On Thu, Dec 1, 2022 at 6:06 PM Joel Sherrill <j...@rtems.org> wrote: > > I'm going to say enthusiastically yes. > > I suspect that other boards will be very similar so this is likely a good > basis to go from. > > If you have links where it is actually for sale and not just described, that > would be great. Seeed Studio: https://www.seeedstudio.com/Sipeed-Maixduino-Kit-for-RISC-V-AI-IoT-p-4047.html?queryID=86b62e47589a0b56236a5c29ab704818&objectID=4047&indexName=bazaar_retailer_products They have a number of other boards and modules, but most have not been in stock for a while. I'm not sure if it's a supply chain problem like the Raspberry Pi, or they are no longer available. Also here: https://www.dfrobot.com/product-1985.html https://www.dfrobot.com/product-1972.html Boards and modules are also available on Aliexpress. > > Chris would likely encourage you to post rtems-tester results. :) I have been using the Rednode.io test framework. I'm currently running a dozen tests, but I would like to try using rtems-tester.
> > And perhaps we can reach out and let them know RTEMS is available on their > hardware. Chris might enjoy poking that bear. lol I'm sure they would not mind. There are ports of rt-thread, nuttx, and no-mmu linux to the SoC. Kendrtye offers a bare metal SDK with the Apache 2.0 license: https://github.com/kendryte/kendryte-standalone-sdk They did have a FreeRTOS SDK, but it is no longer maintained: https://github.com/kendryte/kendryte-freertos-sdk I don't think there are any binary blobs needed, and an RTEMS image can just be flashed to the boards using a "kflash.py" utility that is available through pip, so getting started is easy: Just build RTEMS images and either run on renode or flash via USB cable. Alan > > On Thu, Dec 1, 2022, 1:46 PM Alan Cudmore <alan.cudm...@gmail.com> wrote: >> >> Hi, >> I have been working on a basic BSP for the Kendryte K210 RISC-V CPU, >> and I was wondering if the community members would like me to submit >> it. >> The Kendryte K210 is a dual core 64 bit RISC-V processor with a wealth >> of peripheral I/O, a built-in AI NPU, and 8 Megabytes of on-chip SRAM. >> I like it because it is one of the lowest cost RISC-V CPUs available, >> and it appears to run RTEMS well. >> In addition, my BSP works on the K210 model on the renode.io >> simulator. I believe it would work on QEMU, but it is very close to >> the rv64imafdc_medany riscv BSP variant. >> The changes consist of: >> - A new riscv/riscv variant and associated build option files >> - A new DTB header, since the DTB is included in the BSP similar to >> the polarfire BSP >> - Some code to detect the frequency in bspstart.c >> - A new header file for the k210 >> Because the console uses the same sifive uart as the frdme310arty BSP, >> I factored that out so the Sifive UART can be used in multiple BSP >> variants. It is able to use the existing timer and interrupt code. >> >> In addition to the renode.io simulator, I have run it on the following >> boards: >> - Sipeed MAIX Bit >> - Sipeed MAIXduino (arduino form factor board with ESP32) >> - Sipeed Grove AI hat for Raspberry Pi >> >> Potential negatives: >> - I do not have a BSD licensed device tree source, I created the >> device tree binary from the u-boot distribution. Is it OK to just >> include the device tree binary (similar to the microblaze)? >> - The availability of these boards has not been as good for the last >> year or so, but you can still find them at a relatively low cost. The >> Sipeed company seems to be focusing on another low cost RISC-V SoC, >> which is very interesting as well: The Bouffalo Lab BL808 (example >> here: >> https://wiki.pine64.org/wiki/Ox64) >> >> Is this worth submitting? I don't want to clutter up the tree with >> devices that may become obsolete - we could focus on the upcoming >> round of low cost RISC-V SoCs like the BL808. >> I don't have a specific application I am using it for, but I used it >> as a very inexpensive way to learn RISC-V on a real board. It may be >> of some value to integrate additional peripheral support including the >> AI NPU. >> >> If anyone is interested, I can submit the patches or even provide a >> branch on github. >> >> Thanks, >> Alan >> _______________________________________________ >> devel mailing list >> devel@rtems.org >> http://lists.rtems.org/mailman/listinfo/devel _______________________________________________ devel mailing list devel@rtems.org http://lists.rtems.org/mailman/listinfo/devel
