From fd29ac373f6ff66416b9cf3680afd99c12298f0f Mon Sep 17 00:00:00 2001
From: Tyler Michael Smith <tyler@neuralmagic.com>
Date: Thu, 26 Sep 2024 19:18:26 -0400
Subject: [PATCH] [Bugfix] Fix print_warning_once's line info (#8867)

---
 vllm/utils.py | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/vllm/utils.py b/vllm/utils.py
index b73e3b9bbf68e..a0d2a7e50fc63 100644
--- a/vllm/utils.py
+++ b/vllm/utils.py
@@ -744,7 +744,8 @@ def create_kv_caches_with_random(
 
 @lru_cache
 def print_warning_once(msg: str) -> None:
-    logger.warning(msg)
+    # Set the stacklevel to 2 to print the caller's line info
+    logger.warning(msg, stacklevel=2)
 
 
 @lru_cache(maxsize=None)