Messages in this thread | | | Date | Tue, 5 Oct 2021 13:21:46 -0700 | From | Josh Poimboeuf <> | Subject | Re: [PATCH v8 02/11] x86/tdx: Introduce INTEL_TDX_GUEST config option |
| |
On Tue, Oct 05, 2021 at 07:29:12PM +0200, Borislav Petkov wrote: > On Tue, Oct 05, 2021 at 07:48:25AM -0700, Kuppuswamy, Sathyanarayanan wrote: > > Since x2APIC will always exist in TDX guest case, we have used select to > > enable the support. But since we have dependency on it, I think "depends" > > might be a better choice. > > Right, and while we're on the subject, this looks silly to me too: > > + depends on SECURITY > ... > + select SECURITY_LOCKDOWN_LSM > > because > > Symbol: SECURITY_LOCKDOWN_LSM [=n] > │ Type : bool > │ Defined at security/lockdown/Kconfig:1 > │ Prompt: Basic module for enforcing kernel lockdown > │ Depends on: SECURITY [=n] > ^^^^^^^^^ > > so that symbol already depends on SECURITY. > > And I have SECURITY=n in my config so I still have to go select SECURITY > by hand so that CONFIG_INTEL_TDX_GUEST becomes visible. And when I > select it, SECURITY_LOCKDOWN_LSM gets enabled too. > > But since I have to go select SECURITY, I can just as well enable > SECURITY_LOCKDOWN_LSM in order to have TDX guest support. > > IOW, I don't see the point for the evil "select"s - just make everything > depends on and be done with it. > > Unless there's an aspect I'm missing...
It would also be helpful to explain the dependencies (particularly X86_X2APIC and SECURITY_LOCKDOWN_LSM) in the commit message.
-- Josh
| |