* [PATCH v4 0/2] Enhanced VSTART and VL checks for vector instructions
@ 2025-03-10 2:35 Chao Liu
2025-03-10 2:35 ` [PATCH v4 1/2] target/riscv: refactor VSTART_CHECK_EARLY_EXIT() to accept vl as a parameter Chao Liu
` (2 more replies)
0 siblings, 3 replies; 6+ messages in thread
From: Chao Liu @ 2025-03-10 2:35 UTC (permalink / raw)
To: alistair23
Cc: alistair.francis, bmeng.cn, dbarboza, lc00631, liwei1518, palmer,
qemu-devel, qemu-riscv, zhiwei_liu, zqz00548
Hi, Alistair:
I rebase both patches based on the riscv-to-apply.next branch and tested them.
https://github.com/alistair23/qemu/tree/riscv-to-apply.next
Only the first patch had two conflicts, which were resolved nicely,
and the second patch rebased fine, so you can review it again if necessary,
thanks!
PATCH v3 review:
https://lore.kernel.org/qemu-riscv/71f15782-ec28-4763-8197-c10ec12811ec@ventanamicro.com/
PATCH v2 review:
https://lore.kernel.org/qemu-devel/61e8f7d8-607a-4d63-b9dd-cfbfc840716e@ventanamicro.com/
PATCH v1 review:
https://lore.kernel.org/qemu-devel/CAKmqyKPFYxhK8PANOVzV3FMWxd79wZSJYLWkKOAgEMt_b2KvZA@mail.gmail.com/
--
Regards,
Chao
Chao Liu (2):
target/riscv: refactor VSTART_CHECK_EARLY_EXIT() to accept vl as a
parameter
target/riscv: fix handling of nop for vstart >= vl in some vector
instruction
target/riscv/vcrypto_helper.c | 32 ++++++-------
target/riscv/vector_helper.c | 83 ++++++++++++++++++---------------
target/riscv/vector_internals.c | 4 +-
target/riscv/vector_internals.h | 12 ++---
4 files changed, 69 insertions(+), 62 deletions(-)
--
2.48.1
^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH v4 1/2] target/riscv: refactor VSTART_CHECK_EARLY_EXIT() to accept vl as a parameter
2025-03-10 2:35 [PATCH v4 0/2] Enhanced VSTART and VL checks for vector instructions Chao Liu
@ 2025-03-10 2:35 ` Chao Liu
2025-03-10 2:35 ` [PATCH v4 2/2] target/riscv: fix handling of nop for vstart >= vl in some vector instruction Chao Liu
2025-03-10 22:52 ` [PATCH v4 0/2] Enhanced VSTART and VL checks for vector instructions Alistair Francis
2 siblings, 0 replies; 6+ messages in thread
From: Chao Liu @ 2025-03-10 2:35 UTC (permalink / raw)
To: alistair23
Cc: alistair.francis, bmeng.cn, dbarboza, lc00631, liwei1518, palmer,
qemu-devel, qemu-riscv, zhiwei_liu, zqz00548
Some vector instructions are special, such as the vlm.v instruction,
where setting its vl actually sets evl = (vl + 7) >> 3. To improve
maintainability, we will uniformly use VSTART_CHECK_EARLY_EXIT() to
check for the condition vstart >= vl. This function will also handle
cases involving evl.
Fixes: df4252b2ec ("target/riscv/vector_helpers: do early exit when
vstart >= vl")
Signed-off-by: Chao Liu <lc00631@tecorigin.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
---
target/riscv/vcrypto_helper.c | 32 +++++++--------
target/riscv/vector_helper.c | 69 ++++++++++++++++-----------------
target/riscv/vector_internals.c | 4 +-
target/riscv/vector_internals.h | 12 +++---
4 files changed, 57 insertions(+), 60 deletions(-)
diff --git a/target/riscv/vcrypto_helper.c b/target/riscv/vcrypto_helper.c
index f7423df..1526de9 100644
--- a/target/riscv/vcrypto_helper.c
+++ b/target/riscv/vcrypto_helper.c
@@ -222,7 +222,7 @@ static inline void xor_round_key(AESState *round_state, AESState *round_key)
uint32_t total_elems = vext_get_total_elems(env, desc, 4); \
uint32_t vta = vext_vta(desc); \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (uint32_t i = env->vstart / 4; i < env->vl / 4; i++) { \
AESState round_key; \
@@ -248,7 +248,7 @@ static inline void xor_round_key(AESState *round_state, AESState *round_key)
uint32_t total_elems = vext_get_total_elems(env, desc, 4); \
uint32_t vta = vext_vta(desc); \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (uint32_t i = env->vstart / 4; i < env->vl / 4; i++) { \
AESState round_key; \
@@ -309,7 +309,7 @@ void HELPER(vaeskf1_vi)(void *vd_vptr, void *vs2_vptr, uint32_t uimm,
uint32_t total_elems = vext_get_total_elems(env, desc, 4);
uint32_t vta = vext_vta(desc);
- VSTART_CHECK_EARLY_EXIT(env);
+ VSTART_CHECK_EARLY_EXIT(env, vl);
uimm &= 0b1111;
if (uimm > 10 || uimm == 0) {
@@ -357,7 +357,7 @@ void HELPER(vaeskf2_vi)(void *vd_vptr, void *vs2_vptr, uint32_t uimm,
uint32_t total_elems = vext_get_total_elems(env, desc, 4);
uint32_t vta = vext_vta(desc);
- VSTART_CHECK_EARLY_EXIT(env);
+ VSTART_CHECK_EARLY_EXIT(env, vl);
uimm &= 0b1111;
if (uimm > 14 || uimm < 2) {
@@ -465,7 +465,7 @@ void HELPER(vsha2ms_vv)(void *vd, void *vs1, void *vs2, CPURISCVState *env,
uint32_t total_elems;
uint32_t vta = vext_vta(desc);
- VSTART_CHECK_EARLY_EXIT(env);
+ VSTART_CHECK_EARLY_EXIT(env, env->vl);
for (uint32_t i = env->vstart / 4; i < env->vl / 4; i++) {
if (sew == MO_32) {
@@ -582,7 +582,7 @@ void HELPER(vsha2ch32_vv)(void *vd, void *vs1, void *vs2, CPURISCVState *env,
uint32_t total_elems;
uint32_t vta = vext_vta(desc);
- VSTART_CHECK_EARLY_EXIT(env);
+ VSTART_CHECK_EARLY_EXIT(env, env->vl);
for (uint32_t i = env->vstart / 4; i < env->vl / 4; i++) {
vsha2c_32(((uint32_t *)vs2) + 4 * i, ((uint32_t *)vd) + 4 * i,
@@ -602,7 +602,7 @@ void HELPER(vsha2ch64_vv)(void *vd, void *vs1, void *vs2, CPURISCVState *env,
uint32_t total_elems;
uint32_t vta = vext_vta(desc);
- VSTART_CHECK_EARLY_EXIT(env);
+ VSTART_CHECK_EARLY_EXIT(env, env->vl);
for (uint32_t i = env->vstart / 4; i < env->vl / 4; i++) {
vsha2c_64(((uint64_t *)vs2) + 4 * i, ((uint64_t *)vd) + 4 * i,
@@ -622,7 +622,7 @@ void HELPER(vsha2cl32_vv)(void *vd, void *vs1, void *vs2, CPURISCVState *env,
uint32_t total_elems;
uint32_t vta = vext_vta(desc);
- VSTART_CHECK_EARLY_EXIT(env);
+ VSTART_CHECK_EARLY_EXIT(env, env->vl);
for (uint32_t i = env->vstart / 4; i < env->vl / 4; i++) {
vsha2c_32(((uint32_t *)vs2) + 4 * i, ((uint32_t *)vd) + 4 * i,
@@ -642,7 +642,7 @@ void HELPER(vsha2cl64_vv)(void *vd, void *vs1, void *vs2, CPURISCVState *env,
uint32_t total_elems;
uint32_t vta = vext_vta(desc);
- VSTART_CHECK_EARLY_EXIT(env);
+ VSTART_CHECK_EARLY_EXIT(env, env->vl);
for (uint32_t i = env->vstart / 4; i < env->vl / 4; i++) {
vsha2c_64(((uint64_t *)vs2) + 4 * i, ((uint64_t *)vd) + 4 * i,
@@ -676,7 +676,7 @@ void HELPER(vsm3me_vv)(void *vd_vptr, void *vs1_vptr, void *vs2_vptr,
uint32_t *vs1 = vs1_vptr;
uint32_t *vs2 = vs2_vptr;
- VSTART_CHECK_EARLY_EXIT(env);
+ VSTART_CHECK_EARLY_EXIT(env, env->vl);
for (int i = env->vstart / 8; i < env->vl / 8; i++) {
uint32_t w[24];
@@ -777,7 +777,7 @@ void HELPER(vsm3c_vi)(void *vd_vptr, void *vs2_vptr, uint32_t uimm,
uint32_t *vs2 = vs2_vptr;
uint32_t v1[8], v2[8], v3[8];
- VSTART_CHECK_EARLY_EXIT(env);
+ VSTART_CHECK_EARLY_EXIT(env, env->vl);
for (int i = env->vstart / 8; i < env->vl / 8; i++) {
for (int k = 0; k < 8; k++) {
@@ -802,7 +802,7 @@ void HELPER(vghsh_vv)(void *vd_vptr, void *vs1_vptr, void *vs2_vptr,
uint32_t vta = vext_vta(desc);
uint32_t total_elems = vext_get_total_elems(env, desc, 4);
- VSTART_CHECK_EARLY_EXIT(env);
+ VSTART_CHECK_EARLY_EXIT(env, env->vl);
for (uint32_t i = env->vstart / 4; i < env->vl / 4; i++) {
uint64_t Y[2] = {vd[i * 2 + 0], vd[i * 2 + 1]};
@@ -841,7 +841,7 @@ void HELPER(vgmul_vv)(void *vd_vptr, void *vs2_vptr, CPURISCVState *env,
uint32_t vta = vext_vta(desc);
uint32_t total_elems = vext_get_total_elems(env, desc, 4);
- VSTART_CHECK_EARLY_EXIT(env);
+ VSTART_CHECK_EARLY_EXIT(env, env->vl);
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])};
@@ -879,7 +879,7 @@ void HELPER(vsm4k_vi)(void *vd, void *vs2, uint32_t uimm5, CPURISCVState *env,
uint32_t esz = sizeof(uint32_t);
uint32_t total_elems = vext_get_total_elems(env, desc, esz);
- VSTART_CHECK_EARLY_EXIT(env);
+ VSTART_CHECK_EARLY_EXIT(env, env->vl);
for (uint32_t i = group_start; i < group_end; ++i) {
uint32_t vstart = i * egs;
@@ -937,7 +937,7 @@ void HELPER(vsm4r_vv)(void *vd, void *vs2, CPURISCVState *env, uint32_t desc)
uint32_t esz = sizeof(uint32_t);
uint32_t total_elems = vext_get_total_elems(env, desc, esz);
- VSTART_CHECK_EARLY_EXIT(env);
+ VSTART_CHECK_EARLY_EXIT(env, env->vl);
for (uint32_t i = group_start; i < group_end; ++i) {
uint32_t vstart = i * egs;
@@ -973,7 +973,7 @@ void HELPER(vsm4r_vs)(void *vd, void *vs2, CPURISCVState *env, uint32_t desc)
uint32_t esz = sizeof(uint32_t);
uint32_t total_elems = vext_get_total_elems(env, desc, esz);
- VSTART_CHECK_EARLY_EXIT(env);
+ VSTART_CHECK_EARLY_EXIT(env, env->vl);
for (uint32_t i = group_start; i < group_end; ++i) {
uint32_t vstart = i * egs;
diff --git a/target/riscv/vector_helper.c b/target/riscv/vector_helper.c
index 71b823d..217d2f6 100644
--- a/target/riscv/vector_helper.c
+++ b/target/riscv/vector_helper.c
@@ -260,7 +260,7 @@ vext_ldst_stride(void *vd, void *v0, target_ulong base, target_ulong stride,
uint32_t esz = 1 << log2_esz;
uint32_t vma = vext_vma(desc);
- VSTART_CHECK_EARLY_EXIT(env);
+ VSTART_CHECK_EARLY_EXIT(env, env->vl);
for (i = env->vstart; i < env->vl; env->vstart = ++i) {
k = 0;
@@ -383,10 +383,7 @@ vext_ldst_us(void *vd, target_ulong base, CPURISCVState *env, uint32_t desc,
uint32_t msize = nf * esz;
int mmu_index = riscv_env_mmu_index(env, false);
- if (env->vstart >= evl) {
- env->vstart = 0;
- return;
- }
+ VSTART_CHECK_EARLY_EXIT(env, evl);
#if defined(CONFIG_USER_ONLY)
/*
@@ -544,7 +541,7 @@ vext_ldst_index(void *vd, void *v0, target_ulong base,
uint32_t esz = 1 << log2_esz;
uint32_t vma = vext_vma(desc);
- VSTART_CHECK_EARLY_EXIT(env);
+ VSTART_CHECK_EARLY_EXIT(env, env->vl);
/* load bytes from guest memory */
for (i = env->vstart; i < env->vl; env->vstart = ++i) {
@@ -638,7 +635,7 @@ vext_ldff(void *vd, void *v0, target_ulong base, CPURISCVState *env,
int flags;
void *host;
- VSTART_CHECK_EARLY_EXIT(env);
+ VSTART_CHECK_EARLY_EXIT(env, env->vl);
addr = base + ((env->vstart * nf) << log2_esz);
page_split = -(addr | TARGET_PAGE_MASK);
@@ -1116,7 +1113,7 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, void *vs2, \
uint32_t vta = vext_vta(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
ETYPE s1 = *((ETYPE *)vs1 + H(i)); \
@@ -1150,7 +1147,7 @@ void HELPER(NAME)(void *vd, void *v0, target_ulong s1, void *vs2, \
uint32_t vta = vext_vta(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
ETYPE s2 = *((ETYPE *)vs2 + H(i)); \
@@ -1187,7 +1184,7 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, void *vs2, \
uint32_t vta_all_1s = vext_vta_all_1s(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
ETYPE s1 = *((ETYPE *)vs1 + H(i)); \
@@ -1227,7 +1224,7 @@ void HELPER(NAME)(void *vd, void *v0, target_ulong s1, \
uint32_t vta_all_1s = vext_vta_all_1s(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
ETYPE s2 = *((ETYPE *)vs2 + H(i)); \
@@ -1325,7 +1322,7 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, \
uint32_t vma = vext_vma(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
if (!vm && !vext_elem_mask(v0, i)) { \
@@ -1374,7 +1371,7 @@ void HELPER(NAME)(void *vd, void *v0, target_ulong s1, \
uint32_t vma = vext_vma(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
if (!vm && !vext_elem_mask(v0, i)) { \
@@ -1438,7 +1435,7 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, void *vs2, \
uint32_t vma = vext_vma(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
ETYPE s1 = *((ETYPE *)vs1 + H(i)); \
@@ -1505,7 +1502,7 @@ void HELPER(NAME)(void *vd, void *v0, target_ulong s1, void *vs2, \
uint32_t vma = vext_vma(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
ETYPE s2 = *((ETYPE *)vs2 + H(i)); \
@@ -2054,7 +2051,7 @@ void HELPER(NAME)(void *vd, void *vs1, CPURISCVState *env, \
uint32_t vta = vext_vta(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
ETYPE s1 = *((ETYPE *)vs1 + H(i)); \
@@ -2080,7 +2077,7 @@ void HELPER(NAME)(void *vd, uint64_t s1, CPURISCVState *env, \
uint32_t vta = vext_vta(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
*((ETYPE *)vd + H(i)) = (ETYPE)s1; \
@@ -2105,7 +2102,7 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, void *vs2, \
uint32_t vta = vext_vta(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
ETYPE *vt = (!vext_elem_mask(v0, i) ? vs2 : vs1); \
@@ -2131,7 +2128,7 @@ void HELPER(NAME)(void *vd, void *v0, target_ulong s1, \
uint32_t vta = vext_vta(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
ETYPE s2 = *((ETYPE *)vs2 + H(i)); \
@@ -2178,7 +2175,7 @@ vext_vv_rm_1(void *vd, void *v0, void *vs1, void *vs2,
uint32_t vl, uint32_t vm, int vxrm,
opivv2_rm_fn *fn, uint32_t vma, uint32_t esz)
{
- VSTART_CHECK_EARLY_EXIT(env);
+ VSTART_CHECK_EARLY_EXIT(env, vl);
for (uint32_t i = env->vstart; i < vl; i++) {
if (!vm && !vext_elem_mask(v0, i)) {
@@ -2305,7 +2302,7 @@ vext_vx_rm_1(void *vd, void *v0, target_long s1, void *vs2,
uint32_t vl, uint32_t vm, int vxrm,
opivx2_rm_fn *fn, uint32_t vma, uint32_t esz)
{
- VSTART_CHECK_EARLY_EXIT(env);
+ VSTART_CHECK_EARLY_EXIT(env, vl);
for (uint32_t i = env->vstart; i < vl; i++) {
if (!vm && !vext_elem_mask(v0, i)) {
@@ -3104,7 +3101,7 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, \
uint32_t vma = vext_vma(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
if (!vm && !vext_elem_mask(v0, i)) { \
@@ -3149,7 +3146,7 @@ void HELPER(NAME)(void *vd, void *v0, uint64_t s1, \
uint32_t vma = vext_vma(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
if (!vm && !vext_elem_mask(v0, i)) { \
@@ -3737,7 +3734,7 @@ void HELPER(NAME)(void *vd, void *v0, void *vs2, \
uint32_t vma = vext_vma(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
if (vl == 0) { \
return; \
@@ -4260,7 +4257,7 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, void *vs2, \
uint32_t vma = vext_vma(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
ETYPE s1 = *((ETYPE *)vs1 + H(i)); \
@@ -4302,7 +4299,7 @@ void HELPER(NAME)(void *vd, void *v0, uint64_t s1, void *vs2, \
uint32_t vma = vext_vma(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
ETYPE s2 = *((ETYPE *)vs2 + H(i)); \
@@ -4497,7 +4494,7 @@ void HELPER(NAME)(void *vd, void *v0, uint64_t s1, void *vs2, \
uint32_t vta = vext_vta(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
ETYPE s2 = *((ETYPE *)vs2 + H(i)); \
@@ -4827,7 +4824,7 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, \
uint32_t i; \
int a, b; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
a = vext_elem_mask(vs1, i); \
@@ -5022,7 +5019,7 @@ void HELPER(NAME)(void *vd, void *v0, CPURISCVState *env, uint32_t desc) \
uint32_t vma = vext_vma(desc); \
int i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
if (!vm && !vext_elem_mask(v0, i)) { \
@@ -5059,7 +5056,7 @@ void HELPER(NAME)(void *vd, void *v0, target_ulong s1, void *vs2, \
uint32_t vma = vext_vma(desc); \
target_ulong offset = s1, i_min, i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
i_min = MAX(env->vstart, offset); \
for (i = i_min; i < vl; i++) { \
@@ -5094,7 +5091,7 @@ void HELPER(NAME)(void *vd, void *v0, target_ulong s1, void *vs2, \
uint32_t vma = vext_vma(desc); \
target_ulong i_max, i_min, i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
i_min = MIN(s1 < vlmax ? vlmax - s1 : 0, vl); \
i_max = MAX(i_min, env->vstart); \
@@ -5138,7 +5135,7 @@ static void vslide1up_##BITWIDTH(void *vd, void *v0, uint64_t s1, \
uint32_t vma = vext_vma(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
if (!vm && !vext_elem_mask(v0, i)) { \
@@ -5189,7 +5186,7 @@ static void vslide1down_##BITWIDTH(void *vd, void *v0, uint64_t s1, \
uint32_t vma = vext_vma(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
if (!vm && !vext_elem_mask(v0, i)) { \
@@ -5266,7 +5263,7 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, void *vs2, \
uint64_t index; \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
if (!vm && !vext_elem_mask(v0, i)) { \
@@ -5311,7 +5308,7 @@ void HELPER(NAME)(void *vd, void *v0, target_ulong s1, void *vs2, \
uint64_t index = s1; \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
if (!vm && !vext_elem_mask(v0, i)) { \
@@ -5407,7 +5404,7 @@ void HELPER(NAME)(void *vd, void *v0, void *vs2, \
uint32_t vma = vext_vma(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
if (!vm && !vext_elem_mask(v0, i)) { \
diff --git a/target/riscv/vector_internals.c b/target/riscv/vector_internals.c
index 05b2d01..b490b1d 100644
--- a/target/riscv/vector_internals.c
+++ b/target/riscv/vector_internals.c
@@ -66,7 +66,7 @@ void do_vext_vv(void *vd, void *v0, void *vs1, void *vs2,
uint32_t vma = vext_vma(desc);
uint32_t i;
- VSTART_CHECK_EARLY_EXIT(env);
+ VSTART_CHECK_EARLY_EXIT(env, vl);
for (i = env->vstart; i < vl; i++) {
if (!vm && !vext_elem_mask(v0, i)) {
@@ -92,7 +92,7 @@ void do_vext_vx(void *vd, void *v0, target_long s1, void *vs2,
uint32_t vma = vext_vma(desc);
uint32_t i;
- VSTART_CHECK_EARLY_EXIT(env);
+ VSTART_CHECK_EARLY_EXIT(env, vl);
for (i = env->vstart; i < vl; i++) {
if (!vm && !vext_elem_mask(v0, i)) {
diff --git a/target/riscv/vector_internals.h b/target/riscv/vector_internals.h
index a11cc83..8eee7e5 100644
--- a/target/riscv/vector_internals.h
+++ b/target/riscv/vector_internals.h
@@ -25,11 +25,11 @@
#include "tcg/tcg-gvec-desc.h"
#include "internals.h"
-#define VSTART_CHECK_EARLY_EXIT(env) do { \
- if (env->vstart >= env->vl) { \
- env->vstart = 0; \
- return; \
- } \
+#define VSTART_CHECK_EARLY_EXIT(env, vl) do { \
+ if (env->vstart >= vl) { \
+ env->vstart = 0; \
+ return; \
+ } \
} while (0)
static inline uint32_t vext_nf(uint32_t desc)
@@ -159,7 +159,7 @@ void HELPER(NAME)(void *vd, void *v0, void *vs2, \
uint32_t vma = vext_vma(desc); \
uint32_t i; \
\
- VSTART_CHECK_EARLY_EXIT(env); \
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
\
for (i = env->vstart; i < vl; i++) { \
if (!vm && !vext_elem_mask(v0, i)) { \
--
2.48.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH v4 2/2] target/riscv: fix handling of nop for vstart >= vl in some vector instruction
2025-03-10 2:35 [PATCH v4 0/2] Enhanced VSTART and VL checks for vector instructions Chao Liu
2025-03-10 2:35 ` [PATCH v4 1/2] target/riscv: refactor VSTART_CHECK_EARLY_EXIT() to accept vl as a parameter Chao Liu
@ 2025-03-10 2:35 ` Chao Liu
2025-03-24 4:48 ` Michael Tokarev
2025-03-10 22:52 ` [PATCH v4 0/2] Enhanced VSTART and VL checks for vector instructions Alistair Francis
2 siblings, 1 reply; 6+ messages in thread
From: Chao Liu @ 2025-03-10 2:35 UTC (permalink / raw)
To: alistair23
Cc: alistair.francis, bmeng.cn, dbarboza, lc00631, liwei1518, palmer,
qemu-devel, qemu-riscv, zhiwei_liu, zqz00548
Recently, when I was writing a RISCV test, I found that when VL is set to 0, the
instruction should be nop, but when I tested it, I found that QEMU will treat
all elements as tail elements, and in the case of VTA=1, write all elements
to 1.
After troubleshooting, it was found that the vext_vx_rm_1 function was called in
the vext_vx_rm_2, and then the vext_set_elems_1s function was called to process
the tail element, but only VSTART >= vl was checked in the vext_vx_rm_1
function, which caused the tail element to still be processed even if it was
returned in advance.
So I've made the following change:
Put VSTART_CHECK_EARLY_EXIT(env) at the beginning of the vext_vx_rm_2 function,
so that the VSTART register is checked correctly.
Fixes: df4252b2ec ("target/riscv/vector_helpers: do early exit when
vstart >= vl")
Signed-off-by: Chao Liu <lc00631@tecorigin.com>
Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
---
target/riscv/vector_helper.c | 18 ++++++++++++++----
1 file changed, 14 insertions(+), 4 deletions(-)
diff --git a/target/riscv/vector_helper.c b/target/riscv/vector_helper.c
index 217d2f6..67b3baf 100644
--- a/target/riscv/vector_helper.c
+++ b/target/riscv/vector_helper.c
@@ -2175,8 +2175,6 @@ vext_vv_rm_1(void *vd, void *v0, void *vs1, void *vs2,
uint32_t vl, uint32_t vm, int vxrm,
opivv2_rm_fn *fn, uint32_t vma, uint32_t esz)
{
- VSTART_CHECK_EARLY_EXIT(env, vl);
-
for (uint32_t i = env->vstart; i < vl; i++) {
if (!vm && !vext_elem_mask(v0, i)) {
/* set masked-off elements to 1s */
@@ -2200,6 +2198,8 @@ vext_vv_rm_2(void *vd, void *v0, void *vs1, void *vs2,
uint32_t vta = vext_vta(desc);
uint32_t vma = vext_vma(desc);
+ VSTART_CHECK_EARLY_EXIT(env, vl);
+
switch (env->vxrm) {
case 0: /* rnu */
vext_vv_rm_1(vd, v0, vs1, vs2,
@@ -2302,8 +2302,6 @@ vext_vx_rm_1(void *vd, void *v0, target_long s1, void *vs2,
uint32_t vl, uint32_t vm, int vxrm,
opivx2_rm_fn *fn, uint32_t vma, uint32_t esz)
{
- VSTART_CHECK_EARLY_EXIT(env, vl);
-
for (uint32_t i = env->vstart; i < vl; i++) {
if (!vm && !vext_elem_mask(v0, i)) {
/* set masked-off elements to 1s */
@@ -2327,6 +2325,8 @@ vext_vx_rm_2(void *vd, void *v0, target_long s1, void *vs2,
uint32_t vta = vext_vta(desc);
uint32_t vma = vext_vma(desc);
+ VSTART_CHECK_EARLY_EXIT(env, vl);
+
switch (env->vxrm) {
case 0: /* rnu */
vext_vx_rm_1(vd, v0, s1, vs2,
@@ -4662,6 +4662,8 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, \
uint32_t i; \
TD s1 = *((TD *)vs1 + HD(0)); \
\
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
+ \
for (i = env->vstart; i < vl; i++) { \
TS2 s2 = *((TS2 *)vs2 + HS2(i)); \
if (!vm && !vext_elem_mask(v0, i)) { \
@@ -4750,6 +4752,8 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, \
uint32_t i; \
TD s1 = *((TD *)vs1 + HD(0)); \
\
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
+ \
for (i = env->vstart; i < vl; i++) { \
TS2 s2 = *((TS2 *)vs2 + HS2(i)); \
if (!vm && !vext_elem_mask(v0, i)) { \
@@ -4914,6 +4918,8 @@ static void vmsetm(void *vd, void *v0, void *vs2, CPURISCVState *env,
int i;
bool first_mask_bit = false;
+ VSTART_CHECK_EARLY_EXIT(env, vl);
+
for (i = env->vstart; i < vl; i++) {
if (!vm && !vext_elem_mask(v0, i)) {
/* set masked-off elements to 1s */
@@ -4986,6 +4992,8 @@ void HELPER(NAME)(void *vd, void *v0, void *vs2, CPURISCVState *env, \
uint32_t sum = 0; \
int i; \
\
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
+ \
for (i = env->vstart; i < vl; i++) { \
if (!vm && !vext_elem_mask(v0, i)) { \
/* set masked-off elements to 1s */ \
@@ -5344,6 +5352,8 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, void *vs2, \
uint32_t vta = vext_vta(desc); \
uint32_t num = 0, i; \
\
+ VSTART_CHECK_EARLY_EXIT(env, vl); \
+ \
for (i = env->vstart; i < vl; i++) { \
if (!vext_elem_mask(vs1, i)) { \
continue; \
--
2.48.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH v4 0/2] Enhanced VSTART and VL checks for vector instructions
2025-03-10 2:35 [PATCH v4 0/2] Enhanced VSTART and VL checks for vector instructions Chao Liu
2025-03-10 2:35 ` [PATCH v4 1/2] target/riscv: refactor VSTART_CHECK_EARLY_EXIT() to accept vl as a parameter Chao Liu
2025-03-10 2:35 ` [PATCH v4 2/2] target/riscv: fix handling of nop for vstart >= vl in some vector instruction Chao Liu
@ 2025-03-10 22:52 ` Alistair Francis
2 siblings, 0 replies; 6+ messages in thread
From: Alistair Francis @ 2025-03-10 22:52 UTC (permalink / raw)
To: Chao Liu
Cc: alistair.francis, bmeng.cn, dbarboza, liwei1518, palmer,
qemu-devel, qemu-riscv, zhiwei_liu, zqz00548
On Mon, Mar 10, 2025 at 12:37 PM Chao Liu <lc00631@tecorigin.com> wrote:
>
> Hi, Alistair:
>
> I rebase both patches based on the riscv-to-apply.next branch and tested them.
> https://github.com/alistair23/qemu/tree/riscv-to-apply.next
>
> Only the first patch had two conflicts, which were resolved nicely,
> and the second patch rebased fine, so you can review it again if necessary,
> thanks!
>
> PATCH v3 review:
> https://lore.kernel.org/qemu-riscv/71f15782-ec28-4763-8197-c10ec12811ec@ventanamicro.com/
>
> PATCH v2 review:
> https://lore.kernel.org/qemu-devel/61e8f7d8-607a-4d63-b9dd-cfbfc840716e@ventanamicro.com/
>
> PATCH v1 review:
> https://lore.kernel.org/qemu-devel/CAKmqyKPFYxhK8PANOVzV3FMWxd79wZSJYLWkKOAgEMt_b2KvZA@mail.gmail.com/
>
> --
> Regards,
> Chao
>
> Chao Liu (2):
> target/riscv: refactor VSTART_CHECK_EARLY_EXIT() to accept vl as a
> parameter
> target/riscv: fix handling of nop for vstart >= vl in some vector
> instruction
Thanks!
Applied to riscv-to-apply.next
Alistair
>
> target/riscv/vcrypto_helper.c | 32 ++++++-------
> target/riscv/vector_helper.c | 83 ++++++++++++++++++---------------
> target/riscv/vector_internals.c | 4 +-
> target/riscv/vector_internals.h | 12 ++---
> 4 files changed, 69 insertions(+), 62 deletions(-)
>
> --
> 2.48.1
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH v4 2/2] target/riscv: fix handling of nop for vstart >= vl in some vector instruction
2025-03-10 2:35 ` [PATCH v4 2/2] target/riscv: fix handling of nop for vstart >= vl in some vector instruction Chao Liu
@ 2025-03-24 4:48 ` Michael Tokarev
2025-03-24 11:18 ` Daniel Henrique Barboza
0 siblings, 1 reply; 6+ messages in thread
From: Michael Tokarev @ 2025-03-24 4:48 UTC (permalink / raw)
To: Chao Liu, alistair23
Cc: alistair.francis, bmeng.cn, dbarboza, liwei1518, palmer,
qemu-devel, qemu-riscv, zhiwei_liu, zqz00548, qemu-devel
10.03.2025 05:35, Chao Liu wrote:
> Recently, when I was writing a RISCV test, I found that when VL is set to 0, the
> instruction should be nop, but when I tested it, I found that QEMU will treat
> all elements as tail elements, and in the case of VTA=1, write all elements
> to 1.
>
> After troubleshooting, it was found that the vext_vx_rm_1 function was called in
> the vext_vx_rm_2, and then the vext_set_elems_1s function was called to process
> the tail element, but only VSTART >= vl was checked in the vext_vx_rm_1
> function, which caused the tail element to still be processed even if it was
> returned in advance.
>
> So I've made the following change:
>
> Put VSTART_CHECK_EARLY_EXIT(env) at the beginning of the vext_vx_rm_2 function,
> so that the VSTART register is checked correctly.
>
> Fixes: df4252b2ec ("target/riscv/vector_helpers: do early exit when
> vstart >= vl")
> Signed-off-by: Chao Liu <lc00631@tecorigin.com>
> Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Is this a qemu-stable material (9.2)?
Thanks,
/mjt
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH v4 2/2] target/riscv: fix handling of nop for vstart >= vl in some vector instruction
2025-03-24 4:48 ` Michael Tokarev
@ 2025-03-24 11:18 ` Daniel Henrique Barboza
0 siblings, 0 replies; 6+ messages in thread
From: Daniel Henrique Barboza @ 2025-03-24 11:18 UTC (permalink / raw)
To: Michael Tokarev, Chao Liu, alistair23
Cc: alistair.francis, bmeng.cn, liwei1518, palmer, qemu-devel,
qemu-riscv, zhiwei_liu, zqz00548
On 3/24/25 1:48 AM, Michael Tokarev wrote:
> 10.03.2025 05:35, Chao Liu wrote:
>> Recently, when I was writing a RISCV test, I found that when VL is set to 0, the
>> instruction should be nop, but when I tested it, I found that QEMU will treat
>> all elements as tail elements, and in the case of VTA=1, write all elements
>> to 1.
>>
>> After troubleshooting, it was found that the vext_vx_rm_1 function was called in
>> the vext_vx_rm_2, and then the vext_set_elems_1s function was called to process
>> the tail element, but only VSTART >= vl was checked in the vext_vx_rm_1
>> function, which caused the tail element to still be processed even if it was
>> returned in advance.
>>
>> So I've made the following change:
>>
>> Put VSTART_CHECK_EARLY_EXIT(env) at the beginning of the vext_vx_rm_2 function,
>> so that the VSTART register is checked correctly.
>>
>> Fixes: df4252b2ec ("target/riscv/vector_helpers: do early exit when
>> vstart >= vl")
>> Signed-off-by: Chao Liu <lc00631@tecorigin.com>
>> Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
>
> Is this a qemu-stable material (9.2)?
Yes. Go ahead. Thanks,
Daniel
>
> Thanks,
>
> /mjt
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2025-03-24 11:18 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-03-10 2:35 [PATCH v4 0/2] Enhanced VSTART and VL checks for vector instructions Chao Liu
2025-03-10 2:35 ` [PATCH v4 1/2] target/riscv: refactor VSTART_CHECK_EARLY_EXIT() to accept vl as a parameter Chao Liu
2025-03-10 2:35 ` [PATCH v4 2/2] target/riscv: fix handling of nop for vstart >= vl in some vector instruction Chao Liu
2025-03-24 4:48 ` Michael Tokarev
2025-03-24 11:18 ` Daniel Henrique Barboza
2025-03-10 22:52 ` [PATCH v4 0/2] Enhanced VSTART and VL checks for vector instructions Alistair Francis
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.