The use of STL in shared memory poses a real problem since
(non-smart) pointers are not allowed there.
Is there any solution for containers in shared memory using
smart pointers? Where can I find templates? 11 4851
Użytkownik Michael Schuler napisał, On 2004-01-12 15:01: The use of STL in shared memory poses a real problem since (non-smart) pointers are not allowed there.
Is there any solution for containers in shared memory using smart pointers? Where can I find templates?
AFAIR to use shared memory you should use boost library you can find it at: http://www.boost.org/.
Documentation: http://www.boost.org/libs/smart_ptr/smart_ptr.htm
Best Regards
Darek Ostolski
--
ERROR 1164 HOW IN THE HELL DID YOU GET HERE
Hello,
take a look to <http://tplusplus.sourceforge.net/> resp. at its source code. It uses special
allocators to place the contents of STL containers into shared memory. It's been presented
at OOPSLA 2003.
Cheers,
Philipp.
"Michael Schuler" <Mi*************@fujitsu-siemens.com> wrote in message news:bt**********@news.fujitsu-siemens.com... The use of STL in shared memory poses a real problem since (non-smart) pointers are not allowed there.
Is there any solution for containers in shared memory using smart pointers? Where can I find templates?
On Mon, 12 Jan 2004 15:01:07 +0100, Michael Schuler
<Mi*************@fujitsu-siemens.com> wrote: The use of STL in shared memory poses a real problem since (non-smart) pointers are not allowed there.
Is there any solution for containers in shared memory using smart pointers? Where can I find templates?
If you can allocate your shared memory at a fixed address in each
process' address space, then you just need to write a custom allocator
to allocate from that memory. But if you can't allocate at a fixed
address, then you have to write a complex allocator that uses "offset"
smart pointers (and references!) that store an offset into the shared
memory rather than a pure pointer, and possibly a "shared memory ID"
or similar if you want to have more than one shared memory area in any
process. This is hard to do in a reliable way since C++ doesn't really
support smart references.
Tom
C++ FAQ: http://www.parashift.com/c++-faq-lite/
C FAQ: http://www.eskimo.com/~scs/C-faq/top.html
Thanks Tom,
but to my knowledge if I use an allocator, only the items are in shared
memory not the container itself. In my case teh container also must lie
in shared memory to survive a process failure :-(
tom_usenet wrote: On Mon, 12 Jan 2004 15:01:07 +0100, Michael Schuler <Mi*************@fujitsu-siemens.com> wrote:
The use of STL in shared memory poses a real problem since (non-smart) pointers are not allowed there.
Is there any solution for containers in shared memory using smart pointers? Where can I find templates?
If you can allocate your shared memory at a fixed address in each process' address space, then you just need to write a custom allocator to allocate from that memory. But if you can't allocate at a fixed address, then you have to write a complex allocator that uses "offset" smart pointers (and references!) that store an offset into the shared memory rather than a pure pointer, and possibly a "shared memory ID" or similar if you want to have more than one shared memory area in any process. This is hard to do in a reliable way since C++ doesn't really support smart references.
Tom
C++ FAQ: http://www.parashift.com/c++-faq-lite/ C FAQ: http://www.eskimo.com/~scs/C-faq/top.html
Hello,
that's true, but imposes no problem at all, because there's "placement new",
i.e. given we talk about Unix and given there's some magic "shmAllocator<>",
you can write sth like
...
typedef std::map< int,std::less< int >,shmAllocator< int > > shmIntMap;
shm=shmget(shm_key,sizeof(shmIntMap),IPC_CREAT|070 0);
void *const p=(shmIntMap *)shmat(shm,NULL,0);
shmIntMap *m=new(p) shmIntMap;
...
Then, don't forget to manually call the destructor afterwards.
Now, you might ask, o.k., now both the map and the ints contained within the map
go into the shared memory - but what's about the linked structure that establishes
the tree the map internally consists of, i.e. that glues together the container and
its elements? The answer is the "rebind<>" member template class a conforming
allocator has to have.
Cheers,
Philipp. but to my knowledge if I use an allocator, only the items are in shared memory not the container itself. In my case teh container also must lie in shared memory to survive a process failure :-(
Is there any solution for containers in shared memory using smart pointers? Where can I find templates?
If you can allocate your shared memory at a fixed address in each process' address space, then you just need to write a custom allocator to allocate from that memory. But if you can't allocate at a fixed address, then you have to write a complex allocator that uses "offset" smart pointers (and references!) that store an offset into the shared memory rather than a pure pointer, and possibly a "shared memory ID" or similar if you want to have more than one shared memory area in any process. This is hard to do in a reliable way since C++ doesn't really support smart references.
"void" <chq@nie_spamuj.wp.pl> wrote in message
news:29**************@213.17.192.132... Użytkownik Michael Schuler napisał, On 2004-01-12 15:01:
The use of STL in shared memory poses a real problem since (non-smart) pointers are not allowed there.
Is there any solution for containers in shared memory using smart pointers? Where can I find templates? AFAIR to use shared memory you should use boost library you can find
it at: http://www.boost.org/. Documentation: http://www.boost.org/libs/smart_ptr/smart_ptr.htm
Boost has excellent smart pointers, but does not provide any help with
accessing shared memory.
Jonathan
Thanks Philipp, good hint,
but I still have the problem with pointers inside the container,
whereas pointers pose problems since the shared memory may have
different addresses in different processes :-(
Philipp Bachmann Hello,
that's true, but imposes no problem at all, because there's "placement new", i.e. given we talk about Unix and given there's some magic "shmAllocator<>", you can write sth like
... typedef std::map< int,std::less< int >,shmAllocator< int > > shmIntMap; shm=shmget(shm_key,sizeof(shmIntMap),IPC_CREAT|070 0); void *const p=(shmIntMap *)shmat(shm,NULL,0); shmIntMap *m=new(p) shmIntMap; ...
Then, don't forget to manually call the destructor afterwards.
Now, you might ask, o.k., now both the map and the ints contained within the map go into the shared memory - but what's about the linked structure that establishes the tree the map internally consists of, i.e. that glues together the container and its elements? The answer is the "rebind<>" member template class a conforming allocator has to have.
Cheers, Philipp.
but to my knowledge if I use an allocator, only the items are in shared memory not the container itself. In my case teh container also must lie in shared memory to survive a process failure :-(
Is there any solution for containers in shared memory using smart pointers? Where can I find templates?
If you can allocate your shared memory at a fixed address in each process' address space, then you just need to write a custom allocator to allocate from that memory. But if you can't allocate at a fixed address, then you have to write a complex allocator that uses "offset" smart pointers (and references!) that store an offset into the shared memory rather than a pure pointer, and possibly a "shared memory ID" or similar if you want to have more than one shared memory area in any process. This is hard to do in a reliable way since C++ doesn't really support smart references.
The T++ I've referenced seems to use a shared memory pool of a fixed size,
if I correctly understood its source code. But check this out, please:
<http://allocator.sourceforge.net/>. This looks quite promising, doesn't
it?
Cheers,
Philipp. but I still have the problem with pointers inside the container, whereas pointers pose problems since the shared memory may have different addresses in different processes :-( that's true, but imposes no problem at all, because there's "placement new", i.e. given we talk about Unix and given there's some magic "shmAllocator<>", you can write sth like
... typedef std::map< int,std::less< int >,shmAllocator< int > > shmIntMap; shm=shmget(shm_key,sizeof(shmIntMap),IPC_CREAT|070 0); void *const p=(shmIntMap *)shmat(shm,NULL,0); shmIntMap *m=new(p) shmIntMap; ...
Then, don't forget to manually call the destructor afterwards.
Now, you might ask, o.k., now both the map and the ints contained within the map go into the shared memory - but what's about the linked structure that establishes the tree the map internally consists of, i.e. that glues together the container and its elements? The answer is the "rebind<>" member template class a conforming allocator has to have.but to my knowledge if I use an allocator, only the items are in shared memory not the container itself. In my case teh container also must lie in shared memory to survive a process failure :-( >Is there any solution for containers in shared memory using >smart pointers? Where can I find templates?
If you can allocate your shared memory at a fixed address in each process' address space, then you just need to write a custom allocator to allocate from that memory. But if you can't allocate at a fixed address, then you have to write a complex allocator that uses "offset" smart pointers (and references!) that store an offset into the shared memory rather than a pure pointer, and possibly a "shared memory ID" or similar if you want to have more than one shared memory area in any process. This is hard to do in a reliable way since C++ doesn't really support smart references.
On Tue, 13 Jan 2004 15:05:50 +0100, Michael Schuler
<Mi*************@fujitsu-siemens.com> wrote: Thanks Philipp, good hint,
but I still have the problem with pointers inside the container, whereas pointers pose problems since the shared memory may have different addresses in different processes :-(
Solution 1 (easy if your OS supports it):
Map the shared memory to the same address in every process that
accesses it. Some OSes allow you to at least attempt this, and if it
fails you can bail out.
Solution 2 (a nightmare, not really worth it?):
Assuming your standard library is up to it (the only one I'm aware of
is Dinkumware, Metrowerks might be, libstdc++ and STLport aren't), it
might be possible to write a smart pointer and smart reference that
work using an offset into the shared memory (which will be the same
for every process). You'll obviously have to use those classes for any
other pointers/references you want to put in shared memory. I wrote
some such class templates a while back, with mixed success.
Solution 2 (probably best if you can't map to a fixed address):
Don't use the STL, but your own simpler code. Use explicit offsets,
not pointers.
Tom
C++ FAQ: http://www.parashift.com/c++-faq-lite/
C FAQ: http://www.eskimo.com/~scs/C-faq/top.html
Thanks. Great answer. I will probably map the shared memory at some
fixed address.
tom_usenet wrote: On Tue, 13 Jan 2004 15:05:50 +0100, Michael Schuler <Mi*************@fujitsu-siemens.com> wrote:
Thanks Philipp, good hint,
but I still have the problem with pointers inside the container, whereas pointers pose problems since the shared memory may have different addresses in different processes :-(
Solution 1 (easy if your OS supports it):
Map the shared memory to the same address in every process that accesses it. Some OSes allow you to at least attempt this, and if it fails you can bail out.
Solution 2 (a nightmare, not really worth it?):
Assuming your standard library is up to it (the only one I'm aware of is Dinkumware, Metrowerks might be, libstdc++ and STLport aren't), it might be possible to write a smart pointer and smart reference that work using an offset into the shared memory (which will be the same for every process). You'll obviously have to use those classes for any other pointers/references you want to put in shared memory. I wrote some such class templates a while back, with mixed success.
Solution 2 (probably best if you can't map to a fixed address):
Don't use the STL, but your own simpler code. Use explicit offsets, not pointers.
Tom
C++ FAQ: http://www.parashift.com/c++-faq-lite/ C FAQ: http://www.eskimo.com/~scs/C-faq/top.html
Philipp Bachmann The T++ I've referenced seems to use a shared memory pool of a fixed size, if I correctly understood its source code. But check this out, please: <http://allocator.sourceforge.net/>. This looks quite promising, doesn't it?
Yes, looks great. Thanks for the hint. Cheers, Philipp.
but I still have the problem with pointers inside the container, whereas pointers pose problems since the shared memory may have different addresses in different processes :-(
that's true, but imposes no problem at all, because there's "placement new", i.e. given we talk about Unix and given there's some magic "shmAllocator<>", you can write sth like
... typedef std::map< int,std::less< int >,shmAllocator< int > > shmIntMap; shm=shmget(shm_key,sizeof(shmIntMap),IPC_CREAT|070 0); void *const p=(shmIntMap *)shmat(shm,NULL,0); shmIntMap *m=new(p) shmIntMap; ...
Then, don't forget to manually call the destructor afterwards.
Now, you might ask, o.k., now both the map and the ints contained within the map go into the shared memory - but what's about the linked structure that establishes the tree the map internally consists of, i.e. that glues together the container and its elements? The answer is the "rebind<>" member template class a conforming allocator has to have.
but to my knowledge if I use an allocator, only the items are in shared memory not the container itself. In my case teh container also must lie in shared memory to survive a process failure :-(
>>Is there any solution for containers in shared memory using >>smart pointers? Where can I find templates? > >If you can allocate your shared memory at a fixed address in each >process' address space, then you just need to write a custom allocator >to allocate from that memory. But if you can't allocate at a fixed >address, then you have to write a complex allocator that uses "offset" >smart pointers (and references!) that store an offset into the shared >memory rather than a pure pointer, and possibly a "shared memory ID" >or similar if you want to have more than one shared memory area in any >process. This is hard to do in a reliable way since C++ doesn't really >support smart references. This thread has been closed and replies have been disabled. Please start a new discussion. Similar topics
by: Srijit Kumar Bhadra |
last post by:
Hello,
Here is some sample code with pywin32 build 203 and ctypes 0.9.6.
Best regards,
/Srijit
File: SharedMemCreate_Mutex_win32all.py
# This application should be used with...
|
by: alanrn |
last post by:
I would like to start a dialog on how to implement the equivalent
functionality of UNIX shared memory in .NET. I work with a factory
automation system. The bulk of the system is written in C/C++....
|
by: myren, lord |
last post by:
When I first discovered shared memory (between multiple processes) I
immediately started thinking of how to build my own VM subsystem +
locking mechanisms for a large single block of memory. This...
|
by: phil_gg04 |
last post by:
Dear C++ Experts,
Over the last couple of months I have been writing my first program
using shared memory. It has been something of an "in-at-the-deep-end"
experience, to say the least. At...
|
by: Jeremy |
last post by:
Hi all,
I'm getting very confused about how DB2 uses shared memory and I
wonder if someone could clarify matters for me, please ?
We are running 32bit DB2 V7.2 FP9 under AIX 4.3.3 on a machine...
|
by: Jim |
last post by:
Hello,
I have a broken server that we are going to be moving off to a new server
with a new version of DB2 but here is what I have right now:
RedHat 7.0 (2.2.24smp)
DB2 v6.1.0.40
I am...
|
by: herbert |
last post by:
I am coding a dozen "background" realtime apps for factory automation in .NET
2.0.
The apps need to share a common memory as there are lots of variables to be
shared (and synchronized of...
|
by: llothar |
last post by:
Hello,
i need to manage a heap in shared memory.
Does anybody know about a portable (win32+mac+posix) c implementation
for this.
|
by: Sune |
last post by:
Hi all,
I want to make data stored in-memory (not disk) available to several
processes. My concern is that poorly written C applications with
dangling pointers may(will) damage the data in this...
|
by: ryjfgjl |
last post by:
In our work, we often receive Excel tables with data in the same format. If we want to analyze these data, it can be difficult to analyze them because the data is spread across multiple Excel files...
|
by: emmanuelkatto |
last post by:
Hi All, I am Emmanuel katto from Uganda. I want to ask what challenges you've faced while migrating a website to cloud.
Please let me know.
Thanks!
Emmanuel
|
by: BarryA |
last post by:
What are the essential steps and strategies outlined in the Data Structures and Algorithms (DSA) roadmap for aspiring data scientists? How can individuals effectively utilize this roadmap to progress...
|
by: nemocccc |
last post by:
hello, everyone, I want to develop a software for my android phone for daily needs, any suggestions?
|
by: Hystou |
last post by:
There are some requirements for setting up RAID:
1. The motherboard and BIOS support RAID configuration.
2. The motherboard has 2 or more available SATA protocol SSD/HDD slots (including MSATA, M.2...
|
by: Oralloy |
last post by:
Hello folks,
I am unable to find appropriate documentation on the type promotion of bit-fields when using the generalised comparison operator "<=>".
The problem is that using the GNU compilers,...
|
by: jinu1996 |
last post by:
In today's digital age, having a compelling online presence is paramount for businesses aiming to thrive in a competitive landscape. At the heart of this digital strategy lies an intricately woven...
|
by: Hystou |
last post by:
Overview:
Windows 11 and 10 have less user interface control over operating system update behaviour than previous versions of Windows. In Windows 11 and 10, there is no way to turn off the Windows...
|
by: agi2029 |
last post by:
Let's talk about the concept of autonomous AI software engineers and no-code agents. These AIs are designed to manage the entire lifecycle of a software development project—planning, coding, testing,...
| |