[PATCH, PR 41395] Enable IPA-SRA at -O2 again

[PATCH, PR 41395] Enable IPA-SRA at -O2 again

[Martin Jambor]

Hi,

because I believe all concerns raised in PR 41395 and elsewhere, the
patch below enables IPA-SRA by default at -O2 again.  As slush after
the LTO merge is over, I'd like to commit it soon.

I have bootstrapped and tested this patch on x86_64-linux (with Ada,
LTO enabled, both with "yes" and "release" checking), i686-linux (with
Ada, LTO enabled), ia64-linux (LTO not configured) and sparc-linux
(though on a pre-LTO revision and the tests are still running now).
Uros has bootstrapped and tested it on alpha (though I also know only
about pre-LTO testing).  This makes me believe that the severe
problems caused by it the last time will not happen again.

OK for trunk?

Thanks,

Martin


2009-10-06  Martin Jambor  <mjambor@suse.cz>

	PR bootstrap/41395
	* opts.c (decode_options): Run IPA-SRA at -O2.

Index: mine/gcc/opts.c
===================================================================
--- mine.orig/gcc/opts.c
+++ mine/gcc/opts.c
@@ -922,6 +922,7 @@ decode_options (unsigned int argc, const
   flag_tree_pre = opt2;
   flag_tree_switch_conversion = 1;
   flag_ipa_cp = opt2;
+  flag_ipa_sra = opt2;
 
   /* Track fields in field-sensitive alias analysis.  */
   set_param_value ("max-fields-for-field-sensitive",

Re: [PATCH, PR 41395] Enable IPA-SRA at -O2 again

[Richard Guenther]

On Tue, 6 Oct 2009, Martin Jambor wrote:

> Hi,
> 
> because I believe all concerns raised in PR 41395 and elsewhere, the
> patch below enables IPA-SRA by default at -O2 again.  As slush after
> the LTO merge is over, I'd like to commit it soon.
> 
> I have bootstrapped and tested this patch on x86_64-linux (with Ada,
> LTO enabled, both with "yes" and "release" checking), i686-linux (with
> Ada, LTO enabled), ia64-linux (LTO not configured) and sparc-linux
> (though on a pre-LTO revision and the tests are still running now).
> Uros has bootstrapped and tested it on alpha (though I also know only
> about pre-LTO testing).  This makes me believe that the severe
> problems caused by it the last time will not happen again.
> 
> OK for trunk?

Ok.

Thanks,
Richard.

> 2009-10-06  Martin Jambor  <mjambor@suse.cz>
> 
> 	PR bootstrap/41395
> 	* opts.c (decode_options): Run IPA-SRA at -O2.
> 
> Index: mine/gcc/opts.c
> ===================================================================
> --- mine.orig/gcc/opts.c
> +++ mine/gcc/opts.c
> @@ -922,6 +922,7 @@ decode_options (unsigned int argc, const
>    flag_tree_pre = opt2;
>    flag_tree_switch_conversion = 1;
>    flag_ipa_cp = opt2;
> +  flag_ipa_sra = opt2;
>  
>    /* Track fields in field-sensitive alias analysis.  */
>    set_param_value ("max-fields-for-field-sensitive",
> 
>