mirror of
git://nv-tegra.nvidia.com/linux-nvgpu.git
synced 2025-12-22 17:36:20 +03:00
1f1134544d
Add a script to install unit tests on a target jetson board. The installation consists of copying all the binaries generated by tmake over to that target board and building a source tree that matches the nvgpu code. The reason the systemimage out directory is not used is this is created by the `image createfs' command during the build process. This script will work after just a simple `tmm[p]' Change-Id: I1f2650e666a42c12762ab444159b69ba8fc582f8 Signed-off-by: Alex Waterman <alexw@nvidia.com> Reviewed-on: https://git-master.nvidia.com/r/1850545 Reviewed-by: Philip Elcan <pelcan@nvidia.com> GVS: Gerrit_Virtual_Submit Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com> Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
111 lines
2.6 KiB
Bash
Executable File
111 lines
2.6 KiB
Bash
Executable File
#!/bin/bash
|
|
#
|
|
# Install the unit test binaries, build by tmake, onto a jetson running
|
|
# L4T. The only argument is the IP address of the jetson.
|
|
#
|
|
|
|
usage() {
|
|
echo "Usage:"
|
|
echo ""
|
|
echo " $ $1 [-hs] [--help] [--install-sshkey] <jetson-ip>"
|
|
echo ""
|
|
}
|
|
|
|
# Copy to target jetson. Takes a local path and a remote path relative
|
|
# to ubuntu's $HOME.
|
|
#
|
|
# It helps a lot of you set up an authorized key on the target! Otherwise
|
|
# you may have a lot of typing to do...
|
|
jcp() {
|
|
cmd="rsync -qru $1 $target:$2"
|
|
echo "> $cmd"
|
|
$cmd
|
|
}
|
|
|
|
# Add our public key to the authorized key list on the target host.
|
|
install_ssh_key() {
|
|
ssh-copy-id -f $target
|
|
return $?
|
|
}
|
|
|
|
# Variables which may be set by CLI arguments.
|
|
install_sshkey=
|
|
|
|
# To start with filter out any non-target arguments. Right now that's only -h
|
|
# and -s.
|
|
positionals=()
|
|
while [[ $# -gt 0 ]]
|
|
do
|
|
arg="$1"
|
|
|
|
case $arg in
|
|
-h|--help)
|
|
usage $0
|
|
exit 1
|
|
;;
|
|
-s|--install-sshkey)
|
|
install_sshkey=yes
|
|
shift
|
|
;;
|
|
*)
|
|
positionals+=("$1")
|
|
shift
|
|
;;
|
|
esac
|
|
done
|
|
|
|
set -- "${positionals[@]}"
|
|
|
|
if [ "x$1" == "x" ]
|
|
then
|
|
echo "Missing IP address!"
|
|
usage $0
|
|
exit 1
|
|
fi
|
|
|
|
if [ "x$TOP" == "x" ]
|
|
then
|
|
echo "\$TOP must be set!"
|
|
exit 1
|
|
fi
|
|
|
|
target="ubuntu@$1"
|
|
nvgpu_bins=$TOP/out/*/nvidia/kernel/nvgpu
|
|
|
|
# Install the ssh key if needed.
|
|
if [ "$install_sshkey" == "yes" ]
|
|
then
|
|
echo "Installing our SSH key"
|
|
install_ssh_key $target || exit 1
|
|
fi
|
|
|
|
# Building the necessary directory structure. It may not be present
|
|
# first time this is run.
|
|
ssh $target mkdir -p nvgpu_unit/units
|
|
ssh $target mkdir -p $TOP/kernel
|
|
if [ $? != 0 ]; then
|
|
echo
|
|
echo "!! Unable to make $TOP on the target jetson! This directory needs"
|
|
echo "!! to be present and writable by this script in order for coverage"
|
|
echo "!! tracking to work."
|
|
exit 1
|
|
fi
|
|
|
|
# And copy...
|
|
jcp $nvgpu_bins/userspace-l4t_64/nvgpu_unit nvgpu_unit/nvgpu_unit
|
|
jcp $nvgpu_bins/userspace-l4t_64/libnvgpu_unit-lib.so nvgpu_unit/libnvgpu-unit.so
|
|
jcp $nvgpu_bins/drivers/gpu/nvgpu-l4t_64/libnvgpu-drv.so nvgpu_unit/libnvgpu-drv.so
|
|
jcp $TOP/kernel/nvgpu/userspace/unit.sh nvgpu_unit/unit.sh
|
|
|
|
for unit_dir in `ls $nvgpu_bins/userspace/units`; do
|
|
unit=${unit_dir%-l4t_64}
|
|
jcp $nvgpu_bins/userspace/units/$unit_dir/lib$unit.so nvgpu_unit/units
|
|
done
|
|
|
|
# Set up the necessary coverage files. Basically what we do is recreate just
|
|
# enough of the source/build output here on the local machine over on the
|
|
# target jetson. This means you may
|
|
jcp $nvgpu_bins nvgpu_unit
|
|
jcp $TOP/kernel/nvgpu $TOP/kernel
|
|
jcp $TOP/kernel/nvgpu-next $TOP/kernel
|