xref: /linux/Documentation/trace/kprobetrace.rst (revision eb01fe7abbe2d0b38824d2a93fdb4cc3eaf2ccc1)
1==========================
2Kprobe-based Event Tracing
3==========================
4
5:Author: Masami Hiramatsu
6
7Overview
8--------
9These events are similar to tracepoint-based events. Instead of tracepoints,
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 event tracer, this doesn't need to be activated via
19current_tracer. Instead of that, add probe points via
20/sys/kernel/tracing/kprobe_events, and enable it via
21/sys/kernel/tracing/events/kprobes/<EVENT>/enable.
22
23You can also use /sys/kernel/tracing/dynamic_events instead of
24kprobe_events. That interface will provide unified access to other
25dynamic events too.
26
27Synopsis of kprobe_events
28-------------------------
29::
30
31  p[:[GRP/][EVENT]] [MOD:]SYM[+offs]|MEMADDR [FETCHARGS]	: Set a probe
32  r[MAXACTIVE][:[GRP/][EVENT]] [MOD:]SYM[+0] [FETCHARGS]	: Set a return probe
33  p[:[GRP/][EVENT]] [MOD:]SYM[+0]%return [FETCHARGS]	: Set a return probe
34  -:[GRP/][EVENT]						: Clear a probe
35
36 GRP		: Group name. If omitted, use "kprobes" for it.
37 EVENT		: Event name. If omitted, the event name is generated
38		  based on SYM+offs or MEMADDR.
39 MOD		: Module name which has given SYM.
40 SYM[+offs]	: Symbol+offset where the probe is inserted.
41 SYM%return	: Return address of the symbol
42 MEMADDR	: Address where the probe is inserted.
43 MAXACTIVE	: Maximum number of instances of the specified function that
44		  can be probed simultaneously, or 0 for the default value
45		  as defined in Documentation/trace/kprobes.rst section 1.3.1.
46
47 FETCHARGS	: Arguments. Each probe can have up to 128 args.
48  %REG		: Fetch register REG
49  @ADDR		: Fetch memory at ADDR (ADDR should be in kernel)
50  @SYM[+|-offs]	: Fetch memory at SYM +|- offs (SYM should be a data symbol)
51  $stackN	: Fetch Nth entry of stack (N >= 0)
52  $stack	: Fetch stack address.
53  $argN		: Fetch the Nth function argument. (N >= 1) (\*1)
54  $retval	: Fetch return value.(\*2)
55  $comm		: Fetch current task comm.
56  +|-[u]OFFS(FETCHARG) : Fetch memory at FETCHARG +|- OFFS address.(\*3)(\*4)
57  \IMM		: Store an immediate value to the argument.
58  NAME=FETCHARG : Set NAME as the argument name of FETCHARG.
59  FETCHARG:TYPE : Set TYPE as the type of FETCHARG. Currently, basic types
60		  (u8/u16/u32/u64/s8/s16/s32/s64), hexadecimal types
61		  (x8/x16/x32/x64), "char", "string", "ustring", "symbol", "symstr"
62                  and bitfield are supported.
63
64  (\*1) only for the probe on function entry (offs == 0). Note, this argument access
65        is best effort, because depending on the argument type, it may be passed on
66        the stack. But this only support the arguments via registers.
67  (\*2) only for return probe. Note that this is also best effort. Depending on the
68        return value type, it might be passed via a pair of registers. But this only
69        accesses one register.
70  (\*3) this is useful for fetching a field of data structures.
71  (\*4) "u" means user-space dereference. See :ref:`user_mem_access`.
72
73Function arguments at kretprobe
74-------------------------------
75Function arguments can be accessed at kretprobe using $arg<N> fetcharg. This
76is useful to record the function parameter and return value at once, and
77trace the difference of structure fields (for debuging a function whether it
78correctly updates the given data structure or not).
79See the :ref:`sample<fprobetrace_exit_args_sample>` in fprobe event for how
80it works.
81
82.. _kprobetrace_types:
83
84Types
85-----
86Several types are supported for fetchargs. Kprobe tracer will access memory
87by given type. Prefix 's' and 'u' means those types are signed and unsigned
88respectively. 'x' prefix implies it is unsigned. Traced arguments are shown
89in decimal ('s' and 'u') or hexadecimal ('x'). Without type casting, 'x32'
90or 'x64' is used depends on the architecture (e.g. x86-32 uses x32, and
91x86-64 uses x64).
92
93These value types can be an array. To record array data, you can add '[N]'
94(where N is a fixed number, less than 64) to the base type.
95E.g. 'x16[4]' means an array of x16 (2-byte hex) with 4 elements.
96Note that the array can be applied to memory type fetchargs, you can not
97apply it to registers/stack-entries etc. (for example, '$stack1:x8[8]' is
98wrong, but '+8($stack):x8[8]' is OK.)
99
100Char type can be used to show the character value of traced arguments.
101
102String type is a special type, which fetches a "null-terminated" string from
103kernel space. This means it will fail and store NULL if the string container
104has been paged out. "ustring" type is an alternative of string for user-space.
105See :ref:`user_mem_access` for more info.
106
107The string array type is a bit different from other types. For other base
108types, <base-type>[1] is equal to <base-type> (e.g. +0(%di):x32[1] is same
109as +0(%di):x32.) But string[1] is not equal to string. The string type itself
110represents "char array", but string array type represents "char * array".
111So, for example, +0(%di):string[1] is equal to +0(+0(%di)):string.
112Bitfield is another special type, which takes 3 parameters, bit-width, bit-
113offset, and container-size (usually 32). The syntax is::
114
115 b<bit-width>@<bit-offset>/<container-size>
116
117Symbol type('symbol') is an alias of u32 or u64 type (depends on BITS_PER_LONG)
118which shows given pointer in "symbol+offset" style.
119On the other hand, symbol-string type ('symstr') converts the given address to
120"symbol+offset/symbolsize" style and stores it as a null-terminated string.
121With 'symstr' type, you can filter the event with wildcard pattern of the
122symbols, and you don't need to solve symbol name by yourself.
123For $comm, the default type is "string"; any other type is invalid.
124
125.. _user_mem_access:
126
127User Memory Access
128------------------
129Kprobe events supports user-space memory access. For that purpose, you can use
130either user-space dereference syntax or 'ustring' type.
131
132The user-space dereference syntax allows you to access a field of a data
133structure in user-space. This is done by adding the "u" prefix to the
134dereference syntax. For example, +u4(%si) means it will read memory from the
135address in the register %si offset by 4, and the memory is expected to be in
136user-space. You can use this for strings too, e.g. +u0(%si):string will read
137a string from the address in the register %si that is expected to be in user-
138space. 'ustring' is a shortcut way of performing the same task. That is,
139+0(%si):ustring is equivalent to +u0(%si):string.
140
141Note that kprobe-event provides the user-memory access syntax but it doesn't
142use it transparently. This means if you use normal dereference or string type
143for user memory, it might fail, and may always fail on some architectures. The
144user has to carefully check if the target data is in kernel or user space.
145
146Per-Probe Event Filtering
147-------------------------
148Per-probe event filtering feature allows you to set different filter on each
149probe and gives you what arguments will be shown in trace buffer. If an event
150name is specified right after 'p:' or 'r:' in kprobe_events, it adds an event
151under tracing/events/kprobes/<EVENT>, at the directory you can see 'id',
152'enable', 'format', 'filter' and 'trigger'.
153
154enable:
155  You can enable/disable the probe by writing 1 or 0 on it.
156
157format:
158  This shows the format of this probe event.
159
160filter:
161  You can write filtering rules of this event.
162
163id:
164  This shows the id of this probe event.
165
166trigger:
167  This allows to install trigger commands which are executed when the event is
168  hit (for details, see Documentation/trace/events.rst, section 6).
169
170Event Profiling
171---------------
172You can check the total number of probe hits and probe miss-hits via
173/sys/kernel/tracing/kprobe_profile.
174The first column is event name, the second is the number of probe hits,
175the third is the number of probe miss-hits.
176
177Kernel Boot Parameter
178---------------------
179You can add and enable new kprobe events when booting up the kernel by
180"kprobe_event=" parameter. The parameter accepts a semicolon-delimited
181kprobe events, which format is similar to the kprobe_events.
182The difference is that the probe definition parameters are comma-delimited
183instead of space. For example, adding myprobe event on do_sys_open like below::
184
185  p:myprobe do_sys_open dfd=%ax filename=%dx flags=%cx mode=+4($stack)
186
187should be below for kernel boot parameter (just replace spaces with comma)::
188
189  p:myprobe,do_sys_open,dfd=%ax,filename=%dx,flags=%cx,mode=+4($stack)
190
191
192Usage examples
193--------------
194To add a probe as a new event, write a new definition to kprobe_events
195as below::
196
197  echo 'p:myprobe do_sys_open dfd=%ax filename=%dx flags=%cx mode=+4($stack)' > /sys/kernel/tracing/kprobe_events
198
199This sets a kprobe on the top of do_sys_open() function with recording
2001st to 4th arguments as "myprobe" event. Note, which register/stack entry is
201assigned to each function argument depends on arch-specific ABI. If you unsure
202the ABI, please try to use probe subcommand of perf-tools (you can find it
203under tools/perf/).
204As this example shows, users can choose more familiar names for each arguments.
205::
206
207  echo 'r:myretprobe do_sys_open $retval' >> /sys/kernel/tracing/kprobe_events
208
209This sets a kretprobe on the return point of do_sys_open() function with
210recording return value as "myretprobe" event.
211You can see the format of these events via
212/sys/kernel/tracing/events/kprobes/<EVENT>/format.
213::
214
215  cat /sys/kernel/tracing/events/kprobes/myprobe/format
216  name: myprobe
217  ID: 780
218  format:
219          field:unsigned short common_type;       offset:0;       size:2; signed:0;
220          field:unsigned char common_flags;       offset:2;       size:1; signed:0;
221          field:unsigned char common_preempt_count;       offset:3; size:1;signed:0;
222          field:int common_pid;   offset:4;       size:4; signed:1;
223
224          field:unsigned long __probe_ip; offset:12;      size:4; signed:0;
225          field:int __probe_nargs;        offset:16;      size:4; signed:1;
226          field:unsigned long dfd;        offset:20;      size:4; signed:0;
227          field:unsigned long filename;   offset:24;      size:4; signed:0;
228          field:unsigned long flags;      offset:28;      size:4; signed:0;
229          field:unsigned long mode;       offset:32;      size:4; signed:0;
230
231
232  print fmt: "(%lx) dfd=%lx filename=%lx flags=%lx mode=%lx", REC->__probe_ip,
233  REC->dfd, REC->filename, REC->flags, REC->mode
234
235You can see that the event has 4 arguments as in the expressions you specified.
236::
237
238  echo > /sys/kernel/tracing/kprobe_events
239
240This clears all probe points.
241
242Or,
243::
244
245  echo -:myprobe >> kprobe_events
246
247This clears probe points selectively.
248
249Right after definition, each event is disabled by default. For tracing these
250events, you need to enable it.
251::
252
253  echo 1 > /sys/kernel/tracing/events/kprobes/myprobe/enable
254  echo 1 > /sys/kernel/tracing/events/kprobes/myretprobe/enable
255
256Use the following command to start tracing in an interval.
257::
258
259    # echo 1 > tracing_on
260    Open something...
261    # echo 0 > tracing_on
262
263And you can see the traced information via /sys/kernel/tracing/trace.
264::
265
266  cat /sys/kernel/tracing/trace
267  # tracer: nop
268  #
269  #           TASK-PID    CPU#    TIMESTAMP  FUNCTION
270  #              | |       |          |         |
271             <...>-1447  [001] 1038282.286875: myprobe: (do_sys_open+0x0/0xd6) dfd=3 filename=7fffd1ec4440 flags=8000 mode=0
272             <...>-1447  [001] 1038282.286878: myretprobe: (sys_openat+0xc/0xe <- do_sys_open) $retval=fffffffffffffffe
273             <...>-1447  [001] 1038282.286885: myprobe: (do_sys_open+0x0/0xd6) dfd=ffffff9c filename=40413c flags=8000 mode=1b6
274             <...>-1447  [001] 1038282.286915: myretprobe: (sys_open+0x1b/0x1d <- do_sys_open) $retval=3
275             <...>-1447  [001] 1038282.286969: myprobe: (do_sys_open+0x0/0xd6) dfd=ffffff9c filename=4041c6 flags=98800 mode=10
276             <...>-1447  [001] 1038282.286976: myretprobe: (sys_open+0x1b/0x1d <- do_sys_open) $retval=3
277
278
279Each line shows when the kernel hits an event, and <- SYMBOL means kernel
280returns from SYMBOL(e.g. "sys_open+0x1b/0x1d <- do_sys_open" means kernel
281returns from do_sys_open to sys_open+0x1b).
282