Ignore:
Timestamp:
Oct 12, 2017, 3:15:19 PM (7 years ago)
Author:
Thierry Delisle <tdelisle@…>
Branches:
ADT, aaron-thesis, arm-eh, ast-experimental, cleanup-dtors, deferred_resn, demangler, enum, forall-pointer-decay, jacob/cs343-translation, jenkins-sandbox, master, new-ast, new-ast-unique-expr, new-env, no_list, persistent-indexer, pthread-emulation, qualifiedEnum, resolv-new, with_gc
Children:
0aaac0e
Parents:
b7778c1
Message:

Updated concurrency draft and added new section for implementation.

Location:
doc/proposals/concurrency/text
Files:
5 edited

Legend:

Unmodified
Added
Removed

  • doc/proposals/concurrency/text/basics.tex

    rb7778c1 r3364962  
    2222\begin{figure}
    2323\label{fig:fibonacci-c}
    24 \caption{Different implementations of a fibonacci sequence generator in C.}
    2524\begin{center}
    2625\begin{tabular}{c @{\hskip 0.025in}|@{\hskip 0.025in} c @{\hskip 0.025in}|@{\hskip 0.025in} c}
    … …  
    9190\end{tabular}
    9291\end{center}
     92\caption{Different implementations of a fibonacci sequence generator in C.}
    9393\end{figure}
    9494
    … …  
    9898\begin{figure}
    9999\label{fig:fibonacci-cfa}
    100 \caption{Implementation of fibonacci using coroutines}
    101100\begin{cfacode}
    102101coroutine Fibonacci {
    … …  
    140139}
    141140\end{cfacode}
     141\caption{Implementation of fibonacci using coroutines}
    142142\end{figure}
    143143
    … …  
    199199\begin{figure}
    200200\label{fig:fmt-line}
    201 \caption{Formatting text into lines of 5 blocks of 4 characters.}
    202201\begin{cfacode}[tabsize=3]
    203202//format characters into blocks of 4 and groups of 5 blocks per line
    … …  
    244243}
    245244\end{cfacode}
     245\caption{Formatting text into lines of 5 blocks of 4 characters.}
    246246\end{figure}
    247247

  • doc/proposals/concurrency/text/concurrency.tex

    rb7778c1 r3364962  
    44% ======================================================================
    55% ======================================================================
    6 Several tool can be used to solve concurrency challenges. Since many of these challenges appear with the use of mutable shared-state, some languages and libraries simply disallow mutable shared-state (Erlang~\cite{Erlang}, Haskell~\cite{Haskell}, Akka (Scala)~\cite{Akka}). In these paradigms, interaction among concurrent objects relies on message passing~\cite{Thoth,Harmony,V-Kernel} or other paradigms that closely relate to networking concepts (channels\cit for example). However, in languages that use routine calls as their core abstraction-mechanism, these approaches force a clear distinction between concurrent and non-concurrent paradigms (i.e., message passing versus routine call). This distinction in turn means that, in order to be effective, programmers need to learn two sets of designs patterns. While this distinction can be hidden away in library code, effective use of the librairy still has to take both paradigms into account.
     6Several tool can be used to solve concurrency challenges. Since many of these challenges appear with the use of mutable shared-state, some languages and libraries simply disallow mutable shared-state (Erlang~\cite{Erlang}, Haskell~\cite{Haskell}, Akka (Scala)~\cite{Akka}). In these paradigms, interaction among concurrent objects relies on message passing~\cite{Thoth,Harmony,V-Kernel} or other paradigms closely relate to networking concepts (channels\cit for example). However, in languages that use routine calls as their core abstraction-mechanism, these approaches force a clear distinction between concurrent and non-concurrent paradigms (i.e., message passing versus routine call). This distinction in turn means that, in order to be effective, programmers need to learn two sets of designs patterns. While this distinction can be hidden away in library code, effective use of the librairy still has to take both paradigms into account.
    77
    88Approaches based on shared memory are more closely related to non-concurrent paradigms since they often rely on basic constructs like routine calls and shared objects. At the lowest level, concurrent paradigms are implemented as atomic operations and locks. Many such mechanisms have been proposed, including semaphores~\cite{Dijkstra68b} and path expressions~\cite{Campbell74}. However, for productivity reasons it is desireable to have a higher-level construct be the core concurrency paradigm~\cite{HPP:Study}.
    99
    10 An approach that is worth mentionning because it is gaining in popularity is transactionnal memory~\cite{Dice10}[Check citation]. While this approach is even pursued by system languages like \CC\cit, the performance and feature set is currently too restrictive to be the main concurrency paradigm for general purpose language, which is why it was rejected as the core paradigm for concurrency in \CFA.
    11 
    12 One of the most natural, elegant, and efficient mechanisms for synchronization and communication, especially for shared memory systems, is the \emph{monitor}. Monitors were first proposed by Brinch Hansen~\cite{Hansen73} and later described and extended by C.A.R.~Hoare~\cite{Hoare74}. Many programming languages---e.g., Concurrent Pascal~\cite{ConcurrentPascal}, Mesa~\cite{Mesa}, Modula~\cite{Modula-2}, Turing~\cite{Turing:old}, Modula-3~\cite{Modula-3}, NeWS~\cite{NeWS}, Emerald~\cite{Emerald}, \uC~\cite{Buhr92a} and Java~\cite{Java}---provide monitors as explicit language constructs. In addition, operating-system kernels and device drivers have a monitor-like structure, although they often use lower-level primitives such as semaphores or locks to simulate monitors. For these reasons, this project proposes monitors as the core concurrency-construct.
     10An approach that is worth mentionning because it is gaining in popularity is transactionnal memory~\cite{Dice10}[Check citation]. While this approach is even pursued by system languages like \CC\cit, the performance and feature set is currently too restrictive to be the main concurrency paradigm for systems language, which is why it was rejected as the core paradigm for concurrency in \CFA.
     11
     12One of the most natural, elegant, and efficient mechanisms for synchronization and communication, especially for shared-memory systems, is the \emph{monitor}. Monitors were first proposed by Brinch Hansen~\cite{Hansen73} and later described and extended by C.A.R.~Hoare~\cite{Hoare74}. Many programming languages---e.g., Concurrent Pascal~\cite{ConcurrentPascal}, Mesa~\cite{Mesa}, Modula~\cite{Modula-2}, Turing~\cite{Turing:old}, Modula-3~\cite{Modula-3}, NeWS~\cite{NeWS}, Emerald~\cite{Emerald}, \uC~\cite{Buhr92a} and Java~\cite{Java}---provide monitors as explicit language constructs. In addition, operating-system kernels and device drivers have a monitor-like structure, although they often use lower-level primitives such as semaphores or locks to simulate monitors. For these reasons, this project proposes monitors as the core concurrency-construct.
    1313
    1414\section{Basics}
    15 Non-determinism requires concurrent systems to offer support for mutual-exclusion and synchronisation. Mutual-exclusion is the concept that only a fixed number of threads can access a critical section at any given time, where a critical section is a group of instructions on an associated portion of data that requires the restricted access. On the other hand, synchronization enforces relative ordering of execution and synchronization tools numerous mechanisms to establish timing relationships among threads.
     15Non-determinism requires concurrent systems to offer support for mutual-exclusion and synchronisation. Mutual-exclusion is the concept that only a fixed number of threads can access a critical section at any given time, where a critical section is a group of instructions on an associated portion of data that requires the restricted access. On the other hand, synchronization enforces relative ordering of execution and synchronization tools provide numerous mechanisms to establish timing relationships among threads.
    1616
    1717\subsection{Mutual-Exclusion}
    18 As mentionned above, mutual-exclusion is the guarantee that only a fix number of threads can enter a critical section at once. However, many solution exists for mutual exclusion which vary in terms of performance, flexibility and ease of use. Methods range from low-level locks, which are fast and flexible but require significant attention to be correct, to  higher-level mutual-exclusion methods, which sacrifice some performance in order to improve ease of use. Ease of use comes by either guaranteeing some problems cannot occur (e.g., being deadlock free) or by offering a more explicit coupling between data and corresponding critical section. For example, the \CC \code{std::atomic<T>} which offer an easy way to express mutual-exclusion on a restricted set of operations (.e.g: reading/writing large types atomically). Another challenge with low-level locks is composability. Locks are not composable because it takes careful organising for multiple locks to be used while preventing deadlocks. Easing composability is another feature higher-level mutual-exclusion mechanisms often offer.
     18As mentionned above, mutual-exclusion is the guarantee that only a fix number of threads can enter a critical section at once. However, many solutions exist for mutual exclusion, which vary in terms of performance, flexibility and ease of use. Methods range from low-level locks, which are fast and flexible but require significant attention to be correct, to  higher-level mutual-exclusion methods, which sacrifice some performance in order to improve ease of use. Ease of use comes by either guaranteeing some problems cannot occur (e.g., being deadlock free) or by offering a more explicit coupling between data and corresponding critical section. For example, the \CC \code{std::atomic<T>} offers an easy way to express mutual-exclusion on a restricted set of operations (e.g.: reading/writing large types atomically). Another challenge with low-level locks is composability. Locks have restricted composability because it takes careful organising for multiple locks to be used while preventing deadlocks. Easing composability is another feature higher-level mutual-exclusion mechanisms often offer.
    1919
    2020\subsection{Synchronization}
    21 As for mutual-exclusion, low level synchronisation primitive often offer good performance and good flexibility at the cost of ease of use. Again, higher-level mechanism often simplify usage by adding better coupling between synchronization and data, .eg., message passing, or offering simple solution to otherwise involved challenges. An example of this is barging. As mentionned above synchronization can be expressed as guaranteeing that event \textit{X} always happens before \textit{Y}. Most of the time synchronisation happens around a critical section, where threads most acquire said critical section in a certain order. However, it may also be desired to be able to guarantee that event \textit{Z} does not occur between \textit{X} and \textit{Y}. This is called barging, where event \textit{X} tries to effect event \textit{Y} but anoter thread races to grab the critical section and emits \textit{Z} before \textit{Y}. Preventing or detecting barging is an involved challenge with low-level locks, which can be made much easier by higher-level constructs.
     21As for mutual-exclusion, low-level synchronisation primitives often offer good performance and good flexibility at the cost of ease of use. Again, higher-level mechanism often simplify usage by adding better coupling between synchronization and data, e.g.: message passing, or offering simple solution to otherwise involved challenges. An example is barging. As mentioned above, synchronization can be expressed as guaranteeing that event \textit{X} always happens before \textit{Y}. Most of the time, synchronisation happens around a critical section, where threads must acquire critical sections in a certain order. However, it may also be desirable to guarantee that event \textit{Z} does not occur between \textit{X} and \textit{Y}. Not satisfying this property called barging. For example, where event \textit{X} tries to effect event \textit{Y} but another thread acquires the critical section and emits \textit{Z} before \textit{Y}. Preventing or detecting barging is an involved challenge with low-level locks, which can be made much easier by higher-level constructs. This challenge is often split into two different methods, barging avoidance and barging prevention. Algorithms that use status flags and other flag variables to detect barging threads are said to be using barging avoidance while algorithms that baton-passing locks between threads instead of releasing the locks are said to be using barging prevention.
    2222
    2323% ======================================================================
    … …  
    5555        void ?{}(size_t * this, counter_t & mutex cnt); //conversion
    5656\end{cfacode}
    57 
    58 Here, the constructor(\code{?\{\}}) uses the \code{nomutex} keyword to signify that it does not acquire the monitor mutual-exclusion when constructing. This semantics is because an object not yet constructed should never be shared and therefore does not require mutual exclusion. The prefix increment operator uses \code{mutex} to protect the incrementing process from race conditions. Finally, there is a conversion operator from \code{counter_t} to \code{size_t}. This conversion may or may not require the \code{mutex} keyword depending on whether or not reading an \code{size_t} is an atomic operation.
    59 
    60 Having both \code{mutex} and \code{nomutex} keywords is redundant based on the meaning of a routine having neither of these keywords. For example, given a routine without qualifiers \code{void foo(counter_t & this)}, then it is reasonable that it should default to the safest option \code{mutex}, whereas assuming \code{nomutex} is unsafe and may cause subtle errors. In fact, \code{nomutex} is the "normal" parameter behaviour, with the \code{nomutex} keyword effectively stating explicitly that "this routine is not special". Another alternative is to make having exactly one of these keywords mandatory, which would provide the same semantics but without the ambiguity of supporting routines neither keyword. Mandatory keywords would also have the added benefit of being self-documented but at the cost of extra typing. While there are several benefits to mandatory keywords, they do bring a few challenges. Mandatory keywords in \CFA would imply that the compiler must know without a doubt wheter or not a parameter is a monitor or not. Since \CFA relies heavily on traits as an abstraction mechanism, the distinction between a type that is a monitor and a type that looks like a monitor can become blurred. For this reason, \CFA only has the \code{mutex} keyword.
    61 
    62 
    63 The next semantic decision is to establish when \code{mutex} may be used as a type qualifier. Consider the following declarations:
    64 \begin{cfacode}
    65 int f1(monitor & mutex m);
    66 int f2(const monitor & mutex m);
    67 int f3(monitor ** mutex m);
    68 int f4(monitor * mutex m []);
    69 int f5(graph(monitor*) & mutex m);
    70 \end{cfacode}
    71 The problem is to indentify which object(s) should be acquired. Furthermore, each object needs to be acquired only once. In the case of simple routines like \code{f1} and \code{f2} it is easy to identify an exhaustive list of objects to acquire on entry. Adding indirections (\code{f3}) still allows the compiler and programmer to indentify which object is acquired. However, adding in arrays (\code{f4}) makes it much harder. Array lengths are not necessarily known in C, and even then making sure objects are only acquired once becomes none-trivial. This can be extended to absurd limits like \code{f5}, which uses a graph of monitors. To make the issue tractable, this projects imposes the requirement that a routine may only acquire one monitor per parameter and it must be the type of the parameter with one level of indirection (ignoring potential qualifiers). Also note that while routine \code{f3} can be supported, meaning that monitor \code{**m} is be acquired, passing an array to this routine would be type safe and yet result in undefined behavior because only the first element of the array is acquired. This is specially true for non-copyable objects like monitors, where an array of pointers is simplest way to express a group of monitors. However, this ambiguity is part of the C type-system with respects to arrays. For this reason, \code{mutex} is disallowed in the context where arrays may be passed:
    72 
    73 \begin{cfacode}
    74 int f1(monitor & mutex m);   //Okay : recommanded case
    75 int f2(monitor * mutex m);   //Okay : could be an array but probably not
    76 int f3(monitor mutex m []);  //Not Okay : Array of unkown length
    77 int f4(monitor ** mutex m);  //Not Okay : Could be an array
    78 int f5(monitor * mutex m []); //Not Okay : Array of unkown length
    79 \end{cfacode}
    80 
    81 Unlike object-oriented monitors, where calling a mutex member \emph{implicitly} acquires mutual-exclusion, \CFA uses an explicit mechanism to acquire mutual-exclusion. A consequence of this approach is that it extends naturally to multi-monitor calls.
    82 \begin{cfacode}
    83 int f(MonitorA & mutex a, MonitorB & mutex b);
    84 
    85 MonitorA a;
    86 MonitorB b;
    87 f(a,b);
    88 \end{cfacode}
    89 The capacity to acquire multiple locks before entering a critical section is called \emph{\gls{group-acquire}}. In practice, writing multi-locking routines that do not lead to deadlocks is tricky. Having language support for such a feature is therefore a significant asset for \CFA. In the case presented above, \CFA guarantees that the order of aquisition is consistent across calls to routines using the same monitors as arguments. However, since \CFA monitors use multi-acquisition locks, users can effectively force the acquiring order. For example, notice which routines use \code{mutex}/\code{nomutex} and how this affects aquiring order:
    90 \begin{cfacode}
    91         void foo(A & mutex a, B & mutex b) { //acquire a & b
    92                 ...
    93         }
    94 
    95         void bar(A & mutex a, B & /*nomutex*/ b) { //acquire a
    96                 ... foo(a, b); ... //acquire b
    97         }
    98 
    99         void baz(A & /*nomutex*/ a, B & mutex b) { //acquire b
    100                 ... foo(a, b); ... //acquire a
    101         }
    102 \end{cfacode}
    103 The multi-acquisition monitor lock allows a monitor lock to be acquired by both \code{bar} or \code{baz} and acquired again in \code{foo}. In the calls to \code{bar} and \code{baz} the monitors are acquired in opposite order.
    104 
    105 However, such use leads the lock acquiring order problem. In the example above, the user uses implicit ordering in the case of function \code{foo} but explicit ordering in the case of \code{bar} and \code{baz}. This subtle mistake means that calling these routines concurrently may lead to deadlock and is therefore undefined behavior. As shown on several occasion\cit, solving this problem requires:
    106 \begin{enumerate}
    107         \item Dynamically tracking of the monitor-call order.
    108         \item Implement rollback semantics.
    109 \end{enumerate}
    110 While the first requirement is already a significant constraint on the system, implementing a general rollback semantics in a C-like language is prohibitively complex \cit. In \CFA, users simply need to be carefull when acquiring multiple monitors at the same time.
    111 
    112 Finally, for convenience, monitors support multiple acquiring, that is acquiring a monitor while already holding it does not cause a deadlock. It simply increments an internal counter which is then used to release the monitor after the number of acquires and releases match up. This is particularly usefull when monitor routines use other monitor routines as helpers or for recursions. For example:
    113 \begin{cfacode}
    114 monitor bank {
    115         int money;
    116         log_t usr_log;
    117 };
    118 
    119 void deposit( bank & mutex b, int deposit ) {
    120         b.money += deposit;
    121         b.usr_log | "Adding" | deposit | endl;
    122 }
    123 
    124 void transfer( bank & mutex mybank, bank & mutex yourbank, int me2you) {
    125         deposit( mybank, -me2you );
    126         deposit( yourbank, me2you );
    127 }
    128 \end{cfacode}
    129 
    130 % ======================================================================
    131 % ======================================================================
    132 \subsection{Data semantics} \label{data}
    133 % ======================================================================
    134 % ======================================================================
    135 Once the call semantics are established, the next step is to establish data semantics. Indeed, until now a monitor is used simply as a generic handle but in most cases monitors contain shared data. This data should be intrinsic to the monitor declaration to prevent any accidental use of data without its appropriate protection. For example, here is a complete version of the counter showed in section \ref{call}:
    136 \begin{cfacode}
    137 monitor counter_t {
    138         int value;
    139 };
    140 
    141 void ?{}(counter_t & this) {
    142         this.cnt = 0;
    143 }
    144 
    145 int ?++(counter_t & mutex this) {
    146         return ++this.value;
    147 }
    148 
    149 //need for mutex is platform dependent here
    150 void ?{}(int * this, counter_t & mutex cnt) {
    151         *this = (int)cnt;
    152 }
    153 \end{cfacode}
    154 
    15557This counter is used as follows:
    15658\begin{center}
    … …  
    17173Notice how the counter is used without any explicit synchronisation and yet supports thread-safe semantics for both reading and writting.
    17274
    173 % ======================================================================
    174 % ======================================================================
    175 \subsection{Implementation Details: Interaction with polymorphism}
    176 % ======================================================================
    177 % ======================================================================
    178 Depending on the choice of semantics for when monitor locks are acquired, interaction between monitors and \CFA's concept of polymorphism can be complex to support. However, it is shown that entry-point locking solves most of the issues.
    179 
    180 First of all, interaction between \code{otype} polymorphism and monitors is impossible since monitors do not support copying. Therefore, the main question is how to support \code{dtype} polymorphism. Since a monitor's main purpose is to ensure mutual exclusion when accessing shared data, this implies that mutual exclusion is only required for routines that do in fact access shared data. However, since \code{dtype} polymorphism always handles incomplete types (by definition), no \code{dtype} polymorphic routine can access shared data since the data requires knowledge about the type. Therefore, the only concern when combining \code{dtype} polymorphism and monitors is to protect access to routines.
    181 
    182 Before looking into complex control-flow, it is important to present the difference between the two acquiring options : callsite and entry-point locking, i.e. acquiring the monitors before making a mutex routine call or as the first operation of the mutex routine-call. For example:
    183 \begin{center}
    184 \setlength\tabcolsep{1.5pt}
    185 \begin{tabular}{|c|c|c|}
    186 Code & \gls{callsite-locking} & \gls{entry-point-locking} \\
    187 \CFA & pseudo-code & pseudo-code \\
    188 \hline
    189 \begin{cfacode}[tabsize=3]
    190 void foo(monitor& mutex a){
    191 
    192 
    193 
    194         //Do Work
    195         //...
    196 
    197 }
    198 
    199 void main() {
    200         monitor a;
    201 
    202 
    203 
    204         foo(a);
    205 
    206 }
    207 \end{cfacode} & \begin{pseudo}[tabsize=3]
    208 foo(& a) {
    209 
    210 
    211 
    212         //Do Work
    213         //...
    214 
    215 }
    216 
    217 main() {
    218         monitor a;
    219         //calling routine
    220         //handles concurrency
    221         acquire(a);
    222         foo(a);
    223         release(a);
    224 }
    225 \end{pseudo} & \begin{pseudo}[tabsize=3]
    226 foo(& a) {
    227         //called routine
    228         //handles concurrency
    229         acquire(a);
    230         //Do Work
    231         //...
    232         release(a);
    233 }
    234 
    235 main() {
    236         monitor a;
    237 
    238 
    239 
    240         foo(a);
    241 
    242 }
    243 \end{pseudo}
    244 \end{tabular}
    245 \end{center}
    246 
    247 \Gls{callsite-locking} is inefficient, since any \code{dtype} routine may have to obtain some lock before calling a routine, depending on whether or not the type passed is a monitor. However, with \gls{entry-point-locking} calling a monitor routine becomes exactly the same as calling it from anywhere else.
    248 
    249 Note the \code{mutex} keyword relies on the resolver, which means that in cases where a generic monitor routine is actually desired, writing a mutex routine is possible with the proper trait. This is possible because monitors are designed in terms a trait. For example:
    250 \begin{cfacode}
    251 //Incorrect
    252 //T is not a monitor
    253 forall(dtype T)
    254 void foo(T * mutex t);
    255 
    256 //Correct
    257 //this function only works on monitors
    258 //(any monitor)
    259 forall(dtype T | is_monitor(T))
    260 void bar(T * mutex t));
     75Here, the constructor(\code{?\{\}}) uses the \code{nomutex} keyword to signify that it does not acquire the monitor mutual-exclusion when constructing. This semantics is because an object not yet constructed should never be shared and therefore does not require mutual exclusion. The prefix increment operator uses \code{mutex} to protect the incrementing process from race conditions. Finally, there is a conversion operator from \code{counter_t} to \code{size_t}. This conversion may or may not require the \code{mutex} keyword depending on whether or not reading a \code{size_t} is an atomic operation.
     76
     77For maximum usability, monitors use \gls{multi-acq} semantics, which means a single thread can acquire multiple times the same monitor without deadlock. For example, figure \ref{fig:search} uses recursion and \gls{multi-acq} to print values inside a binary tree.
     78\begin{figure}
     79\label{fig:search}
     80\begin{cfacode}
     81monitor printer { ... };
     82struct tree {
     83        tree * left, right;
     84        char * value;
     85};
     86void print(printer & mutex p, char * v);
     87
     88void print(printer & mutex p, tree * t) {
     89        print(p, t->value);
     90        print(p, t->left );
     91        print(p, t->right);
     92}
     93\end{cfacode}
     94\caption{Recursive printing algorithm using \gls{multi-acq}.}
     95\end{figure}
     96
     97Having both \code{mutex} and \code{nomutex} keywords is redundant based on the meaning of a routine having neither of these keywords. For example, given a routine without qualifiers \code{void foo(counter_t & this)}, then it is reasonable that it should default to the safest option \code{mutex}, whereas assuming \code{nomutex} is unsafe and may cause subtle errors. In fact, \code{nomutex} is the "normal" parameter behaviour, with the \code{nomutex} keyword effectively stating explicitly that "this routine is not special". Another alternative is making exactly one of these keywords mandatory, which would provide the same semantics but without the ambiguity of supporting routines with neither keyword. Mandatory keywords would also have the added benefit of being self-documented but at the cost of extra typing. While there are several benefits to mandatory keywords, they do bring a few challenges. Mandatory keywords in \CFA would imply that the compiler must know without doubt whether or not a parameter is a monitor or not. Since \CFA relies heavily on traits as an abstraction mechanism, the distinction between a type that is a monitor and a type that looks like a monitor can become blurred. For this reason, \CFA only has the \code{mutex} keyword and uses no keyword to mean \code{nomutex}.
     98
     99The next semantic decision is to establish when \code{mutex} may be used as a type qualifier. Consider the following declarations:
     100\begin{cfacode}
     101int f1(monitor & mutex m);
     102int f2(const monitor & mutex m);
     103int f3(monitor ** mutex m);
     104int f4(monitor * mutex m []);
     105int f5(graph(monitor*) & mutex m);
     106\end{cfacode}
     107The problem is to indentify which object(s) should be acquired. Furthermore, each object needs to be acquired only once. In the case of simple routines like \code{f1} and \code{f2} it is easy to identify an exhaustive list of objects to acquire on entry. Adding indirections (\code{f3}) still allows the compiler and programmer to indentify which object is acquired. However, adding in arrays (\code{f4}) makes it much harder. Array lengths are not necessarily known in C, and even then making sure objects are only acquired once becomes none-trivial. This problem can be extended to absurd limits like \code{f5}, which uses a graph of monitors. To make the issue tractable, this project imposes the requirement that a routine may only acquire one monitor per parameter and it must be the type of the parameter with at most one level of indirection (ignoring potential qualifiers). Also note that while routine \code{f3} can be supported, meaning that monitor \code{**m} is be acquired, passing an array to this routine would be type safe and yet result in undefined behavior because only the first element of the array is acquired. However, this ambiguity is part of the C type-system with respects to arrays. For this reason, \code{mutex} is disallowed in the context where arrays may be passed:
     108\begin{cfacode}
     109int f1(monitor & mutex m);   //Okay : recommanded case
     110int f2(monitor * mutex m);   //Okay : could be an array but probably not
     111int f3(monitor mutex m []);  //Not Okay : Array of unkown length
     112int f4(monitor ** mutex m);  //Not Okay : Could be an array
     113int f5(monitor * mutex m []); //Not Okay : Array of unkown length
     114\end{cfacode}
     115Note that not all array functions are actually distinct in the type system sense. However, even the code generation could tell the difference, the extra information is still not sufficient to extend meaningfully the monitor call semantic.
     116
     117Unlike object-oriented monitors, where calling a mutex member \emph{implicitly} acquires mutual-exclusion often receives an object, \CFA uses an explicit mechanism to acquire mutual-exclusion. A consequence of this approach is that it extends naturally to multi-monitor calls.
     118\begin{cfacode}
     119int f(MonitorA & mutex a, MonitorB & mutex b);
     120
     121MonitorA a;
     122MonitorB b;
     123f(a,b);
     124\end{cfacode}
     125The capacity to acquire multiple locks before entering a critical section is called \emph{\gls{bulk-acq}}. In practice, writing multi-locking routines that do not lead to deadlocks is tricky. Having language support for such a feature is therefore a significant asset for \CFA. In the case presented above, \CFA guarantees that the order of aquisition is consistent across calls to routines using the same monitors as arguments. However, since \CFA monitors use \gls{multi-acq} locks, users can effectively force the acquiring order. For example, notice which routines use \code{mutex}/\code{nomutex} and how this affects aquiring order:
     126\begin{cfacode}
     127        void foo(A & mutex a, B & mutex b) { //acquire a & b
     128                ...
     129        }
     130
     131        void bar(A & mutex a, B & /*nomutex*/ b) { //acquire a
     132                ... foo(a, b); ... //acquire b
     133        }
     134
     135        void baz(A & /*nomutex*/ a, B & mutex b) { //acquire b
     136                ... foo(a, b); ... //acquire a
     137        }
     138\end{cfacode}
     139The \gls{multi-acq} monitor lock allows a monitor lock to be acquired by both \code{bar} or \code{baz} and acquired again in \code{foo}. In the calls to \code{bar} and \code{baz} the monitors are acquired in opposite order.
     140
     141However, such use leads to the lock acquiring order problem. In the example above, the user uses implicit ordering in the case of function \code{foo} but explicit ordering in the case of \code{bar} and \code{baz}. This subtle mistake means that calling these routines concurrently may lead to deadlock and is therefore undefined behavior. As shown on several occasion\cit, solving this problem requires:
     142\begin{enumerate}
     143        \item Dynamically tracking of the monitor-call order.
     144        \item Implement rollback semantics.
     145\end{enumerate}
     146While the first requirement is already a significant constraint on the system, implementing a general rollback semantics in a C-like language is prohibitively complex \cit. In \CFA, users simply need to be carefull when acquiring multiple monitors at the same time or only use \gls{bulk-acq} of all the monitors.
     147
     148\Gls{multi-acq} and \gls{bulk-acq} can be used together in interesting ways, for example:
     149\begin{cfacode}
     150monitor bank { ... };
     151
     152void deposit( bank & mutex b, int deposit );
     153
     154void transfer( bank & mutex mybank, bank & mutex yourbank, int me2you) {
     155        deposit( mybank, -me2you );
     156        deposit( yourbank, me2you );
     157}
     158\end{cfacode}
     159This example shows a trivial solution to the bank account transfer problem\cit. Without \gls{multi-acq} and \gls{bulk-acq}, the solution to this problem is much more involved and requires carefull engineering.
     160
     161% ======================================================================
     162% ======================================================================
     163\subsection{Data semantics} \label{data}
     164% ======================================================================
     165% ======================================================================
     166Once the call semantics are established, the next step is to establish data semantics. Indeed, until now a monitor is used simply as a generic handle but in most cases monitors contain shared data. This data should be intrinsic to the monitor declaration to prevent any accidental use of data without its appropriate protection. For example, here is a complete version of the counter showed in section \ref{call}:
     167\begin{cfacode}
     168monitor counter_t {
     169        int value;
     170};
     171
     172void ?{}(counter_t & this) {
     173        this.cnt = 0;
     174}
     175
     176int ?++(counter_t & mutex this) {
     177        return ++this.value;
     178}
     179
     180//need for mutex is platform dependent here
     181void ?{}(int * this, counter_t & mutex cnt) {
     182        *this = (int)cnt;
     183}
    261184\end{cfacode}
    262185
    … …  
    267190% ======================================================================
    268191% ======================================================================
    269 In addition to mutual exclusion, the monitors at the core of \CFA's concurrency can also be used to achieve synchronisation. With monitors, this is generally achieved with internal or external scheduling as in\cit. Since internal scheduling of single monitors is mostly a solved problem, this proposal concentraits on extending internal scheduling to multiple monitors at once. Indeed, like the \gls{group-acquire} semantics, internal scheduling extends to multiple monitors at once in a way that is natural to the user but requires additional complexity on the implementation side.
     192In addition to mutual exclusion, the monitors at the core of \CFA's concurrency can also be used to achieve synchronisation. With monitors, this capability is generally achieved with internal or external scheduling as in\cit. Since internal scheduling within a single monitor is mostly a solved problem, this thesis concentrates on extending internal scheduling to multiple monitors. Indeed, like the \gls{bulk-acq} semantics, internal scheduling extends to multiple monitors in a way that is natural to the user but requires additional complexity on the implementation side.
    270193
    271194First, here is a simple example of such a technique:
    … …  
    278201        void foo(A & mutex a) {
    279202                ...
    280                 // Wait for cooperation from bar()
     203                //Wait for cooperation from bar()
    281204                wait(a.e);
    282205                ...
    … …  
    284207
    285208        void bar(A & mutex a) {
    286                 // Provide cooperation for foo()
     209                //Provide cooperation for foo()
    287210                ...
    288                 // Unblock foo at scope exit
     211                //Unblock foo
    289212                signal(a.e);
    290213        }
    291214\end{cfacode}
    292215
    293 There are two details to note here. First, there \code{signal} is a delayed operation, it only unblocks the waiting thread when it reaches the end of the critical section. This is needed to respect mutual-exclusion. Second, in \CFA, \code{condition} have no particular need to be stored inside a monitor, beyond any software engineering reasons. Here routine \code{foo} waits for the \code{signal} from \code{bar} before making further progress, effectively ensuring a basic ordering.
    294 
    295 An important aspect to take into account here is that \CFA does not allow barging, which means that once function \code{bar} releases the monitor, foo is guaranteed to resume immediately after (unless some other thread waited on the same condition). This guarantees offers the benefit of not having to loop arount waits in order to guarantee that a condition is still met. The main reason \CFA offers this guarantee is that users can easily introduce barging if it becomes a necessity but adding barging prevention or barging avoidance is more involved without language support. Supporting barging prevention as well as extending internal scheduling to multiple monitors is the main source of complexity in the design of \CFA concurrency.
     216There are two details to note here. First, the \code{signal} is a delayed operation, it only unblocks the waiting thread when it reaches the end of the critical section. This semantic is needed to respect mutual-exclusion. Second, in \CFA, a \code{condition} variable can be stored/created independently of a monitor. Here routine \code{foo} waits for the \code{signal} from \code{bar} before making further progress, effectively ensuring a basic ordering.
     217
     218An important aspect of the implementation is that \CFA does not allow barging, which means that once function \code{bar} releases the monitor, foo is guaranteed to resume immediately after (unless some other thread waited on the same condition). This guarantees offers the benefit of not having to loop arount waits in order to guarantee that a condition is still met. The main reason \CFA offers this guarantee is that users can easily introduce barging if it becomes a necessity but adding barging prevention or barging avoidance is more involved without language support. Supporting barging prevention as well as extending internal scheduling to multiple monitors is the main source of complexity in the design of \CFA concurrency.
    296219
    297220% ======================================================================
    … …  
    319242\end{pseudo}
    320243\end{multicols}
    321 The example shows the simple case of having two threads (one for each column) and a single monitor A. One thread acquires before waiting (atomically blocking and releasing A) and the other acquires before signalling. There is an important thing to note here, both \code{wait} and \code{signal} must be called with the proper monitor(s) already acquired. This restriction is hidden on the user side in \uC, as it is a logical requirement for barging prevention.
    322 
    323 A direct extension of the previous example is the \gls{group-acquire} version:
     244The example shows the simple case of having two threads (one for each column) and a single monitor A. One thread acquires before waiting (atomically blocking and releasing A) and the other acquires before signalling. It is important to note here that both \code{wait} and \code{signal} must be called with the proper monitor(s) already acquired. This semantic is a logical requirement for barging prevention.
     245
     246A direct extension of the previous example is a \gls{bulk-acq} version:
    324247
    325248\begin{multicols}{2}
    … …  
    338261\end{pseudo}
    339262\end{multicols}
    340 This version uses \gls{group-acquire} (denoted using the \& symbol), but the presence of multiple monitors does not add a particularly new meaning. Synchronization happens between the two threads in exactly the same way and order. The only difference is that mutual exclusion covers more monitors. On the implementation side, handling multiple monitors does add a degree of complexity as the next few examples demonstrate.
    341 
    342 While deadlock issues can occur when nesting monitors, these issues are only a symptom of the fact that locks, and by extension monitors, are not perfectly composable. However, for monitors as for locks, it is possible to write a program using nesting without encountering any problems if nested is done correctly. For example, the next pseudo-code snippet acquires monitors A then B before waiting while only acquiring B when signalling, effectively avoiding the nested monitor problem.
    343 
     263This version uses \gls{bulk-acq} (denoted using the \& symbol), but the presence of multiple monitors does not add a particularly new meaning. Synchronization happens between the two threads in exactly the same way and order. The only difference is that mutual exclusion covers more monitors. On the implementation side, handling multiple monitors does add a degree of complexity as the next few examples demonstrate.
     264
     265While deadlock issues can occur when nesting monitors, these issues are only a symptom of the fact that locks, and by extension monitors, are not perfectly composable. For monitors, a well known deadlock problem is the Nested Monitor Problem\cit, which occurs when a \code{wait} is made on a thread that holds more than one monitor. For example, the following pseudo-code will run into the nested monitor problem :
    344266\begin{multicols}{2}
    345267\begin{pseudo}
    … …  
    354276
    355277\begin{pseudo}
     278acquire A
     279        acquire B
     280                signal B
     281        release B
     282release A
     283\end{pseudo}
     284\end{multicols}
     285However, for monitors as for locks, it is possible to write a program using nesting without encountering any problems if nesting is done correctly. For example, the next pseudo-code snippet acquires monitors {\sf A} then {\sf B} before waiting, while only acquiring {\sf B} when signalling, effectively avoiding the nested monitor problem.
     286
     287\begin{multicols}{2}
     288\begin{pseudo}
     289acquire A
     290        acquire B
     291                wait B
     292        release B
     293release A
     294\end{pseudo}
     295
     296\columnbreak
     297
     298\begin{pseudo}
    356299
    357300acquire B
    … …  
    362305\end{multicols}
    363306
    364 The next example is where \gls{group-acquire} adds a significant layer of complexity to the internal signalling semantics.
     307The next example is where \gls{bulk-acq} adds a significant layer of complexity to the internal signalling semantics.
    365308
    366309\begin{multicols}{2}
    … …  
    368311\begin{pseudo}[numbers=left]
    369312acquire A
    370         // Code Section 1
    371         acquire A & B
    372                 // Code Section 2
     313        //Code Section 1
     314        acquire A & B
     315                //Code Section 2
    373316                wait A & B
    374                 // Code Section 3
    375         release A & B
    376         // Code Section 4
     317                //Code Section 3
     318        release A & B
     319        //Code Section 4
    377320release A
    378321\end{pseudo}
    … …  
    383326\begin{pseudo}[numbers=left, firstnumber=10]
    384327acquire A
    385         // Code Section 5
    386         acquire A & B
    387                 // Code Section 6
     328        //Code Section 5
     329        acquire A & B
     330                //Code Section 6
    388331                signal A & B
    389                 // Code Section 7
    390         release A & B
    391         // Code Section 8
     332                //Code Section 7
     333        release A & B
     334        //Code Section 8
    392335release A
    393336\end{pseudo}
    … …  
    397340\end{center}
    398341
    399 It is particularly important to pay attention to code sections 8 and 4, which are where the existing semantics of internal scheduling need to be extended for multiple monitors. The root of the problem is that \gls{group-acquire} is used in a context where one of the monitors is already acquired and is why it is important to define the behaviour of the previous pseudo-code. When the signaller thread reaches the location where it should "release A \& B" (line 16), it must actually transfer ownership of monitor B to the waiting thread. This ownership trasnfer is required in order to prevent barging. Since the signalling thread still needs the monitor A, simply waking up the waiting thread is not an option because it would violate mutual exclusion. There are three options:
     342It is particularly important to pay attention to code sections 4 and 8, which are where the existing semantics of internal scheduling need to be extended for multiple monitors. The root of the problem is that \gls{bulk-acq} is used in a context where one of the monitors is already acquired and is why it is important to define the behaviour of the previous pseudo-code. When the signaller thread reaches the location where it should "release A \& B" (line 16), it must actually transfer ownership of monitor B to the waiting thread. This ownership trasnfer is required in order to prevent barging. Since the signalling thread still needs monitor A, simply waking up the waiting thread is not an option because it would violate mutual exclusion. There are three options.
    400343
    401344\subsubsection{Delaying signals}
    402 The first more obvious solution to solve the problem of multi-monitor scheduling is to keep ownership of all locks until the last lock is ready to be transferred. It can be argued that that moment is the correct time to transfer ownership when the last lock is no longer needed because this semantics fits most closely to the behaviour of single monitor scheduling. This solution has the main benefit of transferring ownership of groups of monitors, which simplifies the semantics from mutiple objects to a single group of object, effectively making the existing single monitor semantic viable by simply changing monitors to monitor collections.
     345The first more obvious solution to solve the problem of multi-monitor scheduling is to keep ownership of all locks until the last lock is ready to be transferred. It can be argued that that moment is the correct time to transfer ownership when the last lock is no longer needed because this semantics fits most closely to the behaviour of single monitor scheduling. This solution has the main benefit of transferring ownership of groups of monitors, which simplifies the semantics from mutiple objects to a single group of objects, effectively making the existing single monitor semantic viable by simply changing monitors to monitor groups.
    403346\begin{multicols}{2}
    404347Waiter
    … …  
    424367\end{pseudo}
    425368\end{multicols}
    426 However, this solution can become much more complicated depending on what is executed while secretly holding B (at line 10). Indeed, nothing prevents a user from signalling monitor A on a different condition variable:
    427 \newpage
     369However, this solution can become much more complicated depending on what is executed while secretly holding B (at line 10). Indeed, nothing prevents signalling monitor A on a different condition variable:
    428370\begin{multicols}{2}
    429371Thread 1
    … …  
    467409Note that ordering is not determined by a race condition but by whether signalled threads are enqueued in FIFO or FILO order. However, regardless of the answer, users can move line 15 before line 11 and get the reverse effect.
    468410
    469 In both cases, the threads need to be able to distinguish on a per monitor basis which ones need to be released and which ones need to be transferred. Which means monitors cannot be handled as a single homogenous group.
     411In both cases, the threads need to be able to distinguish, on a per monitor basis, which ones need to be released and which ones need to be transferred, which means monitors cannot be handled as a single homogenous group and therefore invalidates the main benefit of this approach.
    470412
    471413\subsubsection{Dependency graphs}
    472 In the Listing 1 pseudo-code, there is a solution which statisfies both barging prevention and mutual exclusion. If ownership of both monitors is transferred to the waiter when the signaller releases A and then the waiter transfers back ownership of A when it releases it then the problem is solved. Dynamically finding the correct order is therefore the second possible solution. The problem it encounters is that it effectively boils down to resolving a dependency graph of ownership requirements. Here even the simplest of code snippets requires two transfers and it seems to increase in a manner closer to polynomial. For example, the following code, which is just a direct extension to three monitors, requires at least three ownership transfer and has multiple solutions:
     414In the Listing 1 pseudo-code, there is a solution which statisfies both barging prevention and mutual exclusion. If ownership of both monitors is transferred to the waiter when the signaller releases A and then the waiter transfers back ownership of A when it releases it, then the problem is solved. Dynamically finding the correct order is therefore the second possible solution. The problem it encounters is that it effectively boils down to resolving a dependency graph of ownership requirements. Here even the simplest of code snippets requires two transfers and it seems to increase in a manner closer to polynomial. For example, the following code, which is just a direct extension to three monitors, requires at least three ownership transfer and has multiple solutions:
    473415
    474416\begin{multicols}{2}
    … …  
    495437\end{pseudo}
    496438\end{multicols}
    497 Resolving dependency graph being a complex and expensive endeavour, this solution is not the preffered one.
     439
     440\begin{figure}
     441\begin{multicols}{3}
     442Thread $\alpha$
     443\begin{pseudo}[numbers=left, firstnumber=1]
     444acquire A
     445        acquire A & B
     446                wait A & B
     447        release A & B
     448release A
     449\end{pseudo}
     450
     451\columnbreak
     452
     453Thread $\gamma$
     454\begin{pseudo}[numbers=left, firstnumber=1]
     455acquire A
     456        acquire A & B
     457                signal A & B
     458        release A & B
     459        signal A
     460release A
     461\end{pseudo}
     462
     463\columnbreak
     464
     465Thread $\beta$
     466\begin{pseudo}[numbers=left, firstnumber=1]
     467acquire A
     468        wait A
     469release A
     470\end{pseudo}
     471
     472\end{multicols}
     473\caption{Dependency graph}
     474\label{lst:dependency}
     475\end{figure}
     476
     477\begin{figure}
     478\begin{center}
     479\input{dependency}
     480\end{center}
     481\label{fig:dependency}
     482\caption{Dependency graph of the statments in listing \ref{lst:dependency}}
     483\end{figure}
     484
     485Listing \ref{lst:dependency} is the three thread example rewritten for dependency graphs as well as the corresponding dependency graph. Figure \ref{fig:dependency} shows the corresponding dependency graph that results, where every node is a statment of one of the three threads, and the arrows the dependency of that statment. The extra challenge is that this dependency graph is effectively post-mortem, but the run time system needs to be able to build and solve these graphs as the dependency unfolds. Resolving dependency graph being a complex and expensive endeavour, this solution is not the preffered one.
    498486
    499487\subsubsection{Partial signalling} \label{partial-sig}
    … …  
    516504                signal A & B
    517505        release A & B
    518         // ... More code
    519 release A
    520 \end{pseudo}
    521 \end{multicols}
    522 The partial signalling solution transfers ownership of monitor B at lines 10 but does not wake the waiting thread since it is still using monitor A. Only when it reaches line 11 does it actually wakeup the waiting thread. This solution has the benefit that complexity is encapsulated into only two actions, passing monitors to the next owner when they should be release and conditionnaly waking threads if all conditions are met. Contrary to the other solutions, this solution quickly hits an upper bound on complexity of implementation.
     506        //... More code
     507release A
     508\end{pseudo}
     509\end{multicols}
     510The partial signalling solution transfers ownership of monitor B at lines 10 but does not wake the waiting thread since it is still using monitor A. Only when it reaches line 11 does it actually wakeup the waiting thread. This solution has the benefit that complexity is encapsulated into only two actions, passing monitors to the next owner when they should be release and conditionally waking threads if all conditions are met. This solution has a much simpler implementation than a dependency graph solving algorithm which is why it was chosen.
    523511
    524512% ======================================================================
    … …  
    529517An important note is that, until now, signalling a monitor was a delayed operation. The ownership of the monitor is transferred only when the monitor would have otherwise been released, not at the point of the \code{signal} statement. However, in some cases, it may be more convenient for users to immediately transfer ownership to the thread that is waiting for cooperation, which is achieved using the \code{signal_block} routine\footnote{name to be discussed}.
    530518
    531 For example here is an example highlighting the difference in behaviour:
    532 \begin{center}
     519The example in listing \ref{lst:datingservice} highlights the difference in behaviour. As mentioned, \code{signal} only transfers ownership once the current critical section exits, this behaviour cause the need for additional synchronisation when a two-way handshake is needed. To avoid this extraneous synchronisation, the \code{condition} type offers the \code{signal_block} routine which handle two-way handshakes as shown in the example. This removes the need for a second condition variables and simplifies programming. Like every other monitor semantic, \code{signal_block} uses barging prevention which means mutual-exclusion is baton-passed both on the frond-end and the back-end of the call to \code{signal_block}, meaning no other thread can acquire the monitor neither before nor after the call.
     520\begin{figure}
    533521\begin{tabular}{|c|c|}
    534522\code{signal} & \code{signal_block} \\
    535523\hline
    536 \begin{cfacode}
    537 monitor M { int val; };
    538 
    539 void foo(M & mutex m ) {
    540         m.val++;
    541         sout| "Foo:" | m.val |endl;
    542 
    543         wait( c );
    544 
    545         m.val++;
    546         sout| "Foo:" | m.val |endl;
    547 }
    548 
    549 void bar(M & mutex m ) {
    550         m.val++;
    551         sout| "Bar:" | m.val |endl;
    552 
    553         signal( c );
    554 
    555         m.val++;
    556         sout| "Bar:" | m.val |endl;
    557 }
    558 \end{cfacode}&\begin{cfacode}
    559 monitor M { int val; };
    560 
    561 void foo(M & mutex m ) {
    562         m.val++;
    563         sout| "Foo:" | m.val |endl;
    564 
    565         wait( c );
    566 
    567         m.val++;
    568         sout| "Foo:" | m.val |endl;
    569 }
    570 
    571 void bar(M & mutex m ) {
    572         m.val++;
    573         sout| "Bar:" | m.val |endl;
    574 
    575         signal_block( c );
    576 
    577         m.val++;
    578         sout| "Bar:" | m.val |endl;
     524\begin{cfacode}[tabsize=3]
     525monitor DatingService
     526{
     527        //compatibility codes
     528        enum{ CCodes = 20 };
     529
     530        int girlPhoneNo
     531        int boyPhoneNo;
     532};
     533
     534condition girls[CCodes];
     535condition boys [CCodes];
     536condition exchange;
     537
     538int girl(int phoneNo, int ccode)
     539{
     540        //no compatible boy ?
     541        if(empty(boys[ccode]))
     542        {
     543                //wait for boy
     544                wait(girls[ccode]);
     545
     546                //make phone number available
     547                girlPhoneNo = phoneNo;
     548
     549                //wake boy fron chair
     550                signal(exchange);
     551        }
     552        else
     553        {
     554                //make phone number available
     555                girlPhoneNo = phoneNo;
     556
     557                //wake boy
     558                signal(boys[ccode]);
     559
     560                //sit in chair
     561                wait(exchange);
     562        }
     563        return boyPhoneNo;
     564}
     565
     566int boy(int phoneNo, int ccode)
     567{
     568        //same as above
     569        //with boy/girl interchanged
     570}
     571\end{cfacode}&\begin{cfacode}[tabsize=3]
     572monitor DatingService
     573{
     574        //compatibility codes
     575        enum{ CCodes = 20 };
     576
     577        int girlPhoneNo;
     578        int boyPhoneNo;
     579};
     580
     581condition girls[CCodes];
     582condition boys [CCodes];
     583//exchange is not needed
     584
     585int girl(int phoneNo, int ccode)
     586{
     587        //no compatible boy ?
     588        if(empty(boys[ccode]))
     589        {
     590                //wait for boy
     591                wait(girls[ccode]);
     592
     593                //make phone number available
     594                girlPhoneNo = phoneNo;
     595
     596                //wake boy fron chair
     597                signal(exchange);
     598        }
     599        else
     600        {
     601                //make phone number available
     602                girlPhoneNo = phoneNo;
     603
     604                //wake boy
     605                signal_block(boys[ccode]);
     606
     607                //second handshake unnecessary
     608
     609        }
     610        return boyPhoneNo;
     611}
     612
     613int boy(int phoneNo, int ccode)
     614{
     615        //same as above
     616        //with boy/girl interchanged
    579617}
    580618\end{cfacode}
    581619\end{tabular}
    582 \end{center}
    583 Assuming that \code{val} is initialized at 0, that each routine are called from seperate thread and that \code{foo} is always called first. The previous code would yield the following output:
    584 
    585 \begin{center}
    586 \begin{tabular}{|c|c|}
    587 \code{signal} & \code{signal_block} \\
    588 \hline
    589 \begin{pseudo}
    590 Foo: 0
    591 Bar: 1
    592 Bar: 2
    593 Foo: 3
    594 \end{pseudo}&\begin{pseudo}
    595 Foo: 0
    596 Bar: 1
    597 Foo: 2
    598 Bar: 3
    599 \end{pseudo}
    600 \end{tabular}
    601 \end{center}
    602 
    603 As mentionned, \code{signal} only transfers ownership once the current critical section exits, resulting in the second "Bar" line to be printed before the second "Foo" line. On the other hand, \code{signal_block} immediately transfers ownership to \code{bar}, causing an inversion of output. Obviously this means that \code{signal_block} is a blocking call, which will only be resumed once the signalled function exits the critical section.
    604 
    605 % ======================================================================
    606 % ======================================================================
    607 \subsection{Internal scheduling: Implementation} \label{inschedimpl}
    608 % ======================================================================
    609 % ======================================================================
    610 There are several challenges specific to \CFA when implementing internal scheduling. These challenges are direct results of \gls{group-acquire} and loose object definitions. These two constraints are to root cause of most design decisions in the implementation of internal scheduling. Furthermore, to avoid the head-aches of dynamically allocating memory in a concurrent environment, the internal-scheduling design is entirely free of mallocs and other dynamic memory allocation scheme. This is to avoid the chicken and egg problem \cite{Chicken} of having a memory allocator that relies on the threading system and a threading system that relies on the runtime. This extra goal, means that memory management is a constant concern in the design of the system.
    611 
    612 The main memory concern for concurrency is queues. All blocking operations are made by parking threads onto queues. These queues need to be intrinsic\cit to avoid the need memory allocation. This entails that all the fields needed to keep track of all needed information. Since internal scheduling can use an unbound amount of memory (depending on \gls{group-acquire}) statically defining information information in the intrusive fields of threads is insufficient. The only variable sized container that does not require memory allocation is the callstack, which is heavily used in the implementation of internal scheduling. Particularly the GCC extension variable length arrays which is used extensively.
    613 
    614 Since stack allocation is based around scope, the first step of the implementation is to identify the scopes that are available to store the information, and which of these can have a variable length. In the case of external scheduling, the threads and the condition both allow a fixed amount of memory to be stored, while mutex-routines and the actual blocking call allow for an unbound amount (though adding too much to the mutex routine stack size can become expansive faster).
    615 
    616 The following figure is the traditionnal illustration of a monitor :
    617 
    618 \begin{center}
    619 {\resizebox{0.4\textwidth}{!}{\input{monitor}}}
    620 \end{center}
    621 
    622 For \CFA, the previous picture does not have support for blocking multiple monitors on a single condition. To support \gls{group-acquire} two changes to this picture are required. First, it doesn't make sense to tie the condition to a single monitor since blocking two monitors as one would require arbitrarily picking a monitor to hold the condition. Secondly, the object waiting on the conditions and AS-stack cannot simply contain the waiting thread since a single thread can potentially wait on multiple monitors. As mentionned in section \ref{inschedimpl}, the handling in multiple monitors is done by partially passing, which entails that each concerned monitor needs to have a node object. However, for waiting on the condition, since all threads need to wait together, a single object needs to be queued in the condition. Moving out the condition and updating the node types yields :
    623 
    624 \begin{center}
    625 {\resizebox{0.8\textwidth}{!}{\input{int_monitor}}}
    626 \end{center}
    627 
    628 \newpage
    629 
    630 This picture and the proper entry and leave algorithms is the fundamental implementation of internal scheduling.
    631 
    632 \begin{multicols}{2}
    633 Entry
    634 \begin{pseudo}[numbers=left]
    635 if monitor is free
    636         enter
    637 elif I already own the monitor
    638         continue
    639 else
    640         block
    641 increment recursion
    642 
    643 \end{pseudo}
    644 \columnbreak
    645 Exit
    646 \begin{pseudo}[numbers=left, firstnumber=8]
    647 decrement recursion
    648 if recursion == 0
    649         if signal_stack not empty
    650                 set_owner to thread
    651                 if all monitors ready
    652                         wake-up thread
    653 
    654         if entry queue not empty
    655                 wake-up thread
    656 \end{pseudo}
    657 \end{multicols}
    658 
    659 Some important things to notice about the exit routine. The solution discussed in \ref{inschedimpl} can be seen on line 11 of the previous pseudo code. Basically, the solution boils down to having a seperate data structure for the condition queue and the AS-stack, and unconditionally transferring ownership of the monitors but only unblocking the thread when the last monitor has trasnferred ownership. This solution is safe as well as preventing any potential barging.
     620\caption{Dating service example using \code{signal} and \code{signal_block}. }
     621\label{lst:datingservice}
     622\end{figure}
    660623
    661624% ======================================================================
    … …  
    700663\end{tabular}
    701664\end{center}
    702 This method is more constrained and explicit, which may help users tone down the undeterministic nature of concurrency. Indeed, as the following examples demonstrates, external scheduling allows users to wait for events from other threads without the concern of unrelated events occuring. External scheduling can generally be done either in terms of control flow (e.g., \uC) or in terms of data (e.g. Go). Of course, both of these paradigms have their own strenghts and weaknesses but for this project control-flow semantics were chosen to stay consistent with the rest of the languages semantics. Two challenges specific to \CFA arise when trying to add external scheduling with loose object definitions and multi-monitor routines. The previous example shows a simple use \code{_Accept} versus \code{wait}/\code{signal} and its advantages. Note that while other languages often use \code{accept} as the core external scheduling keyword, \CFA uses \code{waitfor} to prevent name collisions with existing socket \acrshort{api}s.
    703 
    704 In the case of internal scheduling, the call to \code{wait} only guarantees that \code{V} is the last routine to access the monitor. This entails that the routine \code{V} may have acquired mutual exclusion several times while routine \code{P} was waiting. On the other hand, external scheduling guarantees that while routine \code{P} was waiting, no routine other than \code{V} could acquire the monitor.
     665This method is more constrained and explicit, which helps users tone down the undeterministic nature of concurrency. Indeed, as the following examples demonstrates, external scheduling allows users to wait for events from other threads without the concern of unrelated events occuring. External scheduling can generally be done either in terms of control flow (e.g., \uC with \code{_Accept}) or in terms of data (e.g. Go with channels). Of course, both of these paradigms have their own strenghts and weaknesses but for this project control-flow semantics were chosen to stay consistent with the rest of the languages semantics. Two challenges specific to \CFA arise when trying to add external scheduling with loose object definitions and multi-monitor routines. The previous example shows a simple use \code{_Accept} versus \code{wait}/\code{signal} and its advantages. Note that while other languages often use \code{accept}/\code{select} as the core external scheduling keyword, \CFA uses \code{waitfor} to prevent name collisions with existing socket \acrshort{api}s.
     666
     667In the case of internal scheduling, the call to \code{wait} only guarantees that \code{V} is the last routine to access the monitor. This entails that a third routine, say \code{isInUse()}, may have acquired mutual exclusion several times while routine \code{P} was waiting. On the other hand, external scheduling guarantees that while routine \code{P} was waiting, no routine other than \code{V} could acquire the monitor.
    705668
    706669% ======================================================================
    … …  
    715678
    716679        void f(A & mutex a);
    717         void f(int a, float b);
    718680        void g(A & mutex a) {
    719                 waitfor(f); // Less obvious which f() to wait for
     681                waitfor(f); //Obvious which f() to wait for
     682        }
     683
     684        void f(A & mutex a, int); // New different F added in scope
     685        void h(A & mutex a) {
     686                waitfor(f); //Less obvious which f() to wait for
    720687        }
    721688\end{cfacode}
    … …  
    728695        if monitor is free
    729696                enter
    730         elif I already own the monitor
     697        elif already own the monitor
    731698                continue
    732699        elif monitor accepts me
    … …  
    738705\end{center}
    739706
    740 For the fist two conditions, it is easy to implement a check that can evaluate the condition in a few instruction. However, a fast check for \pscode{monitor accepts me} is much harder to implement depending on the constraints put on the monitors. Indeed, monitors are often expressed as an entry queue and some acceptor queue as in the following figure:
     707For the first two conditions, it is easy to implement a check that can evaluate the condition in a few instruction. However, a fast check for \pscode{monitor accepts me} is much harder to implement depending on the constraints put on the monitors. Indeed, monitors are often expressed as an entry queue and some acceptor queue as in the following figure:
    741708
    742709\begin{center}
    … …  
    744711\end{center}
    745712
    746 There are other alternatives to these pictures but in the case of this picture implementing a fast accept check is relatively easy. Indeed simply updating a bitmask when the acceptor queue changes is enough to have a check that executes in a single instruction, even with a fairly large number (e.g. 128) of mutex members. This technique cannot be used in \CFA because it relies on the fact that the monitor type declares all the acceptable routines. For OO languages this does not compromise much since monitors already have an exhaustive list of member routines. However, for \CFA this is not the case; routines can be added to a type anywhere after its declaration. Its important to note that the bitmask approach does not actually require an exhaustive list of routines, but it requires a dense unique ordering of routines with an upper-bound and that ordering must be consistent across translation units.
    747 The alternative would be to have a picture more like this one:
     713There are other alternatives to these pictures, but in the case of this picture, implementing a fast accept check is relatively easy. Indeed simply updating a bitmask when the acceptor queue changes is enough to have a check that executes in a single instruction, even with a fairly large number (e.g. 128) of mutex members. This technique cannot be used in \CFA because it relies on the fact that the monitor type declares all the acceptable routines. For OO languages this does not compromise much since monitors already have an exhaustive list of member routines. However, for \CFA this is not the case; routines can be added to a type anywhere after its declaration. Its important to note that the bitmask approach does not actually require an exhaustive list of routines, but it requires a dense unique ordering of routines with an upper-bound and that ordering must be consistent across translation units.
     714The alternative is to have a picture like this one:
    748715
    749716\begin{center}
    … …  
    751718\end{center}
    752719
    753 Not storing the queues inside the monitor means that the storage can vary between routines, allowing for more flexibility and extensions. Storing an array of function-pointers would solve the issue of uniquely identifying acceptable routines. However, the single instruction bitmask compare has been replaced by dereferencing a pointer followed by a linear search. Furthermore, supporting nested external scheduling may now require additionnal searches on calls to waitfor to check if a routine is already queued in.
    754 
    755 At this point we must make a decision between flexibility and performance. Many design decisions in \CFA achieve both flexibility and performance, for example polymorphic routines add significant flexibility but inlining them means the optimizer can easily remove any runtime cost. Here however, the cost of flexibility cannot be trivially removed. In the end, the most flexible approach has been chosen since it allows users to write programs that would otherwise be prohibitively hard to write. This is based on the assumption that writing fast but inflexible locks is closer to a solved problems than writing locks that are as flexible as external scheduling in \CFA.
    756 
    757 Another aspect to consider is what happens if multiple overloads of the same routine are used. For the time being it is assumed that multiple overloads of the same routine are considered as distinct routines. However, this could easily be extended in the future.
     720Not storing the mask inside the monitor means that the storage for the mask information can vary between calls to \code{waitfor}, allowing for more flexibility and extensions. Storing an array of function-pointers would solve the issue of uniquely identifying acceptable routines. However, the single instruction bitmask compare has been replaced by dereferencing a pointer followed by a linear search. Furthermore, supporting nested external scheduling may now require additionnal searches on calls to waitfor to check if a routine is already queued in.
     721
     722Note that in the second picture, tasks need to always keep track of through which routine they are attempting to acquire the monitor and the routine mask needs to have both a function pointer and a set of monitors, as will be discussed in the next section. These details where omitted from the picture for the sake of simplifying the representation.
     723
     724At this point we must make a decision between flexibility and performance. Many design decisions in \CFA achieve both flexibility and performance, for example polymorphic routines add significant flexibility but inlining them means the optimizer can easily remove any runtime cost. Here however, the cost of flexibility cannot be trivially removed. In the end, the most flexible approach has been chosen since it allows users to write programs that would otherwise be prohibitively hard to write. This decision is based on the assumption that writing fast but inflexible locks is closer to a solved problems than writing locks that are as flexible as external scheduling in \CFA.
    758725
    759726% ======================================================================
    … …  
    763730% ======================================================================
    764731
    765 External scheduling, like internal scheduling, becomes orders of magnitude more complex when we start introducing multi-monitor syntax. Even in the simplest possible case some new semantics need to be established:
    766 \begin{cfacode}
    767         mutex struct A {};
    768 
    769         mutex struct B {};
    770 
    771         void g(A & mutex a, B & mutex b) {
    772                 waitfor(f); //ambiguous, which monitor
     732External scheduling, like internal scheduling, becomes significantly more complex when introducing multi-monitor syntax. Even in the simplest possible case, some new semantics need to be established:
     733\begin{cfacode}
     734        monitor M {};
     735
     736        void f(M & mutex a);
     737
     738        void g(M & mutex a, M & mutex b) {
     739                waitfor(f); //ambiguous, keep a pass b or other way around?
    773740        }
    774741\end{cfacode}
    … …  
    777744
    778745\begin{cfacode}
    779         mutex struct A {};
    780 
    781         mutex struct B {};
    782 
    783         void g(A & mutex a, B & mutex b) {
     746        monitor M {};
     747
     748        void f(M & mutex a);
     749
     750        void g(M & mutex a, M & mutex b) {
    784751                waitfor( f, b );
    785752        }
    786753\end{cfacode}
    787754
    788 This is unambiguous. Both locks will be acquired and kept, when routine \code{f} is called the lock for monitor \code{b} will be temporarily transferred from \code{g} to \code{f} (while \code{g} still holds lock \code{a}). This behavior can be extended to multi-monitor waitfor statment as follows.
    789 
    790 \begin{cfacode}
    791         mutex struct A {};
    792 
    793         mutex struct B {};
    794 
    795         void g(A & mutex a, B & mutex b) {
     755This syntax is unambiguous. Both locks are acquired and kept. When routine \code{f} is called, the lock for monitor \code{b} is temporarily transferred from \code{g} to \code{f} (while \code{g} still holds lock \code{a}). This behavior can be extended to multi-monitor waitfor statment as follows.
     756
     757\begin{cfacode}
     758        monitor M {};
     759
     760        void f(M & mutex a, M & mutex b);
     761
     762        void g(M & mutex a, M & mutex b) {
    796763                waitfor( f, a, b);
    797764        }
    798765\end{cfacode}
    799766
    800 Note that the set of monitors passed to the \code{waitfor} statement must be entirely contained in the set of monitor already acquired in the routine. \code{waitfor} used in any other context is Undefined Behaviour.
    801 
    802 An important behavior to note is that what happens when set of monitors only match partially :
     767Note that the set of monitors passed to the \code{waitfor} statement must be entirely contained in the set of monitors already acquired in the routine. \code{waitfor} used in any other context is Undefined Behaviour.
     768
     769An important behavior to note is that what happens when a set of monitors only match partially :
    803770
    804771\begin{cfacode}
    … …  
    815782
    816783        void foo() {
    817                 g(a1, b);
     784                g(a1, b); //block on accept
    818785        }
    819786
    820787        void bar() {
    821                 f(a2, b);
    822         }
    823 \end{cfacode}
    824 
    825 While the equivalent can happen when using internal scheduling, the fact that conditions are branded on first use means that users have to use two different condition variables. In both cases, partially matching monitor sets will not wake-up the waiting thread. It is also important to note that in the case of external scheduling, as for routine calls, the order of parameters is important; \code{waitfor(f,a,b)} and \code{waitfor(f,b,a)} are to distinct waiting condition.
    826 
    827 % ======================================================================
    828 % ======================================================================
    829 \subsection{Implementation Details: External scheduling queues}
    830 % ======================================================================
    831 % ======================================================================
    832 To support multi-monitor external scheduling means that some kind of entry-queues must be used that is aware of both monitors. However, acceptable routines must be aware of the entry queues which means they must be stored inside at least one of the monitors that will be acquired. This in turn adds the requirement a systematic algorithm of disambiguating which queue is relavant regardless of user ordering. The proposed algorithm is to fall back on monitors lock ordering and specify that the monitor that is acquired first is the lock with the relevant entry queue. This assumes that the lock acquiring order is static for the lifetime of all concerned objects but that is a reasonable constraint. This algorithm choice has two consequences, the entry queue of the highest priority monitor is no longer a true FIFO queue and the queue of the lowest priority monitor is both required and probably unused. The queue can no longer be a FIFO queue because instead of simply containing the waiting threads in order arrival, they also contain the second mutex. Therefore, another thread with the same highest priority monitor but a different lowest priority monitor may arrive first but enter the critical section after a thread with the correct pairing. Secondly, since it may not be known at compile time which monitor will be the lowest priority monitor, every monitor needs to have the correct queues even though it is probable that half the multi-monitor queues will go unused for the entire duration of the program.
    833 
    834 
    835 \subsection{Internals}
    836 The complete mask can be pushed to any one, we are in a context where we already have full ownership of (at least) every concerned monitor and therefore monitors will refuse all calls no matter what.
     788                f(a2, b); //fufill cooperation
     789        }
     790\end{cfacode}
     791
     792While the equivalent can happen when using internal scheduling, the fact that conditions are specific to a set of monitors means that users have to use two different condition variables. In both cases, partially matching monitor sets does not wake-up the waiting thread. It is also important to note that in the case of external scheduling, as for routine calls, the order of parameters is important; \code{waitfor(f,a,b)} and \code{waitfor(f,b,a)} are to distinct waiting condition.
     793
     794% ======================================================================
     795% ======================================================================
     796\subsection{Waitfor semantics}
     797% ======================================================================
     798% ======================================================================

  • doc/proposals/concurrency/text/internals.tex

    rb7778c1 r3364962  
    11
    22\chapter{Behind the scene}
     3
     4
     5% ======================================================================
     6% ======================================================================
     7\section{Implementation Details: Interaction with polymorphism}
     8% ======================================================================
     9% ======================================================================
     10Depending on the choice of semantics for when monitor locks are acquired, interaction between monitors and \CFA's concept of polymorphism can be complex to support. However, it is shown that entry-point locking solves most of the issues.
     11
     12First of all, interaction between \code{otype} polymorphism and monitors is impossible since monitors do not support copying. Therefore, the main question is how to support \code{dtype} polymorphism. Since a monitor's main purpose is to ensure mutual exclusion when accessing shared data, this implies that mutual exclusion is only required for routines that do in fact access shared data. However, since \code{dtype} polymorphism always handles incomplete types (by definition), no \code{dtype} polymorphic routine can access shared data since the data requires knowledge about the type. Therefore, the only concern when combining \code{dtype} polymorphism and monitors is to protect access to routines.
     13
     14Before looking into complex control-flow, it is important to present the difference between the two acquiring options : callsite and entry-point locking, i.e. acquiring the monitors before making a mutex routine call or as the first operation of the mutex routine-call. For example:
     15\begin{figure}
     16\label{fig:locking-site}
     17\begin{center}
     18\setlength\tabcolsep{1.5pt}
     19\begin{tabular}{|c|c|c|}
     20Mutex & \gls{callsite-locking} & \gls{entry-point-locking} \\
     21call & pseudo-code & pseudo-code \\
     22\hline
     23\begin{cfacode}[tabsize=3]
     24void foo(monitor& mutex a){
     25
     26        //Do Work
     27        //...
     28
     29}
     30
     31void main() {
     32        monitor a;
     33
     34        foo(a);
     35
     36}
     37\end{cfacode} & \begin{pseudo}[tabsize=3]
     38foo(& a) {
     39
     40        //Do Work
     41        //...
     42
     43}
     44
     45main() {
     46        monitor a;
     47        acquire(a);
     48        foo(a);
     49        release(a);
     50}
     51\end{pseudo} & \begin{pseudo}[tabsize=3]
     52foo(& a) {
     53        acquire(a);
     54        //Do Work
     55        //...
     56        release(a);
     57}
     58
     59main() {
     60        monitor a;
     61
     62        foo(a);
     63
     64}
     65\end{pseudo}
     66\end{tabular}
     67\end{center}
     68\caption{Callsite vs entry-point locking for mutex calls}
     69\end{figure}
     70
     71
     72Note the \code{mutex} keyword relies on the type system, which means that in cases where a generic monitor routine is actually desired, writing a mutex routine is possible with the proper trait, which is possible because monitors are designed in terms a trait. For example:
     73\begin{cfacode}
     74//Incorrect: T is not a monitor
     75forall(dtype T)
     76void foo(T * mutex t);
     77
     78//Correct: this function only works on monitors (any monitor)
     79forall(dtype T | is_monitor(T))
     80void bar(T * mutex t));
     81\end{cfacode}
     82
     83
     84% ======================================================================
     85% ======================================================================
     86\section{Internal scheduling: Implementation} \label{inschedimpl}
     87% ======================================================================
     88% ======================================================================
     89There are several challenges specific to \CFA when implementing internal scheduling. These challenges are direct results of \gls{bulk-acq} and loose object definitions. These two constraints are to root cause of most design decisions in the implementation of internal scheduling. Furthermore, to avoid the head-aches of dynamically allocating memory in a concurrent environment, the internal-scheduling design is entirely free of mallocs and other dynamic memory allocation scheme. This is to avoid the chicken and egg problem \cite{Chicken} of having a memory allocator that relies on the threading system and a threading system that relies on the runtime. This extra goal, means that memory management is a constant concern in the design of the system.
     90
     91The main memory concern for concurrency is queues. All blocking operations are made by parking threads onto queues. These queues need to be intrinsic\cit to avoid the need memory allocation. This entails that all the fields needed to keep track of all needed information. Since internal scheduling can use an unbound amount of memory (depending on \gls{bulk-acq}) statically defining information information in the intrusive fields of threads is insufficient. The only variable sized container that does not require memory allocation is the callstack, which is heavily used in the implementation of internal scheduling. Particularly the GCC extension variable length arrays which is used extensively.
     92
     93Since stack allocation is based around scope, the first step of the implementation is to identify the scopes that are available to store the information, and which of these can have a variable length. In the case of external scheduling, the threads and the condition both allow a fixed amount of memory to be stored, while mutex-routines and the actual blocking call allow for an unbound amount (though adding too much to the mutex routine stack size can become expansive faster).
     94
     95The following figure is the traditionnal illustration of a monitor :
     96
     97\begin{center}
     98{\resizebox{0.4\textwidth}{!}{\input{monitor}}}
     99\end{center}
     100
     101For \CFA, the previous picture does not have support for blocking multiple monitors on a single condition. To support \gls{bulk-acq} two changes to this picture are required. First, it doesn't make sense to tie the condition to a single monitor since blocking two monitors as one would require arbitrarily picking a monitor to hold the condition. Secondly, the object waiting on the conditions and AS-stack cannot simply contain the waiting thread since a single thread can potentially wait on multiple monitors. As mentionned in section \ref{inschedimpl}, the handling in multiple monitors is done by partially passing, which entails that each concerned monitor needs to have a node object. However, for waiting on the condition, since all threads need to wait together, a single object needs to be queued in the condition. Moving out the condition and updating the node types yields :
     102
     103\begin{center}
     104{\resizebox{0.8\textwidth}{!}{\input{int_monitor}}}
     105\end{center}
     106
     107\newpage
     108
     109This picture and the proper entry and leave algorithms is the fundamental implementation of internal scheduling.
     110
     111\begin{multicols}{2}
     112Entry
     113\begin{pseudo}[numbers=left]
     114if monitor is free
     115        enter
     116elif I already own the monitor
     117        continue
     118else
     119        block
     120increment recursion
     121
     122\end{pseudo}
     123\columnbreak
     124Exit
     125\begin{pseudo}[numbers=left, firstnumber=8]
     126decrement recursion
     127if recursion == 0
     128        if signal_stack not empty
     129                set_owner to thread
     130                if all monitors ready
     131                        wake-up thread
     132
     133        if entry queue not empty
     134                wake-up thread
     135\end{pseudo}
     136\end{multicols}
     137
     138Some important things to notice about the exit routine. The solution discussed in \ref{inschedimpl} can be seen on line 11 of the previous pseudo code. Basically, the solution boils down to having a seperate data structure for the condition queue and the AS-stack, and unconditionally transferring ownership of the monitors but only unblocking the thread when the last monitor has trasnferred ownership. This solution is safe as well as preventing any potential barging.
     139
     140% ======================================================================
     141% ======================================================================
     142\section{Implementation Details: External scheduling queues}
     143% ======================================================================
     144% ======================================================================
     145To support multi-monitor external scheduling means that some kind of entry-queues must be used that is aware of both monitors. However, acceptable routines must be aware of the entry queues which means they must be stored inside at least one of the monitors that will be acquired. This in turn adds the requirement a systematic algorithm of disambiguating which queue is relavant regardless of user ordering. The proposed algorithm is to fall back on monitors lock ordering and specify that the monitor that is acquired first is the lock with the relevant entry queue. This assumes that the lock acquiring order is static for the lifetime of all concerned objects but that is a reasonable constraint. This algorithm choice has two consequences, the entry queue of the highest priority monitor is no longer a true FIFO queue and the queue of the lowest priority monitor is both required and probably unused. The queue can no longer be a FIFO queue because instead of simply containing the waiting threads in order arrival, they also contain the second mutex. Therefore, another thread with the same highest priority monitor but a different lowest priority monitor may arrive first but enter the critical section after a thread with the correct pairing. Secondly, since it may not be known at compile time which monitor will be the lowest priority monitor, every monitor needs to have the correct queues even though it is probable that half the multi-monitor queues will go unused for the entire duration of the program.
     146
     147
     148\section{Internals}
     149The complete mask can be pushed to any one, we are in a context where we already have full ownership of (at least) every concerned monitor and therefore monitors will refuse all calls no matter what.

  • doc/proposals/concurrency/text/parallelism.tex

    rb7778c1 r3364962  
    2121
    2222\subsection{Jobs and thread pools}
    23 The approach on the opposite end of the spectrum is to base parallelism on \glspl{pool}. Indeed, \glspl{pool} offer limited flexibility but at the benefit of a simpler user interface. In \gls{pool} based systems, users express parallelism as units of work, called jobs, and a dependency graph (either explicit or implicit) that tie them together. This approach means users need not worry about concurrency but significantly limits the interaction that can occur among jobs. Indeed, any \gls{job} that blocks also blocks the underlying worker, which effectively means the CPU utilization, and therefore throughput, suffers noticeably. It can be argued that a solution to this problem is to use more workers than available cores. However, unless the number of jobs and the number of workers are comparable, having a significant amount of blocked jobs always results in idles cores.
     23An approach on the opposite end of the spectrum is to base parallelism on \glspl{pool}. Indeed, \glspl{pool} offer limited flexibility but at the benefit of a simpler user interface. In \gls{pool} based systems, users express parallelism as units of work, called jobs, and a dependency graph (either explicit or implicit) that tie them together. This approach means users need not worry about concurrency but significantly limit the interaction that can occur among jobs. Indeed, any \gls{job} that blocks also blocks the underlying worker, which effectively means the CPU utilization, and therefore throughput, suffers noticeably. It can be argued that a solution to this problem is to use more workers than available cores. However, unless the number of jobs and the number of workers are comparable, having a significant amount of blocked jobs always results in idles cores.
    2424
    2525The gold standard of this implementation is Intel's TBB library~\cite{TBB}.
    2626
    2727\subsection{Paradigm performance}
    28 While the choice between the three paradigms listed above may have significant performance implication, it is difficult to pindown the performance implications of chosing a model at the language level. Indeed, in many situations one of these paradigms may show better performance but it all strongly depends on the workload. Having a large amount of mostly independent units of work to execute almost guarantess that the \gls{pool} based system has the best performance thanks to the lower memory overhead (i.e., not thread stack per job). However, interactions among jobs can easily exacerbate contention. User-level threads allow fine-grain context switching, which results in better resource utilisation, but context switches is more expansive and the extra control means users need to tweak more variables to get the desired performance. Finally, if the units of uninterrupted work are large enough the paradigm choice is largely amortised by the actual work done.
     28While the choice between the three paradigms listed above may have significant performance implication, it is difficult to pindown the performance implications of chosing a model at the language level. Indeed, in many situations one of these paradigms may show better performance but it all strongly depends on the workload. Having a large amount of mostly independent units of work to execute almost guarantess that the \gls{pool} based system has the best performance thanks to the lower memory overhead (i.e., not thread stack per job). However, interactions among jobs can easily exacerbate contention. User-level threads allow fine-grain context switching, which results in better resource utilisation, but a context switch is more expensive and the extra control means users need to tweak more variables to get the desired performance. Finally, if the units of uninterrupted work are large enough the paradigm choice is largely amortised by the actual work done.
    2929
    3030\TODO
    … …  
    3333
    3434
    35 \subsubsection{Future Work: Machine setup}\label{machine}
    36 While this was not done in the context of this proposal, another important aspect of clusters is affinity. While many common desktop and laptop PCs have homogeneous CPUs, other devices often have more heteregenous setups. For example, system using \acrshort{numa} configurations may benefit from users being able to tie clusters and/or kernel threads to certains CPU cores. OS support for CPU affinity is now common \cit, which means it is both possible and desirable for \CFA to offer an abstraction mechanism for portable CPU affinity.
     35\subsection{Future Work: Machine setup}\label{machine}
     36While this was not done in the context of this thesis, another important aspect of clusters is affinity. While many common desktop and laptop PCs have homogeneous CPUs, other devices often have more heteregenous setups. For example, system using \acrshort{numa} configurations may benefit from users being able to tie clusters and/or kernel threads to certains CPU cores. OS support for CPU affinity is now common \cit, which means it is both possible and desirable for \CFA to offer an abstraction mechanism for portable CPU affinity.
    3737
    3838\subsection{Paradigms}\label{cfaparadigms}

  • doc/proposals/concurrency/text/together.tex

    rb7778c1 r3364962  
    77
    88\section{Threads as monitors}
    9 As it was sbtely alluded in section \ref{threads}, \code{threads} in \CFA are in factor monitors. This means that all the monitors features are available when using threads. For example, here is a very simple two thread pipeline that could be used for a simulator of a game engine :
     9As it was subtely alluded in section \ref{threads}, \code{threads} in \CFA are in fact monitors. This means that all the monitors features are available when using threads. For example, here is a very simple two thread pipeline that could be used for a simulator of a game engine :
    1010\begin{cfacode}
    1111// Visualization declaration
Note: See TracChangeset for help on using the changeset viewer.