URL
https://opencores.org/ocsvn/openrisc/openrisc/trunk
Subversion Repositories openrisc
[/] [openrisc/] [trunk/] [gnu-dev/] [or1k-gcc/] [boehm-gc/] [doc/] [simple_example.html] - Rev 757
Go to most recent revision | Compare with Previous | Blame | View Log
<HTML> <HEAD> <TITLE>Using the Garbage Collector: A simple example</title> </head> <BODY> <H1>Using the Garbage Collector: A simple example</h1> The following consists of step-by-step instructions for building and using the collector. We'll assume a Linux/gcc platform and a single-threaded application. <FONT COLOR=green>The green text contains information about other platforms or scenarios. It can be skipped, especially on first reading</font>. <H2>Building the collector</h2> If you haven't already so, unpack the collector and enter the newly created directory with <PRE> tar xvfz gc<version>.tar.gz cd gc<version> </pre> <P> You can configure, build, and install the collector in a private directory, say /home/xyz/gc, with the following commands: <PRE> ./configure --prefix=/home/xyz/gc --disable-threads make make check make install </pre> Here the "<TT>make check</tt>" command is optional, but highly recommended. It runs a basic correctness test which usually takes well under a minute. <FONT COLOR=green> <H3>Other platforms</h3> On non-Unix, non-Linux platforms, the collector is usually built by copying the appropriate makefile (see the platform-specific README in doc/README.xxx in the distribution) to the file "Makefile" (overwriting the copy of Makefile.direct that was originally there), and then typing "make" (or "nmake" or ...). This builds the library in the source tree. You may want to move it and the files in the include directory to a more convenient place. <P> If you use a makefile that does not require running a configure script, you should first look at the makefile, and adjust any options that are documented there. <P> If your platform provides a "make" utility, that is generally preferred to platform- and compiler- dependent "project" files. (At least that is the strong preference of the would-be maintainer of those project files.) <H3>Threads</h3> If you need thread support, configure the collector with <PRE> --enable-threads=posix --enable-thread-local-alloc --enable-parallel-mark </pre> instead of <TT>--disable-threads</tt> If your target is a real old-fashioned uniprocessor (no "hyperthreading", etc.) you will want to omit <TT>--enable-parallel-mark</tt>. <H3>C++</h3> You will need to include the C++ support, which unfortunately tends to be among the least portable parts of the collector, since it seems to rely on some corner cases of the language. On Linux, it suffices to add <TT>--enable-cplusplus</tt> to the configure options. </font> <H2>Writing the program</h2> You will need a <PRE> #include "gc.h" </pre> at the beginning of every file that allocates memory through the garbage collector. Call <TT>GC_MALLOC</tt> wherever you would have call <TT>malloc</tt>. This initializes memory to zero like <TT>calloc</tt>; there is no need to explicitly clear the result. <P> If you know that an object will not contain pointers to the garbage-collected heap, and you don't need it to be initialized, call <TT>GC_MALLOC_ATOMIC</tt> instead. <P> A function <TT>GC_FREE</tt> is provided but need not be called. For very small objects, your program will probably perform better if you do not call it, and let the collector do its job. <P> A <TT>GC_REALLOC</tt> function behaves like the C library <TT>realloc</tt>. It allocates uninitialized pointer-free memory if the original object was allocated that way. <P> The following program <TT>loop.c</tt> is a trivial example: <PRE> #include "gc.h" #include <assert.h> #include <stdio.h> int main() { int i; GC_INIT(); /* Optional on Linux/X86; see below. */ for (i = 0; i < 10000000; ++i) { int **p = (int **) GC_MALLOC(sizeof(int *)); int *q = (int *) GC_MALLOC_ATOMIC(sizeof(int)); assert(*p == 0); *p = (int *) GC_REALLOC(q, 2 * sizeof(int)); if (i % 100000 == 0) printf("Heap size = %d\n", GC_get_heap_size()); } return 0; } </pre> <FONT COLOR=green> <H3>Interaction with the system malloc</h3> It is usually best not to mix garbage-collected allocation with the system <TT>malloc-free</tt>. If you do, you need to be careful not to store pointers to the garbage-collected heap in memory allocated with the system <TT>malloc</tt>. <H3>Other Platforms</h3> On some other platforms it is necessary to call <TT>GC_INIT()</tt> from the main program, which is presumed to be part of the main executable, not a dynamic library. This can never hurt, and is thus generally good practice. <H3>Threads</h3> For a multithreaded program some more rules apply: <UL> <LI> Files that either allocate through the GC <I>or make thread-related calls</i> should first define the macro <TT>GC_THREADS</tt>, and then include <TT>"gc.h"</tt>. On some platforms this will redefine some threads primitives, e.g. to let the collector keep track of thread creation. <LI> To take advantage of fast thread-local allocation, use the following instead of including <TT>gc.h</tt>: <PRE> #define GC_REDIRECT_TO_LOCAL #include "gc_local_alloc.h" </pre> This will cause GC_MALLOC and GC_MALLOC_ATOMIC to keep per-thread allocation caches, and greatly reduce the number of lock acquisitions during allocation. </ul> <H3>C++</h3> In the case of C++, you need to be especially careful not to store pointers to the garbage-collected heap in areas that are not traced by the collector. The collector includes some <A HREF="gcinterface.html">alternate interfaces</a> to make that easier. <H3>Debugging</h3> Additional debug checks can be performed by defining <TT>GC_DEBUG</tt> before including <TT>gc.h</tt>. Additional options are available if the collector is also built with <TT>--enable-full_debug</tt> and all allocations are performed with <TT>GC_DEBUG</tt> defined. <H3>What if I can't rewrite/recompile my program?</h3> You may be able to build the collector with <TT>--enable-redirect-malloc</tt> and set the <TT>LD_PRELOAD</tt> environment variable to point to the resulting library, thus replacing the standard <TT>malloc</tt> with its garbage-collected counterpart. This is rather platform dependent. See the <A HREF="leak.html">leak detection documentation</a> for some more details. </font> <H2>Compiling and linking</h2> The above application <TT>loop.c</tt> test program can be compiled and linked with <PRE> cc -I/home/xyz/gc/include loop.c /home/xyz/gc/lib/libgc.a -o loop </pre> The <TT>-I</tt> option directs the compiler to the right include directory. In this case, we list the static library directly on the compile line; the dynamic library could have been used instead, provided we arranged for the dynamic loader to find it, e.g. by setting <TT>LD_LIBRARY_PATH</tt>. <FONT COLOR=green> <H3>Threads</h3> On pthread platforms, you will of course also have to link with <TT>-lpthread</tt>, and compile with any thread-safety options required by your compiler. On some platforms, you may also need to link with <TT>-ldl</tt> or <TT>-lrt</tt>. Looking at threadlibs.c in the GC build directory should give you the appropriate list if a plain <TT>-lpthread</tt> doesn't work. </font> <H2>Running the executable</h2> The executable can of course be run normally, e.g. by typing <PRE> ./loop </pre> The operation of the collector is affected by a number of environment variables. For example, setting <TT>GC_PRINT_STATS</tt> produces some GC statistics on stdout. See <TT>README.environment</tt> in the distribution for details. </body> </html>
Go to most recent revision | Compare with Previous | Blame | View Log