Merge branch 'rdma-cq.2' of git://git.infradead.org/users/hch/rdma into 4.5/rdma-cq

Signed-off-by: Doug Ledford <dledford@redhat.com>

Conflicts:
	drivers/infiniband/ulp/srp/ib_srp.c - Conflicts with changes in
	ib_srp.c introduced during 4.4-rc updates

Documentation/kernel-per-CPU-kthreads.txt

@@ -90,7 +90,7 @@ BLOCK_SOFTIRQ:  Do all of the following:
 	from being initiated from tasks that might run on the CPU to
 	be de-jittered.  (It is OK to force this CPU offline and then
 	bring it back online before you start your application.)
-BLOCK_IOPOLL_SOFTIRQ:  Do all of the following:
+IRQ_POLL_SOFTIRQ:  Do all of the following:
 1.	Force block-device interrupts onto some other CPU.
 2.	Initiate any block I/O and block-I/O polling on other CPUs.
 3.	Once your application has started, prevent CPU-hotplug operations

block/Makefile

@@ -5,7 +5,7 @@
 obj-$(CONFIG_BLOCK) := bio.o elevator.o blk-core.o blk-tag.o blk-sysfs.o \
 			blk-flush.o blk-settings.o blk-ioc.o blk-map.o \
 			blk-exec.o blk-merge.o blk-softirq.o blk-timeout.o \
-			blk-iopoll.o blk-lib.o blk-mq.o blk-mq-tag.o \
+			blk-lib.o blk-mq.o blk-mq-tag.o \
 			blk-mq-sysfs.o blk-mq-cpu.o blk-mq-cpumap.o ioctl.o \
 			genhd.o scsi_ioctl.o partition-generic.o ioprio.o \
 			partitions/

drivers/infiniband/Kconfig

@@ -5,6 +5,7 @@ menuconfig INFINIBAND
 	depends on NET
 	depends on INET
 	depends on m || IPV6 != m
+	select IRQ_POLL
 	---help---
 	  Core support for InfiniBand (IB).  Make sure to also select
 	  any protocols you wish to use as well as drivers for your

drivers/infiniband/core/Makefile

@@ -8,7 +8,7 @@ obj-$(CONFIG_INFINIBAND_USER_MAD) +=	ib_umad.o
 obj-$(CONFIG_INFINIBAND_USER_ACCESS) +=	ib_uverbs.o ib_ucm.o \
 					$(user_access-y)
 
-ib_core-y :=			packer.o ud_header.o verbs.o sysfs.o \
+ib_core-y :=			packer.o ud_header.o verbs.o cq.o sysfs.o \
 				device.o fmr_pool.o cache.o netlink.o \
 				roce_gid_mgmt.o
 ib_core-$(CONFIG_INFINIBAND_USER_MEM) += umem.o

drivers/infiniband/core/cq.c

+/*
+ * Copyright (c) 2015 HGST, a Western Digital Company.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ */
+#include <linux/module.h>
+#include <linux/err.h>
+#include <linux/slab.h>
+#include <rdma/ib_verbs.h>
+
+/* # of WCs to poll for with a single call to ib_poll_cq */
+#define IB_POLL_BATCH			16
+
+/* # of WCs to iterate over before yielding */
+#define IB_POLL_BUDGET_IRQ		256
+#define IB_POLL_BUDGET_WORKQUEUE	65536
+
+#define IB_POLL_FLAGS \
+	(IB_CQ_NEXT_COMP | IB_CQ_REPORT_MISSED_EVENTS)
+
+static int __ib_process_cq(struct ib_cq *cq, int budget)
+{
+	int i, n, completed = 0;
+
+	while ((n = ib_poll_cq(cq, IB_POLL_BATCH, cq->wc)) > 0) {
+		for (i = 0; i < n; i++) {
+			struct ib_wc *wc = &cq->wc[i];
+
+			if (wc->wr_cqe)
+				wc->wr_cqe->done(cq, wc);
+			else
+				WARN_ON_ONCE(wc->status == IB_WC_SUCCESS);
+		}
+
+		completed += n;
+
+		if (n != IB_POLL_BATCH ||
+		    (budget != -1 && completed >= budget))
+			break;
+	}
+
+	return completed;
+}
+
+/**
+ * ib_process_direct_cq - process a CQ in caller context
+ * @cq:		CQ to process
+ * @budget:	number of CQEs to poll for
+ *
+ * This function is used to process all outstanding CQ entries on a
+ * %IB_POLL_DIRECT CQ.  It does not offload CQ processing to a different
+ * context and does not ask for completion interrupts from the HCA.
+ *
+ * Note: for compatibility reasons -1 can be passed in %budget for unlimited
+ * polling.  Do not use this feature in new code, it will be removed soon.
+ */
+int ib_process_cq_direct(struct ib_cq *cq, int budget)
+{
+	WARN_ON_ONCE(cq->poll_ctx != IB_POLL_DIRECT);
+
+	return __ib_process_cq(cq, budget);
+}
+EXPORT_SYMBOL(ib_process_cq_direct);
+
+static void ib_cq_completion_direct(struct ib_cq *cq, void *private)
+{
+	WARN_ONCE(1, "got unsolicited completion for CQ 0x%p\n", cq);
+}
+
+static int ib_poll_handler(struct irq_poll *iop, int budget)
+{
+	struct ib_cq *cq = container_of(iop, struct ib_cq, iop);
+	int completed;
+
+	completed = __ib_process_cq(cq, budget);
+	if (completed < budget) {
+		irq_poll_complete(&cq->iop);
+		if (ib_req_notify_cq(cq, IB_POLL_FLAGS) > 0)
+			irq_poll_sched(&cq->iop);
+	}
+
+	return completed;
+}
+
+static void ib_cq_completion_softirq(struct ib_cq *cq, void *private)
+{
+	irq_poll_sched(&cq->iop);
+}
+
+static void ib_cq_poll_work(struct work_struct *work)
+{
+	struct ib_cq *cq = container_of(work, struct ib_cq, work);
+	int completed;
+
+	completed = __ib_process_cq(cq, IB_POLL_BUDGET_WORKQUEUE);
+	if (completed >= IB_POLL_BUDGET_WORKQUEUE ||
+	    ib_req_notify_cq(cq, IB_POLL_FLAGS) > 0)
+		queue_work(ib_comp_wq, &cq->work);
+}
+
+static void ib_cq_completion_workqueue(struct ib_cq *cq, void *private)
+{
+	queue_work(ib_comp_wq, &cq->work);
+}
+
+/**
+ * ib_alloc_cq - allocate a completion queue
+ * @dev:		device to allocate the CQ for
+ * @private:		driver private data, accessible from cq->cq_context
+ * @nr_cqe:		number of CQEs to allocate
+ * @comp_vector:	HCA completion vectors for this CQ
+ * @poll_ctx:		context to poll the CQ from.
+ *
+ * This is the proper interface to allocate a CQ for in-kernel users. A
+ * CQ allocated with this interface will automatically be polled from the
+ * specified context.  The ULP needs must use wr->wr_cqe instead of wr->wr_id
+ * to use this CQ abstraction.
+ */
+struct ib_cq *ib_alloc_cq(struct ib_device *dev, void *private,
+		int nr_cqe, int comp_vector, enum ib_poll_context poll_ctx)
+{
+	struct ib_cq_init_attr cq_attr = {
+		.cqe		= nr_cqe,
+		.comp_vector	= comp_vector,
+	};
+	struct ib_cq *cq;
+	int ret = -ENOMEM;
+
+	cq = dev->create_cq(dev, &cq_attr, NULL, NULL);
+	if (IS_ERR(cq))
+		return cq;
+
+	cq->device = dev;
+	cq->uobject = NULL;
+	cq->event_handler = NULL;
+	cq->cq_context = private;
+	cq->poll_ctx = poll_ctx;
+	atomic_set(&cq->usecnt, 0);
+
+	cq->wc = kmalloc_array(IB_POLL_BATCH, sizeof(*cq->wc), GFP_KERNEL);
+	if (!cq->wc)
+		goto out_destroy_cq;
+
+	switch (cq->poll_ctx) {
+	case IB_POLL_DIRECT:
+		cq->comp_handler = ib_cq_completion_direct;
+		break;
+	case IB_POLL_SOFTIRQ:
+		cq->comp_handler = ib_cq_completion_softirq;
+
+		irq_poll_init(&cq->iop, IB_POLL_BUDGET_IRQ, ib_poll_handler);
+		ib_req_notify_cq(cq, IB_CQ_NEXT_COMP);
+		break;
+	case IB_POLL_WORKQUEUE:
+		cq->comp_handler = ib_cq_completion_workqueue;
+		INIT_WORK(&cq->work, ib_cq_poll_work);
+		ib_req_notify_cq(cq, IB_CQ_NEXT_COMP);
+		break;
+	default:
+		ret = -EINVAL;
+		goto out_free_wc;
+	}
+
+	return cq;
+
+out_free_wc:
+	kfree(cq->wc);
+out_destroy_cq:
+	cq->device->destroy_cq(cq);
+	return ERR_PTR(ret);
+}
+EXPORT_SYMBOL(ib_alloc_cq);
+
+/**
+ * ib_free_cq - free a completion queue
+ * @cq:		completion queue to free.
+ */
+void ib_free_cq(struct ib_cq *cq)
+{
+	int ret;
+
+	if (WARN_ON_ONCE(atomic_read(&cq->usecnt)))
+		return;
+
+	switch (cq->poll_ctx) {
+	case IB_POLL_DIRECT:
+		break;
+	case IB_POLL_SOFTIRQ:
+		irq_poll_disable(&cq->iop);
+		break;
+	case IB_POLL_WORKQUEUE:
+		flush_work(&cq->work);
+		break;
+	default:
+		WARN_ON_ONCE(1);
+	}
+
+	kfree(cq->wc);
+	ret = cq->device->destroy_cq(cq);
+	WARN_ON_ONCE(ret);
+}
+EXPORT_SYMBOL(ib_free_cq);

drivers/infiniband/core/device.c

@@ -58,6 +58,7 @@ struct ib_client_data {
 	bool		  going_down;
 };
 
+struct workqueue_struct *ib_comp_wq;
 struct workqueue_struct *ib_wq;
 EXPORT_SYMBOL_GPL(ib_wq);
 
@@ -954,10 +955,18 @@ static int __init ib_core_init(void)
 	if (!ib_wq)
 		return -ENOMEM;
 
+	ib_comp_wq = alloc_workqueue("ib-comp-wq",
+			WQ_UNBOUND | WQ_HIGHPRI | WQ_MEM_RECLAIM,
+			WQ_UNBOUND_MAX_ACTIVE);
+	if (!ib_comp_wq) {
+		ret = -ENOMEM;
+		goto err;
+	}
+
 	ret = class_register(&ib_class);
 	if (ret) {
 		printk(KERN_WARNING "Couldn't create InfiniBand device class\n");
-		goto err;
+		goto err_comp;
 	}
 
 	ret = ibnl_init();
@@ -972,7 +981,8 @@ static int __init ib_core_init(void)
 
 err_sysfs:
 	class_unregister(&ib_class);
-
+err_comp:
+	destroy_workqueue(ib_comp_wq);
 err:
 	destroy_workqueue(ib_wq);
 	return ret;
@@ -983,6 +993,7 @@ static void __exit ib_core_cleanup(void)
 	ib_cache_cleanup();
 	ibnl_cleanup();
 	class_unregister(&ib_class);
+	destroy_workqueue(ib_comp_wq);
 	/* Make sure that any pending umem accounting work is done. */
 	destroy_workqueue(ib_wq);
 }

drivers/infiniband/ulp/ipoib/ipoib_cm.c

@@ -70,7 +70,6 @@ static struct ib_qp_attr ipoib_cm_err_attr = {
 #define IPOIB_CM_RX_DRAIN_WRID 0xffffffff
 
 static struct ib_send_wr ipoib_cm_rx_drain_wr = {
-	.wr_id = IPOIB_CM_RX_DRAIN_WRID,
 	.opcode = IB_WR_SEND,
 };
 
@@ -223,6 +222,7 @@ static void ipoib_cm_start_rx_drain(struct ipoib_dev_priv *priv)
 	 * error" WC will be immediately generated for each WR we post.
 	 */
 	p = list_entry(priv->cm.rx_flush_list.next, typeof(*p), list);
+	ipoib_cm_rx_drain_wr.wr_id = IPOIB_CM_RX_DRAIN_WRID;
 	if (ib_post_send(p->qp, &ipoib_cm_rx_drain_wr, &bad_wr))
 		ipoib_warn(priv, "failed to post drain wr\n");
 

drivers/infiniband/ulp/iser/iscsi_iser.h

@@ -151,16 +151,12 @@
 					 - ISER_MAX_RX_MISC_PDUS) /	\
 					 (1 + ISER_INFLIGHT_DATAOUTS))
 
-#define ISER_WC_BATCH_COUNT   16
 #define ISER_SIGNAL_CMD_COUNT 32
 
 #define ISER_VER			0x10
 #define ISER_WSV			0x08
 #define ISER_RSV			0x04
 
-#define ISER_FASTREG_LI_WRID		0xffffffffffffffffULL
-#define ISER_BEACON_WRID		0xfffffffffffffffeULL
-
 /**
  * struct iser_hdr - iSER header
  *
@@ -269,7 +265,7 @@ enum iser_desc_type {
 #define ISER_MAX_WRS 7
 
 /**
- * struct iser_tx_desc - iSER TX descriptor (for send wr_id)
+ * struct iser_tx_desc - iSER TX descriptor
  *
  * @iser_header:   iser header
  * @iscsi_header:  iscsi header
@@ -293,6 +289,7 @@ struct iser_tx_desc {
 	u64		             dma_addr;
 	struct ib_sge		     tx_sg[2];
 	int                          num_sge;
+	struct ib_cqe		     cqe;
 	bool			     mapped;
 	u8                           wr_idx;
 	union iser_wr {
@@ -306,9 +303,10 @@ struct iser_tx_desc {
 };
 
 #define ISER_RX_PAD_SIZE	(256 - (ISER_RX_PAYLOAD_SIZE + \
-					sizeof(u64) + sizeof(struct ib_sge)))
+				 sizeof(u64) + sizeof(struct ib_sge) + \
+				 sizeof(struct ib_cqe)))
 /**
- * struct iser_rx_desc - iSER RX descriptor (for recv wr_id)
+ * struct iser_rx_desc - iSER RX descriptor
  *
  * @iser_header:   iser header
  * @iscsi_header:  iscsi header
@@ -323,7 +321,27 @@ struct iser_rx_desc {
 	char		             data[ISER_RECV_DATA_SEG_LEN];
 	u64		             dma_addr;
 	struct ib_sge		     rx_sg;
+	struct ib_cqe		     cqe;
 	char		             pad[ISER_RX_PAD_SIZE];
+} __packed;
+
+/**
+ * struct iser_login_desc - iSER login descriptor
+ *
+ * @req:           pointer to login request buffer
+ * @resp:          pointer to login response buffer
+ * @req_dma:       DMA address of login request buffer
+ * @rsp_dma:      DMA address of login response buffer
+ * @sge:           IB sge for login post recv
+ * @cqe:           completion handler
+ */
+struct iser_login_desc {
+	void                         *req;
+	void                         *rsp;
+	u64                          req_dma;
+	u64                          rsp_dma;
+	struct ib_sge                sge;
+	struct ib_cqe		     cqe;
 } __attribute__((packed));
 
 struct iser_conn;
@@ -333,18 +351,12 @@ struct iscsi_iser_task;
 /**
  * struct iser_comp - iSER completion context
  *
- * @device:     pointer to device handle
  * @cq:         completion queue
- * @wcs:        work completion array
- * @tasklet:    Tasklet handle
  * @active_qps: Number of active QPs attached
  *              to completion context
  */
 struct iser_comp {
-	struct iser_device      *device;
 	struct ib_cq		*cq;
-	struct ib_wc		 wcs[ISER_WC_BATCH_COUNT];
-	struct tasklet_struct	 tasklet;
 	int                      active_qps;
 };
 
@@ -475,10 +487,11 @@ struct iser_fr_pool {
  * @rx_wr:               receive work request for batch posts
  * @device:              reference to iser device
  * @comp:                iser completion context
- * @pi_support:          Indicate device T10-PI support
- * @beacon:              beacon send wr to signal all flush errors were drained
- * @flush_comp:          completes when all connection completions consumed
  * @fr_pool:             connection fast registration poool
+ * @pi_support:          Indicate device T10-PI support
+ * @last:                last send wr to signal all flush errors were drained
+ * @last_cqe:            cqe handler for last wr
+ * @last_comp:           completes when all connection completions consumed
  */
 struct ib_conn {
 	struct rdma_cm_id           *cma_id;
@@ -488,10 +501,12 @@ struct ib_conn {
 	struct ib_recv_wr	     rx_wr[ISER_MIN_POSTED_RX];
 	struct iser_device          *device;
 	struct iser_comp	    *comp;
-	bool			     pi_support;
-	struct ib_send_wr	     beacon;
-	struct completion	     flush_comp;
 	struct iser_fr_pool          fr_pool;
+	bool			     pi_support;
+	struct ib_send_wr	     last;
+	struct ib_cqe		     last_cqe;
+	struct ib_cqe		     reg_cqe;
+	struct completion	     last_comp;
 };
 
 /**
@@ -514,11 +529,7 @@ struct ib_conn {
  * @up_completion:    connection establishment completed
  *                    (state is ISER_CONN_UP)
  * @conn_list:        entry in ig conn list
- * @login_buf:        login data buffer (stores login parameters)
- * @login_req_buf:    login request buffer
- * @login_req_dma:    login request buffer dma address
- * @login_resp_buf:   login response buffer
- * @login_resp_dma:   login response buffer dma address
+ * @login_desc:       login descriptor
  * @rx_desc_head:     head of rx_descs cyclic buffer
  * @rx_descs:         rx buffers array (cyclic buffer)
  * @num_rx_descs:     number of rx descriptors
@@ -541,10 +552,7 @@ struct iser_conn {
 	struct completion	     ib_completion;
 	struct completion	     up_completion;
 	struct list_head	     conn_list;
-
-	char  			     *login_buf;
-	char			     *login_req_buf, *login_resp_buf;
-	u64			     login_req_dma, login_resp_dma;
+	struct iser_login_desc       login_desc;
 	unsigned int 		     rx_desc_head;
 	struct iser_rx_desc	     *rx_descs;
 	u32                          num_rx_descs;
@@ -633,12 +641,14 @@ int iser_conn_terminate(struct iser_conn *iser_conn);
 
 void iser_release_work(struct work_struct *work);
 
-void iser_rcv_completion(struct iser_rx_desc *desc,
-			 unsigned long dto_xfer_len,
-			 struct ib_conn *ib_conn);
-
-void iser_snd_completion(struct iser_tx_desc *desc,
-			 struct ib_conn *ib_conn);
+void iser_err_comp(struct ib_wc *wc, const char *type);
+void iser_login_rsp(struct ib_cq *cq, struct ib_wc *wc);
+void iser_task_rsp(struct ib_cq *cq, struct ib_wc *wc);
+void iser_cmd_comp(struct ib_cq *cq, struct ib_wc *wc);
+void iser_ctrl_comp(struct ib_cq *cq, struct ib_wc *wc);
+void iser_dataout_comp(struct ib_cq *cq, struct ib_wc *wc);
+void iser_reg_comp(struct ib_cq *cq, struct ib_wc *wc);
+void iser_last_comp(struct ib_cq *cq, struct ib_wc *wc);
 
 void iser_task_rdma_init(struct iscsi_iser_task *task);
 
@@ -719,4 +729,28 @@ iser_tx_next_wr(struct iser_tx_desc *tx_desc)
 	return cur_wr;
 }
 
+static inline struct iser_conn *
+to_iser_conn(struct ib_conn *ib_conn)
+{
+	return container_of(ib_conn, struct iser_conn, ib_conn);
+}
+
+static inline struct iser_rx_desc *
+iser_rx(struct ib_cqe *cqe)
+{
+	return container_of(cqe, struct iser_rx_desc, cqe);
+}
+
+static inline struct iser_tx_desc *
+iser_tx(struct ib_cqe *cqe)
+{
+	return container_of(cqe, struct iser_tx_desc, cqe);
+}
+
+static inline struct iser_login_desc *
+iser_login(struct ib_cqe *cqe)
+{
+	return container_of(cqe, struct iser_login_desc, cqe);
+}
+
 #endif

drivers/infiniband/ulp/iser/iser_memory.c

@@ -67,6 +67,11 @@ static struct iser_reg_ops fmr_ops = {
 	.reg_desc_put	= iser_reg_desc_put_fmr,
 };
 
+void iser_reg_comp(struct ib_cq *cq, struct ib_wc *wc)
+{
+	iser_err_comp(wc, "memreg");
+}
+
 int iser_assign_reg_ops(struct iser_device *device)
 {
 	struct ib_device_attr *dev_attr = &device->dev_attr;
@@ -414,12 +419,14 @@ iser_set_prot_checks(struct scsi_cmnd *sc, u8 *mask)
 }
 
 static void
-iser_inv_rkey(struct ib_send_wr *inv_wr, struct ib_mr *mr)
+iser_inv_rkey(struct ib_send_wr *inv_wr,
+	      struct ib_mr *mr,
+	      struct ib_cqe *cqe)
 {
 	u32 rkey;
 
 	inv_wr->opcode = IB_WR_LOCAL_INV;
-	inv_wr->wr_id = ISER_FASTREG_LI_WRID;
+	inv_wr->wr_cqe = cqe;
 	inv_wr->ex.invalidate_rkey = mr->rkey;
 	inv_wr->send_flags = 0;
 	inv_wr->num_sge = 0;
@@ -437,6 +444,7 @@ iser_reg_sig_mr(struct iscsi_iser_task *iser_task,
 {
 	struct iser_tx_desc *tx_desc = &iser_task->desc;
 	struct ib_sig_attrs *sig_attrs = &tx_desc->sig_attrs;
+	struct ib_cqe *cqe = &iser_task->iser_conn->ib_conn.reg_cqe;
 	struct ib_sig_handover_wr *wr;
 	int ret;
 
@@ -448,11 +456,11 @@ iser_reg_sig_mr(struct iscsi_iser_task *iser_task,
 	iser_set_prot_checks(iser_task->sc, &sig_attrs->check_mask);
 
 	if (!pi_ctx->sig_mr_valid)
-		iser_inv_rkey(iser_tx_next_wr(tx_desc), pi_ctx->sig_mr);
+		iser_inv_rkey(iser_tx_next_wr(tx_desc), pi_ctx->sig_mr, cqe);
 
 	wr = sig_handover_wr(iser_tx_next_wr(tx_desc));
 	wr->wr.opcode = IB_WR_REG_SIG_MR;
-	wr->wr.wr_id = ISER_FASTREG_LI_WRID;
+	wr->wr.wr_cqe = cqe;
 	wr->wr.sg_list = &data_reg->sge;
 	wr->wr.num_sge = 1;
 	wr->wr.send_flags = 0;
@@ -485,12 +493,13 @@ static int iser_fast_reg_mr(struct iscsi_iser_task *iser_task,
 			    struct iser_mem_reg *reg)
 {
 	struct iser_tx_desc *tx_desc = &iser_task->desc;
+	struct ib_cqe *cqe = &iser_task->iser_conn->ib_conn.reg_cqe;
 	struct ib_mr *mr = rsc->mr;
 	struct ib_reg_wr *wr;
 	int n;
 
 	if (!rsc->mr_valid)
-		iser_inv_rkey(iser_tx_next_wr(tx_desc), mr);
+		iser_inv_rkey(iser_tx_next_wr(tx_desc), mr, cqe);
 
 	n = ib_map_mr_sg(mr, mem->sg, mem->size, SIZE_4K);
 	if (unlikely(n != mem->size)) {
@@ -501,7 +510,7 @@ static int iser_fast_reg_mr(struct iscsi_iser_task *iser_task,
 
 	wr = reg_wr(iser_tx_next_wr(tx_desc));
 	wr->wr.opcode = IB_WR_REG_MR;
-	wr->wr.wr_id = ISER_FASTREG_LI_WRID;
+	wr->wr.wr_cqe = cqe;
 	wr->wr.send_flags = 0;
 	wr->wr.num_sge = 0;
 	wr->mr = mr;

drivers/infiniband/ulp/srp/ib_srp.h

@@ -66,11 +66,6 @@ enum {
 	SRP_TAG_TSK_MGMT	= 1U << 31,
 
 	SRP_MAX_PAGES_PER_MR	= 512,
-
-	LOCAL_INV_WR_ID_MASK	= 1,
-	FAST_REG_WR_ID_MASK	= 2,
-
-	SRP_LAST_WR_ID		= 0xfffffffcU,
 };
 
 enum srp_target_state {
@@ -128,6 +123,7 @@ struct srp_request {
 	struct srp_direct_buf  *indirect_desc;
 	dma_addr_t		indirect_dma_addr;
 	short			nmdesc;
+	struct ib_cqe		reg_cqe;
 };
 
 /**
@@ -231,6 +227,7 @@ struct srp_iu {
 	void		       *buf;
 	size_t			size;
 	enum dma_data_direction	direction;
+	struct ib_cqe		cqe;
 };
 
 /**

drivers/infiniband/ulp/srpt/ib_srpt.h

@@ -128,36 +128,6 @@ enum {
 	DEFAULT_MAX_RDMA_SIZE = 65536,
 };
 
-enum srpt_opcode {
-	SRPT_RECV,
-	SRPT_SEND,
-	SRPT_RDMA_MID,
-	SRPT_RDMA_ABORT,
-	SRPT_RDMA_READ_LAST,
-	SRPT_RDMA_WRITE_LAST,
-};
-
-static inline u64 encode_wr_id(u8 opcode, u32 idx)
-{
-	return ((u64)opcode << 32) | idx;
-}
-static inline enum srpt_opcode opcode_from_wr_id(u64 wr_id)
-{
-	return wr_id >> 32;
-}
-static inline u32 idx_from_wr_id(u64 wr_id)
-{
-	return (u32)wr_id;
-}
-
-struct rdma_iu {
-	u64		raddr;
-	u32		rkey;
-	struct ib_sge	*sge;
-	u32		sge_cnt;
-	int		mem_id;
-};
-
 /**
  * enum srpt_command_state - SCSI command state managed by SRPT.
  * @SRPT_STATE_NEW:           New command arrived and is being processed.
@@ -189,6 +159,7 @@ enum srpt_command_state {
  * @index: Index of the I/O context in its ioctx_ring array.
  */
 struct srpt_ioctx {
+	struct ib_cqe		cqe;
 	void			*buf;
 	dma_addr_t		dma;
 	uint32_t		index;
@@ -215,32 +186,30 @@ struct srpt_recv_ioctx {
  * @sg:          Pointer to sg-list associated with this I/O context.
  * @sg_cnt:      SG-list size.
  * @mapped_sg_count: ib_dma_map_sg() return value.
- * @n_rdma_ius:  Number of elements in the rdma_ius array.
- * @rdma_ius:    Array with information about the RDMA mapping.
+ * @n_rdma_wrs:  Number of elements in the rdma_wrs array.
+ * @rdma_wrs:    Array with information about the RDMA mapping.
  * @tag:         Tag of the received SRP information unit.
  * @spinlock:    Protects 'state'.
  * @state:       I/O context state.
- * @rdma_aborted: If initiating a multipart RDMA transfer failed, whether
- * 		 the already initiated transfers have finished.
  * @cmd:         Target core command data structure.
  * @sense_data:  SCSI sense data.
  */
 struct srpt_send_ioctx {
 	struct srpt_ioctx	ioctx;
 	struct srpt_rdma_ch	*ch;
-	struct rdma_iu		*rdma_ius;
+	struct ib_rdma_wr	*rdma_wrs;
+	struct ib_cqe		rdma_cqe;
 	struct srp_direct_buf	*rbufs;
 	struct srp_direct_buf	single_rbuf;
 	struct scatterlist	*sg;
 	struct list_head	free_list;
 	spinlock_t		spinlock;
 	enum srpt_command_state	state;
-	bool			rdma_aborted;
 	struct se_cmd		cmd;
 	struct completion	tx_done;
 	int			sg_cnt;
 	int			mapped_sg_count;
-	u16			n_rdma_ius;
+	u16			n_rdma_wrs;
 	u8			n_rdma;
 	u8			n_rbuf;
 	bool			queue_status_only;
@@ -267,9 +236,6 @@ enum rdma_ch_state {
 
 /**
  * struct srpt_rdma_ch - RDMA channel.
- * @wait_queue:    Allows the kernel thread to wait for more work.
- * @thread:        Kernel thread that processes the IB queues associated with
- *                 the channel.
  * @cm_id:         IB CM ID associated with the channel.
  * @qp:            IB queue pair used for communicating over this channel.
  * @cq:            IB completion queue for this channel.
@@ -299,8 +265,6 @@ enum rdma_ch_state {
  * @release_done:  Enables waiting for srpt_release_channel() completion.
  */
 struct srpt_rdma_ch {
-	wait_queue_head_t	wait_queue;
-	struct task_struct	*thread;
 	struct ib_cm_id		*cm_id;
 	struct ib_qp		*qp;
 	struct ib_cq		*cq;

drivers/scsi/Kconfig

@@ -1104,6 +1104,7 @@ config SCSI_IPR
 	tristate "IBM Power Linux RAID adapter support"
 	depends on PCI && SCSI && ATA
 	select FW_LOADER
+	select IRQ_POLL
 	---help---
 	  This driver supports the IBM Power Linux family RAID adapters.
 	  This includes IBM pSeries 5712, 5703, 5709, and 570A, as well

drivers/scsi/be2iscsi/Kconfig

@@ -3,6 +3,7 @@ config BE2ISCSI
 	depends on PCI && SCSI && NET
 	select SCSI_ISCSI_ATTRS
 	select ISCSI_BOOT_SYSFS
+	select IRQ_POLL
 
 	help
 	This driver implements the iSCSI functionality for Emulex

drivers/scsi/be2iscsi/be.h

@@ -20,7 +20,7 @@
 
 #include <linux/pci.h>
 #include <linux/if_vlan.h>
-#include <linux/blk-iopoll.h>
+#include <linux/irq_poll.h>
 #define FW_VER_LEN	32
 #define MCC_Q_LEN	128
 #define MCC_CQ_LEN	256
@@ -101,7 +101,7 @@ struct be_eq_obj {
 	struct beiscsi_hba *phba;
 	struct be_queue_info *cq;
 	struct work_struct work_cqs; /* Work Item */
-	struct blk_iopoll	iopoll;
+	struct irq_poll	iopoll;
 };
 
 struct be_mcc_obj {

drivers/scsi/be2iscsi/be_iscsi.c

@@ -1292,9 +1292,9 @@ static void beiscsi_flush_cq(struct beiscsi_hba *phba)
 
 	for (i = 0; i < phba->num_cpus; i++) {
 		pbe_eq = &phwi_context->be_eq[i];
-		blk_iopoll_disable(&pbe_eq->iopoll);
+		irq_poll_disable(&pbe_eq->iopoll);
 		beiscsi_process_cq(pbe_eq);
-		blk_iopoll_enable(&pbe_eq->iopoll);
+		irq_poll_enable(&pbe_eq->iopoll);
 	}
 }
 

drivers/scsi/be2iscsi/be_main.c

@@ -910,8 +910,7 @@ static irqreturn_t be_isr_msix(int irq, void *dev_id)
 	num_eq_processed = 0;
 	while (eqe->dw[offsetof(struct amap_eq_entry, valid) / 32]
 				& EQE_VALID_MASK) {
-		if (!blk_iopoll_sched_prep(&pbe_eq->iopoll))
-			blk_iopoll_sched(&pbe_eq->iopoll);
+		irq_poll_sched(&pbe_eq->iopoll);
 
 		AMAP_SET_BITS(struct amap_eq_entry, valid, eqe, 0);
 		queue_tail_inc(eq);
@@ -972,8 +971,7 @@ static irqreturn_t be_isr(int irq, void *dev_id)
 			spin_unlock_irqrestore(&phba->isr_lock, flags);
 			num_mcceq_processed++;
 		} else {
-			if (!blk_iopoll_sched_prep(&pbe_eq->iopoll))
-				blk_iopoll_sched(&pbe_eq->iopoll);
+			irq_poll_sched(&pbe_eq->iopoll);
 			num_ioeq_processed++;
 		}
 		AMAP_SET_BITS(struct amap_eq_entry, valid, eqe, 0);
@@ -2295,7 +2293,7 @@ void beiscsi_process_all_cqs(struct work_struct *work)
 	hwi_ring_eq_db(phba, pbe_eq->q.id, 0, 0, 1, 1);
 }
 
-static int be_iopoll(struct blk_iopoll *iop, int budget)
+static int be_iopoll(struct irq_poll *iop, int budget)
 {
 	unsigned int ret;
 	struct beiscsi_hba *phba;
@@ -2306,7 +2304,7 @@ static int be_iopoll(struct blk_iopoll *iop, int budget)
 	pbe_eq->cq_count += ret;
 	if (ret < budget) {
 		phba = pbe_eq->phba;
-		blk_iopoll_complete(iop);
+		irq_poll_complete(iop);
 		beiscsi_log(phba, KERN_INFO,
 			    BEISCSI_LOG_CONFIG | BEISCSI_LOG_IO,
 			    "BM_%d : rearm pbe_eq->q.id =%d\n",
@@ -5293,7 +5291,7 @@ static void beiscsi_quiesce(struct beiscsi_hba *phba,
 
 	for (i = 0; i < phba->num_cpus; i++) {
 		pbe_eq = &phwi_context->be_eq[i];
-		blk_iopoll_disable(&pbe_eq->iopoll);
+		irq_poll_disable(&pbe_eq->iopoll);
 	}
 
 	if (unload_state == BEISCSI_CLEAN_UNLOAD) {
@@ -5579,9 +5577,8 @@ static void beiscsi_eeh_resume(struct pci_dev *pdev)
 
 	for (i = 0; i < phba->num_cpus; i++) {
 		pbe_eq = &phwi_context->be_eq[i];
-		blk_iopoll_init(&pbe_eq->iopoll, be_iopoll_budget,
+		irq_poll_init(&pbe_eq->iopoll, be_iopoll_budget,
 				be_iopoll);
-		blk_iopoll_enable(&pbe_eq->iopoll);
 	}
 
 	i = (phba->msix_enabled) ? i : 0;
@@ -5752,9 +5749,8 @@ static int beiscsi_dev_probe(struct pci_dev *pcidev,
 
 	for (i = 0; i < phba->num_cpus; i++) {
 		pbe_eq = &phwi_context->be_eq[i];
-		blk_iopoll_init(&pbe_eq->iopoll, be_iopoll_budget,
+		irq_poll_init(&pbe_eq->iopoll, be_iopoll_budget,
 				be_iopoll);
-		blk_iopoll_enable(&pbe_eq->iopoll);
 	}
 
 	i = (phba->msix_enabled) ? i : 0;
@@ -5795,7 +5791,7 @@ static int beiscsi_dev_probe(struct pci_dev *pcidev,
 	destroy_workqueue(phba->wq);
 	for (i = 0; i < phba->num_cpus; i++) {
 		pbe_eq = &phwi_context->be_eq[i];
-		blk_iopoll_disable(&pbe_eq->iopoll);
+		irq_poll_disable(&pbe_eq->iopoll);
 	}
 free_twq:
 	beiscsi_clean_port(phba);

drivers/scsi/ipr.c

@@ -3638,7 +3638,7 @@ static struct device_attribute ipr_ioa_reset_attr = {
 	.store = ipr_store_reset_adapter
 };
 
-static int ipr_iopoll(struct blk_iopoll *iop, int budget);
+static int ipr_iopoll(struct irq_poll *iop, int budget);
  /**
  * ipr_show_iopoll_weight - Show ipr polling mode
  * @dev:	class device struct
@@ -3681,34 +3681,33 @@ static ssize_t ipr_store_iopoll_weight(struct device *dev,
 	int i;
 
 	if (!ioa_cfg->sis64) {
-		dev_info(&ioa_cfg->pdev->dev, "blk-iopoll not supported on this adapter\n");
+		dev_info(&ioa_cfg->pdev->dev, "irq_poll not supported on this adapter\n");
 		return -EINVAL;
 	}
 	if (kstrtoul(buf, 10, &user_iopoll_weight))
 		return -EINVAL;
 
 	if (user_iopoll_weight > 256) {
-		dev_info(&ioa_cfg->pdev->dev, "Invalid blk-iopoll weight. It must be less than 256\n");
+		dev_info(&ioa_cfg->pdev->dev, "Invalid irq_poll weight. It must be less than 256\n");
 		return -EINVAL;
 	}
 
 	if (user_iopoll_weight == ioa_cfg->iopoll_weight) {
-		dev_info(&ioa_cfg->pdev->dev, "Current blk-iopoll weight has the same weight\n");
+		dev_info(&ioa_cfg->pdev->dev, "Current irq_poll weight has the same weight\n");
 		return strlen(buf);
 	}
 
 	if (ioa_cfg->iopoll_weight && ioa_cfg->sis64 && ioa_cfg->nvectors > 1) {
 		for (i = 1; i < ioa_cfg->hrrq_num; i++)
-			blk_iopoll_disable(&ioa_cfg->hrrq[i].iopoll);
+			irq_poll_disable(&ioa_cfg->hrrq[i].iopoll);
 	}
 
 	spin_lock_irqsave(shost->host_lock, lock_flags);
 	ioa_cfg->iopoll_weight = user_iopoll_weight;
 	if (ioa_cfg->iopoll_weight && ioa_cfg->sis64 && ioa_cfg->nvectors > 1) {
 		for (i = 1; i < ioa_cfg->hrrq_num; i++) {
-			blk_iopoll_init(&ioa_cfg->hrrq[i].iopoll,
+			irq_poll_init(&ioa_cfg->hrrq[i].iopoll,
 					ioa_cfg->iopoll_weight, ipr_iopoll);
-			blk_iopoll_enable(&ioa_cfg->hrrq[i].iopoll);
 		}
 	}
 	spin_unlock_irqrestore(shost->host_lock, lock_flags);
@@ -5569,7 +5568,7 @@ static int ipr_process_hrrq(struct ipr_hrr_queue *hrr_queue, int budget,
 	return num_hrrq;
 }
 
-static int ipr_iopoll(struct blk_iopoll *iop, int budget)
+static int ipr_iopoll(struct irq_poll *iop, int budget)
 {
 	struct ipr_ioa_cfg *ioa_cfg;
 	struct ipr_hrr_queue *hrrq;
@@ -5585,7 +5584,7 @@ static int ipr_iopoll(struct blk_iopoll *iop, int budget)
 	completed_ops = ipr_process_hrrq(hrrq, budget, &doneq);
 
 	if (completed_ops < budget)
-		blk_iopoll_complete(iop);
+		irq_poll_complete(iop);
 	spin_unlock_irqrestore(hrrq->lock, hrrq_flags);
 
 	list_for_each_entry_safe(ipr_cmd, temp, &doneq, queue) {
@@ -5693,8 +5692,7 @@ static irqreturn_t ipr_isr_mhrrq(int irq, void *devp)
 	if (ioa_cfg->iopoll_weight && ioa_cfg->sis64 && ioa_cfg->nvectors > 1) {
 		if ((be32_to_cpu(*hrrq->hrrq_curr) & IPR_HRRQ_TOGGLE_BIT) ==
 		       hrrq->toggle_bit) {
-			if (!blk_iopoll_sched_prep(&hrrq->iopoll))
-				blk_iopoll_sched(&hrrq->iopoll);
+			irq_poll_sched(&hrrq->iopoll);
 			spin_unlock_irqrestore(hrrq->lock, hrrq_flags);
 			return IRQ_HANDLED;
 		}
@@ -10405,9 +10403,8 @@ static int ipr_probe(struct pci_dev *pdev, const struct pci_device_id *dev_id)
 
 	if (ioa_cfg->iopoll_weight && ioa_cfg->sis64 && ioa_cfg->nvectors > 1) {
 		for (i = 1; i < ioa_cfg->hrrq_num; i++) {
-			blk_iopoll_init(&ioa_cfg->hrrq[i].iopoll,
+			irq_poll_init(&ioa_cfg->hrrq[i].iopoll,
 					ioa_cfg->iopoll_weight, ipr_iopoll);
-			blk_iopoll_enable(&ioa_cfg->hrrq[i].iopoll);
 		}
 	}
 
@@ -10436,7 +10433,7 @@ static void ipr_shutdown(struct pci_dev *pdev)
 	if (ioa_cfg->iopoll_weight && ioa_cfg->sis64 && ioa_cfg->nvectors > 1) {
 		ioa_cfg->iopoll_weight = 0;
 		for (i = 1; i < ioa_cfg->hrrq_num; i++)
-			blk_iopoll_disable(&ioa_cfg->hrrq[i].iopoll);
+			irq_poll_disable(&ioa_cfg->hrrq[i].iopoll);
 	}
 
 	while (ioa_cfg->in_reset_reload) {

drivers/scsi/ipr.h

@@ -32,7 +32,7 @@
 #include <linux/libata.h>
 #include <linux/list.h>
 #include <linux/kref.h>
-#include <linux/blk-iopoll.h>
+#include <linux/irq_poll.h>
 #include <scsi/scsi.h>
 #include <scsi/scsi_cmnd.h>
 
@@ -517,7 +517,7 @@ struct ipr_hrr_queue {
 	u8 allow_cmds:1;
 	u8 removing_ioa:1;
 
-	struct blk_iopoll iopoll;
+	struct irq_poll iopoll;
 };
 
 /* Command packet structure */

include/linux/blk-iopoll.h

-#ifndef BLK_IOPOLL_H
-#define BLK_IOPOLL_H
-
-struct blk_iopoll;
-typedef int (blk_iopoll_fn)(struct blk_iopoll *, int);
-
-struct blk_iopoll {
-	struct list_head list;
-	unsigned long state;
-	unsigned long data;
-	int weight;
-	int max;
-	blk_iopoll_fn *poll;
-};
-
-enum {
-	IOPOLL_F_SCHED		= 0,
-	IOPOLL_F_DISABLE	= 1,
-};
-
-/*
- * Returns 0 if we successfully set the IOPOLL_F_SCHED bit, indicating
- * that we were the first to acquire this iop for scheduling. If this iop
- * is currently disabled, return "failure".
- */
-static inline int blk_iopoll_sched_prep(struct blk_iopoll *iop)
-{
-	if (!test_bit(IOPOLL_F_DISABLE, &iop->state))
-		return test_and_set_bit(IOPOLL_F_SCHED, &iop->state);
-
-	return 1;
-}
-
-static inline int blk_iopoll_disable_pending(struct blk_iopoll *iop)
-{
-	return test_bit(IOPOLL_F_DISABLE, &iop->state);
-}
-
-extern void blk_iopoll_sched(struct blk_iopoll *);
-extern void blk_iopoll_init(struct blk_iopoll *, int, blk_iopoll_fn *);
-extern void blk_iopoll_complete(struct blk_iopoll *);
-extern void __blk_iopoll_complete(struct blk_iopoll *);
-extern void blk_iopoll_enable(struct blk_iopoll *);
-extern void blk_iopoll_disable(struct blk_iopoll *);
-
-#endif

include/linux/interrupt.h

@@ -412,7 +412,7 @@ enum
 	NET_TX_SOFTIRQ,
 	NET_RX_SOFTIRQ,
 	BLOCK_SOFTIRQ,
-	BLOCK_IOPOLL_SOFTIRQ,
+	IRQ_POLL_SOFTIRQ,
 	TASKLET_SOFTIRQ,
 	SCHED_SOFTIRQ,
 	HRTIMER_SOFTIRQ, /* Unused, but kept as tools rely on the

include/linux/irq_poll.h

+#ifndef IRQ_POLL_H
+#define IRQ_POLL_H
+
+struct irq_poll;
+typedef int (irq_poll_fn)(struct irq_poll *, int);
+
+struct irq_poll {
+	struct list_head list;
+	unsigned long state;
+	int weight;
+	irq_poll_fn *poll;
+};
+
+enum {
+	IRQ_POLL_F_SCHED	= 0,
+	IRQ_POLL_F_DISABLE	= 1,
+};
+
+extern void irq_poll_sched(struct irq_poll *);
+extern void irq_poll_init(struct irq_poll *, int, irq_poll_fn *);
+extern void irq_poll_complete(struct irq_poll *);
+extern void irq_poll_enable(struct irq_poll *);
+extern void irq_poll_disable(struct irq_poll *);
+
+#endif

include/rdma/ib_verbs.h

@@ -49,6 +49,7 @@
 #include <linux/scatterlist.h>
 #include <linux/workqueue.h>
 #include <linux/socket.h>
+#include <linux/irq_poll.h>
 #include <uapi/linux/if_ether.h>
 
 #include <linux/atomic.h>
@@ -56,6 +57,7 @@
 #include <asm/uaccess.h>
 
 extern struct workqueue_struct *ib_wq;
+extern struct workqueue_struct *ib_comp_wq;
 
 union ib_gid {
 	u8	raw[16];
@@ -758,7 +760,10 @@ enum ib_wc_flags {
 };
 
 struct ib_wc {
+	union {
 		u64		wr_id;
+		struct ib_cqe	*wr_cqe;
+	};
 	enum ib_wc_status	status;
 	enum ib_wc_opcode	opcode;
 	u32			vendor_err;
@@ -1079,9 +1084,16 @@ struct ib_mw_bind_info {
 	int		mw_access_flags;
 };
 
+struct ib_cqe {
+	void (*done)(struct ib_cq *cq, struct ib_wc *wc);
+};
+
 struct ib_send_wr {
 	struct ib_send_wr      *next;
+	union {
 		u64		wr_id;
+		struct ib_cqe	*wr_cqe;
+	};
 	struct ib_sge	       *sg_list;
 	int			num_sge;
 	enum ib_wr_opcode	opcode;
@@ -1175,7 +1187,10 @@ static inline struct ib_sig_handover_wr *sig_handover_wr(struct ib_send_wr *wr)
 
 struct ib_recv_wr {
 	struct ib_recv_wr      *next;
+	union {
 		u64		wr_id;
+		struct ib_cqe	*wr_cqe;
+	};
 	struct ib_sge	       *sg_list;
 	int			num_sge;
 };
@@ -1307,6 +1322,12 @@ struct ib_ah {
 
 typedef void (*ib_comp_handler)(struct ib_cq *cq, void *cq_context);
 
+enum ib_poll_context {
+	IB_POLL_DIRECT,		/* caller context, no hw completions */
+	IB_POLL_SOFTIRQ,	/* poll from softirq context */
+	IB_POLL_WORKQUEUE,	/* poll from workqueue */
+};
+
 struct ib_cq {
 	struct ib_device       *device;
 	struct ib_uobject      *uobject;
@@ -1315,6 +1336,12 @@ struct ib_cq {
 	void                   *cq_context;
 	int               	cqe;
 	atomic_t          	usecnt; /* count number of work queues */
+	enum ib_poll_context	poll_ctx;
+	struct ib_wc		*wc;
+	union {
+		struct irq_poll		iop;
+		struct work_struct	work;
+	};
 };
 
 struct ib_srq {
@@ -2454,6 +2481,11 @@ static inline int ib_post_recv(struct ib_qp *qp,
 	return qp->device->post_recv(qp, recv_wr, bad_recv_wr);
 }
 
+struct ib_cq *ib_alloc_cq(struct ib_device *dev, void *private,
+		int nr_cqe, int comp_vector, enum ib_poll_context poll_ctx);
+void ib_free_cq(struct ib_cq *cq);
+int ib_process_cq_direct(struct ib_cq *cq, int budget);
+
 /**
  * ib_create_cq - Creates a CQ on the specified device.
  * @device: The device on which to create the CQ.

include/trace/events/irq.h

@@ -15,7 +15,7 @@ struct softirq_action;
 			 softirq_name(NET_TX)		\
 			 softirq_name(NET_RX)		\
 			 softirq_name(BLOCK)		\
-			 softirq_name(BLOCK_IOPOLL)	\
+			 softirq_name(IRQ_POLL)		\
 			 softirq_name(TASKLET)		\
 			 softirq_name(SCHED)		\
 			 softirq_name(HRTIMER)		\

lib/Kconfig

@@ -475,6 +475,11 @@ config DDR
 	  information. This data is useful for drivers handling
 	  DDR SDRAM controllers.
 
+config IRQ_POLL
+	bool "IRQ polling library"
+	help
+	  Helper library to poll interrupt mitigation using polling.
+
 config MPILIB
 	tristate
 	select CLZ_TAB

lib/Makefile

@@ -164,6 +164,7 @@ obj-$(CONFIG_GENERIC_NET_UTILS) += net_utils.o
 
 obj-$(CONFIG_SG_SPLIT) += sg_split.o
 obj-$(CONFIG_STMP_DEVICE) += stmp_device.o
+obj-$(CONFIG_IRQ_POLL) += irq_poll.o
 
 libfdt_files = fdt.o fdt_ro.o fdt_wip.o fdt_rw.o fdt_sw.o fdt_strerror.o \
 	       fdt_empty_tree.o

block/blk-iopoll.c → lib/irq_poll.c

@@ -6,84 +6,84 @@
 #include <linux/module.h>
 #include <linux/init.h>
 #include <linux/bio.h>
-#include <linux/blkdev.h>
 #include <linux/interrupt.h>
 #include <linux/cpu.h>
-#include <linux/blk-iopoll.h>
+#include <linux/irq_poll.h>
 #include <linux/delay.h>
 
-#include "blk.h"
-
-static unsigned int blk_iopoll_budget __read_mostly = 256;
+static unsigned int irq_poll_budget __read_mostly = 256;
 
 static DEFINE_PER_CPU(struct list_head, blk_cpu_iopoll);
 
 /**
- * blk_iopoll_sched - Schedule a run of the iopoll handler
+ * irq_poll_sched - Schedule a run of the iopoll handler
  * @iop:      The parent iopoll structure
  *
  * Description:
- *     Add this blk_iopoll structure to the pending poll list and trigger the
- *     raise of the blk iopoll softirq. The driver must already have gotten a
- *     successful return from blk_iopoll_sched_prep() before calling this.
+ *     Add this irq_poll structure to the pending poll list and trigger the
+ *     raise of the blk iopoll softirq.
  **/
-void blk_iopoll_sched(struct blk_iopoll *iop)
+void irq_poll_sched(struct irq_poll *iop)
 {
 	unsigned long flags;
 
+	if (test_bit(IRQ_POLL_F_DISABLE, &iop->state))
+		return;
+	if (!test_and_set_bit(IRQ_POLL_F_SCHED, &iop->state))
+		return;
+
 	local_irq_save(flags);
 	list_add_tail(&iop->list, this_cpu_ptr(&blk_cpu_iopoll));
-	__raise_softirq_irqoff(BLOCK_IOPOLL_SOFTIRQ);
+	__raise_softirq_irqoff(IRQ_POLL_SOFTIRQ);
 	local_irq_restore(flags);
 }
-EXPORT_SYMBOL(blk_iopoll_sched);
+EXPORT_SYMBOL(irq_poll_sched);
 
 /**
- * __blk_iopoll_complete - Mark this @iop as un-polled again
+ * __irq_poll_complete - Mark this @iop as un-polled again
  * @iop:      The parent iopoll structure
  *
  * Description:
- *     See blk_iopoll_complete(). This function must be called with interrupts
+ *     See irq_poll_complete(). This function must be called with interrupts
  *     disabled.
  **/
-void __blk_iopoll_complete(struct blk_iopoll *iop)
+static void __irq_poll_complete(struct irq_poll *iop)
 {
 	list_del(&iop->list);
 	smp_mb__before_atomic();
-	clear_bit_unlock(IOPOLL_F_SCHED, &iop->state);
+	clear_bit_unlock(IRQ_POLL_F_SCHED, &iop->state);
 }
-EXPORT_SYMBOL(__blk_iopoll_complete);
 
 /**
- * blk_iopoll_complete - Mark this @iop as un-polled again
+ * irq_poll_complete - Mark this @iop as un-polled again
  * @iop:      The parent iopoll structure
  *
  * Description:
  *     If a driver consumes less than the assigned budget in its run of the
  *     iopoll handler, it'll end the polled mode by calling this function. The
- *     iopoll handler will not be invoked again before blk_iopoll_sched_prep()
+ *     iopoll handler will not be invoked again before irq_poll_sched()
  *     is called.
  **/
-void blk_iopoll_complete(struct blk_iopoll *iop)
+void irq_poll_complete(struct irq_poll *iop)
 {
 	unsigned long flags;
 
 	local_irq_save(flags);
-	__blk_iopoll_complete(iop);
+	__irq_poll_complete(iop);
 	local_irq_restore(flags);
 }
-EXPORT_SYMBOL(blk_iopoll_complete);
+EXPORT_SYMBOL(irq_poll_complete);
 
-static void blk_iopoll_softirq(struct softirq_action *h)
+static void irq_poll_softirq(struct softirq_action *h)
 {
 	struct list_head *list = this_cpu_ptr(&blk_cpu_iopoll);
-	int rearm = 0, budget = blk_iopoll_budget;
+	int rearm = 0, budget = irq_poll_budget;
 	unsigned long start_time = jiffies;
 
 	local_irq_disable();
 
 	while (!list_empty(list)) {
-		struct blk_iopoll *iop;
+		struct irq_poll *iop;
 		int work, weight;
 
 		/*
@@ -101,11 +101,11 @@ static void blk_iopoll_softirq(struct softirq_action *h)
 		 * entries to the tail of this list, and only ->poll()
 		 * calls can remove this head entry from the list.
 		 */
-		iop = list_entry(list->next, struct blk_iopoll, list);
+		iop = list_entry(list->next, struct irq_poll, list);
 
 		weight = iop->weight;
 		work = 0;
-		if (test_bit(IOPOLL_F_SCHED, &iop->state))
+		if (test_bit(IRQ_POLL_F_SCHED, &iop->state))
 			work = iop->poll(iop, weight);
 
 		budget -= work;
@@ -121,72 +121,70 @@ static void blk_iopoll_softirq(struct softirq_action *h)
 		 * move the instance around on the list at-will.
 		 */
 		if (work >= weight) {
-			if (blk_iopoll_disable_pending(iop))
-				__blk_iopoll_complete(iop);
+			if (test_bit(IRQ_POLL_F_DISABLE, &iop->state))
+				__irq_poll_complete(iop);
 			else
 				list_move_tail(&iop->list, list);
 		}
 	}
 
 	if (rearm)
-		__raise_softirq_irqoff(BLOCK_IOPOLL_SOFTIRQ);
+		__raise_softirq_irqoff(IRQ_POLL_SOFTIRQ);
 
 	local_irq_enable();
 }
 
 /**
- * blk_iopoll_disable - Disable iopoll on this @iop
+ * irq_poll_disable - Disable iopoll on this @iop
  * @iop:      The parent iopoll structure
  *
  * Description:
  *     Disable io polling and wait for any pending callbacks to have completed.
  **/
-void blk_iopoll_disable(struct blk_iopoll *iop)
+void irq_poll_disable(struct irq_poll *iop)
 {
-	set_bit(IOPOLL_F_DISABLE, &iop->state);
-	while (test_and_set_bit(IOPOLL_F_SCHED, &iop->state))
+	set_bit(IRQ_POLL_F_DISABLE, &iop->state);
+	while (test_and_set_bit(IRQ_POLL_F_SCHED, &iop->state))
 		msleep(1);
-	clear_bit(IOPOLL_F_DISABLE, &iop->state);
+	clear_bit(IRQ_POLL_F_DISABLE, &iop->state);
 }
-EXPORT_SYMBOL(blk_iopoll_disable);
+EXPORT_SYMBOL(irq_poll_disable);
 
 /**
- * blk_iopoll_enable - Enable iopoll on this @iop
+ * irq_poll_enable - Enable iopoll on this @iop
  * @iop:      The parent iopoll structure
  *
  * Description:
  *     Enable iopoll on this @iop. Note that the handler run will not be
  *     scheduled, it will only mark it as active.
  **/
-void blk_iopoll_enable(struct blk_iopoll *iop)
+void irq_poll_enable(struct irq_poll *iop)
 {
-	BUG_ON(!test_bit(IOPOLL_F_SCHED, &iop->state));
+	BUG_ON(!test_bit(IRQ_POLL_F_SCHED, &iop->state));
 	smp_mb__before_atomic();
-	clear_bit_unlock(IOPOLL_F_SCHED, &iop->state);
+	clear_bit_unlock(IRQ_POLL_F_SCHED, &iop->state);
 }
-EXPORT_SYMBOL(blk_iopoll_enable);
+EXPORT_SYMBOL(irq_poll_enable);
 
 /**
- * blk_iopoll_init - Initialize this @iop
+ * irq_poll_init - Initialize this @iop
  * @iop:      The parent iopoll structure
  * @weight:   The default weight (or command completion budget)
  * @poll_fn:  The handler to invoke
  *
  * Description:
- *     Initialize this blk_iopoll structure. Before being actively used, the
- *     driver must call blk_iopoll_enable().
+ *     Initialize and enable this irq_poll structure.
  **/
-void blk_iopoll_init(struct blk_iopoll *iop, int weight, blk_iopoll_fn *poll_fn)
+void irq_poll_init(struct irq_poll *iop, int weight, irq_poll_fn *poll_fn)
 {
 	memset(iop, 0, sizeof(*iop));
 	INIT_LIST_HEAD(&iop->list);
 	iop->weight = weight;
 	iop->poll = poll_fn;
-	set_bit(IOPOLL_F_SCHED, &iop->state);
 }
-EXPORT_SYMBOL(blk_iopoll_init);
+EXPORT_SYMBOL(irq_poll_init);
 
-static int blk_iopoll_cpu_notify(struct notifier_block *self,
+static int irq_poll_cpu_notify(struct notifier_block *self,
 				 unsigned long action, void *hcpu)
 {
 	/*
@@ -199,26 +197,26 @@ static int blk_iopoll_cpu_notify(struct notifier_block *self,
 		local_irq_disable();
 		list_splice_init(&per_cpu(blk_cpu_iopoll, cpu),
 				 this_cpu_ptr(&blk_cpu_iopoll));
-		__raise_softirq_irqoff(BLOCK_IOPOLL_SOFTIRQ);
+		__raise_softirq_irqoff(IRQ_POLL_SOFTIRQ);
 		local_irq_enable();
 	}
 
 	return NOTIFY_OK;
 }
 
-static struct notifier_block blk_iopoll_cpu_notifier = {
-	.notifier_call	= blk_iopoll_cpu_notify,
+static struct notifier_block irq_poll_cpu_notifier = {
+	.notifier_call	= irq_poll_cpu_notify,
 };
 
-static __init int blk_iopoll_setup(void)
+static __init int irq_poll_setup(void)
 {
 	int i;
 
 	for_each_possible_cpu(i)
 		INIT_LIST_HEAD(&per_cpu(blk_cpu_iopoll, i));
 
-	open_softirq(BLOCK_IOPOLL_SOFTIRQ, blk_iopoll_softirq);
-	register_hotcpu_notifier(&blk_iopoll_cpu_notifier);
+	open_softirq(IRQ_POLL_SOFTIRQ, irq_poll_softirq);
+	register_hotcpu_notifier(&irq_poll_cpu_notifier);
 	return 0;
 }
-subsys_initcall(blk_iopoll_setup);
+subsys_initcall(irq_poll_setup);

tools/lib/traceevent/event-parse.c

@@ -3746,7 +3746,7 @@ static const struct flag flags[] = {
 	{ "NET_TX_SOFTIRQ", 2 },
 	{ "NET_RX_SOFTIRQ", 3 },
 	{ "BLOCK_SOFTIRQ", 4 },
-	{ "BLOCK_IOPOLL_SOFTIRQ", 5 },
+	{ "IRQ_POLL_SOFTIRQ", 5 },
 	{ "TASKLET_SOFTIRQ", 6 },
 	{ "SCHED_SOFTIRQ", 7 },
 	{ "HRTIMER_SOFTIRQ", 8 },

tools/perf/util/trace-event-parse.c

@@ -209,7 +209,7 @@ static const struct flag flags[] = {
 	{ "NET_TX_SOFTIRQ", 2 },
 	{ "NET_RX_SOFTIRQ", 3 },
 	{ "BLOCK_SOFTIRQ", 4 },
-	{ "BLOCK_IOPOLL_SOFTIRQ", 5 },
+	{ "IRQ_POLL_SOFTIRQ", 5 },
 	{ "TASKLET_SOFTIRQ", 6 },
 	{ "SCHED_SOFTIRQ", 7 },
 	{ "HRTIMER_SOFTIRQ", 8 },