linux-kernel - Re: [PATCH] rtc: pcf2127: add error message if writing to CLKOUT register fails

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Thu, 20 Jan 2022 18:35:48 -0500
From:   Hugo Villeneuve <hugo@...ovil.com>
To:     Alexandre Belloni <alexandre.belloni@...tlin.com>
Cc:     Alessandro Zummo <a.zummo@...ertech.it>,
        Hugo Villeneuve <hvilleneuve@...onoff.com>,
        linux-rtc@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] rtc: pcf2127: add error message if writing to CLKOUT
 register fails

On Wed, 19 Jan 2022 20:10:20 +0100
Alexandre Belloni <alexandre.belloni@...tlin.com> wrote:

> On 19/01/2022 13:08:45-0500, Hugo Villeneuve wrote:
> > On Wed, 19 Jan 2022 18:37:40 +0100
> > Alexandre Belloni <alexandre.belloni@...tlin.com> wrote:
> > 
> > > On 19/01/2022 12:27:39-0500, Hugo Villeneuve wrote:
> > > > From: Hugo Villeneuve <hvilleneuve@...onoff.com>
> > > > 
> > > > If writing to CLKOUT register fails, the probe operation will be aborted
> > > > without a meaningful error message.
> > > > 
> > > 
> > > The current trend is to remove debug messages, please do not add more :)
> > 
> > Hi,
> > If the read operation fails, the probe function will exit silently, and our RTC chip will not work. In that case, if we parse the dmesg logs, I think we  should have an indication that something went wrong.
> > 
> 
> This is not true, it doesn't fail silently, you'd get:
> rtc-pcf2127: probe of 1-0051 failed with error -121

Well this is certainly true for me because I am not seing the same error message as you :)

Just for context, I have defined a dummy pcf2127 on I2C bus 0 in my device tree (no actual hardware is present). I also added some debug messages to investigate (rtc-pcf2127.c and dd.c files), and here is the dmesg log after issuing "modprobe rtc-pcf2127":

[Thu Jan 20 23:22:20 2022] rtc-pcf2127-i2c 0-0051: pcf2127_i2c_probe
[Thu Jan 20 23:22:20 2022] rtc-pcf2127-i2c 0-0051: pcf2127_probe
[Thu Jan 20 23:22:20 2022] rtc-pcf2127-i2c 0-0051: PORO disabling failed with error -6
[Thu Jan 20 23:22:20 2022] rtc-pcf2127-i2c 0-0051: call_driver_probe probe error: -6

Error code -6 is -ENXIO, and looking at the call_driver_probe() function in dd.c, I now understand why I didn't see the error message (line 531):

	case -ENXIO:
		pr_debug("%s: probe of %s rejects match %d\n",
			 drv->name, dev_name(dev), ret);

So it seems that the return code is different than what you got?


> > I had a case where my dev board was not properly plugged-in, and before this patch, I didn't notice it because of the silent abort of the probe function.
> > 
> 
> Again, not silent.
> 
> > > 
> > > > Signed-off-by: Hugo Villeneuve <hvilleneuve@...onoff.com>
> > > > ---
> > > >  drivers/rtc/rtc-pcf2127.c | 4 +++-
> > > >  1 file changed, 3 insertions(+), 1 deletion(-)
> > > > 
> > > > diff --git a/drivers/rtc/rtc-pcf2127.c b/drivers/rtc/rtc-pcf2127.c
> > > > index 3d1f57e54372..823abe2a7147 100644
> > > > --- a/drivers/rtc/rtc-pcf2127.c
> > > > +++ b/drivers/rtc/rtc-pcf2127.c
> > > > @@ -717,8 +717,10 @@ static int pcf2127_probe(struct device *dev, struct regmap *regmap,
> > > >  	if (!(val & PCF2127_BIT_CLKOUT_OTPR)) {
> > > >  		ret = regmap_set_bits(pcf2127->regmap, PCF2127_REG_CLKOUT,
> > > >  				      PCF2127_BIT_CLKOUT_OTPR);
> > > > -		if (ret < 0)
> > > > +		if (ret < 0) {
> > > > +			dev_err(dev, "writing to CLKOUT register failed\n");
> > > >  			return ret;
> > > > +		}
> > > >  
> > > >  		msleep(100);
> > > >  	}
> > > > -- 
> > > > 2.30.2
> > > > 
> > > 
> > > -- 
> > > Alexandre Belloni, co-owner and COO, Bootlin
> > > Embedded Linux and Kernel engineering
> > > https://bootlin.com
> > > 
> > 
> > 
> > -- 
> > Hugo Villeneuve <hugo@...ovil.com>
> 
> -- 
> Alexandre Belloni, co-owner and COO, Bootlin
> Embedded Linux and Kernel engineering
> https://bootlin.com
> 

-- 
Hugo Villeneuve <hugo@...ovil.com>