MacCMS10's admin template editor (application/admin/controller/Template.php) blocks dangerous PHP functions in template content via a blacklist regex, but the blacklist omitted exec, passthru, popen, show_source, create_function, register_shutdown_function, register_tick_function, and error_log. Combined with ThinkPHP's {if} template tag, which embeds the condition attribute directly into raw PHP (<?php if(condition): ?>), an authenticated administrator could inject a payload such as {if condition="exec('id > /tmp/pwned.txt')"}{/if} to achieve remote code execution. Fixed in commit 71ad3bb29570e110d8e973acff68040a3050ddf0 (2026-06-22), which added the missing functions to the filter.
IOTSmartHome's gui/login.php checkCookie() function builds an authentication query as SELECT * FROM users WHERE ID='<decoded lastLogin cookie>' after base64-decoding the client-supplied lastLogin cookie via safe_decode(), which performs URL-safe base64 decoding with no sanitization of the decoded value before it is concatenated into the SQL string. An unauthenticated attacker can set a lastLogin cookie containing a base64-encoded SQL injection payload (e.g. base64("' OR '1'='1")) to bypass authentication and, via UNION-based injection, extract arbitrary data including user credentials.
Zbtlink router firmware ships an embedded remote-control implant, ENDLESSDOORS, present in every published build across the product line. It is the open-source ycsunjane/rctl tool built in as an OpenWrt package (librctl.so), started at boot and run as root under the process name kworker to blend in with the kernel's [kworker/*] threads. It opens no listening port; it phones home over cleartext TCP to a hardcoded command-and-control server (command channel 7000, interactive-shell callback 7001) with no authentication and no transport encryption, re-attempting contact roughly every 35 seconds. Its command handler passes any received string to popen() as uid=0, and a reserved rctlbash command returns an interactive root shell. Because the channel is unauthenticated and cleartext, control is not limited to whoever planted it: any party that answers at the C2 address, occupies the network path (DNS or route hijack), or acquires the hardcoded fallback domain obtains unauthenticated remote code execution as root.
In Eclipse Theia versions up to and including 1.73.1, the `@theia/filesystem` backend binds `POST /file-upload` in every filesystem-enabled deployment. The handler takes an attacker-supplied absolute path from the multipart `uri` field and calls `fs.move(tmp, target, { overwrite: true })` with no workspace confinement and no authentication. In browser (non-Electron) deployments the connection token is enforced only on WebSocket upgrades; the HTTP middleware in `@theia/core` re-issues the cookie and calls `next()` without rejecting tokenless HTTP requests. Because `multipart/form-data` is a CORS-safelisted request type, a cross-origin web page can trigger the write with no preflight and no credentials, resulting in an unauthenticated arbitrary file write outside the workspace to any absolute path the backend process can write. This can escalate to remote code execution, for example by overwriting a startup-executed file such as `~/.bashrc`. Electron mode uses a separate `ElectronSecurityToken` and is not affected via this path.
Kong Event Gateway versions 1.0.0 through 1.1.1 and 1.2.0 do not enforce key rotation before reaching NIST SP 800-38D recommended usage limit for AES-GCM encryption keys with random nonces when the AWS IAM encryption feature is enabled.
If a producer sends messages at a sustained high rate without key rotation, which only occurs on reboot of the Kong Event Gateway instance, the probability of a nonce collision becomes non-negligible. An authorized consumer who detects a nonce collision can recover parts of plaintext from the affected messages.
New versions 1.1.2 and 1.2.1 enforce automatic key rotation before the recommended usage limit is reached.
In Eclipse Theia versions 0.7.0 and up until including 1.73.1, the `PreferenceUtils.merge` function in `@theia/core` recursively merges preference values without rejecting prototype-related keys (`__proto__`, `constructor`, `prototype`). Because this function is invoked by `PreferenceServiceImpl.doResolve` for every preference resolution across scopes (default, user, workspace, folder), a crafted preference value in a workspace settings file (`.theia/settings.json` or `.vscode/settings.json`) can pollute `Object.prototype` when the user opens the workspace, potentially altering application logic across the Theia process.
In Eclipse Accessibility Tools Framework (ACTF) versions up to 1.6.0 (including source code versions up to v20260630 and ACTF based application miChecker versions up to 3.1.0), it has been identified that an XML External Entity (XXE) vulnerability exists.
If this vulnerability is exploited, a malicious third party could gain access to local resources or internal network resources via computer running applications that use Eclipse ACTF, including miChecker.
In Eclipse Theia versions 1.66.0 and up until including 1.73.1, the `@theia/plugin-ext` backend exposes the `/hostedPlugin/:pluginId/:path(*)` HTTP endpoint, which resolves the requested file path with `path.resolve(localPath, filePath)` without verifying that the resolved path stays within the plugin's directory. An unauthenticated network attacker can send percent-encoded `../` sequences (`%2e%2e%2f`) that decode into the path parameter and escape the plugin directory, allowing arbitrary files readable by the Theia backend process to be retrieved. Plugin IDs are derived deterministically from a plugin's publisher and name, so built-in plugins serve as reliable anchors that require no prior knowledge of the target system.
A privilege escalation vulnerability exists in Rancher's impersonation middleware (pkg/auth/requests/impersonate.go). An authenticated Rancher user with the default user
global role can gain full administrative access to the Rancher control
plane and transitively to all downstream clusters it manages.
This issue affects Rancher: from 2.11.0 before 2.11.16, from 2.12.0 before 2.12.12, from 2.13.0 before 2.13.8, and from 2.14.0 before 2.14.2.
NeuVector through 5.4.9 is can potentially leak information from manager /network/graph API due to missing authentication and cached data containing sensitive information.
The Smash Balloon Social Photo Feed – Easy Social Feeds Plugin plugin for WordPress is vulnerable to Reflected Cross-Site Scripting via REQUEST_URI Query String in all versions up to, and including, 6.11.3 due to insufficient input sanitization and output escaping. This makes it possible for unauthenticated attackers to inject arbitrary web scripts in pages that execute if they can successfully trick a user into performing an action such as clicking on a link.
Denial-of-service vulnerability in M-Files Server versions before 26.5.16015.3 allows an authenticated admin user to cause the M-Files Server process to crash and fail to restart.
The ZTE Smart Life app contains an SQL injection vulnerability that allows attackers to execute UNION SELECT statements to query sensitive data in the feedback.db database across tables, including user accounts, phone numbers, feedback content, and local debug log paths, thereby enabling the theft of local privacy data.
A flaw was found in the Application Subscription controller (multicluster-operators-subscription) of Red Hat Advanced Cluster Management for Kubernetes (ACM). A user with namespace-scoped "edit" privileges in an ACM hub namespace can create a Channel resource pointing to a Helm repository they control and a Subscription resource referencing it. The app-subscription controller fetches and applies the Helm chart contents with its own elevated authority, without verifying whether the subscription creator holds the "open-cluster-management:subscription-admin" role and without restricting applied resources to the subscription namespace. This allows the attacker to include cluster-scoped resources in the Helm chart, such as a ClusterRoleBinding granting the attacker's ServiceAccount the "cluster-admin" ClusterRole. Successful exploitation results in full cluster-admin privilege escalation. This contradicts the ACM documentation which states that non-subscription-admin users should have resources deployed into the subscription namespace only.
A flaw was found in the Multicluster Engine for Kubernetes ClusterCurator controller. A tenant administrator with namespace-scoped privileges can exploit this vulnerability by creating a namespaced ClusterCurator. This action inadvertently grants the tenant administrator the ability to mint a token for a ServiceAccount with cluster-wide administrative authority. This leads to a privilege escalation, allowing the tenant administrator to gain full control over the cluster.
The Layouts for WPBakery plugin for WordPress is vulnerable to unauthorized actions due to a missing capability check on the `Layouts_WPB_Remote::template_sync()` callback registered via `wp_ajax_nopriv_handle_sync` in all versions up to, and including, 1.1.3. This makes it possible for unauthenticated attackers to force the WordPress server to issue outbound HTTP requests to the plugin vendor's external API (`https://www.layoutsforwpbakery.com/wp-json/layoutsforwpbakery/v1/{templates,categories}`) and to write the JSON-decoded responses verbatim into the site's `wp_options` table via `set_transient()` — at any rate the attacker chooses, with no nonce verification, capability check, or rate limiting.
The Backup Migration plugin for WordPress is vulnerable to OS Command Injection in all versions up to, and including, 2.1.5.1 due to insufficient sanitization of the `file` POST parameter on the `restoreBackup()` AJAX handler. The handler applies `esc_attr()` — an HTML-context sanitizer that does not strip shell metacharacters — and concatenates the result, unquoted, into a `php-cli -f … bmi_restore <file> <remote>` command passed to `exec()`. This makes it possible for authenticated attackers, with Administrator-level access (or any user granted the plugin's `do_backups` capability) and above, to execute arbitrary OS commands as the web-server user, bypassing WordPress hardening constants such as `DISALLOW_FILE_EDIT` and `DISALLOW_FILE_MODS` that would otherwise prevent code execution from the admin UI. This is an incomplete fix of CVE-2023-7002, which patched the same pattern only in the `$_POST['url']` path of `handleQuickMigration()`; the equivalent mitigations (`rawurlencode()` + explicit shell-metachar replacement + double-quoting in `exec()`) were never applied to `$backupName`.
The MailChimp Forms by MailMunch plugin for WordPress is vulnerable to unauthorized modification of data due to a missing capability check on the `sign_in()` and `sign_up()` AJAX handlers in all versions up to, and including, 3.2.7. This makes it possible for authenticated attackers, with Subscriber-level access and above, to relink the site's MailMunch integration to an attacker-controlled MailMunch account by submitting attacker-supplied credentials. Once relinked, all subscriber data captured by the plugin's forms is delivered to the attacker, and the forms/landing pages rendered on the site are pulled from the attacker's MailMunch account.
The Search Analytics for WP plugin for WordPress is vulnerable to Cross-Site Request Forgery in all versions up to, and including, 1.4.16. This is due to missing or incorrect nonce validation on the `process_bulk_action()` function of `MWTSA_Stats_Table`. This makes it possible for unauthenticated attackers to delete arbitrary search-term records, including all associated search-history rows, via a forged request granted they can trick a user with access to the plugin's "Search Analytics" dashboard page (Administrator by default) into performing an action such as clicking on a link.
The Simple Yearly Archive plugin for WordPress is vulnerable to Stored Cross-Site Scripting via the `posttype` attribute of the `SimpleYearlyArchive` shortcode in all versions up to, and including, 2.2.4 due to insufficient input sanitization and output escaping on user supplied attributes. This makes it possible for authenticated attackers, with Contributor-level access and above, to inject arbitrary web scripts in pages that will execute whenever a user accesses an injected page.
The Xpro Addons plugin for WordPress is vulnerable to unauthorized creation of data due to a missing capability check on the `get_menu_content_editor()` function in all versions up to, and including, 1.5.1. This makes it possible for authenticated attackers, with Subscriber-level access and above, to create arbitrary published posts of the `xpro_content` custom post type with attacker-controlled titles. The created posts are publicly queryable on the front-end, enabling content injection, SEO spam, and database pollution.
art-template's sub-template resolution logic (src/compile/adapter/resolve-filename.js), used by both the include() and extend() template directives, resolves the target file path via path.resolve(root, filename) with no check afterward that the result remains inside root. Because path.resolve() discards root entirely when filename is an absolute path, and does not block '../' traversal sequences, and the resolved path is passed directly to fs.readFileSync() in loader.js with its contents compiled and rendered, an application that lets a sub-template name be influenced by external input (e.g. a query parameter passed into {{include page}}) allows an attacker to read arbitrary files on disk that the Node process can access.
The Aerie/PlanDev sequencing-server's authorization middleware (sequencing-server/src/app.ts) derives the caller's Hasura session role via getHasuraSession(), which prefers a session_variables object taken directly from the client-supplied JSON request body over the Authorization header's JWT claims, with no verification that the request actually originated from Hasura. By setting {"session_variables":{"x-hasura-role":"aerie_admin"}} in the body of a request to POST /command-expansion/put-expansion with no Authorization header, an unauthenticated attacker satisfies the role check and can insert arbitrary expansion rules into sequencing.expansion_rule, which govern how spacecraft activities are translated into commands. Separately, POST /put-dictionary is explicitly listed in the ENDPOINTS_WHITELIST and is exempt from any authentication, allowing unauthenticated writes of command dictionaries.
Typemill's login endpoint (POST /tm/login, ControllerWebAuth::login()) performs no rate-limiting, failed-attempt counting, or account lockout when captcha is disabled, which is the default configuration. An unauthenticated attacker can send unlimited password-guessing requests against any account, including administrators, with no throttling. The only attempt-counting/lockout logic present in the same file protects an optional secondary email-authcode step and does not apply to the primary password check.
xidown (a yt-dlp/ffmpeg GUI wrapper) builds its yt-dlp command-line invocation (xidown/core/scanner.py and downloader.py) by appending the user-provided or scanned URL as a bare trailing positional argument, with no '--' end-of-options marker and no scheme validation anywhere in the codebase. Because yt-dlp parses any argument beginning with '-' as a CLI option rather than link text, a crafted 'URL' value such as -U (yt-dlp's self-update flag) or --exec=... is parsed as a real yt-dlp option instead of a URL, altering the tool's control flow before its own URL validation runs. Full code execution via --exec was not demonstrated in the single-URL flow tested, but the underlying argument-injection primitive is confirmed and unmitigated across all call sites.
MLflow's AI Gateway accepts an auth_config.api_base value when creating a gateway secret (mlflow/server/handlers.py, _create_gateway_secret) with no validation of scheme, host, or IP range; the value is stored verbatim. The gateway proxy endpoint (mlflow/server/gateway_api.py, raw_proxy) subsequently issues an HTTP request to that stored api_base plus a caller-supplied path and returns the full response body. MLflow's existing SSRF guard, _validate_webhook_url (which blocks non-global and metadata IPs), is never invoked anywhere in this gateway secret/proxy code path. The CreateGatewaySecret action additionally has no entry in the permission-validator map, so it requires only basic authentication rather than any specific scope, meaning any authenticated user — including read-only accounts — can create a secret pointing at an internal address and reach it via the proxy endpoint, potentially exposing cloud-instance IAM credentials via metadata services. This is related to CVE-2026-4035, which addresses a distinct mechanism in the same gateway-secret feature (server-side $ENV_VAR resolution inside the api_key field leaking credentials to the configured upstream); the finding here is an independent missing-validation gap in the api_base destination itself, unaffected by that fix.
Mealie's AsyncSafeTransport SSRF guard (mealie/pkgs/safehttp/transport.py) resolves a target hostname once, checks the resolved IP against private-range rules, but then issues the actual outbound HTTP request using the original hostname, which the underlying async transport re-resolves independently. Because the validated IP is never pinned to the actual connection, a DNS-rebinding attacker (returning a public IP to the validation lookup and a private/metadata IP to the real connection) defeats the guard. This is reachable by any authenticated user via /api/recipes/create/url, /api/recipes/test-scrape-url, and /api/recipes/{slug}/image, and the scraper reflects fetched content back to the requester, allowing an authenticated user to read internal HTTP services and cloud-metadata endpoints.
audiobookshelf's authentication-exemption check (server/routers/Auth.js) matches unauthenticated-allowed GET routes against req.path via a regex requiring a literal /items/:id/cover or /authors/:id/image shape, where req.path retains %2F sequences URL-encoded. Express's router decodes the :id route parameter before handler code runs, so a %2F-encoded '../' sequence in :id (e.g. ..%2f..%2f..%2ftmp%2fpwned) passes the literal-path auth-exemption check while resolving to a real path-traversal payload once decoded. CacheManager.handleCoverCache then joins this decoded value into a cache file path and streams the result before any database-backed ownership check. This bypasses the fix applied for CVE-2025-25205 (which anchored the exemption regex and switched it to req.path) and results in unauthenticated arbitrary file read of any file matching the pattern *_<width>[x<height>].<ext> that the service account can read.
KubeSphere's cluster-controller reconciliation (pkg/utils/clusterclient/clusterclient.go, addCluster) processes every Cluster custom resource's connection configuration and immediately calls Discovery().ServerVersion() against the CRD-specified Kubernetes API endpoint, which is parsed only for URL syntax (url.Parse) with no allow/deny-list for loopback, RFC1918 private ranges, link-local, or cloud-metadata addresses (e.g. 169.254.169.254). A user able to create or update a Cluster CRD can force the controller-manager and apiserver pods to issue outbound requests to arbitrary internal or metadata endpoints.
The Stock-Inventory-Management-System application's login.php assigns raw $_POST username/password values to $_SESSION and builds its authentication query by directly concatenating those session values into a SQL statement with no parameterization or escaping. An unauthenticated remote attacker can submit a payload such as ' OR '1'='1 in the login form to bypass authentication entirely. The same script additionally contains hardcoded administrative credentials (admin/neola) in a post-login conditional check, providing a second, independent full-authentication-bypass path.
Shiori's CheckToken function (internal/domains/auth.go) validates only the JWT's HMAC signature and returns the embedded claims.Account object unmodified, never re-fetching the account from the database. No session store or token-revocation mechanism exists in the codebase. Deleting an account or demoting it from owner to a regular role has no effect on tokens already issued to that account — a deleted or demoted owner's token continues authenticating with its original owner-level privileges until natural expiry, which can be up to 30 days with 'remember me' enabled.
changedetection.io's /login route checks the submitted password against a single PBKDF2-HMAC-SHA256 hash with no per-IP or per-session rate limiting, failed-attempt counter, or lockout (no rate-limiting library is present in requirements.txt). Because the entire application is protected by one shared password with no per-user accounts, a successful brute-force guess grants full administrative access, including the ability to view/regenerate the API token.
changedetection.io's /settings save handler builds an update dict from form.data['application'] and blind-merges it into the stored application settings via .update(). Because WTForms represents an unchecked checkbox as False rather than 'unchanged', and only the 'password' field is special-cased against this problem, a POST to /settings that omits the api_access_token_enabled field (e.g. a minimal scripted request) silently disables API key enforcement for the entire REST API, exposing the full watch list, history, and configuration to unauthenticated requests.
changedetection.io's REST API resources are protected by an @auth.check_token decorator validating the caller's x-api-key header, except the Spec resource registered at /api/v1/full-spec (changedetectionio/api/Spec.py), whose get() method carries neither @auth.check_token nor @validate_openapi_request. An unauthenticated client can retrieve the full merged OpenAPI schema (all endpoint paths, parameters, and registered processor plugins) even when API access control is enabled and every sibling /api/v1/* route correctly requires the key.
The raster Rust crate's crop() function (src/editor.rs) clamps the crop width/height against source dimensions but only clamps the offset_x/offset_y parameters against 0, never against the source width/height. When an offset exceeds the corresponding source dimension, `width2 - offset_x` (or the height equivalent) underflows to a negative i32, which release builds do not trap; the negative value is then cast to usize inside Image::blank()'s Vec::with_capacity() call, triggering a capacity-overflow panic and crashing the process on a single crafted crop request.
imagecli's `carve <ratio>` pipeline operation (Carve::apply() in src/image_ops.rs) only asserts `ratio <= 1.0`, never validating that the ratio is positive. A negative ratio (e.g. -5) causes the computed target width to saturate to 0 via Rust's defined float-to-uint cast, which is then passed to imageproc::seam_carving::shrink_width — a function that panics when given a width below 2, crashing the process. This shares the same missing-input-validation root cause as the sibling `scale` finding in the same file but is an independently fixable, distinct code path.
imagecli's `scale <ratio>` pipeline operation (Scale::apply() in src/image_ops.rs) computes output width/height as (dimension as f32 * ratio) as u32 with no upper-bound validation on the CLI-supplied ratio, which is parsed via nom::number::complete::float with no range check. A large ratio (e.g. 100000) causes an attempted allocation of hundreds of terabytes, aborting the process. Any application embedding imagecli as a library and accepting user-controlled pipeline strings is remotely crashable with a single request.
Pluck CMS's admin panel relies solely on a Referer-header comparison (requestedByTheSameDomain() in data/inc/functions.admin.php, gating every admin.php action) for CSRF protection, with no per-request anti-CSRF token anywhere in the admin area. When a request carries no Referer/Host information, the function's elseif branch returns true, treating the request as same-origin. Because a cross-site attacker page can suppress the Referer header (e.g. via <meta name=referrer content=no-referrer>), it can force an authenticated administrator's browser to submit forged admin actions with no valid Referer, including creating pages with raw HTML (stored XSS via the rendered page) and installing PHP modules/themes (remote code execution).
The SKT Skill Bar plugin for WordPress is vulnerable to Stored Cross-Site Scripting via the `chart_size` attribute of the `skillwrapper` shortcode in all versions up to, and including, 2.6. This is due to insufficient input sanitization and output escaping on the `chart_size` attribute, which is concatenated directly into an inline `<style>` block. This makes it possible for authenticated attackers, with Contributor-level access and above, to inject arbitrary web scripts in pages that will execute whenever a user accesses an injected page.
The AI Chatbot & Workflow Automation by AIWU plugin for WordPress is vulnerable to Sensitive Information Exposure in all versions up to, and including, 1.4.6. This is due to the `getCurrentTaskResults()` method in `modules/workspace/controller.php` being accessible without authentication or authorization checks. The method is not included in the workspace controller's `getNoncedMethods()` array, the base `getPermissions()` returns an empty array, and all AJAX actions are registered with `wp_ajax_nopriv_` hooks (`classes/frame.php:282`). When tasks are created via features like the Bulk Post Generator, the task parameters — including the OpenAI API key in plaintext, AI prompts, keywords, and full AI model configuration — are stored in the database and returned in the JSON response. This makes it possible for unauthenticated attackers to enumerate sequential task IDs and retrieve sensitive configuration data including API keys.
The WPFormify – Stripe Payments with Form and Checkout plugin for WordPress is vulnerable to unauthorized modification and deletion of Stripe payment credentials in all versions up to, and including, 1.1.1. This is due to missing capability checks and nonce verification on the `wpf_stripe_callback_success()` and `wpf_stripe_disconnect()` functions, both hooked to `admin_init`. The `admin_init` hook fires on `admin-post.php` which is accessible without authentication. This makes it possible for unauthenticated attackers to overwrite the site's Stripe API credentials with attacker-controlled values (redirecting payments to the attacker's Stripe account) or disconnect the Stripe integration entirely by deleting the stored credentials.
The LightSync Pro plugin for WordPress is vulnerable to arbitrary file uploads due to missing file type validation in the rest_replace_media() function in all versions up to, and including, 2.1.6. This makes it possible for authenticated attackers, with Author-level access and above, to upload arbitrary files on the affected site's server which may make remote code execution possible.
The Material Dashboard plugin for WordPress is vulnerable to unauthorized access and modification of data due to missing capability checks on the amd_ajax_target_task_manager() function in all versions up to, and including, 1.4.10. This makes it possible for unauthenticated attackers to enumerate all scheduled tasks (potentially exposing PII), execute arbitrary tasks, and delete any task via the public_amd_ajax_handler AJAX action.
The ShopLentor plugin for WordPress is vulnerable to arbitrary function execution via the woolentoropt/v1/custom-action REST API endpoint in all versions up to, and including, 3.3.7. This is due to the handle_action() method passing user-supplied input directly to call_user_func() without an allowlist of permitted callbacks. This makes it possible for authenticated attackers, with Administrator-level access and above, to execute arbitrary PHP callable functions via the 'callback' parameter.
In the Linux kernel, the following vulnerability has been resolved:
xfrm: fix sk_dst_cache double-free in xfrm_user_policy()
xfrm_user_policy() clears the socket dst cache with __sk_dst_reset(),
i.e. the non-atomic __sk_dst_set(sk, NULL): it reads sk_dst_cache with
rcu_dereference_protected(), stores NULL and dst_release()s the old dst.
That is only safe if no other thread modifies sk_dst_cache concurrently.
For a connected UDP socket that does not hold: the transmit fast path
(udp_sendmsg -> sk_dst_check -> sk_dst_reset) resets the cache locklessly
with an atomic xchg(). A per-socket policy change racing a send can make
both sides observe the same old dst and each dst_release() it, dropping
the socket's single reference twice and freeing the xfrm_dst bundle while
it is still referenced:
BUG: KASAN: slab-use-after-free in dst_release
Write of size 4 at addr ffff88801897b6c0 by task exploit/155
Call Trace:
...
dst_release (... ./include/linux/rcuref.h:109)
xfrm_user_policy (./include/net/sock.h:2239 ./include/net/sock.h:2256 net/xfrm/xfrm_state.c:3053)
do_ip_setsockopt (net/ipv4/ip_sockglue.c:1347)
ip_setsockopt (net/ipv4/ip_sockglue.c:1417)
do_sock_setsockopt (net/socket.c:2368)
__sys_setsockopt (net/socket.c:2393)
__x64_sys_setsockopt (net/socket.c:2396)
do_syscall_64 (arch/x86/entry/syscall_64.c:94)
entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:121)
Reachable by an unprivileged user via a user+network namespace.
Use the atomic sk_dst_reset() so the cache is cleared and released with a
single xchg(): whichever side wins releases the dst once, the other sees
NULL and does nothing. Behaviour is otherwise unchanged.
In the Linux kernel, the following vulnerability has been resolved:
xfrm6: clear dst.dev on error to avoid double netdev_put in xfrm6_fill_dst()
On the error path where in6_dev_get(dev) returns NULL, xfrm6_fill_dst()
releases the device reference with netdev_put() but leaves
xdst->u.dst.dev set. dst_destroy() later calls netdev_put(dst->dev)
again, so the same net_device reference is released twice, underflowing
its refcount (ref_tracker WARNING + "unregister_netdevice: waiting for
<dev> to become free").
Clear xdst->u.dst.dev after the netdev_put(), the same way the XFRM
device-offload paths xfrm_dev_state_add() and xfrm_dev_policy_add() in
net/xfrm/xfrm_device.c NULL ->dev when releasing the reference on error.
ref_tracker: reference already released.
ref_tracker: allocated in:
xfrm6_fill_dst (net/ipv6/xfrm6_policy.c:86)
...
udpv6_sendmsg (net/ipv6/udp.c:1696)
...
ref_tracker: freed in:
xfrm6_fill_dst (net/ipv6/xfrm6_policy.c:90)
...
WARNING: lib/ref_tracker.c:322 at ref_tracker_free+0x58b/0x780
dst_destroy (net/core/dst.c:115)
rcu_core
handle_softirqs
...
In the Linux kernel, the following vulnerability has been resolved:
xfrm: policy: preallocate inexact bins before xfrm_hash_rebuild reinsert
xfrm_hash_rebuild()'s first loop preallocates the bins/chains the reinsert
loop needs, so the reinsert (after hlist_del_rcu()) cannot allocate or
fail. But its guard is inverted: it skips policies with prefixlen <
threshold and preallocates for the rest.
prefixlen < threshold is exactly when policy_hash_bysel() returns NULL and
the reinsert takes the allocating xfrm_policy_inexact_insert() path. So the
loop preallocates for the exact policies (which never allocate) and skips
the inexact ones, whose bin/node is then allocated GFP_ATOMIC during
reinsert. On failure the error path only WARN_ONCE()s and continues,
leaving a poisoned bydst node; the next rebuild's hlist_del_rcu()
dereferences LIST_POISON2 and takes a GPF. Reachable under memory pressure,
deterministic via failslab.
Invert the guard so preallocation covers exactly the reinserted policies;
the reinsert then allocates nothing and cannot fail.
Crash:
Oops: general protection fault, probably for non-canonical address
0xfbd59c0000000024: 0000 [#1] SMP KASAN NOPTI
KASAN: maybe wild-memory-access in range [0xdead...]
...
Workqueue: events xfrm_hash_rebuild
RIP: 0010:xfrm_hash_rebuild+0x5b3/0x1190
RAX: dead000000000122 (LIST_POISON2 + offset)
...
Call Trace:
hlist_del_rcu (include/linux/rculist.h:599)
xfrm_hash_rebuild (net/xfrm/xfrm_policy.c:1365)
process_one_work (kernel/workqueue.c:3322)
worker_thread (kernel/workqueue.c:3486)
kthread (kernel/kthread.c:436)
ret_from_fork (arch/x86/kernel/process.c:158)
ret_from_fork_asm (arch/x86/entry/entry_64.S:245)
...
Kernel panic - not syncing: Fatal exception in interrupt
In the Linux kernel, the following vulnerability has been resolved:
ksmbd: validate compound request size before reading StructureSize2
When ksmbd validates a compound (chained) SMB2 request,
ksmbd_smb2_check_message() reads pdu->StructureSize2 without first
checking that the compound element is large enough to contain it.
StructureSize2 is a 2-byte field at offset 64
(__SMB2_HEADER_STRUCTURE_SIZE) from the start of each element.
The compound-walking logic only guarantees that a full 64-byte SMB2
header is present for the trailing element: when NextCommand is 0, len is
reduced to the number of bytes remaining after next_smb2_rcv_hdr_off. A
remote client can craft a compound request whose last element has exactly
64 bytes, so the 2-byte StructureSize2 read at offset 64 extends one byte
past the receive buffer, producing a slab-out-of-bounds read.
BUG: KASAN: slab-out-of-bounds in ksmbd_smb2_check_message (fs/smb/server/smb2misc.c:402)
Read of size 2 at addr ffff888012ae31ac by task kworker/0:1/14
The buggy address is located 172 bytes inside of allocated 173-byte region
Workqueue: ksmbd-io handle_ksmbd_work
Call Trace:
...
kasan_report (mm/kasan/report.c:595)
ksmbd_smb2_check_message (fs/smb/server/smb2misc.c:402)
handle_ksmbd_work (fs/smb/server/server.c:119)
process_one_work (kernel/workqueue.c:3314)
worker_thread (kernel/workqueue.c:3397)
kthread (kernel/kthread.c:436)
ret_from_fork (arch/x86/kernel/process.c:158)
ret_from_fork_asm (arch/x86/entry/entry_64.S:245)
Reject any compound element that is too small to hold StructureSize2
before dereferencing it.
In the Linux kernel, the following vulnerability has been resolved:
gtp: check skb_pull_data() return in gtp1u_send_echo_resp()
gtp1u_send_echo_resp() ignores skb_pull_data()'s return value. Its
caller gtp1u_udp_encap_recv() only guarantees 16 bytes (udphdr +
gtp1_header), but the pull requests 20 (gtp1_header_long + udphdr). For
a 16-19 byte echo request the pull fails and returns NULL without
advancing skb->data; execution continues, and the following skb_push()
plus the IP header pushed by iptunnel_xmit() move skb->data below
skb->head, tripping skb_under_panic().
Fix it by dropping the packet when skb_pull_data() fails.
skbuff: skb_under_panic: ...
kernel BUG at net/core/skbuff.c:214!
Call Trace:
skb_push (net/core/skbuff.c:2648)
iptunnel_xmit (net/ipv4/ip_tunnel_core.c:82)
gtp_encap_recv (drivers/net/gtp.c:701 drivers/net/gtp.c:808 drivers/net/gtp.c:920)
udp_queue_rcv_one_skb (net/ipv4/udp.c:2388)
...
Kernel panic - not syncing: Fatal exception in interrupt
In the Linux kernel, the following vulnerability has been resolved:
nexthop: initialize extack in nh_res_bucket_migrate()
nh_res_bucket_migrate() passes an uninitialized netlink_ext_ack to
call_nexthop_res_bucket_notifiers(). When
nh_notifier_res_bucket_info_init() fails (e.g. the kzalloc returns
-ENOMEM), the error is propagated back before any notifier sets
extack._msg, and the error path formats the stale pointer with
pr_err_ratelimited("%s\n", extack._msg). With CONFIG_INIT_STACK_NONE
this dereferences uninitialized stack memory:
Oops: general protection fault, probably for non-canonical address ...
KASAN: maybe wild-memory-access in range [...]
RIP: 0010:string (lib/vsprintf.c:730)
vsnprintf (lib/vsprintf.c:2945)
_printk (kernel/printk/printk.c:2504)
nh_res_bucket_migrate (net/ipv4/nexthop.c:1816)
nh_res_table_upkeep (net/ipv4/nexthop.c:1866)
rtm_new_nexthop (net/ipv4/nexthop.c:3323)
rtnetlink_rcv_msg (net/core/rtnetlink.c:7076)
netlink_sendmsg (net/netlink/af_netlink.c:1900)
Kernel panic - not syncing: Fatal exception
Zero-initialize extack so _msg is NULL on error paths that never set it.