[PATCH v3] x86/cpufeatures.h: Enable A Few New AVX-512 Features

[PATCH v3] x86/cpufeatures.h: Enable A Few New AVX-512 Features

[Fenghua Yu]

From: Fenghua Yu <fenghua.yu@intel.com>

A few new AVX-512 instruction groups/features are added in cpufeatures.h
for enuermation: AVX512DQ, AVX512BW, and AVX512VL.

Clear the flags in fpu__xstate_clear_all_cpu_caps().

The specification for latest AVX-512 including the features can be found at
https://software.intel.com/sites/default/files/managed/07/b7/319433-023.pdf

Signed-off-by: Fenghua Yu <fenghua.yu@intel.com>
---

I don't enable the flags in kvm. Hopefully kvm guys can pick up the flags
and enable them in kvm.

v3: Clear the flags in fpu__xstate_clear_all_cpu_caps().
v2: Change cpufeature.h to cpufeatures.h to use right file name in latest
kernel.

 arch/x86/include/asm/cpufeatures.h | 3 +++
 arch/x86/kernel/fpu/xstate.c       | 3 +++
 2 files changed, 6 insertions(+)

diff --git a/arch/x86/include/asm/cpufeatures.h b/arch/x86/include/asm/cpufeatures.h
index 1a2811f..9ba2ca9 100644
--- a/arch/x86/include/asm/cpufeatures.h
+++ b/arch/x86/include/asm/cpufeatures.h
@@ -220,6 +220,7 @@
 #define X86_FEATURE_CQM		( 9*32+12) /* Cache QoS Monitoring */
 #define X86_FEATURE_MPX		( 9*32+14) /* Memory Protection Extension */
 #define X86_FEATURE_AVX512F	( 9*32+16) /* AVX-512 Foundation */
+#define X86_FEATURE_AVX512DQ	( 9*32+17) /* AVX-512 DQ Instructions */
 #define X86_FEATURE_RDSEED	( 9*32+18) /* The RDSEED instruction */
 #define X86_FEATURE_ADX		( 9*32+19) /* The ADCX and ADOX instructions */
 #define X86_FEATURE_SMAP	( 9*32+20) /* Supervisor Mode Access Prevention */
@@ -230,6 +231,8 @@
 #define X86_FEATURE_AVX512ER	( 9*32+27) /* AVX-512 Exponential and Reciprocal */
 #define X86_FEATURE_AVX512CD	( 9*32+28) /* AVX-512 Conflict Detection */
 #define X86_FEATURE_SHA_NI	( 9*32+29) /* SHA1/SHA256 Instruction Extensions */
+#define X86_FEATURE_AVX512BW	( 9*32+30) /* AVX-512 BW Instructions */
+#define X86_FEATURE_AVX512VL	( 9*32+31) /* AVX-512 Vector Lengths */
 
 /* Extended state features, CPUID level 0x0000000d:1 (eax), word 10 */
 #define X86_FEATURE_XSAVEOPT	(10*32+ 0) /* XSAVEOPT */
diff --git a/arch/x86/kernel/fpu/xstate.c b/arch/x86/kernel/fpu/xstate.c
index 1b19818..b48ef35 100644
--- a/arch/x86/kernel/fpu/xstate.c
+++ b/arch/x86/kernel/fpu/xstate.c
@@ -59,6 +59,9 @@ void fpu__xstate_clear_all_cpu_caps(void)
 	setup_clear_cpu_cap(X86_FEATURE_AVX512PF);
 	setup_clear_cpu_cap(X86_FEATURE_AVX512ER);
 	setup_clear_cpu_cap(X86_FEATURE_AVX512CD);
+	setup_clear_cpu_cap(X86_FEATURE_AVX512DQ);
+	setup_clear_cpu_cap(X86_FEATURE_AVX512BW);
+	setup_clear_cpu_cap(X86_FEATURE_AVX512VL);
 	setup_clear_cpu_cap(X86_FEATURE_MPX);
 	setup_clear_cpu_cap(X86_FEATURE_XGETBV1);
 	setup_clear_cpu_cap(X86_FEATURE_PKU);
-- 
2.5.0

Re: [PATCH v3] x86/cpufeatures.h: Enable A Few New AVX-512 Features

[Ingo Molnar]

* Fenghua Yu <fenghua.yu@intel.com> wrote:

> +#define X86_FEATURE_AVX512DQ	( 9*32+17) /* AVX-512 DQ Instructions */
> +#define X86_FEATURE_AVX512BW	( 9*32+30) /* AVX-512 BW Instructions */
> +#define X86_FEATURE_AVX512VL	( 9*32+31) /* AVX-512 Vector Lengths */

Yeah, so I don't think it's obvious to people what the DQ/BW/VL extensions are 
precisely, so I changed the text to the following, a bit more verbose 
descriptions:

  #define X86_FEATURE_AVX512DQ  ( 9*32+17) /* AVX-512 DQ (Double/Quad granular) Instructions */
  #define X86_FEATURE_AVX512BW	( 9*32+30) /* AVX-512 BW (Byte/Word granular) Instructions */
  #define X86_FEATURE_AVX512VL	( 9*32+31) /* AVX-512 VL (128/256 Vector Length) Extensions */

Please holler if you disagree!

Thanks,

	Ingo

[tip:x86/asm] x86/cpufeature: Enable new AVX-512 features

[tip-bot for Fenghua Yu]

Commit-ID:  d05004944206cbbf1c453e179768163731c7c6f1
Gitweb:     http://git.kernel.org/tip/d05004944206cbbf1c453e179768163731c7c6f1
Author:     Fenghua Yu <fenghua.yu@intel.com>
AuthorDate: Thu, 10 Mar 2016 19:38:18 -0800
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Sat, 12 Mar 2016 17:30:53 +0100

x86/cpufeature: Enable new AVX-512 features

A few new AVX-512 instruction groups/features are added in cpufeatures.h
for enuermation: AVX512DQ, AVX512BW, and AVX512VL.

Clear the flags in fpu__xstate_clear_all_cpu_caps().

The specification for latest AVX-512 including the features can be found at:

  https://software.intel.com/sites/default/files/managed/07/b7/319433-023.pdf

Note, I didn't enable the flags in KVM. Hopefully the KVM guys can pick up
the flags and enable them in KVM.

Signed-off-by: Fenghua Yu <fenghua.yu@intel.com>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Cc: Gleb Natapov <gleb@kernel.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Quentin Casasnovas <quentin.casasnovas@oracle.com>
Cc: Ravi V Shankar <ravi.v.shankar@intel.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: kvm@vger.kernel.org
Link: http://lkml.kernel.org/r/1457667498-37357-1-git-send-email-fenghua.yu@intel.com
[ Added more detailed feature descriptions. ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/include/asm/cpufeatures.h | 3 +++
 arch/x86/kernel/fpu/xstate.c       | 3 +++
 2 files changed, 6 insertions(+)

diff --git a/arch/x86/include/asm/cpufeatures.h b/arch/x86/include/asm/cpufeatures.h
index d11a3aa..9e0567f 100644
--- a/arch/x86/include/asm/cpufeatures.h
+++ b/arch/x86/include/asm/cpufeatures.h
@@ -220,6 +220,7 @@
 #define X86_FEATURE_CQM		( 9*32+12) /* Cache QoS Monitoring */
 #define X86_FEATURE_MPX		( 9*32+14) /* Memory Protection Extension */
 #define X86_FEATURE_AVX512F	( 9*32+16) /* AVX-512 Foundation */
+#define X86_FEATURE_AVX512DQ	( 9*32+17) /* AVX-512 DQ (Double/Quad granular) Instructions */
 #define X86_FEATURE_RDSEED	( 9*32+18) /* The RDSEED instruction */
 #define X86_FEATURE_ADX		( 9*32+19) /* The ADCX and ADOX instructions */
 #define X86_FEATURE_SMAP	( 9*32+20) /* Supervisor Mode Access Prevention */
@@ -230,6 +231,8 @@
 #define X86_FEATURE_AVX512ER	( 9*32+27) /* AVX-512 Exponential and Reciprocal */
 #define X86_FEATURE_AVX512CD	( 9*32+28) /* AVX-512 Conflict Detection */
 #define X86_FEATURE_SHA_NI	( 9*32+29) /* SHA1/SHA256 Instruction Extensions */
+#define X86_FEATURE_AVX512BW	( 9*32+30) /* AVX-512 BW (Byte/Word granular) Instructions */
+#define X86_FEATURE_AVX512VL	( 9*32+31) /* AVX-512 VL (128/256 Vector Length) Extensions */
 
 /* Extended state features, CPUID level 0x0000000d:1 (eax), word 10 */
 #define X86_FEATURE_XSAVEOPT	(10*32+ 0) /* XSAVEOPT */
diff --git a/arch/x86/kernel/fpu/xstate.c b/arch/x86/kernel/fpu/xstate.c
index d425cda5..6e8354f 100644
--- a/arch/x86/kernel/fpu/xstate.c
+++ b/arch/x86/kernel/fpu/xstate.c
@@ -51,6 +51,9 @@ void fpu__xstate_clear_all_cpu_caps(void)
 	setup_clear_cpu_cap(X86_FEATURE_AVX512PF);
 	setup_clear_cpu_cap(X86_FEATURE_AVX512ER);
 	setup_clear_cpu_cap(X86_FEATURE_AVX512CD);
+	setup_clear_cpu_cap(X86_FEATURE_AVX512DQ);
+	setup_clear_cpu_cap(X86_FEATURE_AVX512BW);
+	setup_clear_cpu_cap(X86_FEATURE_AVX512VL);
 	setup_clear_cpu_cap(X86_FEATURE_MPX);
 	setup_clear_cpu_cap(X86_FEATURE_XGETBV1);
 }