| Message ID | 20240221164112.846409-1-vilasbhat@google.com (mailing list archive) |
|---|---|
| State | Superseded, archived |
| Headers | show |
| Series | [v1] PM: runtime: add tracepoint for runtime_status changes | expand |
On Wed, 21 Feb 2024 16:41:10 +0000 Vilas Bhat <vilasbhat@google.com> wrote: > diff --git a/include/trace/events/rpm.h b/include/trace/events/rpm.h > index 3c716214dab1..f1dc4e95dbce 100644 > --- a/include/trace/events/rpm.h > +++ b/include/trace/events/rpm.h > @@ -101,6 +101,42 @@ TRACE_EVENT(rpm_return_int, > __entry->ret) > ); > > +#define RPM_STATUS_STRINGS \ > + { RPM_INVALID, "RPM_INVALID" }, \ > + { RPM_ACTIVE, "RPM_ACTIVE" }, \ > + { RPM_RESUMING, "RPM_RESUMING" }, \ > + { RPM_SUSPENDED, "RPM_SUSPENDED" }, \ > + { RPM_SUSPENDING, "RPM_SUSPENDING" } > + > +/* > + * ftrace's __print_symbolic requires that all enum values be wrapped in the > + * TRACE_DEFINE_ENUM macro so that the enum value can be encoded in the ftrace > + * ring buffer. > + */ > +TRACE_DEFINE_ENUM(RPM_INVALID); > +TRACE_DEFINE_ENUM(RPM_ACTIVE); > +TRACE_DEFINE_ENUM(RPM_RESUMING); > +TRACE_DEFINE_ENUM(RPM_SUSPENDED); > +TRACE_DEFINE_ENUM(RPM_SUSPENDING); You could do what everyone else does: #define RPM_STATUS_STRINGS \ EM( RPM_INVALID, "RPM_INVALID" ) \ EM( RPM_ACTIVE, "RPM_ACTIVE" ) \ EM( RPM_RESUMING, "RPM_RESUMING" ) \ EM( RPM_SUSPENDED, "RPM_SUSPENDED" ) \ EMe( RPM_SUSPENDING, "RPM_SUSPENDING" ) #undef EM #undef EMe #define EM(a, b) TRACE_DEFINE_ENUM(a); #define EMe(a, b) TRACE_DEFINE_ENUM(a); RPM_STATUS_STRINGS #undef EM #undef EMe #define EM(a, b) { a, b }, #define EMe(a, b) { a, b } > + > +TRACE_EVENT(rpm_status, > + TP_PROTO(struct device *dev, enum rpm_status status), > + TP_ARGS(dev, status), > + > + TP_STRUCT__entry( > + __string(name, dev_name(dev)) > + __field(int, status) > + ), > + > + TP_fast_assign( > + __assign_str(name, dev_name(dev)); > + __entry->status = status; > + ), > + > + TP_printk("%s status=%s", __get_str(name), > + __print_symbolic(__entry->status, RPM_STATUS_STRINGS)) This will be what you want with that last redefine of EM*() -- Steve > +); > + > #endif /* _TRACE_RUNTIME_POWER_H */ > > /* This part must be outside protection */
On Wed, Feb 21, 2024 at 9:12 AM Steven Rostedt <rostedt@goodmis.org> wrote: > > On Wed, 21 Feb 2024 16:41:10 +0000 > Vilas Bhat <vilasbhat@google.com> wrote: > > > > diff --git a/include/trace/events/rpm.h b/include/trace/events/rpm.h > > index 3c716214dab1..f1dc4e95dbce 100644 > > --- a/include/trace/events/rpm.h > > +++ b/include/trace/events/rpm.h > > @@ -101,6 +101,42 @@ TRACE_EVENT(rpm_return_int, > > __entry->ret) > > ); > > > > +#define RPM_STATUS_STRINGS \ > > + { RPM_INVALID, "RPM_INVALID" }, \ > > + { RPM_ACTIVE, "RPM_ACTIVE" }, \ > > + { RPM_RESUMING, "RPM_RESUMING" }, \ > > + { RPM_SUSPENDED, "RPM_SUSPENDED" }, \ > > + { RPM_SUSPENDING, "RPM_SUSPENDING" } > > + > > +/* > > + * ftrace's __print_symbolic requires that all enum values be wrapped in the > > + * TRACE_DEFINE_ENUM macro so that the enum value can be encoded in the ftrace > > + * ring buffer. > > + */ > > +TRACE_DEFINE_ENUM(RPM_INVALID); > > +TRACE_DEFINE_ENUM(RPM_ACTIVE); > > +TRACE_DEFINE_ENUM(RPM_RESUMING); > > +TRACE_DEFINE_ENUM(RPM_SUSPENDED); > > +TRACE_DEFINE_ENUM(RPM_SUSPENDING); > > You could do what everyone else does: > > #define RPM_STATUS_STRINGS \ > EM( RPM_INVALID, "RPM_INVALID" ) \ > EM( RPM_ACTIVE, "RPM_ACTIVE" ) \ > EM( RPM_RESUMING, "RPM_RESUMING" ) \ > EM( RPM_SUSPENDED, "RPM_SUSPENDED" ) \ > EMe( RPM_SUSPENDING, "RPM_SUSPENDING" ) > > #undef EM > #undef EMe > #define EM(a, b) TRACE_DEFINE_ENUM(a); > #define EMe(a, b) TRACE_DEFINE_ENUM(a); > > RPM_STATUS_STRINGS > > #undef EM > #undef EMe > #define EM(a, b) { a, b }, > #define EMe(a, b) { a, b } > Thanks for the comment, Steven. I did notice both methods of defining enum values for tracepoints and chose this method because it felt clearer. Could you clarify on why the method you suggested is preferred? > > + > > +TRACE_EVENT(rpm_status, > > + TP_PROTO(struct device *dev, enum rpm_status status), > > + TP_ARGS(dev, status), > > + > > + TP_STRUCT__entry( > > + __string(name, dev_name(dev)) > > + __field(int, status) > > + ), > > + > > + TP_fast_assign( > > + __assign_str(name, dev_name(dev)); > > + __entry->status = status; > > + ), > > + > > + TP_printk("%s status=%s", __get_str(name), > > + __print_symbolic(__entry->status, RPM_STATUS_STRINGS)) > > This will be what you want with that last redefine of EM*() > > -- Steve > > > > +); > > + > > #endif /* _TRACE_RUNTIME_POWER_H */ > > > > /* This part must be outside protection */ >
On Wed, 21 Feb 2024 09:57:03 -0800 Vilas Bhat <vilasbhat@google.com> wrote: > > You could do what everyone else does: > > > > #define RPM_STATUS_STRINGS \ > > EM( RPM_INVALID, "RPM_INVALID" ) \ > > EM( RPM_ACTIVE, "RPM_ACTIVE" ) \ > > EM( RPM_RESUMING, "RPM_RESUMING" ) \ > > EM( RPM_SUSPENDED, "RPM_SUSPENDED" ) \ > > EMe( RPM_SUSPENDING, "RPM_SUSPENDING" ) > > > > #undef EM > > #undef EMe > > #define EM(a, b) TRACE_DEFINE_ENUM(a); > > #define EMe(a, b) TRACE_DEFINE_ENUM(a); > > > > RPM_STATUS_STRINGS > > > > #undef EM > > #undef EMe > > #define EM(a, b) { a, b }, > > #define EMe(a, b) { a, b } > > > > Thanks for the comment, Steven. I did notice both methods of defining > enum values for tracepoints and chose this method because it felt > clearer. Could you clarify on why the method you suggested is > preferred? > Sure. One big reason: It removes duplication. If you add another enum to the list, you only need to update it in one place. And it prevents the two from getting out of sync. -- Steve
On Wed, Feb 21, 2024 at 10:33 AM Steven Rostedt <rostedt@goodmis.org> wrote: > > On Wed, 21 Feb 2024 09:57:03 -0800 > Vilas Bhat <vilasbhat@google.com> wrote: > > > > You could do what everyone else does: > > > > > > #define RPM_STATUS_STRINGS \ > > > EM( RPM_INVALID, "RPM_INVALID" ) \ > > > EM( RPM_ACTIVE, "RPM_ACTIVE" ) \ > > > EM( RPM_RESUMING, "RPM_RESUMING" ) \ > > > EM( RPM_SUSPENDED, "RPM_SUSPENDED" ) \ > > > EMe( RPM_SUSPENDING, "RPM_SUSPENDING" ) > > > > > > #undef EM > > > #undef EMe > > > #define EM(a, b) TRACE_DEFINE_ENUM(a); > > > #define EMe(a, b) TRACE_DEFINE_ENUM(a); > > > > > > RPM_STATUS_STRINGS > > > > > > #undef EM > > > #undef EMe > > > #define EM(a, b) { a, b }, > > > #define EMe(a, b) { a, b } > > > > > > > Thanks for the comment, Steven. I did notice both methods of defining > > enum values for tracepoints and chose this method because it felt > > clearer. Could you clarify on why the method you suggested is > > preferred? > > > > Sure. One big reason: It removes duplication. > > If you add another enum to the list, you only need to update it in one > place. And it prevents the two from getting out of sync. > > -- Steve That makes sense! Thanks. I will revise the patch and send a new version soon.
diff --git a/drivers/base/power/runtime.c b/drivers/base/power/runtime.c index 05793c9fbb84..d10354847878 100644 --- a/drivers/base/power/runtime.c +++ b/drivers/base/power/runtime.c @@ -94,6 +94,7 @@ static void update_pm_runtime_accounting(struct device *dev) static void __update_runtime_status(struct device *dev, enum rpm_status status) { update_pm_runtime_accounting(dev); + trace_rpm_status(dev, status); dev->power.runtime_status = status; } diff --git a/include/trace/events/rpm.h b/include/trace/events/rpm.h index 3c716214dab1..f1dc4e95dbce 100644 --- a/include/trace/events/rpm.h +++ b/include/trace/events/rpm.h @@ -101,6 +101,42 @@ TRACE_EVENT(rpm_return_int, __entry->ret) ); +#define RPM_STATUS_STRINGS \ + { RPM_INVALID, "RPM_INVALID" }, \ + { RPM_ACTIVE, "RPM_ACTIVE" }, \ + { RPM_RESUMING, "RPM_RESUMING" }, \ + { RPM_SUSPENDED, "RPM_SUSPENDED" }, \ + { RPM_SUSPENDING, "RPM_SUSPENDING" } + +/* + * ftrace's __print_symbolic requires that all enum values be wrapped in the + * TRACE_DEFINE_ENUM macro so that the enum value can be encoded in the ftrace + * ring buffer. + */ +TRACE_DEFINE_ENUM(RPM_INVALID); +TRACE_DEFINE_ENUM(RPM_ACTIVE); +TRACE_DEFINE_ENUM(RPM_RESUMING); +TRACE_DEFINE_ENUM(RPM_SUSPENDED); +TRACE_DEFINE_ENUM(RPM_SUSPENDING); + +TRACE_EVENT(rpm_status, + TP_PROTO(struct device *dev, enum rpm_status status), + TP_ARGS(dev, status), + + TP_STRUCT__entry( + __string(name, dev_name(dev)) + __field(int, status) + ), + + TP_fast_assign( + __assign_str(name, dev_name(dev)); + __entry->status = status; + ), + + TP_printk("%s status=%s", __get_str(name), + __print_symbolic(__entry->status, RPM_STATUS_STRINGS)) +); + #endif /* _TRACE_RUNTIME_POWER_H */ /* This part must be outside protection */
Existing runtime PM ftrace events (`rpm_suspend`, `rpm_resume`, `rpm_return_int`) offer limited visibility into the exact timing of device runtime power state transitions, particularly when asynchronous operations are involved. When the `rpm_suspend` or `rpm_resume` functions are invoked with the `RPM_ASYNC` flag, a return value of 0 i.e., success merely indicates that the device power state request has been queued, not that the device has yet transitioned. A new ftrace event, `rpm_status`, is introduced. This event directly logs the `power.runtime_status` value of a device whenever it changes providing granular tracking of runtime power state transitions regardless of synchronous or asynchronous `rpm_suspend` / `rpm_resume` usage. Signed-off-by: Vilas Bhat <vilasbhat@google.com> --- drivers/base/power/runtime.c | 1 + include/trace/events/rpm.h | 36 ++++++++++++++++++++++++++++++++++++ 2 files changed, 37 insertions(+)