lab_memory Malevolent Memories
- Sunday, February 4 at 11:59 PM 2/4 @ 11:59 PM
Assignment Description
In this lab, you will learn about two memory checking utilities: Valgrind and AddressSanitizer (aka ASAN).
The first utility you will learn about is Valgrind. Valgrind will help you detect memory errors and practice implementing the big three. Valgrind is an extremely useful tool for debugging memory problems and for fixing segfaults or other crashes. This lab is also particularly important because we will be checking for memory errors and leaks on your assignments. You will lose points for memory leaks and/or memory errors (we will also teach you the difference between a memory leak and a memory error). You should check your code with Valgrind/ASAN before handing it in. You should also be aware that Valgrind/ASAN will only detect a leak if your program allocates memory and then fails to deallocate it. It cannot find a leak unless the code containing the leak is executed when the program runs. Thus, you should be sure to test your code thoroughly and check these tests with Valgrind/ASAN.
Valgrind and ASAN do not work well or at all on later versions of macOS! In particular, we have noticed ASAN never reporting errors or errors, and Valgrind always reporting memory leaks.
We strongly recommend working on this assignment on an EWS system
Background on Valgrind
Valgrind is a useful tool to detect memory errors and memory leaks.
Valgrind is a free utility for memory debugging, memory leak detection, and
profiling. It runs only on Linux systems. To prepare your project to be
examined by Valgrind you need to compile and to link it with the debug options
-g
and -O0
. Make sure your Makefile
is using these options when
compiling. In order to test your program with Valgrind you should use the
following command:
valgrind ./yourprogram
To instruct valgrind to also check for memory leaks, run:
valgrind --leak-check=full ./yourprogram
You will see a report about all found mistakes or inconsistencies. Each row of the report starts with the process ID (the process ID is a number assigned by the operating system to a running program). Each error has a description, a stack trace (showing where the error occurred), and other data about the error. It is important to eliminate errors in the order that they occur during execution, since a single error early could cause others later on.
Here is a list of some of the errors that Valgrind can detect and report. (Note that not all of these errors are present in the exercise code.)
Invalid read/write errors. This error will happen when your code reads or writes to a memory address which you did not allocate. Sometimes this error occurs when an array is indexed beyond its boundary, which is referred to as an “overrun” error. Unfortunately, Valgrind is unable to check for locally-allocated arrays (i.e., those that are on the stack.) Overrun checking is only performed for dynamic memory.
Exampleint * arr = new int[6]; arr[10] = -5;
Use of an uninitialized value. This type of error will occur when your code uses a declared variable before any kind of explicit assignment is made to the variable.
Exampleint x; cout << x << endl;
Invalid free error. This occurs when your code attempts to delete allocated memory twice, or delete memory that was not allocated with
new
.Exampleint * x = new int; delete x; delete x;
Mismatched
free() / delete / delete []
. Valgrind keeps track of the method your code uses when allocating memory. If it is deallocated with different method, you will be notified about the error.Exampleint * x = new int[6]; delete x;
Memory leak detection. Valgrind can detect three sources of memory leakage.
- A still reachable block happens when you forget to delete an object,
the pointer to the object still exists, and the memory for object is still
allocated.
Example
int * x = new int[6]; // no corresponding delete x
- A lost block is a little tricky. A pointer to some part of the block of memory still exists, but it is not clear whether it is pointing to the block or is independent of it.
- A definitely lost block occurs when the block is not deleted but no pointer to it is found.
- A still reachable block happens when you forget to delete an object,
the pointer to the object still exists, and the memory for object is still
allocated.
More information about the Valgrind utility can be found at the following links:
- http://www.valgrind.org/docs/manual/quick-start.html
- http://www.valgrind.org/docs/manual/faq.html#faq.reports
- http://www.valgrind.org/docs/manual/manual.html
Fixing Memory Bugs (using Valgrind)
Before fixing the bugs, you’ll need to compile the code:
make
This will create an executable file called allocate, which you can run with:
./allocate
You can then run Valgrind on allocate:
valgrind ./allocate
This works fine for fixing the memory errors, however, to fix the memory leaks,
you’ll need to add --leak-check=full
before ./allocate
:
valgrind --leak-check=full ./allocate
Once you have fixed all the Valgrind errors, you can test your program output using:
./allocate > output.txt
diff output.txt soln_output.txt
Note that most of the work in this lab consists of fixing Valgrind’s errors and memory leaks, rather than the program’s output, which should be correct once the memory errors are fixed.
Background on ASAN
ASAN is another utility used to check for memory-related errors. Unlike Valgrind, which interprets code as it is run, ASAN is built into the executable at compile time. For this reason, ASAN is sometimes faster than Valgrind. However, ASAN does not handle recursion as well as Valgrind.
ASAN checks for a wide variety of things. Here are some things that may pop up when debugging:
Out-of-bounds access to heap, stack, and globals. This error occurs when you allocate some memory and then try to access a region outside your allocated space.
Exampleint * arr = new int[100]; return arr[100];
Use-after-free (dangling pointer reference). This error occurs when you try to use memory you have already freed. It is especially helpful when you have several pointers referring to the same memory.
Exampleint * arr = new int[100]; delete [] arr; return arr[5]; //NOPE
Heap buffer overflow. This error occurs when you go out of bounds within an array created on the heap.
Exampleint * arr = new int[100]; arr[0] = 0; int ret = arr[5 + 100]; //NOPE delete [] arr; return ret;
Stack buffer overflow. This error occurs when you go out of bounds within an array created on the stack.
Exampleint arr[100]; arr[1] = 0; arr[5 + 100]; //NOPE
Use after return. This error occurs when you return a stack variable at the end of a function and try to use it after it’s out of scope.
Exampleint * arr1; void func() { int arr2[100]; arr1 = arr2; } int main() { func(); return arr1[10]; //NOPE }
Double free or Invalid free. Double free occus when you free an heap object twice. Invalid free’s occur when you free a non-heap object.
Exampleint * arr = new int[100]; delete [] arr; delete [] arr; //NOPE
int arr[100]; delete [] arr; //NOPE
Memory leak detection. ASAN can detect three sources of memory leakage.
- A still reachable block happens when you forget to delete an object, the pointer to the object still exists, and the memory for object is still allocated.
- A lost block is a little tricky. A pointer to some part of the block of memory still exists, but it is not clear whether it is pointing to the block or is independent of it.
- A definitely lost block occurs when the block is not deleted but no pointer to it is found.
Fixing Memory Bugs (using ASAN)
NOTE: You cannot used Valgrind and ASAN at the same time. Make sure to run it only on the non-ASAN executable!
Before fixing the bugs, you’ll need to compile the code. ASAN terminates the program upon the first invalid memory access. So, if you have an invalid memory access, you’ll have to fix it before moving on to other errors. This incremental procedure should help you step through memory bugs one at a time.
To compile the code for the lab, run
make
(In this and future assignments, make
will produce 2 versions of each
executable: the “normal” version and a version that has ASAN enabled.)
This will create two executable files named allocate
and allocate-asan
. Run
the ASAN version with
./allocate-asan
which will check for both memory errors and leaks.
Once you have fixed all the memory errors, you can test your program output using:
./allocate > output.txt
diff output.txt soln_output.txt
Note that most of the work in this lab consists of fixing ASAN’s errors and memory leaks, rather than the program output, which should be correct already once the memory errors are fixed.
Further Testing
To test the code further, you may use the provided test program to see whether your program runs fine without any memory leaks. We expect that your code produces no memory leaks for both the allocate program and the test program. To compile the test program, please run the following:
make test
You can test whether the test program successful works without any memory issues by running
valgrind ./test
If you notice no errors or memory leak errors outputted from Valgrind, your test program has successfully run as well. If not, it’s time to Valgrind your way through the code once more :)
To get more specific issues about memory leak releated issues with Valgrind, you may get more information by using:
valgrind --leak-check=full ./test
GDB: A Debugger
While ASAN and Valgrind tell you what went wrong in your program after it has been executed, GDB is a debugging tool that allows you to see what is going on `inside’ your program WHILE it executes! It can also be used after your program has crashed for debugging purposes as well. In particular, Valgrind and ASAN work well for memory related errors, but GDB can handle crashes from those as well as other classes of bugs, such as logic errors. To learn how to use GDB for your lab and mps, visit this page:
Checking Out the Code
After reading this lab specification, the first task is to check out the provided code from the class repository.
To check out your files for the third lab, run the following command in your
cs225git
directory:
git fetch release
git merge release/lab_memory -m "Merging initial lab_memory files"
This should update your directory to contain a new directory called
lab_memory
.
Code Description
For this lab, you will be fixing bugs in course staff’s Student-To-Room allocation program. Since many CS classes are very large (CS 225 has nearly 800 students!), exams are usually spread across several rooms. Before each exam, course staff have to allocate different students to different rooms, so that everyone can take the test with enough space.
For example, if there were only two classrooms of equal size, students in the first half of the alphabet (last name letters A - N) might go to Siebel 1404, while students in the second half of the alphabet (letters M - Z) might go to DCL 1320.
However, with more rooms, this problem becomes more difficult. In the sample situation provided, there are 9 classrooms for the exam, varying in seating capacity from 43 to 70 seats (i.e. 21 to 35 students seated every-other-desk). Although we’ll have to break up the alphabet more, we’d still like to assign students with the same first letter of their last name to the same room, as this makes going to the right room easier.
We’ve provided you the code to solve this problem, however, it has several
memory bugs in it. You’ll have to use Valgrind and/or ASAN, as well as some debugging skills
from lab_debug, to find the bugs and fix
them. Note, there are no bugs in the fileio
namespace.
A reference for the lab is provided for you in Doxygen form.
Committing the Code
Grading Information
The following files are used to grade this assignment:
allocator.cpp
allocator.h
letter.cpp
letter.h
room.cpp
room.h
All other files, including main.cpp
and any testing files you have added
will not be used for grading.