[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[PATCH v4 34/48] target/loongarch: Implement xvclo xvclz
From: |
Song Gao |
Subject: |
[PATCH v4 34/48] target/loongarch: Implement xvclo xvclz |
Date: |
Wed, 30 Aug 2023 16:48:48 +0800 |
This patch includes:
- XVCLO.{B/H/W/D};
- XVCLZ.{B/H/W/D}.
Signed-off-by: Song Gao <gaosong@loongson.cn>
---
target/loongarch/vec.h | 9 +++++++++
target/loongarch/insns.decode | 9 +++++++++
target/loongarch/disas.c | 9 +++++++++
target/loongarch/vec_helper.c | 13 ++-----------
target/loongarch/insn_trans/trans_lasx.c.inc | 9 +++++++++
5 files changed, 38 insertions(+), 11 deletions(-)
diff --git a/target/loongarch/vec.h b/target/loongarch/vec.h
index 67d829f9da..4497cd4a6d 100644
--- a/target/loongarch/vec.h
+++ b/target/loongarch/vec.h
@@ -76,4 +76,13 @@
#define R_SHIFT(a, b) (a >> b)
+#define DO_CLO_B(N) (clz32(~N & 0xff) - 24)
+#define DO_CLO_H(N) (clz32(~N & 0xffff) - 16)
+#define DO_CLO_W(N) (clz32(~N))
+#define DO_CLO_D(N) (clz64(~N))
+#define DO_CLZ_B(N) (clz32(N) - 24)
+#define DO_CLZ_H(N) (clz32(N) - 16)
+#define DO_CLZ_W(N) (clz32(N))
+#define DO_CLZ_D(N) (clz64(N))
+
#endif /* LOONGARCH_VEC_H */
diff --git a/target/loongarch/insns.decode b/target/loongarch/insns.decode
index dc74bae7a5..3175532045 100644
--- a/target/loongarch/insns.decode
+++ b/target/loongarch/insns.decode
@@ -1770,6 +1770,15 @@ xvssrarni_hu_w 0111 01110110 11001 ..... ..... .....
@vv_ui5
xvssrarni_wu_d 0111 01110110 1101 ...... ..... ..... @vv_ui6
xvssrarni_du_q 0111 01110110 111 ....... ..... ..... @vv_ui7
+xvclo_b 0111 01101001 11000 00000 ..... ..... @vv
+xvclo_h 0111 01101001 11000 00001 ..... ..... @vv
+xvclo_w 0111 01101001 11000 00010 ..... ..... @vv
+xvclo_d 0111 01101001 11000 00011 ..... ..... @vv
+xvclz_b 0111 01101001 11000 00100 ..... ..... @vv
+xvclz_h 0111 01101001 11000 00101 ..... ..... @vv
+xvclz_w 0111 01101001 11000 00110 ..... ..... @vv
+xvclz_d 0111 01101001 11000 00111 ..... ..... @vv
+
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/disas.c b/target/loongarch/disas.c
index f043a2f9b6..0fc58735b9 100644
--- a/target/loongarch/disas.c
+++ b/target/loongarch/disas.c
@@ -2196,6 +2196,15 @@ INSN_LASX(xvssrarni_hu_w, vv_i)
INSN_LASX(xvssrarni_wu_d, vv_i)
INSN_LASX(xvssrarni_du_q, vv_i)
+INSN_LASX(xvclo_b, vv)
+INSN_LASX(xvclo_h, vv)
+INSN_LASX(xvclo_w, vv)
+INSN_LASX(xvclo_d, vv)
+INSN_LASX(xvclz_b, vv)
+INSN_LASX(xvclz_h, vv)
+INSN_LASX(xvclz_w, vv)
+INSN_LASX(xvclz_d, vv)
+
INSN_LASX(xvreplgr2vr_b, vr)
INSN_LASX(xvreplgr2vr_h, vr)
INSN_LASX(xvreplgr2vr_w, vr)
diff --git a/target/loongarch/vec_helper.c b/target/loongarch/vec_helper.c
index 852c65716e..789f6b303e 100644
--- a/target/loongarch/vec_helper.c
+++ b/target/loongarch/vec_helper.c
@@ -2161,22 +2161,13 @@ void HELPER(NAME)(void *vd, void *vj, uint32_t desc) \
int i; \
VReg *Vd = (VReg *)vd; \
VReg *Vj = (VReg *)vj; \
+ int oprsz = simd_oprsz(desc); \
\
- for (i = 0; i < LSX_LEN/BIT; i++) \
- { \
+ for (i = 0; i < oprsz / (BIT / 8); i++) { \
Vd->E(i) = DO_OP(Vj->E(i)); \
} \
}
-#define DO_CLO_B(N) (clz32(~N & 0xff) - 24)
-#define DO_CLO_H(N) (clz32(~N & 0xffff) - 16)
-#define DO_CLO_W(N) (clz32(~N))
-#define DO_CLO_D(N) (clz64(~N))
-#define DO_CLZ_B(N) (clz32(N) - 24)
-#define DO_CLZ_H(N) (clz32(N) - 16)
-#define DO_CLZ_W(N) (clz32(N))
-#define DO_CLZ_D(N) (clz64(N))
-
DO_2OP(vclo_b, 8, UB, DO_CLO_B)
DO_2OP(vclo_h, 16, UH, DO_CLO_H)
DO_2OP(vclo_w, 32, UW, DO_CLO_W)
diff --git a/target/loongarch/insn_trans/trans_lasx.c.inc
b/target/loongarch/insn_trans/trans_lasx.c.inc
index dc658fc2cb..4227fbe629 100644
--- a/target/loongarch/insn_trans/trans_lasx.c.inc
+++ b/target/loongarch/insn_trans/trans_lasx.c.inc
@@ -519,6 +519,15 @@ TRANS(xvssrarni_hu_w, LASX, gen_vv_i, 32,
gen_helper_vssrarni_hu_w)
TRANS(xvssrarni_wu_d, LASX, gen_vv_i, 32, gen_helper_vssrarni_wu_d)
TRANS(xvssrarni_du_q, LASX, gen_vv_i, 32, gen_helper_vssrarni_du_q)
+TRANS(xvclo_b, LASX, gen_vv, 32, gen_helper_vclo_b)
+TRANS(xvclo_h, LASX, gen_vv, 32, gen_helper_vclo_h)
+TRANS(xvclo_w, LASX, gen_vv, 32, gen_helper_vclo_w)
+TRANS(xvclo_d, LASX, gen_vv, 32, gen_helper_vclo_d)
+TRANS(xvclz_b, LASX, gen_vv, 32, gen_helper_vclz_b)
+TRANS(xvclz_h, LASX, gen_vv, 32, gen_helper_vclz_h)
+TRANS(xvclz_w, LASX, gen_vv, 32, gen_helper_vclz_w)
+TRANS(xvclz_d, LASX, gen_vv, 32, gen_helper_vclz_d)
+
TRANS(xvreplgr2vr_b, LASX, gvec_dup, 32, MO_8)
TRANS(xvreplgr2vr_h, LASX, gvec_dup, 32, MO_16)
TRANS(xvreplgr2vr_w, LASX, gvec_dup, 32, MO_32)
--
2.39.1
- [PATCH v4 31/48] target/loongarch: Implement xvsrlrn xvsrarn, (continued)
- [PATCH v4 31/48] target/loongarch: Implement xvsrlrn xvsrarn, Song Gao, 2023/08/30
- [PATCH v4 30/48] target/loongarch: Implement xvsrln xvsran, Song Gao, 2023/08/30
- [PATCH v4 36/48] target/loongarch: Implement xvbitclr xvbitset xvbitrev, Song Gao, 2023/08/30
- [PATCH v4 35/48] target/loongarch: Implement xvpcnt, Song Gao, 2023/08/30
- [PATCH v4 33/48] target/loongarch: Implement xvssrlrn xvssrarn, Song Gao, 2023/08/30
- [PATCH v4 34/48] target/loongarch: Implement xvclo xvclz,
Song Gao <=
- [PATCH v4 38/48] target/loongarch: Implement LASX fpu arith instructions, Song Gao, 2023/08/30
- [PATCH v4 42/48] target/loongarch: Implement xvbitsel xvset, Song Gao, 2023/08/30
- [PATCH v4 39/48] target/loongarch: Implement LASX fpu fcvt instructions, Song Gao, 2023/08/30
- [PATCH v4 37/48] target/loongarch: Implement xvfrstp, Song Gao, 2023/08/30
- [PATCH v4 40/48] target/loongarch: Implement xvseq xvsle xvslt, Song Gao, 2023/08/30