* [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions
@ 2021-12-29 2:33 frank.chang
2021-12-29 2:33 ` [PATCH 01/17] target/riscv: rvv-1.0: Add Zve64f extension into RISC-V frank.chang
` (17 more replies)
0 siblings, 18 replies; 37+ messages in thread
From: frank.chang @ 2021-12-29 2:33 UTC (permalink / raw)
To: qemu-devel; +Cc: Frank Chang, qemu-riscv
From: Frank Chang <frank.chang@sifive.com>
In RVV v1.0 spec, several Zve* vector extensions for embedded processors
are defined in Chapter 18.2:
https://github.com/riscv/riscv-v-spec/blob/v1.0/v-spec.adoc#zve-vector-extensions-for-embedded-processors
This patchset implements Zve32f and Zve64f extensions.
The port is available at:
https://github.com/sifive/qemu/tree/rvv-zve32f-zve64f-upstream
Zve32f can be enabled with -cpu option: Zve32f=true and
Zve64f can be enabled with -cpu option: Zve64f=true.
V is not required to be enabled explicitly.
Quote from the inclusion diagram for the six standard vector extensions
from Nick Knight <nick.knight@sifive.com>:
V
|
Zve64d
|
Zve64f
/ \
Zve64x Zve32f
\ /
Zve32x
Note: This patchset depends on other patchsets listed in Based-on
section below so it is not able to be built unless those patchsets
are applied.
Based-on: <20211229021250.29804-1-frank.chang@sifive.com>
Frank Chang (17):
target/riscv: rvv-1.0: Add Zve64f extension into RISC-V
target/riscv: rvv-1.0: Add Zve64f support for configuration insns
target/riscv: rvv-1.0: Add Zve64f support for load and store insns
target/riscv: rvv-1.0: Add Zve64f support for vmulh variant insns
target/riscv: rvv-1.0: Add Zve64f support for vsmul.vv and vsmul.vx
insns
target/riscv: rvv-1.0: Add Zve64f support for scalar fp insns
target/riscv: rvv-1.0: Add Zve64f support for single-width fp
reduction insns
target/riscv: rvv-1.0: Add Zve64f support for widening type-convert
insns
target/riscv: rvv-1.0: Add Zve64f support for narrowing type-convert
insns
target/riscv: rvv-1.0: Allow Zve64f extension to be turned on
target/riscv: rvv-1.0: Add Zve32f extension into RISC-V
target/riscv: rvv-1.0: Add Zve32f support for configuration insns
target/riscv: rvv-1.0: Add Zve32f support for scalar fp insns
target/riscv: rvv-1.0: Add Zve32f support for single-width fp
reduction insns
target/riscv: rvv-1.0: Add Zve32f support for widening type-convert
insns
target/riscv: rvv-1.0: Add Zve32f support for narrowing type-convert
insns
target/riscv: rvv-1.0: Allow Zve32f extension to be turned on
target/riscv/cpu.c | 6 +
target/riscv/cpu.h | 2 +
target/riscv/cpu_helper.c | 5 +-
target/riscv/csr.c | 6 +-
target/riscv/insn_trans/trans_rvv.c.inc | 217 ++++++++++++++++++++----
target/riscv/translate.c | 4 +
6 files changed, 203 insertions(+), 37 deletions(-)
--
2.31.1
^ permalink raw reply [flat|nested] 37+ messages in thread
* [PATCH 01/17] target/riscv: rvv-1.0: Add Zve64f extension into RISC-V
2021-12-29 2:33 [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions frank.chang
@ 2021-12-29 2:33 ` frank.chang
2022-01-17 22:23 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 02/17] target/riscv: rvv-1.0: Add Zve64f support for configuration insns frank.chang
` (16 subsequent siblings)
17 siblings, 1 reply; 37+ messages in thread
From: frank.chang @ 2021-12-29 2:33 UTC (permalink / raw)
To: qemu-devel
Cc: Frank Chang, Palmer Dabbelt, Alistair Francis, Bin Meng,
qemu-riscv
From: Frank Chang <frank.chang@sifive.com>
Signed-off-by: Frank Chang <frank.chang@sifive.com>
---
target/riscv/cpu.c | 4 ++++
target/riscv/cpu.h | 1 +
target/riscv/cpu_helper.c | 5 ++++-
target/riscv/csr.c | 6 +++++-
target/riscv/translate.c | 2 ++
5 files changed, 16 insertions(+), 2 deletions(-)
diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
index 6ef3314bce..01239620ca 100644
--- a/target/riscv/cpu.c
+++ b/target/riscv/cpu.c
@@ -562,6 +562,10 @@ static void riscv_cpu_realize(DeviceState *dev, Error **errp)
}
set_vext_version(env, vext_version);
}
+ if (cpu->cfg.ext_zve64f && !cpu->cfg.ext_f) {
+ error_setg(errp, "Zve64f extension depends upon RVF.");
+ return;
+ }
if (cpu->cfg.ext_j) {
ext |= RVJ;
}
diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h
index dc10f27093..d7b2db2b2f 100644
--- a/target/riscv/cpu.h
+++ b/target/riscv/cpu.h
@@ -317,6 +317,7 @@ struct RISCVCPU {
bool ext_icsr;
bool ext_zfh;
bool ext_zfhmin;
+ bool ext_zve64f;
char *priv_spec;
char *user_spec;
diff --git a/target/riscv/cpu_helper.c b/target/riscv/cpu_helper.c
index 10f3baba53..52d93a41fd 100644
--- a/target/riscv/cpu_helper.c
+++ b/target/riscv/cpu_helper.c
@@ -69,12 +69,15 @@ static RISCVMXL cpu_get_xl(CPURISCVState *env)
void cpu_get_tb_cpu_state(CPURISCVState *env, target_ulong *pc,
target_ulong *cs_base, uint32_t *pflags)
{
+ CPUState *cs = env_cpu(env);
+ RISCVCPU *cpu = RISCV_CPU(cs);
+
uint32_t flags = 0;
*pc = env->pc;
*cs_base = 0;
- if (riscv_has_ext(env, RVV)) {
+ if (riscv_has_ext(env, RVV) || cpu->cfg.ext_zve64f) {
/*
* If env->vl equals to VLMAX, we can use generic vector operation
* expanders (GVEC) to accerlate the vector operations.
diff --git a/target/riscv/csr.c b/target/riscv/csr.c
index 146447eac5..340b9661a8 100644
--- a/target/riscv/csr.c
+++ b/target/riscv/csr.c
@@ -47,7 +47,11 @@ static RISCVException fs(CPURISCVState *env, int csrno)
static RISCVException vs(CPURISCVState *env, int csrno)
{
- if (env->misa_ext & RVV) {
+ CPUState *cs = env_cpu(env);
+ RISCVCPU *cpu = RISCV_CPU(cs);
+
+ if (env->misa_ext & RVV ||
+ cpu->cfg.ext_zve64f) {
#if !defined(CONFIG_USER_ONLY)
if (!env->debugger && !riscv_cpu_vector_enabled(env)) {
return RISCV_EXCP_ILLEGAL_INST;
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 5df6c0d800..8c3b0168b7 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -78,6 +78,7 @@ typedef struct DisasContext {
bool ext_ifencei;
bool ext_zfh;
bool ext_zfhmin;
+ bool ext_zve64f;
bool hlsx;
/* vector extension */
bool vill;
@@ -705,6 +706,7 @@ static void riscv_tr_init_disas_context(DisasContextBase *dcbase, CPUState *cs)
ctx->ext_ifencei = cpu->cfg.ext_ifencei;
ctx->ext_zfh = cpu->cfg.ext_zfh;
ctx->ext_zfhmin = cpu->cfg.ext_zfhmin;
+ ctx->ext_zve64f = cpu->cfg.ext_zve64f;
ctx->vlen = cpu->cfg.vlen;
ctx->elen = cpu->cfg.elen;
ctx->mstatus_hs_fs = FIELD_EX32(tb_flags, TB_FLAGS, MSTATUS_HS_FS);
--
2.31.1
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [PATCH 02/17] target/riscv: rvv-1.0: Add Zve64f support for configuration insns
2021-12-29 2:33 [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions frank.chang
2021-12-29 2:33 ` [PATCH 01/17] target/riscv: rvv-1.0: Add Zve64f extension into RISC-V frank.chang
@ 2021-12-29 2:33 ` frank.chang
2022-01-17 22:21 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 03/17] target/riscv: rvv-1.0: Add Zve64f support for load and store insns frank.chang
` (15 subsequent siblings)
17 siblings, 1 reply; 37+ messages in thread
From: frank.chang @ 2021-12-29 2:33 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-riscv, Frank Chang, Bin Meng, Richard Henderson,
Palmer Dabbelt, Alistair Francis, LIU Zhiwei
From: Frank Chang <frank.chang@sifive.com>
All Zve* extensions support the vector configuration instructions.
Signed-off-by: Frank Chang <frank.chang@sifive.com>
---
target/riscv/insn_trans/trans_rvv.c.inc | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
index 6c285c958b..5b47729a21 100644
--- a/target/riscv/insn_trans/trans_rvv.c.inc
+++ b/target/riscv/insn_trans/trans_rvv.c.inc
@@ -129,7 +129,8 @@ static bool do_vsetvl(DisasContext *s, int rd, int rs1, TCGv s2)
{
TCGv s1, dst;
- if (!require_rvv(s) || !has_ext(s, RVV)) {
+ if (!require_rvv(s) ||
+ !(has_ext(s, RVV) || s->ext_zve64f)) {
return false;
}
@@ -164,7 +165,8 @@ static bool do_vsetivli(DisasContext *s, int rd, TCGv s1, TCGv s2)
{
TCGv dst;
- if (!require_rvv(s) || !has_ext(s, RVV)) {
+ if (!require_rvv(s) ||
+ !(has_ext(s, RVV) || s->ext_zve64f)) {
return false;
}
--
2.31.1
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [PATCH 03/17] target/riscv: rvv-1.0: Add Zve64f support for load and store insns
2021-12-29 2:33 [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions frank.chang
2021-12-29 2:33 ` [PATCH 01/17] target/riscv: rvv-1.0: Add Zve64f extension into RISC-V frank.chang
2021-12-29 2:33 ` [PATCH 02/17] target/riscv: rvv-1.0: Add Zve64f support for configuration insns frank.chang
@ 2021-12-29 2:33 ` frank.chang
2022-01-17 22:27 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 04/17] target/riscv: rvv-1.0: Add Zve64f support for vmulh variant insns frank.chang
` (14 subsequent siblings)
17 siblings, 1 reply; 37+ messages in thread
From: frank.chang @ 2021-12-29 2:33 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-riscv, Frank Chang, Bin Meng, Richard Henderson,
Palmer Dabbelt, Alistair Francis, LIU Zhiwei
From: Frank Chang <frank.chang@sifive.com>
All Zve* extensions support all vector load and store instructions,
except Zve64* extensions do not support EEW=64 for index values when
XLEN=32.
Signed-off-by: Frank Chang <frank.chang@sifive.com>
---
target/riscv/insn_trans/trans_rvv.c.inc | 17 +++++++++++++----
1 file changed, 13 insertions(+), 4 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
index 5b47729a21..820a3387db 100644
--- a/target/riscv/insn_trans/trans_rvv.c.inc
+++ b/target/riscv/insn_trans/trans_rvv.c.inc
@@ -263,10 +263,19 @@ static bool vext_check_st_index(DisasContext *s, int vd, int vs2, int nf,
uint8_t eew)
{
int8_t emul = eew - s->sew + s->lmul;
- return (emul >= -3 && emul <= 3) &&
- require_align(vs2, emul) &&
- require_align(vd, s->lmul) &&
- require_nf(vd, nf, s->lmul);
+ bool ret = (emul >= -3 && emul <= 3) &&
+ require_align(vs2, emul) &&
+ require_align(vd, s->lmul) &&
+ require_nf(vd, nf, s->lmul);
+#ifdef TARGET_RISCV32
+ /*
+ * All Zve* extensions support all vector load and store instructions,
+ * except Zve64* extensions do not support EEW=64 for index values
+ * when XLEN=32. (Section 18.2)
+ */
+ ret &= (!has_ext(s, RVV) && s->ext_zve64f ? eew != MO_64 : true);
+#endif
+ return ret;
}
/*
--
2.31.1
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [PATCH 04/17] target/riscv: rvv-1.0: Add Zve64f support for vmulh variant insns
2021-12-29 2:33 [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions frank.chang
` (2 preceding siblings ...)
2021-12-29 2:33 ` [PATCH 03/17] target/riscv: rvv-1.0: Add Zve64f support for load and store insns frank.chang
@ 2021-12-29 2:33 ` frank.chang
2022-01-17 22:30 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 05/17] target/riscv: rvv-1.0: Add Zve64f support for vsmul.vv and vsmul.vx insns frank.chang
` (13 subsequent siblings)
17 siblings, 1 reply; 37+ messages in thread
From: frank.chang @ 2021-12-29 2:33 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-riscv, Frank Chang, Bin Meng, Richard Henderson,
Palmer Dabbelt, Alistair Francis, LIU Zhiwei
From: Frank Chang <frank.chang@sifive.com>
All Zve* extensions support all vector integer instructions,
except that the vmulh integer multiply variants that return the
high word of the product (vmulh.vv, vmulh.vx, vmulhu.vv, vmulhu.vx,
vmulhsu.vv, vmulhsu.vx) are not included for EEW=64 in Zve64*.
Signed-off-by: Frank Chang <frank.chang@sifive.com>
---
target/riscv/insn_trans/trans_rvv.c.inc | 39 +++++++++++++++++++++----
1 file changed, 33 insertions(+), 6 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
index 820a3387db..658cfbe10e 100644
--- a/target/riscv/insn_trans/trans_rvv.c.inc
+++ b/target/riscv/insn_trans/trans_rvv.c.inc
@@ -1906,14 +1906,41 @@ GEN_OPIVX_TRANS(vmaxu_vx, opivx_check)
GEN_OPIVX_TRANS(vmax_vx, opivx_check)
/* Vector Single-Width Integer Multiply Instructions */
+
+static bool vmulh_vv_check(DisasContext *s, arg_rmrr *a)
+{
+ /*
+ * All Zve* extensions support all vector integer instructions,
+ * except that the vmulh integer multiply variants
+ * that return the high word of the product
+ * (vmulh.vv, vmulh.vx, vmulhu.vv, vmulhu.vx, vmulhsu.vv, vmulhsu.vx)
+ * are not included for EEW=64 in Zve64*. (Section 18.2)
+ */
+ return opivv_check(s, a) &&
+ (!has_ext(s, RVV) && s->ext_zve64f ? s->sew != MO_64 : true);
+}
+
+static bool vmulh_vx_check(DisasContext *s, arg_rmrr *a)
+{
+ /*
+ * All Zve* extensions support all vector integer instructions,
+ * except that the vmulh integer multiply variants
+ * that return the high word of the product
+ * (vmulh.vv, vmulh.vx, vmulhu.vv, vmulhu.vx, vmulhsu.vv, vmulhsu.vx)
+ * are not included for EEW=64 in Zve64*. (Section 18.2)
+ */
+ return opivx_check(s, a) &&
+ (!has_ext(s, RVV) && s->ext_zve64f ? s->sew != MO_64 : true);
+}
+
GEN_OPIVV_GVEC_TRANS(vmul_vv, mul)
-GEN_OPIVV_TRANS(vmulh_vv, opivv_check)
-GEN_OPIVV_TRANS(vmulhu_vv, opivv_check)
-GEN_OPIVV_TRANS(vmulhsu_vv, opivv_check)
+GEN_OPIVV_TRANS(vmulh_vv, vmulh_vv_check)
+GEN_OPIVV_TRANS(vmulhu_vv, vmulh_vv_check)
+GEN_OPIVV_TRANS(vmulhsu_vv, vmulh_vv_check)
GEN_OPIVX_GVEC_TRANS(vmul_vx, muls)
-GEN_OPIVX_TRANS(vmulh_vx, opivx_check)
-GEN_OPIVX_TRANS(vmulhu_vx, opivx_check)
-GEN_OPIVX_TRANS(vmulhsu_vx, opivx_check)
+GEN_OPIVX_TRANS(vmulh_vx, vmulh_vx_check)
+GEN_OPIVX_TRANS(vmulhu_vx, vmulh_vx_check)
+GEN_OPIVX_TRANS(vmulhsu_vx, vmulh_vx_check)
/* Vector Integer Divide Instructions */
GEN_OPIVV_TRANS(vdivu_vv, opivv_check)
--
2.31.1
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [PATCH 05/17] target/riscv: rvv-1.0: Add Zve64f support for vsmul.vv and vsmul.vx insns
2021-12-29 2:33 [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions frank.chang
` (3 preceding siblings ...)
2021-12-29 2:33 ` [PATCH 04/17] target/riscv: rvv-1.0: Add Zve64f support for vmulh variant insns frank.chang
@ 2021-12-29 2:33 ` frank.chang
2022-01-17 22:31 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 06/17] target/riscv: rvv-1.0: Add Zve64f support for scalar fp insns frank.chang
` (12 subsequent siblings)
17 siblings, 1 reply; 37+ messages in thread
From: frank.chang @ 2021-12-29 2:33 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-riscv, Frank Chang, Bin Meng, Richard Henderson,
Palmer Dabbelt, Alistair Francis, LIU Zhiwei
From: Frank Chang <frank.chang@sifive.com>
All Zve* extensions support all vector fixed-point arithmetic
instructions, except that vsmul.vv and vsmul.vx are not supported
for EEW=64 in Zve64*.
Signed-off-by: Frank Chang <frank.chang@sifive.com>
---
target/riscv/insn_trans/trans_rvv.c.inc | 27 +++++++++++++++++++++++--
1 file changed, 25 insertions(+), 2 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
index 658cfbe10e..a1e403fe86 100644
--- a/target/riscv/insn_trans/trans_rvv.c.inc
+++ b/target/riscv/insn_trans/trans_rvv.c.inc
@@ -2121,8 +2121,31 @@ GEN_OPIVX_TRANS(vasub_vx, opivx_check)
GEN_OPIVX_TRANS(vasubu_vx, opivx_check)
/* Vector Single-Width Fractional Multiply with Rounding and Saturation */
-GEN_OPIVV_TRANS(vsmul_vv, opivv_check)
-GEN_OPIVX_TRANS(vsmul_vx, opivx_check)
+
+static bool vsmul_vv_check(DisasContext *s, arg_rmrr *a)
+{
+ /*
+ * All Zve* extensions support all vector fixed-point arithmetic
+ * instructions, except that vsmul.vv and vsmul.vx are not supported
+ * for EEW=64 in Zve64*. (Section 18.2)
+ */
+ return opivv_check(s, a) &&
+ (!has_ext(s, RVV) && s->ext_zve64f ? s->sew != MO_64 : true);
+}
+
+static bool vsmul_vx_check(DisasContext *s, arg_rmrr *a)
+{
+ /*
+ * All Zve* extensions support all vector fixed-point arithmetic
+ * instructions, except that vsmul.vv and vsmul.vx are not supported
+ * for EEW=64 in Zve64*. (Section 18.2)
+ */
+ return opivx_check(s, a) &&
+ (!has_ext(s, RVV) && s->ext_zve64f ? s->sew != MO_64 : true);
+}
+
+GEN_OPIVV_TRANS(vsmul_vv, vsmul_vv_check)
+GEN_OPIVX_TRANS(vsmul_vx, vsmul_vx_check)
/* Vector Single-Width Scaling Shift Instructions */
GEN_OPIVV_TRANS(vssrl_vv, opivv_check)
--
2.31.1
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [PATCH 06/17] target/riscv: rvv-1.0: Add Zve64f support for scalar fp insns
2021-12-29 2:33 [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions frank.chang
` (4 preceding siblings ...)
2021-12-29 2:33 ` [PATCH 05/17] target/riscv: rvv-1.0: Add Zve64f support for vsmul.vv and vsmul.vx insns frank.chang
@ 2021-12-29 2:33 ` frank.chang
2022-01-17 22:39 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 07/17] target/riscv: rvv-1.0: Add Zve64f support for single-width fp reduction insns frank.chang
` (11 subsequent siblings)
17 siblings, 1 reply; 37+ messages in thread
From: frank.chang @ 2021-12-29 2:33 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-riscv, Frank Chang, Bin Meng, Richard Henderson,
Palmer Dabbelt, Alistair Francis, LIU Zhiwei
From: Frank Chang <frank.chang@sifive.com>
Zve64f extension requires the scalar processor to implement the F
extension and implement all vector floating-point instructions for
floating-point operands with EEW=32 (i.e., no widening floating-point
operations).
Signed-off-by: Frank Chang <frank.chang@sifive.com>
---
target/riscv/insn_trans/trans_rvv.c.inc | 41 +++++++++++++++++++------
1 file changed, 31 insertions(+), 10 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
index a1e403fe86..0aa8b7918f 100644
--- a/target/riscv/insn_trans/trans_rvv.c.inc
+++ b/target/riscv/insn_trans/trans_rvv.c.inc
@@ -66,6 +66,17 @@ static bool require_scale_rvf(DisasContext *s)
}
}
+static bool require_zve64f(DisasContext *s)
+{
+ /* RVV + Zve64f = RVV. */
+ if (has_ext(s, RVV)) {
+ return true;
+ }
+
+ /* Zve64f doesn't support FP64. (Section 18.2) */
+ return s->ext_zve64f ? s->sew <= MO_32 : true;
+}
+
/* Destination vector register group cannot overlap source mask register. */
static bool require_vm(int vm, int vd)
{
@@ -2204,7 +2215,8 @@ static bool opfvv_check(DisasContext *s, arg_rmrr *a)
return require_rvv(s) &&
require_rvf(s) &&
vext_check_isa_ill(s) &&
- vext_check_sss(s, a->rd, a->rs1, a->rs2, a->vm);
+ vext_check_sss(s, a->rd, a->rs1, a->rs2, a->vm) &&
+ require_zve64f(s);
}
/* OPFVV without GVEC IR */
@@ -2284,7 +2296,8 @@ static bool opfvf_check(DisasContext *s, arg_rmrr *a)
return require_rvv(s) &&
require_rvf(s) &&
vext_check_isa_ill(s) &&
- vext_check_ss(s, a->rd, a->rs2, a->vm);
+ vext_check_ss(s, a->rd, a->rs2, a->vm) &&
+ require_zve64f(s);
}
/* OPFVF without GVEC IR */
@@ -2501,7 +2514,8 @@ static bool opfv_check(DisasContext *s, arg_rmr *a)
require_rvf(s) &&
vext_check_isa_ill(s) &&
/* OPFV instructions ignore vs1 check */
- vext_check_ss(s, a->rd, a->rs2, a->vm);
+ vext_check_ss(s, a->rd, a->rs2, a->vm) &&
+ require_zve64f(s);
}
static bool do_opfv(DisasContext *s, arg_rmr *a,
@@ -2566,7 +2580,8 @@ static bool opfvv_cmp_check(DisasContext *s, arg_rmrr *a)
return require_rvv(s) &&
require_rvf(s) &&
vext_check_isa_ill(s) &&
- vext_check_mss(s, a->rd, a->rs1, a->rs2);
+ vext_check_mss(s, a->rd, a->rs1, a->rs2) &&
+ require_zve64f(s);
}
GEN_OPFVV_TRANS(vmfeq_vv, opfvv_cmp_check)
@@ -2579,7 +2594,8 @@ static bool opfvf_cmp_check(DisasContext *s, arg_rmrr *a)
return require_rvv(s) &&
require_rvf(s) &&
vext_check_isa_ill(s) &&
- vext_check_ms(s, a->rd, a->rs2);
+ vext_check_ms(s, a->rd, a->rs2) &&
+ require_zve64f(s);
}
GEN_OPFVF_TRANS(vmfeq_vf, opfvf_cmp_check)
@@ -2600,7 +2616,8 @@ static bool trans_vfmv_v_f(DisasContext *s, arg_vfmv_v_f *a)
if (require_rvv(s) &&
require_rvf(s) &&
vext_check_isa_ill(s) &&
- require_align(a->rd, s->lmul)) {
+ require_align(a->rd, s->lmul) &&
+ require_zve64f(s)) {
gen_set_rm(s, RISCV_FRM_DYN);
TCGv_i64 t1;
@@ -3326,7 +3343,8 @@ static bool trans_vfmv_f_s(DisasContext *s, arg_vfmv_f_s *a)
{
if (require_rvv(s) &&
require_rvf(s) &&
- vext_check_isa_ill(s)) {
+ vext_check_isa_ill(s) &&
+ require_zve64f(s)) {
gen_set_rm(s, RISCV_FRM_DYN);
unsigned int ofs = (8 << s->sew);
@@ -3352,7 +3370,8 @@ static bool trans_vfmv_s_f(DisasContext *s, arg_vfmv_s_f *a)
{
if (require_rvv(s) &&
require_rvf(s) &&
- vext_check_isa_ill(s)) {
+ vext_check_isa_ill(s) &&
+ require_zve64f(s)) {
gen_set_rm(s, RISCV_FRM_DYN);
/* The instructions ignore LMUL and vector register group. */
@@ -3403,13 +3422,15 @@ GEN_OPIVI_TRANS(vslidedown_vi, IMM_ZX, vslidedown_vx, slidedown_check)
static bool fslideup_check(DisasContext *s, arg_rmrr *a)
{
return slideup_check(s, a) &&
- require_rvf(s);
+ require_rvf(s) &&
+ require_zve64f(s);
}
static bool fslidedown_check(DisasContext *s, arg_rmrr *a)
{
return slidedown_check(s, a) &&
- require_rvf(s);
+ require_rvf(s) &&
+ require_zve64f(s);
}
GEN_OPFVF_TRANS(vfslide1up_vf, fslideup_check)
--
2.31.1
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [PATCH 07/17] target/riscv: rvv-1.0: Add Zve64f support for single-width fp reduction insns
2021-12-29 2:33 [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions frank.chang
` (5 preceding siblings ...)
2021-12-29 2:33 ` [PATCH 06/17] target/riscv: rvv-1.0: Add Zve64f support for scalar fp insns frank.chang
@ 2021-12-29 2:33 ` frank.chang
2022-01-17 22:50 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 08/17] target/riscv: rvv-1.0: Add Zve64f support for widening type-convert insns frank.chang
` (10 subsequent siblings)
17 siblings, 1 reply; 37+ messages in thread
From: frank.chang @ 2021-12-29 2:33 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-riscv, Frank Chang, Bin Meng, Richard Henderson,
Palmer Dabbelt, Alistair Francis, LIU Zhiwei
From: Frank Chang <frank.chang@sifive.com>
Vector single-width floating-point reduction operations for EEW=32 are
supported for Zve64f extension.
Signed-off-by: Frank Chang <frank.chang@sifive.com>
---
target/riscv/insn_trans/trans_rvv.c.inc | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
index 0aa8b7918f..d7e288b87f 100644
--- a/target/riscv/insn_trans/trans_rvv.c.inc
+++ b/target/riscv/insn_trans/trans_rvv.c.inc
@@ -2935,7 +2935,8 @@ GEN_OPIVV_WIDEN_TRANS(vwredsumu_vs, reduction_widen_check)
static bool freduction_check(DisasContext *s, arg_rmrr *a)
{
return reduction_check(s, a) &&
- require_rvf(s);
+ require_rvf(s) &&
+ require_zve64f(s);
}
GEN_OPFVV_TRANS(vfredsum_vs, freduction_check)
--
2.31.1
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [PATCH 08/17] target/riscv: rvv-1.0: Add Zve64f support for widening type-convert insns
2021-12-29 2:33 [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions frank.chang
` (6 preceding siblings ...)
2021-12-29 2:33 ` [PATCH 07/17] target/riscv: rvv-1.0: Add Zve64f support for single-width fp reduction insns frank.chang
@ 2021-12-29 2:33 ` frank.chang
2022-01-17 22:51 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 09/17] target/riscv: rvv-1.0: Add Zve64f support for narrowing " frank.chang
` (9 subsequent siblings)
17 siblings, 1 reply; 37+ messages in thread
From: frank.chang @ 2021-12-29 2:33 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-riscv, Frank Chang, Bin Meng, Richard Henderson,
Palmer Dabbelt, Alistair Francis, LIU Zhiwei
From: Frank Chang <frank.chang@sifive.com>
Vector widening conversion instructions are provided to and from all
supported integer EEWs for Zve64f extension.
Signed-off-by: Frank Chang <frank.chang@sifive.com>
---
target/riscv/insn_trans/trans_rvv.c.inc | 32 +++++++++++++++++++------
1 file changed, 25 insertions(+), 7 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
index d7e288b87f..9ca8d502b2 100644
--- a/target/riscv/insn_trans/trans_rvv.c.inc
+++ b/target/riscv/insn_trans/trans_rvv.c.inc
@@ -77,6 +77,17 @@ static bool require_zve64f(DisasContext *s)
return s->ext_zve64f ? s->sew <= MO_32 : true;
}
+static bool require_scale_zve64f(DisasContext *s)
+{
+ /* RVV + Zve64f = RVV. */
+ if (has_ext(s, RVV)) {
+ return true;
+ }
+
+ /* Zve64f doesn't support FP64. (Section 18.2) */
+ return s->ext_zve64f ? s->sew <= MO_16 : true;
+}
+
/* Destination vector register group cannot overlap source mask register. */
static bool require_vm(int vm, int vd)
{
@@ -2331,7 +2342,8 @@ static bool opfvv_widen_check(DisasContext *s, arg_rmrr *a)
require_scale_rvf(s) &&
(s->sew != MO_8) &&
vext_check_isa_ill(s) &&
- vext_check_dss(s, a->rd, a->rs1, a->rs2, a->vm);
+ vext_check_dss(s, a->rd, a->rs1, a->rs2, a->vm) &&
+ require_scale_zve64f(s);
}
/* OPFVV with WIDEN */
@@ -2370,7 +2382,8 @@ static bool opfvf_widen_check(DisasContext *s, arg_rmrr *a)
require_scale_rvf(s) &&
(s->sew != MO_8) &&
vext_check_isa_ill(s) &&
- vext_check_ds(s, a->rd, a->rs2, a->vm);
+ vext_check_ds(s, a->rd, a->rs2, a->vm) &&
+ require_scale_zve64f(s);
}
/* OPFVF with WIDEN */
@@ -2400,7 +2413,8 @@ static bool opfwv_widen_check(DisasContext *s, arg_rmrr *a)
require_scale_rvf(s) &&
(s->sew != MO_8) &&
vext_check_isa_ill(s) &&
- vext_check_dds(s, a->rd, a->rs1, a->rs2, a->vm);
+ vext_check_dds(s, a->rd, a->rs1, a->rs2, a->vm) &&
+ require_scale_zve64f(s);
}
/* WIDEN OPFVV with WIDEN */
@@ -2439,7 +2453,8 @@ static bool opfwf_widen_check(DisasContext *s, arg_rmrr *a)
require_scale_rvf(s) &&
(s->sew != MO_8) &&
vext_check_isa_ill(s) &&
- vext_check_dd(s, a->rd, a->rs2, a->vm);
+ vext_check_dd(s, a->rd, a->rs2, a->vm) &&
+ require_scale_zve64f(s);
}
/* WIDEN OPFVF with WIDEN */
@@ -2698,14 +2713,16 @@ static bool opfv_widen_check(DisasContext *s, arg_rmr *a)
static bool opxfv_widen_check(DisasContext *s, arg_rmr *a)
{
return opfv_widen_check(s, a) &&
- require_rvf(s);
+ require_rvf(s) &&
+ require_zve64f(s);
}
static bool opffv_widen_check(DisasContext *s, arg_rmr *a)
{
return opfv_widen_check(s, a) &&
require_scale_rvf(s) &&
- (s->sew != MO_8);
+ (s->sew != MO_8) &&
+ require_scale_zve64f(s);
}
#define GEN_OPFV_WIDEN_TRANS(NAME, CHECK, HELPER, FRM) \
@@ -2756,7 +2773,8 @@ static bool opfxv_widen_check(DisasContext *s, arg_rmr *a)
require_scale_rvf(s) &&
vext_check_isa_ill(s) &&
/* OPFV widening instructions ignore vs1 check */
- vext_check_ds(s, a->rd, a->rs2, a->vm);
+ vext_check_ds(s, a->rd, a->rs2, a->vm) &&
+ require_scale_zve64f(s);
}
#define GEN_OPFXV_WIDEN_TRANS(NAME) \
--
2.31.1
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [PATCH 09/17] target/riscv: rvv-1.0: Add Zve64f support for narrowing type-convert insns
2021-12-29 2:33 [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions frank.chang
` (7 preceding siblings ...)
2021-12-29 2:33 ` [PATCH 08/17] target/riscv: rvv-1.0: Add Zve64f support for widening type-convert insns frank.chang
@ 2021-12-29 2:33 ` frank.chang
2022-01-17 22:53 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 10/17] target/riscv: rvv-1.0: Allow Zve64f extension to be turned on frank.chang
` (8 subsequent siblings)
17 siblings, 1 reply; 37+ messages in thread
From: frank.chang @ 2021-12-29 2:33 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-riscv, Frank Chang, Bin Meng, Richard Henderson,
Palmer Dabbelt, Alistair Francis, LIU Zhiwei
From: Frank Chang <frank.chang@sifive.com>
Vector narrowing conversion instructions are provided to and from all
supported integer EEWs for Zve64f extension.
Signed-off-by: Frank Chang <frank.chang@sifive.com>
---
target/riscv/insn_trans/trans_rvv.c.inc | 9 ++++++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
index 9ca8d502b2..230c475d6c 100644
--- a/target/riscv/insn_trans/trans_rvv.c.inc
+++ b/target/riscv/insn_trans/trans_rvv.c.inc
@@ -2824,14 +2824,16 @@ static bool opfxv_narrow_check(DisasContext *s, arg_rmr *a)
{
return opfv_narrow_check(s, a) &&
require_rvf(s) &&
- (s->sew != MO_64);
+ (s->sew != MO_64) &&
+ require_zve64f(s);
}
static bool opffv_narrow_check(DisasContext *s, arg_rmr *a)
{
return opfv_narrow_check(s, a) &&
require_scale_rvf(s) &&
- (s->sew != MO_8);
+ (s->sew != MO_8) &&
+ require_scale_zve64f(s);
}
#define GEN_OPFV_NARROW_TRANS(NAME, CHECK, HELPER, FRM) \
@@ -2880,7 +2882,8 @@ static bool opxfv_narrow_check(DisasContext *s, arg_rmr *a)
require_scale_rvf(s) &&
vext_check_isa_ill(s) &&
/* OPFV narrowing instructions ignore vs1 check */
- vext_check_sd(s, a->rd, a->rs2, a->vm);
+ vext_check_sd(s, a->rd, a->rs2, a->vm) &&
+ require_scale_zve64f(s);
}
#define GEN_OPXFV_NARROW_TRANS(NAME, HELPER, FRM) \
--
2.31.1
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [PATCH 10/17] target/riscv: rvv-1.0: Allow Zve64f extension to be turned on
2021-12-29 2:33 [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions frank.chang
` (8 preceding siblings ...)
2021-12-29 2:33 ` [PATCH 09/17] target/riscv: rvv-1.0: Add Zve64f support for narrowing " frank.chang
@ 2021-12-29 2:33 ` frank.chang
2022-01-17 22:53 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 11/17] target/riscv: rvv-1.0: Add Zve32f extension into RISC-V frank.chang
` (7 subsequent siblings)
17 siblings, 1 reply; 37+ messages in thread
From: frank.chang @ 2021-12-29 2:33 UTC (permalink / raw)
To: qemu-devel
Cc: Frank Chang, Palmer Dabbelt, Alistair Francis, Bin Meng,
qemu-riscv
From: Frank Chang <frank.chang@sifive.com>
Signed-off-by: Frank Chang <frank.chang@sifive.com>
---
target/riscv/cpu.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
index 01239620ca..38cd11a8ae 100644
--- a/target/riscv/cpu.c
+++ b/target/riscv/cpu.c
@@ -636,6 +636,7 @@ static Property riscv_cpu_properties[] = {
DEFINE_PROP_BOOL("Zicsr", RISCVCPU, cfg.ext_icsr, true),
DEFINE_PROP_BOOL("Zfh", RISCVCPU, cfg.ext_zfh, false),
DEFINE_PROP_BOOL("Zfhmin", RISCVCPU, cfg.ext_zfhmin, false),
+ DEFINE_PROP_BOOL("Zve64f", RISCVCPU, cfg.ext_zve64f, false),
DEFINE_PROP_BOOL("mmu", RISCVCPU, cfg.mmu, true),
DEFINE_PROP_BOOL("pmp", RISCVCPU, cfg.pmp, true),
--
2.31.1
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [PATCH 11/17] target/riscv: rvv-1.0: Add Zve32f extension into RISC-V
2021-12-29 2:33 [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions frank.chang
` (9 preceding siblings ...)
2021-12-29 2:33 ` [PATCH 10/17] target/riscv: rvv-1.0: Allow Zve64f extension to be turned on frank.chang
@ 2021-12-29 2:33 ` frank.chang
2022-01-17 22:54 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 12/17] target/riscv: rvv-1.0: Add Zve32f support for configuration insns frank.chang
` (6 subsequent siblings)
17 siblings, 1 reply; 37+ messages in thread
From: frank.chang @ 2021-12-29 2:33 UTC (permalink / raw)
To: qemu-devel
Cc: Frank Chang, Palmer Dabbelt, Alistair Francis, Bin Meng,
qemu-riscv
From: Frank Chang <frank.chang@sifive.com>
Signed-off-by: Frank Chang <frank.chang@sifive.com>
---
target/riscv/cpu.c | 4 ++--
target/riscv/cpu.h | 1 +
target/riscv/cpu_helper.c | 2 +-
target/riscv/csr.c | 2 +-
target/riscv/translate.c | 2 ++
5 files changed, 7 insertions(+), 4 deletions(-)
diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
index 38cd11a8ae..5e98860a09 100644
--- a/target/riscv/cpu.c
+++ b/target/riscv/cpu.c
@@ -562,8 +562,8 @@ static void riscv_cpu_realize(DeviceState *dev, Error **errp)
}
set_vext_version(env, vext_version);
}
- if (cpu->cfg.ext_zve64f && !cpu->cfg.ext_f) {
- error_setg(errp, "Zve64f extension depends upon RVF.");
+ if ((cpu->cfg.ext_zve32f || cpu->cfg.ext_zve64f) && !cpu->cfg.ext_f) {
+ error_setg(errp, "Zve32f/Zve64f extension depends upon RVF.");
return;
}
if (cpu->cfg.ext_j) {
diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h
index d7b2db2b2f..3f3b3bb062 100644
--- a/target/riscv/cpu.h
+++ b/target/riscv/cpu.h
@@ -317,6 +317,7 @@ struct RISCVCPU {
bool ext_icsr;
bool ext_zfh;
bool ext_zfhmin;
+ bool ext_zve32f;
bool ext_zve64f;
char *priv_spec;
diff --git a/target/riscv/cpu_helper.c b/target/riscv/cpu_helper.c
index 52d93a41fd..7d8b34cf1a 100644
--- a/target/riscv/cpu_helper.c
+++ b/target/riscv/cpu_helper.c
@@ -77,7 +77,7 @@ void cpu_get_tb_cpu_state(CPURISCVState *env, target_ulong *pc,
*pc = env->pc;
*cs_base = 0;
- if (riscv_has_ext(env, RVV) || cpu->cfg.ext_zve64f) {
+ if (riscv_has_ext(env, RVV) || cpu->cfg.ext_zve32f || cpu->cfg.ext_zve64f) {
/*
* If env->vl equals to VLMAX, we can use generic vector operation
* expanders (GVEC) to accerlate the vector operations.
diff --git a/target/riscv/csr.c b/target/riscv/csr.c
index 340b9661a8..7bd3a5d1af 100644
--- a/target/riscv/csr.c
+++ b/target/riscv/csr.c
@@ -51,7 +51,7 @@ static RISCVException vs(CPURISCVState *env, int csrno)
RISCVCPU *cpu = RISCV_CPU(cs);
if (env->misa_ext & RVV ||
- cpu->cfg.ext_zve64f) {
+ cpu->cfg.ext_zve32f || cpu->cfg.ext_zve64f) {
#if !defined(CONFIG_USER_ONLY)
if (!env->debugger && !riscv_cpu_vector_enabled(env)) {
return RISCV_EXCP_ILLEGAL_INST;
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 8c3b0168b7..3d89a6650d 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -78,6 +78,7 @@ typedef struct DisasContext {
bool ext_ifencei;
bool ext_zfh;
bool ext_zfhmin;
+ bool ext_zve32f;
bool ext_zve64f;
bool hlsx;
/* vector extension */
@@ -706,6 +707,7 @@ static void riscv_tr_init_disas_context(DisasContextBase *dcbase, CPUState *cs)
ctx->ext_ifencei = cpu->cfg.ext_ifencei;
ctx->ext_zfh = cpu->cfg.ext_zfh;
ctx->ext_zfhmin = cpu->cfg.ext_zfhmin;
+ ctx->ext_zve32f = cpu->cfg.ext_zve32f;
ctx->ext_zve64f = cpu->cfg.ext_zve64f;
ctx->vlen = cpu->cfg.vlen;
ctx->elen = cpu->cfg.elen;
--
2.31.1
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [PATCH 12/17] target/riscv: rvv-1.0: Add Zve32f support for configuration insns
2021-12-29 2:33 [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions frank.chang
` (10 preceding siblings ...)
2021-12-29 2:33 ` [PATCH 11/17] target/riscv: rvv-1.0: Add Zve32f extension into RISC-V frank.chang
@ 2021-12-29 2:33 ` frank.chang
2022-01-17 22:54 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 13/17] target/riscv: rvv-1.0: Add Zve32f support for scalar fp insns frank.chang
` (5 subsequent siblings)
17 siblings, 1 reply; 37+ messages in thread
From: frank.chang @ 2021-12-29 2:33 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-riscv, Frank Chang, Bin Meng, Richard Henderson,
Palmer Dabbelt, Alistair Francis, LIU Zhiwei
From: Frank Chang <frank.chang@sifive.com>
All Zve* extensions support the vector configuration instructions.
Signed-off-by: Frank Chang <frank.chang@sifive.com>
---
target/riscv/insn_trans/trans_rvv.c.inc | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
index 230c475d6c..c6280c7b0b 100644
--- a/target/riscv/insn_trans/trans_rvv.c.inc
+++ b/target/riscv/insn_trans/trans_rvv.c.inc
@@ -152,7 +152,7 @@ static bool do_vsetvl(DisasContext *s, int rd, int rs1, TCGv s2)
TCGv s1, dst;
if (!require_rvv(s) ||
- !(has_ext(s, RVV) || s->ext_zve64f)) {
+ !(has_ext(s, RVV) || s->ext_zve32f || s->ext_zve64f)) {
return false;
}
@@ -188,7 +188,7 @@ static bool do_vsetivli(DisasContext *s, int rd, TCGv s1, TCGv s2)
TCGv dst;
if (!require_rvv(s) ||
- !(has_ext(s, RVV) || s->ext_zve64f)) {
+ !(has_ext(s, RVV) || s->ext_zve32f || s->ext_zve64f)) {
return false;
}
--
2.31.1
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [PATCH 13/17] target/riscv: rvv-1.0: Add Zve32f support for scalar fp insns
2021-12-29 2:33 [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions frank.chang
` (11 preceding siblings ...)
2021-12-29 2:33 ` [PATCH 12/17] target/riscv: rvv-1.0: Add Zve32f support for configuration insns frank.chang
@ 2021-12-29 2:33 ` frank.chang
2022-01-17 22:55 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 14/17] target/riscv: rvv-1.0: Add Zve32f support for single-width fp reduction insns frank.chang
` (4 subsequent siblings)
17 siblings, 1 reply; 37+ messages in thread
From: frank.chang @ 2021-12-29 2:33 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-riscv, Frank Chang, Bin Meng, Richard Henderson,
Palmer Dabbelt, Alistair Francis, LIU Zhiwei
From: Frank Chang <frank.chang@sifive.com>
Zve32f extension requires the scalar processor to implement the F
extension and implement all vector floating-point instructions for
floating-point operands with EEW=32 (i.e., no widening floating-point
operations).
Signed-off-by: Frank Chang <frank.chang@sifive.com>
---
target/riscv/insn_trans/trans_rvv.c.inc | 21 +++++++++++++++++++++
1 file changed, 21 insertions(+)
diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
index c6280c7b0b..1f5a75eca7 100644
--- a/target/riscv/insn_trans/trans_rvv.c.inc
+++ b/target/riscv/insn_trans/trans_rvv.c.inc
@@ -66,6 +66,17 @@ static bool require_scale_rvf(DisasContext *s)
}
}
+static bool require_zve32f(DisasContext *s)
+{
+ /* RVV + Zve32f = RVV. */
+ if (has_ext(s, RVV)) {
+ return true;
+ }
+
+ /* Zve32f doesn't support FP64. (Section 18.2) */
+ return s->ext_zve32f ? s->sew <= MO_32 : true;
+}
+
static bool require_zve64f(DisasContext *s)
{
/* RVV + Zve64f = RVV. */
@@ -2227,6 +2238,7 @@ static bool opfvv_check(DisasContext *s, arg_rmrr *a)
require_rvf(s) &&
vext_check_isa_ill(s) &&
vext_check_sss(s, a->rd, a->rs1, a->rs2, a->vm) &&
+ require_zve32f(s) &&
require_zve64f(s);
}
@@ -2308,6 +2320,7 @@ static bool opfvf_check(DisasContext *s, arg_rmrr *a)
require_rvf(s) &&
vext_check_isa_ill(s) &&
vext_check_ss(s, a->rd, a->rs2, a->vm) &&
+ require_zve32f(s) &&
require_zve64f(s);
}
@@ -2530,6 +2543,7 @@ static bool opfv_check(DisasContext *s, arg_rmr *a)
vext_check_isa_ill(s) &&
/* OPFV instructions ignore vs1 check */
vext_check_ss(s, a->rd, a->rs2, a->vm) &&
+ require_zve32f(s) &&
require_zve64f(s);
}
@@ -2596,6 +2610,7 @@ static bool opfvv_cmp_check(DisasContext *s, arg_rmrr *a)
require_rvf(s) &&
vext_check_isa_ill(s) &&
vext_check_mss(s, a->rd, a->rs1, a->rs2) &&
+ require_zve32f(s) &&
require_zve64f(s);
}
@@ -2610,6 +2625,7 @@ static bool opfvf_cmp_check(DisasContext *s, arg_rmrr *a)
require_rvf(s) &&
vext_check_isa_ill(s) &&
vext_check_ms(s, a->rd, a->rs2) &&
+ require_zve32f(s) &&
require_zve64f(s);
}
@@ -2632,6 +2648,7 @@ static bool trans_vfmv_v_f(DisasContext *s, arg_vfmv_v_f *a)
require_rvf(s) &&
vext_check_isa_ill(s) &&
require_align(a->rd, s->lmul) &&
+ require_zve32f(s) &&
require_zve64f(s)) {
gen_set_rm(s, RISCV_FRM_DYN);
@@ -3366,6 +3383,7 @@ static bool trans_vfmv_f_s(DisasContext *s, arg_vfmv_f_s *a)
if (require_rvv(s) &&
require_rvf(s) &&
vext_check_isa_ill(s) &&
+ require_zve32f(s) &&
require_zve64f(s)) {
gen_set_rm(s, RISCV_FRM_DYN);
@@ -3393,6 +3411,7 @@ static bool trans_vfmv_s_f(DisasContext *s, arg_vfmv_s_f *a)
if (require_rvv(s) &&
require_rvf(s) &&
vext_check_isa_ill(s) &&
+ require_zve32f(s) &&
require_zve64f(s)) {
gen_set_rm(s, RISCV_FRM_DYN);
@@ -3445,6 +3464,7 @@ static bool fslideup_check(DisasContext *s, arg_rmrr *a)
{
return slideup_check(s, a) &&
require_rvf(s) &&
+ require_zve32f(s) &&
require_zve64f(s);
}
@@ -3452,6 +3472,7 @@ static bool fslidedown_check(DisasContext *s, arg_rmrr *a)
{
return slidedown_check(s, a) &&
require_rvf(s) &&
+ require_zve32f(s) &&
require_zve64f(s);
}
--
2.31.1
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [PATCH 14/17] target/riscv: rvv-1.0: Add Zve32f support for single-width fp reduction insns
2021-12-29 2:33 [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions frank.chang
` (12 preceding siblings ...)
2021-12-29 2:33 ` [PATCH 13/17] target/riscv: rvv-1.0: Add Zve32f support for scalar fp insns frank.chang
@ 2021-12-29 2:33 ` frank.chang
2022-01-17 22:56 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 15/17] target/riscv: rvv-1.0: Add Zve32f support for widening type-convert insns frank.chang
` (3 subsequent siblings)
17 siblings, 1 reply; 37+ messages in thread
From: frank.chang @ 2021-12-29 2:33 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-riscv, Frank Chang, Bin Meng, Richard Henderson,
Palmer Dabbelt, Alistair Francis, LIU Zhiwei
From: Frank Chang <frank.chang@sifive.com>
Vector single-width floating-point reduction operations for EEW=32 are
supported for Zve32f extension.
Signed-off-by: Frank Chang <frank.chang@sifive.com>
---
target/riscv/insn_trans/trans_rvv.c.inc | 1 +
1 file changed, 1 insertion(+)
diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
index 1f5a75eca7..c3f4dabf36 100644
--- a/target/riscv/insn_trans/trans_rvv.c.inc
+++ b/target/riscv/insn_trans/trans_rvv.c.inc
@@ -2974,6 +2974,7 @@ static bool freduction_check(DisasContext *s, arg_rmrr *a)
{
return reduction_check(s, a) &&
require_rvf(s) &&
+ require_zve32f(s) &&
require_zve64f(s);
}
--
2.31.1
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [PATCH 15/17] target/riscv: rvv-1.0: Add Zve32f support for widening type-convert insns
2021-12-29 2:33 [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions frank.chang
` (13 preceding siblings ...)
2021-12-29 2:33 ` [PATCH 14/17] target/riscv: rvv-1.0: Add Zve32f support for single-width fp reduction insns frank.chang
@ 2021-12-29 2:33 ` frank.chang
2022-01-17 22:56 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 16/17] target/riscv: rvv-1.0: Add Zve32f support for narrowing " frank.chang
` (2 subsequent siblings)
17 siblings, 1 reply; 37+ messages in thread
From: frank.chang @ 2021-12-29 2:33 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-riscv, Frank Chang, Bin Meng, Richard Henderson,
Palmer Dabbelt, Alistair Francis, LIU Zhiwei
From: Frank Chang <frank.chang@sifive.com>
Vector widening conversion instructions are provided to and from all
supported integer EEWs for Zve32f extension.
Signed-off-by: Frank Chang <frank.chang@sifive.com>
---
target/riscv/insn_trans/trans_rvv.c.inc | 18 ++++++++++++++++++
1 file changed, 18 insertions(+)
diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
index c3f4dabf36..da0e501f85 100644
--- a/target/riscv/insn_trans/trans_rvv.c.inc
+++ b/target/riscv/insn_trans/trans_rvv.c.inc
@@ -77,6 +77,17 @@ static bool require_zve32f(DisasContext *s)
return s->ext_zve32f ? s->sew <= MO_32 : true;
}
+static bool require_scale_zve32f(DisasContext *s)
+{
+ /* RVV + Zve32f = RVV. */
+ if (has_ext(s, RVV)) {
+ return true;
+ }
+
+ /* Zve32f doesn't support FP64. (Section 18.2) */
+ return s->ext_zve64f ? s->sew <= MO_16 : true;
+}
+
static bool require_zve64f(DisasContext *s)
{
/* RVV + Zve64f = RVV. */
@@ -2356,6 +2367,7 @@ static bool opfvv_widen_check(DisasContext *s, arg_rmrr *a)
(s->sew != MO_8) &&
vext_check_isa_ill(s) &&
vext_check_dss(s, a->rd, a->rs1, a->rs2, a->vm) &&
+ require_scale_zve32f(s) &&
require_scale_zve64f(s);
}
@@ -2396,6 +2408,7 @@ static bool opfvf_widen_check(DisasContext *s, arg_rmrr *a)
(s->sew != MO_8) &&
vext_check_isa_ill(s) &&
vext_check_ds(s, a->rd, a->rs2, a->vm) &&
+ require_scale_zve32f(s) &&
require_scale_zve64f(s);
}
@@ -2427,6 +2440,7 @@ static bool opfwv_widen_check(DisasContext *s, arg_rmrr *a)
(s->sew != MO_8) &&
vext_check_isa_ill(s) &&
vext_check_dds(s, a->rd, a->rs1, a->rs2, a->vm) &&
+ require_scale_zve32f(s) &&
require_scale_zve64f(s);
}
@@ -2467,6 +2481,7 @@ static bool opfwf_widen_check(DisasContext *s, arg_rmrr *a)
(s->sew != MO_8) &&
vext_check_isa_ill(s) &&
vext_check_dd(s, a->rd, a->rs2, a->vm) &&
+ require_scale_zve32f(s) &&
require_scale_zve64f(s);
}
@@ -2731,6 +2746,7 @@ static bool opxfv_widen_check(DisasContext *s, arg_rmr *a)
{
return opfv_widen_check(s, a) &&
require_rvf(s) &&
+ require_zve32f(s) &&
require_zve64f(s);
}
@@ -2739,6 +2755,7 @@ static bool opffv_widen_check(DisasContext *s, arg_rmr *a)
return opfv_widen_check(s, a) &&
require_scale_rvf(s) &&
(s->sew != MO_8) &&
+ require_scale_zve32f(s) &&
require_scale_zve64f(s);
}
@@ -2791,6 +2808,7 @@ static bool opfxv_widen_check(DisasContext *s, arg_rmr *a)
vext_check_isa_ill(s) &&
/* OPFV widening instructions ignore vs1 check */
vext_check_ds(s, a->rd, a->rs2, a->vm) &&
+ require_scale_zve32f(s) &&
require_scale_zve64f(s);
}
--
2.31.1
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [PATCH 16/17] target/riscv: rvv-1.0: Add Zve32f support for narrowing type-convert insns
2021-12-29 2:33 [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions frank.chang
` (14 preceding siblings ...)
2021-12-29 2:33 ` [PATCH 15/17] target/riscv: rvv-1.0: Add Zve32f support for widening type-convert insns frank.chang
@ 2021-12-29 2:33 ` frank.chang
2022-01-17 22:57 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 17/17] target/riscv: rvv-1.0: Allow Zve32f extension to be turned on frank.chang
2022-01-17 12:55 ` [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions Frank Chang
17 siblings, 1 reply; 37+ messages in thread
From: frank.chang @ 2021-12-29 2:33 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-riscv, Frank Chang, Bin Meng, Richard Henderson,
Palmer Dabbelt, Alistair Francis, LIU Zhiwei
From: Frank Chang <frank.chang@sifive.com>
Vector narrowing conversion instructions are provided to and from all
supported integer EEWs for Zve32f extension.
Signed-off-by: Frank Chang <frank.chang@sifive.com>
---
target/riscv/insn_trans/trans_rvv.c.inc | 3 +++
1 file changed, 3 insertions(+)
diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
index da0e501f85..5214cf08db 100644
--- a/target/riscv/insn_trans/trans_rvv.c.inc
+++ b/target/riscv/insn_trans/trans_rvv.c.inc
@@ -2860,6 +2860,7 @@ static bool opfxv_narrow_check(DisasContext *s, arg_rmr *a)
return opfv_narrow_check(s, a) &&
require_rvf(s) &&
(s->sew != MO_64) &&
+ require_zve32f(s) &&
require_zve64f(s);
}
@@ -2868,6 +2869,7 @@ static bool opffv_narrow_check(DisasContext *s, arg_rmr *a)
return opfv_narrow_check(s, a) &&
require_scale_rvf(s) &&
(s->sew != MO_8) &&
+ require_scale_zve32f(s) &&
require_scale_zve64f(s);
}
@@ -2918,6 +2920,7 @@ static bool opxfv_narrow_check(DisasContext *s, arg_rmr *a)
vext_check_isa_ill(s) &&
/* OPFV narrowing instructions ignore vs1 check */
vext_check_sd(s, a->rd, a->rs2, a->vm) &&
+ require_scale_zve32f(s) &&
require_scale_zve64f(s);
}
--
2.31.1
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [PATCH 17/17] target/riscv: rvv-1.0: Allow Zve32f extension to be turned on
2021-12-29 2:33 [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions frank.chang
` (15 preceding siblings ...)
2021-12-29 2:33 ` [PATCH 16/17] target/riscv: rvv-1.0: Add Zve32f support for narrowing " frank.chang
@ 2021-12-29 2:33 ` frank.chang
2022-01-17 22:57 ` Alistair Francis
2022-01-17 12:55 ` [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions Frank Chang
17 siblings, 1 reply; 37+ messages in thread
From: frank.chang @ 2021-12-29 2:33 UTC (permalink / raw)
To: qemu-devel
Cc: Frank Chang, Palmer Dabbelt, Alistair Francis, Bin Meng,
qemu-riscv
From: Frank Chang <frank.chang@sifive.com>
Signed-off-by: Frank Chang <frank.chang@sifive.com>
---
target/riscv/cpu.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
index 5e98860a09..2b54c64f56 100644
--- a/target/riscv/cpu.c
+++ b/target/riscv/cpu.c
@@ -636,6 +636,7 @@ static Property riscv_cpu_properties[] = {
DEFINE_PROP_BOOL("Zicsr", RISCVCPU, cfg.ext_icsr, true),
DEFINE_PROP_BOOL("Zfh", RISCVCPU, cfg.ext_zfh, false),
DEFINE_PROP_BOOL("Zfhmin", RISCVCPU, cfg.ext_zfhmin, false),
+ DEFINE_PROP_BOOL("Zve32f", RISCVCPU, cfg.ext_zve32f, false),
DEFINE_PROP_BOOL("Zve64f", RISCVCPU, cfg.ext_zve64f, false),
DEFINE_PROP_BOOL("mmu", RISCVCPU, cfg.mmu, true),
DEFINE_PROP_BOOL("pmp", RISCVCPU, cfg.pmp, true),
--
2.31.1
^ permalink raw reply related [flat|nested] 37+ messages in thread
* Re: [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions
2021-12-29 2:33 [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions frank.chang
` (16 preceding siblings ...)
2021-12-29 2:33 ` [PATCH 17/17] target/riscv: rvv-1.0: Allow Zve32f extension to be turned on frank.chang
@ 2022-01-17 12:55 ` Frank Chang
17 siblings, 0 replies; 37+ messages in thread
From: Frank Chang @ 2022-01-17 12:55 UTC (permalink / raw)
To: Frank Chang; +Cc: open list:RISC-V, qemu-devel@nongnu.org Developers
[-- Attachment #1: Type: text/plain, Size: 2897 bytes --]
<frank.chang@sifive.com> 於 2021年12月29日 週三 上午10:35寫道：
> From: Frank Chang <frank.chang@sifive.com>
>
> In RVV v1.0 spec, several Zve* vector extensions for embedded processors
> are defined in Chapter 18.2:
>
> https://github.com/riscv/riscv-v-spec/blob/v1.0/v-spec.adoc#zve-vector-extensions-for-embedded-processors
>
> This patchset implements Zve32f and Zve64f extensions.
>
> The port is available at:
> https://github.com/sifive/qemu/tree/rvv-zve32f-zve64f-upstream
>
> Zve32f can be enabled with -cpu option: Zve32f=true and
> Zve64f can be enabled with -cpu option: Zve64f=true.
> V is not required to be enabled explicitly.
>
> Quote from the inclusion diagram for the six standard vector extensions
> from Nick Knight <nick.knight@sifive.com>:
>
> V
> |
> Zve64d
> |
> Zve64f
> / \
> Zve64x Zve32f
> \ /
> Zve32x
>
> Note: This patchset depends on other patchsets listed in Based-on
> section below so it is not able to be built unless those patchsets
> are applied.
>
> Based-on: <20211229021250.29804-1-frank.chang@sifive.com>
>
> Frank Chang (17):
> target/riscv: rvv-1.0: Add Zve64f extension into RISC-V
> target/riscv: rvv-1.0: Add Zve64f support for configuration insns
> target/riscv: rvv-1.0: Add Zve64f support for load and store insns
> target/riscv: rvv-1.0: Add Zve64f support for vmulh variant insns
> target/riscv: rvv-1.0: Add Zve64f support for vsmul.vv and vsmul.vx
> insns
> target/riscv: rvv-1.0: Add Zve64f support for scalar fp insns
> target/riscv: rvv-1.0: Add Zve64f support for single-width fp
> reduction insns
> target/riscv: rvv-1.0: Add Zve64f support for widening type-convert
> insns
> target/riscv: rvv-1.0: Add Zve64f support for narrowing type-convert
> insns
> target/riscv: rvv-1.0: Allow Zve64f extension to be turned on
> target/riscv: rvv-1.0: Add Zve32f extension into RISC-V
> target/riscv: rvv-1.0: Add Zve32f support for configuration insns
> target/riscv: rvv-1.0: Add Zve32f support for scalar fp insns
> target/riscv: rvv-1.0: Add Zve32f support for single-width fp
> reduction insns
> target/riscv: rvv-1.0: Add Zve32f support for widening type-convert
> insns
> target/riscv: rvv-1.0: Add Zve32f support for narrowing type-convert
> insns
> target/riscv: rvv-1.0: Allow Zve32f extension to be turned on
>
> target/riscv/cpu.c | 6 +
> target/riscv/cpu.h | 2 +
> target/riscv/cpu_helper.c | 5 +-
> target/riscv/csr.c | 6 +-
> target/riscv/insn_trans/trans_rvv.c.inc | 217 ++++++++++++++++++++----
> target/riscv/translate.c | 4 +
> 6 files changed, 203 insertions(+), 37 deletions(-)
>
> --
> 2.31.1
>
>
>
ping
[-- Attachment #2: Type: text/html, Size: 3945 bytes --]
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH 02/17] target/riscv: rvv-1.0: Add Zve64f support for configuration insns
2021-12-29 2:33 ` [PATCH 02/17] target/riscv: rvv-1.0: Add Zve64f support for configuration insns frank.chang
@ 2022-01-17 22:21 ` Alistair Francis
0 siblings, 0 replies; 37+ messages in thread
From: Alistair Francis @ 2022-01-17 22:21 UTC (permalink / raw)
To: Frank Chang
Cc: open list:RISC-V, Bin Meng, Richard Henderson,
qemu-devel@nongnu.org Developers, Palmer Dabbelt,
Alistair Francis, LIU Zhiwei
On Wed, Dec 29, 2021 at 12:36 PM <frank.chang@sifive.com> wrote:
>
> From: Frank Chang <frank.chang@sifive.com>
>
> All Zve* extensions support the vector configuration instructions.
>
> Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/insn_trans/trans_rvv.c.inc | 6 ++++--
> 1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
> index 6c285c958b..5b47729a21 100644
> --- a/target/riscv/insn_trans/trans_rvv.c.inc
> +++ b/target/riscv/insn_trans/trans_rvv.c.inc
> @@ -129,7 +129,8 @@ static bool do_vsetvl(DisasContext *s, int rd, int rs1, TCGv s2)
> {
> TCGv s1, dst;
>
> - if (!require_rvv(s) || !has_ext(s, RVV)) {
> + if (!require_rvv(s) ||
> + !(has_ext(s, RVV) || s->ext_zve64f)) {
> return false;
> }
>
> @@ -164,7 +165,8 @@ static bool do_vsetivli(DisasContext *s, int rd, TCGv s1, TCGv s2)
> {
> TCGv dst;
>
> - if (!require_rvv(s) || !has_ext(s, RVV)) {
> + if (!require_rvv(s) ||
> + !(has_ext(s, RVV) || s->ext_zve64f)) {
> return false;
> }
>
> --
> 2.31.1
>
>
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH 01/17] target/riscv: rvv-1.0: Add Zve64f extension into RISC-V
2021-12-29 2:33 ` [PATCH 01/17] target/riscv: rvv-1.0: Add Zve64f extension into RISC-V frank.chang
@ 2022-01-17 22:23 ` Alistair Francis
0 siblings, 0 replies; 37+ messages in thread
From: Alistair Francis @ 2022-01-17 22:23 UTC (permalink / raw)
To: Frank Chang
Cc: Alistair Francis, Bin Meng, Palmer Dabbelt,
qemu-devel@nongnu.org Developers, open list:RISC-V
On Wed, Dec 29, 2021 at 12:34 PM <frank.chang@sifive.com> wrote:
>
> From: Frank Chang <frank.chang@sifive.com>
>
> Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/cpu.c | 4 ++++
> target/riscv/cpu.h | 1 +
> target/riscv/cpu_helper.c | 5 ++++-
> target/riscv/csr.c | 6 +++++-
> target/riscv/translate.c | 2 ++
> 5 files changed, 16 insertions(+), 2 deletions(-)
>
> diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
> index 6ef3314bce..01239620ca 100644
> --- a/target/riscv/cpu.c
> +++ b/target/riscv/cpu.c
> @@ -562,6 +562,10 @@ static void riscv_cpu_realize(DeviceState *dev, Error **errp)
> }
> set_vext_version(env, vext_version);
> }
> + if (cpu->cfg.ext_zve64f && !cpu->cfg.ext_f) {
> + error_setg(errp, "Zve64f extension depends upon RVF.");
> + return;
> + }
> if (cpu->cfg.ext_j) {
> ext |= RVJ;
> }
> diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h
> index dc10f27093..d7b2db2b2f 100644
> --- a/target/riscv/cpu.h
> +++ b/target/riscv/cpu.h
> @@ -317,6 +317,7 @@ struct RISCVCPU {
> bool ext_icsr;
> bool ext_zfh;
> bool ext_zfhmin;
> + bool ext_zve64f;
>
> char *priv_spec;
> char *user_spec;
> diff --git a/target/riscv/cpu_helper.c b/target/riscv/cpu_helper.c
> index 10f3baba53..52d93a41fd 100644
> --- a/target/riscv/cpu_helper.c
> +++ b/target/riscv/cpu_helper.c
> @@ -69,12 +69,15 @@ static RISCVMXL cpu_get_xl(CPURISCVState *env)
> void cpu_get_tb_cpu_state(CPURISCVState *env, target_ulong *pc,
> target_ulong *cs_base, uint32_t *pflags)
> {
> + CPUState *cs = env_cpu(env);
> + RISCVCPU *cpu = RISCV_CPU(cs);
> +
> uint32_t flags = 0;
>
> *pc = env->pc;
> *cs_base = 0;
>
> - if (riscv_has_ext(env, RVV)) {
> + if (riscv_has_ext(env, RVV) || cpu->cfg.ext_zve64f) {
> /*
> * If env->vl equals to VLMAX, we can use generic vector operation
> * expanders (GVEC) to accerlate the vector operations.
> diff --git a/target/riscv/csr.c b/target/riscv/csr.c
> index 146447eac5..340b9661a8 100644
> --- a/target/riscv/csr.c
> +++ b/target/riscv/csr.c
> @@ -47,7 +47,11 @@ static RISCVException fs(CPURISCVState *env, int csrno)
>
> static RISCVException vs(CPURISCVState *env, int csrno)
> {
> - if (env->misa_ext & RVV) {
> + CPUState *cs = env_cpu(env);
> + RISCVCPU *cpu = RISCV_CPU(cs);
> +
> + if (env->misa_ext & RVV ||
> + cpu->cfg.ext_zve64f) {
> #if !defined(CONFIG_USER_ONLY)
> if (!env->debugger && !riscv_cpu_vector_enabled(env)) {
> return RISCV_EXCP_ILLEGAL_INST;
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index 5df6c0d800..8c3b0168b7 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -78,6 +78,7 @@ typedef struct DisasContext {
> bool ext_ifencei;
> bool ext_zfh;
> bool ext_zfhmin;
> + bool ext_zve64f;
> bool hlsx;
> /* vector extension */
> bool vill;
> @@ -705,6 +706,7 @@ static void riscv_tr_init_disas_context(DisasContextBase *dcbase, CPUState *cs)
> ctx->ext_ifencei = cpu->cfg.ext_ifencei;
> ctx->ext_zfh = cpu->cfg.ext_zfh;
> ctx->ext_zfhmin = cpu->cfg.ext_zfhmin;
> + ctx->ext_zve64f = cpu->cfg.ext_zve64f;
> ctx->vlen = cpu->cfg.vlen;
> ctx->elen = cpu->cfg.elen;
> ctx->mstatus_hs_fs = FIELD_EX32(tb_flags, TB_FLAGS, MSTATUS_HS_FS);
> --
> 2.31.1
>
>
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH 03/17] target/riscv: rvv-1.0: Add Zve64f support for load and store insns
2021-12-29 2:33 ` [PATCH 03/17] target/riscv: rvv-1.0: Add Zve64f support for load and store insns frank.chang
@ 2022-01-17 22:27 ` Alistair Francis
2022-01-18 1:37 ` Frank Chang
0 siblings, 1 reply; 37+ messages in thread
From: Alistair Francis @ 2022-01-17 22:27 UTC (permalink / raw)
To: Frank Chang
Cc: open list:RISC-V, Bin Meng, Richard Henderson,
qemu-devel@nongnu.org Developers, Palmer Dabbelt,
Alistair Francis, LIU Zhiwei
On Wed, Dec 29, 2021 at 12:34 PM <frank.chang@sifive.com> wrote:
>
> From: Frank Chang <frank.chang@sifive.com>
>
> All Zve* extensions support all vector load and store instructions,
> except Zve64* extensions do not support EEW=64 for index values when
> XLEN=32.
>
> Signed-off-by: Frank Chang <frank.chang@sifive.com>
> ---
> target/riscv/insn_trans/trans_rvv.c.inc | 17 +++++++++++++----
> 1 file changed, 13 insertions(+), 4 deletions(-)
>
> diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
> index 5b47729a21..820a3387db 100644
> --- a/target/riscv/insn_trans/trans_rvv.c.inc
> +++ b/target/riscv/insn_trans/trans_rvv.c.inc
> @@ -263,10 +263,19 @@ static bool vext_check_st_index(DisasContext *s, int vd, int vs2, int nf,
> uint8_t eew)
> {
> int8_t emul = eew - s->sew + s->lmul;
> - return (emul >= -3 && emul <= 3) &&
> - require_align(vs2, emul) &&
> - require_align(vd, s->lmul) &&
> - require_nf(vd, nf, s->lmul);
> + bool ret = (emul >= -3 && emul <= 3) &&
> + require_align(vs2, emul) &&
> + require_align(vd, s->lmul) &&
> + require_nf(vd, nf, s->lmul);
> +#ifdef TARGET_RISCV32
Don't use hardcoded macros for this, instead use get_xl()
Alistair
> + /*
> + * All Zve* extensions support all vector load and store instructions,
> + * except Zve64* extensions do not support EEW=64 for index values
> + * when XLEN=32. (Section 18.2)
> + */
> + ret &= (!has_ext(s, RVV) && s->ext_zve64f ? eew != MO_64 : true);
> +#endif
> + return ret;
> }
>
> /*
> --
> 2.31.1
>
>
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH 04/17] target/riscv: rvv-1.0: Add Zve64f support for vmulh variant insns
2021-12-29 2:33 ` [PATCH 04/17] target/riscv: rvv-1.0: Add Zve64f support for vmulh variant insns frank.chang
@ 2022-01-17 22:30 ` Alistair Francis
0 siblings, 0 replies; 37+ messages in thread
From: Alistair Francis @ 2022-01-17 22:30 UTC (permalink / raw)
To: Frank Chang
Cc: open list:RISC-V, Bin Meng, Richard Henderson,
qemu-devel@nongnu.org Developers, Palmer Dabbelt,
Alistair Francis, LIU Zhiwei
On Wed, Dec 29, 2021 at 12:36 PM <frank.chang@sifive.com> wrote:
>
> From: Frank Chang <frank.chang@sifive.com>
>
> All Zve* extensions support all vector integer instructions,
> except that the vmulh integer multiply variants that return the
> high word of the product (vmulh.vv, vmulh.vx, vmulhu.vv, vmulhu.vx,
> vmulhsu.vv, vmulhsu.vx) are not included for EEW=64 in Zve64*.
>
> Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/insn_trans/trans_rvv.c.inc | 39 +++++++++++++++++++++----
> 1 file changed, 33 insertions(+), 6 deletions(-)
>
> diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
> index 820a3387db..658cfbe10e 100644
> --- a/target/riscv/insn_trans/trans_rvv.c.inc
> +++ b/target/riscv/insn_trans/trans_rvv.c.inc
> @@ -1906,14 +1906,41 @@ GEN_OPIVX_TRANS(vmaxu_vx, opivx_check)
> GEN_OPIVX_TRANS(vmax_vx, opivx_check)
>
> /* Vector Single-Width Integer Multiply Instructions */
> +
> +static bool vmulh_vv_check(DisasContext *s, arg_rmrr *a)
> +{
> + /*
> + * All Zve* extensions support all vector integer instructions,
> + * except that the vmulh integer multiply variants
> + * that return the high word of the product
> + * (vmulh.vv, vmulh.vx, vmulhu.vv, vmulhu.vx, vmulhsu.vv, vmulhsu.vx)
> + * are not included for EEW=64 in Zve64*. (Section 18.2)
> + */
> + return opivv_check(s, a) &&
> + (!has_ext(s, RVV) && s->ext_zve64f ? s->sew != MO_64 : true);
> +}
> +
> +static bool vmulh_vx_check(DisasContext *s, arg_rmrr *a)
> +{
> + /*
> + * All Zve* extensions support all vector integer instructions,
> + * except that the vmulh integer multiply variants
> + * that return the high word of the product
> + * (vmulh.vv, vmulh.vx, vmulhu.vv, vmulhu.vx, vmulhsu.vv, vmulhsu.vx)
> + * are not included for EEW=64 in Zve64*. (Section 18.2)
> + */
> + return opivx_check(s, a) &&
> + (!has_ext(s, RVV) && s->ext_zve64f ? s->sew != MO_64 : true);
> +}
> +
> GEN_OPIVV_GVEC_TRANS(vmul_vv, mul)
> -GEN_OPIVV_TRANS(vmulh_vv, opivv_check)
> -GEN_OPIVV_TRANS(vmulhu_vv, opivv_check)
> -GEN_OPIVV_TRANS(vmulhsu_vv, opivv_check)
> +GEN_OPIVV_TRANS(vmulh_vv, vmulh_vv_check)
> +GEN_OPIVV_TRANS(vmulhu_vv, vmulh_vv_check)
> +GEN_OPIVV_TRANS(vmulhsu_vv, vmulh_vv_check)
> GEN_OPIVX_GVEC_TRANS(vmul_vx, muls)
> -GEN_OPIVX_TRANS(vmulh_vx, opivx_check)
> -GEN_OPIVX_TRANS(vmulhu_vx, opivx_check)
> -GEN_OPIVX_TRANS(vmulhsu_vx, opivx_check)
> +GEN_OPIVX_TRANS(vmulh_vx, vmulh_vx_check)
> +GEN_OPIVX_TRANS(vmulhu_vx, vmulh_vx_check)
> +GEN_OPIVX_TRANS(vmulhsu_vx, vmulh_vx_check)
>
> /* Vector Integer Divide Instructions */
> GEN_OPIVV_TRANS(vdivu_vv, opivv_check)
> --
> 2.31.1
>
>
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH 05/17] target/riscv: rvv-1.0: Add Zve64f support for vsmul.vv and vsmul.vx insns
2021-12-29 2:33 ` [PATCH 05/17] target/riscv: rvv-1.0: Add Zve64f support for vsmul.vv and vsmul.vx insns frank.chang
@ 2022-01-17 22:31 ` Alistair Francis
0 siblings, 0 replies; 37+ messages in thread
From: Alistair Francis @ 2022-01-17 22:31 UTC (permalink / raw)
To: Frank Chang
Cc: open list:RISC-V, Bin Meng, Richard Henderson,
qemu-devel@nongnu.org Developers, Palmer Dabbelt,
Alistair Francis, LIU Zhiwei
On Wed, Dec 29, 2021 at 12:37 PM <frank.chang@sifive.com> wrote:
>
> From: Frank Chang <frank.chang@sifive.com>
>
> All Zve* extensions support all vector fixed-point arithmetic
> instructions, except that vsmul.vv and vsmul.vx are not supported
> for EEW=64 in Zve64*.
>
> Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/insn_trans/trans_rvv.c.inc | 27 +++++++++++++++++++++++--
> 1 file changed, 25 insertions(+), 2 deletions(-)
>
> diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
> index 658cfbe10e..a1e403fe86 100644
> --- a/target/riscv/insn_trans/trans_rvv.c.inc
> +++ b/target/riscv/insn_trans/trans_rvv.c.inc
> @@ -2121,8 +2121,31 @@ GEN_OPIVX_TRANS(vasub_vx, opivx_check)
> GEN_OPIVX_TRANS(vasubu_vx, opivx_check)
>
> /* Vector Single-Width Fractional Multiply with Rounding and Saturation */
> -GEN_OPIVV_TRANS(vsmul_vv, opivv_check)
> -GEN_OPIVX_TRANS(vsmul_vx, opivx_check)
> +
> +static bool vsmul_vv_check(DisasContext *s, arg_rmrr *a)
> +{
> + /*
> + * All Zve* extensions support all vector fixed-point arithmetic
> + * instructions, except that vsmul.vv and vsmul.vx are not supported
> + * for EEW=64 in Zve64*. (Section 18.2)
> + */
> + return opivv_check(s, a) &&
> + (!has_ext(s, RVV) && s->ext_zve64f ? s->sew != MO_64 : true);
> +}
> +
> +static bool vsmul_vx_check(DisasContext *s, arg_rmrr *a)
> +{
> + /*
> + * All Zve* extensions support all vector fixed-point arithmetic
> + * instructions, except that vsmul.vv and vsmul.vx are not supported
> + * for EEW=64 in Zve64*. (Section 18.2)
> + */
> + return opivx_check(s, a) &&
> + (!has_ext(s, RVV) && s->ext_zve64f ? s->sew != MO_64 : true);
> +}
> +
> +GEN_OPIVV_TRANS(vsmul_vv, vsmul_vv_check)
> +GEN_OPIVX_TRANS(vsmul_vx, vsmul_vx_check)
>
> /* Vector Single-Width Scaling Shift Instructions */
> GEN_OPIVV_TRANS(vssrl_vv, opivv_check)
> --
> 2.31.1
>
>
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH 06/17] target/riscv: rvv-1.0: Add Zve64f support for scalar fp insns
2021-12-29 2:33 ` [PATCH 06/17] target/riscv: rvv-1.0: Add Zve64f support for scalar fp insns frank.chang
@ 2022-01-17 22:39 ` Alistair Francis
0 siblings, 0 replies; 37+ messages in thread
From: Alistair Francis @ 2022-01-17 22:39 UTC (permalink / raw)
To: Frank Chang
Cc: open list:RISC-V, Bin Meng, Richard Henderson,
qemu-devel@nongnu.org Developers, Palmer Dabbelt,
Alistair Francis, LIU Zhiwei
On Wed, Dec 29, 2021 at 12:43 PM <frank.chang@sifive.com> wrote:
>
> From: Frank Chang <frank.chang@sifive.com>
>
> Zve64f extension requires the scalar processor to implement the F
> extension and implement all vector floating-point instructions for
> floating-point operands with EEW=32 (i.e., no widening floating-point
> operations).
>
> Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/insn_trans/trans_rvv.c.inc | 41 +++++++++++++++++++------
> 1 file changed, 31 insertions(+), 10 deletions(-)
>
> diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
> index a1e403fe86..0aa8b7918f 100644
> --- a/target/riscv/insn_trans/trans_rvv.c.inc
> +++ b/target/riscv/insn_trans/trans_rvv.c.inc
> @@ -66,6 +66,17 @@ static bool require_scale_rvf(DisasContext *s)
> }
> }
>
> +static bool require_zve64f(DisasContext *s)
> +{
> + /* RVV + Zve64f = RVV. */
> + if (has_ext(s, RVV)) {
> + return true;
> + }
> +
> + /* Zve64f doesn't support FP64. (Section 18.2) */
> + return s->ext_zve64f ? s->sew <= MO_32 : true;
> +}
> +
> /* Destination vector register group cannot overlap source mask register. */
> static bool require_vm(int vm, int vd)
> {
> @@ -2204,7 +2215,8 @@ static bool opfvv_check(DisasContext *s, arg_rmrr *a)
> return require_rvv(s) &&
> require_rvf(s) &&
> vext_check_isa_ill(s) &&
> - vext_check_sss(s, a->rd, a->rs1, a->rs2, a->vm);
> + vext_check_sss(s, a->rd, a->rs1, a->rs2, a->vm) &&
> + require_zve64f(s);
> }
>
> /* OPFVV without GVEC IR */
> @@ -2284,7 +2296,8 @@ static bool opfvf_check(DisasContext *s, arg_rmrr *a)
> return require_rvv(s) &&
> require_rvf(s) &&
> vext_check_isa_ill(s) &&
> - vext_check_ss(s, a->rd, a->rs2, a->vm);
> + vext_check_ss(s, a->rd, a->rs2, a->vm) &&
> + require_zve64f(s);
> }
>
> /* OPFVF without GVEC IR */
> @@ -2501,7 +2514,8 @@ static bool opfv_check(DisasContext *s, arg_rmr *a)
> require_rvf(s) &&
> vext_check_isa_ill(s) &&
> /* OPFV instructions ignore vs1 check */
> - vext_check_ss(s, a->rd, a->rs2, a->vm);
> + vext_check_ss(s, a->rd, a->rs2, a->vm) &&
> + require_zve64f(s);
> }
>
> static bool do_opfv(DisasContext *s, arg_rmr *a,
> @@ -2566,7 +2580,8 @@ static bool opfvv_cmp_check(DisasContext *s, arg_rmrr *a)
> return require_rvv(s) &&
> require_rvf(s) &&
> vext_check_isa_ill(s) &&
> - vext_check_mss(s, a->rd, a->rs1, a->rs2);
> + vext_check_mss(s, a->rd, a->rs1, a->rs2) &&
> + require_zve64f(s);
> }
>
> GEN_OPFVV_TRANS(vmfeq_vv, opfvv_cmp_check)
> @@ -2579,7 +2594,8 @@ static bool opfvf_cmp_check(DisasContext *s, arg_rmrr *a)
> return require_rvv(s) &&
> require_rvf(s) &&
> vext_check_isa_ill(s) &&
> - vext_check_ms(s, a->rd, a->rs2);
> + vext_check_ms(s, a->rd, a->rs2) &&
> + require_zve64f(s);
> }
>
> GEN_OPFVF_TRANS(vmfeq_vf, opfvf_cmp_check)
> @@ -2600,7 +2616,8 @@ static bool trans_vfmv_v_f(DisasContext *s, arg_vfmv_v_f *a)
> if (require_rvv(s) &&
> require_rvf(s) &&
> vext_check_isa_ill(s) &&
> - require_align(a->rd, s->lmul)) {
> + require_align(a->rd, s->lmul) &&
> + require_zve64f(s)) {
> gen_set_rm(s, RISCV_FRM_DYN);
>
> TCGv_i64 t1;
> @@ -3326,7 +3343,8 @@ static bool trans_vfmv_f_s(DisasContext *s, arg_vfmv_f_s *a)
> {
> if (require_rvv(s) &&
> require_rvf(s) &&
> - vext_check_isa_ill(s)) {
> + vext_check_isa_ill(s) &&
> + require_zve64f(s)) {
> gen_set_rm(s, RISCV_FRM_DYN);
>
> unsigned int ofs = (8 << s->sew);
> @@ -3352,7 +3370,8 @@ static bool trans_vfmv_s_f(DisasContext *s, arg_vfmv_s_f *a)
> {
> if (require_rvv(s) &&
> require_rvf(s) &&
> - vext_check_isa_ill(s)) {
> + vext_check_isa_ill(s) &&
> + require_zve64f(s)) {
> gen_set_rm(s, RISCV_FRM_DYN);
>
> /* The instructions ignore LMUL and vector register group. */
> @@ -3403,13 +3422,15 @@ GEN_OPIVI_TRANS(vslidedown_vi, IMM_ZX, vslidedown_vx, slidedown_check)
> static bool fslideup_check(DisasContext *s, arg_rmrr *a)
> {
> return slideup_check(s, a) &&
> - require_rvf(s);
> + require_rvf(s) &&
> + require_zve64f(s);
> }
>
> static bool fslidedown_check(DisasContext *s, arg_rmrr *a)
> {
> return slidedown_check(s, a) &&
> - require_rvf(s);
> + require_rvf(s) &&
> + require_zve64f(s);
> }
>
> GEN_OPFVF_TRANS(vfslide1up_vf, fslideup_check)
> --
> 2.31.1
>
>
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH 07/17] target/riscv: rvv-1.0: Add Zve64f support for single-width fp reduction insns
2021-12-29 2:33 ` [PATCH 07/17] target/riscv: rvv-1.0: Add Zve64f support for single-width fp reduction insns frank.chang
@ 2022-01-17 22:50 ` Alistair Francis
0 siblings, 0 replies; 37+ messages in thread
From: Alistair Francis @ 2022-01-17 22:50 UTC (permalink / raw)
To: Frank Chang
Cc: open list:RISC-V, Bin Meng, Richard Henderson,
qemu-devel@nongnu.org Developers, Palmer Dabbelt,
Alistair Francis, LIU Zhiwei
On Wed, Dec 29, 2021 at 12:41 PM <frank.chang@sifive.com> wrote:
>
> From: Frank Chang <frank.chang@sifive.com>
>
> Vector single-width floating-point reduction operations for EEW=32 are
> supported for Zve64f extension.
>
> Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/insn_trans/trans_rvv.c.inc | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
> index 0aa8b7918f..d7e288b87f 100644
> --- a/target/riscv/insn_trans/trans_rvv.c.inc
> +++ b/target/riscv/insn_trans/trans_rvv.c.inc
> @@ -2935,7 +2935,8 @@ GEN_OPIVV_WIDEN_TRANS(vwredsumu_vs, reduction_widen_check)
> static bool freduction_check(DisasContext *s, arg_rmrr *a)
> {
> return reduction_check(s, a) &&
> - require_rvf(s);
> + require_rvf(s) &&
> + require_zve64f(s);
> }
>
> GEN_OPFVV_TRANS(vfredsum_vs, freduction_check)
> --
> 2.31.1
>
>
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH 08/17] target/riscv: rvv-1.0: Add Zve64f support for widening type-convert insns
2021-12-29 2:33 ` [PATCH 08/17] target/riscv: rvv-1.0: Add Zve64f support for widening type-convert insns frank.chang
@ 2022-01-17 22:51 ` Alistair Francis
0 siblings, 0 replies; 37+ messages in thread
From: Alistair Francis @ 2022-01-17 22:51 UTC (permalink / raw)
To: Frank Chang
Cc: open list:RISC-V, Bin Meng, Richard Henderson,
qemu-devel@nongnu.org Developers, Palmer Dabbelt,
Alistair Francis, LIU Zhiwei
On Wed, Dec 29, 2021 at 12:42 PM <frank.chang@sifive.com> wrote:
>
> From: Frank Chang <frank.chang@sifive.com>
>
> Vector widening conversion instructions are provided to and from all
> supported integer EEWs for Zve64f extension.
>
> Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/insn_trans/trans_rvv.c.inc | 32 +++++++++++++++++++------
> 1 file changed, 25 insertions(+), 7 deletions(-)
>
> diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
> index d7e288b87f..9ca8d502b2 100644
> --- a/target/riscv/insn_trans/trans_rvv.c.inc
> +++ b/target/riscv/insn_trans/trans_rvv.c.inc
> @@ -77,6 +77,17 @@ static bool require_zve64f(DisasContext *s)
> return s->ext_zve64f ? s->sew <= MO_32 : true;
> }
>
> +static bool require_scale_zve64f(DisasContext *s)
> +{
> + /* RVV + Zve64f = RVV. */
> + if (has_ext(s, RVV)) {
> + return true;
> + }
> +
> + /* Zve64f doesn't support FP64. (Section 18.2) */
> + return s->ext_zve64f ? s->sew <= MO_16 : true;
> +}
> +
> /* Destination vector register group cannot overlap source mask register. */
> static bool require_vm(int vm, int vd)
> {
> @@ -2331,7 +2342,8 @@ static bool opfvv_widen_check(DisasContext *s, arg_rmrr *a)
> require_scale_rvf(s) &&
> (s->sew != MO_8) &&
> vext_check_isa_ill(s) &&
> - vext_check_dss(s, a->rd, a->rs1, a->rs2, a->vm);
> + vext_check_dss(s, a->rd, a->rs1, a->rs2, a->vm) &&
> + require_scale_zve64f(s);
> }
>
> /* OPFVV with WIDEN */
> @@ -2370,7 +2382,8 @@ static bool opfvf_widen_check(DisasContext *s, arg_rmrr *a)
> require_scale_rvf(s) &&
> (s->sew != MO_8) &&
> vext_check_isa_ill(s) &&
> - vext_check_ds(s, a->rd, a->rs2, a->vm);
> + vext_check_ds(s, a->rd, a->rs2, a->vm) &&
> + require_scale_zve64f(s);
> }
>
> /* OPFVF with WIDEN */
> @@ -2400,7 +2413,8 @@ static bool opfwv_widen_check(DisasContext *s, arg_rmrr *a)
> require_scale_rvf(s) &&
> (s->sew != MO_8) &&
> vext_check_isa_ill(s) &&
> - vext_check_dds(s, a->rd, a->rs1, a->rs2, a->vm);
> + vext_check_dds(s, a->rd, a->rs1, a->rs2, a->vm) &&
> + require_scale_zve64f(s);
> }
>
> /* WIDEN OPFVV with WIDEN */
> @@ -2439,7 +2453,8 @@ static bool opfwf_widen_check(DisasContext *s, arg_rmrr *a)
> require_scale_rvf(s) &&
> (s->sew != MO_8) &&
> vext_check_isa_ill(s) &&
> - vext_check_dd(s, a->rd, a->rs2, a->vm);
> + vext_check_dd(s, a->rd, a->rs2, a->vm) &&
> + require_scale_zve64f(s);
> }
>
> /* WIDEN OPFVF with WIDEN */
> @@ -2698,14 +2713,16 @@ static bool opfv_widen_check(DisasContext *s, arg_rmr *a)
> static bool opxfv_widen_check(DisasContext *s, arg_rmr *a)
> {
> return opfv_widen_check(s, a) &&
> - require_rvf(s);
> + require_rvf(s) &&
> + require_zve64f(s);
> }
>
> static bool opffv_widen_check(DisasContext *s, arg_rmr *a)
> {
> return opfv_widen_check(s, a) &&
> require_scale_rvf(s) &&
> - (s->sew != MO_8);
> + (s->sew != MO_8) &&
> + require_scale_zve64f(s);
> }
>
> #define GEN_OPFV_WIDEN_TRANS(NAME, CHECK, HELPER, FRM) \
> @@ -2756,7 +2773,8 @@ static bool opfxv_widen_check(DisasContext *s, arg_rmr *a)
> require_scale_rvf(s) &&
> vext_check_isa_ill(s) &&
> /* OPFV widening instructions ignore vs1 check */
> - vext_check_ds(s, a->rd, a->rs2, a->vm);
> + vext_check_ds(s, a->rd, a->rs2, a->vm) &&
> + require_scale_zve64f(s);
> }
>
> #define GEN_OPFXV_WIDEN_TRANS(NAME) \
> --
> 2.31.1
>
>
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH 09/17] target/riscv: rvv-1.0: Add Zve64f support for narrowing type-convert insns
2021-12-29 2:33 ` [PATCH 09/17] target/riscv: rvv-1.0: Add Zve64f support for narrowing " frank.chang
@ 2022-01-17 22:53 ` Alistair Francis
0 siblings, 0 replies; 37+ messages in thread
From: Alistair Francis @ 2022-01-17 22:53 UTC (permalink / raw)
To: Frank Chang
Cc: open list:RISC-V, Bin Meng, Richard Henderson,
qemu-devel@nongnu.org Developers, Palmer Dabbelt,
Alistair Francis, LIU Zhiwei
On Wed, Dec 29, 2021 at 12:45 PM <frank.chang@sifive.com> wrote:
>
> From: Frank Chang <frank.chang@sifive.com>
>
> Vector narrowing conversion instructions are provided to and from all
> supported integer EEWs for Zve64f extension.
>
> Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/insn_trans/trans_rvv.c.inc | 9 ++++++---
> 1 file changed, 6 insertions(+), 3 deletions(-)
>
> diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
> index 9ca8d502b2..230c475d6c 100644
> --- a/target/riscv/insn_trans/trans_rvv.c.inc
> +++ b/target/riscv/insn_trans/trans_rvv.c.inc
> @@ -2824,14 +2824,16 @@ static bool opfxv_narrow_check(DisasContext *s, arg_rmr *a)
> {
> return opfv_narrow_check(s, a) &&
> require_rvf(s) &&
> - (s->sew != MO_64);
> + (s->sew != MO_64) &&
> + require_zve64f(s);
> }
>
> static bool opffv_narrow_check(DisasContext *s, arg_rmr *a)
> {
> return opfv_narrow_check(s, a) &&
> require_scale_rvf(s) &&
> - (s->sew != MO_8);
> + (s->sew != MO_8) &&
> + require_scale_zve64f(s);
> }
>
> #define GEN_OPFV_NARROW_TRANS(NAME, CHECK, HELPER, FRM) \
> @@ -2880,7 +2882,8 @@ static bool opxfv_narrow_check(DisasContext *s, arg_rmr *a)
> require_scale_rvf(s) &&
> vext_check_isa_ill(s) &&
> /* OPFV narrowing instructions ignore vs1 check */
> - vext_check_sd(s, a->rd, a->rs2, a->vm);
> + vext_check_sd(s, a->rd, a->rs2, a->vm) &&
> + require_scale_zve64f(s);
> }
>
> #define GEN_OPXFV_NARROW_TRANS(NAME, HELPER, FRM) \
> --
> 2.31.1
>
>
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH 10/17] target/riscv: rvv-1.0: Allow Zve64f extension to be turned on
2021-12-29 2:33 ` [PATCH 10/17] target/riscv: rvv-1.0: Allow Zve64f extension to be turned on frank.chang
@ 2022-01-17 22:53 ` Alistair Francis
0 siblings, 0 replies; 37+ messages in thread
From: Alistair Francis @ 2022-01-17 22:53 UTC (permalink / raw)
To: Frank Chang
Cc: Alistair Francis, Bin Meng, Palmer Dabbelt,
qemu-devel@nongnu.org Developers, open list:RISC-V
On Wed, Dec 29, 2021 at 12:44 PM <frank.chang@sifive.com> wrote:
>
> From: Frank Chang <frank.chang@sifive.com>
>
> Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/cpu.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
> index 01239620ca..38cd11a8ae 100644
> --- a/target/riscv/cpu.c
> +++ b/target/riscv/cpu.c
> @@ -636,6 +636,7 @@ static Property riscv_cpu_properties[] = {
> DEFINE_PROP_BOOL("Zicsr", RISCVCPU, cfg.ext_icsr, true),
> DEFINE_PROP_BOOL("Zfh", RISCVCPU, cfg.ext_zfh, false),
> DEFINE_PROP_BOOL("Zfhmin", RISCVCPU, cfg.ext_zfhmin, false),
> + DEFINE_PROP_BOOL("Zve64f", RISCVCPU, cfg.ext_zve64f, false),
> DEFINE_PROP_BOOL("mmu", RISCVCPU, cfg.mmu, true),
> DEFINE_PROP_BOOL("pmp", RISCVCPU, cfg.pmp, true),
>
> --
> 2.31.1
>
>
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH 11/17] target/riscv: rvv-1.0: Add Zve32f extension into RISC-V
2021-12-29 2:33 ` [PATCH 11/17] target/riscv: rvv-1.0: Add Zve32f extension into RISC-V frank.chang
@ 2022-01-17 22:54 ` Alistair Francis
0 siblings, 0 replies; 37+ messages in thread
From: Alistair Francis @ 2022-01-17 22:54 UTC (permalink / raw)
To: Frank Chang
Cc: Alistair Francis, Bin Meng, Palmer Dabbelt,
qemu-devel@nongnu.org Developers, open list:RISC-V
On Wed, Dec 29, 2021 at 12:34 PM <frank.chang@sifive.com> wrote:
>
> From: Frank Chang <frank.chang@sifive.com>
>
> Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/cpu.c | 4 ++--
> target/riscv/cpu.h | 1 +
> target/riscv/cpu_helper.c | 2 +-
> target/riscv/csr.c | 2 +-
> target/riscv/translate.c | 2 ++
> 5 files changed, 7 insertions(+), 4 deletions(-)
>
> diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
> index 38cd11a8ae..5e98860a09 100644
> --- a/target/riscv/cpu.c
> +++ b/target/riscv/cpu.c
> @@ -562,8 +562,8 @@ static void riscv_cpu_realize(DeviceState *dev, Error **errp)
> }
> set_vext_version(env, vext_version);
> }
> - if (cpu->cfg.ext_zve64f && !cpu->cfg.ext_f) {
> - error_setg(errp, "Zve64f extension depends upon RVF.");
> + if ((cpu->cfg.ext_zve32f || cpu->cfg.ext_zve64f) && !cpu->cfg.ext_f) {
> + error_setg(errp, "Zve32f/Zve64f extension depends upon RVF.");
> return;
> }
> if (cpu->cfg.ext_j) {
> diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h
> index d7b2db2b2f..3f3b3bb062 100644
> --- a/target/riscv/cpu.h
> +++ b/target/riscv/cpu.h
> @@ -317,6 +317,7 @@ struct RISCVCPU {
> bool ext_icsr;
> bool ext_zfh;
> bool ext_zfhmin;
> + bool ext_zve32f;
> bool ext_zve64f;
>
> char *priv_spec;
> diff --git a/target/riscv/cpu_helper.c b/target/riscv/cpu_helper.c
> index 52d93a41fd..7d8b34cf1a 100644
> --- a/target/riscv/cpu_helper.c
> +++ b/target/riscv/cpu_helper.c
> @@ -77,7 +77,7 @@ void cpu_get_tb_cpu_state(CPURISCVState *env, target_ulong *pc,
> *pc = env->pc;
> *cs_base = 0;
>
> - if (riscv_has_ext(env, RVV) || cpu->cfg.ext_zve64f) {
> + if (riscv_has_ext(env, RVV) || cpu->cfg.ext_zve32f || cpu->cfg.ext_zve64f) {
> /*
> * If env->vl equals to VLMAX, we can use generic vector operation
> * expanders (GVEC) to accerlate the vector operations.
> diff --git a/target/riscv/csr.c b/target/riscv/csr.c
> index 340b9661a8..7bd3a5d1af 100644
> --- a/target/riscv/csr.c
> +++ b/target/riscv/csr.c
> @@ -51,7 +51,7 @@ static RISCVException vs(CPURISCVState *env, int csrno)
> RISCVCPU *cpu = RISCV_CPU(cs);
>
> if (env->misa_ext & RVV ||
> - cpu->cfg.ext_zve64f) {
> + cpu->cfg.ext_zve32f || cpu->cfg.ext_zve64f) {
> #if !defined(CONFIG_USER_ONLY)
> if (!env->debugger && !riscv_cpu_vector_enabled(env)) {
> return RISCV_EXCP_ILLEGAL_INST;
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index 8c3b0168b7..3d89a6650d 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -78,6 +78,7 @@ typedef struct DisasContext {
> bool ext_ifencei;
> bool ext_zfh;
> bool ext_zfhmin;
> + bool ext_zve32f;
> bool ext_zve64f;
> bool hlsx;
> /* vector extension */
> @@ -706,6 +707,7 @@ static void riscv_tr_init_disas_context(DisasContextBase *dcbase, CPUState *cs)
> ctx->ext_ifencei = cpu->cfg.ext_ifencei;
> ctx->ext_zfh = cpu->cfg.ext_zfh;
> ctx->ext_zfhmin = cpu->cfg.ext_zfhmin;
> + ctx->ext_zve32f = cpu->cfg.ext_zve32f;
> ctx->ext_zve64f = cpu->cfg.ext_zve64f;
> ctx->vlen = cpu->cfg.vlen;
> ctx->elen = cpu->cfg.elen;
> --
> 2.31.1
>
>
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH 12/17] target/riscv: rvv-1.0: Add Zve32f support for configuration insns
2021-12-29 2:33 ` [PATCH 12/17] target/riscv: rvv-1.0: Add Zve32f support for configuration insns frank.chang
@ 2022-01-17 22:54 ` Alistair Francis
0 siblings, 0 replies; 37+ messages in thread
From: Alistair Francis @ 2022-01-17 22:54 UTC (permalink / raw)
To: Frank Chang
Cc: open list:RISC-V, Bin Meng, Richard Henderson,
qemu-devel@nongnu.org Developers, Palmer Dabbelt,
Alistair Francis, LIU Zhiwei
On Wed, Dec 29, 2021 at 12:39 PM <frank.chang@sifive.com> wrote:
>
> From: Frank Chang <frank.chang@sifive.com>
>
> All Zve* extensions support the vector configuration instructions.
>
> Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/insn_trans/trans_rvv.c.inc | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
> index 230c475d6c..c6280c7b0b 100644
> --- a/target/riscv/insn_trans/trans_rvv.c.inc
> +++ b/target/riscv/insn_trans/trans_rvv.c.inc
> @@ -152,7 +152,7 @@ static bool do_vsetvl(DisasContext *s, int rd, int rs1, TCGv s2)
> TCGv s1, dst;
>
> if (!require_rvv(s) ||
> - !(has_ext(s, RVV) || s->ext_zve64f)) {
> + !(has_ext(s, RVV) || s->ext_zve32f || s->ext_zve64f)) {
> return false;
> }
>
> @@ -188,7 +188,7 @@ static bool do_vsetivli(DisasContext *s, int rd, TCGv s1, TCGv s2)
> TCGv dst;
>
> if (!require_rvv(s) ||
> - !(has_ext(s, RVV) || s->ext_zve64f)) {
> + !(has_ext(s, RVV) || s->ext_zve32f || s->ext_zve64f)) {
> return false;
> }
>
> --
> 2.31.1
>
>
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH 13/17] target/riscv: rvv-1.0: Add Zve32f support for scalar fp insns
2021-12-29 2:33 ` [PATCH 13/17] target/riscv: rvv-1.0: Add Zve32f support for scalar fp insns frank.chang
@ 2022-01-17 22:55 ` Alistair Francis
0 siblings, 0 replies; 37+ messages in thread
From: Alistair Francis @ 2022-01-17 22:55 UTC (permalink / raw)
To: Frank Chang
Cc: open list:RISC-V, Bin Meng, Richard Henderson,
qemu-devel@nongnu.org Developers, Palmer Dabbelt,
Alistair Francis, LIU Zhiwei
On Wed, Dec 29, 2021 at 12:48 PM <frank.chang@sifive.com> wrote:
>
> From: Frank Chang <frank.chang@sifive.com>
>
> Zve32f extension requires the scalar processor to implement the F
> extension and implement all vector floating-point instructions for
> floating-point operands with EEW=32 (i.e., no widening floating-point
> operations).
>
> Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/insn_trans/trans_rvv.c.inc | 21 +++++++++++++++++++++
> 1 file changed, 21 insertions(+)
>
> diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
> index c6280c7b0b..1f5a75eca7 100644
> --- a/target/riscv/insn_trans/trans_rvv.c.inc
> +++ b/target/riscv/insn_trans/trans_rvv.c.inc
> @@ -66,6 +66,17 @@ static bool require_scale_rvf(DisasContext *s)
> }
> }
>
> +static bool require_zve32f(DisasContext *s)
> +{
> + /* RVV + Zve32f = RVV. */
> + if (has_ext(s, RVV)) {
> + return true;
> + }
> +
> + /* Zve32f doesn't support FP64. (Section 18.2) */
> + return s->ext_zve32f ? s->sew <= MO_32 : true;
> +}
> +
> static bool require_zve64f(DisasContext *s)
> {
> /* RVV + Zve64f = RVV. */
> @@ -2227,6 +2238,7 @@ static bool opfvv_check(DisasContext *s, arg_rmrr *a)
> require_rvf(s) &&
> vext_check_isa_ill(s) &&
> vext_check_sss(s, a->rd, a->rs1, a->rs2, a->vm) &&
> + require_zve32f(s) &&
> require_zve64f(s);
> }
>
> @@ -2308,6 +2320,7 @@ static bool opfvf_check(DisasContext *s, arg_rmrr *a)
> require_rvf(s) &&
> vext_check_isa_ill(s) &&
> vext_check_ss(s, a->rd, a->rs2, a->vm) &&
> + require_zve32f(s) &&
> require_zve64f(s);
> }
>
> @@ -2530,6 +2543,7 @@ static bool opfv_check(DisasContext *s, arg_rmr *a)
> vext_check_isa_ill(s) &&
> /* OPFV instructions ignore vs1 check */
> vext_check_ss(s, a->rd, a->rs2, a->vm) &&
> + require_zve32f(s) &&
> require_zve64f(s);
> }
>
> @@ -2596,6 +2610,7 @@ static bool opfvv_cmp_check(DisasContext *s, arg_rmrr *a)
> require_rvf(s) &&
> vext_check_isa_ill(s) &&
> vext_check_mss(s, a->rd, a->rs1, a->rs2) &&
> + require_zve32f(s) &&
> require_zve64f(s);
> }
>
> @@ -2610,6 +2625,7 @@ static bool opfvf_cmp_check(DisasContext *s, arg_rmrr *a)
> require_rvf(s) &&
> vext_check_isa_ill(s) &&
> vext_check_ms(s, a->rd, a->rs2) &&
> + require_zve32f(s) &&
> require_zve64f(s);
> }
>
> @@ -2632,6 +2648,7 @@ static bool trans_vfmv_v_f(DisasContext *s, arg_vfmv_v_f *a)
> require_rvf(s) &&
> vext_check_isa_ill(s) &&
> require_align(a->rd, s->lmul) &&
> + require_zve32f(s) &&
> require_zve64f(s)) {
> gen_set_rm(s, RISCV_FRM_DYN);
>
> @@ -3366,6 +3383,7 @@ static bool trans_vfmv_f_s(DisasContext *s, arg_vfmv_f_s *a)
> if (require_rvv(s) &&
> require_rvf(s) &&
> vext_check_isa_ill(s) &&
> + require_zve32f(s) &&
> require_zve64f(s)) {
> gen_set_rm(s, RISCV_FRM_DYN);
>
> @@ -3393,6 +3411,7 @@ static bool trans_vfmv_s_f(DisasContext *s, arg_vfmv_s_f *a)
> if (require_rvv(s) &&
> require_rvf(s) &&
> vext_check_isa_ill(s) &&
> + require_zve32f(s) &&
> require_zve64f(s)) {
> gen_set_rm(s, RISCV_FRM_DYN);
>
> @@ -3445,6 +3464,7 @@ static bool fslideup_check(DisasContext *s, arg_rmrr *a)
> {
> return slideup_check(s, a) &&
> require_rvf(s) &&
> + require_zve32f(s) &&
> require_zve64f(s);
> }
>
> @@ -3452,6 +3472,7 @@ static bool fslidedown_check(DisasContext *s, arg_rmrr *a)
> {
> return slidedown_check(s, a) &&
> require_rvf(s) &&
> + require_zve32f(s) &&
> require_zve64f(s);
> }
>
> --
> 2.31.1
>
>
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH 14/17] target/riscv: rvv-1.0: Add Zve32f support for single-width fp reduction insns
2021-12-29 2:33 ` [PATCH 14/17] target/riscv: rvv-1.0: Add Zve32f support for single-width fp reduction insns frank.chang
@ 2022-01-17 22:56 ` Alistair Francis
0 siblings, 0 replies; 37+ messages in thread
From: Alistair Francis @ 2022-01-17 22:56 UTC (permalink / raw)
To: Frank Chang
Cc: open list:RISC-V, Bin Meng, Richard Henderson,
qemu-devel@nongnu.org Developers, Palmer Dabbelt,
Alistair Francis, LIU Zhiwei
On Wed, Dec 29, 2021 at 12:48 PM <frank.chang@sifive.com> wrote:
>
> From: Frank Chang <frank.chang@sifive.com>
>
> Vector single-width floating-point reduction operations for EEW=32 are
> supported for Zve32f extension.
>
> Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/insn_trans/trans_rvv.c.inc | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
> index 1f5a75eca7..c3f4dabf36 100644
> --- a/target/riscv/insn_trans/trans_rvv.c.inc
> +++ b/target/riscv/insn_trans/trans_rvv.c.inc
> @@ -2974,6 +2974,7 @@ static bool freduction_check(DisasContext *s, arg_rmrr *a)
> {
> return reduction_check(s, a) &&
> require_rvf(s) &&
> + require_zve32f(s) &&
> require_zve64f(s);
> }
>
> --
> 2.31.1
>
>
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH 15/17] target/riscv: rvv-1.0: Add Zve32f support for widening type-convert insns
2021-12-29 2:33 ` [PATCH 15/17] target/riscv: rvv-1.0: Add Zve32f support for widening type-convert insns frank.chang
@ 2022-01-17 22:56 ` Alistair Francis
0 siblings, 0 replies; 37+ messages in thread
From: Alistair Francis @ 2022-01-17 22:56 UTC (permalink / raw)
To: Frank Chang
Cc: open list:RISC-V, Bin Meng, Richard Henderson,
qemu-devel@nongnu.org Developers, Palmer Dabbelt,
Alistair Francis, LIU Zhiwei
On Wed, Dec 29, 2021 at 12:50 PM <frank.chang@sifive.com> wrote:
>
> From: Frank Chang <frank.chang@sifive.com>
>
> Vector widening conversion instructions are provided to and from all
> supported integer EEWs for Zve32f extension.
>
> Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/insn_trans/trans_rvv.c.inc | 18 ++++++++++++++++++
> 1 file changed, 18 insertions(+)
>
> diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
> index c3f4dabf36..da0e501f85 100644
> --- a/target/riscv/insn_trans/trans_rvv.c.inc
> +++ b/target/riscv/insn_trans/trans_rvv.c.inc
> @@ -77,6 +77,17 @@ static bool require_zve32f(DisasContext *s)
> return s->ext_zve32f ? s->sew <= MO_32 : true;
> }
>
> +static bool require_scale_zve32f(DisasContext *s)
> +{
> + /* RVV + Zve32f = RVV. */
> + if (has_ext(s, RVV)) {
> + return true;
> + }
> +
> + /* Zve32f doesn't support FP64. (Section 18.2) */
> + return s->ext_zve64f ? s->sew <= MO_16 : true;
> +}
> +
> static bool require_zve64f(DisasContext *s)
> {
> /* RVV + Zve64f = RVV. */
> @@ -2356,6 +2367,7 @@ static bool opfvv_widen_check(DisasContext *s, arg_rmrr *a)
> (s->sew != MO_8) &&
> vext_check_isa_ill(s) &&
> vext_check_dss(s, a->rd, a->rs1, a->rs2, a->vm) &&
> + require_scale_zve32f(s) &&
> require_scale_zve64f(s);
> }
>
> @@ -2396,6 +2408,7 @@ static bool opfvf_widen_check(DisasContext *s, arg_rmrr *a)
> (s->sew != MO_8) &&
> vext_check_isa_ill(s) &&
> vext_check_ds(s, a->rd, a->rs2, a->vm) &&
> + require_scale_zve32f(s) &&
> require_scale_zve64f(s);
> }
>
> @@ -2427,6 +2440,7 @@ static bool opfwv_widen_check(DisasContext *s, arg_rmrr *a)
> (s->sew != MO_8) &&
> vext_check_isa_ill(s) &&
> vext_check_dds(s, a->rd, a->rs1, a->rs2, a->vm) &&
> + require_scale_zve32f(s) &&
> require_scale_zve64f(s);
> }
>
> @@ -2467,6 +2481,7 @@ static bool opfwf_widen_check(DisasContext *s, arg_rmrr *a)
> (s->sew != MO_8) &&
> vext_check_isa_ill(s) &&
> vext_check_dd(s, a->rd, a->rs2, a->vm) &&
> + require_scale_zve32f(s) &&
> require_scale_zve64f(s);
> }
>
> @@ -2731,6 +2746,7 @@ static bool opxfv_widen_check(DisasContext *s, arg_rmr *a)
> {
> return opfv_widen_check(s, a) &&
> require_rvf(s) &&
> + require_zve32f(s) &&
> require_zve64f(s);
> }
>
> @@ -2739,6 +2755,7 @@ static bool opffv_widen_check(DisasContext *s, arg_rmr *a)
> return opfv_widen_check(s, a) &&
> require_scale_rvf(s) &&
> (s->sew != MO_8) &&
> + require_scale_zve32f(s) &&
> require_scale_zve64f(s);
> }
>
> @@ -2791,6 +2808,7 @@ static bool opfxv_widen_check(DisasContext *s, arg_rmr *a)
> vext_check_isa_ill(s) &&
> /* OPFV widening instructions ignore vs1 check */
> vext_check_ds(s, a->rd, a->rs2, a->vm) &&
> + require_scale_zve32f(s) &&
> require_scale_zve64f(s);
> }
>
> --
> 2.31.1
>
>
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH 16/17] target/riscv: rvv-1.0: Add Zve32f support for narrowing type-convert insns
2021-12-29 2:33 ` [PATCH 16/17] target/riscv: rvv-1.0: Add Zve32f support for narrowing " frank.chang
@ 2022-01-17 22:57 ` Alistair Francis
0 siblings, 0 replies; 37+ messages in thread
From: Alistair Francis @ 2022-01-17 22:57 UTC (permalink / raw)
To: Frank Chang
Cc: open list:RISC-V, Bin Meng, Richard Henderson,
qemu-devel@nongnu.org Developers, Palmer Dabbelt,
Alistair Francis, LIU Zhiwei
On Wed, Dec 29, 2021 at 12:46 PM <frank.chang@sifive.com> wrote:
>
> From: Frank Chang <frank.chang@sifive.com>
>
> Vector narrowing conversion instructions are provided to and from all
> supported integer EEWs for Zve32f extension.
>
> Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/insn_trans/trans_rvv.c.inc | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc
> index da0e501f85..5214cf08db 100644
> --- a/target/riscv/insn_trans/trans_rvv.c.inc
> +++ b/target/riscv/insn_trans/trans_rvv.c.inc
> @@ -2860,6 +2860,7 @@ static bool opfxv_narrow_check(DisasContext *s, arg_rmr *a)
> return opfv_narrow_check(s, a) &&
> require_rvf(s) &&
> (s->sew != MO_64) &&
> + require_zve32f(s) &&
> require_zve64f(s);
> }
>
> @@ -2868,6 +2869,7 @@ static bool opffv_narrow_check(DisasContext *s, arg_rmr *a)
> return opfv_narrow_check(s, a) &&
> require_scale_rvf(s) &&
> (s->sew != MO_8) &&
> + require_scale_zve32f(s) &&
> require_scale_zve64f(s);
> }
>
> @@ -2918,6 +2920,7 @@ static bool opxfv_narrow_check(DisasContext *s, arg_rmr *a)
> vext_check_isa_ill(s) &&
> /* OPFV narrowing instructions ignore vs1 check */
> vext_check_sd(s, a->rd, a->rs2, a->vm) &&
> + require_scale_zve32f(s) &&
> require_scale_zve64f(s);
> }
>
> --
> 2.31.1
>
>
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH 17/17] target/riscv: rvv-1.0: Allow Zve32f extension to be turned on
2021-12-29 2:33 ` [PATCH 17/17] target/riscv: rvv-1.0: Allow Zve32f extension to be turned on frank.chang
@ 2022-01-17 22:57 ` Alistair Francis
0 siblings, 0 replies; 37+ messages in thread
From: Alistair Francis @ 2022-01-17 22:57 UTC (permalink / raw)
To: Frank Chang
Cc: Alistair Francis, Bin Meng, Palmer Dabbelt,
qemu-devel@nongnu.org Developers, open list:RISC-V
On Wed, Dec 29, 2021 at 12:52 PM <frank.chang@sifive.com> wrote:
>
> From: Frank Chang <frank.chang@sifive.com>
>
> Signed-off-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/cpu.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
> index 5e98860a09..2b54c64f56 100644
> --- a/target/riscv/cpu.c
> +++ b/target/riscv/cpu.c
> @@ -636,6 +636,7 @@ static Property riscv_cpu_properties[] = {
> DEFINE_PROP_BOOL("Zicsr", RISCVCPU, cfg.ext_icsr, true),
> DEFINE_PROP_BOOL("Zfh", RISCVCPU, cfg.ext_zfh, false),
> DEFINE_PROP_BOOL("Zfhmin", RISCVCPU, cfg.ext_zfhmin, false),
> + DEFINE_PROP_BOOL("Zve32f", RISCVCPU, cfg.ext_zve32f, false),
> DEFINE_PROP_BOOL("Zve64f", RISCVCPU, cfg.ext_zve64f, false),
> DEFINE_PROP_BOOL("mmu", RISCVCPU, cfg.mmu, true),
> DEFINE_PROP_BOOL("pmp", RISCVCPU, cfg.pmp, true),
> --
> 2.31.1
>
>
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH 03/17] target/riscv: rvv-1.0: Add Zve64f support for load and store insns
2022-01-17 22:27 ` Alistair Francis
@ 2022-01-18 1:37 ` Frank Chang
0 siblings, 0 replies; 37+ messages in thread
From: Frank Chang @ 2022-01-18 1:37 UTC (permalink / raw)
To: Alistair Francis
Cc: open list:RISC-V, Bin Meng, Richard Henderson,
qemu-devel@nongnu.org Developers, Palmer Dabbelt,
Alistair Francis, LIU Zhiwei
[-- Attachment #1: Type: text/plain, Size: 1955 bytes --]
On Tue, Jan 18, 2022 at 6:27 AM Alistair Francis <alistair23@gmail.com>
wrote:
> On Wed, Dec 29, 2021 at 12:34 PM <frank.chang@sifive.com> wrote:
> >
> > From: Frank Chang <frank.chang@sifive.com>
> >
> > All Zve* extensions support all vector load and store instructions,
> > except Zve64* extensions do not support EEW=64 for index values when
> > XLEN=32.
> >
> > Signed-off-by: Frank Chang <frank.chang@sifive.com>
> > ---
> > target/riscv/insn_trans/trans_rvv.c.inc | 17 +++++++++++++----
> > 1 file changed, 13 insertions(+), 4 deletions(-)
> >
> > diff --git a/target/riscv/insn_trans/trans_rvv.c.inc
> b/target/riscv/insn_trans/trans_rvv.c.inc
> > index 5b47729a21..820a3387db 100644
> > --- a/target/riscv/insn_trans/trans_rvv.c.inc
> > +++ b/target/riscv/insn_trans/trans_rvv.c.inc
> > @@ -263,10 +263,19 @@ static bool vext_check_st_index(DisasContext *s,
> int vd, int vs2, int nf,
> > uint8_t eew)
> > {
> > int8_t emul = eew - s->sew + s->lmul;
> > - return (emul >= -3 && emul <= 3) &&
> > - require_align(vs2, emul) &&
> > - require_align(vd, s->lmul) &&
> > - require_nf(vd, nf, s->lmul);
> > + bool ret = (emul >= -3 && emul <= 3) &&
> > + require_align(vs2, emul) &&
> > + require_align(vd, s->lmul) &&
> > + require_nf(vd, nf, s->lmul);
> > +#ifdef TARGET_RISCV32
>
> Don't use hardcoded macros for this, instead use get_xl()
>
> Alistair
>
Thanks for the review.
I'll fix it in my next revision patchset.
Regards,
Frank Chang
>
> > + /*
> > + * All Zve* extensions support all vector load and store
> instructions,
> > + * except Zve64* extensions do not support EEW=64 for index values
> > + * when XLEN=32. (Section 18.2)
> > + */
> > + ret &= (!has_ext(s, RVV) && s->ext_zve64f ? eew != MO_64 : true);
> > +#endif
> > + return ret;
> > }
> >
> > /*
> > --
> > 2.31.1
> >
> >
>
[-- Attachment #2: Type: text/html, Size: 3106 bytes --]
^ permalink raw reply [flat|nested] 37+ messages in thread
end of thread, other threads:[~2022-01-18 1:39 UTC | newest]
Thread overview: 37+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2021-12-29 2:33 [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions frank.chang
2021-12-29 2:33 ` [PATCH 01/17] target/riscv: rvv-1.0: Add Zve64f extension into RISC-V frank.chang
2022-01-17 22:23 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 02/17] target/riscv: rvv-1.0: Add Zve64f support for configuration insns frank.chang
2022-01-17 22:21 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 03/17] target/riscv: rvv-1.0: Add Zve64f support for load and store insns frank.chang
2022-01-17 22:27 ` Alistair Francis
2022-01-18 1:37 ` Frank Chang
2021-12-29 2:33 ` [PATCH 04/17] target/riscv: rvv-1.0: Add Zve64f support for vmulh variant insns frank.chang
2022-01-17 22:30 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 05/17] target/riscv: rvv-1.0: Add Zve64f support for vsmul.vv and vsmul.vx insns frank.chang
2022-01-17 22:31 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 06/17] target/riscv: rvv-1.0: Add Zve64f support for scalar fp insns frank.chang
2022-01-17 22:39 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 07/17] target/riscv: rvv-1.0: Add Zve64f support for single-width fp reduction insns frank.chang
2022-01-17 22:50 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 08/17] target/riscv: rvv-1.0: Add Zve64f support for widening type-convert insns frank.chang
2022-01-17 22:51 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 09/17] target/riscv: rvv-1.0: Add Zve64f support for narrowing " frank.chang
2022-01-17 22:53 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 10/17] target/riscv: rvv-1.0: Allow Zve64f extension to be turned on frank.chang
2022-01-17 22:53 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 11/17] target/riscv: rvv-1.0: Add Zve32f extension into RISC-V frank.chang
2022-01-17 22:54 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 12/17] target/riscv: rvv-1.0: Add Zve32f support for configuration insns frank.chang
2022-01-17 22:54 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 13/17] target/riscv: rvv-1.0: Add Zve32f support for scalar fp insns frank.chang
2022-01-17 22:55 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 14/17] target/riscv: rvv-1.0: Add Zve32f support for single-width fp reduction insns frank.chang
2022-01-17 22:56 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 15/17] target/riscv: rvv-1.0: Add Zve32f support for widening type-convert insns frank.chang
2022-01-17 22:56 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 16/17] target/riscv: rvv-1.0: Add Zve32f support for narrowing " frank.chang
2022-01-17 22:57 ` Alistair Francis
2021-12-29 2:33 ` [PATCH 17/17] target/riscv: rvv-1.0: Allow Zve32f extension to be turned on frank.chang
2022-01-17 22:57 ` Alistair Francis
2022-01-17 12:55 ` [PATCH 00/17] Add RISC-V RVV Zve32f and Zve64f extensions Frank Chang
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).