gcc常用的編譯選項對程式碼的影響(轉)

ba發表於2007-08-12
gcc常用的編譯選項對程式碼的影響(轉)[@more@]測試環境 redhat 6.2

★ 前言

本文討論gcc的一些常用編譯選項對程式碼的影響。當然程式碼變了,
它的記憶體佈局也就會變了,隨之exploit也就要做相應的變動。
gcc的編譯選項實在太多,本文檢了幾個最常用的選項。


★ 演示程式

[alert7@redhat62 alert7]$ cat > test.c
#include
void hi(void)
{
printf("hi");
}

int main(int argc, char *argv[])
{
hi();
return 0;
}


★ 一般情況

[alert7@redhat62 alert7]$ gcc -o test test.c
[alert7@redhat62 alert7]$ wc -c test
11773 test
[alert7@redhat62 alert7]$ gdb -q test
(gdb) disass main
Dump of assembler code for function main:
0x80483e4 : push %ebp
0x80483e5 : mov %esp,%ebp
0x80483e7 : call 0x80483d0
0x80483ec : xor %eax,%eax
0x80483ee : jmp 0x80483f0
0x80483f0 : leave
0x80483f1 : ret
....
End of assembler dump.
(gdb) disass hi
Dump of assembler code for function hi:
0x80483d0 : push %ebp
0x80483d1 : mov %esp,%ebp
0x80483d3 : push $0x8048450
0x80483d8 : call 0x8048308
0x80483dd : add $0x4,%esp
0x80483e0 : leave
0x80483e1 : ret
0x80483e2 : mov %esi,%esi
End of assembler dump.

來看看部分的記憶體映象
(記憶體高址)
+--------+
|bffffbc4| argv的地址(即argv[0]的地址)
0xbffffb84 +--------+
|00000001| argc的值
0xbffffb80 +--------+
|400309cb|main的返回地址
0xbffffb7c +--------+ |bffffb98| 呼叫main函式前的ebp
0xbffffb78 +--------+ |080483ec| hi()的返回地址
0xbffffb74 +--------+
|bffffb78| 呼叫hi()前的esp
0xbffffb70 +--------+
|08048450| "hi"的地址
0xbffffb6c +--------+
| ...... |
(記憶體低址)

leave 指令所做的操作相當於MOV ESP,EBP 然後 POP EBP
ret 指令所做的操作相當於POP EIP


★ -O 編譯選項

With `-O', the compiler tries to reduce code size and execution time.
When you specify `-O', the two options `-fthread-jumps' and
`-fdefer-pop' are turned on
最佳化,減少程式碼大小和執行的時間

[alert7@redhat62 alert7]$ gcc -O -o test test.c
[alert7@redhat62 alert7]$ wc -c test
11757 test
[alert7@redhat62 alert7]$ gdb -q test
(gdb) disass main
Dump of assembler code for function main:
0x80483d8 : push %ebp
0x80483d9 : mov %esp,%ebp
0x80483db : call 0x80483c8
0x80483e0 : xor %eax,%eax
0x80483e2 : leave
0x80483e3 : ret
0x80483e4 : nop
...
End of assembler dump.
(gdb) disass hi
Dump of assembler code for function hi:
0x80483c8 : push %ebp
0x80483c9 : mov %esp,%ebp
0x80483cb : push $0x8048440
0x80483d0 : call 0x8048308
0x80483d5 : leave
0x80483d6 : ret
0x80483d7 : nop
End of assembler dump.

在main()中,把一條jmp指令最佳化掉了,很顯然,這條指令是可以不需要的。
在hi()中,把add $0x4,%esp最佳化掉了,這會不會使stack不平衡呢?
來看看部分的記憶體映象
(記憶體高址)
+--------+
|bffffbc4| argv的地址(即argv[0]的地址)
0xbffffb84 +--------+
|00000001| argc的值
0xbffffb80 +--------+
|400309cb|main的返回地址
0xbffffb7c +--------+ |bffffb98| 呼叫main函式前的ebp
0xbffffb78 +--------+ |080483e0| hi()的返回地址
0xbffffb74 +--------+
|bffffb78| 呼叫hi()前的esp
0xbffffb70 +--------+
|08048440| "hi"的地址
0xbffffb6c +--------+
| ...... |
(記憶體低址)

leave 指令所做的操作相當於把MOV ESP,EBP 然後 POP EBP
看到leave指令操作了沒有,先把ebp--&gtesp,再pop ebp,這樣即使
在過程內堆疊的esp,ebp是不平衡的,但只要返回時候碰到leave指令
就會平衡了,所以把add $0x4,%esp最佳化掉也是沒有問題的。


★ -O2 編譯選項

-O2 Optimize even more. Nearly all supported optimizations that do
not involve a space-speed tradeoff are performed. Loop unrolling
and function inlining are not done, for example. As compared to -O,
this option increases both compilation time and the performance of
the generated code.

[alert7@redhat62 alert7]$ gcc -O2 -o test test.c
[alert7@redhat62 alert7]$ wc -c test
11757 test
[alert7@redhat62 alert7]$ gdb -q test
(gdb) disass main
Dump of assembler code for function main:
0x80483d8 : push %ebp
0x80483d9 : mov %esp,%ebp
0x80483db : call 0x80483c8
0x80483e0 : xor %eax,%eax
0x80483e2 : leave
0x80483e3 : ret
...
0x80483ef : nop
End of assembler dump.
(gdb) disass hi
Dump of assembler code for function hi:
0x80483c8 : push %ebp
0x80483c9 : mov %esp,%ebp
0x80483cb : push $0x8048440
0x80483d0 : call 0x8048308
0x80483d5 : leave
0x80483d6 : ret
0x80483d7 : nop
End of assembler dump.

由於程式比較簡單,再最佳化也沒有好最佳化的了,所以跟-O出來的一樣。


★ -fomit-frame-pointer 編譯選項

-fomit-frame-pointer
Don't keep the frame pointer in a register for functions
that don't need one. This avoids the instructions to save,
set up and restore frame pointers; it also makes an extra
register available in many functions. It also makes
debugging impossible on most machines.

忽略幀指標。這樣在程式就不需要儲存,安裝,和恢復ebp了。這樣ebp也就是一個
free的register了,在函式中就可以隨便使用了。

[alert7@redhat62 alert7]$ gcc -fomit-frame-pointer -o test test.c
[alert7@redhat62 alert7]$ wc -c test
11773 test
[alert7@redhat62 alert7]$ gdb -q test
(gdb) disass main
Dump of assembler code for function main:
0x80483e0 : call 0x80483d0
0x80483e5 : xor %eax,%eax
0x80483e7 : jmp 0x80483f0
0x80483e9 : lea 0x0(%esi,1),%esi
0x80483f0 : ret
....
End of assembler dump.
(gdb) disass hi
Dump of assembler code for function hi:
0x80483d0 : push $0x8048450
0x80483d5 : call 0x8048308
0x80483da : add $0x4,%esp
0x80483dd : ret
0x80483de : mov %esi,%esi
End of assembler dump.

在main()和hi()中都去掉了以下指令
push %ebp
mov %esp,%ebp//這兩條指令安裝
leave//這條指令恢復
來看看部分的記憶體映象
(記憶體高址)
+--------+
|bffffbc4| argv的地址(即argv[0]的地址)
0xbffffb84 +--------+
|00000001| argc的值
0xbffffb80 +--------+
|400309cb|main的返回地址
0xbffffb7c +--------+
|080483e5| hi()的返回地址
0xbffffb78 +--------+
|08048450| "hi"字串的地址
0xbffffb74 +--------+
| ...... |
(記憶體低址)
沒有儲存上層執行環境的ebp.


★ -fomit-frame-pointer && -O2

-fomit-frame-pointer編譯選項去掉了
push %ebp
mov %esp,%ebp//這兩條指令安裝
leave//這條指令恢復
-O2編譯選項去掉了
add $0x4,%esp

兩個加起來會不會這四條指令一起去掉,從而使stack不平衡呢?

[alert7@redhat62 alert7]$ gcc -fomit-frame-pointer -O2 -o test test.c
[alert7@redhat62 alert7]$ wc -c test
11741 test
[alert7@redhat62 alert7]$ gdb -q test
(gdb) disass main
Dump of assembler code for function main:
0x80483d8 : call 0x80483c8
0x80483dd : xor %eax,%eax
0x80483df : ret
End of assembler dump.
(gdb) disass hi
Dump of assembler code for function hi:
0x80483c8 : push $0x8048430
0x80483cd : call 0x8048308
0x80483d2 : add $0x4,%esp
0x80483d5 : ret
0x80483d6 : mov %esi,%esi
End of assembler dump.
來看看部分的記憶體映象
(記憶體高址)
+--------+
|bffffbc4| argv的地址(即argv[0]的地址)
0xbffffb84 +--------+
|00000001| argc的值
0xbffffb80 +--------+
|400309cb|main的返回地址
0xbffffb7c +--------+
|080483dd| hi()的返回地址
0xbffffb78 +--------+
|08048430| "hi"字串的地址
0xbffffb74 +--------+
| ...... |
(記憶體低址)

此時就沒有把add $0x4,%esp最佳化掉,如果最佳化掉的話,整個stack就
會變的不平衡,從而會導致程式出錯。


★ -fPIC 編譯選項

-fPIC If supported for the target machine, emit position-independent
code, suitable for dynamic linking,even if branches need large
displacements.
產生位置無關程式碼(PIC),一般建立共享庫時用到。
在x86上,PIC的程式碼的符號引用都是透過ebx進行操作的。

[alert7@redhat62 alert7]$ gcc -fPIC -o test test.c
[alert7@redhat62 alert7]$ wc -c test
11805 test
[alert7@redhat62 alert7]$ gdb -q test
(gdb) disass main
Dump of assembler code for function main:
0x80483f8 : push %ebp
0x80483f9 : mov %esp,%ebp
0x80483fb : push %ebx
0x80483fc : call 0x8048401
0x8048401 : pop %ebx//取得該指令的地址
0x8048402 : add $0x1093,%ebx//此時ebx裡面存放著是GOT表的地址
0x8048408 : call 0x80483d0
0x804840d : xor %eax,%eax
0x804840f : jmp 0x8048411
0x8048411 : mov 0xfffffffc(%ebp),%ebx
0x8048414 : leave
0x8048415 : ret
...
End of assembler dump.
(gdb) disass hi
Dump of assembler code for function hi:
0x80483d0 : push %ebp
0x80483d1 : mov %esp,%ebp
0x80483d3 : push %ebx
0x80483d4 : call 0x80483d9
0x80483d9 : pop %ebx
0x80483da : add $0x10bb,%ebx
0x80483e0 : lea 0xffffefdc(%ebx),%edx
0x80483e6 : mov %edx,%eax
0x80483e8 : push %eax
0x80483e9 : call 0x8048308
0x80483ee : add $0x4,%esp
0x80483f1 : mov 0xfffffffc(%ebp),%ebx
0x80483f4 : leave
0x80483f5 : ret
0x80483f6 : mov %esi,%esi
End of assembler dump.
來看看部分的記憶體映象

(記憶體高址)
+--------+
|bffffbc4| argv的地址(即argv[0]的地址)
0xbffffb84 +--------+
|00000001| argc的值
0xbffffb80 +--------+
|400309cb|main的返回地址
0xbffffb7c +--------+ |bffffb98| 呼叫main函式前的ebp
0xbffffb78 +--------+ |401081ec| 儲存的ebx
0xbffffb74 +--------+
|0804840d| (存放過call 0x8048401的下一條指令地址)
0xbffffb70 +--------+
|bffffb78| 呼叫hi()前的esp
0xbffffb6c +--------+
|08049494| GOT表地址
0xbffffb68 +--------+
|08048470|(存放過call 0x80483d9的下一條指令地址)
0xbffffb64 +--------+
| ...... |
(記憶體低址)


★ -static 編譯選項

-static
On systems that support dynamic linking, this prevents
linking with the shared libraries. On other systems,
this option has no effect.
把一些函式都靜態的編譯到程式中,而無需動態連結了。

[alert7@redhat62 alert7]$ gcc -o test -static test.c
[alert7@redhat62 alert7]$ wc -c test
962808 test
[alert7@redhat62 alert7]$ gdb -q test
(gdb) disass main
Dump of assembler code for function main:
0x80481b4 : push %ebp
0x80481b5 : mov %esp,%ebp
0x80481b7 : call 0x80481a0
0x80481bc : xor %eax,%eax
0x80481be : jmp 0x80481c0
0x80481c0 : leave
0x80481c1 : ret
...
End of assembler dump.
(gdb) disass hi
Dump of assembler code for function hi:
0x80481a0 : push %ebp
0x80481a1 : mov %esp,%ebp
0x80481a3 : push $0x8071528
0x80481a8 : call 0x804865c
0x80481ad : add $0x4,%esp
0x80481b0 : leave
0x80481b1 : ret
0x80481b2 : mov %esi,%esi
End of assembler dump.
[alert7@redhat62 alert7]$ ldd test
not a dynamic executable
-static出來的程式碼已經沒有PLT了,GOT雖然有,已經全部為0了。

來自 “ ITPUB部落格 ” ,連結:http://blog.itpub.net/10617731/viewspace-950495/,如需轉載,請註明出處,否則將追究法律責任。

相關文章