On Wed, 2024-12-11 at 09:23 +0800, Yan Zhao wrote:
> On Mon, Dec 02, 2024 at 05:03:14PM -0800, Rick Edgecombe wrote:
> ...
> > +u64 tdh_phymem_page_wbinvd_tdr(struct tdx_td *td)
> > +{
> > + struct tdx_module_args args = {};
> > +
> > + args.rcx = tdx_tdr_pa(td) | ((u64)tdx_global_keyid << boot_cpu_data.x86_phys_bits);
> > +
> > + return seamcall(TDH_PHYMEM_PAGE_WBINVD, &args);
> > +}
> > +EXPORT_SYMBOL_GPL(tdh_phymem_page_wbinvd_tdr);
> The tdx_global_keyid is of type u16 in TDX spec and TDX module.
> As Reinette pointed out, u64 could cause overflow.
>
> Do we need to change all keyids to u16, including those in
> tdh.mng.create() in patch 2,
> the global_keyid, tdx_guest_keyid_start in arch/x86/virt/vmx/tdx/tdx.c
> and kvm_tdx->hkid in arch/x86/kvm/vmx/tdx.c ?
It seems like a good idea.
>
> BTW, is it a good idea to move set_hkid_to_hpa() from KVM TDX to x86 common
> header?
>
> static __always_inline hpa_t set_hkid_to_hpa(hpa_t pa, u16 hkid)
> {
> return pa | ((hpa_t)hkid << boot_cpu_data.x86_phys_bits);
> }
Ah, yep.
