View previous topic :: View next topic |
Author |
Message |
PrakashP Veteran
Joined: 27 Oct 2003 Posts: 1249 Location: C.C.A.A., Germania
|
Posted: Wed Aug 11, 2004 2:39 pm Post subject: |
|
|
Well, I already said that the tsc timer won't work. Here is the help text from pm_timer:
Quote: |
CONFIG_X86_PM_TIMER: │
│ │
│ The Power Management Timer is available on all ACPI-capable, │
│ in most cases even if ACPI is unusable or blacklisted. │
│ │
│ This timing source is not affected by powermanagement features │
│ like aggressive processor idling, throttling, frequency and/or │
│ voltage scaling, unlike the commonly used Time Stamp Counter │
│ (TSC) timing source. │
│ │
│ So, if you see messages like 'Losing too many ticks!' in the │
│ kernel logs, and/or you are using this on a notebook which │
│ does not yet have an HPET, you should say "Y" here.
|
So if pm_timer isn't available, hope that hpet works, otherwise you are fsked.
Something else: Have you ever tried to boot with kernel paramter "notsc"? how does this work? |
|
Back to top |
|
|
Acoc Apprentice
Joined: 06 Aug 2004 Posts: 189 Location: New Jersey, USA
|
Posted: Thu Aug 12, 2004 12:24 am Post subject: |
|
|
I was looking at the changelog of 2.6.8-rc4-mm1 and it says there is a tick fix. Could anyone tell me how to use this kernel as I am new to gentoo and how to enable the fix.
John _________________ _There were plenty of times in my century when I was gonna give up, but I never did. Never! Hey, are you even listening to me? Oh, I give up._ Futurama |
|
Back to top |
|
|
thaldyron Apprentice
Joined: 25 Sep 2002 Posts: 227 Location: On Earth
|
Posted: Thu Aug 12, 2004 8:12 am Post subject: |
|
|
PrakashKC wrote: | Well, I already said that the tsc timer won't work. Here is the help text from pm_timer:
Quote: |
CONFIG_X86_PM_TIMER
|
So if pm_timer isn't available, hope that hpet works, otherwise you are fsked.
Something else: Have you ever tried to boot with kernel paramter "notsc"? how does this work? |
I just tried "clock=notsc" but that didn't make any difference.
Where did you get the information about CONFIG_X86_PM_TIMER cos I'm not able to find it and it doesn't seem to be in ".config" either. |
|
Back to top |
|
|
PrakashP Veteran
Joined: 27 Oct 2003 Posts: 1249 Location: C.C.A.A., Germania
|
Posted: Thu Aug 12, 2004 8:15 am Post subject: |
|
|
Well the problem is it is !x86_64, so you won't see it. I am just on x86. As i said, ask at lkml, 'cos this behaviour doesn't seem right. |
|
Back to top |
|
|
themoebius Tux's lil' helper
Joined: 10 Mar 2004 Posts: 117
|
Posted: Thu Aug 12, 2004 9:27 am Post subject: |
|
|
if you google "many lost ticks" you only get 4 different pages, one of them being this thread and the others don't specify a distro, so I'm betting its a gentoo problem. But that wouldn't explain why people are having problems with the mm-sources. Does gentoo mess with the mm-sources?
I think I'll try vanilla and see if that does the trick. |
|
Back to top |
|
|
PrakashP Veteran
Joined: 27 Oct 2003 Posts: 1249 Location: C.C.A.A., Germania
|
Posted: Thu Aug 12, 2004 9:41 am Post subject: |
|
|
Well you don't get many threads as there are not many x86_64 distris out there... |
|
Back to top |
|
|
yxis n00b
Joined: 11 May 2004 Posts: 24 Location: your gnu-spot
|
Posted: Thu Aug 12, 2004 10:27 am Post subject: |
|
|
my lost ticks / hogging prob was solved by disabling generic IDE support in my kernel, and then compiling IDE support for VIA into the kernel, per someone's earlier post in this thread. worked like a charm. (mobo = abit KV8-max3). |
|
Back to top |
|
|
PrakashP Veteran
Joined: 27 Oct 2003 Posts: 1249 Location: C.C.A.A., Germania
|
Posted: Thu Aug 12, 2004 10:32 am Post subject: |
|
|
This method is no cure, just some starnge work-around. Ir may work for you, but I think it just makes it harder to reproduce. Root of all evil *is* the time source. |
|
Back to top |
|
|
Acoc Apprentice
Joined: 06 Aug 2004 Posts: 189 Location: New Jersey, USA
|
Posted: Thu Aug 12, 2004 11:47 pm Post subject: |
|
|
I have tried gentoo-dev 2.6.7-r11 r12 and r13. Only r13 doesn't produce these errors. _________________ _There were plenty of times in my century when I was gonna give up, but I never did. Never! Hey, are you even listening to me? Oh, I give up._ Futurama |
|
Back to top |
|
|
thaldyron Apprentice
Joined: 25 Sep 2002 Posts: 227 Location: On Earth
|
Posted: Fri Aug 13, 2004 9:38 am Post subject: |
|
|
Acoc wrote: | I have tried gentoo-dev 2.6.7-r11 r12 and r13. Only r13 doesn't produce these errors. |
I just tried gentoo-dev-sources-2.6.7-r14 and I still got the same problems.
Disabling the Generic IDE support didn't help either. |
|
Back to top |
|
|
Dephyler n00b
Joined: 01 Jun 2004 Posts: 31
|
Posted: Fri Aug 13, 2004 12:40 pm Post subject: |
|
|
I switch to the development sourcs(not the dev sources) and the ticks are gone. I have not touched the bios, and I use the same kernel config. but everything is fine! i recommend switching to the development-sources.
deph |
|
Back to top |
|
|
dangertools Apprentice
Joined: 29 Dec 2003 Posts: 278 Location: Saalfelden, Austria
|
Posted: Fri Aug 13, 2004 12:44 pm Post subject: |
|
|
Dephyler wrote: | I switch to the development sourcs(not the dev sources) and the ticks are gone. I have not touched the bios, and I use the same kernel config. but everything is fine! i recommend switching to the development-sources. |
As i said some posts above but nobody listened
dangertools |
|
Back to top |
|
|
thaldyron Apprentice
Joined: 25 Sep 2002 Posts: 227 Location: On Earth
|
Posted: Fri Aug 13, 2004 2:46 pm Post subject: |
|
|
dangertools wrote: | Dephyler wrote: | I switch to the development sourcs(not the dev sources) and the ticks are gone. I have not touched the bios, and I use the same kernel config. but everything is fine! i recommend switching to the development-sources. |
As i said some posts above but nobody listened
dangertools |
The development-sources work for me too, so this bug is somewhere in a patch that is contained by the gentoo-dev-sources >= 2.6.7 and mm-sources >= 2.6.8 |
|
Back to top |
|
|
Evangelion Veteran
Joined: 31 May 2002 Posts: 1087 Location: Helsinki, Finland
|
Posted: Fri Aug 13, 2004 7:55 pm Post subject: |
|
|
yxis wrote: | my lost ticks / hogging prob was solved by disabling generic IDE support in my kernel, and then compiling IDE support for VIA into the kernel, per someone's earlier post in this thread. worked like a charm. (mobo = abit KV8-max3). |
If I disable generic IDE on my system, it will not boot _________________ My tech-blog | My other blog |
|
Back to top |
|
|
krobar n00b
Joined: 05 Feb 2004 Posts: 63 Location: illinois
|
Posted: Sat Aug 14, 2004 5:19 am Post subject: |
|
|
i fixed the issue on my system by enabaling the via chipset support in the kernel, don't know if that helps anyone |
|
Back to top |
|
|
dangertools Apprentice
Joined: 29 Dec 2003 Posts: 278 Location: Saalfelden, Austria
|
Posted: Sat Aug 14, 2004 9:39 am Post subject: |
|
|
I have found a BIOS-upgrade for the aspire 1500 series on the panamerican acer website to version 1.06. this update is not available on the european acer website and it has to be installed via DOS. i have upgraded my bios a few minutes ago and the laptop still works as before but i had not the tick problem anymore since using development-sources but with everything enabled like acpi, ide etc. i have built a complete image with a dos bootdisk with cd-rom support. you just have to start the *.bat file on the cdrom and it should upgrade your bios. laptop has to be plugged into AC and the battery has to have more than 30% capacity. perhaps it resolves some of the problems with this laptop for non americans. the batchfile doesn't save the old bios. if something doesn't work don't blame on me but as i said my laptop works still fine.
Here the iso for burning a cdrom
http://members.inode.at/g.schild/DIV/BootDiskAspire1501V106.iso
dangertools |
|
Back to top |
|
|
bverheg n00b
Joined: 01 Sep 2002 Posts: 39 Location: Vlaanderen
|
Posted: Thu Aug 19, 2004 9:36 am Post subject: |
|
|
thaldyron wrote: |
The development-sources work for me too, so this bug is somewhere in a patch that is contained by the gentoo-dev-sources >= 2.6.7 and mm-sources >= 2.6.8 |
I compared the sources and what changed is simply that the "many lost ticks" log messages were added ('many' meaning >100).
I also found out you can pass the kernel a boot option "report_lost_ticks" that will cause logging of lost ticks in all cases.
Before, mm-sources 2.6.7 appeared ok for me, because no "many lost ticks" messages, but when booting it with this option on, I get the following message repeated every 5 seconds.
Code: | ... time.c: Lost 17 timer tick(s)! rip handle_IRQ_event+0x28/0x60) |
This message might give a clue as to the origin of the lost ticks. As I've reported before, on my hardware (Acer Aspire LMi1501) they only occur when I have the 'thermal' module loaded.
EDIT: AND when I have gkrellm2 running.
I expect that on other hardware the source of the lost ticks might be elsewhere.
dangertools: Is your BIOS-upgraded Aspire still working fine? Could you boot it with the report_lost_ticks kernel option and report whether you get any lost ticks messages? |
|
Back to top |
|
|
dangertools Apprentice
Joined: 29 Dec 2003 Posts: 278 Location: Saalfelden, Austria
|
Posted: Thu Aug 19, 2004 10:40 am Post subject: |
|
|
bverheg wrote: | dangertools: Is your BIOS-upgraded Aspire still working fine? Could you boot it with the report_lost_ticks kernel option and report whether you get any lost ticks messages? |
I will give the gentoo-dev-sources a try this evening or tomorrow with and without that bootparam. The lappi is still running fine, no problems yet andit ran a few days through.
dangertools |
|
Back to top |
|
|
thaldyron Apprentice
Joined: 25 Sep 2002 Posts: 227 Location: On Earth
|
Posted: Thu Aug 19, 2004 11:40 am Post subject: |
|
|
I just tried the development-sources-2.6.8.1 with the "report_lost_ticks parameter" and guess what:
Code: | time.c: Lost 7 timer tick(s)! rip ide_intr+0x1e0/0x290 |
The system didn't even boot (or it did and i didn't recognize the other output because this message was repeated all the time).
Next I tried the "notsc" option together with "report_lost_ticks" and the message disappeared. The system seemed to work with the "slight" disadvantage that it shut itself down after a minute because the temperature reached 100°, don't ask me why and I'm not very willing to reproduce that behaviour... |
|
Back to top |
|
|
bverheg n00b
Joined: 01 Sep 2002 Posts: 39 Location: Vlaanderen
|
Posted: Thu Aug 19, 2004 2:10 pm Post subject: |
|
|
I got kernel OOPSes with mm-sources-2.6.8.1.
I guess we will soon see a 2.6.8.2 kernel.
dangertools: thanks, I will try to upgrade the BIOS too. |
|
Back to top |
|
|
dangertools Apprentice
Joined: 29 Dec 2003 Posts: 278 Location: Saalfelden, Austria
|
Posted: Thu Aug 19, 2004 4:04 pm Post subject: |
|
|
i have bootet my 2.6.8.1 developement-sources kernel with report_many_lost ticks and now ask me what it reported to me. many_lost_ticks
but the message changed. it now has something with softirq and other IRQ things in it
looks like this:
Quote: | time.c: Lost 36 timer tick(s)! rip handle_IRQ_event+0x28/0x60)
time.c: Lost 1 timer tick(s)! rip __do_softirq+0x48/0xb0) |
it seems to reduce the messages when stopping gkrellm. it checks the temps and batt status via acpi. it also throws out a lot of ticks when an acpi-event occurs, e.g. inserting or removing the battery.
don't know what to do now.
dangertools |
|
Back to top |
|
|
thaldyron Apprentice
Joined: 25 Sep 2002 Posts: 227 Location: On Earth
|
Posted: Thu Aug 19, 2004 7:37 pm Post subject: |
|
|
dangertools wrote: |
don't know what to do now.
|
Something that still remains a mystery to me is the fact that the 32bit mode (with exactly the same configuration) seems to work flawlessly without producing strange tick messages.
Hence I'm highly convinced that this must a bug somewhere in the amd64 specific kernel part and no bios update etc will help us... |
|
Back to top |
|
|
PrakashP Veteran
Joined: 27 Oct 2003 Posts: 1249 Location: C.C.A.A., Germania
|
Posted: Thu Aug 19, 2004 7:53 pm Post subject: |
|
|
As long as you don't report it at lkml, no one will take care of it... I don#t have an athlon64, so I cannot test. |
|
Back to top |
|
|
ub818 n00b
Joined: 19 Feb 2004 Posts: 66
|
Posted: Wed Aug 25, 2004 8:22 pm Post subject: |
|
|
I keep getting :
Code: | Your time source seems to be instable or some driver is hogging interupts
Losing some ticks... checking if CPU frequency change |
I've re-compiled my kernel many different times trying what worked for some. Nothing seems to work though. This all happened when I started things off with 2.6.7-gentoo-r11.
I've been playing with this error for a few weeks now. |
|
Back to top |
|
|
db9052 n00b
Joined: 11 Jun 2004 Posts: 5
|
Posted: Thu Aug 26, 2004 9:13 am Post subject: |
|
|
dangertools wrote: | I have found a BIOS-upgrade for the aspire 1500 series on the panamerican acer website to version 1.06. this update is not available on the european acer website and it has to be installed via DOS. i have upgraded my bios a few minutes ago and the laptop still works as before but i had not the tick problem anymore since using development-sources but with everything enabled like acpi, ide etc.
dangertools |
I've just installed this BIOS update on my Aspire 1501LCe, however with gentoo-dev-sources I'm still getting lost ticks! |
|
Back to top |
|
|
|