1=========================================
2Uprobe-tracer: Uprobe-based Event Tracing
3=========================================
4
5:Author: Srikar Dronamraju
6
7
8Overview
9--------
10Uprobe based trace events are similar to kprobe based trace events.
11To enable this feature, build your kernel with CONFIG_UPROBE_EVENTS=y.
12
13Similar to the kprobe-event tracer, this doesn't need to be activated via
14current_tracer. Instead of that, add probe points via
15/sys/kernel/debug/tracing/uprobe_events, and enable it via
16/sys/kernel/debug/tracing/events/uprobes/<EVENT>/enable.
17
18However unlike kprobe-event tracer, the uprobe event interface expects the
19user to calculate the offset of the probepoint in the object.
20
21You can also use /sys/kernel/debug/tracing/dynamic_events instead of
22uprobe_events. That interface will provide unified access to other
23dynamic events too.
24
25Synopsis of uprobe_tracer
26-------------------------
27::
28
29  p[:[GRP/][EVENT]] PATH:OFFSET [FETCHARGS] : Set a uprobe
30  r[:[GRP/][EVENT]] PATH:OFFSET [FETCHARGS] : Set a return uprobe (uretprobe)
31  p[:[GRP/][EVENT]] PATH:OFFSET%return [FETCHARGS] : Set a return uprobe (uretprobe)
32  -:[GRP/][EVENT]                           : Clear uprobe or uretprobe event
33
34  GRP           : Group name. If omitted, "uprobes" is the default value.
35  EVENT         : Event name. If omitted, the event name is generated based
36                  on PATH+OFFSET.
37  PATH          : Path to an executable or a library.
38  OFFSET        : Offset where the probe is inserted.
39  OFFSET%return : Offset where the return probe is inserted.
40
41  FETCHARGS     : Arguments. Each probe can have up to 128 args.
42   %REG         : Fetch register REG
43   @ADDR	: Fetch memory at ADDR (ADDR should be in userspace)
44   @+OFFSET	: Fetch memory at OFFSET (OFFSET from same file as PATH)
45   $stackN	: Fetch Nth entry of stack (N >= 0)
46   $stack	: Fetch stack address.
47   $retval	: Fetch return value.(\*1)
48   $comm	: Fetch current task comm.
49   +|-[u]OFFS(FETCHARG) : Fetch memory at FETCHARG +|- OFFS address.(\*2)(\*3)
50   \IMM		: Store an immediate value to the argument.
51   NAME=FETCHARG     : Set NAME as the argument name of FETCHARG.
52   FETCHARG:TYPE     : Set TYPE as the type of FETCHARG. Currently, basic types
53		       (u8/u16/u32/u64/s8/s16/s32/s64), hexadecimal types
54		       (x8/x16/x32/x64), "string" and bitfield are supported.
55
56  (\*1) only for return probe.
57  (\*2) this is useful for fetching a field of data structures.
58  (\*3) Unlike kprobe event, "u" prefix will just be ignored, becuse uprobe
59        events can access only user-space memory.
60
61Types
62-----
63Several types are supported for fetch-args. Uprobe tracer will access memory
64by given type. Prefix 's' and 'u' means those types are signed and unsigned
65respectively. 'x' prefix implies it is unsigned. Traced arguments are shown
66in decimal ('s' and 'u') or hexadecimal ('x'). Without type casting, 'x32'
67or 'x64' is used depends on the architecture (e.g. x86-32 uses x32, and
68x86-64 uses x64).
69String type is a special type, which fetches a "null-terminated" string from
70user space.
71Bitfield is another special type, which takes 3 parameters, bit-width, bit-
72offset, and container-size (usually 32). The syntax is::
73
74 b<bit-width>@<bit-offset>/<container-size>
75
76For $comm, the default type is "string"; any other type is invalid.
77
78
79Event Profiling
80---------------
81You can check the total number of probe hits per event via
82/sys/kernel/debug/tracing/uprobe_profile. The first column is the filename,
83the second is the event name, the third is the number of probe hits.
84
85Usage examples
86--------------
87 * Add a probe as a new uprobe event, write a new definition to uprobe_events
88   as below (sets a uprobe at an offset of 0x4245c0 in the executable /bin/bash)::
89
90    echo 'p /bin/bash:0x4245c0' > /sys/kernel/debug/tracing/uprobe_events
91
92 * Add a probe as a new uretprobe event::
93
94    echo 'r /bin/bash:0x4245c0' > /sys/kernel/debug/tracing/uprobe_events
95
96 * Unset registered event::
97
98    echo '-:p_bash_0x4245c0' >> /sys/kernel/debug/tracing/uprobe_events
99
100 * Print out the events that are registered::
101
102    cat /sys/kernel/debug/tracing/uprobe_events
103
104 * Clear all events::
105
106    echo > /sys/kernel/debug/tracing/uprobe_events
107
108Following example shows how to dump the instruction pointer and %ax register
109at the probed text address. Probe zfree function in /bin/zsh::
110
111    # cd /sys/kernel/debug/tracing/
112    # cat /proc/`pgrep zsh`/maps | grep /bin/zsh | grep r-xp
113    00400000-0048a000 r-xp 00000000 08:03 130904 /bin/zsh
114    # objdump -T /bin/zsh | grep -w zfree
115    0000000000446420 g    DF .text  0000000000000012  Base        zfree
116
1170x46420 is the offset of zfree in object /bin/zsh that is loaded at
1180x00400000. Hence the command to uprobe would be::
119
120    # echo 'p:zfree_entry /bin/zsh:0x46420 %ip %ax' > uprobe_events
121
122And the same for the uretprobe would be::
123
124    # echo 'r:zfree_exit /bin/zsh:0x46420 %ip %ax' >> uprobe_events
125
126.. note:: User has to explicitly calculate the offset of the probe-point
127	in the object.
128
129We can see the events that are registered by looking at the uprobe_events file.
130::
131
132    # cat uprobe_events
133    p:uprobes/zfree_entry /bin/zsh:0x00046420 arg1=%ip arg2=%ax
134    r:uprobes/zfree_exit /bin/zsh:0x00046420 arg1=%ip arg2=%ax
135
136Format of events can be seen by viewing the file events/uprobes/zfree_entry/format.
137::
138
139    # cat events/uprobes/zfree_entry/format
140    name: zfree_entry
141    ID: 922
142    format:
143         field:unsigned short common_type;         offset:0;  size:2; signed:0;
144         field:unsigned char common_flags;         offset:2;  size:1; signed:0;
145         field:unsigned char common_preempt_count; offset:3;  size:1; signed:0;
146         field:int common_pid;                     offset:4;  size:4; signed:1;
147         field:int common_padding;                 offset:8;  size:4; signed:1;
148
149         field:unsigned long __probe_ip;           offset:12; size:4; signed:0;
150         field:u32 arg1;                           offset:16; size:4; signed:0;
151         field:u32 arg2;                           offset:20; size:4; signed:0;
152
153    print fmt: "(%lx) arg1=%lx arg2=%lx", REC->__probe_ip, REC->arg1, REC->arg2
154
155Right after definition, each event is disabled by default. For tracing these
156events, you need to enable it by::
157
158    # echo 1 > events/uprobes/enable
159
160Lets start tracing, sleep for some time and stop tracing.
161::
162
163    # echo 1 > tracing_on
164    # sleep 20
165    # echo 0 > tracing_on
166
167Also, you can disable the event by::
168
169    # echo 0 > events/uprobes/enable
170
171And you can see the traced information via /sys/kernel/debug/tracing/trace.
172::
173
174    # cat trace
175    # tracer: nop
176    #
177    #           TASK-PID    CPU#    TIMESTAMP  FUNCTION
178    #              | |       |          |         |
179                 zsh-24842 [006] 258544.995456: zfree_entry: (0x446420) arg1=446420 arg2=79
180                 zsh-24842 [007] 258545.000270: zfree_exit:  (0x446540 <- 0x446420) arg1=446540 arg2=0
181                 zsh-24842 [002] 258545.043929: zfree_entry: (0x446420) arg1=446420 arg2=79
182                 zsh-24842 [004] 258547.046129: zfree_exit:  (0x446540 <- 0x446420) arg1=446540 arg2=0
183
184Output shows us uprobe was triggered for a pid 24842 with ip being 0x446420
185and contents of ax register being 79. And uretprobe was triggered with ip at
1860x446540 with counterpart function entry at 0x446420.
187