5 ms·
QMK Firmware – An open source firmware for AVR and ARM based keyboards
- sylware 4y agoI would write an assembly-only 64bits RISC-V firmware for keyboards, that without abusing the macro preprocessor of the assembler, with a nearly 0 SDK.
- johndoe0815 4y agoNot a lot of fun if you have to implement a protocol stack for a USB device...
- sylware 4y agoUpfront work is upfront work, and you can "hand compile-ish" linux/BSD usb protocol stacks. I am willing to do it because... no toxic IP "à la" MPEG/ARM ISA tied to RISC-V ISA. But I would target 64bits RISC-V instead of 32bits RISC-V, even though RISC-V 64bits<->32bits seems very good. Now, I need mini-boards with tons of GPIOs (really a lot), a 64bits RISC-V SOC with that USB-C connector for power and data. (I have some, but 32bits RISC-V :'( ).
- johndoe0815 4y agoThere are only a few microcontroller-style RV64 chips on the market. The Kendryte/canaan K210 is probably complete overkill for a keyboard controller (dual 64 bit cores, FPU, neural accelerator, 8 MB SRAM, 48 I/Os) but might still be interesting. Unfortunately, the cores implement an outdated version of the RISC-V specs (different CSRs, different virtual memory management) and the chip itself is only available in a BGA144 package, but there are quite a lot of inexpensive boards using the K210 available, e.g. from Sipeed.
- sylware 4y agoI had a look at the k210 and indeed, completely overkill, but the other RV64 boards are even more overkill. I still have faith though :)