Re: Patch: fix SPEC regressions

[Eric Botcazou <ebotcazou@libertysurf.fr>]

On Monday 28 October 2002 20:40, Dale Johannesen wrote:
> This patch fixes the problems I ran into building SPEC on darwin last
> week.

The installed version of that patch (btw, Dale, you should put the complete 
ChangeLog entry

	* doloop.c (doloop_modify_runtime):  Fix loop count computation
	for unrolled loops.
	* loop.c (loop_invariant_p):  Support calling from unroller.

as the commit message, this makes it much easier to grab the modifications) 
fixes PR optimization/8599, which is a regression from gcc 3.2 present on 
gcc-3_2-branch.

Bootstrapped/regtested on i586-redhat-linux-gnu (c,c++,objc,f77). OK to 
install once the branch is reopen ? OK to install the testcase on mainline as 
well ?

-- 
Eric Botcazou


2002-11-21  Eric Botcazou  <ebotcazou@libertysurf.fr>

	PR optimization/8599

	2002-11-04  Dale Johannesen  <dalej@apple.com>

	* doloop.c (doloop_modify_runtime):  Fix loop count computation
	for unrolled loops.
	* loop.c (loop_invariant_p):  Support calling from unroller.


2002-11-21  Eric Botcazou  <ebotcazou@libertysurf.fr>

	* gcc.dg/i386-unroll-1.c: New test.


/* PR optimization/8599 */
/* { dg-do run { target i?86-*-* } } */
/* { dg-options "-mcpu=k6 -O2 -funroll-loops" } */

void *array[4];

int main ()
{
 int i;

 for (i = 0; i < 4; i++)
   array[i] = 0;
 
 exit (0);
}