Re: The gdb x86 function prologue parser

[Daniel Jacobowitz <drow at false dot org>]

Hmm, Klee's message didn't make it to the list.

On Sun, Jun 12, 2005 at 04:17:38PM +0300, Eli Zaretskii wrote:
> > Cc: Mark Kettenis <mark.kettenis@xs4all.nl>, jason-swarelist@molenda.com,
> >         gdb-patches@sources.redhat.com
> > From: Klee Dienes <klee@mit.edu>
> > Date: Sun, 12 Jun 2005 07:59:01 -0400
> > 
> > (2) gets used by GDB when it wants to know where to set a breakpoint  
> > (the idea being that when you say "break foo", it wants to insert the  
> > breakpoint in foo() after the prologue has been fully executed).
> > 
> > It also gets used where GDB wants to know if it is inside a function  
> > prologue for some reason (for example, so "next" knows if it has  
> > stepped into a new function, or simply jumped interprocedurally).
> 
> Thanks.
> 
> (I'm collecting info for gdbint.texinfo, that's why I asked.)

FYI, the second part is decreasingly true now.  I only see one
reference to in_prologue left in infrun, and I'm pretty sure it's
obsolete.  Most uses are for the first reason.

And the first reason should, IMO, go away.  But it's a lot of work to
support stopping at the very beginning of a function, and still
printing arguments correctly, without help from the compiler.  We don't
have a clear plan on how to do that yet.

-- 
Daniel Jacobowitz
CodeSourcery, LLC