This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH, committed] Fix altivec type
- From: Ziemowit Laski <zlaski at apple dot com>
- To: Bonzini <bonzini at gnu dot org>
- Cc: gcc-patches at gcc dot gnu dot org, rth at redhat dot com, geoffk at apple dot com
- Date: Fri, 19 Mar 2004 12:09:44 -0800
- Subject: Re: [PATCH, committed] Fix altivec type
- References: <20040319131452.GA12785@fencepost>
On 19 Mar, 2004, at 5.14, Bonzini wrote:
Committed as obvious. I found these while checking whether Altivec
is really creating millions of vector types; luckily it seems it does
not, maybe it did in an older version. The macros do not create
any type; instead the altivec attribute always return one of the types
created in rs6000_init_builtins.
Yes, that's right.
I'm CCing Richard because this goes against his recommendation to add
hashing of vector types. It seems very strange to me that he's wrong,
so I'm asking.
I'm CCing Geoff because I have not received any reports from the
automated regression checker, so it took me a few days to find that
there's this typo. Is the checker down?
Yes it is. :-(
BTW, thanks for following up on this. Only yesterday did I discover
that
AltiVec was in fact broken on the FSF mainline. (I believe that the
PowerPC
simulator you were using does _not_ support AltiVec and so you probably
wouldn't have seen the failures.)
--Zem
--------------------------------------------------------------
Ziemowit Laski 1 Infinite Loop, MS 301-2K
Mac OS X Compiler Group Cupertino, CA USA 95014-2083
Apple Computer, Inc. +1.408.974.6229 Fax .5477