2        Real Time Clock (RTC) Drivers for Linux
   3        =======================================
   5When Linux developers talk about a "Real Time Clock", they usually mean
   6something that tracks wall clock time and is battery backed so that it
   7works even with system power off.  Such clocks will normally not track
   8the local time zone or daylight savings time -- unless they dual boot
   9with MS-Windows -- but will instead be set to Coordinated Universal Time
  10(UTC, formerly "Greenwich Mean Time").
  12The newest non-PC hardware tends to just count seconds, like the time(2)
  13system call reports, but RTCs also very commonly represent time using
  14the Gregorian calendar and 24 hour time, as reported by gmtime(3).
  16Linux has two largely-compatible userspace RTC API families you may
  17need to know about:
  19    *   /dev/rtc ... is the RTC provided by PC compatible systems,
  20        so it's not very portable to non-x86 systems.
  22    *   /dev/rtc0, /dev/rtc1 ... are part of a framework that's
  23        supported by a wide variety of RTC chips on all systems.
  25Programmers need to understand that the PC/AT functionality is not
  26always available, and some systems can do much more.  That is, the
  27RTCs use the same API to make requests in both RTC frameworks (using
  28different filenames of course), but the hardware may not offer the
  29same functionality.  For example, not every RTC is hooked up to an
  30IRQ, so they can't all issue alarms; and where standard PC RTCs can
  31only issue an alarm up to 24 hours in the future, other hardware may
  32be able to schedule one any time in the upcoming century.
  35        Old PC/AT-Compatible driver:  /dev/rtc
  36        --------------------------------------
  38All PCs (even Alpha machines) have a Real Time Clock built into them.
  39Usually they are built into the chipset of the computer, but some may
  40actually have a Motorola MC146818 (or clone) on the board. This is the
  41clock that keeps the date and time while your computer is turned off.
  43ACPI has standardized that MC146818 functionality, and extended it in
  44a few ways (enabling longer alarm periods, and wake-from-hibernate).
  45That functionality is NOT exposed in the old driver.
  47However it can also be used to generate signals from a slow 2Hz to a
  48relatively fast 8192Hz, in increments of powers of two. These signals
  49are reported by interrupt number 8. (Oh! So *that* is what IRQ 8 is
  50for...) It can also function as a 24hr alarm, raising IRQ 8 when the
  51alarm goes off. The alarm can also be programmed to only check any
  52subset of the three programmable values, meaning that it could be set to
  53ring on the 30th second of the 30th minute of every hour, for example.
  54The clock can also be set to generate an interrupt upon every clock
  55update, thus generating a 1Hz signal.
  57The interrupts are reported via /dev/rtc (major 10, minor 135, read only
  58character device) in the form of an unsigned long. The low byte contains
  59the type of interrupt (update-done, alarm-rang, or periodic) that was
  60raised, and the remaining bytes contain the number of interrupts since
  61the last read.  Status information is reported through the pseudo-file
  62/proc/driver/rtc if the /proc filesystem was enabled.  The driver has
  63built in locking so that only one process is allowed to have the /dev/rtc
  64interface open at a time.
  66A user process can monitor these interrupts by doing a read(2) or a
  67select(2) on /dev/rtc -- either will block/stop the user process until
  68the next interrupt is received. This is useful for things like
  69reasonably high frequency data acquisition where one doesn't want to
  70burn up 100% CPU by polling gettimeofday etc. etc.
  72At high frequencies, or under high loads, the user process should check
  73the number of interrupts received since the last read to determine if
  74there has been any interrupt "pileup" so to speak. Just for reference, a
  75typical 486-33 running a tight read loop on /dev/rtc will start to suffer
  76occasional interrupt pileup (i.e. > 1 IRQ event since last read) for
  77frequencies above 1024Hz. So you really should check the high bytes
  78of the value you read, especially at frequencies above that of the
  79normal timer interrupt, which is 100Hz.
  81Programming and/or enabling interrupt frequencies greater than 64Hz is
  82only allowed by root. This is perhaps a bit conservative, but we don't want
  83an evil user generating lots of IRQs on a slow 386sx-16, where it might have
  84a negative impact on performance. This 64Hz limit can be changed by writing
  85a different value to /proc/sys/dev/rtc/max-user-freq. Note that the
  86interrupt handler is only a few lines of code to minimize any possibility
  87of this effect.
  89Also, if the kernel time is synchronized with an external source, the 
  90kernel will write the time back to the CMOS clock every 11 minutes. In 
  91the process of doing this, the kernel briefly turns off RTC periodic 
  92interrupts, so be aware of this if you are doing serious work. If you
  93don't synchronize the kernel time with an external source (via ntp or
  94whatever) then the kernel will keep its hands off the RTC, allowing you
  95exclusive access to the device for your applications.
  97The alarm and/or interrupt frequency are programmed into the RTC via
  98various ioctl(2) calls as listed in ./include/linux/rtc.h
  99Rather than write 50 pages describing the ioctl() and so on, it is
 100perhaps more useful to include a small test program that demonstrates
 101how to use them, and demonstrates the features of the driver. This is
 102probably a lot more useful to people interested in writing applications
 103that will be using this driver.  See the code at the end of this document.
 105(The original /dev/rtc driver was written by Paul Gortmaker.)
 108        New portable "RTC Class" drivers:  /dev/rtcN
 109        --------------------------------------------
 111Because Linux supports many non-ACPI and non-PC platforms, some of which
 112have more than one RTC style clock, it needed a more portable solution
 113than expecting a single battery-backed MC146818 clone on every system.
 114Accordingly, a new "RTC Class" framework has been defined.  It offers
 115three different userspace interfaces:
 117    *   /dev/rtcN ... much the same as the older /dev/rtc interface
 119    *   /sys/class/rtc/rtcN ... sysfs attributes support readonly
 120        access to some RTC attributes.
 122    *   /proc/driver/rtc ... the system clock RTC may expose itself
 123        using a procfs interface. If there is no RTC for the system clock,
 124        rtc0 is used by default. More information is (currently) shown
 125        here than through sysfs.
 127The RTC Class framework supports a wide variety of RTCs, ranging from those
 128integrated into embeddable system-on-chip (SOC) processors to discrete chips
 129using I2C, SPI, or some other bus to communicate with the host CPU.  There's
 130even support for PC-style RTCs ... including the features exposed on newer PCs
 131through ACPI.
 133The new framework also removes the "one RTC per system" restriction.  For
 134example, maybe the low-power battery-backed RTC is a discrete I2C chip, but
 135a high functionality RTC is integrated into the SOC.  That system might read
 136the system clock from the discrete RTC, but use the integrated one for all
 137other tasks, because of its greater functionality.
 142The sysfs interface under /sys/class/rtc/rtcN provides access to various
 143rtc attributes without requiring the use of ioctls. All dates and times
 144are in the RTC's timezone, rather than in system time.
 146date:            RTC-provided date
 147hctosys:         1 if the RTC provided the system time at boot via the
 148                 CONFIG_RTC_HCTOSYS kernel option, 0 otherwise
 149max_user_freq:   The maximum interrupt rate an unprivileged user may request
 150                 from this RTC.
 151name:            The name of the RTC corresponding to this sysfs directory
 152since_epoch:     The number of seconds since the epoch according to the RTC
 153time:            RTC-provided time
 154wakealarm:       The time at which the clock will generate a system wakeup
 155                 event. This is a one shot wakeup event, so must be reset
 156                 after wake if a daily wakeup is required. Format is seconds since
 157                 the epoch by default, or if there's a leading +, seconds in the
 158                 future, or if there is a leading +=, seconds ahead of the current
 159                 alarm.
 160offset:          The amount which the rtc clock has been adjusted in firmware.
 161                 Visible only if the driver supports clock offset adjustment.
 162                 The unit is parts per billion, i.e. The number of clock ticks
 163                 which are added to or removed from the rtc's base clock per
 164                 billion ticks. A positive value makes a day pass more slowly,
 165                 longer, and a negative value makes a day pass more quickly.
 170The ioctl() calls supported by /dev/rtc are also supported by the RTC class
 171framework.  However, because the chips and systems are not standardized,
 172some PC/AT functionality might not be provided.  And in the same way, some
 173newer features -- including those enabled by ACPI -- are exposed by the
 174RTC class framework, but can't be supported by the older driver.
 176    *   RTC_RD_TIME, RTC_SET_TIME ... every RTC supports at least reading
 177        time, returning the result as a Gregorian calendar date and 24 hour
 178        wall clock time.  To be most useful, this time may also be updated.
 180    *   RTC_AIE_ON, RTC_AIE_OFF, RTC_ALM_SET, RTC_ALM_READ ... when the RTC
 181        is connected to an IRQ line, it can often issue an alarm IRQ up to
 182        24 hours in the future.  (Use RTC_WKALM_* by preference.)
 184    *   RTC_WKALM_SET, RTC_WKALM_RD ... RTCs that can issue alarms beyond
 185        the next 24 hours use a slightly more powerful API, which supports
 186        setting the longer alarm time and enabling its IRQ using a single
 187        request (using the same model as EFI firmware).
 189    *   RTC_UIE_ON, RTC_UIE_OFF ... if the RTC offers IRQs, the RTC framework
 190        will emulate this mechanism.
 192    *   RTC_PIE_ON, RTC_PIE_OFF, RTC_IRQP_SET, RTC_IRQP_READ ... these icotls
 193        are emulated via a kernel hrtimer.
 195In many cases, the RTC alarm can be a system wake event, used to force
 196Linux out of a low power sleep state (or hibernation) back to a fully
 197operational state.  For example, a system could enter a deep power saving
 198state until it's time to execute some scheduled tasks.
 200Note that many of these ioctls are handled by the common rtc-dev interface.
 201Some common examples:
 203    *   RTC_RD_TIME, RTC_SET_TIME: the read_time/set_time functions will be
 204        called with appropriate values.
 206    *   RTC_ALM_SET, RTC_ALM_READ, RTC_WKALM_SET, RTC_WKALM_RD: gets or sets
 207        the alarm rtc_timer. May call the set_alarm driver function.
 209    *   RTC_IRQP_SET, RTC_IRQP_READ: These are emulated by the generic code.
 211    *   RTC_PIE_ON, RTC_PIE_OFF: These are also emulated by the generic code.
 213If all else fails, check out the tools/testing/selftests/timers/rtctest.c test!