[PATCH v7 18/18] landlock: Document Landlock's network support

[Mickaël Salaün]

On 29/08/2022 19:04, Konstantin Meskhidze wrote:
> Describe network access rules for TCP sockets.
> Add network access example in the tutorial.
> Point out AF_UNSPEC socket family behaviour.
> Point out UDP sockets issues.
> Add kernel configuration support for network.
> 
> Signed-off-by: Konstantin Meskhidze <konstantin.meskhidze at huawei.com>
> ---
> 
> Changes since v6:
> * Adds network support documentaion.
> 
> ---
>   Documentation/userspace-api/landlock.rst | 84 +++++++++++++++++++-----
>   1 file changed, 66 insertions(+), 18 deletions(-)
> 
> diff --git a/Documentation/userspace-api/landlock.rst b/Documentation/userspace-api/landlock.rst
> index 2509c2fbf98f..4b099d1b5a9d 100644
> --- a/Documentation/userspace-api/landlock.rst
> +++ b/Documentation/userspace-api/landlock.rst
> @@ -11,10 +11,10 @@ Landlock: unprivileged access control
>   :Date: August 2022
> 
>   The goal of Landlock is to enable to restrict ambient rights (e.g. global
> -filesystem access) for a set of processes.  Because Landlock is a stackable
> -LSM, it makes possible to create safe security sandboxes as new security layers
> -in addition to the existing system-wide access-controls. This kind of sandbox
> -is expected to help mitigate the security impact of bugs or
> +filesystem or network access) for a set of processes.  Because Landlock
> +is a stackable LSM, it makes possible to create safe security sandboxes as new
> +security layers in addition to the existing system-wide access-controls. This
> +kind of sandbox is expected to help mitigate the security impact of bugs or
>   unexpected/malicious behaviors in user space applications.  Landlock empowers
>   any process, including unprivileged ones, to securely restrict themselves.
> 
> @@ -30,18 +30,20 @@ Landlock rules
> 
>   A Landlock rule describes an action on an object.  An object is currently a
>   file hierarchy, and the related filesystem actions are defined with `access
> -rights`_.  A set of rules is aggregated in a ruleset, which can then restrict
> -the thread enforcing it, and its future children.
> +rights`_.  Since ABI version 3 a port "object" appears with related network actions

Version 3 is wrong here, it should be version 4. Please fix everywhere 
else too.

Why do you quote "object"?

Let's use object (e.g. for filesystem/inode) or data (e.g. for TCP port).


> +for TCP4/TCP6 sockets families.  A set of rules is aggregated in a ruleset, which

TCP4/TCP6 would not make sense for users, please be explicit, and in 
this case "TCP" is enough.

BTW, IP4 and IP6 don't exist, they should be replaced *everywhere* with 
IPv4 and IPv6.


> +can then restrict the thread enforcing it, and its future children.
> 
>   Defining and enforcing a security policy
>   ----------------------------------------
> 
>   We first need to define the ruleset that will contain our rules.  For this
>   example, the ruleset will contain rules that only allow read actions, but write
> -actions will be denied.  The ruleset then needs to handle both of these kind of
> +actions will be denied. The ruleset then needs to handle both of these kind of
>   actions.  This is required for backward and forward compatibility (i.e. the
>   kernel and user space may not know each other's supported restrictions), hence
> -the need to be explicit about the denied-by-default access rights.
> +the need to be explicit about the denied-by-default access rights.  Also ruleset
> +will have network rules for specific ports, so it should handle network actions.
> 
>   .. code-block:: c
> 
> @@ -62,6 +64,9 @@ the need to be explicit about the denied-by-default access rights.
>               LANDLOCK_ACCESS_FS_MAKE_SYM |
>               LANDLOCK_ACCESS_FS_REFER |
>               LANDLOCK_ACCESS_FS_TRUNCATE,
> +        .handled_access_net =
> +            LANDLOCK_ACCESS_NET_BIND_TCP |
> +            LANDLOCK_ACCESS_NET_CONNECT_TCP,
>       };
> 
>   Because we may not know on which kernel version an application will be
> @@ -70,9 +75,9 @@ should try to protect users as much as possible whatever the kernel they are
>   using.  To avoid binary enforcement (i.e. either all security features or
>   none), we can leverage a dedicated Landlock command to get the current version
>   of the Landlock ABI and adapt the handled accesses.  Let's check if we should
> -remove the `LANDLOCK_ACCESS_FS_REFER` or `LANDLOCK_ACCESS_FS_TRUNCATE` access
> -rights, which are only supported starting with the second and third version of
> -the ABI.
> +remove the `LANDLOCK_ACCESS_FS_REFER` or `LANDLOCK_ACCESS_FS_TRUNCATE` or
> +network access rights, which are only supported starting with the second and
> +third version of the ABI.
> 
>   .. code-block:: c
> 
> @@ -87,9 +92,13 @@ the ABI.
>               /* Removes LANDLOCK_ACCESS_FS_REFER for ABI < 2 */
>               ruleset_attr.handled_access_fs &= ~LANDLOCK_ACCESS_FS_REFER;
>               __attribute__((fallthrough));
> +            /* Removes network support for ABI < 2 */
> +            ruleset_attr.handled_access_net = 0;

This is not correct.


>       case 2:
>               /* Removes LANDLOCK_ACCESS_FS_TRUNCATE for ABI < 3 */
>               ruleset_attr.handled_access_fs &= ~LANDLOCK_ACCESS_FS_TRUNCATE;
> +            /* Removes network support for ABI < 3 */
> +            ruleset_attr.handled_access_net = 0;

This part should be for the fourth version.


>       }
> 
>   This enables to create an inclusive ruleset that will contain our rules.
> @@ -129,6 +138,24 @@ descriptor.
>       }
>       err = landlock_add_rule(ruleset_fd, LANDLOCK_RULE_PATH_BENEATH,
>                               &path_beneath, 0);
> +
> +It may also be required to create rules following the same logic as explained
> +for the ruleset creation, by filtering access rights according to the Landlock
> +ABI version.  In this example, this is not required because all of the requested
> +`allowed_access` rights are already available in ABI 1.

This paragraph should not be moved. Furthermore, this hunk remove error 
handling…


> +
> +For network part we can add number of rules containing a port number and actions
> +that a process is allowed to do for certian ports.
> +
> +.. code-block:: c
> +
> +    struct landlock_net_service_attr net_service = {
> +        .allowed_access = LANDLOCK_ACCESS_NET_BIND_TCP,
> +        .port = 8080,
> +    };
> +
> +    err = landlock_add_rule(ruleset_fd, LANDLOCK_RULE_NET_SERVICE,
> +                            &net_service, 0);
>       close(path_beneath.parent_fd);
>       if (err) {
>           perror("Failed to update ruleset");
> @@ -136,13 +163,9 @@ descriptor.
>           return 1;
>       }
> 
> -It may also be required to create rules following the same logic as explained
> -for the ruleset creation, by filtering access rights according to the Landlock
> -ABI version.  In this example, this is not required because all of the requested
> -`allowed_access` rights are already available in ABI 1.
> -

Please add similar standalone code + explanation sections for network here.


>   We now have a ruleset with one rule allowing read access to ``/usr`` while
> -denying all other handled accesses for the filesystem.  The next step is to
> +denying all other handled accesses for the filesystem.  The ruleset also contains
> +a rule allowing to bind current proccess to the port 8080.  The next step is to
>   restrict the current thread from gaining more privileges (e.g. thanks to a SUID
>   binary).
> 
> @@ -280,6 +303,13 @@ It should also be noted that truncating files does not require the
>   system call, this can also be done through :manpage:`open(2)` with the flags
>   `O_RDONLY | O_TRUNC`.
> 
> +AF_UNSPEC socket family
> +-----------------------
> +
> +Sockets of AF_UNSPEC family types are treated as AF_INET(TCP4) socket for bind()
> +hook.  But connect() hook is not allowed by Landlock for AF_UNSPEC sockets. This

Users don't know what is a hook. Such kernel internals are not required 
to explain things.


> +logic prevents from disconnecting already connected sockets.
> +
>   Compatibility
>   =============
> 
> @@ -339,7 +369,7 @@ Access rights
>   -------------
> 
>   .. kernel-doc:: include/uapi/linux/landlock.h
> -    :identifiers: fs_access
> +    :identifiers: fs_access net_access
> 
>   Creating a new ruleset
>   ----------------------
> @@ -358,6 +388,7 @@ Extending a ruleset
> 
>   .. kernel-doc:: include/uapi/linux/landlock.h
>       :identifiers: landlock_rule_type landlock_path_beneath_attr
> +                  landlock_net_service_attr
> 
>   Enforcing a ruleset
>   -------------------
> @@ -406,6 +437,13 @@ Memory usage
>   Kernel memory allocated to create rulesets is accounted and can be restricted
>   by the Documentation/admin-guide/cgroup-v1/memory.rst.
> 
> +UDP sockets restricting
> +-----------------------

I don't think this section is needed. There is a lot of missing access 
types for now and it is not useful to list them all. The "Current 
limitations" section lists limitations about the currently implemented 
access types (e.g. filesystem, TCP).


> +
> +Current network part supports to restrict just TCP sockets type. UPD sockets sandboxing

typo: UDP

> +adds additional issues due to unconnected nature of the protocol. UDP sockets support
> +might come in future Landlock versions.
> +
>   Previous limitations
>   ====================
> 
> @@ -435,6 +473,13 @@ always allowed when using a kernel that only supports the first or second ABI.
>   Starting with the Landlock ABI version 3, it is now possible to securely control
>   truncation thanks to the new `LANDLOCK_ACCESS_FS_TRUNCATE` access right.
> 
> +Network support (ABI < 3)

ABI < 4

> +-------------------------
> +
> +Starting with the Landlock ABI version 3, it is now possible to restrict TCP
> +sockets bind() and connect() syscalls for specific ports allowing processes
> +to establish secure connections.

Try to avoid talking about syscall directly but highlight actions 
instead. Using Landlock doesn't create "secure connections", 
unfortunately. ;)


> +
>   .. _kernel_support:
> 
>   Kernel support
> @@ -453,6 +498,9 @@ still enable it by adding ``lsm=landlock,[...]`` to
>   Documentation/admin-guide/kernel-parameters.rst thanks to the bootloader
>   configuration.
> 
> +To support Landlock's network part, the kernel must be configured with `CONFIG_NET=y`
> +and `CONFIG_INET=y` options.

Might be better:
To be able to explicitly allow TCP operations (e.g., adding a network 
rule with `LANDLOCK_ACCESS_NET_TCP_BIND`), the kernel must support TCP 
(`CONFIG_INET=y`). Otherwise, sys_landlock_add_rule() returns an 
`EAFNOSUPPORT` error, which can safely be ignored because this kind of 
TCP operation is already not possible.


> For TCP6 family sockets `CONFIG_IPV6=y` must be switched on.

This is not required because if IPv6 is not supported there is nothing 
to do for IPv6 sockets.


> +
>   Questions and answers
>   =====================
> 
> --
> 2.25.1
>