[Qemu-devel] [PULL SUBSYSTEM s390x 03/33] s390x/tcg: Implement VECTOR FIND ELEMENT NOT EQUAL

[David Hildenbrand <david@redhat.com>]

Similar to VECTOR FIND ELEMENT EQUAL. Core logic courtesy of Richard H.

Add s390_vec_read_element() that can deal with element sizes.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>
---
 target/s390x/helper.h            |  6 +++
 target/s390x/insn-data.def       |  2 +
 target/s390x/translate_vx.inc.c  | 31 +++++++++++++
 target/s390x/vec.h               | 19 ++++++++
 target/s390x/vec_string_helper.c | 74 ++++++++++++++++++++++++++++++++
 5 files changed, 132 insertions(+)

diff --git a/target/s390x/helper.h b/target/s390x/helper.h
index a1b169b666..fb50b404db 100644
--- a/target/s390x/helper.h
+++ b/target/s390x/helper.h
@@ -224,6 +224,12 @@ DEF_HELPER_FLAGS_4(gvec_vfee32, TCG_CALL_NO_RWG, void, ptr, cptr, cptr, i32)
 DEF_HELPER_5(gvec_vfee_cc8, void, ptr, cptr, cptr, env, i32)
 DEF_HELPER_5(gvec_vfee_cc16, void, ptr, cptr, cptr, env, i32)
 DEF_HELPER_5(gvec_vfee_cc32, void, ptr, cptr, cptr, env, i32)
+DEF_HELPER_FLAGS_4(gvec_vfene8, TCG_CALL_NO_RWG, void, ptr, cptr, cptr, i32)
+DEF_HELPER_FLAGS_4(gvec_vfene16, TCG_CALL_NO_RWG, void, ptr, cptr, cptr, i32)
+DEF_HELPER_FLAGS_4(gvec_vfene32, TCG_CALL_NO_RWG, void, ptr, cptr, cptr, i32)
+DEF_HELPER_5(gvec_vfene_cc8, void, ptr, cptr, cptr, env, i32)
+DEF_HELPER_5(gvec_vfene_cc16, void, ptr, cptr, cptr, env, i32)
+DEF_HELPER_5(gvec_vfene_cc32, void, ptr, cptr, cptr, env, i32)
 
 #ifndef CONFIG_USER_ONLY
 DEF_HELPER_3(servc, i32, env, i64, i64)
diff --git a/target/s390x/insn-data.def b/target/s390x/insn-data.def
index d8907ef6a5..d03c1ee0b3 100644
--- a/target/s390x/insn-data.def
+++ b/target/s390x/insn-data.def
@@ -1197,6 +1197,8 @@
     F(0xe782, VFAE,    VRR_b, V,   0, 0, 0, 0, vfae, 0, IF_VEC)
 /* VECTOR FIND ELEMENT EQUAL */
     F(0xe780, VFEE,    VRR_b, V,   0, 0, 0, 0, vfee, 0, IF_VEC)
+/* VECTOR FIND ELEMENT NOT EQUAL */
+    F(0xe781, VFENE,   VRR_b, V,   0, 0, 0, 0, vfene, 0, IF_VEC)
 
 #ifndef CONFIG_USER_ONLY
 /* COMPARE AND SWAP AND PURGE */
diff --git a/target/s390x/translate_vx.inc.c b/target/s390x/translate_vx.inc.c
index b25afbc011..1ad0b62517 100644
--- a/target/s390x/translate_vx.inc.c
+++ b/target/s390x/translate_vx.inc.c
@@ -2414,3 +2414,34 @@ static DisasJumpType op_vfee(DisasContext *s, DisasOps *o)
     }
     return DISAS_NEXT;
 }
+
+static DisasJumpType op_vfene(DisasContext *s, DisasOps *o)
+{
+    const uint8_t es = get_field(s->fields, m4);
+    const uint8_t m5 = get_field(s->fields, m5);
+    static gen_helper_gvec_3 * const g[3] = {
+        gen_helper_gvec_vfene8,
+        gen_helper_gvec_vfene16,
+        gen_helper_gvec_vfene32,
+    };
+    static gen_helper_gvec_3_ptr * const g_cc[3] = {
+        gen_helper_gvec_vfene_cc8,
+        gen_helper_gvec_vfene_cc16,
+        gen_helper_gvec_vfene_cc32,
+    };
+
+    if (es > ES_32 || m5 & ~0x3) {
+        gen_program_exception(s, PGM_SPECIFICATION);
+        return DISAS_NORETURN;
+    }
+
+    if (extract32(m5, 0, 1)) {
+        gen_gvec_3_ptr(get_field(s->fields, v1), get_field(s->fields, v2),
+                       get_field(s->fields, v3), cpu_env, m5, g_cc[es]);
+        set_cc_static(s);
+    } else {
+        gen_gvec_3_ool(get_field(s->fields, v1), get_field(s->fields, v2),
+                       get_field(s->fields, v3), m5, g[es]);
+    }
+    return DISAS_NEXT;
+}
diff --git a/target/s390x/vec.h b/target/s390x/vec.h
index 3313fb43ee..affc62874c 100644
--- a/target/s390x/vec.h
+++ b/target/s390x/vec.h
@@ -12,6 +12,8 @@
 #ifndef S390X_VEC_H
 #define S390X_VEC_H
 
+#include "tcg/tcg.h"
+
 typedef union S390Vector {
     uint64_t doubleword[2];
     uint32_t word[4];
@@ -70,6 +72,23 @@ static inline uint64_t s390_vec_read_element64(const S390Vector *v, uint8_t enr)
     return v->doubleword[enr];
 }
 
+static inline uint64_t s390_vec_read_element(const S390Vector *v, uint8_t enr,
+                                             uint8_t es)
+{
+    switch (es) {
+    case MO_8:
+        return s390_vec_read_element8(v, enr);
+    case MO_16:
+        return s390_vec_read_element16(v, enr);
+    case MO_32:
+        return s390_vec_read_element32(v, enr);
+    case MO_64:
+        return s390_vec_read_element64(v, enr);
+    default:
+        g_assert_not_reached();
+    }
+}
+
 static inline void s390_vec_write_element8(S390Vector *v, uint8_t enr,
                                            uint8_t data)
 {
diff --git a/target/s390x/vec_string_helper.c b/target/s390x/vec_string_helper.c
index 05ad99e173..0ee3470112 100644
--- a/target/s390x/vec_string_helper.c
+++ b/target/s390x/vec_string_helper.c
@@ -27,6 +27,15 @@ static inline uint64_t zero_search(uint64_t a, uint64_t mask)
     return ~(((a & mask) + mask) | a | mask);
 }
 
+/*
+ * Returns a bit set in the MSB of each element that is not zero,
+ * as defined by the mask.
+ */
+static inline uint64_t nonzero_search(uint64_t a, uint64_t mask)
+{
+    return (((a & mask) + mask) | a) & ~mask;
+}
+
 /*
  * Returns the byte offset for the first match, or 16 for no match.
  */
@@ -209,3 +218,68 @@ void HELPER(gvec_vfee_cc##BITS)(void *v1, const void *v2, const void *v3,      \
 DEF_VFEE_CC_HELPER(8)
 DEF_VFEE_CC_HELPER(16)
 DEF_VFEE_CC_HELPER(32)
+
+static int vfene(void *v1, const void *v2, const void *v3, bool zs, uint8_t es)
+{
+    const uint64_t mask = get_element_lsbs_mask(es);
+    uint64_t a0, a1, b0, b1, e0, e1, z0, z1;
+    uint64_t first_zero = 16;
+    uint64_t first_inequal;
+    bool smaller = false;
+
+    a0 = s390_vec_read_element64(v2, 0);
+    a1 = s390_vec_read_element64(v2, 1);
+    b0 = s390_vec_read_element64(v3, 0);
+    b1 = s390_vec_read_element64(v3, 1);
+    e0 = nonzero_search(a0 ^ b0, mask);
+    e1 = nonzero_search(a1 ^ b1, mask);
+    first_inequal = match_index(e0, e1);
+
+    /* identify the smaller element */
+    if (first_inequal < 16) {
+        uint8_t enr = first_inequal / (1 << es);
+        uint32_t a = s390_vec_read_element(v2, enr, es);
+        uint32_t b = s390_vec_read_element(v3, enr, es);
+
+        smaller = a < b;
+    }
+
+    if (zs) {
+        z0 = zero_search(a0, mask);
+        z1 = zero_search(a1, mask);
+        first_zero = match_index(z0, z1);
+    }
+
+    s390_vec_write_element64(v1, 0, MIN(first_inequal, first_zero));
+    s390_vec_write_element64(v1, 1, 0);
+    if (first_zero == 16 && first_inequal == 16) {
+        return 3;
+    } else if (first_zero < first_inequal) {
+        return 0;
+    }
+    return smaller ? 1 : 2;
+}
+
+#define DEF_VFENE_HELPER(BITS)                                                 \
+void HELPER(gvec_vfene##BITS)(void *v1, const void *v2, const void *v3,        \
+                              uint32_t desc)                                   \
+{                                                                              \
+    const bool zs = extract32(simd_data(desc), 1, 1);                          \
+                                                                               \
+    vfene(v1, v2, v3, zs, MO_##BITS);                                          \
+}
+DEF_VFENE_HELPER(8)
+DEF_VFENE_HELPER(16)
+DEF_VFENE_HELPER(32)
+
+#define DEF_VFENE_CC_HELPER(BITS)                                              \
+void HELPER(gvec_vfene_cc##BITS)(void *v1, const void *v2, const void *v3,     \
+                                 CPUS390XState *env, uint32_t desc)            \
+{                                                                              \
+    const bool zs = extract32(simd_data(desc), 1, 1);                          \
+                                                                               \
+    env->cc_op = vfene(v1, v2, v3, zs, MO_##BITS);                             \
+}
+DEF_VFENE_CC_HELPER(8)
+DEF_VFENE_CC_HELPER(16)
+DEF_VFENE_CC_HELPER(32)
-- 
2.21.0