Rewrite, expand, and update to 2.6.22 state of affairs.

This commit is contained in:
Michael Kerrisk 2007-06-28 06:26:51 +00:00
parent 360ed6b354
commit 09de4bf118
1 changed files with 648 additions and 266 deletions

View File

@ -1,3 +1,7 @@
.\" Copyright (C) 2007 Michael Kerrisk <mtk-manpages@gmx.net>
.\" with some input from Stepan Kasal <kasal@ucw.cz>
.\"
.\" Some content retained from an earlier version of this page:
.\" Copyright (C) 1998 Andries Brouwer (aeb@cwi.nl)
.\" Modifications for 2.2 and 2.4 Copyright (C) 2002 Ian Redfern
.\" <redferni@logica.com>
@ -22,235 +26,488 @@
.\" Formatted or processed versions of this manual, if unaccompanied by
.\" the source, must acknowledge the copyright and authors of this work.
.\"
.TH SYSCALLS 2 2002-01-07 "Linux" "Linux Programmer's Manual"
.TH SYSCALLS 2 2007-07-07 "Linux" "Linux Programmer's Manual"
.SH NAME
none \- list of all system calls
syscalls \- Linux system calls
.SH SYNOPSIS
Linux 2.4 system calls.
Linux system calls.
.SH DESCRIPTION
The system call is the fundamental interface between an application
and the Linux kernel.
As of Linux 2.4.17, there are 1100 system calls
listed in
.IR /usr/src/linux/include/asm-*/unistd.h .
This man page lists those that are common to most platforms.
_llseek(2),
_newselect(2),
_sysctl(2),
accept(2),
access(2),
acct(2),
adjtimex(2),
afs_syscall,
alarm(2),
bdflush(2),
bind(2),
break,
brk(2),
cacheflush(2),
capget(2),
capset(2),
chdir(2),
chmod(2),
chown(2), chown32,
chroot(2),
clone(2),
close(2),
connect(2),
creat(2),
create_module(2),
delete_module(2),
dup(2),
dup2(2),
execve(2),
exit(2),
fchdir(2),
fchmod(2),
fchown(2), fchown32,
fcntl(2), fcntl64,
fdatasync(2),
flock(2),
fork(2),
fstat(2), fstat64,
fstatfs(2),
fsync(2),
ftime,
ftruncate(2), ftruncate64,
get\%_kernel\%_syms(2),
get\%cwd(2),
get\%dents(2), get\%dents64,
get\%egid(2), get\%egid32,
get\%euid(2), get\%euid32,
get\%gid(2), get\%gid32,
get\%groups(2), get\%groups32,
get\%itimer(2),
get\%pagesize(2),
get\%peername(2),
get\%pmsg,
get\%pgid(2),
get\%pgrp(2),
get\%pid(2),
get\%ppid(2),
get\%priority(2),
get\%resgid(2), get\%resgid32,
get\%resuid(2), get\%resuid32,
get\%rlimit(2),
get\%rusage(2),
get\%sid(2),
get\%sockname(2),
get\%sockopt(2),
get\%tid,
get\%timeofday(2),
get\%uid(2), get\%uid32,
gtty,
idle,
init_module(2),
ioctl(2),
io\%perm(2),
iopl(2),
ipc(2),
kill(2),
lchown(2), lchown32,
link(2),
listen(2),
lock,
lseek(2),
lstat(2), lstat64,
madvise(2),
mincore(2),
mkdir(2),
mknod(2),
mlock(2),
mlockall(2),
mmap(2),
modify_ldt(2),
mount(2),
mprotect(2),
mpx,
mremap(2),
msync(2),
munlock(2),
munlockall(2),
munmap(2),
nanosleep(2),
nfsservctl(2),
nice(2),
oldfstat, oldlstat, oldolduname, oldstat, oldumount, olduname,
open(2),
pause(2),
personality(2),
phys,
pipe(2),
pivot_root(2),
poll(2),
prctl(2),
pread(2),
prof, profil,
ptrace(2),
putpmsg,
pwrite(2),
query_module(2),
quotactl(2),
read(2),
readahead,
readdir(2),
readlink(2),
readv(2),
reboot(2),
recv(2), recvfrom(2), recvmsg(2),
rename(2),
rmdir(2),
rt_sigaction,
rt_sigpending,
rt_sigprocmask,
rt_sigqueueinfo,
rt_sigreturn,
rt_sigsuspend,
rt_sigtimedwait,
sched_\%get_\%priority_max(2),
sched_\%get_\%priority_min(2),
sched_\%get\%param(2),
sched_\%get\%scheduler(2),
sched_\%rr_\%get_\%interval(2),
sched_\%set\%param(2),
sched_\%set\%scheduler(2),
sched_\%yield(2),
security,
select(2),
sendfile(2),
send(2), sendmsg(2), sendto(2),
set\%domainname(2),
set\%fsgid(2), set\%fsgid32,
set\%fsuid(2), set\%fsuid32,
set\%gid(2), set\%gid32,
set\%groups(2), set\%groups32,
set\%hostname(2),
set\%itimer(2),
set\%pgid(2),
set\%priority(2),
set\%regid(2), set\%regid32,
set\%resgid(2), set\%resgid32,
set\%resuid(2), set\%resuid32,
set\%reuid(2), set\%reuid32,
set\%rlimit(2),
set\%sid(2),
set\%sockopt(2),
set\%timeofday(2),
set\%uid(2), set\%uid32,
setup(2),
sgetmask(2),
shutdown(2),
sigaction(2),
sigaltstack(2),
signal(2),
sigpending(2),
sigprocmask(2),
sigreturn(2),
sigsuspend(2),
socket(2),
socketcall(2),
socketpair(2),
ssetmask(2),
stat(2), stat64,
statfs(2),
stime(2),
stty,
swapoff(2),
swapon(2),
symlink(2),
sync(2),
sysfs(2),
sysinfo(2),
syslog(2),
time(2),
times(2),
truncate(2), truncate64,
ulimit,
umask(2),
umount(2),
uname(2),
unlink(2),
uselib(2),
ustat(2),
utime(2),
vfork(2),
vhangup(2),
vm86(2),
vm86old,
wait4(2),
waitpid(2),
write(2),
writev(2).
Of the above, 9 are obsolete, namely
getrlimit, oldfstat, oldlstat, oldolduname, oldstat, olduname,
readdir, select and vm86old
This man page lists those system calls that are common to most platforms.
In the list, the
.I Kernel
column indicates the kernel version
for those system calls that were new in Linux 2.2,
or have appeared since that kernel version.
Note the following points:
.IP * 3
Where no kernel version is indicated,
the system call appeared in kernel 2.0 or earlier.
.IP *
Where a system call is marked "2.2"
this means the system call probably appeared in a 2.1.x kernel version,
and first appeared in a stable kernel with 2.2.0.
(No new system calls were added in the stable
kernel 2.0.x series.
The 2.2 kernel was created as a branch of kernel 2.0.21 via the 2.1.x
unstable kernel series)
.IP *
Where a system call is marked "2.4"
this means the system call probably appeared in a 2.3.x kernel version,
and first appeared in a stable kernel with 2.4.0.
(No new system calls were added in the stable
kernel 2.2.x series.
The 2.4 kernel was created as a branch of kernel 2.2.8 via the 2.3.x
unstable kernel series.)
.IP *
Where a system call is marked "2.6"
this means the system call probably appeared in a 2.5.x kernel version,
and first appeared in a stable kernel with 2.6.0.
.IP *
Starting with kernel 2.6.0, the development model changed,
and new system calls may appear in each 2.6.x release.
Kernel 2.6.0 was formed from a branch of kernel 2.4.15 via the 2.5.x
unstable kernel series.
In some cases, a system call appeared in 2.6.x and was also backported
into a 2.4.x release after 2.4.15;
where this is so, the version where the system call appeared
in both of the major kernel series is listed.
(Where a system call is listed as appearing in kernel 2.4.x, x <= 15,
then implicitly it also appeared in the later kernel 2.6.0.)
.PP
The list of system calls that are available as at kernel 2.6.22
(or in a few cases only on older kernels) is as follows:
.\"
.\" Looking at scripts/checksyscalls.sh in the kernel source is
.\" instructive about i386 specifics.
.\"
.TS
l l l
---
l l l.
System call Kernel Notes
\fB_llseek\fP(2)
\fB_newselect\fP(2)
\fB_sysctl\fP(2)
\fBaccept\fP(2)
\fBaccess\fP(2)
\fBacct\fP(2)
\fBadd_key\fP(2) 2.6.11
\fBadjtimex\fP(2)
\fBafs_syscall\fP(2)
\fBalarm\fP(2)
.\" alloc_hugepages (2) appeared in 2.5.36 and went away in 2.5.44
.\" in 2.4.x, the syscall number exists but give ENOSYS
\fBbdflush\fP(2)
\fBbind\fP(2)
\fBbreak\fP(2)
\fBbrk\fP(2)
\fBcacheflush\fP(2) Various architectures;
but not on i386
\fBcapget\fP(2) 2.2
\fBcapset\fP(2) 2.2
\fBchdir\fP(2)
\fBchmod\fP(2)
\fBchown\fP(2)
\fBchown32\fP(2) 2.4
\fBchroot\fP(2)
\fBclock_getres\fP(2) 2.6
\fBclock_gettime\fP(2) 2.6
\fBclock_nanosleep\fP(2) 2.6
\fBclock_settime\fP(2) 2.6
\fBclone\fP(2)
\fBclose\fP(2)
\fBconnect\fP(2)
\fBcreat\fP(2)
\fBcreate_module\fP(2)
\fBdelete_module\fP(2)
\fBdup\fP(2)
\fBdup2\fP(2)
\fBepoll_create\fP(2) 2.6
\fBepoll_ctl\fP(2) 2.6
\fBepoll_pwait\fP(2) 2.6.19
\fBepoll_wait\fP(2) 2.6
\fBeventfd\fP(2) 2.6.22
\fBexecve\fP(2)
\fBexit\fP(2)
\fBexit_group\fP(2) 2.6; 2.4.20
\fBfaccessat\fP(2) 2.6.16
\fBfadvise64\fP(2) 2.6 Implements \fBposix_fadvise\fP(2)
\fBfadvise64_64\fP(2) 2.6 Implements \fBposix_fadvise\fP(2)
\fBfchdir\fP(2)
\fBfchmod\fP(2)
\fBfchmodat\fP(2) 2.6.16
\fBfchown\fP(2)
\fBfchown32\fP(2) 2.4
\fBfchownat\fP(2) 2.6.16
\fBfcntl\fP(2)
\fBfcntl64\fP(2) 2.4
\fBfdatasync\fP(2)
\fBfgetxattr\fP(2) 2.6; 2.4.18
\fBflistxattr\fP(2) 2.6; 2.4.18
\fBflock\fP(2)
\fBfork\fP(2)
.\" free_hugepages (2) appeared in 2.5.36 and went away in 2.5.44
.\" in 2.4.x, the syscall number exists but give ENOSYS
\fBfremovexattr\fP(2) 2.6; 2.4.18
\fBfsetxattr\fP(2) 2.6; 2.4.18
\fBfstat\fP(2)
\fBfstat64\fP(2) 2.4
\fBfstatat64\fP(2) 2.6.16?
\fBfstatfs\fP(2)
\fBfstatfs64\fP(2) 2.6
\fBfsync\fP(2)
\fBftime\fP(2) Implemented in glibc;
see \fBftime\fP(3)
\fBftruncate\fP(2)
\fBftruncate64\fP(2) 2.4
\fBfutex\fP(2) 2.6; 2.4.19
\fBfutimesat\fP(2) 2.6.16
\fBget_kernel_syms\fP(2)
\fBget_mempolicy\fP(2) 2.6.6
\fBget_robust_list\fP(2) 2.6.17
\fBget_thread_area\fP(2) 2.6; 2.4.20
\fBgetcpu\fP(2) 2.6.19
\fBgetcwd\fP(2) 2.2
\fBgetdents\fP(2)
\fBgetdents64\fP(2) 2.4
\fBgetegid\fP(2)
\fBgetegid32\fP(2) 2.4
\fBgeteuid\fP(2)
\fBgeteuid32\fP(2) 2.4
\fBgetgid\fP(2)
\fBgetgid32\fP(2) 2.4
\fBgetgroups\fP(2)
\fBgetgroups32\fP(2) 2.4
\fBgetitimer\fP(2)
\fBgetpeername\fP(2)
\fBgetpagesize\fP(2) Various architectures;
but not on i386
\fBgetpgid\fP(2)
\fBgetpgrp\fP(2)
\fBgetpid\fP(2)
\fBgetpmsg\fP(2) 2.2 Reserved for STREAMS support
\fBgetppid\fP(2)
\fBgetpriority\fP(2)
\fBgetresgid\fP(2) 2.2
\fBgetresgid32\fP(2) 2.4
\fBgetresuid\fP(2) 2.2
\fBgetresuid32\fP(2) 2.4
\fBgetrlimit\fP(2)
\fBgetrusage\fP(2)
\fBgetsid\fP(2)
\fBgetsockname\fP(2)
\fBgetsockopt\fP(2)
\fBgettid\fP(2) 2.4.11
\fBgettimeofday\fP(2)
\fBgetuid\fP(2)
\fBgetuid32\fP(2) 2.4
\fBgetxattr\fP(2) 2.6; 2.4.18
\fBgtty\fP(2)
\fBidle\fP(2)
\fBinit_module\fP(2)
\fBinotify_add_watch\fP(2) 2.6.13
\fBinotify_init\fP(2) 2.6.13
\fBinotify_rm_watch\fP(2) 2.6.13
\fBio_cancel\fP(2) 2.6; 2.4.20
\fBio_destroy\fP(2) 2.6; 2.4.20
\fBio_getevents\fP(2) 2.6; 2.4.20
\fBio_setup\fP(2) 2.6; 2.4.20
\fBio_submit\fP(2) 2.6; 2.4.20
\fBioctl\fP(2)
\fBioperm\fP(2)
\fBiopl\fP(2)
\fBioprio_get\fP(2) 2.6.13
\fBioprio_set\fP(2) 2.6.13
\fBipc\fP(2) Implements System V IPC calls
\fBkexec_load\fP(2) 2.6.7 Was named sys_kexec_load()
from 2.6.7 to 2.6.16
\fBkeyctl\fP(2) 2.6.11
\fBkill\fP(2)
\fBlchown\fP(2) 2.2
\fBlchown32\fP(2) 2.4
\fBlgetxattr\fP(2) 2.6; 2.4.18
\fBlink\fP(2)
\fBlinkat\fP(2) 2.6.16
\fBlisten\fP(2)
\fBlistxattr\fP(2) 2.6; 2.4.18
\fBllistxattr\fP(2) 2.6; 2.4.18
\fBlock\fP(2)
\fBlookup_dcookie\fP(2) 2.6
\fBlremovexattr\fP(2) 2.6; 2.4.18
\fBlseek\fP(2)
\fBlsetxattr\fP(2) 2.6; 2.4.18
\fBlstat\fP(2)
\fBlstat64\fP(2) 2.4
\fBmadvise\fP(2) 2.4
\fBmadvise1\fP(2) 2.4
\fBmbind\fP(2) 2.6.6
.\" \fBmemory_ordering\fP(2) ??? Sparc64
\fBmigrate_pages\fP(2) 2.6.16
\fBmincore\fP(2) 2.4
\fBmkdir\fP(2)
\fBmkdirat\fP(2) 2.6.16
\fBmknod\fP(2)
\fBmknodat\fP(2) 2.6.16
\fBmlock\fP(2)
\fBmlockall\fP(2)
\fBmmap\fP(2)
\fBmmap2\fP(2) 2.4
\fBmodify_ldt\fP(2)
\fBmount\fP(2)
\fBmove_pages\fP(2) 2.6.18
\fBmprotect\fP(2)
\fBmpx\fP(2)
\fBmq_getsetattr\fP(2) 2.6.6 Implements \fBmq_getattr\fP(3)
and \fBmq_setattr\fP(3)
\fBmq_notify\fP(2) 2.6.6
\fBmq_open\fP(2) 2.6.6
\fBmq_timedreceive\fP(2) 2.6.6
\fBmq_timedsend\fP(2) 2.6.6
\fBmq_unlink\fP(2)
\fBmremap\fP(2)
\fBmsgctl\fP(2)
\fBmsgget\fP(2)
\fBmsgrcv\fP(2)
\fBmsgsnd\fP(2)
\fBmsync\fP(2)
\fBmunlock\fP(2)
\fBmunlockall\fP(2)
\fBmunmap\fP(2)
\fBnanosleep\fP(2)
\fBnfsservctl\fP(2) 2.2
\fBnice\fP(2)
\fBoldfstat\fP(2)
\fBoldlstat\fP(2)
\fBoldolduname\fP(2)
\fBoldstat\fP(2)
\fBolduname\fP(2)
\fBopen\fP(2)
\fBopenat\fP(2) 2.6.16
\fBpause\fP(2)
.\" \fBpciconfig_iobase\fP(2) ??? Alpha, PowerPC, ARM
.\" \fBpciconfig_read\fP(2) ??? Alpha, PowerPC, ARM
.\" \fBpciconfig_write\fP(2) ??? Alpha, PowerPC, ARM
\fBpersonality\fP(2)
.\" \fBperfctr\fP(2) ??? Sparc32, Sparc64
.\" \fBperfmonctl\fP(2) ??? IA-64
\fBphys\fP(2) Unimplemented (no slot
since 2.1.116)
\fBpipe\fP(2)
\fBpivot_root\fP(2) 2.5
\fBpoll\fP(2) 2.2
\fBppoll\fP(2) 2.6.16
\fBprctl\fP(2) 2.2
\fBpread64\fP(2) Added as "pread" in 2.2;
renamed "pread64" in 2.6
\fBprof\fP(2)
\fBprofil\fP(2) Implemented in glibc;
see \fBprofil\fP(3)
\fBpselect6\fP(2) 2.6.16 Implements \fBpselect\fP(2)
\fBptrace\fP(2)
\fBputpmsg\fP(2) 2.2 Reserved for STREAMS support
\fBpwrite64\fP(2) Added as "pwrite" in 2.2;
renamed "pwrite64" in 2.6
\fBquery_module\fP(2) 2.2
\fBquotactl\fP(2)
\fBread\fP(2)
\fBreadahead\fP(2) 2.4.13
\fBreaddir\fP(2) Superseded by \fBgetdents\fP(2)
\fBreadlink\fP(2)
\fBreadlinkat\fP(2) 2.6.16
\fBreadv\fP(2)
\fBreboot\fP(2)
\fBrecv\fP(2)
\fBrecvfrom\fP(2)
\fBrecvmsg\fP(2)
\fBremap_file_pages\fP(2) 2.6
\fBremovexattr\fP(2) 2.6; 2.4.18
\fBrename\fP(2)
\fBrenameat\fP(2) 2.6.16
\fBrequest_key\fP(2) 2.6.11
\fBrestart_syscall\fP(2) 2.6
\fBrmdir\fP(2)
\fBrt_sigaction\fP(2) 2.2
\fBrt_sigpending\fP(2) 2.2
\fBrt_sigprocmask\fP(2) 2.2
\fBrt_sigqueueinfo\fP(2) 2.2 Implements \fBsigqueue\fP(2)
\fBrt_sigreturn\fP(2) 2.2
\fBrt_sigsuspend\fP(2) 2.2
\fBrt_sigtimedwait\fP(2) 2.2
\fBsched_get_priority_max\fP(2)
\fBsched_get_priority_min\fP(2)
\fBsched_getaffinity\fP(2) 2.6; 2.4.19
\fBsched_getparam\fP(2)
\fBsched_getscheduler\fP(2)
\fBsched_rr_get_interval\fP(2)
\fBsched_setaffinity\fP(2) 2.6; 2.4.19
\fBsched_setparam\fP(2)
\fBsched_setscheduler\fP(2)
\fBsched_yield\fP(2)
.\" 2.4.11 added security(2), but this was dropped in 2.5.x
.\" \fBsecurity\fP(2)
.\" (2.4.11; not in 2.6)
\fBselect\fP(2)
\fBsemctl\fP(2)
\fBsemget\fP(2)
\fBsemop\fP(2)
\fBsemtimedop\fP(2) 2.6; 2.4.22
\fBsend\fP(2)
\fBsendfile\fP(2) 2.2
\fBsendfile64\fP(2) 2.6; 2.4.19
\fBsendmsg\fP(2)
\fBsendto\fP(2)
\fBset_mempolicy\fP(2) 2.6.6
\fBset_robust_list\fP(2) 2.6.17
\fBset_thread_area\fP(2) 2.6; 2.4.20
\fBset_tid_address\fP(2) 2.6
\fBset_zone_reclaim\fP(2) 2.6.13 Removed in 2.6.16
\fBsetdomainname\fP(2)
\fBsetfsgid\fP(2)
\fBsetfsgid32\fP(2) 2.4
\fBsetfsuid\fP(2)
\fBsetfsuid32\fP(2) 2.4
\fBsetgid\fP(2)
\fBsetgid32\fP(2) 2.4
\fBsetgroups\fP(2)
\fBsetgroups32\fP(2) 2.4
\fBsethostname\fP(2)
\fBsetitimer\fP(2)
\fBsetpgid\fP(2)
\fBsetpriority\fP(2)
\fBsetregid\fP(2)
\fBsetregid32\fP(2) 2.4
\fBsetresgid\fP(2) 2.2
\fBsetresgid32\fP(2) 2.4
\fBsetresuid\fP(2) 2.2
\fBsetresuid32\fP(2) 2.4
\fBsetreuid\fP(2)
\fBsetreuid32\fP(2) 2.4
\fBsetrlimit\fP(2)
\fBsetsid\fP(2)
\fBsetsockopt\fP(2)
\fBsettimeofday\fP(2)
\fBsetuid\fP(2)
\fBsetuid32\fP(2) 2.4
\fBsetup\fP(2) Removed in 2.2
\fBsetxattr\fP(2) 2.6; 2.4.18
\fBsgetmask\fP(2)
\fBshmat\fP(2)
\fBshmctl\fP(2)
\fBshmdt\fP(2)
\fBshmget\fP(2)
\fBshutdown\fP(2)
\fBsigaction\fP(2)
\fBsigaltstack\fP(2) 2.2
\fBsignal\fP(2)
\fBsignalfd\fP(2) 2.6.22
\fBsigpending\fP(2)
\fBsigprocmask\fP(2)
\fBsigreturn\fP(2)
\fBsigsuspend\fP(2)
\fBsocket\fP(2)
\fBsocketcall\fP(2) Implements BSD socket calls
\fBsocketpair\fP(2)
\fBsplice\fP(2) 2.6.17
.\" \fBspu_create\fP(2) 2.6.16 PowerPC only
.\" \fBspu_run\fP(2) 2.6.16 PowerPC only
\fBssetmask\fP(2)
\fBstat\fP(2)
\fBstat64\fP(2) 2.4
\fBstatfs\fP(2)
\fBstatfs64\fP(2) 2.6
\fBstime\fP(2)
\fBstty\fP(2)
\fBswapoff\fP(2)
\fBswapon\fP(2)
\fBsymlink\fP(2)
\fBsymlinkat\fP(2) 2.6.16
\fBsync\fP(2)
\fBsync_file_range\fP(2) 2.6.17
\fBsysfs\fP(2)
\fBsysinfo\fP(2)
\fBsyslog\fP(2) glibc interface is \fBklogctl\fP(3)
\fBtee\fP(2) 2.6.17
\fBtgkill\fP(2) 2.6
\fBtime\fP(2)
\fBtimer_create\fP(2) 2.6
\fBtimer_delete\fP(2) 2.6
\fBtimer_getoverrun\fP(2) 2.6
\fBtimer_gettime\fP(2) 2.6
\fBtimer_settime\fP(2) 2.6
\fBtimerfd\fP(2) 2.6.22
\fBtimes\fP(2)
\fBtkill\fP(2) 2.6; 2.4.22
\fBtruncate\fP(2)
\fBtruncate64\fP(2) 2.4
\fBugetrlimit\fP(2) 2.4
\fBulimit\fP(2) Implemented in glibc;
see \fBulimit\fP(3)
\fBumask\fP(2)
\fBumount\fP(2)
.\" sys_oldumount() -- __NR_umount
\fBumount2\fP(2) 2.2
.\" sys_umount() -- __NR_umount2
\fBuname\fP(2)
\fBunlink\fP(2)
\fBunlinkat\fP(2) 2.6.16
\fBunshare\fP(2) 2.6.16
\fBuselib\fP(2)
\fBustat\fP(2)
\fButime\fP(2)
\fButimensat\fP(2) 2.6.22
\fButimes\fP(2) 2.6
\fBvfork\fP(2)
\fBvhangup\fP(2)
\fBvm86old\fP(2) Superseded by \fBvm86\fP()
\fBvmsplice\fP(2) 2.6.17
\fBvserver\fP(2) 2.6
\fBwait4\fP(2)
\fBwaitid\fP(2) 2.6.10
\fBwaitpid\fP(2)
\fBwrite\fP(2)
\fBwritev\fP(2)
.TE
.PP
Note the following points:
.\" .IP * 3
.\" These system calls are obsolete:
.\" getrlimit,
.\" .\" .\" oldfstat,
.\" .\" .\" oldlstat,
.\" .\" .\" oldolduname,
.\" .\" .\" oldstat,
.\" .\" .\" olduname,
.\" .\" .\" readdir,
.\" .\" .\" select,
.\" .\" .\" and
.\" .\" .\" vm86old
.IP * 3
Although slots are reserved for them in the system call table,
the following system calls are not implemented in the standard kernel:
.BR afs_syscall (2), \" __NR_afs_syscall is 53 on Linux 2.6.22/i386
.BR break (2), \" __NR_break is 17 on Linux 2.6.22/i386
.BR ftime (2), \" __NR_ftime is 35 on Linux 2.6.22/i386
.BR getpmsg (2), \" __NR_getpmsg is 188 on Linux 2.6.22/i386
.BR gtty (2), \" __NR_gtty is 32 on Linux 2.6.22/i386
.BR idle (2), \" __NR_idle is 112 on Linux 2.6.22/i386
.BR lock (2), \" __NR_lock is 53 on Linux 2.6.22/i386
.BR madvise1 (2), \" __NR_madvise1 is 219 on Linux 2.6.22/i386
.BR mpx (2), \" __NR_mpx is 66 on Linux 2.6.22/i386
.BR phys (2), \" Slot has been re-used
.BR prof (2), \" __NR_prof is 44 on Linux 2.6.22/i386
.BR profil (2), \" __NR_profil is 98 on Linux 2.6.22/i386
.BR putpmsg (2), \" __NR_putpmsg is 189 on Linux 2.6.22/i386
.BR security (2),
.\" __NR_putpmsg is 223 on Linux 2.4/i386; absent on 2.6/i386, present
.\" on a a couple og 2.6 architectures
.BR stty (2), \" __NR_stty is 31 on Linux 2.6.22/i386
.BR ulimit (2), \" __NR_ulimit is 58 on Linux 2.6.22/i386
and
.BR vserver (2) \" __NR_vserver is 273 on Linux 2.6.22/i386
(see also
.BR obsolete (2)),
and 15 are unimplemented in the standard kernel, namely
afs_syscall, break, ftime, getpmsg, gtty, idle, lock, mpx, phys,
prof, profil, putpmsg, security, stty and ulimit (see also
.BR unimplemented (2)).
However,
.BR ftime (3),
@ -258,13 +515,21 @@ However,
and
.BR ulimit (3)
exist as library routines.
The slot for phys is in use since 2.1.116 for umount;
phys will never be implemented.
The getpmsg and putpmsg calls are for
kernels patched to support streams, and may never be in the standard
kernel.
The security call is for future use.
The slot for
.BR phys (2)
is in use since kernel 2.1.116 for
.BR umount (2);
.BR phys (2)
will never be implemented.
.IP *
The
.BR getpmsg (2)
and
.BR putpmsg (2)
calls are for kernels patched to support STREAMS,
and may never be in the standard kernel.
.\" The security call is for future use.
.PP
Roughly speaking, the code belonging to the system call
with number __NR_xxx defined in
.I /usr/include/asm/unistd.h
@ -279,52 +544,169 @@ On platforms with
proprietary OS emulation, such as parisc, sparc, sparc64 and alpha,
there are many additional system calls; mips64 also contains a full
set of 32-bit system calls.
Below the details for Linux 2.4.17.
The defines __NR_oldstat and __NR_stat refer to the routines
sys_stat() and sys_newstat(), and similarly for
.I fstat
Over time, changes to the interfaces of some system calls have been
necessary.
One reason for such changes was the need to increase the size of
structures or scalar values passed to the system call.
Because of these changes, there are now multiple implementations
of certain system calls.
These different versions are incompatible at the binary level,
but applications are generally unaware of this: glibc does some magic
to ensure that existing binaries use the version of the system
call that was current when the binary was created,
so that ABI compatibility is preserved.
Examples of systems calls that exist in multiple versions are
the following:
.IP * 3
By now there are three different implementations of
.BR stat (2):
.IR sys_stat ()
(slot
.IR __NR_oldstat ),
.IR sys_newstat ()
(slot
.IR __NR_stat ),
and
.IR lstat .
Similarly, the defines __NR_oldolduname, __NR_olduname and
__NR_uname refer to the routines sys_olduname(), sys_uname()
and sys_newuname().
Thus, __NR_stat and __NR_uname have always referred to the latest
version of the system call, and the older ones are for backward
compatibility.
.IR sys_stat64()
(slot
.IR __NR_stat64 ),
with the last being the most current.
.\" e.g., on 2.6.22/x86: __NR_oldstat 18, __NR_stat 106, __NR_stat64 195
.\" The stat system calls deal with three different data structures,
.\" defined in include/asm-i386/stat.h: __old_kernel_stat, stat, stat64
A similar story applies for
.BR lstat (2)
and
.BR fstat (2).
.IP *
Similarly, the defines
.IR __NR_oldolduname ,
.IR __NR_olduname ,
and
.IR __NR_uname
refer to the routines
.IR sys_olduname (),
.IR sys_uname ()
and
.IR sys_newuname ().
.IP *
In Linux 2.0,
a new version of
.BR vm86 (2)
appeared, with the old and the new kernel routines being named
.IR sys_vm86old ()
and
.IR sys_vm86 ().
.IP *
In Linux 2.4,
a new version of
.BR getrlimit (2)
appeared, with the old and the new kernel routines being named
.IR sys_old_getrlimit ()
(slot
.IR __NR_getrlimit )
and
.IR sys_getrlimit ()
(slot
.IR __NR_ugetrlimit ).
.IP *
Linux 2.4 increased the size of user and group IDs from 16 to 32 bits.
.\" 64 bit off_t changes: ftruncate64, *stat64,
.\" fcntl64 (because of the flock structure), getdents64, *statfs64
To support this change, a range of system calls were added
(e.g.,
.BR chown32 (2),
.BR getuid32 (2),
.BR getgroups32 (2),
.BR setresuid32 (2)),
superseding earlier calls of the same name without the
"32" suffix.
.IP *
Linux 2.4 added support for applications on 32-bit architectures
to access large files (i.e., files for which the sizes and
file offsets can't be represented in 32 bits.)
To support this change, replacements were required for system calls
that deal with file offsets and sizes.
Thus the following system calls were added:
.BR fcntl64 (2),
.BR ftruncate64 (2),
.BR getdents64 (2),
.BR stat64 (2),
.BR statfs64 (2),
and their analogs that work with file descriptors or
symbolic links.
These system calls supersede the older system calls
which, except in the case of the "stat" calls,
have the same name without the "64" suffix.
.sp
On newer platforms that only have 64-bit file access and 32-bit uids
(e.g., alpha, ia64, s390x) there are no *64 or *32 calls.
Where the *64 and *32 calls exist, the other versions are obsolete.
.IP *
The
.I rt_sig*
calls were added in kernel 2.2 to support the addition
of real-time signals (see
.BR signal (7)).
These system calls supersede the older system calls of the same
name without the "rt_" prefix.
.IP *
It is different with
.I select
.B select (2)
and
.IR mmap .
.BR mmap (2).
These use five or more parameters, and caused problems the way
parameter passing on the i386 used to be set up.
Thus, while
other architectures have sys_select() and sys_mmap() corresponding
to __NR_select and __NR_mmap, on i386 one finds old_select()
and old_mmap() (routines that use a pointer to a
other architectures have
.IR sys_select ()
and
.IR sys_mmap ()
corresponding to
.IR __NR_select
and
.IR __NR_mmap ,
on i386 one finds
.IR old_select ()
and
.IR old_mmap ()
(routines that use a pointer to a
parameter block) instead.
These days passing five parameters
is not a problem any more, and there is a __NR__newselect (used by
libc 6) that corresponds directly to sys_select() and similarly __NR_mmap2.
Two other system call numbers, __NR__llseek and __NR__sysctl
have an additional underscore absent in sys_llseek() and sys_sysctl().
Then there is __NR_readdir corresponding to old_readdir(),
which will read at most one directory entry at a time, and is
superseded by sys_getdents().
is not a problem any more, and there is a
.IR __NR__newselect
(used by
libc 6) that corresponds directly to
.IR sys_select ()
and similarly
.IR __NR_mmap2.
.PP
Two system call numbers,
.IR __NR__llseek
and
.IR __NR__sysctl
have an additional underscore absent in
.IR sys_llseek ()
and
.IR sys_sysctl ().
On many platforms, including i386, socket calls are all multiplexed
through socketcall() and System V IPC calls through ipc().
through
.BR socketcall (2)
and System V IPC calls are multiplexed through
.BR ipc (2).
On newer platforms that only have 64-bit file access and 32-bit uids
(e.g., alpha, ia64, s390x) there are no *64 or *32 calls.
Where the *64
and *32 calls exist, the other versions are obsolete.
The chown and lchown system calls were swapped in 2.1.81.
The *64 and
*32 calls were added for kernel 2.4, as were the new versions of
getrlimit and mmap, and the new calls pivot_root, mincore, madvise,
security, gettid and readahead.
In kernel 2.1.81,
.BR lchown (2)
and
.BR chown (2)
were swapped; that is,
.BR lchown (2)
was added with the semantics that were then current for
.BR chown (2),
and the semantics of the latter call were changed to what
they are today.
.SH SEE ALSO
.BR syscall (2),
.BR unimplemented (2)