|
/*
* Copyright (c) 1997, 2022, Oracle and/or its affiliates. All rights reserved.
* DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
*
* This code is free software; you can redistribute it and/or modify it
* under the terms of the GNU General Public License version 2 only, as
* published by the Free Software Foundation.
*
* This code 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
* version 2 for more details (a copy is included in the LICENSE file that
* accompanied this code).
*
* You should have received a copy of the GNU General Public License version
* 2 along with this work; if not, write to the Free Software Foundation,
* Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
*
* Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
* or visit www.oracle.com if you need additional information or have any
* questions.
*
*/
#include "precompiled.hpp"
#include "libadt/vectset.hpp"
#include "memory/allocation.inline.hpp"
#include "memory/resourceArea.hpp"
#include "compiler/compilerDirectives.hpp"
#include "opto/block.hpp"
#include "opto/cfgnode.hpp"
#include "opto/chaitin.hpp"
#include "opto/loopnode.hpp"
#include "opto/machnode.hpp"
#include "opto/matcher.hpp"
#include "opto/opcodes.hpp"
#include "opto/rootnode.hpp"
#include "utilities/copy.hpp"
#include "utilities/powerOfTwo.hpp"
void Block_Array::grow( uint i ) {
assert(i >= Max(), "must be an overflow");
debug_only(_limit = i+1);
if( i < _size ) return;
if( !_size ) {
_size = 1;
_blocks = (Block**)_arena->Amalloc( _size * sizeof(Block*) );
_blocks[0] = NULL;
}
uint old = _size;
_size = next_power_of_2(i);
_blocks = (Block**)_arena->Arealloc( _blocks, old*sizeof(Block*),_size*sizeof(Block*));
Copy::zero_to_bytes( &_blocks[old], (_size-old)*sizeof(Block*) );
}
void Block_List::remove(uint i) {
assert(i < _cnt, "index out of bounds");
Copy::conjoint_words_to_lower((HeapWord*)&_blocks[i+1], (HeapWord*)&_blocks[i], ((_cnt-i-1)*sizeof(Block*)));
pop(); // shrink list by one block
}
void Block_List::insert(uint i, Block *b) {
push(b); // grow list by one block
Copy::conjoint_words_to_higher((HeapWord*)&_blocks[i], (HeapWord*)&_blocks[i+1], ((_cnt-i-1)*sizeof(Block*)));
_blocks[i] = b;
}
#ifndef PRODUCT
void Block_List::print() {
for (uint i=0; i < size(); i++) {
tty->print("B%d ", _blocks[i]->_pre_order);
}
tty->print("size = %d\n", size());
}
#endif
uint Block::code_alignment() const {
// Check for Root block
if (_pre_order == 0) return CodeEntryAlignment;
// Check for Start block
if (_pre_order == 1) return InteriorEntryAlignment;
// Check for loop alignment
if (has_loop_alignment()) return loop_alignment();
return relocInfo::addr_unit(); // no particular alignment
}
uint Block::compute_loop_alignment() {
Node *h = head();
int unit_sz = relocInfo::addr_unit();
if (h->is_Loop() && h->as_Loop()->is_inner_loop()) {
// Pre- and post-loops have low trip count so do not bother with
// NOPs for align loop head. The constants are hidden from tuning
// but only because my "divide by 4" heuristic surely gets nearly
// all possible gain (a "do not align at all" heuristic has a
// chance of getting a really tiny gain).
if (h->is_CountedLoop() && (h->as_CountedLoop()->is_pre_loop() ||
h->as_CountedLoop()->is_post_loop())) {
return (OptoLoopAlignment > 4*unit_sz) ? (OptoLoopAlignment>>2) : unit_sz;
}
// Loops with low backedge frequency should not be aligned.
Node *n = h->in(LoopNode::LoopBackControl)->in(0);
if (n->is_MachIf() && n->as_MachIf()->_prob < 0.01) {
return unit_sz; // Loop does not loop, more often than not!
}
return OptoLoopAlignment; // Otherwise align loop head
}
return unit_sz; // no particular alignment
}
// Compute the size of first 'inst_cnt' instructions in this block.
// Return the number of instructions left to compute if the block has
// less then 'inst_cnt' instructions. Stop, and return 0 if sum_size
// exceeds OptoLoopAlignment.
uint Block::compute_first_inst_size(uint& sum_size, uint inst_cnt,
PhaseRegAlloc* ra) {
uint last_inst = number_of_nodes();
for( uint j = 0; j < last_inst && inst_cnt > 0; j++ ) {
uint inst_size = get_node(j)->size(ra);
if( inst_size > 0 ) {
inst_cnt--;
uint sz = sum_size + inst_size;
if( sz <= (uint)OptoLoopAlignment ) {
// Compute size of instructions which fit into fetch buffer only
// since all inst_cnt instructions will not fit even if we align them.
sum_size = sz;
} else {
return 0;
}
}
}
return inst_cnt;
}
uint Block::find_node( const Node *n ) const {
for( uint i = 0; i < number_of_nodes(); i++ ) {
if( get_node(i) == n )
return i;
}
ShouldNotReachHere();
return 0;
}
// Find and remove n from block list
void Block::find_remove( const Node *n ) {
remove_node(find_node(n));
}
bool Block::contains(const Node *n) const {
return _nodes.contains(n);
}
bool Block::is_trivially_unreachable() const {
return num_preds() <= 1 && !head()->is_Root() && !head()->is_Start();
}
// Return empty status of a block. Empty blocks contain only the head, other
// ideal nodes, and an optional trailing goto.
int Block::is_Empty() const {
// Root or start block is not considered empty
if (head()->is_Root() || head()->is_Start()) {
return not_empty;
}
int success_result = completely_empty;
int end_idx = number_of_nodes() - 1;
// Check for ending goto
if ((end_idx > 0) && (get_node(end_idx)->is_MachGoto())) {
success_result = empty_with_goto;
end_idx--;
}
// Unreachable blocks are considered empty
if (is_trivially_unreachable()) {
return success_result;
}
// Ideal nodes are allowable in empty blocks: skip them Only MachNodes
// turn directly into code, because only MachNodes have non-trivial
// emit() functions.
while ((end_idx > 0) && !get_node(end_idx)->is_Mach()) {
end_idx--;
}
// No room for any interesting instructions?
if (end_idx == 0) {
return success_result;
}
return not_empty;
}
// Return true if the block's code implies that it is likely to be
// executed infrequently. Check to see if the block ends in a Halt or
// a low probability call.
bool Block::has_uncommon_code() const {
Node* en = end();
if (en->is_MachGoto())
en = en->in(0);
if (en->is_Catch())
en = en->in(0);
if (en->is_MachProj() && en->in(0)->is_MachCall()) {
MachCallNode* call = en->in(0)->as_MachCall();
if (call->cnt() != COUNT_UNKNOWN && call->cnt() <= PROB_UNLIKELY_MAG(4)) {
// This is true for slow-path stubs like new_{instance,array},
// slow_arraycopy, complete_monitor_locking, uncommon_trap.
// The magic number corresponds to the probability of an uncommon_trap,
// even though it is a count not a probability.
return true;
}
}
int op = en->is_Mach() ? en->as_Mach()->ideal_Opcode() : en->Opcode();
return op == Op_Halt;
}
// True if block is low enough frequency or guarded by a test which
// mostly does not go here.
bool PhaseCFG::is_uncommon(const Block* block) {
// Initial blocks must never be moved, so are never uncommon.
if (block->head()->is_Root() || block->head()->is_Start()) return false;
// Check for way-low freq
if(block->_freq < BLOCK_FREQUENCY(0.00001f) ) return true;
// Look for code shape indicating uncommon_trap or slow path
if (block->has_uncommon_code()) return true;
const float epsilon = 0.05f;
const float guard_factor = PROB_UNLIKELY_MAG(4) / (1.f - epsilon);
uint uncommon_preds = 0;
uint freq_preds = 0;
uint uncommon_for_freq_preds = 0;
for( uint i=1; i< block->num_preds(); i++ ) {
Block* guard = get_block_for_node(block->pred(i));
// Check to see if this block follows its guard 1 time out of 10000
// or less.
//
// See list of magnitude-4 unlikely probabilities in cfgnode.hpp which
// we intend to be "uncommon", such as slow-path TLE allocation,
// predicted call failure, and uncommon trap triggers.
//
// Use an epsilon value of 5% to allow for variability in frequency
// predictions and floating point calculations. The net effect is
// that guard_factor is set to 9500.
//
// Ignore low-frequency blocks.
// The next check is (guard->_freq < 1.e-5 * 9500.).
if(guard->_freq*BLOCK_FREQUENCY(guard_factor) < BLOCK_FREQUENCY(0.00001f)) {
uncommon_preds++;
} else {
freq_preds++;
if(block->_freq < guard->_freq * guard_factor ) {
uncommon_for_freq_preds++;
}
}
}
if( block->num_preds() > 1 &&
// The block is uncommon if all preds are uncommon or
(uncommon_preds == (block->num_preds()-1) ||
// it is uncommon for all frequent preds.
uncommon_for_freq_preds == freq_preds) ) {
return true;
}
return false;
}
#ifndef PRODUCT
void Block::dump_bidx(const Block* orig, outputStream* st) const {
if (_pre_order) st->print("B%d", _pre_order);
else st->print("N%d", head()->_idx);
if (Verbose && orig != this) {
// Dump the original block's idx
st->print(" (");
orig->dump_bidx(orig, st);
st->print(")");
}
}
void Block::dump_pred(const PhaseCFG* cfg, Block* orig, outputStream* st) const {
if (is_connector()) {
for (uint i=1; i<num_preds(); i++) {
Block *p = cfg->get_block_for_node(pred(i));
p->dump_pred(cfg, orig, st);
}
} else {
dump_bidx(orig, st);
st->print(" ");
}
}
void Block::dump_head(const PhaseCFG* cfg, outputStream* st) const {
// Print the basic block.
dump_bidx(this, st);
st->print(": ");
// Print the outgoing CFG edges.
st->print("#\tout( ");
for( uint i=0; i<_num_succs; i++ ) {
non_connector_successor(i)->dump_bidx(_succs[i], st);
st->print(" ");
}
// Print the incoming CFG edges.
st->print(") <- ");
if( head()->is_block_start() ) {
st->print("in( ");
for (uint i=1; i<num_preds(); i++) {
Node *s = pred(i);
if (cfg != NULL) {
Block *p = cfg->get_block_for_node(s);
p->dump_pred(cfg, p, st);
} else {
while (!s->is_block_start()) {
s = s->in(0);
}
st->print("N%d ", s->_idx );
}
}
st->print(") ");
} else {
st->print("BLOCK HEAD IS JUNK ");
}
// Print loop, if any
const Block *bhead = this; // Head of self-loop
Node *bh = bhead->head();
if ((cfg != NULL) && bh->is_Loop() && !head()->is_Root()) {
LoopNode *loop = bh->as_Loop();
const Block *bx = cfg->get_block_for_node(loop->in(LoopNode::LoopBackControl));
while (bx->is_connector()) {
bx = cfg->get_block_for_node(bx->pred(1));
}
st->print("Loop( B%d-B%d ", bhead->_pre_order, bx->_pre_order);
// Dump any loop-specific bits, especially for CountedLoops.
loop->dump_spec(st);
st->print(")");
} else if (has_loop_alignment()) {
st->print("top-of-loop");
}
// Print frequency and other optimization-relevant information
st->print(" Freq: %g",_freq);
if( Verbose || WizardMode ) {
st->print(" IDom: %d/#%d", _idom ? _idom->_pre_order : 0, _dom_depth);
st->print(" RegPressure: %d",_reg_pressure);
st->print(" IHRP Index: %d",_ihrp_index);
st->print(" FRegPressure: %d",_freg_pressure);
st->print(" FHRP Index: %d",_fhrp_index);
}
st->cr();
}
void Block::dump() const {
dump(NULL);
}
void Block::dump(const PhaseCFG* cfg) const {
dump_head(cfg);
for (uint i=0; i< number_of_nodes(); i++) {
get_node(i)->dump();
}
tty->print("\n");
}
#endif
PhaseCFG::PhaseCFG(Arena* arena, RootNode* root, Matcher& matcher)
: Phase(CFG)
, _root(root)
, _block_arena(arena)
, _regalloc(NULL)
, _scheduling_for_pressure(false)
, _matcher(matcher)
, _node_to_block_mapping(arena)
, _node_latency(NULL)
#ifndef PRODUCT
, _trace_opto_pipelining(C->directive()->TraceOptoPipeliningOption)
#endif
#ifdef ASSERT
, _raw_oops(arena)
#endif
{
ResourceMark rm;
// I'll need a few machine-specific GotoNodes. Make an Ideal GotoNode,
// then Match it into a machine-specific Node. Then clone the machine
// Node on demand.
Node *x = new GotoNode(NULL);
x->init_req(0, x);
_goto = matcher.match_tree(x);
assert(_goto != NULL, "");
_goto->set_req(0,_goto);
// Build the CFG in Reverse Post Order
_number_of_blocks = build_cfg();
_root_block = get_block_for_node(_root);
}
// Build a proper looking CFG. Make every block begin with either a StartNode
// or a RegionNode. Make every block end with either a Goto, If or Return.
// The RootNode both starts and ends it's own block. Do this with a recursive
// backwards walk over the control edges.
uint PhaseCFG::build_cfg() {
VectorSet visited;
// Allocate stack with enough space to avoid frequent realloc
Node_Stack nstack(C->live_nodes() >> 1);
nstack.push(_root, 0);
uint sum = 0; // Counter for blocks
while (nstack.is_nonempty()) {
// node and in's index from stack's top
// 'np' is _root (see above) or RegionNode, StartNode: we push on stack
// only nodes which point to the start of basic block (see below).
Node *np = nstack.node();
// idx > 0, except for the first node (_root) pushed on stack
// at the beginning when idx == 0.
// We will use the condition (idx == 0) later to end the build.
uint idx = nstack.index();
Node *proj = np->in(idx);
const Node *x = proj->is_block_proj();
// Does the block end with a proper block-ending Node? One of Return,
// If or Goto? (This check should be done for visited nodes also).
if (x == NULL) { // Does not end right...
Node *g = _goto->clone(); // Force it to end in a Goto
g->set_req(0, proj);
np->set_req(idx, g);
x = proj = g;
}
if (!visited.test_set(x->_idx)) { // Visit this block once
// Skip any control-pinned middle'in stuff
Node *p = proj;
do {
proj = p; // Update pointer to last Control
p = p->in(0); // Move control forward
} while( !p->is_block_proj() &&
!p->is_block_start() );
// Make the block begin with one of Region or StartNode.
if( !p->is_block_start() ) {
RegionNode *r = new RegionNode( 2 );
r->init_req(1, p); // Insert RegionNode in the way
proj->set_req(0, r); // Insert RegionNode in the way
p = r;
}
// 'p' now points to the start of this basic block
// Put self in array of basic blocks
Block *bb = new (_block_arena) Block(_block_arena, p);
map_node_to_block(p, bb);
map_node_to_block(x, bb);
if( x != p ) { // Only for root is x == p
bb->push_node((Node*)x);
}
// Now handle predecessors
++sum; // Count 1 for self block
uint cnt = bb->num_preds();
for (int i = (cnt - 1); i > 0; i-- ) { // For all predecessors
Node *prevproj = p->in(i); // Get prior input
assert( !prevproj->is_Con(), "dead input not removed" );
// Check to see if p->in(i) is a "control-dependent" CFG edge -
// i.e., it splits at the source (via an IF or SWITCH) and merges
// at the destination (via a many-input Region).
// This breaks critical edges. The RegionNode to start the block
// will be added when <p,i> is pulled off the node stack
if ( cnt > 2 ) { // Merging many things?
assert( prevproj== bb->pred(i),"");
if(prevproj->is_block_proj() != prevproj) { // Control-dependent edge?
// Force a block on the control-dependent edge
Node *g = _goto->clone(); // Force it to end in a Goto
g->set_req(0,prevproj);
p->set_req(i,g);
}
}
nstack.push(p, i); // 'p' is RegionNode or StartNode
}
} else { // Post-processing visited nodes
nstack.pop(); // remove node from stack
// Check if it the fist node pushed on stack at the beginning.
if (idx == 0) break; // end of the build
// Find predecessor basic block
Block *pb = get_block_for_node(x);
// Insert into nodes array, if not already there
if (!has_block(proj)) {
assert( x != proj, "" );
// Map basic block of projection
map_node_to_block(proj, pb);
pb->push_node(proj);
}
// Insert self as a child of my predecessor block
pb->_succs.map(pb->_num_succs++, get_block_for_node(np));
assert( pb->get_node(pb->number_of_nodes() - pb->_num_succs)->is_block_proj(),
"too many control users, not a CFG?" );
}
}
// Return number of basic blocks for all children and self
return sum;
}
// Inserts a goto & corresponding basic block between
// block[block_no] and its succ_no'th successor block
void PhaseCFG::insert_goto_at(uint block_no, uint succ_no) {
// get block with block_no
assert(block_no < number_of_blocks(), "illegal block number");
Block* in = get_block(block_no);
// get successor block succ_no
assert(succ_no < in->_num_succs, "illegal successor number");
Block* out = in->_succs[succ_no];
// Compute frequency of the new block. Do this before inserting
// new block in case succ_prob() needs to infer the probability from
// surrounding blocks.
float freq = in->_freq * in->succ_prob(succ_no);
// get ProjNode corresponding to the succ_no'th successor of the in block
ProjNode* proj = in->get_node(in->number_of_nodes() - in->_num_succs + succ_no)->as_Proj();
// create region for basic block
RegionNode* region = new RegionNode(2);
region->init_req(1, proj);
// setup corresponding basic block
Block* block = new (_block_arena) Block(_block_arena, region);
map_node_to_block(region, block);
C->regalloc()->set_bad(region->_idx);
// add a goto node
Node* gto = _goto->clone(); // get a new goto node
gto->set_req(0, region);
// add it to the basic block
block->push_node(gto);
map_node_to_block(gto, block);
C->regalloc()->set_bad(gto->_idx);
// hook up successor block
block->_succs.map(block->_num_succs++, out);
// remap successor's predecessors if necessary
for (uint i = 1; i < out->num_preds(); i++) {
if (out->pred(i) == proj) out->head()->set_req(i, gto);
}
// remap predecessor's successor to new block
in->_succs.map(succ_no, block);
// Set the frequency of the new block
block->_freq = freq;
// add new basic block to basic block list
add_block_at(block_no + 1, block);
// Update dominator tree information of the new goto block.
block->_idom = in;
block->_dom_depth = in->_dom_depth + 1;
if (out->_idom != in) {
// The successor block was not immediately dominated by the predecessor
// block, so there is no dominator subtree to update.
return;
}
// Update immediate dominator of the successor block.
out->_idom = block;
// Increment the dominator tree depth of the goto block's descendants. These
// are found by a depth-first search starting from the successor block. Two
// domination properties guarantee that only descendant blocks are visited:
// 1) all dominators of a block b must appear in any path from the root to b;
// 2) if a block b does not dominate another block b', b cannot dominate any
// block reachable from b' either.
// The exploration uses header indices as block identifiers, since
// Block::_pre_order might not be unique in the context of this function.
ResourceMark rm;
VectorSet descendants;
descendants.set(block->head()->_idx); // The goto block is a descendant of itself.
Block_List worklist;
worklist.push(out); // Start exploring from the successor block.
while (worklist.size() > 0) {
Block* b = worklist.pop();
// The immediate dominator of b is a descendant, hence b is also a
// descendant. Even though all predecessors of b might not have been visited
// yet, we know that all dominators of b have been already visited (since
// they must appear in any path from the goto block to b).
descendants.set(b->head()->_idx);
b->_dom_depth++;
for (uint i = 0; i < b->_num_succs; i++) {
Block* s = b->_succs[i];
if (s != get_root_block() &&
!descendants.test(s->head()->_idx) &&
// Do not search below non-descendant successors, since any block
// reachable from them cannot be descendant either.
descendants.test(s->_idom->head()->_idx)) {
worklist.push(s);
}
}
}
}
// Does this block end in a multiway branch that cannot have the default case
// flipped for another case?
static bool no_flip_branch(Block *b) {
int branch_idx = b->number_of_nodes() - b->_num_succs-1;
if (branch_idx < 1) {
return false;
}
Node *branch = b->get_node(branch_idx);
if (branch->is_Catch()) {
return true;
}
if (branch->is_Mach()) {
if (branch->is_MachNullCheck()) {
return true;
}
int iop = branch->as_Mach()->ideal_Opcode();
if (iop == Op_FastLock || iop == Op_FastUnlock) {
return true;
}
// Don't flip if branch has an implicit check.
if (branch->as_Mach()->is_TrapBasedCheckNode()) {
return true;
}
}
return false;
}
// Check for NeverBranch at block end. This needs to become a GOTO to the
// true target. NeverBranch are treated as a conditional branch that always
// goes the same direction for most of the optimizer and are used to give a
// fake exit path to infinite loops. At this late stage they need to turn
// into Goto's so that when you enter the infinite loop you indeed hang.
void PhaseCFG::convert_NeverBranch_to_Goto(Block *b) {
// Find true target
int end_idx = b->end_idx();
int idx = b->get_node(end_idx+1)->as_Proj()->_con;
Block *succ = b->_succs[idx];
Node* gto = _goto->clone(); // get a new goto node
gto->set_req(0, b->head());
Node *bp = b->get_node(end_idx);
b->map_node(gto, end_idx); // Slam over NeverBranch
map_node_to_block(gto, b);
C->regalloc()->set_bad(gto->_idx);
b->pop_node(); // Yank projections
b->pop_node(); // Yank projections
b->_succs.map(0,succ); // Map only successor
b->_num_succs = 1;
// remap successor's predecessors if necessary
uint j;
for (j = 1; j < succ->num_preds(); j++) {
if (succ->pred(j)->in(0) == bp) {
succ->head()->set_req(j, gto);
}
}
// Kill alternate exit path
Block* dead = b->_succs[1 - idx];
for (j = 1; j < dead->num_preds(); j++) {
if (dead->pred(j)->in(0) == bp) {
break;
}
}
// Scan through block, yanking dead path from
// all regions and phis.
dead->head()->del_req(j);
for (int k = 1; dead->get_node(k)->is_Phi(); k++) {
dead->get_node(k)->del_req(j);
}
}
// Helper function to move block bx to the slot following b_index. Return
// true if the move is successful, otherwise false
bool PhaseCFG::move_to_next(Block* bx, uint b_index) {
if (bx == NULL) return false;
// Return false if bx is already scheduled.
uint bx_index = bx->_pre_order;
if ((bx_index <= b_index) && (get_block(bx_index) == bx)) {
return false;
}
// Find the current index of block bx on the block list
bx_index = b_index + 1;
while (bx_index < number_of_blocks() && get_block(bx_index) != bx) {
bx_index++;
}
assert(get_block(bx_index) == bx, "block not found");
// If the previous block conditionally falls into bx, return false,
// because moving bx will create an extra jump.
for(uint k = 1; k < bx->num_preds(); k++ ) {
Block* pred = get_block_for_node(bx->pred(k));
if (pred == get_block(bx_index - 1)) {
if (pred->_num_succs != 1) {
return false;
}
}
}
// Reinsert bx just past block 'b'
_blocks.remove(bx_index);
_blocks.insert(b_index + 1, bx);
return true;
}
// Move empty and uncommon blocks to the end.
void PhaseCFG::move_to_end(Block *b, uint i) {
int e = b->is_Empty();
if (e != Block::not_empty) {
if (e == Block::empty_with_goto) {
// Remove the goto, but leave the block.
b->pop_node();
}
// Mark this block as a connector block, which will cause it to be
// ignored in certain functions such as non_connector_successor().
b->set_connector();
}
// Move the empty block to the end, and don't recheck.
_blocks.remove(i);
_blocks.push(b);
}
// Set loop alignment for every block
void PhaseCFG::set_loop_alignment() {
uint last = number_of_blocks();
assert(get_block(0) == get_root_block(), "");
for (uint i = 1; i < last; i++) {
Block* block = get_block(i);
if (block->head()->is_Loop()) {
block->set_loop_alignment(block);
}
}
}
// Make empty basic blocks to be "connector" blocks, Move uncommon blocks
// to the end.
void PhaseCFG::remove_empty_blocks() {
// Move uncommon blocks to the end
uint last = number_of_blocks();
assert(get_block(0) == get_root_block(), "");
for (uint i = 1; i < last; i++) {
Block* block = get_block(i);
if (block->is_connector()) {
break;
}
// Check for NeverBranch at block end. This needs to become a GOTO to the
// true target. NeverBranch are treated as a conditional branch that
// always goes the same direction for most of the optimizer and are used
// to give a fake exit path to infinite loops. At this late stage they
// need to turn into Goto's so that when you enter the infinite loop you
// indeed hang.
if (block->get_node(block->end_idx())->Opcode() == Op_NeverBranch) {
convert_NeverBranch_to_Goto(block);
}
// Look for uncommon blocks and move to end.
if (!C->do_freq_based_layout()) {
if (is_uncommon(block)) {
move_to_end(block, i);
last--; // No longer check for being uncommon!
if (no_flip_branch(block)) { // Fall-thru case must follow?
// Find the fall-thru block
block = get_block(i);
move_to_end(block, i);
last--;
}
// backup block counter post-increment
i--;
}
}
}
// Move empty blocks to the end
last = number_of_blocks();
for (uint i = 1; i < last; i++) {
Block* block = get_block(i);
if (block->is_Empty() != Block::not_empty) {
move_to_end(block, i);
last--;
i--;
}
} // End of for all blocks
}
Block *PhaseCFG::fixup_trap_based_check(Node *branch, Block *block, int block_pos, Block *bnext) {
// Trap based checks must fall through to the successor with
// PROB_ALWAYS.
// They should be an If with 2 successors.
assert(branch->is_MachIf(), "must be If");
assert(block->_num_succs == 2, "must have 2 successors");
// Get the If node and the projection for the first successor.
MachIfNode *iff = block->get_node(block->number_of_nodes()-3)->as_MachIf();
ProjNode *proj0 = block->get_node(block->number_of_nodes()-2)->as_Proj();
ProjNode *proj1 = block->get_node(block->number_of_nodes()-1)->as_Proj();
ProjNode *projt = (proj0->Opcode() == Op_IfTrue) ? proj0 : proj1;
ProjNode *projf = (proj0->Opcode() == Op_IfFalse) ? proj0 : proj1;
// Assert that proj0 and succs[0] match up. Similarly for proj1 and succs[1].
assert(proj0->raw_out(0) == block->_succs[0]->head(), "Mismatch successor 0");
assert(proj1->raw_out(0) == block->_succs[1]->head(), "Mismatch successor 1");
ProjNode *proj_always;
ProjNode *proj_never;
// We must negate the branch if the implicit check doesn't follow
// the branch's TRUE path. Then, the new TRUE branch target will
// be the old FALSE branch target.
if (iff->_prob <= 2*PROB_NEVER) { // There are small rounding errors.
proj_never = projt;
proj_always = projf;
} else {
// We must negate the branch if the trap doesn't follow the
// branch's TRUE path. Then, the new TRUE branch target will
// be the old FALSE branch target.
proj_never = projf;
proj_always = projt;
iff->negate();
}
assert(iff->_prob <= 2*PROB_NEVER, "Trap based checks are expected to trap never!");
// Map the successors properly
block->_succs.map(0, get_block_for_node(proj_never ->raw_out(0))); // The target of the trap.
block->_succs.map(1, get_block_for_node(proj_always->raw_out(0))); // The fall through target.
if (block->get_node(block->number_of_nodes() - block->_num_succs + 1) != proj_always) {
block->map_node(proj_never, block->number_of_nodes() - block->_num_succs + 0);
block->map_node(proj_always, block->number_of_nodes() - block->_num_succs + 1);
}
// Place the fall through block after this block.
Block *bs1 = block->non_connector_successor(1);
if (bs1 != bnext && move_to_next(bs1, block_pos)) {
bnext = bs1;
}
// If the fall through block still is not the next block, insert a goto.
if (bs1 != bnext) {
insert_goto_at(block_pos, 1);
}
return bnext;
}
// Fix up the final control flow for basic blocks.
void PhaseCFG::fixup_flow() {
// Fixup final control flow for the blocks. Remove jump-to-next
// block. If neither arm of an IF follows the conditional branch, we
// have to add a second jump after the conditional. We place the
// TRUE branch target in succs[0] for both GOTOs and IFs.
for (uint i = 0; i < number_of_blocks(); i++) {
Block* block = get_block(i);
block->_pre_order = i; // turn pre-order into block-index
// Connector blocks need no further processing.
if (block->is_connector()) {
assert((i+1) == number_of_blocks() || get_block(i + 1)->is_connector(), "All connector blocks should sink to the end");
continue;
}
assert(block->is_Empty() != Block::completely_empty, "Empty blocks should be connectors");
Block* bnext = (i < number_of_blocks() - 1) ? get_block(i + 1) : NULL;
Block* bs0 = block->non_connector_successor(0);
// Check for multi-way branches where I cannot negate the test to
// exchange the true and false targets.
if (no_flip_branch(block)) {
// Find fall through case - if must fall into its target.
// Get the index of the branch's first successor.
int branch_idx = block->number_of_nodes() - block->_num_succs;
// The branch is 1 before the branch's first successor.
Node *branch = block->get_node(branch_idx-1);
// Handle no-flip branches which have implicit checks and which require
// special block ordering and individual semantics of the 'fall through
// case'.
if ((TrapBasedNullChecks || TrapBasedRangeChecks) &&
branch->is_Mach() && branch->as_Mach()->is_TrapBasedCheckNode()) {
bnext = fixup_trap_based_check(branch, block, i, bnext);
} else {
// Else, default handling for no-flip branches
for (uint j2 = 0; j2 < block->_num_succs; j2++) {
const ProjNode* p = block->get_node(branch_idx + j2)->as_Proj();
if (p->_con == 0) {
// successor j2 is fall through case
if (block->non_connector_successor(j2) != bnext) {
// but it is not the next block => insert a goto
insert_goto_at(i, j2);
}
// Put taken branch in slot 0
if (j2 == 0 && block->_num_succs == 2) {
// Flip targets in succs map
Block *tbs0 = block->_succs[0];
Block *tbs1 = block->_succs[1];
block->_succs.map(0, tbs1);
block->_succs.map(1, tbs0);
}
break;
}
}
}
// Remove all CatchProjs
for (uint j = 0; j < block->_num_succs; j++) {
block->pop_node();
}
} else if (block->_num_succs == 1) {
// Block ends in a Goto?
if (bnext == bs0) {
// We fall into next block; remove the Goto
block->pop_node();
}
} else if(block->_num_succs == 2) { // Block ends in a If?
// Get opcode of 1st projection (matches _succs[0])
// Note: Since this basic block has 2 exits, the last 2 nodes must
// be projections (in any order), the 3rd last node must be
// the IfNode (we have excluded other 2-way exits such as
// CatchNodes already).
MachNode* iff = block->get_node(block->number_of_nodes() - 3)->as_Mach();
ProjNode* proj0 = block->get_node(block->number_of_nodes() - 2)->as_Proj();
ProjNode* proj1 = block->get_node(block->number_of_nodes() - 1)->as_Proj();
// Assert that proj0 and succs[0] match up. Similarly for proj1 and succs[1].
assert(proj0->raw_out(0) == block->_succs[0]->head(), "Mismatch successor 0");
assert(proj1->raw_out(0) == block->_succs[1]->head(), "Mismatch successor 1");
Block* bs1 = block->non_connector_successor(1);
// Check for neither successor block following the current
// block ending in a conditional. If so, move one of the
// successors after the current one, provided that the
// successor was previously unscheduled, but moveable
// (i.e., all paths to it involve a branch).
if (!C->do_freq_based_layout() && bnext != bs0 && bnext != bs1) {
// Choose the more common successor based on the probability
// of the conditional branch.
Block* bx = bs0;
Block* by = bs1;
// _prob is the probability of taking the true path. Make
// p the probability of taking successor #1.
float p = iff->as_MachIf()->_prob;
if (proj0->Opcode() == Op_IfTrue) {
p = 1.0 - p;
}
// Prefer successor #1 if p > 0.5
if (p > PROB_FAIR) {
bx = bs1;
by = bs0;
}
// Attempt the more common successor first
if (move_to_next(bx, i)) {
bnext = bx;
} else if (move_to_next(by, i)) {
bnext = by;
}
}
// Check for conditional branching the wrong way. Negate
// conditional, if needed, so it falls into the following block
// and branches to the not-following block.
// Check for the next block being in succs[0]. We are going to branch
// to succs[0], so we want the fall-thru case as the next block in
// succs[1].
if (bnext == bs0) {
// Fall-thru case in succs[0], should be in succs[1], so flip targets in _succs map
Block* tbs0 = block->_succs[0];
Block* tbs1 = block->_succs[1];
block->_succs.map(0, tbs1);
block->_succs.map(1, tbs0);
// Flip projection for each target
swap(proj0, proj1);
} else if(bnext != bs1) {
// Need a double-branch
// The existing conditional branch need not change.
// Add a unconditional branch to the false target.
// Alas, it must appear in its own block and adding a
// block this late in the game is complicated. Sigh.
insert_goto_at(i, 1);
}
// Make sure we TRUE branch to the target
if (proj0->Opcode() == Op_IfFalse) {
iff->as_MachIf()->negate();
}
block->pop_node(); // Remove IfFalse & IfTrue projections
block->pop_node();
} else {
// Multi-exit block, e.g. a switch statement
// But we don't need to do anything here
}
} // End of for all blocks
}
void PhaseCFG::remove_unreachable_blocks() {
ResourceMark rm;
Block_List unreachable;
// Initialize worklist of unreachable blocks to be removed.
for (uint i = 0; i < number_of_blocks(); i++) {
Block* block = get_block(i);
assert(block->_pre_order == i, "Block::pre_order does not match block index");
if (block->is_trivially_unreachable()) {
unreachable.push(block);
}
}
// Now remove all blocks that are transitively unreachable.
while (unreachable.size() > 0) {
Block* dead = unreachable.pop();
// When this code runs (after PhaseCFG::fixup_flow()), Block::_pre_order
// does not contain pre-order but block-list indices. Ensure they stay
// contiguous by decrementing _pre_order for all elements after 'dead'.
// Block::_rpo does not contain valid reverse post-order indices anymore
// (they are invalidated by block insertions in PhaseCFG::fixup_flow()),
// so there is no need to update them.
for (uint i = dead->_pre_order + 1; i < number_of_blocks(); i++) {
get_block(i)->_pre_order--;
}
_blocks.remove(dead->_pre_order);
_number_of_blocks--;
// Update the successors' predecessor list and push new unreachable blocks.
for (uint i = 0; i < dead->_num_succs; i++) {
Block* succ = dead->_succs[i];
Node* head = succ->head();
for (int j = head->req() - 1; j >= 1; j--) {
if (get_block_for_node(head->in(j)) == dead) {
head->del_req(j);
}
}
if (succ->is_trivially_unreachable()) {
unreachable.push(succ);
}
}
}
}
// postalloc_expand: Expand nodes after register allocation.
//
// postalloc_expand has to be called after register allocation, just
// before output (i.e. scheduling). It only gets called if
// Matcher::require_postalloc_expand is true.
//
// Background:
//
// Nodes that are expandend (one compound node requiring several
// assembler instructions to be implemented split into two or more
// non-compound nodes) after register allocation are not as nice as
// the ones expanded before register allocation - they don't
// participate in optimizations as global code motion. But after
// register allocation we can expand nodes that use registers which
// are not spillable or registers that are not allocated, because the
// old compound node is simply replaced (in its location in the basic
// block) by a new subgraph which does not contain compound nodes any
// more. The scheduler called during output can later on process these
// non-compound nodes.
//
// Implementation:
//
// Nodes requiring postalloc expand are specified in the ad file by using
// a postalloc_expand statement instead of ins_encode. A postalloc_expand
// contains a single call to an encoding, as does an ins_encode
// statement. Instead of an emit() function a postalloc_expand() function
// is generated that doesn't emit assembler but creates a new
// subgraph. The code below calls this postalloc_expand function for each
// node with the appropriate attribute. This function returns the new
// nodes generated in an array passed in the call. The old node,
// potential MachTemps before and potential Projs after it then get
// disconnected and replaced by the new nodes. The instruction
// generating the result has to be the last one in the array. In
// general it is assumed that Projs after the node expanded are
// kills. These kills are not required any more after expanding as
// there are now explicitly visible def-use chains and the Projs are
// removed. This does not hold for calls: They do not only have
// kill-Projs but also Projs defining values. Therefore Projs after
// the node expanded are removed for all but for calls. If a node is
// to be reused, it must be added to the nodes list returned, and it
// will be added again.
//
// Implementing the postalloc_expand function for a node in an enc_class
// is rather tedious. It requires knowledge about many node details, as
// the nodes and the subgraph must be hand crafted. To simplify this,
// adlc generates some utility variables into the postalloc_expand function,
// e.g., holding the operands as specified by the postalloc_expand encoding
// specification, e.g.:
// * unsigned idx_<par_name> holding the index of the node in the ins
// * Node *n_<par_name> holding the node loaded from the ins
// * MachOpnd *op_<par_name> holding the corresponding operand
//
// The ordering of operands can not be determined by looking at a
// rule. Especially if a match rule matches several different trees,
// several nodes are generated from one instruct specification with
// different operand orderings. In this case the adlc generated
// variables are the only way to access the ins and operands
// deterministically.
//
// If assigning a register to a node that contains an oop, don't
// forget to call ra_->set_oop() for the node.
void PhaseCFG::postalloc_expand(PhaseRegAlloc* _ra) {
GrowableArray <Node *> new_nodes(32); // Array with new nodes filled by postalloc_expand function of node.
GrowableArray <Node *> remove(32);
GrowableArray <Node *> succs(32);
unsigned int max_idx = C->unique(); // Remember to distinguish new from old nodes.
DEBUG_ONLY(bool foundNode = false);
// for all blocks
for (uint i = 0; i < number_of_blocks(); i++) {
Block *b = _blocks[i];
// For all instructions in the current block.
for (uint j = 0; j < b->number_of_nodes(); j++) {
Node *n = b->get_node(j);
if (n->is_Mach() && n->as_Mach()->requires_postalloc_expand()) {
#ifdef ASSERT
if (TracePostallocExpand) {
if (!foundNode) {
foundNode = true;
tty->print("POSTALLOC EXPANDING %d %s\n", C->compile_id(),
C->method() ? C->method()->name()->as_utf8() : C->stub_name());
}
tty->print(" postalloc expanding "); n->dump();
if (Verbose) {
tty->print(" with ins:\n");
for (uint k = 0; k < n->len(); ++k) {
if (n->in(k)) { tty->print(" "); n->in(k)->dump(); }
}
}
}
#endif
new_nodes.clear();
// Collect nodes that have to be removed from the block later on.
uint req = n->req();
remove.clear();
for (uint k = 0; k < req; ++k) {
if (n->in(k) && n->in(k)->is_MachTemp()) {
remove.push(n->in(k)); // MachTemps which are inputs to the old node have to be removed.
n->in(k)->del_req(0);
j--;
}
}
// Check whether we can allocate enough nodes. We set a fix limit for
// the size of postalloc expands with this.
uint unique_limit = C->unique() + 40;
if (unique_limit >= _ra->node_regs_max_index()) {
Compile::current()->record_failure("out of nodes in postalloc expand");
return;
}
// Emit (i.e. generate new nodes).
n->as_Mach()->postalloc_expand(&new_nodes, _ra);
assert(C->unique() < unique_limit, "You allocated too many nodes in your postalloc expand.");
// Disconnect the inputs of the old node.
//
// We reuse MachSpillCopy nodes. If we need to expand them, there
// are many, so reusing pays off. If reused, the node already
// has the new ins. n must be the last node on new_nodes list.
if (!n->is_MachSpillCopy()) {
for (int k = req - 1; k >= 0; --k) {
n->del_req(k);
}
}
#ifdef ASSERT
// Check that all nodes have proper operands.
for (int k = 0; k < new_nodes.length(); ++k) {
if (new_nodes.at(k)->_idx < max_idx || !new_nodes.at(k)->is_Mach()) continue; // old node, Proj ...
MachNode *m = new_nodes.at(k)->as_Mach();
for (unsigned int l = 0; l < m->num_opnds(); ++l) {
if (MachOper::notAnOper(m->_opnds[l])) {
outputStream *os = tty;
os->print("Node %s ", m->Name());
os->print("has invalid opnd %d: %p\n", l, m->_opnds[l]);
assert(0, "Invalid operands, see inline trace in hs_err_pid file.");
}
}
}
#endif
// Collect succs of old node in remove (for projections) and in succs (for
// all other nodes) do _not_ collect projections in remove (but in succs)
// in case the node is a call. We need the projections for calls as they are
// associated with registers (i.e. they are defs).
succs.clear();
for (DUIterator k = n->outs(); n->has_out(k); k++) {
if (n->out(k)->is_Proj() && !n->is_MachCall() && !n->is_MachBranch()) {
remove.push(n->out(k));
} else {
succs.push(n->out(k));
}
}
// Replace old node n as input of its succs by last of the new nodes.
for (int k = 0; k < succs.length(); ++k) {
Node *succ = succs.at(k);
for (uint l = 0; l < succ->req(); ++l) {
if (succ->in(l) == n) {
succ->set_req(l, new_nodes.at(new_nodes.length() - 1));
}
}
for (uint l = succ->req(); l < succ->len(); ++l) {
if (succ->in(l) == n) {
succ->set_prec(l, new_nodes.at(new_nodes.length() - 1));
}
}
}
// Index of old node in block.
uint index = b->find_node(n);
// Insert new nodes into block and map them in nodes->blocks array
// and remember last node in n2.
Node *n2 = NULL;
for (int k = 0; k < new_nodes.length(); ++k) {
n2 = new_nodes.at(k);
b->insert_node(n2, ++index);
map_node_to_block(n2, b);
}
// Add old node n to remove and remove them all from block.
remove.push(n);
j--;
#ifdef ASSERT
if (TracePostallocExpand && Verbose) {
tty->print(" removing:\n");
for (int k = 0; k < remove.length(); ++k) {
tty->print(" "); remove.at(k)->dump();
}
tty->print(" inserting:\n");
for (int k = 0; k < new_nodes.length(); ++k) {
tty->print(" "); new_nodes.at(k)->dump();
}
}
#endif
for (int k = 0; k < remove.length(); ++k) {
if (b->contains(remove.at(k))) {
b->find_remove(remove.at(k));
} else {
assert(remove.at(k)->is_Proj() && (remove.at(k)->in(0)->is_MachBranch()), "");
}
}
// If anything has been inserted (n2 != NULL), continue after last node inserted.
// This does not always work. Some postalloc expands don't insert any nodes, if they
// do optimizations (e.g., max(x,x)). In this case we decrement j accordingly.
j = n2 ? b->find_node(n2) : j;
}
}
}
#ifdef ASSERT
if (foundNode) {
tty->print("FINISHED %d %s\n", C->compile_id(),
C->method() ? C->method()->name()->as_utf8() : C->stub_name());
tty->flush();
}
#endif
}
//------------------------------dump-------------------------------------------
#ifndef PRODUCT
void PhaseCFG::_dump_cfg( const Node *end, VectorSet &visited ) const {
const Node *x = end->is_block_proj();
assert( x, "not a CFG" );
// Do not visit this block again
if( visited.test_set(x->_idx) ) return;
// Skip through this block
const Node *p = x;
do {
p = p->in(0); // Move control forward
assert( !p->is_block_proj() || p->is_Root(), "not a CFG" );
} while( !p->is_block_start() );
// Recursively visit
for (uint i = 1; i < p->req(); i++) {
_dump_cfg(p->in(i), visited);
}
// Dump the block
get_block_for_node(p)->dump(this);
}
void PhaseCFG::dump( ) const {
tty->print("\n--- CFG --- %d BBs\n", number_of_blocks());
if (_blocks.size()) { // Did we do basic-block layout?
for (uint i = 0; i < number_of_blocks(); i++) {
const Block* block = get_block(i);
block->dump(this);
}
} else { // Else do it with a DFS
VectorSet visited(_block_arena);
_dump_cfg(_root,visited);
}
}
void PhaseCFG::dump_headers() {
for (uint i = 0; i < number_of_blocks(); i++) {
Block* block = get_block(i);
if (block != NULL) {
block->dump_head(this);
}
}
}
#endif // !PRODUCT
#ifdef ASSERT
void PhaseCFG::verify_memory_writer_placement(const Block* b, const Node* n) const {
if (!n->is_memory_writer()) {
return;
}
CFGLoop* home_or_ancestor = find_block_for_node(n->in(0))->_loop;
bool found = false;
do {
if (b->_loop == home_or_ancestor) {
found = true;
break;
}
home_or_ancestor = home_or_ancestor->parent();
} while (home_or_ancestor != NULL);
assert(found, "block b is not in n's home loop or an ancestor of it");
}
void PhaseCFG::verify_dominator_tree() const {
for (uint i = 0; i < number_of_blocks(); i++) {
Block* block = get_block(i);
assert(block->_dom_depth <= number_of_blocks(), "unexpected dominator tree depth");
if (block == get_root_block()) {
assert(block->_dom_depth == 1, "unexpected root dominator tree depth");
// The root block does not have an immediate dominator, stop checking.
continue;
}
assert(block->_idom != nullptr, "non-root blocks must have immediate dominators");
assert(block->_dom_depth == block->_idom->_dom_depth + 1,
"the dominator tree depth of a node must succeed that of its immediate dominator");
assert(block->num_preds() > 2 || block->_idom == get_block_for_node(block->pred(1)),
"the immediate dominator of a single-predecessor block must be the predecessor");
}
}
void PhaseCFG::verify() const {
// Verify sane CFG
for (uint i = 0; i < number_of_blocks(); i++) {
Block* block = get_block(i);
uint cnt = block->number_of_nodes();
uint j;
for (j = 0; j < cnt; j++) {
Node *n = block->get_node(j);
assert(get_block_for_node(n) == block, "");
if (j >= 1 && n->is_Mach() && n->as_Mach()->ideal_Opcode() == Op_CreateEx) {
assert(j == 1 || block->get_node(j-1)->is_Phi(), "CreateEx must be first instruction in block");
}
verify_memory_writer_placement(block, n);
if (n->needs_anti_dependence_check()) {
verify_anti_dependences(block, n);
}
for (uint k = 0; k < n->req(); k++) {
Node *def = n->in(k);
if (def && def != n) {
Block* def_block = get_block_for_node(def);
assert(def_block || def->is_Con(), "must have block; constants for debug info ok");
// Verify that all definitions dominate their uses (except for virtual
// instructions merging multiple definitions).
assert(n->is_Root() || n->is_Region() || n->is_Phi() || n->is_MachMerge() ||
def_block->dominates(block),
"uses must be dominated by definitions");
// Verify that instructions in the block are in correct order.
// Uses must follow their definition if they are at the same block.
// Mostly done to check that MachSpillCopy nodes are placed correctly
// when CreateEx node is moved in build_ifg_physical().
if (def_block == block && !(block->head()->is_Loop() && n->is_Phi()) &&
// See (+++) comment in reg_split.cpp
!(n->jvms() != NULL && n->jvms()->is_monitor_use(k))) {
bool is_loop = false;
if (n->is_Phi()) {
for (uint l = 1; l < def->req(); l++) {
if (n == def->in(l)) {
is_loop = true;
break; // Some kind of loop
}
}
}
assert(is_loop || block->find_node(def) < j, "uses must follow definitions");
}
}
}
if (n->is_Proj()) {
assert(j >= 1, "a projection cannot be the first instruction in a block");
Node* pred = block->get_node(j - 1);
Node* parent = n->in(0);
assert(parent != NULL, "projections must have a parent");
assert(pred == parent || (pred->is_Proj() && pred->in(0) == parent),
"projections must follow their parents or other sibling projections");
}
}
j = block->end_idx();
Node* bp = (Node*)block->get_node(block->number_of_nodes() - 1)->is_block_proj();
assert(bp, "last instruction must be a block proj");
assert(bp == block->get_node(j), "wrong number of successors for this block");
if (bp->is_Catch()) {
while (block->get_node(--j)->is_MachProj()) {
;
}
assert(block->get_node(j)->is_MachCall(), "CatchProj must follow call");
} else if (bp->is_Mach() && bp->as_Mach()->ideal_Opcode() == Op_If) {
assert(block->_num_succs == 2, "Conditional branch must have two targets");
}
}
verify_dominator_tree();
}
#endif // ASSERT
UnionFind::UnionFind( uint max ) : _cnt(max), _max(max), _indices(NEW_RESOURCE_ARRAY(uint,max)) {
Copy::zero_to_bytes( _indices, sizeof(uint)*max );
}
void UnionFind::extend( uint from_idx, uint to_idx ) {
_nesting.check();
if( from_idx >= _max ) {
uint size = 16;
while( size <= from_idx ) size <<=1;
_indices = REALLOC_RESOURCE_ARRAY( uint, _indices, _max, size );
_max = size;
}
while( _cnt <= from_idx ) _indices[_cnt++] = 0;
_indices[from_idx] = to_idx;
}
void UnionFind::reset( uint max ) {
// Force the Union-Find mapping to be at least this large
extend(max,0);
// Initialize to be the ID mapping.
for( uint i=0; i<max; i++ ) map(i,i);
}
// Straight out of Tarjan's union-find algorithm
uint UnionFind::Find_compress( uint idx ) {
uint cur = idx;
uint next = lookup(cur);
while( next != cur ) { // Scan chain of equivalences
assert( next < cur, "always union smaller" );
cur = next; // until find a fixed-point
next = lookup(cur);
}
// Core of union-find algorithm: update chain of
// equivalences to be equal to the root.
while( idx != next ) {
uint tmp = lookup(idx);
map(idx, next);
idx = tmp;
}
return idx;
}
// Like Find above, but no path compress, so bad asymptotic behavior
uint UnionFind::Find_const( uint idx ) const {
if( idx == 0 ) return idx; // Ignore the zero idx
// Off the end? This can happen during debugging dumps
// when data structures have not finished being updated.
if( idx >= _max ) return idx;
uint next = lookup(idx);
while( next != idx ) { // Scan chain of equivalences
idx = next; // until find a fixed-point
next = lookup(idx);
}
return next;
}
// union 2 sets together.
void UnionFind::Union( uint idx1, uint idx2 ) {
uint src = Find(idx1);
uint dst = Find(idx2);
assert( src, "" );
assert( dst, "" );
assert( src < _max, "oob" );
assert( dst < _max, "oob" );
assert( src < dst, "always union smaller" );
map(dst,src);
}
#ifndef PRODUCT
void Trace::dump( ) const {
tty->print_cr("Trace (freq %f)", first_block()->_freq);
for (Block *b = first_block(); b != NULL; b = next(b)) {
tty->print(" B%d", b->_pre_order);
if (b->head()->is_Loop()) {
tty->print(" (L%d)", b->compute_loop_alignment());
}
if (b->has_loop_alignment()) {
tty->print(" (T%d)", b->code_alignment());
}
}
tty->cr();
}
void CFGEdge::dump( ) const {
tty->print(" B%d --> B%d Freq: %f out:%3d%% in:%3d%% State: ",
from()->_pre_order, to()->_pre_order, freq(), _from_pct, _to_pct);
switch(state()) {
case connected:
tty->print("connected");
break;
case open:
tty->print("open");
break;
case interior:
tty->print("interior");
break;
}
if (infrequent()) {
tty->print(" infrequent");
}
tty->cr();
}
#endif
// Comparison function for edges
static int edge_order(CFGEdge **e0, CFGEdge **e1) {
float freq0 = (*e0)->freq();
float freq1 = (*e1)->freq();
if (freq0 != freq1) {
return freq0 > freq1 ? -1 : 1;
}
int dist0 = (*e0)->to()->_rpo - (*e0)->from()->_rpo;
int dist1 = (*e1)->to()->_rpo - (*e1)->from()->_rpo;
return dist1 - dist0;
}
// Comparison function for edges
extern "C" int trace_frequency_order(const void *p0, const void *p1) {
Trace *tr0 = *(Trace **) p0;
Trace *tr1 = *(Trace **) p1;
Block *b0 = tr0->first_block();
Block *b1 = tr1->first_block();
// The trace of connector blocks goes at the end;
// we only expect one such trace
if (b0->is_connector() != b1->is_connector()) {
return b1->is_connector() ? -1 : 1;
}
// Pull more frequently executed blocks to the beginning
float freq0 = b0->_freq;
float freq1 = b1->_freq;
if (freq0 != freq1) {
return freq0 > freq1 ? -1 : 1;
}
int diff = tr0->first_block()->_rpo - tr1->first_block()->_rpo;
return diff;
}
// Find edges of interest, i.e, those which can fall through. Presumes that
// edges which don't fall through are of low frequency and can be generally
// ignored. Initialize the list of traces.
void PhaseBlockLayout::find_edges() {
// Walk the blocks, creating edges and Traces
uint i;
Trace *tr = NULL;
for (i = 0; i < _cfg.number_of_blocks(); i++) {
Block* b = _cfg.get_block(i);
tr = new Trace(b, next, prev);
traces[tr->id()] = tr;
// All connector blocks should be at the end of the list
if (b->is_connector()) break;
// If this block and the next one have a one-to-one successor
// predecessor relationship, simply append the next block
int nfallthru = b->num_fall_throughs();
while (nfallthru == 1 &&
b->succ_fall_through(0)) {
Block *n = b->_succs[0];
// Skip over single-entry connector blocks, we don't want to
// add them to the trace.
while (n->is_connector() && n->num_preds() == 1) {
n = n->_succs[0];
}
// We see a merge point, so stop search for the next block
if (n->num_preds() != 1) break;
i++;
assert(n == _cfg.get_block(i), "expecting next block");
tr->append(n);
uf->map(n->_pre_order, tr->id());
traces[n->_pre_order] = NULL;
nfallthru = b->num_fall_throughs();
b = n;
}
if (nfallthru > 0) {
// Create a CFGEdge for each outgoing
// edge that could be a fall-through.
for (uint j = 0; j < b->_num_succs; j++ ) {
if (b->succ_fall_through(j)) {
Block *target = b->non_connector_successor(j);
float freq = b->_freq * b->succ_prob(j);
int from_pct = (int) ((100 * freq) / b->_freq);
int to_pct = (int) ((100 * freq) / target->_freq);
edges->append(new CFGEdge(b, target, freq, from_pct, to_pct));
}
}
}
}
// Group connector blocks into one trace
for (i++; i < _cfg.number_of_blocks(); i++) {
Block *b = _cfg.get_block(i);
assert(b->is_connector(), "connector blocks at the end");
tr->append(b);
uf->map(b->_pre_order, tr->id());
traces[b->_pre_order] = NULL;
}
}
// Union two traces together in uf, and null out the trace in the list
void PhaseBlockLayout::union_traces(Trace* updated_trace, Trace* old_trace) {
uint old_id = old_trace->id();
uint updated_id = updated_trace->id();
uint lo_id = updated_id;
uint hi_id = old_id;
// If from is greater than to, swap values to meet
// UnionFind guarantee.
if (updated_id > old_id) {
lo_id = old_id;
hi_id = updated_id;
// Fix up the trace ids
traces[lo_id] = traces[updated_id];
updated_trace->set_id(lo_id);
}
// Union the lower with the higher and remove the pointer
// to the higher.
uf->Union(lo_id, hi_id);
traces[hi_id] = NULL;
}
// Append traces together via the most frequently executed edges
void PhaseBlockLayout::grow_traces() {
// Order the edges, and drive the growth of Traces via the most
// frequently executed edges.
edges->sort(edge_order);
for (int i = 0; i < edges->length(); i++) {
CFGEdge *e = edges->at(i);
if (e->state() != CFGEdge::open) continue;
Block *src_block = e->from();
Block *targ_block = e->to();
// Don't grow traces along backedges?
if (!BlockLayoutRotateLoops) {
if (targ_block->_rpo <= src_block->_rpo) {
targ_block->set_loop_alignment(targ_block);
continue;
}
}
Trace *src_trace = trace(src_block);
Trace *targ_trace = trace(targ_block);
// If the edge in question can join two traces at their ends,
// append one trace to the other.
if (src_trace->last_block() == src_block) {
if (src_trace == targ_trace) {
e->set_state(CFGEdge::interior);
if (targ_trace->backedge(e)) {
// Reset i to catch any newly eligible edge
// (Or we could remember the first "open" edge, and reset there)
i = 0;
}
} else if (targ_trace->first_block() == targ_block) {
e->set_state(CFGEdge::connected);
src_trace->append(targ_trace);
union_traces(src_trace, targ_trace);
}
}
}
}
// Embed one trace into another, if the fork or join points are sufficiently
// balanced.
void PhaseBlockLayout::merge_traces(bool fall_thru_only) {
// Walk the edge list a another time, looking at unprocessed edges.
// Fold in diamonds
for (int i = 0; i < edges->length(); i++) {
CFGEdge *e = edges->at(i);
if (e->state() != CFGEdge::open) continue;
if (fall_thru_only) {
if (e->infrequent()) continue;
}
Block *src_block = e->from();
Trace *src_trace = trace(src_block);
bool src_at_tail = src_trace->last_block() == src_block;
Block *targ_block = e->to();
Trace *targ_trace = trace(targ_block);
bool targ_at_start = targ_trace->first_block() == targ_block;
if (src_trace == targ_trace) {
// This may be a loop, but we can't do much about it.
e->set_state(CFGEdge::interior);
continue;
}
if (fall_thru_only) {
// If the edge links the middle of two traces, we can't do anything.
// Mark the edge and continue.
if (!src_at_tail & !targ_at_start) {
continue;
}
// Don't grow traces along backedges?
if (!BlockLayoutRotateLoops && (targ_block->_rpo <= src_block->_rpo)) {
continue;
}
// If both ends of the edge are available, why didn't we handle it earlier?
assert(src_at_tail ^ targ_at_start, "Should have caught this edge earlier.");
if (targ_at_start) {
// Insert the "targ" trace in the "src" trace if the insertion point
// is a two way branch.
// Better profitability check possible, but may not be worth it.
// Someday, see if the this "fork" has an associated "join";
// then make a policy on merging this trace at the fork or join.
// For example, other things being equal, it may be better to place this
// trace at the join point if the "src" trace ends in a two-way, but
// the insertion point is one-way.
assert(src_block->num_fall_throughs() == 2, "unexpected diamond");
e->set_state(CFGEdge::connected);
src_trace->insert_after(src_block, targ_trace);
union_traces(src_trace, targ_trace);
} else if (src_at_tail) {
if (src_trace != trace(_cfg.get_root_block())) {
e->set_state(CFGEdge::connected);
targ_trace->insert_before(targ_block, src_trace);
union_traces(targ_trace, src_trace);
}
}
} else if (e->state() == CFGEdge::open) {
// Append traces, even without a fall-thru connection.
// But leave root entry at the beginning of the block list.
if (targ_trace != trace(_cfg.get_root_block())) {
e->set_state(CFGEdge::connected);
src_trace->append(targ_trace);
union_traces(src_trace, targ_trace);
}
}
}
}
// Order the sequence of the traces in some desirable way
void PhaseBlockLayout::reorder_traces(int count) {
ResourceArea *area = Thread::current()->resource_area();
Trace ** new_traces = NEW_ARENA_ARRAY(area, Trace *, count);
Block_List worklist;
int new_count = 0;
// Compact the traces.
for (int i = 0; i < count; i++) {
Trace *tr = traces[i];
if (tr != NULL) {
new_traces[new_count++] = tr;
}
}
// The entry block should be first on the new trace list.
Trace *tr = trace(_cfg.get_root_block());
assert(tr == new_traces[0], "entry trace misplaced");
// Sort the new trace list by frequency
qsort(new_traces + 1, new_count - 1, sizeof(new_traces[0]), trace_frequency_order);
// Collect all blocks from existing Traces
_cfg.clear_blocks();
for (int i = 0; i < new_count; i++) {
Trace *tr = new_traces[i];
if (tr != NULL) {
// push blocks onto the CFG list
for (Block* b = tr->first_block(); b != NULL; b = tr->next(b)) {
_cfg.add_block(b);
}
}
}
}
// Order basic blocks based on frequency
PhaseBlockLayout::PhaseBlockLayout(PhaseCFG &cfg)
: Phase(BlockLayout)
, _cfg(cfg) {
ResourceMark rm;
ResourceArea *area = Thread::current()->resource_area();
// List of traces
int size = _cfg.number_of_blocks() + 1;
traces = NEW_ARENA_ARRAY(area, Trace *, size);
memset(traces, 0, size*sizeof(Trace*));
next = NEW_ARENA_ARRAY(area, Block *, size);
memset(next, 0, size*sizeof(Block *));
prev = NEW_ARENA_ARRAY(area, Block *, size);
memset(prev , 0, size*sizeof(Block *));
// List of edges
edges = new GrowableArray<CFGEdge*>;
// Mapping block index --> block_trace
uf = new UnionFind(size);
uf->reset(size);
// Find edges and create traces.
find_edges();
// Grow traces at their ends via most frequent edges.
grow_traces();
// Merge one trace into another, but only at fall-through points.
// This may make diamonds and other related shapes in a trace.
merge_traces(true);
// Run merge again, allowing two traces to be catenated, even if
--> --------------------
--> maximum size reached
--> --------------------
¤ Dauer der Verarbeitung: 0.65 Sekunden
(vorverarbeitet)
¤
|
Haftungshinweis
Die Informationen auf dieser Webseite wurden
nach bestem Wissen sorgfältig zusammengestellt. Es wird jedoch weder Vollständigkeit, noch Richtigkeit,
noch Qualität der bereit gestellten Informationen zugesichert.
Bemerkung:
Die farbliche Syntaxdarstellung ist noch experimentell.
|
