#干了這碗雞湯
生活總是這樣,不能叫人處處都滿意。但我們還要熱情地活下去。人活一生,值得愛的東西很多,不要因為一個不滿意,就灰心。
—— 路遙《平凡的世界》
在Linux中有一個命令我們平時肯定用過,它就是strip,
咳咳,跑題了,不是這個strip。
通過strip可以移除目標文件的符號信息,可以減少目標文件的體積,這里有幾個問題:
什么是符號?
如何使用strip?
strip的作用是什么?
動態(tài)鏈接庫如果被strip后還能被鏈接成功嗎?
靜態(tài)鏈接庫如果被strip后還能被鏈接成功嗎?
什么是符號?
符號可以看作是鏈接中的粘合劑,整個鏈接過程需要基于符號才可以正確完成。鏈接過程的本質(zhì)就是把多個不同的目標文件相互粘到一起,像積木一樣各有凹凸部分,但還是可以拼接成一個整體,這個將多個目標文件粘到一起的東西就是符號??梢詫⒑瘮?shù)和變量統(tǒng)稱為符號,而函數(shù)名和變量名統(tǒng)稱為符號名。
在Linux中可以通過一些命令來查看符號信息:
nm命令:
nm test.o U _GLOBAL_OFFSET_TABLE_0000000000000000 T mainU puts
objdump命令:
objdump -t test.o
file format elf64-x86-64 :
SYMBOL TABLE:0000000000000000 l df *ABS* 0000000000000000 test_c.cc0000000000000000 l d .text 0000000000000000 .text0000000000000000 l d .data 0000000000000000 .data0000000000000000 l d .bss 0000000000000000 .bss0000000000000000 l d .rodata 0000000000000000 .rodata0000000000000000 l d .note.GNU-stack 0000000000000000 .note.GNU-stack0000000000000000 l d .eh_frame 0000000000000000 .eh_frame0000000000000000 l d .comment 0000000000000000 .comment0000000000000000 g F .text 0000000000000017 main0000000000000000 *UND* 0000000000000000 _GLOBAL_OFFSET_TABLE_0000000000000000 *UND* 0000000000000000 puts
readelf命令:
readelf -s test.o
Symbol table '.symtab' contains 12 entries: Num: Value Size Type Bind Vis Ndx Name0: 0000000000000000 0 NOTYPE LOCAL DEFAULT UND1: 0000000000000000 0 FILE LOCAL DEFAULT ABS test_c.cc2: 0000000000000000 0 SECTION LOCAL DEFAULT 13: 0000000000000000 0 SECTION LOCAL DEFAULT 34: 0000000000000000 0 SECTION LOCAL DEFAULT 45: 0000000000000000 0 SECTION LOCAL DEFAULT 56: 0000000000000000 0 SECTION LOCAL DEFAULT 77: 0000000000000000 0 SECTION LOCAL DEFAULT 88: 0000000000000000 0 SECTION LOCAL DEFAULT 69: 0000000000000000 23 FUNC GLOBAL DEFAULT 1 main10: 0000000000000000 0 NOTYPE GLOBAL DEFAULT UND _GLOBAL_OFFSET_TABLE_11: 0000000000000000 0 NOTYPE GLOBAL DEFAULT UND puts
如何使用strip?
在Linux中可以使用man strip查看strip使用方法,最主要的就是移除所有符號的-s參數(shù),用于清除所有的符號信息:
strip -s xxx
在使用strip之前先使用nm查看下可執(zhí)行程序的符號信息:
nm a.out 0000000000200da0 d _DYNAMIC0000000000200fa0 d _GLOBAL_OFFSET_TABLE_000000000000089b t _GLOBAL__sub_I__Z4funcPc0000000000000930 R _IO_stdin_usedw _ITM_deregisterTMCloneTablew _ITM_registerTMCloneTable0000000000000852 t _Z41__static_initialization_and_destruction_0ii00000000000007fa T _Z4funcPc000000000000081c T _Z4funciU _ZNSt8ios_base4InitC1Ev@@GLIBCXX_3.4U _ZNSt8ios_base4InitD1Ev@@GLIBCXX_3.40000000000201020 B _ZSt4cout@@GLIBCXX_3.40000000000000934 r _ZStL19piecewise_construct0000000000201131 b _ZStL8__ioinitU _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@@GLIBCXX_3.40000000000000b24 r __FRAME_END__0000000000000940 r __GNU_EH_FRAME_HDR0000000000201010 D __TMC_END__0000000000201010 B __bss_startU __cxa_atexit@@GLIBC_2.2.5w __cxa_finalize@@GLIBC_2.2.50000000000201000 D __data_start00000000000007b0 t __do_global_dtors_aux0000000000200d98 t __do_global_dtors_aux_fini_array_entry0000000000201008 D __dso_handle0000000000200d88 t __frame_dummy_init_array_entryw __gmon_start__0000000000200d98 t __init_array_end0000000000200d88 t __init_array_start0000000000000920 T __libc_csu_fini00000000000008b0 T __libc_csu_initU __libc_start_main@@GLIBC_2.2.50000000000201010 D _edata0000000000201138 B _end0000000000000924 T _fini0000000000000688 T _init00000000000006f0 T _start0000000000201130 b completed.76980000000000201000 W data_start0000000000000720 t deregister_tm_clones00000000000007f0 t frame_dummy000000000000083d T main0000000000000760 t register_tm_clones
當前這個可執(zhí)行程序的文件大小是8840字節(jié):
-rwxrwxrwx 1 a a 8840 Nov 29 14:54 a.out
使用strip清除符號信息:
~/test$ strip -s a.out
strip后再查看可執(zhí)行文件的符號信息:
~/test$ nm a.outnm: a.out: no symbols
發(fā)現(xiàn)什么符號都沒有了,但還是可以執(zhí)行。
strip后的可執(zhí)行程序文件大小是6120字節(jié):
-rwxrwxrwx 1 a a 6120 Nov 29 14:54 a.out
由此可見通過strip我們可以減少程序的體積。
strip的作用是什么?
前面已經(jīng)大體介紹過,strip最大的作用就是可以減少程序的體積,一般公司對發(fā)布的程序體積要求是極其嚴格的,strip命令是減少程序體積的一個很有效的方法。另一個作用就是提高了安全性,沒有了這些符號,別人分析沒有符號的程序會變得更加困難。
動態(tài)鏈接庫如果被strip后還能被鏈接成功嗎?
先說答案,可以。
先貼出兩段代碼:
// shared.cc
void Print(int a) { std::cout << "Hello World " << a << std::endl; }
// main.cc
void Print(int a);
int main() { Print(666); return 0;}
將shared.cc編成一個動態(tài)鏈接庫:
g++ shared.cc -o shared.so -shared -fPIC
使用readelf查看鏈接庫的符號信息:
readelf -S shared.so There are 28 section headers, starting at offset 0x1aa0:
Section Headers:Name Type Address Offset Size EntSize Flags Link Info Align0] NULL 0000000000000000 00000000 0000000000000000 0000000000000000 0 0 01] .note.gnu.build-i NOTE 00000000000001c8 000001c8 0000000000000024 0000000000000000 A 0 0 42] .gnu.hash GNU_HASH 00000000000001f0 000001f0 000000000000003c 0000000000000000 A 3 0 83] .dynsym DYNSYM 0000000000000230 00000230 00000000000001c8 0000000000000018 A 4 1 84] .dynstr STRTAB 00000000000003f8 000003f8 0000000000000189 0000000000000000 A 0 0 15] .gnu.version VERSYM 0000000000000582 00000582 0000000000000026 0000000000000002 A 3 0 26] .gnu.version_r VERNEED 00000000000005a8 000005a8 0000000000000040 0000000000000000 A 4 2 87] .rela.dyn RELA 00000000000005e8 000005e8 0000000000000108 0000000000000018 A 3 0 88] .rela.plt RELA 00000000000006f0 000006f0 0000000000000078 0000000000000018 AI 3 21 89] .init PROGBITS 0000000000000768 00000768 0000000000000017 0000000000000000 AX 0 0 4.plt PROGBITS 0000000000000780 00000780 0000000000000060 0000000000000010 AX 0 0 16.plt.got PROGBITS 00000000000007e0 000007e0 0000000000000008 0000000000000008 AX 0 0 8.text PROGBITS 00000000000007f0 000007f0 0000000000000181 0000000000000000 AX 0 0 16.fini PROGBITS 0000000000000974 00000974 0000000000000009 0000000000000000 AX 0 0 4.rodata PROGBITS 000000000000097d 0000097d 000000000000000e 0000000000000000 A 0 0 1.eh_frame_hdr PROGBITS 000000000000098c 0000098c 0000000000000034 0000000000000000 A 0 0 4.eh_frame PROGBITS 00000000000009c0 000009c0 00000000000000bc 0000000000000000 A 0 0 8.init_array INIT_ARRAY 0000000000200de0 00000de0 0000000000000010 0000000000000008 WA 0 0 8.fini_array FINI_ARRAY 0000000000200df0 00000df0 0000000000000008 0000000000000008 WA 0 0 8.dynamic DYNAMIC 0000000000200df8 00000df8 00000000000001d0 0000000000000010 WA 4 0 8.got PROGBITS 0000000000200fc8 00000fc8 0000000000000038 0000000000000008 WA 0 0 8.got.plt PROGBITS 0000000000201000 00001000 0000000000000040 0000000000000008 WA 0 0 8.data PROGBITS 0000000000201040 00001040 0000000000000008 0000000000000000 WA 0 0 8.bss NOBITS 0000000000201048 00001048 0000000000000008 0000000000000000 WA 0 0 1.comment PROGBITS 0000000000000000 00001048 0000000000000029 0000000000000001 MS 0 0 1.symtab SYMTAB 0000000000000000 00001078 0000000000000600 0000000000000018 26 46 8.strtab STRTAB 0000000000000000 00001678 0000000000000330 0000000000000000 0 0 1.shstrtab STRTAB 0000000000000000 000019a8 00000000000000f1 0000000000000000 0 0 1Key to Flags:W (write), A (alloc), X (execute), M (merge), S (strings), I (info),L (link order), O (extra OS processing required), G (group), T (TLS),C (compressed), x (unknown), o (OS specific), E (exclude),l (large), p (processor specific)
注意這里有28個符號段,主要有symtab、strtab、dynsym、dynstr段。
strip后再看下符號信息:
readelf -S shared.so There are 26 section headers, starting at offset 0x1158:
Section Headers:Name Type Address Offset Size EntSize Flags Link Info Align0] NULL 0000000000000000 00000000 0000000000000000 0000000000000000 0 0 01] .note.gnu.build-i NOTE 00000000000001c8 000001c8 0000000000000024 0000000000000000 A 0 0 42] .gnu.hash GNU_HASH 00000000000001f0 000001f0 000000000000003c 0000000000000000 A 3 0 83] .dynsym DYNSYM 0000000000000230 00000230 00000000000001c8 0000000000000018 A 4 1 84] .dynstr STRTAB 00000000000003f8 000003f8 0000000000000189 0000000000000000 A 0 0 15] .gnu.version VERSYM 0000000000000582 00000582 0000000000000026 0000000000000002 A 3 0 26] .gnu.version_r VERNEED 00000000000005a8 000005a8 0000000000000040 0000000000000000 A 4 2 87] .rela.dyn RELA 00000000000005e8 000005e8 0000000000000108 0000000000000018 A 3 0 88] .rela.plt RELA 00000000000006f0 000006f0 0000000000000078 0000000000000018 AI 3 21 89] .init PROGBITS 0000000000000768 00000768 0000000000000017 0000000000000000 AX 0 0 4.plt PROGBITS 0000000000000780 00000780 0000000000000060 0000000000000010 AX 0 0 16.plt.got PROGBITS 00000000000007e0 000007e0 0000000000000008 0000000000000008 AX 0 0 8.text PROGBITS 00000000000007f0 000007f0 0000000000000181 0000000000000000 AX 0 0 16.fini PROGBITS 0000000000000974 00000974 0000000000000009 0000000000000000 AX 0 0 4.rodata PROGBITS 000000000000097d 0000097d 000000000000000e 0000000000000000 A 0 0 1.eh_frame_hdr PROGBITS 000000000000098c 0000098c 0000000000000034 0000000000000000 A 0 0 4.eh_frame PROGBITS 00000000000009c0 000009c0 00000000000000bc 0000000000000000 A 0 0 8.init_array INIT_ARRAY 0000000000200de0 00000de0 0000000000000010 0000000000000008 WA 0 0 8.fini_array FINI_ARRAY 0000000000200df0 00000df0 0000000000000008 0000000000000008 WA 0 0 8.dynamic DYNAMIC 0000000000200df8 00000df8 00000000000001d0 0000000000000010 WA 4 0 8.got PROGBITS 0000000000200fc8 00000fc8 0000000000000038 0000000000000008 WA 0 0 8.got.plt PROGBITS 0000000000201000 00001000 0000000000000040 0000000000000008 WA 0 0 8.data PROGBITS 0000000000201040 00001040 0000000000000008 0000000000000000 WA 0 0 8.bss NOBITS 0000000000201048 00001048 0000000000000008 0000000000000000 WA 0 0 1.comment PROGBITS 0000000000000000 00001048 0000000000000029 0000000000000001 MS 0 0 1.shstrtab STRTAB 0000000000000000 00001071 00000000000000e1 0000000000000000 0 0 1Key to Flags:W (write), A (alloc), X (execute), M (merge), S (strings), I (info),L (link order), O (extra OS processing required), G (group), T (TLS),C (compressed), x (unknown), o (OS specific), E (exclude),l (large), p (processor specific)
注意這里有26個符號段,主要有dynsym、dynstr段,這兩個段symtab、strtab被清除掉。
而且依舊可以被鏈接成功并且成功執(zhí)行程序:
~/test$ g++ main.cc -o main ./shared.so;./mainHello World 666
為什么動態(tài)鏈接庫被strip后還可以鏈接成功呢?因為strip只清除普通符號表,會保留動態(tài)符號表,即dynsym、dynstr段,而動態(tài)鏈接依靠的就是動態(tài)符號表。
靜態(tài)鏈接庫如果被strip后還能被鏈接成功嗎?
也是先說答案,合理strip后就可以。
先貼出兩段代碼:
// static.cc
void Print(int a) { std::cout << "Hello World " << a << std::endl; }
void Print(int a);
int main() { Print(666); return 0;}
先將static.cc打包成libsta.a:
gcc -c staticd.cc -o sta.oar -r libsta.a sta.o
查看下靜態(tài)庫的符號:
nm libsta.a
:U _GLOBAL_OFFSET_TABLE_000000000000008f t _GLOBAL__sub_I__Z5Printi0000000000000046 t _Z41__static_initialization_and_destruction_0ii0000000000000000 T _Z5PrintiU _ZNSolsEPFRSoS_EU _ZNSolsEiU _ZNSt8ios_base4InitC1EvU _ZNSt8ios_base4InitD1EvU _ZSt4coutU _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_0000000000000000 r _ZStL19piecewise_construct0000000000000000 b _ZStL8__ioinitU _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKcU __cxa_atexitU __dso_handle
將libsta.a庫strip后發(fā)現(xiàn)什么符號都沒有,且鏈接會失?。?/span>
~/test$ strip -s libsta.a~/test$ nm libsta.asta.o:nm: sta.o: no symbols~/test$ g++ main.cc -o main -L. -lsta; ./main./libsta.a: error adding symbols: Archive has no index; run ranlib to add onecollect2: error: ld returned 1 exit status-bash: ./main: No such file or directory
那難道靜態(tài)鏈接庫就不能strip了嗎?不strip的文件豈不是體積很大?
其實還是可以strip的,但需要合理的使用strip,這里需要換一個strip的參數(shù),就是--strip-unneeded,它確保strip掉的是沒有用的符號,保留用于鏈接的符號,盡管--strip-unneeded不如-s清除的徹底,但是保留了很多有用的信息,確保該鏈接庫是可用的。
strip --strip-unneeded libsta.a nm libsta.a :0000000000000000 T _Z5PrintiU _ZNSolsEPFRSoS_EU _ZNSolsEiU _ZNSt8ios_base4InitC1EvU _ZNSt8ios_base4InitD1EvU _ZSt4coutU _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_U _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKcU __cxa_atexitU __dso_handle
從上面可以看出:通過--strip-unneeded即清除了部分符號的信息,還能保證庫可用,減少程序體積。
關(guān)于strip,今天先介紹到這里,相信大家看完可以對strip理解的更深刻,并能更合理的使用strip。關(guān)于編譯和鏈接,大家可以后臺發(fā)送關(guān)鍵字“程序鏈接”了解更多細節(jié)。
參考資料
https://zhuanlan.zhihu.com/p/72475595
https://xuanxuanblingbling.github.io/ctf/tools/2019/09/06/symbol/
往期推薦
免責聲明:本文內(nèi)容由21ic獲得授權(quán)后發(fā)布,版權(quán)歸原作者所有,本平臺僅提供信息存儲服務(wù)。文章僅代表作者個人觀點,不代表本平臺立場,如有問題,請聯(lián)系我們,謝謝!