|
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [xen master] x86emul: support AVX512_VP2INTERSECT insns
commit fc7f700cf1845d80dee1f4075044a54645aec04e
Author: Jan Beulich <jbeulich@xxxxxxxx>
AuthorDate: Tue Jul 21 14:00:25 2020 +0200
Commit: Jan Beulich <jbeulich@xxxxxxxx>
CommitDate: Tue Jul 21 14:00:25 2020 +0200
x86emul: support AVX512_VP2INTERSECT insns
The standard memory access pattern once again should allow us to go
without a test harness addition beyond the EVEX Disp8-scaling one.
Signed-off-by: Jan Beulich <jbeulich@xxxxxxxx>
Acked-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
---
tools/libxl/libxl_cpuid.c | 1 +
tools/misc/xen-cpuid.c | 2 +-
tools/tests/x86_emulator/evex-disp8.c | 5 +++++
tools/tests/x86_emulator/x86-emulate.h | 1 +
xen/arch/x86/x86_emulate/x86_emulate.c | 8 ++++++++
xen/include/asm-x86/cpufeature.h | 1 +
xen/include/public/arch-x86/cpufeatureset.h | 1 +
xen/tools/gen-cpuid.py | 2 +-
8 files changed, 19 insertions(+), 2 deletions(-)
diff --git a/tools/libxl/libxl_cpuid.c b/tools/libxl/libxl_cpuid.c
index db2f12d115..8b570b7e27 100644
--- a/tools/libxl/libxl_cpuid.c
+++ b/tools/libxl/libxl_cpuid.c
@@ -214,6 +214,7 @@ int libxl_cpuid_parse_config(libxl_cpuid_policy_list
*cpuid, const char* str)
{"avx512-4vnniw",0x00000007, 0, CPUID_REG_EDX, 2, 1},
{"avx512-4fmaps",0x00000007, 0, CPUID_REG_EDX, 3, 1},
+ {"avx512-vp2intersect",0x00000007,0,CPUID_REG_EDX,8, 1},
{"srbds-ctrl", 0x00000007, 0, CPUID_REG_EDX, 9, 1},
{"md-clear", 0x00000007, 0, CPUID_REG_EDX, 10, 1},
{"serialize", 0x00000007, 0, CPUID_REG_EDX, 14, 1},
diff --git a/tools/misc/xen-cpuid.c b/tools/misc/xen-cpuid.c
index 8d8f3532a2..f4bf88e2fa 100644
--- a/tools/misc/xen-cpuid.c
+++ b/tools/misc/xen-cpuid.c
@@ -160,7 +160,7 @@ static const char *const str_7d0[32] =
[ 2] = "avx512_4vnniw", [ 3] = "avx512_4fmaps",
[ 4] = "fsrm",
- /* 8 */ [ 9] = "srbds-ctrl",
+ [ 8] = "avx512-vp2intersect", [ 9] = "srbds-ctrl",
[10] = "md-clear",
/* 12 */ [13] = "tsx-force-abort",
[14] = "serialize",
diff --git a/tools/tests/x86_emulator/evex-disp8.c
b/tools/tests/x86_emulator/evex-disp8.c
index b892ed90da..60c569d5e5 100644
--- a/tools/tests/x86_emulator/evex-disp8.c
+++ b/tools/tests/x86_emulator/evex-disp8.c
@@ -593,6 +593,10 @@ static const struct test avx512_vnni_all[] = {
INSN(pdpwssds, 66, 0f38, 53, vl, d, vl),
};
+static const struct test avx512_vp2intersect_all[] = {
+ INSN(p2intersect, f2, 0f38, 68, vl, dq, vl)
+};
+
static const struct test avx512_vpopcntdq_all[] = {
INSN(popcnt, 66, 0f38, 55, vl, dq, vl)
};
@@ -996,6 +1000,7 @@ void evex_disp8_test(void *instr, struct x86_emulate_ctxt
*ctxt,
RUN(avx512_vbmi, all);
RUN(avx512_vbmi2, all);
RUN(avx512_vnni, all);
+ RUN(avx512_vp2intersect, all);
RUN(avx512_vpopcntdq, all);
if ( cpu_has_avx512f )
diff --git a/tools/tests/x86_emulator/x86-emulate.h
b/tools/tests/x86_emulator/x86-emulate.h
index 6ea6cbc658..fc52edf8ce 100644
--- a/tools/tests/x86_emulator/x86-emulate.h
+++ b/tools/tests/x86_emulator/x86-emulate.h
@@ -168,6 +168,7 @@ static inline bool xcr0_mask(uint64_t mask)
#define cpu_has_movdir64b cp.feat.movdir64b
#define cpu_has_avx512_4vnniw (cp.feat.avx512_4vnniw && xcr0_mask(0xe6))
#define cpu_has_avx512_4fmaps (cp.feat.avx512_4fmaps && xcr0_mask(0xe6))
+#define cpu_has_avx512_vp2intersect (cp.feat.avx512_vp2intersect &&
xcr0_mask(0xe6))
#define cpu_has_serialize cp.feat.serialize
#define cpu_has_avx512_bf16 (cp.feat.avx512_bf16 && xcr0_mask(0xe6))
diff --git a/xen/arch/x86/x86_emulate/x86_emulate.c
b/xen/arch/x86/x86_emulate/x86_emulate.c
index 84bb8e0c9b..44474a268b 100644
--- a/xen/arch/x86/x86_emulate/x86_emulate.c
+++ b/xen/arch/x86/x86_emulate/x86_emulate.c
@@ -488,6 +488,7 @@ static const struct ext0f38_table {
[0x62] = { .simd_size = simd_packed_int, .two_op = 1, .d8s = d8s_bw },
[0x63] = { .simd_size = simd_packed_int, .to_mem = 1, .two_op = 1, .d8s =
d8s_bw },
[0x64 ... 0x66] = { .simd_size = simd_packed_int, .d8s = d8s_vl },
+ [0x68] = { .simd_size = simd_packed_int, .d8s = d8s_vl },
[0x70 ... 0x73] = { .simd_size = simd_packed_int, .d8s = d8s_vl },
[0x75 ... 0x76] = { .simd_size = simd_packed_int, .d8s = d8s_vl },
[0x77] = { .simd_size = simd_packed_fp, .d8s = d8s_vl },
@@ -2005,6 +2006,7 @@ amd_like(const struct x86_emulate_ctxt *ctxt)
#define vcpu_has_enqcmd() (ctxt->cpuid->feat.enqcmd)
#define vcpu_has_avx512_4vnniw() (ctxt->cpuid->feat.avx512_4vnniw)
#define vcpu_has_avx512_4fmaps() (ctxt->cpuid->feat.avx512_4fmaps)
+#define vcpu_has_avx512_vp2intersect() (ctxt->cpuid->feat.avx512_vp2intersect)
#define vcpu_has_serialize() (ctxt->cpuid->feat.serialize)
#define vcpu_has_avx512_bf16() (ctxt->cpuid->feat.avx512_bf16)
@@ -9545,6 +9547,12 @@ x86_emulate(
}
goto simd_zmm;
+ case X86EMUL_OPC_EVEX_F2(0x0f38, 0x68): /* vp2intersect{d,q}
[xyz]mm/mem,[xyz]mm,k+1 */
+ host_and_vcpu_must_have(avx512_vp2intersect);
+ generate_exception_if(evex.opmsk || !evex.r || !evex.R, EXC_UD);
+ op_bytes = 16 << evex.lr;
+ goto avx512f_no_sae;
+
case X86EMUL_OPC_EVEX_66(0x0f38, 0x70): /* vpshldvw
[xyz]mm/mem,[xyz]mm,[xyz]mm{k} */
case X86EMUL_OPC_EVEX_66(0x0f38, 0x72): /* vpshrdvw
[xyz]mm/mem,[xyz]mm,[xyz]mm{k} */
generate_exception_if(!evex.w, EXC_UD);
diff --git a/xen/include/asm-x86/cpufeature.h b/xen/include/asm-x86/cpufeature.h
index f790d5c1f8..4de992781e 100644
--- a/xen/include/asm-x86/cpufeature.h
+++ b/xen/include/asm-x86/cpufeature.h
@@ -128,6 +128,7 @@
/* CPUID level 0x00000007:0.edx */
#define cpu_has_avx512_4vnniw boot_cpu_has(X86_FEATURE_AVX512_4VNNIW)
#define cpu_has_avx512_4fmaps boot_cpu_has(X86_FEATURE_AVX512_4FMAPS)
+#define cpu_has_avx512_vp2intersect
boot_cpu_has(X86_FEATURE_AVX512_VP2INTERSECT)
#define cpu_has_tsx_force_abort boot_cpu_has(X86_FEATURE_TSX_FORCE_ABORT)
#define cpu_has_serialize boot_cpu_has(X86_FEATURE_SERIALIZE)
diff --git a/xen/include/public/arch-x86/cpufeatureset.h
b/xen/include/public/arch-x86/cpufeatureset.h
index fe7492a225..ce3deaa5c7 100644
--- a/xen/include/public/arch-x86/cpufeatureset.h
+++ b/xen/include/public/arch-x86/cpufeatureset.h
@@ -259,6 +259,7 @@ XEN_CPUFEATURE(AMD_PPIN, 8*32+23) /* Protected
Processor Inventory Number
/* Intel-defined CPU features, CPUID level 0x00000007:0.edx, word 9 */
XEN_CPUFEATURE(AVX512_4VNNIW, 9*32+ 2) /*A AVX512 Neural Network Instructions
*/
XEN_CPUFEATURE(AVX512_4FMAPS, 9*32+ 3) /*A AVX512 Multiply Accumulation
Single Precision */
+XEN_CPUFEATURE(AVX512_VP2INTERSECT, 9*32+8) /*a VP2INTERSECT{D,Q} insns */
XEN_CPUFEATURE(SRBDS_CTRL, 9*32+ 9) /* MSR_MCU_OPT_CTRL and
RNGDS_MITG_DIS. */
XEN_CPUFEATURE(MD_CLEAR, 9*32+10) /*A VERW clears microarchitectural
buffers */
XEN_CPUFEATURE(TSX_FORCE_ABORT, 9*32+13) /* MSR_TSX_FORCE_ABORT.RTM_ABORT */
diff --git a/xen/tools/gen-cpuid.py b/xen/tools/gen-cpuid.py
index ffd9529fdf..143ee7cae1 100755
--- a/xen/tools/gen-cpuid.py
+++ b/xen/tools/gen-cpuid.py
@@ -260,7 +260,7 @@ def crunch_numbers(state):
# AVX512 features are built on top of AVX512F
AVX512F: [AVX512DQ, AVX512_IFMA, AVX512PF, AVX512ER, AVX512CD,
AVX512BW, AVX512VL, AVX512_4VNNIW, AVX512_4FMAPS,
- AVX512_VNNI, AVX512_VPOPCNTDQ],
+ AVX512_VNNI, AVX512_VPOPCNTDQ, AVX512_VP2INTERSECT],
# AVX512 extensions acting on vectors of bytes/words are made
# dependents of AVX512BW (as to requiring wider than 16-bit mask
--
generated by git-patchbot for /home/xen/git/xen.git#master
|
 |
Lists.xenproject.org is hosted with RackSpace, monitoring our |