freebsd(13.2): add `netlink/netlink.h` support by dybucc · Pull Request #5326 · rust-lang/libc

dybucc · 2026-07-24T14:42:05Z

Description

This PR updates #3201 with merge conflicts resolved and follows the new plan at
¹.

The patch adds support for netlink.h interfaces in OpenBSD, where there's an
item resolution conflict if we expose the Rust bindings alongside those of
if_mib.h. This set of APIs is "scoped" in C because they live on separate
headers. In rust-lang/libc, we reexport all items at the root crate level, which
makes item resolution fail.

Note this depends on #5325. It won't pass tests but it will build. This is
because the test templates will gather all items in a single file, so item
resolution fails. We can't really skip these items altogether from the tests, so
it may just be necessary to extend ctest to allow skipping module-specific
Rust items.

Checklist

Relevant tests in libc-test/semver have been updated
Commit messages permalink to headers for added or changed API
Placeholder or unstable values like *LAST or *MAX have the standard
doc comment
Tested locally (cargo test -p libc-test --target mytarget); especially
relevant for platforms that may not be checked in CI

@rustbot label +stable-nominated

https://github.com/rust-lang/libc/pull/3201#issuecomment-4736374182 ↩

tgross35 · 2026-07-31T08:12:35Z

The API looks fine from a quick skim, but since there is no hurry, I think it may be worth trying to add support to ctest first so the tricky test setup isn't needed. (It's useful otherwise too.)

Sketched some of that up at #5344

dybucc · 2026-07-31T08:54:38Z

Noted. Since you already pinged some contributor on that issue, I'll wait and
see whether they want to implement it. Otherwise, I'll do it.

tgross35 · 2026-08-04T05:16:21Z

Could you try adding a separate TestGenerator instance like @mbyx recommended in that issue? Since it's split off, that should avoid the need to reexport libc::netlink at test root too.

tgross35 · 2026-08-04T05:16:43Z

Either author or blocked, depending on whether that works.

@rustbot author

rustbot · 2026-08-04T05:16:47Z

Reminder, once the PR becomes ready for a review, use @rustbot ready.

tgross35 · 2026-08-10T03:25:23Z

+                !matches!(
+                    c.ident(),
+                    "CTRL_CMD_UNSPEC"
+                        | "CTRL_CMD_NEWFAMILY"
+                        | "CTRL_CMD_DELFAMILY"
+                        | "CTRL_CMD_GETFAMILY"
+                        | "CTRL_CMD_NEWOPS"
+                        | "CTRL_CMD_DELOPS"
+                        | "CTRL_CMD_GETOPS"
+                        | "CTRL_CMD_NEWMCAST_GRP"


These could be put into functions like is_netlink_const(c: &Const) -> bool so the list can be shared the two places it's needed.

View changes since the review

tgross35 · 2026-08-10T06:48:22Z

    let freebsd14 = matches!(freebsd_ver, Some(n) if n >= 14);
    let freebsd15 = matches!(freebsd_ver, Some(n) if n >= 15);

+    if let Some(net_header) = net_header {


I think you can drop the net_header arg, one invocation of test_freebsd can create all 2-3 ctest objects. I assume you're doing this to share the setup above, so I applied 023c0a3 to make that cfg a bit easier to share.

View changes since the review

Done. I didn't actually mean that. That was just cruft from prior attempts with
a different approach.

tgross35 · 2026-08-10T06:49:43Z

@@ -0,0 +1,6 @@
+//! Directory: `netlink/`
+//!
+//! <https://github.com/freebsd/freebsd-src/tree/df9d6403caa6426e92f5e100602f4d2be474bbae/sys/netlink>


Fyi, no need to permalink these kind of files since we do want to see updates, and the file locations are usually stable enough. As long as there's a permalink in the commit message to show us what it actually looked like at that point in time.

View changes since the review

You mean the files for the upstream directories only, or also the files
referring to the upstream header files?

dybucc · 2026-08-10T09:15:04Z

It seems having separate invocations gets the job done just fine. Still, that
was the same thing I tried in the initial patch. The added complexity comes when
we need to get the SemVer tests to pass.

Those seem to me like they're going to need extending (though it's notably less
work than with ctest as it's all in libc-test's build script.)

tgross35 · 2026-08-10T19:53:58Z

I haven't tried this but do things just work™️ if the semver tests have lines like netlink::NETLINK_ADD_MEMBERSHIP?

dybucc · 2026-08-11T07:09:57Z

I haven't tried this but do things just work™️ if the semver tests have lines
like netlink::NETLINK_ADD_MEMBERSHIP?

@tgross35 They do. Just cleaned up commit history as all tests seem to pass now.

tgross35 · 2026-08-11T08:22:23Z

-        pub(crate) use freebsd::*;
+        // FIXME(1.0,remove): glob reexport should be the default.
+        pub use freebsd::netlink;
+        pub(crate) use freebsd::{
+            net,
+            netinet6,
+            sys,
+            unistd,
+        };


Does this work?

pub(crate) use freebsd::*; pub use netlink;

View changes since the review

tgross35 · 2026-08-11T08:23:19Z

+pub const CTRL_CMD_UNSPEC: c_int = 0;
+pub const CTRL_CMD_NEWFAMILY: c_int = 1;
+pub const CTRL_CMD_DELFAMILY: c_int = 2;
+pub const CTRL_CMD_GETFAMILY: c_int = 3;
+pub const CTRL_CMD_NEWOPS: c_int = 4;
+pub const CTRL_CMD_DELOPS: c_int = 5;
+pub const CTRL_CMD_GETOPS: c_int = 6;
+pub const CTRL_CMD_NEWMCAST_GRP: c_int = 7;
+pub const CTRL_CMD_DELMCAST_GRP: c_int = 8;
+pub const CTRL_CMD_GETMCAST_GRP: c_int = 9;
+pub const CTRL_CMD_GETPOLICY: c_int = 10;
+
+pub const CTRL_ATTR_UNSPEC: c_int = 0;
+pub const CTRL_ATTR_FAMILY_ID: c_int = 1;
+pub const CTRL_ATTR_FAMILY_NAME: c_int = 2;
+pub const CTRL_ATTR_VERSION: c_int = 3;
+pub const CTRL_ATTR_HDRSIZE: c_int = 4;
+pub const CTRL_ATTR_MAXATTR: c_int = 5;
+pub const CTRL_ATTR_OPS: c_int = 6;
+pub const CTRL_ATTR_MCAST_GROUPS: c_int = 7;
+pub const CTRL_ATTR_POLICY: c_int = 8;
+pub const CTRL_ATTR_OP_POLICY: c_int = 9;
+pub const CTRL_ATTR_OP: c_int = 10;
+
+pub const CTRL_ATTR_MCAST_GRP_UNSPEC: c_int = 0;
+pub const CTRL_ATTR_MCAST_GRP_NAME: c_int = 1;
+pub const CTRL_ATTR_MCAST_GRP_ID: c_int = 2;


I think these aren't in the semver file

View changes since the review

tgross35 · 2026-08-11T08:25:17Z

+netlink::netlink::NETLINK_ADD_MEMBERSHIP
+netlink::netlink::NETLINK_AUDIT
+netlink::netlink::NETLINK_BROADCAST_ERROR
+netlink::netlink::NETLINK_CAP_ACK
+netlink::netlink::NETLINK_CONNECTOR
+netlink::netlink::NETLINK_DNRTMSG
+netlink::netlink::NETLINK_DROP_MEMBERSHIP
+netlink::netlink::NETLINK_EXT_ACK
+netlink::netlink::NETLINK_FIB_LOOKUP
+netlink::netlink::NETLINK_FIREWALL
+netlink::netlink::NETLINK_GENERIC
+netlink::netlink::NETLINK_GET_STRICT_CHK
+netlink::netlink::NETLINK_IP6_FW
+netlink::netlink::NETLINK_ISCSI
+netlink::netlink::NETLINK_KOBJECT_UEVENT
+netlink::netlink::NETLINK_LISTEN_ALL_NSID
+netlink::netlink::NETLINK_LIST_MEMBERSHIPS
+netlink::netlink::NETLINK_NETFILTER
+netlink::netlink::NETLINK_NFLOG
+netlink::netlink::NETLINK_NO_ENOBUFS
+netlink::netlink::NETLINK_PKTINFO
+netlink::netlink::NETLINK_ROUTE
+netlink::netlink::NETLINK_RX_RING
+netlink::netlink::NETLINK_SELINUX
+netlink::netlink::NETLINK_SOCK_DIAG
+netlink::netlink::NETLINK_TX_RING
+netlink::netlink::NETLINK_UNUSED
+netlink::netlink::NETLINK_USERSOCK
+netlink::netlink::NETLINK_XFRM
+netlink::netlink::NLMSG_ALIGNTO
+netlink::netlink::NLMSG_DONE
+netlink::netlink::NLMSG_ERROR
+netlink::netlink::NLMSG_NOOP
+netlink::netlink::NLMSG_OVERRUN
+netlink::netlink::NLM_F_ACK
+netlink::netlink::NLM_F_ACK_TLVS
+netlink::netlink::NLM_F_APPEND
+netlink::netlink::NLM_F_ATOMIC
+netlink::netlink::NLM_F_CAPPED
+netlink::netlink::NLM_F_CREATE
+netlink::netlink::NLM_F_DUMP
+netlink::netlink::NLM_F_DUMP_FILTERED
+netlink::netlink::NLM_F_DUMP_INTR
+netlink::netlink::NLM_F_ECHO
+netlink::netlink::NLM_F_EXCL
+netlink::netlink::NLM_F_MATCH
+netlink::netlink::NLM_F_MULTI
+netlink::netlink::NLM_F_NONREC
+netlink::netlink::NLM_F_REPLACE
+netlink::netlink::NLM_F_REQUEST
+netlink::netlink::NLM_F_ROOT
+netlink::netlink::NL_ITEM_ALIGN_SIZE
+netlink::netlink::SOL_NETLINK


It would be better if we can keep all this directly within a netlink module. In the exports, perhaps an in-file mod netlink { pub use ... }

View changes since the review

tgross35 · 2026-08-11T08:28:26Z

+    let mut netlink_cfg = cfg.clone();
+    headers!(netlink_cfg, "netlink/netlink.h",);
+    netlink_cfg
+        .skip_struct(|ty| !matches!(ty.ident(), "sockaddr_nl"))
+        .skip_const(|c| !is_netlink_const(c))
+        .skip_union(|_| true)
+        .skip_alias(|_| true)
+        .skip_static(|_| true)
+        .skip_fn(|_| true)
+        .skip_c_enum(|_| true);
+    ctest::generate_test(&mut netlink_cfg, "../src/lib.rs", "netlink_ctest_output.rs").unwrap();
+
+    // We must restore the above sure-skips because `TestGenerator` shares skips
+    // between cloned instances (here `cfg` and `netlink_cfg`.)
+    cfg.skip_struct(|_| false)
+        .skip_const(|_| false)
+        .skip_union(|_| false)
+        .skip_alias(|_| false)
+        .skip_static(|_| false)
+        .skip_fn(|_| false)
+        .skip_c_enum(|_| false);


Sharing skips isn't intentional, why does this happen?

View changes since the review

tgross35 · 2026-08-11T08:29:38Z

+            // This is tested in a separate ctest invocation because some symbols
+            // `netlink/netlink.h` conflict with some other symbols from `net/if_mib.h`.
+            "sockaddr_nl" => true,


Is this one needed still? I think we should be checking a sockaddr_nl in each ctest invocation

View changes since the review

rustbot · 2026-08-11T17:04:57Z

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

This is an early subset of the Netlink interface, but it proves sufficient for monitoring changes in IP addresses. Coverage can be extended later as needed. See [^1] and [^2]. [^1]: <https://github.com/freebsd/freebsd-src/blob/df9d6403caa6426e92f5e100602f4d2be474bbae/sys/netlink/netlink.h> [^2]: <https://github.com/freebsd/freebsd-src/blob/df9d6403caa6426e92f5e100602f4d2be474bbae/sys/netlink/netlink_generic.h> A small workaround has been necessary in the SemVer tests to ensure we get the right paths to the public submodules for the `netlink/netlink.h` interfaces. Those symbols now are also prepended a `netlink::netlink`. Signed-off-by: Yann Dirson <yann.dirson@vates.fr> Co-authored-by: Yann Dirson <yann.dirson@vates.fr>

Add specific test for `netlink/netlink.h` bindings. This is necessary to avoid conflicts with the bindings for `net/if_mib.h`. libc-test now builds two different `TestGenerator` instances. One of the instances builds tests for the same set of bindings as before this patchset, while the other builds tests only for the `netlink/netlink.h` bindings.

rustbot added O-bsd O-freebsd S-waiting-on-review stable-nominated This PR should be considered for cherry-pick to libc's stable release branch labels Jul 24, 2026

dybucc mentioned this pull request Jul 24, 2026

Netlink support for FreeBSD 13.2 #3201

Open

2 tasks

dybucc force-pushed the freebsd-netlink branch from f53fbf1 to c7886ff Compare July 27, 2026 06:28

rustbot added the A-CI Area: CI-related items label Jul 27, 2026