docs/tracing.txt

   1 = Tracing =
   2
   3 == Introduction ==
   4
   5 This document describes the tracing infrastructure in QEMU and how to use it
   6 for debugging, profiling, and observing execution.
   7
   8 == Quickstart ==
   9
  10 1. Build with the 'simple' trace backend:
  11
  12     ./configure --trace-backend=simple
  13     make
  14
  15 2. Create a file with the events you want to trace:
  16
  17    echo bdrv_aio_readv   > /tmp/events
  18    echo bdrv_aio_writev >> /tmp/events
  19
  20 3. Run the virtual machine to produce a trace file:
  21
  22     qemu -trace events=/tmp/events ... # your normal QEMU invocation
  23
  24 4. Pretty-print the binary trace file:
  25
  26     ./simpletrace.py trace-events trace-*
  27
  28 == Trace events ==
  29
  30 There is a set of static trace events declared in the "trace-events" source
  31 file.  Each trace event declaration names the event, its arguments, and the
  32 format string which can be used for pretty-printing:
  33
  34     qemu_malloc(size_t size, void *ptr) "size %zu ptr %p"
  35     qemu_free(void *ptr) "ptr %p"
  36
  37 The "trace-events" file is processed by the "tracetool" script during build to
  38 generate code for the trace events.  Trace events are invoked directly from
  39 source code like this:
  40
  41     #include "trace.h"  /* needed for trace event prototype */
  42
  43     void *qemu_malloc(size_t size)
  44     {
  45         void *ptr;
  46         if (!size && !allow_zero_malloc()) {
  47             abort();
  48         }
  49         ptr = oom_check(malloc(size ? size : 1));
  50         trace_qemu_malloc(size, ptr);  /* <-- trace event */
  51         return ptr;
  52     }
  53
  54 === Declaring trace events ===
  55
  56 The "tracetool" script produces the trace.h header file which is included by
  57 every source file that uses trace events.  Since many source files include
  58 trace.h, it uses a minimum of types and other header files included to keep the
  59 namespace clean and compile times and dependencies down.
  60
  61 Trace events should use types as follows:
  62
  63  * Use stdint.h types for fixed-size types.  Most offsets and guest memory
  64    addresses are best represented with uint32_t or uint64_t.  Use fixed-size
  65    types over primitive types whose size may change depending on the host
  66    (32-bit versus 64-bit) so trace events don't truncate values or break
  67    the build.
  68
  69  * Use void * for pointers to structs or for arrays.  The trace.h header
  70    cannot include all user-defined struct declarations and it is therefore
  71    necessary to use void * for pointers to structs.
  72
  73    Pointers (including char *) cannot be dereferenced easily (or at all) in
  74    some trace backends.  If pointers are used, ensure they are meaningful by
  75    themselves and do not assume the data they point to will be traced.  Do
  76    not pass in string arguments.
  77
  78  * For everything else, use primitive scalar types (char, int, long) with the
  79    appropriate signedness.
  80
  81 Format strings should reflect the types defined in the trace event.  Take
  82 special care to use PRId64 and PRIu64 for int64_t and uint64_t types,
  83 respectively.  This ensures portability between 32- and 64-bit platforms.  Note
  84 that format strings must begin and end with double quotes.  When using
  85 portability macros, ensure they are preceded and followed by double quotes:
  86 "value %"PRIx64"".
  87
  88 === Hints for adding new trace events ===
  89
  90 1. Trace state changes in the code.  Interesting points in the code usually
  91    involve a state change like starting, stopping, allocating, freeing.  State
  92    changes are good trace events because they can be used to understand the
  93    execution of the system.
  94
  95 2. Trace guest operations.  Guest I/O accesses like reading device registers
  96    are good trace events because they can be used to understand guest
  97    interactions.
  98
  99 3. Use correlator fields so the context of an individual line of trace output
 100    can be understood.  For example, trace the pointer returned by malloc and
 101    used as an argument to free.  This way mallocs and frees can be matched up.
 102    Trace events with no context are not very useful.
 103
 104 4. Name trace events after their function.  If there are multiple trace events
 105    in one function, append a unique distinguisher at the end of the name.
 106
 107 5. If specific trace events are going to be called a huge number of times, this
 108    might have a noticeable performance impact even when the trace events are
 109    programmatically disabled. In this case you should declare the trace event
 110    with the "disable" property, which will effectively disable it at compile
 111    time (using the "nop" backend).
 112
 113 == Generic interface and monitor commands ==
 114
 115 You can programmatically query and control the dynamic state of trace events
 116 through a backend-agnostic interface:
 117
 118 * trace_print_events
 119
 120 * trace_event_set_state
 121   Enables or disables trace events at runtime inside QEMU.
 122   The function returns "true" if the state of the event has been successfully
 123   changed, or "false" otherwise:
 124
 125     #include "trace/control.h"
 126
 127     trace_event_set_state("virtio_irq", true); /* enable */
 128     [...]
 129     trace_event_set_state("virtio_irq", false); /* disable */
 130
 131 Note that some of the backends do not provide an implementation for this
 132 interface, in which case QEMU will just print a warning.
 133
 134 This functionality is also provided through monitor commands:
 135
 136 * info trace-events
 137   View available trace events and their state.  State 1 means enabled, state 0
 138   means disabled.
 139
 140 * trace-event NAME on|off
 141   Enable/disable a given trace event.
 142
 143 The "-trace events=<file>" command line argument can be used to enable the
 144 events listed in <file> from the very beginning of the program. This file must
 145 contain one event name per line.
 146
 147 == Trace backends ==
 148
 149 The "tracetool" script automates tedious trace event code generation and also
 150 keeps the trace event declarations independent of the trace backend.  The trace
 151 events are not tightly coupled to a specific trace backend, such as LTTng or
 152 SystemTap.  Support for trace backends can be added by extending the "tracetool"
 153 script.
 154
 155 The trace backend is chosen at configure time and only one trace backend can
 156 be built into the binary:
 157
 158     ./configure --trace-backend=simple
 159
 160 For a list of supported trace backends, try ./configure --help or see below.
 161
 162 The following subsections describe the supported trace backends.
 163
 164 === Nop ===
 165
 166 The "nop" backend generates empty trace event functions so that the compiler
 167 can optimize out trace events completely.  This is the default and imposes no
 168 performance penalty.
 169
 170 Note that regardless of the selected trace backend, events with the "disable"
 171 property will be generated with the "nop" backend.
 172
 173 === Stderr ===
 174
 175 The "stderr" backend sends trace events directly to standard error.  This
 176 effectively turns trace events into debug printfs.
 177
 178 This is the simplest backend and can be used together with existing code that
 179 uses DPRINTF().
 180
 181 === Simpletrace ===
 182
 183 The "simple" backend supports common use cases and comes as part of the QEMU
 184 source tree.  It may not be as powerful as platform-specific or third-party
 185 trace backends but it is portable.  This is the recommended trace backend
 186 unless you have specific needs for more advanced backends.
 187
 188 ==== Monitor commands ====
 189
 190 * info trace
 191   Display the contents of trace buffer.  This command dumps the trace buffer
 192   with simple formatting.  For full pretty-printing, use the simpletrace.py
 193   script on a binary trace file.
 194
 195   The trace buffer is written into until full.  The full trace buffer is
 196   flushed and emptied.  This means the 'info trace' will display few or no
 197   entries if the buffer has just been flushed.
 198
 199 * trace-file on|off|flush|set <path>
 200   Enable/disable/flush the trace file or set the trace file name.
 201
 202 ==== Analyzing trace files ====
 203
 204 The "simple" backend produces binary trace files that can be formatted with the
 205 simpletrace.py script.  The script takes the "trace-events" file and the binary
 206 trace:
 207
 208     ./simpletrace.py trace-events trace-12345
 209
 210 You must ensure that the same "trace-events" file was used to build QEMU,
 211 otherwise trace event declarations may have changed and output will not be
 212 consistent.
 213
 214 === LTTng Userspace Tracer ===
 215
 216 The "ust" backend uses the LTTng Userspace Tracer library.  There are no
 217 monitor commands built into QEMU, instead UST utilities should be used to list,
 218 enable/disable, and dump traces.
 219
 220 === SystemTap ===
 221
 222 The "dtrace" backend uses DTrace sdt probes but has only been tested with
 223 SystemTap.  When SystemTap support is detected a .stp file with wrapper probes
 224 is generated to make use in scripts more convenient.  This step can also be
 225 performed manually after a build in order to change the binary name in the .stp
 226 probes:
 227
 228     scripts/tracetool --dtrace --stap \
 229                       --binary path/to/qemu-binary \
 230                       --target-type system \
 231                       --target-arch x86_64 \
 232                       <trace-events >qemu.stp