On Tue, 2019-12-17 at 16:17 +0100, Borislav Petkov wrote: > On Sat, Nov 30, 2019 at 01:13:09AM +0200, Jarkko Sakkinen wrote: > > Typo in the subject: > > Subject: Re: [PATCH v24 07/24] x86/cpu/intel: Detect SGX supprt > ^^^^^^ > > From: Sean Christopherson <sean.j.christopherson@xxxxxxxxx> > > > > When the CPU supports SGX, check that the BIOS has enabled SGX and SGX1 > > opcodes are available. Otherwise, all the SGX related capabilities. > > > > In addition, clear X86_FEATURE_SGX_LC also in the case when the launch > > enclave are read-only. This way the feature bit reflects the level that > > Linux supports the launch control. > > > > The check is done for every CPU, not just BSP, in order to verify that > > MSR_IA32_FEATURE_CONTROL is correctly configured on all CPUs. The other > > parts of the kernel, like the enclave driver, expect the same > > configuration from all CPUs. > > > > Signed-off-by: Sean Christopherson <sean.j.christopherson@xxxxxxxxx> > > Co-developed-by: Jarkko Sakkinen <jarkko.sakkinen@xxxxxxxxxxxxxxx> > > Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@xxxxxxxxxxxxxxx> > > --- > > arch/x86/kernel/cpu/intel.c | 41 +++++++++++++++++++++++++++++++++++++ > > 1 file changed, 41 insertions(+) > > ... > > > @@ -761,6 +797,11 @@ static void init_intel(struct cpuinfo_x86 *c) > > if (cpu_has(c, X86_FEATURE_TME)) > > detect_tme(c); > > > > +#ifdef CONFIG_INTEL_SGX > > + if (cpu_has(c, X86_FEATURE_SGX)) > > + detect_sgx(c); > > +#endif > > You can remove the ifdeffery here and put the ifdef around the function > body and drop the __maybe_unused tag: OK, so I remember getting feedback on some unrelated to SGX stuff that I should use __maybe_unused in the past. That is why I used it here instead of ifdeffery. It is used in bunch of places in the kernel. I'm a bit confused when using it is a right thing and when it should be avoided. /Jarkko
