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.
  84
  85 === Hints for adding new trace events ===
  86
  87 1. Trace state changes in the code.  Interesting points in the code usually
  88    involve a state change like starting, stopping, allocating, freeing.  State
  89    changes are good trace events because they can be used to understand the
  90    execution of the system.
  91
  92 2. Trace guest operations.  Guest I/O accesses like reading device registers
  93    are good trace events because they can be used to understand guest
  94    interactions.
  95
  96 3. Use correlator fields so the context of an individual line of trace output
  97    can be understood.  For example, trace the pointer returned by malloc and
  98    used as an argument to free.  This way mallocs and frees can be matched up.
  99    Trace events with no context are not very useful.
 100
 101 4. Name trace events after their function.  If there are multiple trace events
 102    in one function, append a unique distinguisher at the end of the name.
 103
 104 5. If specific trace events are going to be called a huge number of times, this
 105    might have a noticeable performance impact even when the trace events are
 106    programmatically disabled. In this case you should declare the trace event
 107    with the "disable" property, which will effectively disable it at compile
 108    time (using the "nop" backend).
 109
 110 == Generic interface and monitor commands ==
 111
 112 You can programmatically query and control the dynamic state of trace events
 113 through a backend-agnostic interface:
 114
 115 * trace_print_events
 116
 117 * trace_event_set_state
 118   Enables or disables trace events at runtime inside QEMU.
 119   The function returns "true" if the state of the event has been successfully
 120   changed, or "false" otherwise:
 121
 122     #include "trace/control.h"
 123
 124     trace_event_set_state("virtio_irq", true); /* enable */
 125     [...]
 126     trace_event_set_state("virtio_irq", false); /* disable */
 127
 128 Note that some of the backends do not provide an implementation for this
 129 interface, in which case QEMU will just print a warning.
 130
 131 This functionality is also provided through monitor commands:
 132
 133 * info trace-events
 134   View available trace events and their state.  State 1 means enabled, state 0
 135   means disabled.
 136
 137 * trace-event NAME on|off
 138   Enable/disable a given trace event.
 139
 140 The "-trace events=<file>" command line argument can be used to enable the
 141 events listed in <file> from the very beginning of the program. This file must
 142 contain one event name per line.
 143
 144 == Trace backends ==
 145
 146 The "tracetool" script automates tedious trace event code generation and also
 147 keeps the trace event declarations independent of the trace backend.  The trace
 148 events are not tightly coupled to a specific trace backend, such as LTTng or
 149 SystemTap.  Support for trace backends can be added by extending the "tracetool"
 150 script.
 151
 152 The trace backend is chosen at configure time and only one trace backend can
 153 be built into the binary:
 154
 155     ./configure --trace-backend=simple
 156
 157 For a list of supported trace backends, try ./configure --help or see below.
 158
 159 The following subsections describe the supported trace backends.
 160
 161 === Nop ===
 162
 163 The "nop" backend generates empty trace event functions so that the compiler
 164 can optimize out trace events completely.  This is the default and imposes no
 165 performance penalty.
 166
 167 Note that regardless of the selected trace backend, events with the "disable"
 168 property will be generated with the "nop" backend.
 169
 170 === Stderr ===
 171
 172 The "stderr" backend sends trace events directly to standard error.  This
 173 effectively turns trace events into debug printfs.
 174
 175 This is the simplest backend and can be used together with existing code that
 176 uses DPRINTF().
 177
 178 === Simpletrace ===
 179
 180 The "simple" backend supports common use cases and comes as part of the QEMU
 181 source tree.  It may not be as powerful as platform-specific or third-party
 182 trace backends but it is portable.  This is the recommended trace backend
 183 unless you have specific needs for more advanced backends.
 184
 185 ==== Monitor commands ====
 186
 187 * info trace
 188   Display the contents of trace buffer.  This command dumps the trace buffer
 189   with simple formatting.  For full pretty-printing, use the simpletrace.py
 190   script on a binary trace file.
 191
 192   The trace buffer is written into until full.  The full trace buffer is
 193   flushed and emptied.  This means the 'info trace' will display few or no
 194   entries if the buffer has just been flushed.
 195
 196 * trace-file on|off|flush|set <path>
 197   Enable/disable/flush the trace file or set the trace file name.
 198
 199 ==== Analyzing trace files ====
 200
 201 The "simple" backend produces binary trace files that can be formatted with the
 202 simpletrace.py script.  The script takes the "trace-events" file and the binary
 203 trace:
 204
 205     ./simpletrace.py trace-events trace-12345
 206
 207 You must ensure that the same "trace-events" file was used to build QEMU,
 208 otherwise trace event declarations may have changed and output will not be
 209 consistent.
 210
 211 === LTTng Userspace Tracer ===
 212
 213 The "ust" backend uses the LTTng Userspace Tracer library.  There are no
 214 monitor commands built into QEMU, instead UST utilities should be used to list,
 215 enable/disable, and dump traces.
 216
 217 === SystemTap ===
 218
 219 The "dtrace" backend uses DTrace sdt probes but has only been tested with
 220 SystemTap.  When SystemTap support is detected a .stp file with wrapper probes
 221 is generated to make use in scripts more convenient.  This step can also be
 222 performed manually after a build in order to change the binary name in the .stp
 223 probes:
 224
 225     scripts/tracetool --dtrace --stap \
 226                       --binary path/to/qemu-binary \
 227                       --target-type system \
 228                       --target-arch x86_64 \
 229                       <trace-events >qemu.stp