perf(runtime): mark Tokio threads as mimalloc threadpool (#6646)

Upgrade rustfs-mimalloc and rustfs-mimalloc-sys to 0.5.1, then call the new safe wrapper from Tokio worker thread startup so mimalloc can treat runtime threads as threadpool workers.

Keep the hint no-op on Windows, matching RustFS allocator platform boundaries.

Co-authored-by: heihutu <heihutu@gmail.com>

Cargo.lock

@@ -9980,18 +9980,18 @@ dependencies = [
 
 [[package]]
 name = "rustfs-mimalloc"
-version = "0.5.0"
+version = "0.5.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a406f4aa07084301d485beec873af6dccc8e3f8762da244743df92038b1db1a6"
+checksum = "ed388b8a3d55818c32973ded41df5215d33ee0f574d9ca03f2731b3007c3284b"
 dependencies = [
  "rustfs-mimalloc-sys",
 ]
 
 [[package]]
 name = "rustfs-mimalloc-sys"
-version = "0.5.0"
+version = "0.5.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c3051b819175f58445d4c369a72f0ab88149f3885ba8bea2aff3be01f53fe7cd"
+checksum = "1c507265df958f6b63f8ccdd21819374186bab5dbdcc3f9c5cf8bff4b2e3f5d0"
 dependencies = [
  "cc",
 ]

Cargo.toml

@@ -365,8 +365,8 @@ russh-sftp = "2.4.0"
 dav-server = "0.11.0"
 
 # Performance Analysis and Memory Profiling
-rustfs-mimalloc = { version = "0.5.0" }
+rustfs-mimalloc = { version = "0.5.1" }
-rustfs-mimalloc-sys = { version = "0.5.0" }
+rustfs-mimalloc-sys = { version = "0.5.1" }
 hotpath = { version = "0.24.0", default-features = false }
 # Snapshot testing for output format regression detection
 insta = { version = "1.48" }

rustfs/src/server/runtime.rs

@@ -30,6 +30,12 @@ fn compute_default_thread_stack_size() -> usize {
     }
 }
 
+#[inline]
+fn mark_allocator_threadpool_thread() {
+    #[cfg(not(target_os = "windows"))]
+    rustfs_mimalloc::set_current_thread_in_threadpool();
+}
+
 #[cfg(test)]
 mod tests {
     use super::*;
@@ -160,15 +166,19 @@ pub fn tokio_runtime_builder() -> tokio::runtime::Builder {
         rustfs_utils::get_env_usize(rustfs_config::ENV_MAX_IO_EVENTS_PER_TICK, rustfs_config::DEFAULT_MAX_IO_EVENTS_PER_TICK);
     builder.enable_all().max_io_events_per_tick(max_io_events_per_tick);
 
-    // Optional: Simple log of thread start/stop
+    let print_tokio_threads = print_tokio_thread_enable();
-    if print_tokio_thread_enable() {
+    builder.on_thread_start(move || {
-        builder
+        mark_allocator_threadpool_thread();
-            .on_thread_start(|| {
+        if print_tokio_threads {
-                tracing::trace!(thread_id = ?std::thread::current().id(), "worker thread started");
+            tracing::trace!(thread_id = ?std::thread::current().id(), "worker thread started");
-            })
+        }
-            .on_thread_stop(|| {
+    });
-                tracing::trace!(thread_id = ?std::thread::current().id(), "worker thread stopped");
+
-            });
+    // Optional: Simple log of thread stop
+    if print_tokio_threads {
+        builder.on_thread_stop(|| {
+            tracing::trace!(thread_id = ?std::thread::current().id(), "worker thread stopped");
+        });
     }
     if !rustfs_obs::is_production_environment() {
         println!(