[linux-2.6-block.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/sched/task_stack.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/nospec.h>
#include <linux/uprobes.h>
#include <linux/livepatch.h>
#include <linux/syscalls.h>

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

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

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

/*
 * Returns the syscall nr to run (which should match regs->orig_ax) or -1
 * to skip the syscall.
 */
static long syscall_trace_enter(struct pt_regs *regs)
{
	u32 arch = in_ia32_syscall() ? AUDIT_ARCH_I386 : AUDIT_ARCH_X86_64;

	struct thread_info *ti = current_thread_info();
	unsigned long ret = 0;
	bool emulated = false;
	u32 work;

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

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

	if (unlikely(work & _TIF_SYSCALL_EMU))
		emulated = true;

	if ((emulated || (work & _TIF_SYSCALL_TRACE)) &&
	    tracehook_report_syscall_entry(regs))
		return -1L;

	if (emulated)
		return -1L;

#ifdef CONFIG_SECCOMP
	/*
	 * Do seccomp after ptrace, to catch any tracer changes.
	 */
	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;
		}

		ret = __secure_computing(&sd);
		if (ret == -1)
			return ret;
	}
#endif

	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;
}

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

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 EXIT_TO_USERMODE_LOOP_FLAGS 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);

		if (cached_flags & _TIF_PATCH_PENDING)
			klp_update_patch_state(current);

		/* 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);
			rseq_handle_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(current_thread_info()->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 = current_thread_info();
	u32 cached_flags;

	addr_limit_user_check();

	lockdep_assert_irqs_disabled();
	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.
	 *
	 * We also need to clear TS_REGS_POKED_I386: the 32-bit tracer
	 * special case only applies after poking regs and before the
	 * very next return to user mode.
	 */
	ti->status &= ~(TS_COMPAT|TS_I386_REGS_POKED);
#endif

	user_enter_irqoff();
}

#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 = current_thread_info();
	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();

	rseq_syscall(regs);

	/*
	 * 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(unsigned long nr, struct pt_regs *regs)
{
	struct thread_info *ti;

	enter_from_user_mode();
	local_irq_enable();
	ti = current_thread_info();
	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.
	 */
	nr &= __SYSCALL_MASK;
	if (likely(nr < NR_syscalls)) {
		nr = array_index_nospec(nr, NR_syscalls);
		regs->ax = sys_call_table[nr](regs);
	}

	syscall_return_slowpath(regs);
}
#endif

#if defined(CONFIG_X86_32) || defined(CONFIG_IA32_EMULATION)
/*
 * Does a 32-bit syscall.  Called with IRQs on in CONTEXT_KERNEL.  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.
 */
static __always_inline void do_syscall_32_irqs_on(struct pt_regs *regs)
{
	struct thread_info *ti = current_thread_info();
	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)) {
		nr = array_index_nospec(nr, IA32_NR_syscalls);
#ifdef CONFIG_IA32_EMULATION
		regs->ax = ia32_sys_call_table[nr](regs);
#else
		/*
		 * 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);
#endif /* CONFIG_IA32_EMULATION */
	}

	syscall_return_slowpath(regs);
}

/* Handles int $0x80 */
__visible void do_int80_syscall_32(struct pt_regs *regs)
{
	enter_from_user_mode();
	local_irq_enable();
	do_syscall_32_irqs_on(regs);
}

/* 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;

	enter_from_user_mode();

	local_irq_enable();

	/* Fetch EBP from where the vDSO stashed it. */
	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;
		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>
68db0cf1	12	#include <linux/sched/task_stack.h>
1f484aa6 AL	13	#include <linux/mm.h>
	14	#include <linux/smp.h>
	15	#include <linux/errno.h>
	16	#include <linux/ptrace.h>
	17	#include <linux/tracehook.h>
	18	#include <linux/audit.h>
	19	#include <linux/seccomp.h>
	20	#include <linux/signal.h>
	21	#include <linux/export.h>
	22	#include <linux/context_tracking.h>
	23	#include <linux/user-return-notifier.h>
2fbd7af5	24	#include <linux/nospec.h>
1f484aa6	25	#include <linux/uprobes.h>
afb94c9e	26	#include <linux/livepatch.h>
5ea0727b	27	#include <linux/syscalls.h>
1f484aa6 AL	28
	29	#include <asm/desc.h>
	30	#include <asm/traps.h>
710246df	31	#include <asm/vdso.h>
7c0f6ba6	32	#include <linux/uaccess.h>
cd4d09ec	33	#include <asm/cpufeature.h>
1f484aa6 AL	34
	35	#define CREATE_TRACE_POINTS
	36	#include <trace/events/syscalls.h>
	37
feed36cd AL	38	#ifdef CONFIG_CONTEXT_TRACKING
feed36cd AL	39	/* Called on entry from user mode with IRQs off. */
be8a18e2	40	__visible inline void enter_from_user_mode(void)
feed36cd AL	41	{
feed36cd AL	42	CT_WARN_ON(ct_state() != CONTEXT_USER);
2e9d1e15	43	user_exit_irqoff();
feed36cd	44	}
9999c8c0 AL	45	#else
9999c8c0 AL	46	static inline void enter_from_user_mode(void) {}
feed36cd AL	47	#endif
feed36cd AL	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	/*
c87a8517 AL	64	* Returns the syscall nr to run (which should match regs->orig_ax) or -1
c87a8517 AL	65	* to skip the syscall.
1f484aa6	66	*/
c87a8517	67	static long syscall_trace_enter(struct pt_regs *regs)
1f484aa6	68	{
c87a8517 AL	69	u32 arch = in_ia32_syscall() ? AUDIT_ARCH_I386 : AUDIT_ARCH_X86_64;
c87a8517 AL	70
97245d00	71	struct thread_info *ti = current_thread_info();
1f484aa6	72	unsigned long ret = 0;
93e35efb	73	bool emulated = false;
1f484aa6 AL	74	u32 work;
1f484aa6 AL	75
4aabd140 AL	76	if (IS_ENABLED(CONFIG_DEBUG_ENTRY))
4aabd140 AL	77	BUG_ON(regs != task_pt_regs(current));
1f484aa6	78
6aa7de05	79	work = READ_ONCE(ti->flags) & _TIF_WORK_SYSCALL_ENTRY;
1f484aa6	80
93e35efb KC	81	if (unlikely(work & _TIF_SYSCALL_EMU))
	82	emulated = true;
	83
	84	if ((emulated \|\| (work & _TIF_SYSCALL_TRACE)) &&
	85	tracehook_report_syscall_entry(regs))
	86	return -1L;
	87
	88	if (emulated)
	89	return -1L;
	90
1f484aa6 AL	91	#ifdef CONFIG_SECCOMP
1f484aa6 AL	92	/*
93e35efb	93	* Do seccomp after ptrace, to catch any tracer changes.
1f484aa6 AL	94	*/
	95	if (work & _TIF_SECCOMP) {
	96	struct seccomp_data sd;
	97
	98	sd.arch = arch;
	99	sd.nr = regs->orig_ax;
	100	sd.instruction_pointer = regs->ip;
	101	#ifdef CONFIG_X86_64
	102	if (arch == AUDIT_ARCH_X86_64) {
	103	sd.args[0] = regs->di;
	104	sd.args[1] = regs->si;
	105	sd.args[2] = regs->dx;
	106	sd.args[3] = regs->r10;
	107	sd.args[4] = regs->r8;
	108	sd.args[5] = regs->r9;
	109	} else
	110	#endif
	111	{
	112	sd.args[0] = regs->bx;
	113	sd.args[1] = regs->cx;
	114	sd.args[2] = regs->dx;
	115	sd.args[3] = regs->si;
	116	sd.args[4] = regs->di;
	117	sd.args[5] = regs->bp;
	118	}
	119
c87a8517 AL	120	ret = __secure_computing(&sd);
	121	if (ret == -1)
	122	return ret;
1f484aa6 AL	123	}
	124	#endif
	125
1f484aa6 AL	126	if (unlikely(test_thread_flag(TIF_SYSCALL_TRACEPOINT)))
	127	trace_sys_enter(regs, regs->orig_ax);
	128
	129	do_audit_syscall_entry(regs, arch);
	130
	131	return ret ?: regs->orig_ax;
	132	}
	133
39b48e57 AL	134	#define EXIT_TO_USERMODE_LOOP_FLAGS \
39b48e57 AL	135	(_TIF_SIGPENDING \| _TIF_NOTIFY_RESUME \| _TIF_UPROBE \| \
afb94c9e	136	_TIF_NEED_RESCHED \| _TIF_USER_RETURN_NOTIFY \| _TIF_PATCH_PENDING)
72f92478	137
39b48e57 AL	138	static void exit_to_usermode_loop(struct pt_regs *regs, u32 cached_flags)
39b48e57 AL	139	{
c5c46f59 AL	140	/*
c5c46f59 AL	141	* In order to return to user mode, we need to have IRQs off with
afb94c9e	142	* none of EXIT_TO_USERMODE_LOOP_FLAGS set. Several of these flags
c5c46f59 AL	143	* can be set at any time on preemptable kernels if we have IRQs on,
	144	* so we need to loop. Disabling preemption wouldn't help: doing the
	145	* work to clear some of the flags can sleep.
	146	*/
	147	while (true) {
c5c46f59 AL	148	/* We have work to do. */
	149	local_irq_enable();
	150
	151	if (cached_flags & _TIF_NEED_RESCHED)
	152	schedule();
	153
	154	if (cached_flags & _TIF_UPROBE)
	155	uprobe_notify_resume(regs);
	156
43347d56 MB	157	if (cached_flags & _TIF_PATCH_PENDING)
	158	klp_update_patch_state(current);
	159
c5c46f59 AL	160	/* deal with pending signal delivery */
	161	if (cached_flags & _TIF_SIGPENDING)
	162	do_signal(regs);
	163
	164	if (cached_flags & _TIF_NOTIFY_RESUME) {
	165	clear_thread_flag(TIF_NOTIFY_RESUME);
	166	tracehook_notify_resume(regs);
d6761b8f	167	rseq_handle_notify_resume(regs);
c5c46f59 AL	168	}
	169
	170	if (cached_flags & _TIF_USER_RETURN_NOTIFY)
	171	fire_user_return_notifiers();
	172
	173	/* Disable IRQs and retry */
	174	local_irq_disable();
39b48e57	175
97245d00	176	cached_flags = READ_ONCE(current_thread_info()->flags);
39b48e57 AL	177
	178	if (!(cached_flags & EXIT_TO_USERMODE_LOOP_FLAGS))
	179	break;
c5c46f59	180	}
39b48e57 AL	181	}
	182
	183	/* Called with IRQs disabled. */
	184	__visible inline void prepare_exit_to_usermode(struct pt_regs *regs)
	185	{
97245d00	186	struct thread_info *ti = current_thread_info();
39b48e57 AL	187	u32 cached_flags;
39b48e57 AL	188
5ea0727b TG	189	addr_limit_user_check();
5ea0727b TG	190
7a10e2a9	191	lockdep_assert_irqs_disabled();
39b48e57 AL	192	lockdep_sys_exit();
39b48e57 AL	193
4e79e182	194	cached_flags = READ_ONCE(ti->flags);
39b48e57 AL	195
	196	if (unlikely(cached_flags & EXIT_TO_USERMODE_LOOP_FLAGS))
	197	exit_to_usermode_loop(regs, cached_flags);
c5c46f59	198
4e79e182 AL	199	#ifdef CONFIG_COMPAT
	200	/*
	201	* Compat syscalls set TS_COMPAT. Make sure we clear it before
	202	* returning to user mode. We need to clear it after signal
	203	* handling, because syscall restart has a fixup for compat
	204	* syscalls. The fixup is exercised by the ptrace_syscall_32
	205	* selftest.
609c19a3 AL	206	*
	207	* We also need to clear TS_REGS_POKED_I386: the 32-bit tracer
	208	* special case only applies after poking regs and before the
	209	* very next return to user mode.
4e79e182	210	*/
37a8f7c3	211	ti->status &= ~(TS_COMPAT\|TS_I386_REGS_POKED);
4e79e182 AL	212	#endif
4e79e182 AL	213
2e9d1e15	214	user_enter_irqoff();
c5c46f59 AL	215	}
c5c46f59 AL	216
f5e6a975 AL	217	#define SYSCALL_EXIT_WORK_FLAGS \
	218	(_TIF_SYSCALL_TRACE \| _TIF_SYSCALL_AUDIT \| \
	219	_TIF_SINGLESTEP \| _TIF_SYSCALL_TRACEPOINT)
	220
	221	static void syscall_slow_exit_work(struct pt_regs *regs, u32 cached_flags)
	222	{
	223	bool step;
	224
	225	audit_syscall_exit(regs);
	226
	227	if (cached_flags & _TIF_SYSCALL_TRACEPOINT)
	228	trace_sys_exit(regs, regs->ax);
	229
	230	/*
	231	* If TIF_SYSCALL_EMU is set, we only get here because of
	232	* TIF_SINGLESTEP (i.e. this is PTRACE_SYSEMU_SINGLESTEP).
	233	* We already reported this syscall instruction in
	234	* syscall_trace_enter().
	235	*/
	236	step = unlikely(
	237	(cached_flags & (_TIF_SINGLESTEP \| _TIF_SYSCALL_EMU))
	238	== _TIF_SINGLESTEP);
	239	if (step \|\| cached_flags & _TIF_SYSCALL_TRACE)
	240	tracehook_report_syscall_exit(regs, step);
	241	}
	242
c5c46f59 AL	243	/*
	244	* Called with IRQs on and fully valid regs. Returns with IRQs off in a
	245	* state such that we can immediately switch to user mode.
	246	*/
f5e6a975	247	__visible inline void syscall_return_slowpath(struct pt_regs *regs)
c5c46f59	248	{
97245d00	249	struct thread_info *ti = current_thread_info();
c5c46f59	250	u32 cached_flags = READ_ONCE(ti->flags);
c5c46f59 AL	251
	252	CT_WARN_ON(ct_state() != CONTEXT_KERNEL);
	253
460d1245 AL	254	if (IS_ENABLED(CONFIG_PROVE_LOCKING) &&
460d1245 AL	255	WARN(irqs_disabled(), "syscall %ld left IRQs disabled", regs->orig_ax))
c5c46f59 AL	256	local_irq_enable();
c5c46f59 AL	257
d6761b8f MD	258	rseq_syscall(regs);
d6761b8f MD	259
c5c46f59 AL	260	/*
	261	* First do one-time work. If these work items are enabled, we
	262	* want to run them exactly once per syscall exit with IRQs on.
	263	*/
f5e6a975 AL	264	if (unlikely(cached_flags & SYSCALL_EXIT_WORK_FLAGS))
f5e6a975 AL	265	syscall_slow_exit_work(regs, cached_flags);
c5c46f59	266
c5c46f59 AL	267	local_irq_disable();
	268	prepare_exit_to_usermode(regs);
	269	}
bd2d3a3b	270
1e423bff	271	#ifdef CONFIG_X86_64
dfe64506	272	__visible void do_syscall_64(unsigned long nr, struct pt_regs *regs)
1e423bff	273	{
dfe64506	274	struct thread_info *ti;
1e423bff	275
9999c8c0	276	enter_from_user_mode();
1e423bff	277	local_irq_enable();
dfe64506	278	ti = current_thread_info();
1e423bff AL	279	if (READ_ONCE(ti->flags) & _TIF_WORK_SYSCALL_ENTRY)
	280	nr = syscall_trace_enter(regs);
	281
	282	/*
	283	* NB: Native and x32 syscalls are dispatched from the same
	284	* table. The only functional difference is the x32 bit in
	285	* regs->orig_ax, which changes the behavior of some syscalls.
	286	*/
dfe64506 LT	287	nr &= __SYSCALL_MASK;
	288	if (likely(nr < NR_syscalls)) {
	289	nr = array_index_nospec(nr, NR_syscalls);
fa697140	290	regs->ax = sys_call_table[nr](regs);
1e423bff AL	291	}
	292
	293	syscall_return_slowpath(regs);
	294	}
	295	#endif
	296
bd2d3a3b AL	297	#if defined(CONFIG_X86_32) \|\| defined(CONFIG_IA32_EMULATION)
bd2d3a3b AL	298	/*
9999c8c0 AL	299	* Does a 32-bit syscall. Called with IRQs on in CONTEXT_KERNEL. Does
	300	* all entry and exit work and returns with IRQs off. This function is
	301	* extremely hot in workloads that use it, and it's usually called from
33c52129	302	* do_fast_syscall_32, so forcibly inline it to improve performance.
bd2d3a3b	303	*/
a798f091	304	static __always_inline void do_syscall_32_irqs_on(struct pt_regs *regs)
bd2d3a3b	305	{
97245d00	306	struct thread_info *ti = current_thread_info();
bd2d3a3b AL	307	unsigned int nr = (unsigned int)regs->orig_ax;
	308
	309	#ifdef CONFIG_IA32_EMULATION
37a8f7c3	310	ti->status \|= TS_COMPAT;
bd2d3a3b AL	311	#endif
bd2d3a3b AL	312
bd2d3a3b AL	313	if (READ_ONCE(ti->flags) & _TIF_WORK_SYSCALL_ENTRY) {
	314	/*
	315	* Subtlety here: if ptrace pokes something larger than
	316	* 2^32-1 into orig_ax, this truncates it. This may or
	317	* may not be necessary, but it matches the old asm
	318	* behavior.
	319	*/
	320	nr = syscall_trace_enter(regs);
	321	}
	322
33c52129	323	if (likely(nr < IA32_NR_syscalls)) {
2fbd7af5	324	nr = array_index_nospec(nr, IA32_NR_syscalls);
f8781c4a	325	#ifdef CONFIG_IA32_EMULATION
ebeb8c82 DB	326	regs->ax = ia32_sys_call_table[nr](regs);
ebeb8c82 DB	327	#else
bd2d3a3b AL	328	/*
	329	* It's possible that a 32-bit syscall implementation
	330	* takes a 64-bit parameter but nonetheless assumes that
	331	* the high bits are zero. Make sure we zero-extend all
	332	* of the args.
	333	*/
	334	regs->ax = ia32_sys_call_table[nr](
	335	(unsigned int)regs->bx, (unsigned int)regs->cx,
	336	(unsigned int)regs->dx, (unsigned int)regs->si,
	337	(unsigned int)regs->di, (unsigned int)regs->bp);
f8781c4a	338	#endif /* CONFIG_IA32_EMULATION */
bd2d3a3b AL	339	}
	340
	341	syscall_return_slowpath(regs);
	342	}
710246df	343
a798f091 AL	344	/* Handles int $0x80 */
a798f091 AL	345	__visible void do_int80_syscall_32(struct pt_regs *regs)
8b13c255	346	{
9999c8c0	347	enter_from_user_mode();
8b13c255 AL	348	local_irq_enable();
	349	do_syscall_32_irqs_on(regs);
	350	}
	351
5f310f73	352	/* Returns 0 to return using IRET or 1 to return using SYSEXIT/SYSRETL. */
7841b408	353	__visible long do_fast_syscall_32(struct pt_regs *regs)
710246df AL	354	{
	355	/*
	356	* Called using the internal vDSO SYSENTER/SYSCALL32 calling
	357	* convention. Adjust regs so it looks like we entered using int80.
	358	*/
	359
	360	unsigned long landing_pad = (unsigned long)current->mm->context.vdso +
	361	vdso_image_32.sym_int80_landing_pad;
	362
	363	/*
	364	* SYSENTER loses EIP, and even SYSCALL32 needs us to skip forward
	365	* so that 'regs->ip -= 2' lands back on an int $0x80 instruction.
	366	* Fix it up.
	367	*/
	368	regs->ip = landing_pad;
	369
9999c8c0 AL	370	enter_from_user_mode();
9999c8c0 AL	371
710246df	372	local_irq_enable();
9999c8c0 AL	373
9999c8c0 AL	374	/* Fetch EBP from where the vDSO stashed it. */
c68ca678 AL	375	if (
	376	#ifdef CONFIG_X86_64
	377	/*
	378	* Micro-optimization: the pointer we're following is explicitly
	379	* 32 bits, so it can't be out of range.
	380	*/
30bfa7b3	381	__get_user((u32 )&regs->bp,
c68ca678 AL	382	(u32 __user __force *)(unsigned long)(u32)regs->sp)
c68ca678 AL	383	#else
30bfa7b3	384	get_user((u32 )&regs->bp,
c68ca678 AL	385	(u32 __user __force *)(unsigned long)(u32)regs->sp)
	386	#endif
	387	) {
	388
710246df AL	389	/* User code screwed up. */
	390	local_irq_disable();
	391	regs->ax = -EFAULT;
710246df	392	prepare_exit_to_usermode(regs);
7841b408	393	return 0; /* Keep it simple: use IRET. */
710246df	394	}
710246df AL	395
710246df AL	396	/* Now this is just like a normal syscall. */
8b13c255	397	do_syscall_32_irqs_on(regs);
7841b408 AL	398
	399	#ifdef CONFIG_X86_64
	400	/*
	401	* Opportunistic SYSRETL: if possible, try to return using SYSRETL.
	402	* SYSRETL is available on all 64-bit CPUs, so we don't need to
	403	* bother with SYSEXIT.
	404	*
	405	* Unlike 64-bit opportunistic SYSRET, we can't check that CX == IP,
	406	* because the ECX fixup above will ensure that this is essentially
	407	* never the case.
	408	*/
	409	return regs->cs == __USER32_CS && regs->ss == __USER_DS &&
	410	regs->ip == landing_pad &&
	411	(regs->flags & (X86_EFLAGS_RF \| X86_EFLAGS_TF)) == 0;
	412	#else
5f310f73 AL	413	/*
	414	* Opportunistic SYSEXIT: if possible, try to return using SYSEXIT.
	415	*
	416	* Unlike 64-bit opportunistic SYSRET, we can't check that CX == IP,
	417	* because the ECX fixup above will ensure that this is essentially
	418	* never the case.
	419	*
	420	* We don't allow syscalls at all from VM86 mode, but we still
	421	* need to check VM, because we might be returning from sys_vm86.
	422	*/
	423	return static_cpu_has(X86_FEATURE_SEP) &&
	424	regs->cs == __USER_CS && regs->ss == __USER_DS &&
	425	regs->ip == landing_pad &&
	426	(regs->flags & (X86_EFLAGS_RF \| X86_EFLAGS_TF \| X86_EFLAGS_VM)) == 0;
7841b408	427	#endif
710246df	428	}
bd2d3a3b	429	#endif