[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[PATCH v5 10/49] target/ppc: Move Vector Compare Not Equal or Zero to de
|
From: |
matheus . ferst |
|
Subject: |
[PATCH v5 10/49] target/ppc: Move Vector Compare Not Equal or Zero to decodetree |
|
Date: |
Fri, 25 Feb 2022 18:08:57 -0300 |
From: Matheus Ferst <matheus.ferst@eldorado.org.br>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Matheus Ferst <matheus.ferst@eldorado.org.br>
---
target/ppc/helper.h | 9 ++--
target/ppc/insn32.decode | 4 ++
target/ppc/int_helper.c | 50 +++++-----------------
target/ppc/translate/vmx-impl.c.inc | 66 +++++++++++++++++++++++++++--
target/ppc/translate/vmx-ops.c.inc | 3 --
5 files changed, 80 insertions(+), 52 deletions(-)
diff --git a/target/ppc/helper.h b/target/ppc/helper.h
index 67f78b801b..3257203791 100644
--- a/target/ppc/helper.h
+++ b/target/ppc/helper.h
@@ -140,16 +140,13 @@ DEF_HELPER_3(vabsduw, void, avr, avr, avr)
DEF_HELPER_3(vavgsb, void, avr, avr, avr)
DEF_HELPER_3(vavgsh, void, avr, avr, avr)
DEF_HELPER_3(vavgsw, void, avr, avr, avr)
-DEF_HELPER_4(vcmpnezb, void, env, avr, avr, avr)
-DEF_HELPER_4(vcmpnezh, void, env, avr, avr, avr)
-DEF_HELPER_4(vcmpnezw, void, env, avr, avr, avr)
DEF_HELPER_4(vcmpeqfp, void, env, avr, avr, avr)
DEF_HELPER_4(vcmpgefp, void, env, avr, avr, avr)
DEF_HELPER_4(vcmpgtfp, void, env, avr, avr, avr)
DEF_HELPER_4(vcmpbfp, void, env, avr, avr, avr)
-DEF_HELPER_4(vcmpnezb_dot, void, env, avr, avr, avr)
-DEF_HELPER_4(vcmpnezh_dot, void, env, avr, avr, avr)
-DEF_HELPER_4(vcmpnezw_dot, void, env, avr, avr, avr)
+DEF_HELPER_FLAGS_4(VCMPNEZB, TCG_CALL_NO_RWG, void, avr, avr, avr, i32)
+DEF_HELPER_FLAGS_4(VCMPNEZH, TCG_CALL_NO_RWG, void, avr, avr, avr, i32)
+DEF_HELPER_FLAGS_4(VCMPNEZW, TCG_CALL_NO_RWG, void, avr, avr, avr, i32)
DEF_HELPER_4(vcmpeqfp_dot, void, env, avr, avr, avr)
DEF_HELPER_4(vcmpgefp_dot, void, env, avr, avr, avr)
DEF_HELPER_4(vcmpgtfp_dot, void, env, avr, avr, avr)
diff --git a/target/ppc/insn32.decode b/target/ppc/insn32.decode
index 5443ee0394..be9e05cc73 100644
--- a/target/ppc/insn32.decode
+++ b/target/ppc/insn32.decode
@@ -397,6 +397,10 @@ VCMPNEB 000100 ..... ..... ..... . 0000000111 @VC
VCMPNEH 000100 ..... ..... ..... . 0001000111 @VC
VCMPNEW 000100 ..... ..... ..... . 0010000111 @VC
+VCMPNEZB 000100 ..... ..... ..... . 0100000111 @VC
+VCMPNEZH 000100 ..... ..... ..... . 0101000111 @VC
+VCMPNEZW 000100 ..... ..... ..... . 0110000111 @VC
+
## Vector Bit Manipulation Instruction
VCFUGED 000100 ..... ..... ..... 10101001101 @VX
diff --git a/target/ppc/int_helper.c b/target/ppc/int_helper.c
index 734b817b68..f31dba9469 100644
--- a/target/ppc/int_helper.c
+++ b/target/ppc/int_helper.c
@@ -662,46 +662,18 @@ VCF(ux, uint32_to_float32, u32)
VCF(sx, int32_to_float32, s32)
#undef VCF
-#define VCMPNE_DO(suffix, element, etype, cmpzero, record) \
-void helper_vcmpne##suffix(CPUPPCState *env, ppc_avr_t *r, \
- ppc_avr_t *a, ppc_avr_t *b) \
-{ \
- etype ones = (etype)-1; \
- etype all = ones; \
- etype result, none = 0; \
- int i; \
- \
- for (i = 0; i < ARRAY_SIZE(r->element); i++) { \
- if (cmpzero) { \
- result = ((a->element[i] == 0) \
- || (b->element[i] == 0) \
- || (a->element[i] != b->element[i]) ? \
- ones : 0x0); \
- } else { \
- result = (a->element[i] != b->element[i]) ? ones : 0x0; \
- } \
- r->element[i] = result; \
- all &= result; \
- none |= result; \
- } \
- if (record) { \
- env->crf[6] = ((all != 0) << 3) | ((none == 0) << 1); \
- } \
+#define VCMPNEZ(NAME, ELEM) \
+void helper_##NAME(ppc_vsr_t *t, ppc_vsr_t *a, ppc_vsr_t *b, uint32_t desc) \
+{ \
+ for (int i = 0; i < ARRAY_SIZE(t->ELEM); i++) { \
+ t->ELEM[i] = ((a->ELEM[i] == 0) || (b->ELEM[i] == 0) || \
+ (a->ELEM[i] != b->ELEM[i])) ? -1 : 0; \
+ } \
}
-
-/*
- * VCMPNEZ - Vector compare not equal to zero
- * suffix - instruction mnemonic suffix (b: byte, h: halfword, w: word)
- * element - element type to access from vector
- */
-#define VCMPNE(suffix, element, etype, cmpzero) \
- VCMPNE_DO(suffix, element, etype, cmpzero, 0) \
- VCMPNE_DO(suffix##_dot, element, etype, cmpzero, 1)
-VCMPNE(zb, u8, uint8_t, 1)
-VCMPNE(zh, u16, uint16_t, 1)
-VCMPNE(zw, u32, uint32_t, 1)
-#undef VCMPNE_DO
-#undef VCMPNE
+VCMPNEZ(VCMPNEZB, u8)
+VCMPNEZ(VCMPNEZH, u16)
+VCMPNEZ(VCMPNEZW, u32)
+#undef VCMPNEZ
#define VCMPFP_DO(suffix, compare, order, record) \
void helper_vcmp##suffix(CPUPPCState *env, ppc_avr_t *r, \
diff --git a/target/ppc/translate/vmx-impl.c.inc
b/target/ppc/translate/vmx-impl.c.inc
index 7593b93eab..0574bb8bab 100644
--- a/target/ppc/translate/vmx-impl.c.inc
+++ b/target/ppc/translate/vmx-impl.c.inc
@@ -985,10 +985,6 @@ static void glue(gen_, name0##_##name1)(DisasContext *ctx)
\
} \
}
-GEN_VXRFORM(vcmpnezb, 3, 4)
-GEN_VXRFORM(vcmpnezh, 3, 5)
-GEN_VXRFORM(vcmpnezw, 3, 6)
-
static void do_vcmp_rc(int vrt)
{
TCGv_i64 tmp, set, clr;
@@ -1049,6 +1045,68 @@ TRANS_FLAGS2(ISA300, VCMPNEB, do_vcmp, TCG_COND_NE, MO_8)
TRANS_FLAGS2(ISA300, VCMPNEH, do_vcmp, TCG_COND_NE, MO_16)
TRANS_FLAGS2(ISA300, VCMPNEW, do_vcmp, TCG_COND_NE, MO_32)
+static void gen_vcmpnez_vec(unsigned vece, TCGv_vec t, TCGv_vec a, TCGv_vec b)
+{
+ TCGv_vec t0, t1, zero;
+
+ t0 = tcg_temp_new_vec_matching(t);
+ t1 = tcg_temp_new_vec_matching(t);
+ zero = tcg_constant_vec_matching(t, vece, 0);
+
+ tcg_gen_cmp_vec(TCG_COND_EQ, vece, t0, a, zero);
+ tcg_gen_cmp_vec(TCG_COND_EQ, vece, t1, b, zero);
+ tcg_gen_cmp_vec(TCG_COND_NE, vece, t, a, b);
+
+ tcg_gen_or_vec(vece, t, t, t0);
+ tcg_gen_or_vec(vece, t, t, t1);
+
+ tcg_temp_free_vec(t0);
+ tcg_temp_free_vec(t1);
+}
+
+static bool do_vcmpnez(DisasContext *ctx, arg_VC *a, int vece)
+{
+ static const TCGOpcode vecop_list[] = {
+ INDEX_op_cmp_vec, 0
+ };
+ static const GVecGen3 ops[3] = {
+ {
+ .fniv = gen_vcmpnez_vec,
+ .fno = gen_helper_VCMPNEZB,
+ .opt_opc = vecop_list,
+ .vece = MO_8
+ },
+ {
+ .fniv = gen_vcmpnez_vec,
+ .fno = gen_helper_VCMPNEZH,
+ .opt_opc = vecop_list,
+ .vece = MO_16
+ },
+ {
+ .fniv = gen_vcmpnez_vec,
+ .fno = gen_helper_VCMPNEZW,
+ .opt_opc = vecop_list,
+ .vece = MO_32
+ }
+ };
+
+ REQUIRE_INSNS_FLAGS2(ctx, ISA300);
+ REQUIRE_VECTOR(ctx);
+
+ tcg_gen_gvec_3(avr_full_offset(a->vrt), avr_full_offset(a->vra),
+ avr_full_offset(a->vrb), 16, 16, &ops[vece]);
+
+ if (a->rc) {
+ do_vcmp_rc(a->vrt);
+ }
+
+ return true;
+}
+
+TRANS(VCMPNEZB, do_vcmpnez, MO_8)
+TRANS(VCMPNEZH, do_vcmpnez, MO_16)
+TRANS(VCMPNEZW, do_vcmpnez, MO_32)
+
GEN_VXRFORM(vcmpeqfp, 3, 3)
GEN_VXRFORM(vcmpgefp, 3, 7)
GEN_VXRFORM(vcmpgtfp, 3, 11)
diff --git a/target/ppc/translate/vmx-ops.c.inc
b/target/ppc/translate/vmx-ops.c.inc
index 80d460c34e..cb4c5bb953 100644
--- a/target/ppc/translate/vmx-ops.c.inc
+++ b/target/ppc/translate/vmx-ops.c.inc
@@ -184,9 +184,6 @@ GEN_HANDLER2_E(name, str, 0x4, opc2, opc3, 0x00000000,
PPC_NONE, PPC2_ISA300),
GEN_VXRFORM1_300(name, name, #name, opc2, opc3) \
GEN_VXRFORM1_300(name##_dot, name##_, #name ".", opc2, (opc3 | (0x1 << 4)))
-GEN_VXRFORM_300(vcmpnezb, 3, 4)
-GEN_VXRFORM_300(vcmpnezh, 3, 5)
-GEN_VXRFORM_300(vcmpnezw, 3, 6)
GEN_VXRFORM(vcmpeqfp, 3, 3)
GEN_VXRFORM(vcmpgefp, 3, 7)
GEN_VXRFORM(vcmpgtfp, 3, 11)
--
2.25.1
- [PATCH v5 00/49] target/ppc: PowerISA Vector/VSX instruction batch, matheus . ferst, 2022/02/25
- [PATCH v5 01/49] target/ppc: Introduce TRANS*FLAGS macros, matheus . ferst, 2022/02/25
- [PATCH v5 02/49] target/ppc: moved vector even and odd multiplication to decodetree, matheus . ferst, 2022/02/25
- [PATCH v5 03/49] target/ppc: Moved vector multiply high and low to decodetree, matheus . ferst, 2022/02/25
- [PATCH v5 05/49] target/ppc: Implement vmsumcud instruction, matheus . ferst, 2022/02/25
- [PATCH v5 04/49] target/ppc: vmulh* instructions without helpers, matheus . ferst, 2022/02/25
- [PATCH v5 06/49] target/ppc: Implement vmsumudm instruction, matheus . ferst, 2022/02/25
- [PATCH v5 07/49] target/ppc: Move vexts[bhw]2[wd] to decodetree, matheus . ferst, 2022/02/25
- [PATCH v5 08/49] target/ppc: Implement vextsd2q, matheus . ferst, 2022/02/25
- [PATCH v5 09/49] target/ppc: Move Vector Compare Equal/Not Equal/Greater Than to decodetree, matheus . ferst, 2022/02/25
- [PATCH v5 10/49] target/ppc: Move Vector Compare Not Equal or Zero to decodetree,
matheus . ferst <=
- [PATCH v5 12/49] target/ppc: Implement Vector Compare Greater Than Quadword, matheus . ferst, 2022/02/25
- [PATCH v5 11/49] target/ppc: Implement Vector Compare Equal Quadword, matheus . ferst, 2022/02/25
- [PATCH v5 14/49] target/ppc: implement vstri[bh][lr], matheus . ferst, 2022/02/25
- [PATCH v5 13/49] target/ppc: Implement Vector Compare Quadword, matheus . ferst, 2022/02/25
- [PATCH v5 15/49] target/ppc: implement vclrlb, matheus . ferst, 2022/02/25
- [PATCH v5 16/49] target/ppc: implement vclrrb, matheus . ferst, 2022/02/25
- [PATCH v5 17/49] target/ppc: implement vcntmb[bhwd], matheus . ferst, 2022/02/25
- [PATCH v5 19/49] target/ppc: move vs[lr][a][bhwd] to decodetree, matheus . ferst, 2022/02/25
- [PATCH v5 18/49] target/ppc: implement vgnb, matheus . ferst, 2022/02/25
- [PATCH v5 21/49] target/ppc: implement vsrq, matheus . ferst, 2022/02/25