SamsungPassAutofill prior to version 5.2.10.x on Samsung Android devices improperly exports application components, enabling a locally-installed malicious app to access sensitive autofill credential data with user interaction. The CVSS 4.0 vector (AV:L/PR:L/UI:A/VC:H) confirms the attack is constrained to on-device, low-privilege access requiring user-triggered activity, yielding a moderate score of 5.1 despite high confidentiality impact. No public exploit identified at time of analysis, and the vulnerability is not listed in CISA KEV.
Cross-site request forgery in PingFederate's Administrative Console allows remote actors to trick authenticated administrators into unknowingly executing unauthorized actions - such as altering federation trust relationships, adding rogue identity providers, or modifying SP connections - by luring them to load attacker-crafted links while an admin session is active. All PingFederate releases prior to version 13.1 are affected, and the June 2026 release of 13.1 contains the vendor-confirmed fix. Proof-of-concept exploit code is publicly available per the CVSS 4.0 E:P threat metric, and no CISA KEV listing exists at time of analysis.
Stored cross-site scripting in Discourse's moderation review queue allows a low-privileged registered user to execute arbitrary JavaScript in a moderator's browser session. All Discourse deployments running versions prior to 2026.1.6, 2026.5.2, 2026.6.1, or 2026.7.0 are affected, but only when the platform's default Content Security Policy has been modified or disabled. No public exploit identified at time of analysis and no CISA KEV listing; real-world risk is bounded by both the CSP prerequisite and the requirement for moderator interaction.
Heap buffer overflow in entr's run_utility() function allows a local low-privileged user to trigger memory corruption, process abort, and denial of service by supplying crafted command-line arguments or by using the slash-underscore substitution feature, which expands a short token into a longer pathname at runtime. The flaw stems from incorrect use of strlcpy()'s return value - which reflects the total source length rather than bytes written - to advance a destination pointer, causing an unsigned size_t underflow in the remaining-buffer counter and subsequent out-of-bounds heap writes. No public exploit has been identified and this vulnerability is absent from CISA KEV; an upstream fix is available via commit 2467fe0.
Out-of-bounds write in Samsung's libcodec2secqcelpdec.so QCELP audio codec decoder library allows local attackers holding low-privilege execution on affected Samsung devices to corrupt adjacent memory regions, with limited confidentiality, integrity, and availability impact confined to the vulnerable system. The flaw is addressed in SMR Aug-2026 Release 1, Samsung's August 2026 Security Maintenance Release covering all prior unpatched firmware builds. No public exploit code or active exploitation has been identified at time of analysis, and the CVSS 4.0 score of 4.4 reflects the strictly local attack surface and bounded impact scope.
SemClipboardService on Samsung Mobile Devices prior to SMR Aug-2026 Release 1 exposes an authorization bypass that allows a locally authenticated attacker to read clipboard contents without the expected permission check. The flaw is confined to Samsung's proprietary clipboard service and does not affect stock Android builds. No public exploit has been identified at time of analysis, and no active exploitation is listed in CISA KEV.
Device identity spoofing in Samsung Smart Switch's trouble scanning mode allows adjacent network attackers to impersonate trusted devices, potentially enabling interception or manipulation of data during device transfer sessions. All Smart Switch versions prior to 3.7.72.6 are affected, as reported by Samsung Mobile. No public exploit code has been identified at time of analysis, and this CVE does not appear in the CISA KEV catalog.
Process termination via systemd-machined v259 and later enables unprivileged users with active desktop graphical sessions to kill arbitrary system processes, including those running with elevated privileges, by exploiting an improper access control gap in the `register-machine` polkit action. The affected component - systemd-machined - is typically distributed as an optional, non-default package (such as `systemd-container`), and exploitation is strictly limited to desktop graphical login sessions, ruling out remote and SSH-based attack vectors entirely. No public exploit code has been identified at time of analysis, and the vulnerability does not appear in CISA's Known Exploited Vulnerabilities catalog.
Persistent XSS in Typemill CMS 2.x enables authenticated users holding theme-configuration access to store malicious JavaScript URI payloads via Markdown link syntax, which execute in victim browsers upon link click. The vulnerability exists in ParsedownExtension.php and TwigMarkdownExtension.php, where href values in Markdown links are rendered into HTML anchor tags without filtering the javascript: URI scheme, resulting in stored payloads that persist across all page loads. Successful exploitation delivers session cookie theft, authenticated request forgery, and credential harvesting against every visitor who clicks an injected link. No public exploit code or CISA KEV listing has been identified at time of analysis.
Hard-link path traversal in GNU cpio's tar extraction mode enables an attacker-controlled archive to create hard links pointing to files outside the intended extraction directory, defeating the security boundary that `--no-absolute-filenames` is designed to enforce. All GNU cpio versions tracked under CPE cpe:2.3:a:gnu:cpio:*:*:*:*:*:*:*:* are affected when processing tar archives in copy-in mode; the upstream fix exists only as a source commit with no tagged release confirmed at time of analysis. No active exploitation has been identified (no CISA KEV listing, no public exploit code), and exploitation requires a user to actively extract an attacker-supplied archive, materially constraining opportunistic risk.
Terminal escape sequence injection and output spoofing in GNU cpio's archive-listing mode (`cpio -it`) allows an attacker who supplies a crafted archive to manipulate terminal display or forge listing entries on the victim's screen. All versions of GNU cpio are affected per CPE data (cpe:2.3:a:gnu:cpio:*:*:*:*:*:*:*:*), reported by CERT-PL. No public exploit code has been identified at time of analysis and this CVE is not listed in the CISA KEV catalog; however, exploitation requires only basic knowledge of cpio archive format and ANSI escape sequences once a victim can be induced to inspect a crafted archive.
Stack overflow in GNU cpio's make_path function (src/makepath.c) causes a denial of service when a victim extracts a crafted archive. The function uses alloca() to allocate stack memory sized by an attacker-controlled pathname embedded in the archive, allowing unbounded stack growth and a process crash with no confidentiality or integrity impact. No active exploitation or public proof-of-concept has been identified at time of analysis; the upstream fix is available as a single commit.
Out-of-bounds memory copy in QEMU's vhost inflight migration VMState handler allows a privileged attacker with control of the migration channel to crash the destination QEMU process or corrupt heap memory. The root cause is a numeric type mismatch - the inflight buffer size is stored as uint64_t but deserialized by the VMS_VBUFFER load path as signed int32_t, so a crafted value with bit 31 set wraps to a massive size_t, causing qemu_get_buffer() to read far beyond the mmap-backed inflight region. No public exploit code has been identified and the vulnerability is not listed in the CISA KEV catalog; the CVSS score of 4.4 reflects the realistic high-privilege barrier to exploitation.
Improper cryptographic signature verification in Estonian RIA's libdigidocpp core library and its consumer applications - DigiDoc4, DigiDoc for Android, and DigiDoc for iOS - enables bypassing of XAdES signature validity checks through unhandled exceptional conditions during BDOC container parsing. Affected versions span libdigidocpp 4.1.0-4.2.0, DigiDoc4 4.7.0-4.8.1, DigiDoc Android 2.7.0-2.7.1, and DigiDoc iOS 2.8.0, all of which have vendor-released patches. A proof-of-concept is indicated by the CVSS 4.0 E:P exploit maturity modifier, though no active exploitation has been confirmed and this CVE is not listed in CISA KEV.
Discourse's Chat plugin exposes private thread message content to authenticated users who lack access to the private channel. The onebox handler in plugins/chat/lib/chat/onebox_handler.rb resolves Chat::Thread using the route's thread_id before verifying whether the requesting user can access the associated channel_id, allowing a mismatched public-channel/private-thread pair to be submitted via /onebox.json. This is confirmed fixed in versions 2026.1.6, 2026.5.2, 2026.6.1, and 2026.7.0; no public exploit or CISA KEV listing has been identified at time of analysis.
TopicLink authorization bypass in Discourse exposes restricted content metadata to authenticated users who lack permission to view those resources. All Discourse versions prior to 2026.1.6, 2026.5.2, 2026.6.1, and 2026.7.0 are affected. An authenticated user can craft internal links to private messages, restricted topics, or hidden posts and recover their canonicalized slugs or titles via the composer_messages duplicate_lookup API endpoint, even though those resources are not visible to them. No public exploit code or active exploitation has been identified at time of analysis.
Tag name information disclosure in Discourse's discourse-templates plugin exposes hidden tag names to any authenticated user via the /discourse_templates endpoint. The TemplatesSerializer in the plugin bypassed Discourse's Guardian authorization layer when serializing topic tags, causing tags belonging to restricted tag groups - such as staff-only labels - to appear in API responses regardless of the requesting user's permissions. Vendor-released patches are available across four version branches; no public exploit code or active exploitation has been identified.
Authorization bypass in Rocket.Chat's channels.convertToTeam REST endpoint allows any authenticated user holding the create-team permission to convert public channels they do not own or moderate into teams, circumventing the edit-room permission check entirely. Affected are all Rocket.Chat server deployments prior to versions 7.10.14, 8.0.8, 8.1.7, 8.2.7, 8.3.7, 8.4.5, 8.5.2, and 8.6.1. No active exploitation has been confirmed and this vulnerability is not in CISA KEV, but the low exploitation complexity combined with network accessibility warrants timely patching in multi-tenant or community-facing deployments.
HCL BigFix Mobile leaks sensitive application internals through verbose error messages and improper exception handling, enabling authenticated network users to harvest stack traces, internal paths, or configuration details that materially assist reconnaissance. Confirmed by HCL under CVSS 4.3 (AV:N/AC:L/PR:L), the flaw requires only a low-privileged account to exploit and produces no integrity or availability impact - purely a confidentiality concern. No public exploit code or active exploitation (CISA KEV) has been identified at time of analysis.
Missing authorization in ERPNext's Process Statement of Accounts module allows any authenticated low-privilege user to invoke the send_emails function and dispatch automated account statement emails to customers without holding the required role. Versions prior to 15.111.0 (v15 branch) and 16.22.0 (v16 branch) are affected across all ERPNext deployments using this accounts feature. No public exploit code has been identified and the vulnerability is not listed in the CISA KEV catalog; the vendor has released confirmed patches for both active version branches.
HashiCorp Vault's ACL policy engine inconsistently enforces wildcard (glob) deny rules against LIST requests when a trailing slash is appended to the denied path, enabling an authenticated token holder to enumerate secret entry names beneath paths it was intended to be denied access to. Affected versions span multiple Vault Enterprise release trains and Vault Community Edition prior to 2.0.3, with confirmed fixes released across all active branches. No public exploit code or active exploitation has been identified at time of analysis; the CVSS score of 4.3 (Medium) reflects the constrained impact - path enumeration only - and the authenticated access prerequisite.
CyberChef's client-side pretty-recipe parser freezes victim browser tabs via a crafted URL fragment containing large numbers of unmatched quote characters, triggering catastrophic backtracking in a synchronous global regular expression. Any user of CyberChef prior to 11.3.0 who opens a malicious #recipe= URL - whether shared in phishing, chat, or embedded in a page - is subject to a temporary denial-of-service condition. No code execution, data exfiltration, or privilege escalation is possible; impact is confined to browser tab availability. No public exploit identified at time of analysis, and this vulnerability is not listed in CISA KEV.
Unauthenticated access to Apache Ranger Download APIs on versions 2.8.0 and earlier exposes sensitive authorization policy data to any remote attacker with network connectivity. The missing authentication check (CWE-306) means no credentials or privileges are required to invoke these endpoints and retrieve their output. No public exploit code has been identified at time of analysis, but the low complexity and zero-authentication nature of the CVSS vector (AV:N/AC:L/PR:N/UI:N) makes this trivially exploitable by any attacker who can reach the Ranger Admin interface.
Brute-force credential attacks against Apache Ranger's UnixAuth component are possible in all versions up to and including 2.8.0, because the authentication endpoint imposes no rate limiting, account lockout, or anti-automation controls (CWE-307). Any deployment that has explicitly configured UnixAuth - a configuration the vendor labels as not recommended for production - exposes its Ranger admin interface to automated password-guessing from unauthenticated network attackers. No active exploitation is recorded in CISA KEV and no public POC has been identified at time of analysis, but the unauthenticated network vector (AV:N/PR:N/AC:L) makes automated credential attacks trivially executable against any network-reachable instance.
NULL pointer dereference in the shim application's dp.c library exposes Red Hat Enterprise Linux 7, 8, and 9 systems using UEFI Secure Boot to a local denial-of-service condition. An attacker with high local privileges and the ability to interact with the shim boot path can trigger the missing NULL pointer check, potentially disrupting the boot process. No public exploit code has been identified at time of analysis, and the low CVSS score of 3.9 reflects the constrained attack prerequisites, though the changed scope (S:C) signals that the impact crosses from the shim component into the broader boot environment.
Uninitialized-memory and out-of-bounds read in the UpdateHub OTA client of Zephyr RTOS allows a malicious or on-path UpdateHub server to crash the device update thread via a crafted probe response, producing a network-triggerable denial of service. The flaw exists in z_impl_updatehub_probe() within subsys/mgmt/updatehub/updatehub.c, where metadata_copy is allocated with k_malloc (leaving heap unzeroed) and filled via memcpy without a NUL terminator; a subsequent strlen() then scans into uninitialized heap, producing an over-long length fed to json_obj_parse(). No public exploit has been identified at time of analysis and the vulnerability is not listed in the CISA KEV catalog, but it is remotely triggerable against Zephyr devices running UpdateHub without DTLS transport protection.
Socket descriptor exhaustion in Zephyr RTOS's UpdateHub OTA client allows a network on-path attacker - or simply an unreachable update server - to gradually deplete the device's shared net_context pool, causing device-wide networking degradation until reboot. The defect is a resource-management bug (CWE-772) in start_coap_client(): ret was set to -1 immediately after socket creation, making the ret > 0 cleanup gate permanently unreachable, so cleanup_connection() was never invoked on any DTLS or connect failure path. No public exploit exists and the upstream fix is confirmed via Zephyr Project commit and security advisory.
Concurrent access to an unprotected singleton context structure in Zephyr RTOS's UpdateHub OTA subsystem enables two independent callers - the background autohandler and user-triggered update operations - to race through prepare_fds(), producing an out-of-bounds write that corrupts adjacent struct members and disables the firmware-update path. Affected are embedded systems running Zephyr with the UpdateHub subsystem enabled, particularly those built with CONFIG_USERSPACE where unprivileged userspace threads can invoke update syscalls and actively compete with the workqueue handler. No public exploit has been identified at time of analysis, the CVSS base score is 2.5, and the vendor commit confirms no demonstrated path to code execution - impact is limited to availability of the OTA update mechanism.
Physical access to the TBEA TLogger V2.1.0.0B0.0.0.0 (Communication Box 3rd Generation) PCB exposes an unprotected UART debug interface, allowing a physically proximate attacker to intercept the device boot process and runtime debug output without any authentication. Disclosed data - including OS version strings, software versions, network configuration, filesystem paths, and implementation internals - provides a reconnaissance foundation that can materially assist subsequent compromise of the device or connected infrastructure. No public exploit code and no active exploitation (CISA KEV) have been identified at time of analysis.
Super-linear regular-expression complexity in CPython's csv.Sniffer.sniff() allows CPU exhaustion when applications pass attacker-controlled, unbounded CSV input to dialect detection. Any Python application that auto-detects CSV formats from user-supplied data without input length limits is potentially affected across all CPython versions predating the 3.16 fix. No public exploit is identified at time of analysis, and the CVSS 4.0 score of 2.4 reflects a low-severity, local-context vulnerability with limited availability impact.
Man-in-the-middle interception of ECOVACS PRO app traffic is possible due to improper TLS server certificate validation (CWE-295) in both the Android and iOS versions of the application, published by ECOVACS Robotics and coordinated by JPCERT under JVNVU92804348. An attacker with a network-positioned MITM capability can present a rogue certificate that the app accepts without proper validation, enabling retrieval and alteration of all app-server communications, which may include device control commands, account credentials, and home telemetry. No active exploitation is confirmed (not in CISA KEV), and no public exploit code has been identified at time of analysis.
Improper TLS certificate validation in Ecovacs Robotics DEEBOT PRO M1 and K1VAC robot vacuums exposes both devices to man-in-the-middle interception and manipulation of all outbound wget communications. Because certificate verification is disabled, a network-positioned attacker can impersonate any remote server the device contacts, injecting malicious content the device will accept and act upon. Per the vendor description, successful exploitation may result in arbitrary code execution with administrative privileges on the device; no public exploit has been identified at time of analysis.
Cross-site scripting in code-projects Task Management System 1.0 allows remote, unauthenticated attackers to inject malicious JavaScript via at least six unsanitized parameters processed by /user/CommentSave.php, including comment, task_id, mineId, recId, myName, and myImage. When a victim views the affected page or comment, the injected script executes in their browser context, enabling session hijacking, credential theft, or UI redirection. A publicly available proof-of-concept exploit exists, though no active exploitation has been confirmed via CISA KEV listing.
Open redirect in Phoenix LiveView's redirect/2 allows an unauthenticated remote attacker to send a victim's browser to an arbitrary external origin by embedding ASCII tab (\t), line feed (\n), or carriage return (\r) in the :to path argument. The private validate_local_url!/2 guard in lib/phoenix_live_view.ex rejects leading double-slashes and backslashes but does not strip or reject these whitespace control characters; browsers silently discard them before URL resolution, so a payload such as /<TAB>/example.com validates as a local path yet resolves to the scheme-relative URL //example.com. Affected versions span three release trains: 0.5.0–1.0.18, 1.1.0-rc.0–1.1.32, and 1.2.0-rc.0–1.2.8. No public exploit code or CISA KEV listing has been identified at time of analysis.
Uninitialized memory exposure in GNU C Library (glibc) versions 2.0 through 2.43 corrupts the we_wordv member of the wordexp_t structure when wordexp is called with the WRDE_APPEND flag under specific conditions, causing a subsequent wordfree call to abort the calling process. Disclosed by the glibc project as GLIBC-SA-2026-0014 and rooted in CWE-908 (Use of Uninitialized Resource), the flaw carries a CVSS 4.0 base score of 2.1, reflecting meaningful exploitation constraints including a local attack vector and high complexity. No public exploit code and no CISA KEV listing have been identified at time of analysis.
Cross-site scripting in Discourse's discourse-local-dates plugin allows an authenticated attacker to inject arbitrary HTML by embedding crafted markup in local-date format attributes (data-format, data-displayed-timezone, data-timezone), which were interpolated directly into the DOM via template literals instead of safe DOM text APIs. Exploitation is gated by a non-default condition: the site's Content Security Policy must be modified or disabled from the default. No public exploit independent of the vendor fix commits exists at time of analysis, and this is not listed in CISA KEV. The CVSS 4.0 score of 2.0 reflects the constrained attack requirements (AT:P, PR:L, UI:A).
Command injection in Dokploy's Docker build pipeline allows authenticated application editors to execute arbitrary OS commands on the Dokploy host by supplying shell metacharacters in the dockerContextPath build configuration field. The malicious path value flows from the dashboard UI through getDockerContextPath() directly into an unquoted cd shell command executed via execAsync, enabling full host-level code execution under the Dokploy server process. All Dokploy installations prior to version 0.29.13 are affected; no active exploitation or public exploit code has been identified at time of analysis, though exploitation requires only a low-privilege authenticated session.
In the Linux kernel, the following vulnerability has been resolved: drm/dp/mst: fix buffer overflows in sideband chunk accumulation drm_dp_sideband_append_payload() has three related bugs when processing device-provided sideband reply data: 1. Zero-length curchunk_len underflow: msg_len is a 6-bit field taken directly from the DP sideband header. If a device sends msg_len=0, curchunk_len is set to zero. The condition (curchunk_idx >= curchunk_len) is immediately true, and curchunk_len-1 wraps to 255 (u8 underflow). drm_dp_msg_data_crc4() reads 255 bytes from chunk[48], then memcpy() writes 255 bytes into msg[], both far out of bounds. 2. chunk[48] overflow: curchunk_len can reach 63 (6-bit field). chunk[] is only 48 bytes. Multi-iteration payload assembly appends 16-byte blocks until curchunk_idx reaches curchunk_len, writing up to 15 bytes past the end of chunk[] into msg[]. 3. msg[256] overflow: each chunk contributes (curchunk_len-1) bytes to msg[]. No check ensures curlen + (curchunk_len-1) stays within msg[256], so the memcpy can spill into adjacent struct fields. All three are reachable from any DP MST device that can forge sideband reply messages on a physical connection.
In the Linux kernel, the following vulnerability has been resolved: hwmon: (corsair-cpro) Stop device IO before calling hid_hw_stop Calling hid_hw_stop() does not stop the device IO. This results in a race condition between hid_input_report() and the point immediately following the execution of hid_device_io_start() within the driver probe function. If the probe operation fails after "io start" has been initiated, this race condition will result in a UAF vulnerability. Fix the problem by calling hid_device_io_stop() before calling hid_hw_stop().
In the Linux kernel, the following vulnerability has been resolved: hwmon: (nzxt-smart2) Stop device IO before calling hid_hw_stop Calling hid_hw_stop() does not stop the device IO. This results in a race condition between hid_input_report() and the point immediately following the execution of hid_device_io_start() within the driver probe function. If the probe operation fails after "io start" has been initiated, this race condition will result in a UAF vulnerability. Fix the problem by calling hid_device_io_stop() before calling hid_hw_stop().
In the Linux kernel, the following vulnerability has been resolved: ksmbd: defer destroy_previous_session() until after NTLM authentication In ntlm_authenticate(), destroy_previous_session() is called using a user pointer resolved from the client-supplied NTLM blob username field before the NTLMv2 response is validated. An authenticated attacker can set the NTLM blob username to match a victim account and set PreviousSessionId to the victim's session ID; destroy_previous_session() destroys the victim's session while ksmbd_decode_ntlmssp_auth_blob() subsequently rejects the request with -EPERM. Move destroy_previous_session() and the prev_id assignment to after ksmbd_decode_ntlmssp_auth_blob() returns success and use sess->user rather than the pre-authentication lookup result. This matches the ordering already used by krb5_authenticate(), where destroy_previous_session() is called only after ksmbd_krb5_authenticate() returns success.
In the Linux kernel, the following vulnerability has been resolved: btrfs: fix root leak if its reloc root is unexpected in merge_reloc_roots() If we have an unexpected reloc_root for our root, we jump to the out label but never drop the reference we obtained for root, resulting in a leak. Add a missing btrfs_put_root() call.
In the Linux kernel, the following vulnerability has been resolved: wifi: mac80211: free AP_VLAN bc_buf SKBs outside IRQ lock ieee80211_do_stop() removes AP_VLAN packets from the parent AP ps->bc_buf while holding ps->bc_buf.lock with IRQs disabled. It then calls ieee80211_free_txskb() before dropping the lock. ieee80211_free_txskb() is not just a passive SKB release. For SKBs with TX status state it can report a dropped frame through cfg80211/nl80211, and that path can reach netlink tap transmit. This is the same reason the pending queue cleanup in ieee80211_do_stop() already unlinks SKBs under the queue lock and frees them after IRQ state is restored. The buggy scenario involves two paths, with each column showing the order within that path: AP_VLAN management TX: AP_VLAN stop: 1. attach ACK-status state 1. clear the running state 2. queue a multicast SKB on 2. take ps->bc_buf.lock with IRQs parent ps->bc_buf disabled 3. unlink the AP_VLAN SKB 4. call ieee80211_free_txskb() Unlink matching AP_VLAN SKBs from ps->bc_buf under the existing lock, but move them to a local free queue. Drop the lock and restore IRQ state before calling ieee80211_free_txskb(). WARNING: kernel/softirq.c:430 at __local_bh_enable_ip
In the Linux kernel, the following vulnerability has been resolved: wifi: carl9170: bound memcpy length in cmd callback to prevent OOB read When the firmware sends a command response with a length mismatch, carl9170_cmd_callback() logs the mismatch and calls carl9170_restart() but then falls through to memcpy(ar->readbuf, buffer + 4, len - 4). Since len comes from the firmware and can exceed ar->readlen, this copies more data than the readbuf was allocated for. Bound the memcpy to min(len - 4, ar->readlen) so that the response is still completed -- avoiding repeated restarts from queued garbage -- while preventing an overread past the response buffer.
In the Linux kernel, the following vulnerability has been resolved: Bluetooth: btusb: validate Realtek vendor event length btusb_recv_event_realtek() reads the event code at data[0] and the Realtek subevent code at data[2] before deciding whether to consume a vendor event as a coredump. For example, the two-byte event ff 00 contains a complete vendor-event header declaring zero parameters. The old classifier still reads a nonexistent third byte and can misclassify the event as a coredump if the adjacent byte is 0x34. Require the HCI event header and first parameter to be present before inspecting the Realtek subevent code. Short events continue through the normal HCI receive path, which owns their protocol validation.
In the Linux kernel, the following vulnerability has been resolved: drm/dp/mst: fix OOB reads on 2-byte fields in sideband reply parsers Three sideband reply parsers read 16-bit fields as: val = (raw->msg[idx] << 8) | (raw->msg[idx+1]); and check bounds only after the fact. When idx == raw->curlen, raw->msg[idx+1] reads one byte past the received message data into the following struct fields (curchunk_len, curchunk_idx, curlen). Affected functions: - drm_dp_sideband_parse_enum_path_resources_ack() full_payload_bw_number and avail_payload_bw_number fields - drm_dp_sideband_parse_allocate_payload_ack() allocated_pbn field - drm_dp_sideband_parse_query_payload_ack() allocated_pbn field Fix by using a single combined check (idx + 2 > curlen) before each 2-byte read. Since the check is strictly tighter than idx > curlen, no separate step is needed. [added fixes tag]
In the Linux kernel, the following vulnerability has been resolved: drm/amdgpu/sdma6.0: replace BUG_ON() with WARN_ON() There's no need to crash the kernel for these cases. (cherry picked from commit c17a508a7d652da3728f8bbc481bfffe96d65a87)
In the Linux kernel, the following vulnerability has been resolved: drm/amd/display: dce100: skip non-DP stream encoders for DP MST On DCE8-class ASICs (e.g. Bonaire), the resource pool contains digital DIG stream encoders plus one analog DAC encoder. When assigning a stream encoder for a second DisplayPort MST stream, if the preferred digital encoder is already acquired, dce100_find_first_free_match_stream_enc_for_link() falls back to the first free pool entry. That entry may be the analog encoder, whose funcs table lacks DP hooks such as dp_set_stream_attribute. The subsequent atomic commit then dereferences NULL function pointers in link_set_dpms_on() and crashes. Skip encoders without dp_set_stream_attribute when the stream uses a DP signal (including MST). Use dc_is_dp_signal(stream->signal) for the MST fallback path instead of checking only the link connector signal. Tested on: - GPU: AMD Radeon R7 260X (Bonaire / DCE8) - Board: Supermicro C9X299-PG300 - Setup: DP MST daisy chain, hotplug second monitor or have it connected on boot - Kernel: 7.1.3 (issue observed since 6.19) - Result: kernel oops without patch; dual monitors stable with patch (cherry picked from commit 28ec64943e3ee4d9b8d30cea61e380f1429953a8)
In the Linux kernel, the following vulnerability has been resolved: media: airspy: Return queued buffers on start_streaming() failure The vb2 framework hands buffers to the driver via buf_queue() before calling start_streaming(). If start_streaming() returns an error without first returning those buffers via vb2_buffer_done(), vb2_start_streaming() fires WARN_ON(owned_by_drv_count) and the queued buffers leak. airspy_start_streaming() returned -ENODEV early when the USB device had been disconnected (s->udev == NULL) without returning any buffers that buf_queue() had already accepted. Take v4l2_lock first and jump to the existing err_clear_bit label, which already drains s->queued_bufs via vb2_buffer_done(..., VB2_BUF_STATE_QUEUED) before unlocking. This mirrors the uvcvideo fix in commit 4cf3b6fd54eb ("media: uvcvideo: Return queued buffers on start_streaming() failure").
In the Linux kernel, the following vulnerability has been resolved: media: cx23885: add ioremap return check and cleanup Add a check for the return value of pci_ioremap_bar() in cx23885_dev_setup(). If ioremap for BAR0 fails, release the already allocated PCI memory region, decrement the device count, and return -ENODEV. This prevents a potential null pointer dereference and ensures proper cleanup on memory mapping failure.