1 #LyX 1.2 created this file. For more info see http://www.lyx.org/
15 \use_numerical_citations 0
16 \paperorientation portrait
19 \paragraph_separation indent
21 \quotes_language swedish
43 SDCC Compiler User Guide
47 \begin_inset LatexCommand \tableofcontents{}
64 is a Freeware, retargettable, optimizing ANSI-C compiler by
68 designed for 8 bit Microprocessors.
69 The current version targets Intel MCS51 based Microprocessors(8051,8052,
70 etc), Zilog Z80 based MCUs, and the Dallas DS80C390 variant.
71 It can be retargetted for other microprocessors, support for PIC, AVR and
72 186 is under development.
73 The entire source code for the compiler is distributed under GPL.
74 SDCC uses ASXXXX & ASLINK, a Freeware, retargettable assembler & linker.
75 SDCC has extensive language extensions suitable for utilizing various microcont
76 rollers and underlying hardware effectively.
81 In addition to the MCU specific optimizations SDCC also does a host of standard
85 global sub expression elimination,
88 loop optimizations (loop invariant, strength reduction of induction variables
92 constant folding & propagation,
108 For the back-end SDCC uses a global register allocation scheme which should
109 be well suited for other 8 bit MCUs.
114 The peep hole optimizer uses a rule based substitution mechanism which is
120 Supported data-types are:
123 char (8 bits, 1 byte),
126 short and int (16 bits, 2 bytes),
129 long (32 bit, 4 bytes)
136 The compiler also allows
138 inline assembler code
140 to be embedded anywhere in a function.
141 In addition, routines developed in assembly can also be called.
145 SDCC also provides an option (---cyclomatic) to report the relative complexity
147 These functions can then be further optimized, or hand coded in assembly
153 SDCC also comes with a companion source level debugger SDCDB, the debugger
154 currently uses ucSim a freeware simulator for 8051 and other micro-controllers.
159 The latest version can be downloaded from
160 \begin_inset LatexCommand \htmlurl{http://sdcc.sourceforge.net/}
172 All packages used in this compiler system are
180 ; source code for all the sub-packages (pre-processor, assemblers, linkers
181 etc) is distributed with the package.
182 This documentation is maintained using a freeware word processor (LyX).
184 This program is free software; you can redistribute it and/or modify it
185 under the terms of the GNU General Public License as published by the Free
186 Software Foundation; either version 2, or (at your option) any later version.
187 This program is distributed in the hope that it will be useful, but WITHOUT
188 ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
189 FOR A PARTICULAR PURPOSE.
190 See the GNU General Public License for more details.
191 You should have received a copy of the GNU General Public License along
192 with this program; if not, write to the Free Software Foundation, 59 Temple
193 Place - Suite 330, Boston, MA 02111-1307, USA.
194 In other words, you are welcome to use, share and improve this program.
195 You are forbidden to forbid anyone else to use, share and improve what
197 Help stamp out software-hoarding!
200 Typographic conventions
203 Throughout this manual, we will use the following convention.
204 Commands you have to type in are printed in
212 Code samples are printed in
217 Interesting items and new terms are printed in
222 Compatibility with previous versions
225 This version has numerous bug fixes compared with the previous version.
226 But we also introduced some incompatibilities with older versions.
227 Not just for the fun of it, but to make the compiler more stable, efficient
234 short is now equivalent to int (16 bits), it used to be equivalent to char
235 (8 bits) which is not ANSI compliant
238 the default directory where include, library and documention files are stored
239 is now in /usr/local/share
242 char type parameters to vararg functions are casted to int unless explicitly
259 will push a as an int and as a char resp.
266 regextend has been removed
269 option --noregparms has been removed
272 option --stack-after-data has been removed
277 <pending: more incompatibilities?>
283 What do you need before you start installation of SDCC? A computer, and
285 The preferred method of installation is to compile SDCC from source using
287 For Windows some pre-compiled binary distributions are available for your
289 You should have some experience with command line tools and compiler use.
295 The SDCC home page at
296 \begin_inset LatexCommand \htmlurl{http://sdcc.sourceforge.net/}
300 is a great place to find distribution sets.
301 You can also find links to the user mailing lists that offer help or discuss
302 SDCC with other SDCC users.
303 Web links to other SDCC related sites can also be found here.
304 This document can be found in the DOC directory of the source package as
306 Some of the other tools (simulator and assembler) included with SDCC contain
307 their own documentation and can be found in the source distribution.
308 If you want the latest unreleased software, the complete source package
309 is available directly by anonymous CVS on cvs.sdcc.sourceforge.net.
312 Wishes for the future
315 There are (and always will be) some things that could be done.
316 Here are some I can think of:
323 char KernelFunction3(char p) at 0x340;
329 If you can think of some more, please send them to the list.
335 <pending: And then of course a proper index-table
336 \begin_inset LatexCommand \index{index}
346 Install and search paths
349 Linux (and other gcc-builds like Solaris, Cygwin, Mingw and OSX) by default
350 install in /usr/local.
351 You can override this when configuring with --prefix-path.
352 Subdirs used will be bin, share/sdcc/include, share/sdcc/lib and share/sdcc/doc.
354 Windows MSVC and Borland builds will install in one single tree (e.g.
355 /sdcc) with subdirs bin, lib, include and doc.
359 The paths searched when running the compiler are as follows (the first catch
363 Binary files (preprocessor, assembler and linker):
365 - the path of argv[0] (if available)
368 \begin_inset Quotes sld
372 \begin_inset Quotes srd
378 \begin_inset Quotes sld
382 \begin_inset Quotes srd
395 \begin_inset Quotes sld
399 \begin_inset Quotes srd
405 \begin_inset Quotes sld
410 - /usr/local/share/sdcc/include (gcc builds)
412 - path(arv[0])/../include and then /sdcc/include (windoze msvc and borland
420 is auto-appended by the compiler, e.g.
421 small, large, z80, ds390 etc.):
426 \begin_inset Quotes sld
430 \begin_inset Quotes srd
440 \begin_inset Quotes sld
444 \begin_inset Quotes srd
453 - /usr/local/share/sdcc/lib/
459 - path(argv[0])/../lib/
467 (windoze msvc and borland builds)
470 Documentation (although never really searched for, you have to do that yourself
474 \begin_inset Quotes sld
478 \begin_inset Quotes srd
483 - /usr/local/share/sdcc/doc (gcc builds)
485 - /sdcc/doc (windoze msvc and borland builds)
488 So, for windoze it is highly recommended to set the environment variable
489 SDCCHOME to prevent needless usage of -I and -L.
492 Linux and other gcc-based systems (cygwin, mingw, osx)
497 Download the source package
499 either from the SDCC CVS repository or from the
500 \begin_inset LatexCommand \url[nightly snapshots]{http://sdcc.sourceforge.net/snap.php}
506 , it will be named something like sdcc
515 Bring up a command line terminal, such as xterm.
520 Unpack the file using a command like:
523 "tar -xzf sdcc.src.tgz
528 , this will create a sub-directory called sdcc with all of the sources.
531 Change directory into the main SDCC directory, for example type:
548 This configures the package for compilation on your system.
564 All of the source packages will compile, this can take a while.
580 This copies the binary executables, the include files, the libraries and
581 the documentation to the install directories.
585 \layout Subsubsection
587 Windows Install Using a Binary Package
590 Download the binary package and unpack it using your favorite unpacking
591 tool (gunzip, WinZip, etc).
592 This should unpack to a group of sub-directories.
593 An example directory structure after unpacking the mingw package is: c:
599 bin for the executables, c:
619 lib for the include and libraries.
622 Adjust your environment variable PATH to include the location of the bin
623 directory or start sdcc using the full path.
624 \layout Subsubsection
626 Windows Install Using Cygwin and Mingw
629 Follow the instruction in
631 Linux and other gcc-based systems
634 \layout Subsubsection
636 Windows Install Using Microsoft Visual C++ 6.0/NET
641 Download the source package
643 either from the SDCC CVS repository or from the
644 \begin_inset LatexCommand \url[nightly snapshots]{http://sdcc.sourceforge.net/snap.php}
650 , it will be named something like sdcc
657 SDCC is distributed with all the projects, workspaces, and files you need
658 to build it using Visual C++ 6.0/NET.
659 The workspace name is 'sdcc.dsw'.
660 Please note that as it is now, all the executables are created in a folder
664 Once built you need to copy the executables from sdcc
668 bin before runnng SDCC.
673 In order to build SDCC with Visual C++ 6.0/NET you need win32 executables
674 of bison.exe, flex.exe, and gawk.exe.
675 One good place to get them is
676 \begin_inset LatexCommand \url[here]{http://unxutils.sourceforge.net}
684 Download the file UnxUtils.zip.
685 Now you have to install the utilities and setup Visual C++ so it can locate
686 the required programs.
687 Here there are two alternatives (choose one!):
694 a) Extract UnxUtils.zip to your C:
696 hard disk PRESERVING the original paths, otherwise bison won't work.
697 (If you are using WinZip make certain that 'Use folder names' is selected)
701 b) In the Visual C++ IDE click Tools, Options, select the Directory tab,
702 in 'Show directories for:' select 'Executable files', and in the directories
703 window add a new path: 'C:
713 (As a side effect, you get a bunch of Unix utilities that could be useful,
714 such as diff and patch.)
721 This one avoids extracting a bunch of files you may not use, but requires
726 a) Create a directory were to put the tools needed, or use a directory already
734 b) Extract 'bison.exe', 'bison.hairy', 'bison.simple', 'flex.exe', and gawk.exe
735 to such directory WITHOUT preserving the original paths.
736 (If you are using WinZip make certain that 'Use folder names' is not selected)
740 c) Rename bison.exe to '_bison.exe'.
744 d) Create a batch file 'bison.bat' in 'C:
748 ' and add these lines:
768 _bison %1 %2 %3 %4 %5 %6 %7 %8 %9
772 Steps 'c' and 'd' are needed because bison requires by default that the
773 files 'bison.simple' and 'bison.hairy' reside in some weird Unix directory,
774 '/usr/local/share/' I think.
775 So it is necessary to tell bison where those files are located if they
776 are not in such directory.
777 That is the function of the environment variables BISON_SIMPLE and BISON_HAIRY.
781 e) In the Visual C++ IDE click Tools, Options, select the Directory tab,
782 in 'Show directories for:' select 'Executable files', and in the directories
783 window add a new path: 'c:
786 Note that you can use any other path instead of 'c:
788 util', even the path where the Visual C++ tools are, probably: 'C:
792 Microsoft Visual Studio
797 So you don't have to execute step 'e' :)
801 Open 'sdcc.dsw' in Visual Studio, click 'build all', when it finishes copy
802 the executables from sdcc
806 bin, and you can compile using sdcc.
807 \layout Subsubsection
809 Windows Install Using Borland ......
817 Testing out the SDCC Compiler
820 The first thing you should do after installing your SDCC compiler is to
828 at the prompt, and the program should run and tell you the version.
829 If it doesn't run, or gives a message about not finding sdcc program, then
830 you need to check over your installation.
831 Make sure that the sdcc bin directory is in your executable search path
832 defined by the PATH environment setting (see the Trouble-shooting section
834 Make sure that the sdcc program is in the bin folder, if not perhaps something
835 did not install correctly.
841 SDCC binaries are commonly installed in a directory arrangement like this:
849 <lyxtabular version="3" rows="3" columns="2">
851 <column alignment="left" valignment="top" leftline="true" width="0(null)">
852 <column alignment="left" valignment="top" leftline="true" rightline="true" width="0(null)">
853 <row topline="true" bottomline="true">
854 <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
864 <cell alignment="center" valignment="top" topline="true" leftline="true" rightline="true" usebox="none">
871 Holds executables(sdcc, s51, aslink,
880 <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
887 usr/local/share/sdcc/lib
890 <cell alignment="center" valignment="top" topline="true" leftline="true" rightline="true" usebox="none">
903 <row topline="true" bottomline="true">
904 <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
911 usr/local/share/sdcc/include
914 <cell alignment="center" valignment="top" topline="true" leftline="true" rightline="true" usebox="none">
921 Holds common C header files
935 Make sure the compiler works on a very simple example.
936 Type in the following test.c program using your favorite editor:
967 Compile this using the following command:
976 If all goes well, the compiler will generate a test.asm and test.rel file.
977 Congratulations, you've just compiled your first program with SDCC.
978 We used the -c option to tell SDCC not to link the generated code, just
979 to keep things simple for this step.
987 The next step is to try it with the linker.
997 If all goes well the compiler will link with the libraries and produce
998 a test.ihx output file.
1003 (no test.ihx, and the linker generates warnings), then the problem is most
1004 likely that sdcc cannot find the
1008 usr/local/share/sdcc/lib directory
1012 (see the Install trouble-shooting section for suggestions).
1020 The final test is to ensure sdcc can use the
1024 header files and libraries.
1025 Edit test.c and change it to the following:
1045 strcpy(str1, "testing");
1054 Compile this by typing
1061 This should generate a test.ihx output file, and it should give no warnings
1062 such as not finding the string.h file.
1063 If it cannot find the string.h file, then the problem is that sdcc cannot
1064 find the /usr/local/share/sdcc/include directory
1068 (see the Install trouble-shooting section for suggestions).
1071 Install Trouble-shooting
1072 \layout Subsubsection
1074 SDCC does not build correctly.
1077 A thing to try is starting from scratch by unpacking the .tgz source package
1078 again in an empty directory.
1085 ./configure 2>&1 | tee configure.log
1097 make 2>&1 | tee make.log
1103 If anything goes wrong, you can review the log files to locate the problem.
1104 Or a relevant part of this can be attached to an email that could be helpful
1105 when requesting help from the mailing list.
1106 \layout Subsubsection
1109 \begin_inset Quotes sld
1113 \begin_inset Quotes srd
1120 \begin_inset Quotes sld
1124 \begin_inset Quotes srd
1127 command is a script that analyzes your system and performs some configuration
1128 to ensure the source package compiles on your system.
1129 It will take a few minutes to run, and will compile a few tests to determine
1130 what compiler features are installed.
1131 \layout Subsubsection
1134 \begin_inset Quotes sld
1138 \begin_inset Quotes srd
1144 This runs the GNU make tool, which automatically compiles all the source
1145 packages into the final installed binary executables.
1146 \layout Subsubsection
1149 \begin_inset Quotes sld
1153 \begin_inset Quotes erd
1159 This will install the compiler, other executables and libraries in to the
1160 appropriate system directories.
1161 The default is to copy the executables to /usr/local/bin and the libraries
1162 and header files to /usr/local/share/sdcc/lib and /usr/local/share/sdcc/include.
1163 On most systems you will need super-user privilages to do this.
1166 Advanced Install Options
1170 \begin_inset Quotes eld
1174 \begin_inset Quotes erd
1177 command has several options.
1178 The most commonly used option is --prefix=<directory name>, where <directory
1179 name> is the final location for the sdcc executables and libraries, (default
1180 location is /usr/local).
1181 The installation process will create the following directory structure
1182 under the <directory name> specified (if they do not already exist).
1187 bin/ - binary exectables (add to PATH environment variable)
1191 bin/share/sdcc/include/ - include header files
1195 bin/share/sdcc/lib/small/ - Object & library files for small model library
1197 bin/share/sdcc/lib/large/ - Object & library files for large model library
1199 bin/share/sdcc/lib/ds390/ - Object & library files for DS80C390 library
1201 bin/share/sdcc/lib/z80/ - Object & library files for Z80 library
1209 \begin_inset Quotes sld
1212 ./configure --prefix=/usr/local
1213 \begin_inset Quotes erd
1219 will configure the compiler to be installed in directory /usr/local.
1225 SDCC is not just a compiler, but a collection of tools by various developers.
1226 These include linkers, assemblers, simulators and other components.
1227 Here is a summary of some of the components.
1228 Note that the included simulator and assembler have separate documentation
1229 which you can find in the source package in their respective directories.
1230 As SDCC grows to include support for other processors, other packages from
1231 various developers are included and may have their own sets of documentation.
1235 You might want to look at the files which are installed in <installdir>.
1236 At the time of this writing, we find the following programs:
1240 In <installdir>/bin:
1243 sdcc - The compiler.
1246 sdcpp - The C preprocessor.
1249 asx8051 - The assembler for 8051 type processors.
1256 as-gbz80 - The Z80 and GameBoy Z80 assemblers.
1259 aslink -The linker for 8051 type processors.
1266 link-gbz80 - The Z80 and GameBoy Z80 linkers.
1269 s51 - The ucSim 8051 simulator.
1272 sdcdb - The source debugger.
1275 packihx - A tool to pack (compress) Intel hex files.
1278 In <installdir>/share/sdcc/include
1284 In <installdir>/share/sdcc/lib
1287 the sources of the runtime library and the subdirs small large and ds390
1288 with the precompiled relocatables.
1291 In <installdir>/share/sdcc/doc
1297 As development for other processors proceeds, this list will expand to include
1298 executables to support processors like AVR, PIC, etc.
1299 \layout Subsubsection
1304 This is the actual compiler, it in turn uses the c-preprocessor and invokes
1305 the assembler and linkage editor.
1306 \layout Subsubsection
1308 sdcpp - The C-Preprocessor)
1311 The preprocessor is a modified version of the GNU preprocessor.
1312 The C preprocessor is used to pull in #include sources, process #ifdef
1313 statements, #defines and so on.
1314 \layout Subsubsection
1316 asx8051, as-z80, as-gbz80, aslink, link-z80, link-gbz80 - The Assemblers
1320 This is retargettable assembler & linkage editor, it was developed by Alan
1322 John Hartman created the version for 8051, and I (Sandeep) have made some
1323 enhancements and bug fixes for it to work properly with the SDCC.
1324 \layout Subsubsection
1329 S51 is a freeware, opensource simulator developed by Daniel Drotos (
1330 \begin_inset LatexCommand \url{mailto:drdani@mazsola.iit.uni-miskolc.hu}
1335 The simulator is built as part of the build process.
1336 For more information visit Daniel's website at:
1337 \begin_inset LatexCommand \url{http://mazsola.iit.uni-miskolc.hu/~drdani/embedded/s51}
1342 It currently support the core mcs51, the Dallas DS80C390 and the Philips
1344 \layout Subsubsection
1346 sdcdb - Source Level Debugger
1352 <todo: is this thing alive?>
1359 Sdcdb is the companion source level debugger.
1360 The current version of the debugger uses Daniel's Simulator S51, but can
1361 be easily changed to use other simulators.
1368 \layout Subsubsection
1370 Single Source File Projects
1373 For single source file 8051 projects the process is very simple.
1374 Compile your programs with the following command
1377 "sdcc sourcefile.c".
1381 This will compile, assemble and link your source file.
1382 Output files are as follows
1386 sourcefile.asm - Assembler source file created by the compiler
1388 sourcefile.lst - Assembler listing file created by the Assembler
1390 sourcefile.rst - Assembler listing file updated with linkedit information,
1391 created by linkage editor
1393 sourcefile.sym - symbol listing for the sourcefile, created by the assembler
1395 sourcefile.rel - Object file created by the assembler, input to Linkage editor
1397 sourcefile.map - The memory map for the load module, created by the Linker
1399 sourcefile.ihx - The load module in Intel hex format (you can select the
1400 Motorola S19 format with --out-fmt-s19)
1402 sourcefile.cdb - An optional file (with --debug) containing debug information
1405 \layout Subsubsection
1407 Projects with Multiple Source Files
1410 SDCC can compile only ONE file at a time.
1411 Let us for example assume that you have a project containing the following
1416 foo1.c (contains some functions)
1418 foo2.c (contains some more functions)
1420 foomain.c (contains more functions and the function main)
1428 The first two files will need to be compiled separately with the commands:
1460 Then compile the source file containing the
1464 function and link the files together with the following command:
1472 foomain.c\SpecialChar ~
1473 foo1.rel\SpecialChar ~
1485 can be separately compiled as well:
1496 sdcc foomain.rel foo1.rel foo2.rel
1503 The file containing the
1518 file specified in the command line, since the linkage editor processes
1519 file in the order they are presented to it.
1520 \layout Subsubsection
1522 Projects with Additional Libraries
1525 Some reusable routines may be compiled into a library, see the documentation
1526 for the assembler and linkage editor (which are in <installdir>/share/sdcc/doc)
1532 Libraries created in this manner can be included in the command line.
1533 Make sure you include the -L <library-path> option to tell the linker where
1534 to look for these files if they are not in the current directory.
1535 Here is an example, assuming you have the source file
1547 (if that is not the same as your current project):
1554 sdcc foomain.c foolib.lib -L mylib
1565 must be an absolute path name.
1569 The most efficient way to use libraries is to keep seperate modules in seperate
1571 The lib file now should name all the modules.rel files.
1572 For an example see the standard library file
1576 in the directory <installdir>/share/lib/small.
1579 Command Line Options
1580 \layout Subsubsection
1582 Processor Selection Options
1584 \labelwidthstring 00.00.0000
1590 Generate code for the MCS51 (8051) family of processors.
1591 This is the default processor target.
1593 \labelwidthstring 00.00.0000
1599 Generate code for the DS80C390 processor.
1601 \labelwidthstring 00.00.0000
1607 Generate code for the Z80 family of processors.
1609 \labelwidthstring 00.00.0000
1615 Generate code for the GameBoy Z80 processor.
1617 \labelwidthstring 00.00.0000
1623 Generate code for the Atmel AVR processor (In development, not complete).
1625 \labelwidthstring 00.00.0000
1631 Generate code for the PIC 14-bit processors (In development, not complete).
1633 \labelwidthstring 00.00.0000
1639 Generate code for the Toshiba TLCS-900H processor (In development, not
1642 \labelwidthstring 00.00.0000
1648 Generate code for the Philips XA51 processor (In development, not complete).
1649 \layout Subsubsection
1651 Preprocessor Options
1653 \labelwidthstring 00.00.0000
1659 The additional location where the pre processor will look for <..h> or
1660 \begin_inset Quotes eld
1664 \begin_inset Quotes erd
1669 \labelwidthstring 00.00.0000
1675 Command line definition of macros.
1676 Passed to the pre processor.
1678 \labelwidthstring 00.00.0000
1684 Tell the preprocessor to output a rule suitable for make describing the
1685 dependencies of each object file.
1686 For each source file, the preprocessor outputs one make-rule whose target
1687 is the object file name for that source file and whose dependencies are
1688 all the files `#include'd in it.
1689 This rule may be a single line or may be continued with `
1691 '-newline if it is long.
1692 The list of rules is printed on standard output instead of the preprocessed
1696 \labelwidthstring 00.00.0000
1702 Tell the preprocessor not to discard comments.
1703 Used with the `-E' option.
1705 \labelwidthstring 00.00.0000
1716 Like `-M' but the output mentions only the user header files included with
1718 \begin_inset Quotes eld
1722 System header files included with `#include <file>' are omitted.
1724 \labelwidthstring 00.00.0000
1730 Assert the answer answer for question, in case it is tested with a preprocessor
1731 conditional such as `#if #question(answer)'.
1732 `-A-' disables the standard assertions that normally describe the target
1735 \labelwidthstring 00.00.0000
1741 (answer) Assert the answer answer for question, in case it is tested with
1742 a preprocessor conditional such as `#if #question(answer)'.
1743 `-A-' disables the standard assertions that normally describe the target
1746 \labelwidthstring 00.00.0000
1752 Undefine macro macro.
1753 `-U' options are evaluated after all `-D' options, but before any `-include'
1754 and `-imacros' options.
1756 \labelwidthstring 00.00.0000
1762 Tell the preprocessor to output only a list of the macro definitions that
1763 are in effect at the end of preprocessing.
1764 Used with the `-E' option.
1766 \labelwidthstring 00.00.0000
1772 Tell the preprocessor to pass all macro definitions into the output, in
1773 their proper sequence in the rest of the output.
1775 \labelwidthstring 00.00.0000
1786 Like `-dD' except that the macro arguments and contents are omitted.
1787 Only `#define name' is included in the output.
1788 \layout Subsubsection
1792 \labelwidthstring 00.00.0000
1799 the output path resp.
1800 file where everything will be placed
1802 \labelwidthstring 00.00.0000
1812 <absolute path to additional libraries> This option is passed to the linkage
1813 editor's additional libraries search path.
1814 The path name must be absolute.
1815 Additional library files may be specified in the command line.
1816 See section Compiling programs for more details.
1818 \labelwidthstring 00.00.0000
1824 <Value> The start location of the external ram, default value is 0.
1825 The value entered can be in Hexadecimal or Decimal format, e.g.: --xram-loc
1826 0x8000 or --xram-loc 32768.
1828 \labelwidthstring 00.00.0000
1834 <Value> The start location of the code segment, default value 0.
1835 Note when this option is used the interrupt vector table is also relocated
1836 to the given address.
1837 The value entered can be in Hexadecimal or Decimal format, e.g.: --code-loc
1838 0x8000 or --code-loc 32768.
1840 \labelwidthstring 00.00.0000
1846 <Value> By default the stack is placed after the data segment.
1847 Using this option the stack can be placed anywhere in the internal memory
1849 The value entered can be in Hexadecimal or Decimal format, e.g.
1850 --stack-loc 0x20 or --stack-loc 32.
1851 Since the sp register is incremented before a push or call, the initial
1852 sp will be set to one byte prior the provided value.
1853 The provided value should not overlap any other memory areas such as used
1854 register banks or the data segment and with enough space for the current
1857 \labelwidthstring 00.00.0000
1863 <Value> The start location of the internal ram data segment.
1864 The value entered can be in Hexadecimal or Decimal format, eg.
1865 --data-loc 0x20 or --data-loc 32.
1866 (By default, the start location of the internal ram data segment is set
1867 as low as possible in memory, taking into account the used register banks
1868 and the bit segment at address 0x20.
1869 For example if register banks 0 and 1 are used without bit variables, the
1870 data segment will be set, if --data-loc is not used, to location 0x10.)
1872 \labelwidthstring 00.00.0000
1878 <Value> The start location of the indirectly addressable internal ram, default
1880 The value entered can be in Hexadecimal or Decimal format, eg.
1881 --idata-loc 0x88 or --idata-loc 136.
1883 \labelwidthstring 00.00.0000
1892 The linker output (final object code) is in Intel Hex format.
1893 (This is the default option).
1895 \labelwidthstring 00.00.0000
1904 The linker output (final object code) is in Motorola S19 format.
1905 \layout Subsubsection
1909 \labelwidthstring 00.00.0000
1915 Generate code for Large model programs see section Memory Models for more
1917 If this option is used all source files in the project should be compiled
1919 In addition the standard library routines are compiled with small model,
1920 they will need to be recompiled.
1922 \labelwidthstring 00.00.0000
1933 Generate code for Small Model programs see section Memory Models for more
1935 This is the default model.
1936 \layout Subsubsection
1940 \labelwidthstring 00.00.0000
1951 Generate 24-bit flat mode code.
1952 This is the one and only that the ds390 code generator supports right now
1953 and is default when using
1958 See section Memory Models for more details.
1960 \labelwidthstring 00.00.0000
1966 Generate code for the 10 bit stack mode of the Dallas DS80C390 part.
1967 This is the one and only that the ds390 code generator supports right now
1968 and is default when using
1973 In this mode, the stack is located in the lower 1K of the internal RAM,
1974 which is mapped to 0x400000.
1975 Note that the support is incomplete, since it still uses a single byte
1976 as the stack pointer.
1977 This means that only the lower 256 bytes of the potential 1K stack space
1978 will actually be used.
1979 However, this does allow you to reclaim the precious 256 bytes of low RAM
1980 for use for the DATA and IDATA segments.
1981 The compiler will not generate any code to put the processor into 10 bit
1983 It is important to ensure that the processor is in this mode before calling
1984 any re-entrant functions compiled with this option.
1985 In principle, this should work with the
1989 option, but that has not been tested.
1990 It is incompatible with the
1995 It also only makes sense if the processor is in 24 bit contiguous addressing
1998 --model-flat24 option
2001 \layout Subsubsection
2003 Optimization Options
2005 \labelwidthstring 00.00.0000
2011 Will not do global subexpression elimination, this option may be used when
2012 the compiler creates undesirably large stack/data spaces to store compiler
2014 A warning message will be generated when this happens and the compiler
2015 will indicate the number of extra bytes it allocated.
2016 It recommended that this option NOT be used, #pragma\SpecialChar ~
2018 to turn off global subexpression elimination for a given function only.
2020 \labelwidthstring 00.00.0000
2026 Will not do loop invariant optimizations, this may be turned off for reasons
2027 explained for the previous option.
2028 For more details of loop optimizations performed see section Loop Invariants.It
2029 recommended that this option NOT be used, #pragma\SpecialChar ~
2030 NOINVARIANT can be used
2031 to turn off invariant optimizations for a given function only.
2033 \labelwidthstring 00.00.0000
2039 Will not do loop induction optimizations, see section strength reduction
2040 for more details.It is recommended that this option is NOT used, #pragma\SpecialChar ~
2042 ION can be used to turn off induction optimizations for a given function
2045 \labelwidthstring 00.00.0000
2056 Will not generate boundary condition check when switch statements are implement
2057 ed using jump-tables.
2058 See section Switch Statements for more details.
2059 It is recommended that this option is NOT used, #pragma\SpecialChar ~
2061 used to turn off boundary checking for jump tables for a given function
2064 \labelwidthstring 00.00.0000
2073 Will not do loop reversal optimization.
2075 \labelwidthstring 00.00.0000
2081 This will disable the memcpy of initialized data in far space from code
2083 \layout Subsubsection
2087 \labelwidthstring 00.00.0000
2094 will compile and assemble the source, but will not call the linkage editor.
2096 \labelwidthstring 00.00.0000
2102 Run only the C preprocessor.
2103 Preprocess all the C source files specified and output the results to standard
2106 \labelwidthstring 00.00.0000
2117 All functions in the source file will be compiled as
2122 the parameters and local variables will be allocated on the stack.
2123 see section Parameters and Local Variables for more details.
2124 If this option is used all source files in the project should be compiled
2128 \labelwidthstring 00.00.0000
2134 Uses a pseudo stack in the first 256 bytes in the external ram for allocating
2135 variables and passing parameters.
2136 See section on external stack for more details.
2138 \labelwidthstring 00.00.0000
2142 --callee-saves function1[,function2][,function3]....
2145 The compiler by default uses a caller saves convention for register saving
2146 across function calls, however this can cause unneccessary register pushing
2147 & popping when calling small functions from larger functions.
2148 This option can be used to switch the register saving convention for the
2149 function names specified.
2150 The compiler will not save registers when calling these functions, no extra
2151 code will be generated at the entry & exit for these functions to save
2152 & restore the registers used by these functions, this can SUBSTANTIALLY
2153 reduce code & improve run time performance of the generated code.
2154 In the future the compiler (with interprocedural analysis) will be able
2155 to determine the appropriate scheme to use for each function call.
2156 DO NOT use this option for built-in functions such as _muluint..., if this
2157 option is used for a library function the appropriate library function
2158 needs to be recompiled with the same option.
2159 If the project consists of multiple source files then all the source file
2160 should be compiled with the same --callee-saves option string.
2161 Also see #pragma\SpecialChar ~
2164 \labelwidthstring 00.00.0000
2173 When this option is used the compiler will generate debug information, that
2174 can be used with the SDCDB.
2175 The debug information is collected in a file with .cdb extension.
2176 For more information see documentation for SDCDB.
2178 \labelwidthstring 00.00.0000
2184 <filename> This option can be used to use additional rules to be used by
2185 the peep hole optimizer.
2186 See section Peep Hole optimizations for details on how to write these rules.
2188 \labelwidthstring 00.00.0000
2199 Stop after the stage of compilation proper; do not assemble.
2200 The output is an assembler code file for the input file specified.
2202 \labelwidthstring 00.00.0000
2206 -Wa_asmOption[,asmOption]
2209 Pass the asmOption to the assembler.
2211 \labelwidthstring 00.00.0000
2215 -Wl_linkOption[,linkOption]
2218 Pass the linkOption to the linker.
2220 \labelwidthstring 00.00.0000
2229 Integer (16 bit) and long (32 bit) libraries have been compiled as reentrant.
2230 Note by default these libraries are compiled as non-reentrant.
2231 See section Installation for more details.
2233 \labelwidthstring 00.00.0000
2242 This option will cause the compiler to generate an information message for
2243 each function in the source file.
2244 The message contains some
2248 information about the function.
2249 The number of edges and nodes the compiler detected in the control flow
2250 graph of the function, and most importantly the
2252 cyclomatic complexity
2254 see section on Cyclomatic Complexity for more details.
2256 \labelwidthstring 00.00.0000
2265 Floating point library is compiled as reentrant.See section Installation
2268 \labelwidthstring 00.00.0000
2274 The compiler will not overlay parameters and local variables of any function,
2275 see section Parameters and local variables for more details.
2277 \labelwidthstring 00.00.0000
2283 This option can be used when the code generated is called by a monitor
2285 The compiler will generate a 'ret' upon return from the 'main' function.
2286 The default option is to lock up i.e.
2289 \labelwidthstring 00.00.0000
2295 Disable peep-hole optimization.
2297 \labelwidthstring 00.00.0000
2303 Pass the inline assembler code through the peep hole optimizer.
2304 This can cause unexpected changes to inline assembler code, please go through
2305 the peephole optimizer rules defined in the source file tree '<target>/peeph.def
2306 ' before using this option.
2308 \labelwidthstring 00.00.0000
2314 <Value> Causes the linker to check if the internal ram usage is within limits
2317 \labelwidthstring 00.00.0000
2323 <Value> Causes the linker to check if the external ram usage is within limits
2326 \labelwidthstring 00.00.0000
2332 <Value> Causes the linker to check if the code usage is within limits of
2335 \labelwidthstring 00.00.0000
2341 This will prevent the compiler from passing on the default include path
2342 to the preprocessor.
2344 \labelwidthstring 00.00.0000
2350 This will prevent the compiler from passing on the default library path
2353 \labelwidthstring 00.00.0000
2359 Shows the various actions the compiler is performing.
2361 \labelwidthstring 00.00.0000
2367 Shows the actual commands the compiler is executing.
2368 \layout Subsubsection
2370 Intermediate Dump Options
2373 The following options are provided for the purpose of retargetting and debugging
2375 These provided a means to dump the intermediate code (iCode) generated
2376 by the compiler in human readable form at various stages of the compilation
2380 \labelwidthstring 00.00.0000
2386 This option will cause the compiler to dump the intermediate code into
2389 <source filename>.dumpraw
2391 just after the intermediate code has been generated for a function, i.e.
2392 before any optimizations are done.
2393 The basic blocks at this stage ordered in the depth first number, so they
2394 may not be in sequence of execution.
2396 \labelwidthstring 00.00.0000
2402 Will create a dump of iCode's, after global subexpression elimination,
2405 <source filename>.dumpgcse.
2407 \labelwidthstring 00.00.0000
2413 Will create a dump of iCode's, after deadcode elimination, into a file
2416 <source filename>.dumpdeadcode.
2418 \labelwidthstring 00.00.0000
2427 Will create a dump of iCode's, after loop optimizations, into a file named
2430 <source filename>.dumploop.
2432 \labelwidthstring 00.00.0000
2441 Will create a dump of iCode's, after live range analysis, into a file named
2444 <source filename>.dumprange.
2446 \labelwidthstring 00.00.0000
2452 Will dump the life ranges for all symbols.
2454 \labelwidthstring 00.00.0000
2463 Will create a dump of iCode's, after register assignment, into a file named
2466 <source filename>.dumprassgn.
2468 \labelwidthstring 00.00.0000
2474 Will create a dump of the live ranges of iTemp's
2476 \labelwidthstring 00.00.0000
2487 Will cause all the above mentioned dumps to be created.
2490 MCS51/DS390 Storage Class Language Extensions
2493 In addition to the ANSI storage classes SDCC allows the following MCS51
2494 specific storage classes.
2495 \layout Subsubsection
2500 Variables declared with this storage class will be placed in the extern
2506 storage class for Large Memory model, e.g.:
2512 xdata unsigned char xduc;
2513 \layout Subsubsection
2522 storage class for Small Memory model.
2523 Variables declared with this storage class will be allocated in the internal
2531 \layout Subsubsection
2536 Variables declared with this storage class will be allocated into the indirectly
2537 addressable portion of the internal ram of a 8051, e.g.:
2544 \layout Subsubsection
2549 This is a data-type and a storage class specifier.
2550 When a variable is declared as a bit, it is allocated into the bit addressable
2551 memory of 8051, e.g.:
2558 \layout Subsubsection
2563 Like the bit keyword,
2567 signifies both a data-type and storage class, they are used to describe
2568 the special function registers and special bit variables of a 8051, eg:
2574 sfr at 0x80 P0; /* special function register P0 at location 0x80 */
2576 sbit at 0xd7 CY; /* CY (Carry Flag) */
2582 SDCC allows (via language extensions) pointers to explicitly point to any
2583 of the memory spaces of the 8051.
2584 In addition to the explicit pointers, the compiler uses (by default) generic
2585 pointers which can be used to point to any of the memory spaces.
2589 Pointer declaration examples:
2598 /* pointer physically in xternal ram pointing to object in internal ram
2601 data unsigned char * xdata p;
2605 /* pointer physically in code rom pointing to data in xdata space */
2607 xdata unsigned char * code p;
2611 /* pointer physically in code space pointing to data in code space */
2613 code unsigned char * code p;
2617 /* the folowing is a generic pointer physically located in xdata space */
2628 Well you get the idea.
2633 All unqualified pointers are treated as 3-byte (4-byte for the ds390)
2646 The highest order byte of the
2650 pointers contains the data space information.
2651 Assembler support routines are called whenever data is stored or retrieved
2657 These are useful for developing reusable library routines.
2658 Explicitly specifying the pointer type will generate the most efficient
2662 Parameters & Local Variables
2665 Automatic (local) variables and parameters to functions can either be placed
2666 on the stack or in data-space.
2667 The default action of the compiler is to place these variables in the internal
2668 RAM (for small model) or external RAM (for large model).
2669 This in fact makes them
2673 so by default functions are non-reentrant.
2677 They can be placed on the stack either by using the
2681 option or by using the
2685 keyword in the function declaration, e.g.:
2694 unsigned char foo(char i) reentrant
2707 Since stack space on 8051 is limited, the
2715 option should be used sparingly.
2716 Note that the reentrant keyword just means that the parameters & local
2717 variables will be allocated to the stack, it
2721 mean that the function is register bank independent.
2725 Local variables can be assigned storage classes and absolute addresses,
2732 unsigned char foo() {
2738 xdata unsigned char i;
2750 data at 0x31 unsiged char j;
2765 In the above example the variable
2769 will be allocated in the external ram,
2773 in bit addressable space and
2782 or when a function is declared as
2786 this should only be done for static variables.
2789 Parameters however are not allowed any storage class, (storage classes for
2790 parameters will be ignored), their allocation is governed by the memory
2791 model in use, and the reentrancy options.
2797 For non-reentrant functions SDCC will try to reduce internal ram space usage
2798 by overlaying parameters and local variables of a function (if possible).
2799 Parameters and local variables of a function will be allocated to an overlayabl
2800 e segment if the function has
2802 no other function calls and the function is non-reentrant and the memory
2806 If an explicit storage class is specified for a local variable, it will
2810 Note that the compiler (not the linkage editor) makes the decision for overlayin
2812 Functions that are called from an interrupt service routine should be preceded
2813 by a #pragma\SpecialChar ~
2814 NOOVERLAY if they are not reentrant.
2817 Also note that the compiler does not do any processing of inline assembler
2818 code, so the compiler might incorrectly assign local variables and parameters
2819 of a function into the overlay segment if the inline assembler code calls
2820 other c-functions that might use the overlay.
2821 In that case the #pragma\SpecialChar ~
2822 NOOVERLAY should be used.
2825 Parameters and Local variables of functions that contain 16 or 32 bit multiplica
2826 tion or division will NOT be overlayed since these are implemented using
2827 external functions, e.g.:
2837 void set_error(unsigned char errcd)
2853 void some_isr () interrupt 2 using 1
2882 In the above example the parameter
2890 would be assigned to the overlayable segment if the #pragma\SpecialChar ~
2892 not present, this could cause unpredictable runtime behavior when called
2894 The #pragma\SpecialChar ~
2895 NOOVERLAY ensures that the parameters and local variables for
2896 the function are NOT overlayed.
2899 Interrupt Service Routines
2902 SDCC allows interrupt service routines to be coded in C, with some extended
2909 void timer_isr (void) interrupt 2 using 1
2922 The number following the
2926 keyword is the interrupt number this routine will service.
2927 The compiler will insert a call to this routine in the interrupt vector
2928 table for the interrupt number specified.
2933 keyword is used to tell the compiler to use the specified register bank
2934 (8051 specific) when generating code for this function.
2935 Note that when some function is called from an interrupt service routine
2936 it should be preceded by a #pragma\SpecialChar ~
2937 NOOVERLAY if it is not reentrant.
2938 A special note here, int (16 bit) and long (32 bit) integer division, multiplic
2939 ation & modulus operations are implemented using external support routines
2940 developed in ANSI-C, if an interrupt service routine needs to do any of
2941 these operations then the support routines (as mentioned in a following
2942 section) will have to be recompiled using the
2946 option and the source file will need to be compiled using the
2953 If you have multiple source files in your project, interrupt service routines
2954 can be present in any of them, but a prototype of the isr MUST be present
2955 or included in the file that contains the function
2962 Interrupt Numbers and the corresponding address & descriptions for the Standard
2963 8051 are listed below.
2964 SDCC will automatically adjust the interrupt vector table to the maximum
2965 interrupt number specified.
2971 \begin_inset Tabular
2972 <lyxtabular version="3" rows="6" columns="3">
2974 <column alignment="center" valignment="top" leftline="true" width="0(null)">
2975 <column alignment="center" valignment="top" leftline="true" width="0(null)">
2976 <column alignment="center" valignment="top" leftline="true" rightline="true" width="0(null)">
2977 <row topline="true" bottomline="true">
2978 <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
2986 <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
2994 <cell alignment="center" valignment="top" topline="true" leftline="true" rightline="true" usebox="none">
3003 <row topline="true">
3004 <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
3012 <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
3020 <cell alignment="center" valignment="top" topline="true" leftline="true" rightline="true" usebox="none">
3029 <row topline="true">
3030 <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
3038 <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
3046 <cell alignment="center" valignment="top" topline="true" leftline="true" rightline="true" usebox="none">
3055 <row topline="true">
3056 <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
3064 <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
3072 <cell alignment="center" valignment="top" topline="true" leftline="true" rightline="true" usebox="none">
3081 <row topline="true">
3082 <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
3090 <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
3098 <cell alignment="center" valignment="top" topline="true" leftline="true" rightline="true" usebox="none">
3107 <row topline="true" bottomline="true">
3108 <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
3116 <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
3124 <cell alignment="center" valignment="top" topline="true" leftline="true" rightline="true" usebox="none">
3141 If the interrupt service routine is defined without
3145 a register bank or with register bank 0 (using 0), the compiler will save
3146 the registers used by itself on the stack upon entry and restore them at
3147 exit, however if such an interrupt service routine calls another function
3148 then the entire register bank will be saved on the stack.
3149 This scheme may be advantageous for small interrupt service routines which
3150 have low register usage.
3153 If the interrupt service routine is defined to be using a specific register
3158 are save and restored, if such an interrupt service routine calls another
3159 function (using another register bank) then the entire register bank of
3160 the called function will be saved on the stack.
3161 This scheme is recommended for larger interrupt service routines.
3164 Calling other functions from an interrupt service routine is not recommended,
3165 avoid it if possible.
3169 Also see the _naked modifier.
3177 <TODO: this isn't implemented at all!>
3183 A special keyword may be associated with a function declaring it as
3188 SDCC will generate code to disable all interrupts upon entry to a critical
3189 function and enable them back before returning.
3190 Note that nesting critical functions may cause unpredictable results.
3215 The critical attribute maybe used with other attributes like
3223 A special keyword may be associated with a function declaring it as
3232 function modifier attribute prevents the compiler from generating prologue
3233 and epilogue code for that function.
3234 This means that the user is entirely responsible for such things as saving
3235 any registers that may need to be preserved, selecting the proper register
3236 bank, generating the
3240 instruction at the end, etc.
3241 Practically, this means that the contents of the function must be written
3242 in inline assembler.
3243 This is particularly useful for interrupt functions, which can have a large
3244 (and often unnecessary) prologue/epilogue.
3245 For example, compare the code generated by these two functions:
3251 data unsigned char counter;
3253 void simpleInterrupt(void) interrupt 1
3267 void nakedInterrupt(void) interrupt 2 _naked
3300 ; MUST explicitly include ret in _naked function.
3314 For an 8051 target, the generated simpleInterrupt looks like:
3459 whereas nakedInterrupt looks like:
3484 ; MUST explicitly include ret(i) in _naked function.
3490 While there is nothing preventing you from writing C code inside a _naked
3491 function, there are many ways to shoot yourself in the foot doing this,
3492 and is is recommended that you stick to inline assembler.
3495 Functions using private banks
3502 attribute (which tells the compiler to use a register bank other than the
3503 default bank zero) should only be applied to
3507 functions (see note 1 below).
3508 This will in most circumstances make the generated ISR code more efficient
3509 since it will not have to save registers on the stack.
3516 attribute will have no effect on the generated code for a
3520 function (but may occasionally be useful anyway
3526 possible exception: if a function is called ONLY from 'interrupt' functions
3527 using a particular bank, it can be declared with the same 'using' attribute
3528 as the calling 'interrupt' functions.
3529 For instance, if you have several ISRs using bank one, and all of them
3530 call memcpy(), it might make sense to create a specialized version of memcpy()
3531 'using 1', since this would prevent the ISR from having to save bank zero
3532 to the stack on entry and switch to bank zero before calling the function
3539 (pending: I don't think this has been done yet)
3546 function using a non-zero bank will assume that it can trash that register
3547 bank, and will not save it.
3548 Since high-priority interrupts can interrupt low-priority ones on the 8051
3549 and friends, this means that if a high-priority ISR
3553 a particular bank occurs while processing a low-priority ISR
3557 the same bank, terrible and bad things can happen.
3558 To prevent this, no single register bank should be
3562 by both a high priority and a low priority ISR.
3563 This is probably most easily done by having all high priority ISRs use
3564 one bank and all low priority ISRs use another.
3565 If you have an ISR which can change priority at runtime, you're on your
3566 own: I suggest using the default bank zero and taking the small performance
3570 It is most efficient if your ISR calls no other functions.
3571 If your ISR must call other functions, it is most efficient if those functions
3572 use the same bank as the ISR (see note 1 below); the next best is if the
3573 called functions use bank zero.
3574 It is very inefficient to call a function using a different, non-zero bank
3582 Data items can be assigned an absolute address with the
3586 keyword, in addition to a storage class, e.g.:
3592 xdata at 0x8000 unsigned char PORTA_8255 ;
3598 In the above example the PORTA_8255 will be allocated to the location 0x8000
3599 of the external ram.
3600 Note that this feature is provided to give the programmer access to
3604 devices attached to the controller.
3605 The compiler does not actually reserve any space for variables declared
3606 in this way (they are implemented with an equate in the assembler).
3607 Thus it is left to the programmer to make sure there are no overlaps with
3608 other variables that are declared without the absolute address.
3609 The assembler listing file (.lst) and the linker output files (.rst) and
3610 (.map) are a good places to look for such overlaps.
3614 Absolute address can be specified for variables in all storage classes,
3627 The above example will allocate the variable at offset 0x02 in the bit-addressab
3629 There is no real advantage to assigning absolute addresses to variables
3630 in this manner, unless you want strict control over all the variables allocated.
3636 The compiler inserts a call to the C routine
3638 _sdcc__external__startup()
3643 at the start of the CODE area.
3644 This routine is in the runtime library.
3645 By default this routine returns 0, if this routine returns a non-zero value,
3646 the static & global variable initialization will be skipped and the function
3647 main will be invoked Other wise static & global variables will be initialized
3648 before the function main is invoked.
3651 _sdcc__external__startup()
3653 routine to your program to override the default if you need to setup hardware
3654 or perform some other critical operation prior to static & global variable
3658 Inline Assembler Code
3661 SDCC allows the use of in-line assembler with a few restriction as regards
3663 All labels defined within inline assembler code
3671 where nnnn is a number less than 100 (which implies a limit of utmost 100
3672 inline assembler labels
3680 It is strongly recommended that each assembly instruction (including labels)
3681 be placed in a separate line (as the example shows).
3686 command line option is used, the inline assembler code will be passed through
3687 the peephole optimizer.
3688 This might cause some unexpected changes in the inline assembler code.
3689 Please go throught the peephole optimizer rules defined in file
3693 carefully before using this option.
3733 The inline assembler code can contain any valid code understood by the assembler
3734 , this includes any assembler directives and comment lines.
3735 The compiler does not do any validation of the code within the
3745 Inline assembler code cannot reference any C-Labels, however it can reference
3746 labels defined by the inline assembler, e.g.:
3772 ; some assembler code
3792 /* some more c code */
3794 clabel:\SpecialChar ~
3796 /* inline assembler cannot reference this label */
3808 $0003: ;label (can be reference by inline assembler only)
3820 /* some more c code */
3828 In other words inline assembly code can access labels defined in inline
3829 assembly within the scope of the funtion.
3833 The same goes the other way, ie.
3834 labels defines in inline assembly CANNOT be accessed by C statements.
3837 int(16 bit) and long (32 bit) Support
3840 For signed & unsigned int (16 bit) and long (32 bit) variables, division,
3841 multiplication and modulus operations are implemented by support routines.
3842 These support routines are all developed in ANSI-C to facilitate porting
3843 to other MCUs, although some model specific assembler optimations are used.
3844 The following files contain the described routine, all of them can be found
3845 in <installdir>/share/sdcc/lib.
3851 <pending: tabularise this>
3857 _mulsint.c - signed 16 bit multiplication (calls _muluint)
3859 _muluint.c - unsigned 16 bit multiplication
3861 _divsint.c - signed 16 bit division (calls _divuint)
3863 _divuint.c - unsigned 16 bit division
3865 _modsint.c - signed 16 bit modulus (call _moduint)
3867 _moduint.c - unsigned 16 bit modulus
3869 _mulslong.c - signed 32 bit multiplication (calls _mululong)
3871 _mululong.c - unsigned32 bit multiplication
3873 _divslong.c - signed 32 division (calls _divulong)
3875 _divulong.c - unsigned 32 division
3877 _modslong.c - signed 32 bit modulus (calls _modulong)
3879 _modulong.c - unsigned 32 bit modulus
3887 Since they are compiled as
3891 , interrupt service routines should not do any of the above operations.
3892 If this is unavoidable then the above routines will need to be compiled
3897 option, after which the source program will have to be compiled with
3904 Floating Point Support
3907 SDCC supports IEEE (single precision 4bytes) floating point numbers.The floating
3908 point support routines are derived from gcc's floatlib.c and consists of
3909 the following routines:
3915 <pending: tabularise this>
3921 _fsadd.c - add floating point numbers
3923 _fssub.c - subtract floating point numbers
3925 _fsdiv.c - divide floating point numbers
3927 _fsmul.c - multiply floating point numbers
3929 _fs2uchar.c - convert floating point to unsigned char
3931 _fs2char.c - convert floating point to signed char
3933 _fs2uint.c - convert floating point to unsigned int
3935 _fs2int.c - convert floating point to signed int
3937 _fs2ulong.c - convert floating point to unsigned long
3939 _fs2long.c - convert floating point to signed long
3941 _uchar2fs.c - convert unsigned char to floating point
3943 _char2fs.c - convert char to floating point number
3945 _uint2fs.c - convert unsigned int to floating point
3947 _int2fs.c - convert int to floating point numbers
3949 _ulong2fs.c - convert unsigned long to floating point number
3951 _long2fs.c - convert long to floating point number
3959 Note if all these routines are used simultaneously the data space might
3961 For serious floating point usage it is strongly recommended that the large
3968 SDCC allows two memory models for MCS51 code, small and large.
3969 Modules compiled with different memory models should
3973 be combined together or the results would be unpredictable.
3974 The library routines supplied with the compiler are compiled as both small
3976 The compiled library modules are contained in seperate directories as small
3977 and large so that you can link to either set.
3981 When the large model is used all variables declared without a storage class
3982 will be allocated into the external ram, this includes all parameters and
3983 local variables (for non-reentrant functions).
3984 When the small model is used variables without storage class are allocated
3985 in the internal ram.
3988 Judicious usage of the processor specific storage classes and the 'reentrant'
3989 function type will yield much more efficient code, than using the large
3991 Several optimizations are disabled when the program is compiled using the
3992 large model, it is therefore strongly recommdended that the small model
3993 be used unless absolutely required.
3999 The only model supported is Flat 24.
4000 This generates code for the 24 bit contiguous addressing mode of the Dallas
4002 In this mode, up to four meg of external RAM or code space can be directly
4004 See the data sheets at www.dalsemi.com for further information on this part.
4008 In older versions of the compiler, this option was used with the MCS51 code
4014 Now, however, the '390 has it's own code generator, selected by the
4023 Note that the compiler does not generate any code to place the processor
4024 into 24 bitmode (although
4028 in the ds390 libraries will do that for you).
4033 , the boot loader or similar code must ensure that the processor is in 24
4034 bit contiguous addressing mode before calling the SDCC startup code.
4042 option, variables will by default be placed into the XDATA segment.
4047 Segments may be placed anywhere in the 4 meg address space using the usual
4049 Note that if any segments are located above 64K, the -r flag must be passed
4050 to the linker to generate the proper segment relocations, and the Intel
4051 HEX output format must be used.
4052 The -r flag can be passed to the linker by using the option
4056 on the sdcc command line.
4057 However, currently the linker can not handle code segments > 64k.
4060 Defines Created by the Compiler
4063 The compiler creates the following #defines.
4066 SDCC - this Symbol is always defined.
4069 SDCC_mcs51 or SDCC_ds390 or SDCC_z80, etc - depending on the model used
4073 __mcs51 or __ds390 or __z80, etc - depending on the model used (e.g.
4077 SDCC_STACK_AUTO - this symbol is defined when
4084 SDCC_MODEL_SMALL - when
4091 SDCC_MODEL_LARGE - when
4098 SDCC_USE_XSTACK - when
4105 SDCC_STACK_TENBIT - when
4112 SDCC_MODEL_FLAT24 - when
4125 SDCC performs a host of standard optimizations in addition to some MCU specific
4128 \layout Subsubsection
4130 Sub-expression Elimination
4133 The compiler does local and global common subexpression elimination, e.g.:
4148 will be translated to
4164 Some subexpressions are not as obvious as the above example, e.g.:
4178 In this case the address arithmetic a->b[i] will be computed only once;
4179 the equivalent code in C would be.
4195 The compiler will try to keep these temporary variables in registers.
4196 \layout Subsubsection
4198 Dead-Code Elimination
4213 i = 1; \SpecialChar ~
4218 global = 1;\SpecialChar ~
4231 global = 3;\SpecialChar ~
4246 int global; void f ()
4259 \layout Subsubsection
4320 Note: the dead stores created by this copy propagation will be eliminated
4321 by dead-code elimination.
4322 \layout Subsubsection
4327 Two types of loop optimizations are done by SDCC loop invariant lifting
4328 and strength reduction of loop induction variables.
4329 In addition to the strength reduction the optimizer marks the induction
4330 variables and the register allocator tries to keep the induction variables
4331 in registers for the duration of the loop.
4332 Because of this preference of the register allocator, loop induction optimizati
4333 on causes an increase in register pressure, which may cause unwanted spilling
4334 of other temporary variables into the stack / data space.
4335 The compiler will generate a warning message when it is forced to allocate
4336 extra space either on the stack or data space.
4337 If this extra space allocation is undesirable then induction optimization
4338 can be eliminated either for the entire source file (with --noinduction
4339 option) or for a given function only using #pragma\SpecialChar ~
4350 for (i = 0 ; i < 100 ; i ++)
4368 for (i = 0; i < 100; i++)
4378 As mentioned previously some loop invariants are not as apparent, all static
4379 address computations are also moved out of the loop.
4383 Strength Reduction, this optimization substitutes an expression by a cheaper
4390 for (i=0;i < 100; i++)
4410 for (i=0;i< 100;i++) {
4414 ar[itemp1] = itemp2;
4430 The more expensive multiplication is changed to a less expensive addition.
4431 \layout Subsubsection
4436 This optimization is done to reduce the overhead of checking loop boundaries
4437 for every iteration.
4438 Some simple loops can be reversed and implemented using a
4439 \begin_inset Quotes eld
4442 decrement and jump if not zero
4443 \begin_inset Quotes erd
4447 SDCC checks for the following criterion to determine if a loop is reversible
4448 (note: more sophisticated compilers use data-dependency analysis to make
4449 this determination, SDCC uses a more simple minded analysis).
4452 The 'for' loop is of the form
4458 for (<symbol> = <expression> ; <sym> [< | <=] <expression> ; [<sym>++ |
4468 The <for body> does not contain
4469 \begin_inset Quotes eld
4473 \begin_inset Quotes erd
4477 \begin_inset Quotes erd
4483 All goto's are contained within the loop.
4486 No function calls within the loop.
4489 The loop control variable <sym> is not assigned any value within the loop
4492 The loop control variable does NOT participate in any arithmetic operation
4496 There are NO switch statements in the loop.
4497 \layout Subsubsection
4499 Algebraic Simplifications
4502 SDCC does numerous algebraic simplifications, the following is a small sub-set
4503 of these optimizations.
4509 i = j + 0 ; /* changed to */ i = j;
4511 i /= 2; /* changed to */ i >>= 1;
4513 i = j - j ; /* changed to */ i = 0;
4515 i = j / 1 ; /* changed to */ i = j;
4521 Note the subexpressions given above are generally introduced by macro expansions
4522 or as a result of copy/constant propagation.
4523 \layout Subsubsection
4528 SDCC changes switch statements to jump tables when the following conditions
4533 The case labels are in numerical sequence, the labels need not be in order,
4534 and the starting number need not be one or zero.
4540 switch(i) {\SpecialChar ~
4647 Both the above switch statements will be implemented using a jump-table.
4650 The number of case labels is at least three, since it takes two conditional
4651 statements to handle the boundary conditions.
4654 The number of case labels is less than 84, since each label takes 3 bytes
4655 and a jump-table can be utmost 256 bytes long.
4659 Switch statements which have gaps in the numeric sequence or those that
4660 have more that 84 case labels can be split into more than one switch statement
4661 for efficient code generation, e.g.:
4699 If the above switch statement is broken down into two switch statements
4733 case 9: \SpecialChar ~
4743 case 12:\SpecialChar ~
4753 then both the switch statements will be implemented using jump-tables whereas
4754 the unmodified switch statement will not be.
4755 \layout Subsubsection
4757 Bit-shifting Operations.
4760 Bit shifting is one of the most frequently used operation in embedded programmin
4762 SDCC tries to implement bit-shift operations in the most efficient way
4782 generates the following code:
4800 In general SDCC will never setup a loop if the shift count is known.
4840 Note that SDCC stores numbers in little-endian format (i.e.
4841 lowest order first).
4842 \layout Subsubsection
4847 A special case of the bit-shift operation is bit rotation, SDCC recognizes
4848 the following expression to be a left bit-rotation:
4859 i = ((i << 1) | (i >> 7));
4867 will generate the following code:
4883 SDCC uses pattern matching on the parse tree to determine this operation.Variatio
4884 ns of this case will also be recognized as bit-rotation, i.e.:
4890 i = ((i >> 7) | (i << 1)); /* left-bit rotation */
4891 \layout Subsubsection
4896 It is frequently required to obtain the highest order bit of an integral
4897 type (long, int, short or char types).
4898 SDCC recognizes the following expression to yield the highest order bit
4899 and generates optimized code for it, e.g.:
4920 hob = (gint >> 15) & 1;
4933 will generate the following code:
4972 000A E5*01\SpecialChar ~
5000 000C 33\SpecialChar ~
5031 000D E4\SpecialChar ~
5062 000E 13\SpecialChar ~
5093 000F F5*02\SpecialChar ~
5123 Variations of this case however will
5128 It is a standard C expression, so I heartily recommend this be the only
5129 way to get the highest order bit, (it is portable).
5130 Of course it will be recognized even if it is embedded in other expressions,
5137 xyz = gint + ((gint >> 15) & 1);
5143 will still be recognized.
5144 \layout Subsubsection
5149 The compiler uses a rule based, pattern matching and re-writing mechanism
5150 for peep-hole optimization.
5155 a peep-hole optimizer by Christopher W.
5156 Fraser (cwfraser@microsoft.com).
5157 A default set of rules are compiled into the compiler, additional rules
5158 may be added with the
5160 --peep-file <filename>
5163 The rule language is best illustrated with examples.
5191 The above rule will change the following assembly sequence:
5221 Note: All occurrences of a
5225 (pattern variable) must denote the same string.
5226 With the above rule, the assembly sequence:
5244 will remain unmodified.
5248 Other special case optimizations may be added by the user (via
5254 some variants of the 8051 MCU allow only
5263 The following two rules will change all
5285 replace { lcall %1 } by { acall %1 }
5287 replace { ljmp %1 } by { ajmp %1 }
5295 inline-assembler code
5297 is also passed through the peep hole optimizer, thus the peephole optimizer
5298 can also be used as an assembly level macro expander.
5299 The rules themselves are MCU dependent whereas the rule language infra-structur
5300 e is MCU independent.
5301 Peephole optimization rules for other MCU can be easily programmed using
5306 The syntax for a rule is as follows:
5312 rule := replace [ restart ] '{' <assembly sequence> '
5350 <assembly sequence> '
5368 '}' [if <functionName> ] '
5376 <assembly sequence> := assembly instruction (each instruction including
5377 labels must be on a separate line).
5381 The optimizer will apply to the rules one by one from the top in the sequence
5382 of their appearance, it will terminate when all rules are exhausted.
5383 If the 'restart' option is specified, then the optimizer will start matching
5384 the rules again from the top, this option for a rule is expensive (performance)
5385 , it is intended to be used in situations where a transformation will trigger
5386 the same rule again.
5387 An example of this (not a good one, it has side effects) is the following
5414 Note that the replace pattern cannot be a blank, but can be a comment line.
5415 Without the 'restart' option only the inner most 'pop' 'push' pair would
5416 be eliminated, i.e.:
5468 the restart option the rule will be applied again to the resulting code
5469 and then all the pop-push pairs will be eliminated to yield:
5487 A conditional function can be attached to a rule.
5488 Attaching rules are somewhat more involved, let me illustrate this with
5519 The optimizer does a look-up of a function name table defined in function
5524 in the source file SDCCpeeph.c, with the name
5529 If it finds a corresponding entry the function is called.
5530 Note there can be no parameters specified for these functions, in this
5535 is crucial, since the function
5539 expects to find the label in that particular variable (the hash table containin
5540 g the variable bindings is passed as a parameter).
5541 If you want to code more such functions, take a close look at the function
5542 labelInRange and the calling mechanism in source file SDCCpeeph.c.
5543 I know this whole thing is a little kludgey, but maybe some day we will
5544 have some better means.
5545 If you are looking at this file, you will also see the default rules that
5546 are compiled into the compiler, you can add your own rules in the default
5547 set there if you get tired of specifying the --peep-file option.
5553 SDCC supports the following #pragma directives.
5554 This directives are applicable only at a function level.
5557 SAVE - this will save all the current options.
5560 RESTORE - will restore the saved options from the last save.
5561 Note that SAVES & RESTOREs cannot be nested.
5562 SDCC uses the same buffer to save the options each time a SAVE is called.
5565 NOGCSE - will stop global subexpression elimination.
5568 NOINDUCTION - will stop loop induction optimizations.
5571 NOJTBOUND - will not generate code for boundary value checking, when switch
5572 statements are turned into jump-tables.
5575 NOOVERLAY - the compiler will not overlay the parameters and local variables
5579 NOLOOPREVERSE - Will not do loop reversal optimization
5582 EXCLUDE NONE | {acc[,b[,dpl[,dph]]] - The exclude pragma disables generation
5583 of pair of push/pop instruction in ISR function (using interrupt keyword).
5584 The directive should be placed immediately before the ISR function definition
5585 and it affects ALL ISR functions following it.
5586 To enable the normal register saving for ISR functions use #pragma\SpecialChar ~
5587 EXCLUDE\SpecialChar ~
5591 CALLEE-SAVES function1[,function2[,function3...]] - The compiler by default
5592 uses a caller saves convention for register saving across function calls,
5593 however this can cause unneccessary register pushing & popping when calling
5594 small functions from larger functions.
5595 This option can be used to switch the register saving convention for the
5596 function names specified.
5597 The compiler will not save registers when calling these functions, extra
5598 code will be generated at the entry & exit for these functions to save
5599 & restore the registers used by these functions, this can SUBSTANTIALLY
5600 reduce code & improve run time performance of the generated code.
5601 In future the compiler (with interprocedural analysis) will be able to
5602 determine the appropriate scheme to use for each function call.
5603 If --callee-saves command line option is used, the function names specified
5604 in #pragma\SpecialChar ~
5605 CALLEE-SAVES is appended to the list of functions specified inthe
5609 The pragma's are intended to be used to turn-off certain optimizations which
5610 might cause the compiler to generate extra stack / data space to store
5611 compiler generated temporary variables.
5612 This usually happens in large functions.
5613 Pragma directives should be used as shown in the following example, they
5614 are used to control options & optimizations for a given function; pragmas
5615 should be placed before and/or after a function, placing pragma's inside
5616 a function body could have unpredictable results.
5622 #pragma SAVE /* save the current settings */
5624 #pragma NOGCSE /* turnoff global subexpression elimination */
5626 #pragma NOINDUCTION /* turn off induction optimizations */
5648 #pragma RESTORE /* turn the optimizations back on */
5654 The compiler will generate a warning message when extra space is allocated.
5655 It is strongly recommended that the SAVE and RESTORE pragma's be used when
5656 changing options for a function.
5661 <pending: this is messy and incomplete>
5666 Compiler support routines (_gptrget, _mulint etc)
5669 Stdclib functions (puts, printf, strcat etc)
5672 Math functions (sin, pow, sqrt etc)
5675 Interfacing with Assembly Routines
5676 \layout Subsubsection
5678 Global Registers used for Parameter Passing
5681 The compiler always uses the global registers
5689 to pass the first parameter to a routine.
5690 The second parameter onwards is either allocated on the stack (for reentrant
5691 routines or if --stack-auto is used) or in the internal / external ram
5692 (depending on the memory model).
5694 \layout Subsubsection
5696 Assembler Routine(non-reentrant)
5699 In the following example the function cfunc calls an assembler routine asm_func,
5700 which takes two parameters.
5706 extern int asm_func(unsigned char, unsigned char);
5710 int c_func (unsigned char i, unsigned char j)
5718 return asm_func(i,j);
5732 return c_func(10,9);
5740 The corresponding assembler function is:
5746 .globl _asm_func_PARM_2
5810 add a,_asm_func_PARM_2
5846 Note here that the return values are placed in 'dpl' - One byte return value,
5847 'dpl' LSB & 'dph' MSB for two byte values.
5848 'dpl', 'dph' and 'b' for three byte values (generic pointers) and 'dpl','dph','
5849 b' & 'acc' for four byte values.
5852 The parameter naming convention is _<function_name>_PARM_<n>, where n is
5853 the parameter number starting from 1, and counting from the left.
5854 The first parameter is passed in
5855 \begin_inset Quotes eld
5859 \begin_inset Quotes erd
5862 for One bye parameter,
5863 \begin_inset Quotes eld
5867 \begin_inset Quotes erd
5871 \begin_inset Quotes eld
5875 \begin_inset Quotes erd
5879 \begin_inset Quotes eld
5883 \begin_inset Quotes erd
5886 for four bytes, the varible name for the second parameter will be _<function_na
5891 Assemble the assembler routine with the following command:
5898 asx8051 -losg asmfunc.asm
5905 Then compile and link the assembler routine to the C source file with the
5913 sdcc cfunc.c asmfunc.rel
5914 \layout Subsubsection
5916 Assembler Routine(reentrant)
5919 In this case the second parameter onwards will be passed on the stack, the
5920 parameters are pushed from right to left i.e.
5921 after the call the left most parameter will be on the top of the stack.
5928 extern int asm_func(unsigned char, unsigned char);
5932 int c_func (unsigned char i, unsigned char j) reentrant
5940 return asm_func(i,j);
5954 return c_func(10,9);
5962 The corresponding assembler routine is:
6072 The compiling and linking procedure remains the same, however note the extra
6073 entry & exit linkage required for the assembler code, _bp is the stack
6074 frame pointer and is used to compute the offset into the stack for parameters
6075 and local variables.
6081 The external stack is located at the start of the external ram segment,
6082 and is 256 bytes in size.
6083 When --xstack option is used to compile the program, the parameters and
6084 local variables of all reentrant functions are allocated in this area.
6085 This option is provided for programs with large stack space requirements.
6086 When used with the --stack-auto option, all parameters and local variables
6087 are allocated on the external stack (note support libraries will need to
6088 be recompiled with the same options).
6091 The compiler outputs the higher order address byte of the external ram segment
6092 into PORT P2, therefore when using the External Stack option, this port
6093 MAY NOT be used by the application program.
6099 Deviations from the compliancy.
6102 functions are not always reentrant.
6105 structures cannot be assigned values directly, cannot be passed as function
6106 parameters or assigned to each other and cannot be a return value from
6133 s1 = s2 ; /* is invalid in SDCC although allowed in ANSI */
6144 struct s foo1 (struct s parms) /* is invalid in SDCC although allowed in
6166 return rets;/* is invalid in SDCC although allowed in ANSI */
6171 'long long' (64 bit integers) not supported.
6174 'double' precision floating point not supported.
6177 No support for setjmp and longjmp (for now).
6180 Old K&R style function declarations are NOT allowed.
6186 foo(i,j) /* this old style of function declarations */
6188 int i,j; /* are valid in ANSI but not valid in SDCC */
6202 functions declared as pointers must be dereferenced during the call.
6213 /* has to be called like this */
6215 (*foo)(); /* ansi standard allows calls to be made like 'foo()' */
6218 Cyclomatic Complexity
6221 Cyclomatic complexity of a function is defined as the number of independent
6222 paths the program can take during execution of the function.
6223 This is an important number since it defines the number test cases you
6224 have to generate to validate the function.
6225 The accepted industry standard for complexity number is 10, if the cyclomatic
6226 complexity reported by SDCC exceeds 10 you should think about simplification
6227 of the function logic.
6228 Note that the complexity level is not related to the number of lines of
6230 Large functions can have low complexity, and small functions can have large
6236 SDCC uses the following formula to compute the complexity:
6241 complexity = (number of edges in control flow graph) - (number of nodes
6242 in control flow graph) + 2;
6246 Having said that the industry standard is 10, you should be aware that in
6247 some cases it be may unavoidable to have a complexity level of less than
6249 For example if you have switch statement with more than 10 case labels,
6250 each case label adds one to the complexity level.
6251 The complexity level is by no means an absolute measure of the algorithmic
6252 complexity of the function, it does however provide a good starting point
6253 for which functions you might look at for further optimization.
6259 Here are a few guidelines that will help the compiler generate more efficient
6260 code, some of the tips are specific to this compiler others are generally
6261 good programming practice.
6264 Use the smallest data type to represent your data-value.
6265 If it is known in advance that the value is going to be less than 256 then
6266 use a 'char' instead of a 'short' or 'int'.
6269 Use unsigned when it is known in advance that the value is not going to
6271 This helps especially if you are doing division or multiplication.
6274 NEVER jump into a LOOP.
6277 Declare the variables to be local whenever possible, especially loop control
6278 variables (induction).
6281 Since the compiler does not do implicit integral promotion, the programmer
6282 should do an explicit cast when integral promotion is required.
6285 Reducing the size of division, multiplication & modulus operations can reduce
6286 code size substantially.
6287 Take the following code for example.
6293 foobar(unsigned int p1, unsigned char ch)
6297 unsigned char ch1 = p1 % ch ;
6308 For the modulus operation the variable ch will be promoted to unsigned int
6309 first then the modulus operation will be performed (this will lead to a
6310 call to support routine _moduint()), and the result will be casted to a
6312 If the code is changed to
6318 foobar(unsigned int p1, unsigned char ch)
6322 unsigned char ch1 = (unsigned char)p1 % ch ;
6333 It would substantially reduce the code generated (future versions of the
6334 compiler will be smart enough to detect such optimization oppurtunities).
6337 Notes on MCS51 memory layout
6340 The 8051 family of micro controller have a minimum of 128 bytes of internal
6341 memory which is structured as follows
6345 - Bytes 00-1F - 32 bytes to hold up to 4 banks of the registers R7 to R7
6348 - Bytes 20-2F - 16 bytes to hold 128 bit variables and
6350 - Bytes 30-7F - 60 bytes for general purpose use.
6354 Normally the SDCC compiler will only utilise the first bank of registers,
6355 but it is possible to specify that other banks of registers should be used
6356 in interrupt routines.
6357 By default, the compiler will place the stack after the last bank of used
6359 if the first 2 banks of registers are used, it will position the base of
6360 the internal stack at address 16 (0X10).
6361 This implies that as the stack grows, it will use up the remaining register
6362 banks, and the 16 bytes used by the 128 bit variables, and 60 bytes for
6363 general purpose use.
6366 By default, the compiler uses the 60 general purpose bytes to hold "near
6368 The compiler/optimiser may also declare some Local Variables in this area
6373 If any of the 128 bit variables are used, or near data is being used then
6374 care needs to be taken to ensure that the stack does not grow so much that
6375 it starts to over write either your bit variables or "near data".
6376 There is no runtime checking to prevent this from happening.
6379 The amount of stack being used is affected by the use of the "internal stack"
6380 to save registers before a subroutine call is made (--stack-auto will declare
6381 parameters and local variables on the stack) and the number of nested subroutin
6385 If you detect that the stack is over writing you data, then the following
6387 --xstack will cause an external stack to be used for saving registers and
6388 (if --stack-auto is being used) storing parameters and local variables.
6389 However this will produce more code which will be slower to execute.
6393 --stack-loc will allow you specify the start of the stack, i.e.
6394 you could start it after any data in the general purpose area.
6395 However this may waste the memory not used by the register banks and if
6396 the size of the "near data" increases, it may creep into the bottom of
6400 --stack-after-data, similar to the --stack-loc, but it automatically places
6401 the stack after the end of the "near data".
6402 Again this could waste any spare register space.
6405 --data-loc allows you to specify the start address of the near data.
6406 This could be used to move the "near data" further away from the stack
6407 giving it more room to grow.
6408 This will only work if no bit variables are being used and the stack can
6409 grow to use the bit variable space.
6417 If you find that the stack is over writing your bit variables or "near data"
6418 then the approach which best utilised the internal memory is to position
6419 the "near data" after the last bank of used registers or, if you use bit
6420 variables, after the last bit variable by using the --data-loc, e.g.
6421 if two register banks are being used and no bit variables, --data-loc 16,
6422 and use the --stack-after-data option.
6425 If bit variables are being used, another method would be to try and squeeze
6426 the data area in the unused register banks if it will fit, and start the
6427 stack after the last bit variable.
6430 Retargetting for other MCUs.
6433 The issues for retargetting the compiler are far too numerous to be covered
6435 What follows is a brief description of each of the seven phases of the
6436 compiler and its MCU dependency.
6439 Parsing the source and building the annotated parse tree.
6440 This phase is largely MCU independent (except for the language extensions).
6441 Syntax & semantic checks are also done in this phase, along with some initial
6442 optimizations like back patching labels and the pattern matching optimizations
6443 like bit-rotation etc.
6446 The second phase involves generating an intermediate code which can be easy
6447 manipulated during the later phases.
6448 This phase is entirely MCU independent.
6449 The intermediate code generation assumes the target machine has unlimited
6450 number of registers, and designates them with the name iTemp.
6451 The compiler can be made to dump a human readable form of the code generated
6452 by using the --dumpraw option.
6455 This phase does the bulk of the standard optimizations and is also MCU independe
6457 This phase can be broken down into several sub-phases:
6461 Break down intermediate code (iCode) into basic blocks.
6463 Do control flow & data flow analysis on the basic blocks.
6465 Do local common subexpression elimination, then global subexpression elimination
6467 Dead code elimination
6471 If loop optimizations caused any changes then do 'global subexpression eliminati
6472 on' and 'dead code elimination' again.
6475 This phase determines the live-ranges; by live range I mean those iTemp
6476 variables defined by the compiler that still survive after all the optimization
6478 Live range analysis is essential for register allocation, since these computati
6479 on determines which of these iTemps will be assigned to registers, and for
6483 Phase five is register allocation.
6484 There are two parts to this process.
6488 The first part I call 'register packing' (for lack of a better term).
6489 In this case several MCU specific expression folding is done to reduce
6494 The second part is more MCU independent and deals with allocating registers
6495 to the remaining live ranges.
6496 A lot of MCU specific code does creep into this phase because of the limited
6497 number of index registers available in the 8051.
6500 The Code generation phase is (unhappily), entirely MCU dependent and very
6501 little (if any at all) of this code can be reused for other MCU.
6502 However the scheme for allocating a homogenized assembler operand for each
6503 iCode operand may be reused.
6506 As mentioned in the optimization section the peep-hole optimizer is rule
6507 based system, which can reprogrammed for other MCUs.
6510 SDCDB - Source Level Debugger
6513 SDCC is distributed with a source level debugger.
6514 The debugger uses a command line interface, the command repertoire of the
6515 debugger has been kept as close to gdb (the GNU debugger) as possible.
6516 The configuration and build process is part of the standard compiler installati
6517 on, which also builds and installs the debugger in the target directory
6518 specified during configuration.
6519 The debugger allows you debug BOTH at the C source and at the ASM source
6523 Compiling for Debugging
6528 debug option must be specified for all files for which debug information
6530 The complier generates a .cdb file for each of these files.
6531 The linker updates the .cdb file with the address information.
6532 This .cdb is used by the debugger.
6535 How the Debugger Works
6538 When the --debug option is specified the compiler generates extra symbol
6539 information some of which are put into the the assembler source and some
6540 are put into the .cdb file, the linker updates the .cdb file with the address
6541 information for the symbols.
6542 The debugger reads the symbolic information generated by the compiler &
6543 the address information generated by the linker.
6544 It uses the SIMULATOR (Daniel's S51) to execute the program, the program
6545 execution is controlled by the debugger.
6546 When a command is issued for the debugger, it translates it into appropriate
6547 commands for the simulator.
6550 Starting the Debugger
6553 The debugger can be started using the following command line.
6554 (Assume the file you are debugging has the file name foo).
6568 The debugger will look for the following files.
6571 foo.c - the source file.
6574 foo.cdb - the debugger symbol information file.
6577 foo.ihx - the intel hex format object file.
6580 Command Line Options.
6583 --directory=<source file directory> this option can used to specify the
6584 directory search list.
6585 The debugger will look into the directory list specified for source, cdb
6587 The items in the directory list must be separated by ':', e.g.
6588 if the source files can be in the directories /home/src1 and /home/src2,
6589 the --directory option should be --directory=/home/src1:/home/src2.
6590 Note there can be no spaces in the option.
6594 -cd <directory> - change to the <directory>.
6597 -fullname - used by GUI front ends.
6600 -cpu <cpu-type> - this argument is passed to the simulator please see the
6601 simulator docs for details.
6604 -X <Clock frequency > this options is passed to the simulator please see
6605 the simulator docs for details.
6608 -s <serial port file> passed to simulator see the simulator docs for details.
6611 -S <serial in,out> passed to simulator see the simulator docs for details.
6617 As mention earlier the command interface for the debugger has been deliberately
6618 kept as close the GNU debugger gdb, as possible.
6619 This will help the integration with existing graphical user interfaces
6620 (like ddd, xxgdb or xemacs) existing for the GNU debugger.
6621 \layout Subsubsection
6623 break [line | file:line | function | file:function]
6626 Set breakpoint at specified line or function:
6635 sdcdb>break foo.c:100
6639 sdcdb>break foo.c:funcfoo
6640 \layout Subsubsection
6642 clear [line | file:line | function | file:function ]
6645 Clear breakpoint at specified line or function:
6654 sdcdb>clear foo.c:100
6658 sdcdb>clear foo.c:funcfoo
6659 \layout Subsubsection
6664 Continue program being debugged, after breakpoint.
6665 \layout Subsubsection
6670 Execute till the end of the current function.
6671 \layout Subsubsection
6676 Delete breakpoint number 'n'.
6677 If used without any option clear ALL user defined break points.
6678 \layout Subsubsection
6680 info [break | stack | frame | registers ]
6683 info break - list all breakpoints
6686 info stack - show the function call stack.
6689 info frame - show information about the current execution frame.
6692 info registers - show content of all registers.
6693 \layout Subsubsection
6698 Step program until it reaches a different source line.
6699 \layout Subsubsection
6704 Step program, proceeding through subroutine calls.
6705 \layout Subsubsection
6710 Start debugged program.
6711 \layout Subsubsection
6716 Print type information of the variable.
6717 \layout Subsubsection
6722 print value of variable.
6723 \layout Subsubsection
6728 load the given file name.
6729 Note this is an alternate method of loading file for debugging.
6730 \layout Subsubsection
6735 print information about current frame.
6736 \layout Subsubsection
6741 Toggle between C source & assembly source.
6742 \layout Subsubsection
6747 Send the string following '!' to the simulator, the simulator response is
6749 Note the debugger does not interpret the command being sent to the simulator,
6750 so if a command like 'go' is sent the debugger can loose its execution
6751 context and may display incorrect values.
6752 \layout Subsubsection
6759 My name is Bobby Brown"
6762 Interfacing with XEmacs.
6765 Two files (in emacs lisp) are provided for the interfacing with XEmacs,
6766 sdcdb.el and sdcdbsrc.el.
6767 These two files can be found in the $(prefix)/bin directory after the installat
6769 These files need to be loaded into XEmacs for the interface to work.
6770 This can be done at XEmacs startup time by inserting the following into
6771 your '.xemacs' file (which can be found in your HOME directory):
6777 (load-file sdcdbsrc.el)
6783 .xemacs is a lisp file so the () around the command is REQUIRED.
6784 The files can also be loaded dynamically while XEmacs is running, set the
6785 environment variable 'EMACSLOADPATH' to the installation bin directory
6786 (<installdir>/bin), then enter the following command ESC-x load-file sdcdbsrc.
6787 To start the interface enter the following command:
6801 You will prompted to enter the file name to be debugged.
6806 The command line options that are passed to the simulator directly are bound
6807 to default values in the file sdcdbsrc.el.
6808 The variables are listed below, these values maybe changed as required.
6811 sdcdbsrc-cpu-type '51
6814 sdcdbsrc-frequency '11059200
6820 The following is a list of key mapping for the debugger interface.
6828 ;; Current Listing ::
6845 binding\SpecialChar ~
6884 -------\SpecialChar ~
6924 sdcdb-next-from-src\SpecialChar ~
6950 sdcdb-back-from-src\SpecialChar ~
6976 sdcdb-cont-from-src\SpecialChar ~
6986 SDCDB continue command
7002 sdcdb-step-from-src\SpecialChar ~
7028 sdcdb-whatis-c-sexp\SpecialChar ~
7038 SDCDB ptypecommand for data at
7102 sdcdbsrc-delete\SpecialChar ~
7116 SDCDB Delete all breakpoints if no arg
7164 given or delete arg (C-u arg x)
7180 sdcdbsrc-frame\SpecialChar ~
7195 SDCDB Display current frame if no arg,
7244 given or display frame arg
7309 sdcdbsrc-goto-sdcdb\SpecialChar ~
7319 Goto the SDCDB output buffer
7335 sdcdb-print-c-sexp\SpecialChar ~
7346 SDCDB print command for data at
7410 sdcdbsrc-goto-sdcdb\SpecialChar ~
7420 Goto the SDCDB output buffer
7436 sdcdbsrc-mode\SpecialChar ~
7452 Toggles Sdcdbsrc mode (turns it off)
7456 ;; C-c C-f\SpecialChar ~
7464 sdcdb-finish-from-src\SpecialChar ~
7472 SDCDB finish command
7476 ;; C-x SPC\SpecialChar ~
7484 sdcdb-break\SpecialChar ~
7502 Set break for line with point
7504 ;; ESC t\SpecialChar ~
7514 sdcdbsrc-mode\SpecialChar ~
7530 Toggle Sdcdbsrc mode
7532 ;; ESC m\SpecialChar ~
7542 sdcdbsrc-srcmode\SpecialChar ~
7566 The Z80 and gbz80 port
7569 SDCC can target both the Zilog Z80 and the Nintendo Gameboy's Z80-like gbz80.
7570 The port is incomplete - long support is incomplete (mul, div and mod are
7571 unimplimented), and both float and bitfield support is missing.
7572 Apart from that the code generated is correct.
7575 As always, the code is the authoritave reference - see z80/ralloc.c and z80/gen.c.
7576 The stack frame is similar to that generated by the IAR Z80 compiler.
7577 IX is used as the base pointer, HL is used as a temporary register, and
7578 BC and DE are available for holding varibles.
7579 IY is currently unusued.
7580 Return values are stored in HL.
7581 One bad side effect of using IX as the base pointer is that a functions
7582 stack frame is limited to 127 bytes - this will be fixed in a later version.
7588 SDCC has grown to be a large project.
7589 The compiler alone (without the preprocessor, assembler and linker) is
7590 about 40,000 lines of code (blank stripped).
7591 The open source nature of this project is a key to its continued growth
7593 You gain the benefit and support of many active software developers and
7595 Is SDCC perfect? No, that's why we need your help.
7596 The developers take pride in fixing reported bugs.
7597 You can help by reporting the bugs and helping other SDCC users.
7598 There are lots of ways to contribute, and we encourage you to take part
7599 in making SDCC a great software package.
7605 Send an email to the mailing list at 'user-sdcc@sdcc.sourceforge.net' or 'devel-sd
7606 cc@sdcc.sourceforge.net'.
7607 Bugs will be fixed ASAP.
7608 When reporting a bug, it is very useful to include a small test program
7609 which reproduces the problem.
7610 If you can isolate the problem by looking at the generated assembly code,
7611 this can be very helpful.
7612 Compiling your program with the --dumpall option can sometimes be useful
7613 in locating optimization problems.
7619 Sandeep Dutta (sandeep.dutta@usa.net) - SDCC, the compiler, MCS51 code generator,
7622 Alan Baldwin (baldwin@shop-pdp.kent.edu) - Initial version of ASXXXX & ASLINK.
7625 John Hartman (jhartman@compuserve.com) - Porting ASXXX & ASLINK for 8051
7628 Obukhov (dso@usa.net) - malloc & serial i/o routines.
7631 Daniel Drotos (drdani@mazsola.iit.uni-miskolc.hu) - for his Freeware simulator
7633 Malini Dutta(malini_dutta@hotmail.com) - my wife for her patience and support.
7635 Unknown - for the GNU C - preprocessor.
7637 Michael Hope - The Z80 and Z80GB port, 186 development
7639 Kevin Vigor - The DS390 port.
7641 Johan Knol - Lots of fixes and enhancements, DS390/TINI libs.
7643 Scott Datallo - The PIC port.
7649 Thanks to all the other volunteer developers who have helped with coding,
7650 testing, web-page creation, distribution sets, etc.
7651 You know who you are :-)
7658 This document was initially written by Sandeep Dutta
7661 All product names mentioned herein may be trademarks of their respective
7667 \begin_inset LatexCommand \printindex{}