Using OpenCPI Environment Variables
Various environment variables are used to control OpenCPI. All start with the prefix OCPI_.
Nearly all are optional. During installation, some are set when external tools are installed in
non-default locations. Most are only used during development, and a few are used when applications
run. Some of these variables are useful for debugging purposes. The OpenCPI environment is
established when the user sources the opencpi-setup.sh script, in its installed location, which
is the cdk subdirectory of the source installation (git repo).
When the setup script is sourced, the OCPI_CDK_DIR environment variable (which points to
the OpenCPI installation) is set to the directory where the script lives; for example,
~/opencpi/cdk.
The setup command is:
source <location-of-opencpi>/cdk/opencpi-setup.sh -s
So, if OpenCPI is installed at ~/opencpi, the command is:
source ~/opencpi/cdk/opencpi-setup.sh -s
Sourcing this script with no arguments (or with the -h or --help option), will display more
options for special cases. If you prefer the OpenCPI environment to be set up automatically at
login, place this command in the shell (bash) login startup file (usually ~/.profile) after the
export PATH line. Note that this will only take effect when you log in or when you start a new
“login shell” using the -l (ell) option to bash, like:
bash -l
Any persistent site-specific environment variable settings are placed in the user-env.sh file in
the installation directory (OCPI_ROOT_DIR). The file is created automatically during
installation.
OpenCPI currently only supports the bash shell.
The first table below lists variables only relevant to the execution of applications, whether in a development environment or a deployed runtime environment. The second table is for variables only used in a development environment.
Name |
Description |
|
This variable set by OpenCPI automatically and indicates the location of the OpenCPI development or runtime installation. This variable should not be set directly. Even though it has CDK in its name, it is indeed used at runtime to locate the OpenCPI runtime installation. |
|
A colon-separated set of directories to be searched
for runtime artifacts. When referencing the
artifacts exported by a project with sources, be
sure to reference the project’s |
|
The log level (amount of logging) output by the
runtime system. The default is zero (0),
indicating no logging output. The maximum is
20. Commonly useful startup and diagnostic
information (for example, artifact discovery
feedback) is provided at log level 8. Unusual
events are logged at level 4. Logging is to
|
|
The runtime system XML system configuration file. If not specified, the file is located trying these locations in order:
|
|
Allows super-user privileged processes to specify
physical DMA memory to use when the OpenCPI kernel
driver is not used. The format is
|
|
The size (in bytes, decimal, with optional K or M suffix) of the memory pool to be used for buffers in software containers. Must support buffers for all ports used to communicate into or out of the container. |
Name |
Description |
|
A colon-separated set of project directories to be considered in addition to those that are registered. This variable can be used to temporarily augment the project registry when testing a new project. Rarely used. |
|
The location of the current project registry
when dealing with unregistered projects or
ocpidev show commands. By default, this
location is |
All of these variables are optional. If you want permanent settings made when using a given OpenCPI
installation, settings (export commands) may be placed in the user‑env.sh file in the
installation.