[middle-end, patch] Fix PR 36926 - missing tests in indirect inlining

Martin Jambor mjambor@suse.cz
Fri Jul 25 17:04:00 GMT 2008


this is  fix for PR 36926,  a stupid omission in  my implementation of
indirect inlining.  I'm currently bootstrapping and testing it.

Exactly the  same fix  is also  needed for the  tuples version  of the
patch (5/7).  I'll post that in a minute.

I may  not be  able to  report how both  bootstraps and  testings went
until Monday, though.


Subject: Fix PR 36926

2008-07-25  Martin Jambor  <mjambor@suse.cz>

	PR 36926
	* ipa-prop.c (ipa_analyze_call_uses): Call
	ipa_is_ssa_with_stmt_def instead of SSA_NAME_IS_DEFAULT_DEF.

Index: iinln/gcc/ipa-prop.c
--- iinln.orig/gcc/ipa-prop.c
+++ iinln/gcc/ipa-prop.c
@@ -741,7 +741,7 @@ ipa_analyze_call_uses (struct ipa_node_p
      pointer that is a parameter to this function. */
   n1 = PHI_ARG_DEF (def, 0);
   n2 = PHI_ARG_DEF (def, 1);
+  if (!ipa_is_ssa_with_stmt_def (n1) || !ipa_is_ssa_with_stmt_def (n2))
   d1 = SSA_NAME_DEF_STMT (n1);
   d2 = SSA_NAME_DEF_STMT (n2);

More information about the Gcc-patches mailing list