From: Yueh-Ting (eop) Chen <eop.c...@sifive.com>
Signed-off-by: eop Chen <eop.c...@sifive.com>
Reviewed-by: Frank Chang <frank.ch...@sifive.com>
---
target/riscv/insn_trans/trans_rvv.c.inc | 9 +++++++
target/riscv/vector_helper.c | 35 +++++++++++++++++--------
2 files changed, 33 insertions(+), 11 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvv.c.inc
b/target/riscv/insn_trans/trans_rvv.c.inc
index 94a558bf1a..325c5120d4 100644
--- a/target/riscv/insn_trans/trans_rvv.c.inc
+++ b/target/riscv/insn_trans/trans_rvv.c.inc
@@ -712,6 +712,7 @@ static bool ld_us_op(DisasContext *s, arg_r2nfvm *a,
uint8_t eew)
data = FIELD_DP32(data, VDATA, LMUL, emul);
data = FIELD_DP32(data, VDATA, NF, a->nf);
data = FIELD_DP32(data, VDATA, VTA, s->vta);
+ data = FIELD_DP32(data, VDATA, VMA, s->vma);
return ldst_us_trans(a->rd, a->rs1, data, fn, s, false);
}
@@ -750,6 +751,7 @@ static bool st_us_op(DisasContext *s, arg_r2nfvm *a,
uint8_t eew)
data = FIELD_DP32(data, VDATA, LMUL, emul);
data = FIELD_DP32(data, VDATA, NF, a->nf);
data = FIELD_DP32(data, VDATA, VTA, s->vta);
+ data = FIELD_DP32(data, VDATA, VMA, s->vma);
return ldst_us_trans(a->rd, a->rs1, data, fn, s, true);
}
@@ -778,6 +780,7 @@ static bool ld_us_mask_op(DisasContext *s, arg_vlm_v *a,
uint8_t eew)
data = FIELD_DP32(data, VDATA, NF, 1);
/* Mask destination register are always tail-agnostic */
data = FIELD_DP32(data, VDATA, VTA, s->cfg_vta_all_1s);
+ data = FIELD_DP32(data, VDATA, VMA, s->vma);
return ldst_us_trans(a->rd, a->rs1, data, fn, s, false);
}
@@ -797,6 +800,7 @@ static bool st_us_mask_op(DisasContext *s, arg_vsm_v *a,
uint8_t eew)
data = FIELD_DP32(data, VDATA, NF, 1);
/* Mask destination register are always tail-agnostic */
data = FIELD_DP32(data, VDATA, VTA, s->cfg_vta_all_1s);
+ data = FIELD_DP32(data, VDATA, VMA, s->vma);
return ldst_us_trans(a->rd, a->rs1, data, fn, s, true);
}
@@ -869,6 +873,7 @@ static bool ld_stride_op(DisasContext *s, arg_rnfvm *a,
uint8_t eew)
data = FIELD_DP32(data, VDATA, LMUL, emul);
data = FIELD_DP32(data, VDATA, NF, a->nf);
data = FIELD_DP32(data, VDATA, VTA, s->vta);
+ data = FIELD_DP32(data, VDATA, VMA, s->vma);
return ldst_stride_trans(a->rd, a->rs1, a->rs2, data, fn, s, false);
}
@@ -899,6 +904,7 @@ static bool st_stride_op(DisasContext *s, arg_rnfvm *a,
uint8_t eew)
data = FIELD_DP32(data, VDATA, LMUL, emul);
data = FIELD_DP32(data, VDATA, NF, a->nf);
data = FIELD_DP32(data, VDATA, VTA, s->vta);
+ data = FIELD_DP32(data, VDATA, VMA, s->vma);
fn = fns[eew];
if (fn == NULL) {
return false;
@@ -1000,6 +1006,7 @@ static bool ld_index_op(DisasContext *s, arg_rnfvm *a,
uint8_t eew)
data = FIELD_DP32(data, VDATA, LMUL, emul);
data = FIELD_DP32(data, VDATA, NF, a->nf);
data = FIELD_DP32(data, VDATA, VTA, s->vta);
+ data = FIELD_DP32(data, VDATA, VMA, s->vma);
return ldst_index_trans(a->rd, a->rs1, a->rs2, data, fn, s, false);
}
@@ -1053,6 +1060,7 @@ static bool st_index_op(DisasContext *s, arg_rnfvm *a,
uint8_t eew)
data = FIELD_DP32(data, VDATA, LMUL, emul);
data = FIELD_DP32(data, VDATA, NF, a->nf);
data = FIELD_DP32(data, VDATA, VTA, s->vta);
+ data = FIELD_DP32(data, VDATA, VMA, s->vma);
return ldst_index_trans(a->rd, a->rs1, a->rs2, data, fn, s, true);
}
@@ -1119,6 +1127,7 @@ static bool ldff_op(DisasContext *s, arg_r2nfvm *a,
uint8_t eew)
data = FIELD_DP32(data, VDATA, LMUL, emul);
data = FIELD_DP32(data, VDATA, NF, a->nf);
data = FIELD_DP32(data, VDATA, VTA, s->vta);
+ data = FIELD_DP32(data, VDATA, VMA, s->vma);
return ldff_trans(a->rd, a->rs1, data, fn, s);
}
diff --git a/target/riscv/vector_helper.c b/target/riscv/vector_helper.c
index a82ca2fe5a..a92d7dbd3c 100644
--- a/target/riscv/vector_helper.c
+++ b/target/riscv/vector_helper.c
@@ -282,14 +282,18 @@ vext_ldst_stride(void *vd, void *v0, target_ulong base,
uint32_t esz = 1 << log2_esz;
uint32_t total_elems = vext_get_total_elems(env, desc, esz);
uint32_t vta = vext_vta(desc) & is_load;
+ uint32_t vma = vext_vma(desc) & is_load;
for (i = env->vstart; i < env->vl; i++, env->vstart++) {
- if (!vm && !vext_elem_mask(v0, i)) {
- continue;
- }
-
k = 0;
while (k < nf) {
+ if (!vm && !vext_elem_mask(v0, i)) {
+ /* set masked-off elements to 1s */
+ vext_set_elems_1s(vd, vma, (i + k * max_elems) * esz,
+ (i + k * max_elems + 1) * esz);
+ k++;
+ continue;
+ }
target_ulong addr = base + stride * i + (k << log2_esz);
ldst_elem(env, adjust_addr(env, addr), i + k * max_elems, vd, ra);
k++;
@@ -481,15 +485,19 @@ vext_ldst_index(void *vd, void *v0, target_ulong base,
uint32_t esz = 1 << log2_esz;
uint32_t total_elems = vext_get_total_elems(env, desc, esz);
uint32_t vta = vext_vta(desc) & is_load;
+ uint32_t vma = vext_vma(desc) & is_load;
/* load bytes from guest memory */
for (i = env->vstart; i < env->vl; i++, env->vstart++) {
- if (!vm && !vext_elem_mask(v0, i)) {
- continue;
- }
-
k = 0;
while (k < nf) {
+ if (!vm && !vext_elem_mask(v0, i)) {
+ /* set masked-off elements to 1s */
+ vext_set_elems_1s(vd, vma, (i + k * max_elems) * esz,
+ (i + k * max_elems + 1) * esz);
+ k++;
+ continue;
+ }
abi_ptr addr = get_index_addr(base, i, vs2) + (k << log2_esz);
ldst_elem(env, adjust_addr(env, addr), i + k * max_elems, vd, ra);
k++;
@@ -578,6 +586,7 @@ vext_ldff(void *vd, void *v0, target_ulong base,
uint32_t esz = 1 << log2_esz;
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 addr, offset, remain;
/* probe every access*/
@@ -623,10 +632,14 @@ ProbeSuccess:
}
for (i = env->vstart; i < env->vl; i++) {
k = 0;
- if (!vm && !vext_elem_mask(v0, i)) {
- continue;
- }
while (k < nf) {
+ if (!vm && !vext_elem_mask(v0, i)) {
+ /* set masked-off elements to 1s */
+ vext_set_elems_1s(vd, vma, (i + k * max_elems) * esz,
+ (i + k * max_elems + 1) * esz);
+ k++;
+ continue;
+ }
target_ulong addr = base + ((i * nf + k) << log2_esz);
ldst_elem(env, adjust_addr(env, addr), i + k * max_elems, vd, ra);
k++;
--
2.34.2
