PROTEXEC

PROT_EXEC, commonly written as PROT_EXEC, is a memory protection flag defined by POSIX interfaces such as mmap and mprotect. It indicates that the pages in a mapped region may be executed as machine code. The flag is used in combination with other protection flags to control access to a region of virtual memory.

On most Unix-like systems, the protection flags form a bitmask: PROT_READ, PROT_WRITE, PROT_EXEC, and PROT_NONE. Typical values are PROT_READ=0x1, PROT_WRITE=0x2, PROT_EXEC=0x4, PROT_NONE=0x0, though the exact numbers are platform dependent. A region can be mapped with PROT_READ | PROT_EXEC to allow reading and executing, or with PROT_READ | PROT_WRITE to allow code generation, after which protections may be narrowed to PROT_READ | PROT_EXEC.

Usage patterns include support for just-in-time compilers and dynamic code generators, which rely on PROT_EXEC to run generated code. A common approach is to allocate with PROT_WRITE, write the code, then switch to PROT_READ | PROT_EXEC (and drop PROT_WRITE). This aligns with the writeXorExecute (W^X) security principle that many kernels enforce, preventing a region from being both writable and executable at the same time.

Notes: The exact semantics and values of PROT_EXEC are platform-dependent, and some environments impose additional restrictions related to executable memory, NX bits, or executable stacks. When writing portable code, rely on system headers for the constants and follow platform-specific guidance regarding executable memory behavior.