Arbitrary enclave memory overwrite vulnerability in ECall ecall_restore
Description
An arbitrary memory overwrite vulnerability in Asylo versions up to 0.6.0 allows an attacker to make a host call to UntrustedCall. UntrustedCall failed to validate the buffer range within sgx_params and allowed the host to return a pointer that was an address within the enclave memory. This allowed an attacker to read memory values from within the enclave.
AI Insight
LLM-synthesized narrative grounded in this CVE's description and references.
Asylo 0.6.0 and earlier has an arbitrary memory overwrite in UntrustedCall due to missing validation of output buffer range inside sgx_params.
Vulnerability
Asylo versions up to and including 0.6.0 contain an arbitrary memory overwrite vulnerability in the UntrustedCall primitive. The sgx_params->output pointer and sgx_params->output_size are not validated to ensure the output buffer lies outside the enclave memory region. This allows a malicious host to return a pointer that addresses enclave memory, leading to a write past the intended trusted boundary. [1]
Exploitation
An attacker with control over the untrusted host can craft a host call to UntrustedCall and provide an sgx_params->output pointer that points inside the enclave. The missing check allows the host to supply a pointer that overwrites enclave memory without triggering an abort. The attack requires the ability to make arbitrary host calls. [1]
Impact
Successful exploitation permits an untrusted host to write arbitrary data into enclave memory, potentially overwriting sensitive data, modifying code, or altering execution flow. This breaks the confidentiality and integrity guarantees of the enclave, leading to complete compromise of trusted execution. [1]
Mitigation
The fix [1] introduces a call to TrustedPrimitives::IsOutsideEnclave() on the output pointer before use. If the output is inside enclave memory, the enclave aborts with a security error. Users should update to Asylo version 0.6.1 or later. No workaround is available for earlier versions.
AI Insight generated on May 27, 2026. Synthesized from this CVE's description and the cited reference URLs; citations are validated against the source bundle.
Affected products
2- Google LLC/Asylov5Range: 0.6.0
Patches
0No patches discovered yet.
Vulnerability mechanics
AI mechanics synthesis has not run for this CVE yet.
References
1- github.com/google/asylo/commit/83036fd841d33baa7e039f842d131aa7881fdcc2mitrex_refsource_CONFIRM
News mentions
0No linked articles in our index yet.