From ecfeed59ee5a9ab4f0c6db30e85ffcffdf89fecf Mon Sep 17 00:00:00 2001
From: York Jasper Niebuhr <yjnworkstation@gmail.com>
Date: Sat, 25 Oct 2025 23:40:13 +0200
Subject: [PATCH] Made subject multi-CU

---
 finalize/finalize.cpp              | 31 +++++-----
 pinpoint/pinpoint.cpp              |  1 +
 pinpoint/safegcc/safe-input.h      |  9 +++
 pinpoint/stage0/CMakeLists.txt     |  2 +-
 pinpoint/stage0/on_finish_decl.cpp | 93 ++++++++++++++++++++++++++++++
 pinpoint/stage0/stage0.h           | 15 +++++
 subject/CMakeLists.txt             |  3 +-
 subject/main.c                     | 55 ++----------------
 subject/second.c                   |  7 +++
 subject/sub/second.c               |  7 +++
 subject/task_struct.h              | 49 ++++++++++++++++
 11 files changed, 203 insertions(+), 69 deletions(-)
 create mode 100644 pinpoint/safegcc/safe-input.h
 create mode 100644 pinpoint/stage0/on_finish_decl.cpp
 create mode 100644 subject/second.c
 create mode 100644 subject/sub/second.c
 create mode 100644 subject/task_struct.h

diff --git a/finalize/finalize.cpp b/finalize/finalize.cpp
index f7594ac..5400f14 100644
--- a/finalize/finalize.cpp
+++ b/finalize/finalize.cpp
@@ -1,29 +1,26 @@
 #include <iostream>
 
 /*
-<o>.spslr.targets:
-target <uid> <size> <field count>
+<sourcefile>.spslr:
+
+SPSLR_FINALIZE_HEADER <CU filename> <CU uid symbol>
+target <hash> <local uid> <size> <field count>
 f <offset> <size> <flags>
 f <offset> <size> <flags>
 ...
-
-<o>.spslr.ipins:
-(prefix __spslr_pin)
-0 <target uid> <field offset>
-1 <target uid> <field offset>
-2 <target uid> <field offset>
+ipin <label> <target uid> <field offset>
+ipin <label> <target uid> <field offset>
+ipin <label> <target uid> <field offset>
+...
+dpin <symbol> <offset> <level> <target uid>
+dpin <symbol> <offset> <level> <target uid>
 ...
 
-<o>.spslr.dpins:
-myVar1 2 <strip true/false>
-<offset> <lvl> <target uid>
-<offset> <lvl> <target uid>
-myVar2 1 <strip true/false>
-<offset> <lvl> <target uid>
-...
 
-Datapins are randomized in order of their level, from bottom of nest to top
-Datapin strip flag is set if it would not have been preserved in symtab without pinpoint plugin
+
+Datapins for same var/symbol are randomized in order of their level, from bottom of nest to top
+The CU uid symbol helps differentiating between e.g. "file.c" and "sub/file.c" (symbtab has no idea)
+Between CUs, types with the same hash are considered to be the same type
 
 */
 
diff --git a/pinpoint/pinpoint.cpp b/pinpoint/pinpoint.cpp
index e812438..8eeb892 100644
--- a/pinpoint/pinpoint.cpp
+++ b/pinpoint/pinpoint.cpp
@@ -19,6 +19,7 @@ int plugin_init(struct plugin_name_args* plugin_info, struct plugin_gcc_version*
 	register_callback(plugin_info->base_name, PLUGIN_ATTRIBUTES, on_register_attributes, NULL);
 	register_callback(plugin_info->base_name, PLUGIN_FINISH_TYPE, on_finish_type, NULL);
 	register_callback(plugin_info->base_name, PLUGIN_BUILD_COMPONENT_REF, on_preserve_component_ref, NULL);
+	register_callback(plugin_info->base_name, PLUGIN_FINISH_DECL, on_finish_decl, NULL);
 
 	struct register_pass_info separate_offset_pass_info;
 	separate_offset_pass_info.pass = new separate_offset_pass(nullptr);
diff --git a/pinpoint/safegcc/safe-input.h b/pinpoint/safegcc/safe-input.h
new file mode 100644
index 0000000..fe24435
--- /dev/null
+++ b/pinpoint/safegcc/safe-input.h
@@ -0,0 +1,9 @@
+#include <safe-gcc-plugin.h>
+
+#ifndef SAFEGCC_INPUT_H
+#define SAFEGCC_INPUT_H
+
+#include <input.h>
+#include <libiberty.h>
+
+#endif
diff --git a/pinpoint/stage0/CMakeLists.txt b/pinpoint/stage0/CMakeLists.txt
index 7aa01a8..682969f 100644
--- a/pinpoint/stage0/CMakeLists.txt
+++ b/pinpoint/stage0/CMakeLists.txt
@@ -1,3 +1,3 @@
 target_sources(spslr_pinpoint PRIVATE on_register_attributes.cpp on_finish_type.cpp on_preserve_component_ref.cpp
-	target.cpp separator.cpp separate_offset_pass.cpp)
+	on_finish_decl.cpp target.cpp separator.cpp separate_offset_pass.cpp)
 target_include_directories(spslr_pinpoint PRIVATE .)
diff --git a/pinpoint/stage0/on_finish_decl.cpp b/pinpoint/stage0/on_finish_decl.cpp
new file mode 100644
index 0000000..e4bc5bb
--- /dev/null
+++ b/pinpoint/stage0/on_finish_decl.cpp
@@ -0,0 +1,93 @@
+#include <iostream> // TODO
+#include <stage0.h>
+
+#include <pinpoint_error.h>
+#include <safe-input.h>
+
+static std::list<DataPin> pins;
+
+const std::list<DataPin>& dpins() {
+	return pins;
+}
+
+/*
+struct DataPin {
+    struct Component {
+        std::size_t offset;
+        std::size_t level;
+        UID target;
+    };
+
+    std::string symbol;
+    std::list<Component> components;
+};
+*/
+
+/*
+bool type_matches(tree t, hash_set<tree> *target_types)
+{
+    if (!t) return false;
+    t = TYPE_MAIN_VARIANT(t);
+
+    if (target_types->contains(t))
+        return true;
+
+    if (TREE_CODE(t) == RECORD_TYPE || TREE_CODE(t) == UNION_TYPE)
+    {
+        for (tree field = TYPE_FIELDS(t); field; field = TREE_CHAIN(field))
+        {
+            if (TREE_CODE(field) == FIELD_DECL)
+            {
+                tree ftype = TREE_TYPE(field);
+                if (type_matches(ftype, target_types))
+                    return true;
+            }
+        }
+    }
+
+    return false;
+}
+*/
+
+static bool compile_datapin(tree type, DataPin& pin) {
+	// TODO
+	return true;
+}
+
+static void on_static_var(tree var) {
+	tree type = TREE_TYPE(var);
+	if (!type)
+		return;
+
+	DataPin pin;
+	if (!compile_datapin(type, pin))
+		return;
+
+	DECL_PRESERVE_P(var) = 1;
+	DECL_VISIBILITY(var) = VISIBILITY_DEFAULT;
+	DECL_VISIBILITY_SPECIFIED(var) = 1;
+
+	tree symbol_tree = DECL_ASSEMBLER_NAME(var);
+	const char* symbol;
+	if (!symbol_tree || !(symbol = IDENTIFIER_POINTER(symbol_tree)))
+		pinpoint_fatal("on_static_var failed to get symbol of static variable");
+
+	pin.symbol = std::string{ symbol };
+
+	pins.emplace_back(std::move(pin));
+
+	// TODO
+	std::cout << "Pinned: \"" << symbol << "\" (" << lbasename(main_input_filename) << ")" << std::endl;
+}
+
+void on_finish_decl(void* plugin_data, void* user_data) {
+	tree decl = (tree)plugin_data;
+
+	if (TREE_CODE(decl) != VAR_DECL)
+		return;
+
+	if (!TREE_STATIC(decl) || DECL_EXTERNAL(decl))
+		return;
+
+	on_static_var(decl);
+}
diff --git a/pinpoint/stage0/stage0.h b/pinpoint/stage0/stage0.h
index d294022..e8da7a0 100644
--- a/pinpoint/stage0/stage0.h
+++ b/pinpoint/stage0/stage0.h
@@ -2,6 +2,7 @@
 #include <cstddef>
 #include <map>
 #include <string>
+#include <list>
 #include <limits>
 
 #include <safe-tree.h>
@@ -64,9 +65,23 @@ tree make_stage0_ast_separator(UID target, std::size_t offset);
 gimple* make_stage0_gimple_separator(tree lhs, UID target, std::size_t offset);
 bool is_stage0_separator(gimple* stmt, UID& target, std::size_t& offset);
 
+struct DataPin {
+	struct Component {
+		std::size_t offset;
+		std::size_t level;
+		UID target;
+	};
+
+	std::string symbol;
+	std::list<Component> components;
+};
+
+const std::list<DataPin>& dpins();
+
 void on_register_attributes(void* plugin_data, void* user_data);
 void on_finish_type(void* plugin_data, void* user_data);
 void on_preserve_component_ref(void* plugin_data, void* user_data);
+void on_finish_decl(void* plugin_data, void* user_data);
 
 struct separate_offset_pass : gimple_opt_pass {
 	separate_offset_pass(gcc::context* ctxt);
diff --git a/subject/CMakeLists.txt b/subject/CMakeLists.txt
index 7b42e0e..0e4b6d3 100644
--- a/subject/CMakeLists.txt
+++ b/subject/CMakeLists.txt
@@ -1,4 +1,5 @@
-add_executable(subject main.c)
+add_executable(subject main.c second.c sub/second.c)
+target_include_directories(subject PRIVATE .)
 add_dependencies(subject spslr_pinpoint spslr_finalize spslr_selfpatch)
 target_link_libraries(subject PRIVATE spslr_selfpatch)
 target_compile_options(subject PRIVATE -O1 -fplugin=$<TARGET_FILE:spslr_pinpoint> -fdump-tree-separate_offset -fdump-tree-asm_offset)
diff --git a/subject/main.c b/subject/main.c
index 1f3e2d7..737d062 100644
--- a/subject/main.c
+++ b/subject/main.c
@@ -1,54 +1,9 @@
-// test_task_struct.c
-#include <stdio.h>
-#include <stddef.h> // for offsetof
+#include "task_struct.h"
 
-#define container_of(ptr, type, member) ({ \
-    const typeof(((type*)0)->member)* __mptr = (ptr); \
-    (type*)((char*)__mptr - offsetof(type, member)); })
+int second_pid();
+int third_pid();
 
-// Minimal doubly linked list
-struct list_head {
-    struct list_head *next, *prev;
-};
-
-static inline void INIT_LIST_HEAD(struct list_head *list)
-{
-    list->next = list;
-    list->prev = list;
-}
-
-static inline void __list_add(struct list_head *new,
-                              struct list_head *prev,
-                              struct list_head *next)
-{
-    next->prev = new;
-    new->next = next;
-    new->prev = prev;
-    prev->next = new;
-}
-
-static inline void list_add(struct list_head *new, struct list_head *head)
-{
-    __list_add(new, head, head->next);
-}
-
-static inline void list_add_tail(struct list_head *new, struct list_head *head)
-{
-    __list_add(new, head->prev, head);
-}
-
-#define list_entry(ptr, type, member) \
-    container_of(ptr, type, member)
-
-#define list_for_each(pos, head) \
-    for (pos = (head)->next; pos != (head); pos = pos->next)
-
-// A small struct like the Linux kernel's task_struct
-struct task_struct {
-    int pid;
-    const char *comm;
-    struct list_head tasks;  // linkage for global task list
-} __attribute__((spslr));
+struct task_struct global = { .pid = 42, .comm = "main_global" };
 
 int main(void)
 {
@@ -77,6 +32,6 @@ int main(void)
 
 	size_t myOffset = ((size_t)&((struct task_struct*)0)->tasks); // BROKEN, relevancy for kernel unknown
 
-    return 0;
+    return second_pid() * third_pid();
 }
 
diff --git a/subject/second.c b/subject/second.c
new file mode 100644
index 0000000..7f1993a
--- /dev/null
+++ b/subject/second.c
@@ -0,0 +1,7 @@
+#include "task_struct.h"
+
+static struct task_struct global = { .pid = 43, .comm = "third_global" };
+
+int third_pid() {
+	return global.pid;
+}
diff --git a/subject/sub/second.c b/subject/sub/second.c
new file mode 100644
index 0000000..84fd473
--- /dev/null
+++ b/subject/sub/second.c
@@ -0,0 +1,7 @@
+#include "task_struct.h"
+
+static struct task_struct global = { .pid = 0, .comm = "second_global" };
+
+int second_pid() {
+	return global.pid;
+}
diff --git a/subject/task_struct.h b/subject/task_struct.h
new file mode 100644
index 0000000..5cf4f70
--- /dev/null
+++ b/subject/task_struct.h
@@ -0,0 +1,49 @@
+#include <stdio.h>
+#include <stddef.h>
+
+#define container_of(ptr, type, member) ({ \
+	const typeof(((type*)0)->member)* __mptr = (ptr); \
+	(type*)((char*)__mptr - offsetof(type, member)); })
+
+// Minimal doubly linked list
+struct list_head {
+	struct list_head *next, *prev;
+};
+
+static inline void INIT_LIST_HEAD(struct list_head *list)
+{
+	list->next = list;
+	list->prev = list;
+}
+
+static inline void __list_add(struct list_head *new, struct list_head *prev, struct list_head *next)
+{
+	next->prev = new;
+	new->next = next;
+	new->prev = prev;
+	prev->next = new;
+}
+
+static inline void list_add(struct list_head *new, struct list_head *head)
+{
+	__list_add(new, head, head->next);
+}
+
+static inline void list_add_tail(struct list_head *new, struct list_head *head)
+{
+	__list_add(new, head->prev, head);
+}
+
+#define list_entry(ptr, type, member) \
+	container_of(ptr, type, member)
+
+#define list_for_each(pos, head) \
+	for (pos = (head)->next; pos != (head); pos = pos->next)
+
+// A small struct like the Linux kernel's task_struct
+struct task_struct {
+	int pid;
+	const char *comm;
+	struct list_head tasks; // linkage for global task list
+} __attribute__((spslr));
+