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
  ...

9 files changed, 32 insertions, 77 deletions

diff --git a/drivers/infiniband/hw/qib/qib.h b/drivers/infiniband/hw/qib/qib.h
index 83d2349188db..432d6d0fd7f4 100644
--- a/drivers/infiniband/hw/qib/qib.h
+++ b/drivers/infiniband/hw/qib/qib.h
@@ -52,6 +52,7 @@
 #include <linux/kref.h>
 #include <linux/sched.h>
 #include <linux/kthread.h>
+#include <linux/xarray.h>
 #include <rdma/ib_hdrs.h>
 #include <rdma/rdma_vt.h>
 
@@ -1105,8 +1106,7 @@ struct qib_filedata {
 	int rec_cpu_num; /* for cpu affinity; -1 if none */
 };
 
-extern struct list_head qib_dev_list;
-extern spinlock_t qib_devs_lock;
+extern struct xarray qib_dev_table;
 extern struct qib_devdata *qib_lookup(int unit);
 extern u32 qib_cpulist_count;
 extern unsigned long *qib_cpulist;
diff --git a/drivers/infiniband/hw/qib/qib_common.h b/drivers/infiniband/hw/qib/qib_common.h
index a4a1f56ce824..f91f23e02283 100644
--- a/drivers/infiniband/hw/qib/qib_common.h
+++ b/drivers/infiniband/hw/qib/qib_common.h
@@ -57,7 +57,7 @@
  * QIB_VERBOSE_TRACING define as 1 if you want additional tracing in
  * fastpath code
  * QIB_TRACE_REGWRITES define as 1 if you want register writes to be
- * traced in faspath code
+ * traced in fastpath code
  * _QIB_TRACING define as 0 if you want to remove all tracing in a
  * compilation unit
  */
diff --git a/drivers/infiniband/hw/qib/qib_driver.c b/drivers/infiniband/hw/qib/qib_driver.c
index 3117cc5f2a9a..92eeea5679e2 100644
--- a/drivers/infiniband/hw/qib/qib_driver.c
+++ b/drivers/infiniband/hw/qib/qib_driver.c
@@ -49,8 +49,6 @@
  */
 const char ib_qib_version[] = QIB_DRIVER_VERSION "\n";
 
-DEFINE_SPINLOCK(qib_devs_lock);
-LIST_HEAD(qib_dev_list);
 DEFINE_MUTEX(qib_mutex);	/* general driver use */
 
 unsigned qib_ibmtu;
@@ -96,11 +94,11 @@ int qib_count_active_units(void)
 {
 	struct qib_devdata *dd;
 	struct qib_pportdata *ppd;
-	unsigned long flags;
+	unsigned long index, flags;
 	int pidx, nunits_active = 0;
 
-	spin_lock_irqsave(&qib_devs_lock, flags);
-	list_for_each_entry(dd, &qib_dev_list, list) {
+	xa_lock_irqsave(&qib_dev_table, flags);
+	xa_for_each(&qib_dev_table, index, dd) {
 		if (!(dd->flags & QIB_PRESENT) || !dd->kregbase)
 			continue;
 		for (pidx = 0; pidx < dd->num_pports; ++pidx) {
@@ -112,7 +110,7 @@ int qib_count_active_units(void)
 			}
 		}
 	}
-	spin_unlock_irqrestore(&qib_devs_lock, flags);
+	xa_unlock_irqrestore(&qib_dev_table, flags);
 	return nunits_active;
 }
 
@@ -125,13 +123,12 @@ int qib_count_units(int *npresentp, int *nupp)
 {
 	int nunits = 0, npresent = 0, nup = 0;
 	struct qib_devdata *dd;
-	unsigned long flags;
+	unsigned long index, flags;
 	int pidx;
 	struct qib_pportdata *ppd;
 
-	spin_lock_irqsave(&qib_devs_lock, flags);
-
-	list_for_each_entry(dd, &qib_dev_list, list) {
+	xa_lock_irqsave(&qib_dev_table, flags);
+	xa_for_each(&qib_dev_table, index, dd) {
 		nunits++;
 		if ((dd->flags & QIB_PRESENT) && dd->kregbase)
 			npresent++;
@@ -142,8 +139,7 @@ int qib_count_units(int *npresentp, int *nupp)
 				nup++;
 		}
 	}
-
-	spin_unlock_irqrestore(&qib_devs_lock, flags);
+	xa_unlock_irqrestore(&qib_dev_table, flags);
 
 	if (npresentp)
 		*npresentp = npresent;
diff --git a/drivers/infiniband/hw/qib/qib_fs.c b/drivers/infiniband/hw/qib/qib_fs.c
index 1d940a2885c9..ceb42d948412 100644
--- a/drivers/infiniband/hw/qib/qib_fs.c
+++ b/drivers/infiniband/hw/qib/qib_fs.c
@@ -508,8 +508,8 @@ bail:
  */
 static int qibfs_fill_super(struct super_block *sb, void *data, int silent)
 {
-	struct qib_devdata *dd, *tmp;
-	unsigned long flags;
+	struct qib_devdata *dd;
+	unsigned long index;
 	int ret;
 
 	static const struct tree_descr files[] = {
@@ -524,18 +524,12 @@ static int qibfs_fill_super(struct super_block *sb, void *data, int silent)
 		goto bail;
 	}
 
-	spin_lock_irqsave(&qib_devs_lock, flags);
-
-	list_for_each_entry_safe(dd, tmp, &qib_dev_list, list) {
-		spin_unlock_irqrestore(&qib_devs_lock, flags);
+	xa_for_each(&qib_dev_table, index, dd) {
 		ret = add_cntr_files(sb, dd);
 		if (ret)
 			goto bail;
-		spin_lock_irqsave(&qib_devs_lock, flags);
 	}
 
-	spin_unlock_irqrestore(&qib_devs_lock, flags);
-
 bail:
 	return ret;
 }
diff --git a/drivers/infiniband/hw/qib/qib_iba7322.c b/drivers/infiniband/hw/qib/qib_iba7322.c
index ac6a84f11ad0..dd4843379f51 100644
--- a/drivers/infiniband/hw/qib/qib_iba7322.c
+++ b/drivers/infiniband/hw/qib/qib_iba7322.c
@@ -6137,7 +6137,7 @@ static void set_no_qsfp_atten(struct qib_devdata *dd, int change)
 static int setup_txselect(const char *str, const struct kernel_param *kp)
 {
 	struct qib_devdata *dd;
-	unsigned long val;
+	unsigned long index, val;
 	char *n;
 
 	if (strlen(str) >= ARRAY_SIZE(txselect_list)) {
@@ -6153,7 +6153,7 @@ static int setup_txselect(const char *str, const struct kernel_param *kp)
 	}
 	strncpy(txselect_list, str, ARRAY_SIZE(txselect_list) - 1);
 
-	list_for_each_entry(dd, &qib_dev_list, list)
+	xa_for_each(&qib_dev_table, index, dd)
 		if (dd->deviceid == PCI_DEVICE_ID_QLOGIC_IB_7322)
 			set_no_qsfp_atten(dd, 1);
 	return 0;
diff --git a/drivers/infiniband/hw/qib/qib_init.c b/drivers/infiniband/hw/qib/qib_init.c
index 9fd69903ca57..d4fd8a6cff7b 100644
--- a/drivers/infiniband/hw/qib/qib_init.c
+++ b/drivers/infiniband/hw/qib/qib_init.c
@@ -36,7 +36,6 @@
 #include <linux/netdevice.h>
 #include <linux/vmalloc.h>
 #include <linux/delay.h>
-#include <linux/idr.h>
 #include <linux/module.h>
 #include <linux/printk.h>
 #ifdef CONFIG_INFINIBAND_QIB_DCA
@@ -95,7 +94,7 @@ MODULE_PARM_DESC(cc_table_size, "Congestion control table entries 0 (CCA disable
 
 static void verify_interrupt(struct timer_list *);
 
-static struct idr qib_unit_table;
+DEFINE_XARRAY_FLAGS(qib_dev_table, XA_FLAGS_ALLOC | XA_FLAGS_LOCK_IRQ);
 u32 qib_cpulist_count;
 unsigned long *qib_cpulist;
 
@@ -785,21 +784,9 @@ void __attribute__((weak)) qib_disable_wc(struct qib_devdata *dd)
 {
 }
 
-static inline struct qib_devdata *__qib_lookup(int unit)
-{
-	return idr_find(&qib_unit_table, unit);
-}
-
 struct qib_devdata *qib_lookup(int unit)
 {
-	struct qib_devdata *dd;
-	unsigned long flags;
-
-	spin_lock_irqsave(&qib_devs_lock, flags);
-	dd = __qib_lookup(unit);
-	spin_unlock_irqrestore(&qib_devs_lock, flags);
-
-	return dd;
+	return xa_load(&qib_dev_table, unit);
 }
 
 /*
@@ -1046,10 +1033,9 @@ void qib_free_devdata(struct qib_devdata *dd)
 {
 	unsigned long flags;
 
-	spin_lock_irqsave(&qib_devs_lock, flags);
-	idr_remove(&qib_unit_table, dd->unit);
-	list_del(&dd->list);
-	spin_unlock_irqrestore(&qib_devs_lock, flags);
+	xa_lock_irqsave(&qib_dev_table, flags);
+	__xa_erase(&qib_dev_table, dd->unit);
+	xa_unlock_irqrestore(&qib_dev_table, flags);
 
 #ifdef CONFIG_DEBUG_FS
 	qib_dbg_ibdev_exit(&dd->verbs_dev);
@@ -1070,15 +1056,15 @@ u64 qib_int_counter(struct qib_devdata *dd)
 
 u64 qib_sps_ints(void)
 {
-	unsigned long flags;
+	unsigned long index, flags;
 	struct qib_devdata *dd;
 	u64 sps_ints = 0;
 
-	spin_lock_irqsave(&qib_devs_lock, flags);
-	list_for_each_entry(dd, &qib_dev_list, list) {
+	xa_lock_irqsave(&qib_dev_table, flags);
+	xa_for_each(&qib_dev_table, index, dd) {
 		sps_ints += qib_int_counter(dd);
 	}
-	spin_unlock_irqrestore(&qib_devs_lock, flags);
+	xa_unlock_irqrestore(&qib_dev_table, flags);
 	return sps_ints;
 }
 
@@ -1087,12 +1073,9 @@ u64 qib_sps_ints(void)
  * allocator, because the verbs cleanup process both does cleanup and
  * free of the data structure.
  * "extra" is for chip-specific data.
- *
- * Use the idr mechanism to get a unit number for this unit.
  */
 struct qib_devdata *qib_alloc_devdata(struct pci_dev *pdev, size_t extra)
 {
-	unsigned long flags;
 	struct qib_devdata *dd;
 	int ret, nports;
 
@@ -1103,20 +1086,8 @@ struct qib_devdata *qib_alloc_devdata(struct pci_dev *pdev, size_t extra)
 	if (!dd)
 		return ERR_PTR(-ENOMEM);
 
-	INIT_LIST_HEAD(&dd->list);
-
-	idr_preload(GFP_KERNEL);
-	spin_lock_irqsave(&qib_devs_lock, flags);
-
-	ret = idr_alloc(&qib_unit_table, dd, 0, 0, GFP_NOWAIT);
-	if (ret >= 0) {
-		dd->unit = ret;
-		list_add(&dd->list, &qib_dev_list);
-	}
-
-	spin_unlock_irqrestore(&qib_devs_lock, flags);
-	idr_preload_end();
-
+	ret = xa_alloc_irq(&qib_dev_table, &dd->unit, dd, xa_limit_32b,
+			GFP_KERNEL);
 	if (ret < 0) {
 		qib_early_err(&pdev->dev,
 			      "Could not allocate unit ID: error %d\n", -ret);
@@ -1255,8 +1226,6 @@ static int __init qib_ib_init(void)
 	 * These must be called before the driver is registered with
 	 * the PCI subsystem.
 	 */
-	idr_init(&qib_unit_table);
-
 #ifdef CONFIG_INFINIBAND_QIB_DCA
 	dca_register_notify(&dca_notifier);
 #endif
@@ -1281,7 +1250,6 @@ bail_dev:
 #ifdef CONFIG_DEBUG_FS
 	qib_dbg_exit();
 #endif
-	idr_destroy(&qib_unit_table);
 	qib_dev_cleanup();
 bail:
 	return ret;
@@ -1313,7 +1281,7 @@ static void __exit qib_ib_cleanup(void)
 	qib_cpulist_count = 0;
 	kfree(qib_cpulist);
 
-	idr_destroy(&qib_unit_table);
+	WARN_ON(!xa_empty(&qib_dev_table));
 	qib_dev_cleanup();
 }
 
diff --git a/drivers/infiniband/hw/qib/qib_rc.c b/drivers/infiniband/hw/qib/qib_rc.c
index 50dd9811b088..2ac4c67f5ba1 100644
--- a/drivers/infiniband/hw/qib/qib_rc.c
+++ b/drivers/infiniband/hw/qib/qib_rc.c
@@ -933,7 +933,7 @@ void qib_rc_send_complete(struct rvt_qp *qp, struct ib_header *hdr)
 		qp->s_last = s_last;
 		/* see post_send() */
 		barrier();
-		rvt_put_swqe(wqe);
+		rvt_put_qp_swqe(qp, wqe);
 		rvt_qp_swqe_complete(qp,
 				     wqe,
 				     ib_qib_wc_opcode[wqe->wr.opcode],
@@ -975,7 +975,7 @@ static struct rvt_swqe *do_rc_completion(struct rvt_qp *qp,
 	    qib_cmp24(qp->s_sending_psn, qp->s_sending_hpsn) > 0) {
 		u32 s_last;
 
-		rvt_put_swqe(wqe);
+		rvt_put_qp_swqe(qp, wqe);
 		s_last = qp->s_last;
 		if (++s_last >= qp->s_size)
 			s_last = 0;
diff --git a/drivers/infiniband/hw/qib/qib_user_sdma.c b/drivers/infiniband/hw/qib/qib_user_sdma.c
index 31c523b2a9f5..ef19d39a44b1 100644
--- a/drivers/infiniband/hw/qib/qib_user_sdma.c
+++ b/drivers/infiniband/hw/qib/qib_user_sdma.c
@@ -225,8 +225,6 @@ qib_user_sdma_queue_create(struct device *dev, int unit, int ctxt, int sctxt)
 	if (sdma_rb_node) {
 		sdma_rb_node->refcount++;
 	} else {
-		int ret;
-
 		sdma_rb_node = kmalloc(sizeof(
 			struct qib_user_sdma_rb_node), GFP_KERNEL);
 		if (!sdma_rb_node)
@@ -235,8 +233,7 @@ qib_user_sdma_queue_create(struct device *dev, int unit, int ctxt, int sctxt)
 		sdma_rb_node->refcount = 1;
 		sdma_rb_node->pid = current->pid;
 
-		ret = qib_user_sdma_rb_insert(&qib_user_sdma_rb_root,
-					sdma_rb_node);
+		qib_user_sdma_rb_insert(&qib_user_sdma_rb_root, sdma_rb_node);
 	}
 	pq->sdma_rb_node = sdma_rb_node;
 
diff --git a/drivers/infiniband/hw/qib/qib_verbs.h b/drivers/infiniband/hw/qib/qib_verbs.h
index a4426c24b0d1..17bdf8acee2f 100644
--- a/drivers/infiniband/hw/qib/qib_verbs.h
+++ b/drivers/infiniband/hw/qib/qib_verbs.h
@@ -46,7 +46,7 @@
 #include <rdma/ib_pack.h>
 #include <rdma/ib_user_verbs.h>
 #include <rdma/ib_hdrs.h>
-#include <rdma/rdma_vt.h>
+#include <rdma/rdmavt_qp.h>
 #include <rdma/rdmavt_cq.h>
 
 struct qib_ctxtdata;