Hi,
This patch refactors the VRP edge-assertion code to make it
unconditionally traverse SSA-name definitions in order to find suitable
edge assertions to insert. Currently SSA-name definitions get traversed
only when the orginal conditional is a bitwise AND or OR operation,
which is a strange restriction.
At the same time, I made a few of the tests for detecting suitable edge
assertions more generic. Conditionals from which new edge assertions
are deduced include:
if ((x & y) == C) // edge assertions: x != 0, y != 0
if (~x == C) // edge assertion: x == ~C
char y;
...
int x = y;
if (x == C) // edge assertion: y == (char)C
int y;
char x = y;
if (x != C) // edge assertion: y != (int)C
unsigned y;
...
unsigned long x = y;
if (x >= C) // edge assertion: y >= (unsigned)C
int y;
...
unsigned int x = y;
if (x == C) // edge assertion: y == (int)C
int p = q >= r;
if (!p) // edge assertion: q < r
... and any arbitrary combination of such patterns.
I also added the original testcase that motivated me to extend and
ultimately rewrite this code.
I bootstrapped and regtested an earlier version of this patch on
x86_64-unknown-linux-gnu though I will do so again tonight to make sure.
2014-05-03 Patrick Palka <patr...@parcs.ath.cx>
* tree-vrp.c (extract_code_and_val_from_cond_with_ops): Add
sanity test.
(register_edge_assert_for, register_edge_assert_for_1,
register_edge_assert_for_2): Refactor and consolidate
edge-assertion logic into ...
(register_edge_assert_for_2): ... here. Add LIMIT parameter.
Rename to ...
(register_edge_assert_for_1): ... this.
---
gcc/testsuite/gcc.dg/vrp-1.c | 21 ++++
gcc/tree-vrp.c | 252 +++++++++++++++++++------------------------
2 files changed, 134 insertions(+), 139 deletions(-)
create mode 100644 gcc/testsuite/gcc.dg/vrp-1.c
diff --git a/gcc/testsuite/gcc.dg/vrp-1.c b/gcc/testsuite/gcc.dg/vrp-1.c
new file mode 100644
index 0000000..6e1c2f1
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/vrp-1.c
@@ -0,0 +1,21 @@
+// { dg-options "-O2" }
+
+void runtime_error (void) __attribute__ ((noreturn));
+void compiletime_error (void) __attribute__ ((noreturn, error ("")));
+
+static void
+check_equals (int *x, int y)
+{
+ int __p = *x != y;
+ if (__builtin_constant_p (__p) && __p)
+ compiletime_error (); // { dg-error "call to" }
+ if (__p)
+ runtime_error ();
+}
+
+void
+foo (int *x)
+{
+ check_equals (x, 5);
+ check_equals (x, 10);
+}
diff --git a/gcc/tree-vrp.c b/gcc/tree-vrp.c
index 806221c..0d18b42 100644
--- a/gcc/tree-vrp.c
+++ b/gcc/tree-vrp.c
@@ -4721,13 +4721,15 @@ extract_code_and_val_from_cond_with_ops (tree name,
enum tree_code cond_code,
comp_code = swap_tree_comparison (cond_code);
val = cond_op0;
}
- else
+ else if (name == cond_op0)
{
/* The comparison is of the form NAME COMP VAL, so the
comparison code remains unchanged. */
comp_code = cond_code;
val = cond_op1;
}
+ else
+ gcc_unreachable ();
/* Invert the comparison code as necessary. */
if (invert)
@@ -4791,19 +4793,33 @@ masked_increment (double_int val, double_int mask,
double_int sgnbit,
}
/* Try to register an edge assertion for SSA name NAME on edge E for
- the condition COND contributing to the conditional jump pointed to by BSI.
+ the condition COND (composed of COND_CODE, COND_OP0 and COND_OP1)
+ contributing to the conditional jump pointed to by BSI.
+
+ Further, try to recursively register edge assertions for the SSA names in
+ the defining statements of COND's operands. This recursion is limited by
+ LIMIT.
+
Invert the condition COND if INVERT is true.
- Return true if an assertion for NAME could be registered. */
+ Return true if an assertion is registered. */
static bool
-register_edge_assert_for_2 (tree name, edge e, gimple_stmt_iterator bsi,
- enum tree_code cond_code,
+register_edge_assert_for_1 (tree name, edge e, gimple_stmt_iterator bsi,
+ unsigned int limit, enum tree_code cond_code,
tree cond_op0, tree cond_op1, bool invert)
{
tree val;
- enum tree_code comp_code;
+ enum tree_code comp_code, def_rhs_code;
+ gimple def_stmt;
bool retval = false;
+ if (limit == 0)
+ return false;
+
+ /* We only care about SSA_NAMEs. */
+ if (TREE_CODE (name) != SSA_NAME)
+ return false;
+
if (!extract_code_and_val_from_cond_with_ops (name, cond_code,
cond_op0,
cond_op1,
@@ -5348,99 +5364,110 @@ register_edge_assert_for_2 (tree name, edge e,
gimple_stmt_iterator bsi,
}
}
- return retval;
-}
+ /* If COND is effectively an equality test of an SSA_NAME against
+ the value zero or one, then we may be able to assert values
+ for SSA_NAMEs which flow into COND. */
-/* OP is an operand of a truth value expression which is known to have
- a particular value. Register any asserts for OP and for any
- operands in OP's defining statement.
-
- If CODE is EQ_EXPR, then we want to register OP is zero (false),
- if CODE is NE_EXPR, then we want to register OP is nonzero (true). */
-
-static bool
-register_edge_assert_for_1 (tree op, enum tree_code code,
- edge e, gimple_stmt_iterator bsi)
-{
- bool retval = false;
- gimple op_def;
- tree val;
- enum tree_code rhs_code;
-
- /* We only care about SSA_NAMEs. */
- if (TREE_CODE (op) != SSA_NAME)
- return false;
-
- /* We know that OP will have a zero or nonzero value. If OP is used
- more than once go ahead and register an assert for OP. */
- if (live_on_edge (e, op)
- && !has_single_use (op))
- {
- val = build_int_cst (TREE_TYPE (op), 0);
- register_new_assert_for (op, op, code, val, NULL, e, bsi);
- retval = true;
- }
-
- /* Now look at how OP is set. If it's set from a comparison,
- a truth operation or some bit operations, then we may be able
- to register information about the operands of that assignment. */
- op_def = SSA_NAME_DEF_STMT (op);
- if (gimple_code (op_def) != GIMPLE_ASSIGN)
+ def_stmt = SSA_NAME_DEF_STMT (name);
+ if (!is_gimple_assign (def_stmt))
return retval;
- rhs_code = gimple_assign_rhs_code (op_def);
+ def_rhs_code = gimple_assign_rhs_code (def_stmt);
- if (TREE_CODE_CLASS (rhs_code) == tcc_comparison)
+ /* In the case of NAME == C != 0 or NAME != 0, for BIT_AND_EXPR defining
+ statement of NAME we can assert that both operands of the BIT_AND_EXPR
+ have nonzero value. */
+ if (def_rhs_code == BIT_AND_EXPR
+ && ((comp_code == NE_EXPR && integer_zerop (val))
+ || (comp_code == EQ_EXPR && TREE_CODE (val) == INTEGER_CST
+ && integer_nonzerop (val))))
{
- bool invert = (code == EQ_EXPR ? true : false);
- tree op0 = gimple_assign_rhs1 (op_def);
- tree op1 = gimple_assign_rhs2 (op_def);
+ tree op0 = gimple_assign_rhs1 (def_stmt);
+ tree op1 = gimple_assign_rhs2 (def_stmt);
+ tree zero = build_zero_cst (TREE_TYPE (val));
+ retval |= register_edge_assert_for_1 (op0, e, bsi, limit - 1,
+ NE_EXPR, op0, zero, false);
+ retval |= register_edge_assert_for_1 (op1, e, bsi, limit - 1,
+ NE_EXPR, op1, zero, false);
+ }
- if (TREE_CODE (op0) == SSA_NAME)
- retval |= register_edge_assert_for_2 (op0, e, bsi, rhs_code, op0, op1,
- invert);
- if (TREE_CODE (op1) == SSA_NAME)
- retval |= register_edge_assert_for_2 (op1, e, bsi, rhs_code, op0, op1,
- invert);
+ /* In the case of NAME == 0 or NAME != 1, for BIT_IOR_EXPR defining
+ statement of NAME we can assert that both operands of the BIT_IOR_EXPR
+ have zero value. */
+ if (def_rhs_code == BIT_IOR_EXPR
+ && ((comp_code == EQ_EXPR && integer_zerop (val))
+ || (comp_code == NE_EXPR && integer_onep (val)
+ && TYPE_PRECISION (TREE_TYPE (name)) == 1)))
+ {
+ /* For BIT_IOR_EXPR only if NAME == 0 both operands have
+ necessarily zero value, or if type-precision is one. */
+ tree op0 = gimple_assign_rhs1 (def_stmt);
+ tree op1 = gimple_assign_rhs2 (def_stmt);
+ tree zero = build_zero_cst (TREE_TYPE (val));
+ retval |= register_edge_assert_for_1 (op0, e, bsi, limit - 1,
+ EQ_EXPR, op0, zero, false);
+ retval |= register_edge_assert_for_1 (op1, e, bsi, limit - 1,
+ EQ_EXPR, op1, zero, false);
}
- else if ((code == NE_EXPR
- && gimple_assign_rhs_code (op_def) == BIT_AND_EXPR)
- || (code == EQ_EXPR
- && gimple_assign_rhs_code (op_def) == BIT_IOR_EXPR))
+
+ if (def_rhs_code == BIT_NOT_EXPR
+ && (comp_code == EQ_EXPR || comp_code == NE_EXPR)
+ && TREE_CODE (val) == INTEGER_CST)
{
- /* Recurse on each operand. */
- tree op0 = gimple_assign_rhs1 (op_def);
- tree op1 = gimple_assign_rhs2 (op_def);
- if (TREE_CODE (op0) == SSA_NAME
- && has_single_use (op0))
- retval |= register_edge_assert_for_1 (op0, code, e, bsi);
- if (TREE_CODE (op1) == SSA_NAME
- && has_single_use (op1))
- retval |= register_edge_assert_for_1 (op1, code, e, bsi);
+ /* Recurse, inverting VAL. */
+ tree rhs = gimple_assign_rhs1 (def_stmt);
+ tree new_val = fold_build1 (BIT_NOT_EXPR, TREE_TYPE (val), val);
+ retval |= register_edge_assert_for_1 (rhs, e, bsi, limit - 1,
+ comp_code, rhs, new_val, false);
}
- else if (gimple_assign_rhs_code (op_def) == BIT_NOT_EXPR
- && TYPE_PRECISION (TREE_TYPE (gimple_assign_lhs (op_def))) == 1)
+
+ /* In the case of NAME == [01] or NAME != [01], if NAME's defining statement
+ is a TCC_COMPARISON then we can assert the comparison or its negation. */
+ if (TREE_CODE_CLASS (def_rhs_code) == tcc_comparison
+ && (comp_code == EQ_EXPR || comp_code == NE_EXPR)
+ && (integer_zerop (val) || integer_onep (val)))
{
- /* Recurse, flipping CODE. */
- code = invert_tree_comparison (code, false);
- retval |= register_edge_assert_for_1 (gimple_assign_rhs1 (op_def),
- code, e, bsi);
+ tree op0 = gimple_assign_rhs1 (def_stmt);
+ tree op1 = gimple_assign_rhs2 (def_stmt);
+ bool invert = false;
+
+ if ((comp_code == EQ_EXPR && integer_zerop (val))
+ || (comp_code == NE_EXPR && integer_onep (val)))
+ invert = true;
+ retval |= register_edge_assert_for_1 (op0, e, bsi, limit - 1,
+ def_rhs_code, op0, op1, invert);
+ retval |= register_edge_assert_for_1 (op1, e, bsi, limit - 1,
+ def_rhs_code, op0, op1, invert);
}
- else if (gimple_assign_rhs_code (op_def) == SSA_NAME)
+
+ if (def_rhs_code == SSA_NAME)
{
/* Recurse through the copy. */
- retval |= register_edge_assert_for_1 (gimple_assign_rhs1 (op_def),
- code, e, bsi);
+ tree rhs = gimple_assign_rhs1 (def_stmt);
+ retval |= register_edge_assert_for_1 (rhs, e, bsi, limit - 1,
+ comp_code, rhs, val, false);
}
- else if (CONVERT_EXPR_CODE_P (gimple_assign_rhs_code (op_def)))
+
+ if (CONVERT_EXPR_CODE_P (def_rhs_code)
+ && TREE_CODE (val) == INTEGER_CST)
{
- /* Recurse through the type conversion, unless it is a narrowing
- conversion or conversion from non-integral type. */
- tree rhs = gimple_assign_rhs1 (op_def);
+ /* Recurse through the type conversion if sensible. */
+ tree rhs = gimple_assign_rhs1 (def_stmt);
if (INTEGRAL_TYPE_P (TREE_TYPE (rhs))
- && (TYPE_PRECISION (TREE_TYPE (rhs))
- <= TYPE_PRECISION (TREE_TYPE (op))))
- retval |= register_edge_assert_for_1 (rhs, code, e, bsi);
+ && ((comp_code == EQ_EXPR
+ && TYPE_PRECISION (TREE_TYPE (name))
+ >= TYPE_PRECISION (TREE_TYPE (rhs)))
+ || (comp_code == NE_EXPR
+ && TYPE_PRECISION (TREE_TYPE (name))
+ <= TYPE_PRECISION (TREE_TYPE (rhs)))
+ || ((comp_code == GT_EXPR || comp_code == GE_EXPR)
+ && TYPE_UNSIGNED (TREE_TYPE (rhs))
+ && TYPE_UNSIGNED (TREE_TYPE (val)))))
+ {
+ tree new_val = fold_build1 (CONVERT_EXPR, TREE_TYPE (rhs), val);
+ retval |= register_edge_assert_for_1 (rhs, e, bsi, limit - 1,
+ comp_code, rhs, new_val, false);
+ }
}
return retval;
@@ -5455,9 +5482,7 @@ register_edge_assert_for (tree name, edge e,
gimple_stmt_iterator si,
enum tree_code cond_code, tree cond_op0,
tree cond_op1)
{
- tree val;
- enum tree_code comp_code;
- bool retval = false;
+ bool retval;
bool is_else_edge = (e->flags & EDGE_FALSE_VALUE) != 0;
/* Do not attempt to infer anything in names that flow through
@@ -5465,60 +5490,9 @@ register_edge_assert_for (tree name, edge e,
gimple_stmt_iterator si,
if (SSA_NAME_OCCURS_IN_ABNORMAL_PHI (name))
return false;
- if (!extract_code_and_val_from_cond_with_ops (name, cond_code,
- cond_op0, cond_op1,
- is_else_edge,
- &comp_code, &val))
- return false;
-
/* Register ASSERT_EXPRs for name. */
- retval |= register_edge_assert_for_2 (name, e, si, cond_code, cond_op0,
- cond_op1, is_else_edge);
-
-
- /* If COND is effectively an equality test of an SSA_NAME against
- the value zero or one, then we may be able to assert values
- for SSA_NAMEs which flow into COND. */
-
- /* In the case of NAME == 1 or NAME != 0, for BIT_AND_EXPR defining
- statement of NAME we can assert both operands of the BIT_AND_EXPR
- have nonzero value. */
- if (((comp_code == EQ_EXPR && integer_onep (val))
- || (comp_code == NE_EXPR && integer_zerop (val))))
- {
- gimple def_stmt = SSA_NAME_DEF_STMT (name);
-
- if (is_gimple_assign (def_stmt)
- && gimple_assign_rhs_code (def_stmt) == BIT_AND_EXPR)
- {
- tree op0 = gimple_assign_rhs1 (def_stmt);
- tree op1 = gimple_assign_rhs2 (def_stmt);
- retval |= register_edge_assert_for_1 (op0, NE_EXPR, e, si);
- retval |= register_edge_assert_for_1 (op1, NE_EXPR, e, si);
- }
- }
-
- /* In the case of NAME == 0 or NAME != 1, for BIT_IOR_EXPR defining
- statement of NAME we can assert both operands of the BIT_IOR_EXPR
- have zero value. */
- if (((comp_code == EQ_EXPR && integer_zerop (val))
- || (comp_code == NE_EXPR && integer_onep (val))))
- {
- gimple def_stmt = SSA_NAME_DEF_STMT (name);
-
- /* For BIT_IOR_EXPR only if NAME == 0 both operands have
- necessarily zero value, or if type-precision is one. */
- if (is_gimple_assign (def_stmt)
- && (gimple_assign_rhs_code (def_stmt) == BIT_IOR_EXPR
- && (TYPE_PRECISION (TREE_TYPE (name)) == 1
- || comp_code == EQ_EXPR)))
- {
- tree op0 = gimple_assign_rhs1 (def_stmt);
- tree op1 = gimple_assign_rhs2 (def_stmt);
- retval |= register_edge_assert_for_1 (op0, EQ_EXPR, e, si);
- retval |= register_edge_assert_for_1 (op1, EQ_EXPR, e, si);
- }
- }
+ retval = register_edge_assert_for_1 (name, e, si, 5, cond_code,
+ cond_op0, cond_op1, is_else_edge);
return retval;
}
--
1.9.2
