Diff [0fe4e6265f263a0fedaac815f2ed782fe74dabdd:f5c3b6cd1ad780487cd91b5ea931a57c5158864d] for / – Cforall

Jenkinsfile

-              r0fe4e62
+              rf5c3b6c
         arch_name               = ''
         architecture    = ''
         do_alltests             = false
         do_benchmark    = false
 …
                 sh 'make clean > /dev/null'
                 sh 'make > /dev/null 2>&1'
+        }
+        }
         catch (Exception caughtError) {
                 err = caughtError //rethrow error later
 …
 def build() {
         build_stage('Build') {
                 def install_dir = pwd tmp: true
                 //Configure the conpilation (Output is not relevant)
                 //Use the current directory as the installation target so nothing
 …
                 if( !do_benchmark ) return
+                //Write the commit id to Benchmark
+                writeFile  file: 'bench.csv', text:'data=' + gitRefNewValue + ',' + arch_name + ','
                 //Append bench results
                 sh 'make -C src/benchmark --no-print-directory jenkins githash=' + gitRefNewValue + ' arch=' + arch_name + ' | tee bench.json'
+                sh 'make -C src/benchmark --no-print-directory csv-data >> bench.csv'
+        }
+}
 …
                 //Then publish the results
                 sh 'curl -H "Content-Type: application/json" --silent --data @bench.json http://plg2:8082/jenkins/publish > /dev/null || true'
+                sh 'curl --silent --data @bench.csv http://plg2:8082/jenkins/publish > /dev/null || true'
+        }
+}

doc/proposals/concurrency/.gitignore

r0fe4e62	rf5c3b6c
16	16	build/*.out
17	17	build/*.ps
18		~~build/*.pstex~~
19		~~build/*.pstex_t~~
20	18	build/*.tex
21	19	build/*.toc

doc/proposals/concurrency/Makefile

-              r0fe4e62
+              rf5c3b6c
 annex/glossary \
 text/intro \
+text/cforall \
 text/basics \
-text/cforall \
 text/concurrency \
 text/internals \
 text/parallelism \
-text/results \
 text/together \
 text/future \
 …
 }}
+PICTURES = ${addprefix build/, ${addsuffix .pstex, \
+        system \
+        monitor_structs \
+}}
+PICTURES = ${addsuffix .pstex, \
+}
 PROGRAMS = ${addsuffix .tex, \
 …
         build/*.out     \
         build/*.ps      \
-        build/*.pstex   \
         build/*.pstex_t \
         build/*.tex     \
 …
         dvips $< -o $@
 build/${basename ${DOCUMENT}}.dvi : Makefile ${GRAPHS} ${PROGRAMS} ${PICTURES} ${FIGURES} ${SOURCES} ${basename ${DOCUMENT}}.tex ../../LaTeXmacros/common.tex ../../LaTeXmacros/indexstyle annex/local.bib
+build/${basename ${DOCUMENT}}.dvi : Makefile ${GRAPHS} ${PROGRAMS} ${PICTURES} ${FIGURES} ${SOURCES} ${basename ${DOCUMENT}}.tex ../../LaTeXmacros/common.tex ../../LaTeXmacros/indexstyle
         @ if [ ! -r ${basename $@}.ind ] ; then touch ${basename $@}.ind ; fi                           # Conditionally create an empty *.ind (index) file for inclusion until makeindex is run.
 …
         @ -${BibTeX} ${basename $@}
         @ echo "Glossary"
         @ makeglossaries -q -s ${basename $@}.ist ${basename $@}                                                # Make index from *.aux entries and input index at end of document
+        makeglossaries -q -s ${basename $@}.ist ${basename $@}                                          # Make index from *.aux entries and input index at end of document
         @ echo ".dvi generation"
         @ -build/bump_ver.sh

doc/proposals/concurrency/annex/local.bib

-              r0fe4e62
+              rf5c3b6c
         year            = 2017
+}
-@manual{Cpp-Transactions,
-        keywords        = {C++, Transactional Memory},
-        title           = {Technical Specification for C++ Extensions for Transactional Memory},
-        organization= {International Standard ISO/IEC TS 19841:2015 },
-        publisher   = {American National Standards Institute},
-        address = {http://www.iso.org},
-        year            = 2015,
+}
-@article{BankTransfer,
-        keywords        = {Bank Transfer},
-        title   = {Bank Account Transfer Problem},
-        publisher       = {Wiki Wiki Web},
-        address = {http://wiki.c2.com},
-        year            = 2010
+}
-@misc{2FTwoHardThings,
-        keywords        = {Hard Problem},
-        title   = {TwoHardThings},
-        author  = {Martin Fowler},
-        address = {https://martinfowler.com/bliki/TwoHardThings.html},
-        year            = 2009
+}
-@article{IntrusiveData,
-        title           = {Intrusive Data Structures},
-        author  = {Jiri Soukup},
-        journal = {CppReport},
-        year            = 1998,
-        month           = May,
-        volume  = {10/No5.},
-        page            = 22
+}
-@misc{affinityLinux,
-        title           = "{Linux man page - sched\_setaffinity(2)}"
+}
-@misc{affinityWindows,
-        title           = "{Windows (vs.85) - SetThreadAffinityMask function}"
+}
-@misc{affinityFreebsd,
-        title           = "{FreeBSD General Commands Manual - CPUSET(1)}"
+}
-@misc{affinityNetbsd,
-        title           = "{NetBSD Library Functions Manual - AFFINITY(3)}"
+}
-@misc{affinityMacosx,
-        title           = "{Affinity API Release Notes for OS X v10.5}"
+}

doc/proposals/concurrency/figures/int_monitor.fig

-              r0fe4e62
+              rf5c3b6c
 -2
 2
+1 0 1 0 7 50 -1 -1 0.000 0 1 0 0 675.000 2700.000 675 2400 375 2700 675 3000
+4533 2866 4655 3129
+1 0 1 0 7 50 -1 -1 0.000 0 1 0 0 4657.017 2997.000 4655 2873 4533 2997 4655 3121
+1 0 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 2
+2866 4655 3129
+1 0 1 0 7 50 -1 -1 0.000 0 1 0 0 600.000 2625.000 600 2325 300 2625 600 2925
+3225 4500 7425 4800
+3 0 1 -1 -1 0 0 20 0.000 1 0.0000 3375 4650 80 80 3375 4650 3455 4730
+3 0 1 -1 -1 0 0 -1 0.000 1 0.0000 4725 4650 105 105 4725 4650 4830 4755
+3 0 1 -1 -1 0 0 4 0.000 1 0.0000 6225 4650 105 105 6225 4650 6330 4755
+0 -1 0 0 0 12 0.0000 2 135 1035 4950 4725 blocked task\001
+0 -1 0 0 0 12 0.0000 2 135 870 3525 4725 active task\001
+0 -1 0 0 0 12 0.0000 2 180 930 6450 4725 routine ptrs\001
 -6
+4725 2866 4847 3129
+1 0 1 0 7 50 -1 -1 0.000 0 1 0 0 4849.017 2997.000 4847 2873 4725 2997 4847 3121
+1 0 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 2
+2866 4847 3129
+8445 1695 8655 1905
+3 0 1 -1 -1 0 0 -1 0.000 1 0.0000 8550 1800 105 105 8550 1800 8655 1905
+1 -1 0 0 0 10 0.0000 2 75 75 8550 1860 a\001
 -6
+4911 2866 5033 3129
+1 0 1 0 7 50 -1 -1 0.000 0 1 0 0 5035.017 2997.000 5033 2873 4911 2997 5033 3121
+1 0 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 2
+2866 5033 3129
+8445 1395 8655 1605
+3 0 1 -1 -1 0 0 -1 0.000 1 0.0000 8550 1500 105 105 8550 1500 8655 1605
+1 -1 0 0 0 10 0.0000 2 105 90 8550 1560 b\001
 -6
+9027 2866 9149 3129
+1 0 1 0 7 50 -1 -1 0.000 0 0 0 0 9024.983 2997.000 9027 2873 9149 2997 9027 3121
+1 0 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 2
+2866 9027 3129
+3945 1695 4155 1905
+3 0 1 -1 -1 0 0 -1 0.000 1 0.0000 4050 1800 105 105 4050 1800 4155 1905
+1 -1 0 0 0 10 0.0000 2 75 75 4050 1860 a\001
 -6
+9253 2866 9375 3129
+1 0 1 0 7 50 -1 -1 0.000 0 0 0 0 9250.983 2997.000 9253 2873 9375 2997 9253 3121
+1 0 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 2
+2866 9253 3129
+-6
+9478 2866 9600 3129
+1 0 1 0 7 50 -1 -1 0.000 0 0 0 0 9475.983 2997.000 9478 2873 9600 2997 9478 3121
+1 0 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 2
+2866 9478 3129
+3945 1395 4155 1605
+3 0 1 -1 -1 0 0 -1 0.000 1 0.0000 4050 1500 105 105 4050 1500 4155 1605
+1 -1 0 0 0 10 0.0000 2 105 90 4050 1560 b\001
 -6
 3 0 1 -1 -1 0 0 20 0.000 1 0.0000 7650 3675 80 80 7650 3675 7730 3755
 3 0 1 -1 -1 0 0 20 0.000 1 0.0000 3150 3675 80 80 3150 3675 3230 3755
-3 0 1 0 7 50 -1 -1 0.000 1 0.0000 4047 1793 125 125 4047 1793 3929 1752
-3 0 1 0 7 50 -1 -1 0.000 1 0.0000 4050 1500 125 125 4050 1500 3932 1459
-3 0 1 0 7 50 -1 -1 0.000 1 0.0000 8550 1500 125 125 8550 1500 8432 1459
-3 0 1 0 7 50 -1 -1 0.000 1 0.0000 8550 1800 125 125 8550 1800 8432 1759
-3 0 1 0 7 50 -1 -1 0.000 1 0.0000 1200 2850 125 125 1200 2850 1082 2809
-3 0 1 0 7 50 -1 -1 0.000 1 0.0000 900 2850 125 125 900 2850 782 2809
-3 0 1 -1 -1 0 0 4 0.000 1 0.0000 6225 4650 105 105 6225 4650 6330 4755
-3 0 1 -1 -1 0 0 20 0.000 1 0.0000 3150 4650 80 80 3150 4650 3230 4730
-3 0 1 -1 -1 0 0 -1 0.000 1 0.0000 4575 4650 105 105 4575 4650 4680 4755
 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 2
 1950 4200 2100
 …
 4050 3300 4200
 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 2
 3000 1425 3000
+2925 1350 2925
 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 2
 2400 1425 2400
+2325 1350 2325
 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 2
 2700 1500 2925
+2625 1425 2850
 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 2
 2400 1350 2625
+2325 1275 2550
 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 2
+2700 1425 2700
+2625 1350 2625
+3 0 1 0 7 50 -1 -1 0.000 0 0 0 0 0 7
+2775 1275 2645 1125 2645 1050 2775 1125 2905 1275 2905
+2775
+3 0 1 0 7 50 -1 -1 0.000 0 0 0 0 0 7
+2775 900 2645 750 2645 675 2775 750 2905 900 2905
+2775
+3 0 1 0 7 50 -1 -1 0.000 0 0 0 0 0 7
+3000 4725 2870 4575 2870 4500 3000 4575 3130 4725 3130
+3000
+3 0 1 0 7 50 -1 -1 0.000 0 0 0 0 0 7
+3000 5025 2870 4875 2870 4800 3000 4875 3130 5025 3130
+3000
+3 0 1 0 7 50 -1 -1 0.000 0 0 0 0 0 7
+3000 9225 2870 9075 2870 9000 3000 9075 3130 9225 3130
+3000
+3 0 1 0 7 50 -1 -1 0.000 0 0 0 0 0 7
+3000 9525 2870 9375 2870 9300 3000 9375 3130 9525 3130
+3000
+1 0 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 2
+2775 975 2775
+1 0 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 2
+2775 1350 2775
+3 0 1 0 7 50 -1 -1 0.000 0 0 0 0 0 7
+4950 4800 4820 4650 4820 4575 4950 4650 5080 4800 5080
+4950
+1 0 1 0 7 50 -1 -1 0.000 0 0 -1 0 0 2
+4950 4875 4950
+3 0 1 0 7 50 -1 -1 0.000 0 0 0 0 0 7
+4970 3450 4840 3300 4840 3225 4970 3300 5100 3450 5100
+4970
 1 -1 0 0 0 12 0.0000 2 135 315 2850 4275 exit\001
 1 -1 0 0 0 12 0.0000 2 135 315 7350 4275 exit\001
 …
 1 -1 0 0 0 12 0.0000 2 135 495 4050 1275 queue\001
 1 -1 0 0 0 12 0.0000 2 165 420 4050 1050 entry\001
+0 0 50 -1 0 11 0.0000 2 120 705 600 2325 Condition\001
+0 -1 0 0 0 12 0.0000 2 180 930 6450 4725 routine ptrs\001
+0 -1 0 0 0 12 0.0000 2 135 1050 3300 4725 active thread\001
+0 -1 0 0 0 12 0.0000 2 135 1215 4725 4725 blocked thread\001
+0 0 50 -1 0 11 0.0000 2 120 705 450 2250 Condition\001
+0 0 50 -1 0 11 0.0000 2 165 630 3600 5025 signalled\001
+0 0 50 -1 0 11 0.0000 2 165 525 4950 5025 waiting\001

doc/proposals/concurrency/style/cfa-format.tex

-              r0fe4e62
+              rf5c3b6c
 }{}
-\lstnewenvironment{gocode}[1][]{
-  \lstset{
-    language = Golang,
-    style=defaultStyle,
-    #1
+  }
-}{}
 \newcommand{\zero}{\lstinline{zero_t}\xspace}
 \newcommand{\one}{\lstinline{one_t}\xspace}

doc/proposals/concurrency/text/basics.tex

-              r0fe4e62
+              rf5c3b6c
 At its core, concurrency is based on having multiple call-stacks and scheduling among threads of execution executing on these stacks. Concurrency without parallelism only requires having multiple call stacks (or contexts) for a single thread of execution.
+Execution with a single thread and multiple stacks where the thread is self-scheduling deterministically across the stacks is called coroutining. Execution with a single and multiple stacks but where the thread is scheduled by an oracle (non-deterministic from the thread perspective) across the stacks is called concurrency.
+Therefore, a minimal concurrency system can be achieved by creating coroutines, which instead of context switching among each other, always ask an oracle where to context switch next. While coroutines can execute on the caller's stack-frame, stackfull coroutines allow full generality and are sufficient as the basis for concurrency. The aforementioned oracle is a scheduler and the whole system now follows a cooperative threading-model (a.k.a non-preemptive scheduling). The oracle/scheduler can either be a stackless or stackfull entity and correspondingly require one or two context switches to run a different coroutine. In any case, a subset of concurrency related challenges start to appear. For the complete set of concurrency challenges to occur, the only feature missing is preemption.
+A scheduler introduces order of execution uncertainty, while preemption introduces uncertainty about where context-switches occur. Mutual-exclusion and synchronisation are ways of limiting non-determinism in a concurrent system. Now it is important to understand that uncertainty is desireable; uncertainty can be used by runtime systems to significantly increase performance and is often the basis of giving a user the illusion that tasks are running in parallel. Optimal performance in concurrent applications is often obtained by having as much non-determinism as correctness allows.
+Indeed, while execution with a single thread and multiple stacks where the thread is self-scheduling deterministically across the stacks is called coroutining, execution with a single and multiple stacks but where the thread is scheduled by an oracle (non-deterministic from the thread perspective) across the stacks is called concurrency.
+Therefore, a minimal concurrency system can be achieved by creating coroutines, which instead of context switching among each other, always ask an oracle where to context switch next. While coroutines can execute on the caller's stack-frame, stackfull coroutines allow full generality and are sufficient as the basis for concurrency. The aforementioned oracle is a scheduler and the whole system now follows a cooperative threading-model \cit. The oracle/scheduler can either be a stackless or stackfull entity and correspondingly require one or two context switches to run a different coroutine. In any case, a subset of concurrency related challenges start to appear. For the complete set of concurrency challenges to occur, the only feature missing is preemption. Indeed, concurrency challenges appear with non-determinism. Using mutual-exclusion or synchronisation are ways of limiting the lack of determinism in a system. A scheduler introduces order of execution uncertainty, while preemption introduces uncertainty about where context-switches occur. Now it is important to understand that uncertainty is not undesireable; uncertainty can often be used by systems to significantly increase performance and is often the basis of giving a user the illusion that tasks are running in parallel. Optimal performance in concurrent applications is often obtained by having as much non-determinism as correctness allows\cit.
 \section{\protect\CFA 's Thread Building Blocks}
 One of the important features that is missing in C is threading. On modern architectures, a lack of threading is unacceptable\cite{Sutter05, Sutter05b}, and therefore modern programming languages must have the proper tools to allow users to write performant concurrent programs to take advantage of parallelism. As an extension of C, \CFA needs to express these concepts in a way that is as natural as possible to programmers familiar with imperative languages. And being a system-level language means programmers expect to choose precisely which features they need and which cost they are willing to pay.
+One of the important features that is missing in C is threading. On modern architectures, a lack of threading is unacceptable\cite{Sutter05, Sutter05b}, and therefore modern programming languages must have the proper tools to allow users to write performant concurrent and/or parallel programs. As an extension of C, \CFA needs to express these concepts in a way that is as natural as possible to programmers familiar with imperative languages. And being a system-level language means programmers expect to choose precisely which features they need and which cost they are willing to pay.
 \section{Coroutines: A stepping stone}\label{coroutine}
+While the main focus of this proposal is concurrency and parallelism, it is important to address coroutines, which are actually a significant building block of a concurrency system. Coroutines need to deal with context-switches and other context-management operations. Therefore, this proposal includes coroutines both as an intermediate step for the implementation of threads, and a first class feature of \CFA. Furthermore, many design challenges of threads are at least partially present in designing coroutines, which makes the design effort that much more relevant. The core \acrshort{api} of coroutines revolve around two features: independent call stacks and \code{suspend}/\code{resume}.
+While the main focus of this proposal is concurrency and parallelism, it is important to address coroutines, which are actually a significant building block of a concurrency system. Coroutines need to deal with context-switchs and other context-management operations. Therefore, this proposal includes coroutines both as an intermediate step for the implementation of threads, and a first class feature of \CFA. Furthermore, many design challenges of threads are at least partially present in designing coroutines, which makes the design effort that much more relevant. The core \acrshort{api} of coroutines revolve around two features: independent call stacks and \code{suspend}/\code{resume}.
+A good example of a problem made easier with coroutines is genereting the fibonacci sequence. This problem comes with the challenge of decoupling how a sequence is generated and how it is used. Figure \ref{fig:fibonacci-c} shows conventional approaches to writing generators in C. All three of these approach suffer from strong coupling. The left and center approaches require that the generator have knowledge of how the sequence will be used, while the rightmost approach requires to user to hold internal state between calls on behalf of th sequence generator and makes it much harder to handle corner cases like the Fibonacci seed.
 \begin{figure}
+\label{fig:fibonacci-c}
 \begin{center}
 \begin{tabular}{c @{\hskip 0.025in}|@{\hskip 0.025in} c @{\hskip 0.025in}|@{\hskip 0.025in} c}
 …
+        }
+}
-int main() {
-        void print_fib(int n) {
-                printf("%d\n", n);
+        }
-        fibonacci_func(
-, print_fib
-        );
+}
 \end{ccode}&\begin{ccode}[tabsize=2]
 //Using output array
 …
                         f2 = next;
+                }
+                array[i] = next;
+        }
+}
+int main() {
+        int a[10];
+        fibonacci_func(
+, a
+        );
+        for(int i=0;i<10;i++){
+                printf("%d\n", a[i]);
+        }
+                *array = next;
+                array++;
+        }
+}
 \end{ccode}&\begin{ccode}[tabsize=2]
 …
 typedef struct {
         int f1, f2;
 } Iterator_t;
+} iterator_t;
 int fibonacci_state(
         Iterator_t * it
+        iterator_t * it
 ) {
         int f;
         f = it->f1 + it->f2;
         it->f2 = it->f1;
         it->f1 = max(f,1);
+        it->f1 = f;
         return f;
+}
 …
-int main() {
-        Iterator_t it={0,0};
-        for(int i=0;i<10;i++){
-                printf("%d\n",
-                        fibonacci_state(
-                                &it
-                        );
-                );
+        }
+}
 \end{ccode}
 \end{tabular}
 \end{center}
 \caption{Different implementations of a fibonacci sequence generator in C.}
-\label{lst:fibonacci-c}
 \end{figure}
+A good example of a problem made easier with coroutines is generators, like the fibonacci sequence. This problem comes with the challenge of decoupling how a sequence is generated and how it is used. Figure \ref{lst:fibonacci-c} shows conventional approaches to writing generators in C. All three of these approach suffer from strong coupling. The left and center approaches require that the generator have knowledge of how the sequence is used, while the rightmost approach requires holding internal state between calls on behalf of the generator and makes it much harder to handle corner cases like the Fibonacci seed.
+Figure \ref{lst:fibonacci-cfa} is an example of a solution to the fibonnaci problem using \CFA coroutines, where the coroutine stack holds sufficient state for the generation. This solution has the advantage of having very strong decoupling between how the sequence is generated and how it is used. Indeed, this version is as easy to use as the \code{fibonacci_state} solution, while the imlpementation is very similar to the \code{fibonacci_func} example.
+Figure \ref{fig:fibonacci-cfa} is an example of a solution to the fibonnaci problem using \CFA coroutines, using the coroutine stack to hold sufficient state for the generation. This solution has the advantage of having very strong decoupling between how the sequence is generated and how it is used. Indeed, this version is a easy to use as the \code{fibonacci_state} solution, while the imlpementation is very similar to the \code{fibonacci_func} example.
 \begin{figure}
+\label{fig:fibonacci-cfa}
 \begin{cfacode}
 coroutine Fibonacci {
 …
 //main automacically called on first resume
 void main(Fibonacci & this) with (this) {
+void main(Fibonacci & this) {
         int fn1, fn2;           //retained between resumes
         fn  = 0;
         fn1 = fn;
+        this.fn = 0;
+        fn1 = this.fn;
         suspend(this);          //return to last resume
         fn  = 1;
+        this.fn = 1;
         fn2 = fn1;
         fn1 = fn;
+        fn1 = this.fn;
         suspend(this);          //return to last resume
         for ( ;; ) {
                 fn  = fn1 + fn2;
+                this.fn = fn1 + fn2;
                 fn2 = fn1;
                 fn1 = fn;
+                fn1 = this.fn;
                 suspend(this);  //return to last resume
+        }
 …
 \end{cfacode}
 \caption{Implementation of fibonacci using coroutines}
-\label{lst:fibonacci-cfa}
 \end{figure}
+Figure \ref{lst:fmt-line} shows the \code{Format} coroutine which rearranges text in order to group characters into blocks of fixed size. The example takes advantage of resuming coroutines in the constructor to simplify the code and highlights the idea that interesting control flow can occur in the constructor.
+\subsection{Construction}
+One important design challenge for coroutines and threads (shown in section \ref{threads}) is that the runtime system needs to run code after the user-constructor runs to connect the object into the system. In the case of coroutines, this challenge is simpler since there is no non-determinism from preemption or scheduling. However, the underlying challenge remains the same for coroutines and threads.
+The runtime system needs to create the coroutine's stack and more importantly prepare it for the first resumption. The timing of the creation is non-trivial since users both expect to have fully constructed objects once execution enters the coroutine main and to be able to resume the coroutine from the constructor. As regular objects, constructors can leak coroutines before they are ready. There are several solutions to this problem but the chosen options effectively forces the design of the coroutine.
+Furthermore, \CFA faces an extra challenge as polymorphic routines create invisible thunks when casted to non-polymorphic routines and these thunks have function scope. For example, the following code, while looking benign, can run into undefined behaviour because of thunks:
+\begin{cfacode}
+//async: Runs function asynchronously on another thread
+forall(otype T)
+extern void async(void (*func)(T*), T* obj);
+forall(otype T)
+void noop(T *) {}
+void bar() {
+        int a;
+        async(noop, &a);
+}
+\end{cfacode}
+The generated C code\footnote{Code trimmed down for brevity} creates a local thunk to hold type information:
+\begin{ccode}
+extern void async(/* omitted */, void (*func)(void *), void *obj);
+void noop(/* omitted */, void *obj){}
+void bar(){
+        int a;
+        void _thunk0(int *_p0){
+                /* omitted */
+                noop(/* omitted */, _p0);
+        }
+        /* omitted */
+        async(/* omitted */, ((void (*)(void *))(&_thunk0)), (&a));
+}
+\end{ccode}
+The problem in this example is a storage management issue, the function pointer \code{_thunk0} is only valid until the end of the block. This extra challenge limits which solutions are viable because storing the function pointer for too long causes undefined behavior; i.e. the stack based thunk being destroyed before it was used. This challenge is an extension of challenges that come with second-class routines. Indeed, GCC nested routines also have the limitation that the routines cannot be passed outside of the scope of the functions these were declared in. The case of coroutines and threads is simply an extension of this problem to multiple call-stacks.
+\subsection{Alternative: Composition}
+One solution to this challenge is to use composition/containement, where uses add insert a coroutine field which contains the necessary information to manage the coroutine.
+\begin{cfacode}
+struct Fibonacci {
+        int fn; //used for communication
+        coroutine c; //composition
+};
+void ?{}(Fibonacci & this) {
+        this.fn = 0;
+        (this.c){}; //Call constructor to initialize coroutine
+}
+\end{cfacode}
+There are two downsides to this approach. The first, which is relatively minor, made aware of the main routine pointer. This information must either be store in the coroutine runtime data or in its static type structure. When using composition, all coroutine handles have the same static type structure which means the pointer to the main needs to be part of the runtime data. This requirement means the coroutine data must be made larger to store a value that is actually a compile time constant (address of the main routine). The second problem, which is both subtle and significant, is that now users can get the initialisation order of coroutines wrong. Indeed, every field of a \CFA struct is constructed but in declaration order, unless users explicitly write otherwise. This semantics means that users who forget to initialize the coroutine handle may resume the coroutine with an uninitilized object. For coroutines, this is unlikely to be a problem, for threads however, this is a significant problem. Figure \ref{fig:fmt-line} shows the \code{Format} coroutine which rearranges text in order to group characters into blocks of fixed size. This is a good example where the control flow is made much simpler from being able to resume the coroutine from the constructor and highlights the idea that interesting control flow can occor in the constructor.
 \begin{figure}
+\label{fig:fmt-line}
 \begin{cfacode}[tabsize=3]
 //format characters into blocks of 4 and groups of 5 blocks per line
 …
 \end{cfacode}
 \caption{Formatting text into lines of 5 blocks of 4 characters.}
-\label{lst:fmt-line}
 \end{figure}
-\subsection{Construction}
-One important design challenge for coroutines and threads (shown in section \ref{threads}) is that the runtime system needs to run code after the user-constructor runs to connect the fully constructed object into the system. In the case of coroutines, this challenge is simpler since there is no non-determinism from preemption or scheduling. However, the underlying challenge remains the same for coroutines and threads.
-The runtime system needs to create the coroutine's stack and more importantly prepare it for the first resumption. The timing of the creation is non-trivial since users both expect to have fully constructed objects once execution enters the coroutine main and to be able to resume the coroutine from the constructor. As regular objects, constructors can leak coroutines before they are ready. There are several solutions to this problem but the chosen options effectively forces the design of the coroutine.
-Furthermore, \CFA faces an extra challenge as polymorphic routines create invisible thunks when casted to non-polymorphic routines and these thunks have function scope. For example, the following code, while looking benign, can run into undefined behaviour because of thunks:
-\begin{cfacode}
-//async: Runs function asynchronously on another thread
-forall(otype T)
-extern void async(void (*func)(T*), T* obj);
-forall(otype T)
-void noop(T*) {}
-void bar() {
-        int a;
-        async(noop, &a); //start thread running noop with argument a
+}
-\end{cfacode}
-The generated C code\footnote{Code trimmed down for brevity} creates a local thunk to hold type information:
-\begin{ccode}
-extern void async(/* omitted */, void (*func)(void *), void *obj);
-void noop(/* omitted */, void *obj){}
-void bar(){
-        int a;
-        void _thunk0(int *_p0){
-                /* omitted */
-                noop(/* omitted */, _p0);
+        }
-        /* omitted */
-        async(/* omitted */, ((void (*)(void *))(&_thunk0)), (&a));
+}
-\end{ccode}
-The problem in this example is a storage management issue, the function pointer \code{_thunk0} is only valid until the end of the block, which limits the viable solutions because storing the function pointer for too long causes undefined behavior; i.e., the stack-based thunk being destroyed before it can be used. This challenge is an extension of challenges that come with second-class routines. Indeed, GCC nested routines also have the limitation that nested routine cannot be passed outside of the declaration scope. The case of coroutines and threads is simply an extension of this problem to multiple call-stacks.
-\subsection{Alternative: Composition}
-One solution to this challenge is to use composition/containement, where coroutine fields are added to manage the coroutine.
-\begin{cfacode}
-struct Fibonacci {
-        int fn; //used for communication
-        coroutine c; //composition
-};
-void FibMain(void *) {
-        //...
+}
-void ?{}(Fibonacci & this) {
-        this.fn = 0;
-        //Call constructor to initialize coroutine
-        (this.c){myMain};
+}
-\end{cfacode}
-The downside of this approach is that users need to correctly construct the coroutine handle before using it. Like any other objects, doing so the users carefully choose construction order to prevent usage of unconstructed objects. However, in the case of coroutines, users must also pass to the coroutine information about the coroutine main, like in the previous example. This opens the door for user errors and requires extra runtime storage to pass at runtime information that can be known statically.
 \subsection{Alternative: Reserved keyword}
 …
 };
 \end{cfacode}
 The \code{coroutine} keyword means the compiler can find and inject code where needed. The downside of this approach is that it makes coroutine a special case in the language. Users wantint to extend coroutines or build their own for various reasons can only do so in ways offered by the language. Furthermore, implementing coroutines without language supports also displays the power of the programming language used. While this is ultimately the option used for idiomatic \CFA code, coroutines and threads can still be constructed by users without using the language support. The reserved keywords are only present to improve ease of use for the common cases.
+This mean the compiler can solve problems by injecting code where needed. The downside of this approach is that it makes coroutine a special case in the language. Users who would want to extend coroutines or build their own for various reasons can only do so in ways offered by the language. Furthermore, implementing coroutines without language supports also displays the power of the programming language used. While this is ultimately the option used for idiomatic \CFA code, coroutines and threads can both be constructed by users without using the language support. The reserved keywords are only present to improve ease of use for the common cases.
 \subsection{Alternative: Lamda Objects}
 For coroutines as for threads, many implementations are based on routine pointers or function objects\cite{Butenhof97, ANSI14:C++, MS:VisualC++, BoostCoroutines15}. For example, Boost implements coroutines in terms of four functor object types:
+For coroutines as for threads, many implementations are based on routine pointers or function objects\cit. For example, Boost implements coroutines in terms of four functor object types:
 \begin{cfacode}
 asymmetric_coroutine<>::pull_type
 …
 Often, the canonical threading paradigm in languages is based on function pointers, pthread being one of the most well known examples. The main problem of this approach is that the thread usage is limited to a generic handle that must otherwise be wrapped in a custom type. Since the custom type is simple to write in \CFA and solves several issues, added support for routine/lambda based coroutines adds very little.
 A variation of this would be to use a simple function pointer in the same way pthread does for threads :
+A variation of this would be to use an simple function pointer in the same way pthread does for threads :
 \begin{cfacode}
 void foo( coroutine_t cid, void * arg ) {
 …
+}
 \end{cfacode}
 This semantics is more common for thread interfaces than coroutines works equally well. As discussed in section \ref{threads}, this approach is superseeded by static approaches in terms of expressivity.
+This semantic is more common for thread interfaces than coroutines but would work equally well. As discussed in section \ref{threads}, this approach is superseeded by static approaches in terms of expressivity.
 \subsection{Alternative: Trait-based coroutines}
 …
 \end{cfacode}
 In this example, threads of type \code{foo} start execution in the \code{void main(foo &)} routine, which prints \code{"Hello World!"}. While this thesis encourages this approach to enforce strongly-typed programming, users may prefer to use the routine-based thread semantics for the sake of simplicity. With the static semantics it is trivial to write a thread type that takes a function pointer as a parameter and executes it on its stack asynchronously.
+In this example, threads of type \code{foo} start execution in the \code{void main(foo &)} routine, which prints \code{"Hello World!"}. While this thesis encourages this approach to enforce strongly-typed programming, users may prefer to use the routine-based thread semantics for the sake of simplicity. With these semantics it is trivial to write a thread type that takes a function pointer as a parameter and executes it on its stack asynchronously
 \begin{cfacode}
 typedef void (*voidFunc)(int);
 …
 void ?{}(FuncRunner & this, voidFunc inFunc, int arg) {
         this.func = inFunc;
-        this.arg  = arg;
+}
 void main(FuncRunner & this) {
-        //thread starts here and runs the function
         this.func( this.arg );
+}
 \end{cfacode}
 A consequence of the strongly-typed approach to main is that memory layout of parameters and return values to/from a thread are now explicitly specified in the \acrshort{api}.
+An consequence of the strongly typed approach to main is that memory layout of parameters and return values to/from a thread are now explicitly specified in the \acrshort{api}.
 Of course for threads to be useful, it must be possible to start and stop threads and wait for them to complete execution. While using an \acrshort{api} such as \code{fork} and \code{join} is relatively common in the literature, such an interface is unnecessary. Indeed, the simplest approach is to use \acrshort{raii} principles and have threads \code{fork} after the constructor has completed and \code{join} before the destructor runs.
 …
 \end{cfacode}
 This semantic has several advantages over explicit semantics: a thread is always started and stopped exaclty once, users cannot make any progamming errors, and it naturally scales to multiple threads meaning basic synchronisation is very simple.
+This semantic has several advantages over explicit semantics: a thread is always started and stopped exaclty once and users cannot make any progamming errors and it naturally scales to multiple threads meaning basic synchronisation is very simple
 \begin{cfacode}
 …
 \end{cfacode}
 However, one of the drawbacks of this approach is that threads now always form a lattice, that is they are always destroyed in the opposite order of construction because of block structure. This restriction is relaxed by using dynamic allocation, so threads can outlive the scope in which they are created, much like dynamically allocating memory lets objects outlive the scope in which they are created.
+However, one of the drawbacks of this approach is that threads now always form a lattice, that is they are always destroyed in opposite order of construction because of block structure. This restriction is relaxed by using dynamic allocation, so threads can outlive the scope in which they are created, much like dynamically allocating memory lets objects outlive the scope in which they are created
 \begin{cfacode}

doc/proposals/concurrency/text/cforall.tex

-              r0fe4e62
+              rf5c3b6c
 % ======================================================================
 % ======================================================================
 \chapter{Cforall Overview}
+\chapter{Cforall crash course}
 % ======================================================================
 % ======================================================================
 The following is a quick introduction to the \CFA language, specifically tailored to the features needed to support concurrency.
+This thesis presents the design for a set of concurrency features in \CFA. Since it is a new dialect of C, the following is a quick introduction to the language, specifically tailored to the features needed to support concurrency.
+\CFA is a extension of ISO-C and therefore supports all of the same paradigms as C. It is a non-object oriented system language, meaning most of the major abstractions have either no runtime overhead or can be opt-out easily. Like C, the basics of \CFA revolve around structures and routines, which are thin abstractions over machine code. The vast majority of the code produced by the \CFA translator respects memory-layouts and calling-conventions laid out by C. Interestingly, while \CFA is not an object-oriented language, lacking the concept of a receiver (e.g., this), it does have some notion of objects\footnote{C defines the term objects as : ``region of data storage in the execution environment, the contents of which can represent
+values''\cite[3.15]{C11}}, most importantly construction and destruction of objects. Most of the following code examples can be found on the \CFA website \cite{www-cfa}
+\CFA is a extension of ISO-C and therefore supports all of the same paradigms as C. It is a non-object oriented system language, meaning most of the major abstractions have either no runtime overhead or can be opt-out easily. Like C, the basics of \CFA revolve around structures and routines, which are thin abstractions over machine code. The vast majority of the code produced by the \CFA translator respects memory-layouts and calling-conventions laid out by C. Interestingly, while \CFA is not an object-oriented language, lacking the concept of a received (e.g.: this), it does have some notion of objects\footnote{C defines the term objects as : [Where to I get the C11 reference manual?]}, most importantly construction and destruction of objects. Most of the following pieces of code can be found on the \CFA website \cite{www-cfa}
 \section{References}
 Like \CC, \CFA introduces rebindable references providing multiple dereferecing as an alternative to pointers. In regards to concurrency, the semantic difference between pointers and references are not particularly relevant, but since this document uses mostly references, here is a quick overview of the semantics:
+Like \CC, \CFA introduces references as an alternative to pointers. In regards to concurrency, the semantics difference between pointers and references are not particularly relevant but since this document uses mostly references here is a quick overview of the semantics :
 \begin{cfacode}
 int x, *p1 = &x, **p2 = &p1, ***p3 = &p2,
         &r1 = x,    &&r2 = r1,   &&&r3 = r2;
+&r1 = x,    &&r2 = r1,   &&&r3 = r2;
 ***p3 = 3;                                                      //change x
 r3    = 3;                                                      //change x, ***r3
 …
 sizeof(&ar[1]) == sizeof(int *);        //is true, i.e., the size of a reference
 \end{cfacode}
 The important take away from this code example is that references offer a handle to an object, much like pointers, but which is automatically dereferenced for convinience.
+The important thing to take away from this code snippet is that references offer a handle to an object much like pointers but which is automatically derefferenced when convinient.
 \section{Overloading}
 Another important feature of \CFA is function overloading as in Java and \CC, where routines with the same name are selected based on the number and type of the arguments. As well, \CFA uses the return type as part of the selection criteria, as in Ada\cite{Ada}. For routines with multiple parameters and returns, the selection is complex.
+Another important feature of \CFA is function overloading as in Java and \CC, where routine with the same name are selected based on the numbers and type of the arguments. As well, \CFA uses the return type as part of the selection criteria, as in Ada\cite{Ada}. For routines with multiple parameters and returns, the selection is complex.
 \begin{cfacode}
 //selection based on type and number of parameters
 …
 double d = f(4);                //select (2)
 \end{cfacode}
 This feature is particularly important for concurrency since the runtime system relies on creating different types to represent concurrency objects. Therefore, overloading is necessary to prevent the need for long prefixes and other naming conventions that prevent name clashes. As seen in chapter \ref{basics}, routine \code{main} is an example that benefits from overloading.
+This feature is particularly important for concurrency since the runtime system relies on creating different types to represent concurrency objects. Therefore, overloading is necessary to prevent the need for long prefixes and other naming conventions that prevent name clashes. As seen in chapter \ref{basics}, routines main is an example that benefits from overloading.
 \section{Operators}
 Overloading also extends to operators. The syntax for denoting operator-overloading is to name a routine with the symbol of the operator and question marks where the arguments of the operation occur, e.g.:
+Overloading also extends to operators. The syntax for denoting operator-overloading is to name a routine with the symbol of the operator and question marks where the arguments of the operation would be, like so :
 \begin{cfacode}
 int ++? (int op);                       //unary prefix increment
 …
 \section{Parametric Polymorphism}
 Routines in \CFA can also be reused for multiple types. This capability is done using the \code{forall} clause which gives \CFA its name. \code{forall} clauses allow separately compiled routines to support generic usage over multiple types. For example, the following sum function works for any type that supports construction from 0 and addition :
+Routines in \CFA can also be reused for multiple types. This is done using the \code{forall} clause which gives \CFA it's name. \code{forall} clauses allow seperatly compiled routines to support generic usage over multiple types. For example, the following sum function will work for any type which support construction from 0 and addition :
 \begin{cfacode}
 //constraint type, 0 and +
 …
 \end{cfacode}
 Since writing constraints on types can become cumbersome for more constrained functions, \CFA also has the concept of traits. Traits are named collection of constraints that can be used both instead and in addition to regular constraints:
+Since writing constraints on types can become cumbersome for more constrained functions, \CFA also has the concept of traits. Traits are named collection of constraints which can be used both instead and in addition to regular constraints:
 \begin{cfacode}
 trait sumable( otype T ) {
 …
 \section{with Clause/Statement}
 Since \CFA lacks the concept of a receiver, certain functions end-up needing to repeat variable names often. To remove this inconvenience, \CFA provides the \code{with} statement, which opens an aggregate scope making its fields directly accessible (like Pascal).
+Since \CFA lacks the concept of a receiver, certain functions end-up needing to repeat variable names often, to solve this \CFA offers the \code{with} statement which opens an aggregate scope making its fields directly accessible (like Pascal).
 \begin{cfacode}
 struct S { int i, j; };
 int mem(S & this) with (this)           //with clause
+int mem(S & this) with this             //with clause
         i = 1;                                          //this->i
         j = 2;                                          //this->j
 …
         struct S1 { ... } s1;
         struct S2 { ... } s2;
         with (s1)                                       //with statement
+        with s1                                         //with statement
+        {
                 //access fields of s1
                 //without qualification
                 with (s2)                                       //nesting
+                with s2                                 //nesting
+                {
                         //access fields of s1 and s2
 …
+                }
+        }
         with (s1, s2)                           //scopes open in parallel
+        with s1, s2                             //scopes open in parallel
+        {
                 //access fields of s1 and s2

doc/proposals/concurrency/text/concurrency.tex

-              r0fe4e62
+              rf5c3b6c
 % ======================================================================
 % ======================================================================
 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 closely relate to networking concepts (channels\cite{CSP,Go} 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.
+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 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.
 Approaches 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}.
 An approach that is worth mentioning because it is gaining in popularity is transactionnal memory~\cite{Dice10}[Check citation]. While this approach is even pursued by system languages like \CC\cite{Cpp-Transactions}, 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.
+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 systems language, which is why it was rejected as the core paradigm for concurrency in \CFA.
 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.
 …
 \subsection{Synchronization}
 As 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 simpler solution to otherwise involved challenges. As mentioned above, synchronization can be expressed as guaranteeing that event \textit{X} always happens before \textit{Y}. Most of the time, synchronisation happens within a critical section, where threads must acquire mutual-exclusion 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}. The classic exmaple is the thread that finishes using a ressource and unblocks a thread waiting to use the resource, but the unblocked thread must compete again to acquire the resource. 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.
+As 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.
 % ======================================================================
 …
 \end{tabular}
 \end{center}
 Notice how the counter is used without any explicit synchronisation and yet supports thread-safe semantics for both reading and writting, which is similar in usage to \CC \code{atomic} template.
 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 con\-structed 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.
 For maximum usability, monitors use \gls{multi-acq} semantics, which means a single thread can acquire the same monitor multiple times without deadlock. For example, figure \ref{fig:search} uses recursion and \gls{multi-acq} to print values inside a binary tree.
+Notice how the counter is used without any explicit synchronisation and yet supports thread-safe semantics for both reading and writting.
+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 a \code{size_t} is an atomic operation.
+For 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.
 \begin{figure}
 \label{fig:search}
 …
 \end{figure}
 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 making exactly one of these keywords mandatory, which provides 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}.
+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 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}.
 The next semantic decision is to establish when \code{mutex} may be used as a type qualifier. Consider the following declarations:
 …
 int f5(monitor * mutex m []); //Not Okay : Array of unkown length
 \end{cfacode}
 Note that not all array functions are actually distinct in the type system. However, even if the code generation could tell the difference, the extra information is still not sufficient to extend meaningfully the monitor call semantic.
 Unlike object-oriented monitors, where calling a mutex member \emph{implicitly} acquires mutual-exclusion of the receiver object, \CFA uses an explicit mechanism to acquire mutual-exclusion. A consequence of this approach is that it extends naturally to multi-monitor calls.
+Note 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.
+Unlike 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.
 \begin{cfacode}
 int f(MonitorA & mutex a, MonitorB & mutex b);
 …
 f(a,b);
 \end{cfacode}
 While OO monitors could be extended with a mutex qualifier for multiple-monitor calls, no example of this feature could be found. The 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 different routines using the same monitors as arguments. This consistent ordering means acquiring multiple monitors in the way is safe from deadlock. However, users can still force the acquiring order. For example, notice which routines use \code{mutex}/\code{nomutex} and how this affects aquiring order:
+The 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:
 \begin{cfacode}
 void foo(A & mutex a, B & mutex b) { //acquire a & b
 …
 The \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.
 However, 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\cite{Lister77}, solving this problem requires:
+However, 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:
 \begin{enumerate}
         \item Dynamically tracking of the monitor-call order.
         \item Implement rollback semantics.
 \end{enumerate}
 While the first requirement is already a significant constraint on the system, implementing a general rollback semantics in a C-like language is still prohibitively complex \cite{Dice10}. 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. While \CFA provides only a partial solution, many system provide no solution and the \CFA partial solution handles many useful cases.
 For example, \gls{multi-acq} and \gls{bulk-acq} can be used together in interesting ways:
+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 or only use \gls{bulk-acq} of all the monitors.
+\Gls{multi-acq} and \gls{bulk-acq} can be used together in interesting ways, for example:
 \begin{cfacode}
 monitor bank { ... };
 …
+}
 \end{cfacode}
 This example shows a trivial solution to the bank-account transfer-problem\cite{BankTransfer}. Without \gls{multi-acq} and \gls{bulk-acq}, the solution to this problem is much more involved and requires carefull engineering.
 \subsection{\code{mutex} statement} \label{mutex-stmt}
 The call semantics discussed aboved have one software engineering issue, only a named routine can acquire the mutual-exclusion of a set of monitor. \CFA offers the \code{mutex} statement to workaround the need for unnecessary names, avoiding a major software engineering problem\cite{2FTwoHardThings}. Listing \ref{lst:mutex-stmt} shows an example of the \code{mutex} statement, which introduces a new scope in which the mutual-exclusion of a set of monitor is acquired. Beyond naming, the \code{mutex} statement has no semantic difference from a routine call with \code{mutex} parameters.
+This 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.
+\subsubsection{\code{mutex} statement} \label{mutex-stmt}
+The call semantics discussed aboved have one software engineering issue, only a named routine can acquire the mutual-exclusion of a set of monitor. \CFA offers the \code{mutex} statement to workaround the need for unnecessary names, avoiding a major software engineering problem\cit. Listing \ref{lst:mutex-stmt} shows an example of the \code{mutex} statement, which introduces a new scope in which the mutual-exclusion of a set of monitor is acquired. Beyond naming, the \code{mutex} statement has no semantic difference from a routine call with \code{mutex} parameters.
 \begin{figure}
 …
 \end{cfacode}
+Like threads and coroutines, monitors are defined in terms of traits with some additional language support in the form of the \code{monitor} keyword. The monitor trait is :
+\begin{cfacode}
+trait is_monitor(dtype T) {
+        monitor_desc * get_monitor( T & );
+        void ^?{}( T & mutex );
+};
+\end{cfacode}
+Note that the destructor of a monitor must be a \code{mutex} routine. This requirement ensures that the destructor has mutual-exclusion. As with any object, any call to a monitor, using \code{mutex} or otherwise, is Undefined Behaviour after the destructor has run.
+% ======================================================================
+% ======================================================================
+\section{Internal scheduling} \label{intsched}
+% ======================================================================
+% ======================================================================
+In 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 \cite{Hoare74}. 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.
+% ======================================================================
+% ======================================================================
+\section{Internal scheduling} \label{insched}
+% ======================================================================
+% ======================================================================
+In 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.
 First, here is a simple example of such a technique:
 …
 \end{cfacode}
 There 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. The alternative is to return immediately after the call to \code{signal}, which is significantly more restrictive. Second, in \CFA, while it is common to store a \code{condition} as a field of the monitor, 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.
 An important aspect of the implementation is that \CFA does not allow barging, which means that once function \code{bar} releases the monitor, \code{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.
+There 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.
+An 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.
 % ======================================================================
 …
 % ======================================================================
 % ======================================================================
 It is easier to understand the problem of multi-monitor scheduling using a series of pseudo-code. Note that for simplicity in the following snippets of pseudo-code, waiting and signalling is done using an implicit condition variable, like Java built-in monitors. Indeed, \code{wait} statements always use the implicit condition as paremeter and explicitly names the monitors (A and B) associated with the condition. Note that in \CFA, condition variables are tied to a set of monitors on first use (called branding) which means that using internal scheduling with distinct sets of monitors requires one condition variable per set of monitors.
+It is easier to understand the problem of multi-monitor scheduling using a series of pseudo-code. Note that for simplicity in the following snippets of pseudo-code, waiting and signalling is done using an implicit condition variable, like Java built-in monitors. Indeed, \code{wait} statements always use a single condition as paremeter and waits on the monitors associated with the condition.
 \begin{multicols}{2}
 …
 \end{pseudo}
 \end{multicols}
 This version uses \gls{bulk-acq} (denoted using the {\sf\&} 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.
 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. For monitors, a well known deadlock problem is the Nested Monitor Problem \cite{Lister77}, which occurs when a \code{wait} is made by a thread that holds more than one monitor. For example, the following pseudo-code runs into the nested-monitor problem :
+This 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.
+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. 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 :
 \begin{multicols}{2}
 \begin{pseudo}
 …
 \end{pseudo}
 \end{multicols}
-The \code{wait} only releases monitor \code{B} so the signalling thread cannot acquire monitor \code{A} to get to the \code{signal}. Attempting release of all acquired monitors at the \code{wait} results in another set of problems such as releasing monitor \code{C}, which has nothing to do with the \code{signal}.
 However, 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.
 …
 \end{multicols}
+% ======================================================================
+% ======================================================================
+\subsection{Internal Scheduling - in depth}
+% ======================================================================
+% ======================================================================
+A larger example is presented to show complex issuesfor \gls{bulk-acq} and all the implementation options are analyzed. Listing \ref{lst:int-bulk-pseudo} shows an example where \gls{bulk-acq} adds a significant layer of complexity to the internal signalling semantics, and listing \ref{lst:int-bulk-cfa} shows the corresponding \CFA code which implements the pseudo-code in listing \ref{lst:int-bulk-pseudo}. For the purpose of translating the given pseudo-code into \CFA-code any method of introducing monitor into context, other than a \code{mutex} parameter, is acceptable, e.g., global variables, pointer parameters or using locals with the \code{mutex}-statement.
+Listing \ref{lst:int-bulk-pseudo} shows an example where \gls{bulk-acq} adds a significant layer of complexity to the internal signalling semantics. Listing \ref{lst:int-bulk-cfa} shows the corresponding \CFA code which implements the pseudo-code in listing \ref{lst:int-bulk-pseudo}. Note that listing \ref{lst:int-bulk-cfa} uses non-\code{mutex} parameter to introduce monitor \code{b} into context. However, for the purpose of translating the given pseudo-code into \CFA-code any method of introducing new monitors into context, other than a \code{mutex} parameter, is acceptable, e.g. global variables, pointer parameters or using locals with the \code{mutex}-statement.
 \begin{figure}[!b]
 …
 \begin{figure}[!b]
-\begin{center}
-\begin{cfacode}[xleftmargin=.4\textwidth]
-monitor A a;
-monitor B b;
-condition c;
-\end{cfacode}
-\end{center}
 \begin{multicols}{2}
 Waiting thread
 \begin{cfacode}
+mutex(a) {
+monitor A;
+monitor B;
+extern condition c;
+void foo(A & mutex a, B & b) {
         //Code Section 1
         mutex(a, b) {
 …
 Signalling thread
 \begin{cfacode}
+mutex(a) {
+monitor A;
+monitor B;
+extern condition c;
+void foo(A & mutex a, B & b) {
         //Code Section 5
         mutex(a, b) {
 …
 \end{figure}
 The complexity begins at 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 \code{A & B}'' (line 16), it must actually transfer ownership of monitor \code{B} to the waiting thread. This ownership trasnfer is required in order to prevent barging. Since the signalling thread still needs monitor \code{A}, simply waking up the waiting thread is not an option because it violates mutual exclusion. There are three options.
+It 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.
 \subsubsection{Delaying signals}
 The 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 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.
+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 objects, effectively making the existing single monitor semantic viable by simply changing monitors to monitor groups.
 \begin{multicols}{2}
 Waiter
 …
 \end{multicols}
 However, 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:
+\begin{figure}
+\begin{multicols}{3}
+Thread $\alpha$
+\begin{multicols}{2}
+Thread 1
 \begin{pseudo}[numbers=left, firstnumber=1]
 acquire A
 …
 \end{pseudo}
+\columnbreak
+Thread $\gamma$
+\begin{pseudo}[numbers=left, firstnumber=1]
+Thread 2
+\begin{pseudo}[numbers=left, firstnumber=6]
+acquire A
+        wait A
+release A
+\end{pseudo}
+\columnbreak
+Thread 3
+\begin{pseudo}[numbers=left, firstnumber=9]
 acquire A
         acquire A & B
                 signal A & B
         release A & B
+        //Secretly keep B here
         signal A
 release A
+\end{pseudo}
+\columnbreak
+Thread $\beta$
+\begin{pseudo}[numbers=left, firstnumber=1]
+acquire A
+        wait A
+release A
+\end{pseudo}
+\end{multicols}
+\caption{Dependency graph}
+\label{lst:dependency}
+\end{figure}
+//Wakeup thread 1 or 2?
+//Who wakes up the other thread?
+\end{pseudo}
+\end{multicols}
 The goal in this solution is to avoid the need to transfer ownership of a subset of the condition monitors. However, this goal is unreacheable in the previous example. Depending on the order of signals (line 12 and 15) two cases can happen.
 …
 Note 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.
 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 and therefore effectively precludes this approach.
+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 and therefore invalidates the main benefit of this approach.
 \subsubsection{Dependency graphs}
 In the listing \ref{lst:int-bulk-pseudo} 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 \code{A & B} and then the waiter transfers back ownership of \code{A} when it releases it, then the problem is solved (\code{B} is no longer in use at this point). 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:
+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:
 \begin{multicols}{2}
 …
 \begin{figure}
+\begin{multicols}{3}
+Thread $\alpha$
+\begin{pseudo}[numbers=left, firstnumber=1]
+acquire A
+        acquire A & B
+                wait A & B
+        release A & B
+release A
+\end{pseudo}
+\columnbreak
+Thread $\gamma$
+\begin{pseudo}[numbers=left, firstnumber=1]
+acquire A
+        acquire A & B
+                signal A & B
+        release A & B
+        signal A
+release A
+\end{pseudo}
+\columnbreak
+Thread $\beta$
+\begin{pseudo}[numbers=left, firstnumber=1]
+acquire A
+        wait A
+release A
+\end{pseudo}
+\end{multicols}
+\caption{Dependency graph}
+\label{lst:dependency}
+\end{figure}
+\begin{figure}
 \begin{center}
 \input{dependency}
 \end{center}
+\label{fig:dependency}
 \caption{Dependency graph of the statements in listing \ref{lst:dependency}}
-\label{fig:dependency}
 \end{figure}
 Listing \ref{lst:dependency} is the three thread example rewritten for dependency graphs. Figure \ref{fig:dependency} shows the corresponding dependency graph that results, where every node is a statement of one of the three threads, and the arrows the dependency of that statement (e.g., $\alpha1$ must happen before $\alpha2$). The extra challenge is that this dependency graph is effectively post-mortem, but the runtime 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.
+Listing \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 statement of one of the three threads, and the arrows the dependency of that statement. 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.
 \subsubsection{Partial signalling} \label{partial-sig}
+Finally, the solution that is chosen for \CFA is to use partial signalling. Again using listing \ref{lst:int-bulk-pseudo}, 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 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. Furthermore, after being fully implemented, this solution does not appear to have any downsides worth mentionning.
+Finally, the solution that is chosen for \CFA is to use partial signalling. Consider the following case:
+\begin{multicols}{2}
+\begin{pseudo}[numbers=left]
+acquire A
+        acquire A & B
+                wait A & B
+        release A & B
+release A
+\end{pseudo}
+\columnbreak
+\begin{pseudo}[numbers=left, firstnumber=6]
+acquire A
+        acquire A & B
+                signal A & B
+        release A & B
+        //... More code
+release A
+\end{pseudo}
+\end{multicols}
+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 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.
 % ======================================================================
 …
 % ======================================================================
 % ======================================================================
+An 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}.
+The 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.
 \begin{figure}
 \begin{tabular}{|c|c|}
 …
                 girlPhoneNo = phoneNo;
                 //wake boy from chair
+                //wake boy fron chair
                 signal(exchange);
+        }
 …
                 girlPhoneNo = phoneNo;
                 //wake boy from chair
+                //wake boy fron chair
                 signal(exchange);
+        }
 …
 \label{lst:datingservice}
 \end{figure}
-An 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}.
-The 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 requires 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 handles the two-way handshake 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.
 % ======================================================================
 …
 % ======================================================================
 % ======================================================================
 An alternative to internal scheduling is external scheduling, e.g., in \uC.
+An alternative to internal scheduling is to use external scheduling.
 \begin{center}
 \begin{tabular}{|c|c|c|}
 Internal Scheduling & External Scheduling & Go\\
+\begin{tabular}{|c|c|}
+Internal Scheduling & External Scheduling \\
 \hline
 \begin{ucppcode}[tabsize=3]
+\begin{ucppcode}
 _Monitor Semaphore {
         condition c;
 …
 public:
         void P() {
+                if(inUse)
+                        wait(c);
+                if(inUse) wait(c);
                 inUse = true;
+        }
 …
+        }
+}
 \end{ucppcode}&\begin{ucppcode}[tabsize=3]
+\end{ucppcode}&\begin{ucppcode}
 _Monitor Semaphore {
 …
 public:
         void P() {
+                if(inUse)
+                        _Accept(V);
+                if(inUse) _Accept(V);
                 inUse = true;
+        }
 …
+        }
+}
+\end{ucppcode}&\begin{gocode}[tabsize=3]
+type MySem struct {
+        inUse bool
+        c     chan bool
+}
+// acquire
+func (s MySem) P() {
+        if s.inUse {
+                select {
+                case <-s.c:
+                }
+        }
+        s.inUse = true
+}
+// release
+func (s MySem) V() {
+        s.inUse = false
+        //This actually deadlocks
+        //when single thread
+        s.c <- false
+}
+\end{gocode}
+\end{ucppcode}
 \end{tabular}
 \end{center}
 This 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.
 For the \code{P} member above using internal scheduling, the call to \code{wait} only guarantees that \code{V} is the last routine to access the monitor, allowing a third routine, say \code{isInUse()}, acquire mutual exclusion several times while routine \code{P} is waiting. On the other hand, external scheduling guarantees that while routine \code{P} is waiting, no routine other than \code{V} can acquire the monitor.
+This 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.
+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 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.
 % ======================================================================
 …
 % ======================================================================
 % ======================================================================
 In \uC, monitor declarations include an exhaustive list of monitor operations. Since \CFA is not object oriented, monitors become both more difficult to implement and less clear for a user:
+In \uC, monitor declarations include an exhaustive list of monitor operations. Since \CFA is not object oriented it becomes both more difficult to implement but also less clear for the user:
 \begin{cfacode}
 …
 For 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:
-\begin{figure}[H]
 \begin{center}
 {\resizebox{0.4\textwidth}{!}{\input{monitor}}}
 \end{center}
+\label{fig:monitor}
+\end{figure}
+There are other alternatives to these pictures, but in the case of this picture, implementing a fast accept check is relatively easy. Restricted to a fixed number of mutex members, N, the accept check reduces to updating a bitmask when the acceptor queue changes, 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 enumerates (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. It is 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.
+The alternative is to alter the implementeation like this:
+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.
+The alternative is to have a picture like this one:
 \begin{center}
 …
 \end{center}
+Generating a mask dynamically 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 accepted function-pointers replaces the single instruction bitmask compare with dereferencing a pointer followed by a linear search. Furthermore, supporting nested external scheduling (e.g., listing \ref{lst:nest-ext}) may now require additionnal searches on calls to \code{waitfor} statement to check if a routine is already queued in.
+\begin{figure}
+Not 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.
+Note 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.
+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 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.
+% ======================================================================
+% ======================================================================
+\subsection{Multi-monitor scheduling}
+% ======================================================================
+% ======================================================================
+External 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:
 \begin{cfacode}
 monitor M {};
-void foo( M & mutex a ) {}
-void bar( M & mutex b ) {
-        //Nested in the waitfor(bar, c) call
-        waitfor(foo, b);
+}
-void baz( M & mutex c ) {
-        waitfor(bar, c);
+}
-\end{cfacode}
-\caption{Example of nested external scheduling}
-\label{lst:nest-ext}
-\end{figure}
-Note that in the second picture, tasks need to always keep track of 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.
-At this point, a decision must be made 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.
-% ======================================================================
-% ======================================================================
-\subsection{Multi-monitor scheduling}
-% ======================================================================
-% ======================================================================
-External 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:
-\begin{cfacode}
-monitor M {};
 void f(M & mutex a);
 void g(M & mutex b, M & mutex c) {
         waitfor(f); //two monitors M => unkown which to pass to f(M & mutex)
+void g(M & mutex a, M & mutex b) {
+        waitfor(f); //ambiguous, keep a pass b or other way around?
+}
 \end{cfacode}
 …
 \end{cfacode}
 This syntax is unambiguous. Both locks are acquired and kept by \code{g}. 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 \code{waitfor} statement as follows.
+This 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 statement as follows.
 \begin{cfacode}
 …
 Note 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.
 An important behavior to note is when a set of monitors only match partially :
+An important behavior to note is that what happens when a set of monitors only match partially :
 \begin{cfacode}
 …
 \end{cfacode}
 While 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 irrelevant; \code{waitfor(f,a,b)} and \code{waitfor(f,b,a)} are indistinguishable waiting condition.
+While 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.
 % ======================================================================
 …
 % ======================================================================
 Syntactically, the \code{waitfor} statement takes a function identifier and a set of monitors. While the set of monitors can be any list of expression, the function name is more restricted because the compiler validates at compile time the validity of the function type and the parameters used with the \code{waitfor} statement. It checks that the set of monitor passed in matches the requirements for a function call. Listing \ref{lst:waitfor} shows various usage of the waitfor statement and which are acceptable. The choice of the function type is made ignoring any non-\code{mutex} parameter. One limitation of the current implementation is that it does not handle overloading.
+Syntactically, the \code{waitfor} statement takes a function identifier and a set of monitors. While the set of monitors can be any list of expression, the function name is more restricted. This is because the compiler validates at compile time the validity of the waitfor statement. It checks that the set of monitor passed in matches the requirements for a function call. Listing \ref{lst:waitfor} shows various usage of the waitfor statement and which are acceptable. The choice of the function type is made ignoring any non-\code{mutex} parameter. One limitation of the current implementation is that it does not handle overloading.
 \begin{figure}
 \begin{cfacode}
 …
         waitfor(f2, a1, a2); //Incorrect : Mutex arguments don't match
         waitfor(f1, 1);      //Incorrect : 1 not a mutex argument
         waitfor(f9, a1);     //Incorrect : f9 function does not exist
         waitfor(*fp, a1 );   //Incorrect : fp not an identifier
+        waitfor(f4, a1);     //Incorrect : f9 not a function
+        waitfor(*fp, a1 );   //Incorrect : fp not a identifier
         waitfor(f4, a1);     //Incorrect : f4 ambiguous
 …
 \end{figure}
 Finally, for added flexibility, \CFA supports constructing complex \code{waitfor} mask using the \code{or}, \code{timeout} and \code{else}. Indeed, multiple \code{waitfor} can be chained together using \code{or}; this chain forms a single statement that uses baton-pass to any one function that fits one of the function+monitor set passed in. To eanble users to tell which accepted function is accepted, \code{waitfor}s are followed by a statement (including the null statement \code{;}) or a compound statement. When multiple \code{waitfor} are chained together, only the statement corresponding to the accepted function is executed. A \code{waitfor} chain can also be followed by a \code{timeout}, to signify an upper bound on the wait, or an \code{else}, to signify that the call should be non-blocking, that is only check of a matching function call already arrived and return immediately otherwise. Any and all of these clauses can be preceded by a \code{when} condition to dynamically construct the mask based on some current state. Listing \ref{lst:waitfor2}, demonstrates several complex masks and some incorrect ones.
+Finally, for added flexibility, \CFA supports constructing complex waitfor mask using the \code{or}, \code{timeout} and \code{else}. Indeed, multiple \code{waitfor} can be chained together using \code{or}; this chain will form a single statement which will baton-pass to any one function that fits one of the function+monitor set which was passed in. To eanble users to tell which was the accepted function, \code{waitfor}s are followed by a statement (including the null statement \code{;}) or a compound statement. When multiple \code{waitfor} are chained together, only the statement corresponding to the accepted function is executed. A \code{waitfor} chain can also be followed by a \code{timeout}, to signify an upper bound on the wait, or an \code{else}, to signify that the call should be non-blocking, that is only check of a matching function already arrived and return immediately otherwise. Any and all of these clauses can be preceded by a \code{when} condition to dynamically construct the mask based on some current state. Listing \ref{lst:waitfor2}, demonstrates several complex masks and some incorrect ones.
 \begin{figure}
 …
 \label{lst:waitfor2}
 \end{figure}
-% ======================================================================
-% ======================================================================
-\subsection{Waiting for the destructor}
-% ======================================================================
-% ======================================================================
-An interesting use for the \code{waitfor} statement is destructor semantics. Indeed, the \code{waitfor} statement can accept any \code{mutex} routine, which includes the destructor (see section \ref{data}). However, with the semantics discussed until now, waiting for the destructor does not make any sense since using an object after its destructor is called is undefined behaviour. The simplest approach is to disallow \code{waitfor} on a destructor. However, a more expressive approach is to flip execution ordering when waiting for the destructor, meaning that waiting for the destructor allows the destructor to run after the current \code{mutex} routine, similarly to how a condition is signalled.
-\begin{figure}
-\begin{cfacode}
-monitor Executer {};
-struct  Action;
-void ^?{}   (Executer & mutex this);
-void execute(Executer & mutex this, const Action & );
-void run    (Executer & mutex this) {
-        while(true) {
-                   waitfor(execute, this);
-                or waitfor(^?{}   , this) {
-                        break;
+                }
+        }
+}
-\end{cfacode}
-\caption{Example of an executor which executes action in series until the destructor is called.}
-\label{lst:dtor-order}
-\end{figure}
-For example, listing \ref{lst:dtor-order} shows an example of an executor with an infinite loop, which waits for the destructor to break out of this loop. Switching the semantic meaning introduces an idiomatic way to terminate a task and/or wait for its termination via destruction.

doc/proposals/concurrency/text/future.tex

-              r0fe4e62
+              rf5c3b6c
 % ======================================================================
+\section{Flexible Scheduling} \label{futur:sched}
+An important part of concurrency is scheduling. Different scheduling algorithm can affact peformance (both in terms of average and variation). However, no single scheduler is optimal for all workloads and therefore there is value in being able to change the scheduler for given programs. One solution is to offer various tweaking options to users, allowing the scheduler to be adjusted the to requirements of the workload. However, in order to be truly flexible, it would be interesting to allow users to add arbitrary data and arbirary scheduling algorithms to the scheduler. For example, a web server could attach Type-of-Service information to threads and have a ``ToS aware'' scheduling algorithm tailored to this specific web server. This path of flexible schedulers will be explored for \CFA.
+\section{Non-Blocking IO} \label{futur:nbio}
+While most of the parallelism tools
+However, many modern workloads are not bound on computation but on IO operations, an common case being webservers and XaaS (anything as a service). These type of workloads often require significant engineering around amortising costs of blocking IO operations. While improving throughtput of these operations is outside what \CFA can do as a language, it can help users to make better use of the CPU time otherwise spent waiting on IO operations. The current trend is to use asynchronous programming using tools like callbacks and/or futurs and promises\cite. However, while these are valid solutions, they lead to code that is harder to read and maintain because it is much less linear
+\section{Other concurrency tools} \label{futur:tools}
+While monitors offer a flexible and powerful concurent core for \CFA, other concurrency tools are also necessary for a complete multi-paradigm concurrency package. Example of such tools can include simple locks and condition variables, futures and promises\cite{promises}, and executors. These additional features are useful when monitors offer a level of abstraction which is indaquate for certain tasks.
+\section{Implicit threading} \label{futur:implcit}
+Simpler applications can benefit greatly from having implicit parallelism. That is, parallelism that does not rely on the user to write concurrency. This type of parallelism can be achieved both at the language level and at the library level. The cannonical example of implcit parallelism is parallel for loops, which are the simplest example of a divide and conquer algorithm\cite{uC++book}. Listing \ref{lst:parfor} shows three different code examples that accomplish pointwise sums of large arrays. Note that none of these example explicitly declare any concurrency or parallelism objects.
+\begin{figure}
+\begin{center}
+\begin{tabular}[t]{|c|c|c|}
+Sequential & Library Parallel & Language Parallel \\
+\begin{cfacode}[tabsize=3]
+void big_sum(
+        int* a, int* b,
+        int* o,
+        size_t len)
+{
+        for(
+                int i = 0;
+                i < len;
+                ++i )
+        {
+                o[i]=a[i]+b[i];
+        }
+}
+Concurrency and parallelism is still a very active field that strongly benefits from hardware advances. As such certain features that aren't necessarily mature enough in their current state could become relevant in the lifetime of \CFA.
+\section{Non-Blocking IO}
+\section{Other concurrency tools}
+int* a[10000];
+int* b[10000];
+int* c[10000];
+//... fill in a & b
+big_sum(a,b,c,10000);
+\end{cfacode} &\begin{cfacode}[tabsize=3]
+void big_sum(
+        int* a, int* b,
+        int* o,
+        size_t len)
+{
+        range ar(a, a+len);
+        range br(b, b+len);
+        range or(o, o+len);
+        parfor( ai, bi, oi,
+        [](     int* ai,
+                int* bi,
+                int* oi)
+        {
+                oi=ai+bi;
+        });
+}
+\section{Implicit threading}
+% Finally, simpler applications can benefit greatly from having implicit parallelism. That is, parallelism that does not rely on the user to write concurrency. This type of parallelism can be achieved both at the language level and at the system level.
+%
+% \begin{center}
+% \begin{tabular}[t]{|c|c|c|}
+% Sequential & System Parallel & Language Parallel \\
+% \begin{lstlisting}
+% void big_sum(int* a, int* b,
+%                int* out,
+%                size_t length)
+% {
+%       for(int i = 0; i < length; ++i ) {
+%               out[i] = a[i] + b[i];
+%       }
+% }
+%
+%
+%
+%
+%
+% int* a[10000];
+% int* b[10000];
+% int* c[10000];
+% //... fill in a and b ...
+% big_sum(a, b, c, 10000);
+% \end{lstlisting} &\begin{lstlisting}
+% void big_sum(int* a, int* b,
+%                int* out,
+%                size_t length)
+% {
+%       range ar(a, a + length);
+%       range br(b, b + length);
+%       range or(out, out + length);
+%       parfor( ai, bi, oi,
+%       [](int* ai, int* bi, int* oi) {
+%               oi = ai + bi;
+%       });
+% }
+%
+% int* a[10000];
+% int* b[10000];
+% int* c[10000];
+% //... fill in a and b ...
+% big_sum(a, b, c, 10000);
+% \end{lstlisting}&\begin{lstlisting}
+% void big_sum(int* a, int* b,
+%                int* out,
+%                size_t length)
+% {
+%       for (ai, bi, oi) in (a, b, out) {
+%               oi = ai + bi;
+%       }
+% }
+%
+%
+%
+%
+%
+% int* a[10000];
+% int* b[10000];
+% int* c[10000];
+% //... fill in a and b ...
+% big_sum(a, b, c, 10000);
+% \end{lstlisting}
+% \end{tabular}
+% \end{center}
+%
+int* a[10000];
+int* b[10000];
+int* c[10000];
+//... fill in a & b
+big_sum(a,b,c,10000);
+\end{cfacode}&\begin{cfacode}[tabsize=3]
+void big_sum(
+        int* a, int* b,
+        int* o,
+        size_t len)
+{
+        parfor (ai,bi,oi)
+            in (a, b, o )
+        {
+                oi = ai + bi;
+        }
+}
+\section{Multiple Paradigms}
+int* a[10000];
+int* b[10000];
+int* c[10000];
+//... fill in a & b
+big_sum(a,b,c,10000);
+\end{cfacode}
+\end{tabular}
+\end{center}
+\caption{For loop to sum numbers: Sequential, using library parallelism and language parallelism.}
+\label{lst:parfor}
+\end{figure}
+Implicit parallelism is a general solution and therefore has its limitations. However, it is a quick and simple approach to parallelism which may very well be sufficient for smaller applications and reduces the amount of boiler-plate that is needed to start benefiting from parallelism in modern CPUs.
+\section{Transactions}

doc/proposals/concurrency/text/internals.tex

-              r0fe4e62
+              rf5c3b6c
 \chapter{Behind the scene}
-There are several challenges specific to \CFA when implementing concurrency. These challenges are a direct result of \gls{bulk-acq} and loose object-definitions. These two constraints are the root cause of most design decisions in the implementation. Furthermore, to avoid contention from dynamically allocating memory in a concurrent environment, the internal-scheduling design is (almost) entirely free of mallocs. 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.
-The main memory concern for concurrency is queues. All blocking operations are made by parking threads onto queues. The queue design needs to be intrusive\cite{IntrusiveData} to avoid the need for memory allocation, which entails that all the nodes need specific fields to keep track of all needed information. Since many concurrency operations can use an unbound amount of memory (depending on \gls{bulk-acq}), statically defining 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 variable length arrays, which are used extensively.
-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. 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 the later is preferable in terms of performance).
-Note that since the major contributions of this thesis are extending monitor semantics to \gls{bulk-acq} and loose object definitions, any challenges that are not resulting of these characteristiques of \CFA are considered as solved problems and therefore not discussed further.
 % ======================================================================
 % ======================================================================
 \section{Mutex routines}
+\section{Implementation Details: Interaction with polymorphism}
 % ======================================================================
 % ======================================================================
+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.
+The first step towards the monitor implementation is simple mutex-routines using monitors. In the single monitor case, this is done using the entry/exit procedure highlighted in listing \ref{lst:entry1}. This entry/exit procedure does not actually have to be extended to support multiple monitors, indeed it is sufficient to enter/leave monitors one-by-one as long as the order is correct to prevent deadlocks\cite{Havender68}. In \CFA, ordering of monitor relies on memory ordering, this is sufficient because all objects are guaranteed to have distinct non-overlaping memory layouts and mutual-exclusion for a monitor is only defined for its lifetime, meaning that destroying a monitor while it is acquired is undefined behavior. When a mutex call is made, the concerned monitors are agregated into a variable-length pointer array and sorted based on pointer values. This array presists for the entire duration of the mutual-exclusion and its ordering reused extensively.
+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.
+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:
 \begin{figure}
+\begin{multicols}{2}
+Entry
+\begin{pseudo}
+if monitor is free
+        enter
+elif already own the monitor
+        continue
+else
+        block
+increment recursions
+\end{pseudo}
+\columnbreak
+Exit
+\begin{pseudo}
+decrement recursion
+if recursion == 0
+        if entry queue not empty
+                wake-up thread
+\end{pseudo}
+\end{multicols}
+\caption{Initial entry and exit routine for monitors}
+\label{lst:entry1}
+\end{figure}
+\subsection{ Details: Interaction with polymorphism}
+Depending on the choice of semantics for when monitor locks are acquired, interaction between monitors and \CFA's concept of polymorphism can be more complex to support. However, it is shown that entry-point locking solves most of the issues.
+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. 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:
+\begin{figure}[H]
+\label{fig:locking-site}
 \begin{center}
+\setlength\tabcolsep{1.5pt}
 \begin{tabular}{|c|c|c|}
 Mutex & \gls{callsite-locking} & \gls{entry-point-locking} \\
 …
 \end{tabular}
 \end{center}
+\caption{Call-site vs entry-point locking for mutex calls}
+\label{fig:locking-site}
+\caption{Callsite vs entry-point locking for mutex calls}
 \end{figure}
+Note the \code{mutex} keyword relies on the type system, which means that in cases where a generic monitor routine is desired, writing the mutex routine is possible with the proper trait, for example:
+Note 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:
 \begin{cfacode}
 //Incorrect: T may not be monitor
+//Incorrect: T is not a monitor
 forall(dtype T)
 void foo(T * mutex t);
 …
 \end{cfacode}
-Both entry-point and callsite locking are feasible implementations. The current \CFA implementations uses entry-point locking because it requires less work when using \gls{raii}, effectively transferring the burden of implementation to object construction/destruction. The same could be said of callsite locking, the difference being that the later does not necessarily have an existing scope that matches exactly the scope of the mutual exclusion, i.e.: the function body. Furthermore, entry-point locking requires less code generation since any useful routine is called at least as often as it is define, there can be only one entry-point but many callsites.
 % ======================================================================
 % ======================================================================
 \section{Threading} \label{impl:thread}
+\section{Internal scheduling: Implementation} \label{inschedimpl}
 % ======================================================================
 % ======================================================================
+There 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.
 Figure \ref{fig:system1} shows a high-level picture if the \CFA runtime system in regards to concurrency. Each component of the picture is explained in details in the fllowing sections.
+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{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.
+\begin{figure}
+\begin{center}
+{\resizebox{\textwidth}{!}{\input{system.pstex_t}}}
+\end{center}
+\caption{Overview of the entire system}
+\label{fig:system1}
+\end{figure}
+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).
+\subsection{Context Switching}
+As mentionned in section \ref{coroutine}, coroutines are a stepping stone for implementing threading. This is because they share the same mechanism for context-switching between different stacks. To improve performance and simplicity, context-switching is implemented using the following assumption: all context-switches happen inside a specific function call. This assumption means that the context-switch only has to copy the callee-saved registers onto the stack and then switch the stack registers with the ones of the target coroutine/thread. Note that the instruction pointer can be left untouched since the context-switch is always inside the same function. Threads however do not context-switch between each other directly. They context-switch to the scheduler. This method is called a 2-step context-switch and has the advantage of having a clear distinction between user code and the kernel where scheduling and other system operation happen. Obiously, this has the cost of doubling the context-switch cost because threads must context-switch to an intermediate stack. However, the performance of the 2-step context-switch is still superior to a \code{pthread_yield}(see section \ref{results}). additionally, for users in need for optimal performance, it is important to note that having a 2-step context-switch as the default does not prevent \CFA from offering a 1-step context-switch to use manually (or as part of monitors). This option is not currently present in \CFA but the changes required to add it are strictly additive.
+The following figure is the traditionnal illustration of a monitor :
-\subsection{Processors}
-Parallelism in \CFA is built around using processors to specify how much parallelism is desired. \CFA processors are object wrappers around kernel threads, specifically pthreads in the current implementation of \CFA. Indeed, any parallelism must go through operating-system librairies. However, \glspl{uthread} are still the main source of concurrency, processors are simply the underlying source of parallelism. Indeed, processor \glspl{kthread} simply fetch a \glspl{uthread} from the scheduler and run, they are effectively executers for user-threads. The main benefit of this approach is that it offers a well defined boundary between kernel code and user code, for example, kernel thread quiescing, scheduling and interrupt handling. Processors internally use coroutines to take advantage of the existing context-switching semantics.
-\subsection{Stack management}
-One of the challenges of this system is to reduce the footprint as much as possible. Specifically, all pthreads created also have a stack created with them, which should be used as much as possible. Normally, coroutines also create there own stack to run on, however, in the case of the coroutines used for processors, these coroutines run directly on the kernel thread stack, effectively stealing the processor stack. The exception to this rule is the Main Processor, i.e. the initial kernel thread that is given to any program. In order to respect user expectations, the stack of the initial kernel thread, the main stack of the program, is used by the main user thread rather than the main processor.
-\subsection{Preemption} \label{preemption}
-Finally, an important aspect for any complete threading system is preemption. As mentionned in chapter \ref{basics}, preemption introduces an extra degree of uncertainty, which enables users to have multiple threads interleave transparently, rather than having to cooperate among threads for proper scheduling and CPU distribution. Indeed, preemption is desireable because it adds a degree of isolation among threads. In a fully cooperative system, any thread that runs into a long loop can starve other threads, while in a preemptive system starvation can still occur but it does not rely on every thread having to yield or block on a regular basis, which reduces significantly a programmer burden. Obviously, preemption is not optimal for every workload, however any preemptive system can become a cooperative system by making the time-slices extremely large. Which is why \CFA uses a preemptive threading system.
-Preemption in \CFA is based on kernel timers, which are used to run a discrete-event simulation. Every processor keeps track of the current time and registers an expiration time with the preemption system. When the preemption system receives a change in preemption, it sorts these expiration times in a list and sets a kernel timer for the closest one, effectively stepping between preemption events on each signals sent by the timer. These timers use the linux signal {\tt SIGALRM}, which is delivered to the process rather than the kernel-thread. This results in an implementation problem,because when delivering signals to a process, the kernel documentation states that the signal can be delivered to any kernel thread for which the signal is not blocked i.e. :
-\begin{quote}
-A process-directed signal may be delivered to any one of the threads that does not currently have the signal blocked. If more than one of the threads has the signal unblocked, then the kernel chooses an arbitrary thread to which to deliver the signal.
-SIGNAL(7) - Linux Programmer's Manual
-\end{quote}
-For the sake of simplicity and in order to prevent the case of having two threads receiving alarms simultaneously, \CFA programs block the {\tt SIGALRM} signal on every thread except one. Now because of how involontary context-switches are handled, the kernel thread handling {\tt SIGALRM} cannot also be a processor thread.
-Involuntary context-switching is done by sending signal {\tt SIGUSER1} to the corresponding processor and having the thread yield from inside the signal handler. Effectively context-switching away from the signal-handler back to the kernel and the signal-handler frame is eventually unwound when the thread is scheduled again. This approach means that a signal-handler can start on one kernel thread and terminate on a second kernel thread (but the same user thread). It is important to note that signal-handlers save and restore signal masks because user-thread migration can cause signal mask to migrate from one kernel thread to another. This behaviour is only a problem if all kernel threads among which a user thread can migrate differ in terms of signal masks\footnote{Sadly, official POSIX documentation is silent on what distiguishes ``async-signal-safe'' functions from other functions}. However, since the kernel thread hanlding preemption requires a different signal mask, executing user threads on the kernel alarm thread can cause deadlocks. For this reason, the alarm thread is on a tight loop around a system call to \code{sigwaitinfo}, requiring very little CPU time for preemption. One final detail about the alarm thread is how to wake it when additional communication is required (e.g., on thread termination). This unblocking is also done using {\tt SIGALRM}, but sent throught the \code{pthread_sigqueue}. Indeed, \code{sigwait} can differentiate signals sent from \code{pthread_sigqueue} from signals sent from alarms or the kernel.
-\subsection{Scheduler}
-Finally, an aspect that was not mentionned yet is the scheduling algorithm. Currently, the \CFA scheduler uses a single ready queue for all processors, which is the simplest approach to scheduling. Further discussion on scheduling is present in section \label{futur:sched}.
-% ======================================================================
-% ======================================================================
-\section{Internal scheduling} \label{impl:intsched}
-% ======================================================================
-% ======================================================================
-The following figure is the traditional illustration of a monitor (repeated from page~\pageref{fig:monitor} for convenience) :
-\begin{figure}[H]
 \begin{center}
 {\resizebox{0.4\textwidth}{!}{\input{monitor}}}
 \end{center}
-\caption{Traditional illustration of a monitor}
-\label{fig:monitor}
-\end{figure}
+This picture has several components, the two most important being the entry-queue and the AS-stack. The entry-queue is an (almost) FIFO list where threads waiting to enter are parked, while the acceptor-signalor (AS) stack is a FILO list used for threads that have been signalled or otherwise marked as running next.
+For \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 :
-For \CFA, this 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 is non longer helpful to attach the condition to a single monitor. Secondly, the thread waiting on the conditions has to be seperated multiple monitors, which yields :
-\begin{figure}[H]
 \begin{center}
 {\resizebox{0.8\textwidth}{!}{\input{int_monitor}}}
 \end{center}
-\caption{Illustration of \CFA monitor}
-\label{fig:monitor_cfa}
-\end{figure}
+This picture and the proper entry and leave algorithms is the fundamental implementation of internal scheduling (see listing \ref{lst:entry2}). Note that when threads are moved from the condition to the AS-stack, it splits the thread into to pieces. The thread is woken up when all the pieces have moved from the AS-stacks to the active thread seat. In this picture, the threads are split into halves but this is only because there are two monitors in this picture. For a specific signaling operation every monitor needs a piece of thread on its AS-stack.
+\newpage
+\begin{figure}[b]
+This picture and the proper entry and leave algorithms is the fundamental implementation of internal scheduling.
 \begin{multicols}{2}
 Entry
 \begin{pseudo}
+\begin{pseudo}[numbers=left]
 if monitor is free
         enter
 elif already own the monitor
+elif I already own the monitor
         continue
 else
 …
 \columnbreak
 Exit
 \begin{pseudo}
+\begin{pseudo}[numbers=left, firstnumber=8]
 decrement recursion
 if recursion == 0
 …
 \end{pseudo}
 \end{multicols}
-\caption{Entry and exit routine for monitors with internal scheduling}
-\label{lst:entry2}
-\end{figure}
+Some important things to notice about the exit routine. The solution discussed in \ref{intsched} can be seen in the exit routine of listing \ref{lst:entry2}. 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 transferred ownership. This solution is deadlock safe as well as preventing any potential barging. The data structure used for the AS-stack are reused extensively for external scheduling, but in the case of internal scheduling, the data is allocated using variable-length arrays on the callstack of the \code{wait} and \code{signal_block} routines.
+\begin{figure}[H]
+\begin{center}
+{\resizebox{0.8\textwidth}{!}{\input{monitor_structs.pstex_t}}}
+\end{center}
+\caption{Data structures involved in internal/external scheduling}
+\label{fig:structs}
+\end{figure}
+Figure \ref{fig:structs} shows a high level representation of these data-structures. The main idea behind them is that, while figure \ref{fig:monitor_cfa} is a nice illustration in theory, in practice breaking a threads into multiple pieces to put unto intrusive stacks does not make sense. The \code{condition node} is the data structure that is queued into a condition variable and, when signaled, the condition queue is popped and each \code{condition criterion} are moved to the AS-stack. Once all the criterion have be popped from their respective AS-stacks, the thread is woken-up, which is what is shown in listing \ref{lst:entry2}.
+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.
 % ======================================================================
 % ======================================================================
 \section{External scheduling}
+\section{Implementation Details: External scheduling queues}
 % ======================================================================
 % ======================================================================
 Similarly to internal scheduling, external scheduling for multiple monitors relies on the idea that waiting-thread queues are no longer specific to a single monitor, as mentionned in section \ref{extsched}. For internal scheduling, these queues are part of condition variables which are still unique for a given scheduling operation (e.g., no single statment uses multiple conditions). However, in the case of external scheduling, there is no equivalent object which is associated with \code{waitfor} statements. This absence means the queues holding the waiting threads must be stored inside at least one of the monitors that is acquired. The monitors being the only objects that have sufficient lifetime and are available on both sides of the \code{waitfor} statment. This requires an algorithm to choose which monitor holds the relevant queue. It is also important that said algorithm be independent of the order in which users list parameters. The proposed algorithm is to fall back on monitor lock ordering and specify that the monitor that is acquired first is the one with the relevant wainting queue. This assumes that the lock acquiring order is static for the lifetime of all concerned objects but that is a reasonable constraint.
+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.
-This algorithm choice has two consequences :
-\begin{itemize}
-        \item The queue of the highest priority monitor is no longer a true FIFO queue because threads can be moved to the front of the queue. These queues need to contain a set of monitors for each of the waiting threads. Therefore, another thread whose set contains the same highest priority monitor but different lower priority monitors may arrive first but enter the critical section after a thread with the correct pairing.
-        \item The queue of the lowest priority monitor is both required and potentially unused. Indeed, since it is not known at compile time which monitor will be the lowest priority monitor, every monitor needs to have the correct queues even though it is possible that some queues will go unused for the entire duration of the program, for example if a monitor is only used in a specific pair.
-\end{itemize}
+Therefore, the following modifications need to be made to support external scheduling :
+\begin{itemize}
+        \item The threads waiting on the entry-queue need to keep track of which routine is trying to enter, and using which set of monitors. The \code{mutex} routine already has all the required information on its stack so the thread only needs to keep a pointer to that information.
+        \item The monitors need to keep a mask of acceptable routines. This mask contains for each acceptable routine, a routine pointer and an array of monitors to go with it. It also needs storage to keep track of which routine was accepted. Since this information is not specific to any monitor, the monitors actually contain a pointer to an integer on the stack of the waiting thread. Note that the complete mask can be pushed to any owned monitors, regardless of \code{when} statements, the \code{waitfor} statement is used in a context where the thread already has full ownership of (at least) every concerned monitor and therefore monitors will refuse all calls no matter what.
+        \item The entry/exit routine need to be updated as shown in listing \ref{lst:entry3}.
+\end{itemize}
+\subsection{External scheduling - destructors}
+Finally, to support the ordering inversion of destructors, the code generation needs to be modified to use a special entry routine. This routine is needed because of the storage requirements of the call order inversion. Indeed, when waiting for the destructors, storage is need for the waiting context and the lifetime of said storage needs to outlive the waiting operation it is needed for. For regular \code{waitfor} statements, the callstack of the routine itself matches this requirement but it is no longer the case when waiting for the destructor since it is pushed on to the AS-stack for later. The waitfor semantics can then be adjusted correspondingly, as seen in listing \ref{lst:entry-dtor}
+\begin{figure}
+\begin{multicols}{2}
+Entry
+\begin{pseudo}
+if monitor is free
+        enter
+elif already own the monitor
+        continue
+elif matches waitfor mask
+        push criterions to AS-stack
+        continue
+else
+        block
+increment recursion
+\end{pseudo}
+\columnbreak
+Exit
+\begin{pseudo}
+decrement recursion
+if recursion == 0
+        if signal_stack not empty
+                set_owner to thread
+                if all monitors ready
+                        wake-up thread
+                endif
+        endif
+        if entry queue not empty
+                wake-up thread
+        endif
+\end{pseudo}
+\end{multicols}
+\caption{Entry and exit routine for monitors with internal scheduling and external scheduling}
+\label{lst:entry3}
+\end{figure}
+\begin{figure}
+\begin{multicols}{2}
+Destructor Entry
+\begin{pseudo}
+if monitor is free
+        enter
+elif already own the monitor
+        increment recursion
+        return
+create wait context
+if matches waitfor mask
+        reset mask
+        push self to AS-stack
+        baton pass
+else
+        wait
+increment recursion
+\end{pseudo}
+\columnbreak
+Waitfor
+\begin{pseudo}
+if matching thread is already there
+        if found destructor
+                push destructor to AS-stack
+                unlock all monitors
+        else
+                push self to AS-stack
+                baton pass
+        endif
+        return
+endif
+if non-blocking
+        Unlock all monitors
+        Return
+endif
+push self to AS-stack
+set waitfor mask
+block
+return
+\end{pseudo}
+\end{multicols}
+\caption{Pseudo code for the \code{waitfor} routine and the \code{mutex} entry routine for destructors}
+\label{lst:entry-dtor}
+\end{figure}
+\section{Internals}
+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.

doc/proposals/concurrency/text/intro.tex

r0fe4e62	rf5c3b6c
3	3	% ======================================================================
4	4
5		This thesis provides a minimal concurrency \acrshort{api} that is simple, efficient and can be reused to build higher-level features. The simplest possible concurrency system is a thread and a lock but this low-level approach is hard to master. An easier approach for users is to support higher-level constructs as the basis of concurrency. Indeed, for highly productive concurrent programming, high-level approaches are much more popular~\cite{HPP:Study}. Examples are task based, message passing and implicit threading. The high-level approach and its minimal \acrshort{api} are tested in a dialect of C, call \CFA. Furthermore, the proposed \acrshort{api} doubles as an early definition of the \CFA language and library. This thesis also comes with an implementation of the concurrency library for \CFA as well as all the required language features added to the source-to-source translator.
	5	This thesis provides a minimal concurrency \acrshort{api} that is simple, efficient and can be reused to build higher-level features. The simplest possible concurrency system is a thread and a lock but this low-level approach is hard to master. An easier approach for users is to support higher-level constructs as the basis of concurrency. Indeed, for highly productive concurrent programming, high-level approaches are much more popular~\cite{HPP:Study}. Examples are task based, message passing and implicit threading. The high-level approach and its minimal \acrshort{api} are tested in a dialect of C, call \CFA. [Is there value to say that this thesis is also an early definition of the \CFA language and library in regards to concurrency?]
6	6
7	7	There are actually two problems that need to be solved in the design of concurrency for a programming language: which concurrency and which parallelism tools are available to the programmer. While these two concepts are often combined, they are in fact distinct, requiring different tools~\cite{Buhr05a}. Concurrency tools need to handle mutual exclusion and synchronization, while parallelism tools are about performance, cost and resource utilization.

doc/proposals/concurrency/text/parallelism.tex

-              r0fe4e62
+              rf5c3b6c
 Examples of languages that support \glspl{uthread} are Erlang~\cite{Erlang} and \uC~\cite{uC++book}.
 \subsection{Fibers : user-level threads without preemption} \label{fibers}
 A popular varient of \glspl{uthread} is what is often refered to as \glspl{fiber}. However, \glspl{fiber} do not present meaningful semantical differences with \glspl{uthread}. The significant difference between \glspl{uthread} and \glspl{fiber} is the lack of \gls{preemption} in the later one. Advocates of \glspl{fiber} list their high performance and ease of implementation as majors strenghts of \glspl{fiber} but the performance difference between \glspl{uthread} and \glspl{fiber} is controversial, and the ease of implementation, while true, is a weak argument in the context of language design. Therefore this proposal largely ignores fibers.
+\subsection{Fibers : user-level threads without preemption}
+A popular varient of \glspl{uthread} is what is often refered to as \glspl{fiber}. However, \glspl{fiber} do not present meaningful semantical differences with \glspl{uthread}. Advocates of \glspl{fiber} list their high performance and ease of implementation as majors strenghts of \glspl{fiber} but the performance difference between \glspl{uthread} and \glspl{fiber} is controversial, and the ease of implementation, while true, is a weak argument in the context of language design. Therefore this proposal largely ignore fibers.
 An example of a language that uses fibers is Go~\cite{Go}
 …
 \subsection{Paradigm performance}
+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., no 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.
+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 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.
+\TODO
 \section{The \protect\CFA\ Kernel : Processors, Clusters and Threads}\label{kernel}
-\Glspl{cfacluster} have not been fully implmented in the context of this thesis, currently \CFA only supports one \gls{cfacluster}, the initial one. The objective of \gls{cfacluster} is to group \gls{kthread} with identical settings together. \Glspl{uthread} can be scheduled on a \glspl{kthread} of a given \gls{cfacluster}, allowing organization between \glspl{kthread} and \glspl{uthread}. It is important that \glspl{kthread} belonging to a same \glspl{cfacluster} have homogenous settings, otherwise migrating a \gls{uthread} from one \gls{kthread} to the other can cause issues.
 \subsection{Future Work: Machine setup}\label{machine}
 While 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 \cite{affinityLinux, affinityWindows, affinityFreebsd, affinityNetbsd, affinityMacosx} which means it is both possible and desirable for \CFA to offer an abstraction mechanism for portable CPU affinity.
+While 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.
 % \subsection{Paradigms}\label{cfaparadigms}
 % Given these building blocks, it is possible to reproduce all three of the popular paradigms. Indeed, \glspl{uthread} is the default paradigm in \CFA. However, disabling \gls{preemption} on the \gls{cfacluster} means \glspl{cfathread} effectively become \glspl{fiber}. Since several \glspl{cfacluster} with different scheduling policy can coexist in the same application, this allows \glspl{fiber} and \glspl{uthread} to coexist in the runtime of an application. Finally, it is possible to build executors for thread pools from \glspl{uthread} or \glspl{fiber}.
+\subsection{Paradigms}\label{cfaparadigms}
+Given these building blocks, it is possible to reproduce all three of the popular paradigms. Indeed, \glspl{uthread} is the default paradigm in \CFA. However, disabling \glspl{preemption} on the \gls{cfacluster} means \glspl{cfathread} effectively become \glspl{fiber}. Since several \glspl{cfacluster} with different scheduling policy can coexist in the same application, this allows \glspl{fiber} and \glspl{uthread} to coexist in the runtime of an application. Finally, it is possible to build executors for thread pools from \glspl{uthread} or \glspl{fiber}.

doc/proposals/concurrency/text/together.tex

-              r0fe4e62
+              rf5c3b6c
 \section{Threads as monitors}
 As it was subtely alluded in section \ref{threads}, \code{threads} in \CFA are in fact monitors, which means that all monitor 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 :
+As 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 :
 \begin{cfacode}
 // Visualization declaration
 …
+}
 \end{cfacode}
 One of the obvious complaints of the previous code snippet (other than its toy-like simplicity) is that it does not handle exit conditions and just goes on forever. Luckily, the monitor semantics can also be used to clearly enforce a shutdown order in a concise manner :
+One of the obvious complaints of the previous code snippet (other than its toy-like simplicity) is that it does not handle exit conditions and just goes on for ever. Luckily, the monitor semantics can also be used to clearly enforce a shutdown order in a concise manner :
 \begin{cfacode}
 // Visualization declaration
 …
+        }
+}
-// Call destructor for simulator once simulator finishes
-// Call destructor for renderer to signify shutdown
 \end{cfacode}
 \section{Fibers \& Threads}
-As mentionned in section \ref{preemption}, \CFA uses preemptive threads by default but can use fibers on demand. Currently, using fibers is done by adding the following line of code to the program~:
-\begin{cfacode}
-unsigned int default_preemption() {
-        return 0;
+}
-\end{cfacode}
-This function is called by the kernel to fetch the default preemption rate, where 0 signifies an infinite time-slice i.e. no preemption. However, once clusters are fully implemented, it will be possible to create fibers and uthreads in on the same system :
-\begin{figure}
-\begin{cfacode}
-//Cluster forward declaration
-struct cluster;
-//Processor forward declaration
-struct processor;
-//Construct clusters with a preemption rate
-void ?{}(cluster& this, unsigned int rate);
-//Construct processor and add it to cluster
-void ?{}(processor& this, cluster& cluster);
-//Construct thread and schedule it on cluster
-void ?{}(thread& this, cluster& cluster);
-//Declare two clusters
-cluster thread_cluster = { 10`ms };                     //Preempt every 10 ms
-cluster fibers_cluster = { 0 };                         //Never preempt
-//Construct 4 processors
-processor processors[4] = {
-        //2 for the thread cluster
-        thread_cluster;
-        thread_cluster;
-        //2 for the fibers cluster
-        fibers_cluster;
-        fibers_cluster;
-};
-//Declares thread
-thread UThread {};
-void ?{}(UThread& this) {
-        //Construct underlying thread to automatically
-        //be scheduled on the thread cluster
-        (this){ thread_cluster }
+}
-void main(UThread & this);
-//Declares fibers
-thread Fiber {};
-void ?{}(Fiber& this) {
-        //Construct underlying thread to automatically
-        //be scheduled on the fiber cluster
-        (this.__thread){ fibers_cluster }
+}
-void main(Fiber & this);
-\end{cfacode}
-\end{figure}

doc/proposals/concurrency/thesis.tex

-              r0fe4e62
+              rf5c3b6c
 \usepackage[pagewise]{lineno}
 \usepackage{fancyhdr}
-\usepackage{float}
 \renewcommand{\linenumberfont}{\scriptsize\sffamily}
-\usepackage{siunitx}
-\sisetup{ binary-units=true }
 \input{style}                                                   % bespoke macros used in the document
 \usepackage[dvips,plainpages=false,pdfpagelabels,pdfpagemode=UseNone,colorlinks=true,pagebackref=true,linkcolor=blue,citecolor=blue,urlcolor=blue,pagebackref=true,breaklinks=true]{hyperref}
 …
 \input{together}
-\input{results}
 \input{future}

doc/proposals/concurrency/version

r0fe4e62	rf5c3b6c
1		0.1~~1.129~~
	1	0.10.212

src/Common/Debug.h

-              r0fe4e62
+              rf5c3b6c
 #include "SynTree/Declaration.h"
+#define DEBUG
+/// debug codegen a translation unit
+static inline void debugCodeGen( const std::list< Declaration * > & translationUnit, const std::string & label ) {
+        std::list< Declaration * > decls;
+namespace Debug {
+        /// debug codegen a translation unit
+        static inline void codeGen( __attribute__((unused)) const std::list< Declaration * > & translationUnit, __attribute__((unused)) const std::string & label ) {
+        #ifdef DEBUG
+                std::list< Declaration * > decls;
+        filter( translationUnit.begin(), translationUnit.end(), back_inserter( decls ), []( Declaration * decl ) {
+                return ! LinkageSpec::isBuiltin( decl->get_linkage() );
+        });
+                filter( translationUnit.begin(), translationUnit.end(), back_inserter( decls ), []( Declaration * decl ) {
+                        return ! LinkageSpec::isBuiltin( decl->get_linkage() );
+                });
+                std::cerr << "======" << label << "======" << std::endl;
+                CodeGen::generate( decls, std::cerr, false, true );
+        #endif
+        } // dump
+        static inline void treeDump( __attribute__((unused)) const std::list< Declaration * > & translationUnit, __attribute__((unused)) const std::string & label ) {
+        #ifdef DEBUG
+                std::list< Declaration * > decls;
+                filter( translationUnit.begin(), translationUnit.end(), back_inserter( decls ), []( Declaration * decl ) {
+                        return ! LinkageSpec::isBuiltin( decl->get_linkage() );
+                });
+                std::cerr << "======" << label << "======" << std::endl;
+                printAll( decls, std::cerr );
+        #endif
+        } // dump
+}
+        std::cerr << "======" << label << "======" << std::endl;
+        CodeGen::generate( decls, std::cerr, false, true );
+} // dump
 // Local Variables: //

src/Concurrency/Keywords.cc

r0fe4e62	rf5c3b6c
553	553	),
554	554	new ListInit(
555		map_range < std::list<Initializer> > ( args, [](DeclarationWithType var ){
	555	map_range < std::list<Initializer> > ( args, [this](DeclarationWithType var ){
556	556	Type * type = var->get_type()->clone();
557	557	type->set_mutex( false );

src/InitTweak/GenInit.cc

r0fe4e62	rf5c3b6c
214	214	}
215	215	// a type is managed if it appears in the map of known managed types, or if it contains any polymorphism (is a type variable or generic type containing a type variable)
216		return managedTypes.find( SymTab::Mangler::mangle~~Concrete~~( type ) ) != managedTypes.end() \|\| GenPoly::isPolyType( type );
	216	return managedTypes.find( SymTab::Mangler::mangle( type ) ) != managedTypes.end() \|\| GenPoly::isPolyType( type );
217	217	}
218	218
…	…
232	232	Type * type = InitTweak::getPointerBase( params.front()->get_type() );
233	233	assert( type );
234		managedTypes.insert( SymTab::Mangler::mangle~~Concrete~~( type ) );
	234	managedTypes.insert( SymTab::Mangler::mangle( type ) );
235	235	}
236	236	}
…	…
242	242	if ( ObjectDecl * field = dynamic_cast< ObjectDecl * >( member ) ) {
243	243	if ( isManaged( field ) ) {
244		~~// generic parameters should not play a role in determining whether a generic type is constructed - construct all generic types, so that~~
245		~~// polymorphic constructors make generic types managed types~~
246	244	StructInstType inst( Type::Qualifiers(), aggregateDecl );
247		managedTypes.insert( SymTab::Mangler::mangle~~Concrete~~( &inst ) );
	245	managedTypes.insert( SymTab::Mangler::mangle( &inst ) );
248	246	break;
249	247	}

src/InitTweak/InitTweak.cc

-              r0fe4e62
+              rf5c3b6c
         class InitExpander::ExpanderImpl {
         public:
-                virtual ~ExpanderImpl() = default;
                 virtual std::list< Expression * > next( std::list< Expression * > & indices ) = 0;
                 virtual Statement * buildListInit( UntypedExpr * callExpr, std::list< Expression * > & indices ) = 0;
 …
         public:
                 InitImpl( Initializer * init ) : init( init ) {}
-                virtual ~InitImpl() = default;
                 virtual std::list< Expression * > next( __attribute((unused)) std::list< Expression * > & indices ) {
 …
         public:
                 ExprImpl( Expression * expr ) : arg( expr ) {}
+                virtual ~ExprImpl() { delete arg; }
+                ~ExprImpl() { delete arg; }
                 virtual std::list< Expression * > next( std::list< Expression * > & indices ) {

src/ResolvExpr/AlternativeFinder.cc

-              r0fe4e62
+              rf5c3b6c
 #include <memory>                  // for allocator_traits<>::value_type
 #include <utility>                 // for pair
-#include <vector>                  // for vector
 #include "Alternative.h"           // for AltList, Alternative
 …
                 tmpCost.incPoly( -tmpCost.get_polyCost() );
                 if ( tmpCost != Cost::zero ) {
+                // if ( convCost != Cost::zero ) {
                         Type *newType = formalType->clone();
                         env.apply( newType );
 …
 ///     needAssertions.insert( needAssertions.end(), (*tyvar)->get_assertions().begin(), (*tyvar)->get_assertions().end() );
+                }
+        }
+        /// instantiate a single argument by matching actuals from [actualIt, actualEnd) against formalType,
+        /// producing expression(s) in out and their total cost in cost.
+        template< typename AltIterator, typename OutputIterator >
+        bool instantiateArgument( Type * formalType, Initializer * defaultValue, AltIterator & actualIt, AltIterator actualEnd, OpenVarSet & openVars, TypeEnvironment & resultEnv, AssertionSet & resultNeed, AssertionSet & resultHave, const SymTab::Indexer & indexer, Cost & cost, OutputIterator out ) {
+                if ( TupleType * tupleType = dynamic_cast< TupleType * >( formalType ) ) {
+                        // formalType is a TupleType - group actuals into a TupleExpr whose type unifies with the TupleType
+                        std::list< Expression * > exprs;
+                        for ( Type * type : *tupleType ) {
+                                if ( ! instantiateArgument( type, defaultValue, actualIt, actualEnd, openVars, resultEnv, resultNeed, resultHave, indexer, cost, back_inserter( exprs ) ) ) {
+                                        deleteAll( exprs );
+                                        return false;
+                                }
+                        }
+                        *out++ = new TupleExpr( exprs );
+                } else if ( TypeInstType * ttype = Tuples::isTtype( formalType ) ) {
+                        // xxx - mixing default arguments with variadic??
+                        std::list< Expression * > exprs;
+                        for ( ; actualIt != actualEnd; ++actualIt ) {
+                                exprs.push_back( actualIt->expr->clone() );
+                                cost += actualIt->cost;
+                        }
+                        Expression * arg = nullptr;
+                        if ( exprs.size() == 1 && Tuples::isTtype( exprs.front()->get_result() ) ) {
+                                // the case where a ttype value is passed directly is special, e.g. for argument forwarding purposes
+                                // xxx - what if passing multiple arguments, last of which is ttype?
+                                // xxx - what would happen if unify was changed so that unifying tuple types flattened both before unifying lists? then pass in TupleType(ttype) below.
+                                arg = exprs.front();
+                        } else {
+                                arg = new TupleExpr( exprs );
+                        }
+                        assert( arg && arg->get_result() );
+                        if ( ! unify( ttype, arg->get_result(), resultEnv, resultNeed, resultHave, openVars, indexer ) ) {
+                                return false;
+                        }
+                        *out++ = arg;
+                } else if ( actualIt != actualEnd ) {
+                        // both actualType and formalType are atomic (non-tuple) types - if they unify
+                        // then accept actual as an argument, otherwise return false (fail to instantiate argument)
+                        Expression * actual = actualIt->expr;
+                        Type * actualType = actual->get_result();
+                        PRINT(
+                                std::cerr << "formal type is ";
+                                formalType->print( std::cerr );
+                                std::cerr << std::endl << "actual type is ";
+                                actualType->print( std::cerr );
+                                std::cerr << std::endl;
+                        )
+                        if ( ! unify( formalType, actualType, resultEnv, resultNeed, resultHave, openVars, indexer ) ) {
+                                // std::cerr << "unify failed" << std::endl;
+                                return false;
+                        }
+                        // move the expression from the alternative to the output iterator
+                        *out++ = actual;
+                        actualIt->expr = nullptr;
+                        cost += actualIt->cost;
+                        ++actualIt;
+                } else {
+                        // End of actuals - Handle default values
+                        if ( SingleInit *si = dynamic_cast<SingleInit *>( defaultValue )) {
+                                if ( CastExpr * castExpr = dynamic_cast< CastExpr * >( si->get_value() ) ) {
+                                        // so far, only constant expressions are accepted as default values
+                                        if ( ConstantExpr *cnstexpr = dynamic_cast<ConstantExpr *>( castExpr->get_arg() ) ) {
+                                                if ( Constant *cnst = dynamic_cast<Constant *>( cnstexpr->get_constant() ) ) {
+                                                        if ( unify( formalType, cnst->get_type(), resultEnv, resultNeed, resultHave, openVars, indexer ) ) {
+                                                                *out++ = cnstexpr->clone();
+                                                                return true;
+                                                        } // if
+                                                } // if
+                                        } // if
+                                }
+                        } // if
+                        return false;
+                } // if
+                return true;
+        }
+        bool AlternativeFinder::instantiateFunction( std::list< DeclarationWithType* >& formals, const AltList &actuals, bool isVarArgs, OpenVarSet& openVars, TypeEnvironment &resultEnv, AssertionSet &resultNeed, AssertionSet &resultHave, AltList & out ) {
+                simpleCombineEnvironments( actuals.begin(), actuals.end(), resultEnv );
+                // make sure we don't widen any existing bindings
+                for ( TypeEnvironment::iterator i = resultEnv.begin(); i != resultEnv.end(); ++i ) {
+                        i->allowWidening = false;
+                }
+                resultEnv.extractOpenVars( openVars );
+                // flatten actuals so that each actual has an atomic (non-tuple) type
+                AltList exploded;
+                Tuples::explode( actuals, indexer, back_inserter( exploded ) );
+                AltList::iterator actualExpr = exploded.begin();
+                AltList::iterator actualEnd = exploded.end();
+                for ( DeclarationWithType * formal : formals ) {
+                        // match flattened actuals with formal parameters - actuals will be grouped to match
+                        // with formals as appropriate
+                        Cost cost = Cost::zero;
+                        std::list< Expression * > newExprs;
+                        ObjectDecl * obj = strict_dynamic_cast< ObjectDecl * >( formal );
+                        if ( ! instantiateArgument( obj->get_type(), obj->get_init(), actualExpr, actualEnd, openVars, resultEnv, resultNeed, resultHave, indexer, cost, back_inserter( newExprs ) ) ) {
+                                deleteAll( newExprs );
+                                return false;
+                        }
+                        // success - produce argument as a new alternative
+                        assert( newExprs.size() == 1 );
+                        out.push_back( Alternative( newExprs.front(), resultEnv, cost ) );
+                }
+                if ( actualExpr != actualEnd ) {
+                        // there are still actuals remaining, but we've run out of formal parameters to match against
+                        // this is okay only if the function is variadic
+                        if ( ! isVarArgs ) {
+                                return false;
+                        }
+                        out.splice( out.end(), exploded, actualExpr, actualEnd );
+                }
+                return true;
+        }
 …
+        }
+        /// Gets a default value from an initializer, nullptr if not present
+        ConstantExpr* getDefaultValue( Initializer* init ) {
+                if ( SingleInit* si = dynamic_cast<SingleInit*>( init ) ) {
+                        if ( CastExpr* ce = dynamic_cast<CastExpr*>( si->get_value() ) ) {
+                                return dynamic_cast<ConstantExpr*>( ce->get_arg() );
+                        }
+                }
+                return nullptr;
+        }
+        /// State to iteratively build a match of parameter expressions to arguments
+        struct ArgPack {
+                AltList actuals;                 ///< Arguments included in this pack
+                TypeEnvironment env;             ///< Environment for this pack
+                AssertionSet need;               ///< Assertions outstanding for this pack
+                AssertionSet have;               ///< Assertions found for this pack
+                OpenVarSet openVars;             ///< Open variables for this pack
+                unsigned nextArg;                ///< Index of next argument in arguments list
+                std::vector<Alternative> expls;  ///< Exploded actuals left over from last match
+                unsigned nextExpl;               ///< Index of next exploded alternative to use
+                std::vector<unsigned> tupleEls;  /// Number of elements in current tuple element(s)
+                ArgPack(const TypeEnvironment& env, const AssertionSet& need, const AssertionSet& have,
+                                const OpenVarSet& openVars)
+                        : actuals(), env(env), need(need), have(have), openVars(openVars), nextArg(0),
+                          expls(), nextExpl(0), tupleEls() {}
+                /// Starts a new tuple expression
+                void beginTuple() {
+                        if ( ! tupleEls.empty() ) ++tupleEls.back();
+                        tupleEls.push_back(0);
+                }
+                /// Ends a tuple expression, consolidating the appropriate actuals
+                void endTuple() {
+                        // set up new Tuple alternative
+                        std::list<Expression*> exprs;
+                        Cost cost = Cost::zero;
+                        // transfer elements into alternative
+                        for (unsigned i = 0; i < tupleEls.back(); ++i) {
+                                exprs.push_front( actuals.back().expr );
+                                actuals.back().expr = nullptr;
+                                cost += actuals.back().cost;
+                                actuals.pop_back();
+                        }
+                        tupleEls.pop_back();
+                        // build new alternative
+                        actuals.emplace_back( new TupleExpr( exprs ), this->env, cost );
+                }
+                /// Clones and adds an actual, returns this
+                ArgPack& withArg( Expression* expr, Cost cost = Cost::zero ) {
+                        actuals.emplace_back( expr->clone(), this->env, cost );
+                        if ( ! tupleEls.empty() ) ++tupleEls.back();
+                        return *this;
+                }
+        };
+        /// Instantiates an argument to match a formal, returns false if no results left
+        bool instantiateArgument( Type* formalType, Initializer* initializer,
+                        const std::vector< AlternativeFinder >& args,
+                        std::vector<ArgPack>& results, std::vector<ArgPack>& nextResults,
+                        const SymTab::Indexer& indexer ) {
+                if ( TupleType* tupleType = dynamic_cast<TupleType*>( formalType ) ) {
+                        // formalType is a TupleType - group actuals into a TupleExpr
+                        for ( ArgPack& result : results ) { result.beginTuple(); }
+                        for ( Type* type : *tupleType ) {
+                                // xxx - dropping initializer changes behaviour from previous, but seems correct
+                                if ( ! instantiateArgument( type, nullptr, args, results, nextResults, indexer ) )
+                                        return false;
+                        }
+                        for ( ArgPack& result : results ) { result.endTuple(); }
+                        return true;
+                } else if ( TypeInstType* ttype = Tuples::isTtype( formalType ) ) {
+                        // formalType is a ttype, consumes all remaining arguments
+                        // xxx - mixing default arguments with variadic??
+                        std::vector<ArgPack> finalResults{};  /// list of completed tuples
+                        // start tuples
+                        for ( ArgPack& result : results ) {
+                                result.beginTuple();
+                                // use rest of exploded tuple if present
+                                while ( result.nextExpl < result.expls.size() ) {
+                                        const Alternative& actual = result.expls[result.nextExpl];
+                                        result.env.addActual( actual.env, result.openVars );
+                                        result.withArg( actual.expr );
+                                        ++result.nextExpl;
+                                }
+                        }
+                        // iterate until all results completed
+                        while ( ! results.empty() ) {
+                                // add another argument to results
+                                for ( ArgPack& result : results ) {
+                                        // finish result when out of arguments
+                                        if ( result.nextArg >= args.size() ) {
+                                                Type* argType = result.actuals.back().expr->get_result();
+                                                if ( result.tupleEls.back() == 1 && Tuples::isTtype( argType ) ) {
+                                                        // the case where a ttype value is passed directly is special, e.g. for
+                                                        // argument forwarding purposes
+                                                        // xxx - what if passing multiple arguments, last of which is ttype?
+                                                        // xxx - what would happen if unify was changed so that unifying tuple
+                                                        // types flattened both before unifying lists? then pass in TupleType
+                                                        // (ttype) below.
+                                                        result.tupleEls.pop_back();
+                                                } else {
+                                                        // collapse leftover arguments into tuple
+                                                        result.endTuple();
+                                                        argType = result.actuals.back().expr->get_result();
+                                                }
+                                                // check unification for ttype before adding to final
+                                                if ( unify( ttype, argType, result.env, result.need, result.have,
+                                                                result.openVars, indexer ) ) {
+                                                        finalResults.push_back( std::move(result) );
+                                                }
+                                                continue;
+                                        }
+                                        // add each possible next argument
+                                        for ( const Alternative& actual : args[result.nextArg] ) {
+                                                ArgPack aResult = result;  // copy to clone everything
+                                                // add details of actual to result
+                                                aResult.env.addActual( actual.env, aResult.openVars );
+                                                Cost cost = actual.cost;
+                                                // explode argument
+                                                std::vector<Alternative> exploded;
+                                                Tuples::explode( actual, indexer, back_inserter( exploded ) );
+                                                // add exploded argument to tuple
+                                                for ( Alternative& aActual : exploded ) {
+                                                        aResult.withArg( aActual.expr, cost );
+                                                        cost = Cost::zero;
+                                                }
+                                                ++aResult.nextArg;
+                                                nextResults.push_back( std::move(aResult) );
+                                        }
+                                }
+                                // reset for next round
+                                results.swap( nextResults );
+                                nextResults.clear();
+                        }
+                        results.swap( finalResults );
+                        return ! results.empty();
+                }
+                // iterate each current subresult
+                for ( unsigned iResult = 0; iResult < results.size(); ++iResult ) {
+                        ArgPack& result = results[iResult];
+                        if ( result.nextExpl < result.expls.size() ) {
+                                // use remainder of exploded tuple if present
+                                const Alternative& actual = result.expls[result.nextExpl];
+                                result.env.addActual( actual.env, result.openVars );
+                                Type* actualType = actual.expr->get_result();
+                                PRINT(
+                                        std::cerr << "formal type is ";
+                                        formalType->print( std::cerr );
+                                        std::cerr << std::endl << "actual type is ";
+                                        actualType->print( std::cerr );
+                                        std::cerr << std::endl;
+                                )
+                                if ( unify( formalType, actualType, result.env, result.need, result.have,
+                                                result.openVars, indexer ) ) {
+                                        ++result.nextExpl;
+                                        nextResults.push_back( std::move(result.withArg( actual.expr )) );
+                                }
+                                continue;
+                        } else if ( result.nextArg >= args.size() ) {
+                                // use default initializers if out of arguments
+                                if ( ConstantExpr* cnstExpr = getDefaultValue( initializer ) ) {
+                                        if ( Constant* cnst = dynamic_cast<Constant*>( cnstExpr->get_constant() ) ) {
+                                                if ( unify( formalType, cnst->get_type(), result.env, result.need,
+                                                                result.have, result.openVars, indexer ) ) {
+                                                        nextResults.push_back( std::move(result.withArg( cnstExpr )) );
+                                                }
+                                        }
+                                }
+                                continue;
+                        }
+                        // Check each possible next argument
+                        for ( const Alternative& actual : args[result.nextArg] ) {
+                                ArgPack aResult = result;  // copy to clone everything
+                                // add details of actual to result
+                                aResult.env.addActual( actual.env, aResult.openVars );
+                                // explode argument
+                                std::vector<Alternative> exploded;
+                                Tuples::explode( actual, indexer, back_inserter( exploded ) );
+                                if ( exploded.empty() ) {
+                                        // skip empty tuple arguments
+                                        ++aResult.nextArg;
+                                        results.push_back( std::move(aResult) );
+                                        continue;
+                                }
+                                // consider only first exploded actual
+                                const Alternative& aActual = exploded.front();
+                                Type* actualType = aActual.expr->get_result()->clone();
+                                PRINT(
+                                        std::cerr << "formal type is ";
+                                        formalType->print( std::cerr );
+                                        std::cerr << std::endl << "actual type is ";
+                                        actualType->print( std::cerr );
+                                        std::cerr << std::endl;
+                                )
+                                // attempt to unify types
+                                if ( unify( formalType, actualType, aResult.env, aResult.need, aResult.have, aResult.openVars, indexer ) ) {
+                                        // add argument
+                                        aResult.withArg( aActual.expr, actual.cost );
+                                        ++aResult.nextArg;
+                                        if ( exploded.size() > 1 ) {
+                                                // other parts of tuple left over
+                                                aResult.expls = std::move( exploded );
+                                                aResult.nextExpl = 1;
+                                        }
+                                        nextResults.push_back( std::move(aResult) );
+                                }
+                        }
+                }
+                // reset for next parameter
+                results.swap( nextResults );
+                nextResults.clear();
+                return ! results.empty();
+        }
+        template<typename OutputIterator>
+        void AlternativeFinder::makeFunctionAlternatives( const Alternative &func,
+                        FunctionType *funcType, const std::vector< AlternativeFinder > &args,
+                        OutputIterator out ) {
+                OpenVarSet funcOpenVars;
+                AssertionSet funcNeed, funcHave;
+                TypeEnvironment funcEnv( func.env );
+                makeUnifiableVars( funcType, funcOpenVars, funcNeed );
+                // add all type variables as open variables now so that those not used in the parameter
+                // list are still considered open.
+                funcEnv.add( funcType->get_forall() );
+        template< typename OutputIterator >
+        void AlternativeFinder::makeFunctionAlternatives( const Alternative &func, FunctionType *funcType, const AltList &actualAlt, OutputIterator out ) {
+                OpenVarSet openVars;
+                AssertionSet resultNeed, resultHave;
+                TypeEnvironment resultEnv( func.env );
+                makeUnifiableVars( funcType, openVars, resultNeed );
+                resultEnv.add( funcType->get_forall() ); // add all type variables as open variables now so that those not used in the parameter list are still considered open
+                AltList instantiatedActuals; // filled by instantiate function
                 if ( targetType && ! targetType->isVoid() && ! funcType->get_returnVals().empty() ) {
                         // attempt to narrow based on expected target type
                         Type * returnType = funcType->get_returnVals().front()->get_type();
+                        if ( ! unify( returnType, targetType, funcEnv, funcNeed, funcHave, funcOpenVars,
+                                        indexer ) ) {
+                                // unification failed, don't pursue this function alternative
+                        if ( ! unify( returnType, targetType, resultEnv, resultNeed, resultHave, openVars, indexer ) ) {
+                                // unification failed, don't pursue this alternative
                                 return;
+                        }
+                }
+                // iteratively build matches, one parameter at a time
+                std::vector<ArgPack> results{ ArgPack{ funcEnv, funcNeed, funcHave, funcOpenVars } };
+                std::vector<ArgPack> nextResults{};
+                for ( DeclarationWithType* formal : funcType->get_parameters() ) {
+                        ObjectDecl* obj = strict_dynamic_cast< ObjectDecl* >( formal );
+                        if ( ! instantiateArgument(
+                                        obj->get_type(), obj->get_init(), args, results, nextResults, indexer ) )
+                                return;
+                }
+                // filter out results that don't use all the arguments, and aren't variadic
+                std::vector<ArgPack> finalResults{};
+                if ( funcType->get_isVarArgs() ) {
+                        for ( ArgPack& result : results ) {
+                                // use rest of exploded tuple if present
+                                while ( result.nextExpl < result.expls.size() ) {
+                                        const Alternative& actual = result.expls[result.nextExpl];
+                                        result.env.addActual( actual.env, result.openVars );
+                                        result.withArg( actual.expr );
+                                        ++result.nextExpl;
+                                }
+                        }
+                        while ( ! results.empty() ) {
+                                // build combinations for all remaining arguments
+                                for ( ArgPack& result : results ) {
+                                        // keep if used all arguments
+                                        if ( result.nextArg >= args.size() ) {
+                                                finalResults.push_back( std::move(result) );
+                                                continue;
+                                        }
+                                        // add each possible next argument
+                                        for ( const Alternative& actual : args[result.nextArg] ) {
+                                                ArgPack aResult = result; // copy to clone everything
+                                                // add details of actual to result
+                                                aResult.env.addActual( actual.env, aResult.openVars );
+                                                Cost cost = actual.cost;
+                                                // explode argument
+                                                std::vector<Alternative> exploded;
+                                                Tuples::explode( actual, indexer, back_inserter( exploded ) );
+                                                // add exploded argument to arg list
+                                                for ( Alternative& aActual : exploded ) {
+                                                        aResult.withArg( aActual.expr, cost );
+                                                        cost = Cost::zero;
+                                                }
+                                                ++aResult.nextArg;
+                                                nextResults.push_back( std::move(aResult) );
+                                        }
+                                }
+                                // reset for next round
+                                results.swap( nextResults );
+                                nextResults.clear();
+                        }
+                } else {
+                        // filter out results that don't use all the arguments
+                        for ( ArgPack& result : results ) {
+                                if ( result.nextExpl >= result.expls.size() && result.nextArg >= args.size() ) {
+                                        finalResults.push_back( std::move(result) );
+                                }
+                        }
+                }
+                // validate matching combos, add to final result list
+                for ( ArgPack& result : finalResults ) {
+                if ( instantiateFunction( funcType->get_parameters(), actualAlt, funcType->get_isVarArgs(), openVars, resultEnv, resultNeed, resultHave, instantiatedActuals ) ) {
                         ApplicationExpr *appExpr = new ApplicationExpr( func.expr->clone() );
                         Alternative newAlt( appExpr, result.env, sumCost( result.actuals ) );
                         makeExprList( result.actuals, appExpr->get_args() );
+                        Alternative newAlt( appExpr, resultEnv, sumCost( instantiatedActuals ) );
+                        makeExprList( instantiatedActuals, appExpr->get_args() );
                         PRINT(
                                 std::cerr << "instantiate function success: " << appExpr << std::endl;
                                 std::cerr << "need assertions:" << std::endl;
                                 printAssertionSet( result.need, std::cerr, 8 );
+                                printAssertionSet( resultNeed, std::cerr, 8 );
+                        )
                         inferParameters( result.need, result.have, newAlt, result.openVars, out );
+                        inferParameters( resultNeed, resultHave, newAlt, openVars, out );
+                }
+        }
 …
                 if ( funcFinder.alternatives.empty() ) return;
+                std::vector< AlternativeFinder > argAlternatives;
+                findSubExprs( untypedExpr->begin_args(), untypedExpr->end_args(),
+                        back_inserter( argAlternatives ) );
+                std::list< AlternativeFinder > argAlternatives;
+                findSubExprs( untypedExpr->begin_args(), untypedExpr->end_args(), back_inserter( argAlternatives ) );
+                std::list< AltList > possibilities;
+                combos( argAlternatives.begin(), argAlternatives.end(), back_inserter( possibilities ) );
                 // take care of possible tuple assignments
                 // if not tuple assignment, assignment is taken care of as a normal function call
                 Tuples::handleTupleAssignment( *this, untypedExpr, argAlternatives );
+                Tuples::handleTupleAssignment( *this, untypedExpr, possibilities );
                 // find function operators
 …
                                                 Alternative newFunc( *func );
                                                 referenceToRvalueConversion( newFunc.expr );
+                                                makeFunctionAlternatives( newFunc, function, argAlternatives,
+                                                        std::back_inserter( candidates ) );
+                                                for ( std::list< AltList >::iterator actualAlt = possibilities.begin(); actualAlt != possibilities.end(); ++actualAlt ) {
+                                                        // XXX
+                                                        //Designators::check_alternative( function, *actualAlt );
+                                                        makeFunctionAlternatives( newFunc, function, *actualAlt, std::back_inserter( candidates ) );
+                                                }
+                                        }
                                 } else if ( TypeInstType *typeInst = dynamic_cast< TypeInstType* >( func->expr->get_result()->stripReferences() ) ) { // handle ftype (e.g. *? on function pointer)
 …
                                                         Alternative newFunc( *func );
                                                         referenceToRvalueConversion( newFunc.expr );
+                                                        makeFunctionAlternatives( newFunc, function, argAlternatives,
+                                                                std::back_inserter( candidates ) );
+                                                        for ( std::list< AltList >::iterator actualAlt = possibilities.begin(); actualAlt != possibilities.end(); ++actualAlt ) {
+                                                                makeFunctionAlternatives( newFunc, function, *actualAlt, std::back_inserter( candidates ) );
+                                                        } // for
                                                 } // if
                                         } // if
+                                }
+                                // try each function operator ?() with the current function alternative and each of the argument combinations
+                                for ( AltList::iterator funcOp = funcOpFinder.alternatives.begin(); funcOp != funcOpFinder.alternatives.end(); ++funcOp ) {
+                                        // check if the type is pointer to function
+                                        if ( PointerType *pointer = dynamic_cast< PointerType* >( funcOp->expr->get_result()->stripReferences() ) ) {
+                                                if ( FunctionType *function = dynamic_cast< FunctionType* >( pointer->get_base() ) ) {
+                                                        Alternative newFunc( *funcOp );
+                                                        referenceToRvalueConversion( newFunc.expr );
+                                                        for ( std::list< AltList >::iterator actualAlt = possibilities.begin(); actualAlt != possibilities.end(); ++actualAlt ) {
+                                                                AltList currentAlt;
+                                                                currentAlt.push_back( *func );
+                                                                currentAlt.insert( currentAlt.end(), actualAlt->begin(), actualAlt->end() );
+                                                                makeFunctionAlternatives( newFunc, function, currentAlt, std::back_inserter( candidates ) );
+                                                        } // for
+                                                } // if
+                                        } // if
+                                } // for
                         } catch ( SemanticError &e ) {
                                 errors.append( e );
+                        }
                 } // for
-                // try each function operator ?() with each function alternative
-                if ( ! funcOpFinder.alternatives.empty() ) {
-                        // add function alternatives to front of argument list
-                        argAlternatives.insert( argAlternatives.begin(), std::move(funcFinder) );
-                        for ( AltList::iterator funcOp = funcOpFinder.alternatives.begin();
-                                        funcOp != funcOpFinder.alternatives.end(); ++funcOp ) {
-                                try {
-                                        // check if type is a pointer to function
-                                        if ( PointerType* pointer = dynamic_cast<PointerType*>(
-                                                        funcOp->expr->get_result()->stripReferences() ) ) {
-                                                if ( FunctionType* function =
-                                                                dynamic_cast<FunctionType*>( pointer->get_base() ) ) {
-                                                        Alternative newFunc( *funcOp );
-                                                        referenceToRvalueConversion( newFunc.expr );
-                                                        makeFunctionAlternatives( newFunc, function, argAlternatives,
-                                                                std::back_inserter( candidates ) );
+                                                }
+                                        }
-                                } catch ( SemanticError &e ) {
-                                        errors.append( e );
+                                }
+                        }
+                }
                 // Implement SFINAE; resolution errors are only errors if there aren't any non-erroneous resolutions
 …
                 candidates.splice( candidates.end(), alternatives );
+                // use a new list so that alternatives are not examined by addAnonConversions twice.
+                AltList winners;
+                findMinCost( candidates.begin(), candidates.end(), std::back_inserter( winners ) );
+                findMinCost( candidates.begin(), candidates.end(), std::back_inserter( alternatives ) );
                 // function may return struct or union value, in which case we need to add alternatives for implicit
                 // conversions to each of the anonymous members, must happen after findMinCost since anon conversions
                 // are never the cheapest expression
                 for ( const Alternative & alt : winners ) {
+                for ( const Alternative & alt : alternatives ) {
                         addAnonConversions( alt );
+                }
-                alternatives.splice( alternatives.begin(), winners );
                 if ( alternatives.empty() && targetType && ! targetType->isVoid() ) {

src/ResolvExpr/AlternativeFinder.h

-              r0fe4e62
+              rf5c3b6c
           public:
                 AlternativeFinder( const SymTab::Indexer &indexer, const TypeEnvironment &env );
-                AlternativeFinder( const AlternativeFinder& o )
-                        : indexer(o.indexer), alternatives(o.alternatives), env(o.env),
-                          targetType(o.targetType) {}
-                AlternativeFinder( AlternativeFinder&& o )
-                        : indexer(o.indexer), alternatives(std::move(o.alternatives)), env(o.env),
-                          targetType(o.targetType) {}
-                AlternativeFinder& operator= ( const AlternativeFinder& o ) {
-                        if (&o == this) return *this;
-                        // horrific nasty hack to rebind references...
-                        alternatives.~AltList();
-                        new(this) AlternativeFinder(o);
-                        return *this;
+                }
-                AlternativeFinder& operator= ( AlternativeFinder&& o ) {
-                        if (&o == this) return *this;
-                        // horrific nasty hack to rebind references...
-                        alternatives.~AltList();
-                        new(this) AlternativeFinder(std::move(o));
-                        return *this;
+                }
                 void find( Expression *expr, bool adjust = false, bool prune = true, bool failFast = true );
                 /// Calls find with the adjust flag set; adjustment turns array and function types into equivalent pointer types
 …
                 /// Adds alternatives for offsetof expressions, given the base type and name of the member
                 template< typename StructOrUnionType > void addOffsetof( StructOrUnionType *aggInst, const std::string &name );
+                template<typename OutputIterator>
+                void makeFunctionAlternatives( const Alternative &func, FunctionType *funcType, const std::vector< AlternativeFinder >& args, OutputIterator out );
+                bool instantiateFunction( std::list< DeclarationWithType* >& formals, const AltList &actuals, bool isVarArgs, OpenVarSet& openVars, TypeEnvironment &resultEnv, AssertionSet &resultNeed, AssertionSet &resultHave, AltList & out );
+                template< typename OutputIterator >
+                void makeFunctionAlternatives( const Alternative &func, FunctionType *funcType, const AltList &actualAlt, OutputIterator out );
                 template< typename OutputIterator >
                 void inferParameters( const AssertionSet &need, AssertionSet &have, const Alternative &newAlt, OpenVarSet &openVars, OutputIterator out );

src/ResolvExpr/CurrentObject.cc

r0fe4e62	rf5c3b6c
260	260
261	261	AggregateIterator( const std::string & kind, const std::string & name, Type * inst, const MemberList & members ) : kind( kind ), name( name ), inst( inst ), members( members ), curMember( members.begin() ), sub( makeGenericSubstitution( inst ) ) {
262		~~PRINT( std::cerr << "Creating " << kind << "(" << name << ")"; )~~
263	262	init();
264	263	}

src/ResolvExpr/RenameVars.cc

-              r0fe4e62
+              rf5c3b6c
         RenameVars global_renamer;
         RenameVars::RenameVars() : level( 0 ), resetCount( 0 ) {
+        RenameVars::RenameVars() : level( 0 ) {
                 mapStack.push_front( std::map< std::string, std::string >() );
+        }
 …
         void RenameVars::reset() {
                 level = 0;
-                resetCount++;
+        }
 …
                         for ( Type::ForallList::iterator i = type->get_forall().begin(); i != type->get_forall().end(); ++i ) {
                                 std::ostringstream output;
                                 output << "_" << resetCount << "_" << level << "_" << (*i)->get_name();
+                                output << "_" << level << "_" << (*i)->get_name();
                                 std::string newname( output.str() );
                                 mapStack.front()[ (*i)->get_name() ] = newname;

src/ResolvExpr/RenameVars.h

r0fe4e62	rf5c3b6c
48	48	void typeBefore( Type *type );
49	49	void typeAfter( Type *type );
50		int level~~, resetCount~~;
	50	int level;
51	51	std::list< std::map< std::string, std::string > > mapStack;
52	52	};

src/ResolvExpr/TypeEnvironment.cc

-              r0fe4e62
+              rf5c3b6c
+        }
-        void TypeEnvironment::addActual( const TypeEnvironment& actualEnv, OpenVarSet& openVars ) {
-                for ( const EqvClass& c : actualEnv ) {
-                        EqvClass c2 = c;
-                        c2.allowWidening = false;
-                        for ( const std::string& var : c2.vars ) {
-                                openVars[ var ] = c2.data;
+                        }
-                        env.push_back( std::move(c2) );
+                }
+        }
 } // namespace ResolvExpr

src/ResolvExpr/TypeEnvironment.h

-              r0fe4e62
+              rf5c3b6c
                 TypeEnvironment *clone() const { return new TypeEnvironment( *this ); }
-                /// Iteratively adds the environment of a new actual (with allowWidening = false),
-                /// and extracts open variables.
-                void addActual( const TypeEnvironment& actualEnv, OpenVarSet& openVars );
                 typedef std::list< EqvClass >::iterator iterator;
                 iterator begin() { return env.begin(); }

src/SymTab/Mangler.cc

-              r0fe4e62
+              rf5c3b6c
 namespace SymTab {
         std::string Mangler::mangleType( Type * ty ) {
                 Mangler mangler( false, true, true );
+                Mangler mangler( false, true );
                 maybeAccept( ty, mangler );
                 return mangler.get_mangleName();
+        }
+        std::string Mangler::mangleConcrete( Type* ty ) {
+                Mangler mangler( false, false, false );
+                maybeAccept( ty, mangler );
+                return mangler.get_mangleName();
+        }
+        Mangler::Mangler( bool mangleOverridable, bool typeMode, bool mangleGenericParams )
+                : nextVarNum( 0 ), isTopLevel( true ), mangleOverridable( mangleOverridable ), typeMode( typeMode ), mangleGenericParams( mangleGenericParams ) {}
+        Mangler::Mangler( bool mangleOverridable, bool typeMode )
+                : nextVarNum( 0 ), isTopLevel( true ), mangleOverridable( mangleOverridable ), typeMode( typeMode ) {}
         Mangler::Mangler( const Mangler &rhs ) : mangleName() {
 …
                 mangleName << ( refType->get_name().length() + prefix.length() ) << prefix << refType->get_name();
+                if ( mangleGenericParams ) {
+                        std::list< Expression* >& params = refType->get_parameters();
+                        if ( ! params.empty() ) {
+                                mangleName << "_";
+                                for ( std::list< Expression* >::const_iterator param = params.begin(); param != params.end(); ++param ) {
+                                        TypeExpr *paramType = dynamic_cast< TypeExpr* >( *param );
+                                        assertf(paramType, "Aggregate parameters should be type expressions: %s", toString(*param).c_str());
+                                        maybeAccept( paramType->get_type(), *this );
+                                }
+                                mangleName << "_";
+        }
+        void Mangler::mangleGenericRef( ReferenceToType * refType, std::string prefix ) {
+                printQualifiers( refType );
+                std::ostringstream oldName( mangleName.str() );
+                mangleName.clear();
+                mangleName << prefix << refType->get_name();
+                std::list< Expression* >& params = refType->get_parameters();
+                if ( ! params.empty() ) {
+                        mangleName << "_";
+                        for ( std::list< Expression* >::const_iterator param = params.begin(); param != params.end(); ++param ) {
+                                TypeExpr *paramType = dynamic_cast< TypeExpr* >( *param );
+                                assertf(paramType, "Aggregate parameters should be type expressions: %s", toString(*param).c_str());
+                                maybeAccept( paramType->get_type(), *this );
+                        }
+                        mangleName << "_";
+                }
+                oldName << mangleName.str().length() << mangleName.str();
+                mangleName.str( oldName.str() );
+        }
         void Mangler::visit( StructInstType * aggregateUseType ) {
+                mangleRef( aggregateUseType, "s" );
+                if ( typeMode ) mangleGenericRef( aggregateUseType, "s" );
+                else mangleRef( aggregateUseType, "s" );
+        }
         void Mangler::visit( UnionInstType * aggregateUseType ) {
+                mangleRef( aggregateUseType, "u" );
+                if ( typeMode ) mangleGenericRef( aggregateUseType, "u" );
+                else mangleRef( aggregateUseType, "u" );
+        }
 …
                                 varNums[ (*i)->name ] = std::pair< int, int >( nextVarNum++, (int)(*i)->get_kind() );
                                 for ( std::list< DeclarationWithType* >::iterator assert = (*i)->assertions.begin(); assert != (*i)->assertions.end(); ++assert ) {
                                         Mangler sub_mangler( mangleOverridable, typeMode, mangleGenericParams );
+                                        Mangler sub_mangler( mangleOverridable, typeMode );
                                         sub_mangler.nextVarNum = nextVarNum;
                                         sub_mangler.isTopLevel = false;

src/SymTab/Mangler.h

-              r0fe4e62
+              rf5c3b6c
                 /// Mangle syntax tree object; primary interface to clients
                 template< typename SynTreeClass >
             static std::string mangle( SynTreeClass *decl, bool mangleOverridable = true, bool typeMode = false, bool mangleGenericParams = true );
+            static std::string mangle( SynTreeClass *decl, bool mangleOverridable = true, bool typeMode = false );
                 /// Mangle a type name; secondary interface
                 static std::string mangleType( Type* ty );
-                /// Mangle ignoring generic type parameters
-                static std::string mangleConcrete( Type* ty );
                 virtual void visit( ObjectDecl *declaration );
 …
                 bool mangleOverridable;         ///< Specially mangle overridable built-in methods
                 bool typeMode;                  ///< Produce a unique mangled name for a type
-                bool mangleGenericParams;       ///< Include generic parameters in name mangling if true
                 Mangler( bool mangleOverridable, bool typeMode, bool mangleGenericParams );
+                Mangler( bool mangleOverridable, bool typeMode );
                 Mangler( const Mangler & );
                 void mangleDecl( DeclarationWithType *declaration );
                 void mangleRef( ReferenceToType *refType, std::string prefix );
+                void mangleGenericRef( ReferenceToType *refType, std::string prefix );
                 void printQualifiers( Type *type );
 …
         template< typename SynTreeClass >
         std::string Mangler::mangle( SynTreeClass *decl, bool mangleOverridable, bool typeMode, bool mangleGenericParams ) {
                 Mangler mangler( mangleOverridable, typeMode, mangleGenericParams );
+        std::string Mangler::mangle( SynTreeClass *decl, bool mangleOverridable, bool typeMode ) {
+                Mangler mangler( mangleOverridable, typeMode );
                 maybeAccept( decl, mangler );
                 return mangler.get_mangleName();

src/SymTab/Validate.cc

-              r0fe4e62
+              rf5c3b6c
                 HoistStruct::hoistStruct( translationUnit ); // must happen after EliminateTypedef, so that aggregate typedefs occur in the correct order
                 ReturnTypeFixer::fix( translationUnit ); // must happen before autogen
-                acceptAll( translationUnit, epc ); // must happen before VerifyCtorDtorAssign, because void return objects should not exist; before LinkReferenceToTypes because it is an indexer and needs correct types for mangling
                 acceptAll( translationUnit, lrt ); // must happen before autogen, because sized flag needs to propagate to generated functions
                 acceptAll( translationUnit, genericParams );  // check as early as possible - can't happen before LinkReferenceToTypes
+                acceptAll( translationUnit, epc ); // must happen before VerifyCtorDtorAssign, because void return objects should not exist
                 VerifyCtorDtorAssign::verify( translationUnit );  // must happen before autogen, because autogen examines existing ctor/dtors
                 ReturnChecker::checkFunctionReturns( translationUnit );

src/Tuples/TupleAssignment.cc

-              r0fe4e62
+              rf5c3b6c
 #include <memory>                          // for unique_ptr, allocator_trai...
 #include <string>                          // for string
-#include <vector>
 #include "CodeGen/OperatorTable.h"
 …
 #include "ResolvExpr/Resolver.h"           // for resolveCtorInit
 #include "ResolvExpr/TypeEnvironment.h"    // for TypeEnvironment
-#include "ResolvExpr/typeops.h"            // for combos
 #include "SynTree/Declaration.h"           // for ObjectDecl
 #include "SynTree/Expression.h"            // for Expression, CastExpr, Name...
 …
                 // dispatcher for Tuple (multiple and mass) assignment operations
                 TupleAssignSpotter( ResolvExpr::AlternativeFinder & );
                 void spot( UntypedExpr * expr, std::vector<ResolvExpr::AlternativeFinder> &args );
+                void spot( UntypedExpr * expr, const std::list<ResolvExpr::AltList> &possibilities );
           private:
 …
                 struct Matcher {
                   public:
+                        Matcher( TupleAssignSpotter &spotter, const ResolvExpr::AltList& lhs, const
+                                ResolvExpr::AltList& rhs );
+                        Matcher( TupleAssignSpotter &spotter, const ResolvExpr::AltList & alts );
                         virtual ~Matcher() {}
                         virtual void match( std::list< Expression * > &out ) = 0;
 …
                 struct MassAssignMatcher : public Matcher {
                   public:
+                        MassAssignMatcher( TupleAssignSpotter &spotter, const ResolvExpr::AltList& lhs,
+                                const ResolvExpr::AltList& rhs ) : Matcher(spotter, lhs, rhs) {}
+                        MassAssignMatcher( TupleAssignSpotter &spotter, const ResolvExpr::AltList & alts );
                         virtual void match( std::list< Expression * > &out );
                 };
 …
                 struct MultipleAssignMatcher : public Matcher {
                   public:
+                        MultipleAssignMatcher( TupleAssignSpotter &spotter, const ResolvExpr::AltList& lhs,
+                                const ResolvExpr::AltList& rhs ) : Matcher(spotter, lhs, rhs) {}
+                        MultipleAssignMatcher( TupleAssignSpotter &spot, const ResolvExpr::AltList & alts );
                         virtual void match( std::list< Expression * > &out );
                 };
 …
+        }
+        void handleTupleAssignment( ResolvExpr::AlternativeFinder & currentFinder, UntypedExpr * expr,
+                                std::vector<ResolvExpr::AlternativeFinder> &args ) {
+        void handleTupleAssignment( ResolvExpr::AlternativeFinder & currentFinder, UntypedExpr * expr, const std::list<ResolvExpr::AltList> &possibilities ) {
                 TupleAssignSpotter spotter( currentFinder );
                 spotter.spot( expr, args );
+                spotter.spot( expr, possibilities );
+        }
 …
                 : currentFinder(f) {}
+        void TupleAssignSpotter::spot( UntypedExpr * expr,
+                        std::vector<ResolvExpr::AlternativeFinder> &args ) {
+        void TupleAssignSpotter::spot( UntypedExpr * expr, const std::list<ResolvExpr::AltList> &possibilities ) {
                 if (  NameExpr *op = dynamic_cast< NameExpr * >(expr->get_function()) ) {
                         if ( CodeGen::isCtorDtorAssign( op->get_name() ) ) {
+                                fname = op->get_name();
+                                // AlternativeFinder will naturally handle this case case, if it's legal
+                                if ( args.size() == 0 ) return;
+                                // if an assignment only takes 1 argument, that's odd, but maybe someone wrote
+                                // the function, in which case AlternativeFinder will handle it normally
+                                if ( args.size() == 1 && CodeGen::isAssignment( fname ) ) return;
+                                // look over all possible left-hand-sides
+                                for ( ResolvExpr::Alternative& lhsAlt : args[0] ) {
+                                        // skip non-tuple LHS
+                                        if ( ! refToTuple(lhsAlt.expr) ) continue;
+                                        // explode is aware of casts - ensure every LHS expression is sent into explode
+                                        // with a reference cast
+                                        // xxx - this seems to change the alternatives before the normal
+                                        //  AlternativeFinder flow; maybe this is desired?
+                                        if ( ! dynamic_cast<CastExpr*>( lhsAlt.expr ) ) {
+                                                lhsAlt.expr = new CastExpr( lhsAlt.expr,
+                                                                new ReferenceType( Type::Qualifiers(),
+                                                                        lhsAlt.expr->get_result()->clone() ) );
+                               fname = op->get_name();
+                                PRINT( std::cerr << "TupleAssignment: " << fname << std::endl; )
+                                for ( std::list<ResolvExpr::AltList>::const_iterator ali = possibilities.begin(); ali != possibilities.end(); ++ali ) {
+                                        if ( ali->size() == 0 ) continue; // AlternativeFinder will natrually handle this case, if it's legal
+                                        if ( ali->size() <= 1 && CodeGen::isAssignment( op->get_name() ) ) {
+                                                // what does it mean if an assignment takes 1 argument? maybe someone defined such a function, in which case AlternativeFinder will naturally handle it
+                                                continue;
+                                        }
+                                        // explode the LHS so that each field of a tuple-valued-expr is assigned
+                                        ResolvExpr::AltList lhs;
+                                        explode( lhsAlt, currentFinder.get_indexer(), back_inserter(lhs), true );
+                                        for ( ResolvExpr::Alternative& alt : lhs ) {
+                                                // each LHS value must be a reference - some come in with a cast expression,
+                                                // if not just cast to reference here
+                                                if ( ! dynamic_cast<ReferenceType*>( alt.expr->get_result() ) ) {
+                                                        alt.expr = new CastExpr( alt.expr,
+                                                                new ReferenceType( Type::Qualifiers(),
+                                                                        alt.expr->get_result()->clone() ) );
+                                        assert( ! ali->empty() );
+                                        // grab args 2-N and group into a TupleExpr
+                                        const ResolvExpr::Alternative & alt1 = ali->front();
+                                        auto begin = std::next(ali->begin(), 1), end = ali->end();
+                                        PRINT( std::cerr << "alt1 is " << alt1.expr << std::endl; )
+                                        if ( refToTuple(alt1.expr) ) {
+                                                PRINT( std::cerr << "and is reference to tuple" << std::endl; )
+                                                if ( isMultAssign( begin, end ) ) {
+                                                        PRINT( std::cerr << "possible multiple assignment" << std::endl; )
+                                                        matcher.reset( new MultipleAssignMatcher( *this, *ali ) );
+                                                } else {
+                                                        // mass assignment
+                                                        PRINT( std::cerr << "possible mass assignment" << std::endl; )
+                                                        matcher.reset( new MassAssignMatcher( *this,  *ali ) );
+                                                }
+                                        }
-                                        if ( args.size() == 1 ) {
-                                                // mass default-initialization/destruction
-                                                ResolvExpr::AltList rhs{};
-                                                matcher.reset( new MassAssignMatcher( *this, lhs, rhs ) );
                                                 match();
-                                        } else if ( args.size() > 2 ) {
-                                                // expand all possible RHS possibilities
-                                                // TODO build iterative version of this instead of using combos
-                                                std::vector< ResolvExpr::AltList > rhsAlts;
-                                                combos( std::next(args.begin(), 1), args.end(),
-                                                        std::back_inserter( rhsAlts ) );
-                                                for ( const ResolvExpr::AltList& rhsAlt : rhsAlts ) {
-                                                        // multiple assignment
-                                                        ResolvExpr::AltList rhs;
-                                                        explode( rhsAlt, currentFinder.get_indexer(),
-                                                                std::back_inserter(rhs), true );
-                                                        matcher.reset( new MultipleAssignMatcher( *this, lhs, rhs ) );
-                                                        match();
+                                                }
-                                        } else {
-                                                for ( const ResolvExpr::Alternative& rhsAlt : args[1] ) {
-                                                        ResolvExpr::AltList rhs;
-                                                        if ( isTuple(rhsAlt.expr) ) {
-                                                                // multiple assignment
-                                                                explode( rhsAlt, currentFinder.get_indexer(),
-                                                                        std::back_inserter(rhs), true );
-                                                                matcher.reset( new MultipleAssignMatcher( *this, lhs, rhs ) );
-                                                        } else {
-                                                                // mass assignment
-                                                                rhs.push_back( rhsAlt );
-                                                                matcher.reset( new MassAssignMatcher( *this, lhs, rhs ) );
+                                                        }
-                                                        match();
+                                                }
+                                        }
+                                }
 …
                 ResolvExpr::AltList current;
                 // now resolve new assignments
+                for ( std::list< Expression * >::iterator i = new_assigns.begin();
+                                i != new_assigns.end(); ++i ) {
+                for ( std::list< Expression * >::iterator i = new_assigns.begin(); i != new_assigns.end(); ++i ) {
                         PRINT(
                                 std::cerr << "== resolving tuple assign ==" << std::endl;
 …
+                        )
+                        ResolvExpr::AlternativeFinder finder{ currentFinder.get_indexer(),
+                                currentFinder.get_environ() };
+                        ResolvExpr::AlternativeFinder finder( currentFinder.get_indexer(), currentFinder.get_environ() );
                         try {
                                 finder.findWithAdjustment(*i);
 …
                 // combine assignment environments into combined expression environment
                 simpleCombineEnvironments( current.begin(), current.end(), matcher->compositeEnv );
+                currentFinder.get_alternatives().push_front( ResolvExpr::Alternative(
+                        new TupleAssignExpr(solved_assigns, matcher->tmpDecls), matcher->compositeEnv,
+                        ResolvExpr::sumCost( current ) + matcher->baseCost ) );
+        }
+        TupleAssignSpotter::Matcher::Matcher( TupleAssignSpotter &spotter,
+                const ResolvExpr::AltList &lhs, const ResolvExpr::AltList &rhs )
+        : lhs(lhs), rhs(rhs), spotter(spotter),
+          baseCost( ResolvExpr::sumCost( lhs ) + ResolvExpr::sumCost( rhs ) ) {
+                simpleCombineEnvironments( lhs.begin(), lhs.end(), compositeEnv );
+                simpleCombineEnvironments( rhs.begin(), rhs.end(), compositeEnv );
+                currentFinder.get_alternatives().push_front( ResolvExpr::Alternative(new TupleAssignExpr(solved_assigns, matcher->tmpDecls), matcher->compositeEnv, ResolvExpr::sumCost( current  ) + matcher->baseCost ) );
+        }
+        TupleAssignSpotter::Matcher::Matcher( TupleAssignSpotter &spotter, const ResolvExpr::AltList &alts ) : spotter(spotter), baseCost( ResolvExpr::sumCost( alts ) ) {
+                assert( ! alts.empty() );
+                // combine argument environments into combined expression environment
+                simpleCombineEnvironments( alts.begin(), alts.end(), compositeEnv );
+                ResolvExpr::Alternative lhsAlt = alts.front();
+                // explode is aware of casts - ensure every LHS expression is sent into explode with a reference cast
+                if ( ! dynamic_cast< CastExpr * >( lhsAlt.expr ) ) {
+                        lhsAlt.expr = new CastExpr( lhsAlt.expr, new ReferenceType( Type::Qualifiers(), lhsAlt.expr->get_result()->clone() ) );
+                }
+                // explode the lhs so that each field of the tuple-valued-expr is assigned.
+                explode( lhsAlt, spotter.currentFinder.get_indexer(), back_inserter(lhs), true );
+                for ( ResolvExpr::Alternative & alt : lhs ) {
+                        // every LHS value must be a reference - some come in with a cast expression, if it doesn't just cast to reference here.
+                        if ( ! dynamic_cast< ReferenceType * >( alt.expr->get_result() ) ) {
+                                alt.expr = new CastExpr( alt.expr, new ReferenceType( Type::Qualifiers(), alt.expr->get_result()->clone() ) );
+                        }
+                }
+        }
+        TupleAssignSpotter::MassAssignMatcher::MassAssignMatcher( TupleAssignSpotter &spotter, const ResolvExpr::AltList & alts ) : Matcher( spotter, alts ) {
+                assert( alts.size() == 1 || alts.size() == 2 );
+                if ( alts.size() == 2 ) {
+                        rhs.push_back( alts.back() );
+                }
+        }
+        TupleAssignSpotter::MultipleAssignMatcher::MultipleAssignMatcher( TupleAssignSpotter &spotter, const ResolvExpr::AltList & alts ) : Matcher( spotter, alts ) {
+                // explode the rhs so that each field of the tuple-valued-expr is assigned.
+                explode( std::next(alts.begin(), 1), alts.end(), spotter.currentFinder.get_indexer(), back_inserter(rhs), true );
+        }

src/Tuples/Tuples.h

-              r0fe4e62
+              rf5c3b6c
 #include <string>
-#include <vector>
 #include "SynTree/Expression.h"
 …
 namespace Tuples {
         // TupleAssignment.cc
+        void handleTupleAssignment( ResolvExpr::AlternativeFinder & currentFinder, UntypedExpr * assign,
+                std::vector< ResolvExpr::AlternativeFinder >& args );
+        void handleTupleAssignment( ResolvExpr::AlternativeFinder & currentFinder, UntypedExpr * assign, const std::list<ResolvExpr::AltList> & possibilities );
         // TupleExpansion.cc
         /// expands z.[a, b.[x, y], c] into [z.a, z.b.x, z.b.y, z.c], inserting UniqueExprs as appropriate

src/benchmark/Makefile.am

-              r0fe4e62
+              rf5c3b6c
 STATS    = ${TOOLSDIR}stat.py
 repeats  = 30
-TIME_FORMAT = "%E"
-PRINT_FORMAT = '%20s\t'
 .NOTPARALLEL:
 …
 noinst_PROGRAMS =
+all : ctxswitch$(EXEEXT) mutex$(EXEEXT) signal$(EXEEXT) waitfor$(EXEEXT) creation$(EXEEXT)
+bench$(EXEEXT) :
+        @for ccflags in "-debug" "-nodebug"; do \
+                echo ${CC} ${AM_CFLAGS} ${CFLAGS} ${ccflags} @CFA_FLAGS@ -lrt bench.c;\
+                ${CC} ${AM_CFLAGS} ${CFLAGS} $${ccflags} -lrt bench.c;\
+                ./a.out ; \
+        done ; \
+        rm -f ./a.out ;
+%.run : %$(EXEEXT) ${REPEAT}
+        @rm -f .result.log
+        @echo "------------------------------------------------------"
+        @echo $<
+        @${REPEAT} ${repeats} ./a.out | tee -a .result.log
+        @${STATS} .result.log
+        @echo "------------------------------------------------------"
+        @rm -f a.out .result.log
+%.runquiet :
+        @+make $(basename $@)
+csv-data$(EXEEXT):
+        @${CC} ${AM_CFLAGS} ${CFLAGS} ${ccflags} @CFA_FLAGS@ -nodebug -lrt -quiet -DN=50000000 csv-data.c
         @./a.out
+        @rm -f a.out
+%.make :
+        @printf "${PRINT_FORMAT}" $(basename $(subst compile-,,$@))
+        @+/usr/bin/time -f ${TIME_FORMAT} make $(basename $@) 2>&1
+${REPEAT} :
+        @+make -C ${TOOLSDIR} repeat
+## =========================================================================================================
+jenkins$(EXEEXT):
+        @echo "{"
+        @echo -e '\t"githash": "'${githash}'",'
+        @echo -e '\t"arch": "'   ${arch}   '",'
+        @echo -e '\t"compile": {'
+        @+make compile TIME_FORMAT='%e,' PRINT_FORMAT='\t\t\"%s\" :'
+        @echo -e '\t\t"dummy" : {}'
+        @echo -e '\t},'
+        @echo -e '\t"ctxswitch": {'
+        @echo -en '\t\t"coroutine":'
+        @+make ctxswitch-cfa_coroutine.runquiet
+        @echo -en '\t\t,"thread":'
+        @+make ctxswitch-cfa_thread.runquiet
+        @echo -e '\t},'
+        @echo -e '\t"mutex": ['
+        @echo -en '\t\t'
+        @+make mutex-cfa1.runquiet
+        @echo -en '\t\t,'
+        @+make mutex-cfa2.runquiet
+        @echo -e '\t],'
+        @echo -e '\t"scheduling": ['
+        @echo -en '\t\t'
+        @+make signal-cfa1.runquiet
+        @echo -en '\t\t,'
+        @+make signal-cfa2.runquiet
+        @echo -en '\t\t,'
+        @+make waitfor-cfa1.runquiet
+        @echo -en '\t\t,'
+        @+make waitfor-cfa2.runquiet
+        @echo -e '\n\t],'
+        @echo -e '\t"epoch": ' $(shell date +%s)
+        @echo "}"
+        @rm -f ./a.out
 ## =========================================================================================================
 …
 ctxswitch-cfa_coroutine$(EXEEXT):
         @${CC}        ctxswitch/cfa_cor.c   -DBENCH_N=50000000  -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+        ${CC}        ctxswitch/cfa_cor.c   -DBENCH_N=50000000  -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
 ctxswitch-cfa_thread$(EXEEXT):
         @${CC}        ctxswitch/cfa_thrd.c  -DBENCH_N=50000000  -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+        ${CC}        ctxswitch/cfa_thrd.c  -DBENCH_N=50000000  -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
 ctxswitch-upp_coroutine$(EXEEXT):
         @u++          ctxswitch/upp_cor.cc  -DBENCH_N=50000000  -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+        u++          ctxswitch/upp_cor.cc  -DBENCH_N=50000000  -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
 ctxswitch-upp_thread$(EXEEXT):
         @u++          ctxswitch/upp_thrd.cc -DBENCH_N=50000000  -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+        u++          ctxswitch/upp_thrd.cc -DBENCH_N=50000000  -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
 ctxswitch-pthread$(EXEEXT):
+        @@BACKEND_CC@ ctxswitch/pthreads.c  -DBENCH_N=50000000  -I. -lrt -pthread                    ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+        @BACKEND_CC@ ctxswitch/pthreads.c  -DBENCH_N=50000000  -I. -lrt -pthread                    ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+## =========================================================================================================
+creation$(EXEEXT) :\
+        creation-pthread.run            \
+        creation-cfa_coroutine.run      \
+        creation-cfa_thread.run         \
+        creation-upp_coroutine.run      \
+        creation-upp_thread.run
+creation-cfa_coroutine$(EXEEXT):
+        ${CC}        creation/cfa_cor.c   -DBENCH_N=10000000   -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+creation-cfa_thread$(EXEEXT):
+        ${CC}        creation/cfa_thrd.c  -DBENCH_N=10000000   -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+creation-upp_coroutine$(EXEEXT):
+        u++          creation/upp_cor.cc  -DBENCH_N=50000000   -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+creation-upp_thread$(EXEEXT):
+        u++          creation/upp_thrd.cc -DBENCH_N=50000000   -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+creation-pthread$(EXEEXT):
+        @BACKEND_CC@ creation/pthreads.c  -DBENCH_N=250000     -I. -lrt -pthread                    ${AM_CFLAGS} ${CFLAGS} ${ccflags}
 ## =========================================================================================================
 …
 mutex-function$(EXEEXT):
         @@BACKEND_CC@ mutex/function.c    -DBENCH_N=500000000   -I. -lrt -pthread                    ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+        @BACKEND_CC@ mutex/function.c    -DBENCH_N=500000000   -I. -lrt -pthread                    ${AM_CFLAGS} ${CFLAGS} ${ccflags}
 mutex-pthread_lock$(EXEEXT):
         @@BACKEND_CC@ mutex/pthreads.c    -DBENCH_N=50000000    -I. -lrt -pthread                    ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+        @BACKEND_CC@ mutex/pthreads.c    -DBENCH_N=50000000    -I. -lrt -pthread                    ${AM_CFLAGS} ${CFLAGS} ${ccflags}
 mutex-upp$(EXEEXT):
         @u++          mutex/upp.cc        -DBENCH_N=50000000    -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+        u++          mutex/upp.cc        -DBENCH_N=50000000    -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
 mutex-cfa1$(EXEEXT):
         @${CC}        mutex/cfa1.c        -DBENCH_N=5000000     -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+        ${CC}        mutex/cfa1.c        -DBENCH_N=5000000     -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
 mutex-cfa2$(EXEEXT):
         @${CC}        mutex/cfa2.c        -DBENCH_N=5000000     -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+        ${CC}        mutex/cfa2.c        -DBENCH_N=5000000     -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
 mutex-cfa4$(EXEEXT):
         @${CC}        mutex/cfa4.c        -DBENCH_N=5000000     -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+        ${CC}        mutex/cfa4.c        -DBENCH_N=5000000     -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
 ## =========================================================================================================
 …
 signal-upp$(EXEEXT):
         @u++          schedint/upp.cc     -DBENCH_N=5000000     -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+        u++          schedint/upp.cc     -DBENCH_N=5000000     -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
 signal-cfa1$(EXEEXT):
         @${CC}        schedint/cfa1.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+        ${CC}        schedint/cfa1.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
 signal-cfa2$(EXEEXT):
         @${CC}        schedint/cfa2.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+        ${CC}        schedint/cfa2.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
 signal-cfa4$(EXEEXT):
         @${CC}        schedint/cfa4.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+        ${CC}        schedint/cfa4.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
 ## =========================================================================================================
 …
 waitfor-upp$(EXEEXT):
         @u++          schedext/upp.cc     -DBENCH_N=5000000     -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+        u++          schedext/upp.cc     -DBENCH_N=5000000     -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
 waitfor-cfa1$(EXEEXT):
         @${CC}        schedext/cfa1.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+        ${CC}        schedext/cfa1.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
 waitfor-cfa2$(EXEEXT):
         @${CC}        schedext/cfa2.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+        ${CC}        schedext/cfa2.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
 waitfor-cfa4$(EXEEXT):
+        @${CC}        schedext/cfa4.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+## =========================================================================================================
+creation$(EXEEXT) :\
+        creation-pthread.run                    \
+        creation-cfa_coroutine.run              \
+        creation-cfa_coroutine_eager.run        \
+        creation-cfa_thread.run                 \
+        creation-upp_coroutine.run              \
+        creation-upp_thread.run
+creation-cfa_coroutine$(EXEEXT):
+        @${CC}        creation/cfa_cor.c   -DBENCH_N=10000000   -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+creation-cfa_coroutine_eager$(EXEEXT):
+        @${CC}        creation/cfa_cor.c   -DBENCH_N=10000000   -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags} -DEAGER
+creation-cfa_thread$(EXEEXT):
+        @${CC}        creation/cfa_thrd.c  -DBENCH_N=10000000   -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+creation-upp_coroutine$(EXEEXT):
+        @u++          creation/upp_cor.cc  -DBENCH_N=50000000   -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+creation-upp_thread$(EXEEXT):
+        @u++          creation/upp_thrd.cc -DBENCH_N=50000000   -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+creation-pthread$(EXEEXT):
+        @@BACKEND_CC@ creation/pthreads.c  -DBENCH_N=250000     -I. -lrt -pthread                    ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+        ${CC}        schedext/cfa4.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
 ## =========================================================================================================
+compile$(EXEEXT) :\
+        compile-array.make      \
+        compile-attributes.make \
+        compile-empty.make      \
+        compile-expression.make \
+        compile-io.make         \
+        compile-monitor.make    \
+        compile-operators.make  \
+        compile-typeof.make
+%.run : %$(EXEEXT) ${REPEAT}
+        @rm -f .result.log
+        @echo "------------------------------------------------------"
+        @echo $<
+        @${REPEAT} ${repeats} ./a.out | tee -a .result.log
+        @${STATS} .result.log
+        @echo "------------------------------------------------------"
+        @rm -f a.out .result.log
+compile-array$(EXEEXT):
+        @${CC} -nodebug -quiet -fsyntax-only -w ../tests/array.c
+compile-attributes$(EXEEXT):
+        @${CC} -nodebug -quiet -fsyntax-only -w ../tests/attributes.c
+compile-empty$(EXEEXT):
+        @${CC} -nodebug -quiet -fsyntax-only -w compile/empty.c
+compile-expression$(EXEEXT):
+        @${CC} -nodebug -quiet -fsyntax-only -w ../tests/expression.c
+compile-io$(EXEEXT):
+        @${CC} -nodebug -quiet -fsyntax-only -w ../tests/io.c
+compile-monitor$(EXEEXT):
+        @${CC} -nodebug -quiet -fsyntax-only -w ../tests/monitor.c
+compile-operators$(EXEEXT):
+        @${CC} -nodebug -quiet -fsyntax-only -w ../tests/operators.c
+compile-thread$(EXEEXT):
+        @${CC} -nodebug -quiet -fsyntax-only -w ../tests/thread.c
+compile-typeof$(EXEEXT):
+        @${CC} -nodebug -quiet -fsyntax-only -w ../tests/typeof.c
+${REPEAT} :
+        @+make -C ${TOOLSDIR} repeat

src/benchmark/Makefile.in

-              r0fe4e62
+              rf5c3b6c
   esac
 am__tagged_files = $(HEADERS) $(SOURCES) $(TAGS_FILES) $(LISP)
 am__DIST_COMMON = $(srcdir)/Makefile.in compile
+am__DIST_COMMON = $(srcdir)/Makefile.in
 DISTFILES = $(DIST_COMMON) $(DIST_SOURCES) $(TEXINFOS) $(EXTRA_DIST)
 ACLOCAL = @ACLOCAL@
 …
 STATS = ${TOOLSDIR}stat.py
 repeats = 30
-TIME_FORMAT = "%E"
-PRINT_FORMAT = '%20s\t'
 all: all-am
 …
 .NOTPARALLEL:
+all : ctxswitch$(EXEEXT) mutex$(EXEEXT) signal$(EXEEXT) waitfor$(EXEEXT) creation$(EXEEXT)
+bench$(EXEEXT) :
+        @for ccflags in "-debug" "-nodebug"; do \
+                echo ${CC} ${AM_CFLAGS} ${CFLAGS} ${ccflags} @CFA_FLAGS@ -lrt bench.c;\
+                ${CC} ${AM_CFLAGS} ${CFLAGS} $${ccflags} -lrt bench.c;\
+                ./a.out ; \
+        done ; \
+        rm -f ./a.out ;
+csv-data$(EXEEXT):
+        @${CC} ${AM_CFLAGS} ${CFLAGS} ${ccflags} @CFA_FLAGS@ -nodebug -lrt -quiet -DN=50000000 csv-data.c
+        @./a.out
+        @rm -f ./a.out
+ctxswitch$(EXEEXT): \
+        ctxswitch-pthread.run           \
+        ctxswitch-cfa_coroutine.run     \
+        ctxswitch-cfa_thread.run        \
+        ctxswitch-upp_coroutine.run     \
+        ctxswitch-upp_thread.run
+ctxswitch-cfa_coroutine$(EXEEXT):
+        ${CC}        ctxswitch/cfa_cor.c   -DBENCH_N=50000000  -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+ctxswitch-cfa_thread$(EXEEXT):
+        ${CC}        ctxswitch/cfa_thrd.c  -DBENCH_N=50000000  -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+ctxswitch-upp_coroutine$(EXEEXT):
+        u++          ctxswitch/upp_cor.cc  -DBENCH_N=50000000  -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+ctxswitch-upp_thread$(EXEEXT):
+        u++          ctxswitch/upp_thrd.cc -DBENCH_N=50000000  -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+ctxswitch-pthread$(EXEEXT):
+        @BACKEND_CC@ ctxswitch/pthreads.c  -DBENCH_N=50000000  -I. -lrt -pthread                    ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+creation$(EXEEXT) :\
+        creation-pthread.run            \
+        creation-cfa_coroutine.run      \
+        creation-cfa_thread.run         \
+        creation-upp_coroutine.run      \
+        creation-upp_thread.run
+creation-cfa_coroutine$(EXEEXT):
+        ${CC}        creation/cfa_cor.c   -DBENCH_N=10000000   -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+creation-cfa_thread$(EXEEXT):
+        ${CC}        creation/cfa_thrd.c  -DBENCH_N=10000000   -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+creation-upp_coroutine$(EXEEXT):
+        u++          creation/upp_cor.cc  -DBENCH_N=50000000   -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+creation-upp_thread$(EXEEXT):
+        u++          creation/upp_thrd.cc -DBENCH_N=50000000   -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+creation-pthread$(EXEEXT):
+        @BACKEND_CC@ creation/pthreads.c  -DBENCH_N=250000     -I. -lrt -pthread                    ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+mutex$(EXEEXT) :\
+        mutex-function.run      \
+        mutex-pthread_lock.run  \
+        mutex-upp.run           \
+        mutex-cfa1.run          \
+        mutex-cfa2.run          \
+        mutex-cfa4.run
+mutex-function$(EXEEXT):
+        @BACKEND_CC@ mutex/function.c    -DBENCH_N=500000000   -I. -lrt -pthread                    ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+mutex-pthread_lock$(EXEEXT):
+        @BACKEND_CC@ mutex/pthreads.c    -DBENCH_N=50000000    -I. -lrt -pthread                    ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+mutex-upp$(EXEEXT):
+        u++          mutex/upp.cc        -DBENCH_N=50000000    -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+mutex-cfa1$(EXEEXT):
+        ${CC}        mutex/cfa1.c        -DBENCH_N=5000000     -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+mutex-cfa2$(EXEEXT):
+        ${CC}        mutex/cfa2.c        -DBENCH_N=5000000     -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+mutex-cfa4$(EXEEXT):
+        ${CC}        mutex/cfa4.c        -DBENCH_N=5000000     -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+signal$(EXEEXT) :\
+        signal-upp.run          \
+        signal-cfa1.run         \
+        signal-cfa2.run         \
+        signal-cfa4.run
+signal-upp$(EXEEXT):
+        u++          schedint/upp.cc     -DBENCH_N=5000000     -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+signal-cfa1$(EXEEXT):
+        ${CC}        schedint/cfa1.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+signal-cfa2$(EXEEXT):
+        ${CC}        schedint/cfa2.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+signal-cfa4$(EXEEXT):
+        ${CC}        schedint/cfa4.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+waitfor$(EXEEXT) :\
+        waitfor-upp.run         \
+        waitfor-cfa1.run                \
+        waitfor-cfa2.run                \
+        waitfor-cfa4.run
+waitfor-upp$(EXEEXT):
+        u++          schedext/upp.cc     -DBENCH_N=5000000     -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+waitfor-cfa1$(EXEEXT):
+        ${CC}        schedext/cfa1.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+waitfor-cfa2$(EXEEXT):
+        ${CC}        schedext/cfa2.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
+waitfor-cfa4$(EXEEXT):
+        ${CC}        schedext/cfa4.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
 %.run : %$(EXEEXT) ${REPEAT}
 …
         @rm -f a.out .result.log
-%.runquiet :
-        @+make $(basename $@)
-        @./a.out
-        @rm -f a.out
-%.make :
-        @printf "${PRINT_FORMAT}" $(basename $(subst compile-,,$@))
-        @+/usr/bin/time -f ${TIME_FORMAT} make $(basename $@) 2>&1
 ${REPEAT} :
         @+make -C ${TOOLSDIR} repeat
-jenkins$(EXEEXT):
-        @echo "{"
-        @echo -e '\t"githash": "'${githash}'",'
-        @echo -e '\t"arch": "'   ${arch}   '",'
-        @echo -e '\t"compile": {'
-        @+make compile TIME_FORMAT='%e,' PRINT_FORMAT='\t\t\"%s\" :'
-        @echo -e '\t\t"dummy" : {}'
-        @echo -e '\t},'
-        @echo -e '\t"ctxswitch": {'
-        @echo -en '\t\t"coroutine":'
-        @+make ctxswitch-cfa_coroutine.runquiet
-        @echo -en '\t\t,"thread":'
-        @+make ctxswitch-cfa_thread.runquiet
-        @echo -e '\t},'
-        @echo -e '\t"mutex": ['
-        @echo -en '\t\t'
-        @+make mutex-cfa1.runquiet
-        @echo -en '\t\t,'
-        @+make mutex-cfa2.runquiet
-        @echo -e '\t],'
-        @echo -e '\t"scheduling": ['
-        @echo -en '\t\t'
-        @+make signal-cfa1.runquiet
-        @echo -en '\t\t,'
-        @+make signal-cfa2.runquiet
-        @echo -en '\t\t,'
-        @+make waitfor-cfa1.runquiet
-        @echo -en '\t\t,'
-        @+make waitfor-cfa2.runquiet
-        @echo -e '\n\t],'
-        @echo -e '\t"epoch": ' $(shell date +%s)
-        @echo "}"
-ctxswitch$(EXEEXT): \
-        ctxswitch-pthread.run           \
-        ctxswitch-cfa_coroutine.run     \
-        ctxswitch-cfa_thread.run        \
-        ctxswitch-upp_coroutine.run     \
-        ctxswitch-upp_thread.run
-ctxswitch-cfa_coroutine$(EXEEXT):
-        @${CC}        ctxswitch/cfa_cor.c   -DBENCH_N=50000000  -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-ctxswitch-cfa_thread$(EXEEXT):
-        @${CC}        ctxswitch/cfa_thrd.c  -DBENCH_N=50000000  -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-ctxswitch-upp_coroutine$(EXEEXT):
-        @u++          ctxswitch/upp_cor.cc  -DBENCH_N=50000000  -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-ctxswitch-upp_thread$(EXEEXT):
-        @u++          ctxswitch/upp_thrd.cc -DBENCH_N=50000000  -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-ctxswitch-pthread$(EXEEXT):
-        @@BACKEND_CC@ ctxswitch/pthreads.c  -DBENCH_N=50000000  -I. -lrt -pthread                    ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-mutex$(EXEEXT) :\
-        mutex-function.run      \
-        mutex-pthread_lock.run  \
-        mutex-upp.run           \
-        mutex-cfa1.run          \
-        mutex-cfa2.run          \
-        mutex-cfa4.run
-mutex-function$(EXEEXT):
-        @@BACKEND_CC@ mutex/function.c    -DBENCH_N=500000000   -I. -lrt -pthread                    ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-mutex-pthread_lock$(EXEEXT):
-        @@BACKEND_CC@ mutex/pthreads.c    -DBENCH_N=50000000    -I. -lrt -pthread                    ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-mutex-upp$(EXEEXT):
-        @u++          mutex/upp.cc        -DBENCH_N=50000000    -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-mutex-cfa1$(EXEEXT):
-        @${CC}        mutex/cfa1.c        -DBENCH_N=5000000     -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-mutex-cfa2$(EXEEXT):
-        @${CC}        mutex/cfa2.c        -DBENCH_N=5000000     -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-mutex-cfa4$(EXEEXT):
-        @${CC}        mutex/cfa4.c        -DBENCH_N=5000000     -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-signal$(EXEEXT) :\
-        signal-upp.run          \
-        signal-cfa1.run         \
-        signal-cfa2.run         \
-        signal-cfa4.run
-signal-upp$(EXEEXT):
-        @u++          schedint/upp.cc     -DBENCH_N=5000000     -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-signal-cfa1$(EXEEXT):
-        @${CC}        schedint/cfa1.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-signal-cfa2$(EXEEXT):
-        @${CC}        schedint/cfa2.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-signal-cfa4$(EXEEXT):
-        @${CC}        schedint/cfa4.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-waitfor$(EXEEXT) :\
-        waitfor-upp.run         \
-        waitfor-cfa1.run                \
-        waitfor-cfa2.run                \
-        waitfor-cfa4.run
-waitfor-upp$(EXEEXT):
-        @u++          schedext/upp.cc     -DBENCH_N=5000000     -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-waitfor-cfa1$(EXEEXT):
-        @${CC}        schedext/cfa1.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-waitfor-cfa2$(EXEEXT):
-        @${CC}        schedext/cfa2.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-waitfor-cfa4$(EXEEXT):
-        @${CC}        schedext/cfa4.c     -DBENCH_N=500000      -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-creation$(EXEEXT) :\
-        creation-pthread.run                    \
-        creation-cfa_coroutine.run              \
-        creation-cfa_coroutine_eager.run        \
-        creation-cfa_thread.run                 \
-        creation-upp_coroutine.run              \
-        creation-upp_thread.run
-creation-cfa_coroutine$(EXEEXT):
-        @${CC}        creation/cfa_cor.c   -DBENCH_N=10000000   -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-creation-cfa_coroutine_eager$(EXEEXT):
-        @${CC}        creation/cfa_cor.c   -DBENCH_N=10000000   -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags} -DEAGER
-creation-cfa_thread$(EXEEXT):
-        @${CC}        creation/cfa_thrd.c  -DBENCH_N=10000000   -I. -nodebug -lrt -quiet @CFA_FLAGS@ ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-creation-upp_coroutine$(EXEEXT):
-        @u++          creation/upp_cor.cc  -DBENCH_N=50000000   -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-creation-upp_thread$(EXEEXT):
-        @u++          creation/upp_thrd.cc -DBENCH_N=50000000   -I. -nodebug -lrt -quiet             ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-creation-pthread$(EXEEXT):
-        @@BACKEND_CC@ creation/pthreads.c  -DBENCH_N=250000     -I. -lrt -pthread                    ${AM_CFLAGS} ${CFLAGS} ${ccflags}
-compile$(EXEEXT) :\
-        compile-array.make      \
-        compile-attributes.make \
-        compile-empty.make      \
-        compile-expression.make \
-        compile-io.make         \
-        compile-monitor.make    \
-        compile-operators.make  \
-        compile-typeof.make
-compile-array$(EXEEXT):
-        @${CC} -nodebug -quiet -fsyntax-only -w ../tests/array.c
-compile-attributes$(EXEEXT):
-        @${CC} -nodebug -quiet -fsyntax-only -w ../tests/attributes.c
-compile-empty$(EXEEXT):
-        @${CC} -nodebug -quiet -fsyntax-only -w compile/empty.c
-compile-expression$(EXEEXT):
-        @${CC} -nodebug -quiet -fsyntax-only -w ../tests/expression.c
-compile-io$(EXEEXT):
-        @${CC} -nodebug -quiet -fsyntax-only -w ../tests/io.c
-compile-monitor$(EXEEXT):
-        @${CC} -nodebug -quiet -fsyntax-only -w ../tests/monitor.c
-compile-operators$(EXEEXT):
-        @${CC} -nodebug -quiet -fsyntax-only -w ../tests/operators.c
-compile-thread$(EXEEXT):
-        @${CC} -nodebug -quiet -fsyntax-only -w ../tests/thread.c
-compile-typeof$(EXEEXT):
-        @${CC} -nodebug -quiet -fsyntax-only -w ../tests/typeof.c
 # Tell versions [3.59,3.63) of GNU make to not export all variables.

src/benchmark/creation/cfa_cor.c

-              r0fe4e62
+              rf5c3b6c
 coroutine MyCoroutine {};
+void ?{} (MyCoroutine & this) {
+#ifdef EAGER
+        prime(this);
+#endif
+}
+void ?{} (MyCoroutine & this) { prime(this); }
 void main(MyCoroutine & this) {}

src/benchmark/csv-data.c

-              r0fe4e62
+              rf5c3b6c
 // coroutine context switch
 long long int measure_coroutine() {
         const unsigned int NoOfTimes = 50000000;
+        const unsigned int NoOfTimes = N;
         long long int StartTime, EndTime;
 …
 // thread context switch
 long long int measure_thread() {
         const unsigned int NoOfTimes = 50000000;
+        const unsigned int NoOfTimes = N;
         long long int StartTime, EndTime;
 …
 long long int measure_1_monitor_entry() {
         const unsigned int NoOfTimes = 5000000;
+        const unsigned int NoOfTimes = N;
         long long int StartTime, EndTime;
         mon_t mon;
 …
 long long int measure_2_monitor_entry() {
         const unsigned int NoOfTimes = 5000000;
+        const unsigned int NoOfTimes = N;
         long long int StartTime, EndTime;
         mon_t mon1, mon2;
 …
 //-----------------------------------------------------------------------------
 // single internal sched entry
-const unsigned int NoOfTimes = 500000;
 mon_t mon1;
 …
 void side1A( mon_t & mutex a, long long int * out ) {
+        const unsigned int NoOfTimes = 500000;
+        long long int StartTime, EndTime;
+        StartTime = Time();
+        for( int i = 0;; i++ ) {
+                signal(cond1a);
+                if( i > NoOfTimes ) break;
+                wait(cond1b);
+        }
+        EndTime = Time();
+        *out = ( EndTime - StartTime ) / NoOfTimes;
+        long long int StartTime, EndTime;
+        StartTime = Time();
+        for( int i = 0;; i++ ) {
+                signal(&cond1a);
+                if( i > N ) break;
+                wait(&cond1b);
+        }
+        EndTime = Time();
+        *out = ( EndTime - StartTime ) / N;
+}
 void side1B( mon_t & mutex a ) {
         for( int i = 0;; i++ ) {
                 signal(cond1b);
                 if( i > N ) break;
                 wait(cond1a);
+                signal(&cond1b);
+                if( i > N ) break;
+                wait(&cond1a);
+        }
+}
 …
 //-----------------------------------------------------------------------------
 // multi internal sched
+// multi internal sched entry
 mon_t mon2;
 …
 void side2A( mon_t & mutex a, mon_t & mutex b, long long int * out ) {
+        const unsigned int NoOfTimes = 500000;
+        long long int StartTime, EndTime;
+        StartTime = Time();
+        for( int i = 0;; i++ ) {
+                signal(cond2a);
+                if( i > NoOfTimes ) break;
+                wait(cond2b);
+        }
+        EndTime = Time();
+        *out = ( EndTime - StartTime ) / NoOfTimes;
+        long long int StartTime, EndTime;
+        StartTime = Time();
+        for( int i = 0;; i++ ) {
+                signal(&cond2a);
+                if( i > N ) break;
+                wait(&cond2b);
+        }
+        EndTime = Time();
+        *out = ( EndTime - StartTime ) / N;
+}
 void side2B( mon_t & mutex a, mon_t & mutex b ) {
         for( int i = 0;; i++ ) {
                 signal(cond2b);
                 if( i > N ) break;
                 wait(cond2a);
+                signal(&cond2b);
+                if( i > N ) break;
+                wait(&cond2a);
+        }
+}
 …
 //-----------------------------------------------------------------------------
-// single external sched
-volatile int go = 0;
-void __attribute__((noinline)) call( mon_t & mutex m1 ) {}
-long long int  __attribute__((noinline)) wait( mon_t & mutex m1 ) {
-        go = 1;
-        const unsigned int NoOfTimes = 5000000;
-        long long int StartTime, EndTime;
-        StartTime = Time();
-        for (size_t i = 0; i < NoOfTimes; i++) {
-                waitfor(call, m1);
+        }
-        EndTime = Time();
-        go = 0;
-        return ( EndTime - StartTime ) / NoOfTimes;
+}
-thread thrd3 {};
-void ^?{}( thrd3 & mutex this ) {}
-void main( thrd3 & this ) {
-        while(go == 0) { yield(); }
-        while(go == 1) { call(mon1); }
+}
-long long int measure_1_sched_ext() {
-        go = 0;
-        thrd3 t;
-        return wait(mon1);
+}
-//-----------------------------------------------------------------------------
-// multi external sched
-void __attribute__((noinline)) call( mon_t & mutex m1, mon_t & mutex m2 ) {}
-long long int  __attribute__((noinline)) wait( mon_t & mutex m1, mon_t & mutex m2 ) {
-        go = 1;
-        const unsigned int NoOfTimes = 5000000;
-        long long int StartTime, EndTime;
-        StartTime = Time();
-        for (size_t i = 0; i < NoOfTimes; i++) {
-                waitfor(call, m1, m2);
+        }
-        EndTime = Time();
-        go = 0;
-        return ( EndTime - StartTime ) / NoOfTimes;
+}
-thread thrd4 {};
-void ^?{}( thrd4 & mutex this ) {}
-void main( thrd4 & this ) {
-        while(go == 0) { yield(); }
-        while(go == 1) { call(mon1, mon2); }
+}
-long long int measure_2_sched_ext() {
-        go = 0;
-        thrd3 t;
-        return wait(mon1, mon2);
+}
-//-----------------------------------------------------------------------------
 // main loop
 int main()
+{
+        sout | "\tepoch:" | time(NULL) | ',' | endl;
+        sout | "\tctxswitch: {" | endl;
+        sout | "\t\tcoroutine: "| measure_coroutine() | ',' | endl;
+        sout | "\t\tthread:" | measure_thread() | ',' | endl;
+        sout | "\t}," | endl;
+        sout | "\tmutex: ["     | measure_1_monitor_entry()     | ',' | measure_2_monitor_entry()       | "]," | endl;
+        sout | "\tscheduling: ["| measure_1_sched_int()         | ',' | measure_2_sched_int()   | ','  |
+                                          measure_1_sched_ext()         | ',' | measure_2_sched_ext()   | "]," | endl;
+}
+        sout | time(NULL) | ',';
+        sout | measure_coroutine() | ',';
+        sout | measure_thread() | ',';
+        sout | measure_1_monitor_entry() | ',';
+        sout | measure_2_monitor_entry() | ',';
+        sout | measure_1_sched_int() | ',';
+        sout | measure_2_sched_int() | endl;
+}

src/benchmark/schedint/cfa1.c

-              r0fe4e62
+              rf5c3b6c
 void __attribute__((noinline)) call( M & mutex a1 ) {
         signal(c);
+        signal(&c);
+}
 …
         BENCH(
                 for (size_t i = 0; i < n; i++) {
                         wait(c);
+                        wait(&c);
                 },
                 result

src/benchmark/schedint/cfa2.c

-              r0fe4e62
+              rf5c3b6c
 void __attribute__((noinline)) call( M & mutex a1, M & mutex a2 ) {
         signal(c);
+        signal(&c);
+}
 …
         BENCH(
                 for (size_t i = 0; i < n; i++) {
                         wait(c);
+                        wait(&c);
                 },
                 result

src/benchmark/schedint/cfa4.c

-              r0fe4e62
+              rf5c3b6c
 void __attribute__((noinline)) call( M & mutex a1, M & mutex a2, M & mutex a3, M & mutex a4 ) {
         signal(c);
+        signal(&c);
+}
 …
         BENCH(
                 for (size_t i = 0; i < n; i++) {
                         wait(c);
+                        wait(&c);
                 },
                 result

src/libcfa/Makefile.am

-              r0fe4e62
+              rf5c3b6c
 cfa_includedir = $(CFA_INCDIR)
+nobase_cfa_include_HEADERS =    \
+        ${headers}                      \
+        ${stdhdr}                       \
+        math                            \
+        gmp                             \
+        bits/defs.h             \
+        bits/locks.h            \
+        concurrency/invoke.h    \
+        libhdr.h                        \
+        libhdr/libalign.h       \
+        libhdr/libdebug.h       \
+        libhdr/libtools.h
+nobase_cfa_include_HEADERS = ${headers} ${stdhdr} math gmp concurrency/invoke.h
 CLEANFILES = libcfa-prelude.c

src/libcfa/Makefile.in

-              r0fe4e62
+              rf5c3b6c
         containers/result containers/vector concurrency/coroutine \
         concurrency/thread concurrency/kernel concurrency/monitor \
+        ${shell echo stdhdr/*} math gmp bits/defs.h bits/locks.h \
+        concurrency/invoke.h libhdr.h libhdr/libalign.h \
+        libhdr/libdebug.h libhdr/libtools.h
+        ${shell echo stdhdr/*} math gmp concurrency/invoke.h
 HEADERS = $(nobase_cfa_include_HEADERS)
 am__tagged_files = $(HEADERS) $(SOURCES) $(TAGS_FILES) $(LISP)
 …
 stdhdr = ${shell echo stdhdr/*}
 cfa_includedir = $(CFA_INCDIR)
+nobase_cfa_include_HEADERS = \
+        ${headers}                      \
+        ${stdhdr}                       \
+        math                            \
+        gmp                             \
+        bits/defs.h             \
+        bits/locks.h            \
+        concurrency/invoke.h    \
+        libhdr.h                        \
+        libhdr/libalign.h       \
+        libhdr/libdebug.h       \
+        libhdr/libtools.h
+nobase_cfa_include_HEADERS = ${headers} ${stdhdr} math gmp concurrency/invoke.h
 CLEANFILES = libcfa-prelude.c
 all: all-am

src/libcfa/concurrency/alarm.c

-              r0fe4e62
+              rf5c3b6c
         disable_interrupts();
         lock( event_kernel->lock DEBUG_CTX2 );
+        lock( &event_kernel->lock DEBUG_CTX2 );
+        {
                 verify( validate( alarms ) );
 …
+                }
+        }
         unlock( event_kernel->lock );
+        unlock( &event_kernel->lock );
         this->set = true;
         enable_interrupts( DEBUG_CTX );
 …
 void unregister_self( alarm_node_t * this ) {
         disable_interrupts();
         lock( event_kernel->lock DEBUG_CTX2 );
+        lock( &event_kernel->lock DEBUG_CTX2 );
+        {
                 verify( validate( &event_kernel->alarms ) );
                 remove( &event_kernel->alarms, this );
+        }
         unlock( event_kernel->lock );
+        unlock( &event_kernel->lock );
         enable_interrupts( DEBUG_CTX );
         this->set = false;

src/libcfa/concurrency/coroutine.c

r0fe4e62	rf5c3b6c
156	156	this->limit = (char *)libCeiling( (unsigned long)this->storage, 16 ); // minimum alignment
157	157	} // if
158		assertf( this->size >= MinStackSize, "Stack size %zd provides less than minimum of %d bytes for a stack.", this->size, MinStackSize );
	158	assertf( this->size >= MinStackSize, "Stack size %d provides less than minimum of %d bytes for a stack.", this->size, MinStackSize );
159	159
160	160	this->base = (char *)this->limit + this->size;

src/libcfa/concurrency/invoke.h

-              r0fe4e62
+              rf5c3b6c
 //
 #include "bits/defs.h"
 #include "bits/locks.h"
+#include <stdbool.h>
+#include <stdint.h>
 #ifdef __CFORALL__
 …
 #define _INVOKE_H_
+        typedef void (*fptr_t)();
+        typedef int_fast16_t __lock_size_t;
+      #define unlikely(x)    __builtin_expect(!!(x), 0)
+      #define thread_local _Thread_local
+        struct __thread_queue_t {
+                struct thread_desc * head;
+                struct thread_desc ** tail;
+        };
+      typedef void (*fptr_t)();
+        struct __condition_stack_t {
+                struct __condition_criterion_t * top;
+        };
+      struct spinlock {
+            volatile int lock;
+            #ifdef __CFA_DEBUG__
+                  const char * prev_name;
+                  void* prev_thrd;
+            #endif
+      };
+        #ifdef __CFORALL__
+        extern "Cforall" {
+                void ?{}( struct __thread_queue_t & );
+                void append( struct __thread_queue_t &, struct thread_desc * );
+                struct thread_desc * pop_head( struct __thread_queue_t & );
+                struct thread_desc * remove( struct __thread_queue_t &, struct thread_desc ** );
+      struct __thread_queue_t {
+            struct thread_desc * head;
+            struct thread_desc ** tail;
+      };
+                void ?{}( struct __condition_stack_t & );
+                void push( struct __condition_stack_t &, struct __condition_criterion_t * );
+                struct __condition_criterion_t * pop( struct __condition_stack_t & );
+        }
+        #endif
+      struct __condition_stack_t {
+            struct __condition_criterion_t * top;
+      };
+        struct coStack_t {
+                // size of stack
+                size_t size;
+      #ifdef __CFORALL__
+      extern "Cforall" {
+            void ?{}( struct __thread_queue_t & );
+            void append( struct __thread_queue_t *, struct thread_desc * );
+            struct thread_desc * pop_head( struct __thread_queue_t * );
+            struct thread_desc * remove( struct __thread_queue_t *, struct thread_desc ** );
+                // pointer to stack
+                void *storage;
+            void ?{}( struct __condition_stack_t & );
+            void push( struct __condition_stack_t *, struct __condition_criterion_t * );
+            struct __condition_criterion_t * pop( struct __condition_stack_t * );
+                // stack grows towards stack limit
+                void *limit;
+            void ?{}(spinlock & this);
+            void ^?{}(spinlock & this);
+      }
+      #endif
+                // base of stack
+                void *base;
+      struct coStack_t {
+            unsigned int size;                        // size of stack
+            void *storage;                            // pointer to stack
+            void *limit;                              // stack grows towards stack limit
+            void *base;                               // base of stack
+            void *context;                            // address of cfa_context_t
+            void *top;                                // address of top of storage
+            bool userStack;                           // whether or not the user allocated the stack
+      };
+                // address of cfa_context_t
+                void *context;
+      enum coroutine_state { Halted, Start, Inactive, Active, Primed };
+                // address of top of storage
+                void *top;
+      struct coroutine_desc {
+            struct coStack_t stack;                   // stack information of the coroutine
+            const char *name;                         // textual name for coroutine/task, initialized by uC++ generated code
+            int errno_;                               // copy of global UNIX variable errno
+            enum coroutine_state state;               // current execution status for coroutine
+            struct coroutine_desc * starter;          // first coroutine to resume this one
+            struct coroutine_desc * last;             // last coroutine to resume this one
+      };
+                // whether or not the user allocated the stack
+                bool userStack;
+        };
+      struct __waitfor_mask_t {
+            short * accepted;                         // the index of the accepted function, -1 if none
+            struct __acceptable_t * clauses;          // list of acceptable functions, null if any
+            short size;                               // number of acceptable functions
+      };
+        enum coroutine_state { Halted, Start, Inactive, Active, Primed };
+      struct monitor_desc {
+            struct spinlock lock;                     // spinlock to protect internal data
+            struct thread_desc * owner;               // current owner of the monitor
+            struct __thread_queue_t entry_queue;      // queue of threads that are blocked waiting for the monitor
+            struct __condition_stack_t signal_stack;  // stack of conditions to run next once we exit the monitor
+            unsigned int recursion;                   // monitor routines can be called recursively, we need to keep track of that
+            struct __waitfor_mask_t mask;             // mask used to know if some thread is waiting for something while holding the monitor
+            struct __condition_node_t * dtor_node;    // node used to signal the dtor in a waitfor dtor
+      };
+        struct coroutine_desc {
+                // stack information of the coroutine
+                struct coStack_t stack;
+      struct __monitor_group_t {
+            struct monitor_desc ** list;              // currently held monitors
+            short                  size;              // number of currently held monitors
+            fptr_t                 func;              // last function that acquired monitors
+      };
+                // textual name for coroutine/task, initialized by uC++ generated code
+                const char *name;
+      struct thread_desc {
+            // Core threading fields
+            struct coroutine_desc  self_cor;          // coroutine body used to store context
+            struct monitor_desc    self_mon;          // monitor body used for mutual exclusion
+            struct monitor_desc *  self_mon_p;        // pointer to monitor with sufficient lifetime for current monitors
+            struct __monitor_group_t monitors;        // monitors currently held by this thread
+                // copy of global UNIX variable errno
+                int errno_;
+            // Link lists fields
+            struct thread_desc * next;                // instrusive link field for threads
-                // current execution status for coroutine
-                enum coroutine_state state;
-                // first coroutine to resume this one
-                struct coroutine_desc * starter;
-                // last coroutine to resume this one
-                struct coroutine_desc * last;
-        };
-        struct __waitfor_mask_t {
-                // the index of the accepted function, -1 if none
-                short * accepted;
-                // list of acceptable functions, null if any
-                struct __acceptable_t * clauses;
-                // number of acceptable functions
-                __lock_size_t size;
-        };
-        struct monitor_desc {
-                // spinlock to protect internal data
-                struct __spinlock_t lock;
-                // current owner of the monitor
-                struct thread_desc * owner;
-                // queue of threads that are blocked waiting for the monitor
-                struct __thread_queue_t entry_queue;
-                // stack of conditions to run next once we exit the monitor
-                struct __condition_stack_t signal_stack;
-                // monitor routines can be called recursively, we need to keep track of that
-                unsigned int recursion;
-                // mask used to know if some thread is waiting for something while holding the monitor
-                struct __waitfor_mask_t mask;
-                // node used to signal the dtor in a waitfor dtor
-                struct __condition_node_t * dtor_node;
-        };
-        struct __monitor_group_t {
-                // currently held monitors
-                struct monitor_desc ** list;
-                // number of currently held monitors
-                __lock_size_t size;
-                // last function that acquired monitors
-                fptr_t func;
-        };
-        struct thread_desc {
-                // Core threading fields
-                // coroutine body used to store context
-                struct coroutine_desc  self_cor;
-                // monitor body used for mutual exclusion
-                struct monitor_desc    self_mon;
-                // pointer to monitor with sufficient lifetime for current monitors
-                struct monitor_desc *  self_mon_p;
-                // monitors currently held by this thread
-                struct __monitor_group_t monitors;
-                // Link lists fields
-                // instrusive link field for threads
-                struct thread_desc * next;
      };
      #ifdef __CFORALL__
      extern "Cforall" {
                 static inline monitor_desc * ?[?]( const __monitor_group_t & this, ptrdiff_t index ) {
                         return this.list[index];
+                }
+            static inline monitor_desc * ?[?]( const __monitor_group_t & this, ptrdiff_t index ) {
+                  return this.list[index];
+            }
                 static inline bool ?==?( const __monitor_group_t & lhs, const __monitor_group_t & rhs ) {
                         if( (lhs.list != 0) != (rhs.list != 0) ) return false;
                         if( lhs.size != rhs.size ) return false;
                         if( lhs.func != rhs.func ) return false;
+            static inline bool ?==?( const __monitor_group_t & lhs, const __monitor_group_t & rhs ) {
+                  if( (lhs.list != 0) != (rhs.list != 0) ) return false;
+                  if( lhs.size != rhs.size ) return false;
+                  if( lhs.func != rhs.func ) return false;
                         // Check that all the monitors match
                         for( int i = 0; i < lhs.size; i++ ) {
                                 // If not a match, check next function
                                 if( lhs[i] != rhs[i] ) return false;
+                        }
+                  // Check that all the monitors match
+                  for( int i = 0; i < lhs.size; i++ ) {
+                        // If not a match, check next function
+                        if( lhs[i] != rhs[i] ) return false;
+                  }
                         return true;
+                }
+        }
         #endif
+                  return true;
+            }
+      }
+      #endif
 #endif //_INVOKE_H_
 …
 #define _INVOKE_PRIVATE_H_
         struct machine_context_t {
                 void *SP;
                 void *FP;
                 void *PC;
         };
+      struct machine_context_t {
+            void *SP;
+            void *FP;
+            void *PC;
+      };
         // assembler routines that performs the context switch
         extern void CtxInvokeStub( void );
         void CtxSwitch( void * from, void * to ) asm ("CtxSwitch");
+      // assembler routines that performs the context switch
+      extern void CtxInvokeStub( void );
+      void CtxSwitch( void * from, void * to ) asm ("CtxSwitch");
         #if   defined( __x86_64__ )
         #define CtxGet( ctx ) __asm__ ( \
                         "movq %%rsp,%0\n"   \
                         "movq %%rbp,%1\n"   \
                 : "=rm" (ctx.SP), "=rm" (ctx.FP) )
         #elif defined( __i386__ )
         #define CtxGet( ctx ) __asm__ ( \
                         "movl %%esp,%0\n"   \
                         "movl %%ebp,%1\n"   \
                 : "=rm" (ctx.SP), "=rm" (ctx.FP) )
         #endif
+      #if   defined( __x86_64__ )
+      #define CtxGet( ctx ) __asm__ ( \
+                  "movq %%rsp,%0\n"   \
+                  "movq %%rbp,%1\n"   \
+            : "=rm" (ctx.SP), "=rm" (ctx.FP) )
+      #elif defined( __i386__ )
+      #define CtxGet( ctx ) __asm__ ( \
+                  "movl %%esp,%0\n"   \
+                  "movl %%ebp,%1\n"   \
+            : "=rm" (ctx.SP), "=rm" (ctx.FP) )
+      #endif
 #endif //_INVOKE_PRIVATE_H_

src/libcfa/concurrency/kernel

-              r0fe4e62
+              rf5c3b6c
 //-----------------------------------------------------------------------------
 // Locks
+// // Lock the spinlock, spin if already acquired
+// void lock      ( spinlock * DEBUG_CTX_PARAM2 );
+// // Lock the spinlock, yield repeatedly if already acquired
+// void lock_yield( spinlock * DEBUG_CTX_PARAM2 );
+// // Lock the spinlock, return false if already acquired
+// bool try_lock  ( spinlock * DEBUG_CTX_PARAM2 );
+// // Unlock the spinlock
+// void unlock    ( spinlock * );
+void lock      ( spinlock * DEBUG_CTX_PARAM2 );       // Lock the spinlock, spin if already acquired
+void lock_yield( spinlock * DEBUG_CTX_PARAM2 );       // Lock the spinlock, yield repeatedly if already acquired
+bool try_lock  ( spinlock * DEBUG_CTX_PARAM2 );       // Lock the spinlock, return false if already acquired
+void unlock    ( spinlock * );                        // Unlock the spinlock
 struct semaphore {
         __spinlock_t lock;
+        spinlock lock;
         int count;
         __thread_queue_t waiting;
 …
 void  ?{}(semaphore & this, int count = 1);
 void ^?{}(semaphore & this);
 void   P (semaphore & this);
 void   V (semaphore & this);
+void P(semaphore * this);
+void V(semaphore * this);
 …
 // Cluster
 struct cluster {
+        // Ready queue locks
+        __spinlock_t ready_queue_lock;
+        // Ready queue for threads
+        __thread_queue_t ready_queue;
+        // Preemption rate on this cluster
+        unsigned long long int preemption;
+        spinlock ready_queue_lock;                      // Ready queue locks
+        __thread_queue_t ready_queue;                   // Ready queue for threads
+        unsigned long long int preemption;              // Preemption rate on this cluster
 };
 void ?{} (cluster & this);
+void ?{}(cluster & this);
 void ^?{}(cluster & this);
 …
         FinishOpCode action_code;
         thread_desc * thrd;
         __spinlock_t * lock;
         __spinlock_t ** locks;
+        spinlock * lock;
+        spinlock ** locks;
         unsigned short lock_count;
         thread_desc ** thrds;
 …
 struct processor {
         // Main state
+        // Coroutine ctx who does keeps the state of the processor
+        struct processorCtx_t * runner;
+        // Cluster from which to get threads
+        cluster * cltr;
+        // Handle to pthreads
+        pthread_t kernel_thread;
+        struct processorCtx_t * runner;                 // Coroutine ctx who does keeps the state of the processor
+        cluster * cltr;                                 // Cluster from which to get threads
+        pthread_t kernel_thread;                        // Handle to pthreads
         // Termination
+        // Set to true to notify the processor should terminate
+        volatile bool do_terminate;
+        // Termination synchronisation
+        semaphore terminated;
+        volatile bool do_terminate;                     // Set to true to notify the processor should terminate
+        semaphore terminated;                           // Termination synchronisation
         // RunThread data
+        // Action to do after a thread is ran
+        struct FinishAction finish;
+        struct FinishAction finish;                     // Action to do after a thread is ran
         // Preemption data
+        // Node which is added in the discrete event simulaiton
+        struct alarm_node_t * preemption_alarm;
+        // If true, a preemption was triggered in an unsafe region, the processor must preempt as soon as possible
+        bool pending_preemption;
+        struct alarm_node_t * preemption_alarm;         // Node which is added in the discrete event simulaiton
+        bool pending_preemption;                        // If true, a preemption was triggered in an unsafe region, the processor must preempt as soon as possible
 #ifdef __CFA_DEBUG__
+        // Last function to enable preemption on this processor
+        char * last_enable;
+        char * last_enable;                             // Last function to enable preemption on this processor
 #endif
 };
 void  ?{}(processor & this);
 void  ?{}(processor & this, cluster * cltr);
+void ?{}(processor & this);
+void ?{}(processor & this, cluster * cltr);
 void ^?{}(processor & this);

src/libcfa/concurrency/kernel.c

-              r0fe4e62
+              rf5c3b6c
                 LIB_DEBUG_PRINT_SAFE("Kernel : core %p signaling termination\n", &this);
                 this.do_terminate = true;
                 P( this.terminated );
+                P( &this.terminated );
                 pthread_join( this.kernel_thread, NULL );
+        }
 …
+        }
         V( this->terminated );
+        V( &this->terminated );
         LIB_DEBUG_PRINT_SAFE("Kernel : core %p terminated\n", this);
 …
 void finishRunning(processor * this) {
         if( this->finish.action_code == Release ) {
                 unlock( *this->finish.lock );
+                unlock( this->finish.lock );
+        }
         else if( this->finish.action_code == Schedule ) {
 …
+        }
         else if( this->finish.action_code == Release_Schedule ) {
                 unlock( *this->finish.lock );
+                unlock( this->finish.lock );
                 ScheduleThread( this->finish.thrd );
+        }
         else if( this->finish.action_code == Release_Multi ) {
                 for(int i = 0; i < this->finish.lock_count; i++) {
                         unlock( *this->finish.locks[i] );
+                        unlock( this->finish.locks[i] );
+                }
+        }
         else if( this->finish.action_code == Release_Multi_Schedule ) {
                 for(int i = 0; i < this->finish.lock_count; i++) {
                         unlock( *this->finish.locks[i] );
+                        unlock( this->finish.locks[i] );
+                }
                 for(int i = 0; i < this->finish.thrd_count; i++) {
 …
         verifyf( thrd->next == NULL, "Expected null got %p", thrd->next );
         lock(   this_processor->cltr->ready_queue_lock DEBUG_CTX2 );
         append( this_processor->cltr->ready_queue, thrd );
         unlock( this_processor->cltr->ready_queue_lock );
+        lock(   &this_processor->cltr->ready_queue_lock DEBUG_CTX2 );
+        append( &this_processor->cltr->ready_queue, thrd );
+        unlock( &this_processor->cltr->ready_queue_lock );
         verify( disable_preempt_count > 0 );
 …
 thread_desc * nextThread(cluster * this) {
         verify( disable_preempt_count > 0 );
         lock( this->ready_queue_lock DEBUG_CTX2 );
         thread_desc * head = pop_head( this->ready_queue );
         unlock( this->ready_queue_lock );
+        lock( &this->ready_queue_lock DEBUG_CTX2 );
+        thread_desc * head = pop_head( &this->ready_queue );
+        unlock( &this->ready_queue_lock );
         verify( disable_preempt_count > 0 );
         return head;
 …
+}
 void BlockInternal( __spinlock_t * lock ) {
+void BlockInternal( spinlock * lock ) {
         disable_interrupts();
         this_processor->finish.action_code = Release;
 …
+}
 void BlockInternal( __spinlock_t * lock, thread_desc * thrd ) {
+void BlockInternal( spinlock * lock, thread_desc * thrd ) {
         assert(thrd);
         disable_interrupts();
 …
+}
 void BlockInternal(__spinlock_t * locks [], unsigned short count) {
+void BlockInternal(spinlock ** locks, unsigned short count) {
         disable_interrupts();
         this_processor->finish.action_code = Release_Multi;
 …
+}
 void BlockInternal(__spinlock_t * locks [], unsigned short lock_count, thread_desc * thrds [], unsigned short thrd_count) {
+void BlockInternal(spinlock ** locks, unsigned short lock_count, thread_desc ** thrds, unsigned short thrd_count) {
         disable_interrupts();
         this_processor->finish.action_code = Release_Multi_Schedule;
 …
+}
 void LeaveThread(__spinlock_t * lock, thread_desc * thrd) {
+void LeaveThread(spinlock * lock, thread_desc * thrd) {
         verify( disable_preempt_count > 0 );
         this_processor->finish.action_code = thrd ? Release_Schedule : Release;
 …
+}
 static __spinlock_t kernel_abort_lock;
 static __spinlock_t kernel_debug_lock;
+static spinlock kernel_abort_lock;
+static spinlock kernel_debug_lock;
 static bool kernel_abort_called = false;
 …
         // abort cannot be recursively entered by the same or different processors because all signal handlers return when
         // the globalAbort flag is true.
         lock( kernel_abort_lock DEBUG_CTX2 );
+        lock( &kernel_abort_lock DEBUG_CTX2 );
         // first task to abort ?
         if ( !kernel_abort_called ) {                   // not first task to abort ?
                 kernel_abort_called = true;
                 unlock( kernel_abort_lock );
+                unlock( &kernel_abort_lock );
+        }
         else {
                 unlock( kernel_abort_lock );
+                unlock( &kernel_abort_lock );
                 sigset_t mask;
 …
 extern "C" {
         void __lib_debug_acquire() {
                 lock( kernel_debug_lock DEBUG_CTX2 );
+                lock( &kernel_debug_lock DEBUG_CTX2 );
+        }
         void __lib_debug_release() {
                 unlock( kernel_debug_lock );
+                unlock( &kernel_debug_lock );
+        }
+}
 …
 //-----------------------------------------------------------------------------
 // Locks
+void ?{}( spinlock & this ) {
+        this.lock = 0;
+}
+void ^?{}( spinlock & this ) {
+}
+bool try_lock( spinlock * this DEBUG_CTX_PARAM2 ) {
+        return this->lock == 0 && __sync_lock_test_and_set_4( &this->lock, 1 ) == 0;
+}
+void lock( spinlock * this DEBUG_CTX_PARAM2 ) {
+        for ( unsigned int i = 1;; i += 1 ) {
+                if ( this->lock == 0 && __sync_lock_test_and_set_4( &this->lock, 1 ) == 0 ) { break; }
+        }
+        LIB_DEBUG_DO(
+                this->prev_name = caller;
+                this->prev_thrd = this_thread;
+        )
+}
+void lock_yield( spinlock * this DEBUG_CTX_PARAM2 ) {
+        for ( unsigned int i = 1;; i += 1 ) {
+                if ( this->lock == 0 && __sync_lock_test_and_set_4( &this->lock, 1 ) == 0 ) { break; }
+                yield();
+        }
+        LIB_DEBUG_DO(
+                this->prev_name = caller;
+                this->prev_thrd = this_thread;
+        )
+}
+void unlock( spinlock * this ) {
+        __sync_lock_release_4( &this->lock );
+}
 void  ?{}( semaphore & this, int count = 1 ) {
         (this.lock){};
 …
 void ^?{}(semaphore & this) {}
 void P(semaphore & this) {
         lock( this.lock DEBUG_CTX2 );
         this.count -= 1;
         if ( this.count < 0 ) {
+void P(semaphore * this) {
+        lock( &this->lock DEBUG_CTX2 );
+        this->count -= 1;
+        if ( this->count < 0 ) {
                 // queue current task
                 append( this.waiting, (thread_desc *)this_thread );
+                append( &this->waiting, (thread_desc *)this_thread );
                 // atomically release spin lock and block
                 BlockInternal( &this.lock );
+                BlockInternal( &this->lock );
+        }
         else {
             unlock( this.lock );
+        }
+}
 void V(semaphore & this) {
+            unlock( &this->lock );
+        }
+}
+void V(semaphore * this) {
         thread_desc * thrd = NULL;
         lock( this.lock DEBUG_CTX2 );
         this.count += 1;
         if ( this.count <= 0 ) {
+        lock( &this->lock DEBUG_CTX2 );
+        this->count += 1;
+        if ( this->count <= 0 ) {
                 // remove task at head of waiting list
                 thrd = pop_head( this.waiting );
+        }
         unlock( this.lock );
+                thrd = pop_head( &this->waiting );
+        }
+        unlock( &this->lock );
         // make new owner
 …
+}
 void append( __thread_queue_t & this, thread_desc * t ) {
         verify(this.tail != NULL);
         *this.tail = t;
         this.tail = &t->next;
+}
 thread_desc * pop_head( __thread_queue_t & this ) {
         thread_desc * head = this.head;
+void append( __thread_queue_t * this, thread_desc * t ) {
+        verify(this->tail != NULL);
+        *this->tail = t;
+        this->tail = &t->next;
+}
+thread_desc * pop_head( __thread_queue_t * this ) {
+        thread_desc * head = this->head;
         if( head ) {
                 this.head = head->next;
+                this->head = head->next;
                 if( !head->next ) {
                         this.tail = &this.head;
+                        this->tail = &this->head;
+                }
                 head->next = NULL;
 …
+}
 thread_desc * remove( __thread_queue_t & this, thread_desc ** it ) {
+thread_desc * remove( __thread_queue_t * this, thread_desc ** it ) {
         thread_desc * thrd = *it;
         verify( thrd );
 …
         (*it) = thrd->next;
         if( this.tail == &thrd->next ) {
                 this.tail = it;
+        if( this->tail == &thrd->next ) {
+                this->tail = it;
+        }
         thrd->next = NULL;
         verify( (this.head == NULL) == (&this.head == this.tail) );
         verify( *this.tail == NULL );
+        verify( (this->head == NULL) == (&this->head == this->tail) );
+        verify( *this->tail == NULL );
         return thrd;
+}
 …
+}
 void push( __condition_stack_t & this, __condition_criterion_t * t ) {
+void push( __condition_stack_t * this, __condition_criterion_t * t ) {
         verify( !t->next );
         t->next = this.top;
         this.top = t;
+}
 __condition_criterion_t * pop( __condition_stack_t & this ) {
         __condition_criterion_t * top = this.top;
+        t->next = this->top;
+        this->top = t;
+}
+__condition_criterion_t * pop( __condition_stack_t * this ) {
+        __condition_criterion_t * top = this->top;
         if( top ) {
                 this.top = top->next;
+                this->top = top->next;
                 top->next = NULL;
+        }

src/libcfa/concurrency/kernel_private.h

-              r0fe4e62
+              rf5c3b6c
 //Block current thread and release/wake-up the following resources
 void BlockInternal(void);
 void BlockInternal(__spinlock_t * lock);
+void BlockInternal(spinlock * lock);
 void BlockInternal(thread_desc * thrd);
 void BlockInternal(__spinlock_t * lock, thread_desc * thrd);
 void BlockInternal(__spinlock_t * locks [], unsigned short count);
 void BlockInternal(__spinlock_t * locks [], unsigned short count, thread_desc * thrds [], unsigned short thrd_count);
 void LeaveThread(__spinlock_t * lock, thread_desc * thrd);
+void BlockInternal(spinlock * lock, thread_desc * thrd);
+void BlockInternal(spinlock ** locks, unsigned short count);
+void BlockInternal(spinlock ** locks, unsigned short count, thread_desc ** thrds, unsigned short thrd_count);
+void LeaveThread(spinlock * lock, thread_desc * thrd);
 //-----------------------------------------------------------------------------
 …
 struct event_kernel_t {
         alarm_list_t alarms;
         __spinlock_t lock;
+        spinlock lock;
 };

src/libcfa/concurrency/monitor

-              r0fe4e62
+              rf5c3b6c
+}
+// static inline int ?<?(monitor_desc* lhs, monitor_desc* rhs) {
+//      return ((intptr_t)lhs) < ((intptr_t)rhs);
+// }
 struct monitor_guard_t {
         monitor_desc ** m;
         __lock_size_t   count;
+        int count;
         monitor_desc ** prev_mntrs;
         __lock_size_t   prev_count;
+        unsigned short  prev_count;
         fptr_t          prev_func;
 };
 void ?{}( monitor_guard_t & this, monitor_desc ** m, __lock_size_t count, void (*func)() );
+void ?{}( monitor_guard_t & this, monitor_desc ** m, int count, void (*func)() );
 void ^?{}( monitor_guard_t & this );
 …
         monitor_desc * m;
         monitor_desc ** prev_mntrs;
         __lock_size_t   prev_count;
+        unsigned short  prev_count;
         fptr_t          prev_func;
 };
 …
 struct __condition_criterion_t {
+        // Whether or not the criterion is met (True if met)
+        bool ready;
+        // The monitor this criterion concerns
+        monitor_desc * target;
+        // The parent node to which this criterion belongs
+        struct __condition_node_t * owner;
+        // Intrusive linked list Next field
+        __condition_criterion_t * next;
+        bool ready;                                             //Whether or not the criterion is met (True if met)
+        monitor_desc * target;                          //The monitor this criterion concerns
+        struct __condition_node_t * owner;              //The parent node to which this criterion belongs
+        __condition_criterion_t * next;         //Intrusive linked list Next field
 };
 struct __condition_node_t {
+        // Thread that needs to be woken when all criteria are met
+        thread_desc * waiting_thread;
+        // Array of criteria (Criterions are contiguous in memory)
+        __condition_criterion_t * criteria;
+        // Number of criterions in the criteria
+        __lock_size_t count;
+        // Intrusive linked list Next field
+        __condition_node_t * next;
+        // Custom user info accessible before signalling
+        uintptr_t user_info;
+        thread_desc * waiting_thread;                   //Thread that needs to be woken when all criteria are met
+        __condition_criterion_t * criteria;     //Array of criteria (Criterions are contiguous in memory)
+        unsigned short count;                           //Number of criterions in the criteria
+        __condition_node_t * next;                      //Intrusive linked list Next field
+        uintptr_t user_info;                            //Custom user info accessible before signalling
 };
 …
 };
 void ?{}(__condition_node_t & this, thread_desc * waiting_thread, __lock_size_t count, uintptr_t user_info );
+void ?{}(__condition_node_t & this, thread_desc * waiting_thread, unsigned short count, uintptr_t user_info );
 void ?{}(__condition_criterion_t & this );
 void ?{}(__condition_criterion_t & this, monitor_desc * target, __condition_node_t * owner );
 void ?{}( __condition_blocked_queue_t & );
 void append( __condition_blocked_queue_t &, __condition_node_t * );
 __condition_node_t * pop_head( __condition_blocked_queue_t & );
+void append( __condition_blocked_queue_t *, __condition_node_t * );
+__condition_node_t * pop_head( __condition_blocked_queue_t * );
 struct condition {
+        // Link list which contains the blocked threads as-well as the information needed to unblock them
+        __condition_blocked_queue_t blocked;
+        // Array of monitor pointers (Monitors are NOT contiguous in memory)
+        monitor_desc ** monitors;
+        // Number of monitors in the array
+        __lock_size_t monitor_count;
+        __condition_blocked_queue_t blocked;    //Link list which contains the blocked threads as-well as the information needed to unblock them
+        monitor_desc ** monitors;                       //Array of monitor pointers (Monitors are NOT contiguous in memory)
+        unsigned short monitor_count;                   //Number of monitors in the array
 };
 …
+}
               void wait        ( condition & this, uintptr_t user_info = 0 );
               bool signal      ( condition & this );
               bool signal_block( condition & this );
 static inline bool is_empty    ( condition & this ) { return !this.blocked.head; }
          uintptr_t front       ( condition & this );
+void wait( condition * this, uintptr_t user_info = 0 );
+bool signal( condition * this );
+bool signal_block( condition * this );
+static inline bool is_empty( condition * this ) { return !this->blocked.head; }
+uintptr_t front( condition * this );
 //-----------------------------------------------------------------------------

src/libcfa/concurrency/monitor.c

-              r0fe4e62
+              rf5c3b6c
 #include <stdlib>
-#include <inttypes.h>
 #include "libhdr.h"
 …
 // Forward declarations
 static inline void set_owner ( monitor_desc * this, thread_desc * owner );
 static inline void set_owner ( monitor_desc * storage [], __lock_size_t count, thread_desc * owner );
 static inline void set_mask  ( monitor_desc * storage [], __lock_size_t count, const __waitfor_mask_t & mask );
+static inline void set_owner ( monitor_desc ** storage, short count, thread_desc * owner );
+static inline void set_mask  ( monitor_desc ** storage, short count, const __waitfor_mask_t & mask );
 static inline void reset_mask( monitor_desc * this );
 …
 static inline bool is_accepted( monitor_desc * this, const __monitor_group_t & monitors );
 static inline void lock_all  ( __spinlock_t * locks [], __lock_size_t count );
 static inline void lock_all  ( monitor_desc * source [], __spinlock_t * /*out*/ locks [], __lock_size_t count );
 static inline void unlock_all( __spinlock_t * locks [], __lock_size_t count );
 static inline void unlock_all( monitor_desc * locks [], __lock_size_t count );
 static inline void save   ( monitor_desc * ctx [], __lock_size_t count, __spinlock_t * locks [], unsigned int /*out*/ recursions [], __waitfor_mask_t /*out*/ masks [] );
 static inline void restore( monitor_desc * ctx [], __lock_size_t count, __spinlock_t * locks [], unsigned int /*in */ recursions [], __waitfor_mask_t /*in */ masks [] );
 static inline void init     ( __lock_size_t count, monitor_desc * monitors [], __condition_node_t & waiter, __condition_criterion_t criteria [] );
 static inline void init_push( __lock_size_t count, monitor_desc * monitors [], __condition_node_t & waiter, __condition_criterion_t criteria [] );
+static inline void lock_all( spinlock ** locks, unsigned short count );
+static inline void lock_all( monitor_desc ** source, spinlock ** /*out*/ locks, unsigned short count );
+static inline void unlock_all( spinlock ** locks, unsigned short count );
+static inline void unlock_all( monitor_desc ** locks, unsigned short count );
+static inline void save   ( monitor_desc ** ctx, short count, spinlock ** locks, unsigned int * /*out*/ recursions, __waitfor_mask_t * /*out*/ masks );
+static inline void restore( monitor_desc ** ctx, short count, spinlock ** locks, unsigned int * /*in */ recursions, __waitfor_mask_t * /*in */ masks );
+static inline void init     ( int count, monitor_desc ** monitors, __condition_node_t * waiter, __condition_criterion_t * criteria );
+static inline void init_push( int count, monitor_desc ** monitors, __condition_node_t * waiter, __condition_criterion_t * criteria );
 static inline thread_desc *        check_condition   ( __condition_criterion_t * );
 static inline void                 brand_condition   ( condition & );
 static inline [thread_desc *, int] search_entry_queue( const __waitfor_mask_t &, monitor_desc * monitors [], __lock_size_t count );
+static inline void                 brand_condition   ( condition * );
+static inline [thread_desc *, int] search_entry_queue( const __waitfor_mask_t &, monitor_desc ** monitors, int count );
 forall(dtype T | sized( T ))
+static inline __lock_size_t insert_unique( T * array [], __lock_size_t & size, T * val );
+static inline __lock_size_t count_max    ( const __waitfor_mask_t & mask );
+static inline __lock_size_t aggregate    ( monitor_desc * storage [], const __waitfor_mask_t & mask );
+#ifndef __CFA_LOCK_NO_YIELD
+#define DO_LOCK lock_yield
+#else
+#define DO_LOCK lock
+#endif
+static inline short insert_unique( T ** array, short & size, T * val );
+static inline short count_max    ( const __waitfor_mask_t & mask );
+static inline short aggregate    ( monitor_desc ** storage, const __waitfor_mask_t & mask );
 //-----------------------------------------------------------------------------
 …
         __condition_node_t waiter = { thrd, count, user_info };   /* Create the node specific to this wait operation                                     */ \
         __condition_criterion_t criteria[count];                  /* Create the creteria this wait operation needs to wake up                            */ \
         init( count, monitors, waiter, criteria );                /* Link everything together                                                            */ \
+        init( count, monitors, &waiter, criteria );               /* Link everything together                                                            */ \
 #define wait_ctx_primed(thrd, user_info)                        /* Create the necessary information to use the signaller stack                         */ \
         __condition_node_t waiter = { thrd, count, user_info };   /* Create the node specific to this wait operation                                     */ \
         __condition_criterion_t criteria[count];                  /* Create the creteria this wait operation needs to wake up                            */ \
         init_push( count, monitors, waiter, criteria );           /* Link everything together and push it to the AS-Stack                                */ \
+        init_push( count, monitors, &waiter, criteria );          /* Link everything together and push it to the AS-Stack                                */ \
 #define monitor_ctx( mons, cnt )                                /* Define that create the necessary struct for internal/external scheduling operations */ \
         monitor_desc ** monitors = mons;                          /* Save the targeted monitors                                                          */ \
         __lock_size_t count = cnt;                                /* Save the count to a local variable                                                  */ \
+        unsigned short count = cnt;                               /* Save the count to a local variable                                                  */ \
         unsigned int recursions[ count ];                         /* Save the current recursion levels to restore them later                             */ \
         __waitfor_mask_t masks [ count ];                         /* Save the current waitfor masks to restore them later                                */ \
         __spinlock_t *   locks [ count ];                         /* We need to pass-in an array of locks to BlockInternal                               */ \
+        __waitfor_mask_t masks[ count ];                          /* Save the current waitfor masks to restore them later                                */ \
+        spinlock *   locks     [ count ];                         /* We need to pass-in an array of locks to BlockInternal                               */ \
 #define monitor_save    save   ( monitors, count, locks, recursions, masks )
 …
         // Enter single monitor
         static void __enter_monitor_desc( monitor_desc * this, const __monitor_group_t & group ) {
                 // Lock the monitor spinlock
                 DO_LOCK( this->lock DEBUG_CTX2 );
+                // Lock the monitor spinlock, lock_yield to reduce contention
+                lock_yield( &this->lock DEBUG_CTX2 );
                 thread_desc * thrd = this_thread;
 …
                         // Some one else has the monitor, wait in line for it
                         append( this->entry_queue, thrd );
+                        append( &this->entry_queue, thrd );
                         BlockInternal( &this->lock );
 …
                 // Release the lock and leave
                 unlock( this->lock );
+                unlock( &this->lock );
                 return;
+        }
         static void __enter_monitor_dtor( monitor_desc * this, fptr_t func ) {
                 // Lock the monitor spinlock
                 DO_LOCK( this->lock DEBUG_CTX2 );
+                // Lock the monitor spinlock, lock_yield to reduce contention
+                lock_yield( &this->lock DEBUG_CTX2 );
                 thread_desc * thrd = this_thread;
 …
                         set_owner( this, thrd );
                         unlock( this->lock );
+                        unlock( &this->lock );
                         return;
+                }
 …
+                }
                 __lock_size_t count = 1;
+                int count = 1;
                 monitor_desc ** monitors = &this;
                 __monitor_group_t group = { &this, 1, func };
 …
                         // Wake the thread that is waiting for this
                         __condition_criterion_t * urgent = pop( this->signal_stack );
+                        __condition_criterion_t * urgent = pop( &this->signal_stack );
                         verify( urgent );
 …
                         // Some one else has the monitor, wait in line for it
                         append( this->entry_queue, thrd );
+                        append( &this->entry_queue, thrd );
                         BlockInternal( &this->lock );
 …
         // Leave single monitor
         void __leave_monitor_desc( monitor_desc * this ) {
                 // Lock the monitor spinlock, DO_LOCK to reduce contention
                 DO_LOCK( this->lock DEBUG_CTX2 );
+                // Lock the monitor spinlock, lock_yield to reduce contention
+                lock_yield( &this->lock DEBUG_CTX2 );
                 LIB_DEBUG_PRINT_SAFE("Kernel : %10p Leaving mon %p (%p)\n", this_thread, this, this->owner);
 …
                 if( this->recursion != 0) {
                         LIB_DEBUG_PRINT_SAFE("Kernel :  recursion still %d\n", this->recursion);
                         unlock( this->lock );
+                        unlock( &this->lock );
                         return;
+                }
 …
                 // We can now let other threads in safely
                 unlock( this->lock );
+                unlock( &this->lock );
                 //We need to wake-up the thread
 …
                 // Lock the monitor now
                 DO_LOCK( this->lock DEBUG_CTX2 );
+                lock_yield( &this->lock DEBUG_CTX2 );
                 disable_interrupts();
 …
 // relies on the monitor array being sorted
 static inline void enter( __monitor_group_t monitors ) {
         for( __lock_size_t i = 0; i < monitors.size; i++) {
+        for(int i = 0; i < monitors.size; i++) {
                 __enter_monitor_desc( monitors.list[i], monitors );
+        }
 …
 // Leave multiple monitor
 // relies on the monitor array being sorted
 static inline void leave(monitor_desc * monitors [], __lock_size_t count) {
         for( __lock_size_t i = count - 1; i >= 0; i--) {
+static inline void leave(monitor_desc ** monitors, int count) {
+        for(int i = count - 1; i >= 0; i--) {
                 __leave_monitor_desc( monitors[i] );
+        }
 …
 // Ctor for monitor guard
 // Sorts monitors before entering
 void ?{}( monitor_guard_t & this, monitor_desc * m [], __lock_size_t count, fptr_t func ) {
+void ?{}( monitor_guard_t & this, monitor_desc ** m, int count, fptr_t func ) {
         // Store current array
         this.m = m;
 …
         // Save previous thread context
+        this.[prev_mntrs, prev_count, prev_func] = this_thread->monitors.[list, size, func];
+        this.prev_mntrs = this_thread->monitors.list;
+        this.prev_count = this_thread->monitors.size;
+        this.prev_func  = this_thread->monitors.func;
         // Update thread context (needed for conditions)
+        this_thread->monitors.[list, size, func] = [m, count, func];
+        this_thread->monitors.list = m;
+        this_thread->monitors.size = count;
+        this_thread->monitors.func = func;
         // LIB_DEBUG_PRINT_SAFE("MGUARD : enter %d\n", count);
 …
         // Restore thread context
+        this_thread->monitors.[list, size, func] = this.[prev_mntrs, prev_count, prev_func];
+}
+        this_thread->monitors.list = this.prev_mntrs;
+        this_thread->monitors.size = this.prev_count;
+        this_thread->monitors.func = this.prev_func;
+}
 // Ctor for monitor guard
 // Sorts monitors before entering
 void ?{}( monitor_dtor_guard_t & this, monitor_desc * m [], fptr_t func ) {
+void ?{}( monitor_dtor_guard_t & this, monitor_desc ** m, fptr_t func ) {
         // Store current array
         this.m = *m;
         // Save previous thread context
+        this.[prev_mntrs, prev_count, prev_func] = this_thread->monitors.[list, size, func];
+        this.prev_mntrs = this_thread->monitors.list;
+        this.prev_count = this_thread->monitors.size;
+        this.prev_func  = this_thread->monitors.func;
         // Update thread context (needed for conditions)
+        this_thread->monitors.[list, size, func] = [m, 1, func];
+        this_thread->monitors.list = m;
+        this_thread->monitors.size = 1;
+        this_thread->monitors.func = func;
         __enter_monitor_dtor( this.m, func );
+}
 // Dtor for monitor guard
 …
         // Restore thread context
+        this_thread->monitors.[list, size, func] = this.[prev_mntrs, prev_count, prev_func];
+        this_thread->monitors.list = this.prev_mntrs;
+        this_thread->monitors.size = this.prev_count;
+        this_thread->monitors.func = this.prev_func;
+}
 //-----------------------------------------------------------------------------
 // Internal scheduling types
 void ?{}(__condition_node_t & this, thread_desc * waiting_thread, __lock_size_t count, uintptr_t user_info ) {
+void ?{}(__condition_node_t & this, thread_desc * waiting_thread, unsigned short count, uintptr_t user_info ) {
         this.waiting_thread = waiting_thread;
         this.count = count;
 …
+}
 void ?{}(__condition_criterion_t & this, monitor_desc * target, __condition_node_t & owner ) {
+void ?{}(__condition_criterion_t & this, monitor_desc * target, __condition_node_t * owner ) {
         this.ready  = false;
         this.target = target;
         this.owner  = &owner;
+        this.owner  = owner;
         this.next   = NULL;
+}
 …
 //-----------------------------------------------------------------------------
 // Internal scheduling
 void wait( condition & this, uintptr_t user_info = 0 ) {
+void wait( condition * this, uintptr_t user_info = 0 ) {
         brand_condition( this );
         // Check that everything is as expected
         assertf( this.monitors != NULL, "Waiting with no monitors (%p)", this.monitors );
         verifyf( this.monitor_count != 0, "Waiting with 0 monitors (%"PRIiFAST16")", this.monitor_count );
         verifyf( this.monitor_count < 32u, "Excessive monitor count (%"PRIiFAST16")", this.monitor_count );
+        assertf( this->monitors != NULL, "Waiting with no monitors (%p)", this->monitors );
+        verifyf( this->monitor_count != 0, "Waiting with 0 monitors (%i)", this->monitor_count );
+        verifyf( this->monitor_count < 32u, "Excessive monitor count (%i)", this->monitor_count );
         // Create storage for monitor context
         monitor_ctx( this.monitors, this.monitor_count );
+        monitor_ctx( this->monitors, this->monitor_count );
         // Create the node specific to this wait operation
 …
         // Append the current wait operation to the ones already queued on the condition
         // We don't need locks for that since conditions must always be waited on inside monitor mutual exclusion
         append( this.blocked, &waiter );
+        append( &this->blocked, &waiter );
         // Lock all monitors (aggregates the locks as well)
 …
         // Find the next thread(s) to run
         __lock_size_t thread_count = 0;
+        short thread_count = 0;
         thread_desc * threads[ count ];
         __builtin_memset( threads, 0, sizeof( threads ) );
 …
         // Remove any duplicate threads
         for( __lock_size_t i = 0; i < count; i++) {
+        for( int i = 0; i < count; i++) {
                 thread_desc * new_owner = next_thread( monitors[i] );
                 insert_unique( threads, thread_count, new_owner );
 …
+}
 bool signal( condition & this ) {
+bool signal( condition * this ) {
         if( is_empty( this ) ) { return false; }
         //Check that everything is as expected
         verify( this.monitors );
         verify( this.monitor_count != 0 );
+        verify( this->monitors );
+        verify( this->monitor_count != 0 );
         //Some more checking in debug
         LIB_DEBUG_DO(
                 thread_desc * this_thrd = this_thread;
                 if ( this.monitor_count != this_thrd->monitors.size ) {
                         abortf( "Signal on condition %p made with different number of monitor(s), expected %i got %i", &this, this.monitor_count, this_thrd->monitors.size );
+                }
                 for(int i = 0; i < this.monitor_count; i++) {
                         if ( this.monitors[i] != this_thrd->monitors.list[i] ) {
                                 abortf( "Signal on condition %p made with different monitor, expected %p got %i", &this, this.monitors[i], this_thrd->monitors.list[i] );
+                if ( this->monitor_count != this_thrd->monitors.size ) {
+                        abortf( "Signal on condition %p made with different number of monitor(s), expected %i got %i", this, this->monitor_count, this_thrd->monitors.size );
+                }
+                for(int i = 0; i < this->monitor_count; i++) {
+                        if ( this->monitors[i] != this_thrd->monitors.list[i] ) {
+                                abortf( "Signal on condition %p made with different monitor, expected %p got %i", this, this->monitors[i], this_thrd->monitors.list[i] );
+                        }
+                }
         );
         __lock_size_t count = this.monitor_count;
+        unsigned short count = this->monitor_count;
         // Lock all monitors
         lock_all( this.monitors, NULL, count );
+        lock_all( this->monitors, NULL, count );
         //Pop the head of the waiting queue
         __condition_node_t * node = pop_head( this.blocked );
+        __condition_node_t * node = pop_head( &this->blocked );
         //Add the thread to the proper AS stack
 …
                 __condition_criterion_t * crit = &node->criteria[i];
                 assert( !crit->ready );
                 push( crit->target->signal_stack, crit );
+                push( &crit->target->signal_stack, crit );
+        }
         //Release
         unlock_all( this.monitors, count );
+        unlock_all( this->monitors, count );
         return true;
+}
 bool signal_block( condition & this ) {
         if( !this.blocked.head ) { return false; }
+bool signal_block( condition * this ) {
+        if( !this->blocked.head ) { return false; }
         //Check that everything is as expected
         verifyf( this.monitors != NULL, "Waiting with no monitors (%p)", this.monitors );
         verifyf( this.monitor_count != 0, "Waiting with 0 monitors (%"PRIiFAST16")", this.monitor_count );
+        verifyf( this->monitors != NULL, "Waiting with no monitors (%p)", this->monitors );
+        verifyf( this->monitor_count != 0, "Waiting with 0 monitors (%i)", this->monitor_count );
         // Create storage for monitor context
         monitor_ctx( this.monitors, this.monitor_count );
+        monitor_ctx( this->monitors, this->monitor_count );
         // Lock all monitors (aggregates the locks them as well)
 …
         //Find the thread to run
         thread_desc * signallee = pop_head( this.blocked )->waiting_thread;
+        thread_desc * signallee = pop_head( &this->blocked )->waiting_thread;
         set_owner( monitors, count, signallee );
         LIB_DEBUG_PRINT_BUFFER_DECL( "Kernel : signal_block condition %p (s: %p)\n", &this, signallee );
+        LIB_DEBUG_PRINT_BUFFER_DECL( "Kernel : signal_block condition %p (s: %p)\n", this, signallee );
         //Everything is ready to go to sleep
 …
 // Access the user_info of the thread waiting at the front of the queue
 uintptr_t front( condition & this ) {
+uintptr_t front( condition * this ) {
         verifyf( !is_empty(this),
                 "Attempt to access user data on an empty condition.\n"
                 "Possible cause is not checking if the condition is empty before reading stored data."
         );
         return this.blocked.head->user_info;
+        return this->blocked.head->user_info;
+}
 …
         // This statment doesn't have a contiguous list of monitors...
         // Create one!
         __lock_size_t max = count_max( mask );
+        short max = count_max( mask );
         monitor_desc * mon_storage[max];
         __builtin_memset( mon_storage, 0, sizeof( mon_storage ) );
         __lock_size_t actual_count = aggregate( mon_storage, mask );
         LIB_DEBUG_PRINT_BUFFER_DECL( "Kernel : waitfor %d (s: %d, m: %d)\n", actual_count, mask.size, (__lock_size_t)max);
+        short actual_count = aggregate( mon_storage, mask );
+        LIB_DEBUG_PRINT_BUFFER_DECL( "Kernel : waitfor %d (s: %d, m: %d)\n", actual_count, mask.size, (short)max);
         if(actual_count == 0) return;
 …
                                 __condition_criterion_t * dtor_crit = mon2dtor->dtor_node->criteria;
                                 push( mon2dtor->signal_stack, dtor_crit );
+                                push( &mon2dtor->signal_stack, dtor_crit );
                                 unlock_all( locks, count );
 …
         set_mask( monitors, count, mask );
         for( __lock_size_t i = 0; i < count; i++) {
+        for(int i = 0; i < count; i++) {
                 verify( monitors[i]->owner == this_thread );
+        }
 …
+}
 static inline void set_owner( monitor_desc * monitors [], __lock_size_t count, thread_desc * owner ) {
+static inline void set_owner( monitor_desc ** monitors, short count, thread_desc * owner ) {
         monitors[0]->owner     = owner;
         monitors[0]->recursion = 1;
         for( __lock_size_t i = 1; i < count; i++ ) {
+        for( int i = 1; i < count; i++ ) {
                 monitors[i]->owner     = owner;
                 monitors[i]->recursion = 0;
 …
+}
 static inline void set_mask( monitor_desc * storage [], __lock_size_t count, const __waitfor_mask_t & mask ) {
         for( __lock_size_t i = 0; i < count; i++) {
+static inline void set_mask( monitor_desc ** storage, short count, const __waitfor_mask_t & mask ) {
+        for(int i = 0; i < count; i++) {
                 storage[i]->mask = mask;
+        }
 …
         //Check the signaller stack
         LIB_DEBUG_PRINT_SAFE("Kernel :  mon %p AS-stack top %p\n", this, this->signal_stack.top);
         __condition_criterion_t * urgent = pop( this->signal_stack );
+        __condition_criterion_t * urgent = pop( &this->signal_stack );
         if( urgent ) {
                 //The signaller stack is not empty,
 …
         // No signaller thread
         // Get the next thread in the entry_queue
         thread_desc * new_owner = pop_head( this->entry_queue );
+        thread_desc * new_owner = pop_head( &this->entry_queue );
         set_owner( this, new_owner );
 …
 static inline bool is_accepted( monitor_desc * this, const __monitor_group_t & group ) {
         __acceptable_t * it = this->mask.clauses; // Optim
         __lock_size_t count = this->mask.size;
+        int count = this->mask.size;
         // Check if there are any acceptable functions
 …
         // For all acceptable functions check if this is the current function.
         for( __lock_size_t i = 0; i < count; i++, it++ ) {
+        for( short i = 0; i < count; i++, it++ ) {
                 if( *it == group ) {
                         *this->mask.accepted = i;
 …
+}
 static inline void init( __lock_size_t count, monitor_desc * monitors [], __condition_node_t & waiter, __condition_criterion_t criteria [] ) {
         for( __lock_size_t i = 0; i < count; i++) {
+static inline void init( int count, monitor_desc ** monitors, __condition_node_t * waiter, __condition_criterion_t * criteria ) {
+        for(int i = 0; i < count; i++) {
                 (criteria[i]){ monitors[i], waiter };
+        }
         waiter.criteria = criteria;
+}
 static inline void init_push( __lock_size_t count, monitor_desc * monitors [], __condition_node_t & waiter, __condition_criterion_t criteria [] ) {
         for( __lock_size_t i = 0; i < count; i++) {
+        waiter->criteria = criteria;
+}
+static inline void init_push( int count, monitor_desc ** monitors, __condition_node_t * waiter, __condition_criterion_t * criteria ) {
+        for(int i = 0; i < count; i++) {
                 (criteria[i]){ monitors[i], waiter };
                 LIB_DEBUG_PRINT_SAFE( "Kernel :  target %p = %p\n", criteria[i].target, &criteria[i] );
                 push( criteria[i].target->signal_stack, &criteria[i] );
+        }
         waiter.criteria = criteria;
+}
 static inline void lock_all( __spinlock_t * locks [], __lock_size_t count ) {
         for( __lock_size_t i = 0; i < count; i++ ) {
                 DO_LOCK( *locks[i] DEBUG_CTX2 );
+        }
+}
 static inline void lock_all( monitor_desc * source [], __spinlock_t * /*out*/ locks [], __lock_size_t count ) {
         for( __lock_size_t i = 0; i < count; i++ ) {
                 __spinlock_t * l = &source[i]->lock;
                 DO_LOCK( *l DEBUG_CTX2 );
+                push( &criteria[i].target->signal_stack, &criteria[i] );
+        }
+        waiter->criteria = criteria;
+}
+static inline void lock_all( spinlock ** locks, unsigned short count ) {
+        for( int i = 0; i < count; i++ ) {
+                lock_yield( locks[i] DEBUG_CTX2 );
+        }
+}
+static inline void lock_all( monitor_desc ** source, spinlock ** /*out*/ locks, unsigned short count ) {
+        for( int i = 0; i < count; i++ ) {
+                spinlock * l = &source[i]->lock;
+                lock_yield( l DEBUG_CTX2 );
                 if(locks) locks[i] = l;
+        }
+}
+static inline void unlock_all( __spinlock_t * locks [], __lock_size_t count ) {
+        for( __lock_size_t i = 0; i < count; i++ ) {
+                unlock( *locks[i] );
+        }
+}
+static inline void unlock_all( monitor_desc * locks [], __lock_size_t count ) {
+        for( __lock_size_t i = 0; i < count; i++ ) {
+                unlock( locks[i]->lock );
+        }
+}
+static inline void save(
+        monitor_desc * ctx [],
+        __lock_size_t count,
+        __attribute((unused)) __spinlock_t * locks [],
+        unsigned int /*out*/ recursions [],
+        __waitfor_mask_t /*out*/ masks []
+) {
+        for( __lock_size_t i = 0; i < count; i++ ) {
+static inline void unlock_all( spinlock ** locks, unsigned short count ) {
+        for( int i = 0; i < count; i++ ) {
+                unlock( locks[i] );
+        }
+}
+static inline void unlock_all( monitor_desc ** locks, unsigned short count ) {
+        for( int i = 0; i < count; i++ ) {
+                unlock( &locks[i]->lock );
+        }
+}
+static inline void save( monitor_desc ** ctx, short count, __attribute((unused)) spinlock ** locks, unsigned int * /*out*/ recursions, __waitfor_mask_t * /*out*/ masks ) {
+        for( int i = 0; i < count; i++ ) {
                 recursions[i] = ctx[i]->recursion;
                 masks[i]      = ctx[i]->mask;
 …
+}
+static inline void restore(
+        monitor_desc * ctx [],
+        __lock_size_t count,
+        __spinlock_t * locks [],
+        unsigned int /*out*/ recursions [],
+        __waitfor_mask_t /*out*/ masks []
+) {
+static inline void restore( monitor_desc ** ctx, short count, spinlock ** locks, unsigned int * /*out*/ recursions, __waitfor_mask_t * /*out*/ masks ) {
         lock_all( locks, count );
         for( __lock_size_t i = 0; i < count; i++ ) {
+        for( int i = 0; i < count; i++ ) {
                 ctx[i]->recursion = recursions[i];
                 ctx[i]->mask      = masks[i];
 …
+}
 static inline void brand_condition( condition & this ) {
+static inline void brand_condition( condition * this ) {
         thread_desc * thrd = this_thread;
         if( !this.monitors ) {
+        if( !this->monitors ) {
                 // LIB_DEBUG_PRINT_SAFE("Branding\n");
                 assertf( thrd->monitors.list != NULL, "No current monitor to brand condition %p", thrd->monitors.list );
                 this.monitor_count = thrd->monitors.size;
                 this.monitors = malloc( this.monitor_count * sizeof( *this.monitors ) );
                 for( int i = 0; i < this.monitor_count; i++ ) {
                         this.monitors[i] = thrd->monitors.list[i];
+                }
+        }
+}
 static inline [thread_desc *, int] search_entry_queue( const __waitfor_mask_t & mask, monitor_desc * monitors [], __lock_size_t count ) {
         __thread_queue_t & entry_queue = monitors[0]->entry_queue;
+                this->monitor_count = thrd->monitors.size;
+                this->monitors = malloc( this->monitor_count * sizeof( *this->monitors ) );
+                for( int i = 0; i < this->monitor_count; i++ ) {
+                        this->monitors[i] = thrd->monitors.list[i];
+                }
+        }
+}
+static inline [thread_desc *, int] search_entry_queue( const __waitfor_mask_t & mask, monitor_desc ** monitors, int count ) {
+        __thread_queue_t * entry_queue = &monitors[0]->entry_queue;
         // For each thread in the entry-queue
         for(    thread_desc ** thrd_it = &entry_queue.head;
+        for(    thread_desc ** thrd_it = &entry_queue->head;
                 *thrd_it;
                 thrd_it = &(*thrd_it)->next
 …
 forall(dtype T | sized( T ))
 static inline __lock_size_t insert_unique( T * array [], __lock_size_t & size, T * val ) {
+static inline short insert_unique( T ** array, short & size, T * val ) {
         if( !val ) return size;
         for( __lock_size_t i = 0; i <= size; i++) {
+        for(int i = 0; i <= size; i++) {
                 if( array[i] == val ) return size;
+        }
 …
+}
 static inline __lock_size_t count_max( const __waitfor_mask_t & mask ) {
         __lock_size_t max = 0;
         for( __lock_size_t i = 0; i < mask.size; i++ ) {
+static inline short count_max( const __waitfor_mask_t & mask ) {
+        short max = 0;
+        for( int i = 0; i < mask.size; i++ ) {
                 max += mask.clauses[i].size;
+        }
 …
+}
 static inline __lock_size_t aggregate( monitor_desc * storage [], const __waitfor_mask_t & mask ) {
         __lock_size_t size = 0;
         for( __lock_size_t i = 0; i < mask.size; i++ ) {
+static inline short aggregate( monitor_desc ** storage, const __waitfor_mask_t & mask ) {
+        short size = 0;
+        for( int i = 0; i < mask.size; i++ ) {
                 __libcfa_small_sort( mask.clauses[i].list, mask.clauses[i].size );
                 for( __lock_size_t j = 0; j < mask.clauses[i].size; j++) {
+                for( int j = 0; j < mask.clauses[i].size; j++) {
                         insert_unique( storage, size, mask.clauses[i].list[j] );
+                }
 …
+}
 void append( __condition_blocked_queue_t & this, __condition_node_t * c ) {
         verify(this.tail != NULL);
         *this.tail = c;
         this.tail = &c->next;
+}
 __condition_node_t * pop_head( __condition_blocked_queue_t & this ) {
         __condition_node_t * head = this.head;
+void append( __condition_blocked_queue_t * this, __condition_node_t * c ) {
+        verify(this->tail != NULL);
+        *this->tail = c;
+        this->tail = &c->next;
+}
+__condition_node_t * pop_head( __condition_blocked_queue_t * this ) {
+        __condition_node_t * head = this->head;
         if( head ) {
                 this.head = head->next;
+                this->head = head->next;
                 if( !head->next ) {
                         this.tail = &this.head;
+                        this->tail = &this->head;
+                }
                 head->next = NULL;

src/libcfa/concurrency/preemption.c

-              r0fe4e62
+              rf5c3b6c
                 case SI_KERNEL:
                         // LIB_DEBUG_PRINT_SAFE("Kernel : Preemption thread tick\n");
                         lock( event_kernel->lock DEBUG_CTX2 );
+                        lock( &event_kernel->lock DEBUG_CTX2 );
                         tick_preemption();
                         unlock( event_kernel->lock );
+                        unlock( &event_kernel->lock );
                         break;
                 // Signal was not sent by the kernel but by an other thread

src/main.cc

r0fe4e62	rf5c3b6c
206	206	FILE * extras = fopen( libcfap \| treep ? "../prelude/extras.cf" : CFA_LIBDIR "/extras.cf", "r" );
207	207	assertf( extras, "cannot open extras.cf\n" );
208		parse( extras, LinkageSpec::~~Builtin~~C );
	208	parse( extras, LinkageSpec::C );
209	209
210	210	if ( ! libcfap ) {

src/prelude/builtins.c

-              r0fe4e62
+              rf5c3b6c
 } // ?\?
+// FIXME (x \ (unsigned long int)y) relies on X ?\?(T, unsigned long) a function that is neither
+// defined, nor passed as an assertion parameter. Without user-defined conversions, cannot specify
+// X as a type that casts to double, yet it doesn't make sense to write functions with that type
+// signature where X is double.
+// static inline forall( otype T | { void ?{}( T & this, one_t ); T ?*?( T, T ); double ?/?( double, T ); } )
+// double ?\?( T x, signed long int y ) {
+//     if ( y >=  0 ) return (double)(x \ (unsigned long int)y);
+//     else return 1.0 / x \ (unsigned long int)(-y);
+// } // ?\?
+static inline forall( otype T | { void ?{}( T & this, one_t ); T ?*?( T, T ); double ?/?( double, T ); } )
+double ?\?( T x, signed long int y ) {
+    if ( y >=  0 ) return (double)(x \ (unsigned long int)y);
+    else return 1.0 / x \ (unsigned long int)(-y);
+} // ?\?
 static inline long int ?\=?( long int & x, unsigned long int y ) { x = x \ y; return x; }

src/tests/.expect/32/KRfunctions.txt

-              r0fe4e62
+              rf5c3b6c
+__attribute__ ((__nothrow__,__leaf__,__malloc__)) extern void *malloc(unsigned int __size);
+__attribute__ ((__nothrow__,__leaf__)) extern void free(void *__ptr);
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void abort(void);
+__attribute__ ((__nothrow__,__leaf__,__nonnull__(1))) extern signed int atexit(void (*__func)(void));
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void exit(signed int __status);
+extern signed int printf(const char *__restrict __format, ...);
 signed int __f0__Fi_iPCii__1(signed int __a__i_1, const signed int *__b__PCi_1, signed int __c__i_1){
     __attribute__ ((unused)) signed int ___retval_f0__i_1;

src/tests/.expect/32/attributes.txt

-              r0fe4e62
+              rf5c3b6c
+__attribute__ ((__nothrow__,__leaf__,__malloc__)) extern void *malloc(unsigned int __size);
+__attribute__ ((__nothrow__,__leaf__)) extern void free(void *__ptr);
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void abort(void);
+__attribute__ ((__nothrow__,__leaf__,__nonnull__(1))) extern signed int atexit(void (*__func)(void));
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void exit(signed int __status);
+extern signed int printf(const char *__restrict __format, ...);
 signed int __la__Fi___1(){
     __attribute__ ((unused)) signed int ___retval_la__i_1;

src/tests/.expect/32/declarationSpecifier.txt

-              r0fe4e62
+              rf5c3b6c
+__attribute__ ((__nothrow__,__leaf__,__malloc__)) extern void *malloc(unsigned int __size);
+__attribute__ ((__nothrow__,__leaf__)) extern void free(void *__ptr);
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void abort(void);
+__attribute__ ((__nothrow__,__leaf__,__nonnull__(1))) extern signed int atexit(void (*__func)(void));
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void exit(signed int __status);
+extern signed int printf(const char *__restrict __format, ...);
 volatile const signed short int __x1__CVs_1;
 static volatile const signed short int __x2__CVs_1;
 …
+}
 static inline int invoke_main(int argc, char* argv[], char* envp[]) { (void)argc; (void)argv; (void)envp; return __main__Fi_iPPCc__1(argc, argv); }
+__attribute__ ((__nothrow__,__leaf__,__malloc__)) extern void *malloc(unsigned int __size);
+__attribute__ ((__nothrow__,__leaf__)) extern void free(void *__ptr);
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void abort(void);
+__attribute__ ((__nothrow__,__leaf__,__nonnull__(1))) extern signed int atexit(void (*__func)(void));
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void exit(signed int __status);
+extern signed int printf(const char *__restrict __format, ...);
 static inline signed int invoke_main(signed int argc, char **argv, char **envp);
 signed int main(signed int __argc__i_1, char **__argv__PPc_1, char **__envp__PPc_1){

src/tests/.expect/32/extension.txt

-              r0fe4e62
+              rf5c3b6c
+__attribute__ ((__nothrow__,__leaf__,__malloc__)) extern void *malloc(unsigned int __size);
+__attribute__ ((__nothrow__,__leaf__)) extern void free(void *__ptr);
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void abort(void);
+__attribute__ ((__nothrow__,__leaf__,__nonnull__(1))) extern signed int atexit(void (*__func)(void));
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void exit(signed int __status);
+extern signed int printf(const char *__restrict __format, ...);
 __extension__ signed int __a__i_1;
 __extension__ signed int __b__i_1;

src/tests/.expect/32/gccExtensions.txt

-              r0fe4e62
+              rf5c3b6c
+__attribute__ ((__nothrow__,__leaf__,__malloc__)) extern void *malloc(unsigned int __size);
+__attribute__ ((__nothrow__,__leaf__)) extern void free(void *__ptr);
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void abort(void);
+__attribute__ ((__nothrow__,__leaf__,__nonnull__(1))) extern signed int atexit(void (*__func)(void));
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void exit(signed int __status);
+extern signed int printf(const char *__restrict __format, ...);
 extern signed int __x__i_1 asm ( "xx" );
 signed int __main__Fi_iPPCc__1(signed int __argc__i_1, const char **__argv__PPCc_1){
 …
+}
 static inline int invoke_main(int argc, char* argv[], char* envp[]) { (void)argc; (void)argv; (void)envp; return __main__Fi_iPPCc__1(argc, argv); }
+__attribute__ ((__nothrow__,__leaf__,__malloc__)) extern void *malloc(unsigned int __size);
+__attribute__ ((__nothrow__,__leaf__)) extern void free(void *__ptr);
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void abort(void);
+__attribute__ ((__nothrow__,__leaf__,__nonnull__(1))) extern signed int atexit(void (*__func)(void));
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void exit(signed int __status);
+extern signed int printf(const char *__restrict __format, ...);
 static inline signed int invoke_main(signed int argc, char **argv, char **envp);
 signed int main(signed int __argc__i_1, char **__argv__PPc_1, char **__envp__PPc_1){

src/tests/.expect/32/literals.txt

-              r0fe4e62
+              rf5c3b6c
+__attribute__ ((__nothrow__,__leaf__,__malloc__)) extern void *malloc(unsigned int __size);
+__attribute__ ((__nothrow__,__leaf__)) extern void free(void *__ptr);
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void abort(void);
+__attribute__ ((__nothrow__,__leaf__,__nonnull__(1))) extern signed int atexit(void (*__func)(void));
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void exit(signed int __status);
+extern signed int printf(const char *__restrict __format, ...);
 void __for_each__A0_2_0_0____operator_assign__PFd0_Rd0d0____constructor__PF_Rd0____constructor__PF_Rd0d0____destructor__PF_Rd0____operator_assign__PFd1_Rd1d1____constructor__PF_Rd1____constructor__PF_Rd1d1____destructor__PF_Rd1____operator_preincr__PFd0_Rd0____operator_predecr__PFd0_Rd0____operator_equal__PFi_d0d0____operator_notequal__PFi_d0d0____operator_deref__PFRd1_d0__F_d0d0PF_d1___1(__attribute__ ((unused)) void (*_adapterF_9telt_type__P)(void (*__anonymous_object0)(), void *__anonymous_object1), __attribute__ ((unused)) void *(*_adapterFP9telt_type_14titerator_type_M_P)(void (*__anonymous_object2)(), void *__anonymous_object3), __attribute__ ((unused)) signed int (*_adapterFi_14titerator_type14titerator_type_M_PP)(void (*__anonymous_object4)(), void *__anonymous_object5, void *__anonymous_object6), __attribute__ ((unused)) void (*_adapterF14titerator_type_P14titerator_type_P_M)(void (*__anonymous_object7)(), __attribute__ ((unused)) void *___retval__operator_preincr__14titerator_type_1, void *__anonymous_object8), __attribute__ ((unused)) void (*_adapterF_P9telt_type9telt_type__MP)(void (*__anonymous_object9)(), void *__anonymous_object10, void *__anonymous_object11), __attribute__ ((unused)) void (*_adapterF9telt_type_P9telt_type9telt_type_P_MP)(void (*__anonymous_object12)(), __attribute__ ((unused)) void *___retval__operator_assign__9telt_type_1, void *__anonymous_object13, void *__anonymous_object14), __attribute__ ((unused)) void (*_adapterF_P14titerator_type14titerator_type__MP)(void (*__anonymous_object15)(), void *__anonymous_object16, void *__anonymous_object17), __attribute__ ((unused)) void (*_adapterF14titerator_type_P14titerator_type14titerator_type_P_MP)(void (*__anonymous_object18)(), __attribute__ ((unused)) void *___retval__operator_assign__14titerator_type_1, void *__anonymous_object19, void *__anonymous_object20), __attribute__ ((unused)) unsigned long int _sizeof_14titerator_type, __attribute__ ((unused)) unsigned long int _alignof_14titerator_type, __attribute__ ((unused)) unsigned long int _sizeof_9telt_type, __attribute__ ((unused)) unsigned long int _alignof_9telt_type, __attribute__ ((unused)) void *(*___operator_assign__PF14titerator_type_R14titerator_type14titerator_type__1)(void *__anonymous_object21, void *__anonymous_object22), __attribute__ ((unused)) void (*___constructor__PF_R14titerator_type__1)(void *__anonymous_object23), __attribute__ ((unused)) void (*___constructor__PF_R14titerator_type14titerator_type__1)(void *__anonymous_object24, void *__anonymous_object25), __attribute__ ((unused)) void (*___destructor__PF_R14titerator_type__1)(void *__anonymous_object26), __attribute__ ((unused)) void *(*___operator_assign__PF9telt_type_R9telt_type9telt_type__1)(void *__anonymous_object27, void *__anonymous_object28), __attribute__ ((unused)) void (*___constructor__PF_R9telt_type__1)(void *__anonymous_object29), __attribute__ ((unused)) void (*___constructor__PF_R9telt_type9telt_type__1)(void *__anonymous_object30, void *__anonymous_object31), __attribute__ ((unused)) void (*___destructor__PF_R9telt_type__1)(void *__anonymous_object32), __attribute__ ((unused)) void *(*___operator_preincr__PF14titerator_type_R14titerator_type__1)(void *__anonymous_object33), __attribute__ ((unused)) void *(*___operator_predecr__PF14titerator_type_R14titerator_type__1)(void *__anonymous_object34), __attribute__ ((unused)) signed int (*___operator_equal__PFi_14titerator_type14titerator_type__1)(void *__anonymous_object35, void *__anonymous_object36), __attribute__ ((unused)) signed int (*___operator_notequal__PFi_14titerator_type14titerator_type__1)(void *__anonymous_object37, void *__anonymous_object38), __attribute__ ((unused)) void *(*___operator_deref__PFR9telt_type_14titerator_type__1)(void *__anonymous_object39), void *__begin__14titerator_type_1, void *__end__14titerator_type_1, void (*__func__PF_9telt_type__1)(void *__anonymous_object40));
 void __for_each_reverse__A0_2_0_0____operator_assign__PFd0_Rd0d0____constructor__PF_Rd0____constructor__PF_Rd0d0____destructor__PF_Rd0____operator_assign__PFd1_Rd1d1____constructor__PF_Rd1____constructor__PF_Rd1d1____destructor__PF_Rd1____operator_preincr__PFd0_Rd0____operator_predecr__PFd0_Rd0____operator_equal__PFi_d0d0____operator_notequal__PFi_d0d0____operator_deref__PFRd1_d0__F_d0d0PF_d1___1(__attribute__ ((unused)) void (*_adapterF_9telt_type__P)(void (*__anonymous_object41)(), void *__anonymous_object42), __attribute__ ((unused)) void *(*_adapterFP9telt_type_14titerator_type_M_P)(void (*__anonymous_object43)(), void *__anonymous_object44), __attribute__ ((unused)) signed int (*_adapterFi_14titerator_type14titerator_type_M_PP)(void (*__anonymous_object45)(), void *__anonymous_object46, void *__anonymous_object47), __attribute__ ((unused)) void (*_adapterF14titerator_type_P14titerator_type_P_M)(void (*__anonymous_object48)(), __attribute__ ((unused)) void *___retval__operator_preincr__14titerator_type_1, void *__anonymous_object49), __attribute__ ((unused)) void (*_adapterF_P9telt_type9telt_type__MP)(void (*__anonymous_object50)(), void *__anonymous_object51, void *__anonymous_object52), __attribute__ ((unused)) void (*_adapterF9telt_type_P9telt_type9telt_type_P_MP)(void (*__anonymous_object53)(), __attribute__ ((unused)) void *___retval__operator_assign__9telt_type_1, void *__anonymous_object54, void *__anonymous_object55), __attribute__ ((unused)) void (*_adapterF_P14titerator_type14titerator_type__MP)(void (*__anonymous_object56)(), void *__anonymous_object57, void *__anonymous_object58), __attribute__ ((unused)) void (*_adapterF14titerator_type_P14titerator_type14titerator_type_P_MP)(void (*__anonymous_object59)(), __attribute__ ((unused)) void *___retval__operator_assign__14titerator_type_1, void *__anonymous_object60, void *__anonymous_object61), __attribute__ ((unused)) unsigned long int _sizeof_14titerator_type, __attribute__ ((unused)) unsigned long int _alignof_14titerator_type, __attribute__ ((unused)) unsigned long int _sizeof_9telt_type, __attribute__ ((unused)) unsigned long int _alignof_9telt_type, __attribute__ ((unused)) void *(*___operator_assign__PF14titerator_type_R14titerator_type14titerator_type__1)(void *__anonymous_object62, void *__anonymous_object63), __attribute__ ((unused)) void (*___constructor__PF_R14titerator_type__1)(void *__anonymous_object64), __attribute__ ((unused)) void (*___constructor__PF_R14titerator_type14titerator_type__1)(void *__anonymous_object65, void *__anonymous_object66), __attribute__ ((unused)) void (*___destructor__PF_R14titerator_type__1)(void *__anonymous_object67), __attribute__ ((unused)) void *(*___operator_assign__PF9telt_type_R9telt_type9telt_type__1)(void *__anonymous_object68, void *__anonymous_object69), __attribute__ ((unused)) void (*___constructor__PF_R9telt_type__1)(void *__anonymous_object70), __attribute__ ((unused)) void (*___constructor__PF_R9telt_type9telt_type__1)(void *__anonymous_object71, void *__anonymous_object72), __attribute__ ((unused)) void (*___destructor__PF_R9telt_type__1)(void *__anonymous_object73), __attribute__ ((unused)) void *(*___operator_preincr__PF14titerator_type_R14titerator_type__1)(void *__anonymous_object74), __attribute__ ((unused)) void *(*___operator_predecr__PF14titerator_type_R14titerator_type__1)(void *__anonymous_object75), __attribute__ ((unused)) signed int (*___operator_equal__PFi_14titerator_type14titerator_type__1)(void *__anonymous_object76, void *__anonymous_object77), __attribute__ ((unused)) signed int (*___operator_notequal__PFi_14titerator_type14titerator_type__1)(void *__anonymous_object78, void *__anonymous_object79), __attribute__ ((unused)) void *(*___operator_deref__PFR9telt_type_14titerator_type__1)(void *__anonymous_object80), void *__begin__14titerator_type_1, void *__end__14titerator_type_1, void (*__func__PF_9telt_type__1)(void *__anonymous_object81));
 …
+}
 static inline int invoke_main(int argc, char* argv[], char* envp[]) { (void)argc; (void)argv; (void)envp; return __main__Fi___1(); }
+__attribute__ ((__nothrow__,__leaf__,__malloc__)) extern void *malloc(unsigned int __size);
+__attribute__ ((__nothrow__,__leaf__)) extern void free(void *__ptr);
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void abort(void);
+__attribute__ ((__nothrow__,__leaf__,__nonnull__(1))) extern signed int atexit(void (*__func)(void));
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void exit(signed int __status);
+extern signed int printf(const char *__restrict __format, ...);
 static inline signed int invoke_main(signed int argc, char **argv, char **envp);
 signed int main(signed int __argc__i_1, char **__argv__PPc_1, char **__envp__PPc_1){

src/tests/.expect/64/KRfunctions.txt

-              r0fe4e62
+              rf5c3b6c
+__attribute__ ((__nothrow__,__leaf__,__malloc__)) extern void *malloc(unsigned long int __size);
+__attribute__ ((__nothrow__,__leaf__)) extern void free(void *__ptr);
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void abort(void);
+__attribute__ ((__nothrow__,__leaf__,__nonnull__(1))) extern signed int atexit(void (*__func)(void));
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void exit(signed int __status);
+extern signed int printf(const char *__restrict __format, ...);
 signed int __f0__Fi_iPCii__1(signed int __a__i_1, const signed int *__b__PCi_1, signed int __c__i_1){
     __attribute__ ((unused)) signed int ___retval_f0__i_1;

src/tests/.expect/64/attributes.txt

-              r0fe4e62
+              rf5c3b6c
+__attribute__ ((__nothrow__,__leaf__,__malloc__)) extern void *malloc(unsigned long int __size);
+__attribute__ ((__nothrow__,__leaf__)) extern void free(void *__ptr);
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void abort(void);
+__attribute__ ((__nothrow__,__leaf__,__nonnull__(1))) extern signed int atexit(void (*__func)(void));
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void exit(signed int __status);
+extern signed int printf(const char *__restrict __format, ...);
 signed int __la__Fi___1(){
     __attribute__ ((unused)) signed int ___retval_la__i_1;

src/tests/.expect/64/declarationSpecifier.txt

-              r0fe4e62
+              rf5c3b6c
+__attribute__ ((__nothrow__,__leaf__,__malloc__)) extern void *malloc(unsigned long int __size);
+__attribute__ ((__nothrow__,__leaf__)) extern void free(void *__ptr);
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void abort(void);
+__attribute__ ((__nothrow__,__leaf__,__nonnull__(1))) extern signed int atexit(void (*__func)(void));
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void exit(signed int __status);
+extern signed int printf(const char *__restrict __format, ...);
 volatile const signed short int __x1__CVs_1;
 static volatile const signed short int __x2__CVs_1;
 …
+}
 static inline int invoke_main(int argc, char* argv[], char* envp[]) { (void)argc; (void)argv; (void)envp; return __main__Fi_iPPCc__1(argc, argv); }
+__attribute__ ((__nothrow__,__leaf__,__malloc__)) extern void *malloc(unsigned long int __size);
+__attribute__ ((__nothrow__,__leaf__)) extern void free(void *__ptr);
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void abort(void);
+__attribute__ ((__nothrow__,__leaf__,__nonnull__(1))) extern signed int atexit(void (*__func)(void));
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void exit(signed int __status);
+extern signed int printf(const char *__restrict __format, ...);
 static inline signed int invoke_main(signed int argc, char **argv, char **envp);
 signed int main(signed int __argc__i_1, char **__argv__PPc_1, char **__envp__PPc_1){

src/tests/.expect/64/extension.txt

-              r0fe4e62
+              rf5c3b6c
+__attribute__ ((__nothrow__,__leaf__,__malloc__)) extern void *malloc(unsigned long int __size);
+__attribute__ ((__nothrow__,__leaf__)) extern void free(void *__ptr);
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void abort(void);
+__attribute__ ((__nothrow__,__leaf__,__nonnull__(1))) extern signed int atexit(void (*__func)(void));
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void exit(signed int __status);
+extern signed int printf(const char *__restrict __format, ...);
 __extension__ signed int __a__i_1;
 __extension__ signed int __b__i_1;

src/tests/.expect/64/gccExtensions.txt

-              r0fe4e62
+              rf5c3b6c
+__attribute__ ((__nothrow__,__leaf__,__malloc__)) extern void *malloc(unsigned long int __size);
+__attribute__ ((__nothrow__,__leaf__)) extern void free(void *__ptr);
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void abort(void);
+__attribute__ ((__nothrow__,__leaf__,__nonnull__(1))) extern signed int atexit(void (*__func)(void));
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void exit(signed int __status);
+extern signed int printf(const char *__restrict __format, ...);
 extern signed int __x__i_1 asm ( "xx" );
 signed int __main__Fi_iPPCc__1(signed int __argc__i_1, const char **__argv__PPCc_1){
 …
+}
 static inline int invoke_main(int argc, char* argv[], char* envp[]) { (void)argc; (void)argv; (void)envp; return __main__Fi_iPPCc__1(argc, argv); }
+__attribute__ ((__nothrow__,__leaf__,__malloc__)) extern void *malloc(unsigned long int __size);
+__attribute__ ((__nothrow__,__leaf__)) extern void free(void *__ptr);
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void abort(void);
+__attribute__ ((__nothrow__,__leaf__,__nonnull__(1))) extern signed int atexit(void (*__func)(void));
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void exit(signed int __status);
+extern signed int printf(const char *__restrict __format, ...);
 static inline signed int invoke_main(signed int argc, char **argv, char **envp);
 signed int main(signed int __argc__i_1, char **__argv__PPc_1, char **__envp__PPc_1){

src/tests/.expect/64/literals.txt

-              r0fe4e62
+              rf5c3b6c
+__attribute__ ((__nothrow__,__leaf__,__malloc__)) extern void *malloc(unsigned long int __size);
+__attribute__ ((__nothrow__,__leaf__)) extern void free(void *__ptr);
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void abort(void);
+__attribute__ ((__nothrow__,__leaf__,__nonnull__(1))) extern signed int atexit(void (*__func)(void));
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void exit(signed int __status);
+extern signed int printf(const char *__restrict __format, ...);
 void __for_each__A0_2_0_0____operator_assign__PFd0_Rd0d0____constructor__PF_Rd0____constructor__PF_Rd0d0____destructor__PF_Rd0____operator_assign__PFd1_Rd1d1____constructor__PF_Rd1____constructor__PF_Rd1d1____destructor__PF_Rd1____operator_preincr__PFd0_Rd0____operator_predecr__PFd0_Rd0____operator_equal__PFi_d0d0____operator_notequal__PFi_d0d0____operator_deref__PFRd1_d0__F_d0d0PF_d1___1(__attribute__ ((unused)) void (*_adapterF_9telt_type__P)(void (*__anonymous_object0)(), void *__anonymous_object1), __attribute__ ((unused)) void *(*_adapterFP9telt_type_14titerator_type_M_P)(void (*__anonymous_object2)(), void *__anonymous_object3), __attribute__ ((unused)) signed int (*_adapterFi_14titerator_type14titerator_type_M_PP)(void (*__anonymous_object4)(), void *__anonymous_object5, void *__anonymous_object6), __attribute__ ((unused)) void (*_adapterF14titerator_type_P14titerator_type_P_M)(void (*__anonymous_object7)(), __attribute__ ((unused)) void *___retval__operator_preincr__14titerator_type_1, void *__anonymous_object8), __attribute__ ((unused)) void (*_adapterF_P9telt_type9telt_type__MP)(void (*__anonymous_object9)(), void *__anonymous_object10, void *__anonymous_object11), __attribute__ ((unused)) void (*_adapterF9telt_type_P9telt_type9telt_type_P_MP)(void (*__anonymous_object12)(), __attribute__ ((unused)) void *___retval__operator_assign__9telt_type_1, void *__anonymous_object13, void *__anonymous_object14), __attribute__ ((unused)) void (*_adapterF_P14titerator_type14titerator_type__MP)(void (*__anonymous_object15)(), void *__anonymous_object16, void *__anonymous_object17), __attribute__ ((unused)) void (*_adapterF14titerator_type_P14titerator_type14titerator_type_P_MP)(void (*__anonymous_object18)(), __attribute__ ((unused)) void *___retval__operator_assign__14titerator_type_1, void *__anonymous_object19, void *__anonymous_object20), __attribute__ ((unused)) unsigned long int _sizeof_14titerator_type, __attribute__ ((unused)) unsigned long int _alignof_14titerator_type, __attribute__ ((unused)) unsigned long int _sizeof_9telt_type, __attribute__ ((unused)) unsigned long int _alignof_9telt_type, __attribute__ ((unused)) void *(*___operator_assign__PF14titerator_type_R14titerator_type14titerator_type__1)(void *__anonymous_object21, void *__anonymous_object22), __attribute__ ((unused)) void (*___constructor__PF_R14titerator_type__1)(void *__anonymous_object23), __attribute__ ((unused)) void (*___constructor__PF_R14titerator_type14titerator_type__1)(void *__anonymous_object24, void *__anonymous_object25), __attribute__ ((unused)) void (*___destructor__PF_R14titerator_type__1)(void *__anonymous_object26), __attribute__ ((unused)) void *(*___operator_assign__PF9telt_type_R9telt_type9telt_type__1)(void *__anonymous_object27, void *__anonymous_object28), __attribute__ ((unused)) void (*___constructor__PF_R9telt_type__1)(void *__anonymous_object29), __attribute__ ((unused)) void (*___constructor__PF_R9telt_type9telt_type__1)(void *__anonymous_object30, void *__anonymous_object31), __attribute__ ((unused)) void (*___destructor__PF_R9telt_type__1)(void *__anonymous_object32), __attribute__ ((unused)) void *(*___operator_preincr__PF14titerator_type_R14titerator_type__1)(void *__anonymous_object33), __attribute__ ((unused)) void *(*___operator_predecr__PF14titerator_type_R14titerator_type__1)(void *__anonymous_object34), __attribute__ ((unused)) signed int (*___operator_equal__PFi_14titerator_type14titerator_type__1)(void *__anonymous_object35, void *__anonymous_object36), __attribute__ ((unused)) signed int (*___operator_notequal__PFi_14titerator_type14titerator_type__1)(void *__anonymous_object37, void *__anonymous_object38), __attribute__ ((unused)) void *(*___operator_deref__PFR9telt_type_14titerator_type__1)(void *__anonymous_object39), void *__begin__14titerator_type_1, void *__end__14titerator_type_1, void (*__func__PF_9telt_type__1)(void *__anonymous_object40));
 void __for_each_reverse__A0_2_0_0____operator_assign__PFd0_Rd0d0____constructor__PF_Rd0____constructor__PF_Rd0d0____destructor__PF_Rd0____operator_assign__PFd1_Rd1d1____constructor__PF_Rd1____constructor__PF_Rd1d1____destructor__PF_Rd1____operator_preincr__PFd0_Rd0____operator_predecr__PFd0_Rd0____operator_equal__PFi_d0d0____operator_notequal__PFi_d0d0____operator_deref__PFRd1_d0__F_d0d0PF_d1___1(__attribute__ ((unused)) void (*_adapterF_9telt_type__P)(void (*__anonymous_object41)(), void *__anonymous_object42), __attribute__ ((unused)) void *(*_adapterFP9telt_type_14titerator_type_M_P)(void (*__anonymous_object43)(), void *__anonymous_object44), __attribute__ ((unused)) signed int (*_adapterFi_14titerator_type14titerator_type_M_PP)(void (*__anonymous_object45)(), void *__anonymous_object46, void *__anonymous_object47), __attribute__ ((unused)) void (*_adapterF14titerator_type_P14titerator_type_P_M)(void (*__anonymous_object48)(), __attribute__ ((unused)) void *___retval__operator_preincr__14titerator_type_1, void *__anonymous_object49), __attribute__ ((unused)) void (*_adapterF_P9telt_type9telt_type__MP)(void (*__anonymous_object50)(), void *__anonymous_object51, void *__anonymous_object52), __attribute__ ((unused)) void (*_adapterF9telt_type_P9telt_type9telt_type_P_MP)(void (*__anonymous_object53)(), __attribute__ ((unused)) void *___retval__operator_assign__9telt_type_1, void *__anonymous_object54, void *__anonymous_object55), __attribute__ ((unused)) void (*_adapterF_P14titerator_type14titerator_type__MP)(void (*__anonymous_object56)(), void *__anonymous_object57, void *__anonymous_object58), __attribute__ ((unused)) void (*_adapterF14titerator_type_P14titerator_type14titerator_type_P_MP)(void (*__anonymous_object59)(), __attribute__ ((unused)) void *___retval__operator_assign__14titerator_type_1, void *__anonymous_object60, void *__anonymous_object61), __attribute__ ((unused)) unsigned long int _sizeof_14titerator_type, __attribute__ ((unused)) unsigned long int _alignof_14titerator_type, __attribute__ ((unused)) unsigned long int _sizeof_9telt_type, __attribute__ ((unused)) unsigned long int _alignof_9telt_type, __attribute__ ((unused)) void *(*___operator_assign__PF14titerator_type_R14titerator_type14titerator_type__1)(void *__anonymous_object62, void *__anonymous_object63), __attribute__ ((unused)) void (*___constructor__PF_R14titerator_type__1)(void *__anonymous_object64), __attribute__ ((unused)) void (*___constructor__PF_R14titerator_type14titerator_type__1)(void *__anonymous_object65, void *__anonymous_object66), __attribute__ ((unused)) void (*___destructor__PF_R14titerator_type__1)(void *__anonymous_object67), __attribute__ ((unused)) void *(*___operator_assign__PF9telt_type_R9telt_type9telt_type__1)(void *__anonymous_object68, void *__anonymous_object69), __attribute__ ((unused)) void (*___constructor__PF_R9telt_type__1)(void *__anonymous_object70), __attribute__ ((unused)) void (*___constructor__PF_R9telt_type9telt_type__1)(void *__anonymous_object71, void *__anonymous_object72), __attribute__ ((unused)) void (*___destructor__PF_R9telt_type__1)(void *__anonymous_object73), __attribute__ ((unused)) void *(*___operator_preincr__PF14titerator_type_R14titerator_type__1)(void *__anonymous_object74), __attribute__ ((unused)) void *(*___operator_predecr__PF14titerator_type_R14titerator_type__1)(void *__anonymous_object75), __attribute__ ((unused)) signed int (*___operator_equal__PFi_14titerator_type14titerator_type__1)(void *__anonymous_object76, void *__anonymous_object77), __attribute__ ((unused)) signed int (*___operator_notequal__PFi_14titerator_type14titerator_type__1)(void *__anonymous_object78, void *__anonymous_object79), __attribute__ ((unused)) void *(*___operator_deref__PFR9telt_type_14titerator_type__1)(void *__anonymous_object80), void *__begin__14titerator_type_1, void *__end__14titerator_type_1, void (*__func__PF_9telt_type__1)(void *__anonymous_object81));
 …
+}
 static inline int invoke_main(int argc, char* argv[], char* envp[]) { (void)argc; (void)argv; (void)envp; return __main__Fi___1(); }
+__attribute__ ((__nothrow__,__leaf__,__malloc__)) extern void *malloc(unsigned long int __size);
+__attribute__ ((__nothrow__,__leaf__)) extern void free(void *__ptr);
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void abort(void);
+__attribute__ ((__nothrow__,__leaf__,__nonnull__(1))) extern signed int atexit(void (*__func)(void));
+__attribute__ ((__nothrow__,__leaf__,__noreturn__)) extern void exit(signed int __status);
+extern signed int printf(const char *__restrict __format, ...);
 static inline signed int invoke_main(signed int argc, char **argv, char **envp);
 signed int main(signed int __argc__i_1, char **__argv__PPc_1, char **__envp__PPc_1){

src/tests/boundedBuffer.c

-              r0fe4e62
+              rf5c3b6c
 //
+//
 // The contents of this file are covered under the licence agreement in the
 // file "LICENCE" distributed with Cforall.
 //
 // boundedBuffer.c --
 //
+//
+// boundedBuffer.c --
+//
 // Author           : Peter A. Buhr
 // Created On       : Mon Oct 30 12:45:13 2017
 …
 // Last Modified On : Mon Oct 30 23:02:46 2017
 // Update Count     : 9
 //
+//
 #include <stdlib>
 …
 void insert( Buffer & mutex buffer, int elem ) {
         if ( buffer.count == 20 ) wait( buffer.empty );
+        if ( buffer.count == 20 ) wait( &buffer.empty );
         buffer.elements[buffer.back] = elem;
         buffer.back = ( buffer.back + 1 ) % 20;
         buffer.count += 1;
         signal( buffer.full );
+        signal( &buffer.full );
+}
 int remove( Buffer & mutex buffer ) {
         if ( buffer.count == 0 ) wait( buffer.full );
+        if ( buffer.count == 0 ) wait( &buffer.full );
         int elem = buffer.elements[buffer.front];
         buffer.front = ( buffer.front + 1 ) % 20;
         buffer.count -= 1;
         signal( buffer.empty );
+        signal( &buffer.empty );
         return elem;
+}

src/tests/datingService.c

-              r0fe4e62
+              rf5c3b6c
 //                               -*- Mode: C -*-
 //
+//                               -*- Mode: C -*-
+//
 // The contents of this file are covered under the licence agreement in the
 // file "LICENCE" distributed with Cforall.
 //
 // datingService.c --
 //
+//
+// datingService.c --
+//
 // Author           : Peter A. Buhr
 // Created On       : Mon Oct 30 12:56:20 2017
 …
 // Last Modified On : Mon Oct 30 23:02:11 2017
 // Update Count     : 15
 //
+//
 #include <stdlib>                                                                               // random
 …
 #include <thread>
 #include <unistd.h>                                                                             // getpid
+bool empty( condition & c ) {
+        return c.blocked.head == NULL;
+}
 enum { NoOfPairs = 20 };
 …
 unsigned int girl( DatingService & mutex ds, unsigned int PhoneNo, unsigned int ccode ) {
         if ( is_empty( ds.Boys[ccode] ) ) {
                 wait( ds.Girls[ccode] );
+        if ( empty( ds.Boys[ccode] ) ) {
+                wait( &ds.Girls[ccode] );
                 ds.GirlPhoneNo = PhoneNo;
         } else {
                 ds.GirlPhoneNo = PhoneNo;
                 signal_block( ds.Boys[ccode] );
+                signal_block( &ds.Boys[ccode] );
         } // if
         return ds.BoyPhoneNo;
 …
 unsigned int boy( DatingService & mutex ds, unsigned int PhoneNo, unsigned int ccode ) {
         if ( is_empty( ds.Girls[ccode] ) ) {
                 wait( ds.Boys[ccode] );
+        if ( empty( ds.Girls[ccode] ) ) {
+                wait( &ds.Boys[ccode] );
                 ds.BoyPhoneNo = PhoneNo;
         } else {
                 ds.BoyPhoneNo = PhoneNo;
                 signal_block( ds.Girls[ccode] );
+                signal_block( &ds.Girls[ccode] );
         } // if
         return ds.GirlPhoneNo;

src/tests/sched-int-barge.c

-              r0fe4e62
+              rf5c3b6c
         if( action == c.do_wait1 || action == c.do_wait2 ) {
                 c.state = WAIT;
                 wait( cond );
+                wait( &cond );
                 if(c.state != SIGNAL) {
 …
                 c.state = SIGNAL;
                 signal( cond );
                 signal( cond );
+                signal( &cond );
+                signal( &cond );
+        }
         else {

src/tests/sched-int-block.c

-              r0fe4e62
+              rf5c3b6c
 //------------------------------------------------------------------------------
 void wait_op( global_data_t & mutex a, global_data_t & mutex b, unsigned i ) {
         wait( cond, (uintptr_t)this_thread );
+        wait( &cond, (uintptr_t)this_thread );
         yield( random( 10 ) );
 …
         [a.last_thread, b.last_thread, a.last_signaller, b.last_signaller] = this_thread;
         if( !is_empty( cond ) ) {
+        if( !is_empty( &cond ) ) {
                 thread_desc * next = front( cond );
+                thread_desc * next = front( &cond );
                 if( ! signal_block( cond ) ) {
+                if( ! signal_block( &cond ) ) {
                         sout | "ERROR expected to be able to signal" | endl;
                         abort();

src/tests/sched-int-disjoint.c

-              r0fe4e62
+              rf5c3b6c
 // Waiting logic
 bool wait( global_t & mutex m, global_data_t & mutex d ) {
         wait( cond );
+        wait( &cond );
         if( d.state != SIGNAL ) {
                 sout | "ERROR barging!" | endl;
 …
 //------------------------------------------------------------------------------
 // Signalling logic
 void signal( condition & cond, global_t & mutex a, global_data_t & mutex b ) {
+void signal( condition * cond, global_t & mutex a, global_data_t & mutex b ) {
         b.state = SIGNAL;
         signal( cond );
 …
 void logic( global_t & mutex a ) {
         signal( cond, a, data );
+        signal( &cond, a, data );
         yield( random( 10 ) );

src/tests/sched-int-wait.c

-              r0fe4e62
+              rf5c3b6c
 //----------------------------------------------------------------------------------------------------
 // Tools
 void signal( condition & cond, global_t & mutex a, global_t & mutex b ) {
+void signal( condition * cond, global_t & mutex a, global_t & mutex b ) {
         signal( cond );
+}
 void signal( condition & cond, global_t & mutex a, global_t & mutex b, global_t & mutex c ) {
+void signal( condition * cond, global_t & mutex a, global_t & mutex b, global_t & mutex c ) {
         signal( cond );
+}
 void wait( condition & cond, global_t & mutex a, global_t & mutex b ) {
+void wait( condition * cond, global_t & mutex a, global_t & mutex b ) {
         wait( cond );
+}
 void wait( condition & cond, global_t & mutex a, global_t & mutex b, global_t & mutex c ) {
+void wait( condition * cond, global_t & mutex a, global_t & mutex b, global_t & mutex c ) {
         wait( cond );
+}
 …
                 switch( action ) {
                         case 0:
                                 signal( condABC, globalA, globalB, globalC );
+                                signal( &condABC, globalA, globalB, globalC );
                                 break;
                         case 1:
                                 signal( condAB , globalA, globalB );
+                                signal( &condAB , globalA, globalB );
                                 break;
                         case 2:
                                 signal( condBC , globalB, globalC );
+                                signal( &condBC , globalB, globalC );
                                 break;
                         case 3:
                                 signal( condAC , globalA, globalC );
+                                signal( &condAC , globalA, globalC );
                                 break;
                         default:
 …
 void main( WaiterABC & this ) {
         for( int i = 0; i < N; i++ ) {
                 wait( condABC, globalA, globalB, globalC );
+                wait( &condABC, globalA, globalB, globalC );
+        }
 …
 void main( WaiterAB & this ) {
         for( int i = 0; i < N; i++ ) {
                 wait( condAB , globalA, globalB );
+                wait( &condAB , globalA, globalB );
+        }
 …
 void main( WaiterAC & this ) {
         for( int i = 0; i < N; i++ ) {
                 wait( condAC , globalA, globalC );
+                wait( &condAC , globalA, globalC );
+        }
 …
 void main( WaiterBC & this ) {
         for( int i = 0; i < N; i++ ) {
                 wait( condBC , globalB, globalC );
+                wait( &condBC , globalB, globalC );
+        }

src/tests/thread.c

-              r0fe4e62
+              rf5c3b6c
                 yield();
+        }
         V(*this.lock);
+        V(this.lock);
+}
 void main(Second& this) {
         P(*this.lock);
+        P(this.lock);
         for(int i = 0; i < 10; i++) {
                 sout | "Second : Suspend No." | i + 1 | endl;

Context Navigation

Changes in / [0fe4e62:f5c3b6c]

Legend:

Download in other formats: