Bug fixes from mainline (#439)

* Add __syncthreads() to the new FA kernel

* Clearing padding

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>

1 files changed, 2 insertions, 2 deletions

diff --git a/ggml/src/ggml-backend.c b/ggml/src/ggml-backend.c
index 410ab9e5..e191c2d9 100644
--- a/ggml/src/ggml-backend.c
+++ b/ggml/src/ggml-backend.c
@@ -35,7 +35,7 @@ size_t ggml_backend_buft_get_max_size(ggml_backend_buffer_type_t buft) {
     return SIZE_MAX;
 }
 
-GGML_CALL size_t ggml_backend_buft_get_alloc_size(ggml_backend_buffer_type_t buft, struct ggml_tensor * tensor) {
+GGML_CALL size_t ggml_backend_buft_get_alloc_size(ggml_backend_buffer_type_t buft, const struct ggml_tensor * tensor) {
     // get_alloc_size is optional, defaults to ggml_nbytes
     if (buft->iface.get_alloc_size) {
         size_t size = buft->iface.get_alloc_size(buft, tensor);
@@ -114,7 +114,7 @@ size_t ggml_backend_buffer_get_max_size(ggml_backend_buffer_t buffer) {
     return ggml_backend_buft_get_max_size(ggml_backend_buffer_get_type(buffer));
 }
 
-size_t ggml_backend_buffer_get_alloc_size(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor) {
+size_t ggml_backend_buffer_get_alloc_size(ggml_backend_buffer_t buffer, const struct ggml_tensor * tensor) {
     return ggml_backend_buft_get_alloc_size(ggml_backend_buffer_get_type(buffer), tensor);
 }