I guess my patch for musl in 0.5.0 was rejected

[delandtj]

or was that an error and I should reintroduce it ?

[delandtj]

tat was the Patch

diff --git a/crates/polyfuse/src/conn.rs b/crates/polyfuse/src/conn.rs
index d409bfc..808d317 100644
--- a/crates/polyfuse/src/conn.rs
+++ b/crates/polyfuse/src/conn.rs
@@ -1,6 +1,7 @@
 use libc::{c_int, c_void, iovec};
 use std::{
     cmp,
+    convert::TryInto,
     ffi::{OsStr, OsString},
     io,
     mem::{self, MaybeUninit},
@@ -313,19 +314,25 @@ fn receive_fd(reader: &UnixStream) -> io::Result<RawFd> {
     }
     let mut cmsg = MaybeUninit::<Cmsg>::uninit();

-    let mut msg = libc::msghdr {
-        msg_name: ptr::null_mut(),
-        msg_namelen: 0,
-        msg_iov: &mut iov,
-        msg_iovlen: 1,
-        msg_control: cmsg.as_mut_ptr() as *mut c_void,
-        msg_controllen: mem::size_of_val(&cmsg),
-        msg_flags: 0,
-    };
+    // On musl, libc::msghdr has private padding fields; construct via zero-init and set public fields.
+    let mut msg: libc::msghdr = unsafe { mem::zeroed() };
+    msg.msg_name = ptr::null_mut();
+    msg.msg_namelen = 0;
+    msg.msg_iov = &mut iov;
+    msg.msg_iovlen = 1;
+    msg.msg_control = cmsg.as_mut_ptr() as *mut c_void;
+    msg.msg_controllen = mem::size_of::<Cmsg>()
+        .try_into()
+        .expect("invalid control message length");
+    msg.msg_flags = 0;

     syscall! { recvmsg(reader.as_raw_fd(), &mut msg, 0) };

-    if msg.msg_controllen < mem::size_of_val(&cmsg) {
+    if msg.msg_controllen
+        < mem::size_of_val(&cmsg)
+            .try_into()
+            .expect("invalid control message length")
+    {
         return Err(io::Error::new(
             io::ErrorKind::InvalidData,
             "too short control message length",

[delandtj]

as per rust-lang/libc#2344 where musl libc will not expose private padding fields.
tested on both glibc and musl libc

[ubnt-intrepid]

@delandtj I apologize for the lack of clarity in my explanation.

What I meant was that I'd like you to "renew" your PR to the branch for the published version.

Could you please resubmit the same pull request for 0.5.x branch?