Merge tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/gi…

…t/rdma/rdma

Pull rdma updates from Jason Gunthorpe:
 "This has been a slightly more active cycle than normal with ongoing
  core changes and quite a lot of collected driver updates.

   - Various driver fixes for bnxt_re, cxgb4, hns, mlx5, pvrdma, rxe

   - A new data transfer mode for HFI1 giving higher performance

   - Significant functional and bug fix update to the mlx5
     On-Demand-Paging MR feature

   - A chip hang reset recovery system for hns

   - Change mm->pinned_vm to an atomic64

   - Update bnxt_re to support a new 57500 chip

   - A sane netlink 'rdma link add' method for creating rxe devices and
     fixing the various unregistration race conditions in rxe's
     unregister flow

   - Allow lookup up objects by an ID over netlink

   - Various reworking of the core to driver interface:
       - drivers should not assume umem SGLs are in PAGE_SIZE chunks
       - ucontext is accessed via udata not other means
       - start to make the core code responsible for object memory
         allocation
       - drivers should convert struct device to struct ib_device via a
         helper
       - drivers have more tools to avoid use after unregister problems"

* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma: (280 commits)
  net/mlx5: ODP support for XRC transport is not enabled by default in FW
  IB/hfi1: Close race condition on user context disable and close
  RDMA/umem: Revert broken 'off by one' fix
  RDMA/umem: minor bug fix in error handling path
  RDMA/hns: Use GFP_ATOMIC in hns_roce_v2_modify_qp
  cxgb4: kfree mhp after the debug print
  IB/rdmavt: Fix concurrency panics in QP post_send and modify to error
  IB/rdmavt: Fix loopback send with invalidate ordering
  IB/iser: Fix dma_nents type definition
  IB/mlx5: Set correct write permissions for implicit ODP MR
  bnxt_re: Clean cq for kernel consumers only
  RDMA/uverbs: Don't do double free of allocated PD
  RDMA: Handle ucontext allocations by IB/core
  RDMA/core: Fix a WARN() message
  bnxt_re: fix the regression due to changes in alloc_pbl
  IB/mlx4: Increase the timeout for CM cache
  IB/core: Abort page fault handler silently during owning process exit
  IB/mlx5: Validate correct PD before prefetch MR
  IB/mlx5: Protect against prefetch of invalid MR
  RDMA/uverbs: Store PR pointer before it is overwritten
  ...

.clang-format

@@ -240,6 +240,7 @@ ForEachMacros:
   - 'for_each_set_bit'
   - 'for_each_set_bit_from'
   - 'for_each_sg'
+  - 'for_each_sg_dma_page'
   - 'for_each_sg_page'
   - 'for_each_sibling_event'
   - '__for_each_thread'
@@ -360,6 +361,7 @@ ForEachMacros:
   - 'radix_tree_for_each_slot'
   - 'radix_tree_for_each_tagged'
   - 'rbtree_postorder_for_each_entry_safe'
+  - 'rdma_for_each_port'
   - 'resource_list_for_each_entry'
   - 'resource_list_for_each_entry_safe'
   - 'rhl_for_each_entry_rcu'

Documentation/infiniband/user_verbs.txt

@@ -46,11 +46,11 @@ Memory pinning
   I/O targets be kept resident at the same physical address.  The
   ib_uverbs module manages pinning and unpinning memory regions via
   get_user_pages() and put_page() calls.  It also accounts for the
-  amount of memory pinned in the process's locked_vm, and checks that
+  amount of memory pinned in the process's pinned_vm, and checks that
   unprivileged processes do not exceed their RLIMIT_MEMLOCK limit.
 
   Pages that are pinned multiple times are counted each time they are
-  pinned, so the value of locked_vm may be an overestimate of the
+  pinned, so the value of pinned_vm may be an overestimate of the
   number of pages pinned by a process.
 
 /dev files

drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c

@@ -311,7 +311,13 @@ static dma_addr_t __vmw_piter_dma_addr(struct vmw_piter *viter)
 
 static dma_addr_t __vmw_piter_sg_addr(struct vmw_piter *viter)
 {
-	return sg_page_iter_dma_address(&viter->iter);
+	/*
+	 * FIXME: This driver wrongly mixes DMA and CPU SG list iteration and
+	 * needs revision. See
+	 * https://lore.kernel.org/lkml/[email protected]/
+	 */
+	return sg_page_iter_dma_address(
+		container_of(&viter->iter, struct sg_dma_page_iter, base));
 }
 
 

drivers/infiniband/Kconfig

@@ -89,6 +89,7 @@ config INFINIBAND_ADDR_TRANS_CONFIGFS
 	  This allows the user to config the default GID type that the CM
 	  uses for each device, when initiaing new connections.
 
+if INFINIBAND_USER_ACCESS || !INFINIBAND_USER_ACCESS
 source "drivers/infiniband/hw/mthca/Kconfig"
 source "drivers/infiniband/hw/qib/Kconfig"
 source "drivers/infiniband/hw/cxgb3/Kconfig"
@@ -101,6 +102,12 @@ source "drivers/infiniband/hw/ocrdma/Kconfig"
 source "drivers/infiniband/hw/vmw_pvrdma/Kconfig"
 source "drivers/infiniband/hw/usnic/Kconfig"
 source "drivers/infiniband/hw/hns/Kconfig"
+source "drivers/infiniband/hw/bnxt_re/Kconfig"
+source "drivers/infiniband/hw/hfi1/Kconfig"
+source "drivers/infiniband/hw/qedr/Kconfig"
+source "drivers/infiniband/sw/rdmavt/Kconfig"
+source "drivers/infiniband/sw/rxe/Kconfig"
+endif
 
 source "drivers/infiniband/ulp/ipoib/Kconfig"
 
@@ -111,13 +118,5 @@ source "drivers/infiniband/ulp/iser/Kconfig"
 source "drivers/infiniband/ulp/isert/Kconfig"
 
 source "drivers/infiniband/ulp/opa_vnic/Kconfig"
-source "drivers/infiniband/sw/rdmavt/Kconfig"
-source "drivers/infiniband/sw/rxe/Kconfig"
-
-source "drivers/infiniband/hw/hfi1/Kconfig"
-
-source "drivers/infiniband/hw/qedr/Kconfig"
-
-source "drivers/infiniband/hw/bnxt_re/Kconfig"
 
 endif # INFINIBAND

drivers/infiniband/core/Makefile

@@ -15,8 +15,6 @@ ib_core-y :=			packer.o ud_header.o verbs.o cq.o rw.o sysfs.o \
 				nldev.o restrack.o
 
 ib_core-$(CONFIG_SECURITY_INFINIBAND) += security.o
-ib_core-$(CONFIG_INFINIBAND_USER_MEM) += umem.o
-ib_core-$(CONFIG_INFINIBAND_ON_DEMAND_PAGING) += umem_odp.o
 ib_core-$(CONFIG_CGROUP_RDMA) += cgroup.o
 
 ib_cm-y :=			cm.o
@@ -39,3 +37,5 @@ ib_uverbs-y :=			uverbs_main.o uverbs_cmd.o uverbs_marshall.o \
 				uverbs_std_types_flow_action.o uverbs_std_types_dm.o \
 				uverbs_std_types_mr.o uverbs_std_types_counters.o \
 				uverbs_uapi.o uverbs_std_types_device.o
+ib_uverbs-$(CONFIG_INFINIBAND_USER_MEM) += umem.o
+ib_uverbs-$(CONFIG_INFINIBAND_ON_DEMAND_PAGING) += umem_odp.o

drivers/infiniband/core/cache.c

@@ -185,7 +185,7 @@ EXPORT_SYMBOL(ib_cache_gid_parse_type_str);
 
 static struct ib_gid_table *rdma_gid_table(struct ib_device *device, u8 port)
 {
-	return device->cache.ports[port - rdma_start_port(device)].gid;
+	return device->port_data[port].cache.gid;
 }
 
 static bool is_gid_entry_free(const struct ib_gid_table_entry *entry)
@@ -547,21 +547,19 @@ int ib_cache_gid_add(struct ib_device *ib_dev, u8 port,
 	unsigned long mask;
 	int ret;
 
-	if (ib_dev->ops.get_netdev) {
-		idev = ib_dev->ops.get_netdev(ib_dev, port);
-		if (idev && attr->ndev != idev) {
-			union ib_gid default_gid;
+	idev = ib_device_get_netdev(ib_dev, port);
+	if (idev && attr->ndev != idev) {
+		union ib_gid default_gid;
 
-			/* Adding default GIDs in not permitted */
-			make_default_gid(idev, &default_gid);
-			if (!memcmp(gid, &default_gid, sizeof(*gid))) {
-				dev_put(idev);
-				return -EPERM;
-			}
-		}
-		if (idev)
+		/* Adding default GIDs is not permitted */
+		make_default_gid(idev, &default_gid);
+		if (!memcmp(gid, &default_gid, sizeof(*gid))) {
 			dev_put(idev);
+			return -EPERM;
+		}
 	}
+	if (idev)
+		dev_put(idev);
 
 	mask = GID_ATTR_FIND_MASK_GID |
 	       GID_ATTR_FIND_MASK_GID_TYPE |
@@ -765,7 +763,7 @@ static struct ib_gid_table *alloc_gid_table(int sz)
 	return NULL;
 }
 
-static void release_gid_table(struct ib_device *device, u8 port,
+static void release_gid_table(struct ib_device *device,
 			      struct ib_gid_table *table)
 {
 	bool leak = false;
@@ -863,31 +861,27 @@ static void gid_table_reserve_default(struct ib_device *ib_dev, u8 port,
 
 static void gid_table_release_one(struct ib_device *ib_dev)
 {
-	struct ib_gid_table *table;
-	u8 port;
+	unsigned int p;
 
-	for (port = 0; port < ib_dev->phys_port_cnt; port++) {
-		table = ib_dev->cache.ports[port].gid;
-		release_gid_table(ib_dev, port, table);
-		ib_dev->cache.ports[port].gid = NULL;
+	rdma_for_each_port (ib_dev, p) {
+		release_gid_table(ib_dev, ib_dev->port_data[p].cache.gid);
+		ib_dev->port_data[p].cache.gid = NULL;
 	}
 }
 
 static int _gid_table_setup_one(struct ib_device *ib_dev)
 {
-	u8 port;
 	struct ib_gid_table *table;
+	unsigned int rdma_port;
 
-	for (port = 0; port < ib_dev->phys_port_cnt; port++) {
-		u8 rdma_port = port + rdma_start_port(ib_dev);
-
-		table =	alloc_gid_table(
-				ib_dev->port_immutable[rdma_port].gid_tbl_len);
+	rdma_for_each_port (ib_dev, rdma_port) {
+		table = alloc_gid_table(
+			ib_dev->port_data[rdma_port].immutable.gid_tbl_len);
 		if (!table)
 			goto rollback_table_setup;
 
 		gid_table_reserve_default(ib_dev, rdma_port, table);
-		ib_dev->cache.ports[port].gid = table;
+		ib_dev->port_data[rdma_port].cache.gid = table;
 	}
 	return 0;
 
@@ -898,14 +892,11 @@ static int _gid_table_setup_one(struct ib_device *ib_dev)
 
 static void gid_table_cleanup_one(struct ib_device *ib_dev)
 {
-	struct ib_gid_table *table;
-	u8 port;
+	unsigned int p;
 
-	for (port = 0; port < ib_dev->phys_port_cnt; port++) {
-		table = ib_dev->cache.ports[port].gid;
-		cleanup_gid_table_port(ib_dev, port + rdma_start_port(ib_dev),
-				       table);
-	}
+	rdma_for_each_port (ib_dev, p)
+		cleanup_gid_table_port(ib_dev, p,
+				       ib_dev->port_data[p].cache.gid);
 }
 
 static int gid_table_setup_one(struct ib_device *ib_dev)
@@ -983,17 +974,17 @@ const struct ib_gid_attr *rdma_find_gid(struct ib_device *device,
 	unsigned long mask = GID_ATTR_FIND_MASK_GID |
 			     GID_ATTR_FIND_MASK_GID_TYPE;
 	struct ib_gid_attr gid_attr_val = {.ndev = ndev, .gid_type = gid_type};
-	u8 p;
+	unsigned int p;
 
 	if (ndev)
 		mask |= GID_ATTR_FIND_MASK_NETDEV;
 
-	for (p = 0; p < device->phys_port_cnt; p++) {
+	rdma_for_each_port(device, p) {
 		struct ib_gid_table *table;
 		unsigned long flags;
 		int index;
 
-		table = device->cache.ports[p].gid;
+		table = device->port_data[p].cache.gid;
 		read_lock_irqsave(&table->rwlock, flags);
 		index = find_gid(table, gid, &gid_attr_val, false, mask, NULL);
 		if (index >= 0) {
@@ -1025,7 +1016,7 @@ int ib_get_cached_pkey(struct ib_device *device,
 
 	read_lock_irqsave(&device->cache.lock, flags);
 
-	cache = device->cache.ports[port_num - rdma_start_port(device)].pkey;
+	cache = device->port_data[port_num].cache.pkey;
 
 	if (index < 0 || index >= cache->table_len)
 		ret = -EINVAL;
@@ -1043,14 +1034,12 @@ int ib_get_cached_subnet_prefix(struct ib_device *device,
 				u64              *sn_pfx)
 {
 	unsigned long flags;
-	int p;
 
 	if (!rdma_is_port_valid(device, port_num))
 		return -EINVAL;
 
-	p = port_num - rdma_start_port(device);
 	read_lock_irqsave(&device->cache.lock, flags);
-	*sn_pfx = device->cache.ports[p].subnet_prefix;
+	*sn_pfx = device->port_data[port_num].cache.subnet_prefix;
 	read_unlock_irqrestore(&device->cache.lock, flags);
 
 	return 0;
@@ -1073,7 +1062,7 @@ int ib_find_cached_pkey(struct ib_device *device,
 
 	read_lock_irqsave(&device->cache.lock, flags);
 
-	cache = device->cache.ports[port_num - rdma_start_port(device)].pkey;
+	cache = device->port_data[port_num].cache.pkey;
 
 	*index = -1;
 
@@ -1113,7 +1102,7 @@ int ib_find_exact_cached_pkey(struct ib_device *device,
 
 	read_lock_irqsave(&device->cache.lock, flags);
 
-	cache = device->cache.ports[port_num - rdma_start_port(device)].pkey;
+	cache = device->port_data[port_num].cache.pkey;
 
 	*index = -1;
 
@@ -1141,7 +1130,7 @@ int ib_get_cached_lmc(struct ib_device *device,
 		return -EINVAL;
 
 	read_lock_irqsave(&device->cache.lock, flags);
-	*lmc = device->cache.ports[port_num - rdma_start_port(device)].lmc;
+	*lmc = device->port_data[port_num].cache.lmc;
 	read_unlock_irqrestore(&device->cache.lock, flags);
 
 	return ret;
@@ -1159,8 +1148,7 @@ int ib_get_cached_port_state(struct ib_device   *device,
 		return -EINVAL;
 
 	read_lock_irqsave(&device->cache.lock, flags);
-	*port_state = device->cache.ports[port_num
-		- rdma_start_port(device)].port_state;
+	*port_state = device->port_data[port_num].cache.port_state;
 	read_unlock_irqrestore(&device->cache.lock, flags);
 
 	return ret;
@@ -1361,16 +1349,13 @@ static void ib_cache_update(struct ib_device *device,
 
 	write_lock_irq(&device->cache.lock);
 
-	old_pkey_cache = device->cache.ports[port -
-		rdma_start_port(device)].pkey;
+	old_pkey_cache = device->port_data[port].cache.pkey;
 
-	device->cache.ports[port - rdma_start_port(device)].pkey = pkey_cache;
-	device->cache.ports[port - rdma_start_port(device)].lmc = tprops->lmc;
-	device->cache.ports[port - rdma_start_port(device)].port_state =
-		tprops->state;
+	device->port_data[port].cache.pkey = pkey_cache;
+	device->port_data[port].cache.lmc = tprops->lmc;
+	device->port_data[port].cache.port_state = tprops->state;
 
-	device->cache.ports[port - rdma_start_port(device)].subnet_prefix =
-							tprops->subnet_prefix;
+	device->port_data[port].cache.subnet_prefix = tprops->subnet_prefix;
 	write_unlock_irq(&device->cache.lock);
 
 	if (enforce_security)
@@ -1428,27 +1413,17 @@ static void ib_cache_event(struct ib_event_handler *handler,
 
 int ib_cache_setup_one(struct ib_device *device)
 {
-	int p;
+	unsigned int p;
 	int err;
 
 	rwlock_init(&device->cache.lock);
 
-	device->cache.ports =
-		kcalloc(rdma_end_port(device) - rdma_start_port(device) + 1,
-			sizeof(*device->cache.ports),
-			GFP_KERNEL);
-	if (!device->cache.ports)
-		return -ENOMEM;
-
 	err = gid_table_setup_one(device);
-	if (err) {
-		kfree(device->cache.ports);
-		device->cache.ports = NULL;
+	if (err)
 		return err;
-	}
 
-	for (p = 0; p <= rdma_end_port(device) - rdma_start_port(device); ++p)
-		ib_cache_update(device, p + rdma_start_port(device), true);
+	rdma_for_each_port (device, p)
+		ib_cache_update(device, p, true);
 
 	INIT_IB_EVENT_HANDLER(&device->cache.event_handler,
 			      device, ib_cache_event);
@@ -1458,19 +1433,18 @@ int ib_cache_setup_one(struct ib_device *device)
 
 void ib_cache_release_one(struct ib_device *device)
 {
-	int p;
+	unsigned int p;
 
 	/*
 	 * The release function frees all the cache elements.
 	 * This function should be called as part of freeing
 	 * all the device's resources when the cache could no
 	 * longer be accessed.
 	 */
-	for (p = 0; p <= rdma_end_port(device) - rdma_start_port(device); ++p)
-		kfree(device->cache.ports[p].pkey);
+	rdma_for_each_port (device, p)
+		kfree(device->port_data[p].cache.pkey);
 
 	gid_table_release_one(device);
-	kfree(device->cache.ports);
 }
 
 void ib_cache_cleanup_one(struct ib_device *device)

drivers/infiniband/core/cgroup.c

@@ -21,12 +21,11 @@
  * Register with the rdma cgroup. Should be called before
  * exposing rdma device to user space applications to avoid
  * resource accounting leak.
- * Returns 0 on success or otherwise failure code.
  */
-int ib_device_register_rdmacg(struct ib_device *device)
+void ib_device_register_rdmacg(struct ib_device *device)
 {
 	device->cg_device.name = device->name;
-	return rdmacg_register_device(&device->cg_device);
+	rdmacg_register_device(&device->cg_device);
 }
 
 /**

drivers/infiniband/core/cm.c

@@ -4052,8 +4052,7 @@ static void cm_recv_handler(struct ib_mad_agent *mad_agent,
 	atomic_long_inc(&port->counter_group[CM_RECV].
 			counter[attr_id - CM_ATTR_ID_OFFSET]);
 
-	work = kmalloc(sizeof(*work) + sizeof(struct sa_path_rec) * paths,
-		       GFP_KERNEL);
+	work = kmalloc(struct_size(work, path, paths), GFP_KERNEL);
 	if (!work) {
 		ib_free_recv_mad(mad_recv_wc);
 		return;