Add support for setting sysctls

This patch will allow users to specify namespace specific "kernel parameters"
for running inside of a container.

Signed-off-by: Dan Walsh <dwalsh@redhat.com>

contrib/completion/bash/docker

@@ -1671,6 +1671,7 @@ _docker_run() {
 		--shm-size
 		--stop-signal
 		--tmpfs
+		--sysctl
 		--ulimit
 		--user -u
 		--userns

contrib/completion/zsh/_docker

@@ -644,6 +644,7 @@ __docker_subcommand() {
         "($help)--privileged[Give extended privileges to this container]"
         "($help)--read-only[Mount the container's root filesystem as read only]"
         "($help)*--security-opt=[Security options]:security option: "
+        "($help)*--sysctl=-[sysctl options]:sysctl: "
         "($help -t --tty)"{-t,--tty}"[Allocate a pseudo-tty]"
         "($help -u --user)"{-u=,--user=}"[Username or UID]:user:_users"
         "($help)--tmpfs[mount tmpfs]"

daemon/oci_linux.go

@@ -611,6 +611,7 @@ func (daemon *Daemon) createSpec(c *container.Container) (*libcontainerd.Spec, e
 		return nil, fmt.Errorf("linux runtime spec resources: %v", err)
 	}
 	s.Linux.Resources.OOMScoreAdj = &c.HostConfig.OomScoreAdj
+	s.Linux.Sysctl = c.HostConfig.Sysctls
 	if err := setDevices(&s, c); err != nil {
 		return nil, fmt.Errorf("linux runtime spec devices: %v", err)
 	}

docs/reference/api/docker_remote_api.md

@@ -176,6 +176,7 @@ This section lists each version from latest to oldest.  Each listing includes a
 
 [Docker Remote API v1.21](docker_remote_api_v1.21.md) documentation
 
+* `POST /containers/create` and `POST /containers/(id)/start` allow you to configure kernel parameters (sysctls) for use in the container.
 * `GET /volumes` lists volumes from all volume drivers.
 * `POST /volumes/create` to create a volume.
 * `GET /volumes/(name)` get low-level information about a volume.

docs/reference/api/docker_remote_api_v1.21.md

@@ -199,6 +199,7 @@ Create a container
              "RestartPolicy": { "Name": "", "MaximumRetryCount": 0 },
              "NetworkMode": "bridge",
              "Devices": [],
+             "Sysctls": { "net.ipv4.ip_forward": "1" },
              "Ulimits": [{}],
              "LogConfig": { "Type": "json-file", "Config": {} },
              "SecurityOpt": [],
@@ -306,6 +307,10 @@ Json Parameters:
     -   **Devices** - A list of devices to add to the container specified as a JSON object in the
       form
           `{ "PathOnHost": "/dev/deviceName", "PathInContainer": "/dev/deviceName", "CgroupPermissions": "mrw"}`
+    -   **Sysctls** - A list of kernel parameters (sysctls) to set in the container, specified as
+          `{ <name>: <Value> }`, for example:
+	  `{ "net.ipv4.ip_forward": "1" }`
+
     -   **Ulimits** - A list of ulimits to set in the container, specified as
           `{ "Name": <name>, "Soft": <soft limit>, "Hard": <hard limit> }`, for example:
           `Ulimits: { "Name": "nofile", "Soft": 1024, "Hard": 2048 }`
@@ -426,6 +431,9 @@ Return low-level information on the container `id`
 				"Type": "json-file"
 			},
 			"SecurityOpt": null,
+			"Sysctls": {
+			        "net.ipv4.ip_forward": "1"
+			},
 			"VolumesFrom": null,
 			"Ulimits": [{}],
 			"VolumeDriver": ""

docs/reference/commandline/create.md

@@ -82,6 +82,7 @@ Creates a new container.
       --stop-signal="SIGTERM"       Signal to stop a container
       --shm-size=[]                 Size of `/dev/shm`. The format is `<number><unit>`. `number` must be greater than `0`.  Unit is optional and can be `b` (bytes), `k` (kilobytes), `m` (megabytes), or `g` (gigabytes). If you omit the unit, the system uses bytes. If you omit the size entirely, the system uses `64m`.
       --storage-opt=[]              Set storage driver options per container
+      --sysctl[=*[]*]]              Configure namespaced kernel parameters at runtime
       -t, --tty                     Allocate a pseudo-TTY
       -u, --user=""                 Username or UID
       --userns=""                   Container user namespace

docs/reference/commandline/run.md

@@ -84,6 +84,7 @@ parent = "smn_cli"
       --sig-proxy=true              Proxy received signals to the process
       --stop-signal="SIGTERM"       Signal to stop a container
       --storage-opt=[]              Set storage driver options per container
+      --sysctl[=*[]*]]              Configure namespaced kernel parameters at runtime
       -t, --tty                     Allocate a pseudo-TTY
       -u, --user=""                 Username or UID (format: <name|uid>[:<group|gid>])
       --userns=""                   Container user namespace
@@ -620,3 +621,30 @@ If you have set the `--exec-opt isolation=hyperv` option on the Docker `daemon`,
 $ docker run -d --isolation default busybox top
 $ docker run -d --isolation hyperv busybox top
 ```
+
+### Configure namespaced kernel parameters (sysctls) at runtime
+
+The `--sysctl` sets namespaced kernel parameters (sysctls) in the
+container. For example, to turn on IP forwarding in the containers
+network namespace, run this command:
+
+    $ docker run --sysctl net.ipv4.ip_forward=1 someimage
+
+
+> **Note**: Not all sysctls are namespaced. docker does not support changing sysctls
+> inside of a container that also modify the host system. As the kernel 
+> evolves we expect to see more sysctls become namespaced.
+
+#### Currently supported sysctls
+
+  `IPC Namespace`:
+
+  kernel.msgmax, kernel.msgmnb, kernel.msgmni, kernel.sem, kernel.shmall, kernel.shmmax, kernel.shmmni, kernel.shm_rmid_forced
+  Sysctls beginning with fs.mqueue.*
+
+  If you use the `--ipc=host` option these sysctls will not be allowed.
+
+  `Network Namespace`:
+      Sysctls beginning with net.*
+
+  If you use the `--net=host` option using these sysctls will not be allowed.

integration-cli/docker_cli_run_unix_test.go

@@ -4,6 +4,7 @@ package main
 
 import (
 	"bufio"
+	"encoding/json"
 	"fmt"
 	"io/ioutil"
 	"os"
@@ -747,6 +748,37 @@ func (s *DockerSuite) TestRunTmpfsMounts(c *check.C) {
 	}
 }
 
+func (s *DockerSuite) TestRunSysctls(c *check.C) {
+
+	testRequires(c, DaemonIsLinux)
+	var err error
+
+	out, _ := dockerCmd(c, "run", "--sysctl", "net.ipv4.ip_forward=1", "--name", "test", "busybox", "cat", "/proc/sys/net/ipv4/ip_forward")
+	c.Assert(strings.TrimSpace(out), check.Equals, "1")
+
+	out = inspectFieldJSON(c, "test", "HostConfig.Sysctls")
+
+	sysctls := make(map[string]string)
+	err = json.Unmarshal([]byte(out), &sysctls)
+	c.Assert(err, check.IsNil)
+	c.Assert(sysctls["net.ipv4.ip_forward"], check.Equals, "1")
+
+	out, _ = dockerCmd(c, "run", "--sysctl", "net.ipv4.ip_forward=0", "--name", "test1", "busybox", "cat", "/proc/sys/net/ipv4/ip_forward")
+	c.Assert(strings.TrimSpace(out), check.Equals, "0")
+
+	out = inspectFieldJSON(c, "test1", "HostConfig.Sysctls")
+
+	err = json.Unmarshal([]byte(out), &sysctls)
+	c.Assert(err, check.IsNil)
+	c.Assert(sysctls["net.ipv4.ip_forward"], check.Equals, "0")
+
+	runCmd := exec.Command(dockerBinary, "run", "--sysctl", "kernel.foobar=1", "--name", "test2", "busybox", "cat", "/proc/sys/kernel/foobar")
+	out, _, _ = runCommandWithOutput(runCmd)
+	if !strings.Contains(out, "invalid value") {
+		c.Fatalf("expected --sysctl to fail, got %s", out)
+	}
+}
+
 // TestRunSeccompProfileDenyUnshare checks that 'docker run --security-opt seccomp=/tmp/profile.json debian:jessie unshare' exits with operation not permitted.
 func (s *DockerSuite) TestRunSeccompProfileDenyUnshare(c *check.C) {
 	testRequires(c, SameHostDaemon, seccompEnabled, NotArm, Apparmor)

man/docker-create.1.md

@@ -67,6 +67,7 @@ docker-create - Create a new container
 [**--storage-opt**[=*[]*]]
 [**--stop-signal**[=*SIGNAL*]]
 [**--shm-size**[=*[]*]]
+[**--sysctl**[=*[]*]]
 [**-t**|**--tty**]
 [**--tmpfs**[=*[CONTAINER-DIR[:<OPTIONS>]*]]
 [**-u**|**--user**[=*USER*]]
@@ -336,6 +337,21 @@ unit, `b` is used. Set LIMIT to `-1` to enable unlimited swap.
 **--stop-signal**=*SIGTERM*
   Signal to stop a container. Default is SIGTERM.
 
+**--sysctl**=SYSCTL
+  Configure namespaced kernel parameters at runtime
+
+  IPC Namespace - current sysctls allowed:
+
+  kernel.msgmax, kernel.msgmnb, kernel.msgmni, kernel.sem, kernel.shmall, kernel.shmmax, kernel.shmmni, kernel.shm_rmid_forced
+  Sysctls beginning with fs.mqueue.*
+
+  Note: if you use --ipc=host using these sysctls will not be allowed.
+
+  Network Namespace - current sysctls allowed:
+      Sysctls beginning with net.*
+
+  Note: if you use --net=host using these sysctls will not be allowed.
+
 **-t**, **--tty**=*true*|*false*
    Allocate a pseudo-TTY. The default is *false*.
 

man/docker-run.1.md

@@ -71,6 +71,7 @@ docker-run - Run a command in a new container
 [**--stop-signal**[=*SIGNAL*]]
 [**--shm-size**[=*[]*]]
 [**--sig-proxy**[=*true*]]
+[**--sysctl**[=*[]*]]
 [**-t**|**--tty**]
 [**--tmpfs**[=*[CONTAINER-DIR[:<OPTIONS>]*]]
 [**-u**|**--user**[=*USER*]]
@@ -492,6 +493,21 @@ its root filesystem mounted as read only prohibiting any writes.
    `number` must be greater than `0`.  Unit is optional and can be `b` (bytes), `k` (kilobytes), `m`(megabytes), or `g` (gigabytes).
    If you omit the unit, the system uses bytes. If you omit the size entirely, the system uses `64m`.
 
+**--sysctl**=SYSCTL
+  Configure namespaced kernel parameters at runtime
+
+  IPC Namespace - current sysctls allowed:
+
+  kernel.msgmax, kernel.msgmnb, kernel.msgmni, kernel.sem, kernel.shmall, kernel.shmmax, kernel.shmmni, kernel.shm_rmid_forced
+  Sysctls beginning with fs.mqueue.*
+
+  If you use the `--ipc=host` option these sysctls will not be allowed.
+
+  Network Namespace - current sysctls allowed:
+      Sysctls beginning with net.*
+
+  If you use the `--net=host` option these sysctls will not be allowed.
+
 **--sig-proxy**=*true*|*false*
    Proxy received signals to the process (non-TTY mode only). SIGCHLD, SIGSTOP, and SIGKILL are not proxied. The default is *true*.
 
@@ -955,6 +971,23 @@ $ docker run -d --isolation default busybox top
 $ docker run -d --isolation hyperv busybox top
 ```
 
+## Setting Namespaced Kernel Parameters (Sysctls)
+
+The `--sysctl` sets namespaced kernel parameters (sysctls) in the
+container. For example, to turn on IP forwarding in the containers
+network namespace, run this command:
+
+    $ docker run --sysctl net.ipv4.ip_forward=1 someimage
+
+Note:
+
+Not all sysctls are namespaced. docker does not support changing sysctls
+inside of a container that also modify the host system. As the kernel 
+evolves we expect to see more sysctls become namespaced.
+
+See the definition of the `--sysctl` option above for the current list of 
+supported sysctls.
+
 # HISTORY
 April 2014, Originally compiled by William Henry (whenry at redhat dot com)
 based on docker.com source material and internal work.

opts/opts.go

@@ -240,3 +240,35 @@ func ValidateLabel(val string) (string, error) {
 	}
 	return val, nil
 }
+
+// ValidateSysctl validates an sysctl and returns it.
+func ValidateSysctl(val string) (string, error) {
+	validSysctlMap := map[string]bool{
+		"kernel.msgmax":          true,
+		"kernel.msgmnb":          true,
+		"kernel.msgmni":          true,
+		"kernel.sem":             true,
+		"kernel.shmall":          true,
+		"kernel.shmmax":          true,
+		"kernel.shmmni":          true,
+		"kernel.shm_rmid_forced": true,
+	}
+	validSysctlPrefixes := []string{
+		"net.",
+		"fs.mqueue.",
+	}
+	arr := strings.Split(val, "=")
+	if len(arr) < 2 {
+		return "", fmt.Errorf("sysctl '%s' is not whitelisted", val)
+	}
+	if validSysctlMap[arr[0]] {
+		return val, nil
+	}
+
+	for _, vp := range validSysctlPrefixes {
+		if strings.HasPrefix(arr[0], vp) {
+			return val, nil
+		}
+	}
+	return "", fmt.Errorf("sysctl '%s' is not whitelisted", val)
+}

runconfig/opts/parse.go

@@ -42,6 +42,7 @@ func Parse(cmd *flag.FlagSet, args []string) (*container.Config, *container.Host
 		flDevices           = opts.NewListOpts(ValidateDevice)
 
 		flUlimits = NewUlimitOpt(nil)
+		flSysctls = opts.NewMapOpts(nil, opts.ValidateSysctl)
 
 		flPublish           = opts.NewListOpts(nil)
 		flExpose            = opts.NewListOpts(nil)
@@ -127,6 +128,7 @@ func Parse(cmd *flag.FlagSet, args []string) (*container.Config, *container.Host
 	cmd.Var(&flSecurityOpt, []string{"-security-opt"}, "Security Options")
 	cmd.Var(&flStorageOpt, []string{"-storage-opt"}, "Set storage driver options per container")
 	cmd.Var(flUlimits, []string{"-ulimit"}, "Ulimit options")
+	cmd.Var(flSysctls, []string{"-sysctl"}, "Sysctl options")
 	cmd.Var(&flLoggingOpts, []string{"-log-opt"}, "Log driver options")
 
 	cmd.Require(flag.Min, 1)
@@ -430,6 +432,7 @@ func Parse(cmd *flag.FlagSet, args []string) (*container.Config, *container.Host
 		ShmSize:        shmSize,
 		Resources:      resources,
 		Tmpfs:          tmpfs,
+		Sysctls:        flSysctls.GetAll(),
 	}
 
 	// When allocating stdin in attached mode, close stdin at client disconnect