Try   HackMD

Signal trace code

tags: learning

Signal of Linux Kernel

  • POSIX決定Linux中signal如何實現

    • Process中的所有thread共享signal

  • Linux kernel對signal的傳遞過程分成了兩階段

    1. Signal generation: 產生signal, 由kernel在目標process的task_struct更新signal的狀態

    2. Signal delivery: 傳遞signal, kernel將process的control flow交給signal handler

  • Signal已被產生但還沒傳遞時,稱為pending signal

    • 一個signal可以被多次產生,但只能被傳遞一次
    • 可能發生在
      1. process ignore該signal
        Note: SIGKILL. SIGSTOP不能被ignore
      2. process此時沒有在執行

  • task_struct中跟signal相關的data與其structure

    Image Not Showing Possible Reasons
    • The image file may be corrupted
    • The server hosting the image is unavailable
    • The image path is incorrect
    • The image format is not supported
    Learn More →

    可以看到一個process會有兩個sigpending,一個是thread共享的,
    放在signal->shared_pending中,另一個是private的,放在pending中,給kill這類針對所有thread的signal使用

Signal的產生

  • 以下函數都可以產生Signal

    ​send_sig
​send_sig_info
​force_sig
​force_sig_info
​sys_kill
​sys_tkill
​sys_tgkill

  • 以send_sig為例, function在kernel/signal.c中,v5.10.5版本的呼叫流程大概如下圖,而上述的function最終都會呼叫到send_signal

    Image Not Showing Possible Reasons
    • The image file may be corrupted
    • The server hosting the image is unavailable
    • The image path is incorrect
    • The image format is not supported
    Learn More →

  • __send_signal

    1. 用pid_type來判斷要signal要放進signal->shared_pending還是pending
    2. __sigqueue_alloc來創建一個signal queue,加到pending->list中
    3. 如果是SIGKILL
      • signalfd_notify用來通知signalfd有signal來了
      • sigaddset把pending->signal中代表sig的bit改成1
        Note: sigpending->signal是一個64bit的結構,每個bit對應一個signal
      • 最後呼叫complete_signal,用signal_wake_up在要接收signal的thread中設置TIF_SIGPENDING來完成signal的產生
        Note: 如果是SIGKILL,所有thread都會被設置TIF_SIGPENDING
    






















    ​static int __send_signal(int sig, struct kernel_siginfo *info, struct task_struct *t,
​			enum pid_type type, bool force)
​{
​	...
​	//1.
​	pending = (type != PIDTYPE_PID) ? &t->signal->shared_pending : &t->pending;
​	...
​	//2.
​	q = __sigqueue_alloc(sig, t, GFP_ATOMIC, override_rlimit);
​	if (q) {
​		list_add_tail(&q->list, &pending->list);
​	...
​	//3.
​	if ((sig == SIGKILL) || (t->flags & PF_KTHREAD))
​		goto out_set;
​	...
​out_set:
​	signalfd_notify(t, sig);
​	sigaddset(&pending->signal, sig);
​	...
​	complete_signal(sig, t, type);
​​​​    ...
​}

Signal的傳遞

  • 從kernel mode切換回user mode時, kernel都會檢查TIF_SIGPENDING,如果有被set,代表此thread有signal要處理

    • linux用ret_to_user跳回userspace,
    • 第20行檢查TIF_SIGPENDING,若為1就跳到working_pending
    • 第6行do_notify_resume,會呼叫do_signal
    

























    ​/*
​ * Ok, we need to do extra processing, enter the slow path.
​ */
​work_pending:
​	mov    x0, sp                  // 'regs'
​	bl    do_notify_resume         //call do_signal
​#ifdef CONFIG_TRACE_IRQFLAGS
​	bl    trace_hardirqs_on        // enabled while in userspace
​#endif
​	ldr    x1, [tsk, #TSK_TI_FLAGS]    // 再次檢查TIF_SIGPENDING
​	b    finish_ret_to_user
​/*
​ * "slow" syscall return path.
​ */
​ret_to_user:
​	disable_daif
​	gic_prio_kentry_setup tmp=x3
​	ldr    x1, [tsk, #TSK_TI_FLAGS]
​	and    x2, x1, #_TIF_WORK_MASK //檢查TIF_SIGPENDING
​	cbnz    x2, work_pending       //若不為0就跳到working_pending
​finish_ret_to_user:
​	enable_step_tsk x1, x2
​#ifdef CONFIG_GCC_PLUGIN_STACKLEAK
​	bl    stackleak_erase
​#endif
​	kernel_exit 0

  • do_signal

    Image Not Showing Possible Reasons
    • The image file may be corrupted
    • The server hosting the image is unavailable
    • The image path is incorrect
    • The image format is not supported
    Learn More →

    Kernel對signal的處理,主要在兩個function上

    1. 第10行的get_signal: 從queue將此次處理的signal取出,放進一個ksignal的struct中
    2. 第26行的handle_signal: 負責為user space準備好處理signal需要的環境






























static void do_signal(struct pt_regs *regs)
{
    ...
    struct ksignal ksig;
    ...
    /*
     * Get the signal to deliver. When running under ptrace, at this point
     * the debugger may change all of our registers.
     */
    if (get_signal(&ksig)) {
    	/*
    	 * Depending on the signal settings, we may need to revert the
    	 * decision to restart the system call, but skip this if a
    	 * debugger has chosen to restart at a different PC.
    	 */
        if (regs->pc == restart_addr &&
		(retval == -ERESTARTNOHAND ||
		retval == -ERESTART_RESTARTBLOCK ||
		(retval == -ERESTARTSYS &&
		!(ksig.ka.sa.sa_flags & SA_RESTART)))) 
	{
            regs->regs[0] = -EINTR;
	    regs->pc = continue_addr;
	}

	handle_signal(&ksig, regs);
	return;
    }
	...
}
  • get_signal
    1. 第13行將signr指定為要處理的signal number
    2. 第15行dequeue_signal往下執行會呼叫sigdelset跟__sigqueue_free
      • collect_signal中的sigdelset會負責將對應的bit清成0
      • collect_signal中的__sigqueue_free負責將signal從queue中移除
      • 最後recalc_sigpending檢查是否還有待傳遞的signal,沒有就把TIF_SIGPENDING清成0



























bool get_signal(struct ksignal *ksig)
{
    struct sighand_struct *sighand = current->sighand;
    struct signal_struct *signal = current->signal;
    int signr; //最後會是signal number
    ...
    /*
     * Signals generated by the execution of an instruction
     * need to be delivered before any other pending signals
     * so that the instruction pointer in the signal stack
     * frame points to the faulting instruction.
     */
    signr = dequeue_synchronous_signal(&ksig->info);
    if (!signr)
        signr = dequeue_signal(current, &current->blocked, &ksig->info);
    ...
    ksig->sig = signr;
    return ksig->sig > 0;
}

/*ksignal中會包含處理signal需要的所有訊息*/
struct ksignal {
    struct k_sigaction ka; //signal對應的處理方式
    kernel_siginfo_t info; //附加訊息
    int sig;               //signal number
};
//Linux可以用`sigaction`設定對特定signal的處理方式
  • handle_signal負責準備處理signal需要的環境
    1. 保存user space目前的Context
    2. 將執行流程轉交給signal handler
    3. signal handler處理完後,在轉交給kernel,讓kernel將執行權限還給user space




























/*
 * OK, we're invoking a handler
 */	
static void handle_signal(struct ksignal *ksig, struct pt_regs *regs)
{
    sigset_t *oldset = sigmask_to_save();
    int ret;

    /*
     * Perform fixup for the pre-signal frame.
     */
    rseq_signal_deliver(ksig, regs);

    /*
     * Set up the stack frame
     */
    if (ksig->ka.sa.sa_flags & SA_SIGINFO)
    	ret = setup_rt_frame(ksig, oldset, regs);
    else
     	ret = setup_frame(ksig, oldset, regs);

    /*
     * Check that the resulting registers are actually sane.
     */
    ret |= !valid_user_regs(regs);

    signal_setup_done(ret, ksig, 0);
}
  • 其中上面第18行的setup_rt_frame呼叫get_sigframe(下面第4行)負責處理1.的部分
    • 當process從user space進入kernel space
      的時候，user space當下的context就會被保存在kernel space的stack上,由kernel中一個名為pt_regs的struct保存user space在進入kernek前的register
    • 當從kernel space回到user space的時,pt_regs就會丟失，因為每次由進入kernel space的時候,kernel space的stack都一定是空的,因此不能把user space的context保存在kernel上,必須保存在user space的stack上
    • 新增的stack frame layout由get_sigframe規劃































static int
setup_rt_frame(struct ksignal *ksig, sigset_t *set, struct pt_regs *regs)
{
    struct rt_sigframe __user *frame = get_sigframe(ksig, regs, sizeof(*frame));
    int err = 0;

    if (!frame)
    	return 1;

    err |= copy_siginfo_to_user(&frame->info, &ksig->info);
	
    err |= __put_user(0, &frame->sig.uc.uc_flags);
    err |= __put_user(NULL, &frame->sig.uc.uc_link);

    err |= __save_altstack(&frame->sig.uc.uc_stack, regs->ARM_sp);
    err |= setup_sigframe(&frame->sig, regs, set);
    if (err == 0)
    	err = setup_return(regs, ksig, frame->sig.retcode, frame);

    if (err == 0) {
    	/*
    	 * For realtime signals we must also set the second and third
    	 * arguments for the signal handler.
    	 *   -- Peter Maydell <pmaydell@chiark.greenend.org.uk> 2000-12-06
    	 */
    	regs->ARM_r1 = (unsigned long)&frame->info;
    	regs->ARM_r2 = (unsigned long)&frame->sig.uc;
    }

    return err;
}
  • 2.3.則由上面第18行的setup_return(以ARM64為例)完成
    1. 第9行,因為經由setup_sigframe已經將user space的 context存到user中,因此這邊可以直接修改pt_regs來將PC指向signal handler
    2. 按照ARM64的ABI, regs[0]存signal number,regs[29]是stack pointer, regs[30]是Link Register(存放函數的return address)
    3. 第22行將sigtramp指向VDSO的rt_sigreturn(定義在arch/arm64/kernel/vdso/sigreturn.S),並在第24行將return address設為sigtramp,因此signal handler執行完後會接著執行rt_sigreturn

























static void setup_return(struct pt_regs *regs, struct k_sigaction *ka,
                         struct rt_sigframe_user_layout *user, int usig)
{
    __sigrestore_t sigtramp;

    regs->regs[0] = usig;
    regs->sp = (unsigned long)user->sigframe;
    regs->regs[29] = (unsigned long)&user->next_frame->fp;
    regs->pc = (unsigned long)ka->sa.sa_handler;

    if (system_supports_bti()) {
    	regs->pstate &= ~PSR_BTYPE_MASK;
     	regs->pstate |= PSR_BTYPE_C;
    }

    /* TCO (Tag Check Override) always cleared for signal handlers */
    regs->pstate &= ~PSR_TCO_BIT;

    if (ka->sa.sa_flags & SA_RESTORER)
    	sigtramp = ka->sa.sa_restorer;
    else
     	sigtramp = VDSO_SYMBOL(current->mm->context.vdso, sigtramp);

    regs->regs[30] = (unsigned long)sigtramp;
}
  • 使用system call代表會再次切換到kernel mode,由kernle執行sys_rt_sigreturn,呼叫restore_sigframe來恢復user space的執行
































asmlinkage int sys_rt_sigreturn(struct pt_regs *regs)
{
    struct rt_sigframe __user *frame;

    /* Always make any pending restarted system calls return -EINTR */
    current->restart_block.fn = do_no_restart_syscall;

    /*
     * Since we stacked the signal on a 64-bit boundary,
     * then 'sp' should be word aligned here.  If it's
     * not, then the user is trying to mess with us.
     */
    if (regs->ARM_sp & 7)
    	goto badframe;

    frame = (struct rt_sigframe __user *)regs->ARM_sp;

    if (!access_ok(frame, sizeof (*frame)))
    	goto badframe;

    if (restore_sigframe(regs, &frame->sig))
    	goto badframe;

    if (restore_altstack(&frame->sig.uc.uc_stack))
    	goto badframe;

    return regs->ARM_r0;

badframe:
    force_sig(SIGSEGV);
    return 0;
}
Last changed by 
 
許瓈方
0
978

Read more

pwnable.kr Write Up

原始碼

Jan 17, 2025
來拆一拆__libc_start_main

從entry point(start)開始start :::info在剛開始執行時,會把參數以倒序+argc push到stack上 ex. gdb&gt; r aaa 會依序push aaa, binary_path, 0x2(2個arg) (0x4009d0是entry pt) 因此第三行的pop rsi就相當於把stack頂的argc pop給rsi 第四行的mov rdx, rsp相當於把binary_path傳給rdx 第五行對rsp &amp; 0xFF...F0是為了要確保stack是0x10對齊 :::

Feb 5, 2023
SOLACE PubSub+

Introduction 整合跨廠、Data center、Cloud Server及IoT，作為設備之間即時通訊的管理平台Event Management支援EDA (Event-Driven Architecture)偵測特定Event Event Streaming 建立Event mesh以共享數據Event Mesh是一個由事件代理系統(Event Broker)所組成的網路架構層，可以將某應用程式產生的event動態route到其他應用程式(無論他部屬在哪裡) Event Broker接收從Producers產生的事件訊息，並將事件訊息route給事件訂閱者(Consumers) Event Broker可以部屬在Cloud (包含公有雲&amp;私有雲,etc) Software (各類Container或虛擬機管理平台) Appliance

Dec 28, 2022
常用協定及網路相關整理

OSI vs TCP/IP OSI(7&#x5C64;) TCP/IP(4&#x5C64;) &#x7528;&#x9014; ex Protocol 7 Application&#x61C9;&#x7528;&#x5C64;

Jan 16, 2022
Read more from 許瓈方
Published on HackMD