Skip to content

LLM inference in C/C++

C++ ggml

metal : null-check buffer alloc to fix OOM crash (#25371)

Yuri Khrustalev3737e41370da
diff --git a/ggml/src/ggml-metal/ggml-metal.cpp b/ggml/src/ggml-metal/ggml-metal.cpp
index e5b2ee8a55..9756d47050 100644
--- a/ggml/src/ggml-metal/ggml-metal.cpp
+++ b/ggml/src/ggml-metal/ggml-metal.cpp
@@ -204,6 +204,11 @@ static ggml_backend_buffer_t ggml_backend_metal_buffer_type_alloc_buffer(ggml_ba
ggml_metal_device_t ctx_dev = (ggml_metal_device_t)buft->device->context;
ggml_metal_buffer_t res = ggml_metal_buffer_init(ctx_dev, size, shared);
+ if (res == NULL) {
+ GGML_LOG_ERROR("%s: failed to allocate Metal buffer of %zu bytes (out of memory)\n", __func__, size);
+ return NULL;
+ }
+
ggml_backend_buffer_i buf_i = ggml_metal_buffer_is_shared(res)
? ggml_backend_metal_buffer_shared_i
: ggml_backend_metal_buffer_private_i;