coresight.rst (29240B)
1====================================== 2Coresight - HW Assisted Tracing on ARM 3====================================== 4 5 :Author: Mathieu Poirier <mathieu.poirier@linaro.org> 6 :Date: September 11th, 2014 7 8Introduction 9------------ 10 11Coresight is an umbrella of technologies allowing for the debugging of ARM 12based SoC. It includes solutions for JTAG and HW assisted tracing. This 13document is concerned with the latter. 14 15HW assisted tracing is becoming increasingly useful when dealing with systems 16that have many SoCs and other components like GPU and DMA engines. ARM has 17developed a HW assisted tracing solution by means of different components, each 18being added to a design at synthesis time to cater to specific tracing needs. 19Components are generally categorised as source, link and sinks and are 20(usually) discovered using the AMBA bus. 21 22"Sources" generate a compressed stream representing the processor instruction 23path based on tracing scenarios as configured by users. From there the stream 24flows through the coresight system (via ATB bus) using links that are connecting 25the emanating source to a sink(s). Sinks serve as endpoints to the coresight 26implementation, either storing the compressed stream in a memory buffer or 27creating an interface to the outside world where data can be transferred to a 28host without fear of filling up the onboard coresight memory buffer. 29 30At typical coresight system would look like this:: 31 32 ***************************************************************** 33 **************************** AMBA AXI ****************************===|| 34 ***************************************************************** || 35 ^ ^ | || 36 | | * ** 37 0000000 ::::: 0000000 ::::: ::::: @@@@@@@ |||||||||||| 38 0 CPU 0<-->: C : 0 CPU 0<-->: C : : C : @ STM @ || System || 39 |->0000000 : T : |->0000000 : T : : T :<--->@@@@@ || Memory || 40 | #######<-->: I : | #######<-->: I : : I : @@@<-| |||||||||||| 41 | # ETM # ::::: | # PTM # ::::: ::::: @ | 42 | ##### ^ ^ | ##### ^ ! ^ ! . | ||||||||| 43 | |->### | ! | |->### | ! | ! . | || DAP || 44 | | # | ! | | # | ! | ! . | ||||||||| 45 | | . | ! | | . | ! | ! . | | | 46 | | . | ! | | . | ! | ! . | | * 47 | | . | ! | | . | ! | ! . | | SWD/ 48 | | . | ! | | . | ! | ! . | | JTAG 49 *****************************************************************<-| 50 *************************** AMBA Debug APB ************************ 51 ***************************************************************** 52 | . ! . ! ! . | 53 | . * . * * . | 54 ***************************************************************** 55 ******************** Cross Trigger Matrix (CTM) ******************* 56 ***************************************************************** 57 | . ^ . . | 58 | * ! * * | 59 ***************************************************************** 60 ****************** AMBA Advanced Trace Bus (ATB) ****************** 61 ***************************************************************** 62 | ! =============== | 63 | * ===== F =====<---------| 64 | ::::::::: ==== U ==== 65 |-->:: CTI ::<!! === N === 66 | ::::::::: ! == N == 67 | ^ * == E == 68 | ! &&&&&&&&& IIIIIII == L == 69 |------>&& ETB &&<......II I ======= 70 | ! &&&&&&&&& II I . 71 | ! I I . 72 | ! I REP I<.......... 73 | ! I I 74 | !!>&&&&&&&&& II I *Source: ARM ltd. 75 |------>& TPIU &<......II I DAP = Debug Access Port 76 &&&&&&&&& IIIIIII ETM = Embedded Trace Macrocell 77 ; PTM = Program Trace Macrocell 78 ; CTI = Cross Trigger Interface 79 * ETB = Embedded Trace Buffer 80 To trace port TPIU= Trace Port Interface Unit 81 SWD = Serial Wire Debug 82 83While on target configuration of the components is done via the APB bus, 84all trace data are carried out-of-band on the ATB bus. The CTM provides 85a way to aggregate and distribute signals between CoreSight components. 86 87The coresight framework provides a central point to represent, configure and 88manage coresight devices on a platform. This first implementation centers on 89the basic tracing functionality, enabling components such ETM/PTM, funnel, 90replicator, TMC, TPIU and ETB. Future work will enable more 91intricate IP blocks such as STM and CTI. 92 93 94Acronyms and Classification 95--------------------------- 96 97Acronyms: 98 99PTM: 100 Program Trace Macrocell 101ETM: 102 Embedded Trace Macrocell 103STM: 104 System trace Macrocell 105ETB: 106 Embedded Trace Buffer 107ITM: 108 Instrumentation Trace Macrocell 109TPIU: 110 Trace Port Interface Unit 111TMC-ETR: 112 Trace Memory Controller, configured as Embedded Trace Router 113TMC-ETF: 114 Trace Memory Controller, configured as Embedded Trace FIFO 115CTI: 116 Cross Trigger Interface 117 118Classification: 119 120Source: 121 ETMv3.x ETMv4, PTMv1.0, PTMv1.1, STM, STM500, ITM 122Link: 123 Funnel, replicator (intelligent or not), TMC-ETR 124Sinks: 125 ETBv1.0, ETB1.1, TPIU, TMC-ETF 126Misc: 127 CTI 128 129 130Device Tree Bindings 131-------------------- 132 133See Documentation/devicetree/bindings/arm/coresight.txt for details. 134 135As of this writing drivers for ITM, STMs and CTIs are not provided but are 136expected to be added as the solution matures. 137 138 139Framework and implementation 140---------------------------- 141 142The coresight framework provides a central point to represent, configure and 143manage coresight devices on a platform. Any coresight compliant device can 144register with the framework for as long as they use the right APIs: 145 146.. c:function:: struct coresight_device *coresight_register(struct coresight_desc *desc); 147.. c:function:: void coresight_unregister(struct coresight_device *csdev); 148 149The registering function is taking a ``struct coresight_desc *desc`` and 150register the device with the core framework. The unregister function takes 151a reference to a ``struct coresight_device *csdev`` obtained at registration time. 152 153If everything goes well during the registration process the new devices will 154show up under /sys/bus/coresight/devices, as showns here for a TC2 platform:: 155 156 root:~# ls /sys/bus/coresight/devices/ 157 replicator 20030000.tpiu 2201c000.ptm 2203c000.etm 2203e000.etm 158 20010000.etb 20040000.funnel 2201d000.ptm 2203d000.etm 159 root:~# 160 161The functions take a ``struct coresight_device``, which looks like this:: 162 163 struct coresight_desc { 164 enum coresight_dev_type type; 165 struct coresight_dev_subtype subtype; 166 const struct coresight_ops *ops; 167 struct coresight_platform_data *pdata; 168 struct device *dev; 169 const struct attribute_group **groups; 170 }; 171 172 173The "coresight_dev_type" identifies what the device is, i.e, source link or 174sink while the "coresight_dev_subtype" will characterise that type further. 175 176The ``struct coresight_ops`` is mandatory and will tell the framework how to 177perform base operations related to the components, each component having 178a different set of requirement. For that ``struct coresight_ops_sink``, 179``struct coresight_ops_link`` and ``struct coresight_ops_source`` have been 180provided. 181 182The next field ``struct coresight_platform_data *pdata`` is acquired by calling 183``of_get_coresight_platform_data()``, as part of the driver's _probe routine and 184``struct device *dev`` gets the device reference embedded in the ``amba_device``:: 185 186 static int etm_probe(struct amba_device *adev, const struct amba_id *id) 187 { 188 ... 189 ... 190 drvdata->dev = &adev->dev; 191 ... 192 } 193 194Specific class of device (source, link, or sink) have generic operations 195that can be performed on them (see ``struct coresight_ops``). The ``**groups`` 196is a list of sysfs entries pertaining to operations 197specific to that component only. "Implementation defined" customisations are 198expected to be accessed and controlled using those entries. 199 200Device Naming scheme 201-------------------- 202 203The devices that appear on the "coresight" bus were named the same as their 204parent devices, i.e, the real devices that appears on AMBA bus or the platform bus. 205Thus the names were based on the Linux Open Firmware layer naming convention, 206which follows the base physical address of the device followed by the device 207type. e.g:: 208 209 root:~# ls /sys/bus/coresight/devices/ 210 20010000.etf 20040000.funnel 20100000.stm 22040000.etm 211 22140000.etm 230c0000.funnel 23240000.etm 20030000.tpiu 212 20070000.etr 20120000.replicator 220c0000.funnel 213 23040000.etm 23140000.etm 23340000.etm 214 215However, with the introduction of ACPI support, the names of the real 216devices are a bit cryptic and non-obvious. Thus, a new naming scheme was 217introduced to use more generic names based on the type of the device. The 218following rules apply:: 219 220 1) Devices that are bound to CPUs, are named based on the CPU logical 221 number. 222 223 e.g, ETM bound to CPU0 is named "etm0" 224 225 2) All other devices follow a pattern, "<device_type_prefix>N", where : 226 227 <device_type_prefix> - A prefix specific to the type of the device 228 N - a sequential number assigned based on the order 229 of probing. 230 231 e.g, tmc_etf0, tmc_etr0, funnel0, funnel1 232 233Thus, with the new scheme the devices could appear as :: 234 235 root:~# ls /sys/bus/coresight/devices/ 236 etm0 etm1 etm2 etm3 etm4 etm5 funnel0 237 funnel1 funnel2 replicator0 stm0 tmc_etf0 tmc_etr0 tpiu0 238 239Some of the examples below might refer to old naming scheme and some 240to the newer scheme, to give a confirmation that what you see on your 241system is not unexpected. One must use the "names" as they appear on 242the system under specified locations. 243 244Topology Representation 245----------------------- 246 247Each CoreSight component has a ``connections`` directory which will contain 248links to other CoreSight components. This allows the user to explore the trace 249topology and for larger systems, determine the most appropriate sink for a 250given source. The connection information can also be used to establish 251which CTI devices are connected to a given component. This directory contains a 252``nr_links`` attribute detailing the number of links in the directory. 253 254For an ETM source, in this case ``etm0`` on a Juno platform, a typical 255arrangement will be:: 256 257 linaro-developer:~# ls - l /sys/bus/coresight/devices/etm0/connections 258 <file details> cti_cpu0 -> ../../../23020000.cti/cti_cpu0 259 <file details> nr_links 260 <file details> out:0 -> ../../../230c0000.funnel/funnel2 261 262Following the out port to ``funnel2``:: 263 264 linaro-developer:~# ls -l /sys/bus/coresight/devices/funnel2/connections 265 <file details> in:0 -> ../../../23040000.etm/etm0 266 <file details> in:1 -> ../../../23140000.etm/etm3 267 <file details> in:2 -> ../../../23240000.etm/etm4 268 <file details> in:3 -> ../../../23340000.etm/etm5 269 <file details> nr_links 270 <file details> out:0 -> ../../../20040000.funnel/funnel0 271 272And again to ``funnel0``:: 273 274 linaro-developer:~# ls -l /sys/bus/coresight/devices/funnel0/connections 275 <file details> in:0 -> ../../../220c0000.funnel/funnel1 276 <file details> in:1 -> ../../../230c0000.funnel/funnel2 277 <file details> nr_links 278 <file details> out:0 -> ../../../20010000.etf/tmc_etf0 279 280Finding the first sink ``tmc_etf0``. This can be used to collect data 281as a sink, or as a link to propagate further along the chain:: 282 283 linaro-developer:~# ls -l /sys/bus/coresight/devices/tmc_etf0/connections 284 <file details> cti_sys0 -> ../../../20020000.cti/cti_sys0 285 <file details> in:0 -> ../../../20040000.funnel/funnel0 286 <file details> nr_links 287 <file details> out:0 -> ../../../20150000.funnel/funnel4 288 289via ``funnel4``:: 290 291 linaro-developer:~# ls -l /sys/bus/coresight/devices/funnel4/connections 292 <file details> in:0 -> ../../../20010000.etf/tmc_etf0 293 <file details> in:1 -> ../../../20140000.etf/tmc_etf1 294 <file details> nr_links 295 <file details> out:0 -> ../../../20120000.replicator/replicator0 296 297and a ``replicator0``:: 298 299 linaro-developer:~# ls -l /sys/bus/coresight/devices/replicator0/connections 300 <file details> in:0 -> ../../../20150000.funnel/funnel4 301 <file details> nr_links 302 <file details> out:0 -> ../../../20030000.tpiu/tpiu0 303 <file details> out:1 -> ../../../20070000.etr/tmc_etr0 304 305Arriving at the final sink in the chain, ``tmc_etr0``:: 306 307 linaro-developer:~# ls -l /sys/bus/coresight/devices/tmc_etr0/connections 308 <file details> cti_sys0 -> ../../../20020000.cti/cti_sys0 309 <file details> in:0 -> ../../../20120000.replicator/replicator0 310 <file details> nr_links 311 312As described below, when using sysfs it is sufficient to enable a sink and 313a source for successful trace. The framework will correctly enable all 314intermediate links as required. 315 316Note: ``cti_sys0`` appears in two of the connections lists above. 317CTIs can connect to multiple devices and are arranged in a star topology 318via the CTM. See (Documentation/trace/coresight/coresight-ect.rst) 319[#fourth]_ for further details. 320Looking at this device we see 4 connections:: 321 322 linaro-developer:~# ls -l /sys/bus/coresight/devices/cti_sys0/connections 323 <file details> nr_links 324 <file details> stm0 -> ../../../20100000.stm/stm0 325 <file details> tmc_etf0 -> ../../../20010000.etf/tmc_etf0 326 <file details> tmc_etr0 -> ../../../20070000.etr/tmc_etr0 327 <file details> tpiu0 -> ../../../20030000.tpiu/tpiu0 328 329 330How to use the tracer modules 331----------------------------- 332 333There are two ways to use the Coresight framework: 334 3351. using the perf cmd line tools. 3362. interacting directly with the Coresight devices using the sysFS interface. 337 338Preference is given to the former as using the sysFS interface 339requires a deep understanding of the Coresight HW. The following sections 340provide details on using both methods. 341 3421) Using the sysFS interface: 343 344Before trace collection can start, a coresight sink needs to be identified. 345There is no limit on the amount of sinks (nor sources) that can be enabled at 346any given moment. As a generic operation, all device pertaining to the sink 347class will have an "active" entry in sysfs:: 348 349 root:/sys/bus/coresight/devices# ls 350 replicator 20030000.tpiu 2201c000.ptm 2203c000.etm 2203e000.etm 351 20010000.etb 20040000.funnel 2201d000.ptm 2203d000.etm 352 root:/sys/bus/coresight/devices# ls 20010000.etb 353 enable_sink status trigger_cntr 354 root:/sys/bus/coresight/devices# echo 1 > 20010000.etb/enable_sink 355 root:/sys/bus/coresight/devices# cat 20010000.etb/enable_sink 356 1 357 root:/sys/bus/coresight/devices# 358 359At boot time the current etm3x driver will configure the first address 360comparator with "_stext" and "_etext", essentially tracing any instruction 361that falls within that range. As such "enabling" a source will immediately 362trigger a trace capture:: 363 364 root:/sys/bus/coresight/devices# echo 1 > 2201c000.ptm/enable_source 365 root:/sys/bus/coresight/devices# cat 2201c000.ptm/enable_source 366 1 367 root:/sys/bus/coresight/devices# cat 20010000.etb/status 368 Depth: 0x2000 369 Status: 0x1 370 RAM read ptr: 0x0 371 RAM wrt ptr: 0x19d3 <----- The write pointer is moving 372 Trigger cnt: 0x0 373 Control: 0x1 374 Flush status: 0x0 375 Flush ctrl: 0x2001 376 root:/sys/bus/coresight/devices# 377 378Trace collection is stopped the same way:: 379 380 root:/sys/bus/coresight/devices# echo 0 > 2201c000.ptm/enable_source 381 root:/sys/bus/coresight/devices# 382 383The content of the ETB buffer can be harvested directly from /dev:: 384 385 root:/sys/bus/coresight/devices# dd if=/dev/20010000.etb \ 386 of=~/cstrace.bin 387 64+0 records in 388 64+0 records out 389 32768 bytes (33 kB) copied, 0.00125258 s, 26.2 MB/s 390 root:/sys/bus/coresight/devices# 391 392The file cstrace.bin can be decompressed using "ptm2human", DS-5 or Trace32. 393 394Following is a DS-5 output of an experimental loop that increments a variable up 395to a certain value. The example is simple and yet provides a glimpse of the 396wealth of possibilities that coresight provides. 397:: 398 399 Info Tracing enabled 400 Instruction 106378866 0x8026B53C E52DE004 false PUSH {lr} 401 Instruction 0 0x8026B540 E24DD00C false SUB sp,sp,#0xc 402 Instruction 0 0x8026B544 E3A03000 false MOV r3,#0 403 Instruction 0 0x8026B548 E58D3004 false STR r3,[sp,#4] 404 Instruction 0 0x8026B54C E59D3004 false LDR r3,[sp,#4] 405 Instruction 0 0x8026B550 E3530004 false CMP r3,#4 406 Instruction 0 0x8026B554 E2833001 false ADD r3,r3,#1 407 Instruction 0 0x8026B558 E58D3004 false STR r3,[sp,#4] 408 Instruction 0 0x8026B55C DAFFFFFA true BLE {pc}-0x10 ; 0x8026b54c 409 Timestamp Timestamp: 17106715833 410 Instruction 319 0x8026B54C E59D3004 false LDR r3,[sp,#4] 411 Instruction 0 0x8026B550 E3530004 false CMP r3,#4 412 Instruction 0 0x8026B554 E2833001 false ADD r3,r3,#1 413 Instruction 0 0x8026B558 E58D3004 false STR r3,[sp,#4] 414 Instruction 0 0x8026B55C DAFFFFFA true BLE {pc}-0x10 ; 0x8026b54c 415 Instruction 9 0x8026B54C E59D3004 false LDR r3,[sp,#4] 416 Instruction 0 0x8026B550 E3530004 false CMP r3,#4 417 Instruction 0 0x8026B554 E2833001 false ADD r3,r3,#1 418 Instruction 0 0x8026B558 E58D3004 false STR r3,[sp,#4] 419 Instruction 0 0x8026B55C DAFFFFFA true BLE {pc}-0x10 ; 0x8026b54c 420 Instruction 7 0x8026B54C E59D3004 false LDR r3,[sp,#4] 421 Instruction 0 0x8026B550 E3530004 false CMP r3,#4 422 Instruction 0 0x8026B554 E2833001 false ADD r3,r3,#1 423 Instruction 0 0x8026B558 E58D3004 false STR r3,[sp,#4] 424 Instruction 0 0x8026B55C DAFFFFFA true BLE {pc}-0x10 ; 0x8026b54c 425 Instruction 7 0x8026B54C E59D3004 false LDR r3,[sp,#4] 426 Instruction 0 0x8026B550 E3530004 false CMP r3,#4 427 Instruction 0 0x8026B554 E2833001 false ADD r3,r3,#1 428 Instruction 0 0x8026B558 E58D3004 false STR r3,[sp,#4] 429 Instruction 0 0x8026B55C DAFFFFFA true BLE {pc}-0x10 ; 0x8026b54c 430 Instruction 10 0x8026B54C E59D3004 false LDR r3,[sp,#4] 431 Instruction 0 0x8026B550 E3530004 false CMP r3,#4 432 Instruction 0 0x8026B554 E2833001 false ADD r3,r3,#1 433 Instruction 0 0x8026B558 E58D3004 false STR r3,[sp,#4] 434 Instruction 0 0x8026B55C DAFFFFFA true BLE {pc}-0x10 ; 0x8026b54c 435 Instruction 6 0x8026B560 EE1D3F30 false MRC p15,#0x0,r3,c13,c0,#1 436 Instruction 0 0x8026B564 E1A0100D false MOV r1,sp 437 Instruction 0 0x8026B568 E3C12D7F false BIC r2,r1,#0x1fc0 438 Instruction 0 0x8026B56C E3C2203F false BIC r2,r2,#0x3f 439 Instruction 0 0x8026B570 E59D1004 false LDR r1,[sp,#4] 440 Instruction 0 0x8026B574 E59F0010 false LDR r0,[pc,#16] ; [0x8026B58C] = 0x80550368 441 Instruction 0 0x8026B578 E592200C false LDR r2,[r2,#0xc] 442 Instruction 0 0x8026B57C E59221D0 false LDR r2,[r2,#0x1d0] 443 Instruction 0 0x8026B580 EB07A4CF true BL {pc}+0x1e9344 ; 0x804548c4 444 Info Tracing enabled 445 Instruction 13570831 0x8026B584 E28DD00C false ADD sp,sp,#0xc 446 Instruction 0 0x8026B588 E8BD8000 true LDM sp!,{pc} 447 Timestamp Timestamp: 17107041535 448 4492) Using perf framework: 450 451Coresight tracers are represented using the Perf framework's Performance 452Monitoring Unit (PMU) abstraction. As such the perf framework takes charge of 453controlling when tracing gets enabled based on when the process of interest is 454scheduled. When configured in a system, Coresight PMUs will be listed when 455queried by the perf command line tool: 456 457 linaro@linaro-nano:~$ ./perf list pmu 458 459 List of pre-defined events (to be used in -e): 460 461 cs_etm// [Kernel PMU event] 462 463 linaro@linaro-nano:~$ 464 465Regardless of the number of tracers available in a system (usually equal to the 466amount of processor cores), the "cs_etm" PMU will be listed only once. 467 468A Coresight PMU works the same way as any other PMU, i.e the name of the PMU is 469listed along with configuration options within forward slashes '/'. Since a 470Coresight system will typically have more than one sink, the name of the sink to 471work with needs to be specified as an event option. 472On newer kernels the available sinks are listed in sysFS under 473($SYSFS)/bus/event_source/devices/cs_etm/sinks/:: 474 475 root@localhost:/sys/bus/event_source/devices/cs_etm/sinks# ls 476 tmc_etf0 tmc_etr0 tpiu0 477 478On older kernels, this may need to be found from the list of coresight devices, 479available under ($SYSFS)/bus/coresight/devices/:: 480 481 root:~# ls /sys/bus/coresight/devices/ 482 etm0 etm1 etm2 etm3 etm4 etm5 funnel0 483 funnel1 funnel2 replicator0 stm0 tmc_etf0 tmc_etr0 tpiu0 484 root@linaro-nano:~# perf record -e cs_etm/@tmc_etr0/u --per-thread program 485 486As mentioned above in section "Device Naming scheme", the names of the devices could 487look different from what is used in the example above. One must use the device names 488as it appears under the sysFS. 489 490The syntax within the forward slashes '/' is important. The '@' character 491tells the parser that a sink is about to be specified and that this is the sink 492to use for the trace session. 493 494More information on the above and other example on how to use Coresight with 495the perf tools can be found in the "HOWTO.md" file of the openCSD gitHub 496repository [#third]_. 497 4982.1) AutoFDO analysis using the perf tools: 499 500perf can be used to record and analyze trace of programs. 501 502Execution can be recorded using 'perf record' with the cs_etm event, 503specifying the name of the sink to record to, e.g:: 504 505 perf record -e cs_etm/@tmc_etr0/u --per-thread 506 507The 'perf report' and 'perf script' commands can be used to analyze execution, 508synthesizing instruction and branch events from the instruction trace. 509'perf inject' can be used to replace the trace data with the synthesized events. 510The --itrace option controls the type and frequency of synthesized events 511(see perf documentation). 512 513Note that only 64-bit programs are currently supported - further work is 514required to support instruction decode of 32-bit Arm programs. 515 5162.2) Tracing PID 517 518The kernel can be built to write the PID value into the PE ContextID registers. 519For a kernel running at EL1, the PID is stored in CONTEXTIDR_EL1. A PE may 520implement Arm Virtualization Host Extensions (VHE), which the kernel can 521run at EL2 as a virtualisation host; in this case, the PID value is stored in 522CONTEXTIDR_EL2. 523 524perf provides PMU formats that program the ETM to insert these values into the 525trace data; the PMU formats are defined as below: 526 527 "contextid1": Available on both EL1 kernel and EL2 kernel. When the 528 kernel is running at EL1, "contextid1" enables the PID 529 tracing; when the kernel is running at EL2, this enables 530 tracing the PID of guest applications. 531 532 "contextid2": Only usable when the kernel is running at EL2. When 533 selected, enables PID tracing on EL2 kernel. 534 535 "contextid": Will be an alias for the option that enables PID 536 tracing. I.e, 537 contextid == contextid1, on EL1 kernel. 538 contextid == contextid2, on EL2 kernel. 539 540perf will always enable PID tracing at the relevant EL, this is accomplished by 541automatically enable the "contextid" config - but for EL2 it is possible to make 542specific adjustments using configs "contextid1" and "contextid2", E.g. if a user 543wants to trace PIDs for both host and guest, the two configs "contextid1" and 544"contextid2" can be set at the same time: 545 546 perf record -e cs_etm/contextid1,contextid2/u -- vm 547 548 549Generating coverage files for Feedback Directed Optimization: AutoFDO 550--------------------------------------------------------------------- 551 552'perf inject' accepts the --itrace option in which case tracing data is 553removed and replaced with the synthesized events. e.g. 554:: 555 556 perf inject --itrace --strip -i perf.data -o perf.data.new 557 558Below is an example of using ARM ETM for autoFDO. It requires autofdo 559(https://github.com/google/autofdo) and gcc version 5. The bubble 560sort example is from the AutoFDO tutorial (https://gcc.gnu.org/wiki/AutoFDO/Tutorial). 561:: 562 563 $ gcc-5 -O3 sort.c -o sort 564 $ taskset -c 2 ./sort 565 Bubble sorting array of 30000 elements 566 5910 ms 567 568 $ perf record -e cs_etm/@tmc_etr0/u --per-thread taskset -c 2 ./sort 569 Bubble sorting array of 30000 elements 570 12543 ms 571 [ perf record: Woken up 35 times to write data ] 572 [ perf record: Captured and wrote 69.640 MB perf.data ] 573 574 $ perf inject -i perf.data -o inj.data --itrace=il64 --strip 575 $ create_gcov --binary=./sort --profile=inj.data --gcov=sort.gcov -gcov_version=1 576 $ gcc-5 -O3 -fauto-profile=sort.gcov sort.c -o sort_autofdo 577 $ taskset -c 2 ./sort_autofdo 578 Bubble sorting array of 30000 elements 579 5806 ms 580 581 582How to use the STM module 583------------------------- 584 585Using the System Trace Macrocell module is the same as the tracers - the only 586difference is that clients are driving the trace capture rather 587than the program flow through the code. 588 589As with any other CoreSight component, specifics about the STM tracer can be 590found in sysfs with more information on each entry being found in [#first]_:: 591 592 root@genericarmv8:~# ls /sys/bus/coresight/devices/stm0 593 enable_source hwevent_select port_enable subsystem uevent 594 hwevent_enable mgmt port_select traceid 595 root@genericarmv8:~# 596 597Like any other source a sink needs to be identified and the STM enabled before 598being used:: 599 600 root@genericarmv8:~# echo 1 > /sys/bus/coresight/devices/tmc_etf0/enable_sink 601 root@genericarmv8:~# echo 1 > /sys/bus/coresight/devices/stm0/enable_source 602 603From there user space applications can request and use channels using the devfs 604interface provided for that purpose by the generic STM API:: 605 606 root@genericarmv8:~# ls -l /dev/stm0 607 crw------- 1 root root 10, 61 Jan 3 18:11 /dev/stm0 608 root@genericarmv8:~# 609 610Details on how to use the generic STM API can be found here: 611- Documentation/trace/stm.rst [#second]_. 612 613The CTI & CTM Modules 614--------------------- 615 616The CTI (Cross Trigger Interface) provides a set of trigger signals between 617individual CTIs and components, and can propagate these between all CTIs via 618channels on the CTM (Cross Trigger Matrix). 619 620A separate documentation file is provided to explain the use of these devices. 621(Documentation/trace/coresight/coresight-ect.rst) [#fourth]_. 622 623CoreSight System Configuration 624------------------------------ 625 626CoreSight components can be complex devices with many programming options. 627Furthermore, components can be programmed to interact with each other across the 628complete system. 629 630A CoreSight System Configuration manager is provided to allow these complex programming 631configurations to be selected and used easily from perf and sysfs. 632 633See the separate document for further information. 634(Documentation/trace/coresight/coresight-config.rst) [#fifth]_. 635 636 637.. [#first] Documentation/ABI/testing/sysfs-bus-coresight-devices-stm 638 639.. [#second] Documentation/trace/stm.rst 640 641.. [#third] https://github.com/Linaro/perf-opencsd 642 643.. [#fourth] Documentation/trace/coresight/coresight-ect.rst 644 645.. [#fifth] Documentation/trace/coresight/coresight-config.rst