------- Comment #2 from trog24 at comcast dot net 2007-04-12 10:18 -------
Subject: Re: PPC object code generation
Hello,
Thank you for your response.
I have monitored the temperature and it is not overheating.
The manner in which the register is corrupted from a non freeze
crash (I have established that freezes are because of an infinite
loop in an ISR although I am not sure which one; but, suspect the
sound) does not trace from the instruction stream where the crash is
noted which is why I suspect an ISR (memory is good). I have not
figured out a way to trap ISR's under Darwin with a single computer
like I could under OS9 where the problem does not occur (never had a
freeze or DSI running a faster CPU under OS9).
I do not think that the module entry protocol change will fix this
problem; but, such a change will eliminate an area with a potential
to corrupt a register via an errant ISR. I also just wanted to make
note that there are safer ways to transition between modules in an
interrupt driven environment and bug reporting was the only way to
make such a note. I would make the change myself for my system only;
but, do not know gcc well enough to determine where such a change
would occur.
In any case, I do again thank you for your response.
Frank
On Apr 11, 2007, at 8:55 PM, pinskia at gcc dot gnu dot org wrote:
>
>
> ------- Comment #1 from pinskia at gcc dot gnu dot org 2007-04-12
> 04:54 -------
> There is a red zone (stack cushion) for ppc-darwin so that the
> interrupt's
> stack should not be using the space between r1 and r1-220. So
> either someone
> wrote an interrupt incorrectly based on the ABI or you are
> incorrect at what
> you thinking is the real problem. The thing to check next is if
> your machine
> is overheating this can cause these random crashes.
>
> Also I have been using GCC on darwin for about 5 years without any
> issues like
> this except when I know either the memory is bad or the machine was
> overheating.
>
>
> --
>
> pinskia at gcc dot gnu dot org changed:
>
> What |Removed |Added
> ----------------------------------------------------------------------
> ------
> Status|UNCONFIRMED |RESOLVED
> Component|preprocessor |target
> Resolution| |WORKSFORME
>
>
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31544
>
> ------- You are receiving this mail because: -------
> You reported the bug, or are watching the reporter.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31544
