xref: /openbmc/linux/Documentation/trace/kprobetrace.rst (revision 9625f579b7f2536c508a6e741606ca4102c64bc5)
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/debug/tracing/kprobe_events, and enable it via
21/sys/kernel/debug/tracing/events/kprobes/<EVENT>/enable.
22
23You can also use /sys/kernel/debug/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  $comm		: Fetch current task comm.
56  +|-[u]OFFS(FETCHARG) : Fetch memory at FETCHARG +|- OFFS address.(\*3)(\*4)
57  \IMM		: Store an immediate value to the argument.
58  NAME=FETCHARG : Set NAME as the argument name of FETCHARG.
59  FETCHARG:TYPE : Set TYPE as the type of FETCHARG. Currently, basic types
60		  (u8/u16/u32/u64/s8/s16/s32/s64), hexadecimal types
61		  (x8/x16/x32/x64), "string", "ustring", "symbol", "symstr"
62                  and bitfield are supported.
63
64  (\*1) only for the probe on function entry (offs == 0).
65  (\*2) only for return probe.
66  (\*3) this is useful for fetching a field of data structures.
67  (\*4) "u" means user-space dereference. See :ref:`user_mem_access`.
68
69Types
70-----
71Several types are supported for fetchargs. Kprobe tracer will access memory
72by given type. Prefix 's' and 'u' means those types are signed and unsigned
73respectively. 'x' prefix implies it is unsigned. Traced arguments are shown
74in decimal ('s' and 'u') or hexadecimal ('x'). Without type casting, 'x32'
75or 'x64' is used depends on the architecture (e.g. x86-32 uses x32, and
76x86-64 uses x64).
77
78These value types can be an array. To record array data, you can add '[N]'
79(where N is a fixed number, less than 64) to the base type.
80E.g. 'x16[4]' means an array of x16 (2-byte hex) with 4 elements.
81Note that the array can be applied to memory type fetchargs, you can not
82apply it to registers/stack-entries etc. (for example, '$stack1:x8[8]' is
83wrong, but '+8($stack):x8[8]' is OK.)
84
85String type is a special type, which fetches a "null-terminated" string from
86kernel space. This means it will fail and store NULL if the string container
87has been paged out. "ustring" type is an alternative of string for user-space.
88See :ref:`user_mem_access` for more info.
89
90The string array type is a bit different from other types. For other base
91types, <base-type>[1] is equal to <base-type> (e.g. +0(%di):x32[1] is same
92as +0(%di):x32.) But string[1] is not equal to string. The string type itself
93represents "char array", but string array type represents "char * array".
94So, for example, +0(%di):string[1] is equal to +0(+0(%di)):string.
95Bitfield is another special type, which takes 3 parameters, bit-width, bit-
96offset, and container-size (usually 32). The syntax is::
97
98 b<bit-width>@<bit-offset>/<container-size>
99
100Symbol type('symbol') is an alias of u32 or u64 type (depends on BITS_PER_LONG)
101which shows given pointer in "symbol+offset" style.
102On the other hand, symbol-string type ('symstr') converts the given address to
103"symbol+offset/symbolsize" style and stores it as a null-terminated string.
104With 'symstr' type, you can filter the event with wildcard pattern of the
105symbols, and you don't need to solve symbol name by yourself.
106For $comm, the default type is "string"; any other type is invalid.
107
108.. _user_mem_access:
109
110User Memory Access
111------------------
112Kprobe events supports user-space memory access. For that purpose, you can use
113either user-space dereference syntax or 'ustring' type.
114
115The user-space dereference syntax allows you to access a field of a data
116structure in user-space. This is done by adding the "u" prefix to the
117dereference syntax. For example, +u4(%si) means it will read memory from the
118address in the register %si offset by 4, and the memory is expected to be in
119user-space. You can use this for strings too, e.g. +u0(%si):string will read
120a string from the address in the register %si that is expected to be in user-
121space. 'ustring' is a shortcut way of performing the same task. That is,
122+0(%si):ustring is equivalent to +u0(%si):string.
123
124Note that kprobe-event provides the user-memory access syntax but it doesn't
125use it transparently. This means if you use normal dereference or string type
126for user memory, it might fail, and may always fail on some architectures. The
127user has to carefully check if the target data is in kernel or user space.
128
129Per-Probe Event Filtering
130-------------------------
131Per-probe event filtering feature allows you to set different filter on each
132probe and gives you what arguments will be shown in trace buffer. If an event
133name is specified right after 'p:' or 'r:' in kprobe_events, it adds an event
134under tracing/events/kprobes/<EVENT>, at the directory you can see 'id',
135'enable', 'format', 'filter' and 'trigger'.
136
137enable:
138  You can enable/disable the probe by writing 1 or 0 on it.
139
140format:
141  This shows the format of this probe event.
142
143filter:
144  You can write filtering rules of this event.
145
146id:
147  This shows the id of this probe event.
148
149trigger:
150  This allows to install trigger commands which are executed when the event is
151  hit (for details, see Documentation/trace/events.rst, section 6).
152
153Event Profiling
154---------------
155You can check the total number of probe hits and probe miss-hits via
156/sys/kernel/debug/tracing/kprobe_profile.
157The first column is event name, the second is the number of probe hits,
158the third is the number of probe miss-hits.
159
160Kernel Boot Parameter
161---------------------
162You can add and enable new kprobe events when booting up the kernel by
163"kprobe_event=" parameter. The parameter accepts a semicolon-delimited
164kprobe events, which format is similar to the kprobe_events.
165The difference is that the probe definition parameters are comma-delimited
166instead of space. For example, adding myprobe event on do_sys_open like below::
167
168  p:myprobe do_sys_open dfd=%ax filename=%dx flags=%cx mode=+4($stack)
169
170should be below for kernel boot parameter (just replace spaces with comma)::
171
172  p:myprobe,do_sys_open,dfd=%ax,filename=%dx,flags=%cx,mode=+4($stack)
173
174
175Usage examples
176--------------
177To add a probe as a new event, write a new definition to kprobe_events
178as below::
179
180  echo 'p:myprobe do_sys_open dfd=%ax filename=%dx flags=%cx mode=+4($stack)' > /sys/kernel/debug/tracing/kprobe_events
181
182This sets a kprobe on the top of do_sys_open() function with recording
1831st to 4th arguments as "myprobe" event. Note, which register/stack entry is
184assigned to each function argument depends on arch-specific ABI. If you unsure
185the ABI, please try to use probe subcommand of perf-tools (you can find it
186under tools/perf/).
187As this example shows, users can choose more familiar names for each arguments.
188::
189
190  echo 'r:myretprobe do_sys_open $retval' >> /sys/kernel/debug/tracing/kprobe_events
191
192This sets a kretprobe on the return point of do_sys_open() function with
193recording return value as "myretprobe" event.
194You can see the format of these events via
195/sys/kernel/debug/tracing/events/kprobes/<EVENT>/format.
196::
197
198  cat /sys/kernel/debug/tracing/events/kprobes/myprobe/format
199  name: myprobe
200  ID: 780
201  format:
202          field:unsigned short common_type;       offset:0;       size:2; signed:0;
203          field:unsigned char common_flags;       offset:2;       size:1; signed:0;
204          field:unsigned char common_preempt_count;       offset:3; size:1;signed:0;
205          field:int common_pid;   offset:4;       size:4; signed:1;
206
207          field:unsigned long __probe_ip; offset:12;      size:4; signed:0;
208          field:int __probe_nargs;        offset:16;      size:4; signed:1;
209          field:unsigned long dfd;        offset:20;      size:4; signed:0;
210          field:unsigned long filename;   offset:24;      size:4; signed:0;
211          field:unsigned long flags;      offset:28;      size:4; signed:0;
212          field:unsigned long mode;       offset:32;      size:4; signed:0;
213
214
215  print fmt: "(%lx) dfd=%lx filename=%lx flags=%lx mode=%lx", REC->__probe_ip,
216  REC->dfd, REC->filename, REC->flags, REC->mode
217
218You can see that the event has 4 arguments as in the expressions you specified.
219::
220
221  echo > /sys/kernel/debug/tracing/kprobe_events
222
223This clears all probe points.
224
225Or,
226::
227
228  echo -:myprobe >> kprobe_events
229
230This clears probe points selectively.
231
232Right after definition, each event is disabled by default. For tracing these
233events, you need to enable it.
234::
235
236  echo 1 > /sys/kernel/debug/tracing/events/kprobes/myprobe/enable
237  echo 1 > /sys/kernel/debug/tracing/events/kprobes/myretprobe/enable
238
239Use the following command to start tracing in an interval.
240::
241
242    # echo 1 > tracing_on
243    Open something...
244    # echo 0 > tracing_on
245
246And you can see the traced information via /sys/kernel/debug/tracing/trace.
247::
248
249  cat /sys/kernel/debug/tracing/trace
250  # tracer: nop
251  #
252  #           TASK-PID    CPU#    TIMESTAMP  FUNCTION
253  #              | |       |          |         |
254             <...>-1447  [001] 1038282.286875: myprobe: (do_sys_open+0x0/0xd6) dfd=3 filename=7fffd1ec4440 flags=8000 mode=0
255             <...>-1447  [001] 1038282.286878: myretprobe: (sys_openat+0xc/0xe <- do_sys_open) $retval=fffffffffffffffe
256             <...>-1447  [001] 1038282.286885: myprobe: (do_sys_open+0x0/0xd6) dfd=ffffff9c filename=40413c flags=8000 mode=1b6
257             <...>-1447  [001] 1038282.286915: myretprobe: (sys_open+0x1b/0x1d <- do_sys_open) $retval=3
258             <...>-1447  [001] 1038282.286969: myprobe: (do_sys_open+0x0/0xd6) dfd=ffffff9c filename=4041c6 flags=98800 mode=10
259             <...>-1447  [001] 1038282.286976: myretprobe: (sys_open+0x1b/0x1d <- do_sys_open) $retval=3
260
261
262Each line shows when the kernel hits an event, and <- SYMBOL means kernel
263returns from SYMBOL(e.g. "sys_open+0x1b/0x1d <- do_sys_open" means kernel
264returns from do_sys_open to sys_open+0x1b).
265