[deliverable/linux.git] / arch / x86 / entry / common.c

/*
 * common.c - C code for kernel entry and exit
 * Copyright (c) 2015 Andrew Lutomirski
 * GPL v2
 *
 * Based on asm and ptrace code by many authors.  The code here originated
 * in ptrace.c and signal.c.
 */

#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/tracehook.h>
#include <linux/audit.h>
#include <linux/seccomp.h>
#include <linux/signal.h>
#include <linux/export.h>
#include <linux/context_tracking.h>
#include <linux/user-return-notifier.h>
#include <linux/uprobes.h>

#include <asm/desc.h>
#include <asm/traps.h>
#include <asm/vdso.h>
#include <asm/uaccess.h>
#include <asm/cpufeature.h>

#define CREATE_TRACE_POINTS
#include <trace/events/syscalls.h>

static struct thread_info *pt_regs_to_thread_info(struct pt_regs *regs)
{
	unsigned long top_of_stack =
		(unsigned long)(regs + 1) + TOP_OF_KERNEL_STACK_PADDING;
	return (struct thread_info *)(top_of_stack - THREAD_SIZE);
}

#ifdef CONFIG_CONTEXT_TRACKING
/* Called on entry from user mode with IRQs off. */
__visible void enter_from_user_mode(void)
{
	CT_WARN_ON(ct_state() != CONTEXT_USER);
	user_exit();
}
#endif

static void do_audit_syscall_entry(struct pt_regs *regs, u32 arch)
{
#ifdef CONFIG_X86_64
	if (arch == AUDIT_ARCH_X86_64) {
		audit_syscall_entry(regs->orig_ax, regs->di,
				    regs->si, regs->dx, regs->r10);
	} else
#endif
	{
		audit_syscall_entry(regs->orig_ax, regs->bx,
				    regs->cx, regs->dx, regs->si);
	}
}

/*
 * We can return 0 to resume the syscall or anything else to go to phase
 * 2.  If we resume the syscall, we need to put something appropriate in
 * regs->orig_ax.
 *
 * NB: We don't have full pt_regs here, but regs->orig_ax and regs->ax
 * are fully functional.
 *
 * For phase 2's benefit, our return value is:
 * 0:			resume the syscall
 * 1:			go to phase 2; no seccomp phase 2 needed
 * anything else:	go to phase 2; pass return value to seccomp
 */
unsigned long syscall_trace_enter_phase1(struct pt_regs *regs, u32 arch)
{
	struct thread_info *ti = pt_regs_to_thread_info(regs);
	unsigned long ret = 0;
	u32 work;

	if (IS_ENABLED(CONFIG_DEBUG_ENTRY))
		BUG_ON(regs != task_pt_regs(current));

	work = ACCESS_ONCE(ti->flags) & _TIF_WORK_SYSCALL_ENTRY;

#ifdef CONFIG_CONTEXT_TRACKING
	/*
	 * If TIF_NOHZ is set, we are required to call user_exit() before
	 * doing anything that could touch RCU.
	 */
	if (work & _TIF_NOHZ) {
		enter_from_user_mode();
		work &= ~_TIF_NOHZ;
	}
#endif

#ifdef CONFIG_SECCOMP
	/*
	 * Do seccomp first -- it should minimize exposure of other
	 * code, and keeping seccomp fast is probably more valuable
	 * than the rest of this.
	 */
	if (work & _TIF_SECCOMP) {
		struct seccomp_data sd;

		sd.arch = arch;
		sd.nr = regs->orig_ax;
		sd.instruction_pointer = regs->ip;
#ifdef CONFIG_X86_64
		if (arch == AUDIT_ARCH_X86_64) {
			sd.args[0] = regs->di;
			sd.args[1] = regs->si;
			sd.args[2] = regs->dx;
			sd.args[3] = regs->r10;
			sd.args[4] = regs->r8;
			sd.args[5] = regs->r9;
		} else
#endif
		{
			sd.args[0] = regs->bx;
			sd.args[1] = regs->cx;
			sd.args[2] = regs->dx;
			sd.args[3] = regs->si;
			sd.args[4] = regs->di;
			sd.args[5] = regs->bp;
		}

		BUILD_BUG_ON(SECCOMP_PHASE1_OK != 0);
		BUILD_BUG_ON(SECCOMP_PHASE1_SKIP != 1);

		ret = seccomp_phase1(&sd);
		if (ret == SECCOMP_PHASE1_SKIP) {
			regs->orig_ax = -1;
			ret = 0;
		} else if (ret != SECCOMP_PHASE1_OK) {
			return ret;  /* Go directly to phase 2 */
		}

		work &= ~_TIF_SECCOMP;
	}
#endif

	/* Do our best to finish without phase 2. */
	if (work == 0)
		return ret;  /* seccomp and/or nohz only (ret == 0 here) */

#ifdef CONFIG_AUDITSYSCALL
	if (work == _TIF_SYSCALL_AUDIT) {
		/*
		 * If there is no more work to be done except auditing,
		 * then audit in phase 1.  Phase 2 always audits, so, if
		 * we audit here, then we can't go on to phase 2.
		 */
		do_audit_syscall_entry(regs, arch);
		return 0;
	}
#endif

	return 1;  /* Something is enabled that we can't handle in phase 1 */
}

/* Returns the syscall nr to run (which should match regs->orig_ax). */
long syscall_trace_enter_phase2(struct pt_regs *regs, u32 arch,
				unsigned long phase1_result)
{
	struct thread_info *ti = pt_regs_to_thread_info(regs);
	long ret = 0;
	u32 work = ACCESS_ONCE(ti->flags) & _TIF_WORK_SYSCALL_ENTRY;

	if (IS_ENABLED(CONFIG_DEBUG_ENTRY))
		BUG_ON(regs != task_pt_regs(current));

#ifdef CONFIG_SECCOMP
	/*
	 * Call seccomp_phase2 before running the other hooks so that
	 * they can see any changes made by a seccomp tracer.
	 */
	if (phase1_result > 1 && seccomp_phase2(phase1_result)) {
		/* seccomp failures shouldn't expose any additional code. */
		return -1;
	}
#endif

	if (unlikely(work & _TIF_SYSCALL_EMU))
		ret = -1L;

	if ((ret || test_thread_flag(TIF_SYSCALL_TRACE)) &&
	    tracehook_report_syscall_entry(regs))
		ret = -1L;

	if (unlikely(test_thread_flag(TIF_SYSCALL_TRACEPOINT)))
		trace_sys_enter(regs, regs->orig_ax);

	do_audit_syscall_entry(regs, arch);

	return ret ?: regs->orig_ax;
}

long syscall_trace_enter(struct pt_regs *regs)
{
	u32 arch = is_ia32_task() ? AUDIT_ARCH_I386 : AUDIT_ARCH_X86_64;
	unsigned long phase1_result = syscall_trace_enter_phase1(regs, arch);

	if (phase1_result == 0)
		return regs->orig_ax;
	else
		return syscall_trace_enter_phase2(regs, arch, phase1_result);
}

#define EXIT_TO_USERMODE_LOOP_FLAGS				\
	(_TIF_SIGPENDING | _TIF_NOTIFY_RESUME | _TIF_UPROBE |	\
	 _TIF_NEED_RESCHED | _TIF_USER_RETURN_NOTIFY)

static void exit_to_usermode_loop(struct pt_regs *regs, u32 cached_flags)
{
	/*
	 * In order to return to user mode, we need to have IRQs off with
	 * none of _TIF_SIGPENDING, _TIF_NOTIFY_RESUME, _TIF_USER_RETURN_NOTIFY,
	 * _TIF_UPROBE, or _TIF_NEED_RESCHED set.  Several of these flags
	 * can be set at any time on preemptable kernels if we have IRQs on,
	 * so we need to loop.  Disabling preemption wouldn't help: doing the
	 * work to clear some of the flags can sleep.
	 */
	while (true) {
		/* We have work to do. */
		local_irq_enable();

		if (cached_flags & _TIF_NEED_RESCHED)
			schedule();

		if (cached_flags & _TIF_UPROBE)
			uprobe_notify_resume(regs);

		/* deal with pending signal delivery */
		if (cached_flags & _TIF_SIGPENDING)
			do_signal(regs);

		if (cached_flags & _TIF_NOTIFY_RESUME) {
			clear_thread_flag(TIF_NOTIFY_RESUME);
			tracehook_notify_resume(regs);
		}

		if (cached_flags & _TIF_USER_RETURN_NOTIFY)
			fire_user_return_notifiers();

		/* Disable IRQs and retry */
		local_irq_disable();

		cached_flags = READ_ONCE(pt_regs_to_thread_info(regs)->flags);

		if (!(cached_flags & EXIT_TO_USERMODE_LOOP_FLAGS))
			break;

	}
}

/* Called with IRQs disabled. */
__visible inline void prepare_exit_to_usermode(struct pt_regs *regs)
{
	struct thread_info *ti = pt_regs_to_thread_info(regs);
	u32 cached_flags;

	if (IS_ENABLED(CONFIG_PROVE_LOCKING) && WARN_ON(!irqs_disabled()))
		local_irq_disable();

	lockdep_sys_exit();

	cached_flags = READ_ONCE(ti->flags);

	if (unlikely(cached_flags & EXIT_TO_USERMODE_LOOP_FLAGS))
		exit_to_usermode_loop(regs, cached_flags);

#ifdef CONFIG_COMPAT
	/*
	 * Compat syscalls set TS_COMPAT.  Make sure we clear it before
	 * returning to user mode.  We need to clear it *after* signal
	 * handling, because syscall restart has a fixup for compat
	 * syscalls.  The fixup is exercised by the ptrace_syscall_32
	 * selftest.
	 */
	ti->status &= ~TS_COMPAT;
#endif

	user_enter();
}

#define SYSCALL_EXIT_WORK_FLAGS				\
	(_TIF_SYSCALL_TRACE | _TIF_SYSCALL_AUDIT |	\
	 _TIF_SINGLESTEP | _TIF_SYSCALL_TRACEPOINT)

static void syscall_slow_exit_work(struct pt_regs *regs, u32 cached_flags)
{
	bool step;

	audit_syscall_exit(regs);

	if (cached_flags & _TIF_SYSCALL_TRACEPOINT)
		trace_sys_exit(regs, regs->ax);

	/*
	 * If TIF_SYSCALL_EMU is set, we only get here because of
	 * TIF_SINGLESTEP (i.e. this is PTRACE_SYSEMU_SINGLESTEP).
	 * We already reported this syscall instruction in
	 * syscall_trace_enter().
	 */
	step = unlikely(
		(cached_flags & (_TIF_SINGLESTEP | _TIF_SYSCALL_EMU))
		== _TIF_SINGLESTEP);
	if (step || cached_flags & _TIF_SYSCALL_TRACE)
		tracehook_report_syscall_exit(regs, step);
}

/*
 * Called with IRQs on and fully valid regs.  Returns with IRQs off in a
 * state such that we can immediately switch to user mode.
 */
__visible inline void syscall_return_slowpath(struct pt_regs *regs)
{
	struct thread_info *ti = pt_regs_to_thread_info(regs);
	u32 cached_flags = READ_ONCE(ti->flags);

	CT_WARN_ON(ct_state() != CONTEXT_KERNEL);

	if (IS_ENABLED(CONFIG_PROVE_LOCKING) &&
	    WARN(irqs_disabled(), "syscall %ld left IRQs disabled", regs->orig_ax))
		local_irq_enable();

	/*
	 * First do one-time work.  If these work items are enabled, we
	 * want to run them exactly once per syscall exit with IRQs on.
	 */
	if (unlikely(cached_flags & SYSCALL_EXIT_WORK_FLAGS))
		syscall_slow_exit_work(regs, cached_flags);

	local_irq_disable();
	prepare_exit_to_usermode(regs);
}

#ifdef CONFIG_X86_64
__visible void do_syscall_64(struct pt_regs *regs)
{
	struct thread_info *ti = pt_regs_to_thread_info(regs);
	unsigned long nr = regs->orig_ax;

	local_irq_enable();

	if (READ_ONCE(ti->flags) & _TIF_WORK_SYSCALL_ENTRY)
		nr = syscall_trace_enter(regs);

	/*
	 * NB: Native and x32 syscalls are dispatched from the same
	 * table.  The only functional difference is the x32 bit in
	 * regs->orig_ax, which changes the behavior of some syscalls.
	 */
	if (likely((nr & __SYSCALL_MASK) < NR_syscalls)) {
		regs->ax = sys_call_table[nr & __SYSCALL_MASK](
			regs->di, regs->si, regs->dx,
			regs->r10, regs->r8, regs->r9);
	}

	syscall_return_slowpath(regs);
}
#endif

#if defined(CONFIG_X86_32) || defined(CONFIG_IA32_EMULATION)
/*
 * Does a 32-bit syscall.  Called with IRQs on and does all entry and
 * exit work and returns with IRQs off.  This function is extremely hot
 * in workloads that use it, and it's usually called from
 * do_fast_syscall_32, so forcibly inline it to improve performance.
 */
#ifdef CONFIG_X86_32
/* 32-bit kernels use a trap gate for INT80, and the asm code calls here. */
__visible
#else
/* 64-bit kernels use do_syscall_32_irqs_off() instead. */
static
#endif
__always_inline void do_syscall_32_irqs_on(struct pt_regs *regs)
{
	struct thread_info *ti = pt_regs_to_thread_info(regs);
	unsigned int nr = (unsigned int)regs->orig_ax;

#ifdef CONFIG_IA32_EMULATION
	ti->status |= TS_COMPAT;
#endif

	if (READ_ONCE(ti->flags) & _TIF_WORK_SYSCALL_ENTRY) {
		/*
		 * Subtlety here: if ptrace pokes something larger than
		 * 2^32-1 into orig_ax, this truncates it.  This may or
		 * may not be necessary, but it matches the old asm
		 * behavior.
		 */
		nr = syscall_trace_enter(regs);
	}

	if (likely(nr < IA32_NR_syscalls)) {
		/*
		 * It's possible that a 32-bit syscall implementation
		 * takes a 64-bit parameter but nonetheless assumes that
		 * the high bits are zero.  Make sure we zero-extend all
		 * of the args.
		 */
		regs->ax = ia32_sys_call_table[nr](
			(unsigned int)regs->bx, (unsigned int)regs->cx,
			(unsigned int)regs->dx, (unsigned int)regs->si,
			(unsigned int)regs->di, (unsigned int)regs->bp);
	}

	syscall_return_slowpath(regs);
}

#ifdef CONFIG_X86_64
/* Handles INT80 on 64-bit kernels */
__visible void do_syscall_32_irqs_off(struct pt_regs *regs)
{
	local_irq_enable();
	do_syscall_32_irqs_on(regs);
}
#endif

/* Returns 0 to return using IRET or 1 to return using SYSEXIT/SYSRETL. */
__visible long do_fast_syscall_32(struct pt_regs *regs)
{
	/*
	 * Called using the internal vDSO SYSENTER/SYSCALL32 calling
	 * convention.  Adjust regs so it looks like we entered using int80.
	 */

	unsigned long landing_pad = (unsigned long)current->mm->context.vdso +
		vdso_image_32.sym_int80_landing_pad;

	/*
	 * SYSENTER loses EIP, and even SYSCALL32 needs us to skip forward
	 * so that 'regs->ip -= 2' lands back on an int $0x80 instruction.
	 * Fix it up.
	 */
	regs->ip = landing_pad;

	/*
	 * Fetch EBP from where the vDSO stashed it.
	 *
	 * WARNING: We are in CONTEXT_USER and RCU isn't paying attention!
	 */
	local_irq_enable();
	if (
#ifdef CONFIG_X86_64
		/*
		 * Micro-optimization: the pointer we're following is explicitly
		 * 32 bits, so it can't be out of range.
		 */
		__get_user(*(u32 *)&regs->bp,
			    (u32 __user __force *)(unsigned long)(u32)regs->sp)
#else
		get_user(*(u32 *)&regs->bp,
			 (u32 __user __force *)(unsigned long)(u32)regs->sp)
#endif
		) {

		/* User code screwed up. */
		local_irq_disable();
		regs->ax = -EFAULT;
#ifdef CONFIG_CONTEXT_TRACKING
		enter_from_user_mode();
#endif
		prepare_exit_to_usermode(regs);
		return 0;	/* Keep it simple: use IRET. */
	}

	/* Now this is just like a normal syscall. */
	do_syscall_32_irqs_on(regs);

#ifdef CONFIG_X86_64
	/*
	 * Opportunistic SYSRETL: if possible, try to return using SYSRETL.
	 * SYSRETL is available on all 64-bit CPUs, so we don't need to
	 * bother with SYSEXIT.
	 *
	 * Unlike 64-bit opportunistic SYSRET, we can't check that CX == IP,
	 * because the ECX fixup above will ensure that this is essentially
	 * never the case.
	 */
	return regs->cs == __USER32_CS && regs->ss == __USER_DS &&
		regs->ip == landing_pad &&
		(regs->flags & (X86_EFLAGS_RF | X86_EFLAGS_TF)) == 0;
#else
	/*
	 * Opportunistic SYSEXIT: if possible, try to return using SYSEXIT.
	 *
	 * Unlike 64-bit opportunistic SYSRET, we can't check that CX == IP,
	 * because the ECX fixup above will ensure that this is essentially
	 * never the case.
	 *
	 * We don't allow syscalls at all from VM86 mode, but we still
	 * need to check VM, because we might be returning from sys_vm86.
	 */
	return static_cpu_has(X86_FEATURE_SEP) &&
		regs->cs == __USER_CS && regs->ss == __USER_DS &&
		regs->ip == landing_pad &&
		(regs->flags & (X86_EFLAGS_RF | X86_EFLAGS_TF | X86_EFLAGS_VM)) == 0;
#endif
}
#endif
Commit	Line	Data
1f484aa6 AL	1	/*
	2	* common.c - C code for kernel entry and exit
	3	* Copyright (c) 2015 Andrew Lutomirski
	4	* GPL v2
	5	*
	6	* Based on asm and ptrace code by many authors. The code here originated
	7	* in ptrace.c and signal.c.
	8	*/
	9
	10	#include <linux/kernel.h>
	11	#include <linux/sched.h>
	12	#include <linux/mm.h>
	13	#include <linux/smp.h>
	14	#include <linux/errno.h>
	15	#include <linux/ptrace.h>
	16	#include <linux/tracehook.h>
	17	#include <linux/audit.h>
	18	#include <linux/seccomp.h>
	19	#include <linux/signal.h>
	20	#include <linux/export.h>
	21	#include <linux/context_tracking.h>
	22	#include <linux/user-return-notifier.h>
	23	#include <linux/uprobes.h>
	24
	25	#include <asm/desc.h>
	26	#include <asm/traps.h>
710246df AL	27	#include <asm/vdso.h>
710246df AL	28	#include <asm/uaccess.h>
cd4d09ec	29	#include <asm/cpufeature.h>
1f484aa6 AL	30
	31	#define CREATE_TRACE_POINTS
	32	#include <trace/events/syscalls.h>
	33
dd636071 AL	34	static struct thread_info pt_regs_to_thread_info(struct pt_regs regs)
	35	{
	36	unsigned long top_of_stack =
	37	(unsigned long)(regs + 1) + TOP_OF_KERNEL_STACK_PADDING;
	38	return (struct thread_info *)(top_of_stack - THREAD_SIZE);
	39	}
	40
feed36cd AL	41	#ifdef CONFIG_CONTEXT_TRACKING
	42	/* Called on entry from user mode with IRQs off. */
	43	__visible void enter_from_user_mode(void)
	44	{
	45	CT_WARN_ON(ct_state() != CONTEXT_USER);
	46	user_exit();
	47	}
	48	#endif
	49
1f484aa6 AL	50	static void do_audit_syscall_entry(struct pt_regs *regs, u32 arch)
	51	{
	52	#ifdef CONFIG_X86_64
	53	if (arch == AUDIT_ARCH_X86_64) {
	54	audit_syscall_entry(regs->orig_ax, regs->di,
	55	regs->si, regs->dx, regs->r10);
	56	} else
	57	#endif
	58	{
	59	audit_syscall_entry(regs->orig_ax, regs->bx,
	60	regs->cx, regs->dx, regs->si);
	61	}
	62	}
	63
	64	/*
	65	* We can return 0 to resume the syscall or anything else to go to phase
	66	* 2. If we resume the syscall, we need to put something appropriate in
	67	* regs->orig_ax.
	68	*
	69	* NB: We don't have full pt_regs here, but regs->orig_ax and regs->ax
	70	* are fully functional.
	71	*
	72	* For phase 2's benefit, our return value is:
	73	* 0: resume the syscall
	74	* 1: go to phase 2; no seccomp phase 2 needed
	75	* anything else: go to phase 2; pass return value to seccomp
	76	*/
	77	unsigned long syscall_trace_enter_phase1(struct pt_regs *regs, u32 arch)
	78	{
dd636071	79	struct thread_info *ti = pt_regs_to_thread_info(regs);
1f484aa6 AL	80	unsigned long ret = 0;
	81	u32 work;
	82
4aabd140 AL	83	if (IS_ENABLED(CONFIG_DEBUG_ENTRY))
4aabd140 AL	84	BUG_ON(regs != task_pt_regs(current));
1f484aa6	85
dd636071	86	work = ACCESS_ONCE(ti->flags) & _TIF_WORK_SYSCALL_ENTRY;
1f484aa6	87
feed36cd	88	#ifdef CONFIG_CONTEXT_TRACKING
1f484aa6 AL	89	/*
	90	* If TIF_NOHZ is set, we are required to call user_exit() before
	91	* doing anything that could touch RCU.
	92	*/
	93	if (work & _TIF_NOHZ) {
feed36cd	94	enter_from_user_mode();
1f484aa6 AL	95	work &= ~_TIF_NOHZ;
1f484aa6 AL	96	}
feed36cd	97	#endif
1f484aa6 AL	98
	99	#ifdef CONFIG_SECCOMP
	100	/*
	101	* Do seccomp first -- it should minimize exposure of other
	102	* code, and keeping seccomp fast is probably more valuable
	103	* than the rest of this.
	104	*/
	105	if (work & _TIF_SECCOMP) {
	106	struct seccomp_data sd;
	107
	108	sd.arch = arch;
	109	sd.nr = regs->orig_ax;
	110	sd.instruction_pointer = regs->ip;
	111	#ifdef CONFIG_X86_64
	112	if (arch == AUDIT_ARCH_X86_64) {
	113	sd.args[0] = regs->di;
	114	sd.args[1] = regs->si;
	115	sd.args[2] = regs->dx;
	116	sd.args[3] = regs->r10;
	117	sd.args[4] = regs->r8;
	118	sd.args[5] = regs->r9;
	119	} else
	120	#endif
	121	{
	122	sd.args[0] = regs->bx;
	123	sd.args[1] = regs->cx;
	124	sd.args[2] = regs->dx;
	125	sd.args[3] = regs->si;
	126	sd.args[4] = regs->di;
	127	sd.args[5] = regs->bp;
	128	}
	129
	130	BUILD_BUG_ON(SECCOMP_PHASE1_OK != 0);
	131	BUILD_BUG_ON(SECCOMP_PHASE1_SKIP != 1);
	132
	133	ret = seccomp_phase1(&sd);
	134	if (ret == SECCOMP_PHASE1_SKIP) {
	135	regs->orig_ax = -1;
	136	ret = 0;
	137	} else if (ret != SECCOMP_PHASE1_OK) {
	138	return ret; /* Go directly to phase 2 */
	139	}
	140
	141	work &= ~_TIF_SECCOMP;
	142	}
	143	#endif
	144
	145	/* Do our best to finish without phase 2. */
	146	if (work == 0)
	147	return ret; /* seccomp and/or nohz only (ret == 0 here) */
	148
	149	#ifdef CONFIG_AUDITSYSCALL
	150	if (work == _TIF_SYSCALL_AUDIT) {
	151	/*
	152	* If there is no more work to be done except auditing,
	153	* then audit in phase 1. Phase 2 always audits, so, if
	154	* we audit here, then we can't go on to phase 2.
	155	*/
	156	do_audit_syscall_entry(regs, arch);
	157	return 0;
	158	}
	159	#endif
	160
	161	return 1; /* Something is enabled that we can't handle in phase 1 */
162	}
163
164	/* Returns the syscall nr to run (which should match regs->orig_ax). */
165	long syscall_trace_enter_phase2(struct pt_regs *regs, u32 arch,
166	unsigned long phase1_result)
167	{
dd636071	168	struct thread_info *ti = pt_regs_to_thread_info(regs);
1f484aa6	169	long ret = 0;
dd636071	170	u32 work = ACCESS_ONCE(ti->flags) & _TIF_WORK_SYSCALL_ENTRY;
1f484aa6	171
4aabd140 AL	172	if (IS_ENABLED(CONFIG_DEBUG_ENTRY))
4aabd140 AL	173	BUG_ON(regs != task_pt_regs(current));
1f484aa6	174
1f484aa6 AL	175	#ifdef CONFIG_SECCOMP
	176	/*
	177	* Call seccomp_phase2 before running the other hooks so that
	178	* they can see any changes made by a seccomp tracer.
	179	*/
	180	if (phase1_result > 1 && seccomp_phase2(phase1_result)) {
	181	/* seccomp failures shouldn't expose any additional code. */
	182	return -1;
	183	}
	184	#endif
	185
	186	if (unlikely(work & _TIF_SYSCALL_EMU))
	187	ret = -1L;
	188
	189	if ((ret \|\| test_thread_flag(TIF_SYSCALL_TRACE)) &&
	190	tracehook_report_syscall_entry(regs))
	191	ret = -1L;
	192
	193	if (unlikely(test_thread_flag(TIF_SYSCALL_TRACEPOINT)))
	194	trace_sys_enter(regs, regs->orig_ax);
	195
	196	do_audit_syscall_entry(regs, arch);
	197
	198	return ret ?: regs->orig_ax;
	199	}
	200
	201	long syscall_trace_enter(struct pt_regs *regs)
	202	{
	203	u32 arch = is_ia32_task() ? AUDIT_ARCH_I386 : AUDIT_ARCH_X86_64;
	204	unsigned long phase1_result = syscall_trace_enter_phase1(regs, arch);
	205
	206	if (phase1_result == 0)
	207	return regs->orig_ax;
	208	else
	209	return syscall_trace_enter_phase2(regs, arch, phase1_result);
	210	}
	211
39b48e57 AL	212	#define EXIT_TO_USERMODE_LOOP_FLAGS \
	213	(_TIF_SIGPENDING \| _TIF_NOTIFY_RESUME \| _TIF_UPROBE \| \
	214	_TIF_NEED_RESCHED \| _TIF_USER_RETURN_NOTIFY)
72f92478	215
39b48e57 AL	216	static void exit_to_usermode_loop(struct pt_regs *regs, u32 cached_flags)
39b48e57 AL	217	{
c5c46f59 AL	218	/*
	219	* In order to return to user mode, we need to have IRQs off with
	220	* none of _TIF_SIGPENDING, _TIF_NOTIFY_RESUME, _TIF_USER_RETURN_NOTIFY,
	221	* _TIF_UPROBE, or _TIF_NEED_RESCHED set. Several of these flags
	222	* can be set at any time on preemptable kernels if we have IRQs on,
	223	* so we need to loop. Disabling preemption wouldn't help: doing the
	224	* work to clear some of the flags can sleep.
	225	*/
	226	while (true) {
c5c46f59 AL	227	/* We have work to do. */
	228	local_irq_enable();
	229
	230	if (cached_flags & _TIF_NEED_RESCHED)
	231	schedule();
	232
	233	if (cached_flags & _TIF_UPROBE)
	234	uprobe_notify_resume(regs);
	235
	236	/* deal with pending signal delivery */
	237	if (cached_flags & _TIF_SIGPENDING)
	238	do_signal(regs);
	239
	240	if (cached_flags & _TIF_NOTIFY_RESUME) {
	241	clear_thread_flag(TIF_NOTIFY_RESUME);
	242	tracehook_notify_resume(regs);
	243	}
	244
	245	if (cached_flags & _TIF_USER_RETURN_NOTIFY)
	246	fire_user_return_notifiers();
	247
	248	/* Disable IRQs and retry */
	249	local_irq_disable();
39b48e57 AL	250
	251	cached_flags = READ_ONCE(pt_regs_to_thread_info(regs)->flags);
	252
	253	if (!(cached_flags & EXIT_TO_USERMODE_LOOP_FLAGS))
	254	break;
	255
c5c46f59	256	}
39b48e57 AL	257	}
	258
	259	/* Called with IRQs disabled. */
	260	__visible inline void prepare_exit_to_usermode(struct pt_regs *regs)
	261	{
4e79e182	262	struct thread_info *ti = pt_regs_to_thread_info(regs);
39b48e57 AL	263	u32 cached_flags;
	264
	265	if (IS_ENABLED(CONFIG_PROVE_LOCKING) && WARN_ON(!irqs_disabled()))
	266	local_irq_disable();
	267
	268	lockdep_sys_exit();
	269
4e79e182	270	cached_flags = READ_ONCE(ti->flags);
39b48e57 AL	271
	272	if (unlikely(cached_flags & EXIT_TO_USERMODE_LOOP_FLAGS))
	273	exit_to_usermode_loop(regs, cached_flags);
c5c46f59	274
4e79e182 AL	275	#ifdef CONFIG_COMPAT
	276	/*
	277	* Compat syscalls set TS_COMPAT. Make sure we clear it before
	278	* returning to user mode. We need to clear it after signal
	279	* handling, because syscall restart has a fixup for compat
	280	* syscalls. The fixup is exercised by the ptrace_syscall_32
	281	* selftest.
	282	*/
	283	ti->status &= ~TS_COMPAT;
	284	#endif
	285
c5c46f59 AL	286	user_enter();
	287	}
	288
f5e6a975 AL	289	#define SYSCALL_EXIT_WORK_FLAGS \
	290	(_TIF_SYSCALL_TRACE \| _TIF_SYSCALL_AUDIT \| \
	291	_TIF_SINGLESTEP \| _TIF_SYSCALL_TRACEPOINT)
	292
	293	static void syscall_slow_exit_work(struct pt_regs *regs, u32 cached_flags)
	294	{
	295	bool step;
	296
	297	audit_syscall_exit(regs);
	298
	299	if (cached_flags & _TIF_SYSCALL_TRACEPOINT)
	300	trace_sys_exit(regs, regs->ax);
	301
	302	/*
	303	* If TIF_SYSCALL_EMU is set, we only get here because of
	304	* TIF_SINGLESTEP (i.e. this is PTRACE_SYSEMU_SINGLESTEP).
	305	* We already reported this syscall instruction in
	306	* syscall_trace_enter().
	307	*/
	308	step = unlikely(
	309	(cached_flags & (_TIF_SINGLESTEP \| _TIF_SYSCALL_EMU))
	310	== _TIF_SINGLESTEP);
	311	if (step \|\| cached_flags & _TIF_SYSCALL_TRACE)
	312	tracehook_report_syscall_exit(regs, step);
	313	}
	314
c5c46f59 AL	315	/*
	316	* Called with IRQs on and fully valid regs. Returns with IRQs off in a
	317	* state such that we can immediately switch to user mode.
	318	*/
f5e6a975	319	__visible inline void syscall_return_slowpath(struct pt_regs *regs)
c5c46f59 AL	320	{
	321	struct thread_info *ti = pt_regs_to_thread_info(regs);
	322	u32 cached_flags = READ_ONCE(ti->flags);
c5c46f59 AL	323
	324	CT_WARN_ON(ct_state() != CONTEXT_KERNEL);
	325
460d1245 AL	326	if (IS_ENABLED(CONFIG_PROVE_LOCKING) &&
460d1245 AL	327	WARN(irqs_disabled(), "syscall %ld left IRQs disabled", regs->orig_ax))
c5c46f59 AL	328	local_irq_enable();
	329
	330	/*
	331	* First do one-time work. If these work items are enabled, we
	332	* want to run them exactly once per syscall exit with IRQs on.
	333	*/
f5e6a975 AL	334	if (unlikely(cached_flags & SYSCALL_EXIT_WORK_FLAGS))
f5e6a975 AL	335	syscall_slow_exit_work(regs, cached_flags);
c5c46f59	336
c5c46f59 AL	337	local_irq_disable();
	338	prepare_exit_to_usermode(regs);
	339	}
bd2d3a3b	340
1e423bff AL	341	#ifdef CONFIG_X86_64
	342	__visible void do_syscall_64(struct pt_regs *regs)
	343	{
	344	struct thread_info *ti = pt_regs_to_thread_info(regs);
	345	unsigned long nr = regs->orig_ax;
	346
	347	local_irq_enable();
	348
	349	if (READ_ONCE(ti->flags) & _TIF_WORK_SYSCALL_ENTRY)
	350	nr = syscall_trace_enter(regs);
	351
	352	/*
	353	* NB: Native and x32 syscalls are dispatched from the same
	354	* table. The only functional difference is the x32 bit in
	355	* regs->orig_ax, which changes the behavior of some syscalls.
	356	*/
	357	if (likely((nr & __SYSCALL_MASK) < NR_syscalls)) {
	358	regs->ax = sys_call_table[nr & __SYSCALL_MASK](
	359	regs->di, regs->si, regs->dx,
	360	regs->r10, regs->r8, regs->r9);
	361	}
	362
	363	syscall_return_slowpath(regs);
	364	}
	365	#endif
	366
bd2d3a3b AL	367	#if defined(CONFIG_X86_32) \|\| defined(CONFIG_IA32_EMULATION)
bd2d3a3b AL	368	/*
8b13c255	369	* Does a 32-bit syscall. Called with IRQs on and does all entry and
33c52129 AL	370	* exit work and returns with IRQs off. This function is extremely hot
	371	* in workloads that use it, and it's usually called from
	372	* do_fast_syscall_32, so forcibly inline it to improve performance.
bd2d3a3b	373	*/
657c1eea AL	374	#ifdef CONFIG_X86_32
	375	/* 32-bit kernels use a trap gate for INT80, and the asm code calls here. */
	376	__visible
	377	#else
	378	/* 64-bit kernels use do_syscall_32_irqs_off() instead. */
	379	static
	380	#endif
	381	__always_inline void do_syscall_32_irqs_on(struct pt_regs *regs)
bd2d3a3b AL	382	{
	383	struct thread_info *ti = pt_regs_to_thread_info(regs);
	384	unsigned int nr = (unsigned int)regs->orig_ax;
	385
	386	#ifdef CONFIG_IA32_EMULATION
	387	ti->status \|= TS_COMPAT;
	388	#endif
	389
bd2d3a3b AL	390	if (READ_ONCE(ti->flags) & _TIF_WORK_SYSCALL_ENTRY) {
	391	/*
	392	* Subtlety here: if ptrace pokes something larger than
	393	* 2^32-1 into orig_ax, this truncates it. This may or
	394	* may not be necessary, but it matches the old asm
	395	* behavior.
	396	*/
	397	nr = syscall_trace_enter(regs);
	398	}
	399
33c52129	400	if (likely(nr < IA32_NR_syscalls)) {
bd2d3a3b AL	401	/*
	402	* It's possible that a 32-bit syscall implementation
	403	* takes a 64-bit parameter but nonetheless assumes that
	404	* the high bits are zero. Make sure we zero-extend all
	405	* of the args.
	406	*/
	407	regs->ax = ia32_sys_call_table[nr](
	408	(unsigned int)regs->bx, (unsigned int)regs->cx,
	409	(unsigned int)regs->dx, (unsigned int)regs->si,
	410	(unsigned int)regs->di, (unsigned int)regs->bp);
	411	}
	412
	413	syscall_return_slowpath(regs);
	414	}
710246df	415
657c1eea AL	416	#ifdef CONFIG_X86_64
	417	/* Handles INT80 on 64-bit kernels */
	418	__visible void do_syscall_32_irqs_off(struct pt_regs *regs)
8b13c255 AL	419	{
	420	local_irq_enable();
	421	do_syscall_32_irqs_on(regs);
	422	}
657c1eea	423	#endif
8b13c255	424
5f310f73	425	/* Returns 0 to return using IRET or 1 to return using SYSEXIT/SYSRETL. */
7841b408	426	__visible long do_fast_syscall_32(struct pt_regs *regs)
710246df AL	427	{
	428	/*
	429	* Called using the internal vDSO SYSENTER/SYSCALL32 calling
	430	* convention. Adjust regs so it looks like we entered using int80.
	431	*/
	432
	433	unsigned long landing_pad = (unsigned long)current->mm->context.vdso +
	434	vdso_image_32.sym_int80_landing_pad;
	435
	436	/*
	437	* SYSENTER loses EIP, and even SYSCALL32 needs us to skip forward
	438	* so that 'regs->ip -= 2' lands back on an int $0x80 instruction.
	439	* Fix it up.
	440	*/
	441	regs->ip = landing_pad;
	442
	443	/*
30bfa7b3	444	* Fetch EBP from where the vDSO stashed it.
710246df AL	445	*
	446	* WARNING: We are in CONTEXT_USER and RCU isn't paying attention!
	447	*/
	448	local_irq_enable();
c68ca678 AL	449	if (
	450	#ifdef CONFIG_X86_64
	451	/*
	452	* Micro-optimization: the pointer we're following is explicitly
	453	* 32 bits, so it can't be out of range.
	454	*/
30bfa7b3	455	__get_user((u32 )&regs->bp,
c68ca678 AL	456	(u32 __user __force *)(unsigned long)(u32)regs->sp)
c68ca678 AL	457	#else
30bfa7b3	458	get_user((u32 )&regs->bp,
c68ca678 AL	459	(u32 __user __force *)(unsigned long)(u32)regs->sp)
	460	#endif
	461	) {
	462
710246df AL	463	/* User code screwed up. */
	464	local_irq_disable();
	465	regs->ax = -EFAULT;
	466	#ifdef CONFIG_CONTEXT_TRACKING
	467	enter_from_user_mode();
	468	#endif
	469	prepare_exit_to_usermode(regs);
7841b408	470	return 0; /* Keep it simple: use IRET. */
710246df	471	}
710246df AL	472
710246df AL	473	/* Now this is just like a normal syscall. */
8b13c255	474	do_syscall_32_irqs_on(regs);
7841b408 AL	475
	476	#ifdef CONFIG_X86_64
	477	/*
	478	* Opportunistic SYSRETL: if possible, try to return using SYSRETL.
	479	* SYSRETL is available on all 64-bit CPUs, so we don't need to
	480	* bother with SYSEXIT.
	481	*
	482	* Unlike 64-bit opportunistic SYSRET, we can't check that CX == IP,
	483	* because the ECX fixup above will ensure that this is essentially
	484	* never the case.
	485	*/
	486	return regs->cs == __USER32_CS && regs->ss == __USER_DS &&
	487	regs->ip == landing_pad &&
	488	(regs->flags & (X86_EFLAGS_RF \| X86_EFLAGS_TF)) == 0;
	489	#else
5f310f73 AL	490	/*
	491	* Opportunistic SYSEXIT: if possible, try to return using SYSEXIT.
	492	*
	493	* Unlike 64-bit opportunistic SYSRET, we can't check that CX == IP,
	494	* because the ECX fixup above will ensure that this is essentially
	495	* never the case.
	496	*
	497	* We don't allow syscalls at all from VM86 mode, but we still
	498	* need to check VM, because we might be returning from sys_vm86.
	499	*/
	500	return static_cpu_has(X86_FEATURE_SEP) &&
	501	regs->cs == __USER_CS && regs->ss == __USER_DS &&
	502	regs->ip == landing_pad &&
	503	(regs->flags & (X86_EFLAGS_RF \| X86_EFLAGS_TF \| X86_EFLAGS_VM)) == 0;
7841b408	504	#endif
710246df	505	}
bd2d3a3b	506	#endif