README ====== This README discusses issues unique to NuttX configurations for the Spark Core board from Spark Devices (http://www.spark.io). This board features the STM32103CBT6 MCU from STMicro. Microprocessor: 32-bit ARM Cortex M3 at 72MHz STM32F103CBT6 Memory: 120 KB Flash and 20 KB SRAM, 2M serial Flash I/O Pins Out: 37, 17 On the Connector Network: TI CC3000 Wifi Module ADCs: 9 (at 12-bit resolution) Peripherals: 4 timers, 2 I2Cs, 2 SPI ports, 3 USARTs, 2 led's one Blue and one RGB. Other: Sleep, stop, and standby modes; serial wire debug and JTAG interfaces During the development of the SparkCore, the hardware was in limited supply As a work around David Sidrane created a SparkCore Big board (http://nscdg.com/spark/sparkBB.png) that will interface with a maple mini (http://leaflabs.com/docs/hardware/maple-mini.html), and a CC3000BOOST (https://estore.ti.com/CC3000BOOST-CC3000-BoosterPack-P4258.aspx) It breaks out the Tx, Rx to connect to a FTDI TTL-232RG-VREG3V3-WE for the console and wires in the spark LEDs and serial flash to the same I/O as the sparkcore. It has a Jlink compatible Jtag connector on it. Contents ======== - Development Environment - GNU Toolchain Options - IDEs - NuttX EABI "buildroot" Toolchain - NuttX OABI "buildroot" Toolchain - NXFLAT Toolchain - Hardware - Core Pin out - LEDs - Buttons - USARTS and Serial Consoles - DFU and JTAG - Spark -specific Configuration Options - Configurations Development Environment ======================= Either Linux or Cygwin on Windows can be used for the development environment. The source has been built only using the GNU toolchain (see below). Other toolchains will likely cause problems. GNU Toolchain Options ===================== Toolchain Configurations ------------------------ The NuttX make system has been modified to support the following different toolchain options. 1. The CodeSourcery GNU toolchain, 2. The Atollic Toolchain, 3. The devkitARM GNU toolchain, 4. Raisonance GNU toolchain, or 5. The NuttX buildroot Toolchain (see below). All testing has been conducted using the CodeSourcery toolchain for Linux. To use the Atollic, devkitARM, Raisonance GNU, or NuttX buildroot toolchain, you simply need to add one of the following configuration options to your .config (or defconfig) file: CONFIG_ARMV7M_TOOLCHAIN_CODESOURCERYW=n : CodeSourcery under Windows CONFIG_ARMV7M_TOOLCHAIN_CODESOURCERYL=y : CodeSourcery under Linux CONFIG_ARMV7M_TOOLCHAIN_ATOLLIC=y : The Atollic toolchain under Windows CONFIG_ARMV7M_TOOLCHAIN_DEVKITARM=n : devkitARM under Windows CONFIG_ARMV7M_TOOLCHAIN_RAISONANCE=y : Raisonance RIDE7 under Windows CONFIG_ARMV7M_TOOLCHAIN_BUILDROOT=n : NuttX buildroot under Linux or Cygwin (default) If you change the default toolchain, then you may also have to modify the PATH in the setenv.h file if your make cannot find the tools. NOTE: the CodeSourcery (for Windows), Atollic, devkitARM, and Raisonance toolchains are Windows native toolchains. The CodeSourcey (for Linux) and NuttX buildroot toolchains are Cygwin and/or Linux native toolchains. There are several limitations to using a Windows based toolchain in a Cygwin environment. The three biggest are: 1. The Windows toolchain cannot follow Cygwin paths. Path conversions are performed automatically in the Cygwin makefiles using the 'cygpath' utility but you might easily find some new path problems. If so, check out 'cygpath -w' 2. Windows toolchains cannot follow Cygwin symbolic links. Many symbolic links are used in Nuttx (e.g., include/arch). The make system works around these problems for the Windows tools by copying directories instead of linking them. But this can also cause some confusion for you: For example, you may edit a file in a "linked" directory and find that your changes had no effect. That is because you are building the copy of the file in the "fake" symbolic directory. If you use a Windows toolchain, you should get in the habit of making like this: make clean_context all An alias in your .bashrc file might make that less painful. 3. Dependencies are not made when using Windows versions of the GCC. This is because the dependencies are generated using Windows pathes which do not work with the Cygwin make. MKDEP = $(TOPDIR)/tools/mknulldeps.sh The CodeSourcery Toolchain (2009q1) ----------------------------------- The CodeSourcery toolchain (2009q1) does not work with default optimization level of -Os (See Make.defs). It will work with -O0, -O1, or -O2, but not with -Os. The Atollic "Pro" and "Lite" Toolchain -------------------------------------- One problem that I had with the Atollic toolchains is that the provide a gcc.exe and g++.exe in the same bin/ file as their ARM binaries. If the Atollic bin/ path appears in your PATH variable before /usr/bin, then you will get the wrong gcc when you try to build host executables. This will cause to strange, uninterpretable errors build some host binaries in tools/ when you first make. Also, the Atollic toolchains are the only toolchains that have built-in support for the FPU in these configurations. If you plan to use the Cortex-M4 FPU, you will need to use the Atollic toolchain for now. See the FPU section below for more information. The Atollic "Lite" Toolchain ---------------------------- The free, "Lite" version of the Atollic toolchain does not support C++ nor does it support ar, nm, objdump, or objdcopy. If you use the Atollic "Lite" toolchain, you will have to set: CONFIG_HAVE_CXX=n In order to compile successfully. Otherwise, you will get errors like: "C++ Compiler only available in TrueSTUDIO Professional" The make may then fail in some of the post link processing because of some of the other missing tools. The Make.defs file replaces the ar and nm with the default system x86 tool versions and these seem to work okay. Disable all of the following to avoid using objcopy: CONFIG_RRLOAD_BINARY=n CONFIG_INTELHEX_BINARY=n CONFIG_MOTOROLA_SREC=n CONFIG_RAW_BINARY=n devkitARM --------- The devkitARM toolchain includes a version of MSYS make. Make sure that the the paths to Cygwin's /bin and /usr/bin directories appear BEFORE the devkitARM path or will get the wrong version of make. IDEs ==== NuttX is built using command-line make. It can be used with an IDE, but some effort will be required to create the project. Makefile Build -------------- Under Eclipse, it is pretty easy to set up an "empty makefile project" and simply use the NuttX makefile to build the system. That is almost for free under Linux. Under Windows, you will need to set up the "Cygwin GCC" empty makefile project in order to work with Windows (Google for "Eclipse Cygwin" - there is a lot of help on the internet). Using Sourcery CodeBench from http://www.mentor.com/embedded-software/sourcery-tools/sourcery-codebench/overview Download and install the latest version (as of this writting it was sourceryg++-2013.05-64-arm-none-eabi) Import the project from git. File->import->Git-URI, then import a Exiting code as a Makefile progject from the working directory the git clone was done to. Select the Sourcery CodeBench for ARM EABI. N.B. You must do one command line build, before the make will work in CodeBench. Native Build ------------ Here are a few tips before you start that effort: 1) Select the toolchain that you will be using in your .config file 2) Start the NuttX build at least one time from the Cygwin command line before trying to create your project. This is necessary to create certain auto-generated files and directories that will be needed. 3) Set up include pathes: You will need include/, arch/arm/src/stm32, arch/arm/src/common, arch/arm/src/armv7-m, and sched/. 4) All assembly files need to have the definition option -D __ASSEMBLY__ on the command line. Startup files will probably cause you some headaches. The NuttX startup file is arch/arm/src/stm32/stm32_vectors.S. With RIDE, I have to build NuttX one time from the Cygwin command line in order to obtain the pre-built startup object needed by RIDE. NuttX EABI "buildroot" Toolchain ================================ A GNU GCC-based toolchain is assumed. The files */setenv.sh should be modified to point to the correct path to the Cortex-M3 GCC toolchain (if different from the default in your PATH variable). If you have no Cortex-M3 toolchain, one can be downloaded from the NuttX SourceForge download site (https://sourceforge.net/projects/nuttx/files/buildroot/). This GNU toolchain builds and executes in the Linux or Cygwin environment. 1. You must have already configured Nuttx in /nuttx. cd tools ./configure.sh stm32_tiny/ 2. Download the latest buildroot package into 3. unpack the buildroot tarball. The resulting directory may have versioning information on it like buildroot-x.y.z. If so, rename /buildroot-x.y.z to /buildroot. 4. cd /buildroot 5. cp configs/cortexm3-eabi-defconfig-4.6.3 .config 6. make oldconfig 7. make 8. Edit setenv.h, if necessary, so that the PATH variable includes the path to the newly built binaries. See the file configs/README.txt in the buildroot source tree. That has more details PLUS some special instructions that you will need to follow if you are building a Cortex-M3 toolchain for Cygwin under Windows. NOTE: Unfortunately, the 4.6.3 EABI toolchain is not compatible with the the NXFLAT tools. See the top-level TODO file (under "Binary loaders") for more information about this problem. If you plan to use NXFLAT, please do not use the GCC 4.6.3 EABI toochain; instead use the GCC 4.3.3 OABI toolchain. See instructions below. NuttX OABI "buildroot" Toolchain ================================ The older, OABI buildroot toolchain is also available. To use the OABI toolchain: 1. When building the buildroot toolchain, either (1) modify the cortexm3-eabi-defconfig-4.6.3 configuration to use EABI (using 'make menuconfig'), or (2) use an exising OABI configuration such as cortexm3-defconfig-4.3.3 2. Modify the Make.defs file to use the OABI conventions: +CROSSDEV = arm-nuttx-elf- +ARCHCPUFLAGS = -mtune=cortex-m3 -march=armv7-m -mfloat-abi=soft +NXFLATLDFLAGS2 = $(NXFLATLDFLAGS1) -T$(TOPDIR)/binfmt/libnxflat/gnu-nxflat-gotoff.ld -no-check-sections -CROSSDEV = arm-nuttx-eabi- -ARCHCPUFLAGS = -mcpu=cortex-m3 -mthumb -mfloat-abi=soft -NXFLATLDFLAGS2 = $(NXFLATLDFLAGS1) -T$(TOPDIR)/binfmt/libnxflat/gnu-nxflat-pcrel.ld -no-check-sections NXFLAT Toolchain ================ If you are *not* using the NuttX buildroot toolchain and you want to use the NXFLAT tools, then you will still have to build a portion of the buildroot tools -- just the NXFLAT tools. The buildroot with the NXFLAT tools can be downloaded from the NuttX SourceForge download site (https://sourceforge.net/projects/nuttx/files/). This GNU toolchain builds and executes in the Linux or Cygwin environment. 1. You must have already configured Nuttx in /nuttx. cd tools ./configure.sh lpcxpresso-lpc1768/ 2. Download the latest buildroot package into 3. unpack the buildroot tarball. The resulting directory may have versioning information on it like buildroot-x.y.z. If so, rename /buildroot-x.y.z to /buildroot. 4. cd /buildroot 5. cp configs/cortexm3-defconfig-nxflat .config 6. make oldconfig 7. make 8. Edit setenv.h, if necessary, so that the PATH variable includes the path to the newly builtNXFLAT binaries. DFU and JTAG ============ Enbling Support for the DFU Bootloader -------------------------------------- The linker files in these projects can be configured to indicate that you will be loading code using STMicro built-in USB Device Firmware Upgrade (DFU) loader or via some JTAG emulator. You can specify the DFU bootloader by adding the following line: CONFIG_STM32_DFU=y to your .config file. Most of the configurations in this directory are set up to use the DFU loader. If CONFIG_STM32_DFU is defined, the code will not be positioned at the beginning of FLASH (0x08000000) but will be offset to 0x08005000. This offset is needed to make space for the DFU loader and 0x08005000 is where the DFU loader expects to find new applications at boot time. If you need to change that origin for some other bootloader, you will need to edit the file(s) ld.script.dfu for the configuration. For Linux or Mac: ---------------- While on Linux or Mac, we can use dfu-util to upload nuttx binary. 1. Make sure we have installed dfu-util. (From yum, apt-get or build from source.) 2. Start the DFU loader (bootloader) on the Spark board. You do this by resetting the board while holding the "Key" button. Windows should recognize that the DFU loader has been installed. 3. Flash the nuttx.bin to the board use dfu-util. Here's an example: $ dfu-util -a1 -d 1eaf:0003 -D nuttx.bin -R For anything not clear, we can refer to LeafLabs official document: http://leaflabs.com/docs/unix-toolchain.html For Windows: ----------- The DFU SE PC-based software is available from the STMicro website, http://www.st.com. General usage instructions: 1. Convert the NuttX Intel Hex file (nuttx.hex) into a special DFU file (nuttx.dfu)... see below for details. 2. Connect the M3 Wildfire board to your computer using a USB cable. 3. Start the DFU loader on the M3 Wildfire board. You do this by resetting the board while holding the "Key" button. Windows should recognize that the DFU loader has been installed. 3. Run the DFU SE program to load nuttx.dfu into FLASH. What if the DFU loader is not in FLASH? The loader code is available inside of the Demo dirctory of the USBLib ZIP file that can be downloaded from the STMicro Website. You can build it using RIDE (or other toolchains); you will need a JTAG emulator to burn it into FLASH the first time. In order to use STMicro's built-in DFU loader, you will have to get the NuttX binary into a special format with a .dfu extension. The DFU SE PC_based software installation includes a file "DFU File Manager" conversion program that a file in Intel Hex format to the special DFU format. When you successfully build NuttX, you will find a file called nutt.hex in the top-level directory. That is the file that you should provide to the DFU File Manager. You will end up with a file called nuttx.dfu that you can use with the STMicro DFU SE program. Enabling JTAG ------------- If you are not using the DFU, then you will probably also need to enable JTAG support. By default, all JTAG support is disabled but there NuttX configuration options to enable JTAG in various different ways. These configurations effect the setting of the SWJ_CFG[2:0] bits in the AFIO MAPR register. These bits are used to configure the SWJ and trace alternate function I/Os. The SWJ (SerialWire JTAG) supports JTAG or SWD access to the Cortex debug port. The default state in this port is for all JTAG support to be disable. CONFIG_STM32_JTAG_FULL_ENABLE - sets SWJ_CFG[2:0] to 000 which enables full SWJ (JTAG-DP + SW-DP) CONFIG_STM32_JTAG_NOJNTRST_ENABLE - sets SWJ_CFG[2:0] to 001 which enable full SWJ (JTAG-DP + SW-DP) but without JNTRST. CONFIG_STM32_JTAG_SW_ENABLE - sets SWJ_CFG[2:0] to 010 which would set JTAG-DP disabled and SW-DP enabled The default setting (none of the above defined) is SWJ_CFG[2:0] set to 100 which disable JTAG-DP and SW-DP. Hardware ======== The Spark comprises a STM32F103CB 72 Mhz, 128 Flash, 20K Ram, with 37 IO Pins, and a TI CC3000 Wifi Module. It has a 2MB serial flash, onboad regulation and 2 led's one Blue and one RGB. During the development of the SparkCore, the hardware was in limited supply As a work around david_s5 created a SparkCore Big board (http://nscdg.com/spark/sparkBB.png) that will interface with a maple mini (http://leaflabs.com/docs/hardware/maple-mini.html), and a CC3000BOOST (https://estore.ti.com/CC3000BOOST-CC3000-BoosterPack-P4258.aspx) It breaks out the Tx, Rx to connect to a FTDI TTL-232RG-VREG3V3-WE for the console and wires in the spark LEDs and serial flash to the same I/O as the sparkcore. It has a Jlink compatible Jtag connector on it. Core Pin out ============ There are 24 pis on the Spark Core module. Spark Spark Function STM32F103CBT6 Name Pin # Pin # -------- ------ ------------------------------------------------ --------------- RAW JP1-1 Input Power N/A GND JP1-2 GND TX JP1-3 PA[02] USART2_TX/ADC12_IN2/TIM2_CH3 12 RX JP1-4 PA[03] USART2_RX/ADC12_IN3/TIM2_CH4 13 A7 JP1-5 PB[01] ADC12_IN9/TIM3_CH4 19 A6 JP1-6 PB[00] ADC12_IN8/TIM3_CH3 18 A5 JP1-7 PA[07] SPI1_MOSI/ADC12_IN7/TIM3_CH2 17 A4 JP1-8 PA[06] SPI1_MISO/ADC12_IN6/TIM3_CH1 16 A3 JP1-9 PA[05] SPI1_SCK/ADC12_IN5 15 A2 JP1-10 PA[04] SPI1_NSS/USART2_CK/ADC12_IN4 14 A1 JP1-11 PA[01] USART2_RTS/ADC12_IN1/TIM2_CH2 11 A0 JP1-12 PA[00] WKUP/USART2_CTS/ADC12_IN0/TIM2_CH1_ETR 10 +3V3 JP2-1 V3.3 Out of Core NA RST JP2-2 NRST 7 VDDA JP2-3 ADC Voltage 9 GND JP2-4 GND D7 JP2-5 PA[13] JTMS/SWDIO 34 Common with Blue LED LED_USR D6 JP2-6 PA[14] JTCK/SWCLK 37 D5 JP2-7 PA[15] JTDI 38 D4 JP2-8 PB[03] JTDO 39 D3 JP2-9 PB[04] NJTRST 40 D2 JP2-10 PB[05] I2C1_SMBA 41 D1 JP2-11 PB[06] I2C1_SCL/TIM4_CH1 42 D0 JP2-12 PB[07] I2C1_SDA/TIM4_CH2 43 Core Internal IO ================ Spark Function STM32F103CBT6 Name Pin # -------- ------------------------------------------------ --------------- BTN PB[02] BOOT1 20 LED1,D7 PA[13] JTMS/SWDIO 34 LED2 PA[08] USART1_CK/TIM1_CH1/MCO 29 LED3 PA[09] USART1_TX/TIM1_CH2 30 LED4 PA[10] USART1_RX/TIM1_CH3 31 MEM_CS PB[09] TIM4_CH4 46 SST25VF016B Chip Select SPI_CLK PB[13] SPI2_SCK/USART3_CTS/TIM1_CH1N 26 SPI_MISO PB[14] SPI2_MISO/USART3_RTS/TIM1_CH2N 27 SPI_MOSI PB[15] SPI2_MOSI/TIM1_CH3N 28 USB_DISC PB[10] I2C2_SCL/USART3_TX 21 WIFI_CS PB[12] SPI2_NSS/I2C2_SMBA/USART3_CK/TIM1_BKIN 25 CC3000 Chip Select WIFI_EN PB[08] TIM4_CH3 45 CC3000 Module enable WIFI_INT PB[11] I2C2_SDA/USART3_RX 22 CC3000 Host interface SPI interrupt Buttons and LEDs ================ Buttons ------- The Spark has two mechanical buttons. One button is the RESET button connected to the STM32F103CB's reset line via /RST and the other is a generic user configurable button labeled BTN and connected to GPIO PB2/BOOT1. Since on the Spark, BOOT0 is tied to GND it is a moot point that BTN signal is connected to the BOOT1 signal. When a button is pressed it will drive the I/O line to GND. LEDs ---- There are 4 user-controllable LEDs in two packages on board the Spark board: Sigal Location Color GPIO Active ------- ------------ ----------- ----- ----------- LED1 LED_USR Blue LED PA13 High Common With D7 LED2 LED_RGB Red LED PA8 Low LED3 LED_RGB Blue LED PA9 Low LED4 LED_RGB Green LED PA10 Low LED1 is connected to ground and can be illuminated by driving the PA13 output high, it shares the Sparks D7 output. The LED2,LED3 and LED4 are pulled high and can be illuminated by driving the corresponding GPIO output to low. The RGB LEDs are not used by the board port unless CONFIG_ARCH_LEDS is defined. In that case, the usage by the board port is defined in include/board.h and src/up_leds.c. The LEDs are used to encode OS-related events as follows: SYMBOL Meaning LED2 LED3 LED4 red blue green Color ------------------- ----------------------- ------- ------- ------ --------- LED_STARTED NuttX has been started ON OFF OFF Red LED_HEAPALLOCATE Heap has been allocated OFF ON OFF Blue LED_IRQSENABLED Interrupts enabled ON OFF ON Orange LED_STACKCREATED Idle stack created OFF OFF ON Green LED_INIRQ In an interrupt** ON N/C N/C Orange Glow LED_SIGNAL In a signal handler*** N/C ON N/C Blue Glow LED_ASSERTION An assertion failed ON ON ON White LED_PANIC The system has crashed ON N/C N/C Red Flashing LED_IDLE STM32 is is sleep mode (Optional, not used) * If LED2, LED3, LED4 are statically on, then NuttX probably failed to boot and these LEDs will give you some indication of where the failure was ** The normal state is LED4 ON and LED2 faintly glowing. This faint glow is because of timer interupts that result in the LED being illuminated on a small proportion of the time. *** LED3 may also flicker normally if signals are processed. Serial Consoles =============== USART2 ----- If you have a 3.3 V TTL to RS-232 convertor then this is the most convenient serial console to use. UART2 is the default in all of these configurations. USART2 RX PA3 JP1 pin 4 USART2 TX PA2 JP1 pin 3 GND JP1 pin 2 V3.3 JP2 pin 1 Virtual COM Port ---------------- Yet another option is to use UART0 and the USB virtual COM port. This option may be more convenient for long term development, but was painful to use during board bring-up. Spark -specific Configuration Options ============== WIP Configurations ============== Composite: The composite is a super set of all the functions in nsh, usbserial, usbmsc. (usbnsh has not been rung out). Build it with make distclean;cd tools;./configure.sh spark/composite;cd .. then run make menuconfig if you wish to customize things. N.B. Memory is tight, both Flash and RAM are taxed. If you enable debugging you will need to add -Os following the line -g in the line: ifeq ($(CONFIG_DEBUG_SYMBOLS),y) ARCHOPTIMIZATION = -g in the top level Make.degs or the code will not fit. Stack space has been hand optimized using the stack coloring by enabling "Stack usage debug hooks" (CONFIG_DEBUG_STACK) in Build Setup-> Debug Options. I have selected values that have 8-16 bytes of headroom with network debugging on. If you enable more debugging and get a hard fault or any weirdness like commands hanging. Then the Idle, main or Interrupt stack my be too small. Stop the target and have a look a memory for a blown stack: No DEADBEEF at the lowest address of a given stack. Given the RAM memory constraints it is not possible to be running the network and USB CDC/ACM and MSC at the same time. But on the bright side, you can export the FLASH memory to the PC. Write files on the Flash. Reboot and mount the FAT FS and run network code that will have access the files. You can use the scripts/cdc-acm.inf file to install the windows composite device. SPI2 is enabled and support is included for the FAT file system on the 16Mbit (2M) SST25 device and control of the CC3000 on the spark core. When the system boots, you should have a dev/mtdblock0 that can be mounted using the command: mount -t vfat /dev/mtdblock0 /mnt/p0 or /dev/mtdblock0 can be exported as MSC on the USB interface along with a Virtual serial port as a CDC/ACM interface. Use the command conn* and disconn to manage the USB interface. N.B. *If /dev/mtdblock0 is mounted then You must unmount it prior to exporting it via the conn command. Bad things will happen if not. Network control is facilitated by running the c3b (cc3000basic) application. Run c3b from the nsh prompt. +-------------------------------------------+ | Nuttx CC3000 Demo Program | +-------------------------------------------+ 01 - Initialize the CC3000 02 - Show RX & TX buffer sizes, & free RAM 03 - Start Smart Config 04 - Manually connect to AP 05 - Manually add connection profile 06 - List access points 07 - Show CC3000 information 08 - Telnet Type 01-07 to select above option: Select 01. Then use 03 and the TI Smart config application running on an IOS or Android device to configure join your network. Use 07 to see the IP address of the device. (On the next reboot running c3b 01 the CC3000 will automaticaly rejoin the network after the 01 give it a few seconds and enter 07 or 08) Use 08 to start Telnet. Then you can connect to the device using the address listed in command 07. qq will exit the c3b with the telnet deamon running (if started) Slow.... You will be thinking 300 bps. This is because of packet sizes and how the select thread runs in the telnet session. Telnet is not the best showcase for the CC3000, but simply a proof of network connectivity. http POST and GET should be more efficient.