This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Fix debug/49827


This fixes the examples given in the PR building libgcc
for Sparc and CRIS.  A full bootstrap on sparc64-linux
is underway, but it'll take some time to complete.  In
the meantime it's well past the stage1 failure.

Just a silly oversight in processing sequences, after
having copied the code from pre-delay-slot make_edges.


r~



	PR debug/49827
	* dwarf2cfi.c (create_trace_edges): Handle sequences properly.
 
diff --git a/gcc/dwarf2cfi.c b/gcc/dwarf2cfi.c
index fd5f680..9e64de6 100644
--- a/gcc/dwarf2cfi.c
+++ b/gcc/dwarf2cfi.c
@@ -2507,6 +2507,14 @@ create_trace_edges (rtx insn)
 	for (lab = nonlocal_goto_handler_labels; lab; lab = XEXP (lab, 1))
 	  maybe_record_trace_start (XEXP (lab, 0), insn, true);
     }
+  else if (GET_CODE (PATTERN (insn)) == SEQUENCE)
+    {
+      rtx seq = PATTERN (insn);
+      int i, n = XVECLEN (seq, 0);
+      for (i = 0; i < n; ++i)
+	create_trace_edges (XVECEXP (seq, 0, i));
+      return;
+    }
 
   /* Process EH edges.  */
   if (CALL_P (insn) || cfun->can_throw_non_call_exceptions)


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]