This is the mail archive of the gdb-patches@sources.redhat.com mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: RFA: Support libthread_db xregs interface


   Sender: jimb@zenia.home
   Cc: gdb-patches@sources.redhat.com
   From: Jim Blandy <jimb@redhat.com>
   Date: 25 Aug 2004 00:07:07 -0500

Andrew Cagney <cagney@gnu.org> writes:
> > + v:=:const struct regset *:xregs_regset:::0
> > Mark's regset change added both the "regset.h" object and the
> regset_from_core_section architecture method. They, together, replace
> the old *-nat.c:fill_regset et.al. calls.
> > Can we implement the equivalent here for ptrace/thread-db?


   Not sure what you mean.  This change lets a gdbarch object specify a
   regset which the libthread_db support code will then use to read and
   write additional registers beyond those covered by gregset_t and
   fpregset_t.  So this change does provide a regset.h-style regset for
   libthread_db.  That's what you're looking for, right?

I'm not sure, but I presume Andrew is asking you to implement a
regset_from_xxx function for use by ptrace/thread-db, instead of
explicitly adding the xregset to the architecture vector.  Your
current patch leaves the supply_gregset() and supply_fpregset() calls
as they are.  That's fine for now, but in the long run they'll have to
be replaced with regset stuff too.  By using a regset_from_xxx
function we only need a single entry in the architecture vector,
instead of three (or even more).

Right.


For the architecture vector, its clear that we already need some sort of regset_from_xxx method that covers all the regsets. I can imagine it being a function parameterized with an ``enum regset_type'', or an array variable indexed by the same, or even parameterized with a reggroup (but that's getting weird :-). Which ever. With this in place we can deprecate / eliminate supply_regset et.al..

How much of the thread-db should be done now. That's MichaelS/MarkKs call. However it is a relatively straight forward task - just lift the code from corelow.c.

can we do that?
Andrew



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]