mcuee

@tormodvolden Now that libusb 1.0.28 has been released. Maybe this can be merged.

@ndim The default value is in the code since the following commit. * https://github.com/avrdudes/avrdude/commit/dd1255b0c8fcf24e0cf49d04d5cda74785b450c9 Please also refer to PR #933. * #933 But I agree this can be a good...

I have to specify the linuxspi port since the default value does not work. * https://github.com/avrdudes/avrdude/discussions/1654#discussioncomment-10413582 ``` mcuee@raspberrypi400arm64:~/build/avrdude_git/tools $ ./test-avrdude -d 0 -p "-c linuxspi -P /dev/spidev0.1:/dev/gpiochip0:25 -p m328p" Testing...

Current code: maybe it needs to specify the default reset pin number. https://github.com/avrdudes/avrdude/blob/30e19f23324c9337951d6277e054f269690ce718/src/linuxspi.c#L146-L151

> I think we should replace the `@DEFAULT_SPI_PORT@` during the make process, perhaps with the default value `/dev/spidev0.0:/dev/gpiochip0` form the linuxspi.c code and refrain from setting it in the code....

@tormodvolden I think this PR can be changed from draft and ready to be merged.

I tend to think just ignore S390x as no one would probably use avrdude on an S390x machine. But maybe it is interesting to find the root cause.

Closed as not-planned then.

> Would it make sense if the build aborted when it detects it is about to build avrdude for a non little endian system? Sounds like a good idea.

No so sure if there are still relevant embedded Linux system which use big endian system. https://en.wikipedia.org/wiki/Endianness > Some instruction set architectures are "bi-endian" and allow running software of either...