ci: Enable baremetal VFIO integration tests

Relying on a NVIDIA Tesla T4 card present in the SGX machine, this patch
enables baremetal VFIO testing, validated by running several NVIDIA
tools in the guest. The guest image has been prepared to include all the
software needed to run these tests.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
This commit is contained in:
Sebastien Boeuf 2021-03-16 16:11:00 +01:00 committed by Rob Bradford
parent d0c0a98eda
commit 04dc496808
4 changed files with 146 additions and 0 deletions

33
Jenkinsfile vendored
View File

@ -136,6 +136,39 @@ pipeline{
}
}
}
stage ('Worker build VFIO') {
agent { node { label 'bionic-vfio' } }
when { branch 'master' }
stages {
stage ('Checkout') {
steps {
checkout scm
}
}
stage ('Run VFIO integration tests') {
options {
timeout(time: 1, unit: 'HOURS')
}
steps {
sh "scripts/dev_cli.sh tests --integration-vfio"
}
}
stage ('Run VFIO integration tests for musl') {
options {
timeout(time: 1, unit: 'HOURS')
}
steps {
sh "scripts/dev_cli.sh tests --integration-vfio --libc musl"
}
}
}
post {
always {
sh "sudo chown -R jenkins.jenkins ${WORKSPACE}"
deleteDir()
}
}
}
stage ('Worker build - Windows guest') {
agent { node { label 'groovy-win' } }
stages {

View File

@ -182,6 +182,7 @@ cmd_help() {
echo " --cargo Run the cargo tests."
echo " --integration Run the integration tests."
echo " --integration-sgx Run the SGX integration tests."
echo " --integration-vfio Run the VFIO integration tests."
echo " --integration-windows Run the Windows guest integration tests."
echo " --libc Select the C library Cloud Hypervisor will be built against. Default is gnu"
echo " --volumes Hash separated volumes to be exported. Example --volumes /mnt:/mnt#/myvol:/myvol"
@ -298,6 +299,7 @@ cmd_tests() {
cargo=false
integration=false
integration_sgx=false
integration_vfio=false
integration_windows=false
libc="gnu"
arg_vols=""
@ -310,6 +312,7 @@ cmd_tests() {
"--cargo") { cargo=true; } ;;
"--integration") { integration=true; } ;;
"--integration-sgx") { integration_sgx=true; } ;;
"--integration-vfio") { integration_vfio=true; } ;;
"--integration-windows") { integration_windows=true; } ;;
"--libc")
shift
@ -419,6 +422,25 @@ cmd_tests() {
./scripts/run_integration_tests_sgx.sh "$@" || fix_dir_perms $? || exit $?
fi
if [ "$integration_vfio" = true ] ; then
say "Running VFIO integration tests for $target..."
$DOCKER_RUNTIME run \
--workdir "$CTR_CLH_ROOT_DIR" \
--rm \
--privileged \
--security-opt seccomp=unconfined \
--ipc=host \
--net="$CTR_CLH_NET" \
--mount type=tmpfs,destination=/tmp \
--volume /dev:/dev \
--volume "$CLH_ROOT_DIR:$CTR_CLH_ROOT_DIR" $exported_volumes \
--volume "$CLH_INTEGRATION_WORKLOADS:$CTR_CLH_INTEGRATION_WORKLOADS" \
--env USER="root" \
--env CH_LIBC="${libc}" \
"$CTR_IMAGE" \
./scripts/run_integration_tests_vfio.sh "$@" || fix_dir_perms $? || exit $?
fi
if [ "$integration_windows" = true ] ; then
say "Running Windows integration tests for $target..."
$DOCKER_RUNTIME run \

View File

@ -0,0 +1,28 @@
#!/bin/bash
set -x
source $HOME/.cargo/env
source $(dirname "$0")/test-util.sh
process_common_args "$@"
# For now these values are deafult for kvm
features_build=""
features_test="--features integration_tests"
BUILD_TARGET="$(uname -m)-unknown-linux-${CH_LIBC}"
CFLAGS=""
TARGET_CC=""
if [[ "${BUILD_TARGET}" == "x86_64-unknown-linux-musl" ]]; then
TARGET_CC="musl-gcc"
CFLAGS="-I /usr/include/x86_64-linux-musl/ -idirafter /usr/include/"
fi
cargo build --all --release $features_build --target $BUILD_TARGET
strip target/$BUILD_TARGET/release/cloud-hypervisor
export RUST_BACKTRACE=1
time cargo test $features_test "tests::vfio::$test_filter"
RES=$?
exit $RES

View File

@ -88,6 +88,8 @@ mod tests {
const FOCAL_IMAGE_NAME: &str = "focal-server-cloudimg-amd64-custom-20210106-1.raw";
#[cfg(target_arch = "x86_64")]
const FOCAL_SGX_IMAGE_NAME: &str = "focal-server-cloudimg-amd64-sgx.raw";
#[cfg(target_arch = "x86_64")]
const FOCAL_NVIDIA_IMAGE_NAME: &str = "focal-server-cloudimg-amd64-nvidia.raw";
#[cfg(target_arch = "aarch64")]
const BIONIC_IMAGE_NAME: &str = "bionic-server-cloudimg-arm64.raw";
#[cfg(target_arch = "aarch64")]
@ -6323,4 +6325,65 @@ mod tests {
handle_child_output(r, &output);
}
}
#[cfg(target_arch = "x86_64")]
mod vfio {
use crate::tests::*;
#[test]
fn test_nvidia_card() {
let mut focal = UbuntuDiskConfig::new(FOCAL_NVIDIA_IMAGE_NAME.to_string());
let guest = Guest::new(&mut focal);
let mut child = GuestCommand::new(&guest)
.args(&["--cpus", "boot=4"])
.args(&["--memory", "size=4G"])
.args(&["--kernel", guest.fw_path.as_str()])
.args(&["--device", "path=/sys/bus/pci/devices/0000:31:00.0/"])
.default_disks()
.default_net()
.capture_output()
.spawn()
.unwrap();
let r = std::panic::catch_unwind(|| {
guest.wait_vm_boot(None).unwrap();
// Run CUDA sample to validate it can find the device
let device_query_result = guest
.ssh_command(
"sudo /root/NVIDIA_CUDA-11.2_Samples/bin/x86_64/linux/release/deviceQuery",
)
.unwrap();
assert!(device_query_result.contains("Detected 1 CUDA Capable device"));
assert!(device_query_result.contains("Device 0: \"Tesla T4\""));
assert!(device_query_result.contains("Result = PASS"));
// Run NVIDIA DCGM Diagnostics to validate the device is functional
assert_eq!(
guest
.ssh_command("sudo nv-hostengine && echo ok")
.unwrap()
.trim(),
"ok"
);
assert!(guest
.ssh_command("sudo dcgmi discovery -l")
.unwrap()
.contains("Name: Tesla T4"));
assert_eq!(
guest
.ssh_command("sudo dcgmi diag -r 'diagnostic' | grep Pass | wc -l")
.unwrap()
.trim(),
"10"
);
});
let _ = child.kill();
let output = child.wait_with_output().unwrap();
handle_child_output(r, &output);
}
}
}