Re: [avr-gcc-list] Confusing volatile behaviour with 4.6.2

[Georg-Johann Lay]

Weddington, Eric schrieb:
> > -----Original Message-----
> > From: address@hidden [mailto:address@hidden On Behalf Of
>
> Andy
>
> > Warner
> > Sent: Tuesday, November 29, 2011 9:08 AM
> > To: Weddington, Eric
> > Cc: Wim Lewis; address@hidden
> > Subject: Re: [avr-gcc-list] Confusing volatile behaviour with 4.6.2
> >
> > On Tue, Nov 29, 2011 at 7:23 AM, Weddington, Eric
> > <address@hidden> wrote:
> >
> >
> >         Hi Wim,
> >
> >         Do you have a GCC bug report filled out for this issue?
> >
> >
> >
> > If the consensus is that this is a genuine bug (there's an awful lot
>
> of
>
> > false bug reports involving volatiles), I'm happy to do my part and
>
> file
>
> > the report to get it recorded.
>
>
> > From experience there's also a lot of genuine bugs involving volatiles.
> (There's a paper by a friend who's a CS professor at University of Utah
> that goes into quite a lot of detail on a number of different compilers
> and their bugs regarding volatile.)
>
> I'd rather get it in a bug report, so we can get the details recorded
> and not lost in a mailing list, no matter what the outcome. It's
> generally safer that way. :-)

The question on combine vs. avr.md is here:

http://gcc.gnu.org/ml/gcc/2011-11/msg00488.html

I'd say there is enough evidence that it's a bug; but as you see from my 
question, the component is not clear, i.e. if it's "rtl-optimize" or 
"target".

Johann

> Eric