linux-kernel - Re: [PATCH v2 2/2] iio: hx711: fix bug in reset functionality

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] [day] [month] [year] [list]

Message-ID: <20171205131246.GC1958@arbeit>
Date:   Tue, 5 Dec 2017 14:12:46 +0100
From:   Andreas Klinger <ak@...klinger.de>
To:     Jonathan Cameron <jic23@...nel.org>
Cc:     knaack.h@....de, lars@...afoo.de, pmeerw@...erw.net,
        singhalsimran0@...il.com, linux-iio@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 2/2] iio: hx711: fix bug in reset functionality

Hi Jonathan,

see answer to your question below.

Andreas

Jonathan Cameron <jic23@...nel.org> schrieb am Sat, 02. Dec 13:22:
> On Thu, 30 Nov 2017 22:44:45 +0100
> Andreas Klinger <ak@...klinger.de> wrote:
> 
> > return value in hx711_reset() should indicate status of dout
> > otherwise the calling function is reporting an error als false positive
> > 
> > wait up to 1 second until the device is in normal state
> > by changing the channel on every trigger event it turned out that the
> > former 100 ms are not enough for waiting until the device is in normal mode
> 
> What is the ultimate result of this?  I'm trying to assess if we want
> to marked it for stable trees or just apply it ready for the next
> merge window.  It think it means that two reads too close to each
> other will always result in an error for the second one rather than 
> the intended delay and then succeed?

Exactly. If one is accessing to fast the second access will fail. Especially
when using a trigger the second channel is never read without error.

> 
> Thanks,
> 
> Jonathan
> > 
> > Signed-off-by: Andreas Klinger <ak@...klinger.de>
> > ---
> >  drivers/iio/adc/hx711.c | 13 ++++++-------
> >  1 file changed, 6 insertions(+), 7 deletions(-)
> > 
> > diff --git a/drivers/iio/adc/hx711.c b/drivers/iio/adc/hx711.c
> > index d773fd69a76d..eabd2e66aca3 100644
> > --- a/drivers/iio/adc/hx711.c
> > +++ b/drivers/iio/adc/hx711.c
> > @@ -153,15 +153,16 @@ static int hx711_wait_for_ready(struct hx711_data *hx711_data)
> >  	int i, val;
> >  
> >  	/*
> > -	 * a maximum reset cycle time of 56 ms was measured.
> > -	 * we round it up to 100 ms
> > +	 * in some rare cases the reset takes quite a long time
> > +	 * especially when the channel is changed.
> > +	 * Allow up to one second for it
> >  	 */
> >  	for (i = 0; i < 100; i++) {
> >  		val = gpiod_get_value(hx711_data->gpiod_dout);
> >  		if (!val)
> >  			break;
> > -		/* sleep at least 1 ms */
> > -		msleep(1);
> > +		/* sleep at least 10 ms */
> > +		msleep(10);
> >  	}
> >  	if (val)
> >  		return -EIO;
> > @@ -203,9 +204,7 @@ static int hx711_reset(struct hx711_data *hx711_data)
> >  		 * after a dummy read we need to wait vor readiness
> >  		 * for not mixing gain pulses with the clock
> >  		 */
> > -		ret = hx711_wait_for_ready(hx711_data);
> > -		if (ret)
> > -			return ret;
> > +		val = hx711_wait_for_ready(hx711_data);
> >  	}
> >  
> >  	return val;
> 

--