On 12/12/05 03:02, Thomas Gleixner wrote:
> The rebased version of the high resolution patches on top of the
> hrtimers base patch is available from the new project home:

> http://www.tglx.de/projects/hrtimers/2.6.15-rc5/patch-2.6.15-rc5-hrt2.patch

I've tried this on a dual-proc HT Xeon system. On bootup I get the below
"Badness in check_monotonic_clock" message. The machine boots and runs
ok afterwords. Attached is the full dmesg & .config.

This is working very well for me so far; rtc_wakeup works from userspace
& I created a kernel module to start a hrtimer 30k/sec as an experiement.

I also get the occational:
check_periodic_interval: short interval! 49996835.
                bad calibration or timers may be broken.



The "badness in check_monotonic_clock" from dmesg...

<snip>
isapnp: Scanning for PnP cards...
Time: tsc clocksource has been installed.
check_monotonic_clock: monotonic inconsistency detected!
        from         2e5d8586 (777880966) to         2bd16c47 (735145031).
Badness in check_monotonic_clock at kernel/time/timeofday.c:161
 [<c0103dd7>] dump_stack+0x1e/0x22
 [<c0134782>] check_monotonic_clock+0xd4/0xd9
 [<c0134c95>] get_realtime_clock+0xc9/0x10b
 [<c01338b3>] hrtimer_run_queues+0x4b/0x10b
 [<c01258eb>] run_timer_softirq+0x1f/0x1c7
 [<c01215bd>] __do_softirq+0xc1/0xd6
 [<c012161d>] do_softirq+0x4b/0x4d
 [<c0121707>] irq_exit+0x47/0x49
 [<c0110fcd>] smp_apic_timer_interrupt+0x42/0x50
 [<c01039a8>] apic_timer_interrupt+0x1c/0x24
 [<c0100e08>] cpu_idle+0x89/0x9f
 [<c010fcda>] start_secondary+0x17f/0x2ef
 [<00000000>] _stext+0x3feffd68/0x49
 [<dfaa3fb4>] 0xdfaa3fb4
BUG in read_tsc(): TSC went backward!
... Unsynced TSCs?
... [ from 000000003aeeffc4 to 000000003aeedc00 ]
hrtimers: Switched to high resolution mode CPU 0
hrtimers: Switched to high resolution mode CPU 1
hrtimers: Switched to high resolution mode CPU 3
hrtimers: Switched to high resolution mode CPU 2
isapnp: No Plug & Play device found
<snip - continues boot ok>