This is the mail archive of the gdb-patches@sourceware.org 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: [PING][RFA-v2] Fix troubles with watchpoints in DJGPP


Would this be OK then?

Pierre

PS: It could be optimized in the sense that it should only be called
if it has not been called yet.


2009-06-17  Pierre Muller  <muller@ics.u-strasbg.fr>

	* infcmd.c (post_create_inferior): Call breakpoint_re_set after
target
	is pushed for watchpoint promotion to hardware watchpoint.

Index: src/gdb/infcmd.c
===================================================================
RCS file: /cvs/src/src/gdb/infcmd.c,v
retrieving revision 1.245
diff -u -p -r1.245 infcmd.c
--- src/gdb/infcmd.c	7 Jun 2009 16:46:48 -0000	1.245
+++ src/gdb/infcmd.c	8 Jun 2009 07:29:43 -0000
@@ -421,6 +421,13 @@ post_create_inferior (struct target_ops 
 #endif
     }
 
+  /* On systems that load no shared libraries, like DJGPP target,
+     breakpoint_re_set is never called.
+     Call it now so that ordinary watchpoints get a chance to
+     become promoted to hardware watchpoints if the pushed target
+     supports hardware watchpoints.  */
+  breakpoint_re_set ();
+
   observer_notify_inferior_created (target, from_tty);
 }
 

> -----Message d'origine-----
> De?: gdb-patches-owner@sourceware.org [mailto:gdb-patches-
> owner@sourceware.org] De la part de Pedro Alves
> Envoyé?: Wednesday, June 17, 2009 1:00 AM
> À?: Pierre Muller
> Cc?: gdb-patches@sourceware.org; 'Eli Zaretskii'
> Objet?: Re: [PING][RFA-v2] Fix troubles with watchpoints in DJGPP
> 
> On Tuesday 16 June 2009 23:37:38, Pierre Muller wrote:
> >
> > No one reacted to this second version of my patch...
> 
> My reaction was that the patch looked OK,
> but please could you expand the comment some more to
> explain a bit better why we need this:
> 
> +  /* Call breakpoint_re_set to update watchpoints types.  */
> +  breakpoint_re_set ();
> 
> This almost looks like:
> 
> + /* Increment variable by one.  */
> + i++;
> 
> ;-)
> 
> breakpoint_re_set is very likely to be something we
> will be wanting to split further, make smarter and/or
> eliminate, so having its non-obvious uses nicely described
> is a good thing, IMO.
> 
> > I still have other hardware watchpoint related problems
> > that need to be discussed, but this one is really a small patch,
> > no ?
> 
> --
> Pedro Alves


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