llama: free backend compute scratch on cold; guard memory_breakdown

On device release with KV eviction, also call ggml_backend_free_scratch()
on each backend so a cold model drops its Vulkan compute preallocations
(reallocated lazily on the next compute via restore).

Also guard llama_context::memory_breakdown() against a freed scheduler:
evict_kv release resets sched to null, so a cold model that is then torn
down (e.g. terminated by the process manager) hit GGML_ASSERT(sched) in
ggml_backend_sched_get_buffer_type. Skip the compute-buffer accounting
when sched is null.

Assisted-by: Claude

src/llama-context.cpp

@@ -742,6 +742,11 @@ void llama_context::release_device(bool evict_kv) {
         // model holds essentially no VRAM. Rebuilt lazily by sched_reserve() on restore.
         sched.reset();
         sched_need_reserve = true;
+        // finally, free each backend's own compute-scratch (Vulkan prealloc/staging buffers, which
+        // are owned by the backend and survive sched.reset()). Reallocated lazily on next compute.
+        for (auto & backend : backends) {
+            ggml_backend_free_scratch(backend.get());
+        }
     }
 }
 
@@ -3265,6 +3270,9 @@ llama_memory_breakdown llama_context::memory_breakdown() const {
             ret[buft].context += size;
         }
     }
+    // the scheduler (and its compute buffers) may have been freed while the model is cold
+    // (on-demand VRAM eviction, see release_device); it contributes no compute VRAM then.
+    if (sched) {
         if (model.hparams.no_alloc) {
             for (size_t i = 0; i < backends.size(); ++i) {
                 ggml_backend_t             backend = backends[i].get();
@@ -3278,6 +3286,7 @@ llama_memory_breakdown llama_context::memory_breakdown() const {
                 ret[buft].compute += ggml_backend_sched_get_buffer_size(sched.get(), backend);
             }
         }
+    }
     return ret;
 }