Interpretations

Answering questions that may arise related to the meaning of portions of an IEEE standard concerning specific applications.

IEEE Standards Interpretation for IEEE Std 1003.1™-2001 IEEE Standard Standard for Information Technology -- Portable Operating System Interface (POSIX®)

Copyright © 2006 by the Institute of Electrical and Electronics Engineers, Inc. 3 Park Avenue New York, New York 10016-5997 USA All Rights Reserved.

Interpretations are issued to explain and clarify the intent of a standard and do not constitute an alteration to the original standard. In addition, interpretations are not intended to supply consulting information. Permission is hereby granted to download and print one copy of this document. Individuals seeking permission to reproduce and/or distribute this document in its entirety or portions of this document must contact the IEEE Standards Department for the appropriate license. Use of the information contained in this document is at your own risk.

IEEE Standards Department Copyrights and Permissions 445 Hoes Lane, Piscataway, New Jersey 08855-1331, USA

Interpretation Request #70
Topic: select() FD_SETSIZE Relevant Sections: XSH select() Page: 0 Line: 0

Historical kernel implementions of select() have used an auto (stack) declaration of an fd_set (bit array) object and copied the data from user space into this object. Select(), however, has evolved, removing the static limitation of FD_SETSIZE descriptors. Current implementations allow a value for the first parameter, nfds, to be in excess of FD_SETSIZE.

The kernel accomplishes this by internally allocating a bit array sufficient to contain the number of bits needed to support an arbitrary number of bits being selected upon. The length of the copy into the kernel allocated array is based on the value of nfds, rather than on the value of the manifest constant FD_SETSIZE.

As a result, this permits guarding of the FD_SETSIZE definition in , and thus permits user programs to redefine the value prior to including the system header. The user definition of this value is then used in preference to the system default value.

A consequence of this is that it's possible to pass into select() a value for the first parameter - nfds - in excess of the system's FD_SETSIZE default value.

Implementing this way removes the historical limitation of FD_SETSIZE as a fixed maximum value for nfds.

This functionality was first introduced into BSD 4.4-based systems, and is utilized by many existing applications.

The standard states that select() "shall" fail and set errno to [EINVAL] when the nfds argument is less than 0 or greater than FD_SETSIZE.

The ERRORS section of the standard should be changed to accomodate more up-to-date implementations of select() so that errors for nfds > FD_SETSIZE "may fail" .

The wording in the ERRORS section where FD_SETSIZE is mentioned, should be changed to allow nfds arguments greater than the FD_SETSIZE without generating an error. So, the EINVAL error for nfds > FD_SETSIZE shoud change from "shall fail" to "may fail" .

In the "shall fail" list of errors, REMOVE "[EINVAL] The nfds argument is less than 0 or greater than FD_SETSIZE."

REPLACE WITH "[EINVAL] The nfds argument is less than 0."

AND ADD a "may fail" list that states: "Under the following conditions, pselect() and select() "may" fail and set errno to: [EINVAL] The nfds argument is greater than FD_SETSIZE."

REPLACE "[EINVAL] The nfds argument is less than 0 or greater than FD_SETSIZE "

WITH "[EINVAL] The nfds argument is less than 0. [EINVAL] The nfds argument is greater than the default limit for FD_SETSIZE and the default limit for FD_SETSIZE is enforced by the system."

Interpretation Response #70
The standard clearly states that FD_SETSIZE is a fixed limit,and conforming implementations must conform to this

Rationale for Interpretation
The redefinition of FD_SETSIZE by applications is non-conforming behavior, and the effects are not specified in the standard. It was also felt that applications can use the poll() interface for handling larger numbers of file descriptors.