[FMV] Improve test coverage for FMV features. (#166)
[FMV] Improve test coverage for FMV features.
Adds tests for f32mm, f64mm, fp16fml, frintts, predres, rcpc3, rng, sve, sve2,
sve2-bitperm, sve2-sha3, sve2-sm4, wfxt, sb, sm4, sme-f64f64, sme-i16i64, mops.
diff --git a/SingleSource/UnitTests/AArch64/acle-fmv-features.c b/SingleSource/UnitTests/AArch64/acle-fmv-features.c
index 9e26d84..d7b36db 100644
--- a/SingleSource/UnitTests/AArch64/acle-fmv-features.c
+++ b/SingleSource/UnitTests/AArch64/acle-fmv-features.c
@@ -15,58 +15,58 @@
static bool any_fails = false;
#if __HAVE_FUNCTION_MULTI_VERSIONING
-#define CHECK(X, TARGET_GUARD, BODY) \
+#define CHECK(FN_NAME_SUFFIX, FMV_FEATURE, TARGET_GUARD, BODY) \
__attribute__((target(#TARGET_GUARD))) \
- static bool try_##X(void) { \
+ static bool try_##FN_NAME_SUFFIX(void) { \
do \
BODY \
while (0); \
return true; \
} \
- __attribute__((target_version(#X))) \
- static void check_##X(void) { \
- printf("%s\n", #X); \
+ __attribute__((target_version(#FMV_FEATURE))) \
+ static void check_##FN_NAME_SUFFIX(void) { \
+ printf("%s\n", #FMV_FEATURE); \
fflush(stdout); \
- if (!safe_try_feature(try_##X)) { \
+ if (!safe_try_feature(try_##FN_NAME_SUFFIX)) { \
printf("\tFAIL\n"); \
any_fails = true; \
} \
} \
__attribute__((target_version("default"))) \
- static void check_##X(void) { \
- printf("%s\n", #X); \
+ static void check_##FN_NAME_SUFFIX(void) { \
+ printf("%s\n", #FMV_FEATURE); \
fflush(stdout); \
- if (safe_try_feature(try_##X)) { \
+ if (safe_try_feature(try_##FN_NAME_SUFFIX)) { \
printf("\tUPASS\n"); \
any_fails = true; \
} \
}
#else
-#define CHECK(X, TARGET_GUARD, BODY) \
- static void check_##X(void) { \
- printf("%s\n", #X); \
+#define CHECK(FN_NAME_SUFFIX, FMV_FEATURE, TARGET_GUARD, BODY) \
+ static void check_##FN_NAME_SUFFIX(void) { \
+ printf("%s\n", #FMV_FEATURE); \
}
#endif
-CHECK(flagm, flagm, {
+CHECK(flagm, flagm, flagm, {
asm volatile (
"cfinv" "\n"
"cfinv" "\n"
);
})
-CHECK(flagm2, arch=armv8.5-a, {
+CHECK(flagm2, flagm2, arch=armv8.5-a, {
asm volatile (
"axflag" "\n"
"xaflag" "\n"
);
})
-CHECK(dotprod, dotprod, {
+CHECK(dotprod, dotprod, dotprod, {
asm volatile (
"udot v0.4S,v1.16B,v2.16B"
: : : "v0"
);
})
-CHECK(sha3, sha3, {
+CHECK(sha3, sha3, sha3, {
asm volatile (
"fmov d0, #0" "\n"
"fmov d1, #0" "\n"
@@ -74,20 +74,20 @@
: : : "v0"
);
})
-CHECK(rdm, rdm, {
+CHECK(rdm, rdm, rdm, {
asm volatile (
"sqrdmlah s0, s1, s2"
: : : "s0"
);
})
-CHECK(lse, lse, {
+CHECK(lse, lse, lse, {
uint64_t pointee = 0;
asm volatile (
"swp xzr, xzr, [%[pointee]]"
: : [pointee]"r"(&pointee)
);
})
-CHECK(sha2, sha2, {
+CHECK(sha2, sha2, sha2, {
asm volatile (
"fmov d0, #0" "\n"
"fmov d1, #0" "\n"
@@ -96,7 +96,7 @@
: : : "v0"
);
})
-CHECK(aes, aes, {
+CHECK(aes, aes, aes, {
asm volatile (
"fmov d0, #0" "\n"
"fmov d1, #0" "\n"
@@ -104,89 +104,88 @@
: : : "v0"
);
})
-CHECK(pmull, aes, {
+CHECK(pmull, pmull, aes, {
asm volatile (
"fmov d0, #0" "\n"
"pmull v0.1q, v0.1d, v0.1d" "\n"
: : : "v0"
);
})
-CHECK(rcpc, rcpc, {
+CHECK(rcpc, rcpc, rcpc, {
int x;
asm volatile (
"ldaprb w0, [%0]"
: : "r" (&x) : "w0"
);
})
-CHECK(rcpc2, rcpc2, {
+CHECK(rcpc2, rcpc2, arch=armv8.4-a, {
int x;
asm volatile (
"mov x1, %0" "\n"
- // FIXME: rcpc2 instructions are under +rcpc-immo in clang, and not +rcpc2.
- ".inst 0x19400020" // ldapurb w0, [x1]
- : : "r" (&x) : "w0"
+ "ldapurb w0, [x1]" "\n"
+ : : "r" (&x) : "w0", "x1"
);
})
-CHECK(fcma, fcma, {
+CHECK(fcma, fcma, fcma, {
asm volatile (
"fmov d0, #0" "\n"
"fcadd v0.2s, v0.2s, v0.2s, #90" "\n"
: : : "v0"
);
})
-CHECK(jscvt, jscvt, {
+CHECK(jscvt, jscvt, jscvt, {
asm volatile (
"fmov d0, #0" "\n"
"fjcvtzs w1, d0" "\n"
: : : "w1", "d0"
);
})
-CHECK(dpb, arch=armv8.2-a, {
+CHECK(dpb, dpb, arch=armv8.2-a, {
int x;
asm volatile (
"dc cvap, %0"
: : "r" (&x)
);
})
-CHECK(dpb2, arch=armv8.5-a, {
+CHECK(dpb2, dpb2, arch=armv8.5-a, {
int x;
asm volatile (
"dc cvadp, %0"
: : "r" (&x)
);
})
-CHECK(bf16, bf16, {
+CHECK(bf16, bf16, bf16, {
asm volatile (
"bfdot v0.4S,v1.8H,v2.8H"
: : : "v0"
);
})
-CHECK(i8mm, i8mm, {
+CHECK(i8mm, i8mm, i8mm, {
asm volatile (
"sudot v0.4S,v1.16B,v2.4B[0]"
: : : "v0"
);
})
-CHECK(dit, dit, {
+CHECK(dit, dit, dit, {
asm volatile (
"msr DIT, x0"
: : : "x0"
);
})
-CHECK(fp16, fp16, {
+CHECK(fp16, fp16, fp16, {
asm volatile (
"fmov h0, #0"
: : : "v0"
);
})
-CHECK(ssbs2, ssbs, {
+CHECK(ssbs2, ssbs2, ssbs, {
asm volatile (
"mrs x0, SSBS" "\n"
"msr SSBS, x0" "\n"
: : : "x0"
);
})
-CHECK(bti, bti, {
+CHECK(bti, bti, bti, {
// The only test for this requires reading a register that is only
// accessible to EL1.
#ifdef __linux__
@@ -207,35 +206,35 @@
// TODO: implement me on your platform to fix this test!
#endif
})
-CHECK(simd, simd, {
+CHECK(simd, simd, simd, {
asm volatile (
"mov v0.B[0], w0"
: : :
);
})
-CHECK(fp, fp, {
+CHECK(fp, fp, fp, {
asm volatile (
"fmov s0, #0"
: : : "v0"
);
})
-CHECK(crc, crc, {
+CHECK(crc, crc, crc, {
asm volatile ( "crc32b wzr, wzr, wzr");
})
-CHECK(sme, sme, {
+CHECK(sme, sme, sme, {
asm volatile (
"rdsvl x0, #1"
: : : "x0"
);
})
-CHECK(sme2, sme2, {
+CHECK(sme2, sme2, sme2, {
asm volatile (
"smstart za" "\n"
"zero { zt0 }" "\n"
"smstop za" "\n"
);
})
-CHECK(ls64, ls64, {
+CHECK(ls64, ls64, ls64, {
long data[8];
asm volatile (
"ld64b x0, [%0]" "\n"
@@ -247,6 +246,122 @@
: "x0", "x1", "x2", "x3", "x4", "x5", "x6", "x7", "x8", "memory"
);
})
+CHECK(f32mm, f32mm, f32mm, {
+ asm volatile (
+ "fmmla z0.s, z1.s, z2.s"
+ : : : "v0"
+ );
+})
+CHECK(f64mm, f64mm, f64mm, {
+ asm volatile (
+ "fmmla z0.d, z1.d, z2.d"
+ : : : "v0"
+ );
+})
+CHECK(fp16fml, fp16fml, fp16fml, {
+ asm volatile (
+ "fmlal v0.2s, v1.2h, v2.2h"
+ : : : "v0"
+ );
+})
+CHECK(frintts, frintts, arch=armv8.5-a, {
+ asm volatile (
+ "frint32z s0, s1"
+ : : : "v0"
+ );
+})
+CHECK(predres, predres, predres, {
+ asm volatile (
+ "cfp rctx, x0" "\n"
+ "dvp rctx, x1" "\n"
+ "cpp rctx, x2" "\n"
+ );
+})
+CHECK(rcpc3, rcpc3, rcpc3, {
+ long x;
+ asm volatile (
+ "stilp wzr, wzr, [%0]"
+ : : "r" (&x) : "memory"
+ );
+})
+CHECK(rng, rng, rng, {
+ asm volatile (
+ "mrs x0, rndr" "\n"
+ "mrs x1, rndrrs" "\n"
+ : : : "x0", "x1"
+ );
+})
+CHECK(sve, sve, sve, {
+ asm volatile (
+ "fadda s0, p7, s0, z31.s"
+ : : : "v0"
+ );
+})
+CHECK(sve2, sve2, sve2, {
+ asm volatile (
+ "match p15.b, p7/z, z0.b, z1.b"
+ : : : "p15", "cc"
+ );
+})
+CHECK(sve2_bitperm, sve2-bitperm, sve2-bitperm, {
+ asm volatile (
+ "bext z0.s, z1.s, z2.s"
+ : : : "z0"
+ );
+})
+CHECK(sve2_sha3, sve2-sha3, sve2-sha3, {
+ asm volatile (
+ "rax1 z0.d, z1.d, z2.d"
+ : : : "z0"
+ );
+})
+CHECK(sve2_sm4, sve2-sm4, sve2-sm4, {
+ asm volatile (
+ "sm4e z0.s, z0.s, z2.s"
+ : : : "z0"
+ );
+})
+CHECK(wfxt, wfxt, wfxt, {
+ asm volatile (
+ "wfet x0" "\n"
+ "wfit x1" "\n"
+ );
+})
+CHECK(sb, sb, sb, {
+ asm volatile ("sb");
+})
+CHECK(sm4, sm4, sm4, {
+ asm volatile (
+ "sm4e v0.4s, v1.4s"
+ : : : "v0"
+ );
+})
+CHECK(sme_f64f64, sme-f64f64, sme-f64f64, {
+ asm volatile (
+ "fmops za0.d, p0/m, p0/m, z0.d, z0.d"
+ : : : "za"
+ );
+})
+CHECK(sme_i16i64, sme-i16i64, sme-i16i64, {
+ asm volatile (
+ "smopa za0.d, p0/m, p0/m, z0.h, z0.h"
+ : : : "za"
+ );
+})
+CHECK(mops, mops, mops, {
+ long dst[64];
+ long src[64];
+ size_t n = 32;
+ long *to = dst;
+ long *from = src;
+ asm volatile (
+ "cpyfp [%0]!, [%1]!, %2!" "\n"
+ "cpyfm [%0]!, [%1]!, %2!" "\n"
+ "cpyfe [%0]!, [%1]!, %2!" "\n"
+ : "+r" (to), "+r" (from), "+r" (n)
+ : : "cc", "memory"
+ );
+})
static bool safe_try_feature(bool (*try_feature)(void)) {
int child = fork();
@@ -288,6 +403,24 @@
check_sme();
check_sme2();
check_ls64();
+ check_f32mm();
+ check_f64mm();
+ check_fp16fml();
+ check_frintts();
+ check_predres();
+ check_rcpc3();
+ check_rng();
+ check_sve();
+ check_sve2();
+ check_sve2_bitperm();
+ check_sve2_sha3();
+ check_sve2_sm4();
+ check_wfxt();
+ check_sb();
+ check_sm4();
+ check_sme_f64f64();
+ check_sme_i16i64();
+ check_mops();
return any_fails ? -1 : 0;
}
diff --git a/SingleSource/UnitTests/AArch64/acle-fmv-features.reference_output b/SingleSource/UnitTests/AArch64/acle-fmv-features.reference_output
index 4b24c57..f00fba4 100644
--- a/SingleSource/UnitTests/AArch64/acle-fmv-features.reference_output
+++ b/SingleSource/UnitTests/AArch64/acle-fmv-features.reference_output
@@ -25,4 +25,22 @@
sme
sme2
ls64
+f32mm
+f64mm
+fp16fml
+frintts
+predres
+rcpc3
+rng
+sve
+sve2
+sve2-bitperm
+sve2-sha3
+sve2-sm4
+wfxt
+sb
+sm4
+sme-f64f64
+sme-i16i64
+mops
exit 0
