Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Preliminary ESP32 port #517

Merged
merged 10 commits into from Jun 21, 2017
Merged

Preliminary ESP32 port #517

merged 10 commits into from Jun 21, 2017

Conversation

lpereira
Copy link
Collaborator

This patch set adds preliminary support for ESP32, both as a SoC, and as a board. I haven't compiled the documentation yet, but the individual commit messages contain much of what's necessary to build Zephyr for this board. I'm submitting this series now in order to gather feedback and raise awareness in the ESP32 community.

Most of the legwork for an Xtensa port has been made available in previous versions, so basic things such as threads, timers, and synchronization primitives are all working fine. The philosophers sample is working as expected, for instance.

A lot of things are missing from the port, though, including I/O (no GPIO, I2C, SPI, analog I/O, PWM, etc) and connectivity (WiFi, BT, and BLE). We're working on these, and I hope to have patches implementing some of them soon. Connectivity is going to take more work, though.

OpenOCD should work fine (as long as you're using the fork from Espressif).


echo "Flashing"
${esptool} --chip esp32 \
--port /dev/ttyUSB0 \
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

should provide a means to override the port setting, what happens if you had more than one board connected to a system?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

agree, my USB0 is always occupied by something else, I had to change that when testing ESP32.



cmd_flash() {
local esptool=${ESP_IDF_PATH}/components/esptool_py/esptool/esptool.py
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

should allow for esptool path to be set by user if desired, and default to what you have here.

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why?

EMU_PLATFORM ?=
DEBUG_SCRIPT :=
FLASH_SCRIPT := esp32.sh
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

do you expect esp32.sh to do more than flash? if not might rename script to esp32-flash.sh

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

the support scripts we have are designed to do more than flashing, so I think this is in line with other scripts we have, if we add debugging, then it is added to the same script.

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm planning to add a debug command to this script as well, to invoke OpenOCD with the right settings.

Copy link
Collaborator

@galak galak left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Some cleanups for esp32.sh script

@lpereira lpereira force-pushed the esp32 branch 2 times, most recently from 63e8be1 to 541d9c2 Compare June 16, 2017 17:41
@@ -0,0 +1,11 @@
# Kconfig - STM32 MINI A15 board configuration
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

STM32?

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good catch. Fixed.

lpereira and others added 10 commits June 19, 2017 12:50
Three environment variables must be set to use this variant:

   export ZEPHYR_GCC_VARIANT="espressif"
   export ESP_IDF_PATH=/path/to/esp-idf
   export ESPRESSIF_TOOLCHAIN_PATH=/path/to/xtensa-esp32-elf/

ESP-IDF is the SDK provided by Espressif.  It contains, among other things,
the HAL and header files for registers and ROM functions used by the Zephyr
port.  At this stage, with the exception of the HAL library, none of the
binary blobs provided by ESP-IDF are used.  This can be obtained directly
from Espressif, at <https://github.com/espressif/esp-idf>.

Instructions on how to obtain the toolchain are detailed in the README for
ESP-IDF.

Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
Due to the configurable nature of the Xtensa platform, the generic name of
"LX6" cannot be used to describe an SoC as far as Zephyr goes.  So ESP32 is
defined both as a SoC and as a board.

This is based on work by Rajavardhan Gundi.

Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
This is based on the work of Rajavardhan Gundi.

Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
This is a minimal driver enabling console output during the port
bringup.  While the driver works, only one of the three UART devices
are supported, and there isn't any way to change any parameters or
use interrupts.  This will most likely be superceded by a proper
driver after the port has matured.

Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
This header is included by some files provided by ESP-IDF.  Nothing
from this header file is actually used: it's only being added allow
things to compile with the minimal libc.

Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
Until ESP32’s flash cache is utilized, .rodata must be stored in RAM.

Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
Unconditionally use CONFIG_SIMULATOR_XTENSA to determine if XT_SIMULATOR
or XT_BOARD should be defined.

If CONFIG_SYS_CLOCK_HW_CYCLES_PER_SEC, also define XT_CLOCK_FREQ.  This
isn't ideal as the clock frequency might be changed in runtime and this
effectively makes it a constant.

Until we can control the clock frequency in runtime, this will suffice.

Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
The first stage bootloader, part of the ESP32 ROM, already sets up
a stack that's sufficient to execute C programs.  So, instead of
implementing __stack() in assembly, do it in C to simplify things
slightly.

This ESP32-specific initialization will perform the following:

  - Disable the watchdog timer that's enabled by the bootloader
  - Move exception handlers to IRAM
  - Disable normal interrupts
  - Disable the second CPU
  - Zero out the BSS segment

Things that might be performed in the future include setting up the
CPU frequency, memory protection regions, and enabling the flash
cache.

Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
This flashes Zephyr at 0x1000: that's where the first stage bootloader,
part of the ESP32 ROM, expects to find an "image header".

The second-stage bootloader, part of ESP-IDF, isn't used by the Zephyr
port.  However, the bootloader can be used if desired; please refer to
the ESP-IDF documentation on how to set up partitions tables and use
the bootloader.

The following environment variables will affect the ESP32 flashing
process:

  Variable              Default value
  ESP_DEVICE            /dev/ttyUSB0
  ESP_BAUD_RATE         921600
  ESP_FLASH_SIZE        detect
  ESP_FLASH_FREQ        40m
  ESP_FLASH_MODE        dio
  ESP_TOOL              espidf

It's impossible to determine which serial port the ESP32 board is
connected to, as it uses a generic RS232-USB converter.  The default of
/dev/ttyUSB0 is provided as that's often the assigned name on a Linux
machine without any other such converters.

The baud rate of 921600bps is recommended.  If experiencing issues when
flashing, try halving the value a few times (460800, 230400, 115200,
etc).  It might be necessary to change the flash frequency or the flash
mode; please refer to the esptool documentation for guidance on these
settings.

If ${ESP_TOOL} is set to "espidf", the esptool.py script found within
ESP-IDF will be used.  Otherwise, this variable is handled as a path to
the tool.

Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
@lpereira
Copy link
Collaborator Author

Rebased and fixed comments.

config BOARD_ESP32
bool "ESP32 Development Board"
depends on XTENSA
select BOARD_XTENSA
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ugh, at some point we should just use y/n values of CONFIG_SIMULATOR_XTENSA instead of having both CONFIG_SIMULATOR_XTENSA and CONFIG_BOARD_XTENSA

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is sort of done inside xtensa_rtos.h now: if CONFIG_SIMULATOR_XTENSA is defined, XT_SIMULATOR is defined; otherwise, XT_BOARD gets defined. This select can be removed from Kconfig.board.

flash)
cmd_flash "$@"
;;
*)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

do we plan on adding a 'debugserver' target?

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes.

/* The watchdog timer is enabled in the bootloader. We're done booting,
* so disable it.
*/
*wdt_rtc_reg &= ~RTC_CNTL_WDT_FLASHBOOT_MOD_EN;
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

suggest adding a story in the backlog to implement a WDT driver

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sure.

default esp32

config IRQ_OFFLOAD_INTNUM
default 7
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Assuming this value has been tested.
If unsure, run tests/kernel/irq_offload

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

IRQ offload isn't working yet. I've tried with two IRQ nos. that are available for software interrupts, and all I get is an exception. There are more pressing matters at the moment but this is something that have to be fixed eventually.

nashif
nashif approved these changes Jun 21, 2017
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

6 participants