linux/drivers/infiniband/core/rdma_core.h

212 lines
7.5 KiB
C
Raw Normal View History

/*
* Copyright (c) 2005 Topspin Communications. All rights reserved.
* Copyright (c) 2005, 2006 Cisco Systems. All rights reserved.
* Copyright (c) 2005-2017 Mellanox Technologies. All rights reserved.
* Copyright (c) 2005 Voltaire, Inc. All rights reserved.
* Copyright (c) 2005 PathScale, Inc. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#ifndef RDMA_CORE_H
#define RDMA_CORE_H
#include <linux/idr.h>
#include <rdma/uverbs_types.h>
IB/core: Add a generic way to execute an operation on a uobject The ioctl infrastructure treats all user-objects in the same manner. It gets objects ids from the user-space and by using the object type and type attributes mentioned in the object specification, it executes this required method. Passing an object id from the user-space as an attribute is carried out in three stages. The first is carried out before the actual handler and the last is carried out afterwards. The different supported operations are read, write, destroy and create. In the first stage, the former three actions just fetches the object from the repository (by using its id) and locks it. The last action allocates a new uobject. Afterwards, the second stage is carried out when the handler itself carries out the required modification of the object. The last stage is carried out after the handler finishes and commits the result. The former two operations just unlock the object. Destroy calls the "free object" operation, taking into account the object's type and releases the uobject as well. Creation just adds the new uobject to the repository, making the object visible to the application. In order to abstract these details from the ioctl infrastructure layer, we add uverbs_get_uobject_from_context and uverbs_finalize_object functions which corresponds to the first and last stages respectively. Signed-off-by: Matan Barak <matanb@mellanox.com> Reviewed-by: Yishai Hadas <yishaih@mellanox.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2017-08-03 21:06:55 +08:00
#include <rdma/uverbs_ioctl.h>
#include <rdma/ib_verbs.h>
#include <linux/mutex.h>
struct ib_uverbs_device;
void uverbs_destroy_ufile_hw(struct ib_uverbs_file *ufile,
enum rdma_remove_reason reason);
int uobj_destroy(struct ib_uobject *uobj);
/*
* uverbs_uobject_get is called in order to increase the reference count on
* an uobject. This is useful when a handler wants to keep the uobject's memory
* alive, regardless if this uobject is still alive in the context's objects
* repository. Objects are put via uverbs_uobject_put.
*/
void uverbs_uobject_get(struct ib_uobject *uobject);
/*
* In order to indicate we no longer needs this uobject, uverbs_uobject_put
* is called. When the reference count is decreased, the uobject is freed.
* For example, this is used when attaching a completion channel to a CQ.
*/
void uverbs_uobject_put(struct ib_uobject *uobject);
/* Indicate this fd is no longer used by this consumer, but its memory isn't
* necessarily released yet. When the last reference is put, we release the
* memory. After this call is executed, calling uverbs_uobject_get isn't
* allowed.
* This must be called from the release file_operations of the file!
*/
void uverbs_close_fd(struct file *f);
IB/core: Add a generic way to execute an operation on a uobject The ioctl infrastructure treats all user-objects in the same manner. It gets objects ids from the user-space and by using the object type and type attributes mentioned in the object specification, it executes this required method. Passing an object id from the user-space as an attribute is carried out in three stages. The first is carried out before the actual handler and the last is carried out afterwards. The different supported operations are read, write, destroy and create. In the first stage, the former three actions just fetches the object from the repository (by using its id) and locks it. The last action allocates a new uobject. Afterwards, the second stage is carried out when the handler itself carries out the required modification of the object. The last stage is carried out after the handler finishes and commits the result. The former two operations just unlock the object. Destroy calls the "free object" operation, taking into account the object's type and releases the uobject as well. Creation just adds the new uobject to the repository, making the object visible to the application. In order to abstract these details from the ioctl infrastructure layer, we add uverbs_get_uobject_from_context and uverbs_finalize_object functions which corresponds to the first and last stages respectively. Signed-off-by: Matan Barak <matanb@mellanox.com> Reviewed-by: Yishai Hadas <yishaih@mellanox.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2017-08-03 21:06:55 +08:00
/*
* Get an ib_uobject that corresponds to the given id from ufile, assuming
IB/core: Add a generic way to execute an operation on a uobject The ioctl infrastructure treats all user-objects in the same manner. It gets objects ids from the user-space and by using the object type and type attributes mentioned in the object specification, it executes this required method. Passing an object id from the user-space as an attribute is carried out in three stages. The first is carried out before the actual handler and the last is carried out afterwards. The different supported operations are read, write, destroy and create. In the first stage, the former three actions just fetches the object from the repository (by using its id) and locks it. The last action allocates a new uobject. Afterwards, the second stage is carried out when the handler itself carries out the required modification of the object. The last stage is carried out after the handler finishes and commits the result. The former two operations just unlock the object. Destroy calls the "free object" operation, taking into account the object's type and releases the uobject as well. Creation just adds the new uobject to the repository, making the object visible to the application. In order to abstract these details from the ioctl infrastructure layer, we add uverbs_get_uobject_from_context and uverbs_finalize_object functions which corresponds to the first and last stages respectively. Signed-off-by: Matan Barak <matanb@mellanox.com> Reviewed-by: Yishai Hadas <yishaih@mellanox.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2017-08-03 21:06:55 +08:00
* the object is from the given type. Lock it to the required access when
* applicable.
* This function could create (access == NEW), destroy (access == DESTROY)
* or unlock (access == READ || access == WRITE) objects if required.
IB/core: Add support to finalize objects in one transaction The new ioctl based infrastructure either commits or rollbacks all objects of the method as one transaction. In order to do that, we introduce a notion of dealing with a collection of objects that are related to a specific method. This also requires adding a notion of a method and attribute. A method contains a hash of attributes, where each bucket contains several attributes. The attributes are hashed according to their namespace which resides in the four upper bits of the id. For example, an object could be a CQ, which has an action of CREATE_CQ. This action has multiple attributes. For example, the CQ's new handle and the comp_channel. Each layer in this hierarchy - objects, methods and attributes is split into namespaces. The basic example for that is one namespace representing the default entities and another one representing the driver specific entities. When declaring these methods and attributes, we actually declare their specifications. When a method is executed, we actually allocates some space to hold auxiliary information. This auxiliary information contains meta-data about the required objects, such as pointers to their type information, pointers to the uobjects themselves (if exist), etc. The specification, along with the auxiliary information we allocated and filled is given to the finalize_objects function. Signed-off-by: Matan Barak <matanb@mellanox.com> Reviewed-by: Yishai Hadas <yishaih@mellanox.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2017-08-03 21:06:56 +08:00
* The action will be finalized only when uverbs_finalize_object or
* uverbs_finalize_objects are called.
IB/core: Add a generic way to execute an operation on a uobject The ioctl infrastructure treats all user-objects in the same manner. It gets objects ids from the user-space and by using the object type and type attributes mentioned in the object specification, it executes this required method. Passing an object id from the user-space as an attribute is carried out in three stages. The first is carried out before the actual handler and the last is carried out afterwards. The different supported operations are read, write, destroy and create. In the first stage, the former three actions just fetches the object from the repository (by using its id) and locks it. The last action allocates a new uobject. Afterwards, the second stage is carried out when the handler itself carries out the required modification of the object. The last stage is carried out after the handler finishes and commits the result. The former two operations just unlock the object. Destroy calls the "free object" operation, taking into account the object's type and releases the uobject as well. Creation just adds the new uobject to the repository, making the object visible to the application. In order to abstract these details from the ioctl infrastructure layer, we add uverbs_get_uobject_from_context and uverbs_finalize_object functions which corresponds to the first and last stages respectively. Signed-off-by: Matan Barak <matanb@mellanox.com> Reviewed-by: Yishai Hadas <yishaih@mellanox.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2017-08-03 21:06:55 +08:00
*/
struct ib_uobject *
uverbs_get_uobject_from_file(u16 object_id,
struct ib_uverbs_file *ufile,
enum uverbs_obj_access access, s64 id);
IB/core: Add support to finalize objects in one transaction The new ioctl based infrastructure either commits or rollbacks all objects of the method as one transaction. In order to do that, we introduce a notion of dealing with a collection of objects that are related to a specific method. This also requires adding a notion of a method and attribute. A method contains a hash of attributes, where each bucket contains several attributes. The attributes are hashed according to their namespace which resides in the four upper bits of the id. For example, an object could be a CQ, which has an action of CREATE_CQ. This action has multiple attributes. For example, the CQ's new handle and the comp_channel. Each layer in this hierarchy - objects, methods and attributes is split into namespaces. The basic example for that is one namespace representing the default entities and another one representing the driver specific entities. When declaring these methods and attributes, we actually declare their specifications. When a method is executed, we actually allocates some space to hold auxiliary information. This auxiliary information contains meta-data about the required objects, such as pointers to their type information, pointers to the uobjects themselves (if exist), etc. The specification, along with the auxiliary information we allocated and filled is given to the finalize_objects function. Signed-off-by: Matan Barak <matanb@mellanox.com> Reviewed-by: Yishai Hadas <yishaih@mellanox.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2017-08-03 21:06:56 +08:00
/*
* Note that certain finalize stages could return a status:
* (a) alloc_commit could return a failure if the object is committed at the
* same time when the context is destroyed.
* (b) remove_commit could fail if the object wasn't destroyed successfully.
* Since multiple objects could be finalized in one transaction, it is very NOT
* recommended to have several finalize actions which have side effects.
* For example, it's NOT recommended to have a certain action which has both
* a commit action and a destroy action or two destroy objects in the same
* action. The rule of thumb is to have one destroy or commit action with
* multiple lookups.
* The first non zero return value of finalize_object is returned from this
* function. For example, this could happen when we couldn't destroy an
* object.
*/
int uverbs_finalize_object(struct ib_uobject *uobj,
enum uverbs_obj_access access,
bool commit);
IB/core: Add a generic way to execute an operation on a uobject The ioctl infrastructure treats all user-objects in the same manner. It gets objects ids from the user-space and by using the object type and type attributes mentioned in the object specification, it executes this required method. Passing an object id from the user-space as an attribute is carried out in three stages. The first is carried out before the actual handler and the last is carried out afterwards. The different supported operations are read, write, destroy and create. In the first stage, the former three actions just fetches the object from the repository (by using its id) and locks it. The last action allocates a new uobject. Afterwards, the second stage is carried out when the handler itself carries out the required modification of the object. The last stage is carried out after the handler finishes and commits the result. The former two operations just unlock the object. Destroy calls the "free object" operation, taking into account the object's type and releases the uobject as well. Creation just adds the new uobject to the repository, making the object visible to the application. In order to abstract these details from the ioctl infrastructure layer, we add uverbs_get_uobject_from_context and uverbs_finalize_object functions which corresponds to the first and last stages respectively. Signed-off-by: Matan Barak <matanb@mellanox.com> Reviewed-by: Yishai Hadas <yishaih@mellanox.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2017-08-03 21:06:55 +08:00
void setup_ufile_idr_uobject(struct ib_uverbs_file *ufile);
void release_ufile_idr_uobject(struct ib_uverbs_file *ufile);
/*
* This is the runtime description of the uverbs API, used by the syscall
* machinery to validate and dispatch calls.
*/
/*
* Depending on ID the slot pointer in the radix tree points at one of these
* structs.
*/
struct uverbs_api_object {
const struct uverbs_obj_type *type_attrs;
const struct uverbs_obj_type_class *type_class;
u8 disabled:1;
};
struct uverbs_api_ioctl_method {
int(__rcu *handler)(struct uverbs_attr_bundle *attrs);
DECLARE_BITMAP(attr_mandatory, UVERBS_API_ATTR_BKEY_LEN);
u16 bundle_size;
u8 use_stack:1;
u8 driver_method:1;
u8 disabled:1;
u8 has_udata:1;
u8 key_bitmap_len;
u8 destroy_bkey;
};
struct uverbs_api_write_method {
int (*handler)(struct uverbs_attr_bundle *attrs, const char __user *buf,
int in_len, int out_len);
int (*handler_ex)(struct uverbs_attr_bundle *attrs,
struct ib_udata *ucore);
u8 disabled:1;
u8 is_ex:1;
u8 has_udata:1;
u8 has_resp:1;
u8 req_size;
u8 resp_size;
};
struct uverbs_api_attr {
struct uverbs_attr_spec spec;
};
struct uverbs_api {
/* radix tree contains struct uverbs_api_* pointers */
struct radix_tree_root radix;
enum rdma_driver_id driver_id;
unsigned int num_write;
unsigned int num_write_ex;
struct uverbs_api_write_method notsupp_method;
const struct uverbs_api_write_method **write_methods;
const struct uverbs_api_write_method **write_ex_methods;
};
static inline const struct uverbs_api_object *
uapi_get_object(struct uverbs_api *uapi, u16 object_id)
{
return radix_tree_lookup(&uapi->radix, uapi_key_obj(object_id));
}
char *uapi_key_format(char *S, unsigned int key);
struct uverbs_api *uverbs_alloc_api(struct ib_device *ibdev);
void uverbs_disassociate_api_pre(struct ib_uverbs_device *uverbs_dev);
void uverbs_disassociate_api(struct uverbs_api *uapi);
void uverbs_destroy_api(struct uverbs_api *uapi);
void uapi_compute_bundle_size(struct uverbs_api_ioctl_method *method_elm,
unsigned int num_attrs);
RDMA/ucontext: Add a core API for mmaping driver IO memory To support disassociation and PCI hot unplug, we have to track all the VMAs that refer to the device IO memory. When disassociation occurs the VMAs have to be revised to point to the zero page, not the IO memory, to allow the physical HW to be unplugged. The three drivers supporting this implemented three different versions of this algorithm, all leaving something to be desired. This new common implementation has a few differences from the driver versions: - Track all VMAs, including splitting/truncating/etc. Tie the lifetime of the private data allocation to the lifetime of the vma. This avoids any tricks with setting vm_ops which Linus didn't like. (see link) - Support multiple mms, and support properly tracking mmaps triggered by processes other than the one first opening the uverbs fd. This makes fork behavior of disassociation enabled drivers the same as fork support in normal drivers. - Don't use crazy get_task stuff. - Simplify the approach for to racing between vm_ops close and disassociation, fixing the related bugs most of the driver implementations had. Since we are in core code the tracking list can be placed in struct ib_uverbs_ufile, which has a lifetime strictly longer than any VMAs created by mmap on the uverbs FD. Link: https://www.spinics.net/lists/stable/msg248747.html Link: https://lkml.kernel.org/r/CA+55aFxJTV_g46AQPoPXen-UPiqR1HGMZictt7VpC-SMFbm3Cw@mail.gmail.com Signed-off-by: Jason Gunthorpe <jgg@mellanox.com> Signed-off-by: Leon Romanovsky <leonro@mellanox.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2018-09-17 01:43:08 +08:00
void uverbs_user_mmap_disassociate(struct ib_uverbs_file *ufile);
extern const struct uapi_definition uverbs_def_obj_counters[];
extern const struct uapi_definition uverbs_def_obj_cq[];
extern const struct uapi_definition uverbs_def_obj_dm[];
extern const struct uapi_definition uverbs_def_obj_flow_action[];
extern const struct uapi_definition uverbs_def_obj_intf[];
extern const struct uapi_definition uverbs_def_obj_mr[];
extern const struct uapi_definition uverbs_def_write_intf[];
static inline const struct uverbs_api_write_method *
uapi_get_method(const struct uverbs_api *uapi, u32 command)
{
u32 cmd_idx = command & IB_USER_VERBS_CMD_COMMAND_MASK;
if (command & ~(u32)(IB_USER_VERBS_CMD_FLAG_EXTENDED |
IB_USER_VERBS_CMD_COMMAND_MASK))
return ERR_PTR(-EINVAL);
if (command & IB_USER_VERBS_CMD_FLAG_EXTENDED) {
if (cmd_idx >= uapi->num_write_ex)
return ERR_PTR(-EOPNOTSUPP);
return uapi->write_ex_methods[cmd_idx];
}
if (cmd_idx >= uapi->num_write)
return ERR_PTR(-EOPNOTSUPP);
return uapi->write_methods[cmd_idx];
}
#endif /* RDMA_CORE_H */