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 large cycle for RDMA, with several major patch series
  reworking parts of the core code.

   - Rework the so-called 'gid cache' and internal APIs to use a kref'd
     pointer to a struct instead of copying, push this upwards into the
     callers and add more stuff to the struct. The new design avoids
     some ugly races the old one suffered with. This is part of the
     namespace enablement work as the new struct is learning to be
     namespace aware.

   - Various uapi cleanups, moving more stuff to include/uapi and fixing
     some long standing bugs that have recently been discovered.

   - Driver updates for mlx5, mlx4 i40iw, rxe, cxgb4, hfi1, usnic,
     pvrdma, and hns

   - Provide max_send_sge and max_recv_sge attributes to better support
     HW where these values are asymmetric.

   - mlx5 user API 'devx' allows sending commands directly to the device
     FW, instead of trying to cram every wild and niche feature into the
     common API. Sort of like what GPU does.

   - Major write() and ioctl() API rework to cleanly support PCI device
     hot unplug and advance the ioctl conversion work

   - Sparse and compile warning cleanups

   - Add 'const' to the ib_poll_cq() signature, and permit a NULL
     'bad_wr', which is the common use case

   - Various patches to avoid high order allocations across the stack

   - SRQ support for cxgb4, hns and qedr

   - Changes to IPoIB to better follow the netdev model for working with
     struct net_device liftime"

* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma: (312 commits)
  Revert "net/smc: Replace ib_query_gid with rdma_get_gid_attr"
  RDMA/hns: Fix usage of bitmap allocation functions return values
  IB/core: Change filter function return type from int to bool
  IB/core: Update GID entries for netdevice whose mac address changes
  IB/core: Add default GIDs of the bond master netdev
  IB/core: Consider adding default GIDs of bond device
  IB/core: Delete lower netdevice default GID entries in bonding scenario
  IB/core: Avoid confusing del_netdev_default_ips
  IB/core: Add comment for change upper netevent handling
  qedr: Add user space support for SRQ
  qedr: Add support for kernel mode SRQ's
  qedr: Add wrapping generic structure for qpidr and adjust idr routines.
  IB/mlx5: Fix leaking stack memory to userspace
  Update the e-mail address of Bart Van Assche
  IB/ucm: Fix compiling ucm.c
  IB/uverbs: Do not check for device disassociation during ioctl
  IB/uverbs: Remove struct uverbs_root_spec and all supporting code
  IB/uverbs: Use uverbs_api to unmarshal ioctl commands
  IB/uverbs: Use uverbs_alloc for allocations
  IB/uverbs: Add a simple allocator to uverbs_attr_bundle
  ...

1 files changed, 84 insertions, 148 deletions

diff --git a/drivers/infiniband/core/uverbs_main.c b/drivers/infiniband/core/uverbs_main.c
index 2094d136513d..823beca448e1 100644
--- a/drivers/infiniband/core/uverbs_main.c
+++ b/drivers/infiniband/core/uverbs_main.c
@@ -41,8 +41,6 @@
 #include <linux/fs.h>
 #include <linux/poll.h>
 #include <linux/sched.h>
-#include <linux/sched/mm.h>
-#include <linux/sched/task.h>
 #include <linux/file.h>
 #include <linux/cdev.h>
 #include <linux/anon_inodes.h>
@@ -77,7 +75,6 @@ static struct class *uverbs_class;
 static DECLARE_BITMAP(dev_map, IB_UVERBS_MAX_DEVICES);
 
 static ssize_t (*uverbs_cmd_table[])(struct ib_uverbs_file *file,
-				     struct ib_device *ib_dev,
 				     const char __user *buf, int in_len,
 				     int out_len) = {
 	[IB_USER_VERBS_CMD_GET_CONTEXT]		= ib_uverbs_get_context,
@@ -118,7 +115,6 @@ static ssize_t (*uverbs_cmd_table[])(struct ib_uverbs_file *file,
 };
 
 static int (*uverbs_ex_cmd_table[])(struct ib_uverbs_file *file,
-				    struct ib_device *ib_dev,
 				    struct ib_udata *ucore,
 				    struct ib_udata *uhw) = {
 	[IB_USER_VERBS_EX_CMD_CREATE_FLOW]	= ib_uverbs_ex_create_flow,
@@ -138,6 +134,30 @@ static int (*uverbs_ex_cmd_table[])(struct ib_uverbs_file *file,
 static void ib_uverbs_add_one(struct ib_device *device);
 static void ib_uverbs_remove_one(struct ib_device *device, void *client_data);
 
+/*
+ * Must be called with the ufile->device->disassociate_srcu held, and the lock
+ * must be held until use of the ucontext is finished.
+ */
+struct ib_ucontext *ib_uverbs_get_ucontext(struct ib_uverbs_file *ufile)
+{
+	/*
+	 * We do not hold the hw_destroy_rwsem lock for this flow, instead
+	 * srcu is used. It does not matter if someone races this with
+	 * get_context, we get NULL or valid ucontext.
+	 */
+	struct ib_ucontext *ucontext = smp_load_acquire(&ufile->ucontext);
+
+	if (!srcu_dereference(ufile->device->ib_dev,
+			      &ufile->device->disassociate_srcu))
+		return ERR_PTR(-EIO);
+
+	if (!ucontext)
+		return ERR_PTR(-EINVAL);
+
+	return ucontext;
+}
+EXPORT_SYMBOL(ib_uverbs_get_ucontext);
+
 int uverbs_dealloc_mw(struct ib_mw *mw)
 {
 	struct ib_pd *pd = mw->pd;
@@ -154,6 +174,7 @@ static void ib_uverbs_release_dev(struct kobject *kobj)
 	struct ib_uverbs_device *dev =
 		container_of(kobj, struct ib_uverbs_device, kobj);
 
+	uverbs_destroy_api(dev->uapi);
 	cleanup_srcu_struct(&dev->disassociate_srcu);
 	kfree(dev);
 }
@@ -184,7 +205,7 @@ void ib_uverbs_release_ucq(struct ib_uverbs_file *file,
 		}
 		spin_unlock_irq(&ev_file->ev_queue.lock);
 
-		uverbs_uobject_put(&ev_file->uobj_file.uobj);
+		uverbs_uobject_put(&ev_file->uobj);
 	}
 
 	spin_lock_irq(&file->async_file->ev_queue.lock);
@@ -220,20 +241,6 @@ void ib_uverbs_detach_umcast(struct ib_qp *qp,
 	}
 }
 
-static int ib_uverbs_cleanup_ucontext(struct ib_uverbs_file *file,
-				      struct ib_ucontext *context,
-				      bool device_removed)
-{
-	context->closing = 1;
-	uverbs_cleanup_ucontext(context, device_removed);
-	put_pid(context->tgid);
-
-	ib_rdmacg_uncharge(&context->cg_obj, context->device,
-			   RDMACG_RESOURCE_HCA_HANDLE);
-
-	return context->device->dealloc_ucontext(context);
-}
-
 static void ib_uverbs_comp_dev(struct ib_uverbs_device *dev)
 {
 	complete(&dev->comp);
@@ -246,6 +253,8 @@ void ib_uverbs_release_file(struct kref *ref)
 	struct ib_device *ib_dev;
 	int srcu_key;
 
+	release_ufile_idr_uobject(file);
+
 	srcu_key = srcu_read_lock(&file->device->disassociate_srcu);
 	ib_dev = srcu_dereference(file->device->ib_dev,
 				  &file->device->disassociate_srcu);
@@ -338,7 +347,7 @@ static ssize_t ib_uverbs_comp_event_read(struct file *filp, char __user *buf,
 		filp->private_data;
 
 	return ib_uverbs_event_read(&comp_ev_file->ev_queue,
-				    comp_ev_file->uobj_file.ufile, filp,
+				    comp_ev_file->uobj.ufile, filp,
 				    buf, count, pos,
 				    sizeof(struct ib_uverbs_comp_event_desc));
 }
@@ -420,7 +429,9 @@ static int ib_uverbs_async_event_close(struct inode *inode, struct file *filp)
 
 static int ib_uverbs_comp_event_close(struct inode *inode, struct file *filp)
 {
-	struct ib_uverbs_completion_event_file *file = filp->private_data;
+	struct ib_uobject *uobj = filp->private_data;
+	struct ib_uverbs_completion_event_file *file = container_of(
+		uobj, struct ib_uverbs_completion_event_file, uobj);
 	struct ib_uverbs_event *entry, *tmp;
 
 	spin_lock_irq(&file->ev_queue.lock);
@@ -528,7 +539,7 @@ void ib_uverbs_cq_event_handler(struct ib_event *event, void *context_ptr)
 	struct ib_ucq_object *uobj = container_of(event->element.cq->uobject,
 						  struct ib_ucq_object, uobject);
 
-	ib_uverbs_async_handler(uobj->uverbs_file, uobj->uobject.user_handle,
+	ib_uverbs_async_handler(uobj->uobject.ufile, uobj->uobject.user_handle,
 				event->event, &uobj->async_list,
 				&uobj->async_events_reported);
 }
@@ -637,13 +648,13 @@ err_put_refs:
 	return filp;
 }
 
-static bool verify_command_mask(struct ib_device *ib_dev,
-				u32 command, bool extended)
+static bool verify_command_mask(struct ib_uverbs_file *ufile, u32 command,
+				bool extended)
 {
 	if (!extended)
-		return ib_dev->uverbs_cmd_mask & BIT_ULL(command);
+		return ufile->uverbs_cmd_mask & BIT_ULL(command);
 
-	return ib_dev->uverbs_ex_cmd_mask & BIT_ULL(command);
+	return ufile->uverbs_ex_cmd_mask & BIT_ULL(command);
 }
 
 static bool verify_command_idx(u32 command, bool extended)
@@ -713,7 +724,6 @@ static ssize_t ib_uverbs_write(struct file *filp, const char __user *buf,
 {
 	struct ib_uverbs_file *file = filp->private_data;
 	struct ib_uverbs_ex_cmd_hdr ex_hdr;
-	struct ib_device *ib_dev;
 	struct ib_uverbs_cmd_hdr hdr;
 	bool extended;
 	int srcu_key;
@@ -748,24 +758,8 @@ static ssize_t ib_uverbs_write(struct file *filp, const char __user *buf,
 		return ret;
 
 	srcu_key = srcu_read_lock(&file->device->disassociate_srcu);
-	ib_dev = srcu_dereference(file->device->ib_dev,
-				  &file->device->disassociate_srcu);
-	if (!ib_dev) {
-		ret = -EIO;
-		goto out;
-	}
-
-	/*
-	 * Must be after the ib_dev check, as once the RCU clears ib_dev ==
-	 * NULL means ucontext == NULL
-	 */
-	if (!file->ucontext &&
-	    (command != IB_USER_VERBS_CMD_GET_CONTEXT || extended)) {
-		ret = -EINVAL;
-		goto out;
-	}
 
-	if (!verify_command_mask(ib_dev, command, extended)) {
+	if (!verify_command_mask(file, command, extended)) {
 		ret = -EOPNOTSUPP;
 		goto out;
 	}
@@ -773,7 +767,7 @@ static ssize_t ib_uverbs_write(struct file *filp, const char __user *buf,
 	buf += sizeof(hdr);
 
 	if (!extended) {
-		ret = uverbs_cmd_table[command](file, ib_dev, buf,
+		ret = uverbs_cmd_table[command](file, buf,
 						hdr.in_words * 4,
 						hdr.out_words * 4);
 	} else {
@@ -792,7 +786,7 @@ static ssize_t ib_uverbs_write(struct file *filp, const char __user *buf,
 					ex_hdr.provider_in_words * 8,
 					ex_hdr.provider_out_words * 8);
 
-		ret = uverbs_ex_cmd_table[command](file, ib_dev, &ucore, &uhw);
+		ret = uverbs_ex_cmd_table[command](file, &ucore, &uhw);
 		ret = (ret) ? : count;
 	}
 
@@ -804,22 +798,18 @@ out:
 static int ib_uverbs_mmap(struct file *filp, struct vm_area_struct *vma)
 {
 	struct ib_uverbs_file *file = filp->private_data;
-	struct ib_device *ib_dev;
+	struct ib_ucontext *ucontext;
 	int ret = 0;
 	int srcu_key;
 
 	srcu_key = srcu_read_lock(&file->device->disassociate_srcu);
-	ib_dev = srcu_dereference(file->device->ib_dev,
-				  &file->device->disassociate_srcu);
-	if (!ib_dev) {
-		ret = -EIO;
+	ucontext = ib_uverbs_get_ucontext(file);
+	if (IS_ERR(ucontext)) {
+		ret = PTR_ERR(ucontext);
 		goto out;
 	}
 
-	if (!file->ucontext)
-		ret = -ENODEV;
-	else
-		ret = ib_dev->mmap(file->ucontext, vma);
+	ret = ucontext->device->mmap(ucontext, vma);
 out:
 	srcu_read_unlock(&file->device->disassociate_srcu, srcu_key);
 	return ret;
@@ -879,13 +869,12 @@ static int ib_uverbs_open(struct inode *inode, struct file *filp)
 	}
 
 	file->device	 = dev;
-	spin_lock_init(&file->idr_lock);
-	idr_init(&file->idr);
-	file->ucontext	 = NULL;
-	file->async_file = NULL;
 	kref_init(&file->ref);
-	mutex_init(&file->mutex);
-	mutex_init(&file->cleanup_mutex);
+	mutex_init(&file->ucontext_lock);
+
+	spin_lock_init(&file->uobjects_lock);
+	INIT_LIST_HEAD(&file->uobjects);
+	init_rwsem(&file->hw_destroy_rwsem);
 
 	filp->private_data = file;
 	kobject_get(&dev->kobj);
@@ -893,6 +882,11 @@ static int ib_uverbs_open(struct inode *inode, struct file *filp)
 	mutex_unlock(&dev->lists_mutex);
 	srcu_read_unlock(&dev->disassociate_srcu, srcu_key);
 
+	file->uverbs_cmd_mask = ib_dev->uverbs_cmd_mask;
+	file->uverbs_ex_cmd_mask = ib_dev->uverbs_ex_cmd_mask;
+
+	setup_ufile_idr_uobject(file);
+
 	return nonseekable_open(inode, filp);
 
 err_module:
@@ -911,13 +905,7 @@ static int ib_uverbs_close(struct inode *inode, struct file *filp)
 {
 	struct ib_uverbs_file *file = filp->private_data;
 
-	mutex_lock(&file->cleanup_mutex);
-	if (file->ucontext) {
-		ib_uverbs_cleanup_ucontext(file, file->ucontext, false);
-		file->ucontext = NULL;
-	}
-	mutex_unlock(&file->cleanup_mutex);
-	idr_destroy(&file->idr);
+	uverbs_destroy_ufile_hw(file, RDMA_REMOVE_CLOSE);
 
 	mutex_lock(&file->device->lists_mutex);
 	if (!file->is_closed) {
@@ -1006,6 +994,19 @@ static DEVICE_ATTR(abi_version, S_IRUGO, show_dev_abi_version, NULL);
 static CLASS_ATTR_STRING(abi_version, S_IRUGO,
 			 __stringify(IB_USER_VERBS_ABI_VERSION));
 
+static int ib_uverbs_create_uapi(struct ib_device *device,
+				 struct ib_uverbs_device *uverbs_dev)
+{
+	struct uverbs_api *uapi;
+
+	uapi = uverbs_alloc_api(device->driver_specs, device->driver_id);
+	if (IS_ERR(uapi))
+		return PTR_ERR(uapi);
+
+	uverbs_dev->uapi = uapi;
+	return 0;
+}
+
 static void ib_uverbs_add_one(struct ib_device *device)
 {
 	int devnum;
@@ -1048,6 +1049,9 @@ static void ib_uverbs_add_one(struct ib_device *device)
 	rcu_assign_pointer(uverbs_dev->ib_dev, device);
 	uverbs_dev->num_comp_vectors = device->num_comp_vectors;
 
+	if (ib_uverbs_create_uapi(device, uverbs_dev))
+		goto err;
+
 	cdev_init(&uverbs_dev->cdev, NULL);
 	uverbs_dev->cdev.owner = THIS_MODULE;
 	uverbs_dev->cdev.ops = device->mmap ? &uverbs_mmap_fops : &uverbs_fops;
@@ -1067,18 +1071,6 @@ static void ib_uverbs_add_one(struct ib_device *device)
 	if (device_create_file(uverbs_dev->dev, &dev_attr_abi_version))
 		goto err_class;
 
-	if (!device->specs_root) {
-		const struct uverbs_object_tree_def *default_root[] = {
-			uverbs_default_get_objects()};
-
-		uverbs_dev->specs_root = uverbs_alloc_spec_tree(1,
-								default_root);
-		if (IS_ERR(uverbs_dev->specs_root))
-			goto err_class;
-
-		device->specs_root = uverbs_dev->specs_root;
-	}
-
 	ib_set_client_data(device, &uverbs_client, uverbs_dev);
 
 	return;
@@ -1098,44 +1090,6 @@ err:
 	return;
 }
 
-static void ib_uverbs_disassociate_ucontext(struct ib_ucontext *ibcontext)
-{
-	struct ib_device *ib_dev = ibcontext->device;
-	struct task_struct *owning_process  = NULL;
-	struct mm_struct   *owning_mm       = NULL;
-
-	owning_process = get_pid_task(ibcontext->tgid, PIDTYPE_PID);
-	if (!owning_process)
-		return;
-
-	owning_mm = get_task_mm(owning_process);
-	if (!owning_mm) {
-		pr_info("no mm, disassociate ucontext is pending task termination\n");
-		while (1) {
-			put_task_struct(owning_process);
-			usleep_range(1000, 2000);
-			owning_process = get_pid_task(ibcontext->tgid,
-						      PIDTYPE_PID);
-			if (!owning_process ||
-			    owning_process->state == TASK_DEAD) {
-				pr_info("disassociate ucontext done, task was terminated\n");
-				/* in case task was dead need to release the
-				 * task struct.
-				 */
-				if (owning_process)
-					put_task_struct(owning_process);
-				return;
-			}
-		}
-	}
-
-	down_write(&owning_mm->mmap_sem);
-	ib_dev->disassociate_ucontext(ibcontext);
-	up_write(&owning_mm->mmap_sem);
-	mmput(owning_mm);
-	put_task_struct(owning_process);
-}
-
 static void ib_uverbs_free_hw_resources(struct ib_uverbs_device *uverbs_dev,
 					struct ib_device *ib_dev)
 {
@@ -1144,46 +1098,31 @@ static void ib_uverbs_free_hw_resources(struct ib_uverbs_device *uverbs_dev,
 	struct ib_event event;
 
 	/* Pending running commands to terminate */
-	synchronize_srcu(&uverbs_dev->disassociate_srcu);
+	uverbs_disassociate_api_pre(uverbs_dev);
 	event.event = IB_EVENT_DEVICE_FATAL;
 	event.element.port_num = 0;
 	event.device = ib_dev;
 
 	mutex_lock(&uverbs_dev->lists_mutex);
 	while (!list_empty(&uverbs_dev->uverbs_file_list)) {
-		struct ib_ucontext *ucontext;
 		file = list_first_entry(&uverbs_dev->uverbs_file_list,
 					struct ib_uverbs_file, list);
 		file->is_closed = 1;
 		list_del(&file->list);
 		kref_get(&file->ref);
-		mutex_unlock(&uverbs_dev->lists_mutex);
-
-
-		mutex_lock(&file->cleanup_mutex);
-		ucontext = file->ucontext;
-		file->ucontext = NULL;
-		mutex_unlock(&file->cleanup_mutex);
 
-		/* At this point ib_uverbs_close cannot be running
-		 * ib_uverbs_cleanup_ucontext
+		/* We must release the mutex before going ahead and calling
+		 * uverbs_cleanup_ufile, as it might end up indirectly calling
+		 * uverbs_close, for example due to freeing the resources (e.g
+		 * mmput).
 		 */
-		if (ucontext) {
-			/* We must release the mutex before going ahead and
-			 * calling disassociate_ucontext. disassociate_ucontext
-			 * might end up indirectly calling uverbs_close,
-			 * for example due to freeing the resources
-			 * (e.g mmput).
-			 */
-			ib_uverbs_event_handler(&file->event_handler, &event);
-			ib_uverbs_disassociate_ucontext(ucontext);
-			mutex_lock(&file->cleanup_mutex);
-			ib_uverbs_cleanup_ucontext(file, ucontext, true);
-			mutex_unlock(&file->cleanup_mutex);
-		}
+		mutex_unlock(&uverbs_dev->lists_mutex);
 
-		mutex_lock(&uverbs_dev->lists_mutex);
+		ib_uverbs_event_handler(&file->event_handler, &event);
+		uverbs_destroy_ufile_hw(file, RDMA_REMOVE_DRIVER_REMOVE);
 		kref_put(&file->ref, ib_uverbs_release_file);
+
+		mutex_lock(&uverbs_dev->lists_mutex);
 	}
 
 	while (!list_empty(&uverbs_dev->uverbs_events_file_list)) {
@@ -1205,6 +1144,8 @@ static void ib_uverbs_free_hw_resources(struct ib_uverbs_device *uverbs_dev,
 		kill_fasync(&event_file->ev_queue.async_queue, SIGIO, POLL_IN);
 	}
 	mutex_unlock(&uverbs_dev->lists_mutex);
+
+	uverbs_disassociate_api(uverbs_dev->uapi);
 }
 
 static void ib_uverbs_remove_one(struct ib_device *device, void *client_data)
@@ -1232,7 +1173,6 @@ static void ib_uverbs_remove_one(struct ib_device *device, void *client_data)
 		 * cdev was deleted, however active clients can still issue
 		 * commands and close their open files.
 		 */
-		rcu_assign_pointer(uverbs_dev->ib_dev, NULL);
 		ib_uverbs_free_hw_resources(uverbs_dev, device);
 		wait_clients = 0;
 	}
@@ -1241,10 +1181,6 @@ static void ib_uverbs_remove_one(struct ib_device *device, void *client_data)
 		ib_uverbs_comp_dev(uverbs_dev);
 	if (wait_clients)
 		wait_for_completion(&uverbs_dev->comp);
-	if (uverbs_dev->specs_root) {
-		uverbs_free_spec_tree(uverbs_dev->specs_root);
-		device->specs_root = NULL;
-	}
 
 	kobject_put(&uverbs_dev->kobj);
 }