DOCSLIB.ORG

Memory Corruption: Why Protection Is Hard

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Memory Corruption: Why Protection Is Hard Memory Corruption: Why Protection is Hard Mathias Payer, Purdue University http://hexhive.github.io 1 Software is unsafe and insecure ● Low-level languages (C/C++) trade type safety and memory safety for performance – Programmer responsible for all checks ● Large set of legacy and new applications written in C / C++ prone to memory bugs ● Too many bugs to find and fix manually – Protect integrity through safe runtime system 2 3 Memory (un-)safety: invalid deref. Dangling pointer: free(foo); (temporal) *foo = 23; Out-of-bounds pointer: char foo[40]; (spatial) foo[42] = 23; Violation iff: pointer is read, written, or freed 4 Type Confusion vtable*? Dptr class B { Bptr b int b; }; c? class D: B { int c; vtable* virtual void d() {} }; B b D … B *Bptr = new B; c D *Dptr = static_cast<D*>B; Dptr->c = 0x43; // Type confusion! Dptr->d(); // Type confusion! 5 Deployed Defenses 6 Status of deployed defenses ● Data Execution Prevention (DEP) Memory 0x4000x4?? RWXR-X ● Address Space Layout Randomization (ASLR) text ● Stack canaries 0x8000x8?? RWXRW- ● Safe exception handlers data 0xfff0xf?? RWXRW- stack 7 Control-Flow Hijack Attack 8 Control-flow hijack attack ● Attacker modifies code pointer 1 – Information leak: target address 2 3 – Memory safety violation: write ● Control-flow leaves valid graph 4 4' – Reuse existing code – Inject/modify code 9 Control-Flow Hijack Attack int vuln(int usr, int usr2){ Memory void *(func_ptr)(); 1 1 int *q = buf + usr; q … func_ptr = &foo; … buf 2 *q = usr2; … 3 (*func_ptr)(); } func_ptr 2 gadgetcode Attack scenario: code reuse ● Find addresses of gadgets ● Force memory corruption to set up attack ● Leverage gadgets for code-reuse attack ● (Fall back to code injection) Code Heap Stack 11 Attack: buffer overflow to ROP void vuln(char *u1) { // assert(strlen(u1) < MAX) Memory safety Violation char tmp[MAX]; strcpy(tmp, u1); ... Integrity *C } vuln(&exploit); Randomization &C tmp[MAX]don't care saveddon't base care pointer Flow Integrity *&C pointsreturn to address&system() ebp1st after argument: system *u1 call Control-flow 1st argumentnext stack to frame system() Attack hijack Model for memory attacks Memory safety Memory corruption Integrity C *C D *D Randomization &C &D Flow Integrity *&C *&D Code Control-flow Data-only Bad things corruption hijack Control-Flow Integrity 14 Control-Flow Integrity (CFI) CHECK(fn); (*fn)(x); CHECK_RET(); Attackerreturn 7 may corrupt memory, code ptrs. verified when used 15 Three directions for CFI ● Goal: minimize target sets, increase precision 16 Source-based CFI ● Kernel protection crucial for system integrity ● Enforce strict pointer propagation rules – Function pointers can only be assigned – Data pointers to function pointers are forbidden ● Enforce types to protect C++ virtual calls * Fine-Grained Control-Flow Integrity for Kernel Software. Xinyang Ge, Nirupama Talele, Mathias Payer, and Trent Jaeger. In EuroS&P'16 * VTrust: Regaining Trust on Your Virtual Calls. Chao Zhang, Scott A. Carr, Tongxin Li, Yu Ding, Chengyu Song, Mathias Payer, and Dawn Song. In NDSS'16 17 Lockdown*: enforce CFI for binaries ● Fine-grained CFI relies on source code ● Coarse-grained CFI is imprecise ● Goal: enforce fine-grained CFI for binaries – Support legacy, binary code and modularity (libraries) – Leverage precise, dynamic analysis – Enforce stack integrity through shadow stack – Low performance overhead * Fine-Grained Control-Flow Integrity through Binary Hardening Mathias Payer, Antonio Barresi, and Thomas R. Gross. In DIMVA'15 18 Dynamic CFI analysis ● Leverage program's modularity through loader /bin/<exec> /lib/libc.so.6 /lib/lib* exported imported exported imported exported imported - puts puts _dl* funcA ifunc* scanf scanf ... funcB ... funcA mprotect ... ... ... .text .text .text call puts puts: funcA: ... ... ... lea fptr, %eax mprotect: funcB: ... ... ... call *%eax ... symbol table of ELF DSO allowed Control Flow transfer .text section of DSO illegal Control Flow transfer 19 Dynamic CFI analysis ● Leverage program's modularity through loader /bin/<exec> /lib/libc.so.6 /lib/lib* exportedModularityimported increasesexported imported precision.exported imported - puts puts _dl* funcA ifunc* scanf scanf ... funcB ... funcA No sourcemprotect needed.... ... ... .textLeverage context.text of transfers..text call puts puts: funcA: ... ... ... lea fptr, %eax mprotect: funcB: ... ... ... call *%eax ... symbol table of ELF DSO allowed Control Flow transfer .text section of DSO illegal Control Flow transfer 20 Necessity of shadow stack* ● Defenses without stack integrity are broken – Loop through two calls to the same function – Choose any caller as return location ● Shadow stack enforces stack integrity – Attacker restricted to arbitrary targets on the stack – Each target can only be called once, in sequence * Control-Flow Bending: On the Effectiveness of Control-Flow Integrity. Nicholas Carlini, Antonio Barresi, Mathias Payer, David Wagner, and Thomas R. Gross. In Usenix SEC'15 21 printf()-oriented programming* ● Translate program to format string – Memory reads: %s – Memory writes: %n – Conditional: %.*d ● Program counter becomes format string counter – Loops? Overwrite the format specific counter ● Turing-complete domain-specific language * Direct fame to Nicholas Carlini, blame to me 22 Ever heard of brainfuck*? ● > == dataptr++ %1$65535d%1$.*1$d%2$hn ● < == dataptr-- %1$.*1$d %2$hn ● + == *dataptr++ %3$.*3$d %4$hhn ● - == *datapr-- %3$255d%3$.*3$d%4$hhn ● . == putchar(*dataptr) %3$.*3$d%5$hn ● , == getchar(dataptr) %13$.*13$d%4$hn ● [ == if (*dataptr == 0) goto ']' %1$.*1$d%10$.*10$d%2$hn ● ] == if (*dataptr != 0) goto '[' %1$.*1$d%10$.*10$d%2$hn * https://en.wikipedia.org/wiki/Brainfuck 23 Exploitable program void loop() { char* last = output; int* rpc = &progn[pc]; while (*rpc != 0) { // fetch -- decode next instruction sprintf(buf, "%1$.*1$d%1$.*1$d%1$.*1$d%1$.*1$d%1$.*1$d%1$.*1$d%1$.*1$d%1$.*1$d%2$hn", *rpc, (short*)(&real_syms)); // execute -- execute instruction sprintf(buf, *real_syms, ((long long int)array)&0xFFFF, &array, // 1, 2 *array, array, output, // 3, 4, 5 ((long long int)output)&0xFFFF, &output, // 6, 7 &cond, &bf_CGOTO_fmt3[0], // 8, 9 rpc[1], &rpc, 0, *input, // 10, 11, 12, 13 ((long long int)input)&0xFFFF, &input // 14, 15 ); // retire -- update PC sprintf(buf, "12345678%.*d%hn", (int)(((long long int)rpc)&0xFFFF), 0, (short*)&rpc); // for debug: do we need to print? if (output != last) { putchar(output[-1]); last = output; } } } 24 Presenting: printbf* ● Turing complete interpreter ● Relies on format strings ● Allows you to execute stuff http://github.com/HexHive/printbf * Direct fame to Nicholas Carlini, blame to me 25 ● Purdue's Capture the Flag (CTF) team – Compete in international hacking competitions – Gain practical security experience – Competitive, fun, challenging tasks – Open, inclusive environment ● Founded 2014, ~15 active, ~100 interested ● 3rd US academic team, top 50 overall 26 Conclusion 27 Conclusion ● ROP/JOP is key to modern exploits – Leak addresses, connect gadgets, inject code ● Control-flow hijack protection – Shadow stack, precise CFI, and locality – High precision is key for effectiveness ● Low overhead, open-source ● Future work – Protect context of control-flow – Protect data and data-flow, not just control 28 Thank you! Questions? Mathias Payer, Purdue University http://hexhive.github.io 29.
Load more

Recommended publications
  • Defeating Memory Corruption Attacks Via Pointer Taintedness Detection
    Defeating Memory Corruption Attacks via Pointer Taintedness Detection Shuo Chen †, Jun Xu ‡, Nithin Nakka †, Zbigniew Kalbarczyk †, Ravishankar K. Iyer † † Center for Reliable and High-Performance Computing, ‡ Department of Computer Science University of Illinois at Urbana-Champaign, North Carolina State University 1308 W. Main Street, Urbana, IL 61801 Raleigh, NC 27695 {shuochen, nakka, kalbar, iyer}@crhc.uiuc.edu [email protected] Abstract formal methods have been adopted to prevent Most malicious attacks compromise system security programmers from writing insecure software. But despite through memory corruption exploits. Recently proposed substantial research and investment, the state of the art is techniques attempt to defeat these attacks by protecting far from perfect, and as a result, security vulnerabilities are program control data. We have constructed a new class of constantly being discovered in the field. The most direct attacks that can compromise network applications without counter-measure against vulnerabilities in the field is tampering with any control data. These non-control data security patching. Patching, however, is reactive in nature attacks represent a new challenge to system security. In and can only be applied to known vulnerabilities. The long this paper, we propose an architectural technique to latency between bug discovery and patching allows defeat both control data and non-control data attacks attackers to compromise many unpatched systems. An based on the notion of pointer taintedness . A pointer is alternative to patching is runtime vulnerability masking said to be tainted if user input can be used as the pointer that can stop ongoing attacks. Compiler and library value. A security attack is detected whenever a tainted interception techniques have been proposed to mask value is dereferenced during program execution.
    [Show full text]
  • Representing and Reasoning About Dynamic Code
    Representing and reasoning about dynamic code Item Type Proceedings Authors Bartels, Jesse; Stephens, Jon; Debray, Saumya Citation Bartels, J., Stephens, J., & Debray, S. (2020, September). Representing and Reasoning about Dynamic Code. In 2020 35th IEEE/ACM International Conference on Automated Software Engineering (ASE) (pp. 312-323). IEEE. DOI 10.1145/3324884.3416542 Publisher ACM Journal Proceedings - 2020 35th IEEE/ACM International Conference on Automated Software Engineering, ASE 2020 Rights © 2020 Association for Computing Machinery. Download date 23/09/2021 23:26:56 Item License http://rightsstatements.org/vocab/InC/1.0/ Version Final accepted manuscript Link to Item http://hdl.handle.net/10150/652285 Representing and Reasoning about Dynamic Code Jesse Bartels Jon Stephens Saumya Debray Department of Computer Science Department of Computer Science Department of Computer Science The University Of Arizona University Of Texas The University Of Arizona Tucson, AZ 85721, USA Austin, TX 78712, USA Tucson, AZ 85721, USA [email protected] [email protected] [email protected] ABSTRACT and trace dependencies back, into and through the JIT compiler’s Dynamic code, i.e., code that is created or modified at runtime, is code, to understand the data and control flows that influenced the ubiquitous in today’s world. The behavior of dynamic code can JIT compiler’s actions and caused the generation of the problem- depend on the logic of the dynamic code generator in subtle and non- atic code. E.g., for the CVE-2017-5121 bug mentioned above, we obvious ways, e.g., JIT compiler bugs can lead to exploitable vul- might want to perform automated analyses to identify which anal- nerabilities in the resulting JIT-compiled code.
    [Show full text]
  • Sok: Make JIT-Spray Great Again
    SoK: Make JIT-Spray Great Again Robert Gawlik and Thorsten Holz Ruhr-Universitat¨ Bochum Abstract Attacks against client-side programs such as browsers were at first tackled with a non-executable stack to pre- Since the end of the 20th century, it has become clear that vent execution of data on the stack and also with a non- web browsers will play a crucial role in accessing Internet executable heap to stop heap sprays of data being later resources such as the World Wide Web. They evolved executed as code. This defense became widely known into complex software suites that are able to process a as W ⊕ X (Writable xor eXecutable) or Data Execution multitude of data formats. Just-In-Time (JIT) compilation Prevention (DEP) to make any data region non-executable was incorporated to speed up the execution of script code, in 2003 [45, 54]. To counter DEP, attackers started to per- but is also used besides web browsers for performance form code reuse such as Return-Oriented Programming reasons. Attackers happily welcomed JIT in their own (ROP) and many variants [10, 11, 32, 68]. In general, if an way, and until today, JIT compilers are an important target adversary knows the location of static code in the address of various attacks. This includes for example JIT-Spray, space of the vulnerable target, she can prepare a fake stack JIT-based code-reuse attacks and JIT-specific flaws to cir- with addresses of these gadgets. As soon as control of cumvent mitigation techniques in order to simplify the the instruction pointer is gained, these gadgets execute exploitation of memory-corruption vulnerabilities.
    [Show full text]
  • Rockjit: Securing Just-In-Time Compilation Using Modular Control-Flow Integrity
    RockJIT: Securing Just-In-Time Compilation Using Modular Control-Flow Integrity Ben Niu Gang Tan Lehigh University Lehigh University 19 Memorial Dr West 19 Memorial Dr West Bethlehem, PA, 18015 Bethlehem, PA, 18015 [email protected] [email protected] ABSTRACT For performance, modern managed language implementations Managed languages such as JavaScript are popular. For perfor- adopt Just-In-Time (JIT) compilation. Instead of performing pure mance, modern implementations of managed languages adopt Just- interpretation, a JIT compiler dynamically compiles programs into In-Time (JIT) compilation. The danger to a JIT compiler is that an native code and performs optimization on the fly based on informa- attacker can often control the input program and use it to trigger a tion collected through runtime profiling. JIT compilation in man- vulnerability in the JIT compiler to launch code injection or JIT aged languages is the key to high performance, which is often the spraying attacks. In this paper, we propose a general approach only metric when comparing JIT engines, as seen in the case of called RockJIT to securing JIT compilers through Control-Flow JavaScript. Hereafter, we use the term JITted code for native code Integrity (CFI). RockJIT builds a fine-grained control-flow graph that is dynamically generated by a JIT compiler, and code heap for from the source code of the JIT compiler and dynamically up- memory pages that hold JITted code. dates the control-flow policy when new code is generated on the fly. In terms of security, JIT brings its own set of challenges. First, a Through evaluation on Google’s V8 JavaScript engine, we demon- JIT compiler is large and usually written in C/C++, which lacks strate that RockJIT can enforce strong security on a JIT compiler, memory safety.
    [Show full text]
  • Rust: Systems Programmers Can Have Nice Things
    Rust: Systems Programmers Can Have Nice Things Arun Thomas [email protected] @arunthomas EuroBSDCon 2019 On Systems Programming [A] systems programmer has seen the terrors of the world and understood the intrinsic horror of existence -James Mickens, The Night Watch 2 What Makes Systems Software Hard? • Stakes are High: Systems software is critical to enforcing security and safety • Kernels, hypervisors, firmware, bootloaders, embedded software, language runtimes, browsers, … • Usually written in C or C++ for performance • BUT C and C++ are not memory-safe • Memory corruption vulnerabilities abound (and are exploited) • See recent Microsoft study (next slide) 3 Memory Unsafety is a Problem https://msrc-blog.microsoft.com/2019/07/16/a-proactive-approach-to-more-secure-code/ Microsoft found ~70% of CVEs in their products each year continue to be memory safety issues (Matt Miller, MSRC @ Blue Hat IL 2019) 4 Microsoft and Rust 5 Intel and Rust 6 Talk Overview “Systems Programmers Can Have Nice Things” -Robert O’Callahan Random Thoughts On Rust • Why Rust? • Rust for Systems Software • Getting Started with Rust on BSD 7 Why ust? 8 I like C. 9 But it turns out programming languages have evolved in the last 50 years. 10 Rust is a safe, fast, productive systems programming language. 11 Mozilla and Rust • Rust was originally created by Mozilla Research • Initial use case: Servo browser engine • Mozilla began shipping Rust components in Firefox 48 in 2016 • Oxidation is Mozilla’s term for “Rusting out” components • Rust code has improved Firefox’s
    [Show full text]
  • Memory Corruption Vulnerabilities, Part II Gang Tan Penn State University Spring 2019
    Memory Corruption Vulnerabilities, Part II Gang Tan Penn State University Spring 2019 CMPSC 447, Software Security Integer Overflow Vulnerabilities * slides adapted from those by Seacord 3 Integer Overflows An integer overflow occurs when an integer is increased beyond its maximum value or decreased beyond its minimum value Standard integer types (signed) signed char, short int, int, long int, long long int Signed overflow vs unsigned overflow An unsigned overflow occurs when the underlying representation can no longer represent an integer value. A signed overflow occurs when a value is carried over to the sign bit 4 Overflow Examples unsigned int ui; signed int si; ui = UINT_MAX; // 4,294,967,295; ui++; ui = 0 printf(“ui = %u\n", ui); si = INT_MAX; // 2,147,483,647 si++; si = -2,147,483,648 printf(“si = %d\n", si); 5 Overflow Examples, cont’d ui = 0; ui‐‐; ui = 4,294,967,295 printf(“ui = %u\n", ui); si = INT_MIN; // ‐2,147,483,648; si‐‐; si = 2,147,483,647 printf(“si = %d\n", si); 6 Integer Overflow Example int main(int argc, char *const *argv) { unsigned short int total; total = strlen(argv[1]) + strlen(argv[2]) + 1; char *buff = (char *) malloc(total); strcpy(buff, argv[1]); strcat(buff, argv[2]); } What if the total variable is overflowed because of the addition operation? 7 Vulnerability: JPEG Example Based on a real‐world vulnerability in the handling of the comment field in JPEG files void getComment(unsigned int len, char *src) { unsigned int size; size is interpreted as a large size = len ‐ 2; positive value of 0xffffffff
    [Show full text]
  • Practical Control-Flow Integrity
    Practical Control-Flow Integrity by Ben Niu Presented to the Graduate and Research Committee of Lehigh University in Candidacy for the Degree of Doctor of Philosophy in Computer Science Lehigh University January, 2016 Copyright © 2015 Ben Niu. All rights reserved. ii Approved and recommended for acceptance as a dissertation in partial fulfillment of the re- quirements for the degree of Doctor of Philosophy. Ben Niu Practical Control-Flow Integrity Date Gang Tan, Dissertation Director Accepted Date Committee Members Gang Tan (Chair) Mooi-Choo Chuah Michael Spear Stephen McCamant iii ACKNOWLEDGEMENTS I am very grateful to my great advisor Prof. Gang Tan. I had no idea how to do research when I came to Lehigh, and he taught me what to think of, how to think, how to present my thoughts, and how to write them down. His door was always open for discussion, he replied emails during weekends, and he provided detailed guidance on any technical problem. I was constantly surprised at his breadth of knowledge and depth of understanding, and I enjoyed working with him. I also thank my thesis committee members, Prof. Mooi-Choo Chuah, Prof. Michael Spear, and Prof. Stephen McCamant for their courses, research works and insightful comments on my research. Michael’s parallel programming course introduced me lock-free algorithms and his operating system course consolidated my system programming skills. Stephen’s research on SFI provided me a foundation in my research, without which I would be still looking for decent solutions. I appreciate my colleagues at MSRC during my internship at Microsoft, especially Matt & Ken, Suha, Vishal, Swamy, Joe, Axel and Marcus.
    [Show full text]
  • An Examination of the Generic Memory Corruption Exploit Prevention
    AN EXAMINATION OF THE GENERIC MEMORY CORRUPTION EXPLOIT PREVENTION MECHANISMS ON APPLE'S LEOPARD OPERATING SYSTEM Haroon Meer SensePost, Rhodes University [email protected] AN EXAMINATION OF THE GENERIC MEMORY CORRUPTION EXPLOIT PREVENTION MECHANISMS ON APPLE'S LEOPARD OPERATING SYSTEM ABSTRACT The Win32 platform has long been the whipping boy of memory corruption attacks and malware, which has forced Microsoft into implementing aggressive anti-exploitation mechanisms into their newer Operating Systems. Apple's Mac OS X (Leopard) has had a much smoother run, both in the media, and in terms of high profile attacks and the reason for this is less clear. In light of Apple's increased market-share, a comparison between Microsoft's defences and Apple's defences is required as the number of anti-exploitation techniques increases with time. In order to produce a side-by-side comparison, an overview of memory corruption attacks is provided and the common generic anti-exploitation techniques for these attacks are enumerated and described. For each operating system, the quality and effective of each implemented defence is evaluated. The results of the study show that Leopard trails Windows Vista in both the number of defences, as well as the quality and effectiveness of the defences that are implemented. KEY WORDS exploit memory corruption stack heap shellcode overflow ret-2-libc 1 INTRODUCTION This paper will cover the basics of memory corruption exploits, and will then examine how Microsoft Windows Vista and Apple MacOS X Leopard combat these attacks in their default state. The intention is to examine how Apple’s Leopard measures up against the automatic exploit mitigations built into Vista.
    [Show full text]
  • An Evil Copy: How the Loader Betrays You
    An Evil Copy: How the Loader Betrays You Xinyang Ge Mathias Payer Trent Jaeger Microsoft Research Purdue University The Pennsylvania State University [email protected] [email protected] [email protected] Abstract—Dynamic loading is a core feature used on current the value that is being written. Despite significant investment in systems to (i) enable modularity and reuse, (ii) reduce memory bug finding techniques, memory corruption is still an important footprint by sharing code pages of libraries and executables problem, as 745 individual CVEs for 2015 and 692 CVEs for among processes, and (iii) simplify update procedures by elim- 2016 are reported. While not all these vulnerabilities allow an inating the need to recompile executables when a library is attacker to compromise a system with arbitrary code execution, updated. The Executable and Linkable Format (ELF) is a generic many do. specification that describes how executable programs are stitched together from object files produced from source code to libraries Without any defense, attackers inject and execute code to and executables. Programming languages allow fine-grained con- trol over variables, including access and memory protections, so take control of a system through memory corruption vulner- programmers may write defense mechanisms assuming that the abilities. Over the past decade, a set of defense mechanisms permissions specified at the source and/or compiler level will hold have been deployed on commodity systems. Data execution at runtime. prevention [5] is a common defense that enforces code in- tegrity. Code integrity prohibits an attacker from injecting new Unfortunately, information about memory protection is lost code into a running process and is usually enforced by hard- during compilation.
    [Show full text]
  • 413 Advanced Debugging and the Address Sanitizer 03 D.Key
    Developer Tools #WWDC15 Advanced Debugging and the Address Sanitizer Finding your undocumented features Session 413 Mike Swingler Xcode UI Infrastructure Anna Zaks LLVM Program Analysis © 2015 Apple Inc. All rights reserved. Redistribution or public display not permitted without written permission from Apple. Overview Overview View Debugger Overview View Debugger Advanced Breakpoint Actions Overview View Debugger Advanced Breakpoint Actions Address Sanitizer Overview View Debugger Advanced Breakpoint Actions Address Sanitizer Demo View Debugger and Advanced Breakpoints Mike Swingler Xcode UI Infrastructure Summary Summary View Debugger • Focus on troublesome views • Visualize your constraints Summary View Debugger • Focus on troublesome views • Visualize your constraints Advanced Breakpoint Actions • Catch exceptions at throw, print message • Print expressions without adding clutter Summary View Debugger • Focus on troublesome views • Visualize your constraints Advanced Breakpoint Actions • Catch exceptions at throw, print message • Print expressions without adding clutter Address Sanitizer Address Sanitizer Anna Zaks LLVM Program Analysis Memory Corruption Memory Corruption Memory Corruption Is Hard to Debug Hard to consistently reproduce The source of error is often far from its manifestation Language Memory Safety Language Memory Safety Less error prone • Swift • Objective-C Automatic Reference Counting Language Memory Safety Less error prone • Swift • Objective-C Automatic Reference Counting More susceptible to memory issues • Direct
    [Show full text]
  • NLP-EYE: Detecting Memory Corruptions Via Semantic-Aware
    NLP-EYE: Detecting Memory Corruptions via Semantic-Aware Memory Operation Function Identification Jianqiang Wang Siqi Ma ( ) Yuanyuan Zhang ( ) [email protected] [email protected] [email protected] Shanghai Jiao Tong University CSIRO DATA61 Shanghai Jiao Tong University Juanru Li ( ) Zheyu Ma [email protected] [email protected] Shanghai Jiao Tong University Northwestern Polytechnical University Long Mai Tiancheng Chen Dawu Gu [email protected] [email protected] [email protected] Shanghai Jiao Tong University Shanghai Jiao Tong University Shanghai Jiao Tong University Abstract overflow and double-free in real-world software and their Memory corruption vulnerabilities are serious threats to soft- number is steadily increasing. For customized memory opera- ware security, which is often triggered by improper use of tion functions, some private memory operation functions are memory operation functions. The detection of memory cor- poorly implemented and thus carry some memory vulnera- ruptions relies on identifying memory operation functions bilities at birth. On the other hand, developers keep making and examining how it manipulates the memory. Distinguish- common mistakes, such as using the memory after it has been ing memory operation functions is challenging because they released (i.e., the use-after-free vulnerability), during the de- usually come in various forms in real-world software. In this velopment process. Both cases aggravate the emerging of paper, we propose NLP-EYE, an NLP-based memory corrup- memory corruption vulnerabilities, which endow the attack- tion detection system. NLP-EYE is able to identify memory ers higher chance of compromising a computer system.
    [Show full text]
  • Memory Corruption Mitigation Via Hardening and Testing
    Memory corruption mitigation via hardening and testing A Dissertation presented by László Szekeres to The Graduate School in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy in Computer Science Stony Brook University May 2017 Copyright by László Szekeres 2017 Stony Brook University The Graduate School László Szekeres We, the dissertation committe for the above candidate for the Doctor of Philosophy degree, hereby recommend acceptance of this dissertation Dr. R. Sekar – Dissertation Advisor Professor, Department of Computer Science Dr. Scott Stoller – Chairperson of Defense Professor, Department of Computer Science Dr. Long Lu – Committee Member Assistant Professor, Department of Computer Science Dr. Franjo Ivančić – External Committee Member Staff Software Engineer, Google Adjunct Assistant Professor, Columbia University This dissertation is accepted by the Graduate School Charles Taber Dean of the Graduate School ii Abstract of the Dissertation Memory corruption mitigation via hardening and testing by László Szekeres Doctor of Philosophy in Computer Science Stony Brook University 2017 Despite decades of research, memory corruption vulnerabilities continue to be at the fore- front of today’s security threats. There are two main avenues for automating defenses against these vulnerabilities: (a) develop automated analysis/testing techniques to find these bugs, or (b) develop software transformations that block their exploitation, or at least make them more difficult (hardening). This dissertation explores both of these avenues. The thesis begins with a general model of memory corruption attacks and a systematic analysis of previous hardening techniques. We show that there none of the existing solutions protect against control-flow hijacks in a both effective and efficient manner.
    [Show full text]