Re: RFA: ia64 tdep patch

[Kevin Buettner <kevinb at redhat dot com>]

On Oct 23, 10:40pm, Marcel Moolenaar wrote:

> > >From what you said earlier, it sounds as though you'd already
> > implemented something like this for FreeBSD...
> 
> It still requires a special ptrace(2) call, making the distinction
> between flushed and dirty stacked registers explicit to gdb. It also
> doesn't address how to handle coredumps. Having the dirty registers
> in some note section to avoid modifying the state of the process
> represented by the core file is appealing to me. This too makes it
> explicit that dirty registers are special WRT flushed registers.
> 
> Again: by hiding the details, we obfuscate the true state of a
> process. I can't stop thinking how that can make some class of
> failures undebuggable (e.g. bugs related to NaT collections, near
> stack overflow conditions, context switching bugs or maybe even
> setjmp/longjmp bugs).
> 
> Thoughts?

I believe that the Linux/ia64 ptrace() code handles the problems with
NaT collections.  I'm not sure how it behaves with regard to the other
conditions that you mention.  (Though for things like context
switching bugs, that's really not a user space problem.)

IMO, this stuff is really arcane and there's going to be some hair
somewhere.  When I was doing the Linux/ia64 port for GDB, I was happy
that a lot of the hair was in ptrace() so that GDB could be presented
with a somewhat simplified interface.  I also did the now defunct
IA-64 GDB port for AIX and was pleased to find that the kernel authors
seemed to handle these issues in much the same way as Linux.  That
said, I've been expecting some kind of debug interface for IA-64 to
emerge which wants to use a model which is closer to the metal than
what GDB currently expects.  If that's the way you decide to go, I'm
willing to review the necessary IA-64 related patches to make it
happen.

Kevin