]> git.ipfire.org Git - thirdparty/util-linux.git/log
thirdparty/util-linux.git
10 years agobuild-sys: cleanup realtime lib usage
Karel Zak [Fri, 6 Mar 2015 13:20:19 +0000 (14:20 +0100)] 
build-sys: cleanup realtime lib usage

* check for timer_create()
* define dependence on timer_create() for flock
* rename CLOCKGETTIME_LIBS to REALTIME_LIBS

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agolib/timer: use separate file for timers
Karel Zak [Fri, 6 Mar 2015 12:17:20 +0000 (13:17 +0100)] 
lib/timer: use separate file for timers

It seems that static builds require -lpthread for timer_* functions.
It's better to keep it out of our libs (e.g. libmount) to avoid
unnecessary dependence.

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agoMerge branch 'logger-man-syslog-routine' of https://github.com/rgerhards/util-linux
Karel Zak [Fri, 6 Mar 2015 11:36:17 +0000 (12:36 +0100)] 
Merge branch 'logger-man-syslog-routine' of https://github.com/rgerhards/util-linux

* 'logger-man-syslog-routine' of https://github.com/rgerhards/util-linux:
  logger man page: bump man page date to current
  logger man page: fix some "syslog(3) routine" remnants

10 years agologger man page: bump man page date to current
Rainer Gerhards [Fri, 6 Mar 2015 11:32:12 +0000 (12:32 +0100)] 
logger man page: bump man page date to current

10 years agologger man page: fix some "syslog(3) routine" remnants
Rainer Gerhards [Fri, 6 Mar 2015 11:30:56 +0000 (12:30 +0100)] 
logger man page: fix some "syslog(3) routine" remnants

10 years agoMerge branch 'logger-fix-tcp-framing' of https://github.com/rgerhards/util-linux
Karel Zak [Fri, 6 Mar 2015 11:31:48 +0000 (12:31 +0100)] 
Merge branch 'logger-fix-tcp-framing' of https://github.com/rgerhards/util-linux

* 'logger-fix-tcp-framing' of https://github.com/rgerhards/util-linux:
  logger: bugfix: tcp syslog framing is broken, -T unusable
  logger: refactor the way output is written

10 years agologger: use username as the default tag
Karel Zak [Fri, 6 Mar 2015 11:15:23 +0000 (12:15 +0100)] 
logger: use username as the default tag

Reported-by: Rainer Gerhards <rgerhards@adiscon.com>
Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agologger: bugfix: tcp syslog framing is broken, -T unusable
Rainer Gerhards [Fri, 6 Mar 2015 11:12:15 +0000 (12:12 +0100)] 
logger: bugfix: tcp syslog framing is broken, -T unusable

Logger can send via plain tcp syslog if -n -T options are given.
However, the framing is broken so that a syslog receiver can not
know where the first message ends and the next one starts. It
actually looks like no framing at all is used. Plain TCP syslog
framing is described in RFC6587.

This patch adds RFC6587 octet-stuffed framing to TCP syslog. For
local logging, this is always fine, for remote logging this is
NOT recommended by the IETF (the RFC is historic). However, a
full blown RFC5425 TLS sender seems to be out of scope for a tool
like logger IMO.

This patch also refactors the way output is written, seperating
the message format generators from the output writer.

10 years agologger: refactor the way output is written
Rainer Gerhards [Fri, 6 Mar 2015 10:51:31 +0000 (11:51 +0100)] 
logger: refactor the way output is written

Previously, output was written in exactly the same way in three
different places. This is now combined into a single function. This
hopefully makes it easier to adapt to changing output needs.

10 years agologger: remove "interface to syslog(3)" from man page
Karel Zak [Fri, 6 Mar 2015 10:48:23 +0000 (11:48 +0100)] 
logger: remove "interface to syslog(3)" from man page

since v2.26 logger does not use syslog(3) anymore

Reported-by: Rainer Gerhards <rgerhards@adiscon.com>
Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agoMerge branch 'logger-kernel-regression' of https://github.com/rgerhards/util-linux
Karel Zak [Fri, 6 Mar 2015 10:32:17 +0000 (11:32 +0100)] 
Merge branch 'logger-kernel-regression' of https://github.com/rgerhards/util-linux

* 'logger-kernel-regression' of https://github.com/rgerhards/util-linux:
  logger: fix -p kern.* priority is accepted regression
  logger: messages are logged with kern.* priority by default

10 years agologger: fix whitespace and compiler warning [-Wunused-variable]
Karel Zak [Fri, 6 Mar 2015 10:27:16 +0000 (11:27 +0100)] 
logger: fix whitespace and compiler warning [-Wunused-variable]

misc-utils/logger.c: In function ‘syslog_rfc3164’:
misc-utils/logger.c:336:9: warning: unused variable ‘now’ [-Wunused-variable]

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agoMerge branch 'logger-fix-local-timestamp' of https://github.com/rgerhards/util-linux
Karel Zak [Fri, 6 Mar 2015 10:25:56 +0000 (11:25 +0100)] 
Merge branch 'logger-fix-local-timestamp' of https://github.com/rgerhards/util-linux

* 'logger-fix-local-timestamp' of https://github.com/rgerhards/util-linux:
  logger: fix invalid timestamp regression in local format

10 years agologger: add comment
Karel Zak [Fri, 6 Mar 2015 10:05:30 +0000 (11:05 +0100)] 
logger: add comment

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agoMerge branch 'logger-format-fix' of https://github.com/rgerhards/util-linux
Karel Zak [Fri, 6 Mar 2015 10:04:08 +0000 (11:04 +0100)] 
Merge branch 'logger-format-fix' of https://github.com/rgerhards/util-linux

* 'logger-format-fix' of https://github.com/rgerhards/util-linux:
  logger: fix inconsistent format regression when logging locally

10 years agolibblkid: fix compiler warning [-Wunused-variable]
Karel Zak [Fri, 6 Mar 2015 09:57:12 +0000 (10:57 +0100)] 
libblkid: fix compiler warning [-Wunused-variable]

libblkid/src/superblocks/zfs.c: In function 'probe_zfs':
libblkid/src/superblocks/zfs.c:199:11: warning: unused variable 'swab_magic' [-Wunused-variable]

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agoMerge branch 'zfs' of https://github.com/mihu/util-linux
Karel Zak [Fri, 6 Mar 2015 09:53:22 +0000 (10:53 +0100)] 
Merge branch 'zfs' of https://github.com/mihu/util-linux

* 'zfs' of https://github.com/mihu/util-linux:
  zfs: make less syscalls
  blkid: make zfs detection more robust

10 years agologger: fix -p kern.* priority is accepted regression
Rainer Gerhards [Thu, 5 Mar 2015 15:54:34 +0000 (16:54 +0100)] 
logger: fix -p kern.* priority is accepted regression

Pre 2.26, syslog(3) was used for local logging, and it did not
accept kern.* priorities. This is re-enabled by the patch.

10 years agologger: messages are logged with kern.* priority by default
Rainer Gerhards [Thu, 5 Mar 2015 15:44:03 +0000 (16:44 +0100)] 
logger: messages are logged with kern.* priority by default

The default should be user.notice and kern.* should never be used
(syslog(3) forbids this).

This is a severe regression, as messages are now logged to the wrong
bin or not at all. So they get lost and may confuse readers of the
kernel bin.

regression from 2.25.2 to 2.26

10 years agobuild-sys: remove static builds from make-check
Karel Zak [Thu, 5 Mar 2015 14:34:42 +0000 (15:34 +0100)] 
build-sys: remove static builds from make-check

.. temporary, the final solution has to cleanup timer_* functions
usage. The function are unnecessary for libmount, but we use
lib/monotme.c in the library and it probably requires -lrt and *also*
-lpthread for static builds.

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agologger: fix invalid timestamp regression in local format
Rainer Gerhards [Thu, 5 Mar 2015 14:20:50 +0000 (15:20 +0100)] 
logger: fix invalid timestamp regression in local format

Since 1d57503378bdcd838365d625f6d2d0a09da9c29d logger no longer uses
the syslog(3) call. The way the local timestamp is generated did not
match the syslog(3) format. Most importantly, the month name is
formatted based on the user's local. For example:

$ ./logger --stderr  test with logger 2.26.39-eb651-dirty
<5>Mär  5 14:17:47 logger: test with logger 2.26.39-eb651-dirty

"Mär" like in German "März" for "March".

previously:
$ logger --stderr  test with logger 2.25.2
rger: test with logger 2.25.2

In the system log file, this results to the following:

Mar  5 14:17:47 host Mär  5 14:17:47 logger: test with logger 2.26.39-eb651-dirty
Mar  5 14:18:01 host rger: test with logger 2.25.2

This local naming is invalid as of RFC3164. One may argue that
the local log socket traditionally does not have RFC3164 format,
but the timestamp always was as defined in RFC3164 (and along
the lines of the ctime() call). Anything else would also be impractical,
as a syslog parser would otherwise need to know about all
potential locale-specific representations of month names.

This patch corrects the problem and also refactors the timestamp
handling a bit. The same timestamp is needed in local and rfc3164
processing, so there now is a new function to create that stamp.

10 years agologger: fix inconsistent format regression when logging locally
Rainer Gerhards [Wed, 4 Mar 2015 10:17:20 +0000 (11:17 +0100)] 
logger: fix inconsistent format regression when logging locally

The message format when writing to local sockets is inconsistent. Example:

$ ./logger --stderr test
<5>Mär  4 11:03:30 logger: test
$ ./logger -u /dev/log --stderr test
<5>1 2015-03-04T11:03:31.699841+0100 ubuntu1404esp rger - [timeQuality tzKnown="1" isSynced="1" syncAccuracy="29000"] test

The regression was introduced with 4de2e8a03859aaab2c25dc98f33409cd28de6acc
As far as the commit comments and man page indicates, this was meant to affect
remote system logging only, but it also affects local logging when the -u
option is given.

This causes problems with receivers who do not expect full-blown RFC format
on the log socket, like rsyslog. In consequence, this can also affect
log analysis programs and invalidate some of their results.

The patch corrects the behaviour so that the same old-style format is used for
any type of local logging. New-style can always be selected by command line-options.
RFC5424 is still the default for remote logging, as intended in the orignal
commit.

Result with the patch:
$ ./logger --stderr test
<5>Mär  4 11:15:35 logger: test
$ ./logger -u /dev/log --stderr test
<5>Mär  4 11:15:40 logger: test
$ ./logger -u /dev/log --rfc5424 --stderr test
<5>1 2015-03-04T11:21:28.796170+0100 ubuntu1404esp rger - [timeQuality tzKnown="1" isSynced="1" syncAccuracy="27500"] test

10 years agolib/monotonic: fix compiler warnings
Karel Zak [Thu, 5 Mar 2015 12:32:53 +0000 (13:32 +0100)] 
lib/monotonic: fix compiler warnings

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agoagetty: support /usr/lib/os-release too
Karel Zak [Thu, 5 Mar 2015 11:25:28 +0000 (12:25 +0100)] 
agetty: support /usr/lib/os-release too

http://www.freedesktop.org/software/systemd/man/os-release.html

The file /etc/os-release takes precedence over /usr/lib/os-release.
Applications should check for the former, and exclusively use its data
if it exists, and only fall back to /usr/lib/os-release if it is
missing.

Reported-by: Dimitri John Ledkov <dimitri.j.ledkov@intel.com>
Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agoMerge branch 'logger-fix-rfc5424-timestamp' of https://github.com/rgerhards/util...
Karel Zak [Thu, 5 Mar 2015 10:52:00 +0000 (11:52 +0100)] 
Merge branch 'logger-fix-rfc5424-timestamp' of https://github.com/rgerhards/util-linux

* 'logger-fix-rfc5424-timestamp' of https://github.com/rgerhards/util-linux:
  logger: fix invalid timestamp in rfc5425 format

10 years agotailf: use size_t and fwrite()
Karel Zak [Thu, 5 Mar 2015 10:11:53 +0000 (11:11 +0100)] 
tailf: use size_t and fwrite()

Let's use size_t for number of output lines and use fwrite() rather
than while() { putchar() };

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agoflock: fix timeout handler pointer usage
Karel Zak [Thu, 5 Mar 2015 09:47:59 +0000 (10:47 +0100)] 
flock: fix timeout handler pointer usage

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agowhereis: make "-BMS <dir> -f" more obvious in usage()
Karel Zak [Thu, 5 Mar 2015 09:29:10 +0000 (10:29 +0100)] 
whereis: make "-BMS <dir> -f" more obvious in usage()

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agologger: fix -i argument parsing regression
Sami Kerola [Fri, 20 Feb 2015 19:42:34 +0000 (19:42 +0000)] 
logger: fix -i argument parsing regression

With earlier logger it's possible to combine the option -i with other
options, such as -s.  But currently:

$:~> logger -is
logger: failed to parse id: 's'

The changed behaviour breaks existing scripts like dhcpcd-run-hooks from
dhcpcd.

Broken-since: aab5b44405b9a6ada92e419e5a84cc0d1d4afee9
Reference: http://comments.gmane.org/gmane.linux.utilities.util-linux-ng/9683
Reported-by: Juergen Daubert <jue@jue.li>
Reviewed-by: Benno Schulenberg <bensberg@justemail.net>
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
10 years agotailf: ensure file argument really is a file
Sami Kerola [Tue, 17 Feb 2015 21:41:46 +0000 (21:41 +0000)] 
tailf: ensure file argument really is a file

The tailf(1) never worked very well with block or character devices,
sockets, fifos and such.  Now after mmap() is used to find last lines
even the little command used to work for example pipes is broken, so test
the tailf is asked to follow a file and when not fail.  That said
symlinks are OK, as long they point to a file.

Signed-off-by: Sami Kerola <kerolasa@iki.fi>
10 years agotailf: do not allow minus signed last lines argument
Sami Kerola [Sun, 15 Feb 2015 17:12:03 +0000 (17:12 +0000)] 
tailf: do not allow minus signed last lines argument

Before mmap() the command behavior was not completely correct, as
demonstrated below, and after the mmap() it tried to print some eighteen
quintillion lines.

$ tailf -n-1 x
tailf: cannot allocate 18446744073709543424 bytes: Cannot allocate memory

Signed-off-by: Sami Kerola <kerolasa@iki.fi>
10 years agotailf: count last lines correctly at initial print out
Sami Kerola [Sun, 15 Feb 2015 16:52:39 +0000 (16:52 +0000)] 
tailf: count last lines correctly at initial print out

When last lines happen to be greater than string buffer size for fgets()
the number of printed lines resulted to too few.  To avoid miscounts due
insufficient buffer size use mmap() to map the whole file and rewind
until requested number of new lines is found.

Signed-off-by: Sami Kerola <kerolasa@iki.fi>
10 years agotunelp: use parse_switch()
Sami Kerola [Sat, 10 Jan 2015 15:04:27 +0000 (15:04 +0000)] 
tunelp: use parse_switch()

Signed-off-by: Sami Kerola <kerolasa@iki.fi>
10 years agoflock: improve timeout handling
Sami Kerola [Mon, 9 Feb 2015 23:18:09 +0000 (23:18 +0000)] 
flock: improve timeout handling

Signal ALRM raised by the timer, and the timer only, will be considered
as a timeout criteria.

Secondly time interval is made to use monotonic clock.  Documentation of
ITIMER_REAL is unclear whether that time is affected various sources of
clock skew, or does it even tick when system is suspended.

Reviewed-by: Karel Zak <kzak@redhat.com>
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
10 years agologger: fix invalid timestamp in rfc5425 format
Rainer Gerhards [Wed, 4 Mar 2015 17:50:30 +0000 (18:50 +0100)] 
logger: fix invalid timestamp in rfc5425 format

The timestamp is written as

2015-03-04T15:02:02.566782+0100

unfortunately, this is not an RFC3339 timestamp as demanded by rfc5424.
The colon in the time offset field is missing. The correct timestamp is

2015-03-04T15:02:02.566782+01:00

(Note "+0100" vs. "+01:00")

10 years agoflock: add --verbose option
Sami Kerola [Sat, 7 Feb 2015 23:45:24 +0000 (23:45 +0000)] 
flock: add --verbose option

Jenkins script jobs using flock are a great example of a situation in
which one may want an automation to be verbose, so that when unexpected
events happen there is more hints in logs.

Reviewed-by: Benno Schulenberg <bensberg@justemail.net>
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
10 years agowhereis: tell when mandatory option is missing
Sami Kerola [Sat, 7 Feb 2015 21:10:51 +0000 (21:10 +0000)] 
whereis: tell when mandatory option is missing

The <dirs> list is terminated by any arbitrary option, but to be simple
when none is specified complain about -f being missing.

Reviewed-by: Benno Schulenberg <bensberg@justemail.net>
Reviewed-by: Karel Zak <kzak@redhat.com>
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
10 years agobuild-sys: add --enable-usrdir-path
Karel Zak [Tue, 3 Mar 2015 10:36:59 +0000 (11:36 +0100)] 
build-sys: add --enable-usrdir-path

The directories /{sbin,bin} are symliks to /usr/{sbin,bin} on many
systems. This patch add new ./configure option to remove the non-usr
paths from the default $PATH environment variable.

The default $PATH is hardcoded in login(1) and can be overwritten
by /etc/login.defs.

default:

./test_pathnames | grep DEFPATH
       _PATH_DEFPATH /usr/local/bin:/bin:/usr/bin
  _PATH_DEFPATH_ROOT /usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin

--enable-usrdir-path:

./test_pathnames | grep DEFPATH
       _PATH_DEFPATH /usr/local/bin:/usr/bin
  _PATH_DEFPATH_ROOT /usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin

This patch does not modify install paths, you still have to care about
--{bin,lib}dir configure options.

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agolibfdisk: add {first,last}-lba header to sfdisk scritps
Karel Zak [Mon, 2 Mar 2015 12:58:23 +0000 (13:58 +0100)] 
libfdisk: add {first,last}-lba header to sfdisk scritps

The current sfdisk does not allow to create partition that
starts before the default libfdisk First LBA (~1MiB). It
means that

 # sfdisk --dump /dev/sda > foo
 # sfdisk /dev/sdb < foo

does not work on systems where 1st partition does not start at offset
2048.

This patch add new headers to scripts to inform libfdisk about different
First/Last LBA ranges. For example:

 label: gpt
 first-lba: 34

allows to override the library default.

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agozfs: make less syscalls
Michal Humpula [Sat, 28 Feb 2015 20:19:42 +0000 (21:19 +0100)] 
zfs: make less syscalls

10 years agodocs: fstab(5) grammar / English fixes, and some other updates
Peter Cordes [Wed, 25 Feb 2015 02:40:41 +0000 (22:40 -0400)] 
docs: fstab(5) grammar / English fixes, and some other updates

I proofread the whole thing.  I fixed everything that I thought could use
improvement.

various grammar and man page style-guide fixes (commas, word order, etc.).

Reworded a couple things to hopefully make it clear to someone that
didn't already know about fstab.  Re-ordered the intro paragraphs
for easier skimming.  And added an example line.

Expanded on a couple things other things.

Tightened up the wording in some other places to get the point across
faster and in less space.

Thanks to Benno Schulenberg <bensberg@justemail.net>
for several improvements.

Signed-off-by: Peter Cordes <peter@cordes.ca>
10 years agodocs: fix some spelling errors and typos in man pages
Bill Pemberton [Wed, 25 Feb 2015 14:07:59 +0000 (09:07 -0500)] 
docs: fix some spelling errors and typos in man pages

runuser.1: fix spelling implemenation -> implementation
scriptreplay.1: fix spelling overide -> override
unshare.1: fix spelling permamently -> permanently
last.1: fix spelling preferrable -> preferable
lslogins.1: fix spelling priviliges -> privileges
hwclock.8.in: fix spelling transfered -> transferred
prlimit.1: fix typo umlimited -> unlimited
agetty.8: fix typo unnsupported -> unsupported

Signed-off-by: Bill Pemberton <wfp5p@worldbroken.com>
10 years agodocs: update TODO
Karel Zak [Fri, 27 Feb 2015 13:46:43 +0000 (14:46 +0100)] 
docs: update TODO

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agolib/colors: use libtinfo to check terminal capability
Karel Zak [Fri, 27 Feb 2015 12:57:34 +0000 (13:57 +0100)] 
lib/colors: use libtinfo to check terminal capability

The current implementation assumes that all terminals supports colors
and users are forcet to use terminal-colors.d/ to disable colors for
some terminals.

This patch checks for maximal supported colors for the current
terminal and colors are automatically disabled for terminals like
vt100.

The patch moves lib/colors.c from libcommon.la to libtcolors.la to
avoid collisions with another utils.

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agoMerge branch 'fixes' of https://github.com/rudimeier/util-linux
Karel Zak [Thu, 26 Feb 2015 09:43:36 +0000 (10:43 +0100)] 
Merge branch 'fixes' of https://github.com/rudimeier/util-linux

* 'fixes' of https://github.com/rudimeier/util-linux:
  logger: link libsystemd-daemon.so

10 years agoblkid: make zfs detection more robust
Michal Humpula [Wed, 25 Feb 2015 19:25:05 +0000 (20:25 +0100)] 
blkid: make zfs detection more robust

Try to use all the possible uberblock locations.

10 years agocolors: cleanup man pages, add hint to usage()
Karel Zak [Wed, 25 Feb 2015 11:37:55 +0000 (12:37 +0100)] 
colors: cleanup man pages, add hint to usage()

 cfdisk, fdisk, calm dmesg and hexdump

Signed-off-by: Karel Zak <kzak@redhat.com>
Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agosfdisk: add missing --color
Karel Zak [Wed, 25 Feb 2015 11:35:45 +0000 (12:35 +0100)] 
sfdisk: add missing --color

The util already support lib/colors.c stuff, but without command line
option.

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agolib/colors: add USAGE_COLORS_DEFAULT
Karel Zak [Wed, 25 Feb 2015 11:35:05 +0000 (12:35 +0100)] 
lib/colors: add USAGE_COLORS_DEFAULT

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agobuild-sys: add --disable-colors-default
Karel Zak [Wed, 25 Feb 2015 09:06:40 +0000 (10:06 +0100)] 
build-sys: add --disable-colors-default

The current util-linux is to have enabled colorized outputs by
default, this default behavior is possible to change by new configure
option --disable-colors-default.

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agologger: link libsystemd-daemon.so
Ruediger Meier [Tue, 24 Feb 2015 15:26:00 +0000 (16:26 +0100)] 
logger: link libsystemd-daemon.so

Noticed on openSUSE 13.1:
util-linux-2.26/misc-utils/logger.c:735: undefined reference to `sd_booted'

Introduced in d77dc29e.

CC: Sami Kerola <kerolasa@iki.fi>
Signed-off-by: Ruediger Meier <ruediger.meier@ga-group.nl>
10 years agologger: move /dev/log to pathnames.h
Sami Kerola [Sun, 22 Feb 2015 14:41:45 +0000 (14:41 +0000)] 
logger: move /dev/log to pathnames.h

Signed-off-by: Sami Kerola <kerolasa@iki.fi>
10 years agorpmatch: use symbolic value when evaluation return codes
Sami Kerola [Sun, 22 Feb 2015 14:41:40 +0000 (14:41 +0000)] 
rpmatch: use symbolic value when evaluation return codes

Signed-off-by: Sami Kerola <kerolasa@iki.fi>
10 years agoeject: use parse_switch()
Sami Kerola [Sun, 22 Feb 2015 14:41:39 +0000 (14:41 +0000)] 
eject: use parse_switch()

[kzak@redhat.com: - use extended parse_switch()]

Signed-off-by: Sami Kerola <kerolasa@iki.fi>
Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agolib/strutils: extend parse_switch() to accept more options
Karel Zak [Tue, 24 Feb 2015 11:04:22 +0000 (12:04 +0100)] 
lib/strutils: extend parse_switch() to accept more options

* allow to specify more 0|1 pairs
* allow to specify error message

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agolib/strutils: move parse_switch() from setterm(1) to library
Sami Kerola [Sun, 22 Feb 2015 14:41:37 +0000 (14:41 +0000)] 
lib/strutils: move parse_switch() from setterm(1) to library

To allow sharing the code with other utilities.

Signed-off-by: Sami Kerola <kerolasa@iki.fi>
Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agotunelp: remove unnecessary preprocessor directives
Sami Kerola [Sun, 22 Feb 2015 14:41:36 +0000 (14:41 +0000)] 
tunelp: remove unnecessary preprocessor directives

The lp.h included earlier in the tunelp.c has the definitions that were
checked, so these statements could have not been false and such
impossible conditions does not need to be checked.

Signed-off-by: Sami Kerola <kerolasa@iki.fi>
10 years agotunelp: remove get_val() in favour of strtol_or_err()
Sami Kerola [Sun, 22 Feb 2015 14:41:35 +0000 (14:41 +0000)] 
tunelp: remove get_val() in favour of strtol_or_err()

Signed-off-by: Sami Kerola <kerolasa@iki.fi>
10 years agoprlimit: tell in --verbose output which pid got the new limit
Sami Kerola [Sun, 22 Feb 2015 14:41:34 +0000 (14:41 +0000)] 
prlimit: tell in --verbose output which pid got the new limit

Signed-off-by: Sami Kerola <kerolasa@iki.fi>
10 years agolibblkid: cleanup names, don't copy probing results
Karel Zak [Mon, 23 Feb 2015 14:45:41 +0000 (15:45 +0100)] 
libblkid: cleanup names, don't copy probing results

* cleanup function and variable names (s/vals/values/)
* in "safe probe" mode libblkid copies probing result from the first
  attempt, this is unnecessary, with values list we can just move
  values to another list rather than copy all.
* add new debug messages

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agolibblkid: make probing data structures more dynamic
Ondrej Oprala [Tue, 3 Feb 2015 15:30:15 +0000 (16:30 +0100)] 
libblkid: make probing data structures more dynamic

* replace static probing result array with list
* use allocated buffers for probing result variables

[kzak@redhat.com: - rename some functions
                  - clean up \0 terminator usage in variables
                  - remove never used code to convert UUID to lower-case
                  - remove possible memory leaks on errors]

Signed-off-by: Ondrej Oprala <ooprala@redhat.com>
Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agodocs: add terminal-colors requests to TODO
Karel Zak [Thu, 19 Feb 2015 13:31:02 +0000 (14:31 +0100)] 
docs: add terminal-colors requests to TODO

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agotools: add libfdisk docs to kernel.org release generator
Karel Zak [Thu, 19 Feb 2015 12:43:42 +0000 (13:43 +0100)] 
tools: add libfdisk docs to kernel.org release generator

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agobuild-sys: release++ (v2.26) v2.26
Karel Zak [Thu, 19 Feb 2015 12:11:24 +0000 (13:11 +0100)] 
build-sys: release++ (v2.26)

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agodocs: update v2.26-ReleaseNotes
Karel Zak [Thu, 19 Feb 2015 12:10:10 +0000 (13:10 +0100)] 
docs: update v2.26-ReleaseNotes

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agodocs: update AUTHORS file
Karel Zak [Thu, 19 Feb 2015 12:07:55 +0000 (13:07 +0100)] 
docs: update AUTHORS file

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agopo: merge changes
Karel Zak [Thu, 19 Feb 2015 12:06:10 +0000 (13:06 +0100)] 
po: merge changes

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agopo: update vi.po (from translationproject.org)
Trần Ngọc Quân [Thu, 19 Feb 2015 12:02:28 +0000 (13:02 +0100)] 
po: update vi.po (from translationproject.org)

10 years agopo: update uk.po (from translationproject.org)
Yuri Chornoivan [Thu, 19 Feb 2015 12:02:28 +0000 (13:02 +0100)] 
po: update uk.po (from translationproject.org)

10 years agopo: update pl.po (from translationproject.org)
Jakub Bogusz [Thu, 19 Feb 2015 12:02:28 +0000 (13:02 +0100)] 
po: update pl.po (from translationproject.org)

10 years agopo: update nl.po (from translationproject.org)
Benno Schulenberg [Thu, 19 Feb 2015 12:02:28 +0000 (13:02 +0100)] 
po: update nl.po (from translationproject.org)

10 years agopo: update fr.po (from translationproject.org)
David Prévot [Thu, 19 Feb 2015 12:02:28 +0000 (13:02 +0100)] 
po: update fr.po (from translationproject.org)

10 years agopo: update es.po (from translationproject.org)
Antonio Ceballos Roa [Thu, 19 Feb 2015 12:02:28 +0000 (13:02 +0100)] 
po: update es.po (from translationproject.org)

10 years agopo: update de.po (from translationproject.org)
Mario Blättermann [Thu, 19 Feb 2015 12:02:28 +0000 (13:02 +0100)] 
po: update de.po (from translationproject.org)

10 years agopo: update da.po (from translationproject.org)
Joe Hansen [Thu, 19 Feb 2015 12:02:28 +0000 (13:02 +0100)] 
po: update da.po (from translationproject.org)

10 years agopo: update cs.po (from translationproject.org)
Petr Písař [Thu, 19 Feb 2015 12:02:28 +0000 (13:02 +0100)] 
po: update cs.po (from translationproject.org)

10 years agolibfdisk: fix fdisk_partition_get_parent() docs
Karel Zak [Thu, 19 Feb 2015 11:37:23 +0000 (12:37 +0100)] 
libfdisk: fix fdisk_partition_get_parent() docs

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agotests: scsi_debug workaround
Karel Zak [Thu, 19 Feb 2015 11:36:52 +0000 (12:36 +0100)] 
tests: scsi_debug workaround

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agodocs: add cfdisk auto-sort to TODO file
Karel Zak [Thu, 19 Feb 2015 10:56:11 +0000 (11:56 +0100)] 
docs: add cfdisk auto-sort to TODO file

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agolibblkid: set SBMAGIC for nilfs2
Karel Zak [Thu, 19 Feb 2015 10:33:39 +0000 (11:33 +0100)] 
libblkid: set SBMAGIC for nilfs2

Reported-by: Mike Fleetwood <mike.fleetwood@googlemail.com>
Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agosetarch: print error when command runs without arguments
Sami Kerola [Tue, 17 Feb 2015 22:31:48 +0000 (22:31 +0000)] 
setarch: print error when command runs without arguments

$ ./setarch
Segmentation fault (core dumped)
$ coredumpctl gdb
...
Program terminated with signal SIGSEGV, Segmentation fault.
301                     if (*argv[1] != '-') {

Reference: 5edb0ea6bbd57dd916417737f98c9109dc1ecb5b
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
10 years agotests: improve some wordings of the output
Benno Schulenberg [Tue, 17 Feb 2015 20:45:46 +0000 (21:45 +0100)] 
tests: improve some wordings of the output

For one, replace the puzzlingly stern imperative "IGNORE" with
the simple and light-hearted and descriptive "SKIPPED".

Signed-off-by: Benno Schulenberg <bensberg@justemail.net>
10 years agoMerge branch 'hwman2' of https://github.com/jwpi/util-linux
Karel Zak [Thu, 19 Feb 2015 09:54:12 +0000 (10:54 +0100)] 
Merge branch 'hwman2' of https://github.com/jwpi/util-linux

* 'hwman2' of https://github.com/jwpi/util-linux:
  hwclock: improve man-page language

10 years agohwclock: improve man-page language
J William Piggott [Mon, 16 Feb 2015 12:55:31 +0000 (07:55 -0500)] 
hwclock: improve man-page language

Signed-off-by: J William Piggott <elseifthen@gmx.com>
10 years agolibmount: remove assert(arg) from public functions
Karel Zak [Mon, 16 Feb 2015 13:17:54 +0000 (14:17 +0100)] 
libmount: remove assert(arg) from public functions

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agolibsmartcols: remove assert(arg) from public functions
Karel Zak [Mon, 16 Feb 2015 12:49:29 +0000 (13:49 +0100)] 
libsmartcols: remove assert(arg) from public functions

Reported-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agohwclock: improve man-page language
J William Piggott [Sun, 15 Feb 2015 20:39:52 +0000 (15:39 -0500)] 
hwclock: improve man-page language

Signed-off-by: J William Piggott <elseifthen@gmx.com>
10 years agolibmount: improve remount logic
Karel Zak [Mon, 16 Feb 2015 12:38:07 +0000 (13:38 +0100)] 
libmount: improve remount logic

 * fix fs type detection for remount
 * parse mtab only for remount and umount

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agolibmount: add --enable-libmount-force-mountinfo
Karel Zak [Mon, 16 Feb 2015 11:49:49 +0000 (12:49 +0100)] 
libmount: add --enable-libmount-force-mountinfo

The default libmount mtab management depends on mtan symlink. If the
symlink exists than libmount parses /proc/self/mountinfo, otherwise it
parses regular classic /etc/mtab. This is backwardly compatible and
transparent solution.

Unfortunately, this is not robust enough because some broken init
scripts or 3-party mount helpers may remove the symlink and create
regular mtab file. This is pretty bad if initd (systemd) depends on
libmount.

Fortunately we known that mtab is absolutely unwanted on some distros,
so it's fine too ignore mtab at all and don't care about the symlink.

Reported-by: Martin Pitt <martin.pitt@ubuntu.com>
Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agohwclock: man-page errata
J William Piggott [Sun, 15 Feb 2015 18:34:04 +0000 (13:34 -0500)] 
hwclock: man-page errata

Professor Eggert, maintainer of the tz database, reviewed
the new POSIX vs 'RIGHT' man-page section and suggested
the following corrections.

Correct the subdirectory names and remove IANA from the
name of the Time Zone Database.

Signed-off-by: J William Piggott <elseifthen@gmx.com>
10 years agologger: add --socket-errors compatibility option
Sami Kerola [Sun, 15 Feb 2015 09:50:23 +0000 (09:50 +0000)] 
logger: add --socket-errors compatibility option

Hello,

Depending viewpoint this change is either regression fix, or
re-regression in context of none-systemd init.  I ack the change is sent
very late to be part of v2.26, but then again the excess noise was found
only because of -rc1 was tested in sysvinit environment.  IMHO it would
contradict purpose of having rc's if faults will not lead to fixes.

I also want to point out the sysvinit scripts are broken, not the
logger(1), but getting them corrected is practically impossible.
Assuming sysvinit script are further developed by various teams and
distributions who maintain them they should use --socket-error=on in
future, and write scripts that pass without noise.  Meanwhile trying to
be clever when to silence errors seems like a reasonable thing to do.

--->8----
From: Sami Kerola <kerolasa@iki.fi>
Date: Sat, 14 Feb 2015 19:05:55 +0000
Subject: [PATCH] logger: add --socket-errors compatibility option

Now when logger(1) has stopped using openlog() for Unix sockets, in
commit mentioned in reference, the lack of /dev/log detected will report
error accordingly.  According to Gabriele Balducci this makes sysvinit
style boot scripts to print a lot of errors.  So make the logger to
detect whether it should be in compatibility mode, and not report errors
if logging device is missing.  That imitates behavior of glibc openlog().

To allow full control to users the /dev/log error messages can be forced
to on or off.  The automatic error messaging is explained in manual page.

Reference: 1d57503378bdcd838365d625f6d2d0a09da9c29d
Reported-by: Gabriele Balducci <balducci@units.it>
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
10 years agosfdisk: include rpmatch.h
Sami Kerola [Sun, 15 Feb 2015 00:52:49 +0000 (00:52 +0000)] 
sfdisk: include rpmatch.h

Some libc's do not have rpmatch(), and they need the workaround in the
header file.

Signed-off-by: Sami Kerola <kerolasa@iki.fi>
10 years agofsck: fix compiler warning [-Wlogical-not-parentheses]
Karel Zak [Mon, 16 Feb 2015 10:02:53 +0000 (11:02 +0100)] 
fsck: fix compiler warning [-Wlogical-not-parentheses]

Reported-by: Ruediger Meier <sweet_f_a@gmx.de>
Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agoscript.1: Improve documentation of non-interactive behavior
Stanislav Brabec [Fri, 13 Feb 2015 18:59:22 +0000 (19:59 +0100)] 
script.1: Improve documentation of non-interactive behavior

- use exit in .profile to avoid non-logged session after typing "exit"
- document that script can read more from stdin than the command inside:

echo -e 'script -c "read a b ; echo a=\$a b=\$b >defs.sh"\n1 2\necho Done' | sh
Script started, file is typescript
1 2
echo Done
Script done, file is typescript

10 years agoscript: no logging for non-interactive shells
Stanislav Brabec [Wed, 11 Feb 2015 14:07:52 +0000 (15:07 +0100)] 
script: no logging for non-interactive shells

Karel Zak wrote:

> Would be better to learn people to fix their shell initialization
> files to not call script(1) for non-interactive sessions?

It makes sense.

From 8fff32e65399ae07f70e12568d4d5278725673a7 Mon Sep 17 00:00:00 2001
From: Stanislav Brabec <sbrabec@suse.cz>
Date: Wed, 11 Feb 2015 15:02:10 +0100
Subject: [PATCH] script.1: Document behavior in non-interactive shells

And discourage users from such use.

10 years agolibmount: read utab always when read mtab from /proc
Karel Zak [Thu, 12 Feb 2015 12:14:52 +0000 (13:14 +0100)] 
libmount: read utab always when read mtab from /proc

Now libmount reads utab only when mtab filename is no explicitly
specified, but for example:

 mnt_table_parse_mtab(tb, "/proc/self/mountinfo");

ignores utab because filename points to regular file. This is mistake,
we wnat to read utab always when we read mount table from kernel.

Reported-by: Martin Pitt <martin.pitt@ubuntu.com>
Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agofindmnt: don't parse mountinfo twice
Karel Zak [Thu, 12 Feb 2015 11:57:03 +0000 (12:57 +0100)] 
findmnt: don't parse mountinfo twice

We parse /proc/self/mountinfo to initialize cache targets, this step
is unnecessary if --mtab reads all from kernel.

Signed-off-by: Karel Zak <kzak@redhat.com>
10 years agolscpu: add s-Par support
Benjamin Romer [Tue, 10 Feb 2015 21:04:09 +0000 (16:04 -0500)] 
lscpu: add s-Par support

Adds support for the s-Par firmware's hypervisor leaf.

Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
10 years agosulogin, hwclock: use xusleep() instead of usleep()
Sami Kerola [Sun, 8 Feb 2015 20:44:41 +0000 (20:44 +0000)] 
sulogin, hwclock: use xusleep() instead of usleep()

As said in include/c.h the usleep() is marked as obsolete, so do the same
that most of the other util-linux calls do with this interface.

Signed-off-by: Sami Kerola <kerolasa@iki.fi>
10 years agorename: use strrchr() instead of rindex()
Sami Kerola [Sun, 8 Feb 2015 20:44:40 +0000 (20:44 +0000)] 
rename: use strrchr() instead of rindex()

The rindex() is marked legacy in POSIX.1-2001, and apparently Androids
bionic libc does not even have it so it is best not to use the legacy
interface.

Reference: https://lists.gnu.org/archive/html/weechat-dev/2014-02/msg00004.html
Signed-off-by: Sami Kerola <kerolasa@iki.fi>