From mboxrd@z Thu Jan  1 00:00:00 1970
From: Michael Snyder <msnyder@cygnus.com>
To: Fernando Nasser <fnasser@redhat.com>
Cc: Eli Zaretskii <eliz@is.elta.co.il>, kettenis@science.uva.nl, gdb-patches@sources.redhat.com
Subject: Re: [PATCH]: Make Linux use the new unified x86 watchpoint support
Date: Tue, 27 Mar 2001 11:59:00 -0000
Message-id: <3AC0F12B.180C0EDA@cygnus.com>
References: <Pine.SUN.3.91.1010327112540.21272B-100000@is> <3AC0C5B6.3C05@redhat.com> <200103271746.MAA08374@indy.delorie.com> <3AC0D330.3D0EFAF3@redhat.com>
X-SW-Source: 2001-03/msg00510.html

Fernando Nasser wrote:
> 
> Eli Zaretskii wrote:
> >
> > In other words, this kind of global registers is (or was when we
> > discussed it) incompatible with how regcache works.  That is why I was
> > advised not to introduce the debug registers into the regcache.
> 
> I agree.  The user has no reason to be meddling with those (unless
> he/she is debugging a debugger, but he/she could still check the values
> that are loaded there while they are still in GPS registers).
> 
> I think we should consider these debug registers as some piece of
> debugger hardware similar to one of those JTAG devices, a kind of
> CPU-embeded logic analyzer or something of a sort -- some debugger
> device that GDB uses, not as an inferior resource.

Then you need some target-vector-dependent way to fetch them.