Linux-2.6.12-rc2v2.6.12-rc2

Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.

Let it rip!

1 files changed, 41 insertions, 0 deletions

diff --git a/Documentation/time_interpolators.txt b/Documentation/time_interpolators.txt
new file mode 100644
index 000000000000..e3b60854fbc2
--- /dev/null
+++ b/Documentation/time_interpolators.txt
@@ -0,0 +1,41 @@
+Time Interpolators
+------------------
+
+Time interpolators are a base of time calculation between timer ticks and
+allow an accurate determination of time down to the accuracy of the time
+source in nanoseconds.
+
+The architecture specific code typically provides gettimeofday and
+settimeofday under Linux. The time interpolator provides both if an arch
+defines CONFIG_TIME_INTERPOLATION. The arch still must set up timer tick
+operations and call the necessary functions to advance the clock.
+
+With the time interpolator a standardized interface exists for time
+interpolation between ticks. The provided logic is highly scalable
+and has been tested in SMP situations of up to 512 CPUs.
+
+If CONFIG_TIME_INTERPOLATION is defined then the architecture specific code
+(or the device drivers - like HPET) may register time interpolators.
+These are typically defined in the following way:
+
+static struct time_interpolator my_interpolator {
+	.frequency = MY_FREQUENCY,
+	.source = TIME_SOURCE_MMIO32,
+	.shift = 8,		/* scaling for higher accuracy */
+	.drift = -1,		/* Unknown drift */
+	.jitter = 0		/* time source is stable */
+};
+
+void time_init(void)
+{
+	....
+	/* Initialization of the timer *.
+	my_interpolator.address = &my_timer;
+	register_time_interpolator(&my_interpolator);
+	....
+}
+
+For more details see include/linux/timex.h and kernel/timer.c.
+
+Christoph Lameter <christoph@lameter.com>, October 31, 2004
+