[bpf-next,v3] docs/bpf: document BPF ARRAY_OF_MAPS and HASH_OF_MAPS

Commit Message

Donald Hunter Nov. 7, 2022, 1:48 p.m. UTC

Add documentation for the ARRAY_OF_MAPS and HASH_OF_MAPS map types,
including usage and examples.

Signed-off-by: Donald Hunter <[e-mail]>
---
v2 -> v3:
- Update BPF example to show declarative initialisation, as
  suggested by Andrii Nakryiko
- Use LIBBPF_OPTS inline initialisation, as suggested by
  Andrii Nakryiko
- Fix duplicate label warning,
Reported-by: kernel test robot <[e-mail]>

v1 -> v2:
- Fix formatting nits
- Tidy up code snippets as suggested by Maryam Tahhan
---
 Documentation/bpf/map_of_maps.rst | 126 ++++++++++++++++++++++++++++++
 1 file changed, 126 insertions(+)
 create mode 100644 Documentation/bpf/map_of_maps.rst

Comments

Yonghong Song Nov. 7, 2022, 5:46 p.m. UTC | #1

On 11/7/22 5:48 AM, Donald Hunter wrote:
> Add documentation for the ARRAY_OF_MAPS and HASH_OF_MAPS map types,
> including usage and examples.
> 
> Signed-off-by: Donald Hunter <[e-mail]>
> ---
> v2 -> v3:
> - Update BPF example to show declarative initialisation, as
>    suggested by Andrii Nakryiko
> - Use LIBBPF_OPTS inline initialisation, as suggested by
>    Andrii Nakryiko
> - Fix duplicate label warning,
> Reported-by: kernel test robot <[e-mail]>
> 
> v1 -> v2:
> - Fix formatting nits
> - Tidy up code snippets as suggested by Maryam Tahhan
> ---
>   Documentation/bpf/map_of_maps.rst | 126 ++++++++++++++++++++++++++++++
>   1 file changed, 126 insertions(+)
>   create mode 100644 Documentation/bpf/map_of_maps.rst
> 
> diff --git a/Documentation/bpf/map_of_maps.rst b/Documentation/bpf/map_of_maps.rst
> new file mode 100644
> index 000000000000..63e41b06a91d
> --- /dev/null
> +++ b/Documentation/bpf/map_of_maps.rst
> @@ -0,0 +1,126 @@
> +.. SPDX-License-Identifier: GPL-2.0-only
> +.. Copyright (C) 2022 Red Hat, Inc.
> +
> +========================================================
> +BPF_MAP_TYPE_ARRAY_OF_MAPS and BPF_MAP_TYPE_HASH_OF_MAPS
> +========================================================
> +
> +.. note::
> +   - ``BPF_MAP_TYPE_ARRAY_OF_MAPS`` and ``BPF_MAP_TYPE_HASH_OF_MAPS`` were
> +     introduced in kernel version 4.12
> +
> +``BPF_MAP_TYPE_ARRAY_OF_MAPS`` and ``BPF_MAP_TYPE_HASH_OF_MAPS`` provide general
> +purpose support for map in map storage. One level of nesting is supported, where
> +an outer map contains instances of a single type of inner map, for example
> +``array_of_maps->sock_map``.
> +
> +When creating an outer map, an inner map instance is used to initialize the
> +metadata that the outer map holds about its inner maps. This inner map has a
> +separate lifetime from the outer map and can be deleted after the outer map has
> +been created.
> +
> +The outer map supports element update and delete from user space using the
> +syscall API. A BPF program is only allowed to do element lookup in the outer
> +map.

The outer map supports element lookup, update and delete from user space 
using the syscall API.

A BPF program can do element delete for array/hash_of_maps. Please 
double check.

> +
> +.. note::
> +   - Multi-level nesting is not supported.
> +   - Any BPF map type can be used as an inner map, except for
> +     ``BPF_MAP_TYPE_PROG_ARRAY``.
> +   - A BPF program cannot update or delete outer map entries.

A BPF program cannot update outer map entries.

> +
> +For ``BPF_MAP_TYPE_ARRAY_OF_MAPS`` the key is an unsigned 32-bit integer index
> +into the array. The array is a fixed size with ``max_entries`` elements that are
> +zero initialized when created.
> +
> +For ``BPF_MAP_TYPE_HASH_OF_MAPS`` the key type can be chosen when defining the
> +map. The kernel is responsible for allocating and freeing key/value pairs, up to
> +the max_entries limit that you specify. Hash maps use pre-allocation of hash
> +table elements by default. The ``BPF_F_NO_PREALLOC`` flag can be used to disable
> +pre-allocation when it is too memory expensive.
> +
> +Usage
> +=====
> +
> +Kernel BPF Helper
> +-----------------
> +
> +.. c:function::
> +   void *bpf_map_lookup_elem(struct bpf_map *map, const void *key)
> +
> +Inner maps can be retrieved using the ``bpf_map_lookup_elem()`` helper. This
> +helper returns a pointer to the inner map, or ``NULL`` if no entry was found.

bpf_map_delete_elem?

> +
> +Examples
> +========
> +
> +Kernel BPF Example
> +------------------
> +
> +This snippet shows how to create and initialise an array of devmaps in a BPF
> +program. Note that the outer array can only be modified from user space using
> +the syscall API.
> +
> +.. code-block:: c
> +
> +    struct inner_map {
> +            __uint(type, BPF_MAP_TYPE_DEVMAP);
> +            __uint(max_entries, 10);
> +            __type(key, __u32);
> +            __type(value, __u32);
> +    } inner_map1 SEC(".maps"), inner_map2 SEC(".maps");
> +
> +    struct {
> +            __uint(type, BPF_MAP_TYPE_ARRAY_OF_MAPS);
> +            __uint(max_entries, 2);
> +            __type(key, __u32);
> +            __array(values, struct inner_map);
> +    } outer_map SEC(".maps") = {
> +            .values = { &inner_map1,
> +                        &inner_map2 }
> +    };
> +
> +See ``progs/test_bpf_map_in_map.c`` in ``tools/testing/selftests/bpf`` for more

The file name test_bpf_map_in_map.c` does not exist.

> +examples of declarative initialisation of outer maps.
> +
> +User Space
> +----------
> +
> +This snippet shows how to create an array based outer map:
> +
> +.. code-block:: c
> +
> +    int create_outer_array(int inner_fd) {
> +            int fd;
> +
> +            LIBBPF_OPTS(bpf_map_create_opts, opts, .inner_map_fd = inner_fd);

This is declaration. Please put it adjacent to 'int fd'.

> +            fd = bpf_map_create(BPF_MAP_TYPE_ARRAY_OF_MAPS,
> +                                "example_array",       /* name */
> +                                sizeof(__u32),         /* key size */
> +                                sizeof(__u32),         /* value size */
> +                                256,                   /* max entries */
> +                                &opts);                /* create opts */
> +            return fd;
> +    }
> +
> +
> +This snippet shows how to add an inner map to an outer map:
> +
> +.. code-block:: c
> +
> +    int add_devmap(int outer_fd, int index, const char *name) {
> +            int fd;
> +
> +            fd = bpf_map_create(BPF_MAP_TYPE_DEVMAP, name,
> +                                sizeof(__u32), sizeof(__u32), 256, NULL);
> +            if (fd < 0)
> +                    return fd;
> +
> +            return bpf_map_update_elem(outer_fd, &index, &fd, BPF_ANY);
> +    }
> +
> +References
> +==========
> +
> +- https://lore.kernel.org/netdev/[e-mail]/
> +- https://lore.kernel.org/netdev/[e-mail]/

Yonghong Song Nov. 7, 2022, 7:41 p.m. UTC | #2

On 11/7/22 9:46 AM, Yonghong Song wrote:
> 
> 
> On 11/7/22 5:48 AM, Donald Hunter wrote:
>> Add documentation for the ARRAY_OF_MAPS and HASH_OF_MAPS map types,
>> including usage and examples.
>>
>> Signed-off-by: Donald Hunter <[e-mail]>
>> ---
>> v2 -> v3:
>> - Update BPF example to show declarative initialisation, as
>>    suggested by Andrii Nakryiko
>> - Use LIBBPF_OPTS inline initialisation, as suggested by
>>    Andrii Nakryiko
>> - Fix duplicate label warning,
>> Reported-by: kernel test robot <[e-mail]>
>>
>> v1 -> v2:
>> - Fix formatting nits
>> - Tidy up code snippets as suggested by Maryam Tahhan
>> ---
>>   Documentation/bpf/map_of_maps.rst | 126 ++++++++++++++++++++++++++++++
>>   1 file changed, 126 insertions(+)
>>   create mode 100644 Documentation/bpf/map_of_maps.rst
>>
>> diff --git a/Documentation/bpf/map_of_maps.rst 
>> b/Documentation/bpf/map_of_maps.rst
>> new file mode 100644
>> index 000000000000..63e41b06a91d
>> --- /dev/null
>> +++ b/Documentation/bpf/map_of_maps.rst
>> @@ -0,0 +1,126 @@
>> +.. SPDX-License-Identifier: GPL-2.0-only
>> +.. Copyright (C) 2022 Red Hat, Inc.
>> +
>> +========================================================
>> +BPF_MAP_TYPE_ARRAY_OF_MAPS and BPF_MAP_TYPE_HASH_OF_MAPS
>> +========================================================
>> +
>> +.. note::
>> +   - ``BPF_MAP_TYPE_ARRAY_OF_MAPS`` and ``BPF_MAP_TYPE_HASH_OF_MAPS`` 
>> were
>> +     introduced in kernel version 4.12
>> +
>> +``BPF_MAP_TYPE_ARRAY_OF_MAPS`` and ``BPF_MAP_TYPE_HASH_OF_MAPS`` 
>> provide general
>> +purpose support for map in map storage. One level of nesting is 
>> supported, where
>> +an outer map contains instances of a single type of inner map, for 
>> example
>> +``array_of_maps->sock_map``.
>> +
>> +When creating an outer map, an inner map instance is used to 
>> initialize the
>> +metadata that the outer map holds about its inner maps. This inner 
>> map has a
>> +separate lifetime from the outer map and can be deleted after the 
>> outer map has
>> +been created.
>> +
>> +The outer map supports element update and delete from user space 
>> using the
>> +syscall API. A BPF program is only allowed to do element lookup in 
>> the outer
>> +map.
> 
> The outer map supports element lookup, update and delete from user space 
> using the syscall API.
> 
> A BPF program can do element delete for array/hash_of_maps. Please 
> double check.

Okay, I double checked with verifier.c. You are right, only lookup
is supported for bpf programs.

> 
>> +
>> +.. note::
>> +   - Multi-level nesting is not supported.
>> +   - Any BPF map type can be used as an inner map, except for
>> +     ``BPF_MAP_TYPE_PROG_ARRAY``.
>> +   - A BPF program cannot update or delete outer map entries.
> 
> A BPF program cannot update outer map entries.

Yes, only lookup is allowed for bpf programs.

> 
>> +
>> +For ``BPF_MAP_TYPE_ARRAY_OF_MAPS`` the key is an unsigned 32-bit 
>> integer index
>> +into the array. The array is a fixed size with ``max_entries`` 
>> elements that are
>> +zero initialized when created.
>> +
>> +For ``BPF_MAP_TYPE_HASH_OF_MAPS`` the key type can be chosen when 
>> defining the
>> +map. The kernel is responsible for allocating and freeing key/value 
>> pairs, up to
>> +the max_entries limit that you specify. Hash maps use pre-allocation 
>> of hash
>> +table elements by default. The ``BPF_F_NO_PREALLOC`` flag can be used 
>> to disable
>> +pre-allocation when it is too memory expensive.
>> +
>> +Usage
>> +=====
>> +
>> +Kernel BPF Helper
>> +-----------------
>> +
>> +.. c:function::
>> +   void *bpf_map_lookup_elem(struct bpf_map *map, const void *key)
>> +
>> +Inner maps can be retrieved using the ``bpf_map_lookup_elem()`` 
>> helper. This
>> +helper returns a pointer to the inner map, or ``NULL`` if no entry 
>> was found.
> 
> bpf_map_delete_elem?

same here. bpf_map_delete_elem is not available for bpf programs.

> 
>> +
>> +Examples
>> +========
>> +
>> +Kernel BPF Example
>> +------------------
>> +
>> +This snippet shows how to create and initialise an array of devmaps 
>> in a BPF
>> +program. Note that the outer array can only be modified from user 
>> space using
>> +the syscall API.
>> +
>> +.. code-block:: c
>> +
>> +    struct inner_map {
>> +            __uint(type, BPF_MAP_TYPE_DEVMAP);
>> +            __uint(max_entries, 10);
>> +            __type(key, __u32);
>> +            __type(value, __u32);
>> +    } inner_map1 SEC(".maps"), inner_map2 SEC(".maps");
>> +
>> +    struct {
>> +            __uint(type, BPF_MAP_TYPE_ARRAY_OF_MAPS);
>> +            __uint(max_entries, 2);
>> +            __type(key, __u32);
>> +            __array(values, struct inner_map);
>> +    } outer_map SEC(".maps") = {
>> +            .values = { &inner_map1,
>> +                        &inner_map2 }
>> +    };
>> +
>> +See ``progs/test_bpf_map_in_map.c`` in 
>> ``tools/testing/selftests/bpf`` for more
> 
> The file name test_bpf_map_in_map.c` does not exist.
> 
>> +examples of declarative initialisation of outer maps.
>> +
>> +User Space
>> +----------
>> +
>> +This snippet shows how to create an array based outer map:
>> +
>> +.. code-block:: c
>> +
>> +    int create_outer_array(int inner_fd) {
>> +            int fd;
>> +
>> +            LIBBPF_OPTS(bpf_map_create_opts, opts, .inner_map_fd = 
>> inner_fd);
> 
> This is declaration. Please put it adjacent to 'int fd'.
> 
>> +            fd = bpf_map_create(BPF_MAP_TYPE_ARRAY_OF_MAPS,
>> +                                "example_array",       /* name */
>> +                                sizeof(__u32),         /* key size */
>> +                                sizeof(__u32),         /* value size */
>> +                                256,                   /* max entries */
>> +                                &opts);                /* create opts */
>> +            return fd;
>> +    }
>> +
>> +
>> +This snippet shows how to add an inner map to an outer map:
>> +
>> +.. code-block:: c
>> +
>> +    int add_devmap(int outer_fd, int index, const char *name) {
>> +            int fd;
>> +
>> +            fd = bpf_map_create(BPF_MAP_TYPE_DEVMAP, name,
>> +                                sizeof(__u32), sizeof(__u32), 256, 
>> NULL);
>> +            if (fd < 0)
>> +                    return fd;
>> +
>> +            return bpf_map_update_elem(outer_fd, &index, &fd, BPF_ANY);
>> +    }
>> +
>> +References
>> +==========
>> +
>> +- https://lore.kernel.org/netdev/[e-mail]/
>> +- https://lore.kernel.org/netdev/[e-mail]/

Donald Hunter Nov. 8, 2022, 10 a.m. UTC | #3

Yonghong Song <[e-mail]> writes:

> On 11/7/22 9:46 AM, Yonghong Song wrote:
>> On 11/7/22 5:48 AM, Donald Hunter wrote:
>>> +
>>> +The outer map supports element update and delete from user space using the
>>> +syscall API. A BPF program is only allowed to do element lookup in the outer
>>> +map.
>> The outer map supports element lookup, update and delete from user space using the syscall
>> API.
>> A BPF program can do element delete for array/hash_of_maps. Please double check.
>
> Okay, I double checked with verifier.c. You are right, only lookup
> is supported for bpf programs.

Thanks for checking. I do refer to verifier.c to see what helpers are
supported. I will add lookup for userspace.

>>> +
>>> +See ``progs/test_bpf_map_in_map.c`` in ``tools/testing/selftests/bpf`` for more
>> The file name test_bpf_map_in_map.c` does not exist.

Good catch, that's an unfortunate typo. It should be test_btf_map_in_map.c

>>> +examples of declarative initialisation of outer maps.
>>> +
>>> +User Space
>>> +----------
>>> +
>>> +This snippet shows how to create an array based outer map:
>>> +
>>> +.. code-block:: c
>>> +
>>> +    int create_outer_array(int inner_fd) {
>>> +            int fd;
>>> +
>>> +            LIBBPF_OPTS(bpf_map_create_opts, opts, .inner_map_fd = inner_fd);
>> This is declaration. Please put it adjacent to 'int fd'.

Will do. Looking at code in testing/selftests/bpf it seems the preferred style
is to put it above 'int fd;' ?

Patch

diff --git a/Documentation/bpf/map_of_maps.rst b/Documentation/bpf/map_of_maps.rst
new file mode 100644
index 000000000000..63e41b06a91d
--- /dev/null
+++ b/Documentation/bpf/map_of_maps.rst
@@ -0,0 +1,126 @@ 
+.. SPDX-License-Identifier: GPL-2.0-only
+.. Copyright (C) 2022 Red Hat, Inc.
+
+========================================================
+BPF_MAP_TYPE_ARRAY_OF_MAPS and BPF_MAP_TYPE_HASH_OF_MAPS
+========================================================
+
+.. note::
+   - ``BPF_MAP_TYPE_ARRAY_OF_MAPS`` and ``BPF_MAP_TYPE_HASH_OF_MAPS`` were
+     introduced in kernel version 4.12
+
+``BPF_MAP_TYPE_ARRAY_OF_MAPS`` and ``BPF_MAP_TYPE_HASH_OF_MAPS`` provide general
+purpose support for map in map storage. One level of nesting is supported, where
+an outer map contains instances of a single type of inner map, for example
+``array_of_maps->sock_map``.
+
+When creating an outer map, an inner map instance is used to initialize the
+metadata that the outer map holds about its inner maps. This inner map has a
+separate lifetime from the outer map and can be deleted after the outer map has
+been created.
+
+The outer map supports element update and delete from user space using the
+syscall API. A BPF program is only allowed to do element lookup in the outer
+map.
+
+.. note::
+   - Multi-level nesting is not supported.
+   - Any BPF map type can be used as an inner map, except for
+     ``BPF_MAP_TYPE_PROG_ARRAY``.
+   - A BPF program cannot update or delete outer map entries.
+
+For ``BPF_MAP_TYPE_ARRAY_OF_MAPS`` the key is an unsigned 32-bit integer index
+into the array. The array is a fixed size with ``max_entries`` elements that are
+zero initialized when created.
+
+For ``BPF_MAP_TYPE_HASH_OF_MAPS`` the key type can be chosen when defining the
+map. The kernel is responsible for allocating and freeing key/value pairs, up to
+the max_entries limit that you specify. Hash maps use pre-allocation of hash
+table elements by default. The ``BPF_F_NO_PREALLOC`` flag can be used to disable
+pre-allocation when it is too memory expensive.
+
+Usage
+=====
+
+Kernel BPF Helper
+-----------------
+
+.. c:function::
+   void *bpf_map_lookup_elem(struct bpf_map *map, const void *key)
+
+Inner maps can be retrieved using the ``bpf_map_lookup_elem()`` helper. This
+helper returns a pointer to the inner map, or ``NULL`` if no entry was found.
+
+Examples
+========
+
+Kernel BPF Example
+------------------
+
+This snippet shows how to create and initialise an array of devmaps in a BPF
+program. Note that the outer array can only be modified from user space using
+the syscall API.
+
+.. code-block:: c
+
+    struct inner_map {
+            __uint(type, BPF_MAP_TYPE_DEVMAP);
+            __uint(max_entries, 10);
+            __type(key, __u32);
+            __type(value, __u32);
+    } inner_map1 SEC(".maps"), inner_map2 SEC(".maps");
+
+    struct {
+            __uint(type, BPF_MAP_TYPE_ARRAY_OF_MAPS);
+            __uint(max_entries, 2);
+            __type(key, __u32);
+            __array(values, struct inner_map);
+    } outer_map SEC(".maps") = {
+            .values = { &inner_map1,
+                        &inner_map2 }
+    };
+
+See ``progs/test_bpf_map_in_map.c`` in ``tools/testing/selftests/bpf`` for more
+examples of declarative initialisation of outer maps.
+
+User Space
+----------
+
+This snippet shows how to create an array based outer map:
+
+.. code-block:: c
+
+    int create_outer_array(int inner_fd) {
+            int fd;
+
+            LIBBPF_OPTS(bpf_map_create_opts, opts, .inner_map_fd = inner_fd);
+            fd = bpf_map_create(BPF_MAP_TYPE_ARRAY_OF_MAPS,
+                                "example_array",       /* name */
+                                sizeof(__u32),         /* key size */
+                                sizeof(__u32),         /* value size */
+                                256,                   /* max entries */
+                                &opts);                /* create opts */
+            return fd;
+    }
+
+
+This snippet shows how to add an inner map to an outer map:
+
+.. code-block:: c
+
+    int add_devmap(int outer_fd, int index, const char *name) {
+            int fd;
+
+            fd = bpf_map_create(BPF_MAP_TYPE_DEVMAP, name,
+                                sizeof(__u32), sizeof(__u32), 256, NULL);
+            if (fd < 0)
+                    return fd;
+
+            return bpf_map_update_elem(outer_fd, &index, &fd, BPF_ANY);
+    }
+
+References
+==========
+
+- https://lore.kernel.org/netdev/[e-mail]/
+- https://lore.kernel.org/netdev/[e-mail]/