[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>

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

#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)
{
	unsigned long ret = 0;
	u32 work;

	BUG_ON(regs != task_pt_regs(current));

	work = ACCESS_ONCE(current_thread_info()->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)
{
	long ret = 0;
	u32 work = ACCESS_ONCE(current_thread_info()->flags) &
		_TIF_WORK_SYSCALL_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);
}

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);
}

/* Called with IRQs disabled. */
__visible void prepare_exit_to_usermode(struct pt_regs *regs)
{
	if (WARN_ON(!irqs_disabled()))
		local_irq_disable();

	/*
	 * 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) {
		u32 cached_flags =
			READ_ONCE(pt_regs_to_thread_info(regs)->flags);

		if (!(cached_flags & (_TIF_SIGPENDING | _TIF_NOTIFY_RESUME |
				      _TIF_UPROBE | _TIF_NEED_RESCHED |
				      _TIF_USER_RETURN_NOTIFY)))
			break;

		/* 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();
	}

	user_enter();
}

/*
 * 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 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);
	bool step;

	CT_WARN_ON(ct_state() != CONTEXT_KERNEL);

	if (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 (cached_flags & (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_AUDIT |
			    _TIF_SINGLESTEP | _TIF_SYSCALL_TRACEPOINT)) {
		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);
	}

#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);
}
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>
	27
	28	#define CREATE_TRACE_POINTS
	29	#include <trace/events/syscalls.h>
	30
feed36cd AL	31	#ifdef CONFIG_CONTEXT_TRACKING
	32	/* Called on entry from user mode with IRQs off. */
	33	__visible void enter_from_user_mode(void)
	34	{
	35	CT_WARN_ON(ct_state() != CONTEXT_USER);
	36	user_exit();
	37	}
	38	#endif
	39
1f484aa6 AL	40	static void do_audit_syscall_entry(struct pt_regs *regs, u32 arch)
	41	{
	42	#ifdef CONFIG_X86_64
	43	if (arch == AUDIT_ARCH_X86_64) {
	44	audit_syscall_entry(regs->orig_ax, regs->di,
	45	regs->si, regs->dx, regs->r10);
	46	} else
	47	#endif
	48	{
	49	audit_syscall_entry(regs->orig_ax, regs->bx,
	50	regs->cx, regs->dx, regs->si);
	51	}
	52	}
	53
	54	/*
	55	* We can return 0 to resume the syscall or anything else to go to phase
	56	* 2. If we resume the syscall, we need to put something appropriate in
	57	* regs->orig_ax.
	58	*
	59	* NB: We don't have full pt_regs here, but regs->orig_ax and regs->ax
	60	* are fully functional.
	61	*
	62	* For phase 2's benefit, our return value is:
	63	* 0: resume the syscall
	64	* 1: go to phase 2; no seccomp phase 2 needed
	65	* anything else: go to phase 2; pass return value to seccomp
	66	*/
	67	unsigned long syscall_trace_enter_phase1(struct pt_regs *regs, u32 arch)
	68	{
	69	unsigned long ret = 0;
	70	u32 work;
	71
	72	BUG_ON(regs != task_pt_regs(current));
	73
	74	work = ACCESS_ONCE(current_thread_info()->flags) &
	75	_TIF_WORK_SYSCALL_ENTRY;
	76
feed36cd	77	#ifdef CONFIG_CONTEXT_TRACKING
1f484aa6 AL	78	/*
	79	* If TIF_NOHZ is set, we are required to call user_exit() before
	80	* doing anything that could touch RCU.
	81	*/
	82	if (work & _TIF_NOHZ) {
feed36cd	83	enter_from_user_mode();
1f484aa6 AL	84	work &= ~_TIF_NOHZ;
1f484aa6 AL	85	}
feed36cd	86	#endif
1f484aa6 AL	87
	88	#ifdef CONFIG_SECCOMP
	89	/*
	90	* Do seccomp first -- it should minimize exposure of other
	91	* code, and keeping seccomp fast is probably more valuable
	92	* than the rest of this.
	93	*/
	94	if (work & _TIF_SECCOMP) {
	95	struct seccomp_data sd;
	96
	97	sd.arch = arch;
	98	sd.nr = regs->orig_ax;
	99	sd.instruction_pointer = regs->ip;
	100	#ifdef CONFIG_X86_64
	101	if (arch == AUDIT_ARCH_X86_64) {
	102	sd.args[0] = regs->di;
	103	sd.args[1] = regs->si;
	104	sd.args[2] = regs->dx;
	105	sd.args[3] = regs->r10;
	106	sd.args[4] = regs->r8;
	107	sd.args[5] = regs->r9;
	108	} else
	109	#endif
	110	{
	111	sd.args[0] = regs->bx;
	112	sd.args[1] = regs->cx;
	113	sd.args[2] = regs->dx;
	114	sd.args[3] = regs->si;
	115	sd.args[4] = regs->di;
	116	sd.args[5] = regs->bp;
	117	}
	118
	119	BUILD_BUG_ON(SECCOMP_PHASE1_OK != 0);
	120	BUILD_BUG_ON(SECCOMP_PHASE1_SKIP != 1);
	121
	122	ret = seccomp_phase1(&sd);
	123	if (ret == SECCOMP_PHASE1_SKIP) {
	124	regs->orig_ax = -1;
	125	ret = 0;
	126	} else if (ret != SECCOMP_PHASE1_OK) {
	127	return ret; /* Go directly to phase 2 */
	128	}
	129
	130	work &= ~_TIF_SECCOMP;
	131	}
	132	#endif
	133
	134	/* Do our best to finish without phase 2. */
	135	if (work == 0)
	136	return ret; /* seccomp and/or nohz only (ret == 0 here) */
	137
	138	#ifdef CONFIG_AUDITSYSCALL
	139	if (work == _TIF_SYSCALL_AUDIT) {
	140	/*
	141	* If there is no more work to be done except auditing,
	142	* then audit in phase 1. Phase 2 always audits, so, if
	143	* we audit here, then we can't go on to phase 2.
	144	*/
	145	do_audit_syscall_entry(regs, arch);
	146	return 0;
	147	}
	148	#endif
	149
	150	return 1; /* Something is enabled that we can't handle in phase 1 */
151	}
152
153	/* Returns the syscall nr to run (which should match regs->orig_ax). */
154	long syscall_trace_enter_phase2(struct pt_regs *regs, u32 arch,
155	unsigned long phase1_result)
156	{
157	long ret = 0;
158	u32 work = ACCESS_ONCE(current_thread_info()->flags) &
159	_TIF_WORK_SYSCALL_ENTRY;
160
161	BUG_ON(regs != task_pt_regs(current));
162
163	/*
164	* If we stepped into a sysenter/syscall insn, it trapped in
165	* kernel mode; do_debug() cleared TF and set TIF_SINGLESTEP.
166	* If user-mode had set TF itself, then it's still clear from
167	* do_debug() and we need to set it again to restore the user
168	* state. If we entered on the slow path, TF was already set.
169	*/
170	if (work & _TIF_SINGLESTEP)
171	regs->flags \|= X86_EFLAGS_TF;
172
173	#ifdef CONFIG_SECCOMP
174	/*
175	* Call seccomp_phase2 before running the other hooks so that
176	* they can see any changes made by a seccomp tracer.
177	*/
178	if (phase1_result > 1 && seccomp_phase2(phase1_result)) {
179	/* seccomp failures shouldn't expose any additional code. */
180	return -1;
181	}
182	#endif
183
184	if (unlikely(work & _TIF_SYSCALL_EMU))
185	ret = -1L;
186
187	if ((ret \|\| test_thread_flag(TIF_SYSCALL_TRACE)) &&
188	tracehook_report_syscall_entry(regs))
189	ret = -1L;
190
191	if (unlikely(test_thread_flag(TIF_SYSCALL_TRACEPOINT)))
192	trace_sys_enter(regs, regs->orig_ax);
193
194	do_audit_syscall_entry(regs, arch);
195
196	return ret ?: regs->orig_ax;
197	}
198
199	long syscall_trace_enter(struct pt_regs *regs)
200	{
201	u32 arch = is_ia32_task() ? AUDIT_ARCH_I386 : AUDIT_ARCH_X86_64;
202	unsigned long phase1_result = syscall_trace_enter_phase1(regs, arch);
203
204	if (phase1_result == 0)
205	return regs->orig_ax;
206	else
207	return syscall_trace_enter_phase2(regs, arch, phase1_result);
208	}
209
c5c46f59 AL	210	static struct thread_info pt_regs_to_thread_info(struct pt_regs regs)
	211	{
	212	unsigned long top_of_stack =
	213	(unsigned long)(regs + 1) + TOP_OF_KERNEL_STACK_PADDING;
	214	return (struct thread_info *)(top_of_stack - THREAD_SIZE);
	215	}
	216
	217	/* Called with IRQs disabled. */
	218	__visible void prepare_exit_to_usermode(struct pt_regs *regs)
	219	{
	220	if (WARN_ON(!irqs_disabled()))
	221	local_irq_disable();
	222
	223	/*
	224	* In order to return to user mode, we need to have IRQs off with
	225	* none of _TIF_SIGPENDING, _TIF_NOTIFY_RESUME, _TIF_USER_RETURN_NOTIFY,
	226	* _TIF_UPROBE, or _TIF_NEED_RESCHED set. Several of these flags
	227	* can be set at any time on preemptable kernels if we have IRQs on,
	228	* so we need to loop. Disabling preemption wouldn't help: doing the
	229	* work to clear some of the flags can sleep.
	230	*/
	231	while (true) {
	232	u32 cached_flags =
	233	READ_ONCE(pt_regs_to_thread_info(regs)->flags);
	234
	235	if (!(cached_flags & (_TIF_SIGPENDING \| _TIF_NOTIFY_RESUME \|
d132803e AL	236	_TIF_UPROBE \| _TIF_NEED_RESCHED \|
d132803e AL	237	_TIF_USER_RETURN_NOTIFY)))
c5c46f59 AL	238	break;
	239
	240	/* We have work to do. */
	241	local_irq_enable();
	242
	243	if (cached_flags & _TIF_NEED_RESCHED)
	244	schedule();
	245
	246	if (cached_flags & _TIF_UPROBE)
	247	uprobe_notify_resume(regs);
	248
	249	/* deal with pending signal delivery */
	250	if (cached_flags & _TIF_SIGPENDING)
	251	do_signal(regs);
	252
	253	if (cached_flags & _TIF_NOTIFY_RESUME) {
	254	clear_thread_flag(TIF_NOTIFY_RESUME);
	255	tracehook_notify_resume(regs);
	256	}
	257
	258	if (cached_flags & _TIF_USER_RETURN_NOTIFY)
	259	fire_user_return_notifiers();
	260
	261	/* Disable IRQs and retry */
	262	local_irq_disable();
	263	}
	264
	265	user_enter();
	266	}
	267
	268	/*
	269	* Called with IRQs on and fully valid regs. Returns with IRQs off in a
	270	* state such that we can immediately switch to user mode.
	271	*/
	272	__visible void syscall_return_slowpath(struct pt_regs *regs)
	273	{
	274	struct thread_info *ti = pt_regs_to_thread_info(regs);
	275	u32 cached_flags = READ_ONCE(ti->flags);
	276	bool step;
	277
	278	CT_WARN_ON(ct_state() != CONTEXT_KERNEL);
	279
	280	if (WARN(irqs_disabled(), "syscall %ld left IRQs disabled",
	281	regs->orig_ax))
	282	local_irq_enable();
	283
	284	/*
	285	* First do one-time work. If these work items are enabled, we
	286	* want to run them exactly once per syscall exit with IRQs on.
	287	*/
	288	if (cached_flags & (_TIF_SYSCALL_TRACE \| _TIF_SYSCALL_AUDIT \|
	289	_TIF_SINGLESTEP \| _TIF_SYSCALL_TRACEPOINT)) {
	290	audit_syscall_exit(regs);
	291
	292	if (cached_flags & _TIF_SYSCALL_TRACEPOINT)
	293	trace_sys_exit(regs, regs->ax);
	294
	295	/*
	296	* If TIF_SYSCALL_EMU is set, we only get here because of
	297	* TIF_SINGLESTEP (i.e. this is PTRACE_SYSEMU_SINGLESTEP).
	298	* We already reported this syscall instruction in
	299	* syscall_trace_enter().
	300	*/
	301	step = unlikely(
302	(cached_flags & (_TIF_SINGLESTEP \| _TIF_SYSCALL_EMU))
303	== _TIF_SINGLESTEP);
304	if (step \|\| cached_flags & _TIF_SYSCALL_TRACE)
305	tracehook_report_syscall_exit(regs, step);
306	}
307
308	#ifdef CONFIG_COMPAT
309	/*
310	* Compat syscalls set TS_COMPAT. Make sure we clear it before
311	* returning to user mode.
312	*/
313	ti->status &= ~TS_COMPAT;
314	#endif
315
316	local_irq_disable();
317	prepare_exit_to_usermode(regs);
318	}