☠️
EXFILTRATING
DATA...
01010111 10110010 00101001 11000110 01101100 10011011 00010001 11100101 01000010 10101111
[ <-- BACK_TO_SEARCH ]
Qualcomm logo [ WEBSITE ] [ VENDOR STATS ]
CVSS: 7.8 CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H 🐛 CWE-416
👁️
Common Weakness Enumeration - Identifies the specific category of software weakness or flaw.

CVE-2026-57554

CVE-2026-57554 Use-After-Free in Qualcomm
📅 PUBLISHED: 2026-10-06 00:00 UTC 🔄 UPDATED: 2026-10-09 16:35 UTC
[ ZERO-DAY / UNPATCHED ] [ DEFENSE PATCH ↓ ]
💣 PoC: None
👁️
Proof of Concept - Indicates if functional exploit code is publicly available, drastically increasing the risk of immediate attacks.
🎯 EPSS: 0.11%
👁️
0.11% probability
2.8× above the median CVE
Tier: ELEVATED
30-day exploitation window · FIRST.org · updated daily
🥷 MITRE: T1068
👁️
MITRE ATT&CK Tactic - Describes the strategy hackers use to execute this attack.
Country: United States License: Commercial
VULNERABLE: UNKNOWN (no affected version ranges disclosed by any tracked source)
CVSS VECTOR BREAKDOWN 7.8
AV: L Attack Vector Local
AC: L Attack Complexity Low
PR: L Privileges Required Low
UI: N User Interaction None
S: U Scope Unchanged
C: H Confidentiality High
I: H Integrity High
A: H Availability High
7.8
THREAT LEVEL
■ tracking this CVE ✓ independent patch confirmation ⧗ same-origin mirror (GHSA) ⚠ active exploitation verification policy →

Executive Summary

CVE-2026-57554 is a memory-corruption vulnerability (CWE-416, use-after-free class) in Qualcomm Snapdragon G1 Gen 2 Gaming Platform firmware, arising when asynchronous threads access shared performance-counter data simultaneously during FastRPC invocations. It is a local, low-complexity issue requiring low privileges and no user interaction, with CVSS 3.1 base score 7.8 (High) across all three impact metrics (C:H/I:H/A:H). A local attacker on a device exposing the FastRPC interface can corrupt memory in the DSP-adjacent subsystem and potentially escalate to code execution at the privilege level of the affected firmware component.

⚡ AUTOMATE YOUR CVE INTELLIGENCE

This record — and 170,000+ others — are available via the ValtersIT API. One endpoint replaces five: NVD + GitHub CVE + CISA KEV + EPSS + Nuclei + Exploit-DB, unified.

›Patch confidence — UNVERIFIED to CONFIRMED, cross-referenced NVD + RHSA + Debian
›Exploit & PoC — poc_status, poc_url, has_exploit. Not in NVD.
›Shodan dorks — ready-made queries to find exposed vulnerable systems
›Detection rules — YARA · Wazuh · Elastic (all plans) · Sigma (Pro)
›pending_fields — know exactly which of 7 fields NVD hasn't processed yet
›Free backfill — data incomplete when you paid? Re-check within 30 days = 0 credits
PAYG €0.08/lookup Standard €19/mo (was €29) · 5k credits Pro €65/mo (was €99) · 20k credits FOUNDING PRICE · LOCKED FOR LIFE · ENDS 31 DEC 2026
[ FROM €19/MO ]
Free sandbox key available
No signup required

Technical Analysis

The NVD description states: "Memory Corruption when asynchronous threads access shared performance counter data simultaneously during FastRPC invocations." Three technical facts are derivable, and nothing beyond them should be assumed.

First, the vulnerability class is memory corruption with a concurrent-access root cause. CWE-416 (Use-After-Free) is assigned, meaning the corruption is consistent with a reference to a memory object (here, the shared performance-counter data structure) remaining live or being reused after it has been freed, with a second asynchronous thread operating on that stale reference.

Second, the trigger condition is a race. The description explicitly requires asynchronous threads accessing the shared performance-counter data simultaneously during FastRPC invocations. This is a classic time-of-check/time-of-use or unsynchronized-shared-state defect: the performance-counter data is a shared resource whose lifecycle is not adequately protected against concurrent access. Absent proper locking or reference counting, one thread can free or reallocate the object while another still holds a pointer into it. The race window means exploitation is timing-dependent in practice, though CVSS AC:L indicates the vendor assessed the required window as routinely attainable.

Third, the attack surface is the FastRPC invocation path. FastRPC is Qualcomm's remote procedure call mechanism used to invoke functions that execute on the DSP/remote processor from the application processor side. The vulnerable component lives in the firmware/subsystem handling that invocation traffic. That means the attacker must be able to reach the FastRPC invocation path from an unprivileged or low-privileged local context — the typical route is a local application able to issue RPC calls to the relevant remote service.

What is not available in the source data: the specific function or interface name that performs the shared-counter access, the data structure layout, the exact FastRPC service or handle exposed to untrusted callers, and the firmware build in which the flaw resides. No function names, API paths, or file paths are stated in the NVD description. Any such specifics would be fabricated and are deliberately omitted.

The memory-corruption primitive class matters for exploitation: a use-after-free on a shared kernel- or firmware-adjacent object is generally sought after because it can yield both a read/write primitive (via controlled reallocation into a stale slot) and, subsequently, control-flow hijack or data-structure manipulation. The scoring reflects this: Confidentiality, Integrity, and Availability are all rated High, which is more consistent with a corruption primitive that can be leveraged beyond a simple denial of service than with a crash-only bug.

Exploitation Scenarios

No public PoC, exploit code, or Nuclei template exists for this CVE, and none is described in the source data. The attack chain below is inferred strictly from the vulnerability class (CWE-416 with an asynchronous shared-state race on the FastRPC path) and from the CVSS profile (local, low complexity, low privileges, no user interaction). It is a class-level chain, not a reproduction recipe.

1. Reconnaissance / target identification. The attacker needs a target device running the affected Snapdragon G1 Gen 2 Gaming Platform firmware with the relevant FastRPC service reachable. Because the vector is local (AV:L) with low privileges (PR:L), the realistic attacker has already obtained some foothold: a malicious or compromised app, another local user, or a constrained-code process. Reconnaissance on modern Android/Qualcomm devices is typically done by enumerating which remote/DSP services a given app is permitted to open via FastRPC; the attacker establishes whether the specific vulnerable service is accessible from their privilege level.

2. Triggering the flaw class. The attacker drives the vulnerable code path by making repeated, and importantly concurrent, FastRPC invocations from multiple asynchronous threads. The race is between threads operating on the shared performance-counter data: the attacker needs the timing where one thread's operation on the shared counter object (including its release) overlaps another thread's use of that same reference. Success is a matter of racing invocations until the unprotected window is hit; failure is typically a crash (denial of service) rather than control. This is precisely why simultaneous asynchronous invocation is the core of the description.

3. Payload / execution method. At the vulnerability-class level, the goal after triggering the race is to turn stale-pointer access into influence over memory contents. For a use-after-free, that means arranging for the freed region to be reoccupied with attacker-influenced data such that the stale access operates on the attacker's data. The attacker then seeks a corruption outcome — a controlled write or a manipulated pointer/state field — that converts into code execution or privilege gain within the affected firmware component. The source data does not disclose the object layout, the operation performed on the stale reference, or a concrete primitive, so the specific technique cannot be stated.

4. Post-exploitation impact. Given C:H/I:H/A:H, a successful exploit is scored as allowing full loss of confidentiality, integrity, and availability within the affected component's security context. On Qualcomm platforms the compromised context is typically the firmware/DSP-side view of shared memory and service state; the practical effect is code execution at the privilege level of the vulnerable firmware component, which can be a step toward further device compromise. Whether this bridges to full device/OS compromise depends on the memory protections and isolation boundaries between the FastRPC service and the rest of the platform — details not present in the provided data.

Preconditions and difficulty. The attacker must be local, must be able to make FastRPC invocations, and must win a race. AC:L indicates the vendor did not consider precise timing or specialized conditions necessary, but multi-threaded race reliability on real hardware is the principal practical limiter. Two caveats apply: exploitation realism is unproven by public evidence (no PoC), and actual reachability may be constrained by FastRPC service access-control (which services are openable by which apps) — a platform policy not documented in the source data.

Impact Assessment

Business impact concentrates on the mobile/embedded gaming devices and development platforms built on Snapdragon G1 Gen 2, and on any product integrating that platform plus its firmware. Because this is firmware-level memory corruption reachable locally, the affected asset class is the device itself, not a server fleet.

  • Confidentiality (High): a corrupted shared-data structure in the FastRPC path can expose data passing through the subsystem to a lower-privileged attacker.
  • Integrity (High): the corruption primitive allows influencing memory contents, which affects the correctness and trust of the firmware/service state.
  • Availability (High): even a failed race attempt results in memory corruption, so denial-of-service (crash/reset of the affected subsystem) is an expected baseline outcome.
  • Privilege boundary: the exploit is an escalation across a privilege boundary local to the device (PR:L), which makes it a valuable link in a chained device-compromise scenario rather than a standalone remote threat.

Blast radius is limited by the local attack vector and by the fact that it does not traverse the network: it is not wormable and cannot be triggered against a device without an existing local foothold. However, for managed fleets of affected devices — kiosks, handheld gaming devices, OEM evaluation boards — a local escalation of this class is a meaningful post-foothold step. No affected version ranges were published in the source data, which means asset owners cannot currently map the flaw to specific firmware builds; inventory-driven exposure assessment is blocked until the vendor publishes a bulletin or the OEM releases a build identifier.

Detection Guidance

The detection content below targets the observable artifacts of this class of flaw: local FastRPC invocation activity, crashes/resets in the affected subsystem, and anomalous multi-threaded access to the RPC path. Because no exploit code or PoC is public, these rules are behavioral and class-oriented; they are starting points, not attack-specific signatures. No network callback, URL, or file path is known for this CVE, so network detection is necessarily heuristic based on the local-invocation nature of the flaw.

Sigma Rule

title: Suspicious FastRPC Invocation Burst with Subsystem Fault (CVE-2026-57554 class)
id: 8f3b2c1d-5755-4a6e-9c2f-00000057554
status: experimental
description: Detects repeated/concurrent FastRPC invocation activity followed by memory-fault or subsystem reset indicators on Qualcomm Snapdragon platforms. Behavioral approximation; no public PoC exists.
references:
  - https://nvd.nist.gov/vuln/detail/CVE-2026-57554
author: valtersit.com threat intelligence
date: 2026/01/01
logsource:
  product: linux
  service: kernel
detection:
  selection_fastrpc:
    - message|contains: 'fastrpc'
    - message|contains: 'adsprpc'
    - message|contains: 'cdsprpc'
  selection_fault:
    message|contains:
      - 'use-after-free'
      - 'general protection fault'
      - 'unable to handle kernel paging request'
      - 'subsys-restart'
      - 'dsp crash'
  timeframe: 5m
  condition: selection_fastrpc and selection_fault
falsepositives:
  - Legitimate DSP workload crashes during development
  - Vendor diagnostic tools invoking FastRPC services
level: medium
tags:
  - attack.t1068
  - cve.2026.57554

YARA Rule

rule CVE_2026_57554_FastRPC_MemoryCorruption_Artifacts
{
    meta:
        description = "Heuristic detection for artifacts associated with FastRPC shared performance counter memory corruption (CVE-2026-57554). No public PoC; rule matches class-level strings only."
        author = "valtersit.com threat intelligence"
        reference = "https://nvd.nist.gov/vuln/detail/CVE-2026-57554"
        cve = "CVE-2026-57554"
        date = "2026-01-01"
        severity = "high"

    strings:
        $fastrpc1 = "fastrpc" ascii nocase
        $fastrpc2 = "FASTRPC" ascii
        $perf1     = "perf_counter" ascii nocase
        $perf2     = "performance_counter" ascii nocase
        $adsprpc   = "adsprpc" ascii nocase
        $cdsprpc   = "cdsprpc" ascii nocase
        $invoke1   = "fastrpc_invoke" ascii nocase
        $shared    = "shared_perf" ascii nocase

    condition:
        uint16(0) == 0x5A4D or uint32(0) == 0x464C457F
        and (
            (any of ($fastrpc1, $fastrpc2, $adsprpc, $cdsprpc))
            and (any of ($perf1, $perf2, $shared, $invoke1))
        )
}

Suricata IDS Rule

alert ip any any -> any any (msg:"CVE-2026-57554 Potential FastRPC Performance Counter Memory Corruption Activity (local heuristic)"; flow:to_server; content:"fastrpc"; nocase; threshold: type both, track by_src, count 50, seconds 10; classtype:attempted-user; sid:202657554; rev:1; metadata:created_at 2026_01_01, cve CVE-2026-57554, confidence low; reference:url,nvd.nist.gov/vuln/detail/CVE-2026-57554;)

Note: this CVE is a local (AV:L) firmware flaw with no known network component. The rule above is a defensive placeholder for environments that tunnel or proxy FastRPC-adjacent traffic; it is not a reliable indicator on typical devices where FastRPC is entirely on-device.

Elastic Detection

process where host.os.type == "linux" and
  process.name in ("adsprpcd", "cdsprpcd", "fastrpc", "fastrpc_shell_0", "fastrpc_shell_3") and
  process.command_line : ("*invoke*", "*shared*", "*perf*")

Companion EQL for the crash indicator side, to correlate a fault event with FastRPC-adjacent process activity:

sequence by host.id with maxspan=5m
  [ process where host.os.type == "linux" and process.name : ("adsprpcd", "cdsprpcd", "fastrpc*") ]
  [ process where host.os.type == "linux" and process.name == "kernel" and
      process.args : ("*use-after-free*", "*general protection fault*", "*paging request*") ]

Splunk SPL Query

index=main (source="*syslog*" OR source="*dmesg*" OR source="*kernel*")
    ("fastrpc" OR "adsprpc" OR "cdsprpc" OR "FASTRPC")
    ("use-after-free" OR "general protection fault" OR "unable to handle kernel paging request"
     OR "subsys-restart" OR "dsp crash" OR "perf_counter" OR "performance counter")
| stats count min(_time) as firstTime max(_time) as lastTime
        values(host) as hosts
  by source, host
| where count >= 1
| convert ctime(firstTime) ctime(lastTime)
| sort - count

Broadening variant for environments with high DSP activity (error-scan driven), useful for retro-hunting:

index=main ("fastrpc" OR "adsprpc" OR "cdsprpc")
| rex field=_raw "(?<fault>use-after-free|general protection fault|subsys-restart|dsp crash)"
| where isnotnull(fault)
| timechart span=5m count by fault

Wazuh OSSEC Rule

<group name="cve_2026_57554,fastrpc,">
  <rule id="100554" level="10">
    <decoded_as>syslog</decoded_as>
    <match>fastrpc|adsprpc|cdsprpc|FASTRPC</match>
    <regex>use-after-free|general protection fault|unable to handle kernel paging request|subsys-restart|dsp crash</regex>
    <description>CVE-2026-57554 class: FastRPC-related memory fault / subsystem crash detected (possible local memory corruption exploitation attempt).</description>
    <mitre>
      <id>T1068</id>
    </mitre>
    <group>cve_2026_57554,attack.t1068,</group>
  </rule>

  <rule id="100555" level="12" frequency="5" timeframe="300">
    <if_matched_sid>100554</if_matched_sid>
    <same_source_ip />
    <description>CVE-2026-57554 class: repeated FastRPC memory faults within 5 minutes — elevated suspicion of race-condition exploitation attempts.</description>
    <mitre>
      <id>T1068</id>
    </mitre>
    <group>cve_2026_57554,attack.t1068,</group>
  </rule>
</group>

Mitigation & Patch Guidance

Patch confidence: 0/4 (UNPATCHED — no patch evidence from any tracked source). No vendor fix, workaround, or affected version range is available in the provided data. The following are defensive measures appropriate to this vulnerability class; they should be validated against vendor guidance when it becomes available.

Immediate mitigations (compensating controls):

  1. Constrain FastRPC reachability. The attack requires the ability to invoke FastRPC services. Restrict which applications/processes may open FastRPC services by tightening the platform's FastRPC access-control policy (SELinux policy and service ACLs) so that only trusted, signed system components can reach the remote service handling performance-counter data. This directly narrows PR:L exposure.
  2. Reduce local foothold opportunity. Because the vector is local and requires low privileges, apply standard hostile-app defenses: strict app allow-listing, disable side-loading where policy permits, and monitoring for applications that open unusual RPC services.
  3. Monitor for the crash signature. A failed race typically manifests as a subsystem fault. Use the detection content above to alert on repeated DSP/FastRPC faults; a burst of such faults on a single device is consistent with race-based exploitation attempts, not just ordinary instability.
  4. Segment the device class. For managed fleets, treat affected Snapdragon G1 Gen 2 gaming devices as elevated-risk endpoints; apply stricter monitoring and limit access to sensitive data on-device until a fix is available.

Patch / upgrade steps:

  1. Track Qualcomm's monthly Security Bulletin program for the DSP/FastRPC fix covering this CVE and identify the advisory listing the affected Snapdragon G1 Gen 2 platform.
  2. Register affected devices with the device OEM and chipset vendor for firmware-update notification; the fix, when released, will arrive as a firmware/board-support update rather than an OS package.
  3. Patch confidence is 0/4, meaning no patch reference is currently attributable. Do not assume a fix is present based on a device's OS update level — verify the firmware build and its security patch level against vendor advisory mapping.
  4. Where the vulnerability has been fixed downstream (i.e., in an OEM build), ensure the fix is applied to all affected units, including evaluation boards and kiosks that often fall behind the managed fleet update cycle.

Long-term hardening:

  1. Establish a firmware/SBOM inventory for Snapdragon-based devices so that chipset-level CVEs can be mapped to asset fleets — the present inability to map this CVE to affected versions is itself the key governance gap.
  2. Enforce least-privilege RPC design: any service that passes shared data across asynchronous threads should be reviewed for ownership/lifetime discipline; require that future platform integrations validate concurrency safety of shared-buffer RPC paths.
  3. Include chipset/firmware patch convergence in vulnerability-management SLAs, distinct from OS patch SLAs, since chipset fixes ship on a different cadence.
  4. Test the on-device behavior under concurrent RPC load (fuzzing the FastRPC invocation path with parallel callers) as part of device-acceptance testing.

References

Known threat actor campaigns or exploit-in-the-wild evidence: none available. There is no public PoC, no exploit code, and no Nuclei template for this CVE, and the EPSS score of 0.00110 indicates the model assesses exploitation as very unlikely in the near term. Exploitation has not been observed in the wild.

👁️ SHODAN DORKS (THREAT HUNTING)

What is this? Security researchers use these exact search queries on Shodan.io (a search engine for internet-connected devices) to find vulnerable systems globally. You can copy these dorks to see if your organization's IP space is exposed.

product:"Qualcomm"
title:"Qualcomm"
◈ SOURCE Analysis & enrichment: valtersit.com/cve/CVE-2026-57554/ · CVE data: NVD/MITRE (public domain) · © 2026 ValtersIT Hub
>_ SYSTEM_CORE: INTEL_ACTIVE
90% TEAM-CURATED // 10% AUTOMATED INGESTION.
ALL PATCH REFERENCES AND VERSION DATA SOURCED FROM NVD + OSV OFFICIAL REGISTRIES.