loader: Fix GetProcAddr to return proper function pointers for app

Add case of handling null gpu object for global  functions.
For device specifc functions use the dispatch table or  the loader entrypoints
depending on if the trampoline code can be skipped or not.

Conflicts:
	loader/CMakeLists.txt
	vk-generate.py

v2: undo accidental rename of LayerInterceptProcSubcommand
    fix WinDefFileSubcommand due to rebase (olv)

1 files changed, 37 insertions, 5 deletions

diff --git a/loader/loader.c b/loader/loader.c
index 3ce99f01..712118ac 100644
--- a/loader/loader.c
+++ b/loader/loader.c
@@ -42,6 +42,7 @@
 #endif // WIN32
 #include "loader_platform.h"
 #include "table_ops.h"
+#include "gpa_helper.h"
 #include "loader.h"
 #include "vkIcd.h"
 // The following is #included again to catch certain OS-specific functions
@@ -711,6 +712,28 @@ static void loader_init_dispatch_table(VkLayerDispatchTable *tab, PFN_vkGetProcA
         tab->EnumerateLayers = vkEnumerateLayers;
 }
 
+static void *loader_gpa_internal(VkPhysicalGpu gpu, const char * pName)
+{
+    if (gpu == NULL) {
+        return NULL;;
+    }
+    VkBaseLayerObject* gpuw = (VkBaseLayerObject *) gpu;
+    VkLayerDispatchTable * disp_table = * (VkLayerDispatchTable **) gpuw->baseObject;
+    void *addr;
+
+    if (disp_table == NULL)
+        return NULL;
+
+    addr = loader_lookup_dispatch_table(disp_table, pName);
+    if (addr)
+        return addr;
+    else  {
+        if (disp_table->GetProcAddr == NULL)
+            return NULL;
+        return disp_table->GetProcAddr(gpuw->nextObject, pName);
+    }
+}
+
 extern struct loader_icd * loader_get_icd(const VkBaseLayerObject *gpu, uint32_t *gpu_index)
 {
     /*
@@ -971,7 +994,7 @@ extern uint32_t loader_activate_layers(struct loader_icd *icd, uint32_t gpu_inde
     if (!loader_layers_activated(icd, gpu_index)) {
         VkBaseLayerObject *gpuObj = gpu;
         VkBaseLayerObject *nextGpuObj, *baseObj = gpuObj->baseObject;
-        PFN_vkGetProcAddr nextGPA = vkGetProcAddr;
+        PFN_vkGetProcAddr nextGPA = loader_gpa_internal;
 
         count = loader_get_layer_libs(icd, gpu_index, ext_count, ext_names, &pLayerNames);
         if (!count)
@@ -1211,12 +1234,21 @@ LOADER_EXPORT VkResult VKAPI vkEnumerateGpus(
 LOADER_EXPORT void * VKAPI vkGetProcAddr(VkPhysicalGpu gpu, const char * pName)
 {
     if (gpu == NULL) {
-        return NULL;
+
+        /* return entrypoint addresses that are global (in the loader)*/
+        return globalGetProcAddr(pName);
     }
-    VkBaseLayerObject* gpuw = (VkBaseLayerObject *) gpu;
-    VkLayerDispatchTable * disp_table = * (VkLayerDispatchTable **) gpuw->baseObject;
+
     void *addr;
 
+    /* for entrypoints that loader must handle (ie non-dispatchable or create object)
+       make sure the loader entrypoint is returned */
+    addr = loader_non_passthrough_gpa(pName);
+    if (addr)
+        return addr;
+
+    /* return the dispatch table entrypoint for the fastest case */
+    const VkLayerDispatchTable *disp_table = * (VkLayerDispatchTable **) gpu;
     if (disp_table == NULL)
         return NULL;
 
@@ -1226,7 +1258,7 @@ LOADER_EXPORT void * VKAPI vkGetProcAddr(VkPhysicalGpu gpu, const char * pName)
     else  {
         if (disp_table->GetProcAddr == NULL)
             return NULL;
-        return disp_table->GetProcAddr(gpuw->nextObject, pName);
+        return disp_table->GetProcAddr(gpu, pName);
     }
 }