[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>

#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));

	/*
	 * If we stepped into a sysenter/syscall insn, it trapped in
	 * kernel mode; do_debug() cleared TF and set TIF_SINGLESTEP.
	 * If user-mode had set TF itself, then it's still clear from
	 * do_debug() and we need to set it again to restore the user
	 * state.  If we entered on the slow path, TF was already set.
	 */
	if (work & _TIF_SINGLESTEP)
		regs->flags |= X86_EFLAGS_TF;

#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)
{
	u32 cached_flags;

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

	lockdep_sys_exit();

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

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

	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);

#ifdef CONFIG_COMPAT
	/*
	 * Compat syscalls set TS_COMPAT.  Make sure we clear it before
	 * returning to user mode.
	 */
	ti->status &= ~TS_COMPAT;
#endif

	local_irq_disable();
	prepare_exit_to_usermode(regs);
}

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