OpenCSD - CoreSight Trace Decode Library  1.1.1
prog_guide/prog_guide_generic_pkts.md
Go to the documentation of this file.
1 OpenCSD Library - Generic Trace Packet Descriptions {#generic_pkts}
2 ===================================================
3 
4 @brief Interpretation of the Generic Trace output packets.
5 
6 Generic Trace Packets - Collection.
7 -----------------------------------
8 
9 ### Packet interface ###
10 
11 The generic trace packets are the fully decoded output from the trace library.
12 
13 These are delivered to the client application in the form of a callback function. Packets from all trace sources
14 will use the same single callback function, with the CoreSight Trace ID provided to identify the source.
15 
16 The callback is in the form of an interface class ITrcGenElemIn, which has a single function:
17 
18 ~~~{.cpp}
19 virtual ocsd_datapath_resp_t TraceElemIn( const ocsd_trc_index_t index_sop,
20  const uint8_t trc_chan_id,
21  const OcsdTraceElement &elem
22  ) = 0;
23 ~~~
24 
25 The client program will create derived class providing this interface to collect trace packets from the library.
26 
27 The parameters describe the output packet and source channel:
28 |Parameter | Description |
29 |:--------------------------------|:------------------------------------------------------------------------|
30 | `ocsd_trc_index_t index_sop` | Index of the first byte of the trace packet that generated this output. |
31 | `uint8_t trc_chan_id` | The source CoreSight Trace ID. |
32 | `OcsdTraceElement &elem` | The packet class - wraps the `ocsd_generic_trace_elem` structure. |
33 
34 _Note_ : `index_sop` may be the same for multiple output packets. This is due to an one byte atom packet which
35 can represent multiple atoms and hence multiple ranges.
36 
37 The C-API provides a similarly specified callback function definition, with an additional opaque `void *` pointer
38 that the client application may use.
39 
40 ~~~{.c}
41 /** function pointer type for decoder outputs. all protocols, generic data element input */
42 typedef ocsd_datapath_resp_t (* FnTraceElemIn)( const void *p_context,
43  const ocsd_trc_index_t index_sop,
44  const uint8_t trc_chan_id,
45  const ocsd_generic_trace_elem *elem);
46 ~~~
47 
48 ### The Packet Structure ###
49 
50 ~~~{.c}
51 typedef struct _ocsd_generic_trace_elem {
52  ocsd_gen_trc_elem_t elem_type; /* Element type - remaining data interpreted according to this value */
53  ocsd_isa isa; /* instruction set for executed instructions */
54  ocsd_vaddr_t st_addr; /* start address for instruction execution range / inaccessible code address / data address */
55  ocsd_vaddr_t en_addr; /* end address (exclusive) for instruction execution range. */
56  ocsd_pe_context context; /* PE Context */
57  uint64_t timestamp; /* timestamp value for TS element type */
58  uint32_t cycle_count; /* cycle count for explicit cycle count element, or count for element with associated cycle count */
59  ocsd_instr_type last_i_type; /* Last instruction type if instruction execution range */
60  ocsd_instr_subtype last_i_subtype; /* sub type for last instruction in range */
61 
62  //! per element flags
63  union {
64  struct {
65  uint32_t last_instr_exec:1; /* 1 if last instruction in range was executed; */
66  uint32_t last_instr_sz:3; /* size of last instruction in bytes (2/4) */
67  uint32_t has_cc:1; /* 1 if this packet has a valid cycle count included (e.g. cycle count included as part of instruction range packet, always 1 for pure cycle count packet.*/
68  uint32_t cpu_freq_change:1; /* 1 if this packet indicates a change in CPU frequency */
69  uint32_t excep_ret_addr:1; /* 1 if en_addr is the preferred exception return address on exception packet type */
70  uint32_t excep_data_marker:1; /* 1 if the exception entry packet is a data push marker only, with no address information (used typically in v7M trace for marking data pushed onto stack) */
71  uint32_t extended_data:1; /* 1 if the packet extended data pointer is valid. Allows packet extensions for custom decoders, or additional data payloads for data trace. */
72  uint32_t has_ts:1; /* 1 if the packet has an associated timestamp - e.g. SW/STM trace TS+Payload as a single packet */
73  uint32_t last_instr_cond:1; /* 1 if the last instruction was conditional */
74  uint32_t excep_ret_addr_br_tgt:1; /* 1 if exception return address (en_addr) is also the target of a taken branch addr from the previous range. */
75  };
76  uint32_t flag_bits;
77  };
78 
79  //! packet specific payloads
80  union {
81  uint32_t exception_number; /* exception number for exception type packets */
82  trace_event_t trace_event; /* Trace event - trigger etc */
83  trace_on_reason_t trace_on_reason; /* reason for the trace on packet */
84  ocsd_swt_info_t sw_trace_info; /* software trace packet info */
85  uint32_t num_instr_range; /* number of instructions covered by range packet (for T32 this cannot be calculated from en-st/i_size) */
86  unsync_info_t unsync_eot_info; /* additional information for unsync / end-of-trace packets. */
87  trace_marker_payload_t sync_marker; /* marker element - sync later element to position in stream */
88  trace_memtrans_t mem_trans; /* memory transaction packet - transaction event */
89  };
90 
91  const void *ptr_extended_data; /* pointer to extended data buffer (data trace, sw trace payload) / custom structure */
92 
93 } ocsd_generic_trace_elem;
94 ~~~
95 
96 The packet structure contains multiple fields and flag bits. The validity of any of these fields or flags
97 is dependent on the `elem_type` member. The client program must not assume that field values will persist
98 between packets, and must process all valid data during the callback function.
99 
100 The packet reference guide below defines the fields valid for each packet type.
101 
102 --------------------------------------------------------------------------------------------------
103 
104 Generic Trace Packets - Packet Reference.
105 -----------------------------------------
106 
107 This section contains reference descriptions of each of the generic trace packets types define as part of the
108 `ocsd_gen_trc_elem_t` enum value that appears as the first `elem_type` field in the packet structure.
109 
110 The descriptions will include information on which fields in the packets are always valid, optional and any protocol specific information.
111 
112 The tags used in the reference are:-
113 - __packet fields valid__ : fields that are always valid and filled for this packet type.
114 - __packet fields optional__ : fields that _may_ be filled for this packet type.
115  The form `flag -> field` indicates a flag that may be set and the value that is valid if the flag is true
116 - __protocol specific__ : indicates type or fields may be source protocol specific.
117 
118 _Note_: while most of the packets are not protocol specific, there are some protocol differences that mean
119 certain types and fields will differ slightly across protocols. These differences are highlighted in the
120 reference.
121 
122 ### OCSD_GEN_TRC_ELEM_NO_SYNC ###
123 __packet fields valid__: None
124 
125 Element output before the decoder has synchronised with the input stream, or synchronisation is lost.
126 
127 ### OCSD_GEN_TRC_ELEM_INSTR_RANGE ###
128 __packet fields valid__: `isa, st_addr, en_addr, last_i_type, last_i_subtype, last_instr_exec, last_instr_sz, num_instr_range, last_instr_cond`
129 
130 __packet fields optional__: `has_cc -> cycle_count,`
131 
132 __protocol specific__ : ETMv3, PTM
133 
134 This should be the most common packet output for full trace decode. Represents a range of instructions of
135 a single `isa`, executed by the PE. Instruction byte range is from `st_addr` (inclusive) to `en_addr` (exclusive).
136 The total number of instructions executed for the range is given in `num_instr_range`.
137 
138 Information on the last instruction in the range is provided. `last_i_type` shows if the last instruction
139 was a branch or otherwise - which combined with `last_instr_exec` determines if the branch was taken.
140 The last instruction size in bytes is given, to allow clients to quickly determine the address of the last
141 instruction by subtraction from `en_addr`. This value can be 2 or 4 bytes in the T32 instruction set.
142 
143 __ETMv3, PTM__ : These protocols can output a cycle count directly as part of the trace packet that generates
144 the trace range. In this case `has_cc` will be 1 and `cycle_count` will be valid.
145 
146 
147 ### OCSD_GEN_TRC_ELEM_I_RANGE_NOPATH ###
148 __packet fields valid__: `isa, st_addr, en_addr, num_instr_range`
149 
150 `num_instr_range` represents the number of instructions executed in this range, but there is incomplete information
151 as to program execution path from start to end of range.
152 If `num_instr` is 0, then an unknown number of instructions were executed between the start and end of the range.
153 `st_addr` represents the start of execution represented by this packet.
154 `en_addr` represents the address where execution will continue from after the instructions represented by this packet.
155 `isa` represents the ISA for the instruction at `en_addr`.
156 
157 Used when ETMv4 Q elements are being traced.
158 
159 
160 ### OCSD_GEN_TRC_ELEM_ADDR_NACC ###
161 __packet fields valid__: `st_addr`
162 
163 Trace decoder found address in trace that cannot be accessed in the mapped memory images.
164 `st_addr` is the address that cannot be found.
165 
166 Decoder will wait for new address to appear in trace before attempting to restart decoding.
167 
168 
169 ### OCSD_GEN_TRC_ELEM_UNKNOWN ###
170 __packet fields valid__: None
171 
172 Decoder saw invalid packet for protocol being processed. Likely incorrect protocol settings, or corrupted
173 trace data.
174 
175 ### OCSD_GEN_TRC_ELEM_TRACE_ON ###
176 __packet fields valid__: trace_on_reason
177 
178 __packet fields optional__: `has_cc -> cycle_count,`
179 
180 __protocol specific__ : ETMv3, PTM
181 
182 Notification that trace has started / is synced after a discontinuity or at start of trace decode.
183 
184 __ETMv3, PTM__ : These protocols can output a cycle count directly as part of the trace packet that generates
185 the trace on indicator. In this case `has_cc` will be 1 and `cycle_count` will be valid.
186 
187 
188 ### OCSD_GEN_TRC_ELEM_EO_TRACE ###
189 __packet fields valid__: None
190 
191 Marker for end of trace data. Sent once for each CoreSight ID channel.
192 
193 ### OCSD_GEN_TRC_ELEM_PE_CONTEXT ###
194 __packet fields valid__: context
195 
196 __packet fields optional__: `has_cc -> cycle_count,`
197 
198 __protocol specific__ : ETMv3, PTM
199 
200 This packet indicates an update to the PE context - which may be the initial context in a trace stream, or a
201 change since the trace started.
202 
203 The context is contained in a `ocsd_pe_context` structure.
204 
205 ~~~{.c}
206 typedef struct _ocsd_pe_context {
207  ocsd_sec_level security_level; /* security state */
208  ocsd_ex_level exception_level; /* exception level */
209  uint32_t context_id; /* context ID */
210  uint32_t vmid; /* VMID */
211  struct {
212  uint32_t bits64:1; /* 1 if 64 bit operation */
213  uint32_t ctxt_id_valid:1; /* 1 if context ID value valid */
214  uint32_t vmid_valid:1; /* 1 if VMID value is valid */
215  uint32_t el_valid:1; /* 1 if EL value is valid (ETMv4 traces current EL, other protocols do not) */
216  };
217 } ocsd_pe_context;
218 ~~~
219 
220 __ETMv3, PTM__ : These protocols can output a cycle count directly as part of the trace packet that generates
221 the PE context. In this case `has_cc` will be 1 and `cycle_count` will be valid.
222 
223 __ETMv3__ : From ETM 3.5 onwards, exception_level can be set to `ocsd_EL2` when tracing through hypervisor code.
224 On all other occasions this will be set to `ocsd_EL_unknown`.
225 
226 
227 ### OCSD_GEN_TRC_ELEM_ADDR_UNKNOWN ###
228 __packet fields optional__: `has_cc -> cycle_count,`
229 
230 __protocol specific__: ETMv3
231 
232 This packet will only be seen when decoding an ETMv3 protocol source. This indicates that the decoder
233 is waiting for a valid address in order to process trace correctly.
234 
235 The packet can have a cycle count associated with it which the client must account for when tracking cycles used.
236 The packet will be sent once when unknown address occurs. Further `OCSD_GEN_TRC_ELEM_CYCLE_COUNT` packets may follow
237  before the decode receives a valid address to continue decode.
238 
239 
240 ### OCSD_GEN_TRC_ELEM_EXCEPTION ###
241 __packet fields valid__: `exception_number`
242 
243 __packet fields optional__: `has_cc -> cycle_count, excep_ret_addr -> en_addr, excep_data_marker, excep_ret_addr_br_tgt`
244 
245 __protocol specific__: ETMv4, ETMv3, PTM
246 
247 All protocols will include the exception number in the packet.
248 
249 __ETMv4__ : This protocol may provide the preferred return address for the exception - this is the address of
250 the instruction that could be executed on exception return. This address appears in `en_addr` if `excep_ret_addr` = 1.
251 
252 Additionally, this address could also represent the target address of a branch, if the exception occured at the branch target, before any further instructions were execute. If htis is the case then the excep_ret_addr_br_tgt flag will be set. This makes explicit what was previously only implied by teh packet ordered. This information could be used for clients such as perf that branch source/target address pairs.
253 
254 __ETMv3__ : This can set the `excep_data_marker` flag. This indicates that the exception packet is a marker
255 to indicate exception entry in a 7M profile core, for the purposes of tracking data. This will __not__ provide
256 an exception number in this case.
257 
258 __PTM__ : Can have an associated cycle count (`has_cc == 1`), and may provide preferred return address in `en_addr`
259 if `excep_ret_addr` = 1.
260 
261 ### OCSD_GEN_TRC_ELEM_EXCEPTION_RET ###
262 __packet fields valid__: None
263 
264 Marker that a preceding branch was an exception return.
265 
266 ### OCSD_GEN_TRC_ELEM_TIMESTAMP ###
267 __packet fields valid__: `timestamp`
268 
269 __packet fields optional__: `has_cc -> cycle_count,`
270 
271 __protocol specific__: ETMv4, PTM
272 
273 The timestamp packet explicitly provides a timestamp value for the trace stream ID in the callback interface.
274 
275 __PTM__ : This can have an associated cycle count (`has_cc == 1`). For this protocol, the cycle count __is__ part
276 of the cumulative cycle count for the trace session.
277 
278 __ETMv4__ : This can have an associated cycle count (`has_cc == 1`). For this protocl, the cycle coun represents
279 the number of cycles between the previous cycle count packet and this timestamp packet, but __is not__ part of
280 the cumulative cycle count for the trace session.
281 
282 
283 ### OCSD_GEN_TRC_ELEM_CYCLE_COUNT ###
284 __packet fields valid__: `has_cc -> cycle_count`
285 
286 Packet contains a cycle count value. A cycle count value represents the number of cycles passed since the
287 last cycle count value seen. The cycle count value may be associated with a specific packet or instruction
288 range preceding the cycle count packet.
289 
290 Cycle count packets may be added together to build a cumulative count for the trace session.
291 
292 ### OCSD_GEN_TRC_ELEM_EVENT ###
293 __packet fields valid__: `trace_event`
294 
295 This is a hardware event injected into the trace by the ETM/PTM hardware resource programming. See the
296 relevent trace hardware reference manuals for the programming of these events.
297 
298 The `trace_event` is a `trace_event_t` structure that can have an event type - and an event number.
299 
300 ~~~{.c}
301 typedef struct _trace_event_t {
302  uint16_t ev_type; /* event type - unknown (0) trigger (1), numbered event (2)*/
303  uint16_t ev_number; /* event number if numbered event type */
304 } trace_event_t;
305 ~~~
306 
307 The event types depend on the trace hardware:-
308 
309 __ETMv4__ : produces numbered events. The event number is a bitfield of up to four events that occurred.
310 Events 0-3 -> bits 0-3. The bitfield allows a single packet to represent multiple different events occurring.
311 
312 _Note_: The ETMv4 specification has further information on timing of events and event packets. Event 0
313 is also considered a trigger event in ETMv4 hardware, but is not explicitly represented as such in the OCSD protocol.
314 
315 __PTM__, __ETMv3__ : produce trigger events. Event number always set to 0.
316 
317 
318 ### OCSD_GEN_TRC_ELEM_SWTRACE ###
319 __packet fields valid__: `sw_trace_info`
320 
321 __packet fields optional__: `has_ts -> timestamp`, ` extended_data -> ptr_extended_data`
322 
323 The Software trace packet always has a filled in `sw_trace_info` field to describe the current master and channel ID,
324 plus the packet type and size of any payload data.
325 
326 SW trace packets that have a payload will use the extended_data flag and pointer to deliver this data.
327 
328 SW trace packets that include timestamp information will us the `has_ts` flag and fill in the timestamp value.
329 
330 
331 ### OCSD_GEN_TRC_ELEM_SYNC_MARKER ###
332 __packet fields valid__: `sync_marker`
333 
334 Synchronisation marker - marks position in stream of an element that is output later.
335 e.g. a timestamp marker can be output to represent the correct position in the stream for a
336 timestamp packet the is output later.
337 
338 The `sync_marker` field has a structure as shown below.
339 
340 ~~~{.c}
341 typedef enum _trace_sync_marker_t {
342  ELEM_MARKER_TS, /**< Marker for timestamp element */
343 } trace_sync_marker_t;
344 
345 typedef struct _trace_marker_payload_t {
346  trace_sync_marker_t type; /**< type of sync marker */
347  uint32_t value; /**< sync marker value - usage depends on type */
348 } trace_marker_payload_t;
349 ~~~
350 
351 ### OCSD_GEN_TRC_ELEM_MEMTRANS ###
352 __packet fields valid__: `mem_trans`
353 
354 Memory transaction elements may appear in the output stream, if they are not otherwise cancelled
355 by speculative trace packets.
356 
357 The memory transaction field has values as defined in the enum below:-
358 
359 ~~~{.c}
360 typedef enum _memtrans_t {
361  OCSD_MEM_TRANS_TRACE_INIT,/* Trace started while PE in transactional state */
362  OCSD_MEM_TRANS_START, /* Trace after this packet is part of a transactional memory sequence */
363  OCSD_MEM_TRANS_COMMIT, /* Transactional memory sequence valid. */
364  OCSD_MEM_TRANS_FAIL, /* Transactional memory sequence failed - operations since start of transaction have been unwound. */
365 } trace_memtrans_t;
366 ~~~
367 
368 
369 ### OCSD_GEN_TRC_ELEM_CUSTOM ###
370 __packet fields optional__: `extended_data -> ptr_extended_data`,_any others_
371 
372 Custom protocol decoders can use this packet type to provide protocol specific information.
373 
374 Standard fields may be used for similar purposes as defined above, or the extended data pointer can reference
375 other data.
376 
377 --------------------------------------------------------------------------------------------------
378 
379 Generic Trace Packets - Notes on interpretation.
380 ------------------------------------------------
381 
382 The interpretation of the trace output should always be done with reference to the underlying protocol
383 specifications.
384 
385 While the output packets are in general protocol agnostic, there are some inevitable
386 differences related to the underlying protocol that stem from the development of the trace hardware over time.
387 
388 ### OCSD ranges and Trace Atom Packets ###
389 The most common raw trace packet in all the protocols is the Atom packet, and this packet is the basis for most of
390 the `OCSD_GEN_TRC_ELEM_INSTR_RANGE` packets output from the library. A trace range will be output for each atom
391 in the raw trace stream - the `last_instr_exec` flag taking the value of the Atom - 1 for E, 0 for N.
392 
393 `OCSD_GEN_TRC_ELEM_INSTR_RANGE` packets can also be generated for non-atom packets, where flow changes - e.g.
394 exceptions.
395 
396 
397 ### Multi feature OCSD output packets ###
398 Where a raw trace packet contains additional information on top of the basic packet data, then this additional
399 information will be added to the OCSD output packet and flagged accordingly (in the `flag_bits` union in the
400 packet structure).
401 
402 Typically this will be atom+cycle count packets in ETMv3 and PTM protocols. For efficiency and to retain
403 the coupling between the information an `OCSD_GEN_TRC_ELEM_INSTR_RANGE` packet will be output in this case
404 with a `has_cc` flag set and the `cycle_count` value filled.
405 
406 ETMv3 and PTM can add a cycle count to a number of packets, or explicitly emit a cycle count only packet. By
407 contrast ETMv4 only emits cycle count only packets.
408 
409 Clients processing the library output must be aware of these optional additions to the base packet. The
410 OCSD packet descriptions above outline where the additional information can occur.
411 
412 ### Cycle counts ###
413 
414 Cycle counts are cumulative, and represent cycles since the last cycle count output.
415 Explicit cycle count packets are associated with the previous range event, otherwise where a
416 packet includes a cycle count as additional information, then the count is associated with that
417 specific packet - which will often be a range packet.
418 
419 The only exception to this is where the underlying protocol is ETMv4, and a cycle count is included
420 in a timestamp packet. Here the cycle count represents that number of cycles since the last cycle count
421 packet that occurred before the timestamp packet was emitted. This cycle count is not part of the cumulative
422 count. See the ETMv4 specification for further details.
423 
424 
425 ### Correlation - timestamps and cycle counts ###
426 
427 Different trace streams can be correlated using either timestamps, or timestamps plus cycle counts.
428 
429 Both timestamps and cycle counts are enabled by programming ETM control registers, and it is also possible
430 to control the frequency that timestamps appear, or the threshold at which cycle count packets are emitted by
431 additional programming.
432 
433 The output of timestamps and cycle counts increases the amount of trace generated, very significantly when cycle
434 counts are present, so the choice of generating these elements needs to be balanced against the requirement
435 for their use.
436 
437 Decent correlation can be gained by the use of timestamps alone - especially if the source is programmed to
438 produce them more frequently than the default timestamp events. More precise correllation can be performed if
439 the 'gaps' between timestamps can be resolved using cycle counts.
440 
441 Correlation is performed by identifying the same/close timestamp values in two separate trace streams. Cycle counts
442 if present can then be used to resolve the correlation with additional accuracy.
443 
444 
445 
446 
447 
448 
449 
450 
451 
452 
453