* [PATCH 1/2] tree-optimization/105142 - improve maybe_fold_comparisons_from_match_pd fix
@ 2022-07-26 13:41 Richard Biener
0 siblings, 0 replies; 3+ messages in thread
From: Richard Biener @ 2022-07-26 13:41 UTC (permalink / raw)
To: gcc-patches
The following improves on the fix for PR105142 which restricted the
expression lookup used for maybe_fold_comparisons_from_match_pd to
avoid picking up flow-sensitive info for use in places where guarding
conditions do not hold. Instead of not allowing to expand SSA
definitions there the following temporarily clears flow-sensitive
info on the SSA names and restores it when finished matching.
Bootstrapped and tested on x86_64-unknown-linux-gnu.
PR tree-optimization/105142
* gimple-fold.cc (fosa_unwind): New global.
(follow_outer_ssa_edges): When the SSA definition to follow
is does not dominate fosa_bb, temporarily clear flow-sensitive
info. Make sure to not expand stmts with not defined overflow.
(maybe_fold_comparisons_from_match_pd): Set up unwind stack
for follow_outer_ssa_edges and unwind flow-sensitive info
clearing after matching.
---
gcc/gimple-fold.cc | 25 ++++++++++++++++++++++++-
1 file changed, 24 insertions(+), 1 deletion(-)
diff --git a/gcc/gimple-fold.cc b/gcc/gimple-fold.cc
index a1704784bc9..14640f6952a 100644
--- a/gcc/gimple-fold.cc
+++ b/gcc/gimple-fold.cc
@@ -6886,6 +6886,7 @@ and_comparisons_1 (tree type, enum tree_code code1, tree op1a, tree op1b,
}
static basic_block fosa_bb;
+static vec<std::pair<tree, void *> > *fosa_unwind;
static tree
follow_outer_ssa_edges (tree val)
{
@@ -6899,7 +6900,21 @@ follow_outer_ssa_edges (tree val)
&& (def_bb == fosa_bb
|| dominated_by_p (CDI_DOMINATORS, fosa_bb, def_bb))))
return val;
- return NULL_TREE;
+ /* We cannot temporarily rewrite stmts with undefined overflow
+ behavior, so avoid expanding them. */
+ if ((ANY_INTEGRAL_TYPE_P (TREE_TYPE (val))
+ || POINTER_TYPE_P (TREE_TYPE (val)))
+ && !TYPE_OVERFLOW_WRAPS (TREE_TYPE (val)))
+ return NULL_TREE;
+ /* If the definition does not dominate fosa_bb temporarily reset
+ flow-sensitive info. */
+ if (val->ssa_name.info.range_info)
+ {
+ fosa_unwind->safe_push (std::make_pair
+ (val, val->ssa_name.info.range_info));
+ val->ssa_name.info.range_info = NULL;
+ }
+ return val;
}
return val;
}
@@ -6958,9 +6973,14 @@ maybe_fold_comparisons_from_match_pd (tree type, enum tree_code code,
type, gimple_assign_lhs (stmt1),
gimple_assign_lhs (stmt2));
fosa_bb = outer_cond_bb;
+ auto_vec<std::pair<tree, void *>, 8> unwind_stack;
+ fosa_unwind = &unwind_stack;
if (op.resimplify (NULL, (!outer_cond_bb
? follow_all_ssa_edges : follow_outer_ssa_edges)))
{
+ fosa_unwind = NULL;
+ for (auto p : unwind_stack)
+ p.first->ssa_name.info.range_info = p.second;
if (gimple_simplified_result_is_gimple_val (&op))
{
tree res = op.ops[0];
@@ -6982,6 +7002,9 @@ maybe_fold_comparisons_from_match_pd (tree type, enum tree_code code,
return build2 ((enum tree_code)op.code, op.type, op0, op1);
}
}
+ fosa_unwind = NULL;
+ for (auto p : unwind_stack)
+ p.first->ssa_name.info.range_info = p.second;
return NULL_TREE;
}
--
2.35.3
^ permalink raw reply [flat|nested] 3+ messages in thread
* [PATCH 1/2] tree-optimization/105142 - improve maybe_fold_comparisons_from_match_pd fix
@ 2022-07-26 13:41 Richard Biener
0 siblings, 0 replies; 3+ messages in thread
From: Richard Biener @ 2022-07-26 13:41 UTC (permalink / raw)
To: gcc-patches
The following improves on the fix for PR105142 which restricted the
expression lookup used for maybe_fold_comparisons_from_match_pd to
avoid picking up flow-sensitive info for use in places where guarding
conditions do not hold. Instead of not allowing to expand SSA
definitions there the following temporarily clears flow-sensitive
info on the SSA names and restores it when finished matching.
Bootstrapped and tested on x86_64-unknown-linux-gnu.
PR tree-optimization/105142
* gimple-fold.cc (fosa_unwind): New global.
(follow_outer_ssa_edges): When the SSA definition to follow
is does not dominate fosa_bb, temporarily clear flow-sensitive
info. Make sure to not expand stmts with not defined overflow.
(maybe_fold_comparisons_from_match_pd): Set up unwind stack
for follow_outer_ssa_edges and unwind flow-sensitive info
clearing after matching.
---
gcc/gimple-fold.cc | 25 ++++++++++++++++++++++++-
1 file changed, 24 insertions(+), 1 deletion(-)
diff --git a/gcc/gimple-fold.cc b/gcc/gimple-fold.cc
index a1704784bc9..14640f6952a 100644
--- a/gcc/gimple-fold.cc
+++ b/gcc/gimple-fold.cc
@@ -6886,6 +6886,7 @@ and_comparisons_1 (tree type, enum tree_code code1, tree op1a, tree op1b,
}
static basic_block fosa_bb;
+static vec<std::pair<tree, void *> > *fosa_unwind;
static tree
follow_outer_ssa_edges (tree val)
{
@@ -6899,7 +6900,21 @@ follow_outer_ssa_edges (tree val)
&& (def_bb == fosa_bb
|| dominated_by_p (CDI_DOMINATORS, fosa_bb, def_bb))))
return val;
- return NULL_TREE;
+ /* We cannot temporarily rewrite stmts with undefined overflow
+ behavior, so avoid expanding them. */
+ if ((ANY_INTEGRAL_TYPE_P (TREE_TYPE (val))
+ || POINTER_TYPE_P (TREE_TYPE (val)))
+ && !TYPE_OVERFLOW_WRAPS (TREE_TYPE (val)))
+ return NULL_TREE;
+ /* If the definition does not dominate fosa_bb temporarily reset
+ flow-sensitive info. */
+ if (val->ssa_name.info.range_info)
+ {
+ fosa_unwind->safe_push (std::make_pair
+ (val, val->ssa_name.info.range_info));
+ val->ssa_name.info.range_info = NULL;
+ }
+ return val;
}
return val;
}
@@ -6958,9 +6973,14 @@ maybe_fold_comparisons_from_match_pd (tree type, enum tree_code code,
type, gimple_assign_lhs (stmt1),
gimple_assign_lhs (stmt2));
fosa_bb = outer_cond_bb;
+ auto_vec<std::pair<tree, void *>, 8> unwind_stack;
+ fosa_unwind = &unwind_stack;
if (op.resimplify (NULL, (!outer_cond_bb
? follow_all_ssa_edges : follow_outer_ssa_edges)))
{
+ fosa_unwind = NULL;
+ for (auto p : unwind_stack)
+ p.first->ssa_name.info.range_info = p.second;
if (gimple_simplified_result_is_gimple_val (&op))
{
tree res = op.ops[0];
@@ -6982,6 +7002,9 @@ maybe_fold_comparisons_from_match_pd (tree type, enum tree_code code,
return build2 ((enum tree_code)op.code, op.type, op0, op1);
}
}
+ fosa_unwind = NULL;
+ for (auto p : unwind_stack)
+ p.first->ssa_name.info.range_info = p.second;
return NULL_TREE;
}
--
2.35.3
^ permalink raw reply [flat|nested] 3+ messages in thread
* [PATCH 1/2] tree-optimization/105142 - improve maybe_fold_comparisons_from_match_pd fix
@ 2022-07-26 13:41 Richard Biener
0 siblings, 0 replies; 3+ messages in thread
From: Richard Biener @ 2022-07-26 13:41 UTC (permalink / raw)
To: gcc-patches
The following improves on the fix for PR105142 which restricted the
expression lookup used for maybe_fold_comparisons_from_match_pd to
avoid picking up flow-sensitive info for use in places where guarding
conditions do not hold. Instead of not allowing to expand SSA
definitions there the following temporarily clears flow-sensitive
info on the SSA names and restores it when finished matching.
Bootstrapped and tested on x86_64-unknown-linux-gnu.
PR tree-optimization/105142
* gimple-fold.cc (fosa_unwind): New global.
(follow_outer_ssa_edges): When the SSA definition to follow
is does not dominate fosa_bb, temporarily clear flow-sensitive
info. Make sure to not expand stmts with not defined overflow.
(maybe_fold_comparisons_from_match_pd): Set up unwind stack
for follow_outer_ssa_edges and unwind flow-sensitive info
clearing after matching.
---
gcc/gimple-fold.cc | 25 ++++++++++++++++++++++++-
1 file changed, 24 insertions(+), 1 deletion(-)
diff --git a/gcc/gimple-fold.cc b/gcc/gimple-fold.cc
index a1704784bc9..14640f6952a 100644
--- a/gcc/gimple-fold.cc
+++ b/gcc/gimple-fold.cc
@@ -6886,6 +6886,7 @@ and_comparisons_1 (tree type, enum tree_code code1, tree op1a, tree op1b,
}
static basic_block fosa_bb;
+static vec<std::pair<tree, void *> > *fosa_unwind;
static tree
follow_outer_ssa_edges (tree val)
{
@@ -6899,7 +6900,21 @@ follow_outer_ssa_edges (tree val)
&& (def_bb == fosa_bb
|| dominated_by_p (CDI_DOMINATORS, fosa_bb, def_bb))))
return val;
- return NULL_TREE;
+ /* We cannot temporarily rewrite stmts with undefined overflow
+ behavior, so avoid expanding them. */
+ if ((ANY_INTEGRAL_TYPE_P (TREE_TYPE (val))
+ || POINTER_TYPE_P (TREE_TYPE (val)))
+ && !TYPE_OVERFLOW_WRAPS (TREE_TYPE (val)))
+ return NULL_TREE;
+ /* If the definition does not dominate fosa_bb temporarily reset
+ flow-sensitive info. */
+ if (val->ssa_name.info.range_info)
+ {
+ fosa_unwind->safe_push (std::make_pair
+ (val, val->ssa_name.info.range_info));
+ val->ssa_name.info.range_info = NULL;
+ }
+ return val;
}
return val;
}
@@ -6958,9 +6973,14 @@ maybe_fold_comparisons_from_match_pd (tree type, enum tree_code code,
type, gimple_assign_lhs (stmt1),
gimple_assign_lhs (stmt2));
fosa_bb = outer_cond_bb;
+ auto_vec<std::pair<tree, void *>, 8> unwind_stack;
+ fosa_unwind = &unwind_stack;
if (op.resimplify (NULL, (!outer_cond_bb
? follow_all_ssa_edges : follow_outer_ssa_edges)))
{
+ fosa_unwind = NULL;
+ for (auto p : unwind_stack)
+ p.first->ssa_name.info.range_info = p.second;
if (gimple_simplified_result_is_gimple_val (&op))
{
tree res = op.ops[0];
@@ -6982,6 +7002,9 @@ maybe_fold_comparisons_from_match_pd (tree type, enum tree_code code,
return build2 ((enum tree_code)op.code, op.type, op0, op1);
}
}
+ fosa_unwind = NULL;
+ for (auto p : unwind_stack)
+ p.first->ssa_name.info.range_info = p.second;
return NULL_TREE;
}
--
2.35.3
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2022-07-26 13:41 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-07-26 13:41 [PATCH 1/2] tree-optimization/105142 - improve maybe_fold_comparisons_from_match_pd fix Richard Biener
2022-07-26 13:41 Richard Biener
2022-07-26 13:41 Richard Biener
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).