counters;
 }
 
 /**
@@ -2551,7 +2623,8 @@ static struct i40iw_ib_device *i40iw_init_rdma_device(struct i40iw_device *iwdev
 	iwibdev->ibdev.get_dma_mr = i40iw_get_dma_mr;
 	iwibdev->ibdev.reg_user_mr = i40iw_reg_user_mr;
 	iwibdev->ibdev.dereg_mr = i40iw_dereg_mr;
-	iwibdev->ibdev.get_protocol_stats = i40iw_get_protocol_stats;
+	iwibdev->ibdev.alloc_hw_stats = i40iw_alloc_hw_stats;
+	iwibdev->ibdev.get_hw_stats = i40iw_get_hw_stats;
 	iwibdev->ibdev.query_device = i40iw_query_device;
 	iwibdev->ibdev.create_ah = i40iw_create_ah;
 	iwibdev->ibdev.destroy_ah = i40iw_destroy_ah;
diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index fc0320c004a3..432bed510369 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -403,56 +403,55 @@ enum ib_port_speed {
 	IB_SPEED_EDR	= 32
 };
 
-struct ib_protocol_stats {
-	/* TBD... */
-};
-
-struct iw_protocol_stats {
-	u64	ipInReceives;
-	u64	ipInHdrErrors;
-	u64	ipInTooBigErrors;
-	u64	ipInNoRoutes;
-	u64	ipInAddrErrors;
-	u64	ipInUnknownProtos;
-	u64	ipInTruncatedPkts;
-	u64	ipInDiscards;
-	u64	ipInDelivers;
-	u64	ipOutForwDatagrams;
-	u64	ipOutRequests;
-	u64	ipOutDiscards;
-	u64	ipOutNoRoutes;
-	u64	ipReasmTimeout;
-	u64	ipReasmReqds;
-	u64	ipReasmOKs;
-	u64	ipReasmFails;
-	u64	ipFragOKs;
-	u64	ipFragFails;
-	u64	ipFragCreates;
-	u64	ipInMcastPkts;
-	u64	ipOutMcastPkts;
-	u64	ipInBcastPkts;
-	u64	ipOutBcastPkts;
-
-	u64	tcpRtoAlgorithm;
-	u64	tcpRtoMin;
-	u64	tcpRtoMax;
-	u64	tcpMaxConn;
-	u64	tcpActiveOpens;
-	u64	tcpPassiveOpens;
-	u64	tcpAttemptFails;
-	u64	tcpEstabResets;
-	u64	tcpCurrEstab;
-	u64	tcpInSegs;
-	u64	tcpOutSegs;
-	u64	tcpRetransSegs;
-	u64	tcpInErrs;
-	u64	tcpOutRsts;
-};
-
-union rdma_protocol_stats {
-	struct ib_protocol_stats	ib;
-	struct iw_protocol_stats	iw;
-};
+/**
+ * struct rdma_hw_stats
+ * @timestamp - Used by the core code to track when the last update was
+ * @lifespan - Used by the core code to determine how old the counters
+ *   should be before being updated again.  Stored in jiffies, defaults
+ *   to 10 milliseconds, drivers can override the default be specifying
+ *   their own value during their allocation routine.
+ * @name - Array of pointers to static names used for the counters in
+ *   directory.
+ * @num_counters - How many hardware counters there are.  If name is
+ *   shorter than this number, a kernel oops will result.  Driver authors
+ *   are encouraged to leave BUILD_BUG_ON(ARRAY_SIZE(@name) < num_counters)
+ *   in their code to prevent this.
+ * @value - Array of u64 counters that are accessed by the sysfs code and
+ *   filled in by the drivers get_stats routine
+ */
+struct rdma_hw_stats {
+	unsigned long	timestamp;
+	unsigned long	lifespan;
+	const char * const *names;
+	int		num_counters;
+	u64		value[];
+};
+
+#define RDMA_HW_STATS_DEFAULT_LIFESPAN 10
+/**
+ * rdma_alloc_hw_stats_struct - Helper function to allocate dynamic struct
+ *   for drivers.
+ * @names - Array of static const char *
+ * @num_counters - How many elements in array
+ * @lifespan - How many milliseconds between updates
+ */
+static inline struct rdma_hw_stats *rdma_alloc_hw_stats_struct(
+		const char * const *names, int num_counters,
+		unsigned long lifespan)
+{
+	struct rdma_hw_stats *stats;
+
+	stats = kzalloc(sizeof(*stats) + num_counters * sizeof(u64),
+			GFP_KERNEL);
+	if (!stats)
+		return NULL;
+	stats->names = names;
+	stats->num_counters = num_counters;
+	stats->lifespan = msecs_to_jiffies(lifespan);
+
+	return stats;
+}
+
 
 /* Define bits for the various functionality this port needs to be supported by
  * the core.
@@ -1707,8 +1706,29 @@ struct ib_device {
 
 	struct iw_cm_verbs	     *iwcm;
 
-	int		           (*get_protocol_stats)(struct ib_device *device,
-							 union rdma_protocol_stats *stats);
+	/**
+	 * alloc_hw_stats - Allocate a struct rdma_hw_stats and fill in the
+	 *   driver initialized data.  The struct is kfree()'ed by the sysfs
+	 *   core when the device is removed.  A lifespan of -1 in the return
+	 *   struct tells the core to set a default lifespan.
+	 */
+	struct rdma_hw_stats      *(*alloc_hw_stats)(struct ib_device *device,
+						     u8 port_num);
+	/**
+	 * get_hw_stats - Fill in the counter value(s) in the stats struct.
+	 * @index - The index in the value array we wish to have updated, or
+	 *   num_counters if we want all stats updated
+	 * Return codes -
+	 *   < 0 - Error, no counters updated
+	 *   index - Updated the single counter pointed to by index
+	 *   num_counters - Updated all counters (will reset the timestamp
+	 *     and prevent further calls for lifespan milliseconds)
+	 * Drivers are allowed to update all counters in leiu of just the
+	 *   one given in index at their option
+	 */
+	int		           (*get_hw_stats)(struct ib_device *device,
+						   struct rdma_hw_stats *stats,
+						   u8 port, int index);
 	int		           (*query_device)(struct ib_device *device,
 						   struct ib_device_attr *device_attr,
 						   struct ib_udata *udata);
@@ -1926,6 +1946,8 @@ struct ib_device {
 	u8                           node_type;
 	u8                           phys_port_cnt;
 	struct ib_device_attr        attrs;
+	struct attribute_group	     *hw_stats_ag;
+	struct rdma_hw_stats         *hw_stats;
 
 	/**
 	 * The following mandatory functions are used only at device
-- 
cgit v1.2.3