time stamped printks

time stamped printks

[Hinko Kočevar]

Hi all,

I'm building latest GIT tree kernel for CRIS architecture. Right now I'm seeing 
some strange values with printk stamps - at the start they are initialized to 0, 
but quickly change to 42949372 - 0x028F5AFC.

Any thoughts to where to look for fix?


[    0.000000] Linux version 2.6.26-rc1-cris-00180-g21a1022 
(hinkok@alala.kranj.cetrtapot.si) (gcc version 3.2.1 Axis release R64/1.64) #12 
Thu May 8 10:09:47 CEST 2008
[    0.000000] console [ttyS0] enabled
[    0.000000] Setting up paging and the MMU.
[    0.000000] Linux/CRIS port on ETRAX 100LX (c) 2001 Axis Communications AB
[42949372.961619] Built 1 zonelists in Zone order, mobility grouping on.  Total 
pages: 2040
[42949372.961714] Kernel command line: root=/dev/nfs rw ip=dhcp init=/linuxrc 
console=ttyS0,38400
[42949372.963114] PID hash table entries: 64 (order: 6, 256 bytes)
[42949372.963450] ds1302: RTC found.
[42949372.963557] ds1302: SDA, SCL, RST on PB0, PB1, GENIO0
[42949372.963744] rtc_time	: 08:23:41
[42949372.963768] rtc_date	: 2008-05-08
[42949372.963744] Dentry cache hash table entries: 2048 (order: 0, 8192 bytes)

Best regards,
Hinko

-- 
ČETRTA POT, d.o.o., Kranj
Planina 3
4000 Kranj
Slovenia, Europe
Tel. +386 (0) 4 280 66 03
E-mail: hinko.kocevar@cetrtapot.si
Http: www.cetrtapot.si

Re: time stamped printks

[David Miller]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: Text/Plain; charset=utf-8, Size: 900 bytes --]

From: Hinko Kočevar <hinko.kocevar@cetrtapot.si>
Date: Thu, 08 May 2008 10:26:01 +0200

> Hi all,
> 
> I'm building latest GIT tree kernel for CRIS architecture. Right now I'm seeing 
> some strange values with printk stamps - at the start they are initialized to 0, 
> but quickly change to 42949372 - 0x028F5AFC.
> 
> Any thoughts to where to look for fix?

Very early in the boot, the timestamp counter fetching infrastructure
isn't setup yet.  So zero is reported for the timestamp.

So early on in the boot, it is often not even possible to setup
such things yet.  Because infrastructure they need for probing
etc. isn't even available yet.

This happens on basically every single Linux platform, there
isn't really anything to fix.

ÿôèº{.nÇ+‰·Ÿ®‰­†+%ŠËÿ±éݶ\x17¥Šwÿº{.nÇ+‰·¥Š{±þG«éÿŠ{ayº\x1dʇڙë,j\a­¢f£¢·hšïêÿ‘êçz_è®\x03(­éšŽŠÝ¢j"ú\x1a¶^[m§ÿÿ¾\a«þG«éÿ¢¸?™¨è­Ú&£ø§~á¶iO•æ¬z·švØ^\x14\x04\x1a¶^[m§ÿÿÃ\fÿ¶ìÿ¢¸?–I¥

Re: time stamped printks

[Gabriel C]

David Miller wrote:
> From: Hinko Kočevar <hinko.kocevar@cetrtapot.si>
> Date: Thu, 08 May 2008 10:26:01 +0200
> 
>> Hi all,
>>
>> I'm building latest GIT tree kernel for CRIS architecture. Right now I'm seeing 
>> some strange values with printk stamps - at the start they are initialized to 0, 
>> but quickly change to 42949372 - 0x028F5AFC.
>>
>> Any thoughts to where to look for fix?
> 
> Very early in the boot, the timestamp counter fetching infrastructure
> isn't setup yet.  So zero is reported for the timestamp.
> 
> So early on in the boot, it is often not even possible to setup
> such things yet.  Because infrastructure they need for probing
> etc. isn't even available yet.
> 
> This happens on basically every single Linux platform, there
> isn't really anything to fix.

For the early boot 0's it is normal but not for the following strange values :)

I got the same problem on my 64bit box with latest git head and bisected it down to 3e51f33fcc7f55e6df25d15b55ed10c8b4da84cd

http://lkml.org/lkml/2008/5/8/146


Regards,

Gabriel C

Re: time stamped printks

[Hinko Kočevar]

Gabriel C wrote:
> David Miller wrote:
>> From: Hinko Kočevar <hinko.kocevar@cetrtapot.si>
>> Date: Thu, 08 May 2008 10:26:01 +0200
>>
>>> Hi all,
>>>
>>> I'm building latest GIT tree kernel for CRIS architecture. Right now I'm seeing 
>>> some strange values with printk stamps - at the start they are initialized to 0, 
>>> but quickly change to 42949372 - 0x028F5AFC.
>>>
>>> Any thoughts to where to look for fix?
>> Very early in the boot, the timestamp counter fetching infrastructure
>> isn't setup yet.  So zero is reported for the timestamp.
>>
>> So early on in the boot, it is often not even possible to setup
>> such things yet.  Because infrastructure they need for probing
>> etc. isn't even available yet.
>>
>> This happens on basically every single Linux platform, there
>> isn't really anything to fix.

Hmm, I was under the impression that this was not an issue (at least not in 2.6.25).

> 
> For the early boot 0's it is normal but not for the following strange values :)

I agree.

> 
> I got the same problem on my 64bit box with latest git head and bisected it down to 3e51f33fcc7f55e6df25d15b55ed10c8b4da84cd
> 
> http://lkml.org/lkml/2008/5/8/146

Well, cris architecture does not have CONFIG_HAVE_UNSTABLE_SCHED_CLOCK. I guess 
sched rework broker something since it was touched by the above commit, or..:

# CONFIG_CGROUPS is not set
CONFIG_GROUP_SCHED=y
CONFIG_FAIR_GROUP_SCHED=y
# CONFIG_RT_GROUP_SCHED is not set
CONFIG_USER_SCHED=y
# CONFIG_CGROUP_SCHED is not set



 From the 3e51f33fcc7f55e6df25d15b55ed10c8b4da84cd commit Changelog:
  - architectures that have an 'imperfect' hardware clock can set
    CONFIG_HAVE_UNSTABLE_SCHED_CLOCK

How can one do that? Just add it to the .config?

Regards,
Hinko


-- 
ČETRTA POT, d.o.o., Kranj
Planina 3
4000 Kranj
Slovenia, Europe
Tel. +386 (0) 4 280 66 03
E-mail: hinko.kocevar@cetrtapot.si
Http: www.cetrtapot.si