linux-kernel - Re: [PATCH v1 1/4] iio: acpi: Fill data with NULL when iio_get_acpi_device_name_and

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: <ZySN9b9PsRmNlvAu@smile.fi.intel.com>
Date: Fri, 1 Nov 2024 10:14:45 +0200
From: Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
To: Jonathan Cameron <jic23@...nel.org>
Cc: Jonathan Cameron <Jonathan.Cameron@...wei.com>,
	Hans de Goede <hdegoede@...hat.com>, linux-iio@...r.kernel.org,
	linux-kernel@...r.kernel.org, Lars-Peter Clausen <lars@...afoo.de>,
	Dan Carpenter <dan.carpenter@...aro.org>
Subject: Re: [PATCH v1 1/4] iio: acpi: Fill data with NULL when
 iio_get_acpi_device_name_and_data() fails

On Thu, Oct 31, 2024 at 07:17:17PM +0000, Jonathan Cameron wrote:
> On Wed, 30 Oct 2024 18:02:17 +0200
> Andy Shevchenko <andriy.shevchenko@...ux.intel.com> wrote:
> 
> > Fill data with NULL, if provided, when returning NULL from
> > iio_get_acpi_device_name_and_data(). Note, the current users check
> > for name to be valid, except one case which was initially doing
> > like that and has to be fixed separately.
> > 
> > Reported-by: Dan Carpenter <dan.carpenter@...aro.org>
> > Closes: https://lore.kernel.org/r/54fac4a7-b601-40ce-8c00-d94807f5e214@stanley.mountain
> > Fixes: dc60de4eb0a4 ("iio: acpi: Add iio_get_acpi_device_name_and_data() helper function")
> > Signed-off-by: Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
> 
> This is counter intuitive as usual expectation would be no side effects on an
> error return.  How hard to fix all the users to initialize to NULL if they
> care about that?

v2 just has been sent, indeed the result looks much better, thanks for the review!

> There is still a chance we set it to NULL in here anyway, but that should only happen
> if we know the return is good in the sense of no error (missing ACPI etc)
> but not necessarily that dev_name() won't return NULL.
> 
> Don't think dev_name() can currently return NULL but 'maybe' it could...

dev_name() can't return NULL on the properly initialised device (either with
device_add(), or via dev_set_name() call). I do not think we can ever get to
the ->probe() without the above. Tell me, if I'm wrong.

-- 
With Best Regards,
Andy Shevchenko