yuzu/unicorn
1
0
Fork 0
mirror of https://github.com/yuzu-emu/unicorn.git synced 2025-03-29 02:16:59 +00:00
Code Issues Packages Projects Releases Wiki Activity

target/arm: Convert VFP VNMLS to decodetree

Browse source 
Convert the VFP VNMLS instruction to decodetree.

Backports commit c54a416cc6d60efbc79dd37aaf0c8918c05b5815 from qemu
This commit is contained in:
Peter Maydell 2019-06-13 18:04:30 -04:00 committed by Lioncash
parent 67ad40ffa4
commit 638b90de31
No known key found for this signature in database
GPG key ID: 4E3C3CC1031BA9C7
3 changed files with 48 additions and 23 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

42
qemu/target/arm/translate-vfp.inc.c
View file

@ -1360,3 +1360,45 @@ static bool trans_VMLS_dp(DisasContext *s, arg_VMLS_sp *a)
{
return do_vfp_3op_dp(s, gen_VMLS_dp, a->vd, a->vn, a->vm, true);
}
static void gen_VNMLS_sp(TCGContext *tcg_ctx, TCGv_i32 vd, TCGv_i32 vn, TCGv_i32 vm, TCGv_ptr fpst)
{
/*
* VNMLS: -fd + (fn * fm)
* Note that it isn't valid to replace (-A + B) with (B - A) or similar
* plausible looking simplifications because this will give wrong results
* for NaNs.
*/
TCGv_i32 tmp = tcg_temp_new_i32(tcg_ctx);
gen_helper_vfp_muls(tcg_ctx, tmp, vn, vm, fpst);
gen_helper_vfp_negs(tcg_ctx, vd, vd);
gen_helper_vfp_adds(tcg_ctx, vd, vd, tmp, fpst);
tcg_temp_free_i32(tcg_ctx, tmp);
}
static bool trans_VNMLS_sp(DisasContext *s, arg_VNMLS_sp *a)
{
return do_vfp_3op_sp(s, gen_VNMLS_sp, a->vd, a->vn, a->vm, true);
}
static void gen_VNMLS_dp(TCGContext *tcg_ctx, TCGv_i64 vd, TCGv_i64 vn, TCGv_i64 vm, TCGv_ptr fpst)
{
/*
* VNMLS: -fd + (fn * fm)
* Note that it isn't valid to replace (-A + B) with (B - A) or similar
* plausible looking simplifications because this will give wrong results
* for NaNs.
*/
TCGv_i64 tmp = tcg_temp_new_i64(tcg_ctx);
gen_helper_vfp_muld(tcg_ctx, tmp, vn, vm, fpst);
gen_helper_vfp_negd(tcg_ctx, vd, vd);
gen_helper_vfp_addd(tcg_ctx, vd, vd, tmp, fpst);
tcg_temp_free_i64(tcg_ctx, tmp);
}
static bool trans_VNMLS_dp(DisasContext *s, arg_VNMLS_sp *a)
{
return do_vfp_3op_dp(s, gen_VNMLS_dp, a->vd, a->vn, a->vm, true);
}

24
qemu/target/arm/translate.c
View file

@ -1449,19 +1449,6 @@ VFP_OP2(div)
#undef VFP_OP2
static inline void gen_vfp_F1_mul(DisasContext *s, int dp)
{
TCGContext *tcg_ctx = s->uc->tcg_ctx;
/* Like gen_vfp_mul() but put result in F1 */
TCGv_ptr fpst = get_fpstatus_ptr(s, 0);
if (dp) {
gen_helper_vfp_muld(tcg_ctx, s->F1d, s->F0d, s->F1d, fpst);
} else {
gen_helper_vfp_muls(tcg_ctx, s->F1s, s->F0s, s->F1s, fpst);
}
tcg_temp_free_ptr(tcg_ctx, fpst);
}
static inline void gen_vfp_F1_neg(DisasContext *s, int dp)
{
TCGContext *tcg_ctx = s->uc->tcg_ctx;
@ -3240,6 +3227,7 @@ static int disas_vfp_insn(DisasContext *s, uint32_t insn)
switch (op) {
case 0:
case 1:
case 2:
/* Already handled by decodetree */
return 1;
default:
@ -3425,16 +3413,6 @@ static int disas_vfp_insn(DisasContext *s, uint32_t insn)
for (;;) {
/* Perform the calculation. */
switch (op) {
case 2: /* VNMLS: -fd + (fn * fm) */
/* Note that it isn't valid to replace (-A + B) with (B - A)
* or similar plausible looking simplifications
* because this will give wrong results for NaNs.
*/
gen_vfp_F1_mul(s, dp);
gen_mov_F0_vreg(s, dp, rd);
gen_vfp_neg(s, dp);
gen_vfp_add(s, dp);
break;
case 3: /* VNMLA: -fd + -(fn * fm) */
gen_vfp_mul(s, dp);
gen_vfp_F1_neg(s, dp);

5
qemu/target/arm/vfp.decode
View file

@ -107,3 +107,8 @@ VMLS_sp ---- 1110 0.00 .... .... 1010 .1.0 .... \
vm=%vm_sp vn=%vn_sp vd=%vd_sp
VMLS_dp ---- 1110 0.00 .... .... 1011 .1.0 .... \
vm=%vm_dp vn=%vn_dp vd=%vd_dp
VNMLS_sp ---- 1110 0.01 .... .... 1010 .0.0 .... \
vm=%vm_sp vn=%vn_sp vd=%vd_sp
VNMLS_dp ---- 1110 0.01 .... .... 1011 .0.0 .... \
vm=%vm_dp vn=%vn_dp vd=%vd_dp