I know some Java and am right now trying C++ instead and apparently in C++ you can do things like declare an int array of size 6, then change the 10th element of that array, which I'm understanding to be simply the 4th byte after the end of the section of memory that was allocated for the 6-integer array.
So my question is, if I'm careless is it possible to accidentally alter memory in my C++ program that is being used by other programs on my system? Is there an actual risk of seriously messing something up this way? I mean I know you can just restart your computer and clear the memory if you have to, but if I don't do that, there could be some lasting damage.
It depends on your system. Formally, an out of bounds access is
undefined behavior. On a modern general purpose system, each user
process has its own address space, and one process can't modify, or even
read, that of another process (barring shared memory), so unless you're
writing kernel code, you shouldn't be able to break anything outside of
your own process (and non-kernel code can't normally do physical IO, so
I don't see how anything in the hardware could break).
If you're writing kernel code, however, or working on an embedded
processor with no memory mapping or protection, you can literally
destroy hardware with an out of bounds write; if the program is
controlling something like a nuclear power plant, you can even destroy a
lot more than just the machine your code is running on.
Each process is given its own virtual address space, so naturally processes don't see each others memory. Don't forget that even a buffer overrun that is local to your program can have dire consequences - the overrun may cause the program to misbehave and do something that has lasting effect (like deleting all files for example).
This depends on what operating system and environment you are in:
Normal OS (Windows, Linux etc) userspace program: You can only mess up your own process memory. However having really bad luck this can be enough. Imagine for example that you make a call to some function that deletes files. If your memory is corrupted at the time of the call, the parameters to the function might be messed up to mean deletion of something else than you intended. As long as you keep from calling delete file routines etc. in the programs where you test memory handling, this risk is non-existent.
Normal OS, kernel space device driver: You can access system memory and the memory of the currently running process, possibly destroying everything.
Simple embedded OS without memory protection: You can access everything and destroy anything.
Legacey OS without memory protection (Win 3.X, MS-DOS): You can access everyting and destroy anything.
Every program runs in its own address space and one program cannot access (read / modify) any other programs address space (this is a memory management technique called paging).
If you try to access an address in memory that your program cannot read it will cause a segmentation or page fault and your program will crash.
In answer to your question, no permanent damage will be caused.
I'm not sure modern OS (especially win7) allows you to do that. The OS will block the buffer overrun action as you've described
Back in the day (DOS times) some viruses would try to damage hardware by directly programming video or hard drive controller, but even then it was not easy or sure thing. Modern hardware and OSs make it practicaly impossible for user level applications to damage hardware. So program away :) you wont break anything.
There is a different possibility. Having a buffer overrun might allow ill reputed people to exploit that bug and execute arbitrary code in your clients computer. I guess thats bad enough. And the most dangerous part about overrun is that you may not find it even after a seriously excessive test.
Related
I'm a little unclear exactly how these objects function that form bridges between software and hardware. Are they pretty much just software objects that get destroyed if you leak them onto the heap when you terminate the process? Or is there something more to it?
The reason I ask is I forgot to have my initialization routine change its statemachine and hence switch routines resulting in it creating as many "DeviceContexts" and "Devices" as it possibly could and reassigning them to the same pointers (via d3d11createdevice) before I caught my memory leaking at about 2GB.
Then it occurred to me that I really have no clue what it means to fail to release these objects. Is there a hardware component to them that I should be concerned about if these objects were to be leaked such that I need to reset my computer? Or does terminating the process pretty much clean up the mess?
I cold reset my computer regardless just to be sure. But it would be nice to know exactly what happens when you are using low level interfaces like this and you fail to properly destroy/release them.
The operating system will clean up all those device contexts when your program terminates. Otherwise a misbehaved program could bring the system to a standstill.
Your other concern (expressed in a comment) about damaging hardware shouldn't be possible either. If it was a malicious program could wreak all sorts of havoc. You may be possible to damage hardware by directly accessing it, but that sort of access is what the driver (and the device context, which sits between your program and the driver) is for.
I'm trying to understand how the objects (variables, functions, structs, etc) work in c++. In this case I see there are basically two ways of storing them: the stack and the heap. Accordingly, whenever the heap storage is used it needs to be dealocated manually, but if the stack is used, then the dealocation is automaticcally done. so my question is related to the kinds of problems that bad practice might cause the program itself or to the computer. For example:
1.- Let'suposse that I run a program with a recursion solution by using an infinite iteration of functions. Theoretically the program crashes (stack overflow), but does it cause some trouble to the computer itself? (To the RAM maybe or to the SO).
2.- What happens if I forget to dealocate memory on the heap. I mean, does it just cause trouble to the program or it is permanent to the computer in general. I mean it might be that such memory could not be used never again or something.
3.- What are the problems of getting a segmentation fault (the heap).
Some other dangers or cares relevant to this are welcome.
Accordingly, whenever the stack storage is used it needs to be
dealocated manually, but if the heap is used, then the dealocation is
automaticcally done.
When you use stack - local variables in the function - they are deallocated automatically when the function ends (returns).
When you allocate from the heap, the memory allocated remains "in use" until it is freed. If you don't do that, your program, if it runes for long enough and keep allocating "stuff", will use all memory available to it, and eventually fail.
Note that "stackfault" is almost impossible to recover from in an application, because the stack is no longer usable when it's full, and most operations to "recover from error" will involve using SOME stack memory. The processor typically has a special trap to recover from stack fault, but that lands insise the operating system, and if the OS determines the application has run out of stack, it often shows no mercy at all - it just "kills" the application immediately.
1.- Let'suposse that I run a program with a recursion solution by using an infinite iteration of functions. Theoretically the program
crashes (stack overflow), but does it cause some trouble to the
computer itself? (To the RAM maybe or to the SO).
No, the computer itself is not harmed by this in and of itself. There may of course be data-loss if your program wasn't saving something that the user was working on.
Unless the hardware is very badly designed, it's very hard to write code that causes any harm to the computer, beyond loss of stored data (of course, if you write a program that fills the entire hard disk from the first to the last sector, your data will be overwritten with whatever your program fills the disk with - which may well cause the machine to not boot again until you have re-installed an operating system on the disk). But RAM and processors don't get damaged by bad coding (fortunately, as most programmers make mistakes now and again).
2.- What happens if I forget to dealocate memory on the heap. I mean, does it just cause trouble to the program or it is permanent to the
computer in general. I mean it might be that such memory could not be
used never again or something.
Once the program finishes (and most programs that use "too much memory" does terminate in some way or another, at some point).
Of course, how well the operating system and other applications handle "there is no memory at all available" varies a little bit. The operating system in itself is generally OK with it, but some drivers that are badly written may well crash, and thus cause your system to reboot if you are unlucky. Applications are more prone to crashing due to there not being enough memory, because allocations end up with NULL (zero) as the "returned address" when there is no memory available. Using address zero in a modern operating system will almost always lead to a "Segmentation fault" or similar problem (see below for more on that).
But these are extreme cases, most systems are set up such that one application gobbling all available memory will in itself fail before the rest of the system is impacted - not always, and it's certainly not guaranteed that the application "causing" the problem is the first one to be killed if the OS kills applications simply because they "eat a lot of memory". Linux does have a "Out of memory killer", which is a pretty drastic method to ensure the system can continue to work [by some definition of "work"].
3.- What are the problems of getting a segmentation fault (the heap).
Segmentation faults don't directly have anything to do with the heap. The term segmentation fault comes from older operating systems (Unix-style) that used "segments" of memory for different usages, and "Segmentation fault" was when the program went outside it's allocated segment. In modern systems, the memory is split into "pages" - typically 4KB each, but some processors have larger pages, and many modern processors support "large pages" of, for examble, 2MB or 1GB, which is used for large chunks of memory.
Now, if you use an address that points to a page that isn't there (or isn't "yours"), you get a segmentation fault. This, typically will end the application then and there. You can "trap" segmentation fault, but in all operating systems I'm aware of, it's not valid to try to continue from this "trap" - but you could for example store away some files to explain what happened and help troubleshoot the problem later, etc.
Firstly, your understanding of stack/heap allocations is backwards: stack-allocated data is automatically reclaimed when it goes out of scope. Dynamically-allocated data (data allocated with new or malloc), which is generally heap-allocated data, must be manually reclaimed with delete/free. However, you can use C++ destructors (RAII) to automatically reclaim dynamically-allocated resources.
Secondly, the 3 questions you ask have nothing to do with the C++ language, but rather they are only answerable with respect to the environment/operating system you run a C++ program in. Modern operating systems generally isolate processes so that a misbehaving process doesn't trample over OS memory or other running programs. In Linux, for example, each process has its own address space which is allocated by the kernel. If a misbehaving process attempts to write to a memory address outside of its allocated address space, the operating system will send a SIGSEGV (segmentation fault) which usually aborts the process. Older operating systems, such as MS-DOS, didn't have this protection, and so writing to an invalid pointer or triggering a stack overflow could potentially crash the whole operating system.
Likewise, with most mainstream modern operating systems (Linux/UNIX/Windows, etc.), memory leaks (data which is allocated dynamically but never reclaimed) only affect the process which allocated them. When the process terminates, all memory allocated by the process is reclaimed by the OS. But again, this is a feature of the operating system, and has nothing to do with the C++ language. There may be some older operating systems where leaked memory is never reclaimed, even by the OS.
1.- Let'suposse that I run a program with a recursion solution by using an infinite iteration of functions. Theoretically the program crashes (stack overflow), but does it cause some trouble to the computer itself? (To the RAM maybe or to the SO).
A stack overflow should not cause trouble neither to the Operating System nor to the computer. Any modern OS provides an isolated address space to each process. When a process tries to allocate more data in its stack than space is available, the OS detects it (usually via an exception) and terminates the process. This guarantees that no other processes are affected.
2.- What happens if I forget to dealocate memory on the heap. I mean, does it just cause trouble to the program or it is permanent to the computer in general. I mean it might be that such memory could not be used never again or something.
It depends on whether your program is a long running process or not, and the amount of data that you're failing to deallocate. In a long running process (e.g. a server) a recurrent memory leak can lead to thrashing: after some time, your process will be using so much memory that it won't fit in your physical memory. This is not a problem per se, because the OS provides virtual memory but the OS will spend more time moving memory pages from your physical memory to disk than doing useful work. This can affect other processes and it might slow down the system significantly (to the point that it might be better to reboot it).
3.- What are the problems of getting a segmentation fault (the heap).
A Segmentation Fault will crash your process. It's not directly related to the usage of the heap, but rather to accessing a memory region that does not belong to your process (because it's not part of its address space or because it was, but it was freed). Depending on what your process was doing, this can cause other problems: for instance, if the process was writing to a file when the crash happened it's very likely that it will end up corrupt.
First, stack means automatic memory and heap means manual memory. There are ways around both, but that's generally a more advanced question.
On modern operating systems, your application will crash but the operating system and machine as a whole will continue to function. There are of course exceptions to this rule, but they're (again) a more advanced topic.
Allocating from the heap and then not deallocating when you're done just means that your program is still considered to be using the memory even though you're not using it. If left unchecked, your program will fail to allocate memory (out of memory errors). How you handle out-of-memory errors could mean anything from a crash (unhandled error resulting in an unhandled exception or a NULL pointer being accessed and generating a segmentation fault) to odd behavior (caught exception or tested for NULL pointer but no special handling case) to nothing at all (properly handled).
On modern operating systems, the memory will be freed when your application exits.
A segmentation fault in the normal sense will simply crash your application. The operating system may immediately close file or socket handles. It may also perform a dump of your application's memory so that you can try to debug it posthumously with tools designed to do that (more advanced subject).
Alternatively, most (I think?) modern operating systems will use a special method of telling the program that it has done something bad. It is then up to the program's code to decide whether or not it can recover from that or perhaps add additional debug information for the operating system, or whatever really.
I suggest you look into auto pointers (also called smart pointers) for making your heap behave a little bit like a stack -- automatically deallocating memory when you're done using it. If you're using a modern compiler, see std::unique_ptr. If that type name can't be found, then look into the boost library (google). It's a little more advanced but highly valuable knowledge.
Hope this helps.
I want to run my C++ program after flushing the cache, Before running my program I do not know what is there in the cache. Is there some other manner in C++ on Ubuntu via which I may flush my cache before running my program.
EDIT: The motive for flushing the cache is... that each time I run my program I do not want my present data structures to be there in the cache... I mean I want a cold cache... whereby all the accesses are made from the disk.
One way of achieving this is to restart the computer... but considering the number of experiments that I have to run, this is not feasible for me. So, can anyone be kind enough to guide me as to how I can achieve this.
You have no need to flush the cache from your user-mode (non-kernel-mode) program. The OS (Linux, in the case of ubuntu) provides your application with a fresh virtual address space, with no "leftover stuff" from other programs. Without executing special OS system calls, your program can't even get to memory that's used for other applications. So from a cache perspective, your application start from a clean slate, as far as it's concerned. There are cacheflush() system calls (syntax differs by OS), but unless you're doing something out-of-the-ordinary for typical user-mode applications, you can just forget that the cache even exists; it's just there to speed up your program, and the OS manages it via the CPU's MMU, your app does not need to manage it.
You may have also heard about "memory leaks" (memory allocated to your application that your application forgets to free/delete, which is "lost forever" once your application forgets about it). If you're writing a (potentially) long-running program, leaked memory is definitely a concern. But leaked memory is only an issue for the application that leaks it; in modern virtual-memory environments, if application A leaks memory, it doesn't affect application B. And when application A exits, the OS clears out its virtual address space, and any leaked memory is at that point reclaimed by the system and no longer consumes any system resources whatsoever. In many cases, programmers specifically choose to NOT free/delete a memory allocation, knowing that the OS will automatically reclaim the entire amount of the memory when the application exits. There's nothing wrong with that strategy, as long as the program doesn't keep doing that on a repeating basis, exhausting its virtual address space.
This is a common question.
Firstly you have to understand that the caches are never really empty, just like a register is never really empty, it's always there, and it always has a value. The phrase "Flushing the cache" actually refers to writing the cache contents to memory, also called a memory barrier.
see https://en.wikipedia.org/wiki/Memory_barrier
This is not your problem, and so you are using the wrong terminology.
What you really want is to fill the cache with the wrong values. This is harder than it sounds, because you are fighting all the optimisations that normally are your friend. Memcpy'ing a large block of memory (several MB - given the size of todays caches) should normally work though.
However...
You also have file caches and other things that will give your application an unfair advantages. This can be a very complex subject, and is a small project in it's own right.
I'm writing a program (a theorem prover as it happens) whose memory requirement is "as much as possible, please"; that is, it can always do better by using more memory, for practical purposes without upper bound, so what it actually needs to do is use just as much memory as is available, no more and no less. I can figure out how to prioritize data to delete the lowest value stuff when memory runs short; the problem I'm trying to solve is how to tell when this is happening.
Ideally I would like a system call that returns "how much memory is left" or "are we out of memory yet?"; as far as I can tell, no such thing exists?
Of course, malloc can signal out of memory by returning 0 and new can call a handler; these aren't ideal signals, but would be better than nothing. A problem, however, is that I really want to know when physical memory is running out, so I can avoid going deep into swap and thereby making everything grind to a halt; I don't suppose there's any way to ask "are we having to swap yet?" or tell the operating system "don't swap on my account, just fail my requests if it comes to that"?
Another approach would be to find out how much RAM is in the machine, and monitor how much memory the program is using at the moment. As far as I know, there is generally no way to tell the former? I also get the impression there is no reliable way to tell the latter except by wrapping malloc/free with a bookkeeper function (which is then more problematic in C++).
Are there any approaches I'm missing?
The ideal would be a portable solution, but I suspect that's not going to happen. Failing that, a solution that works on Windows and another one that works on Unix would be nice. Failing that, I could get by with a solution that works on Windows and another one that works on Linux.
I think the most useful and flexible way to use all the memory available is to let the user specify how much memory to use.
Let the user write it in a config file or through an interface, then create an allocator (or something similar) that will not provide more than this memory.
That way, you don't have to find statistics about the current computer as this will allways be biased by the fact that the OS could also run other programs as well. Don't even talk about the way the OS will manage cache, the differences between 32 and 64 bit making adress space limit your allocations etc.
In the end, human intelligence (assuming the user know about the context of use) is cheaper to implement when provided by the user.
To find out how much system memory is still unused, under Linux you can parse the file /proc/meminfo and look for a line starting with "MemFree:". Under Windows you can use GlobalMemoryStatusEx http://msdn.microsoft.com/en-us/library/aa366589%28VS.85%29.aspx
Relying on malloc to return 0 when no memory is available might cause problems on Linux, because Linux overcommits memory allocations. malloc will usually return a valid pointer (unless the process is out of virtual address space), but accessing the memory it points to may trigger the "OOM killer", a mechanism that kills your process or another process on the system. The system administrator can tune this behavior.
The best solution I can think of might be to query how many page faults have occurred within, say, the last second. If there's a lot of swapping going on, you should probably release some memory, and if not, you can try allocating more memory.
On Windows, WMI can probably give you some statistics you can use.
But it's a tough problem, since there is no hard limit you can ask the OS for and then stay below. You can keep allocating memory far beyond the point where you've run out of physical memory, which just means you'll cripple your process with excessive swapping.
So the best you can really do is some kind of approximation.
You can keep allocating memory beyond the point where it is useful to do so - i.e. that which requires the OS to swap, or page out important things. The trouble is, it is not necessarily easy to tell where this is.
Also, if your task does any (significant) IO, you will need to have some left for the OS buffers.
I recommend just examining how much there is in the machine, then allocating an amount as a function of that (proportion, or leave some free etc).
I'm writing an caching app that consumes large amounts of memory.
Hopefully, I'll manage my memory well enough, but I'm just thinking about what
to do if I do run out of memory.
If a call to allocate even a simple object fails, is it likely that even a syslog call
will also fail?
EDIT: Ok perhaps I should clarify the question. If malloc or new returns a NULL or 0L value then it essentially means the call failed and it can't give you the memory for some reason. So, what would be the sensible thing to do in that case?
EDIT2: I've just realised that a call to "new" can throw an exception. This could be caught at a higher level so I can perhaps gracefully exit further up. At that point, it may even be possible to recover depending on how much memory is freed. In the least I should by that point hopefully be able to log something. So while I have seen code that checks the value of a pointer after new, it is unnecessary. While in C, you should check the return value for malloc.
Well, if you are in a case where there is a failure to allocate memory, you're going to get a std::bad_alloc exception. The exception causes the stack of your program to be unwound. In all likelihood, the inner loops of your application logic are not going to be handling out of memory conditions, only higher levels of your application should be doing that. Because the stack is getting unwound, a significant chunk of memory is going to be free'd -- which in fact should be almost all the memory used by your program.
The one exception to this is when you ask for a very large (several hundred MB, for example) chunk of memory which cannot be satisfied. When this happens though, there's usually enough smaller chunks of memory remaining which will allow you to gracefully handle the failure.
Stack unwinding is your friend ;)
EDIT: Just realized that the question was also tagged with C -- if that is the case, then you should be having your functions free their internal structures manually when out of memory conditions are found; not to do so is a memory leak.
EDIT2: Example:
#include <iostream>
#include <vector>
void DoStuff()
{
std::vector<int> data;
//insert a whole crapload of stuff into data here.
//Assume std::vector::push_back does the actual throwing
//i.e. data.resize(SOME_LARGE_VALUE_HERE);
}
int main()
{
try
{
DoStuff();
return 0;
}
catch (const std::bad_alloc& ex)
{ //Observe that the local variable `data` no longer exists here.
std::cerr << "Oops. Looks like you need to use a 64 bit system (or "
"get a bigger hard disk) for that calculation!";
return -1;
}
}
EDIT3: Okay, according to commenters there are systems out there which do not follow the standard in this regard. On the other hand, on such systems, you're going to be SOL in any case, so I don't see why they merit discussion. But if you are on such a platform, it is something to keep in mind.
Doesn't this question make assumptions regarding overcommitted memory?
I.e., an out of memory situation might not be recoverable! Even if you have no memory left, calls to malloc and other allocators may still succeed until the program attempts to use the memory. Then, BAM!, some process gets killed by the kernel in order to satisfy memory load.
I don't have any specific experience on Linux, but I spent a lot of time working in video games on games consoles, where running out of memory is verboten, and on Windows-based tools.
On a modern OS, you're most likely to run out of address space. Running out of memory, as such, is basically impossible. So just allocate a large buffer, or buffers, on startup, in order to hold all the data you'll ever need, whilst leaving a small amount for the OS. Writing random junk to these regions would probably be a good idea in order to force the OS to actually assign the memory to your process. If your process survives this attempt to use every byte it's asked for, there's some kind of backing now reserved for all of this stuff, so now you're golden.
Write/steal your own memory manager, and direct it to allocate from these buffers. Then use it, consistently, in your app, or take advantage of gcc's --wrap option to forward calls from malloc and friends appropriately. If you use any libraries that can't be directed to call into your memory manager, junk them, because they'll just get in your way. Lack of overridable memory management calls is evidence of deeper-seated issues; you're better of without this particular component. (Note: even if you're using --wrap, trust me, this is still evidence of a problem! Life is too short to use those libraries that don't let you overload their memory management!)
Once you run out of memory, OK, you're screwed, but you've still got that space you left free before, so if freeing up some of the memory you've asked for is too difficult you can (with care) call system calls to write a message to the system log and then terminate, or whatever. Just make sure to avoid calls to the C library, because they'll probably try to allocate some memory when you least expect it -- programmers who work with systems that have virtualised address spaces are notorious for this kind of thing -- and that's the very thing that has caused the problem in the first place.
This approach might sound like a pain in the arse. Well... it is. But it's straightforward, and it's worth putting in a bit of effort for that. I think there's a Kernighan-and/or-Ritche quote about this.
If your application is likely to allocate large blocks of memory and risks hitting the per-process or VM limits, waiting until an allocation actually fails is a difficult situation from which to recover. By the time malloc returns NULL or new throws std::bad_alloc, things may be too far gone to reliably recover. Depending on your recovery strategy, many operations may still require heap allocations themselves, so you have to be extremely careful on which routines you can rely.
Another strategy you may wish to consider is to query the OS and monitor the available memory, proactively managing your allocations. This way you can avoid allocating a large block if you know it is likely to fail, and will thus have a better chance of recovery.
Also, depending on your memory usage patterns, using a custom allocator may give you better results than the standard built-in malloc. For example, certain allocation patterns can actually lead to memory fragmentation over time, so even though you have free memory, the available blocks in the heap arena may not have an available block of the right size. A good example of this is Firefox, which switched to dmalloc and saw a great increase in memory efficiency.
I don't think that capturing the failure of malloc or new will gain you much in your situation. linux allocates large chunks of virtual pages in malloc by means of mmap. By this you may find yourself in a situation where you allocate much more virtual memory than you have (real + swap).
The program then will only fail much later with a segfault (SIGSEGV) when you write to the first page for which there isn't any place in swap. In theory you could test for such situations by writing a signal handler and then dirtying all pages that you allocate.
But usually this will not help much either, since your application will be in a very bad state long before that: constantly swapping, computing mechanically with your harddisk...
It's possible for writes to the syslog to fail in low memory conditions: there's no way to know that for every platform without looking at the source for the relevant functions. They could need dynamic memory to format strings that are passed in, for instance.
Long before you run out of memory, however, you'll start paging stuff to disk. And when that happens, you can forget any performance advantages from caching.
Personally, I'm convinced by the design behind Varnish: the operating system offers services to solve a lot of the relevant problems, and it makes sense to use those services (minor editing):
So what happens with Squid's elaborate memory management is that it gets into fights with the kernel's elaborate memory management ...
Squid creates a HTTP object in RAM and it gets used some times rapidly after creation. Then after some time it get no more hits and the kernel notices this. Then somebody tries to get memory from the kernel for something and the kernel decides to push those unused pages of memory out to swap space and use the (cache-RAM) more sensibly for some data which is actually used by a program. This however, is done without Squid knowing about it. Squid still thinks that these http objects are in RAM, and they will be, the very second it tries to access them, but until then, the RAM is used for something productive. ...
After some time, Squid will also notice that these objects are unused, and it decides to move them to disk so the RAM can be used for more busy data. So Squid goes out, creates a file and then it writes the http objects to the file.
Here we switch to the high-speed camera: Squid calls write(2), the address it gives is a "virtual address" and the kernel has it marked as "not at home". ...
The kernel tries to find a free page, if there are none, it will take a little used page from somewhere, likely another little used Squid object, write it to the paging ... space on the disk (the "swap area") when that write completes, it will read from another place in the paging pool the data it "paged out" into the now unused RAM page, fix up the paging tables, and retry the instruction which failed. ...
So now Squid has the object in a page in RAM and written to the disk two places: one copy in the operating system's paging space and one copy in the filesystem. ...
Here is how Varnish does it:
Varnish allocate some virtual memory, it tells the operating system to back this memory with space from a disk file. When it needs to send the object to a client, it simply refers to that piece of virtual memory and leaves the rest to the kernel.
If/when the kernel decides it needs to use RAM for something else, the page will get written to the backing file and the RAM page reused elsewhere.
When Varnish next time refers to the virtual memory, the operating system will find a RAM page, possibly freeing one, and read the contents in from the backing file.
And that's it. Varnish doesn't really try to control what is cached in RAM and what is not, the kernel has code and hardware support to do a good job at that, and it does a good job.
You may not need to write caching code at all.
As has been stated, exhausting memory means that all bets are off. IMHO the best method of handling this situation is to fail gracefully (as opposed to simply crashing!). Your cache could allocate a reasonable amount of memory on instantiation. The size of this memory would equate to an amount that, when freed, will allow the program to terminate reasonably. When your cache detects that memory is becoming low then it should release this memory and instigate a graceful shutdown.
I'm writing an caching app that consumes large amounts of memory.
Hopefully, I'll manage my memory well enough, but I'm just thinking about what to do if I do run out of memory.
If you are writing deamon which should run 24/7/365, then you should not use dynamic memory management: preallocate all the memory in advance and manage it using some slab allocator/memory pool mechanism. That will also protect you again the heap fragmentation.
If a call to allocate even a simple object fails, is it likely that even a syslog call will also fail?
Should not. This is partially reason why syslog exists as a syscall: that application can report an error independent of its internal state.
If malloc or new returns a NULL or 0L value then it essentially means the call failed and it can't give you the memory for some reason. So, what would be the sensible thing to do in that case?
I generally try in the situations to properly handle the error condition, applying the general error handling rules. If error happens during initialization - terminate with error, probably configuration error. If error happens during request processing - fail the request with out-of-memory error.
For plain heap memory, malloc() returning 0 generally means:
that you have exhausted the heap and unless your application free some memory, further malloc()s wouldn't succeed.
the wrong allocation size: it is quite common coding error to mix signed and unsigned types when calculating block size. If the size ends up mistakenly negative, passed to malloc() where size_t is expected, it becomes very large number.
So in some sense it is also not wrong to abort() to produce the core file which can be analyzed later to see why the malloc() returned 0. Though I prefer to (1) include the attempted allocation size in the error message and (2) try to proceed further. If application would crash due to other memory problem down the road (*), it would produce core file anyway.
(*) From my experience of making software with dynamic memory management resilient to malloc() errors I see that often malloc() returns 0 not reliably. First attempts returning 0 are followed by a successful malloc() returning valid pointer. But first access to the pointed memory would crash the application. This is my experience on both Linux and HP-UX - and I have seen similar pattern on Solaris 10 too. The behavior isn't unique to Linux. To my knowledge the only way to make an application 100% resilient to memory problems is to preallocate all memory in advance. And that is mandatory for mission critical, safety, life support and carrier grade applications - they are not allowed dynamic memory management past initialization phase.
I don't know why many of the sensible answers are voted down. In most server environments, running out of memory means that you have a leak somewhere, and that it makes little sense to 'free some memory and try to go on'. The nature of C++ and especially the standard library is that it requires allocations all the time. If you are lucky, you might be able to free some memory and execute a clean shutdown, or at least emit a warning.
It is however far more likely that you won't be able to do a thing, unless the allocation that failed was a huge one, and there is still memory available for 'normal' things.
Dan Bernstein is one of the very few guys I know that can implement server software that operates in memory constrained situations.
For most of the rest of us, we should probably design our software that it leaves things in a useful state when it bails out because of an out of memory error.
Unless you are some kind of brain surgeon, there isn't a lot else to do.
Also, very often you won't even get an std::bad_alloc or something like that, you'll just get a pointer in return to your malloc/new, and only die when you actually try to touch all of that memory. This can be prevented by turning off overcommit in the operating system, but still.
Don't count on being able to deal with the SIGSEGV when you touch memory that the kernel hoped you wouldn't be.. I'm not quite sure how this works on the windows side of things, but I bet they do overcommit too.
All in all, this is not one of C++'s strong spots.