Arch Linux

Please read this before reporting a bug:
https://wiki.archlinux.org/title/Bug_reporting_guidelines

Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.

REPEAT: Do NOT report bugs for outdated packages!
Tasklist

FS#6354 - System DateTime Problem

Attached to Project: Arch Linux
Opened by Roberto Spadim (rspadim) - Tuesday, 06 February 2007, 16:49 GMT
Task Type Bug Report
Category Kernel
Status Closed
Assigned To No-one
Architecture not specified
Severity Critical
Priority Normal
Reported Version 0.7.2 Gimmick
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No

Details

i have a new pentium D 820 (dual core), i'm using amd64 since it's 64bits,
i don't know why my datetime is crazy, changing 2 seconds in past and to
seconds in future again



i have sent it to gentto too but i don't have portage so they can't help me: http://bugs.gentoo.org/show_bug.cgi?id=165640
This task depends upon

Closed by  Andreas Radke (AndyRTR)
Wednesday, 07 March 2007, 06:33 GMT
Reason for closing:  Works for me
Comment by Roberto Spadim (rspadim) - Tuesday, 06 February 2007, 16:50 GMT
i'm using arch64 - version 0.8 voodoo
[root@spadim ~]# cat /proc/version
Linux version 2.6.19-ARCH (root@workstation64) (gcc version 4.1.2 20061215 (prerelease)) #1 SMP PREEMPT Thu Jan 11 20:36:34 UTC 2007


Reproducible: Always

Steps to Reproduce:
1.check date and time values from system
2.check time from hardware clock
3.i don't know why but uptime is ok, if i check uptime it only add

Actual Results:
[root@spadim kernel]# date
Ter Fev 6 13:44:58 BRST 2007
[root@spadim kernel]# date
Ter Fev 6 13:44:56 BRST 2007 <-- time get back 2 seconds!
[root@spadim kernel]# date
Ter Fev 6 13:44:59 BRST 2007 <-- time continue ok
[root@spadim kernel]# date
Ter Fev 6 13:44:57 BRST 2007 <-- time get back 2 seconds!
[root@spadim kernel]# date
Ter Fev 6 13:45:00 BRST 2007 <-- time conttinue ok!

my hw clock is ok see:
[root@spadim kernel]# hwclock
Ter 06 Fev 2007 13:46:52 BRST -3.000384 segundos
[root@spadim kernel]# hwclock
Ter 06 Fev 2007 13:46:53 BRST -0.174015 segundos
[root@spadim kernel]# hwclock
Ter 06 Fev 2007 13:46:54 BRST -0.321278 segundos
[root@spadim kernel]# hwclock
Ter 06 Fev 2007 13:46:55 BRST -0.366134 segundos
[root@spadim kernel]# hwclock
Ter 06 Fev 2007 13:46:56 BRST -0.352971 segundos
Ter 06 Fev 2007 13:47:12 BRST -0.292030 segundos
[root@spadim kernel]#
[root@spadim kernel]# hwclock
Ter 06 Fev 2007 13:47:16 BRST -3.220447 segundos <- system is 3 seconds
slower?
Linux version 2.6.19
gcc version 4.1.2 20061215



Expected Results:
[root@spadim kernel]# date
Ter Fev 6 13:44:58 BRST 2007
[root@spadim kernel]# date
Ter Fev 6 13:44:58 BRST 2007 <-- should continue working ok
[root@spadim kernel]# date
Ter Fev 6 13:44:59 BRST 2007 <-- time continue ok
[root@spadim kernel]# date
Ter Fev 6 13:44:59 BRST 2007 <-- should continue working ok
[root@spadim kernel]# date
Ter Fev 6 13:45:00 BRST 2007 <-- time conttinue ok!

my hw clock is ok see:
[root@spadim kernel]# hwclock
Ter 06 Fev 2007 13:46:52 BRST -3.000384 segundos <- should be 0 seconds or <1
second
[root@spadim kernel]# hwclock
Ter 06 Fev 2007 13:46:53 BRST -0.174015 segundos
[root@spadim kernel]# hwclock
Ter 06 Fev 2007 13:46:54 BRST -0.321278 segundos
[root@spadim kernel]# hwclock
Ter 06 Fev 2007 13:46:55 BRST -0.366134 segundos
[root@spadim kernel]# hwclock
Ter 06 Fev 2007 13:46:56 BRST -0.352971 segundos
Ter 06 Fev 2007 13:47:12 BRST -0.292030 segundos
[root@spadim kernel]#
[root@spadim kernel]# hwclock
Ter 06 Fev 2007 13:47:16 BRST -3.220447 segundos <- system is 3 seconds
slower?
Linux version 2.6.19
gcc version 4.1.2 20061215



i don't know if it's a problem with cpu, since it's dual core i don't know if
it can't hangle system time ok, since i have time based programs, i can't run
any program, they get time values with (-1.233 seconds in past and stop
working)
Comment by Roberto Spadim (rspadim) - Tuesday, 06 February 2007, 16:53 GMT
ip traf don't work too, i have an amd atlhon64 3000+ (single core) and i have the same error, since it's a mysql server i didn't see that
Comment by Roberto Spadim (rspadim) - Tuesday, 06 February 2007, 16:55 GMT
intel core 2 duo E6600 work ok (asus p5ld2se core2 version),
pentium d 820 don't (ecs p4m800pro)
Comment by Roberto Spadim (rspadim) - Tuesday, 06 February 2007, 17:27 GMT
very crazy... see

[root@spadim ~]# date +%H:%M:%S.%N
15:24:55.976491000
[root@spadim ~]# date +%H:%M:%S.%N
15:24:58.646311000
[root@spadim ~]# date +%H:%M:%S.%N
15:24:56.492530000
[root@spadim ~]# date +%H:%M:%S.%N
15:24:59.136100000
[root@spadim ~]# date +%H:%M:%S.%N
15:24:57.038754000


with pentium 4 HT 524 it don't happen,
(with the same mother board) but with Damn Small Linux (kernel 2.4)
i think that's a kernel bug
Comment by Jan de Groot (JGC) - Tuesday, 06 February 2007, 18:24 GMT
Yes, experiencing the same thing. In one week time my clock was 15 minutes in the future. Running ntpd helps though.
Do you guys use cpufreq modules to throttle down these CPUs? In my case, it could be due to a lot of CPU throttling between 75% and 100% CPU speed.
Comment by Roberto Spadim (rspadim) - Tuesday, 06 February 2007, 18:31 GMT
i'm not using cpufreq
Comment by Jan de Groot (JGC) - Tuesday, 06 February 2007, 21:56 GMT
OK, then the thing we have in common is a dualcore Intel CPU. It seems adding "noapictimer" to your kernel boot commandline will help out. I don't know if this is still an issue with 2.6.20, but time will tell (either negative or positive, it depends on where your clock goes :P)
Comment by Roberto Spadim (rspadim) - Wednesday, 07 February 2007, 16:24 GMT
i will try noapictimer, with noapic didn't worked, just wait some time to get server down and i will return to you
Comment by Roberto Spadim (rspadim) - Wednesday, 07 February 2007, 16:25 GMT
i will try kernel26-ck and check if it have the same problem
Comment by Roberto Spadim (rspadim) - Thursday, 08 February 2007, 04:12 GMT
with kernel26 and kernel26-ck the same thing occur
with noapic kernel boot but clock stay wrong
with noapictimer kernel don't boot and stop here:

Feb 8 00:11:01 spadim Linux version 2.6.19-ck (root@spadim) (gcc version 4.1.2 20061215 (prerelease)) #1 SMP PREEMPT Wed Feb 7 05:49:25 BRST 2007
Feb 8 00:11:01 spadim Command line: root=/dev/sda3 ro noapictimer
Feb 8 00:11:01 spadim BIOS-provided physical RAM map:
Feb 8 00:11:01 spadim BIOS-e820: 0000000000000000 - 000000000009fc00 (usable)
Feb 8 00:11:01 spadim BIOS-e820: 000000000009fc00 - 00000000000a0000 (reserved)
Feb 8 00:11:01 spadim BIOS-e820: 00000000000e6000 - 0000000000100000 (reserved)
Feb 8 00:11:01 spadim BIOS-e820: 0000000000100000 - 000000003efc0000 (usable)
Feb 8 00:11:01 spadim BIOS-e820: 000000003efc0000 - 000000003efce000 (ACPI data)
Feb 8 00:11:01 spadim BIOS-e820: 000000003efce000 - 000000003eff0000 (ACPI NVS)
Feb 8 00:11:01 spadim BIOS-e820: 000000003eff0000 - 000000003f000000 (reserved)
Feb 8 00:11:01 spadim BIOS-e820: 00000000fec00000 - 00000000fec01000 (reserved)
Feb 8 00:11:01 spadim BIOS-e820: 00000000fee00000 - 00000000fee01000 (reserved)
Feb 8 00:11:01 spadim Entering add_active_range(0, 0, 159) 0 entries of 256 used
Feb 8 00:11:01 spadim Entering add_active_range(0, 256, 257984) 1 entries of 256 used
Feb 8 00:11:01 spadim end_pfn_map = 1043969
Feb 8 00:11:01 spadim DMI 2.3 present.
Feb 8 00:11:01 spadim ACPI: RSDP (v000 ACPIAM ) @ 0x00000000000f8220
Feb 8 00:11:01 spadim ACPI: RSDT (v001 A M I OEMRSDT 0x03000607 MSFT 0x00000097) @ 0x000000003efc0000
Feb 8 00:11:01 spadim ACPI: FADT (v002 A M I OEMFACP 0x03000607 MSFT 0x00000097) @ 0x000000003efc0200
Feb 8 00:11:01 spadim ACPI: MADT (v001 A M I OEMAPIC 0x03000607 MSFT 0x00000097) @ 0x000000003efc0390
Feb 8 00:11:01 spadim ACPI: OEMB (v001 A M I AMI_OEM 0x03000607 MSFT 0x00000097) @ 0x000000003efce040
Feb 8 00:11:01 spadim ACPI: DSDT (v001 12345 12345123 0x00000123 INTL 0x02002026) @ 0x0000000000000000
Feb 8 00:11:01 spadim Entering add_active_range(0, 0, 159) 0 entries of 256 used
Feb 8 00:11:01 spadim Entering add_active_range(0, 256, 257984) 1 entries of 256 used
Feb 8 00:11:01 spadim Zone PFN ranges:
Feb 8 00:11:01 spadim DMA 0 -> 4096
Feb 8 00:11:01 spadim DMA32 4096 -> 1048576
Feb 8 00:11:01 spadim Normal 1048576 -> 1048576
Feb 8 00:11:01 spadim early_node_map[2] active PFN ranges
Feb 8 00:11:01 spadim 0: 0 -> 159
Feb 8 00:11:01 spadim 0: 256 -> 257984
Feb 8 00:11:01 spadim On node 0 totalpages: 257887
Feb 8 00:11:01 spadim DMA zone: 56 pages used for memmap
Feb 8 00:11:01 spadim DMA zone: 1002 pages reserved
Feb 8 00:11:01 spadim DMA zone: 2941 pages, LIFO batch:0
Feb 8 00:11:01 spadim DMA32 zone: 3471 pages used for memmap
Feb 8 00:11:01 spadim DMA32 zone: 250417 pages, LIFO batch:31
Feb 8 00:11:01 spadim Normal zone: 0 pages used for memmap
Feb 8 00:11:01 spadim ACPI: PM-Timer IO Port: 0x808
Feb 8 00:11:01 spadim ACPI: Local APIC address 0xfee00000
Feb 8 00:11:01 spadim ACPI: LAPIC (acpi_id[0x01] lapic_id[0x00] enabled)
Feb 8 00:11:01 spadim Processor #0 (Bootup-CPU)
Feb 8 00:11:01 spadim ACPI: LAPIC (acpi_id[0x02] lapic_id[0x01] enabled)
Feb 8 00:11:01 spadim Processor #1
Feb 8 00:11:01 spadim ACPI: LAPIC (acpi_id[0x03] lapic_id[0x82] disabled)
Feb 8 00:11:01 spadim ACPI: LAPIC (acpi_id[0x04] lapic_id[0x83] disabled)
Feb 8 00:11:01 spadim ACPI: IOAPIC (id[0x02] address[0xfec00000] gsi_base[0])
Feb 8 00:11:01 spadim IOAPIC[0]: apic_id 2, address 0xfec00000, GSI 0-23
Feb 8 00:11:01 spadim ACPI: INT_SRC_OVR (bus 0 bus_irq 0 global_irq 2 dfl dfl)
Feb 8 00:11:01 spadim ACPI: INT_SRC_OVR (bus 0 bus_irq 9 global_irq 9 low level)
Feb 8 00:11:01 spadim ACPI: IRQ0 used by override.
Feb 8 00:11:01 spadim ACPI: IRQ2 used by override.
Feb 8 00:11:01 spadim ACPI: IRQ9 used by override.
Feb 8 00:11:01 spadim Setting APIC routing to physical flat
Feb 8 00:11:01 spadim Using ACPI (MADT) for SMP configuration information
Feb 8 00:11:01 spadim Nosave address range: 000000000009f000 - 00000000000a0000
Feb 8 00:11:01 spadim Nosave address range: 00000000000a0000 - 00000000000e6000
Feb 8 00:11:01 spadim Nosave address range: 00000000000e6000 - 0000000000100000
Feb 8 00:11:01 spadim Allocating PCI resources starting at 40000000 (gap: 3f000000:bfc00000)
Feb 8 00:11:01 spadim SMP: Allowing 4 CPUs, 2 hotplug CPUs
Feb 8 00:11:01 spadim PERCPU: Allocating 30848 bytes of per cpu data
Feb 8 00:11:01 spadim Built 1 zonelists. Total pages: 253358
Feb 8 00:11:01 spadim Kernel command line: root=/dev/sda3 ro
Feb 8 00:11:01 spadim Initializing CPU#0
Feb 8 00:11:01 spadim PID hash table entries: 4096 (order: 12, 32768 bytes)
Feb 8 00:11:01 spadim Console: colour VGA+ 80x25
Feb 8 00:11:01 spadim Dentry cache hash table entries: 131072 (order: 8, 1048576 bytes)
Feb 8 00:11:01 spadim Inode-cache hash table entries: 65536 (order: 7, 524288 bytes)
Feb 8 00:11:01 spadim Checking aperture...
Feb 8 00:11:01 spadim Memory: 1010128k/1031936k available (2355k kernel code, 21324k reserved, 940k data, 224k init)
Feb 8 00:11:01 spadim Calibrating delay using timer specific routine.. 6025.01 BogoMIPS (lpj=3012505)
Feb 8 00:11:01 spadim Security Framework v1.0.0 initialized
Feb 8 00:11:01 spadim Mount-cache hash table entries: 256
Feb 8 00:11:01 spadim CPU: Trace cache: 12K uops, L1 D cache: 16K
Feb 8 00:11:01 spadim CPU: L2 cache: 1024K
Feb 8 00:11:01 spadim using mwait in idle threads.
Feb 8 00:11:01 spadim CPU: Physical Processor ID: 0
Feb 8 00:11:01 spadim CPU: Processor Core ID: 0
Feb 8 00:11:01 spadim CPU0: Thermal monitoring enabled (TM1)
Feb 8 00:11:01 spadim SMP alternatives: switching to UP code
Feb 8 00:11:01 spadim checking if image is initramfs... it is
Feb 8 00:11:01 spadim Freeing initrd memory: 1464k freed
Feb 8 00:11:01 spadim ACPI: Core revision 20060707
Feb 8 00:11:01 spadim ACPI: Looking for DSDT in initramfs... error, file /DSDT.aml not found.
Feb 8 00:11:01 spadim Using local APIC timer interrupts.
Feb 8 00:11:01 spadim result 13437964
Feb 8 00:11:01 spadim Detected 13.437 MHz APIC timer.
Feb 8 00:11:01 spadim SMP alternatives: switching to SMP code
Feb 8 00:11:01 spadim Booting processor 1/2 APIC 0x1
Feb 8 00:11:01 spadim Initializing CPU#1
Feb 8 00:11:01 spadim Calibrating delay using timer specific routine.. 6019.41 BogoMIPS (lpj=3009705)
Feb 8 00:11:01 spadim CPU: Trace cache: 12K uops, L1 D cache: 16K
Feb 8 00:11:01 spadim CPU: L2 cache: 1024K
Feb 8 00:11:01 spadim CPU: Physical Processor ID: 0
Feb 8 00:11:01 spadim CPU: Processor Core ID: 1
Feb 8 00:11:01 spadim CPU1: Thermal monitoring enabled (TM1)
Feb 8 00:11:01 spadim Intel(R) Pentium(R) D CPU 2.80GHz stepping 07
Feb 8 00:11:01 spadim Brought up 2 CPUs
Feb 8 00:11:01 spadim testing NMI watchdog ... OK.
Feb 8 00:11:01 spadim time.c: Using 3.579545 MHz WALL PM GTOD PIT/TSC timer.
Feb 8 00:11:01 spadim time.c: Detected 3010.105 MHz processor.

here kernel stop and don't continue loading, no kernel panic message or keyboard led blinking, just stopped machine
any help? maybe kernel 2.6.20 could work? i didn't tested with DSL since it's a 32bits distrib and 2.4 kernel version, but i will try
Comment by Roberto Spadim (rspadim) - Thursday, 08 February 2007, 04:18 GMT
on intel core 2 duo e6600:

Calibrating delay using timer specific routine.. 4811.91 BogoMIPS (lpj=9623838)
Using local APIC timer interrupts.
Detected 16.695 MHz APIC timer.
Calibrating delay using timer specific routine.. 4808.66 BogoMIPS (lpj=9617330)
time.c: Using 3.579545 MHz WALL PM GTOD PIT/TSC timer.
time.c: Detected 2404.182 MHz processor.
Comment by héctor (hacosta) - Thursday, 08 February 2007, 08:25 GMT
both 2.6.19 (custom) and 2.6.20 (stock) work fine here.
Comment by Roberto Spadim (rspadim) - Thursday, 08 February 2007, 17:58 GMT
where could i get 2.6.20(stock) ?
i tested with damn small linux and clock worked ok, i think that's a kernel problem, i got the pentium4 processor and tested with the same board and worked ok with arch linux, just pentium d 820 didn't.
2.6.20 can be get with abs?
any link?
how could i make an mkinitcpio? i was using mkinit and idon't know how this mkinitcpio work
thanks

i think that sata_via have problems too, since i'm using an NCQ sata2 disk and i get 300 mb/s at cached reads with hdparm, it could be 1500mb/s ou somethink like it, i don't know how to use ahci since i never used mkinitcpio :/
Comment by Roberto Spadim (rspadim) - Thursday, 08 February 2007, 18:10 GMT
check that my motherboard is p4m800pro-m (ecs), i don't remember chipset maybe vt8237 plus or something like it
Comment by Jan de Groot (JGC) - Thursday, 08 February 2007, 18:19 GMT
kernel26 2.6.20 is in testing, you can just download the package from FTP and install it if you like (or uncomment testing, run pacman -Sy, pacman -S kernel26 and revert pacman.conf after that).

About the SATA issue: check the jumper on your harddisk, many disks come with a limitation by jumper that limits it to SATA-1. The cached reads have nothing to do with harddisk speed btw, my value is higher than SATA2 allows ;)
Comment by Roberto Spadim (rspadim) - Thursday, 08 February 2007, 18:22 GMT
[root@spadim log]# pacman -S kernel26
:: kernel26-2.6.19.2-1: is up to date. Upgrade anyway? [Y/n]

it should be 2.6.20?
Comment by Roberto Spadim (rspadim) - Thursday, 08 February 2007, 18:29 GMT
kernel 2.6.20 is only for x86, it didn't exist at x86_64 :/
Comment by Roberto Spadim (rspadim) - Thursday, 08 February 2007, 18:38 GMT
how can i get abs of kernel 2.6.20?
Comment by Roberto Spadim (rspadim) - Saturday, 10 February 2007, 13:55 GMT
using 32 bits clock get ok, but i want 64 bits...
Comment by Roberto Spadim (rspadim) - Thursday, 22 February 2007, 05:45 GMT
i'm with problem yet...
i will try kernel 2.6.20 for 64 bits and put feedback here again
Comment by Andreas Radke (AndyRTR) - Wednesday, 07 March 2007, 06:32 GMT
[andyrtr@workstation64 ~]$ uname -a
Linux workstation64 2.6.20-ARCH #1 SMP PREEMPT Wed Feb 28 22:02:06 UTC 2007 x86_64 Intel(R) Core(TM)2 CPU 6600 @ 2.40GHz GenuineIntel GNU/Linux

[andyrtr@workstation64 ~]$ i=1; while [ $i ]; do date; sleep 1; done
Mi 7. Mär 07:31:07 CET 2007
Mi 7. Mär 07:31:08 CET 2007
Mi 7. Mär 07:31:09 CET 2007
Mi 7. Mär 07:31:10 CET 2007
Mi 7. Mär 07:31:11 CET 2007
Mi 7. Mär 07:31:12 CET 2007
Mi 7. Mär 07:31:13 CET 2007
Mi 7. Mär 07:31:14 CET 2007
Mi 7. Mär 07:31:15 CET 2007
Mi 7. Mär 07:31:16 CET 2007


looks ok here. will close this for now as it's not reproducable. if it happens again ask to reopen it and let it assign to me.

Loading...