diff mbox

, PR ipa/81238, make default target_clone static

Message ID 20170627233413.GA11400@ibm-tiger.the-meissners.org
State New
Headers show

Commit Message

Michael Meissner June 27, 2017, 11:34 p.m. UTC
In going through the target_clone support for the PowerPC, I noticed
that the default target clone function was not explicitly set to private.  This
patch fixes this.  I have checked both the x86_64 and PowerPC target clone
code, and the patch does make it private.

I'm doing bootstrap builds on both little endian power8 and an x86_64 system
right now (both builds are at stage3, so I don't expect any problems).
Assuming there is no regression, can I check this patch in to the trunk?

2017-06-27  Michael Meissner  <meissner@linux.vnet.ibm.com>

	PR ipa/81238
	* multiple_target.c (create_dispatcher_calls): Set the default
	clone to be static, not public.

Comments

Michael Meissner June 27, 2017, 11:56 p.m. UTC | #1
On Tue, Jun 27, 2017 at 07:34:13PM -0400, Michael Meissner wrote:
> In going through the target_clone support for the PowerPC, I noticed
> that the default target clone function was not explicitly set to private.  This
> patch fixes this.  I have checked both the x86_64 and PowerPC target clone
> code, and the patch does make it private.
> 
> I'm doing bootstrap builds on both little endian power8 and an x86_64 system
> right now (both builds are at stage3, so I don't expect any problems).
> Assuming there is no regression, can I check this patch in to the trunk?

Both PowerPC and x86_64 bootstraps have finished with no regressions.
Jeff Law June 28, 2017, 4:32 a.m. UTC | #2
On 06/27/2017 05:34 PM, Michael Meissner wrote:
> In going through the target_clone support for the PowerPC, I noticed
> that the default target clone function was not explicitly set to private.  This
> patch fixes this.  I have checked both the x86_64 and PowerPC target clone
> code, and the patch does make it private.
> 
> I'm doing bootstrap builds on both little endian power8 and an x86_64 system
> right now (both builds are at stage3, so I don't expect any problems).
> Assuming there is no regression, can I check this patch in to the trunk?
> 
> 2017-06-27  Michael Meissner  <meissner@linux.vnet.ibm.com>
> 
> 	PR ipa/81238
> 	* multiple_target.c (create_dispatcher_calls): Set the default
> 	clone to be static, not public.
OK.
jeff
diff mbox

Patch

Index: gcc/multiple_target.c
===================================================================
--- gcc/multiple_target.c	(revision 249710)
+++ gcc/multiple_target.c	(working copy)
@@ -148,6 +148,7 @@  create_dispatcher_calls (struct cgraph_n
 	}
     }
 
+  TREE_PUBLIC (node->decl) = 0;
   symtab->change_decl_assembler_name (node->decl,
 				      clone_function_name (node->decl,
 							   "default"));