A GPU kernel can read sensitive data from another GPU kernel (even from another user or app) through an optimized GPU memory region called _local memory_ on various architectures.

An attacker with specialized hardware and physical access to an impacted device may be able to perform a voltage fault injection attack resulting in compromise of the ASP secure boot potentially leading to arbitrary code execution. 

A division-by-zero error on some AMD processors can potentially return speculative data resulting in loss of confidentiality. 

Certain size values in firmware binary headers could trigger out of bounds reads during signature validation, leading to denial of service or potentially limited leakage of information about out-of-bounds memory contents.

Failure to validate the integer operand in ASP (AMD Secure Processor) bootloader may allow an attacker to introduce an integer overflow in the L2 directory table in SPI flash resulting in a potential denial of service.

IBPB may not prevent return branch predictions from being specified by pre-IBPB branch targets leading to a potential information disclosure.

Insufficient memory cleanup in the AMD Secure Processor (ASP) Trusted Execution Environment (TEE) may allow an authenticated attacker with privileges to generate a valid signed TA and potentially poison the contents of the process memory with attacker controlled data resulting in a loss of confidentiality.

Insufficient verification of missing size check in 'LoadModule' may lead to an out-of-bounds write potentially allowing an attacker with privileges to gain code execution of the OS/kernel by loading a malicious TA.

Improper parameters handling in the AMD Secure Processor (ASP) kernel may allow a privileged attacker to elevate their privileges potentially leading to loss of integrity.

Improper parameters handling in AMD Secure Processor (ASP) drivers may allow a privileged attacker to elevate their privileges potentially leading to loss of integrity.

Execution unit scheduler contention may lead to a side channel vulnerability found on AMD CPU microarchitectures codenamed “Zen 1”, “Zen 2” and “Zen 3” that use simultaneous multithreading (SMT). By measuring the contention level on scheduler queues an attacker may potentially leak sensitive information.

Aliases in the branch predictor may cause some AMD processors to predict the wrong branch type potentially leading to information disclosure.

Mis-trained branch predictions for return instructions may allow arbitrary speculative code execution under certain microarchitecture-dependent conditions.

A potential vulnerability in some AMD processors using frequency scaling may allow an authenticated attacker to execute a timing attack to potentially enable information disclosure.

A malicious or compromised UApp or ABL may be used by an attacker to issue a malformed system call to the Stage 2 Bootloader potentially leading to corrupt memory and code execution.

Insufficient check of the process type in Trusted OS (TOS) may allow an attacker with privileges to enable a lesser privileged process to unmap memory owned by a higher privileged process resulting in a denial of service.

A malicious or compromised UApp or ABL could potentially change the value that the ASP uses for its reserved DRAM, to one outside of the fenced area, potentially leading to data exposure.

Failure to verify the protocol in SMM may allow an attacker to control the protocol and modify SPI flash resulting in a potential arbitrary code execution.

A malicious or compromised UApp or ABL may be used by an attacker to send a malformed system call to the bootloader, resulting in out-of-bounds memory accesses.

An attacker, who gained elevated privileges via some other vulnerability, may be able to read data from Boot ROM resulting in a loss of system integrity.