There doesn't seem to be any reason for getauxval.3 to be
executable...
Signed-off-by: Rasmus Villemoes <rv@rasmusvillemoes.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
The buflen argument to getnetgrent_r has type size_t.
Signed-off-by: Rasmus Villemoes <rv@rasmusvillemoes.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
The header <rpc/des_crypt.h> does not provide rtime();
<rpc/auth_des.h> does, as is also implied in both the NOTES and
EXAMPLE sections.
Signed-off-by: Rasmus Villemoes <rv@rasmusvillemoes.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
The user, passwd and cmd arguments to rexec and rexec_af are all
const.
Signed-off-by: Rasmus Villemoes <rv@rasmusvillemoes.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
re_comp and re_exec take const char* arguments.
Signed-off-by: Rasmus Villemoes <rv@rasmusvillemoes.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
Unlike the BSDs, the second argument of rcmd() and rcmd_af() has
type unsigned short.
The first argument of iruserok_af() has type const void*.
Signed-off-by: Rasmus Villemoes <rv@rasmusvillemoes.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
Add return type for pthread_setschedprio.
Signed-off-by: Rasmus Villemoes <rv@rasmusvillemoes.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
Add return type for pthread_{s,g}etschedparam.
Signed-off-by: Rasmus Villemoes <rv@rasmusvillemoes.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
The name parameter of pthread_getname_np is an output parameter and
hence not const.
Signed-off-by: Rasmus Villemoes <rv@rasmusvillemoes.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
Each of the pthread_attr_get* functions extract some piece of
information from a pthread_attr_t, which is passed by const
reference. Add the const keyword to the prototypes of these
functions.
Signed-off-by: Rasmus Villemoes <rv@rasmusvillemoes.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
The elements of the array sys_errlist are also const.
Also, it is probably not desirable to pretend that errno.h defines
errno in the traditional way (errno(3) correctly contains an explicit
warning against this). I don't know what the best solution is, though.
Signed-off-by: Rasmus Villemoes <rv@rasmusvillemoes.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
The newattr parameter to mq_setattr is const.
Signed-off-by: Rasmus Villemoes <rv@rasmusvillemoes.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
As mentioned further down, malloc_trim returns an integer.
Signed-off-by: Rasmus Villemoes <rv@rasmusvillemoes.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
gnu_dev_makedev, and hence its trivial macro wrapper makedev, takes
two unsigned int parameters; this is consistent with it being the
inverse of (gnu_dev_)major/minor, which return unsigned int.
Signed-off-by: Rasmus Villemoes <rv@rasmusvillemoes.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
The second argument to swapcontext() is const.
Signed-off-by: Rasmus Villemoes <rv@rasmusvillemoes.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
The parameters to inet_makeaddr have type in_addr_t.
Signed-off-by: Rasmus Villemoes <rv@rasmusvillemoes.dk>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
The <asm/page.h> header is not readily available, and the comment
seems to indicate that it is for getting PAGE_SIZE. But it is
never mentioned why one would need to know that, and it is in any
case better obtained using sysconf(), provided by <unistd.h>.
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
The prototype for getrpcport() is obtained by #include'ing
<rpc/rpc.h>. Also, update its prototype.
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
The hostlen and servlen parameters to have type socklet_t.
mtk: The types changed in glibc 2.2, with commit
e10b8512ce40cfb9f706e40edc7ed0ebdd151471
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
Since the term "procfs" is widely used, it seems reasonable to have
a link from that name to proc(5).
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
Since PIDs > /proc/sys/kernel/pid_max are not allocated, this
file thus also imposes a system-wide limit on the number of
threads and processes.
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>