On Saturday, May 18, 2013 07:45:45 AM Viresh Kumar wrote:
On 18 May 2013 05:10, Rafael J. Wysocki rjw@sisk.pl wrote:
Well, I actually meant "can you please verify your belief?". :-)
And that's because I'm wondering why the zero-day build testing doesn't catch this problem. Apparently, it doesn't build .configs with cpufreq governors configured as modules, although I believe it does test "make allmodconfig" for a couple of architectures at least. What gives?
My assumption was wrong. Actually cpufreq_governor.c is never compiled as module, but cpufreq_ondemand is...
And this routine isn't used from cpufreq_ondemand but cpufreq_governor..
But we were lucky that we didn't get a error here and EXPORT_SYMBOL is still required :)
Although not necessarily 3.10 material I suppose?
Rafael