Re: [PATCH v3 3/8] x86, cpu: provide a function topology_num_packages to enumerate #packages

From: Thomas Gleixner
Date: Sat Sep 10 2016 - 09:31:33 EST

B1;2802;0cOn Thu, 8 Sep 2016, Srinivas Pandruvada wrote:

> From: Tim Chen <tim.c.chen@xxxxxxxxxxxxxxx>

$subject: x86, cpu: provide a function topology_num_packages to enumerate #packages

- we switched to the prefix scheme 'x86/subsys'. Please use this.

- this is not related to x86/cpu. x86/topology is the proper prefix.

- Sentence after ':' starts with an uppercase letter.

- please make the subject line short and descriptive.

x86/topology: Provide topology_num_packages()

is completely sufficient, because it's entirely clear that it is a
function and the function name is self explaining.

> We compute the the number of active packages during boot and
> topology update.

We? We do not do anything..... and how is that information useful for the

> Provide a function to export this info for functions that need this
> topology info.

Well, it's obvious that a new function is going to be used by something
which needs it.

In changelogs/comments there is only one thing worse than superflous
informatioin: wrong information.

If you have nothing to say, then omit it instead of forcing the reader to
parse incoherent blurbs for nothing.

> int topology_update_package_map(unsigned int apicid, unsigned int cpu);
> +extern int topology_num_packages(void);
> extern int topology_phys_to_logical_pkg(unsigned int pkg);
> #else
> #define topology_max_packages() (1)

stub function for the !SMP case is missing....