| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| vLLM is an inference and serving engine for large language models. Prior to 0.30.0, a caller can use the request-level media_io_kwargs field to select the GLMGA video backend and supply large values for the fps and max_frames options without a strict work ceiling. GLMGA constructs and deduplicates an attacker-sized pre-decode frame-index list, allowing a compact request and tiny valid video to consume disproportionate CPU time and memory in the shared media-loading executor. This issue is fixed in version 0.30.0. |
| In the Linux kernel, the following vulnerability has been resolved:
IB/mlx5: Fix transport-domain rollback and initialize lb mutex earlier
mlx5_ib_alloc_transport_domain() allocates a transport domain and then
may fail in mlx5_ib_enable_lb(). In that case, the allocated TD is leaked.
Fix this by deallocating the TD when mlx5_ib_enable_lb() returns an
error. Also return 0 explicitly in the no-loopback-capability success
branch, and move dev->lb.mutex initialization to mlx5_ib_stage_init_init(). |
| 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. |
| Inefficient algorithmic complexity in .NET allows an unauthorized attacker to deny service over a network. |
| Asymmetric resource consumption (amplification) in .NET allows an unauthorized attacker to deny service over a network. |
| In the Linux kernel, the following vulnerability has been resolved:
nvmet-tcp: check return value of nvmet_tcp_set_queue_sock
The return value of nvmet_tcp_set_queue_sock() is currently ignored in
nvmet_tcp_tls_handshake_done(). If it fails (e.g., due to the socket
not being in TCP_ESTABLISHED state), the socket callbacks will not be
properly set, leading to queue and socket leakage.
Fix this by capturing the return value and calling
nvmet_tcp_schedule_release_queue() on failure to ensure proper cleanup. |
| In the Linux kernel, the following vulnerability has been resolved:
wifi: mt76: mt7996: Fix possible token leak in mt7996_tx_prepare_skb()
If link_conf or link_sta lookup fails in mt7996_tx_prepare_skb routine,
mt7996 driver leaks an already allocated tx token. Fix the issue
releasing the token in case of error. |
| Improper Handling of Highly Compressed Data (Data Amplification) vulnerability in hexpm hex allows an untrusted package mirror to exhaust the memory or disk space of the machine running mix deps.get, mix deps.update, mix hex.package fetch or mix hex.docs fetch.
Hex treats a mirror configured with HEX_MIRROR_URL, HEX_MIRROR or mirror_url as untrusted: it doesn't send credentials to it, and it checks registry signatures and package checksums that the mirror can't forge. Several response paths consume memory or disk in proportion to what the mirror sends before any of those checks run.
Hex.HTTP.request/5 reads the whole response body into memory and decompresses it with :zlib.gunzip/1 whenever the response's Content-Encoding contains gzip, with no limit on the compressed or decompressed size. Hex doesn't ask for compressed responses, so the server decides whether this step runs. It applies to registry, package, docs and API responses.
Package tarballs have no download size limit. Hex 2.5.1 holds the whole tarball in memory and writes it to the cache, and the development branch streams it to a temporary file; the 16 MiB tarball limit is checked only when the file is unpacked. The tarball is then unpacked, into deps/<app> in 2.5.1 and into a staging directory on the development branch, before its checksums are compared with the registry.
mix hex.docs fetch, and mix hex.docs offline when docs are missing, write the docs tarball to disk and extract it with :mix_hex_erl_tar.extract/2 without the 16 MiB compressed and 128 MiB decompressed limits that mix_hex_tarball:unpack_docs/3 applies. The signed registry doesn't cover docs tarballs, so nothing checks them.
Registry responses are also decompressed before their signature is checked, in hex_core code that Hex vendors; that part is tracked in https://github.com/hexpm/hex_core/security/advisories/GHSA-7c4r-cgf9-9jmq.
This issue affects hex: from 0.11.2 before 2.5.2. |
| IBM WebSphere Application Server - Liberty 20.0.0.12 through 26.0.0.10 could allow a remote attacker to cause a denial of service due to an uncaught exception. |
| An issue in gnutls v3.8.13 causes legitimate CA certificates to be rejected, leading to a Denial of Service (DoS). |
| Improper certificate validation in gnutls v3.8.13 causes the application to reject legitimate certificates for valid users, leading to a Denial of Service (DoS). |
| The Link Layer Control Procedure (LLCP) implementation of the Zephyr software Bluetooth LE Controller retains the receive node that carried an accepted LL_PHY_UPDATE_IND so that it can later be reused for the host notification when the update instant is reached (llcp_rx_node_retain() in subsys/bluetooth/controller/ll_sw/ull_llcp.c, and the node is deliberately not recycled while marked NODE_RX_TYPE_RETAIN). The invalid-PDU arms of llcp_lp_pu_rx() and llcp_rp_pu_rx() in subsys/bluetooth/controller/ll_sw/ull_llcp_phy.c completed the procedure via llcp_lr_complete() / llcp_rr_complete() without first releasing that retained node, so the procedure context — the only remaining reference to the node — was freed while the node was still held out of the receive pool.
A peer device on an established LE connection can drive this deterministically and without pairing or encryption. Against a peripheral it sends LL_PHY_REQ, receives LL_PHY_RSP, sends a valid LL_PHY_UPDATE_IND with an instant a few connection events in the future (so the node becomes retained), and then, before the instant is reached, sends any other LL Control PDU such as LL_LENGTH_REQ; ull_cp_rx() routes it to the active remote PHY Update procedure, which takes the invalid-PDU path. The mirror case applies to a locally initiated PHY Update followed by an LL_REJECT_IND.
In the default configuration (CONFIG_BT_ASSERT and CONFIG_BT_CTLR_ASSERT_DEBUG both default y) the violated invariant in llcp_lr_check_done() / llcp_rr_check_done() triggers a controller assertion, ending in k_oops() (or k_panic()) — a single crafted PDU sequence from radio range faults the device. With those assertions compiled out, each attempt silently leaks one receive PDU node and its memq link; because the controller receive pool is small (PDU_RX_CNT, driven by CONFIG_BT_CTLR_RX_BUFFERS, which defaults to 1) and each attempt costs the attacker only a reconnect, a few repetitions exhaust the pool and leave Bluetooth inoperable until reboot. On releases v3.4.0 through v3.7.x the assertion is never reached, whatever the configuration, so every attempt leaks silently.
The impact is limited to availability: the orphaned node leaves no dangling pointer that is later dereferenced and is never delivered to the host, so there is no memory corruption or information disclosure. The same pull request applies the identical release to the Connection Update and CIS-create procedures, whose invalid-PDU arms had the same omission. |
| FreeRDP from 2.0.0 before 3.32.0 contains an unbounded reassembly vulnerability in WTSProcessChannelData that fails to enforce cumulative fragment limits. Authenticated clients can start a static virtual channel message, withhold the LAST flag, and send endless fragments to exhaust server memory. |
| FreeRDP from 3.0.0-beta1 before 3.32.0 fails to enforce cumulative or absolute size limits when reassembling RDPDR channel fragments in the proxy. Attackers can send a FIRST fragment followed by continuation fragments without LAST, growing per-session heap streams proportional to traffic and exhausting proxy memory. |
| In Apache CXF, the parser for multipart/MTOM attachment part headers did not fully enforce the configured attachment-max-header-size (default 300 characters) and attachment-headers-max-count (default 500) limits. The size limit was applied only to each physical line, not to a header value built from continuation lines or to the combined values of a repeated header. The count limit was checked against the number of distinct header names, not the total number of header lines. A remote, unauthenticated attacker could send a multipart request with very large folded or repeated part headers. The server would then allocate memory without bound, causing a denial of service.
Users are recommended to upgrade to versions 4.2.4 or 4.1.9 or 3.6.13, which fix this issue. |
| Missing release of memory after effective lifetime in Remote Desktop Gateway Service allows an unauthorized attacker to deny service over a network. |
| FreeRDP from 3.0.0 before 3.32.0 contains an uncontrolled resource consumption vulnerability in parse_nla_mode_pdu that allows unauthenticated attackers to exhaust server memory. Attackers can complete TLS and stream bytes not beginning with the DER SEQUENCE tag 0x30, causing unbounded receive buffer growth before NLA authentication. |
| In the Linux kernel, the following vulnerability has been resolved:
usb: typec: tcpci_rt1711h: unregister TCPCI port with devres
rt1711h_probe() registers the TCPCI port before requesting the interrupt
and enabling alert interrupts. If either of those later steps fails, the
probe function returns without unregistering the TCPCI port. The explicit
unregister currently only happens from the remove callback.
Register a devres action immediately after tcpci_register_port() succeeds,
so tcpci_unregister_port() runs on later probe failures and on driver
detach. Drop the remove callback to avoid unregistering the same port
twice.
This issue was identified during our ongoing static-analysis research while
reviewing kernel code. |
| In the Linux kernel, the following vulnerability has been resolved:
tcp: fix stale per-CPU tcp_tw_isn leak enabling ISN prediction
Blamed commit moved the TIME_WAIT-derived ISN from the skb control
block to a per-CPU variable, assuming the value would always be consumed
by tcp_conn_request() for the same packet that wrote it. That assumption
is violated by multiple drop paths between the producer
(__this_cpu_write(tcp_tw_isn, isn) in tcp_v{4,6}_rcv()) and the consumer
(tcp_conn_request()):
- min_ttl / min_hopcount check
- xfrm policy check
- tcp_inbound_hash() MD5/AO mismatch
- tcp_filter() eBPF/SO_ATTACH_FILTER drop
- th->syn && th->fin discard in tcp_rcv_state_process() TCP_LISTEN
- psp_sk_rx_policy_check() in tcp_v{4,6}_do_rcv()
- tcp_checksum_complete() in tcp_v{4,6}_do_rcv()
- tcp_v{4,6}_cookie_check() returning NULL
When a packet is dropped on any of these paths, tcp_tw_isn is left set.
The next SYN processed on the same CPU then consumes the non zero value in
tcp_conn_request(), receiving a potentially predictable ISN.
This patch moves back tcp_tw_isn to skb->cb[], getting rid of the per-cpu
variable.
Note that tcp_v{4,6}_fill_cb() do not set it.
Very litle impact on overall code size/complexity:
$ scripts/bloat-o-meter -t vmlinux.old vmlinux.new
add/remove: 0/0 grow/shrink: 2/1 up/down: 8/-15 (-7)
Function old new delta
tcp_v6_rcv 3038 3042 +4
tcp_v4_rcv 3035 3039 +4
tcp_conn_request 2938 2923 -15
Total: Before=24436060, After=24436053, chg -0.00% |
| In the Linux kernel, the following vulnerability has been resolved:
nfsd: avoid leaking pre-allocated openowner on unconfirmed retry race
When find_or_alloc_open_stateowner() encounters an unconfirmed owner, it
calls release_openowner() and sets oo = NULL. Control then falls through
past the `if (oo)` guard -- which would have freed any pre-allocated
`new` -- and unconditionally executes `new = alloc_stateowner(...)`. If
`new` was already allocated on a prior iteration, the pointer is
silently overwritten and the previous allocation (slab object + owner
name buffer) is leaked.
This requires a race: two NFSv4.0 OPEN threads with the same owner
string, where a concurrent thread inserts a new unconfirmed owner into
the hash between retry iterations. The window is narrow but repeatable
under adversarial conditions.
Fix by adding `goto retry` after `oo = NULL` so the already-allocated
`new` is reused on the next iteration rather than overwritten. |
