Skip to content
Permalink
master
Switch branches/tags
Go to file
@rata
Latest commit 58798e7 Nov 4, 2020 History
This adds the specification for Seccomp Userspace Notification and the
Golang bindings. This contains:
- New fields in the seccomp section to use with seccomp userspace
  notification.
- Additional SeccompState struct containing the container state and file
  descriptors passed for seccomp.

This was discussed in the OCI Weekly Discussion on September 16th,
2020. After review on github, this implementation was changed to the
"Proposal with listenerPath and listenerExtraMetadata". For more
information see:
- #1073 (comment)

Docs presented on the community meeting (for the old implementation
using hooks):
- https://hackmd.io/El8Dd2xrTlCaCG59ns5cwg#September-16-2020
- https://docs.google.com/document/d/1xHw5GQjMj6ZKR-40aKmTWZRkvlPuzMGQRu-YpOFQc30/edit

Documentation for this feature:
- https://www.kernel.org/doc/html/v5.0/userspace-api/seccomp_filter.html#userspace-notification
- man pages: seccomp_user_notif.2 at
  https://git.kernel.org/pub/scm/docs/man-pages/man-pages.git/log/?h=seccomp_user_notif
- brauner's blog:
  https://brauner.github.io/2020/07/23/seccomp-notify.html

This PR is an alternative proposal to PR 1038. While similar in nature,
the main difference is that this PR adds optional metadata to be sent to
the seccomp agent and specifies how the UNIX socket MUST be used.

Signed-off-by: Rodrigo Campos <rodrigo@kinvolk.io>
Signed-off-by: Alban Crequy <alban@kinvolk.io>
Signed-off-by: Mauricio Vásquez <mauricio@kinvolk.io>
49 contributors

Users who have contributed to this file

@wking @mrunalp @hqhq @crosbymichael @vbatts @giuseppe @cyphar @philips @xiaochenshen @laijs @dqminh @thaJeztah

Linux Container Configuration

This document describes the schema for the Linux-specific section of the container configuration. The Linux container specification uses various kernel features like namespaces, cgroups, capabilities, LSM, and filesystem jails to fulfill the spec.

Default Filesystems

The Linux ABI includes both syscalls and several special file paths. Applications expecting a Linux environment will very likely expect these file paths to be set up correctly.

The following filesystems SHOULD be made available in each container's filesystem:

Path Type
/proc proc
/sys sysfs
/dev/pts devpts
/dev/shm tmpfs

Namespaces

A namespace wraps a global system resource in an abstraction that makes it appear to the processes within the namespace that they have their own isolated instance of the global resource. Changes to the global resource are visible to other processes that are members of the namespace, but are invisible to other processes. For more information, see the namespaces(7) man page.

Namespaces are specified as an array of entries inside the namespaces root field. The following parameters can be specified to set up namespaces:

  • type (string, REQUIRED) - namespace type. The following namespace types SHOULD be supported:

    • pid processes inside the container will only be able to see other processes inside the same container or inside the same pid namespace.
    • network the container will have its own network stack.
    • mount the container will have an isolated mount table.
    • ipc processes inside the container will only be able to communicate to other processes inside the same container via system level IPC.
    • uts the container will be able to have its own hostname and domain name.
    • user the container will be able to remap user and group IDs from the host to local users and groups within the container.
    • cgroup the container will have an isolated view of the cgroup hierarchy.
  • path (string, OPTIONAL) - namespace file. This value MUST be an absolute path in the runtime mount namespace. The runtime MUST place the container process in the namespace associated with that path. The runtime MUST generate an error if path is not associated with a namespace of type type.

    If path is not specified, the runtime MUST create a new container namespace of type type.

If a namespace type is not specified in the namespaces array, the container MUST inherit the runtime namespace of that type. If a namespaces field contains duplicated namespaces with same type, the runtime MUST generate an error.

Example

"namespaces": [
    {
        "type": "pid",
        "path": "/proc/1234/ns/pid"
    },
    {
        "type": "network",
        "path": "/var/run/netns/neta"
    },
    {
        "type": "mount"
    },
    {
        "type": "ipc"
    },
    {
        "type": "uts"
    },
    {
        "type": "user"
    },
    {
        "type": "cgroup"
    }
]

User namespace mappings

uidMappings (array of objects, OPTIONAL) describes the user namespace uid mappings from the host to the container. gidMappings (array of objects, OPTIONAL) describes the user namespace gid mappings from the host to the container.

Each entry has the following structure:

  • containerID (uint32, REQUIRED) - is the starting uid/gid in the container.
  • hostID (uint32, REQUIRED) - is the starting uid/gid on the host to be mapped to containerID.
  • size (uint32, REQUIRED) - is the number of ids to be mapped.

The runtime SHOULD NOT modify the ownership of referenced filesystems to realize the mapping. Note that the number of mapping entries MAY be limited by the kernel.

Example

"uidMappings": [
    {
        "containerID": 0,
        "hostID": 1000,
        "size": 32000
    }
],
"gidMappings": [
    {
        "containerID": 0,
        "hostID": 1000,
        "size": 32000
    }
]

Devices

devices (array of objects, OPTIONAL) lists devices that MUST be available in the container. The runtime MAY supply them however it likes (with mknod, by bind mounting from the runtime mount namespace, using symlinks, etc.).

Each entry has the following structure:

  • type (string, REQUIRED) - type of device: c, b, u or p. More info in mknod(1).
  • path (string, REQUIRED) - full path to device inside container. If a file already exists at path that does not match the requested device, the runtime MUST generate an error.
  • major, minor (int64, REQUIRED unless type is p) - major, minor numbers for the device.
  • fileMode (uint32, OPTIONAL) - file mode for the device. You can also control access to devices with cgroups.
  • uid (uint32, OPTIONAL) - id of device owner in the container namespace.
  • gid (uint32, OPTIONAL) - id of device group in the container namespace.

The same type, major and minor SHOULD NOT be used for multiple devices.

Example

"devices": [
    {
        "path": "/dev/fuse",
        "type": "c",
        "major": 10,
        "minor": 229,
        "fileMode": 438,
        "uid": 0,
        "gid": 0
    },
    {
        "path": "/dev/sda",
        "type": "b",
        "major": 8,
        "minor": 0,
        "fileMode": 432,
        "uid": 0,
        "gid": 0
    }
]

Default Devices

In addition to any devices configured with this setting, the runtime MUST also supply:

Control groups

Also known as cgroups, they are used to restrict resource usage for a container and handle device access. cgroups provide controls (through controllers) to restrict cpu, memory, IO, pids, network and RDMA resources for the container. For more information, see the kernel cgroups documentation.

Cgroups Path

cgroupsPath (string, OPTIONAL) path to the cgroups. It can be used to either control the cgroups hierarchy for containers or to run a new process in an existing container.

The value of cgroupsPath MUST be either an absolute path or a relative path.

  • In the case of an absolute path (starting with /), the runtime MUST take the path to be relative to the cgroups mount point.
  • In the case of a relative path (not starting with /), the runtime MAY interpret the path relative to a runtime-determined location in the cgroups hierarchy.

If the value is specified, the runtime MUST consistently attach to the same place in the cgroups hierarchy given the same value of cgroupsPath. If the value is not specified, the runtime MAY define the default cgroups path. Runtimes MAY consider certain cgroupsPath values to be invalid, and MUST generate an error if this is the case.

Implementations of the Spec can choose to name cgroups in any manner. The Spec does not include naming schema for cgroups. The Spec does not support per-controller paths for the reasons discussed in the cgroupv2 documentation. The cgroups will be created if they don't exist.

You can configure a container's cgroups via the resources field of the Linux configuration. Do not specify resources unless limits have to be updated. For example, to run a new process in an existing container without updating limits, resources need not be specified.

Runtimes MAY attach the container process to additional cgroup controllers beyond those necessary to fulfill the resources settings.

Example

"cgroupsPath": "/myRuntime/myContainer",
"resources": {
    "memory": {
    "limit": 100000,
    "reservation": 200000
    },
    "devices": [
        {
            "allow": false,
            "access": "rwm"
        }
    ]
}

Allowed Device list

devices (array of objects, OPTIONAL) configures the allowed device list. The runtime MUST apply entries in the listed order.

Each entry has the following structure:

  • allow (boolean, REQUIRED) - whether the entry is allowed or denied.
  • type (string, OPTIONAL) - type of device: a (all), c (char), or b (block). Unset values mean "all", mapping to a.
  • major, minor (int64, OPTIONAL) - major, minor numbers for the device. Unset values mean "all", mapping to * in the filesystem API.
  • access (string, OPTIONAL) - cgroup permissions for device. A composition of r (read), w (write), and m (mknod).

Example

"devices": [
    {
        "allow": false,
        "access": "rwm"
    },
    {
        "allow": true,
        "type": "c",
        "major": 10,
        "minor": 229,
        "access": "rw"
    },
    {
        "allow": true,
        "type": "b",
        "major": 8,
        "minor": 0,
        "access": "r"
    }
]

Memory

memory (object, OPTIONAL) represents the cgroup subsystem memory and it's used to set limits on the container's memory usage. For more information, see the kernel cgroups documentation about memory.

Values for memory specify the limit in bytes, or -1 for unlimited memory.

  • limit (int64, OPTIONAL) - sets limit of memory usage
  • reservation (int64, OPTIONAL) - sets soft limit of memory usage
  • swap (int64, OPTIONAL) - sets limit of memory+Swap usage
  • kernel (int64, OPTIONAL, NOT RECOMMENDED) - sets hard limit for kernel memory
  • kernelTCP (int64, OPTIONAL, NOT RECOMMENDED) - sets hard limit for kernel TCP buffer memory

The following properties do not specify memory limits, but are covered by the memory controller:

  • swappiness (uint64, OPTIONAL) - sets swappiness parameter of vmscan (See sysctl's vm.swappiness) The values are from 0 to 100. Higher means more swappy.
  • disableOOMKiller (bool, OPTIONAL) - enables or disables the OOM killer. If enabled (false), tasks that attempt to consume more memory than they are allowed are immediately killed by the OOM killer. The OOM killer is enabled by default in every cgroup using the memory subsystem. To disable it, specify a value of true.
  • useHierarchy (bool, OPTIONAL) - enables or disables hierarchical memory accounting. If enabled (true), child cgroups will share the memory limits of this cgroup.

Example

"memory": {
    "limit": 536870912,
    "reservation": 536870912,
    "swap": 536870912,
    "kernel": -1,
    "kernelTCP": -1,
    "swappiness": 0,
    "disableOOMKiller": false
}

CPU

cpu (object, OPTIONAL) represents the cgroup subsystems cpu and cpusets. For more information, see the kernel cgroups documentation about cpusets.

The following parameters can be specified to set up the controller:

  • shares (uint64, OPTIONAL) - specifies a relative share of CPU time available to the tasks in a cgroup
  • quota (int64, OPTIONAL) - specifies the total amount of time in microseconds for which all tasks in a cgroup can run during one period (as defined by period below)
  • period (uint64, OPTIONAL) - specifies a period of time in microseconds for how regularly a cgroup's access to CPU resources should be reallocated (CFS scheduler only)
  • realtimeRuntime (int64, OPTIONAL) - specifies a period of time in microseconds for the longest continuous period in which the tasks in a cgroup have access to CPU resources
  • realtimePeriod (uint64, OPTIONAL) - same as period but applies to realtime scheduler only
  • cpus (string, OPTIONAL) - list of CPUs the container will run in
  • mems (string, OPTIONAL) - list of Memory Nodes the container will run in

Example

"cpu": {
    "shares": 1024,
    "quota": 1000000,
    "period": 500000,
    "realtimeRuntime": 950000,
    "realtimePeriod": 1000000,
    "cpus": "2-3",
    "mems": "0-7"
}

Block IO

blockIO (object, OPTIONAL) represents the cgroup subsystem blkio which implements the block IO controller. For more information, see the kernel cgroups documentation about blkio.

The following parameters can be specified to set up the controller:

  • weight (uint16, OPTIONAL) - specifies per-cgroup weight. This is default weight of the group on all devices until and unless overridden by per-device rules.

  • leafWeight (uint16, OPTIONAL) - equivalents of weight for the purpose of deciding how much weight tasks in the given cgroup has while competing with the cgroup's child cgroups.

  • weightDevice (array of objects, OPTIONAL) - an array of per-device bandwidth weights. Each entry has the following structure:

    • major, minor (int64, REQUIRED) - major, minor numbers for device. For more information, see the mknod(1) man page.
    • weight (uint16, OPTIONAL) - bandwidth weight for the device.
    • leafWeight (uint16, OPTIONAL) - bandwidth weight for the device while competing with the cgroup's child cgroups, CFQ scheduler only

    You MUST specify at least one of weight or leafWeight in a given entry, and MAY specify both.

  • throttleReadBpsDevice, throttleWriteBpsDevice (array of objects, OPTIONAL) - an array of per-device bandwidth rate limits. Each entry has the following structure:

    • major, minor (int64, REQUIRED) - major, minor numbers for device. For more information, see the mknod(1) man page.
    • rate (uint64, REQUIRED) - bandwidth rate limit in bytes per second for the device
  • throttleReadIOPSDevice, throttleWriteIOPSDevice (array of objects, OPTIONAL) - an array of per-device IO rate limits. Each entry has the following structure:

    • major, minor (int64, REQUIRED) - major, minor numbers for device. For more information, see the mknod(1) man page.
    • rate (uint64, REQUIRED) - IO rate limit for the device

Example

"blockIO": {
    "weight": 10,
    "leafWeight": 10,
    "weightDevice": [
        {
            "major": 8,
            "minor": 0,
            "weight": 500,
            "leafWeight": 300
        },
        {
            "major": 8,
            "minor": 16,
            "weight": 500
        }
    ],
    "throttleReadBpsDevice": [
        {
            "major": 8,
            "minor": 0,
            "rate": 600
        }
    ],
    "throttleWriteIOPSDevice": [
        {
            "major": 8,
            "minor": 16,
            "rate": 300
        }
    ]
}

Huge page limits

hugepageLimits (array of objects, OPTIONAL) represents the hugetlb controller which allows to limit the HugeTLB usage per control group and enforces the controller limit during page fault. For more information, see the kernel cgroups documentation about HugeTLB.

Each entry has the following structure:

  • pageSize (string, REQUIRED) - hugepage size The value has the format <size><unit-prefix>B (64KB, 2MB, 1GB), and must match the <hugepagesize> of the corresponding control file found in /sys/fs/cgroup/hugetlb/hugetlb.<hugepagesize>.limit_in_bytes. Values of <unit-prefix> are intended to be parsed using base 1024 ("1KB" = 1024, "1MB" = 1048576, etc).
  • limit (uint64, REQUIRED) - limit in bytes of hugepagesize HugeTLB usage

Example

"hugepageLimits": [
    {
        "pageSize": "2MB",
        "limit": 209715200
    },
    {
        "pageSize": "64KB",
        "limit": 1000000
    }
]

Network

network (object, OPTIONAL) represents the cgroup subsystems net_cls and net_prio. For more information, see the kernel cgroups documentations about net_cls cgroup and net_prio cgroup.

The following parameters can be specified to set up the controller:

  • classID (uint32, OPTIONAL) - is the network class identifier the cgroup's network packets will be tagged with
  • priorities (array of objects, OPTIONAL) - specifies a list of objects of the priorities assigned to traffic originating from processes in the group and egressing the system on various interfaces. The following parameters can be specified per-priority:
    • name (string, REQUIRED) - interface name in runtime network namespace
    • priority (uint32, REQUIRED) - priority applied to the interface

Example

"network": {
    "classID": 1048577,
    "priorities": [
        {
            "name": "eth0",
            "priority": 500
        },
        {
            "name": "eth1",
            "priority": 1000
        }
    ]
}

PIDs

pids (object, OPTIONAL) represents the cgroup subsystem pids. For more information, see the kernel cgroups documentation about pids.

The following parameters can be specified to set up the controller:

  • limit (int64, REQUIRED) - specifies the maximum number of tasks in the cgroup

Example

"pids": {
    "limit": 32771
}

RDMA

rdma (object, OPTIONAL) represents the cgroup subsystem rdma. For more information, see the kernel cgroups documentation about rdma.

The name of the device to limit is the entry key. Entry values are objects with the following properties:

  • hcaHandles (uint32, OPTIONAL) - specifies the maximum number of hca_handles in the cgroup
  • hcaObjects (uint32, OPTIONAL) - specifies the maximum number of hca_objects in the cgroup

You MUST specify at least one of the hcaHandles or hcaObjects in a given entry, and MAY specify both.

Example

"rdma": {
    "mlx5_1": {
        "hcaHandles": 3,
        "hcaObjects": 10000
    },
    "mlx4_0": {
        "hcaObjects": 1000
    },
    "rxe3": {
        "hcaObjects": 10000
    }
}

Unified

unified (object, OPTIONAL) allows cgroup v2 parameters to be to be set and modified for the container.

Each key in the map refers to a file in the cgroup unified hierarchy.

The OCI runtime MUST ensure that the needed cgroup controllers are enabled for the cgroup.

Configuration unknown to the runtime MUST still be written to the relevant file.

The runtime MUST generate an error when the configuration refers to a cgroup controller that is not present or that cannot be enabled.

Example

"unified": {
    "io.max": "259:0 rbps=2097152 wiops=120\n253:0 rbps=2097152 wiops=120",
    "hugetlb.1GB.max": "1073741824"
}

If a controller is enabled on the cgroup v2 hierarchy but the configuration is provided for the cgroup v1 equivalent controller, the runtime MAY attempt a conversion.

If the conversion is not possible the runtime MUST generate an error.

IntelRdt

intelRdt (object, OPTIONAL) represents the Intel Resource Director Technology. If intelRdt is set, the runtime MUST write the container process ID to the tasks file in a proper sub-directory in a mounted resctrl pseudo-filesystem. That sub-directory name is specified by closID parameter. If no mounted resctrl pseudo-filesystem is available in the runtime mount namespace, the runtime MUST generate an error.

If intelRdt is not set, the runtime MUST NOT manipulate any resctrl pseudo-filesystems.

The following parameters can be specified for the container:

  • closID (string, OPTIONAL) - specifies the identity for RDT Class of Service (CLOS). If closID is set, runtimes MUST create closID directory in a mounted resctrl pseudo-filesystem if it doesn't exist. If not set, runtimes MUST use the container ID from start and create the <container-id> directory.

  • l3CacheSchema (string, OPTIONAL) - specifies the schema for L3 cache id and capacity bitmask (CBM). The value SHOULD start with L3: and SHOULD NOT contain newlines.

  • memBwSchema (string, OPTIONAL) - specifies the schema of memory bandwidth per L3 cache id.

    • The value MUST start with MB: and MUST NOT contain newlines.

    • If both l3CacheSchema and memBwSchema are set, runtimes MUST write the combined value to the schemata file in that sub-directory discussed in closID.

    • If l3CacheSchema contains a line beginning with MB:, the value written to schemata file MUST be the non-MB: line(s) from l3CacheSchema and the line from memBWSchema.

    • If either l3CacheSchema or memBwSchema is set, runtimes MUST write the value to the schemata file in the that sub-directory discussed in closID.

    • If neither l3CacheSchema nor memBwSchema is set, runtimes MUST NOT write to schemata files in any resctrl pseudo-filesystems.

    • If closID is set, l3CacheSchema and/or memBwSchema is set, runtimes MUST compare l3CacheSchema and/or memBwSchema value with schemata file, and generate an error if doesn't match.

    • If closID is set, and neither of l3CacheSchema and memBwSchema are set, runtime MUST check if corresponding pre-configured directory closID is present in mounted resctrl. If such pre-configured directory closID exists, runtime MUST assign container to this closID and generate an error if directory does not exist.

Example

Consider a two-socket machine with two L3 caches where the default CBM is 0x7ff and the max CBM length is 11 bits, and minimum memory bandwidth of 10% with a memory bandwidth granularity of 10%.

Tasks inside the container only have access to the "upper" 7/11 of L3 cache on socket 0 and the "lower" 5/11 L3 cache on socket 1, and may use a maximum memory bandwidth of 20% on socket 0 and 70% on socket 1.

"linux": {
    "intelRdt": {
        "closID": "guaranteed_group",
        "l3CacheSchema": "L3:0=7f0;1=1f",
        "memBwSchema": "MB:0=20;1=70"
    }
}

Sysctl

sysctl (object, OPTIONAL) allows kernel parameters to be modified at runtime for the container. For more information, see the sysctl(8) man page.

Example

"sysctl": {
    "net.ipv4.ip_forward": "1",
    "net.core.somaxconn": "256"
}

Seccomp

Seccomp provides application sandboxing mechanism in the Linux kernel. Seccomp configuration allows one to configure actions to take for matched syscalls and furthermore also allows matching on values passed as arguments to syscalls. For more information about Seccomp, see Seccomp kernel documentation. The actions, architectures, and operators are strings that match the definitions in seccomp.h from libseccomp and are translated to corresponding values.

seccomp (object, OPTIONAL)

The following parameters can be specified to set up seccomp:

  • defaultAction (string, REQUIRED) - the default action for seccomp. Allowed values are the same as syscalls[].action.

  • defaultErrnoRet (uint, OPTIONAL) - the errno return code to use. Some actions like SCMP_ACT_ERRNO and SCMP_ACT_TRACE allow to specify the errno code to return. When the action doesn't support an errno, the runtime MUST print and error and fail. If not specified then its default value is EPERM.

  • architectures (array of strings, OPTIONAL) - the architecture used for system calls. A valid list of constants as of libseccomp v2.5.0 is shown below.

    • SCMP_ARCH_X86
    • SCMP_ARCH_X86_64
    • SCMP_ARCH_X32
    • SCMP_ARCH_ARM
    • SCMP_ARCH_AARCH64
    • SCMP_ARCH_MIPS
    • SCMP_ARCH_MIPS64
    • SCMP_ARCH_MIPS64N32
    • SCMP_ARCH_MIPSEL
    • SCMP_ARCH_MIPSEL64
    • SCMP_ARCH_MIPSEL64N32
    • SCMP_ARCH_PPC
    • SCMP_ARCH_PPC64
    • SCMP_ARCH_PPC64LE
    • SCMP_ARCH_S390
    • SCMP_ARCH_S390X
    • SCMP_ARCH_PARISC
    • SCMP_ARCH_PARISC64
    • SCMP_ARCH_RISCV64
  • flags (array of strings, OPTIONAL) - list of flags to use with seccomp(2).

    A valid list of constants is shown below.

    • SECCOMP_FILTER_FLAG_TSYNC
    • SECCOMP_FILTER_FLAG_LOG
    • SECCOMP_FILTER_FLAG_SPEC_ALLOW
  • listenerPath (string, OPTIONAL) - specifies the path of UNIX domain socket over which the runtime will send the container process state data structure when the SCMP_ACT_NOTIFY action is used. This socket MUST use AF_UNIX domain and SOCK_STREAM type. The runtime MUST send exactly one container process state per connection. The connection MUST NOT be reused and it MUST be closed after sending a seccomp state. If sending to this socket fails, the runtime MUST generate an error. If the SCMP_ACT_NOTIFY action is not used this value is ignored.

    The runtime sends the following file descriptors using SCM_RIGHTS and set their names in the fds array of the container process state:

    • seccompFd (string, REQUIRED) is the seccomp file descriptor returned by the seccomp syscall.
  • listenerMetadata (string, OPTIONAL) - specifies an opaque data to pass to the seccomp agent. This string will be sent as the metadata field in the container process state. This field MUST NOT be set if listenerPath is not set.

  • syscalls (array of objects, OPTIONAL) - match a syscall in seccomp. While this property is OPTIONAL, some values of defaultAction are not useful without syscalls entries. For example, if defaultAction is SCMP_ACT_KILL and syscalls is empty or unset, the kernel will kill the container process on its first syscall. Each entry has the following structure:

    • names (array of strings, REQUIRED) - the names of the syscalls. names MUST contain at least one entry.

    • action (string, REQUIRED) - the action for seccomp rules. A valid list of constants as of libseccomp v2.5.0 is shown below.

      • SCMP_ACT_KILL
      • SCMP_ACT_KILL_PROCESS
      • SCMP_ACT_KILL_THREAD
      • SCMP_ACT_TRAP
      • SCMP_ACT_ERRNO
      • SCMP_ACT_TRACE
      • SCMP_ACT_ALLOW
      • SCMP_ACT_LOG
      • SCMP_ACT_NOTIFY
    • errnoRet (uint, OPTIONAL) - the errno return code to use. Some actions like SCMP_ACT_ERRNO and SCMP_ACT_TRACE allow to specify the errno code to return. When the action doesn't support an errno, the runtime MUST print and error and fail. If not specified its default value is EPERM.

    • args (array of objects, OPTIONAL) - the specific syscall in seccomp. Each entry has the following structure:

      • index (uint, REQUIRED) - the index for syscall arguments in seccomp.

      • value (uint64, REQUIRED) - the value for syscall arguments in seccomp.

      • valueTwo (uint64, OPTIONAL) - the value for syscall arguments in seccomp.

      • op (string, REQUIRED) - the operator for syscall arguments in seccomp. A valid list of constants as of libseccomp v2.3.2 is shown below.

        • SCMP_CMP_NE
        • SCMP_CMP_LT
        • SCMP_CMP_LE
        • SCMP_CMP_EQ
        • SCMP_CMP_GE
        • SCMP_CMP_GT
        • SCMP_CMP_MASKED_EQ

Example

"seccomp": {
    "defaultAction": "SCMP_ACT_ALLOW",
    "architectures": [
        "SCMP_ARCH_X86",
        "SCMP_ARCH_X32"
    ],
    "syscalls": [
        {
            "names": [
                "getcwd",
                "chmod"
            ],
            "action": "SCMP_ACT_ERRNO"
        }
    ]
}

The Container Process State

The container process state is a data structure passed via a UNIX socket. The container runtime MUST send the container process state over the UNIX socket as regular payload serialized in JSON and file descriptors MUST be sent using SCM_RIGHTS. The container runtime MAY use several sendmsg(2) calls to send the aforementioned data. If more than one sendmsg(2) is used, the file descriptors MUST be sent only in the first call.

The container process state includes the following properties:

Example sending a single seccompFD file descriptor in the SCM_RIGHTS array:

{
    "ociVersion": "0.2.0",
    "fds": [
        "seccompFd"
    ],
    "pid": 4422,
    "metadata": "MKNOD=/dev/null,/dev/net/tun;BPF_MAP_TYPES=hash,array",
    "state": {
        "ociVersion": "0.2.0",
        "id": "oci-container1",
        "status": "creating",
        "pid": 4422,
        "bundle": "/containers/redis",
        "annotations": {
            "myKey": "myValue"
        }
    }
}

Rootfs Mount Propagation

rootfsPropagation (string, OPTIONAL) sets the rootfs's mount propagation. Its value is either shared, slave, private or unbindable. It's worth noting that a peer group is defined as a group of VFS mounts that propagate events to each other. A nested container is defined as a container launched inside an existing container.

  • shared: the rootfs mount belongs to a new peer group. This means that further mounts (e.g. nested containers) will also belong to that peer group and will propagate events to the rootfs. Note this does not mean that it's shared with the host.
  • slave: the rootfs mount receives propagation events from the host (e.g. if something is mounted on the host it will also appear in the container) but not the other way around.
  • private: the rootfs mount doesn't receive mount propagation events from the host and further mounts in nested containers will be isolated from the host and from the rootfs (even if the nested container rootfsPropagation option is shared).
  • unbindable: the rootfs mount is a private mount that cannot be bind-mounted.

The Shared Subtrees article in the kernel documentation has more information about mount propagation.

Example

"rootfsPropagation": "slave",

Masked Paths

maskedPaths (array of strings, OPTIONAL) will mask over the provided paths inside the container so that they cannot be read. The values MUST be absolute paths in the container namespace.

Example

"maskedPaths": [
    "/proc/kcore"
]

Readonly Paths

readonlyPaths (array of strings, OPTIONAL) will set the provided paths as readonly inside the container. The values MUST be absolute paths in the container namespace.

Example

"readonlyPaths": [
    "/proc/sys"
]

Mount Label

mountLabel (string, OPTIONAL) will set the Selinux context for the mounts in the container.

Example

"mountLabel": "system_u:object_r:svirt_sandbox_file_t:s0:c715,c811"

Personality

personality (object, OPTIONAL) sets the Linux execution personality. For more information see the personality syscall documentation. As most of the options are obsolete and rarely used, and some reduce security, the currently supported set is a small subset of the available options.

  • domain (string, REQUIRED) - the execution domain. The valid list of constants is shown below. LINUX32 will set the uname system call to show a 32 bit CPU type, such as i686.

    • LINUX
    • LINUX32
  • flags (array of strings, OPTIONAL) - the additional flags to apply. Currently no flag values are supported.