An issue was discovered in Espressif ESP-IDF 2.x, 3.0.x through 3.0.9, 3.1.x through 3.1.6, 3.2.x through 3.2.3, and 3.3.x through 3.3.1. An attacker who uses fault injection to physically disrupt the ESP32 CPU can bypass the Secure Boot digest verification at startup, and boot unverified code from flash. The fault injection attack does not disable the Flash Encryption feature, so if the ESP32 is configured with the recommended combination of Secure Boot and Flash Encryption, then the impact is minimized. If the ESP32 is configured without Flash Encryption then successful fault injection allows arbitrary code execution. To protect devices with Flash Encryption and Secure Boot enabled against this attack, a firmware change must be made to permanently enable Flash Encryption in the field if it is not already permanently enabled.

The EAP peer implementation in Espressif ESP-IDF 2.0.0 through 4.0.0 and ESP8266_NONOS_SDK 2.2.0 through 3.1.0 processes EAP Success messages before any EAP method completion or failure, which allows attackers in radio range to cause a denial of service (crash) via a crafted message.

The EAP peer implementation in Espressif ESP-IDF 2.0.0 through 4.0.0 and ESP8266_NONOS_SDK 2.2.0 through 3.1.0 allows the installation of a zero Pairwise Master Key (PMK) after the completion of any EAP authentication method, which allows attackers in radio range to replay, decrypt, or spoof frames via a rogue access point.

An issue was discovered in Espressif ESP-IDF 2.x and 3.x before 3.0.6 and 3.1.x before 3.1.1. Insufficient validation of input data in the 2nd stage bootloader allows a physically proximate attacker to bypass secure boot checks and execute arbitrary code, by crafting an application binary that overwrites a bootloader code segment in process_segment in components/bootloader_support/src/esp_image_format.c. The attack is effective when the flash encryption feature is not enabled, or if the attacker finds a different vulnerability that allows them to write this binary to flash memory.