While parisc would normally have the same behavior as ARM/PowerPC,
they decide to write shim syscall stubs to unpack/realign rather
than expose the padding to userspace.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
Mention a few other system calls that create file descriptors
that display an 'anon_inode' symlink in /proc/PID/fd
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
Krzysztof Kulakowski's patch updated the MPOL_BIND description
to describe current behavior, but the page should retain a
description of the older behavior as well.
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
Behavior of MPOL_BIND has been changed in kernel commit
19770b32609b6bf97a3dece2529089494cbfc549 (Linux 2.6.26). This
patch updates section describing MPOL_BIND to be in sync with
current state of Documentation/vm/numa_memory_policy.txt.
Signed-off-by: Krzysztof Kulakowski <krzysztof.kulakowski@intel.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
Architectures that split 64-bit values across register pairs
usually do so according to their C ABI calling convention (which
means endianness). Add some notes to that effect, and change the
readahead example to show a little endian example (since that is
way more common than big endian).
Also start a new list of syscalls that this issue does not apply
to.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
nanosleep() has a problem if used in a program that catches
signals and those signals are delivered at a very high rate.
Describe the problem, and note that clock_nanosleep(2)
provides a solution.
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
Andreas' patch added a second description of E2BIG that
was (mostly) more detailed than the existing text. Combine
the two texts.
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
The page already (by now) contains a reference to open(2)
for a discussion of open file descriptions. Leave it at that,
since the reader can then deduce how things work.
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
/proc/pid/environ reflects process environment at
*start* of program execution.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
Make the reader's life a little easier by saving them from
having to refer to mq_getattr(3).
Reported-by: Adam Martindale <adam.john.martindale@gmail.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
See __send_signal() in kernel/signal.c:
if (sig >= SIGRTMIN && info->si_code != SI_USER) {
/*
* Queue overflow, abort. We may abort if the
* signal was rt and sent by user using something
* other than kill().
*/
result = TRACE_SIGNAL_OVERFLOW_FAIL;
ret = -EAGAIN;
goto ret;
}
(kill() uses SI_USER, but tkill() and tgkill() use SI_TKILL.)
Signed-off-by: Jann Horn <jannh@google.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
The description of FAT_IOCTL_GET_VOLUME_ID references volume name,
when it really should be volume ID since the volume label is
different in the FAT filesystem. This patch renames the incorrect
volume name references to volume id.
This man page supplies an example display_fat_volume_id.c program
that can be used to read the volume ID (serial number) from a
filesystem. Here is an additional test showing that the volume ID
and volume name are two different entities:
$ dd if=/dev/zero of=fat_volume bs=1M count=1
$ mkfs.fat -v -n MASNEYB fat_volume
[snip]
Volume ID is da8cecf2, volume label MASNEYB .
$ sudo mount -o loop fat_volume /mnt
$ ./display_fat_volume_id /mnt/
Volume ID da8c-ecf2
Signed-off-by: Brian Masney <masneyb@onstation.org>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>