From 355d5584de1129eec1c4043fdee1335010cfabb6 Mon Sep 17 00:00:00 2001 From: "Yueh-Ting (eop) Chen" Date: Mon, 20 Jun 2022 06:51:02 +0000 Subject: target/riscv: rvv: Add mask agnostic for vv instructions According to v-spec, mask agnostic behavior can be either kept as undisturbed or set elements' bits to all 1s. To distinguish the difference of mask policies, QEMU should be able to simulate the mask agnostic behavior as "set mask elements' bits to all 1s". There are multiple possibility for agnostic elements according to v-spec. The main intent of this patch-set tries to add option that can distinguish between mask policies. Setting agnostic elements to all 1s allows QEMU to express this. This is the first commit regarding the optional mask agnostic behavior. Follow-up commits will add this optional behavior for all rvv instructions. Signed-off-by: eop Chen Reviewed-by: Frank Chang Reviewed-by: Weiwei Li Reviewed-by: Alistair Francis Message-Id: <165570784143.17634.35095816584573691-1@git.sr.ht> Signed-off-by: Alistair Francis --- target/riscv/vector_helper.c | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'target/riscv/vector_helper.c') diff --git a/target/riscv/vector_helper.c b/target/riscv/vector_helper.c index a96fc49c71..de895050e0 100644 --- a/target/riscv/vector_helper.c +++ b/target/riscv/vector_helper.c @@ -127,6 +127,11 @@ static inline uint32_t vext_vta(uint32_t desc) return FIELD_EX32(simd_data(desc), VDATA, VTA); } +static inline uint32_t vext_vma(uint32_t desc) +{ + return FIELD_EX32(simd_data(desc), VDATA, VMA); +} + static inline uint32_t vext_vta_all_1s(uint32_t desc) { return FIELD_EX32(simd_data(desc), VDATA, VTA_ALL_1S); @@ -812,10 +817,13 @@ static void do_vext_vv(void *vd, void *v0, void *vs1, void *vs2, uint32_t vl = env->vl; 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; } fn(vd, vs1, vs2, i); -- cgit 1.4.1 From 265ecd4c62a008f3be351a75e1847cee9c71e380 Mon Sep 17 00:00:00 2001 From: "Yueh-Ting (eop) Chen" Date: Mon, 20 Jun 2022 06:51:02 +0000 Subject: target/riscv: rvv: Add mask agnostic for vector load / store instructions Signed-off-by: eop Chen Reviewed-by: Frank Chang Reviewed-by: Weiwei Li Acked-by: Alistair Francis Message-Id: <165570784143.17634.35095816584573691-2@git.sr.ht> Signed-off-by: Alistair Francis --- target/riscv/insn_trans/trans_rvv.c.inc | 5 +++++ target/riscv/vector_helper.c | 35 ++++++++++++++++++++++----------- 2 files changed, 29 insertions(+), 11 deletions(-) (limited to 'target/riscv/vector_helper.c') diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc index 5ec113f6fd..0627eda0c0 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); } @@ -777,6 +778,7 @@ 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); } @@ -866,6 +868,7 @@ 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); } @@ -996,6 +999,7 @@ 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); } @@ -1114,6 +1118,7 @@ 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 de895050e0..e3810d2bc3 100644 --- a/target/riscv/vector_helper.c +++ b/target/riscv/vector_helper.c @@ -283,14 +283,18 @@ 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); + uint32_t vma = vext_vma(desc); 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++; @@ -482,15 +486,19 @@ 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); + uint32_t vma = vext_vma(desc); /* 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++; @@ -579,6 +587,7 @@ 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*/ @@ -624,10 +633,14 @@ 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++; -- cgit 1.4.1 From bce9a636beabbb2c538cdfca49592b04bfbf7e2c Mon Sep 17 00:00:00 2001 From: "Yueh-Ting (eop) Chen" Date: Mon, 20 Jun 2022 06:50:58 +0000 Subject: target/riscv: rvv: Add mask agnostic for vx instructions Signed-off-by: eop Chen Reviewed-by: Frank Chang Reviewed-by: Weiwei Li Acked-by: Alistair Francis Message-Id: <165570784143.17634.35095816584573691-3@git.sr.ht> Signed-off-by: Alistair Francis --- target/riscv/insn_trans/trans_rvv.c.inc | 2 ++ target/riscv/vector_helper.c | 3 +++ 2 files changed, 5 insertions(+) (limited to 'target/riscv/vector_helper.c') diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc index 0627eda0c0..07d86551a9 100644 --- a/target/riscv/insn_trans/trans_rvv.c.inc +++ b/target/riscv/insn_trans/trans_rvv.c.inc @@ -1301,6 +1301,7 @@ static bool opivx_trans(uint32_t vd, uint32_t rs1, uint32_t vs2, uint32_t vm, data = FIELD_DP32(data, VDATA, LMUL, s->lmul); data = FIELD_DP32(data, VDATA, VTA, s->vta); data = FIELD_DP32(data, VDATA, VTA_ALL_1S, s->cfg_vta_all_1s); + data = FIELD_DP32(data, VDATA, VMA, s->vma); desc = tcg_constant_i32(simd_desc(s->cfg_ptr->vlen / 8, s->cfg_ptr->vlen / 8, data)); @@ -1468,6 +1469,7 @@ static bool opivi_trans(uint32_t vd, uint32_t imm, uint32_t vs2, uint32_t vm, data = FIELD_DP32(data, VDATA, LMUL, s->lmul); data = FIELD_DP32(data, VDATA, VTA, s->vta); data = FIELD_DP32(data, VDATA, VTA_ALL_1S, s->cfg_vta_all_1s); + data = FIELD_DP32(data, VDATA, VMA, s->vma); desc = tcg_constant_i32(simd_desc(s->cfg_ptr->vlen / 8, s->cfg_ptr->vlen / 8, data)); diff --git a/target/riscv/vector_helper.c b/target/riscv/vector_helper.c index e3810d2bc3..6be3c4e739 100644 --- a/target/riscv/vector_helper.c +++ b/target/riscv/vector_helper.c @@ -899,10 +899,13 @@ static void do_vext_vx(void *vd, void *v0, target_long s1, void *vs2, uint32_t vl = env->vl; 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; } fn(vd, s1, vs2, i); -- cgit 1.4.1 From fd93045ebfa6ab07ce7017fb4095736c3f6f315a Mon Sep 17 00:00:00 2001 From: "Yueh-Ting (eop) Chen" Date: Mon, 20 Jun 2022 06:50:58 +0000 Subject: target/riscv: rvv: Add mask agnostic for vector integer shift instructions Signed-off-by: eop Chen Reviewed-by: Frank Chang Reviewed-by: Weiwei Li Acked-by: Alistair Francis Message-Id: <165570784143.17634.35095816584573691-4@git.sr.ht> Signed-off-by: Alistair Francis --- target/riscv/insn_trans/trans_rvv.c.inc | 1 + target/riscv/vector_helper.c | 7 +++++++ 2 files changed, 8 insertions(+) (limited to 'target/riscv/vector_helper.c') diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc index 07d86551a9..83b85bb851 100644 --- a/target/riscv/insn_trans/trans_rvv.c.inc +++ b/target/riscv/insn_trans/trans_rvv.c.inc @@ -1901,6 +1901,7 @@ static bool trans_##NAME(DisasContext *s, arg_rmrr *a) \ 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_4_ptr(vreg_ofs(s, a->rd), vreg_ofs(s, 0), \ vreg_ofs(s, a->rs1), \ vreg_ofs(s, a->rs2), cpu_env, \ diff --git a/target/riscv/vector_helper.c b/target/riscv/vector_helper.c index 6be3c4e739..d1daa764b7 100644 --- a/target/riscv/vector_helper.c +++ b/target/riscv/vector_helper.c @@ -1298,10 +1298,13 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, \ uint32_t esz = sizeof(TS1); \ 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; \ } \ TS1 s1 = *((TS1 *)vs1 + HS1(i)); \ @@ -1339,10 +1342,14 @@ void HELPER(NAME)(void *vd, void *v0, target_ulong s1, \ 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; \ } \ TS2 s2 = *((TS2 *)vs2 + HS2(i)); \ -- cgit 1.4.1 From 6e11d7eaa02c5834e5172d25e8c794ac5731b56e Mon Sep 17 00:00:00 2001 From: "Yueh-Ting (eop) Chen" Date: Mon, 20 Jun 2022 06:51:11 +0000 Subject: target/riscv: rvv: Add mask agnostic for vector integer comparison instructions Signed-off-by: eop Chen Reviewed-by: Frank Chang Reviewed-by: Weiwei Li Acked-by: Alistair Francis Message-Id: <165570784143.17634.35095816584573691-5@git.sr.ht> Signed-off-by: Alistair Francis --- target/riscv/insn_trans/trans_rvv.c.inc | 1 + target/riscv/vector_helper.c | 10 ++++++++++ 2 files changed, 11 insertions(+) (limited to 'target/riscv/vector_helper.c') diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc index 83b85bb851..e6aa5295a1 100644 --- a/target/riscv/insn_trans/trans_rvv.c.inc +++ b/target/riscv/insn_trans/trans_rvv.c.inc @@ -1718,6 +1718,7 @@ static bool trans_##NAME(DisasContext *s, arg_rmrr *a) \ data = FIELD_DP32(data, VDATA, VTA, s->vta); \ data = \ FIELD_DP32(data, VDATA, VTA_ALL_1S, s->cfg_vta_all_1s);\ + data = FIELD_DP32(data, VDATA, VMA, s->vma); \ tcg_gen_gvec_4_ptr(vreg_ofs(s, a->rd), vreg_ofs(s, 0), \ vreg_ofs(s, a->rs1), \ vreg_ofs(s, a->rs2), cpu_env, \ diff --git a/target/riscv/vector_helper.c b/target/riscv/vector_helper.c index d1daa764b7..07ce671879 100644 --- a/target/riscv/vector_helper.c +++ b/target/riscv/vector_helper.c @@ -1404,12 +1404,17 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, void *vs2, \ uint32_t vl = env->vl; \ uint32_t total_elems = env_archcpu(env)->cfg.vlen; \ uint32_t vta_all_1s = vext_vta_all_1s(desc); \ + uint32_t vma = vext_vma(desc); \ uint32_t i; \ \ for (i = env->vstart; i < vl; i++) { \ ETYPE s1 = *((ETYPE *)vs1 + H(i)); \ ETYPE s2 = *((ETYPE *)vs2 + H(i)); \ if (!vm && !vext_elem_mask(v0, i)) { \ + /* set masked-off elements to 1s */ \ + if (vma) { \ + vext_set_elem_mask(vd, i, 1); \ + } \ continue; \ } \ vext_set_elem_mask(vd, i, DO_OP(s2, s1)); \ @@ -1462,11 +1467,16 @@ void HELPER(NAME)(void *vd, void *v0, target_ulong s1, void *vs2, \ uint32_t vl = env->vl; \ uint32_t total_elems = env_archcpu(env)->cfg.vlen; \ uint32_t vta_all_1s = vext_vta_all_1s(desc); \ + uint32_t vma = vext_vma(desc); \ uint32_t i; \ \ for (i = env->vstart; i < vl; i++) { \ ETYPE s2 = *((ETYPE *)vs2 + H(i)); \ if (!vm && !vext_elem_mask(v0, i)) { \ + /* set masked-off elements to 1s */ \ + if (vma) { \ + vext_set_elem_mask(vd, i, 1); \ + } \ continue; \ } \ vext_set_elem_mask(vd, i, \ -- cgit 1.4.1 From 72e17a9f86098feb25d2c1f2af043dc5f9f211e3 Mon Sep 17 00:00:00 2001 From: "Yueh-Ting (eop) Chen" Date: Mon, 20 Jun 2022 06:51:11 +0000 Subject: target/riscv: rvv: Add mask agnostic for vector fix-point arithmetic instructions Signed-off-by: eop Chen Reviewed-by: Frank Chang Reviewed-by: Weiwei Li Acked-by: Alistair Francis Message-Id: <165570784143.17634.35095816584573691-6@git.sr.ht> Signed-off-by: Alistair Francis --- target/riscv/vector_helper.c | 26 ++++++++++++++++---------- 1 file changed, 16 insertions(+), 10 deletions(-) (limited to 'target/riscv/vector_helper.c') diff --git a/target/riscv/vector_helper.c b/target/riscv/vector_helper.c index 07ce671879..597fa9c752 100644 --- a/target/riscv/vector_helper.c +++ b/target/riscv/vector_helper.c @@ -2129,10 +2129,12 @@ static inline void vext_vv_rm_1(void *vd, void *v0, void *vs1, void *vs2, CPURISCVState *env, uint32_t vl, uint32_t vm, int vxrm, - opivv2_rm_fn *fn) + opivv2_rm_fn *fn, uint32_t vma, uint32_t esz) { for (uint32_t 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; } fn(vd, vs1, vs2, i, env, vxrm); @@ -2150,23 +2152,24 @@ vext_vv_rm_2(void *vd, void *v0, void *vs1, void *vs2, uint32_t vl = env->vl; uint32_t total_elems = vext_get_total_elems(env, desc, esz); uint32_t vta = vext_vta(desc); + uint32_t vma = vext_vma(desc); switch (env->vxrm) { case 0: /* rnu */ vext_vv_rm_1(vd, v0, vs1, vs2, - env, vl, vm, 0, fn); + env, vl, vm, 0, fn, vma, esz); break; case 1: /* rne */ vext_vv_rm_1(vd, v0, vs1, vs2, - env, vl, vm, 1, fn); + env, vl, vm, 1, fn, vma, esz); break; case 2: /* rdn */ vext_vv_rm_1(vd, v0, vs1, vs2, - env, vl, vm, 2, fn); + env, vl, vm, 2, fn, vma, esz); break; default: /* rod */ vext_vv_rm_1(vd, v0, vs1, vs2, - env, vl, vm, 3, fn); + env, vl, vm, 3, fn, vma, esz); break; } /* set tail elements to 1s */ @@ -2250,10 +2253,12 @@ static inline void vext_vx_rm_1(void *vd, void *v0, target_long s1, void *vs2, CPURISCVState *env, uint32_t vl, uint32_t vm, int vxrm, - opivx2_rm_fn *fn) + opivx2_rm_fn *fn, uint32_t vma, uint32_t esz) { for (uint32_t 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; } fn(vd, s1, vs2, i, env, vxrm); @@ -2271,23 +2276,24 @@ vext_vx_rm_2(void *vd, void *v0, target_long s1, void *vs2, uint32_t vl = env->vl; uint32_t total_elems = vext_get_total_elems(env, desc, esz); uint32_t vta = vext_vta(desc); + uint32_t vma = vext_vma(desc); switch (env->vxrm) { case 0: /* rnu */ vext_vx_rm_1(vd, v0, s1, vs2, - env, vl, vm, 0, fn); + env, vl, vm, 0, fn, vma, esz); break; case 1: /* rne */ vext_vx_rm_1(vd, v0, s1, vs2, - env, vl, vm, 1, fn); + env, vl, vm, 1, fn, vma, esz); break; case 2: /* rdn */ vext_vx_rm_1(vd, v0, s1, vs2, - env, vl, vm, 2, fn); + env, vl, vm, 2, fn, vma, esz); break; default: /* rod */ vext_vx_rm_1(vd, v0, s1, vs2, - env, vl, vm, 3, fn); + env, vl, vm, 3, fn, vma, esz); break; } /* set tail elements to 1s */ -- cgit 1.4.1 From 5b448f44c92f7355ae2b02af0699d22674cc5f6e Mon Sep 17 00:00:00 2001 From: "Yueh-Ting (eop) Chen" Date: Mon, 20 Jun 2022 06:51:12 +0000 Subject: target/riscv: rvv: Add mask agnostic for vector floating-point instructions Signed-off-by: eop Chen Reviewed-by: Frank Chang Reviewed-by: Weiwei Li Acked-by: Alistair Francis Message-Id: <165570784143.17634.35095816584573691-7@git.sr.ht> Signed-off-by: Alistair Francis --- target/riscv/insn_trans/trans_rvv.c.inc | 12 ++++++++++++ target/riscv/vector_helper.c | 26 ++++++++++++++++++++++++++ 2 files changed, 38 insertions(+) (limited to 'target/riscv/vector_helper.c') diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc index e6aa5295a1..8ce3d28603 100644 --- a/target/riscv/insn_trans/trans_rvv.c.inc +++ b/target/riscv/insn_trans/trans_rvv.c.inc @@ -2361,6 +2361,7 @@ static bool trans_##NAME(DisasContext *s, arg_rmrr *a) \ data = FIELD_DP32(data, VDATA, VTA, s->vta); \ data = \ FIELD_DP32(data, VDATA, VTA_ALL_1S, s->cfg_vta_all_1s);\ + data = FIELD_DP32(data, VDATA, VMA, s->vma); \ tcg_gen_gvec_4_ptr(vreg_ofs(s, a->rd), vreg_ofs(s, 0), \ vreg_ofs(s, a->rs1), \ vreg_ofs(s, a->rs2), cpu_env, \ @@ -2446,6 +2447,7 @@ static bool trans_##NAME(DisasContext *s, arg_rmrr *a) \ data = FIELD_DP32(data, VDATA, VTA, s->vta); \ data = FIELD_DP32(data, VDATA, VTA_ALL_1S, \ s->cfg_vta_all_1s); \ + data = FIELD_DP32(data, VDATA, VMA, s->vma); \ return opfvf_trans(a->rd, a->rs1, a->rs2, data, \ fns[s->sew - 1], s); \ } \ @@ -2485,6 +2487,7 @@ static bool trans_##NAME(DisasContext *s, arg_rmrr *a) \ 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_4_ptr(vreg_ofs(s, a->rd), vreg_ofs(s, 0), \ vreg_ofs(s, a->rs1), \ vreg_ofs(s, a->rs2), cpu_env, \ @@ -2525,6 +2528,7 @@ static bool trans_##NAME(DisasContext *s, arg_rmrr *a) \ 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); \ return opfvf_trans(a->rd, a->rs1, a->rs2, data, \ fns[s->sew - 1], s); \ } \ @@ -2562,6 +2566,7 @@ static bool trans_##NAME(DisasContext *s, arg_rmrr *a) \ 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_4_ptr(vreg_ofs(s, a->rd), vreg_ofs(s, 0), \ vreg_ofs(s, a->rs1), \ vreg_ofs(s, a->rs2), cpu_env, \ @@ -2602,6 +2607,7 @@ static bool trans_##NAME(DisasContext *s, arg_rmrr *a) \ 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); \ return opfvf_trans(a->rd, a->rs1, a->rs2, data, \ fns[s->sew - 1], s); \ } \ @@ -2686,6 +2692,7 @@ static bool do_opfv(DisasContext *s, arg_rmr *a, 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, s->cfg_ptr->vlen / 8, @@ -2790,6 +2797,7 @@ static bool trans_vfmv_v_f(DisasContext *s, arg_vfmv_v_f *a) TCGv_i32 desc; uint32_t data = FIELD_DP32(0, VDATA, LMUL, s->lmul); data = FIELD_DP32(data, VDATA, VTA, s->vta); + data = FIELD_DP32(data, VDATA, VMA, s->vma); static gen_helper_vmv_vx * const fns[3] = { gen_helper_vmv_v_x_h, gen_helper_vmv_v_x_w, @@ -2891,6 +2899,7 @@ static bool trans_##NAME(DisasContext *s, arg_rmr *a) \ 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, \ s->cfg_ptr->vlen / 8, \ @@ -2944,6 +2953,7 @@ static bool trans_##NAME(DisasContext *s, arg_rmr *a) \ 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, \ s->cfg_ptr->vlen / 8, \ @@ -3012,6 +3022,7 @@ static bool trans_##NAME(DisasContext *s, arg_rmr *a) \ 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, \ s->cfg_ptr->vlen / 8, \ @@ -3067,6 +3078,7 @@ static bool trans_##NAME(DisasContext *s, arg_rmr *a) \ 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, \ s->cfg_ptr->vlen / 8, \ diff --git a/target/riscv/vector_helper.c b/target/riscv/vector_helper.c index 597fa9c752..315742c6b8 100644 --- a/target/riscv/vector_helper.c +++ b/target/riscv/vector_helper.c @@ -3051,10 +3051,14 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, \ 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; \ } \ do_##NAME(vd, vs1, vs2, i, env); \ @@ -3090,10 +3094,14 @@ void HELPER(NAME)(void *vd, void *v0, uint64_t s1, \ 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; \ } \ do_##NAME(vd, s1, vs2, i, env); \ @@ -3665,6 +3673,7 @@ void HELPER(NAME)(void *vd, void *v0, void *vs2, \ 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; \ \ if (vl == 0) { \ @@ -3672,6 +3681,9 @@ void HELPER(NAME)(void *vd, void *v0, void *vs2, \ } \ 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; \ } \ do_##NAME(vd, vs2, i, env); \ @@ -4182,12 +4194,17 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, void *vs2, \ uint32_t vl = env->vl; \ uint32_t total_elems = env_archcpu(env)->cfg.vlen; \ uint32_t vta_all_1s = vext_vta_all_1s(desc); \ + uint32_t vma = vext_vma(desc); \ uint32_t i; \ \ for (i = env->vstart; i < vl; i++) { \ ETYPE s1 = *((ETYPE *)vs1 + H(i)); \ ETYPE s2 = *((ETYPE *)vs2 + H(i)); \ if (!vm && !vext_elem_mask(v0, i)) { \ + /* set masked-off elements to 1s */ \ + if (vma) { \ + vext_set_elem_mask(vd, i, 1); \ + } \ continue; \ } \ vext_set_elem_mask(vd, i, \ @@ -4215,11 +4232,16 @@ void HELPER(NAME)(void *vd, void *v0, uint64_t s1, void *vs2, \ uint32_t vl = env->vl; \ uint32_t total_elems = env_archcpu(env)->cfg.vlen; \ uint32_t vta_all_1s = vext_vta_all_1s(desc); \ + uint32_t vma = vext_vma(desc); \ uint32_t i; \ \ for (i = env->vstart; i < vl; i++) { \ ETYPE s2 = *((ETYPE *)vs2 + H(i)); \ if (!vm && !vext_elem_mask(v0, i)) { \ + /* set masked-off elements to 1s */ \ + if (vma) { \ + vext_set_elem_mask(vd, i, 1); \ + } \ continue; \ } \ vext_set_elem_mask(vd, i, \ @@ -4342,10 +4364,14 @@ void HELPER(NAME)(void *vd, void *v0, void *vs2, \ 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; \ } \ do_##NAME(vd, vs2, i); \ -- cgit 1.4.1 From 35f2d795f313c43af5851ca2243317c0b0834c6c Mon Sep 17 00:00:00 2001 From: "Yueh-Ting (eop) Chen" Date: Mon, 20 Jun 2022 06:51:11 +0000 Subject: target/riscv: rvv: Add mask agnostic for vector mask instructions Signed-off-by: eop Chen Reviewed-by: Frank Chang Reviewed-by: Weiwei Li Acked-by: Alistair Francis Message-Id: <165570784143.17634.35095816584573691-8@git.sr.ht> Signed-off-by: Alistair Francis --- target/riscv/insn_trans/trans_rvv.c.inc | 3 +++ target/riscv/vector_helper.c | 11 +++++++++++ 2 files changed, 14 insertions(+) (limited to 'target/riscv/vector_helper.c') diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc index 8ce3d28603..c1bd29329e 100644 --- a/target/riscv/insn_trans/trans_rvv.c.inc +++ b/target/riscv/insn_trans/trans_rvv.c.inc @@ -3275,6 +3275,7 @@ static bool trans_##NAME(DisasContext *s, arg_rmr *a) \ data = FIELD_DP32(data, VDATA, LMUL, s->lmul); \ data = \ FIELD_DP32(data, VDATA, VTA_ALL_1S, s->cfg_vta_all_1s);\ + 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, s->cfg_ptr->vlen / 8, \ @@ -3313,6 +3314,7 @@ static bool trans_viota_m(DisasContext *s, arg_viota_m *a) 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); static gen_helper_gvec_3_ptr * const fns[4] = { gen_helper_viota_m_b, gen_helper_viota_m_h, gen_helper_viota_m_w, gen_helper_viota_m_d, @@ -3343,6 +3345,7 @@ static bool trans_vid_v(DisasContext *s, arg_vid_v *a) 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); static gen_helper_gvec_2_ptr * const fns[4] = { gen_helper_vid_v_b, gen_helper_vid_v_h, gen_helper_vid_v_w, gen_helper_vid_v_d, diff --git a/target/riscv/vector_helper.c b/target/riscv/vector_helper.c index 315742c6b8..52518648bb 100644 --- a/target/riscv/vector_helper.c +++ b/target/riscv/vector_helper.c @@ -4879,11 +4879,16 @@ static void vmsetm(void *vd, void *v0, void *vs2, CPURISCVState *env, uint32_t vl = env->vl; uint32_t total_elems = env_archcpu(env)->cfg.vlen; uint32_t vta_all_1s = vext_vta_all_1s(desc); + uint32_t vma = vext_vma(desc); int i; bool first_mask_bit = false; for (i = env->vstart; i < vl; i++) { if (!vm && !vext_elem_mask(v0, i)) { + /* set masked-off elements to 1s */ + if (vma) { + vext_set_elem_mask(vd, i, 1); + } continue; } /* write a zero to all following active elements */ @@ -4944,11 +4949,14 @@ void HELPER(NAME)(void *vd, void *v0, void *vs2, CPURISCVState *env, \ 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 sum = 0; \ int 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 + H(i)) = sum; \ @@ -4975,10 +4983,13 @@ void HELPER(NAME)(void *vd, void *v0, CPURISCVState *env, uint32_t desc) \ 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); \ int 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 + H(i)) = i; \ -- cgit 1.4.1 From edabcd0e0aea2ac8d68931f31fcf8d3b99a28f20 Mon Sep 17 00:00:00 2001 From: "Yueh-Ting (eop) Chen" Date: Mon, 20 Jun 2022 06:51:12 +0000 Subject: target/riscv: rvv: Add mask agnostic for vector permutation instructions Signed-off-by: eop Chen Reviewed-by: Frank Chang Reviewed-by: Weiwei Li Acked-by: Alistair Francis Message-Id: <165570784143.17634.35095816584573691-9@git.sr.ht> Signed-off-by: Alistair Francis --- target/riscv/insn_trans/trans_rvv.c.inc | 1 + target/riscv/vector_helper.c | 26 ++++++++++++++++++++++++-- 2 files changed, 25 insertions(+), 2 deletions(-) (limited to 'target/riscv/vector_helper.c') 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, 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)); \ -- cgit 1.4.1