1ADJTIMEX(8) System Manager's Manual ADJTIMEX(8)
2
6 adjtimex - display or set the kernel time variables
7
9 adjtimex [option]...
10
12 This program gives you raw access to the kernel time variables. Anyone
13 may print out the time variables, but only the superuser may change
14 them.
15 Your computer has two clocks - the "hardware clock" that runs all the
17 time, and the system clock that runs only while the computer is on.
18 Normally, "hwclock --hctosys" should be run at startup to initialize
19 the system clock. The system clock has much better precision (approxi‐
20 mately 1 usec), but the hardware clock probably has better long-term
21 stability. There are three basic strategies for managing these clocks.
22 For a machine connected to the Internet, or equipped with a precision
24 oscillator or radio clock, the best way is to regulate the system clock
25 with ntpd(8). The kernel will automatically update the hardware clock
26 every eleven minutes.
27 In addition, hwclock(8) can be used to approximately correct for a con‐
29 stant drift in the hardware clock. In this case, "hwclock --adjust" is
30 run occasionally. hwclock notes how long it has been since the last
31 adjustment, and nudges the hardware clock forward or back by the appro‐
32 priate amount. The user needs to set the time with "hwclock --set"
33 several times over the course of a few days so hwclock can estimate the
34 drift rate. During that time, ntpd should not be running, or else
35 hwclock will conclude the hardware clock does not drift at all. After
36 you have run "hwclock --set" for the last time, it's okay to start
37 ntpd. Then, "hwclock --systohc" should be run when the machine is shut
38 down. (To see why, suppose the machine runs for a week with ntpd, is
39 shut down for a day, is restarted, and "hwclock --adjust" is run by a
40 startup script. It should only correct for one day's worth of drift.
41 However, it has no way of knowing that ntpd has been adjusting the
42 hardware clock, so it bases its adjustment on the last time hwclock was
43 run.)
44 For a standalone or intermittently connected machine, where it's not
46 possible to run ntpd, you may use adjtimex instead to correct the sys‐
47 tem clock for systematic drift.
48 There are several ways to estimate the drift rate. If your computer
50 can be connected to the net, you might run ntpd for at least several
51 hours and run "adjtimex --print" to learn what values of tick and freq
52 it settled on. Alternately, you could estimate values using as a ref‐
53 erence the CMOS clock (see the --compare and --adjust switches),
54 another host (see --host and --review), or some other source of time
55 (see --watch and --review). You could then add a line to rc.local
56 invoking adjtimex, or configure /etc/init.d/adjtimex or
57 /etc/default/adjtimex, to set those parameters each time you reboot.
58
60 Options may be introduced by either - or --, and unique abbreviations
61 may be used.
62 Here is a summary of the options, grouped by type. Explanations fol‐
64 low.
65 Get/Set Kernel Time Parameters
67 -p --print -t --tick val -f newfreq --frequency newfreq -o val
68 --offset val -s adjustment --singleshot adjustment -S status
69 --status status -m val -R --reset --maxerror val -e val
70 --esterror val -T val --timeconstant val -a[count]
71 --adjust[=count]
72 Estimate Systematic Drifts
74 -c[count] --compare[=count] -i tim --interval tim -l file --log
75 file -h timeserver --host timeserver -w --watch -r[file]
76 --review[=file] -u --utc -d --directisa -n --nointerrupt
77 Informative Output
79 --help -v --version -V --verbose
80 -p, --print
82 Print the current values of the kernel time variables. NOTE:
83 The time is "raw", and may be off by up to one timer tick (10
84 msec). "status" gives the value of the time_status variable in
85 the kernel. For Linux 1.0 and 1.2 kernels, the value is as
86 follows:
87 0 clock is synchronized (so the kernel should
88 periodically set the CMOS clock to match the
89 system clock)
90 1 inserting a leap second at midnight
91 2 deleting a leap second at midnight
92 3 leap second in progress
93 4 leap second has occurred
94 5 clock not externally synchronized (so the
95 kernel should leave the CMOS clock alone)
96 For Linux kernels 2.0 through 2.6, the value is a sum of these:
97 1 PLL updates enabled
98 2 PPS freq discipline enabled
99 4 PPS time discipline enabled
100 8 frequency-lock mode enabled
101 16 inserting leap second
102 32 deleting leap second
103 64 clock unsynchronized
104 128 holding frequency
105 256 PPS signal present
106 512 PPS signal jitter exceeded
107 1024 PPS signal wander exceeded
108 2048 PPS signal calibration error
109 4096 clock hardware fault
110 -t val, --tick val
112 Set the number of microseconds that should be added to the
113 system time for each kernel tick interrupt. For a kernel with
114 USER_HZ=100, there are supposed to be 100 ticks per second, so
115 val should be close to 10000. Increasing val by 1 speeds up the
116 system clock by about 100 ppm, or 8.64 sec/day. tick must be in
117 the range 900000/USER_HZ...1100000/USER_HZ. If val is rejected
118 by the kernel, adjtimex will determine the acceptable range
119 through trial and error and print it. (After completing the
120 search, it will restore the original value.)
121 -f newfreq, --frequency newfreq
123 Set the system clock frequency offset to newfreq. newfreq can
124 be negative or positive, and gives a much finer adjustment than
125 the --tick switch. When USER_HZ=100, the value is scaled such
126 that newfreq = 65536 speeds up the system clock by about 1 ppm,
127 or .0864 sec/day. Thus, all of these are about the same:
128 --tick 9995 --frequency 32768000
129 --tick 10000 --frequency 6553600
130 --tick 10001 --frequency 0
131 --tick 10002 --frequency -6553600
132 --tick 10005 --frequency -32768000
133 To see the acceptable range for newfreq, use --print and look at
134 "tolerance", or try an illegal value (e.g. --tick 0).
135 -s adj, --singleshot adj
137 Slew the system clock by adj usec. (Its rate is changed
138 temporarily by about 1 part in 2000.)
139 -o adj, --offset adj
141 Add a time offset of adj usec. The kernel code adjusts the time
142 gradually by adj, notes how long it has been since the last time
143 offset, and then adjusts the frequency offset to correct for the
144 apparent drift. adj must be in the range -512000...512000.
145 -S status, --status status
147 Set kernel system clock status register to value status. Look
148 here above at the --print switch section for the meaning of
149 status, depending on your kernel.
150 -R, --reset
152 Reset clock status after setting a clock parameter. For early
153 Linux kernels, using the adjtimex(2) system call to set any time
154 parameter the kernel think the clock is synchronized with an
155 external time source, so it sets the kernel variable time_status
156 to TIME_OK. Thereafter, at 11 minute intervals, it will adjust
157 the CMOS clock to match. We prevent this "eleven minute mode"
158 by setting the clock, because that has the side effect of
159 resetting time_status to TIME_BAD. We try not to actually
160 change the clock setting. Kernel versions 2.0.40 and later
161 apparently don't need this. If your kernel does require it, use
162 this option with: -t -T -t -e -m -f -s -o -c -r.
163 -m val, --maxerror val
165 Set maximum error (usec).
166 -e val, --esterror val
168 Set estimated error (usec). The maximum and estimated error are
169 not used by the kernel. They are merely made available to user
170 processes via the adjtimex(2) system call.
171 -T val, --timeconstant val
173 Set phase locked loop (PLL) time constant. val determines the
174 bandwidth or "stiffness" of the PLL. The effective PLL time
175 constant will be a multiple of (2^val). For room-temperature
176 quartz oscillators, David Mills recommends the value 2, which
177 corresponds to a PLL time constant of about 900 sec and a
178 maximum update interval of about 64 sec. The maximum update
179 interval scales directly with the time constant, so that at the
180 maximum time constant of 6, the update interval can be as large
181 as 1024 sec.
182 Values of val between zero and 2 give quick convergence; values
184 between 2 and 6 can be used to reduce network load, but at a
185 modest cost in accuracy.
186 -c[count], --compare[=count]
188 Periodically compare the system clock with the CMOS clock.
189 After the first two calls, print values for tick and frequency
190 offset that would bring the system clock into approximate
191 agreement with the CMOS clock. CMOS clock readings are adjusted
192 for systematic drift using using the correction in /etc/adjtime
193 — see hwclock(8). The interval between comparisons is 10
194 seconds, unless changed by the --interval switch. The optional
195 argument is the number of comparisons. (If the argument is
196 supplied, the "=" is required.) If the CMOS clock and the
197 system clock differ by more than six minutes, adjtimex will try
198 shifting the time from the CMOS clock by some multiple of one
199 hour, up to plus or minus 13 hours in all. This should allow
200 correct operation, including logging, if the --utc switch was
201 used when the CMOS clock is set to local time (or vice-versa),
202 or if summer time has started or stopped since the CMOS clock
203 was last set.
204 -a[count], --adjust[=count]
206 By itself, same as --compare, except the recommended values are
207 actually installed after every third comparison. With --review,
208 the tick and frequency are set to the least-squares estimates.
209 (In the latter case, any count value is ignored.)
210 --force-adjust
212 Override the sanity check that prevents changing the clock rate
213 by more than 500 ppm.
214 -i tim, --interval tim
216 Set the interval in seconds between clock comparisons for the
217 --compare and --adjust options.
218 -u, --utc
220 The CMOS clock is set to UTC (universal time) rather than local
221 time.
222 -d, --directisa
224 To read the CMOS clock accurately, adjtimex usually accesses the
225 clock via the /dev/rtc device driver of the kernel, and makes
226 use of its CMOS update-ended interrupt to detect the beginning
227 of seconds. It will also try /dev/rtc0 (for udev), /dev/misc/rtc
228 (for the obsolete devfs) and possibly others. When the /dev/rtc
229 driver is absent, or when the interrupt is not available,
230 adjtimex can sometimes automatically fallback to a direct access
231 method. This method detects the start of seconds by polling the
232 update-in-progress (UIP) flag of the CMOS clock. You can force
233 this direct access to the CMOS chip with the --directisa switch.
234 Note that the /dev/rtc interrupt method is more accurate, less
236 sensible to perturbations due to system load, cleaner, cheaper,
237 and is generally better than the direct access method. It is
238 advisable to not use the --directisa switch, unless the CMOS
239 chip or the motherboard don't properly provide the necessary
240 interrupt.
241 -n, --nointerrupt
243 Force immediate use of busywait access method, without first
244 waiting for the interrupt timeout.
245 -l[file], --log[=file]
247 Save the current values of the system and CMOS clocks, and
248 optionally a reference time, to file (default
249 /var/log/clocks.log). The reference time is taken from a
250 network timeserver (see the --host switch) or supplied by the
251 user (see the --watch switch).
252 -h timeserver, --host timeserver
254 Use ntpdate to query the given timeserver for the current time.
255 This will fail if timeserver is not running a Network Time
256 Protocol (NTP) server, or if that server is not synchronized.
257 Implies --log.
258 -w, --watch
260 Ask for a keypress when the user knows the time, then ask what
261 that time was, and its approximate accuracy. Implies --log.
262 -r[file], --review[=file]
264 Review the clock log file (default /var/log/clocks.log) and
265 estimate, if possible, the rates of the CMOS and system clocks.
266 Calculate least-squares rates using all suitable log entries.
267 Suggest corrections to adjust for systematic drift. With
268 --adjust, the frequency and tick are set to the suggested
269 values. (The CMOS clock correction is not changed.)
270 -V, --verbose
272 Increase verbosity.
273 --help Print the program options.
275 -v, --version
277 Print the program version.
278
280 If your system clock gained 8 seconds in 24 hours, you could set the
281 tick to 9999, and then it would lose 0.64 seconds a day (that is, 1
282 tick unit = 8.64 seconds per day). To correct the rest of the error,
283 you could set the frequency offset to (2^16)*0.64/.0864 = 485452.
284 Thus, putting the following in rc.local would approximately correct the
285 system clock:
286 adjtimex --tick 9999 --frequency 485452
288
290 adjtimex adjusts only the system clock — the one that runs while the
291 computer is powered up. To set or regulate the CMOS clock, see
292 hwclock(8).
293
295 Steven S. Dick <ssd at nevets.oau.org>, Jim Van Zandt <jrv at
296 comcast.net>.
297
299 date(1L), gettimeofday(2), settimeofday(2), hwclock(8), ntpdate(8),
300 ntpd(8), /usr/src/linux/include/linux/timex.h,
301 /usr/src/linux/include/linux/sched.h, /usr/src/linux/kernel/time.c,
302 /usr/src/linux/kernel/sched.c
303 March 11, 2009 ADJTIMEX(8)