Hi Stephen and Rafael, Sorry, the attachment patch can fix the build failure, which is caused by unset CONFIG_RUNTIME_PM. thanks, -- Ming Lei On Wed, Sep 28, 2011 at 12:51 PM, Stephen Rothwell wrote: > Hi Rafael, > > After merging the pm tree, today's linux-next build (powerpc > ppc64_defconfig) failed like this: > > In file included from include/trace/ftrace.h:567:0, >                 from include/trace/define_trace.h:96, >                 from include/trace/events/rpm.h:99, >                 from kernel/trace/rpm-traces.c:15: > include/trace/events/rpm.h: In function 'ftrace_raw_event_rpm_internal': > include/trace/events/rpm.h:16:1: error: 'struct dev_pm_info' has no member named 'usage_count' > include/trace/events/rpm.h:16:1: error: 'struct dev_pm_info' has no member named 'disable_depth' > include/trace/events/rpm.h:16:1: error: 'struct dev_pm_info' has no member named 'runtime_auto' > include/trace/events/rpm.h:16:1: error: 'struct dev_pm_info' has no member named 'request_pending' > include/trace/events/rpm.h:16:1: error: 'struct dev_pm_info' has no member named 'irq_safe' > include/trace/events/rpm.h:16:1: error: 'struct dev_pm_info' has no member named 'child_count' > In file included from include/trace/ftrace.h:771:0, >                 from include/trace/define_trace.h:96, >                 from include/trace/events/rpm.h:99, >                 from kernel/trace/rpm-traces.c:15: > include/trace/events/rpm.h: In function 'perf_trace_rpm_internal': > include/trace/events/rpm.h:16:1: error: 'struct dev_pm_info' has no member named 'usage_count' > include/trace/events/rpm.h:16:1: error: 'struct dev_pm_info' has no member named 'disable_depth' > include/trace/events/rpm.h:16:1: error: 'struct dev_pm_info' has no member named 'runtime_auto' > include/trace/events/rpm.h:16:1: error: 'struct dev_pm_info' has no member named 'request_pending' > include/trace/events/rpm.h:16:1: error: 'struct dev_pm_info' has no member named 'irq_safe' > include/trace/events/rpm.h:16:1: error: 'struct dev_pm_info' has no member named 'child_count' > > Presumably caused by commit 53b615ccca56 ("PM / Runtime: Introduce trace > points for tracing rpm_* functions"). > > I have used the pm tree from next-20110927 for today. > -- > Cheers, > Stephen Rothwell                    sfr@canb.auug.org.au > http://www.canb.auug.org.au/~sfr/ >