Embedded Working Group Meeting 2024-01-23

Attendance

Write your GH username or Matrix handle here!

  • adamgreig
  • GrantM11235
  • jannic
  • therealprof
  • bartmassey
  • eldruin
  • mabez
  • elpiel
  • newAM

Agenda

  • Announcements
    • svdtools 0.3.9 released
  • Embedded HAL
    • ADC/voltmeter/ammeter traits https://github.com/rust-embedded/embedded-hal/pull/569
      • Having volts and amps seems a bit arbitrary; should we also have temperature and light level and etc etc?
      • There might be some value in such a set of traits but it doesn't seem like they belong in e-h
      • Probably the original plan of an ADC trait is a better fit for embedded-hal, but there are design questions to resolve
    • SPI sharing with fallible CS pins https://github.com/rust-embedded/embedded-hal/issues/573
      • Perhaps the whole shared bus could be poisoned to reduce the runtime overhead
      • But given the rarity of requiring fallible pins and not being able to just wrap them in a panicking wrapper, perhaps option 1 remains the best option for e-h-bus, with users requiring more sophisticated sharing/error handling able to roll their own
      • @GrantM11235 to write a PR
  • svd2rust
  • Continue last week's discussion on interrupt macro syntax across crates

Last Week's Minutes

  • Announcements
    • embedded-hal v1.0.0 released!
      Image Not Showing Possible Reasons
      • The image file may be corrupted
      • The server hosting the image is unavailable
      • The image path is incorrect
      • The image format is not supported
      Learn More →
    • risc-v releases: risc-v 0.11, riscv-rt 0.12, riscv-pac 0.1, riscv-semihosting 0.1
    • linux-embedded-hal 0.4 was released with embedded-hal updated to 1.0
  • Commonality between cortex-m and riscv interrupt macros
    • Could we unify the interrupt-declaring syntax?
    • Currently riscv-rt just has the user export a suitably-named symbol, which would also work for c-m-rt
      • But this doesn't permit the &static mut transform, or ensure that the method can't be called except by hardware as an interrupt, and doesn't check the function type signature
      • We could recommend this for c-m-rt
    • Could c-m-rt's interrupt macro take the interrupt name as an (optional?) argument, so the function could then be named arbitrarily instead?
      • avr-device requires the chip name be passed to the interrupt macro, so couldn't immediately work with this syntax, but maybe could be refactored later to support it
    • What about exception macros (e.g. when used to provide the exceptionframe in a hardfault on c-m-rt) or the entry macro (which also provides the static mut transform)?
    • TBC
Select a repo