Re: [PATCH v2 5/4] ftrace, workqueuetrace: display work name

[KOSAKI Motohiro]

> > ------------------------------
> > # tracer: nop
> > #
> > #           TASK-PID    CPU#    TIMESTAMP  FUNCTION
> > #              | |       |          |         |
> >           <idle>-0     [003]  1540.844960: workqueue_insertion: thread=events/3:30 func=e1000_watchdog_task+0x0/0x6ae [e1000e] work=&adapter->watchdog_task
> >           <idle>-0     [003]  1540.844964: workqueue_insertion: thread=events/3:30 func=vmstat_update+0x0/0x3f work=vmstat_work
> >            <...>-30    [003]  1540.844969: workqueue_handler_entry: thread=events/3:30 latency=0ms func=e1000_watchdog_task+0x0/0x6ae [e1000e] work=&adapter-
> > >watchdog_task
> >            <...>-30    [003]  1540.845003: workqueue_handler_exit: thread=events/3:30 func=e1000_watchdog_task+0x0/0x6ae [e1000e] work=&adapter->watchdog_tas
> > k
> >            <...>-30    [003]  1540.845004: workqueue_handler_entry: thread=events/3:30 latency=0ms func=vmstat_update+0x0/0x3f work=vmstat_work
> >            <...>-30    [003]  1540.845007: workqueue_handler_exit: thread=events/3:30 func=vmstat_update+0x0/0x3f work=vmstat_work
> > 
> > This patch increase kernel-size for work name string table.
> > But actually, lockdep already have work name string table. then if you already use lockdep,
> > you don't get kernel-size increasing by this patch.
> 
> I don't know. Does it happen that frequently for a function to be used
> by several works?
> If so, this feature may be indeed interesting.

No, it's really rare.
in last discuttion, "per work" tracing is one of the key-word.
then, I add this.

but, if frequency is important, I can drop it.



--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/