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

Pull rdma updates from Jason Gunthorpe:
 "This has been a slightly more active cycle than normal with ongoing
  core changes and quite a lot of collected driver updates.

   - Various driver fixes for bnxt_re, cxgb4, hns, mlx5, pvrdma, rxe

   - A new data transfer mode for HFI1 giving higher performance

   - Significant functional and bug fix update to the mlx5
     On-Demand-Paging MR feature

   - A chip hang reset recovery system for hns

   - Change mm->pinned_vm to an atomic64

   - Update bnxt_re to support a new 57500 chip

   - A sane netlink 'rdma link add' method for creating rxe devices and
     fixing the various unregistration race conditions in rxe's
     unregister flow

   - Allow lookup up objects by an ID over netlink

   - Various reworking of the core to driver interface:
       - drivers should not assume umem SGLs are in PAGE_SIZE chunks
       - ucontext is accessed via udata not other means
       - start to make the core code responsible for object memory
         allocation
       - drivers should convert struct device to struct ib_device via a
         helper
       - drivers have more tools to avoid use after unregister problems"

* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma: (280 commits)
  net/mlx5: ODP support for XRC transport is not enabled by default in FW
  IB/hfi1: Close race condition on user context disable and close
  RDMA/umem: Revert broken 'off by one' fix
  RDMA/umem: minor bug fix in error handling path
  RDMA/hns: Use GFP_ATOMIC in hns_roce_v2_modify_qp
  cxgb4: kfree mhp after the debug print
  IB/rdmavt: Fix concurrency panics in QP post_send and modify to error
  IB/rdmavt: Fix loopback send with invalidate ordering
  IB/iser: Fix dma_nents type definition
  IB/mlx5: Set correct write permissions for implicit ODP MR
  bnxt_re: Clean cq for kernel consumers only
  RDMA/uverbs: Don't do double free of allocated PD
  RDMA: Handle ucontext allocations by IB/core
  RDMA/core: Fix a WARN() message
  bnxt_re: fix the regression due to changes in alloc_pbl
  IB/mlx4: Increase the timeout for CM cache
  IB/core: Abort page fault handler silently during owning process exit
  IB/mlx5: Validate correct PD before prefetch MR
  IB/mlx5: Protect against prefetch of invalid MR
  RDMA/uverbs: Store PR pointer before it is overwritten
  ...

1 files changed, 85 insertions, 0 deletions

diff --git a/drivers/infiniband/hw/hfi1/opfn.h b/drivers/infiniband/hw/hfi1/opfn.h
new file mode 100644
index 000000000000..5f2011cabc25
--- /dev/null
+++ b/drivers/infiniband/hw/hfi1/opfn.h
@@ -0,0 +1,85 @@
+/* SPDX-License-Identifier: (GPL-2.0 OR BSD-3-Clause) */
+/*
+ * Copyright(c) 2018 Intel Corporation.
+ *
+ */
+#ifndef _HFI1_OPFN_H
+#define _HFI1_OPFN_H
+
+/**
+ * DOC: Omni Path Feature Negotion (OPFN)
+ *
+ * OPFN is a discovery protocol for Intel Omni-Path fabric that
+ * allows two RC QPs to negotiate a common feature that both QPs
+ * can support. Currently, the only OPA feature that OPFN
+ * supports is TID RDMA.
+ *
+ * Architecture
+ *
+ * OPFN involves the communication between two QPs on the HFI
+ * level on an Omni-Path fabric, and ULPs have no knowledge of
+ * OPFN at all.
+ *
+ * Implementation
+ *
+ * OPFN extends the existing IB RC protocol with the following
+ * changes:
+ * -- Uses Bit 24 (reserved) of DWORD 1 of Base Transport
+ *    Header (BTH1) to indicate that the RC QP supports OPFN;
+ * -- Uses a combination of RC COMPARE_SWAP opcode (0x13) and
+ *    the address U64_MAX (0xFFFFFFFFFFFFFFFF) as an OPFN
+ *    request; The 64-bit data carried with the request/response
+ *    contains the parameters for negotiation and will be
+ *    defined in tid_rdma.c file;
+ * -- Defines IB_WR_RESERVED3 as IB_WR_OPFN.
+ *
+ * The OPFN communication will be triggered when an RC QP
+ * receives a request with Bit 24 of BTH1 set. The responder QP
+ * will then post send an OPFN request with its local
+ * parameters, which will be sent to the requester QP once all
+ * existing requests on the responder QP side have been sent.
+ * Once the requester QP receives the OPFN request, it will
+ * keep a copy of the responder QP's parameters, and return a
+ * response packet with its own local parameters. The responder
+ * QP receives the response packet and keeps a copy of the requester
+ * QP's parameters. After this exchange, each side has the parameters
+ * for both sides and therefore can select the right parameters
+ * for future transactions
+ */
+
+/* STL Verbs Extended */
+#define IB_BTHE_E_SHIFT           24
+#define HFI1_VERBS_E_ATOMIC_VADDR U64_MAX
+
+struct ib_atomic_eth;
+
+enum hfi1_opfn_codes {
+	STL_VERBS_EXTD_NONE = 0,
+	STL_VERBS_EXTD_TID_RDMA,
+	STL_VERBS_EXTD_MAX
+};
+
+struct hfi1_opfn_data {
+	u8 extended;
+	u16 requested;
+	u16 completed;
+	enum hfi1_opfn_codes curr;
+	/* serialize opfn function calls */
+	spinlock_t lock;
+	struct work_struct opfn_work;
+};
+
+/* WR opcode for OPFN */
+#define IB_WR_OPFN IB_WR_RESERVED3
+
+void opfn_send_conn_request(struct work_struct *work);
+void opfn_conn_response(struct rvt_qp *qp, struct rvt_ack_entry *e,
+			struct ib_atomic_eth *ateth);
+void opfn_conn_reply(struct rvt_qp *qp, u64 data);
+void opfn_conn_error(struct rvt_qp *qp);
+void opfn_qp_init(struct rvt_qp *qp, struct ib_qp_attr *attr, int attr_mask);
+void opfn_trigger_conn_request(struct rvt_qp *qp, u32 bth1);
+int opfn_init(void);
+void opfn_exit(void);
+
+#endif /* _HFI1_OPFN_H */