Re: [PATCH] staging: iio: adc: ad7280a: check for devm_kasprint() failure

[Nicholas Mc Guire]

On Mon, Nov 26, 2018 at 04:00:32PM +0300, Dan Carpenter wrote:
> On Mon, Nov 26, 2018 at 10:39:04AM +0100, Nicholas Mc Guire wrote:
> > devm_kasprintf() may return NULL on failure of internal allocation thus
> > the assignments to  attr.name  are not safe if not checked. On error
> > ad7280_attr_init() returns a negative return so -ENOMEM should be
> > OK here (passed on as return value of the probe function).
> > 
> > Signed-off-by: Nicholas Mc Guire <hofrat@xxxxxxxxx>
> > Fixes: 2051f25d2a26 ("iio: adc: New driver for AD7280A Lithium Ion Battery Monitoring System2")
> > ---
> > 
> > Problem located with an experimental coccinelle script
> > 
> > As using   if(!st->iio_attr[cnt].dev_attr.attr.name)  seamed quite
> > unreadable in this case the  (var  == NULL)  variant was used. Not
>                                    ^^
> Why two spaces?

just a typo 

> 
> > sure if there are objections against this (checkpatch.pl issues
> > a CHECK on this).
> > 
> 
> You should just follow checkpatch rules here.  If you don't, someone
> else will just send a patch to make it checkpatch compliant.  One thing
> you could do is at the start of the loop do:
> 
> 	struct iio_dev_attr *attr = &st->iio_attr[cnt];
> 
> Then it becomes:
> 
> 	if (!attr->dev_attr.attr.name)
> 
> It's slightly more readable that way.  Keep in mind that we increment
> cnt++ in the middle of the loop so you'll have to update attr as well.
>
My understanding was that CHECK: notes are not strict rules but
those that may vary from case to case - anyway you solution
sounds reasonable and in any case better than:

       if (!st->iio_attr[cnt].dev_attr.attr.name)

which just looked bad to me.

thx!
hofrat