Re: [PATCH 15/15] ptp: PTP_1588_CLOCK_PCH depends on x86

[Ben Hutchings <ben@decadent.org.uk>]

On Tue, May 14, 2013 at 04:20:09PM +0200, Jiri Slaby wrote:
> On 05/07/2013 09:29 PM, Richard Cochran wrote:
> > On Tue, May 07, 2013 at 04:18:23PM +0200, Jiri Slaby wrote:
> >> From: Jeff Mahoney <jeffm@suse.com>
> >>
> >> The PCH EG20T is only compatible with Intel Atom processors so it
> >> should depend on x86.
> > 
> > This patch has been submitted before,
> > 
> >    https://patchwork.kernel.org/patch/2069071/
> > 
> > and at that time the reaction was that it is good to have drivers
> > cross-compiled, if only for code quality reasons.
> 
> Hmm, then it depends whether the kernel is for users or for developers.
> I, as a user, do not really want to compile drivers with allmodconfig
> which I have no way to load/use.
> 
> And allmodconfig is basically what we, users (SUSE distributors in this
> case) do. So having this driver being built makes our life harder (in
> the meaning we have to have specific rules about disabling unwanted
> drivers from configs).

Wearing my Debian hat, I agree with this.  Perhaps we could define a
CONFIG_BUILD_TEST symbol for people who want to do that, and then
make drivers for hardware that's only found in x86 systems (for
example) depend on X86 || BUILD_TEST.

Ben.

> And, developers should have a testbed where they build 32bit configs.
> And we actually have that for -next AFAIU.

-- 
Ben Hutchings
We get into the habit of living before acquiring the habit of thinking.
                                                              - Albert Camus