xref: /openbmc/u-boot/doc/README.trace (revision 87a62bce)
1# SPDX-License-Identifier: GPL-2.0+
2#
3# Copyright (c) 2013 The Chromium OS Authors.
4
5Tracing in U-Boot
6=================
7
8U-Boot supports a simple tracing feature which allows a record of excecution
9to be collected and sent to a host machine for analysis. At present the
10main use for this is to profile boot time.
11
12
13Overview
14--------
15
16The trace feature uses GCC's instrument-functions feature to trace all
17function entry/exit points. These are then recorded in a memory buffer.
18The memory buffer can be saved to the host over a network link using
19tftpput or by writing to an attached memory device such as MMC.
20
21On the host, the file is first converted with a tool called 'proftool',
22which extracts useful information from it. The resulting trace output
23resembles that emitted by Linux's ftrace feature, so can be visually
24displayed by pytimechart.
25
26
27Quick-start using Sandbox
28-------------------------
29
30Sandbox is a build of U-Boot that can run under Linux so it is a convenient
31way of trying out tracing before you use it on your actual board. To do
32this, follow these steps:
33
34Add the following to include/configs/sandbox.h (if not already there)
35
36#define CONFIG_TRACE
37#define CONFIG_CMD_TRACE
38#define CONFIG_TRACE_BUFFER_SIZE		(16 << 20)
39#define CONFIG_TRACE_EARLY_SIZE		(8 << 20)
40#define CONFIG_TRACE_EARLY
41#define CONFIG_TRACE_EARLY_ADDR		0x00100000
42
43Build sandbox U-Boot with tracing enabled:
44
45$ make FTRACE=1 O=sandbox sandbox_config
46$ make FTRACE=1 O=sandbox
47
48Run sandbox, wait for a bit of trace information to appear, and then capture
49a trace:
50
51$ ./sandbox/u-boot
52
53
54U-Boot 2013.04-rc2-00100-ga72fcef (Apr 17 2013 - 19:25:24)
55
56DRAM:  128 MiB
57trace: enabled
58Using default environment
59
60In:    serial
61Out:   serial
62Err:   serial
63=>trace stats
64	671,406 function sites
65	 69,712 function calls
66	      0 untracked function calls
67	 73,373 traced function calls
68	     16 maximum observed call depth
69	     15 call depth limit
70	 66,491 calls not traced due to depth
71=>trace stats
72	671,406 function sites
73      1,279,450 function calls
74	      0 untracked function calls
75	950,490 traced function calls (333217 dropped due to overflow)
76	     16 maximum observed call depth
77	     15 call depth limit
78      1,275,767 calls not traced due to depth
79=>trace calls 0 e00000
80Call list dumped to 00000000, size 0xae0a40
81=>print
82baudrate=115200
83profbase=0
84profoffset=ae0a40
85profsize=e00000
86stderr=serial
87stdin=serial
88stdout=serial
89
90Environment size: 117/8188 bytes
91=>sb save host 0 trace 0 ${profoffset}
9211405888 bytes written in 10 ms (1.1 GiB/s)
93=>reset
94
95
96Then run proftool to convert the trace information to ftrace format.
97
98$ ./sandbox/tools/proftool -m sandbox/System.map -p trace dump-ftrace >trace.txt
99
100Finally run pytimechart to display it:
101
102$ pytimechart trace.txt
103
104Using this tool you can zoom and pan across the trace, with the function
105calls on the left and little marks representing the start and end of each
106function.
107
108
109CONFIG Options
110--------------
111
112- CONFIG_TRACE
113		Enables the trace feature in U-Boot.
114
115- CONFIG_CMD_TRACE
116		Enables the trace command.
117
118- CONFIG_TRACE_BUFFER_SIZE
119		Size of trace buffer to allocate for U-Boot. This buffer is
120		used after relocation, as a place to put function tracing
121		information. The address of the buffer is determined by
122		the relocation code.
123
124- CONFIG_TRACE_EARLY
125		Define this to start tracing early, before relocation.
126
127- CONFIG_TRACE_EARLY_SIZE
128		Size of 'early' trace buffer. Before U-Boot has relocated
129		it doesn't have a proper trace buffer. On many boards
130		you can define an area of memory to use for the trace
131		buffer until the 'real' trace buffer is available after
132		relocation. The contents of this buffer are then copied to
133		the real buffer.
134
135- CONFIG_TRACE_EARLY_ADDR
136		Address of early trace buffer
137
138
139Building U-Boot with Tracing Enabled
140------------------------------------
141
142Pass 'FTRACE=1' to the U-Boot Makefile to actually instrument the code.
143This is kept as a separate option so that it is easy to enable/disable
144instrumenting from the command line instead of having to change board
145config files.
146
147
148Collecting Trace Data
149---------------------
150
151When you run U-Boot on your board it will collect trace data up to the
152limit of the trace buffer size you have specified. Once that is exhausted
153no more data will be collected.
154
155Collecting trace data has an affect on execution time/performance. You
156will notice this particularly with trvial functions - the overhead of
157recording their execution may even exceed their normal execution time.
158In practice this doesn't matter much so long as you are aware of the
159effect. Once you have done your optimisations, turn off tracing before
160doing end-to-end timing.
161
162The best time to start tracing is right at the beginning of U-Boot. The
163best time to stop tracing is right at the end. In practice it is hard
164to achieve these ideals.
165
166This implementation enables tracing early in board_init_f(). This means
167that it captures most of the board init process, missing only the
168early architecture-specific init. However, it also misses the entire
169SPL stage if there is one.
170
171U-Boot typically ends with a 'bootm' command which loads and runs an
172OS. There is useful trace data in the execution of that bootm
173command. Therefore this implementation provides a way to collect trace
174data after bootm has finished processing, but just before it jumps to
175the OS. In practical terms, U-Boot runs the 'fakegocmd' environment
176variable at this point. This variable should have a short script which
177collects the trace data and writes it somewhere.
178
179Trace data collection relies on a microsecond timer, accesed through
180timer_get_us(). So the first think you should do is make sure that
181this produces sensible results for your board. Suitable sources for
182this timer include high resolution timers, PWMs or profile timers if
183available. Most modern SOCs have a suitable timer for this. Make sure
184that you mark this timer (and anything it calls) with
185__attribute__((no_instrument_function)) so that the trace library can
186use it without causing an infinite loop.
187
188
189Commands
190--------
191
192The trace command has variable sub-commands:
193
194- stats
195		Display tracing statistics
196
197- pause
198		Pause tracing
199
200- resume
201		Resume tracing
202
203- funclist [<addr> <size>]
204		Dump a list of functions into the buffer
205
206- calls  [<addr> <size>]
207		Dump function call trace into buffer
208
209If the address and size are not given, these are obtained from environment
210variables (see below). In any case the environment variables are updated
211after the command runs.
212
213
214Environment Variables
215---------------------
216
217The following are used:
218
219- profbase
220		Base address of trace output buffer
221
222- profoffset
223		Offset of first unwritten byte in trace output buffer
224
225- profsize
226		Size of trace output buffer
227
228All of these are set by the 'trace calls' command.
229
230These variables keep track of the amount of data written to the trace
231output buffer by the 'trace' command. The trace commands which write data
232to the output buffer can use these to specify the buffer to write to, and
233update profoffset each time. This allows successive commands to append data
234to the same buffer, for example:
235
236	trace funclist 10000 e00000
237	trace calls
238
239(the latter command appends more data to the buffer).
240
241
242- fakegocmd
243		Specifies commands to run just before booting the OS. This
244		is a useful time to write the trace data to the host for
245		processing.
246
247
248Writing Out Trace Data
249----------------------
250
251Once the trace data is in an output buffer in memory there are various ways
252to transmit it to the host. Notably you can use tftput to send the data
253over a network link:
254
255fakegocmd=trace pause; usb start; set autoload n; bootp;
256	trace calls 10000000 1000000;
257	tftpput ${profbase} ${profoffset} 192.168.1.4:/tftpboot/calls
258
259This starts up USB (to talk to an attached USB Ethernet dongle), writes
260a trace log to address 10000000 and sends it to a host machine using
261TFTP. After this, U-Boot will boot the OS normally, albeit a little
262later.
263
264
265Converting Trace Output Data
266----------------------------
267
268The trace output data is kept in a binary format which is not documented
269here. To convert it into something useful, you can use proftool.
270
271This tool must be given the U-Boot map file and the trace data received
272from running that U-Boot. It produces a text output file.
273
274Options
275	-m <map_file>
276		Specify U-Boot map file
277
278	-p <trace_file>
279		Specifiy profile/trace file
280
281Commands:
282
283- dump-ftrace
284	Write a text dump of the file in Linux ftrace format to stdout
285
286
287Viewing the Trace Data
288----------------------
289
290You can use pytimechart for this (sudo apt-get pytimechart might work on
291your Debian-style machine, and use your favourite search engine to obtain
292documentation). It expects the file to have a .txt extension. The program
293has terse user interface but is very convenient for viewing U-Boot
294profile information.
295
296
297Workflow Suggestions
298--------------------
299
300The following suggestions may be helpful if you are trying to reduce boot
301time:
302
3031. Enable CONFIG_BOOTSTAGE and CONFIG_BOOTSTAGE_REPORT. This should get
304you are helpful overall snapshot of the boot time.
305
3062. Build U-Boot with tracing and run it. Note the difference in boot time
307(it is common for tracing to add 10% to the time)
308
3093. Collect the trace information as descibed above. Use this to find where
310all the time is being spent.
311
3124. Take a look at that code and see if you can optimise it. Perhaps it is
313possible to speed up the initialisation of a device, or remove an unused
314feature.
315
3165. Rebuild, run and collect again. Compare your results.
317
3186. Keep going until you run out of steam, or your boot is fast enough.
319
320
321Configuring Trace
322-----------------
323
324There are a few parameters in the code that you may want to consider.
325There is a function call depth limit (set to 15 by default). When the
326stack depth goes above this then no tracing information is recorded.
327The maximum depth reached is recorded and displayed by the 'trace stats'
328command.
329
330
331Future Work
332-----------
333
334Tracing could be a little tidier in some areas, for example providing
335run-time configuration options for trace.
336
337Some other features that might be useful:
338
339- Trace filter to select which functions are recorded
340- Sample-based profiling using a timer interrupt
341- Better control over trace depth
342- Compression of trace information
343
344
345Simon Glass <sjg@chromium.org>
346April 2013
347