adapt new mace 0.0.14

.gitignore

@@ -3,9 +3,12 @@
 /scripts/include
 /scripts/lib
 /scripts/lib64
-/scripts/shares/
+/scripts/share/
 /scripts/.gitignore
 *.png
 *.csv
 pyvenv.cfg
 Cargo.lock
+free.txt
+alloc.txt
+*.zst

.vscode/launch.json

@@ -0,0 +1,58 @@
+{
+        // Use IntelliSense to learn about possible attributes.
+        // Hover to view descriptions of existing attributes.
+        // For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387
+        "version": "0.2.0",
+        "configurations": [
+                {
+                        "type": "lldb",
+                        "request": "launch",
+                        "name": "Debug executable 'kv_bench'",
+                        "cargo": {
+                                "args": [
+                                        "build",
+                                        "--bin=kv_bench",
+                                        "--package=kv_bench"
+                                ],
+                                "filter": {
+                                        "name": "kv_bench",
+                                        "kind": "bin"
+                                }
+                        },
+                        "args": [
+                                "--path",
+                                "/home/abby/mace_bench",
+                                "--threads",
+                                "1",
+                                "--iterations",
+                                "100000",
+                                "--mode",
+                                "insert",
+                                "--key-size",
+                                "1024",
+                                "--value-size",
+                                "16"
+                        ],
+                        "cwd": "${workspaceFolder}"
+                },
+                {
+                        "type": "lldb",
+                        "request": "launch",
+                        "name": "Debug unit tests in executable 'kv_bench'",
+                        "cargo": {
+                                "args": [
+                                        "test",
+                                        "--no-run",
+                                        "--bin=kv_bench",
+                                        "--package=kv_bench"
+                                ],
+                                "filter": {
+                                        "name": "kv_bench",
+                                        "kind": "bin"
+                                }
+                        },
+                        "args": [],
+                        "cwd": "${workspaceFolder}"
+                }
+        ]
+}

Cargo.toml

@@ -5,11 +5,16 @@ edition = "2024"
 
 [dependencies]
 mace = { git = "https://github.com/abbycin/mace" }
-clap = { version = "4.5.42", features = ["derive"] }
+clap = { version = "4.5.48", features = ["derive"] }
 rand = "0.9.2"
 log = "0.4.22"
-coreid = { path = "coreid"}
+coreid = { path = "coreid" }
-logger = { path = "logger"}
+logger = { path = "logger" }
+myalloc = { path = "heap_trace" }
+
+[features]
+default = []
+custom_alloc = []
 
 [profile.release]
 lto = true

heap_trace/.gitignore

@@ -0,0 +1,2 @@
+/target
+Cargo.lock

heap_trace/Cargo.toml

@@ -0,0 +1,7 @@
+[package]
+name = "myalloc"
+version = "0.1.0"
+edition = "2024"
+
+[dependencies]
+backtrace = "0.3.76"

heap_trace/src/lib.rs

@@ -0,0 +1,233 @@
+use std::{
+    alloc::{GlobalAlloc, System},
+    cell::Cell,
+    collections::{HashMap, hash_map::Entry},
+    fmt::Display,
+    hash::{DefaultHasher, Hash, Hasher},
+    ptr,
+    sync::{LazyLock, Mutex, atomic::AtomicBool},
+};
+
+pub struct MyAlloc;
+
+fn trace(size: usize, is_alloc: bool) -> Option<String> {
+    let mut key = String::new();
+    backtrace::trace(|f| {
+        backtrace::resolve_frame(f, |sym| {
+            if let Some(filename) = sym.filename()
+                && let Some(line) = sym.lineno()
+            {
+                if let Some(name) = filename.to_str()
+                    && name.contains("mace")
+                {
+                    if name.len() > 10 {
+                        // sometime name maybe empty
+                        let x = format!("{}:{}\n", name, line);
+                        key.extend(x.chars().into_iter());
+                    }
+                }
+            }
+        });
+        true
+    });
+    if !key.is_empty() {
+        let mut lk = G_MAP.lock().unwrap();
+        let tmp = key.clone();
+        match lk.entry(tmp) {
+            Entry::Vacant(v) => {
+                if is_alloc {
+                    v.insert(Status {
+                        nr_alloc: 1,
+                        alloc_size: size,
+                        nr_free: 0,
+                        free_size: 0,
+                    });
+                } else {
+                    v.insert(Status {
+                        nr_alloc: 0,
+                        alloc_size: 0,
+                        nr_free: 1,
+                        free_size: size,
+                    });
+                }
+            }
+            Entry::Occupied(mut o) => {
+                let s = o.get_mut();
+                if is_alloc {
+                    s.nr_alloc += 1;
+                    s.alloc_size += size;
+                } else {
+                    s.nr_free += 1;
+                    s.free_size += size;
+                }
+            }
+        }
+        Some(key)
+    } else {
+        None
+    }
+}
+
+#[derive(Debug)]
+pub struct Status {
+    nr_alloc: usize,
+    alloc_size: usize,
+    nr_free: usize,
+    free_size: usize,
+}
+
+impl Display for Status {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        f.write_fmt(format_args!("{:?}", self))
+    }
+}
+
+static G_STOP: AtomicBool = AtomicBool::new(false);
+
+static G_MAP: LazyLock<Mutex<HashMap<String, Status>>> =
+    LazyLock::new(|| Mutex::new(HashMap::new()));
+
+static G_TRACE: LazyLock<Mutex<HashMap<u64, String>>> =
+    LazyLock::new(|| Mutex::new(HashMap::new()));
+
+const META_LEN: usize = 8;
+
+thread_local! {
+    static G_SELF: Cell<bool> = const { Cell::new(false) };
+}
+
+const fn real_size(layout: &std::alloc::Layout) -> usize {
+    if META_LEN > layout.align() {
+        META_LEN.checked_add(layout.size()).unwrap()
+    } else {
+        layout.align().checked_add(layout.size()).unwrap()
+    }
+}
+
+fn new_layout(layout: std::alloc::Layout) -> std::alloc::Layout {
+    let align = layout.align().max(align_of::<u64>());
+    let sz = real_size(&layout);
+    std::alloc::Layout::from_size_align(sz, align).unwrap()
+}
+
+fn write_hash(x: *mut u8, align: usize, s: Option<String>) -> *mut u8 {
+    let r = unsafe { x.add(META_LEN.max(align)) };
+    if !G_SELF.with(|x| x.get()) {
+        G_SELF.with(|x| x.set(true));
+        let p = x.cast::<u64>();
+        if let Some(s) = s {
+            let mut stat = DefaultHasher::new();
+            s.hash(&mut stat);
+            let h = stat.finish();
+            unsafe { p.write_unaligned(h) };
+            let mut lk = G_TRACE.lock().unwrap();
+            lk.insert(h, s);
+        } else {
+            unsafe { p.write_unaligned(u64::MAX) };
+        }
+        G_SELF.with(|x| x.set(false));
+    }
+    r
+}
+
+fn read_hash(x: *mut u8, align: usize) -> *mut u8 {
+    let (h, p) = unsafe {
+        let p = x.sub(META_LEN.max(align)).cast::<u64>();
+        (p.read_unaligned(), p.cast::<u8>())
+    };
+
+    if h == u64::MAX {
+        return p;
+    }
+    if !G_SELF.with(|x| x.get()) {
+        G_SELF.with(|x| x.set(true));
+        let mut lk = G_TRACE.lock().unwrap();
+        lk.remove(&h);
+        G_SELF.with(|x| x.set(false));
+    }
+    p
+}
+
+unsafe impl GlobalAlloc for MyAlloc {
+    unsafe fn alloc(&self, layout: std::alloc::Layout) -> *mut u8 {
+        let s = if !G_SELF.with(|x| x.get()) && !G_STOP.load(std::sync::atomic::Ordering::Acquire) {
+            G_SELF.with(|x| x.set(true));
+            let x = trace(layout.size(), true);
+            G_SELF.with(|x| x.set(false));
+            x
+        } else {
+            None
+        };
+
+        let new = new_layout(layout);
+        let x = unsafe { System.alloc(new) };
+        write_hash(x, new.align(), s)
+    }
+
+    unsafe fn dealloc(&self, ptr: *mut u8, layout: std::alloc::Layout) {
+        if !G_SELF.with(|x| x.get()) && !G_STOP.load(std::sync::atomic::Ordering::Acquire) {
+            G_SELF.with(|x| x.set(true));
+            trace(layout.size(), false);
+            G_SELF.with(|x| x.set(false));
+        }
+        let new = new_layout(layout);
+        let p = read_hash(ptr, new.align());
+        unsafe { System.dealloc(p, new) };
+    }
+
+    unsafe fn alloc_zeroed(&self, layout: std::alloc::Layout) -> *mut u8 {
+        let p = unsafe { self.alloc(layout) };
+        if !p.is_null() {
+            unsafe { ptr::write_bytes(p, 0, layout.size()) };
+        }
+        p
+    }
+
+    unsafe fn realloc(&self, ptr: *mut u8, layout: std::alloc::Layout, new_size: usize) -> *mut u8 {
+        let s = if !G_SELF.with(|x| x.get()) && !G_STOP.load(std::sync::atomic::Ordering::Acquire) {
+            G_SELF.with(|x| x.set(true));
+            let x = trace(layout.size(), true);
+            G_SELF.with(|x| x.set(false));
+            x
+        } else {
+            None
+        };
+
+        unsafe {
+            let old_layout = new_layout(layout);
+            let raw = ptr.sub(META_LEN.max(old_layout.align()));
+            let new_total_size = META_LEN + new_size;
+
+            let new_raw = System.realloc(raw, old_layout, new_total_size);
+            if new_raw.is_null() {
+                return new_raw;
+            }
+            write_hash(new_raw, old_layout.align(), s)
+        }
+    }
+}
+
+pub fn print_filtered_trace<F>(f: F)
+where
+    F: Fn(&str, &Status),
+{
+    G_STOP.store(true, std::sync::atomic::Ordering::Release);
+    let lk = G_MAP.lock().unwrap();
+    let t = G_TRACE.lock().unwrap();
+
+    for (_, v) in t.iter() {
+        if let Some(s) = lk.get(v) {
+            f(v, s);
+        }
+    }
+}
+
+pub fn print_all_trace<F>(f: F)
+where
+    F: Fn(&str, &Status),
+{
+    G_STOP.store(true, std::sync::atomic::Ordering::Release);
+    let lk = G_MAP.lock().unwrap();
+
+    lk.iter().for_each(|(k, v)| f(k, v));
+}

scripts/init.sh

@@ -2,3 +2,4 @@
 
 python3 -m venv .
 ./bin/pip3 install pandas matplotlib adjustText
+rm -f .gitignore

scripts/mem_analyze.py

@@ -0,0 +1,53 @@
+#!/usr/bin/python3
+
+import sys
+
+assert(len(sys.argv) == 2)
+
+f = open(sys.argv[1])
+lines  = []
+
+allocs = []
+
+while True:
+        line = f.readline()
+        if len(line) < 10:
+                break
+        if line.find('INFO') != -1:
+                continue
+        pos = line.find('Status')
+        if pos < 0:
+                pos = line.find('mace')
+                if pos != 0:
+                        lines.append(line[pos:])
+                else:
+                        lines.append(line)
+        else:
+                cleaned = line[pos+6:].strip().strip('{}')
+                pairs = cleaned.split(',')
+                tl = [tuple(pair.split(': ')) for pair in pairs]
+                tl = [(k.strip(), int(v)) for k, v in tl]
+                allocs.append((tl, ''.join(lines)))
+                lines.clear()
+
+# sort by alloc_size
+allocs.sort(key=lambda x: x[0][1][1], reverse=True)
+
+with open('alloc.txt', 'w') as o:
+        for x in allocs:
+                o.write(f'{x[0]}\n{x[1]}\n')
+# sort by free_size
+allocs.sort(key=lambda x: x[0][3][1], reverse=True)
+
+with open('free.txt', 'w') as o:
+        for x in allocs:
+                o.write(f'{x[0]}\n{x[1]}\n')
+
+alloc_size = 0
+free_size = 0
+
+for x in allocs:
+        alloc_size += x[0][1][1]
+        free_size += x[0][3][1]
+
+print(f"total_alloc {alloc_size} total_free {free_size}")

src/main.rs

@@ -2,6 +2,8 @@ use clap::Parser;
 #[cfg(target_os = "linux")]
 use logger::Logger;
 use mace::{Mace, Options};
+#[cfg(feature = "custom_alloc")]
+use myalloc::{MyAlloc, print_filtered_trace};
 use rand::prelude::*;
 use std::path::Path;
 use std::process::exit;
@@ -9,6 +11,10 @@ use std::sync::Arc;
 use std::thread::JoinHandle;
 use std::time::Instant;
 
+#[cfg(feature = "custom_alloc")]
+#[global_allocator]
+static GLOBAL: MyAlloc = MyAlloc;
+
 #[derive(Parser, Debug)]
 #[command(author, version, about, long_about = None)]
 struct Args {
@@ -40,7 +46,7 @@ struct Args {
 fn main() {
     #[cfg(target_os = "linux")]
     {
-        Logger::init().add_file("/tmp/x.log", true);
+        Logger::init().add_file("/Data/x.log", true);
         log::set_max_level(log::LevelFilter::Info);
     }
     let args = Args::parse();
@@ -71,11 +77,11 @@ fn main() {
     let mut opt = Options::new(path);
     opt.sync_on_write = false;
     opt.tmp_store = args.mode != "get";
-    opt.gc_timeout = 1000 * 60; // make sure GC will not work
     let mut saved = opt.clone();
     saved.tmp_store = false;
     // opt.cache_capacity = 3 << 30; // this is very important for large key-value store
     let mut db = Mace::new(opt.validate().unwrap()).unwrap();
+    db.disable_gc();
 
     let mut rng = rand::rng();
     let value = Arc::new(vec![b'0'; args.value_size]);
@@ -100,8 +106,6 @@ fn main() {
             }
         });
         pre_tx.commit().unwrap();
-        drop(pre_tx);
-        log::info!("=====");
         drop(db);
         // re-open db
         saved.tmp_store = true;
@@ -206,4 +210,7 @@ fn main() {
         ops,
         duration.as_millis()
     );
+    drop(db);
+    #[cfg(feature = "custom_alloc")]
+    print_filtered_trace(|x, y| log::info!("{}{}", x, y));
 }