On Wed, 20 Jan 2021 15:27:27 -0800 Dave Hansen wrote: > On 1/20/21 2:36 PM, Sean Christopherson wrote: > > On Wed, Jan 20, 2021, Dave Hansen wrote: > >> BTW, CONFIG_X86_SGX_VIRTUALIZATION is a pretty porky name. Maybe just > >> CONFIG_X86_SGX_VIRT? > > Mmm, bacon. I used the full "virtualization" to avoid any possible confusion > > with virtual memory. The existing sgx_get_epc_virt_addr() in particular gave me > > pause. > > > > I agree it's long and not consistent since other code in this series uses "virt". > > My thinking was that most shortand versions, e.g. virt_epc, would be used only > > in contexts that are already fairly obvious to be KVM/virtualization related, > > whereas the porcine Kconfig would help establish that context. > > Not a big deal either way. I agree that "virt" can be confusing. > > Considering that: > > +config X86_SGX_VIRTUALIZATION > + depends on ... KVM_INTEL It is already in patch 3: Introduce virtual EPC for use by KVM guests: +config X86_SGX_VIRTUALIZATION + bool "Software Guard eXtensions (SGX) Virtualization" + depends on X86_SGX && KVM_INTEL > > Calling it X86_SGX_KVM doesn't seem horrible either.
