/* Perform doloop optimizations
- Copyright (C) 2004, 2005, 2006 Free Software Foundation, Inc.
+ Copyright (C) 2004, 2005, 2006, 2007 Free Software Foundation, Inc.
Based on code by Michael P. Hayes (m.hayes@elec.canterbury.ac.nz)
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 2, or (at your option) any later
+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
for more details.
You should have received a copy of the GNU General Public License
-along with GCC; see the file COPYING. If not, write to the Free
-Software Foundation, 51 Franklin Street, Fifth Floor, Boston, MA
-02110-1301, USA. */
+along with GCC; see the file COPYING3. If not see
+<http://www.gnu.org/licenses/>. */
#include "config.h"
#include "system.h"
if it is not a decrement and branch jump insn. */
rtx
-doloop_condition_get (rtx pattern)
+doloop_condition_get (rtx doloop_pat)
{
rtx cmp;
rtx inc;
rtx reg;
rtx inc_src;
rtx condition;
+ rtx pattern;
- /* The canonical doloop pattern we expect is:
+ /* The canonical doloop pattern we expect has one of the following
+ forms:
- (parallel [(set (pc) (if_then_else (condition)
- (label_ref (label))
- (pc)))
- (set (reg) (plus (reg) (const_int -1)))
- (additional clobbers and uses)])
+ 1) (parallel [(set (pc) (if_then_else (condition)
+ (label_ref (label))
+ (pc)))
+ (set (reg) (plus (reg) (const_int -1)))
+ (additional clobbers and uses)])
- Some targets (IA-64) wrap the set of the loop counter in
- an if_then_else too.
+ The branch must be the first entry of the parallel (also required
+ by jump.c), and the second entry of the parallel must be a set of
+ the loop counter register. Some targets (IA-64) wrap the set of
+ the loop counter in an if_then_else too.
- In summary, the branch must be the first entry of the
- parallel (also required by jump.c), and the second
- entry of the parallel must be a set of the loop counter
- register. */
+ 2) (set (reg) (plus (reg) (const_int -1))
+ (set (pc) (if_then_else (reg != 0)
+ (label_ref (label))
+ (pc))). */
+
+ pattern = PATTERN (doloop_pat);
if (GET_CODE (pattern) != PARALLEL)
- return 0;
+ {
+ rtx cond;
+
+ /* We expect the decrement to immediately precede the branch. */
+
+ if ((PREV_INSN (doloop_pat) == NULL_RTX)
+ || !INSN_P (PREV_INSN (doloop_pat)))
+ return 0;
- cmp = XVECEXP (pattern, 0, 0);
- inc = XVECEXP (pattern, 0, 1);
+ cmp = pattern;
+ inc = PATTERN (PREV_INSN (doloop_pat));
+ /* We expect the condition to be of the form (reg != 0) */
+ cond = XEXP (SET_SRC (cmp), 0);
+ if (GET_CODE (cond) != NE || XEXP (cond, 1) != const0_rtx)
+ return 0;
+
+ }
+ else
+ {
+ cmp = XVECEXP (pattern, 0, 0);
+ inc = XVECEXP (pattern, 0, 1);
+ }
/* Check for (set (reg) (something)). */
if (GET_CODE (inc) != SET)
if ((XEXP (condition, 0) == reg)
|| (GET_CODE (XEXP (condition, 0)) == PLUS
&& XEXP (XEXP (condition, 0), 0) == reg))
+ {
+ if (GET_CODE (pattern) != PARALLEL)
+ /* The second form we expect:
+
+ (set (reg) (plus (reg) (const_int -1))
+ (set (pc) (if_then_else (reg != 0)
+ (label_ref (label))
+ (pc))).
+
+ is equivalent to the following:
+
+ (parallel [(set (pc) (if_then_else (reg != 1)
+ (label_ref (label))
+ (pc)))
+ (set (reg) (plus (reg) (const_int -1)))
+ (additional clobbers and uses)])
+
+ So we return that form instead.
+ */
+ condition = gen_rtx_fmt_ee (NE, VOIDmode, inc_src, const1_rtx);
+
return condition;
+ }
/* ??? If a machine uses a funny comparison, we could return a
canonicalized form here. */
do_compare_rtx_and_jump (op0, op1, code, 0, mode, NULL_RTX, NULL_RTX, label);
jump = get_last_insn ();
- if (!JUMP_P (jump))
+ if (!jump || !JUMP_P (jump))
{
/* The condition is always false and the jump was optimized out. */
end_sequence ();
seq = get_insns ();
end_sequence ();
- bb = loop_split_edge_with (*e, seq);
+
+ /* There always is at least the jump insn in the sequence. */
+ gcc_assert (seq != NULL_RTX);
+
+ bb = split_edge_and_insert (*e, seq);
*e = single_succ_edge (bb);
if (any_uncondjump_p (jump))
rtx ass = copy_rtx (desc->noloop_assumptions);
basic_block preheader = loop_preheader_edge (loop)->src;
basic_block set_zero
- = loop_split_edge_with (loop_preheader_edge (loop), NULL_RTX);
+ = split_edge (loop_preheader_edge (loop));
basic_block new_preheader
- = loop_split_edge_with (loop_preheader_edge (loop), NULL_RTX);
+ = split_edge (loop_preheader_edge (loop));
edge te;
/* Expand the condition testing the assumptions and if it does not pass,
{
/* All the conditions were simplified to false, remove the
unreachable set_zero block. */
- remove_bb_from_loops (set_zero);
delete_basic_block (set_zero);
}
else
emit_insn_after (sequence, BB_END (set_zero));
set_immediate_dominator (CDI_DOMINATORS, set_zero,
- recount_dominator (CDI_DOMINATORS,
- set_zero));
+ recompute_dominator (CDI_DOMINATORS,
+ set_zero));
}
set_immediate_dominator (CDI_DOMINATORS, new_preheader,
- recount_dominator (CDI_DOMINATORS,
- new_preheader));
+ recompute_dominator (CDI_DOMINATORS,
+ new_preheader));
}
/* Some targets (eg, C4x) need to initialize special looping
{
while (NEXT_INSN (doloop_pat) != NULL_RTX)
doloop_pat = NEXT_INSN (doloop_pat);
- if (JUMP_P (doloop_pat))
- doloop_pat = PATTERN (doloop_pat);
- else
+ if (!JUMP_P (doloop_pat))
doloop_pat = NULL_RTX;
}
return true;
}
-/* This is the main entry point. Process all LOOPS using doloop_optimize. */
+/* This is the main entry point. Process all loops using doloop_optimize. */
void
-doloop_optimize_loops (struct loops *loops)
+doloop_optimize_loops (void)
{
- unsigned i;
+ loop_iterator li;
struct loop *loop;
- for (i = 1; i < loops->num; i++)
+ FOR_EACH_LOOP (li, loop, 0)
{
- loop = loops->parray[i];
- if (!loop)
- continue;
-
doloop_optimize (loop);
}
#ifdef ENABLE_CHECKING
verify_dominators (CDI_DOMINATORS);
- verify_loop_structure (loops);
+ verify_loop_structure ();
#endif
}
#endif /* HAVE_doloop_end */