- Timestamp:
- Aug 2, 2021, 9:43:04 AM (3 years ago)
- Branches:
- ADT, ast-experimental, enum, forall-pointer-decay, jacob/cs343-translation, master, new-ast-unique-expr, pthread-emulation, qualifiedEnum
- Children:
- fa7dbf1
- Parents:
- 06c61e2
- Location:
- doc/theses/andrew_beach_MMath
- Files:
-
- 3 added
- 1 deleted
- 1 edited
Legend:
- Unmodified
- Added
- Removed
-
doc/theses/andrew_beach_MMath/intro.tex
r06c61e2 r417e8ea 2 2 3 3 % The highest level overview of Cforall and EHMs. Get this done right away. 4 This thesis goes overthe design and implementation of the exception handling4 This thesis covers the design and implementation of the exception handling 5 5 mechanism (EHM) of 6 6 \CFA (pronounced sea-for-all and may be written Cforall or CFA). 7 \CFA is a new programming language that extends C, thatmaintains7 \CFA is a new programming language that extends C, which maintains 8 8 backwards-compatibility while introducing modern programming features. 9 9 Adding exception handling to \CFA gives it new ways to handle errors and 10 make otherlarge control-flow jumps.10 make large control-flow jumps. 11 11 12 12 % Now take a step back and explain what exceptions are generally. 13 A language's EHM is a combination of language syntax and run-time 14 components that are used to construct, raise, and handle exceptions, 15 including all control flow. 16 Exceptions are an active mechanism for replacing passive error/return codes and return unions (Go and Rust). 13 17 Exception handling provides dynamic inter-function control flow. 14 18 There are two forms of exception handling covered in this thesis: 15 19 termination, which acts as a multi-level return, 16 20 and resumption, which is a dynamic function call. 21 % PAB: Maybe this sentence was suppose to be deleted? 17 22 Termination handling is much more common, 18 to the extent that it is often seen 19 This seperation is uncommon because termination exception handling is so 20 much more common that it is often assumed. 23 to the extent that it is often seen as the only form of handling. 24 % PAB: I like this sentence better than the next sentence. 25 % This separation is uncommon because termination exception handling is so 26 % much more common that it is often assumed. 21 27 % WHY: Mention other forms of continuation and \cite{CommonLisp} here? 22 A language's EHM is the combination of language syntax and run-time 23 components that are used to construct, raise and handle exceptions, 24 including all control flow. 25 26 Termination exception handling allows control to return to any previous 27 function on the stack directly, skipping any functions between it and the 28 current function. 28 29 Exception handling relies on the concept of nested functions to create handlers that deal with exceptions. 29 30 \begin{center} 30 \input{callreturn} 31 \begin{tabular}[t]{ll} 32 \begin{lstlisting}[aboveskip=0pt,belowskip=0pt,language=CFA,{moredelim=**[is][\color{red}]{@}{@}}] 33 void f( void (*hp)() ) { 34 hp(); 35 } 36 void g( void (*hp)() ) { 37 f( hp ); 38 } 39 void h( int @i@, void (*hp)() ) { 40 void @handler@() { // nested 41 printf( "%d\n", @i@ ); 42 } 43 if ( i == 1 ) hp = handler; 44 if ( i > 0 ) h( i - 1, hp ); 45 else g( hp ); 46 } 47 h( 2, 0 ); 48 \end{lstlisting} 49 & 50 \raisebox{-0.5\totalheight}{\input{handler}} 51 \end{tabular} 31 52 \end{center} 32 33 Resumption exception handling seaches the stack for a handler and then calls 34 it without adding or removing any other stack frames. 35 \todo{Add a diagram showing control flow for resumption.} 53 The nested function @handler@ in the second stack frame is explicitly passed to function @f@. 54 When this handler is called in @f@, it uses the parameter @i@ in the second stack frame, which is accessible by an implicit lexical-link pointer. 55 Setting @hp@ in @h@ at different points in the recursion, results in invoking a different handler. 56 Exception handling extends this idea by eliminating explicit handler passing, and instead, performing a stack search for a handler that matches some criteria (conditional dynamic call), and calls the handler at the top of the stack. 57 It is the runtime search $O(N)$ that differentiates an EHM call (raise) from normal dynamic call $O(1)$ via a function or virtual-member pointer. 58 59 Termination exception handling searches the stack for a handler, unwinds the stack to the frame containing the matching handler, and calling the handler at the top of the stack. 60 \begin{center} 61 \input{termination} 62 \end{center} 63 Note, since the handler can reference variables in @h@, @h@ must remain on the stack for the handler call. 64 After the handler returns, control continues after the lexical location of the handler in @h@ (static return)~\cite[p.~108]{Tennent77}. 65 Unwinding allows recover to any previous 66 function on the stack, skipping any functions between it and the 67 function containing the matching handler. 68 69 Resumption exception handling searches the stack for a handler, does \emph{not} unwind the stack to the frame containing the matching handler, and calls the handler at the top of the stack. 70 \begin{center} 71 \input{resumption} 72 \end{center} 73 After the handler returns, control continues after the resume in @f@ (dynamic return). 74 Not unwinding allows fix up of the problem in @f@ by any previous function on the stack, without disrupting the current set of stack frames. 36 75 37 76 Although a powerful feature, exception handling tends to be complex to set up 38 77 and expensive to use 39 so they areoften limited to unusual or ``exceptional" cases.40 The classic example of this is error handling, exceptions can be used to41 remove error handling logic from the main execution path andwhile paying78 so it is often limited to unusual or ``exceptional" cases. 79 The classic example is error handling, where exceptions are used to 80 remove error handling logic from the main execution path, while paying 42 81 most of the cost only when the error actually occurs. 43 82 … … 49 88 some of the underlying tools used to implement and express exception handling 50 89 in other languages are absent in \CFA. 51 Still the resulting syntax resembles that of other languages:52 \begin{ cfa}53 try{90 Still the resulting basic syntax resembles that of other languages: 91 \begin{lstlisting}[language=CFA,{moredelim=**[is][\color{red}]{@}{@}}] 92 @try@ { 54 93 ... 55 94 T * object = malloc(request_size); 56 95 if (!object) { 57 throwOutOfMemory{fixed_allocation, request_size};96 @throw@ OutOfMemory{fixed_allocation, request_size}; 58 97 } 59 98 ... 60 } catch(OutOfMemory * error) {99 } @catch@ (OutOfMemory * error) { 61 100 ... 62 101 } 63 \end{cfa} 64 102 \end{lstlisting} 65 103 % A note that yes, that was a very fast overview. 66 104 The design and implementation of all of \CFA's EHM's features are … … 69 107 70 108 % The current state of the project and what it contributes. 71 All of these features have been implemented in \CFA, along with 72 a suite of test cases as part of this project. 73 The implementation techniques are generally applicable in other programming 109 The majority of the \CFA EHM is implemented in \CFA, except for a small amount of assembler code. 110 In addition, 111 a suite of tests and performance benchmarks were created as part of this project. 112 The \CFA implementation techniques are generally applicable in other programming 74 113 languages and much of the design is as well. 75 Some parts of the EHM use other features unique to \CFA and these would be 76 harder to replicate in other programming languages. 77 114 Some parts of the EHM use features unique to \CFA, and hence, 115 are harder to replicate in other programming languages. 78 116 % Talk about other programming languages. 79 Some existing programming languages that include EHMs/exception handling 80 include C++, Java and Python. All three examples focus on termination 81 exceptions which unwind the stack as part of the 82 Exceptions also can replace return codes and return unions. 117 Three well known programming languages with EHMs, %/exception handling 118 C++, Java and Python are examined in the performance work. However, these languages focus on termination 119 exceptions, so there is no comparison with resumption. 83 120 84 121 The contributions of this work are: 85 122 \begin{enumerate} 86 123 \item Designing \CFA's exception handling mechanism, adapting designs from 87 other programming languages and the creation ofnew features.88 \item Implementing stack unwinding and the EHM in \CFA, including updating89 the compiler and the run-time environment.90 \item Design ed and implementeda prototype virtual system.124 other programming languages, and creating new features. 125 \item Implementing stack unwinding for the \CFA EHM, including updating 126 the \CFA compiler and run-time environment to generate and execute the EHM code. 127 \item Designing and implementing a prototype virtual system. 91 128 % I think the virtual system and per-call site default handlers are the only 92 129 % "new" features, everything else is a matter of implementation. 130 \item Creating tests and performance benchmarks to compare with EHM's in other languages. 93 131 \end{enumerate} 94 132 95 \todo{I can't figure out a good lead-in to the roadmap.}96 The next section covers the existing state of exceptions.97 The existing state of \CFA is also covered in \autoref{c:existing}.98 The newfeatures are introduced in \autoref{c:features},99 which explainstheir usage and design.100 That is followed by the implementation of th ose features in133 %\todo{I can't figure out a good lead-in to the roadmap.} 134 The thesis is organization as follows. 135 The next section and parts of \autoref{c:existing} cover existing EHMs. 136 New \CFA EHM features are introduced in \autoref{c:features}, 137 covering their usage and design. 138 That is followed by the implementation of these features in 101 139 \autoref{c:implement}. 102 The performance results are examined in \autoref{c:performance}.103 Possibilities to extendthis project are discussed in \autoref{c:future}.140 Performance results are presented in \autoref{c:performance}. 141 Summing up and possibilities for extending this project are discussed in \autoref{c:future}. 104 142 105 143 \section{Background} 106 144 \label{s:background} 107 145 108 Exception handling is not a new concept, 109 with papers on the subject dating back 70s.\cite{Goodenough} 110 111 Early exceptions were often treated as signals. They carried no information 112 except their identity. Ada still uses this system. 146 Exception handling is a well examined area in programming languages, 147 with papers on the subject dating back the 70s~\cite{Goodenough75}. 148 Early exceptions were often treated as signals, which carried no information 149 except their identity. Ada~\cite{Ada} still uses this system. 113 150 114 151 The modern flag-ship for termination exceptions is \Cpp, … … 116 153 in 1990. 117 154 % https://en.cppreference.com/w/cpp/language/history 118 \Cpp has the ability to use any value of any type as an exception. 119 However that seems to immediately pushed aside for classes inherited from 155 While many EHMs have special exception types, 156 \Cpp has the ability to use any type as an exception. 157 However, this generality is not particularly useful, and has been pushed aside for classes, with a convention of inheriting from 120 158 \code{C++}{std::exception}. 121 Although there is a special catch-all syntax it does not allow anything to 122 be done with the caught value bec uase nothing is known about it.123 So instead a base type is defined with somecommon functionality (such as124 the ability to describe the reason the exception was raised) and all125 exceptions have th atfunctionality.126 This seems to be the standard now, as the garentied functionality is worth127 any lost flexibility from limiting it to a single type.128 129 Java was the next popular language to use exceptions.130 Its exception system largely reflects that of \Cpp, except that requires131 you throw a childtype of \code{Java}{java.lang.Throwable}159 While \Cpp has a special catch-all syntax @catch(...)@, there is no way to discriminate its exception type, so nothing can 160 be done with the caught value because nothing is known about it. 161 Instead the base exception-type \code{C++}{std::exception} is defined with common functionality (such as 162 the ability to print a message when the exception is raised but not caught) and all 163 exceptions have this functionality. 164 Having a root exception-type seems to be the standard now, as the guaranteed functionality is worth 165 any lost in flexibility from limiting exceptions types to classes. 166 167 Java~\cite{Java} was the next popular language to use exceptions. 168 Its exception system largely reflects that of \Cpp, except it requires 169 exceptions to be a subtype of \code{Java}{java.lang.Throwable} 132 170 and it uses checked exceptions. 133 Checked exceptions are part of the function interface they are raised from. 134 This includes functions they propogate through, until a handler for that 135 type of exception is found. 136 This makes exception information explicit, which can improve clarity and 171 Checked exceptions are part of a function's interface defining all exceptions it or its called functions raise. 172 Using this information, it is possible to statically verify if a handler exists for all raised exception, \ie no uncaught exceptions. 173 Making exception information explicit, improves clarity and 137 174 safety, but can slow down programming. 138 Some of these, such as dealing with high-order methods or an overly specified 139 throws clause, are technical. However some of the issues are much more 140 human, in that writing/updating all the exception signatures can be enough 141 of a burden people will hack the system to avoid them. 142 Including the ``catch-and-ignore" pattern where a catch block is used without 143 anything to repair or recover from the exception. 144 145 %\subsection 146 Resumption exceptions have been much less popular. 147 Although resumption has a history as old as termination's, very few 175 For example, programming complexity increases when dealing with high-order methods or an overly specified 176 throws clause. However some of the issues are more 177 programming annoyances, such as writing/updating many exception signatures after adding or remove calls. 178 Java programmers have developed multiple programming ``hacks'' to circumvent checked exceptions negating the robustness it is suppose to provide. 179 For example, the ``catch-and-ignore" pattern, where the handler is empty because the exception does not appear relevant to the programmer versus 180 repairing or recovering from the exception. 181 182 %\subsection 183 Resumption exceptions are less popular, 184 although resumption is as old as termination; 185 hence, few 148 186 programming languages have implemented them. 149 187 % http://bitsavers.informatik.uni-stuttgart.de/pdf/xerox/parc/techReports/ 150 188 % CSL-79-3_Mesa_Language_Manual_Version_5.0.pdf 151 Mesa is one programming languages that did. Experiance with Mesa152 is quoted as being one of the reasons resumptions were not189 Mesa~\cite{Mesa} is one programming languages that did. Experience with Mesa 190 is quoted as being one of the reasons resumptions are not 153 191 included in the \Cpp standard. 154 192 % https://en.wikipedia.org/wiki/Exception_handling 155 Since then resumptions have been ignored in the main-stream. 156 157 All of this does call into question the use of resumptions, is 158 something largely rejected decades ago worth revisiting now? 159 Yes, even if it was the right call at the time there have been decades 160 of other developments in computer science that have changed the situation 161 since then. 162 Some of these developments, such as in functional programming's resumption 163 equivalent: algebraic effects\cite{Zhang19}, are directly related to 164 resumptions as well. 165 A complete rexamination of resumptions is beyond a single paper, but it is 166 enough to try them again in \CFA. 193 As a result, resumption has ignored in main-stream programming languages. 194 However, ``what goes around comes around'' and resumption is being revisited now (like user-level threading). 195 While rejecting resumption might have been the right decision in the past, there are decades 196 of developments in computer science that have changed the situation. 197 Some of these developments, such as functional programming's resumption 198 equivalent, algebraic effects\cite{Zhang19}, are enjoying significant success. 199 A complete reexamination of resumptions is beyond this thesis, but their re-emergence is 200 enough to try them in \CFA. 167 201 % Especially considering how much easier they are to implement than 168 202 % termination exceptions. 169 203 170 204 %\subsection 171 Functional languages tend to use other solutions for their primary error172 handling mechanism,exception-like constructs still appear.205 Functional languages tend to use other solutions for their primary EHM, 206 but exception-like constructs still appear. 173 207 Termination appears in error construct, which marks the result of an 174 expression as an error , the result of any expression that tries to use it as175 an error, and so on until an approprate handler is reached.176 Resumption appears in algebr ic effects, where a function dispatches its208 expression as an error; thereafter, the result of any expression that tries to use it is also an 209 error, and so on until an appropriate handler is reached. 210 Resumption appears in algebraic effects, where a function dispatches its 177 211 side-effects to its caller for handling. 178 212 179 213 %\subsection 180 More recently exceptions seem to be vanishing from newer programming 181 languages, replaced by``panic".182 In Rust a panic is just a program level abort that may be implemented by214 Some programming languages have moved to a restricted kind of EHM 215 called ``panic". 216 In Rust~\cite{Rust}, a panic is just a program level abort that may be implemented by 183 217 unwinding the stack like in termination exception handling. 184 218 % https://doc.rust-lang.org/std/panic/fn.catch_unwind.html 185 Go's panic through is very similar to a terminationexcept it only supports219 In Go~\cite{Go}, a panic is very similar to a termination, except it only supports 186 220 a catch-all by calling \code{Go}{recover()}, simplifying the interface at 187 the cost of flex ability.188 189 %\subsection 190 Exception handling's most common use cases are in error handling. 191 Here are some other ways to handle errors and comparisons withexceptions.221 the cost of flexibility. 222 223 %\subsection 224 While exception handling's most common use cases are in error handling, 225 here are other ways to handle errors with comparisons to exceptions. 192 226 \begin{itemize} 193 227 \item\emph{Error Codes}: 194 This pattern uses an enumeration (or just a set of fixed values) to indicate 195 that an error has occured and which error it was. 196 197 There are some issues if a function wants to return an error code and another 198 value. The main issue is that it can be easy to forget checking the error 199 code, which can lead to an error being quitely and implicitly ignored. 200 Some new languages have tools that raise warnings if the return value is 201 discarded to avoid this. 202 It also puts more code on the main execution path. 228 This pattern has a function return an enumeration (or just a set of fixed values) to indicate 229 if an error occurred and possibly which error it was. 230 231 Error codes mix exceptional and normal values, artificially enlarging the type and/or value range. 232 Some languages address this issue by returning multiple values or a tuple, separating the error code from the function result. 233 However, the main issue with error codes is forgetting to checking them, 234 which leads to an error being quietly and implicitly ignored. 235 Some new languages have tools that issue warnings, if the error code is 236 discarded to avoid this problem. 237 Checking error codes also results in bloating the main execution path, especially if an error is not dealt with locally and has to be cascaded down the call stack to a higher-level function.. 238 203 239 \item\emph{Special Return with Global Store}: 204 A function that encounters an error returns some value indicating that it 205 encountered a value but store which error occured in a fixed global location. 206 207 Perhaps the C standard @errno@ is the most famous example of this, 208 where some standard library functions will return some non-value (often a 209 NULL pointer) and set @errno@. 210 211 This avoids the multiple results issue encountered with straight error codes 212 but otherwise many of the same advantages and disadvantages. 213 It does however introduce one other major disadvantage: 214 Everything that uses that global location must agree on all possible errors. 240 Some functions only return a boolean indicating success or failure 241 and store the exact reason for the error in a fixed global location. 242 For example, many C routines return non-zero or -1, indicating success or failure, 243 and write error details into the C standard variable @errno@. 244 245 This approach avoids the multiple results issue encountered with straight error codes 246 but otherwise has many (if not more) of the disadvantages. 247 For example, everything that uses the global location must agree on all possible errors and global variable are unsafe with concurrency. 248 215 249 \item\emph{Return Union}: 216 Replaces error codes with a tagged union.250 This pattern replaces error codes with a tagged union. 217 251 Success is one tag and the errors are another. 218 252 It is also possible to make each possible error its own tag and carry its own … … 220 254 so that one type can be used everywhere in error handling code. 221 255 222 This pattern is very popular in functional or semi-functional language,223 anything withprimitive support for tagged unions (or algebraic data types).256 This pattern is very popular in functional or any semi-functional language with 257 primitive support for tagged unions (or algebraic data types). 224 258 % We need listing Rust/rust to format code snipits from it. 225 259 % Rust's \code{rust}{Result<T, E>} 226 227 The main disadvantage is again it puts code on the main execution path.228 This is also the first technique that allows for more information about an 229 error, other than one of a fix-set of ids, to be sent.230 The y can be missed but some languages can force that they are checked.231 It is also implicitly forced in any languages with checked union access. 260 The main advantage is providing for more information about an 261 error, other than one of a fix-set of ids. 262 While some languages use checked union access to force error-code checking, 263 it is still possible to bypass the checking. 264 The main disadvantage is again significant error code on the main execution path and cascading through called functions. 265 232 266 \item\emph{Handler Functions}: 233 On error the function that produced the error calls another function to 267 This pattern implicitly associates functions with errors. 268 On error, the function that produced the error implicitly calls another function to 234 269 handle it. 235 270 The handler function can be provided locally (passed in as an argument, 236 271 either directly as as a field of a structure/object) or globally (a global 237 272 variable). 238 239 C++ uses this as its fallback system if exception handling fails. 273 C++ uses this approach as its fallback system if exception handling fails, \eg 240 274 \snake{std::terminate_handler} and for a time \snake{std::unexpected_handler} 241 275 242 Handler functions work a lot like resumption exceptions .243 The difference is they are more expencive to set up but cheaper to use, and244 so are more suited to more fequent errors.245 The exception being global handlers if they are rarely change as the time246 in both cases strinks towards zero.276 Handler functions work a lot like resumption exceptions, without the dynamic handler search. 277 Therefore, setting setting up the handler can be more complex/expensive, especially if the handle must be passed through multiple function calls, but cheaper to call $O(1)$, and hence, 278 are more suited to frequent exceptional situations. 279 % The exception being global handlers if they are rarely change as the time 280 % in both cases shrinks towards zero. 247 281 \end{itemize} 248 282 249 283 %\subsection 250 Because of their cost exceptions are rarely used for hot paths of execution. 251 There is an element of self-fulfilling prophocy here as implementation 252 techniques have been designed to make exceptions cheap to set-up at the cost 253 of making them expencive to use. 254 Still, use of exceptions for other tasks is more common in higher-level 255 scripting languages. 256 An iconic example is Python's StopIteration exception which is thrown by 257 an iterator to indicate that it is exausted. Combined with Python's heavy 258 use of the iterator based for-loop. 284 Because of their cost, exceptions are rarely used for hot paths of execution. 285 Therefore, there is an element of self-fulfilling prophecy for implementation 286 techniques to make exceptions cheap to set-up at the cost 287 of expensive usage. 288 This cost differential is less important in higher-level scripting languages, where use of exceptions for other tasks is more common. 289 An iconic example is Python's @StopIteration@ exception that is thrown by 290 an iterator to indicate that it is exhausted, especially when combined with Python's heavy 291 use of the iterator-based for-loop. 259 292 % https://docs.python.org/3/library/exceptions.html#StopIteration
Note: See TracChangeset
for help on using the changeset viewer.