Re: [PATCH] Re: target/7434: [x86_64] ICE in change_address_1, at emit-rtl.c:1934

Jan Hubicka <[email protected]>
Newsgroups gmane.comp.gcc.bugs,gmane.comp.gcc.prs,gmane.comp.gcc.patches
Message-ID <[email protected]>
> On Thu, May 15, 2003 at 12:48:00PM +0200, Jakub Jelinek wrote:
> > It seems the following patch has not made it to gcc-patches ever (was
> > apparently just sent to [email protected]),
> 
> Grr..

Sorry for that.  Will try to double check to where I am sending the
email.  I felt somehow sad that the patch didn't get reviewed, now I see
why...
> 
> > ... though the bug is still present even on the trunk.
> 
> Are you sure?  This looks a lot like PR 10673, which I did fix on
> mainline, but was waiting for 3.3 branch to open to apply there.

I also resolved that patch as equivalent to mine and dropped ours from
SuSE tree and didn't noticed new failures.
> 
> > 	* i386.c (ix86_split_long_move): Fix modes in the precomputed address
> > 	references.
> 
> As for the patch itself, I prefer the use of replace_equiv_address
> in my patch.

I didn't noticed that function until I read your patch.  It definitly is
better choice.  I will check what is going on with the testcase
tomorrow.

Honza
> 
> > 	* gcc.c-torture/compile/20021008-1.c: New test.
> 
> Please do commit the test case.
> 
> 
> r~
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.