[rfc] [4/4] SPU overlay support: Bugfix in remove_breakpoint

["Ulrich Weigand" <uweigand@de.ibm.com>]

Hello,

this fixes a bug we discovered during SPU overlay testing.

The problem occurs when remove_breakpoint attempts to remove a
software breakpoint in an overlay section that is not currently
mapped.  The routine simply goes and writes the saved shadow
contents over that location -- but if another overlay section
is now mapped there, this will clobber its code.

The current behaviour was introduced by Michael Snyder's patch:
http://sourceware.org/ml/gdb-patches/2002-04/msg00149.html
to support hardware breakpoints in overlays.

And I guess if this is a hardware breakpoint, it *should* be
removed -- otherwise the overlay manager would have to fiddle
with hardware breakpoint registers when swapping overlays.

However, for *software* breakpoints this looks definitely 
wrong to me.  So the patch below restores the old behaviour
for those: the shadow contents are restored only if the 
section is still mapped.

Any comments?  I plan on committing this after the rest of
the SPU overlay support patches.

Bye,
Ulrich


ChangeLog:

	* breakpoint.c (remove_breakpoint): Do not remove software
	breakpoints in unmapped overlay sections.

diff -urNp gdb-orig/gdb/breakpoint.c gdb-head/gdb/breakpoint.c
--- gdb-orig/gdb/breakpoint.c	2007-05-06 16:07:05.000000000 +0200
+++ gdb-head/gdb/breakpoint.c	2007-05-07 22:44:15.698329784 +0200
@@ -1585,8 +1585,14 @@ remove_breakpoint (struct bp_location *b
 		 don't know what the overlay manager might do.  */
 	      if (b->loc_type == bp_loc_hardware_breakpoint)
 		val = target_remove_hw_breakpoint (&b->target_info);
-	      else
+
+	      /* However, we should remove *software* breakpoints only
+		 if the section is still mapped, or else we overwrite
+		 wrong code with the saved shadow contents.  */
+	      else if (section_is_mapped (b->section))
 		val = target_remove_breakpoint (&b->target_info);
+	      else
+		val = 0;
 	    }
 	  else
 	    {
-- 
  Dr. Ulrich Weigand
  GNU Toolchain for Linux on System z and Cell BE
  Ulrich.Weigand@de.ibm.com