tree-pass.h (pass_split_functions): Declare.

* tree-pass.h (pass_split_functions): Declare.
	* opts.c (decode_options): Enable function splitting at -O2
	* timevar.def (TV_IPA_FNSPLIT): New macro.
	* ipa-split.c: New file.
	* common.opt (-fpartial-inlining): New flag.
	* Makefile.in (ipa-split.o): New object file.
	* passes.c (init_optimization_passes): Add ipa-split.
	* params.def (max-inline-insns-auto): Reduce max-inline-insns-auto to 40.
	(partial-inlining-entry-probability): New parameters.
	* doc/invoke.texi (-fpartial-inlining): New.
	
	* testsuite/gcc.dg/tree-ssa/ipa-split.c

From-SVN: r161382

gcc/ChangeLog

@@ -1,3 +1,15 @@
+2010-06-25  Jan Hubicka  <jh@suse.cz>
+
+	* tree-pass.h (pass_split_functions): Declare.
+	* opts.c (decode_options): Enable function splitting at -O2
+	* timevar.def (TV_IPA_FNSPLIT): New macro.
+	* ipa-split.c: New file.
+	* common.opt (-fpartial-inlining): New flag.
+	* Makefile.in (ipa-split.o): New object file.
+	* passes.c (init_optimization_passes): Add ipa-split.
+	* params.def (partial-inlining-entry-probability): New parameters.
+	* doc/invoke.texi (-fpartial-inlining): New.
+
 2010-06-25  Manuel López-Ibáñez  <manu@gcc.gnu.org>
 
 	PR 44665

gcc/Makefile.in

@@ -1441,6 +1441,7 @@ OBJS-archive = \
 	cppdefault.o \
 	incpath.o \
 	ipa-cp.o \
+        ipa-split.o \
 	ipa-inline.o \
 	ipa-prop.o \
 	ipa-pure-const.o \
@@ -3003,6 +3004,10 @@ ipa-cp.o : ipa-cp.c $(CONFIG_H) $(SYSTEM_H) coretypes.h  \
    $(TREE_H) $(TARGET_H) $(CGRAPH_H) $(IPA_PROP_H) $(TREE_FLOW_H) \
    $(TREE_PASS_H) $(FLAGS_H) $(TIMEVAR_H) $(DIAGNOSTIC_H) $(TREE_DUMP_H) \
    $(TREE_INLINE_H) $(FIBHEAP_H) $(PARAMS_H) tree-pretty-print.h
+ipa-split.o : ipa-split.c $(CONFIG_H) $(SYSTEM_H) coretypes.h  \
+   $(TREE_H) $(TARGET_H) $(CGRAPH_H) $(IPA_PROP_H) $(TREE_FLOW_H) \
+   $(TREE_PASS_H) $(FLAGS_H) $(TIMEVAR_H) $(DIAGNOSTIC_H) $(TREE_DUMP_H) \
+   $(TREE_INLINE_H) $(FIBHEAP_H) $(PARAMS_H)
 matrix-reorg.o : matrix-reorg.c $(CONFIG_H) $(SYSTEM_H) coretypes.h  \
    $(TM_H) $(TREE_H) $(RTL_H) $(TREE_INLINE_H) $(TREE_FLOW_H) \
    tree-flow-inline.h langhooks.h $(HASHTAB_H) $(TOPLEV_H) $(FLAGS_H) $(GGC_H) \

gcc/common.opt

@@ -884,6 +884,10 @@ foptimize-sibling-calls
 Common Report Var(flag_optimize_sibling_calls) Optimization
 Optimize sibling and tail recursive calls
 
+fpartial-inlining
+Common Report Var(flag_partial_inlining)
+Perform partial inlining
+
 fpre-ipa-mem-report
 Common Report Var(pre_ipa_mem_report)
 Report on memory allocation before interprocedural optimization

gcc/doc/invoke.texi

@@ -362,7 +362,7 @@ Objective-C and Objective-C++ Dialects}.
 -fno-sched-interblock -fno-sched-spec -fno-signed-zeros @gol
 -fno-toplevel-reorder -fno-trapping-math -fno-zero-initialized-in-bss @gol
 -fomit-frame-pointer -foptimize-register-move -foptimize-sibling-calls @gol
--fpeel-loops -fpredictive-commoning -fprefetch-loop-arrays @gol
+-fpartial-inlining -fpeel-loops -fpredictive-commoning -fprefetch-loop-arrays @gol
 -fprofile-correction -fprofile-dir=@var{path} -fprofile-generate @gol
 -fprofile-generate=@var{path} @gol
 -fprofile-use -fprofile-use=@var{path} -fprofile-values @gol
@@ -5870,6 +5870,7 @@ also turns on the following optimization flags:
 -findirect-inlining @gol
 -fipa-sra @gol
 -foptimize-sibling-calls @gol
+-fpartial-inlining @gol
 -fpeephole2 @gol
 -fregmove @gol
 -freorder-blocks  -freorder-functions @gol
@@ -7027,6 +7028,14 @@ This optimization is enabled by default.
 With this option, the compiler will create multiple copies of some
 local variables when unrolling a loop which can result in superior code.
 
+@item -fpartial-inlining
+@opindex fpartial-inlining
+Inline parts of functions.  This option has any effect only
+when inlining itself is turned on by the @option{-finline-functions}
+or @option{-finline-small-functions} options.
+
+Enabled at level @option{-O2}.
+
 @item -fpredictive-commoning
 @opindex fpredictive-commoning
 Perform predictive commoning optimization, i.e., reusing computations

gcc/opts.c

@@ -825,6 +825,7 @@ decode_options (unsigned int argc, const char **argv,
   opt2 = (optimize >= 2);
   flag_inline_small_functions = opt2;
   flag_indirect_inlining = opt2;
+  flag_partial_inlining = opt2;
   flag_thread_jumps = opt2;
   flag_crossjumping = opt2;
   flag_optimize_sibling_calls = opt2;

gcc/params.def

@@ -117,6 +117,12 @@ DEFPARAM (PARAM_EARLY_INLINER_MAX_ITERATIONS,
 	  "The maximum number of nested indirect inlining performed by early inliner",
 	  10, 0, 0)
 
+/* Limit on probability of entry BB.  */
+DEFPARAM (PARAM_PARTIAL_INLINING_ENTRY_PROBABILITY,
+	  "partial-inlining-entry-probability",
+	  "Maximum probability of the entry BB of split region (in percent relative to entry BB of the function) to make partial inlining happen",
+	  70, 0, 0)
+
 /* Limit the number of expansions created by the variable expansion
    optimization to avoid register pressure.  */
 DEFPARAM (PARAM_MAX_VARIABLE_EXPANSIONS,

gcc/passes.c

@@ -795,6 +795,10 @@ init_optimization_passes (void)
           NEXT_PASS (pass_cleanup_eh);
           NEXT_PASS (pass_profile);
           NEXT_PASS (pass_local_pure_const);
+	  /* Split functions creates parts that are not run through
+	     early optimizations again.  It is thus good idea to do this
+	     late.  */
+          NEXT_PASS (pass_split_functions);
 	}
       NEXT_PASS (pass_release_ssa_names);
       NEXT_PASS (pass_rebuild_cgraph_edges);

gcc/testsuite/ChangeLog

@@ -1,3 +1,7 @@
+2010-06-25  Jan Hubicka  <jh@suse.cz>
+
+	* testsuite/gcc.dg/tree-ssa/ipa-split-1.c
+
 2010-06-25  Martin Jambor  <mjambor@suse.cz>
 
 	* g++.dg/ipa/iinline-2.C: New test.

gcc/testsuite/gcc.dg/tree-ssa/ipa-split-1.c

@@ -0,0 +1,29 @@
+/* { dg-do compile } */
+/* { dg-options "-O3 -fdump-tree-fnsplit" } */
+int test2(a)
+{
+   if (a<100)
+     return 1;
+   do_something_big ();
+   do_something_big ();
+   do_something_big ();
+   do_something_big ();
+   do_something_big ();
+   do_something_big ();
+   do_something_big ();
+   do_something_big ();
+   do_something_big ();
+   do_something_big ();
+   do_something_big ();
+   do_something_big ();
+   do_something_big ();
+   return 0;
+}
+
+test()
+{
+  test2(10);
+  test2(20);
+}
+/* { dg-final { scan-tree-dump-times "Splitting function" 1 "fnsplit"} } */
+/* { dg-final { cleanup-tree-dump "fnsplit" } } */

gcc/timevar.def

@@ -52,6 +52,7 @@ DEFTIMEVAR (TV_CGRAPH                , "callgraph construction")
 DEFTIMEVAR (TV_CGRAPHOPT             , "callgraph optimization")
 DEFTIMEVAR (TV_VARPOOL               , "varpool construction")
 DEFTIMEVAR (TV_IPA_CONSTANT_PROP     , "ipa cp")
+DEFTIMEVAR (TV_IPA_FNSPLIT           , "ipa function splitting")
 DEFTIMEVAR (TV_IPA_LTO_GIMPLE_IO     , "ipa lto gimple I/O")
 DEFTIMEVAR (TV_IPA_LTO_DECL_IO       , "ipa lto decl I/O")
 DEFTIMEVAR (TV_IPA_LTO_DECL_INIT_IO  , "ipa lto decl init I/O")

gcc/tree-pass.h

@@ -442,6 +442,7 @@ extern struct gimple_opt_pass pass_build_cgraph_edges;
 extern struct gimple_opt_pass pass_local_pure_const;
 extern struct gimple_opt_pass pass_tracer;
 extern struct gimple_opt_pass pass_warn_unused_result;
+extern struct gimple_opt_pass pass_split_functions;
 
 /* IPA Passes */
 extern struct simple_ipa_opt_pass pass_ipa_function_and_variable_visibility;