Bug#479709: Same behavior with 2.6.26-2
Hello,
after an update to 2.6.26-1-amd64 (Debian 2.6.26-2), I re-enabled
HPET-Support in Bios for testing purposes.
I get the same freeze after some time, so I have to disable HPET again.
For me 2.6.24 works fine, 2.6.25 and later versions are not usable with
HPET support (and chrony).
So maybe someone should reopen this bug.
I'm slightly confused, because
/sys/devices/system/clocksource/clocksource0/current_clocksource
is set to tsc, when HPET is enabled by Bios, but the system is crashing
anyway.
cat /var/log/syslog | grep rtc:
Aug 11 20:08:46 deepthought kernel: [ 0.771453] pnp: the driver
'rtc_cmos' has been registered
Aug 11 20:08:46 deepthought kernel: [ 0.771453] rtc_cmos 00:04: rtc
core: registered rtc_cmos as rtc0
Aug 11 20:08:46 deepthought kernel: [ 0.771453] rtc0: alarms up to
one month
Aug 11 20:08:46 deepthought kernel: [ 0.771453] rtc_cmos 00:04:
driver attached
Aug 11 20:08:46 deepthought kernel: [ 0.771453] rtc_cmos 00:04:
setting system clock to 2008-08-11 18:08:29 UTC (1218478109)
Aug 11 20:43:04 deepthought kernel: [ 2244.526275] rtc: lost 9 interrupts
Aug 11 20:43:05 deepthought kernel: [ 2245.533120] rtc: lost 9 interrupts
freeze at 20:43 with no further messages
cat /var/log/syslog | grep hpet:
Aug 11 20:08:46 deepthought kernel: [ 0.004000] hpet clockevent
registered
Aug 11 20:08:46 deepthought kernel: [ 0.275194] hpet0: at MMIO
0xfed00000, IRQs 2, 8, 0, 0
Aug 11 20:08:46 deepthought kernel: [ 0.275198] hpet0: 4 64-bit
timers, 14318180 Hz
Aug 11 20:08:46 deepthought kernel: [ 0.747453] hpet_resources:
0xfed00000 is busy
best regards
Jörgen
Reply to: