1 ========================== 1 ==========================
2 Kprobe-based Event Tracing 2 Kprobe-based Event Tracing
3 ========================== 3 ==========================
4 4
5 :Author: Masami Hiramatsu 5 :Author: Masami Hiramatsu
6 6
7 Overview 7 Overview
8 -------- 8 --------
9 These events are similar to tracepoint-based e 9 These events are similar to tracepoint-based events. Instead of tracepoints,
10 this is based on kprobes (kprobe and kretprobe 10 this is based on kprobes (kprobe and kretprobe). So it can probe wherever
11 kprobes can probe (this means, all functions e 11 kprobes can probe (this means, all functions except those with
12 __kprobes/nokprobe_inline annotation and those 12 __kprobes/nokprobe_inline annotation and those marked NOKPROBE_SYMBOL).
13 Unlike the tracepoint-based event, this can be 13 Unlike the tracepoint-based event, this can be added and removed
14 dynamically, on the fly. 14 dynamically, on the fly.
15 15
16 To enable this feature, build your kernel with 16 To enable this feature, build your kernel with CONFIG_KPROBE_EVENTS=y.
17 17
18 Similar to the event tracer, this doesn't need 18 Similar to the event tracer, this doesn't need to be activated via
19 current_tracer. Instead of that, add probe poi 19 current_tracer. Instead of that, add probe points via
20 /sys/kernel/tracing/kprobe_events, and enable 20 /sys/kernel/tracing/kprobe_events, and enable it via
21 /sys/kernel/tracing/events/kprobes/<EVENT>/ena 21 /sys/kernel/tracing/events/kprobes/<EVENT>/enable.
22 22
23 You can also use /sys/kernel/tracing/dynamic_e 23 You can also use /sys/kernel/tracing/dynamic_events instead of
24 kprobe_events. That interface will provide uni 24 kprobe_events. That interface will provide unified access to other
25 dynamic events too. 25 dynamic events too.
26 26
27 Synopsis of kprobe_events 27 Synopsis of kprobe_events
28 ------------------------- 28 -------------------------
29 :: 29 ::
30 30
31 p[:[GRP/][EVENT]] [MOD:]SYM[+offs]|MEMADDR [ 31 p[:[GRP/][EVENT]] [MOD:]SYM[+offs]|MEMADDR [FETCHARGS] : Set a probe
32 r[MAXACTIVE][:[GRP/][EVENT]] [MOD:]SYM[+0] [ 32 r[MAXACTIVE][:[GRP/][EVENT]] [MOD:]SYM[+0] [FETCHARGS] : Set a return probe
33 p[:[GRP/][EVENT]] [MOD:]SYM[+0]%return [FETC 33 p[:[GRP/][EVENT]] [MOD:]SYM[+0]%return [FETCHARGS] : Set a return probe
34 -:[GRP/][EVENT] 34 -:[GRP/][EVENT] : Clear a probe
35 35
36 GRP : Group name. If omitted, use 36 GRP : Group name. If omitted, use "kprobes" for it.
37 EVENT : Event name. If omitted, the 37 EVENT : Event name. If omitted, the event name is generated
38 based on SYM+offs or MEMADDR 38 based on SYM+offs or MEMADDR.
39 MOD : Module name which has given 39 MOD : Module name which has given SYM.
40 SYM[+offs] : Symbol+offset where the prob 40 SYM[+offs] : Symbol+offset where the probe is inserted.
41 SYM%return : Return address of the symbol 41 SYM%return : Return address of the symbol
42 MEMADDR : Address where the probe is i 42 MEMADDR : Address where the probe is inserted.
43 MAXACTIVE : Maximum number of instances 43 MAXACTIVE : Maximum number of instances of the specified function that
44 can be probed simultaneously 44 can be probed simultaneously, or 0 for the default value
45 as defined in Documentation/ 45 as defined in Documentation/trace/kprobes.rst section 1.3.1.
46 46
47 FETCHARGS : Arguments. Each probe can ha 47 FETCHARGS : Arguments. Each probe can have up to 128 args.
48 %REG : Fetch register REG 48 %REG : Fetch register REG
49 @ADDR : Fetch memory at ADDR (ADDR s 49 @ADDR : Fetch memory at ADDR (ADDR should be in kernel)
50 @SYM[+|-offs] : Fetch memory at SYM +|- offs 50 @SYM[+|-offs] : Fetch memory at SYM +|- offs (SYM should be a data symbol)
51 $stackN : Fetch Nth entry of stack (N 51 $stackN : Fetch Nth entry of stack (N >= 0)
52 $stack : Fetch stack address. 52 $stack : Fetch stack address.
53 $argN : Fetch the Nth function argum 53 $argN : Fetch the Nth function argument. (N >= 1) (\*1)
54 $retval : Fetch return value.(\*2) 54 $retval : Fetch return value.(\*2)
55 $comm : Fetch current task comm. 55 $comm : Fetch current task comm.
56 +|-[u]OFFS(FETCHARG) : Fetch memory at FETCH 56 +|-[u]OFFS(FETCHARG) : Fetch memory at FETCHARG +|- OFFS address.(\*3)(\*4)
57 \IMM : Store an immediate value to 57 \IMM : Store an immediate value to the argument.
58 NAME=FETCHARG : Set NAME as the argument nam 58 NAME=FETCHARG : Set NAME as the argument name of FETCHARG.
59 FETCHARG:TYPE : Set TYPE as the type of FETC 59 FETCHARG:TYPE : Set TYPE as the type of FETCHARG. Currently, basic types
60 (u8/u16/u32/u64/s8/s16/s32/s 60 (u8/u16/u32/u64/s8/s16/s32/s64), hexadecimal types
61 (x8/x16/x32/x64), VFS layer 61 (x8/x16/x32/x64), VFS layer common type(%pd/%pD), "char",
62 "string", "ustring", "symbol 62 "string", "ustring", "symbol", "symstr" and bitfield are
63 supported. 63 supported.
64 64
65 (\*1) only for the probe on function entry ( 65 (\*1) only for the probe on function entry (offs == 0). Note, this argument access
66 is best effort, because depending on t 66 is best effort, because depending on the argument type, it may be passed on
67 the stack. But this only support the a 67 the stack. But this only support the arguments via registers.
68 (\*2) only for return probe. Note that this 68 (\*2) only for return probe. Note that this is also best effort. Depending on the
69 return value type, it might be passed 69 return value type, it might be passed via a pair of registers. But this only
70 accesses one register. 70 accesses one register.
71 (\*3) this is useful for fetching a field of 71 (\*3) this is useful for fetching a field of data structures.
72 (\*4) "u" means user-space dereference. See 72 (\*4) "u" means user-space dereference. See :ref:`user_mem_access`.
73 73
74 Function arguments at kretprobe 74 Function arguments at kretprobe
75 ------------------------------- 75 -------------------------------
76 Function arguments can be accessed at kretprob 76 Function arguments can be accessed at kretprobe using $arg<N> fetcharg. This
77 is useful to record the function parameter and 77 is useful to record the function parameter and return value at once, and
78 trace the difference of structure fields (for 78 trace the difference of structure fields (for debugging a function whether it
79 correctly updates the given data structure or 79 correctly updates the given data structure or not).
80 See the :ref:`sample<fprobetrace_exit_args_sam 80 See the :ref:`sample<fprobetrace_exit_args_sample>` in fprobe event for how
81 it works. 81 it works.
82 82
83 .. _kprobetrace_types: 83 .. _kprobetrace_types:
84 84
85 Types 85 Types
86 ----- 86 -----
87 Several types are supported for fetchargs. Kpr 87 Several types are supported for fetchargs. Kprobe tracer will access memory
88 by given type. Prefix 's' and 'u' means those 88 by given type. Prefix 's' and 'u' means those types are signed and unsigned
89 respectively. 'x' prefix implies it is unsigne 89 respectively. 'x' prefix implies it is unsigned. Traced arguments are shown
90 in decimal ('s' and 'u') or hexadecimal ('x'). 90 in decimal ('s' and 'u') or hexadecimal ('x'). Without type casting, 'x32'
91 or 'x64' is used depends on the architecture ( 91 or 'x64' is used depends on the architecture (e.g. x86-32 uses x32, and
92 x86-64 uses x64). 92 x86-64 uses x64).
93 93
94 These value types can be an array. To record a 94 These value types can be an array. To record array data, you can add '[N]'
95 (where N is a fixed number, less than 64) to t 95 (where N is a fixed number, less than 64) to the base type.
96 E.g. 'x16[4]' means an array of x16 (2-byte he 96 E.g. 'x16[4]' means an array of x16 (2-byte hex) with 4 elements.
97 Note that the array can be applied to memory t 97 Note that the array can be applied to memory type fetchargs, you can not
98 apply it to registers/stack-entries etc. (for 98 apply it to registers/stack-entries etc. (for example, '$stack1:x8[8]' is
99 wrong, but '+8($stack):x8[8]' is OK.) 99 wrong, but '+8($stack):x8[8]' is OK.)
100 100
101 Char type can be used to show the character va 101 Char type can be used to show the character value of traced arguments.
102 102
103 String type is a special type, which fetches a 103 String type is a special type, which fetches a "null-terminated" string from
104 kernel space. This means it will fail and stor 104 kernel space. This means it will fail and store NULL if the string container
105 has been paged out. "ustring" type is an alter 105 has been paged out. "ustring" type is an alternative of string for user-space.
106 See :ref:`user_mem_access` for more info. 106 See :ref:`user_mem_access` for more info.
107 107
108 The string array type is a bit different from 108 The string array type is a bit different from other types. For other base
109 types, <base-type>[1] is equal to <base-type> 109 types, <base-type>[1] is equal to <base-type> (e.g. +0(%di):x32[1] is same
110 as +0(%di):x32.) But string[1] is not equal to 110 as +0(%di):x32.) But string[1] is not equal to string. The string type itself
111 represents "char array", but string array type 111 represents "char array", but string array type represents "char * array".
112 So, for example, +0(%di):string[1] is equal to 112 So, for example, +0(%di):string[1] is equal to +0(+0(%di)):string.
113 Bitfield is another special type, which takes 113 Bitfield is another special type, which takes 3 parameters, bit-width, bit-
114 offset, and container-size (usually 32). The s 114 offset, and container-size (usually 32). The syntax is::
115 115
116 b<bit-width>@<bit-offset>/<container-size> 116 b<bit-width>@<bit-offset>/<container-size>
117 117
118 Symbol type('symbol') is an alias of u32 or u6 118 Symbol type('symbol') is an alias of u32 or u64 type (depends on BITS_PER_LONG)
119 which shows given pointer in "symbol+offset" s 119 which shows given pointer in "symbol+offset" style.
120 On the other hand, symbol-string type ('symstr 120 On the other hand, symbol-string type ('symstr') converts the given address to
121 "symbol+offset/symbolsize" style and stores it 121 "symbol+offset/symbolsize" style and stores it as a null-terminated string.
122 With 'symstr' type, you can filter the event w 122 With 'symstr' type, you can filter the event with wildcard pattern of the
123 symbols, and you don't need to solve symbol na 123 symbols, and you don't need to solve symbol name by yourself.
124 For $comm, the default type is "string"; any o 124 For $comm, the default type is "string"; any other type is invalid.
125 125
126 VFS layer common type(%pd/%pD) is a special ty 126 VFS layer common type(%pd/%pD) is a special type, which fetches dentry's or
127 file's name from struct dentry's address or st 127 file's name from struct dentry's address or struct file's address.
128 128
129 .. _user_mem_access: 129 .. _user_mem_access:
130 130
131 User Memory Access 131 User Memory Access
132 ------------------ 132 ------------------
133 Kprobe events supports user-space memory acces 133 Kprobe events supports user-space memory access. For that purpose, you can use
134 either user-space dereference syntax or 'ustri 134 either user-space dereference syntax or 'ustring' type.
135 135
136 The user-space dereference syntax allows you t 136 The user-space dereference syntax allows you to access a field of a data
137 structure in user-space. This is done by addin 137 structure in user-space. This is done by adding the "u" prefix to the
138 dereference syntax. For example, +u4(%si) mean 138 dereference syntax. For example, +u4(%si) means it will read memory from the
139 address in the register %si offset by 4, and t 139 address in the register %si offset by 4, and the memory is expected to be in
140 user-space. You can use this for strings too, 140 user-space. You can use this for strings too, e.g. +u0(%si):string will read
141 a string from the address in the register %si 141 a string from the address in the register %si that is expected to be in user-
142 space. 'ustring' is a shortcut way of performi 142 space. 'ustring' is a shortcut way of performing the same task. That is,
143 +0(%si):ustring is equivalent to +u0(%si):stri 143 +0(%si):ustring is equivalent to +u0(%si):string.
144 144
145 Note that kprobe-event provides the user-memor 145 Note that kprobe-event provides the user-memory access syntax but it doesn't
146 use it transparently. This means if you use no 146 use it transparently. This means if you use normal dereference or string type
147 for user memory, it might fail, and may always 147 for user memory, it might fail, and may always fail on some architectures. The
148 user has to carefully check if the target data 148 user has to carefully check if the target data is in kernel or user space.
149 149
150 Per-Probe Event Filtering 150 Per-Probe Event Filtering
151 ------------------------- 151 -------------------------
152 Per-probe event filtering feature allows you t 152 Per-probe event filtering feature allows you to set different filter on each
153 probe and gives you what arguments will be sho 153 probe and gives you what arguments will be shown in trace buffer. If an event
154 name is specified right after 'p:' or 'r:' in 154 name is specified right after 'p:' or 'r:' in kprobe_events, it adds an event
155 under tracing/events/kprobes/<EVENT>, at the d 155 under tracing/events/kprobes/<EVENT>, at the directory you can see 'id',
156 'enable', 'format', 'filter' and 'trigger'. 156 'enable', 'format', 'filter' and 'trigger'.
157 157
158 enable: 158 enable:
159 You can enable/disable the probe by writing 159 You can enable/disable the probe by writing 1 or 0 on it.
160 160
161 format: 161 format:
162 This shows the format of this probe event. 162 This shows the format of this probe event.
163 163
164 filter: 164 filter:
165 You can write filtering rules of this event. 165 You can write filtering rules of this event.
166 166
167 id: 167 id:
168 This shows the id of this probe event. 168 This shows the id of this probe event.
169 169
170 trigger: 170 trigger:
171 This allows to install trigger commands whic 171 This allows to install trigger commands which are executed when the event is
172 hit (for details, see Documentation/trace/ev 172 hit (for details, see Documentation/trace/events.rst, section 6).
173 173
174 Event Profiling 174 Event Profiling
175 --------------- 175 ---------------
176 You can check the total number of probe hits a 176 You can check the total number of probe hits and probe miss-hits via
177 /sys/kernel/tracing/kprobe_profile. 177 /sys/kernel/tracing/kprobe_profile.
178 The first column is event name, the second is 178 The first column is event name, the second is the number of probe hits,
179 the third is the number of probe miss-hits. 179 the third is the number of probe miss-hits.
180 180
181 Kernel Boot Parameter 181 Kernel Boot Parameter
182 --------------------- 182 ---------------------
183 You can add and enable new kprobe events when 183 You can add and enable new kprobe events when booting up the kernel by
184 "kprobe_event=" parameter. The parameter accep 184 "kprobe_event=" parameter. The parameter accepts a semicolon-delimited
185 kprobe events, which format is similar to the 185 kprobe events, which format is similar to the kprobe_events.
186 The difference is that the probe definition pa 186 The difference is that the probe definition parameters are comma-delimited
187 instead of space. For example, adding myprobe 187 instead of space. For example, adding myprobe event on do_sys_open like below::
188 188
189 p:myprobe do_sys_open dfd=%ax filename=%dx f 189 p:myprobe do_sys_open dfd=%ax filename=%dx flags=%cx mode=+4($stack)
190 190
191 should be below for kernel boot parameter (jus 191 should be below for kernel boot parameter (just replace spaces with comma)::
192 192
193 p:myprobe,do_sys_open,dfd=%ax,filename=%dx,f 193 p:myprobe,do_sys_open,dfd=%ax,filename=%dx,flags=%cx,mode=+4($stack)
194 194
195 195
196 Usage examples 196 Usage examples
197 -------------- 197 --------------
198 To add a probe as a new event, write a new def 198 To add a probe as a new event, write a new definition to kprobe_events
199 as below:: 199 as below::
200 200
201 echo 'p:myprobe do_sys_open dfd=%ax filename 201 echo 'p:myprobe do_sys_open dfd=%ax filename=%dx flags=%cx mode=+4($stack)' > /sys/kernel/tracing/kprobe_events
202 202
203 This sets a kprobe on the top of do_sys_open() 203 This sets a kprobe on the top of do_sys_open() function with recording
204 1st to 4th arguments as "myprobe" event. Note, 204 1st to 4th arguments as "myprobe" event. Note, which register/stack entry is
205 assigned to each function argument depends on 205 assigned to each function argument depends on arch-specific ABI. If you unsure
206 the ABI, please try to use probe subcommand of 206 the ABI, please try to use probe subcommand of perf-tools (you can find it
207 under tools/perf/). 207 under tools/perf/).
208 As this example shows, users can choose more f 208 As this example shows, users can choose more familiar names for each arguments.
209 :: 209 ::
210 210
211 echo 'r:myretprobe do_sys_open $retval' >> / 211 echo 'r:myretprobe do_sys_open $retval' >> /sys/kernel/tracing/kprobe_events
212 212
213 This sets a kretprobe on the return point of d 213 This sets a kretprobe on the return point of do_sys_open() function with
214 recording return value as "myretprobe" event. 214 recording return value as "myretprobe" event.
215 You can see the format of these events via 215 You can see the format of these events via
216 /sys/kernel/tracing/events/kprobes/<EVENT>/for 216 /sys/kernel/tracing/events/kprobes/<EVENT>/format.
217 :: 217 ::
218 218
219 cat /sys/kernel/tracing/events/kprobes/mypro 219 cat /sys/kernel/tracing/events/kprobes/myprobe/format
220 name: myprobe 220 name: myprobe
221 ID: 780 221 ID: 780
222 format: 222 format:
223 field:unsigned short common_type; 223 field:unsigned short common_type; offset:0; size:2; signed:0;
224 field:unsigned char common_flags; 224 field:unsigned char common_flags; offset:2; size:1; signed:0;
225 field:unsigned char common_preempt_c 225 field:unsigned char common_preempt_count; offset:3; size:1;signed:0;
226 field:int common_pid; offset:4; 226 field:int common_pid; offset:4; size:4; signed:1;
227 227
228 field:unsigned long __probe_ip; offs 228 field:unsigned long __probe_ip; offset:12; size:4; signed:0;
229 field:int __probe_nargs; offs 229 field:int __probe_nargs; offset:16; size:4; signed:1;
230 field:unsigned long dfd; offs 230 field:unsigned long dfd; offset:20; size:4; signed:0;
231 field:unsigned long filename; offs 231 field:unsigned long filename; offset:24; size:4; signed:0;
232 field:unsigned long flags; offs 232 field:unsigned long flags; offset:28; size:4; signed:0;
233 field:unsigned long mode; offs 233 field:unsigned long mode; offset:32; size:4; signed:0;
234 234
235 235
236 print fmt: "(%lx) dfd=%lx filename=%lx flags 236 print fmt: "(%lx) dfd=%lx filename=%lx flags=%lx mode=%lx", REC->__probe_ip,
237 REC->dfd, REC->filename, REC->flags, REC->mo 237 REC->dfd, REC->filename, REC->flags, REC->mode
238 238
239 You can see that the event has 4 arguments as 239 You can see that the event has 4 arguments as in the expressions you specified.
240 :: 240 ::
241 241
242 echo > /sys/kernel/tracing/kprobe_events 242 echo > /sys/kernel/tracing/kprobe_events
243 243
244 This clears all probe points. 244 This clears all probe points.
245 245
246 Or, 246 Or,
247 :: 247 ::
248 248
249 echo -:myprobe >> kprobe_events 249 echo -:myprobe >> kprobe_events
250 250
251 This clears probe points selectively. 251 This clears probe points selectively.
252 252
253 Right after definition, each event is disabled 253 Right after definition, each event is disabled by default. For tracing these
254 events, you need to enable it. 254 events, you need to enable it.
255 :: 255 ::
256 256
257 echo 1 > /sys/kernel/tracing/events/kprobes/ 257 echo 1 > /sys/kernel/tracing/events/kprobes/myprobe/enable
258 echo 1 > /sys/kernel/tracing/events/kprobes/ 258 echo 1 > /sys/kernel/tracing/events/kprobes/myretprobe/enable
259 259
260 Use the following command to start tracing in 260 Use the following command to start tracing in an interval.
261 :: 261 ::
262 262
263 # echo 1 > tracing_on 263 # echo 1 > tracing_on
264 Open something... 264 Open something...
265 # echo 0 > tracing_on 265 # echo 0 > tracing_on
266 266
267 And you can see the traced information via /sy 267 And you can see the traced information via /sys/kernel/tracing/trace.
268 :: 268 ::
269 269
270 cat /sys/kernel/tracing/trace 270 cat /sys/kernel/tracing/trace
271 # tracer: nop 271 # tracer: nop
272 # 272 #
273 # TASK-PID CPU# TIMESTAMP F 273 # TASK-PID CPU# TIMESTAMP FUNCTION
274 # | | | | 274 # | | | | |
275 <...>-1447 [001] 1038282.286875: 275 <...>-1447 [001] 1038282.286875: myprobe: (do_sys_open+0x0/0xd6) dfd=3 filename=7fffd1ec4440 flags=8000 mode=0
276 <...>-1447 [001] 1038282.286878: 276 <...>-1447 [001] 1038282.286878: myretprobe: (sys_openat+0xc/0xe <- do_sys_open) $retval=fffffffffffffffe
277 <...>-1447 [001] 1038282.286885: 277 <...>-1447 [001] 1038282.286885: myprobe: (do_sys_open+0x0/0xd6) dfd=ffffff9c filename=40413c flags=8000 mode=1b6
278 <...>-1447 [001] 1038282.286915: 278 <...>-1447 [001] 1038282.286915: myretprobe: (sys_open+0x1b/0x1d <- do_sys_open) $retval=3
279 <...>-1447 [001] 1038282.286969: 279 <...>-1447 [001] 1038282.286969: myprobe: (do_sys_open+0x0/0xd6) dfd=ffffff9c filename=4041c6 flags=98800 mode=10
280 <...>-1447 [001] 1038282.286976: 280 <...>-1447 [001] 1038282.286976: myretprobe: (sys_open+0x1b/0x1d <- do_sys_open) $retval=3
281 281
282 282
283 Each line shows when the kernel hits an event, 283 Each line shows when the kernel hits an event, and <- SYMBOL means kernel
284 returns from SYMBOL(e.g. "sys_open+0x1b/0x1d < 284 returns from SYMBOL(e.g. "sys_open+0x1b/0x1d <- do_sys_open" means kernel
285 returns from do_sys_open to sys_open+0x1b). 285 returns from do_sys_open to sys_open+0x1b).
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.