mmap: Fix Windows handle lifetime (#19598)

* ggml: added cleanups in ggml_quantize_free
Add missing cleanup calls for IQ2_S, IQ1_M quantization types and IQ3XS with 512 blocks during quantization cleanup.

* mmap: Fix Windows handle lifetime
Move hMapping from local variable to member variable so it stays alive for the entire lifetime of the mapping.
The file mapping handle must remain valid until UnmapViewOfFile is called.
Fixes cleanup order in destructor.

* Update llama-mmap.cpp

* Update llama-mmap.cpp

Remove trailing whitespace from line 567

diff --git a/src/llama-mmap.cpp b/src/llama-mmap.cpp
index 0261e4c72c9bb7a01f2c80b9439a0fb240615fb0..c03228e9ce22dd380c86fb996a2dc9d706c0c545 100644 (file)
--- a/src/llama-mmap.cpp
+++ b/src/llama-mmap.cpp
@@ -504,6 +504,8 @@ struct llama_mmap::impl {
         }
     }
 #elif defined(_WIN32)
+    HANDLE hMapping = nullptr;
+
     impl(struct llama_file * file, size_t prefetch, bool numa) {
         GGML_UNUSED(numa);
 
@@ -511,7 +513,7 @@ struct llama_mmap::impl {
 
         HANDLE hFile = (HANDLE) _get_osfhandle(file->file_id());
 
-        HANDLE hMapping = CreateFileMappingA(hFile, NULL, PAGE_READONLY, 0, 0, NULL);
+        hMapping = CreateFileMappingA(hFile, NULL, PAGE_READONLY, 0, 0, NULL);
 
         if (hMapping == NULL) {
             DWORD error = GetLastError();
@@ -520,9 +522,9 @@ struct llama_mmap::impl {
 
         addr = MapViewOfFile(hMapping, FILE_MAP_READ, 0, 0, 0);
         DWORD error = GetLastError();
-        CloseHandle(hMapping);
 
         if (addr == NULL) {
+            CloseHandle(hMapping);
             throw std::runtime_error(format("MapViewOfFile failed: %s", llama_format_win_err(error).c_str()));
         }
 
@@ -554,9 +556,17 @@ struct llama_mmap::impl {
     }
 
     ~impl() {
-        if (!UnmapViewOfFile(addr)) {
-            LLAMA_LOG_WARN("warning: UnmapViewOfFile failed: %s\n",
-                    llama_format_win_err(GetLastError()).c_str());
+        if (hMapping) {
+            if (addr) {
+                if (!UnmapViewOfFile(addr)) {
+                    LLAMA_LOG_WARN("warning: UnmapViewOfFile failed: %s\n",
+                            llama_format_win_err(GetLastError()).c_str());
+                }
+            }
+            if (!CloseHandle(hMapping)) {
+                LLAMA_LOG_WARN("warning: CloseHandle failed: %s\n",
+                        llama_format_win_err(GetLastError()).c_str());
+            }
         }
     }
 #else