Re: [PATCH]: Make Linux use the new unified x86 watchpoint support

[Michael Snyder <msnyder at cygnus dot com>]

Mark, this change did in fact fix the problem.
Would you like to check it in?

				Thanks,
				Michael Snyder

Mark Kettenis wrote:
> 
>    Date: Mon, 26 Mar 2001 18:14:00 -0800
>    From: Michael Snyder <msnyder@cygnus.com>
> 
>    Mark Kettenis wrote:
>    >
>    > FYI, I checked this in.  HJ can finally be happy now (although things
>    > probably won't work correctly for multithreaded programs).
> 
>    Mark, this breaks remote i386 targets debugged from linux hosts.
>    STOPPED_BY_WATCHPOINT is unconditionally defined to a function in
>    i386-nat.c, but we may not be debugging a native target.
> 
> STOPPED_BY_WATCHPOINT always has been unconditionally defined.  The
> only difference I can see is that with the old stuff is that I now
> call perror_with_name if the ptrace call fails.
> 
> If I'm right, the attached patch should fix your problems.  I cannot
> test it right now (no access to a Linux/x86 box), but if it works for
> you, feel free to check it in.
> 
> More details in the reply to your other message.
> 
> Mark
> 
> Index: ChangeLog
> from  Mark Kettenis  <kettenis@gnu.org>
> 
>         * i386-linux-nat.c (i386_linux_dr_get): Return 0 if ptrace call
>         fails instead of calling perror_with_name.  This should fix
>         debugging remote i386 targets with a native Linux/x86 GDB.  Add
>         FIXME for this hack.
> 
> Index: i386-linux-nat.c
> ===================================================================
> RCS file: /cvs/src/src/gdb/i386-linux-nat.c,v
> retrieving revision 1.24
> diff -u -r1.24 i386-linux-nat.c
> --- i386-linux-nat.c 2001/03/21 21:22:48 1.24
> +++ i386-linux-nat.c 2001/03/27 08:40:06
> @@ -712,11 +712,20 @@
>       one thread.  */
>    tid = PIDGET (inferior_pid);
> 
> +  /* FIXME: kettenis/2001-03-27: Calling perror_with_name if the
> +     ptrace call fails breaks debugging remote targets.  The correct
> +     way to fix this is to add the hardware breakpoint and watchpoint
> +     stuff to the target vectore.  For now, just return zero if the
> +     ptrace call fails.  */
>    errno = 0;
>    value = ptrace (PT_READ_U, tid,
>                   offsetof (struct user, u_debugreg[regnum]), 0);
>    if (errno != 0)
> +#if 0
>      perror_with_name ("Couldn't read debug register");
> +#else
> +    return 0;
> +#endif
> 
>    return value;
>  }