tools/README.txt
================
This README file addresses the contents of the NuttX tools/ directory.
The tools/ directory contains miscellaneous scripts and host C programs
that are necessary parts of the NuttX build system. These files
include:
README.txt
----------
This file!
Config.mk
---------
This file contains common definitions used by many configuration files.
This file (along with <nuttx>/.config) must be included at the top of
each configuration-specific Make.defs file like:
-include $(TOPDIR)/.config
include $(TOPDIR)/tools/Config.mk
Subsequent logic within the configuration-specific Make.defs file may then
override these default definitions as necessary.
configure.sh
configure.bat
configure.c, cfgparser.c, and cfgparser.h
------------
configure.sh is a bash script that is used to configure NuttX for a given
target board in a environment that supports POSIX paths (Linux, Cygwin,
OSX, or similar). See configs/README.txt or Documentation/NuttxPortingGuide.html
for a description of how to configure NuttX with this script.
configure.c, cfgparser.c, and cfgparser.h can be used to build a work-alike
program as a replacement for configure.sh. This work-alike program would be
used in environments that do not support Bash scripting (such as the Windows
native environment).
configure.bat is a small Windows batch file that can be used as a replacement
for configure.sh in a Windows native environment. configure.bat is actually
just a thin layer that executes configure.exe if it is available. If
configure.exe is not available, then configure.bat will attempt to build it
first.
In order two build configure.exe from configure.c in the Windows native
environment, two assumptions are made:
1) You have installed the MinGW GCC toolchain. This toolchain can be
downloaded from http://www.mingw.org/. Tt is recommended the you not
install the optional MSYS components as there may be conflicts.
2) That path to bin bin/ directory containing mingw-gcc.exe must be
included in the PATH variable.
discover.py
-----------
Example script for discovering devices in the local network.
It is the counter part to apps/netutils/discover
mkconfig.c, cfgdefine.c, and cfgdefine.h
----------------------------------------
These are Cs file that are used to build mkconfig program. The mkconfig
program is used during the initial NuttX build.
When you configure NuttX, you will copy a configuration file called .config
in the top level NuttX directory (See configs/README.txt or
Documentation/NuttxPortingGuide.html). The first time you make NuttX,
the top-level makefile will build the mkconfig executable from mkconfig.c
(using Makefile.host). The top-level Makefile will then execute the
mkconfig program to convert the .config file in the top level directory
into include/nuttx/config.h. config.h is a another version of the
NuttX configuration that can be included by C files.
cmdconfig.c
-----------
This C file can be used to build a utility for comparing two NuttX
configuration files.
kconfig2html.c
--------------
This is a C file that can be used build a utility for converting the
NuttX configuration in the Kconfig files to an HTML document. This
auto-generated documentation will, eventually, replace the manually
updated configuration documentation that is fallling woefully behind.
$ tools/kconfig2html.exe -h
USAGE: tools/kconfig2html [-d] [-a <apps directory>] {-o <out file>] [<Kconfig root>]
tools/kconfig2html [-h]
Where:
-a : Select relative path to the apps/ directory. Theis path is relative
to the <Kconfig directory>. Default: ../apps
-o : Send output to <out file>. Default: Output goes to stdout
-d : Enable debug output
-h : Prints this message and exits
<Kconfig root> is the directory containing the root Kconfig file.
Default <Kconfig directory>: .
mkconfigvars.sh
---------------
The HTML documentation expects to have a copy of the auto-generated
configuration variable documentation Documentation/NuttXConfigVariables.html.
The script mkconfigvars.sh is a simple script that can be used to
re-generated that file as needed.
$ tools/mkconfigvars.sh -h
tools/mkconfigvars.sh is a tool for generation of configuration variable documentation
USAGE: tools/mkconfigvars.sh [-d|h] [-v <major.minor>]
Where:
-v <major.minor>
The NuttX version number expressed as a major and minor number separated
by a period
-d
Enable script debug
-h
show this help message and exit
mkexport.sh and Makefile.export
-------------------------------
These implement part of the top-level Makefile's 'export' target. That
target will bundle up all of the NuttX libraries, header files, and the
startup object into an export-able, binary NuttX distribution. The
Makefile.export is used only by the mkexport.sh script to parse out
options from the top-level Make.defs file.
mkfsdata.pl
-----------
This perl script is used to build the "fake" file system and CGI support
as needed for the apps/netutils/webserver. It is currently used only
by the Makefile at apps/examples/uip. That example serves as an example
of how to configure the uIP webserver "fake" file system.
NOTE: This perl script comes from uIP and was (probably) written
by Adam Dunkels. uIP has a license that is compatible with NuttX.
mkversion.c, cfgdefine.c, and cfgdefine.h
-----------------------------------------
This is C file that is used to build mkversion program. The mkversion
program is used during the initial NuttX build.
When you build NuttX there should be a version file called .version in
the top level NuttX directory (See Documentation/NuttxPortingGuide.html).
The first time you make NuttX, the top-level makefile will build th
mkversion executable from mkversion.c (using Makefile.host). The top-
level Makefile will then execute the mkversion program to convert the
.version file in the top level directory into include/nuttx/version.h.
version.h provides version information that can be included by C files.
mksyscall.c, cvsparser.c, and cvsparser.h
-----------------------------------------
This is a C file that is used to build mksyscall program. The mksyscall
program is used during the initial NuttX build by the logic in the top-
level syscall/ directory.
If you build NuttX as a separately compiled, monolithic kernel and separate
applications, then there is a syscall layer that is used to get from the
user application space to the NuttX kernel space. In the user application
"proxies" for each of the kernel functions are provided. The proxies have
the same function signature as the kernel function, but only execute a
system call.
Within the kernel, there are "stubs" for each of the system calls. The
stubs receive the marshalled system call data, and perform the actually
kernel function call (in kernel-mode) on behalf of the proxy function.
Information about the stubs and proxies is maintained in a comma separated
value (CSV) file in the syscall/ directory. The mksyscall program will
accept this CVS file as input and generate all of the required proxy or
stub files as output. See syscall/README.txt for additonal information.
mksymtab.c, cvsparser.c, and cvsparser.h
----------------------------------------
This is a C file that is used to build symbol tables from common-separated
value (CSV) files. This tool is not used during the NuttX build, but
can be used as needed to generate files.
USAGE: ./mksymtab <cvs-file> <symtab-file>
Where:
<cvs-file> : The path to the input CSV file
<symtab-file>: The path to the output symbol table file
-d : Enable debug output
Example:
cd nuttx/tools
cat ../syscall/syscall.csv ../lib/libc.csv | sort >tmp.csv
./mksymtab.exe tmp.csv tmp.c
mkctags.sh
----------
A script for creating ctags from Ken Pettit. See http://en.wikipedia.org/wiki/Ctags
and http://ctags.sourceforge.net/
nxstyle.c
---------
I am embarassed that this is here. This program is a complete hack
but, unfortunately, it has become so useful to me that I need to keep
it here.
A little background: I have tinkered with pretty printers for some
time and have not been happy with the results. An alternative that
occurred to me would be just a standard checker that examines a C
file that gives warnings for violations of the coding standard.
This turns out to be more difficult that you might think. A pretty
printer understands C syntax: They break the file up into its C
components then reassembles the output in the format. But parsing the
C loses the original file layout and so it not useful in this case.
This program instead, uses a collection of heuristics (i.e., hacks and
bandaids) to examine the C file for obvious violations of the coding
standard. This program is completely ignorant of C syntax; it simply
performs crude pattern matching to check the file.
Usage: nxstyle <path-to-file-to-check>
pic32mx
-------
This directory contains build tools used only for PIC32MX platforms
bdf-convert.c
-------------
This C file is used to build the bdf-converter program. The bdf-converter
program be used to convert fonts in Bitmap Distribution Format (BDF)
into fonts that can be used in the NX graphics system.
Below are general instructions for creating and installing a new font
in the NX graphic system:
1. Locate a font in BDF format,
2. Use the bdf-converter program to convert the BDF font to the NuttX
font format. This will result in a C header file containing
definitions. That header file should be installed at, for example,
graphics/nxfonts/nxfonts_myfont.h.
Create a new NuttX configuration variable. For example, suppose
you define the following variable: CONFIG_NXFONT_MYFONT. Then
you would need to:
3. Define CONFIG_NXFONT_MYFONT=y in your NuttX configuration file.
A font ID number has to be assigned for each new font. The font ID
is defined in the file include/nuttx/nx/nxfonts.h. Those definitions
have to be extended to support your new font. Look at how the font ID
enabled by CONFIG_NXFONT_SANS23X27 is defined and add an ID for your
new font in a similar fashion:
4. include/nuttx/nx/nxfonts.h. Add you new font as a possible system
default font:
#if defined(CONFIG_NXFONT_SANS23X27)
# define NXFONT_DEFAULT FONTID_SANS23X27
#elif defined(CONFIG_NXFONT_MYFONT)
# define NXFONT_DEFAULT FONTID_MYFONT
#endif
Then define the actual font ID. Make sure that the font ID value
is unique:
enum nx_fontid_e
{
FONTID_DEFAULT = 0 /* The default font */
#ifdef CONFIG_NXFONT_SANS23X27
, FONTID_SANS23X27 = 1 /* The 23x27 sans serif font */
#endif
#ifdef CONFIG_NXFONT_MYFONT
, FONTID_MYFONT = 2 /* My shiny, new font */
#endif
...
New Add the font to the NX build system. There are several files that
you have to modify to to this. Look how the build system uses the
font CONFIG_NXFONT_SANS23X27 for examples:
5. nuttx/graphics/Makefile. This file needs logic to auto-generate
a C source file from the header file that you generated with the
the bdf-converter program. Notice NXFONTS_FONTID=2; this must be
set to the same font ID value that you defined in the
include/nuttx/nx/nxfonts.h file.
genfontsources:
ifeq ($(CONFIG_NXFONT_SANS23X27),y)
@$(MAKE) -C nxfonts -f Makefile.sources TOPDIR=$(TOPDIR) NXFONTS_FONTID=1 EXTRADEFINES=$(EXTRADEFINES)
endif
ifeq ($(CONFIG_NXFONT_MYFONT),y)
@$(MAKE) -C nxfonts -f Makefile.sources TOPDIR=$(TOPDIR) NXFONTS_FONTID=2 EXTRADEFINES=$(EXTRADEFINES)
endif
6. nuttx/graphics/nxfonts/Make.defs. Set the make variable NXFSET_CSRCS.
NXFSET_CSRCS determines the name of the font C file to build when
NXFONTS_FONTID=2:
ifeq ($(CONFIG_NXFONT_SANS23X27),y)
NXFSET_CSRCS += nxfonts_bitmaps_sans23x27.c
endif
ifeq ($(CONFIG_NXFONT_MYFONT),y)
NXFSET_CSRCS += nxfonts_bitmaps_myfont.c
endif
7. nuttx/graphics/nxfonts/Makefile.sources. This is the Makefile used
in step 5 that will actually generate the font C file. So, given
your NXFONTS_FONTID=2, it needs to determine a prefix to use for
auto-generated variable and function names and (again) the name of
the auto-generated file to create (this must be the same name that
was used in nuttx/graphics/nxfonts/Make.defs):
ifeq ($(NXFONTS_FONTID),1)
NXFONTS_PREFIX := g_sans23x27_
GEN_CSRC = nxfonts_bitmaps_sans23x27.c
endif
ifeq ($(NXFONTS_FONTID),2)
NXFONTS_PREFIX := g_myfont_
GEN_CSRC = nxfonts_bitmaps_myfont.c
endif
8. graphics/nxfonts/nxfonts_bitmaps.c. This is the file that contains
the generic font structures. It is used as a "template" file by
nuttx/graphics/nxfonts/Makefile.sources to create your customized
font data set.
#if NXFONTS_FONTID == 1
# include "nxfonts_sans23x27.h"
#elif NXFONTS_FONTID == 2
# include "nxfonts_myfont.h"
#else
# error "No font ID specified"
#endif
Where nxfonts_myfont.h is the NuttX font file that we generated in
step 2 using the bdf-converter tool.
9. graphics/nxfonts/nxfonts_getfont.c. Finally, we need to extend the
logic that does the run-time font lookups so that can find our new
font. The lookup function is NXHANDLE nxf_getfonthandle(enum nx_fontid_e fontid).
The new font information needs to be added to data structures used by
that function:
#ifdef CONFIG_NXFONT_SANS23X27
extern const struct nx_fontpackage_s g_sans23x27_package;
#endif
#ifdef CONFIG_NXFONT_MYFONT
extern const struct nx_fontpackage_s g_myfont_package;
#endif
static FAR const struct nx_fontpackage_s *g_fontpackages[] =
{
#ifdef CONFIG_NXFONT_SANS23X27
&g_sans23x27_package,
#endif
#ifdef CONFIG_NXFONT_MYFONT
&g_myfont_package,
#endif
NULL
};
Makefile.host
-------------
This is the makefile that is used to make the mkconfig program from
the mkconfig.c C file, the cmpconfig program from cmpconfig.c C file
the mkversion program from the mkconfig.c C file, or the mksyscall
program from the mksyscall.c file. Usage:
cd tools/
make -f Makefile.host <program>
mkromfsimg.sh
-------------
This script may be used to automate the generate of a ROMFS file system
image. It accepts an rcS script "template" and generates and image that
may be mounted under /etc in the NuttX pseudo file system.
mkdeps.sh
mkdeps.bat
mkdeps.c
mknulldeps.sh
-------------
NuttX uses the GCC compilers capabilities to create Makefile dependencies.
The bash script mkdeps.sh is used to run GCC in order to create the
dependencies. If a NuttX configuration uses the GCC toolchain, its Make.defs
file (see configs/README.txt) will include a line like:
MKDEP = $(TOPDIR)/tools/mkdeps.sh, or
MKDEP = $(TOPDIR)/tools/mkdeps[.exe] (See NOTE below)
If the NuttX configuration does not use a GCC compatible toolchain, then
it cannot use the dependencies and instead it uses mknulldeps.sh:
MKDEP = $(TOPDIR)/tools/mknulldeps.sh
The mknulldeps.sh is a stub script that does essentially nothing.
NOTE: The mk*deps.* files are undergoing change. mkdeps.sh is a bash
script that produces dependencies well for POSIX style hosts (e..g.,
Linux and Cygwin). It does not work well for mixed environments with
a Windows toolchain running in a POSIX style environment (hence, the
mknulldeps.sh script). And, of course, cannot be used in a Windows
nativ environment.
[mkdeps.sh does have an option, --winpath, that purports to convert
the dependencies generated by a Windows toolchain to POSIX format.
However, that is not being used and mostly likely does not cover
all of the conversion cases.]
mkdeps.bat is a simple port of the bash script to run in a Windows
command shell. However, it does not work well either because some
of the common CFLAGS use characters like '=' which are transformed
by the CMD.exe shell.
mkdeps.c generates mkdeps (on Linux) or mkdeps.exe (on Windows).
However, this verison is still under-development. It works well in
the all POSIX environment or in the all Windows environment but also
does not work well in mixed POSIX environment with a Windows toolchain.
In that case, there are still issues with the conversion of things like
'c:\Program Files' to 'c:program files' by bash. Those issues may,
eventually be solvable but for now continue to use mknulldeps.sh in
that mixed environment.
define.sh
define.bat
---------
Different compilers have different conventions for specifying pre-
processor definitions on the compiler command line. This bash
script allows the build system to create create command line definitions
without concern for the particular compiler in use.
The define.bat script is a counterpart for use in the native Windows
build.
incdir.sh
incdir.bat
---------
Different compilers have different conventions for specifying lists
of include file paths on the compiler command line. This incdir.sh
bash script allows the build system to create include file paths without
concern for the particular compiler in use.
The incdir.bat script is a counterpart for use in the native Windows
build. However, there is currently only one compiler supported in
that context: MinGW-GCC.
link.sh
link.bat
copydir.sh
copydir.bat
unlink.sh
unlink.bat
----------
Different file system have different capabilities for symbolic links.
Some windows file systems have no native support for symbolic links.
Cygwin running under windows has special links built in that work with
all cygwin tools. However, they do not work when Windows native tools
are used with cygwin. In that case something different must be done.
If you are building under Linux or under cygwin with a cygwin tool
chain, then your Make.defs file may have definitions like the
following:
DIRLINK = $(TOPDIR)/tools/link.sh
DIRUNLINK = (TOPDIR)/tools/unlink.sh
The first definition is not always present because link.sh is the
default. link.sh is a bash script that performs a normal, Linux-style
symbolic link; unlink.sh is a do-it-all unlinking script.
But if you are building under cygwin using a Windows native toolchain
within a POSIX framework (such as Cygwin), then you will need something
like the following in you Make.defs file:
DIRLINK = $(TOPDIR)/tools/copydir.sh
DIRUNLINK = (TOPDIR)/tools/unlink.sh
copydir.sh will copy the whole directory instead of linking it.
Finally, if you are running in a pure native Windows environment with
a CMD.exe shell, then you will need something like this:
DIRLINK = $(TOPDIR)/tools/copydir.bat
DIRUNLINK = (TOPDIR)/tools/unlink.bat
Note that this will copy directories. ;ink.bat might also be used in
this case. link.bat will attempt to create a symbolic link using the
NTFS mklink.exe command instead of copying files. That logic, however,
has not been verified as of this writing.
kconfig.bat
-----------
Recent versions of NuttX support building NuttX from a native Windows
CMD.exe shell. But kconfig-frontends is a Linux tool and is not yet
available in the pure CMD.exe environment. At this point, there are
only a few options for the Windows user (see the top-level README.txt
file).
You can, with some effort, run the Cygwin kconfig-mconf tool directly
in the CMD.exe shell. In this case, you do not have to modify the
.config file, but there are other complexities: You need to
temporarily set the Cgywin directories in the PATH variable and
then run kconfig-mconf outside of the Make system.
kconfig.bat is a Windows batch file at tools/kconfig.bat that automates
these steps. It is used from the top-level NuttX directory like:
tools/kconfig menuconfig
NOTE: There is an currently an issue with accessing DOS environment
variables from the Cygwin kconfig-mconf running in the CMD.exe shell.
The following change to the top-level Kconfig file seems to work around
these problems:
config APPSDIR
string
- option env="APPSDIR"
+ default "../apps"
mkimage.sh
----------
The creates a downloadable image as needed with the rrload bootloader.
indent.sh
---------
This script can be used to indent .c and .h files in a manner similar
to my coding NuttX coding style. It doesn't do a really good job,
however (see the comments at the top of the indent.sh file).
refresh.sh
----------
This is a bash script that automatics refreshing of board default
configuration (defconfig) files. It does not do anything special
thet you cannot do manually, but is useful for me when I have to
update dozens of confuration files.
Configuration files have to be updated because over time, the
configuration settings change: New configurations are added and
new dependencies are added. So an old configuration file may
not be usable anymore until it is refreshed.
The steps to refresh the file are:
1. Make tools/cmpconfig if it is not already built.
2. Copy the the defconfig file to the top-level NuttX
directory as .config (being careful to save any previous
.config file that you might want to keep!).
3. Execute 'make oldconfig' to update the configuration.
'make oldconfig' will prompt you for each change in the
configuration that requires that you make some decision.
4. Then it runs tools/cmpconfig to show the real differences
between the configuration files. Configuration files are
complex and things can move around so a simple 'diff' between
two configuration files is often not useful. But tools/cmpconfig
will show only the meaningful differences between the two
configuration files.
4. It will edit the .config file to comment out the setting
of the CONFIG_APPS_DIR= setting. This setting should not
be in checked-in defconfig files because the actually must
be determined at the next time that the configuration is
installed.
5. Finally, the refreshed defconfig file is copied back in
place where it can be committed with the next set of
difference to the command line. refresh.sh will prompt
you first to avoid overwriting the defconfig file with
changes that you do not want.
testbuild.sh
------------
This script automates building of a set of configurations. The intent is
simply to assure that the set of configurations build correctly. The -h
option shows the usage:
$ ./testbuild.sh -h
USAGE: ./testbuild.sh [-w|l] [-c|n] [-s] <testlist-file>
USAGE: ./testbuild.sh -h
where
-w|l selects Windows (w) or Linux (l). Default: Linux
-c|n selects Windows native (n) or Cygwin (c). Default Cygwin
-s Use C++ long size_t in new operator. Default unsigned long
-h will show this help test and terminate
<testlist-file> selects the list of configurations to test. No default
Your PATH variable must include the path to both the build tools and the
kconfig-frontends tools
These script needs two pieces of information.
a. A description of the platform that you are testing on. This
description is provided by the optional -w, -l, -c, and -n options.
b. A list of configurations to build. That list is provided by a test
list file. The final, non-optional parameter, <testlist-file>,
provides the path to that file.
The test list file is a sequence of build descriptons, one per line. One
build descriptions consists of two comma separated values. For example:
stm32f429i-disco/nsh,CONFIG_ARMV7M_TOOLCHAIN_GNU_EABIL
The first value is the usual configuration description of the form
form <board-name>/<configuration-name> and must correspond to a
configuration in the nuttx/configs directory.
The second value is valid name for a toolchain configuration to use
when building the configuration. The set of valid toolchain
configuration names depends on the underlying architecture of the
configured board.
NOTE: The environment variable APPSDIR should be set to the relative
path to the application directory when running this script like:
$ export APPSDIR=../apps
zipme.sh
--------
I use this script to create the nuttx-xx.yy.tar.gz tarballs for
release on Bitbucket.org. It is handy because it also does the
kind of clean that you need to do to make a clean code release.