Re: [PATCH] Use current_inferior ()->pid for AIX

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

Aditya Vidyadhar Kamath <ADITYA.VIDYADHAR.KAMATH@ibm.com> wrote:

>AIX is still using inferior_ptid.pid() to get the inferior pid instead
>of the new object current_inferior().With this it is not possible to
>debug any sample program as the it fails with assertion check for
>pid!=0.

I don't quite understand the underlying problem here.  While we may
want to get rid of inferior_ptid at some point in the future, as of
right now inferior_ptid *is* still being used (by many places in the
target stack), and should always be maintained correctly.

Maybe the problem is really that somewhere inferior_ptid isn't being
updated as it should?


As an aside, this part changes semantics of the code:

>@@ -932,11 +932,12 @@
> static void
> pd_deactivate (void)
> {
>+  ptid_t ptdrtn = ptid_t (current_inferior ()->pid);
>   if (!pd_active)
>     return;
>   pthdb_session_destroy (pd_session);
>   
>-  pid_to_prc (&inferior_ptid);
>+  pid_to_prc (&ptdrtn); 
>   pd_active = 0;
> }

The point of this code was to *change* inferior_ptid back from
the per-thread view to the no-threads view as the threading
layer is deactivated.  With your patch, inferior_ptid is no
longer changed at all, which makes the pid_to_prc call useless.

Bye,
Ulrich