From 3cdc8fa128e382136a2398c99c97de96d806c146 Mon Sep 17 00:00:00 2001
From: Roman Gershman <roman@dragonflydb.io>
Date: Tue, 10 Sep 2024 10:26:44 +0300
Subject: [PATCH] chore: add a script that parses allocator tracking logs
 (#3687)

---
 src/core/allocation_tracker.cc         |  3 +-
 tools/parse_allocator_tracking_logs.py | 45 ++++++++++++++++++++++++++
 2 files changed, 47 insertions(+), 1 deletion(-)
 create mode 100755 tools/parse_allocator_tracking_logs.py

diff --git a/src/core/allocation_tracker.cc b/src/core/allocation_tracker.cc
index 7b65c3b22..42b64cecc 100644
--- a/src/core/allocation_tracker.cc
+++ b/src/core/allocation_tracker.cc
@@ -85,7 +85,8 @@ void AllocationTracker::ProcessDelete(void* ptr) {
   if (tracking_.size() == 1 && tracking_.front().sample_odds == 1) {
     size_t usable = mi_usable_size(ptr);
     if (usable <= tracking_.front().upper_bound && usable >= tracking_.front().lower_bound) {
-      LOG(INFO) << "Deallocating " << usable << " bytes (" << ptr << ")";
+      LOG(INFO) << "Deallocating " << usable << " bytes (" << ptr << ")\n"
+                << util::fb2::GetStacktrace();
     }
   }
   inside_tracker_ = false;
diff --git a/tools/parse_allocator_tracking_logs.py b/tools/parse_allocator_tracking_logs.py
new file mode 100755
index 000000000..f10525134
--- /dev/null
+++ b/tools/parse_allocator_tracking_logs.py
@@ -0,0 +1,45 @@
+#!/usr/bin/env python3
+
+"""
+Usage:
+1. First run Dragonfly with tracking allocator enabled. Must be a single allocator range with 100% samping rate to catch both allocations and deallocations.
+2. Finish tracking.
+3. cat /tmp/dragonfly.INFO |  ./parse_allocator_tracking_logs.py
+"""
+import re
+import sys
+
+
+def parse_log(log_lines):
+    memory_map = {}
+
+    allocation_pattern = re.compile(r"Allocating (\d+) bytes \((0x[0-9a-f]+)\)")
+    deallocation_pattern = re.compile(r"Deallocating (\d+) bytes \((0x[0-9a-f]+)\)")
+
+    for line in log_lines:
+        allocation_match = allocation_pattern.search(line)
+        deallocation_match = deallocation_pattern.search(line)
+
+        if allocation_match:
+            size = int(allocation_match.group(1))
+            address = allocation_match.group(2)
+            assert address not in memory_map
+            memory_map[address] = (size, line)
+        elif deallocation_match:
+            size = int(deallocation_match.group(1))
+            address = deallocation_match.group(2)
+            if address in memory_map:
+                assert size == memory_map[address][0]
+                del memory_map[address]
+            else:
+                print(f"Deallocating non existing address: {address} {size}")
+
+    return memory_map
+
+
+if __name__ == "__main__":
+    log_lines = sys.stdin.readlines()
+    memory_map = parse_log(log_lines)
+
+    for address, item in memory_map.items():
+        print(f"Address: {address}, Size: {item[0]} bytes, original line: `{item[1]}`")