Messages in this thread | | | From | "Rafael J. Wysocki" <> | Subject | Re: [PATCH v1 1/2] PM/runtime: introduce trace points for tracing rpm_* functions | Date | Tue, 27 Sep 2011 23:03:34 +0200 |
| |
On Tuesday, September 27, 2011, Steven Rostedt wrote: > On Tue, 2011-09-27 at 22:36 +0200, Rafael J. Wysocki wrote: > > > > TP_printk("%pS:%s ret=%d", (void *)__entry->ip, __get_str(name), > > > __entry->ret) > > > > > > try that. > > > > Well, that certainly will work, but is it the right fix? > > Hehe, yes! The pS means to take a pointer and return the string version > of the address. Like "schedule+0x3a". __entry->ip is the address of the > function that called it __THIS_IP__ but we want to convert that into a > string name. > > We use ip because both perf and trace-cmd should be smart enough to > parse it too. As they both store the kallsyms into the data file. > > The (void *) is used because pS wants a pointer, and we stored the > address as an unsigned long.
OK, good.
I've applied both patches (with the above fix folded into the first one) to linux-pm/pm-runtime (and merged into my linux-next branch), so they will be pushed for 3.2.
Thanks, Rafael
| |