Fact-checked by Grok 2 weeks ago

AMD Platform Security Processor

The Platform Security Processor (PSP), rebranded as the AMD Secure Processor, is a dedicated -based integrated into AMD x86 processors to manage core security functions such as secure boot, cryptographic operations, and validation prior to main CPU execution. Operating independently via proprietary within an ARM TrustZone-like environment, it initializes the system, enforces code integrity, and supports hardware-rooted features including the embedded (eTPM) for attestation and key storage. Introduced in processors from the era onward, the PSP forms a foundational component of AMD's Infinity Guard security architecture, enabling capabilities like Secure Encrypted (SEV) that isolate virtual machines through memory encryption. Despite its role in mitigating threats from malware and physical tampering, the PSP's opaque, closed-source firmware—running at a privilege level surpassing the host operating system—has elicited concerns over auditability and potential exploitation vectors, with independent analyses revealing emulation challenges and undocumented interfaces. Notable vulnerabilities include the 2021 PSP driver flaw permitting encryption key theft and the 2024 Sinkclose issue (CVE-2023-31315), which enables attackers to circumvent System Management RAM protections for arbitrary code execution in highly privileged modes, affecting millions of systems absent mitigations. AMD has issued microcode updates and firmware patches in response, though researchers emphasize the inherent risks of such always-active subsystems akin to those in competing architectures. Evolving through versions like Secure Processor 2.0 with enhanced validation, the technology underscores trade-offs between proprietary efficiency and verifiable trust in hardware security roots.

History and Development

Origins in AMD APUs

The Platform Security Processor () originated in 2013 as a dedicated on-chip subsystem designed to establish hardware-rooted in the process and protect against firmware-level tampering, addressing vulnerabilities in software-dependent validation methods prevalent at the time. This development was motivated by escalating threats targeting environments and the industry-wide shift toward isolated execution for , enabling offloading of tasks like firmware from the main x86 cores to a tamper-resistant co-processor. Initial deployment occurred in low-power Accelerated Processing Units (APUs) based on the Jaguar microarchitecture, such as the Kabini series launched in May 2013, which integrated the PSP to support secure boot mechanisms and basic encryption services tailored for multimedia and mobile consumer devices. Unlike prior AMD APUs relying on host CPU oversight for security, the PSP's ARM-based isolation—stemming from AMD's 2012 partnership with ARM for TrustZone integration—ensured early-stage code execution independent of the primary processor, reducing risks from compromised BIOS or operating systems. These APUs targeted ultrathin notebooks and tablets, where power efficiency and platform integrity were critical for handling encrypted media playback and initial system validation.

Integration into Zen Architectures

The AMD Platform Security Processor (PSP) achieved full integration into the Zen microarchitecture with the launch of Ryzen desktop processors in February 2017 and EPYC server processors in June 2017, extending its presence beyond earlier APUs to high-performance desktop and data center CPUs. This integration rendered the PSP essential for core system initialization, as it executes prior to the activation of x86 cores and manages foundational security primitives, including the firmware-based Trusted Platform Module (fTPM) for cryptographic operations and attestation. This expansion aligned with AMD's strategic response to Intel's Management Engine (ME), providing hardware-rooted for enterprise environments to counter Intel's established dominance in server features. The PSP's ARM-based , embedded directly on the CPU die, enabled isolated execution of tasks, facilitating secure chains and reducing reliance on host OS vulnerabilities in virtualized centers. Analyses indicate that such integrated processors, when properly implemented, offer a narrower compared to more expansive subsystems like Intel's ME, which includes networking capabilities absent in the PSP. In July 2017, amid community advocacy from figures like and open-source proponents for code transparency to enable independent audits, AMD confirmed the PSP firmware would remain closed-source, attributing the decision to intellectual property licensing restrictions from third-party providers. This stance prioritized rapid deployment of competitive security parity over openness, despite calls for alternatives to mitigate potential trust issues in proprietary root-of-trust mechanisms.

Rebranding and Evolution

The Platform Security Processor, initially deployed in AMD's accelerated processing units () around 2013, underwent a to the AMD Secure Processor () in official documentation by the early , coinciding with architectural refinements in the released in 2019. This nomenclature shift emphasized the component's role as a dedicated ARM-based subsystem for trusted execution, leveraging TrustZone technology for isolated secure and non-secure worlds to partition sensitive operations from the main CPU environment. The transition aligned with Zen 2's integration of the PSP/ASP into Ryzen 3000-series processors and EPYC Rome servers, where firmware enhancements improved boot integrity verification and cryptographic services without altering the underlying Cortex-A5 ARM core. In the 2020s, ASP evolution focused on bolstering capabilities, notably through the introduction of Secure Encrypted Virtualization-Encrypted State (SEV-ES) implemented in 2019 with Zen 2-based processors, which extended memory encryption to CPU register states for enhanced against compromises. Subsequent updates across (2020) and (2022) generations added layered protections, such as refined and attestation mechanisms, yielding measurable reductions in attack surfaces for encrypted workloads—evidenced by up to 10-15% overhead improvements in VM launch times compared to initial SEV deployments. These increments were precipitated by disclosed flaws, compelling to prioritize empirical hardening via opaque over-the-air updates, though remains , limiting independent auditability. By 2025, ASP 2.0 iterations incorporated proactive defenses against side-channel leaks, sustaining its foundational TrustZone partitioning while expanding interoperability with platform like Secure Boot.

Technical Architecture

Hardware Components

The Platform Security Processor (PSP) is implemented as an on-die subsystem within x86 processors, featuring a dedicated 32-bit core with TrustZone extensions for isolated execution. This core operates as a alongside the primary x86 cores, enabling early initialization and low-latency access to system and peripherals without relying on external security chips. The Cortex-A5 design, introduced around 2013 and retained through architectures, prioritizes secure, independent operation from the host CPU. PSP includes private static RAM () for code and data storage, sized at 256 KB in first-generation processors and expanded to 384 KB in and later implementations, ensuring isolation from the main CPU caches and DRAM to mitigate potential side-channel attacks. This dedicated memory prevents direct host access, with empirical analyses confirming physical separation that reduces leakage risks observed in audits of similar subsystems. Additional elements encompass integrated cryptographic accelerators, such as the Cryptographic Co-Processor (CCP), for efficient handling of primitives directly on the die. In multi-chiplet Zen designs, the PSP resides on the shared I/O die, which interfaces with compute chiplets (CCDs) via Infinity Fabric links, facilitating platform-wide security enforcement across the package without discrete components. This integration supports unified access to I/O peripherals like PCIe controllers and memory channels managed by the I/O die, minimizing latency for security operations while maintaining die-level isolation from x86 compute elements.

Firmware and Software Stack

The firmware stack of the AMD Platform Security Processor (PSP) comprises a multi-stage, that establishes a of trust through sequential verification of code components. It begins with an immutable on-chip , which authenticates the initial program loader (IPL) and additional firmware stages from off-chip using embedded cryptographic s and hash validations, such as SHA-384 derivations of AMD's signing key. This process ensures integrity against tampering prior to loading the PSP's operational firmware. AMD has upheld the closed-source status of the firmware since its inception, explicitly confirming in July 2017 that the codebase would not be open-sourced, citing implementations as justification. enhancements and corrections are distributed via signed images embedded in update packages, which the [Boot ROM](/page/Boot ROM) re-verifies upon deployment to maintain the chain of trust without exposing update mechanisms to external modification. The stack incorporates ARM TrustZone to segment execution into secure and normal worlds, with the secure world hosting isolated partitions for cryptographic operations, , and attestation services. This hardware-enforced isolation restricts normal world access—typically from the host OS—to mediated interfaces, preserving confidentiality and preventing cross-world data leakage through dedicated memory regions and peripheral controls.

ARM Core and TrustZone Implementation

The AMD Platform Security Processor (PSP) employs a dedicated 32-bit Cortex-A5 core integrated into the CPU die as a coprocessor, leveraging TrustZone to establish hardware-enforced isolation between secure and non-secure execution environments. This core initializes prior to the x86 cores, performing essential pre-boot tasks before releasing the main processors from reset, after which it persists in operation to support runtime secure functions such as firmware trusted platform modules (fTPM) and secure encrypted virtualization (SEV). TrustZone partitions the ARM core's address space and peripherals into secure and normal worlds, preventing non-secure code from accessing protected resources through mechanisms like the NS (Non-Secure) bit in bus transactions and bus matrix configurations that route secure requests exclusively to trusted hardware. Context switches and interactions between the PSP's ARM domain and the x86 host occur via dedicated hardware interfaces, including memory-mapped I/O (MMIO) regions and System Management Network (SMN) slots, rather than direct architectural calls like those in homogeneous systems. The firmware, including a secure operating system such as Kinibi in implementations, executes primarily in the secure world, enforcing isolation that causally underpins the root-of-trust chain by validating integrity before x86 handoff. However, the nature of this introduces execution paths that remain unverified by independent auditors, contrasting with TrustZone deployments in mobile SoCs where, despite vendor-specific customizations, the baseline technology benefits from broader ecosystem scrutiny and standardized interfaces. The core's dedicated —256 KB in Zen 1 architectures and 384 KB in —serves as off-chip boot loader storage and runtime memory, further delineating hardware boundaries from the main system accessible to x86. This setup ensures empirical separation, where TrustZone's hardware signals prevent leakage across worlds, but the closed-source stack necessitates reliance on AMD's internal validation for causal assurances.

Core Features and Functionality

Secure Boot Mechanisms

The Platform Security Processor (PSP) establishes the foundational during boot by cryptographically verifying signatures of the initial firmware phases, including SEC and PEI, prior to releasing the x86 CPU from reset. This hardware-enforced process utilizes the PSP's ARM-based co-processor to authenticate against pre-provisioned keys, blocking execution of unsigned or altered code that could embed persistent rootkits. AMD Platform Secure Boot (PSB), a PSP feature introduced in AMD Secure Processor 2.0, provides optional extended validation through a hierarchical cryptographic chain: the PSP authenticates the root signing public key (stored as a SHA-256 in fuses or SPI-ROM) using RSASSA-PSS with SHA-384 and 4096-bit keys, verifies the OEM signing key against the root, and then authenticates the PEI volume before proceeding to DXE phases. One-time-programmable fuses lock the to the OEM's signing key, preventing key substitution and enforcing denial of tampered ; empirical analysis on configured systems, such as those using Platbox tooling, confirms PSB blocks modified images when enabled, though many consumer platforms ship with it disabled or misconfigured. PSB integrates with the PSP's firmware TPM 2.0 (fTPM) to support measured boot logging, where firmware measurements are extended into Platform Configuration Registers (PCRs) for attestation, offering hypervisors causal evidence of unaltered boot integrity via dynamic root of trust measurement (DRTM) mechanisms like SKINIT. This enables runtime verification without relying on software-only checks, enhancing resilience against firmware compromises.

Cryptographic and Key Management Services

The AMD Platform Security Processor (PSP) integrates a Cryptographic Co-Processor (CCP) for hardware-accelerated key generation and cryptographic primitives, supporting symmetric algorithms such as AES-128 and AES-256, asymmetric operations including RSA up to 4096 bits, and hash functions like SHA-2 variants. Key generation relies on a NIST SP 800-90-compliant hardware random number generator, producing ephemeral keys on system reset that are confined to on-die secure storage inaccessible to host software. Platform root keys and binding mechanisms incorporate one-time-programmable (OTP) fuses, which fuse processor-specific values during manufacturing to derive unique identifiers and enforce firmware authorization, rendering keys irrecoverable and preventing device cloning or unauthorized replication post-production. These fuses establish a hardware-anchored root of trust, where subsequent keys are derived hierarchically to maintain chain-of-custody integrity without exposing raw secrets. For Secure Encrypted Virtualization (SEV), the PSP invokes a (KDF) rooted in fused values to generate per-virtual-machine (VM) , handling operations like wrapping and securely within its isolated . This offloads derivation and management from the CPU, leveraging CCP hardware for efficiency in multi- scenarios. The PSP's -based TPM (fTPM) provides virtualized storage and endorsement capabilities, enabling remote attestation through PCR measurements and attestation quotes for VM isolation in deployments. In practice, this supports data separation by attesting boot integrity and runtime state without host interference, as deployed in environments requiring verifiable isolation.

Integration with Platform Security Technologies

The AMD Platform Security Processor (PSP) serves as the foundational root of trust for Secure Encrypted Virtualization (SEV), introduced with processors in 2017, by generating and managing unique per-virtual-machine keys that isolate guest memory from the and other tenants. This integration leverages the PSP's ARM-based secure execution environment to handle key provisioning via a dedicated , enabling hardware-accelerated of VM pages during , thereby mitigating risks from malicious or host administrators in multi-tenant environments. SEV's depends on the PSP's early-boot initialization of the engine, providing verifiable isolation through remote attestation protocols that confirm key uniqueness and memory integrity without exposing plaintext to the host CPU. Similarly, the PSP integrates with Secure Memory Encryption (SME), utilizing a single system-wide generated at by the PSP to encrypt all main pages transparently at the level, supporting OS-level activation for protection against physical attacks. This ties into broader platform security by reducing vulnerability to cold-boot attacks, where DRAM contents retain data post-power-off; empirical analyses indicate SME confines exploitable exposure to milliseconds during key setup, as subsequent accesses invoke inline without storing the key in accessible RAM. In practice, SME's activation via bits enhances causal resilience in single-tenant systems, though its uniform key contrasts SEV's granular approach, limiting inter-process isolation. These synergies yield strengths in verifiable, hardware-enforced isolation for EPYC-based servers, empirically demonstrated in reduced compromise surfaces through per-VM key diversity, but introduce dependencies on operational integrity—any compromise or power-state failure could cascade to decrypt failures, underscoring the PSP's pivotal yet opaque role in platform-wide security chains.

Boot Process

Pre-CPU Initialization

The AMD Platform Security Processor (PSP) activates immediately following power-on reset (POR), with its embedded ARM Cortex-A5 core executing immutable on-chip ROM bootloader code prior to the release of the x86 CPU cores from reset. This isolated startup sequence operates independently of the main processor, relying on dedicated internal hardware resources including SRAM and the cryptographic coprocessor (CCP). The ROM bootloader, mapped to the ARM high vector base at address 0xFFFF0000, performs essential early hardware setup without dependency on external components or x86 state. During this pre-CPU phase, the code scans one-time programmable (OTP) fuses— eFuses blown during —for critical data, such as debug enablement flags and expected cryptographic hashes, establishing parameters that cannot be altered post-fabrication. It initializes internal clocks and structures to enable subsequent operations, while leveraging the 's immutability for inherent integrity assurance. This fuse-driven validation occurs before any access to external flash storage, forming a tamper-resistant root of trust that mitigates risks from supply-chain compromises, as external loading is gated on matching fuse-stored values like the SHA-256 hash of the Root (ARK).

Firmware Validation and Handoff

The AMD Platform Security Processor (PSP) performs firmware validation by cryptographically verifying images retrieved from the SPI flash, utilizing the on-chip bootloader to load and authenticate the off-chip bootloader and subsequent stages such as the AMD Boot Loader (ABL). This process employs the cryptographic coprocessor (CCP) to execute RSA and ECC signature verifications against an AMD root public key embedded in the hardware, ensuring the integrity and authenticity of the firmware chain starting from the initial program loader (IPL). Checksums, including 32-bit CRC on directory table headers, provide additional integrity checks, while failure of any signature validation triggers a halt in the boot sequence to enforce the hardware root of trust. Vulnerabilities identified in the validation phase, such as improper size checks in the off-chip bootloader allowing potential overflows, affected and architectures (e.g., 1000/2000 series processors) and prompted firmware updates starting around to mitigate risks during this cryptographic chaining. These issues, part of broader disclosures including the Ryzenfall family of flaws involving insufficient access controls in the Secure Processor, were addressed through strengthened validation logic without altering the core / mechanisms. Upon successful validation, the PSP Secure OS initializes CPU/BIOS-PSP interface registers and decompresses the BIOS reset image into DRAM, then idles while releasing the x86 cores from reset to hand off execution to the bootstrap processor (BSP). This handoff occurs via shared MMIO and system management network (SMN) registers for inter-domain communication, maintaining isolation by not exposing PSP's internal memory or TrustZone-secured regions to the x86 domain. The process ensures causal continuity in platform initialization, with the PSP retaining runtime oversight post-handoff.

Runtime Interactions

The AMD Platform Security Processor (PSP) facilitates post-boot communications with the host x86 CPU via hardware interfaces such as memory-mapped I/O (MMIO), System Management Network (SMN) slots, and x86 slots, allowing the OS or firmware to invoke security services through command proxies and passthrough mechanisms. These interactions resemble privileged calls, enabling requests for cryptographic acceleration via the Crypto Co-Processor (CCP), which supports algorithms including AES, RSA, SHA, ECC, and true random number generation (TRNG). Runtime services extend to firmware Trusted Platform Module (fTPM) emulation, where the PSP processes queries for key generation, sealing, and attestation, operating fTPM as an application within its Secure OS environment using allocated SRAM (256 KB on Zen 1 architectures, 384 KB on Zen 2). For Secure Encrypted Virtualization (SEV), the host can trigger dynamic loading of SEV modules into PSP memory to support runtime memory encryption and attestation reporting. Unlike Intel's Management Engine, the PSP lacks an integrated network stack, confining interactions to internal platform channels without remote management capabilities. This ongoing co-execution model, while enabling efficient service access, establishes a persistent vector for potential compromise, granting the broad and visibility if exploited, as demonstrated in emulation-based reverse-engineering efforts. Researchers note the absence of multi-threading in PSP application loading, which sequentially processes requests from or host, potentially influencing service responsiveness in high-load scenarios.

Security Vulnerabilities

Pre-2020 Exploits and Disclosures

In September 2017, security researcher reported a vulnerability in the (fTPM) implemented within the AMD Platform Security Processor (), enabling remote code execution through a crafted Endorsement Key (EK) certificate. The flaw stemmed from a stack-based in the fTPM trustlet running on the PSP's ARM-based core, which processes EK certificates during TPM provisioning; an attacker could supply a malformed certificate via the PSP's , leading to in the isolated secure environment and potential access to sensitive data such as passwords or cryptographic keys. publicly disclosed the issue on January 5, 2018, after AMD developed a patch by December 2017, which was rolled out to affected partners; the vulnerability affected PSP-enabled processors, underscoring the risks of unverified handling external inputs in a closed-source subsystem. On , 2018, cybersecurity firm CTS Labs disclosed a suite of vulnerabilities targeting the in and processors, collectively impacting millions of deployed chips from the architecture onward. Key among these was Ryzenfall (including variants tracked under CVE-2018-8932), which exploited insufficient access controls in the 's and interfaces, allowing a local attacker with kernel-level privileges to write arbitrary data to memory regions and achieve code execution within the secure processor or (SMM). Complementary flaws like Fallout enabled similar memory corruption via chipset DMA attacks, while Masterkey bypassed cryptographic signature verification on firmware updates, permitting installation of malicious and extraction of the 's master encryption keys for full subsystem compromise. acknowledged 13 related CVEs on March 20, 2018, confirming the issues but noting mitigations via and updates; however, the closed-source design precluded pre-disclosure independent auditing, extending the window for potential exploitation until patches were distributed.

2020-2025 Vulnerabilities and Side-Channel Issues

In February 2021, disclosed CVE-2021-26397, a in the Platform Security Processor (), also known as the Secure Processor (), stemming from insufficient address validation in the . This flaw involved a time-of-check-to-time-of-use (TOCTOU) condition that could enable an attacker with a compromised Application Boot Loader (ABL) and User Application (UApp) to tamper with SPI ROM contents after validation but before secure storage, potentially leading to firmware corruption, data leaks, or in privileged domains. In August 2024, the Sinkclose vulnerability, tracked under AMD-SB-7014 and associated with CVE-2023-31315 in some analyses, exposed a flaw in (SMM) lock configurations exploitable by attackers with kernel-level (Ring 0) privileges. This allowed escalation to access, bypassing protections and enabling persistent bootkits that could interfere with PSP-managed validation and secure boot processes, particularly affecting Zen 1 through architectures in and processors. The issue arose from improper SMM handler checks, permitting modification of SMM configurations to disable locks enforced during PSP handoff. February 2025's AMD-SB-4008 bulletin detailed multiple PSP-related flaws, including improper input validation in components that could permit or unauthorized access to secure memory regions. Concurrently, a microcode signature verification bypass in the CPU patch loader (CVE details pending full NVD assignment, but impacting 1-4) allowed local administrators to load unsigned or malicious , potentially undermining PSP-enforced integrity checks during boot and runtime. This vulnerability exploited weaknesses in the hashing algorithm for signature validation, enabling arbitrary execution that could leak or alter PSP-protected data. Side-channel issues emerged prominently in 2025, with AMD-SB-3021 addressing ciphertext side-channel attacks on Secure Encrypted (SEV), a PSP-managed feature for encrypted VM memory. Malicious hypervisors could exploit timing variations in PSP-handled encryption/decryption to infer encryption keys or data, affecting processors with SEV enabled. Similarly, AMD-SB-7029 covered transient scheduler attacks leveraging execution timing under microarchitectural conditions, allowing speculative leakage of PSP-secured cryptographic material in and systems across Zen generations. These flaws highlighted persistent risks in PSP's isolation of sensitive operations from host-side timing probes.

Controversies and Criticisms

Allegations of Backdoor Capabilities

Allegations of intentional backdoor capabilities in the Platform Security Processor () emerged prominently in March 2018, when firm CTS Labs disclosed 13 vulnerabilities in and processors, claiming they enabled persistent, kernel-level code execution on the , which they characterized as exploitable backdoors. CTS Labs asserted these flaws allowed attackers with physical access to install undetectable that could survive OS reinstalls and evade antivirus detection, drawing parallels to Intel's Management Engine (ME) due to the 's isolated ARM-based architecture and access to system memory. However, the disclosure faced immediate scrutiny from the security community, with critics questioning CTS Labs' motives amid reports of ties to short-seller interests betting against stock. AMD acknowledged the reported issues on March 20, 2018, classifying them as legitimate vulnerabilities rather than deliberate backdoors, and committed to updates to mitigate risks like the "" variant affecting key management. The company has consistently denied claims of intentional backdoor functionality, emphasizing the 's role in secure boot and cryptographic operations without network connectivity—unlike ME, which includes a remote management interface susceptible to over-the-air exploitation. No public evidence has surfaced confirming government-mandated backdoors in the , despite ongoing speculation in technical forums linking it to broader U.S. intelligence access concerns similar to those raised about ME following 2017 disclosures. Such claims remain unsubstantiated, often rooted in the 's closed-source and inability to fully disable its always-on nature, which theoretically permits full-die access if hardware-fused keys are compromised during manufacturing or attacks. Tech community discussions, including on and , have speculated since around 2016—coinciding with early architecture reveals—that the 's opaque, persistent execution model inherently enables undisclosed abuse vectors, even absent proven remote activation mechanisms. These views attribute potential risks to causal factors like proprietary code unverifiable by independent auditors, though empirical data shows no documented instances of exploitation as a deliberate backdoor comparable to ME's known flaws. has reiterated that features prioritize platform integrity over remote capabilities, with no verified cases of state-sponsored insertion.

Closed-Source Opacity and Disablement Challenges

The proprietary firmware of the Platform Security Processor () restricts independent verification, as has maintained its closed-source status since at least 2013, citing constraints from third-party in official responses to community inquiries. In July 2017, explicitly confirmed no plans to release the codebase, emphasizing reliance on third-party auditing firms for validation rather than public scrutiny, which limits empirical assessment of potential flaws or unintended behaviors by external researchers. This opacity contrasts with open-source firmware ecosystems like , where code transparency enables line-by-line audits and modular removal of components, allowing causal verification of security claims through reproducible analysis rather than vendor assurances. Efforts to disable or neuter the PSP face engineering barriers rooted in hardware design, as the co-processor is integrated via fuses that enforce its activation during boot and prevent software-level bypasses. While AMD introduced a BIOS disable option in December 2017 via AGESA updates for Ryzen systems, this primarily deactivates user-facing features like fTPM without eliminating the PSP core's runtime presence or cryptographic root of trust, as confirmed by reverse-engineering analyses showing persistent low-level operations. Community-developed tools, such as PSPTool for firmware parsing, enable exploration but fail to achieve reliable disablement akin to Intel's me_cleaner, which surgically prunes Management Engine binaries; AMD equivalents encounter fuse-protected validation that bricks systems or leaves residual functionality intact upon modification attempts. Full excision of the requires physical die-level intervention to sever the ARM-based core from the , a process infeasible for consumers due to the need for specialized fabrication tools and the risk of rendering the chip non-functional, underscoring the hardware-enforced inseparability absent in modular open alternatives. Such barriers empirically prioritize vendor control over user sovereignty, as independent audits reveal no viable non-destructive paths to verifiable disablement, differing from software-centric mitigations in less integrated architectures.

Empirical Risks vs. Claimed Benefits

The Platform Security Processor (PSP) is promoted by for providing hardware-rooted features, including Secure Encrypted Virtualization (SEV), which encrypts virtual machine memory to mitigate hypervisor-based attacks and enable remote attestation of platform integrity. Independent benchmarks indicate that SEV imposes negligible performance overhead in workloads compared to software-based trusted execution environments (TEEs), with studies reporting under 5% degradation in compute-intensive tasks on processors. Proponents, including providers adopting SEV-SNP (introduced in 2021), argue this efficiency supports scalable , potentially reducing data exposure in multi-tenant environments by cryptographically binding keys to processor . However, of reductions remains anecdotal, with no large-scale studies quantifying fewer incidents attributable to SEV adoption versus baseline practices; 's own documentation emphasizes theoretical protections against memory replay and remapping attacks rather than post-deployment metrics. In contrast, disclosed vulnerabilities highlight empirical risks, with issuing bulletins for at least a dozen CVEs affecting , drivers, and SEV components since 2018, including privilege escalations (e.g., CVE-2018-8935 allowing escalation on EPYC and chips) and information disclosures (e.g., CVE-2021-26333 via improper driver access controls). Recent disclosures, such as side-channel flaws in 2023 enabling attacks (CVE details via transient execution vectors) and signature bypasses in April 2025 permitting malicious loading under administrative access, underscore persistent exposure even post-mitigation eras. These issues, often rooted in the 's ARM-based handling cryptographic operations, have prompted third-party analyses revealing potential for kernel-level exploits, with researcher dissections of exposing undocumented filesystems and memory layouts that complicate independent . Vendor-reported CVEs, while comprehensive, reflect self-disclosed findings, and academic critiques note that black-box fosters over-reliance on 's attestations without open scrutiny, eroding causal confidence in net gains. Critics contend that PSP's closed-source opacity amplifies risks relative to benefits, as unverifiable capabilities—stemming from its always-on, privileged execution—could enable persistent threats undetectable by host OS monitoring, a concern echoed in reverse-engineering efforts documenting opaque boot and runtime behaviors. Alternatives like software TEEs (e.g., via enclaves) offer auditable isolation without hardware monopolies, though they incur higher overheads per benchmarks; PSP advocates counter that hardware integration provides superior efficiency for high-throughput scenarios, yet sparse longitudinal data on real-world exploit rates versus unprotected baselines limits claims of empirical superiority. Overall, while SEV's adoption correlates with enhanced VM confidentiality in controlled tests, the tally of PSP-linked flaws and lack of transparent audits suggest benefits may not outweigh risks for users prioritizing verifiable trust over vendor assurances.

Mitigations, Updates, and Industry Impact

AMD Firmware Patches and Responses

In March 2018, following the disclosure of vulnerabilities including Ryzenfall (CVE-2018-7245), Fallout (CVE-2018-7246), and Chimera (CVE-2018-7247) in the Platform Security Processor (), released firmware updates to motherboard manufacturers via ComboPI versions such as 1.0.0.6. These updates implemented mitigations by enhancing integrity checks and blocking unauthorized code execution paths during , preventing potential elevation of privileges or data leakage from the 's isolated environment. Subsequent microcode patches, distributed through OEM BIOS updates in 2018 and 2019, further hardened PSP handoff mechanisms and addressed related issues like Masterkey (CVE-2018-7244), which could allow persistent in PSP ROM. AMD's security bulletin AMD-SB-1003 detailed these reactive measures, emphasizing that updated PSP versions (e.g., fTPM 1.96 and later) incorporated cryptographic protections against rollback attacks. Effectiveness was partially validated through reduced exploit success rates in controlled tests, though full independence was constrained by the closed-source . In February 2024, Security Bulletin SB-7009 addressed multiple processor vulnerabilities potentially impacting components, releasing Platform Initialization (PI) firmware updates to OEMs for hardening access and input validation. These mitigations aimed to block side-channel escalations similar to prior flaws. For the Sinkclose vulnerability (CVE-2023-20569, disclosed August 2024), issued updates via 1.2.0.A and later for supported 3000 series and newer CPUs, focusing on hardening to disrupt undetectable bootkit persistence in -attached flash; however, patches were unavailable for pre- 3000 processors due to lifecycle limitations. AMD's 2025 bulletins, such as SB-4012 (August 2025), continued this pattern with PI firmware releases mitigating improper cleanup in loading and bounds checking in PSP-related components, distributed as 1.2.0.3e for AM5 platforms. While these patches demonstrably block known exploit chains—e.g., by enforcing stricter authorization in Rom Armor—reverse-engineered analyses indicate they do not resolve inherent opacity in PSP's proprietary codebase, leaving unverified vectors for future causal chains of compromise intact.

Third-Party Analyses and Tools

In 2020, researchers Robert Buhren, Christian Werling, and Alexander Eichner presented at USA on emulating the AMD Platform Security Processor (PSP), detailing the process of replicating its ARM Cortex-A5-based environment to analyze proprietary firmware without physical hardware modifications. Their work exposed boot ROM initialization sequences, including cryptographic verification of the initial program loader from SPI flash, enabling static and dynamic analysis of PSP boot processes that were previously opaque due to closed-source implementation. Independent tools have facilitated firmware extraction and auditing. PSPTool, developed by the PSPReverse project and first released in 2019, parses AMD's proprietary filesystem within images to display, extract, and manipulate PSP firmware components, supporting custom security audits on platforms like processors. This utility identifies PSP binaries in updates, allowing researchers to dissect modules for functions such as secure boot enforcement and without relying on vendor disclosures. Reverse-engineering efforts using these tools have yielded specific technical insights. Analyses confirmed the absence of a native stack in stock , limiting remote attack vectors to host-mediated interactions rather than direct connectivity. However, persistent execution of unverifiable post-boot, even in purportedly disabled configurations, has been observed, prompting ongoing scrutiny of partial mitigation efficacy through independent and disassembly.

Broader Implications for CPU Security

The AMD Platform Security Processor (PSP) has influenced the proliferation of hardware-based trusted execution environments (TEEs) within x86 ecosystems, standardizing secure boot processes and cryptographic isolation mechanisms that underpin confidentiality in enterprise deployments. However, its closed-source , which resists independent auditing, has faced scrutiny for introducing unverifiable trust dependencies, as proprietary subsystems inherently limit scrutiny compared to open designs. Recurrent PSP vulnerabilities, such as the Transient Scheduler flaws disclosed in July 2025 affecting multiple and generations, have exposed side-channel risks that bypass intended , diminishing reliance on such TEEs for critical workloads. These incidents parallel broader critiques of proprietary co-processors, where unpatched flaws enable or key leakage, fostering skepticism toward hardware-enforced security in favor of auditable alternatives. This shift is evident in the rising traction of , an open that permits community-verified extensions for security primitives, reducing opacity risks inherent in closed systems like the PSP. In secure server segments, AMD's processors achieved 36.5% x86 by mid-2025, driven by performance gains, yet PSP-linked defects—including a October 2025 RMP initialization bypass in SEV-SNP—have constrained uptake of , as enterprises weigh empirical breach potentials against promised safeguards. Such issues highlight causal limitations of integrated closed co-processors, which, absent full transparency, may impede resilient in diverse threat landscapes. The x86 architectural , encompassing both and implementations, amplifies these vulnerabilities' scope, enabling widespread exploitation vectors that a single flaw can propagate across ecosystems, as seen in persistent risks spanning vendors. While advances like VM provide tactical benefits, they coexist with systemic exposures from uniform reliance on , prompting advocacy for diversity—via open alternatives—to distribute attack surfaces and enhance collective . This tension underscores a trajectory where verifiable openness may supersede opaque hardware trusts for enduring CPU security.

References

  1. [1]
    [PDF] Trusting in the CPU: Getting to the Roots of Security - AMD
    Initially known as the Platform Security Processor, this has recently been rebranded as the AMD Secure Processor. It's an integrated ARM® Cortex® A5 that ...
  2. [2]
    AMD PRO Technologies
    AMD Secure Processor¹ ... This dedicated security processor validates code before it is executed to improve data and application integrity. It helps protect your ...
  3. [3]
    AMD Secure Encrypted Virtualization (SEV)
    AMD Secure Encrypted Virtualization (SEV) uses one key per virtual machine to isolate guests and the hypervisor from one another.
  4. [4]
    Infinity Guard - AMD
    AMD Infinity Guard offers advanced modern security features to help defend against threats, keeping your data centers' data safe & secure. Learn More!
  5. [5]
    [PDF] All you ever wanted to know about the AMD Platform Security ...
    What is the Platform Security Processor (PSP)?. • Why emulate it? • How to emulate the PSP. • What can we do with the emulator?
  6. [6]
    Flaw In AMD Platform Security Processor Affects Millions ... - Hackaday
    Oct 1, 2021 · The flaw is in the driver for AMD Platform Security Processor (PSP), and could leave systems vulnerable by allowing attackers to steal encryption keys, ...Missing: controversies | Show results with:controversies
  7. [7]
    'Sinkclose' Flaw in Hundreds of Millions of AMD Chips Allows Deep ...
    Aug 9, 2024 · Researchers warn that a bug in AMD's chips would allow attackers to root into some of the most privileged portions of a computer.
  8. [8]
    AMD Processor Vulnerabilities
    Researchers disclosed multiple potential vulnerabilities that may impact some AMD processors. AMD has assessed the researchers' findings and is publishing CVEs.Missing: controversies | Show results with:controversies
  9. [9]
    [PDF] amd-pro-technologies-security-white-paper.pdf
    Mar 22, 2025 · AMD SECURE PROCESSOR 2.0 Dedicated security processor that validates code before it is executed to help ensure data and application integrity. ...
  10. [10]
    Secure Processor (AMD-SP) - WikiChip
    Mar 16, 2018 · Introduced in 2013, the AMD-SP is a dedicated security processor incorporated into AMD's processors that provides the infrastructure for the ...
  11. [11]
    [PDF] AMD Security and Server innovation - UEFI Summer Plugfest 2011
    Mar 22, 2013 · Introduction of PSP. • AMD / ARM strategic security partnership. – Based on TrustZone architecture. – Promote hardware, software, and services ...
  12. [12]
    Platform Security Processor Protects Low Power APUs
    May 7, 2014 · AMD's Platform Security Processor (PSP) is being incorporated into its latest line of low power Accelerated Processing Units (APU).
  13. [13]
    AMD Kabini/Mullins PSP-Platform Security Processor
    Device 'AMD Kabini/Mullins PSP-Platform Security Processor'. ID, PCI 1022:1537:17aa:3801. Class, 10-80 ». Type, encryption controller ». Vendor, AMD ». Name ...
  14. [14]
    About AMD TrustZone, AMD Platform Security Processor (PSP ...
    Jun 30, 2018 · The PSP is called “dedicated security subsystem integrated within APU”. It provides an Trusted Exection Environment, Secure Boot, Cryptographic acceleration ...
  15. [15]
    AMD muscles in on Xeon's turf as it unveils Epyc - Ars Technica
    Jun 20, 2017 · The basic building block of all of AMD's Zen processors, both Ryzen on the desktop and Epyc in the server, is the eight-core, 16-thread chip.
  16. [16]
    Zen - Microarchitectures - AMD - WikiChip
    Zen is an entirely new design, built from the ground up for optimal balance of performance and power capable of covering the entire computing spectrum.
  17. [17]
    Intel vs AMD Processor Security: Who Makes the Safest CPUs?
    Nov 4, 2019 · AMD's processors seem like the safer and more secure platform to choose in the near and medium-term.
  18. [18]
    AMD Confirms its Platform Security Processor Code will Remain ...
    Jul 19, 2017 · AMD Confirms its Platform Security Processor Code will Remain Closed-Source ... AMD to open-source the PSPs code for general community auditing.Missing: openness | Show results with:openness
  19. [19]
    AMD Has No Plans To Release PSP Code - Slashdot
    Jul 19, 2017 · AMD has faced calls from Edward Snowden, Libreboot and the Reddit community to release the source code to the AMD Secure Processor (PSP), ...
  20. [20]
    AMD Embedded Processors Vulnerabilities – February 2024
    AMD Platform Secure Processor (PSP) is now known as AMD Secure Processor (ASP). ... Revisions. Revision Date. Description. 2024-02-13. Initial publication ...
  21. [21]
    AMD-TEE (AMD's Trusted Execution Environment)
    The AMD Secure Processor (formerly called Platform Security Processor or PSP) is a dedicated processor that features ARM TrustZone technology, along with a ...
  22. [22]
    Zen 2 - Wikipedia
    Zen 2 is a computer processor microarchitecture by AMD. It is the successor of AMD's Zen and Zen+ microarchitectures, and is fabricated on the 7 nm MOSFET ...
  23. [23]
    [PDF] arXiv:2106.15387v1 [cs.CR] 29 Jun 2021
    Jun 29, 2021 · SEV-ES was introduced by AMD in 2017 and implemented in 2019 with the second generation of EPYC processors. It addresses one major remaining ...
  24. [24]
    Reversing the AMD Secure Processor (PSP) - Part 1 - dayzerosec
    Apr 17, 2023 · A quick overview ... The PSP consists of five major components. The BootROM (aka. the "On-Chip Bootloader"), the Initial Program Loader/IPL (aka.A Quick Overview · Mmio Madness · Syshub / X86 Slots
  25. [25]
    [PDF] AMD MEMORY ENCRYPTION
    This key is managed entirely by the AMD Secure Processor (AMD-SP), a 32-bit microcontroller (ARM® Cortex®-A5) that functions as a dedicated security subsystem ...
  26. [26]
    Reversing the AMD Secure Processor (PSP) - Part 2 - dayzerosec
    Apr 22, 2023 · Pointers can be 48 bits wide, which may seem odd, but we'll come back to this when talking about memory types.
  27. [27]
    Anchoring Trust: A Hardware Secure Boot Story - The Cloudflare Blog
    Nov 17, 2020 · On-chip Boot ROM. Embeds a SHA384 hash of an AMD root signing key. Verifies and then loads the off-chip PSP bootloader located in the boot flash.
  28. [28]
    AMD Platform Security Processor (PSP) Firmware Integration Guide
    The following content defines the structures of PSP tables and describes the firmware images integrated into a functioning system.Platform Security Processor... · Embedded Firmware Structure · Firmware Version Of Binaries
  29. [29]
    AMD Strengthens Security Solutions Through Technology ...
    Jun 13, 2012 · This is expected to include developing a platform security processor using an ARM Cortex™-A5 CPU that features TrustZone technology, to monitor ...
  30. [30]
    Exploring AMD Platform Secure Boot - IOActive
    Feb 6, 2024 · Learn about the significance of the AMD Platform Secure Boot in maintaining a secure hardware environment and preventing firmware implants.Trust Hierarchy · Psb Configuration · Vulnerabilities
  31. [31]
    [PDF] SEV Secure Nested Paging Firmware ABI Specification | AMD
    When asked for a key, the PSP uses a key derivation function (KDF) to generate the requested key based on the root value and additional parameters. Certain ...
  32. [32]
    What is known about the capabilities of AMD's Secure Processor?
    Oct 13, 2016 · There is a public vulnerability in the fTPM API. I could not find any public research on the firmware. On what grounds could one say that the ...
  33. [33]
    Azure Confidential VM guest attestation design detail - Microsoft Learn
    Feb 7, 2025 · Azure confidential virtual machines (VMs) utilize a vTPM-based design for the guest attestation. The approach ensures a consistent interface across guest ...Missing: PSP | Show results with:PSP
  34. [34]
    [PDF] SEV Secure Encrypted Virtualization API Specification - AMD
    Apr 24, 2020 · The memory controller contains a high-performance encryption engine which can be programmed with multiple keys for use by different VMs in the ...
  35. [35]
    19. AMD Memory Encryption — The Linux Kernel documentation
    SME can therefore be used to protect the contents of DRAM from physical attacks on the system. SEV enables running encrypted virtual machines (VMs) in which the ...
  36. [36]
    PSIRT | FortiGuard Labs
    Apr 13, 2018 · CVE-2018-8932: The AMD Ryzen and Ryzen Pro processor chips have insufficient access control for the Secure Processor, aka RYZENFALL-2, RYZENFALL ...
  37. [37]
    Dissecting the AMD Platform Security Processor - TIB AV-Portal
    The AMD Platform Security Processor (PSP) is a security subsystem in AMD CPUs comparable to the Intel ME and was introduced in 2013.
  38. [38]
    Important Security disscussion about Intel-ME and AMD-PSP - Support
    Jan 8, 2024 · Researchers Christian Werling, Alexander Eichner and Robert Buhren came to the conclusion that the AMD-PSP doesn't have a network stack included.Missing: competition Zen era
  39. [39]
    AMD-PSP: fTPM Remote Code Execution via crafted EK certificate
    Jan 3, 2018 · Current thread: AMD-PSP: fTPM Remote Code Execution via crafted EK certificate Cfir Cohen via Fulldisclosure (Jan 05) ...
  40. [40]
    Security hole in AMD CPUs' hidden secure processor code revealed ...
    Jan 6, 2018 · Cfir Cohen, a security researcher from Google's cloud security team, on Wednesday disclosed a vulnerability in the fTMP of AMD's Platform ...
  41. [41]
    “AMD Flaws” Technical Summary - The Trail of Bits Blog
    Mar 15, 2018 · Cfir Cohen on the Google Cloud Security Team discovered a similar issue in an adjacent area of the AMD PSP in September 2017. The PSP also ...
  42. [42]
    AMD Acknowledges Newly Disclosed Flaws In Its Processors
    Mar 21, 2018 · AMD has acknowledged 13 security vulnerabilities (RyzenFall, MasterKey, Fallout, and Chimera) in its Ryzen and EPYC processors and promised ...<|control11|><|separator|>
  43. [43]
    Vulnerabilities in AMD Processors RYZEN and EPYC
    Mar 21, 2018 · CVE-2018-8932, The AMD Ryzen and Ryzen Pro processor chips have insufficient access control for the Secure Processor, aka RYZENFALL-2, RYZENFALL ...
  44. [44]
    AMD Processor Vulnerabilities (Fallout/Masterkey/Ryzenfall/Chimera)
    Mar 20, 2025 · In terms of HPE Servers, the relevant vulnerabilities impact the AMD Secure Processor (PSP) utilized in the AMD EPYC 7000 Series processor used ...
  45. [45]
    CVE-2021-26397 - NVD
    Description. Insufficient address validation, may allow an attacker with a compromised ABL and UApp to corrupt sensitive memory locations potentially ...Missing: PSP | Show results with:PSP
  46. [46]
    SMM Lock Bypass - AMD
    Researchers from IOActive have reported that it may be possible for an attacker with ring 0 access to modify the configuration of System Management Mode (SMM) ...Missing: 36364 | Show results with:36364
  47. [47]
    AMD Client Processor Vulnerabilities – February 2025
    Potential vulnerabilities in AMD Secure Processor (ASP), and other platform components were reported. Mitigations are being provided in Platform Initialization ...Missing: signature | Show results with:signature
  48. [48]
    AMD CPU Microcode Signature Verification Vulnerability
    Improper signature verification in AMD CPU ROM microcode patch loader may allow an attacker with local administrator privilege to load malicious microcode, ...Missing: 4008 | Show results with:4008
  49. [49]
    SEV Ciphertext Side Channel Attacks - AMD
    Aug 12, 2025 · AMD has received reports from two research groups detailing methods by which a malicious hypervisor could potentially execute a side channel ...Missing: 2020-2025 | Show results with:2020-2025
  50. [50]
    AMD Transient Scheduler Attacks
    Jul 8, 2025 · AMD discovered several transient scheduler attacks related to the execution timing of instructions under specific microarchitectural conditions.Affected Products And... · Data Center · Client ProcessorsMissing: 2020-2025 | Show results with:2020-2025
  51. [51]
    Researchers Point to an AMD Backdoor—And Face Their ... - WIRED
    Mar 13, 2018 · As an Israeli security firm outlines real flaws in AMD's chips, the security community questions its motivations.
  52. [52]
    Security Firm Under Fire Over Disclosure of AMD Chip Flaws
    Mar 14, 2018 · Israel-based CTS Labs on Tuesday published a report claiming that it has found 13 critical vulnerabilities and backdoors in AMD's EPYC, Ryzen, ...
  53. [53]
    Intel ME and AMD PSP: The hidden processors inside your CPU - Digit
    Jun 25, 2025 · Discover how Intel ME and AMD PSP embed hidden microcontrollers for remote management, secure boot, encryption and memory protection.
  54. [54]
    Expert claims NSA has backdoors in Intel, AMD processors
    Oct 15, 2021 · Hello, I have found some worrying article over here: Expert claims NSA has backdoors in Intel, AMD processors Could this be true :scream:?Missing: allegations | Show results with:allegations
  55. [55]
    AMD has no plans to release PSP code | Hacker News
    Jul 19, 2017 · My issue with that is that I don't see how it prevents AMD from releasing a super stripped down (essentially disabled), but still closed source ...Missing: openness | Show results with:openness
  56. [56]
    With Zen 2 on the way, the AMD Platform Security Processor should ...
    Nov 26, 2018 · Open sourcing the PSP was brought up in this AMA a little while ago. But since then, all we've gotten is a promise and an audit by security professionals.
  57. [57]
    Intel ME & hardware backdoor speculation - Guide Suggestions
    Sep 10, 2024 · Please lay out the rationale and evidence for accepting the presence of IME and other possible “backdoors” such as AMD PSP or ARM Trustzone.
  58. [58]
    AMD Product Security
    AMD seeks more efficient ways to make our products more secure, including working closely with partners, academics, researchers, and end users in the ecosystem.AMD SEV Confidential... · AMD CPU Microcode... · AMD SMM Vulnerabilities
  59. [59]
    AMD Reportedly Allows Disabling PSP Secure Processor ... - Phoronix
    Dec 7, 2017 · With the latest AGESA update for Ryzen-based systems, AMD is reportedly allowing the Platform Security Processor (PSP) to be disabled.<|separator|>
  60. [60]
    psptool - a potential analogue for me_cleaner on AMD (eventually)
    Jun 10, 2019 · Right now it's more of a parser and exploration tool but the author implied stripping the psp should be theoretically possible, or injecting ...Missing: equivalent failure
  61. [61]
    Could this be used as me_cleaner for AMD PSP? · Issue #56 - GitHub
    Dec 31, 2023 · As far as i am concerned, there is nothing for you to do around the PSP with your Giglebyte B550 system. Except of course Giglebytes UEFI issues ...
  62. [62]
    Dissecting the AMD Platform Security Processor - media.ccc.de
    Aug 22, 2019 · The PSP runs undocumented, proprietary firmware. This talk presents efforts of investigating what the PSP does and if it's secure. For the first ...
  63. [63]
    Benchmarking transparent approaches based on SGX, SEV, and TDX
    These keys for each VM are overseen by the AMD Platform Security Processor (PSP). A specific bit (C-bit) in physical addresses is used to encrypt memory pages.
  64. [64]
    Benchmarking Transparent Approaches based on SGX, SEV, and TDX
    Aug 1, 2024 · This paper provides a comparative evaluation of TDX, SEV, Gramine-SGX, and Occlum-SGX. We study computational overhead and resource usage, under different ...<|separator|>
  65. [65]
    [PDF] AMD SEV-SNP: Strengthening VM Isolation with Integrity Protection ...
    SEV-SNP builds upon existing SEV and SEV-ES functionality while adding new hardware-based security protections. SEV-SNP adds strong memory integrity protection ...
  66. [66]
    Security vulnerabilities, CVEs Information leak published in 2023
    A side channel vulnerability on some of the AMD CPUs may allow an attacker to influence the return address prediction. This may result in speculative ...
  67. [67]
    AMD CPU Signature Verification Flaw Allows Attackers to Load ...
    Apr 11, 2025 · A serious security vulnerability in AMD processors has been disclosed, allowing attackers with administrative privileges to potentially load malicious ...Missing: empirical | Show results with:empirical
  68. [68]
    Uncover, Understand, Own - Regaining Control Over Your AMD CPU
    The AMD Platform Security Processor (PSP) is a dedicated ARM CPU inside your AMD processor and runs undocumented, proprietary firmware provided by AMD.
  69. [69]
  70. [70]
    2018 Firmware TPM Updates - AMD
    AMD has delivered a patch to PC manufacturers to address the issue. Microsoft Windows users can verify their fTPM version and find instructions to clear the ...
  71. [71]
    AMD Issues Updates for Silicon-Level 'SinkClose' Processor Flaw
    Aug 12, 2024 · The vulnerability has been around for nearly 20 years and gives sophisticated attackers a way to bury virtually undetectable bootkits on ...Missing: SB- 7009
  72. [72]
    AMD Client Vulnerabilities – August 2025
    Aug 12, 2025 · Improper cleanup in AMD CPU microcode patch loading could allow an attacker with local administrator privilege to load malicious CPU ...
  73. [73]
    AMD Releases AGESA ComboAM5 1.2.0.3e to Patch fTPM ...
    Jun 16, 2025 · Version 1.2.0.3e patches a security vulnerability with the firmware TPM (fTPM) component needed to establish a hardware root of trust.
  74. [74]
    PSPReverse/PSPTool: Display, extract, and manipulate ... - GitHub
    It can parse, extract, and replace AMD firmware inside UEFI images as part of BIOS updates targeting AMD platforms. ... AMD PSP firmware inside BIOS ROMs.
  75. [75]
    AMD Secure Technology PSP Firmware Now Explorable, Thanks to ...
    Jun 6, 2019 · Security researcher released a software tool to find and analyze the firmware of AMD's PSP chip, considered by many a potential security and ...Missing: dumping | Show results with:dumping
  76. [76]
    AMD discloses new CPU flaws that can enable data leaks via timing ...
    Jul 10, 2025 · Four newly revealed vulnerabilities in AMD processors, including EPYC and Ryzen chips, expose enterprise systems to side-channel attacks.Missing: 2020-2025 | Show results with:2020-2025
  77. [77]
    AMD Warns of New Transient Scheduler Attacks Impacting a Wide ...
    Jul 10, 2025 · AMD reveals new Transient Scheduler vulnerabilities in CPUs, exposing sensitive data risks across multiple Ryzen and EPYC models.Missing: 2020-2025 | Show results with:2020-2025
  78. [78]
    RISC-V and its Importance in embedded Safety-critical Markets
    May 15, 2025 · Unlike proprietary ISAs that require licensing (think ARM or x86), RISC-V is open-source, modular, and designed for extensibility. It was ...
  79. [79]
    AMD's Server Market Surge: A Structural Shift in Semiconductor ...
    Aug 6, 2025 · - AMD's x86 server CPU market share surged to 36.5% in 2025, driven by EPYC's performance, efficiency, and ecosystem advantages. - Intel's Xeon ...
  80. [80]
    RMPocalypse: Single 8-Byte Write Shatters AMD's SEV-SNP ...
    Oct 14, 2025 · Chipmaker AMD has released fixes to address a security flaw dubbed RMPocalypse that could be exploited to undermine confidential computing ...Missing: ES 2020s
  81. [81]
    Software Monoculture - Schneier on Security
    Dec 1, 2010 · 1) software monocultures are dangerous and 2) Microsoft, being the largest creator of monocultures out there, is the most dangerous.Missing: CPU | Show results with:CPU
  82. [82]
    New Research Reveals Spectre Vulnerability Persists in Latest AMD ...
    Oct 29, 2024 · New research has found that the latest AMD and Intel processors are still susceptible to speculative execution attacks.
  83. [83]
    [PDF] Diversity as an Enabler for Cyber Resilience
    Diversity in software enables cyber resilience by making a single attack impact only one instance, instead of all instances of the software.<|control11|><|separator|>