diff --git a/src/runtime/go.mod b/src/runtime/go.mod index ab3fdcbc13..07d3932279 100644 --- a/src/runtime/go.mod +++ b/src/runtime/go.mod @@ -29,6 +29,7 @@ require ( github.com/hashicorp/go-multierror v1.0.0 github.com/intel-go/cpuid v0.0.0-20210602155658-5747e5cec0d9 github.com/mdlayher/vsock v0.0.0-20191108225356-d9c65923cb8f + github.com/opencontainers/image-spec v1.0.2 // indirect github.com/opencontainers/runc v1.0.3 github.com/opencontainers/runtime-spec v1.0.3-0.20210326190908-1c3f411f0417 github.com/opencontainers/selinux v1.8.2 @@ -58,7 +59,8 @@ require ( ) replace ( - github.com/opencontainers/runc => github.com/opencontainers/runc v1.0.1 + github.com/opencontainers/image-spec => github.com/opencontainers/image-spec v1.0.2 + github.com/opencontainers/runc => github.com/opencontainers/runc v1.0.3 github.com/uber-go/atomic => go.uber.org/atomic v1.5.1 google.golang.org/genproto => google.golang.org/genproto v0.0.0-20180817151627-c66870c02cf8 ) diff --git a/src/runtime/go.sum b/src/runtime/go.sum index da4fcc3fe0..3bc3ba026c 100644 --- a/src/runtime/go.sum +++ b/src/runtime/go.sum @@ -691,12 +691,10 @@ github.com/opencontainers/go-digest v1.0.0-rc1/go.mod h1:cMLVZDEM3+U2I4VmLI6N8jQ github.com/opencontainers/go-digest v1.0.0-rc1.0.20180430190053-c9281466c8b2/go.mod h1:cMLVZDEM3+U2I4VmLI6N8jQYUd2OVphdqWwCJHrFt2s= github.com/opencontainers/go-digest v1.0.0 h1:apOUWs51W5PlhuyGyz9FCeeBIOUDA/6nW8Oi/yOhh5U= github.com/opencontainers/go-digest v1.0.0/go.mod h1:0JzlMkj0TRzQZfJkVvzbP0HBR3IKzErnv2BNG4W4MAM= -github.com/opencontainers/image-spec v1.0.0/go.mod h1:BtxoFyWECRxE4U/7sNtV5W15zMzWCbyJoFRP3s7yZA0= -github.com/opencontainers/image-spec v1.0.1/go.mod h1:BtxoFyWECRxE4U/7sNtV5W15zMzWCbyJoFRP3s7yZA0= -github.com/opencontainers/image-spec v1.0.2-0.20211117181255-693428a734f5 h1:q37d91F6BO4Jp1UqWiun0dUFYaqv6WsKTLTCaWv+8LY= -github.com/opencontainers/image-spec v1.0.2-0.20211117181255-693428a734f5/go.mod h1:BtxoFyWECRxE4U/7sNtV5W15zMzWCbyJoFRP3s7yZA0= -github.com/opencontainers/runc v1.0.1 h1:G18PGckGdAm3yVQRWDVQ1rLSLntiniKJ0cNRT2Tm5gs= -github.com/opencontainers/runc v1.0.1/go.mod h1:aTaHFFwQXuA71CiyxOdFFIorAoemI04suvGRQFzWTD0= +github.com/opencontainers/image-spec v1.0.2 h1:9yCKha/T5XdGtO0q9Q9a6T5NUCsTn/DrBg0D7ufOcFM= +github.com/opencontainers/image-spec v1.0.2/go.mod h1:BtxoFyWECRxE4U/7sNtV5W15zMzWCbyJoFRP3s7yZA0= +github.com/opencontainers/runc v1.0.3 h1:1hbqejyQWCJBvtKAfdO0b1FmaEf2z/bxnjqbARass5k= +github.com/opencontainers/runc v1.0.3/go.mod h1:aTaHFFwQXuA71CiyxOdFFIorAoemI04suvGRQFzWTD0= github.com/opencontainers/runtime-spec v0.1.2-0.20190507144316-5b71a03e2700/go.mod h1:jwyrGlmzljRJv/Fgzds9SsS/C5hL+LL3ko9hs6T5lQ0= github.com/opencontainers/runtime-spec v1.0.1/go.mod h1:jwyrGlmzljRJv/Fgzds9SsS/C5hL+LL3ko9hs6T5lQ0= github.com/opencontainers/runtime-spec v1.0.2-0.20190207185410-29686dbc5559/go.mod h1:jwyrGlmzljRJv/Fgzds9SsS/C5hL+LL3ko9hs6T5lQ0= diff --git a/src/runtime/vendor/github.com/opencontainers/image-spec/specs-go/v1/annotations.go b/src/runtime/vendor/github.com/opencontainers/image-spec/specs-go/v1/annotations.go index 581cf7cdfa..35d8108958 100644 --- a/src/runtime/vendor/github.com/opencontainers/image-spec/specs-go/v1/annotations.go +++ b/src/runtime/vendor/github.com/opencontainers/image-spec/specs-go/v1/annotations.go @@ -53,10 +53,4 @@ const ( // AnnotationDescription is the annotation key for the human-readable description of the software packaged in the image. AnnotationDescription = "org.opencontainers.image.description" - - // AnnotationBaseImageDigest is the annotation key for the digest of the image's base image. - AnnotationBaseImageDigest = "org.opencontainers.image.base.digest" - - // AnnotationBaseImageName is the annotation key for the image reference of the image's base image. - AnnotationBaseImageName = "org.opencontainers.image.base.name" ) diff --git a/src/runtime/vendor/github.com/opencontainers/image-spec/specs-go/v1/config.go b/src/runtime/vendor/github.com/opencontainers/image-spec/specs-go/v1/config.go index ffff4b6d18..fe799bd698 100644 --- a/src/runtime/vendor/github.com/opencontainers/image-spec/specs-go/v1/config.go +++ b/src/runtime/vendor/github.com/opencontainers/image-spec/specs-go/v1/config.go @@ -89,20 +89,9 @@ type Image struct { // Architecture is the CPU architecture which the binaries in this image are built to run on. Architecture string `json:"architecture"` - // Variant is the variant of the specified CPU architecture which image binaries are intended to run on. - Variant string `json:"variant,omitempty"` - // OS is the name of the operating system which the image is built to run on. OS string `json:"os"` - // OSVersion is an optional field specifying the operating system - // version, for example on Windows `10.0.14393.1066`. - OSVersion string `json:"os.version,omitempty"` - - // OSFeatures is an optional field specifying an array of strings, - // each listing a required OS feature (for example on Windows `win32k`). - OSFeatures []string `json:"os.features,omitempty"` - // Config defines the execution parameters which should be used as a base when running a container using the image. Config ImageConfig `json:"config,omitempty"` diff --git a/src/runtime/vendor/github.com/opencontainers/image-spec/specs-go/v1/mediatype.go b/src/runtime/vendor/github.com/opencontainers/image-spec/specs-go/v1/mediatype.go index 4f35ac134f..bad7bb97f4 100644 --- a/src/runtime/vendor/github.com/opencontainers/image-spec/specs-go/v1/mediatype.go +++ b/src/runtime/vendor/github.com/opencontainers/image-spec/specs-go/v1/mediatype.go @@ -34,10 +34,6 @@ const ( // referenced by the manifest. MediaTypeImageLayerGzip = "application/vnd.oci.image.layer.v1.tar+gzip" - // MediaTypeImageLayerZstd is the media type used for zstd compressed - // layers referenced by the manifest. - MediaTypeImageLayerZstd = "application/vnd.oci.image.layer.v1.tar+zstd" - // MediaTypeImageLayerNonDistributable is the media type for layers referenced by // the manifest but with distribution restrictions. MediaTypeImageLayerNonDistributable = "application/vnd.oci.image.layer.nondistributable.v1.tar" @@ -47,11 +43,6 @@ const ( // restrictions. MediaTypeImageLayerNonDistributableGzip = "application/vnd.oci.image.layer.nondistributable.v1.tar+gzip" - // MediaTypeImageLayerNonDistributableZstd is the media type for zstd - // compressed layers referenced by the manifest but with distribution - // restrictions. - MediaTypeImageLayerNonDistributableZstd = "application/vnd.oci.image.layer.nondistributable.v1.tar+zstd" - // MediaTypeImageConfig specifies the media type for the image configuration. MediaTypeImageConfig = "application/vnd.oci.image.config.v1+json" ) diff --git a/src/runtime/vendor/github.com/opencontainers/image-spec/specs-go/version.go b/src/runtime/vendor/github.com/opencontainers/image-spec/specs-go/version.go index 31f99cf645..0d9543f160 100644 --- a/src/runtime/vendor/github.com/opencontainers/image-spec/specs-go/version.go +++ b/src/runtime/vendor/github.com/opencontainers/image-spec/specs-go/version.go @@ -25,7 +25,7 @@ const ( VersionPatch = 2 // VersionDev indicates development branch. Releases will be empty string. - VersionDev = "-dev" + VersionDev = "" ) // Version is the specification version that the package types support. diff --git a/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/fs/cpu.go b/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/fs/cpu.go index 31c1c874ea..7ca8b8a8d4 100644 --- a/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/fs/cpu.go +++ b/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/fs/cpu.go @@ -4,6 +4,7 @@ package fs import ( "bufio" + "errors" "fmt" "os" "strconv" @@ -11,6 +12,7 @@ import ( "github.com/opencontainers/runc/libcontainer/cgroups" "github.com/opencontainers/runc/libcontainer/cgroups/fscommon" "github.com/opencontainers/runc/libcontainer/configs" + "golang.org/x/sys/unix" ) type CpuGroup struct{} @@ -71,15 +73,33 @@ func (s *CpuGroup) Set(path string, r *configs.Resources) error { return fmt.Errorf("the minimum allowed cpu-shares is %d", sharesRead) } } + + var period string if r.CpuPeriod != 0 { - if err := cgroups.WriteFile(path, "cpu.cfs_period_us", strconv.FormatUint(r.CpuPeriod, 10)); err != nil { - return err + period = strconv.FormatUint(r.CpuPeriod, 10) + if err := cgroups.WriteFile(path, "cpu.cfs_period_us", period); err != nil { + // Sometimes when the period to be set is smaller + // than the current one, it is rejected by the kernel + // (EINVAL) as old_quota/new_period exceeds the parent + // cgroup quota limit. If this happens and the quota is + // going to be set, ignore the error for now and retry + // after setting the quota. + if !errors.Is(err, unix.EINVAL) || r.CpuQuota == 0 { + return err + } + } else { + period = "" } } if r.CpuQuota != 0 { if err := cgroups.WriteFile(path, "cpu.cfs_quota_us", strconv.FormatInt(r.CpuQuota, 10)); err != nil { return err } + if period != "" { + if err := cgroups.WriteFile(path, "cpu.cfs_period_us", period); err != nil { + return err + } + } } return s.SetRtSched(path, r) } diff --git a/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/fs2/hugetlb.go b/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/fs2/hugetlb.go index 3f513975bd..96390a224f 100644 --- a/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/fs2/hugetlb.go +++ b/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/fs2/hugetlb.go @@ -30,10 +30,7 @@ func setHugeTlb(dirPath string, r *configs.Resources) error { } func statHugeTlb(dirPath string, stats *cgroups.Stats) error { - hugePageSizes, err := cgroups.GetHugePageSize() - if err != nil { - return errors.Wrap(err, "failed to fetch hugetlb info") - } + hugePageSizes, _ := cgroups.GetHugePageSize() hugetlbStats := cgroups.HugetlbStats{} for _, pagesize := range hugePageSizes { diff --git a/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/systemd/common.go b/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/systemd/common.go index 3506c82746..06c6c2e95a 100644 --- a/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/systemd/common.go +++ b/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/systemd/common.go @@ -310,6 +310,14 @@ func getUnitName(c *configs.Cgroup) string { return c.Name } +// This code should be in sync with getUnitName. +func getUnitType(unitName string) string { + if strings.HasSuffix(unitName, ".slice") { + return "Slice" + } + return "Scope" +} + // isDbusError returns true if the error is a specific dbus error. func isDbusError(err error, name string) bool { if err != nil { @@ -388,10 +396,10 @@ func resetFailedUnit(cm *dbusConnManager, name string) { } } -func getUnitProperty(cm *dbusConnManager, unitName string, propertyName string) (*systemdDbus.Property, error) { +func getUnitTypeProperty(cm *dbusConnManager, unitName string, unitType string, propertyName string) (*systemdDbus.Property, error) { var prop *systemdDbus.Property err := cm.retryOnDisconnect(func(c *systemdDbus.Conn) (Err error) { - prop, Err = c.GetUnitPropertyContext(context.TODO(), unitName, propertyName) + prop, Err = c.GetUnitTypePropertyContext(context.TODO(), unitName, unitType, propertyName) return Err }) return prop, err diff --git a/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/systemd/dbus.go b/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/systemd/dbus.go index a70a9df43e..0690564556 100644 --- a/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/systemd/dbus.go +++ b/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/systemd/dbus.go @@ -4,6 +4,7 @@ package systemd import ( "context" + "fmt" "sync" systemdDbus "github.com/coreos/go-systemd/v22/dbus" @@ -54,7 +55,10 @@ func (d *dbusConnManager) getConnection() (*systemdDbus.Conn, error) { conn, err := d.newConnection() if err != nil { - return nil, err + // When dbus-user-session is not installed, we can't detect whether we should try to connect to user dbus or system dbus, so d.dbusRootless is set to false. + // This may fail with a cryptic error "read unix @->/run/systemd/private: read: connection reset by peer: unknown." + // https://github.com/moby/moby/issues/42793 + return nil, fmt.Errorf("failed to connect to dbus (hint: for rootless containers, maybe you need to install dbus-user-session package, see https://github.com/opencontainers/runc/blob/master/docs/cgroup-v2.md): %w", err) } dbusC = conn return conn, nil diff --git a/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/systemd/v1.go b/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/systemd/v1.go index 1a8e1e3c6c..cd4720c5d4 100644 --- a/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/systemd/v1.go +++ b/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/systemd/v1.go @@ -6,6 +6,7 @@ import ( "errors" "os" "path/filepath" + "reflect" "strings" "sync" @@ -345,6 +346,11 @@ func (m *legacyManager) freezeBeforeSet(unitName string, r *configs.Resources) ( // Special case for SkipDevices, as used by Kubernetes to create pod // cgroups with allow-all device policy). if r.SkipDevices { + if r.SkipFreezeOnSet { + // Both needsFreeze and needsThaw are false. + return + } + // No need to freeze if SkipDevices is set, and either // (1) systemd unit does not (yet) exist, or // (2) it has DevicePolicy=auto and empty DeviceAllow list. @@ -353,15 +359,20 @@ func (m *legacyManager) freezeBeforeSet(unitName string, r *configs.Resources) ( // a non-existent unit returns default properties, // and settings in (2) are the defaults. // - // Do not return errors from getUnitProperty, as they alone + // Do not return errors from getUnitTypeProperty, as they alone // should not prevent Set from working. - devPolicy, e := getUnitProperty(m.dbus, unitName, "DevicePolicy") + + unitType := getUnitType(unitName) + + devPolicy, e := getUnitTypeProperty(m.dbus, unitName, unitType, "DevicePolicy") if e == nil && devPolicy.Value == dbus.MakeVariant("auto") { - devAllow, e := getUnitProperty(m.dbus, unitName, "DeviceAllow") - if e == nil && devAllow.Value == dbus.MakeVariant([]deviceAllowEntry{}) { - needsFreeze = false - needsThaw = false - return + devAllow, e := getUnitTypeProperty(m.dbus, unitName, unitType, "DeviceAllow") + if e == nil { + if rv := reflect.ValueOf(devAllow.Value.Value()); rv.Kind() == reflect.Slice && rv.Len() == 0 { + needsFreeze = false + needsThaw = false + return + } } } } diff --git a/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/systemd/v2.go b/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/systemd/v2.go index 8176ce5b2e..55273b722c 100644 --- a/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/systemd/v2.go +++ b/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/cgroups/systemd/v2.go @@ -5,7 +5,6 @@ package systemd import ( "fmt" "math" - "os" "path/filepath" "strconv" "strings" @@ -307,9 +306,10 @@ func (m *unifiedManager) Destroy() error { return err } - // XXX this is probably not needed, systemd should handle it - err := os.Remove(m.path) - if err != nil && !os.IsNotExist(err) { + // systemd 239 do not remove sub-cgroups. + err := cgroups.RemovePath(m.path) + // cgroups.RemovePath has handled ErrNotExist + if err != nil { return err } diff --git a/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/configs/cgroup_linux.go b/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/configs/cgroup_linux.go index a1e7f0afd4..5ea9d940ce 100644 --- a/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/configs/cgroup_linux.go +++ b/src/runtime/vendor/github.com/opencontainers/runc/libcontainer/configs/cgroup_linux.go @@ -131,4 +131,16 @@ type Resources struct { // // NOTE it is impossible to start a container which has this flag set. SkipDevices bool `json:"-"` + + // SkipFreezeOnSet is a flag for cgroup manager to skip the cgroup + // freeze when setting resources. Only applicable to systemd legacy + // (i.e. cgroup v1) manager (which uses freeze by default to avoid + // spurious permission errors caused by systemd inability to update + // device rules in a non-disruptive manner). + // + // If not set, a few methods (such as looking into cgroup's + // devices.list and querying the systemd unit properties) are used + // during Set() to figure out whether the freeze is required. Those + // methods may be relatively slow, thus this flag. + SkipFreezeOnSet bool `json:"-"` } diff --git a/src/runtime/vendor/modules.txt b/src/runtime/vendor/modules.txt index f52a619fca..f663a80330 100644 --- a/src/runtime/vendor/modules.txt +++ b/src/runtime/vendor/modules.txt @@ -221,10 +221,11 @@ github.com/mitchellh/mapstructure github.com/moby/sys/mountinfo # github.com/opencontainers/go-digest v1.0.0 github.com/opencontainers/go-digest -# github.com/opencontainers/image-spec v1.0.2-0.20211117181255-693428a734f5 +# github.com/opencontainers/image-spec v1.0.2 => github.com/opencontainers/image-spec v1.0.2 +## explicit github.com/opencontainers/image-spec/specs-go github.com/opencontainers/image-spec/specs-go/v1 -# github.com/opencontainers/runc v1.0.3 => github.com/opencontainers/runc v1.0.1 +# github.com/opencontainers/runc v1.0.3 => github.com/opencontainers/runc v1.0.3 ## explicit github.com/opencontainers/runc/libcontainer/cgroups github.com/opencontainers/runc/libcontainer/cgroups/devices @@ -457,6 +458,7 @@ k8s.io/apimachinery/pkg/api/resource # k8s.io/cri-api v0.23.0-alpha.4 ## explicit k8s.io/cri-api/pkg/apis/runtime/v1alpha2 -# github.com/opencontainers/runc => github.com/opencontainers/runc v1.0.1 +# github.com/opencontainers/image-spec => github.com/opencontainers/image-spec v1.0.2 +# github.com/opencontainers/runc => github.com/opencontainers/runc v1.0.3 # github.com/uber-go/atomic => go.uber.org/atomic v1.5.1 # google.golang.org/genproto => google.golang.org/genproto v0.0.0-20180817151627-c66870c02cf8