From d04cfaf2f5844455c8e773edf0d7f02919d3249c Mon Sep 17 00:00:00 2001
From: Francis Couture-Harpin <git@compilade.net>
Date: Mon, 18 Mar 2024 21:26:08 -0400
Subject: [PATCH] llama : fix llama_output_reserve nullptr deref when new_size
 is 0

---
 llama.cpp | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/llama.cpp b/llama.cpp
index 075b1937a..faf65e339 100644
--- a/llama.cpp
+++ b/llama.cpp
@@ -9195,7 +9195,7 @@ static void llama_output_reserve(llama_context & lctx, int32_t n_outputs) {
 
     // alloc only when more than the current capacity is required
     // TODO: also consider shrinking the buffer
-    if (prev_size < new_size) {
+    if (!lctx.buf_output || prev_size < new_size) {
         if (lctx.buf_output) {
 #ifndef NDEBUG
             // This doesn't happen often, but may be annoying in some cases (like the HellaSwag benchmark)