Merge branch 'fix-binary-name' into 'main'

Use BinaryName for v1 containerd runtime config

See merge request nvidia/container-toolkit/container-toolkit!159

CHANGELOG.md

@@ -0,0 +1,166 @@
+# NVIDIA Container Toolkit Changelog
+
+## v1.10.0-rc.3
+
+* Use default config instead of raising an error if config file cannot be found
+* Ignore NVIDIA_REQUIRE_JETPACK* environment variables for requirement checks
+* Fix bug in detection of Tegra systems where `/sys/devices/soc0/family` is ignored
+* Fix bug where links to devices were detected as devices
+* [libnvida-container] Fix bug introduced when adding libcudadebugger.so to list of libraries
+
+## v1.10.0-rc.2
+
+* Add support for NVIDIA_REQUIRE_* checks for cuda version and arch to csv mode
+* Switch to debug logging to reduce log verbosity
+* Support logging to logs requested in command line
+* Fix bug when launching containers with relative root path (e.g. using containerd)
+* Allow low-level runtime path to be set explicitly as nvidia-container-runtime.runtimes option
+* Fix failure to locate low-level runtime if PATH envvar is unset
+* Replace experimental option for NVIDIA Container Runtime with nvidia-container-runtime.mode = csv option
+* Use csv as default mode on Tegra systems without NVML
+* Add --version flag to all CLIs
+* [libnvidia-container] Bump libtirpc to 1.3.2
+* [libnvidia-container] Fix bug when running host ldconfig using glibc compiled with a non-standard prefix
+* [libnvidia-container] Add libcudadebugger.so to list of compute libraries
+
+## v1.10.0-rc.1
+
+* Include nvidia-ctk CLI in installed binaries
+* Add experimental option to NVIDIA Container Runtime
+
+## v1.9.0
+
+* [libnvidia-container] Add additional check for Tegra in /sys/.../family file in CLI
+* [libnvidia-container] Update jetpack-specific CLI option to only load Base CSV files by default
+* [libnvidia-container] Fix bug (from 1.8.0) when mounting GSP firmware into containers without /lib to /usr/lib symlinks
+* [libnvidia-container] Update nvml.h to CUDA 11.6.1 nvML_DEV 11.6.55
+* [libnvidia-container] Update switch statement to include new brands from latest nvml.h
+* [libnvidia-container] Process all --require flags on Jetson platforms
+* [libnvidia-container] Fix long-standing issue with running ldconfig on Debian systems
+
+## v1.8.1
+
+* [libnvidia-container] Fix bug in determining cgroup root when running in nested containers
+* [libnvidia-container] Fix permission issue when determining cgroup version
+
+## v1.8.0
+
+* Promote 1.8.0-rc.2-1 to 1.8.0
+
+## v1.8.0-rc.2
+
+* Remove support for building amazonlinux1 packages
+
+## v1.8.0-rc.1
+
+* [libnvidia-container] Add support for cgroupv2
+* Release toolkit-container images from nvidia-container-toolkit repository
+
+## v1.7.0
+
+* Promote 1.7.0-rc.1-1 to 1.7.0
+ * Bump Golang version to 1.16.4
+
+## v1.7.0-rc.1
+
+* Specify containerd runtime type as string in config tools to remove dependency on containerd package
+* Add supported-driver-capabilities config option to allow for a subset of all driver capabilities to be specified
+
+## v1.6.0
+
+* Promote 1.6.0-rc.3-1 to 1.6.0
+ * Fix unnecessary logging to stderr instead of configured nvidia-container-runtime log file
+
+## v1.6.0-rc.3
+
+* Add supported-driver-capabilities config option to the nvidia-container-toolkit
+* Move OCI and command line checks for runtime to internal oci package
+
+## v1.6.0-rc.2
+
+* Use relative path to OCI specification file (config.json) if bundle path is not specified as an argument to the nvidia-container-runtime
+
+## v1.6.0-rc.1
+
+* Add AARCH64 package for Amazon Linux 2
+* Include nvidia-container-runtime into nvidia-container-toolkit package
+
+## v1.5.1
+
+* Fix bug where Docker Swarm device selection is ignored if NVIDIA_VISIBLE_DEVICES is also set
+* Improve unit testing by using require package and adding coverage reports
+* Remove unneeded go dependencies by running go mod tidy
+* Move contents of pkg directory to cmd for CLI tools
+* Ensure make binary target explicitly sets GOOS
+
+## v1.5.0
+
+* Add dependence on libnvidia-container-tools >= 1.4.0
+* Add golang check targets to Makefile
+* Add Jenkinsfile definition for build targets
+* Move docker.mk to docker folder
+
+## v1.4.2
+
+* Add dependence on libnvidia-container-tools >= 1.3.3
+
+## v1.4.1
+
+* Ignore NVIDIA_VISIBLE_DEVICES for containers with insufficent privileges
+* Add dependence on libnvidia-container-tools >= 1.3.2
+
+## v1.4.0
+
+* Add 'compute' capability to list of defaults
+* Add dependence on libnvidia-container-tools >= 1.3.1
+
+## v1.3.0
+
+* Promote 1.3.0-rc.2-1 to 1.3.0
+* Add dependence on libnvidia-container-tools >= 1.3.0
+
+## v1.3.0-rc.2
+
+* 2c180947 Add more tests for new semantics with device list from volume mounts
+* 7c003857 Refactor accepting device lists from volume mounts as a boolean
+
+## v1.3.0-rc.1
+
+* b50d86c1 Update build system to accept a TAG variable for things like rc.x
+* fe65573b Add common CI tests for things like golint, gofmt, unit tests, etc.
+* da6fbb34 Revert "Add ability to merge envars of the form NVIDIA_VISIBLE_DEVICES_*"
+* a7fb3330 Flip build-all targets to run automatically on merge requests
+* 8b248b66 Rename github.com/NVIDIA/container-toolkit to nvidia-container-toolkit
+* da36874e Add new config options to pull device list from mounted files instead of ENVVAR
+
+## v1.2.1
+
+* 4e6e0ed4 Add 'ngx' to list of*all* driver capabilities
+* 2f4af743 List config.toml as a config file in the RPM SPEC
+
+## v1.2.0
+
+*  8e0aab46 Fix repo listed in changelog for debian distributions
+*  320bb6e4 Update dependence on libnvidia-container to 1.2.0
+*  6cfc8097 Update package license to match source license
+*  e7dc3cbb Fix debian copyright file
+*  d3aee3e0 Add the 'ngx' driver capability
+
+## v1.1.2
+
+* c32237f3 Add support for parsing Linux Capabilities for older OCI specs
+
+## v1.1.1
+
+* d202aded Update dependence to libnvidia-container 1.1.1
+
+## v1.1.0
+
+* 4e4de762 Update build system to support multi-arch builds
+* fcc1d116 Add support for MIG (Multi-Instance GPUs)
+* d4ff0416 Add ability to merge envars of the form NVIDIA_VISIBLE_DEVICES_*
+* 60f165ad Add no-pivot option to toolkit
+
+## v1.0.5
+
+* Initial release. Replaces older package nvidia-container-runtime-hook. (Closes: #XXXXXX)

docker/Dockerfile.amazonlinux

@@ -27,6 +27,7 @@ ENV GOPATH /go
 ENV PATH $GOPATH/bin:/usr/local/go/bin:$PATH
 
 # packaging
+ARG PKG_NAME
 ARG PKG_VERS
 ARG PKG_REV
 
@@ -63,6 +64,8 @@ COPY packaging/rpm .
 CMD arch=$(uname -m) && \
     rpmbuild --clean --target=$arch -bb \
              -D "_topdir $PWD" \
+             -D "release_date $(date +'%a %b %d %Y')" \
+             -D "git_commit ${GIT_COMMIT}" \
              -D "version $VERSION" \
              -D "libnvidia_container_version ${VERSION}-${RELEASE}" \
              -D "release $RELEASE" \

docker/Dockerfile.centos

@@ -27,6 +27,7 @@ ENV GOPATH /go
 ENV PATH $GOPATH/bin:/usr/local/go/bin:$PATH
 
 # packaging
+ARG PKG_NAME
 ARG PKG_VERS
 ARG PKG_REV
 
@@ -61,6 +62,8 @@ COPY packaging/rpm .
 CMD arch=$(uname -m) && \
     rpmbuild --clean --target=$arch -bb \
              -D "_topdir $PWD" \
+             -D "release_date $(date +'%a %b %d %Y')" \
+             -D "git_commit ${GIT_COMMIT}" \
              -D "version $VERSION" \
              -D "libnvidia_container_version ${VERSION}-${RELEASE}" \
              -D "release $RELEASE" \

docker/Dockerfile.debian

@@ -32,6 +32,7 @@ ENV GOPATH /go
 ENV PATH $GOPATH/bin:/usr/local/go/bin:$PATH
 
 # packaging
+ARG PKG_NAME
 ARG PKG_VERS
 ARG PKG_REV
 
@@ -64,9 +65,11 @@ RUN if [ "$(lsb_release -cs)" = "jessie" ]; then \
 WORKDIR $DIST_DIR
 COPY packaging/debian ./debian
 
-RUN sed -i "s;@VERSION@;${REVISION};" debian/changelog && \
-    dch --changelog debian/changelog --append "Bump libnvidia-container dependency to ${REVISION}}" && \
-    dch --changelog debian/changelog -r "" && \
+RUN dch --create --package="${PKG_NAME}" \
+        --newversion "${REVISION}" \
+            "See https://gitlab.com/nvidia/container-toolkit/container-toolkit/-/blob/${GIT_COMMIT}/CHANGELOG.md for the changelog" && \
+    dch --append "Bump libnvidia-container dependency to ${REVISION}" && \
+    dch -r "" && \
     if [ "$REVISION" != "$(dpkg-parsechangelog --show-field=Version)" ]; then exit 1; fi
 
 CMD export DISTRIB="$(lsb_release -cs)" && \

docker/Dockerfile.opensuse-leap

@@ -25,6 +25,7 @@ ENV GOPATH /go
 ENV PATH $GOPATH/bin:/usr/local/go/bin:$PATH
 
 # packaging
+ARG PKG_NAME
 ARG PKG_VERS
 ARG PKG_REV
 
@@ -59,6 +60,8 @@ COPY packaging/rpm .
 CMD arch=$(uname -m) && \
     rpmbuild --clean --target=$arch -bb \
              -D "_topdir $PWD" \
+             -D "release_date $(date +'%a %b %d %Y')" \
+             -D "git_commit ${GIT_COMMIT}" \
              -D "version $VERSION" \
              -D "libnvidia_container_version ${VERSION}-${RELEASE}" \
              -D "release $RELEASE" \

docker/Dockerfile.ubuntu

@@ -30,6 +30,7 @@ ENV GOPATH /go
 ENV PATH $GOPATH/bin:/usr/local/go/bin:$PATH
 
 # packaging
+ARG PKG_NAME
 ARG PKG_VERS
 ARG PKG_REV
 
@@ -57,9 +58,11 @@ COPY config/config.toml.${CONFIG_TOML_SUFFIX} $DIST_DIR/config.toml
 WORKDIR $DIST_DIR
 COPY packaging/debian ./debian
 
-RUN sed -i "s;@VERSION@;${REVISION};" debian/changelog && \
-    dch --changelog debian/changelog --append "Bump libnvidia-container dependency to ${REVISION}}" && \
-    dch --changelog debian/changelog -r "" && \
+RUN dch --create --package="${PKG_NAME}" \
+        --newversion "${REVISION}" \
+            "See https://gitlab.com/nvidia/container-toolkit/container-toolkit/-/blob/${GIT_COMMIT}/CHANGELOG.md for the changelog" && \
+    dch --append "Bump libnvidia-container dependency to ${REVISION}" && \
+    dch -r "" && \
     if [ "$REVISION" != "$(dpkg-parsechangelog --show-field=Version)" ]; then exit 1; fi
 
 CMD export DISTRIB="$(lsb_release -cs)" && \

docker/docker.mk

@@ -128,6 +128,7 @@ docker-build-%:
 	    --progress=plain \
 	    --build-arg BASEIMAGE="$(BASEIMAGE)" \
 	    --build-arg GOLANG_VERSION="$(GOLANG_VERSION)" \
+	    --build-arg PKG_NAME="$(LIB_NAME)" \
 	    --build-arg PKG_VERS="$(LIB_VERSION)" \
 	    --build-arg PKG_REV="$(PKG_REV)" \
 	    --build-arg CONFIG_TOML_SUFFIX="$(CONFIG_TOML_SUFFIX)" \

internal/config/config.go

@@ -61,7 +61,7 @@ func GetConfig() (*Config, error) {
 
 	tomlFile, err := os.Open(configFilePath)
 	if err != nil {
-		return nil, fmt.Errorf("failed to open config file %v: %v", configFilePath, err)
+		return getDefaultConfig(), nil
 	}
 	defer tomlFile.Close()
 

internal/config/image/cuda_image.go

@@ -29,6 +29,7 @@ const (
 	envCUDAVersion      = "CUDA_VERSION"
 	envNVRequirePrefix  = "NVIDIA_REQUIRE_"
 	envNVRequireCUDA    = envNVRequirePrefix + "CUDA"
+	envNVRequireJetpack = envNVRequirePrefix + "JETPACK"
 	envNVDisableRequire = "NVIDIA_DISABLE_REQUIRE"
 )
 
@@ -84,7 +85,7 @@ func (i CUDA) GetRequirements() ([]string, error) {
 	// All variables with the "NVIDIA_REQUIRE_" prefix are passed to nvidia-container-cli
 	var requirements []string
 	for name, value := range i {
-		if strings.HasPrefix(name, envNVRequirePrefix) {
+		if strings.HasPrefix(name, envNVRequirePrefix) && !strings.HasPrefix(name, envNVRequireJetpack) {
 			requirements = append(requirements, value)
 		}
 	}

internal/config/image/cuda_image_test.go

@@ -69,3 +69,55 @@ func TestParseMajorMinorVersionInvalid(t *testing.T) {
 		})
 	}
 }
+
+func TestGetRequirements(t *testing.T) {
+	testCases := []struct {
+		description  string
+		env          []string
+		requirements []string
+	}{
+		{
+			description:  "NVIDIA_REQUIRE_JETPACK is ignored",
+			env:          []string{"NVIDIA_REQUIRE_JETPACK=csv-mounts=all"},
+			requirements: nil,
+		},
+		{
+			description:  "NVIDIA_REQUIRE_JETPACK_HOST_MOUNTS is ignored",
+			env:          []string{"NVIDIA_REQUIRE_JETPACK_HOST_MOUNTS=base-only"},
+			requirements: nil,
+		},
+		{
+			description:  "single requirement set",
+			env:          []string{"NVIDIA_REQUIRE_CUDA=cuda>=11.6"},
+			requirements: []string{"cuda>=11.6"},
+		},
+		{
+			description:  "requirements are concatenated requirement set",
+			env:          []string{"NVIDIA_REQUIRE_CUDA=cuda>=11.6", "NVIDIA_REQUIRE_BRAND=brand=tesla"},
+			requirements: []string{"cuda>=11.6", "brand=tesla"},
+		},
+		{
+			description:  "legacy image",
+			env:          []string{"CUDA_VERSION=11.6"},
+			requirements: []string{"cuda>=11.6"},
+		},
+		{
+			description:  "legacy image with additional requirement",
+			env:          []string{"CUDA_VERSION=11.6", "NVIDIA_REQUIRE_BRAND=brand=tesla"},
+			requirements: []string{"cuda>=11.6", "brand=tesla"},
+		},
+	}
+
+	for _, tc := range testCases {
+		t.Run(tc.description, func(t *testing.T) {
+			image, err := NewCUDAImageFromEnv(tc.env)
+			require.NoError(t, err)
+
+			requirements, err := image.GetRequirements()
+			require.NoError(t, err)
+			require.ElementsMatch(t, tc.requirements, requirements)
+
+		})
+
+	}
+}

internal/info/info.go

@@ -48,8 +48,8 @@ func IsTegraSystem() (bool, string) {
 		return true, fmt.Sprintf("%v found", tegraReleaseFile)
 	}
 
-	if info, err := os.Stat(tegraFamilyFile); err != nil || !info.IsDir() {
-		return false, fmt.Sprintf("%v not found", tegraFamilyFile)
+	if info, err := os.Stat(tegraFamilyFile); err != nil || info.IsDir() {
+		return false, fmt.Sprintf("%v file not found", tegraFamilyFile)
 	}
 
 	contents, err := os.ReadFile(tegraFamilyFile)

internal/lookup/device.go

@@ -42,11 +42,11 @@ func NewCharDeviceLocator(logger *logrus.Logger, root string) Locator {
 
 // assertCharDevice checks whether the specified path is a char device and returns an error if this is not the case.
 func assertCharDevice(filename string) error {
-	info, err := os.Stat(filename)
+	info, err := os.Lstat(filename)
 	if err != nil {
 		return fmt.Errorf("error getting info: %v", err)
 	}
-	if info.Mode()|os.ModeCharDevice == 0 {
+	if info.Mode()&os.ModeCharDevice == 0 {
 		return fmt.Errorf("%v is not a char device", filename)
 	}
 	return nil

packaging/debian/changelog

@@ -1,211 +0,0 @@
-nvidia-container-toolkit (1.10.0~rc.2-1) experimental; urgency=medium
-
-  * Add support for NVIDIA_REQUIRE_* checks for cuda version and arch to csv mode
-  * Switch to debug logging to reduce log verbosity
-  * Support logging to logs requested in command line
-  * Fix bug when launching containers with relative root path (e.g. using containerd)
-  * Allow low-level runtime path to be set explicitly as nvidia-container-runtime.runtimes option
-  * Fix failure to locate low-level runtime if PATH envvar is unset
-  * Replace experimental option for NVIDIA Container Runtime with nvidia-container-runtime.mode = csv option
-  * Use csv as default mode on Tegra systems without NVML
-  * Add --version flag to all CLIs
-  * [libnvidia-container] Bump libtirpc to 1.3.2
-  * [libnvidia-container] Fix bug when running host ldconfig using glibc compiled with a non-standard prefix
-  * [libnvidia-container] Add libcudadebugger.so to list of compute libraries
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com>  Fri, 13 May 2022 13:41:10 +0200
-
-nvidia-container-toolkit (1.10.0~rc.1-1) experimental; urgency=medium
-
-  * Include nvidia-ctk CLI in installed binaries
-  * Add experimental option to NVIDIA Container Runtime
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com>  Thu, 24 Mar 2022 13:22:24 +0200
-
-nvidia-container-toolkit (1.9.0-1) UNRELEASED; urgency=medium
-
- * [libnvidia-container] Add additional check for Tegra in /sys/.../family file in CLI
- * [libnvidia-container] Update jetpack-specific CLI option to only load Base CSV files by default
- * [libnvidia-container] Fix bug (from 1.8.0) when mounting GSP firmware into containers without /lib to /usr/lib symlinks
- * [libnvidia-container] Update nvml.h to CUDA 11.6.1 nvML_DEV 11.6.55
- * [libnvidia-container] Update switch statement to include new brands from latest nvml.h
- * [libnvidia-container] Process all --require flags on Jetson platforms
- * [libnvidia-container] Fix long-standing issue with running ldconfig on Debian systems
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com> Fri, 18 Mar 2022 06:10:56 +0200
-
-nvidia-container-toolkit (1.8.1-1) UNRELEASED; urgency=medium
-
- * [libnvidia-container] Fix bug in determining cgroup root when running in nested containers
- * [libnvidia-container] Fix permission issue when determining cgroup version
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com> Mon, 14 Feb 2022 09:53:26 +0100
-
-nvidia-container-toolkit (1.8.0-1) UNRELEASED; urgency=medium
-
- * Promote 1.8.0~rc.2-1 to 1.8.0-1
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com> Fri, 04 Feb 2022 06:04:39 +0100
-
-nvidia-container-toolkit (1.8.0~rc.2-1) UNRELEASED; urgency=medium
-
-  * Remove support for building amazonlinux1 packages
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com> Thu, 20 Jan 2022 13:38:48 +0100
-
-nvidia-container-toolkit (1.8.0~rc.1-1) UNRELEASED; urgency=medium
-
-  * [libnvidia-container] Add support for cgroupv2
-  * Release toolkit-container images from nvidia-container-toolkit repository
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com> Wed, 08 Dec 2021 12:56:49 +0100
-
-nvidia-container-toolkit (1.7.0-1) UNRELEASED; urgency=medium
-
-  * Promote 1.7.0~rc.1-1 to 1.7.0-1
-  * Bump Golang version to 1.16.4
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com>  Tue, 30 Nov 2021 14:11:55 +0100
-
-nvidia-container-toolkit (1.7.0~rc.1-1) experimental; urgency=medium
-
-  * Specify containerd runtime type as string in config tools to remove dependency on containerd package
-  * Add supported-driver-capabilities config option to allow for a subset of all driver capabilities to be specified
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com>  Thu, 25 Nov 2021 11:36:29 +0100
-
-nvidia-container-toolkit (1.6.0-1) UNRELEASED; urgency=medium
-
-  * Promote 1.6.0~rc.3-1 to 1.6.0-1
-  * Fix unnecessary logging to stderr instead of configured nvidia-container-runtime log file
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com>  Wed, 17 Nov 2021 09:25:15 +0100
-
-nvidia-container-toolkit (1.6.0~rc.3-1) experimental; urgency=medium
-
-  * Add supported-driver-capabilities config option to the nvidia-container-toolkit
-  * Move OCI and command line checks for runtime to internal oci package
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com>  Mon, 15 Nov 2021 13:02:23 +0100
-
-nvidia-container-toolkit (1.6.0~rc.2-1) experimental; urgency=medium
-
-  * Use relative path to OCI specification file (config.json) if bundle path is not specified as an argument to the nvidia-container-runtime
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com>  Fri, 05 Nov 2021 12:24:05 +0200
-
-nvidia-container-toolkit (1.6.0~rc.1-1) experimental; urgency=medium
-
-  * Add AARCH64 package for Amazon Linux 2
-  * Include nvidia-container-runtime into nvidia-container-toolkit package
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com>  Mon, 06 Sep 2021 12:24:05 +0200
-
-nvidia-container-toolkit (1.5.1-1) UNRELEASED; urgency=medium
-
-  * Fix bug where Docker Swarm device selection is ignored if
-    NVIDIA_VISIBLE_DEVICES is also set
-  * Improve unit testing by using require package and adding coverage reports
-  * Remove unneeded go dependencies by running go mod tidy
-  * Move contents of pkg directory to cmd for CLI tools
-  * Ensure make binary target explicitly sets GOOS
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com>  Mon, 14 Jun 2021 09:00:00 -0700
-
-nvidia-container-toolkit (1.5.0-1) UNRELEASED; urgency=medium
-
-  * Add dependence on libnvidia-container-tools >= 1.4.0
-  * Add golang check targets to Makefile
-  * Add Jenkinsfile definition for build targets
-  * Move docker.mk to docker folder
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com>  Thu, 29 Apr 2021 03:12:43 -0700
-
-nvidia-container-toolkit (1.4.2-1) UNRELEASED; urgency=medium
-
-  * Add dependence on libnvidia-container-tools >= 1.3.3
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com>  Fri, 05 Feb 2021 02:24:36 -0700
-
-nvidia-container-toolkit (1.4.1-1) UNRELEASED; urgency=medium
-
-  * Ignore NVIDIA_VISIBLE_DEVICES for containers with insufficent privileges
-  * Add dependence on libnvidia-container-tools >= 1.3.2
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com>  Mon, 25 Jan 2021 02:18:04 -0700
-
-nvidia-container-toolkit (1.4.0-1) UNRELEASED; urgency=medium
-
-  * Add 'compute' capability to list of defaults
-  * Add dependence on libnvidia-container-tools >= 1.3.1
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com>  Fri, 11 Dec 2020 18:29:23 -0700
-
-nvidia-container-toolkit (1.3.0-1) UNRELEASED; urgency=medium
-
-  * Promote 1.3.0~rc.2-1 to 1.3.0-1
-  * Add dependence on libnvidia-container-tools >= 1.3.0
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com>  Wed, 16 Sep 2020 13:40:29 -0700
-
-nvidia-container-toolkit (1.3.0~rc.2-1) experimental; urgency=medium
-
-  * 2c180947 Add more tests for new semantics with device list from volume mounts
-  * 7c003857 Refactor accepting device lists from volume mounts as a boolean
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com>  Mon, 10 Aug 2020 15:05:34 -0700
-
-nvidia-container-toolkit (1.3.0~rc.1-1) experimental; urgency=medium
-
-  * b50d86c1 Update build system to accept a TAG variable for things like rc.x
-  * fe65573b Add common CI tests for things like golint, gofmt, unit tests, etc.
-  * da6fbb34 Revert "Add ability to merge envars of the form NVIDIA_VISIBLE_DEVICES_*"
-  * a7fb3330 Flip build-all targets to run automatically on merge requests
-  * 8b248b66 Rename github.com/NVIDIA/container-toolkit to nvidia-container-toolkit
-  * da36874e Add new config options to pull device list from mounted files instead of ENVVAR
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com>  Fri, 24 Jul 2020 22:21:49 -0700
-
-nvidia-container-toolkit (1.2.1-1) UNRELEASED; urgency=medium
-
-  * 4e6e0ed4 Add 'ngx' to list of *all* driver capabilities
-  * 2f4af743 List config.toml as a config file in the RPM SPEC
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com>  Wed, 22 Jul 2020 15:36:12 -0700
-
-nvidia-container-toolkit (1.2.0-1) UNRELEASED; urgency=medium
-
-  *  8e0aab46 Fix repo listed in changelog for debian distributions
-  *  320bb6e4 Update dependence on libnvidia-container to 1.2.0
-  *  6cfc8097 Update package license to match source license
-  *  e7dc3cbb Fix debian copyright file
-  *  d3aee3e0 Add the 'ngx' driver capability
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com>  Wed, 08 Jul 2020 18:11:19 -0700
-
-nvidia-container-toolkit (1.1.2-1) UNRELEASED; urgency=medium
-
-  * c32237f3 Add support for parsing Linux Capabilities for older OCI specs
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com>  Wed, 03 Jun 2020 12:05:32 -0700
-
-nvidia-container-toolkit (1.1.1-1) UNRELEASED; urgency=medium
-
-  * d202aded Update dependence to libnvidia-container 1.1.1
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com>  Tue, 19 May 2020 12:05:32 -0700
-
-nvidia-container-toolkit (1.1.0-1) UNRELEASED; urgency=medium
-
-  * 4e4de762 Update build system to support multi-arch builds
-  * fcc1d116 Add support for MIG (Multi-Instance GPUs)
-  * d4ff0416 Add ability to merge envars of the form NVIDIA_VISIBLE_DEVICES_*
-  * 60f165ad Add no-pivot option to toolkit
-
- -- NVIDIA CORPORATION <cudatools@nvidia.com>  Fri, 15 May 2020 12:05:32 -0700
-
-nvidia-container-toolkit (1.0.5-1) UNRELEASED; urgency=medium
-
-  * Initial release. Replaces older package nvidia-container-runtime-hook. (Closes: #XXXXXX)
-
- -- Rajat Chopra <rajatc@nvidia.com>  Wed, 10 Jul 2019 11:31:11 -0700

packaging/rpm/SPECS/nvidia-container-toolkit.spec

@@ -67,133 +67,7 @@ rm -f %{_bindir}/nvidia-container-runtime-hook
 /usr/share/containers/oci/hooks.d/oci-nvidia-hook.json
 
 %changelog
-* Fri May 13 2022 NVIDIA CORPORATION <cudatools@nvidia.com> 1.10.0-0.1.rc.2
-- Add support for NVIDIA_REQUIRE_* checks for cuda version and arch to csv mode
-- Switch to debug logging to reduce log verbosity
-- Support logging to logs requested in command line
-- Fix bug when launching containers with relative root path (e.g. using containerd)
-- Allow low-level runtime path to be set explicitly as nvidia-container-runtime.runtimes option
-- Fix failure to locate low-level runtime if PATH envvar is unset
-- Replace experimental option for NVIDIA Container Runtime with nvidia-container-runtime.mode = csv option
-- Use csv as default mode on Tegra systems without NVML
-- Add --version flag to all CLIs
-- [libnvidia-container] Bump libtirpc to 1.3.2
-- [libnvidia-container] Fix bug when running host ldconfig using glibc compiled with a non-standard prefix
-- [libnvidia-container] Add libcudadebugger.so to list of compute libraries
-
-* Thu Mar 24 2022 NVIDIA CORPORATION <cudatools@nvidia.com> 1.10.0-0.1.rc.1
-- Include nvidia-ctk CLI in installed binaries
-- Add experimental option to NVIDIA Container Runtime
-
-* Fri Mar 18 2022 NVIDIA CORPORATION <cudatools@nvidia.com> 1.9.0-1
-- [libnvidia-container] Add additional check for Tegra in /sys/.../family file in CLI
-- [libnvidia-container] Update jetpack-specific CLI option to only load Base CSV files by default
-- [libnvidia-container] Fix bug (from 1.8.0) when mounting GSP firmware into containers without /lib to /usr/lib symlinks
-- [libnvidia-container] Update nvml.h to CUDA 11.6.1 nvML_DEV 11.6.55
-- [libnvidia-container] Update switch statement to include new brands from latest nvml.h
-- [libnvidia-container] Process all --require flags on Jetson platforms
-- [libnvidia-container] Fix long-standing issue with running ldconfig on Debian systems
-
-* Mon Feb 14 2022 NVIDIA CORPORATION <cudatools@nvidia.com> 1.8.1-1
-- [libnvidia-container] Fix bug in determining cgroup root when running in nested containers
-- [libnvidia-container] Fix permission issue when determining cgroup version
-
-* Fri Feb 04 2022 NVIDIA CORPORATION <cudatools@nvidia.com> 1.8.0-1
-- Promote 1.8.0-0.1.rc.2 to 1.8.0-1
-
-* Thu Jan 20 2022 NVIDIA CORPORATION <cudatools@nvidia.com> 1.8.0-0.1.rc.2
-- Remove support for building amazonlinux1 packages
-
-* Wed Dec 08 2021 NVIDIA CORPORATION <cudatools@nvidia.com> 1.8.0-0.1.rc.1
-- [libnvidia-container] Add support for cgroupv2
-- Release toolkit-container images from nvidia-container-toolkit repository
-
-* Tue Nov 30 2021 NVIDIA CORPORATION <cudatools@nvidia.com> 1.7.0-1
-- Promote 1.7.0~rc.1-1 to 1.7.0-1
-- Bump Golang version to 1.16.4
-
-* Thu Nov 25 2021 NVIDIA CORPORATION <cudatools@nvidia.com> 1.7.0-0.1.rc.1
-- Specify containerd runtime type as string in config tools to remove dependency on containerd package
-- Add supported-driver-capabilities config option to allow for a subset of all driver capabilities to be specified
-
-* Wed Nov 17 2021 NVIDIA CORPORATION <cudatools@nvidia.com> 1.6.0-1
-- Promote 1.6.0-0.1.rc.3 to 1.6.0-1
-- Fix unnecessary logging to stderr instead of configured nvidia-container-runtime log file
-
-* Mon Nov 15 2021 NVIDIA CORPORATION <cudatools@nvidia.com> 1.6.0-0.1.rc.3
-
-- Add supported-driver-capabilities config option to the nvidia-container-toolkit
-- Move OCI and command line checks for runtime to internal oci package
-
-* Fri Nov 05 2021 NVIDIA CORPORATION <cudatools@nvidia.com> 1.6.0-0.1.rc.2
-
-- Use relative path to OCI specification file (config.json) if bundle path is not specified as an argument to the nvidia-container-runtime
-
-* Mon Sep 06 2021 NVIDIA CORPORATION <cudatools@nvidia.com> 1.6.0-0.1.rc.1
-
-- Add AARCH64 package for Amazon Linux 2
-- Include nvidia-container-runtime into nvidia-container-toolkit package
-
-* Mon Jun 14 2021 NVIDIA CORPORATION <cudatools@nvidia.com> 1.5.1-1
-
-- Fix bug where Docker Swarm device selection is ignored if NVIDIA_VISIBLE_DEVICES is also set
-- Improve unit testing by using require package and adding coverage reports
-- Remove unneeded go dependencies by running go mod tidy
-- Move contents of pkg directory to cmd for CLI tools
-- Ensure make binary target explicitly sets GOOS
-
-* Thu Apr 29 2021 NVIDIA CORPORATION <cudatools@nvidia.com> 1.5.0-1
-- Add dependence on libnvidia-container-tools >= 1.4.0
-- Add golang check targets to Makefile
-- Add Jenkinsfile definition for build targets
-- Move docker.mk to docker folder
-
-* Fri Feb 05 2021 NVIDIA CORPORATION <cudatools@nvidia.com> 1.4.2-1
-- Add dependence on libnvidia-container-tools >= 1.3.3
-
-* Mon Jan 25 2021 NVIDIA CORPORATION <cudatools@nvidia.com> 1.4.1-1
-- Ignore NVIDIA_VISIBLE_DEVICES for containers with insufficent privileges
-- Add dependence on libnvidia-container-tools >= 1.3.2
-
-* Fri Dec 11 2020 NVIDIA CORPORATION <cudatools@nvidia.com> 1.4.0-1
-- Add 'compute' capability to list of defaults
-- Add dependence on libnvidia-container-tools >= 1.3.1
-
-* Wed Sep 16 2020 NVIDIA CORPORATION <cudatools@nvidia.com> 1.3.0-1
-- Promote 1.3.0-0.1.rc.2 to 1.3.0-1
-- Add dependence on libnvidia-container-tools >= 1.3.0
-
-* Mon Aug 10 2020 NVIDIA CORPORATION <cudatools@nvidia.com> 1.3.0-0.1.rc.2
-- 2c180947 Add more tests for new semantics with device list from volume mounts
-- 7c003857 Refactor accepting device lists from volume mounts as a boolean
-
-* Fri Jul 24 2020 NVIDIA CORPORATION <cudatools@nvidia.com> 1.3.0-0.1.rc.1
-- b50d86c1 Update build system to accept a TAG variable for things like rc.x
-- fe65573b Add common CI tests for things like golint, gofmt, unit tests, etc.
-- da6fbb34 Revert "Add ability to merge envars of the form NVIDIA_VISIBLE_DEVICES_*"
-- a7fb3330 Flip build-all targets to run automatically on merge requests
-- 8b248b66 Rename github.com/NVIDIA/container-toolkit to nvidia-container-toolkit
-- da36874e Add new config options to pull device list from mounted files instead of ENVVAR
-
-* Wed Jul 22 2020 NVIDIA CORPORATION <cudatools@nvidia.com> 1.2.1-1
-- 4e6e0ed4 Add 'ngx' to list of *all* driver capabilities
-- 2f4af743 List config.toml as a config file in the RPM SPEC
-
-* Wed Jul 08 2020 NVIDIA CORPORATION <cudatools@nvidia.com> 1.2.0-1
-- 8e0aab46 Fix repo listed in changelog for debian distributions
-- 320bb6e4 Update dependence on libnvidia-container to 1.2.0
-- 6cfc8097 Update package license to match source license
-- e7dc3cbb Fix debian copyright file
-- d3aee3e0 Add the 'ngx' driver capability
-
-* Wed Jun 03 2020 NVIDIA CORPORATION <cudatools@nvidia.com> 1.1.2-1
-- c32237f3 Add support for parsing Linux Capabilities for older OCI specs
-
-* Tue May 19 2020 NVIDIA CORPORATION <cudatools@nvidia.com> 1.1.1-1
-- d202aded Update dependence to libnvidia-container 1.1.1
-
-* Fri May 15 2020 NVIDIA CORPORATION <cudatools@nvidia.com> 1.1.0-1
-- 4e4de762 Update build system to support multi-arch builds
-- fcc1d116 Add support for MIG (Multi-Instance GPUs)
-- d4ff0416 Add ability to merge envars of the form NVIDIA_VISIBLE_DEVICES_*
-- 60f165ad Add no-pivot option to toolkit
+# As of 1.10.0-1 we generate the release information automatically
+* %{release_date} NVIDIA CORPORATION <cudatools@nvidia.com> %{version}-%{release}
+- See https://gitlab.com/nvidia/container-toolkit/container-toolkit/-/blob/%{git_commit}/CHANGELOG.md
+- Bump libnvidia-container dependency to libnvidia-container-tools >= %{libnvidia_container_version}

tools/container/containerd/config.go

@@ -28,9 +28,8 @@ type UpdateReverter interface {
 
 type config struct {
 	*toml.Tree
-	version   int64
-	cri       string
-	binaryKey string
+	version int64
+	cri     string
 }
 
 // update adds the specified runtime class to the the containerd config.
@@ -48,7 +47,10 @@ func (config *config) update(runtimeClass string, runtimeType string, runtimeBin
 		config.SetPath(runtimeClassPath, runc)
 	}
 
-	config.initRuntime(runtimeClassPath, runtimeType, runtimeBinary)
+	config.initRuntime(runtimeClassPath, runtimeType, "BinaryName", runtimeBinary)
+	if config.version == 1 {
+		config.initRuntime(runtimeClassPath, runtimeType, "Runtime", runtimeBinary)
+	}
 
 	if setAsDefault {
 		defaultRuntimeNamePath := config.defaultRuntimeNamePath()
@@ -83,7 +85,7 @@ func (config *config) revert(runtimeClass string) {
 
 // initRuntime creates a runtime config if it does not exist and ensures that the
 // runtimes binary path is specified.
-func (config *config) initRuntime(path []string, runtimeType string, binary string) {
+func (config *config) initRuntime(path []string, runtimeType string, binaryKey string, binary string) {
 	if config.GetPath(path) == nil {
 		config.SetPath(append(path, "runtime_type"), runtimeType)
 		config.SetPath(append(path, "runtime_root"), "")
@@ -91,7 +93,7 @@ func (config *config) initRuntime(path []string, runtimeType string, binary stri
 		config.SetPath(append(path, "privileged_without_host_devices"), false)
 	}
 
-	binaryPath := append(path, "options", config.binaryKey)
+	binaryPath := append(path, "options", binaryKey)
 	config.SetPath(binaryPath, binary)
 }
 
@@ -99,10 +101,6 @@ func (config config) runcPath() []string {
 	return config.runtimeClassPath("runc")
 }
 
-func (config config) runtimeClassBinaryPath(runtimeClass string) []string {
-	return append(config.runtimeClassPath(runtimeClass), "options", config.binaryKey)
-}
-
 func (config config) runtimeClassPath(runtimeClass string) []string {
 	return append(config.containerdPath(), "runtimes", runtimeClass)
 }

tools/container/containerd/config_v1.go

@@ -31,10 +31,9 @@ type configV1 struct {
 func newConfigV1(cfg *toml.Tree) UpdateReverter {
 	c := configV1{
 		config: config{
-			Tree:      cfg,
-			version:   1,
-			cri:       "cri",
-			binaryKey: "Runtime",
+			Tree:    cfg,
+			version: 1,
+			cri:     "cri",
 		},
 	}
 
@@ -68,7 +67,8 @@ func (config *configV1) Update(o *options) error {
 
 		log.Warnf("Setting default_runtime is deprecated")
 		defaultRuntimePath := append(config.containerdPath(), "default_runtime")
-		config.initRuntime(defaultRuntimePath, o.runtimeType, runtimeBinary)
+		config.initRuntime(defaultRuntimePath, o.runtimeType, "Runtime", runtimeBinary)
+		config.initRuntime(defaultRuntimePath, o.runtimeType, "BinaryName", runtimeBinary)
 	}
 	return nil
 }
@@ -85,6 +85,14 @@ func (config *configV1) Revert(o *options) error {
 			}
 		}
 	}
+	if runtime, ok := config.GetPath(append(defaultRuntimeOptionsPath, "BinaryName")).(string); ok {
+		for _, runtimeBinary := range o.getRuntimeBinaries() {
+			if path.Base(runtimeBinary) == path.Base(runtime) {
+				config.DeletePath(append(defaultRuntimeOptionsPath, "BinaryName"))
+				break
+			}
+		}
+	}
 
 	if options, ok := config.GetPath(defaultRuntimeOptionsPath).(*toml.Tree); ok {
 		if len(options.Keys()) == 0 {

tools/container/containerd/config_v1_test.go

@@ -110,7 +110,7 @@ func TestUpdateV1ConfigDefaultRuntime(t *testing.T) {
 		if tc.expectedDefaultRuntimeBinary == nil {
 			require.Nil(t, defaultRuntime, "%d: %v", i, tc)
 		} else {
-			expected, err := runtimeTomlConfigV1(tc.expectedDefaultRuntimeBinary.(string))
+			expected, err := defaultRuntimeTomlConfigV1(tc.expectedDefaultRuntimeBinary.(string))
 			require.NoError(t, err, "%d: %v", i, tc)
 
 			configContents, _ := toml.Marshal(defaultRuntime.(*toml.Tree))
@@ -291,7 +291,7 @@ func TestRevertV1Config(t *testing.T) {
 								"nvidia":              runtimeMapV1("/test/runtime/dir/nvidia-container-runtime"),
 								"nvidia-experimental": runtimeMapV1("/test/runtime/dir/nvidia-container-runtime-experimental"),
 							},
-							"default_runtime":      runtimeMapV1("/test/runtime/dir/nvidia-container-runtime"),
+							"default_runtime":      defaultRuntimeV1("/test/runtime/dir/nvidia-container-runtime"),
 							"default_runtime_name": "nvidia",
 						},
 					},
@@ -325,6 +325,23 @@ func runtimeTomlConfigV1(binary string) (*toml.Tree, error) {
 	return toml.TreeFromMap(runtimeMapV1(binary))
 }
 
+func defaultRuntimeTomlConfigV1(binary string) (*toml.Tree, error) {
+	return toml.TreeFromMap(defaultRuntimeV1(binary))
+}
+
+func defaultRuntimeV1(binary string) map[string]interface{} {
+	return map[string]interface{}{
+		"runtime_type":                    runtimeType,
+		"runtime_root":                    "",
+		"runtime_engine":                  "",
+		"privileged_without_host_devices": false,
+		"options": map[string]interface{}{
+			"BinaryName": binary,
+			"Runtime":    binary,
+		},
+	}
+}
+
 func runtimeMapV1(binary string) map[string]interface{} {
 	return map[string]interface{}{
 		"runtime_type":                    runtimeType,
@@ -332,7 +349,8 @@ func runtimeMapV1(binary string) map[string]interface{} {
 		"runtime_engine":                  "",
 		"privileged_without_host_devices": false,
 		"options": map[string]interface{}{
-			"Runtime": binary,
+			"BinaryName": binary,
+			"Runtime":    binary,
 		},
 	}
 }
@@ -359,6 +377,7 @@ func runcRuntimeConfigMapV1(binary string) map[string]interface{} {
 		"privileged_without_host_devices": true,
 		"options": map[string]interface{}{
 			"runc-option": "value",
+			"BinaryName":  binary,
 			"Runtime":     binary,
 		},
 	}

tools/container/containerd/config_v2.go

@@ -28,10 +28,9 @@ type configV2 struct {
 func newConfigV2(cfg *toml.Tree) UpdateReverter {
 	c := configV2{
 		config: config{
-			Tree:      cfg,
-			version:   2,
-			cri:       "io.containerd.grpc.v1.cri",
-			binaryKey: "BinaryName",
+			Tree:    cfg,
+			version: 2,
+			cri:     "io.containerd.grpc.v1.cri",
 		},
 	}
 

versions.mk

@@ -14,7 +14,7 @@
 
 LIB_NAME := nvidia-container-toolkit
 LIB_VERSION := 1.10.0
-LIB_TAG := rc.2
+LIB_TAG := rc.3
 
 # Specify the nvidia-docker2 and nvidia-container-runtime package versions.
 # Note: The tag is automatically specified to match LIB_TAG.