rtc.txt 10.6 KB
Newer Older
1 2 3
=======================================
Real Time Clock (RTC) Drivers for Linux
=======================================
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34

When Linux developers talk about a "Real Time Clock", they usually mean
something that tracks wall clock time and is battery backed so that it
works even with system power off.  Such clocks will normally not track
the local time zone or daylight savings time -- unless they dual boot
with MS-Windows -- but will instead be set to Coordinated Universal Time
(UTC, formerly "Greenwich Mean Time").

The newest non-PC hardware tends to just count seconds, like the time(2)
system call reports, but RTCs also very commonly represent time using
the Gregorian calendar and 24 hour time, as reported by gmtime(3).

Linux has two largely-compatible userspace RTC API families you may
need to know about:

    *	/dev/rtc ... is the RTC provided by PC compatible systems,
	so it's not very portable to non-x86 systems.

    *	/dev/rtc0, /dev/rtc1 ... are part of a framework that's
	supported by a wide variety of RTC chips on all systems.

Programmers need to understand that the PC/AT functionality is not
always available, and some systems can do much more.  That is, the
RTCs use the same API to make requests in both RTC frameworks (using
different filenames of course), but the hardware may not offer the
same functionality.  For example, not every RTC is hooked up to an
IRQ, so they can't all issue alarms; and where standard PC RTCs can
only issue an alarm up to 24 hours in the future, other hardware may
be able to schedule one any time in the upcoming century.


35 36
Old PC/AT-Compatible driver:  /dev/rtc
--------------------------------------
Linus Torvalds's avatar
Linus Torvalds committed
37 38 39 40 41 42

All PCs (even Alpha machines) have a Real Time Clock built into them.
Usually they are built into the chipset of the computer, but some may
actually have a Motorola MC146818 (or clone) on the board. This is the
clock that keeps the date and time while your computer is turned off.

43 44 45 46
ACPI has standardized that MC146818 functionality, and extended it in
a few ways (enabling longer alarm periods, and wake-from-hibernate).
That functionality is NOT exposed in the old driver.

Linus Torvalds's avatar
Linus Torvalds committed
47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83
However it can also be used to generate signals from a slow 2Hz to a
relatively fast 8192Hz, in increments of powers of two. These signals
are reported by interrupt number 8. (Oh! So *that* is what IRQ 8 is
for...) It can also function as a 24hr alarm, raising IRQ 8 when the
alarm goes off. The alarm can also be programmed to only check any
subset of the three programmable values, meaning that it could be set to
ring on the 30th second of the 30th minute of every hour, for example.
The clock can also be set to generate an interrupt upon every clock
update, thus generating a 1Hz signal.

The interrupts are reported via /dev/rtc (major 10, minor 135, read only
character device) in the form of an unsigned long. The low byte contains
the type of interrupt (update-done, alarm-rang, or periodic) that was
raised, and the remaining bytes contain the number of interrupts since
the last read.  Status information is reported through the pseudo-file
/proc/driver/rtc if the /proc filesystem was enabled.  The driver has
built in locking so that only one process is allowed to have the /dev/rtc
interface open at a time.

A user process can monitor these interrupts by doing a read(2) or a
select(2) on /dev/rtc -- either will block/stop the user process until
the next interrupt is received. This is useful for things like
reasonably high frequency data acquisition where one doesn't want to
burn up 100% CPU by polling gettimeofday etc. etc.

At high frequencies, or under high loads, the user process should check
the number of interrupts received since the last read to determine if
there has been any interrupt "pileup" so to speak. Just for reference, a
typical 486-33 running a tight read loop on /dev/rtc will start to suffer
occasional interrupt pileup (i.e. > 1 IRQ event since last read) for
frequencies above 1024Hz. So you really should check the high bytes
of the value you read, especially at frequencies above that of the
normal timer interrupt, which is 100Hz.

Programming and/or enabling interrupt frequencies greater than 64Hz is
only allowed by root. This is perhaps a bit conservative, but we don't want
an evil user generating lots of IRQs on a slow 386sx-16, where it might have
84 85 86 87
a negative impact on performance. This 64Hz limit can be changed by writing
a different value to /proc/sys/dev/rtc/max-user-freq. Note that the
interrupt handler is only a few lines of code to minimize any possibility
of this effect.
Linus Torvalds's avatar
Linus Torvalds committed
88 89 90 91 92 93 94 95 96 97 98 99 100 101 102

Also, if the kernel time is synchronized with an external source, the 
kernel will write the time back to the CMOS clock every 11 minutes. In 
the process of doing this, the kernel briefly turns off RTC periodic 
interrupts, so be aware of this if you are doing serious work. If you
don't synchronize the kernel time with an external source (via ntp or
whatever) then the kernel will keep its hands off the RTC, allowing you
exclusive access to the device for your applications.

The alarm and/or interrupt frequency are programmed into the RTC via
various ioctl(2) calls as listed in ./include/linux/rtc.h
Rather than write 50 pages describing the ioctl() and so on, it is
perhaps more useful to include a small test program that demonstrates
how to use them, and demonstrates the features of the driver. This is
probably a lot more useful to people interested in writing applications
103 104 105 106 107
that will be using this driver.  See the code at the end of this document.

(The original /dev/rtc driver was written by Paul Gortmaker.)


108 109
New portable "RTC Class" drivers:  /dev/rtcN
--------------------------------------------
110 111 112 113 114 115 116 117 118 119 120 121

Because Linux supports many non-ACPI and non-PC platforms, some of which
have more than one RTC style clock, it needed a more portable solution
than expecting a single battery-backed MC146818 clone on every system.
Accordingly, a new "RTC Class" framework has been defined.  It offers
three different userspace interfaces:

    *	/dev/rtcN ... much the same as the older /dev/rtc interface

    *	/sys/class/rtc/rtcN ... sysfs attributes support readonly
	access to some RTC attributes.

122 123 124
    *	/proc/driver/rtc ... the system clock RTC may expose itself
	using a procfs interface. If there is no RTC for the system clock,
	rtc0 is used by default. More information is (currently) shown
125 126 127 128 129 130 131 132 133 134 135 136 137 138
	here than through sysfs.

The RTC Class framework supports a wide variety of RTCs, ranging from those
integrated into embeddable system-on-chip (SOC) processors to discrete chips
using I2C, SPI, or some other bus to communicate with the host CPU.  There's
even support for PC-style RTCs ... including the features exposed on newer PCs
through ACPI.

The new framework also removes the "one RTC per system" restriction.  For
example, maybe the low-power battery-backed RTC is a discrete I2C chip, but
a high functionality RTC is integrated into the SOC.  That system might read
the system clock from the discrete RTC, but use the integrated one for all
other tasks, because of its greater functionality.

139
SYSFS interface
140 141 142 143 144 145
---------------

The sysfs interface under /sys/class/rtc/rtcN provides access to various
rtc attributes without requiring the use of ioctls. All dates and times
are in the RTC's timezone, rather than in system time.

146 147 148
================ ==============================================================
date  	   	 RTC-provided date
hctosys   	 1 if the RTC provided the system time at boot via the
149
		 CONFIG_RTC_HCTOSYS kernel option, 0 otherwise
150
max_user_freq	 The maximum interrupt rate an unprivileged user may request
151
		 from this RTC.
152 153 154 155
name		 The name of the RTC corresponding to this sysfs directory
since_epoch	 The number of seconds since the epoch according to the RTC
time		 RTC-provided time
wakealarm	 The time at which the clock will generate a system wakeup
156
		 event. This is a one shot wakeup event, so must be reset
157 158 159 160 161
		 after wake if a daily wakeup is required. Format is seconds
		 since the epoch by default, or if there's a leading +, seconds
		 in the future, or if there is a leading +=, seconds ahead of
		 the current alarm.
offset		 The amount which the rtc clock has been adjusted in firmware.
162 163 164 165 166
		 Visible only if the driver supports clock offset adjustment.
		 The unit is parts per billion, i.e. The number of clock ticks
		 which are added to or removed from the rtc's base clock per
		 billion ticks. A positive value makes a day pass more slowly,
		 longer, and a negative value makes a day pass more quickly.
167 168
*/nvmem		 The non volatile storage exported as a raw file, as described
		 in Documentation/nvmem/nvmem.txt
169
================ ==============================================================
170

171
IOCTL interface
172 173
---------------

174 175 176 177 178 179 180 181 182 183 184 185
The ioctl() calls supported by /dev/rtc are also supported by the RTC class
framework.  However, because the chips and systems are not standardized,
some PC/AT functionality might not be provided.  And in the same way, some
newer features -- including those enabled by ACPI -- are exposed by the
RTC class framework, but can't be supported by the older driver.

    *	RTC_RD_TIME, RTC_SET_TIME ... every RTC supports at least reading
	time, returning the result as a Gregorian calendar date and 24 hour
	wall clock time.  To be most useful, this time may also be updated.

    *	RTC_AIE_ON, RTC_AIE_OFF, RTC_ALM_SET, RTC_ALM_READ ... when the RTC
	is connected to an IRQ line, it can often issue an alarm IRQ up to
186
	24 hours in the future.  (Use RTC_WKALM_* by preference.)
187

188
    *	RTC_WKALM_SET, RTC_WKALM_RD ... RTCs that can issue alarms beyond
189 190 191 192
	the next 24 hours use a slightly more powerful API, which supports
	setting the longer alarm time and enabling its IRQ using a single
	request (using the same model as EFI firmware).

193 194
    *	RTC_UIE_ON, RTC_UIE_OFF ... if the RTC offers IRQs, the RTC framework
	will emulate this mechanism.
195

196 197
    *	RTC_PIE_ON, RTC_PIE_OFF, RTC_IRQP_SET, RTC_IRQP_READ ... these icotls
	are emulated via a kernel hrtimer.
198 199 200 201 202

In many cases, the RTC alarm can be a system wake event, used to force
Linux out of a low power sleep state (or hibernation) back to a fully
operational state.  For example, a system could enter a deep power saving
state until it's time to execute some scheduled tasks.
Linus Torvalds's avatar
Linus Torvalds committed
203

204 205
Note that many of these ioctls are handled by the common rtc-dev interface.
Some common examples:
206 207 208 209

    *	RTC_RD_TIME, RTC_SET_TIME: the read_time/set_time functions will be
	called with appropriate values.

210 211
    *	RTC_ALM_SET, RTC_ALM_READ, RTC_WKALM_SET, RTC_WKALM_RD: gets or sets
	the alarm rtc_timer. May call the set_alarm driver function.
212

213
    *	RTC_IRQP_SET, RTC_IRQP_READ: These are emulated by the generic code.
214

215
    *	RTC_PIE_ON, RTC_PIE_OFF: These are also emulated by the generic code.
216

217
If all else fails, check out the tools/testing/selftests/timers/rtctest.c test!