hw/rdma: Implementation of generic rdma device layers

This layer is composed of two sub-modules, backend and resource manager.
Backend sub-module is responsible for all the interaction with IB layers
such as ibverbs and umad (external libraries).
Resource manager is a collection of functions and structures to manage
RDMA resources such as QPs, CQs and MRs.
Reviewed-by: Dotan Barak <dotanb@mellanox.com>
Reviewed-by: Zhu Yanjun <yanjun.zhu@oracle.com>
Signed-off-by: Yuval Shaia <yuval.shaia@oracle.com>
Signed-off-by: Marcel Apfelbaum <marcel@redhat.com>

Makefile.objs

@@ -130,6 +130,7 @@ trace-events-subdirs += hw/block/dataplane
 trace-events-subdirs += hw/char
 trace-events-subdirs += hw/intc
 trace-events-subdirs += hw/net
+trace-events-subdirs += hw/rdma
 trace-events-subdirs += hw/virtio
 trace-events-subdirs += hw/audio
 trace-events-subdirs += hw/misc

configure

@@ -1572,7 +1572,7 @@ disabled with --disable-FEATURE, default is enabled if available:
   hax             HAX acceleration support
   hvf             Hypervisor.framework acceleration support
   whpx            Windows Hypervisor Platform acceleration support
-  rdma            RDMA-based migration support
+  rdma            Enable RDMA-based migration and PVRDMA support
   vde             support for vde network
   netmap          support for netmap network
   linux-aio       Linux AIO support
@@ -2923,15 +2923,16 @@ if test "$rdma" != "no" ; then
 #include <rdma/rdma_cma.h>
 int main(void) { return 0; }
 EOF
-  rdma_libs="-lrdmacm -libverbs"
+  rdma_libs="-lrdmacm -libverbs -libumad"
   if compile_prog "" "$rdma_libs" ; then
     rdma="yes"
+    libs_softmmu="$libs_softmmu $rdma_libs"
   else
     if test "$rdma" = "yes" ; then
         error_exit \
-            " OpenFabrics librdmacm/libibverbs not present." \
+            " OpenFabrics librdmacm/libibverbs/libibumad not present." \
             " Your options:" \
-            "  (1) Fast: Install infiniband packages from your distro." \
+            "  (1) Fast: Install infiniband packages (devel) from your distro." \
             "  (2) Cleanest: Install libraries from www.openfabrics.org" \
             "  (3) Also: Install softiwarp if you don't have RDMA hardware"
     fi

hw/rdma/Makefile.objs

 ifeq ($(CONFIG_RDMA),y)
-obj-$(CONFIG_PCI) += rdma_utils.o
+obj-$(CONFIG_PCI) += rdma_utils.o rdma_backend.o rdma_rm.o
 endif

hw/rdma/rdma_backend.h

+/*
+ *  RDMA device: Definitions of Backend Device functions
+ *
+ * Copyright (C) 2018 Oracle
+ * Copyright (C) 2018 Red Hat Inc
+ *
+ * Authors:
+ *     Yuval Shaia <yuval.shaia@oracle.com>
+ *     Marcel Apfelbaum <marcel@redhat.com>
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2 or later.
+ * See the COPYING file in the top-level directory.
+ *
+ */
+
+#ifndef RDMA_BACKEND_H
+#define RDMA_BACKEND_H
+
+#include <qapi/error.h>
+#include "rdma_rm_defs.h"
+#include "rdma_backend_defs.h"
+
+/* Add definition for QP0 and QP1 as there is no userspace enums for them */
+enum ibv_special_qp_type {
+    IBV_QPT_SMI = 0,
+    IBV_QPT_GSI = 1,
+};
+
+static inline union ibv_gid *rdma_backend_gid(RdmaBackendDev *dev)
+{
+    return &dev->gid;
+}
+
+static inline uint32_t rdma_backend_qpn(const RdmaBackendQP *qp)
+{
+    return qp->ibqp ? qp->ibqp->qp_num : 0;
+}
+
+static inline uint32_t rdma_backend_mr_lkey(const RdmaBackendMR *mr)
+{
+    return mr->ibmr ? mr->ibmr->lkey : 0;
+}
+
+static inline uint32_t rdma_backend_mr_rkey(const RdmaBackendMR *mr)
+{
+    return mr->ibmr ? mr->ibmr->rkey : 0;
+}
+
+int rdma_backend_init(RdmaBackendDev *backend_dev,
+                      RdmaDeviceResources *rdma_dev_res,
+                      const char *backend_device_name, uint8_t port_num,
+                      uint8_t backend_gid_idx, struct ibv_device_attr *dev_attr,
+                      Error **errp);
+void rdma_backend_fini(RdmaBackendDev *backend_dev);
+void rdma_backend_register_comp_handler(void (*handler)(int status,
+                                        unsigned int vendor_err, void *ctx));
+void rdma_backend_unregister_comp_handler(void);
+
+int rdma_backend_query_port(RdmaBackendDev *backend_dev,
+                            struct ibv_port_attr *port_attr);
+int rdma_backend_create_pd(RdmaBackendDev *backend_dev, RdmaBackendPD *pd);
+void rdma_backend_destroy_pd(RdmaBackendPD *pd);
+
+int rdma_backend_create_mr(RdmaBackendMR *mr, RdmaBackendPD *pd, uint64_t addr,
+                           size_t length, int access);
+void rdma_backend_destroy_mr(RdmaBackendMR *mr);
+
+int rdma_backend_create_cq(RdmaBackendDev *backend_dev, RdmaBackendCQ *cq,
+                           int cqe);
+void rdma_backend_destroy_cq(RdmaBackendCQ *cq);
+void rdma_backend_poll_cq(RdmaDeviceResources *rdma_dev_res, RdmaBackendCQ *cq);
+
+int rdma_backend_create_qp(RdmaBackendQP *qp, uint8_t qp_type,
+                           RdmaBackendPD *pd, RdmaBackendCQ *scq,
+                           RdmaBackendCQ *rcq, uint32_t max_send_wr,
+                           uint32_t max_recv_wr, uint32_t max_send_sge,
+                           uint32_t max_recv_sge);
+int rdma_backend_qp_state_init(RdmaBackendDev *backend_dev, RdmaBackendQP *qp,
+                               uint8_t qp_type, uint32_t qkey);
+int rdma_backend_qp_state_rtr(RdmaBackendDev *backend_dev, RdmaBackendQP *qp,
+                              uint8_t qp_type, union ibv_gid *dgid,
+                              uint32_t dqpn, uint32_t rq_psn, uint32_t qkey,
+                              bool use_qkey);
+int rdma_backend_qp_state_rts(RdmaBackendQP *qp, uint8_t qp_type,
+                              uint32_t sq_psn, uint32_t qkey, bool use_qkey);
+void rdma_backend_destroy_qp(RdmaBackendQP *qp);
+
+void rdma_backend_post_send(RdmaBackendDev *backend_dev,
+                            RdmaBackendQP *qp, uint8_t qp_type,
+                            struct ibv_sge *sge, uint32_t num_sge,
+                            union ibv_gid *dgid, uint32_t dqpn, uint32_t dqkey,
+                            void *ctx);
+void rdma_backend_post_recv(RdmaBackendDev *backend_dev,
+                            RdmaDeviceResources *rdma_dev_res,
+                            RdmaBackendQP *qp, uint8_t qp_type,
+                            struct ibv_sge *sge, uint32_t num_sge, void *ctx);
+
+#endif

hw/rdma/rdma_rm.h

+/*
+ * RDMA device: Definitions of Resource Manager functions
+ *
+ * Copyright (C) 2018 Oracle
+ * Copyright (C) 2018 Red Hat Inc
+ *
+ * Authors:
+ *     Yuval Shaia <yuval.shaia@oracle.com>
+ *     Marcel Apfelbaum <marcel@redhat.com>
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2 or later.
+ * See the COPYING file in the top-level directory.
+ *
+ */
+
+#ifndef RDMA_RM_H
+#define RDMA_RM_H
+
+#include <qapi/error.h>
+#include "rdma_backend_defs.h"
+#include "rdma_rm_defs.h"
+
+int rdma_rm_init(RdmaDeviceResources *dev_res, struct ibv_device_attr *dev_attr,
+                 Error **errp);
+void rdma_rm_fini(RdmaDeviceResources *dev_res);
+
+int rdma_rm_alloc_pd(RdmaDeviceResources *dev_res, RdmaBackendDev *backend_dev,
+                     uint32_t *pd_handle, uint32_t ctx_handle);
+RdmaRmPD *rdma_rm_get_pd(RdmaDeviceResources *dev_res, uint32_t pd_handle);
+void rdma_rm_dealloc_pd(RdmaDeviceResources *dev_res, uint32_t pd_handle);
+
+int rdma_rm_alloc_mr(RdmaDeviceResources *dev_res, uint32_t pd_handle,
+                     uint64_t guest_start, size_t guest_length, void *host_virt,
+                     int access_flags, uint32_t *mr_handle, uint32_t *lkey,
+                     uint32_t *rkey);
+RdmaRmMR *rdma_rm_get_mr(RdmaDeviceResources *dev_res, uint32_t mr_handle);
+void rdma_rm_dealloc_mr(RdmaDeviceResources *dev_res, uint32_t mr_handle);
+
+int rdma_rm_alloc_uc(RdmaDeviceResources *dev_res, uint32_t pfn,
+                     uint32_t *uc_handle);
+RdmaRmUC *rdma_rm_get_uc(RdmaDeviceResources *dev_res, uint32_t uc_handle);
+void rdma_rm_dealloc_uc(RdmaDeviceResources *dev_res, uint32_t uc_handle);
+
+int rdma_rm_alloc_cq(RdmaDeviceResources *dev_res, RdmaBackendDev *backend_dev,
+                     uint32_t cqe, uint32_t *cq_handle, void *opaque);
+RdmaRmCQ *rdma_rm_get_cq(RdmaDeviceResources *dev_res, uint32_t cq_handle);
+void rdma_rm_req_notify_cq(RdmaDeviceResources *dev_res, uint32_t cq_handle,
+                           bool notify);
+void rdma_rm_dealloc_cq(RdmaDeviceResources *dev_res, uint32_t cq_handle);
+
+int rdma_rm_alloc_qp(RdmaDeviceResources *dev_res, uint32_t pd_handle,
+                     uint8_t qp_type, uint32_t max_send_wr,
+                     uint32_t max_send_sge, uint32_t send_cq_handle,
+                     uint32_t max_recv_wr, uint32_t max_recv_sge,
+                     uint32_t recv_cq_handle, void *opaque, uint32_t *qpn);
+RdmaRmQP *rdma_rm_get_qp(RdmaDeviceResources *dev_res, uint32_t qpn);
+int rdma_rm_modify_qp(RdmaDeviceResources *dev_res, RdmaBackendDev *backend_dev,
+                      uint32_t qp_handle, uint32_t attr_mask,
+                      union ibv_gid *dgid, uint32_t dqpn,
+                      enum ibv_qp_state qp_state, uint32_t qkey,
+                      uint32_t rq_psn, uint32_t sq_psn);
+void rdma_rm_dealloc_qp(RdmaDeviceResources *dev_res, uint32_t qp_handle);
+
+int rdma_rm_alloc_cqe_ctx(RdmaDeviceResources *dev_res, uint32_t *cqe_ctx_id,
+                          void *ctx);
+void *rdma_rm_get_cqe_ctx(RdmaDeviceResources *dev_res, uint32_t cqe_ctx_id);
+void rdma_rm_dealloc_cqe_ctx(RdmaDeviceResources *dev_res, uint32_t cqe_ctx_id);
+
+#endif

hw/rdma/trace-events

+# See docs/tracing.txt for syntax documentation.
+
+#hw/rdma/rdma_backend.c
+create_ah_cache_hit(uint64_t subnet, uint64_t net_id) "subnet = 0x%"PRIx64" net_id = 0x%"PRIx64
+create_ah_cache_miss(uint64_t subnet, uint64_t net_id) "subnet = 0x%"PRIx64" net_id = 0x%"PRIx64