This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Unaligned block moves and MEM_ALIGN re-broken
- From: kenner at vlsi1 dot ultra dot nyu dot edu (Richard Kenner)
- To: dje at watson dot ibm dot com
- Cc: gcc at gcc dot gnu dot org
- Date: Fri, 25 Jan 02 15:00:46 EST
- Subject: Re: Unaligned block moves and MEM_ALIGN re-broken
With your changes to the way memattrs alignment is initialized, GCC
imposes a stricter, mode-based alignment on the MEM than the TYPE
itself defines for that MEM.
You've made that claim before and I still don't agree with it.
compute_record_mode in stor-layout.c is just one place that shows how
a mode is validated.
We do indeed require that an object of an integral mode be at an alignment
defined by GET_MODE_ALIGNMENT.
The only time when this isn't true is when STRICT_ALIGNMENT is not defined.
but in that case, why does the alignment matter?
This is a regression in GCC which you have introduced.
Can you be specific as to what has regressed?