Container Security vulnerabilities

A flaw was found in Podman. This issue may allow an attacker to create a specially crafted container that, when configured to share the same IPC with at least one other container, can create a large number of IPC resources in /dev/shm. The malicious container will continue to exhaust resources until it is out-of-memory (OOM) killed. While the malicious container's cgroup will be removed, the IPC resources it created are not. Those resources are tied to the IPC namespace that will not be removed until all containers using it are stopped, and one non-malicious container is holding the namespace open. The malicious container is restarted, either automatically or by attacker control, repeating the process and increasing the amount of memory consumed. With a container configured to restart always, such as `podman run --restart=always`, this can result in a memory-based denial of service of the system.

Moby is an open-source project created by Docker for software containerization. A security vulnerability has been detected in certain versions of Docker Engine, which could allow an attacker to bypass authorization plugins (AuthZ) under specific circumstances. The base likelihood of this being exploited is low. Using a specially-crafted API request, an Engine API client could make the daemon forward the request or response to an authorization plugin without the body. In certain circumstances, the authorization plugin may allow a request which it would have otherwise denied if the body had been forwarded to it. A security issue was discovered In 2018, where an attacker could bypass AuthZ plugins using a specially crafted API request. This could lead to unauthorized actions, including privilege escalation. Although this issue was fixed in Docker Engine v18.09.1 in January 2019, the fix was not carried forward to later major versions, resulting in a regression. Anyone who depends on authorization plugins that introspect the request and/or response body to make access control decisions is potentially impacted. Docker EE v19.03.x and all versions of Mirantis Container Runtime are not vulnerable. docker-ce v27.1.1 containes patches to fix the vulnerability. Patches have also been merged into the master, 19.03, 20.0, 23.0, 24.0, 25.0, 26.0, and 26.1 release branches. If one is unable to upgrade immediately, avoid using AuthZ plugins and/or restrict access to the Docker API to trusted parties, following the principle of least privilege.

In Docker Desktop before v4.29.0, an attacker who has gained access to the Docker Desktop VM through a container breakout can further escape to the host by passing extensions and dashboard related IPC messages. Docker Desktop v4.29.0 https://docs.docker.com/desktop/release-notes/#4290 fixes the issue on MacOS, Linux and Windows with Hyper-V backend. As exploitation requires "Allow only extensions distributed through the Docker Marketplace" to be disabled, Docker Desktop  v4.31.0 https://docs.docker.com/desktop/release-notes/#4310  additionally changes the default configuration to enable this setting by default.

In Docker Desktop on Windows before v4.31.0 allows a user in the docker-users group to cause a Windows Denial-of-Service through the exec-path Docker daemon config option in Windows containers mode.

A security regression (CVE-2006-5051) was discovered in OpenSSH's server (sshd). There is a race condition which can lead sshd to handle some signals in an unsafe manner. An unauthenticated, remote attacker may be able to trigger it by failing to authenticate within a set time period.

When generating the systemd service units for the docker snap (and other similar snaps), snapd does not specify Delegate=yes - as a result systemd will move processes from the containers created and managed by these snaps into the cgroup of the main daemon within the snap itself when reloading system units. This may grant additional privileges to a container within the snap that were not originally intended.

A flaw was found in cri-o. A malicious container can create a symbolic link to arbitrary files on the host via directory traversal (“../“). This flaw allows the container to read and write to arbitrary files on the host system.

Argo CD is a declarative, GitOps continuous delivery tool for Kubernetes. It has been discovered that an unprivileged pod in a different namespace on the same cluster could connect to the Redis server on port 6379. Despite having installed the latest version of the VPC CNI plugin on the EKS cluster, it requires manual enablement through configuration to enforce network policies. This raises concerns that many clients might unknowingly have open access to their Redis servers. This vulnerability could lead to Privilege Escalation to the level of cluster controller, or to information leakage, affecting anyone who does not have strict access controls on their Redis instance. This issue has been patched in version(s) 2.8.19, 2.9.15 and 2.10.10.

A flaw was found in cri-o, where an arbitrary systemd property can be injected via a Pod annotation. Any user who can create a pod with an arbitrary annotation may perform an arbitrary action on the host system.

Moby is an open source container framework that is a key component of Docker Engine, Docker Desktop, and other distributions of container tooling or runtimes. In 26.0.0, IPv6 is not disabled on network interfaces, including those belonging to networks where `--ipv6=false`. An container with an `ipvlan` or `macvlan` interface will normally be configured to share an external network link with the host machine. Because of this direct access, (1) Containers may be able to communicate with other hosts on the local network over link-local IPv6 addresses, (2) if router advertisements are being broadcast over the local network, containers may get SLAAC-assigned addresses, and (3) the interface will be a member of IPv6 multicast groups. This means interfaces in IPv4-only networks present an unexpectedly and unnecessarily increased attack surface. The issue is patched in 26.0.2. To completely disable IPv6 in a container, use `--sysctl=net.ipv6.conf.all.disable_ipv6=1` in the `docker create` or `docker run` command. Or, in the service configuration of a `compose` file.

Malicious code was discovered in the upstream tarballs of xz, starting with version 5.6.0. Through a series of complex obfuscations, the liblzma build process extracts a prebuilt object file from a disguised test file existing in the source code, which is then used to modify specific functions in the liblzma code. This results in a modified liblzma library that can be used by any software linked against this library, intercepting and modifying the data interaction with this library.

Moby is an open source container framework that is a key component of Docker Engine, Docker Desktop, and other distributions of container tooling or runtimes. Moby's networking implementation allows for many networks, each with their own IP address range and gateway, to be defined. This feature is frequently referred to as custom networks, as each network can have a different driver, set of parameters and thus behaviors. When creating a network, the `--internal` flag is used to designate a network as _internal_. The `internal` attribute in a docker-compose.yml file may also be used to mark a network _internal_, and other API clients may specify the `internal` parameter as well. When containers with networking are created, they are assigned unique network interfaces and IP addresses. The host serves as a router for non-internal networks, with a gateway IP that provides SNAT/DNAT to/from container IPs. Containers on an internal network may communicate between each other, but are precluded from communicating with any networks the host has access to (LAN or WAN) as no default route is configured, and firewall rules are set up to drop all outgoing traffic. Communication with the gateway IP address (and thus appropriately configured host services) is possible, and the host may communicate with any container IP directly. In addition to configuring the Linux kernel's various networking features to enable container networking, `dockerd` directly provides some services to container networks. Principal among these is serving as a resolver, enabling service discovery, and resolution of names from an upstream resolver. When a DNS request for a name that does not correspond to a container is received, the request is forwarded to the configured upstream resolver. This request is made from the container's network namespace: the level of access and routing of traffic is the same as if the request was made by the container itself. As a consequence of this design, containers solely attached to an internal network will be unable to resolve names using the upstream resolver, as the container itself is unable to communicate with that nameserver. Only the names of containers also attached to the internal network are able to be resolved. Many systems run a local forwarding DNS resolver. As the host and any containers have separate loopback devices, a consequence of the design described above is that containers are unable to resolve names from the host's configured resolver, as they cannot reach these addresses on the host loopback device. To bridge this gap, and to allow containers to properly resolve names even when a local forwarding resolver is used on a loopback address, `dockerd` detects this scenario and instead forward DNS requests from the host namework namespace. The loopback resolver then forwards the requests to its configured upstream resolvers, as expected. Because `dockerd` forwards DNS requests to the host loopback device, bypassing the container network namespace's normal routing semantics entirely, internal networks can unexpectedly forward DNS requests to an external nameserver. By registering a domain for which they control the authoritative nameservers, an attacker could arrange for a compromised container to exfiltrate data by encoding it in DNS queries that will eventually be answered by their nameservers. Docker Desktop is not affected, as Docker Desktop always runs an internal resolver on a RFC 1918 address. Moby releases 26.0.0, 25.0.4, and 23.0.11 are patched to prevent forwarding any DNS requests from internal networks. As a workaround, run containers intended to be solely attached to internal networks with a custom upstream address, which will force all upstream DNS queries to be resolved from the container's network namespace.

A flaw was found in Buildah (and subsequently Podman Build) which allows containers to mount arbitrary locations on the host filesystem into build containers. A malicious Containerfile can use a dummy image with a symbolic link to the root filesystem as a mount source and cause the mount operation to mount the host root filesystem inside the RUN step. The commands inside the RUN step will then have read-write access to the host filesystem, allowing for full container escape at build time.

VMware ESXi contains an out-of-bounds write vulnerability. A malicious actor with privileges within the VMX process may trigger an out-of-bounds write leading to an escape of the sandbox.

Moby is an open-source project created by Docker to enable software containerization. The classic builder cache system is prone to cache poisoning if the image is built FROM scratch. Also, changes to some instructions (most important being HEALTHCHECK and ONBUILD) would not cause a cache miss. An attacker with the knowledge of the Dockerfile someone is using could poison their cache by making them pull a specially crafted image that would be considered as a valid cache candidate for some build steps. 23.0+ users are only affected if they explicitly opted out of Buildkit (DOCKER_BUILDKIT=0 environment variable) or are using the /build API endpoint. All users on versions older than 23.0 could be impacted. Image build API endpoint (/build) and ImageBuild function from github.com/docker/docker/client is also affected as it the uses classic builder by default. Patches are included in 24.0.9 and 25.0.2 releases.

BuildKit is a toolkit for converting source code to build artifacts in an efficient, expressive and repeatable manner. In addition to running containers as build steps, BuildKit also provides APIs for running interactive containers based on built images. It was possible to use these APIs to ask BuildKit to run a container with elevated privileges. Normally, running such containers is only allowed if special `security.insecure` entitlement is enabled both by buildkitd configuration and allowed by the user initializing the build request. The issue has been fixed in v0.12.5 . Avoid using BuildKit frontends from untrusted sources.

BuildKit is a toolkit for converting source code to build artifacts in an efficient, expressive and repeatable manner. A malicious BuildKit frontend or Dockerfile using RUN --mount could trick the feature that removes empty files created for the mountpoints into removing a file outside the container, from the host system. The issue has been fixed in v0.12.5. Workarounds include avoiding using BuildKit frontends from an untrusted source or building an untrusted Dockerfile containing RUN --mount feature.

BuildKit is a toolkit for converting source code to build artifacts in an efficient, expressive and repeatable manner. Two malicious build steps running in parallel sharing the same cache mounts with subpaths could cause a race condition that can lead to files from the host system being accessible to the build container. The issue has been fixed in v0.12.5. Workarounds include, avoiding using BuildKit frontend from an untrusted source or building an untrusted Dockerfile containing cache mounts with --mount=type=cache,source=... options.

BuildKit is a toolkit for converting source code to build artifacts in an efficient, expressive and repeatable manner. A malicious BuildKit client or frontend could craft a request that could lead to BuildKit daemon crashing with a panic. The issue has been fixed in v0.12.5. As a workaround, avoid using BuildKit frontends from untrusted sources.

runc is a CLI tool for spawning and running containers on Linux according to the OCI specification. In runc 1.1.11 and earlier, due to an internal file descriptor leak, an attacker could cause a newly-spawned container process (from runc exec) to have a working directory in the host filesystem namespace, allowing for a container escape by giving access to the host filesystem ("attack 2"). The same attack could be used by a malicious image to allow a container process to gain access to the host filesystem through runc run ("attack 1"). Variants of attacks 1 and 2 could be also be used to overwrite semi-arbitrary host binaries, allowing for complete container escapes ("attack 3a" and "attack 3b"). runc 1.1.12 includes patches for this issue.

A flaw was found in CRI-O that involves an experimental annotation leading to a container being unconfined. This may allow a pod to specify and get any amount of memory/cpu, circumventing the kubernetes scheduler and potentially resulting in a denial of service in the node.

A security issue was discovered in Kubernetes where a user that can create pods and persistent volumes on Windows nodes may be able to escalate to admin privileges on those nodes. Kubernetes clusters are only affected if they are using an in-tree storage plugin for Windows nodes.

Windows Common Log File System Driver Elevation of Privilege Vulnerability

A security issue was discovered in Kubernetes where a user that can create pods on Windows nodes running kubernetes-csi-proxy may be able to escalate to admin privileges on those nodes. Kubernetes clusters are only affected if they include Windows nodes running kubernetes-csi-proxy.

A security issue was discovered in Kubernetes where a user that can create pods on Windows nodes may be able to escalate to admin privileges on those nodes. Kubernetes clusters are only affected if they include Windows nodes.

A security issue was discovered in Kubernetes where a user that can create pods on Windows nodes may be able to escalate to admin privileges on those nodes. Kubernetes clusters are only affected if they include Windows nodes.

Code injection via nginx.ingress.kubernetes.io/permanent-redirect annotation.

Ingress nginx annotation injection causes arbitrary command execution.

This flaw makes curl overflow a heap based buffer in the SOCKS5 proxy handshake. When curl is asked to pass along the host name to the SOCKS5 proxy to allow that to resolve the address instead of it getting done by curl itself, the maximum length that host name can be is 255 bytes. If the host name is detected to be longer, curl switches to local name resolving and instead passes on the resolved address only. Due to this bug, the local variable that means "let the host resolve the name" could get the wrong value during a slow SOCKS5 handshake, and contrary to the intention, copy the too long host name to the target buffer instead of copying just the resolved address there. The target buffer being a heap based buffer, and the host name coming from the URL that curl has been told to operate with.

A vulnerability was found in cri-o. This issue allows the addition of arbitrary lines into /etc/passwd by use of a specially crafted environment variable.

Docker Desktop before 4.23.0 allows an unprivileged user to bypass Enhanced Container Isolation (ECI) restrictions via the debug shell which remains accessible for a short time window after launching Docker Desktop. The affected functionality is available for Docker Business customers only and assumes an environment where users are not granted local root or Administrator privileges. This issue has been fixed in Docker Desktop 4.23.0. Affected Docker Desktop versions: from 4.13.0 before 4.23.0.

Local privilege escalation vulnerability in Ubuntu Kernels overlayfs ovl_copy_up_meta_inode_data skip permission checks when calling ovl_do_setxattr on Ubuntu kernels

On Ubuntu kernels carrying both c914c0e27eb0 and "UBUNTU: SAUCE: overlayfs: Skip permission checking for trusted.overlayfs.* xattrs", an unprivileged user may set privileged extended attributes on the mounted files, leading them to be set on the upper files without the appropriate security checks.

Docker Desktop before 4.6.0 on Windows allows attackers to delete any file through the hyperv/destroy dockerBackendV2 API via a symlink in the DataFolder parameter, a different vulnerability than CVE-2022-26659.

A Time-of-check Time-of-use (TOCTOU) flaw was found in podman. This issue may allow a malicious user to replace a normal file in a volume with a symlink while exporting the volume, allowing for access to arbitrary files on the host file system.

Docker Desktop before 4.17.0 allows an unprivileged user to bypass Enhanced Container Isolation (ECI) restrictions by setting the Docker host to docker.raw.sock, or npipe:////.pipe/docker_engine_linux on Windows, via the -H (--host) CLI flag or the DOCKER_HOST environment variable and launch containers without the additional hardening features provided by ECI. This would not affect already running containers, nor containers launched through the usual approach (without Docker's raw socket). The affected functionality is available for Docker Business customers only and assumes an environment where users are not granted local root or Administrator privileges. This issue has been fixed in Docker Desktop 4.17.0. Affected Docker Desktop versions: from 4.13.0 before 4.17.0.

runc through 1.1.4 has Incorrect Access Control leading to Escalation of Privileges, related to libcontainer/rootfs_linux.go. To exploit this, an attacker must be able to spawn two containers with custom volume-mount configurations, and be able to run custom images. NOTE: this issue exists because of a CVE-2019-19921 regression.

Due to a vulnerability in the io_uring subsystem, it is possible to leak kernel memory information to the user process. timens_install calls current_is_single_threaded to determine if the current process is single-threaded, but this call does not consider io_uring's io_worker threads, thus it is possible to insert a time namespace's vvar page to process's memory space via a page fault. When this time namespace is destroyed, the vvar page is also freed, but not removed from the process' memory, and a next page allocated by the kernel will be still available from the user-space process and can leak memory contents via this (read-only) use-after-free vulnerability. We recommend upgrading past version 5.10.161 or commit  788d0824269bef539fe31a785b1517882eafed93 https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/io_uring

containerd is an open source container runtime. A bug was found in containerd prior to versions 1.6.18 and 1.5.18 where supplementary groups are not set up properly inside a container. If an attacker has direct access to a container and manipulates their supplementary group access, they may be able to use supplementary group access to bypass primary group restrictions in some cases, potentially gaining access to sensitive information or gaining the ability to execute code in that container. Downstream applications that use the containerd client library may be affected as well. This bug has been fixed in containerd v1.6.18 and v.1.5.18. Users should update to these versions and recreate containers to resolve this issue. Users who rely on a downstream application that uses containerd's client library should check that application for a separate advisory and instructions. As a workaround, ensure that the `"USER $USERNAME"` Dockerfile instruction is not used. Instead, set the container entrypoint to a value similar to `ENTRYPOINT ["su", "-", "user"]` to allow `su` to properly set up supplementary groups.

A use after free vulnerability exists in the ALSA PCM package in the Linux Kernel. SNDRV_CTL_IOCTL_ELEM_{READ|WRITE}32 is missing locks that can be used in a use-after-free that can result in a priviledge escalation to gain ring0 access from the system user. We recommend upgrading past commit 56b88b50565cd8b946a2d00b0c83927b7ebb055e

There is a logic error in io_uring's implementation which can be used to trigger a use-after-free vulnerability leading to privilege escalation. In the io_prep_async_work function the assumption that the last io_grab_identity call cannot return false is not true, and in this case the function will use the init_cred or the previous linked requests identity to do operations instead of using the current identity. This can lead to reference counting issues causing use-after-free. We recommend upgrading past version 5.10.161.

An issue was discovered in ksmbd in the Linux kernel 5.15 through 5.19 before 5.19.2. There is a heap-based buffer overflow in set_ntacl_dacl, related to use of SMB2_QUERY_INFO_HE after a malformed SMB2_SET_INFO_HE command.

An issue was discovered in ksmbd in the Linux kernel 5.15 through 5.19 before 5.19.2. fs/ksmbd/smb2pdu.c omits a kfree call in certain smb2_handle_negotiate error conditions, aka a memory leak.

Apache Commons Text performs variable interpolation, allowing properties to be dynamically evaluated and expanded. The standard format for interpolation is "${prefix:name}", where "prefix" is used to locate an instance of org.apache.commons.text.lookup.StringLookup that performs the interpolation. Starting with version 1.5 and continuing through 1.9, the set of default Lookup instances included interpolators that could result in arbitrary code execution or contact with remote servers. These lookups are: - "script" - execute expressions using the JVM script execution engine (javax.script) - "dns" - resolve dns records - "url" - load values from urls, including from remote servers Applications using the interpolation defaults in the affected versions may be vulnerable to remote code execution or unintentional contact with remote servers if untrusted configuration values are used. Users are recommended to upgrade to Apache Commons Text 1.10.0, which disables the problematic interpolators by default.

Incorrect handling of the supplementary groups in the CRI-O container engine might lead to sensitive information disclosure or possible data modification if an attacker has direct access to the affected container where supplementary groups are used to set access permissions and is able to execute a binary code in that container.

Onedev is an open source, self-hosted Git Server with CI/CD and Kanban. When using Docker-based job executors, the Docker socket (e.g. /var/run/docker.sock on Linux) is mounted into each Docker step. Users that can define and trigger CI/CD jobs on a project could use this to control the Docker daemon on the host machine. This is a known dangerous pattern, as it can be used to break out of Docker containers and, in most cases, gain root privileges on the host system. This issue allows regular (non-admin) users to potentially take over the build infrastructure of a OneDev instance. Attackers need to have an account (or be able to register one) and need permission to create a project. Since code.onedev.io has the right preconditions for this to be exploited by remote attackers, it could have been used to hijack builds of OneDev itself, e.g. by injecting malware into the docker images that are built and pushed to Docker Hub. The impact is increased by this as described before. Users are advised to upgrade to 7.3.0 or higher. There are no known workarounds for this issue.

An incorrect handling of the supplementary groups in the Podman container engine might lead to the sensitive information disclosure or possible data modification if an attacker has direct access to the affected container where supplementary groups are used to set access permissions and is able to execute a binary code in that container.

containerd is an open source container runtime. A bug was found in the containerd's CRI implementation where programs inside a container can cause the containerd daemon to consume memory without bound during invocation of the `ExecSync` API. This can cause containerd to consume all available memory on the computer, denying service to other legitimate workloads. Kubernetes and crictl can both be configured to use containerd's CRI implementation; `ExecSync` may be used when running probes or when executing processes via an "exec" facility. This bug has been fixed in containerd 1.6.6 and 1.5.13. Users should update to these versions to resolve the issue. Users unable to upgrade should ensure that only trusted images and commands are used.

A vulnerability was found in CRI-O that causes memory or disk space exhaustion on the node for anyone with access to the Kube API. The ExecSync request runs commands in a container and logs the output of the command. This output is then read by CRI-O after command execution, and it is read in a manner where the entire file corresponding to the output of the command is read in. Thus, if the output of the command is large it is possible to exhaust the memory or the disk space of the node when CRI-O reads the output of the command. The highest threat from this vulnerability is system availability.

A flaw was found in cri-o, where containers were incorrectly started with non-empty default permissions. A vulnerability was found in Moby (Docker Engine) where containers started incorrectly with non-empty inheritable Linux process capabilities. This flaw allows an attacker with access to programs with inheritable file capabilities to elevate those capabilities to the permitted set when execve(2) runs.