Related
I want a simple C method to be able to run hex bytecode on a Linux 64 bit machine. Here's the C program that I have:
char code[] = "\x48\x31\xc0";
#include <stdio.h>
int main(int argc, char **argv)
{
int (*func) ();
func = (int (*)()) code;
(int)(*func)();
printf("%s\n","DONE");
}
The code that I am trying to run ("\x48\x31\xc0") I obtained by writting this simple assembly program (it's not supposed to really do anything)
.text
.globl _start
_start:
xorq %rax, %rax
and then compiling and objdump-ing it to obtain the bytecode.
However, when I run my C program I get a segmentation fault. Any ideas?
Machine code has to be in an executable page. Your char code[] is in the read+write data section, without exec permission, so the code cannot be executed from there.
Here is a simple example of allocating an executable page with mmap:
#include <stdio.h>
#include <string.h>
#include <sys/mman.h>
int main ()
{
char code[] = {
0x8D, 0x04, 0x37, // lea eax,[rdi+rsi]
0xC3 // ret
};
int (*sum) (int, int) = NULL;
// allocate executable buffer
sum = mmap (0, sizeof(code), PROT_READ|PROT_WRITE|PROT_EXEC,
MAP_PRIVATE|MAP_ANONYMOUS, -1, 0);
// copy code to buffer
memcpy (sum, code, sizeof(code));
// doesn't actually flush cache on x86, but ensure memcpy isn't
// optimized away as a dead store.
__builtin___clear_cache (sum, sum + sizeof(sum)); // GNU C
// run code
int a = 2;
int b = 3;
int c = sum (a, b);
printf ("%d + %d = %d\n", a, b, c);
}
See another answer on this question for details about __builtin___clear_cache.
Until recent Linux kernel versions (sometime before 5.4), you could simply compile with gcc -z execstack - that would make all pages executable, including read-only data (.rodata), and read-write data (.data) where char code[] = "..." goes.
Now -z execstack only applies to the actual stack, so it currently works only for non-const local arrays. i.e. move char code[] = ... into main.
See Linux default behavior against `.data` section for the kernel change, and Unexpected exec permission from mmap when assembly files included in the project for the old behaviour: enabling Linux's READ_IMPLIES_EXEC process for that program. (In Linux 5.4, that Q&A shows you'd only get READ_IMPLIES_EXEC for a missing PT_GNU_STACK, like a really old binary; modern GCC -z execstack would set PT_GNU_STACK = RWX metadata in the executable, which Linux 5.4 would handle as making only the stack itself executable. At some point before that, PT_GNU_STACK = RWX did result in READ_IMPLIES_EXEC.)
The other option is to make system calls at runtime to copy into an executable page, or change permissions on the page it's in. That's still more complicated than using a local array to get GCC to copy code into executable stack memory.
(I don't know if there's an easy way to enable READ_IMPLIES_EXEC under modern kernels. Having no GNU-stack attribute at all in an ELF binary does that for 32-bit code, but not 64-bit.)
Yet another option is __attribute__((section(".text"))) const char code[] = ...;
Working example: https://godbolt.org/z/draGeh.
If you need the array to be writeable, e.g. for shellcode that inserts some zeros into strings, you could maybe link with ld -N. But probably best to use -z execstack and a local array.
Two problems in the question:
exec permission on the page, because you used an array that will go in the noexec read+write .data section.
your machine code doesn't end with a ret instruction so even if it did run, execution would fall into whatever was next in memory instead of returning.
And BTW, the REX prefix is totally redundant. "\x31\xc0" xor eax,eax has exactly the same effect as xor rax,rax.
You need the page containing the machine code to have execute permission. x86-64 page tables have a separate bit for execute separate from read permission, unlike legacy 386 page tables.
The easiest way to get static arrays to be in read+exec memory was to compile with gcc -z execstack. (Used to make the stack and other sections executable, now only the stack).
Until recently (2018 or 2019), the standard toolchain (binutils ld) would put section .rodata into the same ELF segment as .text, so they'd both have read+exec permission. Thus using const char code[] = "..."; was sufficient for executing manually-specified bytes as data, without execstack.
But on my Arch Linux system with GNU ld (GNU Binutils) 2.31.1, that's no longer the case. readelf -a shows that the .rodata section went into an ELF segment with .eh_frame_hdr and .eh_frame, and it only has Read permission. .text goes in a segment with Read + Exec, and .data goes in a segment with Read + Write (along with the .got and .got.plt). (What's the difference of section and segment in ELF file format)
I assume this change is to make ROP and Spectre attacks harder by not having read-only data in executable pages where sequences of useful bytes could be used as "gadgets" that end with the bytes for a ret or jmp reg instruction.
// TODO: use char code[] = {...} inside main, with -z execstack, for current Linux
// Broken on recent Linux, used to work without execstack.
#include <stdio.h>
// can be non-const if you use gcc -z execstack. static is also optional
static const char code[] = {
0x8D, 0x04, 0x37, // lea eax,[rdi+rsi] // retval = a+b;
0xC3 // ret
};
static const char ret0_code[] = "\x31\xc0\xc3"; // xor eax,eax ; ret
// the compiler will append a 0 byte to terminate the C string,
// but that's fine. It's after the ret.
int main () {
// void* cast is easier to type than a cast to function pointer,
// and in C can be assigned to any other pointer type. (not C++)
int (*sum) (int, int) = (void*)code;
int (*ret0)(void) = (void*)ret0_code;
// run code
int c = sum (2, 3);
return ret0();
}
On older Linux systems: gcc -O3 shellcode.c && ./a.out (Works because of const on global/static arrays)
On Linux before 5.5 (or so) gcc -O3 -z execstack shellcode.c && ./a.out (works because of -zexecstack regardless of where your machine code is stored). Fun fact: gcc allows -zexecstack with no space, but clang only accepts clang -z execstack.
These also work on Windows, where read-only data goes in .rdata instead of .rodata.
The compiler-generated main looks like this (from objdump -drwC -Mintel). You can run it inside gdb and set breakpoints on code and ret0_code
(I actually used gcc -no-pie -O3 -zexecstack shellcode.c hence the addresses near 401000
0000000000401020 <main>:
401020: 48 83 ec 08 sub rsp,0x8 # stack aligned by 16 before a call
401024: be 03 00 00 00 mov esi,0x3
401029: bf 02 00 00 00 mov edi,0x2 # 2 args
40102e: e8 d5 0f 00 00 call 402008 <code> # note the target address in the next page
401033: 48 83 c4 08 add rsp,0x8
401037: e9 c8 0f 00 00 jmp 402004 <ret0_code> # optimized tailcall
Or use system calls to modify page permissions
Instead of compiling with gcc -zexecstack, you can instead use mmap(PROT_EXEC) to allocate new executable pages, or mprotect(PROT_EXEC) to change existing pages to executable. (Including pages holding static data.) You also typically want at least PROT_READ and sometimes PROT_WRITE, of course.
Using mprotect on a static array means you're still executing the code from a known location, maybe making it easier to set a breakpoint on it.
On Windows you can use VirtualAlloc or VirtualProtect.
Telling the compiler that data is executed as code
Normally compilers like GCC assume that data and code are separate. This is like type-based strict aliasing, but even using char* doesn't make it well-defined to store into a buffer and then call that buffer as a function pointer.
In GNU C, you also need to use __builtin___clear_cache(buf, buf + len) after writing machine code bytes to a buffer, because the optimizer doesn't treat dereferencing a function pointer as reading bytes from that address. Dead-store elimination can remove the stores of machine code bytes into a buffer, if the compiler proves that the store isn't read as data by anything. https://codegolf.stackexchange.com/questions/160100/the-repetitive-byte-counter/160236#160236 and https://godbolt.org/g/pGXn3B has an example where gcc really does do this optimization, because gcc "knows about" malloc.
(And on non-x86 architectures where I-cache isn't coherent with D-cache, it actually will do any necessary cache syncing. On x86 it's purely a compile-time optimization blocker and doesn't expand to any instructions itself.)
Re: the weird name with three underscores: It's the usual __builtin_name pattern, but name is __clear_cache.
My edit on #AntoineMathys's answer added this.
In practice GCC/clang don't "know about" mmap(MAP_ANONYMOUS) the way they know about malloc. So in practice the optimizer will assume that the memcpy into the buffer might be read as data by the non-inline function call through the function pointer, even without __builtin___clear_cache(). (Unless you declared the function type as __attribute__((const)).)
On x86, where I-cache is coherent with data caches, having the stores happen in asm before the call is sufficient for correctness. On other ISAs, __builtin___clear_cache() will actually emit special instructions as well as ensuring the right compile-time ordering.
It's good practice to include it when copying code into a buffer because it doesn't cost performance, and stops hypothetical future compilers from breaking your code. (e.g. if they do understand that mmap(MAP_ANONYMOUS) gives newly-allocated anonymous memory that nothing else has a pointer to, just like malloc.)
With current GCC, I was able to provoke GCC into really doing an optimization we don't want by using __attribute__((const)) to tell the optimizer sum() is a pure function (that only reads its args, not global memory). GCC then knows sum() can't read the result of the memcpy as data.
With another memcpy into the same buffer after the call, GCC does dead-store elimination into just the 2nd store after the call. This results in no store before the first call so it executes the 00 00 add [rax], al bytes, segfaulting.
// demo of a problem on x86 when not using __builtin___clear_cache
#include <stdio.h>
#include <string.h>
#include <sys/mman.h>
int main ()
{
char code[] = {
0x8D, 0x04, 0x37, // lea eax,[rdi+rsi]
0xC3 // ret
};
__attribute__((const)) int (*sum) (int, int) = NULL;
// copy code to executable buffer
sum = mmap (0,sizeof(code),PROT_READ|PROT_WRITE|PROT_EXEC,
MAP_PRIVATE|MAP_ANON,-1,0);
memcpy (sum, code, sizeof(code));
//__builtin___clear_cache(sum, sum + sizeof(code));
int c = sum (2, 3);
//printf ("%d + %d = %d\n", a, b, c);
memcpy(sum, (char[]){0x31, 0xc0, 0xc3, 0}, 4); // xor-zero eax, ret, padding for a dword store
//__builtin___clear_cache(sum, sum + 4);
return sum(2,3);
}
Compiled on the Godbolt compiler explorer with GCC9.2 -O3
main:
push rbx
xor r9d, r9d
mov r8d, -1
mov ecx, 34
mov edx, 7
mov esi, 4
xor edi, edi
sub rsp, 16
call mmap
mov esi, 3
mov edi, 2
mov rbx, rax
call rax # call before store
mov DWORD PTR [rbx], 12828721 # 0xC3C031 = xor-zero eax, ret
add rsp, 16
pop rbx
ret # no 2nd call, CSEd away because const and same args
Passing different args would have gotten another call reg, but even with __builtin___clear_cache the two sum(2,3) calls can CSE. __attribute__((const)) doesn't respect changes to the machine code of a function. Don't do it. It's safe if you're going to JIT the function once and then call many times, though.
Uncommenting the first __clear_cache results in
mov DWORD PTR [rax], -1019804531 # lea; ret
call rax
mov DWORD PTR [rbx], 12828721 # xor-zero; ret
... still CSE and use the RAX return value
The first store is there because of __clear_cache and the sum(2,3) call. (Removing the first sum(2,3) call does let dead-store elimination happen across the __clear_cache.)
The second store is there because the side-effect on the buffer returned by mmap is assumed to be important, and that's the final value main leaves.
Godbolt's ./a.out option to run the program still seems to always fail (exit status of 255); maybe it sandboxes JITing? It works on my desktop with __clear_cache and crashes without.
mprotect on a page holding existing C variables.
You can also give a single existing page read+write+exec permission. This is an alternative to compiling with -z execstack
You don't need __clear_cache on a page holding read-only C variables because there's no store to optimize away. You would still need it for initializing a local buffer (on the stack). Otherwise GCC will optimize away the initializer for this private buffer that a non-inline function call definitely doesn't have a pointer to. (Escape analysis). It doesn't consider the possibility that the buffer might hold the machine code for the function unless you tell it that via __builtin___clear_cache.
#include <stdio.h>
#include <sys/mman.h>
#include <stdint.h>
// can be non-const if you want, we're using mprotect
static const char code[] = {
0x8D, 0x04, 0x37, // lea eax,[rdi+rsi] // retval = a+b;
0xC3 // ret
};
static const char ret0_code[] = "\x31\xc0\xc3";
int main () {
// void* cast is easier to type than a cast to function pointer,
// and in C can be assigned to any other pointer type. (not C++)
int (*sum) (int, int) = (void*)code;
int (*ret0)(void) = (void*)ret0_code;
// hard-coding x86's 4k page size for simplicity.
// also assume that `code` doesn't span a page boundary and that ret0_code is in the same page.
uintptr_t page = (uintptr_t)code & -4095ULL; // round down
mprotect((void*)page, 4096, PROT_READ|PROT_EXEC|PROT_WRITE); // +write in case the page holds any writeable C vars that would crash later code.
// run code
int c = sum (2, 3);
return ret0();
}
I used PROT_READ|PROT_EXEC|PROT_WRITE in this example so it works regardless of where your variable is. If it was a local on the stack and you left out PROT_WRITE, call would fail after making the stack read only when it tried to push a return address.
Also, PROT_WRITE lets you test shellcode that self-modifies, e.g. to edit zeros into its own machine code, or other bytes it was avoiding.
$ gcc -O3 shellcode.c # without -z execstack
$ ./a.out
$ echo $?
0
$ strace ./a.out
...
mprotect(0x55605aa3f000, 4096, PROT_READ|PROT_WRITE|PROT_EXEC) = 0
exit_group(0) = ?
+++ exited with 0 +++
If I comment out the mprotect, it does segfault with recent versions of GNU Binutils ld which no longer put read-only constant data into the same ELF segment as the .text section.
If I did something like ret0_code[2] = 0xc3;, I would need __builtin___clear_cache(ret0_code+2, ret0_code+2) after that to make sure the store wasn't optimized away, but if I don't modify the static arrays then it's not needed after mprotect. It is needed after mmap+memcpy or manual stores, because we want to execute bytes that have been written in C (with memcpy).
You need to include the assembly in-line via a special compiler directive so that it'll properly end up in a code segment. See this guide, for example: http://www.ibiblio.org/gferg/ldp/GCC-Inline-Assembly-HOWTO.html
Your machine code may be all right, but your CPU objects.
Modern CPUs manage memory in segments. In normal operation, the operating system loads a new program into a program-text segment and sets up a stack in a data segment. The operating system tells the CPU never to run code in a data segment. Your code is in code[], in a data segment. Thus the segfault.
This will take some effort.
Your code variable is stored in the .data section of your executable:
$ readelf -p .data exploit
String dump of section '.data':
[ 10] H1À
H1À is the value of your variable.
The .data section is not executable:
$ readelf -S exploit
There are 30 section headers, starting at offset 0x1150:
Section Headers:
[Nr] Name Type Address Offset
Size EntSize Flags Link Info Align
[...]
[24] .data PROGBITS 0000000000601010 00001010
0000000000000014 0000000000000000 WA 0 0 8
All 64-bit processors I'm familiar with support non-executable pages natively in the pagetables. Most newer 32-bit processors (the ones that support PAE) provide enough extra space in their pagetables for the operating system to emulate hardware non-executable pages. You'll need to run either an ancient OS or an ancient processor to get a .data section marked executable.
Because these are just flags in the executable, you ought to be able to set the X flag through some other mechanism, but I don't know how to do so. And your OS might not even let you have pages that are both writable and executable.
You may need to set the page executable before you may call it.
On MS-Windows, see the VirtualProtect -function.
URL: http://msdn.microsoft.com/en-us/library/windows/desktop/aa366898%28v=vs.85%29.aspx
Sorry, I couldn't follow above examples which are complicated.
So, I created an elegant solution for executing hex code from C.
Basically, you could use asm and .word keywords to place your instructions in hex format.
See below example:
asm volatile(".rept 1024\n"
CNOP
".endr\n");
where CNOP is defined as below:
#define ".word 0x00010001 \n"
Basically, c.nop instruction was not supported by my current assembler. So, I defined CNOP as the hex equivalent of c.nop with proper syntax and used inside asm, with which I was aware of.
.rept <NUM> .endr will basically, repeat the instruction NUM times.
This solution is working and verified.
I'm trying to write an ELF executable loader for x86-64 Linux, similar to this, which was implemented on ARM. Chris Rossbach's advanced OS class includes a lab that does basically what I want to do. My goal is to load a simple (statically-linked) "hello world" type binary into my process's memory and run it without execveing. I have successfully mmap'd the ELF file, set up the stack, and jumped to the ELF's entry point (_start).
// put ELF file into memory. This is just one line of a complex
// for() loop that loads the binary from a file.
mmap((void*)program_header.p_vaddr, program_header.p_memsz, map, MAP_PRIVATE|MAP_FIXED, elffd, program_header.p_offset);
newstack = mmap(NULL, 4096, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANONYMOUS, 0, 0); // Map a page for the stack
if((long)newstack < 0) {
fprintf(stderr, "ERROR: mmap returned error when allocating stack, %s\n", strerror(errno));
exit(1);
}
topstack = (unsigned long*)((unsigned char*)newstack+4096); // Top of new stack
*((unsigned long*)topstack-1) = 0; // Set up the stack
*((unsigned long*)topstack-2) = 0; // with argc, argv[], etc.
*((unsigned long*)topstack-3) = 0;
*((unsigned long*)topstack-4) = argv[1];
*((unsigned long*)topstack-5) = 1;
asm("mov %0,%%rsp\n" // Install new stack pointer
"xor %%rax, %%rax\n" // Zero registers
"xor %%rbx, %%rbx\n"
"xor %%rcx, %%rcx\n"
"xor %%rdx, %%rdx\n"
"xor %%rsi, %%rsi\n"
"xor %%rdi, %%rdi\n"
"xor %%r8, %%r8\n"
"xor %%r9, %%r9\n"
"xor %%r10, %%r10\n"
"xor %%r11, %%r11\n"
"xor %%r12, %%r12\n"
"xor %%r13, %%r13\n"
"xor %%r14, %%r14\n"
:
: "r"(topstack-5)
:"rax", "rbx", "rcx", "rdx", "rsi", "rdi", "r8", "r9", "r10", "r11", "r12", "r13", "r14");
asm("push %%rax\n"
"pop %%rax\n"
:
:
: "rax");
asm("mov %0,%%rax\n" // Jump to the entry point of the loaded ELF file
"jmp *%%rax\n"
:
: "r"(jump_target)
: );
I then step through this code in gdb. I've pasted the first few instructions of the startup code below. Everything works great until the first push instruction (starred). The push causes a segfault.
0x60026000 xor %ebp,%ebp
0x60026002 mov %rdx,%r9
0x60026005 pop %rsi
0x60026006 mov %rsp,%rdx
0x60026009 and $0xfffffffffffffff0,%rsp
0x6002600d * push %rax
0x6002600e push %rsp
0x6002600f mov $0x605f4990,%r8
I have tried:
Using the stack from the original process.
mmaping a new stack (as in the above code): (1) and (2) both cause segfaults.
pushing and poping to/from the stack before jmping to the loaded ELF file. This does not cause a segfault.
Changing the protection flags for the stack in the second mmap to PROT_READ | PROT_WRITE | PROT_EXEC. This doesn't make a difference.
I suspect this maybe has something to do with the segment descriptors (maybe?). It seems like the code from the ELF file that I'm loading does not have write access to the stack segment, no matter where it is located. I have not tried to modify the segment descriptor for the newly loaded binary or change the architectural segment registers. Is this necessary? Does anybody know how to fix this?
It turned out that when I was stepping through the loaded code in gdb, the debugger would consistently blow by the first push instruction when I typed nexti and instead continue execution. It was not in fact the push instruction that was causing the segfault but a much later instruction in the C library start code. The problem was caused by a failed call to mmap in the initial binary load that I didn't error check.
Regarding gdb randomly deciding to continue execution instead of stepping: this can be fixed by loading the symbols from the target executable after jumping to the newly loaded executable.
I've run into a problem (repetitively) with various company's' embedded linux products where GPL source code from them does not match what is actually running on a system. It's "close", but not quite right, especially with respect to the standard C library they use.
Isn't that a violation of the GPL?
Often this mismatch results in a programmer (like me) cross compiling only to have the device reply cryptically "file not found" or something similar when the program is run.
I'm not alone with this kind of problem -- For many people have threads directly and indirectly related to the problem: eg:
Compile parameters for MIPS based codesourcery toolchain?
And I've run into the problem on Sony devices, D-link, and many others. It's very common.
Making a new library is not a good solution, since most systems are ROMFS only, and LD_LIBRARY_PATH is sometimes broken -- so that installing a new library on the device wastes very limited memory and often won't work.
If I knew what the right source code version of the library was, I could go around the manufacturer's carelessness and compile it from the original developer's tree; but how can I find out which version I need when all I have is the binary of the library itself?
For example: I ran elfread -a libc.so.0 on a DSL modem's libc (see below); but I don't see anything here that could tell me exactly which libc it was...
How can I find the name of the source code, or an identifier from the library's binary so I can create a cross compiler using that library? eg: Can anyone tell me what source code this library came from, and how they know?
ELF Header:
Magic: 7f 45 4c 46 01 02 01 00 00 00 00 00 00 00 00 00
Class: ELF32
Data: 2's complement, big endian
Version: 1 (current)
OS/ABI: UNIX - System V
ABI Version: 0
Type: DYN (Shared object file)
Machine: MIPS R3000
Version: 0x1
Entry point address: 0x5a60
Start of program headers: 52 (bytes into file)
Start of section headers: 0 (bytes into file)
Flags: 0x1007, noreorder, pic, cpic, o32, mips1
Size of this header: 52 (bytes)
Size of program headers: 32 (bytes)
Number of program headers: 4
Size of section headers: 0 (bytes)
Number of section headers: 0
Section header string table index: 0
There are no sections in this file.
There are no sections to group in this file.
Program Headers:
Type Offset VirtAddr PhysAddr FileSiz MemSiz Flg Align
REGINFO 0x0000b4 0x000000b4 0x000000b4 0x00018 0x00018 R 0x4
LOAD 0x000000 0x00000000 0x00000000 0x2c9ee 0x2c9ee R E 0x1000
LOAD 0x02c9f0 0x0006c9f0 0x0006c9f0 0x009a0 0x040b8 RW 0x1000
DYNAMIC 0x0000cc 0x000000cc 0x000000cc 0x0579a 0x0579a RWE 0x4
Dynamic section at offset 0xcc contains 19 entries:
Tag Type Name/Value
0x0000000e (SONAME) Library soname: [libc.so.0]
0x00000004 (HASH) 0x18c
0x00000005 (STRTAB) 0x3e9c
0x00000006 (SYMTAB) 0x144c
0x0000000a (STRSZ) 6602 (bytes)
0x0000000b (SYMENT) 16 (bytes)
0x00000015 (DEBUG) 0x0
0x00000003 (PLTGOT) 0x6ce20
0x00000011 (REL) 0x5868
0x00000012 (RELSZ) 504 (bytes)
0x00000013 (RELENT) 8 (bytes)
0x70000001 (MIPS_RLD_VERSION) 1
0x70000005 (MIPS_FLAGS) NOTPOT
0x70000006 (MIPS_BASE_ADDRESS) 0x0
0x7000000a (MIPS_LOCAL_GOTNO) 11
0x70000011 (MIPS_SYMTABNO) 677
0x70000012 (MIPS_UNREFEXTNO) 17
0x70000013 (MIPS_GOTSYM) 0x154
0x00000000 (NULL) 0x0
There are no relocations in this file.
The decoding of unwind sections for machine type MIPS R3000 is not currently supported.
Histogram for bucket list length (total of 521 buckets):
Length Number % of total Coverage
0 144 ( 27.6%)
1 181 ( 34.7%) 27.1%
2 130 ( 25.0%) 66.0%
3 47 ( 9.0%) 87.1%
4 12 ( 2.3%) 94.3%
5 5 ( 1.0%) 98.1%
6 1 ( 0.2%) 99.0%
7 1 ( 0.2%) 100.0%
No version information found in this file.
Primary GOT:
Canonical gp value: 00074e10
Reserved entries:
Address Access Initial Purpose
0006ce20 -32752(gp) 00000000 Lazy resolver
0006ce24 -32748(gp) 80000000 Module pointer (GNU extension)
Local entries:
Address Access Initial
0006ce28 -32744(gp) 00070000
0006ce2c -32740(gp) 00030000
0006ce30 -32736(gp) 00000000
0006ce34 -32732(gp) 00010000
0006ce38 -32728(gp) 0006d810
0006ce3c -32724(gp) 0006d814
0006ce40 -32720(gp) 00020000
0006ce44 -32716(gp) 00000000
0006ce48 -32712(gp) 00000000
Global entries:
Address Access Initial Sym.Val. Type Ndx Name
0006ce4c -32708(gp) 000186c0 000186c0 FUNC bad section index[ 6] __fputc_unlocked
0006ce50 -32704(gp) 000211a4 000211a4 FUNC bad section index[ 6] sigprocmask
0006ce54 -32700(gp) 0001e2b4 0001e2b4 FUNC bad section index[ 6] free
0006ce58 -32696(gp) 00026940 00026940 FUNC bad section index[ 6] raise
...
truncated listing
....
Note:
The rest of this post is a blog showing how I came to ask the question above and to put useful information about the subject in one place.
Don't bother reading it unless you want to know I actually did research the question... in gory detail... and how NOT to answer my question.
The proper (theoretical) way to get a libc program running on (for example) a D-link modem would simply be to get the TRUE source code for the product from the manufacturer, and compile against those libraries.... (It's GPL !? right, so the law is on our side, right?)
For example: I just bought a D-Link DSL-520B modem and a 526B modem -- but found out after the fact that the manufacturer "forgot" to supply linux source code for the 520B but does have it for the 526B. I checked all of the DSL-5xxB devices online for source code & toolchains, finding to my delight that ALL of them (including 526B) -- contain the SAME pre-compiled libc.so.0 with MD5sum of 6ed709113ce615e9f170aafa0eac04a6 . So in theory, all supported modems in the DSL-5xxB family seemed to use the same libc library... and I hoped I might be able to use that library.
But after I figured out how to get the DSL modem itself to send me a copy of the installed /lib/libc.so.0 library -- I found to my disgust that they ALL use a library with MD5 sum of b8d492decc8207e724a0822641205078 . In NEITHER of the modems I bought (supported or not) was found the same library as contained in the source code toolchain.
To verify the toolchain from D-link was defective, I didn't compile a program (the toolchain wouldn't run on my PC anyway as it was the wrong binary format) -- but I found the toolchain had some pre-compiled mips binaries in it already; so I simply downloaded one to the modem and chmod +x -- and (surprise) I got the message "file not found." when I tried to run it ... It won't run.
So, I knew the toolchains were no good immediately, but not exactly why.
I decided to get a newer verson of MIPS GCC (binary version) that should have less bugs, more features and which is supported on most PC platforms. This is the way to GO!
See: Kernel.org pre-compiled binaries
I upgraded to gcc 4.9.0 after selecting the older "mips" verson from the above site to get the right FTP page; and set my shells' PATH variable to the /bin directory of the cross compiler once installed.
Then I copied all the header files and libraries from the D-link source code package into the new cross compiler just to verify that it could compile D-link libc binaries. And it did on the first try, compiling "hello world!" with no warnings or errors into a mips 32 big endian binary.
( START EDIT: ) #ChrisStratton points out in the comments (after this post) that my test of the toolchain is inadequate, and that using a newer GCC with an older library -- even though it links properly -- is flawed as a test. I wish there was a way to give him points for his comments -- I've become convinced that he's right; although that makes what D-link did even a worse problem -- for there's no way to know from the binaries on the modem which GCC they actually used. The GCC used for the kernel isn't necessarily the same used in user space.
In order to test the new compiler's compatibility with the modems and also make tools so I could get a copy of the actual libraries found on the modem: ( END EDIT ) I wrote a program that doesn't use the C library at all (but in two parts): It ran just fine... and the code is attached to show how it can be done.
The first listing is an assembly language program to bypass linking the standard C libraries on MIPS; and the second listing is a program meant to create an octal number dump of a binary file/stream using only the linux kernel. eg: It enables copying/pasting or scripting of binary data over telnet, netcat, etc... via ash/bash or busybox :) like a poor man's uucp.
// substart.S MIPS assembly language bypass of libc startup code
// it just calls main, and then jumps to the exit function
.text
.globl __start
__start: .ent __start
.frame $29, 32, $31
.set noreorder
.cpload $25
.set reorder
.cprestore 16
jal main
j exit
.end __start
// end substart.S
...and...
// octdump.c
// To compile w/o libc :
// mips-linux-gcc stubstart.S octdump.c -nostdlib -o octdump
// To compile with working libc (eg: x86 system) :
// gcc octdump.c -o octdump_x86
#include <syscall.h>
#include <errno.h>
#include <sys/types.h>
int* __errno_location(void) { return &errno; }
#ifdef _syscall1
// define three unix functions (exit,read,write) in terms of unix syscall macros.
_syscall1( void, exit, int, status );
_syscall3( ssize_t, read, int, fd, void*, buf, size_t, count );
_syscall3( ssize_t, write, int, fd, const void*, buf, size_t, count );
#endif
#include <unistd.h>
void oct( unsigned char c ) {
unsigned int n = c;
int m=6;
static unsigned char oval[6]={'\\','\\','0','0','0','0'};
if (n < 64) { m-=1; n <<= 3; }
if (n < 64) { m-=1; n <<= 3; }
if (n < 64) { m-=1; n <<= 3; }
oval[5]='0'+(n&7);
oval[4]='0'+((n>>3)&7);
oval[3]='0'+((n>>6)&7);
write( STDOUT_FILENO, oval, m );
}
int main(void) {
char buffer[255];
int count=1;
int i;
while (count>0) {
count=read( STDIN_FILENO, buffer, 17 );
if (count>0) write( STDOUT_FILENO, "echo -ne $'",11 );
for (i=0; i<count; ++i) oct( buffer[i] );
if (count>0) write( STDOUT_FILENO, "'\n", 2 );
}
write( STDOUT_FILENO,"#\n",2);
return 0;
}
Once mips' octdump was saved (chmod +x) as /var/octdump on the modem, it ran without errors.
(use your imagination about how I got it on there... Dlink's TFTP, & friends are broken.)
I was able to use octdump to copy all the dynamic libraries off the DSL modem and examine them, using an automated script to avoid copy/pasting by hand.
#!/bin/env python
# octget.py
# A program to upload a file off an embedded linux device via telnet
import socket
import time
import sys
import string
if len( sys.argv ) != 4 :
raise ValueError, "Usage: octget.py IP_OF_MODEM passwd path_to_file_to_get"
o = socket.socket( socket.AF_INET, socket.SOCK_STREAM )
o.connect((sys.argv[1],23)) # The IP address of the DSL modem.
time.sleep(1)
sys.stderr.write( o.recv(1024) )
o.send("admin\r\n");
time.sleep(0.1)
sys.stderr.write( o.recv(1024) )
o.send(sys.argv[2]+"\r\n")
time.sleep(0.1)
o.send("sh\r\n")
time.sleep(0.1)
sys.stderr.write( o.recv(1024) )
o.send("cd /var\r\n")
time.sleep(0.1)
sys.stderr.write( o.recv(1024) )
o.send("./octdump.x < "+sys.argv[3]+"\r\n" );
sys.stderr.write( o.recv(21) )
get="y"
while get and not ('#' in get):
get = o.recv(4096)
get = get.translate( None, '\r' )
sys.stdout.write( get )
time.sleep(0.5)
o.close()
The DSL520B modem had the following libraries...
libcrypt.so.0 libpsi.so libutil.so.0 ld-uClibc.so.0 libc.so.0 libdl.so.0 libpsixml.so
... and I thought I might cross compile using these libraries since (at least in theory) -- GCC could link against them; and my problem might be solved.
I made very sure to erase all the incompatible .so libraries from gcc-4.9.0/mips-linux/mips-linux/lib, but kept the generic crt..o files; then I copied the modem's libraries into the cross compiler directory.
But even though the kernel version of the source code, and the kernel version of the modem matched -- GCC found undefined symbols in the crt files.... So, either the generic crt files or the modem libraries themselves are somehow defective... and I don't know why. Without knowing how to get the full library version of the ? ucLibc ? library, I'm not sure how I can get the CORRECT source code to recompile the libraries and the crt's from scratch.
I have an ARM cross compiler in Ubuntu(arm-linux-gnueabi-gcc) and the default archtecture is ARMv7. However, I want to compile an ARMv5 binary. I do this by giving the compiler the -march=armv5te option.
So far, so good. Since my ARM system uses BusyBox, I have to compile my binary statically linked. So I give gcc the -static option.
However, I have a problem with libc.a which the linker links to my ARMv5 binary. This file is compiled with the ARMv7 architecture option. So, even if I cross-compile my ARM binary with ARMv5, I can't run it on my BusyBox based ARMv5 box.
How can I solve this problem?
Where can I get the ARMv5 libc.a static library, and how can I link it?
Thank you in advance.
You have two choices,
Get the right compiler.
Write your own 'C' Library.
Get the right compiler.
You are always safest to have a compiler match your system. This applies to x86 Linux and various distributions. You are lucky if different compilers work. It is more difficult when you cross-compile as often the compiler will not be automatically synced. Try to run a program on a 1999 x86 Mandrake Linux compiled on your 2014 Ubuntu system.
As well as instruction compatibility (which you have identified), there are ABI and OS dependencies. Specifically, the armv7 is most likely hardfloat (has floating point FPU and register call convention) and you need a softfloat (emulated FPU). The specific glibc (or ucLibc) has specific calls and expectations of the Linux OS. For instance, the way threads works has changed over time.
Write your own
You can always use -fno-builtin and -ffreestanding as well as -static. Then you can not use any libc functions, but you can program them your self.
There are external source, like Mark Martinec's snprintf and building blocks like write() which is easy to implement,
#define _SYS_IOCTL_H 1
#include <linux/unistd.h>
#include <linux/ioctl.h>
static inline int write(int fd, void *buf, int len)
{
int rval;
asm volatile ("mov r0, %1\n\t"
"mov r1, %2\n\t"
"mov r2, %3\n\t"
"mov r7, %4\n\t"
"swi #0\n\t"
"mov %0, r0\n\t"
: "=r" (rval)
: "r" (fd),
"r" (buf),
"r" (len),
"Ir" (__NR_write)
: "r0", "r1", "r2", "r7");
return rval;
}
static inline void exit(int status)
{
asm volatile ("mov r0, %0\n\t"
"mov r7, %1\n\t"
"swi #0\n\t"
: : "r" (status),
"Ir" (__NR_exit)
: "r0", "r7");
}
You have to add your own start-up machinery taken care of by the 'C' library,
/* Called from assembler startup. */
int main (int argc, char*argv[])
{
write(STDOUT, "Hello world\n", sizeof("Hello world\n"));
return 0;
}
/* Wrapper for main return code. */
void __attribute__ ((unused)) estart (int argc, char*argv[])
{
int rval = main(argc,argv);
exit(rval);
}
/* Setup arguments for estart [like main()]. */
void __attribute__ ((naked)) _start (void)
{
asm(" sub lr, lr, lr\n" /* Clear the link register. */
" ldr r0, [sp]\n" /* Get argc... */
" add r1, sp, #4\n" /* ... and argv ... */
" b estart\n" /* Let's go! */
);
}
If this is too daunting, because you need to implement a lot of functionality, then you can try and get various library source and rebuild them with -fno-builtin and make sure that the libraries do not get linked with the Ubuntu libraries, which are incompatible.
Projects like crosstool-ng can allow you to build a correct compiler (maybe with more advanced code generation) that suits the armv5 system exactly. This may seem like a pain, but the alternatives above aren't easy either.
I would like to use ld's --build-id option in order to add build information to my binary. However, I'm not sure how to make this information available inside the program. Assume I want to write a program that writes a backtrace every time an exception occurs, and a script that parses this information. The script reads the symbol table of the program and searches for the addresses printed in the backtrace (I'm forced to use such a script because the program is statically linked and backtrace_symbols is not working). In order for the script to work correctly I need to match build version of the program with the build version of the program which created the backtrace. How can I print the build version of the program (located in the .note.gnu.build-id elf section) from the program itself?
How can I print the build version of the program (located in the .note.gnu.build-id elf section) from the program itself?
You need to read the ElfW(Ehdr) (at the beginning of the file) to find program headers in your binary (.e_phoff and .e_phnum will tell you where program headers are, and how many of them to read).
You then read program headers, until you find PT_NOTE segment of your program. That segment will tell you offset to the beginning of all the notes in your binary.
You then need to read the ElfW(Nhdr) and skip the rest of the note (total size of the note is sizeof(Nhdr) + .n_namesz + .n_descsz, properly aligned), until you find a note with .n_type == NT_GNU_BUILD_ID.
Once you find NT_GNU_BUILD_ID note, skip past its .n_namesz, and read the .n_descsz bytes to read the actual build-id.
You can verify that you are reading the right data by comparing what you read with the output of readelf -n a.out.
P.S.
If you are going to go through the trouble to decode build-id as above, and if your executable is not stripped, it may be better for you to just decode and print symbol names instead (i.e. to replicate what backtrace_symbols does) -- it's actually easier to do than decoding ELF notes, because the symbol table contains fixed-sized entries.
Basically, this is the code I've written based on answer given to my question. In order to compile the code I had to make some changes and I hope it will work for as many types of platforms as possible. However, it was tested only on one build machine. One of the assumptions I used was that the program was built on the machine which runs it so no point in checking endianness compatibility between the program and the machine.
user#:~/$ uname -s -r -m -o
Linux 3.2.0-45-generic x86_64 GNU/Linux
user#:~/$ g++ test.cpp -o test
user#:~/$ readelf -n test | grep Build
Build ID: dc5c4682e0282e2bd8bc2d3b61cfe35826aa34fc
user#:~/$ ./test
Build ID: dc5c4682e0282e2bd8bc2d3b61cfe35826aa34fc
#include <elf.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/mman.h>
#include <sys/stat.h>
#if __x86_64__
# define ElfW(type) Elf64_##type
#else
# define ElfW(type) Elf32_##type
#endif
/*
detecting build id of a program from its note section
http://stackoverflow.com/questions/17637745/can-a-program-read-its-own-elf-section
http://www.scs.stanford.edu/histar/src/pkg/uclibc/utils/readelf.c
http://www.sco.com/developers/gabi/2000-07-17/ch5.pheader.html#note_section
*/
int main (int argc, char* argv[])
{
char *thefilename = argv[0];
FILE *thefile;
struct stat statbuf;
ElfW(Ehdr) *ehdr = 0;
ElfW(Phdr) *phdr = 0;
ElfW(Nhdr) *nhdr = 0;
if (!(thefile = fopen(thefilename, "r"))) {
perror(thefilename);
exit(EXIT_FAILURE);
}
if (fstat(fileno(thefile), &statbuf) < 0) {
perror(thefilename);
exit(EXIT_FAILURE);
}
ehdr = (ElfW(Ehdr) *)mmap(0, statbuf.st_size,
PROT_READ|PROT_WRITE, MAP_PRIVATE, fileno(thefile), 0);
phdr = (ElfW(Phdr) *)(ehdr->e_phoff + (size_t)ehdr);
while (phdr->p_type != PT_NOTE)
{
++phdr;
}
nhdr = (ElfW(Nhdr) *)(phdr->p_offset + (size_t)ehdr);
while (nhdr->n_type != NT_GNU_BUILD_ID)
{
nhdr = (ElfW(Nhdr) *)((size_t)nhdr + sizeof(ElfW(Nhdr)) + nhdr->n_namesz + nhdr->n_descsz);
}
unsigned char * build_id = (unsigned char *)malloc(nhdr->n_descsz);
memcpy(build_id, (void *)((size_t)nhdr + sizeof(ElfW(Nhdr)) + nhdr->n_namesz), nhdr->n_descsz);
printf(" Build ID: ");
for (int i = 0 ; i < nhdr->n_descsz ; ++i)
{
printf("%02x",build_id[i]);
}
free(build_id);
printf("\n");
return 0;
}
Yes, a program can read its own .note.gnu.build-id. The important piece is the dl_iterate_phdr function.
I've used this technique in Mesa (the OpenGL/Vulkan implementation) to read its own build-id for use with the on-disk shader cache.
I've extracted those bits into a separate project[1] for easy use by others.
[1] https://github.com/mattst88/build-id