]>
Commit | Line | Data |
---|---|---|
b60a8f50 | 1 | ==================== Changes in man-pages-3.72 ==================== |
5346e714 | 2 | |
257fc8d2 | 3 | Released: 2014-09-07, Mountain View |
5346e714 MK |
4 | |
5 | ||
125baa35 MK |
6 | Contributors |
7 | ------------ | |
26c33493 | 8 | |
5fb27aeb | 9 | The following people contributed patches/fixes or (noted in brackets |
808daa68 | 10 | in the changelog below) reports, notes, and ideas that have been |
5fb27aeb | 11 | incorporated in changes in this release: |
b0aca9e9 | 12 | |
1b679f0a MK |
13 | Christian von Roques <roques@mti.ag> |
14 | Holger Hans Peter Freyther <holger@moiji-mobile.com> | |
15 | Michael Haardt <michael@moria.de> | |
16 | Michael Kerrisk <mtk.manpages@gmail.com> | |
17 | Mike Frysinger <vapier@gentoo.org> | |
18 | Peter Schiffer <pschiffe@redhat.com> | |
19 | Rusty Russell <rusty@rustcorp.com.au> | |
20 | Sorin Dumitru <sdumitru@ixiacom.com> | |
b3206068 MK |
21 | |
22 | Apologies if I missed anyone! | |
23 | ||
24 | ||
324197b5 MK |
25 | New and rewritten pages |
26 | ----------------------- | |
27 | ||
1b679f0a MK |
28 | memusage.1 |
29 | Peter Schiffer, Michael Kerrisk [Jan Chaloupka] | |
30 | New page for glibc memusage(1) command | |
12e73609 | 31 | |
1b679f0a MK |
32 | memusagestat.1 |
33 | Peter Schiffer [Jan Chaloupka, Michael Kerrisk] | |
34 | New page for glibc memusagestat(1) command | |
324197b5 | 35 | |
1b679f0a MK |
36 | mtrace.1 |
37 | Peter Schiffer [Jan Chaloupka] | |
38 | New page describing the glibc mtrace(1) command | |
65266f17 | 39 | |
65266f17 | 40 | |
19b7ee90 MK |
41 | Changes to individual pages |
42 | --------------------------- | |
65266f17 | 43 | |
1b679f0a MK |
44 | connect.2 |
45 | Michael Haardt | |
46 | Note that a new socket should be used if connect() fails | |
47 | ||
48 | fcntl.2 | |
49 | Michael Kerrisk | |
50 | One must define _GNU_SOURCE to get the F_OFD_* definitions | |
51 | ||
52 | poll.2, select.2 | |
53 | Rusty Russell | |
54 | Fix erroneous description of "available for write". | |
55 | POSIX says: "POLLOUT Normal data may be written without | |
4b2dbe83 | 56 | blocking.". This "may" is misleading, see the POSIX |
1b679f0a MK |
57 | write page: |
58 | ||
59 | Write requests to a pipe or FIFO shall be handled in the | |
60 | same way as a regular file with the following exceptions: | |
61 | ... | |
62 | If the O_NONBLOCK flag is clear, a write request may cause | |
63 | the thread to block, but on normal completion it shall | |
64 | return nbyte. | |
65 | ... | |
66 | When attempting to write to a file descriptor (other than a | |
67 | pipe or FIFO) that supports non-blocking writes and cannot | |
68 | accept the data immediately: | |
69 | ||
70 | If the O_NONBLOCK flag is clear, write() shall block the | |
71 | calling thread until the data can be accepted. | |
72 | ||
73 | If the O_NONBLOCK flag is set, write() shall not block the | |
74 | thread. If some data can be written without blocking the | |
75 | thread, write() shall write what it can and return the | |
76 | number of bytes written. Otherwise, it shall return -1 and | |
77 | set errno to [EAGAIN]. | |
78 | ||
79 | The net result is that write() of more than 1 byte on a | |
80 | socket, pipe or FIFO which is "ready" may block: write() | |
81 | (unlike read!) will attempt to write the entire buffer and | |
82 | only return a short write under exceptional circumstances. | |
83 | ||
84 | Indeed, this is the behaviour we see in Linux: | |
85 | ||
86 | https://github.com/rustyrussell/ccan/commit/897626152d12d7fd13a8feb36989eb5c8c1f3485 | |
87 | https://plus.google.com/103188246877163594460/posts/BkTGTMHDFgZ | |
88 | ||
89 | errno.3 | |
90 | Michael Kerrisk | |
91 | SEE ALSO: add errno(1) | |
92 | ||
93 | rtnetlink.3 | |
94 | Holger Hans Peter Freyther | |
95 | Fix parameters for the send() call in the example | |
96 | ||
97 | inotify.7 | |
98 | Michael Kerrisk | |
99 | IN_OPEN and IN_CLOSE_NOWRITE can also occur for directories | |
100 | Michael Kerrisk | |
101 | IN_CLOSE_WRITE occurs only for files (not monitored directory) | |
102 | Michael Kerrisk | |
103 | IN_MODIFY is generated for files only (not monitored directories) | |
104 | Michael Kerrisk | |
105 | IN_ACCESS occurs only for files inside directories | |
106 | IN_ACCESS does not occur for monitored directory. | |
107 | ||
108 | packet.7 | |
109 | Sorin Dumitru | |
110 | Fix include file | |
111 | It looks like most of the socket options from this man pages | |
112 | are not defined in <netpacket/packet.h>. They are defined in | |
113 | <linux/if_packet.h> so we should include that one. |