This is the mail archive of the gdb-patches@sourceware.cygnus.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]

Re: `long double' support for ix86 targets


   Date: Fri, 3 Mar 2000 13:08:07 -0700
   From: Kevin Buettner <kevinb@cygnus.com>

   On Mar 3, 11:25am, Jim Kingdon wrote:

   > > Kevin's changes to findvar.c:extract_floating() and store_floating()
   > > together with some further analysis have convinced me that for all but
   > > one ix86 targets 96-bit long doubles of type &floatformat_i387_ext are
   > > the right thing.
   > 
   > That analysis looks sound to me.
   > 
   > Unless you get some feedback from others, I'd go with that patch.

   Mark's patch looks good to me too.  I am curious though to see how OSF/1
   will be handled.  (That was the one case that Mark had identified where
   overrides will be needed.  I think it'd be good to apply a patch for
   the OSF/1 targets at the same time that he applies his other patch.)

Well, since sizeof (long double) == sizeof (double) on that platform,
the following could make sense:

  #define TARGET_LONG_DOUBLE_FORMAT &floatformat_ieee_double_little
  #define TARGET_LONG_DOUBLE_BITS 64

However, I think I'll leave the current status quo and just put

  /* FIXME: bla bla.  */
  #undef TARGET_LONG_DOUBLE_FORMAT
  #undef TARGET_LONG_DOUBLE_BITS

after Including "i386/tm-i386m3.h" in `config/i386/tm-i386mk.h'.  Of
course this will be checked in at the same time as the other patch.

Mark

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