141 lines · plain
1=======================================2Real Time Clock (RTC) Drivers for Linux3=======================================4 5When Linux developers talk about a "Real Time Clock", they usually mean6something that tracks wall clock time and is battery backed so that it7works even with system power off. Such clocks will normally not track8the local time zone or daylight savings time -- unless they dual boot9with MS-Windows -- but will instead be set to Coordinated Universal Time10(UTC, formerly "Greenwich Mean Time").11 12The newest non-PC hardware tends to just count seconds, like the time(2)13system call reports, but RTCs also very commonly represent time using14the Gregorian calendar and 24 hour time, as reported by gmtime(3).15 16Linux has two largely-compatible userspace RTC API families you may17need to know about:18 19 * /dev/rtc ... is the RTC provided by PC compatible systems,20 so it's not very portable to non-x86 systems.21 22 * /dev/rtc0, /dev/rtc1 ... are part of a framework that's23 supported by a wide variety of RTC chips on all systems.24 25Programmers need to understand that the PC/AT functionality is not26always available, and some systems can do much more. That is, the27RTCs use the same API to make requests in both RTC frameworks (using28different filenames of course), but the hardware may not offer the29same functionality. For example, not every RTC is hooked up to an30IRQ, so they can't all issue alarms; and where standard PC RTCs can31only issue an alarm up to 24 hours in the future, other hardware may32be able to schedule one any time in the upcoming century.33 34 35Old PC/AT-Compatible driver: /dev/rtc36--------------------------------------37 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 may40actually have a Motorola MC146818 (or clone) on the board. This is the41clock that keeps the date and time while your computer is turned off.42 43ACPI has standardized that MC146818 functionality, and extended it in44a few ways (enabling longer alarm periods, and wake-from-hibernate).45That functionality is NOT exposed in the old driver.46 47However it can also be used to generate signals from a slow 2Hz to a48relatively fast 8192Hz, in increments of powers of two. These signals49are reported by interrupt number 8. (Oh! So *that* is what IRQ 8 is50for...) It can also function as a 24hr alarm, raising IRQ 8 when the51alarm goes off. The alarm can also be programmed to only check any52subset of the three programmable values, meaning that it could be set to53ring 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 clock55update, thus generating a 1Hz signal.56 57The interrupts are reported via /dev/rtc (major 10, minor 135, read only58character device) in the form of an unsigned long. The low byte contains59the type of interrupt (update-done, alarm-rang, or periodic) that was60raised, and the remaining bytes contain the number of interrupts since61the last read. Status information is reported through the pseudo-file62/proc/driver/rtc if the /proc filesystem was enabled. The driver has63built in locking so that only one process is allowed to have the /dev/rtc64interface open at a time.65 66A user process can monitor these interrupts by doing a read(2) or a67select(2) on /dev/rtc -- either will block/stop the user process until68the next interrupt is received. This is useful for things like69reasonably high frequency data acquisition where one doesn't want to70burn up 100% CPU by polling gettimeofday etc. etc.71 72At high frequencies, or under high loads, the user process should check73the number of interrupts received since the last read to determine if74there has been any interrupt "pileup" so to speak. Just for reference, a75typical 486-33 running a tight read loop on /dev/rtc will start to suffer76occasional interrupt pileup (i.e. > 1 IRQ event since last read) for77frequencies above 1024Hz. So you really should check the high bytes78of the value you read, especially at frequencies above that of the79normal timer interrupt, which is 100Hz.80 81Programming and/or enabling interrupt frequencies greater than 64Hz is82only allowed by root. This is perhaps a bit conservative, but we don't want83an evil user generating lots of IRQs on a slow 386sx-16, where it might have84a negative impact on performance. This 64Hz limit can be changed by writing85a different value to /proc/sys/dev/rtc/max-user-freq. Note that the86interrupt handler is only a few lines of code to minimize any possibility87of this effect.88 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 you93don't synchronize the kernel time with an external source (via ntp or94whatever) then the kernel will keep its hands off the RTC, allowing you95exclusive access to the device for your applications.96 97The alarm and/or interrupt frequency are programmed into the RTC via98various ioctl(2) calls as listed in ./include/linux/rtc.h99Rather than write 50 pages describing the ioctl() and so on, it is100perhaps more useful to include a small test program that demonstrates101how to use them, and demonstrates the features of the driver. This is102probably a lot more useful to people interested in writing applications103that will be using this driver. See the code at the end of this document.104 105(The original /dev/rtc driver was written by Paul Gortmaker.)106 107 108New portable "RTC Class" drivers: /dev/rtcN109--------------------------------------------110 111Because Linux supports many non-ACPI and non-PC platforms, some of which112have more than one RTC style clock, it needed a more portable solution113than expecting a single battery-backed MC146818 clone on every system.114Accordingly, a new "RTC Class" framework has been defined. It offers115three different userspace interfaces:116 117 * /dev/rtcN ... much the same as the older /dev/rtc interface118 119 * /sys/class/rtc/rtcN ... sysfs attributes support readonly120 access to some RTC attributes.121 122 * /proc/driver/rtc ... the system clock RTC may expose itself123 using a procfs interface. If there is no RTC for the system clock,124 rtc0 is used by default. More information is (currently) shown125 here than through sysfs.126 127The RTC Class framework supports a wide variety of RTCs, ranging from those128integrated into embeddable system-on-chip (SOC) processors to discrete chips129using I2C, SPI, or some other bus to communicate with the host CPU. There's130even support for PC-style RTCs ... including the features exposed on newer PCs131through ACPI.132 133The new framework also removes the "one RTC per system" restriction. For134example, maybe the low-power battery-backed RTC is a discrete I2C chip, but135a high functionality RTC is integrated into the SOC. That system might read136the system clock from the discrete RTC, but use the integrated one for all137other tasks, because of its greater functionality.138 139Check out tools/testing/selftests/rtc/rtctest.c for an example usage of the140ioctl interface.141