Signed-off-by: Sean Christopherson <sean.j.christopherson(a)intel.com&gt; --- drivers/platform/x86/intel_sgx/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/platform/x86/intel_sgx/Kconfig b/drivers/platform/x86/intel_sgx/Kconfig index 9641728..442a8ef 100644 --- a/drivers/platform/x86/intel_sgx/Kconfig +++ b/drivers/platform/x86/intel_sgx/Kconfig @@ -5,7 +5,7 @@ config INTEL_SGX tristate "Intel(R) SGX Driver" default n - depends on X86 + depends on CPU_SUP_INTEL select MMU_NOTIFIER ---help--- Intel(R) SGX is a set of CPU instructions that can be used by -- 2.7.4 _______________________________________________ intel-sgx-kernel-dev mailing list intel-sgx-kernel-dev(a)lists.01.org https://lists.01.org/mailman/listinfo/intel-sgx-kernel-dev

Jarkko Sakkinen <jarkko.sakkinen(a)linux.intel.com&gt; wrote: > On Thu, Jul 20, 2017 at 12:53:43PM -0700, Sean Christopherson wrote: > > Signed-off-by: Sean Christopherson <sean.j.christopherson(a)intel.com&gt; > > --- > > drivers/platform/x86/intel_sgx/Kconfig | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/drivers/platform/x86/intel_sgx/Kconfig b/drivers/platform/x86/intel_sgx/Kconfig > > index 9641728..442a8ef 100644 > > --- a/drivers/platform/x86/intel_sgx/Kconfig > > +++ b/drivers/platform/x86/intel_sgx/Kconfig > > @@ -5,7 +5,7 @@ > > config INTEL_SGX > > tristate "Intel(R) SGX Driver" > > default n > > - depends on X86 > > + depends on CPU_SUP_INTEL > > select MMU_NOTIFIER > > ---help--- > > Intel(R) SGX is a set of CPU instructions that can be used by > > -- > > 2.7.4 > > > > _______________________________________________ > > intel-sgx-kernel-dev mailing list > > intel-sgx-kernel-dev(a)lists.01.org > > https://lists.01.org/mailman/listinfo/intel-sgx-kernel-dev > > > We need to depend on X86_64 for first upstream version because > > 1. My in-kernel LE supports only 64-bit *initially* for *simplicity*. It > can be extended later on. > 2. There is no a clean way to setup PAT when SGX driver is running in > 32-bit mode because symbols like reserve_memtype() are not exported > for kernel modules [1]. The best way to support 32-bit would be to > use [2] but it needs to be extended to support WB. Right now it is > only used by DRM and therefore has only WC-support. For simplicity, > it is better to clean the hacks for 32-bit environment and move > later on to [2]. > > Because of these reasons I think the line should be > > depends on CPU_SUP_INTEL && X86_64 > > [1] arch/x86/mm/pat.c > [2] include/linux/io-mapping.h > > /Jarkko > On Thu, Jul 20, 2017 at 12:53:43PM -0700, Sean Christopherson wrote: > > Signed-off-by: Sean Christopherson <sean.j.christopherson(a)intel.com&gt; > > --- > > drivers/platform/x86/intel_sgx/Kconfig | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/drivers/platform/x86/intel_sgx/Kconfig b/drivers/platform/x86/intel_sgx/Kconfig > > index 9641728..442a8ef 100644 > > --- a/drivers/platform/x86/intel_sgx/Kconfig > > +++ b/drivers/platform/x86/intel_sgx/Kconfig > > @@ -5,7 +5,7 @@ > > config INTEL_SGX > > tristate "Intel(R) SGX Driver" > > default n > > - depends on X86 > > + depends on CPU_SUP_INTEL > > select MMU_NOTIFIER > > ---help--- > > Intel(R) SGX is a set of CPU instructions that can be used by > > -- > > 2.7.4 > > > > _______________________________________________ > > intel-sgx-kernel-dev mailing list > > intel-sgx-kernel-dev(a)lists.01.org > > https://lists.01.org/mailman/listinfo/intel-sgx-kernel-dev > > > We need to depend on X86_64 for first upstream version because > > 1. My in-kernel LE supports only 64-bit *initially* for *simplicity*. It > can be extended later on. > 2. There is no a clean way to setup PAT when SGX driver is running in > 32-bit mode because symbols like reserve_memtype() are not exported > for kernel modules [1]. The best way to support 32-bit would be to > use [2] but it needs to be extended to support WB. Right now it is > only used by DRM and therefore has only WC-support. For simplicity, > it is better to clean the hacks for 32-bit environment and move > later on to [2]. > > Because of these reasons I think the line should be > > depends on CPU_SUP_INTEL && X86_64 > > [1] arch/x86/mm/pat.c > [2] include/linux/io-mapping.h > > /Jarkko I don't think we have to worry about the PAT, my understanding is that the memtype from the PRMRR overrides everything, e.g. PAT, EPT, MTRR...