Let's keep conversation here so other may benefit. You can try to reaching me on slack mynewt general channel.
Jerzy pt., 27 cze 2025 o 11:42 Keith P <keithcpere...@gmail.com> napisał(a): > Hi Jerzy, > > The build tools are available here: > https://gitlab.com/shaktiproject/software/riscv-tools > The code is ran on an FPGA (Arty A7-100T), there is a qemu vm but it > doesn't work correctly. > > I realised that the _start may already be defined while looking closer at > the Hifive1 startup code, I'll make those changes. > Is it possible that we can get on a call if you are free on the weekend? > > Regards, > Keith > > On Thu, Jun 26, 2025 at 3:49 PM Jerzy Kasenberg < > jerzy.kasenb...@codecoup.pl> > wrote: > > > Hi Keith, > > > > Your code defines _start as staring point and it also defines this > > function. > > This function initializes bss and calls init then it ends with busy loop. > > > > To make it link more code you could change your startup function to > > different name > > then after your init (or at the end of this function) you could call > _start > > function that is present > > in baselib.c and that would start more mynewt code. > > > > Keep in mind that mynewt expect maynew_main function (not main) to be > > starting point of application. > > > > regards > > Jerzy > > > > > > czw., 26 cze 2025 o 10:18 Jerzy Kasenberg <jerzy.kasenb...@codecoup.pl> > > napisał(a): > > > > > Hi Keith, > > > > > > I did port of hifive1 few years back but I don't have setup ready. > > > Can you tell me what system you are using for development and how to > get > > > build tools you have? > > > If I can setup environment again I'm pretty sure we can make this work. > > > Originally compiler used for HiFive1 board was downloaded from SiFive > > site > > > I don't know > > > it this is the one you are using and if the one I used is still there > > > > > > regards > > > Jerzy > > > > > > > > > śr., 25 cze 2025 o 09:37 Keith P <keithcpere...@gmail.com> napisał(a): > > > > > >> Hello everyone, > > >> > > >> I am an intern at a lab and don't really get how to continue with > > porting, > > >> I have made a few files and copied some code from the mcu's original > SDK > > >> <https://gitlab.com/shaktiproject/software/shakti-sdk>. > > >> > > >> I am porting to the Shakti line of MCU's, they are fully open-source > > RISCV > > >> based MCUs. > > >> > > >> I have followed the fairly short porting guide and have been trying to > > use > > >> the FE310 bsp implementation by SiFive as a template/guide since that > is > > >> also a RISCV board. > > >> > > >> My task as of now is to get UART working. > > >> > > >> Here are the directories I've added: > > >> repos/apache-mynewt-core/hw/bsp/shakti_parashu/ > > >> ├── bsp.yml > > >> ├── include > > >> │ └── bsp > > >> │ └── bsp.h > > >> ├── parashu_boot.ld > > >> ├── parashu.ld > > >> ├── pkg.yml > > >> ├── src > > >> │ └── hal_bsp.c > > >> └── syscfg.yml > > >> > > >> 3 directories, 7 files > > >> > > >> repos/apache-mynewt-core/hw/mcu/shakti/ > > >> ├── parashu > > >> │ ├── include > > >> │ │ ├── clint.h > > >> │ │ ├── defines.h > > >> │ │ ├── mcu > > >> │ │ │ ├── parashu.h > > >> │ │ │ ├── parashu_hal.h > > >> │ │ │ ├── platform.h > > >> │ │ │ └── sys_clock.h > > >> │ │ ├── plic.h > > >> │ │ └── traps.h > > >> │ ├── pkg.yml > > >> │ ├── src > > >> │ │ ├── arch > > >> │ │ │ └── rv32imacu > > >> │ │ │ └── start.S > > >> │ │ ├── clint.c > > >> │ │ ├── hal_flash.c > > >> │ │ ├── hal_os_tick.c > > >> │ │ ├── hal_reset_cause.c > > >> │ │ ├── hal_system.c > > >> │ │ ├── hal_uart.c > > >> │ │ ├── init.c > > >> │ │ ├── plic.c > > >> │ │ ├── sys_clock.c > > >> │ │ └── traps.c > > >> │ └── syscfg.yml > > >> └── pkg.yml > > >> > > >> 6 directories, 22 files > > >> > > >> repos/apache-mynewt-core/kernel/os/src/arch/shakti/ > > >> ├── start.S > > >> └── trap.S > > >> > > >> 0 directories, 2 files > > >> > > >> repos/apache-mynewt-core/compiler/riscv64-unknown-elf-shakti/ > > >> ├── compiler.yml > > >> └── pkg.yml > > >> > > >> 0 directories, 2 files > > >> > > >> I forked the mynewt-core repo and uploaded my changes to the fork > > >> <https://github.com/Keith-C-P/shakti-mynewt-core>. > > >> > > >> The code compiles, but the dump looks all wrong as major sections are > > >> missing, like main and others > > >> > > >> I did compile mynewt for both native (linux) and for hifive1, i ran > > >> hifive1 > > >> on a qemu vm as I don't have access to the actual board, but both > worked > > >> fine. > > >> > > >> I'll be honest I don't know what questions to ask as I am just > starting > > to > > >> get into the embedded field and wanted a difficult project so I can > > learn > > >> fast. > > >> > > >> Any and all help would be greatly appreciated > > >> Thanks in advance > > >> > > >> Regards, > > >> Keith > > >> > > > > > >