Skip to contentSkip to conversation

2.6.25.2 - Jiffies/Time jumping back and forth (Regereesion from 2.6.24)

DISCUSSION Martin Knoblauch 3 participants 3 messages May 7, 2008 – May 8, 2008

lore
Share & tools
lore
#0 lore raw

----- Original Message ----

Quoted 37 lines

From: Martin Knoblauch <[email protected]>
To: Bart Van Assche <[email protected]>
Cc: [email protected]; [email protected]
Sent: Wednesday, May 7, 2008 5:02:23 PM
Subject: Re: 2.6.25.2 - Jiffies/Time jumping back and forth (Regereesion from 2.6.24)

----- Original Message ----

Quoted 5 lines

From: Bart Van Assche
To: Martin Knoblauch
Cc: [email protected]; [email protected]
Sent: Wednesday, May 7, 2008 3:06:02 PM
Subject: Re: 2.6.25.2 - Jiffies/Time jumping back and forth (Regereesion from

2.6.24)

Quoted 6 lines · Martin Knoblauch
Quoted 2 lines

any news on the one below? I upgraded to 2.6.25.2 and the problem still

persists. Would adding a bugzilla actually help ? :-)

Quoted 1 line
Quoted 8 lines
Quoted 1 line

http://www.ussg.iu.edu/hypermail/linux/kernel/0804.2/1144.html

I noticed a similar issue on 2.6.22 and 2.6.24 kernels: with ACPI disabled Linux' clock jumps backward and forward early during boot. See also the kernel messages from the attachment in comment #13 of
http://bugzilla.kernel.org/show_bug.cgi?id=9936.

Bart,

Hi Bart,

no this looks different. As soon as the timesource is detected, I see non-zero timestamps
in dmesg, but they look smaller and jump in both directions. And I have never seen this before
2.6.25.

Quoted 4 lines

It actually might be a good idea to get the dmesg output from 2.6.24.

Cheers
Martin

interesting - maybe related to clocksources. The HPET seems to be detected later in 2.6.25:

2.6.24:

[    0.000000] Initializing CPU#0
[    0.000000] PID hash table entries: 4096 (order: 12, 32768 bytes)
-->[    0.000000] hpet clockevent registered
-->[    0.000000] TSC calibrated against HPET
[  150.233392] time.c: Detected 2327.502 MHz processor.
[  150.235525] Console: colour VGA+ 80x25
[  150.235532] console [tty0] enabled
[  150.235546] Lock dependency validator: Copyright (c) 2006 Red Hat, Inc., Ingo Molnar
[  150.235548] ... MAX_LOCKDEP_SUBCLASSES:    8
[  150.235549] ... MAX_LOCK_DEPTH:          30
[  150.235550] ... MAX_LOCKDEP_KEYS:        2048
[  150.235551] ... CLASSHASH_SIZE:           1024
[  150.235552] ... MAX_LOCKDEP_ENTRIES:     8192
[  150.235553] ... MAX_LOCKDEP_CHAINS:      16384
[  150.235554] ... CHAINHASH_SIZE:          8192
[  150.235556]  memory used by lock dependency info: 1712 kB
[  150.235557]  per task-struct memory footprint: 2160 bytes
[  150.235559] Checking aperture...
[  150.235568] PCI-DMA: Using software bounce buffering for IO (SWIOTLB)
[  150.267666] Placing software IO TLB between 0x123c000 - 0x523c000
[  150.377395] Memory: 8086092k/9437184k available (2901k kernel code, 301896k reserved, 1534k data, 772k init)
[  150.454897] Calibrating delay using timer specific routine.. 4659.71 BogoMIPS (lpj=9319423)

2.6.25.2:

[    0.000000] Initializing CPU#0
[    0.000000] PID hash table entries: 4096 (order: 12, 32768 bytes)
-->[    0.000000] TSC calibrated against PM_TIMER
[    0.000004] time.c: Detected 2327.497 MHz processor.
[    0.002079] Console: colour VGA+ 80x25
[    0.002086] console [tty0] enabled
[    0.002107] Lock dependency validator: Copyright (c) 2006 Red Hat, Inc., Ingo Molnar
[    0.002109] ... MAX_LOCKDEP_SUBCLASSES:    8
[    0.002110] ... MAX_LOCK_DEPTH:          48
[    0.002111] ... MAX_LOCKDEP_KEYS:        2048
[    0.002112] ... CLASSHASH_SIZE:           1024
[    0.002114] ... MAX_LOCKDEP_ENTRIES:     8192
[    0.002115] ... MAX_LOCKDEP_CHAINS:      16384
[    0.002116] ... CHAINHASH_SIZE:          8192
[    0.002117]  memory used by lock dependency info: 1712 kB
[    0.002119]  per task-struct memory footprint: 3456 bytes
[    0.002121] Checking aperture...
[    0.002130] PCI-DMA: Using software bounce buffering for IO (SWIOTLB)
[    0.004000] Placing software IO TLB between 0x15093000 - 0x19093000
[    0.004000] Memory: 8110236k/9437184k available (2950k kernel code, 277752k reserved, 1567k data, 800k init)
[    0.004000] CPA: page pool initialized 1 of 1 pages preallocated
-->[    0.004000] hpet clockevent registered
[    0.080008] Calibrating delay using timer specific routine.. 4660.31 BogoMIPS (lpj=9320632)

Any ideas how to force hpet earlier? Specifying hpet=force and/or clocksource=hpet did not help.

Cheers
Martin

#1 lore raw
Quoted 17 lines · Martin Knoblauch
Quoted 10 lines

no this looks different. As soon as the timesource is detected, I see non-zero timestamps
in dmesg, but they look smaller and jump in both directions. And I have never seen this before
2.6.25.

It actually might be a good idea to get the dmesg output from 2.6.24.

Cheers
Martin

interesting - maybe related to clocksources. The HPET seems to be detected later in 2.6.25:

2.6.24:

[    0.000000] Initializing CPU#0
[    0.000000] PID hash table entries: 4096 (order: 12, 32768 bytes)
Quoted 8 lines

-->[ 0.000000] hpet clockevent registered
-->[ 0.000000] TSC calibrated against HPET

2.6.25.2:

[    0.000000] Initializing CPU#0
[    0.000000] PID hash table entries: 4096 (order: 12, 32768 bytes)

-->[ 0.000000] TSC calibrated against PM_TIMER

That's odd indeed. Is HPET not registered at all in 2.6.25 or does it just happen late ?

please provide the output of:
/sys/devices/system/clocksource/clocksource0/available_clocksource /sys/devices/system/clocksource/clocksource0/current_clocksource

Thanks,
tglx

#2 lore raw
Quoted 26 lines · Thomas Gleixner
Quoted 16 lines · Martin Knoblauch
Quoted 10 lines

no this looks different. As soon as the timesource is detected, I see non-zero timestamps
in dmesg, but they look smaller and jump in both directions. And I have never seen this before
2.6.25.

It actually might be a good idea to get the dmesg output from 2.6.24.

Cheers
Martin

interesting - maybe related to clocksources. The HPET seems to be detected later in 2.6.25:

2.6.24:

[    0.000000] Initializing CPU#0
[    0.000000] PID hash table entries: 4096 (order: 12, 32768 bytes)
Quoted 8 lines

-->[ 0.000000] hpet clockevent registered
-->[ 0.000000] TSC calibrated against HPET

2.6.25.2:

[    0.000000] Initializing CPU#0
[    0.000000] PID hash table entries: 4096 (order: 12, 32768 bytes)

-->[ 0.000000] TSC calibrated against PM_TIMER

Quoted 5 lines

That's odd indeed. Is HPET not registered at all in 2.6.25 or does it just happen late ?

please provide the output of:
/sys/devices/system/clocksource/clocksource0/available_clocksource /sys/devices/system/clocksource/clocksource0/current_clocksource

I was about to report / ask bout this issue too.

I see the same output for >=2.6.25 , TSC calibrated against PM_TIMER and not against HPET anymore.

[    0.000000] TSC calibrated against PM_TIMER
[    0.000016] time.c: Detected 2499.949 MHz processor.
...
[    0.000999] hpet clockevent registered

So it seems to register to late in >=2.6.25 kernels ?

cat /sys/devices/system/clocksource/clocksource0/available_clocksource tsc hpet acpi_pm jiffies

cat /sys/devices/system/clocksource/clocksource0/current_clocksource tsc

Quoted 3 lines

Thanks,
tglx

Regards,

Gabriel C