Subversion Repositories HelenOS-doc

Rev

Rev 180 | Go to most recent revision | Details | Last modification | View Log | RSS feed

Rev Author Line No. Line
179 decky 1
\documentclass{llncs}
2
\usepackage{graphicx}
3
 
4
\title{A Road to a Formally Verified General-Purpose Operating System}
5
\author{Martin D\v{e}ck\'{y}}
6
\institute{Department of Software Engineering\\
7
        Faculty of Mathematics and Physics, Charles University\\
8
        Malostransk\'{e} n\'{a}m\v{e}st\'{i} 25, Prague 1, 118~00, Czech Republic\\
9
        \email{martin.decky@dsrg.mff.cuni.cz}}
10
 
11
\begin{document}
12
    \maketitle
13
 
14
    \begin{abstract}
15
        Methods of formal description and verification represent a viable way for achieving
16
        fundamentally bug-free software. However, in reality only a small subset of the existing operating
17
        systems were ever formally verified, despite the fact that an operating system is a critical part
18
        of almost any other software system. This paper summarizes the challenges involved in formal verification
19
        of operating systems, points out several key design choices which should make the formal verification
20
        easier and presents a work-in-progress and initial experiences with formal verification of HelenOS,
21
        a state-of-the-art microkernel-based operating system, which, however, was not designed specifically
22
        with formal verification in mind, as this is mostly prohibitive due to time and budget constrains.
23
 
24
        The contribution of this paper is the shift of focus from attempts to use a single ``silver-bullet''
25
        formal verification method which would be able to verify everything to a combination of multiple
26
        formalisms and techniques to cover various aspects of the verification process. The formally verified
27
        operating system is the emerging property of the combination.
28
    \end{abstract}
29
 
30
    \section{Introduction}
31
        \label{introduction}
32
        In the context of formal verification of software, it is always important to model the outer
33
        environment with a proper level of abstraction. Weak assumptions on the environment make the formal
34
        verification less feasible, because they yield many degrees of freedom which render
35
        the properties which we want to verify overly complex. But strong assumptions on the environment are likewise
36
        not without a price: Any practical usability of the verification results depends on the question
37
        whether we are really able to create a physical environment which satisfies the assumptions.
38
 
39
        There are some assumptions which are universal for any formal verification method. For the sake of
40
        simplicity let us call them \emph{idealized hardware}. Let us also cover the computational model
41
        and basic mathematical principles by this term.
42
 
43
        What is \emph{idealized hardware}? We usually assume that the software is executed by a ``perfect'' hardware
44
        which works exactly according to its specification. Every divergence between the intended and actual
45
        behavior of software counts as a software bug and never as an issue in some hardware component. We
46
        implicitly assume that the hardware is perfect in the sense of physical realization and all probabilities
47
        of accidental mechanical and electrical failures are expected to be zero.
48
 
49
        It is trivial to see that the \emph{idealized hardware} assumptions can never hold in real world.
50
        Accidental mechanical and electrical failures are inevitable in principle and their probability can
51
        never be a mathematical zero. Does this mean that any formal verification method is ultimately futile
52
        since we are unable to ensure validity of the most elementary assumptions? Probably not. We just need
53
        to be aware about the inherent and fundamental limitations of the formal methods which result from
54
        the shortcomings of the physical world.
55
 
56
        The more precisely we model the software environment of the software under discussion, the more precisely
57
        we can calculate or estimate the resulting impact of the imperfections of the physical world on our software
58
        system. This is the most important motivation for formal reasoning about the correctness of the operating system.
59
 
60
        \medskip
61
 
62
        In more detail, operating systems have a somewhat special position among all software:
63
 
64
        \begin{itemize}
65
            \item \emph{Operating systems are usually designed to run on bare hardware.} This means that except
66
                  the \emph{idealized hardware} assumptions we do not have to take any or almost any extra assumptions
67
                  into account.
68
            \item \emph{Operating systems create the lowest software layer and provide services to essentially
69
                  all other software.} Considering the principle of recursion, the properties of the
70
                  operating systems which we prove or disprove form the assumptions for the upper layers
71
                  of software. Thus the dependability of end-user and enterprise software systems is limited
72
                  by the dependability of the operating system.
73
            \item \emph{Operating systems are non-trivial software on their own.} The way
74
                  they are generally designed and programmed (spanning both the kernel and user mode,
75
                  manipulating execution contexts and concurrency, handling critical hardware-related
76
                  operations) represent significant and interesting challenges for the formal verification
77
                  methods and tools.
78
        \end{itemize}
79
 
80
        Even in the informal understanding, the dependability of an operating system greatly determines the perceived
81
        dependability of the entire software stack. This led to several research initiatives in the recent years
82
        which target the creation of a formally verified operating systems from scratch~\cite{seL4}.
83
 
84
        However, 98~\%\footnote{98~\% of client computers connected to the Internet as of January 2010~\cite{marketshare}.} of
85
        the market share of general-purpose operating systems is taken by Windows, Mac~OS~X and Linux.
86
        These systems were clearly not designed with formal verification in mind from the very beginning.
87
        The situation on the embedded, real-time and special-purpose operating systems market is probably different,
88
        but it is unlikely that the situation in the large domain of desktops and servers is going to change
89
        very rapidly in the near future.
90
 
91
        Therefore we need to face the challenges of applying formal methods on an existing code base in the domain
92
        of general-purpose operating systems. Fortunately, the software design qualities of the general-purpose
93
        operating systems gradually improve over time. We can see then novel approaches in the operating systems
94
        research from the late 1980s and 1990s (microkernel design, user space file system and device drivers, etc.)
95
        to slowly emerge in the originally monolithic operating systems. We can also see better code quality thanks
96
        to improved software engineering (code review, proper escalation management, etc.).
97
 
98
        \medskip
99
 
100
        This paper proposes an approach and presents a work-in-progress case study of formal verification
101
        of an general-purpose research operating system, which was also not created specifically with formal
102
        verification in mind from the very beginning, but it was designed according to state-of-the-art operating systems
103
        principles.
104
 
105
        \medskip
106
 
107
        \noindent\textbf{Structure of the Paper.} In Section \ref{context} we introduce the case study in more detail and explain
108
        why we believe it is relevant. In Section \ref{analysis} we discuss our proposal of the combination of formal methods
109
        and tools. In Section \ref{evaluation} we present the preliminary evaluation of our efforts and estimate the complexity
110
        of the imminent next steps we want to take according to our proposal. Finally, in Section \ref{conclusion}
111
        we present the conclusion of the paper.
112
 
113
    \section{Context}
114
        \label{context}
115
        Two very common schemes of operating system design are \emph{monolithic design} and \emph{microkernel design}.
116
        Without going into much detail of any specific operating system, we can define the monolithic design as
117
        a preference to put numerous aspects of the core operating system functionality into the kernel,
118
        while microkernel design is a preference to keep the kernel small, with just a minimal set of features.
119
 
120
        The features which are missing from the kernel in the microkernel design are implemented in user space, usually
121
        by means of libraries, servers and/or software components.
122
 
123
        \subsection{HelenOS}
124
            \label{helenos}
125
            \emph{HelenOS} is a general-purpose research operating system which is being developed at Charles
126
            University in Prague. The source code is available under the BSD open source license and can be
127
            freely downloaded from the project web site~\cite{helenos}. The authors of the code base are
128
            both from the academia and from the open source community (several contributors are employed
129
            as Solaris kernel developers and many are freelance IT professionals). We consistently strive to support
130
            the research and also the practical motivations for developing the system.
131
 
132
            HelenOS uses a preemptive priority-feedback scheduler, it supports SMP hardware and it is
133
            designed to be highly portable. Currently it runs on 7 distinct hardware architectures, including the
134
            most common IA-32, x86-64 (AMD64), IA-64, SPARC~v9 and PowerPC. It also runs on ARMv7 and MIPS,
135
            but currently only in simulators and not on physical hardware.
136
 
137
            Although HelenOS is still far from being an everyday replacement for Linux or Windows due to the lack
138
            of end-user applications (whose development is extremely time-consuming, but unfortunately of
139
            no scientific value), the essential foundations such as file system support and TCP/IP networking
140
            are already in place.
141
 
142
            HelenOS does not currently target embedded devices (although the ARMv7 port can be easily modified to
143
            run on various embedded boards) and does not implement real-time scheduling and synchronization.
144
            Many development projects such as task snapshoting and migration, generic device driver
145
            framework, support for MMU-less platforms and performance monitoring are currently underway.
146
 
147
            \medskip
148
 
149
            HelenOS has a microkernel multiserver design, but the guiding principles of the HelenOS design are
150
            actually more elaborate:
151
 
152
            \begin{itemize}
153
                \item \emph{(Microkernel principle)} Every functionality of the operating system that does not
154
                      have to be necessary implemented in the kernel should be implemented in user space. This
155
                      implies that subsystems such as the file system, device drivers (except those which are
156
                      essential for the basic kernel functionality), naming and trading services, networking,
157
                      human interface and similar features should be implemented in user space.
158
                \item \emph{(Full-fledged principle)} Features which need to be implemented in kernel should
159
                      be designed with full-fledged algorithms and data structures. In contrast
160
                      to several other microkernel operating systems, where the authors have deliberately chosen
161
                      the most simplistic approach (static memory allocation, na\"{\i}ve algorithms, simple data
162
                      structures), HelenOS microkernel tries to use the most advanced and suitable means available.
163
                      It contains features such as AVL and B+ trees, hashing tables, SLAB memory allocator, multiple
164
                      in-kernel synchronization primitives, fine-grained locking and so on.
165
                \item \emph{(Multiserver principle)} Subsystems in user space should be decomposed with the smallest
166
                      reasonable granularity. Each unit of decomposition should be encapsulated in a separate task.
167
                      The tasks represent software components with isolated address spaces. From the design point of
168
                      view the kernel can be seen as a separate component, too.
169
                \item \emph{(Split of mechanism and policy)} The kernel should only provide low-level mechanisms,
170
                      while the high-level policies which are built upon these mechanisms should be defined in
171
                      user space. This also implies that the policies should be easily replaceable while keeping
172
                      the low-level mechanisms intact.
173
                \item \emph{(Encapsulation principle)} The communication between the tasks/components should be
174
                      implemented only via a set of well-defined interfaces. In the user-to-user case the preferred
175
                      communication mechanism is HelenOS IPC, which provides reasonable mix of abstraction and
176
                      performance (RPC-like primitives combined with implicit memory sharing for large data
177
                      transfers). In case of synchronous user-to-kernel communication the usual syscalls are used.
178
                      HelenOS IPC is used again for asynchronous kernel-to-user communication.
179
                \item \emph{(Portability principle)} The design and implementation should always maintain a high
180
                      level of platform neutrality and portability. Platform-specific code in the kernel, core
181
                      libraries and tasks implementing low-level functionality (e.g. device drivers) should be
182
                      clearly separated from the generic code (either by component decomposition, naming and trading,
183
                      or at least by internal compile-time APIs).
184
            \end{itemize}
185
 
186
            As these design guiding principles suggest, the size of the HelenOS microkernel is considerably larger
187
            compared to ``scrupulous'' microkernel implementations. The average footprint of the kernel ranges from
188
            569~KiB when all logging messages, asserts, symbol resolution and debugging kernel console are compiled
189
            in, down to 198~KiB for a non-debugging production build. We do not believe that the raw size
190
            of the microkernel is a relevant quality criterion per se, without taking the actual feature set
191
            into account.
192
 
193
            \medskip
194
 
195
            The run-time architecture of HelenOS is inherently dynamic. The bindings between the components are
196
            not created at compile-time, but during the bootstrap process and can be modified to a large degree
197
            also during normal operation mode of the system (via human interaction and external events). This
198
            creates particularly interesting challenges for describing the design of the system by many formalisms.
199
 
200
            \medskip
201
 
202
            Yet another set of obstacles for reasoning about the properties of HelenOS lies in the design of
203
            the ubiquitous HelenOS IPC mechanism and the associated threading model. The IPC is inherently
204
            asynchronous with constant message buffers in the kernel and dynamic buffers in user space.
205
            It uses the notions of uni-directional bindings, mandatory pairing of requests and replies,
206
            binding establishment and abolishment handshakes, memory sharing and fast message forwarding.
207
 
208
            For easier management of the asynchronous messages and the possibility to process multiple
209
            messages from different peers without the usual kernel threading overhead, the core user space
210
            library manages the execution flow by so-called \emph{fibrils}. A fibril is a user-space-managed thread with
211
            cooperative scheduling. A different fibril is scheduled every time the current fibril is
212
            about to be blocked while sending out IPC requests (because the kernel buffers of the addressee
213
            are full) or while waiting on an IPC reply. This allows different execution flows within the
214
            same thread to process multiple requests and replies. To safeguard proper sequencing
215
            of IPC messages and provide synchronization, special fibril-aware synchronization primitives
216
            (mutexes, condition variables, etc.) are available.
217
 
218
            Because of the cooperative nature of fibrils, they might cause severe performance under-utilization
219
            in SMP configurations and system-wide bottlenecks. As multicore processors are more and more
220
            common nowadays, that would be a substantial design flaw. Therefore the fibrils can be also freely
221
            (and to some degree even automatically) combined with the usual kernel threads, which provide
222
            preemptive scheduling and true parallelism on SMP machines. Needless to say, this combination is
223
            also a grand challenge for the formal reasoning.
224
 
225
            \medskip
226
 
227
            To sum up, the choice of HelenOS as our case study is based on the fact that it was not designed
228
            in advance with formal verification in mind. This is similar to most general-purpose operating
229
            systems in common use. At the same time, it does not have an obsolete design and is non-trivial.
230
 
231
        \subsection{The C Programming Language}
232
            A large majority of operating systems is coded in the C programming language. HelenOS is no exception
233
            to this. The choice of C in the case of kernel is usually well-motivated -- the language was designed
234
            specifically for implementing system software~\cite{c}. It is reasonably low-level in the sense that it allows
235
            to access the memory and other hardware resources with similar effectiveness as from assembler.
236
            It also requires almost no run-time support and it exports many features of the von Neumann hardware
237
            architecture to the programmer in a very straightforward, but still relatively portable way.
238
 
239
            However, what is the biggest advantage of C in terms of run-time performance is also the biggest weakness
240
            for formal reasoning. The permissive memory access model of C, the lack of any reference safety
241
            enforcement, the weak type system and generally little semantic information in the code -- all these
242
            properties do not allow to make many general assumptions about the code.
243
 
244
            Programming languages which target controlled environments such as Java or C\(\sharp\) are
245
            generally easier for formal reasoning because they provide a well-known set of primitives
246
            and language constructs for object ownership, threading and synchronization. Many non-imperative
247
            programming languages can be even considered to be a form of ``executable specification'' and thus
248
            very suitable for formal reasoning. In C, almost everything is left to the programmer who
249
            is free to set the rules.
250
 
251
            \medskip
252
 
253
            The reasons for using C in the user space of HelenOS (and other established operating systems) is
254
            probably much more questionable. Except for the core libraries and services (such as device drivers),
255
            C might be easily replaced by any high-level and perhaps even non-imperative programming language.
256
            The reasons for using C in this context is mostly historical.
257
 
258
            However, as we have stated in Section \ref{introduction}, the way general-purpose operating systems
259
            are implemented changes only slowly and therefore any propositions which require radical modification
260
            of the existing code base before committing to the formal verification are not realistic.
261
 
262
    \section{Analysis}
263
        \label{analysis}
264
 
265
        \begin{figure}[t]
266
            \begin{center}
267
                \resizebox*{120mm}{!}{\includegraphics{diag}}
268
                \caption{Overview of the formalisms and tools proposed.}
269
                \label{fig:diag}
270
            \end{center}
271
        \end{figure}
272
 
273
        In this section, we analyze the properties we would like to check in a general-purpose
274
        operating system. Each set of properties usually requires a specific verification method,
275
        tool or toolchain.
276
 
277
        Our approach will be mostly bottom-to-top, or, in other words, from the lower levels of abstraction
278
        to the higher levels of abstraction. If the verification fails on a lower level, it usually
279
        does not make much sense to continue with the higher levels of abstraction until the issues
280
        are tackled. A structured overview of the formalisms and tools can be seen on Figure \ref{fig:diag}.
281
 
282
        Please note that the titles of the following sections do not follow any particular established
283
        taxonomy of verification methods. We have simply chosen the names to be descriptive.
284
 
285
        \subsection{C Language Compiler and Continuous Integration Tool}
286
            \label{clang}
287
            The initial levels of abstraction do not go far from the C source code. First, we would certainly like to
288
            know whether our code base is compliant with the programming language specification and passes
289
            only the basic semantic checks (proper number and types of arguments passed to functions, etc.).
290
            It is perhaps not very surprising that these decisions can be make by any plain C compiler.
291
            However, with the current implementation of HelenOS even this is not quite trivial.
292
            Besides the requirement to support 7 hardware platforms, the system's compile-time configuration
293
            can be also affected by approximately 65 configuration options, most of which are booleans, the rest
294
            are enumerated types.
295
 
296
            These configuration options are bound by logical propositions in conjunctive or disjunctive
297
            normal forms and while some options are freely configurable, the value of others gets inferred
298
            by the build system of HelenOS. Thus, the overall number of distinct configurations in which
299
            HelenOS can be compiled is at least one order of magnitude larger than the plain number
300
            of supported hardware platforms.
301
 
302
            Various configuration options affect conditional compilation and linking. The programmers
303
            are used to make sure that the source code compiles and links fine with respect to the
304
            most common and obvious configurations, but the unforeseen interaction of the less common
305
            configuration options might cause linking or even compilation errors.
306
 
307
            \medskip
308
 
309
            A straightforward solution is to generate all distinct configurations, starting from the
310
            open variables and inferring the others. This can be part of the continuous integration
311
            process which would try to compile and link the sources in all distinct configurations.
312
 
313
            If we want to be really pedantic, we should also make sure that we run all the relevant higher
314
            level verification methods on all configurations generated by this step. That would certainly
315
            require to multiply the time required by the verification methods at least by the number
316
            of the distinct configurations. Constraining the set of configurations to just the most
317
            representative ones is perhaps a reasonable compromise.
318
 
319
        \subsection{Regression and Unit Tests}
320
            Although some people would argue whether testing is a formal verification method, we still include
321
            it into the big picture. Running regression and unit tests which are part of HelenOS code base
322
            in the continuous integration process is fairly easy. The only complication lies in the technicalities:
323
            We need to setup an automated network of physical machines and simulators which can run the
324
            appropriate compilation outputs for the specific platforms. We need to be able to reboot
325
            them remotely and distribute the boot images to them. And last but not least, we need to be
326
            able to gather the results from them.
327
 
328
        \subsection{Instrumentation}
329
            Instrumentation tools for detecting memory leaks, performance bottlenecks and soft-deadlocks
330
            are also not usually considered to be formal verification tools. They are also rarely utilized
331
            on regular basis as part of the continuous integration process. But again, it might be helpful
332
            to just mention them in the context of regression and unit tests.
333
 
334
            If some regression or unit tests fail, they sometimes do not give sufficient information to
335
            tell immediately what is the root cause of the issue. In that case running the faulting tests
336
            on manually or automatically instrumented executable code might provide more data and point
337
            more directly to the actual bug.
338
 
339
        \subsection{Verifying C Language Compiler}
340
            C language compilers are traditionally not considered to be formal verification tools. Many people
341
            just say that C compilers are good at generating executable code, but do not care much about the semantics
342
            of the source code (on the other hand, formal verification tools usually do not generate any executable code
343
            at all). However, with recent development in the compiler domain, the old paradigms are shifting.
344
 
345
            As the optimization passes and general maturity of the compilers improve over time,
346
            the compilers try to extract and use more and more semantic information from the source code.
347
            The C language is quite poor on explicit semantic information, but the verifying compilers
348
            try to rely on vendor-specific language extensions and on the fact that some semantic information
349
            can be added to the source code without changing the resulting executable code.
350
 
351
            The checks done by the verifying compilers cannot result in fatal errors in the usual cases (they
352
            are just warnings). Firstly, the compilers still need to successfully compile a well-formed C source
353
            code compliant to some older standard (e.g. C89) even when it is not up with the current quality
354
            expectations. Old legacy source code should still pass the compilation as it did decades ago.
355
 
356
            Secondly, the checks run by the verifying compilers are usually not based on abstract interpretation
357
            or exhaustive traversal of a model state space. They are mostly realized as abstract syntax tree
358
            transformations much in the line with the supporting routines of the compilation process (data
359
            and control flow graph analysis, dead code elimination, register allocation, etc.) and the evaluation
360
            function is basically the matching of antipatterns of common programming bugs.
361
 
362
            The checks are usually conservative. The verifying compilers identify code constructs which are suspicious,
363
            which might arise out of programmer's bad intuition and so on, but even these code snippets cannot be
364
            tagged as definitive bugs (since the programmer can be simply in a position where he/she really wants to
365
            do something very strange, but nevertheless legitimate). It is upon the programmer
366
            to examine the root cause of the compiler warning, tell whether it is really a bug or just a false
367
            positive and fix the issue by either amending some additional semantic information (e.g. adding an
368
            explicit typecast or a vendor-specific language extension) or rewriting the code.
369
 
370
        \subsection{Static Analyzer}
371
            Static analyzers try to go deeper than verifying compilers. Besides detecting common antipatterns of
372
            bugs, they also use techniques such as abstract interpretation to check for more complex properties.
373
 
374
            Most commercial static analyzers come with a predefined set of properties which cannot be easily changed.
375
            They are coupled with the commonly used semantics of the environment and generate domain-specific models
376
            of the software based not only on the syntax of the source code, but also based on the assumptions derived
377
            from the memory access model, allocation and deallocation rules, tracking of references and tracking of
378
            concurrency locks.
379
 
380
            The biggest advantage of static analyzers is that they can be easily included in the development or
381
            continuous integration process as an additional automated step, very similar to the verifying compilers.
382
            No definition of properties is needed and false positives can be relatively easily eliminated by amending
383
            some explicit additional information to the source code within the boundaries of the programming language.
384
 
385
            The authors of static analyzers claim large quantities of bugs detected or prevented~\cite{billion}, but static
386
            analyzers are still relatively limited by the kind of bugs they are designed to detect. They
387
            are usually good at pointing out common issues with security implications (specific types of buffer
388
            and stack overruns, usage of well-known functions in an unsafe way, clear cases of forgotten
389
            deallocation of resources and release of locks, etc.). Unfortunately, many static analyzers
390
            only analyze a single-threaded control flow and are thus unable to detect concurrency issues
391
            such as deadlocks.
392
 
393
        \subsection{Static Verifier}
394
            There is one key difference between a static analyzer and a static verifier: Static verifiers
395
            allow the user to specify one's own properties, in terms of preconditions, postconditions and
396
            invariants in the code. Many static verifiers also target true multithreaded usage patterns
397
            and have the capability to check proper locking order, hand-over-hand locking and even liveliness.
398
 
399
            In the context of an operating system kernel and core libraries two kinds of properties are
400
            common:
401
 
402
            \begin{itemize}
403
                \item \emph{Consistency constrains:} These properties define the correct way how data is supposed
404
                      to be manipulated by some related set of subroutines. Checking for these
405
                      properties ensures that data structures and internal states will not get corrupt due
406
                      to bugs in the functions and methods which are designed to manipulate them.
407
                \item \emph{Interface enforcement:} These properties define the correct semantics by which
408
                      a set of subroutines should be used by the rest of the code. Checking for these properties
409
                      ensures that some API is always used by the rest of the code in a specified way
410
                      and all possible error states are detected and reported.
411
            \end{itemize}
412
 
413
        \subsection{Model Checker}
414
            \label{modelcheck}
415
            On the first sight it does not seem to be reasonable to consider general model checkers as
416
            relevant independent tools for formal verification of an existing operating system. While many
417
            different tools use model checkers as their backends, verifying a complete model of the entire
418
            system created by hand seems to be infeasible both in the sense of time required for the model
419
            creation and resources required by the checker to finish the exhaustive traversal of the model's
420
            address space.
421
 
422
            Nevertheless, model checkers on their own can still serve a good job verifying abstract
423
            properties of key algorithms without dealing with the technical details of the implementation.
424
            Various properties of synchronization algorithms, data structures and communication protocols
425
            can be verified in the most generic conditions by model checkers, answering the
426
            question whether they are designed properly in theory.
427
 
428
            If the implementation of these algorithms and protocols do not behave correctly, we can be sure
429
            that the root cause is in the non-compliance between the design and implementation and not a
430
            fundamental flaw of the design itself.
431
 
432
        \subsection{Behavior Checker}
433
            All previously mentioned verification methods were targeting internal properties of the operating system
434
            components. If we are moving to a higher-level abstraction in order to specify correct interaction of the encapsulated
435
            components in terms of interface compatibility and communication, we can utilize \emph{Behavior Protocols}~\cite{bp}.
436
 
437
            To gain the knowledge about the architecture of the whole operating system in terms of software
438
            component composition and bindings, we can use \emph{Architecture Description Language}~\cite{adl}.
439
            This language has the possibility to capture interface types (with method signatures), primitive
440
            components (in terms of provided and required interfaces), composite components (an architectural
441
            compositions of primitive components) and the bindings between the respective interfaces of the
442
            components.
443
 
444
            Unfortunately, the description is usually static, which is not quite suitable for the dynamic
445
            nature of HelenOS and other operating systems. This limitation can be circumvented by considering a relevant
446
            snapshot of the dynamic run-time architecture. This snapshot fixed in time is equivalent to
447
            a statically defined architecture.
448
 
449
            Behavior Protocol checkers can target three main goals:
450
 
451
            \begin{itemize}
452
                \item \emph{Horizontal compliance:} Also called \emph{compatibility}. The goal is to check
453
                      whether the specifications of components that are bound together are semantically
454
                      compatible. All required interfaces need to be bound to provided interfaces and
455
                      the communication between the components cannot lead to \emph{no activity} (a deadlock)
456
                      or a \emph{bad activity} (a livelock).
457
                \item \emph{Vertical compliance:} Also called \emph{substituability}. The goal is to check whether
458
                      it is possible to replace a set of primitive components that are nested inside a composite
459
                      component by the composite component itself. In other words, this compliance can answer the
460
                      questions whether the architecture description of the system is sound with respect to the hierarchical
461
                      composition of the components.
462
                \item \emph{Compliance between the specification and the implementation:} Using various means
463
                      for generating artificial environments for an isolated component the checker is able to
464
                      partially answer the question whether the implementation of the component is an instantiation
465
                      of the component specification.
466
            \end{itemize}
467
 
468
            Horizontal and vertical compliance checking can be done exhaustively. This is a fundamental property
469
            which allows the reasoning about the dependability of the entire component-based operating system.
470
            Assuming that the lower-level verification methods (described in Sections \ref{clang} to \ref{modelcheck})
471
            prove some specific properties of the primitive components, we can be sure that the composition of
472
            the primitive components into composite components and ultimately into the whole operating system
473
            does not break these properties.
474
 
475
            The feasibility of many lower-level verification methods from Sections \ref{clang} to \ref{modelcheck}
476
            depends largely on the size and complexity of the code under verification. If the entire operating
477
            system is decomposed into primitive components with a reasonable granularity, it is more likely that the
478
            individual primitive components can be verified against a large number of properties. Thanks to the
479
            recursive component composition we can then be sure that these properties also hold for the entire system.
480
            In monolithic operating systems without a clear decomposition we cannot do this inference and we
481
            need to verify the lower-level properties over a much larger code base (e.g. the whole monolithic kernel).
482
 
483
            \medskip
484
 
485
            The compliance between the behavior specification and the actual behavior of the implementation is, unfortunately,
486
            the missing link in the chain. This compliance cannot be easily verified in an exhaustive manner. If there is
487
            a discrepancy between the specified and the actual behavior of the components, we cannot conclude anything about
488
            the properties holding in the entire system.
489
 
490
            However, there is one way how to improve the situation: \emph{code generation}. If we generate implementation
491
            from the specification, the compliance between them is axiomatic. If we are able to generate enough
492
            code from the specification to run into the hand-written ``business code'' where we check for
493
            the lower-level properties, the conclusions about the component composition are going to hold.
494
 
495
        \subsection{Behavior Description Generator}
496
            To conclude our path towards higher abstractions we can utilize tools that can
497
            generate the behavior descriptions from \emph{textual use cases} written in a domain-constrained English.
498
            These generated artifacts can be then compared (e.g. via vertical compliance checking) with the formal
499
            specification. Although the comparison might not provide clean-cut results, it can still be
500
            helpful to confront the more-or-less informal user expectations on the system with the exact formal description.
501
 
502
        \subsection{Summary}
503
            \label{missing}
504
            So far, we have proposed a compact combination of formal methods which start at the level of C programming
505
            language, offer the possibility to check for the presence of various common antipatterns, to check for generic
506
            algorithm-related properties, consistency constrains, interface enforcements and conclude with a framework
507
            to make these properties hold even in the case of a large operating system composed from many
508
            components of compliant behavior.
509
 
510
            We are also able to fill in some of the missing pieces by other software engineering approaches
511
            such as regression and unit testing and instrumentation.
512
 
513
            \medskip
514
 
515
            We have spoken only about the functional properties. In general, we cannot apply the same formalisms
516
            and methods on extra-functional properties (e.g. timing properties, performance properties, etc.).
517
            And although it probably does make a good sense to reason about component composition for the extra-functional
518
            properties, the exact relation might be different compared to the functional properties.
519
 
520
            The extra-functional properties need to be tackled by our future work.
521
 
522
    \section{Evaluation}
523
        \label{evaluation}
524
        This section copies the structure of the previous Section \ref{analysis} and adds HelenOS-specific
525
        evaluation of the the proposed formalisms and tools. As this is still largely a work-in-progress,
526
        in many cases just the initial observations can be made.
527
 
528
        \subsection{Verifying C Language Compiler and Continuous Integration Tool}
529
            The primary C compiler used by HelenOS is \emph{GNU GCC 4.4.3} (all platforms)~\cite{gcc} and \emph{Clang 2.6.0}
530
            (IA-32)~\cite{clang}. We have taken some effort to support also \emph{ICC} and \emph{Sun Studio} C compilers,
531
            but the compatibility with these compilers in not guaranteed.
532
 
533
            The whole code base is compiled with the \texttt{-Wall} and \texttt{-Wextra} compilation options. These options turn on
534
            most of the verification checks of the compilers. The compilers trip on common bug antipatterns such
535
            as implicit typecasting of pointer types, comparison of signed and unsigned integer values (danger
536
            of unchecked overflows), the usage of uninitialized variables, the presence of unused local variables,
537
            NULL-pointer dereferencing (determined by conservative local control flow analysis), functions
538
            with non-void return typed that do not return any value and so on. We treat all compilation warnings
539
            as fatal errors, thus the code base must pass without any warnings.
540
 
541
            We also turn on several more specific and strict checks. These checks helped to discover several
542
            latent bugs in the source code:
543
 
544
            \begin{itemize}
545
                \item \texttt{-Wfloat-equal} Check for exact equality comparison between floating point values. The
546
                      usage of equal comparator on floats is usually misguided due to the inherent computational errors
547
                      of floats.
548
                \item \texttt{-Wcast-align} Check for code which casts pointers to a type with a stricter alignment
549
                      requirement. On many RISC-based platforms this can cause run-time unaligned access exceptions.
550
                \item \texttt{-Wconversion} Check for code where the implicit type conversion (e.g. from float to integer,
551
                      between signed and unsigned integers or between integers with different number of bits) can
552
                      cause the actual value to change.
553
            \end{itemize}
554
 
555
            To enhance the semantic information in the source code, we use GCC-specific language extensions to annotate
556
            some particular kernel and core library routines:
557
 
558
            \begin{itemize}
559
                \item \texttt{\_\_attribute\_\_((noreturn))} Functions marked in this way never finish from the point of view
560
                      of the current sequential execution flow. The most common case are the routines which restore previously saved
561
                      execution context.
562
                \item \texttt{\_\_attribute\_\_((returns\_twice))} Functions marked in this way may return multiple times from
563
                      the point of view of the current sequential execution flow. This is the case of routines which save the current
564
                      execution context (first the function returns as usual, but the function can eventually ``return again''
565
                      when the context is being restored).
566
            \end{itemize}
567
 
568
            The use of these extensions has pointed out to several hard-to-debug bugs on the IA-64 platform.
569
 
570
            \medskip
571
 
572
            The automated continuous integration building system is currently work-in-progress. Thus, we do not
573
            test all possible configurations of HelenOS with each changeset yet. Currently only
574
            a representative set of 14 configurations (at least one for each supported platform) is tested by hand
575
            by the developers before committing any non-trivial changeset.
576
 
577
            From occasional tests of other configurations by hand and the frequency of compilation, linkage and
578
            even run-time problems we conclude that the automated testing of all feasible configurations will
579
            be very beneficial.
580
 
581
        \subsection{Regression and Unit Tests}
582
            As already stated in the previous section, the continuous integration building system has not been finished
583
            yet. Therefore regression and unit tests are executed occasionally by hand, which is time consuming
584
            and prone to human omissions. An automated approach is definitively going to be very helpful.
585
 
586
        \subsection{Instrumentation}
587
            We are in the process of implementing our own code instrumentation framework which is motivated mainly
588
            by the need to support MMU-less architectures in the future. But this framework might be also very helpful
589
            in detecting memory and generic resource leaks. We have not tried \emph{Valgrind}~\cite{valgrind} or any similar existing tool
590
            because of the estimated complexity to adopt it for the usage in HelenOS.
591
 
592
            HelenOS was also scanned by \emph{Coverity}~\cite{coverity} in 2006 when no errors were detected. However, since that
593
            time the code base has not been analyzed by Coverity.
594
 
595
        \subsection{Static Analyzer}
596
            The integration of various static analyzers into the HelenOS continuous integration process is underway.
597
            For the initial evaluation we have used \emph{Stanse}~\cite{stanse} and \emph{Clang Analyzer}~\cite{clanganalyzer}.
598
            Both of them showed to be moderately helpful to point out instances of unreachable dead code, use of language
599
            constructs which have ambiguous semantics in C and one case of possible NULL-pointer dereference.
600
 
601
            The open framework of Clang seems to be very promising for implementing domain-specific checks (and at
602
            the same time it is also a very promising compiler framework). Our mid-term goal is to incorporate some of the features
603
            of Stanse and VCC (see Section \ref{staticverifier2}) into Clang Analyzer.
604
 
605
        \subsection{Static Verifier}
606
            \label{staticverifier2}
607
            We have started to extend the source code of HelenOS kernel with annotations understood
608
            by \emph{Frama-C}~\cite{framac} and \emph{VCC}~\cite{vcc}. Initially we have targeted simple kernel data structures
609
            (doubly-linked circular lists) and basic locking operations. Currently we are evaluating the initial experiences
610
            and we are trying to identify the most suitable methodology, but we expect quite promising results.
611
 
612
            As the VCC is based on the Microsoft C++ Compiler, which does not support many GCC extensions, we have been
613
            faced with the requirement to preprocess the source code to be syntactically accepted by VCC. This turned out
614
            to be feasible.
615
 
616
        \subsection{Model Checker}
617
            We are in the process of creating models of kernel wait queues (basic HelenOS kernel synchronization
618
            primitive) and futexes (basic user space thread synchronization primitive) using \emph{Promela} and
619
            verify several formal properties (deadlock freedom, fairness) in \emph{Spin}~\cite{spin}. As both the Promela language
620
            and the Spin model checker are mature and commonly used tools for such purposes, we expect no major problems
621
            with this approach. Because both synchronization primitives are relatively complex, utilizing a model checker
622
            should provide a much more trustworthy proof of the required properties than ``paper and pencil''.
623
 
624
        \subsection{Behavior Checker}
625
            We have created an architecture description in ADL language similar to SOFA ADL~\cite{adl} for the majority of the
626
            HelenOS components and created the Behavior Protocol specification of these components. The architecture
627
            is a snapshot of the dynamic architecture just after a successful bootstrap.
628
 
629
            Both the architecture and behavior description is readily available as part of the source code repository
630
            of HelenOS, including tools which can preprocess the Behavior Protocols according to the architecture description
631
            and create an output suitable for \emph{bp2promela} checker~\cite{bp}.
632
 
633
            As the resulting complexity of the description is larger than any of the previously published case studies
634
            on Behavior Protocols (compare to~\cite{cocome}), our current work-in-progress is to optimize and fine-tune the bp2promela
635
            checker to process the input.
636
 
637
            \medskip
638
 
639
            We have not started to generate code from the architecture description so far because of time constrains. However,
640
            we believe that this is a very promising way to go.
641
 
642
        \subsection{Behavior Description Generator}
643
            We have not tackled the issue of behavior description generation yet, although tools such as \emph{Procasor}~\cite{procasor} are readily
644
            available. We do not consider it our priority at this time.
645
 
646
    \section{Conclusion}
647
        \label{conclusion}
648
        In this paper we propose a complex combination of various formal verification methods and tools
649
        to achieve the verification of an entire general-purpose operating system. The proposed approach generally follows
650
        a bottom-to-top path, starting with low-level checks using state-of-the-art verifying C language compilers,
651
        following by static analyzers and static verifiers. In specific contexts regression and unit tests,
652
        code instrumentation and model checkers for the sake of verification of key algorithms
653
        are utilized.
654
 
655
        Thanks to the properties of state-of-the-art microkernel multiserver operating
656
        system design (e.g. software component encapsulation and composition), we demonstrate
657
        that it should be feasible to successfully verify larger and more complex operating systems than
658
        in the case of monolithic designs. We use formal component architecture and behavior
659
        description for the closure. The final goal -- a formally verified operating system -- is the
660
        emerging property of the combination of the various methods.
661
 
662
        \medskip
663
 
664
        The contribution of this paper is the shift of focus from attempts to use a single ``silver-bullet''
665
        method for formal verification of an operating system to a combination of multiple methods supported
666
        by a suitable architecture of the operating system.
667
 
668
        We also argue that the approach should be suitable for the mainstream
669
        general-purpose operating systems in the near future, because they are gradually
670
        incorporating more microkernel-based features and fine-grained software components.
671
 
672
        Although the evaluation of the proposed approach on HelenOS is still work-in-progress, the
673
        preliminary results and estimates are promising.
674
 
675
        \medskip
676
 
677
        \noindent\textbf{Acknowledgments.} The author would like to express his gratitude to all contributors of the HelenOS
678
        project. Without their vision and dedication the work on this paper would be almost impossible
679
 
680
        This work was partially supported by the Ministry of Education of the Czech Republic
681
        (grant MSM\-0021620838).
682
 
683
    \begin{thebibliography}{99}
684
        \bibitem{billion}Bessey A., Block K., Chelf B., Chou A., Fulton B., Hallem S., Henri-Gros C., Kamsky A., McPeak S., Engler D.: \emph{A Few Billion Lines of Code Later: Using Static Analysis to Find Bugs in the Real World}, Communications of the ACM, Vol. 53 No. 2, pp 66-75, 2010
685
        \bibitem{bp}Kofron J.: \emph{Checking Software Component Behavior Using Behavior Protocols and Spin}, in proceedings of Applied Computing 2007, Seoul, Korea, pp. 1513-1517, 2007
686
        \bibitem{gcc}GCC, the GNU Compiler Collection, \texttt{http://gcc.gnu.org/}
687
        \bibitem{clang}Clang: a C language family frontend for LLVM, \texttt{http://clang.llvm.org/}
688
        \bibitem{clanganalyzer}Clang Static Analyzer, \texttt{http://clang-analyzer.llvm.org/}
689
        \bibitem{cocome}Bulej L., Bures T., Coupaye T., Decky M., Jezek P., Parizek P., Plasil F., Poch T., Rivierre N., Sery O., Tuma P.: \emph{CoCoME in Fractal}, chapter in The Common Component Modeling Example: Comparing Software Component Models, Springer-Verlag, LNCS 5153, 2008
690
        \bibitem{coverity}Coverity, \texttt{http://scan.coverity.com/}
691
        \bibitem{procasor}Mencl V.: \emph{Deriving Behavior Specifications from Textual Use Cases}, in Proceedings of Workshop on Intelligent Technologies for Software Engineering (WITSE04, Sep 21, 2004, part of ASE 2004), Linz, Austria, pp. 331 - 341, Oesterreichische Computer Gesellschaft, 2004
692
        \bibitem{framac}Frama-C, \texttt{http://frama-c.cea.fr/}
693
        \bibitem{c}Lawlis P. K: \emph{Guidelines for Choosing a Computer Language: Support for the Visionary Organization}, Ada Information Clearinghouse, \texttt{http://archive.adaic.com/docs/reports/lawlis/k.htm}, 1998
694
        \bibitem{helenos}HelenOS Project, \texttt{http://www.helenos.org/}
695
        \bibitem{adl}Oplustil T.: \emph{Inheritance in Architecture Description Languages}, reviewed section of Proceedings of the Week of Doctoral Students 2003 conference (WDS 2003), Charles University, Prague, Czech Republic, 2003, pp. 124 - 131, 2003
696
        \bibitem{marketshare}Operating System Market Share, \texttt{http://marketshare.hitslink.com/report\-.aspx?qprid=8\&qptimeframe=M\&qpsp=132}, retrieved on 2010-02-28
697
        \bibitem{seL4}Klein G., Elphinstone K., Heiser G., Andronick J., Cock D., Derrin P., Elkaduwe D., Engelhardt K., Kolanski R., Norrish M., Sewell T., Tuch H., Winwood S.: \emph{seL4: Formal verification of an OS kernel}, Proceedings of the 22nd ACM Symposium on Operating Systems Principles, Big Sky, MT, USA, 2009
698
        \bibitem{spin}Spin, \texttt{http://spinroot.com/}
699
        \bibitem{stanse}Stanse: Static Analysis Framework for C code, \texttt{http://stanse.fi.muni.cz/}
700
        \bibitem{valgrind}Valgrind, \texttt{http://valgrind.org/}
701
        \bibitem{vcc}VCC, \texttt{http://vcc.codeplex.com/}
702
    \end{thebibliography}
703
\end{document}