Merge branch 'linus' into tracing/hw-branch-tracing
[sfrench/cifs-2.6.git] / kernel / trace / trace.h
index e7fbc826f1e93391faa03cee96d84c68094b4e7a..9e15802cca9f5c81fb4723ac1e4b24e9a63e2a31 100644 (file)
@@ -123,7 +123,6 @@ struct userstack_entry {
 struct bprint_entry {
        struct trace_entry      ent;
        unsigned long           ip;
-       int                     depth;
        const char              *fmt;
        u32                     buf[];
 };
@@ -131,7 +130,6 @@ struct bprint_entry {
 struct print_entry {
        struct trace_entry      ent;
        unsigned long           ip;
-       int                     depth;
        char                    buf[];
 };
 
@@ -184,6 +182,12 @@ struct trace_power {
        struct power_trace      state_data;
 };
 
+enum kmemtrace_type_id {
+       KMEMTRACE_TYPE_KMALLOC = 0,     /* kmalloc() or kfree(). */
+       KMEMTRACE_TYPE_CACHE,           /* kmem_cache_*(). */
+       KMEMTRACE_TYPE_PAGES,           /* __get_free_pages() and friends. */
+};
+
 struct kmemtrace_alloc_entry {
        struct trace_entry      ent;
        enum kmemtrace_type_id type_id;
@@ -202,6 +206,19 @@ struct kmemtrace_free_entry {
        const void *ptr;
 };
 
+struct syscall_trace_enter {
+       struct trace_entry      ent;
+       int                     nr;
+       unsigned long           args[];
+};
+
+struct syscall_trace_exit {
+       struct trace_entry      ent;
+       int                     nr;
+       unsigned long           ret;
+};
+
+
 /*
  * trace_flag_type is an enumeration that holds different
  * states when a trace occurs. These are:
@@ -315,6 +332,10 @@ extern void __ftrace_bad_type(void);
                          TRACE_KMEM_ALLOC);    \
                IF_ASSIGN(var, ent, struct kmemtrace_free_entry,        \
                          TRACE_KMEM_FREE);     \
+               IF_ASSIGN(var, ent, struct syscall_trace_enter,         \
+                         TRACE_SYSCALL_ENTER);                         \
+               IF_ASSIGN(var, ent, struct syscall_trace_exit,          \
+                         TRACE_SYSCALL_EXIT);                          \
                __ftrace_bad_type();                                    \
        } while (0)
 
@@ -468,6 +489,8 @@ trace_current_buffer_lock_reserve(unsigned char type, unsigned long len,
                                  unsigned long flags, int pc);
 void trace_current_buffer_unlock_commit(struct ring_buffer_event *event,
                                        unsigned long flags, int pc);
+void trace_nowake_buffer_unlock_commit(struct ring_buffer_event *event,
+                                       unsigned long flags, int pc);
 
 struct trace_entry *tracing_get_trace_entry(struct trace_array *tr,
                                                struct trace_array_cpu *data);
@@ -547,7 +570,7 @@ struct tracer_switch_ops {
 };
 #endif /* CONFIG_CONTEXT_SWITCH_TRACER */
 
-extern char *trace_find_cmdline(int pid);
+extern void trace_find_cmdline(int pid, char comm[]);
 
 #ifdef CONFIG_DYNAMIC_FTRACE
 extern unsigned long ftrace_update_tot_cnt;
@@ -583,9 +606,9 @@ extern int trace_selftest_startup_hw_branches(struct tracer *trace,
 extern void *head_page(struct trace_array_cpu *data);
 extern long ns2usecs(cycle_t nsec);
 extern int
-trace_vbprintk(unsigned long ip, int depth, const char *fmt, va_list args);
+trace_vbprintk(unsigned long ip, const char *fmt, va_list args);
 extern int
-trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args);
+trace_vprintk(unsigned long ip, const char *fmt, va_list args);
 
 extern unsigned long trace_flags;
 
@@ -669,6 +692,8 @@ enum trace_iterator_flags {
        TRACE_ITER_PRINTK_MSGONLY       = 0x10000,
        TRACE_ITER_CONTEXT_INFO         = 0x20000, /* Print pid/cpu/time */
        TRACE_ITER_LATENCY_FMT          = 0x40000,
+       TRACE_ITER_GLOBAL_CLK           = 0x80000,
+       TRACE_ITER_SLEEP_TIME           = 0x100000,
 };
 
 /*
@@ -761,22 +786,89 @@ enum {
        TRACE_EVENT_TYPE_RAW            = 2,
 };
 
+struct ftrace_event_field {
+       struct list_head        link;
+       char                    *name;
+       char                    *type;
+       int                     offset;
+       int                     size;
+};
+
 struct ftrace_event_call {
-       char            *name;
-       char            *system;
-       struct dentry   *dir;
-       int             enabled;
-       int             (*regfunc)(void);
-       void            (*unregfunc)(void);
-       int             id;
-       int             (*raw_init)(void);
-       int             (*show_format)(struct trace_seq *s);
+       char                    *name;
+       char                    *system;
+       struct dentry           *dir;
+       int                     enabled;
+       int                     (*regfunc)(void);
+       void                    (*unregfunc)(void);
+       int                     id;
+       int                     (*raw_init)(void);
+       int                     (*show_format)(struct trace_seq *s);
+       int                     (*define_fields)(void);
+       struct list_head        fields;
+       struct filter_pred      **preds;
+
+#ifdef CONFIG_EVENT_PROFILE
+       atomic_t        profile_count;
+       int             (*profile_enable)(struct ftrace_event_call *);
+       void            (*profile_disable)(struct ftrace_event_call *);
+#endif
 };
 
+struct event_subsystem {
+       struct list_head        list;
+       const char              *name;
+       struct dentry           *entry;
+       struct filter_pred      **preds;
+};
+
+#define events_for_each(event)                                         \
+       for (event = __start_ftrace_events;                             \
+            (unsigned long)event < (unsigned long)__stop_ftrace_events; \
+            event++)
+
+#define MAX_FILTER_PRED 8
+
+struct filter_pred;
+
+typedef int (*filter_pred_fn_t) (struct filter_pred *pred, void *event);
+
+struct filter_pred {
+       filter_pred_fn_t fn;
+       u64 val;
+       char *str_val;
+       int str_len;
+       char *field_name;
+       int offset;
+       int not;
+       int or;
+       int compound;
+       int clear;
+};
+
+int trace_define_field(struct ftrace_event_call *call, char *type,
+                      char *name, int offset, int size);
+extern void filter_free_pred(struct filter_pred *pred);
+extern void filter_print_preds(struct filter_pred **preds,
+                              struct trace_seq *s);
+extern int filter_parse(char **pbuf, struct filter_pred *pred);
+extern int filter_add_pred(struct ftrace_event_call *call,
+                          struct filter_pred *pred);
+extern void filter_free_preds(struct ftrace_event_call *call);
+extern int filter_match_preds(struct ftrace_event_call *call, void *rec);
+extern void filter_free_subsystem_preds(struct event_subsystem *system);
+extern int filter_add_subsystem_pred(struct event_subsystem *system,
+                                    struct filter_pred *pred);
+
 void event_trace_printk(unsigned long ip, const char *fmt, ...);
 extern struct ftrace_event_call __start_ftrace_events[];
 extern struct ftrace_event_call __stop_ftrace_events[];
 
+#define for_each_event(event)                                          \
+       for (event = __start_ftrace_events;                             \
+            (unsigned long)event < (unsigned long)__stop_ftrace_events; \
+            event++)
+
 extern const char *__start___trace_bprintk_fmt[];
 extern const char *__stop___trace_bprintk_fmt[];