Hello!
I need to handle stack overflow in Linux from programs written in C/C++. I've surprisingly found that it's not not easy... In Windows 2000 I'm using structured exception handling for this purpose, but what mechanism should be used in Linux? Writing handlers for all signals that can be handled didn't help me. The result of program execution is always Segmentation fault.
Thanks in advance,
Foma
This Question has been solved and asker verified All Experts Exchange premium technology solutions are available to subscription members.
Experts Exchange has been collecting answers to technology questions since 1996…3 million and counting! If you have a question, chances are we already have your answer.
If you can't find the exact answer you're looking for, ask our exclusive community of 50,000 experts. You’ll get a personalized answer from a trusted professional.
Thousands of free tech tips, tricks, how-to’s and tutorials are available in our peer reviewed articles section. See for yourself how smart our experts are, no login required.
Save solutions to your questions, answers you’ve discovered through searching plus helpful articles in your personal knowledgebase for easy future access.
Access the answers to your technology questions today.
30-day free trial. Register in 60 seconds.
Members of the expert community talk about why the experience at Experts Exchange is different than what you will find anywhere else.
Try it out and discover for yourself.
30-day free trial. Register in 60 seconds.
Join the community of experts here and help other tech pros by answering question in your area of expertise. You can earn FREE access to all Experts Exchange's premium features and resources.
Hi Stefan!
Actually, lots of program need to handle stack overflow. For example i'm writing interpreter for a programming language. And of course, a user can write a program in this language that will cause infinite number of function calls, which leads to stack overflow.
Hm, I'm sure Java virtual machine, for instance, handle this problem correctly on any platform, including Linux.
Best regards,
Foma
Foma,
True, but I don't think the Java virtual machine uses the system's stack for calling java methods.
Your interpreter should probably not recursive itself. I'm quite sure it is possible to calculate the maximum stack size.
A simple method is to call
pmap $$
...and then you see the current process's stack size + location. You can then easily add a call to check the stack to each function invocation. Make sure you have a bit of stack to spare when you trigger the overflow exception.
That's a classic for a simple assembly subroutine, something like (GCC's AT&T syntax):
cmp $_stack_limit,%esp
ble _overflow
ret
...whereas "overflow" is a C routine which handles the stack overflow itself. You could even have the ASM in a little macro which you call in all the critical functions, as the overhead is minimal.
Catching SIGSEGV should work, except that signal handlers are executed on the current stack. So, when the stack overflows the signal handler will cause a segmentation fault too. To avoid this, you will have to establish an alternate signal stack using sigaltstack() and pass the SA_ONSTACK flag to sigaction().
you can also give libsigsegv a try. here's the URL: http://libsigsegv.sourcefo
-- Adil
I strongly discourage catching segmentation faults. Chances are high that you get hundreds of them in a very short time. And you have no guarantee that they are related to your stack problem.
The worst case would be that you're missing a core dump which was created by a NULL reference or something similar, and pass on to the user a totally wrong "stack overflow message".
Thank you very much for your responces!
To mtmike:
Thank you. Simple solution and it works. Unfortunately I cannot distinguish stack overflow from other segmentation violations…
To Sadrul:
Nice library. It allows handling exactly stack overflows. It distinguishes arbitrary segmentation fault and stack overflow using some heuristics (I saw it in its source code, but I didn’t manage to look into it yet). Unfortunately, this library is not suitable to me because I need to return from signal handler to program automatically, what is achieved by using sigaction with SA_SIGINFO. But the library uses sigaction w/o this flag for Linux…
To stefan73:
I understand your position, but I can’t agree with you. Yes, in general it’s better not to get segmentation faults at all and write programs in such a manner that you will never get access violation. But sometimes it’s very hard or impossible… Concerning my program, I do my own memory management and SIGSEGV signal (when the address in selected diapason) is normal behavior of the program. When SIGSEGV signal is got I perform mmap call to correct the problem. In other case it can be stack overflow or error in program. I can suspect that it is stack overflow and make an effort to release resources. If during this step SIGSEGV arises, that means error in program and I just terminate the program… I think it not an ideal solution, but it is satisfactory for me
Business Accounts
Answer for Membership
by: stefan73Posted on 2004-10-15 at 04:49:00ID: 12318014
Hi Foma,
I'm not aware of any tools which prevent stack overflows. You can set the stack limit with "limit" or "ulimit", but you cannot really detect them as such.
What are you putting on the stack which causes an overflow? Normally it's better to revise the initial design than to chase after error symptoms later.
Cheers!
Stefan