linux-kernel - Re: [PATCH 03/16] iio: adc: max1027: Push only the requested samples

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]

Message-ID: <20210830110756.733d5201@jic23-huawei>
Date:   Mon, 30 Aug 2021 11:07:56 +0100
From:   Jonathan Cameron <jic23@...nel.org>
To:     "Sa, Nuno" <Nuno.Sa@...log.com>
Cc:     Miquel Raynal <miquel.raynal@...tlin.com>,
        Lars-Peter Clausen <lars@...afoo.de>,
        Thomas Petazzoni <thomas.petazzoni@...tlin.com>,
        "linux-iio@...r.kernel.org" <linux-iio@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 03/16] iio: adc: max1027: Push only the requested
 samples

On Fri, 20 Aug 2021 07:10:48 +0000
"Sa, Nuno" <Nuno.Sa@...log.com> wrote:

> > -----Original Message-----
> > From: Miquel Raynal <miquel.raynal@...tlin.com>
> > Sent: Wednesday, August 18, 2021 1:11 PM
> > To: Jonathan Cameron <jic23@...nel.org>; Lars-Peter Clausen
> > <lars@...afoo.de>
> > Cc: Thomas Petazzoni <thomas.petazzoni@...tlin.com>; linux-
> > iio@...r.kernel.org; linux-kernel@...r.kernel.org; Miquel Raynal
> > <miquel.raynal@...tlin.com>
> > Subject: [PATCH 03/16] iio: adc: max1027: Push only the requested
> > samples
> > 
> > [External]
> > 
> > When a triggered scan occurs, the identity of the desired channels is
> > known in indio_dev->active_scan_mask. Instead of reading and
> > pushing to
> > the IIO buffers all channels each time, scan the minimum amount of
> > channels (0 to maximum requested chan, to be exact) and only
> > provide the
> > samples requested by the user.
> > 
> > For example, if the user wants channels 1, 4 and 5, all channels from
> > 0 to 5 will be scanned but only the desired channels will be pushed to
> > the IIO buffers.
> > 
> > Signed-off-by: Miquel Raynal <miquel.raynal@...tlin.com>
> > ---
> >  drivers/iio/adc/max1027.c | 25 +++++++++++++++++++++----
> >  1 file changed, 21 insertions(+), 4 deletions(-)
> > 
> > diff --git a/drivers/iio/adc/max1027.c b/drivers/iio/adc/max1027.c
> > index b753658bb41e..8ab660f596b5 100644
> > --- a/drivers/iio/adc/max1027.c
> > +++ b/drivers/iio/adc/max1027.c
> > @@ -360,6 +360,9 @@ static int max1027_set_trigger_state(struct
> > iio_trigger *trig, bool state)
> >  	struct max1027_state *st = iio_priv(indio_dev);
> >  	int ret;
> > 
> > +	if (bitmap_empty(indio_dev->active_scan_mask, indio_dev-  
> > >masklength))  
> > +		return -EINVAL;
> > +  
> 
> I'm not sure this can actually happen. If you try to enable the buffer
> with no scan element, it should give you an error before you reach
> this point...
> 
> >  	if (state) {
> >  		/* Start acquisition on cnvst */
> >  		st->reg = MAX1027_SETUP_REG |
> > MAX1027_CKS_MODE0 |
> > @@ -368,9 +371,12 @@ static int max1027_set_trigger_state(struct
> > iio_trigger *trig, bool state)
> >  		if (ret < 0)
> >  			return ret;
> > 
> > -		/* Scan from 0 to max */
> > -		st->reg = MAX1027_CONV_REG | MAX1027_CHAN(0) |
> > -			  MAX1027_SCAN_N_M | MAX1027_TEMP;
> > +		/*
> > +		 * Scan from 0 to the highest requested channel. The
> > temperature
> > +		 * could be avoided but it simplifies a bit the logic.
> > +		 */
> > +		st->reg = MAX1027_CONV_REG |
> > MAX1027_SCAN_0_N | MAX1027_TEMP;
> > +		st->reg |= MAX1027_CHAN(fls(*indio_dev-  
> > >active_scan_mask) - 2);  
> >  		ret = spi_write(st->spi, &st->reg, 1);
> >  		if (ret < 0)
> >  			return ret;
> > @@ -391,11 +397,22 @@ static irqreturn_t
> > max1027_trigger_handler(int irq, void *private)
> >  	struct iio_poll_func *pf = private;
> >  	struct iio_dev *indio_dev = pf->indio_dev;
> >  	struct max1027_state *st = iio_priv(indio_dev);
> > +	unsigned int scanned_chans = fls(*indio_dev-  
> > >active_scan_mask);  
> > +	u16 *buf = st->buffer;  
> 
> I think sparse will complain here. buffer is a __be16 restricted
> type so you should not mix those... 
> > +	unsigned int bit;
> > 
> >  	pr_debug("%s(irq=%d, private=0x%p)\n", __func__, irq,
> > private);in/20210818_miquel_raynal_bring_software_triggers_support_to_max1027_like_adcs.mbx
> > 
> >  	/* fill buffer with all channel */
> > -	spi_read(st->spi, st->buffer, indio_dev->masklength * 2);
> > +	spi_read(st->spi, st->buffer, scanned_chans * 2);
> > +
> > +	/* Only keep the channels selected by the user */
> > +	for_each_set_bit(bit, indio_dev->active_scan_mask,
> > +			 indio_dev->masklength) {
> > +		if (buf[0] != st->buffer[bit])
> > +			buf[0] = st->buffer[bit];  
> 
> Since we are here, when looking into the driver, I realized
> that st->buffer is not DMA safe. In IIO, we kind of want to enforce
> that all buffers that are passed to spi/i2c buses are safe... Maybe
> this is something you can include in your series.

Why is it not?  st->buffer is result of a devm_kmalloc_array() call and
that should provide a DMA safe buffer as I understand it.

> 
> - Nuno Sá
> 
>  > +		buf++;
> > +	}
> > 
> >  	iio_push_to_buffers(indio_dev, st->buffer);
> > 
> > --
> > 2.27.0  
>