scsi: mpi3mr: Add io_uring interface support in I/O-polled mode

Add support for the io_uring interface in I/O-polled mode.

This feature is disabled in the driver by default. To enable the feature, a
module parameter "poll_queues" has to be set with the desired number of
polling queues.

When the feature is enabled, the driver reserves a certain number of
operational queue pairs for the poll_queues either from the available queue
pairs or creates additional queue pairs based on the operational queue
availability.

The Polling queues will have corresponding IRQ and ISR functions as similar
to default queues. However, the IRQ line is disabled by the driver for
poll_queues.

Link: https://lore.kernel.org/r/20211220141159.16117-22-sreekanth.reddy@broadcom.com
Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

3 files changed, 210 insertions, 38 deletions

diff --git a/drivers/scsi/mpi3mr/mpi3mr.h b/drivers/scsi/mpi3mr/mpi3mr.h
index 8dd669fa5efb..64783a8e05aa 100644
--- a/drivers/scsi/mpi3mr/mpi3mr.h
+++ b/drivers/scsi/mpi3mr/mpi3mr.h
@@ -220,6 +220,12 @@ enum mpi3mr_reset_reason {
 	MPI3MR_RESET_FROM_FIRMWARE = 27,
 };
 
+/* Queue type definitions */
+enum queue_type {
+	MPI3MR_DEFAULT_QUEUE = 0,
+	MPI3MR_POLL_QUEUE,
+};
+
 /**
  * struct mpi3mr_compimg_ver - replica of component image
  * version defined in mpi30_image.h in host endianness
@@ -331,6 +337,7 @@ struct op_req_qinfo {
  * @pend_ios: Number of IOs pending in HW for this queue
  * @enable_irq_poll: Flag to indicate polling is enabled
  * @in_use: Queue is handled by poll/ISR
+ * @qtype: Type of queue (types defined in enum queue_type)
  */
 struct op_reply_qinfo {
 	u16 ci;
@@ -345,6 +352,7 @@ struct op_reply_qinfo {
 	atomic_t pend_ios;
 	bool enable_irq_poll;
 	atomic_t in_use;
+	enum queue_type qtype;
 };
 
 /**
@@ -703,6 +711,9 @@ struct scmd_priv {
  * @driver_info: Driver, Kernel, OS information to firmware
  * @change_count: Topology change count
  * @op_reply_q_offset: Operational reply queue offset with MSIx
+ * @default_qcount: Total Default queues
+ * @active_poll_qcount: Currently active poll queue count
+ * @requested_poll_qcount: User requested poll queue count
  */
 struct mpi3mr_ioc {
 	struct list_head list;
@@ -839,6 +850,10 @@ struct mpi3mr_ioc {
 	struct mpi3_driver_info_layout driver_info;
 	u16 change_count;
 	u16 op_reply_q_offset;
+
+	u16 default_qcount;
+	u16 active_poll_qcount;
+	u16 requested_poll_qcount;
 };
 
 /**
@@ -940,5 +955,8 @@ void mpi3mr_flush_delayed_cmd_lists(struct mpi3mr_ioc *mrioc);
 void mpi3mr_check_rh_fault_ioc(struct mpi3mr_ioc *mrioc, u32 reason_code);
 void mpi3mr_print_fault_info(struct mpi3mr_ioc *mrioc);
 void mpi3mr_check_rh_fault_ioc(struct mpi3mr_ioc *mrioc, u32 reason_code);
+int mpi3mr_process_op_reply_q(struct mpi3mr_ioc *mrioc,
+	struct op_reply_qinfo *op_reply_q);
+int mpi3mr_blk_mq_poll(struct Scsi_Host *shost, unsigned int queue_num);
 
 #endif /*MPI3MR_H_INCLUDED*/
diff --git a/drivers/scsi/mpi3mr/mpi3mr_fw.c b/drivers/scsi/mpi3mr/mpi3mr_fw.c
index a4decaa89e8d..0fa66b34f0a6 100644
--- a/drivers/scsi/mpi3mr/mpi3mr_fw.c
+++ b/drivers/scsi/mpi3mr/mpi3mr_fw.c
@@ -16,6 +16,10 @@ static int mpi3mr_setup_admin_qpair(struct mpi3mr_ioc *mrioc);
 static void mpi3mr_process_factsdata(struct mpi3mr_ioc *mrioc,
 	struct mpi3_ioc_facts_data *facts_data);
 
+static int poll_queues;
+module_param(poll_queues, int, 0444);
+MODULE_PARM_DESC(poll_queues, "Number of queues for io_uring poll mode. (Range 1 - 126)");
+
 #if defined(writeq) && defined(CONFIG_64BIT)
 static inline void mpi3mr_writeq(__u64 b, volatile void __iomem *addr)
 {
@@ -461,10 +465,21 @@ mpi3mr_get_reply_desc(struct op_reply_qinfo *op_reply_q, u32 reply_ci)
 	return reply_desc;
 }
 
-static int mpi3mr_process_op_reply_q(struct mpi3mr_ioc *mrioc,
-	struct mpi3mr_intr_info *intr_info)
+/**
+ * mpi3mr_process_op_reply_q - Operational reply queue handler
+ * @mrioc: Adapter instance reference
+ * @op_reply_q: Operational reply queue info
+ *
+ * Checks the specific operational reply queue and drains the
+ * reply queue entries until the queue is empty and process the
+ * individual reply descriptors.
+ *
+ * Return: 0 if queue is already processed,or number of reply
+ *	    descriptors processed.
+ */
+int mpi3mr_process_op_reply_q(struct mpi3mr_ioc *mrioc,
+	struct op_reply_qinfo *op_reply_q)
 {
-	struct op_reply_qinfo *op_reply_q = intr_info->op_reply_q;
 	struct op_req_qinfo *op_req_q;
 	u32 exp_phase;
 	u32 reply_ci;
@@ -515,7 +530,7 @@ static int mpi3mr_process_op_reply_q(struct mpi3mr_ioc *mrioc,
 		 * Ensure remaining completion happens from threaded ISR.
 		 */
 		if (num_op_reply > mrioc->max_host_ios) {
-			intr_info->op_reply_q->enable_irq_poll = true;
+			op_reply_q->enable_irq_poll = true;
 			break;
 		}
 
@@ -530,6 +545,34 @@ static int mpi3mr_process_op_reply_q(struct mpi3mr_ioc *mrioc,
 	return num_op_reply;
 }
 
+/**
+ * mpi3mr_blk_mq_poll - Operational reply queue handler
+ * @shost: SCSI Host reference
+ * @queue_num: Request queue number (w.r.t OS it is hardware context number)
+ *
+ * Checks the specific operational reply queue and drains the
+ * reply queue entries until the queue is empty and process the
+ * individual reply descriptors.
+ *
+ * Return: 0 if queue is already processed,or number of reply
+ *	    descriptors processed.
+ */
+int mpi3mr_blk_mq_poll(struct Scsi_Host *shost, unsigned int queue_num)
+{
+	int num_entries = 0;
+	struct mpi3mr_ioc *mrioc;
+
+	mrioc = (struct mpi3mr_ioc *)shost->hostdata;
+
+	if ((mrioc->reset_in_progress || mrioc->prepare_for_reset))
+		return 0;
+
+	num_entries = mpi3mr_process_op_reply_q(mrioc,
+			&mrioc->op_reply_qinfo[queue_num]);
+
+	return num_entries;
+}
+
 static irqreturn_t mpi3mr_isr_primary(int irq, void *privdata)
 {
 	struct mpi3mr_intr_info *intr_info = privdata;
@@ -550,7 +593,8 @@ static irqreturn_t mpi3mr_isr_primary(int irq, void *privdata)
 	if (!midx)
 		num_admin_replies = mpi3mr_process_admin_reply_q(mrioc);
 	if (intr_info->op_reply_q)
-		num_op_reply = mpi3mr_process_op_reply_q(mrioc, intr_info);
+		num_op_reply = mpi3mr_process_op_reply_q(mrioc,
+		    intr_info->op_reply_q);
 
 	if (num_admin_replies || num_op_reply)
 		return IRQ_HANDLED;
@@ -621,9 +665,10 @@ static irqreturn_t mpi3mr_isr_poll(int irq, void *privdata)
 			mpi3mr_process_admin_reply_q(mrioc);
 		if (intr_info->op_reply_q)
 			num_op_reply +=
-			    mpi3mr_process_op_reply_q(mrioc, intr_info);
+			    mpi3mr_process_op_reply_q(mrioc,
+				intr_info->op_reply_q);
 
-		usleep_range(mrioc->irqpoll_sleep, 10 * mrioc->irqpoll_sleep);
+		usleep_range(MPI3MR_IRQ_POLL_SLEEP, 10 * MPI3MR_IRQ_POLL_SLEEP);
 
 	} while (atomic_read(&intr_info->op_reply_q->pend_ios) &&
 	    (num_op_reply < mrioc->max_host_ios));
@@ -667,6 +712,25 @@ static inline int mpi3mr_request_irq(struct mpi3mr_ioc *mrioc, u16 index)
 	return retval;
 }
 
+static void mpi3mr_calc_poll_queues(struct mpi3mr_ioc *mrioc, u16 max_vectors)
+{
+	if (!mrioc->requested_poll_qcount)
+		return;
+
+	/* Reserved for Admin and Default Queue */
+	if (max_vectors > 2 &&
+		(mrioc->requested_poll_qcount < max_vectors - 2)) {
+		ioc_info(mrioc,
+		    "enabled polled queues (%d) msix (%d)\n",
+		    mrioc->requested_poll_qcount, max_vectors);
+	} else {
+		ioc_info(mrioc,
+		    "disabled polled queues (%d) msix (%d) because of no resources for default queue\n",
+		    mrioc->requested_poll_qcount, max_vectors);
+		mrioc->requested_poll_qcount = 0;
+	}
+}
+
 /**
  * mpi3mr_setup_isr - Setup ISR for the controller
  * @mrioc: Adapter instance reference
@@ -679,51 +743,72 @@ static inline int mpi3mr_request_irq(struct mpi3mr_ioc *mrioc, u16 index)
 static int mpi3mr_setup_isr(struct mpi3mr_ioc *mrioc, u8 setup_one)
 {
 	unsigned int irq_flags = PCI_IRQ_MSIX;
-	int max_vectors;
+	int max_vectors, min_vec;
 	int retval;
 	int i;
-	struct irq_affinity desc = { .pre_vectors =  1};
+	struct irq_affinity desc = { .pre_vectors =  1, .post_vectors = 1 };
 
 	if (mrioc->is_intr_info_set)
 		return 0;
 
 	mpi3mr_cleanup_isr(mrioc);
 
-	if (setup_one || reset_devices)
+	if (setup_one || reset_devices) {
 		max_vectors = 1;
-	else {
+		retval = pci_alloc_irq_vectors(mrioc->pdev,
+		    1, max_vectors, irq_flags);
+		if (retval < 0) {
+			ioc_err(mrioc, "cannot allocate irq vectors, ret %d\n",
+			    retval);
+			goto out_failed;
+		}
+	} else {
 		max_vectors =
-		    min_t(int, mrioc->cpu_count + 1, mrioc->msix_count);
+		    min_t(int, mrioc->cpu_count + 1 +
+			mrioc->requested_poll_qcount, mrioc->msix_count);
+
+		mpi3mr_calc_poll_queues(mrioc, max_vectors);
 
 		ioc_info(mrioc,
 		    "MSI-X vectors supported: %d, no of cores: %d,",
 		    mrioc->msix_count, mrioc->cpu_count);
 		ioc_info(mrioc,
-		    "MSI-x vectors requested: %d\n", max_vectors);
-	}
+		    "MSI-x vectors requested: %d poll_queues %d\n",
+		    max_vectors, mrioc->requested_poll_qcount);
+
+		desc.post_vectors = mrioc->requested_poll_qcount;
+		min_vec = desc.pre_vectors + desc.post_vectors;
+		irq_flags |= PCI_IRQ_AFFINITY | PCI_IRQ_ALL_TYPES;
+
+		retval = pci_alloc_irq_vectors_affinity(mrioc->pdev,
+			min_vec, max_vectors, irq_flags, &desc);
+
+		if (retval < 0) {
+			ioc_err(mrioc, "cannot allocate irq vectors, ret %d\n",
+			    retval);
+			goto out_failed;
+		}
 
-	irq_flags |= PCI_IRQ_AFFINITY | PCI_IRQ_ALL_TYPES;
 
-	mrioc->op_reply_q_offset = (max_vectors > 1) ? 1 : 0;
-	retval = pci_alloc_irq_vectors_affinity(mrioc->pdev,
-				1, max_vectors, irq_flags, &desc);
-	if (retval < 0) {
-		ioc_err(mrioc, "Cannot alloc irq vectors\n");
-		goto out_failed;
-	}
-	if (retval != max_vectors) {
-		ioc_info(mrioc,
-		    "allocated vectors (%d) are less than configured (%d)\n",
-		    retval, max_vectors);
 		/*
 		 * If only one MSI-x is allocated, then MSI-x 0 will be shared
 		 * between Admin queue and operational queue
 		 */
-		if (retval == 1)
+		if (retval == min_vec)
 			mrioc->op_reply_q_offset = 0;
+		else if (retval != (max_vectors)) {
+			ioc_info(mrioc,
+			    "allocated vectors (%d) are less than configured (%d)\n",
+			    retval, max_vectors);
+		}
 
 		max_vectors = retval;
+		mrioc->op_reply_q_offset = (max_vectors > 1) ? 1 : 0;
+
+		mpi3mr_calc_poll_queues(mrioc, max_vectors);
+
 	}
+
 	mrioc->intr_info = kzalloc(sizeof(struct mpi3mr_intr_info) * max_vectors,
 	    GFP_KERNEL);
 	if (!mrioc->intr_info) {
@@ -1511,10 +1596,11 @@ static void mpi3mr_free_op_reply_q_segments(struct mpi3mr_ioc *mrioc, u16 q_idx)
 static int mpi3mr_delete_op_reply_q(struct mpi3mr_ioc *mrioc, u16 qidx)
 {
 	struct mpi3_delete_reply_queue_request delq_req;
+	struct op_reply_qinfo *op_reply_q = mrioc->op_reply_qinfo + qidx;
 	int retval = 0;
 	u16 reply_qid = 0, midx;
 
-	reply_qid = mrioc->op_reply_qinfo[qidx].qid;
+	reply_qid = op_reply_q->qid;
 
 	midx = REPLY_QUEUE_IDX_TO_MSIX_IDX(qidx, mrioc->op_reply_q_offset);
 
@@ -1524,6 +1610,9 @@ static int mpi3mr_delete_op_reply_q(struct mpi3mr_ioc *mrioc, u16 qidx)
 		goto out;
 	}
 
+	(op_reply_q->qtype == MPI3MR_DEFAULT_QUEUE) ? mrioc->default_qcount-- :
+	    mrioc->active_poll_qcount--;
+
 	memset(&delq_req, 0, sizeof(delq_req));
 	mutex_lock(&mrioc->init_cmds.mutex);
 	if (mrioc->init_cmds.state & MPI3MR_CMD_PENDING) {
@@ -1748,8 +1837,26 @@ static int mpi3mr_create_op_reply_q(struct mpi3mr_ioc *mrioc, u16 qidx)
 	create_req.host_tag = cpu_to_le16(MPI3MR_HOSTTAG_INITCMDS);
 	create_req.function = MPI3_FUNCTION_CREATE_REPLY_QUEUE;
 	create_req.queue_id = cpu_to_le16(reply_qid);
-	create_req.flags = MPI3_CREATE_REPLY_QUEUE_FLAGS_INT_ENABLE_ENABLE;
-	create_req.msix_index = cpu_to_le16(mrioc->intr_info[midx].msix_index);
+
+	if (midx < (mrioc->intr_info_count - mrioc->requested_poll_qcount))
+		op_reply_q->qtype = MPI3MR_DEFAULT_QUEUE;
+	else
+		op_reply_q->qtype = MPI3MR_POLL_QUEUE;
+
+	if (op_reply_q->qtype == MPI3MR_DEFAULT_QUEUE) {
+		create_req.flags =
+			MPI3_CREATE_REPLY_QUEUE_FLAGS_INT_ENABLE_ENABLE;
+		create_req.msix_index =
+			cpu_to_le16(mrioc->intr_info[midx].msix_index);
+	} else {
+		create_req.msix_index = cpu_to_le16(mrioc->intr_info_count - 1);
+		ioc_info(mrioc, "create reply queue(polled): for qid(%d), midx(%d)\n",
+			reply_qid, midx);
+		if (!mrioc->active_poll_qcount)
+			disable_irq_nosync(pci_irq_vector(mrioc->pdev,
+			    mrioc->intr_info_count - 1));
+	}
+
 	if (mrioc->enable_segqueue) {
 		create_req.flags |=
 		    MPI3_CREATE_REQUEST_QUEUE_FLAGS_SEGMENTED_SEGMENTED;
@@ -1790,6 +1897,9 @@ static int mpi3mr_create_op_reply_q(struct mpi3mr_ioc *mrioc, u16 qidx)
 	if (midx < mrioc->intr_info_count)
 		mrioc->intr_info[midx].op_reply_q = op_reply_q;
 
+	(op_reply_q->qtype == MPI3MR_DEFAULT_QUEUE) ? mrioc->default_qcount++ :
+	    mrioc->active_poll_qcount++;
+
 out_unlock:
 	mrioc->init_cmds.state = MPI3MR_CMD_NOTUSED;
 	mutex_unlock(&mrioc->init_cmds.mutex);
@@ -1970,8 +2080,10 @@ static int mpi3mr_create_op_queues(struct mpi3mr_ioc *mrioc)
 		goto out_failed;
 	}
 	mrioc->num_op_reply_q = mrioc->num_op_req_q = i;
-	ioc_info(mrioc, "Successfully created %d Operational Q pairs\n",
-	    mrioc->num_op_reply_q);
+	ioc_info(mrioc,
+	    "successfully created %d operational queue pairs(default/polled) queue = (%d/%d)\n",
+	    mrioc->num_op_reply_q, mrioc->default_qcount,
+	    mrioc->active_poll_qcount);
 
 	return retval;
 out_failed:
@@ -2019,7 +2131,7 @@ int mpi3mr_op_request_post(struct mpi3mr_ioc *mrioc,
 	if (mpi3mr_check_req_qfull(op_req_q)) {
 		midx = REPLY_QUEUE_IDX_TO_MSIX_IDX(
 		    reply_qidx, mrioc->op_reply_q_offset);
-		mpi3mr_process_op_reply_q(mrioc, &mrioc->intr_info[midx]);
+		mpi3mr_process_op_reply_q(mrioc, mrioc->intr_info[midx].op_reply_q);
 
 		if (mpi3mr_check_req_qfull(op_req_q)) {
 			retval = -EAGAIN;
@@ -3465,6 +3577,10 @@ int mpi3mr_setup_resources(struct mpi3mr_ioc *mrioc)
 	    mrioc->sysif_regs, memap_sz);
 	ioc_info(mrioc, "Number of MSI-X vectors found in capabilities: (%d)\n",
 	    mrioc->msix_count);
+
+	if (!reset_devices && poll_queues > 0)
+		mrioc->requested_poll_qcount = min_t(int, poll_queues,
+				mrioc->msix_count - 2);
 	return retval;
 
 out_failed:
@@ -3826,6 +3942,8 @@ void mpi3mr_memset_buffers(struct mpi3mr_ioc *mrioc)
 	u16 i;
 
 	mrioc->change_count = 0;
+	mrioc->active_poll_qcount = 0;
+	mrioc->default_qcount = 0;
 	if (mrioc->admin_req_base)
 		memset(mrioc->admin_req_base, 0, mrioc->admin_req_q_sz);
 	if (mrioc->admin_reply_base)
diff --git a/drivers/scsi/mpi3mr/mpi3mr_os.c b/drivers/scsi/mpi3mr/mpi3mr_os.c
index d893c6d61a60..8bf1b590cc58 100644
--- a/drivers/scsi/mpi3mr/mpi3mr_os.c
+++ b/drivers/scsi/mpi3mr/mpi3mr_os.c
@@ -3049,17 +3049,49 @@ static int mpi3mr_bios_param(struct scsi_device *sdev,
  * mpi3mr_map_queues - Map queues callback handler
  * @shost: SCSI host reference
  *
- * Call the blk_mq_pci_map_queues with from which operational
- * queue the mapping has to be done
+ * Maps default and poll queues.
  *
- * Return: return of blk_mq_pci_map_queues
+ * Return: return zero.
  */
 static int mpi3mr_map_queues(struct Scsi_Host *shost)
 {
 	struct mpi3mr_ioc *mrioc = shost_priv(shost);
+	int i, qoff, offset;
+	struct blk_mq_queue_map *map = NULL;
+
+	offset = mrioc->op_reply_q_offset;
+
+	for (i = 0, qoff = 0; i < HCTX_MAX_TYPES; i++) {
+		map = &shost->tag_set.map[i];
+
+		map->nr_queues  = 0;
+
+		if (i == HCTX_TYPE_DEFAULT)
+			map->nr_queues = mrioc->default_qcount;
+		else if (i == HCTX_TYPE_POLL)
+			map->nr_queues = mrioc->active_poll_qcount;
+
+		if (!map->nr_queues) {
+			BUG_ON(i == HCTX_TYPE_DEFAULT);
+			continue;
+		}
+
+		/*
+		 * The poll queue(s) doesn't have an IRQ (and hence IRQ
+		 * affinity), so use the regular blk-mq cpu mapping
+		 */
+		map->queue_offset = qoff;
+		if (i != HCTX_TYPE_POLL)
+			blk_mq_pci_map_queues(map, mrioc->pdev, offset);
+		else
+			blk_mq_map_queues(map);
+
+		qoff += map->nr_queues;
+		offset += map->nr_queues;
+	}
+
+	return 0;
 
-	return blk_mq_pci_map_queues(&shost->tag_set.map[HCTX_TYPE_DEFAULT],
-	    mrioc->pdev, mrioc->op_reply_q_offset);
 }
 
 /**
@@ -3873,6 +3905,7 @@ static struct scsi_host_template mpi3mr_driver_template = {
 	.eh_host_reset_handler		= mpi3mr_eh_host_reset,
 	.bios_param			= mpi3mr_bios_param,
 	.map_queues			= mpi3mr_map_queues,
+	.mq_poll                        = mpi3mr_blk_mq_poll,
 	.no_write_same			= 1,
 	.can_queue			= 1,
 	.this_id			= -1,
@@ -4105,6 +4138,9 @@ mpi3mr_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	}
 
 	shost->nr_hw_queues = mrioc->num_op_reply_q;
+	if (mrioc->active_poll_qcount)
+		shost->nr_maps = 3;
+
 	shost->can_queue = mrioc->max_host_ios;
 	shost->sg_tablesize = MPI3MR_SG_DEPTH;
 	shost->max_id = mrioc->facts.max_perids + 1;