[PATCH] Fix alignment bug

Gabriel Dos Reis gdr@integrable-solutions.net
Sun Feb 29 19:37:00 GMT 2004


Eric Botcazou <ebotcazou@act-europe.fr> writes:

| > It's PR 12761.  It was closed as WONTFIX ;(
| 
| Thanks.  I'm not sure it is related though, since it appears to be 
| connected to arrays, not records.
| 
| FYI, I recently fixed another long-standing alignment problem on SPARC
| 
| 2003-12-01  Eric Botcazou  <ebotcazou@libertysurf.fr>
| 
| 	PR middle-end/7847
| 	* expr.c (expand_expr) [normal_inner_ref]: When 'offset' is non-zero,
| 	do not recheck that 'op0' is a MEM.  Move comment.  When testing for
| 	unaligned objects, take also into account the alignment of 'op0' and
| 	'mode1' if 'op0' is a MEM.
| 
| so you might want to try it too.

Would it be appropriate for gcc-3_3-branch?

-- Gaby



More information about the Gcc-patches mailing list