Mateusz Bronk
83655bc3ce
Follows up on 590856d (removing the Linux LE).
Removes all of whitelist management and LE facilities from the AESM and SDK.
Leaves only skeleton API stubs behind (for partial ABI compatibility).
!BREAKING CHANGES!
- Launch-related stub(`libsgx_launch.so`) and simulation (`libsgx_launch_sim.so`) libraries
removed from the SGX SDK.
- Removed AESM support for the deprecated Linux SGX out-of-tree (OOT) driver
(will no longer attempt an enclave load if OOT driver is detected)
- AESM APIs for launch control and whitelist management will now
return SGX_ERROR_FEATURE_NOT_SUPPORTED:
Affected APIs:
* get_launch_token(...)
* sgx_get_whitelist_size(...)
* sgx_get_whitelist(...)
* sgx_register_wl_cert_chain(...)
- Deprecated `sgx_uae_launch.h` SDK header
- Marked init token inputs `reserved` in the relevant loader APIs (no longer in use)
---------
Co-authored-by: Krzysztof1 Wisniewski <krzysztof1.wisniewski@intel.com>
Signed-off-by: Mateusz Bronk <mateusz.bronk@intel.com>
Intel(R) Software Guard Extensions Protected Code Loader (Intel(R) SGX PCL) for Linux* OS
Introduction
Intel(R) SGX PCL is intended to protect Intellectual Property (IP) within the code for Intel(R) SGX enclave applications running on the Linux* OS.
Problem: Intel(R) SGX provides integrity of code and confidentiality and integrity of data at run-time. However, it does NOT provide confidentiality of code offline as a binary file on disk. Adversaries can reverse engineer the binary enclave shared object.
Solution: The enclave shared object (.so) is encrypted at build time. It is decrypted at enclave load time.
Intel(R) SGX PCL provides:
-
sgx_encrypt: A tool to encrypt the shared object at build time.
See sources at sdk\encrypt_enclave.
-
libsgx_pcl.a: A library that is statically linked to the enclave and enables the decryption of the enclave at load time.
See sources at sdk\protected_code_loader.
-
SampleEnclavePCL: Sample code which demonstrates how the tool and lib need to be integrated into an existing enclave project.
Purpose of this code sample:
Enclave writers should compare SampleEnclave and SampleEnclavePCL. This demonstrates how the Intel(R) SGX PCL is to be integrated into the project of the enclave writer.
Build and test the Intel(R) SGX PCL with the sample code
- Install Intel(R) Software Guard Extensions (Intel(R) SGX) SDK for Linux* OS
- Enclave test key(two options): a. Install openssl first, then the project will generate a test key<Enclave_private_test.pem>/<Seal_private_test.pem> automatically when you build the project. b. Rename your test key(3072-bit RSA private key) to <Enclave_private_test.pem>/<Seal_private_test.pem> and put it under the / folder.
- Make sure your environment is set: $ source ${sgx-sdk-install-path}/environment
- Build the project with the prepared Makefile: a. Hardware Mode, Debug build: $ make b. Hardware Mode, Pre-release build: $ make SGX_PRERELEASE=1 SGX_DEBUG=0 c. Hardware Mode, Release build: $ make SGX_DEBUG=0 d. Simulation Mode, Debug build: $ make SGX_MODE=SIM e. Simulation Mode, Pre-release build: $ make SGX_MODE=SIM SGX_PRERELEASE=1 SGX_DEBUG=0 f. Simulation Mode, Release build: $ make SGX_MODE=SIM SGX_DEBUG=0
- Execute the binary directly: $ ./app
- Remember to "make clean" before switching build mode
Launch token initialization
If using libsgx-enclave-common or sgxpsw under version 2.4, an initialized variable launch_token needs to be passed as the 3rd parameter of API sgx_create_enclave. For example,
sgx_launch_token_t launch_token = {0}; sgx_create_enclave(ENCLAVE_FILENAME, SGX_DEBUG_FLAG, launch_token, NULL, &global_eid, NULL);
! NOTICE: support for launch_token has been removed in version 2.28 of the SDK. ! The 'launch_token' parameter is now reserved (a 'nullptr' may be passed in its place)