Notes - MCS
Secure Execution Environments
Notes - MCS
Secure Execution Environments
  • Secure Execution Environments
  • Introduction
    • Trusted Computing Base (TCB)
    • TEE (Trusted Execution Environment)
    • Can you trust the operating system?
  • Security in Operating Systems
    • Operating system
    • Virtual machines and hypervisors
    • Computational model
    • Access control
    • Protection with capabilities
    • Unix file protection ACLs
    • Windows NTFS file protection
    • Unix file protection ACLs
    • Privilege elevation
    • Privilege reduction
    • Linux login
  • Virtualization on Intel Processors
    • Modes of Operation
    • Virtual memory
    • How to put assembly instructions inside C code
    • A more elaborate example
    • Useful assembly instructions
  • Intel Software Guard Extensions
    • What is SGX (Software Guard eXtensions)?
    • SGX Enclave Memory
    • Guidelines for designing applications using SGX
    • Performance Overhead
    • SDK compilation modes
    • Writing Enclave Functions
  • ARM TrustZone
    • SoC and IP
    • ARM TrustZone
    • Worlds
    • Architecture
    • TrustZone bootstrap
  • Linux Kernel Namespaces
    • Namespaces
    • Advantages
    • Process Namespace
    • Network namespace
    • Mount namespace
    • UTS namespace
    • User namespace
  • LXC Linux Containers
    • Container
    • LXC containers
  • AppArmor
    • Purpose
    • Enforcement
    • Benefits
    • Attack prevention
    • Enforcement policies
    • Enforcement modes
    • Logging and auditing
    • Profiles
  • TPM (Trusted Platform Module)
    • Trusted Platform Module (TPM)
    • History
    • Cryptographic Concepts
    • Use cases
    • TPM Software Stack (TSS)
    • TPM concepts
    • Entities
    • Key management
    • Restrict signatures
    • Sessions
    • Authorization roles
  • Bootstrap security
    • AEGIS
    • Trusted computing
    • Root of Trust Measurements
    • Trusted Computing Platform Alliance (TCPA)
    • TPM-based attestation
    • Trusted Platform identity credentials
    • UEFI (Unified Extensible Firmware Interface)
    • NSA Boot Security
    • UEFI secure boot & TPM measurements
    • Intel Trusted Execution Technology (TXT)
    • Smartcards
      • Java Cards
      • OpenCard Framework (OCF)
      • Cryptographic services
Powered by GitBook
On this page
  • Objects
  • Subjects
  • Mandatory access controls
  • Protection with ACLs (Access Control Lists)
  1. Security in Operating Systems

Access control

Last updated 1 year ago

The OS kernel is an access control monitor.

  • Controls all interactions of subjects with protected objects.

Objects

  • Hardware.

  • Entities of the computational model.

Subjects

  • Usually local processes.

    • Through the system called API.

    • A system call (or syscall) is not an ordinary function call.

  • But also messages from other hosts.

Mandatory access controls

OS kernels have plenty of mandatory access control policies.

  • They are part of the computational model logic.

  • They cannot be overruled not even by administrators.

    • Unless they change the OS kernel behaviour.

Examples:

  • The kernel runs in CPU-privileged modes, user applications run in non-privileged modes.

  • Separation of virtual memory areas.

  • Inter-process signalling.

  • Interpretation of files’ access control protections.

Protection with ACLs (Access Control Lists)

Each object has an ACL.

  • It says which subjects can do what.

An ACL can be discretionary or mandatory.

  • When mandatory it cannot be modified.

  • When discretionary it can be tailored.

An ACL is checked when an activity, on behalf of a subject, wants to manipulate the object.

  • If the manipulation request is not authorized by the ACL, the access is denied.

  • The OS kernel is responsible for enforcing ACL-based protection.