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 smaller cycle than normal. One new driver was
  accepted, which is unusual, and at least one more driver remains in
  review on the list.

  Summary:

   - Driver fixes for hns, hfi1, nes, rxe, i40iw, mlx5, cxgb4,
     vmw_pvrdma

   - Many patches from MatthewW converting radix tree and IDR users to
     use xarray

   - Introduction of tracepoints to the MAD layer

   - Build large SGLs at the start for DMA mapping and get the driver to
     split them

   - Generally clean SGL handling code throughout the subsystem

   - Support for restricting RDMA devices to net namespaces for
     containers

   - Progress to remove object allocation boilerplate code from drivers

   - Change in how the mlx5 driver shows representor ports linked to VFs

   - mlx5 uapi feature to access the on chip SW ICM memory

   - Add a new driver for 'EFA'. This is HW that supports user space
     packet processing through QPs in Amazon's cloud"

* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma: (186 commits)
  RDMA/ipoib: Allow user space differentiate between valid dev_port
  IB/core, ipoib: Do not overreact to SM LID change event
  RDMA/device: Don't fire uevent before device is fully initialized
  lib/scatterlist: Remove leftover from sg_page_iter comment
  RDMA/efa: Add driver to Kconfig/Makefile
  RDMA/efa: Add the efa module
  RDMA/efa: Add EFA verbs implementation
  RDMA/efa: Add common command handlers
  RDMA/efa: Implement functions that submit and complete admin commands
  RDMA/efa: Add the ABI definitions
  RDMA/efa: Add the com service API definitions
  RDMA/efa: Add the efa_com.h file
  RDMA/efa: Add the efa.h header file
  RDMA/efa: Add EFA device definitions
  RDMA: Add EFA related definitions
  RDMA/umem: Remove hugetlb flag
  RDMA/bnxt_re: Use core helpers to get aligned DMA address
  RDMA/i40iw: Use core helpers to get aligned DMA address within a supported page size
  RDMA/verbs: Add a DMA iterator to return aligned contiguous memory blocks
  RDMA/umem: Add API to find best driver supported page size in an MR
  ...

drivers/infiniband/sw/rdmavt/ah.c

@@ -89,36 +89,29 @@ EXPORT_SYMBOL(rvt_check_ah);
 
 /**
  * rvt_create_ah - create an address handle
- * @pd: the protection domain
+ * @ibah: the IB address handle
  * @ah_attr: the attributes of the AH
  * @create_flags: create address handle flags (see enum rdma_create_ah_flags)
  * @udata: pointer to user's input output buffer information.
  *
  * This may be called from interrupt context.
  *
- * Return: newly allocated ah
+ * Return: 0 on success
  */
-struct ib_ah *rvt_create_ah(struct ib_pd *pd,
-			    struct rdma_ah_attr *ah_attr,
-			    u32 create_flags,
-			    struct ib_udata *udata)
+int rvt_create_ah(struct ib_ah *ibah, struct rdma_ah_attr *ah_attr,
+		  u32 create_flags, struct ib_udata *udata)
 {
-	struct rvt_ah *ah;
-	struct rvt_dev_info *dev = ib_to_rvt(pd->device);
+	struct rvt_ah *ah = ibah_to_rvtah(ibah);
+	struct rvt_dev_info *dev = ib_to_rvt(ibah->device);
 	unsigned long flags;
 
-	if (rvt_check_ah(pd->device, ah_attr))
-		return ERR_PTR(-EINVAL);
-
-	ah = kmalloc(sizeof(*ah), GFP_ATOMIC);
-	if (!ah)
-		return ERR_PTR(-ENOMEM);
+	if (rvt_check_ah(ibah->device, ah_attr))
+		return -EINVAL;
 
 	spin_lock_irqsave(&dev->n_ahs_lock, flags);
 	if (dev->n_ahs_allocated == dev->dparms.props.max_ah) {
 		spin_unlock_irqrestore(&dev->n_ahs_lock, flags);
-		kfree(ah);
-		return ERR_PTR(-ENOMEM);
+		return -ENOMEM;
 	}
 
 	dev->n_ahs_allocated++;
@@ -129,35 +122,32 @@ struct ib_ah *rvt_create_ah(struct ib_pd *pd,
 	atomic_set(&ah->refcount, 0);
 
 	if (dev->driver_f.notify_new_ah)
-		dev->driver_f.notify_new_ah(pd->device, ah_attr, ah);
+		dev->driver_f.notify_new_ah(ibah->device, ah_attr, ah);
 
-	return &ah->ibah;
+	return 0;
 }
 
 /**
  * rvt_destory_ah - Destory an address handle
  * @ibah: address handle
  * @destroy_flags: destroy address handle flags (see enum rdma_destroy_ah_flags)
+ * @udata: user data or NULL for kernel object
  *
  * Return: 0 on success
  */
-int rvt_destroy_ah(struct ib_ah *ibah, u32 destroy_flags)
+void rvt_destroy_ah(struct ib_ah *ibah, u32 destroy_flags)
 {
 	struct rvt_dev_info *dev = ib_to_rvt(ibah->device);
 	struct rvt_ah *ah = ibah_to_rvtah(ibah);
 	unsigned long flags;
 
-	if (atomic_read(&ah->refcount) != 0)
-		return -EBUSY;
+	WARN_ON_ONCE(atomic_read(&ah->refcount));
 
 	spin_lock_irqsave(&dev->n_ahs_lock, flags);
 	dev->n_ahs_allocated--;
 	spin_unlock_irqrestore(&dev->n_ahs_lock, flags);
 
 	rdma_destroy_ah_attr(&ah->attr);
-	kfree(ah);
-
-	return 0;
 }
 
 /**

drivers/infiniband/sw/rdmavt/ah.h

@@ -50,11 +50,9 @@
 
 #include <rdma/rdma_vt.h>
 
-struct ib_ah *rvt_create_ah(struct ib_pd *pd,
-			    struct rdma_ah_attr *ah_attr,
-			    u32 create_flags,
-			    struct ib_udata *udata);
-int rvt_destroy_ah(struct ib_ah *ibah, u32 destroy_flags);
+int rvt_create_ah(struct ib_ah *ah, struct rdma_ah_attr *ah_attr,
+		  u32 create_flags, struct ib_udata *udata);
+void rvt_destroy_ah(struct ib_ah *ibah, u32 destroy_flags);
 int rvt_modify_ah(struct ib_ah *ibah, struct rdma_ah_attr *ah_attr);
 int rvt_query_ah(struct ib_ah *ibah, struct rdma_ah_attr *ah_attr);
 

drivers/infiniband/sw/rdmavt/cq.c

@@ -168,7 +168,6 @@ static void send_complete(struct work_struct *work)
  * rvt_create_cq - create a completion queue
  * @ibdev: the device this completion queue is attached to
  * @attr: creation attributes
- * @context: unused by the QLogic_IB driver
  * @udata: user data for libibverbs.so
  *
  * Called by ib_create_cq() in the generic verbs code.
@@ -178,7 +177,6 @@ static void send_complete(struct work_struct *work)
  */
 struct ib_cq *rvt_create_cq(struct ib_device *ibdev,
 			    const struct ib_cq_init_attr *attr,
-			    struct ib_ucontext *context,
 			    struct ib_udata *udata)
 {
 	struct rvt_dev_info *rdi = ib_to_rvt(ibdev);
@@ -232,7 +230,7 @@ struct ib_cq *rvt_create_cq(struct ib_device *ibdev,
 	if (udata && udata->outlen >= sizeof(__u64)) {
 		int err;
 
-		cq->ip = rvt_create_mmap_info(rdi, sz, context, wc);
+		cq->ip = rvt_create_mmap_info(rdi, sz, udata, wc);
 		if (!cq->ip) {
 			ret = ERR_PTR(-ENOMEM);
 			goto bail_wc;
@@ -299,12 +297,13 @@ done:
 /**
  * rvt_destroy_cq - destroy a completion queue
  * @ibcq: the completion queue to destroy.
+ * @udata: user data or NULL for kernel object
  *
  * Called by ib_destroy_cq() in the generic verbs code.
  *
  * Return: always 0
  */
-int rvt_destroy_cq(struct ib_cq *ibcq)
+int rvt_destroy_cq(struct ib_cq *ibcq, struct ib_udata *udata)
 {
 	struct rvt_cq *cq = ibcq_to_rvtcq(ibcq);
 	struct rvt_dev_info *rdi = cq->rdi;

drivers/infiniband/sw/rdmavt/cq.h

@@ -53,9 +53,8 @@
 
 struct ib_cq *rvt_create_cq(struct ib_device *ibdev,
 			    const struct ib_cq_init_attr *attr,
-			    struct ib_ucontext *context,
 			    struct ib_udata *udata);
-int rvt_destroy_cq(struct ib_cq *ibcq);
+int rvt_destroy_cq(struct ib_cq *ibcq, struct ib_udata *udata);
 int rvt_req_notify_cq(struct ib_cq *ibcq, enum ib_cq_notify_flags notify_flags);
 int rvt_resize_cq(struct ib_cq *ibcq, int cqe, struct ib_udata *udata);
 int rvt_poll_cq(struct ib_cq *ibcq, int num_entries, struct ib_wc *entry);

drivers/infiniband/sw/rdmavt/mmap.c

@@ -49,6 +49,7 @@
 #include <linux/vmalloc.h>
 #include <linux/mm.h>
 #include <asm/pgtable.h>
+#include <rdma/uverbs_ioctl.h>
 #include "mmap.h"
 
 /**
@@ -150,18 +151,19 @@ done:
  * rvt_create_mmap_info - allocate information for hfi1_mmap
  * @rdi: rvt dev struct
  * @size: size in bytes to map
- * @context: user context
+ * @udata: user data (must be valid!)
  * @obj: opaque pointer to a cq, wq etc
  *
  * Return: rvt_mmap struct on success
  */
-struct rvt_mmap_info *rvt_create_mmap_info(struct rvt_dev_info *rdi,
-					   u32 size,
-					   struct ib_ucontext *context,
-					   void *obj)
+struct rvt_mmap_info *rvt_create_mmap_info(struct rvt_dev_info *rdi, u32 size,
+					   struct ib_udata *udata, void *obj)
 {
 	struct rvt_mmap_info *ip;
 
+	if (!udata)
+		return ERR_PTR(-EINVAL);
+
 	ip = kmalloc_node(sizeof(*ip), GFP_KERNEL, rdi->dparms.node);
 	if (!ip)
 		return ip;
@@ -177,7 +179,9 @@ struct rvt_mmap_info *rvt_create_mmap_info(struct rvt_dev_info *rdi,
 
 	INIT_LIST_HEAD(&ip->pending_mmaps);
 	ip->size = size;
-	ip->context = context;
+	ip->context =
+		container_of(udata, struct uverbs_attr_bundle, driver_udata)
+			->context;
 	ip->obj = obj;
 	kref_init(&ip->ref);
 

drivers/infiniband/sw/rdmavt/mmap.h

@@ -53,10 +53,8 @@
 void rvt_mmap_init(struct rvt_dev_info *rdi);
 void rvt_release_mmap_info(struct kref *ref);
 int rvt_mmap(struct ib_ucontext *context, struct vm_area_struct *vma);
-struct rvt_mmap_info *rvt_create_mmap_info(struct rvt_dev_info *rdi,
-					   u32 size,
-					   struct ib_ucontext *context,
-					   void *obj);
+struct rvt_mmap_info *rvt_create_mmap_info(struct rvt_dev_info *rdi, u32 size,
+					   struct ib_udata *udata, void *obj);
 void rvt_update_mmap_info(struct rvt_dev_info *rdi, struct rvt_mmap_info *ip,
 			  u32 size, void *obj);
 

drivers/infiniband/sw/rdmavt/mr.c

@@ -392,7 +392,7 @@ struct ib_mr *rvt_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 	if (IS_ERR(umem))
 		return (void *)umem;
 
-	n = umem->nmap;
+	n = ib_umem_num_pages(umem);
 
 	mr = __rvt_alloc_mr(n, pd);
 	if (IS_ERR(mr)) {
@@ -548,7 +548,7 @@ bool rvt_ss_has_lkey(struct rvt_sge_state *ss, u32 lkey)
  *
  * Returns 0 on success.
  */
-int rvt_dereg_mr(struct ib_mr *ibmr)
+int rvt_dereg_mr(struct ib_mr *ibmr, struct ib_udata *udata)
 {
 	struct rvt_mr *mr = to_imr(ibmr);
 	int ret;
@@ -575,9 +575,8 @@ out:
  *
  * Return: the memory region on success, otherwise return an errno.
  */
-struct ib_mr *rvt_alloc_mr(struct ib_pd *pd,
-			   enum ib_mr_type mr_type,
-			   u32 max_num_sg)
+struct ib_mr *rvt_alloc_mr(struct ib_pd *pd, enum ib_mr_type mr_type,
+			   u32 max_num_sg, struct ib_udata *udata)
 {
 	struct rvt_mr *mr;
 

drivers/infiniband/sw/rdmavt/mr.h

@@ -78,10 +78,9 @@ struct ib_mr *rvt_get_dma_mr(struct ib_pd *pd, int acc);
 struct ib_mr *rvt_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 			      u64 virt_addr, int mr_access_flags,
 			      struct ib_udata *udata);
-int rvt_dereg_mr(struct ib_mr *ibmr);
-struct ib_mr *rvt_alloc_mr(struct ib_pd *pd,
-			   enum ib_mr_type mr_type,
-			   u32 max_num_sg);
+int rvt_dereg_mr(struct ib_mr *ibmr, struct ib_udata *udata);
+struct ib_mr *rvt_alloc_mr(struct ib_pd *pd, enum ib_mr_type mr_type,
+			   u32 max_num_sg, struct ib_udata *udata);
 int rvt_map_mr_sg(struct ib_mr *ibmr, struct scatterlist *sg,
 		  int sg_nents, unsigned int *sg_offset);
 struct ib_fmr *rvt_alloc_fmr(struct ib_pd *pd, int mr_access_flags,

drivers/infiniband/sw/rdmavt/pd.c

@@ -51,15 +51,13 @@
 /**
  * rvt_alloc_pd - allocate a protection domain
  * @ibpd: PD
- * @context: optional user context
  * @udata: optional user data
  *
  * Allocate and keep track of a PD.
  *
  * Return: 0 on success
  */
-int rvt_alloc_pd(struct ib_pd *ibpd, struct ib_ucontext *context,
-		 struct ib_udata *udata)
+int rvt_alloc_pd(struct ib_pd *ibpd, struct ib_udata *udata)
 {
 	struct ib_device *ibdev = ibpd->device;
 	struct rvt_dev_info *dev = ib_to_rvt(ibdev);
@@ -93,10 +91,11 @@ bail:
 /**
  * rvt_dealloc_pd - Free PD
  * @ibpd: Free up PD
+ * @udata: Valid user data or NULL for kernel object
  *
  * Return: always 0
  */
-void rvt_dealloc_pd(struct ib_pd *ibpd)
+void rvt_dealloc_pd(struct ib_pd *ibpd, struct ib_udata *udata)
 {
 	struct rvt_dev_info *dev = ib_to_rvt(ibpd->device);
 

drivers/infiniband/sw/rdmavt/pd.h

@@ -50,8 +50,7 @@
 
 #include <rdma/rdma_vt.h>
 
-int rvt_alloc_pd(struct ib_pd *pd, struct ib_ucontext *context,
-		 struct ib_udata *udata);
-void rvt_dealloc_pd(struct ib_pd *ibpd);
+int rvt_alloc_pd(struct ib_pd *pd, struct ib_udata *udata);
+void rvt_dealloc_pd(struct ib_pd *ibpd, struct ib_udata *udata);
 
 #endif          /* DEF_RDMAVTPD_H */

drivers/infiniband/sw/rdmavt/qp.c

@@ -623,13 +623,7 @@ static void rvt_clear_mr_refs(struct rvt_qp *qp, int clr_sends)
 		while (qp->s_last != qp->s_head) {
 			struct rvt_swqe *wqe = rvt_get_swqe_ptr(qp, qp->s_last);
 
-			rvt_put_swqe(wqe);
-
-			if (qp->ibqp.qp_type == IB_QPT_UD ||
-			    qp->ibqp.qp_type == IB_QPT_SMI ||
-			    qp->ibqp.qp_type == IB_QPT_GSI)
-				atomic_dec(&ibah_to_rvtah(
-						wqe->ud_wr.ah)->refcount);
+			rvt_put_qp_swqe(qp, wqe);
 			if (++qp->s_last >= qp->s_size)
 				qp->s_last = 0;
 			smp_wmb(); /* see qp_set_savail */
@@ -957,8 +951,6 @@ struct ib_qp *rvt_create_qp(struct ib_pd *ibpd,
 	size_t sg_list_sz;
 	struct ib_qp *ret = ERR_PTR(-ENOMEM);
 	struct rvt_dev_info *rdi = ib_to_rvt(ibpd->device);
-	struct rvt_ucontext *ucontext = rdma_udata_to_drv_context(
-		udata, struct rvt_ucontext, ibucontext);
 	void *priv = NULL;
 	size_t sqsize;
 
@@ -1131,8 +1123,7 @@ struct ib_qp *rvt_create_qp(struct ib_pd *ibpd,
 		} else {
 			u32 s = sizeof(struct rvt_rwq) + qp->r_rq.size * sz;
 
-			qp->ip = rvt_create_mmap_info(rdi, s,
-						      &ucontext->ibucontext,
+			qp->ip = rvt_create_mmap_info(rdi, s, udata,
 						      qp->r_rq.wq);
 			if (!qp->ip) {
 				ret = ERR_PTR(-ENOMEM);
@@ -1617,7 +1608,7 @@ inval:
  *
  * Return: 0 on success.
  */
-int rvt_destroy_qp(struct ib_qp *ibqp)
+int rvt_destroy_qp(struct ib_qp *ibqp, struct ib_udata *udata)
 {
 	struct rvt_qp *qp = ibqp_to_rvtqp(ibqp);
 	struct rvt_dev_info *rdi = ib_to_rvt(ibqp->device);
@@ -2018,8 +2009,7 @@ static int rvt_post_one_wr(struct rvt_qp *qp,
 	 * opportunity to adjust PSN values based on internal checks.
 	 */
 	log_pmtu = qp->log_pmtu;
-	if (qp->ibqp.qp_type != IB_QPT_UC &&
-	    qp->ibqp.qp_type != IB_QPT_RC) {
+	if (qp->allowed_ops == IB_OPCODE_UD) {
 		struct rvt_ah *ah = ibah_to_rvtah(wqe->ud_wr.ah);
 
 		log_pmtu = ah->log_pmtu;
@@ -2067,8 +2057,7 @@ static int rvt_post_one_wr(struct rvt_qp *qp,
 	return 0;
 
 bail_inval_free_ref:
-	if (qp->ibqp.qp_type != IB_QPT_UC &&
-	    qp->ibqp.qp_type != IB_QPT_RC)
+	if (qp->allowed_ops == IB_OPCODE_UD)
 		atomic_dec(&ibah_to_rvtah(ud_wr(wr)->ah)->refcount);
 bail_inval_free:
 	/* release mr holds */
@@ -2691,11 +2680,7 @@ void rvt_send_complete(struct rvt_qp *qp, struct rvt_swqe *wqe,
 	qp->s_last = last;
 	/* See post_send() */
 	barrier();
-	rvt_put_swqe(wqe);
-	if (qp->ibqp.qp_type == IB_QPT_UD ||
-	    qp->ibqp.qp_type == IB_QPT_SMI ||
-	    qp->ibqp.qp_type == IB_QPT_GSI)
-		atomic_dec(&ibah_to_rvtah(wqe->ud_wr.ah)->refcount);
+	rvt_put_qp_swqe(qp, wqe);
 
 	rvt_qp_swqe_complete(qp,
 			     wqe,

drivers/infiniband/sw/rdmavt/qp.h

@@ -48,7 +48,7 @@
  *
  */
 
-#include <rdma/rdma_vt.h>
+#include <rdma/rdmavt_qp.h>
 
 int rvt_driver_qp_init(struct rvt_dev_info *rdi);
 void rvt_qp_exit(struct rvt_dev_info *rdi);
@@ -57,7 +57,7 @@ struct ib_qp *rvt_create_qp(struct ib_pd *ibpd,
 			    struct ib_udata *udata);
 int rvt_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
 		  int attr_mask, struct ib_udata *udata);
-int rvt_destroy_qp(struct ib_qp *ibqp);
+int rvt_destroy_qp(struct ib_qp *ibqp, struct ib_udata *udata);
 int rvt_query_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
 		 int attr_mask, struct ib_qp_init_attr *init_attr);
 int rvt_post_recv(struct ib_qp *ibqp, const struct ib_recv_wr *wr,

drivers/infiniband/sw/rdmavt/rc.c

@@ -45,7 +45,7 @@
  *
  */
 
-#include <rdma/rdma_vt.h>
+#include <rdma/rdmavt_qp.h>
 #include <rdma/ib_hdrs.h>
 
 /*

drivers/infiniband/sw/rdmavt/srq.c

@@ -71,31 +71,24 @@ void rvt_driver_srq_init(struct rvt_dev_info *rdi)
  * @srq_init_attr: the attributes of the SRQ
  * @udata: data from libibverbs when creating a user SRQ
  *
- * Return: Allocated srq object
+ * Return: 0 on success
  */
-struct ib_srq *rvt_create_srq(struct ib_pd *ibpd,
-			      struct ib_srq_init_attr *srq_init_attr,
-			      struct ib_udata *udata)
+int rvt_create_srq(struct ib_srq *ibsrq, struct ib_srq_init_attr *srq_init_attr,
+		   struct ib_udata *udata)
 {
-	struct rvt_dev_info *dev = ib_to_rvt(ibpd->device);
-	struct rvt_ucontext *ucontext = rdma_udata_to_drv_context(
-		udata, struct rvt_ucontext, ibucontext);
-	struct rvt_srq *srq;
+	struct rvt_dev_info *dev = ib_to_rvt(ibsrq->device);
+	struct rvt_srq *srq = ibsrq_to_rvtsrq(ibsrq);
 	u32 sz;
-	struct ib_srq *ret;
+	int ret;
 
 	if (srq_init_attr->srq_type != IB_SRQT_BASIC)
-		return ERR_PTR(-EOPNOTSUPP);
+		return -EOPNOTSUPP;
 
 	if (srq_init_attr->attr.max_sge == 0 ||
 	    srq_init_attr->attr.max_sge > dev->dparms.props.max_srq_sge ||
 	    srq_init_attr->attr.max_wr == 0 ||
 	    srq_init_attr->attr.max_wr > dev->dparms.props.max_srq_wr)
-		return ERR_PTR(-EINVAL);
-
-	srq = kzalloc_node(sizeof(*srq), GFP_KERNEL, dev->dparms.node);
-	if (!srq)
-		return ERR_PTR(-ENOMEM);
+		return -EINVAL;
 
 	/*
 	 * Need to use vmalloc() if we want to support large #s of entries.
@@ -109,7 +102,7 @@ struct ib_srq *rvt_create_srq(struct ib_pd *ibpd,
 		vzalloc_node(sizeof(struct rvt_rwq) + srq->rq.size * sz,
 			     dev->dparms.node);
 	if (!srq->rq.wq) {
-		ret = ERR_PTR(-ENOMEM);
+		ret = -ENOMEM;
 		goto bail_srq;
 	}
 
@@ -118,23 +111,18 @@ struct ib_srq *rvt_create_srq(struct ib_pd *ibpd,
 	 * See rvt_mmap() for details.
 	 */
 	if (udata && udata->outlen >= sizeof(__u64)) {
-		int err;
 		u32 s = sizeof(struct rvt_rwq) + srq->rq.size * sz;
 
-		srq->ip =
-		    rvt_create_mmap_info(dev, s, &ucontext->ibucontext,
-					 srq->rq.wq);
+		srq->ip = rvt_create_mmap_info(dev, s, udata, srq->rq.wq);
 		if (!srq->ip) {
-			ret = ERR_PTR(-ENOMEM);
+			ret = -ENOMEM;
 			goto bail_wq;
 		}
 
-		err = ib_copy_to_udata(udata, &srq->ip->offset,
+		ret = ib_copy_to_udata(udata, &srq->ip->offset,
 				       sizeof(srq->ip->offset));
-		if (err) {
-			ret = ERR_PTR(err);
+		if (ret)
 			goto bail_ip;
-		}
 	}
 
 	/*
@@ -146,7 +134,7 @@ struct ib_srq *rvt_create_srq(struct ib_pd *ibpd,
 	spin_lock(&dev->n_srqs_lock);
 	if (dev->n_srqs_allocated == dev->dparms.props.max_srq) {
 		spin_unlock(&dev->n_srqs_lock);
-		ret = ERR_PTR(-ENOMEM);
+		ret = -ENOMEM;
 		goto bail_ip;
 	}
 
@@ -159,14 +147,13 @@ struct ib_srq *rvt_create_srq(struct ib_pd *ibpd,
 		spin_unlock_irq(&dev->pending_lock);
 	}
 
-	return &srq->ibsrq;
+	return 0;
 
 bail_ip:
 	kfree(srq->ip);
 bail_wq:
 	vfree(srq->rq.wq);
 bail_srq:
-	kfree(srq);
 	return ret;
 }
 
@@ -338,9 +325,8 @@ int rvt_query_srq(struct ib_srq *ibsrq, struct ib_srq_attr *attr)
  * rvt_destroy_srq - destory an srq
  * @ibsrq: srq object to destroy
  *
- * Return always 0
  */
-int rvt_destroy_srq(struct ib_srq *ibsrq)
+void rvt_destroy_srq(struct ib_srq *ibsrq, struct ib_udata *udata)
 {
 	struct rvt_srq *srq = ibsrq_to_rvtsrq(ibsrq);
 	struct rvt_dev_info *dev = ib_to_rvt(ibsrq->device);
@@ -352,7 +338,4 @@ int rvt_destroy_srq(struct ib_srq *ibsrq)
 		kref_put(&srq->ip->ref, rvt_release_mmap_info);
 	else
 		vfree(srq->rq.wq);
-	kfree(srq);
-
-	return 0;
 }

drivers/infiniband/sw/rdmavt/srq.h

@@ -50,13 +50,12 @@
 
 #include <rdma/rdma_vt.h>
 void rvt_driver_srq_init(struct rvt_dev_info *rdi);
-struct ib_srq *rvt_create_srq(struct ib_pd *ibpd,
-			      struct ib_srq_init_attr *srq_init_attr,
-			      struct ib_udata *udata);
+int rvt_create_srq(struct ib_srq *ibsrq, struct ib_srq_init_attr *srq_init_attr,
+		   struct ib_udata *udata);
 int rvt_modify_srq(struct ib_srq *ibsrq, struct ib_srq_attr *attr,
 		   enum ib_srq_attr_mask attr_mask,
 		   struct ib_udata *udata);
 int rvt_query_srq(struct ib_srq *ibsrq, struct ib_srq_attr *attr);
-int rvt_destroy_srq(struct ib_srq *ibsrq);
+void rvt_destroy_srq(struct ib_srq *ibsrq, struct ib_udata *udata);
 
 #endif          /* DEF_RVTSRQ_H */

drivers/infiniband/sw/rdmavt/trace_qp.h

@@ -51,7 +51,7 @@
 #include <linux/trace_seq.h>
 
 #include <rdma/ib_verbs.h>
-#include <rdma/rdma_vt.h>
+#include <rdma/rdmavt_qp.h>
 
 #undef TRACE_SYSTEM
 #define TRACE_SYSTEM rvt_qp

drivers/infiniband/sw/rdmavt/trace_rc.h

@@ -51,7 +51,7 @@
 #include <linux/trace_seq.h>
 
 #include <rdma/ib_verbs.h>
-#include <rdma/rdma_vt.h>
+#include <rdma/rdmavt_qp.h>
 
 #undef TRACE_SYSTEM
 #define TRACE_SYSTEM rvt_rc

drivers/infiniband/sw/rdmavt/trace_tx.h

@@ -51,7 +51,7 @@
 #include <linux/trace_seq.h>
 
 #include <rdma/ib_verbs.h>
-#include <rdma/rdma_vt.h>
+#include <rdma/rdmavt_qp.h>
 
 #undef TRACE_SYSTEM
 #define TRACE_SYSTEM rvt_tx

drivers/infiniband/sw/rdmavt/vt.c

@@ -425,7 +425,10 @@ static const struct ib_device_ops rvt_dev_ops = {
 	.req_notify_cq = rvt_req_notify_cq,
 	.resize_cq = rvt_resize_cq,
 	.unmap_fmr = rvt_unmap_fmr,
+
+	INIT_RDMA_OBJ_SIZE(ib_ah, rvt_ah, ibah),
 	INIT_RDMA_OBJ_SIZE(ib_pd, rvt_pd, ibpd),
+	INIT_RDMA_OBJ_SIZE(ib_srq, rvt_srq, ibsrq),
 	INIT_RDMA_OBJ_SIZE(ib_ucontext, rvt_ucontext, ibucontext),
 };
 

drivers/infiniband/sw/rxe/rxe_cq.c

@@ -82,7 +82,7 @@ static void rxe_send_complete(unsigned long data)
 }
 
 int rxe_cq_from_init(struct rxe_dev *rxe, struct rxe_cq *cq, int cqe,
-		     int comp_vector, struct ib_ucontext *context,
+		     int comp_vector, struct ib_udata *udata,
 		     struct rxe_create_cq_resp __user *uresp)
 {
 	int err;
@@ -94,7 +94,7 @@ int rxe_cq_from_init(struct rxe_dev *rxe, struct rxe_cq *cq, int cqe,
 		return -ENOMEM;
 	}
 
-	err = do_mmap_info(rxe, uresp ? &uresp->mi : NULL, context,
+	err = do_mmap_info(rxe, uresp ? &uresp->mi : NULL, udata,
 			   cq->queue->buf, cq->queue->buf_size, &cq->queue->ip);
 	if (err) {
 		vfree(cq->queue->buf);
@@ -115,13 +115,13 @@ int rxe_cq_from_init(struct rxe_dev *rxe, struct rxe_cq *cq, int cqe,
 }
 
 int rxe_cq_resize_queue(struct rxe_cq *cq, int cqe,
-			struct rxe_resize_cq_resp __user *uresp)
+			struct rxe_resize_cq_resp __user *uresp,
+			struct ib_udata *udata)
 {
 	int err;
 
 	err = rxe_queue_resize(cq->queue, (unsigned int *)&cqe,
-			       sizeof(struct rxe_cqe),
-			       cq->queue->ip ? cq->queue->ip->context : NULL,
+			       sizeof(struct rxe_cqe), udata,
 			       uresp ? &uresp->mi : NULL, NULL, &cq->cq_lock);
 	if (!err)
 		cq->ibcq.cqe = cqe;

drivers/infiniband/sw/rxe/rxe_hdr.h

@@ -643,7 +643,7 @@ struct rxe_atmeth {
 	__be32			rkey;
 	__be64			swap_add;
 	__be64			comp;
-} __attribute__((__packed__));
+} __packed;
 
 static inline u64 __atmeth_va(void *arg)
 {

drivers/infiniband/sw/rxe/rxe_loc.h

@@ -53,11 +53,12 @@ int rxe_cq_chk_attr(struct rxe_dev *rxe, struct rxe_cq *cq,
 		    int cqe, int comp_vector);
 
 int rxe_cq_from_init(struct rxe_dev *rxe, struct rxe_cq *cq, int cqe,
-		     int comp_vector, struct ib_ucontext *context,
+		     int comp_vector, struct ib_udata *udata,
 		     struct rxe_create_cq_resp __user *uresp);
 
 int rxe_cq_resize_queue(struct rxe_cq *cq, int new_cqe,
-			struct rxe_resize_cq_resp __user *uresp);
+			struct rxe_resize_cq_resp __user *uresp,
+			struct ib_udata *udata);
 
 int rxe_cq_post(struct rxe_cq *cq, struct rxe_cqe *cqe, int solicited);
 
@@ -91,10 +92,8 @@ struct rxe_mmap_info {
 
 void rxe_mmap_release(struct kref *ref);
 
-struct rxe_mmap_info *rxe_create_mmap_info(struct rxe_dev *dev,
-					   u32 size,
-					   struct ib_ucontext *context,
-					   void *obj);
+struct rxe_mmap_info *rxe_create_mmap_info(struct rxe_dev *dev, u32 size,
+					   struct ib_udata *udata, void *obj);
 
 int rxe_mmap(struct ib_ucontext *context, struct vm_area_struct *vma);
 
@@ -224,13 +223,12 @@ int rxe_srq_chk_attr(struct rxe_dev *rxe, struct rxe_srq *srq,
 		     struct ib_srq_attr *attr, enum ib_srq_attr_mask mask);
 
 int rxe_srq_from_init(struct rxe_dev *rxe, struct rxe_srq *srq,
-		      struct ib_srq_init_attr *init,
-		      struct ib_ucontext *context,
+		      struct ib_srq_init_attr *init, struct ib_udata *udata,
 		      struct rxe_create_srq_resp __user *uresp);
 
 int rxe_srq_from_attr(struct rxe_dev *rxe, struct rxe_srq *srq,
 		      struct ib_srq_attr *attr, enum ib_srq_attr_mask mask,
-		      struct rxe_modify_srq_cmd *ucmd);
+		      struct rxe_modify_srq_cmd *ucmd, struct ib_udata *udata);
 
 void rxe_dealloc(struct ib_device *ib_dev);
 

drivers/infiniband/sw/rxe/rxe_mmap.c

@@ -36,6 +36,7 @@
 #include <linux/mm.h>
 #include <linux/errno.h>
 #include <asm/pgtable.h>
+#include <rdma/uverbs_ioctl.h>
 
 #include "rxe.h"
 #include "rxe_loc.h"
@@ -140,13 +141,14 @@ done:
 /*
  * Allocate information for rxe_mmap
  */
-struct rxe_mmap_info *rxe_create_mmap_info(struct rxe_dev *rxe,
-					   u32 size,
-					   struct ib_ucontext *context,
-					   void *obj)
+struct rxe_mmap_info *rxe_create_mmap_info(struct rxe_dev *rxe, u32 size,
+					   struct ib_udata *udata, void *obj)
 {
 	struct rxe_mmap_info *ip;
 
+	if (!udata)
+		return ERR_PTR(-EINVAL);
+
 	ip = kmalloc(sizeof(*ip), GFP_KERNEL);
 	if (!ip)
 		return NULL;
@@ -165,7 +167,9 @@ struct rxe_mmap_info *rxe_create_mmap_info(struct rxe_dev *rxe,
 
 	INIT_LIST_HEAD(&ip->pending_mmaps);
 	ip->info.size = size;
-	ip->context = context;
+	ip->context =
+		container_of(udata, struct uverbs_attr_bundle, driver_udata)
+			->context;
 	ip->obj = obj;
 	kref_init(&ip->ref);
 

drivers/infiniband/sw/rxe/rxe_mr.c

@@ -179,7 +179,7 @@ int rxe_mem_init_user(struct rxe_pd *pd, u64 start,
 	}
 
 	mem->umem = umem;
-	num_buf = umem->nmap;
+	num_buf = ib_umem_num_pages(umem);
 
 	rxe_mem_init(access, mem);
 
@@ -199,6 +199,12 @@ int rxe_mem_init_user(struct rxe_pd *pd, u64 start,
 		buf = map[0]->buf;
 
 		for_each_sg_page(umem->sg_head.sgl, &sg_iter, umem->nmap, 0) {
+			if (num_buf >= RXE_BUF_PER_MAP) {
+				map++;
+				buf = map[0]->buf;
+				num_buf = 0;
+			}
+
 			vaddr = page_address(sg_page_iter_page(&sg_iter));
 			if (!vaddr) {
 				pr_warn("null vaddr\n");
@@ -211,11 +217,6 @@ int rxe_mem_init_user(struct rxe_pd *pd, u64 start,
 			num_buf++;
 			buf++;
 
-			if (num_buf >= RXE_BUF_PER_MAP) {
-				map++;
-				buf = map[0]->buf;
-				num_buf = 0;
-			}
 		}
 	}
 

drivers/infiniband/sw/rxe/rxe_net.c

@@ -338,13 +338,13 @@ static void prepare_ipv6_hdr(struct dst_entry *dst, struct sk_buff *skb,
 	ip6h->payload_len = htons(skb->len - sizeof(*ip6h));
 }
 
-static int prepare4(struct rxe_pkt_info *pkt, struct sk_buff *skb,
-		    struct rxe_av *av)
+static int prepare4(struct rxe_pkt_info *pkt, struct sk_buff *skb)
 {
 	struct rxe_qp *qp = pkt->qp;
 	struct dst_entry *dst;
 	bool xnet = false;
 	__be16 df = htons(IP_DF);
+	struct rxe_av *av = rxe_get_av(pkt);
 	struct in_addr *saddr = &av->sgid_addr._sockaddr_in.sin_addr;
 	struct in_addr *daddr = &av->dgid_addr._sockaddr_in.sin_addr;
 
@@ -364,11 +364,11 @@ static int prepare4(struct rxe_pkt_info *pkt, struct sk_buff *skb,
 	return 0;
 }
 
-static int prepare6(struct rxe_pkt_info *pkt, struct sk_buff *skb,
-		    struct rxe_av *av)
+static int prepare6(struct rxe_pkt_info *pkt, struct sk_buff *skb)
 {
 	struct rxe_qp *qp = pkt->qp;
 	struct dst_entry *dst;
+	struct rxe_av *av = rxe_get_av(pkt);
 	struct in6_addr *saddr = &av->sgid_addr._sockaddr_in6.sin6_addr;
 	struct in6_addr *daddr = &av->dgid_addr._sockaddr_in6.sin6_addr;
 
@@ -392,16 +392,15 @@ static int prepare6(struct rxe_pkt_info *pkt, struct sk_buff *skb,
 int rxe_prepare(struct rxe_pkt_info *pkt, struct sk_buff *skb, u32 *crc)
 {
 	int err = 0;
-	struct rxe_av *av = rxe_get_av(pkt);
 
-	if (av->network_type == RDMA_NETWORK_IPV4)
-		err = prepare4(pkt, skb, av);
-	else if (av->network_type == RDMA_NETWORK_IPV6)
-		err = prepare6(pkt, skb, av);
+	if (skb->protocol == htons(ETH_P_IP))
+		err = prepare4(pkt, skb);
+	else if (skb->protocol == htons(ETH_P_IPV6))
+		err = prepare6(pkt, skb);
 
 	*crc = rxe_icrc_hdr(pkt, skb);
 
-	if (ether_addr_equal(skb->dev->dev_addr, av->dmac))
+	if (ether_addr_equal(skb->dev->dev_addr, rxe_get_av(pkt)->dmac))
 		pkt->mask |= RXE_LOOPBACK_MASK;
 
 	return err;
@@ -422,23 +421,20 @@ static void rxe_skb_tx_dtor(struct sk_buff *skb)
 
 int rxe_send(struct rxe_pkt_info *pkt, struct sk_buff *skb)
 {
-	struct rxe_av *av;
 	int err;
 
-	av = rxe_get_av(pkt);
-
 	skb->destructor = rxe_skb_tx_dtor;
 	skb->sk = pkt->qp->sk->sk;
 
 	rxe_add_ref(pkt->qp);
 	atomic_inc(&pkt->qp->skb_out);
 
-	if (av->network_type == RDMA_NETWORK_IPV4) {
+	if (skb->protocol == htons(ETH_P_IP)) {
 		err = ip_local_out(dev_net(skb_dst(skb)->dev), skb->sk, skb);
-	} else if (av->network_type == RDMA_NETWORK_IPV6) {
+	} else if (skb->protocol == htons(ETH_P_IPV6)) {
 		err = ip6_local_out(dev_net(skb_dst(skb)->dev), skb->sk, skb);
 	} else {
-		pr_err("Unknown layer 3 protocol: %d\n", av->network_type);
+		pr_err("Unknown layer 3 protocol: %d\n", skb->protocol);
 		atomic_dec(&pkt->qp->skb_out);
 		rxe_drop_ref(pkt->qp);
 		kfree_skb(skb);
@@ -462,7 +458,7 @@ struct sk_buff *rxe_init_packet(struct rxe_dev *rxe, struct rxe_av *av,
 				int paylen, struct rxe_pkt_info *pkt)
 {
 	unsigned int hdr_len;
-	struct sk_buff *skb;
+	struct sk_buff *skb = NULL;
 	struct net_device *ndev;
 	const struct ib_gid_attr *attr;
 	const int port_num = 1;
@@ -470,7 +466,6 @@ struct sk_buff *rxe_init_packet(struct rxe_dev *rxe, struct rxe_av *av,
 	attr = rdma_get_gid_attr(&rxe->ib_dev, port_num, av->grh.sgid_index);
 	if (IS_ERR(attr))
 		return NULL;
-	ndev = attr->ndev;
 
 	if (av->network_type == RDMA_NETWORK_IPV4)
 		hdr_len = ETH_HLEN + sizeof(struct udphdr) +
@@ -479,15 +474,26 @@ struct sk_buff *rxe_init_packet(struct rxe_dev *rxe, struct rxe_av *av,
 		hdr_len = ETH_HLEN + sizeof(struct udphdr) +
 			sizeof(struct ipv6hdr);
 
+	rcu_read_lock();
+	ndev = rdma_read_gid_attr_ndev_rcu(attr);
+	if (IS_ERR(ndev)) {
+		rcu_read_unlock();
+		goto out;
+	}
 	skb = alloc_skb(paylen + hdr_len + LL_RESERVED_SPACE(ndev),
 			GFP_ATOMIC);
 
-	if (unlikely(!skb))
+	if (unlikely(!skb)) {
+		rcu_read_unlock();
 		goto out;
+	}
 
-	skb_reserve(skb, hdr_len + LL_RESERVED_SPACE(rxe->ndev));
+	skb_reserve(skb, hdr_len + LL_RESERVED_SPACE(ndev));
 
+	/* FIXME: hold reference to this netdev until life of this skb. */
 	skb->dev	= ndev;
+	rcu_read_unlock();
+
 	if (av->network_type == RDMA_NETWORK_IPV4)
 		skb->protocol = htons(ETH_P_IP);
 	else

drivers/infiniband/sw/rxe/rxe_pool.c

@@ -52,12 +52,12 @@ struct rxe_type_info rxe_type_info[RXE_NUM_TYPES] = {
 	[RXE_TYPE_AH] = {
 		.name		= "rxe-ah",
 		.size		= sizeof(struct rxe_ah),
-		.flags		= RXE_POOL_ATOMIC,
+		.flags		= RXE_POOL_ATOMIC | RXE_POOL_NO_ALLOC,
 	},
 	[RXE_TYPE_SRQ] = {
 		.name		= "rxe-srq",
 		.size		= sizeof(struct rxe_srq),
-		.flags		= RXE_POOL_INDEX,
+		.flags		= RXE_POOL_INDEX | RXE_POOL_NO_ALLOC,
 		.min_index	= RXE_MIN_SRQ_INDEX,
 		.max_index	= RXE_MAX_SRQ_INDEX,
 	},

drivers/infiniband/sw/rxe/rxe_qp.c

@@ -217,8 +217,7 @@ static void rxe_qp_init_misc(struct rxe_dev *rxe, struct rxe_qp *qp,
 }
 
 static int rxe_qp_init_req(struct rxe_dev *rxe, struct rxe_qp *qp,
-			   struct ib_qp_init_attr *init,
-			   struct ib_ucontext *context,
+			   struct ib_qp_init_attr *init, struct ib_udata *udata,
 			   struct rxe_create_qp_resp __user *uresp)
 {
 	int err;
@@ -254,7 +253,7 @@ static int rxe_qp_init_req(struct rxe_dev *rxe, struct rxe_qp *qp,
 	if (!qp->sq.queue)
 		return -ENOMEM;
 
-	err = do_mmap_info(rxe, uresp ? &uresp->sq_mi : NULL, context,
+	err = do_mmap_info(rxe, uresp ? &uresp->sq_mi : NULL, udata,
 			   qp->sq.queue->buf, qp->sq.queue->buf_size,
 			   &qp->sq.queue->ip);
 
@@ -287,7 +286,7 @@ static int rxe_qp_init_req(struct rxe_dev *rxe, struct rxe_qp *qp,
 
 static int rxe_qp_init_resp(struct rxe_dev *rxe, struct rxe_qp *qp,
 			    struct ib_qp_init_attr *init,
-			    struct ib_ucontext *context,
+			    struct ib_udata *udata,
 			    struct rxe_create_qp_resp __user *uresp)
 {
 	int err;
@@ -308,7 +307,7 @@ static int rxe_qp_init_resp(struct rxe_dev *rxe, struct rxe_qp *qp,
 		if (!qp->rq.queue)
 			return -ENOMEM;
 
-		err = do_mmap_info(rxe, uresp ? &uresp->rq_mi : NULL, context,
+		err = do_mmap_info(rxe, uresp ? &uresp->rq_mi : NULL, udata,
 				   qp->rq.queue->buf, qp->rq.queue->buf_size,
 				   &qp->rq.queue->ip);
 		if (err) {
@@ -344,8 +343,6 @@ int rxe_qp_from_init(struct rxe_dev *rxe, struct rxe_qp *qp, struct rxe_pd *pd,
 	struct rxe_cq *rcq = to_rcq(init->recv_cq);
 	struct rxe_cq *scq = to_rcq(init->send_cq);
 	struct rxe_srq *srq = init->srq ? to_rsrq(init->srq) : NULL;
-	struct rxe_ucontext *ucontext =
-		rdma_udata_to_drv_context(udata, struct rxe_ucontext, ibuc);
 
 	rxe_add_ref(pd);
 	rxe_add_ref(rcq);
@@ -360,11 +357,11 @@ int rxe_qp_from_init(struct rxe_dev *rxe, struct rxe_qp *qp, struct rxe_pd *pd,
 
 	rxe_qp_init_misc(rxe, qp, init);
 
-	err = rxe_qp_init_req(rxe, qp, init, &ucontext->ibuc, uresp);
+	err = rxe_qp_init_req(rxe, qp, init, udata, uresp);
 	if (err)
 		goto err1;
 
-	err = rxe_qp_init_resp(rxe, qp, init, &ucontext->ibuc, uresp);
+	err = rxe_qp_init_resp(rxe, qp, init, udata, uresp);
 	if (err)
 		goto err2;
 

drivers/infiniband/sw/rxe/rxe_queue.c

@@ -36,18 +36,15 @@
 #include "rxe_loc.h"
 #include "rxe_queue.h"
 
-int do_mmap_info(struct rxe_dev *rxe,
-		 struct mminfo __user *outbuf,
-		 struct ib_ucontext *context,
-		 struct rxe_queue_buf *buf,
-		 size_t buf_size,
-		 struct rxe_mmap_info **ip_p)
+int do_mmap_info(struct rxe_dev *rxe, struct mminfo __user *outbuf,
+		 struct ib_udata *udata, struct rxe_queue_buf *buf,
+		 size_t buf_size, struct rxe_mmap_info **ip_p)
 {
 	int err;
 	struct rxe_mmap_info *ip = NULL;
 
 	if (outbuf) {
-		ip = rxe_create_mmap_info(rxe, buf_size, context, buf);
+		ip = rxe_create_mmap_info(rxe, buf_size, udata, buf);
 		if (!ip)
 			goto err1;
 
@@ -153,12 +150,9 @@ static int resize_finish(struct rxe_queue *q, struct rxe_queue *new_q,
 	return 0;
 }
 
-int rxe_queue_resize(struct rxe_queue *q,
-		     unsigned int *num_elem_p,
-		     unsigned int elem_size,
-		     struct ib_ucontext *context,
-		     struct mminfo __user *outbuf,
-		     spinlock_t *producer_lock,
+int rxe_queue_resize(struct rxe_queue *q, unsigned int *num_elem_p,
+		     unsigned int elem_size, struct ib_udata *udata,
+		     struct mminfo __user *outbuf, spinlock_t *producer_lock,
 		     spinlock_t *consumer_lock)
 {
 	struct rxe_queue *new_q;
@@ -170,7 +164,7 @@ int rxe_queue_resize(struct rxe_queue *q,
 	if (!new_q)
 		return -ENOMEM;
 
-	err = do_mmap_info(new_q->rxe, outbuf, context, new_q->buf,
+	err = do_mmap_info(new_q->rxe, outbuf, udata, new_q->buf,
 			   new_q->buf_size, &new_q->ip);
 	if (err) {
 		vfree(new_q->buf);

drivers/infiniband/sw/rxe/rxe_queue.h

@@ -76,12 +76,9 @@ struct rxe_queue {
 	unsigned int		index_mask;
 };
 
-int do_mmap_info(struct rxe_dev *rxe,
-		 struct mminfo __user *outbuf,
-		 struct ib_ucontext *context,
-		 struct rxe_queue_buf *buf,
-		 size_t buf_size,
-		 struct rxe_mmap_info **ip_p);
+int do_mmap_info(struct rxe_dev *rxe, struct mminfo __user *outbuf,
+		 struct ib_udata *udata, struct rxe_queue_buf *buf,
+		 size_t buf_size, struct rxe_mmap_info **ip_p);
 
 void rxe_queue_reset(struct rxe_queue *q);
 
@@ -89,10 +86,8 @@ struct rxe_queue *rxe_queue_init(struct rxe_dev *rxe,
 				 int *num_elem,
 				 unsigned int elem_size);
 
-int rxe_queue_resize(struct rxe_queue *q,
-		     unsigned int *num_elem_p,
-		     unsigned int elem_size,
-		     struct ib_ucontext *context,
+int rxe_queue_resize(struct rxe_queue *q, unsigned int *num_elem_p,
+		     unsigned int elem_size, struct ib_udata *udata,
 		     struct mminfo __user *outbuf,
 		     /* Protect producers while resizing queue */
 		     spinlock_t *producer_lock,

drivers/infiniband/sw/rxe/rxe_srq.c

@@ -99,8 +99,7 @@ err1:
 }
 
 int rxe_srq_from_init(struct rxe_dev *rxe, struct rxe_srq *srq,
-		      struct ib_srq_init_attr *init,
-		      struct ib_ucontext *context,
+		      struct ib_srq_init_attr *init, struct ib_udata *udata,
 		      struct rxe_create_srq_resp __user *uresp)
 {
 	int err;
@@ -128,7 +127,7 @@ int rxe_srq_from_init(struct rxe_dev *rxe, struct rxe_srq *srq,
 
 	srq->rq.queue = q;
 
-	err = do_mmap_info(rxe, uresp ? &uresp->mi : NULL, context, q->buf,
+	err = do_mmap_info(rxe, uresp ? &uresp->mi : NULL, udata, q->buf,
 			   q->buf_size, &q->ip);
 	if (err) {
 		vfree(q->buf);
@@ -149,7 +148,7 @@ int rxe_srq_from_init(struct rxe_dev *rxe, struct rxe_srq *srq,
 
 int rxe_srq_from_attr(struct rxe_dev *rxe, struct rxe_srq *srq,
 		      struct ib_srq_attr *attr, enum ib_srq_attr_mask mask,
-		      struct rxe_modify_srq_cmd *ucmd)
+		      struct rxe_modify_srq_cmd *ucmd, struct ib_udata *udata)
 {
 	int err;
 	struct rxe_queue *q = srq->rq.queue;
@@ -163,11 +162,8 @@ int rxe_srq_from_attr(struct rxe_dev *rxe, struct rxe_srq *srq,
 		mi = u64_to_user_ptr(ucmd->mmap_info_addr);
 
 		err = rxe_queue_resize(q, &attr->max_wr,
-				       rcv_wqe_size(srq->rq.max_sge),
-				       srq->rq.queue->ip ?
-						srq->rq.queue->ip->context :
-						NULL,
-				       mi, &srq->rq.producer_lock,
+				       rcv_wqe_size(srq->rq.max_sge), udata, mi,
+				       &srq->rq.producer_lock,
 				       &srq->rq.consumer_lock);
 		if (err)
 			goto err2;

drivers/infiniband/sw/rxe/rxe_verbs.c

@@ -176,8 +176,7 @@ static int rxe_port_immutable(struct ib_device *dev, u8 port_num,
 	return 0;
 }
 
-static int rxe_alloc_pd(struct ib_pd *ibpd, struct ib_ucontext *context,
-			struct ib_udata *udata)
+static int rxe_alloc_pd(struct ib_pd *ibpd, struct ib_udata *udata)
 {
 	struct rxe_dev *rxe = to_rdev(ibpd->device);
 	struct rxe_pd *pd = to_rpd(ibpd);
@@ -185,37 +184,31 @@ static int rxe_alloc_pd(struct ib_pd *ibpd, struct ib_ucontext *context,
 	return rxe_add_to_pool(&rxe->pd_pool, &pd->pelem);
 }
 
-static void rxe_dealloc_pd(struct ib_pd *ibpd)
+static void rxe_dealloc_pd(struct ib_pd *ibpd, struct ib_udata *udata)
 {
 	struct rxe_pd *pd = to_rpd(ibpd);
 
 	rxe_drop_ref(pd);
 }
 
-static struct ib_ah *rxe_create_ah(struct ib_pd *ibpd,
-				   struct rdma_ah_attr *attr,
-				   u32 flags,
-				   struct ib_udata *udata)
+static int rxe_create_ah(struct ib_ah *ibah, struct rdma_ah_attr *attr,
+			 u32 flags, struct ib_udata *udata)
 
 {
 	int err;
-	struct rxe_dev *rxe = to_rdev(ibpd->device);
-	struct rxe_pd *pd = to_rpd(ibpd);
-	struct rxe_ah *ah;
+	struct rxe_dev *rxe = to_rdev(ibah->device);
+	struct rxe_ah *ah = to_rah(ibah);
 
 	err = rxe_av_chk_attr(rxe, attr);
 	if (err)
-		return ERR_PTR(err);
+		return err;
 
-	ah = rxe_alloc(&rxe->ah_pool);
-	if (!ah)
-		return ERR_PTR(-ENOMEM);
-
-	rxe_add_ref(pd);
-	ah->pd = pd;
+	err = rxe_add_to_pool(&rxe->ah_pool, &ah->pelem);
+	if (err)
+		return err;
 
 	rxe_init_av(attr, &ah->av);
-	return &ah->ibah;
+	return 0;
 }
 
 static int rxe_modify_ah(struct ib_ah *ibah, struct rdma_ah_attr *attr)
@@ -242,13 +235,11 @@ static int rxe_query_ah(struct ib_ah *ibah, struct rdma_ah_attr *attr)
 	return 0;
 }
 
-static int rxe_destroy_ah(struct ib_ah *ibah, u32 flags)
+static void rxe_destroy_ah(struct ib_ah *ibah, u32 flags)
 {
 	struct rxe_ah *ah = to_rah(ibah);
 
-	rxe_drop_ref(ah->pd);
 	rxe_drop_ref(ah);
-	return 0;
 }
 
 static int post_one_recv(struct rxe_rq *rq, const struct ib_recv_wr *ibwr)
@@ -298,21 +289,18 @@ err1:
 	return err;
 }
 
-static struct ib_srq *rxe_create_srq(struct ib_pd *ibpd,
-				     struct ib_srq_init_attr *init,
-				     struct ib_udata *udata)
+static int rxe_create_srq(struct ib_srq *ibsrq, struct ib_srq_init_attr *init,
+			  struct ib_udata *udata)
 {
 	int err;
-	struct rxe_dev *rxe = to_rdev(ibpd->device);
-	struct rxe_pd *pd = to_rpd(ibpd);
-	struct rxe_ucontext *ucontext =
-		rdma_udata_to_drv_context(udata, struct rxe_ucontext, ibuc);
-	struct rxe_srq *srq;
+	struct rxe_dev *rxe = to_rdev(ibsrq->device);
+	struct rxe_pd *pd = to_rpd(ibsrq->pd);
+	struct rxe_srq *srq = to_rsrq(ibsrq);
 	struct rxe_create_srq_resp __user *uresp = NULL;
 
 	if (udata) {
 		if (udata->outlen < sizeof(*uresp))
-			return ERR_PTR(-EINVAL);
+			return -EINVAL;
 		uresp = udata->outbuf;
 	}
 
@@ -320,28 +308,24 @@ static struct ib_srq *rxe_create_srq(struct ib_pd *ibpd,
 	if (err)
 		goto err1;
 
-	srq = rxe_alloc(&rxe->srq_pool);
-	if (!srq) {
-		err = -ENOMEM;
+	err = rxe_add_to_pool(&rxe->srq_pool, &srq->pelem);
+	if (err)
 		goto err1;
-	}
 
-	rxe_add_index(srq);
 	rxe_add_ref(pd);
 	srq->pd = pd;
 
-	err = rxe_srq_from_init(rxe, srq, init, &ucontext->ibuc, uresp);
+	err = rxe_srq_from_init(rxe, srq, init, udata, uresp);
 	if (err)
 		goto err2;
 
-	return &srq->ibsrq;
+	return 0;
 
 err2:
 	rxe_drop_ref(pd);
-	rxe_drop_index(srq);
 	rxe_drop_ref(srq);
 err1:
-	return ERR_PTR(err);
+	return err;
 }
 
 static int rxe_modify_srq(struct ib_srq *ibsrq, struct ib_srq_attr *attr,
@@ -366,7 +350,7 @@ static int rxe_modify_srq(struct ib_srq *ibsrq, struct ib_srq_attr *attr,
 	if (err)
 		goto err1;
 
-	err = rxe_srq_from_attr(rxe, srq, attr, mask, &ucmd);
+	err = rxe_srq_from_attr(rxe, srq, attr, mask, &ucmd, udata);
 	if (err)
 		goto err1;
 
@@ -389,7 +373,7 @@ static int rxe_query_srq(struct ib_srq *ibsrq, struct ib_srq_attr *attr)
 	return 0;
 }
 
-static int rxe_destroy_srq(struct ib_srq *ibsrq)
+static void rxe_destroy_srq(struct ib_srq *ibsrq, struct ib_udata *udata)
 {
 	struct rxe_srq *srq = to_rsrq(ibsrq);
 
@@ -397,10 +381,7 @@ static int rxe_destroy_srq(struct ib_srq *ibsrq)
 		rxe_queue_cleanup(srq->rq.queue);
 
 	rxe_drop_ref(srq->pd);
-	rxe_drop_index(srq);
 	rxe_drop_ref(srq);
-
-	return 0;
 }
 
 static int rxe_post_srq_recv(struct ib_srq *ibsrq, const struct ib_recv_wr *wr,
@@ -509,7 +490,7 @@ static int rxe_query_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
 	return 0;
 }
 
-static int rxe_destroy_qp(struct ib_qp *ibqp)
+static int rxe_destroy_qp(struct ib_qp *ibqp, struct ib_udata *udata)
 {
 	struct rxe_qp *qp = to_rqp(ibqp);
 
@@ -799,7 +780,6 @@ err1:
 
 static struct ib_cq *rxe_create_cq(struct ib_device *dev,
 				   const struct ib_cq_init_attr *attr,
-				   struct ib_ucontext *context,
 				   struct ib_udata *udata)
 {
 	int err;
@@ -826,8 +806,8 @@ static struct ib_cq *rxe_create_cq(struct ib_device *dev,
 		goto err1;
 	}
 
-	err = rxe_cq_from_init(rxe, cq, attr->cqe, attr->comp_vector,
-			       context, uresp);
+	err = rxe_cq_from_init(rxe, cq, attr->cqe, attr->comp_vector, udata,
+			       uresp);
 	if (err)
 		goto err2;
 
@@ -839,7 +819,7 @@ err1:
 	return ERR_PTR(err);
 }
 
-static int rxe_destroy_cq(struct ib_cq *ibcq)
+static int rxe_destroy_cq(struct ib_cq *ibcq, struct ib_udata *udata)
 {
 	struct rxe_cq *cq = to_rcq(ibcq);
 
@@ -866,7 +846,7 @@ static int rxe_resize_cq(struct ib_cq *ibcq, int cqe, struct ib_udata *udata)
 	if (err)
 		goto err1;
 
-	err = rxe_cq_resize_queue(cq, cqe, uresp);
+	err = rxe_cq_resize_queue(cq, cqe, uresp, udata);
 	if (err)
 		goto err1;
 
@@ -990,7 +970,7 @@ err2:
 	return ERR_PTR(err);
 }
 
-static int rxe_dereg_mr(struct ib_mr *ibmr)
+static int rxe_dereg_mr(struct ib_mr *ibmr, struct ib_udata *udata)
 {
 	struct rxe_mem *mr = to_rmr(ibmr);
 
@@ -1001,9 +981,8 @@ static int rxe_dereg_mr(struct ib_mr *ibmr)
 	return 0;
 }
 
-static struct ib_mr *rxe_alloc_mr(struct ib_pd *ibpd,
-				  enum ib_mr_type mr_type,
-				  u32 max_num_sg)
+static struct ib_mr *rxe_alloc_mr(struct ib_pd *ibpd, enum ib_mr_type mr_type,
+				  u32 max_num_sg, struct ib_udata *udata)
 {
 	struct rxe_dev *rxe = to_rdev(ibpd->device);
 	struct rxe_pd *pd = to_rpd(ibpd);
@@ -1176,7 +1155,10 @@ static const struct ib_device_ops rxe_dev_ops = {
 	.reg_user_mr = rxe_reg_user_mr,
 	.req_notify_cq = rxe_req_notify_cq,
 	.resize_cq = rxe_resize_cq,
+
+	INIT_RDMA_OBJ_SIZE(ib_ah, rxe_ah, ibah),
 	INIT_RDMA_OBJ_SIZE(ib_pd, rxe_pd, ibpd),
+	INIT_RDMA_OBJ_SIZE(ib_srq, rxe_srq, ibsrq),
 	INIT_RDMA_OBJ_SIZE(ib_ucontext, rxe_ucontext, ibuc),
 };
 

drivers/infiniband/sw/rxe/rxe_verbs.h

@@ -71,8 +71,8 @@ struct rxe_pd {
 };
 
 struct rxe_ah {
-	struct rxe_pool_entry	pelem;
 	struct ib_ah		ibah;
+	struct rxe_pool_entry	pelem;
 	struct rxe_pd		*pd;
 	struct rxe_av		av;
 };
@@ -120,8 +120,8 @@ struct rxe_rq {
 };
 
 struct rxe_srq {
-	struct rxe_pool_entry	pelem;
 	struct ib_srq		ibsrq;
+	struct rxe_pool_entry	pelem;
 	struct rxe_pd		*pd;
 	struct rxe_rq		rq;
 	u32			srq_num;