chore: Fix cippy errors

aya-log-parser/src/lib.rs

@@ -1,3 +1,6 @@
+// We implement our own formatter here and we pass literal strings on purpose.
+#![allow(clippy::literal_string_with_formatting_args)]
+
 use std::str;
 
 use aya_log_common::DisplayHint;

aya-obj/src/btf/relocation.rs

@@ -1068,7 +1068,7 @@ impl ComputedRelocation {
                     }
                     BtfType::Enum64(en) => {
                         let variant = &en.variants[accessor.index];
-                        (variant.value_high as u64) << 32 | variant.value_low as u64
+                        ((variant.value_high as u64) << 32) | variant.value_low as u64
                     }
                     // candidate selection ensures that rel_kind == local_kind == target_kind
                     _ => unreachable!(),

aya/src/maps/lpm_trie.rs

@@ -63,7 +63,7 @@ pub struct LpmTrie<T, K, V> {
 /// let ipaddr = Ipv4Addr::new(8,8,8,8);
 /// let key =  Key::new(16, u32::from(ipaddr).to_be());
 /// ```
-#[repr(packed)]
+#[repr(C, packed)]
 pub struct Key<K: Pod> {
     prefix_len: u32,
     data: K,

aya/src/sys/mod.rs

@@ -93,6 +93,10 @@ fn syscall(call: Syscall<'_>) -> SysResult<i64> {
     #[cfg(test)]
     return TEST_SYSCALL.with(|test_impl| unsafe { test_impl.borrow()(call) });
 
+    // The type of integer taken by `ioctl` is different in glibc (i64) and
+    // musl (i32). musl builds would complain about useless conversion.
+    // `libc::ioctl` returns i32 on x86_64 while `libc::syscall` returns i64.
+    #[allow(clippy::useless_conversion)]
     #[cfg_attr(test, allow(unreachable_code))]
     {
         let ret = unsafe {
@@ -109,8 +113,6 @@ fn syscall(call: Syscall<'_>) -> SysResult<i64> {
                 } => libc::syscall(SYS_perf_event_open, &attr, pid, cpu, group, flags),
                 Syscall::PerfEventIoctl { fd, request, arg } => {
                     let ret = libc::ioctl(fd.as_raw_fd(), request.try_into().unwrap(), arg);
-                    // `libc::ioctl` returns i32 on x86_64 while `libc::syscall` returns i64.
-                    #[allow(clippy::useless_conversion)]
                     ret.into()
                 }
             }

ebpf/aya-ebpf/src/maps/lpm_trie.rs

@@ -18,7 +18,7 @@ pub struct LpmTrie<K, V> {
 
 unsafe impl<K: Sync, V: Sync> Sync for LpmTrie<K, V> {}
 
-#[repr(packed)]
+#[repr(C, packed)]
 pub struct Key<K> {
     /// Represents the number of bits matched against.
     pub prefix_len: u32,

ebpf/aya-ebpf/src/maps/perf/perf_event_array.rs

@@ -51,7 +51,7 @@ impl<T> PerfEventArray<T> {
     }
 
     pub fn output_at_index<C: EbpfContext>(&self, ctx: &C, index: u32, data: &T, flags: u32) {
-        let flags = u64::from(flags) << 32 | u64::from(index);
+        let flags = (u64::from(flags) << 32) | u64::from(index);
         unsafe {
             bpf_perf_event_output(
                 ctx.as_ptr(),

ebpf/aya-ebpf/src/maps/perf/perf_event_byte_array.rs

@@ -48,7 +48,7 @@ impl PerfEventByteArray {
     }
 
     pub fn output_at_index<C: EbpfContext>(&self, ctx: &C, index: u32, data: &[u8], flags: u32) {
-        let flags = u64::from(flags) << 32 | u64::from(index);
+        let flags = (u64::from(flags) << 32) | u64::from(index);
         unsafe {
             bpf_perf_event_output(
                 ctx.as_ptr(),

xtask/public-api/aya-ebpf.txt

@@ -263,7 +263,7 @@ pub fn aya_ebpf::maps::hash_map::PerCpuHashMap<K, V>::borrow_mut(&mut self) -> &
 impl<T> core::convert::From<T> for aya_ebpf::maps::hash_map::PerCpuHashMap<K, V>
 pub fn aya_ebpf::maps::hash_map::PerCpuHashMap<K, V>::from(t: T) -> T
 pub mod aya_ebpf::maps::lpm_trie
-#[repr(packed)] pub struct aya_ebpf::maps::lpm_trie::Key<K>
+#[repr(C, packed)] pub struct aya_ebpf::maps::lpm_trie::Key<K>
 pub aya_ebpf::maps::lpm_trie::Key::data: K
 pub aya_ebpf::maps::lpm_trie::Key::prefix_len: u32
 impl<K> aya_ebpf::maps::lpm_trie::Key<K>

xtask/public-api/aya.txt

@@ -264,7 +264,7 @@ pub fn aya::maps::hash_map::PerCpuHashMap<T, K, V>::borrow_mut(&mut self) -> &mu
 impl<T> core::convert::From<T> for aya::maps::hash_map::PerCpuHashMap<T, K, V>
 pub fn aya::maps::hash_map::PerCpuHashMap<T, K, V>::from(t: T) -> T
 pub mod aya::maps::lpm_trie
-#[repr(packed)] pub struct aya::maps::lpm_trie::Key<K: aya::Pod>
+#[repr(C, packed)] pub struct aya::maps::lpm_trie::Key<K: aya::Pod>
 impl<K: aya::Pod> aya::maps::lpm_trie::Key<K>
 pub fn aya::maps::lpm_trie::Key<K>::data(&self) -> K
 pub fn aya::maps::lpm_trie::Key<K>::new(prefix_len: u32, data: K) -> Self