This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: addressof and alpha_expand_block_move
- To: law at cygnus dot com
- Subject: Re: addressof and alpha_expand_block_move
- From: Richard Henderson <rth at cygnus dot com>
- Date: Sat, 24 Jan 1998 15:17:19 -0800
- Cc: Mike Stump <mrs at wrs dot com>, egcs at cygnus dot com, rth at cygnus dot com
- References: <199801221959.LAA25910@kankakee.wrs.com> <29218.885619723@hurl.cygnus.com>
- Reply-To: Richard Henderson <rth at cygnus dot com>
On Fri, Jan 23, 1998 at 10:28:43PM -0700, Jeffrey A Law wrote:
> ie, what happens if the expansion of ADDRESSOF creates an invalid address --
> like a base + const_int, where the const_int won't fit into the range allowed
> on the target (which can be quite small in some cases).
This is already handled by the addressof code. Without my patch
it almost alway only tries to replace with a reg, and even then
it goes through the whole validate_change thing.
What wasn't happening in the alpha case is the reg set and memory
reference being combined back into a single instruction.
r~