historical-speck.lore.kernel.org archive mirror
 help / color / mirror / Atom feed
* [MODERATED] [qemu PATCH] mds
@ 2019-03-04 11:04 Paolo Bonzini
  0 siblings, 0 replies; only message in thread
From: Paolo Bonzini @ 2019-03-04 11:04 UTC (permalink / raw)
  To: speck

md-clear is a new CPUID bit which is set when microcode provides the
mechanism to invoke a flush of various exploitable CPU buffers by invoking
the VERW instruction.  Add the new feature, and pass it down to
Hypervisor.framework guests.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
	The last hunk is only needed for OS X, but anyway this is going
	to be the patch that will be committed upstream.

 target/i386/cpu.c           | 2 +-
 target/i386/cpu.h           | 1 +
 target/i386/hvf/x86_cpuid.c | 3 ++-
 3 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/target/i386/cpu.c b/target/i386/cpu.c
index d990070c59..16da90562c 100644
--- a/target/i386/cpu.c
+++ b/target/i386/cpu.c
@@ -1075,7 +1075,7 @@ static FeatureWordInfo feature_word_info[FEATURE_WORDS] = {
         .feat_names = {
             NULL, NULL, "avx512-4vnniw", "avx512-4fmaps",
             NULL, NULL, NULL, NULL,
-            NULL, NULL, NULL, NULL,
+            NULL, NULL, "md-clear", NULL,
             NULL, NULL, NULL, NULL,
             NULL, NULL, NULL, NULL,
             NULL, NULL, NULL, NULL,
diff --git a/target/i386/cpu.h b/target/i386/cpu.h
index 26412f15eb..cbfab1a421 100644
--- a/target/i386/cpu.h
+++ b/target/i386/cpu.h
@@ -692,6 +692,7 @@ typedef uint32_t FeatureWordArray[FEATURE_WORDS];
 
 #define CPUID_7_0_EDX_AVX512_4VNNIW (1U << 2) /* AVX512 Neural Network Instructions */
 #define CPUID_7_0_EDX_AVX512_4FMAPS (1U << 3) /* AVX512 Multiply Accumulation Single Precision */
+#define CPUID_7_0_EDX_MD_CLEAR      (1U << 10) /* Microarchitectural Data Clear */
 #define CPUID_7_0_EDX_SPEC_CTRL     (1U << 26) /* Speculation Control */
 #define CPUID_7_0_EDX_ARCH_CAPABILITIES (1U << 29)  /*Arch Capabilities*/
 #define CPUID_7_0_EDX_SPEC_CTRL_SSBD  (1U << 31) /* Speculative Store Bypass Disable */
diff --git a/target/i386/hvf/x86_cpuid.c b/target/i386/hvf/x86_cpuid.c
index 9874a46e92..f76ba50424 100644
--- a/target/i386/hvf/x86_cpuid.c
+++ b/target/i386/hvf/x86_cpuid.c
@@ -103,7 +103,8 @@ uint32_t hvf_get_supported_cpuid(uint32_t func, uint32_t idx,
             }
 
             ecx &= CPUID_7_0_ECX_AVX512BMI | CPUID_7_0_ECX_AVX512_VPOPCNTDQ;
-            edx &= CPUID_7_0_EDX_AVX512_4VNNIW | CPUID_7_0_EDX_AVX512_4FMAPS;
+            edx &= CPUID_7_0_EDX_AVX512_4VNNIW | CPUID_7_0_EDX_AVX512_4FMAPS | \
+                   CPUID_7_0_EDX_MD_CLEAR;
         } else {
             ebx = 0;
             ecx = 0;
-- 
2.20.1

^ permalink raw reply related	[flat|nested] only message in thread

only message in thread, other threads:[~2019-03-04 11:04 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-03-04 11:04 [MODERATED] [qemu PATCH] mds Paolo Bonzini

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).