Damien George
Damien George
> Im also happy to take a look this afternoon and post a second PR Great, thanks, I'll leave it to you. > I also have a small test script...
See related #16770.
Zephyr 4.0.0 is now supported, see #16879.
Thanks for raising this PR. This definitely needs to be fixed, and the approach here looks reasonable, although I'd like to test it myself.
@TheMDev can you comment how this PR compares with the approach in #373?
> I think I might prefer if instead of just simplifying each instance, we actually moved the logic to a separate `_is_valid_timeout(timeout)` static method Yes, I think that's a good...
> Would a PR that bakes in the extensions set an MPY file is built for be accepted? > > I thought about adding an extra raw code/sub-code element with...
@agatti this PR in general looks good, and to make progress with it let's put aside the issue of RISC-V extensions, including floating point. That's anyway being discussed in #17635....
Trying to build this locally I get a few errors: ``` /usr/lib/gcc/riscv64-elf/14.1.0/rv64imac/lp64/libgcc.a(_clzsi2.o): in function `__clzdi2': /build/riscv64-elf-gcc/src/gcc-14.1.0/libgcc/libgcc2.c:690:(.text.__clzdi2+0x18): relocation truncated to fit: R_RISCV_HI20 against symbol `__clz_tab' defined in .rodata.__clz_tab section in /usr/lib/gcc/riscv64-elf/14.1.0/rv64imac/lp64/libgcc.a(_clz.o)...
OK, thanks for the information about medany. It's unfortunate that RV64 is awkward in this way. Is it possible to pass a flag to qemu to change the memory layout?...