removing freebsd 11 EOL proposal. #3139

devnexen · 2023-03-05T10:21:25Z

No description provided.

rustbot · 2023-03-05T10:21:30Z

(rustbot has picked a reviewer for you, use r? to override)

JohnTitor · 2023-03-05T14:30:47Z

Note that there're some overrides with #2406.

bors · 2023-03-08T09:56:53Z

☔ The latest upstream changes (presumably #3037) made this pull request unmergeable. Please resolve the merge conflicts.

bors · 2023-05-06T11:14:24Z

☔ The latest upstream changes (presumably #2999) made this pull request unmergeable. Please resolve the merge conflicts.

workingjubilee · 2023-07-26T20:38:36Z

FCP: Drop support for FreeBSD 10 and 11 from std rust#89058
Rust toolchain update: Update FreeBSD toolchain to 12.3 rust#97944

@devnexen Please resolve the merge conflict.

asomers · 2023-07-26T21:34:52Z

build.rs

-        Some(13) if libc_ci => set_cfg("freebsd13"),
-        Some(14) if libc_ci => set_cfg("freebsd14"),
-        Some(_) | None => set_cfg("freebsd11"),
+        Some(12) if libc_ci => println!("cargo:rustc-cfg=freebsd12"),


This looks like a merge conflict. You should use set_cfg instead of println!

asomers · 2023-07-26T21:38:12Z

src/unix/bsd/freebsdlike/freebsd/freebsd11/mod.rs

-// APIs that were changed after FreeBSD 11
-
-// The type of `nlink_t` changed from `u16` to `u64` in FreeBSD 12:
-pub type nlink_t = u16;


All of the stuff that was in this file should now be moved into freebsdlike/freebsd/mod.rs For example, move all three nlink_t definitions from freebsd{12,13,14}.rs into that file.

workingjubilee · 2023-07-28T01:24:08Z

I started a conversation on Zulip to see how people feel about this.

tgross35 · 2025-08-10T05:33:40Z

This is basically 100% conflicted, but @devnexen would you be up for rebasing this?

EOL was 5 years ago so I can't imagine anyone would object at this point.

tgross35 · 2025-08-11T00:59:32Z

build.rs

    };

    match which_freebsd {
        x if x < 10 => panic!("FreeBSD older than 10 is not supported"),


Could you update this comment?

tgross35 · 2025-08-11T01:00:56Z

I haven't looked through this in detail yet but the changes look correct. @asomers would you be able to review?

asomers · 2025-08-13T16:32:47Z

I haven't looked through this in detail yet but the changes look correct. @asomers would you be able to review?

In theory I agree with this. I'll try to review it next Saturday. We'll want to run a few dozen crates' test suites with this change, too.

asomers

These are the only two problems I've found in my testing so far. But that begs the question: how did you test these changes?

asomers · 2025-08-17T16:02:42Z

src/unix/bsd/freebsdlike/mod.rs

-        all(target_os = "freebsd", any(freebsd11, freebsd10)),
-        link_name = "kevent@FBSD_1.0"
-    )]
-    pub fn kevent(


Did you really mean to remove this function? I think you should just remove the link_name instead. The function is definitely needed.

asomers · 2025-08-17T16:30:26Z

src/unix/bsd/freebsdlike/freebsd/mod.rs


+pub type nlink_t = u64;
+pub type dev_t = u64;
+pub type ino_t = c_ulong;


This should be u64. Otherwise stat is broken on x86.

Suggested change

pub type ino_t = c_ulong;

pub type ino_t = u64;

asomers · 2025-09-07T20:12:41Z

Your latest changes fix the two bugs that I found. What about you? How did you test these changes?

devnexen · 2025-09-07T20:14:59Z

I test within a VM the vast majority of time, sometimes when I can t I look at freebsd sources and count on CI to see. Not ideal I know.

asomers · 2025-09-07T20:28:10Z

But what did you do within that VM? What projects did you test? How did you patch them to use libc main branch?

devnexen · 2025-09-07T20:32:01Z

Usually, I just do cargo test --manifest-path libc-test/Cargo.toml. I did not try to use in any project.

asomers · 2025-09-07T22:54:28Z

Usually, I just do cargo test --manifest-path libc-test/Cargo.toml. I did not try to use in any project.

For a change like this, that really isn't sufficient. You need to run the test suites of a selection of other crates, patched to use this branch of libc.

devnexen · 2025-09-07T23:13:59Z

Usually, I just do cargo test --manifest-path libc-test/Cargo.toml. I did not try to use in any project.

For a change like this, that really isn't sufficient. You need to run the test suites of a selection of other crates, patched to use this branch of libc.

Ok I ll try

devnexen · 2025-09-08T21:51:45Z

@asomers, here a test with nix crate

[dcarlier@ ~/Contribs/nix]$ cargo test --all-features
    Updating git repository `https://github.com/devnexen/libc`
     Locking 1 package to latest compatible version
      Adding libc v1.0.0-alpha.1 (https://github.com/devnexen/libc?branch=remove_fbsd11_support#3f8c5074)
   Compiling libc v0.2.172
   Compiling nix v0.30.1 (/home/dcarlier/Contribs/nix)
   Compiling getrandom v0.3.3
   Compiling errno v0.3.13
   Compiling parking_lot_core v0.9.11
   Compiling rustix v1.0.8
   Compiling rand_core v0.9.3
   Compiling rand_chacha v0.9.0
   Compiling parking_lot v0.12.4
   Compiling rand v0.9.2
   Compiling sysctl v0.4.6
   Compiling tempfile v3.21.0
error[E0308]: mismatched types
  --> src/sys/select.rs:25:53
   |
25 |         usize::try_from(fd).map_or(false, |fd| fd < FD_SETSIZE),
   |                                                --   ^^^^^^^^^^ expected `usize`, found `i32`
   |                                                |
   |                                                expected because this is `usize`
   |
help: you can convert an `i32` to a `usize` and panic if the converted value doesn't fit
   |
25 |         usize::try_from(fd).map_or(false, |fd| fd < FD_SETSIZE.try_into().unwrap()),
   |                                                               ++++++++++++++++++++

error[E0308]: mismatched types
    --> src/sys/select.rs:113:65
     |
 113 |             range: 0..highest.map(|h| h as usize + 1).unwrap_or(FD_SETSIZE),
     |                                                       --------- ^^^^^^^^^^ expected `usize`, found `i32`
     |                                                       |
     |                                                       arguments to this method are incorrect
     |
help: the return type of this call is `i32` due to the type of the argument passed
    --> src/sys/select.rs:113:23
     |
 113 |             range: 0..highest.map(|h| h as usize + 1).unwrap_or(FD_SETSIZE),
     |                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^----------^
     |                                                                 |
     |                                                                 this argument influences the return type of `unwrap_or`
note: method defined here
    --> /home/dcarlier/.rustup/toolchains/nightly-x86_64-unknown-freebsd/lib/rustlib/src/rust/library/core/src/option.rs:1031:18
     |
1031 |     pub const fn unwrap_or(self, default: T) -> T
     |                  ^^^^^^^^^
help: you can convert an `i32` to a `usize` and panic if the converted value doesn't fit
     |
 113 |             range: 0..highest.map(|h| h as usize + 1).unwrap_or(FD_SETSIZE.try_into().unwrap()),
     |                                                                           ++++++++++++++++++++

error[E0308]: mismatched types
    --> src/unistd.rs:1100:41
     |
1100 |     unsafe { libc::execv(path.as_ptr(), args_p.as_ptr()) };
     |              -----------                ^^^^^^^^^^^^^^^ types differ in mutability
     |              |
     |              arguments to this function are incorrect
     |
     = note: expected raw pointer `*const *mut i8`
                found raw pointer `*const *const i8`
note: function defined here
    --> /home/dcarlier/.cargo/git/checkouts/libc-42f01ba267c5d37b/3f8c507/src/unix/mod.rs:1033:12
     |
1033 |     pub fn execv(prog: *const c_char, argv: *const *mut c_char) -> c_int;
     |            ^^^^^

error[E0308]: arguments to this function are incorrect
    --> src/unistd.rs:1126:14
     |
1126 |     unsafe { libc::execve(path.as_ptr(), args_p.as_ptr(), env_p.as_ptr()) };
     |              ^^^^^^^^^^^^
     |
note: types differ in mutability
    --> src/unistd.rs:1126:42
     |
1126 |     unsafe { libc::execve(path.as_ptr(), args_p.as_ptr(), env_p.as_ptr()) };
     |                                          ^^^^^^^^^^^^^^^
     = note: expected raw pointer `*const *mut i8`
                found raw pointer `*const *const i8`
note: types differ in mutability
    --> src/unistd.rs:1126:59
     |
1126 |     unsafe { libc::execve(path.as_ptr(), args_p.as_ptr(), env_p.as_ptr()) };
     |                                                           ^^^^^^^^^^^^^^
     = note: expected raw pointer `*const *mut i8`
                found raw pointer `*const *const i8`
note: function defined here
    --> /home/dcarlier/.cargo/git/checkouts/libc-42f01ba267c5d37b/3f8c507/src/unix/mod.rs:1034:12
     |
1034 |     pub fn execve(prog: *const c_char, argv: *const *mut c_char, envp: *const *mut c_char)
     |            ^^^^^^

error[E0308]: mismatched types
    --> src/unistd.rs:1147:46
     |
1147 |     unsafe { libc::execvp(filename.as_ptr(), args_p.as_ptr()) };
     |              ------------                    ^^^^^^^^^^^^^^^ types differ in mutability
     |              |
     |              arguments to this function are incorrect
     |
     = note: expected raw pointer `*const *mut i8`
                found raw pointer `*const *const i8`
note: function defined here
    --> /home/dcarlier/.cargo/git/checkouts/libc-42f01ba267c5d37b/3f8c507/src/unix/mod.rs:1036:12
     |
1036 |     pub fn execvp(c: *const c_char, argv: *const *mut c_char) -> c_int;
     |            ^^^^^^

error[E0308]: arguments to this function are incorrect
    --> src/unistd.rs:1197:14
     |
1197 |     unsafe { libc::fexecve(fd.as_fd().as_raw_fd(), args_p.as_ptr(), env_p.as_ptr()) };
     |              ^^^^^^^^^^^^^
     |
note: types differ in mutability
    --> src/unistd.rs:1197:52
     |
1197 |     unsafe { libc::fexecve(fd.as_fd().as_raw_fd(), args_p.as_ptr(), env_p.as_ptr()) };
     |                                                    ^^^^^^^^^^^^^^^
     = note: expected raw pointer `*const *mut i8`
                found raw pointer `*const *const i8`
note: types differ in mutability
    --> src/unistd.rs:1197:69
     |
1197 |     unsafe { libc::fexecve(fd.as_fd().as_raw_fd(), args_p.as_ptr(), env_p.as_ptr()) };
     |                                                                     ^^^^^^^^^^^^^^
     = note: expected raw pointer `*const *mut i8`
                found raw pointer `*const *const i8`
note: function defined here
    --> /home/dcarlier/.cargo/git/checkouts/libc-42f01ba267c5d37b/3f8c507/src/unix/bsd/freebsdlike/mod.rs:1530:12
     |
1530 |     pub fn fexecve(fd: c_int, argv: *const *mut c_char, envp: *const *mut c_char) -> c_int;
     |            ^^^^^^^

error[E0308]: mismatched types
   --> src/macros.rs:72:35
    |
 72 |                       const $Flag = libc::$Flag $(as $cast)*;
    |                                     ^^^^^^^^^^^ expected `i32`, found `i16`
    |
   ::: src/spawn.rs:203:1
    |
203 | / libc_bitflags!(
204 | |     /// Process attributes to be changed in the new process image when invoking [`posix_spawn`]
205 | |     /// or [`posix_spawnp`]. See
206 | |     /// [posix_spawn](https://pubs.opengroup.org/onlinepubs/9699919799/functions/posix_spawn.html).
...   |
227 | | );
    | | -
    | | |
    | |_arguments to this function are incorrect
    |   in this macro invocation
    |
note: associated function defined here
   --> src/spawn.rs:203:1
    |
203 | / libc_bitflags!(
204 | |     /// Process attributes to be changed in the new process image when invoking [`posix_spawn`]
205 | |     /// or [`posix_spawnp`]. See
206 | |     /// [posix_spawn](https://pubs.opengroup.org/onlinepubs/9699919799/functions/posix_spawn.html).
...   |
227 | | );
    | |_^
    = note: this error originates in the macro `libc_bitflags` (in Nightly builds, run with -Z macro-backtrace for more info)

For more information about this error, try `rustc --explain E0308`.
error: could not compile `nix` (lib) due to 10 previous errors

tgross35 · 2025-10-29T20:24:32Z

I'm a bit confused: where did the breakage in the above example come from here? The default FreeBSD version didn't change, so I'm not sure where the errors are coming from. AIUI this just deletes the FreeBSD 11 module then commonizes 12+.

devnexen · 2025-10-29T20:27:53Z

I do not think it s related to this PR in particular but might have more to do with the breaking changes of the future libc major version.

tgross35 · 2025-10-29T20:35:16Z

Ah, I forgot about that bit. @asomers what were your concerns with breakage? AIUI there should be no delta for anybody not using freebsd11.

If you just want to verify there isn't any accidental breakage then it's possible to run cargo rustdoc -- --output-format json -Zunstable-options before and after this patch (setting RUST_LIBC_UNSTABLE_FREEBSD_VERSION as needed) and saving off the file, then run cargo semver-checks --baseline-rustdoc ... --current-rustdoc ....

asomers

@tgross35 the breakage shown in the Nix crate came from switching libc-0.2 to libc-1.0. But it's important for the submitter to check that nix doesn't break any worse on libc-1.0 with this PR than without it.
And I agree that there shouldn't be any breakage as a result of removing support for the FreeBSD 10 and 11 ABIs from libc. However, there may be as a result of mistakes in the PR. The fact that the submitter didn't catch the accidental removal of kevent or changing the type of ino_t suggests that he didn't test it very closely.

asomers · 2025-10-29T21:02:56Z

src/unix/bsd/mod.rs

-    )]
    pub fn rand() -> c_int;
-    #[cfg_attr(
-        all(target_os = "freebsd", any(freebsd12, freebsd11, freebsd10)),


Why are you removing this link name? FreeBSD 12 still needs it.

asomers · 2025-10-29T21:03:48Z

src/unix/bsd/mod.rs

    pub fn abs(i: c_int) -> c_int;
    pub fn labs(i: c_long) -> c_long;
-    #[cfg_attr(
-        all(target_os = "freebsd", any(freebsd12, freebsd11, freebsd10)),


Here, too. Why remove the link name?

asomers · 2025-10-29T21:06:13Z

src/unix/bsd/mod.rs

    )]
-    #[cfg_attr(
-        all(target_os = "freebsd", any(freebsd12, freebsd11, freebsd10)),
-        link_name = "wait4@FBSD_1.0"


Like the others, I think that removing this link_name was unintended. However, it is coincidentally the right thing to do. It never should've been there in the first place. wait4 hasn't ever been modified since FreeBSD first adopted ELF symbol versions.

rustbot · 2025-10-29T22:04:19Z

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.

rustbot assigned JohnTitor Mar 5, 2023

rustbot added the S-waiting-on-review label Mar 5, 2023

devnexen marked this pull request as ready for review March 5, 2023 10:34

devnexen force-pushed the remove_fbsd11_support branch from e30b5ac to f01623a Compare March 8, 2023 12:11

JohnTitor added breakage-candidate S-blocked and removed S-waiting-on-review labels Mar 12, 2023

devnexen force-pushed the remove_fbsd11_support branch from f01623a to 9e7d63b Compare July 26, 2023 20:42

asomers suggested changes Jul 26, 2023

View reviewed changes

devnexen force-pushed the remove_fbsd11_support branch 3 times, most recently from ce5b55f to 72de88e Compare July 26, 2023 22:46

workingjubilee mentioned this pull request Jul 28, 2023

Raise libc's FreeBSD ABI to 12 #2406

Merged

devnexen marked this pull request as draft August 10, 2025 06:07

devnexen force-pushed the remove_fbsd11_support branch from 72de88e to e3d1192 Compare August 10, 2025 06:07

rustbot added O-bsd O-unix labels Aug 10, 2025

devnexen force-pushed the remove_fbsd11_support branch 2 times, most recently from 3fe771e to 662e1b4 Compare August 10, 2025 06:24

rustbot added the O-x86 label Aug 10, 2025

devnexen force-pushed the remove_fbsd11_support branch 3 times, most recently from 1c0de03 to 5eb24f0 Compare August 10, 2025 07:02

tgross35 removed the S-blocked label Aug 10, 2025

tgross35 reviewed Aug 11, 2025

View reviewed changes

devnexen force-pushed the remove_fbsd11_support branch from 5eb24f0 to 2da0438 Compare August 11, 2025 04:44

asomers suggested changes Aug 17, 2025

View reviewed changes

devnexen force-pushed the remove_fbsd11_support branch 3 times, most recently from 2c962b8 to cb008d7 Compare August 17, 2025 17:10

tgross35 unassigned JohnTitor Aug 27, 2025

devnexen requested a review from asomers August 30, 2025 06:33

devnexen force-pushed the remove_fbsd11_support branch from cb008d7 to 3f8c507 Compare September 7, 2025 20:34

removing freebsd 11 EOL proposal.

d7e87b8

asomers suggested changes Oct 29, 2025

View reviewed changes

devnexen force-pushed the remove_fbsd11_support branch from 3f8c507 to 72a72a7 Compare October 29, 2025 22:04

devnexen force-pushed the remove_fbsd11_support branch 4 times, most recently from 952d8d8 to d7e87b8 Compare October 29, 2025 22:54

removing freebsd 11 EOL proposal. #3139

Are you sure you want to change the base?

removing freebsd 11 EOL proposal. #3139

Uh oh!

Conversation

devnexen commented Mar 5, 2023

Uh oh!

rustbot commented Mar 5, 2023

Uh oh!

JohnTitor commented Mar 5, 2023

Uh oh!

bors commented Mar 8, 2023

Uh oh!

bors commented May 6, 2023

Uh oh!

workingjubilee commented Jul 26, 2023

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

workingjubilee commented Jul 28, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

tgross35 commented Aug 10, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

tgross35 commented Aug 11, 2025

Uh oh!

asomers commented Aug 13, 2025

Uh oh!

asomers left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

asomers commented Sep 7, 2025

Uh oh!

devnexen commented Sep 7, 2025

Uh oh!

asomers commented Sep 7, 2025

Uh oh!

devnexen commented Sep 7, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

asomers commented Sep 7, 2025

Uh oh!

devnexen commented Sep 7, 2025

Uh oh!

devnexen commented Sep 8, 2025

Uh oh!

tgross35 commented Oct 29, 2025

Uh oh!

devnexen commented Oct 29, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

tgross35 commented Oct 29, 2025

Uh oh!

asomers left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rustbot commented Oct 29, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants

workingjubilee commented Jul 28, 2023 •

edited

Loading

devnexen commented Sep 7, 2025 •

edited

Loading

devnexen commented Oct 29, 2025 •

edited

Loading