CVE-2020-15257

containerd is an industry-standard container runtime and is available as a daemon for Linux and Windows. In containerd before versions 1.3.9 and 1.4.3, the containerd-shim API is improperly exposed to host network containers. Access controls for the shim’s API socket verified that the connecting process had an effective UID of 0, but did not otherwise restrict access to the abstract Unix domain socket. This would allow malicious containers running in the same network namespace as the shim, with an effective UID of 0 but otherwise reduced privileges, to cause new processes to be run with elevated privileges. This vulnerability has been fixed in containerd 1.3.9 and 1.4.3. Users should update to these versions as soon as they are released. It should be noted that containers started with an old version of containerd-shim should be stopped and restarted, as running containers will continue to be vulnerable even after an upgrade. If you are not providing the ability for untrusted users to start containers in the same network namespace as the shim (typically the "host" network namespace, for example with docker run --net=host or hostNetwork: true in a Kubernetes pod) and run with an effective UID of 0, you are not vulnerable to this issue. If you are running containers with a vulnerable configuration, you can deny access to all abstract sockets with AppArmor by adding a line similar to deny unix addr=@**, to your policy. It is best practice to run containers with a reduced set of privileges, with a non-zero UID, and with isolated namespaces. The containerd maintainers strongly advise against sharing namespaces with the host. Reducing the set of isolation mechanisms used for a container necessarily increases that container's privilege, regardless of what container runtime is used for running that container.

containerd is an industry-standard container runtime and is available as a daemon for Linux and Windows. In containerd before versions 1.3.9 and 1.4.3, the containerd-shim API is improperly exposed to host network containers. Access controls for the shim’s API socket verified that the connecting process had an effective UID of 0, but did not otherwise restrict access to the abstract Unix domain socket. This would allow malicious containers running in the same network namespace as the shim, with an effective UID of 0 but otherwise reduced privileges, to cause new processes to be run with elevated privileges. This vulnerability has been fixed in containerd 1.3.9 and 1.4.3. Users should update to these versions as soon as they are released. It should be noted that containers started with an old version of containerd-shim should be stopped and restarted, as running containers will continue to be vulnerable even after an upgrade. If you are not providing the ability for untrusted users to start containers in the same network namespace as the shim (typically the "host" network namespace, for example with docker run --net=host or hostNetwork: true in a Kubernetes pod) and run with an effective UID of 0, you are not vulnerable to this issue. If you are running containers with a vulnerable configuration, you can deny access to all abstract sockets with AppArmor by adding a line similar to deny unix addr=@**, to your policy. It is best practice to run containers with a reduced set of privileges, with a non-zero UID, and with isolated namespaces. The containerd maintainers strongly advise against sharing namespaces with the host. Reducing the set of isolation mechanisms used for a container necessarily increases that container's privilege, regardless of what container runtime is used for running that container.

containerd-shim API Exposed to Host Network Containers in github.com/containerd/containerd

## Impact Access controls for the shim’s API socket verified that the connecting process had an effective UID of 0, but did not otherwise restrict access to the abstract Unix domain socket. This would allow malicious containers running in the same network namespace as the shim, with an effective UID of 0 but otherwise reduced privileges, to cause new processes to be run with elevated privileges. ### Specific Go Packages Affected github.com/containerd/containerd/cmd ## Patches This vulnerability has been fixed in containerd 1.3.9 and 1.4.3. Users should update to these versions as soon as they are released. It should be noted that containers started with an old version of containerd-shim should be stopped and restarted, as running containers will continue to be vulnerable even after an upgrade. ## Workarounds If you are not providing the ability for untrusted users to start containers in the same network namespace as the shim (typically the "host" network namespace, for example with `docker run --net=host` or `hostNetwork: true` in a Kubernetes pod) and run with an effective UID of 0, you are not vulnerable to this issue. If you are running containers with a vulnerable configuration, you can deny access to all abstract sockets with AppArmor by adding a line similar to `deny unix addr=@**,` to your policy. It is best practice to run containers with a reduced set of privileges, with a non-zero UID, and with isolated namespaces. The containerd maintainers strongly advise against sharing namespaces with the host. Reducing the set of isolation mechanisms used for a container necessarily increases that container's privilege, regardless of what container runtime is used for running that container. ## Credits The containerd maintainers would like to thank Jeff Dileo of NCC Group for responsibly disclosing this issue in accordance with the [containerd security policy](https://github.com/containerd/project/blob/master/SECURITY.md) and for reviewing the patch. ## For more information If you have any questions or comments about this advisory: * [Open an issue](https://github.com/containerd/containerd/issues/new/choose) * Email us at security@containerd.io if you think you’ve found a security bug.

containerd es un tiempo de ejecución de contenedor estándar de la industria y está disponible como demonio para Linux y Windows. En containerd anterior a las versiones 1.3.9 y 1.4.3, la API containerd-shim está expuesta inapropiadamente a los contenedores de red del host. Los controles de acceso para el socket de la API de shim verificaron que el proceso de conexión tuviera un UID efectivo de 0, pero no restringieron de otra manera el acceso al socket de dominio Unix abstracto. Esto permitiría que los contenedores maliciosos se ejecuten en el mismo espacio de nombres de red que el shim, con un UID efectivo de 0 pero con privilegios reducidos, para causar que nuevos procesos se ejecuten con privilegios elevados. Esta vulnerabilidad se ha corregido en containerd versiones 1.3.9 y 1.4.3. Los usuarios deben actualizar a estas versiones tan pronto como se publiquen. Cabe señalar que los contenedores iniciados con una versión anterior de containerd-shim deben detenerse y reiniciarse, ya que los contenedores en ejecución seguirán siendo vulnerables inclusive después de una actualización. Si no proporciona la capacidad para que los usuarios que no son de confianza inicien contenedores en el mismo espacio de nombres de red que el shim (normalmente el espacio de nombres de red "host", por ejemplo, con docker run --net=host o hostNetwork: true en un pod de Kubernetes) y ejecutar con un UID efectivo de 0, no es vulnerable a este problema. Si está ejecutando contenedores con una configuración vulnerable, puede denegar el acceso a todos los sockets abstractos con AppArmor agregando una línea similar a denegar unix addr=@**, para su política. Es una buena práctica ejecutar contenedores con un conjunto reducido de privilegios, con un UID distinto de cero y con espacios de nombres aislados. Los encargados de mantenimiento de contenedores no aconsejan compartir espacios de nombres con el host. Reducir el conjunto de mecanismos de aislamiento usados para un contenedor necesariamente aumenta el privilegio de ese contenedor, independientemente del tiempo de ejecución del contenedor que se use para ejecutar ese contenedor