Re: [PATCH v10 12/15] platform/x86: dell-smbios: Add filtering support

["Pali Rohár" <pali.rohar@gmail.com>]

On Friday 05 January 2018 14:32:54 Mario.Limonciello@dell.com wrote:
> 
> 
> > -----Original Message-----
> > From: platform-driver-x86-owner@vger.kernel.org [mailto:platform-driver-x86-
> > owner@vger.kernel.org] On Behalf Of Pali Rohár
> > Sent: Friday, January 5, 2018 5:13 AM
> > To: Limonciello, Mario <Mario_Limonciello@Dell.com>
> > Cc: dvhart@infradead.org; Andy Shevchenko <andy.shevchenko@gmail.com>;
> > LKML <linux-kernel@vger.kernel.org>; platform-driver-x86@vger.kernel.org; Andy
> > Lutomirski <luto@kernel.org>; quasisec@google.com; rjw@rjwysocki.net;
> > mjg59@google.com; hch@lst.de; Greg KH <greg@kroah.com>; Alan Cox
> > <gnomes@lxorguk.ukuu.org.uk>
> > Subject: Re: [PATCH v10 12/15] platform/x86: dell-smbios: Add filtering support
> > 
> > I know that this patch is already applied and merged, but I spotted this
> > problem:
> > 
> > On Thursday 19 October 2017 12:50:15 Mario Limonciello wrote:
> > > +/* calls that are explicitly blacklisted */
> > > +static struct smbios_call call_blacklist[] = {
> > > +	{0x0000, 01, 07}, /* manufacturing use */
> > > +	{0x0000, 06, 05}, /* manufacturing use */
> > > +	{0x0000, 11, 03}, /* write once */
> > > +	{0x0000, 11, 07}, /* write once */
> > 
> > Numbers prefixed by zero means that they are in octal notation, right?
> Is that how the kernel interprets an integer prefix by zero?

No, this is how C language define it. See e.g. C11 standard, section
6.4.4.1 Integer constants:

decimal-constant:
	nonzero-digit
	decimal-constant digit

octal-constant:
	0
	octal-constant octal-digit

So in C decimal number cannot start with digit zero.

I think the place where octal numbers are used are in permissions (0777)

> I prefixed by zero for readability, they're supposed to be decimal.
> 
> > This can lead to misunderstanding, confusion or problems in future...
> > 
> > Can we have all numbers either in hexadecimal or decimal notation?
> 
> Could you elaborate more why this is problematic the way it is?

Currently it is not problem as 7 is same number in octal (07) and
decimal (7). representation. But e.g. octal 077 is 63 in decimal.

> Are you meaning you would rather see this?
> 	{0x0000, 1, 7}, /* manufacturing use */
> 	{0x0000, 6, 5}, /* manufacturing use */
> 	{0x0000, 11, 3}, /* write once */
> 	{0x0000, 11, 7}, /* write once */

Yes, this is better. If you need to achieve alignment then use spaces.
Really, not leading zeros.

> That seems less readable to me but should interpret the same way.

Example:

{0x000, 077, 7},
{0x000, 007, 7},

is **not** same as

{0x000,  77, 7},
{0x000,   7, 7},

As first number in first section is (decimal) 63, not (decimal) 77.

> Perhaps it would be better if you submit a patch with what is clearer to
> you.
> 
> > 
> > > +	{0x0000, 11, 11}, /* write once */
> > > +	{0x0000, 19, -1}, /* diagnostics */
> > > +	/* handled by kernel: dell-laptop */
> > > +	{0x0000, CLASS_INFO, SELECT_RFKILL},
> > > +	{0x0000, CLASS_KBD_BACKLIGHT, SELECT_KBD_BACKLIGHT},
> > > +};
> > 
> > --
> > Pali Rohár
> > pali.rohar@gmail.com

-- 
Pali Rohár
pali.rohar@gmail.com