pjs/tools/footprint/foldelf.cpp

338 строки
8.9 KiB
C++

/* -*- Mode: C++; indent-tabs-mode: nil; c-basic-offset: 4 -*-
*
* The contents of this file are subject to the Mozilla Public
* License Version 1.1 (the "License"); you may not use this file
* except in compliance with the License. You may obtain a copy of
* the License at http://www.mozilla.org/MPL/
*
* Software distributed under the License is distributed on an "AS
* IS" basis, WITHOUT WARRANTY OF ANY KIND, either express oqr
* implied. See the License for the specific language governing
* rights and limitations under the License.
*
* The Original Code is foldelf.cpp, released November 28, 2000.
*
* The Initial Developer of the Original Code is Netscape
* Communications Corporation. Portions created by Netscape are
* Copyright (C) 2000 Netscape Communications Corporation. All
* Rights Reserved.
*
* Contributor(s):
* Chris Waterson <waterson@netscape.com>
*
* This program reads an ELF file and computes information about
* redundancies.
*
*
*/
#include <algorithm>
#include <fstream>
#include <string>
#include <vector>
#include <map>
#include <elf.h>
#include <sys/mman.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <unistd.h>
#include <errno.h>
#include <getopt.h>
//----------------------------------------------------------------------
char* opt_type;
char* opt_section;
//----------------------------------------------------------------------
static void
hexdump(ostream& out, const char* bytes, size_t count)
{
hex(out);
size_t off = 0;
while (off < count) {
out.form("%08lx: ", off);
const char* p = bytes + off;
int j = 0;
while (j < 16) {
out.form("%02x", p[j++] & 0xff);
if (j + off >= count)
break;
out.form("%02x ", p[j++] & 0xff);
if (j + off >= count)
break;
}
// Pad
for (; j < 16; ++j)
out << ((j%2) ? " " : " ");
for (j = 0; j < 16; ++j) {
if (j + off < count)
out.put(isprint(p[j]) ? p[j] : '.');
}
out << endl;
off += 16;
}
}
//----------------------------------------------------------------------
int
verify_elf_header(const Elf32_Ehdr* hdr)
{
if (hdr->e_ident[EI_MAG0] != ELFMAG0
|| hdr->e_ident[EI_MAG1] != ELFMAG1
|| hdr->e_ident[EI_MAG2] != ELFMAG2
|| hdr->e_ident[EI_MAG3] != ELFMAG3) {
cerr << "not an elf file" << endl;
return -1;
}
if (hdr->e_ident[EI_CLASS] != ELFCLASS32) {
cerr << "not a 32-bit elf file" << endl;
return -1;
}
if (hdr->e_ident[EI_DATA] != ELFDATA2LSB) {
cerr << "not a little endian elf file" << endl;
return -1;
}
if (hdr->e_ident[EI_VERSION] != EV_CURRENT) {
cerr << "incompatible version" << endl;
return -1;
}
return 0;
}
//----------------------------------------------------------------------
class elf_symbol : public Elf32_Sym
{
public:
elf_symbol(const Elf32_Sym& sym)
{ ::memcpy(static_cast<Elf32_Sym*>(this), &sym, sizeof(Elf32_Sym)); }
friend bool operator==(const elf_symbol& lhs, const elf_symbol& rhs) {
return 0 == ::memcmp(static_cast<const Elf32_Sym*>(&lhs),
static_cast<const Elf32_Sym*>(&rhs),
sizeof(Elf32_Sym)); }
};
//----------------------------------------------------------------------
static const char*
st_bind(unsigned char info)
{
switch (ELF32_ST_BIND(info)) {
case STB_LOCAL: return "local";
case STB_GLOBAL: return "global";
case STB_WEAK: return "weak";
default: return "unknown";
}
}
static const char*
st_type(unsigned char info)
{
switch (ELF32_ST_TYPE(info)) {
case STT_NOTYPE: return "none";
case STT_OBJECT: return "object";
case STT_FUNC: return "func";
case STT_SECTION: return "section";
case STT_FILE: return "file";
default: return "unknown";
}
}
static unsigned char
st_type(const char* type)
{
if (strcmp(type, "none") == 0) {
return STT_NOTYPE;
}
else if (strcmp(type, "object") == 0) {
return STT_OBJECT;
}
else if (strcmp(type, "func") == 0) {
return STT_FUNC;
}
else {
return 0;
}
}
//----------------------------------------------------------------------
typedef vector<elf_symbol> elf_symbol_table;
typedef map< basic_string<char>, elf_symbol_table > elf_text_map;
void
process_mapping(char* mapping, size_t size)
{
const Elf32_Ehdr* ehdr = reinterpret_cast<Elf32_Ehdr*>(mapping);
if (verify_elf_header(ehdr) < 0)
return;
// find the section headers
const Elf32_Shdr* shdrs = reinterpret_cast<Elf32_Shdr*>(mapping + ehdr->e_shoff);
// find the section header string table, .shstrtab
const Elf32_Shdr* shstrtabsh = shdrs + ehdr->e_shstrndx;
const char* shstrtab = mapping + shstrtabsh->sh_offset;
// find the sections we care about
const Elf32_Shdr *symtabsh, *strtabsh, *textsh;
int textndx;
for (int i = 0; i < ehdr->e_shnum; ++i) {
basic_string<char> name(shstrtab + shdrs[i].sh_name);
if (name == opt_section) {
textsh = shdrs + i;
textndx = i;
}
else if (name == ".symtab") {
symtabsh = shdrs + i;
}
else if (name == ".strtab") {
strtabsh = shdrs + i;
}
}
// find the .strtab
char* strtab = mapping + strtabsh->sh_offset;
// find the .text
char* text = mapping + textsh->sh_offset;
int textaddr = textsh->sh_addr;
// find the symbol table
int nentries = symtabsh->sh_size / sizeof(Elf32_Sym);
Elf32_Sym* symtab = reinterpret_cast<Elf32_Sym*>(mapping + symtabsh->sh_offset);
// look for symbols in the .text section
elf_text_map textmap;
for (int i = 0; i < nentries; ++i) {
const Elf32_Sym* sym = symtab + i;
if (sym->st_shndx == textndx &&
ELF32_ST_TYPE(sym->st_info) == st_type(opt_type) &&
sym->st_size) {
basic_string<char> functext(text + sym->st_value - textaddr, sym->st_size);
elf_symbol_table& syms = textmap[functext];
if (syms.end() == find(syms.begin(), syms.end(), elf_symbol(*sym)))
syms.insert(syms.end(), *sym);
}
}
int uniquebytes = 0, totalbytes = 0;
int uniquecount = 0, totalcount = 0;
for (elf_text_map::const_iterator entry = textmap.begin();
entry != textmap.end();
++entry) {
const elf_symbol_table& syms = entry->second;
if (syms.size() <= 1)
continue;
int sz = syms.begin()->st_size;
uniquebytes += sz;
totalbytes += sz * syms.size();
uniquecount += 1;
totalcount += syms.size();
for (elf_symbol_table::const_iterator sym = syms.begin(); sym != syms.end(); ++sym)
cout << strtab + sym->st_name << endl;
dec(cout);
cout << syms.size() << " copies of " << sz << " bytes";
cout << " (" << ((syms.size() - 1) * sz) << " redundant bytes)" << endl;
hexdump(cout, entry->first.data(), entry->first.size());
cout << endl;
}
dec(cout);
cout << "bytes unique=" << uniquebytes << ", total=" << totalbytes << endl;
cout << "entries unique=" << uniquecount << ", total=" << totalcount << endl;
}
void
process_file(const char* name)
{
int fd = open(name, O_RDWR);
if (fd >= 0) {
struct stat statbuf;
if (fstat(fd, &statbuf) >= 0) {
size_t size = statbuf.st_size;
void* mapping = mmap(0, size, PROT_READ, MAP_SHARED, fd, 0);
if (mapping != MAP_FAILED) {
process_mapping(static_cast<char*>(mapping), size);
munmap(mapping, size);
}
}
close(fd);
}
}
static void
usage()
{
cerr << "foldelf [--section=<section>] [--type=<type>] [file ...]\n\
--section, -s the section of the ELF file to scan; defaults\n\
to ``.text''. Valid values include any section\n\
of the ELF file.\n\
--type, -t the type of object to examine in the section;\n\
defaults to ``func''. Valid values include\n\
``none'', ``func'', or ``object''.\n";
}
static struct option opts[] = {
{ "type", required_argument, 0, 't' },
{ "section", required_argument, 0, 's' },
{ "help", no_argument, 0, '?' },
{ 0, 0, 0, 0 }
};
int
main(int argc, char* argv[])
{
while (1) {
int option_index = 0;
int c = getopt_long(argc, argv, "t:s:", opts, &option_index);
if (c < 0) break;
switch (c) {
case 't':
opt_type = optarg;
break;
case 's':
opt_section = optarg;
break;
case '?':
usage();
break;
}
}
for (int i = optind; i < argc; ++i)
process_file(argv[i]);
return 0;
}