Backport DMABUF support from Upstream

drivers/infiniband/core/umem_dmabuf.c

@@ -163,12 +163,63 @@ struct ib_umem_dmabuf *ib_umem_dmabuf_get(struct ib_device *device,
 }
 EXPORT_SYMBOL(ib_umem_dmabuf_get);
 
+static void
+ib_umem_dmabuf_unsupported_move_notify(struct dma_buf_attachment *attach)
+{
+	struct ib_umem_dmabuf *umem_dmabuf = attach->importer_priv;
+
+	ibdev_warn_ratelimited(umem_dmabuf->umem.ibdev,
+			       "Invalidate callback should not be called when memory is pinned\n");
+}
+
+static struct dma_buf_attach_ops ib_umem_dmabuf_attach_pinned_ops = {
+	.allow_peer2peer = true,
+	.move_notify = ib_umem_dmabuf_unsupported_move_notify,
+};
+
+struct ib_umem_dmabuf *ib_umem_dmabuf_get_pinned(struct ib_device *device,
+						 unsigned long offset,
+						 size_t size, int fd,
+						 int access)
+{
+	struct ib_umem_dmabuf *umem_dmabuf;
+	int err;
+
+	umem_dmabuf = ib_umem_dmabuf_get(device, offset, size, fd, access,
+					 &ib_umem_dmabuf_attach_pinned_ops);
+	if (IS_ERR(umem_dmabuf))
+		return umem_dmabuf;
+
+	dma_resv_lock(umem_dmabuf->attach->dmabuf->resv, NULL);
+	err = dma_buf_pin(umem_dmabuf->attach);
+	if (err)
+		goto err_release;
+	umem_dmabuf->pinned = 1;
+
+	err = ib_umem_dmabuf_map_pages(umem_dmabuf);
+	if (err)
+		goto err_unpin;
+	dma_resv_unlock(umem_dmabuf->attach->dmabuf->resv);
+
+	return umem_dmabuf;
+
+err_unpin:
+	dma_buf_unpin(umem_dmabuf->attach);
+err_release:
+	dma_resv_unlock(umem_dmabuf->attach->dmabuf->resv);
+	ib_umem_release(&umem_dmabuf->umem);
+	return ERR_PTR(err);
+}
+EXPORT_SYMBOL(ib_umem_dmabuf_get_pinned);
+
 void ib_umem_dmabuf_release(struct ib_umem_dmabuf *umem_dmabuf)
 {
 	struct dma_buf *dmabuf = umem_dmabuf->attach->dmabuf;
 
 	dma_resv_lock(dmabuf->resv, NULL);
 	ib_umem_dmabuf_unmap_pages(umem_dmabuf);
+	if (umem_dmabuf->pinned)
+		dma_buf_unpin(umem_dmabuf->attach);
 	dma_resv_unlock(dmabuf->resv);
 
 	dma_buf_detach(dmabuf, umem_dmabuf->attach);

drivers/infiniband/core/uverbs_ioctl.c

@@ -337,6 +337,14 @@ static int uverbs_process_attr(struct bundle_priv *pbundle,
 
 		break;
 
+	case UVERBS_ATTR_TYPE_RAW_FD:
+		if (uattr->attr_data.reserved || uattr->len != 0 ||
+		    uattr->data_s64 < INT_MIN || uattr->data_s64 > INT_MAX)
+			return -EINVAL;
+		/* _uverbs_get_const_signed() is the accessor */
+		e->ptr_attr.data = uattr->data_s64;
+		break;
+
 	case UVERBS_ATTR_TYPE_IDRS_ARRAY:
 		return uverbs_process_idrs_array(pbundle, attr_uapi,
 						 &e->objs_arr_attr, uattr,

drivers/infiniband/hw/efa/efa.h

@@ -139,6 +139,10 @@ int efa_create_cq(struct ib_cq *ibcq, const struct ib_cq_init_attr *attr,
 struct ib_mr *efa_reg_mr(struct ib_pd *ibpd, u64 start, u64 length,
 			 u64 virt_addr, int access_flags,
 			 struct ib_udata *udata);
+struct ib_mr *efa_reg_user_mr_dmabuf(struct ib_pd *ibpd, u64 start,
+				     u64 length, u64 virt_addr,
+				     int fd, int access_flags,
+				     struct ib_udata *udata);
 int efa_dereg_mr(struct ib_mr *ibmr, struct ib_udata *udata);
 int efa_get_port_immutable(struct ib_device *ibdev, u32 port_num,
 			   struct ib_port_immutable *immutable);

drivers/infiniband/hw/efa/efa_main.c

@@ -266,6 +266,7 @@ static const struct ib_device_ops efa_dev_ops = {
 	.query_port = efa_query_port,
 	.query_qp = efa_query_qp,
 	.reg_user_mr = efa_reg_mr,
+	.reg_user_mr_dmabuf = efa_reg_user_mr_dmabuf,
 
 	INIT_RDMA_OBJ_SIZE(ib_ah, efa_ah, ibah),
 	INIT_RDMA_OBJ_SIZE(ib_cq, efa_cq, ibcq),

drivers/infiniband/hw/efa/efa_verbs.c

@@ -3,6 +3,8 @@
  * Copyright 2018-2021 Amazon.com, Inc. or its affiliates. All rights reserved.
  */
 
+#include <linux/dma-buf.h>
+#include <linux/dma-resv.h>
 #include <linux/vmalloc.h>
 #include <linux/log2.h>
 
@@ -1490,26 +1492,18 @@ static int efa_create_pbl(struct efa_dev *dev,
 	return 0;
 }
 
-struct ib_mr *efa_reg_mr(struct ib_pd *ibpd, u64 start, u64 length,
-			 u64 virt_addr, int access_flags,
-			 struct ib_udata *udata)
+static struct efa_mr *efa_alloc_mr(struct ib_pd *ibpd, int access_flags,
+				   struct ib_udata *udata)
 {
 	struct efa_dev *dev = to_edev(ibpd->device);
-	struct efa_com_reg_mr_params params = {};
-	struct efa_com_reg_mr_result result = {};
-	struct pbl_context pbl;
 	int supp_access_flags;
-	unsigned int pg_sz;
 	struct efa_mr *mr;
-	int inline_size;
-	int err;
 
 	if (udata && udata->inlen &&
 	    !ib_is_udata_cleared(udata, 0, sizeof(udata->inlen))) {
 		ibdev_dbg(&dev->ibdev,
 			  "Incompatible ABI params, udata not cleared\n");
-		err = -EINVAL;
-		goto err_out;
+		return ERR_PTR(-EINVAL);
 	}
 
 	supp_access_flags =
@@ -1521,23 +1515,26 @@ struct ib_mr *efa_reg_mr(struct ib_pd *ibpd, u64 start, u64 length,
 		ibdev_dbg(&dev->ibdev,
 			  "Unsupported access flags[%#x], supported[%#x]\n",
 			  access_flags, supp_access_flags);
-		err = -EOPNOTSUPP;
-		goto err_out;
+		return ERR_PTR(-EOPNOTSUPP);
 	}
 
 	mr = kzalloc(sizeof(*mr), GFP_KERNEL);
-	if (!mr) {
-		err = -ENOMEM;
-		goto err_out;
-	}
+	if (!mr)
+		return ERR_PTR(-ENOMEM);
 
-	mr->umem = ib_umem_get(ibpd->device, start, length, access_flags);
-	if (IS_ERR(mr->umem)) {
-		err = PTR_ERR(mr->umem);
-		ibdev_dbg(&dev->ibdev,
-			  "Failed to pin and map user space memory[%d]\n", err);
-		goto err_free;
-	}
+	return mr;
+}
+
+static int efa_register_mr(struct ib_pd *ibpd, struct efa_mr *mr, u64 start,
+			   u64 length, u64 virt_addr, int access_flags)
+{
+	struct efa_dev *dev = to_edev(ibpd->device);
+	struct efa_com_reg_mr_params params = {};
+	struct efa_com_reg_mr_result result = {};
+	struct pbl_context pbl;
+	unsigned int pg_sz;
+	int inline_size;
+	int err;
 
 	params.pd = to_epd(ibpd)->pdn;
 	params.iova = virt_addr;
@@ -1548,10 +1545,9 @@ struct ib_mr *efa_reg_mr(struct ib_pd *ibpd, u64 start, u64 length,
 				       dev->dev_attr.page_size_cap,
 				       virt_addr);
 	if (!pg_sz) {
-		err = -EOPNOTSUPP;
 		ibdev_dbg(&dev->ibdev, "Failed to find a suitable page size in page_size_cap %#llx\n",
 			  dev->dev_attr.page_size_cap);
-		goto err_unmap;
+		return -EOPNOTSUPP;
 	}
 
 	params.page_shift = order_base_2(pg_sz);
@@ -1565,31 +1561,100 @@ struct ib_mr *efa_reg_mr(struct ib_pd *ibpd, u64 start, u64 length,
 	if (params.page_num <= inline_size) {
 		err = efa_create_inline_pbl(dev, mr, &params);
 		if (err)
-			goto err_unmap;
+			return err;
 
 		err = efa_com_register_mr(&dev->edev, &params, &result);
 		if (err)
-			goto err_unmap;
+			return err;
 	} else {
 		err = efa_create_pbl(dev, &pbl, mr, &params);
 		if (err)
-			goto err_unmap;
+			return err;
 
 		err = efa_com_register_mr(&dev->edev, &params, &result);
 		pbl_destroy(dev, &pbl);
 
 		if (err)
-			goto err_unmap;
+			return err;
 	}
 
 	mr->ibmr.lkey = result.l_key;
 	mr->ibmr.rkey = result.r_key;
 	mr->ibmr.length = length;
 	ibdev_dbg(&dev->ibdev, "Registered mr[%d]\n", mr->ibmr.lkey);
 
+	return 0;
+}
+
+struct ib_mr *efa_reg_user_mr_dmabuf(struct ib_pd *ibpd, u64 start,
+				     u64 length, u64 virt_addr,
+				     int fd, int access_flags,
+				     struct ib_udata *udata)
+{
+	struct efa_dev *dev = to_edev(ibpd->device);
+	struct ib_umem_dmabuf *umem_dmabuf;
+	struct efa_mr *mr;
+	int err;
+
+	mr = efa_alloc_mr(ibpd, access_flags, udata);
+	if (IS_ERR(mr)) {
+		err = PTR_ERR(mr);
+		goto err_out;
+	}
+
+	umem_dmabuf = ib_umem_dmabuf_get_pinned(ibpd->device, start, length, fd,
+						access_flags);
+	if (IS_ERR(umem_dmabuf)) {
+		err = PTR_ERR(umem_dmabuf);
+		ibdev_dbg(&dev->ibdev, "Failed to get dmabuf umem[%d]\n", err);
+		goto err_free;
+	}
+
+	mr->umem = &umem_dmabuf->umem;
+	err = efa_register_mr(ibpd, mr, start, length, virt_addr, access_flags);
+	if (err)
+		goto err_release;
+
+	return &mr->ibmr;
+
+err_release:
+	ib_umem_release(mr->umem);
+err_free:
+	kfree(mr);
+err_out:
+	atomic64_inc(&dev->stats.reg_mr_err);
+	return ERR_PTR(err);
+}
+
+struct ib_mr *efa_reg_mr(struct ib_pd *ibpd, u64 start, u64 length,
+			 u64 virt_addr, int access_flags,
+			 struct ib_udata *udata)
+{
+	struct efa_dev *dev = to_edev(ibpd->device);
+	struct efa_mr *mr;
+	int err;
+
+	mr = efa_alloc_mr(ibpd, access_flags, udata);
+	if (IS_ERR(mr)) {
+		err = PTR_ERR(mr);
+		goto err_out;
+	}
+
+	mr->umem = ib_umem_get(ibpd->device, start, length, access_flags);
+	if (IS_ERR(mr->umem)) {
+		err = PTR_ERR(mr->umem);
+		ibdev_dbg(&dev->ibdev,
+			  "Failed to pin and map user space memory[%d]\n", err);
+		goto err_free;
+	}
+
+	err = efa_register_mr(ibpd, mr, start, length, virt_addr, access_flags);
+	if (err)
+		goto err_release;
+
 	return &mr->ibmr;
 
-err_unmap:
+err_release:
 	ib_umem_release(mr->umem);
 err_free:
 	kfree(mr);