[GPU] Fix oob in reorder for BMG (#28200)

### Details:
 - Fix for border_gpu.zyx_bsv16fsv16 case
 - case when x size less than tile size was not handled

src/plugins/intel_gpu/src/kernel_selector/cl_kernels/reorder_data_bfyx_to_blocked_format.cl

@@ -26,6 +26,18 @@
                                         } \
                                     }
 
+#define FUNC_LOAD_LEFTOVERS(inner, outer)    unroll_for (uint lh = 0; lh < outer; ++lh) { \
+                                        const uint input_idx = INPUT0_GET_TILED_INDEX(INPUT0_TILED_ORDER); \
+                                        INPUTVTYPE read_data; \
+                                        unroll_for (uint lw = 0; lw < inner; ++lw) { \
+                                            read_data[lw] = input[input_idx + lw]; \
+                                        } \
+                                        unroll_for (uint lw = 0; lw < inner; ++lw) { \
+                                            const uint dst = local_buf_offset + lw; \
+                                            transpose_buf[dst][lh] = read_data[lw]; \
+                                        } \
+                                    }
+
 #define FUNC_VSTORE(loop)           unroll_for (uint lw = 0; lw < loop; ++lw) { \
                                         const uint output_idx = output_idx_tile + (lw * x_pitch); \
                                         VSTORE(TO_OUTPUTVTYPE(transpose_buf[local_buf_offset + lw]), 0, output + output_idx); \
@@ -109,7 +121,15 @@ KERNEL (reorder_data_bfyx_to_blocked_format)(
 
     if (F_NO_REMAINDER_CONDITION) {
         // read and transpose
+#ifdef X_REMAINDER_CONDITION
+        if (X_NO_REMAINDER_CONDITION) {
+            FUNC_VLOAD(TILE_SIZE, TILE_SIZE)
+        } else {
+            FUNC_LOAD_LEFTOVERS(X_REMAINDER_SIZE, TILE_SIZE)
+        }
+#else
         FUNC_VLOAD(TILE_SIZE, TILE_SIZE)
+#endif
 
         // write to ddr
 #ifdef X_REMAINDER_CONDITION
@@ -125,7 +145,15 @@ KERNEL (reorder_data_bfyx_to_blocked_format)(
 #ifdef F_REMAINDER_CONDITION
     else if (F_REMAINDER_CONDITION) {
         // read and transpose
+    #ifdef X_REMAINDER_CONDITION
+        if (X_NO_REMAINDER_CONDITION) {
+            FUNC_VLOAD(TILE_SIZE, F_REMAINDER_SIZE)
+        } else {
+            FUNC_LOAD_LEFTOVERS(X_REMAINDER_SIZE, F_REMAINDER_SIZE)
+        }
+    #else
         FUNC_VLOAD(TILE_SIZE, F_REMAINDER_SIZE)
+    #endif
 
         // write to ddr
     #ifdef X_REMAINDER_CONDITION