[commit/6.0?] Sanity check in baseclass_offset

[commit/6.0?] Sanity check in baseclass_offset

[Daniel Jacobowitz]

If we don't have, for whatever reason, debugging information for a base
class, TYPE_VPTR_FIELDNO may be -1.  TYPE_FIELDS may also be NULL.  When we
access 0[-1], we segfault.  This patch conditionalizes the sanity check with
another sanity check.

OK to move this onto the 6.0 branch, given schedule?  Checked into HEAD, no
regressions.

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer

2003-08-22  Daniel Jacobowitz  <drow@mvista.com>

	* gnu-v3-abi.c (gnuv3_baseclass_offset): Check whether
	TYPE_VPTR_FIELDNO is valid.

--- gdb-6.0/gdb/gnu-v3-abi.c.orig	2003-08-22 10:51:25.000000000 -0400
+++ gdb-6.0/gdb/gnu-v3-abi.c	2003-08-22 10:53:13.000000000 -0400
@@ -412,10 +412,15 @@ gnuv3_baseclass_offset (struct type *typ
      v3 C++ ABI Section 2.4.I.2.b.  Fortunately the ABI guarantees that the
      vtable pointer will be located at the beginning of the object, so we can
      bypass the casting.  Verify that the TYPE_VPTR_FIELDNO is in fact at the
-     start of whichever baseclass it resides in, as a sanity measure.  */
+     start of whichever baseclass it resides in, as a sanity measure - iff
+     we have debugging information for that baseclass.  */
 
   vbasetype = TYPE_VPTR_BASETYPE (type);
-  if (TYPE_FIELD_BITPOS (vbasetype, TYPE_VPTR_FIELDNO (vbasetype)) != 0)
+  if (TYPE_VPTR_FIELDNO (vbasetype) < 0)
+    fill_in_vptr_fieldno (vbasetype);
+
+  if (TYPE_VPTR_FIELDNO (vbasetype) >= 0
+      && TYPE_FIELD_BITPOS (vbasetype, TYPE_VPTR_FIELDNO (vbasetype)) != 0)
     error ("Illegal vptr offset in class %s",
 	   TYPE_NAME (vbasetype) ? TYPE_NAME (vbasetype) : "<unknown>");
 

Re: [commit/6.0?] Sanity check in baseclass_offset

[Andrew Cagney]

> If we don't have, for whatever reason, debugging information for a base
> class, TYPE_VPTR_FIELDNO may be -1.  TYPE_FIELDS may also be NULL.  When we
> access 0[-1], we segfault.  This patch conditionalizes the sanity check with
> another sanity check.
> 
> OK to move this onto the 6.0 branch, given schedule?  Checked into HEAD, no
> regressions.

I think so.

Andrew

Re: [commit/6.0?] Sanity check in baseclass_offset

[Daniel Jacobowitz]

On Sun, Aug 24, 2003 at 08:29:06AM -0400, Andrew Cagney wrote:
> >If we don't have, for whatever reason, debugging information for a base
> >class, TYPE_VPTR_FIELDNO may be -1.  TYPE_FIELDS may also be NULL.  When we
> >access 0[-1], we segfault.  This patch conditionalizes the sanity check 
> >with
> >another sanity check.
> >
> >OK to move this onto the 6.0 branch, given schedule?  Checked into HEAD, no
> >regressions.
> 
> I think so.

Checked in on the branch, thanks.

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer