1 files changed, 146 insertions, 83 deletions

diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c
index 527e17e..57501d9 100644
--- a/kernel/trace/trace_syscalls.c
+++ b/kernel/trace/trace_syscalls.c
@@ -14,6 +14,43 @@ static int sys_refcount_exit;
 static DECLARE_BITMAP(enabled_enter_syscalls, NR_syscalls);
 static DECLARE_BITMAP(enabled_exit_syscalls, NR_syscalls);
 
+extern unsigned long __start_syscalls_metadata[];
+extern unsigned long __stop_syscalls_metadata[];
+
+static struct syscall_metadata **syscalls_metadata;
+
+static struct syscall_metadata *find_syscall_meta(unsigned long syscall)
+{
+	struct syscall_metadata *start;
+	struct syscall_metadata *stop;
+	char str[KSYM_SYMBOL_LEN];
+
+
+	start = (struct syscall_metadata *)__start_syscalls_metadata;
+	stop = (struct syscall_metadata *)__stop_syscalls_metadata;
+	kallsyms_lookup(syscall, NULL, NULL, NULL, str);
+
+	for ( ; start < stop; start++) {
+		/*
+		 * Only compare after the "sys" prefix. Archs that use
+		 * syscall wrappers may have syscalls symbols aliases prefixed
+		 * with "SyS" instead of "sys", leading to an unwanted
+		 * mismatch.
+		 */
+		if (start->name && !strcmp(start->name + 3, str + 3))
+			return start;
+	}
+	return NULL;
+}
+
+static struct syscall_metadata *syscall_nr_to_meta(int nr)
+{
+	if (!syscalls_metadata || nr >= NR_syscalls || nr < 0)
+		return NULL;
+
+	return syscalls_metadata[nr];
+}
+
 enum print_line_t
 print_syscall_enter(struct trace_iterator *iter, int flags)
 {
@@ -30,7 +67,7 @@ print_syscall_enter(struct trace_iterator *iter, int flags)
 	if (!entry)
 		goto end;
 
-	if (entry->enter_id != ent->type) {
+	if (entry->enter_event->id != ent->type) {
 		WARN_ON_ONCE(1);
 		goto end;
 	}
@@ -85,7 +122,7 @@ print_syscall_exit(struct trace_iterator *iter, int flags)
 		return TRACE_TYPE_HANDLED;
 	}
 
-	if (entry->exit_id != ent->type) {
+	if (entry->exit_event->id != ent->type) {
 		WARN_ON_ONCE(1);
 		return TRACE_TYPE_UNHANDLED;
 	}
@@ -103,24 +140,19 @@ extern char *__bad_type_size(void);
 #define SYSCALL_FIELD(type, name)					\
 	sizeof(type) != sizeof(trace.name) ?				\
 		__bad_type_size() :					\
-		#type, #name, offsetof(typeof(trace), name), sizeof(trace.name)
+		#type, #name, offsetof(typeof(trace), name),		\
+		sizeof(trace.name), is_signed_type(type)
 
 int syscall_enter_format(struct ftrace_event_call *call, struct trace_seq *s)
 {
 	int i;
-	int nr;
 	int ret;
-	struct syscall_metadata *entry;
+	struct syscall_metadata *entry = call->data;
 	struct syscall_trace_enter trace;
 	int offset = offsetof(struct syscall_trace_enter, args);
 
-	nr = syscall_name_to_nr(call->data);
-	entry = syscall_nr_to_meta(nr);
-
-	if (!entry)
-		return 0;
-
-	ret = trace_seq_printf(s, "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n",
+	ret = trace_seq_printf(s, "\tfield:%s %s;\toffset:%zu;\tsize:%zu;"
+			       "\tsigned:%u;\n",
 			       SYSCALL_FIELD(int, nr));
 	if (!ret)
 		return 0;
@@ -130,8 +162,10 @@ int syscall_enter_format(struct ftrace_event_call *call, struct trace_seq *s)
 				        entry->args[i]);
 		if (!ret)
 			return 0;
-		ret = trace_seq_printf(s, "\toffset:%d;\tsize:%zu;\n", offset,
-				       sizeof(unsigned long));
+		ret = trace_seq_printf(s, "\toffset:%d;\tsize:%zu;"
+				       "\tsigned:%u;\n", offset,
+				       sizeof(unsigned long),
+				       is_signed_type(unsigned long));
 		if (!ret)
 			return 0;
 		offset += sizeof(unsigned long);
@@ -163,8 +197,10 @@ int syscall_exit_format(struct ftrace_event_call *call, struct trace_seq *s)
 	struct syscall_trace_exit trace;
 
 	ret = trace_seq_printf(s,
-			       "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n"
-			       "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n",
+			       "\tfield:%s %s;\toffset:%zu;\tsize:%zu;"
+			       "\tsigned:%u;\n"
+			       "\tfield:%s %s;\toffset:%zu;\tsize:%zu;"
+			       "\tsigned:%u;\n",
 			       SYSCALL_FIELD(int, nr),
 			       SYSCALL_FIELD(long, ret));
 	if (!ret)
@@ -176,22 +212,19 @@ int syscall_exit_format(struct ftrace_event_call *call, struct trace_seq *s)
 int syscall_enter_define_fields(struct ftrace_event_call *call)
 {
 	struct syscall_trace_enter trace;
-	struct syscall_metadata *meta;
+	struct syscall_metadata *meta = call->data;
 	int ret;
-	int nr;
 	int i;
 	int offset = offsetof(typeof(trace), args);
 
-	nr = syscall_name_to_nr(call->data);
-	meta = syscall_nr_to_meta(nr);
-
-	if (!meta)
-		return 0;
-
 	ret = trace_define_common_fields(call);
 	if (ret)
 		return ret;
 
+	ret = trace_define_field(call, SYSCALL_FIELD(int, nr), FILTER_OTHER);
+	if (ret)
+		return ret;
+
 	for (i = 0; i < meta->nb_args; i++) {
 		ret = trace_define_field(call, meta->types[i],
 					 meta->args[i], offset,
@@ -212,7 +245,11 @@ int syscall_exit_define_fields(struct ftrace_event_call *call)
 	if (ret)
 		return ret;
 
-	ret = trace_define_field(call, SYSCALL_FIELD(long, ret), 0,
+	ret = trace_define_field(call, SYSCALL_FIELD(int, nr), FILTER_OTHER);
+	if (ret)
+		return ret;
+
+	ret = trace_define_field(call, SYSCALL_FIELD(long, ret),
 				 FILTER_OTHER);
 
 	return ret;
@@ -239,8 +276,8 @@ void ftrace_syscall_enter(struct pt_regs *regs, long id)
 
 	size = sizeof(*entry) + sizeof(unsigned long) * sys_data->nb_args;
 
-	event = trace_current_buffer_lock_reserve(&buffer, sys_data->enter_id,
-						  size, 0, 0);
+	event = trace_current_buffer_lock_reserve(&buffer,
+			sys_data->enter_event->id, size, 0, 0);
 	if (!event)
 		return;
 
@@ -271,8 +308,8 @@ void ftrace_syscall_exit(struct pt_regs *regs, long ret)
 	if (!sys_data)
 		return;
 
-	event = trace_current_buffer_lock_reserve(&buffer, sys_data->exit_id,
-				sizeof(*entry), 0, 0);
+	event = trace_current_buffer_lock_reserve(&buffer,
+			sys_data->exit_event->id, sizeof(*entry), 0, 0);
 	if (!event)
 		return;
 
@@ -285,14 +322,12 @@ void ftrace_syscall_exit(struct pt_regs *regs, long ret)
 		trace_current_buffer_unlock_commit(buffer, event, 0, 0);
 }
 
-int reg_event_syscall_enter(void *ptr)
+int reg_event_syscall_enter(struct ftrace_event_call *call)
 {
 	int ret = 0;
 	int num;
-	char *name;
 
-	name = (char *)ptr;
-	num = syscall_name_to_nr(name);
+	num = ((struct syscall_metadata *)call->data)->syscall_nr;
 	if (num < 0 || num >= NR_syscalls)
 		return -ENOSYS;
 	mutex_lock(&syscall_trace_lock);
@@ -309,13 +344,11 @@ int reg_event_syscall_enter(void *ptr)
 	return ret;
 }
 
-void unreg_event_syscall_enter(void *ptr)
+void unreg_event_syscall_enter(struct ftrace_event_call *call)
 {
 	int num;
-	char *name;
 
-	name = (char *)ptr;
-	num = syscall_name_to_nr(name);
+	num = ((struct syscall_metadata *)call->data)->syscall_nr;
 	if (num < 0 || num >= NR_syscalls)
 		return;
 	mutex_lock(&syscall_trace_lock);
@@ -326,14 +359,12 @@ void unreg_event_syscall_enter(void *ptr)
 	mutex_unlock(&syscall_trace_lock);
 }
 
-int reg_event_syscall_exit(void *ptr)
+int reg_event_syscall_exit(struct ftrace_event_call *call)
 {
 	int ret = 0;
 	int num;
-	char *name;
 
-	name = (char *)ptr;
-	num = syscall_name_to_nr(name);
+	num = ((struct syscall_metadata *)call->data)->syscall_nr;
 	if (num < 0 || num >= NR_syscalls)
 		return -ENOSYS;
 	mutex_lock(&syscall_trace_lock);
@@ -350,13 +381,11 @@ int reg_event_syscall_exit(void *ptr)
 	return ret;
 }
 
-void unreg_event_syscall_exit(void *ptr)
+void unreg_event_syscall_exit(struct ftrace_event_call *call)
 {
 	int num;
-	char *name;
 
-	name = (char *)ptr;
-	num = syscall_name_to_nr(name);
+	num = ((struct syscall_metadata *)call->data)->syscall_nr;
 	if (num < 0 || num >= NR_syscalls)
 		return;
 	mutex_lock(&syscall_trace_lock);
@@ -367,13 +396,44 @@ void unreg_event_syscall_exit(void *ptr)
 	mutex_unlock(&syscall_trace_lock);
 }
 
-struct trace_event event_syscall_enter = {
-	.trace			= print_syscall_enter,
-};
+int init_syscall_trace(struct ftrace_event_call *call)
+{
+	int id;
+
+	id = register_ftrace_event(call->event);
+	if (!id)
+		return -ENODEV;
+	call->id = id;
+	INIT_LIST_HEAD(&call->fields);
+	return 0;
+}
+
+int __init init_ftrace_syscalls(void)
+{
+	struct syscall_metadata *meta;
+	unsigned long addr;
+	int i;
+
+	syscalls_metadata = kzalloc(sizeof(*syscalls_metadata) *
+					NR_syscalls, GFP_KERNEL);
+	if (!syscalls_metadata) {
+		WARN_ON(1);
+		return -ENOMEM;
+	}
+
+	for (i = 0; i < NR_syscalls; i++) {
+		addr = arch_syscall_addr(i);
+		meta = find_syscall_meta(addr);
+		if (!meta)
+			continue;
+
+		meta->syscall_nr = i;
+		syscalls_metadata[i] = meta;
+	}
 
-struct trace_event event_syscall_exit = {
-	.trace			= print_syscall_exit,
-};
+	return 0;
+}
+core_initcall(init_ftrace_syscalls);
 
 #ifdef CONFIG_EVENT_PROFILE
 
@@ -387,8 +447,10 @@ static void prof_syscall_enter(struct pt_regs *regs, long id)
 	struct syscall_metadata *sys_data;
 	struct syscall_trace_enter *rec;
 	unsigned long flags;
+	char *trace_buf;
 	char *raw_data;
 	int syscall_nr;
+	int rctx;
 	int size;
 	int cpu;
 
@@ -412,41 +474,42 @@ static void prof_syscall_enter(struct pt_regs *regs, long id)
 	/* Protect the per cpu buffer, begin the rcu read side */
 	local_irq_save(flags);
 
+	rctx = perf_swevent_get_recursion_context();
+	if (rctx < 0)
+		goto end_recursion;
+
 	cpu = smp_processor_id();
 
-	if (in_nmi())
-		raw_data = rcu_dereference(trace_profile_buf_nmi);
-	else
-		raw_data = rcu_dereference(trace_profile_buf);
+	trace_buf = rcu_dereference(perf_trace_buf);
 
-	if (!raw_data)
+	if (!trace_buf)
 		goto end;
 
-	raw_data = per_cpu_ptr(raw_data, cpu);
+	raw_data = per_cpu_ptr(trace_buf, cpu);
 
 	/* zero the dead bytes from align to not leak stack to user */
 	*(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL;
 
 	rec = (struct syscall_trace_enter *) raw_data;
 	tracing_generic_entry_update(&rec->ent, 0, 0);
-	rec->ent.type = sys_data->enter_id;
+	rec->ent.type = sys_data->enter_event->id;
 	rec->nr = syscall_nr;
 	syscall_get_arguments(current, regs, 0, sys_data->nb_args,
 			       (unsigned long *)&rec->args);
-	perf_tp_event(sys_data->enter_id, 0, 1, rec, size);
+	perf_tp_event(sys_data->enter_event->id, 0, 1, rec, size);
 
 end:
+	perf_swevent_put_recursion_context(rctx);
+end_recursion:
 	local_irq_restore(flags);
 }
 
-int reg_prof_syscall_enter(char *name)
+int prof_sysenter_enable(struct ftrace_event_call *call)
 {
 	int ret = 0;
 	int num;
 
-	num = syscall_name_to_nr(name);
-	if (num < 0 || num >= NR_syscalls)
-		return -ENOSYS;
+	num = ((struct syscall_metadata *)call->data)->syscall_nr;
 
 	mutex_lock(&syscall_trace_lock);
 	if (!sys_prof_refcount_enter)
@@ -462,13 +525,11 @@ int reg_prof_syscall_enter(char *name)
 	return ret;
 }
 
-void unreg_prof_syscall_enter(char *name)
+void prof_sysenter_disable(struct ftrace_event_call *call)
 {
 	int num;
 
-	num = syscall_name_to_nr(name);
-	if (num < 0 || num >= NR_syscalls)
-		return;
+	num = ((struct syscall_metadata *)call->data)->syscall_nr;
 
 	mutex_lock(&syscall_trace_lock);
 	sys_prof_refcount_enter--;
@@ -484,7 +545,9 @@ static void prof_syscall_exit(struct pt_regs *regs, long ret)
 	struct syscall_trace_exit *rec;
 	unsigned long flags;
 	int syscall_nr;
+	char *trace_buf;
 	char *raw_data;
+	int rctx;
 	int size;
 	int cpu;
 
@@ -510,17 +573,19 @@ static void prof_syscall_exit(struct pt_regs *regs, long ret)
 
 	/* Protect the per cpu buffer, begin the rcu read side */
 	local_irq_save(flags);
+
+	rctx = perf_swevent_get_recursion_context();
+	if (rctx < 0)
+		goto end_recursion;
+
 	cpu = smp_processor_id();
 
-	if (in_nmi())
-		raw_data = rcu_dereference(trace_profile_buf_nmi);
-	else
-		raw_data = rcu_dereference(trace_profile_buf);
+	trace_buf = rcu_dereference(perf_trace_buf);
 
-	if (!raw_data)
+	if (!trace_buf)
 		goto end;
 
-	raw_data = per_cpu_ptr(raw_data, cpu);
+	raw_data = per_cpu_ptr(trace_buf, cpu);
 
 	/* zero the dead bytes from align to not leak stack to user */
 	*(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL;
@@ -528,24 +593,24 @@ static void prof_syscall_exit(struct pt_regs *regs, long ret)
 	rec = (struct syscall_trace_exit *)raw_data;
 
 	tracing_generic_entry_update(&rec->ent, 0, 0);
-	rec->ent.type = sys_data->exit_id;
+	rec->ent.type = sys_data->exit_event->id;
 	rec->nr = syscall_nr;
 	rec->ret = syscall_get_return_value(current, regs);
 
-	perf_tp_event(sys_data->exit_id, 0, 1, rec, size);
+	perf_tp_event(sys_data->exit_event->id, 0, 1, rec, size);
 
 end:
+	perf_swevent_put_recursion_context(rctx);
+end_recursion:
 	local_irq_restore(flags);
 }
 
-int reg_prof_syscall_exit(char *name)
+int prof_sysexit_enable(struct ftrace_event_call *call)
 {
 	int ret = 0;
 	int num;
 
-	num = syscall_name_to_nr(name);
-	if (num < 0 || num >= NR_syscalls)
-		return -ENOSYS;
+	num = ((struct syscall_metadata *)call->data)->syscall_nr;
 
 	mutex_lock(&syscall_trace_lock);
 	if (!sys_prof_refcount_exit)
@@ -561,13 +626,11 @@ int reg_prof_syscall_exit(char *name)
 	return ret;
 }
 
-void unreg_prof_syscall_exit(char *name)
+void prof_sysexit_disable(struct ftrace_event_call *call)
 {
 	int num;
 
-	num = syscall_name_to_nr(name);
-	if (num < 0 || num >= NR_syscalls)
-		return;
+	num = ((struct syscall_metadata *)call->data)->syscall_nr;
 
 	mutex_lock(&syscall_trace_lock);
 	sys_prof_refcount_exit--;