Documentation/trace/uprobetracer.txt

   1             Uprobe-tracer: Uprobe-based Event Tracing
   2             =========================================
   3
   4            Documentation written by Srikar Dronamraju
   5
   6
   7 Overview
   8 --------
   9 Uprobe based trace events are similar to kprobe based trace events.
  10 To enable this feature, build your kernel with CONFIG_UPROBE_EVENT=y.
  11
  12 Similar to the kprobe-event tracer, this doesn't need to be activated via
  13 current_tracer. Instead of that, add probe points via
  14 /sys/kernel/debug/tracing/uprobe_events, and enable it via
  15 /sys/kernel/debug/tracing/events/uprobes/<EVENT>/enabled.
  16
  17 However unlike kprobe-event tracer, the uprobe event interface expects the
  18 user to calculate the offset of the probepoint in the object.
  19
  20 Synopsis of uprobe_tracer
  21 -------------------------
  22   p[:[GRP/]EVENT] PATH:SYMBOL[+offs] [FETCHARGS] : Set a uprobe
  23   r[:[GRP/]EVENT] PATH:SYMBOL[+offs] [FETCHARGS] : Set a return uprobe (uretprobe)
  24   -:[GRP/]EVENT                                  : Clear uprobe or uretprobe event
  25
  26   GRP           : Group name. If omitted, "uprobes" is the default value.
  27   EVENT         : Event name. If omitted, the event name is generated based
  28                   on SYMBOL+offs.
  29   PATH          : Path to an executable or a library.
  30   SYMBOL[+offs] : Symbol+offset where the probe is inserted.
  31
  32   FETCHARGS     : Arguments. Each probe can have up to 128 args.
  33    %REG         : Fetch register REG
  34
  35 Event Profiling
  36 ---------------
  37 You can check the total number of probe hits and probe miss-hits via
  38 /sys/kernel/debug/tracing/uprobe_profile.
  39 The first column is event name, the second is the number of probe hits,
  40 the third is the number of probe miss-hits.
  41
  42 Usage examples
  43 --------------
  44  * Add a probe as a new uprobe event, write a new definition to uprobe_events
  45 as below: (sets a uprobe at an offset of 0x4245c0 in the executable /bin/bash)
  46
  47     echo 'p: /bin/bash:0x4245c0' > /sys/kernel/debug/tracing/uprobe_events
  48
  49  * Add a probe as a new uretprobe event:
  50
  51     echo 'r: /bin/bash:0x4245c0' > /sys/kernel/debug/tracing/uprobe_events
  52
  53  * Unset registered event:
  54
  55     echo '-:bash_0x4245c0' >> /sys/kernel/debug/tracing/uprobe_events
  56
  57  * Print out the events that are registered:
  58
  59     cat /sys/kernel/debug/tracing/uprobe_events
  60
  61  * Clear all events:
  62
  63     echo > /sys/kernel/debug/tracing/uprobe_events
  64
  65 Following example shows how to dump the instruction pointer and %ax register
  66 at the probed text address. Probe zfree function in /bin/zsh:
  67
  68     # cd /sys/kernel/debug/tracing/
  69     # cat /proc/`pgrep zsh`/maps | grep /bin/zsh | grep r-xp
  70     00400000-0048a000 r-xp 00000000 08:03 130904 /bin/zsh
  71     # objdump -T /bin/zsh | grep -w zfree
  72     0000000000446420 g    DF .text  0000000000000012  Base        zfree
  73
  74   0x46420 is the offset of zfree in object /bin/zsh that is loaded at
  75   0x00400000. Hence the command to uprobe would be:
  76
  77     # echo 'p:zfree_entry /bin/zsh:0x46420 %ip %ax' > uprobe_events
  78
  79   And the same for the uretprobe would be:
  80
  81     # echo 'r:zfree_exit /bin/zsh:0x46420 %ip %ax' >> uprobe_events
  82
  83 Please note: User has to explicitly calculate the offset of the probe-point
  84 in the object. We can see the events that are registered by looking at the
  85 uprobe_events file.
  86
  87     # cat uprobe_events
  88     p:uprobes/zfree_entry /bin/zsh:0x00046420 arg1=%ip arg2=%ax
  89     r:uprobes/zfree_exit /bin/zsh:0x00046420 arg1=%ip arg2=%ax
  90
  91 Format of events can be seen by viewing the file events/uprobes/zfree_entry/format
  92
  93     # cat events/uprobes/zfree_entry/format
  94     name: zfree_entry
  95     ID: 922
  96     format:
  97          field:unsigned short common_type;         offset:0;  size:2; signed:0;
  98          field:unsigned char common_flags;         offset:2;  size:1; signed:0;
  99          field:unsigned char common_preempt_count; offset:3;  size:1; signed:0;
 100          field:int common_pid;                     offset:4;  size:4; signed:1;
 101          field:int common_padding;                 offset:8;  size:4; signed:1;
 102
 103          field:unsigned long __probe_ip;           offset:12; size:4; signed:0;
 104          field:u32 arg1;                           offset:16; size:4; signed:0;
 105          field:u32 arg2;                           offset:20; size:4; signed:0;
 106
 107     print fmt: "(%lx) arg1=%lx arg2=%lx", REC->__probe_ip, REC->arg1, REC->arg2
 108
 109 Right after definition, each event is disabled by default. For tracing these
 110 events, you need to enable it by:
 111
 112     # echo 1 > events/uprobes/enable
 113
 114 Lets disable the event after sleeping for some time.
 115
 116     # sleep 20
 117     # echo 0 > events/uprobes/enable
 118
 119 And you can see the traced information via /sys/kernel/debug/tracing/trace.
 120
 121     # cat trace
 122     # tracer: nop
 123     #
 124     #           TASK-PID    CPU#    TIMESTAMP  FUNCTION
 125     #              | |       |          |         |
 126                  zsh-24842 [006] 258544.995456: zfree_entry: (0x446420) arg1=446420 arg2=79
 127                  zsh-24842 [007] 258545.000270: zfree_exit:  (0x446540 <- 0x446420) arg1=446540 arg2=0
 128                  zsh-24842 [002] 258545.043929: zfree_entry: (0x446420) arg1=446420 arg2=79
 129                  zsh-24842 [004] 258547.046129: zfree_exit:  (0x446540 <- 0x446420) arg1=446540 arg2=0
 130
 131 Output shows us uprobe was triggered for a pid 24842 with ip being 0x446420
 132 and contents of ax register being 79. And uretprobe was triggered with ip at
 133 0x446540 with counterpart function entry at 0x446420.