APM Patches

[Liam J. Foy]

On Wed, 07 Jul 2004 19:42:35 +0300
Niki Denev <nike_d at cytexbg.com> wrote:

> Niki Denev writes:
> 
> > M. Warner Losh writes:
> > 
> >> In message: <20040704204353.26b4496b.liamfoy at sepulcrum.org>
> >>             "Liam J. Foy" <liamfoy at sepulcrum.org> writes:
> >> : Hey guys,
> >> : 
> >> : 	Since it was decided (http://lists.freebsd.org/pipermail/freebsd-acpi/2004-June/000352.html)
> >> : we are going to stick with apm -l producing -1 and not 255 which is stated in the handbook would one 
> >> : of you guys please commit:
> >> : 
> >> : --- /usr/src/usr.sbin/apm/apm.8	Thu Jun 24 17:32:55 2004
> >> : +++ /liamfoy/apm.8	Thu Jun 24 17:32:27 2004
> >> : @@ -106,7 +106,7 @@
> >> :  state respectively.
> >> :  .It Fl t
> >> :  Display the estimated remaining battery lifetime in seconds.  If
> >> : -it is unknown, 255 is displayed.
> >> : +it is unknown, -1 is displayed.
> >> :  .It Fl Z
> >> :  Transition the system into standby mode.  This mode uses less power than
> >> :  full power mode, but more than suspend mode.  Some laptops support
> >> : 
> >> : ----
> >> : 
> >> : Another patch I would like you guys to review is this. Currently apm -t will output
> >> : 0 when it cannot find a valid rate or the full battery time(as the comment mentions).
> >> : I think it should return -1 (unknown) to reflect an error, which is stated in the man page.
> >> : It should not return 0 since we do not have 0 seconds left, we have an unknown value
> >> : remaining. Either that or the man page it edited. I believe the following patch should 
> >> : be commited really.
> >> : 
> >> : The patch is:
> >> : 
> >> : --- /usr/src/sys/dev/acpica/acpi_cmbat.c	Sun Jul  4 20:41:43 2004
> >> : +++ /home/liamfoy/acpi_cmbat.c	Sun Jul  4 20:39:14 2004
> >> : @@ -536,7 +536,7 @@
> >> :  	    bat[i]->min = (bat[i]->full_charge_time * bat[i]->cap) / 100;
> >> :  	} else {
> >> :  	    /* Couldn't find valid rate and full battery time */
> >> : -	    bat[i]->min = 0;
> >> : +	    bat[i]->min = -1;
> >> :  	}
> >> :  	total_min += bat[i]->min;
> >> :  	total_cap += bat[i]->cap;
> >> 
> >> I don't like this patch, since we use ->min later for math...
> >> 
> >> Warner
> > 
> > What about this ?
> > 
> > --- sys/dev/acpica/acpi_cmbat.c.orig	Mon Jul  5 15:15:28 2004
> > +++ sys/dev/acpica/acpi_cmbat.c	Mon Jul  5 16:37:02 2004
> > @@ -655,7 +655,7 @@
> >  	battinfo->state = ACPI_BATT_STAT_NOT_PRESENT;
> >      } else {
> >  	battinfo->cap = sc->cap;
> > -	battinfo->min = sc->min;
> > +	battinfo->min = sc->min ? sc->min : -1;
> >  	battinfo->state = sc->bst.state;
> >      }
> >  
> > 
> > 
> > --
> > Regards,
> > Niki
> 
> 
> So, is this Ok or not? :)
> I'm running with this for a few days... no probs...
> But also the only time that this is used probably is in apm(8) ?
> It seems to not affect the math with ->min, by maybe it is not the cleanest
> solution to this cosmetic problem?

I personally like this patch. It does not effect the math and is much
better than returning 0 which would indicate to the user he/she has not
battery life left. I believe this should be changed.

Depends on what the other guys think.
> :)
> 
> --
> Cheers,
> Niki
> 
> 


-- 
-Liam J. Foy
http://liamfoy.kerneled.org
"Love is like maths -- the idea is simple but can be quite complicated."