class: big, middle # ECE 7420 / ENGI 9807: Security .title[ .lecture[Lecture 3:] .title[Buffer overflows] ] --- # Today ### More memory issues: buffer overflows * stack smashing * heap spraying ### Mitigations --- # Buffers ### Useful things! ??? It's been said that most of computing is a matter of transforming things from one representation into another so that we can do computation (and then, likely, to transform those results into another representation!). In order to do that, we often need to store information somewhere... like in a buffer! -- ### Demo: [sum.c](sum.c) * a program that loads and sums some integers * example: [numbers.dat](numbers.dat) * (also see [Makefile](Makefile)) ??? This example program loads integers from a file and adds their values together. Some things to note as we walk through this example: * low-level file I/O functions from the C standard library * "hex" tools (binary viewing / manipulation using hex representations) * endianness --- # Buffer problems -- ### Q: what if we load too much data? For example: * [big.dat](big.dat) * [error.dat](error.dat) ??? If we load too much data into our buffer, we **overwrite adjacent data**. The consequences of this depend on **what the adjacent memory holds**! --- # Buffer overflow ### Without bounds checks... memory corruption! -- ### What is the consequence of this corruption? -- * depends on **how much** we overflow the buffer ??? When we attempt to process [big.dat](big.dat), we see a `SIGSEGV` in the C standard libary's `memcpy` function. This is because we are copying in so much data that we walk right out of **a virtual memory allocation**. We can investigate this with a debugger, examining the **call stack** to see who called what. When we attempt to process [error.dat](error.dat), we still see a `SIGSEGV`, but it's for a much more interesting reason. When we try to investigate _this_ case with a debugger, we can't even see **a call stack**! Why is this? -- * depends **where** the overflowed buffer is ??? Loading 0x10 (i.e., 16) integers from the file [error.dat](error.dat) _overflows_ the eight-integer (i.e., 32 B) buffer that we are loading into. This overwrites whatever comes after the buffer. The significance of that depends on where the buffer we're loading into is located! --- <img src="../call-stack.svg" height="600" align="right"/> # The call stack ### Given what we know about stacks... -- ### If we overflow a local variable... -- ### What happens? -- [error.dat](error.dat) ??? Loading 16 integers from the file [error.dat](error.dat) overwrites whatever comes after the buffer we're loading into. In the case of a local variable stored on the stack, this might be **another local variable**, but it might also be any of the things that live on the stack between functions' local variables. For example, it might be a **return address**! --- # Stack smashing ### Changing return addresses can cause crashes -- ### But can we get even more creative? ??? We can _always_ get more creative. 🙂 -- [malice.dat](malice.dat) (compiled from [sh.s](sh.s): assembly for FreeBSD, Linux and OpenBSD) -- This is known as "shellcode", as it "pops a shell" ??? "Popping a shell" can be a beachhead in an attack on a real system, as an attacker can then execute arbitrary commands. It can also be a demonstration that the attacker _could_ execute arbitrary commands if they wanted to. -- ##### See: ["Smashing The Stack For Fun And Profit"](https://insecure.org/stf/smashstack.html) by "Aleph One" --- # What just happened? -- ### Payload -- * loaded attacker-provided code into memory -- * all ready to be executed by... -- ### Control-flow highjacking -- * in this case, overwriting the return address -- (two birds, one stone) -- * in other cases: other attacks! --- # Prevention ### How can we prevent stack smashing? -- * write perfect software! -- * _memory-safe_ languages -- (**partial** answer) ??? We will talk about memory safety in the next couple of lectures. We will also explore some of the tools that we used in today's lecture in Thursday's lab! --- # Mitigations ### How can we prevent/reduce stack smashing? -- * stack canaries: `-fstack-protector` ??? A _stack canary_, like a [canary in a coal mine](https://www.smithsonianmag.com/smart-news/story-real-canary-coal-mine-180961570/) (fun picture [here](http://history.alberta.ca/energyheritage/coal/the-early-development-of-the-coal-industry-1874-1914/early-methods-and-technology/canaries-in-the-coal-mine.aspx)), is something that can be checked to see if conditions are too dangerous to continue normal operations. In the case of a canary, it would faint from carbon dioxide before humans would, sending a signal that the mine wasn't safe. In the case of a stack, **random values** can be written to the stack in between functions' allocations. Code is inserted to check this "canary" value **when returning from a function** to ensure that **it hasn't been overwritten**. -- * non-executable stacks -- ([we needed `-z execstack` to demo!](Makefile)) -- * `W^X`: memory regions writable **or** executable (limitations?) ??? Marking memory as _non-executable_ is something that wasn't possible on 32-bit x86 computers, but _is_ possible on **64-bit x86_64 computers**. This functionality can be used to prevent executable stacks (always a good idea!) and/or a full `W^X` policy. -- * ASLR: address space layout randomization (more later) -- ### ... and more to follow --- # The attacker strikes back -- ### Guessing precise addresses is hard -- NOP sleds, relative addressing -- ### Shellcode authors avoid zeroes (why?) -- ### Is shellcode easy to spot? -- See: [English shellcode](https://www.cs.jhu.edu/~sam/ccs243-mason.pdf)* .footnote[ * "English Shellcode", Mason, Small, Monrose and MacManus, in _CCS '09: Proceedings of the 16th ACM conference on Computer and communications security_, 2009. DOI: [10.1145/1653662.1653725](https://dx.doi.org/10.1145/1653662.1653725) ] --- # Summary ### Buffer overflows #### Stack smashing #### Heap spraying ### Mitigations ... with more to follow! --- class: big, middle The End.