On Fri, May 15, 2026, Ackerley Tng wrote: > Sean Christopherson <[email protected]> writes: > > > > > [...snip...] > > > >> +#ifdef CONFIG_USERFAULTFD > >> +static bool kvm_gmem_can_userfault(struct vm_area_struct *vma, vm_flags_t > >> vm_flags) > >> +{ > >> + struct inode *inode = file_inode(vma->vm_file); > >> + > >> + /* > >> + * Only support userfaultfd for guest_memfd with INIT_SHARED flag. > >> + * This ensures the memory can be mapped to userspace. > >> + */ > > Is the principle here that any memory that is allowed to be mapped to > userspace can be userfault-ed?
Yes? It's more the opposite though: memory that can't be SHARED and thus can't be mapped into userspace can't possibly support userfaultfd. Six of one, half dozen of the other, but I think it's important to highlight that there's no judgment call being made, i.e. we aren't deciding to support only SHARED just because we're lazy, it's a hard requirement.
