*Hello everyone, Tiago,*
I have made progress with the build issue I was encountering and am now
"stuck" at what I believe is one of the final stages. I am experiencing the
following compilation error:
LD: nuttx
riscv-none-elf-ld: warning: /home/felipe-moura/nuttxspace/nuttx-felipe/nuttx
has a LOAD segment with RWX permissions
riscv-none-elf-ld: /home/felipe-moura/nuttxspace/nuttx-felipe/staging/
libarch.a(sleep_modes.o): in function
`esp_set_deep_sleep_wake_stub_default_entry':
sleep_modes.c:(.rtc.text.4+0x0): undefined reference to
`_rtc_force_fast_end'
riscv-none-elf-ld: sleep_modes.c:(.rtc.text.4+0x4): undefined reference to
`_rtc_text_start'
riscv-none-elf-ld: sleep_modes.c:(.rtc.text.4+0x8): undefined reference to
`_rtc_force_fast_end'
riscv-none-elf-ld: sleep_modes.c:(.rtc.text.4+0xc): undefined reference to
`_rtc_text_start'
riscv-none-elf-ld:
/home/felipe-moura/nuttxspace/nuttx-felipe/staging/libarch.a(sleep_cpu.o):
in function `esp_sleep_cpu_retention':
sleep_cpu.c:(.iram1.8+0x200): undefined reference to
`rv_core_critical_regs_save'
riscv-none-elf-ld: sleep_cpu.c:(.iram1.8+0x21e): undefined reference to
`rv_core_critical_regs_restore'
riscv-none-elf-ld: sleep_cpu.c:(.iram1.8+0x226): undefined reference to
`rv_core_critical_regs_restore'
make[1]: *** [Makefile:189: nuttx] Error 1
make: *** [tools/Unix.mk:551: nuttx] Error 2
I would greatly appreciate your assistance.
Focusing on the item _rtc_text_start, I see that it is defined in the file:
arch/risc-v/src/chip/esp-hal-3rdparty/components/esp_system/ld/esp32c6/
sections.ld.in
However, I am unsure how to add this file to the complete linking process.
I added this file to hal_esp32c6.mk, but I received the following error:
LD: nuttx
riscv-none-elf-ld:/home/felipe-moura/nuttxspace/nuttx-felipe/arch/risc-v/src
/chip/esp-hal-3rdparty/components/esp_system/ld/esp32c6/sections.ld.in.tmp:
113 cannot move location counter backwards (from 40809800 to 40800000)
make[1]: *** [Makefile:189: nuttx] Error 1
make: *** [tools/Unix.mk:551: nuttx] Error 2
Does anyone have any tips or suggestions that could help me move forward
with this process?
Thank you in advance for your help.
*Best regards,*
Em sex., 25 de out. de 2024 às 12:02, Tiago Medicci Serrano <
tiago.medi...@gmail.com> escreveu:
> Hi Felipe,
>
> The RTC GPIO is a feature of the GPIO/RTC driver, so a substitute for this
> function should be part
> of `nuttx/arch/risc-v/src/common/espressif/esp_rtc_gpio.c`. That being
> said, you can either implement it (and define it as a macro, on HAL) or you
> can either ignore these functions related to GPIO wakeup from sleep for now
> (using, for instance, the `#ifndef __NuttX__`). It's up to you: if the GPIO
> wakeup isn't mandatory, it'd go baby steps, removing this feature for now.
>
> Best regards,
>
> Em sex., 25 de out. de 2024 às 09:33, Felipe Moura Oliveira <
> moura....@gmail.com> escreveu:
>
> > Hello Tiago,
> >
> > Thank you for your assistance earlier.
> >
> > I would like to discuss further the necessity of the
> > esp_driver_gpio/include/rtc_io.h file, which is only available in the
> IDF.
> >
> > In sleep_modes.c, there is the following code snippet:
> > esp_err_t esp_sleep_enable_ext1_wakeup_io(uint64_t io_mask,
> > esp_sleep_ext1_wakeup_mode_t level_mode)
> > {
> > if (io_mask == 0 && level_mode > ESP_EXT1_WAKEUP_ANY_HIGH) {
> > return ESP_ERR_INVALID_ARG;
> > }
> > // Translate bit map of GPIO numbers into the bit map of RTC IO numbers
> > uint32_t rtc_gpio_mask = 0;
> > for (int gpio = 0; io_mask; ++gpio, io_mask >>= 1) {
> > if ((io_mask & 1) == 0) {
> > continue;
> > }
> > if (!esp_sleep_is_valid_wakeup_gpio(gpio)) {
> > ESP_LOGE(TAG, "Not an RTC IO: GPIO%d", gpio);
> > return ESP_ERR_INVALID_ARG;
> > }
> > rtc_gpio_mask |= BIT(rtc_io_number_get(gpio));
> > }
> >
> > I believe this code is essential for compilation. The function
> > rtc_io_number_get is located at the following path in the IDF:
> > components/esp_driver_gpio/include/driver/rtc_io.h.
> >
> > However, the esp_driver_gpio folder is not present in the 3rd-party
> > directory. In this case, what is the recommended procedure to proceed?
> >
> > Thank you again for your support.
> >
> > Em sex., 25 de out. de 2024 às 08:52, Tiago Medicci Serrano <
> > tiago.medi...@gmail.com> escreveu:
> >
> > > Hi Felipe,
> > >
> > > `esp_private/pm_impl.h` is from the component `esp_pm` of ESP-IDF,
> which
> > is
> > > a driver directly. Same for `esp_driver_gpio/include/rtc_io.h` (here, a
> > > detail: I wasn't able to find this path on IDF, so make sure you are
> > > checking under the `release/v5.1` branch, which the HAL was based). The
> > HAL
> > > repository doesn't contain ESP-IDF drivers, which are always
> implemented
> > on
> > > NuttX. For `esp_cpu.h`, it's already on HAL (at
> > > `esp-hal-3rdparty/components/esp_hw_support/include/esp_cpu.h`).
> > >
> > > The files under `esp-hal-3rdparty/components/esp_hw_support`
> > > (`sleep_modes.c`, for instance) MAY be used to implement the driver.
> This
> > > is, mostly, a wrapping layer under the function of `esp_rom` and `hal`
> > > components. However, they still may refer to ESP-IDF drivers. That
> being
> > > said, you can check if removing such dependencies is possible. We
> usually
> > > do that with conditional macros. A good starting point is to check for
> > > `#ifdef __NuttX__` on these sources and headers: we use them to
> > remove/add
> > > headers and functions that aren't available on NuttX/needed by NuttX.
> > >
> > > Best regards,
> > >
> > > Em qui., 24 de out. de 2024 às 19:46, Felipe Moura Oliveira <
> > > moura....@gmail.com> escreveu:
> > >
> > > > Hello all,
> > > >
> > > > Tiago,
> > > >
> > > > I have a question regarding the porting process. Specifically, I need
> > to
> > > > add the sleep_modes.c file to the compilation process. Am I on the
> > right
> > > > path with this approach?
> > > >
> > > > To achieve this, I modified the hal_esp32c6.mk file by adding
> > > > sleep_modes.c.
> > > > However, I am facing some issues: within sleep_modes.c, some #include
> > > > statements reference files that are not available in the 3rd-party
> > > > directory but are only present in the IDF. For example:
> > > >
> > > > - esp_private/pm_impl.h
> > > > - esp_cpu.h
> > > > - components/esp_driver_gpio/include/rtc_io.h
> > > >
> > > > I need those includes because sleep_modes.c uses function from there.
> > > >
> > > > Could you please advise if it's possible to incorporate these IDF
> > folders
> > > > into the 3rd-party directory, or if that approach is not recommended?
> > > >
> > > > Thank you for your assistance.
> > > >
> > > > Best regards,
> > > >
> > > > Em qua., 23 de out. de 2024 às 15:21, Tiago Medicci Serrano <
> > > > tiago.medi...@gmail.com> escreveu:
> > > >
> > > > > Hi,
> > > > >
> > > > > Just another note to guide your development: we don't develop
> > anything
> > > on
> > > > > `esp-hal-3rdparty`: most of the code we use from it is derived from
> > the
> > > > > path
> > > > >
> > > > >
> > > >
> > >
> >
> https://github.com/espressif/esp-hal-3rdparty/tree/release/v5.1.c/components/hal
> > > > > (essentially, header files from
> `components/hal/<chip>/include/hal`,
> > > > except
> > > > > for some of the sources files that use the hal functions at
> > > > > `components/hal` to implement some function we may use). You can
> > take a
> > > > > look at the branch `
> > > > >
> > https://github.com/espressif/esp-hal-3rdparty/commits/release/v5.1.c`
> <https://github.com/espressif/esp-hal-3rdparty/commits/release/v5.1.c>
> > <https://github.com/espressif/esp-hal-3rdparty/commits/release/v5.1.c>
> > > <https://github.com/espressif/esp-hal-3rdparty/commits/release/v5.1.c>
> > > > <
> https://github.com/espressif/esp-hal-3rdparty/commits/release/v5.1.c>
> > > > > <
> > https://github.com/espressif/esp-hal-3rdparty/commits/release/v5.1.c>
> > > > and
> > > > > especially the commits beginning
> > > > > with faaa46ebfb37fba4250de831efbbf2862958c344 to check the kind of
> > > > changes
> > > > > we do in hal. Essentially, we create a wrapper layer to make it
> > > > independent
> > > > > of the host OS.
> > > > >
> > > > > *Drivers must always be implemented on NuttX!*
> > > > >
> > > > > That being said, you can either 1) try to adapt the implementation
> of
> > > > other
> > > > > devices to use the HAL functions or 2) get some inspiration on how
> > IDF
> > > > > implements PM. I'd stick with the 2nd option before simply trying
> to
> > > > adapt
> > > > > an existing implementation (from ESP32-C3 legacy or from
> > > ESP32/ESP32-S3).
> > > > >
> > > > > Best regards,
> > > > >
> > > > > Em qua., 23 de out. de 2024 às 13:35, Felipe Moura Oliveira <
> > > > > moura....@gmail.com> escreveu:
> > > > >
> > > > > > Hello Tiago.
> > > > > >
> > > > > > Thank you for clarification, I will follow your instructions.
> > > > > >
> > > > > > Em qua., 23 de out. de 2024 às 13:01, Tiago Medicci Serrano <
> > > > > > tiago.medi...@gmail.com> escreveu:
> > > > > >
> > > > > > > Hi Felipe,
> > > > > > >
> > > > > > > Just complementing: use your forked version of the HAL during
> the
> > > > > > > development. As soon as you have a working version, you can
> > submit
> > > on
> > > > > the
> > > > > > > official repository.
> > > > > > >
> > > > > > > Best regards,
> > > > > > >
> > > > > > > Em qua., 23 de out. de 2024 às 12:58, Tiago Medicci Serrano <
> > > > > > > tiago.medi...@gmail.com> escreveu:
> > > > > > >
> > > > > > > > Hi Felipe,
> > > > > > > >
> > > > > > > > Thanks for asking! You should follow the path using the
> > functions
> > > > on
> > > > > > > > esp-hal-3rdparty. You can fork the repository and use
> > > > > > > >
> > > `ESP_HAL_3RDPARTY_VERSION=b4c723a119344b4b71d69819019d55637fb570fd
> > > > > > > > ESP_HAL_3RDPARTY_URL="g...@github.com:
> > > > tmedicci/esp-hal-3rdparty.git"`
> > > > > > env
> > > > > > > > vars (globally or before the make command) to use your
> version
> > of
> > > > the
> > > > > > > HAL.
> > > > > > > >
> > > > > > > > I just wanted to share some thoughts: using the HAL enables
> us
> > to
> > > > > make
> > > > > > > the
> > > > > > > > feature available for all the Espressif devices (please don't
> > > > bother
> > > > > > with
> > > > > > > > that now, as soon as you submit it, we can test for the other
> > > > devices
> > > > > > and
> > > > > > > > eventually make any adjustments). Although it's possible to
> > > > > reimplement
> > > > > > > > these functions on NuttX, we don't recommend that because of
> > the
> > > > > first
> > > > > > > > statement (the ability to use for the other devices): in that
> > > case
> > > > we
> > > > > > > > wouldn't be able to support the feature. You can take a look
> on
> > > IDF
> > > > > for
> > > > > > > > some inspiration on how it's implemented.
> > > > > > > >
> > > > > > > > Please let me know if you have any questions.
> > > > > > > >
> > > > > > > > Best regards,
> > > > > > > >
> > > > > > > > Em qua., 23 de out. de 2024 às 12:38, Felipe Moura Oliveira <
> > > > > > > > moura....@gmail.com> escreveu:
> > > > > > > >
> > > > > > > >> Hello everyone,
> > > > > > > >>
> > > > > > > >> Our project has reached a point where we need the Power
> > Manager
> > > > > > > >> functionality, which is not yet available in the ESP32C6. I
> am
> > > > > > studying
> > > > > > > to
> > > > > > > >> start the implementation, but I am confused about the
> correct
> > > way
> > > > to
> > > > > > > >> approach this.
> > > > > > > >>
> > > > > > > >> There is a Power Manager implementation for the ESP32C3, but
> > it
> > > > > seems
> > > > > > to
> > > > > > > >> follow the legacy methodology. In fact, the file
> > esp32c3_idle.c
> > > is
> > > > > > > located
> > > > > > > >> in the path arch/risc-v/src/esp32c3-legacy. In this
> > methodology,
> > > > > > > functions
> > > > > > > >> are re-implemented from the Espressif driver. For example,
> the
> > > > > > function
> > > > > > > >> esp32c3_light_sleep_start(), which is currently available in
> > > > > > > >> esp32c6/esp-hal-3rdparty/.../sleep_modes.c.
> > > > > > > >>
> > > > > > > >> I tried using the implementations from esp-hal-3rdparty to
> > > develop
> > > > > the
> > > > > > > >> Power Manager for the ESP32C6 in NuttX, but the files still
> > have
> > > > > many
> > > > > > > >> dependencies on items that are not present in NuttX.
> > > > > > > >>
> > > > > > > >> My question is: Should I try to make the Power Manager work
> > > using
> > > > > the
> > > > > > > >> functions in esp-hal-3rdparty, even if it requires making
> > > changes
> > > > to
> > > > > > the
> > > > > > > >> 3rd-party code? Or should I re-implement the functions
> within
> > > > NuttX,
> > > > > > > even
> > > > > > > >> if, in this case, there are functions with the same scope
> and
> > > > > > > >> implementation in different files, where one is compiled,
> and
> > > the
> > > > > > other
> > > > > > > is
> > > > > > > >> not?
> > > > > > > >>
> > > > > > > >>
> > > > > > > >> --
> > > > > > > >> *Felipe Moura de Oliveira*
> > > > > > > >> *Universidade Federal de Minas Gerais*
> > > > > > > >> Linkedin <
> > https://www.linkedin.com/in/felipe-oliveira-75a651a0>
> > > > > > > >> <https://twitter.com/FelipeMOliveir?lang=pt-br>
> > > > > > > >>
> > > > > > > >
> > > > > > > >
> > > > > > > > --
> > > > > > > > Tiago Medicci Serrano
> > > > > > > >
> > > > > > > > Embedded Software Engineer
> > > > > > > > MSc Electronics/Microelectronics
> > > > > > > > m: +55 (19) 981403886 <+55+(19)+981403886>
> > > > > > > > e: tiago.medi...@gmail.com
> > > > > > > > a: Campinas, Brazil
> > > > > > > > Follow me:
> > > > > > > > <https://www.linkedin.com/in/tiago-serrano-924458b6>
> > > > > > > > <https://github.com/tmedicci>
> > > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > --
> > > > > > > Tiago Medicci Serrano
> > > > > > >
> > > > > > > Embedded Software Engineer
> > > > > > > MSc Electronics/Microelectronics
> > > > > > > m: +55 (19) 981403886 <+55+(19)+981403886>
> > > > > > > e: tiago.medi...@gmail.com
> > > > > > > a: Campinas, Brazil
> > > > > > > Follow me:
> > > > > > > <https://www.linkedin.com/in/tiago-serrano-924458b6>
> > > > > > > <https://github.com/tmedicci>
> > > > > > >
> > > > > >
> > > > > >
> > > > > > --
> > > > > > *Felipe Moura de Oliveira*
> > > > > > *Universidade Federal de Minas Gerais*
> > > > > > Linkedin <https://www.linkedin.com/in/felipe-oliveira-75a651a0>
> > > > > > <https://twitter.com/FelipeMOliveir?lang=pt-br>
> > > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Tiago Medicci Serrano
> > > > >
> > > > > Embedded Software Engineer
> > > > > MSc Electronics/Microelectronics
> > > > > m: +55 (19) 981403886 <+55+(19)+981403886>
> > > > > e: tiago.medi...@gmail.com
> > > > > a: Campinas, Brazil
> > > > > Follow me:
> > > > > <https://www.linkedin.com/in/tiago-serrano-924458b6>
> > > > > <https://github.com/tmedicci>
> > > > >
> > > >
> > > >
> > > > --
> > > > *Felipe Moura de Oliveira*
> > > > *Universidade Federal de Minas Gerais*
> > > > Linkedin <https://www.linkedin.com/in/felipe-oliveira-75a651a0>
> > > > <https://twitter.com/FelipeMOliveir?lang=pt-br>
> > > >
> > >
> > >
> > > --
> > > Tiago Medicci Serrano
> > >
> > > Embedded Software Engineer
> > > MSc Electronics/Microelectronics
> > > m: +55 (19) 981403886 <+55+(19)+981403886>
> > > e: tiago.medi...@gmail.com
> > > a: Campinas, Brazil
> > > Follow me:
> > > <https://www.linkedin.com/in/tiago-serrano-924458b6>
> > > <https://github.com/tmedicci>
> > >
> >
> >
> > --
> > *Felipe Moura de Oliveira*
> > *Universidade Federal de Minas Gerais*
> > Linkedin <https://www.linkedin.com/in/felipe-oliveira-75a651a0>
> > <https://twitter.com/FelipeMOliveir?lang=pt-br>
> >
>
--
*Felipe Moura de Oliveira*
*Universidade Federal de Minas Gerais*
Linkedin <https://www.linkedin.com/in/felipe-oliveira-75a651a0>
<https://twitter.com/FelipeMOliveir?lang=pt-br>
