[net-next-2.6.git] / arch / x86_64 / kernel / ptrace.c

/* ptrace.c */
/* By Ross Biro 1/23/92 */
/*
 * Pentium III FXSR, SSE support
 *	Gareth Hughes <gareth@valinux.com>, May 2000
 * 
 * x86-64 port 2000-2002 Andi Kleen
 */

#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/mm.h>
#include <linux/smp.h>
#include <linux/errno.h>
#include <linux/ptrace.h>
#include <linux/user.h>
#include <linux/security.h>
#include <linux/audit.h>
#include <linux/seccomp.h>
#include <linux/signal.h>

#include <asm/uaccess.h>
#include <asm/pgtable.h>
#include <asm/system.h>
#include <asm/processor.h>
#include <asm/i387.h>
#include <asm/debugreg.h>
#include <asm/ldt.h>
#include <asm/desc.h>
#include <asm/proto.h>
#include <asm/ia32.h>

/*
 * does not yet catch signals sent when the child dies.
 * in exit.c or in signal.c.
 */

/*
 * Determines which flags the user has access to [1 = access, 0 = no access].
 * Prohibits changing ID(21), VIP(20), VIF(19), VM(17), IOPL(12-13), IF(9).
 * Also masks reserved bits (63-22, 15, 5, 3, 1).
 */
#define FLAG_MASK 0x54dd5UL

/* set's the trap flag. */
#define TRAP_FLAG 0x100UL

/*
 * eflags and offset of eflags on child stack..
 */
#define EFLAGS offsetof(struct pt_regs, eflags)
#define EFL_OFFSET ((int)(EFLAGS-sizeof(struct pt_regs)))

/*
 * this routine will get a word off of the processes privileged stack. 
 * the offset is how far from the base addr as stored in the TSS.  
 * this routine assumes that all the privileged stacks are in our
 * data space.
 */   
static inline unsigned long get_stack_long(struct task_struct *task, int offset)
{
	unsigned char *stack;

	stack = (unsigned char *)task->thread.rsp0;
	stack += offset;
	return (*((unsigned long *)stack));
}

/*
 * this routine will put a word on the processes privileged stack. 
 * the offset is how far from the base addr as stored in the TSS.  
 * this routine assumes that all the privileged stacks are in our
 * data space.
 */
static inline long put_stack_long(struct task_struct *task, int offset,
	unsigned long data)
{
	unsigned char * stack;

	stack = (unsigned char *) task->thread.rsp0;
	stack += offset;
	*(unsigned long *) stack = data;
	return 0;
}

#define LDT_SEGMENT 4

unsigned long convert_rip_to_linear(struct task_struct *child, struct pt_regs *regs)
{
	unsigned long addr, seg;

	addr = regs->rip;
	seg = regs->cs & 0xffff;

	/*
	 * We'll assume that the code segments in the GDT
	 * are all zero-based. That is largely true: the
	 * TLS segments are used for data, and the PNPBIOS
	 * and APM bios ones we just ignore here.
	 */
	if (seg & LDT_SEGMENT) {
		u32 *desc;
		unsigned long base;

		down(&child->mm->context.sem);
		desc = child->mm->context.ldt + (seg & ~7);
		base = (desc[0] >> 16) | ((desc[1] & 0xff) << 16) | (desc[1] & 0xff000000);

		/* 16-bit code segment? */
		if (!((desc[1] >> 22) & 1))
			addr &= 0xffff;
		addr += base;
		up(&child->mm->context.sem);
	}
	return addr;
}

static int is_setting_trap_flag(struct task_struct *child, struct pt_regs *regs)
{
	int i, copied;
	unsigned char opcode[15];
	unsigned long addr = convert_rip_to_linear(child, regs);

	copied = access_process_vm(child, addr, opcode, sizeof(opcode), 0);
	for (i = 0; i < copied; i++) {
		switch (opcode[i]) {
		/* popf and iret */
		case 0x9d: case 0xcf:
			return 1;

			/* CHECKME: 64 65 */

		/* opcode and address size prefixes */
		case 0x66: case 0x67:
			continue;
		/* irrelevant prefixes (segment overrides and repeats) */
		case 0x26: case 0x2e:
		case 0x36: case 0x3e:
		case 0x64: case 0x65:
		case 0xf2: case 0xf3:
			continue;

		case 0x40 ... 0x4f:
			if (regs->cs != __USER_CS)
				/* 32-bit mode: register increment */
				return 0;
			/* 64-bit mode: REX prefix */
			continue;

			/* CHECKME: f2, f3 */

		/*
		 * pushf: NOTE! We should probably not let
		 * the user see the TF bit being set. But
		 * it's more pain than it's worth to avoid
		 * it, and a debugger could emulate this
		 * all in user space if it _really_ cares.
		 */
		case 0x9c:
		default:
			return 0;
		}
	}
	return 0;
}

static void set_singlestep(struct task_struct *child)
{
	struct pt_regs *regs = task_pt_regs(child);

	/*
	 * Always set TIF_SINGLESTEP - this guarantees that
	 * we single-step system calls etc..  This will also
	 * cause us to set TF when returning to user mode.
	 */
	set_tsk_thread_flag(child, TIF_SINGLESTEP);

	/*
	 * If TF was already set, don't do anything else
	 */
	if (regs->eflags & TRAP_FLAG)
		return;

	/* Set TF on the kernel stack.. */
	regs->eflags |= TRAP_FLAG;

	/*
	 * ..but if TF is changed by the instruction we will trace,
	 * don't mark it as being "us" that set it, so that we
	 * won't clear it by hand later.
	 */
	if (is_setting_trap_flag(child, regs))
		return;

	child->ptrace |= PT_DTRACE;
}

static void clear_singlestep(struct task_struct *child)
{
	/* Always clear TIF_SINGLESTEP... */
	clear_tsk_thread_flag(child, TIF_SINGLESTEP);

	/* But touch TF only if it was set by us.. */
	if (child->ptrace & PT_DTRACE) {
		struct pt_regs *regs = task_pt_regs(child);
		regs->eflags &= ~TRAP_FLAG;
		child->ptrace &= ~PT_DTRACE;
	}
}

/*
 * Called by kernel/ptrace.c when detaching..
 *
 * Make sure the single step bit is not set.
 */
void ptrace_disable(struct task_struct *child)
{ 
	clear_singlestep(child);
}

static int putreg(struct task_struct *child,
	unsigned long regno, unsigned long value)
{
	unsigned long tmp; 
	
	/* Some code in the 64bit emulation may not be 64bit clean.
	   Don't take any chances. */
	if (test_tsk_thread_flag(child, TIF_IA32))
		value &= 0xffffffff;
	switch (regno) {
		case offsetof(struct user_regs_struct,fs):
			if (value && (value & 3) != 3)
				return -EIO;
			child->thread.fsindex = value & 0xffff; 
			return 0;
		case offsetof(struct user_regs_struct,gs):
			if (value && (value & 3) != 3)
				return -EIO;
			child->thread.gsindex = value & 0xffff;
			return 0;
		case offsetof(struct user_regs_struct,ds):
			if (value && (value & 3) != 3)
				return -EIO;
			child->thread.ds = value & 0xffff;
			return 0;
		case offsetof(struct user_regs_struct,es): 
			if (value && (value & 3) != 3)
				return -EIO;
			child->thread.es = value & 0xffff;
			return 0;
		case offsetof(struct user_regs_struct,ss):
			if ((value & 3) != 3)
				return -EIO;
			value &= 0xffff;
			return 0;
		case offsetof(struct user_regs_struct,fs_base):
			if (value >= TASK_SIZE_OF(child))
				return -EIO;
			child->thread.fs = value;
			return 0;
		case offsetof(struct user_regs_struct,gs_base):
			if (value >= TASK_SIZE_OF(child))
				return -EIO;
			child->thread.gs = value;
			return 0;
		case offsetof(struct user_regs_struct, eflags):
			value &= FLAG_MASK;
			tmp = get_stack_long(child, EFL_OFFSET); 
			tmp &= ~FLAG_MASK; 
			value |= tmp;
			break;
		case offsetof(struct user_regs_struct,cs): 
			if ((value & 3) != 3)
				return -EIO;
			value &= 0xffff;
			break;
	}
	put_stack_long(child, regno - sizeof(struct pt_regs), value);
	return 0;
}

static unsigned long getreg(struct task_struct *child, unsigned long regno)
{
	unsigned long val;
	switch (regno) {
		case offsetof(struct user_regs_struct, fs):
			return child->thread.fsindex;
		case offsetof(struct user_regs_struct, gs):
			return child->thread.gsindex;
		case offsetof(struct user_regs_struct, ds):
			return child->thread.ds;
		case offsetof(struct user_regs_struct, es):
			return child->thread.es; 
		case offsetof(struct user_regs_struct, fs_base):
			return child->thread.fs;
		case offsetof(struct user_regs_struct, gs_base):
			return child->thread.gs;
		default:
			regno = regno - sizeof(struct pt_regs);
			val = get_stack_long(child, regno);
			if (test_tsk_thread_flag(child, TIF_IA32))
				val &= 0xffffffff;
			return val;
	}

}

long arch_ptrace(struct task_struct *child, long request, long addr, long data)
{
	long i, ret;
	unsigned ui;

	switch (request) {
	/* when I and D space are separate, these will need to be fixed. */
	case PTRACE_PEEKTEXT: /* read word at location addr. */ 
	case PTRACE_PEEKDATA:
		ret = generic_ptrace_peekdata(child, addr, data);
		break;

	/* read the word at location addr in the USER area. */
	case PTRACE_PEEKUSR: {
		unsigned long tmp;

		ret = -EIO;
		if ((addr & 7) ||
		    addr > sizeof(struct user) - 7)
			break;

		switch (addr) { 
		case 0 ... sizeof(struct user_regs_struct) - sizeof(long):
			tmp = getreg(child, addr);
			break;
		case offsetof(struct user, u_debugreg[0]):
			tmp = child->thread.debugreg0;
			break;
		case offsetof(struct user, u_debugreg[1]):
			tmp = child->thread.debugreg1;
			break;
		case offsetof(struct user, u_debugreg[2]):
			tmp = child->thread.debugreg2;
			break;
		case offsetof(struct user, u_debugreg[3]):
			tmp = child->thread.debugreg3;
			break;
		case offsetof(struct user, u_debugreg[6]):
			tmp = child->thread.debugreg6;
			break;
		case offsetof(struct user, u_debugreg[7]):
			tmp = child->thread.debugreg7;
			break;
		default:
			tmp = 0;
			break;
		}
		ret = put_user(tmp,(unsigned long __user *) data);
		break;
	}

	/* when I and D space are separate, this will have to be fixed. */
	case PTRACE_POKETEXT: /* write the word at location addr. */
	case PTRACE_POKEDATA:
		ret = 0;
		if (access_process_vm(child, addr, &data, sizeof(data), 1) == sizeof(data))
			break;
		ret = -EIO;
		break;

	case PTRACE_POKEUSR: /* write the word at location addr in the USER area */
	{
		int dsize = test_tsk_thread_flag(child, TIF_IA32) ? 3 : 7;
		ret = -EIO;
		if ((addr & 7) ||
		    addr > sizeof(struct user) - 7)
			break;

		switch (addr) { 
		case 0 ... sizeof(struct user_regs_struct) - sizeof(long):
			ret = putreg(child, addr, data);
			break;
		/* Disallows to set a breakpoint into the vsyscall */
		case offsetof(struct user, u_debugreg[0]):
			if (data >= TASK_SIZE_OF(child) - dsize) break;
			child->thread.debugreg0 = data;
			ret = 0;
			break;
		case offsetof(struct user, u_debugreg[1]):
			if (data >= TASK_SIZE_OF(child) - dsize) break;
			child->thread.debugreg1 = data;
			ret = 0;
			break;
		case offsetof(struct user, u_debugreg[2]):
			if (data >= TASK_SIZE_OF(child) - dsize) break;
			child->thread.debugreg2 = data;
			ret = 0;
			break;
		case offsetof(struct user, u_debugreg[3]):
			if (data >= TASK_SIZE_OF(child) - dsize) break;
			child->thread.debugreg3 = data;
			ret = 0;
			break;
		case offsetof(struct user, u_debugreg[6]):
				  if (data >> 32)
				break; 
			child->thread.debugreg6 = data;
			ret = 0;
			break;
		case offsetof(struct user, u_debugreg[7]):
			/* See arch/i386/kernel/ptrace.c for an explanation of
			 * this awkward check.*/
			data &= ~DR_CONTROL_RESERVED;
			for(i=0; i<4; i++)
				if ((0x5554 >> ((data >> (16 + 4*i)) & 0xf)) & 1)
					break;
			if (i == 4) {
			  child->thread.debugreg7 = data;
			  if (data)
			  	set_tsk_thread_flag(child, TIF_DEBUG);
			  else
			  	clear_tsk_thread_flag(child, TIF_DEBUG);
			  ret = 0;
		  	}
		  break;
		}
		break;
	}
	case PTRACE_SYSCALL: /* continue and stop at next (return from) syscall */
	case PTRACE_CONT:    /* restart after signal. */

		ret = -EIO;
		if (!valid_signal(data))
			break;
		if (request == PTRACE_SYSCALL)
			set_tsk_thread_flag(child,TIF_SYSCALL_TRACE);
		else
			clear_tsk_thread_flag(child,TIF_SYSCALL_TRACE);
		clear_tsk_thread_flag(child, TIF_SINGLESTEP);
		child->exit_code = data;
		/* make sure the single step bit is not set. */
		clear_singlestep(child);
		wake_up_process(child);
		ret = 0;
		break;

#ifdef CONFIG_IA32_EMULATION
		/* This makes only sense with 32bit programs. Allow a
		   64bit debugger to fully examine them too. Better
		   don't use it against 64bit processes, use
		   PTRACE_ARCH_PRCTL instead. */
	case PTRACE_SET_THREAD_AREA: {
		struct user_desc __user *p;
		int old; 
		p = (struct user_desc __user *)data;
		get_user(old,  &p->entry_number); 
		put_user(addr, &p->entry_number);
		ret = do_set_thread_area(&child->thread, p);
		put_user(old,  &p->entry_number); 
		break;
	case PTRACE_GET_THREAD_AREA:
		p = (struct user_desc __user *)data;
		get_user(old,  &p->entry_number); 
		put_user(addr, &p->entry_number);
		ret = do_get_thread_area(&child->thread, p);
		put_user(old,  &p->entry_number); 
		break;
	} 
#endif
		/* normal 64bit interface to access TLS data. 
		   Works just like arch_prctl, except that the arguments
		   are reversed. */
	case PTRACE_ARCH_PRCTL: 
		ret = do_arch_prctl(child, data, addr);
		break;

/*
 * make the child exit.  Best I can do is send it a sigkill. 
 * perhaps it should be put in the status that it wants to 
 * exit.
 */
	case PTRACE_KILL:
		ret = 0;
		if (child->exit_state == EXIT_ZOMBIE)	/* already dead */
			break;
		clear_tsk_thread_flag(child, TIF_SINGLESTEP);
		child->exit_code = SIGKILL;
		/* make sure the single step bit is not set. */
		clear_singlestep(child);
		wake_up_process(child);
		break;

	case PTRACE_SINGLESTEP:    /* set the trap flag. */
		ret = -EIO;
		if (!valid_signal(data))
			break;
		clear_tsk_thread_flag(child,TIF_SYSCALL_TRACE);
		set_singlestep(child);
		child->exit_code = data;
		/* give it a chance to run. */
		wake_up_process(child);
		ret = 0;
		break;

	case PTRACE_DETACH:
		/* detach a process that was attached. */
		ret = ptrace_detach(child, data);
		break;

	case PTRACE_GETREGS: { /* Get all gp regs from the child. */
	  	if (!access_ok(VERIFY_WRITE, (unsigned __user *)data,
			       sizeof(struct user_regs_struct))) {
			ret = -EIO;
			break;
		}
		ret = 0;
		for (ui = 0; ui < sizeof(struct user_regs_struct); ui += sizeof(long)) {
			ret |= __put_user(getreg(child, ui),(unsigned long __user *) data);
			data += sizeof(long);
		}
		break;
	}

	case PTRACE_SETREGS: { /* Set all gp regs in the child. */
		unsigned long tmp;
	  	if (!access_ok(VERIFY_READ, (unsigned __user *)data,
			       sizeof(struct user_regs_struct))) {
			ret = -EIO;
			break;
		}
		ret = 0;
		for (ui = 0; ui < sizeof(struct user_regs_struct); ui += sizeof(long)) {
			ret = __get_user(tmp, (unsigned long __user *) data);
			if (ret)
				break;
			ret = putreg(child, ui, tmp);
			if (ret)
				break;
			data += sizeof(long);
		}
		break;
	}

	case PTRACE_GETFPREGS: { /* Get the child extended FPU state. */
		if (!access_ok(VERIFY_WRITE, (unsigned __user *)data,
			       sizeof(struct user_i387_struct))) {
			ret = -EIO;
			break;
		}
		ret = get_fpregs((struct user_i387_struct __user *)data, child);
		break;
	}

	case PTRACE_SETFPREGS: { /* Set the child extended FPU state. */
		if (!access_ok(VERIFY_READ, (unsigned __user *)data,
			       sizeof(struct user_i387_struct))) {
			ret = -EIO;
			break;
		}
		set_stopped_child_used_math(child);
		ret = set_fpregs(child, (struct user_i387_struct __user *)data);
		break;
	}

	default:
		ret = ptrace_request(child, request, addr, data);
		break;
	}
	return ret;
}

static void syscall_trace(struct pt_regs *regs)
{

#if 0
	printk("trace %s rip %lx rsp %lx rax %d origrax %d caller %lx tiflags %x ptrace %x\n",
	       current->comm,
	       regs->rip, regs->rsp, regs->rax, regs->orig_rax, __builtin_return_address(0),
	       current_thread_info()->flags, current->ptrace); 
#endif

	ptrace_notify(SIGTRAP | ((current->ptrace & PT_TRACESYSGOOD)
				? 0x80 : 0));
	/*
	 * this isn't the same as continuing with a signal, but it will do
	 * for normal use.  strace only continues with a signal if the
	 * stopping signal is not SIGTRAP.  -brl
	 */
	if (current->exit_code) {
		send_sig(current->exit_code, current, 1);
		current->exit_code = 0;
	}
}

asmlinkage void syscall_trace_enter(struct pt_regs *regs)
{
	/* do the secure computing check first */
	secure_computing(regs->orig_rax);

	if (test_thread_flag(TIF_SYSCALL_TRACE)
	    && (current->ptrace & PT_PTRACED))
		syscall_trace(regs);

	if (unlikely(current->audit_context)) {
		if (test_thread_flag(TIF_IA32)) {
			audit_syscall_entry(AUDIT_ARCH_I386,
					    regs->orig_rax,
					    regs->rbx, regs->rcx,
					    regs->rdx, regs->rsi);
		} else {
			audit_syscall_entry(AUDIT_ARCH_X86_64,
					    regs->orig_rax,
					    regs->rdi, regs->rsi,
					    regs->rdx, regs->r10);
		}
	}
}

asmlinkage void syscall_trace_leave(struct pt_regs *regs)
{
	if (unlikely(current->audit_context))
		audit_syscall_exit(AUDITSC_RESULT(regs->rax), regs->rax);

	if ((test_thread_flag(TIF_SYSCALL_TRACE)
	     || test_thread_flag(TIF_SINGLESTEP))
	    && (current->ptrace & PT_PTRACED))
		syscall_trace(regs);
}
Commit	Line	Data
1da177e4 LT	1	/* ptrace.c */
	2	/* By Ross Biro 1/23/92 */
	3	/*
	4	* Pentium III FXSR, SSE support
	5	* Gareth Hughes <gareth@valinux.com>, May 2000
	6	*
	7	* x86-64 port 2000-2002 Andi Kleen
	8	*/
	9
	10	#include <linux/kernel.h>
	11	#include <linux/sched.h>
	12	#include <linux/mm.h>
	13	#include <linux/smp.h>
1da177e4 LT	14	#include <linux/errno.h>
	15	#include <linux/ptrace.h>
	16	#include <linux/user.h>
	17	#include <linux/security.h>
	18	#include <linux/audit.h>
	19	#include <linux/seccomp.h>
7ed20e1a	20	#include <linux/signal.h>
1da177e4 LT	21
	22	#include <asm/uaccess.h>
	23	#include <asm/pgtable.h>
	24	#include <asm/system.h>
	25	#include <asm/processor.h>
	26	#include <asm/i387.h>
	27	#include <asm/debugreg.h>
	28	#include <asm/ldt.h>
	29	#include <asm/desc.h>
	30	#include <asm/proto.h>
	31	#include <asm/ia32.h>
	32
	33	/*
	34	* does not yet catch signals sent when the child dies.
	35	* in exit.c or in signal.c.
	36	*/
	37
60923df3 CE	38	/*
	39	* Determines which flags the user has access to [1 = access, 0 = no access].
	40	* Prohibits changing ID(21), VIP(20), VIF(19), VM(17), IOPL(12-13), IF(9).
	41	* Also masks reserved bits (63-22, 15, 5, 3, 1).
	42	*/
	43	#define FLAG_MASK 0x54dd5UL
1da177e4 LT	44
	45	/* set's the trap flag. */
	46	#define TRAP_FLAG 0x100UL
	47
	48	/*
	49	* eflags and offset of eflags on child stack..
	50	*/
	51	#define EFLAGS offsetof(struct pt_regs, eflags)
	52	#define EFL_OFFSET ((int)(EFLAGS-sizeof(struct pt_regs)))
	53
	54	/*
	55	* this routine will get a word off of the processes privileged stack.
	56	* the offset is how far from the base addr as stored in the TSS.
	57	* this routine assumes that all the privileged stacks are in our
	58	* data space.
	59	*/
	60	static inline unsigned long get_stack_long(struct task_struct *task, int offset)
	61	{
	62	unsigned char *stack;
	63
	64	stack = (unsigned char *)task->thread.rsp0;
	65	stack += offset;
	66	return (((unsigned long )stack));
	67	}
	68
	69	/*
	70	* this routine will put a word on the processes privileged stack.
	71	* the offset is how far from the base addr as stored in the TSS.
	72	* this routine assumes that all the privileged stacks are in our
	73	* data space.
	74	*/
	75	static inline long put_stack_long(struct task_struct *task, int offset,
	76	unsigned long data)
	77	{
	78	unsigned char * stack;
	79
	80	stack = (unsigned char *) task->thread.rsp0;
	81	stack += offset;
	82	(unsigned long ) stack = data;
	83	return 0;
	84	}
	85
e502cdd6 AK	86	#define LDT_SEGMENT 4
	87
	88	unsigned long convert_rip_to_linear(struct task_struct child, struct pt_regs regs)
	89	{
	90	unsigned long addr, seg;
	91
	92	addr = regs->rip;
	93	seg = regs->cs & 0xffff;
	94
	95	/*
	96	* We'll assume that the code segments in the GDT
	97	* are all zero-based. That is largely true: the
	98	* TLS segments are used for data, and the PNPBIOS
	99	* and APM bios ones we just ignore here.
	100	*/
	101	if (seg & LDT_SEGMENT) {
	102	u32 *desc;
	103	unsigned long base;
	104
	105	down(&child->mm->context.sem);
	106	desc = child->mm->context.ldt + (seg & ~7);
	107	base = (desc[0] >> 16) \| ((desc[1] & 0xff) << 16) \| (desc[1] & 0xff000000);
	108
	109	/* 16-bit code segment? */
	110	if (!((desc[1] >> 22) & 1))
	111	addr &= 0xffff;
	112	addr += base;
	113	up(&child->mm->context.sem);
	114	}
	115	return addr;
	116	}
	117
2ade2920	118	static int is_setting_trap_flag(struct task_struct child, struct pt_regs regs)
e502cdd6 AK	119	{
e502cdd6 AK	120	int i, copied;
2ade2920	121	unsigned char opcode[15];
e502cdd6 AK	122	unsigned long addr = convert_rip_to_linear(child, regs);
	123
	124	copied = access_process_vm(child, addr, opcode, sizeof(opcode), 0);
	125	for (i = 0; i < copied; i++) {
	126	switch (opcode[i]) {
2ade2920 CE	127	/* popf and iret */
2ade2920 CE	128	case 0x9d: case 0xcf:
e502cdd6 AK	129	return 1;
	130
	131	/* CHECKME: 64 65 */
	132
	133	/* opcode and address size prefixes */
	134	case 0x66: case 0x67:
	135	continue;
	136	/* irrelevant prefixes (segment overrides and repeats) */
	137	case 0x26: case 0x2e:
	138	case 0x36: case 0x3e:
	139	case 0x64: case 0x65:
d4d35854	140	case 0xf2: case 0xf3:
e502cdd6 AK	141	continue;
e502cdd6 AK	142
e502cdd6	143	case 0x40 ... 0x4f:
a752d719 CE	144	if (regs->cs != __USER_CS)
	145	/* 32-bit mode: register increment */
	146	return 0;
	147	/* 64-bit mode: REX prefix */
e502cdd6 AK	148	continue;
e502cdd6 AK	149
d4d35854	150	/* CHECKME: f2, f3 */
e502cdd6 AK	151
	152	/*
	153	* pushf: NOTE! We should probably not let
	154	* the user see the TF bit being set. But
	155	* it's more pain than it's worth to avoid
	156	* it, and a debugger could emulate this
	157	* all in user space if it _really_ cares.
	158	*/
	159	case 0x9c:
	160	default:
	161	return 0;
	162	}
	163	}
	164	return 0;
	165	}
	166
aa85b9af AK	167	static void set_singlestep(struct task_struct *child)
aa85b9af AK	168	{
bb049232	169	struct pt_regs *regs = task_pt_regs(child);
aa85b9af AK	170
	171	/*
	172	* Always set TIF_SINGLESTEP - this guarantees that
	173	* we single-step system calls etc.. This will also
	174	* cause us to set TF when returning to user mode.
	175	*/
	176	set_tsk_thread_flag(child, TIF_SINGLESTEP);
	177
	178	/*
	179	* If TF was already set, don't do anything else
	180	*/
	181	if (regs->eflags & TRAP_FLAG)
	182	return;
	183
	184	/* Set TF on the kernel stack.. */
	185	regs->eflags \|= TRAP_FLAG;
	186
e502cdd6 AK	187	/*
	188	* ..but if TF is changed by the instruction we will trace,
	189	* don't mark it as being "us" that set it, so that we
	190	* won't clear it by hand later.
e502cdd6	191	*/
2ade2920	192	if (is_setting_trap_flag(child, regs))
e502cdd6 AK	193	return;
e502cdd6 AK	194
aa85b9af AK	195	child->ptrace \|= PT_DTRACE;
	196	}
	197
	198	static void clear_singlestep(struct task_struct *child)
	199	{
	200	/* Always clear TIF_SINGLESTEP... */
	201	clear_tsk_thread_flag(child, TIF_SINGLESTEP);
	202
	203	/* But touch TF only if it was set by us.. */
	204	if (child->ptrace & PT_DTRACE) {
bb049232	205	struct pt_regs *regs = task_pt_regs(child);
aa85b9af AK	206	regs->eflags &= ~TRAP_FLAG;
	207	child->ptrace &= ~PT_DTRACE;
	208	}
	209	}
	210
1da177e4 LT	211	/*
	212	* Called by kernel/ptrace.c when detaching..
	213	*
	214	* Make sure the single step bit is not set.
	215	*/
	216	void ptrace_disable(struct task_struct *child)
	217	{
aa85b9af	218	clear_singlestep(child);
1da177e4 LT	219	}
	220
	221	static int putreg(struct task_struct *child,
	222	unsigned long regno, unsigned long value)
	223	{
	224	unsigned long tmp;
	225
	226	/* Some code in the 64bit emulation may not be 64bit clean.
	227	Don't take any chances. */
	228	if (test_tsk_thread_flag(child, TIF_IA32))
	229	value &= 0xffffffff;
	230	switch (regno) {
	231	case offsetof(struct user_regs_struct,fs):
	232	if (value && (value & 3) != 3)
	233	return -EIO;
	234	child->thread.fsindex = value & 0xffff;
	235	return 0;
	236	case offsetof(struct user_regs_struct,gs):
	237	if (value && (value & 3) != 3)
	238	return -EIO;
	239	child->thread.gsindex = value & 0xffff;
	240	return 0;
	241	case offsetof(struct user_regs_struct,ds):
	242	if (value && (value & 3) != 3)
	243	return -EIO;
	244	child->thread.ds = value & 0xffff;
	245	return 0;
	246	case offsetof(struct user_regs_struct,es):
	247	if (value && (value & 3) != 3)
	248	return -EIO;
	249	child->thread.es = value & 0xffff;
	250	return 0;
	251	case offsetof(struct user_regs_struct,ss):
	252	if ((value & 3) != 3)
	253	return -EIO;
	254	value &= 0xffff;
	255	return 0;
	256	case offsetof(struct user_regs_struct,fs_base):
84929801	257	if (value >= TASK_SIZE_OF(child))
f6b8d477	258	return -EIO;
1da177e4 LT	259	child->thread.fs = value;
	260	return 0;
	261	case offsetof(struct user_regs_struct,gs_base):
84929801	262	if (value >= TASK_SIZE_OF(child))
f6b8d477	263	return -EIO;
1da177e4 LT	264	child->thread.gs = value;
	265	return 0;
	266	case offsetof(struct user_regs_struct, eflags):
	267	value &= FLAG_MASK;
	268	tmp = get_stack_long(child, EFL_OFFSET);
	269	tmp &= ~FLAG_MASK;
	270	value \|= tmp;
	271	break;
	272	case offsetof(struct user_regs_struct,cs):
	273	if ((value & 3) != 3)
	274	return -EIO;
	275	value &= 0xffff;
	276	break;
	277	}
	278	put_stack_long(child, regno - sizeof(struct pt_regs), value);
	279	return 0;
	280	}
	281
	282	static unsigned long getreg(struct task_struct *child, unsigned long regno)
	283	{
	284	unsigned long val;
	285	switch (regno) {
	286	case offsetof(struct user_regs_struct, fs):
	287	return child->thread.fsindex;
	288	case offsetof(struct user_regs_struct, gs):
	289	return child->thread.gsindex;
	290	case offsetof(struct user_regs_struct, ds):
	291	return child->thread.ds;
	292	case offsetof(struct user_regs_struct, es):
	293	return child->thread.es;
	294	case offsetof(struct user_regs_struct, fs_base):
	295	return child->thread.fs;
	296	case offsetof(struct user_regs_struct, gs_base):
	297	return child->thread.gs;
	298	default:
	299	regno = regno - sizeof(struct pt_regs);
	300	val = get_stack_long(child, regno);
	301	if (test_tsk_thread_flag(child, TIF_IA32))
	302	val &= 0xffffffff;
	303	return val;
	304	}
	305
	306	}
	307
481bed45	308	long arch_ptrace(struct task_struct *child, long request, long addr, long data)
1da177e4	309	{
1da177e4 LT	310	long i, ret;
	311	unsigned ui;
	312
1da177e4 LT	313	switch (request) {
	314	/* when I and D space are separate, these will need to be fixed. */
	315	case PTRACE_PEEKTEXT: /* read word at location addr. */
76647323 AD	316	case PTRACE_PEEKDATA:
76647323 AD	317	ret = generic_ptrace_peekdata(child, addr, data);
1da177e4	318	break;
1da177e4 LT	319
	320	/* read the word at location addr in the USER area. */
	321	case PTRACE_PEEKUSR: {
	322	unsigned long tmp;
	323
	324	ret = -EIO;
	325	if ((addr & 7) \|\|
	326	addr > sizeof(struct user) - 7)
	327	break;
	328
	329	switch (addr) {
c4d1fcf3	330	case 0 ... sizeof(struct user_regs_struct) - sizeof(long):
1da177e4 LT	331	tmp = getreg(child, addr);
	332	break;
	333	case offsetof(struct user, u_debugreg[0]):
	334	tmp = child->thread.debugreg0;
	335	break;
	336	case offsetof(struct user, u_debugreg[1]):
	337	tmp = child->thread.debugreg1;
	338	break;
	339	case offsetof(struct user, u_debugreg[2]):
	340	tmp = child->thread.debugreg2;
	341	break;
	342	case offsetof(struct user, u_debugreg[3]):
	343	tmp = child->thread.debugreg3;
	344	break;
	345	case offsetof(struct user, u_debugreg[6]):
	346	tmp = child->thread.debugreg6;
	347	break;
	348	case offsetof(struct user, u_debugreg[7]):
	349	tmp = child->thread.debugreg7;
	350	break;
	351	default:
	352	tmp = 0;
	353	break;
	354	}
	355	ret = put_user(tmp,(unsigned long __user *) data);
	356	break;
	357	}
	358
	359	/* when I and D space are separate, this will have to be fixed. */
	360	case PTRACE_POKETEXT: /* write the word at location addr. */
	361	case PTRACE_POKEDATA:
	362	ret = 0;
	363	if (access_process_vm(child, addr, &data, sizeof(data), 1) == sizeof(data))
	364	break;
	365	ret = -EIO;
	366	break;
	367
	368	case PTRACE_POKEUSR: /* write the word at location addr in the USER area */
84929801 SS	369	{
84929801 SS	370	int dsize = test_tsk_thread_flag(child, TIF_IA32) ? 3 : 7;
1da177e4 LT	371	ret = -EIO;
	372	if ((addr & 7) \|\|
	373	addr > sizeof(struct user) - 7)
	374	break;
	375
	376	switch (addr) {
c4d1fcf3	377	case 0 ... sizeof(struct user_regs_struct) - sizeof(long):
1da177e4 LT	378	ret = putreg(child, addr, data);
	379	break;
	380	/* Disallows to set a breakpoint into the vsyscall */
	381	case offsetof(struct user, u_debugreg[0]):
84929801	382	if (data >= TASK_SIZE_OF(child) - dsize) break;
1da177e4 LT	383	child->thread.debugreg0 = data;
	384	ret = 0;
	385	break;
	386	case offsetof(struct user, u_debugreg[1]):
84929801	387	if (data >= TASK_SIZE_OF(child) - dsize) break;
1da177e4 LT	388	child->thread.debugreg1 = data;
	389	ret = 0;
	390	break;
	391	case offsetof(struct user, u_debugreg[2]):
84929801	392	if (data >= TASK_SIZE_OF(child) - dsize) break;
1da177e4 LT	393	child->thread.debugreg2 = data;
	394	ret = 0;
	395	break;
	396	case offsetof(struct user, u_debugreg[3]):
84929801	397	if (data >= TASK_SIZE_OF(child) - dsize) break;
1da177e4 LT	398	child->thread.debugreg3 = data;
	399	ret = 0;
	400	break;
	401	case offsetof(struct user, u_debugreg[6]):
	402	if (data >> 32)
	403	break;
	404	child->thread.debugreg6 = data;
	405	ret = 0;
	406	break;
	407	case offsetof(struct user, u_debugreg[7]):
	408	/* See arch/i386/kernel/ptrace.c for an explanation of
	409	* this awkward check.*/
893efca9 JB	410	data &= ~DR_CONTROL_RESERVED;
	411	for(i=0; i<4; i++)
	412	if ((0x5554 >> ((data >> (16 + 4*i)) & 0xf)) & 1)
1da177e4 LT	413	break;
1da177e4 LT	414	if (i == 4) {
d3a4f48d SE	415	child->thread.debugreg7 = data;
	416	if (data)
	417	set_tsk_thread_flag(child, TIF_DEBUG);
	418	else
	419	clear_tsk_thread_flag(child, TIF_DEBUG);
1da177e4	420	ret = 0;
d3a4f48d	421	}
1da177e4 LT	422	break;
	423	}
	424	break;
84929801	425	}
1da177e4	426	case PTRACE_SYSCALL: /* continue and stop at next (return from) syscall */
aa85b9af	427	case PTRACE_CONT: /* restart after signal. */
1da177e4 LT	428
1da177e4 LT	429	ret = -EIO;
7ed20e1a	430	if (!valid_signal(data))
1da177e4 LT	431	break;
	432	if (request == PTRACE_SYSCALL)
	433	set_tsk_thread_flag(child,TIF_SYSCALL_TRACE);
	434	else
	435	clear_tsk_thread_flag(child,TIF_SYSCALL_TRACE);
	436	clear_tsk_thread_flag(child, TIF_SINGLESTEP);
	437	child->exit_code = data;
aa85b9af AK	438	/* make sure the single step bit is not set. */
aa85b9af AK	439	clear_singlestep(child);
1da177e4 LT	440	wake_up_process(child);
	441	ret = 0;
	442	break;
1da177e4 LT	443
	444	#ifdef CONFIG_IA32_EMULATION
	445	/* This makes only sense with 32bit programs. Allow a
	446	64bit debugger to fully examine them too. Better
	447	don't use it against 64bit processes, use
	448	PTRACE_ARCH_PRCTL instead. */
	449	case PTRACE_SET_THREAD_AREA: {
	450	struct user_desc __user *p;
	451	int old;
	452	p = (struct user_desc __user *)data;
	453	get_user(old, &p->entry_number);
	454	put_user(addr, &p->entry_number);
	455	ret = do_set_thread_area(&child->thread, p);
	456	put_user(old, &p->entry_number);
	457	break;
	458	case PTRACE_GET_THREAD_AREA:
	459	p = (struct user_desc __user *)data;
	460	get_user(old, &p->entry_number);
	461	put_user(addr, &p->entry_number);
	462	ret = do_get_thread_area(&child->thread, p);
	463	put_user(old, &p->entry_number);
	464	break;
	465	}
	466	#endif
	467	/* normal 64bit interface to access TLS data.
	468	Works just like arch_prctl, except that the arguments
	469	are reversed. */
	470	case PTRACE_ARCH_PRCTL:
	471	ret = do_arch_prctl(child, data, addr);
	472	break;
	473
	474	/*
	475	* make the child exit. Best I can do is send it a sigkill.
	476	* perhaps it should be put in the status that it wants to
	477	* exit.
	478	*/
aa85b9af	479	case PTRACE_KILL:
1da177e4 LT	480	ret = 0;
	481	if (child->exit_state == EXIT_ZOMBIE) /* already dead */
	482	break;
	483	clear_tsk_thread_flag(child, TIF_SINGLESTEP);
	484	child->exit_code = SIGKILL;
	485	/* make sure the single step bit is not set. */
aa85b9af	486	clear_singlestep(child);
1da177e4 LT	487	wake_up_process(child);
1da177e4 LT	488	break;
1da177e4	489
aa85b9af	490	case PTRACE_SINGLESTEP: /* set the trap flag. */
1da177e4	491	ret = -EIO;
7ed20e1a	492	if (!valid_signal(data))
1da177e4 LT	493	break;
1da177e4 LT	494	clear_tsk_thread_flag(child,TIF_SYSCALL_TRACE);
aa85b9af	495	set_singlestep(child);
1da177e4 LT	496	child->exit_code = data;
	497	/* give it a chance to run. */
	498	wake_up_process(child);
	499	ret = 0;
	500	break;
1da177e4 LT	501
	502	case PTRACE_DETACH:
	503	/* detach a process that was attached. */
	504	ret = ptrace_detach(child, data);
	505	break;
	506
	507	case PTRACE_GETREGS: { /* Get all gp regs from the child. */
	508	if (!access_ok(VERIFY_WRITE, (unsigned __user *)data,
	509	sizeof(struct user_regs_struct))) {
	510	ret = -EIO;
	511	break;
	512	}
	513	ret = 0;
	514	for (ui = 0; ui < sizeof(struct user_regs_struct); ui += sizeof(long)) {
	515	ret \|= __put_user(getreg(child, ui),(unsigned long __user *) data);
	516	data += sizeof(long);
	517	}
	518	break;
	519	}
	520
	521	case PTRACE_SETREGS: { /* Set all gp regs in the child. */
	522	unsigned long tmp;
	523	if (!access_ok(VERIFY_READ, (unsigned __user *)data,
	524	sizeof(struct user_regs_struct))) {
	525	ret = -EIO;
	526	break;
	527	}
	528	ret = 0;
	529	for (ui = 0; ui < sizeof(struct user_regs_struct); ui += sizeof(long)) {
f49481bc AK	530	ret = __get_user(tmp, (unsigned long __user *) data);
	531	if (ret)
	532	break;
	533	ret = putreg(child, ui, tmp);
	534	if (ret)
	535	break;
1da177e4 LT	536	data += sizeof(long);
	537	}
	538	break;
	539	}
	540
	541	case PTRACE_GETFPREGS: { /* Get the child extended FPU state. */
	542	if (!access_ok(VERIFY_WRITE, (unsigned __user *)data,
	543	sizeof(struct user_i387_struct))) {
	544	ret = -EIO;
	545	break;
	546	}
	547	ret = get_fpregs((struct user_i387_struct __user *)data, child);
	548	break;
	549	}
	550
	551	case PTRACE_SETFPREGS: { /* Set the child extended FPU state. */
	552	if (!access_ok(VERIFY_READ, (unsigned __user *)data,
	553	sizeof(struct user_i387_struct))) {
	554	ret = -EIO;
	555	break;
	556	}
	557	set_stopped_child_used_math(child);
	558	ret = set_fpregs(child, (struct user_i387_struct __user *)data);
	559	break;
	560	}
	561
	562	default:
	563	ret = ptrace_request(child, request, addr, data);
	564	break;
	565	}
1da177e4 LT	566	return ret;
	567	}
	568
	569	static void syscall_trace(struct pt_regs *regs)
	570	{
	571
	572	#if 0
	573	printk("trace %s rip %lx rsp %lx rax %d origrax %d caller %lx tiflags %x ptrace %x\n",
	574	current->comm,
	575	regs->rip, regs->rsp, regs->rax, regs->orig_rax, __builtin_return_address(0),
	576	current_thread_info()->flags, current->ptrace);
	577	#endif
	578
	579	ptrace_notify(SIGTRAP \| ((current->ptrace & PT_TRACESYSGOOD)
	580	? 0x80 : 0));
	581	/*
	582	* this isn't the same as continuing with a signal, but it will do
	583	* for normal use. strace only continues with a signal if the
	584	* stopping signal is not SIGTRAP. -brl
	585	*/
	586	if (current->exit_code) {
	587	send_sig(current->exit_code, current, 1);
	588	current->exit_code = 0;
	589	}
	590	}
	591
	592	asmlinkage void syscall_trace_enter(struct pt_regs *regs)
	593	{
	594	/* do the secure computing check first */
	595	secure_computing(regs->orig_rax);
	596
2fd6f58b DW	597	if (test_thread_flag(TIF_SYSCALL_TRACE)
	598	&& (current->ptrace & PT_PTRACED))
	599	syscall_trace(regs);
	600
488f2eac DW	601	if (unlikely(current->audit_context)) {
488f2eac DW	602	if (test_thread_flag(TIF_IA32)) {
5411be59	603	audit_syscall_entry(AUDIT_ARCH_I386,
488f2eac DW	604	regs->orig_rax,
	605	regs->rbx, regs->rcx,
	606	regs->rdx, regs->rsi);
	607	} else {
5411be59	608	audit_syscall_entry(AUDIT_ARCH_X86_64,
488f2eac DW	609	regs->orig_rax,
	610	regs->rdi, regs->rsi,
	611	regs->rdx, regs->r10);
	612	}
	613	}
1da177e4 LT	614	}
	615
	616	asmlinkage void syscall_trace_leave(struct pt_regs *regs)
	617	{
	618	if (unlikely(current->audit_context))
5411be59	619	audit_syscall_exit(AUDITSC_RESULT(regs->rax), regs->rax);
1da177e4 LT	620
	621	if ((test_thread_flag(TIF_SYSCALL_TRACE)
	622	\|\| test_thread_flag(TIF_SINGLESTEP))
	623	&& (current->ptrace & PT_PTRACED))
	624	syscall_trace(regs);
	625	}