public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
* [PATCH] tree-optimization/106938 - cleanup abnormal edges after inlining
@ 2022-09-14 12:50 Richard Biener
  0 siblings, 0 replies; only message in thread
From: Richard Biener @ 2022-09-14 12:50 UTC (permalink / raw)
  To: gcc-patches

After inlining and IPA transforms we run fixup_cfg to fixup CFG
effects in other functions.  But that fails to clean abnormal
edges from non-pure/const calls which might no longer be necessary
when ->calls_setjmp is false.  The following ensures this happens
and refactors things so we call EH/abnormal cleanup only on the
last stmt in a block.

Bootstrapped and tested on x86_64-unknown-linux-gnu, pushed.

	PR tree-optimization/106938
	* tree-cfg.cc (execute_fixup_cfg): Purge dead abnormal
	edges for all last stmts in a block.  Do EH cleanup
	only on the last stmt in a block.

	* gcc.dg/pr106938.c: New testcase.
---
 gcc/testsuite/gcc.dg/pr106938.c | 36 +++++++++++++++++++++++++++++++++
 gcc/tree-cfg.cc                 | 13 ++++++------
 2 files changed, 43 insertions(+), 6 deletions(-)
 create mode 100644 gcc/testsuite/gcc.dg/pr106938.c

diff --git a/gcc/testsuite/gcc.dg/pr106938.c b/gcc/testsuite/gcc.dg/pr106938.c
new file mode 100644
index 00000000000..7365a8c29fb
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/pr106938.c
@@ -0,0 +1,36 @@
+/* { dg-do compile } */
+/* { dg-options "-O1 -fno-ipa-pure-const -fno-tree-ccp -Wuninitialized" } */
+
+int n;
+
+void
+undefined (void);
+
+__attribute__ ((returns_twice)) int
+zero (void)
+{
+  return 0;
+}
+
+void
+bar (int)
+{
+  int i;
+
+  for (i = 0; i < -1; ++i)
+    n = 0;
+}
+
+__attribute__ ((simd)) void
+foo (void)
+{
+  int uninitialized;
+
+  undefined ();
+
+  while (uninitialized < 1) /* { dg-warning "uninitialized" } */
+    {
+      bar (zero ());
+      ++uninitialized;
+    }
+}
diff --git a/gcc/tree-cfg.cc b/gcc/tree-cfg.cc
index 5cb6db337f9..48c5c4de51f 100644
--- a/gcc/tree-cfg.cc
+++ b/gcc/tree-cfg.cc
@@ -9828,16 +9828,12 @@ execute_fixup_cfg (void)
 	      int flags = gimple_call_flags (stmt);
 	      if (flags & (ECF_CONST | ECF_PURE | ECF_LOOPING_CONST_OR_PURE))
 		{
-		  if (gimple_purge_dead_abnormal_call_edges (bb))
-		    todo |= TODO_cleanup_cfg;
-
 		  if (gimple_in_ssa_p (cfun))
 		    {
 		      todo |= TODO_update_ssa | TODO_cleanup_cfg;
 		      update_stmt (stmt);
 		    }
 		}
-
 	      if (flags & ECF_NORETURN
 		  && fixup_noreturn_call (stmt))
 		todo |= TODO_cleanup_cfg;
@@ -9867,10 +9863,15 @@ execute_fixup_cfg (void)
 		}
 	    }
 
-	  if (maybe_clean_eh_stmt (stmt)
+	  gsi_next (&gsi);
+	}
+      if (gimple *last = last_stmt (bb))
+	{
+	  if (maybe_clean_eh_stmt (last)
 	      && gimple_purge_dead_eh_edges (bb))
 	    todo |= TODO_cleanup_cfg;
-	  gsi_next (&gsi);
+	  if (gimple_purge_dead_abnormal_call_edges (bb))
+	    todo |= TODO_cleanup_cfg;
 	}
 
       /* If we have a basic block with no successors that does not
-- 
2.35.3

^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2022-09-14 12:50 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-09-14 12:50 [PATCH] tree-optimization/106938 - cleanup abnormal edges after inlining Richard Biener

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).