bhyve/riscv kernel part

[br]

Summary

Basic VMM support for RISC-V.

Lots of TODOs indicated within the code, but working for me on single core, single guest VM instance.

Page: https://wiki.freebsd.org/riscv/bhyve

Test Plan

tested in Spike

./spike -m4096 -d --isa RV64IMAFDCH_zicntr_zihpm_sstc --kernel /usr/obj/usr/home/br/dev/freebsd/riscv.riscv64/sys/GENERIC/kernel.bin ../../opensbi/build/platform/generic/firmware/fw_jump.elf

and then

bhyve -m 2560 -o bootrom=/kernel.bin -o console=stdio -s 4,ahci-hd,/bin/ls test

(in both host&guest kernels I have a small mdroot embedded)

[jrtc27]

Without wanting to give a detailed review of the whole thing, I will say that implementing SBI should be done in userspace, it doesn't need to be in the kernel (and really needs to be in userspace to be able to handle the console I/O calls), just like arm64 has PSCI in userspace. You want to just have it send up a VM_EXITCODE_ECALL.

[br]

In D45553#1039259, @jrtc27 wrote:

Without wanting to give a detailed review of the whole thing, I will say that implementing SBI should be done in userspace, it doesn't need to be in the kernel (and really needs to be in userspace to be able to handle the console I/O calls), just like arm64 has PSCI in userspace. You want to just have it send up a VM_EXITCODE_ECALL.

Good idea. Agree, we can move it to userspace.

[br]

• fix macro HSTATUS_VGEIN_S
• rename vmm_call_hyp() to vmm_switch() for world switch

[br]

Handle entire SBI in userspace based on jrtc27@ suggestion. We may need however to handle the IPI or TIME (in case of no SSTC) extensions here in kernel for speed (?), so leave the no-op vmm_sbi.c file for now for debugging purposes.

[mhorne]

Hi, just a preliminary look, much of the vmm code is beyond me. I hope to test this stuff in a couple weeks!

Two things which could be extracted and committed now:

• Addition of new SBI implementation IDs
• Addition of fpe.c. I am hoping to tackle kernel FPU context soon, which can make use of what you have added here.

sys/riscv/include/vmm.h

2–31

Optional, but lately it seems we are encouraging new copyright additions to include proper SPDX tag and omit the boilerplate.

sys/riscv/riscv/fpe.c

178 ↗

(On Diff #139701)

sys/riscv/vmm/vmm_ktr.h

70

I guess there is still no shared directory for common vmm code?

sys/riscv/vmm/vmm_riscv.c

406

Seems like handled should just be a bool.

sys/riscv/vmm/vmm_switch.S

203

vmm_switch LGTM.

[mhorne]

share/mk/bsd.cpu.mk

354 ↗

(On Diff #139929)

This one is for userspace, and we might not even need to enable 'h' there?

[br]

Regenerate

[br]

Address comments

share/mk/bsd.cpu.mk

354 ↗

(On Diff #139929)

Right we don't need this for userspace. Thanks.

sys/riscv/vmm/vmm_ktr.h

70

no such directory yet

sys/riscv/vmm/vmm_riscv.c

406

Fixed!