| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| Pillow is a Python imaging library. From 8.1.0 until 13.0.0, Pillow can allocate an APNG disposal background at the full canvas size before applying the decompression-bomb check. A crafted APNG with a small first frame, a very large declared canvas, and background disposal or first-frame OP_PREVIOUS reaches PngImageFile._seek while Image.open() positions frame zero. The full-canvas allocation occurs inside Image.open() without requiring pixel loading or frame iteration. This can exhaust memory and deny service. This issue is fixed in version 13.0.0. |
| russh is a Rust SSH client and server library that includes an SSH agent protocol implementation. Prior to 0.64.1, the agent server spawns a separate long-lived task for every SSH_AGENT_CONSTRAIN_LIFETIME record in a constrained add-key request. The loop accepts repeated lifetime records, clones the key blob into each task, and does not roll back tasks when later constraint parsing fails. An attacker that can speak the agent protocol locally or through agent forwarding can exhaust memory and tasks until the agent process is starved or terminated, denying SSH agent service. This issue is fixed in version 0.64.1. |
| Uncontrolled recursion in the SVG parser and renderer of the Qt SVG module (QtSvg) in Qt Group Qt allows attackers to cause a denial of service (application crash via stack exhaustion) via a crafted SVG file with deeply nested elements. |
| AIOHTTP is an asynchronous HTTP client/server framework for asyncio and Python. Prior to 3.14.4, keepalive timeout handling is absent during the initial portion of a request. An attacker can keep an idle connection open indefinitely before sending any request data, consuming server connection capacity. This issue is fixed in version 3.14.4. |
| Cpanel::JSON::XS versions before 4.80 for Perl allow memory exhaustion by buffering incremental input past max_size in incr_parse.
The parser appends each chunk to its buffer. In scalar or list context, it compares max_size with how far it has progressed, not with the buffer length. A chunk that starts with a NUL byte stops the parser before it progresses, and a call in void context skips the check, so the buffer grows without limit.
A peer streaming data to an incr_parse consumer can exhaust process memory regardless of max_size. |
| undici is an HTTP client library for Node.js. In versions 6.5.0 through 6.29.0, 7.0.0 through 7.30.0, and 8.0.0 through 8.12.1, its EventSource client can be driven to exhaust process memory by a malicious or compromised Server-Sent Events peer. By sending millions of tiny completed data fields with no blank event separator, or a single unfinished line split across many chunks, the parser retains a growing amount of per-field and per-chunk state that the byte-based maxEventSize and maxResponseSize limits do not bound, and because maxResponseSize is enforced on the compressed body before decompression, a small compressed response can expand into enough retained memory to terminate the process. An application only needs to maintain an EventSource connection to an untrusted peer, which controls only the response, and availability is the only impact. The issue is fixed in undici 6.29.1, 7.30.1, and 8.12.2, which bound the retained parser state independently of the byte limits, and users should upgrade to one of those versions or later. There is no workaround. |
| undici is an HTTP client library for Node.js. In versions 5.15.0 through 6.29.0, 7.0.0 through 7.30.0, and 8.0.0 through 8.12.1, its WebSocket client, including the bundled WebSocket exposed by Node.js, can be driven to exhaust process memory by a malicious or compromised WebSocket server. For every PING frame received, the client automatically writes a PONG response and ignores the backpressure signal from the socket write, so a server that streams small PING frames while never reading the client's PONGs makes the outbound write queue and heap grow without bound until the Node.js process runs out of memory and terminates. The attack is remote and unauthenticated, needs only a normal WebSocket connection, and is asymmetric, since small inbound PINGs produce a much larger retained outbound queue, with availability the only impact. The issue is fixed in undici 6.29.1, 7.30.1, and 8.12.2, which pause WebSocket frame parsing when the automatic PONG response hits socket backpressure and resume on drain, bounding the outbound queue, and users should upgrade to one of those versions or later. There is no workaround. |
| RabbitMQ is a messaging and streaming broker. Prior to versions 4.3.0, 4.2.6, 4.1.11, 4.0.20, and 3.13.15, The content-header BodySize (a uint64) was stored without validation against max_message_size. The size check ran only when assembly completed. By declaring body_size = 2^63-1 and then streaming fragments, a client ensured that check_msg_size never fired, so the accumulated body size went unbounded. A reader process accumulates memory until the memory alarm fires, degrading all publishers cluster-wide, or until the node runs out of memory. The memory alarm provides only partial mitigation, since it is reactive rather than preventive. AMQP 0-9-1 is the most widely used protocol, and any publisher can trigger this condition. Preconditions include Any authenticated AMQP 0-9-1 client with publish permission can exploit this.. This issue is fixed in versions 4.3.0, 4.2.6, 4.1.11, 4.0.20, and 3.13.15. |
| RabbitMQ is a messaging and streaming broker. Prior to versions 3.13.15, 4.0.20, 4.1.11, 4.2.6, and 4.3.0, add_vhost/2 calls rabbit_data_coercion:atomize_keys/1 (the unsafe variant using binary_to_atom) on the vhost metadata map. The 20 MB management body limit fits ~1M+ short keys. Admin-only. An administrator importing a crafted definitions file can crash the node in a single request: a vhosts entry with ~1M unique metadata keys exhausts the atom table during import. Preconditions include administrator tag. This issue is fixed in versions 3.13.15, 4.0.20, 4.1.11, 4.2.6, and 4.3.0. |
| RabbitMQ is a messaging and streaming broker. Prior to versions 3.13.15, 4.0.20, 4.1.11, 4.2.6, 4.3.0, When a binding is created on an x-jms-topic exchange, add_binding/3 reads the rjms_erlang_selector argument and passes it through erl_scan:string/1 then erl_parse:parse_term/1. erl_scan:string/1 interns every atom literal it tokenizes. validate_binding/2 is a no-op (-> ok.), there is no length cap, and the surrounding try/catch cannot reclaim atoms. The Java JMS client compiles selectors client-side, but the server does not enforce this , a raw AMQP client can send arbitrary selector strings. An authenticated low-privilege AMQP user confined to one vhost can crash the entire broker node (cross-tenant DoS) in <100 bind calls. Preconditions include rabbitmq_jms_topic_exchange plugin enabled (bundled; required for any JMS deployment) Authenticated AMQP user with read on an x-jms-topic exchange + write on a queue (or configure to declare both). This issue is fixed in versions 3.13.15, 4.0.20, 4.1.11, 4.2.6, 4.3.0. |
| RabbitMQ is a messaging and streaming broker. Prior to versions 3.13.15, 4.0.20, 4.1.11, 4.2.6, and 4.3.0, add_binding/3 parses the routing key as an integer weight N and computes ring positions with lists:seq(NextN0, NextN0 + N - 1). validate_binding/2 only checks N >= 1 , no upper bound. The resulting list is stored in the exchange's Khepri record, replicated cluster-wide, and reloaded on restart. A user with write permission on a consistent-hash exchange and read on a queue can create a binding whose routing key (the hash-ring weight) is an arbitrarily large integer. The broker allocates a list of that many integers via lists:seq/2 and persists it to Khepri across all cluster nodes , a single binding with weight 100000000 allocates ~800 MB on every node and survives restarts. Preconditions include rabbitmq_consistent_hash_exchange plugin enabled write permission on a consistent-hash exchange + read on a queue (standard binding perms). This issue is fixed in versions 3.13.15, 4.0.20, 4.1.11, 4.2.6, and 4.3.0. |
| RabbitMQ is a messaging and streaming broker. Prior to versions 4.1.11, 4.2.6, and 4.3.0, validate_partitions only checks that the requested partition count is at least 1, with no upper bound. A large count such as lists:seq(0, 500000000) allocates roughly 8GB. Preconditions include The rabbitmq_stream_management plugin must be enabled. The caller needs the management tag and access to the target vhost.. This issue is fixed in versions 4.1.11, 4.2.6, and 4.3.0. |
| RabbitMQ is a messaging and streaming broker. Prior to versions 3.13.15, 4.0.20, 4.1.11, 4.2.6, and 4.3.1, get_chunk_selector/1 calls binary_to_atom on the raw client-supplied <<"chunk_selector">> property from post-auth subscribe and resolve_offset_spec frames, with no whitelist and no existing guard. An authenticated stream client with read access to any stream can crash the broker node. Preconditions include rabbitmq_stream plugin enabled Authenticated stream-protocol user with read access to at least one stream. This issue is fixed in versions 3.13.15, 4.0.20, 4.1.11, 4.2.6, and 4.3.1. |
| RabbitMQ is a messaging and streaming broker. Prior to versions 3.13.15, 4.0.20, 4.1.11, and 4.2.6, parse_array_primitive/2 for constructor 0x45 (list0) returns an element with byte-width B = 0. The enclosing array32 parser at line 148 reads a 4-byte Count from the wire and loops Count times consuming B bytes each , with B = 0, no input is consumed and the loop builds a list of Count empty elements bounded only by the 32-bit field. The SASL-mechanisms / SASL-init frame is parsed by amqp10_framing:decode_bin/1 from rabbit_amqp_reader.erl:412 before authentication completes. The pre-auth incoming_max_frame_size (default 8192 bytes) caps the frame, not the Count field, so a 19-byte payload with Count = 0xFFFFFFFF is accepted. No max_heap_size is set on the reader process. An unauthenticated network attacker can crash any RabbitMQ node that has the AMQP 1.0 listener enabled (default port 5672) by sending a single ~19-byte frame. The reader process attempts to build a list of ~4 billion empty elements, exhausting heap memory and terminating the Erlang VM. All tenants and protocols on the node lose service. Preconditions include Network reachability to the AMQP listener (port 5672, enabled by default) No authentication required. This issue is fixed in versions 3.13.15, 4.0.20, 4.1.11, and 4.2.6. |
| In the Linux kernel, the following vulnerability has been resolved:
wifi: nl80211: reject oversized EMA RNR lists
nl80211_parse_rnr_elems() stores the parsed element count in a
u8-backed cfg80211_rnr_elems::cnt field and uses that count to size
the flexible array allocation.
Reject nested NL80211_ATTR_EMA_RNR_ELEMS input once the count reaches
255, before incrementing it again. This keeps the parser aligned with
the data structure it fills and matches the existing bound check used
by nl80211_parse_mbssid_elems(). |
| In the Linux kernel, the following vulnerability has been resolved:
vsock/virtio: fix potential unbounded skb queue
virtio_transport_inc_rx_pkt() checks vvs->rx_bytes + len > vvs->buf_alloc.
virtio_transport_recv_enqueue() skips coalescing for packets
with VIRTIO_VSOCK_SEQ_EOM.
If fed with packets with len == 0 and VIRTIO_VSOCK_SEQ_EOM,
a very large number of packets can be queued
because vvs->rx_bytes stays at 0.
Fix this by estimating the skb metadata size:
(Number of skbs in the queue) * SKB_TRUESIZE(0) |
| In the Linux kernel, the following vulnerability has been resolved:
netfilter: nf_conncount: increase the connection clean up limit to 64
After the optimization to only perform one GC per jiffy, a new problem
was introduced. If more than 8 new connections are tracked per jiffy the
list won't be cleaned up fast enough possibly reaching the limit
wrongly.
In order to prevent this issue, only skip the GC if it was already
triggered during the same jiffy and the increment is lower than the
clean up limit. In addition, increase the clean up limit to 64
connections to avoid triggering GC too often and do more effective GCs.
This has been tested using a HTTP server and several
performance tools while having nft_connlimit/xt_connlimit or OVS limit
configured.
Output of slowhttptest + OVS limit at 52000 connections:
slow HTTP test status on 340th second:
initializing: 0
pending: 432
connected: 51998
error: 0
closed: 0
service available: YES |
| In the Linux kernel, the following vulnerability has been resolved:
crypto: pcrypt - Fix handling of MAY_BACKLOG requests
MAY_BACKLOG requests can return EBUSY. Handle them by checking
for that value and filtering out EINPROGRESS notifications. |
| In the Linux kernel, the following vulnerability has been resolved:
netfilter: flowtable: strictly check for maximum number of actions
The maximum number of flowtable hardware offload actions in IPv6 is:
* ethernet mangling (4 payload actions, 2 for each ethernet address)
* SNAT (4 payload actions)
* DNAT (4 payload actions)
* Double VLAN (4 vlan actions, 2 for popping vlan, and 2 for pushing)
for QinQ.
* Redirect (1 action)
Which makes 17, while the maximum is 16. But act_ct supports for tunnels
actions too. Note that payload action operates at 32-bit word level, so
mangling an IPv6 address takes 4 payload actions.
Update flow_action_entry_next() calls to check for the maximum number of
supported actions.
While at it, rise the maximum number of actions per flow from 16 to 24
so this works fine with IPv6 setups. |
| SGLang through 0.5.22 with the Mooncake disaggregation backend accepts unauthenticated decode registration messages on the Prefill rank ZMQ PULL socket, allowing denial of service. Network-adjacent attackers can send many registrations with unique mooncake_session_id values and large nested-buffer frames to exhaust Prefill memory until the scheduler is OOM-killed. |
