l4x_hybrid_return for ovlscreen driver

[Adam Lackorzynski]

Hi Marc,

On Wed Nov 21, 2007 at 14:38:54 +0100, Marc CHALAND wrote:
> 2007/11/21, Adam Lackorzynski <adam at os.inf.tu-dresden.de>:
> >
> > Are you using 2k or 4k stack in Fiasco? If 2k, try 4k.
> 
> I've tried 4k without success. Should I compile all l4env ?

No, this is just a kernel matter.

> > Also, there's a
> > stack depth debugging feature, i.e. when you look at the thread list
> > there's figure how much stack is used. You could look at those figure
> > too.
> 
> When it crashes, X has got a figure of 1016.

It's not necessarily the X process. What are the highest numbers and to
which task do they belong?

> I join fiasco binary. pf occurs at 0xf0020d8d. Is there something to exploit ?

Please send me (in private mail) the fiasco.image file. The one you
attached is the stripped version which gives me difficulties finding the
correspondig source location.

> I also tried to not use assembler IPC shortcut, but fiasco crashes at boot time.

Uh? That should not happen and it works for me. Could you send me the
config you used for this setup?

> > I'd suspect inlining is using a bit more stack than without
> > inlining.
> 
> I would have think the opposite as parameters, ebp and eip should be
> stacked as with inline code, it shouldn't ?

Inlining will duplicate code and thus use more stack, at least that's
our experience. I guess gcc could be smarter about stack usage but
well...



Adam
-- 
Adam                 adam at os.inf.tu-dresden.de
  Lackorzynski         http://os.inf.tu-dresden.de/~adam/