@haui

I couldn't resist to have a quick look into the device tree file(s) in the kernel sources of Bianbu Linux:

see here: https://gitee.com/bianbu-linux/linux-6.6/blob/bl-v2.0.y/arch/riscv/boot/dts/spacemit/k1-x.dtsi

On a quick look (didn't check entry by entry) the 'riscv,isa-extensions' entries per cpu seems to match your table.

Anyway as I said before, I think this does not need to reflect the real capabilities of the SOC too.

#riscv #spacemit #spacemitk1 #spacemitx60 #devicetree #linux #bananapif3 #bpif3

@haui

Regarding this I found an article from RedHat:

see https://research.redhat.com/blog/article/risc-v-extensions-whats-available-and-how-to-find-it/

According to the article the DeviceTree can be used to get that information, but since that's just a configuration describing the hardware it could be incomplete too.

#riscv #spacemit #spacemitk1 #bananapif3 #bpif3 #linux #devicetree

@haui

I do not think the SOC is missing those extensions.

See here: https://www.kernel.org/doc/html/v6.12-rc4/arch/riscv/uabi.html

According to the kernel documentation about isa lines in /proc/cpuinfo:

"... the absence of an extension in these lines does not necessarily mean the hardware does not support that feature. The running kernel may not recognize the extension, or may have deliberately removed it from the listing."

#riscv #spacemit #spacemitk1 #bananapif3 #bpif3 #linux #kernel