mirror/gcc
1
0
Fork 0
mirror of https://github.com/gcc-mirror/gcc.git synced 2024-11-21 13:40:47 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
master
gcc/gcc/tree-ssa-forwprop.cc
Christoph Müllner 1c4d39ada3 forwprop: Try to blend two isomorphic VEC_PERM sequences 
This extends forwprop by yet another VEC_PERM optimization:
It attempts to blend two isomorphic vector sequences by using the
redundancy in the lane utilization in these sequences.
This redundancy in lane utilization comes from the way how specific
scalar statements end up vectorized: two VEC_PERMs on top, binary operations
on both of them, and a final VEC_PERM to create the result.
Here is an example of this sequence:

  v_in = {e0, e1, e2, e3}
  v_1 = VEC_PERM <v_in, v_in, {0, 2, 0, 2}>
  // v_1 = {e0, e2, e0, e2}
  v_2 = VEC_PERM <v_in, v_in, {1, 3, 1, 3}>
  // v_2 = {e1, e3, e1, e3}

  v_x = v_1 + v_2
  // v_x = {e0+e1, e2+e3, e0+e1, e2+e3}
  v_y = v_1 - v_2
  // v_y = {e0-e1, e2-e3, e0-e1, e2-e3}

  v_out = VEC_PERM <v_x, v_y, {0, 1, 6, 7}>
  // v_out = {e0+e1, e2+e3, e0-e1, e2-e3}

To remove the redundancy, lanes 2 and 3 can be freed, which allows to
change the last statement into:
  v_out' = VEC_PERM <v_x, v_y, {0, 1, 4, 5}>
  // v_out' = {e0+e1, e2+e3, e0-e1, e2-e3}

The cost of eliminating the redundancy in the lane utilization is that
lowering the VEC PERM expression could get more expensive because of
tighter packing of the lanes.  Therefore this optimization is not done
alone, but in only in case we identify two such sequences that can be
blended.

Once all candidate sequences have been identified, we try to blend them,
so that we can use the freed lanes for the second sequence.
On success we convert 2x (2x BINOP + 1x VEC_PERM) to
2x VEC_PERM + 2x BINOP + 2x VEC_PERM traded for 4x VEC_PERM + 2x BINOP.

The implemented transformation reuses (rewrites) the statements
of the first sequence and the last VEC_PERM of the second sequence.
The remaining four statements of the second statment are left untouched
and will be eliminated by DCE later.

This targets x264_pixel_satd_8x4, which calculates the sum of absolute
transformed differences (SATD) using Hadamard transformation.
We have seen 8% speedup on SPEC's x264 on a 5950X (x86-64) and 7%
speedup on an AArch64 machine.

Bootstrapped and reg-tested on x86-64 and AArch64 (all languages).

gcc/ChangeLog:

	* tree-ssa-forwprop.cc (struct _vec_perm_simplify_seq): New data
	structure to store analysis results of a vec perm simplify sequence.
	(get_vect_selector_index_map): Helper to get an index map from the
	provided vector permute selector.
	(recognise_vec_perm_simplify_seq): Helper to recognise a
	vec perm simplify sequence.
	(narrow_vec_perm_simplify_seq): Helper to pack the lanes more
	tight.
	(can_blend_vec_perm_simplify_seqs_p): Test if two vec perm
	sequences can be blended.
	(calc_perm_vec_perm_simplify_seqs): Helper to calculate the new
	permutation indices.
	(blend_vec_perm_simplify_seqs): Helper to blend two vec perm
	simplify sequences.
	(process_vec_perm_simplify_seq_list): Helper to process a list
	of vec perm simplify sequences.
	(append_vec_perm_simplify_seq_list): Helper to add a vec perm
	simplify sequence to the list.
	(pass_forwprop::execute): Integrate new functionality.

gcc/testsuite/ChangeLog:

	* gcc.dg/tree-ssa/satd-hadamard.c: New test.
	* gcc.dg/tree-ssa/vector-10.c: New test.
	* gcc.dg/tree-ssa/vector-8.c: New test.
	* gcc.dg/tree-ssa/vector-9.c: New test.
	* gcc.target/aarch64/sve/satd-hadamard.c: New test.

Signed-off-by: Christoph Müllner <christoph.muellner@vrull.eu>
2024-11-21 13:38:54 +01:00

4810 lines
150 KiB
C++
Raw Permalink Blame History

/* Forward propagation of expressions for single use variables.
Copyright (C) 2004-2024 Free Software Foundation, Inc.
This file is part of GCC.
GCC is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 3, or (at your option)
any later version.
GCC 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 General Public License for more details.
You should have received a copy of the GNU General Public License
along with GCC; see the file COPYING3. If not see
<http://www.gnu.org/licenses/>. */
#define INCLUDE_MEMORY
#include "config.h"
#include "system.h"
#include "coretypes.h"
#include "backend.h"
#include "rtl.h"
#include "tree.h"
#include "gimple.h"
#include "cfghooks.h"
#include "tree-pass.h"
#include "ssa.h"
#include "expmed.h"
#include "optabs-query.h"
#include "gimple-pretty-print.h"
#include "fold-const.h"
#include "stor-layout.h"
#include "gimple-iterator.h"
#include "gimple-fold.h"
#include "tree-eh.h"
#include "gimplify.h"
#include "gimplify-me.h"
#include "tree-cfg.h"
#include "expr.h"
#include "tree-dfa.h"
#include "tree-ssa-propagate.h"
#include "tree-ssa-dom.h"
#include "tree-ssa-strlen.h"
#include "builtins.h"
#include "tree-cfgcleanup.h"
#include "cfganal.h"
#include "optabs-tree.h"
#include "insn-config.h"
#include "recog.h"
#include "cfgloop.h"
#include "tree-vectorizer.h"
#include "tree-vector-builder.h"
#include "vec-perm-indices.h"
#include "internal-fn.h"
#include "cgraph.h"
#include "tree-ssa.h"
#include "gimple-range.h"
#include "tree-ssa-dce.h"
/* This pass propagates the RHS of assignment statements into use
sites of the LHS of the assignment. It's basically a specialized
form of tree combination. It is hoped all of this can disappear
when we have a generalized tree combiner.
One class of common cases we handle is forward propagating a single use
variable into a COND_EXPR.
bb0:
x = a COND b;
if (x) goto ... else goto ...
Will be transformed into:
bb0:
if (a COND b) goto ... else goto ...
Similarly for the tests (x == 0), (x != 0), (x == 1) and (x != 1).
Or (assuming c1 and c2 are constants):
bb0:
x = a + c1;
if (x EQ/NEQ c2) goto ... else goto ...
Will be transformed into:
bb0:
if (a EQ/NEQ (c2 - c1)) goto ... else goto ...
Similarly for x = a - c1.
Or
bb0:
x = !a
if (x) goto ... else goto ...
Will be transformed into:
bb0:
if (a == 0) goto ... else goto ...
Similarly for the tests (x == 0), (x != 0), (x == 1) and (x != 1).
For these cases, we propagate A into all, possibly more than one,
COND_EXPRs that use X.
Or
bb0:
x = (typecast) a
if (x) goto ... else goto ...
Will be transformed into:
bb0:
if (a != 0) goto ... else goto ...
(Assuming a is an integral type and x is a boolean or x is an
integral and a is a boolean.)
Similarly for the tests (x == 0), (x != 0), (x == 1) and (x != 1).
For these cases, we propagate A into all, possibly more than one,
COND_EXPRs that use X.
In addition to eliminating the variable and the statement which assigns
a value to the variable, we may be able to later thread the jump without
adding insane complexity in the dominator optimizer.
Also note these transformations can cascade. We handle this by having
a worklist of COND_EXPR statements to examine. As we make a change to
a statement, we put it back on the worklist to examine on the next
iteration of the main loop.
A second class of propagation opportunities arises for ADDR_EXPR
nodes.
ptr = &x->y->z;
res = *ptr;
Will get turned into
res = x->y->z;
Or
ptr = (type1*)&type2var;
res = *ptr
Will get turned into (if type1 and type2 are the same size
and neither have volatile on them):
res = VIEW_CONVERT_EXPR<type1>(type2var)
Or
ptr = &x[0];
ptr2 = ptr + <constant>;
Will get turned into
ptr2 = &x[constant/elementsize];
Or
ptr = &x[0];
offset = index * element_size;
offset_p = (pointer) offset;
ptr2 = ptr + offset_p
Will get turned into:
ptr2 = &x[index];
Or
ssa = (int) decl
res = ssa & 1
Provided that decl has known alignment >= 2, will get turned into
res = 0
We also propagate casts into SWITCH_EXPR and COND_EXPR conditions to
allow us to remove the cast and {NOT_EXPR,NEG_EXPR} into a subsequent
{NOT_EXPR,NEG_EXPR}.
This will (of course) be extended as other needs arise. */
/* Data structure that contains simplifiable vectorized permute sequences.
See recognise_vec_perm_simplify_seq () for a description of the sequence. */
struct _vec_perm_simplify_seq
{
/* Defining stmts of vectors in the sequence. */
gassign *v_1_stmt;
gassign *v_2_stmt;
gassign *v_x_stmt;
gassign *v_y_stmt;
/* Final permute statment. */
gassign *stmt;
/* New selector indices for stmt. */
tree new_sel;
/* Elements of each vector and selector. */
unsigned int nelts;
};
typedef struct _vec_perm_simplify_seq *vec_perm_simplify_seq;
static bool forward_propagate_addr_expr (tree, tree, bool);
/* Set to true if we delete dead edges during the optimization. */
static bool cfg_changed;
static tree rhs_to_tree (tree type, gimple *stmt);
static bitmap to_purge;
/* Const-and-copy lattice. */
static vec<tree> lattice;
/* Set the lattice entry for NAME to VAL. */
static void
fwprop_set_lattice_val (tree name, tree val)
{
if (TREE_CODE (name) == SSA_NAME)
{
if (SSA_NAME_VERSION (name) >= lattice.length ())
{
lattice.reserve (num_ssa_names - lattice.length ());
lattice.quick_grow_cleared (num_ssa_names);
}
lattice[SSA_NAME_VERSION (name)] = val;
/* As this now constitutes a copy duplicate points-to
and range info appropriately. */
if (TREE_CODE (val) == SSA_NAME)
maybe_duplicate_ssa_info_at_copy (name, val);
}
}
/* Invalidate the lattice entry for NAME, done when releasing SSA names. */
static void
fwprop_invalidate_lattice (tree name)
{
if (name
&& TREE_CODE (name) == SSA_NAME
&& SSA_NAME_VERSION (name) < lattice.length ())
lattice[SSA_NAME_VERSION (name)] = NULL_TREE;
}
/* Get the statement we can propagate from into NAME skipping
trivial copies. Returns the statement which defines the
propagation source or NULL_TREE if there is no such one.
If SINGLE_USE_ONLY is set considers only sources which have
a single use chain up to NAME. If SINGLE_USE_P is non-null,
it is set to whether the chain to NAME is a single use chain
or not. SINGLE_USE_P is not written to if SINGLE_USE_ONLY is set. */
static gimple *
get_prop_source_stmt (tree name, bool single_use_only, bool *single_use_p)
{
bool single_use = true;
do {
gimple *def_stmt = SSA_NAME_DEF_STMT (name);
if (!has_single_use (name))
{
single_use = false;
if (single_use_only)
return NULL;
}
/* If name is defined by a PHI node or is the default def, bail out. */
if (!is_gimple_assign (def_stmt))
return NULL;
/* If def_stmt is a simple copy, continue looking. */
if (gimple_assign_rhs_code (def_stmt) == SSA_NAME)
name = gimple_assign_rhs1 (def_stmt);
else
{
if (!single_use_only && single_use_p)
*single_use_p = single_use;
return def_stmt;
}
} while (1);
}
/* Checks if the destination ssa name in DEF_STMT can be used as
propagation source. Returns true if so, otherwise false. */
static bool
can_propagate_from (gimple *def_stmt)
{
gcc_assert (is_gimple_assign (def_stmt));
/* If the rhs has side-effects we cannot propagate from it. */
if (gimple_has_volatile_ops (def_stmt))
return false;
/* If the rhs is a load we cannot propagate from it. */
if (TREE_CODE_CLASS (gimple_assign_rhs_code (def_stmt)) == tcc_reference
|| TREE_CODE_CLASS (gimple_assign_rhs_code (def_stmt)) == tcc_declaration)
return false;
/* Constants can be always propagated. */
if (gimple_assign_single_p (def_stmt)
&& is_gimple_min_invariant (gimple_assign_rhs1 (def_stmt)))
return true;
/* We cannot propagate ssa names that occur in abnormal phi nodes. */
if (stmt_references_abnormal_ssa_name (def_stmt))
return false;
/* If the definition is a conversion of a pointer to a function type,
then we cannot apply optimizations as some targets require
function pointers to be canonicalized and in this case this
optimization could eliminate a necessary canonicalization. */
if (CONVERT_EXPR_CODE_P (gimple_assign_rhs_code (def_stmt)))
{
tree rhs = gimple_assign_rhs1 (def_stmt);
if (FUNCTION_POINTER_TYPE_P (TREE_TYPE (rhs)))
return false;
}
return true;
}
/* Remove a chain of dead statements starting at the definition of
NAME. The chain is linked via the first operand of the defining statements.
If NAME was replaced in its only use then this function can be used
to clean up dead stmts. The function handles already released SSA
names gracefully.
Returns true if cleanup-cfg has to run. */
static bool
remove_prop_source_from_use (tree name)
{
gimple_stmt_iterator gsi;
gimple *stmt;
bool cfg_changed = false;
do {
basic_block bb;
if (SSA_NAME_IN_FREE_LIST (name)
|| SSA_NAME_IS_DEFAULT_DEF (name)
|| !has_zero_uses (name))
return cfg_changed;
stmt = SSA_NAME_DEF_STMT (name);
if (gimple_code (stmt) == GIMPLE_PHI
|| gimple_has_side_effects (stmt))
return cfg_changed;
bb = gimple_bb (stmt);
gsi = gsi_for_stmt (stmt);
unlink_stmt_vdef (stmt);
if (gsi_remove (&gsi, true))
bitmap_set_bit (to_purge, bb->index);
fwprop_invalidate_lattice (gimple_get_lhs (stmt));
release_defs (stmt);
name = is_gimple_assign (stmt) ? gimple_assign_rhs1 (stmt) : NULL_TREE;
} while (name && TREE_CODE (name) == SSA_NAME);
return cfg_changed;
}
/* Return the rhs of a gassign *STMT in a form of a single tree,
converted to type TYPE.
This should disappear, but is needed so we can combine expressions and use
the fold() interfaces. Long term, we need to develop folding and combine
routines that deal with gimple exclusively . */
static tree
rhs_to_tree (tree type, gimple *stmt)
{
location_t loc = gimple_location (stmt);
enum tree_code code = gimple_assign_rhs_code (stmt);
switch (get_gimple_rhs_class (code))
{
case GIMPLE_TERNARY_RHS:
return fold_build3_loc (loc, code, type, gimple_assign_rhs1 (stmt),
gimple_assign_rhs2 (stmt),
gimple_assign_rhs3 (stmt));
case GIMPLE_BINARY_RHS:
return fold_build2_loc (loc, code, type, gimple_assign_rhs1 (stmt),
gimple_assign_rhs2 (stmt));
case GIMPLE_UNARY_RHS:
return build1 (code, type, gimple_assign_rhs1 (stmt));
case GIMPLE_SINGLE_RHS:
return gimple_assign_rhs1 (stmt);
default:
gcc_unreachable ();
}
}
/* Combine OP0 CODE OP1 in the context of a COND_EXPR. Returns
the folded result in a form suitable for COND_EXPR_COND or
NULL_TREE, if there is no suitable simplified form. If
INVARIANT_ONLY is true only gimple_min_invariant results are
considered simplified. */
static tree
combine_cond_expr_cond (gimple *stmt, enum tree_code code, tree type,
tree op0, tree op1, bool invariant_only)
{
tree t;
gcc_assert (TREE_CODE_CLASS (code) == tcc_comparison);
fold_defer_overflow_warnings ();
t = fold_binary_loc (gimple_location (stmt), code, type, op0, op1);
if (!t)
{
fold_undefer_overflow_warnings (false, NULL, 0);
return NULL_TREE;
}
/* Require that we got a boolean type out if we put one in. */
gcc_assert (TREE_CODE (TREE_TYPE (t)) == TREE_CODE (type));
/* Canonicalize the combined condition for use in a COND_EXPR. */
t = canonicalize_cond_expr_cond (t);
/* Bail out if we required an invariant but didn't get one. */
if (!t || (invariant_only && !is_gimple_min_invariant (t)))
{
fold_undefer_overflow_warnings (false, NULL, 0);
return NULL_TREE;
}
bool nowarn = warning_suppressed_p (stmt, OPT_Wstrict_overflow);
fold_undefer_overflow_warnings (!nowarn, stmt, 0);
return t;
}
/* Combine the comparison OP0 CODE OP1 at LOC with the defining statements
of its operand. Return a new comparison tree or NULL_TREE if there
were no simplifying combines. */
static tree
forward_propagate_into_comparison_1 (gimple *stmt,
enum tree_code code, tree type,
tree op0, tree op1)
{
tree tmp = NULL_TREE;
tree rhs0 = NULL_TREE, rhs1 = NULL_TREE;
bool single_use0_p = false, single_use1_p = false;
/* For comparisons use the first operand, that is likely to
simplify comparisons against constants. */
if (TREE_CODE (op0) == SSA_NAME)
{
gimple *def_stmt = get_prop_source_stmt (op0, false, &single_use0_p);
if (def_stmt && can_propagate_from (def_stmt))
{
enum tree_code def_code = gimple_assign_rhs_code (def_stmt);
bool invariant_only_p = !single_use0_p;
rhs0 = rhs_to_tree (TREE_TYPE (op1), def_stmt);
/* Always combine comparisons or conversions from booleans. */
if (TREE_CODE (op1) == INTEGER_CST
&& ((CONVERT_EXPR_CODE_P (def_code)
&& TREE_CODE (TREE_TYPE (TREE_OPERAND (rhs0, 0)))
== BOOLEAN_TYPE)
|| TREE_CODE_CLASS (def_code) == tcc_comparison))
invariant_only_p = false;
tmp = combine_cond_expr_cond (stmt, code, type,
rhs0, op1, invariant_only_p);
if (tmp)
return tmp;
}
}
/* If that wasn't successful, try the second operand. */
if (TREE_CODE (op1) == SSA_NAME)
{
gimple *def_stmt = get_prop_source_stmt (op1, false, &single_use1_p);
if (def_stmt && can_propagate_from (def_stmt))
{
rhs1 = rhs_to_tree (TREE_TYPE (op0), def_stmt);
tmp = combine_cond_expr_cond (stmt, code, type,
op0, rhs1, !single_use1_p);
if (tmp)
return tmp;
}
}
/* If that wasn't successful either, try both operands. */
if (rhs0 != NULL_TREE
&& rhs1 != NULL_TREE)
tmp = combine_cond_expr_cond (stmt, code, type,
rhs0, rhs1,
!(single_use0_p && single_use1_p));
return tmp;
}
/* Propagate from the ssa name definition statements of the assignment
from a comparison at *GSI into the conditional if that simplifies it.
Returns 1 if the stmt was modified and 2 if the CFG needs cleanup,
otherwise returns 0. */
static int
forward_propagate_into_comparison (gimple_stmt_iterator *gsi)
{
gimple *stmt = gsi_stmt (*gsi);
tree tmp;
bool cfg_changed = false;
tree type = TREE_TYPE (gimple_assign_lhs (stmt));
tree rhs1 = gimple_assign_rhs1 (stmt);
tree rhs2 = gimple_assign_rhs2 (stmt);
/* Combine the comparison with defining statements. */
tmp = forward_propagate_into_comparison_1 (stmt,
gimple_assign_rhs_code (stmt),
type, rhs1, rhs2);
if (tmp && useless_type_conversion_p (type, TREE_TYPE (tmp)))
{
gimple_assign_set_rhs_from_tree (gsi, tmp);
fold_stmt (gsi);
update_stmt (gsi_stmt (*gsi));
if (TREE_CODE (rhs1) == SSA_NAME)
cfg_changed |= remove_prop_source_from_use (rhs1);
if (TREE_CODE (rhs2) == SSA_NAME)
cfg_changed |= remove_prop_source_from_use (rhs2);
return cfg_changed ? 2 : 1;
}
return 0;
}
/* Propagate from the ssa name definition statements of COND_EXPR
in GIMPLE_COND statement STMT into the conditional if that simplifies it.
Returns zero if no statement was changed, one if there were
changes and two if cfg_cleanup needs to run. */
static int
forward_propagate_into_gimple_cond (gcond *stmt)
{
tree tmp;
enum tree_code code = gimple_cond_code (stmt);
bool cfg_changed = false;
tree rhs1 = gimple_cond_lhs (stmt);
tree rhs2 = gimple_cond_rhs (stmt);
/* We can do tree combining on SSA_NAME and comparison expressions. */
if (TREE_CODE_CLASS (gimple_cond_code (stmt)) != tcc_comparison)
return 0;
tmp = forward_propagate_into_comparison_1 (stmt, code,
boolean_type_node,
rhs1, rhs2);
if (tmp
&& is_gimple_condexpr_for_cond (tmp))
{
if (dump_file)
{
fprintf (dump_file, " Replaced '");
print_gimple_expr (dump_file, stmt, 0);
fprintf (dump_file, "' with '");
print_generic_expr (dump_file, tmp);
fprintf (dump_file, "'\n");
}
gimple_cond_set_condition_from_tree (stmt, unshare_expr (tmp));
update_stmt (stmt);
if (TREE_CODE (rhs1) == SSA_NAME)
cfg_changed |= remove_prop_source_from_use (rhs1);
if (TREE_CODE (rhs2) == SSA_NAME)
cfg_changed |= remove_prop_source_from_use (rhs2);
return (cfg_changed || is_gimple_min_invariant (tmp)) ? 2 : 1;
}
/* Canonicalize _Bool == 0 and _Bool != 1 to _Bool != 0 by swapping edges. */
if ((TREE_CODE (TREE_TYPE (rhs1)) == BOOLEAN_TYPE
|| (INTEGRAL_TYPE_P (TREE_TYPE (rhs1))
&& TYPE_PRECISION (TREE_TYPE (rhs1)) == 1))
&& ((code == EQ_EXPR
&& integer_zerop (rhs2))
|| (code == NE_EXPR
&& integer_onep (rhs2))))
{
basic_block bb = gimple_bb (stmt);
gimple_cond_set_code (stmt, NE_EXPR);
gimple_cond_set_rhs (stmt, build_zero_cst (TREE_TYPE (rhs1)));
EDGE_SUCC (bb, 0)->flags ^= (EDGE_TRUE_VALUE|EDGE_FALSE_VALUE);
EDGE_SUCC (bb, 1)->flags ^= (EDGE_TRUE_VALUE|EDGE_FALSE_VALUE);
return 1;
}
return 0;
}
/* We've just substituted an ADDR_EXPR into stmt. Update all the
relevant data structures to match. */
static void
tidy_after_forward_propagate_addr (gimple *stmt)
{
/* We may have turned a trapping insn into a non-trapping insn. */
if (maybe_clean_or_replace_eh_stmt (stmt, stmt))
bitmap_set_bit (to_purge, gimple_bb (stmt)->index);
if (TREE_CODE (gimple_assign_rhs1 (stmt)) == ADDR_EXPR)
recompute_tree_invariant_for_addr_expr (gimple_assign_rhs1 (stmt));
}
/* NAME is a SSA_NAME representing DEF_RHS which is of the form
ADDR_EXPR <whatever>.
Try to forward propagate the ADDR_EXPR into the use USE_STMT.
Often this will allow for removal of an ADDR_EXPR and INDIRECT_REF
node or for recovery of array indexing from pointer arithmetic.
Return true if the propagation was successful (the propagation can
be not totally successful, yet things may have been changed). */
static bool
forward_propagate_addr_expr_1 (tree name, tree def_rhs,
gimple_stmt_iterator *use_stmt_gsi,
bool single_use_p)
{
tree lhs, rhs, rhs2, array_ref;
gimple *use_stmt = gsi_stmt (*use_stmt_gsi);
enum tree_code rhs_code;
bool res = true;
gcc_assert (TREE_CODE (def_rhs) == ADDR_EXPR);
lhs = gimple_assign_lhs (use_stmt);
rhs_code = gimple_assign_rhs_code (use_stmt);
rhs = gimple_assign_rhs1 (use_stmt);
/* Do not perform copy-propagation but recurse through copy chains. */
if (TREE_CODE (lhs) == SSA_NAME
&& rhs_code == SSA_NAME)
return forward_propagate_addr_expr (lhs, def_rhs, single_use_p);
/* The use statement could be a conversion. Recurse to the uses of the
lhs as copyprop does not copy through pointer to integer to pointer
conversions and FRE does not catch all cases either.
Treat the case of a single-use name and
a conversion to def_rhs type separate, though. */
if (TREE_CODE (lhs) == SSA_NAME
&& CONVERT_EXPR_CODE_P (rhs_code))
{
/* If there is a point in a conversion chain where the types match
so we can remove a conversion re-materialize the address here
and stop. */
if (single_use_p
&& useless_type_conversion_p (TREE_TYPE (lhs), TREE_TYPE (def_rhs)))
{
gimple_assign_set_rhs1 (use_stmt, unshare_expr (def_rhs));
gimple_assign_set_rhs_code (use_stmt, TREE_CODE (def_rhs));
return true;
}
/* Else recurse if the conversion preserves the address value. */
if ((INTEGRAL_TYPE_P (TREE_TYPE (lhs))
|| POINTER_TYPE_P (TREE_TYPE (lhs)))
&& (TYPE_PRECISION (TREE_TYPE (lhs))
>= TYPE_PRECISION (TREE_TYPE (def_rhs))))
return forward_propagate_addr_expr (lhs, def_rhs, single_use_p);
return false;
}
/* If this isn't a conversion chain from this on we only can propagate
into compatible pointer contexts. */
if (!types_compatible_p (TREE_TYPE (name), TREE_TYPE (def_rhs)))
return false;
/* Propagate through constant pointer adjustments. */
if (TREE_CODE (lhs) == SSA_NAME
&& rhs_code == POINTER_PLUS_EXPR
&& rhs == name
&& TREE_CODE (gimple_assign_rhs2 (use_stmt)) == INTEGER_CST)
{
tree new_def_rhs;
/* As we come here with non-invariant addresses in def_rhs we need
to make sure we can build a valid constant offsetted address
for further propagation. Simply rely on fold building that
and check after the fact. */
new_def_rhs = fold_build2 (MEM_REF, TREE_TYPE (TREE_TYPE (rhs)),
def_rhs,
fold_convert (ptr_type_node,
gimple_assign_rhs2 (use_stmt)));
if (TREE_CODE (new_def_rhs) == MEM_REF
&& !is_gimple_mem_ref_addr (TREE_OPERAND (new_def_rhs, 0)))
return false;
new_def_rhs = build1 (ADDR_EXPR, TREE_TYPE (rhs), new_def_rhs);
/* Recurse. If we could propagate into all uses of lhs do not
bother to replace into the current use but just pretend we did. */
if (forward_propagate_addr_expr (lhs, new_def_rhs, single_use_p))
return true;
if (useless_type_conversion_p (TREE_TYPE (lhs),
TREE_TYPE (new_def_rhs)))
gimple_assign_set_rhs_with_ops (use_stmt_gsi, TREE_CODE (new_def_rhs),
new_def_rhs);
else if (is_gimple_min_invariant (new_def_rhs))
gimple_assign_set_rhs_with_ops (use_stmt_gsi, NOP_EXPR, new_def_rhs);
else
return false;
gcc_assert (gsi_stmt (*use_stmt_gsi) == use_stmt);
update_stmt (use_stmt);
return true;
}
/* Now strip away any outer COMPONENT_REF/ARRAY_REF nodes from the LHS.
ADDR_EXPR will not appear on the LHS. */
tree *lhsp = gimple_assign_lhs_ptr (use_stmt);
while (handled_component_p (*lhsp))
lhsp = &TREE_OPERAND (*lhsp, 0);
lhs = *lhsp;
/* Now see if the LHS node is a MEM_REF using NAME. If so,
propagate the ADDR_EXPR into the use of NAME and fold the result. */
if (TREE_CODE (lhs) == MEM_REF
&& TREE_OPERAND (lhs, 0) == name)
{
tree def_rhs_base;
poly_int64 def_rhs_offset;
/* If the address is invariant we can always fold it. */
if ((def_rhs_base = get_addr_base_and_unit_offset (TREE_OPERAND (def_rhs, 0),
&def_rhs_offset)))
{
poly_offset_int off = mem_ref_offset (lhs);
tree new_ptr;
off += def_rhs_offset;
if (TREE_CODE (def_rhs_base) == MEM_REF)
{
off += mem_ref_offset (def_rhs_base);
new_ptr = TREE_OPERAND (def_rhs_base, 0);
}
else
new_ptr = build_fold_addr_expr (def_rhs_base);
TREE_OPERAND (lhs, 0) = new_ptr;
TREE_OPERAND (lhs, 1)
= wide_int_to_tree (TREE_TYPE (TREE_OPERAND (lhs, 1)), off);
tidy_after_forward_propagate_addr (use_stmt);
/* Continue propagating into the RHS if this was not the only use. */
if (single_use_p)
return true;
}
/* If the LHS is a plain dereference and the value type is the same as
that of the pointed-to type of the address we can put the
dereferenced address on the LHS preserving the original alias-type. */
else if (integer_zerop (TREE_OPERAND (lhs, 1))
&& ((gimple_assign_lhs (use_stmt) == lhs
&& useless_type_conversion_p
(TREE_TYPE (TREE_OPERAND (def_rhs, 0)),
TREE_TYPE (gimple_assign_rhs1 (use_stmt))))
|| types_compatible_p (TREE_TYPE (lhs),
TREE_TYPE (TREE_OPERAND (def_rhs, 0))))
/* Don't forward anything into clobber stmts if it would result
in the lhs no longer being a MEM_REF. */
&& (!gimple_clobber_p (use_stmt)
|| TREE_CODE (TREE_OPERAND (def_rhs, 0)) == MEM_REF))
{
tree *def_rhs_basep = &TREE_OPERAND (def_rhs, 0);
tree new_offset, new_base, saved, new_lhs;
while (handled_component_p (*def_rhs_basep))
def_rhs_basep = &TREE_OPERAND (*def_rhs_basep, 0);
saved = *def_rhs_basep;
if (TREE_CODE (*def_rhs_basep) == MEM_REF)
{
new_base = TREE_OPERAND (*def_rhs_basep, 0);
new_offset = fold_convert (TREE_TYPE (TREE_OPERAND (lhs, 1)),
TREE_OPERAND (*def_rhs_basep, 1));
}
else
{
new_base = build_fold_addr_expr (*def_rhs_basep);
new_offset = TREE_OPERAND (lhs, 1);
}
*def_rhs_basep = build2 (MEM_REF, TREE_TYPE (*def_rhs_basep),
new_base, new_offset);
TREE_THIS_VOLATILE (*def_rhs_basep) = TREE_THIS_VOLATILE (lhs);
TREE_SIDE_EFFECTS (*def_rhs_basep) = TREE_SIDE_EFFECTS (lhs);
TREE_THIS_NOTRAP (*def_rhs_basep) = TREE_THIS_NOTRAP (lhs);
new_lhs = unshare_expr (TREE_OPERAND (def_rhs, 0));
*lhsp = new_lhs;
TREE_THIS_VOLATILE (new_lhs) = TREE_THIS_VOLATILE (lhs);
TREE_SIDE_EFFECTS (new_lhs) = TREE_SIDE_EFFECTS (lhs);
*def_rhs_basep = saved;
tidy_after_forward_propagate_addr (use_stmt);
/* Continue propagating into the RHS if this was not the
only use. */
if (single_use_p)
return true;
}
else
/* We can have a struct assignment dereferencing our name twice.
Note that we didn't propagate into the lhs to not falsely
claim we did when propagating into the rhs. */
res = false;
}
/* Strip away any outer COMPONENT_REF, ARRAY_REF or ADDR_EXPR
nodes from the RHS. */
tree *rhsp = gimple_assign_rhs1_ptr (use_stmt);
if (TREE_CODE (*rhsp) == ADDR_EXPR)
rhsp = &TREE_OPERAND (*rhsp, 0);
while (handled_component_p (*rhsp))
rhsp = &TREE_OPERAND (*rhsp, 0);
rhs = *rhsp;
/* Now see if the RHS node is a MEM_REF using NAME. If so,
propagate the ADDR_EXPR into the use of NAME and fold the result. */
if (TREE_CODE (rhs) == MEM_REF
&& TREE_OPERAND (rhs, 0) == name)
{
tree def_rhs_base;
poly_int64 def_rhs_offset;
if ((def_rhs_base = get_addr_base_and_unit_offset (TREE_OPERAND (def_rhs, 0),
&def_rhs_offset)))
{
poly_offset_int off = mem_ref_offset (rhs);
tree new_ptr;
off += def_rhs_offset;
if (TREE_CODE (def_rhs_base) == MEM_REF)
{
off += mem_ref_offset (def_rhs_base);
new_ptr = TREE_OPERAND (def_rhs_base, 0);
}
else
new_ptr = build_fold_addr_expr (def_rhs_base);
TREE_OPERAND (rhs, 0) = new_ptr;
TREE_OPERAND (rhs, 1)
= wide_int_to_tree (TREE_TYPE (TREE_OPERAND (rhs, 1)), off);
fold_stmt_inplace (use_stmt_gsi);
tidy_after_forward_propagate_addr (use_stmt);
return res;
}
/* If the RHS is a plain dereference and the value type is the same as
that of the pointed-to type of the address we can put the
dereferenced address on the RHS preserving the original alias-type. */
else if (integer_zerop (TREE_OPERAND (rhs, 1))
&& ((gimple_assign_rhs1 (use_stmt) == rhs
&& useless_type_conversion_p
(TREE_TYPE (gimple_assign_lhs (use_stmt)),
TREE_TYPE (TREE_OPERAND (def_rhs, 0))))
|| types_compatible_p (TREE_TYPE (rhs),
TREE_TYPE (TREE_OPERAND (def_rhs, 0)))))
{
tree *def_rhs_basep = &TREE_OPERAND (def_rhs, 0);
tree new_offset, new_base, saved, new_rhs;
while (handled_component_p (*def_rhs_basep))
def_rhs_basep = &TREE_OPERAND (*def_rhs_basep, 0);
saved = *def_rhs_basep;
if (TREE_CODE (*def_rhs_basep) == MEM_REF)
{
new_base = TREE_OPERAND (*def_rhs_basep, 0);
new_offset = fold_convert (TREE_TYPE (TREE_OPERAND (rhs, 1)),
TREE_OPERAND (*def_rhs_basep, 1));
}
else
{
new_base = build_fold_addr_expr (*def_rhs_basep);
new_offset = TREE_OPERAND (rhs, 1);
}
*def_rhs_basep = build2 (MEM_REF, TREE_TYPE (*def_rhs_basep),
new_base, new_offset);
TREE_THIS_VOLATILE (*def_rhs_basep) = TREE_THIS_VOLATILE (rhs);
TREE_SIDE_EFFECTS (*def_rhs_basep) = TREE_SIDE_EFFECTS (rhs);
TREE_THIS_NOTRAP (*def_rhs_basep) = TREE_THIS_NOTRAP (rhs);
new_rhs = unshare_expr (TREE_OPERAND (def_rhs, 0));
*rhsp = new_rhs;
TREE_THIS_VOLATILE (new_rhs) = TREE_THIS_VOLATILE (rhs);
TREE_SIDE_EFFECTS (new_rhs) = TREE_SIDE_EFFECTS (rhs);
*def_rhs_basep = saved;
fold_stmt_inplace (use_stmt_gsi);
tidy_after_forward_propagate_addr (use_stmt);
return res;
}
}
/* If the use of the ADDR_EXPR is not a POINTER_PLUS_EXPR, there
is nothing to do. */
if (gimple_assign_rhs_code (use_stmt) != POINTER_PLUS_EXPR
|| gimple_assign_rhs1 (use_stmt) != name)
return false;
/* The remaining cases are all for turning pointer arithmetic into
array indexing. They only apply when we have the address of
element zero in an array. If that is not the case then there
is nothing to do. */
array_ref = TREE_OPERAND (def_rhs, 0);
if ((TREE_CODE (array_ref) != ARRAY_REF
|| TREE_CODE (TREE_TYPE (TREE_OPERAND (array_ref, 0))) != ARRAY_TYPE
|| TREE_CODE (TREE_OPERAND (array_ref, 1)) != INTEGER_CST)
&& TREE_CODE (TREE_TYPE (array_ref)) != ARRAY_TYPE)
return false;
rhs2 = gimple_assign_rhs2 (use_stmt);
/* Optimize &x[C1] p+ C2 to &x p+ C3 with C3 = C1 * element_size + C2. */
if (TREE_CODE (rhs2) == INTEGER_CST)
{
tree new_rhs = build1_loc (gimple_location (use_stmt),
ADDR_EXPR, TREE_TYPE (def_rhs),
fold_build2 (MEM_REF,
TREE_TYPE (TREE_TYPE (def_rhs)),
unshare_expr (def_rhs),
fold_convert (ptr_type_node,
rhs2)));
gimple_assign_set_rhs_from_tree (use_stmt_gsi, new_rhs);
use_stmt = gsi_stmt (*use_stmt_gsi);
update_stmt (use_stmt);
tidy_after_forward_propagate_addr (use_stmt);
return true;
}
return false;
}
/* STMT is a statement of the form SSA_NAME = ADDR_EXPR <whatever>.
Try to forward propagate the ADDR_EXPR into all uses of the SSA_NAME.
Often this will allow for removal of an ADDR_EXPR and INDIRECT_REF
node or for recovery of array indexing from pointer arithmetic.
PARENT_SINGLE_USE_P tells if, when in a recursive invocation, NAME was
the single use in the previous invocation. Pass true when calling
this as toplevel.
Returns true, if all uses have been propagated into. */
static bool
forward_propagate_addr_expr (tree name, tree rhs, bool parent_single_use_p)
{
imm_use_iterator iter;
gimple *use_stmt;
bool all = true;
bool single_use_p = parent_single_use_p && has_single_use (name);
FOR_EACH_IMM_USE_STMT (use_stmt, iter, name)
{
bool result;
tree use_rhs;
/* If the use is not in a simple assignment statement, then
there is nothing we can do. */
if (!is_gimple_assign (use_stmt))
{
if (!is_gimple_debug (use_stmt))
all = false;
continue;
}
gimple_stmt_iterator gsi = gsi_for_stmt (use_stmt);
result = forward_propagate_addr_expr_1 (name, rhs, &gsi,
single_use_p);
/* If the use has moved to a different statement adjust
the update machinery for the old statement too. */
if (use_stmt != gsi_stmt (gsi))
{
update_stmt (use_stmt);
use_stmt = gsi_stmt (gsi);
}
update_stmt (use_stmt);
all &= result;
/* Remove intermediate now unused copy and conversion chains. */
use_rhs = gimple_assign_rhs1 (use_stmt);
if (result
&& TREE_CODE (gimple_assign_lhs (use_stmt)) == SSA_NAME
&& TREE_CODE (use_rhs) == SSA_NAME
&& has_zero_uses (gimple_assign_lhs (use_stmt)))
{
gimple_stmt_iterator gsi = gsi_for_stmt (use_stmt);
fwprop_invalidate_lattice (gimple_get_lhs (use_stmt));
release_defs (use_stmt);
gsi_remove (&gsi, true);
}
}
return all && has_zero_uses (name);
}
/* Helper function for simplify_gimple_switch. Remove case labels that
have values outside the range of the new type. */
static void
simplify_gimple_switch_label_vec (gswitch *stmt, tree index_type,
vec<std::pair<int, int> > &edges_to_remove)
{
unsigned int branch_num = gimple_switch_num_labels (stmt);
auto_vec<tree> labels (branch_num);
unsigned int i, len;
/* Collect the existing case labels in a VEC, and preprocess it as if
we are gimplifying a GENERIC SWITCH_EXPR. */
for (i = 1; i < branch_num; i++)
labels.quick_push (gimple_switch_label (stmt, i));
preprocess_case_label_vec_for_gimple (labels, index_type, NULL);
/* If any labels were removed, replace the existing case labels
in the GIMPLE_SWITCH statement with the correct ones.
Note that the type updates were done in-place on the case labels,
so we only have to replace the case labels in the GIMPLE_SWITCH
if the number of labels changed. */
len = labels.length ();
if (len < branch_num - 1)
{
bitmap target_blocks;
edge_iterator ei;
edge e;
/* Corner case: *all* case labels have been removed as being
out-of-range for INDEX_TYPE. Push one label and let the
CFG cleanups deal with this further. */
if (len == 0)
{
tree label, elt;
label = CASE_LABEL (gimple_switch_default_label (stmt));
elt = build_case_label (build_int_cst (index_type, 0), NULL, label);
labels.quick_push (elt);
len = 1;
}
for (i = 0; i < labels.length (); i++)
gimple_switch_set_label (stmt, i + 1, labels[i]);
for (i++ ; i < branch_num; i++)
gimple_switch_set_label (stmt, i, NULL_TREE);
gimple_switch_set_num_labels (stmt, len + 1);
/* Cleanup any edges that are now dead. */
target_blocks = BITMAP_ALLOC (NULL);
for (i = 0; i < gimple_switch_num_labels (stmt); i++)
{
tree elt = gimple_switch_label (stmt, i);
basic_block target = label_to_block (cfun, CASE_LABEL (elt));
bitmap_set_bit (target_blocks, target->index);
}
for (ei = ei_start (gimple_bb (stmt)->succs); (e = ei_safe_edge (ei)); )
{
if (! bitmap_bit_p (target_blocks, e->dest->index))
edges_to_remove.safe_push (std::make_pair (e->src->index,
e->dest->index));
else
ei_next (&ei);
}
BITMAP_FREE (target_blocks);
}
}
/* STMT is a SWITCH_EXPR for which we attempt to find equivalent forms of
the condition which we may be able to optimize better. */
static bool
simplify_gimple_switch (gswitch *stmt,
vec<std::pair<int, int> > &edges_to_remove)
{
/* The optimization that we really care about is removing unnecessary
casts. That will let us do much better in propagating the inferred
constant at the switch target. */
tree cond = gimple_switch_index (stmt);
if (TREE_CODE (cond) == SSA_NAME)
{
gimple *def_stmt = SSA_NAME_DEF_STMT (cond);
if (gimple_assign_cast_p (def_stmt))
{
tree def = gimple_assign_rhs1 (def_stmt);
if (TREE_CODE (def) != SSA_NAME)
return false;
/* If we have an extension or sign-change that preserves the
values we check against then we can copy the source value into
the switch. */
tree ti = TREE_TYPE (def);
if (INTEGRAL_TYPE_P (ti)
&& TYPE_PRECISION (ti) <= TYPE_PRECISION (TREE_TYPE (cond)))
{
size_t n = gimple_switch_num_labels (stmt);
tree min = NULL_TREE, max = NULL_TREE;
if (n > 1)
{
min = CASE_LOW (gimple_switch_label (stmt, 1));
if (CASE_HIGH (gimple_switch_label (stmt, n - 1)))
max = CASE_HIGH (gimple_switch_label (stmt, n - 1));
else
max = CASE_LOW (gimple_switch_label (stmt, n - 1));
}
if ((!min || int_fits_type_p (min, ti))
&& (!max || int_fits_type_p (max, ti)))
{
gimple_switch_set_index (stmt, def);
simplify_gimple_switch_label_vec (stmt, ti,
edges_to_remove);
update_stmt (stmt);
return true;
}
}
}
}
return false;
}
/* For pointers p2 and p1 return p2 - p1 if the
difference is known and constant, otherwise return NULL. */
static tree
constant_pointer_difference (tree p1, tree p2)
{
int i, j;
#define CPD_ITERATIONS 5
tree exps[2][CPD_ITERATIONS];
tree offs[2][CPD_ITERATIONS];
int cnt[2];
for (i = 0; i < 2; i++)
{
tree p = i ? p1 : p2;
tree off = size_zero_node;
gimple *stmt;
enum tree_code code;
/* For each of p1 and p2 we need to iterate at least
twice, to handle ADDR_EXPR directly in p1/p2,
SSA_NAME with ADDR_EXPR or POINTER_PLUS_EXPR etc.
on definition's stmt RHS. Iterate a few extra times. */
j = 0;
do
{
if (!POINTER_TYPE_P (TREE_TYPE (p)))
break;
if (TREE_CODE (p) == ADDR_EXPR)
{
tree q = TREE_OPERAND (p, 0);
poly_int64 offset;
tree base = get_addr_base_and_unit_offset (q, &offset);
if (base)
{
q = base;
if (maybe_ne (offset, 0))
off = size_binop (PLUS_EXPR, off, size_int (offset));
}
if (TREE_CODE (q) == MEM_REF
&& TREE_CODE (TREE_OPERAND (q, 0)) == SSA_NAME)
{
p = TREE_OPERAND (q, 0);
off = size_binop (PLUS_EXPR, off,
wide_int_to_tree (sizetype,
mem_ref_offset (q)));
}
else
{
exps[i][j] = q;
offs[i][j++] = off;
break;
}
}
if (TREE_CODE (p) != SSA_NAME)
break;
exps[i][j] = p;
offs[i][j++] = off;
if (j == CPD_ITERATIONS)
break;
stmt = SSA_NAME_DEF_STMT (p);
if (!is_gimple_assign (stmt) || gimple_assign_lhs (stmt) != p)
break;
code = gimple_assign_rhs_code (stmt);
if (code == POINTER_PLUS_EXPR)
{
if (TREE_CODE (gimple_assign_rhs2 (stmt)) != INTEGER_CST)
break;
off = size_binop (PLUS_EXPR, off, gimple_assign_rhs2 (stmt));
p = gimple_assign_rhs1 (stmt);
}
else if (code == ADDR_EXPR || CONVERT_EXPR_CODE_P (code))
p = gimple_assign_rhs1 (stmt);
else
break;
}
while (1);
cnt[i] = j;
}
for (i = 0; i < cnt[0]; i++)
for (j = 0; j < cnt[1]; j++)
if (exps[0][i] == exps[1][j])
return size_binop (MINUS_EXPR, offs[0][i], offs[1][j]);
return NULL_TREE;
}
/* *GSI_P is a GIMPLE_CALL to a builtin function.
Optimize
memcpy (p, "abcd", 4);
memset (p + 4, ' ', 3);
into
memcpy (p, "abcd ", 7);
call if the latter can be stored by pieces during expansion.
Optimize
memchr ("abcd", a, 4) == 0;
or
memchr ("abcd", a, 4) != 0;
to
(a == 'a' || a == 'b' || a == 'c' || a == 'd') == 0
or
(a == 'a' || a == 'b' || a == 'c' || a == 'd') != 0
Also canonicalize __atomic_fetch_op (p, x, y) op x
to __atomic_op_fetch (p, x, y) or
__atomic_op_fetch (p, x, y) iop x
to __atomic_fetch_op (p, x, y) when possible (also __sync). */
static bool
simplify_builtin_call (gimple_stmt_iterator *gsi_p, tree callee2)
{
gimple *stmt1, *stmt2 = gsi_stmt (*gsi_p);
enum built_in_function other_atomic = END_BUILTINS;
enum tree_code atomic_op = ERROR_MARK;
tree vuse = gimple_vuse (stmt2);
if (vuse == NULL)
return false;
stmt1 = SSA_NAME_DEF_STMT (vuse);
tree res;
switch (DECL_FUNCTION_CODE (callee2))
{
case BUILT_IN_MEMCHR:
if (gimple_call_num_args (stmt2) == 3
&& (res = gimple_call_lhs (stmt2)) != nullptr
&& use_in_zero_equality (res) != nullptr
&& CHAR_BIT == 8
&& BITS_PER_UNIT == 8)
{
tree ptr = gimple_call_arg (stmt2, 0);
if (TREE_CODE (ptr) != ADDR_EXPR
|| TREE_CODE (TREE_OPERAND (ptr, 0)) != STRING_CST)
break;
unsigned HOST_WIDE_INT slen
= TREE_STRING_LENGTH (TREE_OPERAND (ptr, 0));
/* It must be a non-empty string constant. */
if (slen < 2)
break;
/* For -Os, only simplify strings with a single character. */
if (!optimize_bb_for_speed_p (gimple_bb (stmt2))
&& slen > 2)
break;
tree size = gimple_call_arg (stmt2, 2);
/* Size must be a constant which is <= UNITS_PER_WORD and
<= the string length. */
if (TREE_CODE (size) != INTEGER_CST)
break;
if (!tree_fits_uhwi_p (size))
break;
unsigned HOST_WIDE_INT sz = tree_to_uhwi (size);
if (sz == 0 || sz > UNITS_PER_WORD || sz >= slen)
break;
tree ch = gimple_call_arg (stmt2, 1);
location_t loc = gimple_location (stmt2);
if (!useless_type_conversion_p (char_type_node,
TREE_TYPE (ch)))
ch = fold_convert_loc (loc, char_type_node, ch);
const char *p = TREE_STRING_POINTER (TREE_OPERAND (ptr, 0));
unsigned int isize = sz;
tree *op = XALLOCAVEC (tree, isize);
for (unsigned int i = 0; i < isize; i++)
{
op[i] = build_int_cst (char_type_node, p[i]);
op[i] = fold_build2_loc (loc, EQ_EXPR, boolean_type_node,
op[i], ch);
}
for (unsigned int i = isize - 1; i >= 1; i--)
op[i - 1] = fold_convert_loc (loc, boolean_type_node,
fold_build2_loc (loc,
BIT_IOR_EXPR,
boolean_type_node,
op[i - 1],
op[i]));
res = fold_convert_loc (loc, TREE_TYPE (res), op[0]);
gimplify_and_update_call_from_tree (gsi_p, res);
return true;
}
break;
case BUILT_IN_MEMSET:
if (gimple_call_num_args (stmt2) != 3
|| gimple_call_lhs (stmt2)
|| CHAR_BIT != 8
|| BITS_PER_UNIT != 8)
break;
else
{
tree callee1;
tree ptr1, src1, str1, off1, len1, lhs1;
tree ptr2 = gimple_call_arg (stmt2, 0);
tree val2 = gimple_call_arg (stmt2, 1);
tree len2 = gimple_call_arg (stmt2, 2);
tree diff, vdef, new_str_cst;
gimple *use_stmt;
unsigned int ptr1_align;
unsigned HOST_WIDE_INT src_len;
char *src_buf;
use_operand_p use_p;
if (!tree_fits_shwi_p (val2)
|| !tree_fits_uhwi_p (len2)
|| compare_tree_int (len2, 1024) == 1)
break;
if (is_gimple_call (stmt1))
{
/* If first stmt is a call, it needs to be memcpy
or mempcpy, with string literal as second argument and
constant length. */
callee1 = gimple_call_fndecl (stmt1);
if (callee1 == NULL_TREE
|| !fndecl_built_in_p (callee1, BUILT_IN_NORMAL)
|| gimple_call_num_args (stmt1) != 3)
break;
if (DECL_FUNCTION_CODE (callee1) != BUILT_IN_MEMCPY
&& DECL_FUNCTION_CODE (callee1) != BUILT_IN_MEMPCPY)
break;
ptr1 = gimple_call_arg (stmt1, 0);
src1 = gimple_call_arg (stmt1, 1);
len1 = gimple_call_arg (stmt1, 2);
lhs1 = gimple_call_lhs (stmt1);
if (!tree_fits_uhwi_p (len1))
break;
str1 = string_constant (src1, &off1, NULL, NULL);
if (str1 == NULL_TREE)
break;
if (!tree_fits_uhwi_p (off1)
|| compare_tree_int (off1, TREE_STRING_LENGTH (str1) - 1) > 0
|| compare_tree_int (len1, TREE_STRING_LENGTH (str1)
- tree_to_uhwi (off1)) > 0
|| TREE_CODE (TREE_TYPE (str1)) != ARRAY_TYPE
|| TYPE_MODE (TREE_TYPE (TREE_TYPE (str1)))
!= TYPE_MODE (char_type_node))
break;
}
else if (gimple_assign_single_p (stmt1))
{
/* Otherwise look for length 1 memcpy optimized into
assignment. */
ptr1 = gimple_assign_lhs (stmt1);
src1 = gimple_assign_rhs1 (stmt1);
if (TREE_CODE (ptr1) != MEM_REF
|| TYPE_MODE (TREE_TYPE (ptr1)) != TYPE_MODE (char_type_node)
|| !tree_fits_shwi_p (src1))
break;
ptr1 = build_fold_addr_expr (ptr1);
STRIP_USELESS_TYPE_CONVERSION (ptr1);
callee1 = NULL_TREE;
len1 = size_one_node;
lhs1 = NULL_TREE;
off1 = size_zero_node;
str1 = NULL_TREE;
}
else
break;
diff = constant_pointer_difference (ptr1, ptr2);
if (diff == NULL && lhs1 != NULL)
{
diff = constant_pointer_difference (lhs1, ptr2);
if (DECL_FUNCTION_CODE (callee1) == BUILT_IN_MEMPCPY
&& diff != NULL)
diff = size_binop (PLUS_EXPR, diff,
fold_convert (sizetype, len1));
}
/* If the difference between the second and first destination pointer
is not constant, or is bigger than memcpy length, bail out. */
if (diff == NULL
|| !tree_fits_uhwi_p (diff)
|| tree_int_cst_lt (len1, diff)
|| compare_tree_int (diff, 1024) == 1)
break;
/* Use maximum of difference plus memset length and memcpy length
as the new memcpy length, if it is too big, bail out. */
src_len = tree_to_uhwi (diff);
src_len += tree_to_uhwi (len2);
if (src_len < tree_to_uhwi (len1))
src_len = tree_to_uhwi (len1);
if (src_len > 1024)
break;
/* If mempcpy value is used elsewhere, bail out, as mempcpy
with bigger length will return different result. */
if (lhs1 != NULL_TREE
&& DECL_FUNCTION_CODE (callee1) == BUILT_IN_MEMPCPY
&& (TREE_CODE (lhs1) != SSA_NAME
|| !single_imm_use (lhs1, &use_p, &use_stmt)
|| use_stmt != stmt2))
break;
/* If anything reads memory in between memcpy and memset
call, the modified memcpy call might change it. */
vdef = gimple_vdef (stmt1);
if (vdef != NULL
&& (!single_imm_use (vdef, &use_p, &use_stmt)
|| use_stmt != stmt2))
break;
ptr1_align = get_pointer_alignment (ptr1);
/* Construct the new source string literal. */
src_buf = XALLOCAVEC (char, src_len + 1);
if (callee1)
memcpy (src_buf,
TREE_STRING_POINTER (str1) + tree_to_uhwi (off1),
tree_to_uhwi (len1));
else
src_buf[0] = tree_to_shwi (src1);
memset (src_buf + tree_to_uhwi (diff),
tree_to_shwi (val2), tree_to_uhwi (len2));
src_buf[src_len] = '\0';
/* Neither builtin_strncpy_read_str nor builtin_memcpy_read_str
handle embedded '\0's. */
if (strlen (src_buf) != src_len)
break;
rtl_profile_for_bb (gimple_bb (stmt2));
/* If the new memcpy wouldn't be emitted by storing the literal
by pieces, this optimization might enlarge .rodata too much,
as commonly used string literals couldn't be shared any
longer. */
if (!can_store_by_pieces (src_len,
builtin_strncpy_read_str,
src_buf, ptr1_align, false))
break;
new_str_cst = build_string_literal (src_len, src_buf);
if (callee1)
{
/* If STMT1 is a mem{,p}cpy call, adjust it and remove
memset call. */
if (lhs1 && DECL_FUNCTION_CODE (callee1) == BUILT_IN_MEMPCPY)
gimple_call_set_lhs (stmt1, NULL_TREE);
gimple_call_set_arg (stmt1, 1, new_str_cst);
gimple_call_set_arg (stmt1, 2,
build_int_cst (TREE_TYPE (len1), src_len));
update_stmt (stmt1);
unlink_stmt_vdef (stmt2);
gsi_replace (gsi_p, gimple_build_nop (), false);
fwprop_invalidate_lattice (gimple_get_lhs (stmt2));
release_defs (stmt2);
if (lhs1 && DECL_FUNCTION_CODE (callee1) == BUILT_IN_MEMPCPY)
{
fwprop_invalidate_lattice (lhs1);
release_ssa_name (lhs1);
}
return true;
}
else
{
/* Otherwise, if STMT1 is length 1 memcpy optimized into
assignment, remove STMT1 and change memset call into
memcpy call. */
gimple_stmt_iterator gsi = gsi_for_stmt (stmt1);
if (!is_gimple_val (ptr1))
ptr1 = force_gimple_operand_gsi (gsi_p, ptr1, true, NULL_TREE,
true, GSI_SAME_STMT);
tree fndecl = builtin_decl_explicit (BUILT_IN_MEMCPY);
gimple_call_set_fndecl (stmt2, fndecl);
gimple_call_set_fntype (as_a <gcall *> (stmt2),
TREE_TYPE (fndecl));
gimple_call_set_arg (stmt2, 0, ptr1);
gimple_call_set_arg (stmt2, 1, new_str_cst);
gimple_call_set_arg (stmt2, 2,
build_int_cst (TREE_TYPE (len2), src_len));
unlink_stmt_vdef (stmt1);
gsi_remove (&gsi, true);
fwprop_invalidate_lattice (gimple_get_lhs (stmt1));
release_defs (stmt1);
update_stmt (stmt2);
return false;
}
}
break;
#define CASE_ATOMIC(NAME, OTHER, OP) \
case BUILT_IN_##NAME##_1: \
case BUILT_IN_##NAME##_2: \
case BUILT_IN_##NAME##_4: \
case BUILT_IN_##NAME##_8: \
case BUILT_IN_##NAME##_16: \
atomic_op = OP; \
other_atomic \
= (enum built_in_function) (BUILT_IN_##OTHER##_1 \
+ (DECL_FUNCTION_CODE (callee2) \
- BUILT_IN_##NAME##_1)); \
goto handle_atomic_fetch_op;
CASE_ATOMIC (ATOMIC_FETCH_ADD, ATOMIC_ADD_FETCH, PLUS_EXPR)
CASE_ATOMIC (ATOMIC_FETCH_SUB, ATOMIC_SUB_FETCH, MINUS_EXPR)
CASE_ATOMIC (ATOMIC_FETCH_AND, ATOMIC_AND_FETCH, BIT_AND_EXPR)
CASE_ATOMIC (ATOMIC_FETCH_XOR, ATOMIC_XOR_FETCH, BIT_XOR_EXPR)
CASE_ATOMIC (ATOMIC_FETCH_OR, ATOMIC_OR_FETCH, BIT_IOR_EXPR)
CASE_ATOMIC (SYNC_FETCH_AND_ADD, SYNC_ADD_AND_FETCH, PLUS_EXPR)
CASE_ATOMIC (SYNC_FETCH_AND_SUB, SYNC_SUB_AND_FETCH, MINUS_EXPR)
CASE_ATOMIC (SYNC_FETCH_AND_AND, SYNC_AND_AND_FETCH, BIT_AND_EXPR)
CASE_ATOMIC (SYNC_FETCH_AND_XOR, SYNC_XOR_AND_FETCH, BIT_XOR_EXPR)
CASE_ATOMIC (SYNC_FETCH_AND_OR, SYNC_OR_AND_FETCH, BIT_IOR_EXPR)
CASE_ATOMIC (ATOMIC_ADD_FETCH, ATOMIC_FETCH_ADD, MINUS_EXPR)
CASE_ATOMIC (ATOMIC_SUB_FETCH, ATOMIC_FETCH_SUB, PLUS_EXPR)
CASE_ATOMIC (ATOMIC_XOR_FETCH, ATOMIC_FETCH_XOR, BIT_XOR_EXPR)
CASE_ATOMIC (SYNC_ADD_AND_FETCH, SYNC_FETCH_AND_ADD, MINUS_EXPR)
CASE_ATOMIC (SYNC_SUB_AND_FETCH, SYNC_FETCH_AND_SUB, PLUS_EXPR)
CASE_ATOMIC (SYNC_XOR_AND_FETCH, SYNC_FETCH_AND_XOR, BIT_XOR_EXPR)
#undef CASE_ATOMIC
handle_atomic_fetch_op:
if (gimple_call_num_args (stmt2) >= 2 && gimple_call_lhs (stmt2))
{
tree lhs2 = gimple_call_lhs (stmt2), lhsc = lhs2;
tree arg = gimple_call_arg (stmt2, 1);
gimple *use_stmt, *cast_stmt = NULL;
use_operand_p use_p;
tree ndecl = builtin_decl_explicit (other_atomic);
if (ndecl == NULL_TREE || !single_imm_use (lhs2, &use_p, &use_stmt))
break;
if (gimple_assign_cast_p (use_stmt))
{
cast_stmt = use_stmt;
lhsc = gimple_assign_lhs (cast_stmt);
if (lhsc == NULL_TREE
|| !INTEGRAL_TYPE_P (TREE_TYPE (lhsc))
|| (TYPE_PRECISION (TREE_TYPE (lhsc))
!= TYPE_PRECISION (TREE_TYPE (lhs2)))
|| !single_imm_use (lhsc, &use_p, &use_stmt))
{
use_stmt = cast_stmt;
cast_stmt = NULL;
lhsc = lhs2;
}
}
bool ok = false;
tree oarg = NULL_TREE;
enum tree_code ccode = ERROR_MARK;
tree crhs1 = NULL_TREE, crhs2 = NULL_TREE;
if (is_gimple_assign (use_stmt)
&& gimple_assign_rhs_code (use_stmt) == atomic_op)
{
if (gimple_assign_rhs1 (use_stmt) == lhsc)
oarg = gimple_assign_rhs2 (use_stmt);
else if (atomic_op != MINUS_EXPR)
oarg = gimple_assign_rhs1 (use_stmt);
}
else if (atomic_op == MINUS_EXPR
&& is_gimple_assign (use_stmt)
&& gimple_assign_rhs_code (use_stmt) == PLUS_EXPR
&& TREE_CODE (arg) == INTEGER_CST
&& (TREE_CODE (gimple_assign_rhs2 (use_stmt))
== INTEGER_CST))
{
tree a = fold_convert (TREE_TYPE (lhs2), arg);
tree o = fold_convert (TREE_TYPE (lhs2),
gimple_assign_rhs2 (use_stmt));
if (wi::to_wide (a) == wi::neg (wi::to_wide (o)))
ok = true;
}
else if (atomic_op == BIT_AND_EXPR || atomic_op == BIT_IOR_EXPR)
;
else if (gimple_code (use_stmt) == GIMPLE_COND)
{
ccode = gimple_cond_code (use_stmt);
crhs1 = gimple_cond_lhs (use_stmt);
crhs2 = gimple_cond_rhs (use_stmt);
}
else if (is_gimple_assign (use_stmt))
{
if (gimple_assign_rhs_class (use_stmt) == GIMPLE_BINARY_RHS)
{
ccode = gimple_assign_rhs_code (use_stmt);
crhs1 = gimple_assign_rhs1 (use_stmt);
crhs2 = gimple_assign_rhs2 (use_stmt);
}
else if (gimple_assign_rhs_code (use_stmt) == COND_EXPR)
{
tree cond = gimple_assign_rhs1 (use_stmt);
if (COMPARISON_CLASS_P (cond))
{
ccode = TREE_CODE (cond);
crhs1 = TREE_OPERAND (cond, 0);
crhs2 = TREE_OPERAND (cond, 1);
}
}
}
if (ccode == EQ_EXPR || ccode == NE_EXPR)
{
/* Deal with x - y == 0 or x ^ y == 0
being optimized into x == y and x + cst == 0
into x == -cst. */
tree o = NULL_TREE;
if (crhs1 == lhsc)
o = crhs2;
else if (crhs2 == lhsc)
o = crhs1;
if (o && atomic_op != PLUS_EXPR)
oarg = o;
else if (o
&& TREE_CODE (o) == INTEGER_CST
&& TREE_CODE (arg) == INTEGER_CST)
{
tree a = fold_convert (TREE_TYPE (lhs2), arg);
o = fold_convert (TREE_TYPE (lhs2), o);
if (wi::to_wide (a) == wi::neg (wi::to_wide (o)))
ok = true;
}
}
if (oarg && !ok)
{
if (operand_equal_p (arg, oarg, 0))
ok = true;
else if (TREE_CODE (arg) == SSA_NAME
&& TREE_CODE (oarg) == SSA_NAME)
{
tree oarg2 = oarg;
if (gimple_assign_cast_p (SSA_NAME_DEF_STMT (oarg)))
{
gimple *g = SSA_NAME_DEF_STMT (oarg);
oarg2 = gimple_assign_rhs1 (g);
if (TREE_CODE (oarg2) != SSA_NAME
|| !INTEGRAL_TYPE_P (TREE_TYPE (oarg2))
|| (TYPE_PRECISION (TREE_TYPE (oarg2))
!= TYPE_PRECISION (TREE_TYPE (oarg))))
oarg2 = oarg;
}
if (gimple_assign_cast_p (SSA_NAME_DEF_STMT (arg)))
{
gimple *g = SSA_NAME_DEF_STMT (arg);
tree rhs1 = gimple_assign_rhs1 (g);
/* Handle e.g.
x.0_1 = (long unsigned int) x_4(D);
_2 = __atomic_fetch_add_8 (&vlong, x.0_1, 0);
_3 = (long int) _2;
_7 = x_4(D) + _3; */
if (rhs1 == oarg || rhs1 == oarg2)
ok = true;
/* Handle e.g.
x.18_1 = (short unsigned int) x_5(D);
_2 = (int) x.18_1;
_3 = __atomic_fetch_xor_2 (&vshort, _2, 0);
_4 = (short int) _3;
_8 = x_5(D) ^ _4;
This happens only for char/short. */
else if (TREE_CODE (rhs1) == SSA_NAME
&& INTEGRAL_TYPE_P (TREE_TYPE (rhs1))
&& (TYPE_PRECISION (TREE_TYPE (rhs1))
== TYPE_PRECISION (TREE_TYPE (lhs2))))
{
g = SSA_NAME_DEF_STMT (rhs1);
if (gimple_assign_cast_p (g)
&& (gimple_assign_rhs1 (g) == oarg
|| gimple_assign_rhs1 (g) == oarg2))
ok = true;
}
}
if (!ok && arg == oarg2)
/* Handle e.g.
_1 = __sync_fetch_and_add_4 (&v, x_5(D));
_2 = (int) _1;
x.0_3 = (int) x_5(D);
_7 = _2 + x.0_3; */
ok = true;
}
}
if (ok)
{
tree new_lhs = make_ssa_name (TREE_TYPE (lhs2));
gimple_call_set_lhs (stmt2, new_lhs);
gimple_call_set_fndecl (stmt2, ndecl);
gimple_stmt_iterator gsi = gsi_for_stmt (use_stmt);
if (ccode == ERROR_MARK)
gimple_assign_set_rhs_with_ops (&gsi, cast_stmt
? NOP_EXPR : SSA_NAME,
new_lhs);
else
{
crhs1 = new_lhs;
crhs2 = build_zero_cst (TREE_TYPE (lhs2));
if (gimple_code (use_stmt) == GIMPLE_COND)
{
gcond *cond_stmt = as_a <gcond *> (use_stmt);
gimple_cond_set_lhs (cond_stmt, crhs1);
gimple_cond_set_rhs (cond_stmt, crhs2);
}
else if (gimple_assign_rhs_class (use_stmt)
== GIMPLE_BINARY_RHS)
{
gimple_assign_set_rhs1 (use_stmt, crhs1);
gimple_assign_set_rhs2 (use_stmt, crhs2);
}
else
{
gcc_checking_assert (gimple_assign_rhs_code (use_stmt)
== COND_EXPR);
tree cond = build2 (ccode, boolean_type_node,
crhs1, crhs2);
gimple_assign_set_rhs1 (use_stmt, cond);
}
}
update_stmt (use_stmt);
if (atomic_op != BIT_AND_EXPR
&& atomic_op != BIT_IOR_EXPR
&& !stmt_ends_bb_p (stmt2))
{
/* For the benefit of debug stmts, emit stmt(s) to set
lhs2 to the value it had from the new builtin.
E.g. if it was previously:
lhs2 = __atomic_fetch_add_8 (ptr, arg, 0);
emit:
new_lhs = __atomic_add_fetch_8 (ptr, arg, 0);
lhs2 = new_lhs - arg;
We also keep cast_stmt if any in the IL for
the same reasons.
These stmts will be DCEd later and proper debug info
will be emitted.
This is only possible for reversible operations
(+/-/^) and without -fnon-call-exceptions. */
gsi = gsi_for_stmt (stmt2);
tree type = TREE_TYPE (lhs2);
if (TREE_CODE (arg) == INTEGER_CST)
arg = fold_convert (type, arg);
else if (!useless_type_conversion_p (type, TREE_TYPE (arg)))
{
tree narg = make_ssa_name (type);
gimple *g = gimple_build_assign (narg, NOP_EXPR, arg);
gsi_insert_after (&gsi, g, GSI_NEW_STMT);
arg = narg;
}
enum tree_code rcode;
switch (atomic_op)
{
case PLUS_EXPR: rcode = MINUS_EXPR; break;
case MINUS_EXPR: rcode = PLUS_EXPR; break;
case BIT_XOR_EXPR: rcode = atomic_op; break;
default: gcc_unreachable ();
}
gimple *g = gimple_build_assign (lhs2, rcode, new_lhs, arg);
gsi_insert_after (&gsi, g, GSI_NEW_STMT);
update_stmt (stmt2);
}
else
{
/* For e.g.
lhs2 = __atomic_fetch_or_8 (ptr, arg, 0);
after we change it to
new_lhs = __atomic_or_fetch_8 (ptr, arg, 0);
there is no way to find out the lhs2 value (i.e.
what the atomic memory contained before the operation),
values of some bits are lost. We have checked earlier
that we don't have any non-debug users except for what
we are already changing, so we need to reset the
debug stmts and remove the cast_stmt if any. */
imm_use_iterator iter;
FOR_EACH_IMM_USE_STMT (use_stmt, iter, lhs2)
if (use_stmt != cast_stmt)
{
gcc_assert (is_gimple_debug (use_stmt));
gimple_debug_bind_reset_value (use_stmt);
update_stmt (use_stmt);
}
if (cast_stmt)
{
gsi = gsi_for_stmt (cast_stmt);
gsi_remove (&gsi, true);
}
update_stmt (stmt2);
release_ssa_name (lhs2);
}
}
}
break;
default:
break;
}
return false;
}
/* Given a ssa_name in NAME see if it was defined by an assignment and
set CODE to be the code and ARG1 to the first operand on the rhs and ARG2
to the second operand on the rhs. */
static inline void
defcodefor_name (tree name, enum tree_code *code, tree *arg1, tree *arg2)
{
gimple *def;
enum tree_code code1;
tree arg11;
tree arg21;
tree arg31;
enum gimple_rhs_class grhs_class;
code1 = TREE_CODE (name);
arg11 = name;
arg21 = NULL_TREE;
arg31 = NULL_TREE;
grhs_class = get_gimple_rhs_class (code1);
if (code1 == SSA_NAME)
{
def = SSA_NAME_DEF_STMT (name);
if (def && is_gimple_assign (def)
&& can_propagate_from (def))
{
code1 = gimple_assign_rhs_code (def);
arg11 = gimple_assign_rhs1 (def);
arg21 = gimple_assign_rhs2 (def);
arg31 = gimple_assign_rhs3 (def);
}
}
else if (grhs_class != GIMPLE_SINGLE_RHS)
code1 = ERROR_MARK;
*code = code1;
*arg1 = arg11;
if (arg2)
*arg2 = arg21;
if (arg31)
*code = ERROR_MARK;
}
/* Recognize rotation patterns. Return true if a transformation
applied, otherwise return false.
We are looking for X with unsigned type T with bitsize B, OP being
+, | or ^, some type T2 wider than T. For:
(X << CNT1) OP (X >> CNT2) iff CNT1 + CNT2 == B
((T) ((T2) X << CNT1)) OP ((T) ((T2) X >> CNT2)) iff CNT1 + CNT2 == B
transform these into:
X r<< CNT1
Or for:
(X << Y) OP (X >> (B - Y))
(X << (int) Y) OP (X >> (int) (B - Y))
((T) ((T2) X << Y)) OP ((T) ((T2) X >> (B - Y)))
((T) ((T2) X << (int) Y)) OP ((T) ((T2) X >> (int) (B - Y)))
(X << Y) | (X >> ((-Y) & (B - 1)))
(X << (int) Y) | (X >> (int) ((-Y) & (B - 1)))
((T) ((T2) X << Y)) | ((T) ((T2) X >> ((-Y) & (B - 1))))
((T) ((T2) X << (int) Y)) | ((T) ((T2) X >> (int) ((-Y) & (B - 1))))
transform these into (last 2 only if ranger can prove Y < B
or Y = N * B):
X r<< Y
or
X r<< (& & (B - 1))
The latter for the forms with T2 wider than T if ranger can't prove Y < B.
Or for:
(X << (Y & (B - 1))) | (X >> ((-Y) & (B - 1)))
(X << (int) (Y & (B - 1))) | (X >> (int) ((-Y) & (B - 1)))
((T) ((T2) X << (Y & (B - 1)))) | ((T) ((T2) X >> ((-Y) & (B - 1))))
((T) ((T2) X << (int) (Y & (B - 1)))) \
| ((T) ((T2) X >> (int) ((-Y) & (B - 1))))
transform these into:
X r<< (Y & (B - 1))
Note, in the patterns with T2 type, the type of OP operands
might be even a signed type, but should have precision B.
Expressions with & (B - 1) should be recognized only if B is
a power of 2. */
static bool
simplify_rotate (gimple_stmt_iterator *gsi)
{
gimple *stmt = gsi_stmt (*gsi);
tree arg[2], rtype, rotcnt = NULL_TREE;
tree def_arg1[2], def_arg2[2];
enum tree_code def_code[2];
tree lhs;
int i;
bool swapped_p = false;
gimple *g;
gimple *def_arg_stmt[2] = { NULL, NULL };
int wider_prec = 0;
bool add_masking = false;
arg[0] = gimple_assign_rhs1 (stmt);
arg[1] = gimple_assign_rhs2 (stmt);
rtype = TREE_TYPE (arg[0]);
/* Only create rotates in complete modes. Other cases are not
expanded properly. */
if (!INTEGRAL_TYPE_P (rtype)
|| !type_has_mode_precision_p (rtype))
return false;
for (i = 0; i < 2; i++)
{
defcodefor_name (arg[i], &def_code[i], &def_arg1[i], &def_arg2[i]);
if (TREE_CODE (arg[i]) == SSA_NAME)
def_arg_stmt[i] = SSA_NAME_DEF_STMT (arg[i]);
}
/* Look through narrowing (or same precision) conversions. */
if (CONVERT_EXPR_CODE_P (def_code[0])
&& CONVERT_EXPR_CODE_P (def_code[1])
&& INTEGRAL_TYPE_P (TREE_TYPE (def_arg1[0]))
&& INTEGRAL_TYPE_P (TREE_TYPE (def_arg1[1]))
&& TYPE_PRECISION (TREE_TYPE (def_arg1[0]))
== TYPE_PRECISION (TREE_TYPE (def_arg1[1]))
&& TYPE_PRECISION (TREE_TYPE (def_arg1[0])) >= TYPE_PRECISION (rtype)
&& has_single_use (arg[0])
&& has_single_use (arg[1]))
{
wider_prec = TYPE_PRECISION (TREE_TYPE (def_arg1[0]));
for (i = 0; i < 2; i++)
{
arg[i] = def_arg1[i];
defcodefor_name (arg[i], &def_code[i], &def_arg1[i], &def_arg2[i]);
if (TREE_CODE (arg[i]) == SSA_NAME)
def_arg_stmt[i] = SSA_NAME_DEF_STMT (arg[i]);
}
}
else
{
/* Handle signed rotate; the RSHIFT_EXPR has to be done
in unsigned type but LSHIFT_EXPR could be signed. */
i = (def_code[0] == LSHIFT_EXPR || def_code[0] == RSHIFT_EXPR);
if (CONVERT_EXPR_CODE_P (def_code[i])
&& (def_code[1 - i] == LSHIFT_EXPR || def_code[1 - i] == RSHIFT_EXPR)
&& INTEGRAL_TYPE_P (TREE_TYPE (def_arg1[i]))
&& TYPE_PRECISION (rtype) == TYPE_PRECISION (TREE_TYPE (def_arg1[i]))
&& has_single_use (arg[i]))
{
arg[i] = def_arg1[i];
defcodefor_name (arg[i], &def_code[i], &def_arg1[i], &def_arg2[i]);
if (TREE_CODE (arg[i]) == SSA_NAME)
def_arg_stmt[i] = SSA_NAME_DEF_STMT (arg[i]);
}
}
/* One operand has to be LSHIFT_EXPR and one RSHIFT_EXPR. */
for (i = 0; i < 2; i++)
if (def_code[i] != LSHIFT_EXPR && def_code[i] != RSHIFT_EXPR)
return false;
else if (!has_single_use (arg[i]))
return false;
if (def_code[0] == def_code[1])
return false;
/* If we've looked through narrowing conversions before, look through
widening conversions from unsigned type with the same precision
as rtype here. */
if (TYPE_PRECISION (TREE_TYPE (def_arg1[0])) != TYPE_PRECISION (rtype))
for (i = 0; i < 2; i++)
{
tree tem;
enum tree_code code;
defcodefor_name (def_arg1[i], &code, &tem, NULL);
if (!CONVERT_EXPR_CODE_P (code)
|| !INTEGRAL_TYPE_P (TREE_TYPE (tem))
|| TYPE_PRECISION (TREE_TYPE (tem)) != TYPE_PRECISION (rtype))
return false;
def_arg1[i] = tem;
}
/* Both shifts have to use the same first operand. */
if (!operand_equal_for_phi_arg_p (def_arg1[0], def_arg1[1])
|| !types_compatible_p (TREE_TYPE (def_arg1[0]),
TREE_TYPE (def_arg1[1])))
{
if ((TYPE_PRECISION (TREE_TYPE (def_arg1[0]))
!= TYPE_PRECISION (TREE_TYPE (def_arg1[1])))
|| (TYPE_UNSIGNED (TREE_TYPE (def_arg1[0]))
== TYPE_UNSIGNED (TREE_TYPE (def_arg1[1]))))
return false;
/* Handle signed rotate; the RSHIFT_EXPR has to be done
in unsigned type but LSHIFT_EXPR could be signed. */
i = def_code[0] != RSHIFT_EXPR;
if (!TYPE_UNSIGNED (TREE_TYPE (def_arg1[i])))
return false;
tree tem;
enum tree_code code;
defcodefor_name (def_arg1[i], &code, &tem, NULL);
if (!CONVERT_EXPR_CODE_P (code)
|| !INTEGRAL_TYPE_P (TREE_TYPE (tem))
|| TYPE_PRECISION (TREE_TYPE (tem)) != TYPE_PRECISION (rtype))
return false;
def_arg1[i] = tem;
if (!operand_equal_for_phi_arg_p (def_arg1[0], def_arg1[1])
|| !types_compatible_p (TREE_TYPE (def_arg1[0]),
TREE_TYPE (def_arg1[1])))
return false;
}
else if (!TYPE_UNSIGNED (TREE_TYPE (def_arg1[0])))
return false;
/* CNT1 + CNT2 == B case above. */
if (tree_fits_uhwi_p (def_arg2[0])
&& tree_fits_uhwi_p (def_arg2[1])
&& tree_to_uhwi (def_arg2[0])
+ tree_to_uhwi (def_arg2[1]) == TYPE_PRECISION (rtype))
rotcnt = def_arg2[0];
else if (TREE_CODE (def_arg2[0]) != SSA_NAME
|| TREE_CODE (def_arg2[1]) != SSA_NAME)
return false;
else
{
tree cdef_arg1[2], cdef_arg2[2], def_arg2_alt[2];
enum tree_code cdef_code[2];
gimple *def_arg_alt_stmt[2] = { NULL, NULL };
int check_range = 0;
gimple *check_range_stmt = NULL;
/* Look through conversion of the shift count argument.
The C/C++ FE cast any shift count argument to integer_type_node.
The only problem might be if the shift count type maximum value
is equal or smaller than number of bits in rtype. */
for (i = 0; i < 2; i++)
{
def_arg2_alt[i] = def_arg2[i];
defcodefor_name (def_arg2[i], &cdef_code[i],
&cdef_arg1[i], &cdef_arg2[i]);
if (CONVERT_EXPR_CODE_P (cdef_code[i])
&& INTEGRAL_TYPE_P (TREE_TYPE (cdef_arg1[i]))
&& TYPE_PRECISION (TREE_TYPE (cdef_arg1[i]))
> floor_log2 (TYPE_PRECISION (rtype))
&& type_has_mode_precision_p (TREE_TYPE (cdef_arg1[i])))
{
def_arg2_alt[i] = cdef_arg1[i];
if (TREE_CODE (def_arg2[i]) == SSA_NAME)
def_arg_alt_stmt[i] = SSA_NAME_DEF_STMT (def_arg2[i]);
defcodefor_name (def_arg2_alt[i], &cdef_code[i],
&cdef_arg1[i], &cdef_arg2[i]);
}
else
def_arg_alt_stmt[i] = def_arg_stmt[i];
}
for (i = 0; i < 2; i++)
/* Check for one shift count being Y and the other B - Y,
with optional casts. */
if (cdef_code[i] == MINUS_EXPR
&& tree_fits_shwi_p (cdef_arg1[i])
&& tree_to_shwi (cdef_arg1[i]) == TYPE_PRECISION (rtype)
&& TREE_CODE (cdef_arg2[i]) == SSA_NAME)
{
tree tem;
enum tree_code code;
if (cdef_arg2[i] == def_arg2[1 - i]
|| cdef_arg2[i] == def_arg2_alt[1 - i])
{
rotcnt = cdef_arg2[i];
check_range = -1;
if (cdef_arg2[i] == def_arg2[1 - i])
check_range_stmt = def_arg_stmt[1 - i];
else
check_range_stmt = def_arg_alt_stmt[1 - i];
break;
}
defcodefor_name (cdef_arg2[i], &code, &tem, NULL);
if (CONVERT_EXPR_CODE_P (code)
&& INTEGRAL_TYPE_P (TREE_TYPE (tem))
&& TYPE_PRECISION (TREE_TYPE (tem))
> floor_log2 (TYPE_PRECISION (rtype))
&& type_has_mode_precision_p (TREE_TYPE (tem))
&& (tem == def_arg2[1 - i]
|| tem == def_arg2_alt[1 - i]))
{
rotcnt = tem;
check_range = -1;
if (tem == def_arg2[1 - i])
check_range_stmt = def_arg_stmt[1 - i];
else
check_range_stmt = def_arg_alt_stmt[1 - i];
break;
}
}
/* The above sequence isn't safe for Y being 0,
because then one of the shifts triggers undefined behavior.
This alternative is safe even for rotation count of 0.
One shift count is Y and the other (-Y) & (B - 1).
Or one shift count is Y & (B - 1) and the other (-Y) & (B - 1). */
else if (cdef_code[i] == BIT_AND_EXPR
&& pow2p_hwi (TYPE_PRECISION (rtype))
&& tree_fits_shwi_p (cdef_arg2[i])
&& tree_to_shwi (cdef_arg2[i])
== TYPE_PRECISION (rtype) - 1
&& TREE_CODE (cdef_arg1[i]) == SSA_NAME
&& gimple_assign_rhs_code (stmt) == BIT_IOR_EXPR)
{
tree tem;
enum tree_code code;
defcodefor_name (cdef_arg1[i], &code, &tem, NULL);
if (CONVERT_EXPR_CODE_P (code)
&& INTEGRAL_TYPE_P (TREE_TYPE (tem))
&& TYPE_PRECISION (TREE_TYPE (tem))
> floor_log2 (TYPE_PRECISION (rtype))
&& type_has_mode_precision_p (TREE_TYPE (tem)))
defcodefor_name (tem, &code, &tem, NULL);
if (code == NEGATE_EXPR)
{
if (tem == def_arg2[1 - i] || tem == def_arg2_alt[1 - i])
{
rotcnt = tem;
check_range = 1;
if (tem == def_arg2[1 - i])
check_range_stmt = def_arg_stmt[1 - i];
else
check_range_stmt = def_arg_alt_stmt[1 - i];
break;
}
tree tem2;
defcodefor_name (tem, &code, &tem2, NULL);
if (CONVERT_EXPR_CODE_P (code)
&& INTEGRAL_TYPE_P (TREE_TYPE (tem2))
&& TYPE_PRECISION (TREE_TYPE (tem2))
> floor_log2 (TYPE_PRECISION (rtype))
&& type_has_mode_precision_p (TREE_TYPE (tem2)))
{
if (tem2 == def_arg2[1 - i]
|| tem2 == def_arg2_alt[1 - i])
{
rotcnt = tem2;
check_range = 1;
if (tem2 == def_arg2[1 - i])
check_range_stmt = def_arg_stmt[1 - i];
else
check_range_stmt = def_arg_alt_stmt[1 - i];
break;
}
}
else
tem2 = NULL_TREE;
if (cdef_code[1 - i] == BIT_AND_EXPR
&& tree_fits_shwi_p (cdef_arg2[1 - i])
&& tree_to_shwi (cdef_arg2[1 - i])
== TYPE_PRECISION (rtype) - 1
&& TREE_CODE (cdef_arg1[1 - i]) == SSA_NAME)
{
if (tem == cdef_arg1[1 - i]
|| tem2 == cdef_arg1[1 - i])
{
rotcnt = def_arg2[1 - i];
break;
}
tree tem3;
defcodefor_name (cdef_arg1[1 - i], &code, &tem3, NULL);
if (CONVERT_EXPR_CODE_P (code)
&& INTEGRAL_TYPE_P (TREE_TYPE (tem3))
&& TYPE_PRECISION (TREE_TYPE (tem3))
> floor_log2 (TYPE_PRECISION (rtype))
&& type_has_mode_precision_p (TREE_TYPE (tem3)))
{
if (tem == tem3 || tem2 == tem3)
{
rotcnt = def_arg2[1 - i];
break;
}
}
}
}
}
if (check_range && wider_prec > TYPE_PRECISION (rtype))
{
if (TREE_CODE (rotcnt) != SSA_NAME)
return false;
int_range_max r;
range_query *q = get_range_query (cfun);
if (q == get_global_range_query ())
q = enable_ranger (cfun);
if (!q->range_of_expr (r, rotcnt, check_range_stmt))
{
if (check_range > 0)
return false;
r.set_varying (TREE_TYPE (rotcnt));
}
int prec = TYPE_PRECISION (TREE_TYPE (rotcnt));
signop sign = TYPE_SIGN (TREE_TYPE (rotcnt));
wide_int min = wide_int::from (TYPE_PRECISION (rtype), prec, sign);
wide_int max = wide_int::from (wider_prec - 1, prec, sign);
if (check_range < 0)
max = min;
int_range<1> r2 (TREE_TYPE (rotcnt), min, max);
r.intersect (r2);
if (!r.undefined_p ())
{
if (check_range > 0)
{
int_range_max r3;
for (int i = TYPE_PRECISION (rtype) + 1; i < wider_prec;
i += TYPE_PRECISION (rtype))
{
int j = i + TYPE_PRECISION (rtype) - 2;
min = wide_int::from (i, prec, sign);
max = wide_int::from (MIN (j, wider_prec - 1),
prec, sign);
int_range<1> r4 (TREE_TYPE (rotcnt), min, max);
r3.union_ (r4);
}
r.intersect (r3);
if (!r.undefined_p ())
return false;
}
add_masking = true;
}
}
if (rotcnt == NULL_TREE)
return false;
swapped_p = i != 1;
}
if (!useless_type_conversion_p (TREE_TYPE (def_arg2[0]),
TREE_TYPE (rotcnt)))
{
g = gimple_build_assign (make_ssa_name (TREE_TYPE (def_arg2[0])),
NOP_EXPR, rotcnt);
gsi_insert_before (gsi, g, GSI_SAME_STMT);
rotcnt = gimple_assign_lhs (g);
}
if (add_masking)
{
g = gimple_build_assign (make_ssa_name (TREE_TYPE (rotcnt)),
BIT_AND_EXPR, rotcnt,
build_int_cst (TREE_TYPE (rotcnt),
TYPE_PRECISION (rtype) - 1));
gsi_insert_before (gsi, g, GSI_SAME_STMT);
rotcnt = gimple_assign_lhs (g);
}
lhs = gimple_assign_lhs (stmt);
if (!useless_type_conversion_p (rtype, TREE_TYPE (def_arg1[0])))
lhs = make_ssa_name (TREE_TYPE (def_arg1[0]));
g = gimple_build_assign (lhs,
((def_code[0] == LSHIFT_EXPR) ^ swapped_p)
? LROTATE_EXPR : RROTATE_EXPR, def_arg1[0], rotcnt);
if (!useless_type_conversion_p (rtype, TREE_TYPE (def_arg1[0])))
{
gsi_insert_before (gsi, g, GSI_SAME_STMT);
g = gimple_build_assign (gimple_assign_lhs (stmt), NOP_EXPR, lhs);
}
gsi_replace (gsi, g, false);
return true;
}
/* Check whether an array contains a valid ctz table. */
static bool
check_ctz_array (tree ctor, unsigned HOST_WIDE_INT mulc,
HOST_WIDE_INT &zero_val, unsigned shift, unsigned bits)
{
tree elt, idx;
unsigned HOST_WIDE_INT i, mask;
unsigned matched = 0;
mask = ((HOST_WIDE_INT_1U << (bits - shift)) - 1) << shift;
zero_val = 0;
FOR_EACH_CONSTRUCTOR_ELT (CONSTRUCTOR_ELTS (ctor), i, idx, elt)
{
if (TREE_CODE (idx) != INTEGER_CST || TREE_CODE (elt) != INTEGER_CST)
return false;
if (i > bits * 2)
return false;
unsigned HOST_WIDE_INT index = tree_to_shwi (idx);
HOST_WIDE_INT val = tree_to_shwi (elt);
if (index == 0)
{
zero_val = val;
matched++;
}
if (val >= 0 && val < bits && (((mulc << val) & mask) >> shift) == index)
matched++;
if (matched > bits)
return true;
}
return false;
}
/* Check whether a string contains a valid ctz table. */
static bool
check_ctz_string (tree string, unsigned HOST_WIDE_INT mulc,
HOST_WIDE_INT &zero_val, unsigned shift, unsigned bits)
{
unsigned HOST_WIDE_INT len = TREE_STRING_LENGTH (string);
unsigned HOST_WIDE_INT mask;
unsigned matched = 0;
const unsigned char *p = (const unsigned char *) TREE_STRING_POINTER (string);
if (len < bits || len > bits * 2)
return false;
mask = ((HOST_WIDE_INT_1U << (bits - shift)) - 1) << shift;
zero_val = p[0];
for (unsigned i = 0; i < len; i++)
if (p[i] < bits && (((mulc << p[i]) & mask) >> shift) == i)
matched++;
return matched == bits;
}
/* Recognize count trailing zeroes idiom.
The canonical form is array[((x & -x) * C) >> SHIFT] where C is a magic
constant which when multiplied by a power of 2 creates a unique value
in the top 5 or 6 bits. This is then indexed into a table which maps it
to the number of trailing zeroes. Array[0] is returned so the caller can
emit an appropriate sequence depending on whether ctz (0) is defined on
the target. */
static bool
optimize_count_trailing_zeroes (tree array_ref, tree x, tree mulc,
tree tshift, HOST_WIDE_INT &zero_val)
{
tree type = TREE_TYPE (array_ref);
tree array = TREE_OPERAND (array_ref, 0);
gcc_assert (TREE_CODE (mulc) == INTEGER_CST);
gcc_assert (TREE_CODE (tshift) == INTEGER_CST);
tree input_type = TREE_TYPE (x);
unsigned input_bits = tree_to_shwi (TYPE_SIZE (input_type));
/* Check the array element type is not wider than 32 bits and the input is
an unsigned 32-bit or 64-bit type. */
if (TYPE_PRECISION (type) > 32 || !TYPE_UNSIGNED (input_type))
return false;
if (input_bits != 32 && input_bits != 64)
return false;
if (!direct_internal_fn_supported_p (IFN_CTZ, input_type, OPTIMIZE_FOR_BOTH))
return false;
/* Check the lower bound of the array is zero. */
tree low = array_ref_low_bound (array_ref);
if (!low || !integer_zerop (low))
return false;
unsigned shiftval = tree_to_shwi (tshift);
/* Check the shift extracts the top 5..7 bits. */
if (shiftval < input_bits - 7 || shiftval > input_bits - 5)
return false;
tree ctor = ctor_for_folding (array);
if (!ctor)
return false;
unsigned HOST_WIDE_INT val = tree_to_uhwi (mulc);
if (TREE_CODE (ctor) == CONSTRUCTOR)
return check_ctz_array (ctor, val, zero_val, shiftval, input_bits);
if (TREE_CODE (ctor) == STRING_CST
&& TYPE_PRECISION (type) == CHAR_TYPE_SIZE)
return check_ctz_string (ctor, val, zero_val, shiftval, input_bits);
return false;
}
/* Match.pd function to match the ctz expression. */
extern bool gimple_ctz_table_index (tree, tree *, tree (*)(tree));
static bool
simplify_count_trailing_zeroes (gimple_stmt_iterator *gsi)
{
gimple *stmt = gsi_stmt (*gsi);
tree array_ref = gimple_assign_rhs1 (stmt);
tree res_ops[3];
HOST_WIDE_INT zero_val;
gcc_checking_assert (TREE_CODE (array_ref) == ARRAY_REF);
if (!gimple_ctz_table_index (TREE_OPERAND (array_ref, 1), &res_ops[0], NULL))
return false;
if (optimize_count_trailing_zeroes (array_ref, res_ops[0],
res_ops[1], res_ops[2], zero_val))
{
tree type = TREE_TYPE (res_ops[0]);
HOST_WIDE_INT ctz_val = 0;
HOST_WIDE_INT type_size = tree_to_shwi (TYPE_SIZE (type));
bool zero_ok
= CTZ_DEFINED_VALUE_AT_ZERO (SCALAR_INT_TYPE_MODE (type), ctz_val) == 2;
int nargs = 2;
/* If the input value can't be zero, don't special case ctz (0). */
if (tree_expr_nonzero_p (res_ops[0]))
{
zero_ok = true;
zero_val = 0;
ctz_val = 0;
nargs = 1;
}
/* Skip if there is no value defined at zero, or if we can't easily
return the correct value for zero. */
if (!zero_ok)
return false;
if (zero_val != ctz_val && !(zero_val == 0 && ctz_val == type_size))
return false;
gimple_seq seq = NULL;
gimple *g;
gcall *call
= gimple_build_call_internal (IFN_CTZ, nargs, res_ops[0],
nargs == 1 ? NULL_TREE
: build_int_cst (integer_type_node,
ctz_val));
gimple_set_location (call, gimple_location (stmt));
gimple_set_lhs (call, make_ssa_name (integer_type_node));
gimple_seq_add_stmt (&seq, call);
tree prev_lhs = gimple_call_lhs (call);
/* Emit ctz (x) & 31 if ctz (0) is 32 but we need to return 0. */
if (zero_val == 0 && ctz_val == type_size)
{
g = gimple_build_assign (make_ssa_name (integer_type_node),
BIT_AND_EXPR, prev_lhs,
build_int_cst (integer_type_node,
type_size - 1));
gimple_set_location (g, gimple_location (stmt));
gimple_seq_add_stmt (&seq, g);
prev_lhs = gimple_assign_lhs (g);
}
g = gimple_build_assign (gimple_assign_lhs (stmt), NOP_EXPR, prev_lhs);
gimple_seq_add_stmt (&seq, g);
gsi_replace_with_seq (gsi, seq, true);
return true;
}
return false;
}
/* Combine an element access with a shuffle. Returns true if there were
any changes made, else it returns false. */
static bool
simplify_bitfield_ref (gimple_stmt_iterator *gsi)
{
gimple *stmt = gsi_stmt (*gsi);
gimple *def_stmt;
tree op, op0, op1;
tree elem_type, type;
tree p, m, tem;
unsigned HOST_WIDE_INT nelts, idx;
poly_uint64 size, elem_size;
enum tree_code code;
op = gimple_assign_rhs1 (stmt);
gcc_checking_assert (TREE_CODE (op) == BIT_FIELD_REF);
op0 = TREE_OPERAND (op, 0);
if (TREE_CODE (op0) != SSA_NAME
|| TREE_CODE (TREE_TYPE (op0)) != VECTOR_TYPE)
return false;
def_stmt = get_prop_source_stmt (op0, false, NULL);
if (!def_stmt || !can_propagate_from (def_stmt))
return false;
op1 = TREE_OPERAND (op, 1);
code = gimple_assign_rhs_code (def_stmt);
elem_type = TREE_TYPE (TREE_TYPE (op0));
type = TREE_TYPE (op);
/* Also handle vector type.
.i.e.
_7 = VEC_PERM_EXPR <_1, _1, { 2, 3, 2, 3 }>;
_11 = BIT_FIELD_REF <_7, 64, 0>;
to
_11 = BIT_FIELD_REF <_1, 64, 64>. */
size = tree_to_poly_uint64 (TYPE_SIZE (type));
if (maybe_ne (bit_field_size (op), size))
return false;
elem_size = tree_to_poly_uint64 (TYPE_SIZE (elem_type));
if (code != VEC_PERM_EXPR
|| !constant_multiple_p (bit_field_offset (op), elem_size, &idx))
return false;
m = gimple_assign_rhs3 (def_stmt);
if (TREE_CODE (m) != VECTOR_CST
|| !VECTOR_CST_NELTS (m).is_constant (&nelts))
return false;
/* One element. */
if (known_eq (size, elem_size))
idx = TREE_INT_CST_LOW (VECTOR_CST_ELT (m, idx)) % (2 * nelts);
else
{
unsigned HOST_WIDE_INT nelts_op;
if (!constant_multiple_p (size, elem_size, &nelts_op)
|| !pow2p_hwi (nelts_op))
return false;
/* Clamp vec_perm_expr index. */
unsigned start = TREE_INT_CST_LOW (vector_cst_elt (m, idx)) % (2 * nelts);
unsigned end = TREE_INT_CST_LOW (vector_cst_elt (m, idx + nelts_op - 1))
% (2 * nelts);
/* Be in the same vector. */
if ((start < nelts) != (end < nelts))
return false;
for (unsigned HOST_WIDE_INT i = 1; i != nelts_op; i++)
{
/* Continuous area. */
if (TREE_INT_CST_LOW (vector_cst_elt (m, idx + i)) % (2 * nelts) - 1
!= TREE_INT_CST_LOW (vector_cst_elt (m, idx + i - 1))
% (2 * nelts))
return false;
}
/* Alignment not worse than before. */
if (start % nelts_op)
return false;
idx = start;
}
if (idx < nelts)
p = gimple_assign_rhs1 (def_stmt);
else
{
p = gimple_assign_rhs2 (def_stmt);
idx -= nelts;
}
tem = build3 (BIT_FIELD_REF, TREE_TYPE (op),
p, op1, bitsize_int (idx * elem_size));
gimple_assign_set_rhs1 (stmt, tem);
fold_stmt (gsi);
update_stmt (gsi_stmt (*gsi));
return true;
}
/* Determine whether applying the 2 permutations (mask1 then mask2)
gives back one of the input. */
static int
is_combined_permutation_identity (tree mask1, tree mask2)
{
tree mask;
unsigned HOST_WIDE_INT nelts, i, j;
bool maybe_identity1 = true;
bool maybe_identity2 = true;
gcc_checking_assert (TREE_CODE (mask1) == VECTOR_CST
&& TREE_CODE (mask2) == VECTOR_CST);
/* For VLA masks, check for the following pattern:
v1 = VEC_PERM_EXPR (v0, ..., mask1)
v2 = VEC_PERM_EXPR (v1, ..., mask2)
-->
v2 = v0
if mask1 == mask2 == {nelts - 1, nelts - 2, ...}. */
if (operand_equal_p (mask1, mask2, 0)
&& !VECTOR_CST_NELTS (mask1).is_constant ())
{
vec_perm_builder builder;
if (tree_to_vec_perm_builder (&builder, mask1))
{
poly_uint64 nelts = TYPE_VECTOR_SUBPARTS (TREE_TYPE (mask1));
vec_perm_indices sel (builder, 1, nelts);
if (sel.series_p (0, 1, nelts - 1, -1))
return 1;
}
}
mask = fold_ternary (VEC_PERM_EXPR, TREE_TYPE (mask1), mask1, mask1, mask2);
if (mask == NULL_TREE || TREE_CODE (mask) != VECTOR_CST)
return 0;
if (!VECTOR_CST_NELTS (mask).is_constant (&nelts))
return 0;
for (i = 0; i < nelts; i++)
{
tree val = VECTOR_CST_ELT (mask, i);
gcc_assert (TREE_CODE (val) == INTEGER_CST);
j = TREE_INT_CST_LOW (val) & (2 * nelts - 1);
if (j == i)
maybe_identity2 = false;
else if (j == i + nelts)
maybe_identity1 = false;
else
return 0;
}
return maybe_identity1 ? 1 : maybe_identity2 ? 2 : 0;
}
/* Combine a shuffle with its arguments. Returns 1 if there were any
changes made, 2 if cfg-cleanup needs to run. Else it returns 0. */
static int
simplify_permutation (gimple_stmt_iterator *gsi)
{
gimple *stmt = gsi_stmt (*gsi);
gimple *def_stmt = NULL;
tree op0, op1, op2, op3, arg0, arg1;
enum tree_code code, code2 = ERROR_MARK;
bool single_use_op0 = false;
gcc_checking_assert (gimple_assign_rhs_code (stmt) == VEC_PERM_EXPR);
op0 = gimple_assign_rhs1 (stmt);
op1 = gimple_assign_rhs2 (stmt);
op2 = gimple_assign_rhs3 (stmt);
if (TREE_CODE (op2) != VECTOR_CST)
return 0;
if (TREE_CODE (op0) == VECTOR_CST)
{
code = VECTOR_CST;
arg0 = op0;
}
else if (TREE_CODE (op0) == SSA_NAME)
{
def_stmt = get_prop_source_stmt (op0, false, &single_use_op0);
if (!def_stmt)
return 0;
code = gimple_assign_rhs_code (def_stmt);
if (code == VIEW_CONVERT_EXPR)
{
tree rhs = gimple_assign_rhs1 (def_stmt);
tree name = TREE_OPERAND (rhs, 0);
if (TREE_CODE (name) != SSA_NAME)
return 0;
if (!has_single_use (name))
single_use_op0 = false;
/* Here we update the def_stmt through this VIEW_CONVERT_EXPR,
but still keep the code to indicate it comes from
VIEW_CONVERT_EXPR. */
def_stmt = SSA_NAME_DEF_STMT (name);
if (!def_stmt || !is_gimple_assign (def_stmt))
return 0;
if (gimple_assign_rhs_code (def_stmt) != CONSTRUCTOR)
return 0;
}
if (!can_propagate_from (def_stmt))
return 0;
arg0 = gimple_assign_rhs1 (def_stmt);
}
else
return 0;
/* Two consecutive shuffles. */
if (code == VEC_PERM_EXPR)
{
tree orig;
int ident;
if (op0 != op1)
return 0;
op3 = gimple_assign_rhs3 (def_stmt);
if (TREE_CODE (op3) != VECTOR_CST)
return 0;
ident = is_combined_permutation_identity (op3, op2);
if (!ident)
return 0;
orig = (ident == 1) ? gimple_assign_rhs1 (def_stmt)
: gimple_assign_rhs2 (def_stmt);
gimple_assign_set_rhs1 (stmt, unshare_expr (orig));
gimple_assign_set_rhs_code (stmt, TREE_CODE (orig));
gimple_set_num_ops (stmt, 2);
update_stmt (stmt);
return remove_prop_source_from_use (op0) ? 2 : 1;
}
else if (code == CONSTRUCTOR
|| code == VECTOR_CST
|| code == VIEW_CONVERT_EXPR)
{
if (op0 != op1)
{
if (TREE_CODE (op0) == SSA_NAME && !single_use_op0)
return 0;
if (TREE_CODE (op1) == VECTOR_CST)
arg1 = op1;
else if (TREE_CODE (op1) == SSA_NAME)
{
gimple *def_stmt2 = get_prop_source_stmt (op1, true, NULL);
if (!def_stmt2)
return 0;
code2 = gimple_assign_rhs_code (def_stmt2);
if (code2 == VIEW_CONVERT_EXPR)
{
tree rhs = gimple_assign_rhs1 (def_stmt2);
tree name = TREE_OPERAND (rhs, 0);
if (TREE_CODE (name) != SSA_NAME)
return 0;
if (!has_single_use (name))
return 0;
def_stmt2 = SSA_NAME_DEF_STMT (name);
if (!def_stmt2 || !is_gimple_assign (def_stmt2))
return 0;
if (gimple_assign_rhs_code (def_stmt2) != CONSTRUCTOR)
return 0;
}
else if (code2 != CONSTRUCTOR && code2 != VECTOR_CST)
return 0;
if (!can_propagate_from (def_stmt2))
return 0;
arg1 = gimple_assign_rhs1 (def_stmt2);
}
else
return 0;
}
else
{
/* Already used twice in this statement. */
if (TREE_CODE (op0) == SSA_NAME && num_imm_uses (op0) > 2)
return 0;
arg1 = arg0;
}
/* If there are any VIEW_CONVERT_EXPRs found when finding permutation
operands source, check whether it's valid to transform and prepare
the required new operands. */
if (code == VIEW_CONVERT_EXPR || code2 == VIEW_CONVERT_EXPR)
{
/* Figure out the target vector type to which operands should be
converted. If both are CONSTRUCTOR, the types should be the
same, otherwise, use the one of CONSTRUCTOR. */
tree tgt_type = NULL_TREE;
if (code == VIEW_CONVERT_EXPR)
{
gcc_assert (gimple_assign_rhs_code (def_stmt) == CONSTRUCTOR);
code = CONSTRUCTOR;
tgt_type = TREE_TYPE (arg0);
}
if (code2 == VIEW_CONVERT_EXPR)
{
tree arg1_type = TREE_TYPE (arg1);
if (tgt_type == NULL_TREE)
tgt_type = arg1_type;
else if (tgt_type != arg1_type)
return 0;
}
if (!VECTOR_TYPE_P (tgt_type))
return 0;
tree op2_type = TREE_TYPE (op2);
/* Figure out the shrunk factor. */
poly_uint64 tgt_units = TYPE_VECTOR_SUBPARTS (tgt_type);
poly_uint64 op2_units = TYPE_VECTOR_SUBPARTS (op2_type);
if (maybe_gt (tgt_units, op2_units))
return 0;
unsigned int factor;
if (!constant_multiple_p (op2_units, tgt_units, &factor))
return 0;
/* Build the new permutation control vector as target vector. */
vec_perm_builder builder;
if (!tree_to_vec_perm_builder (&builder, op2))
return 0;
vec_perm_indices indices (builder, 2, op2_units);
vec_perm_indices new_indices;
if (new_indices.new_shrunk_vector (indices, factor))
{
tree mask_type = tgt_type;
if (!VECTOR_INTEGER_TYPE_P (mask_type))
{
tree elem_type = TREE_TYPE (mask_type);
unsigned elem_size = TREE_INT_CST_LOW (TYPE_SIZE (elem_type));
tree int_type = build_nonstandard_integer_type (elem_size, 0);
mask_type = build_vector_type (int_type, tgt_units);
}
op2 = vec_perm_indices_to_tree (mask_type, new_indices);
}
else
return 0;
/* Convert the VECTOR_CST to the appropriate vector type. */
if (tgt_type != TREE_TYPE (arg0))
arg0 = fold_build1 (VIEW_CONVERT_EXPR, tgt_type, arg0);
else if (tgt_type != TREE_TYPE (arg1))
arg1 = fold_build1 (VIEW_CONVERT_EXPR, tgt_type, arg1);
}
/* VIEW_CONVERT_EXPR should be updated to CONSTRUCTOR before. */
gcc_assert (code == CONSTRUCTOR || code == VECTOR_CST);
/* Shuffle of a constructor. */
bool ret = false;
tree res_type
= build_vector_type (TREE_TYPE (TREE_TYPE (arg0)),
TYPE_VECTOR_SUBPARTS (TREE_TYPE (op2)));
tree opt = fold_ternary (VEC_PERM_EXPR, res_type, arg0, arg1, op2);
if (!opt
|| (TREE_CODE (opt) != CONSTRUCTOR && TREE_CODE (opt) != VECTOR_CST))
return 0;
/* Found VIEW_CONVERT_EXPR before, need one explicit conversion. */
if (res_type != TREE_TYPE (op0))
{
tree name = make_ssa_name (TREE_TYPE (opt));
gimple *ass_stmt = gimple_build_assign (name, opt);
gsi_insert_before (gsi, ass_stmt, GSI_SAME_STMT);
opt = build1 (VIEW_CONVERT_EXPR, TREE_TYPE (op0), name);
}
gimple_assign_set_rhs_from_tree (gsi, opt);
update_stmt (gsi_stmt (*gsi));
if (TREE_CODE (op0) == SSA_NAME)
ret = remove_prop_source_from_use (op0);
if (op0 != op1 && TREE_CODE (op1) == SSA_NAME)
ret |= remove_prop_source_from_use (op1);
return ret ? 2 : 1;
}
return 0;
}
/* Get the BIT_FIELD_REF definition of VAL, if any, looking through
conversions with code CONV_CODE or update it if still ERROR_MARK.
Return NULL_TREE if no such matching def was found. */
static tree
get_bit_field_ref_def (tree val, enum tree_code &conv_code)
{
if (TREE_CODE (val) != SSA_NAME)
return NULL_TREE ;
gimple *def_stmt = get_prop_source_stmt (val, false, NULL);
if (!def_stmt)
return NULL_TREE;
enum tree_code code = gimple_assign_rhs_code (def_stmt);
if (code == FLOAT_EXPR
|| code == FIX_TRUNC_EXPR
|| CONVERT_EXPR_CODE_P (code))
{
tree op1 = gimple_assign_rhs1 (def_stmt);
if (conv_code == ERROR_MARK)
conv_code = code;
else if (conv_code != code)
return NULL_TREE;
if (TREE_CODE (op1) != SSA_NAME)
return NULL_TREE;
def_stmt = SSA_NAME_DEF_STMT (op1);
if (! is_gimple_assign (def_stmt))
return NULL_TREE;
code = gimple_assign_rhs_code (def_stmt);
}
if (code != BIT_FIELD_REF)
return NULL_TREE;
return gimple_assign_rhs1 (def_stmt);
}
/* Recognize a VEC_PERM_EXPR. Returns true if there were any changes. */
static bool
simplify_vector_constructor (gimple_stmt_iterator *gsi)
{
gimple *stmt = gsi_stmt (*gsi);
tree op, orig[2], type, elem_type;
unsigned elem_size, i;
unsigned HOST_WIDE_INT nelts;
unsigned HOST_WIDE_INT refnelts;
enum tree_code conv_code;
constructor_elt *elt;
op = gimple_assign_rhs1 (stmt);
type = TREE_TYPE (op);
gcc_checking_assert (TREE_CODE (op) == CONSTRUCTOR
&& TREE_CODE (type) == VECTOR_TYPE);
if (!TYPE_VECTOR_SUBPARTS (type).is_constant (&nelts))
return false;
elem_type = TREE_TYPE (type);
elem_size = TREE_INT_CST_LOW (TYPE_SIZE (elem_type));
orig[0] = NULL;
orig[1] = NULL;
conv_code = ERROR_MARK;
bool maybe_ident = true;
bool maybe_blend[2] = { true, true };
tree one_constant = NULL_TREE;
tree one_nonconstant = NULL_TREE;
auto_vec<tree> constants;
constants.safe_grow_cleared (nelts, true);
auto_vec<std::pair<unsigned, unsigned>, 64> elts;
FOR_EACH_VEC_SAFE_ELT (CONSTRUCTOR_ELTS (op), i, elt)
{
tree ref, op1;
unsigned int elem;
if (i >= nelts)
return false;
/* Look for elements extracted and possibly converted from
another vector. */
op1 = get_bit_field_ref_def (elt->value, conv_code);
if (op1
&& TREE_CODE ((ref = TREE_OPERAND (op1, 0))) == SSA_NAME
&& VECTOR_TYPE_P (TREE_TYPE (ref))
&& useless_type_conversion_p (TREE_TYPE (op1),
TREE_TYPE (TREE_TYPE (ref)))
&& constant_multiple_p (bit_field_offset (op1),
bit_field_size (op1), &elem)
&& TYPE_VECTOR_SUBPARTS (TREE_TYPE (ref)).is_constant (&refnelts))
{
unsigned int j;
for (j = 0; j < 2; ++j)
{
if (!orig[j])
{
if (j == 0
|| useless_type_conversion_p (TREE_TYPE (orig[0]),
TREE_TYPE (ref)))
break;
}
else if (ref == orig[j])
break;
}
/* Found a suitable vector element. */
if (j < 2)
{
orig[j] = ref;
if (elem != i || j != 0)
maybe_ident = false;
if (elem != i)
maybe_blend[j] = false;
elts.safe_push (std::make_pair (j, elem));
continue;
}
/* Else fallthru. */
}
/* Handle elements not extracted from a vector.
1. constants by permuting with constant vector
2. a unique non-constant element by permuting with a splat vector */
if (orig[1]
&& orig[1] != error_mark_node)
return false;
orig[1] = error_mark_node;
if (CONSTANT_CLASS_P (elt->value))
{
if (one_nonconstant)
return false;
if (!one_constant)
one_constant = elt->value;
constants[i] = elt->value;
}
else
{
if (one_constant)
return false;
if (!one_nonconstant)
one_nonconstant = elt->value;
else if (!operand_equal_p (one_nonconstant, elt->value, 0))
return false;
}
elts.safe_push (std::make_pair (1, i));
maybe_ident = false;
}
if (i < nelts)
return false;
if (! orig[0]
|| ! VECTOR_TYPE_P (TREE_TYPE (orig[0])))
return false;
refnelts = TYPE_VECTOR_SUBPARTS (TREE_TYPE (orig[0])).to_constant ();
/* We currently do not handle larger destination vectors. */
if (refnelts < nelts)
return false;
if (maybe_ident)
{
tree conv_src_type
= (nelts != refnelts
? (conv_code != ERROR_MARK
? build_vector_type (TREE_TYPE (TREE_TYPE (orig[0])), nelts)
: type)
: TREE_TYPE (orig[0]));
if (conv_code != ERROR_MARK
&& !supportable_convert_operation (conv_code, type, conv_src_type,
&conv_code))
{
/* Only few targets implement direct conversion patterns so try
some simple special cases via VEC_[UN]PACK[_FLOAT]_LO_EXPR. */
optab optab;
insn_code icode;
tree halfvectype, dblvectype;
enum tree_code unpack_op;
if (!BYTES_BIG_ENDIAN)
unpack_op = (FLOAT_TYPE_P (TREE_TYPE (type))
? VEC_UNPACK_FLOAT_LO_EXPR
: VEC_UNPACK_LO_EXPR);
else
unpack_op = (FLOAT_TYPE_P (TREE_TYPE (type))
? VEC_UNPACK_FLOAT_HI_EXPR
: VEC_UNPACK_HI_EXPR);
/* Conversions between DFP and FP have no special tree code
but we cannot handle those since all relevant vector conversion
optabs only have a single mode. */
if (CONVERT_EXPR_CODE_P (conv_code)
&& FLOAT_TYPE_P (TREE_TYPE (type))
&& (DECIMAL_FLOAT_TYPE_P (TREE_TYPE (type))
!= DECIMAL_FLOAT_TYPE_P (TREE_TYPE (conv_src_type))))
return false;
if (CONVERT_EXPR_CODE_P (conv_code)
&& (2 * TYPE_PRECISION (TREE_TYPE (TREE_TYPE (orig[0])))
== TYPE_PRECISION (TREE_TYPE (type)))
&& mode_for_vector (as_a <scalar_mode>
(TYPE_MODE (TREE_TYPE (TREE_TYPE (orig[0])))),
nelts * 2).exists ()
&& (dblvectype
= build_vector_type (TREE_TYPE (TREE_TYPE (orig[0])),
nelts * 2))
/* Only use it for vector modes or for vector booleans
represented as scalar bitmasks. See PR95528. */
&& (VECTOR_MODE_P (TYPE_MODE (dblvectype))
|| VECTOR_BOOLEAN_TYPE_P (dblvectype))
&& (optab = optab_for_tree_code (unpack_op,
dblvectype,
optab_default))
&& ((icode = optab_handler (optab, TYPE_MODE (dblvectype)))
!= CODE_FOR_nothing)
&& (insn_data[icode].operand[0].mode == TYPE_MODE (type)))
{
gimple_seq stmts = NULL;
tree dbl;
if (refnelts == nelts)
{
/* ??? Paradoxical subregs don't exist, so insert into
the lower half of a wider zero vector. */
dbl = gimple_build (&stmts, BIT_INSERT_EXPR, dblvectype,
build_zero_cst (dblvectype), orig[0],
bitsize_zero_node);
}
else if (refnelts == 2 * nelts)
dbl = orig[0];
else
dbl = gimple_build (&stmts, BIT_FIELD_REF, dblvectype,
orig[0], TYPE_SIZE (dblvectype),
bitsize_zero_node);
gsi_insert_seq_before (gsi, stmts, GSI_SAME_STMT);
gimple_assign_set_rhs_with_ops (gsi, unpack_op, dbl);
}
else if (CONVERT_EXPR_CODE_P (conv_code)
&& (TYPE_PRECISION (TREE_TYPE (TREE_TYPE (orig[0])))
== 2 * TYPE_PRECISION (TREE_TYPE (type)))
&& mode_for_vector (as_a <scalar_mode>
(TYPE_MODE
(TREE_TYPE (TREE_TYPE (orig[0])))),
nelts / 2).exists ()
&& (halfvectype
= build_vector_type (TREE_TYPE (TREE_TYPE (orig[0])),
nelts / 2))
/* Only use it for vector modes or for vector booleans
represented as scalar bitmasks. See PR95528. */
&& (VECTOR_MODE_P (TYPE_MODE (halfvectype))
|| VECTOR_BOOLEAN_TYPE_P (halfvectype))
&& (optab = optab_for_tree_code (VEC_PACK_TRUNC_EXPR,
halfvectype,
optab_default))
&& ((icode = optab_handler (optab, TYPE_MODE (halfvectype)))
!= CODE_FOR_nothing)
&& (insn_data[icode].operand[0].mode == TYPE_MODE (type)))
{
gimple_seq stmts = NULL;
tree low = gimple_build (&stmts, BIT_FIELD_REF, halfvectype,
orig[0], TYPE_SIZE (halfvectype),
bitsize_zero_node);
tree hig = gimple_build (&stmts, BIT_FIELD_REF, halfvectype,
orig[0], TYPE_SIZE (halfvectype),
TYPE_SIZE (halfvectype));
gsi_insert_seq_before (gsi, stmts, GSI_SAME_STMT);
gimple_assign_set_rhs_with_ops (gsi, VEC_PACK_TRUNC_EXPR,
low, hig);
}
else
return false;
update_stmt (gsi_stmt (*gsi));
return true;
}
if (nelts != refnelts)
{
gassign *lowpart
= gimple_build_assign (make_ssa_name (conv_src_type),
build3 (BIT_FIELD_REF, conv_src_type,
orig[0], TYPE_SIZE (conv_src_type),
bitsize_zero_node));
gsi_insert_before (gsi, lowpart, GSI_SAME_STMT);
orig[0] = gimple_assign_lhs (lowpart);
}
if (conv_code == ERROR_MARK)
{
tree src_type = TREE_TYPE (orig[0]);
if (!useless_type_conversion_p (type, src_type))
{
gcc_assert (known_eq (TYPE_VECTOR_SUBPARTS (type),
TYPE_VECTOR_SUBPARTS (src_type))
&& useless_type_conversion_p (TREE_TYPE (type),
TREE_TYPE (src_type)));
tree rhs = build1 (VIEW_CONVERT_EXPR, type, orig[0]);
orig[0] = make_ssa_name (type);
gassign *assign = gimple_build_assign (orig[0], rhs);
gsi_insert_before (gsi, assign, GSI_SAME_STMT);
}
gimple_assign_set_rhs_from_tree (gsi, orig[0]);
}
else
gimple_assign_set_rhs_with_ops (gsi, conv_code, orig[0],
NULL_TREE, NULL_TREE);
}
else
{
/* If we combine a vector with a non-vector avoid cases where
we'll obviously end up with more GIMPLE stmts which is when
we'll later not fold this to a single insert into the vector
and we had a single extract originally. See PR92819. */
if (nelts == 2
&& refnelts > 2
&& orig[1] == error_mark_node
&& !maybe_blend[0])
return false;
tree mask_type, perm_type, conv_src_type;
perm_type = TREE_TYPE (orig[0]);
conv_src_type = (nelts == refnelts
? perm_type
: build_vector_type (TREE_TYPE (perm_type), nelts));
if (conv_code != ERROR_MARK
&& !supportable_convert_operation (conv_code, type, conv_src_type,
&conv_code))
return false;
/* Now that we know the number of elements of the source build the
permute vector.
??? When the second vector has constant values we can shuffle
it and its source indexes to make the permutation supported.
For now it mimics a blend. */
vec_perm_builder sel (refnelts, refnelts, 1);
bool all_same_p = true;
for (i = 0; i < elts.length (); ++i)
{
sel.quick_push (elts[i].second + elts[i].first * refnelts);
all_same_p &= known_eq (sel[i], sel[0]);
}
/* And fill the tail with "something". It's really don't care,
and ideally we'd allow VEC_PERM to have a smaller destination
vector. As a heuristic:
(a) if what we have so far duplicates a single element, make the
tail do the same
(b) otherwise preserve a uniform orig[0]. This facilitates
later pattern-matching of VEC_PERM_EXPR to a BIT_INSERT_EXPR. */
for (; i < refnelts; ++i)
sel.quick_push (all_same_p
? sel[0]
: (elts[0].second == 0 && elts[0].first == 0
? 0 : refnelts) + i);
vec_perm_indices indices (sel, orig[1] ? 2 : 1, refnelts);
machine_mode vmode = TYPE_MODE (perm_type);
if (!can_vec_perm_const_p (vmode, vmode, indices))
return false;
mask_type
= build_vector_type (build_nonstandard_integer_type (elem_size, 1),
refnelts);
if (GET_MODE_CLASS (TYPE_MODE (mask_type)) != MODE_VECTOR_INT
|| maybe_ne (GET_MODE_SIZE (TYPE_MODE (mask_type)),
GET_MODE_SIZE (TYPE_MODE (perm_type))))
return false;
tree op2 = vec_perm_indices_to_tree (mask_type, indices);
bool converted_orig1 = false;
gimple_seq stmts = NULL;
if (!orig[1])
orig[1] = orig[0];
else if (orig[1] == error_mark_node
&& one_nonconstant)
{
/* ??? We can see if we can safely convert to the original
element type. */
converted_orig1 = conv_code != ERROR_MARK;
orig[1] = gimple_build_vector_from_val (&stmts, UNKNOWN_LOCATION,
converted_orig1
? type : perm_type,
one_nonconstant);
}
else if (orig[1] == error_mark_node)
{
/* ??? See if we can convert the vector to the original type. */
converted_orig1 = conv_code != ERROR_MARK;
unsigned n = converted_orig1 ? nelts : refnelts;
tree_vector_builder vec (converted_orig1
? type : perm_type, n, 1);
for (unsigned i = 0; i < n; ++i)
if (i < nelts && constants[i])
vec.quick_push (constants[i]);
else
/* ??? Push a don't-care value. */
vec.quick_push (one_constant);
orig[1] = vec.build ();
}
tree blend_op2 = NULL_TREE;
if (converted_orig1)
{
/* Make sure we can do a blend in the target type. */
vec_perm_builder sel (nelts, nelts, 1);
for (i = 0; i < elts.length (); ++i)
sel.quick_push (elts[i].first
? elts[i].second + nelts : i);
vec_perm_indices indices (sel, 2, nelts);
machine_mode vmode = TYPE_MODE (type);
if (!can_vec_perm_const_p (vmode, vmode, indices))
return false;
mask_type
= build_vector_type (build_nonstandard_integer_type (elem_size, 1),
nelts);
if (GET_MODE_CLASS (TYPE_MODE (mask_type)) != MODE_VECTOR_INT
|| maybe_ne (GET_MODE_SIZE (TYPE_MODE (mask_type)),
GET_MODE_SIZE (TYPE_MODE (type))))
return false;
blend_op2 = vec_perm_indices_to_tree (mask_type, indices);
}
tree orig1_for_perm
= converted_orig1 ? build_zero_cst (perm_type) : orig[1];
tree res = gimple_build (&stmts, VEC_PERM_EXPR, perm_type,
orig[0], orig1_for_perm, op2);
if (nelts != refnelts)
res = gimple_build (&stmts, BIT_FIELD_REF,
conv_code != ERROR_MARK ? conv_src_type : type,
res, TYPE_SIZE (type), bitsize_zero_node);
if (conv_code != ERROR_MARK)
res = gimple_build (&stmts, conv_code, type, res);
else if (!useless_type_conversion_p (type, TREE_TYPE (res)))
{
gcc_assert (known_eq (TYPE_VECTOR_SUBPARTS (type),
TYPE_VECTOR_SUBPARTS (perm_type))
&& useless_type_conversion_p (TREE_TYPE (type),
TREE_TYPE (perm_type)));
res = gimple_build (&stmts, VIEW_CONVERT_EXPR, type, res);
}
/* Blend in the actual constant. */
if (converted_orig1)
res = gimple_build (&stmts, VEC_PERM_EXPR, type,
res, orig[1], blend_op2);
gsi_insert_seq_before (gsi, stmts, GSI_SAME_STMT);
gimple_assign_set_rhs_with_ops (gsi, SSA_NAME, res);
}
update_stmt (gsi_stmt (*gsi));
return true;
}
/* Prepare a TARGET_MEM_REF ref so that it can be subsetted as
lvalue. This splits out an address computation stmt before *GSI
and returns a MEM_REF wrapping the address. */
static tree
prepare_target_mem_ref_lvalue (tree ref, gimple_stmt_iterator *gsi)
{
if (TREE_CODE (TREE_OPERAND (ref, 0)) == ADDR_EXPR)
mark_addressable (TREE_OPERAND (TREE_OPERAND (ref, 0), 0));
tree ptrtype = build_pointer_type (TREE_TYPE (ref));
tree tem = make_ssa_name (ptrtype);
gimple *new_stmt
= gimple_build_assign (tem, build1 (ADDR_EXPR, TREE_TYPE (tem),
unshare_expr (ref)));
gsi_insert_before (gsi, new_stmt, GSI_SAME_STMT);
ref = build2_loc (EXPR_LOCATION (ref),
MEM_REF, TREE_TYPE (ref), tem,
build_int_cst (TREE_TYPE (TREE_OPERAND (ref, 1)), 0));
return ref;
}
/* Rewrite the vector load at *GSI to component-wise loads if the load
is only used in BIT_FIELD_REF extractions with eventual intermediate
widening. */
static void
optimize_vector_load (gimple_stmt_iterator *gsi)
{
gimple *stmt = gsi_stmt (*gsi);
tree lhs = gimple_assign_lhs (stmt);
tree rhs = gimple_assign_rhs1 (stmt);
/* Gather BIT_FIELD_REFs to rewrite, looking through
VEC_UNPACK_{LO,HI}_EXPR. */
use_operand_p use_p;
imm_use_iterator iter;
bool rewrite = true;
auto_vec<gimple *, 8> bf_stmts;
auto_vec<tree, 8> worklist;
worklist.quick_push (lhs);
do
{
tree def = worklist.pop ();
unsigned HOST_WIDE_INT def_eltsize
= TREE_INT_CST_LOW (TYPE_SIZE (TREE_TYPE (TREE_TYPE (def))));
FOR_EACH_IMM_USE_FAST (use_p, iter, def)
{
gimple *use_stmt = USE_STMT (use_p);
if (is_gimple_debug (use_stmt))
continue;
if (!is_gimple_assign (use_stmt))
{
rewrite = false;
break;
}
enum tree_code use_code = gimple_assign_rhs_code (use_stmt);
tree use_rhs = gimple_assign_rhs1 (use_stmt);
if (use_code == BIT_FIELD_REF
&& TREE_OPERAND (use_rhs, 0) == def
/* If its on the VEC_UNPACK_{HI,LO}_EXPR
def need to verify it is element aligned. */
&& (def == lhs
|| (known_eq (bit_field_size (use_rhs), def_eltsize)
&& constant_multiple_p (bit_field_offset (use_rhs),
def_eltsize)
/* We can simulate the VEC_UNPACK_{HI,LO}_EXPR
via a NOP_EXPR only for integral types.
??? Support VEC_UNPACK_FLOAT_{HI,LO}_EXPR. */
&& INTEGRAL_TYPE_P (TREE_TYPE (use_rhs)))))
{
bf_stmts.safe_push (use_stmt);
continue;
}
/* Walk through one level of VEC_UNPACK_{LO,HI}_EXPR. */
if (def == lhs
&& (use_code == VEC_UNPACK_HI_EXPR
|| use_code == VEC_UNPACK_LO_EXPR)
&& use_rhs == lhs)
{
worklist.safe_push (gimple_assign_lhs (use_stmt));
continue;
}
rewrite = false;
break;
}
if (!rewrite)
break;
}
while (!worklist.is_empty ());
if (!rewrite)
{
gsi_next (gsi);
return;
}
/* We now have all ultimate uses of the load to rewrite in bf_stmts. */
/* Prepare the original ref to be wrapped in adjusted BIT_FIELD_REFs.
For TARGET_MEM_REFs we have to separate the LEA from the reference. */
tree load_rhs = rhs;
if (TREE_CODE (load_rhs) == TARGET_MEM_REF)
load_rhs = prepare_target_mem_ref_lvalue (load_rhs, gsi);
/* Rewrite the BIT_FIELD_REFs to be actual loads, re-emitting them at
the place of the original load. */
for (gimple *use_stmt : bf_stmts)
{
tree bfr = gimple_assign_rhs1 (use_stmt);
tree new_rhs = unshare_expr (load_rhs);
if (TREE_OPERAND (bfr, 0) != lhs)
{
/* When the BIT_FIELD_REF is on the promoted vector we have to
adjust it and emit a conversion afterwards. */
gimple *def_stmt
= SSA_NAME_DEF_STMT (TREE_OPERAND (bfr, 0));
enum tree_code def_code
= gimple_assign_rhs_code (def_stmt);
/* The adjusted BIT_FIELD_REF is of the promotion source
vector size and at half of the offset... */
new_rhs = fold_build3 (BIT_FIELD_REF,
TREE_TYPE (TREE_TYPE (lhs)),
new_rhs,
TYPE_SIZE (TREE_TYPE (TREE_TYPE (lhs))),
size_binop (EXACT_DIV_EXPR,
TREE_OPERAND (bfr, 2),
bitsize_int (2)));
/* ... and offsetted by half of the vector if VEC_UNPACK_HI_EXPR. */
if (def_code == (!BYTES_BIG_ENDIAN
? VEC_UNPACK_HI_EXPR : VEC_UNPACK_LO_EXPR))
TREE_OPERAND (new_rhs, 2)
= size_binop (PLUS_EXPR, TREE_OPERAND (new_rhs, 2),
size_binop (EXACT_DIV_EXPR,
TYPE_SIZE (TREE_TYPE (lhs)),
bitsize_int (2)));
tree tem = make_ssa_name (TREE_TYPE (TREE_TYPE (lhs)));
gimple *new_stmt = gimple_build_assign (tem, new_rhs);
location_t loc = gimple_location (use_stmt);
gimple_set_location (new_stmt, loc);
gsi_insert_before (gsi, new_stmt, GSI_SAME_STMT);
/* Perform scalar promotion. */
new_stmt = gimple_build_assign (gimple_assign_lhs (use_stmt),
NOP_EXPR, tem);
gimple_set_location (new_stmt, loc);
gsi_insert_before (gsi, new_stmt, GSI_SAME_STMT);
}
else
{
/* When the BIT_FIELD_REF is on the original load result
we can just wrap that. */
tree new_rhs = fold_build3 (BIT_FIELD_REF, TREE_TYPE (bfr),
unshare_expr (load_rhs),
TREE_OPERAND (bfr, 1),
TREE_OPERAND (bfr, 2));
gimple *new_stmt = gimple_build_assign (gimple_assign_lhs (use_stmt),
new_rhs);
location_t loc = gimple_location (use_stmt);
gimple_set_location (new_stmt, loc);
gsi_insert_before (gsi, new_stmt, GSI_SAME_STMT);
}
gimple_stmt_iterator gsi2 = gsi_for_stmt (use_stmt);
unlink_stmt_vdef (use_stmt);
gsi_remove (&gsi2, true);
}
/* Finally get rid of the intermediate stmts. */
gimple *use_stmt;
FOR_EACH_IMM_USE_STMT (use_stmt, iter, lhs)
{
if (is_gimple_debug (use_stmt))
{
if (gimple_debug_bind_p (use_stmt))
{
gimple_debug_bind_reset_value (use_stmt);
update_stmt (use_stmt);
}
continue;
}
gimple_stmt_iterator gsi2 = gsi_for_stmt (use_stmt);
unlink_stmt_vdef (use_stmt);
release_defs (use_stmt);
gsi_remove (&gsi2, true);
}
/* And the original load. */
release_defs (stmt);
gsi_remove (gsi, true);
}
/* Primitive "lattice" function for gimple_simplify. */
static tree
fwprop_ssa_val (tree name)
{
/* First valueize NAME. */
if (TREE_CODE (name) == SSA_NAME
&& SSA_NAME_VERSION (name) < lattice.length ())
{
tree val = lattice[SSA_NAME_VERSION (name)];
if (val)
name = val;
}
/* We continue matching along SSA use-def edges for SSA names
that are not single-use. Currently there are no patterns
that would cause any issues with that. */
return name;
}
/* Get an index map from the provided vector permute selector
and return the number of unique indices.
E.g.: { 1, 3, 1, 3 } -> <0, 1, 0, 1>, 2
{ 0, 2, 0, 2 } -> <0, 1, 0, 1>, 2
{ 3, 2, 1, 0 } -> <0, 1, 2, 3>, 4. */
static unsigned int
get_vect_selector_index_map (tree sel, vec<unsigned int> *index_map)
{
gcc_assert (VECTOR_CST_NELTS (sel).is_constant ());
unsigned int nelts = VECTOR_CST_NELTS (sel).to_constant ();
unsigned int n = 0;
for (unsigned int i = 0; i < nelts; i++)
{
/* Extract the i-th value from the selector. */
tree sel_cst_tree = VECTOR_CST_ELT (sel, i);
unsigned int sel_cst = TREE_INT_CST_LOW (sel_cst_tree);
unsigned int j = 0;
for (; j <= i; j++)
{
tree prev_sel_cst_tree = VECTOR_CST_ELT (sel, j);
unsigned int prev_sel_cst
= TREE_INT_CST_LOW (prev_sel_cst_tree);
if (prev_sel_cst == sel_cst)
break;
}
index_map->quick_push (j);
n += (i == j) ? 1 : 0;
}
return n;
}
/* Search for opportunities to free half of the lanes in the following pattern:
v_in = {e0, e1, e2, e3}
v_1 = VEC_PERM <v_in, v_in, {0, 2, 0, 2}>
// v_1 = {e0, e2, e0, e2}
v_2 = VEC_PERM <v_in, v_in, {1, 3, 1, 3}>
// v_2 = {e1, e3, e1, e3}
v_x = v_1 + v_2
// v_x = {e0+e1, e2+e3, e0+e1, e2+e3}
v_y = v_1 - v_2
// v_y = {e0-e1, e2-e3, e0-e1, e2-e3}
v_out = VEC_PERM <v_x, v_y, {0, 1, 6, 7}>
// v_out = {e0+e1, e2+e3, e0-e1, e2-e3}
The last statement could be simplified to:
v_out' = VEC_PERM <v_x, v_y, {0, 1, 4, 5}>
// v_out' = {e0+e1, e2+e3, e0-e1, e2-e3}
Characteristic properties:
- v_1 and v_2 are created from the same input vector v_in and introduce the
lane duplication (in the selection operand) that we can eliminate.
- v_x and v_y are results from lane-preserving operations that use v_1 and
v_2 as inputs.
- v_out is created by selecting from duplicated lanes. */
static bool
recognise_vec_perm_simplify_seq (gassign *stmt, vec_perm_simplify_seq *seq)
{
gcc_checking_assert (stmt);
gcc_checking_assert (gimple_assign_rhs_code (stmt) == VEC_PERM_EXPR);
basic_block bb = gimple_bb (stmt);
/* Decompose the final vec permute statement. */
tree v_x = gimple_assign_rhs1 (stmt);
tree v_y = gimple_assign_rhs2 (stmt);
tree sel = gimple_assign_rhs3 (stmt);
if (!VECTOR_CST_NELTS (sel).is_constant ()
|| TREE_CODE (v_x) != SSA_NAME
|| TREE_CODE (v_y) != SSA_NAME
|| !has_single_use (v_x)
|| !has_single_use (v_y))
return false;
unsigned int nelts = VECTOR_CST_NELTS (sel).to_constant ();
/* Lookup the definition of v_x and v_y. */
gassign *v_x_stmt = dyn_cast<gassign *> (SSA_NAME_DEF_STMT (v_x));
gassign *v_y_stmt = dyn_cast<gassign *> (SSA_NAME_DEF_STMT (v_y));
if (!v_x_stmt || gimple_bb (v_x_stmt) != bb
|| !v_y_stmt || gimple_bb (v_y_stmt) != bb)
return false;
/* Check the operations that define v_x and v_y. */
if (TREE_CODE_CLASS (gimple_assign_rhs_code (v_x_stmt)) != tcc_binary
|| TREE_CODE_CLASS (gimple_assign_rhs_code (v_y_stmt)) != tcc_binary)
return false;
tree v_x_1 = gimple_assign_rhs1 (v_x_stmt);
tree v_x_2 = gimple_assign_rhs2 (v_x_stmt);
tree v_y_1 = gimple_assign_rhs1 (v_y_stmt);
tree v_y_2 = gimple_assign_rhs2 (v_y_stmt);
if (v_x_stmt == v_y_stmt
|| TREE_CODE (v_x_1) != SSA_NAME
|| TREE_CODE (v_x_2) != SSA_NAME
|| num_imm_uses (v_x_1) != 2
|| num_imm_uses (v_x_2) != 2)
return false;
if (v_x_1 != v_y_1 || v_x_2 != v_y_2)
{
/* Allow operands of commutative operators to swap. */
if (commutative_tree_code (gimple_assign_rhs_code (v_x_stmt)))
{
/* Keep v_x_1 the first operand for non-commutative operators. */
v_x_1 = gimple_assign_rhs2 (v_x_stmt);
v_x_2 = gimple_assign_rhs1 (v_x_stmt);
if (v_x_1 != v_y_1 || v_x_2 != v_y_2)
return false;
}
else if (commutative_tree_code (gimple_assign_rhs_code (v_y_stmt)))
{
if (v_x_1 != v_y_2 || v_x_2 != v_y_1)
return false;
}
else
return false;
}
gassign *v_1_stmt = dyn_cast<gassign *> (SSA_NAME_DEF_STMT (v_x_1));
gassign *v_2_stmt = dyn_cast<gassign *> (SSA_NAME_DEF_STMT (v_x_2));
if (!v_1_stmt || gimple_bb (v_1_stmt) != bb
|| !v_2_stmt || gimple_bb (v_2_stmt) != bb)
return false;
if (gimple_assign_rhs_code (v_1_stmt) != VEC_PERM_EXPR
|| gimple_assign_rhs_code (v_2_stmt) != VEC_PERM_EXPR)
return false;
/* Decompose initial VEC_PERM_EXPRs. */
tree v_in = gimple_assign_rhs1 (v_1_stmt);
tree v_1_sel = gimple_assign_rhs3 (v_1_stmt);
tree v_2_sel = gimple_assign_rhs3 (v_2_stmt);
if (v_in != gimple_assign_rhs2 (v_1_stmt)
|| v_in != gimple_assign_rhs1 (v_2_stmt)
|| v_in != gimple_assign_rhs2 (v_2_stmt))
return false;
if (!VECTOR_CST_NELTS (v_1_sel).is_constant ()
|| !VECTOR_CST_NELTS (v_2_sel).is_constant ())
return false;
if (nelts != VECTOR_CST_NELTS (v_1_sel).to_constant ()
|| nelts != VECTOR_CST_NELTS (v_2_sel).to_constant ())
return false;
/* Now check permutation selection operands. */
auto_vec<unsigned int> v_1_lane_map, v_2_lane_map;
v_1_lane_map.reserve (nelts);
v_2_lane_map.reserve (nelts);
unsigned int v_1_lanes, v_2_lanes;
v_1_lanes = get_vect_selector_index_map (v_1_sel, &v_1_lane_map);
v_2_lanes = get_vect_selector_index_map (v_2_sel, &v_2_lane_map);
/* Check if we could free up half of the lanes. */
if (v_1_lanes != v_2_lanes || v_1_lanes > (nelts / 2))
return false;
/* Create the new selector. */
vec_perm_builder new_sel_perm (nelts, nelts, 1);
for (unsigned int i = 0; i < nelts; i++)
{
/* Extract the i-th value from the selector. */
tree sel_cst_tree = VECTOR_CST_ELT (sel, i);
unsigned int sel_cst = TREE_INT_CST_LOW (sel_cst_tree);
unsigned int j;
if (sel_cst < nelts)
j = v_1_lane_map[sel_cst];
else
j = v_2_lane_map[sel_cst - nelts] + nelts;
new_sel_perm.quick_push (j);
if (dump_file && (dump_flags & TDF_DETAILS))
{
fprintf (dump_file, "%u", j);
if (i != (nelts -1))
fprintf (dump_file, ", ");
}
}
vec_perm_indices new_indices (new_sel_perm, 2, nelts);
tree vectype = TREE_TYPE (gimple_assign_lhs (stmt));
machine_mode vmode = TYPE_MODE (vectype);
if (!can_vec_perm_const_p (vmode, vmode, new_indices, false))
return false;
*seq = XNEW (struct _vec_perm_simplify_seq);
(*seq)->stmt = stmt;
(*seq)->v_1_stmt = v_1_stmt;
(*seq)->v_2_stmt = v_2_stmt;
(*seq)->v_x_stmt = v_x_stmt;
(*seq)->v_y_stmt = v_y_stmt;
(*seq)->nelts = nelts;
(*seq)->new_sel = vect_gen_perm_mask_checked (vectype, new_indices);
if (dump_file)
{
fprintf (dump_file, "Found vec perm simplify sequence ending with: ");
print_gimple_stmt (dump_file, stmt, 0);
}
return true;
}
/* Reduce the lane consumption of a simplifiable vec perm sequence. */
static void
narrow_vec_perm_simplify_seq (const vec_perm_simplify_seq &seq)
{
gassign *stmt = seq->stmt;
if (dump_file && (dump_flags & TDF_DETAILS))
{
fprintf (dump_file, "Updating VEC_PERM statment:\n");
fprintf (dump_file, "Old stmt: ");
print_gimple_stmt (dump_file, stmt, 0);
}
/* Update the last VEC_PERM statement. */
gimple_assign_set_rhs3 (stmt, seq->new_sel);
update_stmt (stmt);
if (dump_file && (dump_flags & TDF_DETAILS))
{
fprintf (dump_file, "New stmt: ");
print_gimple_stmt (dump_file, stmt, 0);
}
}
/* Test if we can blend two simplifiable vec permute sequences.
NEED_SWAP will be set, if sequences must be swapped for blending. */
static bool
can_blend_vec_perm_simplify_seqs_p (vec_perm_simplify_seq seq1,
vec_perm_simplify_seq seq2,
bool *need_swap)
{
unsigned int nelts = seq1->nelts;
basic_block bb = gimple_bb (seq1->stmt);
gcc_assert (gimple_bb (seq2->stmt) == bb);
/* BBs and number of elements must be equal. */
if (gimple_bb (seq2->stmt) != bb || seq2->nelts != nelts)
return false;
/* We need vectors of the same type. */
if (TREE_TYPE (gimple_assign_lhs (seq1->stmt))
!= TREE_TYPE (gimple_assign_lhs (seq2->stmt)))
return false;
/* We require isomorphic operators. */
if (((gimple_assign_rhs_code (seq1->v_x_stmt)
!= gimple_assign_rhs_code (seq2->v_x_stmt))
|| (gimple_assign_rhs_code (seq1->v_y_stmt)
!= gimple_assign_rhs_code (seq2->v_y_stmt))))
return false;
/* We cannot have any dependencies between the sequences.
For merging, we will reuse seq1->v_1_stmt and seq1->v_2_stmt.
seq1's v_in is defined before these statements, but we need
to check if seq2's v_in is defined before them as well.
Further, we will reuse seq2->stmt. We need to ensure that
seq1->v_x_stmt and seq1->v_y_stmt are before it.
Note, that we don't need to check the BBs here, because all
statements of both sequences have to be in the same BB.
*/
tree seq2_v_in = gimple_assign_rhs1 (seq2->v_1_stmt);
if (TREE_CODE (seq2_v_in) != SSA_NAME)
return false;
gassign *seq2_v_in_stmt = dyn_cast<gassign *> (SSA_NAME_DEF_STMT (seq2_v_in));
if (!seq2_v_in_stmt || gimple_bb (seq2_v_in_stmt) != bb
|| (gimple_uid (seq2_v_in_stmt) > gimple_uid (seq1->v_1_stmt))
|| (gimple_uid (seq1->v_x_stmt) > gimple_uid (seq2->stmt))
|| (gimple_uid (seq1->v_y_stmt) > gimple_uid (seq2->stmt)))
{
tree seq1_v_in = gimple_assign_rhs1 (seq1->v_1_stmt);
if (TREE_CODE (seq1_v_in) != SSA_NAME)
return false;
gassign *seq1_v_in_stmt
= dyn_cast<gassign *> (SSA_NAME_DEF_STMT (seq1_v_in));
/* Let's try to see if we succeed when swapping the sequences. */
if (!seq1_v_in_stmt || gimple_bb (seq1_v_in_stmt)
|| (gimple_uid (seq1_v_in_stmt) > gimple_uid (seq2->v_1_stmt))
|| (gimple_uid (seq2->v_x_stmt) > gimple_uid (seq1->stmt))
|| (gimple_uid (seq2->v_y_stmt) > gimple_uid (seq1->stmt)))
return false;
*need_swap = true;
}
else
*need_swap = false;
if (dump_file && (dump_flags & TDF_DETAILS))
fprintf (dump_file, "Found vec perm simplify sequence pair.\n");
return true;
}
/* Calculate the permutations for blending the two given vec permute
sequences. This may fail if the resulting permutation is not
supported. */
static bool
calc_perm_vec_perm_simplify_seqs (vec_perm_simplify_seq seq1,
vec_perm_simplify_seq seq2,
vec_perm_indices *seq2_stmt_indices,
vec_perm_indices *seq1_v_1_stmt_indices,
vec_perm_indices *seq1_v_2_stmt_indices)
{
unsigned int i;
unsigned int nelts = seq1->nelts;
auto_vec<int> lane_assignment;
lane_assignment.create (2 * nelts);
/* Mark all lanes as free. */
lane_assignment.quick_grow_cleared (2 * nelts);
/* Reserve lanes for seq1. */
for (i = 0; i < nelts; i++)
{
unsigned int l = TREE_INT_CST_LOW (VECTOR_CST_ELT (seq1->new_sel, i));
lane_assignment[l] = 1;
}
/* Reserve lanes for seq2 and calculate selector for seq2->stmt. */
vec_perm_builder seq2_stmt_sel_perm (nelts, nelts, 1);
for (i = 0; i < nelts; i++)
{
unsigned int l = TREE_INT_CST_LOW (VECTOR_CST_ELT (seq2->new_sel, i));
while (lane_assignment[l] != 0)
l++;
lane_assignment[l] = 2;
seq2_stmt_sel_perm.quick_push (l);
}
seq2_stmt_indices->new_vector (seq2_stmt_sel_perm, 2, nelts);
tree vectype = TREE_TYPE (gimple_assign_lhs (seq2->stmt));
machine_mode vmode = TYPE_MODE (vectype);
if (!can_vec_perm_const_p (vmode, vmode, *seq2_stmt_indices, false))
return false;
/* Calculate selectors for seq1->v_1_stmt and seq1->v_2_stmt. */
vec_perm_builder seq1_v_1_stmt_sel_perm (nelts, nelts, 1);
vec_perm_builder seq1_v_2_stmt_sel_perm (nelts, nelts, 1);
for (i = 0; i < nelts; i++)
{
bool use_seq1 = lane_assignment[i] == 1;
tree s1 = gimple_assign_rhs3 (use_seq1 ? seq1->v_1_stmt
: seq2->v_1_stmt);
tree s2 = gimple_assign_rhs3 (use_seq1 ? seq1->v_2_stmt
: seq2->v_2_stmt);
unsigned int l1 = TREE_INT_CST_LOW (VECTOR_CST_ELT (s1, i)) % nelts;
unsigned int l2 = TREE_INT_CST_LOW (VECTOR_CST_ELT (s2, i)) % nelts;
seq1_v_1_stmt_sel_perm.quick_push (l1 + (use_seq1 ? 0 : nelts));
seq1_v_2_stmt_sel_perm.quick_push (l2 + (use_seq1 ? 0 : nelts));
}
seq1_v_1_stmt_indices->new_vector (seq1_v_1_stmt_sel_perm, 2, nelts);
vectype = TREE_TYPE (gimple_assign_lhs (seq1->v_1_stmt));
vmode = TYPE_MODE (vectype);
if (!can_vec_perm_const_p (vmode, vmode, *seq1_v_1_stmt_indices, false))
return false;
seq1_v_2_stmt_indices->new_vector (seq1_v_2_stmt_sel_perm, 2, nelts);
vectype = TREE_TYPE (gimple_assign_lhs (seq1->v_2_stmt));
vmode = TYPE_MODE (vectype);
if (!can_vec_perm_const_p (vmode, vmode, *seq1_v_2_stmt_indices, false))
return false;
return true;
}
/* Blend the two given simplifiable vec permute sequences using the
given permutations. */
static void
blend_vec_perm_simplify_seqs (vec_perm_simplify_seq seq1,
vec_perm_simplify_seq seq2,
const vec_perm_indices &seq2_stmt_indices,
const vec_perm_indices &seq1_v_1_stmt_indices,
const vec_perm_indices &seq1_v_2_stmt_indices)
{
/* We don't need to adjust seq1->stmt because its lanes consumption
was already narrowed before entering this function. */
/* Adjust seq2->stmt: copy RHS1/RHS2 from seq1->stmt and set new sel. */
if (dump_file && (dump_flags & TDF_DETAILS))
{
fprintf (dump_file, "Updating VEC_PERM statment:\n");
fprintf (dump_file, "Old stmt: ");
print_gimple_stmt (dump_file, seq2->stmt, 0);
}
gimple_assign_set_rhs1 (seq2->stmt, gimple_assign_rhs1 (seq1->stmt));
gimple_assign_set_rhs2 (seq2->stmt, gimple_assign_rhs2 (seq1->stmt));
tree vectype = TREE_TYPE (gimple_assign_lhs (seq2->stmt));
tree sel = vect_gen_perm_mask_checked (vectype, seq2_stmt_indices);
gimple_assign_set_rhs3 (seq2->stmt, sel);
update_stmt (seq2->stmt);
if (dump_file && (dump_flags & TDF_DETAILS))
{
fprintf (dump_file, "New stmt: ");
print_gimple_stmt (dump_file, seq2->stmt, 0);
}
/* Adjust seq1->v_1_stmt: copy RHS2 from seq2->v_1_stmt and set new sel. */
if (dump_file && (dump_flags & TDF_DETAILS))
{
fprintf (dump_file, "Updating VEC_PERM statment:\n");
fprintf (dump_file, "Old stmt: ");
print_gimple_stmt (dump_file, seq1->v_1_stmt, 0);
}
gimple_assign_set_rhs2 (seq1->v_1_stmt, gimple_assign_rhs1 (seq2->v_1_stmt));
vectype = TREE_TYPE (gimple_assign_lhs (seq1->v_1_stmt));
sel = vect_gen_perm_mask_checked (vectype, seq1_v_1_stmt_indices);
gimple_assign_set_rhs3 (seq1->v_1_stmt, sel);
update_stmt (seq1->v_1_stmt);
if (dump_file && (dump_flags & TDF_DETAILS))
{
fprintf (dump_file, "New stmt: ");
print_gimple_stmt (dump_file, seq1->v_1_stmt, 0);
}
/* Adjust seq1->v_2_stmt: copy RHS2 from seq2->v_2_stmt and set new sel. */
if (dump_file && (dump_flags & TDF_DETAILS))
{
fprintf (dump_file, "Updating VEC_PERM statment:\n");
fprintf (dump_file, "Old stmt: ");
print_gimple_stmt (dump_file, seq1->v_2_stmt, 0);
}
gimple_assign_set_rhs2 (seq1->v_2_stmt, gimple_assign_rhs1 (seq2->v_2_stmt));
vectype = TREE_TYPE (gimple_assign_lhs (seq1->v_2_stmt));
sel = vect_gen_perm_mask_checked (vectype, seq1_v_2_stmt_indices);
gimple_assign_set_rhs3 (seq1->v_2_stmt, sel);
update_stmt (seq1->v_2_stmt);
if (dump_file && (dump_flags & TDF_DETAILS))
{
fprintf (dump_file, "New stmt: ");
print_gimple_stmt (dump_file, seq1->v_2_stmt, 0);
}
/* At this point, we have four unmodified seq2 stmts, which will be
eliminated by DCE. */
if (dump_file)
fprintf (dump_file, "Vec perm simplify sequences have been blended.\n\n");
}
/* Try to blend narrowed vec_perm_simplify_seqs pairwise.
The provided list will be empty after this call. */
static void
process_vec_perm_simplify_seq_list (vec<vec_perm_simplify_seq> *l)
{
unsigned int i, j;
vec_perm_simplify_seq seq1, seq2;
if (l->is_empty ())
return;
if (dump_file && (dump_flags & TDF_DETAILS))
fprintf (dump_file, "Processing %u vec perm simplify sequences.\n",
l->length ());
FOR_EACH_VEC_ELT (*l, i, seq1)
{
if (i + 1 < l->length ())
{
FOR_EACH_VEC_ELT_FROM (*l, j, seq2, i + 1)
{
bool swap = false;
if (can_blend_vec_perm_simplify_seqs_p (seq1, seq2, &swap))
{
vec_perm_indices seq2_stmt_indices;
vec_perm_indices seq1_v_1_stmt_indices;
vec_perm_indices seq1_v_2_stmt_indices;
if (calc_perm_vec_perm_simplify_seqs (swap ? seq2 : seq1,
swap ? seq1 : seq2,
&seq2_stmt_indices,
&seq1_v_1_stmt_indices,
&seq1_v_2_stmt_indices))
{
/* Narrow lane usage. */
narrow_vec_perm_simplify_seq (seq1);
narrow_vec_perm_simplify_seq (seq2);
/* Blend sequences. */
blend_vec_perm_simplify_seqs (swap ? seq2 : seq1,
swap ? seq1 : seq2,
seq2_stmt_indices,
seq1_v_1_stmt_indices,
seq1_v_2_stmt_indices);
/* We can use unordered_remove as we break the loop. */
l->unordered_remove (j);
XDELETE (seq2);
break;
}
}
}
}
/* We don't need to call l->remove for seq1. */
XDELETE (seq1);
}
l->truncate (0);
}
static void
append_vec_perm_simplify_seq_list (vec<vec_perm_simplify_seq> *l,
const vec_perm_simplify_seq &seq)
{
/* If no space on list left, then process the list. */
if (!l->space (1))
process_vec_perm_simplify_seq_list (l);
l->quick_push (seq);
}
/* Main entry point for the forward propagation and statement combine
optimizer. */
namespace {
const pass_data pass_data_forwprop =
{
GIMPLE_PASS, /* type */
"forwprop", /* name */
OPTGROUP_NONE, /* optinfo_flags */
TV_TREE_FORWPROP, /* tv_id */
( PROP_cfg | PROP_ssa ), /* properties_required */
0, /* properties_provided */
0, /* properties_destroyed */
0, /* todo_flags_start */
TODO_update_ssa, /* todo_flags_finish */
};
class pass_forwprop : public gimple_opt_pass
{
public:
pass_forwprop (gcc::context *ctxt)
: gimple_opt_pass (pass_data_forwprop, ctxt)
{}
/* opt_pass methods: */
opt_pass * clone () final override { return new pass_forwprop (m_ctxt); }
bool gate (function *) final override { return flag_tree_forwprop; }
unsigned int execute (function *) final override;
}; // class pass_forwprop
unsigned int
pass_forwprop::execute (function *fun)
{
unsigned int todoflags = 0;
cfg_changed = false;
calculate_dominance_info (CDI_DOMINATORS);
/* Combine stmts with the stmts defining their operands. Do that
in an order that guarantees visiting SSA defs before SSA uses. */
lattice.create (num_ssa_names);
lattice.quick_grow_cleared (num_ssa_names);
int *postorder = XNEWVEC (int, n_basic_blocks_for_fn (fun));
int postorder_num = pre_and_rev_post_order_compute_fn (fun, NULL,
postorder, false);
int *bb_to_rpo = XNEWVEC (int, last_basic_block_for_fn (fun));
for (int i = 0; i < postorder_num; ++i)
{
bb_to_rpo[postorder[i]] = i;
edge_iterator ei;
edge e;
FOR_EACH_EDGE (e, ei, BASIC_BLOCK_FOR_FN (fun, postorder[i])->succs)
e->flags &= ~EDGE_EXECUTABLE;
}
single_succ_edge (BASIC_BLOCK_FOR_FN (fun, ENTRY_BLOCK))->flags
|= EDGE_EXECUTABLE;
auto_vec<gimple *, 4> to_fixup;
auto_vec<gimple *, 32> to_remove;
auto_vec<unsigned, 32> to_remove_defs;
auto_vec<std::pair<int, int>, 10> edges_to_remove;
auto_bitmap simple_dce_worklist;
auto_bitmap need_ab_cleanup;
to_purge = BITMAP_ALLOC (NULL);
auto_vec<vec_perm_simplify_seq, 8> vec_perm_simplify_seq_list;
for (int i = 0; i < postorder_num; ++i)
{
gimple_stmt_iterator gsi;
basic_block bb = BASIC_BLOCK_FOR_FN (fun, postorder[i]);
edge_iterator ei;
edge e;
/* Skip processing not executable blocks. We could improve
single_use tracking by at least unlinking uses from unreachable
blocks but since blocks with uses are not processed in a
meaningful order this is probably not worth it. */
bool any = false;
FOR_EACH_EDGE (e, ei, bb->preds)
{
if ((e->flags & EDGE_EXECUTABLE)
/* We can handle backedges in natural loops correctly but
for irreducible regions we have to take all backedges
conservatively when we did not visit the source yet. */
|| (bb_to_rpo[e->src->index] > i
&& !dominated_by_p (CDI_DOMINATORS, e->src, e->dest)))
{
any = true;
break;
}
}
if (!any)
continue;
/* Record degenerate PHIs in the lattice. */
for (gphi_iterator si = gsi_start_phis (bb); !gsi_end_p (si);
gsi_next (&si))
{
gphi *phi = si.phi ();
tree res = gimple_phi_result (phi);
if (virtual_operand_p (res))
continue;
tree first = NULL_TREE;
bool all_same = true;
edge_iterator ei;
edge e;
FOR_EACH_EDGE (e, ei, bb->preds)
{
/* Ignore not executable forward edges. */
if (!(e->flags & EDGE_EXECUTABLE))
{
if (bb_to_rpo[e->src->index] < i)
continue;
/* Avoid equivalences from backedges - while we might
be able to make irreducible regions reducible and
thus turning a back into a forward edge we do not
want to deal with the intermediate SSA issues that
exposes. */
all_same = false;
}
tree use = PHI_ARG_DEF_FROM_EDGE (phi, e);
if (use == res)
/* The PHI result can also appear on a backedge, if so
we can ignore this case for the purpose of determining
the singular value. */
;
else if (! first)
first = use;
else if (! operand_equal_p (first, use, 0))
{
all_same = false;
break;
}
}
if (all_same)
{
if (may_propagate_copy (res, first))
to_remove_defs.safe_push (SSA_NAME_VERSION (res));
fwprop_set_lattice_val (res, first);
}
}
/* Apply forward propagation to all stmts in the basic-block.
Note we update GSI within the loop as necessary. */
unsigned int uid = 1;
for (gsi = gsi_start_bb (bb); !gsi_end_p (gsi); )
{
gimple *stmt = gsi_stmt (gsi);
tree lhs, rhs;
enum tree_code code;
gimple_set_uid (stmt, uid++);
if (!is_gimple_assign (stmt))
{
process_vec_perm_simplify_seq_list (&vec_perm_simplify_seq_list);
gsi_next (&gsi);
continue;
}
lhs = gimple_assign_lhs (stmt);
rhs = gimple_assign_rhs1 (stmt);
code = gimple_assign_rhs_code (stmt);
if (TREE_CODE (lhs) != SSA_NAME
|| has_zero_uses (lhs))
{
process_vec_perm_simplify_seq_list (&vec_perm_simplify_seq_list);
gsi_next (&gsi);
continue;
}
/* If this statement sets an SSA_NAME to an address,
try to propagate the address into the uses of the SSA_NAME. */
if ((code == ADDR_EXPR
/* Handle pointer conversions on invariant addresses
as well, as this is valid gimple. */
|| (CONVERT_EXPR_CODE_P (code)
&& TREE_CODE (rhs) == ADDR_EXPR
&& POINTER_TYPE_P (TREE_TYPE (lhs))))
&& TREE_CODE (TREE_OPERAND (rhs, 0)) != TARGET_MEM_REF)
{
tree base = get_base_address (TREE_OPERAND (rhs, 0));
if ((!base
|| !DECL_P (base)
|| decl_address_invariant_p (base))
&& !stmt_references_abnormal_ssa_name (stmt)
&& forward_propagate_addr_expr (lhs, rhs, true))
{
fwprop_invalidate_lattice (gimple_get_lhs (stmt));
release_defs (stmt);
gsi_remove (&gsi, true);
}
else
gsi_next (&gsi);
}
else if (code == POINTER_PLUS_EXPR)
{
tree off = gimple_assign_rhs2 (stmt);
if (TREE_CODE (off) == INTEGER_CST
&& can_propagate_from (stmt)
&& !simple_iv_increment_p (stmt)
/* ??? Better adjust the interface to that function
instead of building new trees here. */
&& forward_propagate_addr_expr
(lhs,
build1_loc (gimple_location (stmt),
ADDR_EXPR, TREE_TYPE (rhs),
fold_build2 (MEM_REF,
TREE_TYPE (TREE_TYPE (rhs)),
rhs,
fold_convert (ptr_type_node,
off))), true))
{
fwprop_invalidate_lattice (gimple_get_lhs (stmt));
release_defs (stmt);
gsi_remove (&gsi, true);
}
else if (is_gimple_min_invariant (rhs))
{
/* Make sure to fold &a[0] + off_1 here. */
fold_stmt_inplace (&gsi);
update_stmt (stmt);
if (gimple_assign_rhs_code (stmt) == POINTER_PLUS_EXPR)
gsi_next (&gsi);
}
else
gsi_next (&gsi);
}
else if (TREE_CODE (TREE_TYPE (lhs)) == COMPLEX_TYPE
&& gimple_assign_load_p (stmt)
&& !gimple_has_volatile_ops (stmt)
&& TREE_CODE (rhs) != TARGET_MEM_REF
&& TREE_CODE (rhs) != BIT_FIELD_REF
&& !stmt_can_throw_internal (fun, stmt))
{
/* Rewrite loads used only in real/imagpart extractions to
component-wise loads. */
use_operand_p use_p;
imm_use_iterator iter;
bool rewrite = true;
FOR_EACH_IMM_USE_FAST (use_p, iter, lhs)
{
gimple *use_stmt = USE_STMT (use_p);
if (is_gimple_debug (use_stmt))
continue;
if (!is_gimple_assign (use_stmt)
|| (gimple_assign_rhs_code (use_stmt) != REALPART_EXPR
&& gimple_assign_rhs_code (use_stmt) != IMAGPART_EXPR)
|| TREE_OPERAND (gimple_assign_rhs1 (use_stmt), 0) != lhs)
{
rewrite = false;
break;
}
}
if (rewrite)
{
gimple *use_stmt;
FOR_EACH_IMM_USE_STMT (use_stmt, iter, lhs)
{
if (is_gimple_debug (use_stmt))
{
if (gimple_debug_bind_p (use_stmt))
{
gimple_debug_bind_reset_value (use_stmt);
update_stmt (use_stmt);
}
continue;
}
tree new_rhs = build1 (gimple_assign_rhs_code (use_stmt),
TREE_TYPE (TREE_TYPE (rhs)),
unshare_expr (rhs));
gimple *new_stmt
= gimple_build_assign (gimple_assign_lhs (use_stmt),
new_rhs);
location_t loc = gimple_location (use_stmt);
gimple_set_location (new_stmt, loc);
gimple_stmt_iterator gsi2 = gsi_for_stmt (use_stmt);
unlink_stmt_vdef (use_stmt);
gsi_remove (&gsi2, true);
gsi_insert_before (&gsi, new_stmt, GSI_SAME_STMT);
}
release_defs (stmt);
gsi_remove (&gsi, true);
}
else
gsi_next (&gsi);
}
else if (TREE_CODE (TREE_TYPE (lhs)) == VECTOR_TYPE
&& (TYPE_MODE (TREE_TYPE (lhs)) == BLKmode
/* After vector lowering rewrite all loads, but
initially do not since this conflicts with
vector CONSTRUCTOR to shuffle optimization. */
|| (fun->curr_properties & PROP_gimple_lvec))
&& gimple_assign_load_p (stmt)
&& !gimple_has_volatile_ops (stmt)
&& !stmt_can_throw_internal (fun, stmt)
&& (!VAR_P (rhs) || !DECL_HARD_REGISTER (rhs)))
optimize_vector_load (&gsi);
else if (code == COMPLEX_EXPR)
{
/* Rewrite stores of a single-use complex build expression
to component-wise stores. */
use_operand_p use_p;
gimple *use_stmt, *def1, *def2;
tree rhs2;
if (single_imm_use (lhs, &use_p, &use_stmt)
&& gimple_store_p (use_stmt)
&& !gimple_has_volatile_ops (use_stmt)
&& is_gimple_assign (use_stmt)
&& (TREE_CODE (TREE_TYPE (gimple_assign_lhs (use_stmt)))
== COMPLEX_TYPE)
&& (TREE_CODE (gimple_assign_lhs (use_stmt))
!= TARGET_MEM_REF))
{
tree use_lhs = gimple_assign_lhs (use_stmt);
if (auto_var_p (use_lhs))
DECL_NOT_GIMPLE_REG_P (use_lhs) = 1;
tree new_lhs = build1 (REALPART_EXPR,
TREE_TYPE (TREE_TYPE (use_lhs)),
unshare_expr (use_lhs));
gimple *new_stmt = gimple_build_assign (new_lhs, rhs);
location_t loc = gimple_location (use_stmt);
gimple_set_location (new_stmt, loc);
gimple_set_vuse (new_stmt, gimple_vuse (use_stmt));
gimple_set_vdef (new_stmt, make_ssa_name (gimple_vop (fun)));
SSA_NAME_DEF_STMT (gimple_vdef (new_stmt)) = new_stmt;
gimple_set_vuse (use_stmt, gimple_vdef (new_stmt));
gimple_stmt_iterator gsi2 = gsi_for_stmt (use_stmt);
gsi_insert_before (&gsi2, new_stmt, GSI_SAME_STMT);
new_lhs = build1 (IMAGPART_EXPR,
TREE_TYPE (TREE_TYPE (use_lhs)),
unshare_expr (use_lhs));
gimple_assign_set_lhs (use_stmt, new_lhs);
gimple_assign_set_rhs1 (use_stmt, gimple_assign_rhs2 (stmt));
update_stmt (use_stmt);
release_defs (stmt);
gsi_remove (&gsi, true);
}
/* Rewrite a component-wise load of a complex to a complex
load if the components are not used separately. */
else if (TREE_CODE (rhs) == SSA_NAME
&& has_single_use (rhs)
&& ((rhs2 = gimple_assign_rhs2 (stmt)), true)
&& TREE_CODE (rhs2) == SSA_NAME
&& has_single_use (rhs2)
&& (def1 = SSA_NAME_DEF_STMT (rhs),
gimple_assign_load_p (def1))
&& (def2 = SSA_NAME_DEF_STMT (rhs2),
gimple_assign_load_p (def2))
&& (gimple_vuse (def1) == gimple_vuse (def2))
&& !gimple_has_volatile_ops (def1)
&& !gimple_has_volatile_ops (def2)
&& !stmt_can_throw_internal (fun, def1)
&& !stmt_can_throw_internal (fun, def2)
&& gimple_assign_rhs_code (def1) == REALPART_EXPR
&& gimple_assign_rhs_code (def2) == IMAGPART_EXPR
&& operand_equal_p (TREE_OPERAND (gimple_assign_rhs1
(def1), 0),
TREE_OPERAND (gimple_assign_rhs1
(def2), 0)))
{
tree cl = TREE_OPERAND (gimple_assign_rhs1 (def1), 0);
gimple_assign_set_rhs_from_tree (&gsi, unshare_expr (cl));
gcc_assert (gsi_stmt (gsi) == stmt);
gimple_set_vuse (stmt, gimple_vuse (def1));
gimple_set_modified (stmt, true);
gimple_stmt_iterator gsi2 = gsi_for_stmt (def1);
gsi_remove (&gsi, false);
gsi_insert_after (&gsi2, stmt, GSI_SAME_STMT);
}
else
gsi_next (&gsi);
}
else if (code == CONSTRUCTOR
&& VECTOR_TYPE_P (TREE_TYPE (rhs))
&& TYPE_MODE (TREE_TYPE (rhs)) == BLKmode
&& CONSTRUCTOR_NELTS (rhs) > 0
&& (!VECTOR_TYPE_P (TREE_TYPE (CONSTRUCTOR_ELT (rhs, 0)->value))
|| (TYPE_MODE (TREE_TYPE (CONSTRUCTOR_ELT (rhs, 0)->value))
!= BLKmode)))
{
/* Rewrite stores of a single-use vector constructors
to component-wise stores if the mode isn't supported. */
use_operand_p use_p;
gimple *use_stmt;
if (single_imm_use (lhs, &use_p, &use_stmt)
&& gimple_store_p (use_stmt)
&& !gimple_has_volatile_ops (use_stmt)
&& !stmt_can_throw_internal (fun, use_stmt)
&& is_gimple_assign (use_stmt))
{
tree elt_t = TREE_TYPE (CONSTRUCTOR_ELT (rhs, 0)->value);
unsigned HOST_WIDE_INT elt_w
= tree_to_uhwi (TYPE_SIZE (elt_t));
unsigned HOST_WIDE_INT n
= tree_to_uhwi (TYPE_SIZE (TREE_TYPE (rhs)));
tree use_lhs = gimple_assign_lhs (use_stmt);
if (auto_var_p (use_lhs))
DECL_NOT_GIMPLE_REG_P (use_lhs) = 1;
else if (TREE_CODE (use_lhs) == TARGET_MEM_REF)
{
gimple_stmt_iterator gsi2 = gsi_for_stmt (use_stmt);
use_lhs = prepare_target_mem_ref_lvalue (use_lhs, &gsi2);
}
for (unsigned HOST_WIDE_INT bi = 0; bi < n; bi += elt_w)
{
unsigned HOST_WIDE_INT ci = bi / elt_w;
tree new_rhs;
if (ci < CONSTRUCTOR_NELTS (rhs))
new_rhs = CONSTRUCTOR_ELT (rhs, ci)->value;
else
new_rhs = build_zero_cst (elt_t);
tree new_lhs = build3 (BIT_FIELD_REF,
elt_t,
unshare_expr (use_lhs),
bitsize_int (elt_w),
bitsize_int (bi));
gimple *new_stmt = gimple_build_assign (new_lhs, new_rhs);
location_t loc = gimple_location (use_stmt);
gimple_set_location (new_stmt, loc);
gimple_set_vuse (new_stmt, gimple_vuse (use_stmt));
gimple_set_vdef (new_stmt,
make_ssa_name (gimple_vop (fun)));
SSA_NAME_DEF_STMT (gimple_vdef (new_stmt)) = new_stmt;
gimple_set_vuse (use_stmt, gimple_vdef (new_stmt));
gimple_stmt_iterator gsi2 = gsi_for_stmt (use_stmt);
gsi_insert_before (&gsi2, new_stmt, GSI_SAME_STMT);
}
gimple_stmt_iterator gsi2 = gsi_for_stmt (use_stmt);
unlink_stmt_vdef (use_stmt);
release_defs (use_stmt);
gsi_remove (&gsi2, true);
release_defs (stmt);
gsi_remove (&gsi, true);
}
else
gsi_next (&gsi);
}
else if (code == VEC_PERM_EXPR)
{
/* Find vectorized sequences, where we can reduce the lane
utilization. The narrowing will be donw later and only
if we find a pair of sequences that can be blended. */
gassign *assign = dyn_cast <gassign *> (stmt);
vec_perm_simplify_seq seq;
if (recognise_vec_perm_simplify_seq (assign, &seq))
append_vec_perm_simplify_seq_list (&vec_perm_simplify_seq_list,
seq);
gsi_next (&gsi);
}
else
gsi_next (&gsi);
}
process_vec_perm_simplify_seq_list (&vec_perm_simplify_seq_list);
/* Combine stmts with the stmts defining their operands.
Note we update GSI within the loop as necessary. */
for (gsi = gsi_start_bb (bb); !gsi_end_p (gsi); gsi_next (&gsi))
{
gimple *stmt = gsi_stmt (gsi);
/* Mark stmt as potentially needing revisiting. */
gimple_set_plf (stmt, GF_PLF_1, false);
bool can_make_abnormal_goto = (is_gimple_call (stmt)
&& stmt_can_make_abnormal_goto (stmt));
/* Substitute from our lattice. We need to do so only once. */
bool substituted_p = false;
use_operand_p usep;
ssa_op_iter iter;
FOR_EACH_SSA_USE_OPERAND (usep, stmt, iter, SSA_OP_USE)
{
tree use = USE_FROM_PTR (usep);
tree val = fwprop_ssa_val (use);
if (val && val != use)
{
if (!is_gimple_debug (stmt))
bitmap_set_bit (simple_dce_worklist, SSA_NAME_VERSION (use));
if (may_propagate_copy (use, val))
{
propagate_value (usep, val);
substituted_p = true;
}
}
}
if (substituted_p
&& is_gimple_assign (stmt)
&& gimple_assign_rhs_code (stmt) == ADDR_EXPR)
recompute_tree_invariant_for_addr_expr (gimple_assign_rhs1 (stmt));
if (substituted_p
&& can_make_abnormal_goto
&& !stmt_can_make_abnormal_goto (stmt))
bitmap_set_bit (need_ab_cleanup, bb->index);
bool changed;
do
{
gimple *orig_stmt = stmt = gsi_stmt (gsi);
bool was_noreturn = (is_gimple_call (stmt)
&& gimple_call_noreturn_p (stmt));
changed = false;
auto_vec<tree, 8> uses;
FOR_EACH_SSA_USE_OPERAND (usep, stmt, iter, SSA_OP_USE)
if (uses.space (1))
uses.quick_push (USE_FROM_PTR (usep));
if (fold_stmt (&gsi, fwprop_ssa_val, simple_dce_worklist))
{
changed = true;
stmt = gsi_stmt (gsi);
/* Cleanup the CFG if we simplified a condition to
true or false. */
if (gcond *cond = dyn_cast <gcond *> (stmt))
if (gimple_cond_true_p (cond)
|| gimple_cond_false_p (cond))
cfg_changed = true;
/* Queue old uses for simple DCE if not debug statement. */
if (!is_gimple_debug (stmt))
for (tree use : uses)
if (TREE_CODE (use) == SSA_NAME
&& !SSA_NAME_IS_DEFAULT_DEF (use))
bitmap_set_bit (simple_dce_worklist,
SSA_NAME_VERSION (use));
}
if (changed || substituted_p)
{
if (maybe_clean_or_replace_eh_stmt (orig_stmt, stmt))
bitmap_set_bit (to_purge, bb->index);
if (!was_noreturn
&& is_gimple_call (stmt) && gimple_call_noreturn_p (stmt))
to_fixup.safe_push (stmt);
update_stmt (stmt);
substituted_p = false;
}
switch (gimple_code (stmt))
{
case GIMPLE_ASSIGN:
{
tree rhs1 = gimple_assign_rhs1 (stmt);
enum tree_code code = gimple_assign_rhs_code (stmt);
if (TREE_CODE_CLASS (code) == tcc_comparison)
{
int did_something;
did_something = forward_propagate_into_comparison (&gsi);
if (maybe_clean_or_replace_eh_stmt (stmt, gsi_stmt (gsi)))
bitmap_set_bit (to_purge, bb->index);
if (did_something == 2)
cfg_changed = true;
changed = did_something != 0;
}
else if ((code == PLUS_EXPR
|| code == BIT_IOR_EXPR
|| code == BIT_XOR_EXPR)
&& simplify_rotate (&gsi))
changed = true;
else if (code == VEC_PERM_EXPR)
{
int did_something = simplify_permutation (&gsi);
if (did_something == 2)
cfg_changed = true;
changed = did_something != 0;
}
else if (code == BIT_FIELD_REF)
changed = simplify_bitfield_ref (&gsi);
else if (code == CONSTRUCTOR
&& TREE_CODE (TREE_TYPE (rhs1)) == VECTOR_TYPE)
changed = simplify_vector_constructor (&gsi);
else if (code == ARRAY_REF)
changed = simplify_count_trailing_zeroes (&gsi);
break;
}
case GIMPLE_SWITCH:
changed = simplify_gimple_switch (as_a <gswitch *> (stmt),
edges_to_remove);
break;
case GIMPLE_COND:
{
int did_something = forward_propagate_into_gimple_cond
(as_a <gcond *> (stmt));
if (did_something == 2)
cfg_changed = true;
changed = did_something != 0;
break;
}
case GIMPLE_CALL:
{
tree callee = gimple_call_fndecl (stmt);
if (callee != NULL_TREE
&& fndecl_built_in_p (callee, BUILT_IN_NORMAL))
changed = simplify_builtin_call (&gsi, callee);
break;
}
default:;
}
if (changed)
{
/* If the stmt changed then re-visit it and the statements
inserted before it. */
for (; !gsi_end_p (gsi); gsi_prev (&gsi))
if (gimple_plf (gsi_stmt (gsi), GF_PLF_1))
break;
if (gsi_end_p (gsi))
gsi = gsi_start_bb (bb);
else
gsi_next (&gsi);
}
}
while (changed);
/* Stmt no longer needs to be revisited. */
stmt = gsi_stmt (gsi);
gcc_checking_assert (!gimple_plf (stmt, GF_PLF_1));
gimple_set_plf (stmt, GF_PLF_1, true);
/* Fill up the lattice. */
if (gimple_assign_single_p (stmt))
{
tree lhs = gimple_assign_lhs (stmt);
tree rhs = gimple_assign_rhs1 (stmt);
if (TREE_CODE (lhs) == SSA_NAME)
{
tree val = lhs;
if (TREE_CODE (rhs) == SSA_NAME)
val = fwprop_ssa_val (rhs);
else if (is_gimple_min_invariant (rhs))
val = rhs;
/* If we can propagate the lattice-value mark the
stmt for removal. */
if (val != lhs
&& may_propagate_copy (lhs, val))
to_remove_defs.safe_push (SSA_NAME_VERSION (lhs));
fwprop_set_lattice_val (lhs, val);
}
}
else if (gimple_nop_p (stmt))
to_remove.safe_push (stmt);
}
/* Substitute in destination PHI arguments. */
FOR_EACH_EDGE (e, ei, bb->succs)
for (gphi_iterator gsi = gsi_start_phis (e->dest);
!gsi_end_p (gsi); gsi_next (&gsi))
{
gphi *phi = gsi.phi ();
use_operand_p use_p = PHI_ARG_DEF_PTR_FROM_EDGE (phi, e);
tree arg = USE_FROM_PTR (use_p);
if (TREE_CODE (arg) != SSA_NAME
|| virtual_operand_p (arg))
continue;
tree val = fwprop_ssa_val (arg);
if (val != arg
&& may_propagate_copy (arg, val, !(e->flags & EDGE_ABNORMAL)))
propagate_value (use_p, val);
}
/* Mark outgoing exectuable edges. */
if (edge e = find_taken_edge (bb, NULL))
{
e->flags |= EDGE_EXECUTABLE;
if (EDGE_COUNT (bb->succs) > 1)
cfg_changed = true;
}
else
{
FOR_EACH_EDGE (e, ei, bb->succs)
e->flags |= EDGE_EXECUTABLE;
}
}
free (postorder);
free (bb_to_rpo);
lattice.release ();
/* First remove chains of stmts where we check no uses remain. */
simple_dce_from_worklist (simple_dce_worklist, to_purge);
auto remove = [](gimple *stmt)
{
if (dump_file && (dump_flags & TDF_DETAILS))
{
fprintf (dump_file, "Removing dead stmt ");
print_gimple_stmt (dump_file, stmt, 0);
fprintf (dump_file, "\n");
}
gimple_stmt_iterator gsi = gsi_for_stmt (stmt);
if (gimple_code (stmt) == GIMPLE_PHI)
remove_phi_node (&gsi, true);
else
{
unlink_stmt_vdef (stmt);
gsi_remove (&gsi, true);
release_defs (stmt);
}
};
/* Then remove stmts we know we can remove even though we did not
substitute in dead code regions, so uses can remain. Do so in reverse
order to make debug stmt creation possible. */
while (!to_remove_defs.is_empty())
{
tree def = ssa_name (to_remove_defs.pop ());
/* For example remove_prop_source_from_use can remove stmts queued
for removal. Deal with this gracefully. */
if (!def)
continue;
gimple *stmt = SSA_NAME_DEF_STMT (def);
remove (stmt);
}
/* Wipe other queued stmts that do not have SSA defs. */
while (!to_remove.is_empty())
{
gimple *stmt = to_remove.pop ();
remove (stmt);
}
/* Fixup stmts that became noreturn calls. This may require splitting
blocks and thus isn't possible during the walk. Do this
in reverse order so we don't inadvertedly remove a stmt we want to
fixup by visiting a dominating now noreturn call first. */
while (!to_fixup.is_empty ())
{
gimple *stmt = to_fixup.pop ();
if (dump_file && dump_flags & TDF_DETAILS)
{
fprintf (dump_file, "Fixing up noreturn call ");
print_gimple_stmt (dump_file, stmt, 0);
fprintf (dump_file, "\n");
}
cfg_changed |= fixup_noreturn_call (stmt);
}
cfg_changed |= gimple_purge_all_dead_eh_edges (to_purge);
cfg_changed |= gimple_purge_all_dead_abnormal_call_edges (need_ab_cleanup);
BITMAP_FREE (to_purge);
/* Remove edges queued from switch stmt simplification. */
for (auto ep : edges_to_remove)
{
basic_block src = BASIC_BLOCK_FOR_FN (fun, ep.first);
basic_block dest = BASIC_BLOCK_FOR_FN (fun, ep.second);
edge e;
if (src && dest && (e = find_edge (src, dest)))
{
free_dominance_info (CDI_DOMINATORS);
remove_edge (e);
cfg_changed = true;
}
}
if (get_range_query (fun) != get_global_range_query ())
disable_ranger (fun);
if (cfg_changed)
todoflags |= TODO_cleanup_cfg;
return todoflags;
}
} // anon namespace
gimple_opt_pass *
make_pass_forwprop (gcc::context *ctxt)
{
return new pass_forwprop (ctxt);
}
Reference in New Issue View Git Blame Copy Permalink