[Bug 1313024] Re: rcS(5) UTC option documentation error
Geoffrey Hager
gchager at gmail.com
Sat Apr 26 07:18:40 UTC 2014
** Description changed:
PROBLEM DESCRIPTION:
Quoting from
- http://manpages.ubuntu.com/manpages/trusty/en/man8/hwclock.8.html:
+ http://manpages.ubuntu.com/manpages/trusty/en/man5/rcS.5.html:
"OPTIONS … UTC This is used to govern how the hardware real
time clock is interpreted when it is read (e.g., at boot time, for the
purpose of setting the system clock) and when it is written (e.g., at
shutdown). If this option is set to no then the system clock is assumed
to be set to local time. If the option is set to yes then the system
clock is assumed to be set to something approximating
Coordinated Universal Time (UTC)."
I discovered that the last two sentences describe the opposite of the
actual behavior (which for me did in fact result in a good deal of
confusion & problems & extra time configuring my system to use rtcwake
with my HW clock set to UTC). These two sentences should refer to the
‘hardware real time clock’ instead of the ‘system clock’. The actual
behavior is illustrated below (along with my pertinent system
information).
-
PROPOSED CORRECTION:
These last two sentences should refer to the ‘hardware real time clock’
instead of the ‘system clock’, i.e.:
"If this option is set to no then the hardware real time clock is
assumed to be set to local time. If the option is set to yes then
the hardware real time clock is assumed to be set to Coordinated
Universal Time (UTC)."
-
BACKGROUND:
If “UTC=yes” is set in /etc/default/rcS:
- # assume that the BIOS clock is set to UTC time (recommended)
- UTC=yes
+ # assume that the BIOS clock is set to UTC time (recommended)
+ UTC=yes
… then the script in /etc/init/hwclock.conf calls:
- hwclock --rtc=/dev/rtc0 --systz --utc --noadjfile
+ hwclock --rtc=/dev/rtc0 --systz --utc --noadjfile
… and the script in /etc/init/hwclock-save.conf calls:
- hwclock --rtc=/dev/rtc0 --systohc --utc --noadjfile
+ hwclock --rtc=/dev/rtc0 --systohc --utc --noadjfile
Otherwise (if ! $UTC=yes) then those respective hwclock calls are made
with the ‘--localtime’ option instead of ‘--utc’.
-
SUMMARY of the actual observed behavior:
‘hwclock --systohc --localtime’ writes the system time to the HW clock
without any timezone translation (i.e., assumes that the HW clock is set
to local time).
‘hwclock --systohc --utc’ translates the system time from local to UTC
based on the system timezone configuration and writes the translated
time (UTC) to the HW clock (i.e., assumes that the HW clock is set to
UTC).
‘hwclock --systz --localtime’ writes the HW clock time to the system
time without any timezone translation (i.e., assumes that the HW clock
is set to local time).
‘hwclock --systz --utc’ translates the HW clock time from UTC to local
based on the system timezone configuration and writes the translated
time (local) to the system time (i.e., assumes that the HW clock is set
to UTC).
(Note that hwclock does behave consistently with its manpage.)
-
MY SYSTEM INFORMATION:
# lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 14.04 LTS
Release: 14.04
Codename: trusty
# uname -srvmpio
Linux 3.13.0-24-generic #46-Ubuntu SMP Thu Apr 10 19:11:08 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux
-
HOW TO DEMONSTRATE THE ACTUAL BEHAVIOR:
# date
Sat Apr 26 01:47:22 EDT 2014
# cat /proc/driver/rtc | grep rtc
rtc_time : 05:47:24
rtc_date : 2014-04-26
# hwclock --rtc=/dev/rtc0 --systohc --localtime --noadjfile; cat /proc/driver/rtc | grep rtc
rtc_time : 01:47:37
rtc_date : 2014-04-26
# hwclock --rtc=/dev/rtc0 --systohc --utc --noadjfile; cat /proc/driver/rtc | grep rtc
rtc_time : 05:47:45
rtc_date : 2014-04-26
# cat /proc/driver/rtc | grep rtc; hwclock --rtc=/dev/rtc0 --systz --utc --noadjfile; date
rtc_time : 06:09:10
rtc_date : 2014-04-26
Sat Apr 26 02:09:10 EDT 2014
# cat /proc/driver/rtc | grep rtc; hwclock --rtc=/dev/rtc0 --systz --localtime --noadjfile; date
rtc_time : 06:09:36
rtc_date : 2014-04-26
Sat Apr 26 06:09:36 EDT 2014
** Summary changed:
- rcS(5) UTC option documentation error
+ rcS(5) UTC option manpage documentation error
--
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to sysvinit in Ubuntu.
https://bugs.launchpad.net/bugs/1313024
Title:
rcS(5) UTC option manpage documentation error
Status in “sysvinit” package in Ubuntu:
New
Bug description:
PROBLEM DESCRIPTION:
Quoting from
http://manpages.ubuntu.com/manpages/trusty/en/man5/rcS.5.html:
"OPTIONS … UTC This is used to govern how the hardware real
time clock is interpreted when it is read (e.g., at boot time, for the
purpose of setting the system clock) and when it is written (e.g.,
at shutdown). If this option is set to no then the system clock is
assumed to be set to local time. If the option is set to yes then
the system clock is assumed to be set to something
approximating Coordinated Universal Time (UTC)."
I discovered that the last two sentences describe the opposite of the
actual behavior (which for me did in fact result in a good deal of
confusion & problems & extra time configuring my system to use rtcwake
with my HW clock set to UTC). These two sentences should refer to the
‘hardware real time clock’ instead of the ‘system clock’. The actual
behavior is illustrated below (along with my pertinent system
information).
PROPOSED CORRECTION:
These last two sentences should refer to the ‘hardware real time
clock’ instead of the ‘system clock’, i.e.:
"If this option is set to no then the hardware real time clock is
assumed to be set to local time. If the option is set to yes then
the hardware real time clock is assumed to be set to Coordinated
Universal Time (UTC)."
BACKGROUND:
If “UTC=yes” is set in /etc/default/rcS:
# assume that the BIOS clock is set to UTC time (recommended)
UTC=yes
… then the script in /etc/init/hwclock.conf calls:
hwclock --rtc=/dev/rtc0 --systz --utc --noadjfile
… and the script in /etc/init/hwclock-save.conf calls:
hwclock --rtc=/dev/rtc0 --systohc --utc --noadjfile
Otherwise (if ! $UTC=yes) then those respective hwclock calls are made
with the ‘--localtime’ option instead of ‘--utc’.
SUMMARY of the actual observed behavior:
‘hwclock --systohc --localtime’ writes the system time to the HW clock
without any timezone translation (i.e., assumes that the HW clock is
set to local time).
‘hwclock --systohc --utc’ translates the system time from local to UTC
based on the system timezone configuration and writes the translated
time (UTC) to the HW clock (i.e., assumes that the HW clock is set to
UTC).
‘hwclock --systz --localtime’ writes the HW clock time to the system
time without any timezone translation (i.e., assumes that the HW clock
is set to local time).
‘hwclock --systz --utc’ translates the HW clock time from UTC to local
based on the system timezone configuration and writes the translated
time (local) to the system time (i.e., assumes that the HW clock is
set to UTC).
(Note that hwclock does behave consistently with its manpage.)
MY SYSTEM INFORMATION:
# lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 14.04 LTS
Release: 14.04
Codename: trusty
# uname -srvmpio
Linux 3.13.0-24-generic #46-Ubuntu SMP Thu Apr 10 19:11:08 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux
HOW TO DEMONSTRATE THE ACTUAL BEHAVIOR:
# date
Sat Apr 26 01:47:22 EDT 2014
# cat /proc/driver/rtc | grep rtc
rtc_time : 05:47:24
rtc_date : 2014-04-26
# hwclock --rtc=/dev/rtc0 --systohc --localtime --noadjfile; cat /proc/driver/rtc | grep rtc
rtc_time : 01:47:37
rtc_date : 2014-04-26
# hwclock --rtc=/dev/rtc0 --systohc --utc --noadjfile; cat /proc/driver/rtc | grep rtc
rtc_time : 05:47:45
rtc_date : 2014-04-26
# cat /proc/driver/rtc | grep rtc; hwclock --rtc=/dev/rtc0 --systz --utc --noadjfile; date
rtc_time : 06:09:10
rtc_date : 2014-04-26
Sat Apr 26 02:09:10 EDT 2014
# cat /proc/driver/rtc | grep rtc; hwclock --rtc=/dev/rtc0 --systz --localtime --noadjfile; date
rtc_time : 06:09:36
rtc_date : 2014-04-26
Sat Apr 26 06:09:36 EDT 2014
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/sysvinit/+bug/1313024/+subscriptions
More information about the foundations-bugs
mailing list