[Qemu-devel] [RFC PATCH v1 11/22] target/i386: reimplement (V)PANDN, (V)ANDNPS, (V)ANDNPD

[Jan Bobek]

Use the gvec infrastructure to achieve the desired functionality.

Signed-off-by: Jan Bobek <address@hidden>
---
 target/i386/ops_sse.h        |  4 ----
 target/i386/ops_sse_header.h |  2 --
 target/i386/translate.c      | 26 ++++++++++++++++++++++++--
 3 files changed, 24 insertions(+), 8 deletions(-)

diff --git a/target/i386/ops_sse.h b/target/i386/ops_sse.h
index 68dbeda047..84562a4536 100644
--- a/target/i386/ops_sse.h
+++ b/target/i386/ops_sse.h
@@ -353,8 +353,6 @@ static inline int satsw(int x)
 #define FMAXUB(a, b) ((a) > (b)) ? (a) : (b)
 #define FMAXSW(a, b) ((int16_t)(a) > (int16_t)(b)) ? (a) : (b)
 
-#define FANDN(a, b) ((~(a)) & (b))
-
 #define FCMPGTB(a, b) ((int8_t)(a) > (int8_t)(b) ? -1 : 0)
 #define FCMPGTW(a, b) ((int16_t)(a) > (int16_t)(b) ? -1 : 0)
 #define FCMPGTL(a, b) ((int32_t)(a) > (int32_t)(b) ? -1 : 0)
@@ -394,8 +392,6 @@ SSE_HELPER_B(helper_pmaxub, FMAXUB)
 SSE_HELPER_W(helper_pminsw, FMINSW)
 SSE_HELPER_W(helper_pmaxsw, FMAXSW)
 
-SSE_HELPER_Q(helper_pandn, FANDN)
-
 SSE_HELPER_B(helper_pcmpgtb, FCMPGTB)
 SSE_HELPER_W(helper_pcmpgtw, FCMPGTW)
 SSE_HELPER_L(helper_pcmpgtl, FCMPGTL)
diff --git a/target/i386/ops_sse_header.h b/target/i386/ops_sse_header.h
index a98b9f8f3f..abd00ca69d 100644
--- a/target/i386/ops_sse_header.h
+++ b/target/i386/ops_sse_header.h
@@ -86,8 +86,6 @@ SSE_HELPER_B(pmaxub, FMAXUB)
 SSE_HELPER_W(pminsw, FMINSW)
 SSE_HELPER_W(pmaxsw, FMAXSW)
 
-SSE_HELPER_Q(pandn, FANDN)
-
 SSE_HELPER_B(pcmpgtb, FCMPGTB)
 SSE_HELPER_W(pcmpgtw, FCMPGTW)
 SSE_HELPER_L(pcmpgtl, FCMPGTL)
diff --git a/target/i386/translate.c b/target/i386/translate.c
index cfe285e3e5..69e9514679 100644
--- a/target/i386/translate.c
+++ b/target/i386/translate.c
@@ -2757,7 +2757,7 @@ static const SSEFunc_0_epp sse_op_table1[256][4] = {
     [0x52] = { gen_helper_rsqrtps, NULL, gen_helper_rsqrtss, NULL },
     [0x53] = { gen_helper_rcpps, NULL, gen_helper_rcpss, NULL },
     [0x54] = { SSE_TOMBSTONE, SSE_TOMBSTONE }, /* andps, andpd */
-    [0x55] = { gen_helper_pandn_xmm, gen_helper_pandn_xmm }, /* andnps, andnpd 
*/
+    [0x55] = { SSE_TOMBSTONE, SSE_TOMBSTONE }, /* andnps, andnpd */
     [0x56] = { SSE_TOMBSTONE, SSE_TOMBSTONE }, /* orps, orpd */
     [0x57] = { SSE_TOMBSTONE, SSE_TOMBSTONE }, /* xorps, xorpd */
     [0x58] = SSE_FOP(add),
@@ -2829,7 +2829,7 @@ static const SSEFunc_0_epp sse_op_table1[256][4] = {
     [0xdc] = MMX_OP2(paddusb),
     [0xdd] = MMX_OP2(paddusw),
     [0xde] = MMX_OP2(pmaxub),
-    [0xdf] = MMX_OP2(pandn),
+    [0xdf] = { SSE_TOMBSTONE, SSE_TOMBSTONE },
     [0xe0] = MMX_OP2(pavgb),
     [0xe1] = MMX_OP2(psraw),
     [0xe2] = MMX_OP2(psrad),
@@ -3177,6 +3177,17 @@ static inline void gen_gvec_ld_modrm_3(CPUX86State *env, 
DisasContext *s,
 #define gen_vandpd_xmm gen_vpand_xmm
 #define gen_vandpd_ymm gen_vpand_ymm
 
+#define gen_pandn_mm(env, s, modrm)   gen_gvec_ld_modrm_mm  ((env), (s), 
(modrm), MO_64, tcg_gen_gvec_andc, 0121)
+#define gen_pandn_xmm(env, s, modrm)  gen_gvec_ld_modrm_xmm ((env), (s), 
(modrm), MO_64, tcg_gen_gvec_andc, 0121)
+#define gen_vpandn_xmm(env, s, modrm) gen_gvec_ld_modrm_vxmm((env), (s), 
(modrm), MO_64, tcg_gen_gvec_andc, 0132)
+#define gen_vpandn_ymm(env, s, modrm) gen_gvec_ld_modrm_vymm((env), (s), 
(modrm), MO_64, tcg_gen_gvec_andc, 0132)
+#define gen_andnps_xmm  gen_pandn_xmm
+#define gen_vandnps_xmm gen_vpandn_xmm
+#define gen_vandnps_ymm gen_vpandn_ymm
+#define gen_andnpd_xmm  gen_pandn_xmm
+#define gen_vandnpd_xmm gen_vpandn_xmm
+#define gen_vandnpd_ymm gen_vpandn_ymm
+
 #define gen_por_mm(env, s, modrm)   gen_gvec_ld_modrm_mm  ((env), (s), 
(modrm), MO_64, tcg_gen_gvec_or, 0112)
 #define gen_por_xmm(env, s, modrm)  gen_gvec_ld_modrm_xmm ((env), (s), 
(modrm), MO_64, tcg_gen_gvec_or, 0112)
 #define gen_vpor_xmm(env, s, modrm) gen_gvec_ld_modrm_vxmm((env), (s), 
(modrm), MO_64, tcg_gen_gvec_or, 0123)
@@ -3301,6 +3312,17 @@ static void gen_sse(CPUX86State *env, DisasContext *s, 
int b)
     case 0x54 | M_0F | P_66 | VEX_128: gen_vandpd_xmm(env, s, modrm); return;
     case 0x54 | M_0F | P_66 | VEX_256: gen_vandpd_ymm(env, s, modrm); return;
 
+    case 0xdf | M_0F:                  gen_pandn_mm(env, s, modrm); return;
+    case 0xdf | M_0F | P_66:           gen_pandn_xmm(env, s, modrm); return;
+    case 0xdf | M_0F | P_66 | VEX_128: gen_vpandn_xmm(env, s, modrm); return;
+    case 0xdf | M_0F | P_66 | VEX_256: gen_vpandn_ymm(env, s, modrm); return;
+    case 0x55 | M_0F:                  gen_andnps_xmm(env, s, modrm); return;
+    case 0x55 | M_0F | VEX_128:        gen_vandnps_xmm(env, s, modrm); return;
+    case 0x55 | M_0F | VEX_256:        gen_vandnps_ymm(env, s, modrm); return;
+    case 0x55 | M_0F | P_66:           gen_andnpd_xmm(env, s, modrm); return;
+    case 0x55 | M_0F | P_66 | VEX_128: gen_vandnpd_xmm(env, s, modrm); return;
+    case 0x55 | M_0F | P_66 | VEX_256: gen_vandnpd_ymm(env, s, modrm); return;
+
     case 0xeb | M_0F:                  gen_por_mm(env, s, modrm); return;
     case 0xeb | M_0F | P_66:           gen_por_xmm(env, s, modrm); return;
     case 0xeb | M_0F | P_66 | VEX_128: gen_vpor_xmm(env, s, modrm); return;
-- 
2.20.1