readdir64

readdir64 is a function in the C standard library used to read directory entries from an open directory stream. It is the 64-bit counterpart to readdir, returning entries through a dirent64 structure that uses 64-bit fields for inode numbers and offsets. This enables correct handling of large directories and large-files ecosystems on filesystems that require 64-bit addressing.

The typical prototype is struct dirent64 *readdir64(DIR *dirp); where dirp is a pointer to a directory stream obtained from opendir or similar functions. The function returns a pointer to a statically allocated dirent64 structure containing information about the next directory entry, or NULL on end of directory or error. When NULL is returned, errno may be set to indicate an error; if end of directory is reached, errno is usually zero.

The dirent64 structure differs from the older dirent structure by using 64-bit fields, such as d_ino (inode number) and d_off (offset of the next entry in the directory), in addition to the usual entry name and length fields. This design supports directories that exceed the limits of 32-bit offsets and inode numbers.

Compatibility and usage considerations vary by system. Some platforms provide readdir64 for 32-bit and 64-bit builds to ensure large-file support, while others offer readdir64 primarily for historical compatibility. On many modern 64-bit systems, readdir64 and readdir may be functionally equivalent for typical usage, though readdir64 remains the correct choice when explicit 64-bit directory entry semantics are required.