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: Is this a bug in doloop.c?

From: James E Wilson <wilson at specifixinc dot com>
To: John Lu <jlu at lsil dot com>
Cc: gcc at gcc dot gnu dot org
Date: Mon, 30 Aug 2004 18:13:20 -0700
Subject: Re: Is this a bug in doloop.c?
References: <200408252003.i7PK3h419759@dts0.lsil.com>

John Lu wrote:

in order for the loop to iterate N times.  Currently, if the
exit test is:
	(if_then_else (ge (match_operand:QI 0 "register_operand" "")
	                  (const_int 0))
the looping register is set to N-1.

The point of this code is to detect loops that exit when the iterator reaches -1. So checking for const1_rtx instead of const0_rtx makes no sense here.

However, looking at current code in the loop-doloop.c file, I see that there have been a number of changes. decrement_count has been replaced with increment_count, and it is now set to false here. This file has been largely rewritten.

It is likely that your problem is already fixed on mainline. However, I can't be sure of this, as you didn't give us a testcase, or much explanation of what was going wrong. -- Jim Wilson, GNU Tools Support, http://www.SpecifixInc.com

References:
- Is this a bug in doloop.c?
  - From: John Lu

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