re PR bootstrap/82832 (Broken PGO LTO bootstrap on x86_64 after r254379)

	PR bootstrap/82832
	* ipa-inline-transform.c (update_noncloned_frequencies): Always
	scale.
	(inline_transform): Likewise.
	* predict.c (counts_to_freqs): Remove useless conditional.
	* profile-count.h (profile_count::apply_scale): Move sanity check.
	* tree-inline.c (copy_bb): Always scale.
	(copy_cfg_body): Likewise.

From-SVN: r254452

diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index b21c5fce2ad9735baf56fb8700fe1fce1a6fb27c..26afa77f726622ace592c9f7a578c7c26840ddf6 100644 (file)
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,3 +1,14 @@
+2017-11-06  Jan Hubicka  <hubicka@ucw.cz>
+
+       PR bootstrap/82832
+       * ipa-inline-transform.c (update_noncloned_frequencies): Always
+       scale.
+       (inline_transform): Likewise.
+       * predict.c (counts_to_freqs): Remove useless conditional.
+       * profile-count.h (profile_count::apply_scale): Move sanity check.
+       * tree-inline.c (copy_bb): Always scale.
+       (copy_cfg_body): Likewise.
+
 2017-11-06  Christophe Lyon  <christophe.lyon@linaro.org>
 
        PR target/67591

diff --git a/gcc/ipa-inline-transform.c b/gcc/ipa-inline-transform.c
index 886e8edd473b40229cd162550313637ac5ceb7cf..e367df7df0e5c507f7ebf2bd8da86462cf82b4b4 100644 (file)
--- a/gcc/ipa-inline-transform.c
+++ b/gcc/ipa-inline-transform.c
@@ -59,7 +59,18 @@ update_noncloned_frequencies (struct cgraph_node *node,
                              profile_count den)
 {
   struct cgraph_edge *e;
-  bool scale = (num == profile_count::zero () || den > 0);
+
+  /* We always must scale to be sure counters end up compatible.
+     If den is zero, just force it nonzero and hope for reasonable
+     approximation.
+     When num is forced nonzero, also update den, so we do not scale profile
+     to 0.   */
+  if (!(num == den)
+      && !(den.force_nonzero () == den))
+    {
+      den = den.force_nonzero ();
+      num = num.force_nonzero ();
+    }
 
   /* We do not want to ignore high loop nest after freq drops to 0.  */
   if (!freq_scale)
@@ -71,19 +82,16 @@ update_noncloned_frequencies (struct cgraph_node *node,
         e->frequency = CGRAPH_FREQ_MAX;
       if (!e->inline_failed)
         update_noncloned_frequencies (e->callee, freq_scale, num, den);
-      if (scale)
-       e->count = e->count.apply_scale (num, den);
+      e->count = e->count.apply_scale (num, den);
     }
   for (e = node->indirect_calls; e; e = e->next_callee)
     {
       e->frequency = e->frequency * (gcov_type) freq_scale / CGRAPH_FREQ_BASE;
       if (e->frequency > CGRAPH_FREQ_MAX)
         e->frequency = CGRAPH_FREQ_MAX;
-      if (scale)
-       e->count = e->count.apply_scale (num, den);
+      e->count = e->count.apply_scale (num, den);
     }
-  if (scale)
-    node->count = node->count.apply_scale (num, den);
+  node->count = node->count.apply_scale (num, den);
 }
 
 /* We removed or are going to remove the last call to NODE.
@@ -692,7 +700,10 @@ inline_transform (struct cgraph_node *node)
 
          basic_block bb;
          FOR_ALL_BB_FN (bb, cfun)
-           bb->count = bb->count.apply_scale (num, den);
+           if (num == profile_count::zero ())
+             bb->count = bb->count.global0 ();
+           else
+             bb->count = bb->count.apply_scale (num, den);
          ENTRY_BLOCK_PTR_FOR_FN (cfun)->count = node->count;
        }
       todo = optimize_inline_calls (current_function_decl);

diff --git a/gcc/predict.c b/gcc/predict.c
index cf42ccbd903b9c5063de30702c514897d28a2e9c..0fd2b72e135841c7ce57381e877ae8c32f85e16e 100644 (file)
--- a/gcc/predict.c
+++ b/gcc/predict.c
@@ -3324,8 +3324,7 @@ counts_to_freqs (void)
   basic_block bb;
 
   FOR_BB_BETWEEN (bb, ENTRY_BLOCK_PTR_FOR_FN (cfun), NULL, next_bb)
-    if (!(bb->count < true_count_max))
-      true_count_max = true_count_max.max (bb->count);
+    true_count_max = true_count_max.max (bb->count);
 
   cfun->cfg->count_max = true_count_max;
 

diff --git a/gcc/profile-count.h b/gcc/profile-count.h
index d793d11c830db216db5f0152570c5058b8060ab0..f16bbcb2b9f77cf7838f95df3305d627302622d4 100644 (file)
--- a/gcc/profile-count.h
+++ b/gcc/profile-count.h
@@ -949,9 +949,9 @@ public:
        return num;
       if (!initialized_p () || !num.initialized_p () || !den.initialized_p ())
        return profile_count::uninitialized ();
-      gcc_checking_assert (den.m_val);
       if (num == den)
        return *this;
+      gcc_checking_assert (den.m_val);
 
       profile_count ret;
       uint64_t val;

diff --git a/gcc/tree-inline.c b/gcc/tree-inline.c
index 040799c778c25866e78d77de505ce657155c95a6..7ef107247b55400cf7b7f46280107dcb187c3aba 100644 (file)
--- a/gcc/tree-inline.c
+++ b/gcc/tree-inline.c
@@ -1770,8 +1770,18 @@ copy_bb (copy_body_data *id, basic_block bb,
   basic_block copy_basic_block;
   tree decl;
   basic_block prev;
-  bool scale = !num.initialized_p ()
-              || (den.nonzero_p () || num == profile_count::zero ());
+
+  /* We always must scale to be sure counters end up compatible.
+     If den is zero, just force it nonzero and hope for reasonable
+     approximation.
+     When num is forced nonzero, also update den, so we do not scale profile
+     to 0.   */
+  if (!(num == den)
+      && !(den.force_nonzero () == den))
+    {
+      den = den.force_nonzero ();
+      num = num.force_nonzero ();
+    }
 
   /* Search for previous copied basic block.  */
   prev = bb->prev_bb;
@@ -1781,10 +1791,7 @@ copy_bb (copy_body_data *id, basic_block bb,
   /* create_basic_block() will append every new block to
      basic_block_info automatically.  */
   copy_basic_block = create_basic_block (NULL, (basic_block) prev->aux);
-  if (scale)
-    copy_basic_block->count = bb->count.apply_scale (num, den);
-  else if (num.initialized_p ())
-    copy_basic_block->count = bb->count;
+  copy_basic_block->count = bb->count.apply_scale (num, den);
 
   copy_gsi = gsi_start_bb (copy_basic_block);
 
@@ -2691,6 +2698,18 @@ copy_cfg_body (copy_body_data * id, profile_count,
   profile_count den = ENTRY_BLOCK_PTR_FOR_FN (src_cfun)->count;
   profile_count num = entry_block_map->count;
 
+  /* We always must scale to be sure counters end up compatible.
+     If den is zero, just force it nonzero and hope for reasonable
+     approximation.
+     When num is forced nonzero, also update den, so we do not scale profile
+     to 0.   */
+  if (!(num == den)
+      && !(den.force_nonzero () == den))
+    {
+      den = den.force_nonzero ();
+      num = num.force_nonzero ();
+    }
+
   cfun_to_copy = id->src_cfun = DECL_STRUCT_FUNCTION (callee_fndecl);
 
   /* Register specific tree functions.  */