[timer] max timeout

[timer] max timeout

[sebastien person]

Hi,

is there a max timeout to respect when I use mod_timer ? or add_timer ?

Is it bad to do the following call ?

	mod_timer(&timer, jiffies+(0.1*HZ));

that might fire the timer 1/10 second later.

Thanks.

sebastien person

Re: [timer] max timeout

[Andi Kleen]

On Wed, May 23, 2001 at 04:28:01PM +0200, sebastien person wrote:
> Is it bad to do the following call ?
> 
> 	mod_timer(&timer, jiffies+(0.1*HZ));

Yes very bad. gcc will generate a floating point add for that, corrupting
the user process' floating point context. 


-Andi

Re: [timer] max timeout

[Andrzej Krzysztofowicz]

"sebastien person wrote:"
> Is it bad to do the following call ?
> 
> 	mod_timer(&timer, jiffies+(0.1*HZ));

Yes, it is bad. Don't use floating point in the kernel if you don't need.

> that might fire the timer 1/10 second later.

HZ/10 is much better ...

-- 
=======================================================================
  Andrzej M. Krzysztofowicz               ankry@mif.pg.gda.pl
  phone (48)(58) 347 14 61
Faculty of Applied Phys. & Math.,   Technical University of Gdansk

Re: [timer] max timeout

[sebastien person]

Le Wed, 23 May 2001 16:58:15 +0200 (MET DST)
Andrzej Krzysztofowicz <ankry@pg.gda.pl> a ecrit :

> "sebastien person wrote:"
> > Is it bad to do the following call ?
> > 
> > 	mod_timer(&timer, jiffies+(0.1*HZ));
> 
> Yes, it is bad. Don't use floating point in the kernel if you don't
need.

So, there is a good solution to fire the timer after 100 ms ?
And is there any max limitation on the expires value ?

e.g. what is the biggest time period we could have between timer call and
the execution of the timer function ?



> 
> > that might fire the timer 1/10 second later.
> 
> HZ/10 is much better ...
> 
> -- 
> =======================================================================
>   Andrzej M. Krzysztofowicz               ankry@mif.pg.gda.pl
>   phone (48)(58) 347 14 61
> Faculty of Applied Phys. & Math.,   Technical University of Gdansk
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel"
in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

Re: [timer] max timeout

[george anzinger]

sebastien person wrote:
> 
> Hi,
> 
> is there a max timeout to respect when I use mod_timer ? or add_timer ?
> 
> Is it bad to do the following call ?
> 
>         mod_timer(&timer, jiffies+(0.1*HZ));
> 
> that might fire the timer 1/10 second later.
> 
> Thanks.
> 
More than enough on the fp.  Now the other question.  

The time value is in jiffies (aka 1/Hz sec.).  The max value (in a 32
bit system) is 0x7ffffff.  This value is added to the current value of
jiffies to get the the absolute timer expire time.  While the value is
unsigned (and thus you could go higher) the compare code (timer_before()
and timer_after()) depend on the subtraction (of jiffies from expire) to
be of the correct sign.  To insure this you must keep the timeout values
sign clear (or if you don't like to think of an unsigned as having a
sign, then the highest order bit must be zero).

George

Re: [timer] max timeout

[Mark Frazer]

george anzinger <george@mvista.com> [01/05/25 17:04]:
> More than enough on the fp.  Now the other question.  
> 
> The time value is in jiffies (aka 1/Hz sec.).  The max value (in a 32
> bit system) is 0x7ffffff.  This value is added to the current value of
> jiffies to get the the absolute timer expire time.  While the value is
> unsigned (and thus you could go higher) the compare code (timer_before()
> and timer_after()) depend on the subtraction (of jiffies from expire) to
> be of the correct sign.  To insure this you must keep the timeout values
> sign clear (or if you don't like to think of an unsigned as having a
> sign, then the highest order bit must be zero).
> 
> George

The output of `find . -type f | xargs grep 'jiffies +'` would suggest
that there are a few latent bugs as jiffies grows to values near the
top of its range.  I guess this hasn't turned up as 0x7fffffff / (100 *
3600 * 24) = 248.55.

Re: [timer] max timeout

[Horst von Brand]

Mark Frazer <mark@somanetworks.com> said:

[...]

> The output of `find . -type f | xargs grep 'jiffies +'` would suggest
> that there are a few latent bugs as jiffies grows to values near the
> top of its range.  I guess this hasn't turned up as 0x7fffffff / (100 *
> 3600 * 24) = 248.55.

There were patches floating around a _long_ time ago (2.1-ish AFAIR) that
allowed you to start your system at any jiffy. Many people checked out what
happens on jiffie rollover then. Might be a good idea to do that again...
-- 
Dr. Horst H. von Brand                       mailto:vonbrand@inf.utfsm.cl
Departamento de Informatica                     Fono: +56 32 654431
Universidad Tecnica Federico Santa Maria              +56 32 654239
Casilla 110-V, Valparaiso, Chile                Fax:  +56 32 797513