This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: bug in cse.c
- From: Richard Henderson <rth at redhat dot com>
- To: Leehod Baruch <LEEHOD at il dot ibm dot com>
- Cc: gcc at gcc dot gnu dot org, Mircea Namolaru <NAMOLARU at il dot ibm dot com>, Ayal Zaks <ZAKS at il dot ibm dot com>, MustafA at il dot ibm dot com
- Date: Wed, 29 Sep 2004 09:28:50 -0700
- Subject: Re: bug in cse.c
- References: <OFC73BE0FC.A552B0D5-ONC2256F1D.00411961-C2256F1D.006098A5@il.ibm.com>
On Tue, Sep 28, 2004 at 07:36:53PM +0200, Leehod Baruch wrote:
> (insn 1289 1288 1264 14 (set (reg:SI 839)
> (subreg:SI (reg:DI 558) 0)) -1 (nil)
> (nil))
...
> The high subreg of register 558 is really a const_int 0 ...
If this is a big-endian target, (subreg x 0) *is* the high subreg.
> The reason for this mistake is that the analysis didn't take into
> account that the computer may use the big Endean method.
Well, subreg_lsb certainly does examine endianness, so all I
can say without more information is that your analysis is wrong.
You'll have to give a much better description of the problem,
including a test case.
r~