[opensuse-kernel] FastTSC calibration failed
Hi, I get this message not always, but often during boot. Here is what #cat /var/log/messages | grep TSC shows me. From november 7th, this message exists. (since update to this kernel: 3.7.0-rc4-4-desktop i686) Nov 11 11:24:40 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration failed Nov 11 11:24:40 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2 EeePc-Rob-SFN9:~ # cat /proc/cpuinfo | grep tsc flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx constant_tsc arch_perfmon pebs bts aperfmperf pni dtes64 monitor ds_cpl est tm2 ssse3 xtpr pdcm movbe lahf_lm dtherm flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx constant_tsc arch_perfmon pebs bts aperfmperf pni dtes64 monitor ds_cpl est tm2 ssse3 xtpr pdcm movbe lahf_lm dtherm Is there anything i can/have to do about it? Kind regards. -- Have a nice day, Oddball. OS: Linux 3.7.0-rc4-4-desktop i686 Huidige gebruiker: oddball@EeePc-Rob-SFN9 Systeem: openSUSE 12.2 (i586) KDE: 4.9.3 "release 520" -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
Am 11.11.2012 11:46, schrieb Oddball:
Hi,
I get this message not always, but often during boot. Here is what #cat /var/log/messages | grep TSC shows me. From november 7th, this message exists. (since update to this kernel: 3.7.0-rc4-4-desktop i686)
Nov 11 11:24:40 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration failed Nov 11 11:24:40 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2
That's just a diagnostic message.
Is there anything i can/have to do about it?
AFAIK, you would need to put a different CPU into your machine to get rid of this message. -- Stefan Seyfried "If your lighter runs out of fluid or flint and stops making fire, and you can't be bothered to figure out about lighter fluid or flint, that is not Zippo's fault." -- bkw -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
Op 11-11-12 13:31, Stefan Seyfried schreef:
Am 11.11.2012 11:46, schrieb Oddball:
Hi,
I get this message not always, but often during boot. Here is what #cat /var/log/messages | grep TSC shows me. From november 7th, this message exists. (since update to this kernel: 3.7.0-rc4-4-desktop i686)
Nov 11 11:24:40 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration failed Nov 11 11:24:40 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2 That's just a diagnostic message.
Is there anything i can/have to do about it? AFAIK, you would need to put a different CPU into your machine to get rid of this message.
Yes, well, i thought more like putting the 3.7.0-rc3-2-desktop i686 back. Because it happened only from 7 nov, the day this kernel was installed. And also i thought it was appreciated to give some feedback about these daily kernels. Unfortunately it is not listed anymore in HEAD. Alternative is now: 3.4.18-1.1-i686. -- Have a nice day, Oddball. OS: Linux 3.7.0-rc4-4-desktop i686 Huidige gebruiker: oddball@EeePc-Rob-SFN9 Systeem: openSUSE 12.2 (i586) KDE: 4.9.3 "release 520" -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
On 11/11/2012 06:31 AM, Stefan Seyfried wrote:
Am 11.11.2012 11:46, schrieb Oddball:
Hi,
I get this message not always, but often during boot. Here is what #cat /var/log/messages | grep TSC shows me. From november 7th, this message exists. (since update to this kernel: 3.7.0-rc4-4-desktop i686)
Nov 11 11:24:40 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration failed Nov 11 11:24:40 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2
That's just a diagnostic message.
Is there anything i can/have to do about it?
AFAIK, you would need to put a different CPU into your machine to get rid of this message.
Some of us see the following: tsc: Fast TSC calibration using PIT tsc: Detected 2000.279 MHz processor tsc: Marking TSC unstable due to TSCs unsynchronized Calibrating delay loop (skipped), value calculated using timer frequency.. 4000.55 BogoMIPS (lpj=8001116) This message arises on an HP laptop with an AMD dual-core CPU running at 2 GHz. As it gets the CPU frequency correct, I have ignored the message for years. Larry -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
On Sunday 11 November 2012 13:31:18 Stefan Seyfried wrote:
Am 11.11.2012 11:46, schrieb Oddball:
Hi,
I get this message not always, but often during boot. Here is what #cat /var/log/messages | grep TSC shows me. From november 7th, this message exists. (since update to this kernel: 3.7.0-rc4-4-desktop i686)
Nov 11 11:24:40 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration failed Nov 11 11:24:40 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2
That's just a diagnostic message.
But why does it show up _sometimes_ ? Either the TSC is stable or it is not. Regards Oliver -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
On 11/11/12 13:04, Larry Finger wrote:
On 11/11/2012 06:31 AM, Stefan Seyfried wrote:
Am 11.11.2012 11:46, schrieb Oddball:
Hi,
I get this message not always, but often during boot. Here is what #cat /var/log/messages | grep TSC shows me. From november 7th, this message exists. (since update to this kernel: 3.7.0-rc4-4-desktop i686)
Nov 11 11:24:40 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration failed Nov 11 11:24:40 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2
That's just a diagnostic message.
Is there anything i can/have to do about it?
AFAIK, you would need to put a different CPU into your machine to get rid of this message.
Some of us see the following:
tsc: Fast TSC calibration using PIT tsc: Detected 2000.279 MHz processor tsc: Marking TSC unstable due to TSCs unsynchronized Calibrating delay loop (skipped), value calculated using timer frequency.. 4000.55 BogoMIPS (lpj=8001116)
Use hpet.. boot with clocksource=hpet . -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
Am 11.11.2012 18:30, schrieb Oliver Neukum:
On Sunday 11 November 2012 13:31:18 Stefan Seyfried wrote:
That's just a diagnostic message.
But why does it show up _sometimes_ ? Either the TSC is stable or it is not.
Well, YOU are the kernel hacker, not me :-) I'd guess that the old kernel for some reason did not use all available C-States or such, and the newer one has support for more of the chipset's features. -- Stefan Seyfried "If your lighter runs out of fluid or flint and stops making fire, and you can't be bothered to figure out about lighter fluid or flint, that is not Zippo's fault." -- bkw -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
On Monday 12 November 2012 06:36:44 Stefan Seyfried wrote:
Am 11.11.2012 18:30, schrieb Oliver Neukum:
On Sunday 11 November 2012 13:31:18 Stefan Seyfried wrote:
That's just a diagnostic message.
But why does it show up _sometimes_ ? Either the TSC is stable or it is not.
Well, YOU are the kernel hacker, not me :-)
I'd guess that the old kernel for some reason did not use all available C-States or such, and the newer one has support for more of the chipset's features.
Yes, it can be explained that a kernel version makes a difference. But I read the bug report to say that the problem shows up sometimes with the same kernel version and sometimes not. That would be problematic. Oddball, could you clarify? Regards Oliver -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
Op 12-11-12 08:04, Oliver Neukum schreef:
Am 11.11.2012 18:30, schrieb Oliver Neukum:
On Sunday 11 November 2012 13:31:18 Stefan Seyfried wrote:
That's just a diagnostic message. But why does it show up _sometimes_ ? Either the TSC is stable or it is not. Well, YOU are the kernel hacker, not me :-)
I'd guess that the old kernel for some reason did not use all available C-States or such, and the newer one has support for more of the chipset's features. Yes, it can be explained that a kernel version makes a difference. But I read the bug report to say that the problem shows up sometimes with the same kernel version and sometimes not. That would be
On Monday 12 November 2012 06:36:44 Stefan Seyfried wrote: problematic.
Oddball, could you clarify?
Regards Oliver
I can not say more. Sometimes the message shows. And sometimes it does not. I installed kernel 3.4.18-1-desktop i686 yesterday. This one does not show the message. Does it make sense to reinstall kernel 3.7.0-rc4-4-desktop i686? -- Have a nice day, Oddball. OS: Linux 3.4.18-1-desktop i686 Huidige gebruiker: oddball@EeePc-Rob-SFN9 Systeem: openSUSE 12.2 (i586) KDE: 4.9.3 "release 520" -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
On Monday 12 November 2012 09:12:18 Oddball wrote:
I can not say more. Sometimes the message shows. And sometimes it does not.
Was there a kernel version which sometimes showed the message and sometimes did not show the message? Regards Oliver -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
On Monday 12 November 2012 09:12:18 Oddball wrote:
Op 12-11-12 08:04, Oliver Neukum schreef:
Am 11.11.2012 18:30, schrieb Oliver Neukum:
On Sunday 11 November 2012 13:31:18 Stefan Seyfried wrote:
That's just a diagnostic message. But why does it show up _sometimes_ ? Either the TSC is stable or it is not. Well, YOU are the kernel hacker, not me :-)
I'd guess that the old kernel for some reason did not use all available C-States or such, and the newer one has support for more of the chipset's features. Yes, it can be explained that a kernel version makes a difference. But I read the bug report to say that the problem shows up sometimes with the same kernel version and sometimes not. That would be
On Monday 12 November 2012 06:36:44 Stefan Seyfried wrote: problematic.
Oddball, could you clarify?
Regards Oliver
I can not say more. Sometimes the message shows. And sometimes it does not. I installed kernel 3.4.18-1-desktop i686 yesterday. This one does not show the message.
Does it make sense to reinstall kernel 3.7.0-rc4-4-desktop i686?
Hi Jiri, this phenomenon worries me a bit. It seems like the kernel under some circumstances fails to detect instability of the TSC. I am afraid that would do bad things to time keeping. What needs to be done? Regards Oliver -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On 2012-11-12 15:34, Oliver Neukum wrote:
Hi Jiri,
this phenomenon worries me a bit. It seems like the kernel under some circumstances fails to detect instability of the TSC. I am afraid that would do bad things to time keeping. What needs to be done?
I have seen this problem years ago; I had bad problems with the clock and related, that were solved on a later release and much better when I changed main machine. There were several bugzillas on the issues. - -- Cheers / Saludos, Carlos E. R. (from 11.4 x86_64 "Celadon" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.16 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://www.enigmail.net/ iF4EAREIAAYFAlChTe8ACgkQja8UbcUWM1zzVAEAm96Y+NZ8btTB8BtH1Wlm1q9+ Usw1L5fzE7UZNJC7ZdAA/3IGRYxUG8FA6voHipFYWkvZLZQTXeqFuGGXNiPIPdTU =v7o5 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
Hi, After installing the new daily kernel, the same message showed up. The previous week, there was no message. But looking at the output here, is that the same was happening all the time, tsc gave just no message at boot-time about it.
EeePc-Rob-SFN9:~ # cat /var/log/messages | grep TSC Nov 7 19:35:14 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration using PIT Nov 7 19:35:14 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 7 19:48:18 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration using PIT Nov 7 19:48:18 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 7 19:52:39 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration using PIT Nov 7 19:52:39 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 7 20:53:25 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration using PIT Nov 7 20:53:25 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 7 21:02:50 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration using PIT Nov 7 21:02:50 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 7 21:18:33 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration failed Nov 7 21:18:33 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 7 21:21:35 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration failed Nov 7 21:21:35 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 7 21:25:34 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration using PIT Nov 7 21:25:34 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 7 21:37:07 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration failed Nov 7 21:37:07 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 8 22:47:11 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration using PIT Nov 8 22:47:11 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 9 07:48:36 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration using PIT Nov 9 07:48:36 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 10 10:42:19 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration using PIT Nov 10 10:42:19 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 10 11:35:24 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration failed Nov 10 11:35:24 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 11 11:20:53 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration failed Nov 11 11:20:53 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 11 11:24:40 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration failed Nov 11 11:24:40 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 11 15:38:52 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration using PIT Nov 11 15:38:52 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 11 20:45:27 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration failed Nov 11 20:45:27 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2 Replacing the kernel by the default 3.4 Nov 11 20:55:05 EeePc-Rob-SFN9 kernel: Fast TSC calibration using PIT Nov 11 20:55:05 EeePc-Rob-SFN9 kernel: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 12 07:13:51 EeePc-Rob-SFN9 kernel: Fast TSC calibration using PIT Nov 12 07:13:51 EeePc-Rob-SFN9 kernel: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 12 08:44:00 EeePc-Rob-SFN9 kernel: Fast TSC calibration using PIT Nov 12 08:44:00 EeePc-Rob-SFN9 kernel: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 12 16:31:47 EeePc-Rob-SFN9 kernel: Fast TSC calibration using PIT Nov 12 16:31:47 EeePc-Rob-SFN9 kernel: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 12 19:02:26 EeePc-Rob-SFN9 kernel: Fast TSC calibration using PIT Nov 12 19:02:26 EeePc-Rob-SFN9 kernel: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 12 21:35:09 EeePc-Rob-SFN9 kernel: Fast TSC calibration using PIT Nov 12 21:35:09 EeePc-Rob-SFN9 kernel: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 12 22:26:52 EeePc-Rob-SFN9 kernel: Fast TSC calibration using PIT Nov 12 22:26:52 EeePc-Rob-SFN9 kernel: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 13 07:25:48 EeePc-Rob-SFN9 kernel: Fast TSC calibration using PIT Nov 13 07:25:48 EeePc-Rob-SFN9 kernel: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 14 07:47:45 EeePc-Rob-SFN9 kernel: Fast TSC calibration using PIT Nov 14 07:47:45 EeePc-Rob-SFN9 kernel: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 15 07:17:16 EeePc-Rob-SFN9 kernel: Fast TSC calibration using PIT Nov 15 07:17:16 EeePc-Rob-SFN9 kernel: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 15 13:33:51 EeePc-Rob-SFN9 kernel: Fast TSC calibration using PIT Nov 15 13:33:51 EeePc-Rob-SFN9 kernel: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 15 14:38:13 EeePc-Rob-SFN9 kernel: Fast TSC calibration using PIT Nov 15 14:38:13 EeePc-Rob-SFN9 kernel: Marking TSC unstable due to TSC halts in idle states deeper than C2 Nov 16 11:40:43 EeePc-Rob-SFN9 kernel: Fast TSC calibration using PIT Nov 16 11:40:43 EeePc-Rob-SFN9 kernel: Marking TSC unstable due to TSC halts in idle states deeper than C2 Replacing default 3.4 with daily: 3.7.0-rc5-2-desktop i686
Nov 16 16:35:37 EeePc-Rob-SFN9 kernel: tsc: Fast TSC calibration failed Nov 16 16:35:37 EeePc-Rob-SFN9 kernel: tsc: Marking TSC unstable due to TSC halts in idle states deeper than C2 EeePc-Rob-SFN9:~ #
If you look at the dates, you can see the previous Daily, the default 3.4, and the new daily: 3.7.0-rc5-2-desktop i686 Today is boot, and replacing kernel reboot time. So this is the "what doesn't know, doesn't hurt" scenario i guess. Anyway it makes no sense replacing the kernel by a previous one, just to 'not' see an error message, now does it? so logging is there, but no error message... (replacing the cpu is not an option in an eee-pc ..lol..) EeePc-Rob-SFN9:~ # cat /proc/cpuinfo | grep tsc flags : fpu vme de pse *tsc* msr pae mce cx8 apic sep mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx *constant_tsc* arch_perfmon pebs bts aperfmperf pni dtes64 monitor ds_cpl est tm2 ssse3 xtpr pdcm movbe lahf_lm dtherm -- Have a nice day, Oddball. OS: Linux 3.7.0-rc5-2-desktop i686 Huidige gebruiker: oddball@EeePc-Rob-SFN9 Systeem: openSUSE 12.2 (i586) KDE: 4.9.3 "release 520" -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
participants (6)
-
Carlos E. R. -
Cristian Rodríguez -
Larry Finger -
Oddball -
Oliver Neukum -
Stefan Seyfried