Lack of stack protection exploit mechanisms in ASP Secure OS Trusted Execution Environment (TEE) may allow a privileged attacker with access to AMD signing keys to c006Frrupt the return address, causing a stack-based buffer overrun, potentially leading to a denial of service.

Insufficient access controls in ASP kernel may allow a privileged attacker with access to AMD signing keys and the BIOS menu or UEFI shell to map DRAM regions in protected areas, potentially leading to a loss of platform integrity.

A malicious attacker in x86 can misconfigure the Trusted Memory Regions (TMRs), which may allow the attacker to set an arbitrary address range for the TMR, potentially leading to a loss of integrity and availability.

An out of bounds memory write when processing the AMD PSP1 Configuration Block (APCB) could allow an attacker with access the ability to modify the BIOS image, and the ability to sign the resulting image, to potentially modify the APCB block resulting in arbitrary code execution.

Improper validation in a model specific register (MSR) could allow a malicious program with ring0 access to modify SMM configuration while SMI lock is enabled, potentially leading to arbitrary code execution.

Improper restriction of write operations in SNP firmware could allow a malicious hypervisor to potentially overwrite a guest's memory or UMC seed resulting in loss of confidentiality and integrity.

Improper input validation in SEV-SNP could allow a malicious hypervisor to read or overwrite guest memory potentially leading to data leakage or data corruption.

Improper restriction of write operations in SNP firmware could allow a malicious hypervisor to overwrite a guest's UMC seed potentially allowing reading of memory from a decommissioned guest.

A potential weakness in AMD SPI protection features may allow a malicious attacker with Ring0 (kernel mode) access to bypass the native System Management Mode (SMM) ROM protections.

An out of bounds write vulnerability in the AMD Radeon™ user mode driver for DirectX® 11 could allow an attacker with access to a malformed shader to potentially achieve arbitrary code execution.