[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[PULL 23/44] target/riscv: rvv: Add mask agnostic for vector permutation
|
From: |
Alistair Francis |
|
Subject: |
[PULL 23/44] target/riscv: rvv: Add mask agnostic for vector permutation instructions |
|
Date: |
Wed, 7 Sep 2022 10:03:32 +0200 |
From: "Yueh-Ting (eop) Chen" <eop.chen@sifive.com>
Signed-off-by: eop Chen <eop.chen@sifive.com>
Reviewed-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Weiwei Li <liweiwei@iscas.ac.cn>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <165570784143.17634.35095816584573691-9@git.sr.ht>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
---
target/riscv/vector_helper.c | 26 +++++++++++++++++++++++--
target/riscv/insn_trans/trans_rvv.c.inc | 1 +
2 files changed, 25 insertions(+), 2 deletions(-)
diff --git a/target/riscv/vector_helper.c b/target/riscv/vector_helper.c
index 52518648bb..d224861c2c 100644
--- a/target/riscv/vector_helper.c
+++ b/target/riscv/vector_helper.c
@@ -5018,11 +5018,14 @@ void HELPER(NAME)(void *vd, void *v0, target_ulong s1,
void *vs2, \
uint32_t esz = sizeof(ETYPE); \
uint32_t total_elems = vext_get_total_elems(env, desc, esz); \
uint32_t vta = vext_vta(desc); \
+ uint32_t vma = vext_vma(desc); \
target_ulong offset = s1, i_min, i; \
\
i_min = MAX(env->vstart, offset); \
for (i = i_min; i < vl; i++) { \
if (!vm && !vext_elem_mask(v0, i)) { \
+ /* set masked-off elements to 1s */ \
+ vext_set_elems_1s(vd, vma, i * esz, (i + 1) * esz); \
continue; \
} \
*((ETYPE *)vd + H(i)) = *((ETYPE *)vs2 + H(i - offset)); \
@@ -5047,13 +5050,17 @@ void HELPER(NAME)(void *vd, void *v0, target_ulong s1,
void *vs2, \
uint32_t esz = sizeof(ETYPE); \
uint32_t total_elems = vext_get_total_elems(env, desc, esz); \
uint32_t vta = vext_vta(desc); \
+ uint32_t vma = vext_vma(desc); \
target_ulong i_max, i; \
\
i_max = MAX(MIN(s1 < vlmax ? vlmax - s1 : 0, vl), env->vstart); \
for (i = env->vstart; i < i_max; ++i) { \
- if (vm || vext_elem_mask(v0, i)) { \
- *((ETYPE *)vd + H(i)) = *((ETYPE *)vs2 + H(i + s1)); \
+ if (!vm && !vext_elem_mask(v0, i)) { \
+ /* set masked-off elements to 1s */ \
+ vext_set_elems_1s(vd, vma, i * esz, (i + 1) * esz); \
+ continue; \
} \
+ *((ETYPE *)vd + H(i)) = *((ETYPE *)vs2 + H(i + s1)); \
} \
\
for (i = i_max; i < vl; ++i) { \
@@ -5083,10 +5090,13 @@ static void vslide1up_##BITWIDTH(void *vd, void *v0,
target_ulong s1, \
uint32_t esz = sizeof(ETYPE); \
uint32_t total_elems = vext_get_total_elems(env, desc, esz); \
uint32_t vta = vext_vta(desc); \
+ uint32_t vma = vext_vma(desc); \
uint32_t i; \
\
for (i = env->vstart; i < vl; i++) { \
if (!vm && !vext_elem_mask(v0, i)) { \
+ /* set masked-off elements to 1s */ \
+ vext_set_elems_1s(vd, vma, i * esz, (i + 1) * esz); \
continue; \
} \
if (i == 0) { \
@@ -5128,10 +5138,13 @@ static void vslide1down_##BITWIDTH(void *vd, void *v0,
target_ulong s1, \
uint32_t esz = sizeof(ETYPE); \
uint32_t total_elems = vext_get_total_elems(env, desc, esz); \
uint32_t vta = vext_vta(desc); \
+ uint32_t vma = vext_vma(desc); \
uint32_t i; \
\
for (i = env->vstart; i < vl; i++) { \
if (!vm && !vext_elem_mask(v0, i)) { \
+ /* set masked-off elements to 1s */ \
+ vext_set_elems_1s(vd, vma, i * esz, (i + 1) * esz); \
continue; \
} \
if (i == vl - 1) { \
@@ -5199,11 +5212,14 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, void
*vs2, \
uint32_t esz = sizeof(TS2); \
uint32_t total_elems = vext_get_total_elems(env, desc, esz); \
uint32_t vta = vext_vta(desc); \
+ uint32_t vma = vext_vma(desc); \
uint64_t index; \
uint32_t i; \
\
for (i = env->vstart; i < vl; i++) { \
if (!vm && !vext_elem_mask(v0, i)) { \
+ /* set masked-off elements to 1s */ \
+ vext_set_elems_1s(vd, vma, i * esz, (i + 1) * esz); \
continue; \
} \
index = *((TS1 *)vs1 + HS1(i)); \
@@ -5239,11 +5255,14 @@ void HELPER(NAME)(void *vd, void *v0, target_ulong s1,
void *vs2, \
uint32_t esz = sizeof(ETYPE); \
uint32_t total_elems = vext_get_total_elems(env, desc, esz); \
uint32_t vta = vext_vta(desc); \
+ uint32_t vma = vext_vma(desc); \
uint64_t index = s1; \
uint32_t i; \
\
for (i = env->vstart; i < vl; i++) { \
if (!vm && !vext_elem_mask(v0, i)) { \
+ /* set masked-off elements to 1s */ \
+ vext_set_elems_1s(vd, vma, i * esz, (i + 1) * esz); \
continue; \
} \
if (index >= vlmax) { \
@@ -5318,10 +5337,13 @@ void HELPER(NAME)(void *vd, void *v0, void *vs2,
\
uint32_t esz = sizeof(ETYPE); \
uint32_t total_elems = vext_get_total_elems(env, desc, esz); \
uint32_t vta = vext_vta(desc); \
+ uint32_t vma = vext_vma(desc); \
uint32_t i; \
\
for (i = env->vstart; i < vl; i++) { \
if (!vm && !vext_elem_mask(v0, i)) { \
+ /* set masked-off elements to 1s */ \
+ vext_set_elems_1s(vd, vma, i * esz, (i + 1) * esz); \
continue; \
} \
*((ETYPE *)vd + HD(i)) = *((DTYPE *)vs2 + HS1(i)); \
diff --git a/target/riscv/insn_trans/trans_rvv.c.inc
b/target/riscv/insn_trans/trans_rvv.c.inc
index c1bd29329e..e58208f363 100644
--- a/target/riscv/insn_trans/trans_rvv.c.inc
+++ b/target/riscv/insn_trans/trans_rvv.c.inc
@@ -3891,6 +3891,7 @@ static bool int_ext_op(DisasContext *s, arg_rmr *a,
uint8_t seq)
data = FIELD_DP32(data, VDATA, VM, a->vm);
data = FIELD_DP32(data, VDATA, LMUL, s->lmul);
data = FIELD_DP32(data, VDATA, VTA, s->vta);
+ data = FIELD_DP32(data, VDATA, VMA, s->vma);
tcg_gen_gvec_3_ptr(vreg_ofs(s, a->rd), vreg_ofs(s, 0),
vreg_ofs(s, a->rs2), cpu_env,
--
2.37.2
- [PULL 11/44] target/riscv: Simplify the check in hmode to reuse the check in riscv_csrrw_check, (continued)
- [PULL 11/44] target/riscv: Simplify the check in hmode to reuse the check in riscv_csrrw_check, Alistair Francis, 2022/09/07
- [PULL 15/44] target/riscv: rvv: Add mask agnostic for vv instructions, Alistair Francis, 2022/09/07
- [PULL 14/44] docs: List kvm as a supported accelerator on RISC-V, Alistair Francis, 2022/09/07
- [PULL 16/44] target/riscv: rvv: Add mask agnostic for vector load / store instructions, Alistair Francis, 2022/09/07
- [PULL 17/44] target/riscv: rvv: Add mask agnostic for vx instructions, Alistair Francis, 2022/09/07
- [PULL 19/44] target/riscv: rvv: Add mask agnostic for vector integer comparison instructions, Alistair Francis, 2022/09/07
- [PULL 18/44] target/riscv: rvv: Add mask agnostic for vector integer shift instructions, Alistair Francis, 2022/09/07
- [PULL 21/44] target/riscv: rvv: Add mask agnostic for vector floating-point instructions, Alistair Francis, 2022/09/07
- [PULL 20/44] target/riscv: rvv: Add mask agnostic for vector fix-point arithmetic instructions, Alistair Francis, 2022/09/07
- [PULL 22/44] target/riscv: rvv: Add mask agnostic for vector mask instructions, Alistair Francis, 2022/09/07
- [PULL 23/44] target/riscv: rvv: Add mask agnostic for vector permutation instructions,
Alistair Francis <=
- [PULL 25/44] target/riscv: Add Zihintpause support, Alistair Francis, 2022/09/07
- [PULL 24/44] target/riscv: rvv: Add option 'rvv_ma_all_1s' to enable optional mask agnostic behavior, Alistair Francis, 2022/09/07
- [PULL 26/44] hw/riscv: remove 'fdt' param from riscv_setup_rom_reset_vec(), Alistair Francis, 2022/09/07
- [PULL 29/44] hw/riscv: microchip_pfsoc: fix kernel panics due to missing peripherals, Alistair Francis, 2022/09/07
- [PULL 28/44] hw/riscv: opentitan: bump opentitan version, Alistair Francis, 2022/09/07
- [PULL 31/44] hw/riscv: virt: fix uart node name, Alistair Francis, 2022/09/07
- [PULL 35/44] target/riscv: Add xicondops in ISA entry, Alistair Francis, 2022/09/07
- [PULL 43/44] hw/riscv: virt: Add PMU DT node to the device tree, Alistair Francis, 2022/09/07
- [PULL 36/44] target/riscv: Use official extension names for AIA CSRs, Alistair Francis, 2022/09/07
- [PULL 34/44] hw/core: fix platform bus node name, Alistair Francis, 2022/09/07