This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: More help to my port.

On Mon, 2003-09-08 at 01:58, Anders Ãdland wrote:
> Only the first insn is checked with GO_IF_LEGITIMATE_ADDRESS:
> (set (mem:SI (reg:SI 76))
>       (reg:SI 83))
> The other two are not.

I pointed you at reload.  Did you look at it?  You should be stepping
through the part of reload I pointed you at to see what is happening. 

The .greg dump file should be particularly instructive here, because
there should be a list of reloads generated for the bad instruction.  If
this list of reloads is missing or incomplete, then that is where the
problem lies.

> Do I have to define LEGITIMIZE_RELOAD_ADDRESS as Alexandre proposed?

It shouldn't necessary.  I believe there are some rare cases where you
might need to define it for correctness reasons, but these cases should
really be considered reload bugs, and I don't believe your case is one
of them.
-- 
Jim Wilson, GNU Tools Support, http://www.SpecifixInc.com
Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]