yuzu/unicorn
1
0
Fork 0
mirror of https://github.com/yuzu-emu/unicorn.git synced 2025-04-19 12:31:47 +00:00
Code Issues Packages Projects Releases Wiki Activity

target/arm: Convert VCMLA, VCADD size field to MO_* in decode

Browse source 
The VCMLA and VCADD insns have a size field which is 0 for fp16
and 1 for fp32 (note that this is the reverse of the Neon 3-same
encoding!). Convert it to MO_* values in decode for consistency.

Backports d186a4854c04e9832907b0b4240a47731da20993
This commit is contained in:
Peter Maydell 2021-03-01 18:23:31 -05:00 committed by Lioncash
parent 61abec1908
commit 16ad0d93d9
2 changed files with 24 additions and 16 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

18
qemu/target/arm/neon-shared.decode
View file

@ -34,11 +34,17 @@
%vd_dp 22:1 12:4
%vd_sp 12:4 22:1
VCMLA 1111 110 rot:2 . 1 size:1 .... .... 1000 . q:1 . 0 .... \
vm=%vm_dp vn=%vn_dp vd=%vd_dp
# For VCMLA/VCADD insns, convert the single-bit size field
# which is 0 for fp16 and 1 for fp32 into a MO_* constant.
# (Note that this is the reverse of the sense of the 1-bit size
# field in the 3same_fp Neon insns.)
%vcadd_size 20:1 !function=plus1
VCADD 1111 110 rot:1 1 . 0 size:1 .... .... 1000 . q:1 . 0 .... \
vm=%vm_dp vn=%vn_dp vd=%vd_dp
VCMLA 1111 110 rot:2 . 1 . .... .... 1000 . q:1 . 0 .... \
vm=%vm_dp vn=%vn_dp vd=%vd_dp size=%vcadd_size
VCADD 1111 110 rot:1 1 . 0 . .... .... 1000 . q:1 . 0 .... \
vm=%vm_dp vn=%vn_dp vd=%vd_dp size=%vcadd_size
# VUDOT and VSDOT
VDOT 1111 110 00 . 10 .... .... 1101 . q:1 . u:1 .... \
@ -51,9 +57,9 @@ VFML 1111 110 0 s:1 . 10 .... .... 1000 . 1 . 1 .... \
vm=%vm_dp vn=%vn_dp vd=%vd_dp q=1
VCMLA_scalar 1111 1110 0 . rot:2 .... .... 1000 . q:1 index:1 0 vm:4 \
vn=%vn_dp vd=%vd_dp size=0
vn=%vn_dp vd=%vd_dp size=1
VCMLA_scalar 1111 1110 1 . rot:2 .... .... 1000 . q:1 . 0 .... \
vm=%vm_dp vn=%vn_dp vd=%vd_dp size=1 index=0
vm=%vm_dp vn=%vn_dp vd=%vd_dp size=2 index=0
VDOT_scalar 1111 1110 0 . 10 .... .... 1101 . q:1 index:1 u:1 rm:4 \
vm=%vm_dp vn=%vn_dp vd=%vd_dp

22
qemu/target/arm/translate-neon.inc.c
View file

@ -173,7 +173,7 @@ static bool trans_VCMLA(DisasContext *s, arg_VCMLA *a)
TCGContext *tcg_ctx = s->uc->tcg_ctx;
if (!dc_isar_feature(aa32_vcma, s)
|| (!a->size && !dc_isar_feature(aa32_fp16_arith, s))) {
|| (a->size == MO_16 && !dc_isar_feature(aa32_fp16_arith, s))) {
return false;
}
@ -192,8 +192,9 @@ static bool trans_VCMLA(DisasContext *s, arg_VCMLA *a)
}
opr_sz = (1 + a->q) * 8;
fpst = fpstatus_ptr(tcg_ctx, a->size == 0 ? FPST_STD_F16 : FPST_STD);
fn_gvec_ptr = a->size ? gen_helper_gvec_fcmlas : gen_helper_gvec_fcmlah;
fpst = fpstatus_ptr(tcg_ctx, a->size == MO_16 ? FPST_STD_F16 : FPST_STD);
fn_gvec_ptr = (a->size == MO_16) ?
gen_helper_gvec_fcmlah : gen_helper_gvec_fcmlas;
tcg_gen_gvec_3_ptr(tcg_ctx, vfp_reg_offset(1, a->vd),
vfp_reg_offset(1, a->vn),
vfp_reg_offset(1, a->vm),
@ -211,7 +212,7 @@ static bool trans_VCADD(DisasContext *s, arg_VCADD *a)
TCGContext *tcg_ctx = s->uc->tcg_ctx;
if (!dc_isar_feature(aa32_vcma, s)
|| (!a->size && !dc_isar_feature(aa32_fp16_arith, s))) {
|| (a->size == MO_16 && !dc_isar_feature(aa32_fp16_arith, s))) {
return false;
}
@ -230,8 +231,9 @@ static bool trans_VCADD(DisasContext *s, arg_VCADD *a)
}
opr_sz = (1 + a->q) * 8;
fpst = fpstatus_ptr(tcg_ctx, a->size == 0 ? FPST_STD_F16 : FPST_STD);
fn_gvec_ptr = a->size ? gen_helper_gvec_fcadds : gen_helper_gvec_fcaddh;
fpst = fpstatus_ptr(tcg_ctx, a->size == MO_16 ? FPST_STD_F16 : FPST_STD);
fn_gvec_ptr = (a->size == MO_16) ?
gen_helper_gvec_fcaddh : gen_helper_gvec_fcadds;
tcg_gen_gvec_3_ptr(tcg_ctx, vfp_reg_offset(1, a->vd),
vfp_reg_offset(1, a->vn),
vfp_reg_offset(1, a->vm),
@ -316,7 +318,7 @@ static bool trans_VCMLA_scalar(DisasContext *s, arg_VCMLA_scalar *a)
if (!dc_isar_feature(aa32_vcma, s)) {
return false;
}
if (a->size == 0 && !dc_isar_feature(aa32_fp16_arith, s)) {
if (a->size == MO_16 && !dc_isar_feature(aa32_fp16_arith, s)) {
return false;
}
@ -334,10 +336,10 @@ static bool trans_VCMLA_scalar(DisasContext *s, arg_VCMLA_scalar *a)
return true;
}
fn_gvec_ptr = (a->size ? gen_helper_gvec_fcmlas_idx
: gen_helper_gvec_fcmlah_idx);
fn_gvec_ptr = (a->size == MO_16) ?
gen_helper_gvec_fcmlah_idx : gen_helper_gvec_fcmlas_idx;
opr_sz = (1 + a->q) * 8;
fpst = fpstatus_ptr(tcg_ctx, a->size == 0 ? FPST_STD_F16 : FPST_STD);
fpst = fpstatus_ptr(tcg_ctx, a->size == MO_16 ? FPST_STD_F16 : FPST_STD);
tcg_gen_gvec_3_ptr(tcg_ctx, vfp_reg_offset(1, a->vd),
vfp_reg_offset(1, a->vn),
vfp_reg_offset(1, a->vm),