lttng: add system call instrumentation probe

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

drivers/staging/lttng/lttng-syscalls.c

@@ -0,0 +1,438 @@
+/*
+ * lttng-syscalls.c
+ *
+ * Copyright 2010-2011 (c) - Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
+ *
+ * LTTng syscall probes.
+ *
+ * Dual LGPL v2.1/GPL v2 license.
+ */
+
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/compat.h>
+#include <asm/ptrace.h>
+#include <asm/syscall.h>
+
+#include "ltt-events.h"
+
+#ifndef CONFIG_COMPAT
+static inline int is_compat_task(void)
+{
+	return 0;
+}
+#endif
+
+static
+void syscall_entry_probe(void *__data, struct pt_regs *regs, long id);
+
+/*
+ * Take care of NOARGS not supported by mainline.
+ */
+#define DECLARE_EVENT_CLASS_NOARGS(name, tstruct, assign, print)
+#define DEFINE_EVENT_NOARGS(template, name)
+#define TRACE_EVENT_NOARGS(name, struct, assign, print)
+
+/*
+ * Create LTTng tracepoint probes.
+ */
+#define LTTNG_PACKAGE_BUILD
+#define CREATE_TRACE_POINTS
+#define TP_MODULE_OVERRIDE
+#define TRACE_INCLUDE_PATH ../instrumentation/syscalls/headers
+
+#define PARAMS(args...)	args
+
+#undef TRACE_SYSTEM
+
+/* Hijack probe callback for system calls */
+#undef TP_PROBE_CB
+#define TP_PROBE_CB(_template)		&syscall_entry_probe
+#define SC_TRACE_EVENT(_name, _proto, _args, _struct, _assign, _printk)	\
+	TRACE_EVENT(_name, PARAMS(_proto), PARAMS(_args),\
+		PARAMS(_struct), PARAMS(_assign), PARAMS(_printk))
+#define SC_DECLARE_EVENT_CLASS_NOARGS(_name, _struct, _assign, _printk)	\
+	DECLARE_EVENT_CLASS_NOARGS(_name, PARAMS(_struct), PARAMS(_assign),\
+		PARAMS(_printk))
+#define SC_DEFINE_EVENT_NOARGS(_template, _name)			\
+	DEFINE_EVENT_NOARGS(_template, _name)
+#define TRACE_SYSTEM syscalls_integers
+#include "instrumentation/syscalls/headers/syscalls_integers.h"
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM syscalls_pointers
+#include "instrumentation/syscalls/headers/syscalls_pointers.h"
+#undef TRACE_SYSTEM
+#undef SC_TRACE_EVENT
+#undef SC_DECLARE_EVENT_CLASS_NOARGS
+#undef SC_DEFINE_EVENT_NOARGS
+
+#define TRACE_SYSTEM syscalls_unknown
+#include "instrumentation/syscalls/headers/syscalls_unknown.h"
+#undef TRACE_SYSTEM
+
+/* For compat syscalls */
+#undef _TRACE_SYSCALLS_integers_H
+#undef _TRACE_SYSCALLS_pointers_H
+
+/* Hijack probe callback for system calls */
+#undef TP_PROBE_CB
+#define TP_PROBE_CB(_template)		&syscall_entry_probe
+#define SC_TRACE_EVENT(_name, _proto, _args, _struct, _assign, _printk)	\
+	TRACE_EVENT(compat_##_name, PARAMS(_proto), PARAMS(_args),	\
+		PARAMS(_struct), PARAMS(_assign),			\
+		PARAMS(_printk))
+#define SC_DECLARE_EVENT_CLASS_NOARGS(_name, _struct, _assign, _printk) \
+	DECLARE_EVENT_CLASS_NOARGS(compat_##_name, PARAMS(_struct),	\
+		PARAMS(_assign), PARAMS(_printk))
+#define SC_DEFINE_EVENT_NOARGS(_template, _name)			\
+	DEFINE_EVENT_NOARGS(compat_##_template, compat_##_name)
+#define TRACE_SYSTEM compat_syscalls_integers
+#include "instrumentation/syscalls/headers/compat_syscalls_integers.h"
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM compat_syscalls_pointers
+#include "instrumentation/syscalls/headers/compat_syscalls_pointers.h"
+#undef TRACE_SYSTEM
+#undef SC_TRACE_EVENT
+#undef SC_DECLARE_EVENT_CLASS_NOARGS
+#undef SC_DEFINE_EVENT_NOARGS
+#undef TP_PROBE_CB
+
+#undef TP_MODULE_OVERRIDE
+#undef LTTNG_PACKAGE_BUILD
+#undef CREATE_TRACE_POINTS
+
+struct trace_syscall_entry {
+	void *func;
+	const struct lttng_event_desc *desc;
+	const struct lttng_event_field *fields;
+	unsigned int nrargs;
+};
+
+#define CREATE_SYSCALL_TABLE
+
+#undef TRACE_SYSCALL_TABLE
+#define TRACE_SYSCALL_TABLE(_template, _name, _nr, _nrargs)	\
+	[ _nr ] = {						\
+		.func = __event_probe__##_template,		\
+		.nrargs = (_nrargs),				\
+		.fields = __event_fields___##_template,		\
+		.desc = &__event_desc___##_name,		\
+	},
+
+static const struct trace_syscall_entry sc_table[] = {
+#include "instrumentation/syscalls/headers/syscalls_integers.h"
+#include "instrumentation/syscalls/headers/syscalls_pointers.h"
+};
+
+#undef TRACE_SYSCALL_TABLE
+#define TRACE_SYSCALL_TABLE(_template, _name, _nr, _nrargs)	\
+	[ _nr ] = {						\
+		.func = __event_probe__##compat_##_template,	\
+		.nrargs = (_nrargs),				\
+		.fields = __event_fields___##compat_##_template,\
+		.desc = &__event_desc___##compat_##_name,	\
+	},
+
+/* Create compatibility syscall table */
+const struct trace_syscall_entry compat_sc_table[] = {
+#include "instrumentation/syscalls/headers/compat_syscalls_integers.h"
+#include "instrumentation/syscalls/headers/compat_syscalls_pointers.h"
+};
+
+#undef CREATE_SYSCALL_TABLE
+
+static void syscall_entry_unknown(struct ltt_event *event,
+	struct pt_regs *regs, unsigned int id)
+{
+	unsigned long args[UNKNOWN_SYSCALL_NRARGS];
+
+	syscall_get_arguments(current, regs, 0, UNKNOWN_SYSCALL_NRARGS, args);
+	if (unlikely(is_compat_task()))
+		__event_probe__compat_sys_unknown(event, id, args);
+	else
+		__event_probe__sys_unknown(event, id, args);
+}
+
+void syscall_entry_probe(void *__data, struct pt_regs *regs, long id)
+{
+	struct ltt_channel *chan = __data;
+	struct ltt_event *event, *unknown_event;
+	const struct trace_syscall_entry *table, *entry;
+	size_t table_len;
+
+	if (unlikely(is_compat_task())) {
+		table = compat_sc_table;
+		table_len = ARRAY_SIZE(compat_sc_table);
+		unknown_event = chan->sc_compat_unknown;
+	} else {
+		table = sc_table;
+		table_len = ARRAY_SIZE(sc_table);
+		unknown_event = chan->sc_unknown;
+	}
+	if (unlikely(id >= table_len)) {
+		syscall_entry_unknown(unknown_event, regs, id);
+		return;
+	}
+	if (unlikely(is_compat_task()))
+		event = chan->compat_sc_table[id];
+	else
+		event = chan->sc_table[id];
+	if (unlikely(!event)) {
+		syscall_entry_unknown(unknown_event, regs, id);
+		return;
+	}
+	entry = &table[id];
+	WARN_ON_ONCE(!entry);
+
+	switch (entry->nrargs) {
+	case 0:
+	{
+		void (*fptr)(void *__data) = entry->func;
+
+		fptr(event);
+		break;
+	}
+	case 1:
+	{
+		void (*fptr)(void *__data, unsigned long arg0) = entry->func;
+		unsigned long args[1];
+
+		syscall_get_arguments(current, regs, 0, entry->nrargs, args);
+		fptr(event, args[0]);
+		break;
+	}
+	case 2:
+	{
+		void (*fptr)(void *__data,
+			unsigned long arg0,
+			unsigned long arg1) = entry->func;
+		unsigned long args[2];
+
+		syscall_get_arguments(current, regs, 0, entry->nrargs, args);
+		fptr(event, args[0], args[1]);
+		break;
+	}
+	case 3:
+	{
+		void (*fptr)(void *__data,
+			unsigned long arg0,
+			unsigned long arg1,
+			unsigned long arg2) = entry->func;
+		unsigned long args[3];
+
+		syscall_get_arguments(current, regs, 0, entry->nrargs, args);
+		fptr(event, args[0], args[1], args[2]);
+		break;
+	}
+	case 4:
+	{
+		void (*fptr)(void *__data,
+			unsigned long arg0,
+			unsigned long arg1,
+			unsigned long arg2,
+			unsigned long arg3) = entry->func;
+		unsigned long args[4];
+
+		syscall_get_arguments(current, regs, 0, entry->nrargs, args);
+		fptr(event, args[0], args[1], args[2], args[3]);
+		break;
+	}
+	case 5:
+	{
+		void (*fptr)(void *__data,
+			unsigned long arg0,
+			unsigned long arg1,
+			unsigned long arg2,
+			unsigned long arg3,
+			unsigned long arg4) = entry->func;
+		unsigned long args[5];
+
+		syscall_get_arguments(current, regs, 0, entry->nrargs, args);
+		fptr(event, args[0], args[1], args[2], args[3], args[4]);
+		break;
+	}
+	case 6:
+	{
+		void (*fptr)(void *__data,
+			unsigned long arg0,
+			unsigned long arg1,
+			unsigned long arg2,
+			unsigned long arg3,
+			unsigned long arg4,
+			unsigned long arg5) = entry->func;
+		unsigned long args[6];
+
+		syscall_get_arguments(current, regs, 0, entry->nrargs, args);
+		fptr(event, args[0], args[1], args[2],
+			args[3], args[4], args[5]);
+		break;
+	}
+	default:
+		break;
+	}
+}
+
+/* noinline to diminish caller stack size */
+static
+int fill_table(const struct trace_syscall_entry *table, size_t table_len,
+	struct ltt_event **chan_table, struct ltt_channel *chan, void *filter)
+{
+	const struct lttng_event_desc *desc;
+	unsigned int i;
+
+	/* Allocate events for each syscall, insert into table */
+	for (i = 0; i < table_len; i++) {
+		struct lttng_kernel_event ev;
+		desc = table[i].desc;
+
+		if (!desc) {
+			/* Unknown syscall */
+			continue;
+		}
+		/*
+		 * Skip those already populated by previous failed
+		 * register for this channel.
+		 */
+		if (chan_table[i])
+			continue;
+		memset(&ev, 0, sizeof(ev));
+		strncpy(ev.name, desc->name, LTTNG_SYM_NAME_LEN);
+		ev.name[LTTNG_SYM_NAME_LEN - 1] = '\0';
+		ev.instrumentation = LTTNG_KERNEL_NOOP;
+		chan_table[i] = ltt_event_create(chan, &ev, filter,
+						desc);
+		if (!chan_table[i]) {
+			/*
+			 * If something goes wrong in event registration
+			 * after the first one, we have no choice but to
+			 * leave the previous events in there, until
+			 * deleted by session teardown.
+			 */
+			return -EINVAL;
+		}
+	}
+	return 0;
+}
+
+int lttng_syscalls_register(struct ltt_channel *chan, void *filter)
+{
+	struct lttng_kernel_event ev;
+	int ret;
+
+	wrapper_vmalloc_sync_all();
+
+	if (!chan->sc_table) {
+		/* create syscall table mapping syscall to events */
+		chan->sc_table = kzalloc(sizeof(struct ltt_event *)
+					* ARRAY_SIZE(sc_table), GFP_KERNEL);
+		if (!chan->sc_table)
+			return -ENOMEM;
+	}
+
+#ifdef CONFIG_COMPAT
+	if (!chan->compat_sc_table) {
+		/* create syscall table mapping compat syscall to events */
+		chan->compat_sc_table = kzalloc(sizeof(struct ltt_event *)
+					* ARRAY_SIZE(compat_sc_table), GFP_KERNEL);
+		if (!chan->compat_sc_table)
+			return -ENOMEM;
+	}
+#endif
+	if (!chan->sc_unknown) {
+		const struct lttng_event_desc *desc =
+			&__event_desc___sys_unknown;
+
+		memset(&ev, 0, sizeof(ev));
+		strncpy(ev.name, desc->name, LTTNG_SYM_NAME_LEN);
+		ev.name[LTTNG_SYM_NAME_LEN - 1] = '\0';
+		ev.instrumentation = LTTNG_KERNEL_NOOP;
+		chan->sc_unknown = ltt_event_create(chan, &ev, filter,
+						    desc);
+		if (!chan->sc_unknown) {
+			return -EINVAL;
+		}
+	}
+
+	if (!chan->sc_compat_unknown) {
+		const struct lttng_event_desc *desc =
+			&__event_desc___compat_sys_unknown;
+
+		memset(&ev, 0, sizeof(ev));
+		strncpy(ev.name, desc->name, LTTNG_SYM_NAME_LEN);
+		ev.name[LTTNG_SYM_NAME_LEN - 1] = '\0';
+		ev.instrumentation = LTTNG_KERNEL_NOOP;
+		chan->sc_compat_unknown = ltt_event_create(chan, &ev, filter,
+							   desc);
+		if (!chan->sc_compat_unknown) {
+			return -EINVAL;
+		}
+	}
+
+	if (!chan->sc_exit) {
+		const struct lttng_event_desc *desc =
+			&__event_desc___exit_syscall;
+
+		memset(&ev, 0, sizeof(ev));
+		strncpy(ev.name, desc->name, LTTNG_SYM_NAME_LEN);
+		ev.name[LTTNG_SYM_NAME_LEN - 1] = '\0';
+		ev.instrumentation = LTTNG_KERNEL_NOOP;
+		chan->sc_exit = ltt_event_create(chan, &ev, filter,
+						 desc);
+		if (!chan->sc_exit) {
+			return -EINVAL;
+		}
+	}
+
+	ret = fill_table(sc_table, ARRAY_SIZE(sc_table),
+			chan->sc_table, chan, filter);
+	if (ret)
+		return ret;
+#ifdef CONFIG_COMPAT
+	ret = fill_table(compat_sc_table, ARRAY_SIZE(compat_sc_table),
+			chan->compat_sc_table, chan, filter);
+	if (ret)
+		return ret;
+#endif
+	ret = tracepoint_probe_register("sys_enter",
+			(void *) syscall_entry_probe, chan);
+	if (ret)
+		return ret;
+	/*
+	 * We change the name of sys_exit tracepoint due to namespace
+	 * conflict with sys_exit syscall entry.
+	 */
+	ret = tracepoint_probe_register("sys_exit",
+			(void *) __event_probe__exit_syscall,
+			chan->sc_exit);
+	if (ret) {
+		WARN_ON_ONCE(tracepoint_probe_unregister("sys_enter",
+			(void *) syscall_entry_probe, chan));
+	}
+	return ret;
+}
+
+/*
+ * Only called at session destruction.
+ */
+int lttng_syscalls_unregister(struct ltt_channel *chan)
+{
+	int ret;
+
+	if (!chan->sc_table)
+		return 0;
+	ret = tracepoint_probe_unregister("sys_exit",
+			(void *) __event_probe__exit_syscall,
+			chan->sc_exit);
+	if (ret)
+		return ret;
+	ret = tracepoint_probe_unregister("sys_enter",
+			(void *) syscall_entry_probe, chan);
+	if (ret)
+		return ret;
+	/* ltt_event destroy will be performed by ltt_session_destroy() */
+	kfree(chan->sc_table);
+#ifdef CONFIG_COMPAT
+	kfree(chan->compat_sc_table);
+#endif
+	return 0;
+}