IB/core: Add new ioctl interface

In this ioctl interface, processing the command starts from
properties of the command and fetching the appropriate user objects
before calling the handler.

Parsing and validation is done according to a specifier declared by
the driver's code. In the driver, all supported objects are declared.
These objects are separated to different object namepsaces. Dividing
objects to namespaces is done at initialization by using the higher
bits of the object ids. This initialization can mix objects declared
in different places to one parsing tree using in this ioctl interface.

For each object we list all supported methods. Similarly to objects,
methods are separated to method namespaces too. Namespacing is done
similarly to the objects case. This could be used in order to add
methods to an existing object.

Each method has a specific handler, which could be either a default
handler or a driver specific handler.
Along with the handler, a bunch of attributes are specified as well.
Similarly to objects and method, attributes are namespaced and hashed
by their ids at initialization too. All supported attributes are
subject to automatic fetching and validation. These attributes include
the command, response and the method's related objects' ids.

When these entities (objects, methods and attributes) are used, the
high bits of the entities ids are used in order to calculate the hash
bucket index. Then, these high bits are masked out in order to have a
zero based index. Since we use these high bits for both bucketing and
namespacing, we get a compact representation and O(1) array access.
This is mandatory for efficient dispatching.

Each attribute has a type (PTR_IN, PTR_OUT, IDR and FD) and a length.
Attributes could be validated through some attributes, like:
(*) Minimum size / Exact size
(*) Fops for FD
(*) Object type for IDR

If an IDR/fd attribute is specified, the kernel also states the object
type and the required access (NEW, WRITE, READ or DESTROY).
All uobject/fd management is done automatically by the infrastructure,
meaning - the infrastructure will fail concurrent commands that at
least one of them requires concurrent access (WRITE/DESTROY),
synchronize actions with device removals (dissociate context events)
and take care of reference counting (increase/decrease) for concurrent
actions invocation. The reference counts on the actual kernel objects
shall be handled by the handlers.

 objects
+--------+
|        |
|        |   methods                                                                +--------+
|        |   ns         method      method_spec                           +-----+   |len     |
+--------+  +------+[d]+-------+   +----------------+[d]+------------+    |attr1+-> |type    |
| object +> |method+-> | spec  +-> +  attr_buckets  +-> |default_chain+--> +-----+   |idr_type|
+--------+  +------+   |handler|   |                |   +------------+    |attr2|   |access  |
|        |  |      |   +-------+   +----------------+   |driver chain|    +-----+   +--------+
|        |  |      |                                    +------------+
|        |  +------+
|        |
|        |
|        |
|        |
|        |
|        |
|        |
|        |
|        |
|        |
+--------+

[d] = Hash ids to groups using the high order bits

The right types table is also chosen by using the high bits from
the ids. Currently we have either default or driver specific groups.

Once validation and object fetching (or creation) completed, we call
the handler:
int (*handler)(struct ib_device *ib_dev, struct ib_uverbs_file *ufile,
               struct uverbs_attr_bundle *ctx);

ctx bundles attributes of different namespaces. Each element there
is an array of attributes which corresponds to one namespaces of
attributes. For example, in the usually used case:

 ctx                               core
+----------------------------+     +------------+
| core:                      +---> | valid      |
+----------------------------+     | cmd_attr   |
| driver:                    |     +------------+
|----------------------------+--+  | valid      |
                                |  | cmd_attr   |
                                |  +------------+
                                |  | valid      |
                                |  | obj_attr   |
                                |  +------------+
                                |
                                |  drivers
                                |  +------------+
                                +> | valid      |
                                   | cmd_attr   |
                                   +------------+
                                   | valid      |
                                   | cmd_attr   |
                                   +------------+
                                   | valid      |
                                   | obj_attr   |
                                   +------------+
Signed-off-by: Matan Barak <matanb@mellanox.com>
Reviewed-by: Yishai Hadas <yishaih@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>

drivers/infiniband/core/Makefile

@@ -32,4 +32,4 @@ ib_umad-y :=			user_mad.o
 ib_ucm-y :=			ucm.o
 
 ib_uverbs-y :=			uverbs_main.o uverbs_cmd.o uverbs_marshall.o \
-				rdma_core.o uverbs_std_types.o
+				rdma_core.o uverbs_std_types.o uverbs_ioctl.o

drivers/infiniband/core/rdma_core.c

@@ -36,10 +36,56 @@
 #include <rdma/uverbs_types.h>
 #include <linux/rcupdate.h>
 #include <rdma/uverbs_ioctl.h>
+#include <rdma/rdma_user_ioctl.h>
 #include "uverbs.h"
 #include "core_priv.h"
 #include "rdma_core.h"
 
+int uverbs_ns_idx(u16 *id, unsigned int ns_count)
+{
+	int ret = (*id & UVERBS_ID_NS_MASK) >> UVERBS_ID_NS_SHIFT;
+
+	if (ret >= ns_count)
+		return -EINVAL;
+
+	*id &= ~UVERBS_ID_NS_MASK;
+	return ret;
+}
+
+const struct uverbs_object_spec *uverbs_get_object(const struct ib_device *ibdev,
+						   uint16_t object)
+{
+	const struct uverbs_root_spec *object_hash = ibdev->specs_root;
+	const struct uverbs_object_spec_hash *objects;
+	int ret = uverbs_ns_idx(&object, object_hash->num_buckets);
+
+	if (ret < 0)
+		return NULL;
+
+	objects = object_hash->object_buckets[ret];
+
+	if (object >= objects->num_objects)
+		return NULL;
+
+	return objects->objects[object];
+}
+
+const struct uverbs_method_spec *uverbs_get_method(const struct uverbs_object_spec *object,
+						   uint16_t method)
+{
+	const struct uverbs_method_spec_hash *methods;
+	int ret = uverbs_ns_idx(&method, object->num_buckets);
+
+	if (ret < 0)
+		return NULL;
+
+	methods = object->method_buckets[ret];
+	if (method >= methods->num_methods)
+		return NULL;
+
+	return methods->methods[method];
+}
+
 void uverbs_uobject_get(struct ib_uobject *uobject)
 {
 	kref_get(&uobject->ref);

drivers/infiniband/core/rdma_core.h

@@ -43,6 +43,11 @@
 #include <rdma/ib_verbs.h>
 #include <linux/mutex.h>
 
+int uverbs_ns_idx(u16 *id, unsigned int ns_count);
+const struct uverbs_object_spec *uverbs_get_object(const struct ib_device *ibdev,
+						   uint16_t object);
+const struct uverbs_method_spec *uverbs_get_method(const struct uverbs_object_spec *object,
+						   uint16_t method);
 /*
  * These functions initialize the context and cleanups its uobjects.
  * The context has a list of objects which is protected by a mutex

include/rdma/ib_verbs.h

@@ -2348,6 +2348,8 @@ struct ib_device {
 	void (*get_dev_fw_str)(struct ib_device *, char *str);
 	const struct cpumask *(*get_vector_affinity)(struct ib_device *ibdev,
 						     int comp_vector);
+
+	struct uverbs_root_spec		*specs_root;
 };
 
 struct ib_client {

include/rdma/uverbs_ioctl.h

@@ -43,6 +43,8 @@
 
 enum uverbs_attr_type {
 	UVERBS_ATTR_TYPE_NA,
+	UVERBS_ATTR_TYPE_PTR_IN,
+	UVERBS_ATTR_TYPE_PTR_OUT,
 	UVERBS_ATTR_TYPE_IDR,
 	UVERBS_ATTR_TYPE_FD,
 };
@@ -54,29 +56,110 @@ enum uverbs_obj_access {
 	UVERBS_ACCESS_DESTROY
 };
 
+enum {
+	UVERBS_ATTR_SPEC_F_MANDATORY	= 1U << 0,
+	/* Support extending attributes by length */
+	UVERBS_ATTR_SPEC_F_MIN_SZ	= 1U << 1,
+};
+
 struct uverbs_attr_spec {
 	enum uverbs_attr_type		type;
-	struct {
-		/*
-		 * higher bits mean the namespace and lower bits mean
-		 * the type id within the namespace.
-		 */
-		u16			obj_type;
-		u8			access;
-	} obj;
+	union {
+		u16				len;
+		struct {
+			/*
+			 * higher bits mean the namespace and lower bits mean
+			 * the type id within the namespace.
+			 */
+			u16			obj_type;
+			u8			access;
+		} obj;
+	};
+	/* Combination of bits from enum UVERBS_ATTR_SPEC_F_XXXX */
+	u8				flags;
 };
 
 struct uverbs_attr_spec_hash {
 	size_t				num_attrs;
+	unsigned long			*mandatory_attrs_bitmask;
 	struct uverbs_attr_spec		attrs[0];
 };
 
+struct uverbs_attr_bundle;
+struct ib_uverbs_file;
+
+enum {
+	/*
+	 * Action marked with this flag creates a context (or root for all
+	 * objects).
+	 */
+	UVERBS_ACTION_FLAG_CREATE_ROOT = 1U << 0,
+};
+
+struct uverbs_method_spec {
+	/* Combination of bits from enum UVERBS_ACTION_FLAG_XXXX */
+	u32						flags;
+	size_t						num_buckets;
+	size_t						num_child_attrs;
+	int (*handler)(struct ib_device *ib_dev, struct ib_uverbs_file *ufile,
+		       struct uverbs_attr_bundle *ctx);
+	struct uverbs_attr_spec_hash		*attr_buckets[0];
+};
+
+struct uverbs_method_spec_hash {
+	size_t					num_methods;
+	struct uverbs_method_spec		*methods[0];
+};
+
+struct uverbs_object_spec {
+	const struct uverbs_obj_type		*type_attrs;
+	size_t					num_buckets;
+	struct uverbs_method_spec_hash		*method_buckets[0];
+};
+
+struct uverbs_object_spec_hash {
+	size_t					num_objects;
+	struct uverbs_object_spec		*objects[0];
+};
+
+struct uverbs_root_spec {
+	size_t					num_buckets;
+	struct uverbs_object_spec_hash		*object_buckets[0];
+};
+
+/* =================================================
+ *              Parsing infrastructure
+ * =================================================
+ */
+
+struct uverbs_ptr_attr {
+	union {
+		u64		data;
+		void	__user *ptr;
+	};
+	u16		len;
+	/* Combination of bits from enum UVERBS_ATTR_F_XXXX */
+	u16		flags;
+};
+
 struct uverbs_obj_attr {
+	/* pointer to the kernel descriptor -> type, access, etc */
+	const struct uverbs_obj_type	*type;
 	struct ib_uobject		*uobject;
+	/* fd or id in idr of this object */
+	int				id;
 };
 
 struct uverbs_attr {
-	struct uverbs_obj_attr	obj_attr;
+	/*
+	 * pointer to the user-space given attribute, in order to write the
+	 * new uobject's id or update flags.
+	 */
+	struct ib_uverbs_attr __user	*uattr;
+	union {
+		struct uverbs_ptr_attr	ptr_attr;
+		struct uverbs_obj_attr	obj_attr;
+	};
 };
 
 struct uverbs_attr_bundle_hash {

include/uapi/rdma/rdma_user_ioctl.h

@@ -43,6 +43,39 @@
 /* Legacy name, for user space application which already use it */
 #define IB_IOCTL_MAGIC		RDMA_IOCTL_MAGIC
 
+#define RDMA_VERBS_IOCTL \
+	_IOWR(RDMA_IOCTL_MAGIC, 1, struct ib_uverbs_ioctl_hdr)
+
+#define UVERBS_ID_NS_MASK 0xF000
+#define UVERBS_ID_NS_SHIFT 12
+
+enum {
+	/* User input */
+	UVERBS_ATTR_F_MANDATORY = 1U << 0,
+	/*
+	 * Valid output bit should be ignored and considered set in
+	 * mandatory fields. This bit is kernel output.
+	 */
+	UVERBS_ATTR_F_VALID_OUTPUT = 1U << 1,
+};
+
+struct ib_uverbs_attr {
+	__u16 attr_id;		/* command specific type attribute */
+	__u16 len;		/* only for pointers */
+	__u16 flags;		/* combination of UVERBS_ATTR_F_XXXX */
+	__u16 reserved;
+	__u64 data;		/* ptr to command, inline data or idr/fd */
+};
+
+struct ib_uverbs_ioctl_hdr {
+	__u16 length;
+	__u16 object_id;
+	__u16 method_id;
+	__u16 num_attrs;
+	__u64 reserved;
+	struct ib_uverbs_attr  attrs[0];
+};
+
 /*
  * General blocks assignments
  * It is closed on purpose do not expose it it user space