RE: [PATCH V1 2/6] regulator: da9062: DA9062 regulator driver

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 18 April 2015 12:48 Mark Brown wrote:

> On Fri, Apr 17, 2015 at 03:23:32PM +0100, S Twiss wrote:
> 
> > +/* Regulator interrupt handlers */
> > +static irqreturn_t da9062_ldo_lim_event(int irq, void *data)
> > +{
> > +	struct da9062_regulators *regulators = data;
> > +	struct da9062 *hw = regulators->regulator[0].hw;
> > +	struct da9062_regulator *regl;
> > +	int bits, i, ret;
> > +
> > +	ret = regmap_read(hw->regmap, DA9062AA_STATUS_D, &bits);
> > +	if (ret < 0)
> > +		return IRQ_NONE;
> 
> Please log an error for this, if we're having trouble talking to the
> device that seems like a serious issue.

Will do that with a dev_err()

-       if (ret < 0)
+       if (ret < 0) {
+               dev_err(hw->dev,
+                       "Failed to read LDO overcurrent indicator\n");
                return IRQ_NONE;
+       }

> 
> > +	for (i = regulators->n_regulators - 1; i >= 0; i--) {
> > +		regl = &regulators->regulator[i];
> > +		if (regl->info->oc_event.reg != DA9062AA_STATUS_D)
> > +			continue;
> > +
> > +		if (BIT(regl->info->oc_event.lsb) & bits)
> > +			regulator_notifier_call_chain(regl->rdev,
> > +
> 	REGULATOR_EVENT_OVER_CURRENT, NULL);
> > +	}
> > +
> > +	return IRQ_HANDLED;
> 
> This will return IRQ_HANDLED even if none of the regulators were
> flagginng an event.
> 

Thanks,
I will refactor that part with a default error path and only return IRQ_HANDLED
when there is a hit on the notifier call.

> > +static irqreturn_t da9062_vdd_warn_event(int irq, void *data)
> > +{
> > +	return IRQ_HANDLED;
> > +}
> 
> Ignoring an interrupt is not usefully handling it - at the *least* this
> should be generating a log message.
> 

Not much use -- yes.

I've taken another look at this and I've decided to move it into the core.
The VDD_WARN applies to the system voltage and is not specific to regulators.
So it would make more sense for this to exist in the core driver instead -- this
would need to exist at all times and if it stays in the regulators it could be
moduled out of existence.

This will mean messing with a few files to fix this one.

drivers/mfd/da9062-core.c
regulator/da9062-regulator.c
include/linux/mfd/da9062/core.h

I'll resend this as PATCH V2.

> > +static struct da9062_regulators_pdata *da9062_parse_regulators_dt(
> > +		struct platform_device *pdev,
> > +		struct of_regulator_match **reg_matches)
> > +{
> > +	struct da9062_regulators_pdata *pdata;
> > +	struct da9062_regulator_data *rdata;
> > +	struct device_node *node;
> > +	int i, n, num;
> > +
> > +	node = of_get_child_by_name(pdev->dev.parent->of_node,
> "regulators");
> > +	if (!node) {
> > +		dev_err(&pdev->dev, "Regulators device node not
> found\n");
> > +		return ERR_PTR(-ENODEV);
> > +	}
> > +
> > +	num = of_regulator_match(&pdev->dev, node, da9062_matches,
> > +				 ARRAY_SIZE(da9062_matches));
> 
> Don't open code this, describe the DT names in the regualtor_desc and
> let the core register.
> 

Okay. I think I am getting this.
As of v3.18 there are  newer parts to regulator_desc from the commit 
a0c7b16 "regulator: of: Provide simplified DT parsing method"
The search function is now in the core. 

Am I on on the right track with this one?

> > +	if (IS_ERR(pdata) || pdata->n_regulators == 0) {
> > +		dev_err(&pdev->dev,
> > +			"No regulators defined for the platform\n");
> > +		return PTR_ERR(pdata);
> > +	}
> > +
> > +	n_regulators = ARRAY_SIZE(local_regulator_info),
> 
> This is broken, the set of regulators in the silicon is not a property
> of the platform.  The driver should just register all the regualtors
> that are present in the silicon.  I'm fairly sure I've been through this
> before...
> 
> > +	ret = request_threaded_irq(irq,
> > +				   NULL, da9062_vdd_warn_event,
> > +				   IRQF_TRIGGER_LOW | IRQF_ONESHOT,
> > +				   "VDD_WARN", regulators);
> 
> devm_request_threaded_irq().

yes -- this will come as part of the refactoring for the irq handler
static irqreturn_t da9062_vdd_warn_event(int irq, void *data)
into the core.

Regards,
Steve

--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html




[Index of Archives]     [Linux Media Devel]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [Linux Wireless Networking]     [Linux Omap]

  Powered by Linux