Re: A tiny bug in linear_ptab_walker.cc

[Yan Zheng]
> Hi all.
> In file kernel/src/generic/linear_ptab_walker.cc line 386

> --code--
>         else if (! tpg->is_subtree (t_space, t_size+1))
>         {
>         /*
>          * There alredy exists a larger mapping.  Just
>          * continue.  BEWARE: This may cause extension of
>          * access rights to be refused even though they are
>          * perfectly legal.  I.e. if all the mappings in the
>          * subtree of the sender's address space are valid.
>          *
>          * NOTE: There have been discussions about changing
>          * the specifications so that the larger mapping is
>          * removed.
>          */
>         printf ("map_fpage(): Larger mapping already exists.\n");
>         enter_kdebug ("warning: larger mapping exists");
>         goto Next_receiver_entry;
>         }
> --code--


> I think the variable t_size should increase by one before goto
> Next_receiver_entry.

Hmmm... yes.  You're right.  Thanks.  As you can see from the comment
though, the semantics of such over-mappings are probably going to
change.

	eSk