UNIX标准及实现

UNIX标准及实现

引言

    在UNIX编程环境和C程序设计语言的标准化方面已经做了很多工作。虽然UNIX应用程序在不同的UNIX操作系统版本之间进行移植相当容易，但是20世纪80年代UNIX版本的剧增以及它们之间差别的扩大，导致很多大用户(例如美国政府)呼吁对其进行标堆化。

    本章首先将介绍过去20年来进行的各种标准化工作，然后讨论这些UNIX编程标准对本书所列举的各种UNIX操作系统实现的影响。所有标准化工作的一个重要部分是对每种实现必须定义的各种限制进行说明，所以我们将说明这些限制以及确定其值的各种方法。

UNIX标准化

ISO C

    In late 1989, ANSI Standard X3.159-1989 for the C programming language was approved. This standard was also adoptedas International Standard ISO/IEC 9899:1990. ANSI is the American National Standards Institute, the U.S. member in the International Organization for Standardization(ISO). IEC stands for the International Electrotechnical Commission.

    The C standard is now maintained and developed by the ISO/IEC international standardization working group for the C programming language, known as ISO/IEC JTC1/SC22/WG14, or WG14 for short. The intent of the ISO C standard is to provide portability of conformingC programs to a wide variety of operating systems, not only the UNIX System. This standard defines not only the syntax and semanticsof the programming language but also a standard library [Chapter 7 of ISO 1999; Plauger 1992; Appendix B of Kernighan and Ritchie 1988]. This library is important because all contemporaryUNIX systems, such as the ones described in this book, provide the library routines that are specified in the C standard.

    In 1999, the ISO C standard was updated and approvedas ISO/IEC 9899:1999, largely to improve support for applications that perform numerical processing. The changes don’t affect the POSIX interfaces described in this book, except for the addition of the restrict keyword to some of the function prototypes. This keyword is used to tell the compiler which pointer references can be optimized, by indicating that the object to which the pointer refers is accessed in the function only via that pointer

    Since 1999, three technical corrigendahave been published to correct errors in the ISO C standard—one in 2001, one in 2004, and one in 2007. As with most standards, there is a delay between the standard’s approval and the modification of software to conform to it. As each vendor’scompilation systems evolve, they add more support for the latest version of the ISO C standard

A summary of the current level of conformanceof gcc to the 1999 version of the ISO C standard is available at http://gcc.gnu.org/c99status.html. Although the C standard was updated in 2011, we deal only with the 1999 version in this text, because the other standards haven’t yet caught up with the relevantchanges.

    The ISO C library can be divided into 24 areas, based on the headers defined by the standard (see Figure 2.1). The POSIX.1 standard includes these headers, as well as others. As Figure 2.1 shows, all of these headers are supported by the four implementations (FreeBSD 8.0, Linux 3.2.0, Mac OS X 10.6.8, and Solaris 10) that are described later in this chapter.

IEEE POSIX

     POSIX is a family of standards initially developed by the IEEE(Institute of Electrical and Electronics Engineers). POSIXstands for Portable Operating System Interface. It originallyreferred only to the IEEE Standard 1003.1-1988 — the operating system interface — but was later extended to include many of the standards and draft standards with the 1003 designation, including the shell and utilities (1003.2).

    Of specific interest to this book is the 1003.1 operating system interface standard, whose goal is to promotethe portability of applications among various UNIX System environments. This standard defines the services that an operating system must provide if it is to be ‘‘POSIX compliant,’’ and has been adoptedby most computer vendors. Although the 1003.1 standard is based on the UNIX operating system, the standard is not restricted to UNIX and UNIX-like systems. Indeed, some vendors supplying proprietaryoperating systems claim that these systems have been made POSIX compliant, while still leaving all their proprietaryfeaturesin place.

    Because the 1003.1 standard specifiesan interface and not an implementation, no distinctionis made between system calls and library functions. All the routines in the standard are called functions.

    Standards are continually evolving, and the 1003.1 standard is no exception. The 1988 version, IEEE Standard 1003.1-1988, was modified and submitted to the International Organization for Standardization. No new interfaces or features were added, but the text was revised. The resulting document was published as IEEE Standard 1003.1-1990 [IEEE 1990]. This is also International Standard ISO/IEC 9945-1:1990. This standard was commonly referred to as POSIX.1, a term which we’ll use in this text to refer to the different versions of the standard.

    The IEEE 1003.1 working group continued to make changes to the standard. In 1996, a revisedversion of the IEEE 1003.1 standard was published. It included the 1003.1-1990 standard, the 1003.1b-1993 real-time extensions standard, and the interfaces for multithreaded programming, called pthreads for POSIX threads. This version of the standard was also published as International Standard ISO/IEC 9945-1:1996. More real- time interfaces were added in 1999 with the publication of IEEE Standard 1003.1d-1999. A year later, IEEE Standard 1003.1j-2000 was published, including even more real-time interfaces, and IEEE Standard 1003.1q-2000 was published, adding event-tracing extensions to the standard.

    The 2001 version of 1003.1 departed from the prior versions in that it combined several 1003.1 amendments, the 1003.2 standard, and portionsof the Single UNIX Specification (SUS), Version 2 (more on this later). The resulting standard, IEEE Standard 1003.1-2001, included the following other standards:

• ISO/IEC 9945-1 (IEEE Standard 1003.1-1996), which includes
• IEEE Standard 1003.1-1990
• IEEE Standard 1003.1b-1993 (real-time extensions)
• IEEE Standard 1003.1c-1995 (pthreads)
• IEEE Standard 1003.1i-1995 (real-time technical corrigenda)
• IEEE P1003.1a draft standard (system interface amendment)
• IEEE Standard 1003.1d-1999 (advanced real-time extensions)
• IEEE Standard 1003.1j-2000 (more advanced real-time extensions)
• IEEE Standard 1003.1q-2000 (tracing)
• Parts of IEEE Standard 1003.1g-2000 (protocol-independent interfaces)
• ISO/IEC 9945-2 (IEEE Standard 1003.2-1993)
• IEEE P1003.2b draft standard (shell and utilities amendment)
• IEEE Standard 1003.2d-1994 (batch extensions)
• The Base Specifications of the Single UNIX Specification, version 2, which include
• System Interface Definitions, Issue 5
• Commands and Utilities, Issue 5
• System Interfaces and Headers, Issue5
• Open Group Technical Standard, Networking Services, Issue 5.2
• ISO/IEC 9899:1999, Programming Languages–C

    In 2004, the POSIX.1 specificationwas updated with technical corrections; more comprehensivechanges were made in 2008 and released as Issue7 of the Base Specifications. ISO approvedthis version at the end of 2008 and published it in 2009 as International Standard ISO/IEC 9945:2009. It is based on several other standards:

• IEEE Standard 1003.1, 2004 Edition
• Open Group Technical Standard, 2006, Extended API Set, Parts 1–4
• ISO/IEC 9899:1999, including corrigenda

    Figure 2.2, Figure 2.3, and Figure 2.4 summarize the required and optional headers as specified by POSIX.1. Because POSIX.1 includes the ISO C standard library functions, it also requires the headers listed in Figure 2.1. All four figures summarize which headers are included in the implementations discussed in this book.

Header	FreeBSD 8.0	Linux 3.2.0	Mac OS X 10.6.8	Solaris 10	Description
<aio.h>	•	•	•	•	asynchronous I/O
<cpio.h>	•	•	•	•	cpio archive values
<dirent.h>	•	•	•	•	directory entries
<dlfcn.h>	•	•	•	•	dynamic linking
<fcntl.h>	•	•	•	•	file control
<fnmatch.h>	•	•	•	•	filename-matching types
<glob.h>	•	•	•	•	pathname pattern-matching and generation
<grp.h>	•	•	•	•	group file
<iconv.h>	•	•	•	•	codeset conversion utility
<langinfo.h>	•	•	•	•	language information constants
<monetary.h>	•	•	•	•	monetarytypes and functions
<netdb.h>	•	•	•	•	network database operations
<nl_types.h>	•	•	•	•	message catalogs
<poll.h>	•	•	•	•	poll function
<pthread.h>	•	•	•	•	threads
<pwd.h>	•	•	•	•	password file
<regex.h>	•	•	•	•	regular expressions
<sched.h>	•	•	•	•	execution scheduling
<semaphore.h>	•	•	•	•	semaphores
<strings.h>	•	•	•	•	string operations
<tar.h>	•	•	•	•	tar archive values
<termios.h>	•	•	•	•	terminal I/O
<unistd.h>	•	•	•	•	symbolic constants
<wordexp.h>	•	•	•	•	word-expansion definitions
<arpa/inet.h>	•	•	•	•	Internet definitions
<net/if.h>	•	•	•	•	socket local interfaces
<netinet/in.h>	•	•	•	•	Internet address family
<netinet/tcp.h>	•	•	•	•	Transmission Control Protocol definitions
<sys/mman.h>	•	•	•	•	memory management declarations
<sys/select.h>	•	•	•	•	select function
<sys/socket.h>	•	•	•	•	sockets interface
<sys/stat.h>	•	•	•	•	file status
<sys/statvfs.h>	•	•	•	•	file system information
<sys/times.h>	•	•	•	•	process times
<sys/types.h>	•	•	•	•	primitive system data types
<sys/un.h>	•	•	•	•	UNIX domain socket definitions
<sys/utsname.h>	•	•	•	•	system name
<sys/wait.h>	•	•	•	•	process control

Figure 2.2 Required headers defined by the POSIX standard

    In this text we describe the 2008 edition of POSIX.1. Its interfaces are divided into required ones and optional ones. The optional interfaces are further divided into 40 sections, based on functionality. The sections containing nonobsoleteprogramming interfaces are summarized in Figure 2.5 with their respective option codes. Option codes are two- to three-character abbreviationsthat identifythe interfaces that belong to each functional area and highlight text describing aspects of the standard that depend on the support of a particular option. Many options deal with real-time extensions.

Header	FreeBSD 8.0	Linux 3.2.0	Mac OS X 10.6.8	Solaris 10	Description
<fmtmsg.h>	•	•	•	•	message display structures
<ftw.h>	•	•	•	•	file tree walking
<libgen.h>	•	•	•	•	pathname management functions
<ndbm.h>	•		•	•	database operations
<search.h>	•	•	•	•	search tables
<syslog.h>	•	•	•	•	system error logging
<utmpx.h>	•	•	•	•	user accounting database
<sys/ipc.h>	•	•	•	•	IPC
<sys/msg.h>	•	•	•	•	XSI message queues
<sys/resource.h>	•	•	•	•	resource operations
<sys/sem.h>	•	•	•	•	XSI semaphores
<sys/shm.h>	•	•	•	•	XSI shared memory
<sys/time.h>	•	•	•	•	time types
<sys/uio.h>	•	•	•	•	vector I/O operations

Figure 2.3 XSI option headers defined by the POSIX standard

Header	FreeBSD 8.0	Linux 3.2.0	Mac OS X 10.6.8	Solaris 10	Description
<mqueue.h>	•	•	•	•	message queues
<spawn.h>	•	•	•	•	real-time spawn interface

Figure 2.4 Optional headers defined by the POSIX standard

Code	SUS  mandatory	Symbolic constant	Description
ADV		_POSIX_ADVISORY_INFO	advisory information (real-time)
CPT		_POSIX_CPUTIME	process CPU time clocks (real-time)
FSC		_POSIX_FSYNC	file synchronization
IP6		_POSIX_IPV6	IPv6 interfaces
ML		_POSIX_MEMLOCK	process memory locking (real-time)
MLR		_POSIX_MEMLOCK_RANGE	memory range locking (real-time)
MON		_POSIX_MONOTONIC_CLOCK	monotonicclock (real-time)
MSG		_POSIX_MESSAGE_PASSING	message passing (real-time)
MX		_ _STDC_IEC_559_ _	IEC 60559 floating-point option
PIO		_POSIX_PRIORITIZED_IO	prioritized input and output
PS		_POSIX_PRIORITY_SCHEDULING	process scheduling(real-time)
RPI		_POSIX_THREAD_ROBUST_PRIO_INHERIT	robust mutex priority inheritance(real-time)
RPP		_POSIX_THREAD_ROBUST_PRIO_PROTECT	robust mutexpriority protection (real-time)
RS		_POSIX_RAW_SOCKETS	raw sockets
SHM		_POSIX_SHARED_MEMORY_OBJECTS	shared memory objects (real-time)
SIO		_POSIX_SYNCHRONIZED_IO	synchronizedinput and output (real-time)
SPN		_POSIX_SPAWN	spawn(real-time)
SS		POSIX_SPORADIC_SERVER	process sporadicserver (real-time)
TCT		_POSIX_THREAD_CPUTIME	thread CPU time clocks (real-time)
TPI		_POSIX_THREAD_PRIO_INHERIT	nonrobust mutex priority inheritance (real-time)
TPP		_POSIX_THREAD_PRIO_PROTECT	nonrobust mutex priority protection (real-time)
TPS		_POSIX_THREAD_PRIORITY_SCHEDULING	thread execution scheduling(real-time)
TSA		_POSIX_THREAD_ATTR_STACKADDR	thread stack address attribute
TSH		_POSIX_THREAD_PROCESS_SHARED	thread process-shared synchronization
TSP		_POSIX_THREAD_SPORADIC_SERVER	thread sporadicserver (real-time)
TSS		_POSIX_THREAD_ATTR_STACKSIZE	thread stack size address
TYM		_POSIX_TYPED_MEMORY_OBJECTS	typed memory objects (real-time)
XSI		_XOPEN_UNIX	X/Open interfaces

Figure 2.5 POSIX.1 optional interface groups and codes

    POSIX.1 does not include the notionof a superuser. Instead, certain operations require ‘‘appropriate privileges,’’ although POSIX.1 leaves the definition of this term up to the implementation. UNIX systems that conform to the Department of Defense’s security guidelineshave many levels of security. In this text, however, we use the traditional terminologyandrefer to operations that require superuser privilege.

    After more than twenty years of work, the standards are matureand stable. The POSIX.1 standard is maintained by an open working group known as the Austin Group (http://www.opengroup.org/austin). To ensure that they are still relevant, the standards need to be either updated or reaffirmed every so often.

Single UNIX Specification

    The Single UNIX Specification, a superset of the POSIX.1 standard, specifiesadditional interfaces that extend the functionality provided by the POSIX.1 specification. POSIX.1 is equivalentto the Base Specifications portion of the Single UNIX Specification.

    The X/Open System Interfaces (XSI) option in POSIX.1 describes optional interfaces and defines which optional portionsof POSIX.1 must be supported for an implementation to be deemedXSI conforming. These include file synchronization, thread stack address and size attributes, thread process-shared synchronization, and the _XOPEN_UNIX symbolic constant (marked "SUS mandatory" in Figure 2.5). Only XSI- conforming implementations can be called UNIX systems.

The Open Group owns the UNIX trademark and uses the Single UNIX Specification to define the interfaces an implementation must support to call itself a UNIX system. Vendorsmust file conformance statements, pass test suitesto verify conformance, and licensethe right to use the UNIX trademark

    Several of the interfaces that are optional for XSI-conforming systems are divided into option groups based on common functionality, as follows:

• Encryption: denotedby the _XOPEN_CRYPT symbolic constant
• Real-time: denotedby the _XOPEN_REALTIME symbolic constant
• Advanced real-time
• Real-time threads: denoted by _XOPEN_REALTIME_THREADS
• Advanced real-time threads

    The Single UNIX Specification is a publication of The Open Group, which was formed in 1996 as a merger of X/Open and the Open Software Foundation (OSF), both industry consortia. X/Open used to publish the X/Open Portability Guide, which adoptedspecific standards and filled in the gaps where functionality was missing. The goal of these guides was to improve application portability beyond what was possible by merely conforming to published standards.

    The first version of the Single UNIX Specification was published by X/Open in 1994. It was also known as "Spec 1170", because it contained roughly1,170 interfaces. It grew out of the Common Open Software Environment (COSE) initiative, whose goal was to improve application portability across all implementations of the UNIX operating system. The COSE group — Sun, IBM, HP, Novell/USL, and OSF—went further than endorsingstandards by including interfaces used by common commercial applications. The resulting 1,170 interfaces were selected from these applications, and also included the X/Open Common Application Environment (CAE), Issue 4 (known as "XPG4" as a historical reference to its predecessor, the X/Open Portability Guide), the System V Interface Definition (SVID), Edition 3, Level 1 interfaces, and the OSF Application Environment Specification (AES) Full Use interfaces.

    The second version of the Single UNIX Specification was published by The Open Group in 1997. The new version added support for threads, real-time interfaces, 64-bit processing, large files, and enhanced multibyte character processing.

    The third version of the Single UNIX Specification (SUSv3) was published by The Open Group in 2001. The Base Specificationsof SUSv3 are the same as IEEE Standard 1003.1-2001 and are divided into four sections: Base Definitions, System Interfaces, Shell and Utilities, and Rationale. SUSv3 also includes X/Open Curses Issue 4, Version 2, but this specification is not part of POSIX.1.

    In 2002, ISO approvedthe IEEE Standard 1003.1-2001 as International Standard ISO/IEC 9945:2002. The Open Group updated the 1003.1 standard again in 2003 to include technical corrections, and ISO approved this as International Standard ISO/IEC 9945:2003. In April 2004, The Open Group published the Single UNIX Specification, Version 3, 2004 Edition. It merged more technical corrections into the main text of the standard.

    In 2008, the Single UNIX Specification was updated, including corrections and newinterfaces, removing obsolete interfaces, and marking other interfaces as being obsolescent in preparation for future removal. Additionally, some previously optionalinterfaces were promoted to nonoptional status, including asynchronous I/O, barriers,clock selection, memory-mapped files, memory protection, reader–writer locks, real-time signals, POSIX semaphores, spinlocks, thread-safe functions, threads, timeouts,and timers. The resulting standard is known as Issue7 of the Base Specifications, and is the same as POSIX.1-2008. The Open Group bundledthis version with an updatedX/Open Cursesspecification and released them as version 4 of the Single UNIXSpecification in 2010. We’ll refer to this as SUSv4.

UNIX系统实现

    The previous section described ISO C, IEEE POSIX, and the Single UNIX Specification — three standards originallycreated by independentorganizations. Standards, however, are interface specifications. How do these standards relate to the real world? These standards are taken by vendors and turned into actual implementations. In this book, we are interested in both these standards and their implementation.

    Section 1.1 of McKusick et al. [1996] gives a detailed history (and a nice picture) of the UNIX System family tree(in Figure 2.5.1). Everything starts from the Sixth Edition (1976) and Seventh Edition (1979) of the UNIX Time-Sharing System on the PDP-11 (usually called Version 6 and Version 7, respectively). These were the first releases widely distributed outside of Bell Laboratories. Three branches of the tree evolved.

    1) One at AT&T that led to System III and System V, the so-called commercial versions of the UNIX System.

    2) One at the University of California at Berkeley that led to the 4.xBSD implementations.

    3) The research version of the UNIX System, developed at the Computing Science Research Center of AT&T Bell Laboratories, that led to the UNIX Time-Sharing System 8th Edition, 9th Edition, and ended with the 10th Edition in 1990.

Figure 2.5.1   the UNIX System family tree

1. First Edition (V1)
2. |
3. Second Edition (V2)
4. |
5. Third Edition (V3)
6. |
7. Fourth Edition (V4)
8. |
9. Fifth Edition (V5)
10. |
11. Sixth Edition (V6) -----*
12. \ |
13. \ |
14. \ |
15. Seventh Edition (V7) |
16. \ |
17. \ 1BSD
18. 32V |
19. \ 2BSD---------------*
20. \ / |
21. \ / |
22. \/ |
23. 3BSD |
24. | |
25. 4.0BSD 2.7.9BSD
26. | |
27. 4.1BSD --------------> 2.8BSD
28. | |
29. 4.1aBSD -----------\ |
30. | \ |
31. 4.1bBSD \ |
32. | \ |
33. *------ 4.1cBSD --------------> 2.9BSD
34. / | |
35. Eighth Edition | 2.9BSD-Seismo
36. | | |
37. +----<--- 4.2BSD 2.9.1BSD
38. | | |
39. +----<--- 4.3BSD -------------> 2.10BSD
40. | | / |
41. Ninth Edition | / 2.10.1BSD
42. | 4.3BSD Tahoe-----+ |
43. | | \ |
44. | | \ |
45. v | 2.11BSD
46. Tenth Edition | |
47. | 2.11BSD rev #430
48. 4.3BSD NET/1 |
49. | v
50. 4.3BSD Reno
51. |
52. *---------- 4.3BSD NET/2 -------------------+-------------*
53. | | | |
54. 386BSD 0.0 | | BSD/386 ALPHA
55. | | | |
56. 386BSD 0.1 ------------>+ | BSD/386 0.3.[13]
57. | \ | 4.4BSD Alpha |
58. | 386BSD 1.0 | | BSD/386 0.9.[34]
59. | | 4.4BSD |
60. | | / | |
61. | | 4.4BSD-Encumbered | |
62. | NetBSD 0.8 | BSD/386 1.0
63. | | | |
64. FreeBSD 1.0 NetBSD 0.9 | BSD/386 1.1
65. | | .----- 4.4BSD Lite |
66. FreeBSD 1.1 | / / | \ |
67. | | / / | \ |
68. FreeBSD 1.1.5 .---|--------' / | \ |
69. | / | / | \ |
70. FreeBSD 1.1.5.1 / | / | \ |
71. | / NetBSD 1.0 <-' | \ |
72. | / | | \ |
73. FreeBSD 2.0 <--' | | BSD/OS 2.0
74. | \ | |
75. FreeBSD 2.0.5 \ | BSD/OS 2.0.1
76. | .-----\------------- 4.4BSD Lite2 |
77. | | \ | | | | |
78. | | .-----|------Rhapsody | | | |
79. | | | | NetBSD 1.3 | | |
80. | | | | OpenBSD 2.3 | |
81. | | | | BSD/OS 3.0 |
82. FreeBSD 2.1 | | | |
83. | | | | NetBSD 1.1 ------. BSD/OS 2.1
84. | FreeBSD 2.1.5 | | | \ |
85. | | | | NetBSD 1.2 \ BSD/OS 3.0
86. | FreeBSD 2.1.6 | | | \ OpenBSD 2.0 |
87. | | | | | \ | |
88. | FreeBSD 2.1.6.1 | | | \ | |
89. | | | | | \ | |
90. | FreeBSD 2.1.7 | | | | | |
91. | | | | | NetBSD 1.2.1 | |
92. | FreeBSD 2.1.7.1 | | | | |
93. | | | | | |
94. | | | | | |
95. *-FreeBSD 2.2 | | | | |
96. | \ | | | | |
97. | FreeBSD 2.2.1 | | | | |
98. | | | | | | |
99. | FreeBSD 2.2.2 | | | OpenBSD 2.1 |
100. | | | | | | |
101. | FreeBSD 2.2.5 | | | | |
102. | | | | | OpenBSD 2.2 |
103. | | | | NetBSD 1.3 | |
104. | FreeBSD 2.2.6 | | | | | |
105. | | | | | NetBSD 1.3.1 | BSD/OS 3.1
106. | | | | | | OpenBSD 2.3 |
107. | | | | | NetBSD 1.3.2 | |
108. | FreeBSD 2.2.7 | | | | | |
109. | | | | | | | BSD/OS 4.0
110. | v | | | | | |
111. | FreeBSD 2.2.8 | | | | | |
112. | | | | | OpenBSD 2.4 |
113. FreeBSD 3.0 <--------* | | v | |
114. | | | NetBSD 1.3.3 | |
115. *---FreeBSD 3.1 | | | |
116. | | | | | BSD/OS 4.0.1
117. | FreeBSD 3.2----* | NetBSD 1.4 OpenBSD 2.5 |
118. | | | | | | | | |
119. | | | | | | | | |
120. | | | | | | | | |
121. | FreeBSD 3.3 | | | | NetBSD 1.4.1 | |
122. | | | | | | | OpenBSD 2.6 |
123. | FreeBSD 3.4 | | | | | | |
124. | | | | | | | | BSD/OS 4.1
125. FreeBSD 4.0 | | | | | NetBSD 1.4.2 | |
126. | | | | | | | | |
127. | | | | | | | | |
128. | FreeBSD 3.5 | | | | | OpenBSD 2.7 |
129. | | | | | | | | |
130. | FreeBSD 3.5.1 | | | | | | |
131. | | | | | | | |
132. *---FreeBSD 4.1 | | | | | | |
133. | | | | (?) | | | |
134. | FreeBSD 4.1.1 | | / | | | |
135. | | | | / | | | |
136. | FreeBSD 4.2 Darwin/ | NetBSD 1.4.3 | |
137. | | Mac OS X | OpenBSD 2.8 BSD/OS 4.2
138. | | | | | |
139. | | | | | |
140. | | 10.0 NetBSD 1.5 | |
141. | FreeBSD 4.3 | | | | |
142. | | | | | OpenBSD 2.9 |
143. | | | | NetBSD 1.5.1 | |
144. | | | | | | |
145. | FreeBSD 4.4-. | | NetBSD 1.5.2 | |
146. | | | Mac OS X | | | |
147. | | | 10.1 | | OpenBSD 3.0 |
148. | FreeBSD 4.5 | | | | | |
149. | | \ | | | | BSD/OS 4.3
150. | FreeBSD 4.6 \ | | | OpenBSD 3.1 |
151. | | \ | | NetBSD 1.5.3 | |
152. | FreeBSD 4.6.2 Mac OS X | | |
153. | | 10.2 | | |
154. | FreeBSD 4.7 | | | |
155. | | | NetBSD 1.6 OpenBSD 3.2 |
156. | FreeBSD 4.8 | | | | |
157. | | | | NetBSD 1.6.1 | |
158. | |--------. | | | OpenBSD 3.3 BSD/OS 5.0
159. | | \ | | | | |
160. | FreeBSD 4.9 | | | | OpenBSD 3.4 BSD/OS 5.1 ISE
161. | | | | | | |
162. | | | | | NetBSD 1.6.2 |
163. | | | | | | |
164. | | | | | | OpenBSD 3.5
165. | | | | | v |
166. | FreeBSD 4.10 | | | |
167. | | | | | |
168. | FreeBSD 4.11 | | | |
169. | | | | | |
170. | v `-|------|-----------------|---------------------.
171. | | | | \
172. FreeBSD 5.0 | | | |
173. | | | | |
174. FreeBSD 5.1 | | | DragonFly 1.0
175. | \ | | | |
176. | ----- Mac OS X | | |
177. | 10.3 | | |
178. FreeBSD 5.2 | | | |
179. | | | | | |
180. | FreeBSD 5.2.1 | | | |
181. | | | | |
182. *---FreeBSD 5.3 | | | |
183. | | | | OpenBSD 3.6 |
184. | | | NetBSD 2.0 | |
185. | | | | | | | DragonFly 1.2.0
186. | | Mac OS X | | NetBSD 2.0.2 | |
187. | FreeBSD 5.4 10.4 | | | | |
188. | | | | | | OpenBSD 3.7 |
189. | V | | | NetBSD 2.0.3 | |
190. | | | | | | |
191. *---FreeBSD 6.0 | | | v OpenBSD 3.8 |
192. | | | | | | |
193. | V | | \ | |
194. | | | NetBSD 2.1 | |
195. | | | | |
196. | | NetBSD 3.0 | |
197. | | | | DragonFly 1.4.0
198. | | | | |
199. FreeBSD 7 -current | NetBSD -current OpenBSD -current |
200. | | | | |
201. v v v v v

SVR4

    UNIX System V Release 4 (SVR4) was a product of AT&T's UNIX System Laboratories (USL, formerly AT&T's UNIX Software Operation).  SVR4 merged functionality from AT&T UNIX System V Release 3.2 (SVR3.2), the SunOS operating system from Sun Microsystems, the 4.3BSD release from the University of California, and the Xenix system from Microsoft into one coherentoperating system. (Xenix was originallydeveloped from Version 7, with many features later taken from System V.) The SVR4 source code was released in late 1989, with the first end-user copies becoming available during 1990. SVR4 conformed to both the POSIX 1003.1 standard and the X/Open Portability Guide, Issue3 (XPG3).

    AT&T also published the System V Interface Definition (SVID) [AT&T 1989]. Issue 3 of the SVID specifiedthe functionality that an operating system must offer to qualifyas a conforming implementation of UNIX System V Release 4. As with POSIX.1, the SVID specified an interface, not an implementation. No distinctionwas made in the SVID between system calls and library functions. The reference manual for an actual implementation of SVR4 must be consulted to see this distinction [AT&T 1990e].

4.4BSD

    The Berkeley Software Distribution (BSD) releases were produced and distributed by the Computer Systems Research Group (CSRG) at the University of California at Berkeley. 4.2BSD was released in 1983 and 4.3BSD in 1986. Both of these releases ran on the VAX minicomputer. The next release, 4.3BSD Tahoe in 1988, also ran on a particular minicomputer called the Tahoe.This was followed in 1990 with the 4.3BSD Reno release; 4.3BSD Reno supported many of the POSIX.1 features.

    The original BSD systems contained proprietaryAT&T source code and were covered by AT&T licenses. To obtain the source code to the BSD system you had to have a UNIX source license from AT&T. This changed as more and more of the AT&T source code was replaced over the years with non-AT&T source code and as many of the new features added to the Berkeley system were derived from non-AT&T sources.

    In 1989, Berkeley identifiedmuch of the non-AT&T source code in the 4.3BSD Tahoe release and made it publicly availableas the BSD Networking Software, Release 1.0. Release 2.0 of the BSD Networking Software followed in 1991, which was derived from the 4.3BSD Reno release. The intent was that most, if not all, of the 4.4BSD system would be free of AT&T license restrictions, thus making the source code availableto all.

    4.4BSD-Lite was intended to be the final release from the CSRG. Its introductionwas delayed, however, because of legal battles with USL. Once the legal differences were resolved, 4.4BSD-Lite was released in 1994, fully unencumbered, so no UNIX source license was needed to receive it. The CSRG followed this with a bug-fix release in 1995. This release, 4.4BSD-Lite, release 2, was the final version of BSD from the CSRG. (This version of BSD is described in the book by McKusick et al. [1996].)

    The UNIX system development done at Berkeley started with PDP-11s, then moved to the VAX minicomputer, and then to other so-called workstations. During the early 1990s, support was provided to Berkeley for the popular 80386-based personal computers, leading to what is called 386BSD. This support was provided by Bill Jolitz and was documented in a series of monthly articles in Dr. Dobb’s Journal throughout1991. Much of this code appeared in the BSD Networking Software, Release 2.0.

FreeBSD

    FreeBSD is based on the 4.4BSD-Lite operating system. The FreeBSD project was formed tocarry onthe BSD line after the Computing Science Research Group (CSRG) at the University of California at Berkeley decided to end its work on the BSD versions of the UNIX operating system, and the 386BSD project seemed to be neglectedfor too long.

    All software produced by the FreeBSD project is freely available in both binary and source forms. The FreeBSD 8.0 operating system was one of the four operating systems used to test the examples in this book.

Several other BSD-based free operating systems are available.  The NetBSD project (http://www.netbsd.org) is similar to the FreeBSD project, but emphasizesportability between hardware platforms. The OpenBSD project (http://www.openbsd.org) is similar to FreeBSD but places a greater emphasis on security.

Linux

    Linux is an operating system that provides a rich programming environment similar to that of a UNIX System; it is freely available under the GNU Public License. The popularityof Linux is somewhat of a phenomenon in the computer industry. Linux is distinguishedby often being the first operating system to support new hardware.

    Linux was created in 1991 by Linus Torvalds as a replacement for MINIX. A grass-roots effort then sprang up, wherebymany developers across the world volunteeredtheir time to use and enhance it. 

Mac OS X

    Mac OS X is based on entirelydifferent technology than priorversions. The core operating system is called ‘‘Darwin,’’ andis based on a combination of the Mach kernel (Accetta et al. [1986]), the FreeBSD operating system, and an object-oriented framework for drivers and other kernel extensions. As of version 10.5, the Intel port of Mac OS X has been certifiedto be a UNIX system. (For more information on UNIX certification, see http://www.opengroup.org/certification/idx/unix.html.)

Solaris

    Solaris is the version of the UNIX System developed by Sun Microsystems (now Oracle). Solaris is based on System V Release 4, but includes more than fifteen years of enhancements from the engineers at Sun Microsystems. It is arguablythe only commerciallysuccessful SVR4 descendant, and is formally certifiedto be a UNIX system.

    In 2005, Sun Microsystems released most of the Solaris operating system source code to the public as part of the OpenSolaris open source operating system in an attempt to build an external developer community around Solaris.

Other UNIX System

Other versions of the UNIX system that have been certifiedin the past include

• AIX, IBM’s version of the UNIX System

• HP-UX, Hewlett-Packard’s version of the UNIX System

• IRIX, the UNIX System version shipped by Silicon Graphics

• UnixWare, the UNIX System descendedfrom SVR4 sold by SCO

标准和实现的关系

    The standardsthat we’ve mentioned define a subset of any actual system. The focus of this book is on four real systems: FreeBSD 8.0, Linux 3.2.0, Mac OS X 10.6.8, and Solaris 10. Although only Mac OS X and Solaris can call themselves UNIX systems, all four provide a similar programming environment. Because all four are POSIX compliantto varying degrees, we will also concentrate on the features required by the POSIX.1 standard, noting any differences between POSIX and the actual implementations of these four systems. Those features and routines that are specific to only a particular implementation are clearly marked. We’ll also note any features that are required on UNIX systems but are optional on other POSIX-conforming systems.

    Be aware that the implementations provide backward compatibility for features in earlier releases, such as SVR3.2 and 4.3BSD. For example, Solaris supports both the POSIX.1 specification for nonblocking I/O (O_NONBLOCK) and the traditional System V method (O_NDELAY). In this text, we’ll use only the POSIX.1 feature, although we’ll mention the nonstandard feature that it replaces. Similarly, both SVR3.2 and 4.3BSD provided reliablesignals in a way that differs from the POSIX.1 standard. In Chapter 10 we describe only the POSIX.1 signal mechanism.

限制

    The implementations define many magic numbers and constants. Many of these have been hard coded into programs or were determined using ad hoc techniques. With the various standardizationefforts that we’ve described, more portable methods are now provided to determine these magic numbers and implementation-defined limits, greatly improving the portability of software written for the UNIX environment.

    Two types of limits are needed:

    1) Compile-time limits (e.g., what’s the largest value of a short integer?)

    2) Runtime limits (e.g., how many bytes in a filename?)

    Compile-time limits can be defined in headers that any program can include at compile time. But runtime limits require the process to call a function to obtain the limit’s value.

    Additionally, some limits can be fixed on a given implementation—and could therefore be defined statically in a header—yet vary on another implementation and would require a runtime function call. An example of this type of limit is the maximum number of bytes in a filename. Before SVR4, System V historicallyallowed only 14 bytes in a filename, whereas BSD-derived systems increased this number to 255. Most UNIX System implementations these days support multiple file system types, and each type has its own limit. This is the case of a runtime limit that depends on where in the file system the file in question is located. A filename in the root file system, for example, could have a 14-byte limit, whereas a filename in another file system could have a 255-byte limit.

    To solve these problems, three types of limits are provided:

    1) Compile-time limits (headers)

    2) Runtime limits not associated with a file or directory (the sysconf function)

    3) Runtime limits that are associated with a file or a directory (the pathconf and

fpathconf functions)

    To further confusethings, if a particular runtime limit does not vary on a given system, it can be defined statically in a header. If it is not defined in a header, however, the application must call one of the three conf functions (which we describe shortly) to determine its value at runtime.

Name	Description	Minimum  acceptable value	Typical value
CHAR_BIT	bits in a char	8	8
CHAR_MAX	max value of char	(see later)	127
CHAR_MIN	min value of char	(see later)	-128
SCHAR_MAX	max value of signed char	127	127
SCHAR_MIN	min value of signed char	−127	−128
UCHAR_MAX	max value of unsigned char	255	255
INT_MAX	max value of int	32,767	2,147,483,647
INT_MIN	min value of int	−32,767	−2,147,483,648
UINT_MAX	max value of unsigned int	65,535	4,294,967,295
SHRT_MAX	max value of short	32,767	32,767
SHRT_MIN	min value of short	−32,767	−32,768
USHRT_MAX	max value of unsigned short	65,535	65,535
LONG_MAX	max value of long	2,147,483,647	2,147,483,647
LONG_MIN	min value of long	−2,147,483,647	−2,147,483,648
ULONG_MAX	max value of unsigned long	4,294,967,295	4,294,967,295
LLONG_MAX	max value of long long	9,223,372,036,854,775,807	9,223,372,036,854,775,807
LLONG_MIN	min value of long long	−9,223,372,036,854,775,807	−9,223,372,036,854,775,808
ULLONG_MAX	max value of unsigned long long	18,446,744,073,709,551,615	18,446,744,073,709,551,615
MB_LEN_MAX	max number of bytes multibyte character constant	1	6

Figure 2.6 Sizes of integral values from <limits.h>

ISO C限制

    All of the compile-time limits defined by ISO C are defined in the file <limits.h> (see Figure 2.6). These constants don’t change in a given system. The third column in Figure 2.6 shows the minimum acceptable values from the ISO C standard. This allows for a system with 16-bit integers using one’s-complementarithmetic. The fourth column shows the values from a Linux system with 32-bit integers using two’s- complement arithmetic. Note thatnone of the unsigned data types has a minimum value, as this value must be 0 for an unsigned data type. On a 64-bit system, the values for long integer maximums match the maximum values for long long integers.

    One difference that we will encounteris whether a system provides signed or unsigned character values. From the fourth column in Figure 2.6, we see that this particular system uses signed characters. We see that CHAR_MIN equals SCHAR_MIN and that CHAR_MAX equals SCHAR_MAX. If the system uses unsigned characters, we would have CHAR_MIN equal to 0 and CHAR_MAX equal to UCHAR_MAX.

    The floating-point data types in the header <float.h> have a similar set of definitions. Anyone doing serious floating-point work should examine this file.

    Although the ISO C standard specifiesminimum acceptable values for integral data types, POSIX.1 makes extensionsto the C standard. To conformto POSIX.1, an implementation must support a minimum value of 2,147,483,647 for INT_MAX, −2,147,483,647 for INT_MIN, and 4,294,967,295 for UINT_MAX. Because POSIX.1 requires implementations to support an 8-bit char, CHAR_BIT must be 8, SCHAR_MIN must be −128, SCHAR_MAX must be 127, and UCHAR_MAX must be 255.

    Another ISO C constant that we’ll encounteris FOPEN_MAX, the minimum number of standard I/O streams that the implementation guaranteescan be open at once. This constant is found in the <stdio.h> header, and its minimum value is 8. The POSIX.1 value STREAM_MAX, if defined, must have the same value as FOPEN_MAX.

    ISO C also defines the constant TMP_MAX in <stdio.h>. It is the maximum number of unique filenames generated by the tmpnam function. We’ll have more to say about this constant in Section 5.13.

    Although ISO C defines the constant FILENAME_MAX, we avoid using it, because POSIX.1 provides better alternatives(NAME_MAX and PATH_MAX). We’ll see these constants shortly.

    Figure 2.7shows the values of FILENAME_MAX, FOPEN_MAX, and TMP_MAX on the four platforms we discuss in this book.

Limit	FreeBSD 8.0	Linux 3.2.0	Mac OS X 10.6.8	Solaris 10
FOPEN_MAX	20	16	20	20
TMP_MAX	308,915,776	238,328	308,915,776	17,576
FILENAME_MAX	1024	4096	1024	1024

Figure 2.7 ISO limits on various platforms

POSIX限制

    POSIX.1 defines numerousconstants that deal with implementation limits of the operating system. Unfortunately, this is one of the more confusing aspectsof POSIX.1. Although POSIX.1 defines numerouslimits and constants, we’ll concernourselves with only the ones that affect the base POSIX.1 interfaces. These limits and constants are divided into the following seven categories:

    1) Numerical limits: LONG_BIT, SSIZE_MAX, and WORD_BIT

    2) Minimum values: the 25 constants in Figure 2.8

    3) Maximum value: _POSIX_CLOCKRES_MIN

    4) Runtime increasable values: CHARCLASS_NAME_MAX, COLL_WEIGHTS_MAX, LINE_MAX, NGROUPS_MAX, and RE_DUP_MAX

    5) Runtime invariantvalues, possibly indeterminate: the 17 constants in Figure 2.9 (plus an additional four constants introduced in Section 12.2 and three constants introduced in Section 14.5)

    6) Other invariant  values: NL_ARGMAX, NL_MSGMAX, NL_SETMAX, and NL_TEXTMAX

    7) Pathname variable  values: FILESIZEBITS, LINK_MAX, MAX_CANON, MAX_INPUT, NAME_MAX, PATH_MAX, PIPE_BUF, and SYMLINK_MAX

Name	Description: minimum acceptable value for maximum ...	Value
_POSIX_ARG_MAX	length of arguments to exec functions	4,096
_POSIX_CHILD_MAX	number of child processes at a time per real user ID	25
_POSIX_DELAYTIMER_MAX	number of timer expiration overruns	32
_POSIX_HOST_NAME_MAX	length of a host name as returned by gethostname	255
_POSIX_LINK_MAX	number of links to a file	8
_POSIX_LOGIN_NAME_MAX	length of a login name	9
_POSIX_MAX_CANON	number of bytes on a terminal’s canonicalinput queue	255
_POSIX_MAX_INPUT	space available on a terminal’s input queue	255
_POSIX_NAME_MAX	number of bytes in a filename, not including the terminating null	14
_POSIX_NGROUPS_MAX	number of simultaneous supplementarygroup IDs per process	8
_POSIX_OPEN_MAX	maximum number of open files per process	20
_POSIX_PATH_MAX	number of bytes in a pathname, including the terminating null	256
_POSIX_PIPE_BUF	number of bytes that can be written atomicallyto a pipe	512
_POSIX_RE_DUP_MAX	number of repeated occurrencesof a basic regular expression permittedby the regexec and regcomp functions when using the interval notation\{m,n\}	255
_POSIX_RTSIG_MAX	number of real-time signal numbers reservedfor applications	8
_POSIX_SEM_NSEMS_MAX	number of semaphores a process can have in use at one time	256
_POSIX_SEM_VALUE_MAX	value a semaphore can hold	32,767
_POSIX_SIGQUEUE_MAX	number of queued signals a process can send and have pending	32
_POSIX_SSIZE_MAX	value that can be stored in ssize_t object	32,767
_POSIX_STREAM_MAX	number of standard I/O streams a process can have open at once	8
_POSIX_SYMLINK_MAX	number of bytes in a symbolic link	255
_POSIX_SYMLOOP_MAX	number of symbolic links that can be traversedduring pathname resolution	8
_POSIX_TIMER_MAX	number of timers per process	32
_POSIX_TTY_NAME_MAX	length of a terminal device name, including the terminating null	9
_POSIX_TZNAME_MAX	number of bytes for the name of a time zone	6

Figure 2.8 POSIX.1 minimum values from <limits.h>

    Of these limits and constants, some may be defined in <limits.h>, and others may or may not be defined, depending on certain conditions. We describe the limits and constants that may or may not be defined in Section 2.5.4, when we describe the sysconf, pathconf, and fpathconf functions. The 25 minimum values are shown in Figure 2.8.

    These minimum values do not change from one system to another. They specifythe most restrictivevalues for these features. A conformingPOSIX.1 implementation must provide values that are at least this large. This is why they are called minimums, although their names all contain MAX. Also, to ensure portability, a strictlyconforming application must not require a larger value. We describe what each of these constants refers to as we proceedthrough the text.

A strictlyconforming POSIX application is different from an application that is merelyPOSIX conforming. A POSIX-conforming application uses only interfaces defined in IEEE Standard 1003.1-2008. A strictly conforming POSIX application must meet further restrictions, such as not relying on any undefined behavior, not using any obsolescentinterfaces, and not requiringvalues of constants larger than the minimums shown in Figure 2.8.

Name	Description	Minimum acceptable value
ARG_MAX	maximum length of arguments to exec functions	_POSIX_ARG_MAX
ATEXIT_MAX	maximum number of functions that can be registered with the atexit function	32
CHILD_MAX	maximum number of child processes per real user ID	_POSIX_CHILD_MAX
DELAYTIMER_MAX	maximum number of timer expiration overruns	_POSIX_DELAYTIMER_MAX
HOST_NAME_MAX	maximum length of a host name as returned by gethostname	_POSIX_HOST_NAME_MAX
LOGIN_NAME_MAX	maximum length of a login name	_POSIX_LOGIN_NAME_MAX
OPEN_MAX	one more than the maximum value assigned to a newly created file descriptor 赋予新建文件描述符的最大值+1	_POSIX_OPEN_MAX
PAGESIZE	system memory page size, in bytes	1
RTSIG_MAX	maximum number of real-time signals reserved for application use	_POSIX_RTSIG_MAX
SEM_NSEMS_MAX	maximum number of semaphoresa process can use	_POSIX_SEM_NSEMS_MAX
SEM_VALUE_MAX	maximum value of a semaphore	_POSIX_SEM_VALUE_MAX
SIGQUEUE_MAX	maximum number of signals that can be queued for a process	_POSIX_SIGQUEUE_MAX
STREAM_MAX	maximum number of standard I/O streams a process can have open at once	_POSIX_STREAM_MAX
SYMLOOP_MAX	number of symboliclinks that can be traversedduring pathname resolution	_POSIX_SYMLOOP_MAX
TIMER_MAX	maximum number of timers per process	_POSIX_TIMER_MAX
TTY_NAME_MAX	length of a terminal device name, including the terminating null	_POSIX_TTY_NAME_MAX
TZNAME_MAX	number of bytes for the name of a time zon	_POSIX_TZNAME_MAX

Figure 2.9 POSIX.1 runtime invariant values from <limits.h>

    Unfortunately, some of these invariant minimum values are too small to be of practical use. For example, most UNIX systems today provide far more than 20 open files per process. Also, the minimum limit of 256 for _POSIX_PATH_MAX is too small. Pathnames can exceed this limit. This means that we can’t use the two constants _POSIX_OPEN_MAX and _POSIX_PATH_MAX as array sizes at compile time.

    Each of the 25 invariant minimum values in Figure Figure 2.8 has an associated implementation value whose name is formed by removing the _POSIX_ prefix from the name in Figure 2.8. The names without the leading _POSIX_ were intended to be the actual values that a given implementation supports. (These 25 implementation values are from items 1, 4, 5, and 7 from our list earlier in this section: 2 of the runtime increasable values, 15 of the runtime invariant values, and 7 of the pathname variable values, along with SSIZE_MAX from the numeric values.) The problem is that not all of the 25 implementation values are guaranteed to be defined in the <limits.h> header.

    For example, a particular value may not be included in the header if its actual value for a given process depends on the amount of memory on the system. If the values are not defined in the header, we can’t use them as array bounds at compile time. To determine the actual implementation value at runtime, POSIX.1 decided to provide three functions for us to call—sysconf, pathconf, and fpathconf. There is still a problem, however, because some of the values are defined by POSIX.1 as being possibly ‘‘indeterminate’’ (logically infinite). This means that the value has no practicalupper bound. On Solaris, for example, the number of functions you can register with atexit to be run when a process ends is limited only by the amount of memory on the system. Thus ATEXIT_MAX is considered indeterminateon Solaris. We’ll return to this problem of indeterminate runtime limits in Section 2.5.5.

XSI限制

    The XSI option also defines constants representingimplementation limits. They include:

    1) Minimum values: the five constants in Figure 2.10

    2) Runtime invariantvalues, possibly indeterminate: IOV_MAX and PAGE_SIZE

    The minimum values are listed in Figure 2.10. The last two illustratethe situationin which the POSIX.1 minimums were too small—presumablyto allow for embeddedPOSIX.1 implementations—so symbols with larger minimum values were added for XSI-conforming systems.

Name	Description	Minimum acceptable value	Typical value
NL_LANGMAX	maximum number of bytes in LANG environment variable	14	14
NZERO	default process priority	20	20
_XOPEN_IOV_MAX	maximum number of iovec structures that can be used with readv or writev	16	16
_XOPEN_NAME_MAX	number of bytes in a filename	255	255
_XOPEN_PATH_MAX	number of bytes in a pathname	1024	1024

Figure 2.10 XSI minimum values from <limits.h>

函数sysconfi、pathconf、fpathconf

    We’ve listed variousminimum values that an implementation must support, but how do we find out the limits that a particular system actually supports? As we mentioned earlier, some of these limits might be availableat compile time; others must be determined at runtime. We’ve also mentioned that some limits don’t change in a given system, whereasothers can change because they are associated with a file or directory. The runtime limits are obtained by calling one of the following three functions.            

1. #include<unistd.h>
2.  
3. long sysconf(int name);
4.  
5. long pathconf(constchar*pathname,int name);
6.  
7. long fpathconf(int fd,int name);
8.  
9. All three return: corresponding value if OK,−1 on error (see later)

    The difference between the last two functions is that one takes a pathname as its argument and the other takes a file descriptor argument.

    Figure 2.11 lists the name arguments that sysconf uses to identifysystem limits. Constants beginning with _SC_ are used as arguments to sysconf to identify the runtime limit. Figure 2.12 lists the name arguments that are used by pathconf and fpathconf to identify system limits. Constants beginning with _PC_ are used asarguments to pathconf and fpathconf to identify the runtime limit.

Name of limit	Description	name argument
ARG_MAX	maximum length, in bytes, of arguments to the exec functions	_SC_ARG_MAX
ATEXIT_MAX	maximum number of functions that can be registered with the atexit function	_SC_ATEXIT_MAX
CHILD_MAX	maximum number of processes per real user ID	_SC_CHILD_MAX
clock ticks/second	number of clock ticksper second	_SC_CLK_TCK
COLL_WEIGHTS_MAX	maximum number of weightsthat can be assignedto an entry of the LC_COLLATE order keyword in the locale definition file	_SC_COLL_WEIGHTS_MAX
DELAYTIMER_MAX	maximum number of timer expiration overruns	_SC_DELAYTIMER_MAX
HOST_NAME_MAX	maximum length of a host name as returned by gethostname	_SC_HOST_NAME_MAX
IOV_MAX	maximum number of iovec structures that can be used with readv or writev	_SC_IOV_MAX
LINE_MAX	maximum length of a utility’s input line	_SC_LINE_MAX
LOGIN_NAME_MAX	maximum length of a login name	_SC_LOGIN_NAME_MAX
NGROUPS_MAX	maximum number of simultaneous supplementaryprocess group IDs per process	_SC_NGROUPS_MAX
OPEN_MAX	one more than the maximum value assigned to a newly created file descriptor	_SC_OPEN_MAX
PAGESIZE	system memory page size, in bytes	_SC_PAGESIZE
PAGE_SIZE	system memory page size, in bytes	_SC_PAGE_SIZE
RE_DUP_MAX	number of repeated occurrencesof a basic regular expression permittedby the regexec and regcomp functions when using the interval notation\{m,n\}	_SC_RE_DUP_MAX
RTSIG_MAX	maximum number of real-time signals reservedfor application use	_SC_RTSIG_MAX
SEM_NSEMS_MAX	maximum number of semaphoresa process can use at one time	_SC_SEM_NSEMS_MAX
SEM_VALUE_MAX	maximum value of a semaphore	_SC_SEM_VALUE_MAX
SIGQUEUE_MAX	maximum number of signals that can be queued for a process	_SC_SIGQUEUE_MAX
STREAM_MAX	maximum number of standard I/O streams per process at any given time; if defined, it must have the same value as FOPEN_MAX	_SC_STREAM_MAX
SYMLOOP_MAX	number of symboliclinks that can be traversedduring pathname resolution	_SC_SYMLOOP_MAX
TIMER_MAX	maximum number of timers per process	_SC_TIMER_MAX
TTY_NAME_MAX	length of a terminal device name, including the terminating null	_SC_TTY_NAME_MAX
TZNAME_MAX	maximum number of bytes for a time zone name	_SC_TZNAME_MAX

Figure 2.11 Limits and name arguments to sysconf

 

    We need to look in more detail at the different return values from these three functions.

    1) All three functions return −1 and set errno to EINVAL if the name isn’t one of the appropriate constants. The third column in Figures 2.11 and 2.12 lists the limit constants we’ll deal with throughout the rest of this book.

    2) Some names can return either the value of the variable (a return value ≥ 0) or an indicationthat the value is indeterminate. An indeterminatevalue is indicatedby returning −1 and not changing the value of errno.

    3) The value returned for _SC_CLK_TCK is the number of clock ticks per second, for use with the return values from the times function (Section 8.17).

    Some restrictionsapply to the pathconf pathname argument and the fpathconf fd argument. If any of these restrictionsisn’t met, the results are undefined.

    1) The referenced file for _PC_MAX_CANON and _PC_MAX_INPUT must be a terminalfile.  

    2) The referenced file for _PC_LINK_MAX and _PC_TIMESTAMP_RESOLUTION can be either a file or a directory. If the referenced file is a directory, the return value applies to the directory itself, not to the filename entrieswithin the directory.  

    3) The referenced file for _PC_FILESIZEBITS and _PC_NAME_MAX must be a directory. The return value applies tofilenames within the directory.

    4) The referenced file for _PC_PATH_MAX must be a directory. The value returned is the maximum length of a relativepathname when the specified directory is the working directory. (Unfortunately, this isn’t the real maximum length of an absolute pathname, which is what we want to know. We’ll return to this problem in Section 2.5.5.)

    5) The referenced file for _PC_PIPE_BUF must be a pipe, FIFO, or directory. In the first two cases (pipe or FIFO), the return value is the limit for the referenced pipe or FIFO. For the other case (a directory), the return value is the limit for any FIFO created in that directory. 

    6) The referenced file for _PC_SYMLINK_MAX must be a directory. The value returned is the maximum length of the string that a symbolic link in that directory can contain.

Name of limit	Description	name argument
FILESIZEBITS	minimum number of bits needed to represent, as a signed integer value, the maximum size of a regularfile allowed in the specified directory	_PC_FILESIZEBITS
LINK_MAX	maximum value of a file’s link count	_PC_LINK_MAX
MAX_CANON	maximum number of bytes on a terminal’s canonicalinput queue	_PC_MAX_CANON
MAX_INPUT	number of bytes for which space is available on terminal’s input queue	_PC_MAX_INPUT
NAME_MAX	maximum number of bytes in a filename (does not include a null at end)	_PC_NAME_MAX
PATH_MAX	maximum number of bytes in a relative pathname, including the terminating null	_PC_PATH_MAX
PIPE_BUF	maximum number of bytes that can be written atomicallyto a pipe	_PC_PIPE_BUF
_POSIX_TIMESTAMP_RESOLUTION	resolutionin nanoseconds for file timestamps	_PC_TIMESTAMP_RESOLUTION
SYMLINK_MAX	number of bytes in a symboliclink	_PC_SYMLINK_MAX

Figure 2.12 Limits and name arguments to pathconf and fpathconf

Example

    The awk(1) program shown in Figure 2.13 builds a C program that prints the value of each pathconf and sysconf symbol.

1. # firewaywei 2016年 08月 18日 星期四 21:22:13 CST
2. #!/usr/bin/awk -f
3. BEGIN {
4. printf("#include \"apue.h\"\n")
5. printf("#include <errno.h>\n")
6. printf("#include <limits.h>\n")
7. printf("\n")
8. printf("static void pr_sysconf(char *, int);\n")
9. printf("static void pr_pathconf(char *, char *, int);\n")
10. printf("\n")
11. printf("int\n")
12. printf("main(int argc, char *argv[])\n")
13. printf("{\n")
14. printf("\tif (argc != 2)\n")
15. printf("\t\terr_quit(\"usage: a.out <dirname>\");\n\n")
16. FS="\t+"
17. while(getline <"sysconf.sym">0){
18. printf("#ifdef %s\n", $1)
19. printf("\tprintf(\"%s defined to be %%ld\\n\", (long)%s + 0);\n",
20. $1, $1)
21. printf("#else\n")
22. printf("\tprintf(\"no symbol for %s\\n\");\n", $1)
23. printf("#endif\n")
24. printf("#ifdef %s\n", $2)
25. printf("\tpr_sysconf(\"%s =\", %s);\n", $1, $2)
26. printf("#else\n")
27. printf("\tprintf(\"no symbol for %s\\n\");\n", $2)
28. printf("#endif\n")
29. printf("\n")
30. }
31. close("sysconf.sym")
32. printf("\n\n")
33. while(getline <"pathconf.sym">0){
34. printf("#ifdef %s\n", $1)
35. printf("\tprintf(\"%s defined to be %%ld\\n\", (long)%s+0);\n",
36. $1, $1)
37. printf("#else\n")
38. printf("\tprintf(\"no symbol for %s\\n\");\n", $1)
39. printf("#endif\n")
40. printf("#ifdef %s\n", $2)
41. printf("\tpr_pathconf(\"%s =\", argv[1], %s);\n", $1, $2)
42. printf("#else\n")
43. printf("\tprintf(\"no symbol for %s\\n\");\n", $2)
44. printf("#endif\n")
45. printf("\n")
46. }
47. close("pathconf.sym")
48. exit
49. }
51. END {
52. printf("\texit(0);\n")
53. printf("}\n\n")
54. printf("static void\n")
55. printf("pr_sysconf(char *mesg, int name)\n")
56. printf("{\n")
57. printf("\tlong val;\n\n")
58. printf("\tfputs(mesg, stdout);\n")
59. printf("\terrno = 0;\n")
60. printf("\tif ((val = sysconf(name)) < 0) {\n")
61. printf("\t\tif (errno != 0) {\n")
62. printf("\t\t\tif (errno == EINVAL)\n")
63. printf("\t\t\t\tfputs(\" (not supported)\\n\", stdout);\n")
64. printf("\t\t\telse\n")
65. printf("\t\t\t\terr_sys(\"sysconf error\");\n")
66. printf("\t\t} else {\n")
67. printf("\t\t\tfputs(\" (no limit)\\n\", stdout);\n")
68. printf("\t\t}\n")
69. printf("\t} else {\n")
70. printf("\t\tprintf(\" %%ld\\n\", val);\n")
71. printf("\t}\n")
72. printf("}\n\n")
73. printf("static void\n")
74. printf("pr_pathconf(char *mesg, char *path, int name)\n")
75. printf("{\n")
76. printf("\tlong val;\n")
77. printf("\n")
78. printf("\tfputs(mesg, stdout);\n")
79. printf("\terrno = 0;\n")
80. printf("\tif ((val = pathconf(path, name)) < 0) {\n")
81. printf("\t\tif (errno != 0) {\n")
82. printf("\t\t\tif (errno == EINVAL)\n")
83. printf("\t\t\t\tfputs(\" (not supported)\\n\", stdout);\n")
84. printf("\t\t\telse\n")
85. printf("\t\t\t\terr_sys(\"pathconf error, path = %%s\", path);\n")
86. printf("\t\t} else {\n")
87. printf("\t\t\tfputs(\" (no limit)\\n\", stdout);\n")
88. printf("\t\t}\n")
89. printf("\t} else {\n")
90. printf("\t\tprintf(\" %%ld\\n\", val);\n")
91. printf("\t}\n")
92. printf("}\n")
93. }

Figure 2.13 Build C program to print all supported configuration limits

    The awk program reads two input files—pathconf.sym and sysconf.sym—that contain lists of the limit name and symbol, separated by tabs. All symbols are not defined on every platform, so the awk program surrounds each call to pathconf and sysconf with the necessary #ifdef statements.

    For example, the awk program transformsa line in the input file that looks like

    NAME_MAX     _PC_NAME_MAX

    into the following C code:

1. #ifdef NAME_MAX
2. printf("NAME_MAX defined to be %ld\n",(long)NAME_MAX+0);
3. #else
4. printf("no symbol for NAME_MAX\n");
5. #endif
6. #ifdef _PC_NAME_MAX
7. pr_pathconf("NAME_MAX =", argv[1], _PC_NAME_MAX);
8. #else
9. printf("no symbol for _PC_NAME_MAX\n");
10. #endif

    The program in Figure 2.14, generated by the awk program, prints all these limits, handling the case in which a limit is not defined.

1. /**
2. * 文件名: standards/conf.c
3. * 内容: 由makeconf.awk脚本自动生成
4. * 时间：2016年 08月 19日 星期五 21:10:48 CST
5. * 作者：firewaywei
6. */
7. #include"apue.h"
8. #include<errno.h>
9. #include<limits.h>
11. staticvoid pr_sysconf(char*,int);
12. staticvoid pr_pathconf(char*,char*,int);
14. int
15. main(int argc,char*argv[])
16. {
17. if(argc !=2)
18. err_quit("usage: a.out <dirname>");
20. #ifdef ARG_MAX
21. printf("ARG_MAX defined to be %ld\n",(long)ARG_MAX +0);
22. #else
23. printf("no symbol for ARG_MAX\n");
24. #endif
25. #ifdef _SC_ARG_MAX
26. pr_sysconf("ARG_MAX =", _SC_ARG_MAX);
27. #else
28. printf("no symbol for _SC_ARG_MAX\n");
29. #endif
31. #ifdef ATEXIT_MAX
32. printf("ATEXIT_MAX defined to be %ld\n",(long)ATEXIT_MAX +0);
33. #else
34. printf("no symbol for ATEXIT_MAX\n");
35. #endif
36. #ifdef _SC_ATEXIT_MAX
37. pr_sysconf("ATEXIT_MAX =", _SC_ATEXIT_MAX);
38. #else
39. printf("no symbol for _SC_ATEXIT_MAX\n");
40. #endif
42. #ifdef CHARCLASS_NAME_MAX
43. printf("CHARCLASS_NAME_MAX defined to be %ld\n",(long)CHARCLASS_NAME_MAX +0);
44. #else
45. printf("no symbol for CHARCLASS_NAME_MAX\n");
46. #endif
47. #ifdef _SC_CHARCLASS_NAME_MAX
48. pr_sysconf("CHARCLASS_NAME_MAX =", _SC_CHARCLASS_NAME_MAX);
49. #else
50. printf("no symbol for _SC_CHARCLASS_NAME_MAX\n");
51. #endif
53. #ifdef CHILD_MAX
54. printf("CHILD_MAX defined to be %ld\n",(long)CHILD_MAX +0);
55. #else
56. printf("no symbol for CHILD_MAX\n");
57. #endif
58. #ifdef _SC_CHILD_MAX
59. pr_sysconf("CHILD_MAX =", _SC_CHILD_MAX);
60. #else
61. printf("no symbol for _SC_CHILD_MAX\n");
62. #endif
64. #ifdef CLOCKTICKSPERSECOND /*clock ticks/second*/
65. printf("CLOCKTICKSPERSECOND /*clock ticks/second*/ defined to be %ld\n",(long)CLOCKTICKSPERSECOND /*clock ticks/second*/+0);
66. #else
67. printf("no symbol for CLOCKTICKSPERSECOND /*clock ticks/second*/\n");
68. #endif
69. #ifdef _SC_CLK_TCK
70. pr_sysconf("CLOCKTICKSPERSECOND /*clock ticks/second*/ =", _SC_CLK_TCK);
71. #else
72. printf("no symbol for _SC_CLK_TCK\n");
73. #endif
75. #ifdef COLL_WEIGHTS_MAX
76. printf("COLL_WEIGHTS_MAX defined to be %ld\n",(long)COLL_WEIGHTS_MAX +0);
77. #else
78. printf("no symbol for COLL_WEIGHTS_MAX\n");
79. #endif
80. #ifdef _SC_COLL_WEIGHTS_MAX
81. pr_sysconf("COLL_WEIGHTS_MAX =", _SC_COLL_WEIGHTS_MAX);
82. #else
83. printf("no symbol for _SC_COLL_WEIGHTS_MAX\n");
84. #endif
86. #ifdef DELAYTIMER_MAX
87. printf("DELAYTIMER_MAX defined to be %ld\n",(long)DELAYTIMER_MAX +0);
88. #else
89. printf("no symbol for DELAYTIMER_MAX\n");
90. #endif
91. #ifdef _SC_DELAYTIMER_MAX
92. pr_sysconf("DELAYTIMER_MAX =", _SC_DELAYTIMER_MAX);
93. #else
94. printf("no symbol for _SC_DELAYTIMER_MAX\n");
95. #endif
97. #ifdef HOST_NAME_MAX
98. printf("HOST_NAME_MAX defined to be %ld\n",(long)HOST_NAME_MAX +0);
99. #else
100. printf("no symbol for HOST_NAME_MAX\n");
101. #endif
102. #ifdef _SC_HOST_NAME_MAX
103. pr_sysconf("HOST_NAME_MAX =", _SC_HOST_NAME_MAX);
104. #else
105. printf("no symbol for _SC_HOST_NAME_MAX\n");
106. #endif
108. #ifdef IOV_MAX
109. printf("IOV_MAX defined to be %ld\n",(long)IOV_MAX +0);
110. #else
111. printf("no symbol for IOV_MAX\n");
112. #endif
113. #ifdef _SC_IOV_MAX
114. pr_sysconf("IOV_MAX =", _SC_IOV_MAX);
115. #else
116. printf("no symbol for _SC_IOV_MAX\n");
117. #endif
119. #ifdef LINE_MAX
120. printf("LINE_MAX defined to be %ld\n",(long)LINE_MAX +0);
121. #else
122. printf("no symbol for LINE_MAX\n");
123. #endif
124. #ifdef _SC_LINE_MAX
125. pr_sysconf("LINE_MAX =", _SC_LINE_MAX);
126. #else
127. printf("no symbol for _SC_LINE_MAX\n");
128. #endif
130. #ifdef LOGIN_NAME_MAX
131. printf("LOGIN_NAME_MAX defined to be %ld\n",(long)LOGIN_NAME_MAX +0);
132. #else
133. printf("no symbol for LOGIN_NAME_MAX\n");
134. #endif
135. #ifdef _SC_LOGIN_NAME_MAX
136. pr_sysconf("LOGIN_NAME_MAX =", _SC_LOGIN_NAME_MAX);
137. #else
138. printf("no symbol for _SC_LOGIN_NAME_MAX\n");
139. #endif
141. #ifdef NGROUPS_MAX
142. printf("NGROUPS_MAX defined to be %ld\n",(long)NGROUPS_MAX +0);
143. #else
144. printf("no symbol for NGROUPS_MAX\n");
145. #endif
146. #ifdef _SC_NGROUPS_MAX
147. pr_sysconf("NGROUPS_MAX =", _SC_NGROUPS_MAX);
148. #else
149. printf("no symbol for _SC_NGROUPS_MAX\n");
150. #endif
152. #ifdef OPEN_MAX
153. printf("OPEN_MAX defined to be %ld\n",(long)OPEN_MAX +0);
154. #else
155. printf("no symbol for OPEN_MAX\n");
156. #endif
157. #ifdef _SC_OPEN_MAX
158. pr_sysconf("OPEN_MAX =", _SC_OPEN_MAX);
159. #else
160. printf("no symbol for _SC_OPEN_MAX\n");
161. #endif
163. #ifdef PAGESIZE
164. printf("PAGESIZE defined to be %ld\n",(long)PAGESIZE +0);
165. #else
166. printf("no symbol for PAGESIZE\n");
167. #endif
168. #ifdef _SC_PAGESIZE
169. pr_sysconf("PAGESIZE =", _SC_PAGESIZE);
170. #else
171. printf("no symbol for _SC_PAGESIZE\n");
172. #endif
174. #ifdef PAGE_SIZE
175. printf("PAGE_SIZE defined to be %ld\n",(long)PAGE_SIZE +0);
176. #else
177. printf("no symbol for PAGE_SIZE\n");
178. #endif
179. #ifdef _SC_PAGE_SIZE
180. pr_sysconf("PAGE_SIZE =", _SC_PAGE_SIZE);
181. #else
182. printf("no symbol for _SC_PAGE_SIZE\n");
183. #endif
185. #ifdef RE_DUP_MAX
186. printf("RE_DUP_MAX defined to be %ld\n",(long)RE_DUP_MAX +0);
187. #else
188. printf("no symbol for RE_DUP_MAX\n");
189. #endif
190. #ifdef _SC_RE_DUP_MAX
191. pr_sysconf("RE_DUP_MAX =", _SC_RE_DUP_MAX);
192. #else
193. printf("no symbol for _SC_RE_DUP_MAX\n");
194. #endif
196. #ifdef RTSIG_MAX
197. printf("RTSIG_MAX defined to be %ld\n",(long)RTSIG_MAX +0);
198. #else
199. printf("no symbol for RTSIG_MAX\n");
200. #endif
201. #ifdef _SC_RTSIG_MAX
202. pr_sysconf("RTSIG_MAX =", _SC_RTSIG_MAX);
203. #else
204. printf("no symbol for _SC_RTSIG_MAX\n");
205. #endif
207. #ifdef SEM_NSEMS_MAX
208. printf("SEM_NSEMS_MAX defined to be %ld\n",(long)SEM_NSEMS_MAX +0);
209. #else
210. printf("no symbol for SEM_NSEMS_MAX\n");
211. #endif
212. #ifdef _SC_SEM_NSEMS_MAX
213. pr_sysconf("SEM_NSEMS_MAX =", _SC_SEM_NSEMS_MAX);
214. #else
215. printf("no symbol for _SC_SEM_NSEMS_MAX\n");
216. #endif
218. #ifdef SEM_VALUE_MAX
219. printf("SEM_VALUE_MAX defined to be %ld\n",(long)SEM_VALUE_MAX +0);
220. #else
221. printf("no symbol for SEM_VALUE_MAX\n");
222. #endif
223. #ifdef _SC_SEM_VALUE_MAX
224. pr_sysconf("SEM_VALUE_MAX =", _SC_SEM_VALUE_MAX);
225. #else
226. printf("no symbol for _SC_SEM_VALUE_MAX\n");
227. #endif
229. #ifdef SIGQUEUE_MAX
230. printf("SIGQUEUE_MAX defined to be %ld\n",(long)SIGQUEUE_MAX +0);
231. #else
232. printf("no symbol for SIGQUEUE_MAX\n");
233. #endif
234. #ifdef _SC_SIGQUEUE_MAX
235. pr_sysconf("SIGQUEUE_MAX =", _SC_SIGQUEUE_MAX);
236. #else
237. printf("no symbol for _SC_SIGQUEUE_MAX\n");
238. #endif
240. #ifdef STREAM_MAX
241. printf("STREAM_MAX defined to be %ld\n",(long)STREAM_MAX +0);
242. #else
243. printf("no symbol for STREAM_MAX\n");
244. #endif
245. #ifdef _SC_STREAM_MAX
246. pr_sysconf("STREAM_MAX =", _SC_STREAM_MAX);
247. #else
248. printf("no symbol for _SC_STREAM_MAX\n");
249. #endif
251. #ifdef SYMLOOP_MAX
252. printf("SYMLOOP_MAX defined to be %ld\n",(long)SYMLOOP_MAX +0);
253. #else
254. printf("no symbol for SYMLOOP_MAX\n");
255. #endif
256. #ifdef _SC_SYMLOOP_MAX
257. pr_sysconf("SYMLOOP_MAX =", _SC_SYMLOOP_MAX);
258. #else
259. printf("no symbol for _SC_SYMLOOP_MAX\n");
260. #endif
262. #ifdef TIMER_MAX
263. printf("TIMER_MAX defined to be %ld\n",(long)TIMER_MAX +0);
264. #else
265. printf("no symbol for TIMER_MAX\n");
266. #endif
267. #ifdef _SC_TIMER_MAX
268. pr_sysconf("TIMER_MAX =", _SC_TIMER_MAX);
269. #else
270. printf("no symbol for _SC_TIMER_MAX\n");
271. #endif
273. #ifdef TTY_NAME_MAX
274. printf("TTY_NAME_MAX defined to be %ld\n",(long)TTY_NAME_MAX +0);
275. #else
276. printf("no symbol for TTY_NAME_MAX\n");
277. #endif
278. #ifdef _SC_TTY_NAME_MAX
279. pr_sysconf("TTY_NAME_MAX =", _SC_TTY_NAME_MAX);
280. #else
281. printf("no symbol for _SC_TTY_NAME_MAX\n");
282. #endif
284. #ifdef TZNAME_MAX
285. printf("TZNAME_MAX defined to be %ld\n",(long)TZNAME_MAX +0);
286. #else
287. printf("no symbol for TZNAME_MAX\n");
288. #endif
289. #ifdef _SC_TZNAME_MAX
290. pr_sysconf("TZNAME_MAX =", _SC_TZNAME_MAX);
291. #else
292. printf("no symbol for _SC_TZNAME_MAX\n");
293. #endif
297. #ifdef FILESIZEBITS
298. printf("FILESIZEBITS defined to be %ld\n",(long)FILESIZEBITS+0);
299. #else
300. printf("no symbol for FILESIZEBITS\n");
301. #endif
302. #ifdef _PC_FILESIZEBITS
303. pr_pathconf("FILESIZEBITS =", argv[1], _PC_FILESIZEBITS);
304. #else
305. printf("no symbol for _PC_FILESIZEBITS\n");
306. #endif
308. #ifdef LINK_MAX
309. printf("LINK_MAX defined to be %ld\n",(long)LINK_MAX+0);
310. #else
311. printf("no symbol for LINK_MAX\n");
312. #endif
313. #ifdef _PC_LINK_MAX
314. pr_pathconf("LINK_MAX =", argv[1], _PC_LINK_MAX);
315. #else
316. printf("no symbol for _PC_LINK_MAX\n");
317. #endif
319. #ifdef MAX_CANON
320. printf("MAX_CANON defined to be %ld\n",(long)MAX_CANON+0);
321. #else
322. printf("no symbol for MAX_CANON\n");
323. #endif
324. #ifdef _PC_MAX_CANON
325. pr_pathconf("MAX_CANON =", argv[1], _PC_MAX_CANON);
326. #else
327. printf("no symbol for _PC_MAX_CANON\n");
328. #endif
330. #ifdef MAX_INPUT
331. printf("MAX_INPUT defined to be %ld\n",(long)MAX_INPUT+0);
332. #else
333. printf("no symbol for MAX_INPUT\n");
334. #endif
335. #ifdef _PC_MAX_INPUT
336. pr_pathconf("MAX_INPUT =", argv[1], _PC_MAX_INPUT);
337. #else
338. printf("no symbol for _PC_MAX_INPUT\n");
339. #endif
341. #ifdef NAME_MAX
342. printf("NAME_MAX defined to be %ld\n",(long)NAME_MAX+0);
343. #else
344. printf("no symbol for NAME_MAX\n");
345. #endif
346. #ifdef _PC_NAME_MAX
347. pr_pathconf("NAME_MAX =", argv[1], _PC_NAME_MAX);
348. #else
349. printf("no symbol for _PC_NAME_MAX\n");
350. #endif
352. #ifdef PATH_MAX
353. printf("PATH_MAX defined to be %ld\n",(long)PATH_MAX+0);
354. #else
355. printf("no symbol for PATH_MAX\n");
356. #endif
357. #ifdef _PC_PATH_MAX
358. pr_pathconf("PATH_MAX =", argv[1], _PC_PATH_MAX);
359. #else
360. printf("no symbol for _PC_PATH_MAX\n");
361. #endif
363. #ifdef PIPE_BUF
364. printf("PIPE_BUF defined to be %ld\n",(long)PIPE_BUF+0);
365. #else
366. printf("no symbol for PIPE_BUF\n");
367. #endif
368. #ifdef _PC_PIPE_BUF
369. pr_pathconf("PIPE_BUF =", argv[1], _PC_PIPE_BUF);
370. #else
371. printf("no symbol for _PC_PIPE_BUF\n");
372. #endif
374. #ifdef SYMLINK_MAX
375. printf("SYMLINK_MAX defined to be %ld\n",(long)SYMLINK_MAX+0);
376. #else
377. printf("no symbol for SYMLINK_MAX\n");
378. #endif
379. #ifdef _PC_SYMLINK_MAX
380. pr_pathconf("SYMLINK_MAX =", argv[1], _PC_SYMLINK_MAX);
381. #else
382. printf("no symbol for _PC_SYMLINK_MAX\n");
383. #endif
385. #ifdef _POSIX_TIMESTAMP_RESOLUTION
386. printf("_POSIX_TIMESTAMP_RESOLUTION defined to be %ld\n",(long)_POSIX_TIMESTAMP_RESOLUTION+0);
387. #else
388. printf("no symbol for _POSIX_TIMESTAMP_RESOLUTION\n");
389. #endif
390. #ifdef _PC_TIMESTAMP_RESOLUTION
391. pr_pathconf("_POSIX_TIMESTAMP_RESOLUTION =", argv[1], _PC_TIMESTAMP_RESOLUTION);
392. #else
393. printf("no symbol for _PC_TIMESTAMP_RESOLUTION\n");
394. #endif
396. exit(0);
397. }
399. staticvoid
400. pr_sysconf(char*mesg,int name)
401. {
402. long val;
404. fputs(mesg, stdout);
405. errno =0;
406. if((val = sysconf(name))<0){
407. if(errno !=0){
408. if(errno == EINVAL)
409. fputs(" (not supported)\n", stdout);
410. else
411. err_sys("sysconf error");
412. }else{
413. fputs(" (no limit)\n", stdout);
414. }
415. }else{
416. printf(" %ld\n", val);
417. }
418. }
420. staticvoid
421. pr_pathconf(char*mesg,char*path,int name)
422. {
423. long val;
425. fputs(mesg, stdout);
426. errno =0;
427. if((val = pathconf(path, name))<0){
428. if(errno !=0){
429. if(errno == EINVAL)
430. fputs(" (not supported)\n", stdout);
431. else
432. err_sys("pathconf error, path = %s", path);
433. }else{
434. fputs(" (no limit)\n", stdout);
435. }
436. }else{
437. printf(" %ld\n", val);
438. }
439. }

Figure 2.14 Print all possible sysconf and pathconf values

    Figure 2.15 summarizes the results from Figure 2.14 for the four systems we discuss in this book. The entry‘‘no symbol’’ means that the system doesn’t provide a corresponding_SC or _PC symbol to query the value of the constant. Thus the limit is undefined in this case. In contrast, the entry ‘‘unsupported’’ means that the symbol is defined by the system but unrecognized by the sysconf or pathconf functions. The entry ‘‘no limit’’ means that the system defines no limit for the constant, but this doesn’t mean that the limit is infinite; it just means that the limit is indeterminite.

				Solaris 10
Limit	FreeBSD 8.0	Linux 3.2.0	Mac OS X  10.6.8	UFS file system	PCFS file system
ARG_MAX	262,144	2,097,152	262,144	2,096,640	2,096,640
ATEXIT_MAX	32	2,147,483,647	2,147,483,647	no limit	no limit
CHARCLASS_NAME_MAX	no symbol	2,048	14	14	14
CHILD_MAX	1,760	47,211	266	8,021	8,021
clock ticks/second	128	100	100	100	100
COLL_WEIGHTS_MAX	0	255	2	10	10
FILESIZEBITS	64	64	64	41	unsupported
HOST_NAME_MAX	255	64	255	255	255
IOV_MAX	1,024	1,024	1024	16	16
LINE_MAX	2,048	2,048	2,048	2,048	2,048
LINK_MAX	32,767	65,000	32,767	32,767	1
LOGIN_NAME_MAX	17	256	255	9	9
MAX_CANON	255	255	1,024	256	256
MAX_INPUT	255	255	1,024	512	512
NAME_MAX	255	255	255	255	8
NGROUPS_MAX	1,023	65,536	16	16	16
OPEN_MAX	3,520	1,024	256	256	256
PAGESIZE	4,096	4,096	4,096	8,192	8,192
PAGE_SIZE	4,096	4,096	4,096	8,192	8,192
PATH_MAX	1,024	4,096	1,024	1,024	1,024
PIPE_BUF	512	4,096	512	5,120	5,120
RE_DUP_MAX	255	32,767	255	255	255
STREAM_MAX	3,520	16	20	256	256
SYMLINK_MAX	1,024	no limit	255	1,024	1,024
SYMLOOP_MAX	32	no limit	32	20	20
TTY_NAME_MAX	255	32	255	128	128
TZNAME_MAX	255	6	255	no limit	no limit

Figure 2.15 Examples of configuration limits

Beware that some limits are reported incorrectly. For example, on Linux, SYMLOOP_MAX is reportedly unlimited, but an examination of the source code revealsthat there is actually a hard-coded limit of 40 for the number of consecutivesymbolic links traversedin the absence ofa loop (see the follow_link function in fs/namei.c).     
    Another potential source of inaccuracyin Linux is that the pathconf and fpathconf functions are implemented in the C library. The configuration limits returned by these functions depend on the underlyingfile system type, so if your file system is unknown to the C library, the functions return an educatedguess.

我们在linux平台上， 对Figure 2.14源码进行gcc -E conf.c -o conf.i, 看一下预处理以后的文件代码：

2. # ......
3. staticvoid pr_sysconf(char*,int);
4. staticvoid pr_pathconf(char*,char*,int);
6. int
7. main(int argc,char*argv[])
8. {
9. if(argc !=2)
10. err_quit("usage: a.out <dirname>");
15. printf("no symbol for ARG_MAX\n");
18. pr_sysconf("ARG_MAX =", _SC_ARG_MAX);
26. printf("no symbol for ATEXIT_MAX\n");
29. pr_sysconf("ATEXIT_MAX =", _SC_ATEXIT_MAX);
35. printf("CHARCLASS_NAME_MAX defined to be %ld\n",(long)2048+0);
40. pr_sysconf("CHARCLASS_NAME_MAX =", _SC_CHARCLASS_NAME_MAX);
48. printf("no symbol for CHILD_MAX\n");
51. pr_sysconf("CHILD_MAX =", _SC_CHILD_MAX);
59. printf("no symbol for CLOCKTICKSPERSECOND /*clock ticks/second*/\n");
62. pr_sysconf("CLOCKTICKSPERSECOND /*clock ticks/second*/ =", _SC_CLK_TCK);
68. printf("COLL_WEIGHTS_MAX defined to be %ld\n",(long)255+0);
73. pr_sysconf("COLL_WEIGHTS_MAX =", _SC_COLL_WEIGHTS_MAX);
79. printf("DELAYTIMER_MAX defined to be %ld\n",(long)2147483647+0);
84. pr_sysconf("DELAYTIMER_MAX =", _SC_DELAYTIMER_MAX);
90. printf("HOST_NAME_MAX defined to be %ld\n",(long)64+0);
95. pr_sysconf("HOST_NAME_MAX =", _SC_HOST_NAME_MAX);
101. printf("IOV_MAX defined to be %ld\n",(long)1024+0);
106. pr_sysconf("IOV_MAX =", _SC_IOV_MAX);
112. printf("LINE_MAX defined to be %ld\n",(long)2048+0);
117. pr_sysconf("LINE_MAX =", _SC_LINE_MAX);
123. printf("LOGIN_NAME_MAX defined to be %ld\n",(long)256+0);
128. pr_sysconf("LOGIN_NAME_MAX =", _SC_LOGIN_NAME_MAX);
134. printf("NGROUPS_MAX defined to be %ld\n",(long)65536+0);
139. pr_sysconf("NGROUPS_MAX =", _SC_NGROUPS_MAX);
147. printf("no symbol for OPEN_MAX\n");
150. pr_sysconf("OPEN_MAX =", _SC_OPEN_MAX);
158. printf("no symbol for PAGESIZE\n");
161. pr_sysconf("PAGESIZE =", _SC_PAGESIZE);
169. printf("no symbol for PAGE_SIZE\n");
172. pr_sysconf("PAGE_SIZE =", _SC_PAGESIZE);
178. printf("RE_DUP_MAX defined to be %ld\n",(long)(0x7fff)+0);
183. pr_sysconf("RE_DUP_MAX =", _SC_RE_DUP_MAX);
189. printf("RTSIG_MAX defined to be %ld\n",(long)32+0);
194. pr_sysconf("RTSIG_MAX =", _SC_RTSIG_MAX);
202. printf("no symbol for SEM_NSEMS_MAX\n");
205. pr_sysconf("SEM_NSEMS_MAX =", _SC_SEM_NSEMS_MAX);
211. printf("SEM_VALUE_MAX defined to be %ld\n",(long)(2147483647)+0);
216. pr_sysconf("SEM_VALUE_MAX =", _SC_SEM_VALUE_MAX);
224. printf("no symbol for SIGQUEUE_MAX\n");
227. pr_sysconf("SIGQUEUE_MAX =", _SC_SIGQUEUE_MAX);
235. printf("no symbol for STREAM_MAX\n");
238. pr_sysconf("STREAM_MAX =", _SC_STREAM_MAX);
246. printf("no symbol for SYMLOOP_MAX\n");
249. pr_sysconf("SYMLOOP_MAX =", _SC_SYMLOOP_MAX);
257. printf("no symbol for TIMER_MAX\n");
260. pr_sysconf("TIMER_MAX =", _SC_TIMER_MAX);
266. printf("TTY_NAME_MAX defined to be %ld\n",(long)32+0);
271. pr_sysconf("TTY_NAME_MAX =", _SC_TTY_NAME_MAX);
279. printf("no symbol for TZNAME_MAX\n");
282. pr_sysconf("TZNAME_MAX =", _SC_TZNAME_MAX);
283. # 294 "conf.c"
284. printf("no symbol for FILESIZEBITS\n");
287. pr_pathconf("FILESIZEBITS =", argv[1], _PC_FILESIZEBITS);
295. printf("no symbol for LINK_MAX\n");
298. pr_pathconf("LINK_MAX =", argv[1], _PC_LINK_MAX);
304. printf("MAX_CANON defined to be %ld\n",(long)255+0);
309. pr_pathconf("MAX_CANON =", argv[1], _PC_MAX_CANON);
315. printf("MAX_INPUT defined to be %ld\n",(long)255+0);
320. pr_pathconf("MAX_INPUT =", argv[1], _PC_MAX_INPUT);
326. printf("NAME_MAX defined to be %ld\n",(long)255+0);
331. pr_pathconf("NAME_MAX =", argv[1], _PC_NAME_MAX);
337. printf("PATH_MAX defined to be %ld\n",(long)4096+0);
342. pr_pathconf("PATH_MAX =", argv[1], _PC_PATH_MAX);
348. printf("PIPE_BUF defined to be %ld\n",(long)4096+0);
353. pr_pathconf("PIPE_BUF =", argv[1], _PC_PIPE_BUF);
361. printf("no symbol for SYMLINK_MAX\n");
364. pr_pathconf("SYMLINK_MAX =", argv[1], _PC_SYMLINK_MAX);
372. printf("no symbol for _POSIX_TIMESTAMP_RESOLUTION\n");
377. printf("no symbol for _PC_TIMESTAMP_RESOLUTION\n");
380. exit(0);
381. }
383. staticvoid
384. pr_sysconf(char*mesg,int name)
385. {
386. long val;
388. fputs(mesg, stdout);
389. (*__errno_location ())=0;
390. if((val = sysconf(name))<0){
391. if((*__errno_location ())!=0){
392. if((*__errno_location ())==22)
393. fputs(" (not supported)\n", stdout);
394. else
395. err_sys("sysconf error");
396. }else{
397. fputs(" (no limit)\n", stdout);
398. }
399. }else{
400. printf(" %ld\n", val);
401. }
402. }
404. staticvoid
405. pr_pathconf(char*mesg,char*path,int name)
406. {
407. long val;
409. fputs(mesg, stdout);
410. (*__errno_location ())=0;
411. if((val = pathconf(path, name))<0){
412. if((*__errno_location ())!=0){
413. if((*__errno_location ())==22)
414. fputs(" (not supported)\n", stdout);
415. else
416. err_sys("pathconf error, path = %s", path);
417. }else{
418. fputs(" (no limit)\n", stdout);
419. }
420. }else{
421. printf(" %ld\n", val);
422. }
423. }

    结果如下： 

no symbol for ARG_MAX
ARG_MAX = 2097152
no symbol for ATEXIT_MAX
ATEXIT_MAX = 2147483647
CHARCLASS_NAME_MAX defined to be 2048
CHARCLASS_NAME_MAX = 2048
no symbol for CHILD_MAX
CHILD_MAX = 7884
no symbol for CLOCKTICKSPERSECOND /*clock ticks/second*/
CLOCKTICKSPERSECOND /*clock ticks/second*/ = 100
COLL_WEIGHTS_MAX defined to be 255
COLL_WEIGHTS_MAX = 255
DELAYTIMER_MAX defined to be 2147483647
DELAYTIMER_MAX = 2147483647
HOST_NAME_MAX defined to be 64
HOST_NAME_MAX = 64
IOV_MAX defined to be 1024
IOV_MAX = 1024
LINE_MAX defined to be 2048
LINE_MAX = 2048
LOGIN_NAME_MAX defined to be 256
LOGIN_NAME_MAX = 256
NGROUPS_MAX defined to be 65536
NGROUPS_MAX = 65536
no symbol for OPEN_MAX
OPEN_MAX = 1024
no symbol for PAGESIZE
PAGESIZE = 4096
no symbol for PAGE_SIZE
PAGE_SIZE = 4096
RE_DUP_MAX defined to be 32767
RE_DUP_MAX = 32767
RTSIG_MAX defined to be 32
RTSIG_MAX = 32
no symbol for SEM_NSEMS_MAX
SEM_NSEMS_MAX = (no limit)
SEM_VALUE_MAX defined to be 2147483647
SEM_VALUE_MAX = 2147483647
no symbol for SIGQUEUE_MAX
SIGQUEUE_MAX = 7884
no symbol for STREAM_MAX
STREAM_MAX = 16
no symbol for SYMLOOP_MAX
SYMLOOP_MAX = (no limit)
no symbol for TIMER_MAX
TIMER_MAX = (no limit)
TTY_NAME_MAX defined to be 32
TTY_NAME_MAX = 32
no symbol for TZNAME_MAX
TZNAME_MAX = 6
no symbol for FILESIZEBITS
FILESIZEBITS = 64
no symbol for LINK_MAX
LINK_MAX = 65000
MAX_CANON defined to be 255
MAX_CANON = 255
MAX_INPUT defined to be 255
MAX_INPUT = 255
NAME_MAX defined to be 255
NAME_MAX = 255
PATH_MAX defined to be 4096
PATH_MAX = 4096
PIPE_BUF defined to be 4096
PIPE_BUF = 4096
no symbol for SYMLINK_MAX
SYMLINK_MAX = (no limit)
no symbol for _POSIX_TIMESTAMP_RESOLUTION
no symbol for _PC_TIMESTAMP_RESOLUTION

不确定的运行时限制

    We mentioned that some of the limits can be indeterminate. The problem we encounter is that if these limits aren’t defined in the <limits.h> header, we can’t use them at compile time. But they might not be defined at runtime if their value is indeterminate! Let’s look at two specific cases: allocating storage for a pathname and determining the number of file descriptors.

Pathnames

    Many programs need to allocatestorage for a pathname. Typically, the storage has been allocated at compile time, and variousmagic numbers—few of which are the correct value — have been used by different programs as the array size: 256, 512, 1024, or the standard I/O constant BUFSIZ. The 4.3BSD constant MAXPATHLEN in the header <sys/param.h> is the correct value, but many 4.3BSD applications didn’t use it.

    POSIX.1 tries to help with the PATH_MAX value, but if this value is indeterminate, we’re still out of luck. Figure 2.16 shows a function that we’ll use throughoutthis text to allocate storage dynamically for a pathname

1. /**
2. * 文件名: lib/openmax.c
3. * 内容：
4. * 时间: 2016年 08月 19日 星期五 15:23:51 CST
5. * 作者：firewaywei
6. *
7. */
8. #include"apue.h"
9. #include<errno.h>
10. #include<limits.h>
12. #ifdef OPEN_MAX
13. staticlong openmax = OPEN_MAX;
14. #else
15. staticlong openmax =0;
16. #endif
18. /*
19. * If OPEN_MAX is indeterminate, this might be inadequate.
20. */
21. #define OPEN_MAX_GUESS 256
23. long
24. open_max(void)
25. {
26. if(openmax ==0)
27. {/* first time through */
28. errno =0;
29. if((openmax = sysconf(_SC_OPEN_MAX))<0)
30. {
31. if(errno ==0)
32. {
33. openmax = OPEN_MAX_GUESS;/* it's indeterminate */
34. }
35. else
36. {
37. err_sys("sysconf error for _SC_OPEN_MAX");
38. }
39. }
40. }
41. return(openmax);
42. }

 Figure 2.16 Dynamically allocate space for a pathname

     If the constant PATH_MAX is defined in <limits.h>, then we’re all set. If it’s not, then we need to call pathconf. The value returned by pathconf is the maximum size of a relative pathname when the first argument is the working directory, so we specify the root as the first argument and add 1 to the result. If pathconf indicates that PATH_MAX is indeterminate, we have to puntand just guess a value. 

    Versions of POSIX.1 priorto 2001 were unclear as to whether PATH_MAX included a null byte at the end of the pathname. If the operating system implementation conforms to one of these prior versions and doesn’t conform to any version of the Single UNIX Specification (which does require the terminating null byte to be included), we need to add 1 to the amountof memory we allocate for a pathname, just to be on the safe side. 

    The correct way to handle the case of an indeterminate result depends on how the allocated space is being used. If we are allocating space for a call to getcwd, for example — to return the absolute pathname of the current working directory; see Section 4.23—and if the allocated space is too small, an error is returned and errno is set to ERANGE. We could then increase the allocated space by calling realloc (see Section 7.8 and Exercise 4.16) and try again. We could keep doing this until the call to getcwd succeeded

Maximum Number of Open Files

    A common sequence of code in a daemon process — a process that runs in the background, not connected to a terminal—is one that closes all open files. Some programs have the following code sequence, assuming the constant NOFILE was defined in the <sys/param.h> header:

1. #include<sys/param.h>
3. for(i =0; i < NOFILE; i++)
4. {
5. close(i);
6. }

    Other programs use the constant _NFILE that some versions of <stdio.h> provide as the upper limit. Some hard code the upper limit as 20. However, none of these approachesis portable.   

    We would hope to use the POSIX.1 value OPEN_MAX to determine this value portably, but if the value is indeterminate, we still have a problem. If we wrote the following code and if OPEN_MAX was indeterminate, the loop would never execute, since sysconf would return −1:

1. #include<unistd.h>
3. for(i =0; i < sysconf(_SC_OPEN_MAX); i++)
4. {
5. close(i);
6. }

    Our best option in this case is just to close all descriptors up to some arbitrarylimit — say, 256. We show this technique in Figure 2.17. As with our pathname example, this strategy is not guaranteed to work for all cases, but it’s the best we can do without using a more exoticapproach.

1. /**
2. * 文件名: lib/openmax.c
3. * 内容：
4. * 时间: 2016年 08月 19日 星期五 15:23:51 CST
5. * 作者：firewaywei
6. *
7. */
8. #include"apue.h"
9. #include<errno.h>
10. #include<limits.h>
12. #ifdef OPEN_MAX
13. staticlong openmax = OPEN_MAX;
14. #else
15. staticlong openmax =0;
16. #endif
18. /*
19. * If OPEN_MAX is indeterminate, this might be inadequate.
20. */
21. #define OPEN_MAX_GUESS 256
23. long
24. open_max(void)
25. {
26. if(openmax ==0)
27. {/* first time through */
28. errno =0;
29. if((openmax = sysconf(_SC_OPEN_MAX))<0)
30. {
31. if(errno ==0)
32. {
33. openmax = OPEN_MAX_GUESS;/* it's indeterminate */
34. }
35. else
36. {
37. err_sys("sysconf error for _SC_OPEN_MAX");
38. }
39. }
40. }
41. return(openmax);
42. }

Figure 2.17 Determine the number of file descriptors

    We might be tempted to call close until we get an error return, but the error return from close (EBADF) doesn’t distinguish between an invalid descriptor and a descriptor that wasn’t open. If we tried this technique and descriptor 9 was not open but descriptor 10 was, we would stop on 9 and never close 10. The dup function (Section 3.12) does return a specific error when OPEN_MAX is exceeded, but duplicatinga descriptor a couple of hundred times is an extremeway to determine this value. 

    Some implementations will return LONG_MAX for limit values that are effectivelyunlimited. Such is the case with the Linux limit for ATEXIT_MAX (see Figure 2.15). This isn’t a good idea, because it can cause programs to behave badly.

    For example, we can use the ulimit command built into the Bourne-again shell to change the maximum number of files our processes can have open at one time. This generally requires special (superuser) privilegesif the limit is to be effectivelyunlimited. But once set to infinite, sysconf will report LONG_MAX as the limit for OPEN_MAX. A program that relies on this value as the upper bound of file descriptors to close, as shown in Figure 2.17, will waste a lot of time trying to close 2,147,483,647 file descriptors, most of which aren’t even in use. 

    Systems that support the XSI option in the Single UNIX Specification will provide the getrlimit(2) function (Section 7.11). It can be used to return the maximum number of descriptors that a process can have open. With it, we can detectthat there is no configured upper bound to the number of open files our processes can open, so we can avoid this problem.

The OPEN_MAX value is called runtime invariant by POSIX, meaning that its value should not change during the lifetime of a process. But on systems that support the XSI option, we can call the setrlimit(2) function (Section 7.11) to change this value for a running process. (This value can also be changed from the C shell with the limit command, and from the Bourne, Bourne-again, Debian Almquist, and Korn shells with the ulimit command.) If our system supports this functionality, we could change the function in Figure 2.17 to call sysconf every time it is called, not just the first time.

选项

    We saw the list of POSIX.1 options in Figure 2.5 and discussed XSI option groups in Section 2.2.3. If we are to write portable applications that depend on any of these optionally supported features, we need a portable way to determine whether an implementation supports a given option.

    Just as with limits (Section 2.5), POSIX.1 defines three ways to do this.

    1) Compile-time options are defined in <unistd.h>. 

    2) Runtime options that are not associated with a file or a directory are identified with the sysconf function. 

    3) Runtime options that are associated with a file or a directory are discovered by calling either the pathconf or the fpathconf function.

    The options include the symbols listed in the third column of Figure 2.5, as well as the symbols listed in Figures 2.19 and 2.18. If the symbolic constant is not defined, we must use sysconf, pathconf, or fpathconf to determine whether the option is supported. In this case, the name argument to the function is formed by replacing the _POSIX at the beginning of the symbol with _SC or _PC. For constants that begin with _XOPEN, the name argument is formed by prepending the string with _SC or _PC. For example, if the constant _POSIX_RAW_SOCKETS is undefined, we can call sysconf with the name argument set to _SC_RAW_SOCKETS to determine whether the platform supports the raw sockets option. If the constant _XOPEN_UNIX is undefined, we can call sysconf with the name argument set to _SC_XOPEN_UNIX to determine whether the platform supports the XSI option interfaces.

    For each option, we have three possibilities for a platform’s support status.

    1) If the symbolic constant is either undefined or defined to have the value −1, then the corresponding option is unsupported by the platform at compile time. It is possible to run an old application on a newer system where the option is supported, so a runtime check might indicate the option is supported even though the option wasn’t supported at the time the application was compiled. 

    2) If the symbolic constant is defined to be greater than zero, then the corresponding option is supported. 

    3) If the symbolic constant is defined to be equal to zero, then we must call sysconf, pathconf, or fpathconf to determine whether the option is supported

    The symbolic constants used with pathconf and fpathconf are summarized in Figure 2.18. Figure 2.19 summarizes the nonobsoleteoptions and their symbolic constants that can be used with sysconf, in addition to those listed in Figure 2.5. Note that we omitoptions associated with utility commands. 

Name of option	Indicates ...	name argument
_POSIX_CHOWN_RESTRICTED	whether use of chown is restricted	_PC_CHOWN_RESTRICTED
_POSIX_NO_TRUNC	whether filenames longer than NAME_MAX	_PC_NO_TRUNC
_POSIX_VDISABLE	if defined, terminal special characters can be disabled with this value	_PC_VDISABLE
_POSIX_ASYNC_IO	whether asynchronous I/O can be used with the associated file	_PC_ASYNC_IO
_POSIX_PRIO_IO	whether prioritized I/O can be used with the associated file	_PC_PRIO_IO
_POSIX_SYNC_IO	whether synchronized I/O can be used with the associated file	_PC_SYNC_IO
_POSIX2_SYMLINKS	whether symbolic links are supported in the directory	_PC_2_SYMLINKS

Figure 2.18  Options and name arguments to pathconf and fpathconf

Name of option	Indicates ...	name argument
_POSIX_ASYNCHRONOUS_IO	whether the implementation supports POSIX asynchronous I/O	_SC_ASYNCHRONOUS_IO
_POSIX_BARRIERS	whether the implementation supports barriers	_SC_BARRIERS
_POSIX_CLOCK_SELECTION	whether the implementation supports clock selection	_SC_CLOCK_SELECTION
_POSIX_JOB_CONTROL	whether the implementation supports job control	_SC_JOB_CONTROL
_POSIX_MAPPED_FILES	whether the implementation supports memory-mapped files	_SC_MAPPED_FILES
_POSIX_MEMORY_PROTECTION	whether the implementation supports memory protection	_SC_MEMORY_PROTECTION
_POSIX_READER_WRITER_LOCKS	whether the implementation supports reader–writer locks	_SC_READER_WRITER_LOCKS
_POSIX_REALTIME_SIGNALS	whether the implementation supports real-time signals	_SC_REALTIME_SIGNALS
_POSIX_SAVED_IDS	whether the implementation supports the saved set-user-ID and the saved set-group-ID	_SC_SAVED_IDS
_POSIX_SEMAPHORES	whether the implementation supports POSIX semaphores	_SC_SEMAPHORES
_POSIX_SHELL	whether the implementation supports the POSIX shell	_SC_SHELL
_POSIX_SPIN_LOCKS	whether the implementation supports spin locks	_SC_SPIN_LOCKS
_POSIX_THREAD_SAFE_FUNCTIONS	whether the implementation supports thread-safe functions	_SC_THREAD_SAFE_FUNCTIONS
_POSIX_THREADS	whether the implementation supports threads	_SC_THREADS
_POSIX_TIMEOUTS	whether the implementation supports timeout-based variants of selected functions	_SC_TIMEOUTS
_POSIX_TIMERS	whether the implementation supports timers	_SC_TIMERS
_POSIX_VERSION	the POSIX.1 version	_SC_VERSION
_XOPEN_CRYPT	whether the implementation supports the XSI encryption option group	_SC_XOPEN_CRYPT
_XOPEN_REALTIME	whether the implementation supports the XSI real-time option group	_SC_XOPEN_REALTIME
_XOPEN_REALTIME_THREADS	whether the implementation supports the XSI real-time threads option group	_SC_XOPEN_REALTIME_THREADS
_XOPEN_SHM	whether the implementation supports the XSI shared memory option group	_SC_XOPEN_SHM
_XOPEN_VERSION	the XSI version	_SC_XOPEN_VERSION

Figure 2.19 Options and name arguments to sysconf

    As with the system limits, there are several points to noteregarding how options are treated by sysconf, pathconf, and fpathconf.

    1) The value returned for _SC_VERSION indicatesthe four-digit year and two-digit month of the standard. This value can be 198808L, 199009L, 199506L, or some other value for a later version of the standard. The value associated with Version 3 of the Single UNIX Specification is 200112L (the 2001 edition of POSIX.1). The value associated with Version 4 of the Single UNIX Specification (the 2008 edition of POSIX.1) is 200809L. 

    2) The value returned for _SC_XOPEN_VERSION indicatesthe version of the XSI that the system supports. The value associated with Version 3 of the Single UNIX Specification is 600. The value associated with Version 4 of the Single UNIX Specification (the 2008 edition of POSIX.1) is 700. 

    3) The values _SC_JOB_CONTROL, _SC_SAVED_IDS, and _PC_VDISABLE no longer representoptional features. Although XPG4 and priorversions of the Single UNIX Specification required that these features be supported, Version 3 of the Single UNIX Specification is the earliest version where these features are no longer optional in POSIX.1. These symbols are retained for backward compatibility. 

    4) Platforms conformingto POSIX.1-2008 are also required to support the following options:

• _POSIX_ASYNCHRONOUS_IO
• _POSIX_BARRIERS
• _POSIX_CLOCK_SELECTION
• _POSIX_MAPPED_FILES
• _POSIX_MEMORY_PROTECTION
• _POSIX_READER_WRITER_LOCKS
• _POSIX_REALTIME_SIGNALS
• _POSIX_SEMAPHORES
• _POSIX_SPIN_LOCKS
• _POSIX_THREAD_SAFE_FUNCTIONS
• _POSIX_THREADS
• _POSIX_TIMEOUTS
• _POSIX_TIMERS

    These constants are defined to have the value 200809L. Their corresponding _SC symbols are also retained for backward compatibility.

    5) _PC_CHOWN_RESTRICTED and _PC_NO_TRUNC return −1 without changing errno if the feature is not supported for the specified pathname or fd. On all POSIX-conforming systems, the return value will be greater than zero (indicating that the feature is supported). 

    6) The referencedfile for _PC_CHOWN_RESTRICTED must be either a file or a directory. If it is a directory, the return value indicateswhether this option applies to files within that directory. 

    7) The referencedfile for _PC_NO_TRUNC and _PC_2_SYMLINKS must be a directory. 

    8) For _PC_NO_TRUNC, the return value applies to filenames within the directory. 

    9) The referenced file for _PC_VDISABLE must be a terminalfile. 

    10) For _PC_ASYNC_IO, _PC_PRIO_IO, and _PC_SYNC_IO, the referenced file must not be a directory.

    Figure 2.20 shows several configuration options and their corresponding values on the four sample systems we discuss in this text. An entry is ‘‘unsupported’’ if the system defines the symbolic constant but it has a value of −1, or if it has a value of 0 but the corresponding sysconf or pathconf call returned −1. It is interesting to see that some system implementations haven’t yet caught up to the latest version of the Single UNIX Specification.

				Solaris 10
Limit	FreeBSD  8.0	Linux  3.2.0	Mac OS X  10.6.8	UFS file system	PCFS file system
_POSIX_CHOWN_RESTRICTED	1	1	200112	1	1
_POSIX_JOB_CONTROL	1	1	200112	1	1
_POSIX_NO_TRUNC	1	1	200112	1	unsupported
_POSIX_SAVED_IDS	unsupported	1	200112	1	1
_POSIX_THREADS	200112	200809	200112	200112	200112
_POSIX_VDISABLE	255	0	255	0	0
_POSIX_VERSION	200112	200809	200112	200112	200112
_XOPEN_UNIX	unsupported	1	1	1	1
_XOPEN_VERSION	unsupported	700	600	600	600

Figure 2.20 Examples of configuration options

功能测试宏

    The headers define numerousPOSIX.1 and XSI symbols, as we’ve described. Even so, most implementations can add their own definitions to these headers, in addition tothe POSIX.1 and XSI definitions. If we want to compile a program so that it depends only on the POSIX definitions and doesn’t conflict with any implementation-defined constants, we need to define the constant _POSIX_C_SOURCE. All the POSIX.1 headers use this constant to exclude any implementation-defined definitions when _POSIX_C_SOURCE is defined.

Older versions of the POSIX.1 standard defined the _POSIX_SOURCE constant. This was supersededby the _POSIX_C_SOURCE constant in the 2001 version of POSIX.1.

    The constants _POSIX_C_SOURCE and _XOPEN_SOURCE are called feature test macros. All feature test macros begin with an underscore. When used, they are typically defined in the cc command, as in

cc -D_POSIX_C_SOURCE=200809L  file.c

    This causes the feature test macro to be defined before any header files are included by the C program. If we want to use only the POSIX.1 definitions, we can also set the first line of a source file to

1. #define _POSIX_C_SOURCE 200809L

    To enable the XSI option of Version 4 of the Single UNIX Specification, we need to define the constant _XOPEN_SOURCE to be 700. Besidesenabling the XSI option, this has the same effect as defining _POSIX_C_SOURCE to be 200809L as far as POSIX.1 functionality is concerned. 

    The Single UNIX Specification defines the c99 utility as the interface to the C compilationenvironment. With it we can compile a file as follows:

c99 -D_XOPEN_SOURCE=700 file.c -o file

To enable the 1999 ISO C extensions in the gcc C compiler, we use the -std=c99 option, as in

gcc -D_XOPEN_SOURCE=700 -std=c99 file.c -o file

基本系统数据类型

    Historically, certainC data types have been associated with certain UNIX system variables. For example, major and minor device numbers have historically been stored in a 16-bit short integer, with 8 bits for the major device number and 8 bits for the minor device number. But many larger systems need more than 256 values for these device numbers, so a different technique is needed. (Indeed, the 32-bit version of Solaris uses 32 bits for the device number: 14 bits for the major and 18 bits for the minor.) 

    The header <sys/types.h> defines some implementation-dependent data types, called the primitive system data types. More of these data types are defined in other headers as well. These data types are defined in the headers with the C typedef facility. Most end in _t. Figure 2.21 lists many of the primitive system data types that we’ll encounter in this text. 

Type	Description
clock_t	counter of clock ticks(process time)
comp_t	compressed clock ticks(not defined by POSIX.1)
dev_t	device numbers (major and minor)
fd_set	file descriptor sets
fpos_t	file position
gid_t	numeric group IDs
ino_t	i-node numbers
mode_t	file type, file creation mode
nlink_t	link counts for directory entries
off_t	file sizes and offsets (signed) (lseek)
pid_t	process IDs and process group IDs (signed)
pthread_t	thread IDs
ptrdiff_t	result of subtractingtwo pointers (signed)
rlim_t	resource limits
sig_atomic_t	data type that can be accessed atomically
sigset_t	signal set
size_t	sizes of objects (such as strings) (unsigned)
ssize_t	functions that return a count of bytes (signed) (read, write)
time_t	counter of seconds of calendar time
uid_t	numericuser IDs
wchar_t	can represent all distinctcharacter codes

Figure 2.21 Some common primitive system data types

    By defining these data types this way, we do not build into our programs implementation details that can change from one system to another. We describe what each of these data types is used for when we encounter them later in the text.

标准之间的冲突

    All in all, these various standards fit together nicely. Our main concern is any differences between the ISO C standard and POSIX.1, since the Base Specifications of the Single UNIX Specification and POSIX.1 are one and the same. Conflicts are unintended, but if they should arise, POSIX.1 defers to the ISO C standard. However, there are some differences.

    ISO C defines the function clock to return the amount of CPU time used by a process. The value returned is a clock_t value, but ISO C doesn’t specify its units. To convert this value to seconds, we divideit by CLOCKS_PER_SEC, which is defined in the <time.h> header. POSIX.1 defines the function times that returns both the CPU time (for the caller and all its terminated children) and the clock time. All these time values are clock_t values. The sysconf function is used to obtain the number of clock ticks per second for use with the return values from the times function. What we have is the same data type (clock_t) used to hold measurements of time defined with different units by ISO C and POSIX.1. The difference can be seen in Solaris, where clock returns microseconds (hence CLOCKS_PER_SEC is 1 million), whereassysconf returns the value 100 for clock ticks per second. Thus we must take care when using variables of type clock_t so that we don’t mix variables with different units. 

    Another area of potential conflict is when the ISO C standard specifies a function, but doesn’t specify it as strongly as POSIX.1 does. This is the case for functions that require a different implementation in a POSIX environment (with multiple processes) than in an ISO C environment (where very little can be assumed about the host operating system). Nevertheless, POSIX-compliant systems implement the ISO C function for compatibility. The signal function is an example. If we unknowingly use the signal function provided by Solaris (hoping to write portable code that can be run in ISO C environments and under older UNIX systems), it will provide semantics different from the POSIX.1 sigaction function. We’ll have more to say about the signal function in Chapter 10.

小结

     Much has happened with the standardization of the UNIX programming environment over the past two and a half decades. We’ve described the dominantstandards — ISO C, POSIX, and the Single UNIX Specification—and their effect on the four platforms that we’ll examine in this text—FreeBSD, Linux, Mac OS X, and Solaris. These standards try to define certainparameters that can change with each implementation, but we’ve seen that these limits are imperfect. We’ll encounter many of these limits and magic constants as we proceed through the text. 

    The bibliographyspecifies how to obtain copies of the standards discussed in this chapter.

来自为知笔记(Wiz)