xref: /openbmc/linux/Documentation/trace/kprobetrace.rst (revision 4b0aaacee51eb6592a03fdefd5ce97558518e291)
1==========================
2Kprobe-based Event Tracing
3==========================
4
5:Author: Masami Hiramatsu
6
7Overview
8--------
9These events are similar to tracepoint based events. Instead of Tracepoint,
10this is based on kprobes (kprobe and kretprobe). So it can probe wherever
11kprobes can probe (this means, all functions except those with
12__kprobes/nokprobe_inline annotation and those marked NOKPROBE_SYMBOL).
13Unlike the Tracepoint based event, this can be added and removed
14dynamically, on the fly.
15
16To enable this feature, build your kernel with CONFIG_KPROBE_EVENTS=y.
17
18Similar to the events tracer, this doesn't need to be activated via
19current_tracer. Instead of that, add probe points via
20/sys/kernel/debug/tracing/kprobe_events, and enable it via
21/sys/kernel/debug/tracing/events/kprobes/<EVENT>/enable.
22
23
24Synopsis of kprobe_events
25-------------------------
26::
27
28  p[:[GRP/]EVENT] [MOD:]SYM[+offs]|MEMADDR [FETCHARGS]	: Set a probe
29  r[MAXACTIVE][:[GRP/]EVENT] [MOD:]SYM[+0] [FETCHARGS]	: Set a return probe
30  -:[GRP/]EVENT						: Clear a probe
31
32 GRP		: Group name. If omitted, use "kprobes" for it.
33 EVENT		: Event name. If omitted, the event name is generated
34		  based on SYM+offs or MEMADDR.
35 MOD		: Module name which has given SYM.
36 SYM[+offs]	: Symbol+offset where the probe is inserted.
37 MEMADDR	: Address where the probe is inserted.
38 MAXACTIVE	: Maximum number of instances of the specified function that
39		  can be probed simultaneously, or 0 for the default value
40		  as defined in Documentation/kprobes.txt section 1.3.1.
41
42 FETCHARGS	: Arguments. Each probe can have up to 128 args.
43  %REG		: Fetch register REG
44  @ADDR		: Fetch memory at ADDR (ADDR should be in kernel)
45  @SYM[+|-offs]	: Fetch memory at SYM +|- offs (SYM should be a data symbol)
46  $stackN	: Fetch Nth entry of stack (N >= 0)
47  $stack	: Fetch stack address.
48  $retval	: Fetch return value.(*)
49  $comm		: Fetch current task comm.
50  +|-offs(FETCHARG) : Fetch memory at FETCHARG +|- offs address.(**)
51  NAME=FETCHARG : Set NAME as the argument name of FETCHARG.
52  FETCHARG:TYPE : Set TYPE as the type of FETCHARG. Currently, basic types
53		  (u8/u16/u32/u64/s8/s16/s32/s64), hexadecimal types
54		  (x8/x16/x32/x64), "string" and bitfield are supported.
55
56  (*) only for return probe.
57  (**) this is useful for fetching a field of data structures.
58
59Types
60-----
61Several types are supported for fetch-args. Kprobe tracer will access memory
62by given type. Prefix 's' and 'u' means those types are signed and unsigned
63respectively. 'x' prefix implies it is unsigned. Traced arguments are shown
64in decimal ('s' and 'u') or hexadecimal ('x'). Without type casting, 'x32'
65or 'x64' is used depends on the architecture (e.g. x86-32 uses x32, and
66x86-64 uses x64).
67String type is a special type, which fetches a "null-terminated" string from
68kernel space. This means it will fail and store NULL if the string container
69has been paged out.
70Bitfield is another special type, which takes 3 parameters, bit-width, bit-
71offset, and container-size (usually 32). The syntax is::
72
73 b<bit-width>@<bit-offset>/<container-size>
74
75For $comm, the default type is "string"; any other type is invalid.
76
77
78Per-Probe Event Filtering
79-------------------------
80Per-probe event filtering feature allows you to set different filter on each
81probe and gives you what arguments will be shown in trace buffer. If an event
82name is specified right after 'p:' or 'r:' in kprobe_events, it adds an event
83under tracing/events/kprobes/<EVENT>, at the directory you can see 'id',
84'enable', 'format', 'filter' and 'trigger'.
85
86enable:
87  You can enable/disable the probe by writing 1 or 0 on it.
88
89format:
90  This shows the format of this probe event.
91
92filter:
93  You can write filtering rules of this event.
94
95id:
96  This shows the id of this probe event.
97
98trigger:
99  This allows to install trigger commands which are executed when the event is
100  hit (for details, see Documentation/trace/events.rst, section 6).
101
102Event Profiling
103---------------
104You can check the total number of probe hits and probe miss-hits via
105/sys/kernel/debug/tracing/kprobe_profile.
106The first column is event name, the second is the number of probe hits,
107the third is the number of probe miss-hits.
108
109
110Usage examples
111--------------
112To add a probe as a new event, write a new definition to kprobe_events
113as below::
114
115  echo 'p:myprobe do_sys_open dfd=%ax filename=%dx flags=%cx mode=+4($stack)' > /sys/kernel/debug/tracing/kprobe_events
116
117This sets a kprobe on the top of do_sys_open() function with recording
1181st to 4th arguments as "myprobe" event. Note, which register/stack entry is
119assigned to each function argument depends on arch-specific ABI. If you unsure
120the ABI, please try to use probe subcommand of perf-tools (you can find it
121under tools/perf/).
122As this example shows, users can choose more familiar names for each arguments.
123::
124
125  echo 'r:myretprobe do_sys_open $retval' >> /sys/kernel/debug/tracing/kprobe_events
126
127This sets a kretprobe on the return point of do_sys_open() function with
128recording return value as "myretprobe" event.
129You can see the format of these events via
130/sys/kernel/debug/tracing/events/kprobes/<EVENT>/format.
131::
132
133  cat /sys/kernel/debug/tracing/events/kprobes/myprobe/format
134  name: myprobe
135  ID: 780
136  format:
137          field:unsigned short common_type;       offset:0;       size:2; signed:0;
138          field:unsigned char common_flags;       offset:2;       size:1; signed:0;
139          field:unsigned char common_preempt_count;       offset:3; size:1;signed:0;
140          field:int common_pid;   offset:4;       size:4; signed:1;
141
142          field:unsigned long __probe_ip; offset:12;      size:4; signed:0;
143          field:int __probe_nargs;        offset:16;      size:4; signed:1;
144          field:unsigned long dfd;        offset:20;      size:4; signed:0;
145          field:unsigned long filename;   offset:24;      size:4; signed:0;
146          field:unsigned long flags;      offset:28;      size:4; signed:0;
147          field:unsigned long mode;       offset:32;      size:4; signed:0;
148
149
150  print fmt: "(%lx) dfd=%lx filename=%lx flags=%lx mode=%lx", REC->__probe_ip,
151  REC->dfd, REC->filename, REC->flags, REC->mode
152
153You can see that the event has 4 arguments as in the expressions you specified.
154::
155
156  echo > /sys/kernel/debug/tracing/kprobe_events
157
158This clears all probe points.
159
160Or,
161::
162
163  echo -:myprobe >> kprobe_events
164
165This clears probe points selectively.
166
167Right after definition, each event is disabled by default. For tracing these
168events, you need to enable it.
169::
170
171  echo 1 > /sys/kernel/debug/tracing/events/kprobes/myprobe/enable
172  echo 1 > /sys/kernel/debug/tracing/events/kprobes/myretprobe/enable
173
174And you can see the traced information via /sys/kernel/debug/tracing/trace.
175::
176
177  cat /sys/kernel/debug/tracing/trace
178  # tracer: nop
179  #
180  #           TASK-PID    CPU#    TIMESTAMP  FUNCTION
181  #              | |       |          |         |
182             <...>-1447  [001] 1038282.286875: myprobe: (do_sys_open+0x0/0xd6) dfd=3 filename=7fffd1ec4440 flags=8000 mode=0
183             <...>-1447  [001] 1038282.286878: myretprobe: (sys_openat+0xc/0xe <- do_sys_open) $retval=fffffffffffffffe
184             <...>-1447  [001] 1038282.286885: myprobe: (do_sys_open+0x0/0xd6) dfd=ffffff9c filename=40413c flags=8000 mode=1b6
185             <...>-1447  [001] 1038282.286915: myretprobe: (sys_open+0x1b/0x1d <- do_sys_open) $retval=3
186             <...>-1447  [001] 1038282.286969: myprobe: (do_sys_open+0x0/0xd6) dfd=ffffff9c filename=4041c6 flags=98800 mode=10
187             <...>-1447  [001] 1038282.286976: myretprobe: (sys_open+0x1b/0x1d <- do_sys_open) $retval=3
188
189
190Each line shows when the kernel hits an event, and <- SYMBOL means kernel
191returns from SYMBOL(e.g. "sys_open+0x1b/0x1d <- do_sys_open" means kernel
192returns from do_sys_open to sys_open+0x1b).
193
194