From 98ef80442c470a0ebf75afff64d2d78d01ae60cf Mon Sep 17 00:00:00 2001 From: Michael Kerrisk Date: Sun, 19 Apr 2015 11:27:01 +0200 Subject: [PATCH] Start of man-pages-4.00: updating Changes and Changes.old --- Changes | 450 +----------------------------------------------- Changes.old | 483 ++++++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 485 insertions(+), 448 deletions(-) diff --git a/Changes b/Changes index 987143eb4..b235eec99 100644 --- a/Changes +++ b/Changes @@ -1,6 +1,6 @@ -==================== Changes in man-pages-3.83 ==================== +==================== Changes in man-pages-4.00 ==================== -Released: 2015-04-19, Munich +Released: ????-??-??, Munich Contributors @@ -10,36 +10,6 @@ The following people contributed patches/fixes or (noted in brackets in the changelog below) reports, notes, and ideas that have been incorporated in changes in this release: -Andreas Baak -Andreas Dilger -cdlscpmv -Cyrill Gorcunov -Darrick J. Wong -David Rientjes -Dima Tisnek -Eric Sandeen -Fabien Pichot -Felix Sedlmeier -Gleb Fotengauer-Malinovskiy -Heinrich Schuchardt -Jann Horn -Jon Grant -Jonny Grant -Kees Cook -Masanari Iida -Ma Shimiao -Michael Kerrisk -Nikos Mavrogiannopoulos -Omar Sandoval -Pierre Chifflier -Robin H. Johnson -Rob Landley -Theodore Ts'o -Vlastimil Babka -Walter Harms -William Woodruff -YOSHIFUJI Hideaki -Zeng Linggang Apologies if I missed anyone! @@ -47,435 +17,19 @@ Apologies if I missed anyone! New and rewritten pages ----------------------- -posix_madvise.3 - Michael Kerrisk - New page documenting posix_madvise(3) - -ftw.3 - Michael Kerrisk - Reorganize the page to give primacy to nftw() - nftw() is the better API, and POSIX.1-2008 marks ftw() obsolete. Newly documented interfaces in existing pages --------------------------------------------- -getdents.2 - Michael Kerrisk - Document getdents64() - -mount.2 - Michael Kerrisk, Theodore Ts'o [Eric Sandeen, Andreas Dilger, - Omar Sandoval, Darrick J. Wong] - Document MS_LAZYTIME - Based on Ted T'so's commit message 0ae45f63d4e - -proc.5 - Michael Kerrisk - Document /proc/sys/kernel/randomize_va_space - Michael Kerrisk - Document /proc/PID/fdinfo epoll format - Michael Kerrisk - Describe /proc/PID/fdinfo eventfd format - Michael Kerrisk - Document /proc/PID/fdinfo signalfd format - New and changed links --------------------- -newfstatat.2 - Michael Kerrisk - New link to fstatat64.2 - -prlimit64.2 - Michael Kerrisk - New link to getrlimit.2 - Global changes -------------- -Various secion 3 math pages - Michael Kerrisk - Note that these functions are also in POSIX.1-2001 and POSIX.1-2008 - Changes to individual pages --------------------------- -getent.1 - Robin H. Johnson - Document options - The options to getent are now documented, after being around for - nearly a decade without changes. - Michael Kerrisk - Document help and version options - -fallocate.2 - Michael Kerrisk - Fix kernel version number for FALLOC_FL_ZERO_RANGE - FALLOC_FL_ZERO_RANGE as added in 3.15, not 3.14. - Michael Kerrisk - Note that SMB3 added FALLOC_FL_ZERO_RANGE support in Linux 3.17 - -getdents.2 - Michael Kerrisk [Dima Tisnek] - Update timestamp - Really just a marker to record the reporters of bugs - that stemmed from the fact that the page did not - document getdents64(). I'll fix things up in the changelog. - - See https://bugzilla.kernel.org/show_bug.cgi?id=14795 - -getrlimit.2 - Michael Kerrisk - Note that the underlying system call for prlimit() is prlimit64() - Michael Kerrisk - Remove "_FILE_OFFSET_BITS == 64" from prlimit() FTM requirements - "_FILE_OFFSET_BITS == 64" is not needed to get the prlimit() - declaration. - -ioctl_list.2 - Nikos Mavrogiannopoulos - SIOCADDRT accepts in6_rtmsg in INET6 protocol - Heinrich Schuchardt - TFD_IOC_SET_TICKS - timerfd_create.2 mentions TFD_IOC_SET_TICKS. We should add it to - ioctl_list.2, too. - -llseek.2 - Michael Kerrisk - Advise reader to use lseek(2) instead - Michael Kerrisk - llseek() exists on 32-bit platforms to support seeking to large offsets - -madvise.2 - David Rientjes - Specify MADV_REMOVE returns EINVAL for hugetlbfs - madvise(2) actually returns with error EINVAL for MADV_REMOVE - when used for hugetlb VMAs, not EOPNOTSUPP, and this has been - the case since MADV_REMOVE was introduced in commit f6b3ec238d12 - ("madvise(MADV_REMOVE): remove pages from tmpfs shm backing - store"). Specify the exact behavior. - Michael Kerrisk - SEE ALSO: add posix_madvise(2) - -poll.2 - Michael Kerrisk [Andreas Baak] - SEE ALSO: add epoll(7) - -posix_fadvise.2 - Michael Kerrisk - Add "C library/kernel ABI differences" subsection - -pread.2 - Michael Kerrisk - Add "C library/kernel ABI differences" subsection - -seccomp.2 - Michael Kerrisk [Pierre Chifflier, Kees Cook] - Note that seccomp_data is read-only - -stat.2 - Michael Kerrisk - Add some details on various "stat" versions - Three versions of "stat" appeared on 32-bit systems, - dealing with structures of different (increasing) sizes. - Explain some of the details, and also note that the - situation is simpler on modern 64-bit architectures. - Michael Kerrisk - Add a little more detail on history of three versions of *stat() - Michael Kerrisk - Add mention of newfstatat() - The underlying system call for fstatat() is newfstatat() - on some architectures. - -symlink.2 - Michael Kerrisk [Jonny Grant] - ERRORS: add linkpath=="" case for ENOENT - -syscalls.2 - Michael Kerrisk - Remove prlimit() - There really is only the prlimit64() system call. - See also commit 7b08c45b7df. - Michael Kerrisk - Add some details about the "multiple versions of system calls" - The multiple-system-call-version phenomenon is particular a - feature of older 32-bit platforms. Hint at that fact in the text. - -timerfd_create.2 - Cyrill Gorcunov [Michael Kerrisk] - Document TFD_IOC_SET_TICKS ioctl() operation - Michael Kerrisk - Add some details to C library/kernel ABI differences - -unshare.2 - Michael Kerrisk [Fabien Pichot] - Remove mention of "System V" from discussion of CLONE_NEWIPC - These days, CLONE_NEWIPC also affects POSIX message queues. - -asprintf.3 - Zeng Linggang - ATTRIBUTES: Note functions that are thread-safe - -carg.3 - Michael Kerrisk - Add introductory sentence explaining what these functions calculate - -ccos.3 - Ma Shimiao - ATTRIBUTES: Note functions that are thread-safe - Michael Kerrisk - Add introductory sentence explaining what these functions calculate - -ccosh.3 - Michael Kerrisk - Add introductory sentence explaining what these functions calculate - -cexp.3 - Ma Shimiao - ATTRIBUTES: Note functions that are thread-safe - -clock.3 - Ma Shimiao - ATTRIBUTES: Note functions that is thread-safe - -clog.3 - Michael Kerrisk - Add introductory sentence explaining what these functions calculate - -csin.3 - Ma Shimiao - ATTRIBUTES: Note functions that are thread-safe - Michael Kerrisk - Add introductory sentence explaining what these functions calculate - -csinh.3 - Ma Shimiao - ATTRIBUTES: Note functions that are thread-safe - Michael Kerrisk - Add introductory sentence explaining what these functions calculate - -csqrt.3 - Ma Shimiao - ATTRIBUTES: Note functions that are thread-safe - Michael Kerrisk - Simplify description of what these functions calculate - -ctan.3 - Ma Shimiao - ATTRIBUTES: Note functions that are thread-safe - Michael Kerrisk - Add introductory sentence explaining what these functions calculate - -ctanh.3 - Ma Shimiao - ATTRIBUTES: Note functions that are thread-safe - Michael Kerrisk - Add introductory sentence explaining what these functions calculate - -ctime.3 - Zeng Linggang - ATTRIBUTES: Note functions that aren't thread-safe - -exec.3 - Michael Kerrisk - SYNOPSIS: Clarify calling signature for execl() and execlp() - Michael Kerrisk [Andreas Baak] - Correct prototype for execle() - Make the prototype shown into correct C. - -ftw.3 - Michael Kerrisk [Felix Sedlmeier] - ftw() and nftw() differ for the non-stat-able symlink case - The POSIX specification of ftw() says that an un-stat-able - symlink may yield either FTW_NS or FTW_SL. The specification - of nftw() does not carry this statement. - Michael Kerrisk - CONFORMING TO: add POSIX.1-2008 - Michael Kerrisk - Update POSIX version references in NOTES - -getcwd.3 - Jann Horn [Michael Kerrisk] - Note behavior for unreachable current working directory - Michael Kerrisk - Add ENOMEM error - -gethostbyname.3 - Michael Kerrisk [Jonny Grant] - Clarify that NO_ADDRESS and NO_DATA are synonyms - Michael Kerrisk - Add some detail for NO_DATA - Text consistent with POSIX and FreeBSD's gethostbyname() man page. - Zeng Linggang - ATTRIBUTES: Note functions that aren't thread-safe - -getnetent.3 - Zeng Linggang - ATTRIBUTES: Note functions that aren't thread-safe - -get_nprocs_conf.3 - Zeng Linggang - ATTRIBUTES: Note functions that are thread-safe - -getutent.3 - Zeng Linggang - ATTRIBUTES: Note functions that aren't thread-safe - -glob.3 - Zeng Linggang - ATTRIBUTES: Note functions that aren't thread-safe - -insque.3 - Ma Shimiao - ATTRIBUTES: Note functions that are thread-safe - -login.3 - Zeng Linggang - ATTRIBUTES: Note functions that aren't thread-safe - -lseek64.3 - Michael Kerrisk - Clarify details with respect to 32-bit and 64-bit systems - -malloc.3 - Michael Kerrisk - Add ENOMEM error - -mbsnrtowcs.3 - Zeng Linggang - ATTRIBUTES: Note function that isn't thread-safe - -mbsrtowcs.3 - Zeng Linggang - ATTRIBUTES: Note function that isn't thread-safe - -mq_notify.3 - Michael Kerrisk - Add "C library/kernel ABI differences" subsection - -mq_open.3 - Michael Kerrisk [Fabien Pichot] - NOTES: explain differences from the underlying system call - The check for the slash at the start of a pathname is done in glibc - -openpty.3 - Zeng Linggang - ATTRIBUTES: Note functions that aren't thread-safe - -perror.3 - Zeng Linggang - ATTRIBUTES: Note function that is thread-safe - -posix_memalign.3 - Zeng Linggang - ATTRIBUTES: Note functions that aren't thread-safe - -printf.3 - Zeng Linggang - ATTRIBUTES: Note functions that are thread-safe - Walter Harms [Michael Kerrisk] - Simplify the example code - -qsort.3 - Michael Kerrisk [Rob Landley] - alphasort() and versionsort() are not suitable for 'compar' - In glibc 2.10, the prototypes of alphasort() and versionsort() - were changed so that the arguments switched from 'const void *' to - 'const struct dirent **', to match the POSIX.1-2008 specification - of alphasort(). As such, compiler warnings will result if - these functions are used as the arguments of qsort(). - - warning: passing argument 4 of 'qsort' from incompatible - pointer type - expected '__compar_fn_t' but argument is of type - 'int (*)(const struct dirent **, const struct dirent **)' - - Therefore, remove the ancient NOTES text suggesting that - alphasort() and versionsort() can be used as suitable - 'compar' arguments for qsort(). - -realpath.3 - Michael Kerrisk [Jon Grant] - Add ENOMEM error - -scandir.3 - Michael Kerrisk - glibc 2.10 changed the argument types for alphasort() and versionsort() - Zeng Linggang - ATTRIBUTES: Note functions that are thread-safe - -scanf.3 - Zeng Linggang - ATTRIBUTES: Note functions that are thread-safe - -setnetgrent.3 - Zeng Linggang - ATTRIBUTES: Note functions that aren't thread-safe - -significand.3 - Ma Shimiao - ATTRIBUTES: Note functions that are thread-safe - -strcasecmp.3 - Michael Kerrisk [Jonny Grant] - Clarify that strcasecmp() does a byte-wise comparison - Michael Kerrisk - CONFORMING TO: add POSIX.1-2008 - -unlocked_stdio.3 - Zeng Linggang - ATTRIBUTES: Note functions that aren't thread-safe - -updwtmp.3 - Zeng Linggang - ATTRIBUTES: Note functions that aren't thread-safe - -wcrtomb.3 - Zeng Linggang - ATTRIBUTES: Note function that isn't thread-safe - -wcsnrtombs.3 - Zeng Linggang - ATTRIBUTES: Note function that isn't thread-safe - -wcsrtombs.3 - Zeng Linggang - ATTRIBUTES: Note function that isn't thread-safe - -wordexp.3 - Zeng Linggang - ATTRIBUTES: Note functions that aren't thread-safe - -wprintf.3 - Zeng Linggang - ATTRIBUTES: Note functions that are thread-safe - -proc.5 - Michael Kerrisk - Describe "mnt_id" field of /proc/PID/fdinfo - Michael Kerrisk - Note that abstract sockets are included in /proc/net/unix - Michael Kerrisk - Update description /proc/sys/unix 'Type' field - The existing text was very crufty. UNIX domain sockets - support more than SOCK_STREAM for a _very_ long time now. - Michael Kerrisk - Add some detail to /proc/PID/timers - Michael Kerrisk [Vlastimil Babka] - Enhance discussion of /proc/PID/status 'VmSwap' field - Based on https://lwn.net/Articles/634978/ - -epoll.7 - Michael Kerrisk - SEE ALSO: add poll(2) and select(2) - -icmp.7 - YOSHIFUJI Hideaki/吉藤英明 - Document net.ipv4.ping_group_range knob - -nptl.7 - Michael Kerrisk - Add reference to timer_create(2) diff --git a/Changes.old b/Changes.old index b81deddfa..1d31e19a0 100644 --- a/Changes.old +++ b/Changes.old @@ -37383,3 +37383,486 @@ ld.so.8 Michael Kerrisk Note the use of /lib64 and /usr/lib64 on some 64-bit architectures + + +==================== Changes in man-pages-3.83 ==================== + +Released: 2015-04-19, Munich + + +Contributors +------------ + +The following people contributed patches/fixes or (noted in brackets +in the changelog below) reports, notes, and ideas that have been +incorporated in changes in this release: + +Andreas Baak +Andreas Dilger +cdlscpmv +Cyrill Gorcunov +Darrick J. Wong +David Rientjes +Dima Tisnek +Eric Sandeen +Fabien Pichot +Felix Sedlmeier +Gleb Fotengauer-Malinovskiy +Heinrich Schuchardt +Jann Horn +Jon Grant +Jonny Grant +Kees Cook +Masanari Iida +Ma Shimiao +Michael Kerrisk +Nikos Mavrogiannopoulos +Omar Sandoval +Pierre Chifflier +Robin H. Johnson +Rob Landley +Theodore Ts'o +Vlastimil Babka +Walter Harms +William Woodruff +YOSHIFUJI Hideaki +Zeng Linggang + +Apologies if I missed anyone! + + +New and rewritten pages +----------------------- + +posix_madvise.3 + Michael Kerrisk + New page documenting posix_madvise(3) + +ftw.3 + Michael Kerrisk + Reorganize the page to give primacy to nftw() + nftw() is the better API, and POSIX.1-2008 marks ftw() obsolete. + +Newly documented interfaces in existing pages +--------------------------------------------- + +getdents.2 + Michael Kerrisk + Document getdents64() + +mount.2 + Michael Kerrisk, Theodore Ts'o [Eric Sandeen, Andreas Dilger, + Omar Sandoval, Darrick J. Wong] + Document MS_LAZYTIME + Based on Ted T'so's commit message 0ae45f63d4e + +proc.5 + Michael Kerrisk + Document /proc/sys/kernel/randomize_va_space + Michael Kerrisk + Document /proc/PID/fdinfo epoll format + Michael Kerrisk + Describe /proc/PID/fdinfo eventfd format + Michael Kerrisk + Document /proc/PID/fdinfo signalfd format + + +New and changed links +--------------------- + +newfstatat.2 + Michael Kerrisk + New link to fstatat64.2 + +prlimit64.2 + Michael Kerrisk + New link to getrlimit.2 + + +Global changes +-------------- + +Various secion 3 math pages + Michael Kerrisk + Note that these functions are also in POSIX.1-2001 and POSIX.1-2008 + + +Changes to individual pages +--------------------------- + +getent.1 + Robin H. Johnson + Document options + The options to getent are now documented, after being around for + nearly a decade without changes. + Michael Kerrisk + Document help and version options + +fallocate.2 + Michael Kerrisk + Fix kernel version number for FALLOC_FL_ZERO_RANGE + FALLOC_FL_ZERO_RANGE as added in 3.15, not 3.14. + Michael Kerrisk + Note that SMB3 added FALLOC_FL_ZERO_RANGE support in Linux 3.17 + +getdents.2 + Michael Kerrisk [Dima Tisnek] + Update timestamp + Really just a marker to record the reporters of bugs + that stemmed from the fact that the page did not + document getdents64(). I'll fix things up in the changelog. + + See https://bugzilla.kernel.org/show_bug.cgi?id=14795 + +getrlimit.2 + Michael Kerrisk + Note that the underlying system call for prlimit() is prlimit64() + Michael Kerrisk + Remove "_FILE_OFFSET_BITS == 64" from prlimit() FTM requirements + "_FILE_OFFSET_BITS == 64" is not needed to get the prlimit() + declaration. + +ioctl_list.2 + Nikos Mavrogiannopoulos + SIOCADDRT accepts in6_rtmsg in INET6 protocol + Heinrich Schuchardt + TFD_IOC_SET_TICKS + timerfd_create.2 mentions TFD_IOC_SET_TICKS. We should add it to + ioctl_list.2, too. + +llseek.2 + Michael Kerrisk + Advise reader to use lseek(2) instead + Michael Kerrisk + llseek() exists on 32-bit platforms to support seeking to large offsets + +madvise.2 + David Rientjes + Specify MADV_REMOVE returns EINVAL for hugetlbfs + madvise(2) actually returns with error EINVAL for MADV_REMOVE + when used for hugetlb VMAs, not EOPNOTSUPP, and this has been + the case since MADV_REMOVE was introduced in commit f6b3ec238d12 + ("madvise(MADV_REMOVE): remove pages from tmpfs shm backing + store"). Specify the exact behavior. + Michael Kerrisk + SEE ALSO: add posix_madvise(2) + +poll.2 + Michael Kerrisk [Andreas Baak] + SEE ALSO: add epoll(7) + +posix_fadvise.2 + Michael Kerrisk + Add "C library/kernel ABI differences" subsection + +pread.2 + Michael Kerrisk + Add "C library/kernel ABI differences" subsection + +seccomp.2 + Michael Kerrisk [Pierre Chifflier, Kees Cook] + Note that seccomp_data is read-only + +stat.2 + Michael Kerrisk + Add some details on various "stat" versions + Three versions of "stat" appeared on 32-bit systems, + dealing with structures of different (increasing) sizes. + Explain some of the details, and also note that the + situation is simpler on modern 64-bit architectures. + Michael Kerrisk + Add a little more detail on history of three versions of *stat() + Michael Kerrisk + Add mention of newfstatat() + The underlying system call for fstatat() is newfstatat() + on some architectures. + +symlink.2 + Michael Kerrisk [Jonny Grant] + ERRORS: add linkpath=="" case for ENOENT + +syscalls.2 + Michael Kerrisk + Remove prlimit() + There really is only the prlimit64() system call. + See also commit 7b08c45b7df. + Michael Kerrisk + Add some details about the "multiple versions of system calls" + The multiple-system-call-version phenomenon is particular a + feature of older 32-bit platforms. Hint at that fact in the text. + +timerfd_create.2 + Cyrill Gorcunov [Michael Kerrisk] + Document TFD_IOC_SET_TICKS ioctl() operation + Michael Kerrisk + Add some details to C library/kernel ABI differences + +unshare.2 + Michael Kerrisk [Fabien Pichot] + Remove mention of "System V" from discussion of CLONE_NEWIPC + These days, CLONE_NEWIPC also affects POSIX message queues. + +asprintf.3 + Zeng Linggang + ATTRIBUTES: Note functions that are thread-safe + +carg.3 + Michael Kerrisk + Add introductory sentence explaining what these functions calculate + +ccos.3 + Ma Shimiao + ATTRIBUTES: Note functions that are thread-safe + Michael Kerrisk + Add introductory sentence explaining what these functions calculate + +ccosh.3 + Michael Kerrisk + Add introductory sentence explaining what these functions calculate + +cexp.3 + Ma Shimiao + ATTRIBUTES: Note functions that are thread-safe + +clock.3 + Ma Shimiao + ATTRIBUTES: Note functions that is thread-safe + +clog.3 + Michael Kerrisk + Add introductory sentence explaining what these functions calculate + +csin.3 + Ma Shimiao + ATTRIBUTES: Note functions that are thread-safe + Michael Kerrisk + Add introductory sentence explaining what these functions calculate + +csinh.3 + Ma Shimiao + ATTRIBUTES: Note functions that are thread-safe + Michael Kerrisk + Add introductory sentence explaining what these functions calculate + +csqrt.3 + Ma Shimiao + ATTRIBUTES: Note functions that are thread-safe + Michael Kerrisk + Simplify description of what these functions calculate + +ctan.3 + Ma Shimiao + ATTRIBUTES: Note functions that are thread-safe + Michael Kerrisk + Add introductory sentence explaining what these functions calculate + +ctanh.3 + Ma Shimiao + ATTRIBUTES: Note functions that are thread-safe + Michael Kerrisk + Add introductory sentence explaining what these functions calculate + +ctime.3 + Zeng Linggang + ATTRIBUTES: Note functions that aren't thread-safe + +exec.3 + Michael Kerrisk + SYNOPSIS: Clarify calling signature for execl() and execlp() + Michael Kerrisk [Andreas Baak] + Correct prototype for execle() + Make the prototype shown into correct C. + +ftw.3 + Michael Kerrisk [Felix Sedlmeier] + ftw() and nftw() differ for the non-stat-able symlink case + The POSIX specification of ftw() says that an un-stat-able + symlink may yield either FTW_NS or FTW_SL. The specification + of nftw() does not carry this statement. + Michael Kerrisk + CONFORMING TO: add POSIX.1-2008 + Michael Kerrisk + Update POSIX version references in NOTES + +getcwd.3 + Jann Horn [Michael Kerrisk] + Note behavior for unreachable current working directory + Michael Kerrisk + Add ENOMEM error + +gethostbyname.3 + Michael Kerrisk [Jonny Grant] + Clarify that NO_ADDRESS and NO_DATA are synonyms + Michael Kerrisk + Add some detail for NO_DATA + Text consistent with POSIX and FreeBSD's gethostbyname() man page. + Zeng Linggang + ATTRIBUTES: Note functions that aren't thread-safe + +getnetent.3 + Zeng Linggang + ATTRIBUTES: Note functions that aren't thread-safe + +get_nprocs_conf.3 + Zeng Linggang + ATTRIBUTES: Note functions that are thread-safe + +getutent.3 + Zeng Linggang + ATTRIBUTES: Note functions that aren't thread-safe + +glob.3 + Zeng Linggang + ATTRIBUTES: Note functions that aren't thread-safe + +insque.3 + Ma Shimiao + ATTRIBUTES: Note functions that are thread-safe + +login.3 + Zeng Linggang + ATTRIBUTES: Note functions that aren't thread-safe + +lseek64.3 + Michael Kerrisk + Clarify details with respect to 32-bit and 64-bit systems + +malloc.3 + Michael Kerrisk + Add ENOMEM error + +mbsnrtowcs.3 + Zeng Linggang + ATTRIBUTES: Note function that isn't thread-safe + +mbsrtowcs.3 + Zeng Linggang + ATTRIBUTES: Note function that isn't thread-safe + +mq_notify.3 + Michael Kerrisk + Add "C library/kernel ABI differences" subsection + +mq_open.3 + Michael Kerrisk [Fabien Pichot] + NOTES: explain differences from the underlying system call + The check for the slash at the start of a pathname is done in glibc + +openpty.3 + Zeng Linggang + ATTRIBUTES: Note functions that aren't thread-safe + +perror.3 + Zeng Linggang + ATTRIBUTES: Note function that is thread-safe + +posix_memalign.3 + Zeng Linggang + ATTRIBUTES: Note functions that aren't thread-safe + +printf.3 + Zeng Linggang + ATTRIBUTES: Note functions that are thread-safe + Walter Harms [Michael Kerrisk] + Simplify the example code + +qsort.3 + Michael Kerrisk [Rob Landley] + alphasort() and versionsort() are not suitable for 'compar' + In glibc 2.10, the prototypes of alphasort() and versionsort() + were changed so that the arguments switched from 'const void *' to + 'const struct dirent **', to match the POSIX.1-2008 specification + of alphasort(). As such, compiler warnings will result if + these functions are used as the arguments of qsort(). + + warning: passing argument 4 of 'qsort' from incompatible + pointer type + expected '__compar_fn_t' but argument is of type + 'int (*)(const struct dirent **, const struct dirent **)' + + Therefore, remove the ancient NOTES text suggesting that + alphasort() and versionsort() can be used as suitable + 'compar' arguments for qsort(). + +realpath.3 + Michael Kerrisk [Jon Grant] + Add ENOMEM error + +scandir.3 + Michael Kerrisk + glibc 2.10 changed the argument types for alphasort() and versionsort() + Zeng Linggang + ATTRIBUTES: Note functions that are thread-safe + +scanf.3 + Zeng Linggang + ATTRIBUTES: Note functions that are thread-safe + +setnetgrent.3 + Zeng Linggang + ATTRIBUTES: Note functions that aren't thread-safe + +significand.3 + Ma Shimiao + ATTRIBUTES: Note functions that are thread-safe + +strcasecmp.3 + Michael Kerrisk [Jonny Grant] + Clarify that strcasecmp() does a byte-wise comparison + Michael Kerrisk + CONFORMING TO: add POSIX.1-2008 + +unlocked_stdio.3 + Zeng Linggang + ATTRIBUTES: Note functions that aren't thread-safe + +updwtmp.3 + Zeng Linggang + ATTRIBUTES: Note functions that aren't thread-safe + +wcrtomb.3 + Zeng Linggang + ATTRIBUTES: Note function that isn't thread-safe + +wcsnrtombs.3 + Zeng Linggang + ATTRIBUTES: Note function that isn't thread-safe + +wcsrtombs.3 + Zeng Linggang + ATTRIBUTES: Note function that isn't thread-safe + +wordexp.3 + Zeng Linggang + ATTRIBUTES: Note functions that aren't thread-safe + +wprintf.3 + Zeng Linggang + ATTRIBUTES: Note functions that are thread-safe + +proc.5 + Michael Kerrisk + Describe "mnt_id" field of /proc/PID/fdinfo + Michael Kerrisk + Note that abstract sockets are included in /proc/net/unix + Michael Kerrisk + Update description /proc/sys/unix 'Type' field + The existing text was very crufty. UNIX domain sockets + support more than SOCK_STREAM for a _very_ long time now. + Michael Kerrisk + Add some detail to /proc/PID/timers + Michael Kerrisk [Vlastimil Babka] + Enhance discussion of /proc/PID/status 'VmSwap' field + Based on https://lwn.net/Articles/634978/ + +epoll.7 + Michael Kerrisk + SEE ALSO: add poll(2) and select(2) + +icmp.7 + YOSHIFUJI Hideaki/吉藤英明 + Document net.ipv4.ping_group_range knob + +nptl.7 + Michael Kerrisk + Add reference to timer_create(2)