mirror of
https://github.com/yuzu-emu/unicorn.git
synced 2025-01-22 22:51:01 +00:00
target/arm: Reorg NEON VLD/VST all elements
Instead of shifts and masks, use direct loads and stores from the neon register file. Mirror the iteration structure of the ARM pseudocode more closely. Correct the parameters of the VLD2 A2 insn. Note that this includes a bugfix for handling of the insn "VLD2 (multiple 2-element structures)" -- we were using an incorrect stride value. Backports commit ac55d00709e78cd39dfa298dcaac7aecb58762e8 from qemu
This commit is contained in:
Richard Henderson
Lioncash
parent
a2239b9f5b
commit
1cab7a41ac
|
|
@ -1682,6 +1682,29 @@ static TCGv_i32 neon_load_reg(DisasContext *s, int reg, int pass)
|
|||
return tmp;
|
||||
}
|
||||
|
||||
static void neon_load_element64(DisasContext *s, TCGv_i64 var, int reg, int ele, TCGMemOp mop)
|
||||
{
|
||||
TCGContext *tcg_ctx = s->uc->tcg_ctx;
|
||||
long offset = neon_element_offset(reg, ele, mop & MO_SIZE);
|
||||
|
||||
switch (mop) {
|
||||
case MO_UB:
|
||||
tcg_gen_ld8u_i64(tcg_ctx, var, tcg_ctx->cpu_env, offset);
|
||||
break;
|
||||
case MO_UW:
|
||||
tcg_gen_ld16u_i64(tcg_ctx, var, tcg_ctx->cpu_env, offset);
|
||||
break;
|
||||
case MO_UL:
|
||||
tcg_gen_ld32u_i64(tcg_ctx, var, tcg_ctx->cpu_env, offset);
|
||||
break;
|
||||
case MO_Q:
|
||||
tcg_gen_ld_i64(tcg_ctx, var, tcg_ctx->cpu_env, offset);
|
||||
break;
|
||||
default:
|
||||
g_assert_not_reached();
|
||||
}
|
||||
}
|
||||
|
||||
static void neon_store_reg(DisasContext *s, int reg, int pass, TCGv_i32 var)
|
||||
{
|
||||
TCGContext *tcg_ctx = s->uc->tcg_ctx;
|
||||
|
|
@ -1689,6 +1712,29 @@ static void neon_store_reg(DisasContext *s, int reg, int pass, TCGv_i32 var)
|
|||
tcg_temp_free_i32(tcg_ctx, var);
|
||||
}
|
||||
|
||||
static void neon_store_element64(DisasContext *s, int reg, int ele, TCGMemOp size, TCGv_i64 var)
|
||||
{
|
||||
TCGContext *tcg_ctx = s->uc->tcg_ctx;
|
||||
long offset = neon_element_offset(reg, ele, size);
|
||||
|
||||
switch (size) {
|
||||
case MO_8:
|
||||
tcg_gen_st8_i64(tcg_ctx, var, tcg_ctx->cpu_env, offset);
|
||||
break;
|
||||
case MO_16:
|
||||
tcg_gen_st16_i64(tcg_ctx, var, tcg_ctx->cpu_env, offset);
|
||||
break;
|
||||
case MO_32:
|
||||
tcg_gen_st32_i64(tcg_ctx, var, tcg_ctx->cpu_env, offset);
|
||||
break;
|
||||
case MO_64:
|
||||
tcg_gen_st_i64(tcg_ctx, var, tcg_ctx->cpu_env, offset);
|
||||
break;
|
||||
default:
|
||||
g_assert_not_reached();
|
||||
}
|
||||
}
|
||||
|
||||
static inline void neon_load_reg64(DisasContext *s, TCGv_i64 var, int reg)
|
||||
{
|
||||
TCGContext *tcg_ctx = s->uc->tcg_ctx;
|
||||
|
|
@ -5030,16 +5076,16 @@ static struct {
|
|||
int interleave;
|
||||
int spacing;
|
||||
} const neon_ls_element_type[11] = {
|
||||
{4, 4, 1},
|
||||
{4, 4, 2},
|
||||
{1, 4, 1},
|
||||
{1, 4, 2},
|
||||
{4, 1, 1},
|
||||
{4, 2, 1},
|
||||
{3, 3, 1},
|
||||
{3, 3, 2},
|
||||
{2, 2, 2},
|
||||
{1, 3, 1},
|
||||
{1, 3, 2},
|
||||
{3, 1, 1},
|
||||
{1, 1, 1},
|
||||
{2, 2, 1},
|
||||
{2, 2, 2},
|
||||
{1, 2, 1},
|
||||
{1, 2, 2},
|
||||
{2, 1, 1}
|
||||
};
|
||||
|
||||
|
|
@ -5061,6 +5107,8 @@ static int disas_neon_ls_insn(DisasContext *s, uint32_t insn)
|
|||
int shift;
|
||||
int n;
|
||||
int vec_size;
|
||||
int mmu_idx;
|
||||
TCGMemOp endian;
|
||||
TCGv_i32 addr;
|
||||
TCGv_i32 tmp;
|
||||
TCGv_i32 tmp2;
|
||||
|
|
@ -5082,6 +5130,8 @@ static int disas_neon_ls_insn(DisasContext *s, uint32_t insn)
|
|||
rn = (insn >> 16) & 0xf;
|
||||
rm = insn & 0xf;
|
||||
load = (insn & (1 << 21)) != 0;
|
||||
endian = s->be_data;
|
||||
mmu_idx = get_mem_index(s);
|
||||
if ((insn & (1 << 23)) == 0) {
|
||||
/* Load store all elements. */
|
||||
op = (insn >> 8) & 0xf;
|
||||
|
|
@ -5106,104 +5156,34 @@ static int disas_neon_ls_insn(DisasContext *s, uint32_t insn)
|
|||
nregs = neon_ls_element_type[op].nregs;
|
||||
interleave = neon_ls_element_type[op].interleave;
|
||||
spacing = neon_ls_element_type[op].spacing;
|
||||
if (size == 3 && (interleave | spacing) != 1)
|
||||
if (size == 3 && (interleave | spacing) != 1) {
|
||||
return 1;
|
||||
}
|
||||
tmp64 = tcg_temp_new_i64(tcg_ctx);
|
||||
addr = tcg_temp_new_i32(tcg_ctx);
|
||||
tmp2 = tcg_const_i32(tcg_ctx, 1 << size);
|
||||
load_reg_var(s, addr, rn);
|
||||
stride = (1 << size) * interleave;
|
||||
for (reg = 0; reg < nregs; reg++) {
|
||||
if (interleave > 2 || (interleave == 2 && nregs == 2)) {
|
||||
load_reg_var(s, addr, rn);
|
||||
tcg_gen_addi_i32(tcg_ctx, addr, addr, (1 << size) * reg);
|
||||
} else if (interleave == 2 && nregs == 4 && reg == 2) {
|
||||
load_reg_var(s, addr, rn);
|
||||
tcg_gen_addi_i32(tcg_ctx, addr, addr, 1 << size);
|
||||
}
|
||||
if (size == 3) {
|
||||
tmp64 = tcg_temp_new_i64(tcg_ctx);
|
||||
if (load) {
|
||||
gen_aa32_ld64(s, tmp64, addr, get_mem_index(s));
|
||||
neon_store_reg64(s, tmp64, rd);
|
||||
} else {
|
||||
neon_load_reg64(s, tmp64, rd);
|
||||
gen_aa32_st64(s, tmp64, addr, get_mem_index(s));
|
||||
}
|
||||
tcg_temp_free_i64(tcg_ctx, tmp64);
|
||||
tcg_gen_addi_i32(tcg_ctx, addr, addr, stride);
|
||||
} else {
|
||||
for (pass = 0; pass < 2; pass++) {
|
||||
if (size == 2) {
|
||||
if (load) {
|
||||
tmp = tcg_temp_new_i32(tcg_ctx);
|
||||
gen_aa32_ld32u(s, tmp, addr, get_mem_index(s));
|
||||
neon_store_reg(s, rd, pass, tmp);
|
||||
} else {
|
||||
tmp = neon_load_reg(s, rd, pass);
|
||||
gen_aa32_st32(s, tmp, addr, get_mem_index(s));
|
||||
tcg_temp_free_i32(tcg_ctx, tmp);
|
||||
}
|
||||
tcg_gen_addi_i32(tcg_ctx, addr, addr, stride);
|
||||
} else if (size == 1) {
|
||||
if (load) {
|
||||
tmp = tcg_temp_new_i32(tcg_ctx);
|
||||
gen_aa32_ld16u(s, tmp, addr, get_mem_index(s));
|
||||
tcg_gen_addi_i32(tcg_ctx, addr, addr, stride);
|
||||
tmp2 = tcg_temp_new_i32(tcg_ctx);
|
||||
gen_aa32_ld16u(s, tmp2, addr, get_mem_index(s));
|
||||
tcg_gen_addi_i32(tcg_ctx, addr, addr, stride);
|
||||
tcg_gen_shli_i32(tcg_ctx, tmp2, tmp2, 16);
|
||||
tcg_gen_or_i32(tcg_ctx, tmp, tmp, tmp2);
|
||||
tcg_temp_free_i32(tcg_ctx, tmp2);
|
||||
neon_store_reg(s, rd, pass, tmp);
|
||||
} else {
|
||||
tmp = neon_load_reg(s, rd, pass);
|
||||
tmp2 = tcg_temp_new_i32(tcg_ctx);
|
||||
tcg_gen_shri_i32(tcg_ctx, tmp2, tmp, 16);
|
||||
gen_aa32_st16(s, tmp, addr, get_mem_index(s));
|
||||
tcg_temp_free_i32(tcg_ctx, tmp);
|
||||
tcg_gen_addi_i32(tcg_ctx, addr, addr, stride);
|
||||
gen_aa32_st16(s, tmp2, addr, get_mem_index(s));
|
||||
tcg_temp_free_i32(tcg_ctx, tmp2);
|
||||
tcg_gen_addi_i32(tcg_ctx, addr, addr, stride);
|
||||
}
|
||||
} else /* size == 0 */ {
|
||||
if (load) {
|
||||
tmp2 = NULL;
|
||||
for (n = 0; n < 4; n++) {
|
||||
tmp = tcg_temp_new_i32(tcg_ctx);
|
||||
gen_aa32_ld8u(s, tmp, addr, get_mem_index(s));
|
||||
tcg_gen_addi_i32(tcg_ctx, addr, addr, stride);
|
||||
if (n == 0) {
|
||||
tmp2 = tmp;
|
||||
} else {
|
||||
tcg_gen_shli_i32(tcg_ctx, tmp, tmp, n * 8);
|
||||
tcg_gen_or_i32(tcg_ctx, tmp2, tmp2, tmp);
|
||||
tcg_temp_free_i32(tcg_ctx, tmp);
|
||||
}
|
||||
}
|
||||
neon_store_reg(s, rd, pass, tmp2);
|
||||
} else {
|
||||
tmp2 = neon_load_reg(s, rd, pass);
|
||||
for (n = 0; n < 4; n++) {
|
||||
tmp = tcg_temp_new_i32(tcg_ctx);
|
||||
if (n == 0) {
|
||||
tcg_gen_mov_i32(tcg_ctx, tmp, tmp2);
|
||||
} else {
|
||||
tcg_gen_shri_i32(tcg_ctx, tmp, tmp2, n * 8);
|
||||
}
|
||||
gen_aa32_st8(s, tmp, addr, get_mem_index(s));
|
||||
tcg_temp_free_i32(tcg_ctx, tmp);
|
||||
tcg_gen_addi_i32(tcg_ctx, addr, addr, stride);
|
||||
}
|
||||
tcg_temp_free_i32(tcg_ctx, tmp2);
|
||||
}
|
||||
for (n = 0; n < 8 >> size; n++) {
|
||||
int xs;
|
||||
for (xs = 0; xs < interleave; xs++) {
|
||||
int tt = rd + reg + spacing * xs;
|
||||
|
||||
if (load) {
|
||||
gen_aa32_ld_i64(s, tmp64, addr, mmu_idx, endian | size);
|
||||
neon_store_element64(s, tt, n, size, tmp64);
|
||||
} else {
|
||||
neon_load_element64(s, tmp64, tt, n, size);
|
||||
gen_aa32_st_i64(s, tmp64, addr, mmu_idx, endian | size);
|
||||
}
|
||||
tcg_gen_add_i32(tcg_ctx, addr, addr, tmp2);
|
||||
}
|
||||
}
|
||||
rd += spacing;
|
||||
}
|
||||
tcg_temp_free_i32(tcg_ctx, addr);
|
||||
stride = nregs * 8;
|
||||
tcg_temp_free_i32(tcg_ctx, tmp2);
|
||||
tcg_temp_free_i64(tcg_ctx, tmp64);
|
||||
stride = nregs * interleave * 8;
|
||||
} else {
|
||||
size = (insn >> 10) & 3;
|
||||
if (size == 3) {
|
||||
|
|
|
|||
Loading…
Reference in a new issue