Merge remote-tracking branch 'remotes/rth/tags/pull-tcg-20200317' into staging

Fix tcg/i386 bug vs sari_vec.
Fix tcg-runtime-gvec.c vs i386 without avx.

# gpg: Signature made Tue 17 Mar 2020 18:58:14 GMT
# gpg:                using RSA key 7A481E78868B4DB6A85A05C064DF38E8AF7E215F
# gpg:                issuer "richard.henderson@linaro.org"
# gpg: Good signature from "Richard Henderson <richard.henderson@linaro.org>" [full]
# Primary key fingerprint: 7A48 1E78 868B 4DB6 A85A  05C0 64DF 38E8 AF7E 215F

* remotes/rth/tags/pull-tcg-20200317:
  tcg: Remove tcg-runtime-gvec.c DO_CMP0
  tcg: Tidy tcg-runtime-gvec.c DUP*
  tcg: Tidy tcg-runtime-gvec.c types
  tcg: Remove CONFIG_VECTOR16
  tcg/i386: Bound shift count expanding sari_vec

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

3 files changed, 122 insertions, 241 deletions

diff --git a/accel/tcg/tcg-runtime-gvec.c b/accel/tcg/tcg-runtime-gvec.c
index 5b1902d..ca44970 100644
--- a/accel/tcg/tcg-runtime-gvec.c
+++ b/accel/tcg/tcg-runtime-gvec.c
@@ -24,48 +24,6 @@
 #include "tcg/tcg-gvec-desc.h"
 
 
-/* Virtually all hosts support 16-byte vectors.  Those that don't can emulate
- * them via GCC's generic vector extension.  This turns out to be simpler and
- * more reliable than getting the compiler to autovectorize.
- *
- * In tcg-op-gvec.c, we asserted that both the size and alignment of the data
- * are multiples of 16.
- *
- * When the compiler does not support all of the operations we require, the
- * loops are written so that we can always fall back on the base types.
- */
-#ifdef CONFIG_VECTOR16
-typedef uint8_t vec8 __attribute__((vector_size(16)));
-typedef uint16_t vec16 __attribute__((vector_size(16)));
-typedef uint32_t vec32 __attribute__((vector_size(16)));
-typedef uint64_t vec64 __attribute__((vector_size(16)));
-
-typedef int8_t svec8 __attribute__((vector_size(16)));
-typedef int16_t svec16 __attribute__((vector_size(16)));
-typedef int32_t svec32 __attribute__((vector_size(16)));
-typedef int64_t svec64 __attribute__((vector_size(16)));
-
-#define DUP16(X)  { X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, X }
-#define DUP8(X)   { X, X, X, X, X, X, X, X }
-#define DUP4(X)   { X, X, X, X }
-#define DUP2(X)   { X, X }
-#else
-typedef uint8_t vec8;
-typedef uint16_t vec16;
-typedef uint32_t vec32;
-typedef uint64_t vec64;
-
-typedef int8_t svec8;
-typedef int16_t svec16;
-typedef int32_t svec32;
-typedef int64_t svec64;
-
-#define DUP16(X)  X
-#define DUP8(X)   X
-#define DUP4(X)   X
-#define DUP2(X)   X
-#endif /* CONFIG_VECTOR16 */
-
 static inline void clear_high(void *d, intptr_t oprsz, uint32_t desc)
 {
     intptr_t maxsz = simd_maxsz(desc);
@@ -83,8 +41,8 @@ void HELPER(gvec_add8)(void *d, void *a, void *b, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec8)) {
-        *(vec8 *)(d + i) = *(vec8 *)(a + i) + *(vec8 *)(b + i);
+    for (i = 0; i < oprsz; i += sizeof(uint8_t)) {
+        *(uint8_t *)(d + i) = *(uint8_t *)(a + i) + *(uint8_t *)(b + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -94,8 +52,8 @@ void HELPER(gvec_add16)(void *d, void *a, void *b, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec16)) {
-        *(vec16 *)(d + i) = *(vec16 *)(a + i) + *(vec16 *)(b + i);
+    for (i = 0; i < oprsz; i += sizeof(uint16_t)) {
+        *(uint16_t *)(d + i) = *(uint16_t *)(a + i) + *(uint16_t *)(b + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -105,8 +63,8 @@ void HELPER(gvec_add32)(void *d, void *a, void *b, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec32)) {
-        *(vec32 *)(d + i) = *(vec32 *)(a + i) + *(vec32 *)(b + i);
+    for (i = 0; i < oprsz; i += sizeof(uint32_t)) {
+        *(uint32_t *)(d + i) = *(uint32_t *)(a + i) + *(uint32_t *)(b + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -116,8 +74,8 @@ void HELPER(gvec_add64)(void *d, void *a, void *b, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(vec64 *)(d + i) = *(vec64 *)(a + i) + *(vec64 *)(b + i);
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(uint64_t *)(d + i) = *(uint64_t *)(a + i) + *(uint64_t *)(b + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -125,11 +83,10 @@ void HELPER(gvec_add64)(void *d, void *a, void *b, uint32_t desc)
 void HELPER(gvec_adds8)(void *d, void *a, uint64_t b, uint32_t desc)
 {
     intptr_t oprsz = simd_oprsz(desc);
-    vec8 vecb = (vec8)DUP16(b);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec8)) {
-        *(vec8 *)(d + i) = *(vec8 *)(a + i) + vecb;
+    for (i = 0; i < oprsz; i += sizeof(uint8_t)) {
+        *(uint8_t *)(d + i) = *(uint8_t *)(a + i) + (uint8_t)b;
     }
     clear_high(d, oprsz, desc);
 }
@@ -137,11 +94,10 @@ void HELPER(gvec_adds8)(void *d, void *a, uint64_t b, uint32_t desc)
 void HELPER(gvec_adds16)(void *d, void *a, uint64_t b, uint32_t desc)
 {
     intptr_t oprsz = simd_oprsz(desc);
-    vec16 vecb = (vec16)DUP8(b);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec16)) {
-        *(vec16 *)(d + i) = *(vec16 *)(a + i) + vecb;
+    for (i = 0; i < oprsz; i += sizeof(uint16_t)) {
+        *(uint16_t *)(d + i) = *(uint16_t *)(a + i) + (uint16_t)b;
     }
     clear_high(d, oprsz, desc);
 }
@@ -149,11 +105,10 @@ void HELPER(gvec_adds16)(void *d, void *a, uint64_t b, uint32_t desc)
 void HELPER(gvec_adds32)(void *d, void *a, uint64_t b, uint32_t desc)
 {
     intptr_t oprsz = simd_oprsz(desc);
-    vec32 vecb = (vec32)DUP4(b);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec32)) {
-        *(vec32 *)(d + i) = *(vec32 *)(a + i) + vecb;
+    for (i = 0; i < oprsz; i += sizeof(uint32_t)) {
+        *(uint32_t *)(d + i) = *(uint32_t *)(a + i) + (uint32_t)b;
     }
     clear_high(d, oprsz, desc);
 }
@@ -161,11 +116,10 @@ void HELPER(gvec_adds32)(void *d, void *a, uint64_t b, uint32_t desc)
 void HELPER(gvec_adds64)(void *d, void *a, uint64_t b, uint32_t desc)
 {
     intptr_t oprsz = simd_oprsz(desc);
-    vec64 vecb = (vec64)DUP2(b);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(vec64 *)(d + i) = *(vec64 *)(a + i) + vecb;
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(uint64_t *)(d + i) = *(uint64_t *)(a + i) + b;
     }
     clear_high(d, oprsz, desc);
 }
@@ -175,8 +129,8 @@ void HELPER(gvec_sub8)(void *d, void *a, void *b, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec8)) {
-        *(vec8 *)(d + i) = *(vec8 *)(a + i) - *(vec8 *)(b + i);
+    for (i = 0; i < oprsz; i += sizeof(uint8_t)) {
+        *(uint8_t *)(d + i) = *(uint8_t *)(a + i) - *(uint8_t *)(b + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -186,8 +140,8 @@ void HELPER(gvec_sub16)(void *d, void *a, void *b, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec16)) {
-        *(vec16 *)(d + i) = *(vec16 *)(a + i) - *(vec16 *)(b + i);
+    for (i = 0; i < oprsz; i += sizeof(uint16_t)) {
+        *(uint16_t *)(d + i) = *(uint16_t *)(a + i) - *(uint16_t *)(b + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -197,8 +151,8 @@ void HELPER(gvec_sub32)(void *d, void *a, void *b, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec32)) {
-        *(vec32 *)(d + i) = *(vec32 *)(a + i) - *(vec32 *)(b + i);
+    for (i = 0; i < oprsz; i += sizeof(uint32_t)) {
+        *(uint32_t *)(d + i) = *(uint32_t *)(a + i) - *(uint32_t *)(b + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -208,8 +162,8 @@ void HELPER(gvec_sub64)(void *d, void *a, void *b, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(vec64 *)(d + i) = *(vec64 *)(a + i) - *(vec64 *)(b + i);
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(uint64_t *)(d + i) = *(uint64_t *)(a + i) - *(uint64_t *)(b + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -217,11 +171,10 @@ void HELPER(gvec_sub64)(void *d, void *a, void *b, uint32_t desc)
 void HELPER(gvec_subs8)(void *d, void *a, uint64_t b, uint32_t desc)
 {
     intptr_t oprsz = simd_oprsz(desc);
-    vec8 vecb = (vec8)DUP16(b);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec8)) {
-        *(vec8 *)(d + i) = *(vec8 *)(a + i) - vecb;
+    for (i = 0; i < oprsz; i += sizeof(uint8_t)) {
+        *(uint8_t *)(d + i) = *(uint8_t *)(a + i) - (uint8_t)b;
     }
     clear_high(d, oprsz, desc);
 }
@@ -229,11 +182,10 @@ void HELPER(gvec_subs8)(void *d, void *a, uint64_t b, uint32_t desc)
 void HELPER(gvec_subs16)(void *d, void *a, uint64_t b, uint32_t desc)
 {
     intptr_t oprsz = simd_oprsz(desc);
-    vec16 vecb = (vec16)DUP8(b);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec16)) {
-        *(vec16 *)(d + i) = *(vec16 *)(a + i) - vecb;
+    for (i = 0; i < oprsz; i += sizeof(uint16_t)) {
+        *(uint16_t *)(d + i) = *(uint16_t *)(a + i) - (uint16_t)b;
     }
     clear_high(d, oprsz, desc);
 }
@@ -241,11 +193,10 @@ void HELPER(gvec_subs16)(void *d, void *a, uint64_t b, uint32_t desc)
 void HELPER(gvec_subs32)(void *d, void *a, uint64_t b, uint32_t desc)
 {
     intptr_t oprsz = simd_oprsz(desc);
-    vec32 vecb = (vec32)DUP4(b);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec32)) {
-        *(vec32 *)(d + i) = *(vec32 *)(a + i) - vecb;
+    for (i = 0; i < oprsz; i += sizeof(uint32_t)) {
+        *(uint32_t *)(d + i) = *(uint32_t *)(a + i) - (uint32_t)b;
     }
     clear_high(d, oprsz, desc);
 }
@@ -253,11 +204,10 @@ void HELPER(gvec_subs32)(void *d, void *a, uint64_t b, uint32_t desc)
 void HELPER(gvec_subs64)(void *d, void *a, uint64_t b, uint32_t desc)
 {
     intptr_t oprsz = simd_oprsz(desc);
-    vec64 vecb = (vec64)DUP2(b);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(vec64 *)(d + i) = *(vec64 *)(a + i) - vecb;
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(uint64_t *)(d + i) = *(uint64_t *)(a + i) - b;
     }
     clear_high(d, oprsz, desc);
 }
@@ -267,8 +217,8 @@ void HELPER(gvec_mul8)(void *d, void *a, void *b, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec8)) {
-        *(vec8 *)(d + i) = *(vec8 *)(a + i) * *(vec8 *)(b + i);
+    for (i = 0; i < oprsz; i += sizeof(uint8_t)) {
+        *(uint8_t *)(d + i) = *(uint8_t *)(a + i) * *(uint8_t *)(b + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -278,8 +228,8 @@ void HELPER(gvec_mul16)(void *d, void *a, void *b, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec16)) {
-        *(vec16 *)(d + i) = *(vec16 *)(a + i) * *(vec16 *)(b + i);
+    for (i = 0; i < oprsz; i += sizeof(uint16_t)) {
+        *(uint16_t *)(d + i) = *(uint16_t *)(a + i) * *(uint16_t *)(b + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -289,8 +239,8 @@ void HELPER(gvec_mul32)(void *d, void *a, void *b, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec32)) {
-        *(vec32 *)(d + i) = *(vec32 *)(a + i) * *(vec32 *)(b + i);
+    for (i = 0; i < oprsz; i += sizeof(uint32_t)) {
+        *(uint32_t *)(d + i) = *(uint32_t *)(a + i) * *(uint32_t *)(b + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -300,8 +250,8 @@ void HELPER(gvec_mul64)(void *d, void *a, void *b, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(vec64 *)(d + i) = *(vec64 *)(a + i) * *(vec64 *)(b + i);
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(uint64_t *)(d + i) = *(uint64_t *)(a + i) * *(uint64_t *)(b + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -309,11 +259,10 @@ void HELPER(gvec_mul64)(void *d, void *a, void *b, uint32_t desc)
 void HELPER(gvec_muls8)(void *d, void *a, uint64_t b, uint32_t desc)
 {
     intptr_t oprsz = simd_oprsz(desc);
-    vec8 vecb = (vec8)DUP16(b);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec8)) {
-        *(vec8 *)(d + i) = *(vec8 *)(a + i) * vecb;
+    for (i = 0; i < oprsz; i += sizeof(uint8_t)) {
+        *(uint8_t *)(d + i) = *(uint8_t *)(a + i) * (uint8_t)b;
     }
     clear_high(d, oprsz, desc);
 }
@@ -321,11 +270,10 @@ void HELPER(gvec_muls8)(void *d, void *a, uint64_t b, uint32_t desc)
 void HELPER(gvec_muls16)(void *d, void *a, uint64_t b, uint32_t desc)
 {
     intptr_t oprsz = simd_oprsz(desc);
-    vec16 vecb = (vec16)DUP8(b);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec16)) {
-        *(vec16 *)(d + i) = *(vec16 *)(a + i) * vecb;
+    for (i = 0; i < oprsz; i += sizeof(uint16_t)) {
+        *(uint16_t *)(d + i) = *(uint16_t *)(a + i) * (uint16_t)b;
     }
     clear_high(d, oprsz, desc);
 }
@@ -333,11 +281,10 @@ void HELPER(gvec_muls16)(void *d, void *a, uint64_t b, uint32_t desc)
 void HELPER(gvec_muls32)(void *d, void *a, uint64_t b, uint32_t desc)
 {
     intptr_t oprsz = simd_oprsz(desc);
-    vec32 vecb = (vec32)DUP4(b);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec32)) {
-        *(vec32 *)(d + i) = *(vec32 *)(a + i) * vecb;
+    for (i = 0; i < oprsz; i += sizeof(uint32_t)) {
+        *(uint32_t *)(d + i) = *(uint32_t *)(a + i) * (uint32_t)b;
     }
     clear_high(d, oprsz, desc);
 }
@@ -345,11 +292,10 @@ void HELPER(gvec_muls32)(void *d, void *a, uint64_t b, uint32_t desc)
 void HELPER(gvec_muls64)(void *d, void *a, uint64_t b, uint32_t desc)
 {
     intptr_t oprsz = simd_oprsz(desc);
-    vec64 vecb = (vec64)DUP2(b);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(vec64 *)(d + i) = *(vec64 *)(a + i) * vecb;
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(uint64_t *)(d + i) = *(uint64_t *)(a + i) * b;
     }
     clear_high(d, oprsz, desc);
 }
@@ -359,8 +305,8 @@ void HELPER(gvec_neg8)(void *d, void *a, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec8)) {
-        *(vec8 *)(d + i) = -*(vec8 *)(a + i);
+    for (i = 0; i < oprsz; i += sizeof(uint8_t)) {
+        *(uint8_t *)(d + i) = -*(uint8_t *)(a + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -370,8 +316,8 @@ void HELPER(gvec_neg16)(void *d, void *a, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec16)) {
-        *(vec16 *)(d + i) = -*(vec16 *)(a + i);
+    for (i = 0; i < oprsz; i += sizeof(uint16_t)) {
+        *(uint16_t *)(d + i) = -*(uint16_t *)(a + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -381,8 +327,8 @@ void HELPER(gvec_neg32)(void *d, void *a, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec32)) {
-        *(vec32 *)(d + i) = -*(vec32 *)(a + i);
+    for (i = 0; i < oprsz; i += sizeof(uint32_t)) {
+        *(uint32_t *)(d + i) = -*(uint32_t *)(a + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -392,8 +338,8 @@ void HELPER(gvec_neg64)(void *d, void *a, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(vec64 *)(d + i) = -*(vec64 *)(a + i);
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(uint64_t *)(d + i) = -*(uint64_t *)(a + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -499,8 +445,8 @@ void HELPER(gvec_not)(void *d, void *a, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(vec64 *)(d + i) = ~*(vec64 *)(a + i);
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(uint64_t *)(d + i) = ~*(uint64_t *)(a + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -510,8 +456,8 @@ void HELPER(gvec_and)(void *d, void *a, void *b, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(vec64 *)(d + i) = *(vec64 *)(a + i) & *(vec64 *)(b + i);
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(uint64_t *)(d + i) = *(uint64_t *)(a + i) & *(uint64_t *)(b + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -521,8 +467,8 @@ void HELPER(gvec_or)(void *d, void *a, void *b, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(vec64 *)(d + i) = *(vec64 *)(a + i) | *(vec64 *)(b + i);
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(uint64_t *)(d + i) = *(uint64_t *)(a + i) | *(uint64_t *)(b + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -532,8 +478,8 @@ void HELPER(gvec_xor)(void *d, void *a, void *b, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(vec64 *)(d + i) = *(vec64 *)(a + i) ^ *(vec64 *)(b + i);
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(uint64_t *)(d + i) = *(uint64_t *)(a + i) ^ *(uint64_t *)(b + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -543,8 +489,8 @@ void HELPER(gvec_andc)(void *d, void *a, void *b, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(vec64 *)(d + i) = *(vec64 *)(a + i) &~ *(vec64 *)(b + i);
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(uint64_t *)(d + i) = *(uint64_t *)(a + i) &~ *(uint64_t *)(b + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -554,8 +500,8 @@ void HELPER(gvec_orc)(void *d, void *a, void *b, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(vec64 *)(d + i) = *(vec64 *)(a + i) |~ *(vec64 *)(b + i);
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(uint64_t *)(d + i) = *(uint64_t *)(a + i) |~ *(uint64_t *)(b + i);
     }
     clear_high(d, oprsz, desc);
 }
@@ -565,8 +511,8 @@ void HELPER(gvec_nand)(void *d, void *a, void *b, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(vec64 *)(d + i) = ~(*(vec64 *)(a + i) & *(vec64 *)(b + i));
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(uint64_t *)(d + i) = ~(*(uint64_t *)(a + i) & *(uint64_t *)(b + i));
     }
     clear_high(d, oprsz, desc);
 }
@@ -576,8 +522,8 @@ void HELPER(gvec_nor)(void *d, void *a, void *b, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(vec64 *)(d + i) = ~(*(vec64 *)(a + i) | *(vec64 *)(b + i));
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(uint64_t *)(d + i) = ~(*(uint64_t *)(a + i) | *(uint64_t *)(b + i));
     }
     clear_high(d, oprsz, desc);
 }
@@ -587,8 +533,8 @@ void HELPER(gvec_eqv)(void *d, void *a, void *b, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(vec64 *)(d + i) = ~(*(vec64 *)(a + i) ^ *(vec64 *)(b + i));
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(uint64_t *)(d + i) = ~(*(uint64_t *)(a + i) ^ *(uint64_t *)(b + i));
     }
     clear_high(d, oprsz, desc);
 }
@@ -596,11 +542,10 @@ void HELPER(gvec_eqv)(void *d, void *a, void *b, uint32_t desc)
 void HELPER(gvec_ands)(void *d, void *a, uint64_t b, uint32_t desc)
 {
     intptr_t oprsz = simd_oprsz(desc);
-    vec64 vecb = (vec64)DUP2(b);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(vec64 *)(d + i) = *(vec64 *)(a + i) & vecb;
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(uint64_t *)(d + i) = *(uint64_t *)(a + i) & b;
     }
     clear_high(d, oprsz, desc);
 }
@@ -608,11 +553,10 @@ void HELPER(gvec_ands)(void *d, void *a, uint64_t b, uint32_t desc)
 void HELPER(gvec_xors)(void *d, void *a, uint64_t b, uint32_t desc)
 {
     intptr_t oprsz = simd_oprsz(desc);
-    vec64 vecb = (vec64)DUP2(b);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(vec64 *)(d + i) = *(vec64 *)(a + i) ^ vecb;
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(uint64_t *)(d + i) = *(uint64_t *)(a + i) ^ b;
     }
     clear_high(d, oprsz, desc);
 }
@@ -620,11 +564,10 @@ void HELPER(gvec_xors)(void *d, void *a, uint64_t b, uint32_t desc)
 void HELPER(gvec_ors)(void *d, void *a, uint64_t b, uint32_t desc)
 {
     intptr_t oprsz = simd_oprsz(desc);
-    vec64 vecb = (vec64)DUP2(b);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(vec64 *)(d + i) = *(vec64 *)(a + i) | vecb;
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(uint64_t *)(d + i) = *(uint64_t *)(a + i) | b;
     }
     clear_high(d, oprsz, desc);
 }
@@ -635,8 +578,8 @@ void HELPER(gvec_shl8i)(void *d, void *a, uint32_t desc)
     int shift = simd_data(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec8)) {
-        *(vec8 *)(d + i) = *(vec8 *)(a + i) << shift;
+    for (i = 0; i < oprsz; i += sizeof(uint8_t)) {
+        *(uint8_t *)(d + i) = *(uint8_t *)(a + i) << shift;
     }
     clear_high(d, oprsz, desc);
 }
@@ -647,8 +590,8 @@ void HELPER(gvec_shl16i)(void *d, void *a, uint32_t desc)
     int shift = simd_data(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec16)) {
-        *(vec16 *)(d + i) = *(vec16 *)(a + i) << shift;
+    for (i = 0; i < oprsz; i += sizeof(uint16_t)) {
+        *(uint16_t *)(d + i) = *(uint16_t *)(a + i) << shift;
     }
     clear_high(d, oprsz, desc);
 }
@@ -659,8 +602,8 @@ void HELPER(gvec_shl32i)(void *d, void *a, uint32_t desc)
     int shift = simd_data(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec32)) {
-        *(vec32 *)(d + i) = *(vec32 *)(a + i) << shift;
+    for (i = 0; i < oprsz; i += sizeof(uint32_t)) {
+        *(uint32_t *)(d + i) = *(uint32_t *)(a + i) << shift;
     }
     clear_high(d, oprsz, desc);
 }
@@ -671,8 +614,8 @@ void HELPER(gvec_shl64i)(void *d, void *a, uint32_t desc)
     int shift = simd_data(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(vec64 *)(d + i) = *(vec64 *)(a + i) << shift;
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(uint64_t *)(d + i) = *(uint64_t *)(a + i) << shift;
     }
     clear_high(d, oprsz, desc);
 }
@@ -683,8 +626,8 @@ void HELPER(gvec_shr8i)(void *d, void *a, uint32_t desc)
     int shift = simd_data(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec8)) {
-        *(vec8 *)(d + i) = *(vec8 *)(a + i) >> shift;
+    for (i = 0; i < oprsz; i += sizeof(uint8_t)) {
+        *(uint8_t *)(d + i) = *(uint8_t *)(a + i) >> shift;
     }
     clear_high(d, oprsz, desc);
 }
@@ -695,8 +638,8 @@ void HELPER(gvec_shr16i)(void *d, void *a, uint32_t desc)
     int shift = simd_data(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec16)) {
-        *(vec16 *)(d + i) = *(vec16 *)(a + i) >> shift;
+    for (i = 0; i < oprsz; i += sizeof(uint16_t)) {
+        *(uint16_t *)(d + i) = *(uint16_t *)(a + i) >> shift;
     }
     clear_high(d, oprsz, desc);
 }
@@ -707,8 +650,8 @@ void HELPER(gvec_shr32i)(void *d, void *a, uint32_t desc)
     int shift = simd_data(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec32)) {
-        *(vec32 *)(d + i) = *(vec32 *)(a + i) >> shift;
+    for (i = 0; i < oprsz; i += sizeof(uint32_t)) {
+        *(uint32_t *)(d + i) = *(uint32_t *)(a + i) >> shift;
     }
     clear_high(d, oprsz, desc);
 }
@@ -719,8 +662,8 @@ void HELPER(gvec_shr64i)(void *d, void *a, uint32_t desc)
     int shift = simd_data(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(vec64 *)(d + i) = *(vec64 *)(a + i) >> shift;
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(uint64_t *)(d + i) = *(uint64_t *)(a + i) >> shift;
     }
     clear_high(d, oprsz, desc);
 }
@@ -731,8 +674,8 @@ void HELPER(gvec_sar8i)(void *d, void *a, uint32_t desc)
     int shift = simd_data(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec8)) {
-        *(svec8 *)(d + i) = *(svec8 *)(a + i) >> shift;
+    for (i = 0; i < oprsz; i += sizeof(uint8_t)) {
+        *(int8_t *)(d + i) = *(int8_t *)(a + i) >> shift;
     }
     clear_high(d, oprsz, desc);
 }
@@ -743,8 +686,8 @@ void HELPER(gvec_sar16i)(void *d, void *a, uint32_t desc)
     int shift = simd_data(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec16)) {
-        *(svec16 *)(d + i) = *(svec16 *)(a + i) >> shift;
+    for (i = 0; i < oprsz; i += sizeof(uint16_t)) {
+        *(int16_t *)(d + i) = *(int16_t *)(a + i) >> shift;
     }
     clear_high(d, oprsz, desc);
 }
@@ -755,8 +698,8 @@ void HELPER(gvec_sar32i)(void *d, void *a, uint32_t desc)
     int shift = simd_data(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec32)) {
-        *(svec32 *)(d + i) = *(svec32 *)(a + i) >> shift;
+    for (i = 0; i < oprsz; i += sizeof(uint32_t)) {
+        *(int32_t *)(d + i) = *(int32_t *)(a + i) >> shift;
     }
     clear_high(d, oprsz, desc);
 }
@@ -767,8 +710,8 @@ void HELPER(gvec_sar64i)(void *d, void *a, uint32_t desc)
     int shift = simd_data(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        *(svec64 *)(d + i) = *(svec64 *)(a + i) >> shift;
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        *(int64_t *)(d + i) = *(int64_t *)(a + i) >> shift;
     }
     clear_high(d, oprsz, desc);
 }
@@ -917,39 +860,30 @@ void HELPER(gvec_sar64v)(void *d, void *a, void *b, uint32_t desc)
     clear_high(d, oprsz, desc);
 }
 
-/* If vectors are enabled, the compiler fills in -1 for true.
-   Otherwise, we must take care of this by hand.  */
-#ifdef CONFIG_VECTOR16
-# define DO_CMP0(X)  X
-#else
-# define DO_CMP0(X)  -(X)
-#endif
-
 #define DO_CMP1(NAME, TYPE, OP)                                            \
 void HELPER(NAME)(void *d, void *a, void *b, uint32_t desc)                \
 {                                                                          \
     intptr_t oprsz = simd_oprsz(desc);                                     \
     intptr_t i;                                                            \
     for (i = 0; i < oprsz; i += sizeof(TYPE)) {                            \
-        *(TYPE *)(d + i) = DO_CMP0(*(TYPE *)(a + i) OP *(TYPE *)(b + i));  \
+        *(TYPE *)(d + i) = -(*(TYPE *)(a + i) OP *(TYPE *)(b + i));        \
     }                                                                      \
     clear_high(d, oprsz, desc);                                            \
 }
 
 #define DO_CMP2(SZ) \
-    DO_CMP1(gvec_eq##SZ, vec##SZ, ==)    \
-    DO_CMP1(gvec_ne##SZ, vec##SZ, !=)    \
-    DO_CMP1(gvec_lt##SZ, svec##SZ, <)    \
-    DO_CMP1(gvec_le##SZ, svec##SZ, <=)   \
-    DO_CMP1(gvec_ltu##SZ, vec##SZ, <)    \
-    DO_CMP1(gvec_leu##SZ, vec##SZ, <=)
+    DO_CMP1(gvec_eq##SZ, uint##SZ##_t, ==)    \
+    DO_CMP1(gvec_ne##SZ, uint##SZ##_t, !=)    \
+    DO_CMP1(gvec_lt##SZ, int##SZ##_t, <)      \
+    DO_CMP1(gvec_le##SZ, int##SZ##_t, <=)     \
+    DO_CMP1(gvec_ltu##SZ, uint##SZ##_t, <)    \
+    DO_CMP1(gvec_leu##SZ, uint##SZ##_t, <=)
 
 DO_CMP2(8)
 DO_CMP2(16)
 DO_CMP2(32)
 DO_CMP2(64)
 
-#undef DO_CMP0
 #undef DO_CMP1
 #undef DO_CMP2
 
@@ -1450,11 +1384,11 @@ void HELPER(gvec_bitsel)(void *d, void *a, void *b, void *c, uint32_t desc)
     intptr_t oprsz = simd_oprsz(desc);
     intptr_t i;
 
-    for (i = 0; i < oprsz; i += sizeof(vec64)) {
-        vec64 aa = *(vec64 *)(a + i);
-        vec64 bb = *(vec64 *)(b + i);
-        vec64 cc = *(vec64 *)(c + i);
-        *(vec64 *)(d + i) = (bb & aa) | (cc & ~aa);
+    for (i = 0; i < oprsz; i += sizeof(uint64_t)) {
+        uint64_t aa = *(uint64_t *)(a + i);
+        uint64_t bb = *(uint64_t *)(b + i);
+        uint64_t cc = *(uint64_t *)(c + i);
+        *(uint64_t *)(d + i) = (bb & aa) | (cc & ~aa);
     }
     clear_high(d, oprsz, desc);
 }
diff --git a/configure b/configure
index 78d59b7..206d22c 100755
--- a/configure
+++ b/configure
@@ -5765,58 +5765,6 @@ if  test "$plugins" = "yes" &&
 fi
 
 ########################################
-# See if 16-byte vector operations are supported.
-# Even without a vector unit the compiler may expand these.
-# There is a bug in old GCC for PPC that crashes here.
-# Unfortunately it's the system compiler for Centos 7.
-
-cat > $TMPC << EOF
-typedef unsigned char U1 __attribute__((vector_size(16)));
-typedef unsigned short U2 __attribute__((vector_size(16)));
-typedef unsigned int U4 __attribute__((vector_size(16)));
-typedef unsigned long long U8 __attribute__((vector_size(16)));
-typedef signed char S1 __attribute__((vector_size(16)));
-typedef signed short S2 __attribute__((vector_size(16)));
-typedef signed int S4 __attribute__((vector_size(16)));
-typedef signed long long S8 __attribute__((vector_size(16)));
-static U1 a1, b1;
-static U2 a2, b2;
-static U4 a4, b4;
-static U8 a8, b8;
-static S1 c1;
-static S2 c2;
-static S4 c4;
-static S8 c8;
-static int i;
-void helper(void *d, void *a, int shift, int i);
-void helper(void *d, void *a, int shift, int i)
-{
-  *(U1 *)(d + i) = *(U1 *)(a + i) << shift;
-  *(U2 *)(d + i) = *(U2 *)(a + i) << shift;
-  *(U4 *)(d + i) = *(U4 *)(a + i) << shift;
-  *(U8 *)(d + i) = *(U8 *)(a + i) << shift;
-}
-int main(void)
-{
-  a1 += b1; a2 += b2; a4 += b4; a8 += b8;
-  a1 -= b1; a2 -= b2; a4 -= b4; a8 -= b8;
-  a1 *= b1; a2 *= b2; a4 *= b4; a8 *= b8;
-  a1 &= b1; a2 &= b2; a4 &= b4; a8 &= b8;
-  a1 |= b1; a2 |= b2; a4 |= b4; a8 |= b8;
-  a1 ^= b1; a2 ^= b2; a4 ^= b4; a8 ^= b8;
-  a1 <<= i; a2 <<= i; a4 <<= i; a8 <<= i;
-  a1 >>= i; a2 >>= i; a4 >>= i; a8 >>= i;
-  c1 >>= i; c2 >>= i; c4 >>= i; c8 >>= i;
-  return 0;
-}
-EOF
-
-vector16=no
-if compile_prog "" "" ; then
-  vector16=yes
-fi
-
-########################################
 # See if __attribute__((alias)) is supported.
 # This false for Xcode 9, but has been remedied for Xcode 10.
 # Unfortunately, travis uses Xcode 9 by default.
@@ -7446,10 +7394,6 @@ if test "$atomic64" = "yes" ; then
   echo "CONFIG_ATOMIC64=y" >> $config_host_mak
 fi
 
-if test "$vector16" = "yes" ; then
-  echo "CONFIG_VECTOR16=y" >> $config_host_mak
-fi
-
 if test "$attralias" = "yes" ; then
   echo "CONFIG_ATTRIBUTE_ALIAS=y" >> $config_host_mak
 fi
diff --git a/tcg/i386/tcg-target.inc.c b/tcg/i386/tcg-target.inc.c
index cdedcb2..223dba9 100644
--- a/tcg/i386/tcg-target.inc.c
+++ b/tcg/i386/tcg-target.inc.c
@@ -3391,12 +3391,15 @@ static void expand_vec_sari(TCGType type, unsigned vece,
 
     case MO_64:
         if (imm <= 32) {
-            /* We can emulate a small sign extend by performing an arithmetic
+            /*
+             * We can emulate a small sign extend by performing an arithmetic
              * 32-bit shift and overwriting the high half of a 64-bit logical
-             * shift (note that the ISA says shift of 32 is valid).
+             * shift.  Note that the ISA says shift of 32 is valid, but TCG
+             * does not, so we have to bound the smaller shift -- we get the
+             * same result in the high half either way.
              */
             t1 = tcg_temp_new_vec(type);
-            tcg_gen_sari_vec(MO_32, t1, v1, imm);
+            tcg_gen_sari_vec(MO_32, t1, v1, MIN(imm, 31));
             tcg_gen_shri_vec(MO_64, v0, v1, imm);
             vec_gen_4(INDEX_op_x86_blend_vec, type, MO_32,
                       tcgv_vec_arg(v0), tcgv_vec_arg(v0),