Sponsored Content
Top Forums Programming Shared memory and C++ Objects (cont) Post 14548 by wizard on Friday 1st of February 2002 03:36:33 PM
Old 02-01-2002
Ok. Apparently I understood you correctly in the first post. First of all, when you create an object containing a virtual function, you do not create a virtual function table. At compile time, a pointer to the virtual function is created and placed into a module wide virtual function table which is a table of pointers to virtual functions defined within the compiled object. If you can get the value of the pointer pointing to the virtual function table, it will only have meaning in the executable that contains the instantiation of Class1, in this case object Obj1.

Secondly, if you create a shared memory segment and place a pointer to the virtual function, the virtual function table, or a pointer to anything inside the address space of ProcessA, you are NOT "sharing the object". You are passing a virtual memory address with meaning in ProcessA, to ProcessB where it has no meaning. When ProcessB dereferences the pointer, you will more than likely get a SIGSEGV.

Shared memory is used for passing DATA objects, not executable text (i.e. a virtual function) or a pointer.

Obviously, I don't know why ProcessB needs to access the Obj1 object defined in ProcessA. If the Class1 template (with virtual functions and all) is declared inside both source modules (or more likely in a header file and included in the source), then you could have separate instantiations of the template in both processes. If you need to share data values between the instantiated objects, then you could pass the data values using shared memory, but you cannot pass a pointer to the executable text of the virtual function. You don't need to as you would have the same executable text in each process.

Now, that said, you should look at the load() system call in AIX. I've used this to dynamically create modules and load them into the process address space of a running process. Realize that this is a non-portable system call. Since I don't know what your ultimate goal is here, this is the only possible option I can come up with.
 

10 More Discussions You Might Find Interesting

1. Programming

Runtime Linking shared Objects

I'm runtime linking (dlopen and dlsym) to a shared object (library) I've created and after a number of function calls into the library the program core dumps (Illegal operation). This only occurs during runtime linking. If I use the same library and dynamically link during compile time everything... (3 Replies)
Discussion started by: dneely
3 Replies

2. UNIX for Dummies Questions & Answers

Shared Objects

Hi Friends ! I have a library, say libxyz.a. To view all the object files in the archive, i issued the command : ar -t libxyz.a which displayed all the object files it contains. Now, I would like to know the functions in each object file. Is there any such command that displays... (3 Replies)
Discussion started by: mrgubbala
3 Replies

3. Programming

Linking with shared objects

hi all ! Do I need all the shared objects to be present while compiling my code which has reference to a only one shared object, which in turn refers to another shared object. for example I want to compile example.c which refers to sample.so sample.so has refrence to anothersample.so do... (2 Replies)
Discussion started by: disclaimer
2 Replies

4. UNIX for Advanced & Expert Users

Shared Objects

Hi. Does anyone know by how much a text size of an executable(on ibm) would grow if you link one shared object(library)? Is it a constant number or it depends on a .so that is linked? (3 Replies)
Discussion started by: Yura
3 Replies

5. UNIX for Advanced & Expert Users

debugging shared objects

Hi, i am trying to debug a binary which is using a shared lib. but i could not succeed in tracking the code flow in the classes defined in this library. i get: class MyClass <opaque> error i followed the instructions in the link below:... (0 Replies)
Discussion started by: yakari
0 Replies

6. AIX

Wrong Shared objects getting loaded

I have two envoirmets(Envoirment A and Envoirment B) running on same server(AIX vesion 5.3).Both have different groups.I am facing a strange problem.Shared objects of one envoirment (Envoirment A)are getting loaded into the second(Envoirment B).So the servers that have dependency on shared objects... (2 Replies)
Discussion started by: nitin@tcs
2 Replies

7. Programming

g++ with -frepo and shared objects...

G'day, I have been working with a large application that makes extensive use of templates. When compiled under Unix (with g++), this sees some rather impressive bloat. I have been trying to make a temporary quick-fix by using the -frepo option, which results in dramatically smaller shared... (0 Replies)
Discussion started by: Elric of Grans
0 Replies

8. Programming

Creation and Accessing Shared Objects (.so)

Hi, I am looking for references about creating and accessing Shared Objects (.so) through C/C++ on Unix / Linux platforms. Is it possible and where can I find the info. Thanks Phil (1 Reply)
Discussion started by: phil nascimento
1 Replies

9. Linux

Make file for shared objects

dear Experts, please help, actually i am trying to create a .so(shared object through make file through ld) i am not understaning how to proceed i have tried like through command like i can do it in 2 step like my progam :test2.c $gcc -fPIC -c test2.c $ld -shared -soname test2.so -o... (1 Reply)
Discussion started by: vin_pll
1 Replies

10. UNIX for Advanced & Expert Users

Shared objects -urgent please help me out

Hi All...... I have my tool in my one server lets say E1 and same tool I tried to install in E2 server so everything is fine but, while executing the my tool for example... $ ./batch At that time Im getting this following error. ./batch: error while loading shared libraries: libqabwvcd.so:... (3 Replies)
Discussion started by: ksrivani
3 Replies
MREMAP(2)						     Linux Programmer's Manual							 MREMAP(2)

NAME
mremap - re-map a virtual memory address SYNOPSIS
#include <unistd.h> #include <sys/mman.h> void * mremap(void *old_address, size_t old_size , size_t new_size, unsigned long flags); DESCRIPTION
mremap expands (or shrinks) an existing memory mapping, potentially moving it at the same time (controlled by the flags argument and the available virtual address space). old_address is the old address of the virtual memory block that you want to expand (or shrink). Note that old_address has to be page aligned. old_size is the old size of the virtual memory block. new_size is the requested size of the virtual memory block after the resize. The flags argument is a bitmap of flags. In Linux the memory is divided into pages. A user process has (one or) several linear virtual memory segments. Each virtual memory seg- ment has one or more mappings to real memory pages (in the page table). Each virtual memory segment has its own protection (access rights), which may cause a segmentation violation if the memory is accessed incorrectly (e.g., writing to a read-only segment). Accessing virtual memory outside of the segments will also cause a segmentation violation. mremap uses the Linux page table scheme. mremap changes the mapping between virtual addresses and memory pages. This can be used to implement a very efficient realloc. FLAGS
MREMAP_MAYMOVE indicates if the operation should fail, or change the virtual address if the resize cannot be done at the current virtual address. RETURN VALUE
On success mremap returns a pointer to the new virtual memory area. On error, -1 is returned, and errno is set appropriately. ERRORS
EINVAL An invalid argument was given. Most likely old_address was not page aligned. EFAULT "Segmentation fault." Some address in the range old_address to old_address+old_size is an invalid virtual memory address for this process. You can also get EFAULT even if there exist mappings that cover the whole address space requested, but those mappings are of different types. EAGAIN The memory segment is locked and cannot be re-mapped. ENOMEM The memory area cannot be expanded at the current virtual address, and the MREMAP_MAYMOVE flag is not set in flags. Or, there is not enough (virtual) memory available. NOTES
With current glibc includes, in order to get the definition of MREMAP_MAYMOVE, you need to define _GNU_SOURCE before including <sys/mman.h>. CONFORMING TO
This call is Linux-specific, and should not be used in programs intended to be portable. 4.2BSD had a (never actually implemented) mremap(2) call with completely different semantics. SEE ALSO
getpagesize(2), realloc(3), malloc(3), brk(2), sbrk(2), mmap(2) Your favorite OS text book for more information on paged memory. (Modern Operating Systems by Andrew S. Tannenbaum, Inside Linux by Ran- dolf Bentson, The Design of the UNIX Operating System by Maurice J. Bach.) Linux 1.3.87 1996-04-12 MREMAP(2)
All times are GMT -4. The time now is 10:09 PM.
Unix & Linux Forums Content Copyright 1993-2022. All Rights Reserved.
Privacy Policy