Re: [PATCH tracing/kprobes 4/7] tracing/kprobes: Add event profiling support

[Frederic Weisbecker <fweisbec@gmail.com>]

On Fri, Sep 11, 2009 at 03:30:24PM -0400, Masami Hiramatsu wrote:
>> Note that the end-result must be u64 aligned for perf ring buffer.
>> And this is a bit tricky.
>> What is inserted in the perf ring buffer is:
>>
>> raw_trace + (u32)raw_trace_size
>>
>> So we must ensure that sizeof(raw_trace) + sizeof(u32)
>> is well u64 aligned.
>>
>> We don't insert the trace_size ourself though, this is done
>> from kernel/perf_counter.c
>>
>> But we need to handle the size of the size (sorry) in the final
>> alignment.
>> To sum-up: sizeof(raw_trace) doesn't need (shouldn't) to be u64
>> aligned but sizeof(raw_trace) + sizeof(u32) must be.
>>
>> Given this aligned size, we then substract it by sizeof(u32)
>> to have the needed size of the raw entry.
>>
>> This result gives you the size of char raw_data[], which
>> is also the same size passed in perf_tpcounter_event().
>>
>> See?
>
> Ah, I see. So the size to write to perf_tpcounter_event must be
> '(a multiple number of sizeof(u64)) - sizeof(u32)', right?


Exactly.
To simplify I guess the raw events just needs to be u32 aligned :)


> (Hmm, why would not perf_counter align data by itself? :)


Because that would require it to copy the data into a seperate
u64 aligned buffer.



>>
>> That's why we have this in trace/ftrace.h:
>>
>> __data_size = "the real entry data size"
>> __entry_size = ALIGN(__data_size + sizeof(*entry) + sizeof(u32), sizeof(u64));
>> __entry_size -= sizeof(u32);
>>
>> do {
>> 		char raw_data[__entry_size];
>> 		...
>> 		perf_tpcounter_event(event_call->id, __addr, __count, entry,
>> 			     __entry_size);
>> 		...
>> } while (0);
>
> Ok, I'll do that.


Thanks!