Skip to content

Commit

Permalink
target/loongarch: Implement LASX fpu arith instructions
Browse files Browse the repository at this point in the history 
This patch includes:
- XVF{ADD/SUB/MUL/DIV}.{S/D};
- XVF{MADD/MSUB/NMADD/NMSUB}.{S/D};
- XVF{MAX/MIN}.{S/D};
- XVF{MAXA/MINA}.{S/D};
- XVFLOGB.{S/D};
- XVFCLASS.{S/D};
- XVF{SQRT/RECIP/RSQRT}.{S/D}.

Signed-off-by: Song Gao <gaosong@loongson.cn>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20230914022645.1151356-47-gaosong@loongson.cn>
  • Loading branch information
@gaosong-loongson
gaosong-loongson committed Sep 20, 2023
1 parent abee168 commit c9caf15
Show file tree
Hide file tree
Showing 4 changed files with 158 additions and 16 deletions.
46 changes: 46 additions & 0 deletions target/loongarch/disas.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -1708,6 +1708,11 @@ static void output_v_i_x(DisasContext *ctx, arg_v_i *a, const char *mnemonic)
output(ctx, mnemonic, "x%d, 0x%x", a->vd, a->imm);
}

static void output_vvvv_x(DisasContext *ctx, arg_vvvv *a, const char *mnemonic)
{
output(ctx, mnemonic, "x%d, x%d, x%d, x%d", a->vd, a->vj, a->vk, a->va);
}

static void output_vvv_x(DisasContext *ctx, arg_vvv * a, const char *mnemonic)
{
output(ctx, mnemonic, "x%d, x%d, x%d", a->vd, a->vj, a->vk);
Expand Down Expand Up @@ -2240,6 +2245,47 @@ INSN_LASX(xvfrstp_h, vvv)
INSN_LASX(xvfrstpi_b, vv_i)
INSN_LASX(xvfrstpi_h, vv_i)

INSN_LASX(xvfadd_s, vvv)
INSN_LASX(xvfadd_d, vvv)
INSN_LASX(xvfsub_s, vvv)
INSN_LASX(xvfsub_d, vvv)
INSN_LASX(xvfmul_s, vvv)
INSN_LASX(xvfmul_d, vvv)
INSN_LASX(xvfdiv_s, vvv)
INSN_LASX(xvfdiv_d, vvv)

INSN_LASX(xvfmadd_s, vvvv)
INSN_LASX(xvfmadd_d, vvvv)
INSN_LASX(xvfmsub_s, vvvv)
INSN_LASX(xvfmsub_d, vvvv)
INSN_LASX(xvfnmadd_s, vvvv)
INSN_LASX(xvfnmadd_d, vvvv)
INSN_LASX(xvfnmsub_s, vvvv)
INSN_LASX(xvfnmsub_d, vvvv)

INSN_LASX(xvfmax_s, vvv)
INSN_LASX(xvfmax_d, vvv)
INSN_LASX(xvfmin_s, vvv)
INSN_LASX(xvfmin_d, vvv)

INSN_LASX(xvfmaxa_s, vvv)
INSN_LASX(xvfmaxa_d, vvv)
INSN_LASX(xvfmina_s, vvv)
INSN_LASX(xvfmina_d, vvv)

INSN_LASX(xvflogb_s, vv)
INSN_LASX(xvflogb_d, vv)

INSN_LASX(xvfclass_s, vv)
INSN_LASX(xvfclass_d, vv)

INSN_LASX(xvfsqrt_s, vv)
INSN_LASX(xvfsqrt_d, vv)
INSN_LASX(xvfrecip_s, vv)
INSN_LASX(xvfrecip_d, vv)
INSN_LASX(xvfrsqrt_s, vv)
INSN_LASX(xvfrsqrt_d, vv)

INSN_LASX(xvreplgr2vr_b, vr)
INSN_LASX(xvreplgr2vr_h, vr)
INSN_LASX(xvreplgr2vr_w, vr)
Expand Down
75 changes: 63 additions & 12 deletions target/loongarch/insn_trans/trans_vec.c.inc
View file
Original file line number Diff line number Diff line change
Expand Up @@ -33,6 +33,10 @@ static bool check_vec(DisasContext *ctx, uint32_t oprsz)
static bool gen_vvvv_ptr_vl(DisasContext *ctx, arg_vvvv *a, uint32_t oprsz,
gen_helper_gvec_4_ptr *fn)
{
if (!check_vec(ctx, oprsz)) {
return true;
}

tcg_gen_gvec_4_ptr(vec_full_offset(a->vd),
vec_full_offset(a->vj),
vec_full_offset(a->vk),
Expand All @@ -45,13 +49,15 @@ static bool gen_vvvv_ptr_vl(DisasContext *ctx, arg_vvvv *a, uint32_t oprsz,
static bool gen_vvvv_ptr(DisasContext *ctx, arg_vvvv *a,
gen_helper_gvec_4_ptr *fn)
{
if (!check_vec(ctx, 16)) {
return true;
}

return gen_vvvv_ptr_vl(ctx, a, 16, fn);
}

static bool gen_xxxx_ptr(DisasContext *ctx, arg_vvvv *a,
gen_helper_gvec_4_ptr *fn)
{
return gen_vvvv_ptr_vl(ctx, a, 32, fn);
}

static bool gen_vvvv_vl(DisasContext *ctx, arg_vvvv *a, uint32_t oprsz,
gen_helper_gvec_4 *fn)
{
Expand All @@ -76,6 +82,9 @@ static bool gen_vvvv(DisasContext *ctx, arg_vvvv *a,
static bool gen_vvv_ptr_vl(DisasContext *ctx, arg_vvv *a, uint32_t oprsz,
gen_helper_gvec_3_ptr *fn)
{
if (!check_vec(ctx, oprsz)) {
return true;
}
tcg_gen_gvec_3_ptr(vec_full_offset(a->vd),
vec_full_offset(a->vj),
vec_full_offset(a->vk),
Expand All @@ -87,13 +96,15 @@ static bool gen_vvv_ptr_vl(DisasContext *ctx, arg_vvv *a, uint32_t oprsz,
static bool gen_vvv_ptr(DisasContext *ctx, arg_vvv *a,
gen_helper_gvec_3_ptr *fn)
{
if (!check_vec(ctx, 16)) {
return true;
}

return gen_vvv_ptr_vl(ctx, a, 16, fn);
}

static bool gen_xxx_ptr(DisasContext *ctx, arg_vvv *a,
gen_helper_gvec_3_ptr *fn)
{
return gen_vvv_ptr_vl(ctx, a, 32, fn);
}

static bool gen_vvv_vl(DisasContext *ctx, arg_vvv *a, uint32_t oprsz,
gen_helper_gvec_3 *fn)
{
Expand Down Expand Up @@ -121,6 +132,10 @@ static bool gen_xxx(DisasContext *ctx, arg_vvv *a, gen_helper_gvec_3 *fn)
static bool gen_vv_ptr_vl(DisasContext *ctx, arg_vv *a, uint32_t oprsz,
gen_helper_gvec_2_ptr *fn)
{
if (!check_vec(ctx, oprsz)) {
return true;
}

tcg_gen_gvec_2_ptr(vec_full_offset(a->vd),
vec_full_offset(a->vj),
cpu_env,
Expand All @@ -131,13 +146,15 @@ static bool gen_vv_ptr_vl(DisasContext *ctx, arg_vv *a, uint32_t oprsz,
static bool gen_vv_ptr(DisasContext *ctx, arg_vv *a,
gen_helper_gvec_2_ptr *fn)
{
if (!check_vec(ctx, 16)) {
return true;
}

return gen_vv_ptr_vl(ctx, a, 16, fn);
}

static bool gen_xx_ptr(DisasContext *ctx, arg_vv *a,
gen_helper_gvec_2_ptr *fn)
{
return gen_vv_ptr_vl(ctx, a, 32, fn);
}

static bool gen_vv_vl(DisasContext *ctx, arg_vv *a, uint32_t oprsz,
gen_helper_gvec_2 *fn)
{
Expand Down Expand Up @@ -4313,6 +4330,14 @@ TRANS(vfmul_s, LSX, gen_vvv_ptr, gen_helper_vfmul_s)
TRANS(vfmul_d, LSX, gen_vvv_ptr, gen_helper_vfmul_d)
TRANS(vfdiv_s, LSX, gen_vvv_ptr, gen_helper_vfdiv_s)
TRANS(vfdiv_d, LSX, gen_vvv_ptr, gen_helper_vfdiv_d)
TRANS(xvfadd_s, LASX, gen_xxx_ptr, gen_helper_vfadd_s)
TRANS(xvfadd_d, LASX, gen_xxx_ptr, gen_helper_vfadd_d)
TRANS(xvfsub_s, LASX, gen_xxx_ptr, gen_helper_vfsub_s)
TRANS(xvfsub_d, LASX, gen_xxx_ptr, gen_helper_vfsub_d)
TRANS(xvfmul_s, LASX, gen_xxx_ptr, gen_helper_vfmul_s)
TRANS(xvfmul_d, LASX, gen_xxx_ptr, gen_helper_vfmul_d)
TRANS(xvfdiv_s, LASX, gen_xxx_ptr, gen_helper_vfdiv_s)
TRANS(xvfdiv_d, LASX, gen_xxx_ptr, gen_helper_vfdiv_d)

TRANS(vfmadd_s, LSX, gen_vvvv_ptr, gen_helper_vfmadd_s)
TRANS(vfmadd_d, LSX, gen_vvvv_ptr, gen_helper_vfmadd_d)
Expand All @@ -4322,29 +4347,55 @@ TRANS(vfnmadd_s, LSX, gen_vvvv_ptr, gen_helper_vfnmadd_s)
TRANS(vfnmadd_d, LSX, gen_vvvv_ptr, gen_helper_vfnmadd_d)
TRANS(vfnmsub_s, LSX, gen_vvvv_ptr, gen_helper_vfnmsub_s)
TRANS(vfnmsub_d, LSX, gen_vvvv_ptr, gen_helper_vfnmsub_d)
TRANS(xvfmadd_s, LASX, gen_xxxx_ptr, gen_helper_vfmadd_s)
TRANS(xvfmadd_d, LASX, gen_xxxx_ptr, gen_helper_vfmadd_d)
TRANS(xvfmsub_s, LASX, gen_xxxx_ptr, gen_helper_vfmsub_s)
TRANS(xvfmsub_d, LASX, gen_xxxx_ptr, gen_helper_vfmsub_d)
TRANS(xvfnmadd_s, LASX, gen_xxxx_ptr, gen_helper_vfnmadd_s)
TRANS(xvfnmadd_d, LASX, gen_xxxx_ptr, gen_helper_vfnmadd_d)
TRANS(xvfnmsub_s, LASX, gen_xxxx_ptr, gen_helper_vfnmsub_s)
TRANS(xvfnmsub_d, LASX, gen_xxxx_ptr, gen_helper_vfnmsub_d)

TRANS(vfmax_s, LSX, gen_vvv_ptr, gen_helper_vfmax_s)
TRANS(vfmax_d, LSX, gen_vvv_ptr, gen_helper_vfmax_d)
TRANS(vfmin_s, LSX, gen_vvv_ptr, gen_helper_vfmin_s)
TRANS(vfmin_d, LSX, gen_vvv_ptr, gen_helper_vfmin_d)
TRANS(xvfmax_s, LASX, gen_xxx_ptr, gen_helper_vfmax_s)
TRANS(xvfmax_d, LASX, gen_xxx_ptr, gen_helper_vfmax_d)
TRANS(xvfmin_s, LASX, gen_xxx_ptr, gen_helper_vfmin_s)
TRANS(xvfmin_d, LASX, gen_xxx_ptr, gen_helper_vfmin_d)

TRANS(vfmaxa_s, LSX, gen_vvv_ptr, gen_helper_vfmaxa_s)
TRANS(vfmaxa_d, LSX, gen_vvv_ptr, gen_helper_vfmaxa_d)
TRANS(vfmina_s, LSX, gen_vvv_ptr, gen_helper_vfmina_s)
TRANS(vfmina_d, LSX, gen_vvv_ptr, gen_helper_vfmina_d)
TRANS(xvfmaxa_s, LASX, gen_xxx_ptr, gen_helper_vfmaxa_s)
TRANS(xvfmaxa_d, LASX, gen_xxx_ptr, gen_helper_vfmaxa_d)
TRANS(xvfmina_s, LASX, gen_xxx_ptr, gen_helper_vfmina_s)
TRANS(xvfmina_d, LASX, gen_xxx_ptr, gen_helper_vfmina_d)

TRANS(vflogb_s, LSX, gen_vv_ptr, gen_helper_vflogb_s)
TRANS(vflogb_d, LSX, gen_vv_ptr, gen_helper_vflogb_d)
TRANS(xvflogb_s, LASX, gen_xx_ptr, gen_helper_vflogb_s)
TRANS(xvflogb_d, LASX, gen_xx_ptr, gen_helper_vflogb_d)

TRANS(vfclass_s, LSX, gen_vv_ptr, gen_helper_vfclass_s)
TRANS(vfclass_d, LSX, gen_vv_ptr, gen_helper_vfclass_d)
TRANS(xvfclass_s, LASX, gen_xx_ptr, gen_helper_vfclass_s)
TRANS(xvfclass_d, LASX, gen_xx_ptr, gen_helper_vfclass_d)

TRANS(vfsqrt_s, LSX, gen_vv_ptr, gen_helper_vfsqrt_s)
TRANS(vfsqrt_d, LSX, gen_vv_ptr, gen_helper_vfsqrt_d)
TRANS(vfrecip_s, LSX, gen_vv_ptr, gen_helper_vfrecip_s)
TRANS(vfrecip_d, LSX, gen_vv_ptr, gen_helper_vfrecip_d)
TRANS(vfrsqrt_s, LSX, gen_vv_ptr, gen_helper_vfrsqrt_s)
TRANS(vfrsqrt_d, LSX, gen_vv_ptr, gen_helper_vfrsqrt_d)
TRANS(xvfsqrt_s, LASX, gen_xx_ptr, gen_helper_vfsqrt_s)
TRANS(xvfsqrt_d, LASX, gen_xx_ptr, gen_helper_vfsqrt_d)
TRANS(xvfrecip_s, LASX, gen_xx_ptr, gen_helper_vfrecip_s)
TRANS(xvfrecip_d, LASX, gen_xx_ptr, gen_helper_vfrecip_d)
TRANS(xvfrsqrt_s, LASX, gen_xx_ptr, gen_helper_vfrsqrt_s)
TRANS(xvfrsqrt_d, LASX, gen_xx_ptr, gen_helper_vfrsqrt_d)

TRANS(vfcvtl_s_h, LSX, gen_vv_ptr, gen_helper_vfcvtl_s_h)
TRANS(vfcvth_s_h, LSX, gen_vv_ptr, gen_helper_vfcvth_s_h)
Expand Down
41 changes: 41 additions & 0 deletions target/loongarch/insns.decode
View file
Original file line number Diff line number Diff line change
Expand Up @@ -1816,6 +1816,47 @@ xvfrstp_h 0111 01010010 10111 ..... ..... ..... @vvv
xvfrstpi_b 0111 01101001 10100 ..... ..... ..... @vv_ui5
xvfrstpi_h 0111 01101001 10101 ..... ..... ..... @vv_ui5

xvfadd_s 0111 01010011 00001 ..... ..... ..... @vvv
xvfadd_d 0111 01010011 00010 ..... ..... ..... @vvv
xvfsub_s 0111 01010011 00101 ..... ..... ..... @vvv
xvfsub_d 0111 01010011 00110 ..... ..... ..... @vvv
xvfmul_s 0111 01010011 10001 ..... ..... ..... @vvv
xvfmul_d 0111 01010011 10010 ..... ..... ..... @vvv
xvfdiv_s 0111 01010011 10101 ..... ..... ..... @vvv
xvfdiv_d 0111 01010011 10110 ..... ..... ..... @vvv

xvfmadd_s 0000 10100001 ..... ..... ..... ..... @vvvv
xvfmadd_d 0000 10100010 ..... ..... ..... ..... @vvvv
xvfmsub_s 0000 10100101 ..... ..... ..... ..... @vvvv
xvfmsub_d 0000 10100110 ..... ..... ..... ..... @vvvv
xvfnmadd_s 0000 10101001 ..... ..... ..... ..... @vvvv
xvfnmadd_d 0000 10101010 ..... ..... ..... ..... @vvvv
xvfnmsub_s 0000 10101101 ..... ..... ..... ..... @vvvv
xvfnmsub_d 0000 10101110 ..... ..... ..... ..... @vvvv

xvfmax_s 0111 01010011 11001 ..... ..... ..... @vvv
xvfmax_d 0111 01010011 11010 ..... ..... ..... @vvv
xvfmin_s 0111 01010011 11101 ..... ..... ..... @vvv
xvfmin_d 0111 01010011 11110 ..... ..... ..... @vvv

xvfmaxa_s 0111 01010100 00001 ..... ..... ..... @vvv
xvfmaxa_d 0111 01010100 00010 ..... ..... ..... @vvv
xvfmina_s 0111 01010100 00101 ..... ..... ..... @vvv
xvfmina_d 0111 01010100 00110 ..... ..... ..... @vvv

xvflogb_s 0111 01101001 11001 10001 ..... ..... @vv
xvflogb_d 0111 01101001 11001 10010 ..... ..... @vv

xvfclass_s 0111 01101001 11001 10101 ..... ..... @vv
xvfclass_d 0111 01101001 11001 10110 ..... ..... @vv

xvfsqrt_s 0111 01101001 11001 11001 ..... ..... @vv
xvfsqrt_d 0111 01101001 11001 11010 ..... ..... @vv
xvfrecip_s 0111 01101001 11001 11101 ..... ..... @vv
xvfrecip_d 0111 01101001 11001 11110 ..... ..... @vv
xvfrsqrt_s 0111 01101001 11010 00001 ..... ..... @vv
xvfrsqrt_d 0111 01101001 11010 00010 ..... ..... @vv

xvreplgr2vr_b 0111 01101001 11110 00000 ..... ..... @vr
xvreplgr2vr_h 0111 01101001 11110 00001 ..... ..... @vr
xvreplgr2vr_w 0111 01101001 11110 00010 ..... ..... @vr
Expand Down
12 changes: 8 additions & 4 deletions target/loongarch/vec_helper.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -2452,9 +2452,10 @@ void HELPER(NAME)(void *vd, void *vj, void *vk, \
VReg *Vd = (VReg *)vd; \
VReg *Vj = (VReg *)vj; \
VReg *Vk = (VReg *)vk; \
int oprsz = simd_oprsz(desc); \
\
vec_clear_cause(env); \
for (i = 0; i < LSX_LEN/BIT; i++) { \
for (i = 0; i < oprsz / (BIT / 8); i++) { \
Vd->E(i) = FN(Vj->E(i), Vk->E(i), &env->fp_status); \
vec_update_fcsr0(env, GETPC()); \
} \
Expand Down Expand Up @@ -2486,9 +2487,10 @@ void HELPER(NAME)(void *vd, void *vj, void *vk, void *va, \
VReg *Vj = (VReg *)vj; \
VReg *Vk = (VReg *)vk; \
VReg *Va = (VReg *)va; \
int oprsz = simd_oprsz(desc); \
\
vec_clear_cause(env); \
for (i = 0; i < LSX_LEN/BIT; i++) { \
for (i = 0; i < oprsz / (BIT / 8); i++) { \
Vd->E(i) = FN(Vj->E(i), Vk->E(i), Va->E(i), flags, &env->fp_status); \
vec_update_fcsr0(env, GETPC()); \
} \
Expand All @@ -2512,9 +2514,10 @@ void HELPER(NAME)(void *vd, void *vj, \
int i; \
VReg *Vd = (VReg *)vd; \
VReg *Vj = (VReg *)vj; \
int oprsz = simd_oprsz(desc); \
\
vec_clear_cause(env); \
for (i = 0; i < LSX_LEN/BIT; i++) { \
for (i = 0; i < oprsz / (BIT / 8); i++) { \
Vd->E(i) = FN(env, Vj->E(i)); \
} \
}
Expand Down Expand Up @@ -2544,8 +2547,9 @@ void HELPER(NAME)(void *vd, void *vj, \
int i; \
VReg *Vd = (VReg *)vd; \
VReg *Vj = (VReg *)vj; \
int oprsz = simd_oprsz(desc); \
\
for (i = 0; i < LSX_LEN/BIT; i++) { \
for (i = 0; i < oprsz / (BIT / 8); i++) { \
Vd->E(i) = FN(env, Vj->E(i)); \
} \
}
Expand Down

0 comments on commit c9caf15

Please sign in to comment.