vmm: Start reconciling amd64 and arm64 copies of vmm_dev.c

[markj]

Summary

Most of the code in vmm_dev.c and vmm.c can and should be shared between
amd64 and arm64 (and eventually riscv) rather than being duplicated. To
the end of adding a shared implementation in sys/dev/vmm, this patch
eliminates most of the differences between the two copies of vmm_dev.c.

• Remove an unneeded cdefs.h include.
• Simplify the amd64 implementation of vcpu_unlock_one().
• Simplify the arm64 implementation of vcpu_lock_one().
• Pass buffer sizes to alloc_memseg() and get_memseg() on arm64. On amd64 this is needed for compat ioctls, but these functions should be merged.
• Make devmem_mmap_single() stricter on arm64.

[corvink]

sys/arm64/vmm/vmm_dev.c

274

Shouldn't we check that len <= mseg->name?

[markj]

sys/arm64/vmm/vmm_dev.c

274

In general it's up to the caller to pass the correct len, as mseg might be a struct vm_memseg or a struct vm_memseg_fbsd12.

Are you asking me to add KASSERT(len <= sizeof(mseg->name)) or so? That's fine, but I'd probably put that in a separate commit; this patch is just about minimizing differences between amd64 and arm64.

[corvink]

sys/arm64/vmm/vmm_dev.c

274

Yes, a kassert might be a good idea.

[markj]

sys/arm64/vmm/vmm_dev.c

274

Actually, that assertion would be wrong in the COMPAT12 case, struct vm_memseg_fbsd12's name field is larger than the one in struct vm_memseg. (That is a bit surprising really.) So, I think I prefer to leave the code as it is.