Re: shared memory object classes

New Message Reply Date view Thread view Subject view Author view

Jim Helman (jimh++at++surreal)
Fri, 24 Feb 95 18:13:20 -0800


> 1. vtbls may not work correctly, that is, objects with virtual
> methods become dangerous to use, because their
> table look-ups might point to different address spaces

This is true, but, in the most common Performer pfSharedArena case,
all the C++ code is in place before we fork (pfConfig), so there isn't
any problem.

problems occur when you have processes in which the code for your C++
objects lives at different addresses in different processes, in
particular if you are sharing memory between unrelated execuatables
(e.g. pfDataPools) and potentially in related processes if you load
C++ objects via DSO's after a fork. But even in these cases, I
believe that with some ld work one can usually get the C++ DSO's
located at the same addresses in both executables without conflicts.

rgds,

-jim helman

jimh++at++surreal.asd.sgi.com
415/390-1151


New Message Reply Date view Thread view Subject view Author view

This archive was generated by hypermail 2.0b2 on Mon Aug 10 1998 - 17:51:00 PDT

This message has been cleansed for anti-spam protection. Replace '++at++' in any mail addresses with the '@' symbol.