This patch includes:
- XVMADD.{B/H/W/D};
- XVMSUB.{B/H/W/D};
- XVMADDW{EV/OD}.{H.B/W.H/D.W/Q.D}[U];
- XVMADDW{EV/OD}.{H.BU.B/W.HU.H/D.WU.W/Q.DU.D}.
Signed-off-by: Song Gao <gaos...@loongson.cn>
---
target/loongarch/disas.c | 34 ++++++++++
target/loongarch/insn_trans/trans_lasx.c.inc | 69 ++++++++++++++++++++
target/loongarch/insn_trans/trans_lsx.c.inc | 2 +
target/loongarch/insns.decode | 34 ++++++++++
target/loongarch/vec.h | 3 +
target/loongarch/vec_helper.c | 33 +++++-----
6 files changed, 160 insertions(+), 15 deletions(-)
diff --git a/target/loongarch/disas.c b/target/loongarch/disas.c
index e5f9a6bcdf..b115fe8315 100644
--- a/target/loongarch/disas.c
+++ b/target/loongarch/disas.c
@@ -1928,6 +1928,40 @@ INSN_LASX(xvmulwod_w_hu_h, vvv)
INSN_LASX(xvmulwod_d_wu_w, vvv)
INSN_LASX(xvmulwod_q_du_d, vvv)
+INSN_LASX(xvmadd_b, vvv)
+INSN_LASX(xvmadd_h, vvv)
+INSN_LASX(xvmadd_w, vvv)
+INSN_LASX(xvmadd_d, vvv)
+INSN_LASX(xvmsub_b, vvv)
+INSN_LASX(xvmsub_h, vvv)
+INSN_LASX(xvmsub_w, vvv)
+INSN_LASX(xvmsub_d, vvv)
+
+INSN_LASX(xvmaddwev_h_b, vvv)
+INSN_LASX(xvmaddwev_w_h, vvv)
+INSN_LASX(xvmaddwev_d_w, vvv)
+INSN_LASX(xvmaddwev_q_d, vvv)
+INSN_LASX(xvmaddwod_h_b, vvv)
+INSN_LASX(xvmaddwod_w_h, vvv)
+INSN_LASX(xvmaddwod_d_w, vvv)
+INSN_LASX(xvmaddwod_q_d, vvv)
+INSN_LASX(xvmaddwev_h_bu, vvv)
+INSN_LASX(xvmaddwev_w_hu, vvv)
+INSN_LASX(xvmaddwev_d_wu, vvv)
+INSN_LASX(xvmaddwev_q_du, vvv)
+INSN_LASX(xvmaddwod_h_bu, vvv)
+INSN_LASX(xvmaddwod_w_hu, vvv)
+INSN_LASX(xvmaddwod_d_wu, vvv)
+INSN_LASX(xvmaddwod_q_du, vvv)
+INSN_LASX(xvmaddwev_h_bu_b, vvv)
+INSN_LASX(xvmaddwev_w_hu_h, vvv)
+INSN_LASX(xvmaddwev_d_wu_w, vvv)
+INSN_LASX(xvmaddwev_q_du_d, vvv)
+INSN_LASX(xvmaddwod_h_bu_b, vvv)
+INSN_LASX(xvmaddwod_w_hu_h, vvv)
+INSN_LASX(xvmaddwod_d_wu_w, vvv)
+INSN_LASX(xvmaddwod_q_du_d, vvv)
+
INSN_LASX(xvreplgr2vr_b, vr)
INSN_LASX(xvreplgr2vr_h, vr)
INSN_LASX(xvreplgr2vr_w, vr)
diff --git a/target/loongarch/insn_trans/trans_lasx.c.inc
b/target/loongarch/insn_trans/trans_lasx.c.inc
index 1b07d3ce3a..2c2fae91b9 100644
--- a/target/loongarch/insn_trans/trans_lasx.c.inc
+++ b/target/loongarch/insn_trans/trans_lasx.c.inc
@@ -275,6 +275,75 @@ TRANS(xvmulwod_h_bu_b, gvec_vvv, 32, MO_8, do_vmulwod_u_s)
TRANS(xvmulwod_w_hu_h, gvec_vvv, 32, MO_16, do_vmulwod_u_s)
TRANS(xvmulwod_d_wu_w, gvec_vvv, 32, MO_32, do_vmulwod_u_s)
+TRANS(xvmadd_b, gvec_vvv, 32, MO_8, do_vmadd)
+TRANS(xvmadd_h, gvec_vvv, 32, MO_16, do_vmadd)
+TRANS(xvmadd_w, gvec_vvv, 32, MO_32, do_vmadd)
+TRANS(xvmadd_d, gvec_vvv, 32, MO_64, do_vmadd)
+TRANS(xvmsub_b, gvec_vvv, 32, MO_8, do_vmsub)
+TRANS(xvmsub_h, gvec_vvv, 32, MO_16, do_vmsub)
+TRANS(xvmsub_w, gvec_vvv, 32, MO_32, do_vmsub)
+TRANS(xvmsub_d, gvec_vvv, 32, MO_64, do_vmsub)
+
+TRANS(xvmaddwev_h_b, gvec_vvv, 32, MO_8, do_vmaddwev_s)
+TRANS(xvmaddwev_w_h, gvec_vvv, 32, MO_16, do_vmaddwev_s)
+TRANS(xvmaddwev_d_w, gvec_vvv, 32, MO_32, do_vmaddwev_s)
+
+#define XVMADD_Q(NAME, FN, idx1, idx2) \
+static bool trans_## NAME(DisasContext *ctx, arg_vvv * a) \
+{ \
+ TCGv_i64 rh, rl, arg1, arg2, th, tl; \
+ int i; \
+ \
+ CHECK_VEC; \
+ \
+ rh = tcg_temp_new_i64(); \
+ rl = tcg_temp_new_i64(); \
+ arg1 = tcg_temp_new_i64(); \
+ arg2 = tcg_temp_new_i64(); \
+ th = tcg_temp_new_i64(); \
+ tl = tcg_temp_new_i64(); \
+ \
+ for (i = 0; i < 2; i++) { \
+ get_vreg64(arg1, a->vj, idx1 + i * 2); \
+ get_vreg64(arg2, a->vk, idx2 + i * 2); \
+ get_vreg64(rh, a->vd, 1 + i * 2); \
+ get_vreg64(rl, a->vd, 0 + i * 2); \
+ \
+ tcg_gen_## FN ##_i64(tl, th, arg1, arg2); \
+ tcg_gen_add2_i64(rl, rh, rl, rh, tl, th); \
+ \
+ set_vreg64(rh, a->vd, 1 + i * 2); \
+ set_vreg64(rl, a->vd, 0 + i * 2); \
+ } \
+ \
+ return true; \
+}
+
+XVMADD_Q(xvmaddwev_q_d, muls2, 0, 0)
+XVMADD_Q(xvmaddwod_q_d, muls2, 1, 1)
+XVMADD_Q(xvmaddwev_q_du, mulu2, 0, 0)
+XVMADD_Q(xvmaddwod_q_du, mulu2, 1, 1)
+XVMADD_Q(xvmaddwev_q_du_d, mulus2, 0, 0)
+XVMADD_Q(xvmaddwod_q_du_d, mulus2, 1, 1)
+
+TRANS(xvmaddwod_h_b, gvec_vvv, 32, MO_8, do_vmaddwod_s)
+TRANS(xvmaddwod_w_h, gvec_vvv, 32, MO_16, do_vmaddwod_s)
+TRANS(xvmaddwod_d_w, gvec_vvv, 32, MO_32, do_vmaddwod_s)
+
+TRANS(xvmaddwev_h_bu, gvec_vvv, 32, MO_8, do_vmaddwev_u)
+TRANS(xvmaddwev_w_hu, gvec_vvv, 32, MO_16, do_vmaddwev_u)
+TRANS(xvmaddwev_d_wu, gvec_vvv, 32, MO_32, do_vmaddwev_u)
+TRANS(xvmaddwod_h_bu, gvec_vvv, 32, MO_8, do_vmaddwod_u)
+TRANS(xvmaddwod_w_hu, gvec_vvv, 32, MO_16, do_vmaddwod_u)
+TRANS(xvmaddwod_d_wu, gvec_vvv, 32, MO_32, do_vmaddwod_u)
+
+TRANS(xvmaddwev_h_bu_b, gvec_vvv, 32, MO_8, do_vmaddwev_u_s)
+TRANS(xvmaddwev_w_hu_h, gvec_vvv, 32, MO_16, do_vmaddwev_u_s)
+TRANS(xvmaddwev_d_wu_w, gvec_vvv, 32, MO_32, do_vmaddwev_u_s)
+TRANS(xvmaddwod_h_bu_b, gvec_vvv, 32, MO_8, do_vmaddwod_u_s)
+TRANS(xvmaddwod_w_hu_h, gvec_vvv, 32, MO_16, do_vmaddwod_u_s)
+TRANS(xvmaddwod_d_wu_w, gvec_vvv, 32, MO_32, do_vmaddwod_u_s)
+
TRANS(xvreplgr2vr_b, gvec_dup, 32, MO_8)
TRANS(xvreplgr2vr_h, gvec_dup, 32, MO_16)
TRANS(xvreplgr2vr_w, gvec_dup, 32, MO_32)
diff --git a/target/loongarch/insn_trans/trans_lsx.c.inc
b/target/loongarch/insn_trans/trans_lsx.c.inc
index cc97866ef9..f2f7c7a9aa 100644
--- a/target/loongarch/insn_trans/trans_lsx.c.inc
+++ b/target/loongarch/insn_trans/trans_lsx.c.inc
@@ -2333,6 +2333,8 @@ static bool trans_## NAME (DisasContext *ctx, arg_vvv *a)
\
{ \
TCGv_i64 rh, rl, arg1, arg2, th, tl; \
\
+ CHECK_VEC; \
+ \
rh = tcg_temp_new_i64(); \
rl = tcg_temp_new_i64(); \
arg1 = tcg_temp_new_i64(); \
diff --git a/target/loongarch/insns.decode b/target/loongarch/insns.decode
index 0f9ebe641f..d6fb51ae64 100644
--- a/target/loongarch/insns.decode
+++ b/target/loongarch/insns.decode
@@ -1511,6 +1511,40 @@ xvmulwod_w_hu_h 0111 01001010 00101 ..... ..... .....
@vvv
xvmulwod_d_wu_w 0111 01001010 00110 ..... ..... ..... @vvv
xvmulwod_q_du_d 0111 01001010 00111 ..... ..... ..... @vvv
+xvmadd_b 0111 01001010 10000 ..... ..... ..... @vvv
+xvmadd_h 0111 01001010 10001 ..... ..... ..... @vvv
+xvmadd_w 0111 01001010 10010 ..... ..... ..... @vvv
+xvmadd_d 0111 01001010 10011 ..... ..... ..... @vvv
+xvmsub_b 0111 01001010 10100 ..... ..... ..... @vvv
+xvmsub_h 0111 01001010 10101 ..... ..... ..... @vvv
+xvmsub_w 0111 01001010 10110 ..... ..... ..... @vvv
+xvmsub_d 0111 01001010 10111 ..... ..... ..... @vvv
+
+xvmaddwev_h_b 0111 01001010 11000 ..... ..... ..... @vvv
+xvmaddwev_w_h 0111 01001010 11001 ..... ..... ..... @vvv
+xvmaddwev_d_w 0111 01001010 11010 ..... ..... ..... @vvv
+xvmaddwev_q_d 0111 01001010 11011 ..... ..... ..... @vvv
+xvmaddwod_h_b 0111 01001010 11100 ..... ..... ..... @vvv
+xvmaddwod_w_h 0111 01001010 11101 ..... ..... ..... @vvv
+xvmaddwod_d_w 0111 01001010 11110 ..... ..... ..... @vvv
+xvmaddwod_q_d 0111 01001010 11111 ..... ..... ..... @vvv
+xvmaddwev_h_bu 0111 01001011 01000 ..... ..... ..... @vvv
+xvmaddwev_w_hu 0111 01001011 01001 ..... ..... ..... @vvv
+xvmaddwev_d_wu 0111 01001011 01010 ..... ..... ..... @vvv
+xvmaddwev_q_du 0111 01001011 01011 ..... ..... ..... @vvv
+xvmaddwod_h_bu 0111 01001011 01100 ..... ..... ..... @vvv
+xvmaddwod_w_hu 0111 01001011 01101 ..... ..... ..... @vvv
+xvmaddwod_d_wu 0111 01001011 01110 ..... ..... ..... @vvv
+xvmaddwod_q_du 0111 01001011 01111 ..... ..... ..... @vvv
+xvmaddwev_h_bu_b 0111 01001011 11000 ..... ..... ..... @vvv
+xvmaddwev_w_hu_h 0111 01001011 11001 ..... ..... ..... @vvv
+xvmaddwev_d_wu_w 0111 01001011 11010 ..... ..... ..... @vvv
+xvmaddwev_q_du_d 0111 01001011 11011 ..... ..... ..... @vvv
+xvmaddwod_h_bu_b 0111 01001011 11100 ..... ..... ..... @vvv
+xvmaddwod_w_hu_h 0111 01001011 11101 ..... ..... ..... @vvv
+xvmaddwod_d_wu_w 0111 01001011 11110 ..... ..... ..... @vvv
+xvmaddwod_q_du_d 0111 01001011 11111 ..... ..... ..... @vvv
+
xvreplgr2vr_b 0111 01101001 11110 00000 ..... ..... @vr
xvreplgr2vr_h 0111 01101001 11110 00001 ..... ..... @vr
xvreplgr2vr_w 0111 01101001 11110 00010 ..... ..... @vr
diff --git a/target/loongarch/vec.h b/target/loongarch/vec.h
index c371a59a2e..1abc6a3da0 100644
--- a/target/loongarch/vec.h
+++ b/target/loongarch/vec.h
@@ -62,4 +62,7 @@
#define DO_MUL(a, b) (a * b)
+#define DO_MADD(a, b, c) (a + b * c)
+#define DO_MSUB(a, b, c) (a - b * c)
+
#endif /* LOONGARCH_VEC_H */
diff --git a/target/loongarch/vec_helper.c b/target/loongarch/vec_helper.c
index 804fbc6969..367b794853 100644
--- a/target/loongarch/vec_helper.c
+++ b/target/loongarch/vec_helper.c
@@ -508,17 +508,16 @@ DO_ODD_U_S(vmulwod_h_bu_b, 16, H, UH, B, UB, DO_MUL)
DO_ODD_U_S(vmulwod_w_hu_h, 32, W, UW, H, UH, DO_MUL)
DO_ODD_U_S(vmulwod_d_wu_w, 64, D, UD, W, UW, DO_MUL)
-#define DO_MADD(a, b, c) (a + b * c)
-#define DO_MSUB(a, b, c) (a - b * c)
-
#define VMADDSUB(NAME, BIT, E, DO_OP) \
void HELPER(NAME)(void *vd, void *vj, void *vk, uint32_t v) \
{ \
- int i; \
+ int i, len; \
VReg *Vd = (VReg *)vd; \
VReg *Vj = (VReg *)vj; \
VReg *Vk = (VReg *)vk; \
- for (i = 0; i < LSX_LEN/BIT; i++) { \
+ \
+ len = (simd_oprsz(v) == 16) ? LSX_LEN : LASX_LEN; \
+ for (i = 0; i < len / BIT; i++) { \
Vd->E(i) = DO_OP(Vd->E(i), Vj->E(i) ,Vk->E(i)); \
} \
}
@@ -535,13 +534,14 @@ VMADDSUB(vmsub_d, 64, D, DO_MSUB)
#define VMADDWEV(NAME, BIT, E1, E2, DO_OP) \
void HELPER(NAME)(void *vd, void *vj, void *vk, uint32_t v) \
{ \
- int i; \
+ int i, len; \
VReg *Vd = (VReg *)vd; \
VReg *Vj = (VReg *)vj; \
VReg *Vk = (VReg *)vk; \
typedef __typeof(Vd->E1(0)) TD; \
\
- for (i = 0; i < LSX_LEN/BIT; i++) { \
+ len = (simd_oprsz(v) == 16) ? LSX_LEN : LASX_LEN; \
+ for (i = 0; i < len / BIT; i++) { \
Vd->E1(i) += DO_OP((TD)Vj->E2(2 * i), (TD)Vk->E2(2 * i)); \
} \
}
@@ -556,13 +556,14 @@ VMADDWEV(vmaddwev_d_wu, 64, UD, UW, DO_MUL)
#define VMADDWOD(NAME, BIT, E1, E2, DO_OP) \
void HELPER(NAME)(void *vd, void *vj, void *vk, uint32_t v) \
{ \
- int i; \
+ int i, len; \
VReg *Vd = (VReg *)vd; \
VReg *Vj = (VReg *)vj; \
VReg *Vk = (VReg *)vk; \
typedef __typeof(Vd->E1(0)) TD; \
\
- for (i = 0; i < LSX_LEN/BIT; i++) { \
+ len = (simd_oprsz(v) == 16) ? LSX_LEN : LASX_LEN; \
+ for (i = 0; i < len / BIT; i++) { \
Vd->E1(i) += DO_OP((TD)Vj->E2(2 * i + 1), \
(TD)Vk->E2(2 * i + 1)); \
} \
@@ -578,14 +579,15 @@ VMADDWOD(vmaddwod_d_wu, 64, UD, UW, DO_MUL)
#define VMADDWEV_U_S(NAME, BIT, ES1, EU1, ES2, EU2, DO_OP) \
void HELPER(NAME)(void *vd, void *vj, void *vk, uint32_t v) \
{ \
- int i; \
+ int i, len; \
VReg *Vd = (VReg *)vd; \
VReg *Vj = (VReg *)vj; \
VReg *Vk = (VReg *)vk; \
typedef __typeof(Vd->ES1(0)) TS1; \
typedef __typeof(Vd->EU1(0)) TU1; \
\
- for (i = 0; i < LSX_LEN/BIT; i++) { \
+ len = (simd_oprsz(v) == 16) ? LSX_LEN : LASX_LEN; \
+ for (i = 0; i < len / BIT; i++) { \
Vd->ES1(i) += DO_OP((TU1)Vj->EU2(2 * i), \
(TS1)Vk->ES2(2 * i)); \
} \
@@ -598,16 +600,17 @@ VMADDWEV_U_S(vmaddwev_d_wu_w, 64, D, UD, W, UW, DO_MUL)
#define VMADDWOD_U_S(NAME, BIT, ES1, EU1, ES2, EU2, DO_OP) \
void HELPER(NAME)(void *vd, void *vj, void *vk, uint32_t v) \
{ \
- int i; \
+ int i, len; \
VReg *Vd = (VReg *)vd; \
VReg *Vj = (VReg *)vj; \
VReg *Vk = (VReg *)vk; \
typedef __typeof(Vd->ES1(0)) TS1; \
typedef __typeof(Vd->EU1(0)) TU1; \
\
- for (i = 0; i < LSX_LEN/BIT; i++) { \
- Vd->ES1(i) += DO_OP((TU1)Vj->EU2(2 * i + 1), \
- (TS1)Vk->ES2(2 * i + 1)); \
+ len = (simd_oprsz(v) == 16) ? LSX_LEN : LASX_LEN; \
+ for (i = 0; i < len / BIT; i++) { \
+ Vd->ES1(i) += DO_OP((TU1)Vj->EU2(2 * i + 1), \
+ (TS1)Vk->ES2(2 * i + 1)); \
} \
}
--
2.39.1
