Re: [PATCH v3 06/11] iio: core: merge buffer/ & scan_elements/ attributes

[Andy Shevchenko <andy.shevchenko@gmail.com>]

On Thu, Feb 4, 2021 at 3:41 PM Alexandru Ardelean
<ardeleanalex@gmail.com> wrote:
> On Wed, Feb 3, 2021 at 12:04 PM Andy Shevchenko
> <andy.shevchenko@gmail.com> wrote:
> > On Mon, Feb 1, 2021 at 5:28 PM Alexandru Ardelean
> > <alexandru.ardelean@analog.com> wrote:

...

> > > +       group->attrs = kcalloc(buffer_attrcount + 1,
> > > +                              sizeof(struct attribute *), GFP_KERNEL);
> > > +       if (!group->attrs)
> > > +               return -ENOMEM;
> > > +
> > > +       memcpy(group->attrs, buffer_attrs,
> > > +              buffer_attrcount * sizeof(struct attribute *));
> >
> > kmemdup() ?
> > Perhaps introduce kmemdup_array().
>
> doesn't add much benefit from what i can tell;
> and it complicates things with the fact that we need to add the extra
> null terminator element;
> [1] if we kmemdup(buffer_attrcount + 1) , the copy an extra element we
> don't need, which needs to be null-ed

Ah, I see now. Thanks for pointing it out!


> > > +       group->attrs = kcalloc(scan_el_attrcount + 1,
> > > +                              sizeof(struct attribute *), GFP_KERNEL);
> > > +       if (!group->attrs) {
> > > +               ret = -ENOMEM;
> > > +               goto error_free_buffer_attrs;
> > > +       }
> > > +
> > > +       memcpy(group->attrs, &buffer_attrs[buffer_attrcount],
> > > +              scan_el_attrcount * sizeof(struct attribute *));
> >
> > Ditto.
>
> continuing from [1]
> here it may be worse, because kmemdup() would copy 1 element from
> undefined memory;


-- 
With Best Regards,
Andy Shevchenko