Presentation is loading. Please wait.

Presentation is loading. Please wait.

Code Reviews James Walden Northern Kentucky University.

Similar presentations


Presentation on theme: "Code Reviews James Walden Northern Kentucky University."— Presentation transcript:

1 Code Reviews James Walden Northern Kentucky University

2 CSC 666: Secure Software Engineering Topics 1.Types of Reviews 2.Code Review Process 3.Checklists 4.Prioritizing Code to Review

3 CSC 666: Secure Software Engineering Code Reviews Inspection of source code by one or more people who aren’t the author of the code.  Goal: Identify defects for later removal.  People: Moderator, reviewers.  Scope: Module or small set of classes.  Time: 1-2 hours; <1kloc

4 CSC 666: Secure Software Engineering Benefits of Code Reviews 1.Find defects sooner in the lifecycle. (IBM finds 82% of defects before testing.) 2.Find defects with less effort than testing. (IBM—rev: 3.5 hrs/defect, test: 15-25 hrs/defect.) 3.Find different defects than testing. (Can identify design and requirements problems too.) 4.Educate developers about security bugs. (Developers frequently make the same mistakes.)

5 CSC 666: Secure Software Engineering PCI DSS 6.3.7.b Obtain and review policies to confirm that all custom application code changes for web applications must be reviewed (using either manual or automated processes) as follows: ƒ Code changes are reviewed by individuals other then the originating code author, and by individuals who are knowledgeable in code revie techniques and secure coding practices. ƒ Code reviews ensure code is developed accordin to secure coding guidelines such as the Open Web Security Project Guide (see PCI DSS Requirement 6.5). ƒ Appropriate corrections are implemented prior to release. ƒ Code review results are reviewed and approved by management prior to release.

6 CSC 666: Secure Software Engineering Inspections  Most formal process.  Thorough coverage with separated roles.  Use checklists to focus on specified goals.  Collect metrics to track defects.  Determine whether further inspections of revised software needed at end of meeting.  Extensive documentation of effectiveness.  One study found 16-20 defects/kloc with inspections compared with 3 defects/kloc in informal walkthrough.

7 CSC 666: Secure Software Engineering Roles in Reviews 1.Moderator Manages meeting; follows up on issues. 2.Reader Paraphrases code during meeting. Not the author. 3.Recorder Records bugs discovered. 4.Author Provides context for code; answers questions. Makes corrections after code review.

8 CSC 666: Secure Software Engineering Walkthroughs  Less formal process.  Author leads meeting and describes code.  Focus on author needs, not quality goals.  No checklists used or metrics gathered.  Quality varies widely.  Walkthrough quality depends solely on author.  Useful for educating developers about code, provide high level view of design and defects.

9 CSC 666: Secure Software Engineering Code Review Process Planning Author Moderator Prep Everyone Meeting Everyone Rework Author Follow-up Author Moderator

10 CSC 666: Secure Software Engineering Planning 1.Author initiates Planning once code ready. 2.A Moderator is assigned to inspection. 3.Author and Moderator assemble inspection pkg. 4.Moderator identifies other participants. Planning Author Moderator Prep Everyone Meeting Everyone Rework Author Follow-up Author Moderator

11 CSC 666: Secure Software Engineering Preparation 1.Reviewers examine inspection package. 2.Reviewers use checklists and analysis tools. 3.Reviewers mark bugs found. Planning Author Moderator Prep Everyone Meeting Everyone Rework Author Follow-up Author Moderator

12 CSC 666: Secure Software Engineering Meeting 1.Reader describes code in own words. 2.Reviewers comment and ask questions. 3.Recorder notes all potential bugs, suggestions. 4.Team appraises code at meeting conclusion. Planning Author Moderator Prep Everyone Meeting Everyone Rework Author Follow-up Author Moderator

13 CSC 666: Secure Software Engineering Rework Author addresses issues recorded at meeting. Planning Author Moderator Prep Everyone Meeting Everyone Rework Author Follow-up Author Moderator

14 CSC 666: Secure Software Engineering Follow-up 1.Moderator meets with Author about rework. 2.Moderator verifies all changes made correctly. 3.Author checks in corrected code. Planning Author Moderator Prep Everyone Meeting Everyone Rework Author Follow-up Author Moderator

15 CSC 666: Secure Software Engineering Formality Spectrum ReviewPlanningPrepMeetingReworkFollowup InspectionYes Team ReviewYes No WalkthroughYesNoYes No Pair Programming YesNoCon- tinuous Yes Peer Deskcheck NoYesPossiblyYesNo Ad Hoc Review No Yes No

16 CSC 666: Secure Software Engineering Code Review Tips 1.Know your limits. Typical review speed is 150-200 lines/hour. Limit meeting length to 1-2 hours. 2.Know what bugs to look for. Checklists Static analysis tools 3.Use tools. Simple tools: grep, findstr Code viewers Static analysis tools 4.Require preparation before the meeting.

17 CSC 666: Secure Software Engineering Checklists Security reviews should include checklists of common problems, including: 1.SQL injection 2.Cross-site scripting 3.Input validation bugs 4.Checking return values 5.Resource name canonicalization 6.Race conditions

18 CSC 666: Secure Software Engineering Code Review Problems 1.Requires substantial expertise in area of programming and security to be effective. 2.Human readers are fallible and will miss mistakes. 3.Code reviews are slow. Unreviewed legacy code will take time to review.

19 CSC 666: Secure Software Engineering Prioritizing Code If you can’t review everything, review  Code that runs with privileged mode.  Code that listens on globally accessible sockets.  Code that is accessible w/o authentication.  Code with a history of vulnerabilities.  Code that handles sensitive data.  Complex code.  Code that changes frequently.

20 CSC 666: Secure Software Engineering Reviewing for SQL Injection

21 CSC 666: Secure Software Engineering Ex: Zune infinite loop on 12/31/08 year = 1980; while (days > 365) { if (IsLeapYear(year)) { if (days > 366) { days -= 366; year += 1; } } else { days -= 365; year += 1; } }

22 CSC 666: Secure Software Engineering Key Points  Roles  Moderator  Reader  Recorder  Author  Process  Planning  Preparation  Meeting  Re-work  Followup

23 CSC 666: Secure Software Engineering References 1.Brian Chess and Jacob West, Secure Programming with Static Analysis, Addison-Wesley, 2007. 2.Michael Howard, “A Process for Performing Security Code Reviews.” IEEE Security & Privacy, July 2006. 3.Eoin Keary et. al., OWASP Code Review Guide 1.1, http://www.owasp.org/index.php/Category:OWASP_Co de_Review_Project, 2008. http://www.owasp.org/index.php/Category:OWASP_Co de_Review_Project 4.Steve McConnell, Code Complete, 2/e, Microsoft Press, 2004. 5.Gary McGraw, Software Security, Addison-Wesley, 2006. 6.PCI Security Standards Council, PCI DSS Requirements and Security Assessment Procedures, v1.2, 2008. 7.Karl Wiegers, Peer Reviews in Software, Addison- Wesley, 2002.

24 Static Analysis James Walden Northern Kentucky University

25 CSC 666: Secure Software Engineering Topics 1.Why Static Analysis? 2.False Positives and Negatives 3.Static Analysis Internals 4.Using the Tools

26 CSC 666: Secure Software Engineering What is Static Analysis? Static = without program execution  Includes everything except testing.  Usually used to refer to compiler type tools. Examples  Static type checking  Vulnerability detection tools  Formal methods

27 CSC 666: Secure Software Engineering Why Static Analysis? 1.Code reviews require substantial expertise in secure programming. 2.Human readers are fallible and will miss mistakes. 3.Code reviews are slow. Unreviewed legacy code will take time to review.

28 CSC 666: Secure Software Engineering Verification Techniques Assurance Cost Penetration Testing Formal Verification Security Testing Code Review Static Analysis

29 CSC 666: Secure Software Engineering False Negatives and Positives False Positives  Tool reports bugs in code that aren’t there.  Complex control or data flow can confuse tools. False Negatives  Tool fails to discover bugs that are there.  Code complexity or lack of rules to check.

30 CSC 666: Secure Software Engineering False Negatives and Positives Mistakes False Negatives False Positives Check Heuristics

31 CSC 666: Secure Software Engineering Static Analyis Approaches 1.Standard compiler warnings and type checking. 2.Lexing source checkers that look for bad names like strcpy() and gets(). 3.Parsing source code checkers. 4.Parsing checkers with annotations. 5.Formal proof based program verification.

32 CSC 666: Secure Software Engineering Static Analysis Internals  Parser  Model Builder  Analysis Engine

33 CSC 666: Secure Software Engineering Parser  Convert programming language to AST.  Must have a parser for each language that tool supports. Abstract Syntax Tree

34 CSC 666: Secure Software Engineering Control Flow Graph nConsec = 0 s1 = getHexChar(a) s2 = getHexChar(b) return nConsec if(a > b)

35 CSC 666: Secure Software Engineering Data Flow with SSA Source Code: if (bytesRead < 8) { tail = (byte) bytesRead; } SSA Form: if (bytesRead 1 < 8) { tail 2 = (byte) bytesRead 1 ; } tail 3 = φ(tail 1, tail 2 );

36 CSC 666: Secure Software Engineering Taint Propagation Track flow of data from source to sink.  Source: where data comes into program.  Sink: function that consumes the data. Vulnerabilities reported if  Data comes from an untrusted source.  Data consumed by a dangerous sink.  No function between source and sink makes the data safe.

37 CSC 666: Secure Software Engineering Tainting SQL Injection Example $link = mysql_connect($DB_HOST, $DB_USERNAME, $DB_PASSWORD) or die ("Couldn't connect: ". mysql_error()); mysql_select_db($DB_DATABASE); $username = $_GET[‘username’]; $password = $_GET[‘password’]; $query = "select count(*) from users where username = '$username' and password = '$password'"; $result = mysql_query($query); Source Sink

38 CSC 666: Secure Software Engineering Local vs. Global Analysis Local Analysis: Analysis of an individual function, a.k.a. intraprocedural analysis. Global Analysis: Follows control and data flow between functions, a.k.a. interprocedural analysis.

39 CSC 666: Secure Software Engineering Rules Security knowledge base for tool.  Identify data sources.  Identify data sinks.  Model behavior of validation functions.  Check for dangerous configurations.  Check control flow (i.e. every lock released.) Customize for process + project  Check coding style is obeyed.  Check for custom functions, standards.

40 CSC 666: Secure Software Engineering Static Analysis Tools Simple search (lexing) tools  Flawfinder  ITS4  RATS Parsing Tools  Coverity SAVE  HP Fortify Source Code Analyzer  IBM Rational AppScan Source Edition  Klocwork Insight  FindBugs  splint

41 CSC 666: Secure Software Engineering Using the Tools Who runs the tools?  Developers  Security team When do you run the tool?  While code is being written (IDE integration)  Before code check-in  After each build  After major milestones What do you do with the results?  Support code review process.  Support security metrics.  Use to decide if project should be released.

42 CSC 666: Secure Software Engineering Code Reviews Review Code Run Tool Review Results Fix Bugs Update Rules

43 CSC 666: Secure Software Engineering Static Analysis Metrics  Vulnerability density (vulns/KLOC)  Vulnerabilities divided by severity  Critical, high, medium, low  Vulnerability types  Injection, XSS, race conditions, etc.  Vulnerability dwell  How long bug remains in code after detection.  Audit coverage  Percentage of code covered by reviews.

44 CSC 666: Secure Software Engineering Evolution of a Single Project

45 Key Points  Static Analysis  Source code needed, but not execution.  Fast, repeatable, objective analysis.  Sources and Sinks  Malicious input enters program via sources.  Exploitation occurs at the sink.  Mistakes  False positives  False negatives CSC 666: Secure Software Engineering

46 References 1.Brian Chess and Jacob West, Secure Programming with Static Analysis, Addison- Wesley, 2007. 2.Eoin Keary et. al., OWASP Code Review Guide 1.1, http://www.owasp.org/index.php/Category:OW ASP_Code_Review_Project, 2008. http://www.owasp.org/index.php/Category:OW ASP_Code_Review_Project 3.Gary McGraw, Software Security, Addison- Wesley, 2006. 4.PCI Security Standards Council, PCI DSS Requirements and Security Assessment Procedures, v1.2, 2008. 5.Karl Wiegers, Peer Reviews in Software, Addison-Wesley, 2002.


Download ppt "Code Reviews James Walden Northern Kentucky University."

Similar presentations


Ads by Google