This is the mail archive of the gcc@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: *_BY_PIECES_P flaw

From: James E Wilson <wilson at specifixinc dot com>
To: Jakub Jelinek <jakub at redhat dot com>
Cc: gcc at gcc dot gnu dot org
Date: Fri, 06 Aug 2004 13:54:32 -0700
Subject: Re: *_BY_PIECES_P flaw
References: <20040805134718.GV8296@devserv.devel.redhat.com>

Jakub Jelinek wrote:

I think best would be to add a max_size argument to this function,
which would be MOVE_MAX_PIECES + 1 or STORE_MAX_PIECES + 1 depending
on in which macro or routine it is used.

I think all of the uses of MOVE_MAX in expr.c are pointless, since we will never consider anything bigger than MOVE_MAX_PIECES. However, since we have STORE_MAX_PIECES, and it may be different than MOVE_MAX_PIECES, I think your suggestion is right. We need to pass in the appropriate *_MAX_PIECES value to move_by_pieces_ninsns.

Such change would need perhaps some corrections to i386, x86_64 and SH
MOVE_RATIO/CLEAR_RATIO/MOVE_BY_PIECES_P settings.

The x86 CLEAR_RATIO should perhaps be the same as the x86 MOVE_RATIO. And of course the MOVE_BY_PIECES_P macro in the sh port has to change if we add another option to move_by_pieces_ninsns. Otherwise, it isn't clear that anything else has to change. -- Jim Wilson, GNU Tools Support, http://www.SpecifixInc.com

Follow-Ups:
- Re: *_BY_PIECES_P flaw
  - From: David Edelsohn

References:
- *_BY_PIECES_P flaw
  - From: Jakub Jelinek

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