Skip to contentSkip to conversation

athlon x2 + 2.6.14 + SMP = fast clock

DISCUSSION Christopher Mulcahy 5 participants 8 messages Nov 8, 2005 – Nov 22, 2005

lore
Share & tools
lore
#0 lore raw

I am running 2.6.14 SMP on an dual-core athlon x2 3800. The system clock runs at roughly twice normal speed.

'cat /proc/interrupts' returns:
CPU0 CPU1

  0:   30834589   31107050    IO-APIC-edge  timer
  1:        267      72633    IO-APIC-edge  i8042
  8:          0          0    IO-APIC-edge  rtc
 12:       8662    2216469    IO-APIC-edge  i8042
 15:      43599   10919362    IO-APIC-edge  ide1
 17:       1577     423629   IO-APIC-level  libata
 18:      61263   47314795   IO-APIC-level  ehci_hcd:usb1,
ohci_hcd:usb2, ohci_hcd:usb3
 19:       2026     485190   IO-APIC-level  eth0
 20:         54      14118   IO-APIC-level  ATI IXP
 21:          0          2   IO-APIC-level  acpi, ohci1394
NMI:       2611       3571 
LOC:   30969789   30969781 
ERR:        146
MIS:          0

On my other SMP systems ( admittedly 4-way 2.4.30 ) timer interrupts are processed only by CPU0.

Perhaps unrelated, my kernel log is filled with these messages:
Nov 8 00:44:19 jones kernel: [301738.837187] APIC error on CPU0: 40(40) Nov 8 00:44:19 jones kernel: [301738.849596] APIC error on CPU1: 40(40) Nov 8 02:35:25 jones kernel: [305067.785826] APIC error on CPU0: 40(40) Nov 8 02:35:25 jones kernel: [305067.808330] APIC error on CPU1: 40(40) Nov 8 03:34:47 jones kernel: [306846.571532] APIC error on CPU0: 40(40) Nov 8 03:34:47 jones kernel: [306846.598922] APIC error on CPU1: 40(40) Nov 8 03:50:47 jones kernel: [307325.974012] APIC error on CPU0: 40(40) Nov 8 03:50:47 jones kernel: [307326.002707] APIC error on CPU1: 40(40) Nov 8 03:50:47 jones kernel: [307325.983997] APIC error on CPU0: 40(40) Nov 8 03:50:47 jones kernel: [307326.012689] APIC error on CPU1: 40(40)

Not sure if this is relevant:
http://leaf.dragonflybsd.org/mailarchive/commits/2005-11/msg00038.html

If you want dmesg, .config or other, contact me off list, I am not subscribed.

Chris

#1 lore raw
Quoted 2 lines · Christopher Mulcahy

I am running 2.6.14 SMP on an dual-core athlon x2 3800. The system clock runs at roughly twice normal speed.

Is this a new regression or did the problem occur with 2.6.13 or older kernels?

Would you mind opening a kernel bug and attaching your dmesg and config?

http://bugzilla.kernel.org

Also try booting w/ "idle=poll" to see if that doesn't clear up the issue.

thanks
-john

#2 lore raw
Quoted 6 lines · john stultz
Quoted 2 lines · Christopher Mulcahy

I am running 2.6.14 SMP on an dual-core athlon x2 3800. The system clock runs at roughly twice normal speed.

Is this a new regression or did the problem occur with 2.6.13 or older kernels?

This is a new-machine.
The only other kernel it has seen is the distro-install-kernel ( 2.6.12 uni-processor (ubuntu-5.10) ) ( this kernel does not have a problem, but it is not SMP )

I will try to find time to build 2.4.13 and 2.4.12 SMP kernels with the ~same config to see if they have the same problem. ( I presume I could then attach these findings to the original bugzilla report? )

Quoted 5 lines

Would you mind opening a kernel bug and attaching your dmesg and config?

http://bugzilla.kernel.org

will do.

Quoted 4 lines

Also try booting w/ "idle=poll" to see if that doesn't clear up the issue.

Tried that without results.

Quoted 6 lines

thanks
-john

#3 lore raw
Quoted 7 lines · Christopher Mulcahy
Quoted 6 lines · john stultz
Quoted 2 lines · Christopher Mulcahy

I am running 2.6.14 SMP on an dual-core athlon x2 3800. The system clock runs at roughly twice normal speed.

Is this a new regression or did the problem occur with 2.6.13 or older kernels?

Quoted 8 lines

This is a new-machine.
The only other kernel it has seen is the distro-install-kernel ( 2.6.12 uni-processor (ubuntu-5.10) ) ( this kernel does not have a problem, but it is not SMP )

I will try to find time to build 2.4.13 and 2.4.12 SMP kernels with the ~same config to see if they have the same problem. ( I presume I could then attach these findings to the original bugzilla report? )

There are a few similar sounding bugs out there:
If its an ATI chipset, check out
http://bugzilla.kernel.org/show_bug.cgi?id=3927

If its an nvidia chipset, check out
http://bugzilla.kernel.org/show_bug.cgi?id=3341

Quoted 5 lines
Quoted 4 lines

Would you mind opening a kernel bug and attaching your dmesg and config?

http://bugzilla.kernel.org

will do.

Please tag me as the owner when you do.

Quoted 5 lines
Quoted 4 lines

Also try booting w/ "idle=poll" to see if that doesn't clear up the issue.

Tried that without results.

Bummer. Your box may not boot, but trying noapic might help as well.

thanks
-john

#4 lore raw

On Tue, 08 Nov 2005 14:59:25 -0800
john stultz <[email protected]> mentioned:

Quoted 17 lines · Christopher Mulcahy
Quoted 7 lines · john stultz
Quoted 6 lines · Christopher Mulcahy
Quoted 2 lines

I am running 2.6.14 SMP on an dual-core athlon x2 3800. The system clock runs at roughly twice normal speed.

Is this a new regression or did the problem occur with 2.6.13 or older kernels?

Quoted 8 lines

This is a new-machine.
The only other kernel it has seen is the distro-install-kernel ( 2.6.12 uni-processor (ubuntu-5.10) ) ( this kernel does not have a problem, but it is not SMP )

I will try to find time to build 2.4.13 and 2.4.12 SMP kernels with the ~same config to see if they have the same problem. ( I presume I could then attach these findings to the original bugzilla report? )

Quoted 6 lines

There are a few similar sounding bugs out there:
If its an ATI chipset, check out
http://bugzilla.kernel.org/show_bug.cgi?id=3927

If its an nvidia chipset, check out
http://bugzilla.kernel.org/show_bug.cgi?id=3341

My machine's clock runs about 2X from normal speed. Could you try my patch which I just posted a hour ago?
http://marc.theaimsgroup.com/?l=linux-kernel&m=113249769027262&w=2

The patch will detect whether IO-APCI timer interupt is generated too fast and try to use a legacy i8259A IRQ instead.

It might help. It also worked on 2.4.31 kernel for me.

Quoted 27 lines
Quoted 7 lines
Quoted 6 lines

Would you mind opening a kernel bug and attaching your dmesg and config?

http://bugzilla.kernel.org

will do.

Please tag me as the owner when you do.

Quoted 5 lines
Quoted 4 lines

Also try booting w/ "idle=poll" to see if that doesn't clear up the issue.

Tried that without results.

Bummer. Your box may not boot, but trying noapic might help as well.

thanks

(quoted diff) +0 -2
-john
 
-

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Signature · 1 line
Akira Tsukamoto <[email protected], [email protected]>
#5 lore raw

I am using arch amd64.

I applied this patch.

http://bugzilla.kernel.org/attachment.cgi?id=6061&action=view

It applies to 2.6.14.2 so long as you remove the static declaration of 'int disable_timer_pin_1'

It appears to have solved my problem.

Chris

Quoted 66 lines · Akira Tsukamoto

On Tue, 08 Nov 2005 14:59:25 -0800
john stultz <[email protected]> mentioned:

Quoted 17 lines · Christopher Mulcahy
Quoted 7 lines · john stultz
Quoted 6 lines · Christopher Mulcahy
Quoted 2 lines

I am running 2.6.14 SMP on an dual-core athlon x2 3800. The system clock runs at roughly twice normal speed.

Is this a new regression or did the problem occur with 2.6.13 or older kernels?

Quoted 8 lines

This is a new-machine.
The only other kernel it has seen is the distro-install-kernel ( 2.6.12 uni-processor (ubuntu-5.10) ) ( this kernel does not have a problem, but it is not SMP )

I will try to find time to build 2.4.13 and 2.4.12 SMP kernels with the ~same config to see if they have the same problem. ( I presume I could then attach these findings to the original bugzilla report? )

Quoted 6 lines

There are a few similar sounding bugs out there:
If its an ATI chipset, check out
http://bugzilla.kernel.org/show_bug.cgi?id=3927

If its an nvidia chipset, check out
http://bugzilla.kernel.org/show_bug.cgi?id=3341

My machine's clock runs about 2X from normal speed. Could you try my patch which I just posted a hour ago?
http://marc.theaimsgroup.com/?l=linux-kernel&m=113249769027262&w=2

The patch will detect whether IO-APCI timer interupt is generated too fast and try to use a legacy i8259A IRQ instead.

It might help. It also worked on 2.4.31 kernel for me.

Quoted 27 lines
Quoted 7 lines
Quoted 6 lines

Would you mind opening a kernel bug and attaching your dmesg and config?

http://bugzilla.kernel.org

will do.

Please tag me as the owner when you do.

Quoted 5 lines
Quoted 4 lines

Also try booting w/ "idle=poll" to see if that doesn't clear up the issue.

Tried that without results.

Bummer. Your box may not boot, but trying noapic might help as well.

thanks

(quoted diff) +0 -2
-john
 
-

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

#6 lore raw
Quoted 10 lines · Christopher Mulcahy

I am using arch amd64.

I applied this patch.

http://bugzilla.kernel.org/attachment.cgi?id=6061&action=view

It applies to 2.6.14.2 so long as you remove the static declaration of 'int disable_timer_pin_1'

It appears to have solved my problem.

I wonder is this is x86_64 only problem?

I'm having the same problem on dual xeon 1.8GHz i686 with HT enabled, kernel 2.6.14.2-4smp. Clock runs twice fast. Previously I was using 2.6.11 kernel with no such problem.

cmdline is: acpi=ht nmi_watchdog=1

Unfortunately I can't do any testing right now on this machine but I'll try acpi_skip_timer_override cmdline option as soon as I can.

Quoted 1 line

Chris

00:00.0 Host bridge: Intel Corporation E7500 Memory Controller Hub (rev 03) 00:02.0 PCI bridge: Intel Corporation E7500/E7501 Hub Interface B PCI-to-PCI Bridge (rev 03)
00:1d.0 USB Controller: Intel Corporation 82801CA/CAM USB (Hub #1) (rev 02) 00:1e.0 PCI bridge: Intel Corporation 82801 PCI Bridge (rev 42) 00:1f.0 ISA bridge: Intel Corporation 82801CA LPC Interface Controller (rev
02)
00:1f.1 IDE interface: Intel Corporation 82801CA Ultra ATA Storage Controller (rev 02)
00:1f.3 SMBus: Intel Corporation 82801CA/CAM SMBus Controller (rev 02) 01:01.0 Ethernet controller: Intel Corporation 82557/8/9 [Ethernet Pro 100] (rev 10)
01:02.0 VGA compatible controller: ATI Technologies Inc Rage XL (rev 27) 02:1c.0 PIC: Intel Corporation 82870P2 P64H2 I/OxAPIC (rev 03) 02:1d.0 PCI bridge: Intel Corporation 82870P2 P64H2 Hub PCI Bridge (rev 03) 02:1e.0 PIC: Intel Corporation 82870P2 P64H2 I/OxAPIC (rev 03) 02:1f.0 PCI bridge: Intel Corporation 82870P2 P64H2 Hub PCI Bridge (rev 03) 03:03.0 Fibre Channel: QLogic Corp. QLA2312 Fibre Channel Adapter (rev 02) 04:01.0 Ethernet controller: Intel Corporation 82545EM Gigabit Ethernet Controller (Copper) (rev 01)

# cat /proc/interrupts
           CPU0       CPU1       CPU2       CPU3
  0:   47938745          0          0          0    IO-APIC-edge  timer
  2:          0          0          0          0          XT-PIC  cascade
  4:      41666          0          0          0    IO-APIC-edge  serial
  8:          3          0          0          0    IO-APIC-edge  rtc
145:          0          0          0          0   IO-APIC-level  
uhci_hcd:usb1
153:   49053282          0          0          0   IO-APIC-level  eth0
169:    3839290          0          0          0   IO-APIC-level  qla2300
177:   45129167          0          0          0   IO-APIC-level  eth1
NMI:   47939763   47939734   47939732   47939731
LOC:   47943502   47944012   47944469   47944277
ERR:          0
MIS:          0
Signature · 2 lines
Arkadiusz Miśkiewicz                    PLD/Linux Team
http://www.t17.ds.pwr.wroc.pl/~misiek/  http://ftp.pld-linux.org/
#7 lore raw

On Mon, 21 Nov 2005 10:28:03 +0100
Arkadiusz Miskiewicz <[email protected]> mentioned:

Quoted 6 lines

I wonder is this is x86_64 only problem?

I'm having the same problem on dual xeon 1.8GHz i686 with HT enabled, kernel 2.6.14.2-4smp. Clock runs twice fast. Previously I was using 2.6.11 kernel with no such problem.

My laptop, clock runs twice fast but has PentiumM, so it shouldn't be x86_64 specific.

Quoted 7 lines

cmdline is: acpi=ht nmi_watchdog=1

Unfortunately I can't do any testing right now on this machine but I'll try acpi_skip_timer_override cmdline option as soon as I can.

Quoted 1 line

Chris

Signature · 1 line
Akira Tsukamoto <[email protected]> <>