RE: [PATCH 3/4] power_supply: Introduce PSE compliant algorithm

[Tc, Jenny]

> From: Sebastian Reichel [mailto:sre@xxxxxxxxxx]
> Sent: Tuesday, July 08, 2014 9:26 PM
> To: Tc, Jenny
> Cc: linux-kernel@xxxxxxxxxxxxxxx; Dmitry Eremin-Solenikov; Pavel Machek; Stephen
> Rothwell; Anton Vorontsov; David Woodhouse; David Cohen; Pallala, Ramakrishna
> Subject: Re: [PATCH 3/4] power_supply: Introduce PSE compliant algorithm
> 
> On Tue, Jul 08, 2014 at 06:07:29AM +0000, Tc, Jenny wrote:
> > > > +static int get_tempzone(struct psy_pse_chrg_prof *pse_mod_bprof,
> > > > +		int temp)
> > > > +{
> > > > +	int i = 0;
> > > > +	int temp_range_cnt;
> > > > +
> > > > +	temp_range_cnt =  min_t(u16, pse_mod_bprof->temp_mon_ranges,
> > > > +					BATT_TEMP_NR_RNG);
> > > > +	if ((temp < pse_mod_bprof->temp_low_lim) ||
> > > > +		(temp > pse_mod_bprof->temp_mon_range[0].temp_up_lim))
> > > > +		return -EINVAL;
> > > > +
> > > > +	for (i = 0; i < temp_range_cnt; ++i)
> > > > +		if (temp > pse_mod_bprof->temp_mon_range[i].temp_up_lim)
> > > > +			break;
> > > > +	return i-1;
> > > > +}
> > >
> > > pse_mod_bprof->temp_mon_ranges > BATT_TEMP_NR_RNG is not allowed, so
> > > I suggest to print an error and return some error code.
> > >
> > min_t takes care of the upper bound. The algorithm process
> > BATT_TEMP_NR_RNG even if the actual number of zones are greater than this.
> 
> Right, the function will not fail, but the zone information table is truncated. I would
> expect at least warning about that. I think it doesn't hurt to have a small function,
> which validates the zone data as good as possible. Using incorrect temperature
> zones is a safety thread and we should try our best to avoid exploding batteries ;)
> 
> Maybe something like that:
> 
> static bool check_tempzones(struct psy_pse_chrg_prof *pse_mod_bprof) {
>     int i = 0;
>     int last_temp = ;
> 
>     /* check size */
>     if (BATT_TEMP_NR_RNG > pse_mod_bprof->temp_mon_ranges)
>         return false;

This is in a way good to have, OK to implement the same.

But KO with below suggestion. This doesn't guarantee safety. IMHO
Safety is 1/0 - SAFE or NOT SAFE. No half safety.

To ensure complete safety, measures should be taken at the entry point- where data
is read from external source. Since the algorithm gets the data from internal
kernel component (power_supply_charger.c), it trust the data. Since the data
is originated from battery identification driver, the safety should be ensured at
that level.

> 
>     /* check zone order */
>     for (i = 0; i < pse_mod_bprof->temp_mon_ranges; i++) {
>         if (last_temp < pse_mod_bprof->temp_mon_range[i].temp_up_lim)
>             return false;
>         last_temp = pse_mod_bprof->temp_mon_range[i].temp_up_lim;
>     }
> 
>     return true;
> }
> 
> -- Sebastian
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/