[DynInst_API:] Some documentation patches + code sample

[Alin MÃndroc <mindroc.alin@xxxxxxxxx>]

Hi everyone,

I have attached some code patches to the documentation, as following:

 CodeSource.tex: Added underscore in front of the names of the members of the "Hint" struct, as they are named in the source code

 Iterating.tex: Changed "findModule" to "findModuleByName"; also, quotes around "foo" came out as bogus unicode chars, so I replaced them

 LineInformation.tex: Fixed link for code sample which pointed to wrong section

 Symtab.C, Symtab.h: Added capital "O" for "getNumberofSymbols" and "getNumberofRegions", as they are named in the documentation

Also, an example that I have missed in your documentation was how to parse a file and get the assembly instructions for each function, a compilable piece of code together with build guidelines, like the sample in the manual for parseAPI which shows how to generate a CFG.

So I have also attached a code sample with this functionality, which can be built exactly like in the ParseAPI manual, section 3.1. I am not sure where to place it in the manual, so I have attached only the source code.

Regards,

Alin

#include <iostream>
#include "CodeObject.h"
#include "InstructionDecoder.h"

using namespace std;
using namespace Dyninst;
using namespace ParseAPI;
using namespace InstructionAPI;

int main(int argc, char **argv){
	if(argc != 2){
		printf("Usage: %s <binary path>\n", argv[0]);
		return -1;
	}

	char *binaryPath = argv[1];

	SymtabCodeSource *sts;
	CodeObject *co;
	CodeRegion *cr;
	Instruction::Ptr instr;

	SymtabAPI::Symtab *symTab;
	std::string binaryPathStr(binaryPath);
	bool isParsable = SymtabAPI::Symtab::openFile(symTab, binaryPathStr);

	if(isParsable == false){
		const char *error = "error: file can not be parsed";
		cout << error;
		return - 1;
	}

	sts = new SymtabCodeSource(binaryPath);
	co = new CodeObject(sts);

	//parse the binary given as a command line arg
	co->parse();

	//get list of all functions in the binary
	const CodeObject::funclist &all = co->funcs();
	if(all.size() == 0){
		const char *error = "error: no functions in file";
		cout << error;
		return - 1;
	}

	auto fit = all.begin();
	Function *f = *fit;

	//create an Instruction decoder which will convert the binary opcodes to strings
	InstructionDecoder decoder(f->isrc()->getPtrToInstruction(f->addr()),
			InstructionDecoder::maxInstructionLength,
			f->region()->getArch());

	for(;fit != all.end(); ++fit){
		Function *f = *fit;

		//get address of entry point for current function
		Address crtAddr = f->addr();
		int instr_count = 0;

		instr = decoder.decode((unsigned char *)f->isrc()->getPtrToInstruction(crtAddr));
		auto fbl = f->blocks().end();
		fbl--;
		Block *b = *fbl;
		Address lastAddr = b->last();

		//if current function has zero instructions, don't output it
		if(crtAddr == lastAddr)
			continue;

		cout << "\n\n\"" << f->name() << "\" :";
		while(crtAddr < lastAddr){
			//decode current instruction
			instr = decoder.decode((unsigned char *)f->isrc()->getPtrToInstruction(crtAddr));

			cout << "\n" << hex << crtAddr;
			cout << ": \"" << instr->format() << "\"";

			//go to the address of the next instruction
			crtAddr += instr->size();
			instr_count++;
		}
	}

	return 0;
}

Attachment: patches.zip
Description: Zip archive