This commit is contained in:
Shiz 2018-04-12 00:14:36 +02:00
commit 6081557aa2
9 changed files with 444 additions and 0 deletions

2
.gitignore vendored Normal file
View File

@ -0,0 +1,2 @@
/bin
/obj

44
Makefile Normal file
View File

@ -0,0 +1,44 @@
LDRDIR = ldr
OBJDIR = obj
BINDIR = bin
SRCDIR = src
DATADIR = data
COPTFLAGS=-Os -fvisibility=hidden -mpreferred-stack-boundary=2 -fwhole-program \
-ffast-math -funsafe-math-optimizations -fno-stack-protector -fomit-frame-pointer \
-fno-exceptions -fno-unwind-tables -fno-asynchronous-unwind-tables
CXXOPTFLAGS=$(COPTFLAGS) \
-fno-rtti -fno-enforce-eh-specs -fnothrow-opt -fno-use-cxa-get-exception-ptr \
-fno-implicit-templates -fno-threadsafe-statics -fno-use-cxa-atexit
ASFLAGS=-f elf -I $(LDRDIR)/
CFLAGS=-Wall -Wextra -Wpedantic -std=c99 $(COPTFLAGS) -nostartfiles -fno-PIC
CXXFLAGS=-Wall -Wextra -Wpedantic -std=c++11 $(CXXOPTFLAGS) -nostartfiles -fno-PIC
LIBS=-lGL -ldl -lc
LDFLAGS=--oformat=binary -T ldr/link.ld
.PHONY: all
all: $(BINDIR)/test
.PHONY: clean
clean:
rm -rf $(OBJDIR)/* $(BINDIR)/*
.SECONDARY:
$(OBJDIR)/%.o: $(SRCDIR)/%.c
$(CC) $(CFLAGS) -c $^ -o $@
$(OBJDIR)/%.o.syms: $(OBJDIR)/%.o
readelf -s $^ | grep UND | sed 1d | awk '{ print $$8 }' > $@
$(OBJDIR)/symbols.%.s: $(OBJDIR)/%.o.syms
$(LDRDIR)/mksyms $(LIBS) $$(cat $^) > $@
$(OBJDIR)/header.%.o: $(OBJDIR)/symbols.%.s $(LDRDIR)/header.s $(LDRDIR)/loader.s
nasm $(ASFLAGS) $< -o $@
$(BINDIR)/%: $(OBJDIR)/%.o $(OBJDIR)/header.%.o
$(LD) $(LDFLAGS) $^ -o $@

78
ldr/header.s Normal file
View File

@ -0,0 +1,78 @@
%define ORIGIN 0x400000
extern _size
[section .header]
header:
; e_ident
db 0x7F, "ELF" ; EI_MAG0-EI_MAG3
db 1 ; EI_CLASS: 1 = 32-bit
db 1 ; EI_DATA: 1 = LSB
db 1 ; EI_VERSION
db 3 ; EI_OSABI: 3 = Linux
db 1 ; EI_OSABIVERSION
times 7 db 0 ; EI_PAD, ld.so is a busta and won't let us use our leet group tags for padding bytes :(
; e_type: 2 = executable
dw 2
; e_machine: 3 = x86
dw 3
; e_version
dd 1
; e_entry
dd _start
; e_phoff
dd (.segments - header)
; e_shoff
dd 0
; e_flags
dd 0
; e_ehsize
dw (.segments - header)
; e_phentsize
dw (.segments.load - .segments.dynamic)
.segments:
.segments.dynamic:
; {e_phnum: 2, e_shentsize: 0}, p_type: 2 = PT_DYNAMIC
dd 2
; {e_shnum: <junk>, e_shstrnd: <junk>}, p_offset
dd (.dynamic - header)
; p_vaddr
dd .dynamic
; p_paddr
dd 0
; p_filesz
dd (.dynamic.end - .dynamic)
; p_memsz
dd (.dynamic.end - .dynamic)
; p_flags, p_align
dq 0
.segments.load:
; p_type: 1 = PT_LOAD
dd 1
; p_offset
dd 0
; p_vaddr
dd ORIGIN
; p_paddr
dd 0
; p_filesz
dd _size
; p_memsz
dd _size
; p_flags: 1 = execute, 4 = read
dd (1 | 2 | 4)
; p_align
dd 0x1000
.segments.end:
.dynamic:
.dynamic.strtab:
; d_tag: 5 = DT_STRTAB
dd 5
; d_un.d_ptr
dd _symbols
.dynamic.symtab:
; this is required to be present or ld.so will crash, but it can be bogus
; d_tag: 6 = DT_SYMTAB
dd 6
; d_un.d_ptr
dd 0

24
ldr/link.ld Normal file
View File

@ -0,0 +1,24 @@
OUTPUT_FORMAT(binary)
SECTIONS {
. = 0x400000;
.header : { *(.header) }
.text : {
*(.text .text.* .rdata .rdata.* .rodata .rodata.*)
}
.data : {
*(.data .data.* .bss .bss.* .tdata .tdata.* .tbss .tbss.*)
}
.dynamic : { *(.dynamic) } :all :dyn
.dynstuff : { *(.symtab .strtab .shstrtab .rel.text .got.plt .gnu.linkonce.* .plt .plt.got .interp) } :all
/DISCARD/ : {
*(.*)
}
_size = . - 0x400000;
}

130
ldr/loader.s Normal file
View File

@ -0,0 +1,130 @@
[section .text]
%define LM_NAME_OFFSET 0x4
%define LM_NEXT_OFFSET 0xC
%define LM_ADDR_OFFSET 0
%define LM_INFO_OFFSET 0x20
%define LM_NBUCKETS_OFFSET 0x178
%define LM_GNU_BUCKETS_OFFSET 0x188
%define LM_GNU_CHAIN_ZERO_OFFSET 0x18C
%define DT_VALUE_OFFSET 0x4
%define DYN_PTR_OFFSET 0x4
%define DT_SYMTAB 0x6
%define DT_SYMSIZE_SHIFT 4
strcmp: ; (const char *s1 (esi), const char *s2 (edi))
push esi
push edi
.cmp: lodsb
or al, al
jz .done
sub al, [edi]
jnz .done
inc edi
jmp .cmp
.done: pop edi
pop esi
ret
basename: ; (const char *s (esi))
push esi
push edi
mov edi, esi
.cmp: lodsb
or al, al
jz .done
cmp al, 47 ; '/'
cmove edi, esi
jmp .cmp
.done: mov eax, edi
pop edi
pop esi
ret
link_symbol: ; (struct link_map *entry, uint32_t *h)
mov ecx, esi
; eax = *h % entry->l_nbuckets
mov eax, [ecx]
xor edx, edx
mov ebx, [ebp + LM_NBUCKETS_OFFSET]
div ebx
; eax = entry->l_gnu_buckets[eax]
mov eax, [ebp + LM_GNU_BUCKETS_OFFSET]
mov eax, [eax + edx * 4]
; *h |= 1
or word [ecx], 1
.check_bucket: ; edx = entry->l_gnu_chain_zero[eax] | 1
mov edx, [ebp + LM_GNU_CHAIN_ZERO_OFFSET]
mov edx, [edx + eax * 4]
or edx, 1
; check if this is our symbol
cmp edx, [ecx]
je .found
inc eax
jmp .check_bucket
.found: ; it is! edx = entry->l_info[DT_SYMTAB]->d_un.d_ptr
mov edx, [ebp + LM_INFO_OFFSET + DT_SYMTAB * 4]
mov edx, [edx + DYN_PTR_OFFSET]
; edx = edx[eax].dt_value + entry->l_addr
shl eax, DT_SYMSIZE_SHIFT
mov edx, [edx + eax + DT_VALUE_OFFSET]
add edx, [ebp + LM_ADDR_OFFSET]
sub edx, ecx
sub edx, 4
; finally, write it back!
mov [ecx], edx
ret
link: ; (struct link_map *root, char *symtable)
mov eax, [esp+4]
mov esi, [esp+8]
.do_library: ; null library name means end of symbol table, we're done
cmp byte [esi], 0
jz .done
; setup start of map again
mov ebp, eax
push eax
.find_map_entry: ; compare basename(entry->l_name) to lib name, if so we got a match
push esi
mov esi, [ebp + LM_NAME_OFFSET]
call basename
mov edi, eax
pop esi
call strcmp
jz .process_map_entry
; no match, next entry it is!
mov ebp, [ebp + LM_NEXT_OFFSET]
jmp .find_map_entry
.process_map_entry: ; skip past the name in the symbol table now to get to the symbols
lodsb
or al, al
jnz .process_map_entry
.do_symbols: ; null byte means end of symbols for this library!
cmp byte [esi], 0
jz .next_library
inc esi
call link_symbol
add esi, 4
jmp .do_symbols
.next_library: pop eax
inc esi
jmp .do_library
.done: ret
extern main
_start:
push _symbols
push eax
call link
call main
int3

80
ldr/mksyms Executable file
View File

@ -0,0 +1,80 @@
#!/usr/bin/env python3
import sys
import os.path
import subprocess
import itertools
import argparse
def hash_djb2(s):
h = 5381
for c in s:
h = (h * 33 + ord(c)) & 0xFFFFFFFF
return h
def output_x86(libraries):
shorts = { l: l.split('.', 1)[0].lower() for l in libraries }
print('%include "header.s"')
print('.dynamic.needed:')
for library in libraries:
print('dd 1')
print('dd (_symbols.{} - _symbols)'.format(shorts[library]))
print('.dynamic.end:')
print('_symbols:')
for library, symbols in libraries.items():
print('\t_symbols.{}: db "{}",0'.format(shorts[library], library))
for sym in symbols:
hash = hash_djb2(sym)
print("""
\t\tglobal {name}
\t\t{name}: db 0xE9
\t\t dd 0x{hash:x}
""".strip('\n').format(name=sym, hash=hash))
print('\tdb 0')
print('db 0')
print('%include "loader.s"')
def get_cc_paths():
output = subprocess.check_output(['cc', '-print-search-dirs'], stderr=subprocess.DEVNULL)
paths = {}
for entry in output.decode('utf-8').splitlines():
category, path = entry.split(': ', 1)
path = path.lstrip('=')
paths[category] = list(set(os.path.realpath(p) for p in path.split(':') if os.path.isdir(p)))
return paths
def find_symbol(paths, libraries, symbol):
output = subprocess.check_output(['scanelf', '-B', '-F' '%s %S', '-s', '+{}'.format(symbol)] + paths, stderr=subprocess.DEVNULL)
for entry in output.decode('utf-8').splitlines():
sym, soname, path = entry.split(' ', 2)
if symbol in sym.split(',') and any(soname.startswith(l) for l in libraries):
return soname
def main():
parser = argparse.ArgumentParser()
parser.add_argument('-a', '--architecture', default='x86', help='architecture to generate asm code for')
parser.add_argument('-l', '--library', metavar='LIB', default=['c'], action='append', help='libraries to link against')
parser.add_argument('symbols', metavar='SYM', nargs='+', help='symbol to search for')
args = parser.parse_args()
paths = get_cc_paths()
symbols = {}
libraries = ['lib{}.so'.format(l) for l in args.library]
for symbol in args.symbols:
library = find_symbol(paths['libraries'], libraries, symbol)
if not library:
sys.stderr.write('could not find symbol: {}\n'.format(symbol))
sys.exit(1)
symbols.setdefault(library, [])
symbols[library].append(symbol)
output_x86(symbols)
if __name__ == '__main__':
main()

68
smol Executable file
View File

@ -0,0 +1,68 @@
#!/bin/sh
set -e
COPTFLAGS=$(cat <<'EOF'
-Os -fvisibility=hidden -mpreferred-stack-boundary=2 -fwhole-program
-ffast-math -funsafe-math-optimizations -fno-stack-protector -fomit-frame-pointer
-fno-exceptions -fno-unwind-tables -fno-asynchronous-unwind-tables
EOF
)
CXXOPTFLAGS=$(cat <<EOF
$CXXOPTFLAGS
-fno-rtti -fno-enforce-eh-specs -fnothrow-opt -fno-use-cxa-get-exception-ptr
-fno-implicit-templates -fno-threadsafe-statics -fno-use-cxa-atexit
EOF
)
CC="${CC:-cc}"
CXX="${CXX:-c++}"
CFLAGS="-Wall -Wextra -Wpedantic -std=c99 $COPTFLAGS -nostartfiles -fno-PIC"
CXXFLAGS="-Wall -Wextra -Wpedantic -std=c++11 $CXXOPTFLAGS -nostartfiles -fno-PIC"
ASFLAGS="-f elf -I ldr/"
LDFLAGS="--oformat=binary -T ldr/link.ld"
mkdir -p bin
rm -rf obj
mkdir -p obj
cleared=
mksym_args=
files=
symfiles=
output=s.out
i=1
add_out() {
out="obj/$i"
files="$files $out"
i=$((i+1))
}
add_syms() {
readelf -s "$1" | grep UND | sed 1d | awk '{ print $8 }' > $1.syms
symfiles="$symfiles $1.syms"
}
for a; do
if test -z "$cleared"; then
set --; cleared=1
fi
case "$a" in
-o) output="$a";;
-l*) mksym_args="$mksym_args $a";;
-*) set -- "$@" "$a";;
*.c)
add_out
$CC $CFLAGS "$@" -c "$a" -o "$out"
add_syms "$out";;
*.cxx|*.cc|*.cpp)
add_out
$CXX $CXXFLAGS "$@" -c "$a" -o "$out"
add_syms "$out";;
*) printf "not sure what to do with input file: %s, bailing\n" "$a" &>2; exit 1;;
esac
done
./ldr/mksyms $mksym_args $(cat $symfiles) > obj/$output.syms.s
nasm $ASFLAGS obj/$output.syms.s -o obj/$output.header.o
ld $LDFLAGS obj/$output.header.o $files -o bin/$output

9
src/hello.c Normal file
View File

@ -0,0 +1,9 @@
#include <stdlib.h>
#include <stdio.h>
const char *f = "foo";
int main(void) {
printf("hello world %s\n", f);
exit(42);
}

9
src/sdl.c Normal file
View File

@ -0,0 +1,9 @@
#include <SDL2/SDL.h>
int main(void) {
SDL_Init(SDL_INIT_VIDEO);
SDL_Window *w = SDL_CreateWindow("nice", SDL_WINDOWPOS_UNDEFINED, SDL_WINDOWPOS_UNDEFINED, 1280, 720, SDL_WINDOW_OPENGL);
SDL_Delay(3000);
SDL_DestroyWindow(w);
return 0;
}