README 7.2 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187
  1. Graphene Library OS with Intel SGX Support
  2. A Linux-compatible Library OS for Multi-Process Applications
  3. 1. WHAT IS GRAPHENE?
  4. Graphene Library OS is a project to provided lightweight guest OSes with
  5. support for Linux multi-process applications. Comparable to virtual
  6. machines, Graphene can run applications in an isolated environment, with
  7. virtualization benefits such as guest customization, platform independence
  8. and migration.
  9. Graphene Library OS supports native, unmodified Linux appliations upon
  10. any platform that Graphene Library OS has been ported to. Currently,
  11. Graphene Library OS is successfully ported to Linux, FreeBSD and Intel SGX
  12. enclaves upon Linux platforms.
  13. With the Intel SGX support, Graphene Library OS can secure a critical
  14. application in a hardware encrypted memory region. Graphene Library OS can
  15. protect applications against malicious system stack, with minimal porting
  16. effort.
  17. Graphene Library OS is a work published in Eurosys 2014. For more
  18. information. see the paper: Tsai, et al, "Cooperation and Security Isolation
  19. of Library OSes for Multi-Process Applications", Eurosys 2014.
  20. 2. HOW TO BUILD GRAPHENE?
  21. Graphene Library OS is consist of five parts:
  22. - Instrumented GNU Library C
  23. - LibOS (a shared library named "libsysdb.so")
  24. - PAL, a.k.a Platform Adaption Layer (a shared library named "libpal.so")
  25. - Reference monitor (a shared library named "libpal_sec.so")
  26. - Minor kernel customization and kernel modules
  27. Graphene Library OS currently only works on x86_64 architecture.
  28. Graphene Library OS is tested to be compiling and running on Ubuntu 12.04/14.04
  29. (both server and desktop version), along with Linux kernel 3.5/3.14.
  30. We recommand to build and install Graphene with the same host platform.
  31. Other distributions of 64-bit Linux can potentially, but the result is not
  32. guaranteed. If you find Graphene not working on other distributions, please
  33. contact us with a detailed bug report.
  34. The following packages are required for building Graphene: (can be installed
  35. with 'apt-get install')
  36. - build-essential
  37. - autoconf
  38. - gawk
  39. To build the system, simply run the following commands in the root of the
  40. source tree:
  41. make
  42. make install
  43. (Add Graphene kernel as a boot option by commands like "update-grub")
  44. (reboot and choose the Graphene kernel)
  45. Please note that the building process may pause before building the Linux
  46. kernel, because it requires you to provide a sensible configuration file
  47. (.config). The Graphene kernel requires the following options to be enabled
  48. in the configuration:
  49. - CONFIG_GRAPHENE=y
  50. - CONFIG_GRAPHENE_BULK_IPC=y
  51. - CONFIG_GRAPHENE_ISOLATE=y
  52. Each part of Graphene can be built separately in the subdirectories.
  53. To build Graphene library OS with debug symbols, run "make DEBUG=1" instead of
  54. "make".
  55. For more details about the building and installation, see the Graphene github
  56. Wiki page: <https://github.com/oscarlab/graphene/wiki>.
  57. 2-1. BUILD WITH INTEL SGX SUPPORT
  58. To build Graphene Library OS with Intel SGX support, run "make SGX=1" instead
  59. of "make". "DEBUG=1" can be used to build with debug symbols. Using "make SGX=1"
  60. in the test or regression directory will automatically generate the enclave
  61. signatures (in .sig files).
  62. A 3072-bit RSA private key (PEM format) is required for signing the enclaves.
  63. The default enclave key is placed in 'host/Linux-SGX/signer/enclave-key.pem',
  64. or the key can be specified through environment variable 'SGX_ENCLAVE_KEY'
  65. when building Graphene with Intel SGX support. If you don't have a private key,
  66. create it with the following command:
  67. openssl genrsa -3 -out enclave-key.pem 3072
  68. After signing the enclaves, users may ship the application files with the
  69. built Graphene Library OS, along with a SGX-specific manifest (.manifest.sgx
  70. files) and the signatures, to the Intel SGX-enanled hosts. The Intel SGX
  71. Linux SDK is required for running Graphene Library OS. Download and install
  72. from the official Intel github repositories:
  73. <https://github.com/01org/linux-sgx>
  74. <https://github.com/01org/linux-sgx-driver>
  75. A Linux driver must be installed before runing Graphene Library OS in enclaves.
  76. Simply run the following command to build the driver:
  77. cd Pal/src/host/Linux-SGX/sgx-driver
  78. make
  79. (The console will be prompted to ask for the path of Intel SGX driver code)
  80. sudo ./load.sh
  81. Finally generating the runtime enclave tokens by running "make SGX_RUN=1".
  82. 3. HOW TO RUN AN APPLICATION IN GRAPHENE?
  83. Graphene library OS uses PAL (libpal.so) as a loader to bootstrap an
  84. application in the library OS. To start Graphene, PAL (libpal.so) will have
  85. to be run as an executable, with the name of the program, and a "manifest
  86. file" given from the command line. Graphene provides three options for
  87. spcifying the programs and manifest files:
  88. option 1: (automatic manifest)
  89. [PATH TO Pal/src]/pal [PROGRAM] [ARGUMENTS]...
  90. (Manifest file: "[PROGRAM].manifest" or "manifest")
  91. option 2: (given manifest)
  92. [PATH TO Pal/src]/pal [MANIFEST] [ARGUMENTS]...
  93. option 3: (manifest as a script)
  94. [PATH TO MANIFEST]/[MANIFEST] [ARGUMENTS]...
  95. (Manifest must have "#![PATH_TO_PAL]/libpal.so" as the first line)
  96. Using "libpal.so" as loader to start Graphene will not attach the applications
  97. to the Graphene reference monitor. Tha applications will have better
  98. performance, but no strong security isolation. To attach the applications to
  99. the Graphene reference monitor, Graphene must be started with the PAL
  100. reference monitor loader (libpal_sec.so). Graphene provides three options for
  101. spcifying the programs and manifest files to the loader:
  102. option 4: (automatic manifest - with reference monitor)
  103. [PATH TO Pal/src]/pal_sec [PROGRAM] [ARGUMENTS]...
  104. (Manifest file: "[PROGRAM].manifest" or "manifest")
  105. option 5: (given manifest - with reference monitor)
  106. [PATH TO Pal/src]/pal_sec [MANIFEST] [ARGUMENTS]...
  107. option 6: (manifest as a script - with reference monitor)
  108. [PATH TO MANIFEST]/[MANIFEST] [ARGUMENTS]...
  109. (Manifest must have "#![PATH TO Pal/src]/pal_sec" as the first line)
  110. Although manifest files are optional for Graphene, running an application
  111. usually requires some minimal configuration in its manifest file. A
  112. sensible manifest file will include paths to the library OS and GNU
  113. library C, environment variables such as LD_LIBRARY_PATH, file systems to
  114. be mounted, and isolation rules to be enforced in the reference monitor.
  115. Here is an example of manifest files:
  116. loader.preload = file:LibOS/shim/src/libsysdb.so
  117. loader.env.LDL_LIBRAY_PATH = /lib
  118. fs.mount.glibc.type = chroot
  119. fs.mount.glibc.path = /lib
  120. fs.mount.glibc.uri = file:LibOS/build
  121. More examples can be found in the test directories (LibOS/shim/test). We have
  122. also tested several commercial applications such as GCC, Bash and Apache,
  123. and the manifest files that bootstrap them in Graphene are provided in the
  124. individual directories.
  125. For more information and the detail of the manifest syntax, see the Graphene
  126. github Wiki page: <https://github.com/oscarlab/graphene/wiki>.
  127. 4. HOW TO CONTACT THE MAINTAINER?
  128. For any questions or bug reports, please contact us:
  129. Chia-Che Tsai <chitsai@cs.stonybrook.edu>
  130. Donald Porter <porter@cs.stonybrook.edu>
  131. or post an issue on our github repository:
  132. <https://github.com/oscarlab/graphene/issues>