Warning : many lost ticks, time source seems to be instable

LordPhobos · n00b Joined: 22 May 2003 Posts: 15

Has anyone found a fix for this yet? I'm running the 2.6.10-r7 Gentoo-Dev kernel on my eMachines M6805 (AMD64) laptop and I'm still getting this "lost ticks" message. Would definitely like to fix it if a solution has been found.

vyzivus · Posted: Mon Feb 28, 2005 12:18 am Post subject:

Interesting, I thought that they fixed the problem already. I don't have the 'lost ticks' messages in my dmesg anymore - they stopped when I tried gentoo-dev-sources-2.6.9-r14 if i remember correctly.

Lately I'm having issues with my keyboard - it does not seem to react on keypress (even on key release - the key is 'stuck' and repeat until I punch it again). The keyboard seems to start working correctly when I shut down gkrellm and gnome-battery-applet. This is very interesting because I remember that the 'lost ticks' messages stopped when I did exactly this (stopped both gkrellm and that battery-panel).

I think that these two bugs are related and (maybe only in my case) connected somehow to buggy ACPI driver. If this is the case then why did the 'lost ticks' message stopped spawning in my log? I don't know. Hope someone out there does :wink:

_________________
I thought what I'd do was, I'd pretend I was one of those deaf-mutes or should I?

DiscreteOnyxOrb · n00b Joined: 08 Apr 2004 Posts: 24

Can anyone else confirm that upgrading to gentoo-dev-sources-2.6.9-r14 fixes the problem? I'm on a K8V Deluxe and my particular incarnation of the problem seems to be tied to my VMware habit. I have ACPI compeltely turned off. vyzivus, what hardware are you using?

vyzivus · Posted: Tue Mar 01, 2005 12:35 pm Post subject:

I have VIA chipset, that's for sure. But I can't detect my motherboard anyhow! I have acer aspire 1501lmi laptop, here is the lspci output:

ShadowHawkBV · Guru Joined: 27 Mar 2004 Posts: 352

This problem seems to be caused by numerous things. Mine was caused by

USB 1.1 support in the BIOS being on. as well as HID support and VIA 82XX sound being activated in the kernel as *.

If memory serves, the 2.6.9+ or 2.6.10+ kernels stopped the reporting of these error messages. There is an option to turn the error message back on in the menuconfig, if it is still occuring. Everyone seems to have an opinion on what is causing it, but no one solution seems to fix everyone.

Good luck.
_________________
This space for rent... Well maybe to give away.. Heck.. i'll pay you to take it.

Lost Linux Neophyte
Intel i7-1065G7
Intel i7-8565U
Intel Atom Cherry Trail
AMD Phenom(tm) II X4 955
Pure 64bit frustration :-)

ahorn · Guru Joined: 01 Nov 2004 Posts: 366 Location: Ol' europe

what is with the kernel parameter 'notsc' and 'nohpet', some1 tested it?
_________________
Too many connections. Please try again later.

Entropius · n00b Joined: 24 Aug 2004 Posts: 16

In my case it's not just error messages -- the clock is losing about twenty minutes a day, making me late for work yesterday.

I'm running a 32-bit Gentoo system on an AMD64 laptop (Via motherboard), if that helps. Enabling the "power management timer support" in the kernel and recompiling now ... we'll see if that helps.

alaa · n00b Joined: 23 Feb 2004 Posts: 57

I'm running 2.6.11-gentoo-r7 and have been having problems running ndiswrapper on my AMD64 machine. Whenever I do that it causes a process called events/0 to take up 99% cpu.

dmesg shows:

numbaonestunna · n00b Joined: 05 Mar 2003 Posts: 48

but i have the *exact* same problem, and it *pisses* me off. Anyone have any ideas? Stupid ndiswrapper!

dangertools · Posted: Tue May 17, 2005 3:19 pm Post subject:

stupid ndiswrapper? don't use it, support it or pay for driverloader! ndiswrapper devs don't have amd64

loftwyr · Posted: Wed May 18, 2005 2:15 am Post subject:

I have tried both ndiiswrapper and driverloader. I get the same problem with both.
_________________
My emerge --info
Have you run revdep-rebuild lately? It's in gentoolkit and it's worth a shot if things don't work well.
Celebrating 5 years of Gentoo-ing.

new2gen2 · n00b Joined: 27 May 2005 Posts: 1

I am afraid I got bad news for everyone that has this problem!
I have 4 different Linux 64 bit OSs on one 80GB SATA hard drive.
Because the CMOS date & time are effected so are the OSs!
Ive traced back the problem back to the AMD64 2004 install CD
which I downloaded & burned on April 10 2005.
I am now in the process of starting all over again.
DANG!!! That was the only Gentoo I got to work out of the last 4 years.
Try booting the install CD again and watch for it or simply do a
"dmesg | less".
:oops:

UPDATE: It was the Universal AMD64 install CD (above date)that has the problem.
Started over by downloading and burning Minimum AMD64 CD then
seperately downloading stage3 and latest portage.
e.c.

dangertools · Posted: Fri May 27, 2005 6:38 am Post subject:

my installation one year ago was using the minimal liveCD 2004.2 or 2004.1, don't remember exactly, from stage1. atm i'm on 2005.0 and i am running a vanilla kernel. ndiswrapper fires my wifi. my "lost ticks output" is as follows:

puggy · Posted: Fri May 27, 2005 7:43 pm Post subject:

There is a fix for this in vanilla 2.6.12 release candidate for this. It's the only thing that appears in make oldconfig going from rc3 to 4 for me.
_________________
Where there's open source , there's a way.

paulisdead · Guru Joined: 10 Apr 2002 Posts: 510 Location: Seattle, WA

Sorry to drench up an old topic, but I've got this problem too now.

I just upgraded to a dual core 4400 and 2x 7800GTXs, and now when the system is under a heavy load, the system clock runs too fast. At first, it could get really bad to the point where a video I was watching was running like it was in fast forward, but after tweaking around with it, I only got ahead about 4 minutes overnight while doing repeated xorg compiles. I updated my bios to the latest one available for the neo4 SLI, and have tried both the 2.6.12.1 vanilla and the 2.6.12-gentoo-r1 kernels with SMP enabled. Both cores are properly detected, and loads are distributed between them. Here's the exact error I see in dmesg

paulisdead · Guru Joined: 10 Apr 2002 Posts: 510 Location: Seattle, WA

OK, well I know it's not specific to the MSI Neo4. I swapped it out for the DFI Nf4 UT, and am still having my clock run fast. I also gave the 2.6.12-git8 kernel a shot, and switched to udev in the process, and haven't seen any improvement.
_________________
"we should make it a law that all geeks have dates" - Linus

vyzivus · Posted: Tue Jun 28, 2005 8:01 pm Post subject:

I have emerged gentoo-sources-2.6.12-r2 just now, turned off HPET timers in kernel, turned on 'Processor type & feats > PM timer' and succesfully got rid of that last 'lost ticks' message. I also modprobed ac and battery (ACPI-stuff, that caused the keyboard key-ignorance problem), and it's working for now. Yes :-D

I heard that VIA opened their specifications to the world, maybe this is the first fruit?

_________________
I thought what I'd do was, I'd pretend I was one of those deaf-mutes or should I?

paulisdead · Guru Joined: 10 Apr 2002 Posts: 510 Location: Seattle, WA

paulisdead · Guru Joined: 10 Apr 2002 Posts: 510 Location: Seattle, WA

It looks like I finally fixed it. I just compiled xorg without any clock drift for the first time I got this chip. I noticed my CPU temp went down, so it was probably running the chip itself faster too. I decided to try a lot of boot options at once, so I'm not sure what did it, but I'm on 2.6.13-rc1 and booted the kernel up with these options

c0un7d0wn · n00b Joined: 03 Oct 2004 Posts: 10

I have the same problem both in 32 bit and 64 bit. Using no_timer_check fixed the problem in 64 bit but not on 32 bit.

no_timer_check is available in >=2.6.12 only.

pjp · Administrator Joined: 16 Apr 2002 Posts: 20067

A shame there hasn't been a definitive problem/solution. I've had the lost ticks/instable time source/hogging interrupts message only 3 times prior to now, and I didn't even know it. The only reason I noticed now, is because my USB drive isn't recognized anymore. No hardware changes, and no kernel changes in quite a while (trying another now).

EDIT:

*points and laughs at the idiot*

Never underestimate Layer 1. The USB cable was plugged into a different PC

However, upgrading the did kill my sound.
_________________
Quis separabit? Quo animo?

blotto · kernel /2.6.12-r4 noapictimer root=/dev/hda3

Hi
I solved my timing issue and someone elses sound problems with this

gralves · Posted: Sun Sep 11, 2005 5:17 am Post subject:

For everyone whit AMD64x2 processors having this problem:

It appears to be a kernel bug:

http://bugzilla.kernel.org/show_bug.cgi?id=5105

tsaiming · n00b Joined: 23 May 2005 Posts: 27

Same problem on AMD64 (not dual) on aspire 1522Wlmi, with a little difference, this problem is often (maybe always) related to the keyboard issuse (lost or repeated keystroke).
I've tried all the solution here, nothing worked, I have tried suse 9.2 (32 bit) Live cd and the keyboard (then the "lost tick") stuffs don't appear!
Have someone installed gentoo 64 on the same machine?

ycUygB1 · Posted: Thu Sep 15, 2005 2:54 pm Post subject: boot with notsc

After reading the bugzilla thread, the consensus seems to be that booting with notsc will solve the problem.