This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] PR target/18916; Function arg passing mem align fixes.
On Wed, Jan 05, 2005 at 09:54:05PM +1030, Alan Modra wrote:
> I didn't consider taking the address of these wretched misaligned
> params. :-( Still, the patch has some merit in that at least you
> now can use misaligned params in their functions, whereas previously
> you were liable to get a segfault.
Well, no. Not on true STRICT_ALIGNMENT targets. Just setting the
alignment low isn't enough to make "movdf" do the right thing. You
only get away with this because powerpc can handle (semi-)misaligned
doubles.
I'll agree that having Correct alignments on the mems is good though.
That will be used in emit_block_copy when we move the thing into its
proper location.
> Yes, I guess this will still fail. It's not much different from
> creating a misaligned structure pointer some other way, though.
No, it's very different.
If I use __attribute__((packed)) in some screwy way, that's my fault.
Here I'm just trying to pass some parameters to a function. I'm not
using any language extensions whatsoever; the program is strictly
conforming.
r~