Merge tag 'perf-core-for-mingo-5.5-20191128' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/urgent

Pull perf/core improvements and fixes from Arnaldo Carvalho de Melo:

perf script:

  Adrian Hunter:

  - Fix brstackinsn for AUXTRACE.

  - Fix invalid LBR/binary mismatch error.

perf diff:

  Arnaldo Carvalho de Melo:

  - Use llabs() with 64-bit values, fixing the build in some 32-bit
    architectures.

perf pmu:

  Andi Kleen:

  - Use file system cache to optimize sysfs access.

x86:

  Adrian Hunter:

  - Add some more Intel instructions to the opcode map and to the perf
    test entry:

      gf2p8affineinvqb, gf2p8affineqb, gf2p8mulb, v4fmaddps,
      v4fmaddss, v4fnmaddps, v4fnmaddss, vaesdec, vaesdeclast, vaesenc,
      vaesenclast, vcvtne2ps2bf16, vcvtneps2bf16, vdpbf16ps,
      vgf2p8affineinvqb, vgf2p8affineqb, vgf2p8mulb, vp2intersectd,
      vp2intersectq, vp4dpwssd, vp4dpwssds, vpclmulqdq, vpcompressb,
      vpcompressw, vpdpbusd, vpdpbusds, vpdpwssd, vpdpwssds, vpexpandb,
      vpexpandw, vpopcntb, vpopcntd, vpopcntq, vpopcntw, vpshldd, vpshldq,
      vpshldvd, vpshldvq, vpshldvw, vpshldw, vpshrdd, vpshrdq, vpshrdvd,
      vpshrdvq, vpshrdvw, vpshrdw, vpshufbitqmb.

perf affinity:

  Andi Kleen:

  - Add infrastructure to save/restore affinity

perf maps:

  Arnaldo Carvalho de Melo:

  - Merge 'struct maps' with 'struct map_groups', as there is a
    1x1 relationship, simplifying code overal.

perf build:

  Jiri Olsa:

  - Allow to link with libbpf dynamicaly.

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>

1 files changed, 655 insertions, 0 deletions

diff --git a/tools/perf/arch/x86/tests/insn-x86-dat-src.c b/tools/perf/arch/x86/tests/insn-x86-dat-src.c
index dd85a3afd9ce..ddbf07c50bb8 100644
--- a/tools/perf/arch/x86/tests/insn-x86-dat-src.c
+++ b/tools/perf/arch/x86/tests/insn-x86-dat-src.c
@@ -510,6 +510,82 @@ int main(void)
 	asm volatile("vrsqrt14ss %xmm24,%xmm25,%xmm26{%k7}");
 	asm volatile("vrsqrt14sd %xmm24,%xmm25,%xmm26{%k7}");
 
+	/* AVX-512: Op code 0f 38 50 */
+
+	asm volatile("vpdpbusd %xmm1, %xmm2, %xmm3");
+	asm volatile("vpdpbusd %ymm1, %ymm2, %ymm3");
+	asm volatile("vpdpbusd %zmm1, %zmm2, %zmm3");
+	asm volatile("vpdpbusd 0x12345678(%rax,%rcx,8),%zmm2,%zmm3");
+	asm volatile("vpdpbusd 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 38 51 */
+
+	asm volatile("vpdpbusds %xmm1, %xmm2, %xmm3");
+	asm volatile("vpdpbusds %ymm1, %ymm2, %ymm3");
+	asm volatile("vpdpbusds %zmm1, %zmm2, %zmm3");
+	asm volatile("vpdpbusds 0x12345678(%rax,%rcx,8),%zmm2,%zmm3");
+	asm volatile("vpdpbusds 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 38 52 */
+
+	asm volatile("vdpbf16ps %xmm1, %xmm2, %xmm3");
+	asm volatile("vdpbf16ps %ymm1, %ymm2, %ymm3");
+	asm volatile("vdpbf16ps %zmm1, %zmm2, %zmm3");
+	asm volatile("vdpbf16ps 0x12345678(%rax,%rcx,8),%zmm2,%zmm3");
+	asm volatile("vdpbf16ps 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	asm volatile("vpdpwssd %xmm1, %xmm2, %xmm3");
+	asm volatile("vpdpwssd %ymm1, %ymm2, %ymm3");
+	asm volatile("vpdpwssd %zmm1, %zmm2, %zmm3");
+	asm volatile("vpdpwssd 0x12345678(%rax,%rcx,8),%zmm2,%zmm3");
+	asm volatile("vpdpwssd 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	asm volatile("vp4dpwssd (%rax), %zmm0, %zmm4");
+	asm volatile("vp4dpwssd (%eax), %zmm0, %zmm4");
+	asm volatile("vp4dpwssd 0x12345678(%rax,%rcx,8),%zmm0,%zmm4");
+	asm volatile("vp4dpwssd 0x12345678(%eax,%ecx,8),%zmm0,%zmm4");
+
+	/* AVX-512: Op code 0f 38 53 */
+
+	asm volatile("vpdpwssds %xmm1, %xmm2, %xmm3");
+	asm volatile("vpdpwssds %ymm1, %ymm2, %ymm3");
+	asm volatile("vpdpwssds %zmm1, %zmm2, %zmm3");
+	asm volatile("vpdpwssds 0x12345678(%rax,%rcx,8),%zmm2,%zmm3");
+	asm volatile("vpdpwssds 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	asm volatile("vp4dpwssds (%rax), %zmm0, %zmm4");
+	asm volatile("vp4dpwssds (%eax), %zmm0, %zmm4");
+	asm volatile("vp4dpwssds 0x12345678(%rax,%rcx,8),%zmm0,%zmm4");
+	asm volatile("vp4dpwssds 0x12345678(%eax,%ecx,8),%zmm0,%zmm4");
+
+	/* AVX-512: Op code 0f 38 54 */
+
+	asm volatile("vpopcntb %xmm1, %xmm2");
+	asm volatile("vpopcntb %ymm1, %ymm2");
+	asm volatile("vpopcntb %zmm1, %zmm2");
+	asm volatile("vpopcntb 0x12345678(%rax,%rcx,8),%zmm2");
+	asm volatile("vpopcntb 0x12345678(%eax,%ecx,8),%zmm2");
+
+	asm volatile("vpopcntw %xmm1, %xmm2");
+	asm volatile("vpopcntw %ymm1, %ymm2");
+	asm volatile("vpopcntw %zmm1, %zmm2");
+	asm volatile("vpopcntw 0x12345678(%rax,%rcx,8),%zmm2");
+	asm volatile("vpopcntw 0x12345678(%eax,%ecx,8),%zmm2");
+
+	/* AVX-512: Op code 0f 38 55 */
+
+	asm volatile("vpopcntd %xmm1, %xmm2");
+	asm volatile("vpopcntd %ymm1, %ymm2");
+	asm volatile("vpopcntd %zmm1, %zmm2");
+	asm volatile("vpopcntd 0x12345678(%rax,%rcx,8),%zmm2");
+	asm volatile("vpopcntd 0x12345678(%eax,%ecx,8),%zmm2");
+
+	asm volatile("vpopcntq %xmm1, %xmm2");
+	asm volatile("vpopcntq %ymm1, %ymm2");
+	asm volatile("vpopcntq %zmm1, %zmm2");
+	asm volatile("vpopcntq 0x12345678(%rax,%rcx,8),%zmm2");
+	asm volatile("vpopcntq 0x12345678(%eax,%ecx,8),%zmm2");
+
 	/* AVX-512: Op code 0f 38 59 */
 
 	asm volatile("vpbroadcastq %xmm4,%xmm6");
@@ -526,6 +602,34 @@ int main(void)
 	asm volatile("vbroadcasti32x8 (%rcx),%zmm28");
 	asm volatile("vbroadcasti64x4 (%rcx),%zmm26");
 
+	/* AVX-512: Op code 0f 38 62 */
+
+	asm volatile("vpexpandb %xmm1, %xmm2");
+	asm volatile("vpexpandb %ymm1, %ymm2");
+	asm volatile("vpexpandb %zmm1, %zmm2");
+	asm volatile("vpexpandb 0x12345678(%rax,%rcx,8),%zmm2");
+	asm volatile("vpexpandb 0x12345678(%eax,%ecx,8),%zmm2");
+
+	asm volatile("vpexpandw %xmm1, %xmm2");
+	asm volatile("vpexpandw %ymm1, %ymm2");
+	asm volatile("vpexpandw %zmm1, %zmm2");
+	asm volatile("vpexpandw 0x12345678(%rax,%rcx,8),%zmm2");
+	asm volatile("vpexpandw 0x12345678(%eax,%ecx,8),%zmm2");
+
+	/* AVX-512: Op code 0f 38 63 */
+
+	asm volatile("vpcompressb %xmm1, %xmm2");
+	asm volatile("vpcompressb %ymm1, %ymm2");
+	asm volatile("vpcompressb %zmm1, %zmm2");
+	asm volatile("vpcompressb %zmm2,0x12345678(%rax,%rcx,8)");
+	asm volatile("vpcompressb %zmm2,0x12345678(%eax,%ecx,8)");
+
+	asm volatile("vpcompressw %xmm1, %xmm2");
+	asm volatile("vpcompressw %ymm1, %ymm2");
+	asm volatile("vpcompressw %zmm1, %zmm2");
+	asm volatile("vpcompressw %zmm2,0x12345678(%rax,%rcx,8)");
+	asm volatile("vpcompressw %zmm2,0x12345678(%eax,%ecx,8)");
+
 	/* AVX-512: Op code 0f 38 64 */
 
 	asm volatile("vpblendmd %zmm26,%zmm27,%zmm28");
@@ -541,6 +645,76 @@ int main(void)
 	asm volatile("vpblendmb %zmm26,%zmm27,%zmm28");
 	asm volatile("vpblendmw %zmm26,%zmm27,%zmm28");
 
+	/* AVX-512: Op code 0f 38 68 */
+
+	asm volatile("vp2intersectd %xmm1, %xmm2, %k3");
+	asm volatile("vp2intersectd %ymm1, %ymm2, %k3");
+	asm volatile("vp2intersectd %zmm1, %zmm2, %k3");
+	asm volatile("vp2intersectd 0x12345678(%rax,%rcx,8),%zmm2,%k3");
+	asm volatile("vp2intersectd 0x12345678(%eax,%ecx,8),%zmm2,%k3");
+
+	asm volatile("vp2intersectq %xmm1, %xmm2, %k3");
+	asm volatile("vp2intersectq %ymm1, %ymm2, %k3");
+	asm volatile("vp2intersectq %zmm1, %zmm2, %k3");
+	asm volatile("vp2intersectq 0x12345678(%rax,%rcx,8),%zmm2,%k3");
+	asm volatile("vp2intersectq 0x12345678(%eax,%ecx,8),%zmm2,%k3");
+
+	/* AVX-512: Op code 0f 38 70 */
+
+	asm volatile("vpshldvw %xmm1, %xmm2, %xmm3");
+	asm volatile("vpshldvw %ymm1, %ymm2, %ymm3");
+	asm volatile("vpshldvw %zmm1, %zmm2, %zmm3");
+	asm volatile("vpshldvw 0x12345678(%rax,%rcx,8),%zmm2,%zmm3");
+	asm volatile("vpshldvw 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 38 71 */
+
+	asm volatile("vpshldvd %xmm1, %xmm2, %xmm3");
+	asm volatile("vpshldvd %ymm1, %ymm2, %ymm3");
+	asm volatile("vpshldvd %zmm1, %zmm2, %zmm3");
+	asm volatile("vpshldvd 0x12345678(%rax,%rcx,8),%zmm2,%zmm3");
+	asm volatile("vpshldvd 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	asm volatile("vpshldvq %xmm1, %xmm2, %xmm3");
+	asm volatile("vpshldvq %ymm1, %ymm2, %ymm3");
+	asm volatile("vpshldvq %zmm1, %zmm2, %zmm3");
+	asm volatile("vpshldvq 0x12345678(%rax,%rcx,8),%zmm2,%zmm3");
+	asm volatile("vpshldvq 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 38 72 */
+
+	asm volatile("vcvtne2ps2bf16 %xmm1, %xmm2, %xmm3");
+	asm volatile("vcvtne2ps2bf16 %ymm1, %ymm2, %ymm3");
+	asm volatile("vcvtne2ps2bf16 %zmm1, %zmm2, %zmm3");
+	asm volatile("vcvtne2ps2bf16 0x12345678(%rax,%rcx,8),%zmm2,%zmm3");
+	asm volatile("vcvtne2ps2bf16 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	asm volatile("vcvtneps2bf16 %xmm1, %xmm2");
+	asm volatile("vcvtneps2bf16 %ymm1, %xmm2");
+	asm volatile("vcvtneps2bf16 %zmm1, %ymm2");
+	asm volatile("vcvtneps2bf16 0x12345678(%rax,%rcx,8),%ymm2");
+	asm volatile("vcvtneps2bf16 0x12345678(%eax,%ecx,8),%ymm2");
+
+	asm volatile("vpshrdvw %xmm1, %xmm2, %xmm3");
+	asm volatile("vpshrdvw %ymm1, %ymm2, %ymm3");
+	asm volatile("vpshrdvw %zmm1, %zmm2, %zmm3");
+	asm volatile("vpshrdvw 0x12345678(%rax,%rcx,8),%zmm2,%zmm3");
+	asm volatile("vpshrdvw 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 38 73 */
+
+	asm volatile("vpshrdvd %xmm1, %xmm2, %xmm3");
+	asm volatile("vpshrdvd %ymm1, %ymm2, %ymm3");
+	asm volatile("vpshrdvd %zmm1, %zmm2, %zmm3");
+	asm volatile("vpshrdvd 0x12345678(%rax,%rcx,8),%zmm2,%zmm3");
+	asm volatile("vpshrdvd 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	asm volatile("vpshrdvq %xmm1, %xmm2, %xmm3");
+	asm volatile("vpshrdvq %ymm1, %ymm2, %ymm3");
+	asm volatile("vpshrdvq %zmm1, %zmm2, %zmm3");
+	asm volatile("vpshrdvq 0x12345678(%rax,%rcx,8),%zmm2,%zmm3");
+	asm volatile("vpshrdvq 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
 	/* AVX-512: Op code 0f 38 75 */
 
 	asm volatile("vpermi2b %zmm24,%zmm25,%zmm26");
@@ -613,6 +787,14 @@ int main(void)
 	asm volatile("vpermb %zmm26,%zmm27,%zmm28");
 	asm volatile("vpermw %zmm26,%zmm27,%zmm28");
 
+	/* AVX-512: Op code 0f 38 8f */
+
+	asm volatile("vpshufbitqmb %xmm1, %xmm2, %k3");
+	asm volatile("vpshufbitqmb %ymm1, %ymm2, %k3");
+	asm volatile("vpshufbitqmb %zmm1, %zmm2, %k3");
+	asm volatile("vpshufbitqmb 0x12345678(%rax,%rcx,8),%zmm2,%k3");
+	asm volatile("vpshufbitqmb 0x12345678(%eax,%ecx,8),%zmm2,%k3");
+
 	/* AVX-512: Op code 0f 38 90 */
 
 	asm volatile("vpgatherdd %xmm2,0x02(%rbp,%xmm7,2),%xmm1");
@@ -627,6 +809,40 @@ int main(void)
 	asm volatile("vpgatherqd 0x7b(%rbp,%zmm27,8),%ymm26{%k1}");
 	asm volatile("vpgatherqq 0x7b(%rbp,%zmm27,8),%zmm26{%k1}");
 
+	/* AVX-512: Op code 0f 38 9a */
+
+	asm volatile("vfmsub132ps %xmm1, %xmm2, %xmm3");
+	asm volatile("vfmsub132ps %ymm1, %ymm2, %ymm3");
+	asm volatile("vfmsub132ps %zmm1, %zmm2, %zmm3");
+	asm volatile("vfmsub132ps 0x12345678(%rax,%rcx,8),%zmm2,%zmm3");
+	asm volatile("vfmsub132ps 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	asm volatile("vfmsub132pd %xmm1, %xmm2, %xmm3");
+	asm volatile("vfmsub132pd %ymm1, %ymm2, %ymm3");
+	asm volatile("vfmsub132pd %zmm1, %zmm2, %zmm3");
+	asm volatile("vfmsub132pd 0x12345678(%rax,%rcx,8),%zmm2,%zmm3");
+	asm volatile("vfmsub132pd 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	asm volatile("v4fmaddps (%rax), %zmm0, %zmm4");
+	asm volatile("v4fmaddps (%eax), %zmm0, %zmm4");
+	asm volatile("v4fmaddps 0x12345678(%rax,%rcx,8),%zmm0,%zmm4");
+	asm volatile("v4fmaddps 0x12345678(%eax,%ecx,8),%zmm0,%zmm4");
+
+	/* AVX-512: Op code 0f 38 9b */
+
+	asm volatile("vfmsub132ss %xmm1, %xmm2, %xmm3");
+	asm volatile("vfmsub132ss 0x12345678(%rax,%rcx,8),%xmm2,%xmm3");
+	asm volatile("vfmsub132ss 0x12345678(%eax,%ecx,8),%xmm2,%xmm3");
+
+	asm volatile("vfmsub132sd %xmm1, %xmm2, %xmm3");
+	asm volatile("vfmsub132sd 0x12345678(%rax,%rcx,8),%xmm2,%xmm3");
+	asm volatile("vfmsub132sd 0x12345678(%eax,%ecx,8),%xmm2,%xmm3");
+
+	asm volatile("v4fmaddss (%rax), %xmm0, %xmm4");
+	asm volatile("v4fmaddss (%eax), %xmm0, %xmm4");
+	asm volatile("v4fmaddss 0x12345678(%rax,%rcx,8),%xmm0,%xmm4");
+	asm volatile("v4fmaddss 0x12345678(%eax,%ecx,8),%xmm0,%xmm4");
+
 	/* AVX-512: Op code 0f 38 a0 */
 
 	asm volatile("vpscatterdd %zmm28,0x7b(%rbp,%zmm29,8){%k1}");
@@ -647,6 +863,40 @@ int main(void)
 	asm volatile("vscatterqps %ymm6,0x7b(%rbp,%zmm29,8){%k1}");
 	asm volatile("vscatterqpd %zmm28,0x7b(%rbp,%zmm29,8){%k1}");
 
+	/* AVX-512: Op code 0f 38 aa */
+
+	asm volatile("vfmsub213ps %xmm1, %xmm2, %xmm3");
+	asm volatile("vfmsub213ps %ymm1, %ymm2, %ymm3");
+	asm volatile("vfmsub213ps %zmm1, %zmm2, %zmm3");
+	asm volatile("vfmsub213ps 0x12345678(%rax,%rcx,8),%zmm2,%zmm3");
+	asm volatile("vfmsub213ps 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	asm volatile("vfmsub213pd %xmm1, %xmm2, %xmm3");
+	asm volatile("vfmsub213pd %ymm1, %ymm2, %ymm3");
+	asm volatile("vfmsub213pd %zmm1, %zmm2, %zmm3");
+	asm volatile("vfmsub213pd 0x12345678(%rax,%rcx,8),%zmm2,%zmm3");
+	asm volatile("vfmsub213pd 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	asm volatile("v4fnmaddps (%rax), %zmm0, %zmm4");
+	asm volatile("v4fnmaddps (%eax), %zmm0, %zmm4");
+	asm volatile("v4fnmaddps 0x12345678(%rax,%rcx,8),%zmm0,%zmm4");
+	asm volatile("v4fnmaddps 0x12345678(%eax,%ecx,8),%zmm0,%zmm4");
+
+	/* AVX-512: Op code 0f 38 ab */
+
+	asm volatile("vfmsub213ss %xmm1, %xmm2, %xmm3");
+	asm volatile("vfmsub213ss 0x12345678(%rax,%rcx,8),%xmm2,%xmm3");
+	asm volatile("vfmsub213ss 0x12345678(%eax,%ecx,8),%xmm2,%xmm3");
+
+	asm volatile("vfmsub213sd %xmm1, %xmm2, %xmm3");
+	asm volatile("vfmsub213sd 0x12345678(%rax,%rcx,8),%xmm2,%xmm3");
+	asm volatile("vfmsub213sd 0x12345678(%eax,%ecx,8),%xmm2,%xmm3");
+
+	asm volatile("v4fnmaddss (%rax), %xmm0, %xmm4");
+	asm volatile("v4fnmaddss (%eax), %xmm0, %xmm4");
+	asm volatile("v4fnmaddss 0x12345678(%rax,%rcx,8),%xmm0,%xmm4");
+	asm volatile("v4fnmaddss 0x12345678(%eax,%ecx,8),%xmm0,%xmm4");
+
 	/* AVX-512: Op code 0f 38 b4 */
 
 	asm volatile("vpmadd52luq %zmm26,%zmm27,%zmm28");
@@ -685,6 +935,50 @@ int main(void)
 	asm volatile("vrsqrt28ss %xmm28,%xmm29,%xmm30{%k7}");
 	asm volatile("vrsqrt28sd %xmm25,%xmm26,%xmm27{%k7}");
 
+	/* AVX-512: Op code 0f 38 cf */
+
+	asm volatile("gf2p8mulb %xmm1, %xmm3");
+	asm volatile("gf2p8mulb 0x12345678(%rax,%rcx,8),%xmm3");
+	asm volatile("gf2p8mulb 0x12345678(%eax,%ecx,8),%xmm3");
+
+	asm volatile("vgf2p8mulb %xmm1, %xmm2, %xmm3");
+	asm volatile("vgf2p8mulb %ymm1, %ymm2, %ymm3");
+	asm volatile("vgf2p8mulb %zmm1, %zmm2, %zmm3");
+	asm volatile("vgf2p8mulb 0x12345678(%rax,%rcx,8),%zmm2,%zmm3");
+	asm volatile("vgf2p8mulb 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 38 dc */
+
+	asm volatile("vaesenc %xmm1, %xmm2, %xmm3");
+	asm volatile("vaesenc %ymm1, %ymm2, %ymm3");
+	asm volatile("vaesenc %zmm1, %zmm2, %zmm3");
+	asm volatile("vaesenc 0x12345678(%rax,%rcx,8),%zmm2,%zmm3");
+	asm volatile("vaesenc 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 38 dd */
+
+	asm volatile("vaesenclast %xmm1, %xmm2, %xmm3");
+	asm volatile("vaesenclast %ymm1, %ymm2, %ymm3");
+	asm volatile("vaesenclast %zmm1, %zmm2, %zmm3");
+	asm volatile("vaesenclast 0x12345678(%rax,%rcx,8),%zmm2,%zmm3");
+	asm volatile("vaesenclast 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 38 de */
+
+	asm volatile("vaesdec %xmm1, %xmm2, %xmm3");
+	asm volatile("vaesdec %ymm1, %ymm2, %ymm3");
+	asm volatile("vaesdec %zmm1, %zmm2, %zmm3");
+	asm volatile("vaesdec 0x12345678(%rax,%rcx,8),%zmm2,%zmm3");
+	asm volatile("vaesdec 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 38 df */
+
+	asm volatile("vaesdeclast %xmm1, %xmm2, %xmm3");
+	asm volatile("vaesdeclast %ymm1, %ymm2, %ymm3");
+	asm volatile("vaesdeclast %zmm1, %zmm2, %zmm3");
+	asm volatile("vaesdeclast 0x12345678(%rax,%rcx,8),%zmm2,%zmm3");
+	asm volatile("vaesdeclast 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
 	/* AVX-512: Op code 0f 3a 03 */
 
 	asm volatile("valignd $0x12,%zmm28,%zmm29,%zmm30");
@@ -804,6 +1098,13 @@ int main(void)
 	asm volatile("vshufi32x4 $0x12,%zmm25,%zmm26,%zmm27");
 	asm volatile("vshufi64x2 $0x12,%zmm28,%zmm29,%zmm30");
 
+	/* AVX-512: Op code 0f 3a 44 */
+
+	asm volatile("vpclmulqdq $0x12,%xmm1,%xmm2,%xmm3");
+	asm volatile("vpclmulqdq $0x12,%ymm1,%ymm2,%ymm3");
+	asm volatile("vpclmulqdq $0x12,%zmm1,%zmm2,%zmm3");
+	asm volatile("vpclmulqdq $0x12,%zmm25,%zmm26,%zmm27");
+
 	/* AVX-512: Op code 0f 3a 50 */
 
 	asm volatile("vrangeps $0x12,%zmm25,%zmm26,%zmm27");
@@ -844,6 +1145,62 @@ int main(void)
 	asm volatile("vfpclassss $0x12,%xmm27,%k5");
 	asm volatile("vfpclasssd $0x12,%xmm30,%k5");
 
+	/* AVX-512: Op code 0f 3a 70 */
+
+	asm volatile("vpshldw $0x12,%xmm1,%xmm2,%xmm3");
+	asm volatile("vpshldw $0x12,%ymm1,%ymm2,%ymm3");
+	asm volatile("vpshldw $0x12,%zmm1,%zmm2,%zmm3");
+	asm volatile("vpshldw $0x12,%zmm25,%zmm26,%zmm27");
+
+	/* AVX-512: Op code 0f 3a 71 */
+
+	asm volatile("vpshldd $0x12,%xmm1,%xmm2,%xmm3");
+	asm volatile("vpshldd $0x12,%ymm1,%ymm2,%ymm3");
+	asm volatile("vpshldd $0x12,%zmm1,%zmm2,%zmm3");
+	asm volatile("vpshldd $0x12,%zmm25,%zmm26,%zmm27");
+
+	asm volatile("vpshldq $0x12,%xmm1,%xmm2,%xmm3");
+	asm volatile("vpshldq $0x12,%ymm1,%ymm2,%ymm3");
+	asm volatile("vpshldq $0x12,%zmm1,%zmm2,%zmm3");
+	asm volatile("vpshldq $0x12,%zmm25,%zmm26,%zmm27");
+
+	/* AVX-512: Op code 0f 3a 72 */
+
+	asm volatile("vpshrdw $0x12,%xmm1,%xmm2,%xmm3");
+	asm volatile("vpshrdw $0x12,%ymm1,%ymm2,%ymm3");
+	asm volatile("vpshrdw $0x12,%zmm1,%zmm2,%zmm3");
+	asm volatile("vpshrdw $0x12,%zmm25,%zmm26,%zmm27");
+
+	/* AVX-512: Op code 0f 3a 73 */
+
+	asm volatile("vpshrdd $0x12,%xmm1,%xmm2,%xmm3");
+	asm volatile("vpshrdd $0x12,%ymm1,%ymm2,%ymm3");
+	asm volatile("vpshrdd $0x12,%zmm1,%zmm2,%zmm3");
+	asm volatile("vpshrdd $0x12,%zmm25,%zmm26,%zmm27");
+
+	asm volatile("vpshrdq $0x12,%xmm1,%xmm2,%xmm3");
+	asm volatile("vpshrdq $0x12,%ymm1,%ymm2,%ymm3");
+	asm volatile("vpshrdq $0x12,%zmm1,%zmm2,%zmm3");
+	asm volatile("vpshrdq $0x12,%zmm25,%zmm26,%zmm27");
+
+	/* AVX-512: Op code 0f 3a ce */
+
+	asm volatile("gf2p8affineqb $0x12,%xmm1,%xmm3");
+
+	asm volatile("vgf2p8affineqb $0x12,%xmm1,%xmm2,%xmm3");
+	asm volatile("vgf2p8affineqb $0x12,%ymm1,%ymm2,%ymm3");
+	asm volatile("vgf2p8affineqb $0x12,%zmm1,%zmm2,%zmm3");
+	asm volatile("vgf2p8affineqb $0x12,%zmm25,%zmm26,%zmm27");
+
+	/* AVX-512: Op code 0f 3a cf */
+
+	asm volatile("gf2p8affineinvqb $0x12,%xmm1,%xmm3");
+
+	asm volatile("vgf2p8affineinvqb $0x12,%xmm1,%xmm2,%xmm3");
+	asm volatile("vgf2p8affineinvqb $0x12,%ymm1,%ymm2,%ymm3");
+	asm volatile("vgf2p8affineinvqb $0x12,%zmm1,%zmm2,%zmm3");
+	asm volatile("vgf2p8affineinvqb $0x12,%zmm25,%zmm26,%zmm27");
+
 	/* AVX-512: Op code 0f 72 (Grp13) */
 
 	asm volatile("vprord $0x12,%zmm25,%zmm26");
@@ -1946,6 +2303,69 @@ int main(void)
 	asm volatile("vrsqrt14ss %xmm4,%xmm5,%xmm6{%k7}");
 	asm volatile("vrsqrt14sd %xmm4,%xmm5,%xmm6{%k7}");
 
+	/* AVX-512: Op code 0f 38 50 */
+
+	asm volatile("vpdpbusd %xmm1, %xmm2, %xmm3");
+	asm volatile("vpdpbusd %ymm1, %ymm2, %ymm3");
+	asm volatile("vpdpbusd %zmm1, %zmm2, %zmm3");
+	asm volatile("vpdpbusd 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 38 51 */
+
+	asm volatile("vpdpbusds %xmm1, %xmm2, %xmm3");
+	asm volatile("vpdpbusds %ymm1, %ymm2, %ymm3");
+	asm volatile("vpdpbusds %zmm1, %zmm2, %zmm3");
+	asm volatile("vpdpbusds 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 38 52 */
+
+	asm volatile("vdpbf16ps %xmm1, %xmm2, %xmm3");
+	asm volatile("vdpbf16ps %ymm1, %ymm2, %ymm3");
+	asm volatile("vdpbf16ps %zmm1, %zmm2, %zmm3");
+	asm volatile("vdpbf16ps 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	asm volatile("vpdpwssd %xmm1, %xmm2, %xmm3");
+	asm volatile("vpdpwssd %ymm1, %ymm2, %ymm3");
+	asm volatile("vpdpwssd %zmm1, %zmm2, %zmm3");
+	asm volatile("vpdpwssd 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	asm volatile("vp4dpwssd (%eax), %zmm0, %zmm4");
+	asm volatile("vp4dpwssd 0x12345678(%eax,%ecx,8),%zmm0,%zmm4");
+
+	/* AVX-512: Op code 0f 38 53 */
+
+	asm volatile("vpdpwssds %xmm1, %xmm2, %xmm3");
+	asm volatile("vpdpwssds %ymm1, %ymm2, %ymm3");
+	asm volatile("vpdpwssds %zmm1, %zmm2, %zmm3");
+	asm volatile("vpdpwssds 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	asm volatile("vp4dpwssds (%eax), %zmm0, %zmm4");
+	asm volatile("vp4dpwssds 0x12345678(%eax,%ecx,8),%zmm0,%zmm4");
+
+	/* AVX-512: Op code 0f 38 54 */
+
+	asm volatile("vpopcntb %xmm1, %xmm2");
+	asm volatile("vpopcntb %ymm1, %ymm2");
+	asm volatile("vpopcntb %zmm1, %zmm2");
+	asm volatile("vpopcntb 0x12345678(%eax,%ecx,8),%zmm2");
+
+	asm volatile("vpopcntw %xmm1, %xmm2");
+	asm volatile("vpopcntw %ymm1, %ymm2");
+	asm volatile("vpopcntw %zmm1, %zmm2");
+	asm volatile("vpopcntw 0x12345678(%eax,%ecx,8),%zmm2");
+
+	/* AVX-512: Op code 0f 38 55 */
+
+	asm volatile("vpopcntd %xmm1, %xmm2");
+	asm volatile("vpopcntd %ymm1, %ymm2");
+	asm volatile("vpopcntd %zmm1, %zmm2");
+	asm volatile("vpopcntd 0x12345678(%eax,%ecx,8),%zmm2");
+
+	asm volatile("vpopcntq %xmm1, %xmm2");
+	asm volatile("vpopcntq %ymm1, %ymm2");
+	asm volatile("vpopcntq %zmm1, %zmm2");
+	asm volatile("vpopcntq 0x12345678(%eax,%ecx,8),%zmm2");
+
 	/* AVX-512: Op code 0f 38 59 */
 
 	asm volatile("vpbroadcastq %xmm4,%xmm6");
@@ -1962,6 +2382,30 @@ int main(void)
 	asm volatile("vbroadcasti32x8 (%ecx),%zmm6");
 	asm volatile("vbroadcasti64x4 (%ecx),%zmm6");
 
+	/* AVX-512: Op code 0f 38 62 */
+
+	asm volatile("vpexpandb %xmm1, %xmm2");
+	asm volatile("vpexpandb %ymm1, %ymm2");
+	asm volatile("vpexpandb %zmm1, %zmm2");
+	asm volatile("vpexpandb 0x12345678(%eax,%ecx,8),%zmm2");
+
+	asm volatile("vpexpandw %xmm1, %xmm2");
+	asm volatile("vpexpandw %ymm1, %ymm2");
+	asm volatile("vpexpandw %zmm1, %zmm2");
+	asm volatile("vpexpandw 0x12345678(%eax,%ecx,8),%zmm2");
+
+	/* AVX-512: Op code 0f 38 63 */
+
+	asm volatile("vpcompressb %xmm1, %xmm2");
+	asm volatile("vpcompressb %ymm1, %ymm2");
+	asm volatile("vpcompressb %zmm1, %zmm2");
+	asm volatile("vpcompressb %zmm2,0x12345678(%eax,%ecx,8)");
+
+	asm volatile("vpcompressw %xmm1, %xmm2");
+	asm volatile("vpcompressw %ymm1, %ymm2");
+	asm volatile("vpcompressw %zmm1, %zmm2");
+	asm volatile("vpcompressw %zmm2,0x12345678(%eax,%ecx,8)");
+
 	/* AVX-512: Op code 0f 38 64 */
 
 	asm volatile("vpblendmd %zmm4,%zmm5,%zmm6");
@@ -1977,6 +2421,66 @@ int main(void)
 	asm volatile("vpblendmb %zmm4,%zmm5,%zmm6");
 	asm volatile("vpblendmw %zmm4,%zmm5,%zmm6");
 
+	/* AVX-512: Op code 0f 38 68 */
+
+	asm volatile("vp2intersectd %xmm1, %xmm2, %k3");
+	asm volatile("vp2intersectd %ymm1, %ymm2, %k3");
+	asm volatile("vp2intersectd %zmm1, %zmm2, %k3");
+	asm volatile("vp2intersectd 0x12345678(%eax,%ecx,8),%zmm2,%k3");
+
+	asm volatile("vp2intersectq %xmm1, %xmm2, %k3");
+	asm volatile("vp2intersectq %ymm1, %ymm2, %k3");
+	asm volatile("vp2intersectq %zmm1, %zmm2, %k3");
+	asm volatile("vp2intersectq 0x12345678(%eax,%ecx,8),%zmm2,%k3");
+
+	/* AVX-512: Op code 0f 38 70 */
+
+	asm volatile("vpshldvw %xmm1, %xmm2, %xmm3");
+	asm volatile("vpshldvw %ymm1, %ymm2, %ymm3");
+	asm volatile("vpshldvw %zmm1, %zmm2, %zmm3");
+	asm volatile("vpshldvw 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 38 71 */
+
+	asm volatile("vpshldvd %xmm1, %xmm2, %xmm3");
+	asm volatile("vpshldvd %ymm1, %ymm2, %ymm3");
+	asm volatile("vpshldvd %zmm1, %zmm2, %zmm3");
+	asm volatile("vpshldvd 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	asm volatile("vpshldvq %xmm1, %xmm2, %xmm3");
+	asm volatile("vpshldvq %ymm1, %ymm2, %ymm3");
+	asm volatile("vpshldvq %zmm1, %zmm2, %zmm3");
+	asm volatile("vpshldvq 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 38 72 */
+
+	asm volatile("vcvtne2ps2bf16 %xmm1, %xmm2, %xmm3");
+	asm volatile("vcvtne2ps2bf16 %ymm1, %ymm2, %ymm3");
+	asm volatile("vcvtne2ps2bf16 %zmm1, %zmm2, %zmm3");
+	asm volatile("vcvtne2ps2bf16 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	asm volatile("vcvtneps2bf16 %xmm1, %xmm2");
+	asm volatile("vcvtneps2bf16 %ymm1, %xmm2");
+	asm volatile("vcvtneps2bf16 %zmm1, %ymm2");
+	asm volatile("vcvtneps2bf16 0x12345678(%eax,%ecx,8),%ymm2");
+
+	asm volatile("vpshrdvw %xmm1, %xmm2, %xmm3");
+	asm volatile("vpshrdvw %ymm1, %ymm2, %ymm3");
+	asm volatile("vpshrdvw %zmm1, %zmm2, %zmm3");
+	asm volatile("vpshrdvw 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 38 73 */
+
+	asm volatile("vpshrdvd %xmm1, %xmm2, %xmm3");
+	asm volatile("vpshrdvd %ymm1, %ymm2, %ymm3");
+	asm volatile("vpshrdvd %zmm1, %zmm2, %zmm3");
+	asm volatile("vpshrdvd 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	asm volatile("vpshrdvq %xmm1, %xmm2, %xmm3");
+	asm volatile("vpshrdvq %ymm1, %ymm2, %ymm3");
+	asm volatile("vpshrdvq %zmm1, %zmm2, %zmm3");
+	asm volatile("vpshrdvq 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
 	/* AVX-512: Op code 0f 38 75 */
 
 	asm volatile("vpermi2b %zmm4,%zmm5,%zmm6");
@@ -2048,6 +2552,13 @@ int main(void)
 	asm volatile("vpermb %zmm4,%zmm5,%zmm6");
 	asm volatile("vpermw %zmm4,%zmm5,%zmm6");
 
+	/* AVX-512: Op code 0f 38 8f */
+
+	asm volatile("vpshufbitqmb %xmm1, %xmm2, %k3");
+	asm volatile("vpshufbitqmb %ymm1, %ymm2, %k3");
+	asm volatile("vpshufbitqmb %zmm1, %zmm2, %k3");
+	asm volatile("vpshufbitqmb 0x12345678(%eax,%ecx,8),%zmm2,%k3");
+
 	/* AVX-512: Op code 0f 38 90 */
 
 	asm volatile("vpgatherdd %xmm2,0x02(%ebp,%xmm7,2),%xmm1");
@@ -2062,6 +2573,32 @@ int main(void)
 	asm volatile("vpgatherqd 0x7b(%ebp,%zmm7,8),%ymm6{%k1}");
 	asm volatile("vpgatherqq 0x7b(%ebp,%zmm7,8),%zmm6{%k1}");
 
+	/* AVX-512: Op code 0f 38 9a */
+
+	asm volatile("vfmsub132ps %xmm1, %xmm2, %xmm3");
+	asm volatile("vfmsub132ps %ymm1, %ymm2, %ymm3");
+	asm volatile("vfmsub132ps %zmm1, %zmm2, %zmm3");
+	asm volatile("vfmsub132ps 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	asm volatile("vfmsub132pd %xmm1, %xmm2, %xmm3");
+	asm volatile("vfmsub132pd %ymm1, %ymm2, %ymm3");
+	asm volatile("vfmsub132pd %zmm1, %zmm2, %zmm3");
+	asm volatile("vfmsub132pd 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	asm volatile("v4fmaddps (%eax), %zmm0, %zmm4");
+	asm volatile("v4fmaddps 0x12345678(%eax,%ecx,8),%zmm0,%zmm4");
+
+	/* AVX-512: Op code 0f 38 9b */
+
+	asm volatile("vfmsub132ss %xmm1, %xmm2, %xmm3");
+	asm volatile("vfmsub132ss 0x12345678(%eax,%ecx,8),%xmm2,%xmm3");
+
+	asm volatile("vfmsub132sd %xmm1, %xmm2, %xmm3");
+	asm volatile("vfmsub132sd 0x12345678(%eax,%ecx,8),%xmm2,%xmm3");
+
+	asm volatile("v4fmaddss (%eax), %xmm0, %xmm4");
+	asm volatile("v4fmaddss 0x12345678(%eax,%ecx,8),%xmm0,%xmm4");
+
 	/* AVX-512: Op code 0f 38 a0 */
 
 	asm volatile("vpscatterdd %zmm6,0x7b(%ebp,%zmm7,8){%k1}");
@@ -2082,6 +2619,32 @@ int main(void)
 	asm volatile("vscatterqps %ymm6,0x7b(%ebp,%zmm7,8){%k1}");
 	asm volatile("vscatterqpd %zmm6,0x7b(%ebp,%zmm7,8){%k1}");
 
+	/* AVX-512: Op code 0f 38 aa */
+
+	asm volatile("vfmsub213ps %xmm1, %xmm2, %xmm3");
+	asm volatile("vfmsub213ps %ymm1, %ymm2, %ymm3");
+	asm volatile("vfmsub213ps %zmm1, %zmm2, %zmm3");
+	asm volatile("vfmsub213ps 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	asm volatile("vfmsub213pd %xmm1, %xmm2, %xmm3");
+	asm volatile("vfmsub213pd %ymm1, %ymm2, %ymm3");
+	asm volatile("vfmsub213pd %zmm1, %zmm2, %zmm3");
+	asm volatile("vfmsub213pd 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	asm volatile("v4fnmaddps (%eax), %zmm0, %zmm4");
+	asm volatile("v4fnmaddps 0x12345678(%eax,%ecx,8),%zmm0,%zmm4");
+
+	/* AVX-512: Op code 0f 38 ab */
+
+	asm volatile("vfmsub213ss %xmm1, %xmm2, %xmm3");
+	asm volatile("vfmsub213ss 0x12345678(%eax,%ecx,8),%xmm2,%xmm3");
+
+	asm volatile("vfmsub213sd %xmm1, %xmm2, %xmm3");
+	asm volatile("vfmsub213sd 0x12345678(%eax,%ecx,8),%xmm2,%xmm3");
+
+	asm volatile("v4fnmaddss (%eax), %xmm0, %xmm4");
+	asm volatile("v4fnmaddss 0x12345678(%eax,%ecx,8),%xmm0,%xmm4");
+
 	/* AVX-512: Op code 0f 38 b4 */
 
 	asm volatile("vpmadd52luq %zmm4,%zmm5,%zmm6");
@@ -2120,6 +2683,44 @@ int main(void)
 	asm volatile("vrsqrt28ss %xmm5,%xmm6,%xmm7{%k7}");
 	asm volatile("vrsqrt28sd %xmm5,%xmm6,%xmm7{%k7}");
 
+	/* AVX-512: Op code 0f 38 cf */
+
+	asm volatile("gf2p8mulb %xmm1, %xmm3");
+	asm volatile("gf2p8mulb 0x12345678(%eax,%ecx,8),%xmm3");
+
+	asm volatile("vgf2p8mulb %xmm1, %xmm2, %xmm3");
+	asm volatile("vgf2p8mulb %ymm1, %ymm2, %ymm3");
+	asm volatile("vgf2p8mulb %zmm1, %zmm2, %zmm3");
+	asm volatile("vgf2p8mulb 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 38 dc */
+
+	asm volatile("vaesenc %xmm1, %xmm2, %xmm3");
+	asm volatile("vaesenc %ymm1, %ymm2, %ymm3");
+	asm volatile("vaesenc %zmm1, %zmm2, %zmm3");
+	asm volatile("vaesenc 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 38 dd */
+
+	asm volatile("vaesenclast %xmm1, %xmm2, %xmm3");
+	asm volatile("vaesenclast %ymm1, %ymm2, %ymm3");
+	asm volatile("vaesenclast %zmm1, %zmm2, %zmm3");
+	asm volatile("vaesenclast 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 38 de */
+
+	asm volatile("vaesdec %xmm1, %xmm2, %xmm3");
+	asm volatile("vaesdec %ymm1, %ymm2, %ymm3");
+	asm volatile("vaesdec %zmm1, %zmm2, %zmm3");
+	asm volatile("vaesdec 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 38 df */
+
+	asm volatile("vaesdeclast %xmm1, %xmm2, %xmm3");
+	asm volatile("vaesdeclast %ymm1, %ymm2, %ymm3");
+	asm volatile("vaesdeclast %zmm1, %zmm2, %zmm3");
+	asm volatile("vaesdeclast 0x12345678(%eax,%ecx,8),%zmm2,%zmm3");
+
 	/* AVX-512: Op code 0f 3a 03 */
 
 	asm volatile("valignd $0x12,%zmm5,%zmm6,%zmm7");
@@ -2239,6 +2840,12 @@ int main(void)
 	asm volatile("vshufi32x4 $0x12,%zmm5,%zmm6,%zmm7");
 	asm volatile("vshufi64x2 $0x12,%zmm5,%zmm6,%zmm7");
 
+	/* AVX-512: Op code 0f 3a 44 */
+
+	asm volatile("vpclmulqdq $0x12,%xmm1,%xmm2,%xmm3");
+	asm volatile("vpclmulqdq $0x12,%ymm1,%ymm2,%ymm3");
+	asm volatile("vpclmulqdq $0x12,%zmm1,%zmm2,%zmm3");
+
 	/* AVX-512: Op code 0f 3a 50 */
 
 	asm volatile("vrangeps $0x12,%zmm5,%zmm6,%zmm7");
@@ -2279,6 +2886,54 @@ int main(void)
 	asm volatile("vfpclassss $0x12,%xmm7,%k5");
 	asm volatile("vfpclasssd $0x12,%xmm7,%k5");
 
+	/* AVX-512: Op code 0f 3a 70 */
+
+	asm volatile("vpshldw $0x12,%xmm1,%xmm2,%xmm3");
+	asm volatile("vpshldw $0x12,%ymm1,%ymm2,%ymm3");
+	asm volatile("vpshldw $0x12,%zmm1,%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 3a 71 */
+
+	asm volatile("vpshldd $0x12,%xmm1,%xmm2,%xmm3");
+	asm volatile("vpshldd $0x12,%ymm1,%ymm2,%ymm3");
+	asm volatile("vpshldd $0x12,%zmm1,%zmm2,%zmm3");
+
+	asm volatile("vpshldq $0x12,%xmm1,%xmm2,%xmm3");
+	asm volatile("vpshldq $0x12,%ymm1,%ymm2,%ymm3");
+	asm volatile("vpshldq $0x12,%zmm1,%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 3a 72 */
+
+	asm volatile("vpshrdw $0x12,%xmm1,%xmm2,%xmm3");
+	asm volatile("vpshrdw $0x12,%ymm1,%ymm2,%ymm3");
+	asm volatile("vpshrdw $0x12,%zmm1,%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 3a 73 */
+
+	asm volatile("vpshrdd $0x12,%xmm1,%xmm2,%xmm3");
+	asm volatile("vpshrdd $0x12,%ymm1,%ymm2,%ymm3");
+	asm volatile("vpshrdd $0x12,%zmm1,%zmm2,%zmm3");
+
+	asm volatile("vpshrdq $0x12,%xmm1,%xmm2,%xmm3");
+	asm volatile("vpshrdq $0x12,%ymm1,%ymm2,%ymm3");
+	asm volatile("vpshrdq $0x12,%zmm1,%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 3a ce */
+
+	asm volatile("gf2p8affineqb $0x12,%xmm1,%xmm3");
+
+	asm volatile("vgf2p8affineqb $0x12,%xmm1,%xmm2,%xmm3");
+	asm volatile("vgf2p8affineqb $0x12,%ymm1,%ymm2,%ymm3");
+	asm volatile("vgf2p8affineqb $0x12,%zmm1,%zmm2,%zmm3");
+
+	/* AVX-512: Op code 0f 3a cf */
+
+	asm volatile("gf2p8affineinvqb $0x12,%xmm1,%xmm3");
+
+	asm volatile("vgf2p8affineinvqb $0x12,%xmm1,%xmm2,%xmm3");
+	asm volatile("vgf2p8affineinvqb $0x12,%ymm1,%ymm2,%ymm3");
+	asm volatile("vgf2p8affineinvqb $0x12,%zmm1,%zmm2,%zmm3");
+
 	/* AVX-512: Op code 0f 72 (Grp13) */
 
 	asm volatile("vprord $0x12,%zmm5,%zmm6");