Uptime counter

Uptime counter

[Markus Hästbacka]

Hi list,
I wonder does any kernel branch have a uptime counter that doesn't stop
counting at 497 days? Or is a patch needed for the job to
2.{0,2,4,6} kernel?

	Markus

Re: Uptime counter

[Matti Aarnio]

On Sun, Feb 01, 2004 at 10:41:41PM +0200, Markus Hästbacka wrote:
> Hi list,
> I wonder does any kernel branch have a uptime counter that doesn't stop
> counting at 497 days? Or is a patch needed for the job to
> 2.{0,2,4,6} kernel?

Any 64 bit machine since day 1,  but also 2.6 at i386 does work.

> 	Markus

/Matti Aarnio

Re: Uptime counter

[Måns Rullgård]

Markus Hästbacka <midian@ihme.org> writes:

> I wonder does any kernel branch have a uptime counter that doesn't
> stop counting at 497 days? Or is a patch needed for the job to
> 2.{0,2,4,6} kernel?

Why don't you try it?

Jokes aside, why do you care?

-- 
Måns Rullgård
mru@kth.se

Re: Uptime counter

[Christian Borntraeger]

Markus Hästbacka wrote:
> Hi list,
> I wonder does any kernel branch have a uptime counter that doesn't stop
> counting at 497 days? Or is a patch needed for the job to
> 2.{0,2,4,6} kernel?

In 2.6 there is no 497 days limit, as jiffies are now 64 bit. 

By the way: Having a machine with more than 497 days of uptime normally 
shows a serios lack of security awareness..

cheers 

Christian

Re: Uptime counter

[Markus Hästbacka]

On Sun, 1 Feb 2004, Christian Borntraeger wrote:
>
> In 2.6 there is no 497 days limit, as jiffies are now 64 bit.
>
Ok, I just would be intrested in a patch for 2.0 and 2.4 to get these
jiffies to 64 bit.
> By the way: Having a machine with more than 497 days of uptime normally
> shows a serios lack of security awareness..
>
I know, but running a 2.0.x machine with that kind of uptime isn't really
that bad, thought if the machine has alot of accounts it wouldn't be that
great idea.

But anyway, thanks for the information!

	Markus

Re: Uptime counter

[Matthew Reppert]

[-- Attachment #1: Type: text/plain, Size: 1109 bytes --]

On Sun, 2004-02-01 at 16:07, Ludootje wrote:
> On Sun, 2004-02-01 at 20:51, Matti Aarnio wrote:
> > On Sun, Feb 01, 2004 at 10:41:41PM +0200, Markus Hästbacka wrote:
> > > Hi list,
> > > I wonder does any kernel branch have a uptime counter that doesn't stop
> > > counting at 497 days? Or is a patch needed for the job to
> > > 2.{0,2,4,6} kernel?
> > 
> > Any 64 bit machine since day 1,  but also 2.6 at i386 does work.
> > 
> > > 	Markus
> > 
> > /Matti Aarnio
> 
> It's the first time I hear about the uptime being resetted after 497 days,
> why is this? Is this a mistake or are their reasons for it?

On 32-bit architectures, the uptime counter is only 32 bits wide. Each
"tick" of the counter is worth 1/HZ seconds (IIRC). So, you can get the
number of seconds this will hold with simple math (2^32 * 1/HZ, HZ
being 100 on i386). This is about 497.1 days.

Of course, on 64-bit architectures, the counter will hold 4 billion
times that, which is about as long as the Earth has existed. Apparently
2.6 has come up with a way to deal with this on 32-bit architectures.

Matt

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: Uptime counter

[David Weinehall]

On Sun, Feb 01, 2004 at 11:16:56PM +0200, Markus Hästbacka wrote:
> On Sun, 1 Feb 2004, Christian Borntraeger wrote:
> >
> > In 2.6 there is no 497 days limit, as jiffies are now 64 bit.
> >
> Ok, I just would be intrested in a patch for 2.0 and 2.4 to get these
> jiffies to 64 bit.
> > By the way: Having a machine with more than 497 days of uptime normally
> > shows a serios lack of security awareness..
> >
> I know, but running a 2.0.x machine with that kind of uptime isn't really
> that bad, thought if the machine has alot of accounts it wouldn't be that
> great idea.
 
Well, you're soon going to reboot to install the upcoming 2.0.40, right?
And I promise to release 2.0.41 before you've had 497 days of uptime
with that one... :-)

> But anyway, thanks for the information!


Regards: David Weinehall
-- 
 /) David Weinehall <tao@acc.umu.se> /) Northern lights wander      (\
//  Maintainer of the v2.0 kernel   //  Dance across the winter sky //
\)  http://www.acc.umu.se/~tao/    (/   Full colour fire           (/

Re: Uptime counter

[Tomasz Torcz]

On Sun, Feb 01, 2004 at 10:02:06PM +0100, Christian Borntraeger wrote:
> By the way: Having a machine with more than 497 days of uptime normally 
> shows a serios lack of security awareness..

 Most kernel problems can be solved with simple modules like this:
 http://toxygen.net/hotfixes/mremap.c 
 Userspace upgrades obviously do not require reboots.

-- 
Tomasz Torcz                                                       72->|   80->|
zdzichu@irc.-nie.spam-.pl                                          72->|   80->|
|> Playing:  ...

Re: Uptime counter

[Markus Hästbacka]

On Sun, 1 Feb 2004, David Weinehall wrote:
> On Sun, Feb 01, 2004 at 11:16:56PM +0200, Markus Hästbacka wrote:
> > On Sun, 1 Feb 2004, Christian Borntraeger wrote:
> > >
> > > In 2.6 there is no 497 days limit, as jiffies are now 64 bit.
> > >
> > Ok, I just would be intrested in a patch for 2.0 and 2.4 to get these
> > jiffies to 64 bit.
> > > By the way: Having a machine with more than 497 days of uptime normally
> > > shows a serios lack of security awareness..
> > >
> > I know, but running a 2.0.x machine with that kind of uptime isn't really
> > that bad, thought if the machine has alot of accounts it wouldn't be that
> > great idea.
>
> Well, you're soon going to reboot to install the upcoming 2.0.40, right?
> And I promise to release 2.0.41 before you've had 497 days of uptime
> with that one... :-)
>
Of course :)
But when you'll stop releasing stuff, then it's time to see that :)
> > But anyway, thanks for the information!
>
>
> Regards: David Weinehall
> --
>  /) David Weinehall <tao@acc.umu.se> /) Northern lights wander      (\
> //  Maintainer of the v2.0 kernel   //  Dance across the winter sky //
> \)  http://www.acc.umu.se/~tao/    (/   Full colour fire           (/
>
Btw, when is it coming? :-)

	Markus

Re: Uptime counter

[David Weinehall]

On Sun, Feb 01, 2004 at 11:34:39PM +0200, Markus Hästbacka wrote:
> On Sun, 1 Feb 2004, David Weinehall wrote:
[snip]
> > Well, you're soon going to reboot to install the upcoming 2.0.40, right?
> > And I promise to release 2.0.41 before you've had 497 days of uptime
> > with that one... :-)
> >
> Of course :)
> But when you'll stop releasing stuff, then it's time to see that :)

When I stop releasing stuff, it's time to upgrade to the 3.x kernel...

[snip]

> Btw, when is it coming? :-)

When I've got enough feedback that 2.0.40-rc8 is working...  Soon, very
soon.  Unless, of course, you were talking about 2.0.41, which is quite
some time away... :-)


Regards: David Weinehall
-- 
 /) David Weinehall <tao@acc.umu.se> /) Northern lights wander      (\
//  Maintainer of the v2.0 kernel   //  Dance across the winter sky //
\)  http://www.acc.umu.se/~tao/    (/   Full colour fire           (/

Re: Uptime counter

[Guillermo Menguez Alvarez]

> Ok, I just would be intrested in a patch for 2.0 and 2.4 to get these
> jiffies to 64 bit.

Maybe it's what you are looking for:

http://www.plumlocosoft.com/kernel/

It's the -ck patchset (now maintained by Eric Hustvedt for the 2.4.x
series), they have a patch to add 64 bit jiffies ot it seems so.

Regards,
Guillermo.

-- 
Usuario Linux #212057 - Maquinas Linux #98894, #130864 y #168988
Proyecto LONIX: http://lonix.sourceforge.net
Lagrimas en la Lluvia: http://www.lagrimasenlalluvia.com

Re: Uptime counter

[Markus Hästbacka]

On Sun, 1 Feb 2004, David Weinehall wrote:
> On Sun, Feb 01, 2004 at 11:34:39PM +0200, Markus Hästbacka wrote:
> > On Sun, 1 Feb 2004, David Weinehall wrote:
> [snip]
> > > Well, you're soon going to reboot to install the upcoming 2.0.40, right?
> > > And I promise to release 2.0.41 before you've had 497 days of uptime
> > > with that one... :-)
> > >
> > Of course :)
> > But when you'll stop releasing stuff, then it's time to see that :)
>
> When I stop releasing stuff, it's time to upgrade to the 3.x kernel...
>
Never! This boxen is sticking with 2.0 :-)
> [snip]
>
> > Btw, when is it coming? :-)
>
> When I've got enough feedback that 2.0.40-rc8 is working...  Soon, very
> soon.  Unless, of course, you were talking about 2.0.41, which is quite
> some time away... :-)
>
I see no problems with it, maybe I should do some tests or something? :)

	Markus

Re: Uptime counter

[Ludootje]

On Sun, 2004-02-01 at 20:51, Matti Aarnio wrote:
> On Sun, Feb 01, 2004 at 10:41:41PM +0200, Markus Hästbacka wrote:
> > Hi list,
> > I wonder does any kernel branch have a uptime counter that doesn't stop
> > counting at 497 days? Or is a patch needed for the job to
> > 2.{0,2,4,6} kernel?
> 
> Any 64 bit machine since day 1,  but also 2.6 at i386 does work.
> 
> > 	Markus
> 
> /Matti Aarnio

It's the first time I hear about the uptime being resetted after 497 days,
why is this? Is this a mistake or are their reasons for it?

Thanks,
Ludootje

Re: Uptime counter

[Ludootje]

On Sun, 2004-02-01 at 21:02, Christian 
> By the way: Having a machine with more than 497 days of
> uptime normally shows a serios lack of security awareness..
> 
> cheers 
> 
> Christian

If there hasn't been a remote root exploit in the kernel for 497
days I don't see what the problem is...

Ludootje

Re: Uptime counter

[David Weinehall]

On Mon, Feb 02, 2004 at 12:07:40AM +0200, Markus Hästbacka wrote:
> On Sun, 1 Feb 2004, David Weinehall wrote:
> > On Sun, Feb 01, 2004 at 11:34:39PM +0200, Markus Hästbacka wrote:
> > > On Sun, 1 Feb 2004, David Weinehall wrote:
> > [snip]
> > > > Well, you're soon going to reboot to install the upcoming 2.0.40, right?
> > > > And I promise to release 2.0.41 before you've had 497 days of uptime
> > > > with that one... :-)
> > > >
> > > Of course :)
> > > But when you'll stop releasing stuff, then it's time to see that :)
> >
> > When I stop releasing stuff, it's time to upgrade to the 3.x kernel...
> >
> Never! This boxen is sticking with 2.0 :-)
> > [snip]
> >
> > > Btw, when is it coming? :-)
> >
> > When I've got enough feedback that 2.0.40-rc8 is working...  Soon, very
> > soon.  Unless, of course, you were talking about 2.0.41, which is quite
> > some time away... :-)
> >
> I see no problems with it, maybe I should do some tests or something? :)

Well, since most of the changes in the latest kernels involve
networking, trying it with various different network-adapters would be
interesting, and stress-testing the network-code in general.

If you have the hardware or a really good confidence, a recent
2.2-kernel to compare with and sufficient knowledge of C, have a look at
the network-drivers for a2065 and ariadne, both of which lack the
padding-fixes the other adapters have, since I didn't want to touch that
mess...


Regards: David Weinehall
-- 
 /) David Weinehall <tao@acc.umu.se> /) Northern lights wander      (\
//  Maintainer of the v2.0 kernel   //  Dance across the winter sky //
\)  http://www.acc.umu.se/~tao/    (/   Full colour fire           (/

Re: Uptime counter

[Paul Jakma]

On Sun, 1 Feb 2004, Markus Hästbacka wrote:

> I know, but running a 2.0.x machine with that kind of uptime isn't
> really that bad, thought if the machine has alot of accounts it
> wouldn't be that great idea.

2.0? 2.2 has been out more than long enough for boxes running such 
kernels to have rolled over several times already.

security: if the box concerned is heavily firewalled and partitioned 
away from rest of internet, whats the problem? :)

> But anyway, thanks for the information!
> 
> 	Markus

regards,
-- 
Paul Jakma	paul@clubi.ie	paul@jakma.org	Key ID: 64A2FF6A
	warning: do not ever send email to spam@dishone.st
Fortune:
"Probably the best operating system in the world is the [operating system]
 made for the PDP-11 by Bell Laboratories." - Ted Nelson, October 1977

Re: Uptime counter

[Markus Hästbacka]

On Sun, 1 Feb 2004, David Weinehall wrote:

> On Mon, Feb 02, 2004 at 12:07:40AM +0200, Markus Hästbacka wrote:
[..snip..]
>
> Well, since most of the changes in the latest kernels involve
> networking, trying it with various different network-adapters would be
> interesting, and stress-testing the network-code in general.
>
> If you have the hardware or a really good confidence, a recent
> 2.2-kernel to compare with and sufficient knowledge of C, have a look at
> the network-drivers for a2065 and ariadne, both of which lack the
> padding-fixes the other adapters have, since I didn't want to touch that
> mess...
>
I don't have that hardware to test on, but I can stress test it tomorrow,
give the box a _high_ load, and that way help testing.

Thanks for the suggestions.

	Markus

Re: Uptime counter

[Markus Hästbacka]

On Sun, 1 Feb 2004, Paul Jakma wrote:
>
> 2.0? 2.2 has been out more than long enough for boxes running such
> kernels to have rolled over several times already.
>
Well, not for boxen with 2.0.40-rc8 (Check release date @ kernel.org and
be suprised :-)
> security: if the box concerned is heavily firewalled and partitioned
> away from rest of internet, whats the problem? :)
>
Nah, it's a webserver!

	Markus

Re: Uptime counter

[David Weinehall]

On Mon, Feb 02, 2004 at 12:59:26AM +0200, Markus Hästbacka wrote:
> On Sun, 1 Feb 2004, Paul Jakma wrote:
> >
> > 2.0? 2.2 has been out more than long enough for boxes running such
> > kernels to have rolled over several times already.
> >
> Well, not for boxen with 2.0.40-rc8 (Check release date @ kernel.org and
> be suprised :-)

And for more amusement, check release-date for 2.0.40-rc6 first, then
-rc7...

> > security: if the box concerned is heavily firewalled and partitioned
> > away from rest of internet, whats the problem? :)
> >
> Nah, it's a webserver!

Cool.  I'm assuming it's NOT an SMP-machine?


Regards: David Weinehall
-- 
 /) David Weinehall <tao@acc.umu.se> /) Northern lights wander      (\
//  Maintainer of the v2.0 kernel   //  Dance across the winter sky //
\)  http://www.acc.umu.se/~tao/    (/   Full colour fire           (/

Re: Uptime counter

[Wakko Warner]

> 2.0? 2.2 has been out more than long enough for boxes running such 
> kernels to have rolled over several times already.

Definately, I have a 2.2.13 box that's rolled 3 times.  I have to use utmp
to get the uptime.

[wakko@rod:/home/wakko] last -xf /var/run/utmp runlevel
runlevel (to lvl 5)                    Thu Nov 18 22:36 - 18:48 (1535+20:12)

utmp begins Thu Nov 18 22:36:07 1999
[wakko@rod:/home/wakko] 
  6:49pm  up 44 days, 15:23h,  3 users,  load average: 0.00, 0.00, 0.00
[wakko@rod:/home/wakko] 

It is firewalled by the way.
  
-- 
 Lab tests show that use of micro$oft causes cancer in lab animals

Re: Uptime counter

[Markus Hästbacka]

On Mon, 2 Feb 2004, David Weinehall wrote:
> On Mon, Feb 02, 2004 at 12:59:26AM +0200, Markus Hästbacka wrote:
> > On Sun, 1 Feb 2004, Paul Jakma wrote:
> > >
> > > 2.0? 2.2 has been out more than long enough for boxes running such
> > > kernels to have rolled over several times already.
> > >
> > Well, not for boxen with 2.0.40-rc8 (Check release date @ kernel.org and
> > be suprised :-)
>
> And for more amusement, check release-date for 2.0.40-rc6 first, then
> -rc7...
>
Yep. :-)
> > > security: if the box concerned is heavily firewalled and partitioned
> > > away from rest of internet, whats the problem? :)
> > >
> > Nah, it's a webserver!
>
> Cool.  I'm assuming it's NOT an SMP-machine?
>
Yep, a non-SMP box. (I hope I someday even will own a smp-box, *dreams* :)

	Markus

Re: Uptime counter

[David Weinehall]

On Mon, Feb 02, 2004 at 09:20:13AM +0200, Markus Hästbacka wrote:
> On Mon, 2 Feb 2004, David Weinehall wrote:
> > On Mon, Feb 02, 2004 at 12:59:26AM +0200, Markus Hästbacka wrote:
> > > On Sun, 1 Feb 2004, Paul Jakma wrote:
> > > >
> > > > 2.0? 2.2 has been out more than long enough for boxes running such
> > > > kernels to have rolled over several times already.
> > > >
> > > Well, not for boxen with 2.0.40-rc8 (Check release date @ kernel.org and
> > > be suprised :-)
> >
> > And for more amusement, check release-date for 2.0.40-rc6 first, then
> > -rc7...
> >
> Yep. :-)
> > > > security: if the box concerned is heavily firewalled and partitioned
> > > > away from rest of internet, whats the problem? :)
> > > >
> > > Nah, it's a webserver!
> >
> > Cool.  I'm assuming it's NOT an SMP-machine?
> >
> Yep, a non-SMP box. (I hope I someday even will own a smp-box, *dreams* :)

Well, if you ever get one, don't even attempt to run a 2.0-kernel on it.
While possible, it's about as effecient as buying a Ferrari and then try
to push it to its limits in London, high traffic hours...


Regards: David Weinehall
-- 
 /) David Weinehall <tao@acc.umu.se> /) Northern lights wander      (\
//  Maintainer of the v2.0 kernel   //  Dance across the winter sky //
\)  http://www.acc.umu.se/~tao/    (/   Full colour fire           (/

Re: Uptime counter

[Markus Hästbacka]

On Mon, 2 Feb 2004, David Weinehall wrote:
[..snip..]
> > Yep, a non-SMP box. (I hope I someday even will own a smp-box, *dreams* :)
>
> Well, if you ever get one, don't even attempt to run a 2.0-kernel on it.
> While possible, it's about as effecient as buying a Ferrari and then try
> to push it to its limits in London, high traffic hours...
>
Right, though, I think 10.x is out before I can get a SMP box. :)
I would describe it with putting a ferrari engine in a lada.

	Markus

Re: Uptime counter

[Valdis.Kletnieks]

[-- Attachment #1: Type: text/plain, Size: 291 bytes --]

On Sun, 01 Feb 2004 22:02:06 +0100, you said:

> By the way: Having a machine with more than 497 days of uptime normally 
> shows a serios lack of security awareness..

Or an embedded system that you don't really feel like rebooting, like
an HVAC controller or a deep-space probe....

[-- Attachment #2: Type: application/pgp-signature, Size: 226 bytes --]

Re: Uptime counter

[Mike Fedyk]

On Mon, Feb 02, 2004 at 12:54:42PM -0500, Valdis.Kletnieks@vt.edu wrote:
> Or an embedded system that you don't really feel like rebooting, like
> an HVAC controller or a deep-space probe....

Quick!  Get a satalite dish! ;)

Re: Uptime counter

[Bill Davidsen]

Christian Borntraeger wrote:
> Markus Hästbacka wrote:
> 
>>Hi list,
>>I wonder does any kernel branch have a uptime counter that doesn't stop
>>counting at 497 days? Or is a patch needed for the job to
>>2.{0,2,4,6} kernel?
> 
> 
> In 2.6 there is no 497 days limit, as jiffies are now 64 bit. 
> 
> By the way: Having a machine with more than 497 days of uptime normally 
> shows a serios lack of security awareness..

Think embedded systems...

-- 
bill davidsen <davidsen@tmr.com>
   CTO TMR Associates, Inc
   Doing interesting things with small computers since 1979