RE: [PATCH] rtc-pm8xxx: Clear Alarm register on resume

[maggarwa]

Comments inline.

-----Original Message-----
From: Alexandre Belloni <alexandre.belloni@xxxxxxxxxxx> 
Sent: Tuesday, March 17, 2020 5:25 PM
To: maggarwa@xxxxxxxxxxxxxx
Cc: a.zummo@xxxxxxxxxxxx; linux-rtc@xxxxxxxxxxxxxxx;
linux-kernel@xxxxxxxxxxxxxxx
Subject: Re: [PATCH] rtc-pm8xxx: Clear Alarm register on resume

On 17/03/2020 10:20:36+0530, maggarwa@xxxxxxxxxxxxxx wrote:
> Hi,
> 
> Comments inline.
> 
> 
> Thanks & Regards,
> Mohit
> 
> -----Original Message-----
> From: Alexandre Belloni <alexandre.belloni@xxxxxxxxxxx>
> Sent: Monday, March 16, 2020 3:59 PM
> To: Mohit Aggarwal <maggarwa@xxxxxxxxxxxxxx>
> Cc: a.zummo@xxxxxxxxxxxx; linux-rtc@xxxxxxxxxxxxxxx; 
> linux-kernel@xxxxxxxxxxxxxxx
> Subject: Re: [PATCH] rtc-pm8xxx: Clear Alarm register on resume
> 
> Hi,
> 
> On 16/03/2020 12:41:28+0530, Mohit Aggarwal wrote:
> > Currently, alarm register is not cleared on resume leading to reboot 
> > during power off charging mode.
> > 
> > Change-Id: Ie2e6bbab8aa46e4e9b9cc984181ffab557cbbdae
> 
> No Change-Id upstream please.
> [Mohit]: Will fix in next patch.
> 
> > Signed-off-by: Mohit Aggarwal <maggarwa@xxxxxxxxxxxxxx>
> > ---
> >  drivers/rtc/rtc-pm8xxx.c | 13 ++++++++++++-
> >  1 file changed, 12 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/rtc/rtc-pm8xxx.c b/drivers/rtc/rtc-pm8xxx.c 
> > index
> > bbe013f..96e7985 100644
> > --- a/drivers/rtc/rtc-pm8xxx.c
> > +++ b/drivers/rtc/rtc-pm8xxx.c
> > @@ -1,5 +1,5 @@
> >  // SPDX-License-Identifier: GPL-2.0-only
> > -/* Copyright (c) 2010-2011, 2019, The Linux Foundation. All rights 
> > reserved. */
> > +/* Copyright (c) 2010-2011, 2019-2020, The Linux Foundation. All 
> > +rights reserved. */
> >  
> >  #include <linux/of.h>
> >  #include <linux/module.h>
> > @@ -301,6 +301,7 @@ static int pm8xxx_rtc_alarm_irq_enable(struct 
> > device
> *dev, unsigned int enable)
> >  	struct pm8xxx_rtc *rtc_dd = dev_get_drvdata(dev);
> >  	const struct pm8xxx_rtc_regs *regs = rtc_dd->regs;
> >  	unsigned int ctrl_reg;
> > +	u8 value[NUM_8_BIT_RTC_REGS] = {0};
> >  
> >  	spin_lock_irqsave(&rtc_dd->ctrl_reg_lock, irq_flags);
> >  
> > @@ -319,6 +320,16 @@ static int pm8xxx_rtc_alarm_irq_enable(struct 
> > device
> *dev, unsigned int enable)
> >  		goto rtc_rw_fail;
> >  	}
> >  
> > +	/* Clear Alarm register */
> > +	if (!enable) {
> > +		rc = regmap_bulk_write(rtc_dd->regmap, regs->alarm_rw,
> value,
> > +					sizeof(value));
> 
> This is not properly aligned.
> [Mohit]: I don't see any alignment issue at my end. I can see proper 
> tabs are present.
> 

The alignment should match the opening parenthesis.
[Mohit]: Thanks for clarification. Uploaded new patch.

> > +		if (rc) {
> > +			dev_err(dev, "Write to RTC ALARM register
> failed\n");
> 
> Is that error message necessary? What would be the user action after 
> seeing that in the logs? Will the logs actually be seen?
> [Mohit]: In case issue in question reproduces even after this change 
> then for debugging purposes user can look out for this error log in 
> kernel logs which can help to triage the issue.
> 

Who in the field on the final product will see this error message? For
debugging purposes, regmap already provides plenty of tracing facilities.
[Mohit]: This is mostly for debugging issues reported during internal tests
using kernel logs. Also added the debug error log to remain in-line with
other error logs in file.


--
Alexandre Belloni, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com