Add RISC-V vector cryptographic instruction set support
Subsystems (4)
cover Add RISC-V vector cryptographic instruction set support
This patchset provides an implementation for Zvkb, Zvkned, Zvknh, Zvksh, Zvkg, and Zvksed of the draft RISC-V vector cryptography extensions as per the 20230303 version of the specification(1) (1fcbb30). Please note that the Zvkt data-independent execution latency extension has not been implemented, and we would recommend not using these patches in an environment where timing attacks are an issue.
Work performed by Dickon, Lawrence, Nazar, Kiran, and William from Codethink sponsored by SiFive, as well as Max Chou and Frank Chang from SiFive.
For convenience we have created a git repo with our patches on top of a recent master. https://github.com/CodethinkLabs/qemu-ct
1. https://github.com/riscv/riscv-crypto/releases
Dickon Hood (2):
qemu/bitops.h: Limit rotate amounts
target/riscv: Add vrol.[vv,vx] and vror.[vv,vx,vi] decoding,
translation and execution support
Kiran Ostrolenk (8):
target/riscv: Refactor some of the generic vector functionality target/riscv: Refactor some of the generic vector functionality target/riscv: Refactor some of the generic vector functionality target/riscv: Refactor some of the generic vector functionality target/riscv: Add vsha2ms.vv decoding, translation and execution
support
target/riscv: Add zvksh cpu property
target/riscv: Add vsm3c.vi decoding, translation and execution support target/riscv: Expose zvksh cpu property
Lawrence Hunter (17):
target/riscv: Add vclmul.vv decoding, translation and execution
support
target/riscv: Add vclmul.vx decoding, translation and execution
support
target/riscv: Add vclmulh.vv decoding, translation and execution
support
target/riscv: Add vclmulh.vx decoding, translation and execution
support
target/riscv: Add vaesef.vv decoding, translation and execution
support
target/riscv: Add vaesef.vs decoding, translation and execution
support
target/riscv: Add vaesdf.vv decoding, translation and execution
support
target/riscv: Add vaesdf.vs decoding, translation and execution
support
target/riscv: Add vaesdm.vv decoding, translation and execution
support
target/riscv: Add vaesdm.vs decoding, translation and execution
support
target/riscv: Add vaesz.vs decoding, translation and execution support target/riscv: Add vsha2c[hl].vv decoding, translation and execution
support
target/riscv: Add vsm3me.vv decoding, translation and execution
support
target/riscv: Add zvkg cpu property
target/riscv: Add vgmul.vv decoding, translation and execution support target/riscv: Add vghsh.vv decoding, translation and execution support target/riscv: Expose zvkg cpu property
Max Chou (5):
crypto: Create sm4_subword
crypto: Add SM4 constant parameter CK
target/riscv: Add zvksed cfg property
target/riscv: Add Zvksed support
target/riscv: Expose Zvksed property
Nazar Kazakov (10):
target/riscv: Add zvkb cpu property
target/riscv: Add vrev8.v decoding, translation and execution support target/riscv: Add vandn.[vv,vx] decoding, translation and execution
support
target/riscv: Expose zvkb cpu property
target/riscv: Add zvkned cpu property
target/riscv: Add vaeskf1.vi decoding, translation and execution
support
target/riscv: Add vaeskf2.vi decoding, translation and execution
support
target/riscv: Expose zvkned cpu property
target/riscv: Add zvknh cpu properties
target/riscv: Expose zvknh cpu properties
William Salmon (3):
target/riscv: Add vbrev8.v decoding, translation and execution support target/riscv: Add vaesem.vv decoding, translation and execution
support
target/riscv: Add vaesem.vs decoding, translation and execution
support
accel/tcg/tcg-runtime-gvec.c | 11 + accel/tcg/tcg-runtime.h | 1 + crypto/sm4.c | 10 + include/crypto/sm4.h | 9 + include/qemu/bitops.h | 24 +- target/arm/tcg/crypto_helper.c | 10 +- target/riscv/cpu.c | 36 + target/riscv/cpu.h | 7 + target/riscv/helper.h | 71 ++ target/riscv/insn32.decode | 49 + target/riscv/insn_trans/trans_rvv.c.inc | 93 +- target/riscv/insn_trans/trans_rvzvkb.c.inc | 220 ++++ target/riscv/insn_trans/trans_rvzvkg.c.inc | 40 + target/riscv/insn_trans/trans_rvzvkned.c.inc | 170 +++ target/riscv/insn_trans/trans_rvzvknh.c.inc | 84 ++ target/riscv/insn_trans/trans_rvzvksed.c.inc | 57 + target/riscv/insn_trans/trans_rvzvksh.c.inc | 43 + target/riscv/meson.build | 4 +- target/riscv/op_helper.c | 5 + target/riscv/translate.c | 6 + target/riscv/vcrypto_helper.c | 1001 ++++++++++++++++++ target/riscv/vector_helper.c | 240 +---- target/riscv/vector_internals.c | 81 ++ target/riscv/vector_internals.h | 222 ++++ 24 files changed, 2192 insertions(+), 302 deletions(-) create mode 100644 target/riscv/insn_trans/trans_rvzvkb.c.inc create mode 100644 target/riscv/insn_trans/trans_rvzvkg.c.inc create mode 100644 target/riscv/insn_trans/trans_rvzvkned.c.inc create mode 100644 target/riscv/insn_trans/trans_rvzvknh.c.inc create mode 100644 target/riscv/insn_trans/trans_rvzvksed.c.inc create mode 100644 target/riscv/insn_trans/trans_rvzvksh.c.inc create mode 100644 target/riscv/vcrypto_helper.c create mode 100644 target/riscv/vector_internals.c create mode 100644 target/riscv/vector_internals.h
Signature · 1 line
2.39.2
PATCH 1/45 target/riscv: Add zvkb cpu property
From: Nazar Kazakov <[email protected]>
--- target/riscv/cpu.c | 13 +++++++++++++ target/riscv/cpu.h | 1 + 2 files changed, 14 insertions(+)
target/riscv/cpu.c +13 -0
@@ -109,6 +109,7 @@ static const struct isa_ext_data isa_edata_arr[] = {ISA_EXT_DATA_ENTRY (zve64d, true,PRIV_VERSION_1_12_0 , ext_zve64d),ISA_EXT_DATA_ENTRY (zvfh, true,PRIV_VERSION_1_12_0 , ext_zvfh),ISA_EXT_DATA_ENTRY (zvfhmin, true,PRIV_VERSION_1_12_0 , ext_zvfhmin), +ISA_EXT_DATA_ENTRY (zvkb, true,PRIV_VERSION_1_12_0 , ext_zvkb),ISA_EXT_DATA_ENTRY (zhinx, true,PRIV_VERSION_1_12_0 , ext_zhinx),ISA_EXT_DATA_ENTRY (zhinxmin, true,PRIV_VERSION_1_12_0 , ext_zhinxmin), ISA_EXT_DATA_ENTRY(smaia, true, PRIV_VERSION_1_12_0, ext_smaia),
@@ -1211,6 +1212,18 @@ static void riscv_cpu_realize(DeviceState *dev, Error **errp)return ; } +/* +* In principle zve*x would also suffice here, were they supported +* in qemu +*/ +if (cpu->cfg.ext_zvkb && + !(cpu->cfg.ext_zve32f || cpu->cfg.ext_zve64f || + cpu->cfg.ext_zve64d || cpu->cfg.ext_v)) { +error_setg ( + errp,"Vector crypto extensions require V or Zve* extensions" ); +return ; + } +#ifndef CONFIG_USER_ONLY if (cpu->cfg.pmu_num) {if (!riscv_pmu_init (cpu, cpu->cfg.pmu_num) && cpu->cfg.ext_sscofpmf) {
target/riscv/cpu.h +1 -0
@@ -470,6 +470,7 @@ struct RISCVCPUConfig {bool ext_zve32f;bool ext_zve64f;bool ext_zve64d; +bool ext_zvkb;bool ext_zmmul;bool ext_zvfh;bool ext_zvfhmin;
Signature · 1 line
2.39.2
PATCH 2/45 target/riscv: Refactor some of the generic vector functionality
In reply to Lawrence HunterFrom: Kiran Ostrolenk <[email protected]>
Summary of refactoring:
* take some functions/macros out of vector_helper and put them in a new module called vector_internals
* factor the non SEW-specific stuff out of GEN_OPIVV_TRANS into function opivv_trans (similar to opivi_trans)
All this refactoring ensures more functions/macros can be used by both vector and vector-crypto helpers (latter implemented in proceeding commit).
--- target/riscv/insn_trans/trans_rvv.c.inc | 62 ++++++------ target/riscv/meson.build | 1 + target/riscv/vector_helper.c | 120 +----------------------- target/riscv/vector_internals.c | 57 +++++++++++ target/riscv/vector_internals.h | 118 +++++++++++++++++++++++ 5 files changed, 209 insertions(+), 149 deletions(-) create mode 100644 target/riscv/vector_internals.c create mode 100644 target/riscv/vector_internals.h
target/riscv/insn_trans/trans_rvv.c.inc +32 -30
@@ -1643,38 +1643,40 @@ GEN_OPIWX_WIDEN_TRANS(vwadd_wx) GEN_OPIWX_WIDEN_TRANS(vwsubu_wx) GEN_OPIWX_WIDEN_TRANS(vwsub_wx) +static bool opivv_trans(uint32_t vd, uint32_t vs1, uint32_t vs2, uint32_t vm, + gen_helper_gvec_4_ptr *fn, DisasContext *s) +{ + uint32_t data = 0; + TCGLabel *over = gen_new_label(); + tcg_gen_brcondi_tl(TCG_COND_EQ, cpu_vl, 0, over); + tcg_gen_brcond_tl(TCG_COND_GEU, cpu_vstart, cpu_vl, over); + + data = FIELD_DP32(data, VDATA, VM, vm); + data = FIELD_DP32(data, VDATA, LMUL, s->lmul); + data = FIELD_DP32(data, VDATA, VTA, s->vta); + data = FIELD_DP32(data, VDATA, VTA_ALL_1S, s->cfg_vta_all_1s); + data = FIELD_DP32(data, VDATA, VMA, s->vma); + tcg_gen_gvec_4_ptr(vreg_ofs(s, vd), vreg_ofs(s, 0), vreg_ofs(s, vs1), + vreg_ofs(s, vs2), cpu_env, s->cfg_ptr->vlen / 8, + s->cfg_ptr->vlen / 8, data, fn); + mark_vs_dirty(s); + gen_set_label(over); + return true; +} + /* Vector Integer Add-with-Carry / Subtract-with-Borrow Instructions */ /* OPIVV without GVEC IR */ -#define GEN_OPIVV_TRANS(NAME, CHECK) \ -static bool trans_##NAME(DisasContext *s, arg_rmrr *a) \ -{ \ - if (CHECK(s, a)) { \ - uint32_t data = 0; \ - static gen_helper_gvec_4_ptr * const fns[4] = { \ - gen_helper_##NAME##_b, gen_helper_##NAME##_h, \ - gen_helper_##NAME##_w, gen_helper_##NAME##_d, \ - }; \ - TCGLabel *over = gen_new_label(); \ - tcg_gen_brcondi_tl(TCG_COND_EQ, cpu_vl, 0, over); \ - tcg_gen_brcond_tl(TCG_COND_GEU, cpu_vstart, cpu_vl, over); \ - \ - data = FIELD_DP32(data, VDATA, VM, a->vm); \ - data = FIELD_DP32(data, VDATA, LMUL, s->lmul); \ - data = FIELD_DP32(data, VDATA, VTA, s->vta); \ - data = \ - FIELD_DP32(data, VDATA, VTA_ALL_1S, s->cfg_vta_all_1s);\ - data = FIELD_DP32(data, VDATA, VMA, s->vma); \ - tcg_gen_gvec_4_ptr(vreg_ofs(s, a->rd), vreg_ofs(s, 0), \ - vreg_ofs(s, a->rs1), \ - vreg_ofs(s, a->rs2), cpu_env, \ - s->cfg_ptr->vlen / 8, \ - s->cfg_ptr->vlen / 8, data, \ - fns[s->sew]); \ - mark_vs_dirty(s); \ - gen_set_label(over); \ - return true; \ - } \ - return false; \ +#define GEN_OPIVV_TRANS(NAME, CHECK) \ +static bool trans_##NAME(DisasContext *s, arg_rmrr *a) \ +{ \ + if (CHECK(s, a)) { \ + static gen_helper_gvec_4_ptr * const fns[4] = { \ + gen_helper_##NAME##_b, gen_helper_##NAME##_h, \ + gen_helper_##NAME##_w, gen_helper_##NAME##_d, \ + }; \ + return opivv_trans(a->rd, a->rs1, a->rs2, a->vm, fns[s->sew], s);\ + } \ + return false; \ } /*
target/riscv/meson.build +1 -0
@@ -16,6 +16,7 @@ riscv_ss.add(files( 'gdbstub.c', 'op_helper.c', 'vector_helper.c', + 'vector_internals.c', 'bitmanip_helper.c', 'translate.c', 'm128_helper.c',
target/riscv/vector_helper.c +1 -119
162 lines not on this page · raw message
target/riscv/vector_internals.c new file +57 -0
57 lines not on this page · raw message
target/riscv/vector_internals.h new file +118 -0
118 lines not on this page · raw message
Signature · 1 line
2.39.2
PATCH 3/45 target/riscv: Add vclmul.vv decoding, translation and execution support
In reply to Lawrence Hunter--- target/riscv/helper.h | 3 ++ target/riscv/insn32.decode | 3 ++ target/riscv/insn_trans/trans_rvzvkb.c.inc | 40 +++++++++++++++++++++ target/riscv/meson.build | 3 +- target/riscv/translate.c | 1 + target/riscv/vcrypto_helper.c | 42 ++++++++++++++++++++++ 6 files changed, 91 insertions(+), 1 deletion(-) create mode 100644 target/riscv/insn_trans/trans_rvzvkb.c.inc create mode 100644 target/riscv/vcrypto_helper.c
target/riscv/helper.h +3 -0
@@ -1142,3 +1142,6 @@ DEF_HELPER_FLAGS_1(aes64im, TCG_CALL_NO_RWG_SE, tl, tl) DEF_HELPER_FLAGS_3(sm4ed, TCG_CALL_NO_RWG_SE, tl, tl, tl, tl)DEF_HELPER_FLAGS_3 (sm4ks,TCG_CALL_NO_RWG_SE , tl, tl, tl, tl) + +/* Vector crypto functions */ +DEF_HELPER_6 (vclmul_vv, void, ptr, ptr, ptr, ptr, env, i32)
target/riscv/insn32.decode +3 -0
@@ -908,3 +908,6 @@ sm4ks .. 11010 ..... ..... 000 ..... 0110011 @k_aes # *** RV32 Zicond Standard Extension *** czero_eqz 0000111 ..... ..... 101 ..... 0110011 @r czero_nez 0000111 ..... ..... 111 ..... 0110011 @r + +# *** RV64 Zvkb vector crypto extension *** +vclmul_vv 001100 . ..... ..... 010 ..... 1010111 @r_vm
target/riscv/insn_trans/trans_rvzvkb.c.inc new file +40 -0
@@ -0,0 +1,40 @@ +/* + * RISC-V translation routines for the Zvkb Extension. + * + * Copyright (C) 2023 SiFive, Inc. + * Written by Codethink Ltd and SiFive. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2 or later, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program. If not, see <http://www.gnu.org/licenses/>. + */ + +#define GEN_VV_MASKED_TRANS(NAME, CHECK) \ +static bool trans_##NAME(DisasContext *s, arg_rmrr *a) \ +{ \ + if (CHECK(s, a)) { \ + return opivv_trans(a->rd, a->rs1, a->rs2, a->vm, \ + gen_helper_##NAME, s); \ + } \ + return false; \ +} + +static bool zvkb_vv_check(DisasContext *s, arg_rmrr *a) +{ + return opivv_check(s, a) && s->cfg_ptr->ext_zvkb == true; +} + +static bool vclmul_vv_check(DisasContext *s, arg_rmrr *a) +{ + return zvkb_vv_check(s, a) && s->sew == MO_64; +} + +GEN_VV_MASKED_TRANS(vclmul_vv, vclmul_vv_check)
target/riscv/meson.build +2 -1
@@ -20,7 +20,8 @@ riscv_ss.add(files( 'bitmanip_helper.c', 'translate.c', 'm128_helper.c', - 'crypto_helper.c' + 'crypto_helper.c', + 'vcrypto_helper.c' )) riscv_ss.add(when: 'CONFIG_KVM', if_true: files('kvm.c'), if_false: files('kvm-stub.c'))
target/riscv/translate.c +1 -0
@@ -1083,6 +1083,7 @@ static uint32_t opcode_at(DisasContextBase *dcbase, target_ulong pc)#include "insn_trans/trans_rvzicbo.c.inc" #include "insn_trans/trans_rvzfh.c.inc" #include "insn_trans/trans_rvk.c.inc" +#include "insn_trans/trans_rvzvkb.c.inc" #include "insn_trans/trans_privileged.c.inc" #include "insn_trans/trans_svinval.c.inc" #include "decode-xthead.c.inc"
target/riscv/vcrypto_helper.c new file +42 -0
42 lines not on this page · raw message
Signature · 1 line
2.39.2
PATCH 4/45 target/riscv: Refactor some of the generic vector functionality
In reply to Lawrence HunterFrom: Kiran Ostrolenk <[email protected]>
This refactoring ensures these functions/macros can be used by both vector and vector-crypto helpers (latter implemented in proceeding commit).
--- target/riscv/vector_helper.c | 36 --------------------------------- target/riscv/vector_internals.c | 24 ++++++++++++++++++++++ target/riscv/vector_internals.h | 16 +++++++++++++++ 3 files changed, 40 insertions(+), 36 deletions(-)
target/riscv/vector_helper.c +0 -36
@@ -721,8 +721,6 @@ GEN_VEXT_VV(vsub_vv_h, 2) GEN_VEXT_VV(vsub_vv_w, 4)GEN_VEXT_VV (vsub_vv_d,8 ) -typedef void opivx2_fn (void *vd,target_long s1,void *vs2,int i); -/* * (T1)s1 gives the real operator type. * (TX1)(T1)s1 expands the operator type of widen or narrow operations.
@@ -747,40 +745,6 @@ RVVCALL(OPIVX2, vrsub_vx_h, OP_SSS_H, H2, H2, DO_RSUB) RVVCALL(OPIVX2, vrsub_vx_w, OP_SSS_W, H4, H4, DO_RSUB) RVVCALL(OPIVX2, vrsub_vx_d, OP_SSS_D, H8, H8, DO_RSUB) -static void do_vext_vx (void *vd,void *v0,target_long s1,void *vs2, -CPURISCVState *env,uint32_t desc, -opivx2_fn fn,uint32_t esz) -{ -uint32_t vm =vext_vm (desc); -uint32_t vl = env->vl; -uint32_t total_elems =vext_get_total_elems (env, desc, esz); -uint32_t vta =vext_vta (desc); -uint32_t vma =vext_vma (desc); -uint32_t i; - -for (i = env->vstart; i < vl; i++) { -if (!vm && !vext_elem_mask (v0, i)) { -/* set masked-off elements to 1s */ -vext_set_elems_1s (vd, vma, i * esz, (i +1 ) * esz); -continue ; - } -fn (vd, s1, vs2, i); - } - env->vstart =0 ; -/* set tail elements to 1s */ -vext_set_elems_1s (vd, vta, vl * esz, total_elems * esz); -} - -/* generate the helpers for OPIVX */ -#define GEN_VEXT_VX (NAME ,ESZ ) \ -void HELPER(NAME)(void *vd, void *v0, target_ulong s1, \ - void *vs2, CPURISCVState *env, \ - uint32_t desc) \ -{ \ - do_vext_vx(vd, v0, s1, vs2, env, desc, \ - do_##NAME, ESZ); \ -} - GEN_VEXT_VX(vadd_vx_b, 1) GEN_VEXT_VX(vadd_vx_h, 2)GEN_VEXT_VX (vadd_vx_w,4 )
target/riscv/vector_internals.c +24 -0
@@ -55,3 +55,27 @@ void do_vext_vv(void *vd, void *v0, void *vs1, void *vs2,/* set tail elements to 1s */ vext_set_elems_1s (vd, vta, vl * esz, total_elems * esz); } + +void do_vext_vx (void *vd,void *v0,target_long s1,void *vs2, +CPURISCVState *env,uint32_t desc, +opivx2_fn fn,uint32_t esz) +{ +uint32_t vm =vext_vm (desc); +uint32_t vl = env->vl; +uint32_t total_elems =vext_get_total_elems (env, desc, esz); +uint32_t vta =vext_vta (desc); +uint32_t vma =vext_vma (desc); +uint32_t i; + +for (i = env->vstart; i < vl; i++) { +if (!vm && !vext_elem_mask (v0, i)) { +/* set masked-off elements to 1s */ +vext_set_elems_1s (vd, vma, i * esz, (i +1 ) * esz); +continue ; + } +fn (vd, s1, vs2, i); + } + env->vstart =0 ; +/* set tail elements to 1s */ +vext_set_elems_1s (vd, vta, vl * esz, total_elems * esz); +}
target/riscv/vector_internals.h +16 -0
@@ -115,4 +115,20 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, \ do_##NAME, ESZ); \ } +typedef void opivx2_fn (void *vd,target_long s1,void *vs2,int i); + +void do_vext_vx (void *vd,void *v0,target_long s1,void *vs2, +CPURISCVState *env,uint32_t desc, +opivx2_fn fn,uint32_t esz); + +/* generate the helpers for OPIVX */ +#define GEN_VEXT_VX (NAME ,ESZ ) \ +void HELPER(NAME)(void *vd, void *v0, target_ulong s1, \ + void *vs2, CPURISCVState *env, \ + uint32_t desc) \ +{ \ + do_vext_vx(vd, v0, s1, vs2, env, desc, \ + do_##NAME, ESZ); \ +} +#endif /* TARGET_RISCV_VECTOR_INTERNALS_H */
Signature · 1 line
2.39.2
Re: [PATCH 04/45] target/riscv: Refactor some of the generic vector functionality
Quoted 6 lines · Lawrence Hunter
From: Kiran Ostrolenk <[email protected]>
This refactoring ensures these functions/macros can be used by both vector and vector-crypto helpers (latter implemented in proceeding commit).
Quoted 8 lines
Signed-off-by Kiran Ostrolenk <[email protected]>---
target/riscv/vector_helper.c | 36 ------------------------------- --
target/riscv/vector_internals.c | 24 ++++++++++++++++++++++target/riscv/vector_internals.h| 16 +++++++++++++++3 files changed, 40 insertions(+), 36 deletions(-)
Quoted 119 lines
target/riscv/vector_helper.c +0
-0@@ -721,8 +721,6 @@ GEN_VEXT_VV(vsub_vv_h, 2) GEN_VEXT_VV(vsub_vv_w, 4) GEN_VEXT_VV(vsub_vv_d, 8)target/riscv/vector_helper.c +0
-2-typedef void opivx2_fn (void *vd,target_long s1,void *vs2,int i); -/*
* (T1)s1 gives the real operator type.
* (TX1)(T1)s1 expands the operator type of widen or narrow operations.target/riscv/vector_helper.c +0
-0@@ -747,40 +745,6 @@ RVVCALL(OPIVX2, vrsub_vx_h, OP_SSS_H, H2, H2, DO_RSUB) RVVCALL(OPIVX2, vrsub_vx_w, OP_SSS_W, H4, H4, DO_RSUB) RVVCALL(OPIVX2, vrsub_vx_d, OP_SSS_D, H8, H8, DO_RSUB)
-static void do_vext_vx(void *vd, void *v0,target_longs1, void *vs2,
- CPURISCVState *env,uint32_tdesc, -opivx2_fnfn,uint32_tesz)target/riscv/vector_helper.c +0
-31-{ - uint32_t vm = vext_vm(desc); - uint32_t vl = env->vl; - uint32_t total_elems = vext_get_total_elems(env, desc, esz); - uint32_t vta = vext_vta(desc); - uint32_t vma = vext_vma(desc); - uint32_t i; - - for (i = env->vstart; i < vl; i++) { - if (!vm && !vext_elem_mask(v0, i)) { - /* set masked-off elements to 1s */ - vext_set_elems_1s(vd, vma, i * esz, (i + 1) * esz); - continue; - } - fn(vd, s1, vs2, i); - } - env->vstart = 0; - /* set tail elements to 1s */ - vext_set_elems_1s(vd, vta, vl * esz, total_elems * esz); -} - -/* generate the helpers for OPIVX */ -#define GEN_VEXT_VX (NAME ,ESZ ) \ -void HELPER(NAME)(void *vd, void *v0, target_ulong s1, \ - void *vs2, CPURISCVState *env, \ - uint32_t desc) \ -{ \ - do_vext_vx(vd, v0, s1, vs2, env, desc, \ - do_##NAME, ESZ); \ -} -GEN_VEXT_VX(vadd_vx_b, 1)
GEN_VEXT_VX(vadd_vx_h, 2)
GEN_VEXT_VX(vadd_vx_w, 4)target/riscv/vector_internals.c +0
-0@@ -55,3 +55,27 @@ void do_vext_vv(void *vd, void *v0, void *vs1,void *vs2, /* set tail elements to 1s */ vext_set_elems_1s(vd, vta, vl * esz, total_elems * esz);}
target/riscv/vector_internals.c +24
-0+ +void do_vext_vx (void *vd,void *v0,target_long s1,void *vs2, + CPURISCVState *env, uint32_t desc, + opivx2_fn fn, uint32_t esz) +{ + uint32_t vm = vext_vm(desc); + uint32_t vl = env->vl; + uint32_t total_elems = vext_get_total_elems(env, desc, esz); + uint32_t vta = vext_vta(desc); + uint32_t vma = vext_vma(desc); + uint32_t i; + + for (i = env->vstart; i < vl; i++) { + if (!vm && !vext_elem_mask(v0, i)) { + /* set masked-off elements to 1s */ + vext_set_elems_1s(vd, vma, i * esz, (i + 1) * esz); + continue; + } + fn(vd, s1, vs2, i); + } + env->vstart = 0; + /* set tail elements to 1s */ + vext_set_elems_1s(vd, vta, vl * esz, total_elems * esz); +}target/riscv/vector_internals.h +0
-0@@ -115,4 +115,20 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, \ do_##NAME, ESZ); \ }target/riscv/vector_internals.h +16
-0+typedef void opivx2_fn (void *vd,target_long s1,void *vs2,int i); + +void do_vext_vx (void *vd,void *v0,target_long s1,void *vs2, + CPURISCVState *env, uint32_t desc, + opivx2_fn fn, uint32_t esz); + +/* generate the helpers for OPIVX */ +#define GEN_VEXT_VX (NAME ,ESZ ) \ +void HELPER(NAME)(void *vd, void *v0, target_ulong s1, \ + void *vs2, CPURISCVState *env, \ + uint32_t desc) \ +{ \ + do_vext_vx(vd, v0, s1, vs2, env, desc, \ + do_##NAME, ESZ); \ +} +#endif /*
TARGET_RISCV_VECTOR_INTERNALS_H*/
PATCH 5/45 target/riscv: Add vclmul.vx decoding, translation and execution support
In reply to Lawrence Hunter--- target/riscv/helper.h | 1 + target/riscv/insn32.decode | 1 + target/riscv/insn_trans/trans_rvzvkb.c.inc | 48 ++++++++++++++++++++++ target/riscv/vcrypto_helper.c | 2 + 4 files changed, 52 insertions(+)
target/riscv/helper.h +1 -0
@@ -1145,3 +1145,4 @@ DEF_HELPER_FLAGS_3(sm4ks, TCG_CALL_NO_RWG_SE, tl, tl, tl, tl)/* Vector crypto functions */ DEF_HELPER_6(vclmul_vv, void, ptr, ptr, ptr, ptr, env, i32) +DEF_HELPER_6 (vclmul_vx, void, ptr, ptr, tl, ptr, env, i32)
target/riscv/insn32.decode +1 -0
@@ -911,3 +911,4 @@ czero_nez 0000111 ..... ..... 111 ..... 0110011 @r # *** RV64 Zvkb vector crypto extension *** vclmul_vv 001100 . ..... ..... 010 ..... 1010111 @r_vm +vclmul_vx 001100 . ..... ..... 110 ..... 1010111 @r_vm
target/riscv/insn_trans/trans_rvzvkb.c.inc +48 -0
51 lines not on this page · raw message
target/riscv/vcrypto_helper.c +2 -0
@@ -40,3 +40,5 @@ static uint64_t clmul64(uint64_t y, uint64_t x) RVVCALL(OPIVV2, vclmul_vv, OP_UUU_D, H8, H8, H8, clmul64) GEN_VEXT_VV(vclmul_vv, 8) +RVVCALL(OPIVX2, vclmul_vx, OP_UUU_D, H8, H8, clmul64) +GEN_VEXT_VX (vclmul_vx,8 )
Signature · 1 line
2.39.2
PATCH 6/45 target/riscv: Add vclmulh.vv decoding, translation and execution support
In reply to Lawrence Hunter--- target/riscv/helper.h | 1 + target/riscv/insn32.decode | 1 + target/riscv/insn_trans/trans_rvzvkb.c.inc | 1 + target/riscv/vcrypto_helper.c | 13 +++++++++++++ 4 files changed, 16 insertions(+)
target/riscv/helper.h +1 -0
@@ -1146,3 +1146,4 @@ DEF_HELPER_FLAGS_3(sm4ks, TCG_CALL_NO_RWG_SE, tl, tl, tl, tl)/* Vector crypto functions */ DEF_HELPER_6(vclmul_vv, void, ptr, ptr, ptr, ptr, env, i32) DEF_HELPER_6(vclmul_vx, void, ptr, ptr, tl, ptr, env, i32) +DEF_HELPER_6 (vclmulh_vv, void, ptr, ptr, ptr, ptr, env, i32)
target/riscv/insn32.decode +1 -0
@@ -912,3 +912,4 @@ czero_nez 0000111 ..... ..... 111 ..... 0110011 @r # *** RV64 Zvkb vector crypto extension *** vclmul_vv 001100 . ..... ..... 010 ..... 1010111 @r_vm vclmul_vx 001100 . ..... ..... 110 ..... 1010111 @r_vm +vclmulh_vv 001101 . ..... ..... 010 ..... 1010111 @r_vm
target/riscv/insn_trans/trans_rvzvkb.c.inc +1 -0
@@ -38,6 +38,7 @@ static bool vclmul_vv_check(DisasContext *s, arg_rmrr *a) } GEN_VV_MASKED_TRANS(vclmul_vv, vclmul_vv_check) +GEN_VV_MASKED_TRANS(vclmulh_vv, vclmul_vv_check) #define GEN_VX_MASKED_TRANS(NAME, CHECK) \ static bool trans_##NAME(DisasContext *s, arg_rmrr *a) \
target/riscv/vcrypto_helper.c +13 -0
@@ -38,7 +38,20 @@ static uint64_t clmul64(uint64_t y, uint64_t x)return result; } +static uint64_t clmulh64 (uint64_t y,uint64_t x) +{ +uint64_t result =0 ; +for (int j =63 ; j >=1 ; j--) { +if ((y >> j) &1 ) { + result ^= (x >> (64 - j)); + } + } +return result; +} + RVVCALL(OPIVV2, vclmul_vv, OP_UUU_D, H8, H8, H8, clmul64) GEN_VEXT_VV(vclmul_vv, 8) RVVCALL(OPIVX2, vclmul_vx, OP_UUU_D, H8, H8, clmul64) GEN_VEXT_VX(vclmul_vx, 8) +RVVCALL(OPIVV2, vclmulh_vv, OP_UUU_D, H8, H8, H8, clmulh64) +GEN_VEXT_VV (vclmulh_vv,8 )
Signature · 1 line
2.39.2
PATCH 7/45 target/riscv: Add vclmulh.vx decoding, translation and execution support
In reply to Lawrence Hunter--- target/riscv/helper.h | 1 + target/riscv/insn32.decode | 1 + target/riscv/insn_trans/trans_rvzvkb.c.inc | 1 + target/riscv/vcrypto_helper.c | 2 ++ 4 files changed, 5 insertions(+)
target/riscv/helper.h +1 -0
@@ -1147,3 +1147,4 @@ DEF_HELPER_FLAGS_3(sm4ks, TCG_CALL_NO_RWG_SE, tl, tl, tl, tl) DEF_HELPER_6(vclmul_vv, void, ptr, ptr, ptr, ptr, env, i32) DEF_HELPER_6(vclmul_vx, void, ptr, ptr, tl, ptr, env, i32) DEF_HELPER_6(vclmulh_vv, void, ptr, ptr, ptr, ptr, env, i32) +DEF_HELPER_6 (vclmulh_vx, void, ptr, ptr, tl, ptr, env, i32)
target/riscv/insn32.decode +1 -0
@@ -913,3 +913,4 @@ czero_nez 0000111 ..... ..... 111 ..... 0110011 @r vclmul_vv 001100 . ..... ..... 010 ..... 1010111 @r_vm vclmul_vx 001100 . ..... ..... 110 ..... 1010111 @r_vm vclmulh_vv 001101 . ..... ..... 010 ..... 1010111 @r_vm +vclmulh_vx 001101 . ..... ..... 110 ..... 1010111 @r_vm
target/riscv/insn_trans/trans_rvzvkb.c.inc +1 -0
@@ -87,3 +87,4 @@ static bool vclmul_vx_check(DisasContext *s, arg_rmrr *a) } GEN_VX_MASKED_TRANS(vclmul_vx, vclmul_vx_check) +GEN_VX_MASKED_TRANS(vclmulh_vx, vclmul_vx_check)
target/riscv/vcrypto_helper.c +2 -0
@@ -55,3 +55,5 @@ RVVCALL(OPIVX2, vclmul_vx, OP_UUU_D, H8, H8, clmul64) GEN_VEXT_VX(vclmul_vx, 8) RVVCALL(OPIVV2, vclmulh_vv, OP_UUU_D, H8, H8, H8, clmulh64) GEN_VEXT_VV(vclmulh_vv, 8) +RVVCALL(OPIVX2, vclmulh_vx, OP_UUU_D, H8, H8, clmulh64) +GEN_VEXT_VX (vclmulh_vx,8 )
Signature · 1 line
2.39.2
PATCH 8/45 target/riscv: Refactor some of the generic vector functionality
In reply to Lawrence HunterFrom: Kiran Ostrolenk <[email protected]>
This refactoring ensures these functions/macros can be used by both vector and vector-crypto helpers (latter implemented in proceeding commit).
Also moves the checks out of do_opiv{v,x,i}_gvec{,_shift} functions and into the corresponding macros. This enables the functions to be reused in proceeding commit without check duplication.
--- target/riscv/insn_trans/trans_rvv.c.inc | 28 +++++++-------- target/riscv/vector_helper.c | 45 ----------------------- target/riscv/vector_internals.h | 48 +++++++++++++++++++++++++ 3 files changed, 60 insertions(+), 61 deletions(-)
target/riscv/insn_trans/trans_rvv.c.inc +12 -16
@@ -1187,9 +1187,6 @@ do_opivv_gvec(DisasContext *s, arg_rmrr *a, GVecGen3Fn *gvec_fn, gen_helper_gvec_4_ptr *fn) { TCGLabel *over = gen_new_label(); - if (!opivv_check(s, a)) { - return false; - } tcg_gen_brcondi_tl(TCG_COND_EQ, cpu_vl, 0, over); tcg_gen_brcond_tl(TCG_COND_GEU, cpu_vstart, cpu_vl, over);
@@ -1223,6 +1220,9 @@ static bool trans_##NAME(DisasContext *s, arg_rmrr *a) \ gen_helper_##NAME##_b, gen_helper_##NAME##_h, \ gen_helper_##NAME##_w, gen_helper_##NAME##_d, \ }; \ + if (!opivv_check(s, a)) { \ + return false; \ + } \ return do_opivv_gvec(s, a, tcg_gen_gvec_##SUF, fns[s->sew]); \ }
@@ -1282,10 +1282,6 @@ static inline bool do_opivx_gvec(DisasContext *s, arg_rmrr *a, GVecGen2sFn *gvec_fn, gen_helper_opivx *fn) { - if (!opivx_check(s, a)) { - return false; - } - if (a->vm && s->vl_eq_vlmax && !(s->vta && s->lmul < 0)) { TCGv_i64 src1 = tcg_temp_new_i64();
@@ -1307,6 +1303,9 @@ static bool trans_##NAME(DisasContext *s, arg_rmrr *a) \ gen_helper_##NAME##_b, gen_helper_##NAME##_h, \ gen_helper_##NAME##_w, gen_helper_##NAME##_d, \ }; \ + if (!opivx_check(s, a)) { \ + return false; \ + } \ return do_opivx_gvec(s, a, tcg_gen_gvec_##SUF, fns[s->sew]); \ }
@@ -1439,10 +1438,6 @@ static inline bool do_opivi_gvec(DisasContext *s, arg_rmrr *a, GVecGen2iFn *gvec_fn, gen_helper_opivx *fn, imm_mode_t imm_mode) { - if (!opivx_check(s, a)) { - return false; - } - if (a->vm && s->vl_eq_vlmax && !(s->vta && s->lmul < 0)) { gvec_fn(s->sew, vreg_ofs(s, a->rd), vreg_ofs(s, a->rs2), extract_imm(s, a->rs1, imm_mode), MAXSZ(s), MAXSZ(s));
@@ -1460,6 +1455,9 @@ static bool trans_##NAME(DisasContext *s, arg_rmrr *a) \ gen_helper_##OPIVX##_b, gen_helper_##OPIVX##_h, \ gen_helper_##OPIVX##_w, gen_helper_##OPIVX##_d, \ }; \ + if (!opivx_check(s, a)) { \ + return false; \ + } \ return do_opivi_gvec(s, a, tcg_gen_gvec_##SUF, \ fns[s->sew], IMM_MODE); \ }
@@ -1785,10 +1783,6 @@ static inline bool do_opivx_gvec_shift(DisasContext *s, arg_rmrr *a, GVecGen2sFn32 *gvec_fn, gen_helper_opivx *fn) { - if (!opivx_check(s, a)) { - return false; - } - if (a->vm && s->vl_eq_vlmax && !(s->vta && s->lmul < 0)) { TCGv_i32 src1 = tcg_temp_new_i32();
@@ -1810,7 +1804,9 @@ static bool trans_##NAME(DisasContext *s, arg_rmrr *a) \ gen_helper_##NAME##_b, gen_helper_##NAME##_h, \ gen_helper_##NAME##_w, gen_helper_##NAME##_d, \ }; \ - \ + if (!opivx_check(s, a)) { \ + return false; \ + } \ return do_opivx_gvec_shift(s, a, tcg_gen_gvec_##SUF, fns[s->sew]); \ }
target/riscv/vector_helper.c +0 -45
74 lines not on this page · raw message
target/riscv/vector_internals.h +48 -0
69 lines not on this page · raw message
Signature · 1 line
2.39.2
PATCH 9/45 qemu/bitops.h: Limit rotate amounts
In reply to Lawrence HunterFrom: Dickon Hood <[email protected]>
Rotates have been fixed up to only allow for reasonable rotate amounts (ie, no rotates >7 on an 8b value etc.) This fixes a problem with riscv vector rotate instructions.
--- include/qemu/bitops.h | 24 ++++++++++++++++-------- 1 file changed, 16 insertions(+), 8 deletions(-)
include/qemu/bitops.h +16 -8
@@ -218,7 +218,8 @@ static inline unsigned long find_first_zero_bit(const unsigned long *addr,*/ static inline uint8_t rol8 (uint8_t word,unsigned int shift) { -return (word << shift) | (word >> ((8 - shift) &7 )); + shift &=7 ; +return (word << shift) | (word >> (8 - shift)); }/**
@@ -228,7 +229,8 @@ static inline uint8_t rol8(uint8_t word, unsigned int shift)*/ static inline uint8_t ror8 (uint8_t word,unsigned int shift) { -return (word >> shift) | (word << ((8 - shift) &7 )); + shift &=7 ; +return (word >> shift) | (word << (8 - shift)); }/**
@@ -238,7 +240,8 @@ static inline uint8_t ror8(uint8_t word, unsigned int shift)*/ static inline uint16_t rol16 (uint16_t word,unsigned int shift) { -return (word << shift) | (word >> ((16 - shift) &15 )); + shift &=15 ; +return (word << shift) | (word >> (16 - shift)); }/**
@@ -248,7 +251,8 @@ static inline uint16_t rol16(uint16_t word, unsigned int shift)*/ static inline uint16_t ror16 (uint16_t word,unsigned int shift) { -return (word >> shift) | (word << ((16 - shift) &15 )); + shift &=15 ; +return (word >> shift) | (word << (16 - shift)); }/**
@@ -258,7 +262,8 @@ static inline uint16_t ror16(uint16_t word, unsigned int shift)*/ static inline uint32_t rol32 (uint32_t word,unsigned int shift) { -return (word << shift) | (word >> ((32 - shift) &31 )); + shift &=31 ; +return (word << shift) | (word >> (32 - shift)); }/**
@@ -268,7 +273,8 @@ static inline uint32_t rol32(uint32_t word, unsigned int shift)*/ static inline uint32_t ror32 (uint32_t word,unsigned int shift) { -return (word >> shift) | (word << ((32 - shift) &31 )); + shift &=31 ; +return (word >> shift) | (word << (32 - shift)); }/**
@@ -278,7 +284,8 @@ static inline uint32_t ror32(uint32_t word, unsigned int shift)*/ static inline uint64_t rol64 (uint64_t word,unsigned int shift) { -return (word << shift) | (word >> ((64 - shift) &63 )); + shift &=63 ; +return (word << shift) | (word >> (64 - shift)); }/**
@@ -288,7 +295,8 @@ static inline uint64_t rol64(uint64_t word, unsigned int shift)*/ static inline uint64_t ror64 (uint64_t word,unsigned int shift) { -return (word >> shift) | (word << ((64 - shift) &63 )); + shift &=63 ; +return (word >> shift) | (word << (64 - shift)); }/**
Signature · 1 line
2.39.2
PATCH 10/45 target/riscv: Add vrol.[vv,vx] and vror.[vv,vx,vi] decoding, translation and execution support
In reply to Lawrence HunterFrom: Dickon Hood <[email protected]>
Add an implementation of the vrol.* and vror.* instructions, with mappings between the RISC-V instructions and their internal TCG accelerated implmentations.
There are some missing ror helpers, so I've bodged it by converting them to rols.
--- target/riscv/helper.h | 20 ++++++++ target/riscv/insn32.decode | 7 +++ target/riscv/insn_trans/trans_rvv.c.inc | 3 ++ target/riscv/insn_trans/trans_rvzvkb.c.inc | 57 ++++++++++++++++++++++ target/riscv/vcrypto_helper.c | 36 ++++++++++++++ 5 files changed, 123 insertions(+)
target/riscv/helper.h +20 -0
@@ -1148,3 +1148,23 @@ DEF_HELPER_6(vclmul_vv, void, ptr, ptr, ptr, ptr, env, i32) DEF_HELPER_6(vclmul_vx, void, ptr, ptr, tl, ptr, env, i32) DEF_HELPER_6(vclmulh_vv, void, ptr, ptr, ptr, ptr, env, i32) DEF_HELPER_6(vclmulh_vx, void, ptr, ptr, tl, ptr, env, i32) + +DEF_HELPER_6(vror_vv_b, void, ptr, ptr, ptr, ptr, env, i32) +DEF_HELPER_6(vror_vv_h, void, ptr, ptr, ptr, ptr, env, i32) +DEF_HELPER_6(vror_vv_w, void, ptr, ptr, ptr, ptr, env, i32) +DEF_HELPER_6(vror_vv_d, void, ptr, ptr, ptr, ptr, env, i32) + +DEF_HELPER_6(vror_vx_b, void, ptr, ptr, tl, ptr, env, i32) +DEF_HELPER_6(vror_vx_h, void, ptr, ptr, tl, ptr, env, i32) +DEF_HELPER_6(vror_vx_w, void, ptr, ptr, tl, ptr, env, i32) +DEF_HELPER_6(vror_vx_d, void, ptr, ptr, tl, ptr, env, i32) + +DEF_HELPER_6(vrol_vv_b, void, ptr, ptr, ptr, ptr, env, i32) +DEF_HELPER_6(vrol_vv_h, void, ptr, ptr, ptr, ptr, env, i32) +DEF_HELPER_6(vrol_vv_w, void, ptr, ptr, ptr, ptr, env, i32) +DEF_HELPER_6(vrol_vv_d, void, ptr, ptr, ptr, ptr, env, i32) + +DEF_HELPER_6(vrol_vx_b, void, ptr, ptr, tl, ptr, env, i32) +DEF_HELPER_6(vrol_vx_h, void, ptr, ptr, tl, ptr, env, i32) +DEF_HELPER_6(vrol_vx_w, void, ptr, ptr, tl, ptr, env, i32) +DEF_HELPER_6 (vrol_vx_d, void, ptr, ptr, tl, ptr, env, i32)
target/riscv/insn32.decode +7 -0
@@ -37,6 +37,7 @@ %imm_u 12:s20 !function=ex_shift_12 %imm_bs 30:2 !function=ex_shift_3 %imm_rnum 20:4 +%imm_z6 26:1 15:5 # Argument sets: &empty
@@ -82,6 +83,7 @@ @r_vm ...... vm:1 ..... ..... ... ..... ....... &rmrr %rs2 %rs1 %rd @r_vm_1 ...... . ..... ..... ... ..... ....... &rmrr vm=1 %rs2 %rs1 %rd @r_vm_0 ...... . ..... ..... ... ..... ....... &rmrr vm=0 %rs2 %rs1 %rd +@r2_zimm6 ..... . vm:1 ..... ..... ... ..... ....... &rmrr %rs2 rs1=%imm_z6 %rd @r2_zimm11 . zimm:11 ..... ... ..... ....... %rs1 %rd @r2_zimm10 .. zimm:10 ..... ... ..... ....... %rs1 %rd @r2_s ....... ..... ..... ... ..... ....... %rs2 %rs1
@@ -914,3 +916,8 @@ vclmul_vv 001100 . ..... ..... 010 ..... 1010111 @r_vm vclmul_vx 001100 . ..... ..... 110 ..... 1010111 @r_vm vclmulh_vv 001101 . ..... ..... 010 ..... 1010111 @r_vm vclmulh_vx 001101 . ..... ..... 110 ..... 1010111 @r_vm +vrol_vv 010101 . ..... ..... 000 ..... 1010111 @r_vm +vrol_vx 010101 . ..... ..... 100 ..... 1010111 @r_vm +vror_vv 010100 . ..... ..... 000 ..... 1010111 @r_vm +vror_vx 010100 . ..... ..... 100 ..... 1010111 @r_vm +vror_vi 01010. . ..... ..... 011 ..... 1010111 @r2_zimm6
target/riscv/insn_trans/trans_rvv.c.inc +3 -0
@@ -1374,6 +1374,7 @@ GEN_OPIVX_GVEC_TRANS(vrsub_vx, rsubs) typedef enum { IMM_ZX, /* Zero-extended */ IMM_SX, /* Sign-extended */ + IMM_ZIMM6, /* Truncate to 6 bits */ IMM_TRUNC_SEW, /* Truncate to log(SEW) bits */ IMM_TRUNC_2SEW, /* Truncate to log(2*SEW) bits */ } imm_mode_t;
@@ -1389,6 +1390,8 @@ static int64_t extract_imm(DisasContext *s, uint32_t imm, imm_mode_t imm_mode) return extract64(imm, 0, s->sew + 3); case IMM_TRUNC_2SEW: return extract64(imm, 0, s->sew + 4); + case IMM_ZIMM6: + return extract64(imm, 0, 6); default: g_assert_not_reached(); }
target/riscv/insn_trans/trans_rvzvkb.c.inc +57 -0
60 lines not on this page · raw message
target/riscv/vcrypto_helper.c +36 -0
@@ -57,3 +57,39 @@ RVVCALL(OPIVV2, vclmulh_vv, OP_UUU_D, H8, H8, H8, clmulh64) GEN_VEXT_VV(vclmulh_vv, 8) RVVCALL(OPIVX2, vclmulh_vx, OP_UUU_D, H8, H8, clmulh64) GEN_VEXT_VX(vclmulh_vx, 8) + +RVVCALL(OPIVV2, vror_vv_b, OP_UUU_B, H1, H1, H1, ror8) +RVVCALL(OPIVV2, vror_vv_h, OP_UUU_H, H2, H2, H2, ror16) +RVVCALL(OPIVV2, vror_vv_w, OP_UUU_W, H4, H4, H4, ror32) +RVVCALL(OPIVV2, vror_vv_d, OP_UUU_D, H8, H8, H8, ror64) +GEN_VEXT_VV(vror_vv_b, 1) +GEN_VEXT_VV(vror_vv_h, 2) +GEN_VEXT_VV(vror_vv_w, 4) +GEN_VEXT_VV(vror_vv_d, 8) + +RVVCALL(OPIVX2, vror_vx_b, OP_UUU_B, H1, H1, ror8) +RVVCALL(OPIVX2, vror_vx_h, OP_UUU_H, H2, H2, ror16) +RVVCALL(OPIVX2, vror_vx_w, OP_UUU_W, H4, H4, ror32) +RVVCALL(OPIVX2, vror_vx_d, OP_UUU_D, H8, H8, ror64) +GEN_VEXT_VX(vror_vx_b, 1) +GEN_VEXT_VX(vror_vx_h, 2) +GEN_VEXT_VX(vror_vx_w, 4) +GEN_VEXT_VX(vror_vx_d, 8) + +RVVCALL(OPIVV2, vrol_vv_b, OP_UUU_B, H1, H1, H1, rol8) +RVVCALL(OPIVV2, vrol_vv_h, OP_UUU_H, H2, H2, H2, rol16) +RVVCALL(OPIVV2, vrol_vv_w, OP_UUU_W, H4, H4, H4, rol32) +RVVCALL(OPIVV2, vrol_vv_d, OP_UUU_D, H8, H8, H8, rol64) +GEN_VEXT_VV(vrol_vv_b, 1) +GEN_VEXT_VV(vrol_vv_h, 2) +GEN_VEXT_VV(vrol_vv_w, 4) +GEN_VEXT_VV(vrol_vv_d, 8) + +RVVCALL(OPIVX2, vrol_vx_b, OP_UUU_B, H1, H1, rol8) +RVVCALL(OPIVX2, vrol_vx_h, OP_UUU_H, H2, H2, rol16) +RVVCALL(OPIVX2, vrol_vx_w, OP_UUU_W, H4, H4, rol32) +RVVCALL(OPIVX2, vrol_vx_d, OP_UUU_D, H8, H8, rol64) +GEN_VEXT_VX(vrol_vx_b, 1) +GEN_VEXT_VX(vrol_vx_h, 2) +GEN_VEXT_VX(vrol_vx_w, 4) +GEN_VEXT_VX (vrol_vx_d,8 )
Signature · 1 line
2.39.2
PATCH 11/45 target/riscv: Refactor some of the generic vector functionality
In reply to Lawrence HunterFrom: Kiran Ostrolenk <[email protected]>
This refactoring ensures these functions/macros can be used by both vector and vector-crypto helpers (latter implemented in proceeding commit).
--- target/riscv/vector_helper.c | 39 -------------------------------- target/riscv/vector_internals.h | 40 +++++++++++++++++++++++++++++++++ 2 files changed, 40 insertions(+), 39 deletions(-)
target/riscv/vector_helper.c +0 -39
@@ -3412,11 +3412,6 @@ GEN_VEXT_VF(vfwnmsac_vf_h, 4)GEN_VEXT_VF (vfwnmsac_vf_w,8 ) /* Vector Floating-Point Square-Root Instruction */ -/* (TD, T2, TX2) */ -#define OP_UU_H uint16_t, uint16_t, uint16_t -#define OP_UU_W uint32_t, uint32_t, uint32_t -#define OP_UU_D uint64_t, uint64_t, uint64_t -#define OPFVV1 (NAME ,TD ,T2 ,TX2 ,HD ,HS2 ,OP ) \ static void do_##NAME(void *vd, void *vs2, int i, \ CPURISCVState *env) \
@@ -4109,40 +4104,6 @@ GEN_VEXT_CMP_VF(vmfge_vf_w, uint32_t, H4, vmfge32)GEN_VEXT_CMP_VF (vmfge_vf_d, uint64_t,H8 , vmfge64) /* Vector Floating-Point Classify Instruction */ -#define OPIVV1 (NAME ,TD ,T2 ,TX2 ,HD ,HS2 ,OP ) \ -static void do_##NAME(void *vd, void *vs2, int i) \ -{ \ - TX2 s2 = *((T2 *)vs2 + HS2(i)); \ - *((TD *)vd + HD(i)) = OP(s2); \ -} - -#define GEN_VEXT_V (NAME ,ESZ ) \ -void HELPER(NAME)(void *vd, void *v0, void *vs2, \ - CPURISCVState *env, uint32_t desc) \ -{ \ - uint32_t vm = vext_vm(desc); \ - uint32_t vl = env->vl; \ - uint32_t total_elems = \ - vext_get_total_elems(env, desc, ESZ); \ - uint32_t vta = vext_vta(desc); \ - uint32_t vma = vext_vma(desc); \ - uint32_t i; \ - \ - for (i = env->vstart; i < vl; i++) { \ - if (!vm && !vext_elem_mask(v0, i)) { \ -/* set masked-off elements to 1s */ \ -vext_set_elems_1s (vd, vma, i *ESZ , \ - (i +1 ) *ESZ ); \ -continue ; \ - } \ - do_##NAME(vd, vs2, i); \ - } \ - env->vstart = 0; \ -/* set tail elements to 1s */ \ -vext_set_elems_1s (vd, vta, vl *ESZ , \ - total_elems *ESZ ); \ -} -target_ulong fclass_h (uint64_t frs1) {float16 f = frs1;
target/riscv/vector_internals.h +40 -0
@@ -121,12 +121,52 @@ void vext_set_elems_1s(void *base, uint32_t is_agnostic, uint32_t cnt,/* expand macro args before macro */ #define RVVCALL (macro, ...) macro(__VA_ARGS__) +/* (TD, T2, TX2) */ +#define OP_UU_B uint8_t, uint8_t, uint8_t +#define OP_UU_H uint16_t, uint16_t, uint16_t +#define OP_UU_W uint32_t, uint32_t, uint32_t +#define OP_UU_D uint64_t, uint64_t, uint64_t +/* (TD, T1, T2, TX1, TX2) */ #define OP_UUU_B uint8_t, uint8_t, uint8_t, uint8_t, uint8_t#define OP_UUU_H uint16_t, uint16_t, uint16_t, uint16_t, uint16_t#define OP_UUU_W uint32_t, uint32_t, uint32_t, uint32_t, uint32_t#define OP_UUU_D uint64_t, uint64_t, uint64_t, uint64_t, uint64_t +#define OPIVV1 (NAME ,TD ,T2 ,TX2 ,HD ,HS2 ,OP ) \ +static void do_##NAME(void *vd, void *vs2, int i) \ +{ \ + TX2 s2 = *((T2 *)vs2 + HS2(i)); \ + *((TD *)vd + HD(i)) = OP(s2); \ +} + +#define GEN_VEXT_V (NAME ,ESZ ) \ +void HELPER(NAME)(void *vd, void *v0, void *vs2, \ + CPURISCVState *env, uint32_t desc) \ +{ \ + uint32_t vm = vext_vm(desc); \ + uint32_t vl = env->vl; \ + uint32_t total_elems = \ + vext_get_total_elems(env, desc, ESZ); \ + uint32_t vta = vext_vta(desc); \ + uint32_t vma = vext_vma(desc); \ + uint32_t i; \ + \ + for (i = env->vstart; i < vl; i++) { \ + if (!vm && !vext_elem_mask(v0, i)) { \ +/* set masked-off elements to 1s */ \ +vext_set_elems_1s (vd, vma, i *ESZ , \ + (i +1 ) *ESZ ); \ +continue ; \ + } \ + do_##NAME(vd, vs2, i); \ + } \ + env->vstart = 0; \ +/* set tail elements to 1s */ \ +vext_set_elems_1s (vd, vta, vl *ESZ , \ + total_elems *ESZ ); \ +} +/* operation of two vector elements */ typedef void opivv2_fn (void *vd,void *vs1,void *vs2,int i);
Signature · 1 line
2.39.2
PATCH 12/45 target/riscv: Add vbrev8.v decoding, translation and execution support
In reply to Lawrence HunterFrom: William Salmon <[email protected]>
--- target/riscv/helper.h | 5 +++ target/riscv/insn32.decode | 1 + target/riscv/insn_trans/trans_rvzvkb.c.inc | 38 ++++++++++++++++++++++ target/riscv/vcrypto_helper.c | 21 ++++++++++++ 4 files changed, 65 insertions(+)
target/riscv/helper.h +5 -0
@@ -1168,3 +1168,8 @@ DEF_HELPER_6(vrol_vx_b, void, ptr, ptr, tl, ptr, env, i32) DEF_HELPER_6(vrol_vx_h, void, ptr, ptr, tl, ptr, env, i32) DEF_HELPER_6(vrol_vx_w, void, ptr, ptr, tl, ptr, env, i32) DEF_HELPER_6(vrol_vx_d, void, ptr, ptr, tl, ptr, env, i32) + +DEF_HELPER_5(vbrev8_v_b, void, ptr, ptr, ptr, env, i32) +DEF_HELPER_5(vbrev8_v_h, void, ptr, ptr, ptr, env, i32) +DEF_HELPER_5(vbrev8_v_w, void, ptr, ptr, ptr, env, i32) +DEF_HELPER_5 (vbrev8_v_d, void, ptr, ptr, ptr, env, i32)
target/riscv/insn32.decode +1 -0
@@ -921,3 +921,4 @@ vrol_vx 010101 . ..... ..... 100 ..... 1010111 @r_vm vror_vv 010100 . ..... ..... 000 ..... 1010111 @r_vm vror_vx 010100 . ..... ..... 100 ..... 1010111 @r_vm vror_vi 01010. . ..... ..... 011 ..... 1010111 @r2_zimm6 +vbrev8_v 010010 . ..... 01000 010 ..... 1010111 @r2_vm
target/riscv/insn_trans/trans_rvzvkb.c.inc +38 -0
@@ -145,3 +145,41 @@ GEN_OPIVX_GVEC_SHIFT_TRANS_CHECK(vrol_vx, rotls, zvkb_vx_check) GEN_OPIVV_GVEC_TRANS_CHECK(vror_vv, rotrv, zvkb_vv_check) GEN_OPIVX_GVEC_SHIFT_TRANS_CHECK(vror_vx, rotrs, zvkb_vx_check) GEN_OPIVI_GVEC_TRANS_CHECK(vror_vi, IMM_ZIMM6, vror_vx, rotri, zvkb_vx_check) + +#define GEN_OPIV_TRANS(NAME, CHECK) \ +static bool trans_##NAME(DisasContext *s, arg_rmr * a) \ +{ \ + if (CHECK(s, a)) { \ + uint32_t data = 0; \ + static gen_helper_gvec_3_ptr * const fns[4] = { \ + gen_helper_##NAME##_b, gen_helper_##NAME##_h, \ + gen_helper_##NAME##_w, gen_helper_##NAME##_d, \ + }; \ + TCGLabel *over = gen_new_label(); \ + tcg_gen_brcond_tl(TCG_COND_GEU, cpu_vstart, cpu_vl, over); \ + \ + data = FIELD_DP32(data, VDATA, VM, a->vm); \ + data = FIELD_DP32(data, VDATA, LMUL, s->lmul); \ + data = FIELD_DP32(data, VDATA, VTA, s->vta); \ + data = FIELD_DP32(data, VDATA, VTA_ALL_1S, s->cfg_vta_all_1s); \ + data = FIELD_DP32(data, VDATA, VMA, s->vma); \ + tcg_gen_gvec_3_ptr(vreg_ofs(s, a->rd), vreg_ofs(s, 0), \ + vreg_ofs(s, a->rs2), cpu_env, \ + s->cfg_ptr->vlen / 8, s->cfg_ptr->vlen / 8, \ + data, fns[s->sew]); \ + mark_vs_dirty(s); \ + gen_set_label(over); \ + return true; \ + } \ + return false; \ +} + +static bool vxrev8_check(DisasContext *s, arg_rmr *a) +{ + return s->cfg_ptr->ext_zvkb == true && + require_rvv(s) && + vext_check_isa_ill(s) && + vext_check_ss(s, a->rd, a->rs2, a->vm); +} + +GEN_OPIV_TRANS(vbrev8_v, vxrev8_check)
target/riscv/vcrypto_helper.c +21 -0
@@ -93,3 +93,24 @@ GEN_VEXT_VX(vrol_vx_b, 1) GEN_VEXT_VX(vrol_vx_h, 2) GEN_VEXT_VX(vrol_vx_w, 4) GEN_VEXT_VX(vrol_vx_d, 8) + +static uint64_t brev8 (uint64_t val) +{ + val = ((val &0x5555555555555555ull ) <<1 ) + | ((val &0xAAAAAAAAAAAAAAAAull ) >>1 ); + val = ((val &0x3333333333333333ull ) <<2 ) + | ((val &0xCCCCCCCCCCCCCCCCull ) >>2 ); + val = ((val &0x0F0F0F0F0F0F0F0Full ) <<4 ) + | ((val &0xF0F0F0F0F0F0F0F0ull ) >>4 ); + +return val; +} + +RVVCALL(OPIVV1, vbrev8_v_b, OP_UU_B, H1, H1, brev8) +RVVCALL(OPIVV1, vbrev8_v_h, OP_UU_H, H2, H2, brev8) +RVVCALL(OPIVV1, vbrev8_v_w, OP_UU_W, H4, H4, brev8) +RVVCALL(OPIVV1, vbrev8_v_d, OP_UU_D, H8, H8, brev8) +GEN_VEXT_V(vbrev8_v_b, 1) +GEN_VEXT_V(vbrev8_v_h, 2) +GEN_VEXT_V(vbrev8_v_w, 4) +GEN_VEXT_V (vbrev8_v_d,8 )
Signature · 1 line
2.39.2
PATCH 13/45 target/riscv: Add vrev8.v decoding, translation and execution support
In reply to Lawrence HunterFrom: Nazar Kazakov <[email protected]>
--- target/riscv/helper.h | 4 ++++ target/riscv/insn32.decode | 1 + target/riscv/insn_trans/trans_rvzvkb.c.inc | 1 + target/riscv/vcrypto_helper.c | 11 +++++++++++ 4 files changed, 17 insertions(+)
target/riscv/helper.h +4 -0
@@ -1169,6 +1169,10 @@ DEF_HELPER_6(vrol_vx_h, void, ptr, ptr, tl, ptr, env, i32) DEF_HELPER_6(vrol_vx_w, void, ptr, ptr, tl, ptr, env, i32) DEF_HELPER_6(vrol_vx_d, void, ptr, ptr, tl, ptr, env, i32) +DEF_HELPER_5(vrev8_v_b, void, ptr, ptr, ptr, env, i32) +DEF_HELPER_5(vrev8_v_h, void, ptr, ptr, ptr, env, i32) +DEF_HELPER_5(vrev8_v_w, void, ptr, ptr, ptr, env, i32) +DEF_HELPER_5(vrev8_v_d, void, ptr, ptr, ptr, env, i32) DEF_HELPER_5(vbrev8_v_b, void, ptr, ptr, ptr, env, i32) DEF_HELPER_5(vbrev8_v_h, void, ptr, ptr, ptr, env, i32)DEF_HELPER_5 (vbrev8_v_w, void, ptr, ptr, ptr, env, i32)
target/riscv/insn32.decode +1 -0
@@ -922,3 +922,4 @@ vror_vv 010100 . ..... ..... 000 ..... 1010111 @r_vm vror_vx 010100 . ..... ..... 100 ..... 1010111 @r_vm vror_vi 01010. . ..... ..... 011 ..... 1010111 @r2_zimm6 vbrev8_v 010010 . ..... 01000 010 ..... 1010111 @r2_vm +vrev8_v 010010 . ..... 01001 010 ..... 1010111 @r2_vm
target/riscv/insn_trans/trans_rvzvkb.c.inc +1 -0
@@ -183,3 +183,4 @@ static bool vxrev8_check(DisasContext *s, arg_rmr *a) } GEN_OPIV_TRANS(vbrev8_v, vxrev8_check) +GEN_OPIV_TRANS(vrev8_v, vxrev8_check)
target/riscv/vcrypto_helper.c +11 -0
@@ -20,6 +20,7 @@#include "qemu/osdep.h" #include "qemu/host-utils.h" #include "qemu/bitops.h" +#include "qemu/bswap.h" #include "cpu.h" #include "exec/memop.h" #include "exec/exec-all.h"
@@ -114,3 +115,13 @@ GEN_VEXT_V(vbrev8_v_b, 1) GEN_VEXT_V(vbrev8_v_h, 2) GEN_VEXT_V(vbrev8_v_w, 4) GEN_VEXT_V(vbrev8_v_d, 8) + +#define DO_IDENTITY (a) (a) +RVVCALL(OPIVV1, vrev8_v_b, OP_UU_B, H1, H1, DO_IDENTITY) +RVVCALL(OPIVV1, vrev8_v_h, OP_UU_H, H2, H2, bswap16) +RVVCALL(OPIVV1, vrev8_v_w, OP_UU_W, H4, H4, bswap32) +RVVCALL(OPIVV1, vrev8_v_d, OP_UU_D, H8, H8, bswap64) +GEN_VEXT_V(vrev8_v_b, 1) +GEN_VEXT_V(vrev8_v_h, 2) +GEN_VEXT_V(vrev8_v_w, 4) +GEN_VEXT_V (vrev8_v_d,8 )
Signature · 1 line
2.39.2
PATCH 14/45 target/riscv: Add vandn.[vv,vx] decoding, translation and execution support
In reply to Lawrence HunterFrom: Nazar Kazakov <[email protected]>
--- accel/tcg/tcg-runtime-gvec.c | 11 +++++++ accel/tcg/tcg-runtime.h | 1 + target/riscv/helper.h | 9 ++++++ target/riscv/insn32.decode | 2 ++ target/riscv/insn_trans/trans_rvzvkb.c.inc | 34 ++++++++++++++++++++++ target/riscv/vcrypto_helper.c | 19 ++++++++++++ 6 files changed, 76 insertions(+)
accel/tcg/tcg-runtime-gvec.c +11 -0
@@ -550,6 +550,17 @@ void HELPER(gvec_ands)(void *d, void *a, uint64_t b, uint32_t desc)clear_high (d, oprsz, desc); } +void HELPER (gvec_andsc )(void *d,void *a,uint64_t b,uint32_t desc) +{ +intptr_t oprsz =simd_oprsz (desc); +intptr_t i; + +for (i =0 ; i < oprsz; i +=sizeof (uint64_t )) { + *(uint64_t *)(d + i) = *(uint64_t *)(a + i) & ~b; + } +clear_high (d, oprsz, desc); +} +void HELPER (gvec_xors )(void *d,void *a,uint64_t b,uint32_t desc) {intptr_t oprsz =simd_oprsz (desc);
accel/tcg/tcg-runtime.h +1 -0
@@ -217,6 +217,7 @@ DEF_HELPER_FLAGS_4(gvec_nor, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(gvec_eqv, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(gvec_ands, TCG_CALL_NO_RWG, void, ptr, ptr, i64, i32) +DEF_HELPER_FLAGS_4(gvec_andsc, TCG_CALL_NO_RWG, void, ptr, ptr, i64, i32) DEF_HELPER_FLAGS_4(gvec_xors, TCG_CALL_NO_RWG, void, ptr, ptr, i64, i32)DEF_HELPER_FLAGS_4 (gvec_ors,TCG_CALL_NO_RWG , void, ptr, ptr, i64, i32)
target/riscv/helper.h +9 -0
@@ -1177,3 +1177,12 @@ DEF_HELPER_5(vbrev8_v_b, void, ptr, ptr, ptr, env, i32) DEF_HELPER_5(vbrev8_v_h, void, ptr, ptr, ptr, env, i32) DEF_HELPER_5(vbrev8_v_w, void, ptr, ptr, ptr, env, i32) DEF_HELPER_5(vbrev8_v_d, void, ptr, ptr, ptr, env, i32) + +DEF_HELPER_6(vandn_vv_b, void, ptr, ptr, ptr, ptr, env, i32) +DEF_HELPER_6(vandn_vv_h, void, ptr, ptr, ptr, ptr, env, i32) +DEF_HELPER_6(vandn_vv_w, void, ptr, ptr, ptr, ptr, env, i32) +DEF_HELPER_6(vandn_vv_d, void, ptr, ptr, ptr, ptr, env, i32) +DEF_HELPER_6(vandn_vx_b, void, ptr, ptr, tl, ptr, env, i32) +DEF_HELPER_6(vandn_vx_h, void, ptr, ptr, tl, ptr, env, i32) +DEF_HELPER_6(vandn_vx_w, void, ptr, ptr, tl, ptr, env, i32) +DEF_HELPER_6 (vandn_vx_d, void, ptr, ptr, tl, ptr, env, i32)
target/riscv/insn32.decode +2 -0
@@ -923,3 +923,5 @@ vror_vx 010100 . ..... ..... 100 ..... 1010111 @r_vm vror_vi 01010. . ..... ..... 011 ..... 1010111 @r2_zimm6 vbrev8_v 010010 . ..... 01000 010 ..... 1010111 @r2_vm vrev8_v 010010 . ..... 01001 010 ..... 1010111 @r2_vm +vandn_vv 000001 . ..... ..... 000 ..... 1010111 @r_vm +vandn_vx 000001 . ..... ..... 100 ..... 1010111 @r_vm
target/riscv/insn_trans/trans_rvzvkb.c.inc +34 -0
@@ -146,6 +146,40 @@ GEN_OPIVV_GVEC_TRANS_CHECK(vror_vv, rotrv, zvkb_vv_check) GEN_OPIVX_GVEC_SHIFT_TRANS_CHECK(vror_vx, rotrs, zvkb_vx_check) GEN_OPIVI_GVEC_TRANS_CHECK(vror_vi, IMM_ZIMM6, vror_vx, rotri, zvkb_vx_check) + +static void tcg_gen_gvec_andsc(unsigned vece, uint32_t dofs, uint32_t aofs, + TCGv_i64 c, uint32_t oprsz, uint32_t maxsz) +{ + static GVecGen2s g = { + .fni8 = tcg_gen_andc_i64, + .fniv = tcg_gen_andc_vec, + .fno = gen_helper_gvec_andsc, + .prefer_i64 = TCG_TARGET_REG_BITS == 64, + }; + + g.vece = vece; + + tcg_gen_dup_i64(vece, c, c); + tcg_gen_gvec_2s(dofs, aofs, oprsz, maxsz, c, &g); +} + +#define GEN_OPIVX_GVEC_TRANS_CHECK(NAME, SUF, CHECK) \ +static bool trans_##NAME(DisasContext *s, arg_rmrr *a) \ +{ \ + if (CHECK(s, a)) { \ + static gen_helper_opivx * const fns[4] = { \ + gen_helper_##NAME##_b, gen_helper_##NAME##_h, \ + gen_helper_##NAME##_w, gen_helper_##NAME##_d, \ + }; \ + return do_opivx_gvec(s, a, tcg_gen_gvec_##SUF, fns[s->sew]); \ + } \ + return false; \ +} + +/* vandn.v[vx] */ +GEN_OPIVV_GVEC_TRANS_CHECK(vandn_vv, andc, zvkb_vv_check) +GEN_OPIVX_GVEC_TRANS_CHECK(vandn_vx, andsc, zvkb_vx_check) + #define GEN_OPIV_TRANS(NAME, CHECK) \ static bool trans_##NAME(DisasContext *s, arg_rmr * a) \ { \
target/riscv/vcrypto_helper.c +19 -0
@@ -125,3 +125,22 @@ GEN_VEXT_V(vrev8_v_b, 1) GEN_VEXT_V(vrev8_v_h, 2) GEN_VEXT_V(vrev8_v_w, 4) GEN_VEXT_V(vrev8_v_d, 8) + +#define DO_ANDN (a, b) ((a) & ~(b)) +RVVCALL(OPIVV2, vandn_vv_b, OP_UUU_B, H1, H1, H1, DO_ANDN) +RVVCALL(OPIVV2, vandn_vv_h, OP_UUU_H, H2, H2, H2, DO_ANDN) +RVVCALL(OPIVV2, vandn_vv_w, OP_UUU_W, H4, H4, H4, DO_ANDN) +RVVCALL(OPIVV2, vandn_vv_d, OP_UUU_D, H8, H8, H8, DO_ANDN) +GEN_VEXT_VV(vandn_vv_b, 1) +GEN_VEXT_VV(vandn_vv_h, 2) +GEN_VEXT_VV(vandn_vv_w, 4) +GEN_VEXT_VV(vandn_vv_d, 8) + +RVVCALL(OPIVX2, vandn_vx_b, OP_UUU_B, H1, H1, DO_ANDN) +RVVCALL(OPIVX2, vandn_vx_h, OP_UUU_H, H2, H2, DO_ANDN) +RVVCALL(OPIVX2, vandn_vx_w, OP_UUU_W, H4, H4, DO_ANDN) +RVVCALL(OPIVX2, vandn_vx_d, OP_UUU_D, H8, H8, DO_ANDN) +GEN_VEXT_VX(vandn_vx_b, 1) +GEN_VEXT_VX(vandn_vx_h, 2) +GEN_VEXT_VX(vandn_vx_w, 4) +GEN_VEXT_VX (vandn_vx_d,8 )
Signature · 1 line
2.39.2
PATCH 15/45 target/riscv: Expose zvkb cpu property
In reply to Lawrence HunterFrom: Nazar Kazakov <[email protected]>
--- target/riscv/cpu.c | 3 +++ 1 file changed, 3 insertions(+)
target/riscv/cpu.c +3 -0
@@ -1469,6 +1469,9 @@ static Property riscv_cpu_extensions[] = {DEFINE_PROP_BOOL ("x-zvfh" ,RISCVCPU , cfg.ext_zvfh, false),DEFINE_PROP_BOOL ("x-zvfhmin" ,RISCVCPU , cfg.ext_zvfhmin, false), +/* Vector cryptography extensions */ +DEFINE_PROP_BOOL ("x-zvkb" ,RISCVCPU , cfg.ext_zvkb, false), + DEFINE_PROP_END_OF_LIST(), };
Signature · 1 line
2.39.2
PATCH 16/45 target/riscv: Add zvkned cpu property
In reply to Lawrence HunterFrom: Nazar Kazakov <[email protected]>
--- target/riscv/cpu.c | 3 ++- target/riscv/cpu.h | 1 + 2 files changed, 3 insertions(+), 1 deletion(-)
target/riscv/cpu.c +2 -1
@@ -110,6 +110,7 @@ static const struct isa_ext_data isa_edata_arr[] = {ISA_EXT_DATA_ENTRY (zvfh, true,PRIV_VERSION_1_12_0 , ext_zvfh),ISA_EXT_DATA_ENTRY (zvfhmin, true,PRIV_VERSION_1_12_0 , ext_zvfhmin),ISA_EXT_DATA_ENTRY (zvkb, true,PRIV_VERSION_1_12_0 , ext_zvkb), +ISA_EXT_DATA_ENTRY (zvkned, true,PRIV_VERSION_1_12_0 , ext_zvkned),ISA_EXT_DATA_ENTRY (zhinx, true,PRIV_VERSION_1_12_0 , ext_zhinx),ISA_EXT_DATA_ENTRY (zhinxmin, true,PRIV_VERSION_1_12_0 , ext_zhinxmin), ISA_EXT_DATA_ENTRY(smaia, true, PRIV_VERSION_1_12_0, ext_smaia),
@@ -1216,7 +1217,7 @@ static void riscv_cpu_realize(DeviceState *dev, Error **errp)* In principle zve*x would also suffice here, were they supported * in qemu */ -if (cpu->cfg.ext_zvkb && +if ((cpu->cfg.ext_zvkb || cpu->cfg.ext_zvkned) && !(cpu->cfg.ext_zve32f || cpu->cfg.ext_zve64f || cpu->cfg.ext_zve64d || cpu->cfg.ext_v)) { error_setg(
target/riscv/cpu.h +1 -0
@@ -471,6 +471,7 @@ struct RISCVCPUConfig {bool ext_zve64f;bool ext_zve64d;bool ext_zvkb; +bool ext_zvkned;bool ext_zmmul;bool ext_zvfh;bool ext_zvfhmin;
Signature · 1 line
2.39.2
PATCH 17/45 target/riscv: Add vaesef.vv decoding, translation and execution support
In reply to Lawrence Hunter--- target/riscv/helper.h | 3 + target/riscv/insn32.decode | 4 ++ target/riscv/insn_trans/trans_rvzvkned.c.inc | 72 +++++++++++++++++++ target/riscv/op_helper.c | 5 ++ target/riscv/translate.c | 1 + target/riscv/vcrypto_helper.c | 73 ++++++++++++++++++++ 6 files changed, 158 insertions(+) create mode 100644 target/riscv/insn_trans/trans_rvzvkned.c.inc
target/riscv/helper.h +3 -0
@@ -1,5 +1,6 @@/* Exceptions */ DEF_HELPER_2(raise_exception, noreturn, env, i32) +DEF_HELPER_2 (restore_cpu_and_raise_exception, noreturn, env, i32) /* Floating Point - rounding mode */DEF_HELPER_FLAGS_2 (set_rounding_mode,TCG_CALL_NO_WG , void, env, i32)
@@ -1186,3 +1187,5 @@ DEF_HELPER_6(vandn_vx_b, void, ptr, ptr, tl, ptr, env, i32) DEF_HELPER_6(vandn_vx_h, void, ptr, ptr, tl, ptr, env, i32) DEF_HELPER_6(vandn_vx_w, void, ptr, ptr, tl, ptr, env, i32) DEF_HELPER_6(vandn_vx_d, void, ptr, ptr, tl, ptr, env, i32) + +DEF_HELPER_4 (vaesef_vv, void, ptr, ptr, env, i32)
target/riscv/insn32.decode +4 -0
@@ -75,6 +75,7 @@ @r_rm ....... ..... ..... ... ..... ....... %rs2 %rs1 %rm %rd @r2_rm ....... ..... ..... ... ..... ....... %rs1 %rm %rd @r2 ....... ..... ..... ... ..... ....... &r2 %rs1 %rd +@r2_vm_1 ...... . ..... ..... ... ..... ....... &rmr vm=1 %rs2 %rd @r2_nfvm ... ... vm:1 ..... ..... ... ..... ....... &r2nfvm %nf %rs1 %rd @r2_vm ...... vm:1 ..... ..... ... ..... ....... &rmr %rs2 %rd @r1_vm ...... vm:1 ..... ..... ... ..... ....... %rd
@@ -925,3 +926,6 @@ vbrev8_v 010010 . ..... 01000 010 ..... 1010111 @r2_vm vrev8_v 010010 . ..... 01001 010 ..... 1010111 @r2_vm vandn_vv 000001 . ..... ..... 000 ..... 1010111 @r_vm vandn_vx 000001 . ..... ..... 100 ..... 1010111 @r_vm + +# *** RV64 Zvkned vector crypto extension *** +vaesef_vv 101000 1 ..... 00011 010 ..... 1110111 @r2_vm_1
target/riscv/insn_trans/trans_rvzvkned.c.inc new file +72 -0
72 lines not on this page · raw message
target/riscv/op_helper.c +5 -0
@@ -38,6 +38,11 @@ void helper_raise_exception(CPURISCVState *env, uint32_t exception)riscv_raise_exception (env, exception,0 ); } +void helper_restore_cpu_and_raise_exception (CPURISCVState *env,uint32_t exception) +{ +riscv_raise_exception (env, exception,GETPC ()); +} +target_ulong helper_csrr (CPURISCVState *env,int csr) {/*
target/riscv/translate.c +1 -0
@@ -1084,6 +1084,7 @@ static uint32_t opcode_at(DisasContextBase *dcbase, target_ulong pc)#include "insn_trans/trans_rvzfh.c.inc" #include "insn_trans/trans_rvk.c.inc" #include "insn_trans/trans_rvzvkb.c.inc" +#include "insn_trans/trans_rvzvkned.c.inc" #include "insn_trans/trans_privileged.c.inc" #include "insn_trans/trans_svinval.c.inc" #include "decode-xthead.c.inc"
target/riscv/vcrypto_helper.c +73 -0
82 lines not on this page · raw message
Signature · 1 line
2.39.2
PATCH 18/45 target/riscv: Add vaesef.vs decoding, translation and execution support
In reply to Lawrence Hunter--- target/riscv/helper.h | 1 + target/riscv/insn32.decode | 1 + target/riscv/insn_trans/trans_rvzvkned.c.inc | 20 +++++++++++ target/riscv/vcrypto_helper.c | 36 ++++++++++++++++++++ 4 files changed, 58 insertions(+)
target/riscv/helper.h +1 -0
@@ -1189,3 +1189,4 @@ DEF_HELPER_6(vandn_vx_w, void, ptr, ptr, tl, ptr, env, i32) DEF_HELPER_6(vandn_vx_d, void, ptr, ptr, tl, ptr, env, i32) DEF_HELPER_4(vaesef_vv, void, ptr, ptr, env, i32) +DEF_HELPER_4 (vaesef_vs, void, ptr, ptr, env, i32)
target/riscv/insn32.decode +1 -0
@@ -929,3 +929,4 @@ vandn_vx 000001 . ..... ..... 100 ..... 1010111 @r_vm # *** RV64 Zvkned vector crypto extension *** vaesef_vv 101000 1 ..... 00011 010 ..... 1110111 @r2_vm_1 +vaesef_vs 101001 1 ..... 00011 010 ..... 1110111 @r2_vm_1
target/riscv/insn_trans/trans_rvzvkned.c.inc +20 -0
@@ -69,4 +69,24 @@ static bool vaes_check_vv(DisasContext *s, arg_rmr *a) require_align(a->rs2, s->lmul) && s->vstart % 4 == 0 && s->sew == MO_32; } + +static bool vaes_check_overlap(DisasContext *s, int vd, int vs2) +{ + int8_t op_size = s->lmul <= 0 ? 1 : 1 << s->lmul; + return !is_overlapped(vd, op_size, vs2, 1); +} + +static bool vaes_check_vs(DisasContext *s, arg_rmr *a) +{ + return vaes_check_overlap(s, a->rd, a->rs2) && + MAXSZ(s) >= (128 / 8) && /* EGW in bytes */ + s->cfg_ptr->ext_zvkned == true && + require_rvv(s) && + vext_check_isa_ill(s) && + require_align(a->rd, s->lmul) && + s->vstart % 4 == 0 && + s->sew == MO_32; +} + GEN_V_UNMASKED_TRANS(vaesef_vv, vaes_check_vv) +GEN_V_UNMASKED_TRANS(vaesef_vs, vaes_check_vs)
target/riscv/vcrypto_helper.c +36 -0
@@ -214,6 +214,42 @@ void HELPER(NAME)(void *vd_vptr, void *vs2_vptr, CPURISCVState *env, \vext_set_elems_1s (vd, vta, vl *4 , total_elems *4 ); \ } +#define GEN_ZVKNED_HELPER_VS (NAME , ...) \ +void HELPER(NAME)(void *vd_vptr, void *vs2_vptr, CPURISCVState *env, \ + uint32_t desc) \ +{ \ + uint64_t *vd = vd_vptr; \ + uint64_t *vs2 = vs2_vptr; \ + uint32_t vl = env->vl; \ + uint32_t total_elems = vext_get_total_elems(env, desc, 4); \ + uint32_t vta = vext_vta(desc); \ + \ + for (uint32_t i = env->vstart / 4; i < env->vl / 4; i++) { \ + uint64_t round_key[2] = { \ + cpu_to_le64(vs2[0]), \ + cpu_to_le64(vs2[1]), \ + }; \ + uint8_t round_state[4][4]; \ + cpu_to_le64s(vd + i * 2 + 0); \ + cpu_to_le64s(vd + i * 2 + 1); \ + for (int j = 0; j < 16; j++) { \ + round_state[j / 4][j % 4] = ((uint8_t *)(vd + i * 2))[j]; \ + } \ + __VA_ARGS__; \ + for (int j = 0; j < 16; j++) { \ + ((uint8_t *)(vd + i * 2))[j] = round_state[j / 4][j % 4]; \ + } \ + le64_to_cpus(vd + i * 2 + 0); \ + le64_to_cpus(vd + i * 2 + 1); \ + } \ + env->vstart = 0; \ +/* set tail elements to 1s */ \ +vext_set_elems_1s (vd, vta, vl *4 , total_elems *4 ); \ +} + GEN_ZVKNED_HELPER_VV(vaesef_vv, aes_sub_bytes(round_state);aes_shift_bytes (round_state); xor_round_key(round_state, (uint8_t *)round_key);) +GEN_ZVKNED_HELPER_VS(vaesef_vs, aes_sub_bytes(round_state); +aes_shift_bytes (round_state); + xor_round_key(round_state, (uint8_t *)round_key);)
Signature · 1 line
2.39.2
PATCH 19/45 target/riscv: Add vaesdf.vv decoding, translation and execution support
In reply to Lawrence Hunter--- target/riscv/helper.h | 1 + target/riscv/insn32.decode | 1 + target/riscv/insn_trans/trans_rvzvkned.c.inc | 1 + target/riscv/vcrypto_helper.c | 31 ++++++++++++++++++++ 4 files changed, 34 insertions(+)
target/riscv/helper.h +1 -0
@@ -1190,3 +1190,4 @@ DEF_HELPER_6(vandn_vx_d, void, ptr, ptr, tl, ptr, env, i32) DEF_HELPER_4(vaesef_vv, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesef_vs, void, ptr, ptr, env, i32) +DEF_HELPER_4 (vaesdf_vv, void, ptr, ptr, env, i32)
target/riscv/insn32.decode +1 -0
@@ -930,3 +930,4 @@ vandn_vx 000001 . ..... ..... 100 ..... 1010111 @r_vm # *** RV64 Zvkned vector crypto extension *** vaesef_vv 101000 1 ..... 00011 010 ..... 1110111 @r2_vm_1 vaesef_vs 101001 1 ..... 00011 010 ..... 1110111 @r2_vm_1 +vaesdf_vv 101000 1 ..... 00001 010 ..... 1110111 @r2_vm_1
target/riscv/insn_trans/trans_rvzvkned.c.inc +1 -0
@@ -90,3 +90,4 @@ static bool vaes_check_vs(DisasContext *s, arg_rmr *a) GEN_V_UNMASKED_TRANS(vaesef_vv, vaes_check_vv) GEN_V_UNMASKED_TRANS(vaesef_vs, vaes_check_vs) +GEN_V_UNMASKED_TRANS(vaesdf_vv, vaes_check_vv)
target/riscv/vcrypto_helper.c +31 -0
@@ -181,6 +181,34 @@ static inline void xor_round_key(uint8_t round_state[4][4], uint8_t *round_key) } } +static inline void aes_inv_sub_bytes (uint8_t round_state[4 ][4 ]) +{ +for (int j =0 ; j <16 ; j++) { + round_state[j /4 ][j %4 ] = AES_isbox[round_state[j /4 ][j %4 ]]; + } +} + +static inline void aes_inv_shift_bytes (uint8_t round_state[4 ][4 ]) +{ +uint8_t temp; + temp = round_state[3 ][1 ]; + round_state[3 ][1 ] = round_state[2 ][1 ]; + round_state[2 ][1 ] = round_state[1 ][1 ]; + round_state[1 ][1 ] = round_state[0 ][1 ]; + round_state[0 ][1 ] = temp; + temp = round_state[0 ][2 ]; + round_state[0 ][2 ] = round_state[2 ][2 ]; + round_state[2 ][2 ] = temp; + temp = round_state[1 ][2 ]; + round_state[1 ][2 ] = round_state[3 ][2 ]; + round_state[3 ][2 ] = temp; + temp = round_state[0 ][3 ]; + round_state[0 ][3 ] = round_state[1 ][3 ]; + round_state[1 ][3 ] = round_state[2 ][3 ]; + round_state[2 ][3 ] = round_state[3 ][3 ]; + round_state[3 ][3 ] = temp; +} +#define GEN_ZVKNED_HELPER_VV (NAME , ...) \ void HELPER(NAME)(void *vd_vptr, void *vs2_vptr, CPURISCVState *env, \ uint32_t desc) \
@@ -253,3 +281,6 @@ GEN_ZVKNED_HELPER_VV(vaesef_vv, aes_sub_bytes(round_state); GEN_ZVKNED_HELPER_VS(vaesef_vs, aes_sub_bytes(round_state);aes_shift_bytes (round_state); xor_round_key(round_state, (uint8_t *)round_key);) +GEN_ZVKNED_HELPER_VV(vaesdf_vv, aes_inv_shift_bytes(round_state); +aes_inv_sub_bytes (round_state); + xor_round_key(round_state, (uint8_t *)round_key);)
Signature · 1 line
2.39.2
PATCH 20/45 target/riscv: Add vaesdf.vs decoding, translation and execution support
In reply to Lawrence Hunter--- target/riscv/helper.h | 1 + target/riscv/insn32.decode | 1 + target/riscv/insn_trans/trans_rvzvkned.c.inc | 1 + target/riscv/vcrypto_helper.c | 3 +++ 4 files changed, 6 insertions(+)
target/riscv/helper.h +1 -0
@@ -1191,3 +1191,4 @@ DEF_HELPER_6(vandn_vx_d, void, ptr, ptr, tl, ptr, env, i32) DEF_HELPER_4(vaesef_vv, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesef_vs, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesdf_vv, void, ptr, ptr, env, i32) +DEF_HELPER_4 (vaesdf_vs, void, ptr, ptr, env, i32)
target/riscv/insn32.decode +1 -0
@@ -931,3 +931,4 @@ vandn_vx 000001 . ..... ..... 100 ..... 1010111 @r_vm vaesef_vv 101000 1 ..... 00011 010 ..... 1110111 @r2_vm_1 vaesef_vs 101001 1 ..... 00011 010 ..... 1110111 @r2_vm_1 vaesdf_vv 101000 1 ..... 00001 010 ..... 1110111 @r2_vm_1 +vaesdf_vs 101001 1 ..... 00001 010 ..... 1110111 @r2_vm_1
target/riscv/insn_trans/trans_rvzvkned.c.inc +1 -0
@@ -91,3 +91,4 @@ static bool vaes_check_vs(DisasContext *s, arg_rmr *a) GEN_V_UNMASKED_TRANS(vaesef_vv, vaes_check_vv) GEN_V_UNMASKED_TRANS(vaesef_vs, vaes_check_vs) GEN_V_UNMASKED_TRANS(vaesdf_vv, vaes_check_vv) +GEN_V_UNMASKED_TRANS(vaesdf_vs, vaes_check_vs)
target/riscv/vcrypto_helper.c +3 -0
@@ -284,3 +284,6 @@ GEN_ZVKNED_HELPER_VS(vaesef_vs, aes_sub_bytes(round_state); GEN_ZVKNED_HELPER_VV(vaesdf_vv, aes_inv_shift_bytes(round_state);aes_inv_sub_bytes (round_state); xor_round_key(round_state, (uint8_t *)round_key);) +GEN_ZVKNED_HELPER_VS(vaesdf_vs, aes_inv_shift_bytes(round_state); +aes_inv_sub_bytes (round_state); + xor_round_key(round_state, (uint8_t *)round_key);)
Signature · 1 line
2.39.2
PATCH 21/45 target/riscv: Add vaesdm.vv decoding, translation and execution support
In reply to Lawrence Hunter--- target/riscv/helper.h | 1 + target/riscv/insn32.decode | 1 + target/riscv/insn_trans/trans_rvzvkned.c.inc | 1 + target/riscv/vcrypto_helper.c | 36 ++++++++++++++++++++ 4 files changed, 39 insertions(+)
target/riscv/helper.h +1 -0
@@ -1192,3 +1192,4 @@ DEF_HELPER_4(vaesef_vv, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesef_vs, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesdf_vv, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesdf_vs, void, ptr, ptr, env, i32) +DEF_HELPER_4 (vaesdm_vv, void, ptr, ptr, env, i32)
target/riscv/insn32.decode +1 -0
@@ -932,3 +932,4 @@ vaesef_vv 101000 1 ..... 00011 010 ..... 1110111 @r2_vm_1 vaesef_vs 101001 1 ..... 00011 010 ..... 1110111 @r2_vm_1 vaesdf_vv 101000 1 ..... 00001 010 ..... 1110111 @r2_vm_1 vaesdf_vs 101001 1 ..... 00001 010 ..... 1110111 @r2_vm_1 +vaesdm_vv 101000 1 ..... 00000 010 ..... 1110111 @r2_vm_1
target/riscv/insn_trans/trans_rvzvkned.c.inc +1 -0
@@ -92,3 +92,4 @@ GEN_V_UNMASKED_TRANS(vaesef_vv, vaes_check_vv) GEN_V_UNMASKED_TRANS(vaesef_vs, vaes_check_vs) GEN_V_UNMASKED_TRANS(vaesdf_vv, vaes_check_vv) GEN_V_UNMASKED_TRANS(vaesdf_vs, vaes_check_vs) +GEN_V_UNMASKED_TRANS(vaesdm_vv, vaes_check_vv)
target/riscv/vcrypto_helper.c +36 -0
@@ -209,6 +209,38 @@ static inline void aes_inv_shift_bytes(uint8_t round_state[4][4]) round_state[3 ][3 ] = temp; } +static inline uint8_t xtime (uint8_t x) +{ +return (x <<1 ) ^ (((x >>7 ) &1 ) *0x1b ); +} + +static inline uint8_t multiply (uint8_t x,uint8_t y) +{ +return (((y &1 ) * x) ^ ((y >>1 &1 ) *xtime (x)) ^ + ((y >>2 &1 ) *xtime (xtime (x))) ^ + ((y >>3 &1 ) *xtime (xtime (xtime (x)))) ^ + ((y >>4 &1 ) *xtime (xtime (xtime (xtime (x)))))); +} + +static inline void aes_inv_mix_cols (uint8_t round_state[4 ][4 ]) +{ +uint8_t a, b, c, d; +for (int j =0 ; j <4 ; ++j) { + a = round_state[j][0 ]; + b = round_state[j][1 ]; + c = round_state[j][2 ]; + d = round_state[j][3 ]; + round_state[j][0 ] =multiply (a,0x0e ) ^multiply (b,0x0b ) ^ +multiply (c,0x0d ) ^multiply (d,0x09 ); + round_state[j][1 ] =multiply (a,0x09 ) ^multiply (b,0x0e ) ^ +multiply (c,0x0b ) ^multiply (d,0x0d ); + round_state[j][2 ] =multiply (a,0x0d ) ^multiply (b,0x09 ) ^ +multiply (c,0x0e ) ^multiply (d,0x0b ); + round_state[j][3 ] =multiply (a,0x0b ) ^multiply (b,0x0d ) ^ +multiply (c,0x09 ) ^multiply (d,0x0e ); + } +} +#define GEN_ZVKNED_HELPER_VV (NAME , ...) \ void HELPER(NAME)(void *vd_vptr, void *vs2_vptr, CPURISCVState *env, \ uint32_t desc) \
@@ -287,3 +319,7 @@ GEN_ZVKNED_HELPER_VV(vaesdf_vv, aes_inv_shift_bytes(round_state); GEN_ZVKNED_HELPER_VS(vaesdf_vs, aes_inv_shift_bytes(round_state);aes_inv_sub_bytes (round_state); xor_round_key(round_state, (uint8_t *)round_key);) +GEN_ZVKNED_HELPER_VV(vaesdm_vv, aes_inv_shift_bytes(round_state); +aes_inv_sub_bytes (round_state); +xor_round_key (round_state, (uint8_t *)round_key); + aes_inv_mix_cols(round_state);)
Signature · 1 line
2.39.2
PATCH 22/45 target/riscv: Add vaesdm.vs decoding, translation and execution support
In reply to Lawrence Hunter--- target/riscv/helper.h | 1 + target/riscv/insn32.decode | 1 + target/riscv/insn_trans/trans_rvzvkned.c.inc | 1 + target/riscv/vcrypto_helper.c | 4 ++++ 4 files changed, 7 insertions(+)
target/riscv/helper.h +1 -0
@@ -1193,3 +1193,4 @@ DEF_HELPER_4(vaesef_vs, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesdf_vv, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesdf_vs, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesdm_vv, void, ptr, ptr, env, i32) +DEF_HELPER_4 (vaesdm_vs, void, ptr, ptr, env, i32)
target/riscv/insn32.decode +1 -0
@@ -933,3 +933,4 @@ vaesef_vs 101001 1 ..... 00011 010 ..... 1110111 @r2_vm_1 vaesdf_vv 101000 1 ..... 00001 010 ..... 1110111 @r2_vm_1 vaesdf_vs 101001 1 ..... 00001 010 ..... 1110111 @r2_vm_1 vaesdm_vv 101000 1 ..... 00000 010 ..... 1110111 @r2_vm_1 +vaesdm_vs 101001 1 ..... 00000 010 ..... 1110111 @r2_vm_1
target/riscv/insn_trans/trans_rvzvkned.c.inc +1 -0
@@ -93,3 +93,4 @@ GEN_V_UNMASKED_TRANS(vaesef_vs, vaes_check_vs) GEN_V_UNMASKED_TRANS(vaesdf_vv, vaes_check_vv) GEN_V_UNMASKED_TRANS(vaesdf_vs, vaes_check_vs) GEN_V_UNMASKED_TRANS(vaesdm_vv, vaes_check_vv) +GEN_V_UNMASKED_TRANS(vaesdm_vs, vaes_check_vs)
target/riscv/vcrypto_helper.c +4 -0
@@ -323,3 +323,7 @@ GEN_ZVKNED_HELPER_VV(vaesdm_vv, aes_inv_shift_bytes(round_state);aes_inv_sub_bytes (round_state);xor_round_key (round_state, (uint8_t *)round_key); aes_inv_mix_cols(round_state);) +GEN_ZVKNED_HELPER_VS(vaesdm_vs, aes_inv_shift_bytes(round_state); +aes_inv_sub_bytes (round_state); +xor_round_key (round_state, (uint8_t *)round_key); + aes_inv_mix_cols(round_state);)
Signature · 1 line
2.39.2
PATCH 23/45 target/riscv: Add vaesz.vs decoding, translation and execution support
In reply to Lawrence Hunter--- target/riscv/helper.h | 1 + target/riscv/insn32.decode | 1 + target/riscv/insn_trans/trans_rvzvkned.c.inc | 1 + target/riscv/vcrypto_helper.c | 2 ++ 4 files changed, 5 insertions(+)
target/riscv/helper.h +1 -0
@@ -1194,3 +1194,4 @@ DEF_HELPER_4(vaesdf_vv, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesdf_vs, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesdm_vv, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesdm_vs, void, ptr, ptr, env, i32) +DEF_HELPER_4 (vaesz_vs, void, ptr, ptr, env, i32)
target/riscv/insn32.decode +1 -0
@@ -934,3 +934,4 @@ vaesdf_vv 101000 1 ..... 00001 010 ..... 1110111 @r2_vm_1 vaesdf_vs 101001 1 ..... 00001 010 ..... 1110111 @r2_vm_1 vaesdm_vv 101000 1 ..... 00000 010 ..... 1110111 @r2_vm_1 vaesdm_vs 101001 1 ..... 00000 010 ..... 1110111 @r2_vm_1 +vaesz_vs 101001 1 ..... 00111 010 ..... 1110111 @r2_vm_1
target/riscv/insn_trans/trans_rvzvkned.c.inc +1 -0
@@ -94,3 +94,4 @@ GEN_V_UNMASKED_TRANS(vaesdf_vv, vaes_check_vv) GEN_V_UNMASKED_TRANS(vaesdf_vs, vaes_check_vs) GEN_V_UNMASKED_TRANS(vaesdm_vv, vaes_check_vv) GEN_V_UNMASKED_TRANS(vaesdm_vs, vaes_check_vs) +GEN_V_UNMASKED_TRANS(vaesz_vs, vaes_check_vs)
target/riscv/vcrypto_helper.c +2 -0
@@ -327,3 +327,5 @@ GEN_ZVKNED_HELPER_VS(vaesdm_vs, aes_inv_shift_bytes(round_state);aes_inv_sub_bytes (round_state);xor_round_key (round_state, (uint8_t *)round_key); aes_inv_mix_cols(round_state);) +GEN_ZVKNED_HELPER_VS (vaesz_vs, + xor_round_key(round_state, (uint8_t *)round_key);)
Signature · 1 line
2.39.2
PATCH 24/45 target/riscv: Add vaesem.vv decoding, translation and execution support
In reply to Lawrence HunterFrom: William Salmon <[email protected]>
--- target/riscv/helper.h | 1 + target/riscv/insn32.decode | 1 + target/riscv/insn_trans/trans_rvzvkned.c.inc | 1 + target/riscv/vcrypto_helper.c | 17 +++++++++++++++++ 4 files changed, 20 insertions(+)
target/riscv/helper.h +1 -0
@@ -1192,6 +1192,7 @@ DEF_HELPER_4(vaesef_vv, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesef_vs, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesdf_vv, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesdf_vs, void, ptr, ptr, env, i32) +DEF_HELPER_4(vaesem_vv, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesdm_vv, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesdm_vs, void, ptr, ptr, env, i32)DEF_HELPER_4 (vaesz_vs, void, ptr, ptr, env, i32)
target/riscv/insn32.decode +1 -0
@@ -932,6 +932,7 @@ vaesef_vv 101000 1 ..... 00011 010 ..... 1110111 @r2_vm_1 vaesef_vs 101001 1 ..... 00011 010 ..... 1110111 @r2_vm_1 vaesdf_vv 101000 1 ..... 00001 010 ..... 1110111 @r2_vm_1 vaesdf_vs 101001 1 ..... 00001 010 ..... 1110111 @r2_vm_1 +vaesem_vv 101000 1 ..... 00010 010 ..... 1110111 @r2_vm_1 vaesdm_vv 101000 1 ..... 00000 010 ..... 1110111 @r2_vm_1 vaesdm_vs 101001 1 ..... 00000 010 ..... 1110111 @r2_vm_1 vaesz_vs 101001 1 ..... 00111 010 ..... 1110111 @r2_vm_1
target/riscv/insn_trans/trans_rvzvkned.c.inc +1 -0
@@ -95,3 +95,4 @@ GEN_V_UNMASKED_TRANS(vaesdf_vs, vaes_check_vs) GEN_V_UNMASKED_TRANS(vaesdm_vv, vaes_check_vv) GEN_V_UNMASKED_TRANS(vaesdm_vs, vaes_check_vs) GEN_V_UNMASKED_TRANS(vaesz_vs, vaes_check_vs) +GEN_V_UNMASKED_TRANS(vaesem_vv, vaes_check_vv)
target/riscv/vcrypto_helper.c +17 -0
@@ -241,6 +241,20 @@ static inline void aes_inv_mix_cols(uint8_t round_state[4][4]) } } +static inline void aes_mix_cols (uint8_t round_state[4 ][4 ]) +{ +uint8_t a, b; +for (int j =0 ; j <4 ; ++j) { + a = round_state[j][0 ]; + b = round_state[j][0 ] ^ round_state[j][1 ] ^ round_state[j][2 ] ^ + round_state[j][3 ]; + round_state[j][0 ] ^=xtime (round_state[j][0 ] ^ round_state[j][1 ]) ^ b; + round_state[j][1 ] ^=xtime (round_state[j][1 ] ^ round_state[j][2 ]) ^ b; + round_state[j][2 ] ^=xtime (round_state[j][2 ] ^ round_state[j][3 ]) ^ b; + round_state[j][3 ] ^=xtime (round_state[j][3 ] ^ a) ^ b; + } +} +#define GEN_ZVKNED_HELPER_VV (NAME , ...) \ void HELPER(NAME)(void *vd_vptr, void *vs2_vptr, CPURISCVState *env, \ uint32_t desc) \
@@ -319,6 +333,9 @@ GEN_ZVKNED_HELPER_VV(vaesdf_vv, aes_inv_shift_bytes(round_state); GEN_ZVKNED_HELPER_VS(vaesdf_vs, aes_inv_shift_bytes(round_state);aes_inv_sub_bytes (round_state); xor_round_key(round_state, (uint8_t *)round_key);) +GEN_ZVKNED_HELPER_VV(vaesem_vv, aes_shift_bytes(round_state); +aes_sub_bytes (round_state);aes_mix_cols (round_state); + xor_round_key(round_state, (uint8_t *)round_key);) GEN_ZVKNED_HELPER_VV(vaesdm_vv, aes_inv_shift_bytes(round_state);aes_inv_sub_bytes (round_state);xor_round_key (round_state, (uint8_t *)round_key);
Signature · 1 line
2.39.2
PATCH 25/45 target/riscv: Add vaesem.vs decoding, translation and execution support
In reply to Lawrence HunterFrom: William Salmon <[email protected]>
--- target/riscv/helper.h | 1 + target/riscv/insn32.decode | 1 + target/riscv/insn_trans/trans_rvzvkned.c.inc | 1 + target/riscv/vcrypto_helper.c | 3 +++ 4 files changed, 6 insertions(+)
target/riscv/helper.h +1 -0
@@ -1193,6 +1193,7 @@ DEF_HELPER_4(vaesef_vs, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesdf_vv, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesdf_vs, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesem_vv, void, ptr, ptr, env, i32) +DEF_HELPER_4(vaesem_vs, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesdm_vv, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesdm_vs, void, ptr, ptr, env, i32)DEF_HELPER_4 (vaesz_vs, void, ptr, ptr, env, i32)
target/riscv/insn32.decode +1 -0
@@ -933,6 +933,7 @@ vaesef_vs 101001 1 ..... 00011 010 ..... 1110111 @r2_vm_1 vaesdf_vv 101000 1 ..... 00001 010 ..... 1110111 @r2_vm_1 vaesdf_vs 101001 1 ..... 00001 010 ..... 1110111 @r2_vm_1 vaesem_vv 101000 1 ..... 00010 010 ..... 1110111 @r2_vm_1 +vaesem_vs 101001 1 ..... 00010 010 ..... 1110111 @r2_vm_1 vaesdm_vv 101000 1 ..... 00000 010 ..... 1110111 @r2_vm_1 vaesdm_vs 101001 1 ..... 00000 010 ..... 1110111 @r2_vm_1 vaesz_vs 101001 1 ..... 00111 010 ..... 1110111 @r2_vm_1
target/riscv/insn_trans/trans_rvzvkned.c.inc +1 -0
@@ -96,3 +96,4 @@ GEN_V_UNMASKED_TRANS(vaesdm_vv, vaes_check_vv) GEN_V_UNMASKED_TRANS(vaesdm_vs, vaes_check_vs) GEN_V_UNMASKED_TRANS(vaesz_vs, vaes_check_vs) GEN_V_UNMASKED_TRANS(vaesem_vv, vaes_check_vv) +GEN_V_UNMASKED_TRANS(vaesem_vs, vaes_check_vs)
target/riscv/vcrypto_helper.c +3 -0
@@ -336,6 +336,9 @@ GEN_ZVKNED_HELPER_VS(vaesdf_vs, aes_inv_shift_bytes(round_state); GEN_ZVKNED_HELPER_VV(vaesem_vv, aes_shift_bytes(round_state);aes_sub_bytes (round_state);aes_mix_cols (round_state); xor_round_key(round_state, (uint8_t *)round_key);) +GEN_ZVKNED_HELPER_VS(vaesem_vs, aes_shift_bytes(round_state); +aes_sub_bytes (round_state);aes_mix_cols (round_state); + xor_round_key(round_state, (uint8_t *)round_key);) GEN_ZVKNED_HELPER_VV(vaesdm_vv, aes_inv_shift_bytes(round_state);aes_inv_sub_bytes (round_state);xor_round_key (round_state, (uint8_t *)round_key);
Signature · 1 line
2.39.2
PATCH 26/45 target/riscv: Add vaeskf1.vi decoding, translation and execution support
In reply to Lawrence HunterFrom: Nazar Kazakov <[email protected]>
--- target/riscv/helper.h | 1 + target/riscv/insn32.decode | 1 + target/riscv/insn_trans/trans_rvzvkned.c.inc | 58 ++++++++++++++++++++ target/riscv/vcrypto_helper.c | 44 +++++++++++++++ 4 files changed, 104 insertions(+)
target/riscv/helper.h +1 -0
@@ -1197,3 +1197,4 @@ DEF_HELPER_4(vaesem_vs, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesdm_vv, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesdm_vs, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesz_vs, void, ptr, ptr, env, i32) +DEF_HELPER_5 (vaeskf1_vi, void, ptr, ptr, i32, env, i32)
target/riscv/insn32.decode +1 -0
@@ -937,3 +937,4 @@ vaesem_vs 101001 1 ..... 00010 010 ..... 1110111 @r2_vm_1 vaesdm_vv 101000 1 ..... 00000 010 ..... 1110111 @r2_vm_1 vaesdm_vs 101001 1 ..... 00000 010 ..... 1110111 @r2_vm_1 vaesz_vs 101001 1 ..... 00111 010 ..... 1110111 @r2_vm_1 +vaeskf1_vi 100010 1 ..... ..... 010 ..... 1110111 @r_vm_1
target/riscv/insn_trans/trans_rvzvkned.c.inc +58 -0
61 lines not on this page · raw message
target/riscv/vcrypto_helper.c +44 -0
47 lines not on this page · raw message
Signature · 1 line
2.39.2
PATCH 27/45 target/riscv: Add vaeskf2.vi decoding, translation and execution support
In reply to Lawrence HunterFrom: Nazar Kazakov <[email protected]>
--- target/riscv/helper.h | 1 + target/riscv/insn32.decode | 1 + target/riscv/insn_trans/trans_rvzvkned.c.inc | 13 +++++ target/riscv/vcrypto_helper.c | 59 ++++++++++++++++++++ 4 files changed, 74 insertions(+)
target/riscv/helper.h +1 -0
@@ -1198,3 +1198,4 @@ DEF_HELPER_4(vaesdm_vv, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesdm_vs, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesz_vs, void, ptr, ptr, env, i32) DEF_HELPER_5(vaeskf1_vi, void, ptr, ptr, i32, env, i32) +DEF_HELPER_5 (vaeskf2_vi, void, ptr, ptr, i32, env, i32)
target/riscv/insn32.decode +1 -0
@@ -938,3 +938,4 @@ vaesdm_vv 101000 1 ..... 00000 010 ..... 1110111 @r2_vm_1 vaesdm_vs 101001 1 ..... 00000 010 ..... 1110111 @r2_vm_1 vaesz_vs 101001 1 ..... 00111 010 ..... 1110111 @r2_vm_1 vaeskf1_vi 100010 1 ..... ..... 010 ..... 1110111 @r_vm_1 +vaeskf2_vi 101010 1 ..... ..... 010 ..... 1110111 @r_vm_1
target/riscv/insn_trans/trans_rvzvkned.c.inc +13 -0
@@ -154,4 +154,17 @@ static bool vaeskf1_check(DisasContext *s, arg_vaeskf1_vi * a) require_align(a->rs2, s->lmul); } +static bool vaeskf2_check(DisasContext *s, arg_vaeskf2_vi *a) +{ + return s->cfg_ptr->ext_zvkned == true && + require_rvv(s) && + vext_check_isa_ill(s) && + MAXSZ(s) >= (128 / 8) && /* EGW in bytes */ + s->vstart % 4 == 0 && + s->sew == MO_32 && + require_align(a->rd, s->lmul) && + require_align(a->rs2, s->lmul); +} + GEN_VI_UNMASKED_TRANS(vaeskf1_vi, vaeskf1_check, 4) +GEN_VI_UNMASKED_TRANS(vaeskf2_vi, vaeskf2_check, 4)
target/riscv/vcrypto_helper.c +59 -0
62 lines not on this page · raw message
Signature · 1 line
2.39.2
PATCH 28/45 target/riscv: Expose zvkned cpu property
In reply to Lawrence HunterFrom: Nazar Kazakov <[email protected]>
--- target/riscv/cpu.c | 1 + 1 file changed, 1 insertion(+)
target/riscv/cpu.c +1 -0
@@ -1472,6 +1472,7 @@ static Property riscv_cpu_extensions[] = {/* Vector cryptography extensions */ DEFINE_PROP_BOOL ("x-zvkb" ,RISCVCPU , cfg.ext_zvkb, false), +DEFINE_PROP_BOOL ("x-zvkned" ,RISCVCPU , cfg.ext_zvkned, false), DEFINE_PROP_END_OF_LIST(), };
Signature · 1 line
2.39.2
PATCH 29/45 target/riscv: Add zvknh cpu properties
In reply to Lawrence HunterFrom: Nazar Kazakov <[email protected]>
--- target/riscv/cpu.c | 11 ++++++++++- target/riscv/cpu.h | 2 ++ 2 files changed, 12 insertions(+), 1 deletion(-)
target/riscv/cpu.c +10 -1
@@ -111,6 +111,8 @@ static const struct isa_ext_data isa_edata_arr[] = {ISA_EXT_DATA_ENTRY (zvfhmin, true,PRIV_VERSION_1_12_0 , ext_zvfhmin),ISA_EXT_DATA_ENTRY (zvkb, true,PRIV_VERSION_1_12_0 , ext_zvkb),ISA_EXT_DATA_ENTRY (zvkned, true,PRIV_VERSION_1_12_0 , ext_zvkned), +ISA_EXT_DATA_ENTRY (zvknha, true,PRIV_VERSION_1_12_0 , ext_zvknha), +ISA_EXT_DATA_ENTRY (zvknhb, true,PRIV_VERSION_1_12_0 , ext_zvknhb),ISA_EXT_DATA_ENTRY (zhinx, true,PRIV_VERSION_1_12_0 , ext_zhinx),ISA_EXT_DATA_ENTRY (zhinxmin, true,PRIV_VERSION_1_12_0 , ext_zhinxmin), ISA_EXT_DATA_ENTRY(smaia, true, PRIV_VERSION_1_12_0, ext_smaia),
@@ -1217,7 +1219,7 @@ static void riscv_cpu_realize(DeviceState *dev, Error **errp)* In principle zve*x would also suffice here, were they supported * in qemu */ -if ((cpu->cfg.ext_zvkb || cpu->cfg.ext_zvkned) && +if ((cpu->cfg.ext_zvkb || cpu->cfg.ext_zvkned || cpu->cfg.ext_zvknha) && !(cpu->cfg.ext_zve32f || cpu->cfg.ext_zve64f || cpu->cfg.ext_zve64d || cpu->cfg.ext_v)) { error_setg(
@@ -1225,6 +1227,13 @@ static void riscv_cpu_realize(DeviceState *dev, Error **errp)return ; } +if (cpu->cfg.ext_zvknhb && + !(cpu->cfg.ext_zve64f || cpu->cfg.ext_zve64d || cpu->cfg.ext_v)) { +error_setg (errp, +"Zvknhb extension requires V or Zve64{f,d} extensions" ); +return ; + } +#ifndef CONFIG_USER_ONLY if (cpu->cfg.pmu_num) {if (!riscv_pmu_init (cpu, cpu->cfg.pmu_num) && cpu->cfg.ext_sscofpmf) {
target/riscv/cpu.h +2 -0
@@ -472,6 +472,8 @@ struct RISCVCPUConfig {bool ext_zve64d;bool ext_zvkb;bool ext_zvkned; +bool ext_zvknha; +bool ext_zvknhb;bool ext_zmmul;bool ext_zvfh;bool ext_zvfhmin;
Signature · 1 line
2.39.2
PATCH 30/45 target/riscv: Add vsha2ms.vv decoding, translation and execution support
In reply to Lawrence HunterFrom: Kiran Ostrolenk <[email protected]>
--- target/riscv/helper.h | 2 + target/riscv/insn32.decode | 3 + target/riscv/insn_trans/trans_rvzvknh.c.inc | 82 +++++++++++++++++++++ target/riscv/translate.c | 1 + target/riscv/vcrypto_helper.c | 74 +++++++++++++++++++ 5 files changed, 162 insertions(+) create mode 100644 target/riscv/insn_trans/trans_rvzvknh.c.inc
target/riscv/helper.h +2 -0
@@ -1199,3 +1199,5 @@ DEF_HELPER_4(vaesdm_vs, void, ptr, ptr, env, i32) DEF_HELPER_4(vaesz_vs, void, ptr, ptr, env, i32) DEF_HELPER_5(vaeskf1_vi, void, ptr, ptr, i32, env, i32) DEF_HELPER_5(vaeskf2_vi, void, ptr, ptr, i32, env, i32) + +DEF_HELPER_5 (vsha2ms_vv, void, ptr, ptr, ptr, env, i32)
target/riscv/insn32.decode +3 -0
@@ -939,3 +939,6 @@ vaesdm_vs 101001 1 ..... 00000 010 ..... 1110111 @r2_vm_1 vaesz_vs 101001 1 ..... 00111 010 ..... 1110111 @r2_vm_1 vaeskf1_vi 100010 1 ..... ..... 010 ..... 1110111 @r_vm_1 vaeskf2_vi 101010 1 ..... ..... 010 ..... 1110111 @r_vm_1 + +# *** RV64 Zvknh vector crypto extension *** +vsha2ms_vv 101101 1 ..... ..... 010 ..... 1110111 @r_vm_1
target/riscv/insn_trans/trans_rvzvknh.c.inc new file +82 -0
82 lines not on this page · raw message
target/riscv/translate.c +1 -0
@@ -1085,6 +1085,7 @@ static uint32_t opcode_at(DisasContextBase *dcbase, target_ulong pc)#include "insn_trans/trans_rvk.c.inc" #include "insn_trans/trans_rvzvkb.c.inc" #include "insn_trans/trans_rvzvkned.c.inc" +#include "insn_trans/trans_rvzvknh.c.inc" #include "insn_trans/trans_privileged.c.inc" #include "insn_trans/trans_svinval.c.inc" #include "decode-xthead.c.inc"
target/riscv/vcrypto_helper.c +74 -0
77 lines not on this page · raw message
Signature · 1 line
2.39.2
PATCH 31/45 target/riscv: Add vsha2c[hl].vv decoding, translation and execution support
In reply to Lawrence Hunter--- target/riscv/helper.h | 2 + target/riscv/insn32.decode | 2 + target/riscv/insn_trans/trans_rvzvknh.c.inc | 2 + target/riscv/vcrypto_helper.c | 140 ++++++++++++++++++++ 4 files changed, 146 insertions(+)
target/riscv/helper.h +2 -0
@@ -1201,3 +1201,5 @@ DEF_HELPER_5(vaeskf1_vi, void, ptr, ptr, i32, env, i32) DEF_HELPER_5(vaeskf2_vi, void, ptr, ptr, i32, env, i32) DEF_HELPER_5(vsha2ms_vv, void, ptr, ptr, ptr, env, i32) +DEF_HELPER_5(vsha2ch_vv, void, ptr, ptr, ptr, env, i32) +DEF_HELPER_5 (vsha2cl_vv, void, ptr, ptr, ptr, env, i32)
target/riscv/insn32.decode +2 -0
@@ -942,3 +942,5 @@ vaeskf2_vi 101010 1 ..... ..... 010 ..... 1110111 @r_vm_1 # *** RV64 Zvknh vector crypto extension *** vsha2ms_vv 101101 1 ..... ..... 010 ..... 1110111 @r_vm_1 +vsha2ch_vv 101110 1 ..... ..... 010 ..... 1110111 @r_vm_1 +vsha2cl_vv 101111 1 ..... ..... 010 ..... 1110111 @r_vm_1
target/riscv/insn_trans/trans_rvzvknh.c.inc +2 -0
@@ -80,3 +80,5 @@ static bool vsha_check(DisasContext *s, arg_rmrr *a) } GEN_VV_UNMASKED_TRANS(vsha2ms_vv, vsha_check, 4) +GEN_VV_UNMASKED_TRANS(vsha2cl_vv, vsha_check, 4) +GEN_VV_UNMASKED_TRANS(vsha2ch_vv, vsha_check, 4)
target/riscv/vcrypto_helper.c +140 -0
143 lines not on this page · raw message
Signature · 1 line
2.39.2
PATCH 32/45 target/riscv: Expose zvknh cpu properties
In reply to Lawrence HunterFrom: Nazar Kazakov <[email protected]>
--- target/riscv/cpu.c | 2 ++ 1 file changed, 2 insertions(+)
target/riscv/cpu.c +2 -0
@@ -1482,6 +1482,8 @@ static Property riscv_cpu_extensions[] = {/* Vector cryptography extensions */ DEFINE_PROP_BOOL ("x-zvkb" ,RISCVCPU , cfg.ext_zvkb, false),DEFINE_PROP_BOOL ("x-zvkned" ,RISCVCPU , cfg.ext_zvkned, false), +DEFINE_PROP_BOOL ("x-zvknha" ,RISCVCPU , cfg.ext_zvknha, false), +DEFINE_PROP_BOOL ("x-zvknhb" ,RISCVCPU , cfg.ext_zvknhb, false), DEFINE_PROP_END_OF_LIST(), };
Signature · 1 line
2.39.2
PATCH 33/45 target/riscv: Add zvksh cpu property
In reply to Lawrence HunterFrom: Kiran Ostrolenk <[email protected]>
--- target/riscv/cpu.c | 4 +++- target/riscv/cpu.h | 1 + 2 files changed, 4 insertions(+), 1 deletion(-)
target/riscv/cpu.c +3 -1
@@ -113,6 +113,7 @@ static const struct isa_ext_data isa_edata_arr[] = {ISA_EXT_DATA_ENTRY (zvkned, true,PRIV_VERSION_1_12_0 , ext_zvkned),ISA_EXT_DATA_ENTRY (zvknha, true,PRIV_VERSION_1_12_0 , ext_zvknha),ISA_EXT_DATA_ENTRY (zvknhb, true,PRIV_VERSION_1_12_0 , ext_zvknhb), +ISA_EXT_DATA_ENTRY (zvksh, true,PRIV_VERSION_1_12_0 , ext_zvksh),ISA_EXT_DATA_ENTRY (zhinx, true,PRIV_VERSION_1_12_0 , ext_zhinx),ISA_EXT_DATA_ENTRY (zhinxmin, true,PRIV_VERSION_1_12_0 , ext_zhinxmin), ISA_EXT_DATA_ENTRY(smaia, true, PRIV_VERSION_1_12_0, ext_smaia),
@@ -1219,7 +1220,8 @@ static void riscv_cpu_realize(DeviceState *dev, Error **errp)* In principle zve*x would also suffice here, were they supported * in qemu */ -if ((cpu->cfg.ext_zvkb || cpu->cfg.ext_zvkned || cpu->cfg.ext_zvknha) && +if ((cpu->cfg.ext_zvkb || cpu->cfg.ext_zvkned || cpu->cfg.ext_zvknha || + cpu->cfg.ext_zvksh) && !(cpu->cfg.ext_zve32f || cpu->cfg.ext_zve64f || cpu->cfg.ext_zve64d || cpu->cfg.ext_v)) { error_setg(
target/riscv/cpu.h +1 -0
@@ -474,6 +474,7 @@ struct RISCVCPUConfig {bool ext_zvkned;bool ext_zvknha;bool ext_zvknhb; +bool ext_zvksh;bool ext_zmmul;bool ext_zvfh;bool ext_zvfhmin;
Signature · 1 line
2.39.2
PATCH 34/45 target/riscv: Add vsm3me.vv decoding, translation and execution support
In reply to Lawrence Hunter--- target/riscv/helper.h | 2 ++ target/riscv/insn32.decode | 3 ++ target/riscv/insn_trans/trans_rvzvksh.c.inc | 37 +++++++++++++++++++ target/riscv/translate.c | 1 + target/riscv/vcrypto_helper.c | 39 +++++++++++++++++++++ 5 files changed, 82 insertions(+) create mode 100644 target/riscv/insn_trans/trans_rvzvksh.c.inc
target/riscv/helper.h +2 -0
@@ -1203,3 +1203,5 @@ DEF_HELPER_5(vaeskf2_vi, void, ptr, ptr, i32, env, i32) DEF_HELPER_5(vsha2ms_vv, void, ptr, ptr, ptr, env, i32) DEF_HELPER_5(vsha2ch_vv, void, ptr, ptr, ptr, env, i32) DEF_HELPER_5(vsha2cl_vv, void, ptr, ptr, ptr, env, i32) + +DEF_HELPER_5 (vsm3me_vv, void, ptr, ptr, ptr, env, i32)
target/riscv/insn32.decode +3 -0
@@ -944,3 +944,6 @@ vaeskf2_vi 101010 1 ..... ..... 010 ..... 1110111 @r_vm_1 vsha2ms_vv 101101 1 ..... ..... 010 ..... 1110111 @r_vm_1 vsha2ch_vv 101110 1 ..... ..... 010 ..... 1110111 @r_vm_1 vsha2cl_vv 101111 1 ..... ..... 010 ..... 1110111 @r_vm_1 + +# *** RV64 Zvksh vector crypto extension *** +vsm3me_vv 100000 1 ..... ..... 010 ..... 1110111 @r_vm_1
target/riscv/insn_trans/trans_rvzvksh.c.inc new file +37 -0
@@ -0,0 +1,37 @@ +/* + * RISC-V translation routines for the Zvksh Extension. + * + * Copyright (C) 2023 SiFive, Inc. + * Written by Codethink Ltd and SiFive. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2 or later, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program. If not, see <http://www.gnu.org/licenses/>. + */ + +static inline bool vsm3_check(DisasContext *s, arg_rmrr *a) +{ + int mult = 1 << MAX(s->lmul, 0); + return s->cfg_ptr->ext_zvksh == true && + require_rvv(s) && + vext_check_isa_ill(s) && + !is_overlapped(a->rd, mult, a->rs2, mult) && + MAXSZ(s) >= (256 / 8) && /* EGW in bytes */ + s->vstart % 8 == 0 && + s->sew == MO_32; +} + +static inline bool vsm3me_check(DisasContext *s, arg_rmrr *a) +{ + return vsm3_check(s, a) && vext_check_sss(s, a->rd, a->rs1, a->rs2, a->vm); +} + +GEN_VV_UNMASKED_TRANS(vsm3me_vv, vsm3me_check, 8)
target/riscv/translate.c +1 -0
@@ -1086,6 +1086,7 @@ static uint32_t opcode_at(DisasContextBase *dcbase, target_ulong pc)#include "insn_trans/trans_rvzvkb.c.inc" #include "insn_trans/trans_rvzvkned.c.inc" #include "insn_trans/trans_rvzvknh.c.inc" +#include "insn_trans/trans_rvzvksh.c.inc" #include "insn_trans/trans_privileged.c.inc" #include "insn_trans/trans_svinval.c.inc" #include "decode-xthead.c.inc"
target/riscv/vcrypto_helper.c +39 -0
@@ -666,3 +666,42 @@ void HELPER(vsha2cl_vv)(void *vd, void *vs1, void *vs2, CPURISCVState *env,vext_set_elems_1s (vd, vta, env->vl * esz, total_elems * esz); env->vstart =0 ; } + +static inline uint32_t p1 (uint32_t x) +{ +return x ^rol32 (x,15 ) ^rol32 (x,23 ); +} + +static inline uint32_t zvksh_w (uint32_t m16,uint32_t m9,uint32_t m3, +uint32_t m13,uint32_t m6) +{ +return p1 (m16 ^ m9 ^rol32 (m3,15 )) ^rol32 (m13,7 ) ^ m6; +} + +void HELPER (vsm3me_vv )(void *vd_vptr,void *vs1_vptr,void *vs2_vptr, +CPURISCVState *env,uint32_t desc) +{ +uint32_t esz =memop_size (FIELD_EX64 (env->vtype,VTYPE ,VSEW )); +uint32_t total_elems =vext_get_total_elems (env, desc, esz); +uint32_t vta =vext_vta (desc); +uint32_t *vd = vd_vptr; +uint32_t *vs1 = vs1_vptr; +uint32_t *vs2 = vs2_vptr; + +for (int i = env->vstart /8 ; i < env->vl /8 ; i++) { +uint32_t w[24 ]; +for (int j =0 ; j <8 ; j++) { + w[j] =bswap32 (vs1[H4 ((i *8 ) + j)]); + w[j +8 ] =bswap32 (vs2[H4 ((i *8 ) + j)]); + } +for (int j =0 ; j <8 ; j++) { + w[j +16 ] = +zvksh_w (w[j], w[j +7 ], w[j +13 ], w[j +3 ], w[j +10 ]); + } +for (int j =0 ; j <8 ; j++) { + vd[(i *8 ) + j] =bswap32 (w[H4 (j +16 )]); + } + } +vext_set_elems_1s (vd_vptr, vta, env->vl * esz, total_elems * esz); + env->vstart =0 ; +}
Signature · 1 line
2.39.2
PATCH 35/45 target/riscv: Add vsm3c.vi decoding, translation and execution support
In reply to Lawrence HunterFrom: Kiran Ostrolenk <[email protected]>
--- target/riscv/helper.h | 1 + target/riscv/insn32.decode | 1 + target/riscv/insn_trans/trans_rvzvksh.c.inc | 6 ++ target/riscv/vcrypto_helper.c | 95 +++++++++++++++++++++ 4 files changed, 103 insertions(+)
target/riscv/helper.h +1 -0
@@ -1205,3 +1205,4 @@ DEF_HELPER_5(vsha2ch_vv, void, ptr, ptr, ptr, env, i32) DEF_HELPER_5(vsha2cl_vv, void, ptr, ptr, ptr, env, i32) DEF_HELPER_5(vsm3me_vv, void, ptr, ptr, ptr, env, i32) +DEF_HELPER_5 (vsm3c_vi, void, ptr, ptr, i32, env, i32)
target/riscv/insn32.decode +1 -0
@@ -947,3 +947,4 @@ vsha2cl_vv 101111 1 ..... ..... 010 ..... 1110111 @r_vm_1 # *** RV64 Zvksh vector crypto extension *** vsm3me_vv 100000 1 ..... ..... 010 ..... 1110111 @r_vm_1 +vsm3c_vi 101011 1 ..... ..... 010 ..... 1110111 @r_vm_1
target/riscv/insn_trans/trans_rvzvksh.c.inc +6 -0
@@ -34,4 +34,10 @@ static inline bool vsm3me_check(DisasContext *s, arg_rmrr *a) return vsm3_check(s, a) && vext_check_sss(s, a->rd, a->rs1, a->rs2, a->vm); } +static inline bool vsm3c_check(DisasContext *s, arg_rmrr *a) +{ + return vsm3_check(s, a) && vext_check_ss(s, a->rd, a->rs2, a->vm); +} + GEN_VV_UNMASKED_TRANS(vsm3me_vv, vsm3me_check, 8) +GEN_VI_UNMASKED_TRANS(vsm3c_vi, vsm3c_check, 8)
target/riscv/vcrypto_helper.c +95 -0
98 lines not on this page · raw message
Signature · 1 line
2.39.2
PATCH 36/45 target/riscv: Expose zvksh cpu property
In reply to Lawrence HunterFrom: Kiran Ostrolenk <[email protected]>
--- target/riscv/cpu.c | 1 + 1 file changed, 1 insertion(+)
target/riscv/cpu.c +1 -0
@@ -1486,6 +1486,7 @@ static Property riscv_cpu_extensions[] = {DEFINE_PROP_BOOL ("x-zvkned" ,RISCVCPU , cfg.ext_zvkned, false),DEFINE_PROP_BOOL ("x-zvknha" ,RISCVCPU , cfg.ext_zvknha, false),DEFINE_PROP_BOOL ("x-zvknhb" ,RISCVCPU , cfg.ext_zvknhb, false), +DEFINE_PROP_BOOL ("x-zvksh" ,RISCVCPU , cfg.ext_zvksh, false), DEFINE_PROP_END_OF_LIST(), };
Signature · 1 line
2.39.2
PATCH 37/45 target/riscv: Add zvkg cpu property
In reply to Lawrence Hunter--- target/riscv/cpu.c | 5 +++-- target/riscv/cpu.h | 1 + 2 files changed, 4 insertions(+), 2 deletions(-)
target/riscv/cpu.c +3 -2
@@ -110,6 +110,7 @@ static const struct isa_ext_data isa_edata_arr[] = {ISA_EXT_DATA_ENTRY (zvfh, true,PRIV_VERSION_1_12_0 , ext_zvfh),ISA_EXT_DATA_ENTRY (zvfhmin, true,PRIV_VERSION_1_12_0 , ext_zvfhmin),ISA_EXT_DATA_ENTRY (zvkb, true,PRIV_VERSION_1_12_0 , ext_zvkb), +ISA_EXT_DATA_ENTRY (zvkg, true,PRIV_VERSION_1_12_0 , ext_zvkg),ISA_EXT_DATA_ENTRY (zvkned, true,PRIV_VERSION_1_12_0 , ext_zvkned),ISA_EXT_DATA_ENTRY (zvknha, true,PRIV_VERSION_1_12_0 , ext_zvknha), ISA_EXT_DATA_ENTRY(zvknhb, true, PRIV_VERSION_1_12_0, ext_zvknhb),
@@ -1220,8 +1221,8 @@ static void riscv_cpu_realize(DeviceState *dev, Error **errp)* In principle zve*x would also suffice here, were they supported * in qemu */ -if ((cpu->cfg.ext_zvkb || cpu->cfg.ext_zvkned || cpu->cfg.ext_zvknha || - cpu->cfg.ext_zvksh) && +if ((cpu->cfg.ext_zvkb || cpu->cfg.ext_zvkg || cpu->cfg.ext_zvkned || + cpu->cfg.ext_zvknha || cpu->cfg.ext_zvksh) && !(cpu->cfg.ext_zve32f || cpu->cfg.ext_zve64f || cpu->cfg.ext_zve64d || cpu->cfg.ext_v)) { error_setg(
target/riscv/cpu.h +1 -0
@@ -471,6 +471,7 @@ struct RISCVCPUConfig {bool ext_zve64f;bool ext_zve64d;bool ext_zvkb; +bool ext_zvkg;bool ext_zvkned;bool ext_zvknha;bool ext_zvknhb;
Signature · 1 line
2.39.2
PATCH 38/45 target/riscv: Add vgmul.vv decoding, translation and execution support
In reply to Lawrence Hunter--- target/riscv/helper.h | 2 ++ target/riscv/insn32.decode | 3 ++ target/riscv/insn_trans/trans_rvzvkg.c.inc | 30 +++++++++++++++++++ target/riscv/translate.c | 1 + target/riscv/vcrypto_helper.c | 34 ++++++++++++++++++++++ 5 files changed, 70 insertions(+) create mode 100644 target/riscv/insn_trans/trans_rvzvkg.c.inc
target/riscv/helper.h +2 -0
@@ -1206,3 +1206,5 @@ DEF_HELPER_5(vsha2cl_vv, void, ptr, ptr, ptr, env, i32) DEF_HELPER_5(vsm3me_vv, void, ptr, ptr, ptr, env, i32) DEF_HELPER_5(vsm3c_vi, void, ptr, ptr, i32, env, i32) + +DEF_HELPER_4 (vgmul_vv, void, ptr, ptr, env, i32)
target/riscv/insn32.decode +3 -0
@@ -948,3 +948,6 @@ vsha2cl_vv 101111 1 ..... ..... 010 ..... 1110111 @r_vm_1 # *** RV64 Zvksh vector crypto extension *** vsm3me_vv 100000 1 ..... ..... 010 ..... 1110111 @r_vm_1 vsm3c_vi 101011 1 ..... ..... 010 ..... 1110111 @r_vm_1 + +# *** RV64 Zvkg vector crypto extension *** +vgmul_vv 101000 1 ..... 10001 010 ..... 1110111 @r2_vm_1
target/riscv/insn_trans/trans_rvzvkg.c.inc new file +30 -0
@@ -0,0 +1,30 @@ +/* + * RISC-V translation routines for the Zvkg Extension. + * + * Copyright (C) 2023 SiFive, Inc. + * Written by Codethink Ltd and SiFive. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2 or later, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program. If not, see <http://www.gnu.org/licenses/>. + */ + +static bool vgmul_check(DisasContext *s, arg_rmr *a) +{ + return s->cfg_ptr->ext_zvkg == true && + vext_check_isa_ill(s) && + require_rvv(s) && + MAXSZ(s) >= (128 / 8) && /* EGW in bytes */ + vext_check_ss(s, a->rd, a->rs2, a->vm) && + s->vstart % 4 == 0 && s->sew == MO_32; +} + +GEN_V_UNMASKED_TRANS(vgmul_vv, vgmul_check)
target/riscv/translate.c +1 -0
@@ -1087,6 +1087,7 @@ static uint32_t opcode_at(DisasContextBase *dcbase, target_ulong pc)#include "insn_trans/trans_rvzvkned.c.inc" #include "insn_trans/trans_rvzvknh.c.inc" #include "insn_trans/trans_rvzvksh.c.inc" +#include "insn_trans/trans_rvzvkg.c.inc" #include "insn_trans/trans_privileged.c.inc" #include "insn_trans/trans_svinval.c.inc" #include "decode-xthead.c.inc"
target/riscv/vcrypto_helper.c +34 -0
@@ -800,3 +800,37 @@ void HELPER(vsm3c_vi)(void *vd_vptr, void *vs2_vptr, uint32_t uimm,vext_set_elems_1s (vd_vptr, vta, env->vl * esz, total_elems * esz); env->vstart =0 ; } + +void HELPER (vgmul_vv )(void *vd_vptr,void *vs2_vptr, +CPURISCVState *env,uint32_t desc) +{ +uint64_t *vd = vd_vptr; +uint64_t *vs2 = vs2_vptr; +uint32_t vta =vext_vta (desc); +uint32_t total_elems =vext_get_total_elems (env, desc,4 ); + +for (uint32_t i = env->vstart /4 ; i < env->vl /4 ; i++) { +uint64_t Y [2 ] = {brev8 (vd[i *2 +0 ]),brev8 (vd[i *2 +1 ])}; +uint64_t H [2 ] = {brev8 (vs2[i *2 +0 ]),brev8 (vs2[i *2 +1 ])}; +uint64_t Z [2 ] = {0 ,0 }; + +for (uint j =0 ; j <128 ; j++) { +if ((Y [j /64 ] >> (j %64 )) &1 ) { +Z [0 ] ^=H [0 ]; +Z [1 ] ^=H [1 ]; + } +bool reduce = ((H [1 ] >>63 ) &1 ); +H [1 ] =H [1 ] <<1 |H [0 ] >>63 ; +H [0 ] =H [0 ] <<1 ; +if (reduce) { +H [0 ] ^=0x87 ; + } + } + + vd[i *2 +0 ] =brev8 (Z [0 ]); + vd[i *2 +1 ] =brev8 (Z [1 ]); + } +/* set tail elements to 1s */ +vext_set_elems_1s (vd, vta, env->vl *4 , total_elems *4 ); + env->vstart =0 ; +}
Signature · 1 line
2.39.2
PATCH 39/45 target/riscv: Add vghsh.vv decoding, translation and execution support
In reply to Lawrence Hunter--- target/riscv/helper.h | 1 + target/riscv/insn32.decode | 1 + target/riscv/insn_trans/trans_rvzvkg.c.inc | 10 ++++++ target/riscv/vcrypto_helper.c | 38 ++++++++++++++++++++++ 4 files changed, 50 insertions(+)
target/riscv/helper.h +1 -0
@@ -1207,4 +1207,5 @@ DEF_HELPER_5(vsha2cl_vv, void, ptr, ptr, ptr, env, i32) DEF_HELPER_5(vsm3me_vv, void, ptr, ptr, ptr, env, i32) DEF_HELPER_5(vsm3c_vi, void, ptr, ptr, i32, env, i32) +DEF_HELPER_5(vghsh_vv, void, ptr, ptr, ptr, env, i32)DEF_HELPER_4 (vgmul_vv, void, ptr, ptr, env, i32)
target/riscv/insn32.decode +1 -0
@@ -950,4 +950,5 @@ vsm3me_vv 100000 1 ..... ..... 010 ..... 1110111 @r_vm_1 vsm3c_vi 101011 1 ..... ..... 010 ..... 1110111 @r_vm_1 # *** RV64 Zvkg vector crypto extension *** +vghsh_vv 101100 1 ..... ..... 010 ..... 1110111 @r_vm_1 vgmul_vv 101000 1 ..... 10001 010 ..... 1110111 @r2_vm_1
target/riscv/insn_trans/trans_rvzvkg.c.inc +10 -0
@@ -28,3 +28,13 @@ static bool vgmul_check(DisasContext *s, arg_rmr *a) } GEN_V_UNMASKED_TRANS(vgmul_vv, vgmul_check) + +static bool vghsh_check(DisasContext *s, arg_rmrr *a) +{ + return s->cfg_ptr->ext_zvkg == true && + opivv_check(s, a) && + MAXSZ(s) >= (128 / 8) && /* EGW in bytes */ + s->vstart % 4 == 0 && s->sew == MO_32; +} + +GEN_VV_UNMASKED_TRANS(vghsh_vv, vghsh_check, 4)
target/riscv/vcrypto_helper.c +38 -0
@@ -801,6 +801,44 @@ void HELPER(vsm3c_vi)(void *vd_vptr, void *vs2_vptr, uint32_t uimm, env->vstart =0 ; } +void HELPER (vghsh_vv )(void *vd_vptr,void *vs1_vptr,void *vs2_vptr, +CPURISCVState *env,uint32_t desc) +{ +uint64_t *vd = vd_vptr; +uint64_t *vs1 = vs1_vptr; +uint64_t *vs2 = vs2_vptr; +uint32_t vta =vext_vta (desc); +uint32_t total_elems =vext_get_total_elems (env, desc,4 ); + +for (uint32_t i = env->vstart /4 ; i < env->vl /4 ; i++) { +uint64_t Y [2 ] = {vd[i *2 +0 ], vd[i *2 +1 ]}; +uint64_t H [2 ] = {brev8 (vs2[i *2 +0 ]),brev8 (vs2[i *2 +1 ])}; +uint64_t X [2 ] = {vs1[i *2 +0 ], vs1[i *2 +1 ]}; +uint64_t Z [2 ] = {0 ,0 }; + +uint64_t S [2 ] = {brev8 (Y [0 ] ^X [0 ]),brev8 (Y [1 ] ^X [1 ])}; + +for (uint j =0 ; j <128 ; j++) { +if ((S [j /64 ] >> (j %64 )) &1 ) { +Z [0 ] ^=H [0 ]; +Z [1 ] ^=H [1 ]; + } +bool reduce = ((H [1 ] >>63 ) &1 ); +H [1 ] =H [1 ] <<1 |H [0 ] >>63 ; +H [0 ] =H [0 ] <<1 ; +if (reduce) { +H [0 ] ^=0x87 ; + } + } + + vd[i *2 +0 ] =brev8 (Z [0 ]); + vd[i *2 +1 ] =brev8 (Z [1 ]); + } +/* set tail elements to 1s */ +vext_set_elems_1s (vd, vta, env->vl *4 , total_elems *4 ); + env->vstart =0 ; +} +void HELPER (vgmul_vv )(void *vd_vptr,void *vs2_vptr,CPURISCVState *env,uint32_t desc) {
Signature · 1 line
2.39.2
PATCH 40/45 target/riscv: Expose zvkg cpu property
In reply to Lawrence Hunter--- target/riscv/cpu.c | 1 + 1 file changed, 1 insertion(+)
target/riscv/cpu.c +1 -0
@@ -1484,6 +1484,7 @@ static Property riscv_cpu_extensions[] = {/* Vector cryptography extensions */ DEFINE_PROP_BOOL ("x-zvkb" ,RISCVCPU , cfg.ext_zvkb, false), +DEFINE_PROP_BOOL ("x-zvkg" ,RISCVCPU , cfg.ext_zvkg, false),DEFINE_PROP_BOOL ("x-zvkned" ,RISCVCPU , cfg.ext_zvkned, false),DEFINE_PROP_BOOL ("x-zvknha" ,RISCVCPU , cfg.ext_zvknha, false), DEFINE_PROP_BOOL("x-zvknhb", RISCVCPU, cfg.ext_zvknhb, false),
Signature · 1 line
2.39.2
PATCH 41/45 crypto: Create sm4_subword
In reply to Lawrence HunterFrom: Max Chou <[email protected]>
- Share sm4_subword between different targets.
--- include/crypto/sm4.h | 8 ++++++++ target/arm/tcg/crypto_helper.c | 10 ++-------- 2 files changed, 10 insertions(+), 8 deletions(-)
include/crypto/sm4.h +8 -0
@@ -3,4 +3,12 @@extern const uint8_t sm4_sbox[256 ]; +static inline uint32_t sm4_subword (uint32_t word) +{ +return sm4_sbox[word &0xff ] | + sm4_sbox[(word >>8 ) &0xff ] <<8 | + sm4_sbox[(word >>16 ) &0xff ] <<16 | + sm4_sbox[(word >>24 ) &0xff ] <<24 ; +} +#endif
target/arm/tcg/crypto_helper.c +2 -8
@@ -707,10 +707,7 @@ static void do_crypto_sm4e(uint64_t *rd, uint64_t *rn, uint64_t *rm)CR_ST_WORD (d, (i +3 ) %4 ) ^CR_ST_WORD (n, i); - t = sm4_sbox[t &0xff ] | - sm4_sbox[(t >>8 ) &0xff ] <<8 | - sm4_sbox[(t >>16 ) &0xff ] <<16 | - sm4_sbox[(t >>24 ) &0xff ] <<24 ; + t =sm4_subword (t);CR_ST_WORD (d, i) ^= t ^rol32 (t,2 ) ^rol32 (t,10 ) ^rol32 (t,18 ) ^rol32 (t,24 );
@@ -744,10 +741,7 @@ static void do_crypto_sm4ekey(uint64_t *rd, uint64_t *rn, uint64_t *rm)CR_ST_WORD (d, (i +3 ) %4 ) ^CR_ST_WORD (m, i); - t = sm4_sbox[t &0xff ] | - sm4_sbox[(t >>8 ) &0xff ] <<8 | - sm4_sbox[(t >>16 ) &0xff ] <<16 | - sm4_sbox[(t >>24 ) &0xff ] <<24 ; + t =sm4_subword (t);CR_ST_WORD (d, i) ^= t ^rol32 (t,13 ) ^rol32 (t,23 ); }
Signature · 1 line
2.39.2
PATCH 42/45 crypto: Add SM4 constant parameter CK
In reply to Lawrence HunterFrom: Max Chou <[email protected]>
--- crypto/sm4.c | 10 ++++++++++ include/crypto/sm4.h | 1 + 2 files changed, 11 insertions(+)
crypto/sm4.c +10 -0
@@ -47,3 +47,13 @@ uint8_t const sm4_sbox[] = { 0x79, 0xee, 0x5f, 0x3e, 0xd7, 0xcb, 0x39, 0x48, }; +uint32_t const sm4_ck[] = { +0x00070e15 ,0x1c232a31 ,0x383f464d ,0x545b6269 , +0x70777e85 ,0x8c939aa1 ,0xa8afb6bd ,0xc4cbd2d9 , +0xe0e7eef5 ,0xfc030a11 ,0x181f262d ,0x343b4249 , +0x50575e65 ,0x6c737a81 ,0x888f969d ,0xa4abb2b9 , +0xc0c7ced5 ,0xdce3eaf1 ,0xf8ff060d ,0x141b2229 , +0x30373e45 ,0x4c535a61 ,0x686f767d ,0x848b9299 , +0xa0a7aeb5 ,0xbcc3cad1 ,0xd8dfe6ed ,0xf4fb0209 , +0x10171e25 ,0x2c333a41 ,0x484f565d ,0x646b7279 +};
include/crypto/sm4.h +1 -0
@@ -2,6 +2,7 @@#define QEMU_SM4_H extern const uint8_t sm4_sbox[256 ]; +extern const uint32_t sm4_ck[32 ];static inline uint32_t sm4_subword (uint32_t word) {
Signature · 1 line
2.39.2
PATCH 43/45 target/riscv: Add zvksed cfg property
In reply to Lawrence HunterFrom: Max Chou <[email protected]>
--- target/riscv/cpu.c | 3 ++- target/riscv/cpu.h | 1 + 2 files changed, 3 insertions(+), 1 deletion(-)
target/riscv/cpu.c +2 -1
@@ -114,6 +114,7 @@ static const struct isa_ext_data isa_edata_arr[] = {ISA_EXT_DATA_ENTRY (zvkned, true,PRIV_VERSION_1_12_0 , ext_zvkned),ISA_EXT_DATA_ENTRY (zvknha, true,PRIV_VERSION_1_12_0 , ext_zvknha),ISA_EXT_DATA_ENTRY (zvknhb, true,PRIV_VERSION_1_12_0 , ext_zvknhb), +ISA_EXT_DATA_ENTRY (zvksed, true,PRIV_VERSION_1_12_0 , ext_zvksed),ISA_EXT_DATA_ENTRY (zvksh, true,PRIV_VERSION_1_12_0 , ext_zvksh),ISA_EXT_DATA_ENTRY (zhinx, true,PRIV_VERSION_1_12_0 , ext_zhinx), ISA_EXT_DATA_ENTRY(zhinxmin, true, PRIV_VERSION_1_12_0, ext_zhinxmin),
@@ -1222,7 +1223,7 @@ static void riscv_cpu_realize(DeviceState *dev, Error **errp)* in qemu */ if ((cpu->cfg.ext_zvkb || cpu->cfg.ext_zvkg || cpu->cfg.ext_zvkned || - cpu->cfg.ext_zvknha || cpu->cfg.ext_zvksh) && + cpu->cfg.ext_zvknha || cpu->cfg.ext_zvksed || cpu->cfg.ext_zvksh) && !(cpu->cfg.ext_zve32f || cpu->cfg.ext_zve64f || cpu->cfg.ext_zve64d || cpu->cfg.ext_v)) { error_setg(
target/riscv/cpu.h +1 -0
@@ -475,6 +475,7 @@ struct RISCVCPUConfig {bool ext_zvkned;bool ext_zvknha;bool ext_zvknhb; +bool ext_zvksed;bool ext_zvksh;bool ext_zmmul;bool ext_zvfh;
Signature · 1 line
2.39.2
PATCH 44/45 target/riscv: Add Zvksed support
In reply to Lawrence HunterFrom: Max Chou <[email protected]>
- add vsm4k, vsm4r instructions
[[email protected]: Moved SM4 functions from crypto_helper.c to vcrypto_helper.c] [[email protected]: Added alignment checks, refactored code to use macros, and minor style changes]
--- target/riscv/helper.h | 4 + target/riscv/insn32.decode | 5 + target/riscv/insn_trans/trans_rvzvksed.c.inc | 57 +++++++++ target/riscv/translate.c | 1 + target/riscv/vcrypto_helper.c | 127 +++++++++++++++++++ 5 files changed, 194 insertions(+) create mode 100644 target/riscv/insn_trans/trans_rvzvksed.c.inc
target/riscv/helper.h +4 -0
@@ -1209,3 +1209,7 @@ DEF_HELPER_5(vsm3c_vi, void, ptr, ptr, i32, env, i32) DEF_HELPER_5(vghsh_vv, void, ptr, ptr, ptr, env, i32) DEF_HELPER_4(vgmul_vv, void, ptr, ptr, env, i32) + +DEF_HELPER_5(vsm4k_vi, void, ptr, ptr, i32, env, i32) +DEF_HELPER_4(vsm4r_vv, void, ptr, ptr, env, i32) +DEF_HELPER_4 (vsm4r_vs, void, ptr, ptr, env, i32)
target/riscv/insn32.decode +5 -0
@@ -952,3 +952,8 @@ vsm3c_vi 101011 1 ..... ..... 010 ..... 1110111 @r_vm_1 # *** RV64 Zvkg vector crypto extension *** vghsh_vv 101100 1 ..... ..... 010 ..... 1110111 @r_vm_1 vgmul_vv 101000 1 ..... 10001 010 ..... 1110111 @r2_vm_1 + +# *** RV64 Zvksed vector crypto extension *** +vsm4k_vi 100001 1 ..... ..... 010 ..... 1110111 @r_vm_1 +vsm4r_vv 101000 1 ..... 10000 010 ..... 1110111 @r2_vm_1 +vsm4r_vs 101001 1 ..... 10000 010 ..... 1110111 @r2_vm_1
target/riscv/insn_trans/trans_rvzvksed.c.inc new file +57 -0
57 lines not on this page · raw message
target/riscv/translate.c +1 -0
@@ -1088,6 +1088,7 @@ static uint32_t opcode_at(DisasContextBase *dcbase, target_ulong pc)#include "insn_trans/trans_rvzvknh.c.inc" #include "insn_trans/trans_rvzvksh.c.inc" #include "insn_trans/trans_rvzvkg.c.inc" +#include "insn_trans/trans_rvzvksed.c.inc" #include "insn_trans/trans_privileged.c.inc" #include "insn_trans/trans_svinval.c.inc" #include "decode-xthead.c.inc"
target/riscv/vcrypto_helper.c +127 -0
136 lines not on this page · raw message
Signature · 1 line
2.39.2
PATCH 45/45 target/riscv: Expose Zvksed property
In reply to Lawrence HunterFrom: Max Chou <[email protected]>
--- target/riscv/cpu.c | 1 + 1 file changed, 1 insertion(+)
target/riscv/cpu.c +1 -0
@@ -1489,6 +1489,7 @@ static Property riscv_cpu_extensions[] = {DEFINE_PROP_BOOL ("x-zvkned" ,RISCVCPU , cfg.ext_zvkned, false),DEFINE_PROP_BOOL ("x-zvknha" ,RISCVCPU , cfg.ext_zvknha, false),DEFINE_PROP_BOOL ("x-zvknhb" ,RISCVCPU , cfg.ext_zvknhb, false), +DEFINE_PROP_BOOL ("x-zvksed" ,RISCVCPU , cfg.ext_zvksed, false),DEFINE_PROP_BOOL ("x-zvksh" ,RISCVCPU , cfg.ext_zvksh, false), DEFINE_PROP_END_OF_LIST(),
Signature · 1 line
2.39.2
Quoted 6 lines · Lawrence Hunter
This patchset provides an implementation for Zvkb, Zvkned, Zvknh, Zvksh, Zvkg, and Zvksed of the draft RISC-V vector cryptography extensions as per the 20230303 version of the specification(1) (1fcbb30). Please note that the Zvkt data-independent execution latency extension has not been implemented, and we would recommend not using these patches in an environment where timing attacks are an issue.
Work performed by Dickon, Lawrence, Nazar, Kiran, and William from Codethink sponsored by SiFive, as well as Max Chou and Frank Chang from SiFive.
For convenience we have created a git repo with our patches on top of a recent master. https://github.com/CodethinkLabs/qemu-ct
I did test and review this patchset.
Since most of my comments affect multiple patches I have summarized them here in one email.
Observations that only affect a single patch will be sent in response to the corresponding email.
I have tested this series with the OpenSSL PR for Zvk that can be found here:
https://github.com/openssl/openssl/pull/20149
I ran with all Zvk* extensions enabled (using Zvkg for GCM) and with Zvkb only (using Zvkb for GCM).
All tests succeed. Note, however, that the test coverage is limited (e.g. no .vv instructions, vstart is always zero).
When sending out a follow-up version (even if it just introduces a minimal fix), then consider using patchset versioning (e.g. git format-patch -v2 ...).
It might be a matter of taste, but I would prefer a series that groups and orders the commits differently:
a) independent changes to the existing code (refactoring only, but
no new features) - one commit per topic
b) introduction of new functionality - one commit per extension
A series using such a commit granularity and order would be easier to maintain and review (and not result in 45 patches).
Also, the refactoring changes could land before Zvk freezes if maintainers decide to do so.
So far all translation files in target/riscv/insn_trans/* contain multiple extensions if they are related.
I think we should follow this pattern and use a common trans_zvk.c.inc file.
All patches to insn32.decode have comments of the form "RV64 Zvk* vector crypto extension".
What is the point of the "RV64"? I would simply remove that.
All instructions set "env->vstart = 0;" at the end.
I don't think that this is correct (the specification does not require this).
The tests of the reserved encodings are not consistent:
* Zvknh does a dynamic test (query tcg_gen_*())
* Zvkned does a dynamic test (tcg_gen_*())
* Zvkg does not test for (vl%EGS == 0)
The vl CSR can only be updated by the vset{i}vl{i} instructions. The same applies to the vstart CSR and the vtype CSR that holds vsew, vlmul and other fields.
The current code tests the VSTART/SEW value using "s->vstart % 4 == 0"/"s->sew == MO_32".
Why is it not possible to do the same with VL, i.e. "s->vl % 4 == 0" (after adding it to DisasContext)?
Also, I would introduce named constants or macros for the EGS values to avoid magic constants in the code
(some extensions do that - e.g. ZVKSED_EGS).
BR
Christoph
Quoted 119 lines
1. https://github.com/riscv/riscv-crypto/releases
Dickon Hood (2):
qemu/bitops.h: Limit rotate amounts
target/riscv: Add vrol.[vv,vx] and vror.[vv,vx,vi] decoding,
translation and execution supportKiran Ostrolenk (8):
target/riscv: Refactor some of the generic vector functionality target/riscv: Refactor some of the generic vector functionality target/riscv: Refactor some of the generic vector functionality target/riscv: Refactor some of the generic vector functionality target/riscv: Add vsha2ms.vv decoding, translation and execution
support
target/riscv: Add zvksh cpu property
target/riscv: Add vsm3c.vi decoding, translation and execution support target/riscv: Expose zvksh cpu propertyLawrence Hunter (17):
target/riscv: Add vclmul.vv decoding, translation and execution
support
target/riscv: Add vclmul.vx decoding, translation and execution
support
target/riscv: Add vclmulh.vv decoding, translation and execution
support
target/riscv: Add vclmulh.vx decoding, translation and execution
support
target/riscv: Add vaesef.vv decoding, translation and execution
support
target/riscv: Add vaesef.vs decoding, translation and execution
support
target/riscv: Add vaesdf.vv decoding, translation and execution
support
target/riscv: Add vaesdf.vs decoding, translation and execution
support
target/riscv: Add vaesdm.vv decoding, translation and execution
support
target/riscv: Add vaesdm.vs decoding, translation and execution
support
target/riscv: Add vaesz.vs decoding, translation and execution support target/riscv: Add vsha2c[hl].vv decoding, translation and execution
support
target/riscv: Add vsm3me.vv decoding, translation and execution
support
target/riscv: Add zvkg cpu property
target/riscv: Add vgmul.vv decoding, translation and execution support target/riscv: Add vghsh.vv decoding, translation and execution support target/riscv: Expose zvkg cpu propertyMax Chou (5):
crypto: Createsm4_subword
crypto: Add SM4 constant parameter CK
target/riscv: Add zvksed cfg property
target/riscv: Add Zvksed support
target/riscv: Expose Zvksed propertyNazar Kazakov (10):
target/riscv: Add zvkb cpu property
target/riscv: Add vrev8.v decoding, translation and execution support target/riscv: Add vandn.[vv,vx] decoding, translation and execution
support
target/riscv: Expose zvkb cpu property
target/riscv: Add zvkned cpu property
target/riscv: Add vaeskf1.vi decoding, translation and execution
support
target/riscv: Add vaeskf2.vi decoding, translation and execution
support
target/riscv: Expose zvkned cpu property
target/riscv: Add zvknh cpu properties
target/riscv: Expose zvknh cpu propertiesWilliam Salmon (3):
target/riscv: Add vbrev8.v decoding, translation and execution support target/riscv: Add vaesem.vv decoding, translation and execution
support
target/riscv: Add vaesem.vs decoding, translation and execution
supportaccel/tcg/tcg-runtime-gvec.c | 11 + accel/tcg/tcg-runtime.h | 1 + crypto/sm4.c | 10 + include/crypto/sm4.h | 9 + include/qemu/bitops.h | 24 +- target/arm/tcg/crypto_helper.c | 10 +- target/riscv/cpu.c | 36 + target/riscv/cpu.h | 7 + target/riscv/helper.h | 71 ++ target/riscv/insn32.decode | 49 + target/riscv/insn_trans/trans_rvv.c.inc | 93 +- target/riscv/insn_trans/trans_rvzvkb.c.inc | 220 ++++ target/riscv/insn_trans/trans_rvzvkg.c.inc | 40 + target/riscv/insn_trans/trans_rvzvkned.c.inc | 170 +++ target/riscv/insn_trans/trans_rvzvknh.c.inc | 84 ++ target/riscv/insn_trans/trans_rvzvksed.c.inc | 57 + target/riscv/insn_trans/trans_rvzvksh.c.inc | 43 + target/riscv/meson.build | 4 +- target/riscv/op_helper.c | 5 + target/riscv/translate.c | 6 + target/riscv/vcrypto_helper.c | 1001 ++++++++++++++++++ target/riscv/vector_helper.c | 240 +---- target/riscv/vector_internals.c | 81 ++ target/riscv/vector_internals.h | 222 ++++ 24 files changed, 2192 insertions(+), 302 deletions(-) create mode 100644 target/riscv/insn_trans/trans_rvzvkb.c.inc create mode 100644 target/riscv/insn_trans/trans_rvzvkg.c.inc create mode 100644 target/riscv/insn_trans/trans_rvzvkned.c.inc create mode 100644 target/riscv/insn_trans/trans_rvzvknh.c.inc create mode 100644 target/riscv/insn_trans/trans_rvzvksed.c.inc create mode 100644 target/riscv/insn_trans/trans_rvzvksh.c.inc create mode 100644 target/riscv/vcrypto_helper.c create mode 100644 target/riscv/vector_internals.c create mode 100644 target/riscv/vector_internals.h--
2.39.2
Quoted 16 lines · Christoph Müllner
Quoted 6 lines · Lawrence Hunter
This patchset provides an implementation for Zvkb, Zvkned, Zvknh, Zvksh, Zvkg, and Zvksed of the draft RISC-V vector cryptography extensions as per the 20230303 version of the specification(1) (1fcbb30). Please note that the Zvkt data-independent execution latency extension has not been implemented, and we would recommend not using these patches in an environment where timing attacks are an issue.
Work performed by Dickon, Lawrence, Nazar, Kiran, and William from Codethink sponsored by SiFive, as well as Max Chou and Frank Chang from SiFive.
For convenience we have created a git repo with our patches on top of a recent master. https://github.com/CodethinkLabs/qemu-ct
I did test and review this patchset.
Since most of my comments affect multiple patches I have summarized them here in one email.
Observations that only affect a single patch will be sent in response to the corresponding email.I have tested this series with the OpenSSL PR for Zvk that can be found here:
Quoted 8 lines
https://github.com/openssl/openssl/pull/20149I ran with all Zvk* extensions enabled (using Zvkg for GCM) and with Zvkb only (using Zvkb for GCM).
All tests succeed. Note, however, that the test coverage is limited (e.g. no .vv instructions, vstart is always zero).When sending out a follow-up version (even if it just introduces a minimal fix), then consider using patchset versioning (e.g. git format-patch -v2 ...).
Ok, will do
Quoted 9 lines
It might be a matter of taste, but I would prefer a series that groups and orders the commits differently:
a) independent changes to the existing code (refactoring only, but
no new features) - one commit per topic
b) introduction of new functionality - one commit per extension
A series using such a commit granularity and order would be easier to maintain and review (and not result in 45 patches).
Also, the refactoring changes could land before Zvk freezes if maintainers decide to do so.
Makes sense, will do
Quoted 3 lines
So far all translation files in target/riscv/insn_trans/* contain multiple extensions if they are related.
I think we should follow this pattern and use a common trans_zvk.c.inc file.
Agree, will do
Quoted 3 lines
All patches to insn32.decode have comments of the form "RV64 Zvk* vector crypto extension".
What is the point of the "RV64"? I would simply remove that.
Ok, will remove it
Quoted 2 lines
All instructions set "
env->vstart= 0;" at the end.
I don't think that this is correct (the specification does not require this).
That's from vector spec: "All vector instructions are defined to begin execution with the element number given in the vstart CSR, leaving earlier elements in the destination vector undisturbed, and to reset the vstart CSR to zero at the end of execution." - from "3.7. Vector Start Index CSR vstart"
Quoted 4 lines
The tests of the reserved encodings are not consistent:
* Zvknh does a dynamic test (query tcg_gen_*())
* Zvkned does a dynamic test (tcg_gen_*())
* Zvkg does not test for (vl%EGS == 0)
Zvkg also does dynamic test, by calling macros GEN_V_UNMASKED_TRANS and GEN_VV_UNMASKED_TRANS
Quoted 7 lines
The vl CSR can only be updated by the vset{i}vl{i} instructions. The same applies to the vstart CSR and the vtype CSR that holds vsew, vlmul and other fields.
The current code tests the VSTART/SEW value using "s->vstart% 4 == 0"/"s->sew ==MO_32".
Why is it not possible to do the same with VL, i.e. "s->vl% 4 == 0" (after adding it to DisasContext)?
vl can also be updated by another instruction - from vector spec "3.5. Vector Length Register vl" - "The XLEN-bit-wide read-only vl CSR can only be updated by the vset{i}vl{i} instructions, and the fault-only-first vector load instruction variants." So just because of that fault-only-first instruction we need dynamic checks.
vstart is just another CSR -- software can write to it, but probably shouldn't. Whether that's ever going to be useful outside testing ISA conformance tests or not I don't know, but it's clearly read-write (also section 3.7).
Quoted 3 lines
Also, I would introduce named constants or macros for the EGS values to avoid magic constants in the code
(some extensions do that - e.g.ZVKSED_EGS).
Makes sense, will do
Best,
Lawrence
Quoted 70 lines · Lawrence Hunter
Quoted 32 lines · Christoph Müllner
Quoted 9 lines · Lawrence Hunter
This patchset provides an implementation for Zvkb, Zvkned, Zvknh, Zvksh, Zvkg, and Zvksed of the draft RISC-V vector cryptography extensions as per the 20230303 version of the specification(1) (1fcbb30). Please note that the Zvkt data-independent execution latency extension has not been implemented, and we would recommend not using these patches in an environment where timing attacks are an issue.
Quoted 6 lines
Work performed by Dickon, Lawrence, Nazar, Kiran, and William from Codethink sponsored by SiFive, as well as Max Chou and Frank Chang from SiFive.
For convenience we have created a git repo with our patches on top of a recent master. https://github.com/CodethinkLabs/qemu-ct
I did test and review this patchset.
Since most of my comments affect multiple patches I have summarized them here in one email.
Observations that only affect a single patch will be sent in response to the corresponding email.I have tested this series with the OpenSSL PR for Zvk that can be found here:
https://github.com/openssl/openssl/pull/20149I ran with all Zvk* extensions enabled (using Zvkg for GCM) and with Zvkb only (using Zvkb for GCM).
All tests succeed. Note, however, that the test coverage is limited (e.g. no .vv instructions, vstart is always zero).Quoted 4 lines
When sending out a follow-up version (even if it just introduces a minimal fix),
then consider using patchset versioning (e.g. git format-patch -v2 ...).Ok, will do
Quoted 9 lines
It might be a matter of taste, but I would prefer a series that groups and orders the commits differently:
a) independent changes to the existing code (refactoring only, but
no new features) - one commit per topic
b) introduction of new functionality - one commit per extension
A series using such a commit granularity and order would be easier to maintain and review (and not result in 45 patches).
Also, the refactoring changes could land before Zvk freezes if maintainers decide to do so.Makes sense, will do
Quoted 4 lines
So far all translation files in target/riscv/insn_trans/* contain multiple extensions if they are related.
I think we should follow this pattern and use a common trans_zvk.c.inc file.Agree, will do
Quoted 3 lines
All patches to insn32.decode have comments of the form "RV64 Zvk* vector crypto extension".
What is the point of the "RV64"? I would simply remove that.Ok, will remove it
Quoted 3 lines
All instructions set "
env->vstart= 0;" at the end.
I don't think that this is correct (the specification does not require this).
Quoted 5 lines
That's from vector spec: "All vector instructions are defined to begin execution with the element number given in the vstart CSR, leaving earlier elements in the destination vector undisturbed, and to reset the vstart CSR to zero at the end of execution." - from "3.7. Vector Start Index CSR vstart"
Yes, you are right.
I just created a PR for the Zvk* spec to clarify this:
https://github.com/riscv/riscv-crypto/pull/308
Quoted 36 lines
Quoted 5 lines
The tests of the reserved encodings are not consistent:
* Zvknh does a dynamic test (query tcg_gen_*())
* Zvkned does a dynamic test (tcg_gen_*())
* Zvkg does not test for (vl%EGS == 0)Zvkg also does dynamic test, by calling macros
GEN_V_UNMASKED_TRANSandGEN_VV_UNMASKED_TRANSQuoted 7 lines
The vl CSR can only be updated by the vset{i}vl{i} instructions. The same applies to the vstart CSR and the vtype CSR that holds vsew, vlmul and other fields.
The current code tests the VSTART/SEW value using "s->vstart% 4 == 0"/"s->sew ==MO_32".
Why is it not possible to do the same with VL, i.e. "s->vl% 4 == 0" (after adding it to DisasContext)?vl can also be updated by another instruction - from vector spec "3.5. Vector Length Register vl" - "The XLEN-bit-wide read-only vl CSR can only be updated by the vset{i}vl{i} instructions, and the fault-only-first vector load instruction variants." So just because of that fault-only-first instruction we need dynamic checks.
vstart is just another CSR -- software can write to it, but probably shouldn't. Whether that's ever going to be useful outside testing ISA conformance tests or not I don't know, but it's clearly read-write (also section 3.7).
Quoted 3 lines
Also, I would introduce named constants or macros for the EGS values to avoid magic constants in the code
(some extensions do that - e.g.ZVKSED_EGS).Makes sense, will do
Best,
Lawrence