diff options
author | Frederic Weisbecker <fweisbec@gmail.com> | 2008-11-26 19:46:33 -0500 |
---|---|---|
committer | Ingo Molnar <mingo@elte.hu> | 2008-11-27 04:59:14 -0500 |
commit | 83a8df618eb04bd2819a758f3b409b1449862434 (patch) | |
tree | f6a8d6a75e2c94d7c0cfb475f11d5bd4adc20cf8 /kernel | |
parent | c7cc77307669336a08928ab8668bdb3f3bcc021b (diff) |
tracing/function-graph-tracer: enhancements for the trace output
Impact: enhance the output of the graph-tracer
This patch applies some ideas of Ingo Molnar and Steven Rostedt.
* Output leaf functions in one line with parenthesis, semicolon and duration
output.
* Add a second column (after cpu) for an overhead sign.
if duration > 100 us, "!"
if duration > 10 us, "+"
else " "
* Print output in us with remaining nanosec: u.n
* Print duration on the right end, following the indentation of the functions.
Use also visual clues: "-" on entry call (no duration to output) and "+" on
return (duration output).
The name of the tracer has been fixed as well: function-branch becomes
function_branch.
Here is an example of the new output:
CPU[000] dequeue_entity() { -
CPU[000] update_curr() { -
CPU[000] update_min_vruntime(); + 0.512 us
CPU[000] } + 1.504 us
CPU[000] clear_buddies(); + 0.481 us
CPU[000] update_min_vruntime(); + 0.504 us
CPU[000] } + 4.557 us
CPU[000] hrtick_update() { -
CPU[000] hrtick_start_fair(); + 0.489 us
CPU[000] } + 1.443 us
CPU[000] + } + 14.655 us
CPU[000] + } + 15.678 us
CPU[000] + } + 16.686 us
CPU[000] msecs_to_jiffies(); + 0.481 us
CPU[000] put_prev_task_fair(); + 0.504 us
CPU[000] pick_next_task_fair(); + 0.482 us
CPU[000] pick_next_task_rt(); + 0.504 us
CPU[000] pick_next_task_fair(); + 0.481 us
CPU[000] pick_next_task_idle(); + 0.489 us
CPU[000] _spin_trylock(); + 0.655 us
CPU[000] _spin_unlock(); + 0.609 us
CPU[000] ------------8<---------- thread bash-2794 ------------8<----------
CPU[000] finish_task_switch() { -
CPU[000] _spin_unlock_irq(); + 0.722 us
CPU[000] } + 2.369 us
CPU[000] ! } + 501972.605 us
CPU[000] ! } + 501973.763 us
CPU[000] copy_from_read_buf() { -
CPU[000] _spin_lock_irqsave(); + 0.670 us
CPU[000] _spin_unlock_irqrestore(); + 0.699 us
CPU[000] copy_to_user() { -
CPU[000] might_fault() { -
CPU[000] __might_sleep(); + 0.503 us
CPU[000] } + 1.632 us
CPU[000] __copy_to_user_ll(); + 0.542 us
CPU[000] } + 3.858 us
CPU[000] tty_audit_add_data() { -
CPU[000] _spin_lock_irq(); + 0.609 us
CPU[000] _spin_unlock_irq(); + 0.624 us
CPU[000] } + 3.196 us
CPU[000] _spin_lock_irqsave(); + 0.624 us
CPU[000] _spin_unlock_irqrestore(); + 0.625 us
CPU[000] + } + 13.611 us
CPU[000] copy_from_read_buf() { -
CPU[000] _spin_lock_irqsave(); + 0.624 us
CPU[000] _spin_unlock_irqrestore(); + 0.616 us
CPU[000] } + 2.820 us
CPU[000]
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Diffstat (limited to 'kernel')
-rw-r--r-- | kernel/trace/trace_functions_graph.c | 168 |
1 files changed, 159 insertions, 9 deletions
diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index d31d695174aa..b958d60377b3 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c | |||
@@ -14,6 +14,10 @@ | |||
14 | #include "trace.h" | 14 | #include "trace.h" |
15 | 15 | ||
16 | #define TRACE_GRAPH_INDENT 2 | 16 | #define TRACE_GRAPH_INDENT 2 |
17 | /* Spaces between function call and time duration */ | ||
18 | #define TRACE_GRAPH_TIMESPACE_ENTRY " " | ||
19 | /* Spaces between function call and closing braces */ | ||
20 | #define TRACE_GRAPH_TIMESPACE_RET " " | ||
17 | 21 | ||
18 | #define TRACE_GRAPH_PRINT_OVERRUN 0x1 | 22 | #define TRACE_GRAPH_PRINT_OVERRUN 0x1 |
19 | static struct tracer_opt trace_opts[] = { | 23 | static struct tracer_opt trace_opts[] = { |
@@ -63,26 +67,130 @@ static int verif_pid(struct trace_seq *s, pid_t pid, int cpu) | |||
63 | last_pid[cpu] = pid; | 67 | last_pid[cpu] = pid; |
64 | comm = trace_find_cmdline(pid); | 68 | comm = trace_find_cmdline(pid); |
65 | 69 | ||
66 | return trace_seq_printf(s, "\nCPU[%03d]" | 70 | return trace_seq_printf(s, "\nCPU[%03d] " |
67 | " ------------8<---------- thread %s-%d" | 71 | " ------------8<---------- thread %s-%d" |
68 | " ------------8<----------\n\n", | 72 | " ------------8<----------\n\n", |
69 | cpu, comm, pid); | 73 | cpu, comm, pid); |
70 | } | 74 | } |
71 | 75 | ||
76 | static bool | ||
77 | trace_branch_is_leaf(struct trace_iterator *iter, | ||
78 | struct ftrace_graph_ent_entry *curr) | ||
79 | { | ||
80 | struct ring_buffer_iter *ring_iter; | ||
81 | struct ring_buffer_event *event; | ||
82 | struct ftrace_graph_ret_entry *next; | ||
83 | |||
84 | ring_iter = iter->buffer_iter[iter->cpu]; | ||
85 | |||
86 | if (!ring_iter) | ||
87 | return false; | ||
88 | |||
89 | event = ring_buffer_iter_peek(iter->buffer_iter[iter->cpu], NULL); | ||
90 | |||
91 | if (!event) | ||
92 | return false; | ||
93 | |||
94 | next = ring_buffer_event_data(event); | ||
95 | |||
96 | if (next->ent.type != TRACE_GRAPH_RET) | ||
97 | return false; | ||
98 | |||
99 | if (curr->ent.pid != next->ent.pid || | ||
100 | curr->graph_ent.func != next->ret.func) | ||
101 | return false; | ||
102 | |||
103 | return true; | ||
104 | } | ||
105 | |||
106 | |||
107 | static inline int | ||
108 | print_graph_duration(unsigned long long duration, struct trace_seq *s) | ||
109 | { | ||
110 | unsigned long nsecs_rem = do_div(duration, 1000); | ||
111 | return trace_seq_printf(s, "+ %llu.%lu us\n", duration, nsecs_rem); | ||
112 | } | ||
113 | |||
114 | /* Signal a overhead of time execution to the output */ | ||
115 | static int | ||
116 | print_graph_overhead(unsigned long long duration, struct trace_seq *s) | ||
117 | { | ||
118 | /* Duration exceeded 100 msecs */ | ||
119 | if (duration > 100000ULL) | ||
120 | return trace_seq_printf(s, "! "); | ||
121 | |||
122 | /* Duration exceeded 10 msecs */ | ||
123 | if (duration > 10000ULL) | ||
124 | return trace_seq_printf(s, "+ "); | ||
125 | |||
126 | return trace_seq_printf(s, " "); | ||
127 | } | ||
128 | |||
129 | /* Case of a leaf function on its call entry */ | ||
72 | static enum print_line_t | 130 | static enum print_line_t |
73 | print_graph_entry(struct ftrace_graph_ent *call, struct trace_seq *s, | 131 | print_graph_entry_leaf(struct trace_iterator *iter, |
74 | struct trace_entry *ent, int cpu) | 132 | struct ftrace_graph_ent_entry *entry, struct trace_seq *s) |
75 | { | 133 | { |
134 | struct ftrace_graph_ret_entry *ret_entry; | ||
135 | struct ftrace_graph_ret *graph_ret; | ||
136 | struct ring_buffer_event *event; | ||
137 | struct ftrace_graph_ent *call; | ||
138 | unsigned long long duration; | ||
76 | int i; | 139 | int i; |
77 | int ret; | 140 | int ret; |
78 | 141 | ||
79 | if (!verif_pid(s, ent->pid, cpu)) | 142 | event = ring_buffer_read(iter->buffer_iter[iter->cpu], NULL); |
143 | ret_entry = ring_buffer_event_data(event); | ||
144 | graph_ret = &ret_entry->ret; | ||
145 | call = &entry->graph_ent; | ||
146 | duration = graph_ret->rettime - graph_ret->calltime; | ||
147 | |||
148 | /* Overhead */ | ||
149 | ret = print_graph_overhead(duration, s); | ||
150 | if (!ret) | ||
80 | return TRACE_TYPE_PARTIAL_LINE; | 151 | return TRACE_TYPE_PARTIAL_LINE; |
81 | 152 | ||
82 | ret = trace_seq_printf(s, "CPU[%03d] ", cpu); | 153 | /* Function */ |
154 | for (i = 0; i < call->depth * TRACE_GRAPH_INDENT; i++) { | ||
155 | ret = trace_seq_printf(s, " "); | ||
156 | if (!ret) | ||
157 | return TRACE_TYPE_PARTIAL_LINE; | ||
158 | } | ||
159 | |||
160 | ret = seq_print_ip_sym(s, call->func, 0); | ||
161 | if (!ret) | ||
162 | return TRACE_TYPE_PARTIAL_LINE; | ||
163 | |||
164 | ret = trace_seq_printf(s, "();"); | ||
83 | if (!ret) | 165 | if (!ret) |
84 | return TRACE_TYPE_PARTIAL_LINE; | 166 | return TRACE_TYPE_PARTIAL_LINE; |
85 | 167 | ||
168 | /* Duration */ | ||
169 | ret = trace_seq_printf(s, TRACE_GRAPH_TIMESPACE_ENTRY); | ||
170 | if (!ret) | ||
171 | return TRACE_TYPE_PARTIAL_LINE; | ||
172 | |||
173 | ret = print_graph_duration(duration, s); | ||
174 | if (!ret) | ||
175 | return TRACE_TYPE_PARTIAL_LINE; | ||
176 | |||
177 | return TRACE_TYPE_HANDLED; | ||
178 | } | ||
179 | |||
180 | static enum print_line_t | ||
181 | print_graph_entry_nested(struct ftrace_graph_ent_entry *entry, | ||
182 | struct trace_seq *s) | ||
183 | { | ||
184 | int i; | ||
185 | int ret; | ||
186 | struct ftrace_graph_ent *call = &entry->graph_ent; | ||
187 | |||
188 | /* No overhead */ | ||
189 | ret = trace_seq_printf(s, " "); | ||
190 | if (!ret) | ||
191 | return TRACE_TYPE_PARTIAL_LINE; | ||
192 | |||
193 | /* Function */ | ||
86 | for (i = 0; i < call->depth * TRACE_GRAPH_INDENT; i++) { | 194 | for (i = 0; i < call->depth * TRACE_GRAPH_INDENT; i++) { |
87 | ret = trace_seq_printf(s, " "); | 195 | ret = trace_seq_printf(s, " "); |
88 | if (!ret) | 196 | if (!ret) |
@@ -93,26 +201,62 @@ print_graph_entry(struct ftrace_graph_ent *call, struct trace_seq *s, | |||
93 | if (!ret) | 201 | if (!ret) |
94 | return TRACE_TYPE_PARTIAL_LINE; | 202 | return TRACE_TYPE_PARTIAL_LINE; |
95 | 203 | ||
96 | ret = trace_seq_printf(s, "() {\n"); | 204 | ret = trace_seq_printf(s, "() {"); |
97 | if (!ret) | 205 | if (!ret) |
98 | return TRACE_TYPE_PARTIAL_LINE; | 206 | return TRACE_TYPE_PARTIAL_LINE; |
207 | |||
208 | /* No duration to print at this state */ | ||
209 | ret = trace_seq_printf(s, TRACE_GRAPH_TIMESPACE_ENTRY "-\n"); | ||
210 | if (!ret) | ||
211 | return TRACE_TYPE_PARTIAL_LINE; | ||
212 | |||
99 | return TRACE_TYPE_HANDLED; | 213 | return TRACE_TYPE_HANDLED; |
100 | } | 214 | } |
101 | 215 | ||
102 | static enum print_line_t | 216 | static enum print_line_t |
217 | print_graph_entry(struct ftrace_graph_ent_entry *field, struct trace_seq *s, | ||
218 | struct trace_iterator *iter, int cpu) | ||
219 | { | ||
220 | int ret; | ||
221 | struct trace_entry *ent = iter->ent; | ||
222 | |||
223 | if (!verif_pid(s, ent->pid, cpu)) | ||
224 | return TRACE_TYPE_PARTIAL_LINE; | ||
225 | |||
226 | ret = trace_seq_printf(s, "CPU[%03d] ", cpu); | ||
227 | if (!ret) | ||
228 | return TRACE_TYPE_PARTIAL_LINE; | ||
229 | |||
230 | if (trace_branch_is_leaf(iter, field)) | ||
231 | return print_graph_entry_leaf(iter, field, s); | ||
232 | else | ||
233 | return print_graph_entry_nested(field, s); | ||
234 | |||
235 | } | ||
236 | |||
237 | static enum print_line_t | ||
103 | print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, | 238 | print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, |
104 | struct trace_entry *ent, int cpu) | 239 | struct trace_entry *ent, int cpu) |
105 | { | 240 | { |
106 | int i; | 241 | int i; |
107 | int ret; | 242 | int ret; |
243 | unsigned long long duration = trace->rettime - trace->calltime; | ||
108 | 244 | ||
245 | /* Pid */ | ||
109 | if (!verif_pid(s, ent->pid, cpu)) | 246 | if (!verif_pid(s, ent->pid, cpu)) |
110 | return TRACE_TYPE_PARTIAL_LINE; | 247 | return TRACE_TYPE_PARTIAL_LINE; |
111 | 248 | ||
249 | /* Cpu */ | ||
112 | ret = trace_seq_printf(s, "CPU[%03d] ", cpu); | 250 | ret = trace_seq_printf(s, "CPU[%03d] ", cpu); |
113 | if (!ret) | 251 | if (!ret) |
114 | return TRACE_TYPE_PARTIAL_LINE; | 252 | return TRACE_TYPE_PARTIAL_LINE; |
115 | 253 | ||
254 | /* Overhead */ | ||
255 | ret = print_graph_overhead(duration, s); | ||
256 | if (!ret) | ||
257 | return TRACE_TYPE_PARTIAL_LINE; | ||
258 | |||
259 | /* Closing brace */ | ||
116 | for (i = 0; i < trace->depth * TRACE_GRAPH_INDENT; i++) { | 260 | for (i = 0; i < trace->depth * TRACE_GRAPH_INDENT; i++) { |
117 | ret = trace_seq_printf(s, " "); | 261 | ret = trace_seq_printf(s, " "); |
118 | if (!ret) | 262 | if (!ret) |
@@ -123,10 +267,16 @@ print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, | |||
123 | if (!ret) | 267 | if (!ret) |
124 | return TRACE_TYPE_PARTIAL_LINE; | 268 | return TRACE_TYPE_PARTIAL_LINE; |
125 | 269 | ||
126 | ret = trace_seq_printf(s, "%llu\n", trace->rettime - trace->calltime); | 270 | /* Duration */ |
271 | ret = trace_seq_printf(s, TRACE_GRAPH_TIMESPACE_RET); | ||
272 | if (!ret) | ||
273 | return TRACE_TYPE_PARTIAL_LINE; | ||
274 | |||
275 | ret = print_graph_duration(duration, s); | ||
127 | if (!ret) | 276 | if (!ret) |
128 | return TRACE_TYPE_PARTIAL_LINE; | 277 | return TRACE_TYPE_PARTIAL_LINE; |
129 | 278 | ||
279 | /* Overrun */ | ||
130 | if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERRUN) { | 280 | if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERRUN) { |
131 | ret = trace_seq_printf(s, " (Overruns: %lu)\n", | 281 | ret = trace_seq_printf(s, " (Overruns: %lu)\n", |
132 | trace->overrun); | 282 | trace->overrun); |
@@ -146,7 +296,7 @@ print_graph_function(struct trace_iterator *iter) | |||
146 | case TRACE_GRAPH_ENT: { | 296 | case TRACE_GRAPH_ENT: { |
147 | struct ftrace_graph_ent_entry *field; | 297 | struct ftrace_graph_ent_entry *field; |
148 | trace_assign_type(field, entry); | 298 | trace_assign_type(field, entry); |
149 | return print_graph_entry(&field->graph_ent, s, entry, | 299 | return print_graph_entry(field, s, iter, |
150 | iter->cpu); | 300 | iter->cpu); |
151 | } | 301 | } |
152 | case TRACE_GRAPH_RET: { | 302 | case TRACE_GRAPH_RET: { |
@@ -160,7 +310,7 @@ print_graph_function(struct trace_iterator *iter) | |||
160 | } | 310 | } |
161 | 311 | ||
162 | static struct tracer graph_trace __read_mostly = { | 312 | static struct tracer graph_trace __read_mostly = { |
163 | .name = "function-graph", | 313 | .name = "function_graph", |
164 | .init = graph_trace_init, | 314 | .init = graph_trace_init, |
165 | .reset = graph_trace_reset, | 315 | .reset = graph_trace_reset, |
166 | .print_line = print_graph_function, | 316 | .print_line = print_graph_function, |