Modify the HCL_ProcessReadings() function to try to always provide
a valid sample. Instead of dropping a sample outside of the expected
delay, provide its assumed quality level as a small integer (relative to
already accumulated samples), and let the caller decide what quality is
acceptable.
Add a new parameter to limit the negative value of the step state
variable. It's set as a maximum delay in number of updates before the
actual step applied to the quantile estimate starts growing from the
minimum step when the input value is consistently larger or smaller than
the estimate.
This prevents the algorithm from effectively becoming the slower 1U
variant if the quantile estimate is stable most of the time.
Set it to 100 updates for the NTP delay and 1000 updates for the hwclock
delay. An option could be added later to make it configurable.
Extend the interval of accepted delays by half of the quantile minimum
step in both directions to make room for floating-point errors in the
quantile calculation and an error that will be intentionally added in
the next commit.
Estimate the 1st and 2nd 10-quantile of the reading delay and accept
only readings between them unless the error of the offset predicted from
previous samples is larger than the minimum reading error. With the 25
PHC readings per ioctl it should combine about 2-3 readings.
This should improve hwclock tracking and synchronization stability when
a PHC reading delay occasionally falls below the normal expected
minimum, or all readings in the batch are delayed significantly (e.g.
due to high PCIe load).
Move processing of PHC readings from sys_linux to hwclock, where
statistics can be collected and filtering improved.
In the PHC refclock driver accumulate the samples even if not in the
external timestamping mode to update the context which will be needed
for improved filtering.
On some HW it seems it's possible to get an occasional bad reading of
the PHC (with normal delay), or in a worse case the clock can step due
to a HW/driver bug, which triggers reset of the HW clock instance. To
avoid having a bad estimate of the frequency when the next (good) sample
is accumulated, drop also the last sample which triggered the reset.
Fix a sign error in conversion of HW time to local time, which caused
the jitter to be amplified instead of reduced. NTP with HW timestamping
should now be more stable and able to ignore occasionally delayed
readings of PHC.
It was never used for anything and messages in debug output already
include filenames, which can be easily grepped if there is a need
to see log messages only from a particular file.
For now, when converting a raw timestamp, return error of the last
sample as the maximum error of the timestamp. This is needed to include
the PHC reading delay in the NTP dispersion.
