len0rd/FreeRTOS-Kernel
1
0
Fork 0
mirror of https://github.com/FreeRTOS/FreeRTOS-Kernel.git synced 2025-08-19 09:38:32 -04:00
Code Issues Projects Releases Packages Wiki Activity

Update the MSVC and MingW demos to use the latest FreeRTOS+Trace recorder library.

Browse source
This commit is contained in:
Richard Barry 2017-05-29 21:45:31 +00:00
parent f289bfb388
commit 35f5990e7a
13 changed files with 1194 additions and 919 deletions
Download patch file Download diff file Expand all files Collapse all files

520
FreeRTOS/Demo/WIN32-MingW/Trace_Recorder_Configuration/trcConfig.h
View file

@ -1,475 +1,159 @@
/*******************************************************************************
* Tracealyzer v2.7.0 Recorder Library
/*******************************************************************************
* Trace Recorder Library for Tracealyzer v3.1.2
* Percepio AB, www.percepio.com
*
* trcConfig.h
*
* Configuration parameters for the trace recorder library. Before using the
* trace recorder library, please check that the default settings are
* appropriate for your system, and if necessary adjust these. Most likely, you
* will need to adjust the NTask, NISR, NQueue, NMutex and NSemaphore values to
* reflect the number of such objects in your system. These may be
* over-approximated, although larger values values implies more RAM usage.
* Main configuration parameters for the trace recorder library.
* More settings can be found in trcStreamingConfig.h and trcSnapshotConfig.h.
*
* Read more at http://percepio.com/2016/10/05/rtos-tracing/
*
* Terms of Use
* This software is copyright Percepio AB. The recorder library is free for
* use together with Percepio products. You may distribute the recorder library
* in its original form, including modifications in trcHardwarePort.c/.h
* given that these modification are clearly marked as your own modifications
* and documented in the initial comment section of these source files.
* This software is the intellectual property of Percepio AB and may not be
* sold or in other ways commercially redistributed without explicit written
* permission by Percepio AB.
* This file is part of the trace recorder library (RECORDER), which is the
* intellectual property of Percepio AB (PERCEPIO) and provided under a
* license as follows.
* The RECORDER may be used free of charge for the purpose of recording data
* intended for analysis in PERCEPIO products. It may not be used or modified
* for other purposes without explicit permission from PERCEPIO.
* You may distribute the RECORDER in its original source code form, assuming
* this text (terms of use, disclaimer, copyright notice) is unchanged. You are
* allowed to distribute the RECORDER with minor modifications intended for
* configuration or porting of the RECORDER, e.g., to allow using it on a
* specific processor, processor family or with a specific communication
* interface. Any such modifications should be documented directly below
* this comment block.
*
* Disclaimer
* The trace tool and recorder library is being delivered to you AS IS and
* Percepio AB makes no warranty as to its use or performance. Percepio AB does
* not and cannot warrant the performance or results you may obtain by using the
* software or documentation. Percepio AB make no warranties, express or
* implied, as to noninfringement of third party rights, merchantability, or
* fitness for any particular purpose. In no event will Percepio AB, its
* technology partners, or distributors be liable to you for any consequential,
* incidental or special damages, including any lost profits or lost savings,
* even if a representative of Percepio AB has been advised of the possibility
* of such damages, or for any claim by any third party. Some jurisdictions do
* not allow the exclusion or limitation of incidental, consequential or special
* damages, or the exclusion of implied warranties or limitations on how long an
* implied warranty may last, so the above limitations may not apply to you.
* The RECORDER is being delivered to you AS IS and PERCEPIO makes no warranty
* as to its use or performance. PERCEPIO does not and cannot warrant the
* performance or results you may obtain by using the RECORDER or documentation.
* PERCEPIO make no warranties, express or implied, as to noninfringement of
* third party rights, merchantability, or fitness for any particular purpose.
* In no event will PERCEPIO, its technology partners, or distributors be liable
* to you for any consequential, incidental or special damages, including any
* lost profits or lost savings, even if a representative of PERCEPIO has been
* advised of the possibility of such damages, or for any claim by any third
* party. Some jurisdictions do not allow the exclusion or limitation of
* incidental, consequential or special damages, or the exclusion of implied
* warranties or limitations on how long an implied warranty may last, so the
* above limitations may not apply to you.
*
* Tabs are used for indent in this file (1 tab = 4 spaces)
*
* Copyright Percepio AB, 2014.
* Copyright Percepio AB, 2016.
* www.percepio.com
******************************************************************************/
#ifndef TRCCONFIG_H
#define TRCCONFIG_H
#ifndef TRC_CONFIG_H
#define TRC_CONFIG_H
/******************************************************************************
* SELECTED_PORT
*
* Macro that specifies what hardware port that should be used.
* Available ports are:
*
* Port Name Code Official OS supported
* PORT_APPLICATION_DEFINED -2 - -
* PORT_NOT_SET -1 - -
* PORT_HWIndependent 0 Yes Any
* PORT_Win32 1 Yes FreeRTOS on Win32
* PORT_Atmel_AT91SAM7 2 No Any
* PORT_Atmel_UC3A0 3 No Any
* PORT_ARM_CortexM 4 Yes Any
* PORT_Renesas_RX600 5 Yes Any
* PORT_Microchip_dsPIC_AND_PIC24 6 Yes Any
* PORT_TEXAS_INSTRUMENTS_TMS570 7 No Any
* PORT_TEXAS_INSTRUMENTS_MSP430 8 No Any
* PORT_MICROCHIP_PIC32MX 9 Yes Any
* PORT_XILINX_PPC405 10 No FreeRTOS
* PORT_XILINX_PPC440 11 No FreeRTOS
* PORT_XILINX_MICROBLAZE 12 No Any
* PORT_NXP_LPC210X 13 No Any
* PORT_MICROCHIP_PIC32MZ 14 Yes Any
* PORT_ARM_CORTEX_A9 15 No Any
*****************************************************************************/
#ifndef WIN32
// Set the port setting here!
#define SELECTED_PORT PORT_NOT_SET
#if (SELECTED_PORT == PORT_NOT_SET)
#error "You need to define SELECTED_PORT here!"
#endif
#else
// For Win32 demo projects this is set automatically
#define SELECTED_PORT PORT_Win32
#ifdef __cplusplus
extern "C" {
#endif
/******************************************************************************
* FREERTOS_VERSION
*
* Specify what version of FreeRTOS that is used. This is necessary compensate
* for renamed symbols in the FreeRTOS kernel (does not build if incorrect).
*
* FREERTOS_VERSION_7_3_OR_7_4 (= 1) If using FreeRTOS v7.3.0 - v7.4.2
* FREERTOS_VERSION_7_5_OR_7_6 (= 2) If using FreeRTOS v7.5.0 - v7.6.0
* FREERTOS_VERSION_8_0_OR_LATER (= 3) If using FreeRTOS v8.0.0 or later
*****************************************************************************/
#define FREERTOS_VERSION FREERTOS_VERSION_8_0_OR_LATER
#include "trcPortDefines.h"
/******************************************************************************
* TRACE_RECORDER_STORE_MODE
* Include of processor header file
*
* Macro which should be defined as one of:
* - TRACE_STORE_MODE_RING_BUFFER
* - TRACE_STORE_MODE_STOP_WHEN_FULL
* Default is TRACE_STORE_MODE_RING_BUFFER.
*
* With TRACE_RECORDER_STORE_MODE set to TRACE_STORE_MODE_RING_BUFFER, the
* events are stored in a ring buffer, i.e., where the oldest events are
* overwritten when the buffer becomes full. This allows you to get the last
* events leading up to an interesting state, e.g., an error, without having
* to store the whole run since startup.
*
* When TRACE_RECORDER_STORE_MODE is TRACE_STORE_MODE_STOP_WHEN_FULL, the
* recording is stopped when the buffer becomes full. This is useful for
* recording events following a specific state, e.g., the startup sequence.
* Here you may need to include the header file for your processor. This is
* required at least for the ARM Cortex-M port, that uses the ARM CMSIS API.
* Try that in case of build problems. Otherwise, remove the #error line below.
*****************************************************************************/
#define TRACE_RECORDER_STORE_MODE TRACE_STORE_MODE_RING_BUFFER
//#error "Trace Recorder: Please include your processor's header file here and remove this line."
/*******************************************************************************
* TRACE_SCHEDULING_ONLY
* Configuration Macro: TRC_CFG_HARDWARE_PORT
*
* Macro which should be defined as an integer value.
* Specify what hardware port to use (i.e., the "timestamping driver").
*
* If this setting is enabled (= 1), only scheduling events are recorded.
* If disabled (= 0), all events are recorded.
* All ARM Cortex-M MCUs are supported by "TRC_HARDWARE_PORT_ARM_Cortex_M".
* This port uses the DWT cycle counter for Cortex-M3/M4/M7 devices, which is
* available on most such devices. In case your device don't have DWT support,
* you will get an error message opening the trace. In that case, you may
* force the recorder to use SysTick timestamping instead, using this define:
*
* Users of FreeRTOS+Trace Free Edition only displays scheduling events, so this
* option can be used to avoid storing unsupported events.
* #define TRC_CFG_ARM_CM_USE_SYSTICK
*
* Default value is 0 (store all enabled events).
* For ARM Cortex-M0/M0+ devices, SysTick mode is used automatically.
*
* See trcHardwarePort.h for available ports and information on how to
* define your own port, if not already present.
******************************************************************************/
#define TRACE_SCHEDULING_ONLY 0
#define TRC_CFG_HARDWARE_PORT TRC_HARDWARE_PORT_Win32
/*******************************************************************************
* EVENT_BUFFER_SIZE
* Configuration Macro: TRC_CFG_RECORDER_MODE
*
* Macro which should be defined as an integer value.
* Specify what recording mode to use. Snapshot means that the data is saved in
* an internal RAM buffer, for later upload. Streaming means that the data is
* transferred continuously to the host PC.
*
* This defines the capacity of the event buffer, i.e., the number of records
* it may store. Most events use one record (4 byte), although some events
* require multiple 4-byte records. You should adjust this to the amount of RAM
* available in the target system.
* For more information, see http://percepio.com/2016/10/05/rtos-tracing/
* and the Tracealyzer User Manual.
*
* Default value is 1000, which means that 4000 bytes is allocated for the
* event buffer.
* Values:
* TRC_RECORDER_MODE_SNAPSHOT
* TRC_RECORDER_MODE_STREAMING
******************************************************************************/
#define EVENT_BUFFER_SIZE 15000
#define TRC_CFG_RECORDER_MODE TRC_RECORDER_MODE_SNAPSHOT
/*******************************************************************************
* NTask, NISR, NQueue, NSemaphore, NMutex
* Configuration Macro: TRC_CFG_RECORDER_BUFFER_ALLOCATION
*
* A group of macros which should be defined as integer values, zero or larger.
* Specifies how the recorder buffer is allocated (also in case of streaming, in
* port using the recorder's internal temporary buffer)
*
* These define the capacity of the Object Property Table, i.e., the maximum
* number of objects active at any given point, within each object class (e.g.,
* task, queue, semaphore, ...).
* Values:
* TRC_RECORDER_BUFFER_ALLOCATION_STATIC - Static allocation (internal)
* TRC_RECORDER_BUFFER_ALLOCATION_DYNAMIC - Malloc in vTraceEnable
* TRC_RECORDER_BUFFER_ALLOCATION_CUSTOM - Use vTraceSetRecorderDataBuffer
*
* If tasks or other other objects are deleted in your system, this
* setting does not limit the total amount of objects created, only the number
* of objects that have been successfully created but not yet deleted.
*
* Using too small values will cause vTraceError to be called, which stores an
* error message in the trace that is shown when opening the trace file.
*
* It can be wise to start with large values for these constants,
* unless you are very confident on these numbers. Then do a recording and
* check the actual usage by selecting View menu -> Trace Details ->
* Resource Usage -> Object Table.
* Static and dynamic mode does the allocation for you, either in compile time
* (static) or in runtime (malloc).
* The custom mode allows you to control how and where the allocation is made,
* for details see TRC_ALLOC_CUSTOM_BUFFER and vTraceSetRecorderDataBuffer().
******************************************************************************/
#define NTask 100
#define NISR 60
#define NQueue 60
#define NSemaphore 60
#define NMutex 60
#define NTimer 200
#define NEventGroup 60
#define TRC_CFG_RECORDER_BUFFER_ALLOCATION TRC_RECORDER_BUFFER_ALLOCATION_STATIC
/******************************************************************************
* INCLUDE_MEMMANG_EVENTS
* TRC_CFG_FREERTOS_VERSION
*
* Macro which should be defined as either zero (0) or one (1).
* Specify what version of FreeRTOS that is used (don't change unless using the
* trace recorder library with an older version of FreeRTOS).
*
* This controls if malloc and free calls should be traced. Set this to zero to
* exclude malloc/free calls, or one (1) to include such events in the trace.
*
* Default value is 1.
* TRC_FREERTOS_VERSION_7_3_OR_7_4 If using FreeRTOS v7.3.0 - v7.4.2
* TRC_FREERTOS_VERSION_7_5_OR_7_6 If using FreeRTOS v7.5.0 - v7.6.0
* TRC_FREERTOS_VERSION_8_X If using FreeRTOS v8.X.X
* TRC_FREERTOS_VERSION_9_X If using FreeRTOS v9.X.X
*****************************************************************************/
#define INCLUDE_MEMMANG_EVENTS 1
#define TRC_CFG_FREERTOS_VERSION TRC_FREERTOS_VERSION_9_X
/******************************************************************************
* INCLUDE_USER_EVENTS
* TRC_CFG_MAX_ISR_NESTING
*
* Macro which should be defined as either zero (0) or one (1).
* Defines how many levels of interrupt nesting the recorder can handle, in
* case multiple ISRs are traced and ISR nesting is possible. If this
* is exceeded, the particular ISR will not be traced and the recorder then
* logs an error message. This setting is used to allocate an internal stack
* for keeping track of the previous execution context (4 byte per entry).
*
* If this is zero (0) the code for creating User Events is excluded to
* reduce code size. User Events are application-generated events, like
* "printf" but for the trace log instead of console output. User Events are
* much faster than a printf and can therefore be used in timing critical code.
* See vTraceUserEvent() and vTracePrintF() in trcUser.h
* This value must be a non-zero positive constant, at least 1.
*
* Default value is 1.
*
* Note that User Events are only displayed in Professional Edition.
* Default value: 8
*****************************************************************************/
#define INCLUDE_USER_EVENTS 1
#define TRC_CFG_MAX_ISR_NESTING 8
/*****************************************************************************
* INCLUDE_ISR_TRACING
*
* Macro which should be defined as either zero (0) or one (1).
*
* If this is zero (0), the code for recording Interrupt Service Routines is
* excluded to reduce code size.
*
* Default value is 1.
*
* Note, if the kernel has no central interrupt dispatcher, recording ISRs
* require that you insert calls to vTraceStoreISRBegin and vTraceStoreISREnd
* in your interrupt handlers.
*****************************************************************************/
#define INCLUDE_ISR_TRACING 1
/*****************************************************************************
* INCLUDE_READY_EVENTS
*
* Macro which should be defined as either zero (0) or one (1).
*
* If one (1), events are recorded when tasks enter scheduling state "ready".
* This uses a lot of space in the event buffer, so excluding "ready events"
* will allow for longer traces. Including ready events however allows for
* showing the initial pending time before tasks enter the execution state, and
* for presenting accurate response times.
*
* Default value is 1.
*****************************************************************************/
#define INCLUDE_READY_EVENTS 1
/*****************************************************************************
* INCLUDE_NEW_TIME_EVENTS
*
* Macro which should be defined as either zero (0) or one (1).
*
* If this is zero (1), events will be generated whenever the OS clock is
* increased.
*
* Default value is 0.
*****************************************************************************/
#define INCLUDE_NEW_TIME_EVENTS 1
/******************************************************************************
* INCLUDE_FLOAT_SUPPORT
*
* Macro which should be defined as either zero (0) or one (1).
*
* If this is zero (0), all references to floating point values are removed,
* in case floating point values are not supported by the platform used.
* Floating point values are only used in vTracePrintF and its subroutines, to
* store float (%f) or double (%lf) arguments.
*
* vTracePrintF can be used with integer and string arguments in either case.
*
* Default value is 1.
*****************************************************************************/
#define INCLUDE_FLOAT_SUPPORT 0
/******************************************************************************
* INCLUDE_OBJECT_DELETE
*
* Macro which should be defined as either zero (0) or one (1).
*
* This must be enabled (1) if tasks, queues or other
* traced kernel objects are deleted at runtime. If no deletes are made, this
* can be set to 0 in order to exclude the delete-handling code.
*
* Default value is 1.
*****************************************************************************/
#define INCLUDE_OBJECT_DELETE 1
/*******************************************************************************
* SYMBOL_TABLE_SIZE
*
* Macro which should be defined as an integer value.
*
* This defines the capacity of the symbol table, in bytes. This symbol table
* stores User Events labels and names of deleted tasks, queues, or other kernel
* objects. If you don't use User Events or delete any kernel
* objects you set this to a very low value. The minimum recommended value is 4.
* A size of zero (0) is not allowed since a zero-sized array may result in a
* 32-bit pointer, i.e., using 4 bytes rather than 0.
*
* Default value is 800.
******************************************************************************/
#define SYMBOL_TABLE_SIZE 5000
#if (SYMBOL_TABLE_SIZE == 0)
#error "SYMBOL_TABLE_SIZE may not be zero!"
/* Specific configuration, depending on Streaming/Snapshot mode */
#if (TRC_CFG_RECORDER_MODE == TRC_RECORDER_MODE_SNAPSHOT)
#include "trcSnapshotConfig.h"
#elif (TRC_CFG_RECORDER_MODE == TRC_RECORDER_MODE_STREAMING)
#include "trcStreamingConfig.h"
#endif
/******************************************************************************
* NameLenTask, NameLenQueue, ...
*
* Macros that specify the maximum lengths (number of characters) for names of
* kernel objects, such as tasks and queues. If longer names are used, they will
* be truncated when stored in the recorder.
*****************************************************************************/
#define NameLenTask 15
#define NameLenISR 15
#define NameLenQueue 15
#define NameLenSemaphore 15
#define NameLenMutex 15
#define NameLenTimer 15
#define NameLenEventGroup 15
/******************************************************************************
* TRACE_DATA_ALLOCATION
*
* This defines how to allocate the recorder data structure, i.e., using a
* static declaration or using a dynamic allocation in runtime (malloc).
*
* Should be one of these two options:
* - TRACE_DATA_ALLOCATION_STATIC (default)
* - TRACE_DATA_ALLOCATION_DYNAMIC
*
* Using static allocation has the benefits of compile-time errors if the buffer
* is too large (too large constants in trcConfig.h) and no need to call the
* initialization routine (xTraceInitTraceData).
*
* Using dynamic allocation may give more flexibility in some cases.
*****************************************************************************/
#define TRACE_DATA_ALLOCATION TRACE_DATA_ALLOCATION_STATIC
/******************************************************************************
*** ADVANCED SETTINGS ********************************************************
******************************************************************************
* The remaining settings are not necessary to modify but allows for optimizing
* the recorder setup for your specific needs, e.g., to exclude events that you
* are not interested in, in order to get longer traces.
*****************************************************************************/
/******************************************************************************
* HEAP_SIZE_BELOW_16M
*
* An integer constant that can be used to reduce the buffer usage of memory
* allocation events (malloc/free). This value should be 1 if the heap size is
* below 16 MB (2^24 byte), and you can live with reported addresses showing the
* lower 24 bits only. If 0, you get the full 32-bit addresses.
*
* Default value is 0.
******************************************************************************/
#define HEAP_SIZE_BELOW_16M 0
/******************************************************************************
* USE_LINKER_PRAGMA
*
* Macro which should be defined as an integer value, default is 0.
*
* If this is 1, the header file "recorderdata_linker_pragma.h" is included just
* before the declaration of RecorderData (in trcBase.c), i.e., the trace data
* structure. This allows the user to specify a pragma with linker options.
*
* Example (for IAR Embedded Workbench and NXP LPC17xx):
* #pragma location="AHB_RAM_MEMORY"
*
* This example instructs the IAR linker to place RecorderData in another RAM
* bank, the AHB RAM. This can also be used for other compilers with a similar
* pragmas for linker options.
*
* Note that this only applies if using static allocation, see below.
******************************************************************************/
#define USE_LINKER_PRAGMA 0
/******************************************************************************
* USE_IMPLICIT_IFE_RULES
*
* Macro which should be defined as either zero (0) or one (1).
* Default is 1.
*
* Tracealyzer groups the events into actor instances, based on context-switches
* and a definition of "Instance Finish Events", or IFEs. These are kernel calls
* considered to be the last event in a task instance. Some kernel calls are
* considered IFEs by default (e.g., delay functions), but it is also possible
* to specify this individually for each task (see vTraceTaskInstanceFinish).
*
* If USE_IMPLICIT_IFE_RULES is one (1), the default IFEs will be enabled, which
* gives a "typical" grouping of events into instances. You can combine this
* with calls to vTraceTaskInstanceFinish for specific tasks.
*
* If USE_IMPLICIT_IFE_RULES is zero (0), the implicit IFEs are disabled and all
* events withing each task is then shown as a single instance, unless you call
* vTraceTaskInstanceFinish() at suitable locations to mark the IFEs.
*****************************************************************************/
#define USE_IMPLICIT_IFE_RULES 1
/******************************************************************************
* USE_16BIT_OBJECT_HANDLES
*
* Macro which should be defined as either zero (0) or one (1).
*
* If set to 0 (zero), the recorder uses 8-bit handles to identify kernel
* objects such as tasks and queues. This limits the supported number of
* concurrently active objects to 255 of each type (object class).
*
* If set to 1 (one), the recorder uses 16-bit handles to identify kernel
* objects such as tasks and queues. This limits the supported number of
* concurrent objects to 65535 of each type (object class). However, since the
* object property table is limited to 64 KB, the practical limit is about
* 3000 objects in total.
*
* Default is 0.
*
* NOTE: An object with handle above 255 will use an extra 4-byte record in
* the event buffer whenever referenced. Moreover, some internal tables in the
* recorder gets larger when using 16-bit handles. The additional RAM usage is
* 5-10 byte plus 1 byte per kernel object i.e., task, queue, mutex, etc.
*****************************************************************************/
#define USE_16BIT_OBJECT_HANDLES 0
/******************************************************************************
* USE_TRACE_ASSERT
*
* Macro which should be defined as either zero (0) or one (1).
* Default is 1.
*
* If this is one (1), the TRACE_ASSERT macro will verify that a condition is
* true. If the condition is false, vTraceError() will be called.
* This is used on several places in the recorder code for sanity checks on
* parameters. Can be switched off to reduce CPU usage of the tracing.
*****************************************************************************/
#define USE_TRACE_ASSERT 1
/*******************************************************************************
* USE_SEPARATE_USER_EVENT_BUFFER
*
* Macro which should be defined as an integer value.
* Default is zero (0).
*
* This enables and disables the use of the separate user event buffer. Using
* this separate buffer has the benefit of not overwriting the user events with
* kernel events (usually generated at a much higher rate), i.e., when using
* ring-buffer mode.
*
* Note: When using the separate user event buffer, you may get an artificial
* task instance named "Unknown actor". This is added as a placeholder when the
* user event history is longer than the task scheduling history.
******************************************************************************/
#define USE_SEPARATE_USER_EVENT_BUFFER 0
/*******************************************************************************
* USER_EVENT_BUFFER_SIZE
*
* Macro which should be defined as an integer value.
*
* This defines the capacity of the user event buffer, in number of slots.
* A single user event can use between 1 and X slots, depending on the data.
*
* Only in use if USE_SEPARATE_USER_EVENT_BUFFER is set to 1.
******************************************************************************/
#define USER_EVENT_BUFFER_SIZE 10
/*******************************************************************************
* USER_EVENT_CHANNELS
*
* Macro which should be defined as an integer value.
*
* This defines the number of allowed user event channels.
*
* Only in use if USE_SEPARATE_USER_EVENT_BUFFER is set to 1.
******************************************************************************/
#define CHANNEL_FORMAT_PAIRS 32
#ifdef __cplusplus
}
#endif
#endif /* _TRC_CONFIG_H */

472
FreeRTOS/Demo/WIN32-MingW/Trace_Recorder_Configuration/trcSnapshotConfig.h Normal file
View file

@ -0,0 +1,472 @@
/*******************************************************************************
* Trace Recorder Library for Tracealyzer v3.1.2
* Percepio AB, www.percepio.com
*
* trcSnapshotConfig.h
*
* Configuration parameters for trace recorder library in snapshot mode.
* Read more at http://percepio.com/2016/10/05/rtos-tracing/
*
* Terms of Use
* This file is part of the trace recorder library (RECORDER), which is the
* intellectual property of Percepio AB (PERCEPIO) and provided under a
* license as follows.
* The RECORDER may be used free of charge for the purpose of recording data
* intended for analysis in PERCEPIO products. It may not be used or modified
* for other purposes without explicit permission from PERCEPIO.
* You may distribute the RECORDER in its original source code form, assuming
* this text (terms of use, disclaimer, copyright notice) is unchanged. You are
* allowed to distribute the RECORDER with minor modifications intended for
* configuration or porting of the RECORDER, e.g., to allow using it on a
* specific processor, processor family or with a specific communication
* interface. Any such modifications should be documented directly below
* this comment block.
*
* Disclaimer
* The RECORDER is being delivered to you AS IS and PERCEPIO makes no warranty
* as to its use or performance. PERCEPIO does not and cannot warrant the
* performance or results you may obtain by using the RECORDER or documentation.
* PERCEPIO make no warranties, express or implied, as to noninfringement of
* third party rights, merchantability, or fitness for any particular purpose.
* In no event will PERCEPIO, its technology partners, or distributors be liable
* to you for any consequential, incidental or special damages, including any
* lost profits or lost savings, even if a representative of PERCEPIO has been
* advised of the possibility of such damages, or for any claim by any third
* party. Some jurisdictions do not allow the exclusion or limitation of
* incidental, consequential or special damages, or the exclusion of implied
* warranties or limitations on how long an implied warranty may last, so the
* above limitations may not apply to you.
*
* Tabs are used for indent in this file (1 tab = 4 spaces)
*
* Copyright Percepio AB, 2017.
* www.percepio.com
******************************************************************************/
#ifndef TRC_SNAPSHOT_CONFIG_H
#define TRC_SNAPSHOT_CONFIG_H
#define TRC_SNAPSHOT_MODE_RING_BUFFER (0x01)
#define TRC_SNAPSHOT_MODE_STOP_WHEN_FULL (0x02)
/******************************************************************************
* TRC_CFG_SNAPSHOT_MODE
*
* Macro which should be defined as one of:
* - TRC_SNAPSHOT_MODE_RING_BUFFER
* - TRC_SNAPSHOT_MODE_STOP_WHEN_FULL
* Default is TRC_SNAPSHOT_MODE_RING_BUFFER.
*
* With TRC_CFG_SNAPSHOT_MODE set to TRC_SNAPSHOT_MODE_RING_BUFFER, the
* events are stored in a ring buffer, i.e., where the oldest events are
* overwritten when the buffer becomes full. This allows you to get the last
* events leading up to an interesting state, e.g., an error, without having
* to store the whole run since startup.
*
* When TRC_CFG_SNAPSHOT_MODE is TRC_SNAPSHOT_MODE_STOP_WHEN_FULL, the
* recording is stopped when the buffer becomes full. This is useful for
* recording events following a specific state, e.g., the startup sequence.
*****************************************************************************/
#define TRC_CFG_SNAPSHOT_MODE TRC_SNAPSHOT_MODE_RING_BUFFER
/*******************************************************************************
* TRC_CFG_SCHEDULING_ONLY
*
* Macro which should be defined as an integer value.
*
* If this setting is enabled (= 1), only scheduling events are recorded.
* If disabled (= 0), all events are recorded.
*
* For users of Tracealyzer Free Edition, that only displays scheduling events, this
* option can be used to avoid storing other events.
*
* Default value is 0 (store all enabled events).
*
******************************************************************************/
#define TRC_CFG_SCHEDULING_ONLY 0
/*******************************************************************************
* TRC_CFG_EVENT_BUFFER_SIZE
*
* Macro which should be defined as an integer value.
*
* This defines the capacity of the event buffer, i.e., the number of records
* it may store. Most events use one record (4 byte), although some events
* require multiple 4-byte records. You should adjust this to the amount of RAM
* available in the target system.
*
* Default value is 1000, which means that 4000 bytes is allocated for the
* event buffer.
******************************************************************************/
#define TRC_CFG_EVENT_BUFFER_SIZE 15000
/*******************************************************************************
* TRC_CFG_NTASK, TRC_CFG_NISR, TRC_CFG_NQUEUE, TRC_CFG_NSEMAPHORE...
*
* A group of macros which should be defined as integer values, zero or larger.
*
* These define the capacity of the Object Property Table, i.e., the maximum
* number of objects active at any given point, within each object class (e.g.,
* task, queue, semaphore, ...).
*
* If tasks or other objects are deleted in your system, this
* setting does not limit the total amount of objects created, only the number
* of objects that have been successfully created but not yet deleted.
*
* Using too small values will cause vTraceError to be called, which stores an
* error message in the trace that is shown when opening the trace file. The
* error message can also be retrieved using xTraceGetLastError.
*
* It can be wise to start with large values for these constants,
* unless you are very confident on these numbers. Then do a recording and
* check the actual usage by selecting View menu -> Trace Details ->
* Resource Usage -> Object Table.
******************************************************************************/
#define TRC_CFG_NTASK 150
#define TRC_CFG_NISR 90
#define TRC_CFG_NQUEUE 90
#define TRC_CFG_NSEMAPHORE 90
#define TRC_CFG_NMUTEX 90
#define TRC_CFG_NTIMER 250
#define TRC_CFG_NEVENTGROUP 90
/******************************************************************************
* TRC_CFG_INCLUDE_MEMMANG_EVENTS
*
* Macro which should be defined as either zero (0) or one (1).
*
* This controls if malloc and free calls should be traced. Set this to zero (0)
* to exclude malloc/free calls, or one (1) to include such events in the trace.
*
* Default value is 1.
*****************************************************************************/
#define TRC_CFG_INCLUDE_MEMMANG_EVENTS 1
/******************************************************************************
* TRC_CFG_INCLUDE_USER_EVENTS
*
* Macro which should be defined as either zero (0) or one (1).
*
* If this is zero (0) the code for creating User Events is excluded to
* reduce code size. User Events are application-generated events, like
* "printf" but for the trace log and the formatting is done offline, by the
* Tracealyzer visualization tool. User Events are much faster than a printf
* and can therefore be used in timing critical code.
*
* Default value is 1.
*****************************************************************************/
#define TRC_CFG_INCLUDE_USER_EVENTS 1
/*****************************************************************************
* TRC_CFG_INCLUDE_ISR_TRACING
*
* Macro which should be defined as either zero (0) or one (1).
*
* If this is zero (0), the code for recording Interrupt Service Routines is
* excluded, in order to reduce code size.
*
* Default value is 1.
*
* Note: tracing ISRs requires that you insert calls to vTraceStoreISRBegin
* and vTraceStoreISREnd in your interrupt handlers.
*****************************************************************************/
#define TRC_CFG_INCLUDE_ISR_TRACING 1
/*****************************************************************************
* TRC_CFG_INCLUDE_READY_EVENTS
*
* Macro which should be defined as either zero (0) or one (1).
*
* If one (1), events are recorded when tasks enter scheduling state "ready".
* This allows Tracealyzer to show the initial pending time before tasks enter
* the execution state, and present accurate response times.
* If zero (0), "ready events" are not created, which allows for recording
* longer traces in the same amount of RAM.
*
* Default value is 1.
*****************************************************************************/
#define TRC_CFG_INCLUDE_READY_EVENTS 1
/*****************************************************************************
* TRC_CFG_INCLUDE_OSTICK_EVENTS
*
* Macro which should be defined as either zero (0) or one (1).
*
* If this is one (1), events will be generated whenever the OS clock is
* increased. If zero (0), OS tick events are not generated, which allows for
* recording longer traces in the same amount of RAM.
*
* Default value is 0.
*****************************************************************************/
#define TRC_CFG_INCLUDE_OSTICK_EVENTS 1
/******************************************************************************
* TRC_CFG_INCLUDE_FLOAT_SUPPORT
*
* Macro which should be defined as either zero (0) or one (1).
*
* If this is zero (0), the support for logging floating point values in
* vTracePrintF is stripped out, in case floating point values are not used or
* supported by the platform used.
*
* Floating point values are only used in vTracePrintF and its subroutines, to
* allow for storing float (%f) or double (%lf) arguments.
*
* vTracePrintF can be used with integer and string arguments in either case.
*
* Default value is 0.
*****************************************************************************/
#define TRC_CFG_INCLUDE_FLOAT_SUPPORT 0
/******************************************************************************
* TRC_CFG_INCLUDE_OBJECT_DELETE
*
* Macro which should be defined as either zero (0) or one (1).
*
* This must be enabled (1) if tasks, queues or other
* traced kernel objects are deleted at runtime. If no deletes are made, this
* can be set to 0 in order to exclude the delete-handling code.
*
* Default value is 1.
*****************************************************************************/
#define TRC_CFG_INCLUDE_OBJECT_DELETE 1
/*******************************************************************************
* TRC_CFG_SYMBOL_TABLE_SIZE
*
* Macro which should be defined as an integer value.
*
* This defines the capacity of the symbol table, in bytes. This symbol table
* stores User Events labels and names of deleted tasks, queues, or other kernel
* objects. If you don't use User Events or delete any kernel
* objects you set this to a very low value. The minimum recommended value is 4.
* A size of zero (0) is not allowed since a zero-sized array may result in a
* 32-bit pointer, i.e., using 4 bytes rather than 0.
*
* Default value is 800.
******************************************************************************/
#define TRC_CFG_SYMBOL_TABLE_SIZE 5000
#if (TRC_CFG_SYMBOL_TABLE_SIZE == 0)
#error "TRC_CFG_SYMBOL_TABLE_SIZE may not be zero!"
#endif
/******************************************************************************
* TRC_CFG_NAME_LEN_TASK, TRC_CFG_NAME_LEN_QUEUE, ...
*
* Macros that specify the maximum lengths (number of characters) for names of
* kernel objects, such as tasks and queues. If longer names are used, they will
* be truncated when stored in the recorder.
*****************************************************************************/
#define TRC_CFG_NAME_LEN_TASK 15
#define TRC_CFG_NAME_LEN_ISR 15
#define TRC_CFG_NAME_LEN_QUEUE 15
#define TRC_CFG_NAME_LEN_SEMAPHORE 15
#define TRC_CFG_NAME_LEN_MUTEX 15
#define TRC_CFG_NAME_LEN_TIMER 15
#define TRC_CFG_NAME_LEN_EVENTGROUP 15
/******************************************************************************
*** ADVANCED SETTINGS ********************************************************
******************************************************************************
* The remaining settings are not necessary to modify but allows for optimizing
* the recorder setup for your specific needs, e.g., to exclude events that you
* are not interested in, in order to get longer traces.
*****************************************************************************/
/******************************************************************************
* TRC_CFG_HEAP_SIZE_BELOW_16M
*
* An integer constant that can be used to reduce the buffer usage of memory
* allocation events (malloc/free). This value should be 1 if the heap size is
* below 16 MB (2^24 byte), and you can live with reported addresses showing the
* lower 24 bits only. If 0, you get the full 32-bit addresses.
*
* Default value is 0.
******************************************************************************/
#define TRC_CFG_HEAP_SIZE_BELOW_16M 0
/******************************************************************************
* TRC_CFG_USE_IMPLICIT_IFE_RULES
*
* Macro which should be defined as either zero (0) or one (1).
* Default is 1.
*
* Tracealyzer groups the events into "instances" based on Instance Finish
* Events (IFEs), produced either by default rules or calls to the recorder
* functions vTraceInstanceFinishedNow and vTraceInstanceFinishedNext.
*
* If TRC_CFG_USE_IMPLICIT_IFE_RULES is one (1), the default IFE rules is
* used, resulting in a "typical" grouping of events into instances.
* If these rules don't give appropriate instances in your case, you can
* override the default rules using vTraceInstanceFinishedNow/Next for one
* or several tasks. The default IFE rules are then disabled for those tasks.
*
* If TRC_CFG_USE_IMPLICIT_IFE_RULES is zero (0), the implicit IFE rules are
* disabled globally. You must then call vTraceInstanceFinishedNow or
* vTraceInstanceFinishedNext to manually group the events into instances,
* otherwise the tasks will appear a single long instance.
*
* The default IFE rules count the following events as "instance finished":
* - Task delay, delay until
* - Task suspend
* - Blocking on "input" operations, i.e., when the task is waiting for the
* next a message/signal/event. But only if this event is blocking.
*
* For details, see trcSnapshotKernelPort.h and look for references to the
* macro trcKERNEL_HOOKS_SET_TASK_INSTANCE_FINISHED.
*****************************************************************************/
#define TRC_CFG_USE_IMPLICIT_IFE_RULES 1
/******************************************************************************
* TRC_CFG_USE_16BIT_OBJECT_HANDLES
*
* Macro which should be defined as either zero (0) or one (1).
*
* If set to 0 (zero), the recorder uses 8-bit handles to identify kernel
* objects such as tasks and queues. This limits the supported number of
* concurrently active objects to 255 of each type (tasks, queues, mutexes,
* etc.) Note: 255, not 256, since handle 0 is reserved.
*
* If set to 1 (one), the recorder uses 16-bit handles to identify kernel
* objects such as tasks and queues. This limits the supported number of
* concurrent objects to 65535 of each type (object class). However, since the
* object property table is limited to 64 KB, the practical limit is about
* 3000 objects in total.
*
* Default is 0 (8-bit handles)
*
* NOTE: An object with handle above 255 will use an extra 4-byte record in
* the event buffer whenever the object is referenced. Moreover, some internal
* tables in the recorder gets slightly larger when using 16-bit handles.
*****************************************************************************/
#define TRC_CFG_USE_16BIT_OBJECT_HANDLES 0
/******************************************************************************
* TRC_CFG_USE_TRACE_ASSERT
*
* Macro which should be defined as either zero (0) or one (1).
* Default is 1.
*
* If this is one (1), the TRACE_ASSERT macro (used at various locations in the
* trace recorder) will verify that a relevant condition is true.
* If the condition is false, prvTraceError() will be called, which stops the
* recording and stores an error message that is displayed when opening the
* trace in Tracealyzer.
*
* This is used on several places in the recorder code for sanity checks on
* parameters. Can be switched off to reduce the footprint of the tracing, but
* we recommend to have it enabled initially.
*****************************************************************************/
#define TRC_CFG_USE_TRACE_ASSERT 1
/*******************************************************************************
* TRC_CFG_USE_SEPARATE_USER_EVENT_BUFFER
*
* Macro which should be defined as an integer value.
*
* Set TRC_CFG_USE_SEPARATE_USER_EVENT_BUFFER to 1 to enable the
* separate user event buffer (UB).
* In this mode, user events are stored separately from other events,
* e.g., RTOS events. Thereby you can get a much longer history of
* user events as they don't need to share the buffer space with more
* frequent events.
*
* The UB is typically used with the snapshot ring-buffer mode, so the
* recording can continue when the main buffer gets full. And since the
* main buffer then overwrites the earliest events, Tracealyzer displays
* "Unknown Actor" instead of task scheduling for periods with UB data only.
*
* In UB mode, user events are structured as UB channels, which contains
* a channel name and a default format string. Register a UB channel using
* xTraceRegisterUBChannel.
*
* Events and data arguments are written using vTraceUBEvent and
* vTraceUBData. They are designed to provide efficient logging of
* repeating events, using the same format string within each channel.
*
* Examples:
*
* traceString chn1 = xTraceRegisterString("Channel 1");
* traceString fmt1 = xTraceRegisterString("Event!");
* traceUBChannel UBCh1 = xTraceRegisterUBChannel(chn1, fmt1);
*
* traceString chn2 = xTraceRegisterString("Channel 2");
* traceString fmt2 = xTraceRegisterString("X: %d, Y: %d");
* traceUBChannel UBCh2 = xTraceRegisterUBChannel(chn2, fmt2);
*
* // Result in "[Channel 1] Event!"
* vTraceUBEvent(UBCh1);
*
* // Result in "[Channel 2] X: 23, Y: 19"
* vTraceUBData(UBCh2, 23, 19);
*
* You can also use the other user event functions, like vTracePrintF.
* as they are then rerouted to the UB instead of the main event buffer.
* vTracePrintF then looks up the correct UB channel based on the
* provided channel name and format string, or creates a new UB channel
* if no match is found. The format string should therefore not contain
* "random" messages but mainly format specifiers. Random strings should
* be stored using %s and with the string as an argument.
*
* // Creates a new UB channel ("Channel 2", "%Z: %d")
* vTracePrintF(chn2, "%Z: %d", value1);
*
* // Finds the existing UB channel
* vTracePrintF(chn2, "%Z: %d", value2);
******************************************************************************/
#define TRC_CFG_USE_SEPARATE_USER_EVENT_BUFFER 0
/*******************************************************************************
* TRC_CFG_SEPARATE_USER_EVENT_BUFFER_SIZE
*
* Macro which should be defined as an integer value.
*
* This defines the capacity of the user event buffer (UB), in number of slots.
* A single user event can use multiple slots, depending on the arguments.
*
* Only applicable if TRC_CFG_USE_SEPARATE_USER_EVENT_BUFFER is 1.
******************************************************************************/
#define TRC_CFG_SEPARATE_USER_EVENT_BUFFER_SIZE 200
/*******************************************************************************
* TRC_CFG_UB_CHANNELS
*
* Macro which should be defined as an integer value.
*
* This defines the number of User Event Buffer Channels (UB channels).
* These are used to structure the events when using the separate user
* event buffer, and contains both a User Event Channel (the name) and
* a default format string for the channel.
*
* Only applicable if TRC_CFG_USE_SEPARATE_USER_EVENT_BUFFER is 1.
******************************************************************************/
#define TRC_CFG_UB_CHANNELS 32
/*******************************************************************************
* TRC_CFG_ISR_TAILCHAINING_THRESHOLD
*
* Macro which should be defined as an integer value.
*
* If tracing multiple ISRs, this setting allows for accurate display of the
* context-switching also in cases when the ISRs execute in direct sequence.
*
* vTraceStoreISREnd normally assumes that the ISR returns to the previous
* context, i.e., a task or a preempted ISR. But if another traced ISR
* executes in direct sequence, Tracealyzer may incorrectly display a minimal
* fragment of the previous context in between the ISRs.
*
* By using TRC_CFG_ISR_TAILCHAINING_THRESHOLD you can avoid this. This is
* however a threshold value that must be measured for your specific setup.
* See http://percepio.com/2014/03/21/isr_tailchaining_threshold/
*
* The default setting is 0, meaning "disabled" and that you may get an
* extra fragments of the previous context in between tail-chained ISRs.
*
* Note: This setting has separate definitions in trcSnapshotConfig.h and
* trcStreamingConfig.h, since it is affected by the recorder mode.
******************************************************************************/
#define TRC_CFG_ISR_TAILCHAINING_THRESHOLD 0
#endif /*TRC_SNAPSHOT_CONFIG_H*/