Re: [PATCH v2 06/19] OpenRISC: PTrace

[Jonas Bonn]

On Tue, 2011-07-05 at 17:42 +0200, Arnd Bergmann wrote:
> On Saturday 02 July 2011, Jonas Bonn wrote:
> > Signed-off-by: Jonas Bonn <jonas@xxxxxxxxxxxx>
> > ---
> >  arch/openrisc/include/asm/ptrace.h |  124 ++++++++
> >  arch/openrisc/kernel/ptrace.c      |  574 ++++++++++++++++++++++++++++++++++++
> >  arch/openrisc/kernel/ptrace.h      |   37 +++
> >  3 files changed, 735 insertions(+), 0 deletions(-)
> >  create mode 100644 arch/openrisc/include/asm/ptrace.h
> >  create mode 100644 arch/openrisc/kernel/ptrace.c
> >  create mode 100644 arch/openrisc/kernel/ptrace.h
> 
> This patch still contains a bunch of commented out code that wants to be
> cleaned up. I think you have done that in most other patches that needed
> the same cleanup.
> 
> 	Arnd

I've been looking a bit at the ptrace stuff the last couple of days.
One question that occurred to me was:  is it mandatory to export the
"struct pt_regs" info in ptrace.h to userspace?

I'm not sure I like the layout of our struct pt_regs.  It would be
better to lay out the registers in a way that mirrors the order that
they are saved during exception/syscall entry in order to make better
use of the cache.

Since the registers can be exported via a regset to userspace anyway
(which doesn't necessarily need to look like pt_regs), I don't see that
userspace really needs how we lay out the registers on the stack.  By
not exporting pt_regs, I am free to change the layout... if it's
exported, it becomes ABI.

Comments?

/Jonas

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/