Re: [MiNT] bug in gcc!

[Eero Tamminen <oak@helsinkinet.fi>]

Hi,

On Monday 18 January 2010, Miro Kropacek wrote:
> 2010/1/18 Vincent Rivière <vincent.riviere@freesbee.fr>
> > Miro Kropacek wrote:
> >>            (reg:SI 0 %d0))) 139 {*addsi3_internal} (nil))
> >> jdsample.c:141: internal compiler error: in
> >> reload_cse_simplify_operands, at postreload.c:396
> >> Please submit a full bug report,
> >
> > I don' understand much, but it looks like this:
> > http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41176
>
> Or this: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40836 but I don't
> understand much more either :)

Looking at the bug reports...

What compiler flags were used when this happens?
(by the gcc specs file and source C flags)

Does the issue go away if you toggle frame-pointers or
disable -fschedule-insns[2] ?

If it's e.g. latter, I think it could be disabled from the specs file as
it's rarely used (unlike the framepointer switch), but that may affect
produced code performance (would need to be tested to see whether
it's an issue in practice).


	- Eero