github/kata-containers
1
0
Fork 2
mirror of https://github.com/kata-containers/kata-containers.git synced 2026-04-08 04:53:18 +00:00
Code Issues Projects Releases Wiki Activity

nvidia: add nvrc.smi.srs=1 to default nvidia kernel params

Browse Source 
The attestation-agent no longer sets nvidia devices to ready
automatically. Instead, we should use nvrc for this. Since this is
required for all nvidia workloads, add it to the default nv kernel
params.

With bounce buffers, the timing of attesting a device versus setting it
to ready is not so important.

Signed-off-by: Tobin Feldman-Fitzthum <tfeldmanfitz@nvidia.com>
This commit is contained in:
Tobin Feldman-Fitzthum
2026-04-07 13:56:04 +00:00
parent 7385938c57
commit e40abcf72d
4 changed files with 6 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
src/runtime/Makefile
View File

@@ -489,6 +489,9 @@ ifneq (,$(QEMUCMD))
KERNELPARAMS_NV += "pci=nocrs"
KERNELPARAMS_NV += "pci=assign-busses"
KERNELPARAMS_CONFIDENTIAL_NV = $(KERNELPARAMS_NV)
KERNELPARAMS_CONFIDENTIAL_NV += "nvrc.smi.srs=1"
# Setting this to false can lead to cgroup leakages in the host
# Best practice for production is to set this to true
DEFSANDBOXCGROUPONLY_NV = true
@@ -661,6 +664,7 @@ USER_VARS += DEFAULTMEMORY_NV
USER_VARS += DEFAULTVFIOPORT_NV
USER_VARS += DEFAULTPCIEROOTPORT_NV
USER_VARS += KERNELPARAMS_NV
USER_VARS += KERNELPARAMS_CONFIDENTIAL_NV
USER_VARS += KERNELVERITYPARAMS_NV
USER_VARS += KERNELVERITYPARAMS_CONFIDENTIAL_NV
USER_VARS += DEFAULTTIMEOUT_NV

2
src/runtime/config/configuration-qemu-nvidia-gpu-snp.toml.in
View File

@@ -90,7 +90,7 @@ snp_guest_policy = 196608
# may stop the virtual machine from booting.
# To see the list of default parameters, enable hypervisor debug, create a
# container and look for 'default-kernel-parameters' log entries.
kernel_params = "@KERNELPARAMS_NV@"
kernel_params = "@KERNELPARAMS_CONFIDENTIAL_NV@"
# Optional dm-verity parameters (comma-separated key=value list):
# root_hash=...,salt=...,data_blocks=...,data_block_size=...,hash_block_size=...

2
src/runtime/config/configuration-qemu-nvidia-gpu-tdx.toml.in
View File

@@ -67,7 +67,7 @@ valid_hypervisor_paths = @QEMUTDXEXPERIMENTALVALIDHYPERVISORPATHS@
# may stop the virtual machine from booting.
# To see the list of default parameters, enable hypervisor debug, create a
# container and look for 'default-kernel-parameters' log entries.
kernel_params = "@KERNELPARAMS_NV@"
kernel_params = "@KERNELPARAMS_CONFIDENTIAL_NV@"
# Optional dm-verity parameters (comma-separated key=value list):
# root_hash=...,salt=...,data_blocks=...,data_block_size=...,hash_block_size=...

8
tests/integration/kubernetes/k8s-nvidia-cuda.bats
View File

@@ -29,14 +29,6 @@ setup() {
envsubst < "${pod_yaml_in}" > "${pod_yaml}"
if [ "${TEE}" = "true" ]; then
kernel_params_annotation="io.katacontainers.config.hypervisor.kernel_params"
kernel_params_value="nvrc.smi.srs=1"
set_metadata_annotation "${pod_yaml}" \
"${kernel_params_annotation}" \
"${kernel_params_value}"
fi
policy_settings_dir="$(create_tmp_policy_settings_dir "${pod_config_dir}")"
add_requests_to_policy_settings "${policy_settings_dir}" "ReadStreamRequest"