/* Functions to determine/estimate number of iterations of a loop.
- Copyright (C) 2004, 2005, 2006, 2007, 2008 Free Software Foundation,
- Inc.
-
+ Copyright (C) 2004, 2005, 2006, 2007, 2008, 2009, 2010
+ 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/>. */
#include "coretypes.h"
#include "tm.h"
#include "tree.h"
-#include "rtl.h"
#include "tm_p.h"
-#include "hard-reg-set.h"
#include "basic-block.h"
#include "output.h"
#include "diagnostic.h"
+#include "tree-pretty-print.h"
+#include "gimple-pretty-print.h"
#include "intl.h"
#include "tree-flow.h"
#include "tree-dump.h"
*var = op0;
/* Always sign extend the offset. */
- off = double_int_sext (tree_to_double_int (op1),
- TYPE_PRECISION (type));
+ off = tree_to_double_int (op1);
+ if (negate)
+ off = double_int_neg (off);
+ off = double_int_sext (off, TYPE_PRECISION (type));
mpz_set_double_int (offset, off, false);
break;
return;
default:
return;
- }
+ }
mpz_init (offc0);
mpz_init (offc1);
overflow decreases the appropriate offset by M, and underflow
increases it by M. The above inequality would not necessarily be
true if
-
+
-- VARX + OFFX underflows and VARX + OFFC0 does not, or
VARX + OFFC0 overflows, but VARX + OFFX does not.
This may only happen if OFFX < OFFC0.
/* Stores the bounds on the value of the expression X - Y in LOOP to BNDS.
The subtraction is considered to be performed in arbitrary precision,
without overflows.
-
+
We do not attempt to be too clever regarding the value ranges of X and
Y; most of the time, they are just integers or ssa names offsetted by
integer. However, we try to use the information contained in the
}
/* Derives the upper bound BND on the number of executions of loop with exit
- condition S * i <> C, assuming that the loop is not infinite. If
+ condition S * i <> C, assuming that this exit is taken. If
NO_OVERFLOW is true, then the control variable of the loop does not
overflow. If NO_OVERFLOW is true or BNDS.below >= 0, then BNDS.up
contains the upper bound on the value of C. */
/* Determines number of iterations of loop whose ending condition
is IV <> FINAL. TYPE is the type of the iv. The number of
- iterations is stored to NITER. NEVER_INFINITE is true if
+ iterations is stored to NITER. EXIT_MUST_BE_TAKEN is true if
we know that the exit must be taken eventually, i.e., that the IV
ever reaches the value FINAL (we derived this earlier, and possibly set
NITER->assumptions to make sure this is the case). BNDS contains the
static bool
number_of_iterations_ne (tree type, affine_iv *iv, tree final,
- struct tree_niter_desc *niter, bool never_infinite,
+ struct tree_niter_desc *niter, bool exit_must_be_taken,
bounds *bnds)
{
tree niter_type = unsigned_type_for (type);
build_int_cst (niter_type, 1), bits);
s = fold_binary_to_constant (RSHIFT_EXPR, niter_type, s, bits);
- if (!never_infinite)
+ if (!exit_must_be_taken)
{
- /* If we cannot assume that the loop is not infinite, record the
+ /* If we cannot assume that the exit is taken eventually, record the
assumptions for divisibility of c. */
assumption = fold_build2 (FLOOR_MOD_EXPR, niter_type, c, d);
assumption = fold_build2 (EQ_EXPR, boolean_type_node,
niter->assumptions = fold_build2 (TRUTH_AND_EXPR, boolean_type_node,
niter->assumptions, assumption);
}
-
+
c = fold_build2 (EXACT_DIV_EXPR, niter_type, c, d);
tmp = fold_build2 (MULT_EXPR, niter_type, c, inverse (s, bound));
niter->niter = fold_build2 (BIT_AND_EXPR, niter_type, tmp, bound);
of the final value does not overflow are recorded in NITER. If we
find the final value, we adjust DELTA and return TRUE. Otherwise
we return false. BNDS bounds the value of IV1->base - IV0->base,
- and will be updated by the same amount as DELTA. */
+ and will be updated by the same amount as DELTA. EXIT_MUST_BE_TAKEN is
+ true if we know that the exit must be taken eventually. */
static bool
number_of_iterations_lt_to_ne (tree type, affine_iv *iv0, affine_iv *iv1,
struct tree_niter_desc *niter,
tree *delta, tree step,
- bounds *bnds)
+ bool exit_must_be_taken, bounds *bnds)
{
tree niter_type = TREE_TYPE (step);
tree mod = fold_build2 (FLOOR_MOD_EXPR, niter_type, *delta, step);
tree tmod;
mpz_t mmod;
tree assumption = boolean_true_node, bound, noloop;
- bool ret = false;
+ bool ret = false, fv_comp_no_overflow;
tree type1 = type;
if (POINTER_TYPE_P (type))
type1 = sizetype;
mpz_set_double_int (mmod, tree_to_double_int (mod), true);
mpz_neg (mmod, mmod);
+ /* If the induction variable does not overflow and the exit is taken,
+ then the computation of the final value does not overflow. This is
+ also obviously the case if the new final value is equal to the
+ current one. Finally, we postulate this for pointer type variables,
+ as the code cannot rely on the object to that the pointer points being
+ placed at the end of the address space (and more pragmatically,
+ TYPE_{MIN,MAX}_VALUE is not defined for pointers). */
+ if (integer_zerop (mod) || POINTER_TYPE_P (type))
+ fv_comp_no_overflow = true;
+ else if (!exit_must_be_taken)
+ fv_comp_no_overflow = false;
+ else
+ fv_comp_no_overflow =
+ (iv0->no_overflow && integer_nonzerop (iv0->step))
+ || (iv1->no_overflow && integer_nonzerop (iv1->step));
+
if (integer_nonzerop (iv0->step))
{
/* The final value of the iv is iv1->base + MOD, assuming that this
computation does not overflow, and that
iv0->base <= iv1->base + MOD. */
- if (!iv0->no_overflow && !integer_zerop (mod))
+ if (!fv_comp_no_overflow)
{
bound = fold_build2 (MINUS_EXPR, type1,
TYPE_MAX_VALUE (type1), tmod);
- if (POINTER_TYPE_P (type))
- bound = fold_convert (type, bound);
assumption = fold_build2 (LE_EXPR, boolean_type_node,
iv1->base, bound);
if (integer_zerop (assumption))
/* The final value of the iv is iv0->base - MOD, assuming that this
computation does not overflow, and that
iv0->base - MOD <= iv1->base. */
- if (!iv1->no_overflow && !integer_zerop (mod))
+ if (!fv_comp_no_overflow)
{
bound = fold_build2 (PLUS_EXPR, type1,
TYPE_MIN_VALUE (type1), tmod);
- if (POINTER_TYPE_P (type))
- bound = fold_convert (type, bound);
assumption = fold_build2 (GE_EXPR, boolean_type_node,
iv0->base, bound);
if (integer_zerop (assumption))
if (!integer_nonzerop (assumption))
niter->assumptions = fold_build2 (TRUTH_AND_EXPR, boolean_type_node,
niter->assumptions, assumption);
-
+
iv0->no_overflow = true;
iv1->no_overflow = true;
return true;
/* We are going to compute the number of iterations as
(iv1->base - iv0->base + step - 1) / step, computed in the unsigned
- variant of TYPE. This formula only works if
-
+ variant of TYPE. This formula only works if
+
-step + 1 <= (iv1->base - iv0->base) <= MAX - step + 1
-
+
(where MAX is the maximum value of the unsigned variant of TYPE, and
the computations in this formula are performed in full precision
(without overflows).
and for loops iv0->base < iv1->base - step * i the condition
iv0->base < iv1->base + step, due to loop header copying, which enable us
to prove the lower bound.
-
+
The upper bound is more complicated. Unless the expressions for initial
and final value themselves contain enough information, we usually cannot
derive it from the context. */
if (rolls_p && no_overflow_p)
return;
-
+
type1 = type;
if (POINTER_TYPE_P (type))
type1 = sizetype;
}
/* And then we can compute iv0->base - diff, and compare it with
- iv1->base. */
- mbzl = fold_build2 (MINUS_EXPR, type1,
+ iv1->base. */
+ mbzl = fold_build2 (MINUS_EXPR, type1,
fold_convert (type1, iv0->base), diff);
mbzr = fold_convert (type1, iv1->base);
}
/* Determines number of iterations of loop whose ending condition
is IV0 < IV1. TYPE is the type of the iv. The number of
iterations is stored to NITER. BNDS bounds the difference
- IV1->base - IV0->base. */
+ IV1->base - IV0->base. EXIT_MUST_BE_TAKEN is true if we know
+ that the exit must be taken eventually. */
static bool
number_of_iterations_lt (tree type, affine_iv *iv0, affine_iv *iv1,
struct tree_niter_desc *niter,
- bool never_infinite ATTRIBUTE_UNUSED,
- bounds *bnds)
+ bool exit_must_be_taken, bounds *bnds)
{
tree niter_type = unsigned_type_for (type);
tree delta, step, s;
or
for (i = iv1->base; i > iv0->base; i--).
-
+
In both cases # of iterations is iv1->base - iv0->base, assuming that
iv1->base >= iv0->base.
transform the condition to != comparison. In particular, this will be
the case if DELTA is constant. */
if (number_of_iterations_lt_to_ne (type, iv0, iv1, niter, &delta, step,
- bnds))
+ exit_must_be_taken, bnds))
{
affine_iv zps;
/* Determines number of iterations of loop whose ending condition
is IV0 <= IV1. TYPE is the type of the iv. The number of
- iterations is stored to NITER. NEVER_INFINITE is true if
+ iterations is stored to NITER. EXIT_MUST_BE_TAKEN is true if
we know that this condition must eventually become false (we derived this
earlier, and possibly set NITER->assumptions to make sure this
is the case). BNDS bounds the difference IV1->base - IV0->base. */
static bool
number_of_iterations_le (tree type, affine_iv *iv0, affine_iv *iv1,
- struct tree_niter_desc *niter, bool never_infinite,
+ struct tree_niter_desc *niter, bool exit_must_be_taken,
bounds *bnds)
{
tree assumption;
/* Say that IV0 is the control variable. Then IV0 <= IV1 iff
IV0 < IV1 + 1, assuming that IV1 is not equal to the greatest
value of the type. This we must know anyway, since if it is
- equal to this value, the loop rolls forever. */
+ equal to this value, the loop rolls forever. We do not check
+ this condition for pointer type ivs, as the code cannot rely on
+ the object to that the pointer points being placed at the end of
+ the address space (and more pragmatically, TYPE_{MIN,MAX}_VALUE is
+ not defined for pointers). */
- if (!never_infinite)
+ if (!exit_must_be_taken && !POINTER_TYPE_P (type))
{
if (integer_nonzerop (iv0->step))
assumption = fold_build2 (NE_EXPR, boolean_type_node,
bounds_add (bnds, double_int_one, type1);
- return number_of_iterations_lt (type, iv0, iv1, niter, never_infinite, bnds);
+ return number_of_iterations_lt (type, iv0, iv1, niter, exit_must_be_taken,
+ bnds);
}
/* Dumps description of affine induction variable IV to FILE. */
exited (including possibly non-returning function calls, exceptions, etc.)
-- in this case we can use the information whether the control induction
variables can overflow or not in a more efficient way.
-
+
The results (number of iterations and assumptions as described in
comments at struct tree_niter_desc in tree-flow.h) are stored to NITER.
Returns false if it fails to determine number of iterations, true if it
affine_iv *iv1, struct tree_niter_desc *niter,
bool only_exit)
{
- bool never_infinite, ret;
+ bool exit_must_be_taken = false, ret;
bounds bnds;
/* The meaning of these assumptions is this:
code = swap_tree_comparison (code);
}
- if (!only_exit)
- {
- /* If this is not the only possible exit from the loop, the information
- that the induction variables cannot overflow as derived from
- signedness analysis cannot be relied upon. We use them e.g. in the
- following way: given loop for (i = 0; i <= n; i++), if i is
- signed, it cannot overflow, thus this loop is equivalent to
- for (i = 0; i < n + 1; i++); however, if n == MAX, but the loop
- is exited in some other way before i overflows, this transformation
- is incorrect (the new loop exits immediately). */
- iv0->no_overflow = false;
- iv1->no_overflow = false;
- }
-
if (POINTER_TYPE_P (type))
{
/* Comparison of pointers is undefined unless both iv0 and iv1 point
to the same object. If they do, the control variable cannot wrap
(as wrap around the bounds of memory will never return a pointer
that would be guaranteed to point to the same object, even if we
- avoid undefined behavior by casting to size_t and back). The
- restrictions on pointer arithmetics and comparisons of pointers
- ensure that using the no-overflow assumptions is correct in this
- case even if ONLY_EXIT is false. */
+ avoid undefined behavior by casting to size_t and back). */
iv0->no_overflow = true;
iv1->no_overflow = true;
}
- /* If the control induction variable does not overflow, the loop obviously
- cannot be infinite. */
- if (!integer_zerop (iv0->step) && iv0->no_overflow)
- never_infinite = true;
- else if (!integer_zerop (iv1->step) && iv1->no_overflow)
- never_infinite = true;
- else
- never_infinite = false;
+ /* If the control induction variable does not overflow and the only exit
+ from the loop is the one that we analyze, we know it must be taken
+ eventually. */
+ if (only_exit)
+ {
+ if (!integer_zerop (iv0->step) && iv0->no_overflow)
+ exit_must_be_taken = true;
+ else if (!integer_zerop (iv1->step) && iv1->no_overflow)
+ exit_must_be_taken = true;
+ }
/* We can handle the case when neither of the sides of the comparison is
invariant, provided that the test is NE_EXPR. This rarely occurs in
niter->max = double_int_zero;
return true;
}
-
+
/* OK, now we know we have a senseful loop. Handle several cases, depending
on what comparison operator is used. */
bound_difference (loop, iv1->base, iv0->base, &bnds);
case NE_EXPR:
gcc_assert (integer_zerop (iv1->step));
ret = number_of_iterations_ne (type, iv0, iv1->base, niter,
- never_infinite, &bnds);
+ exit_must_be_taken, &bnds);
break;
case LT_EXPR:
- ret = number_of_iterations_lt (type, iv0, iv1, niter, never_infinite,
+ ret = number_of_iterations_lt (type, iv0, iv1, niter, exit_must_be_taken,
&bnds);
break;
case LE_EXPR:
- ret = number_of_iterations_le (type, iv0, iv1, niter, never_infinite,
+ ret = number_of_iterations_le (type, iv0, iv1, niter, exit_must_be_taken,
&bnds);
break;
default:
return false;
}
-
+
op0 = gimple_cond_lhs (stmt);
op1 = gimple_cond_rhs (stmt);
type = TREE_TYPE (op0);
if (TREE_CODE (type) != INTEGER_TYPE
&& !POINTER_TYPE_P (type))
return false;
-
+
if (!simple_iv (loop, loop_containing_stmt (stmt), op0, &iv0, false))
return false;
if (!simple_iv (loop, loop_containing_stmt (stmt), op1, &iv1, false))
{
const char *wording;
location_t loc = gimple_location (stmt);
-
+
/* We can provide a more specific warning if one of the operator is
constant and the other advances by +1 or -1. */
if (!integer_zerop (iv1.step)
? N_("assuming that the loop is not infinite")
: N_("cannot optimize possibly infinite loops");
else
- wording =
+ wording =
flag_unsafe_loop_optimizations
? N_("assuming that the loop counter does not overflow")
: N_("cannot optimize loop, the loop counter may overflow");
- if (LOCATION_LINE (loc) > 0)
- warning (OPT_Wunsafe_loop_optimizations, "%H%s", &loc, gettext (wording));
- else
- warning (OPT_Wunsafe_loop_optimizations, "%s", gettext (wording));
+ warning_at ((LOCATION_LINE (loc) > 0) ? loc : input_location,
+ OPT_Wunsafe_loop_optimizations, "%s", gettext (wording));
}
return flag_unsafe_loop_optimizations;
finite_loop_p (struct loop *loop)
{
unsigned i;
- VEC (edge, heap) *exits = get_loop_exit_edges (loop);
+ VEC (edge, heap) *exits;
edge ex;
struct tree_niter_desc desc;
bool finite = false;
loop->num);
return true;
}
-
+
exits = get_loop_exit_edges (loop);
for (i = 0; VEC_iterate (edge, exits, i, ex); i++)
{
if (!bb
|| !flow_bb_inside_loop_p (loop, bb))
return NULL;
-
+
if (gimple_code (stmt) == GIMPLE_PHI)
{
if (bb == loop->header)
* the initial value of the phi node is constant
* the value of the phi node in the next iteration can be derived from the
value in the current iteration by a chain of operations with constants.
-
+
If such phi node exists, it is returned, otherwise NULL is returned. */
static gimple
return phi;
}
-/* Given an expression X, then
-
+/* Given an expression X, then
+
* if X is NULL_TREE, we return the constant BASE.
* otherwise X is a SSA name, whose value in the considered loop is derived
by a chain of operations with constant from a result of a phi node in
tree niter = NULL_TREE, aniter;
*exit = NULL;
+
+ /* Loops with multiple exits are expensive to handle and less important. */
+ if (!flag_expensive_optimizations
+ && VEC_length (edge, exits) > 1)
+ return chrec_dont_know;
+
for (i = 0; VEC_iterate (edge, exits, i, ex); i++)
{
if (!just_once_each_iteration_p (loop, ex->src))
/* Returns a constant upper bound on the value of expression VAL. VAL
is considered to be unsigned. If its type is signed, its value must
be nonnegative. */
-
+
static double_int
derive_constant_upper_bound (tree val)
{
/* Returns a constant upper bound on the value of expression OP0 CODE OP1,
whose type is TYPE. The expression is considered to be unsigned. If
its type is signed, its value must be nonnegative. */
-
+
static double_int
derive_constant_upper_bound_ops (tree type, tree op0,
enum tree_code code, tree op1)
return max;
return derive_constant_upper_bound_assign (stmt);
- default:
+ default:
return max;
}
}
allocated structure. If this is the case, the array may be allocated larger
than its upper bound implies. */
-static bool
+bool
array_at_struct_end_p (tree ref)
{
tree base = get_base_address (ref);
its declaration. */
if (!base || !INDIRECT_REF_P (base))
return false;
-
+
for (;handled_component_p (ref); ref = parent)
{
parent = TREE_OPERAND (ref, 0);
low = array_ref_low_bound (base);
high = array_ref_up_bound (base);
-
+
/* The case of nonconstant bounds could be handled, but it would be
complicated. */
if (TREE_CODE (low) != INTEGER_CST
next = fold_binary (PLUS_EXPR, type, low, step);
else
next = fold_binary (PLUS_EXPR, type, high, step);
-
+
if (tree_int_cst_compare (low, next) <= 0
&& tree_int_cst_compare (next, high) <= 0)
return true;
gimple_stmt_iterator bsi;
basic_block bb;
bool reliable;
-
+
bbs = get_loop_body (loop);
for (i = 0; i < loop->num_nodes; i++)
true, true, true);
}
VEC_free (edge, heap, exits);
-
+
infer_loop_bounds_from_undefined (loop);
/* If we have a measured profile, use it to estimate the number of
static bool
n_of_executions_at_most (gimple stmt,
- struct nb_iter_bound *niter_bound,
+ struct nb_iter_bound *niter_bound,
tree niter)
{
double_int bound = niter_bound->bound;
/* We know that NITER_BOUND->stmt is executed at most NITER_BOUND->bound + 1
times. This means that:
-
+
-- if NITER_BOUND->is_exit is true, then everything before
NITER_BOUND->stmt is executed at most NITER_BOUND->bound + 1
times, and everything after it at most NITER_BOUND->bound times.
enough with respect to the step and initial condition in order to
keep the evolution confined in TYPEs bounds. Return true when the
iv is known to overflow or when the property is not computable.
-
+
USE_OVERFLOW_SEMANTICS is true if this function should assume that
the rules for overflow of the given language apply (e.g., that signed
arithmetics in C does not overflow). */
bool
-scev_probably_wraps_p (tree base, tree step,
+scev_probably_wraps_p (tree base, tree step,
gimple at_stmt, struct loop *loop,
bool use_overflow_semantics)
{
We used to test for the following situation that frequently appears
during address arithmetics:
-
+
D.1621_13 = (long unsigned intD.4) D.1620_12;
D.1622_14 = D.1621_13 * 8;
D.1623_15 = (doubleD.29 *) D.1622_14;