LTO/WPA: Ensure that output_offload_tables only writes table once [PR116535]

When ltrans was written concurrently, e.g. via -flto=N (N > 1, assuming
sufficient partiations, e.g., via -flto-partition=max), output_offload_tables
wrote the output tables once per fork.

	PR lto/116535

gcc/ChangeLog:

	* lto-cgraph.cc (output_offload_tables): Remove offload_ frees.
	* lto-streamer-out.cc (lto_output): Make call to it depend on
	lto_get_out_decl_state ()->output_offload_tables_p.
	* lto-streamer.h (struct lto_out_decl_state): Add
	output_offload_tables_p field.
	* tree-pass.h (ipa_write_optimization_summaries): Add bool argument.
	* passes.cc (ipa_write_summaries_1): Add bool
	output_offload_tables_p arg.
	(ipa_write_summaries): Update call.
	(ipa_write_optimization_summaries): Accept output_offload_tables_p.

gcc/lto/ChangeLog:

	* lto.cc (stream_out): Update call to
	ipa_write_optimization_summaries to pass true for first partition.

diff --git a/gcc/lto-cgraph.cc b/gcc/lto-cgraph.cc
index 6395033ab9df2721122ccaf7b5dffb85f571e3ef..1492409427c97b175c840c9a43c6f93a373430bf 100644 (file)
--- a/gcc/lto-cgraph.cc
+++ b/gcc/lto-cgraph.cc
@@ -1139,16 +1139,6 @@ output_offload_tables (void)
 
   streamer_write_uhwi_stream (ob->main_stream, 0);
   lto_destroy_simple_output_block (ob);
-
-  /* In WHOPR mode during the WPA stage the joint offload tables need to be
-     streamed to one partition only.  That's why we free offload_funcs and
-     offload_vars after the first call of output_offload_tables.  */
-  if (flag_wpa)
-    {
-      vec_free (offload_funcs);
-      vec_free (offload_vars);
-      vec_free (offload_ind_funcs);
-    }
 }
 
 /* Verify the partitioning of NODE.  */

diff --git a/gcc/lto-streamer-out.cc b/gcc/lto-streamer-out.cc
index 523d6dad221e92d8c059f193cc8252a5d5314663..a4b171358d41de345e0951eb46a6f93e79f2f49f 100644 (file)
--- a/gcc/lto-streamer-out.cc
+++ b/gcc/lto-streamer-out.cc
@@ -2829,7 +2829,8 @@ lto_output (void)
      statements using the statement UIDs.  */
   output_symtab ();
 
-  output_offload_tables ();
+  if (lto_get_out_decl_state ()->output_offload_tables_p)
+    output_offload_tables ();
 
   if (flag_checking)
     {

diff --git a/gcc/lto-streamer.h b/gcc/lto-streamer.h
index 79c44d2cae718fad98e35ca95eebf6aebc79a3fe..4da1a3efe033738ae1479e34b8023a51418f66c9 100644 (file)
--- a/gcc/lto-streamer.h
+++ b/gcc/lto-streamer.h
@@ -531,6 +531,9 @@ struct lto_out_decl_state
 
   /* True if decl state is compressed.  */
   bool compressed;
+
+  /* True if offload tables should be output. */
+  bool output_offload_tables_p;
 };
 
 typedef struct lto_out_decl_state *lto_out_decl_state_ptr;

diff --git a/gcc/lto/lto.cc b/gcc/lto/lto.cc
index 52dd436fd9a14475e2bdbb4959b5d3cbf36e19cb..1ee215d8f1d371955c2517c3230dfcc25e0a3571 100644 (file)
--- a/gcc/lto/lto.cc
+++ b/gcc/lto/lto.cc
@@ -178,7 +178,7 @@ stream_out (char *temp_filename, lto_symtab_encoder_t encoder, int part)
 
   gcc_assert (!dump_file);
   streamer_dump_file = dump_begin (TDI_lto_stream_out, NULL, part);
-  ipa_write_optimization_summaries (encoder);
+  ipa_write_optimization_summaries (encoder, part == 0);
 
   free (CONST_CAST (char *, file->filename));
 

diff --git a/gcc/passes.cc b/gcc/passes.cc
index d73f8ba97b64b2ea5d14d9fe4b2a971a4d92efa4..057850f4decbc7d4a6f805b9efe49d4feb910d65 100644 (file)
--- a/gcc/passes.cc
+++ b/gcc/passes.cc
@@ -2829,11 +2829,13 @@ ipa_write_summaries_2 (opt_pass *pass, struct lto_out_decl_state *state)
    summaries.  SET is the set of nodes to be written.  */
 
 static void
-ipa_write_summaries_1 (lto_symtab_encoder_t encoder)
+ipa_write_summaries_1 (lto_symtab_encoder_t encoder,
+                      bool output_offload_tables_p)
 {
   pass_manager *passes = g->get_passes ();
   struct lto_out_decl_state *state = lto_new_out_decl_state ();
   state->symtab_node_encoder = encoder;
+  state->output_offload_tables_p = output_offload_tables_p;
 
   lto_output_init_mode_table ();
   lto_push_out_decl_state (state);
@@ -2897,7 +2899,8 @@ ipa_write_summaries (void)
     if (vnode->need_lto_streaming)
       lto_set_symtab_encoder_in_partition (encoder, vnode);
 
-  ipa_write_summaries_1 (compute_ltrans_boundary (encoder));
+  ipa_write_summaries_1 (compute_ltrans_boundary (encoder),
+                        flag_generate_offload);
 
   free (order);
   if (streamer_dump_file)
@@ -2952,10 +2955,12 @@ ipa_write_optimization_summaries_1 (opt_pass *pass,
    NULL, write out all summaries of all nodes. */
 
 void
-ipa_write_optimization_summaries (lto_symtab_encoder_t encoder)
+ipa_write_optimization_summaries (lto_symtab_encoder_t encoder,
+                                 bool output_offload_tables_p)
 {
   struct lto_out_decl_state *state = lto_new_out_decl_state ();
   state->symtab_node_encoder = encoder;
+  state->output_offload_tables_p = output_offload_tables_p;
 
   lto_output_init_mode_table ();
   lto_push_out_decl_state (state);

diff --git a/gcc/tree-pass.h b/gcc/tree-pass.h
index 3a0cf13089e2f15d7263112366e9ad380cf68604..a928cbe4557368ec483919a06cd3d29d733a7b66 100644 (file)
--- a/gcc/tree-pass.h
+++ b/gcc/tree-pass.h
@@ -684,7 +684,8 @@ extern void emergency_dump_function (void);
 extern void print_current_pass (FILE *);
 extern void debug_pass (void);
 extern void ipa_write_summaries (void);
-extern void ipa_write_optimization_summaries (struct lto_symtab_encoder_d *);
+extern void ipa_write_optimization_summaries (struct lto_symtab_encoder_d *,
+                                             bool);
 extern void ipa_read_summaries (void);
 extern void ipa_read_optimization_summaries (void);
 extern void register_one_dump_file (opt_pass *);