[Qemu-devel] [PULL 16/19] target-tilegx: Implement v?int_* instructions.

[Richard Henderson <rth@twiddle.net>]

From: Chen Gang <gang.chen.5i5j@gmail.com>

Signed-off-by: Chen Gang <gang.chen.5i5j@gmail.com>
Message-Id: <1443956491-26850-2-git-send-email-gang.chen.5i5j@gmail.com>
Signed-off-by: Richard Henderson <rth@twiddle.net>
---
 target-tilegx/helper.h      |  5 +++++
 target-tilegx/simd_helper.c | 48 +++++++++++++++++++++++++++++++++++++++++++++
 target-tilegx/translate.c   | 14 +++++++++++++
 3 files changed, 67 insertions(+)

diff --git a/target-tilegx/helper.h b/target-tilegx/helper.h
index 82d84f1..c58ee20 100644
--- a/target-tilegx/helper.h
+++ b/target-tilegx/helper.h
@@ -10,6 +10,11 @@ DEF_HELPER_FLAGS_3(cmula, TCG_CALL_NO_RWG_SE, i64, i64, i64, i64)
 DEF_HELPER_FLAGS_3(cmulaf, TCG_CALL_NO_RWG_SE, i64, i64, i64, i64)
 DEF_HELPER_FLAGS_4(cmul2, TCG_CALL_NO_RWG_SE, i64, i64, i64, int, int)
 
+DEF_HELPER_FLAGS_2(v1int_h, TCG_CALL_NO_RWG_SE, i64, i64, i64)
+DEF_HELPER_FLAGS_2(v1int_l, TCG_CALL_NO_RWG_SE, i64, i64, i64)
+DEF_HELPER_FLAGS_2(v2int_h, TCG_CALL_NO_RWG_SE, i64, i64, i64)
+DEF_HELPER_FLAGS_2(v2int_l, TCG_CALL_NO_RWG_SE, i64, i64, i64)
+
 DEF_HELPER_FLAGS_2(v1multu, TCG_CALL_NO_RWG_SE, i64, i64, i64)
 DEF_HELPER_FLAGS_2(v1shl, TCG_CALL_NO_RWG_SE, i64, i64, i64)
 DEF_HELPER_FLAGS_2(v1shru, TCG_CALL_NO_RWG_SE, i64, i64, i64)
diff --git a/target-tilegx/simd_helper.c b/target-tilegx/simd_helper.c
index 2cff43c..d294671 100644
--- a/target-tilegx/simd_helper.c
+++ b/target-tilegx/simd_helper.c
@@ -100,3 +100,51 @@ uint64_t helper_v2shrs(uint64_t a, uint64_t b)
     }
     return r;
 }
+
+uint64_t helper_v1int_h(uint64_t a, uint64_t b)
+{
+    uint64_t r = 0;
+    int i;
+
+    for (i = 0; i < 32; i += 8) {
+        r = deposit64(r, 2 * i + 8, 8, extract64(a, i + 32, 8));
+        r = deposit64(r, 2 * i, 8, extract64(b, i + 32, 8));
+    }
+    return r;
+}
+
+uint64_t helper_v1int_l(uint64_t a, uint64_t b)
+{
+    uint64_t r = 0;
+    int i;
+
+    for (i = 0; i < 32; i += 8) {
+        r = deposit64(r, 2 * i + 8, 8, extract64(a, i, 8));
+        r = deposit64(r, 2 * i, 8, extract64(b, i, 8));
+    }
+    return r;
+}
+
+uint64_t helper_v2int_h(uint64_t a, uint64_t b)
+{
+    uint64_t r = 0;
+    int i;
+
+    for (i = 0; i < 32; i += 16) {
+        r = deposit64(r, 2 * i + 16, 16, extract64(a, i + 32, 16));
+        r = deposit64(r, 2 * i, 16, extract64(b, i + 32, 16));
+    }
+    return r;
+}
+
+uint64_t helper_v2int_l(uint64_t a, uint64_t b)
+{
+    uint64_t r = 0;
+    int i;
+
+    for (i = 0; i < 32; i += 16) {
+        r = deposit64(r, 2 * i + 16, 16, extract64(a, i, 16));
+        r = deposit64(r, 2 * i, 16, extract64(b, i, 16));
+    }
+    return r;
+}
diff --git a/target-tilegx/translate.c b/target-tilegx/translate.c
index 9bb8857..034cbc2 100644
--- a/target-tilegx/translate.c
+++ b/target-tilegx/translate.c
@@ -1260,10 +1260,17 @@ static TileExcp gen_rrr_opcode(DisasContext *dc, unsigned opext,
     case OE_RRR(V1DOTPUS, 0, X0):
     case OE_RRR(V1DOTPU, 0, X0):
     case OE_RRR(V1DOTP, 0, X0):
+        return TILEGX_EXCP_OPCODE_UNIMPLEMENTED;
     case OE_RRR(V1INT_H, 0, X0):
     case OE_RRR(V1INT_H, 0, X1):
+        gen_helper_v1int_h(tdest, tsrca, tsrcb);
+        mnemonic = "v1int_h";
+        break;
     case OE_RRR(V1INT_L, 0, X0):
     case OE_RRR(V1INT_L, 0, X1):
+        gen_helper_v1int_l(tdest, tsrca, tsrcb);
+        mnemonic = "v1int_l";
+        break;
     case OE_RRR(V1MAXU, 0, X0):
     case OE_RRR(V1MAXU, 0, X1):
     case OE_RRR(V1MINU, 0, X0):
@@ -1329,10 +1336,17 @@ static TileExcp gen_rrr_opcode(DisasContext *dc, unsigned opext,
     case OE_RRR(V2CMPNE, 0, X1):
     case OE_RRR(V2DOTPA, 0, X0):
     case OE_RRR(V2DOTP, 0, X0):
+        return TILEGX_EXCP_OPCODE_UNIMPLEMENTED;
     case OE_RRR(V2INT_H, 0, X0):
     case OE_RRR(V2INT_H, 0, X1):
+        gen_helper_v2int_h(tdest, tsrca, tsrcb);
+        mnemonic = "v2int_h";
+        break;
     case OE_RRR(V2INT_L, 0, X0):
     case OE_RRR(V2INT_L, 0, X1):
+        gen_helper_v2int_l(tdest, tsrca, tsrcb);
+        mnemonic = "v2int_l";
+        break;
     case OE_RRR(V2MAXS, 0, X0):
     case OE_RRR(V2MAXS, 0, X1):
     case OE_RRR(V2MINS, 0, X0):
-- 
2.4.3