From ec396f1dce7af881fcf8e067bd5f5a26b3110eaf Mon Sep 17 00:00:00 2001
From: Jason Hiser <jdhiser@gmail.com>
Date: Tue, 7 Aug 2018 09:41:21 -0400
Subject: [PATCH] revised cfg builder to build blocks for apparently
unreachable code.
Former-commit-id: c04807a76e1358c5b454fb64dd19a13fde642fd0
---
.gitignore | 2 +-
libIRDB/include/cfg/BasicBlock.hpp | 3 +-
libIRDB/include/cfg/CFG.hpp | 15 +++---
libIRDB/src/cfg/BasicBlock.cpp | 4 +-
libIRDB/src/cfg/CFG.cpp | 83 +++++++++++++++++++++---------
5 files changed, 71 insertions(+), 36 deletions(-)
diff --git a/.gitignore b/.gitignore
index 10e73a2eb..5477298a9 100644
--- a/.gitignore
+++ b/.gitignore
@@ -9,4 +9,4 @@ lib/
manifest.txt.config
tools/meds2pdb/meds2pdb
plugins_install/*.exe
-
+*.swp
diff --git a/libIRDB/include/cfg/BasicBlock.hpp b/libIRDB/include/cfg/BasicBlock.hpp
index 42a9a4bf8..7e6475280 100644
--- a/libIRDB/include/cfg/BasicBlock.hpp
+++ b/libIRDB/include/cfg/BasicBlock.hpp
@@ -53,8 +53,7 @@ class BasicBlock_t
protected:
- void BuildBlock(Function_t* func,
- Instruction_t* insn,
+ void BuildBlock( Instruction_t* insn,
const std::map<Instruction_t*,BasicBlock_t*> &insn2block_map
);
diff --git a/libIRDB/include/cfg/CFG.hpp b/libIRDB/include/cfg/CFG.hpp
index 647f916aa..ae2c6d0cd 100644
--- a/libIRDB/include/cfg/CFG.hpp
+++ b/libIRDB/include/cfg/CFG.hpp
@@ -26,24 +26,27 @@ class ControlFlowGraph_t
{
public:
ControlFlowGraph_t(Function_t* func);
-
BasicBlock_t* GetEntry() const { return entry; }
Function_t* GetFunction() const { return function; }
+ BasicBlockSet_t& GetBlocks() { return blocks; }
+ const BasicBlockSet_t& GetBlocks() const { return blocks; }
void dump(std::ostream &os=std::cout) const { os<<*this; }
- protected:
+ private:
+ // methods
void Build(Function_t *func);
+ void alloc_blocks(const InstructionSet_t &starts, map<Instruction_t*,BasicBlock_t*>& insn2block_map);
+ void build_blocks(const map<Instruction_t*,BasicBlock_t*>& insn2block_map);
+ void find_unblocked_instructions(InstructionSet_t &starts, Function_t* func);
- private:
- std::set<BasicBlock_t*> blocks;
+ // data
+ BasicBlockSet_t blocks;
BasicBlock_t* entry;
Function_t* function;
/* friends */
public:
friend std::ostream& operator<<(std::ostream& os, const ControlFlowGraph_t& cfg);
- BasicBlockSet_t& GetBlocks() { return blocks; }
- const BasicBlockSet_t& GetBlocks() const { return blocks; }
};
diff --git a/libIRDB/src/cfg/BasicBlock.cpp b/libIRDB/src/cfg/BasicBlock.cpp
index bb0e5c7dc..ccba987cd 100644
--- a/libIRDB/src/cfg/BasicBlock.cpp
+++ b/libIRDB/src/cfg/BasicBlock.cpp
@@ -37,11 +37,11 @@ BasicBlock_t::BasicBlock_t()
void BasicBlock_t::BuildBlock
(
- Function_t* func,
Instruction_t* insn,
const map<Instruction_t*,BasicBlock_t*> &insn2block_map
)
{
+ const auto &func=insn->GetFunction();
assert(insn);
/* loop through the instructions for this block */
while(insn)
@@ -116,7 +116,7 @@ void BasicBlock_t::BuildBlock
break;
/* check for a fallthrough out of the function */
- if(ft_insn && !is_in_container(func->GetInstructions(),ft_insn))
+ if(ft_insn && ft_insn->GetFunction() != func) // !is_in_container(func->GetInstructions(),ft_insn))
break;
diff --git a/libIRDB/src/cfg/CFG.cpp b/libIRDB/src/cfg/CFG.cpp
index 2e1b80c83..cfbc731f8 100644
--- a/libIRDB/src/cfg/CFG.cpp
+++ b/libIRDB/src/cfg/CFG.cpp
@@ -25,10 +25,12 @@
using namespace std;
using namespace libIRDB;
+#define ALLOF(a) begin(a),end(a)
+
/*
* FindTargets - locate all possible instructions that are the target of a jump instruction
*/
-static set<Instruction_t*> FindBlockStarts(Function_t* func)
+static InstructionSet_t FindBlockStarts(Function_t* func)
{
InstructionSet_t targets;
@@ -80,55 +82,86 @@ static set<Instruction_t*> FindBlockStarts(Function_t* func)
return targets;
}
-
-
-
ControlFlowGraph_t::ControlFlowGraph_t(Function_t* func) :
entry(NULL), function(func)
{
Build(func);
}
-void ControlFlowGraph_t::Build(Function_t* func)
-{
- set<Instruction_t*> starts=FindBlockStarts(func);
-
- map<Instruction_t*,BasicBlock_t*> insn2block_map;
+void ControlFlowGraph_t::alloc_blocks(const InstructionSet_t &starts, map<Instruction_t*,BasicBlock_t*>& insn2block_map)
+{
/* create a basic block for each instruction that starts a block */
- for( set<Instruction_t*>::const_iterator it=starts.begin();
- it!=starts.end();
- ++it
- )
+ for(const auto &insn : starts)
{
- Instruction_t* insn=*it;
- BasicBlock_t* newblock=new BasicBlock_t;
+ if(is_in_container(insn2block_map,insn)) // already allocated
+ continue;
- /* record the entry block */
- if(insn==func->GetEntryPoint())
- entry=newblock;
+ auto newblock=new BasicBlock_t;
assert( insn && newblock );
blocks.insert(newblock);
insn2block_map[insn]=newblock;
}
+}
+
+void ControlFlowGraph_t::build_blocks(const map<Instruction_t*,BasicBlock_t*>& insn2block_map)
+{
/* Ask the basic block to set the fields for each block that need to be set */
- for( map<Instruction_t*,BasicBlock_t*>::const_iterator it=insn2block_map.begin();
- it!=insn2block_map.end();
- ++it
- )
+ for(const auto &it : insn2block_map)
{
- Instruction_t* insn=(*it).first;
- BasicBlock_t* block=(*it).second;
+ const auto insn=it.first;
+ const auto block=it.second;
+
+ if(block->GetInstructions().size()>0) // already built
+ continue;
assert(insn && block);
- block->BuildBlock(func, insn, insn2block_map);
+ block->BuildBlock(insn, insn2block_map);
}
+}
+
+void ControlFlowGraph_t::find_unblocked_instructions(InstructionSet_t &starts, Function_t* func)
+{
+ auto mapped_instructions=InstructionSet_t();
+ auto missed_instructions=InstructionSet_t();
+ for(const auto block : GetBlocks())
+ mapped_instructions.insert(ALLOF(block->GetInstructions()));
+
+ auto my_inserter=inserter(missed_instructions,missed_instructions.end());
+ set_difference(ALLOF(func->GetInstructions()), ALLOF(mapped_instructions), my_inserter);
+ starts.insert(ALLOF(missed_instructions));
+}
+
+
+
+void ControlFlowGraph_t::Build(Function_t* func)
+{
+ auto starts=FindBlockStarts(func);
+
+ auto insn2block_map=map<Instruction_t*,BasicBlock_t*> ();
+
+ alloc_blocks(starts, insn2block_map);
+ build_blocks(insn2block_map);
+ /* record the entry block */
+ entry=insn2block_map[func->GetEntryPoint()];
+
+ /* most functions are done now. */
+ /* however, if a function has a (direct) side entrance,
+ * some code may appear unreachable and not be placed in
+ * a block -- here, we detect that code and create a
+ * new basic block for every instruction, as any may have a side entrance
+ */
+ /* note: side entrances may miss a block start */
+ /* in code that appears reachable from the entrance?! */
+ find_unblocked_instructions(starts, func);
+ alloc_blocks(starts, insn2block_map);
+ build_blocks(insn2block_map);
}
--
GitLab