svn commit: r216385 - head/sys/i386/xen
Colin Percival
cperciva at FreeBSD.org
Sat Dec 11 22:33:34 UTC 2010
Author: cperciva
Date: Sat Dec 11 22:33:33 2010
New Revision: 216385
URL: http://svn.freebsd.org/changeset/base/216385
Log:
Reduce the Xen timecounter from 1GHz to 2^-9 GHz, thereby increasing the
timecounter period from 2^32 ns (~4.3s) to 2^41 ns (~36m39s). Some time
sharing systems can skip clock interrupts for a few seconds when under
load (e.g., if we've recently used more than our fair share of CPU and
someone else wants a burst of CPU) and we were losing time in quanta of
2^32 ns due to timecounter wrapping.
Increasing the timecounter period up to 2^41 ns is definitely overkill,
but we still have microsecond timecounter precision, and anyone using
paravirtualized hardware when they need submicrosecond timing is crazy.
Modified:
head/sys/i386/xen/clock.c
Modified: head/sys/i386/xen/clock.c
==============================================================================
--- head/sys/i386/xen/clock.c Sat Dec 11 22:13:29 2010 (r216384)
+++ head/sys/i386/xen/clock.c Sat Dec 11 22:33:33 2010 (r216385)
@@ -523,7 +523,8 @@ startrtclock()
set_cyc2ns_scale(cpu_khz/1000);
tsc_freq = cpu_khz * 1000;
- timer_freq = xen_timecounter.tc_frequency = 1000000000LL;
+ timer_freq = 1000000000LL;
+ xen_timecounter.tc_frequency = timer_freq >> 9;
tc_init(&xen_timecounter);
rdtscll(alarm);
@@ -830,7 +831,7 @@ xen_get_timecount(struct timecounter *tc
clk = shadow->system_timestamp + get_nsec_offset(shadow);
- return (uint32_t)(clk);
+ return (uint32_t)(clk >> 9);
}
More information about the svn-src-head
mailing list