Skip to content

Latest commit

 

History

History
189 lines (144 loc) · 8.59 KB

README.md

File metadata and controls

189 lines (144 loc) · 8.59 KB
Raw

Nx_Iperf application description

This application provides an example of Azure RTOS NetXDuo stack usage . It shows performance when using different modes : TCP_server, UDP_server, TCP_client and UDP_client. The main entry function tx_application_define() is then called by ThreadX during kernel start, at this stage, all NetXDuo resources are created.

  • A NX_PACKET_POOL AppPool is allocated
  • A NX_IP instance IpInstance using that pool is initialized
  • A NX_PACKET_POOL WebServerPool is allocated
  • The ARP, ICMP and protocols (TCP and UDP) are enabled for the NX_IP instance
  • A DHCP client is created.

The application creates 1 thread :

  • AppMainThread (priority 4, PreemtionThreashold 4) : created with the TX_AUTO_START flag to start automatically.

The AppMainThread starts and perform the following actions:

  • Starts the DHCP client
  • Waits for the IP address resolution
  • Resumes the nx_iperf_entry

The nx_iperf_entry, once started:

  • Creates a NetXDuo Iperf demo web page.

The application then creates 4 threads with the same priorities :

  • thread_tcp_rx_iperf (priority 1, PreemtionThreashold 1) : created with the TX_AUTO_START flag to start automatically.
  • thread_tcp_tx_iperf (priority 1, PreemtionThreashold 1) : created with the TX_AUTO_START flag to start automatically.
  • thread_udp_rx_iperf (priority 1, PreemtionThreashold 1) : created with the TX_AUTO_START flag to start automatically.
  • thread_udp_tx_iperf (priority 1, PreemtionThreashold 1) : created with the TX_AUTO_START flag to start automatically.

Expected success behavior

Error behaviors

  • The Red LED is toggling to indicate any error that have occurred.

Assumptions if any

  • The Application is using the DHCP to acquire IP address, thus a DHCP server should be reachable by the board in the LAN used to test the application.
  • The application is configuring the Ethernet IP with a static predefined MAC Address, make sure to change it in case multiple boards are connected on the same LAN to avoid any potential network traffic issues.
  • The MAC Address is defined in the main.c`
void MX_ETH_Init(void)
{

  /* USER CODE BEGIN ETH_Init 0 */

  /* USER CODE END ETH_Init 0 */

  /* USER CODE BEGIN ETH_Init 1 */

  /* USER CODE END ETH_Init 1 */
  heth.Instance = ETH;
  heth.Init.MACAddr[0] =   0x00;
  heth.Init.MACAddr[1] =   0x80;
  heth.Init.MACAddr[2] =   0xE1;
  heth.Init.MACAddr[3] =   0x00;
  heth.Init.MACAddr[4] =   0x00;
  heth.Init.MACAddr[5] =   0x00;

Known limitations

  • The packet pool is not optimized. It can be less than that by reducing NX_PACKET_POOL_SIZE in file "app_netxduo.h" and NX_APP_MEM_POOL_SIZE in file "app_azure_rtos_config.h". This update can decrease NetXDuo performance.

Notes

  1. If the user code size exceeds the DTCM-RAM size or starts from internal cacheable memories (SRAM1 and SRAM2), that is shared between several processors, then it is highly recommended to enable the CPU cache and maintain its coherence at application level. The address and the size of cacheable buffers (shared between CPU and other masters) must be properly updated to be aligned to cache line size (32 bytes).

ThreadX usage hints

  • ThreadX uses the Systick as time base, thus it is mandatory that the HAL uses a separate time base through the TIM IPs.

  • ThreadX is configured with 100 ticks/sec by default, this should be taken into account when using delays or timeouts at application. It is always possible to reconfigure it in the "tx_user.h", the "TX_TIMER_TICKS_PER_SECOND" define,but this should be reflected in "tx_initialize_low_level.s" file too.

  • ThreadX is disabling all interrupts during kernel start-up to avoid any unexpected behavior, therefore all system related calls (HAL, BSP) should be done either at the beginning of the application or inside the thread entry functions.

  • ThreadX offers the "tx_application_define()" function, that is automatically called by the tx_kernel_enter() API. It is highly recommended to use it to create all applications ThreadX related resources (threads, semaphores, memory pools...) but it should not in any way contain a system API call (HAL or BSP).

  • Using dynamic memory allocation requires to apply some changes to the linker file. ThreadX needs to pass a pointer to the first free memory location in RAM to the tx_application_define() function, using the "first_unused_memory" argument. This require changes in the linker files to expose this memory location.

    • For EWARM add the following section into the .icf file:
    place in RAM_region    { last section FREE_MEM };
    • For MDK-ARM:
    either define the RW_IRAM1 region in the ".sct" file
or modify the line below in "tx_low_level_initilize.s to match the memory region being used
    LDR r1, =|Image$$RW_IRAM1$$ZI$$Limit|
    • For STM32CubeIDE add the following section into the .ld file:
    ._threadx_heap :
  {
     . = ALIGN(8);
     __RAM_segment_used_end__ = .;
     . = . + 64K;
     . = ALIGN(8);
   } >RAM AT> RAM

    The simplest way to provide memory for ThreadX is to define a new section, see ._threadx_heap above.
In the example above the ThreadX heap size is set to 64KBytes.
The ._threadx_heap must be located between the .bss and the ._user_heap_stack sections in the linker script.
Caution: Make sure that ThreadX does not need more than the provided heap memory (64KBytes in this example).
Read more in STM32CubeIDE User Guide, chapter: "Linker script".
    • The "tx_initialize_low_level.s" should be also modified to enable the "USE_DYNAMIC_MEMORY_ALLOCATION" flag.

NetX Duo usage hints

  • The NetXDuo application needs to allocate the NX_PACKET pool in a dedicated section that is configured as below is an example of the section declaration for different IDEs.
  • For EWARM ".icf" file 
    define symbol __ICFEDIT_region_NXDATA_start__  = 0x20064000;
define symbol __ICFEDIT_region_NXDATA_end__   = 0x2007FFFF;
define region NXAppPool_region  = mem:[from __ICFEDIT_region_NXDATA_start__ to __ICFEDIT_region_NXDATA_end__];
place in NXAppPool_region { section .NXAppPoolSection};
    • For MDK-ARM
     RW_NXDriverSection 0x20064000 0x1A00  {

*(.NetXPoolSection) }

+ For STM32CubeIDE ".ld" file

.nx_section 0x20064000 (NOLOAD): { *(.NXAppPoolSection) } >RAM


this section is then used in the <code> app_azure_rtos.c</code> file to force the <code>nx_byte_pool_buffer</code> allocation.

/* USER CODE BEGIN NX_Pool_Buffer */

#if defined ( ICCARM ) /* IAR Compiler */ #pragma location = ".NetXPoolSection"

#elif defined ( __CC_ARM ) /* MDK ARM Compiler */ attribute((section(".NetXPoolSection")))

#elif defined ( GNUC ) /* GNU Compiler */ attribute((section(".NetXPoolSection"))) #endif

/* USER CODE END NX_Pool_Buffer */ __ALIGN_BEGIN static UCHAR nx_byte_pool_buffer[NX_APP_MEM_POOL_SIZE] __ALIGN_END; static TX_BYTE_POOL nx_app_byte_pool;

For more details about the MPU configuration please refer to the [AN4838](https://www.st.com/resource/en/application_note/dm00272912-managing-memory-protection-unit-in-stm32-mcus-stmicroelectronics.pdf)

### <b>Keywords</b>

RTOS, Network, ThreadX, NetXDuo, UART

### <b>Hardware and Software environment</b>

  - This application runs on STM32F767ZIxx devices
  - This application has been tested with STMicroelectronics STM32F767ZI-NUCLEO board MB1137 Revision B-01
    and can be easily tailored to any other supported device and development board.

  - This application uses USART3 to display logs, the hyperterminal configuration is as follows:
      - BaudRate = 115200 baud
      - Word Length = 8 Bits
      - Stop Bit = 1
      - Parity = None
      - Flow control = None

###  <b>How to use it ?</b>

In order to make the program work, you must do the following :

 - Open your preferred toolchain
 - Rebuild all files and load your image into target memory
 - Run the application
 - Run the [iperf tool] by following steps described in this link https://docs.microsoft.com/en-us/azure/rtos/netx-duo/netx-duo-iperf/chapter3 .