]>
Commit | Line | Data |
---|---|---|
00b27da3 CD |
1 | ========================================= |
2 | Uprobe-tracer: Uprobe-based Event Tracing | |
3 | ========================================= | |
decc6bfb | 4 | |
00b27da3 | 5 | :Author: Srikar Dronamraju |
decc6bfb | 6 | |
f3f096cf SD |
7 | |
8 | Overview | |
9 | -------- | |
10 | Uprobe based trace events are similar to kprobe based trace events. | |
6b0b7551 | 11 | To enable this feature, build your kernel with CONFIG_UPROBE_EVENTS=y. |
f3f096cf SD |
12 | |
13 | Similar to the kprobe-event tracer, this doesn't need to be activated via | |
14 | current_tracer. Instead of that, add probe points via | |
15 | /sys/kernel/debug/tracing/uprobe_events, and enable it via | |
e50891d6 | 16 | /sys/kernel/debug/tracing/events/uprobes/<EVENT>/enable. |
f3f096cf SD |
17 | |
18 | However unlike kprobe-event tracer, the uprobe event interface expects the | |
decc6bfb | 19 | user to calculate the offset of the probepoint in the object. |
f3f096cf | 20 | |
0597c49c MH |
21 | You can also use /sys/kernel/debug/tracing/dynamic_events instead of |
22 | uprobe_events. That interface will provide unified access to other | |
23 | dynamic events too. | |
24 | ||
f3f096cf SD |
25 | Synopsis of uprobe_tracer |
26 | ------------------------- | |
00b27da3 CD |
27 | :: |
28 | ||
306cfe20 NK |
29 | p[:[GRP/]EVENT] PATH:OFFSET [FETCHARGS] : Set a uprobe |
30 | r[:[GRP/]EVENT] PATH:OFFSET [FETCHARGS] : Set a return uprobe (uretprobe) | |
31 | -:[GRP/]EVENT : Clear uprobe or uretprobe event | |
f3f096cf | 32 | |
decc6bfb AA |
33 | GRP : Group name. If omitted, "uprobes" is the default value. |
34 | EVENT : Event name. If omitted, the event name is generated based | |
306cfe20 | 35 | on PATH+OFFSET. |
decc6bfb | 36 | PATH : Path to an executable or a library. |
306cfe20 | 37 | OFFSET : Offset where the probe is inserted. |
f3f096cf | 38 | |
decc6bfb AA |
39 | FETCHARGS : Arguments. Each probe can have up to 128 args. |
40 | %REG : Fetch register REG | |
b079d374 | 41 | @ADDR : Fetch memory at ADDR (ADDR should be in userspace) |
b7e0bf34 | 42 | @+OFFSET : Fetch memory at OFFSET (OFFSET from same file as PATH) |
b079d374 NK |
43 | $stackN : Fetch Nth entry of stack (N >= 0) |
44 | $stack : Fetch stack address. | |
e65f7ae7 | 45 | $retval : Fetch return value.(\*1) |
35abb67d | 46 | $comm : Fetch current task comm. |
e65f7ae7 | 47 | +|-[u]OFFS(FETCHARG) : Fetch memory at FETCHARG +|- OFFS address.(\*2)(\*3) |
6218bf9f | 48 | \IMM : Store an immediate value to the argument. |
b079d374 NK |
49 | NAME=FETCHARG : Set NAME as the argument name of FETCHARG. |
50 | FETCHARG:TYPE : Set TYPE as the type of FETCHARG. Currently, basic types | |
17ce3dc7 MH |
51 | (u8/u16/u32/u64/s8/s16/s32/s64), hexadecimal types |
52 | (x8/x16/x32/x64), "string" and bitfield are supported. | |
b079d374 | 53 | |
e65f7ae7 MH |
54 | (\*1) only for return probe. |
55 | (\*2) this is useful for fetching a field of data structures. | |
56 | (\*3) Unlike kprobe event, "u" prefix will just be ignored, becuse uprobe | |
57 | events can access only user-space memory. | |
b079d374 NK |
58 | |
59 | Types | |
60 | ----- | |
61 | Several types are supported for fetch-args. Uprobe tracer will access memory | |
62 | by given type. Prefix 's' and 'u' means those types are signed and unsigned | |
bdca79c2 MH |
63 | respectively. 'x' prefix implies it is unsigned. Traced arguments are shown |
64 | in decimal ('s' and 'u') or hexadecimal ('x'). Without type casting, 'x32' | |
65 | or 'x64' is used depends on the architecture (e.g. x86-32 uses x32, and | |
66 | x86-64 uses x64). | |
b079d374 NK |
67 | String type is a special type, which fetches a "null-terminated" string from |
68 | user space. | |
69 | Bitfield is another special type, which takes 3 parameters, bit-width, bit- | |
00b27da3 | 70 | offset, and container-size (usually 32). The syntax is:: |
b079d374 NK |
71 | |
72 | b<bit-width>@<bit-offset>/<container-size> | |
73 | ||
35abb67d OS |
74 | For $comm, the default type is "string"; any other type is invalid. |
75 | ||
f3f096cf SD |
76 | |
77 | Event Profiling | |
78 | --------------- | |
b689269b SD |
79 | You can check the total number of probe hits per event via |
80 | /sys/kernel/debug/tracing/uprobe_profile. The first column is the filename, | |
81 | the second is the event name, the third is the number of probe hits. | |
f3f096cf SD |
82 | |
83 | Usage examples | |
84 | -------------- | |
decc6bfb | 85 | * Add a probe as a new uprobe event, write a new definition to uprobe_events |
00b27da3 | 86 | as below (sets a uprobe at an offset of 0x4245c0 in the executable /bin/bash):: |
decc6bfb | 87 | |
7058763b | 88 | echo 'p /bin/bash:0x4245c0' > /sys/kernel/debug/tracing/uprobe_events |
decc6bfb | 89 | |
00b27da3 | 90 | * Add a probe as a new uretprobe event:: |
decc6bfb | 91 | |
7058763b | 92 | echo 'r /bin/bash:0x4245c0' > /sys/kernel/debug/tracing/uprobe_events |
decc6bfb | 93 | |
00b27da3 | 94 | * Unset registered event:: |
f3f096cf | 95 | |
7058763b | 96 | echo '-:p_bash_0x4245c0' >> /sys/kernel/debug/tracing/uprobe_events |
f3f096cf | 97 | |
00b27da3 | 98 | * Print out the events that are registered:: |
f3f096cf | 99 | |
decc6bfb | 100 | cat /sys/kernel/debug/tracing/uprobe_events |
f3f096cf | 101 | |
00b27da3 | 102 | * Clear all events:: |
f3f096cf | 103 | |
decc6bfb AA |
104 | echo > /sys/kernel/debug/tracing/uprobe_events |
105 | ||
106 | Following example shows how to dump the instruction pointer and %ax register | |
00b27da3 | 107 | at the probed text address. Probe zfree function in /bin/zsh:: |
f3f096cf SD |
108 | |
109 | # cd /sys/kernel/debug/tracing/ | |
decc6bfb | 110 | # cat /proc/`pgrep zsh`/maps | grep /bin/zsh | grep r-xp |
f3f096cf SD |
111 | 00400000-0048a000 r-xp 00000000 08:03 130904 /bin/zsh |
112 | # objdump -T /bin/zsh | grep -w zfree | |
113 | 0000000000446420 g DF .text 0000000000000012 Base zfree | |
114 | ||
00b27da3 CD |
115 | 0x46420 is the offset of zfree in object /bin/zsh that is loaded at |
116 | 0x00400000. Hence the command to uprobe would be:: | |
decc6bfb AA |
117 | |
118 | # echo 'p:zfree_entry /bin/zsh:0x46420 %ip %ax' > uprobe_events | |
119 | ||
00b27da3 | 120 | And the same for the uretprobe would be:: |
f3f096cf | 121 | |
decc6bfb | 122 | # echo 'r:zfree_exit /bin/zsh:0x46420 %ip %ax' >> uprobe_events |
f3f096cf | 123 | |
00b27da3 CD |
124 | .. note:: User has to explicitly calculate the offset of the probe-point |
125 | in the object. | |
126 | ||
127 | We can see the events that are registered by looking at the uprobe_events file. | |
128 | :: | |
f3f096cf SD |
129 | |
130 | # cat uprobe_events | |
decc6bfb AA |
131 | p:uprobes/zfree_entry /bin/zsh:0x00046420 arg1=%ip arg2=%ax |
132 | r:uprobes/zfree_exit /bin/zsh:0x00046420 arg1=%ip arg2=%ax | |
ec83db0f | 133 | |
00b27da3 CD |
134 | Format of events can be seen by viewing the file events/uprobes/zfree_entry/format. |
135 | :: | |
ec83db0f | 136 | |
decc6bfb AA |
137 | # cat events/uprobes/zfree_entry/format |
138 | name: zfree_entry | |
ec83db0f SD |
139 | ID: 922 |
140 | format: | |
decc6bfb AA |
141 | field:unsigned short common_type; offset:0; size:2; signed:0; |
142 | field:unsigned char common_flags; offset:2; size:1; signed:0; | |
143 | field:unsigned char common_preempt_count; offset:3; size:1; signed:0; | |
144 | field:int common_pid; offset:4; size:4; signed:1; | |
145 | field:int common_padding; offset:8; size:4; signed:1; | |
ec83db0f | 146 | |
decc6bfb AA |
147 | field:unsigned long __probe_ip; offset:12; size:4; signed:0; |
148 | field:u32 arg1; offset:16; size:4; signed:0; | |
149 | field:u32 arg2; offset:20; size:4; signed:0; | |
ec83db0f SD |
150 | |
151 | print fmt: "(%lx) arg1=%lx arg2=%lx", REC->__probe_ip, REC->arg1, REC->arg2 | |
f3f096cf SD |
152 | |
153 | Right after definition, each event is disabled by default. For tracing these | |
00b27da3 | 154 | events, you need to enable it by:: |
f3f096cf SD |
155 | |
156 | # echo 1 > events/uprobes/enable | |
157 | ||
78a89463 | 158 | Lets start tracing, sleep for some time and stop tracing. |
00b27da3 | 159 | :: |
decc6bfb | 160 | |
78a89463 | 161 | # echo 1 > tracing_on |
f3f096cf | 162 | # sleep 20 |
78a89463 LC |
163 | # echo 0 > tracing_on |
164 | ||
165 | Also, you can disable the event by:: | |
166 | ||
f3f096cf SD |
167 | # echo 0 > events/uprobes/enable |
168 | ||
169 | And you can see the traced information via /sys/kernel/debug/tracing/trace. | |
00b27da3 | 170 | :: |
f3f096cf SD |
171 | |
172 | # cat trace | |
173 | # tracer: nop | |
174 | # | |
175 | # TASK-PID CPU# TIMESTAMP FUNCTION | |
176 | # | | | | | | |
decc6bfb AA |
177 | zsh-24842 [006] 258544.995456: zfree_entry: (0x446420) arg1=446420 arg2=79 |
178 | zsh-24842 [007] 258545.000270: zfree_exit: (0x446540 <- 0x446420) arg1=446540 arg2=0 | |
179 | zsh-24842 [002] 258545.043929: zfree_entry: (0x446420) arg1=446420 arg2=79 | |
180 | zsh-24842 [004] 258547.046129: zfree_exit: (0x446540 <- 0x446420) arg1=446540 arg2=0 | |
181 | ||
182 | Output shows us uprobe was triggered for a pid 24842 with ip being 0x446420 | |
183 | and contents of ax register being 79. And uretprobe was triggered with ip at | |
184 | 0x446540 with counterpart function entry at 0x446420. |