qemu/target/arm/tcg/translate-sme.c

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

376 lines
12 KiB
C
Raw Normal View History

/*
* AArch64 SME translation
*
* Copyright (c) 2022 Linaro, Ltd
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, see <http://www.gnu.org/licenses/>.
*/
#include "qemu/osdep.h"
#include "translate.h"
#include "translate-a64.h"
/*
* Include the generated decoder.
*/
#include "decode-sme.c.inc"
/*
* Resolve tile.size[index] to a host pointer, where tile and index
* are always decoded together, dependent on the element size.
*/
static TCGv_ptr get_tile_rowcol(DisasContext *s, int esz, int rs,
int tile_index, bool vertical)
{
int tile = tile_index >> (4 - esz);
int index = esz == MO_128 ? 0 : extract32(tile_index, 0, 4 - esz);
int pos, len, offset;
TCGv_i32 tmp;
TCGv_ptr addr;
/* Compute the final index, which is Rs+imm. */
tmp = tcg_temp_new_i32();
tcg_gen_trunc_tl_i32(tmp, cpu_reg(s, rs));
tcg_gen_addi_i32(tmp, tmp, index);
/* Prepare a power-of-two modulo via extraction of @len bits. */
len = ctz32(streaming_vec_reg_size(s)) - esz;
if (!len) {
/*
* SVL is 128 and the element size is 128. There is exactly
* one 128x128 tile in the ZA storage, and so we calculate
* (Rs + imm) MOD 1, which is always 0. We need to special case
* this because TCG doesn't allow deposit ops with len 0.
*/
tcg_gen_movi_i32(tmp, 0);
} else if (vertical) {
/*
* Compute the byte offset of the index within the tile:
* (index % (svl / size)) * size
* = (index % (svl >> esz)) << esz
* Perform the power-of-two modulo via extraction of the low @len bits.
* Perform the multiply by shifting left by @pos bits.
* Perform these operations simultaneously via deposit into zero.
*/
pos = esz;
tcg_gen_deposit_z_i32(tmp, tmp, pos, len);
/*
* For big-endian, adjust the indexed column byte offset within
* the uint64_t host words that make up env->zarray[].
*/
if (HOST_BIG_ENDIAN && esz < MO_64) {
tcg_gen_xori_i32(tmp, tmp, 8 - (1 << esz));
}
} else {
/*
* Compute the byte offset of the index within the tile:
* (index % (svl / size)) * (size * sizeof(row))
* = (index % (svl >> esz)) << (esz + log2(sizeof(row)))
*/
pos = esz + ctz32(sizeof(ARMVectorReg));
tcg_gen_deposit_z_i32(tmp, tmp, pos, len);
/* Row slices are always aligned and need no endian adjustment. */
}
/* The tile byte offset within env->zarray is the row. */
offset = tile * sizeof(ARMVectorReg);
/* Include the byte offset of zarray to make this relative to env. */
offset += offsetof(CPUARMState, zarray);
tcg_gen_addi_i32(tmp, tmp, offset);
/* Add the byte offset to env to produce the final pointer. */
addr = tcg_temp_new_ptr();
tcg_gen_ext_i32_ptr(addr, tmp);
tcg_gen_add_ptr(addr, addr, tcg_env);
return addr;
}
/*
* Resolve tile.size[0] to a host pointer.
* Used by e.g. outer product insns where we require the entire tile.
*/
static TCGv_ptr get_tile(DisasContext *s, int esz, int tile)
{
TCGv_ptr addr = tcg_temp_new_ptr();
int offset;
offset = tile * sizeof(ARMVectorReg) + offsetof(CPUARMState, zarray);
tcg_gen_addi_ptr(addr, tcg_env, offset);
return addr;
}
static bool trans_ZERO(DisasContext *s, arg_ZERO *a)
{
if (!dc_isar_feature(aa64_sme, s)) {
return false;
}
if (sme_za_enabled_check(s)) {
gen_helper_sme_zero(tcg_env, tcg_constant_i32(a->imm),
tcg_constant_i32(streaming_vec_reg_size(s)));
}
return true;
}
static bool trans_MOVA(DisasContext *s, arg_MOVA *a)
{
static gen_helper_gvec_4 * const h_fns[5] = {
gen_helper_sve_sel_zpzz_b, gen_helper_sve_sel_zpzz_h,
gen_helper_sve_sel_zpzz_s, gen_helper_sve_sel_zpzz_d,
gen_helper_sve_sel_zpzz_q
};
static gen_helper_gvec_3 * const cz_fns[5] = {
gen_helper_sme_mova_cz_b, gen_helper_sme_mova_cz_h,
gen_helper_sme_mova_cz_s, gen_helper_sme_mova_cz_d,
gen_helper_sme_mova_cz_q,
};
static gen_helper_gvec_3 * const zc_fns[5] = {
gen_helper_sme_mova_zc_b, gen_helper_sme_mova_zc_h,
gen_helper_sme_mova_zc_s, gen_helper_sme_mova_zc_d,
gen_helper_sme_mova_zc_q,
};
TCGv_ptr t_za, t_zr, t_pg;
TCGv_i32 t_desc;
int svl;
if (!dc_isar_feature(aa64_sme, s)) {
return false;
}
if (!sme_smza_enabled_check(s)) {
return true;
}
t_za = get_tile_rowcol(s, a->esz, a->rs, a->za_imm, a->v);
t_zr = vec_full_reg_ptr(s, a->zr);
t_pg = pred_full_reg_ptr(s, a->pg);
svl = streaming_vec_reg_size(s);
t_desc = tcg_constant_i32(simd_desc(svl, svl, 0));
if (a->v) {
/* Vertical slice -- use sme mova helpers. */
if (a->to_vec) {
zc_fns[a->esz](t_zr, t_za, t_pg, t_desc);
} else {
cz_fns[a->esz](t_za, t_zr, t_pg, t_desc);
}
} else {
/* Horizontal slice -- reuse sve sel helpers. */
if (a->to_vec) {
h_fns[a->esz](t_zr, t_za, t_zr, t_pg, t_desc);
} else {
h_fns[a->esz](t_za, t_zr, t_za, t_pg, t_desc);
}
}
return true;
}
static bool trans_LDST1(DisasContext *s, arg_LDST1 *a)
{
typedef void GenLdSt1(TCGv_env, TCGv_ptr, TCGv_ptr, TCGv, TCGv_i32);
/*
* Indexed by [esz][be][v][mte][st], which is (except for load/store)
* also the order in which the elements appear in the function names,
* and so how we must concatenate the pieces.
*/
#define FN_LS(F) { gen_helper_sme_ld1##F, gen_helper_sme_st1##F }
#define FN_MTE(F) { FN_LS(F), FN_LS(F##_mte) }
#define FN_HV(F) { FN_MTE(F##_h), FN_MTE(F##_v) }
#define FN_END(L, B) { FN_HV(L), FN_HV(B) }
static GenLdSt1 * const fns[5][2][2][2][2] = {
FN_END(b, b),
FN_END(h_le, h_be),
FN_END(s_le, s_be),
FN_END(d_le, d_be),
FN_END(q_le, q_be),
};
#undef FN_LS
#undef FN_MTE
#undef FN_HV
#undef FN_END
TCGv_ptr t_za, t_pg;
TCGv_i64 addr;
uint32_t desc;
bool be = s->be_data == MO_BE;
bool mte = s->mte_active[0];
if (!dc_isar_feature(aa64_sme, s)) {
return false;
}
if (!sme_smza_enabled_check(s)) {
return true;
}
t_za = get_tile_rowcol(s, a->esz, a->rs, a->za_imm, a->v);
t_pg = pred_full_reg_ptr(s, a->pg);
addr = tcg_temp_new_i64();
tcg_gen_shli_i64(addr, cpu_reg(s, a->rm), a->esz);
tcg_gen_add_i64(addr, addr, cpu_reg_sp(s, a->rn));
if (!mte) {
addr = clean_data_tbi(s, addr);
}
desc = make_svemte_desc(s, streaming_vec_reg_size(s), 1, a->esz, a->st, 0);
fns[a->esz][be][a->v][mte][a->st](tcg_env, t_za, t_pg, addr,
tcg_constant_i32(desc));
return true;
}
typedef void GenLdStR(DisasContext *, TCGv_ptr, int, int, int, int);
static bool do_ldst_r(DisasContext *s, arg_ldstr *a, GenLdStR *fn)
{
int svl = streaming_vec_reg_size(s);
int imm = a->imm;
TCGv_ptr base;
if (!sme_za_enabled_check(s)) {
return true;
}
/* ZA[n] equates to ZA0H.B[n]. */
base = get_tile_rowcol(s, MO_8, a->rv, imm, false);
fn(s, base, 0, svl, a->rn, imm * svl);
return true;
}
TRANS_FEAT(LDR, aa64_sme, do_ldst_r, a, gen_sve_ldr)
TRANS_FEAT(STR, aa64_sme, do_ldst_r, a, gen_sve_str)
static bool do_adda(DisasContext *s, arg_adda *a, MemOp esz,
gen_helper_gvec_4 *fn)
{
int svl = streaming_vec_reg_size(s);
uint32_t desc = simd_desc(svl, svl, 0);
TCGv_ptr za, zn, pn, pm;
if (!sme_smza_enabled_check(s)) {
return true;
}
za = get_tile(s, esz, a->zad);
zn = vec_full_reg_ptr(s, a->zn);
pn = pred_full_reg_ptr(s, a->pn);
pm = pred_full_reg_ptr(s, a->pm);
fn(za, zn, pn, pm, tcg_constant_i32(desc));
return true;
}
TRANS_FEAT(ADDHA_s, aa64_sme, do_adda, a, MO_32, gen_helper_sme_addha_s)
TRANS_FEAT(ADDVA_s, aa64_sme, do_adda, a, MO_32, gen_helper_sme_addva_s)
TRANS_FEAT(ADDHA_d, aa64_sme_i16i64, do_adda, a, MO_64, gen_helper_sme_addha_d)
TRANS_FEAT(ADDVA_d, aa64_sme_i16i64, do_adda, a, MO_64, gen_helper_sme_addva_d)
static bool do_outprod(DisasContext *s, arg_op *a, MemOp esz,
gen_helper_gvec_5 *fn)
{
int svl = streaming_vec_reg_size(s);
uint32_t desc = simd_desc(svl, svl, a->sub);
TCGv_ptr za, zn, zm, pn, pm;
if (!sme_smza_enabled_check(s)) {
return true;
}
za = get_tile(s, esz, a->zad);
zn = vec_full_reg_ptr(s, a->zn);
zm = vec_full_reg_ptr(s, a->zm);
pn = pred_full_reg_ptr(s, a->pn);
pm = pred_full_reg_ptr(s, a->pm);
fn(za, zn, zm, pn, pm, tcg_constant_i32(desc));
return true;
}
static bool do_outprod_fpst(DisasContext *s, arg_op *a, MemOp esz,
ARMFPStatusFlavour e_fpst,
gen_helper_gvec_5_ptr *fn)
{
int svl = streaming_vec_reg_size(s);
uint32_t desc = simd_desc(svl, svl, a->sub);
TCGv_ptr za, zn, zm, pn, pm, fpst;
if (!sme_smza_enabled_check(s)) {
return true;
}
za = get_tile(s, esz, a->zad);
zn = vec_full_reg_ptr(s, a->zn);
zm = vec_full_reg_ptr(s, a->zm);
pn = pred_full_reg_ptr(s, a->pn);
pm = pred_full_reg_ptr(s, a->pm);
fpst = fpstatus_ptr(e_fpst);
fn(za, zn, zm, pn, pm, fpst, tcg_constant_i32(desc));
return true;
}
target/arm: Handle denormals correctly for FMOPA (widening) The FMOPA (widening) SME instruction takes pairs of half-precision floating point values, widens them to single-precision, does a two-way dot product and accumulates the results into a single-precision destination. We don't quite correctly handle the FPCR bits FZ and FZ16 which control flushing of denormal inputs and outputs. This is because at the moment we pass a single float_status value to the helper function, which then uses that configuration for all the fp operations it does. However, because the inputs to this operation are float16 and the outputs are float32 we need to use the fp_status_f16 for the float16 input widening but the normal fp_status for everything else. Otherwise we will apply the flushing control FPCR.FZ16 to the 32-bit output rather than the FPCR.FZ control, and incorrectly flush a denormal output to zero when we should not (or vice-versa). (In commit 207d30b5fdb5b we tried to fix the FZ handling but didn't get it right, switching from "use FPCR.FZ for everything" to "use FPCR.FZ16 for everything".) Pass the CPU env to the sme_fmopa_h helper instead of an fp_status pointer, and have the helper pass an extra fp_status into the f16_dotadd() function so that we can use the right status for the right parts of this operation. Cc: qemu-stable@nongnu.org Fixes: 207d30b5fdb5 ("target/arm: Use FPST_F16 for SME FMOPA (widening)") Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2373 Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
2024-08-01 12:15:03 +03:00
static bool do_outprod_env(DisasContext *s, arg_op *a, MemOp esz,
gen_helper_gvec_5_ptr *fn)
{
int svl = streaming_vec_reg_size(s);
uint32_t desc = simd_desc(svl, svl, a->sub);
TCGv_ptr za, zn, zm, pn, pm;
if (!sme_smza_enabled_check(s)) {
return true;
}
za = get_tile(s, esz, a->zad);
zn = vec_full_reg_ptr(s, a->zn);
zm = vec_full_reg_ptr(s, a->zm);
pn = pred_full_reg_ptr(s, a->pn);
pm = pred_full_reg_ptr(s, a->pm);
fn(za, zn, zm, pn, pm, tcg_env, tcg_constant_i32(desc));
return true;
}
TRANS_FEAT(FMOPA_h, aa64_sme, do_outprod_env, a,
MO_32, gen_helper_sme_fmopa_h)
TRANS_FEAT(FMOPA_s, aa64_sme, do_outprod_fpst, a,
MO_32, FPST_FPCR, gen_helper_sme_fmopa_s)
TRANS_FEAT(FMOPA_d, aa64_sme_f64f64, do_outprod_fpst, a,
MO_64, FPST_FPCR, gen_helper_sme_fmopa_d)
TRANS_FEAT(BFMOPA, aa64_sme, do_outprod_env, a, MO_32, gen_helper_sme_bfmopa)
TRANS_FEAT(SMOPA_s, aa64_sme, do_outprod, a, MO_32, gen_helper_sme_smopa_s)
TRANS_FEAT(UMOPA_s, aa64_sme, do_outprod, a, MO_32, gen_helper_sme_umopa_s)
TRANS_FEAT(SUMOPA_s, aa64_sme, do_outprod, a, MO_32, gen_helper_sme_sumopa_s)
TRANS_FEAT(USMOPA_s, aa64_sme, do_outprod, a, MO_32, gen_helper_sme_usmopa_s)
TRANS_FEAT(SMOPA_d, aa64_sme_i16i64, do_outprod, a, MO_64, gen_helper_sme_smopa_d)
TRANS_FEAT(UMOPA_d, aa64_sme_i16i64, do_outprod, a, MO_64, gen_helper_sme_umopa_d)
TRANS_FEAT(SUMOPA_d, aa64_sme_i16i64, do_outprod, a, MO_64, gen_helper_sme_sumopa_d)
TRANS_FEAT(USMOPA_d, aa64_sme_i16i64, do_outprod, a, MO_64, gen_helper_sme_usmopa_d)