1========================== 2Kprobe-based Event Tracing 3========================== 4 5:Author: Masami Hiramatsu 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>/enable. 22 23You can also use /sys/kernel/debug/tracing/dynamic_events instead of 24kprobe_events. That interface will provide unified access to other 25dynamic events too. 26 27Synopsis of kprobe_events 28------------------------- 29:: 30 31 p[:[GRP/]EVENT] [MOD:]SYM[+offs]|MEMADDR [FETCHARGS] : Set a probe 32 r[MAXACTIVE][:[GRP/]EVENT] [MOD:]SYM[+0] [FETCHARGS] : Set a return probe 33 -:[GRP/]EVENT : Clear a probe 34 35 GRP : Group name. If omitted, use "kprobes" for it. 36 EVENT : Event name. If omitted, the event name is generated 37 based on SYM+offs or MEMADDR. 38 MOD : Module name which has given SYM. 39 SYM[+offs] : Symbol+offset where the probe is inserted. 40 MEMADDR : Address where the probe is inserted. 41 MAXACTIVE : Maximum number of instances of the specified function that 42 can be probed simultaneously, or 0 for the default value 43 as defined in Documentation/staging/kprobes.rst section 1.3.1. 44 45 FETCHARGS : Arguments. Each probe can have up to 128 args. 46 %REG : Fetch register REG 47 @ADDR : Fetch memory at ADDR (ADDR should be in kernel) 48 @SYM[+|-offs] : Fetch memory at SYM +|- offs (SYM should be a data symbol) 49 $stackN : Fetch Nth entry of stack (N >= 0) 50 $stack : Fetch stack address. 51 $argN : Fetch the Nth function argument. (N >= 1) (\*1) 52 $retval : Fetch return value.(\*2) 53 $comm : Fetch current task comm. 54 +|-[u]OFFS(FETCHARG) : Fetch memory at FETCHARG +|- OFFS address.(\*3)(\*4) 55 \IMM : Store an immediate value to the argument. 56 NAME=FETCHARG : Set NAME as the argument name of FETCHARG. 57 FETCHARG:TYPE : Set TYPE as the type of FETCHARG. Currently, basic types 58 (u8/u16/u32/u64/s8/s16/s32/s64), hexadecimal types 59 (x8/x16/x32/x64), "string", "ustring" and bitfield 60 are supported. 61 62 (\*1) only for the probe on function entry (offs == 0). 63 (\*2) only for return probe. 64 (\*3) this is useful for fetching a field of data structures. 65 (\*4) "u" means user-space dereference. See :ref:`user_mem_access`. 66 67Types 68----- 69Several types are supported for fetch-args. Kprobe tracer will access memory 70by given type. Prefix 's' and 'u' means those types are signed and unsigned 71respectively. 'x' prefix implies it is unsigned. Traced arguments are shown 72in decimal ('s' and 'u') or hexadecimal ('x'). Without type casting, 'x32' 73or 'x64' is used depends on the architecture (e.g. x86-32 uses x32, and 74x86-64 uses x64). 75These value types can be an array. To record array data, you can add '[N]' 76(where N is a fixed number, less than 64) to the base type. 77E.g. 'x16[4]' means an array of x16 (2bytes hex) with 4 elements. 78Note that the array can be applied to memory type fetchargs, you can not 79apply it to registers/stack-entries etc. (for example, '$stack1:x8[8]' is 80wrong, but '+8($stack):x8[8]' is OK.) 81String type is a special type, which fetches a "null-terminated" string from 82kernel space. This means it will fail and store NULL if the string container 83has been paged out. "ustring" type is an alternative of string for user-space. 84See :ref:`user_mem_access` for more info.. 85The string array type is a bit different from other types. For other base 86types, <base-type>[1] is equal to <base-type> (e.g. +0(%di):x32[1] is same 87as +0(%di):x32.) But string[1] is not equal to string. The string type itself 88represents "char array", but string array type represents "char * array". 89So, for example, +0(%di):string[1] is equal to +0(+0(%di)):string. 90Bitfield is another special type, which takes 3 parameters, bit-width, bit- 91offset, and container-size (usually 32). The syntax is:: 92 93 b<bit-width>@<bit-offset>/<container-size> 94 95Symbol type('symbol') is an alias of u32 or u64 type (depends on BITS_PER_LONG) 96which shows given pointer in "symbol+offset" style. 97For $comm, the default type is "string"; any other type is invalid. 98 99.. _user_mem_access: 100 101User Memory Access 102------------------ 103Kprobe events supports user-space memory access. For that purpose, you can use 104either user-space dereference syntax or 'ustring' type. 105 106The user-space dereference syntax allows you to access a field of a data 107structure in user-space. This is done by adding the "u" prefix to the 108dereference syntax. For example, +u4(%si) means it will read memory from the 109address in the register %si offset by 4, and the memory is expected to be in 110user-space. You can use this for strings too, e.g. +u0(%si):string will read 111a string from the address in the register %si that is expected to be in user- 112space. 'ustring' is a shortcut way of performing the same task. That is, 113+0(%si):ustring is equivalent to +u0(%si):string. 114 115Note that kprobe-event provides the user-memory access syntax but it doesn't 116use it transparently. This means if you use normal dereference or string type 117for user memory, it might fail, and may always fail on some archs. The user 118has to carefully check if the target data is in kernel or user space. 119 120Per-Probe Event Filtering 121------------------------- 122Per-probe event filtering feature allows you to set different filter on each 123probe and gives you what arguments will be shown in trace buffer. If an event 124name is specified right after 'p:' or 'r:' in kprobe_events, it adds an event 125under tracing/events/kprobes/<EVENT>, at the directory you can see 'id', 126'enable', 'format', 'filter' and 'trigger'. 127 128enable: 129 You can enable/disable the probe by writing 1 or 0 on it. 130 131format: 132 This shows the format of this probe event. 133 134filter: 135 You can write filtering rules of this event. 136 137id: 138 This shows the id of this probe event. 139 140trigger: 141 This allows to install trigger commands which are executed when the event is 142 hit (for details, see Documentation/trace/events.rst, section 6). 143 144Event Profiling 145--------------- 146You can check the total number of probe hits and probe miss-hits via 147/sys/kernel/debug/tracing/kprobe_profile. 148The first column is event name, the second is the number of probe hits, 149the third is the number of probe miss-hits. 150 151Kernel Boot Parameter 152--------------------- 153You can add and enable new kprobe events when booting up the kernel by 154"kprobe_event=" parameter. The parameter accepts a semicolon-delimited 155kprobe events, which format is similar to the kprobe_events. 156The difference is that the probe definition parameters are comma-delimited 157instead of space. For example, adding myprobe event on do_sys_open like below 158 159 p:myprobe do_sys_open dfd=%ax filename=%dx flags=%cx mode=+4($stack) 160 161should be below for kernel boot parameter (just replace spaces with comma) 162 163 p:myprobe,do_sys_open,dfd=%ax,filename=%dx,flags=%cx,mode=+4($stack) 164 165 166Usage examples 167-------------- 168To add a probe as a new event, write a new definition to kprobe_events 169as below:: 170 171 echo 'p:myprobe do_sys_open dfd=%ax filename=%dx flags=%cx mode=+4($stack)' > /sys/kernel/debug/tracing/kprobe_events 172 173This sets a kprobe on the top of do_sys_open() function with recording 1741st to 4th arguments as "myprobe" event. Note, which register/stack entry is 175assigned to each function argument depends on arch-specific ABI. If you unsure 176the ABI, please try to use probe subcommand of perf-tools (you can find it 177under tools/perf/). 178As this example shows, users can choose more familiar names for each arguments. 179:: 180 181 echo 'r:myretprobe do_sys_open $retval' >> /sys/kernel/debug/tracing/kprobe_events 182 183This sets a kretprobe on the return point of do_sys_open() function with 184recording return value as "myretprobe" event. 185You can see the format of these events via 186/sys/kernel/debug/tracing/events/kprobes/<EVENT>/format. 187:: 188 189 cat /sys/kernel/debug/tracing/events/kprobes/myprobe/format 190 name: myprobe 191 ID: 780 192 format: 193 field:unsigned short common_type; offset:0; size:2; signed:0; 194 field:unsigned char common_flags; offset:2; size:1; signed:0; 195 field:unsigned char common_preempt_count; offset:3; size:1;signed:0; 196 field:int common_pid; offset:4; size:4; signed:1; 197 198 field:unsigned long __probe_ip; offset:12; size:4; signed:0; 199 field:int __probe_nargs; offset:16; size:4; signed:1; 200 field:unsigned long dfd; offset:20; size:4; signed:0; 201 field:unsigned long filename; offset:24; size:4; signed:0; 202 field:unsigned long flags; offset:28; size:4; signed:0; 203 field:unsigned long mode; offset:32; size:4; signed:0; 204 205 206 print fmt: "(%lx) dfd=%lx filename=%lx flags=%lx mode=%lx", REC->__probe_ip, 207 REC->dfd, REC->filename, REC->flags, REC->mode 208 209You can see that the event has 4 arguments as in the expressions you specified. 210:: 211 212 echo > /sys/kernel/debug/tracing/kprobe_events 213 214This clears all probe points. 215 216Or, 217:: 218 219 echo -:myprobe >> kprobe_events 220 221This clears probe points selectively. 222 223Right after definition, each event is disabled by default. For tracing these 224events, you need to enable it. 225:: 226 227 echo 1 > /sys/kernel/debug/tracing/events/kprobes/myprobe/enable 228 echo 1 > /sys/kernel/debug/tracing/events/kprobes/myretprobe/enable 229 230Use the following command to start tracing in an interval. 231:: 232 233 # echo 1 > tracing_on 234 Open something... 235 # echo 0 > tracing_on 236 237And you can see the traced information via /sys/kernel/debug/tracing/trace. 238:: 239 240 cat /sys/kernel/debug/tracing/trace 241 # tracer: nop 242 # 243 # TASK-PID CPU# TIMESTAMP FUNCTION 244 # | | | | | 245 <...>-1447 [001] 1038282.286875: myprobe: (do_sys_open+0x0/0xd6) dfd=3 filename=7fffd1ec4440 flags=8000 mode=0 246 <...>-1447 [001] 1038282.286878: myretprobe: (sys_openat+0xc/0xe <- do_sys_open) $retval=fffffffffffffffe 247 <...>-1447 [001] 1038282.286885: myprobe: (do_sys_open+0x0/0xd6) dfd=ffffff9c filename=40413c flags=8000 mode=1b6 248 <...>-1447 [001] 1038282.286915: myretprobe: (sys_open+0x1b/0x1d <- do_sys_open) $retval=3 249 <...>-1447 [001] 1038282.286969: myprobe: (do_sys_open+0x0/0xd6) dfd=ffffff9c filename=4041c6 flags=98800 mode=10 250 <...>-1447 [001] 1038282.286976: myretprobe: (sys_open+0x1b/0x1d <- do_sys_open) $retval=3 251 252 253Each line shows when the kernel hits an event, and <- SYMBOL means kernel 254returns from SYMBOL(e.g. "sys_open+0x1b/0x1d <- do_sys_open" means kernel 255returns from do_sys_open to sys_open+0x1b). 256