From: Richard Biener <rguenther@suse.de>
To: Juzhe-Zhong <juzhe.zhong@rivai.ai>
Cc: gcc-patches@gcc.gnu.org, richard.sandiford@arm.com
Subject: Re: [PATCH] cleanup: make all cond_len_* and mask_len_* consistent on the order of mask and len
Date: Fri, 21 Jul 2023 06:20:19 +0000 (UTC) [thread overview]
Message-ID: <nycvar.YFH.7.77.849.2307210620081.12935@jbgna.fhfr.qr> (raw)
In-Reply-To: <20230721020900.298554-1-juzhe.zhong@rivai.ai>
On Fri, 21 Jul 2023, Juzhe-Zhong wrote:
> This patch is depending on:
> https://gcc.gnu.org/pipermail/gcc-patches/2023-July/625121.html
>
> Hi, Richard and Richi.
>
> This patch is to align the order of mask and len.
>
> Currently, According to this piece code:
> if (final_len && final_mask)
> call = gimple_build_call_internal (
> IFN_LEN_MASK_GATHER_LOAD, 7, dataref_ptr,
> vec_offset, scale, zero, final_mask, final_len,
> bias);
>
> You can see the order of mask and len, is {mask,len,bias}.
> "mask" comes before "len". The reason of this order is that we want to
> reuse the current codes of MASK_GATHER_LOAD/MASK_SCATTER_STORE.
>
> Same situation for COND_LEN_*, we want to reuse the codes of COND_*.
>
> Reusing codes from the existing MASK_* or COND_* can allow us not to
> change the codes too much and make the codes elegant and easy to maintain && read.
>
> To avoid any confusions of auto-vectorization patterns that includes both mask and len,
>
> this patch align the order of mask and len for both Gimple IR and RTL pattern into
>
> {mask, len, bias} to make everything cleaner and more elegant.
>
> Bootstrap and Regression is on the way.
OK.
> gcc/ChangeLog:
>
> * config/riscv/autovec.md: Align order of mask and len.
> * config/riscv/riscv-v.cc (expand_load_store): Ditto.
> (expand_gather_scatter): Ditto.
> * doc/md.texi: Ditto.
> * internal-fn.cc (add_len_and_mask_args): Ditto.
> (add_mask_and_len_args): Ditto.
> (expand_partial_load_optab_fn): Ditto.
> (expand_partial_store_optab_fn): Ditto.
> (expand_scatter_store_optab_fn): Ditto.
> (expand_gather_load_optab_fn): Ditto.
> (internal_fn_len_index): Ditto.
> (internal_fn_mask_index): Ditto.
> (internal_len_load_store_bias): Ditto.
> * tree-vect-stmts.cc (vectorizable_store): Ditto.
> (vectorizable_load): Ditto.
>
> ---
> gcc/config/riscv/autovec.md | 96 ++++++++++++++++++-------------------
> gcc/config/riscv/riscv-v.cc | 12 ++---
> gcc/doc/md.texi | 36 +++++++-------
> gcc/internal-fn.cc | 50 +++++++++----------
> gcc/tree-vect-stmts.cc | 8 ++--
> 5 files changed, 101 insertions(+), 101 deletions(-)
>
> diff --git a/gcc/config/riscv/autovec.md b/gcc/config/riscv/autovec.md
> index 7eb96d42c18..d899922586a 100644
> --- a/gcc/config/riscv/autovec.md
> +++ b/gcc/config/riscv/autovec.md
> @@ -25,9 +25,9 @@
> (define_expand "mask_len_load<mode><vm>"
> [(match_operand:V 0 "register_operand")
> (match_operand:V 1 "memory_operand")
> - (match_operand 2 "autovec_length_operand")
> - (match_operand 3 "const_0_operand")
> - (match_operand:<VM> 4 "vector_mask_operand")]
> + (match_operand:<VM> 2 "vector_mask_operand")
> + (match_operand 3 "autovec_length_operand")
> + (match_operand 4 "const_0_operand")]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_load_store (operands, true);
> @@ -37,9 +37,9 @@
> (define_expand "mask_len_store<mode><vm>"
> [(match_operand:V 0 "memory_operand")
> (match_operand:V 1 "register_operand")
> - (match_operand 2 "autovec_length_operand")
> - (match_operand 3 "const_0_operand")
> - (match_operand:<VM> 4 "vector_mask_operand")]
> + (match_operand:<VM> 2 "vector_mask_operand")
> + (match_operand 3 "autovec_length_operand")
> + (match_operand 4 "const_0_operand")]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_load_store (operands, false);
> @@ -67,9 +67,9 @@
> (match_operand:RATIO64I 2 "register_operand")
> (match_operand 3 "<RATIO64:gs_extension>")
> (match_operand 4 "<RATIO64:gs_scale>")
> - (match_operand 5 "autovec_length_operand")
> - (match_operand 6 "const_0_operand")
> - (match_operand:<RATIO64:VM> 7 "vector_mask_operand")]
> + (match_operand:<RATIO64:VM> 5 "vector_mask_operand")
> + (match_operand 6 "autovec_length_operand")
> + (match_operand 7 "const_0_operand")]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_gather_scatter (operands, true);
> @@ -82,9 +82,9 @@
> (match_operand:RATIO32I 2 "register_operand")
> (match_operand 3 "<RATIO32:gs_extension>")
> (match_operand 4 "<RATIO32:gs_scale>")
> - (match_operand 5 "autovec_length_operand")
> - (match_operand 6 "const_0_operand")
> - (match_operand:<RATIO32:VM> 7 "vector_mask_operand")]
> + (match_operand:<RATIO32:VM> 5 "vector_mask_operand")
> + (match_operand 6 "autovec_length_operand")
> + (match_operand 7 "const_0_operand")]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_gather_scatter (operands, true);
> @@ -97,9 +97,9 @@
> (match_operand:RATIO16I 2 "register_operand")
> (match_operand 3 "<RATIO16:gs_extension>")
> (match_operand 4 "<RATIO16:gs_scale>")
> - (match_operand 5 "autovec_length_operand")
> - (match_operand 6 "const_0_operand")
> - (match_operand:<RATIO16:VM> 7 "vector_mask_operand")]
> + (match_operand:<RATIO16:VM> 5 "vector_mask_operand")
> + (match_operand 6 "autovec_length_operand")
> + (match_operand 7 "const_0_operand")]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_gather_scatter (operands, true);
> @@ -112,9 +112,9 @@
> (match_operand:RATIO8I 2 "register_operand")
> (match_operand 3 "<RATIO8:gs_extension>")
> (match_operand 4 "<RATIO8:gs_scale>")
> - (match_operand 5 "autovec_length_operand")
> - (match_operand 6 "const_0_operand")
> - (match_operand:<RATIO8:VM> 7 "vector_mask_operand")]
> + (match_operand:<RATIO8:VM> 5 "vector_mask_operand")
> + (match_operand 6 "autovec_length_operand")
> + (match_operand 7 "const_0_operand")]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_gather_scatter (operands, true);
> @@ -127,9 +127,9 @@
> (match_operand:RATIO4I 2 "register_operand")
> (match_operand 3 "<RATIO4:gs_extension>")
> (match_operand 4 "<RATIO4:gs_scale>")
> - (match_operand 5 "autovec_length_operand")
> - (match_operand 6 "const_0_operand")
> - (match_operand:<RATIO4:VM> 7 "vector_mask_operand")]
> + (match_operand:<RATIO4:VM> 5 "vector_mask_operand")
> + (match_operand 6 "autovec_length_operand")
> + (match_operand 7 "const_0_operand")]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_gather_scatter (operands, true);
> @@ -142,9 +142,9 @@
> (match_operand:RATIO2I 2 "register_operand")
> (match_operand 3 "<RATIO2:gs_extension>")
> (match_operand 4 "<RATIO2:gs_scale>")
> - (match_operand 5 "autovec_length_operand")
> - (match_operand 6 "const_0_operand")
> - (match_operand:<RATIO2:VM> 7 "vector_mask_operand")]
> + (match_operand:<RATIO2:VM> 5 "vector_mask_operand")
> + (match_operand 6 "autovec_length_operand")
> + (match_operand 7 "const_0_operand")]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_gather_scatter (operands, true);
> @@ -161,9 +161,9 @@
> (match_operand:RATIO1 2 "register_operand")
> (match_operand 3 "<RATIO1:gs_extension>")
> (match_operand 4 "<RATIO1:gs_scale>")
> - (match_operand 5 "autovec_length_operand")
> - (match_operand 6 "const_0_operand")
> - (match_operand:<RATIO1:VM> 7 "vector_mask_operand")]
> + (match_operand:<RATIO1:VM> 5 "vector_mask_operand")
> + (match_operand 6 "autovec_length_operand")
> + (match_operand 7 "const_0_operand")]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_gather_scatter (operands, true);
> @@ -180,9 +180,9 @@
> (match_operand 2 "<RATIO64:gs_extension>")
> (match_operand 3 "<RATIO64:gs_scale>")
> (match_operand:RATIO64 4 "register_operand")
> - (match_operand 5 "autovec_length_operand")
> - (match_operand 6 "const_0_operand")
> - (match_operand:<RATIO64:VM> 7 "vector_mask_operand")]
> + (match_operand:<RATIO64:VM> 5 "vector_mask_operand")
> + (match_operand 6 "autovec_length_operand")
> + (match_operand 7 "const_0_operand")]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_gather_scatter (operands, false);
> @@ -195,9 +195,9 @@
> (match_operand 2 "<RATIO32:gs_extension>")
> (match_operand 3 "<RATIO32:gs_scale>")
> (match_operand:RATIO32 4 "register_operand")
> - (match_operand 5 "autovec_length_operand")
> - (match_operand 6 "const_0_operand")
> - (match_operand:<RATIO32:VM> 7 "vector_mask_operand")]
> + (match_operand:<RATIO32:VM> 5 "vector_mask_operand")
> + (match_operand 6 "autovec_length_operand")
> + (match_operand 7 "const_0_operand")]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_gather_scatter (operands, false);
> @@ -210,9 +210,9 @@
> (match_operand 2 "<RATIO16:gs_extension>")
> (match_operand 3 "<RATIO16:gs_scale>")
> (match_operand:RATIO16 4 "register_operand")
> - (match_operand 5 "autovec_length_operand")
> - (match_operand 6 "const_0_operand")
> - (match_operand:<RATIO16:VM> 7 "vector_mask_operand")]
> + (match_operand:<RATIO16:VM> 5 "vector_mask_operand")
> + (match_operand 6 "autovec_length_operand")
> + (match_operand 7 "const_0_operand")]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_gather_scatter (operands, false);
> @@ -225,9 +225,9 @@
> (match_operand 2 "<RATIO8:gs_extension>")
> (match_operand 3 "<RATIO8:gs_scale>")
> (match_operand:RATIO8 4 "register_operand")
> - (match_operand 5 "autovec_length_operand")
> - (match_operand 6 "const_0_operand")
> - (match_operand:<RATIO8:VM> 7 "vector_mask_operand")]
> + (match_operand:<RATIO8:VM> 5 "vector_mask_operand")
> + (match_operand 6 "autovec_length_operand")
> + (match_operand 7 "const_0_operand")]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_gather_scatter (operands, false);
> @@ -240,9 +240,9 @@
> (match_operand 2 "<RATIO4:gs_extension>")
> (match_operand 3 "<RATIO4:gs_scale>")
> (match_operand:RATIO4 4 "register_operand")
> - (match_operand 5 "autovec_length_operand")
> - (match_operand 6 "const_0_operand")
> - (match_operand:<RATIO4:VM> 7 "vector_mask_operand")]
> + (match_operand:<RATIO4:VM> 5 "vector_mask_operand")
> + (match_operand 6 "autovec_length_operand")
> + (match_operand 7 "const_0_operand")]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_gather_scatter (operands, false);
> @@ -255,9 +255,9 @@
> (match_operand 2 "<RATIO2:gs_extension>")
> (match_operand 3 "<RATIO2:gs_scale>")
> (match_operand:RATIO2 4 "register_operand")
> - (match_operand 5 "autovec_length_operand")
> - (match_operand 6 "const_0_operand")
> - (match_operand:<RATIO2:VM> 7 "vector_mask_operand")]
> + (match_operand:<RATIO2:VM> 5 "vector_mask_operand")
> + (match_operand 6 "autovec_length_operand")
> + (match_operand 7 "const_0_operand")]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_gather_scatter (operands, false);
> @@ -274,9 +274,9 @@
> (match_operand 2 "<RATIO1:gs_extension>")
> (match_operand 3 "<RATIO1:gs_scale>")
> (match_operand:RATIO1 4 "register_operand")
> - (match_operand 5 "autovec_length_operand")
> - (match_operand 6 "const_0_operand")
> - (match_operand:<RATIO1:VM> 7 "vector_mask_operand")]
> + (match_operand:<RATIO1:VM> 5 "vector_mask_operand")
> + (match_operand 6 "autovec_length_operand")
> + (match_operand 7 "const_0_operand")]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_gather_scatter (operands, false);
> diff --git a/gcc/config/riscv/riscv-v.cc b/gcc/config/riscv/riscv-v.cc
> index e338be151d3..0ae87936c38 100644
> --- a/gcc/config/riscv/riscv-v.cc
> +++ b/gcc/config/riscv/riscv-v.cc
> @@ -3093,13 +3093,13 @@ expand_select_vl (rtx *ops)
> emit_insn (gen_no_side_effects_vsetvl_rtx (rvv_mode, ops[0], ops[1]));
> }
>
> -/* Expand LEN_MASK_{LOAD,STORE}. */
> +/* Expand MASK_LEN_{LOAD,STORE}. */
> void
> expand_load_store (rtx *ops, bool is_load)
> {
> poly_int64 value;
> - rtx len = ops[2];
> - rtx mask = ops[4];
> + rtx mask = ops[2];
> + rtx len = ops[3];
> machine_mode mode = GET_MODE (ops[0]);
>
> if (poly_int_rtx_p (len, &value) && known_eq (value, GET_MODE_NUNITS (mode)))
> @@ -3235,6 +3235,8 @@ expand_gather_scatter (rtx *ops, bool is_load)
> rtx ptr, vec_offset, vec_reg, len, mask;
> bool zero_extend_p;
> int scale_log2;
> + rtx mask = ops[5];
> + rtx len = ops[6];
> if (is_load)
> {
> vec_reg = ops[0];
> @@ -3242,8 +3244,6 @@ expand_gather_scatter (rtx *ops, bool is_load)
> vec_offset = ops[2];
> zero_extend_p = INTVAL (ops[3]);
> scale_log2 = exact_log2 (INTVAL (ops[4]));
> - len = ops[5];
> - mask = ops[7];
> }
> else
> {
> @@ -3252,8 +3252,6 @@ expand_gather_scatter (rtx *ops, bool is_load)
> vec_offset = ops[1];
> zero_extend_p = INTVAL (ops[2]);
> scale_log2 = exact_log2 (INTVAL (ops[3]));
> - len = ops[5];
> - mask = ops[7];
> }
>
> machine_mode vec_mode = GET_MODE (vec_reg);
> diff --git a/gcc/doc/md.texi b/gcc/doc/md.texi
> index f864936d54a..9693b6bfe79 100644
> --- a/gcc/doc/md.texi
> +++ b/gcc/doc/md.texi
> @@ -5042,12 +5042,12 @@ of the result should be set to zero.
>
> @cindex @code{mask_len_gather_load@var{m}@var{n}} instruction pattern
> @item @samp{mask_len_gather_load@var{m}@var{n}}
> -Like @samp{gather_load@var{m}@var{n}}, but takes an extra length operand (operand 5),
> -a bias operand (operand 6) as well as a mask operand (operand 7). Similar to mask_len_load,
> -the instruction loads at most (operand 5 + operand 6) elements from memory.
> +Like @samp{gather_load@var{m}@var{n}}, but takes an extra mask operand (operand 5),
> +a len operand (operand 6) as well as a bias operand (operand 7). Similar to mask_len_load,
> +the instruction loads at most (operand 6 + operand 7) elements from memory.
> Bit @var{i} of the mask is set if element @var{i} of the result should
> be loaded from memory and clear if element @var{i} of the result should be undefined.
> -Mask elements @var{i} with @var{i} > (operand 5 + operand 6) are ignored.
> +Mask elements @var{i} with @var{i} > (operand 6 + operand 7) are ignored.
>
> @cindex @code{scatter_store@var{m}@var{n}} instruction pattern
> @item @samp{scatter_store@var{m}@var{n}}
> @@ -5080,11 +5080,11 @@ of the result should be stored to memory.
>
> @cindex @code{mask_len_scatter_store@var{m}@var{n}} instruction pattern
> @item @samp{mask_len_scatter_store@var{m}@var{n}}
> -Like @samp{scatter_store@var{m}@var{n}}, but takes an extra length operand (operand 5),
> -a bias operand (operand 6) as well as a mask operand (operand 7). The instruction stores
> -at most (operand 5 + operand 6) elements of (operand 4) to memory.
> +Like @samp{scatter_store@var{m}@var{n}}, but takes an extra mask operand (operand 5),
> +a len operand (operand 6) as well as a bias operand (operand 7). The instruction stores
> +at most (operand 6 + operand 7) elements of (operand 4) to memory.
> Bit @var{i} of the mask is set if element @var{i} of (operand 4) should be stored.
> -Mask elements @var{i} with @var{i} > (operand 5 + operand 6) are ignored.
> +Mask elements @var{i} with @var{i} > (operand 6 + operand 7) are ignored.
>
> @cindex @code{vec_set@var{m}} instruction pattern
> @item @samp{vec_set@var{m}}
> @@ -5324,15 +5324,15 @@ This pattern is not allowed to @code{FAIL}.
> @cindex @code{mask_len_load@var{m}@var{n}} instruction pattern
> @item @samp{mask_len_load@var{m}@var{n}}
> Perform a masked load from the memory location pointed to by operand 1
> -into register operand 0. (operand 2 + operand 3) elements are loaded from
> +into register operand 0. (operand 3 + operand 4) elements are loaded from
> memory and other elements in operand 0 are set to undefined values.
> This is a combination of len_load and maskload.
> -Operands 0 and 1 have mode @var{m}, which must be a vector mode. Operand 2
> +Operands 0 and 1 have mode @var{m}, which must be a vector mode. Operand 3
> has whichever integer mode the target prefers. A mask is specified in
> -operand 4 which must be of type @var{n}. The mask has lower precedence than
> +operand 2 which must be of type @var{n}. The mask has lower precedence than
> the length and is itself subject to length masking,
> -i.e. only mask indices < (operand 2 + operand 3) are used.
> -Operand 3 conceptually has mode @code{QI}.
> +i.e. only mask indices < (operand 3 + operand 4) are used.
> +Operand 4 conceptually has mode @code{QI}.
>
> Operand 2 can be a variable or a constant amount. Operand 4 specifies a
> constant bias: it is either a constant 0 or a constant -1. The predicate on
> @@ -5351,14 +5351,14 @@ This pattern is not allowed to @code{FAIL}.
> @cindex @code{mask_len_store@var{m}@var{n}} instruction pattern
> @item @samp{mask_len_store@var{m}@var{n}}
> Perform a masked store from vector register operand 1 into memory operand 0.
> -(operand 2 + operand 3) elements are stored to memory
> +(operand 3 + operand 4) elements are stored to memory
> and leave the other elements of operand 0 unchanged.
> This is a combination of len_store and maskstore.
> -Operands 0 and 1 have mode @var{m}, which must be a vector mode. Operand 2 has whichever
> -integer mode the target prefers. A mask is specified in operand 4 which must be
> +Operands 0 and 1 have mode @var{m}, which must be a vector mode. Operand 3 has whichever
> +integer mode the target prefers. A mask is specified in operand 2 which must be
> of type @var{n}. The mask has lower precedence than the length and is itself subject to
> -length masking, i.e. only mask indices < (operand 2 + operand 3) are used.
> -Operand 3 conceptually has mode @code{QI}.
> +length masking, i.e. only mask indices < (operand 3 + operand 4) are used.
> +Operand 4 conceptually has mode @code{QI}.
>
> Operand 2 can be a variable or a constant amount. Operand 3 specifies a
> constant bias: it is either a constant 0 or a constant -1. The predicate on
> diff --git a/gcc/internal-fn.cc b/gcc/internal-fn.cc
> index 969bfe9ab67..0c153a4ee0d 100644
> --- a/gcc/internal-fn.cc
> +++ b/gcc/internal-fn.cc
> @@ -298,10 +298,10 @@ get_multi_vector_move (tree array_type, convert_optab optab)
> return convert_optab_handler (optab, imode, vmode);
> }
>
> -/* Add len and mask arguments according to the STMT. */
> +/* Add mask and len arguments according to the STMT. */
>
> static unsigned int
> -add_len_and_mask_args (expand_operand *ops, unsigned int opno, gcall *stmt)
> +add_mask_and_len_args (expand_operand *ops, unsigned int opno, gcall *stmt)
> {
> internal_fn ifn = gimple_call_internal_fn (stmt);
> int len_index = internal_fn_len_index (ifn);
> @@ -309,6 +309,13 @@ add_len_and_mask_args (expand_operand *ops, unsigned int opno, gcall *stmt)
> int bias_index = len_index + 1;
> int mask_index = internal_fn_mask_index (ifn);
> /* The order of arguments are always {len,bias,mask}. */
> + if (mask_index >= 0)
> + {
> + tree mask = gimple_call_arg (stmt, mask_index);
> + rtx mask_rtx = expand_normal (mask);
> + create_input_operand (&ops[opno++], mask_rtx,
> + TYPE_MODE (TREE_TYPE (mask)));
> + }
> if (len_index >= 0)
> {
> tree len = gimple_call_arg (stmt, len_index);
> @@ -320,13 +327,6 @@ add_len_and_mask_args (expand_operand *ops, unsigned int opno, gcall *stmt)
> rtx bias = expand_normal (biast);
> create_input_operand (&ops[opno++], bias, QImode);
> }
> - if (mask_index >= 0)
> - {
> - tree mask = gimple_call_arg (stmt, mask_index);
> - rtx mask_rtx = expand_normal (mask);
> - create_input_operand (&ops[opno++], mask_rtx,
> - TYPE_MODE (TREE_TYPE (mask)));
> - }
> return opno;
> }
>
> @@ -2944,7 +2944,7 @@ expand_partial_load_optab_fn (internal_fn ifn, gcall *stmt, convert_optab optab)
> target = expand_expr (lhs, NULL_RTX, VOIDmode, EXPAND_WRITE);
> create_output_operand (&ops[i++], target, TYPE_MODE (type));
> create_fixed_operand (&ops[i++], mem);
> - i = add_len_and_mask_args (ops, i, stmt);
> + i = add_mask_and_len_args (ops, i, stmt);
> expand_insn (icode, i, ops);
>
> if (!rtx_equal_p (target, ops[0].value))
> @@ -2986,7 +2986,7 @@ expand_partial_store_optab_fn (internal_fn ifn, gcall *stmt, convert_optab optab
> reg = expand_normal (rhs);
> create_fixed_operand (&ops[i++], mem);
> create_input_operand (&ops[i++], reg, TYPE_MODE (type));
> - i = add_len_and_mask_args (ops, i, stmt);
> + i = add_mask_and_len_args (ops, i, stmt);
> expand_insn (icode, i, ops);
> }
>
> @@ -3566,7 +3566,7 @@ expand_scatter_store_optab_fn (internal_fn, gcall *stmt, direct_optab optab)
> create_integer_operand (&ops[i++], TYPE_UNSIGNED (TREE_TYPE (offset)));
> create_integer_operand (&ops[i++], scale_int);
> create_input_operand (&ops[i++], rhs_rtx, TYPE_MODE (TREE_TYPE (rhs)));
> - i = add_len_and_mask_args (ops, i, stmt);
> + i = add_mask_and_len_args (ops, i, stmt);
>
> insn_code icode = convert_optab_handler (optab, TYPE_MODE (TREE_TYPE (rhs)),
> TYPE_MODE (TREE_TYPE (offset)));
> @@ -3595,7 +3595,7 @@ expand_gather_load_optab_fn (internal_fn, gcall *stmt, direct_optab optab)
> create_input_operand (&ops[i++], offset_rtx, TYPE_MODE (TREE_TYPE (offset)));
> create_integer_operand (&ops[i++], TYPE_UNSIGNED (TREE_TYPE (offset)));
> create_integer_operand (&ops[i++], scale_int);
> - i = add_len_and_mask_args (ops, i, stmt);
> + i = add_mask_and_len_args (ops, i, stmt);
> insn_code icode = convert_optab_handler (optab, TYPE_MODE (TREE_TYPE (lhs)),
> TYPE_MODE (TREE_TYPE (offset)));
> expand_insn (icode, i, ops);
> @@ -4603,8 +4603,6 @@ internal_fn_len_index (internal_fn fn)
> {
> case IFN_LEN_LOAD:
> case IFN_LEN_STORE:
> - case IFN_MASK_LEN_LOAD:
> - case IFN_MASK_LEN_STORE:
> return 2;
>
> case IFN_MASK_LEN_GATHER_LOAD:
> @@ -4633,6 +4631,8 @@ internal_fn_len_index (internal_fn fn)
> return 4;
>
> case IFN_COND_LEN_NEG:
> + case IFN_MASK_LEN_LOAD:
> + case IFN_MASK_LEN_STORE:
> return 3;
>
> default:
> @@ -4652,12 +4652,12 @@ internal_fn_mask_index (internal_fn fn)
> case IFN_MASK_LOAD_LANES:
> case IFN_MASK_STORE:
> case IFN_MASK_STORE_LANES:
> + case IFN_MASK_LEN_LOAD:
> + case IFN_MASK_LEN_STORE:
> return 2;
>
> case IFN_MASK_GATHER_LOAD:
> case IFN_MASK_SCATTER_STORE:
> - case IFN_MASK_LEN_LOAD:
> - case IFN_MASK_LEN_STORE:
> case IFN_MASK_LEN_GATHER_LOAD:
> case IFN_MASK_LEN_SCATTER_STORE:
> return 4;
> @@ -4742,17 +4742,18 @@ internal_check_ptrs_fn_supported_p (internal_fn ifn, tree type,
> && insn_operand_matches (icode, 4, GEN_INT (align)));
> }
>
> -/* Return the supported bias for IFN which is either IFN_LEN_LOAD
> - or IFN_LEN_STORE. For now we only support the biases of 0 and -1
> - (in case 0 is not an allowable length for len_load or len_store).
> - If none of the biases match what the backend provides, return
> - VECT_PARTIAL_BIAS_UNSUPPORTED. */
> +/* Return the supported bias for IFN which is either IFN_{LEN_,MASK_LEN_,}LOAD
> + or IFN_{LEN_,MASK_LEN_,}STORE. For now we only support the biases of 0 and
> + -1 (in case 0 is not an allowable length for {len_,mask_len_}load or
> + {len_,mask_len_}store). If none of the biases match what the backend
> + provides, return VECT_PARTIAL_BIAS_UNSUPPORTED. */
>
> signed char
> internal_len_load_store_bias (internal_fn ifn, machine_mode mode)
> {
> optab optab = direct_internal_fn_optab (ifn);
> insn_code icode = direct_optab_handler (optab, mode);
> + int bias_no = 3;
>
> if (icode == CODE_FOR_nothing)
> {
> @@ -4770,14 +4771,15 @@ internal_len_load_store_bias (internal_fn ifn, machine_mode mode)
> optab = direct_internal_fn_optab (IFN_MASK_LEN_STORE);
> }
> icode = convert_optab_handler (optab, mode, mask_mode);
> + bias_no = 4;
> }
>
> if (icode != CODE_FOR_nothing)
> {
> /* For now we only support biases of 0 or -1. Try both of them. */
> - if (insn_operand_matches (icode, 3, GEN_INT (0)))
> + if (insn_operand_matches (icode, bias_no, GEN_INT (0)))
> return 0;
> - if (insn_operand_matches (icode, 3, GEN_INT (-1)))
> + if (insn_operand_matches (icode, bias_no, GEN_INT (-1)))
> return -1;
> }
>
> diff --git a/gcc/tree-vect-stmts.cc b/gcc/tree-vect-stmts.cc
> index caa4c974cb9..d5b4f020332 100644
> --- a/gcc/tree-vect-stmts.cc
> +++ b/gcc/tree-vect-stmts.cc
> @@ -9087,8 +9087,8 @@ vectorizable_store (vec_info *vinfo,
> if (partial_ifn == IFN_MASK_LEN_STORE)
> call = gimple_build_call_internal (IFN_MASK_LEN_STORE, 6,
> dataref_ptr, ptr,
> - final_len, bias,
> - final_mask, vec_oprnd);
> + final_mask, final_len,
> + bias, vec_oprnd);
> else
> call
> = gimple_build_call_internal (IFN_LEN_STORE, 5,
> @@ -10654,8 +10654,8 @@ vectorizable_load (vec_info *vinfo,
> if (partial_ifn == IFN_MASK_LEN_LOAD)
> call = gimple_build_call_internal (IFN_MASK_LEN_LOAD,
> 5, dataref_ptr,
> - ptr, final_len,
> - bias, final_mask);
> + ptr, final_mask,
> + final_len, bias);
> else
> call = gimple_build_call_internal (IFN_LEN_LOAD, 4,
> dataref_ptr, ptr,
>
--
Richard Biener <rguenther@suse.de>
SUSE Software Solutions Germany GmbH, Frankenstrasse 146, 90461 Nuernberg,
Germany; GF: Ivo Totev, Andrew Myers, Andrew McDonald, Boudien Moerman;
HRB 36809 (AG Nuernberg)
next prev parent reply other threads:[~2023-07-21 6:20 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-07-21 2:09 Juzhe-Zhong
2023-07-21 6:20 ` Richard Biener [this message]
2023-07-21 8:38 ` =?gb18030?B?TGVodWEgRGluZw==?=
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=nycvar.YFH.7.77.849.2307210620081.12935@jbgna.fhfr.qr \
--to=rguenther@suse.de \
--cc=gcc-patches@gcc.gnu.org \
--cc=juzhe.zhong@rivai.ai \
--cc=richard.sandiford@arm.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).