From: "Günther Noack" <gnoack3000@gmail.com>
To: Tingmao Wang <m@maowtm.org>
Cc: "Alejandro Colomar" <alx@kernel.org>,
"Mickaël Salaün" <mic@digikod.net>,
linux-man@vger.kernel.org
Subject: Re: [PATCH v2] landlock.7, landlock_*.2: Document LANDLOCK_ADD_RULE_QUIET
Date: Sat, 29 Aug 2026 19:24:45 +0200 [thread overview]
Message-ID: <20260829.acedcd1feb62@gnoack.org> (raw)
In-Reply-To: <20260829151304.101952-1-m@maowtm.org>
Thank you very much, Tingmao!
Documentation is the same as in the kernel docs, and renders fine. I
left a few smaller comments below on individual points. Apart from
these this looks good. :)
On Sat, Aug 29, 2026 at 04:13:03PM +0100, Tingmao Wang wrote:
> LANDLOCK_ADD_RULE_QUIET is a new feature introduced in Landlock ABI
> version 10, merged in kernel v7.2 [1]. This patch copies relevant
> kernel documentation into man-pages.
>
> Link: <https://patch.msgid.link/cover.1781228815.git.m@maowtm.org/> [1]
> Signed-off-by: Tingmao Wang <m@maowtm.org>
> ---
>
> Changes in v2:
> - Fix missing .RE, and fix EINVAL label being incorrectly formatted
> - Fix missed API bump in the example program (abi = MIN(abi, 10);)
>
> Hi,
>
> For context, I'm the author of the quiet flag feature and this is my
> first man-pages patch. @Günther or @Mickaël, can one of you do a quick review?
>
> All text in this patch is copied from the kernel source except this
> bit:
> .TP
> .B EINVAL
> .I flags
> is not 0 or one of the allowed values.
>
> (the kernel says "%EINVAL: @flags is not valid", I decided to make it
> more precise)
>
> man/man2/landlock_add_rule.2 | 48 ++++++++++++++++++++++++++++--
> man/man2/landlock_create_ruleset.2 | 48 ++++++++++++++++++++++++++++++
> man/man7/landlock.7 | 29 +++++++++++++++++-
> 3 files changed, 121 insertions(+), 4 deletions(-)
>
> diff --git a/man/man2/landlock_add_rule.2 b/man/man2/landlock_add_rule.2
> index fe01a98d9..c848c9b4b 100644
> --- a/man/man2/landlock_add_rule.2
> +++ b/man/man2/landlock_add_rule.2
> @@ -120,7 +120,44 @@ .SH DESCRIPTION
> and it will automatically translate to binding on the related port range.
> .P
> .I flags
> -must be 0.
> +can either be 0 or contain:
> +.TP
> +.BR LANDLOCK_ADD_RULE_QUIET " (since Landlock ABI version 10)"
> +Together with the
> +.I quiet_*
> +fields in
> +.IR "struct landlock_ruleset_attr" ,
> +this flag controls whether Landlock will log audit messages when
> +access to the objects covered by this rule is denied by this layer.
> +.IP
> +If logging is enabled, when Landlock denies an access,
> +it will suppress the log if all of the following are true:
> +.RS
> +.IP \[bu] 3
> +this layer is the innermost layer that denied the access;
> +.IP \[bu]
> +all accesses denied by this layer are part of the
> +.I quiet_*
> +fields in the related
> +.IR "struct landlock_ruleset_attr" ;
> +.IP \[bu]
> +the object (or one of its parents, for filesystem rules) is
> +marked as "quiet" via
> +.BR LANDLOCK_ADD_RULE_QUIET .
> +.RE
> +.IP
> +Because logging is only suppressed by a layer if the layer denies
> +access,
(I suspect Alejandro will bring it up as well; man pages use "semantic
line breaks" trying to break lines after logical parts of a sentence,
e.g.
Because logging is only suppressed by a layer
if the layer denies access,
etc.)
> +a sandboxed program cannot use this flag to "hide" access denials,
> +without denying itself the access in the first place.
> +.IP
> +The effect of this flag does not depend on the value of
> +.I allowed_access
> +in the passed in
> +.IR rule_attr .
> +When this flag is present, the caller is also allowed to pass in an
> +empty
(Also here, please use semantic line breaks so that "empty" does not
stand on its own line.)
> +.IR allowed_access .
> .SH RETURN VALUE
> On success,
> .BR landlock_add_rule ()
> @@ -159,7 +196,7 @@ .SH ERRORS
> .TP
> .B EINVAL
> .I flags
> -is not 0.
> +is not 0 or one of the allowed values.
> .TP
> .B EINVAL
> The rule accesses are inconsistent (i.e.,
> @@ -181,10 +218,15 @@ .SH ERRORS
> .IR \%struct\~landlock_net_port_attr ,
> the port number is greater than 65535.
> .TP
> +.B EINVAL
> +.B LANDLOCK_ADD_RULE_QUIET
> +is passed but the ruleset has no quiet access bits set for the
> +corresponding rule type.
(Maybe semantic line breaks as well)
> +.TP
> .B ENOMSG
> Empty accesses (i.e.,
> .I rule_attr\->allowed_access
> -is 0).
> +is 0) and no flags.
> .TP
> .B EOPNOTSUPP
> Landlock is supported by the kernel but disabled at boot time.
> diff --git a/man/man2/landlock_create_ruleset.2 b/man/man2/landlock_create_ruleset.2
> index 2a33fa4b5..8611e3aba 100644
> --- a/man/man2/landlock_create_ruleset.2
> +++ b/man/man2/landlock_create_ruleset.2
> @@ -45,6 +45,9 @@ .SH DESCRIPTION
> __u64 handled_access_fs;
> __u64 handled_access_net;
> __u64 scoped;
> + __u64 quiet_access_fs;
> + __u64 quiet_access_net;
> + __u64 quiet_scoped;
> };
> .EE
> .in
> @@ -70,6 +73,17 @@ .SH DESCRIPTION
> in
> .BR landlock (7)).
> .IP
> +.I quiet_access_fs
> +is a bitmask of filesystem actions which should not be logged if
> +per-object quiet flag is set.
Maybe add a "the" before "per-object" here? Unlike in the kernel
docs, this is a full sentence here, so a more complete sentence is
probably in order? (Same for quiet_access_net and quiet_scoped below
as well.)
> +.IP
> +.I quiet_access_net
> +is a bitmask of network actions which should not be logged if
> +per-object quiet flag is set.
> +.IP
> +.I quiet_scoped
> +is a bitmask of scoped actions which should not be logged.
> +.IP
> This structure defines a set of
> .IR "handled access rights" ,
> a set of actions on different object types,
> @@ -100,6 +114,29 @@ .SH DESCRIPTION
> a wide range or all access rights that they know about at build time
> (and that they have tested with a kernel that supported them all).
> .IP
> +.I quiet_access_fs
> +and
> +.I quiet_access_net
> +are bitmasks of actions for which a denial by this layer will not
> +trigger a log if the corresponding object (or its children, for
> +filesystem rules) is marked with the "quiet" bit via
(line breaks)
> +.BR LANDLOCK_ADD_RULE_QUIET ,
> +even if logging would normally take place per
> +.BR landlock_restrict_self (2)
> +flags.
> +.I quiet_scoped
> +is similar, except that it does not require marking any objects as quiet
> +-
Should maybe be a \[em]?
> +if the ruleset is created with any bits set in
> +.IR quiet_scoped ,
> +then denial
> +of such scoped resources will not trigger any log.
> +These 3 fields are available since Landlock ABI version 10
> +(see
> +.B Quiet rule flag
> +in
> +.BR landlock (7)).
Compared to the header file, you dropped the sentence
@quiet_access_fs, @quiet_access_net and @quiet_scoped must be a
subset of @handled_access_fs, @handled_access_net and @scoped
respectively.
This looks unintentional?
> +.IP
> This structure can grow in future Landlock versions.
> .P
> .I size
> @@ -204,6 +241,17 @@ .SH ERRORS
> or
> .BR LANDLOCK_CREATE_RULESET_ERRATA .
> .TP
> +.B EINVAL
> +.IR quiet_access_fs ,
> +.IR quiet_access_net ,
> +or
> +.I quiet_scoped
> +is not a subset of the corresponding
> +.IR handled_access_fs ,
> +.IR handled_access_net ,
> +or
> +.IR scoped .
> +.TP
> .B ENOMSG
> Empty accesses (i.e.,
> .I attr
> diff --git a/man/man7/landlock.7 b/man/man7/landlock.7
> index 880dd5058..402e35cbe 100644
> --- a/man/man7/landlock.7
> +++ b/man/man7/landlock.7
> @@ -456,6 +456,31 @@ .SS Truncating files
> It is also possible to pass such file descriptors between processes,
> keeping their Landlock properties,
> even when these processes do not have an enforced Landlock ruleset.
> +.SS Quiet rule flag
> +Starting with the Landlock ABI version 10,
> +it is possible to selectively suppress logs for specific denied
> +accesses on a per-object basis with the
> +.B LANDLOCK_ADD_RULE_QUIET
> +flag of
> +.BR landlock_add_rule (2),
> +in combination with the
> +.B quiet_access_fs
> +and
> +.B quiet_access_net
> +fields
> +of
> +.IR "struct landlock_ruleset_attr" .
> +It is also now possible to suppress
> +logs for scope accesses via the
> +.B quiet_scoped
> +field of
> +.IR "struct landlock_ruleset_attr" .
> +The object is marked as quiet within a ruleset when at least one
> +.BR landlock_add_rule (2)
> +call is made for it with the
> +.B LANDLOCK_ADD_RULE_QUIET
> +flag, additional add-rule calls for the same object without this flag
> +do not clear it.
> .SH VERSIONS
> Landlock was introduced in Linux 5.13.
> .P
> @@ -500,6 +525,8 @@ .SH VERSIONS
> 8 7.0 LANDLOCK_RESTRICT_SELF_TSYNC
> _ _ _
> 9 7.1 LANDLOCK_ACCESS_FS_RESOLVE_UNIX
> +_ _ _
> +10 7.2 LANDLOCK_ADD_RULE_QUIET
> .TE
> .P
> Users should use the Landlock ABI version rather than the kernel version
> @@ -622,7 +649,7 @@ .SH EXAMPLES
> perror("Unable to use Landlock");
> return; /* Graceful fallback: Do nothing. */
> }
> -abi = MIN(abi, 9);
> +abi = MIN(abi, 10);
Please also add an entry to the array in the example (even though it
is the same as the entry before, in this case). Otherwise, the
example has an out-of-bounds array access when this is 10.
> \&
> /* Only use the available rights in the ruleset. */
> attr.handled_access_fs &= landlock_fs_access_rights[abi \- 1];
> --
> 2.55.0
>
Thanks,
–Günther
next prev parent reply other threads:[~2026-08-29 17:24 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-08-29 15:13 [PATCH v2] landlock.7, landlock_*.2: Document LANDLOCK_ADD_RULE_QUIET Tingmao Wang
2026-08-29 17:24 ` Günther Noack [this message]
2026-08-29 21:16 ` Alejandro Colomar
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20260829.acedcd1feb62@gnoack.org \
--to=gnoack3000@gmail.com \
--cc=alx@kernel.org \
--cc=linux-man@vger.kernel.org \
--cc=m@maowtm.org \
--cc=mic@digikod.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox