1 Kprobe-based Event Tracing
2 ==========================
4 Documentation is written by Masami Hiramatsu
9 These events are similar to tracepoint based events. Instead of Tracepoint,
10 this is based on kprobes (kprobe and kretprobe). So it can probe wherever
11 kprobes can probe (this means, all functions except those with
12 __kprobes/nokprobe_inline annotation and those marked NOKPROBE_SYMBOL).
13 Unlike the Tracepoint based event, this can be added and removed
14 dynamically, on the fly.
16 To enable this feature, build your kernel with CONFIG_KPROBE_EVENTS=y.
18 Similar to the events tracer, this doesn't need to be activated via
19 current_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>/enabled.
24 Synopsis of kprobe_events
25 -------------------------
26 p[:[GRP/]EVENT] [MOD:]SYM[+offs]|MEMADDR [FETCHARGS] : Set a probe
27 r[:[GRP/]EVENT] [MOD:]SYM[+0] [FETCHARGS] : Set a return probe
28 -:[GRP/]EVENT : Clear a probe
30 GRP : Group name. If omitted, use "kprobes" for it.
31 EVENT : Event name. If omitted, the event name is generated
32 based on SYM+offs or MEMADDR.
33 MOD : Module name which has given SYM.
34 SYM[+offs] : Symbol+offset where the probe is inserted.
35 MEMADDR : Address where the probe is inserted.
37 FETCHARGS : Arguments. Each probe can have up to 128 args.
38 %REG : Fetch register REG
39 @ADDR : Fetch memory at ADDR (ADDR should be in kernel)
40 @SYM[+|-offs] : Fetch memory at SYM +|- offs (SYM should be a data symbol)
41 $stackN : Fetch Nth entry of stack (N >= 0)
42 $stack : Fetch stack address.
43 $retval : Fetch return value.(*)
44 $comm : Fetch current task comm.
45 +|-offs(FETCHARG) : Fetch memory at FETCHARG +|- offs address.(**)
46 NAME=FETCHARG : Set NAME as the argument name of FETCHARG.
47 FETCHARG:TYPE : Set TYPE as the type of FETCHARG. Currently, basic types
48 (u8/u16/u32/u64/s8/s16/s32/s64), hexadecimal types
49 (x8/x16/x32/x64), "string" and bitfield are supported.
51 (*) only for return probe.
52 (**) this is useful for fetching a field of data structures.
56 Several types are supported for fetch-args. Kprobe tracer will access memory
57 by given type. Prefix 's' and 'u' means those types are signed and unsigned
58 respectively. 'x' prefix implies it is unsigned. Traced arguments are shown
59 in decimal ('s' and 'u') or hexadecimal ('x'). Without type casting, 'x32'
60 or 'x64' is used depends on the architecture (e.g. x86-32 uses x32, and
62 String type is a special type, which fetches a "null-terminated" string from
63 kernel space. This means it will fail and store NULL if the string container
65 Bitfield is another special type, which takes 3 parameters, bit-width, bit-
66 offset, and container-size (usually 32). The syntax is;
68 b<bit-width>@<bit-offset>/<container-size>
70 For $comm, the default type is "string"; any other type is invalid.
73 Per-Probe Event Filtering
74 -------------------------
75 Per-probe event filtering feature allows you to set different filter on each
76 probe and gives you what arguments will be shown in trace buffer. If an event
77 name is specified right after 'p:' or 'r:' in kprobe_events, it adds an event
78 under tracing/events/kprobes/<EVENT>, at the directory you can see 'id',
79 'enabled', 'format' and 'filter'.
82 You can enable/disable the probe by writing 1 or 0 on it.
85 This shows the format of this probe event.
88 You can write filtering rules of this event.
91 This shows the id of this probe event.
96 You can check the total number of probe hits and probe miss-hits via
97 /sys/kernel/debug/tracing/kprobe_profile.
98 The first column is event name, the second is the number of probe hits,
99 the third is the number of probe miss-hits.
104 To add a probe as a new event, write a new definition to kprobe_events
107 echo 'p:myprobe do_sys_open dfd=%ax filename=%dx flags=%cx mode=+4($stack)' > /sys/kernel/debug/tracing/kprobe_events
109 This sets a kprobe on the top of do_sys_open() function with recording
110 1st to 4th arguments as "myprobe" event. Note, which register/stack entry is
111 assigned to each function argument depends on arch-specific ABI. If you unsure
112 the ABI, please try to use probe subcommand of perf-tools (you can find it
114 As this example shows, users can choose more familiar names for each arguments.
116 echo 'r:myretprobe do_sys_open $retval' >> /sys/kernel/debug/tracing/kprobe_events
118 This sets a kretprobe on the return point of do_sys_open() function with
119 recording return value as "myretprobe" event.
120 You can see the format of these events via
121 /sys/kernel/debug/tracing/events/kprobes/<EVENT>/format.
123 cat /sys/kernel/debug/tracing/events/kprobes/myprobe/format
127 field:unsigned short common_type; offset:0; size:2; signed:0;
128 field:unsigned char common_flags; offset:2; size:1; signed:0;
129 field:unsigned char common_preempt_count; offset:3; size:1;signed:0;
130 field:int common_pid; offset:4; size:4; signed:1;
132 field:unsigned long __probe_ip; offset:12; size:4; signed:0;
133 field:int __probe_nargs; offset:16; size:4; signed:1;
134 field:unsigned long dfd; offset:20; size:4; signed:0;
135 field:unsigned long filename; offset:24; size:4; signed:0;
136 field:unsigned long flags; offset:28; size:4; signed:0;
137 field:unsigned long mode; offset:32; size:4; signed:0;
140 print fmt: "(%lx) dfd=%lx filename=%lx flags=%lx mode=%lx", REC->__probe_ip,
141 REC->dfd, REC->filename, REC->flags, REC->mode
143 You can see that the event has 4 arguments as in the expressions you specified.
145 echo > /sys/kernel/debug/tracing/kprobe_events
147 This clears all probe points.
151 echo -:myprobe >> kprobe_events
153 This clears probe points selectively.
155 Right after definition, each event is disabled by default. For tracing these
156 events, you need to enable it.
158 echo 1 > /sys/kernel/debug/tracing/events/kprobes/myprobe/enable
159 echo 1 > /sys/kernel/debug/tracing/events/kprobes/myretprobe/enable
161 And you can see the traced information via /sys/kernel/debug/tracing/trace.
163 cat /sys/kernel/debug/tracing/trace
166 # TASK-PID CPU# TIMESTAMP FUNCTION
168 <...>-1447 [001] 1038282.286875: myprobe: (do_sys_open+0x0/0xd6) dfd=3 filename=7fffd1ec4440 flags=8000 mode=0
169 <...>-1447 [001] 1038282.286878: myretprobe: (sys_openat+0xc/0xe <- do_sys_open) $retval=fffffffffffffffe
170 <...>-1447 [001] 1038282.286885: myprobe: (do_sys_open+0x0/0xd6) dfd=ffffff9c filename=40413c flags=8000 mode=1b6
171 <...>-1447 [001] 1038282.286915: myretprobe: (sys_open+0x1b/0x1d <- do_sys_open) $retval=3
172 <...>-1447 [001] 1038282.286969: myprobe: (do_sys_open+0x0/0xd6) dfd=ffffff9c filename=4041c6 flags=98800 mode=10
173 <...>-1447 [001] 1038282.286976: myretprobe: (sys_open+0x1b/0x1d <- do_sys_open) $retval=3
176 Each line shows when the kernel hits an event, and <- SYMBOL means kernel
177 returns from SYMBOL(e.g. "sys_open+0x1b/0x1d <- do_sys_open" means kernel
178 returns from do_sys_open to sys_open+0x1b).