238 continue; // Give up is reference is beyond 4K page size |
238 continue; // Give up is reference is beyond 4K page size |
239 } |
239 } |
240 } |
240 } |
241 |
241 |
242 // Check ctrl input to see if the null-check dominates the memory op |
242 // Check ctrl input to see if the null-check dominates the memory op |
243 Block *cb = cfg->_bbs[mach->_idx]; |
243 Block *cb = cfg->get_block_for_node(mach); |
244 cb = cb->_idom; // Always hoist at least 1 block |
244 cb = cb->_idom; // Always hoist at least 1 block |
245 if( !was_store ) { // Stores can be hoisted only one block |
245 if( !was_store ) { // Stores can be hoisted only one block |
246 while( cb->_dom_depth > (_dom_depth + 1)) |
246 while( cb->_dom_depth > (_dom_depth + 1)) |
247 cb = cb->_idom; // Hoist loads as far as we want |
247 cb = cb->_idom; // Hoist loads as far as we want |
248 // The non-null-block should dominate the memory op, too. Live |
248 // The non-null-block should dominate the memory op, too. Live |
263 vidx = j; |
263 vidx = j; |
264 // Ignore DecodeN val which could be hoisted to where needed. |
264 // Ignore DecodeN val which could be hoisted to where needed. |
265 if( is_decoden ) continue; |
265 if( is_decoden ) continue; |
266 } |
266 } |
267 // Block of memory-op input |
267 // Block of memory-op input |
268 Block *inb = cfg->_bbs[mach->in(j)->_idx]; |
268 Block *inb = cfg->get_block_for_node(mach->in(j)); |
269 Block *b = this; // Start from nul check |
269 Block *b = this; // Start from nul check |
270 while( b != inb && b->_dom_depth > inb->_dom_depth ) |
270 while( b != inb && b->_dom_depth > inb->_dom_depth ) |
271 b = b->_idom; // search upwards for input |
271 b = b->_idom; // search upwards for input |
272 // See if input dominates null check |
272 // See if input dominates null check |
273 if( b != inb ) |
273 if( b != inb ) |
274 break; |
274 break; |
275 } |
275 } |
276 if( j > 0 ) |
276 if( j > 0 ) |
277 continue; |
277 continue; |
278 Block *mb = cfg->_bbs[mach->_idx]; |
278 Block *mb = cfg->get_block_for_node(mach); |
279 // Hoisting stores requires more checks for the anti-dependence case. |
279 // Hoisting stores requires more checks for the anti-dependence case. |
280 // Give up hoisting if we have to move the store past any load. |
280 // Give up hoisting if we have to move the store past any load. |
281 if( was_store ) { |
281 if( was_store ) { |
282 Block *b = mb; // Start searching here for a local load |
282 Block *b = mb; // Start searching here for a local load |
283 // mach use (faulting) trying to hoist |
283 // mach use (faulting) trying to hoist |
304 if( e->is_MachNullCheck() && e->in(1) == mach ) |
304 if( e->is_MachNullCheck() && e->in(1) == mach ) |
305 continue; // Already being used as a NULL check |
305 continue; // Already being used as a NULL check |
306 |
306 |
307 // Found a candidate! Pick one with least dom depth - the highest |
307 // Found a candidate! Pick one with least dom depth - the highest |
308 // in the dom tree should be closest to the null check. |
308 // in the dom tree should be closest to the null check. |
309 if( !best || |
309 if (best == NULL || cfg->get_block_for_node(mach)->_dom_depth < cfg->get_block_for_node(best)->_dom_depth) { |
310 cfg->_bbs[mach->_idx]->_dom_depth < cfg->_bbs[best->_idx]->_dom_depth ) { |
|
311 best = mach; |
310 best = mach; |
312 bidx = vidx; |
311 bidx = vidx; |
313 |
|
314 } |
312 } |
315 } |
313 } |
316 // No candidate! |
314 // No candidate! |
317 if( !best ) return; |
315 if (best == NULL) { |
|
316 return; |
|
317 } |
318 |
318 |
319 // ---- Found an implicit null check |
319 // ---- Found an implicit null check |
320 extern int implicit_null_checks; |
320 extern int implicit_null_checks; |
321 implicit_null_checks++; |
321 implicit_null_checks++; |
322 |
322 |
323 if( is_decoden ) { |
323 if( is_decoden ) { |
324 // Check if we need to hoist decodeHeapOop_not_null first. |
324 // Check if we need to hoist decodeHeapOop_not_null first. |
325 Block *valb = cfg->_bbs[val->_idx]; |
325 Block *valb = cfg->get_block_for_node(val); |
326 if( this != valb && this->_dom_depth < valb->_dom_depth ) { |
326 if( this != valb && this->_dom_depth < valb->_dom_depth ) { |
327 // Hoist it up to the end of the test block. |
327 // Hoist it up to the end of the test block. |
328 valb->find_remove(val); |
328 valb->find_remove(val); |
329 this->add_inst(val); |
329 this->add_inst(val); |
330 cfg->_bbs.map(val->_idx,this); |
330 cfg->map_node_to_block(val, this); |
331 // DecodeN on x86 may kill flags. Check for flag-killing projections |
331 // DecodeN on x86 may kill flags. Check for flag-killing projections |
332 // that also need to be hoisted. |
332 // that also need to be hoisted. |
333 for (DUIterator_Fast jmax, j = val->fast_outs(jmax); j < jmax; j++) { |
333 for (DUIterator_Fast jmax, j = val->fast_outs(jmax); j < jmax; j++) { |
334 Node* n = val->fast_out(j); |
334 Node* n = val->fast_out(j); |
335 if( n->is_MachProj() ) { |
335 if( n->is_MachProj() ) { |
336 cfg->_bbs[n->_idx]->find_remove(n); |
336 cfg->get_block_for_node(n)->find_remove(n); |
337 this->add_inst(n); |
337 this->add_inst(n); |
338 cfg->_bbs.map(n->_idx,this); |
338 cfg->map_node_to_block(n, this); |
339 } |
339 } |
340 } |
340 } |
341 } |
341 } |
342 } |
342 } |
343 // Hoist the memory candidate up to the end of the test block. |
343 // Hoist the memory candidate up to the end of the test block. |
344 Block *old_block = cfg->_bbs[best->_idx]; |
344 Block *old_block = cfg->get_block_for_node(best); |
345 old_block->find_remove(best); |
345 old_block->find_remove(best); |
346 add_inst(best); |
346 add_inst(best); |
347 cfg->_bbs.map(best->_idx,this); |
347 cfg->map_node_to_block(best, this); |
348 |
348 |
349 // Move the control dependence |
349 // Move the control dependence |
350 if (best->in(0) && best->in(0) == old_block->_nodes[0]) |
350 if (best->in(0) && best->in(0) == old_block->_nodes[0]) |
351 best->set_req(0, _nodes[0]); |
351 best->set_req(0, _nodes[0]); |
352 |
352 |
353 // Check for flag-killing projections that also need to be hoisted |
353 // Check for flag-killing projections that also need to be hoisted |
354 // Should be DU safe because no edge updates. |
354 // Should be DU safe because no edge updates. |
355 for (DUIterator_Fast jmax, j = best->fast_outs(jmax); j < jmax; j++) { |
355 for (DUIterator_Fast jmax, j = best->fast_outs(jmax); j < jmax; j++) { |
356 Node* n = best->fast_out(j); |
356 Node* n = best->fast_out(j); |
357 if( n->is_MachProj() ) { |
357 if( n->is_MachProj() ) { |
358 cfg->_bbs[n->_idx]->find_remove(n); |
358 cfg->get_block_for_node(n)->find_remove(n); |
359 add_inst(n); |
359 add_inst(n); |
360 cfg->_bbs.map(n->_idx,this); |
360 cfg->map_node_to_block(n, this); |
361 } |
361 } |
362 } |
362 } |
363 |
363 |
364 Compile *C = cfg->C; |
364 Compile *C = cfg->C; |
365 // proj==Op_True --> ne test; proj==Op_False --> eq test. |
365 // proj==Op_True --> ne test; proj==Op_False --> eq test. |
386 // Since schedule-local needs precise def-use info, we need to correct |
386 // Since schedule-local needs precise def-use info, we need to correct |
387 // it as well. |
387 // it as well. |
388 Node *old_tst = proj->in(0); |
388 Node *old_tst = proj->in(0); |
389 MachNode *nul_chk = new (C) MachNullCheckNode(old_tst->in(0),best,bidx); |
389 MachNode *nul_chk = new (C) MachNullCheckNode(old_tst->in(0),best,bidx); |
390 _nodes.map(end_idx(),nul_chk); |
390 _nodes.map(end_idx(),nul_chk); |
391 cfg->_bbs.map(nul_chk->_idx,this); |
391 cfg->map_node_to_block(nul_chk, this); |
392 // Redirect users of old_test to nul_chk |
392 // Redirect users of old_test to nul_chk |
393 for (DUIterator_Last i2min, i2 = old_tst->last_outs(i2min); i2 >= i2min; --i2) |
393 for (DUIterator_Last i2min, i2 = old_tst->last_outs(i2min); i2 >= i2min; --i2) |
394 old_tst->last_out(i2)->set_req(0, nul_chk); |
394 old_tst->last_out(i2)->set_req(0, nul_chk); |
395 // Clean-up any dead code |
395 // Clean-up any dead code |
396 for (uint i3 = 0; i3 < old_tst->req(); i3++) |
396 for (uint i3 = 0; i3 < old_tst->req(); i3++) |
530 return n; |
530 return n; |
531 } |
531 } |
532 |
532 |
533 |
533 |
534 //------------------------------set_next_call---------------------------------- |
534 //------------------------------set_next_call---------------------------------- |
535 void Block::set_next_call( Node *n, VectorSet &next_call, Block_Array &bbs ) { |
535 void Block::set_next_call( Node *n, VectorSet &next_call, PhaseCFG* cfg) { |
536 if( next_call.test_set(n->_idx) ) return; |
536 if( next_call.test_set(n->_idx) ) return; |
537 for( uint i=0; i<n->len(); i++ ) { |
537 for( uint i=0; i<n->len(); i++ ) { |
538 Node *m = n->in(i); |
538 Node *m = n->in(i); |
539 if( !m ) continue; // must see all nodes in block that precede call |
539 if( !m ) continue; // must see all nodes in block that precede call |
540 if( bbs[m->_idx] == this ) |
540 if (cfg->get_block_for_node(m) == this) { |
541 set_next_call( m, next_call, bbs ); |
541 set_next_call(m, next_call, cfg); |
|
542 } |
542 } |
543 } |
543 } |
544 } |
544 |
545 |
545 //------------------------------needed_for_next_call--------------------------- |
546 //------------------------------needed_for_next_call--------------------------- |
546 // Set the flag 'next_call' for each Node that is needed for the next call to |
547 // Set the flag 'next_call' for each Node that is needed for the next call to |
547 // be scheduled. This flag lets me bias scheduling so Nodes needed for the |
548 // be scheduled. This flag lets me bias scheduling so Nodes needed for the |
548 // next subroutine call get priority - basically it moves things NOT needed |
549 // next subroutine call get priority - basically it moves things NOT needed |
549 // for the next call till after the call. This prevents me from trying to |
550 // for the next call till after the call. This prevents me from trying to |
550 // carry lots of stuff live across a call. |
551 // carry lots of stuff live across a call. |
551 void Block::needed_for_next_call(Node *this_call, VectorSet &next_call, Block_Array &bbs) { |
552 void Block::needed_for_next_call(Node *this_call, VectorSet &next_call, PhaseCFG* cfg) { |
552 // Find the next control-defining Node in this block |
553 // Find the next control-defining Node in this block |
553 Node* call = NULL; |
554 Node* call = NULL; |
554 for (DUIterator_Fast imax, i = this_call->fast_outs(imax); i < imax; i++) { |
555 for (DUIterator_Fast imax, i = this_call->fast_outs(imax); i < imax; i++) { |
555 Node* m = this_call->fast_out(i); |
556 Node* m = this_call->fast_out(i); |
556 if( bbs[m->_idx] == this && // Local-block user |
557 if(cfg->get_block_for_node(m) == this && // Local-block user |
557 m != this_call && // Not self-start node |
558 m != this_call && // Not self-start node |
558 m->is_MachCall() ) |
559 m->is_MachCall() ) |
559 call = m; |
560 call = m; |
560 break; |
561 break; |
561 } |
562 } |
562 if (call == NULL) return; // No next call (e.g., block end is near) |
563 if (call == NULL) return; // No next call (e.g., block end is near) |
563 // Set next-call for all inputs to this call |
564 // Set next-call for all inputs to this call |
564 set_next_call(call, next_call, bbs); |
565 set_next_call(call, next_call, cfg); |
565 } |
566 } |
566 |
567 |
567 //------------------------------add_call_kills------------------------------------- |
568 //------------------------------add_call_kills------------------------------------- |
568 void Block::add_call_kills(MachProjNode *proj, RegMask& regs, const char* save_policy, bool exclude_soe) { |
569 void Block::add_call_kills(MachProjNode *proj, RegMask& regs, const char* save_policy, bool exclude_soe) { |
569 // Fill in the kill mask for the call |
570 // Fill in the kill mask for the call |
579 } |
580 } |
580 } |
581 } |
581 |
582 |
582 |
583 |
583 //------------------------------sched_call------------------------------------- |
584 //------------------------------sched_call------------------------------------- |
584 uint Block::sched_call( Matcher &matcher, Block_Array &bbs, uint node_cnt, Node_List &worklist, GrowableArray<int> &ready_cnt, MachCallNode *mcall, VectorSet &next_call ) { |
585 uint Block::sched_call( Matcher &matcher, PhaseCFG* cfg, uint node_cnt, Node_List &worklist, GrowableArray<int> &ready_cnt, MachCallNode *mcall, VectorSet &next_call ) { |
585 RegMask regs; |
586 RegMask regs; |
586 |
587 |
587 // Schedule all the users of the call right now. All the users are |
588 // Schedule all the users of the call right now. All the users are |
588 // projection Nodes, so they must be scheduled next to the call. |
589 // projection Nodes, so they must be scheduled next to the call. |
589 // Collect all the defined registers. |
590 // Collect all the defined registers. |
598 // Collect defined registers |
599 // Collect defined registers |
599 regs.OR(n->out_RegMask()); |
600 regs.OR(n->out_RegMask()); |
600 // Check for scheduling the next control-definer |
601 // Check for scheduling the next control-definer |
601 if( n->bottom_type() == Type::CONTROL ) |
602 if( n->bottom_type() == Type::CONTROL ) |
602 // Warm up next pile of heuristic bits |
603 // Warm up next pile of heuristic bits |
603 needed_for_next_call(n, next_call, bbs); |
604 needed_for_next_call(n, next_call, cfg); |
604 |
605 |
605 // Children of projections are now all ready |
606 // Children of projections are now all ready |
606 for (DUIterator_Fast jmax, j = n->fast_outs(jmax); j < jmax; j++) { |
607 for (DUIterator_Fast jmax, j = n->fast_outs(jmax); j < jmax; j++) { |
607 Node* m = n->fast_out(j); // Get user |
608 Node* m = n->fast_out(j); // Get user |
608 if( bbs[m->_idx] != this ) continue; |
609 if(cfg->get_block_for_node(m) != this) { |
|
610 continue; |
|
611 } |
609 if( m->is_Phi() ) continue; |
612 if( m->is_Phi() ) continue; |
610 int m_cnt = ready_cnt.at(m->_idx)-1; |
613 int m_cnt = ready_cnt.at(m->_idx)-1; |
611 ready_cnt.at_put(m->_idx, m_cnt); |
614 ready_cnt.at_put(m->_idx, m_cnt); |
612 if( m_cnt == 0 ) |
615 if( m_cnt == 0 ) |
613 worklist.push(m); |
616 worklist.push(m); |
621 |
624 |
622 // Set all registers killed and not already defined by the call. |
625 // Set all registers killed and not already defined by the call. |
623 uint r_cnt = mcall->tf()->range()->cnt(); |
626 uint r_cnt = mcall->tf()->range()->cnt(); |
624 int op = mcall->ideal_Opcode(); |
627 int op = mcall->ideal_Opcode(); |
625 MachProjNode *proj = new (matcher.C) MachProjNode( mcall, r_cnt+1, RegMask::Empty, MachProjNode::fat_proj ); |
628 MachProjNode *proj = new (matcher.C) MachProjNode( mcall, r_cnt+1, RegMask::Empty, MachProjNode::fat_proj ); |
626 bbs.map(proj->_idx,this); |
629 cfg->map_node_to_block(proj, this); |
627 _nodes.insert(node_cnt++, proj); |
630 _nodes.insert(node_cnt++, proj); |
628 |
631 |
629 // Select the right register save policy. |
632 // Select the right register save policy. |
630 const char * save_policy; |
633 const char * save_policy; |
631 switch (op) { |
634 switch (op) { |
754 uint i3; |
757 uint i3; |
755 for(i3 = 0; i3<phi_cnt; i3++ ) { // For all pre-scheduled |
758 for(i3 = 0; i3<phi_cnt; i3++ ) { // For all pre-scheduled |
756 Node *n = _nodes[i3]; // Get pre-scheduled |
759 Node *n = _nodes[i3]; // Get pre-scheduled |
757 for (DUIterator_Fast jmax, j = n->fast_outs(jmax); j < jmax; j++) { |
760 for (DUIterator_Fast jmax, j = n->fast_outs(jmax); j < jmax; j++) { |
758 Node* m = n->fast_out(j); |
761 Node* m = n->fast_out(j); |
759 if( cfg->_bbs[m->_idx] ==this ) { // Local-block user |
762 if (cfg->get_block_for_node(m) == this) { // Local-block user |
760 int m_cnt = ready_cnt.at(m->_idx)-1; |
763 int m_cnt = ready_cnt.at(m->_idx)-1; |
761 ready_cnt.at_put(m->_idx, m_cnt); // Fix ready count |
764 ready_cnt.at_put(m->_idx, m_cnt); // Fix ready count |
762 } |
765 } |
763 } |
766 } |
764 } |
767 } |
787 Node* d = delay.pop(); |
790 Node* d = delay.pop(); |
788 worklist.push(d); |
791 worklist.push(d); |
789 } |
792 } |
790 |
793 |
791 // Warm up the 'next_call' heuristic bits |
794 // Warm up the 'next_call' heuristic bits |
792 needed_for_next_call(_nodes[0], next_call, cfg->_bbs); |
795 needed_for_next_call(_nodes[0], next_call, cfg); |
793 |
796 |
794 #ifndef PRODUCT |
797 #ifndef PRODUCT |
795 if (cfg->trace_opto_pipelining()) { |
798 if (cfg->trace_opto_pipelining()) { |
796 for (uint j=0; j<_nodes.size(); j++) { |
799 for (uint j=0; j<_nodes.size(); j++) { |
797 Node *n = _nodes[j]; |
800 Node *n = _nodes[j]; |
798 int idx = n->_idx; |
801 int idx = n->_idx; |
799 tty->print("# ready cnt:%3d ", ready_cnt.at(idx)); |
802 tty->print("# ready cnt:%3d ", ready_cnt.at(idx)); |
800 tty->print("latency:%3d ", cfg->_node_latency->at_grow(idx)); |
803 tty->print("latency:%3d ", cfg->get_latency_for_node(n)); |
801 tty->print("%4d: %s\n", idx, n->Name()); |
804 tty->print("%4d: %s\n", idx, n->Name()); |
802 } |
805 } |
803 } |
806 } |
804 #endif |
807 #endif |
805 |
808 |
823 _nodes.map(phi_cnt++,n); // Schedule him next |
826 _nodes.map(phi_cnt++,n); // Schedule him next |
824 |
827 |
825 #ifndef PRODUCT |
828 #ifndef PRODUCT |
826 if (cfg->trace_opto_pipelining()) { |
829 if (cfg->trace_opto_pipelining()) { |
827 tty->print("# select %d: %s", n->_idx, n->Name()); |
830 tty->print("# select %d: %s", n->_idx, n->Name()); |
828 tty->print(", latency:%d", cfg->_node_latency->at_grow(n->_idx)); |
831 tty->print(", latency:%d", cfg->get_latency_for_node(n)); |
829 n->dump(); |
832 n->dump(); |
830 if (Verbose) { |
833 if (Verbose) { |
831 tty->print("# ready list:"); |
834 tty->print("# ready list:"); |
832 for( uint i=0; i<worklist.size(); i++ ) { // Inspect entire worklist |
835 for( uint i=0; i<worklist.size(); i++ ) { // Inspect entire worklist |
833 Node *n = worklist[i]; // Get Node on worklist |
836 Node *n = worklist[i]; // Get Node on worklist |
838 } |
841 } |
839 |
842 |
840 #endif |
843 #endif |
841 if( n->is_MachCall() ) { |
844 if( n->is_MachCall() ) { |
842 MachCallNode *mcall = n->as_MachCall(); |
845 MachCallNode *mcall = n->as_MachCall(); |
843 phi_cnt = sched_call(matcher, cfg->_bbs, phi_cnt, worklist, ready_cnt, mcall, next_call); |
846 phi_cnt = sched_call(matcher, cfg, phi_cnt, worklist, ready_cnt, mcall, next_call); |
844 continue; |
847 continue; |
845 } |
848 } |
846 |
849 |
847 if (n->is_Mach() && n->as_Mach()->has_call()) { |
850 if (n->is_Mach() && n->as_Mach()->has_call()) { |
848 RegMask regs; |
851 RegMask regs; |
849 regs.Insert(matcher.c_frame_pointer()); |
852 regs.Insert(matcher.c_frame_pointer()); |
850 regs.OR(n->out_RegMask()); |
853 regs.OR(n->out_RegMask()); |
851 |
854 |
852 MachProjNode *proj = new (matcher.C) MachProjNode( n, 1, RegMask::Empty, MachProjNode::fat_proj ); |
855 MachProjNode *proj = new (matcher.C) MachProjNode( n, 1, RegMask::Empty, MachProjNode::fat_proj ); |
853 cfg->_bbs.map(proj->_idx,this); |
856 cfg->map_node_to_block(proj, this); |
854 _nodes.insert(phi_cnt++, proj); |
857 _nodes.insert(phi_cnt++, proj); |
855 |
858 |
856 add_call_kills(proj, regs, matcher._c_reg_save_policy, false); |
859 add_call_kills(proj, regs, matcher._c_reg_save_policy, false); |
857 } |
860 } |
858 |
861 |
859 // Children are now all ready |
862 // Children are now all ready |
860 for (DUIterator_Fast i5max, i5 = n->fast_outs(i5max); i5 < i5max; i5++) { |
863 for (DUIterator_Fast i5max, i5 = n->fast_outs(i5max); i5 < i5max; i5++) { |
861 Node* m = n->fast_out(i5); // Get user |
864 Node* m = n->fast_out(i5); // Get user |
862 if( cfg->_bbs[m->_idx] != this ) continue; |
865 if (cfg->get_block_for_node(m) != this) { |
|
866 continue; |
|
867 } |
863 if( m->is_Phi() ) continue; |
868 if( m->is_Phi() ) continue; |
864 if (m->_idx >= max_idx) { // new node, skip it |
869 if (m->_idx >= max_idx) { // new node, skip it |
865 assert(m->is_MachProj() && n->is_Mach() && n->as_Mach()->has_call(), "unexpected node types"); |
870 assert(m->is_MachProj() && n->is_Mach() && n->as_Mach()->has_call(), "unexpected node types"); |
866 continue; |
871 continue; |
867 } |
872 } |
915 } |
920 } |
916 } |
921 } |
917 } |
922 } |
918 |
923 |
919 //------------------------------catch_cleanup_find_cloned_def------------------ |
924 //------------------------------catch_cleanup_find_cloned_def------------------ |
920 static Node *catch_cleanup_find_cloned_def(Block *use_blk, Node *def, Block *def_blk, Block_Array &bbs, int n_clone_idx) { |
925 static Node *catch_cleanup_find_cloned_def(Block *use_blk, Node *def, Block *def_blk, PhaseCFG* cfg, int n_clone_idx) { |
921 assert( use_blk != def_blk, "Inter-block cleanup only"); |
926 assert( use_blk != def_blk, "Inter-block cleanup only"); |
922 |
927 |
923 // The use is some block below the Catch. Find and return the clone of the def |
928 // The use is some block below the Catch. Find and return the clone of the def |
924 // that dominates the use. If there is no clone in a dominating block, then |
929 // that dominates the use. If there is no clone in a dominating block, then |
925 // create a phi for the def in a dominating block. |
930 // create a phi for the def in a dominating block. |
941 if( j == def_blk->_num_succs ) { |
946 if( j == def_blk->_num_succs ) { |
942 // Block at same level in dom-tree is not a successor. It needs a |
947 // Block at same level in dom-tree is not a successor. It needs a |
943 // PhiNode, the PhiNode uses from the def and IT's uses need fixup. |
948 // PhiNode, the PhiNode uses from the def and IT's uses need fixup. |
944 Node_Array inputs = new Node_List(Thread::current()->resource_area()); |
949 Node_Array inputs = new Node_List(Thread::current()->resource_area()); |
945 for(uint k = 1; k < use_blk->num_preds(); k++) { |
950 for(uint k = 1; k < use_blk->num_preds(); k++) { |
946 inputs.map(k, catch_cleanup_find_cloned_def(bbs[use_blk->pred(k)->_idx], def, def_blk, bbs, n_clone_idx)); |
951 Block* block = cfg->get_block_for_node(use_blk->pred(k)); |
|
952 inputs.map(k, catch_cleanup_find_cloned_def(block, def, def_blk, cfg, n_clone_idx)); |
947 } |
953 } |
948 |
954 |
949 // Check to see if the use_blk already has an identical phi inserted. |
955 // Check to see if the use_blk already has an identical phi inserted. |
950 // If it exists, it will be at the first position since all uses of a |
956 // If it exists, it will be at the first position since all uses of a |
951 // def are processed together. |
957 // def are processed together. |
963 |
969 |
964 // If an existing PhiNode was not found, make a new one. |
970 // If an existing PhiNode was not found, make a new one. |
965 if (fixup == NULL) { |
971 if (fixup == NULL) { |
966 Node *new_phi = PhiNode::make(use_blk->head(), def); |
972 Node *new_phi = PhiNode::make(use_blk->head(), def); |
967 use_blk->_nodes.insert(1, new_phi); |
973 use_blk->_nodes.insert(1, new_phi); |
968 bbs.map(new_phi->_idx, use_blk); |
974 cfg->map_node_to_block(new_phi, use_blk); |
969 for (uint k = 1; k < use_blk->num_preds(); k++) { |
975 for (uint k = 1; k < use_blk->num_preds(); k++) { |
970 new_phi->set_req(k, inputs[k]); |
976 new_phi->set_req(k, inputs[k]); |
971 } |
977 } |
972 fixup = new_phi; |
978 fixup = new_phi; |
973 } |
979 } |
1003 } |
1009 } |
1004 |
1010 |
1005 //------------------------------catch_cleanup_inter_block--------------------- |
1011 //------------------------------catch_cleanup_inter_block--------------------- |
1006 // Fix all input edges in use that reference "def". The use is in a different |
1012 // Fix all input edges in use that reference "def". The use is in a different |
1007 // block than the def. |
1013 // block than the def. |
1008 static void catch_cleanup_inter_block(Node *use, Block *use_blk, Node *def, Block *def_blk, Block_Array &bbs, int n_clone_idx) { |
1014 static void catch_cleanup_inter_block(Node *use, Block *use_blk, Node *def, Block *def_blk, PhaseCFG* cfg, int n_clone_idx) { |
1009 if( !use_blk ) return; // Can happen if the use is a precedence edge |
1015 if( !use_blk ) return; // Can happen if the use is a precedence edge |
1010 |
1016 |
1011 Node *new_def = catch_cleanup_find_cloned_def(use_blk, def, def_blk, bbs, n_clone_idx); |
1017 Node *new_def = catch_cleanup_find_cloned_def(use_blk, def, def_blk, cfg, n_clone_idx); |
1012 catch_cleanup_fix_all_inputs(use, def, new_def); |
1018 catch_cleanup_fix_all_inputs(use, def, new_def); |
1013 } |
1019 } |
1014 |
1020 |
1015 //------------------------------call_catch_cleanup----------------------------- |
1021 //------------------------------call_catch_cleanup----------------------------- |
1016 // If we inserted any instructions between a Call and his CatchNode, |
1022 // If we inserted any instructions between a Call and his CatchNode, |
1017 // clone the instructions on all paths below the Catch. |
1023 // clone the instructions on all paths below the Catch. |
1018 void Block::call_catch_cleanup(Block_Array &bbs, Compile* C) { |
1024 void Block::call_catch_cleanup(PhaseCFG* cfg, Compile* C) { |
1019 |
1025 |
1020 // End of region to clone |
1026 // End of region to clone |
1021 uint end = end_idx(); |
1027 uint end = end_idx(); |
1022 if( !_nodes[end]->is_Catch() ) return; |
1028 if( !_nodes[end]->is_Catch() ) return; |
1023 // Start of region to clone |
1029 // Start of region to clone |
1055 out->push(n->fast_out(j1)); |
1061 out->push(n->fast_out(j1)); |
1056 } |
1062 } |
1057 uint max = out->size(); |
1063 uint max = out->size(); |
1058 for (uint j = 0; j < max; j++) {// For all users |
1064 for (uint j = 0; j < max; j++) {// For all users |
1059 Node *use = out->pop(); |
1065 Node *use = out->pop(); |
1060 Block *buse = bbs[use->_idx]; |
1066 Block *buse = cfg->get_block_for_node(use); |
1061 if( use->is_Phi() ) { |
1067 if( use->is_Phi() ) { |
1062 for( uint k = 1; k < use->req(); k++ ) |
1068 for( uint k = 1; k < use->req(); k++ ) |
1063 if( use->in(k) == n ) { |
1069 if( use->in(k) == n ) { |
1064 Node *fixup = catch_cleanup_find_cloned_def(bbs[buse->pred(k)->_idx], n, this, bbs, n_clone_idx); |
1070 Block* block = cfg->get_block_for_node(buse->pred(k)); |
|
1071 Node *fixup = catch_cleanup_find_cloned_def(block, n, this, cfg, n_clone_idx); |
1065 use->set_req(k, fixup); |
1072 use->set_req(k, fixup); |
1066 } |
1073 } |
1067 } else { |
1074 } else { |
1068 if (this == buse) { |
1075 if (this == buse) { |
1069 catch_cleanup_intra_block(use, n, this, beg, n_clone_idx); |
1076 catch_cleanup_intra_block(use, n, this, beg, n_clone_idx); |
1070 } else { |
1077 } else { |
1071 catch_cleanup_inter_block(use, buse, n, this, bbs, n_clone_idx); |
1078 catch_cleanup_inter_block(use, buse, n, this, cfg, n_clone_idx); |
1072 } |
1079 } |
1073 } |
1080 } |
1074 } // End for all users |
1081 } // End for all users |
1075 |
1082 |
1076 } // End of for all Nodes in cloned area |
1083 } // End of for all Nodes in cloned area |