Merge tag 'for-linus-2' of git://git.kernel.org/pub/scm/linux/kernel/git/dledford/rdma

Pull second round of rdma updates from Doug Ledford:
 "This can be split out into just two categories:

   - fixes to the RDMA R/W API in regards to SG list length limits
     (about 5 patches)

   - fixes/features for the Intel hfi1 driver (everything else)

  The hfi1 driver is still being brought to full feature support by
  Intel, and they have a lot of people working on it, so that amounts to
  almost the entirety of this pull request"

* tag 'for-linus-2' of git://git.kernel.org/pub/scm/linux/kernel/git/dledford/rdma: (84 commits)
  IB/hfi1: Add cache evict LRU list
  IB/hfi1: Fix memory leak during unexpected shutdown
  IB/hfi1: Remove unneeded mm argument in remove function
  IB/hfi1: Consistently call ops->remove outside spinlock
  IB/hfi1: Use evict mmu rb operation
  IB/hfi1: Add evict operation to the mmu rb handler
  IB/hfi1: Fix TID caching actions
  IB/hfi1: Make the cache handler own its rb tree root
  IB/hfi1: Make use of mm consistent
  IB/hfi1: Fix user SDMA racy user request claim
  IB/hfi1: Fix error condition that needs to clean up
  IB/hfi1: Release node on insert failure
  IB/hfi1: Validate SDMA user iovector count
  IB/hfi1: Validate SDMA user request index
  IB/hfi1: Use the same capability state for all shared contexts
  IB/hfi1: Prevent null pointer dereference
  IB/hfi1: Rename TID mmu_rb_* functions
  IB/hfi1: Remove unneeded empty check in hfi1_mmu_rb_unregister()
  IB/hfi1: Restructure hfi1_file_open
  IB/hfi1: Make iovec loop index easy to understand
  ...

6 files changed, 97 insertions, 27 deletions

diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index 94a0bc5b5bdd..8e90dd28bb75 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -1490,6 +1490,10 @@ struct ib_rwq_ind_table_init_attr {
 	struct ib_wq	**ind_tbl;
 };
 
+/*
+ * @max_write_sge: Maximum SGE elements per RDMA WRITE request.
+ * @max_read_sge:  Maximum SGE elements per RDMA READ request.
+ */
 struct ib_qp {
 	struct ib_device       *device;
 	struct ib_pd	       *pd;
@@ -1511,6 +1515,8 @@ struct ib_qp {
 	void                  (*event_handler)(struct ib_event *, void *);
 	void		       *qp_context;
 	u32			qp_num;
+	u32			max_write_sge;
+	u32			max_read_sge;
 	enum ib_qp_type		qp_type;
 	struct ib_rwq_ind_table *rwq_ind_tbl;
 };
diff --git a/include/rdma/opa_port_info.h b/include/rdma/opa_port_info.h
index 2b95c2c336eb..9303e0e4f508 100644
--- a/include/rdma/opa_port_info.h
+++ b/include/rdma/opa_port_info.h
@@ -33,11 +33,6 @@
 #if !defined(OPA_PORT_INFO_H)
 #define OPA_PORT_INFO_H
 
-/* Temporary until HFI driver is updated */
-#ifndef USE_PI_LED_ENABLE
-#define USE_PI_LED_ENABLE 0
-#endif
-
 #define OPA_PORT_LINK_MODE_NOP	0		/* No change */
 #define OPA_PORT_LINK_MODE_OPA	4		/* Port mode is OPA */
 
@@ -274,23 +269,12 @@ enum port_info_field_masks {
 	OPA_PI_MASK_MTU_CAP                       = 0x0F,
 };
 
-#if USE_PI_LED_ENABLE
 struct opa_port_states {
 	u8     reserved;
 	u8     ledenable_offlinereason;   /* 1 res, 1 bit, 6 bits */
 	u8     reserved2;
 	u8     portphysstate_portstate;   /* 4 bits, 4 bits */
 };
-#define PI_LED_ENABLE_SUP 1
-#else
-struct opa_port_states {
-	u8     reserved;
-	u8     offline_reason;            /* 2 res, 6 bits */
-	u8     reserved2;
-	u8     portphysstate_portstate;   /* 4 bits, 4 bits */
-};
-#define PI_LED_ENABLE_SUP 0
-#endif
 
 struct opa_port_state_info {
 	struct opa_port_states port_states;
diff --git a/include/rdma/rdma_vt.h b/include/rdma/rdma_vt.h
index 9c9a27d42aaa..e31502107a58 100644
--- a/include/rdma/rdma_vt.h
+++ b/include/rdma/rdma_vt.h
@@ -158,6 +158,7 @@ struct rvt_driver_params {
 	u32 max_mad_size;
 	u8 qos_shift;
 	u8 max_rdma_atomic;
+	u8 reserved_operations;
 };
 
 /* Protection domain */
@@ -351,6 +352,9 @@ struct rvt_dev_info {
 	/* Driver specific properties */
 	struct rvt_driver_params dparms;
 
+	/* post send table */
+	const struct rvt_operation_params *post_parms;
+
 	struct rvt_mregion __rcu *dma_mr;
 	struct rvt_lkey_table lkey_table;
 
@@ -484,6 +488,9 @@ void rvt_unregister_device(struct rvt_dev_info *rvd);
 int rvt_check_ah(struct ib_device *ibdev, struct ib_ah_attr *ah_attr);
 int rvt_init_port(struct rvt_dev_info *rdi, struct rvt_ibport *port,
 		  int port_index, u16 *pkey_table);
+int rvt_fast_reg_mr(struct rvt_qp *qp, struct ib_mr *ibmr, u32 key,
+		    int access);
+int rvt_invalidate_rkey(struct rvt_qp *qp, u32 rkey);
 int rvt_rkey_ok(struct rvt_qp *qp, struct rvt_sge *sge,
 		u32 len, u64 vaddr, u32 rkey, int acc);
 int rvt_lkey_ok(struct rvt_lkey_table *rkt, struct rvt_pd *pd,
diff --git a/include/rdma/rdmavt_mr.h b/include/rdma/rdmavt_mr.h
index 5edffdca8c53..6b3c6c8b6b77 100644
--- a/include/rdma/rdmavt_mr.h
+++ b/include/rdma/rdmavt_mr.h
@@ -81,6 +81,7 @@ struct rvt_mregion {
 	u32 mapsz;              /* size of the map array */
 	u8  page_shift;         /* 0 - non unform/non powerof2 sizes */
 	u8  lkey_published;     /* in global table */
+	atomic_t lkey_invalid;	/* true if current lkey is invalid */
 	struct completion comp; /* complete when refcount goes to zero */
 	atomic_t refcount;
 	struct rvt_segarray *map[0];    /* the segments */
diff --git a/include/rdma/rdmavt_qp.h b/include/rdma/rdmavt_qp.h
index 6d23b879416a..bd34d0b56bf7 100644
--- a/include/rdma/rdmavt_qp.h
+++ b/include/rdma/rdmavt_qp.h
@@ -145,6 +145,12 @@
 	(RVT_PROCESS_SEND_OK | RVT_FLUSH_SEND)
 
 /*
+ * Internal send flags
+ */
+#define RVT_SEND_RESERVE_USED           IB_SEND_RESERVED_START
+#define RVT_SEND_COMPLETION_ONLY	(IB_SEND_RESERVED_START << 1)
+
+/*
  * Send work request queue entry.
  * The size of the sg_list is determined when the QP is created and stored
  * in qp->s_max_sge.
@@ -216,23 +222,43 @@ struct rvt_mmap_info {
  * to send a RDMA read response or atomic operation.
  */
 struct rvt_ack_entry {
-	u8 opcode;
-	u8 sent;
+	struct rvt_sge rdma_sge;
+	u64 atomic_data;
 	u32 psn;
 	u32 lpsn;
-	union {
-		struct rvt_sge rdma_sge;
-		u64 atomic_data;
-	};
+	u8 opcode;
+	u8 sent;
 };
 
 #define	RC_QP_SCALING_INTERVAL	5
 
-/*
- * Variables prefixed with s_ are for the requester (sender).
- * Variables prefixed with r_ are for the responder (receiver).
- * Variables prefixed with ack_ are for responder replies.
+#define RVT_OPERATION_PRIV        0x00000001
+#define RVT_OPERATION_ATOMIC      0x00000002
+#define RVT_OPERATION_ATOMIC_SGE  0x00000004
+#define RVT_OPERATION_LOCAL       0x00000008
+#define RVT_OPERATION_USE_RESERVE 0x00000010
+
+#define RVT_OPERATION_MAX (IB_WR_RESERVED10 + 1)
+
+/**
+ * rvt_operation_params - op table entry
+ * @length - the length to copy into the swqe entry
+ * @qpt_support - a bit mask indicating QP type support
+ * @flags - RVT_OPERATION flags (see above)
  *
+ * This supports table driven post send so that
+ * the driver can have differing an potentially
+ * different sets of operations.
+ *
+ **/
+
+struct rvt_operation_params {
+	size_t length;
+	u32 qpt_support;
+	u32 flags;
+};
+
+/*
  * Common variables are protected by both r_rq.lock and s_lock in that order
  * which only happens in modify_qp() or changing the QP 'state'.
  */
@@ -307,6 +333,7 @@ struct rvt_qp {
 	u32 s_next_psn;         /* PSN for next request */
 	u32 s_avail;            /* number of entries avail */
 	u32 s_ssn;              /* SSN of tail entry */
+	atomic_t s_reserved_used; /* reserved entries in use */
 
 	spinlock_t s_lock ____cacheline_aligned_in_smp;
 	u32 s_flags;
@@ -343,6 +370,8 @@ struct rvt_qp {
 	struct rvt_sge_state s_ack_rdma_sge;
 	struct timer_list s_timer;
 
+	atomic_t local_ops_pending; /* number of fast_reg/local_inv reqs */
+
 	/*
 	 * This sge list MUST be last. Do not add anything below here.
 	 */
@@ -436,6 +465,49 @@ static inline struct rvt_rwqe *rvt_get_rwqe_ptr(struct rvt_rq *rq, unsigned n)
 		  rq->max_sge * sizeof(struct ib_sge)) * n);
 }
 
+/**
+ * rvt_qp_wqe_reserve - reserve operation
+ * @qp - the rvt qp
+ * @wqe - the send wqe
+ *
+ * This routine used in post send to record
+ * a wqe relative reserved operation use.
+ */
+static inline void rvt_qp_wqe_reserve(
+	struct rvt_qp *qp,
+	struct rvt_swqe *wqe)
+{
+	wqe->wr.send_flags |= RVT_SEND_RESERVE_USED;
+	atomic_inc(&qp->s_reserved_used);
+}
+
+/**
+ * rvt_qp_wqe_unreserve - clean reserved operation
+ * @qp - the rvt qp
+ * @wqe - the send wqe
+ *
+ * This decrements the reserve use count.
+ *
+ * This call MUST precede the change to
+ * s_last to insure that post send sees a stable
+ * s_avail.
+ *
+ * An smp_mp__after_atomic() is used to insure
+ * the compiler does not juggle the order of the s_last
+ * ring index and the decrementing of s_reserved_used.
+ */
+static inline void rvt_qp_wqe_unreserve(
+	struct rvt_qp *qp,
+	struct rvt_swqe *wqe)
+{
+	if (unlikely(wqe->wr.send_flags & RVT_SEND_RESERVE_USED)) {
+		wqe->wr.send_flags &= ~RVT_SEND_RESERVE_USED;
+		atomic_dec(&qp->s_reserved_used);
+		/* insure no compiler re-order up to s_last change */
+		smp_mb__after_atomic();
+	}
+}
+
 extern const int  ib_rvt_state_ops[];
 
 struct rvt_dev_info;
diff --git a/include/uapi/rdma/hfi/hfi1_user.h b/include/uapi/rdma/hfi/hfi1_user.h
index 98bebf8bef55..d15e7289d835 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 1
+#define HFI1_USER_SWMINOR 2
 
 /*
  * We will encode the major/minor inside a single 32bit version number.