[patch] Remove -fconserve-space

[patch] Remove -fconserve-space

[Steven Bosscher]

[-- Attachment #1: Type: text/plain, Size: 166 bytes --]

Hello,

The attached patch removes the -fconserve-space flag, as discussed last week.
Bootstrapped&tested on powerpc64-unknown-linux-gnu. OK for trunk?

Ciao!
Steven

[-- Attachment #2: remove_fconserve_space.diff.txt --]
[-- Type: text/plain, Size: 3917 bytes --]

gcc/
	* doc/invoke.texi (fconserve-space): Remove documentation.

c-family/
	* c.opt (fconserve-space): Turn into a no-op.

cp/
	* decl.c: Do not include output.h.
	(start_decl): Remove code for flag_conserve_space.

testsuite/
	* g++.old-deja/g++.brendan/array1.C: Remove -fconserve-space flag.

Index: doc/invoke.texi
===================================================================
--- doc/invoke.texi	(revision 188182)
+++ doc/invoke.texi	(working copy)
@@ -182,7 +182,7 @@ in the following sections.
 @item C++ Language Options
 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
 @gccoptlist{-fabi-version=@var{n}  -fno-access-control  -fcheck-new @gol
--fconserve-space  -fconstexpr-depth=@var{n}  -ffriend-injection @gol
+-fconstexpr-depth=@var{n}  -ffriend-injection @gol
 -fno-elide-constructors @gol
 -fno-enforce-eh-specs @gol
 -ffor-scope  -fno-for-scope  -fno-gnu-keywords @gol
@@ -1947,18 +1947,6 @@ return value even without this option.  In all oth
 exhaustion is signalled by throwing @code{std::bad_alloc}.  See also
 @samp{new (nothrow)}.
 
-@item -fconserve-space
-@opindex fconserve-space
-Put uninitialized or run-time-initialized global variables into the
-common segment, as C does.  This saves space in the executable at the
-cost of not diagnosing duplicate definitions.  If you compile with this
-flag and your program mysteriously crashes after @code{main()} has
-completed, you may have an object that is being destroyed twice because
-two definitions are merged.
-
-This option is no longer useful on most targets, now that support has
-been added for putting variables into BSS without making them common.
-
 @item -fconstexpr-depth=@var{n}
 @opindex fconstexpr-depth
 Set the maximum nested evaluation depth for C++11 constexpr functions
Index: c-family/c.opt
===================================================================
--- c-family/c.opt	(revision 188182)
+++ c-family/c.opt	(working copy)
@@ -762,7 +762,7 @@ Allow the arguments of the '?' operator to have di
 
 fconserve-space
 C++ ObjC++ Var(flag_conserve_space)
-Reduce the size of object files
+Does nothing.  Preserved for backward compatibility.
 
 fconstant-string-class=
 ObjC ObjC++ Joined MissingArgError(no class name specified with %qs)
Index: cp/decl.c
===================================================================
--- cp/decl.c	(revision 188182)
+++ cp/decl.c	(working copy)
@@ -39,7 +39,6 @@ along with GCC; see the file COPYING3.  If not see
 #include "tree-inline.h"
 #include "decl.h"
 #include "intl.h"
-#include "output.h" /* for have_global_bss_p */
 #include "toplev.h"
 #include "hashtab.h"
 #include "tm_p.h"
@@ -4528,18 +4527,6 @@ start_decl (const cp_declarator *declarator,
   if (decl == error_mark_node)
     return error_mark_node;
 
-  /* Tell the back end to use or not use .common as appropriate.  If we say
-     -fconserve-space, we want this to save .data space, at the expense of
-     wrong semantics.  If we say -fno-conserve-space, we want this to
-     produce errors about redefs; to do this we force variables into the
-     data segment.  */
-  if (flag_conserve_space
-      && TREE_CODE (decl) == VAR_DECL
-      && TREE_PUBLIC (decl)
-      && !DECL_THREAD_LOCAL_P (decl)
-      && !have_global_bss_p ())
-    DECL_COMMON (decl) = 1;
-
   if (TREE_CODE (decl) == VAR_DECL
       && DECL_NAMESPACE_SCOPE_P (decl) && !TREE_PUBLIC (decl) && !was_public
       && !DECL_THIS_STATIC (decl) && !DECL_ARTIFICIAL (decl))
Index: testsuite/g++.old-deja/g++.brendan/array1.C
===================================================================
--- testsuite/g++.old-deja/g++.brendan/array1.C	(revision 188182)
+++ testsuite/g++.old-deja/g++.brendan/array1.C	(working copy)
@@ -1,5 +1,5 @@
 // { dg-do assemble  }
-// { dg-options "-fconserve-space -fcommon" }
+// { dg-options "-fcommon" }
 // GROUPS passed array-bindings
 
 extern "C" int printf (const char *, ...);

Re: [patch] Remove -fconserve-space

[Jason Merrill]

On 06/06/2012 02:06 PM, Steven Bosscher wrote:
> -// { dg-options "-fconserve-space -fcommon" }
> +// { dg-options "-fcommon" }

I think we can just remove the dg-options; -fcommon has no effect for 
C++.  OK with that change.

Jason

Re: [patch] Remove -fconserve-space

[Gerald Pfeifer]

On Wed, 6 Jun 2012, Steven Bosscher wrote:
> The attached patch removes the -fconserve-space flag, as discussed last 
> week. Bootstrapped&tested on powerpc64-unknown-linux-gnu. OK for trunk?

How about the following for the release notes?

(Happy to use a different/better rationale.  This one's borrowed
from the old documentation. ;)

Gerald
Index: changes.html
===================================================================
RCS file: /cvs/gcc/wwwdocs/htdocs/gcc-4.8/changes.html,v
retrieving revision 1.17
diff -u -3 -p -r1.17 changes.html
--- changes.html	20 Aug 2012 12:23:39 -0000	1.17
+++ changes.html	26 Aug 2012 00:11:29 -0000
@@ -38,7 +38,7 @@ explicit use of vector types may be inco
 built with older versions of GCC.  Auto-vectorized code is not affected
 by this change.</p>
 
-<h2>General Optimizer Improvements</h2>
+<h2>General Optimizer Improvements (and Changes)</h2>
 
   <ul>
     <li>A new option <code>-ftree-partial-pre</code> was added to control
@@ -46,6 +46,9 @@ by this change.</p>
       This option is enabled by default at the <code>-O3</code> optimization
       level, and it makes PRE more aggressive.
     </li>
+    <li>The option <code>-fconserve-space</code> has been removed; it
+    was no longer useful on most targets since GCC supports putting
+    variables into BSS without making them common.</li>
   </ul>
 
 

Re: [patch] Remove -fconserve-space

[Gerald Pfeifer]

Hi there,

any comments on my proposed release notes entry below?

Gerald

On Sun, 26 Aug 2012, Gerald Pfeifer wrote:
>> The attached patch removes the -fconserve-space flag, as discussed last 
>> week. Bootstrapped&tested on powerpc64-unknown-linux-gnu. OK for trunk?
> 
> How about the following for the release notes?
> 
> (Happy to use a different/better rationale.  This one's borrowed
> from the old documentation. ;)
> 
> Gerald
> Index: changes.html
> ===================================================================
> RCS file: /cvs/gcc/wwwdocs/htdocs/gcc-4.8/changes.html,v
> retrieving revision 1.17
> diff -u -3 -p -r1.17 changes.html
> --- changes.html	20 Aug 2012 12:23:39 -0000	1.17
> +++ changes.html	26 Aug 2012 00:11:29 -0000
> @@ -38,7 +38,7 @@ explicit use of vector types may be inco
>  built with older versions of GCC.  Auto-vectorized code is not affected
>  by this change.</p>
>  
> -<h2>General Optimizer Improvements</h2>
> +<h2>General Optimizer Improvements (and Changes)</h2>
>  
>    <ul>
>      <li>A new option <code>-ftree-partial-pre</code> was added to control
> @@ -46,6 +46,9 @@ by this change.</p>
>        This option is enabled by default at the <code>-O3</code> optimization
>        level, and it makes PRE more aggressive.
>      </li>
> +    <li>The option <code>-fconserve-space</code> has been removed; it
> +    was no longer useful on most targets since GCC supports putting
> +    variables into BSS without making them common.</li>
>    </ul>

Re: [patch] Remove -fconserve-space

[Jason Merrill]

That looks good to me.

Jason