This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: [PATCH] Fix PR middle-end/26717

From: Paolo Bonzini <paolo dot bonzini at lu dot unisi dot ch>
To: Andrew Pinski <pinskia at physics dot uc dot edu>
Cc: gcc-patches at gcc dot gnu dot org, Uros Bizjak <uros at kss-loka dot si>
Date: Thu, 23 Mar 2006 15:51:10 +0100
Subject: Re: [PATCH] Fix PR middle-end/26717
References: <1143109898.4422790a12d4f@ssl.kss-loka.si> <7AF0FE9A-DE8A-4505-B3EB-6F8ECF808C19@physics.uc.edu>

I just noticed this, vectors cause the same issue.
testcase:
typedef float __attribute__((vector_size(16))) ftype;

ftype f(ftype a, ftype b)
{
  ftype c = a/a;
  return c/b;
}

Yes, so a function like the get_one_constant I wrote for PR26821 could definitely help in general.

Or, we may want to punt and use TREE_CODE (type) == REAL_TYPE instead of FLOAT_TYPE_P or TREE_CODE (type) != COMPLEX_TYPE.

Paolo

Follow-Ups:
- Re: [PATCH] Fix PR middle-end/26717
  - From: Richard Guenther

References:
- [PATCH] Fix PR middle-end/26717
  - From: Uros Bizjak
- Re: [PATCH] Fix PR middle-end/26717
  - From: Andrew Pinski

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]