scsi: hisi_sas: relocate smp sg map

Currently we use DQ lock to protect delivery of DQ entry one by one.

To optimise to allow more than one slot to be built for a single DQ in
parallel, we need to remove the DQ lock when preparing slots, prior to
delivery.

To achieve this, we rearrange the slot build order to ensure that once
we allocate a slot for a task, we do cannot fail to deliver the task.

In this patch, we rearrange the slot building for SMP tasks to ensure
that sg mapping part (which can fail) happens before we allocate the
slot in the DQ.

Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

4 files changed, 51 insertions, 106 deletions

diff --git a/drivers/scsi/hisi_sas/hisi_sas_main.c b/drivers/scsi/hisi_sas/hisi_sas_main.c
index ff5b8d7..0ce7c71 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_main.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_main.c
@@ -319,7 +319,8 @@ static int hisi_sas_task_prep(struct sas_task *task, struct hisi_sas_dq
 	struct hisi_sas_cmd_hdr	*cmd_hdr_base;
 	struct asd_sas_port *sas_port = device->port;
 	struct device *dev = hisi_hba->dev;
-	int dlvry_queue_slot, dlvry_queue, n_elem = 0, rc, slot_idx;
+	int dlvry_queue_slot, dlvry_queue, rc, slot_idx;
+	int  n_elem = 0, n_elem_req = 0, n_elem_resp = 0;
 	unsigned long flags;
 
 	if (!sas_port) {
@@ -358,6 +359,8 @@ static int hisi_sas_task_prep(struct sas_task *task, struct hisi_sas_dq
 	}
 
 	if (!sas_protocol_ata(task->task_proto)) {
+		unsigned int req_len, resp_len;
+
 		if (task->num_scatter) {
 			n_elem = dma_map_sg(dev, task->scatter,
 					    task->num_scatter, task->data_dir);
@@ -365,6 +368,29 @@ static int hisi_sas_task_prep(struct sas_task *task, struct hisi_sas_dq
 				rc = -ENOMEM;
 				goto prep_out;
 			}
+		} else if (task->task_proto & SAS_PROTOCOL_SMP) {
+			n_elem_req = dma_map_sg(dev, &task->smp_task.smp_req,
+						1, DMA_TO_DEVICE);
+			if (!n_elem_req) {
+				rc = -ENOMEM;
+				goto prep_out;
+			}
+			req_len = sg_dma_len(&task->smp_task.smp_req);
+			if (req_len & 0x3) {
+				rc = -EINVAL;
+				goto err_out_dma_unmap;
+			}
+			n_elem_resp = dma_map_sg(dev, &task->smp_task.smp_resp,
+						 1, DMA_FROM_DEVICE);
+			if (!n_elem_req) {
+				rc = -ENOMEM;
+				goto err_out_dma_unmap;
+			}
+			resp_len = sg_dma_len(&task->smp_task.smp_resp);
+			if (resp_len & 0x3) {
+				rc = -EINVAL;
+				goto err_out_dma_unmap;
+			}
 		}
 	} else
 		n_elem = task->num_scatter;
@@ -375,11 +401,9 @@ static int hisi_sas_task_prep(struct sas_task *task, struct hisi_sas_dq
 						    device);
 	else
 		rc = hisi_sas_slot_index_alloc(hisi_hba, &slot_idx);
-	if (rc) {
-		spin_unlock_irqrestore(&hisi_hba->lock, flags);
-		goto err_out;
-	}
 	spin_unlock_irqrestore(&hisi_hba->lock, flags);
+	if (rc)
+		goto err_out_dma_unmap;
 
 	rc = hisi_hba->hw->get_free_slot(hisi_hba, dq);
 	if (rc)
@@ -458,14 +482,22 @@ err_out_tag:
 	spin_lock_irqsave(&hisi_hba->lock, flags);
 	hisi_sas_slot_index_free(hisi_hba, slot_idx);
 	spin_unlock_irqrestore(&hisi_hba->lock, flags);
-err_out:
-	dev_err(dev, "task prep: failed[%d]!\n", rc);
-	if (!sas_protocol_ata(task->task_proto))
-		if (n_elem)
-			dma_unmap_sg(dev, task->scatter,
-				     task->num_scatter,
-				     task->data_dir);
+err_out_dma_unmap:
+	if (!sas_protocol_ata(task->task_proto)) {
+		if (task->num_scatter) {
+			dma_unmap_sg(dev, task->scatter, task->num_scatter,
+			     task->data_dir);
+		} else if (task->task_proto & SAS_PROTOCOL_SMP) {
+			if (n_elem_req)
+				dma_unmap_sg(dev, &task->smp_task.smp_req,
+					     1, DMA_TO_DEVICE);
+			if (n_elem_resp)
+				dma_unmap_sg(dev, &task->smp_task.smp_resp,
+					     1, DMA_FROM_DEVICE);
+		}
+	}
 prep_out:
+	dev_err(dev, "task prep: failed[%d]!\n", rc);
 	return rc;
 }
 
diff --git a/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c
index 84a0ccc..3769d70 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c
@@ -974,38 +974,17 @@ static int prep_smp_v1_hw(struct hisi_hba *hisi_hba,
 	struct sas_task *task = slot->task;
 	struct hisi_sas_cmd_hdr *hdr = slot->cmd_hdr;
 	struct domain_device *device = task->dev;
-	struct device *dev = hisi_hba->dev;
 	struct hisi_sas_port *port = slot->port;
-	struct scatterlist *sg_req, *sg_resp;
+	struct scatterlist *sg_req;
 	struct hisi_sas_device *sas_dev = device->lldd_dev;
 	dma_addr_t req_dma_addr;
-	unsigned int req_len, resp_len;
-	int elem, rc;
+	unsigned int req_len;
 
-	/*
-	* DMA-map SMP request, response buffers
-	*/
 	/* req */
 	sg_req = &task->smp_task.smp_req;
-	elem = dma_map_sg(dev, sg_req, 1, DMA_TO_DEVICE);
-	if (!elem)
-		return -ENOMEM;
 	req_len = sg_dma_len(sg_req);
 	req_dma_addr = sg_dma_address(sg_req);
 
-	/* resp */
-	sg_resp = &task->smp_task.smp_resp;
-	elem = dma_map_sg(dev, sg_resp, 1, DMA_FROM_DEVICE);
-	if (!elem) {
-		rc = -ENOMEM;
-		goto err_out_req;
-	}
-	resp_len = sg_dma_len(sg_resp);
-	if ((req_len & 0x3) || (resp_len & 0x3)) {
-		rc = -EINVAL;
-		goto err_out_resp;
-	}
-
 	/* create header */
 	/* dw0 */
 	hdr->dw0 = cpu_to_le32((port->id << CMD_HDR_PORT_OFF) |
@@ -1027,14 +1006,6 @@ static int prep_smp_v1_hw(struct hisi_hba *hisi_hba,
 	hdr->sts_buffer_addr = cpu_to_le64(hisi_sas_status_buf_addr_dma(slot));
 
 	return 0;
-
-err_out_resp:
-	dma_unmap_sg(dev, &slot->task->smp_task.smp_resp, 1,
-		     DMA_FROM_DEVICE);
-err_out_req:
-	dma_unmap_sg(dev, &slot->task->smp_task.smp_req, 1,
-		     DMA_TO_DEVICE);
-	return rc;
 }
 
 static int prep_ssp_v1_hw(struct hisi_hba *hisi_hba,
diff --git a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c
index 9e68731..ac0a0b2 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c
@@ -1721,37 +1721,16 @@ static int prep_smp_v2_hw(struct hisi_hba *hisi_hba,
 	struct sas_task *task = slot->task;
 	struct hisi_sas_cmd_hdr *hdr = slot->cmd_hdr;
 	struct domain_device *device = task->dev;
-	struct device *dev = hisi_hba->dev;
 	struct hisi_sas_port *port = slot->port;
-	struct scatterlist *sg_req, *sg_resp;
+	struct scatterlist *sg_req;
 	struct hisi_sas_device *sas_dev = device->lldd_dev;
 	dma_addr_t req_dma_addr;
-	unsigned int req_len, resp_len;
-	int elem, rc;
+	unsigned int req_len;
 
-	/*
-	* DMA-map SMP request, response buffers
-	*/
 	/* req */
 	sg_req = &task->smp_task.smp_req;
-	elem = dma_map_sg(dev, sg_req, 1, DMA_TO_DEVICE);
-	if (!elem)
-		return -ENOMEM;
-	req_len = sg_dma_len(sg_req);
 	req_dma_addr = sg_dma_address(sg_req);
-
-	/* resp */
-	sg_resp = &task->smp_task.smp_resp;
-	elem = dma_map_sg(dev, sg_resp, 1, DMA_FROM_DEVICE);
-	if (!elem) {
-		rc = -ENOMEM;
-		goto err_out_req;
-	}
-	resp_len = sg_dma_len(sg_resp);
-	if ((req_len & 0x3) || (resp_len & 0x3)) {
-		rc = -EINVAL;
-		goto err_out_resp;
-	}
+	req_len = sg_dma_len(&task->smp_task.smp_req);
 
 	/* create header */
 	/* dw0 */
@@ -1775,14 +1754,6 @@ static int prep_smp_v2_hw(struct hisi_hba *hisi_hba,
 	hdr->sts_buffer_addr = cpu_to_le64(hisi_sas_status_buf_addr_dma(slot));
 
 	return 0;
-
-err_out_resp:
-	dma_unmap_sg(dev, &slot->task->smp_task.smp_resp, 1,
-		     DMA_FROM_DEVICE);
-err_out_req:
-	dma_unmap_sg(dev, &slot->task->smp_task.smp_req, 1,
-		     DMA_TO_DEVICE);
-	return rc;
 }
 
 static int prep_ssp_v2_hw(struct hisi_hba *hisi_hba,
diff --git a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
index 492c3be..cea5354 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
@@ -986,38 +986,17 @@ static int prep_smp_v3_hw(struct hisi_hba *hisi_hba,
 	struct sas_task *task = slot->task;
 	struct hisi_sas_cmd_hdr *hdr = slot->cmd_hdr;
 	struct domain_device *device = task->dev;
-	struct device *dev = hisi_hba->dev;
 	struct hisi_sas_port *port = slot->port;
-	struct scatterlist *sg_req, *sg_resp;
+	struct scatterlist *sg_req;
 	struct hisi_sas_device *sas_dev = device->lldd_dev;
 	dma_addr_t req_dma_addr;
-	unsigned int req_len, resp_len;
-	int elem, rc;
+	unsigned int req_len;
 
-	/*
-	 * DMA-map SMP request, response buffers
-	 */
 	/* req */
 	sg_req = &task->smp_task.smp_req;
-	elem = dma_map_sg(dev, sg_req, 1, DMA_TO_DEVICE);
-	if (!elem)
-		return -ENOMEM;
 	req_len = sg_dma_len(sg_req);
 	req_dma_addr = sg_dma_address(sg_req);
 
-	/* resp */
-	sg_resp = &task->smp_task.smp_resp;
-	elem = dma_map_sg(dev, sg_resp, 1, DMA_FROM_DEVICE);
-	if (!elem) {
-		rc = -ENOMEM;
-		goto err_out_req;
-	}
-	resp_len = sg_dma_len(sg_resp);
-	if ((req_len & 0x3) || (resp_len & 0x3)) {
-		rc = -EINVAL;
-		goto err_out_resp;
-	}
-
 	/* create header */
 	/* dw0 */
 	hdr->dw0 = cpu_to_le32((port->id << CMD_HDR_PORT_OFF) |
@@ -1040,14 +1019,6 @@ static int prep_smp_v3_hw(struct hisi_hba *hisi_hba,
 	hdr->sts_buffer_addr = cpu_to_le64(hisi_sas_status_buf_addr_dma(slot));
 
 	return 0;
-
-err_out_resp:
-	dma_unmap_sg(dev, &slot->task->smp_task.smp_resp, 1,
-		     DMA_FROM_DEVICE);
-err_out_req:
-	dma_unmap_sg(dev, &slot->task->smp_task.smp_req, 1,
-		     DMA_TO_DEVICE);
-	return rc;
 }
 
 static int prep_ata_v3_hw(struct hisi_hba *hisi_hba,