On Wed, Mar 23, 2022 at 2:05 PM liuhongt via Gcc-patches
<gcc-patches@gcc.gnu.org> wrote:
>
> In validate_subreg, both (subreg:V2HF (reg:SI) 0)
> and (subreg:V8HF (reg:V2HF) 0) are valid, but not
> for (subreg:V8HF (reg:SI) 0) which causes ICE.
>
> Ideally it should be handled in validate_subreg to support
> subreg for all modes available in TARGET_CAN_CHANGE_MODE_CLASS, but
> that would be too risky in stage4, so the patch is a walkround in the
> backend to force_reg operands before lowpart_subreg for expanders or
> pre_reload splitters.
>
> Bootstrapped and regtest on x86_64-pc-linux-gnu{-m32,}.
> Also with native on SPR.
> Ok for trunk?
>
> gcc/ChangeLog:
>
> PR target/104976
> * config/i386/sse.md (ssePSmodelower): New.
> (*avx_cmp<mode>3_ltint_not): Force_reg operand before
> lowpart_subreg to avoid NULL_RTX.
> (<avx512>_fmaddc_<mode>_mask1<round_expand_name>,
> <avx512>_fcmaddc_<mode>_mask1<round_expand_name>,
> fma_<mode>_fmaddc_bcst, fma_<mode>_fcmaddc_bcst,
> <avx512>_<complexopname>_<mode>_mask<round_name>,
> avx512fp16_fcmaddcsh_v8hf_mask1<round_expand_name>,
> avx512fp16_fcmaddcsh_v8hf_mask3<round_expand_name>,
> avx512fp16_fmaddcsh_v8hf_mask3<round_expand_name>,
> avx512fp16_fmaddcsh_v8hf_mask3<round_expand_name>,
> float<floatunssuffix><mode>v4hf2,
> float<floatunssuffix>v2div2hf2,
> fix<fixunssuffix>_truncv4hf<mode>2,
> fix<fixunssuffix>_truncv2hfv2di2, extendv4hf<mode>2,
> extendv2hfv2df2,
> trunc<mode>v4hf2,truncv2dfv2hf2,
> *avx512bw_permvar_truncv16siv16hi_1,
> *avx512bw_permvar_truncv16siv16hi_1_hf,
> *avx512f_permvar_truncv8siv8hi_1,
> *avx512f_permvar_truncv8siv8hi_1_hf,
> *avx512f_vpermvar_truncv8div8si_1,
> *avx512f_permvar_truncv32hiv32qi_1,
> *avx512f_permvar_truncv16hiv16qi_1,
> *avx512f_permvar_truncv4div4si_1,
> *avx512f_pshufb_truncv8hiv8qi_1,
> *avx512f_pshufb_truncv4siv4hi_1,
> *avx512f_pshufd_truncv2div2si_1,
> sdot_prod<mode>, avx2_pblend<ssemodesuffix>_1,
> ashrv2di3,ashrv2di3,usdot_prod<mode>): Ditto.
>
> gcc/testsuite/ChangeLog:
>
> * gcc.target/i386/pr104976.c: New test.
> * gcc.target/i386/avx512fp16-vfcmaddcph-1a.c: Scan either
> vblendps or masked vmovaps.
> * gcc.target/i386/avx512fp16-vfmaddcph-1a.c: Ditto
> * gcc.target/i386/avx512fp16vl-vfcmaddcph-1a.c: Ditto.
> * gcc.target/i386/avx512fp16vl-vfmaddcph-1a.c: Ditto.
> ---
> gcc/config/i386/sse.md | 236 +++++++++++++-----
> .../i386/avx512fp16-vfcmaddcph-1a.c | 2 +-
> .../gcc.target/i386/avx512fp16-vfmaddcph-1a.c | 2 +-
> .../i386/avx512fp16vl-vfcmaddcph-1a.c | 4 +-
> .../i386/avx512fp16vl-vfmaddcph-1a.c | 4 +-
> gcc/testsuite/gcc.target/i386/pr104976.c | 13 +
> 6 files changed, 196 insertions(+), 65 deletions(-)
> create mode 100644 gcc/testsuite/gcc.target/i386/pr104976.c
>
> diff --git a/gcc/config/i386/sse.md b/gcc/config/i386/sse.md
> index 6f7af2f21d6..a9e18d38323 100644
> --- a/gcc/config/i386/sse.md
> +++ b/gcc/config/i386/sse.md
> @@ -1059,6 +1059,18 @@ (define_mode_attr ssePSmode
> (V4DF "V8SF") (V2DF "V4SF")
> (V32HF "V16SF") (V16HF "V8SF") (V8HF "V4SF")])
>
> +(define_mode_attr ssePSmodelower
> + [(V16SI "v16sf") (V8DF "v16sf")
> + (V16SF "v16sf") (V8DI "v16sf")
> + (V64QI "v16sf") (V32QI "v8sf") (V16QI "v4sf")
> + (V32HI "v16sf") (V16HI "v8sf") (V8HI "v4sf")
> + (V8SI "v8sf") (V4SI "v4sf")
> + (V4DI "v8sf") (V2DI "v4sf")
> + (V4TI "v16sf") (V2TI "v8sf") (V1TI "v4sf")
> + (V8SF "v8sf") (V4SF "v4sf")
> + (V4DF "v8sf") (V2DF "v4sf")
> + (V32HF "v16sf") (V16HF "v8sf") (V8HF "v4sf")])
> +
> (define_mode_attr ssePSmode2
> [(V8DI "V8SF") (V4DI "V4SF")])
>
> @@ -3617,6 +3629,9 @@ (define_insn_and_split "*avx_cmp<mode>3_ltint_not"
> operands[1] = force_reg (<ssebytemode>mode,
> gen_lowpart (<ssebytemode>mode, operands[1]));
> operands[2] = gen_lowpart (<ssebytemode>mode, operands[2]);
> +
> + if (!MEM_P (operands[3]))
> + operands[3] = force_reg (<ssebytemode>mode, operands[3]);
> operands[3] = lowpart_subreg (<MODE>mode, operands[3], <ssebytemode>mode);
> })
>
> @@ -6319,7 +6334,7 @@ (define_expand
> "<avx512>_fmaddc_<mode>_mask1<round_expand_name>"
> (match_operand:<avx512fmaskcmode> 4 "register_operand")]
> "TARGET_AVX512FP16 && <round_mode512bit_condition>"
> {
> - rtx op0, op1;
> + rtx op0, op1, dest;
> if (<round_embedded_complex>)
> emit_insn (gen_<avx512>_fmaddc_<mode>_mask<round_expand_name> (
> operands[0], operands[1], operands[2], operands[3],
> @@ -6328,9 +6343,16 @@ (define_expand
> "<avx512>_fmaddc_<mode>_mask1<round_expand_name>"
> emit_insn (gen_<avx512>_fmaddc_<mode>_mask (operands[0],
> operands[1], operands[2], operands[3], operands[4]));
>
> - op0 = lowpart_subreg (<ssePSmode>mode, operands[0], <MODE>mode);
> + op0 = lowpart_subreg (<ssePSmode>mode,
> + force_reg (<MODE>mode, operands[0]),
> + <MODE>mode);
> + dest = gen_reg_rtx (<ssePSmode>mode);
> + if (!MEM_P (operands[1]))
> + operands[1] = force_reg (<MODE>mode, operands[1]);
> op1 = lowpart_subreg (<ssePSmode>mode, operands[1], <MODE>mode);
> - emit_insn (gen_<complexmove>_mask (op0, op0, op1, operands[4]));
> + emit_insn (gen_<complexmove>_mask (dest, op0, op1, operands[4]));
> + emit_move_insn (operands[0],
> + lowpart_subreg (<MODE>mode, dest, <ssePSmode>mode));
> DONE;
> })
>
> @@ -6356,7 +6378,7 @@ (define_expand
> "<avx512>_fcmaddc_<mode>_mask1<round_expand_name>"
> (match_operand:<avx512fmaskcmode> 4 "register_operand")]
> "TARGET_AVX512FP16 && <round_mode512bit_condition>"
> {
> - rtx op0, op1;
> + rtx op0, op1, dest;
> if (<round_embedded_complex>)
> emit_insn (gen_<avx512>_fcmaddc_<mode>_mask<round_expand_name> (
> operands[0], operands[1], operands[2], operands[3],
> @@ -6367,9 +6389,16 @@ (define_expand
> "<avx512>_fcmaddc_<mode>_mask1<round_expand_name>"
> operands[1], operands[2], operands[3], operands[4]));
> }
>
> - op0 = lowpart_subreg (<ssePSmode>mode, operands[0], <MODE>mode);
> + op0 = lowpart_subreg (<ssePSmode>mode,
> + force_reg (<MODE>mode, operands[0]),
> + <MODE>mode);
> + dest = gen_reg_rtx (<ssePSmode>mode);
> + if (!MEM_P (operands[1]))
> + operands[1] = force_reg (<MODE>mode, operands[1]);
> op1 = lowpart_subreg (<ssePSmode>mode, operands[1], <MODE>mode);
> - emit_insn (gen_<complexmove>_mask (op0, op0, op1, operands[4]));
> + emit_insn (gen_<complexmove>_mask (dest, op0, op1, operands[4]));
> + emit_move_insn (operands[0],
> + lowpart_subreg (<MODE>mode, dest, <ssePSmode>mode));
> DONE;
> })
>
> @@ -6479,17 +6508,23 @@ (define_insn_and_split "fma_<mode>_fmaddc_bcst"
> (match_operand:<ssePSmode> 2 "bcst_vector_operand") 0)
> (match_operand:VF_AVX512FP16VL 3 "vector_operand")]
> UNSPEC_COMPLEX_FMA))]
> - "TARGET_AVX512FP16"
> + "TARGET_AVX512FP16 && ix86_pre_reload_split ()"
> "#"
> "&& 1"
> - [(set (match_dup 0)
> - (unspec:<ssePSmode>
> - [(match_dup 1) (match_dup 2) (match_dup 3)]
> - UNSPEC_COMPLEX_FMA_PAIR))]
> + [(const_int 0)]
> {
> - operands[0] = lowpart_subreg (<ssePSmode>mode, operands[0], <MODE>mode);
> + if (!MEM_P (operands[1]))
> + operands[1] = force_reg (<MODE>mode, operands[1]);
> + if (!MEM_P (operands[3]))
> + operands[3] = force_reg (<MODE>mode, operands[3]);
> operands[1] = lowpart_subreg (<ssePSmode>mode, operands[1], <MODE>mode);
> operands[3] = lowpart_subreg (<ssePSmode>mode, operands[3], <MODE>mode);
> + rtx dest = gen_reg_rtx (<ssePSmode>mode);
> + emit_insn (gen_fma_fmaddc_<ssePSmodelower>_pair (dest, operands[1],
> + operands[2],
> operands[3]));
> + emit_move_insn (operands[0],
> + lowpart_subreg (<MODE>mode, dest, <ssePSmode>mode));
> + DONE;
> })
>
> (define_insn_and_split "fma_<mode>_fcmaddc_bcst"
> @@ -6500,17 +6535,24 @@ (define_insn_and_split "fma_<mode>_fcmaddc_bcst"
> (match_operand:<ssePSmode> 2 "bcst_vector_operand") 0)
> (match_operand:VF_AVX512FP16VL 3 "vector_operand")]
> UNSPEC_COMPLEX_FCMA))]
> - "TARGET_AVX512FP16"
> + "TARGET_AVX512FP16 && ix86_pre_reload_split ()"
> "#"
> "&& 1"
> - [(set (match_dup 0)
> - (unspec:<ssePSmode>
> - [(match_dup 1) (match_dup 2) (match_dup 3)]
> - UNSPEC_COMPLEX_FCMA_PAIR))]
> + [(const_int 0)]
> {
> - operands[0] = lowpart_subreg (<ssePSmode>mode, operands[0], <MODE>mode);
> + if (!MEM_P (operands[1]))
> + operands[1] = force_reg (<MODE>mode, operands[1]);
> + if (!MEM_P (operands[3]))
> + operands[3] = force_reg (<MODE>mode, operands[3]);
> operands[1] = lowpart_subreg (<ssePSmode>mode, operands[1], <MODE>mode);
> operands[3] = lowpart_subreg (<ssePSmode>mode, operands[3], <MODE>mode);
> + rtx dest = gen_reg_rtx (<ssePSmode>mode);
> + emit_insn (gen_fma_fcmaddc_<ssePSmodelower>_pair (dest, operands[1],
> + operands[2],
> + operands[3]));
> + emit_move_insn (operands[0],
> + lowpart_subreg (<MODE>mode, dest, <ssePSmode>mode));
> + DONE;
> })
>
> (define_insn "<avx512>_<complexopname>_<mode>_mask<round_name>"
> @@ -6650,7 +6692,7 @@ (define_expand
> "avx512fp16_fcmaddcsh_v8hf_mask3<round_expand_name>"
> (match_operand:QI 4 "register_operand")]
> "TARGET_AVX512FP16 && <round_mode512bit_condition>"
> {
> - rtx op0, op1;
> + rtx dest, op0, op1;
>
> if (<round_embedded_complex>)
> emit_insn (gen_avx512fp16_fcmaddcsh_v8hf_mask<round_expand_name> (
> @@ -6660,9 +6702,15 @@ (define_expand
> "avx512fp16_fcmaddcsh_v8hf_mask3<round_expand_name>"
> emit_insn (gen_avx512fp16_fcmaddcsh_v8hf_mask (operands[0],
> operands[1], operands[2], operands[3], operands[4]));
>
> - op0 = lowpart_subreg (V4SFmode, operands[0], V8HFmode);
> + dest = gen_reg_rtx (V4SFmode);
> + op0 = lowpart_subreg (V4SFmode,
> + force_reg (V8HFmode, operands[0]),
> + V8HFmode);
> + if (!MEM_P (operands[3]))
> + operands[3] = force_reg (V8HFmode, operands[3]);
> op1 = lowpart_subreg (V4SFmode, operands[3], V8HFmode);
> - emit_insn (gen_sse_movss (op0, op1, op0));
> + emit_insn (gen_sse_movss (dest, op1, op0));
> + emit_move_insn (operands[0], lowpart_subreg (V8HFmode, dest, V4SFmode));
> DONE;
> })
>
> @@ -6674,7 +6722,7 @@ (define_expand
> "avx512fp16_fmaddcsh_v8hf_mask3<round_expand_name>"
> (match_operand:QI 4 "register_operand")]
> "TARGET_AVX512FP16 && <round_mode512bit_condition>"
> {
> - rtx op0, op1;
> + rtx dest, op0, op1;
>
> if (<round_embedded_complex>)
> emit_insn (gen_avx512fp16_fmaddcsh_v8hf_mask<round_expand_name> (
> @@ -6684,9 +6732,15 @@ (define_expand
> "avx512fp16_fmaddcsh_v8hf_mask3<round_expand_name>"
> emit_insn (gen_avx512fp16_fmaddcsh_v8hf_mask (operands[0],
> operands[1], operands[2], operands[3], operands[4]));
>
> - op0 = lowpart_subreg (V4SFmode, operands[0], V8HFmode);
> + dest = gen_reg_rtx (V4SFmode);
> + op0 = lowpart_subreg (V4SFmode,
> + force_reg (V8HFmode, operands[0]),
> + V8HFmode);
> + if (!MEM_P (operands[3]))
> + operands[3] = force_reg (V8HFmode, operands[3]);
> op1 = lowpart_subreg (V4SFmode, operands[3], V8HFmode);
> - emit_insn (gen_sse_movss (op0, op1, op0));
> + emit_insn (gen_sse_movss (dest, op1, op0));
> + emit_move_insn (operands[0], lowpart_subreg (V8HFmode, dest, V4SFmode));
> DONE;
> })
>
> @@ -6796,9 +6850,10 @@ (define_expand "float<floatunssuffix><mode>v4hf2"
> (match_operand:VI4_128_8_256 1 "vector_operand")))]
> "TARGET_AVX512FP16 && TARGET_AVX512VL"
> {
> - operands[0] = lowpart_subreg (V8HFmode, operands[0], V4HFmode);
> - emit_insn (gen_avx512fp16_float<floatunssuffix><mode>v4hf2 (operands[0],
> + rtx dest = gen_reg_rtx (V8HFmode);
> + emit_insn (gen_avx512fp16_float<floatunssuffix><mode>v4hf2 (dest,
> operands[1]));
> + emit_move_insn (operands[0], lowpart_subreg (V4HFmode, dest, V8HFmode));
> DONE;
> })
>
> @@ -6871,9 +6926,10 @@ (define_expand "float<floatunssuffix>v2div2hf2"
> (match_operand:V2DI 1 "vector_operand")))]
> "TARGET_AVX512FP16 && TARGET_AVX512VL"
> {
> - operands[0] = lowpart_subreg (V8HFmode, operands[0], V2HFmode);
> - emit_insn (gen_avx512fp16_float<floatunssuffix>v2div2hf2 (operands[0],
> + rtx dest = gen_reg_rtx (V8HFmode);
> + emit_insn (gen_avx512fp16_float<floatunssuffix>v2div2hf2 (dest,
> operands[1]));
> + emit_move_insn (operands[0], lowpart_subreg (V2HFmode, dest, V8HFmode));
> DONE;
> })
>
> @@ -7004,7 +7060,9 @@ (define_expand "fix<fixunssuffix>_truncv4hf<mode>2"
> {
> if (!MEM_P (operands[1]))
> {
> - operands[1] = lowpart_subreg (V8HFmode, operands[1], V4HFmode);
> + operands[1] = lowpart_subreg (V8HFmode,
> + force_reg (V4HFmode, operands[1]),
> + V4HFmode);
> emit_insn (gen_avx512fp16_fix<fixunssuffix>_trunc<mode>2 (operands[0],
> operands[1]));
> DONE;
> @@ -7041,7 +7099,9 @@ (define_expand "fix<fixunssuffix>_truncv2hfv2di2"
> {
> if (!MEM_P (operands[1]))
> {
> - operands[1] = lowpart_subreg (V8HFmode, operands[1], V2HFmode);
> + operands[1] = lowpart_subreg (V8HFmode,
> + force_reg (V2HFmode, operands[1]),
> + V2HFmode);
> emit_insn (gen_avx512fp16_fix<fixunssuffix>_truncv2di2 (operands[0],
> operands[1]));
> DONE;
> @@ -7120,7 +7180,9 @@ (define_expand "extendv4hf<mode>2"
> {
> if (!MEM_P (operands[1]))
> {
> - operands[1] = lowpart_subreg (V8HFmode, operands[1], V4HFmode);
> + operands[1] = lowpart_subreg (V8HFmode,
> + force_reg (V4HFmode, operands[1]),
> + V4HFmode);
> emit_insn (gen_avx512fp16_float_extend_ph<mode>2
> (operands[0], operands[1]));
> DONE;
> @@ -7157,7 +7219,9 @@ (define_expand "extendv2hfv2df2"
> {
> if (!MEM_P (operands[1]))
> {
> - operands[1] = lowpart_subreg (V8HFmode, operands[1], V2HFmode);
> + operands[1] = lowpart_subreg (V8HFmode,
> + force_reg (V2HFmode, operands[1]),
> + V2HFmode);
> emit_insn (gen_avx512fp16_float_extend_phv2df2
> (operands[0], operands[1]));
> DONE;
> @@ -7207,8 +7271,10 @@ (define_expand "trunc<mode>v4hf2"
> (float_truncate:V4HF (match_operand:VF4_128_8_256 1
> "vector_operand")))]
> "TARGET_AVX512FP16 && TARGET_AVX512VL"
> {
> - operands[0] = lowpart_subreg (V8HFmode, operands[0], V4HFmode);
> - emit_insn (gen_avx512fp16_trunc<mode>v4hf2 (operands[0], operands[1]));
> + rtx dest = gen_reg_rtx (V8HFmode);
> +
> + emit_insn (gen_avx512fp16_trunc<mode>v4hf2 (dest, operands[1]));
> + emit_move_insn (operands[0], lowpart_subreg (V4HFmode, dest, V8HFmode));
> DONE;
> })
>
> @@ -7286,8 +7352,9 @@ (define_expand "truncv2dfv2hf2"
> (float_truncate:V2HF (match_operand:V2DF 1 "vector_operand")))]
> "TARGET_AVX512FP16 && TARGET_AVX512VL"
> {
> - operands[0] = lowpart_subreg (V8HFmode, operands[0], V2HFmode);
> - emit_insn (gen_avx512fp16_truncv2dfv2hf2 (operands[0], operands[1]));
> + rtx dest = gen_reg_rtx (V8HFmode);
> + emit_insn (gen_avx512fp16_truncv2dfv2hf2 (dest, operands[1]));
> + emit_move_insn (operands[0], lowpart_subreg (V2HFmode, dest, V8HFmode));
> DONE;
> })
>
> @@ -13384,7 +13451,9 @@ (define_insn_and_split
> "*avx512bw_permvar_truncv16siv16hi_1"
> "&& 1"
> [(set (match_dup 0)
> (truncate:V16HI (match_dup 1)))]
> - "operands[1] = lowpart_subreg (V16SImode, operands[1], V32HImode);")
> + "operands[1] = lowpart_subreg (V16SImode,
> + force_reg (V32HImode, operands[1]),
> + V32HImode);")
>
> (define_insn_and_split "*avx512bw_permvar_truncv16siv16hi_1_hf"
> [(set (match_operand:V16HF 0 "nonimmediate_operand")
> @@ -13408,8 +13477,18 @@ (define_insn_and_split
> "*avx512bw_permvar_truncv16siv16hi_1_hf"
> [(set (match_dup 0)
> (truncate:V16HI (match_dup 1)))]
> {
> - operands[0] = lowpart_subreg (V16HImode, operands[0], V16HFmode);
> - operands[1] = lowpart_subreg (V16SImode, operands[1], V32HImode);
> + operands[1] = lowpart_subreg (V16SImode,
> + force_reg (V32HImode, operands[1]),
> + V32HImode);
> + if (MEM_P (operands[0]))
> + operands[0] = lowpart_subreg (V16HImode, operands[0], V16HFmode);
> + else
> + {
> + rtx op0 = gen_reg_rtx (V16HImode);
> + emit_insn (gen_truncv16siv16hi2 (op0, operands[1]));
> + emit_move_insn (operands[0], lowpart_subreg (V16HFmode, op0,
> V16HImode));
> + DONE;
> + }
> })
>
>
> @@ -13429,7 +13508,9 @@ (define_insn_and_split
> "*avx512f_permvar_truncv8siv8hi_1"
> "&& 1"
> [(set (match_dup 0)
> (truncate:V8HI (match_dup 1)))]
> - "operands[1] = lowpart_subreg (V8SImode, operands[1], V16HImode);")
> + "operands[1] = lowpart_subreg (V8SImode,
> + force_reg (V16HImode, operands[1]),
> + V16HImode);")
>
> (define_insn_and_split "*avx512f_permvar_truncv8siv8hi_1_hf"
> [(set (match_operand:V8HF 0 "nonimmediate_operand")
> @@ -13449,8 +13530,18 @@ (define_insn_and_split
> "*avx512f_permvar_truncv8siv8hi_1_hf"
> [(set (match_dup 0)
> (truncate:V8HI (match_dup 1)))]
> {
> - operands[0] = lowpart_subreg (V8HImode, operands[0], V8HFmode);
> - operands[1] = lowpart_subreg (V8SImode, operands[1], V16HImode);
> + operands[1] = lowpart_subreg (V8SImode,
> + force_reg (V16HImode, operands[1]),
> + V16HImode);
> + if (MEM_P (operands[0]))
> + operands[0] = lowpart_subreg (V8HImode, operands[0], V8HFmode);
> + else
> + {
> + rtx op0 = gen_reg_rtx (V8HImode);
> + emit_insn (gen_truncv8siv8hi2 (op0, operands[1]));
> + emit_move_insn (operands[0], lowpart_subreg (V8HFmode, op0, V8HImode));
> + DONE;
> + }
> })
>
> (define_insn_and_split "*avx512f_vpermvar_truncv8div8si_1"
> @@ -13469,7 +13560,9 @@ (define_insn_and_split
> "*avx512f_vpermvar_truncv8div8si_1"
> "&& 1"
> [(set (match_dup 0)
> (truncate:V8SI (match_dup 1)))]
> - "operands[1] = lowpart_subreg (V8DImode, operands[1], V16SImode);")
> + "operands[1] = lowpart_subreg (V8DImode,
> + force_reg (V16SImode, operands[1]),
> + V16SImode);")
>
> (define_insn "avx512f_<code><pmov_src_lower><mode>2_mask"
> [(set (match_operand:PMOV_DST_MODE_1 0 "nonimmediate_operand" "=v,m")
> @@ -13539,7 +13632,9 @@ (define_insn_and_split
> "*avx512f_permvar_truncv32hiv32qi_1"
> "&& 1"
> [(set (match_dup 0)
> (truncate:V32QI (match_dup 1)))]
> - "operands[1] = lowpart_subreg (V32HImode, operands[1], V64QImode);")
> + "operands[1] = lowpart_subreg (V32HImode,
> + force_reg (V64QImode, operands[1]),
> + V64QImode);")
>
> (define_insn "avx512bw_<code>v32hiv32qi2_mask"
> [(set (match_operand:V32QI 0 "nonimmediate_operand" "=v,m")
> @@ -13607,7 +13702,9 @@ (define_insn_and_split
> "*avx512f_permvar_truncv16hiv16qi_1"
> "&& 1"
> [(set (match_dup 0)
> (truncate:V16QI (match_dup 1)))]
> - "operands[1] = lowpart_subreg (V16HImode, operands[1], V32QImode);")
> + "operands[1] = lowpart_subreg (V16HImode,
> + force_reg (V32QImode, operands[1]),
> + V32QImode);")
>
> (define_insn_and_split "*avx512f_permvar_truncv4div4si_1"
> [(set (match_operand:V4SI 0 "nonimmediate_operand")
> @@ -13623,7 +13720,9 @@ (define_insn_and_split
> "*avx512f_permvar_truncv4div4si_1"
> "&& 1"
> [(set (match_dup 0)
> (truncate:V4SI (match_dup 1)))]
> - "operands[1] = lowpart_subreg (V4DImode, operands[1], V8SImode);")
> + "operands[1] = lowpart_subreg (V4DImode,
> + force_reg (V8SImode, operands[1]),
> + V8SImode);")
>
> (define_insn "<avx512>_<code><ssedoublemodelower><mode>2_mask"
> [(set (match_operand:PMOV_DST_MODE_2 0 "nonimmediate_operand" "=v,m")
> @@ -13701,7 +13800,9 @@ (define_insn_and_split
> "*avx512f_pshufb_truncv8hiv8qi_1"
> [(const_int 0)]
> {
> rtx op1 = gen_reg_rtx (V8QImode);
> - operands[1] = lowpart_subreg (V8HImode, operands[1], V16QImode);
> + operands[1] = lowpart_subreg (V8HImode,
> + force_reg (V16QImode, operands[1]),
> + V16QImode);
> emit_insn (gen_truncv8hiv8qi2 (op1, operands[1]));
> emit_move_insn (operands[0], lowpart_subreg (DImode, op1, V8QImode));
> DONE;
> @@ -14079,7 +14180,9 @@ (define_insn_and_split
> "*avx512f_pshufb_truncv4siv4hi_1"
> [(const_int 0)]
> {
> rtx op1 = gen_reg_rtx (V4HImode);
> - operands[1] = lowpart_subreg (V4SImode, operands[1], V16QImode);
> + operands[1] = lowpart_subreg (V4SImode,
> + force_reg (V16QImode, operands[1]),
> + V16QImode);
> emit_insn (gen_truncv4siv4hi2 (op1, operands[1]));
> emit_move_insn (operands[0], lowpart_subreg (DImode, op1, V4HImode));
> DONE;
> @@ -14326,7 +14429,9 @@ (define_insn_and_split
> "*avx512f_pshufd_truncv2div2si_1"
> [(const_int 0)]
> {
> rtx op1 = gen_reg_rtx (V2SImode);
> - operands[1] = lowpart_subreg (V2DImode, operands[1], V4SImode);
> + operands[1] = lowpart_subreg (V2DImode,
> + force_reg (V4SImode, operands[1]),
> + V4SImode);
> emit_insn (gen_truncv2div2si2 (op1, operands[1]));
> emit_move_insn (operands[0], lowpart_subreg (DImode, op1, V2SImode));
> DONE;
> @@ -15360,8 +15465,12 @@ (define_expand "sdot_prod<mode>"
> || (<MODE_SIZE> < 64
> && ((TARGET_AVX512VNNI && TARGET_AVX512VL) || TARGET_AVXVNNI)))
> {
> - operands[1] = lowpart_subreg (<sseunpackmode>mode, operands[1],
> <MODE>mode);
> - operands[2] = lowpart_subreg (<sseunpackmode>mode, operands[2],
> <MODE>mode);
> + operands[1] = lowpart_subreg (<sseunpackmode>mode,
> + force_reg (<MODE>mode, operands[1]),
> + <MODE>mode);
> + operands[2] = lowpart_subreg (<sseunpackmode>mode,
> + force_reg (<MODE>mode, operands[2]),
> + <MODE>mode);
> emit_insn (gen_rtx_SET (operands[0], operands[3]));
> emit_insn (gen_vpdpwssd_<SDOT_VPDP_SUF> (operands[0], operands[3],
> operands[1], operands[2]));
> @@ -21376,11 +21485,12 @@ (define_expand "avx2_pblend<ssemodesuffix>_1"
> emit_insn (gen_avx2_pblend<ssemodesuffix> (tmp, operands[1],
> operands[2], blendw_idx));
>
> - operands[0] = lowpart_subreg (V8SImode, operands[0], <MODE>mode);
> + rtx op0 = gen_reg_rtx (V8SImode);
> tmp = lowpart_subreg (V8SImode, tmp, <MODE>mode);
> operands[1] = lowpart_subreg (V8SImode, operands[1], <MODE>mode);
> - emit_insn (gen_avx2_pblenddv8si (operands[0], operands[1],
> + emit_insn (gen_avx2_pblenddv8si (op0, operands[1],
> tmp, blendd_idx));
> + emit_move_insn (operands[0], lowpart_subreg (<MODE>mode, op0,
> V8SImode));
> }
>
> DONE;
> @@ -24205,7 +24315,9 @@ (define_expand "ashrv2di3"
> vec_perm_builder sel (4, 4, 1);
> sel.quick_grow (4);
> rtx arg0, arg1;
> - rtx op1 = lowpart_subreg (V4SImode, operands[1], V2DImode);
> + rtx op1 = lowpart_subreg (V4SImode,
> + force_reg (V2DImode, operands[1]),
> + V2DImode);
> rtx target = gen_reg_rtx (V4SImode);
> if (UINTVAL (operands[2]) >= 63)
> {
> @@ -24271,9 +24383,11 @@ (define_expand "ashrv2di3"
> else
> {
> rtx temp = gen_reg_rtx (V4SImode);
> - emit_insn (gen_ashrv4si3 (temp, lowpart_subreg (V4SImode,
> - operands[1],
> - V2DImode),
> + emit_insn (gen_ashrv4si3 (temp,
> + lowpart_subreg (V4SImode,
> + force_reg (V2DImode,
> +
> operands[1]),
> + V2DImode),
> GEN_INT (31)));
> zero_or_all_ones = gen_reg_rtx (V4SImode);
> emit_insn (gen_sse2_pshufd_1 (zero_or_all_ones, temp,
> @@ -27768,8 +27882,12 @@ (define_expand "usdot_prod<mode>"
> ||((TARGET_AVX512VNNI && TARGET_AVX512VL)
> || TARGET_AVXVNNI))"
> {
> - operands[1] = lowpart_subreg (<VI1SI>mode, operands[1], <MODE>mode);
> - operands[2] = lowpart_subreg (<VI1SI>mode, operands[2], <MODE>mode);
> + operands[1] = lowpart_subreg (<VI1SI>mode,
> + force_reg (<MODE>mode, operands[1]),
> + <MODE>mode);
> + operands[2] = lowpart_subreg (<VI1SI>mode,
> + force_reg (<MODE>mode, operands[2]),
> + <MODE>mode);
> emit_insn (gen_rtx_SET (operands[0], operands[3]));
> emit_insn (gen_vpdpbusd_<vi1si> (operands[0], operands[3],
> operands[1], operands[2]));
> diff --git a/gcc/testsuite/gcc.target/i386/avx512fp16-vfcmaddcph-1a.c
> b/gcc/testsuite/gcc.target/i386/avx512fp16-vfcmaddcph-1a.c
> index cd39b7f99ff..851561ad416 100644
> --- a/gcc/testsuite/gcc.target/i386/avx512fp16-vfcmaddcph-1a.c
> +++ b/gcc/testsuite/gcc.target/i386/avx512fp16-vfcmaddcph-1a.c
> @@ -6,7 +6,7 @@
> /* { dg-final { scan-assembler-times "vfcmaddcph\[
> \\t\]+\{rn-sae\}\[^\{\n\]*%zmm\[0-9\]+\[^\n\r]*%zmm\[0-9\]+\[^\n\r]*%zmm\[0-9\]+(?:\n|\[
> \\t\]+#)" 1 } } */
> /* { dg-final { scan-assembler-times "vfcmaddcph\[
> \\t\]+\{rn-sae\}\[^\{\n\]*%zmm\[0-9\]+\[^\n\r]*%zmm\[0-9\]+\[^\n\r]*%zmm\[0-9\]+\{%k\[0-9\]\}\[^\n\r]*(?:\n|\[
> \\t\]+#)" 2 } } */
> /* { dg-final { scan-assembler-times "vfcmaddcph\[
> \\t\]+\{rz-sae\}\[^\{\n\]*%zmm\[0-9\]+\[^\n\r]*%zmm\[0-9\]+\[^\n\r]*%zmm\[0-9\]+\{%k\[0-9\]\}\{z\}\[^\n\r]*(?:\n|\[
> \\t\]+#)" 1 } } */
> -/* { dg-final { scan-assembler-times "vblendmps\[
> \\t\]+%zmm\[0-9\]+\[^\n\r]*%zmm\[0-9\]+\[^\n\r]*%zmm\[0-9\]+\{%k\[0-9\]\}(?:\n|\[
> \\t\]+#)" 2 } } */
> +/* { dg-final { scan-assembler-times "(?:vblendmps\[
> \\t\]+%zmm\[0-9\]+|vmovaps\[
> \\t\]+)\[^\n\r]*%zmm\[0-9\]+\[^\n\r]*%zmm\[0-9\]+\{%k\[0-9\]\}(?:\n|\[
> \\t\]+#)" 2 } } */
>
> #include <immintrin.h>
>
> diff --git a/gcc/testsuite/gcc.target/i386/avx512fp16-vfmaddcph-1a.c
> b/gcc/testsuite/gcc.target/i386/avx512fp16-vfmaddcph-1a.c
> index 859b215ab17..dfcc6e5f93c 100644
> --- a/gcc/testsuite/gcc.target/i386/avx512fp16-vfmaddcph-1a.c
> +++ b/gcc/testsuite/gcc.target/i386/avx512fp16-vfmaddcph-1a.c
> @@ -6,7 +6,7 @@
> /* { dg-final { scan-assembler-times "vfmaddcph\[
> \\t\]+\{rn-sae\}\[^\{\n\]*%zmm\[0-9\]+\[^\n\r]*%zmm\[0-9\]+\[^\n\r]*%zmm\[0-9\]+(?:\n|\[
> \\t\]+#)" 1 } } */
> /* { dg-final { scan-assembler-times "vfmaddcph\[
> \\t\]+\{rn-sae\}\[^\{\n\]*%zmm\[0-9\]+\[^\n\r]*%zmm\[0-9\]+\[^\n\r]*%zmm\[0-9\]+\{%k\[0-9\]\}\[^\n\r]*(?:\n|\[
> \\t\]+#)" 2 } } */
> /* { dg-final { scan-assembler-times "vfmaddcph\[
> \\t\]+\{rz-sae\}\[^\{\n\]*%zmm\[0-9\]+\[^\n\r]*%zmm\[0-9\]+\[^\n\r]*%zmm\[0-9\]+\{%k\[0-9\]\}\{z\}\[^\n\r]*(?:\n|\[
> \\t\]+#)" 1 } } */
> -/* { dg-final { scan-assembler-times "vblendmps\[
> \\t\]+%zmm\[0-9\]+\[^\n\r]*%zmm\[0-9\]+\[^\n\r]*%zmm\[0-9\]+\{%k\[0-9\]\}(?:\n|\[
> \\t\]+#)" 2 } } */
> +/* { dg-final { scan-assembler-times "(?:vblendmps\[
> \\t\]+%zmm\[0-9\]+|vmovaps\[
> \\t\]+)\[^\n\r]*%zmm\[0-9\]+\[^\n\r]*%zmm\[0-9\]+\{%k\[0-9\]\}(?:\n|\[
> \\t\]+#)" 2 } } */
>
> #include <immintrin.h>
>
> diff --git a/gcc/testsuite/gcc.target/i386/avx512fp16vl-vfcmaddcph-1a.c
> b/gcc/testsuite/gcc.target/i386/avx512fp16vl-vfcmaddcph-1a.c
> index 627241c0403..93edb92984e 100644
> --- a/gcc/testsuite/gcc.target/i386/avx512fp16vl-vfcmaddcph-1a.c
> +++ b/gcc/testsuite/gcc.target/i386/avx512fp16vl-vfcmaddcph-1a.c
> @@ -3,11 +3,11 @@
> /* { dg-final { scan-assembler-times "vfcmaddcph\[
> \\t\]+\[^\{\n\]*%ymm\[0-9\]+\[^\n\r]*%ymm\[0-9\]+(?:\n|\[ \\t\]+#)" 1 } } */
> /* { dg-final { scan-assembler-times "vfcmaddcph\[
> \\t\]+%ymm\[0-9\]+\[^\n\r]*%ymm\[0-9\]+\[^\n\r]*%ymm\[0-9\]+\{%k\[0-9\]\}(?:\n|\[
> \\t\]+#)" 2 } } */
> /* { dg-final { scan-assembler-times "vfcmaddcph\[
> \\t\]+%ymm\[0-9\]+\[^\n\r]*%ymm\[0-9\]+\[^\n\r]*%ymm\[0-9\]+\{%k\[0-9\]\}\{z\}\[^\n\r]*(?:\n|\[
> \\t\]+#)" 1 } } */
> -/* { dg-final { scan-assembler-times "vblendmps\[
> \\t\]+%ymm\[0-9\]+\[^\n\r]*%ymm\[0-9\]+\[^\n\r]*%ymm\[0-9\]+\{%k\[0-9\]\}(?:\n|\[
> \\t\]+#)" 1 } } */
> +/* { dg-final { scan-assembler-times "(?:vblendmps\[
> \\t\]+%ymm\[0-9\]+|vmovaps\[
> \\t\]+)\[^\n\r]*%ymm\[0-9\]+\[^\n\r]*%ymm\[0-9\]+\{%k\[0-9\]\}(?:\n|\[
> \\t\]+#)" 1 } } */
> /* { dg-final { scan-assembler-times "vfcmaddcph\[
> \\t\]+%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+(?:\n|\[
> \\t\]+#)" 1 } } */
> /* { dg-final { scan-assembler-times "vfcmaddcph\[
> \\t\]+%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+\{%k\[0-9\]\}(?:\n|\[
> \\t\]+#)" 2 } } */
> /* { dg-final { scan-assembler-times "vfcmaddcph\[
> \\t\]+%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+\{%k\[0-9\]\}\{z\}\[^\n\r]*(?:\n|\[
> \\t\]+#)" 1 } } */
> -/* { dg-final { scan-assembler-times "vblendmps\[
> \\t\]+%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+\{%k\[0-9\]\}(?:\n|\[
> \\t\]+#)" 1 } } */
> +/* { dg-final { scan-assembler-times "(?:vblendmps\[
> \\t\]+%xmm\[0-9\]+|vmovaps\[
> \\t\]+)\[^\n\r]*%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+\{%k\[0-9\]\}(?:\n|\[
> \\t\]+#)" 1 } } */
>
> #include <immintrin.h>
>
> diff --git a/gcc/testsuite/gcc.target/i386/avx512fp16vl-vfmaddcph-1a.c
> b/gcc/testsuite/gcc.target/i386/avx512fp16vl-vfmaddcph-1a.c
> index 75cba83f4ed..7dfbb3e1e19 100644
> --- a/gcc/testsuite/gcc.target/i386/avx512fp16vl-vfmaddcph-1a.c
> +++ b/gcc/testsuite/gcc.target/i386/avx512fp16vl-vfmaddcph-1a.c
> @@ -3,11 +3,11 @@
> /* { dg-final { scan-assembler-times "vfmaddcph\[
> \\t\]+\[^\{\n\]*%ymm\[0-9\]+\[^\n\r]*%ymm\[0-9\]+(?:\n|\[ \\t\]+#)" 1 } } */
> /* { dg-final { scan-assembler-times "vfmaddcph\[
> \\t\]+%ymm\[0-9\]+\[^\n\r]*%ymm\[0-9\]+\[^\n\r]*%ymm\[0-9\]+\{%k\[0-9\]\}(?:\n|\[
> \\t\]+#)" 2 } } */
> /* { dg-final { scan-assembler-times "vfmaddcph\[
> \\t\]+%ymm\[0-9\]+\[^\n\r]*%ymm\[0-9\]+\[^\n\r]*%ymm\[0-9\]+\{%k\[0-9\]\}\{z\}\[^\n\r]*(?:\n|\[
> \\t\]+#)" 1 } } */
> -/* { dg-final { scan-assembler-times "vblendmps\[
> \\t\]+%ymm\[0-9\]+\[^\n\r]*%ymm\[0-9\]+\[^\n\r]*%ymm\[0-9\]+\{%k\[0-9\]\}(?:\n|\[
> \\t\]+#)" 1 } } */
> +/* { dg-final { scan-assembler-times "(?:vblendmps\[
> \\t\]+%ymm\[0-9\]+|vmovaps\[
> \\t\]+)\[^\n\r]*%ymm\[0-9\]+\[^\n\r]*%ymm\[0-9\]+\{%k\[0-9\]\}(?:\n|\[
> \\t\]+#)" 1 } } */
> /* { dg-final { scan-assembler-times "vfmaddcph\[
> \\t\]+%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+(?:\n|\[
> \\t\]+#)" 1 } } */
> /* { dg-final { scan-assembler-times "vfmaddcph\[
> \\t\]+%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+\{%k\[0-9\]\}(?:\n|\[
> \\t\]+#)" 2 } } */
> /* { dg-final { scan-assembler-times "vfmaddcph\[
> \\t\]+%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+\{%k\[0-9\]\}\{z\}\[^\n\r]*(?:\n|\[
> \\t\]+#)" 1 } } */
> -/* { dg-final { scan-assembler-times "vblendmps\[
> \\t\]+%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+\{%k\[0-9\]\}(?:\n|\[
> \\t\]+#)" 1 } } */
> +/* { dg-final { scan-assembler-times "(?:vblendmps\[
> \\t\]+%xmm\[0-9\]+|vmovaps\[
> \\t\]+)\[^\n\r]*%xmm\[0-9\]+\[^\n\r]*%xmm\[0-9\]+\{%k\[0-9\]\}(?:\n|\[
> \\t\]+#)" 1 } } */
>
> #include <immintrin.h>
>
> diff --git a/gcc/testsuite/gcc.target/i386/pr104976.c
> b/gcc/testsuite/gcc.target/i386/pr104976.c
> new file mode 100644
> index 00000000000..af8e75abd34
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/i386/pr104976.c
> @@ -0,0 +1,13 @@
> +/* { dg-do compile } */
> +/* { dg-options "-mavx512fp16 -mavx512vl -O2" } */
> +
> +void
> +foo (int a, float b, double* c)
> +{
> + union {
> + _Float16 f[2];
> + int i;} x;
> + x.i = a;
> + c[0] = x.f[0] * 0.2;
> + c[1] = x.f[1] * 0.2;
> +}
> --
> 2.18.1
>
--
BR,
Hongtao
