1 Kprobe-based Event Tracing 2 ========================== 3 4 Documentation is written by Masami Hiramatsu 5 6 7Overview 8-------- 9These events are similar to tracepoint based events. Instead of Tracepoint, 10this is based on kprobes (kprobe and kretprobe). So it can probe wherever 11kprobes can probe (this means, all functions except those with 12__kprobes/nokprobe_inline annotation and those marked NOKPROBE_SYMBOL). 13Unlike the Tracepoint based event, this can be added and removed 14dynamically, on the fly. 15 16To enable this feature, build your kernel with CONFIG_KPROBE_EVENTS=y. 17 18Similar to the events tracer, this doesn't need to be activated via 19current_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. 22 23 24Synopsis of kprobe_events 25------------------------- 26 p[:[GRP/]EVENT] [MOD:]SYM[+offs]|MEMADDR [FETCHARGS] : Set a probe 27 r[MAXACTIVE][:[GRP/]EVENT] [MOD:]SYM[+0] [FETCHARGS] : Set a return probe 28 -:[GRP/]EVENT : Clear a probe 29 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. 36 MAXACTIVE : Maximum number of instances of the specified function that 37 can be probed simultaneously, or 0 for the default value 38 as defined in Documentation/kprobes.txt section 1.3.1. 39 40 FETCHARGS : Arguments. Each probe can have up to 128 args. 41 %REG : Fetch register REG 42 @ADDR : Fetch memory at ADDR (ADDR should be in kernel) 43 @SYM[+|-offs] : Fetch memory at SYM +|- offs (SYM should be a data symbol) 44 $stackN : Fetch Nth entry of stack (N >= 0) 45 $stack : Fetch stack address. 46 $retval : Fetch return value.(*) 47 $comm : Fetch current task comm. 48 +|-offs(FETCHARG) : Fetch memory at FETCHARG +|- offs address.(**) 49 NAME=FETCHARG : Set NAME as the argument name of FETCHARG. 50 FETCHARG:TYPE : Set TYPE as the type of FETCHARG. Currently, basic types 51 (u8/u16/u32/u64/s8/s16/s32/s64), hexadecimal types 52 (x8/x16/x32/x64), "string" and bitfield are supported. 53 54 (*) only for return probe. 55 (**) this is useful for fetching a field of data structures. 56 57Types 58----- 59Several types are supported for fetch-args. Kprobe tracer will access memory 60by given type. Prefix 's' and 'u' means those types are signed and unsigned 61respectively. 'x' prefix implies it is unsigned. Traced arguments are shown 62in decimal ('s' and 'u') or hexadecimal ('x'). Without type casting, 'x32' 63or 'x64' is used depends on the architecture (e.g. x86-32 uses x32, and 64x86-64 uses x64). 65String type is a special type, which fetches a "null-terminated" string from 66kernel space. This means it will fail and store NULL if the string container 67has been paged out. 68Bitfield is another special type, which takes 3 parameters, bit-width, bit- 69offset, and container-size (usually 32). The syntax is; 70 71 b<bit-width>@<bit-offset>/<container-size> 72 73For $comm, the default type is "string"; any other type is invalid. 74 75 76Per-Probe Event Filtering 77------------------------- 78 Per-probe event filtering feature allows you to set different filter on each 79probe and gives you what arguments will be shown in trace buffer. If an event 80name is specified right after 'p:' or 'r:' in kprobe_events, it adds an event 81under tracing/events/kprobes/<EVENT>, at the directory you can see 'id', 82'enabled', 'format' and 'filter'. 83 84enabled: 85 You can enable/disable the probe by writing 1 or 0 on it. 86 87format: 88 This shows the format of this probe event. 89 90filter: 91 You can write filtering rules of this event. 92 93id: 94 This shows the id of this probe event. 95 96 97Event Profiling 98--------------- 99 You can check the total number of probe hits and probe miss-hits via 100/sys/kernel/debug/tracing/kprobe_profile. 101 The first column is event name, the second is the number of probe hits, 102the third is the number of probe miss-hits. 103 104 105Usage examples 106-------------- 107To add a probe as a new event, write a new definition to kprobe_events 108as below. 109 110 echo 'p:myprobe do_sys_open dfd=%ax filename=%dx flags=%cx mode=+4($stack)' > /sys/kernel/debug/tracing/kprobe_events 111 112 This sets a kprobe on the top of do_sys_open() function with recording 1131st to 4th arguments as "myprobe" event. Note, which register/stack entry is 114assigned to each function argument depends on arch-specific ABI. If you unsure 115the ABI, please try to use probe subcommand of perf-tools (you can find it 116under tools/perf/). 117As this example shows, users can choose more familiar names for each arguments. 118 119 echo 'r:myretprobe do_sys_open $retval' >> /sys/kernel/debug/tracing/kprobe_events 120 121 This sets a kretprobe on the return point of do_sys_open() function with 122recording return value as "myretprobe" event. 123 You can see the format of these events via 124/sys/kernel/debug/tracing/events/kprobes/<EVENT>/format. 125 126 cat /sys/kernel/debug/tracing/events/kprobes/myprobe/format 127name: myprobe 128ID: 780 129format: 130 field:unsigned short common_type; offset:0; size:2; signed:0; 131 field:unsigned char common_flags; offset:2; size:1; signed:0; 132 field:unsigned char common_preempt_count; offset:3; size:1;signed:0; 133 field:int common_pid; offset:4; size:4; signed:1; 134 135 field:unsigned long __probe_ip; offset:12; size:4; signed:0; 136 field:int __probe_nargs; offset:16; size:4; signed:1; 137 field:unsigned long dfd; offset:20; size:4; signed:0; 138 field:unsigned long filename; offset:24; size:4; signed:0; 139 field:unsigned long flags; offset:28; size:4; signed:0; 140 field:unsigned long mode; offset:32; size:4; signed:0; 141 142 143print fmt: "(%lx) dfd=%lx filename=%lx flags=%lx mode=%lx", REC->__probe_ip, 144REC->dfd, REC->filename, REC->flags, REC->mode 145 146 You can see that the event has 4 arguments as in the expressions you specified. 147 148 echo > /sys/kernel/debug/tracing/kprobe_events 149 150 This clears all probe points. 151 152 Or, 153 154 echo -:myprobe >> kprobe_events 155 156 This clears probe points selectively. 157 158 Right after definition, each event is disabled by default. For tracing these 159events, you need to enable it. 160 161 echo 1 > /sys/kernel/debug/tracing/events/kprobes/myprobe/enable 162 echo 1 > /sys/kernel/debug/tracing/events/kprobes/myretprobe/enable 163 164 And you can see the traced information via /sys/kernel/debug/tracing/trace. 165 166 cat /sys/kernel/debug/tracing/trace 167# tracer: nop 168# 169# TASK-PID CPU# TIMESTAMP FUNCTION 170# | | | | | 171 <...>-1447 [001] 1038282.286875: myprobe: (do_sys_open+0x0/0xd6) dfd=3 filename=7fffd1ec4440 flags=8000 mode=0 172 <...>-1447 [001] 1038282.286878: myretprobe: (sys_openat+0xc/0xe <- do_sys_open) $retval=fffffffffffffffe 173 <...>-1447 [001] 1038282.286885: myprobe: (do_sys_open+0x0/0xd6) dfd=ffffff9c filename=40413c flags=8000 mode=1b6 174 <...>-1447 [001] 1038282.286915: myretprobe: (sys_open+0x1b/0x1d <- do_sys_open) $retval=3 175 <...>-1447 [001] 1038282.286969: myprobe: (do_sys_open+0x0/0xd6) dfd=ffffff9c filename=4041c6 flags=98800 mode=10 176 <...>-1447 [001] 1038282.286976: myretprobe: (sys_open+0x1b/0x1d <- do_sys_open) $retval=3 177 178 179 Each line shows when the kernel hits an event, and <- SYMBOL means kernel 180returns from SYMBOL(e.g. "sys_open+0x1b/0x1d <- do_sys_open" means kernel 181returns from do_sys_open to sys_open+0x1b). 182 183