On 9/10/20 7:45 AM, Tom Lendacky wrote:
On 9/9/20 9:22 PM, Krish Sadhukhan wrote:
Some AMD hardware platforms enforce cache coherency across encryption
domains.
Add this hardware feature as a CPUID feature to the kernel.
Signed-off-by: Krish Sadhukhan <krish.sadhukhan@xxxxxxxxxx>
---
arch/x86/include/asm/cpufeatures.h | 1 +
arch/x86/kernel/cpu/amd.c | 3 +++
2 files changed, 4 insertions(+)
diff --git a/arch/x86/include/asm/cpufeatures.h
b/arch/x86/include/asm/cpufeatures.h
index 81335e6fe47d..0e5b27ee5931 100644
--- a/arch/x86/include/asm/cpufeatures.h
+++ b/arch/x86/include/asm/cpufeatures.h
@@ -293,6 +293,7 @@
#define X86_FEATURE_FENCE_SWAPGS_USER (11*32+ 4) /* "" LFENCE in
user entry SWAPGS path */
#define X86_FEATURE_FENCE_SWAPGS_KERNEL (11*32+ 5) /* "" LFENCE
in kernel entry SWAPGS path */
#define X86_FEATURE_SPLIT_LOCK_DETECT (11*32+ 6) /* #AC for
split lock */
+#define X86_FEATURE_HW_CACHE_COHERENCY (11*32+ 7) /* AMD
hardware-enforced cache coherency */
/* Intel-defined CPU features, CPUID level 0x00000007:1 (EAX),
word 12 */
#define X86_FEATURE_AVX512_BF16 (12*32+ 5) /* AVX512
BFLOAT16 instructions */
diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c
index 4507ededb978..698884812989 100644
--- a/arch/x86/kernel/cpu/amd.c
+++ b/arch/x86/kernel/cpu/amd.c
@@ -632,6 +632,9 @@ static void early_detect_mem_encrypt(struct
cpuinfo_x86 *c)
*/
c->x86_phys_bits -= (cpuid_ebx(CPUID_AMD_SME) >> 6) & 0x3f;
+ if (cpuid_eax(CPUID_AMD_SME) & 0x400)
+ set_cpu_cap(c, X86_FEATURE_HW_CACHE_COHERENCY);
Why not add this to arch/x86/kernel/cpu/scattered.c?
The reason why I put it in amd.c is because it's AMD-specific, though I
know we have SME and SEV in scattered.c. Shouldn't SME and SEV features
be ideally placed in AMD-specific files and scattered.c be used for
common CPUID features ?
Thanks,
Tom
+
if (IS_ENABLED(CONFIG_X86_32))
goto clear_all;
