Closed t-moe closed 5 months ago
Hei @japaric or @Urhengulas Can you look at this draft?
I've removed the udf
instruction to trigger a hard fault and instead use semihosting.
(To my knowledge, riscv has no such thing as an udf
instruction, to manually trigger a hardfault.)
That way we can get support for riscv and later a lot more platforms (see https://github.com/taiki-e/semihosting#platform-support).
Sorry this is taking so long. Unfortunately we can't look at this before the week of December 11.
Hi! Sorry for the delay. The draft looks good in principle - just a few comments.
There are some some CI failures to look at - a loop that doesn't loop, and a changelog update that isn't in the right format.
Hi! What is the status on this PR?
@SergioGasquez I'll probably abandon this, as I have created embedded-test now which is much more under our control. Also I feel that the testing framework shouldn't be coupled to whether you use defmt or not.
I'll also create panic_semihosting_abort which does what I tried to do in this PR with panic_probe.
Abandoning this, as I want to focus on embedded-test (which is now part of probe-rs)
This PR adds riscv32 support to defmt-test and panic-probe. Additionally it adds the
print-log
feature to panic-probe.In a separate PR I will:
use cortex_m_rt as _;
(as it is no longer included in defmt-test)